Early endothelial activation at the blood–nerve barrier defines a hallmark of ALS

Early endothelial activation at the blood–nerve barrier defines a hallmark of ALS | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results Early endothelial activation at the blood–nerve barrier defines a hallmark of ALS View ORCID Profile Ganesh Parameshwar Bhat , Santo Diprima , Cristian G Beccaria , Paola Podini , Aurora Badaloni , Alessia Motta , Ilaria Brambilla , Luis Martins , Daniela Boselli , Stefano de Pretis , Giovanni Nardo , View ORCID Profile Caterina Bendotti , View ORCID Profile Nilo Riva , View ORCID Profile Matteo Iannacone , Angelo Quattrini , Dario Bonanomi doi: https://doi.org/10.1101/2025.08.31.673224 Ganesh Parameshwar Bhat 1 Division of Neuroscience, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Ganesh Parameshwar Bhat Santo Diprima 1 Division of Neuroscience, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy 2 Center for Omics Sciences, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Cristian G Beccaria 3 Division of Immunology, Transplantation and Infectious Diseases, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Paola Podini 4 Institute of Experimental Neurology, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Aurora Badaloni 1 Division of Neuroscience, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Alessia Motta 1 Division of Neuroscience, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Ilaria Brambilla 1 Division of Neuroscience, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Luis Martins 1 Division of Neuroscience, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Daniela Boselli 5 FRACTAL-Flow cytometry Resource Advanced Cytometry Technical Applications Laboratory, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Stefano de Pretis 2 Center for Omics Sciences, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Giovanni Nardo 6 Laboratory of Molecular Neurobiology and Preclinical Therapy, Department of Neuroscience, Istituto di Ricerche Farmacologiche Mario Negri IRCCS , 20156 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Caterina Bendotti 6 Laboratory of Molecular Neurobiology and Preclinical Therapy, Department of Neuroscience, Istituto di Ricerche Farmacologiche Mario Negri IRCCS , 20156 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Caterina Bendotti Nilo Riva 7 3rd Neurology Unit and Motor Neuron Diseases Center, Fondazione IRCCS Istituto Neurologico Carlo Besta , 20133 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Nilo Riva Matteo Iannacone 3 Division of Immunology, Transplantation and Infectious Diseases, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Matteo Iannacone Angelo Quattrini 4 Institute of Experimental Neurology, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site Dario Bonanomi 1 Division of Neuroscience, IRCCS San Raffaele Scientific Institute , 20132 Milan, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: bonanomi.dario{at}hsr.it Abstract Full Text Info/History Metrics Preview PDF ABSTRACT Vascular defects are common in Amyotrophic Lateral Sclerosis (ALS). The prevailing view is that breakdown of the blood–brain and blood–spinal cord barriers contributes to neurodegeneration. Here, we reveal that selective and early vulnerability of peripheral nerve endothelium —manifested as endothelial cell activation and interlinked blood–nerve barrier dysfunction— constitutes a core feature of ALS pathogenesis, arising before vascular alterations in the central nervous system (CNS) and motor neuron pathology. Vascular changes in ALS patients have been largely studied in postmortem samples, limiting insight into their onset, causes, and pathogenic role. Surveying diagnostic motor nerve samples from ALS patients across clinical stages, we observed endothelial damage that preceded axonal loss and demyelination, marking vascular dysfunction as an early disease event. Similar ultrastructural abnormalities were detected in pre-symptomatic ALS mouse models ( SOD1 and TARDBP mutants). Notably, endothelial cells became dysfunctional even when not carrying mutant TDP-43, indicating they respond to non-cell autonomous disease signals. Transcriptional, histological, and functional analyses revealed that these alterations were largely confined to peripheral nerves, while spinal cord vessels exhibited delayed and more focal changes. Single-cell sequencing identified ALS-susceptible endothelial cell subsets prone to a pro-inflammatory phenotype and impaired blood–nerve barrier function, increasing permeability via the transcellular route. These changes coincided with reactivity of nerve-resident macrophages and neutrophil infiltration. Neutrophil depletion attenuated endothelial activation and barrier leakiness, mitigating axonopathy in ALS mice. Our work unmasks the greater susceptibility of the peripheral nerve vasculature in ALS relative to the CNS. The early activation of peripheral nerve endothelium, combined with its potential reversibility, identifies a therapeutic window and suggests strategies for targeting the vascular–immune axis to protect the motor system. INTRODUCTION The architecture, composition, and molecular properties of vascular beds reflect the structural and functional requirements of the surrounding tissue 1 . Endothelial cells (ECs) that line these vessels embody this diversity and specialization 2 , 3 . Far from being an inert interface, the endothelium is metabolically active and plastic. Beyond their circulatory role in delivering oxygen and nutrients, ECs —together with associated cells in tissue-specific vascular niches— maintain organ homeostasis, contributing to vascular regulation, immune surveillance, and tissue repair 4 – 6 . At the same time, ECs mount maladaptive responses to pathological stimuli that exacerbate disease 7 . This specialization and adaptive capacity also shape how ECs respond to pathophysiological stress 7 . These observations raise the question of whether the mechanisms that confer EC specialization and plasticity may also render them selectively susceptible to damage, generating focal patterns of vulnerability. A prototypical example of highly specialized endothelium is found in the vascular barriers of the CNS — the blood-brain barrier (BBB) and blood-spinal cord barrier (BSCB). CNS ECs acquire a barrier phenotype characterized by tight junctions, low rates of transcellular transport, restricted leukocyte adhesion molecule expression, and specialized transporter systems, all of which are essential for maintaining CNS homeostasis 8 – 10 . These barriers are disrupted in neurological disorders, including ALS, where BBB and BSCB defects are associated with neuroinflammation and motor neuron dysfunction 11 – 13 . In ALS patients and mouse models, BSCB abnormalities include reduced tight junction protein expression, decreased pericyte and astrocytic coverage, and activation of perivascular fibroblasts 14 – 23 . Increased levels of serum proteins have been detected in the cerebrospinal fluid of ALS patients, indicating BBB/BSCB dysfunction 24 – 30 , although these changes do not clearly align with disease aggressiveness 31 or the spatial pattern of motor neuron loss 32 . In ALS rodent models, BSCB impairment has been observed prior to motor neuron degeneration 19 , 20 , 22 , 33 – 35 , whereas other studies reported an intact barrier to systemically administered IgG 36 . While the relationship between CNS barrier leakage and neurodegeneration remains to be resolved, the expression of pathogenic gene variants in ECs is sufficient to trigger vascular dysfunction —including BBB/BSCB alterations 37 – 39 — and motor neuron degeneration 40 via non-cell autonomous disease mechanisms 41 . Although ALS research has primarily focused on CNS vascular damage and barrier dysfunction, spinal motor neurons —the main disease targets— extend most of their volume into the peripheral nervous system (PNS), where motor axons travel long distances within peripheral nerves to innervate skeletal muscles and autonomic ganglia 42 . This unique positioning at the CNS–PNS interface, together with the considerable heterogeneity in ALS onset and progression 43 , 44 , raises questions about where the disease initiates and how it propagates 45 , 46 . Early pathogenic events, including direct axonal damage and a “dying-back” phenomenon from the PNS, have been proposed 47 – 49 and likely act in concert with “dying-forward” mechanisms of disease spreading from the CNS 45 . Supporting PNS involvement, transcriptomic analyses and histopathological assessment of diagnostic motor nerve biopsies from ALS patients have revealed early alterations, including accumulation of pathological phopho-TDP-43 aggregates in axon fibers and Schwann cells —the PNS glia— before neurodegeneration 50 , 51 . Schwann cells can trigger neuroinflammation via chemokine secretion 52 , while both nerve-resident macrophages and infiltrating immune cells appear to play context-dependent roles that may be either protective or deleterious 53 – 57 . Collectively, these observations underscore the critical role of the peripheral nerve microenvironment in ALS pathogenesis. In the PNS, motor axons run alongside capillaries within the endoneurial compartment —the innermost layer of the nerve— whose ECs form the blood–nerve barrier (BNB), analogous to CNS barriers but comparatively less well studied 58 , 59 . The BNB shares key features with CNS barriers, including endothelial tight junctions, low rates of transcellular transport, and pericyte coverage 60 , but, unlike CNS barriers, it lacks astrocytic endfeet —critical components of neurovascular units 61 . The baseline permeability of the BNB is higher than that of the BBB 62 , 63 , and its ECs exhibit enhanced angiogenic capacity, supporting regeneration after injury 64 , 65 —a distinctive feature of PNS nerves that is absent in the mammalian CNS. The importance of the BNB for normal nerve physiology is underscored by the fact that its alteration is associated with several peripheral neuropathies, including diabetic and inflammatory neuropathies 58 , 59 . Although BNB changes in ALS remain poorly explored, studies in SOD1 mutant rat models have shown endoneurial edema and macrophage activation in motor nerves, indicating early BNB impairment that precedes axonal degeneration 66 . We show that peripheral nerves harbor a vulnerable vasculature in ALS. Endothelial activation —defined by an inflammatory and procoagulant phenotype 7 — coincided with BNB disruption, driven primarily by aberrant transcellular permeability in motor nerves before CNS vascular pathology and neurodegeneration. Venous capillaries were intrinsically more susceptible, displaying an elevated inflammatory tone that distinguished them from CNS counterparts. ECs became dysfunctional even in the absence of ALS-linked mutations, indicating that they responded to non-cell-autonomous disease signals. These vascular changes were accompanied by remodeling of nerve-resident macrophages and neutrophil infiltration, which further compromised barrier integrity. Collectively, these findings identify the peripheral nerve endothelium as an active sensor and amplifier of ALS pathology, revealing a previously underappreciated site of vulnerability in motor neuron disease. RESULTS Early microvascular changes in peripheral nerves of ALS patients and mouse models Retrospective transmission electron microscopy (TEM) analysis of motor nerve ( obturator n. ) biopsies collected to assist in diagnosis of motor neuron disease revealed a high incidence of microvascular abnormalities in clinically confirmed cases of sporadic ALS (n=12) ( Fig. 1a, i ). Nerve samples were categorized from “normal” to “severe” based on the extent of axonal loss, assessed in semithin sections ( Fig. 1d-h ). Biopsies from non-ALS neuropathy or myopathy patients, without evidence of demyelination or axonal pathology, served as controls (n=3). In ALS nerves, ECs lining endoneurial blood vessels ( Fig. 1c ) exhibited abnormal accumulation of pinocytotic vesicles, tight junctions (TJ) alterations, marked hypertrophy leading to luminal narrowing, central microfilament bundles resembling atherosclerotic lesions 67 , detachment from the basal lamina, and hyperplasia ( Fig. 1d’-h’, i and Suppl. Fig. 1a-l ). While the overall severity of endothelial damage correlated with axonal pathology, the aberrant increase in cytoplasmic vesicles —indicative of enhanced transcytosis (transcellular transport)— was already evident in histologically “normal” ALS cases, prior to significant axonal loss ( Fig. 1e, e’, f, f’, i, j ). Conversely, TJ opening —expected to increase paracellular permeability— along with EC hyperplasia and basal lamina detachment, were observed in cases with more extensive axonal loss ( Fig. 1h, h’, i and Suppl. Fig. 1j-k ). Download figure Open in new tab Supplementary Fig. 1: Ultrastructural endothelial defects in patient motor nerve biopsies. TEM images of control (myopathy) obturator nerve biopsy ( a ) and sporadic ALS cases of varying severity ( b-g’ , i-l ). Accumulation of pinocytic vesicles (closed arrowheads) is visible in all ALS samples. TJs (arrows) are partially open in severe cases ( j , k ). Hypertrophy is evident in d , d’ , e , f (double open arrowheads span the endothelium). EC hyperplasia is visible in g , l (nu, nuclei). Microfilament bundles (mf) are present in d’ , j , k . Boxed areas in a , b , c , l are magnified in Fig. 1d’, e’, f’, h’ , respectively. Boxed areas in d , g, i are magnified in d’ , g’ i’ . h , Toluidine Blue-stained semithin transverse sections of diagnostic obturator nerve biopsies from severe ALS case with marked axon loss (asterisks). This sporadic patient carried the SOD1 p.G148S mutation while genetic changes were not identified for the other cases. Scale bars: a, 5μm; b, c, d, g, i, l, 2μm; g’, i’, 1μm; d’, e, f, j, k, 500nm; h, 20μm Download figure Open in new tab Fig. 1: Microvascular alterations in peripheral nerves of ALS patients and mouse models. a , Diagnostic workflow for ALS based on histopathological analysis of motor nerve ( obturator n .) biopsies. b , Schematic of quadriceps motor nerve and saphenous sensory nerve in mouse hindlimb. c , Schematic of peripheral nerve showing blood-nerve-barrier localization in endoneurial microvessels and permeable vasculature of the epineurial compartment. d-h , Toluidine Blue-stained semithin transverse sections of diagnostic obturator nerve biopsies from non-ALS (control, d ) and sporadic ALS cases ( e-h ) with varying degrees of fiber loss (asterisks), reflecting different stages of disease progression. d’-h’ , Corresponding TEM images of endoneurial ECs. Pinocytotic vesicles (closed arrowheads) accumulate in ALS nerves. TJ (arrows) are compromised in severe cases ( h’ ). EC hypertrophy, open arrowhead, e’ . Microfilaments (mf; g’ ). Multiple EC nuclei (nu; h’ ) indicate hyperplasia. i , Heatmap of severity score for axonopathy (fiber loss, demyelination) and morphological alterations in endoneurial ECs from obturator nerve biopsies. Controls are non-ALS neuropathy (np) and myopathy (myo) cases with normal nerve fiber morphology. Case numbers are reported (see Suppl. Fig. 1 ). j , Vesicle density in human motor nerve ECs. ALS “normal” cases are included in the mild group; “moderate" ALS cases in the severe group. Mean ± SEM. Sample size and statistics for all figures are reported in Methods. k , l , Vesicle density in motor nerve ECs from SOD1 G93A ( k ) and TDP-43 Q331K ( l ) mice at progressive disease stages. Mean ± SEM m , Timeline of disease progression in SOD1 G93A and TDP-43 Q331K mouse models. n-r , TEM images of endoneurial ECs from control ( n ), SOD1 G93A ( o , p , q ) and TDP43 Q331K ( r ) motor nerves at the indicated stages. ECs in mutants show increased cytoplasmic vesicles (closed arrowheads, o-r ), hypertrophy (open arrowheads, o , r ), hyperplasia (nu, nuclei, p ). TJs (arrows) are maintained through the early-symptomatic stage ( q , r ). Scale bars: d-h, 20μm; d’-h’, 500nm; n-r, 500nm. A spectrum of EC alterations similar to those observed in human ALS nerve biopsies was identified by TEM in the microvasculature of motor nerves ( quadriceps n. ) of two transgenic mouse models of ALS, SOD1 G93A and TDP-43 Q331K ( Fig. 1b, n-r and Suppl. Fig. 2i-n, u-w ). These models develop progressive motor neuron disease with distinct temporal patterns, as defined by histological and behavioral assays ( Fig. 1m , Suppl. Fig. 2g, h and Suppl. Fig. 3o-q ) 68 – 70 . Axonal loss in SOD1 G93A motor nerves began approximately at 90 days of age (postnatal, P90; onset stage) and worsened by P120 (early-symptomatic) ( Suppl. Fig. 2a-f ) 53 . In TDP-43 Q331K mice, which moderately overexpress human mutant TDP-43 without mislocalization and potentially mirror an early stage of ALS 68 , 70 , only a few degenerating axons were observed by the onset of motor symptoms at P120 ( Suppl. Fig. 2o-t ). As noted in patient nerve biopsies, endoneurial ECs in both SOD1 G93A and TDP-43 Q331K motor nerves showed a conspicuous increase in cytoplasmic vesicle density by the onset stage ( Fig. 1k, l, o-r and Suppl. Fig. 2m’, n, u’, v’ ), while TJs remained relatively preserved through the early-symptomatic phase ( Fig. 1q, r and Suppl. Fig. 2m’, u’, v’ ). These defects were never detected in control motor nerves from non-transgenic littermates, nor in sensory nerves ( saphenous n. ) of mutant mice, supporting the regional specificity of ALS pathology in these models ( Fig. 1b, n and data not shown) 69 . Download figure Open in new tab Supplementary Fig. 2: Endothelial alterations precede axonal degeneration in SOD1 G93A and TDP-43 Q331K ALS mice. a-d , Semithin transverse sections of control ( a ) and SOD1 G93A ( b-d ) quadriceps motor nerves during disease progression. Arrowheads point to degenerating fibers. e , f , Axon density ( e ) and percentage of degenerating axons ( f ) assessed in semithin sections from control and SOD1 G93A motor nerves. Mean ± SEM. g , Rotarod motor performance of control and SOD1 G93A mice. Mean ± SEM; normalized to body weight, BDW). h , Body weight measurement in control and SOD1 G93A mice. Mean ± SEM. i-n , TEM images of endoneurial vessels in motor nerves of control ( i ) and SOD1 G93A ( j-n ) mice at the indicated stage. Boxed areas in i-l are magnified in Fig. 1n-q . Abundant cytoplasmic vesicles are visible in SOD1 G93A ECs (arrowheads in m’ , n ). TJs are preserved (arrows, m’ ). o-r , Semithin transverse sections of control ( o ) and TDP-43 Q331K ( p-r ) motor nerves during disease progression. (arrowhead in r points to degenerating fiber). s , t , Axon density ( s ) and percentage of degenerating axons ( t ) assessed in semithin sections from control and TDP-43 Q331K motor nerves. Mean ± SEM. Mild axonal loss occurs in the early-symptomatic phase (P180), with only a few actively degenerating fibers visible (compare to SOD1 G93A , f ). Increased inter-fiber distance in r is a sign of edema. u-w , TEM images of endoneurial vessels in motor nerves of TDP-43 Q331K mice at P120 (onset) and P180 (early-symptomatic). Boxed areas in u , v are magnified in u’ , v’ , respectively. Arrowheads indicate cytoplasmic vesicles. Arrows point to preserved TJs. Double open arrowheads mark hypertrophy ( v’ ). The boxed area in w is magnified in Fig. 1r . Scale bars: a-d, o-r, 20μm; i, j, k, l, m, u, v, w, 2μm; m’, n, u’, v’, 500nm Download figure Open in new tab Supplementary Fig. 3: Enhanced microvascular barrier defects in SOD1 G93A peripheral nerves compared to the spinal cord. a-b’ , Longitudinal views of sciatic nerves from the same control and SOD1 G93A mice at P135 as in Fig. 1k-l ’ (where spinal cords are shown). Sulfo-biotin leakage is widespread in nerves. Boxed areas in a , b are magnified in a’ , b’ . c-f , Spinal cords from control and SOD1 G93A mice at P60 ( d ), P90 ( e ), and P150 ( f ) following sulfo-biotin injection. The tracer remains confined to the vasculature until late disease stage, when focal leakage is observed (arrows, f ). Vessels are labeled with Podocalyxin. g-j , Control and SOD1 G93A spinal cords showing motor neurons in the ventral horn (NeuN) and reactive astrocytes (GFAP). Motor neuron loss and astrocytosis are conspicuous in the symptomatic phase (P150, j ). k-n , Control and SOD1 G93A spinal cords showing proliferation/activation of microglia (IBA1) in the ventral horn of mutant mice. DAPI, nuclei. o , Number of NeuN + motor neurons in spinal cord sections of control and SOD1 G93A mice at progressive stages, normalized to average control value. p , q , Quantification of astrocytes (GFAP + ) ( p ) and microglia (IBA1 + ) ( q ) in control and SOD1 G93A spinal cords. Immunoreactivity area normalized to average control value. r-u , Spinal cord ( r , s ) and sciatic nerve ( t , u ) vessels from P120 control and SOD1 G93A mice stained for CD31. Vessel dilation is frequent in SOD1 G93A nerves ( u , arrows) but rare in the spinal cord ( s, arrow). DAPI, nuclei. v , w , Vessel diameter in spinal cord ( v ) and sciatic nerve ( w ) of control and SOD1 G93A mice across disease stages, normalized to average control value. x , 3D rendering from two-photon imaging of sciatic nerve vasculature stained for CD31/Podocalyxin in P90 control and SOD1 G93A mice. Mutants show discrete architectural changes, including increased anastomotic branching (arrows). y , Cell type composition inferred from bulk RNA-seq data using CIBERSORTx confirms efficient EC purification. Color scale indicates cell abundance. Tabula Muris and sciatic nerve scRNA-seq datasets were used to deconvolve spinal cord and nerve EC transcriptomes, respectively (see Methods). z , GSEA identifies upregulated (red) and downregulated (blue) pathways in SOD1 G93A at P90 and P120. Scale bars: a, b, 200μm; a’, b’, 50μm; c-n, 100μm; r-u, 50μm; x, 250μm. These findings indicate that endothelial alterations in the peripheral nerve microvasculature arise at the pre-symptomatic stage, preceding axonal degeneration, and may represent a shared pathogenic feature of both familial and sporadic ALS, as observed across different transgenic mouse models and in patients with distinct clinical presentations. Heightened vulnerability of peripheral nerve endothelium in ALS The structural changes identified in the endothelium of human and mouse peripheral nerves point to a compromised blood-nerve barrier (BNB) ( Fig. 1c ). To assess vascular barrier integrity, we intravenously administered the permeability tracer sulfo-NHS-biotin 64 , 71 , 72 to SOD1 G93A mice at different disease stages and examined its distribution in sciatic nerves, which carry both motor and sensory fibers ( Fig. 2a-e ). In control mice, sulfo-biotin was confined to the lumen of endoneurial vessels, confirming an intact BNB ( Fig. 2b ). In contrast, focal leakage into the nerve parenchyma was already evident in SOD1 G93A mice at presymptomatic stage (P60) and worsened from onset (P90) onward ( Fig. 2c-f and Suppl. Fig. 3a-b ’). BNB dysfunction was limited to motor nerves, with no leakage observed in sensory nerves at any stage ( Fig. 2h-j ’), aligning with the pattern of ultrastructural EC alterations. Notably, in the spinal cord of SOD1 G93A mice, sulfo-biotin was retained in the microvasculature until the symptomatic phase (P120-140) ( Fig. 2g ), when only sporadic leaky hotspots became visible despite widespread astrogliosis, microglial activation, and motor neuron loss ( Fig. 2k, l’ and Suppl. Fig. 3c-q ). In parallel, SOD1 G93A sciatic nerves showed progressive capillary dilation and tortuosity, features associated with inflammation, hypoxia, and barrier permeability in other conditions 73 , 74 . By contrast, spinal cord capillary size and morphology remained largely unchanged, with only localized alterations observed at advanced stages ( Suppl. Fig. 3r-x ). Together, these findings demonstrate early BNB disruption, preceding blood-spinal cord barrier (BSCB) impairment 19 , 20 , 33 . Download figure Open in new tab Fig. 2: Blood-nerve barrier disruption and transcriptional dysregulation in the peripheral nerve endothelium of SOD1 G93A ALS mice. a , Vascular permeability assessed through intravenous injection of sulfo-NHS-biotin. b-e , Longitudinal views of sciatic nerves from control ( b ) and SOD1 G93A ( c-e ) mice injected with sulfo-biotin (550 Da). The tracer is retained in BNB-forming endoneurial vessels (CD31 + ) in controls ( b , arrow) but extravasates in mutants ( c-e ). Leakage is focal at pre-symptomatic stage (P60; c , arrow) and widespread from onset (P90) onward ( d , e ). Sulfo-biotin accumulates in the epineurium that contains permeable vessels ( b , arrowhead). f , g Extravascular sulfo-biotin signal in control and SOD1 G93A sciatic nerves ( f ) and spinal cord ( g ) at progressive stages, normalized to average control value. h-j’ , Vascular leakage in quadriceps motor nerve of mice injected with sulfo-biotin at P70 ( i , i’ ; open arrows). The tracer is retained in endoneurial vessels (CD31 + ) of controls ( h , h’ , closed arrows) and SOD1 G93A saphenous sensory nerve ( j , j’ + , closed arrows). Sulfo-biotin is shown separately in h’-j’ . Arrowheads point to epineurium. k-l’ , Spinal cord transverse sections from control ( k , k’ ) and SOD1 G93A ( l , l’ ) mice injected with sulfo-biotin at P135. Boxed areas in k , l are magnified in k’ , l’ . The tracer is confined to vessels (CD31 + ) in controls. Focal leakage is visible in SOD1 G93A ( l’ , arrow). DAPI labels nuclei ( k , l ). NeuN labels neurons ( k’ , l’ ). m , Schematic of bulk RNA-seq of ECs FACS-purified separately from sciatic nerves and spinal cord using anti-CD31 antibody. n , o , Volcano plots of EC genes upregulated (red) or downregulated (blue) (padj <0.1) in SOD1 G93A sciatic nerve ( n ) or spinal cord ( o ) vs. control at P90 and P120. Selected genes are labeled. p , q , GO enrichment from upregulated ( p ) and downregulated ( q ) genes (padj ≤0.05) in SOD1 G93A sciatic nerve vs. control at P90 and P120. Spinal cord DEGs are too few for GO analysis. r , Average expression heatmaps (z-score) of representative genes for deregulated biological processes. Scale bars: b-e, 50μm; h-j’, 200μm; k, l, 200μm; k’, l’, 50μm. We asked whether the enhanced susceptibility of the BNB in SOD1 G93A mice, compared to the BSCB, was reflected in distinct transcriptional changes in the endothelium. To address this, we isolated ECs from the sciatic nerve and spinal cord of SOD1 G93A mice and non-transgenic controls at P90 and P120 using fluorescence-activated cell sorting (FACS) with an anti-CD31 antibody, followed by bulk RNA-sequencing ( Fig. 2m ). The sciatic nerve epineurium was mechanically removed to enrich BNB-forming endoneurial ECs. After confirming EC population purity ( Suppl. Fig. 3y ), differential gene expression (DGE) analysis showed divergent transcriptional responses of sciatic nerve versus spinal cord ECs in SOD1 G93A mice. In the sciatic nerve, hundreds of genes were affected (both upregulated and downregulated) at P90, with further changes observed at P120 ( Fig. 2n ). In contrast, only a few genes were deregulated in the spinal cord at either time point ( Fig. 2o ). In nerve ECs, some differentially expressed genes (DEGs) changed progressively between P90 and P120, while others appeared only at a single time point, reflecting an evolving and partly stage-specific endothelial response. Gene Ontology (GO) annotation of nerve DEGs and Gene Set Enrichment Analysis (GSEA) pointed to impaired BNB function, already evident at P90, consistent with early vascular leakage in SOD1 G93A nerves ( Fig. 2p-r and Suppl. Fig. 3z ). This impairment was indicated by decreased expression of solute transporters ( Slco1c1 , Slc2a1 , Slc7a1 , Slc19a3 , Slc45a3 ), key components of Wnt and Notch pathways required for barrier maintenance ( Axin2 , Reck , Lef1 , Dkk3 , Sfrp5 , Dll4 , Hey1 , Hes1 ), as well as selected junction proteins ( Jcad , Tjp2 , Cldn5 ) and endothelial barrier modulators ( Bsg , Spock2 , Mfsd2a ) ( Fig. 2q, r ). Downregulation of vasoprotective factors ( Klf2 , Klf4 , Igfbp3 , Igfbp6 , Lrp1 , Plat ) indicated altered vascular homeostasis ( Fig. 2r ). Among metabolism-related DEGs, those involved in lipid biosynthesis, transport, and processing were especially affected. Decreased cholesterol availability was inferred from reduced expression of biosynthetic enzymes ( Hmgcs1 , Hmgcr , Sqle , Msmo1 ) and uptake receptors/transporters ( Ldlr , Abca2 , Pltp ), along with increased expression of efflux transporters ( Abcg1 , Scarb1 , Abca1 ) ( Fig. 2r ). A similar change in cholesterol homeostasis has been observed in SOD1 G93A motor neurons, highlighting a shared metabolic imbalance 75 . In parallel, GO analysis of upregulated DEGs revealed an inflammatory profile in SOD1 G93A nerve ECs, evident at P90 and more pronounced at P120 ( Fig. 2p, r ). This signature included pro-inflammatory cytokines/chemokines ( Il6 , Cxcl1 , Cxcl10 , Ccl2 ), receptors ( Il1r1 , Tacr1/NK1 , Tlr2 , Trl3 ), key signaling regulators –particularly within the TNFα-NFκB axis– and cell adhesion molecules ( Vcam1 , Icam1 , Sele ), which are expressed on activated endothelium and involved in immune cell invasion 76 ( Fig. 2r ). Conversely, downregulation of a type-I interferon signature suggested suppression of immunoregulatory mechanisms active in capillaries of healthy nerves, as well as in the brain and other organs 64 , 77 ( Fig. 2q, r ). In summary, peripheral nerve ECs in SOD1 G93A mice are more sensitive than those in the spinal cord, adopting a pro-inflammatory, activated state 76 , 78 with concomitant BNB loss before axonal or motor neuron degeneration. Dysregulated transcytosis drives blood-nerve barrier breakdown Several lines of evidence indicate that paracellular transport in the nerve endothelium remains relatively preserved during the early phase of ALS. Structural alterations in EC contacts were only evident in severe human cases (Fig.1h’, i), with only few junctional genes found downregulated in SOD1 G93A mice ( Fig. 2r ). Furthermore, the distribution and levels of the adherens junction protein VE-cadherin (CDH5) did not change in SOD1 G93A nerves ( Suppl. Fig. 4d-f ) and the TJ component ZO-1 decreased only in late-stage symptomatic mice (P140) ( Suppl. Fig. 4g-i ). In contrast, the increased density of cytoplasmic vesicles prior to disease onset ( Fig. 1i-l ) suggests that dysregulated transcellular transport may drive BNB impairment. To investigate this possibility, we immunostained SOD1 G93A sciatic nerves for PLVAP, which forms stomatal and fenestral diaphragms on ECs, and regulates trans-endothelial transport 79 , and Claudin-5 (CLDN5), a TJ component expressed by barrier-forming ECs. In control nerves, PLVAP was restricted to permeable vessels of the epineurium, which correspondingly lacked CLDN5. Conversely, BNB endoneurial vessels, in which both paracellular and transcellular permeability are suppressed, expressed CLDN5 but not PLVAP ( Fig. 3a ) 60 , 64 . These non-overlapping patterns of PLVAP and CLDN5 were drastically altered in SOD1 G93A nerves ( Fig. 3b ). At the pre-symptomatic stage (P60), ectopic expression of PLVAP was observed in endoneurial vessels that retained normal levels of CLDN5. PLVAP induction was initially focal and became progressively widespread at the symptomatic stage (P120) when CLDN5 expression began to decline ( Fig. 3c-g, j, k ). The aberrant upregulation of PLVAP in CLDN5 + ECs was exclusive to motor fiber-carrying nerves of SOD1 G93A mice and absent in pure sensory nerves ( Suppl. Fig. 4a-c ). Notably, PLVAP was never induced in spinal cord microvessels, despite a localized decrease in CLDN5 in the ventral horn by P120, suggesting TJ alterations ( Fig. 3h-i’, l, m ). These results point to distinct mechanisms underlying the response of PNS and CNS barriers to ALS pathology. Download figure Open in new tab Supplementary Fig. 4: Deregulated transcytosis at the blood-nerve barrier precedes endothelial junctional alterations. a-b’ , Endoneurial vessels in quadriceps (motor) ( a , a’ ) and saphenous (sensory) ( b , b’ ) nerves from P60 SOD1 G93A mice stained for PLVAP, CLDN5, and PODXL (marks all vessels). DAPI, nuclei. a’ , b’ show only PLVAP and CLDN5 staining. c , PLVAP signal in endoneurial vessels of P60 control and SOD1 G93A motor nerves, normalized to average control value. d , e , Endoneurial vessels in control ( d ) and SOD1 G93A ( e ) sciatic nerves at P120 stained for VE-cadherin (CDH5) and PODXL (marks all vessels). f , CDH5 signal in endoneurial vessels of control and SOD1 G93A sciatic nerves, normalized to average control value. g-h’’ , ZO-1 staining in endoneurial vessels of control ( g ) and SOD1 G93A ( h , h’ , h” ) sciatic nerves at progressive stages. Control is P90. PODXL, pan-vessel marker. ZO1 decreases in SOD1 G93A at advanced stage ( h” , arrow). i , ZO-1 signal in endoneurial vessels of control and SOD1 G93A sciatic nerves during disease progression, normalized to average control value. j-k” , Fibrinogen staining in endoneurial vessels of control ( j ) and SOD1 G93A ( k , k’ , k” ) sciatic nerves at progressive stages. Control is P90. CD31, pan-vessel marker. Fibrinogen extravasation is evident at P90 and increases at P140 (arrows). l , Extravascular fibrinogen signal in control and SOD1 G93A sciatic nerves at progressive stages, normalized to average control value. m , n , S aphenous (sensory, m ) and quadriceps (motor, n ) nerves from P90 SOD1 G93A mice injected systemically with HRP, detected by tyramide amplification. HRP is confined to vessels in sensory nerves but extravasates in motor nerves and is phagocytosed by macrophages ( n , arrowheads). DAPI, nuclei. o , Extravascular HRP-tyramide signal in control and SOD1 G93A motor nerves (P90), normalized to average control value. p-q , Spinal cord vessels from control ( p ) and SOD1 G93A ( q ) mice at P120 stained for MFSD2A, PLVAP and CD31. MFSD2A expression remains unchanged in mutants; PLVAP is not induced. r , MFSD2A signal in spinal cord vessels of control and SOD1 G93A mice at P120, normalized to average control value. s-u’ , Sciatic nerves from control ( s ) and SOD1 G93A at P90 ( t ) and P120 ( u ) stained for Caveolin-1 (CAV1), PLVAP and CD31. Dashed contours outline the endoneurium. Boxed areas in s , t , u are magnified in s’ , t’ , u’ , showing only CAV1 and CD31. Representative dashed lines spanning vessel width in s’ , t’ , u’ used to measure CAV1 signal profile in v . v , CAV1 signal intensity profiles. In control vessels, CAV1 shows two distinct peaks at the plasma membrane; in SOD1 G93A nerves, a diffuse profile indicates increased cytoplasmic distribution. w-x’ , Dorsal root ganglion (DRG) vessels from wild-type mice stained for PLVAP, CLDN5 and CD31 ( w ). Higher magnification views show PLVAP and CD31 ( x ), and CAV1 with DAPI (nuclei) ( x’ ). Representative dashed line spanning vessel width in x’ used to measure CAV1 signal profile in y . y , CAV1 signal intensity profiles in PLVAP + DRG vessels show a diffuse pattern, consistent with cytoplasmic localization. Scale bars: a-b’, 50μm; d, e, 50μm; g-h”, 50μm; j-k”, 50μm; m, n, 100μm; p, q, 100μm; s-u, 100μm; s’-u’, 25μm; w, 100μm; x, x’, 25μm. Download figure Open in new tab Fig. 3: Transcytosis impairment in the nerve endothelium but not spinal cord of SOD1 G93A ALS mice. a , b , Sciatic nerves from control ( a ) and SOD1 G93A mice at P90 stained for PLVAP and CLDN5. DAPI, nuclei. CLDN5 + endoneurial vessels are PLVAP − in controls ( a , arrow) but express PLVAP in mutants ( b , arrows). Permeable epineurial vessels express PLVAP in both genotypes (arrowheads). Dashed contours outline the endoneurium. Boxed areas in a and b are magnified in c and e , respectively. c-f’ , Ectopic expression of PLVAP in endoneurial vessels of SOD1 G93A sciatic nerves during disease course (P60, P90, P120) ( d-f’ , arrows), but not in controls ( c ). Podocalyxin (PODXL) marks all vessels; DAPI, nuclei. Lower panels ( c’-f’ ) show only PLVAP and Claudin-5 staining. Dashed contours outline the endoneurium. Arrowheads point to epineurial vessels. g , Schematic of progressive BNB impairment in SOD1 G93A nerves. Transcytosis increases at the presymptomatic stage, marked by focal PLVAP upregulation in endoneurial vessels, which transition into a hybrid PLVAP + /CLDN5 + state. At the symptomatic stage, PLVAP expression spreads and CLDN5 declines, indicating tight junction impairment. h-i’ , Spinal cords from control ( h , h’ ) and SOD1 G93A ( i , i’ ) mice at P120 stained for PLVAP and CLDN5. CD31 marks all vessels; DAPI, nuclei. Lower panels ( h’ , i’ ) show only PLVAP and CLDN5 staining. PLVAP is not upregulated in mutants, while focal tight junction loss is observed ( i , i’ , arrows). j , k , PLVAP ( j ) and CLDN5 ( k ) signal in endoneurial vessels of control and SOD1 G93A sciatic nerves during disease progression, normalized to average control value. l , m, PLVAP ( l ) and CLDN5 ( m ) signal in spinal cord vessels of control and SOD1 G93A mice at P120. n , o , Sciatic nerves from P90 control ( n ) of SOD1 G93A ( o ) mice injected systemically with HRP, detected by Tyramide amplification. DAPI, nuclei. HRP is confined to vessels in controls ( n , arrowheads) but extravasate in mutant nerves ( o , arrows). p , Extravascular HRP-tyramide signal in control and SOD1 G93A sciatic nerves (P90), normalized to average control value. q , r , Motor nerves from P120 control ( q ) and SOD1 G93A ( r ) mice injected systemically with BSA-FITC. The tracer extravasates and is phagocytosed by macrophages (IBA1 + ) in mutants ( r , arrowheads). CD31 marks vessels; DAPI, nuclei. s , Extravascular BSA-FITC signal in control and SOD1 G93A motor/sciatic nerves (P120), normalized to average control value. t , Western blotting for phoshoY14-CAV1, total CAV1 and CLDN5 in sciatic nerve lysates from control and SOD1 G93A mice at P90. Vinculin is a loading control. u-v’ , Endoneurial sciatic nerve vessels from P120 control ( u , u’ ) and SOD1 G93A ( v , v’ ) mice stained for MFSD2A, PLVAP and CD31. MFSD2A is shown separately in u’ , v’ . MFSD2A expression decreases in mutant vessels that upregulate PLVAP (arrows). A subset of vessels retain MFSD2A expression (arrowheads). w , MFSD2A signal in endoneurial vessels of P120 control and SOD1 G93A sciatic nerves, normalized to average control value. x , Scatter plot showing MFSD2A (x-axis) and PLVAP (y-axis) signal intensity in individual endoneurial vessels of control and SOD1 G93A sciatic nerve. y , qRT-PCR for MFSD2A in control and ALS obturator motor nerve biopsies. Mean (normalized to control) ± SEM. Scale bars: a, b, 500μm; c-f’, 50μm; h-i’, 100μm; n, o, 200μm; q, r, 100μm; u-v’, 100μm. Endoneurial ECs that induced PLVAP expression at an early disease stage exhibited abnormal transcytosis. This was monitored through intravenous injection of horseradish peroxidase (HRP-tyramide) and fluorescent-albumin (BSA-FITC), established markers of transcellular transport 60 , 80 – 82 . In control mice, the tracers were confined to the lumen of endoneurial vessels, reflecting a low transcytosis rate ( Fig. 3n, q ). By contrast, in SOD1 G93A sciatic and motor nerves —but not sensory nerves— both tracers diffused into the parenchyma and were subsequently phagocytosed by macrophages ( Fig. 3o, p, r, s and Suppl. Fig. 4m-o ). Leakage of plasma proteins in the endoneurium was further supported by accumulation of extravascular fibrinogen in SOD1 G93A nerves by onset stage ( Suppl. Fig. 4j-l ). A main route of endothelial transcytosis occurs via caveolae 83 , whose formation is enhanced by Src-dependent phosphorylation of caveolin-1 (CAV1), their key structural component 84 , 85 . Consistent with increased transcytosis at the BNB, the levels of both total and phosphorylated (pY14) CAV1 were elevated in SOD1 G93A sciatic nerves, while CLDN5 was unaffected ( Fig. 3t ). The subcellular distribution of CAV1 reflected BNB permeability changes: in control endoneurial ECs, CAV1 localized to the plasma membrane, similar to BBB and BSCB capillaries 86 , but redistributed to the cytoplasm in PLVAP + ECs of SOD1 G93A nerves ( Suppl. Fig. 4s-v and data not shown). This pattern resembled that observed in wild-type dorsal root ganglia vessels, which are naturally permeable (PLVAP + ) ( Suppl. Fig. 4w-y ). Increased caveolae-mediated transcytosis in SOD1 G93A endoneurial vessels was further evidenced by a reduction in MFSD2A —a key inhibitor of the pathway 87 , 88 — correlating with PLVAP induction ( Fig. 3u-x ; see also Fig. 2r , heatmap), whereas MFSD2A expression in the spinal cord remained unchanged ( Suppl. Fig. 4p-r ). MFSD2A levels were also decreased in motor nerve biopsies from ALS patients ( Fig. 3y ). We conclude that the early increase in BNB permeability in ALS nerves is primarily driven by a nerve-specific enhancement of transcytosis, which is not observed in the spinal cord. Peripheral nerves harbor ALS-susceptible capillaries with distinctive para-inflammatory tone The observation that BNB leakage and PLVAP induction in endoneurial vessels of SOD1 G93A mice first appeared in sparse hotspots raises the question of whether specific EC subtypes are inherently more susceptible to ALS. To address this, we performed droplet-based single-cell RNA-seq (scRNA-seq) on CD31 + ECs FACS-isolated from the sciatic nerves of onset-stage (P90) SOD1 G93A mice and non-transgenic controls ( Fig. 4a, b ). Unsupervised subclustering of BNB-forming endoneurial ECs ( Cldn5 + ) identified a cell subset with increased Plvap expression in SOD1 G93A samples ( Fig. 4c-e ). This activated population corresponded to capillary-venous ECs expressing carbonic anhydrase Car4 64 , 77 ( Fig. 4e, f and Suppl. Fig. 5a, b ). The other two endoneurial clusters were “capillary-barrier” ECs, characterized by junctional and transport genes ( Scl2a1 , Mfsd2a , Spock2 ), and arterial ECs ( Gja4 , Stmn2 , Sema3g ) ( Fig. 4f and Suppl. Fig. 5b ). The remaining EC clusters were assigned to epineural vessels ( Plvap + in controls), including veins ( Nr2f2 , Ackr1 , Sepl , Vwf ) and capillaries ( Rgcc , Cd300lg , Thrsp , Cxcl12 ; “capillary-Plvap + ”). A smaller “capillary-Plvap − ” cluster, possibly associated with epi/perineurial compartments 64 , lacked Plvap and expressed genes such as Gpihbp1 , Aqp7 and Timp4 ( Fig. 4f and Suppl. Fig. 5b ). The relative abundance of EC subtypes was similar between SOD1 G93A and control nerves, except for a modest expansion of the capillary-venous cluster (p=0.0495) ( Fig. 4g ). Most genes upregulated in SOD1 G93A ECs were found in the capillary-venous subtype, whereas downregulated genes were more prominent in capillary-barrier ECs ( Fig. 4h ), as confirmed by Expression Weighted Celltype Enrichment (EWCE) analysis of bulk RNA-seq DEGs ( Fig. 4i ). Download figure Open in new tab Supplementary Fig. 5: Single-cell profiling of the nerve endothelium in control and SOD1 G93A mice. a , UMAP plots of control and SOD1 G93A nerve EC datasets color-coded by subtype. b , Expression heatmap (z-score) of EC subtype-marker genes defined in a previous sciatic nerve scRNA-seq dataset{Bhat, 2024 #1}. c , Pathological gene signature identified in venous capillaries, visualized on UMAP plots of control and SOD1 G93A nerve ECs using thresholded binary labeling (CellID). d , UMAP plots of control and SOD1 G93A nerve ECs showing expression of representative genes from the capillary-venous pathological signature. e , Expression heatmap (z-score) of top-100 marker genes for EC subtypes from healthy sciatic nerve EC atlas. f , Expression of EC subtype-marker genes visualized in UMAP plots of healthy sciatic nerve EC atlas. g , h , PLVAP signal in CAR4 + and CAR4 − endoneurial vessels of sciatic nerves from P120 control and SOD1 G93A mice ( g ) and P180 control and TDP-43 Q331K mice ( h ), normalized to average control value. x-m’ , Sciatic nerves from control ( i, i’ ), SOD1 G93A ( j-k’ ), and TDP-43 Q331K ( l-m’ ) mice at the indicated stages. Control corresponds to P120. CAR4 + /CLDN5 + venous capillaries (arrows) upregulate PLVAP in mutants. CLDN5 + /CAR4 − vessels remain PLVAP − (arrowheads). i’-m’ show only CAR4 and PLVAP staining. Scale bars: i-m’, 100μm. Download figure Open in new tab Fig. 4: Primed state and vulnerability of capillary subtypes in peripheral nerves of ALS mice. a , Workflow for scRNA-seq of mouse sciatic nerve endothelium. b , c , Uniform manifold approximation and projection (UMAP) plot of integrated control and SOD1 G93A nerve ECs, color coded by genotype ( b ) or by endoneurial (orange) vs. epineurial (gray) identity ( c ). The dataset comprises 2611 cells (control, 1250 cells; SOD1 G93A , 1361cells). d , Reclustering of endoneurial ECs from integrated control and SOD1 G93A dataset. Cluster 2 has high Plvap expression in SOD1 G93A . e , UMAP plots showing expression of Cldn5 , Plvap and Car4 in control and SOD1 G93A endoneurial ECs. f , UMAP plot of integrated control and SOD1 G93A dataset color-coded by EC subtype. Capillary-venous ECs correspond to venous-Plvap − ECs annotated in a previous nerve dataset{Bhat, 2024 #1}. g , Relative abundance of EC subtypes in control and SOD1 G93A nerves. h , Number of upregulated (red) and downregulated (blue) DEGs (padj <0.1) identified by scRNA-seq in nerve EC subtypes in SOD1 G93A vs. control, normalized by cluster size. i , Distribution of upregulated (red) and downregulated (blue) DEGs (padj <0.1, concordant at P90-P120) from bulk RNA-seq of SOD1 G93A vs. control sciatic nerve ECs computed by EWCE. j , UMAP plot of ECs from healthy sciatic nerve color-coded by subtype. The atlas comprises 6 independent scRNA-seq datasets. The capillary-arterial subtype was not resolved in our scRNA-seq dataset alone ( Fig. 4f ) k , Dendrogram for hierarchical clustering of EC subtypes from healthy sciatic nerve atlas based on the top 200 markers for each celltype. l , Dot plot showing the expression of EC subtype markers from healthy sciatic nerve atlas. m , Expression of “venous signature” in nerve EC subtypes from healthy sciatic nerve atlas. The gene set includes 58 venous markers identified in various organs{Kalucka, 2020 #2} with padj <10 -8 in nerve capillary-venous ECs. n-p’ , ICAM1/CLDN5 co-staining identifies venous capillaries (arrows) in sciatic nerves from P60 control ( n , n’ ), SOD1 G93A ( o , o’ ) and TDP-43 Q331K ( p , p’ ) mice. Venous capillaries become PLVAP + in mutants. CLDN5 + /ICAM1 − vessels remain PLVAP − (arrowheads). n’, o’ , p’ show only ICAM1 and CLDN5 staining. Dashed contours outline the endoneurium. q-r , CAR4/CLDN5 co-staining identifies venous capillaries (arrows) in sciatic nerves from P120 control ( q ) and SOD1 G93A ( r ) mice. CAR4 + vessels become PLVAP + in SOD1 G93A . CAR4 − vessels remain PLVAP − (arrowheads) Dashed contours outline the endoneurium. s , t , Percentage of CAR4 + vs. CAR4 − endoneurial vessels expressing PLVAP in control, SOD1 G93A ( s ) or TDP-43 Q331K ( t ) sciatic nerves at the indicated stages. u , Expression heatmap (z-score) of capillary-venous pathological signature across EC subtypes from controls and SOD1 G93A nerves. The signature includes 50 genes upregulated in SOD1 G93A capillary-venous ECs (padj <0.05). The dendrogram illustrates hierarchical clustering of EC subtypes in the two conditions. v , Expression of the nerve capillary-venous pathological signature across EC subtypes from nerve and spinal cord in the indicated conditions. w , x , Log2 fold-change (Log2FC) vs. Log2FC scatterplot comparing DEGs (padj <0.1, colored) in ECs from healthy nerve vs. spinal cord (x-axis) and SOD1 G93A vs. control nerve (y-axis) at P90 ( w ) and P120 ( x ). Selected DEGs are labeled. y , GO analysis of DEGs (P90 and P120 combined) from the comparisons in w , x . Genes downregulated in healthy nerve and upregulated in SOD1 G93A (top/left quadrants) did not yield significant GO results. z , Distribution of DEGs (from comparisons in w , x ) across EC subtypes in healthy sciatic nerve, based on expression strength index (ESI, see Methods). Scale bars: n-p’, 100μm; q, r, 100μm. To better characterize steady-state features of these EC subtypes, we generated a scRNA-seq atlas of the healthy nerve endothelium by integrating our control samples with five published adult mouse sciatic nerve datasets 64 , 89 – 92 . This yielded a total of ∼5000 ECs, divided into seven clusters, which were annotated based on previously defined markers 64 , 77 ( Fig. 4j and Suppl. Fig. 5e ). The four endoneurial EC subtypes exhibited an arteriovenous zonation similar to the brain vasculature 93 , encompassing capillary-venous, capillary-barrier, capillary-arterial, and arterial populations ( Fig. 4j ). Hierarchical clustering revealed a close relationship between capillary-venous and capillary-barrier ECs, the two subtypes most affected in SOD1 G93A nerves ( Fig. 4k ). Accordingly, capillary-venous ECs expressed vascular barrier genes ( Cldn5, Ocln , Tjp1 , Bsg , Spock2 , Mfsd2a , Lef1 , Scl2a1 ), though often at lower levels than bona fide barrier capillaries ( Fig. 4l and Suppl. Fig. 5e, f ). At the same time, they exhibited a venous phenotype defined by highly conserved markers ( Apoe , Vcam1 , Vwf , Ctla2a , Ptgs1 , Tmsb10 , Icam1 ), along with others shared across venous beds in different organs ( Car4 , Lcn2 , Ch25h , Ptgis , Adgrg6 ) 77 ( Fig. 4l, m and Suppl. Fig. 5e, f ). In addition, enrichment of ribosomal protein genes suggested elevated basal translational activity ( Suppl. Fig. 5e ). Venous capillaries in the sciatic nerve were identified by co-staining for the venous adhesion molecule ICAM1 77 and the barrier marker CLDN5 ( Fig. 4l, n, n’ and Suppl. Fig. 5e ). Consistent with the predicted susceptibility of this EC subtype, PLVAP upregulation was restricted to ICAM1 + /CLDN5 + vessels in pre-symptomatic SOD1 G93A and TDP-43 Q331K mice ( Fig. 4o-p ’). However, because ICAM1 levels increased in mutants ( Fig. 5b and Suppl. Fig. 11h ), we instead labeled venous capillaries with CAR4, which remained stable throughout disease progression ( Fig. 4e, l, q ). As expected, ectopic PLVAP expression was again confined to CAR4 + /CLDN5 + capillaries in both SOD1 G93A and TDP-43 Q331K models at pre-symptomatic stage, and this association persisted as the proportion of PLVAP + endoneurial vessels rose in the symptomatic phase ( Fig. 4r-t , Suppl. Fig. 5g-m ’). Download figure Open in new tab Fig. 5: Endothelial activation in peripheral nerves of SOD1 G93A ALS mice parallels myeloid remodeling but is independent of monocyte infiltration. a , b , VWF ( a ) and ICAM1 ( b ) signal in endoneurial vessels of control and SOD1 G93A sciatic nerves at P60 and P120, normalized to average control value. c-f , Sciatic nerves from control ( c ) and SOD1 G93A ( d-f ) mice at the indicated stages stained for VCAM1. CD31 marks all vessels. VCAM1 is upregulated in endoneurial vessels of mutants (arrows). g , VCAM1 signal in endoneurial vessels of control and SOD1 G93A sciatic nerves during disease progression, normalized to average control value. h-k , Sciatic nerves of control ( h ) and SOD1 G93A ( i-k ) mice at the indicated stages, expressing Cx3cr1 GFP / Ccr2 RFP reporter. CD31 marks vessels. In controls, elongated GFP + MΦ associate with vessels ( h , arrowheads). In SOD1 G93A nerves, rare globular GFP + MΦ appear at P60 ( i , white arrows) and increase in the parenchyma at P90 ( j ) and P120 ( k ). RFP + monocytes infiltrate SOD1 G93A nerves at P120 ( k , red arrows); some acquire GFP expression ( k , yellow arrows). Panels are magnified from Suppl. Fig. 7l . l , Number of GFP + , RFP + , GFP + /RFP + double-positive MΦ in control and SOD1 G93A sciatic nerves at the indicated stages. Mean ± SEM. m , Number of leukocytes (CD45 + ) measured by flow cytometry in control and SOD1 G93A sciatic nerves at P120. Mean ± SEM. n , t-SNE density maps of multi-parametric flow cytometry profiling of sciatic nerve myeloid cells showing changes in abundance of discrete populations in SOD1 G93A mice compared to controls (P120). MΦ subsets 1-6; Mon, monocytes; DC, dendritic cells; Neu, neutrophils. o , Frequency of myeloid populations identified by unsupervised t-SNE analysis (gated from CD11b + cells). Small MΦ5 and MΦ6 clusters that did not display significant changes in mutants are not shown. Mean ± SEM. p , Flow cytometry plots from control and SOD1 G93A (P120) sciatic nerves showing resident macrophages (rMΦ, green gate), activated macrophages (aMΦ, organe gate), and monocytes (Mon, red gate) discriminated based on F4/80 and Ly6c levels (schematic, left panel). q , Frequency of rMΦ, aMΦ and monocytes (gated from CD11b + cells) in control and SOD1 G93A nerves. Mean ± SEM. r , Normalized mean fluorescence intensity (MFI) of selected inflammatory markers upregulated in SOD1 G93A nerve macrophages. Since aMΦ and monocytes are nearly absent in control nerves, the total F4/80 + population is used for comparison with SOD1 G93A . s , Frequency of rMΦ, aMΦ and monocytes (gated from CD11b + cells) in control, SOD1 G93A , and SOD1 G93A ; Ccr2 -/- sciatic nerves. Mean ± SEM. Flow cytometry plots are shown in Suppl. Fig. 8f . t-y , Sciatic nerves from control ( t , u ), SOD1 G93A ( v , w ), and SOD1 G93A ; Ccr2 -/- ( x , y ) mice at P120 stained for PLVAP and CLDN5 (DAPI, nuclei) ( t , v , x ), or VCAM1 (CD31, all vessels) ( u , w , y ). Dashed contours outline the endoneurium. PLVAP and VCAM1 are upregulated in CNDN5 + vessels in SOD1 G93A and SOD1 G93A ; Ccr2 -/- nerves (arrows). Some vessels remain PLVAP − (arrowheads). z , PLVAP and VCAM1 signal in endoneurial vessels of control, SOD1 G93A , and SOD1 G93A ; Ccr2 -/- sciatic nerves, normalized to average control value. Scale bars: c-f, 100μm; h-k, 50μm; t, v, x, 200μm; u, w, y, 100μm. Next, we defined a “pathological” signature enriched in SOD1 G93A capillary-venous ECs, comprising genes linked to vascular inflammation, permeability and activation ( Plvap , Vcam1 , Lrg1 , Tacr1 ), along with others with less defined endothelial functions (e.g., water channel-forming Aqp1 ) and ribosomal protein genes ( Fig. 4u and Suppl. Fig. 5c, d ). SOD1 G93A capillary-barrier ECs also upregulated this gene set, albeit to a lesser extent ( Fig. 4u ). Unexpectedly, based on this signature control capillary-venous ECs clustered with mutant barrier ECs ( Fig. 4u ), suggesting a basal state of activation further amplified in disease. To assess whether the ALS signature identified in nerve capillary-venous ECs was shared across other pathological conditions, we turned to experimental autoimmune encephalomyelitis (EAE), a model characterized by BSCB breakdown and inflammatory changes primarily affecting venous ECs 94 . Using scRNA-seq data from spinal cord ECs of healthy controls, Freund’s adjuvant (CFA)-immunized mice, and mice with acute or chronic EAE, we annotated Cldn5 + EC clusters corresponding to capillary-venous, capillary-barrier, and arterial subtypes ( Suppl. Fig. 6a-c ) 94 . Among these, only venous capillaries in CFA and EAE spinal cords, but not in healthy controls, upregulated the pathological gene set, though at much lower levels than in nerves ( Fig. 4v ). While key mediators of vascular inflammation and damage, such as Vcam1 and Lrg1 , were induced in both tissues, other genes were specific to venous capillaries in SOD1 G93A nerves (e.g., Plvap , Tacr1 , Aqp1 , Aplnr ) ( Suppl. Fig. 6d ). That venous ECs in CFA mice paralleled changes seen in EAE ( Fig. 4v ) underscores the sensitivity of the signature, as CFA/pertussis toxin —used as a control for EAE— per se increases vascular permeability and immune activation 95 – 97 . Furthermore, this comparison reinforced the notion that capillary-venous ECs in healthy nerves exhibit an activated gene profile, comparable to spinal cords of CFA-treated and EAE mice ( Fig. 4v ). Download figure Open in new tab Supplementary Fig. 6: Distinct specialization and reactivity of venous capillaries in spinal cord and nerve. a , UMAP plot of integrated spinal cord scRNA-seq EC dataset{Shahriar, 2024 #3} showing expression of Cldn5 . Only Cldn5 + cells from the original dataset were retained for the analysis. b , Split UMAP plots for healthy, CFA-treated, acute EAE (16-17 days postimmunization, d.p.i.), and early chronic EAE (28-29 d.p.i.) color-coded by unsupervised cell clusters. Clusters 1, 2, and 3 correspond to capillary-barrier, artery, and capillary-venous ECs, respectively. c , Dot plot showing expression of marker genes used for EC cluster annotation. d , Dot plot showing expression of top DEGs from the nerve capillary-venous pathological signature (padj ≤0.05 and Log2FC ≥0.5 in SOD1 G93A vs. control) in spinal cord EC subtypes under healthy, CFA-treated, acute, and chronic EAE conditions. e , GSEA of biological processes in capillary-venous ECs from healthy nerve (from nerve EC atlas) versus spinal cord{Shahriar, 2024 #3}. f , GO enrichment of biological processes in ECs from healthy nerve and spinal cord based on bulk RNA-seq (control samples). We therefore asked whether the nerve vasculature exhibits a basal activation state. Compared with spinal cord ECs, healthy nerve ECs were enriched for inflammatory pathways ( Fig. 2m and Suppl. Fig. 6f ), which were further upregulated in SOD1 G93A ( Fig. 4w, x , top-right quadrants and Fig. 4y ) and predominantly detected in capillary-venous ECs ( Fig. 4z ). In contrast, nerve-enriched genes downregulated in SOD1 G93A were associated with vascular homeostasis and enriched in capillary-barrier and arterial ECs ( Fig. 4w, x , bottom-right quadrants; Fig. 4y, z ). Altogether, these findings suggest that the nerve endothelium maintains a basal para-inflammatory tone 98 , partly driven by the distinct transcriptional profile of capillary-venous ECs compared with their spinal cord counterparts ( Suppl. Fig. 6e ), possibly making the nerve vasculature more susceptible to pathological changes in ALS. Endothelial activation precedes and is independent of monocyte infiltration in ALS nerves In line with our transcriptional analysis pointing to inflammatory activation of the nerve endothelium, immunostaining of SOD1 G93A sciatic nerves revealed increased expression of VWF, ICAM1, and VCAM1—key mediators of vascular inflammation and immune cell recruitment— in endoneurial microvessels from the pre-symptomatic stage (P60) onward ( Fig. 5a-g and Suppl. Fig. 7a-e ”). To explore the spatiotemporal relationship between EC activation and remodeling of the myeloid compartment in peripheral nerves, we crossed SOD1 G93A with the Cx3cr1 GFP / Ccr2 RFP dual-reporter allele ( Fig. 5h-k and Suppl. Fig. 7l ). This model distinguished GFP + resident macrophages expressing fractalkine receptor CX3CR1 from RFP + infiltrating inflammatory monocytes expressing CCR2, the receptor for CCL2 —a monocyte chemoattractant upregulated in SOD1 G93A nerve ECs ( Fig. 2r ). In control sciatic nerves, GFP + resident macrophages with an elongated morphology were closely associated with endoneurial vessels, while blood-derived RFP + monocytes were virtually absent ( Fig. 5h ). In SOD1 G93A nerves, GFP + macrophages began to increase in numbers at the presymptomatic stage (P60) ( Fig. 5i, l ). By disease onset (P90), many had adopted a globular shape consistent with phagocytic activity and were scattered in the nerve parenchyma, away from vessels ( Fig. 5j, l ). At this stage, RFP + monocytes began infiltrating, and by the early-symptomatic phase (P120) both populations had expanded further. A subset of RFP + monocytes co-expressed GFP, reflecting their differentiation into activated macrophages ( Fig. 5k, l ). Download figure Open in new tab Supplementary Fig. 7: Endothelial activation in ALS nerves, but not spinal cord, coincides with resident macrophage activation. a , b , Upregulation of VWF in CAR4 + capillaries of SOD1 G93A vs. control sciatic nerves at P60 (arrows). Arrowheads point to VWF + /CAR4 − vessels. c-e” , Sciatic nerves from control ( c-c” ) and SOD1 G93A mice at P60 ( d-d” ) and P120 ( e-e” ). ICAM1 and VWF are upregulated in PLVAP+ endoneurial vessels in mutants (arrows). VWF also increases in PLVAP − vessels (arrowheads, d-d” ). ICAM1 and VWF staining are shown separately in c’-e’ and c”-e” , respectively. f , g , VCAM1 marks larger veins in both control ( f ) and SOD1 G93A ( g ) spinal cords (arrowheads) and is not upregulated in the microvasculature of mutants at P120. CD31 marks all vessels. h , VCAM1 signal in control and SOD1 G93A spinal cord vasculature, normalized to average control value. i , j , VWF marks larger veins in both control ( i ) and SOD1 G93A ( j ) spinal cord (arrowheads) and is not upregulated in the microvasculature of mutants at P120. CD31 marks all vessels. k , VWF signal in control and SOD1 G93A spinal cord vasculature, normalized to average control value. l , Sciatic nerves of control and SOD1 G93A mice at the indicated stages, expressing Cx3cr1 GFP / Ccr2 RFP reporter. CD31 marks vessels. Boxed areas are magnified in Fig. 5h-k . m , Spinal cord of control and SOD1 G93A mice (P120) expressing Cx3cr1 GFP / Ccr2 RFP reporter. GFP + microglia expand in mutants while RFP + monocytes are absent. n-p’ , Motor nerves from P60 control ( n , n’ ) and SOD1 G93A ( o , o’ ) mice, or SOD1 G93A sensory nerves ( p , p’ ). Focal induction of VCAM1 in endoneurial vessels is visible in mutant motor nerves (arrowhead in o’ ), but not sensory nerves. IBA1 labels macrophages (arrowheads). Boxed areas in n , o , p are magnified in n’ , o’ , p’ . q , VCAM1 signal in endoneurial vessels of P60 control and SOD1 G93A motor nerves, normalized to average control value. r-s’ , Motor nerves from P60 control and SOD1 G93A mice expressing Cx3cr1 GFP / Ccr2 RFP reporter. VCAM1 marks activated endoneurial vessels in mutants (arrows in s , s’ ). Nearby GFP + resident macrophages remain elongated but increase in numbers compared to controls (arrowheads). VCAM1 staining is show separately in r’ , s’ . t , Number of GFP + , RFP + , GFP + /RFP + double-positive MΦ in control and SOD1 G93A motor nerves (P60). Mean ± SEM. Scale bars: a, b, 100μm; c-e”, 100μm; f, g, 200μm; I, j, 100μm; l, 100μm; m, 100μm; n, o, p, 100μm; n’, o’, p’, 50μm; r-s’, 50μm. We concluded that in SOD1 G93A nerves carrying motor fibers ( sciatic and quadriceps n. ), but not in sensory nerves, EC activation marked by VCAM1 induction coincided with the early expansion and activation of resident macrophages, whereas infiltration of inflammatory monocytes occurred later ( Fig. 5g, l and Suppl. Fig. 7n-t ). In contrast, EC activation was not observed in the spinal cord of symptomatic SOD1 G93A mice, as VCAM1 and VWF were not induced in the microvasculature but remained confined to larger veins ( Suppl. Fig. 7f-k ). At this stage, microglia marked by Cx3cr1 GFP expanded without recruitment of Ccr2 RFP monocytes ( Suppl. Fig. 7m ). Given the activated EC profile and increased leukocyte counts in SOD1 G93A peripheral nerves ( Fig. 5m ), we used high-dimensional spectral flow cytometry to profile the immune compartment in sciatic nerves from control and mutant mice (P120) ( Fig. 5n, o and Suppl. Fig. 8a ). The expansion of leukocytes was not due to lymphocytic infiltration, as B cells (CD19 + ) and T cells (CD4 + and CD8 + ) were not detected ( Suppl. Fig. 8b-e ). Instead, unsupervised analysis of myeloid composition using t-distributed stochastic neighbor embedding (t-SNE) dimensionality reduction revealed a substantial increase in the frequencies of monocytes, specific macrophage subpopulations, and neutrophils in SOD1 G93A nerves, while dendritic cells remained stable ( Fig. 5n-o ). Macrophage clusters MF1 and MF2, exhibiting a resting phenotype, were reduced in SOD1 G93A mice, whereas MF3 and MF4 with an activated phenotype —marked by high expression of IA-IE, CD64, CD68, CX3CR1, CD86 and CD80— were enriched ( Fig. 5n, o and Suppl. Fig. 8a ). Considering the role of infiltrating monocytes and macrophages in modulating ALS progression 53 , 54 , 99 , 100 , we conducted a supervised analysis to further characterize these lineages, distinguishing three main subpopulations based on F4/80 and Ly6c expression: tissue-resident MF (rMF; F4/80 + Ly6c − ), activated MF (aMF) displaying intermediate Ly6c levels (F4/80 + Ly6c low ), and monocytes (F4/80 low Ly6c hi ) ( Fig. 5p ). In healthy sciatic nerves, most CD11b + myeloid cells were rMF, but this subpopulation declined in SOD1 G93A mice ( Fig. 5p, q ) and began to express pro-inflammatory markers such as IA-IE, CD11c and CD64 ( Fig. 5r ). In parallel, inflammatory monocytes increased, along with a substantial expansion of the aMF subset, which upregulated several pro-inflammatory molecules (e.g., CD80, CD86, IA-IE, CD11c, CD64, CD68 and CCR2) ( Fig. 5r ). The aMΦ population likely comprised both infiltrating monocyte-derived macrophages (CCR2 + ) and resident macrophages that may have transitioned toward a pro-inflammatory state, in contrast to the pro-resolving traits typically associated with Ly6c low macrophages in other contexts 53 , 54 , 101 , 102 . Collectively, these findings reveal a shift of the macrophage compartment toward an activated state in ALS-affected nerves. Download figure Open in new tab Supplementary Fig. 8: Nerve-resident macrophages drive immune remodeling in SOD1 G93A mice without lymphocyte recruitment. a , t-SNE maps of nerve myeloid cells showing antigen expression profiles used to identify populations by unsupervised clustering. b , Flow cytometry plots showing near-absent CD19 + B cells in sciatic nerves of control and SOD1 G93A mice at P120. c , Frequency of CD11b + myeloid cells (left) and CD19 + B cells (right) in control and SOD1 G93A nerves. Mean ± SEM. d , Flow cytometry plots showing negligible presence of CD4 + and CD8 + T cells in sciatic nerves of control and SOD1 G93A mice at P120. e , Frequency of CD4 + (left) and CD8 + T cells (right) in control and SOD1 G93A nerves. Mean ± SEM. f , Flow cytometry plots showing resident macrophages (rMΦ), activated macrophages (aMΦ), and monocytes (Mon) discriminated based on F4/80 and Ly6c levels in sciatic nerves from control, SOD1 G93A , and SOD1 G93A ; Ccr2 -/- mice at P120. Monocyte infiltration is prevented in SOD1 G93A ; Ccr2 -/- mice. The percentage of aMΦ is not reduced in SOD1 G93A ; Ccr2 -/- compared to SOD1 G93A indicating that they primarily derive from rMΦ. g-i , Macrophages identified by F4/80 staining in sciatic nerves from control, SOD1 G93A , and SOD1 G93A ; Ccr2 -/- mice at P120. CD31 marks vessels; DAPI, nuclei. Macrophages are elongated in controls (arrowheads) and become globular/phagocytic in mutants (arrows). j , Macrophage counts in SOD1 G93A , and SOD1 G93A ; Ccr2 -/- sciatic nerves based on F4/80 staining. Mean (normalized to control) ± SEM. Scale bar: g-i, 100μm. In light of these dynamics, we investigated whether EC activation was sustained by monocytes/macrophages infiltration by crossing SOD1 G93A mice with a Ccr2 knockout allele, which prevents monocyte recruitment to the sites of inflammation 103 . As expected, the numbers of inflammatory monocytes in sciatic nerves were drastically reduced in SOD1 G93A ; Ccr2 -/- mice compared to SOD1 G93A mutants at P120 ( Fig. 5s and Suppl. Fig. 8f ). However, there were no significant differences in the frequencies of rMΦ and aMΦ populations in the two models, further supporting the idea that aMΦ were not derived from inflammatory monocytes, but likely emerged from rMΦ that had acquired a pro-inflammatory profile ( Fig. 5s and Suppl. Fig. 8f ). Notably, despite near-complete absence of inflammatory monocytes and their derivatives in SOD1 G93A ; Ccr2 -/- sciatic nerves, EC activation and BNB impairment — reflected by ectopic induction of VCAM1 and PLVAP on endoneurial vessels, respectively— occurred to the same extent as in SOD1 G93A mice ( Fig. 5t-z ). Thus, ALS-induced vascular inflammation in peripheral nerves precedes extensive reconfiguration and activation of tissue-resident macrophages and develops independently of circulating monocyte infiltration, which takes place only after disease onset. Whether the earliest instances of focal EC activation and resident macrophage responses are interdependent or just temporally aligned remains to be addressed. Neutrophils sustain endothelial activation and contribute to axonal pathology in ALS peripheral nerves High-dimensional unsupervised and supervised spectral flow cytometry analysis of leukocytes in sciatic nerve revealed a significant increase in neutrophils (CD11b + Ly6G + ) in SOD1 G93A mice at P120 ( Fig. 5n, o ; Fig. 6a, b ). Since this approach did not differentiate between circulating and extravasated neutrophils, we performed tissue immunostaining for Ly6g and Ly-6B.2 (7/4) to assess neutrophil infiltration into the nerve parenchyma. In control mice, extravasated neutrophils were virtually absent. In contrast, they were present in the sciatic and motor nerves of SOD1 G93A mice as early as P60 and became increasingly abundant by disease onset (P90) ( Fig. 6c, d ). However, in the spinal cord of mutants, neutrophils remained rare, even at advanced stages, and were confined to the vessel lumen or meninges without infiltrating the parenchyma ( Suppl. Fig. 9a-c ). The accumulation of neutrophils in SOD1 G93A peripheral nerves after P90 followed the early induction of inflammatory markers on endoneurial vessels, which became evident at P60 ( Fig. 5d ), suggesting that their initial ingress from the circulation was mediated by focal activation of the nerve endothelium 104 , 105 . Download figure Open in new tab Suppl. Fig. 9: Neutrophil infiltration in SOD1 G93A nerves, but not spinal cord, contributes to vascular damage. a , b , Neutrophils identified by Ly-6B.2 (7/4) staining in the spinal cord of control ( a ) and SOD1 G93A ( b ) mice at P120. Few neutrophils (arrowheads) are seen within the vasculature or associated with the meninges (a, left; b, middle) in both genotypes. CD31 marks vessels; DAPI, nuclei. Dashed contours outline the spinal cord parenchyma. c , Number of neutrophils infiltrating the spinal cord of control and SOD1 G93A mice. Extravascular neutrophils are rare, with no significant difference between genotypes. d , Percentage of neutrophils in peripheral blood assessed during anti-Ly6G treatment. Circulating neutrophils are similar between control and SOD1 G93A mice and are depleted by anti-Ly6G treatment. e , Flow cytometry plots showing increased neutrophils (CD11b + Ly6G + ) in the sciatic nerves of SOD1 G93A mice (P120) and their depletion following anti-Ly6G treatment. f , Frequency of neutrophils in sciatic nerves of control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. g-i , Neutrophils [Ly-6B.2 (7/4), arrowheads] detected in SOD1 G93A quadriceps nerves ( h ), but not in control ( g ) or SOD1 G93A /anti-Ly6G-treated ( i ) mice at P120. CD31, vessels. j-l , PLVAP induction in CLDN5 + endoneurial vessels of SOD1 G93A quadriceps nerves at P120 ( k , arrows) is reduced in SOD1 G93A /anti-Ly6G-treated mice ( l ). PLVAP is absent in control ( j ). CD31, vessels. m , Neutrophil counts [Ly-6B.2 (7/4) staining] in motor nerves of control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. Mean ± SEM. n , o , PLVAP ( n ) and CLDN5 ( o ) signal in motor nerve endoneurial vessels at P120, normalized to average control value. p-r’ , TEM images of endoneurial vessels from control ( p , p’ ), SOD1 G93A /isotype-treated ( q , q’ ) and SOD1 G93A /anti-Ly6G-treated ( r , r’ ) motor nerves. Boxed areas in p , q , r are magnified in p’ , q’ , r’ . Pinocytotic vesicles (arrowheads) are frequent in SOD1 G93A ( q’ ) but not in controls ( p’ ) and anti-Ly6G-treated mutants ( r’ ). Partial opening of EC junctions (arrows) is observed in SOD1 G93A ( q’ ) mutants but not in controls ( p’ ) and anti-Ly6G-treated mice ( r’ ). s , Number of ChAT + motor neurons in spinal cord sections of control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. Mean (normalized to control) ± SEM. t , u , Quantification of astrocytes (GFAP + ) ( t ) and microglia (IBA1 + ) ( u ) in the spinal cord of control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. Mean (normalized to control) ± SEM. v , All-limb grip strength test in control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. Mean ± SEM. w , Rotarod motor performance test in control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. Mean ± SEM; normalized to body weight, BDW). x , Body weight measurement over time in control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice. Mean (normalized to initial BDW) ± SEM. Anti-Ly6G treatment started at 12 weeks (P90). Scale bars: a-b, 100μm; g-l, 100μm; p, q, 1μm; r, 2μm; p’, q’, r’, 500nm. Download figure Open in new tab Fig. 6: Neutrophil depletion reverses vascular damage and mitigates axonal degeneration in SOD1 G93A ALS mice. a , b , Representative flow cytometry plots ( a ) and corresponding frequency measurements ( b ) showing increase of neutrophils (CD11b + Ly6G + ) in sciatic nerves of SOD1 G93A mice at P120. c , Neutrophils (Ly6G + , arrowheads) infiltrate SOD1G93A motor nerves at P90. CD31 marks vessels; DAPI, nuclei. d , Number of extravasated neutrophils in motor nerves of control and SOD1 G93A mice at different stages. Mean ± SEM. e , Timeline of neutrophil depletion with anti-Ly6G treatment. f-h , Neutrophils detected by Ly-6B.2 (7/4) staining in sciatic nerves of control ( f ), SOD1 G93A /isotype antibody-treated ( g ) and SOD1 G93A /anti-Ly6G-treated ( h ) mice at P120. Infiltrating neutrophils (arrowheads) are depleted by anti-Ly6G treatment. CD31, vessels; DAPI, nuclei. i-k , VCAM1 induction in endoneurial vessels of SOD1 G93 sciatic nerves ( j , arrows; compare to control mice, i ) is rescued by anti-Ly6G treatment ( k ). l-n , PLVAP upregulation in CLDN5 + vessels of SOD1 G93A sciatic nerves ( m , arrows; compare to control mice, l ) is rescued by anti-Ly6G treatment ( n ). Dashed contours outline the endoneurium. o , Neutrophil counts (Ly-6B.2 7/4 and Ly6G staining) in sciatic nerves of control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. Mean ± SEM. p , VCAM1 signal in sciatic nerve endoneurial vessels at P120, normalized to average control value. q , Macrophages counts (F4/80 staining) in sciatic nerves at P120, normalized to average control value. r , s , PLVAP ( r ) and CLDN5 ( s ) signal in sciatic nerve endoneurial vessels at P120, normalized to average control value. t , Vesicle density in motor nerve ECs at P120 assessed by TEM. u , t-SNE density maps of multi-parametric flow cytometry profiling of myeloid cells from sciatic nerves of control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. MΦ subsets 1-6; Mon, monocytes; DC, dendritic cells; Neu, neutrophils. v , Frequency of neutrophils, monocytes, resident (non-activated) MΦ1 and activated MΦ3 subsets identified by unsupervised t-SNE analysis (gated from CD11b + cells). Mean ± SEM. w , Toluidine Blue-stained semithin sections of quadriceps (motor) nerves from control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120. Degenerating fibers (arrows) are reduced by anti-Ly6G treatment. x , y , Axon density ( x ) and percentage of degenerating axons ( y ) assessed in semithin sections from control and SOD1 G93A motor nerves. Mean ± SEM. z , Model: an exacerbating loop between nerve EC activation (EC A ) and neutrophil (Neu) infiltration promotes vascular damage and neurodegeneration. EC dysfunction reverses upon neutrophil depletion (EC R , resting). Neutrophils amplify EC responses and modulate activation of tissue-resident macrophages (MΦ A ) independently of monocyte (Mon) infiltration. Monocytes do not contribute to EC activation. Potential interactions between EC and MΦ activation remain to be defined. Scale bars: c, 100μm; f-n, 100μm; w, 20μm. Given that infiltrating neutrophils are known to increase vascular permeability 106 , 107 , we tested their contribution to sustained EC activation and BNB dysfunction. SOD1 G93A mice were treated with an anti-Ly6G depleting antibody, or isotype control via intraperitoneal injection from P90 to P120, targeting the window of significant neutrophil accumulation ( Fig. 6e ). The treatment efficiently depleted neutrophils from blood and nerve tissue, as confirmed by flow cytometry and immunofluorescence for the Ly-6B.2 (7/4) neutrophil marker, which showed a substantial reduction of neutrophil infiltration in SOD1 G93A sciatic nerves ( Fig. 6f-h, o and Suppl. Fig. 9d-f ). Neutrophil depletion significantly improved vascular pathology in sciatic and motor nerves of SOD1 G93A mice. The aberrant induction of VCAM1 and PLVAP in endoneurial vessels was substantially attenuated, while CLDN5 expression was partially rescued in anti-Ly6G-treated mutants ( Fig. 6i-n, p, r, s and Suppl. Fig. 9g-o ). Furthermore, TEM revealed a marked reduction in pinocytotic vesicles in motor nerve ECs, along with restored, sealed TJs ( Fig. 6t and Suppl. Fig. 9p-r ’). Thus, preventing neutrophil infiltration in peripheral nerves promotes vascular integrity. We next evaluated the relationship between neutrophil infiltration and activation of the myeloid compartment in SOD1 G93A nerves. Neutrophil depletion mitigated macrophage dysregulation, leading to a reduction in F4/80 + MΦ counts in the affected tissues ( Fig. 6q and Suppl. Fig. 10a-c ). High-dimensional unsupervised spectral flow cytometry revealed a reduction in the activated macrophage subset (MF3), accompanied by a corresponding increase in the resting resident pool (MF1), while the frequency of inflammatory monocytes remained unchanged ( Fig. 6u-v and Suppl. Fig. 10f-h ). Supervised analysis further confirmed the increase in rMΦ (F4/80 + Ly6C − ) and a parallel decrease in aMΦ (F4/80 + Ly6C low ) subpopulations, with no significant changes in infiltrating monocytes ( Suppl. Fig. 10d, e ). Conversely, neutrophil numbers and extravasation remained elevated in the sciatic nerves of SOD1 G93A ; Ccr2 -/- mice, indicating that neutrophil recruitment occurred independently of monocyte infiltration ( Suppl. Fig. 10i-n ). Download figure Open in new tab Suppl. Fig. 10: Neutrophils modulate macrophage activation in nerves independently of monocyte recruitment. a-c , Macrophages identified by F4/80 staining in sciatic nerves from control ( a ), SOD1 G93A /isotype-treated ( b ) and SOD1 G93A /anti-Ly6G-treated ( c ) mice at P120. DAPI, nuclei. Quantification shown in Fig. 6q . d , Flow cytometry plots from sciatic nerves of control, SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice at P120 showing resident macrophages (rMΦ, green gate), activated macrophages (aMΦ, orange gate), and monocytes (red gate) discriminated based on F4/80 and Ly6c levels. e , Frequency of rMΦ, aMΦ and monocytes (gated from CD11b + cells) from sciatic nerves. Mean ± SEM. f , t-SNE maps of nerve myeloid cells showing antigen expression profiles used for unsupervised clustering. g , Normalized MFI of surface markers used to define myeloid cell clusters in unsupervised analysis. h , Frequency of MΦ2-6 subsets in sciatic nerves identified by unsupervised t-SNE analysis (gated from CD11b + cells). MΦ2 are non-activated (resident) macrophages, reduced in SOD1 G93A mutants. MΦ4 and MΦ6 are activated subsets that increase in mutants. No significant differences are observed between SOD1 G93A /isotype-treated and SOD1 G93A /anti-Ly6G-treated mice. Oher cell clusters are shown in Fig. 6u, v . Mean ± SEM. i-k , Neutrophils identified by Ly-6B.2 (7/4) staining in sciatic nerves from control ( i ), SOD1 G93A ( j ), and SOD1 G93A ; Ccr2 -/- ( k ) mice at P120. CD31 marks vessels; DAPI, nuclei. Infiltrating neutrophils (arrowheads) are present in both mutant genotypes ( j , k ) but not in controls ( i ). l , Neutrophil counts in sciatic nerves based on Ly-6B.2 (7/4) staining. Neutrophils increase in both SOD1 G93A and SOD1 G93A ; Ccr2 -/- mutants, with no significant difference between the two genotypes. Mean (normalized to control) ± SEM. m , Flow cytometry plots from the sciatic nerves of control, SOD1 G93A , and SOD1 G93A ; Ccr2 -/- mice at P120 showing a comparable increase of neutrophils (CD11b + Ly6G + ) in both mutant genotypes. n , Frequency of neutrophils in sciatic nerves of control, SOD1 G93A , and SOD1 G93A ; Ccr2 -/- . Mean ± SEM. Scale bars: a-c, 200μm; i-k, 100μm. Because vascular inflammation and BNB impairment improved upon neutrophil depletion, we examined whether this translated into amelioration of ALS-like phenotypes in SOD1 G93A mice. Semithin motor nerve sections revealed fewer degenerating axon fibers in neutrophil-depleted mutants compared to untreated controls at the symptomatic stage (P120), although total axon numbers remained unchanged ( Fig. 6w-y ). Amelioration of axonal pathology was accompanied by a moderate but significant preservation of motor neurons, while astrogliosis and microgliosis remained elevated ( Suppl. Fig. 9s-u ). However, these improvements did not alter overall disease progression in SOD1 G93A mice, as motor performance — assessed by grip strength and rotarod tests— and body weight decline were not rescued ( Suppl. Fig. 9v-x ). We conclude that EC activation in ALS peripheral nerves is a potentially reversible process, as shown by its attenuation upon neutrophil depletion. Our findings suggest a detrimental feedback loop between EC activation and neutrophil infiltration that perturbs resident macrophages, disrupts the BNB, and accelerates axonal degeneration independently of monocyte recruitment ( Fig. 6z ). Non-cell autonomous impact of ALS mutations on the nerve endothelium We have detected ultrastructural abnormalities accompanied by early upregulation of PLVAP in the nerve endothelium of TDP-43 Q331K mice, similar to SOD1 G93A mutants, demonstrating that these changes are not limited to an individual ALS model (Fig.1l, r and Fig. 4t ). Unlike SOD1 G93A , which is ubiquitously expressed 69 , TDP-43 Q331K is driven by the murine prion promoter, active in neurons and glia 68 , 108 , 109 . Immunostaining for Myc-tagged TDP-43 Q331K showed nuclear localization in spinal cord parenchymal cells, but never in ECs ( Fig. 7a, b, e ). A similar pattern was observed in the sciatic nerve, where rare Myc + nuclei—possibly fibroblasts— were detected, with no expression in ECs ( Fig. 7c-e ). These findings were further confirmed by quantitative RT-PCR for human TDP-43 Q331K on whole sciatic nerve or spinal cord extracts. In line with the immunostaining results, mutant transcript levels were higher in the spinal cord than in the sciatic nerve but remained undetectable in FACS-purified ECs from either tissue ( Fig. 7f ). We concluded that TDP-43 Q331K is absent from ECs in the CNS and PNS, consistent with the reported activity of the transgenic promoter. Download figure Open in new tab Fig. 7: Endothelial abnormalities arise non-cell autonomously in TDP-43 Q331K nerves a , b , Spinal cords of control ( a ) and TDP-43 Q331K ( b ) mice stained for Myc-Tag to reveal the TDP-43 Q331K transgene. TDP-43 Q331K is detected in nuclei of parenchymal cells (neurons and glia, closed arrowheads in inset in b ) but not in EC nuclei (open arrowheads in inset). Myc-Tag does not label non-transgenic controls ( a ). COL4A1 marks vessels; DAPI, nuclei. c , d , Transverse sections of control ( c ) and TDP-43 Q331K ( d ) sciatic nerves stained for Myc-Tag. TDP-43 Q331K is not detected in EC nuclei (open arrowheads in inset in d ). Sparse Myc-Tag + nuclei are visible in the nerve parenchyma (arrowheads). CLDN5 marks endoneurial vessels; DAPI, nuclei. e , Percentage of Myc-tag + nuclei (green, Q331K + ) in the spinal cord and sciatic nerves of control and TDP-43 Q331K mice. EC nuclei are Myc-tag − . f , TDP-43 Q331K transgene expression measured by qRT-PCR in whole spinal cords and sciatic nerves of control and TDP-43 Q331K mice, or in FACS-purified ECs from each tissue. Values represent 2^–ΔCt normalized to the EC-negative fraction. Average cycle threshold (Ct) are indicated above each bar. g-j’ , Endoneurial vessels of sciatic nerves from P120 control ( g , g’ ) and TDP-43 Q331K mice ( h , h’ ), or sensory ( i , i’ ) and motor ( j , j’ ) TDP-43 Q331K nerves stained for PLVAP and CLDN5. PLVAP is upregulated in mutant sciatic and motor nerves ( h , h’ , j , j’ , arrows). PLVAP and CLDN5 are shown separately in g’-j’ . Dashed contours in i and j outline the endoneurium. CD31 marks vessels; DAPI, nuclei. k-n , Endoneurial vessels of sciatic, sensory, and motor nerves of controls or TDP-43 Q331K mice stained for VCAM1. VCAM1 is upregulated in TDP-43 Q331K sciatic and motor nerves ( l , n , arrows). CD31, vessels; DAPI, nuclei. o , p , PLVAP ( o ) and CLDN5 ( p ) signal in sciatic nerve endoneurial vessels at P120, normalized to average control value. q , VCAM1 signal in sciatic nerve endoneurial vessels at P120, normalized to average control value. r , Sciatic nerves of P180 control (upper panel) and TDP-43 Q331K (lower panel) mice injected with sulfo-biotin. The tracer leaks from vessels into the nerve parenchyma in mutants. CD31, vessels; DAPI, nuclei. s , Extravascular sulfo-biotin signal in control and TDP-43 Q331K sciatic nerves, normalized to average control value. t-u , Endoneurial vessels of motor nerves from presymptomatic P60 TDP-43 Q331K mice ( u ) upregulate VCAM1 (arrows). At this stage, IBA1 + macrophages (arrowheads) are comparable between control ( t ) and mutant ( u ) nerves. v , VCAM1 signal in motor nerve endoneurial vessels at P60, normalized to average control value. w , x , Neutrophils identified by Ly-6B.2 (7/4) staining in motor nerves of control ( w ) and TDP-43 Q331K ( x ) mice. Extravascular neutrophils are present in mutants ( x , closed arrowheads). Dashed contours outline the endoneurium. Few neutrophils are visible in the epineurium of controls ( w, open arrowheads). y , Neutrophil counts in motor and sciatic nerves of control and TDP-43 Q331K mice at P180. Mean ± SEM. z , Model: The nerve endothelium senses and amplifies focal ALS-related nerve damage, promoting neuroinflammation and neurodegeneration. EC activation and BNB disruption are driven by non-cell autonomous ( TDP-43 Q331K ) and cell-autonomous ( SOD1 G93A ) mechanisms. Aberrant upregulation of transcytosis during the presymptomatic phase precedes TJ opening. Scale bars: a, b, 100μm; inset in b, 25μm; c, d, 25μm; inset in d, 5μm; g-n, 100μm; r, 200μm; t, u, 100μm; w, x, 100μm Despite this expression pattern, in sciatic and motor nerves of TDP-43 Q331K mice at onset stage (P120; Fig. 1m ), but not in sensory nerves, we observed the full spectrum of endothelial abnormalities identified in SOD1 G93A mutants, including upregulation of PLVAP, which preceded changes in CLDN5 levels ( Fig. 7g-j’, o, p ) —supporting TEM findings showing early accumulation of pinocytotic vesicles in ECs before the loosening of TJs ( Fig. 1l, r and Suppl. Fig. 2u-v ’). EC activation, marked by VCAM1 induction in endoneurial vessels, was already significant in pre-symptomatic mutants (P60), anticipating the increase in macrophage numbers, which became conspicuous at later stages ( Fig. 7k-n, q, t-v and Suppl. Fig. 11a-e ). Vascular inflammation in TDP-43 Q331K nerves was further evidenced by elevated expression of VWF and ICAM1 in endoneurial ECs, coinciding with PLVAP upregulation ( Suppl. Fig. 11f-h ). Similar to SOD1 G93A mice, TDP-43 Q331K mutants displayed BNB leakage of sulfo-biotin in motor and sciatic nerves, but not sensory nerves ( Fig. 7r, s and Suppl. Fig. 1i-l ). In contrast, BSCB permeability remained intact despite astro-microgliosis and motor neuron loss in the spinal cord of symptomatic TDP-43 Q331K mice (P180) ( Suppl. Fig. 11m-s ). At this stage, neutrophils were found to extravasate into motor and sciatic nerves ( Fig. 7w-y and data not shown). In TDP-43 Q331K mice, endothelial phenotypes were penetrant and reached a severity comparable to SOD1 G93A in the affected areas but were less widespread throughout the nerve. The more focal nature of this damage likely reflects the milder motor neuron pathology of TDP-43 Q331K mice compared to the SOD1 G93A model 68 . Download figure Open in new tab Suppl. Fig. 11: Vascular defects in peripheral nerves but not the spinal cord of TDP-43 Q331K mice a , b , Sciatic nerve endoneurial vessels upregulate VCAM1 in P60 TDP-43 Q331K mice ( b , arrows) compared to controls ( a ). Macrophages (IBA1 + , arrowheads) are comparable between genotypes at this stage. CD31 marks vessels; DAPI, nuclei. c , d , Macrophages (F4/80 + , arrowheads) acquire a globular/phagocytic morphology in P120 TDP-43 Q331K sciatic nerves ( d ) compared to controls ( c ). CD31, vessels; DAPI, nuclei. e , Macrophage counts (F4/80 + ) in sciatic nerves of control and TDP-43 Q331K mice at P120, normalized to average control value. f , Upregulation of VWF and PLVAP in sciatic nerve endoneurial vessels from P120 TDP-43 Q331K mice (arrows). Some PLVAP vessels maintain low VWF expression (open arrowheads). CD31 marks all vessels. g , h , VWF ( g ) and ICAM1 ( h ) signal in endoneurial vessels of control and TDP-43 Q331K sciatic nerves at P120, normalized to average control value. i-l , Sciatic nerves from control mice ( i ), and sciatic ( j ), saphenous (sensory, k ), and quadriceps (motor, l ) nerves from P180 TDP-43 Q331K mice injected with sulfo-biotin. The tracer diffuses into the sciatic and motor nerve parenchyma in mutants ( j , i , arrows). CD31, vessels; DAPI, nuclei. m , Spinal cords from P180 control and TDP-43 Q331K injected with sulfo-biotin. The tracer is retained in the vasculature in both genotypes. CD31, vessels; DAPI, nuclei. n , Extravascular sulfo-biotin signal in control and TDP-43 Q331K spinal cords at P180, normalized to average control value. o , p , Spinal cords from P180 control ( o ) and TDP-43 Q331K ( p ) mice stained for GFAP and IBA1 to reveal reactive astrocytes (open arrowheads) and microglia (closed arrowheads), respectively. q , r , Quantification of astrocytes (GFAP + ) ( q ) and microglia (IBA1 + ) ( r ) in P180 control and TDP-43 Q331K spinal cords. Immunoreactivity area normalized to average control value. s , Number of ChAT + motor neurons in spinal cord hemisections of P180 control and TDP Q331K mice. t , Schematic of hTDP-43 Q331K transgene driven by the murine prion promoter and flanked by loxP sites. Bottom: Experimental timeline for Cre-dependent excision of TDP-43 Q331K from ECs with Cdh5 CreER . u-v’ , Spinal cord vessels labeled by lox-STOP-lox (LSL) Tomato reporter driven by Cdh5 CreER in control ( u ) and TDP-43 Q331K mice ( v , v’ ). Myc-Tag marks TDP-43 Q331K -positive nuclei in mutants ( v’ , closed arrowheads) but is absent in EC nuclei ( v’ , open arrowheads). The boxed area in v in magnified in v’ . CD31, vessels. w , Cdh5 CreER -mediated recombination in sciatic nerve ECs (CD31 + ) confirmed by vascular Tomato reporter labeling (closed arrowheads). Tomato also labels the Cdh5 + perineurium (CD31 − , open arrowheads). x , PLVAP induction (arrows) in CLDN5 + endoneurial sciatic nerve vessels is comparable between TDP-43 Q331K and Cdh5 CreER ; TDP-43 Q331K mutants (P120). CD31, marks all vessels; DAPI, nuclei. Lower panels show only PLVAP and CLDN5 staining. y , z , PLVAP ( y ) and VCAM1 ( z ) signal in endoneurial vessels of control, TDP-43 Q331K , and Cdh5 CreER ; TDP-43 Q331K sciatic nerves at P120, normalized to average control value. Scale bars: a-d, 100μm; f, 100μm; i, j, 500μm; k, l, 100μm; m, 100μm; o, p, 100μm; u, v, 100μm; v’, 50μm; w, 100μm; x, 100μm. Thus, despite lacking TDP-43 Q331K expression, ECs in peripheral nerves developed vascular abnormalities similar to those observed in SOD1 G93A mice, where the mutant transgene is present ubiquitously. To further confirm this finding, we took advantage of the design of the TDP-43 Q331K construct, which is flanked by loxP sites, to excise the mutant transgene specifically from ECs via Cre-mediated recombination using Cdh5 CreER 110 . Cdh5 CreER ; TDP-43 Q331K mice were treated with tamoxifen at P30 and analyzed at P135 (post-onset) ( Suppl. Fig. 11t-w ). EC activation and BNB alterations in these mutants remained comparable to TDP-43 Q331K mice ( Suppl. Fig. 11x-z and data not shown), providing additional evidence that the nerve endothelium becomes dysfunctional even in the absence of a direct effect of ALS genes in ECs. These results imply that vascular defects in peripheral nerves can arise non-cell autonomously, with ECs responding to ALS-related changes and signals derived from their tissue microenvironment. Notably, axonal degeneration was minimal at the onset of vascular abnormalities in TDP-43 Q331K mice ( Suppl. Fig. 2o-t ), underscoring the high sensitivity of the nerve endothelium to ALS-associated pathology. DISCUSSION Peripheral nerve abnormalities —including axonal injury, defects in retrograde transport, neuromuscular junction alterations, and neuroinflammation— have long been considered central features of disease progression 47 , 48 , 111 – 113 . In this study, we identify the BNB as both a sensor and an active facilitator of ALS progression. We demonstrate that disease-associated alterations in nerve ECs are not only central to the pathophysiology but also reversible through targeted modulation of immune cell infiltration, opening potential therapeutic avenues to delay disease progression. Here, we provide evidence that nerve ECs not only react to early peripheral pathology but also may actively propagate disease signals, thereby amplifying neuroinflammation and axonal degeneration. In addition, we delineate EC-specific molecular signatures that may serve as biomarkers of BNB dysfunction in ALS and potentially in other PNS disorders. Our findings align with the concept that ECs are sensitive to perturbations in both systemic circulation and their local tissue environment 114 , and mount rapid responses that can become maladaptive in chronic disease states 7 , 115 . The peripheral nerve vasculature is a gateway for motor neuron degeneration Diagnostic nerve biopsies obtained from living patients —unlike postmortem specimens, which have typically been used to study ALS-related vascular changes in the brain and spinal cord— provid a unique window into the early phases of disease 51 , 116 . Unexpectedly, in motor nerves we observed vascular alterations that preceded overt axonal loss, supporting the early involvement of BNB dysfunction in ALS pathogenesis. The vulnerability of the PNS vasculature is supported by findings of BNB alterations in vasculitic and inflammatory neuropathies 58 , 59 , 117 , 118 , as well as in chemotherapy-induced peripheral neuropathy, where excessive vasoconstriction in nerve blood vessels causes hypoxia and pain 119 . In line with these observations, BNB disruption in presymptomatic ALS mice results in leakage of plasma proteins such as fibrinogen, which impairs axonal regeneration and myelination 120 and activates innate immune pathways that exacerbate neuronal injury 121 , 122 . BBB and BSCB disruption in ALS has been studied mainly through histology and permeability assays, while transcriptional characterization has been lacking 14 , 19 , 20 , 123 . By profiling ECs from spinal cord and peripheral nerves, we observed minimal transcriptional changes in spinal cord even at symptomatic stages, when astro- and microgliosis are already evident. In contrast, peripheral nerve ECs showed robust alterations already at presymptomatic phases. Notably, nerve ECs a higher “activated” tone at baseline compared to spinal cord ECs, which may predispose them to earlier and stronger pathological responses. Under ALS conditions, this baseline priming becomes amplified, with nerve ECs showing increased expression of inflammatory mediators such as VCAM1 and ICAM1, which are broadly implicated in vascular inflammation 76 , 124 , 125 . By comparison, neither VCAM1, ICAM1, nor PLVAP —a marker linked to increased transcellular permeability— are induced in spinal cord ECs of mutant mice, highlighting the selective vulnerability of peripheral nerve vasculature. Consistent with this, gene ontology analyses highlighted enrichment of NF-κB and TNF-α signaling pathways, which are major regulators of EC activation 76 , 126 , 127 . Notably, endothelial-specific inhibition of NF-κB has been shown to reduce tissue inflammation 128 , suggesting a promising therapeutic approach for ALS-related vascular dysfunction. Organotypic endothelial identity underlies vulnerability of peripheral nerve vasculature in ALS Single-cell RNA-sequencing studies have revealed extensive heterogeneity among ECs across species and disease states 3 , 77 , 129 . We found that in ALS peripheral nerves, capillary endothelial subtypes are the primary responders to disease: venous capillaries show pronounced upregulation of genes linked to inflammation, activation, and increased permeability, whereas barrier capillaries largely downregulate genes involved in transport and vascular homeostasis. Venous capillaries are known to be susceptible in CNS vascular pathologies. For instance, they are the site of origin for cerebral cavernous malformations 130 , 131 and show the strongest inflammatory and neoangiogenic response and contribute to in EAE 94 , 132 . We identified equivalent populations of capillary venous ECs in both nerve and spinal cord based on shared markers; however, the nerve subset exhibited a distinctive baseline profile, marked by elevated expression of pro-inflammatory and activation-associated genes. In ALS, these nerve venous ECs further upregulated a set of genes (e.g., Plvap , Vwf , Aqp1 , Tacr1, Aplnr ) not induced in the corresponding spinal cord subset during EAE inflammation 94 , revealing a unique, nerve-specific transcriptional response to ALS-related damage. Venous markers VWF and VCAM1, which are elevated in aged CNS vasculature and implicated in EC activation, atherosclerosis, and stroke 114 , 133 – 135 , increased in the pre-symptomatic phase exclusively in motor and sciatic nerve ECs of both SOD1 and TARDBP mutant mice, but remained unchanged in the spinal cord vasculature. A similar pattern was observed for TACR1 (NK1R), which mediates inflammatory signaling and cytokine production 136 , and for LRG1, an IL6-induced secreted glycoprotein involved in cardiovascular disease and innate immunity that promotes neutrophil chemotaxis and fibrosis 137 . Collectively, these findings highlight the activated state of the nerve endothelium consistent with the organotypic features of the capillary–venous EC subset and its selective vulnerability in ALS. Enhanced transcellular permeability underlies early BNB disruption The two cardinal mechanisms of barrier maintenance —transcellular and paracellular transport— appear differentially sensitive to pathological stress. Our data indicate that enhanced transcellular route, rather than TJ disruption, is the primary driver of BNB breakdown in presymptomatic ALS, with structural alterations in EC contacts becoming evident only at more advanced stages. In pre-symptomatic ALS models, focal PLVAP activation in nerve vasculature marks early vascular insult. This is accompanied by selective downregulation of MFSD2A, a key inhibitor of caveolae-mediated transcytosis, and increased cytoplasmic localization of CAV1, consistent with electron microscopy evidence of elevated vesicular transport. Collectively, these changes confirm a pathological surge in transcytosis preceding junctional breakdown. Similar dynamics are observed in stroke models, where transcellular leakage precedes TJ disruption 138 . Notably, PLVAP+ vessels still retain tight junction markers such as CLDN5, and we observe ECs in a hybrid PLVAP − /CLDN5 + state, suggesting a partial transition toward a non-barrier phenotype rather than complete transcriptional reprogramming. This aligns with evidence of a reduction in Wnt tone, a key driver of barrier induction and maintenance 72 , 139 , 140 , despite the pathway remain active —as evidenced by LEF1 expression in ALS ECs (data not shown)— indicating that ECs retain core barrier identity while selectively deregulating specific molecular components that perturb BNB integrity. Together, these observations establish a mechanistic framework in which early, graded dysregulation of transcellular permeability underlies selective BNB vulnerability in ALS. This early response of the transcytotic pathway provides a potential window for therapeutic intervention. Temporary modulation of transcellular permeability could facilitate targeted delivery of neuroprotective factors 60 , and strategies such as overexpression of the transcytosis suppressor MFSD2A or inhibition of vesicular transport have demonstrated efficacy in rescuing BBB alterations in brain disorders 141 – 143 . Endothelial–immune crosstalk drives peripheral nerve inflammation in ALS Endothelial–immune interactions actively shape tissue inflammation and disease progression in neurological disorders, including ALS 144 , 145 . In ALS nerves, endothelial activation drives early upregulation of adhesion molecules and inflammatory mediators, facilitating immune cell recruitment. Recruited immune cells then feed-back on ECs, amplifying barrier dysfunction and sustaining inflammation 106 , 146 . High-dimensional flow cytometry revealed expansion of neutrophils, inflammatory monocytes, and activated macrophages in the peripheral nerves of SOD1 G93A mice, paralleling EC transcriptional shifts toward a pro-inflammatory state. Resident macrophages, essential for maintaining vascular homeostasis in the PNS 60 , acquire an activated phenotype in ALS, consistent with previous reports 53 , 54 , 57 . Our analysis revealed that resident macrophages, normally perivascular in healthy nerve, transition from a homeostatic to an activated state in situ, without recruitment of circulating monocytes. Activated macrophages detach from vessels, become phagocytic, and undergo an inflammatory switch, accompanied by relocalization within the endoneurium. This process may be driven by downregulation of CX3CL1 on endoneurial ECs, a ligand for CX3CR1 expressed by resident macrophages, and a genetic modifier in ALS 147 . Nerve EC activation was accompanied by substantial infiltration of neutrophils that modulate the activation state of macrophage populations. Neutrophils localize near degenerating axons and vessels, amplifying vascular dysfunction and inflammation 106 , and compromising barrier integrity via protease and ROS release and downregulation of TJ proteins such as CLDN5 148 . Neutrophil depletion restored BNB integrity and alleviated axons degeneration, without however improving motor symptoms. These results highlights neutrophils as key modulators of ALS pathology in peripheral nerves and suggests that therapeutic strategies targeting neutrophils, alone or combined with interventions to normalize EC function, may offer a promising avenue 55 , 149 . Non–cell-autonomous ALS signals trigger endothelial activation Mutant TDP-43 can cell-autonomously regulate angiogenesis and barrier function, and its overexpression in brain ECs induces neuroinflammation and BBB alterations 37 – 39 , 150 . However, pTDP-43 aggregates in the spinal cord do not consistently correlate with BSCB leakage 32 , leaving it unclear whether vascular dysfunction arises from direct effects in ECs or indirectly via signals from neurons, glia, or other surrounding cells. In our study, EC activation in the TDP-43 Q331K model occurs in motor —but not sensory— nerves prior to detectable axonal degeneration, even in the absence of mutant protein in ECs. These observations indicate that ECs respond non–cell autonomously to ALS-associated signals, likely originating from vulnerable motor axons or the surrounding microenvironment in a pre-degenerative state. In SOD1 G93A mice, where all cells express the mutant protein, vascular pathology is quantitatively more severe although all hallmarks of vascular dysfunction are shared across these two models, suggesting that both cell-autonomous and non–cell autonomous mechanisms contribute to peripheral EC activation and BNB disruption in ALS. Together, our data indicate that peripheral nerve ECs act as sensors of early ALS-related damage, integrating pathological cues from the tissue microenvironment and mounting maladaptive responses. Elucidating the nature of these activating signals and the relative contributions of intrinsic and extrinsic mechanisms to endothelial dysfunction in the PNS remains an important open question, with direct relevance for defining pathological pathways and identifying potential windows for therapeutic intervention. Author contributions G.P.B. designed experiments, collected and analyzed data with the help of A.B., A.M., I.B., and L.M. S.D. performed computational analysis with the help of SdP. C.B. performed high-dimensional flow cytometry and data analysis. P.P. conducted nerve morphology and electron microscopy studies. D.Bos. performed cell sorting. G.N. and C.B. provided nerve samples from mutant mice. N.R. selected and analyzed clinical cases. M.I. provided critical comments for immunological studies. A.Q. collected human nerve samples with the help of N.R, and analyzed morphological data. D.B. supervised the study, analyzed data and wrote the manuscript with G.P.B. MATERIALS AND METHODS Sample size and Statistical analysis Statistical significance is reported as follows: ns, not significant (p >0.05); * p <0.05; ** p <0.01; *** p <0.001; **** p <0.0001. Fig. 1j: N, TEM images (≥3 from each motor nerve biopsy). One-way ANOVA with Dunnett’s test, p<0.001 Mild (includes 2 ALS “normal” and 4 “mild” cases) vs. Control (3 non-ALS neuropathies/myopathy cases); p<0.0001 Severe (includes 2 ALS “moderate” and 4 “severe” cases) vs. Control. Fig. 1k: N, TEM images from motor nerves, control (8 nerves), SOD1 G93A P60 (4), P90 (6), P120 (8). One-way ANOVA with Dunnett’s test, (ns) p=0.74 SOD1 G93A P60 vs. control; p<0.0001 SOD1 G93A P60 and P120 vs. control. Fig. 1l: N, TEM images from motor nerves, control (9 nerves P60-P180), TDP-43 Q331K P60 (2), P120 (4), P180 (48). One-way ANOVA with Dunnett’s test, p<0.0001 TDP-43 Q331K P60, P120, and P180 vs. control. Suppl. Fig. 2e, f: N, sections/mice: control (9/7 mice at P60-P120), SOD1 G93A P60 (9/4), P90 (11/5), P120-150 (6/4). One-way ANOVA with Dunnett’s test. Total fiber counts ( e ), (ns) p=0.24 SOD1 G93A P60 vs. control; p<0.05 SOD1 G93A P90 vs. control; p<0.001 SOD1 G93A P120-150 vs. control. Degenerating fibers/total ( f ), (ns) p=0.66 SOD1 G93A P60 vs. control; p<0.0001 SOD1 G93A P90 and P120-150 vs. control. Suppl. Fig. 2g, h: N, mice: 16-26 per genotype at each stage. One-way ANOVA with Sidak’s test. Rotarod test ( g ), (ns) p=0.94 SOD1 G93A vs. control at P60; (ns) p=0.16 SOD1 G93A vs. control at P90; p<0.0001 SOD1 G93A vs. control at P120. BDW measurement ( h ), (ns) p=0.42 SOD1 G93A vs. control at P60; (ns) p=0.43 SOD1 G93A vs. control at P90; p<0.0001 SOD1 G93A vs. control at P120. Suppl. Fig. 2s, t: N, sections/mice: control (13/6 mice at P60-P120), TDP-43 Q331K P60 (16/6), P120 (6/4), P180 (6/6). One-way ANOVA with Dunnett’s test. Total fiber counts ( s ), (ns) p=0.79 TDP-43 Q331K P60 vs. control; (ns) p=0.45 TDP-43 Q331K P120 vs. control; p<0.01 TDP-43 Q331K P180 vs. control. Degenerating fibers/total ( t ), p<0.01 TDP-43 Q331K P60 vs. control, p<0.001 TDP-43 Q331K P120 and P180 vs. control. Fig. 2f: N, tiled regions of interest (ROIs) spanning the sciatic nerve parenchyma from 5-6 mice per genotype/stage. Control values are pooled from P60-P120. One-way ANOVA with Dunnett’s test, p<0.05 SOD1 G93A P60 vs. control; p<0.0001 SOD1 G93A P90 and P120 vs. control. Fig. 2g: N, tiled ROIs spanning the spinal cord (4 per spinal cord) from 5-6 mice per genotype/stage. Control values are pooled from P60-P120. One-way ANOVA with Dunnett’s test, (ns) p=0.98 SOD1 G93A P60 vs. control; (ns) p=0.74 SOD1 G93A P90 vs. control; p<0.0001 SOD1 G93A P120 vs. control. Suppl. Fig. 3o: N, spinal cord images (∼10 per spinal cord) from 5-6 mice per genotype/stage. Control values are pooled from P60-P150. One-way ANOVA with Dunnett’s test, (ns) p=0.67 SOD1 G93A P60 vs. control; (ns) p=0.22 SOD1 G93A P90 vs. control; p<0.0001 SOD1 G93A P150 vs. control. Suppl. Fig. 3p, q: N, tiled ROIs spanning spinal cords from 6 control mice (P60-P150), 2 SOD1 G93A P60, 2 SOD1 G93A P90, and 3 SOD1 G93A P150. One-way ANOVA with Dunnett’s test. GFAP staining ( p ), (ns) p=0.58 SOD1 G93A P60 vs. control; p<0.0001 SOD1 G93A P90 and P120 vs. control. IBA1 staining ( q ), (ns) p=0.74 SOD1 G93A P60 vs. control; p<0.01 SOD1 G93A P90 vs. control; p<0.0001 SOD1 G93A P150 vs. control. Suppl. Fig. 3v: N, spinal cord vessels from 5-8 mice per genotype/stage. Control values are pooled from P60-P120. One-way ANOVA with Dunnett’s test, (ns) p=0.89 SOD1 G93A P60 vs. control; p<0.01 SOD1 G93A P90 vs. control; p<0.0001 SOD1 G93A P120 vs. control. Suppl. Fig. 3w: N, sciatic nerve vessels from 7-12 mice per genotype/stage. Control values are pooled from P60-P120. One-way ANOVA with Dunnett’s test, (ns) p=0.12 SOD1 G93A P60 vs. control; (ns) p=0.82 SOD1 G93A P90 vs. control; (ns) p=0.43 SOD1 G93A P120 vs. control. Fig. 3j, k: N, vessels from tiled ROIs spanning sciatic nerves from 5-6 mice per genotype/stage. Control values are pooled from P60-P120. One-way ANOVA with Dunnett’s test. PLVAP staining ( j ), p<0.0001 SOD1 G93A P60, P90 and P120 vs. control. CLDN5 staining ( k ): (ns) p=0.72 SOD1 G93A P60 vs. control; (ns) p=0.079 SOD1 G93A P90 vs. control; p<0.0001 SOD1 G93A P120 vs. control. Fig. 3l, m: N, spinal cord vessels from 8 control and 9 SOD1 G93A mice. Unpaired t test. PLVAP staining ( l ), (ns) p=0.56 SOD1 G93A vs. control. CLDN5 staining ( m ): p<0.0001 SOD1 G93A vs. control. Fig. 3p: N, tiled regions ROIs spanning the sciatic nerve parenchyma from 4 control and 5 SOD1 G93A mice. Unpaired t test, p<0.0001 SOD1 G93A vs. control. Fig. 3s: N, tiled regions ROIs spanning the nerve parenchyma from 5 control and 6 SOD1 G93A mice. Pooled measurements from sciatic and motor nerves. Unpaired t test, p<0.0001 SOD1 G93A vs. control. Fig. 3w, x: N, vessels from tiled ROIs spanning sciatic nerves from 3 control and 4 SOD1 G93A mice. Unpaired t test, p<0.0001 SOD1 G93A vs. control. Fig. 3y: N, nerve biopsies from control (2) and ALS (3) cases. p<0.01 ALS vs. control. Suppl. Fig. 4c: N, vessels from tiled ROIs spanning sciatic nerves from 6 mice per genotype/stage. Unpaired t test, p<0.0001 SOD1 G93A vs. control. Suppl. Fig. 4f: N, vessels from tiled ROIs spanning sciatic nerves from 4 mice per genotype/stage. Unpaired t test, (ns) p=0.22 SOD1 G93A vs. control. Suppl. Fig. 4i: N, vessels from tiled ROIs spanning sciatic nerves from 4-5 mice per genotype/stage. Control values are pooled from P60-P190. One-way ANOVA with Dunnett’s test, (ns) p>0.99 SOD1 G93A P60 vs. control; (ns) p=0.73 SOD1 G93A P90 vs. control; p<0.0001 SOD1 G93A P140 vs. control. Suppl. Fig. 4l: N, ROIs spanning the sciatic nerve parenchyma from 4-5 mice per genotype/stage. One-way ANOVA with Dunnett’s test, (ns) p=0.07 P60 vs. control; p<0.0001 SOD1 G93A P90 and P140 vs. control. Suppl. Fig. 4o: N, tiled regions ROIs spanning the sciatic nerve parenchyma from 3 mice per genotype. Unpaired t test, p<0.0001 SOD1 G93A vs. control. Suppl. Fig. 4r: N, spinal cord vessels from 3 mice per genotype. Unpaired t test, (ns) p=0.06 SOD1 G93A vs. control. Suppl. Fig. 4v: N, vessels (18-26) from 4-5 mice per genotype/stage. Suppl. Fig. 4y: N, vessels (23) from 2 wild-type mice. Fig. 4s, t: N, CLDN5+ vessels (>100) from 4-5 mice per genotype/stage. Controls are pooled from different stages. Suppl. Fig. 5g, h: N, tiled regions ROIs spanning the sciatic nerve parenchyma from 4-5 mice per genotype/stage. One-way ANOVA with Sidak’s test. (ns) p>0.9 CAR4 − vs. CAR4 + control; p<0.0001 CAR4 − vs. CAR4 + SOD1 G93A ; p<0.001 CAR4 − vs. CAR4 + TDP-43 Q331K . Fig. 5a, b: N, vessels from tiled ROIs spanning sciatic nerves from 5 control and 4 SOD1 G93A mice at P60, and 4 control and 6 SOD1 G93A mice at P120. One-way ANOVA with Sidak’s test, p<0.0001 SOD1 G93A vs. control. Fig. 5g: N, vessels from tiled ROIs spanning sciatic nerves from 7-8 mice per genotype/stage. Control values are pooled from different stages. One-way ANOVA with Dunnett’s test, p=0.0035 SOD1 G93A P60 vs. control; p<0.0001 SOD1 G93A P90 and P120 vs. control. Fig. 5l: N, reporter-labeled cells (MF) counted in 55-80 fields of view spanning sciatic nerves from 3-4 mice per genotype/stage. Control values are from P60. Statistical annotations are color-coded by cell labeling. Mixed-effect analysis with Tukey’s test, SOD1 G93A P60 vs. control; SOD1 G93A P90 vs. P60; SOD1 G93A P120 vs. P90. Fig. 5m: N, 3 independent sciatic nerve preparations per genotype, each containing nerves pooled from 2–4 mice. Unpaired t test, p<0.01 SOD1 G93A vs. control. Fig. 5o: N, 4 independent sciatic nerve preparations per genotype, each containing nerves pooled from 2–4 mice. Unpaired t test, (ns) p=0.14 DC SOD1 G93A vs. control; p<0.0001 MF3 SOD1 G93A vs. control; p<0.01 other cell clusters SOD1 G93A vs. control. Fig. 5q: N, sciatic nerves from 6 mice per genotype. Data pooled from 3 independent experiments. Unpaired t test, p<0.0001 rMF SOD1 G93A vs. control; p<0. 01 aMF SOD1 G93A vs. control; p<0.001 monocytes SOD1 G93A vs. control. Fig. 5s: N, sciatic nerves from 4 mice per genotype. One-way ANOVA with Tukey’s test for indicated pairwise comparisons. Fig. 5z: N, vessels from tiled ROIs spanning sciatic nerves from 3-4 mice per genotype. One-way ANOVA with Sidak’s test, (ns) p=0.8 PLVAP SOD1 G93A ; Ccr2 -/- vs. SOD1 G93A ; (ns) p=0.31 VCAM1 SOD1 G93A ; Ccr2 -/- vs. SOD1 G93A . Suppl. Fig. 7h: N, spinal cord vessels from 3 mice per genotype. Unpaired t test, (ns) p=0.32 SOD1 G93A vs. control. Suppl. Fig. 7k: N, spinal cord vessels from 3 control and 4 SOD1 G93A mice. Unpaired t test, (ns) p=0.52 SOD1 G93A vs. control. Suppl. Fig. 7q: N, vessels from tiled ROIs spanning sciatic nerves (6 control, 7 SOD1 G93A ). Unpaired t test, p<0.0001 SOD1 G93A vs. control. Suppl. Fig. 7t: N, reporter-labeled cells (MF) counted in 22-36 fields of view spanning motor nerves (3 control, 4 SOD1 G93A ). Statistical annotations are color-coded by cell labeling. Mixed-effect analysis with Sidak’s test, SOD1 G93A vs. control. Suppl. Fig. 8c, e: N, sciatic nerves from 5 mice per genotype. Unpaired t test, SOD1 G93A vs. control. Suppl. Fig. 8j: N, fields of view for MF counts, spanning sciatic nerves from 5 control, 5 SOD1 G93A , and 6 SOD1 G93A ; Ccr2 -/- mice. One-way ANOVA with Sidak’s test, p<0.01 SOD1 G93A ; Ccr2 -/- vs. SOD1 G93A . Fig. 6b: N, sciatic nerves from 11 control and 7 SOD1 G93A mice. Data pooled from 3 independent experiments. Unpaired t test, p<0.01 SOD1 G93A vs. control. Fig. 6d: N, Ly-6B.2 (7/4)-FITC + neutrophils counted in 35-53 fields of view spanning motor nerves from 4-6 mice per genotype/stage. One-way ANOVA with Dunnett’s test, (ns) p=0.103 SOD1 G93A P60 vs. control; p<0.0001 SOD1 G93A P90 and P120 vs. control. Fig. 6o: N, Ly-6B.2 (7/4)-FITC + (or Ly6G + ) neutrophils counted in 19-31 fields of view spanning sciatic nerves from 8 control and 11 SOD1 G93A /isotype or anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; p=0.07 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Fig. 6p: N, vessels from tiled ROIs spanning sciatic nerves from 8 (control) and 11 ( SOD1 G93A /isotype or anti-Ly6G-treated) mice. One-way ANOVA with Tukey’s test, p<0.0001 all pairwise comparisons. Fig. 6q: N, fields of view for MF counts, spanning sciatic nerves from 8 control and 11 SOD1 G93A /isotype or anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 all pairwise comparisons. Fig. 6r: N, vessels from tiled ROIs spanning sciatic nerves from 8 control and 11 SOD1 G93A /isotype or anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; (ns) p=0.0504 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Fig. 6s: N, vessels from tiled ROIs spanning sciatic nerves from 8 control and 11 SOD1 G93A /isotype or anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. control; p<0. 001 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Fig. 6t: N, TEM images from sciatic nerves of 9 control, 13 SOD1 G93A /isotype treated, and 17 SOD1 G93A /anti-Ly6G-treated) mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; (ns) p=0.367 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Fig. 6v: N, sciatic nerves from 6 mice per genotype. One-way ANOVA with Tukey’s test. Neutrophils , p=0.013 SOD1 G93A /isotype-treated vs. control; p=0.20 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.001 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated; Monocytes , p=0.01 SOD1 G93A /isotype-treated vs. control; p<0.001 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.25 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated; MF1 , p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated; MF3 , p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.001 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.001 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Fig. 6x, y: N, motor nerves from 9 control, 13 SOD1 G93A /isotype treated, and 17 or SOD1 G93A /anti-Ly6G-treatedmice. One-way ANOVA with Tukey’s test. Total fiber counts ( x ), p<0.001 SOD1 G93A /isotype-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.32 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Degenerating fibers/total ( y ), p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.001 SOD1 G93A /anti-Ly6G-treated vs. control; p=0.0014 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9c: N, Ly-6B.2 (7/4)-FITC + neutrophils counted in 35-46 fields of view of spinal cords from 4 control and 6 SOD1 G93A mice. Unpaired t test, (ns) p=0.302 SOD1 G93A vs. control. Suppl. Fig. 9f: N, sciatic nerves from 6 control, 7 SOD1 G93A /isotype treated, and 8 SOD1 G93A /anti-Ly6G-treated mice. Data pooled from 2 independent experiments. One-way ANOVA with Tukey’s test, p<0.01 SOD1 G93A /isotype-treated vs. control; (ns) p=0.66 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.001 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9m: N, Ly-6B.2 (7/4)-FITC + (or Ly6G + ) neutrophils counted in 13-21 fields of view spanning motor nerves from 8 control and 11 SOD1 G93A /isotype or anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9n: N, vessels from tiled ROIs spanning sciatic nerves from 8 control and 11 SOD1 G93A /isotype or anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.01 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9o: N, vessels from tiled ROIs spanning sciatic nerves from 8 control and 11 SOD1 G93A /isotype or anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.43 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9s: N, spinal cord images from 11 control, 11 SOD1 G93A /isotype-treated, and 13 SOD1 G93A /anti-Ly6G-treated) mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9t, u: N, tiled ROIs spanning spinal cords from 6 control, 7 SOD1 G93A /isotype treated, and 8 SOD1 G93A /anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test. GFAP staining ( t ), p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.209 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. IBA1 staining ( u ), p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.82 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9w: N, mice; 16 control, 17 SOD1 G93A /isotype treated, and 18 SOD1 G93A /anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.0001 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.98 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9w: N, mice; 16 control, 17 SOD1 G93A /isotype treated, and 18 SOD1 G93A /anti-Ly6G-treated mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A /isotype-treated vs. control; p0.99 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Suppl. Fig. 9x: N, mice; 18 control, 15 SOD1 G93A /isotype treated, and 18 SOD1 G93A /anti-Ly6G-treated mice. Unpaired t test, p<0.01 SOD1 G93A /anti-Ly6G-treated vs. control at 16 weeks; p<0.05 p0.05 all other pairwise comparisons. Suppl. Fig. 10e: N, sciatic nerves from 7 control, 6 SOD1 G93A /isotype treated, and 6 SOD1 G93A /anti-Ly6G-treated mice. Data pooled from 2 independent experiments One-way ANOVA with Tukey’s test. rMF , p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. aMF , p<0.05 SOD1 G93A /isotype-treated vs. control; (ns) p=0.89 SOD1 G93A /anti-Ly6G-treated vs. control; p<0.05 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Mon , p<0.001 SOD1 G93A /isotype-treated vs. control; p<0.01 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.55 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Fig. 10h: N, sciatic nerves from 6 mice per genotype. One-way ANOVA with Tukey’s test. MF2 , p<0.01 SOD1 G93A /isotype-treated vs. control; p<0.001 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.17 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated; MF4 , p<0.0001 SOD1 G93A /isotype-treated vs. control; p<0.001 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.39 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated; MF5 , (ns) p=0.11 SOD1 G93A /isotype-treated vs. control; (ns) p=0.94 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.066 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated; MF6 , p<0.01 SOD1 G93A /isotype-treated vs. control; (ns) p=0.12 SOD1 G93A /anti-Ly6G-treated vs. control; (ns) p=0.35 SOD1 G93A /anti-Ly6G-treated vs. SOD1 G93A /isotype-treated. Fig. 10l: N, Ly-6B.2 (7/4)-FITC + neutrophils counted in 19-29 fields of view spanning sciatic nerves from 4 control, 6 SOD1 G93A , and 6 SOD1 G93A ; Ccr2 -/- mice. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A vs. control; p<0.0001 SOD1 G93A ; Ccr2 -/- vs. control; (ns) p=0.87 SOD1 G93A ; Ccr2 -/- vs. SOD1 G93A . Fig. 10n: N, sciatic nerves from 4 mice per genotype. One-way ANOVA with Tukey’s test, p<0.0001 SOD1 G93A vs. control; p500 total nuclei counted in sciatic nerve or spinal cord sections from 3 mice per genotype. All EC nuclei observed in the spinal cords (n=122) or sciatic nerves (n=41) of TDP-43 Q331K mice were Myc-Tag − . Fig. 7f: N, technical replicates from 2 independent experiments. Fig. 7o, p, q: N, vessels from tiled ROIs spanning sciatic nerves from 10 mice per genotype. Unpaired t test. ( o ) p<0.0001 ( p ) (ns) p=0.0517 ( q ) p<0.0001 TDP-43 Q331K vs. control. Fig. 7s: N, tiled ROIs spanning the sciatic nerve parenchyma from 4 control and 6 TDP-43 Q331K mice. Unpaired t test, p<0.0001 TDP-43 Q331K vs. control. Fig. 7v: N, vessels from tiled ROIs spanning motor nerves from 5 control and 6 TDP-43 Q331K mice. Unpaired t test, p<0.0001 TDP-43 Q331K vs. control. Fig. 7y: N, Ly-6B.2 (7/4)-FITC + neutrophils counted in 29-37 fields of view spanning sciatic and motor nerves from 5 control and 6 TDP-43 Q331K mice. Unpaired t test, p<0.0001 TDP-43 Q331K vs. control. Suppl. Fig. 11e: N, fields of view for MF counts, spanning sciatic nerves from 3 control and 4 TDP-43 Q331K mice. Unpaired t test, p<0.0001 TDP-43 Q331K vs. control. Suppl. Fig. 11g, h: N, vessels from tiled ROIs spanning sciatic nerves from 7 control and 10 TDP-43 Q331K mice. Unpaired t test, p<0.0001 TDP-43 Q331K vs. control. Suppl. Fig. 11n: N, tiled ROIs spanning the sciatic nerve parenchyma from 4 control and 6 TDP-43 Q331K mice. Unpaired t test, (ns) p=0.26 TDP-43 Q331K vs. control. Suppl. Fig. 11q, r: N, tiled ROIs spanning spinal cords from 4 control and 6 TDP-43 Q331K mice. Unpaired t test, p<0.0001 TDP-43 Q331K vs. control. Suppl. Fig. 11s: N, tiled ROIs spanning spinal cords from 7 control and 5 TDP-43 Q331K mice. Unpaired t test, (ns) p=0.78 TDP-43 Q331K vs control. Suppl. Fig. 11y, z: N, vessels from tiled ROIs spanning sciatic nerves from 2 control, 2 TDP-43 Q331K , and 4 Cdh5 CreER ; TDP-43 Q331K . One-way ANOVA with Tukey’s test. ( y ) p<0.0001 TDP-43 Q331K vs. control; p<0.0001 Cdh5 CreER ; TDP-43 Q331K vs. control; (ns) p=0.12 Cdh5 CreER ; TDP-43 Q331K vs. TDP-43 Q331K . ( z ) p<0.0001 TDP-43 Q331K vs. control; p<0.0001 Cdh5 CreER ; TDP-43 Q331K vs. control; (ns) p=0.061 Cdh5 CreER ; TDP-43 Q331K vs. TDP-43 Q331K . Mice SOD1 G93A (B6.Cg-Tg(SOD1*G93A)1Gur/J; Jax #004435) 69 . SOD1 G93A transgene copy numbers across generations were controlled by quantitative PCR using with SYBR select master mix (Themo, Cat #4472908) on genomic DNA with primers (forward) 5’-GAAACCGCGACTAACAATCAAAGTG-3’ and (reverse) 5’-CATCAGCCCTAATCCATCTGATGC-3’ for SOD1 , and (forward) 5’-CAGGTGGGCTCCAGCATT-3’ and (reverse) 5’-TCACCAAGTGTCATTTCTGCCTTTG-3’ for APOB as a normalizer. Male and female SOD1 G93A mice were classified as presymptomatic (60 ± 5 days), onset (90 ± 5 days), or symptomatic (120 ± 5 days) based on neuroscore 151 , body weight measurements, and behavioral tests, followed by post-mortem motor neuron and axonal fiber counts. This progression is consistent with previous studies 53 . For consistency, mice are referred to as P60, P90, and P120, with the specific age indicated when analyzed at other time points. TDP-43 Q331K (6.Cg-Tg(Prnp-TARDBP*Q331K)103Dwc/J; Jax #017933) 68 were staged as presymptomatic (60 ± 10 days), onset (120 ± 10 days), and symptomatic (180 ± 10 days) 68 , 70 , with both males and females included in the analysis. Cdh5 CreER (Cdh5(PAC)-CreERT2) 110 . R26-Tomato LSL (Ai14; Jax #007914) 152 . Ccr2 RFP Cx3cr1 GFP (Jax #032127) 153 . Ccr2 -/- (Jax #004999) 103 . Tamoxifen (Tmx; Sigma-Aldrich Cat# T5648) was dissolved at 10mg/ml in 1:9 ethanol:corn oil (Sigma-Aldrich Cat# S5007) and administered intraperitoneally (i.p.) to Cdh5 CreER ; TDP-43 Q331K mice immediately after weaning for 7 consecutive days at 100mg/kg body weight (BDW). All mouse lines were maintained on C57Bl6/N strain. Both males and females were analyzed unless otherwise specified. Mice were maintained in pathogen-free facilities under standard housing conditions with continuous access to food and water. All experimental procedures and handling were approved by the Animal Research Committee of IRCCS San Raffaele Hospital and performed in compliance with IACUC guidelines. Behavioral analysis Rotarod test: All mice were trained for 1 week before testing. Mice were placed on an accelerated rotarod (Ugo Basile) from 5 to 40 rotations per minute for a maximum of 300 seconds. A minimum of 5 trials were performed in each session, and the longest latency to fall, normalized to BDW, was recorded 70 . Grip test: aximal force (g) exerted by all four limbs on the grid was measured at least 5 times using a grip strength meter (Ugo Basile). The highest and lowest values were discarded, and the average of the remaining measurements was recorded. BDW was measured every 2 weeks from 2 months of age until the symptomatic stage. For behavioral studies involving neutrophil depletion, only male SOD1 G93A mice were included. Human Biopsy samples The anterior motor branch of the obturator nerve was sampled and processed according to standardized protocols 116 . Motor nerves from non-ALS neuropathy or myopathy cases were used as controls after applying established exclusion criteria 116 . All biopsies were obtained from Institute of Experimental Neurology (INSPE) tissue bank. Patients provided informed consent to the study. All experimental protocols were approved by San Raffaele Scientific Institute Ethical Committee (Milan, Italy). Axonopathy (demyelination and degeneration) was assessed on semithin sections as part of the clinical diagnostic routine. Alterations in endoneurial vessels (including accumulation of pinocytotic vesicles, tight junction changes, cellular hypertrophy and hyperplasia, basal lamina detachment, and microfilament accumulation) were scored blindly on corresponding TEM sections by two independent researchers, and the average score for each feature was calculated. Electron microscopy and morphological studies Mice were sacrificed by cervical dislocation and sciatic nerve, femoral nerves with quadriceps and saphenous nerve branches were collected and fixed with 2% glutaraldehyde in 0.12M phosphate buffer, postfixed with 1% osmium tetroxide, dehydrated with alcohol and embedded in Epon (Fluka). Transverse semithin sections (0.5–1μm) of mouse and human nerve samples were stained with toluidine blue and examined by light microscopy. Ultrathin (100–120 nm) sections were stained with uranyl acetate and lead citrate, and examined by TEM (Talos L120C G2). Immunohistochemistry After dissection, nerves were fixed in 4% paraformaldehyde (PFA) on ice for 2 hours, washed extensively in PBS, cryoprotected in 30% sucrose (in PBS) overnight (ON), embedded in OCT (Sakura Cat# 4583) and cryosectioned at 16-18μm thickness for longitudinal sections and 10μm thickness for transverse sections. For spinal cord analysis, mice were trans-cardiac perfused with saline followed by 4% PFA and dissected and further fixed with 4% PFA ON, cryoprotected with 30% sucrose (in PBS) ON and cryosectioned at 20-25μm thickness. For immunostaining, nerve sections were permeabilized with 0.2% Triton X-100 and blocked in 5% donkey serum diluted in PBS for 1 hr. Spinal cord sections were permeabilized with 0.3% Triton X-100 and blocked in 5% donkey serum diluted in PBS for 1 hr. Primary antibodies were incubated in same solution ON at 4°C. After 3 washes in PBS of 5 min each, secondary antibodies were incubated in 0.1% Triton X-100 and 5% donkey serum diluted in PBS at RT for 2 hrs. After 3 washes in PBS, sections were stained with DAPI and mounted with Fluoromount Aqueous Mounting Medium (Sigma-Aldrich Cat# F4680). For PLVAP, Claudin-5, ZO-1 and VE-cadherin immunostaining, nerves were embedded directly in OCT and frozen on dry ice cooled 2-methylbutane. Sections were fixed with pre-cooled methanol for 5-7 min at 4°C and incubated with primary and secondary antibodies diluted in 1% BSA and 0.01% Triton X-100. For Choline Acetyl Transferase (ChAT) and Myc Immunostaining spinal cords sections were subjected to antigen retrieval with Tris-EDTA buffer for 10 min at 95°C (10mM Tris base, 1 mM EDTA, 0.05% Tween 20 pH 9.0). Samples were routinely imaged with Olympus Fluoview FV3000RS or Leica TCS SP8 SMD FLIM Laser Scanning Confocal microscopes. MAVIG RS-G4 resonant scanning confocal microscope (VivaScope GmbH Research) was used for imaging the entire length of longitudinal nerve sections. Primary antibodies used for immunostaining were: rat anti-CD31 (1:200; BD Pharm Cat# 550274), goat anti-CD31 (1:1000; R&D Systems Cat# AF3628), rabbit anti-RFP (1:1000; MBL Cat# PM005), rat anti-PLVAP (1:100; BD Pharm Cat# 553849), rabbit anti-claudin5 (1:400, Thermo Fisher Scientific Cat# 34-1600), goat anti-Podocalyxin (1:400; R&D Systems Cat# AF1556), rat anti-CD144 (1:100; BD Pharm Cat# 550548), rabbit anti-VWF (1:200; Dako Cat# A008229-5), chicken anti-GFP (1:500; Abcam Cat# ab13970), rabbit anti-GFP (1:1000; Thermo Fisher Scientific Cat# A6455), rabbit anti-Fibrinogen (1:10000; Dako Cat# A0080), rabbit anti ZO-1 (1:300; Thermo Fisher Scientific Cat# 40-2300), rat anti-VCAM1 (1:100; BD Pharm Cat# 550547), goat anti-ICAM1 (1:150; R&D Systems Cat# AF796), goat anti-Car4 (1:400; R&D Systems Cat# AF2414), , Rabbit anti-Iba1 (1:500; Wako Cat# 019-19741), rabbit anti-GFAP (1:400; Dako Cat# Z0334), mouse anti-NeuN (1:100; Millipore Cat# MAB377), rabbit anti-ChAT (1:2000; Abcam Cat#AB178850), mouse anti-Myc (1:100; Cell signaling Cat#2276), rat anti-F4/80 (1:250; Abcam Cat# AB6640), rabbit anti-Caveolin (1:300; Cell signaling Cat# 3267), rabbit anti-Collagen IV (1:500; Biorad Cat# 2150-1470), rat anti-Ly6g (1:100; BD Pharm Cat# 551459), rat anti-7/4 FITC (1:200; Abcam Cat# AB53453), rabbit anti-Mfsd2A (1:400, Cell signaling Cat# 80302). Secondary antibodies: goat or donkey anti-rabbit/rat/goat/mouse Alexa Fluor 488-, 555-, 647-conjugated (1:500; Thermo Fisher Scientific). Whole nerve staining and tissue clearing Nerves were fixed overnight in 4% PFA, and the epineurium was manually removed. Nerves were incubated in PBSGT (0.2% gelatin, 0.5% TritonX-100, 0.01% Thimerosal, Sigma-Aldrich Cat# T8784 in PBS) for 2 days at RT on a rocker. Primary antibodies were diluted in PBSGT with 10mg/mL saponin and incubated at RT for 7 days on a rocker. Samples were washed 6 times over one day with PBSGT, then incubated with secondary antibodies in PBSGT with 10mg/mL saponin for 2 days at 37°C on a rocker. After 6 washes in PBSGT, nerves were stored in PBS at 4°C until clearing with X-CLARITY Tissue Clearing System (Logos Biosystems). Samples were incubated in Hydrogel Solution (Logos Bio-systems Cat# C1310X) for 24 hours at 4°C, transferred in X-CLARITY Polymerization System for 3 hrs at 37°C, then washed several times in Electrophoretic Tissue Clearing Solution (Logos Biosystems Cat# C13001). Nerves were clarified using the X-CLARITYTissue Clearing System II (Logos Biosystems C30001) to remove lipids from the tissue, applying a current of 0.6A at 37°C for 4 hours. Samples were washed extensively in PBS and incubated in X-Clarity Mounting Solution (Logos Bio-systems Cat# C13101) for 1 hour at RT before imaging with a Nikon A1 MP multiphoton microscope with Coherent Chameleon Ultra II using Apo LWD 25 X (NA 1.1) Water Plan objective. 3D-image stacks were analyzed with Arivis 4D software. Primary antibodies used for whole nerve staining were a mix of goat anti-CD31 (1:500; R&D Systems Cat# AF3628) and goat anti-Podocalyxin (1:1000; R&D Systems Cat# AF1556). Secondary antibodies (from Thermo Fisher Scientific) were donkey anti-goat AlexaFluor-555 (1:1000; Cat# A21432). Vascular permeability assays Adult mice received 200μl of EZ-Link Sulfo-NHS-LC-Biotin (Thermo Fisher Scientific Cat# 21335) (20mg/ml in PBS) via tail vein injection and were sacrificed after 30 min to allow circulation of the tracer. Nerves and spinal cords were harvested for immunostaining and the dye was detected using AlexaFluor 555- or 647-streptavidin (1:500; Thermo Fisher Scientific Cat# S32355, S32357). Mice that received BSA-FITC (Invitrogen Cat# A23015) (1.5mg/ml in PBS; 10μl/g BDW) via tail vein injection were sacrificed after 2-3 hours for nerve harvesting 60 . Mice tail vein injected with HRP (Sigma Aldrich Cat# P8250; 500mg/kg BDW in PBS) were sacrificed after 30 min. HRP was detected on nerve sections using TSA Fluorescein system (Akoya Biosciences Cat# NEK701A001KT). Endothelial cell isolation from sciatic nerve Sciatic nerves were harvested in DMEM/F-12 medium (Gibco Cat# 11039021). The epineurium was manually stripped before dissociation for bulk RNA-seq in order to enrich for endoneurial ECs, while it was not removed for scRNA-seq to capture EC from the entire nerve vasculature. Nerves were pooled by genotype, minced on ice in DMEM/F-12 using a razor blade and rotated for 30 min at 37°C in dissociation buffer [Collagenase Type II (50ml/ml from 200mg/ml stock; Gibco Cat# 17101015), Hyaluyronidase (1:50 from 1% stock; Sigma-Aldrich Cat# H3506), DNase I (2.5μg/ml; Sigma-Aldrich Cat# DN25) in DMEM/F- 12] with trituration by pipetting every 10 min. The reaction was blocked with 5% FBS/2mM EDTA in PBS w/o Ca2 + /Mg2 + . The cell suspension was passed through a 70mm cell strainer and centrifuged at 300 xg for 10min (4°C). The pellet was washed twice in FACS buffer [5% FBS in PBS w/o Ca2 + /Mg2 + ], resuspended in 0.5% BSA/PBS w/o Ca2 +/ Mg2 + and incubated with myelin removal beads (Miltenyi Biotech Cat# 130-096-731) for 15 min, and passed through LS columns (Miltenyi Biotech Cat# 130-042-041) to remove myelin debris. The flowthrough cell suspension was collected and stained with rat anti-CD31 APC (1:50; BD Pharmingen Cat#551262) and rat anti-CD45.2 V405 (1:250; BD Biosciences Cat# 560697) at 4°C for 30 min with intermittent tapping. Samples were washed twice with FACS buffer and resuspended in the same buffer. ECs (CD31 + /CD45 − ) were sorted with BD FACSAria Fusion flow cytometer into 400μl of Trizol-LS (Thermo Fisher Scientific Cat # 10296010) for bulk RNA-seq or DMEM/F-12 medium for scRNA-seq. EC isolation from spinal cord Spinal cords were dissected from the same mice used for sciatic nerve EC isolation described above and pooled by genotype. Each spinal cord was chopped on ice into 3-4 pieces in DMEM/F-12 medium. Tissue dissociation was performed with the Miltenyi Adult Brain Dissociation kit (Cat# 130-107-677). Samples were incubated with the dissociation enzyme mix, placed on Miltenyi gentleMACS Octo dissociator for 30 min, and passed through a 70μm cell strainer. After red blood cell removal, samples were resuspended in 2% BSA/PBS w/o Ca2 + /Mg2 + and stained with with anti-CD31 and anti-CD45.2 antibodies for FACS-purification of CD31 + /CD45 − ECs as indicated above. ECs were sorted in 400μl of Trizol-LS for total RNA isolation and bulk RNA-seq. Bulk RNA-sequencing and bioinformatic analysis For each replicate, 6-7 sciatic nerves and 3-4 spinal cords were pooled by genotype and dissociated as described above: Control (non-transgenic) P90 sciatic nerves (3 independent replicates), SOD1 G93A P90 sciatic nerves (3 replicates), Control P120 sciatic nerves (3 independent replicates), SOD1 G93A P120 sciatic nerves (3 replicates), Control P90 spinal cords (3 replicates), SOD1 G93A P90 spinal cords (3 replicates), Control P120 spinal cords (4 replicates), SOD1 G93A P120 sciatic nerves (4 replicates). An average of 20,000 ECs were purified by FACS for each replicate. Total RNA was prepared as per manufacturer’s instructions and resuspended in 10-12ml molecular-grade water. After quantification and quality control with Qubit 3.0 Fluorometer (Thermo Fisher Scientific) and Agilent 2100 Bioanalyzer, RNA-seq libraries were generated using the SMART-Seq Ultra Low Input RNA Kit (Illumina) and sequenced on an Illumina NovaSeq6000 system (single-end, 100bp read length). Sequencing reads were trimmed using Trimmomatic (v0.39) 154 to remove sequencing adapters and exclude low-quality reads, and aligned to the mm10 reference genome using STAR aligner (v2.5.3a) 155 . Transcript quantification (total counts per gene) was performed using FeatureCounts (v1.6.4) 156 based on mapping to exons annotated in GENCODE Mus musculus basic gene annotation (version M16), excluding multimapping reads. For each comparison, only genes with average expression greater than 3 RPKM (Reads Per Kilobase per Million mapped reads) across samples were retained for downstream analysis. Differential gene expression analysis was conducted using the DESeq2 package (v1.30.1) 157 . To visualize gene expression patterns, RPKM values were transformed to z-scores across samples using the pheatmap R package. GO analysis of DEGs (padj ≤0.05) was performed using Enrichr 158 . Enriched terms (padj <0.01) were selected across relevant gene-set libraries. The resulting list was further curated to prioritize the most significant and representative processes, excluding redundant or overly generic terms. GSEA was performed using all expressed genes (cpm > 1) as input, with the clusterProfiler package (v4.2.2) 159 . Enrichment, volcano, GSEA and expression Dot plot were generated with ggplot2 (v4.2.3). EWCE (Expression Weighted Cell-type Enrichment) package (v1.2.0) 160 was applied to DEGs showing consistent changes at P90 and P120 using the integrated control/ SOD1 G93A EC nerve scRNAseq dataset as a reference. Cell-type abundance was inferred from bulk RNA-seq results with CIBERSORTx19 161 using raw count matrices as input. For spinal cord samples, reference profiles were derived from the Brain Tabula Muris ( https://figshare.com/articles/dataset/5829687 ). For sciatic nerve samples, reference signatures were obtained from an in-house sciatic nerve single-cell atlas. Single-cell sequencing and bioinformatic analysis Approximately 20 intact sciatic nerves were pooled for each group (Control vs. SOD1 G93A and dissociated as described above. Single-cell suspensions were stained as indicated above with a mix of antibodies to label ECs (rat anti-CD31 APC; 1:50) along with myeloid cells (rat anti-CD45.2 V405; 1: 400), mesenchymal cells (rabbit anti-NG2 Alexa Fluor488; Millipore Cat# AB5320A4; 1:200), and Schwann cells (mouse anti O4 PE; Miltenyi Biotec Cat# 130-117-357; 1:250). All CD31⁺ ECs were sorted, along with a proportion of other labeled and unlabeled cells, to ensure representative inclusion of various nerve cell types without over-representing the most abundant population, particularly Schwann cells. Sorted cells were counted manually and with an automatic cell counter, and resuspended in 2% FBS/DMEM/F-12 at a concentration of 1000 cells/ml. Cells were subsequently encapsulated and processed using the Chromium platform (10X Genomics) with the Chromium Single Cell 3’ Library & Gel Bead Kit v3. After quality controls and quantification on TapeStation instrument (Agilent), libraries were sequenced on Illumina NovaSeq6000 system. From 3 independent cell purification and scRNA-seq experiments, we generated 2 Control and 4 SOD1 G93A samples. Raw sequencing data were demultiplexed using the mkfastq application from CellRanger (v6.1.2, 10x Genomics). Sequencing libraries were aligned to the 10x Genomics default mm10 reference genome (version 2020-A) using the CellRanger pipeline 162 . The resulting raw gene expression matrices were processed using the Seurat R package (v5.0.0) 163 , with default parameters unless otherwise specified. Only cells (unique barcodes) that expressed a minimum of 300 genes (considering only genes detected in at least 5 cells) and contained over 1000 unique molecular identifiers (UMIs) were retained. Cells with more than 15% of total reads mapping to mitochondrial genes were excluded. After filtering, a total of 35,793 single cells were retained, comprising 16,962 cells from non-transgenic control samples and 18,831 cells from SOD1 G93A samples. Following quality control, datasets were integrated using Seurat’s Harmony-based integration (v0.1.1) 164 with SCTransform normalization 165 and batch correction regressing out mitochondrial gene expression to remove technical variation. Principal component analysis was performed (with RunPCA) and the number of PCs was selected after inspection with ElbowPlot (nPCs=50). Cell clustering was performed using FindNeighbors and FindClusters with a resolution of 0.1. Cell clusters were annotated based on expression of canonical marker genes and visualized on UMAP plots with the “FeaturePlot” and DimPlot functions. From this dataset, ECs (3,062 cells) were extracted, re-integrated and re-clustered following the procedure described above with a clustering resolution of 0.5. Rare contaminating cells were removed. Marker genes for each EC subcluster, and DEGs between control and SOD1 G93A samples for each EC subtype cluster were detected using the FindAllMarkers function (thresh.use=0.1, min.pct=0.01, min.diff.pct=-Inf e test.use = "wilcox"). Endoneurial EC selected based on known marker genes 64 were re-integrated and re-clustered with a resolution of 0.2. Marker gene detection, cell-type annotation and DEGs analysis were performed as described above. To assess gene expression patterns, average log-normalized expression values were transformed to z-scores across cell populations using the pheatmap R package. Gene signatures were generated using the AddModuleScore function in Seurat to quantify module expression scores across cells. Single cell data were visualized using UMAP. Alluvial plot was generated using the geom_alluvium function from the ggalluvial R package (v0.12.3) 166 . To compile a single-cell atlas of the sciatic nerve endothelium, Seurat v4.0 workflow was used to integrate 5 previously published scRNA-seq datasets of the intact mouse sciatic nerve 64 , 89 – 92 , along with the control nerve samples obtained in this study. Variable genes (n=2000) were identified in each sample using the FindVariableFeatures function. Integration anchors were computed with FindIntegrationAnchors (dims = 1:30) and used to integrate the datasets via the IntegrateData function. The integrated dataset was then scaled and centered using ScaleData , regressing out unwanted sources of variation such as mitochondrial gene expression and dataset of origin. Cell clustering was performed using FindNeighbors and FindClusters with a resolution of 0.1. ECs annotated based on the expression of marker genes were extracted and re-integrated following the same workflow, using a clustering resolution of 0.2. The contribution of each dataset was as follows: Bhat et al. (514 ECs), Gerber et al. (1544 ECs), Zhao et al. (828 ECs), Toma et al. (380 ECs), Wolbert et al. (488 ECs), this study (1201 ECs). Small cell clusters that could not be annotated with known EC-subtype marker genes were eliminated yielding to a total of 4894 ECs in the atlas. Dendrogram visualization of hierarchical clustering of EC subtypes was performed with pheatmap R package (v1.0.12) 167 using the mean expression values of the top 200 marker for each subtype cluster, ranked by log2 fold change. The reanalysis of scRNA-seq dataset of spinal cord endothelium (GSE210776) 94 was performed with the standard Seurat workflow (v5.0.0). Integrated analysis was carried out using the IntegrateLayers function, with CCAIntegration method. The integrated object was then used to compute a nearest-neighbor graph (FindNeighbors, dims = 1:30, reduction = "integrated.cca"). Cell clustering was performed using FindClusters at resolution 0.1. UMAP was applied for visualization (RunUMAP, dims = 1:30, reduction = "integrated.cca"). Cldn5⁺ ECs were extracted and re-clustered applying the same workflow with 0.15 resolution. To compare the transcriptomic profiles of sciatic nerve vs. spinal cord ECs we performed cross-dataset label transfer using the FindTransferAnchors and TransferData functions (dims = 1:30) from the Seurat package to projected EC subtype annotations from the sciatic nerve EC atlas onto the spinal cord single-cell dataset. RNA purification and quantitative RT-PCR from human motor nerve biopsies Nerve biopsies flash-frozen in liquid nitrogen were stored at the San Raffaele Hospital-INSPE tissue bank. Tissue homogenates were prepared using Precellys Evolution tissue homogenizer equipped with Cryolys Evolution cooling system (Bertin Technologies). Briefly, frozen nerves were transferred into tubes containing beads and 500µl of TRIzol (Invitrogen, Cat# 15596026) and homogenized at 6000 rpm for 4 × 30 s with 60 s pauses. Lysates were centrifuged, and supernatants were collected for RNA extraction following the standard TRIzol protocol. cDNA was synthesized from 0.3-1mg RNA with M-MLV reverse transcriptase (Thermo Fisher Scientific Cat# 28025013) and random primers (Thermo Fisher Scientific Cat# N8080127), diluted to 40μl in molecular biology-grade water and used for real-time quantitative PCR with SYBR Select Master Mix (Thermo Fisher Scientific Cat# 4472913). RT-PCR reactions were performed in triplicate. Primers: MFSD2A forward 5’-CTCCTGGGCCATCATGCTCTC-3’; reverse 5’-GGCCACCAAAGATGAGAAA-3’. GAPDH forward 5’-ATCACCATCTTCCAGGAGCGAG-3’; reverse 5’-GGGCAGAGATGATGACCCTTTTG-3’. Western blotting Mouse nerves were flash-frozen with liquid nitrogen immediately after harvesting, triturated manually with pestels on dry ice, and resuspended in SDS-lysis buffer (150mM NaCl, 50mM TrisHCl pH7.4, 10mM EDTA, 1% Tx-100, 0.1% SDS; 150μl/ nerve) supplemented with protease (Roche Cat# 4693132001) and phosphatase (Thermo Fisher Scientific Cat# 78420) inhibitor cocktails, followed by bath sonication for 30 seconds at full power. Lysates were incubated on a thermomixer at 700 rpm for 30 min at 4°C and centrifuged at 13,000g for 15 min (4°C). Supernatants were collected and quantified with Bradford protein assay (Bio-Rad Cat# 5000006). Protein samples were run on handcast 8-10% SDS-PAGE gels or NuPAGE 4-12% Bis-Tris gels (Thermo Fisher Scientific) and transferred on nitrocellulose membranes (Amersham/Merck Cat# GE10600002). Antibodies used for immunoblotting were: goat mouse anti-Vinculin (1:5000, Millipore Cat# 05-386), mouse anti-Caveolin (1:1000; BD Pharm Cat# 610407), mouse anti-pY14 Caveolin (1:1000; BD Pharm Cat# 611339), rabbit anti-Claudin5 (1:1000, Thermo Fisher Scientific Cat# 34-1600). In vivo neutrophil depletion Neutrophils were depleted by intraperitoneal injection of InVivoMAb anti-mouse Ly6G antibody (clone 1A8; Catalog #BE0075-1, BioXcell) at a dose of 10 mg/kg BDW every four days. Control mice received an equivalent dose of InVivoMAb rat IgG2a isotype control antibody, anti-trinitrophenol (clone 2A3; Catalog #BE0089, BioXcell). Neutrophil depletion was confirmed by flow cytometry analysis of peripheral blood collected from the retro-orbital sinus of anesthetized mice the day before and the day after antibody administration. Blood was drawn into sodium heparin-coated hematocrit capillaries, and red blood cells were lysed using ACK Lysing Buffer. The resulting single-cell suspensions were washed twice with PBS and used for subsequent staining and analysis. Only males were used for this experiment to ensure consistency in body weight measurement and behavioral tests across groups. Immune cell isolation and flow cytometry Single-cell suspensions from mouse sciatic nerves were prepared as described above. For flow cytometry staining, cells were plated at a maximum density of 2×10 6 cells per well in a 96-well U-bottom tissue culture plate. Cell viability was assessed with Viability 405/520 fixable dye (Miltenyi, Cat #130-130-404) or LIVE/DEAD Fixable Far-Red dye (Invitrogen, Cat #L34973). Staining of surface markers was carried out at 4°C for 30 min in the presence of anti-CD16/32 blocking antibody (Fc-block, BD Biosciences #553142) with Brilliant Stain buffer (BD Biosciences Cat #566349) according to manufacturer instructions, using the following antibodies: anti-CD45 (clone 30-F11, BioLegend Cat #103111), anti-CD45.2 (clone 104, BD Biosciences Cat #564616), anti-CD4 (clone RM4-5, eBioscience Cat #48-0042-82), anti-CD8a (clone 53-6.7, BD Biosciences Cat #612898), anti-CD80 (clone 16-10A1, BioLegend Cat #104738), anti-CD86 (clone GL1, BD Bioscience Cat #564199), anti-IA-IE (clone M5/114.15.2, BD Biosciences Cat #748846), anti-H2-kB (clone AF6-88.5, BD Biosciences Cat #742861), anti-F4/80 (clone BM8, BioLegend Cat #123110), anti-CD11c (clone N418, BioLegend Cat #117319), anti-CD11b (clone M1/70, BD Pharmingen Cat #557960), anti-CD68 (clone FA-11, BioLegend Cat #137015), anti-CCR5 (clone HM-CCR5, BioLegend Cat #107005), anti-CX3CR1 (clone SA011F11, BD Biosciences Cat #149029), anti-Ly6G (clone 1A8, BioLegend Cat #127624), anti-CCR3 (clone J073E5, BioLegend Cat #144508), anti-Ly6C (clone HK1.4, BioLegend Cat #128014). Flow cytometry was performed in FACS buffer [2mM EDTA, 2% FBS (Corning) in PBS] using BD FACSymphony A5 Cell Analyzer or Cytek Aurora. Flow cytometry data were analyzed with FlowJo Software (Treestar). Image Analysis, quantification and statistics Images used for quantification were acquired using the same microscope settings and adjusted to the same background levels whenever possible. Images were processed with Fiji/ImageJ, Adobe Photoshop and Adobe Illustrator. 3D rendering and videos were generated with Arivis 4D software. Mean intensity values and px area of various antigens were measured with Fiji/ImageJ on a binary mask of either CD31 or Podocalyxin staining, as indicated. Vessel width was measured by line tool from Fiji/ImageJ and plotted in px length. For permeability assays the extravascular intensity value of sulfo-biotin/HRP/BSA in nerves or spinal cord were quantified after subtracting its signal from vessels masked by CD31. CAV1 staining were generated from individual confocal images of control or SOD1G93A nerve ECs acquired at 60× magnification with 1.5× optical zoom. Microglia and astrocyte activation in SOD1 G93A and TDP43 Q331K mice were quantified by measuring IBA1 + or GFAP + px area, respectively. Macrophages or neutrophils were manually counted by F4-80 or Ly-6B.2 (7/4) staining, respectively, within the same px area across samples. CAR4 + and PLVAP + vessels were counted manually within random areas of the same size in control, SOD1 G93A or TDP43 Q331K nerve sections. Profile intensity plots for Motor neurons were manually counted in lumbar spinal cord sections based on NeuN or ChAT staining. Myc + nuclei were counted within a fixed area of spinal cord or sciatic nerve transverse sections from TDP43 Q331K mice. GFP + or RFP + cells labeled by Cx3cr1 GFP / Ccr2 RFP reporter were counted after anti-GFP/RFP staining within the same area of longitudinal nerve sections across control and mutant samples. Nerve fibers were counted manually in areas of 300 px 2 from semithin sections. Pinocytotic vesicles were manually counted in TEM images of ECs and normalized by cytoplasmic area. Statistical analysis was performed with GraphPad Prism V8.4.0 software. Statistical details are found in the figure legends. Acknowledgements We thank Ralf Adams for Cdh5(PAC)-CreERT2 mice; Cesare Covino and the Advanced Microscopy Laboratory (ALEMBIC) of San Raffaele Hospital for expertise and instrumentation. Funding sources: AriSLA Foundation pilot grant (NeVALS), Italian Ministry of Health GR-2016-02364327, Giovanni Armenise-Harvard Foundation Career Development Award. Funder Information Declared AriSLA Foundation pilot grant (NeVALS) Italian Ministry of Health , GR-2016-02364327 Giovanni Armenise-Harvard Foundation, https://ror.org/04q644746 REFERENCES 1. ↵ Augustin , H.G. , and Koh , G.Y . ( 2017 ). Organotypic vasculature: From descriptive heterogeneity to functional pathophysiology . Science 357 . doi: 10.1126/science.aal2379 . OpenUrl Abstract / FREE Full Text 2. ↵ Aird , W.C . ( 2012 ). Endothelial cell heterogeneity . Cold Spring Harb Perspect Med 2 , a006429 . doi: 10.1101/cshperspect.a006429 . OpenUrl Abstract / FREE Full Text 3. ↵ Trimm , E. , and Red-Horse , K . ( 2023 ). Vascular endothelial cell development and diversity . Nat Rev Cardiol 20 , 197 – 210 . doi: 10.1038/s41569-022-00770-1 . OpenUrl CrossRef PubMed 4. ↵ Amersfoort , J. , Eelen , G. , and Carmeliet , P . ( 2022 ). Immunomodulation by endothelial cells - partnering up with the immune system? Nat Rev Immunol 22 , 576 – 588 . doi: 10.1038/s41577-022-00694-4 . OpenUrl CrossRef PubMed 5. Dejana , E. , Hirschi , K.K. , and Simons , M . ( 2017 ). The molecular basis of endothelial cell plasticity . Nat Commun 8 , 14361 . doi: 10.1038/ncomms14361 . OpenUrl CrossRef PubMed 6. ↵ Rafii , S. , Butler , J.M. , and Ding , B.S . ( 2016 ). Angiocrine functions of organ-specific endothelial cells . Nature 529 , 316 – 325 . doi: 10.1038/nature17040 . OpenUrl CrossRef PubMed 7. ↵ Aird , W.C . ( 2005 ). Spatial and temporal dynamics of the endothelium . J Thromb Haemost 3 , 1392 – 1406 . doi: 10.1111/j.1538-7836.2005.01328.x . OpenUrl CrossRef PubMed Web of Science 8. ↵ Bartanusz , V. , Jezova , D. , Alajajian , B. , and Digicaylioglu , M . ( 2011 ). The blood-spinal cord barrier: morphology and clinical implications . Ann Neurol 70 , 194 – 206 . doi: 10.1002/ana.22421 . OpenUrl CrossRef PubMed 9. Lacoste , B. , Prat , A. , Freitas-Andrade , M. , and Gu , C . ( 2024 ). The Blood-Brain Barrier: Composition, Properties, and Roles in Brain Health . Cold Spring Harb Perspect Biol . doi: 10.1101/cshperspect.a041422 . OpenUrl Abstract / FREE Full Text 10. ↵ Profaci , C.P. , Munji , R.N. , Pulido , R.S. , and Daneman , R . ( 2020 ). The blood-brain barrier in health and disease: Important unanswered questions . J Exp Med 217 . doi: 10.1084/jem.20190062 . OpenUrl CrossRef 11. ↵ Sweeney , M.D. , Zhao , Z. , Montagne , A. , Nelson , A.R. , and Zlokovic , B.V . ( 2019 ). Blood-Brain Barrier: From Physiology to Disease and Back . Physiol Rev 99 , 21 – 78 . doi: 10.1152/physrev.00050.2017 . OpenUrl CrossRef PubMed 12. Garbuzova-Davis , S. , Rodrigues , M.C. , Hernandez-Ontiveros , D.G. , Louis , M.K. , Willing , A.E. , Borlongan , C.V. , and Sanberg , P.R . ( 2011 ). Amyotrophic lateral sclerosis: a neurovascular disease . Brain Res 1398 , 113 – 125 . doi: 10.1016/j.brainres.2011.04.049 . OpenUrl CrossRef PubMed 13. ↵ Steinruecke , M. , Lonergan , R.M. , Selvaraj , B.T. , Chandran , S. , Diaz-Castro , B. , and Stavrou , M . ( 2023 ). Blood-CNS barrier dysfunction in amyotrophic lateral sclerosis: Proposed mechanisms and clinical implications . J Cereb Blood Flow Metab 43 , 642 – 654 . doi: 10.1177/0271678X231153281 . OpenUrl CrossRef PubMed 14. ↵ Garbuzova-Davis , S. , Haller , E. , Saporta , S. , Kolomey , I. , Nicosia , S.V. , and Sanberg , P.R . ( 2007 ). Ultrastructure of blood-brain barrier and blood-spinal cord barrier in SOD1 mice modeling ALS . Brain Res 1157 , 126 – 137 . doi: 10.1016/j.brainres.2007.04.044 . OpenUrl CrossRef PubMed Web of Science 15. Garbuzova-Davis , S. , Hernandez-Ontiveros , D.G. , Rodrigues , M.C. , Haller , E. , Frisina-Deyo , A. , Mirtyl , S. , Sallot , S. , Saporta , S. , Borlongan , C.V. , and Sanberg , P.R . ( 2012 ). Impaired blood-brain/spinal cord barrier in ALS patients . Brain Res 1469 , 114 – 128 . doi: 10.1016/j.brainres.2012.05.056 . OpenUrl CrossRef PubMed 16. Manberg , A. , Skene , N. , Sanders , F. , Trusohamn , M. , Remnestal , J. , Szczepinska , A. , Aksoylu , I.S. , Lonnerberg , P. , Ebarasi , L. , Wouters , S. , et al. ( 2021 ). Altered perivascular fibroblast activity precedes ALS disease onset . Nat Med 27 , 640 – 646 . doi: 10.1038/s41591-021-01295-9 . OpenUrl CrossRef PubMed 17. Winkler , E.A. , Sengillo , J.D. , Bell , R.D. , Wang , J. , and Zlokovic , B.V . ( 2012 ). Blood-spinal cord barrier pericyte reductions contribute to increased capillary permeability . J Cereb Blood Flow Metab 32 , 1841 – 1852 . doi: 10.1038/jcbfm.2012.113 . OpenUrl CrossRef PubMed 18. Winkler , E.A. , Sengillo , J.D. , Sullivan , J.S. , Henkel , J.S. , Appel , S.H. , and Zlokovic , B.V . ( 2013 ). Blood-spinal cord barrier breakdown and pericyte reductions in amyotrophic lateral sclerosis . Acta Neuropathol 125 , 111 – 120 . doi: 10.1007/s00401-012-1039-8 . OpenUrl CrossRef PubMed Web of Science 19. ↵ Zhong , Z. , Deane , R. , Ali , Z. , Parisi , M. , Shapovalov , Y. , O’Banion , M.K. , Stojanovic , K. , Sagare , A. , Boillee , S. , Cleveland , D.W. , and Zlokovic , B.V . ( 2008 ). ALS-causing SOD1 mutants generate vascular changes prior to motor neuron degeneration . Nat Neurosci 11 , 420 – 422 . doi: 10.1038/nn2073 . OpenUrl CrossRef PubMed Web of Science 20. ↵ Garbuzova-Davis , S. , Saporta , S. , Haller , E. , Kolomey , I. , Bennett , S.P. , Potter , H. , and Sanberg , P.R . ( 2007 ). Evidence of compromised blood-spinal cord barrier in early and late symptomatic SOD1 mice modeling ALS . PLoS One 2 , e1205 . doi: 10.1371/journal.pone.0001205 . OpenUrl CrossRef PubMed 21. Henkel , J.S. , Beers , D.R. , Wen , S. , Bowser , R. , and Appel , S.H . ( 2009 ). Decreased mRNA expression of tight junction proteins in lumbar spinal cords of patients with ALS . Neurology 72 , 1614 – 1616 . doi: 10.1212/WNL.0b013e3181a41228 . OpenUrl CrossRef PubMed 22. ↵ Sasaki , S . ( 2015 ). Alterations of the blood-spinal cord barrier in sporadic amyotrophic lateral sclerosis . Neuropathology 35 , 518 – 528 . doi: 10.1111/neup.12221 . OpenUrl CrossRef PubMed 23. ↵ Yamadera , M. , Fujimura , H. , Inoue , K. , Toyooka , K. , Mori , C. , Hirano , H. , and Sakoda , S . ( 2015 ). Microvascular disturbance with decreased pericyte coverage is prominent in the ventral horn of patients with amyotrophic lateral sclerosis . Amyotroph Lateral Scler Frontotemporal Degener 16 , 393 – 401 . doi: 10.3109/21678421.2015.1011663 . OpenUrl CrossRef PubMed 24. ↵ Annunziata , P. , and Volpi , N . ( 1985 ). High levels of C3c in the cerebrospinal fluid from amyotrophic lateral sclerosis patients . Acta Neurol Scand 72 , 61 – 64 . doi: 10.1111/j.1600-0404.1985.tb01548.x . OpenUrl CrossRef PubMed 25. Brettschneider , J. , Petzold , A. , Sussmuth , S.D. , Ludolph , A.C. , and Tumani , H . ( 2006 ). Axonal damage markers in cerebrospinal fluid are increased in ALS . Neurology 66 , 852 – 856 . doi: 10.1212/01.wnl.0000203120.85850.54 . OpenUrl Abstract / FREE Full Text 26. Donnenfeld , H. , Kascsak , R.J. , and Bartfeld , H . ( 1984 ). Deposits of IgG and C3 in the spinal cord and motor cortex of ALS patients . J Neuroimmunol 6 , 51 – 57 . doi: 10.1016/0165-5728(84)90042-0 . OpenUrl CrossRef PubMed Web of Science 27. Engelhardt , J.I. , and Appel , S.H . ( 1990 ). IgG reactivity in the spinal cord and motor cortex in amyotrophic lateral sclerosis . Arch Neurol 47 , 1210 – 1216 . doi: 10.1001/archneur.1990.00530110068019 . OpenUrl CrossRef PubMed Web of Science 28. Guiloff , R.J. , McGregor , B. , Thompson , E. , Blackwood , W. , and Paul , E . ( 1980 ). Motor neurone disease with elevated cerebrospinal fluid protein . J Neurol Neurosurg Psychiatry 43 , 390 – 396 . doi: 10.1136/jnnp.43.5.390 . OpenUrl Abstract / FREE Full Text 29. Leonardi , A. , Abbruzzese , G. , Arata , L. , Cocito , L. , and Vische , M . ( 1984 ). Cerebrospinal fluid (CSF) findings in amyotrophic lateral sclerosis . J Neurol 231 , 75 – 78 . doi: 10.1007/BF00313720 . OpenUrl CrossRef PubMed Web of Science 30. ↵ Troost , D. , Van den Oord , J.J. , and Vianney de Jong , J.M. ( 1990 ). Immunohistochemical characterization of the inflammatory infiltrate in amyotrophic lateral sclerosis . Neuropathol Appl Neurobiol 16 , 401 – 410 . doi: 10.1111/j.1365-2990.1990.tb01276.x . OpenUrl CrossRef PubMed Web of Science 31. ↵ Prell , T. , Vlad , B. , Gaur , N. , Stubendorff , B. , and Grosskreutz , J . ( 2021 ). Blood-Brain Barrier Disruption Is Not Associated With Disease Aggressiveness in Amyotrophic Lateral Sclerosis . Front Neurosci 15 , 656456 . doi: 10.3389/fnins.2021.656456 . OpenUrl CrossRef PubMed 32. ↵ Waters , S. , Swanson , M.E.V. , Dieriks , B.V. , Zhang , Y.B. , Grimsey , N.L. , Murray , H.C. , Turner , C. , Waldvogel , H.J. , Faull , R.L.M. , An , J. , et al. ( 2021 ). Blood-spinal cord barrier leakage is independent of motor neuron pathology in ALS . Acta Neuropathol Commun 9 , 144 . doi: 10.1186/s40478-021-01244-0 . OpenUrl CrossRef PubMed 33. ↵ Miyazaki , K. , Ohta , Y. , Nagai , M. , Morimoto , N. , Kurata , T. , Takehisa , Y. , Ikeda , Y. , Matsuura , T. , and Abe , K . ( 2011 ). Disruption of neurovascular unit prior to motor neuron degeneration in amyotrophic lateral sclerosis . J Neurosci Res 89 , 718 – 728 . doi: 10.1002/jnr.22594 . OpenUrl CrossRef PubMed 34. Nicaise , C. , Mitrecic , D. , Demetter , P. , De Decker , R. , Authelet , M. , Boom , A. , and Pochet , R. ( 2009 ). Impaired blood-brain and blood-spinal cord barriers in mutant SOD1-linked ALS rat . Brain Res 1301 , 152 – 162 . doi: 10.1016/j.brainres.2009.09.018 . OpenUrl CrossRef PubMed Web of Science 35. ↵ Ouali Alami , N. , Tang , L. , Wiesner , D. , Commisso , B. , Bayer , D. , Weishaupt , J. , Dupuis , L. , Wong , P. , Baumann , B. , Wirth , T. , et al. ( 2020 ). Multiplexed chemogenetics in astrocytes and motoneurons restore blood-spinal cord barrier in ALS . Life Sci Alliance 3 . doi: 10.26508/lsa.201900571 . OpenUrl Abstract / FREE Full Text 36. ↵ Bien-Ly , N. , Boswell , C.A. , Jeet , S. , Beach , T.G. , Hoyte , K. , Luk , W. , Shihadeh , V. , Ulufatu , S. , Foreman , O. , Lu , Y. , et al. ( 2015 ). Lack of Widespread BBB Disruption in Alzheimer’s Disease Models: Focus on Therapeutic Antibodies . Neuron 88 , 289 – 297 . doi: 10.1016/j.neuron.2015.09.036 . OpenUrl CrossRef PubMed 37. ↵ Arribas , V. , Onetti , Y. , Ramiro-Pareta , M. , Villacampa , P. , Beck , H. , Alberola , M. , Esteve-Codina , A. , Merkel , A. , Sperandio , M. , Martinez-Estrada , O.M. , et al. ( 2024 ). Endothelial TDP-43 controls sprouting angiogenesis and vascular barrier integrity, and its deletion triggers neuroinflammation . JCI Insight 9 . doi: 10.1172/jci.insight.177819 . OpenUrl CrossRef 38. Cheemala , A. , Kimble , A.L. , Burrage , E.N. , Helming , S.B. , Tyburski , J.D. , Leclair , N.K. , Omar , O.M. , Zuberi , A.R. , Murphy , M. , Jellison , E.R. , et al. ( 2025 ). Amyotrophic lateral sclerosis and frontotemporal dementia mutation reduces endothelial TDP-43 and causes blood-brain barrier defects . Sci Adv 11 , eads0505 . doi: 10.1126/sciadv.ads0505 . OpenUrl CrossRef PubMed 39. ↵ Omar , O.M.F. , Kimble , A.L. , Cheemala , A. , Tyburski , J.D. , Pandey , S. , Wu , Q. , Reese , B. , Jellison , E.R. , Hao , B. , Li , Y. , et al. ( 2025 ). Endothelial TDP-43 depletion disrupts core blood-brain barrier pathways in neurodegeneration . Nat Neurosci 28 , 973 – 984 . doi: 10.1038/s41593-025-01914-5 . OpenUrl CrossRef PubMed 40. ↵ Sullivan , J.M. , Bagnell , A.M. , Alevy , J. , Avila , E.M. , Mihaljevic , L. , Saavedra-Rivera , P.C. , Kong , L. , Huh , J.S. , McCray , B.A. , Aisenberg , W.H. , et al. ( 2024 ). Gain-of-function mutations of TRPV4 acting in endothelial cells drive blood-CNS barrier breakdown and motor neuron degeneration in mice . Sci Transl Med 16 , eadk1358 . doi: 10.1126/scitranslmed.adk1358 . OpenUrl CrossRef PubMed 41. ↵ Van Harten , A.C.M. , Phatnani , H. , and Przedborski , S. ( 2021 ). Non-cell-autonomous pathogenic mechanisms in amyotrophic lateral sclerosis . Trends Neurosci 44 , 658 – 668 . doi: 10.1016/j.tins.2021.04.008 . OpenUrl CrossRef PubMed 42. ↵ Bonanomi , D. , and Pfaff , S.L . ( 2010 ). Motor axon pathfinding . Cold Spring Harb Perspect Biol 2 , a001735 . doi: 10.1101/cshperspect.a001735 . OpenUrl Abstract / FREE Full Text 43. ↵ Grad , L.I. , Rouleau , G.A. , Ravits , J. , and Cashman , N.R . ( 2017 ). Clinical Spectrum of Amyotrophic Lateral Sclerosis (ALS) . Cold Spring Harb Perspect Med 7 . doi: 10.1101/cshperspect.a024117 . OpenUrl Abstract / FREE Full Text 44. ↵ Chio , A. , Calvo , A. , Moglia , C. , Mazzini , L. , Mora , G. , and group, P.s. ( 2011 ). Phenotypic heterogeneity of amyotrophic lateral sclerosis: a population based study . J Neurol Neurosurg Psychiatry 82 , 740 – 746 . doi: 10.1136/jnnp.2010.235952 . OpenUrl Abstract / FREE Full Text 45. ↵ Baker , M.R . ( 2014 ). ALS--dying forward, backward or outward? Nat Rev Neurol 10 , 660 . doi: 10.1038/nrneurol.2013.221-c1 . OpenUrl CrossRef PubMed 46. ↵ Ravits , J.M. , and La Spada , A.R. ( 2009 ). ALS motor phenotype heterogeneity, focality, and spread: deconstructing motor neuron degeneration . Neurology 73 , 805 – 811 . doi: 10.1212/WNL.0b013e3181b6bbbd . OpenUrl Abstract / FREE Full Text 47. ↵ Coleman , M.P . ( 2022 ). Axon Biology in ALS: Mechanisms of Axon Degeneration and Prospects for Therapy . Neurotherapeutics 19 , 1133 – 1144 . doi: 10.1007/s13311-022-01297-6 . OpenUrl CrossRef PubMed 48. ↵ Fischer , L.R. , Culver , D.G. , Tennant , P. , Davis , A.A. , Wang , M. , Castellano-Sanchez , A. , Khan , J. , Polak , M.A. , and Glass , J.D . ( 2004 ). Amyotrophic lateral sclerosis is a distal axonopathy: evidence in mice and man . Exp Neurol 185 , 232 – 240 . doi: 10.1016/j.expneurol.2003.10.004 . OpenUrl CrossRef PubMed Web of Science 49. ↵ White , M.A. , Lin , Z. , Kim , E. , Henstridge , C.M. , Pena Altamira , E. , Hunt , C.K. , Burchill , E. , Callaghan , I. , Loreto , A. , Brown-Wright , H. , et al. ( 2019 ). Sarm1 deletion suppresses TDP-43-linked motor neuron degeneration and cortical spine loss . Acta Neuropathol Commun 7 , 166 . doi: 10.1186/s40478-019-0800-9 . OpenUrl CrossRef 50. ↵ Riva , N. , Clarelli , F. , Domi , T. , Cerri , F. , Gallia , F. , Trimarco , A. , Brambilla , P. , Lunetta , C. , Lazzerini , A. , Lauria , G. , et al. ( 2016 ). Unraveling gene expression profiles in peripheral motor nerve from amyotrophic lateral sclerosis patients: insights into pathogenesis . Sci Rep 6 , 39297 . doi: 10.1038/srep39297 . OpenUrl CrossRef 51. ↵ Riva , N. , Gentile , F. , Cerri , F. , Gallia , F. , Podini , P. , Dina , G. , Falzone , Y.M. , Fazio , R. , Lunetta , C. , Calvo , A. , et al. ( 2022 ). Phosphorylated TDP-43 aggregates in peripheral motor nerves of patients with amyotrophic lateral sclerosis . Brain 145 , 276 – 284 . doi: 10.1093/brain/awab285 . OpenUrl CrossRef PubMed 52. ↵ Trias , E. , Kovacs , M. , King , P.H. , Si , Y. , Kwon , Y. , Varela , V. , Ibarburu , S. , Moura , I.C. , Hermine , O. , Beckman , J.S. , and Barbeito , L . ( 2020 ). Schwann cells orchestrate peripheral nerve inflammation through the expression of CSF1, IL-34, and SCF in amyotrophic lateral sclerosis . Glia 68 , 1165 – 1181 . doi: 10.1002/glia.23768 . OpenUrl CrossRef PubMed 53. ↵ Chiot , A. , Zaidi , S. , Iltis , C. , Ribon , M. , Berriat , F. , Schiaffino , L. , Jolly , A. , de la Grange , P. , Mallat , M. , Bohl , D. , et al. ( 2020 ). Modifying macrophages at the periphery has the capacity to change microglial reactivity and to extend ALS survival . Nat Neurosci 23 , 1339 – 1351 . doi: 10.1038/s41593-020-00718-z . OpenUrl CrossRef PubMed 54. ↵ Shiraishi , W. , Yamasaki , R. , Hashimoto , Y. , Ko , S. , Kobayakawa , Y. , Isobe , N. , Matsushita , T. , and Kira , J.I . ( 2021 ). Clearance of peripheral nerve misfolded mutant protein by infiltrated macrophages correlates with motor neuron disease progression . Sci Rep 11 , 16438 . doi: 10.1038/s41598-021-96064-6 . OpenUrl CrossRef PubMed 55. ↵ Trias , E. , King , P.H. , Si , Y. , Kwon , Y. , Varela , V. , Ibarburu , S. , Kovacs , M. , Moura , I.C. , Beckman , J.S. , Hermine , O. , and Barbeito , L . ( 2018 ). Mast cells and neutrophils mediate peripheral motor pathway degeneration in ALS . JCI Insight 3 . doi: 10.1172/jci.insight.123249 . OpenUrl CrossRef 56. Nardo , G. , Trolese , M.C. , de Vito , G. , Cecchi , R. , Riva , N. , Dina , G. , Heath , P.R. , Quattrini , A. , Shaw , P.J. , Piazza , V. , and Bendotti , C. ( 2016 ). Immune response in peripheral axons delays disease progression in SOD1(G93A) mice . J Neuroinflammation 13 , 261 . doi: 10.1186/s12974-016-0732-2 . OpenUrl CrossRef PubMed 57. ↵ Chiu , I.M. , Phatnani , H. , Kuligowski , M. , Tapia , J.C. , Carrasco , M.A. , Zhang , M. , Maniatis , T. , and Carroll , M.C . ( 2009 ). Activation of innate and humoral immunity in the peripheral nervous system of ALS transgenic mice . Proc Natl Acad Sci U S A 106 , 20960 – 20965 . doi: 10.1073/pnas.0911405106 . OpenUrl Abstract / FREE Full Text 58. ↵ Reinhold , A.K. , and Rittner , H.L . ( 2017 ). Barrier function in the peripheral and central nervous system-a review . Pflugers Arch 469 , 123 – 134 . doi: 10.1007/s00424-016-1920-8 . OpenUrl CrossRef PubMed 59. ↵ Ubogu , E.E . ( 2020 ). Biology of the human blood-nerve barrier in health and disease . Exp Neurol 328 , 113272 . doi: 10.1016/j.expneurol.2020.113272 . OpenUrl CrossRef PubMed 60. ↵ Malong , L. , Napoli , I. , Casal , G. , White , I.J. , Stierli , S. , Vaughan , A. , Cattin , A.L. , Burden , J.J. , Hng , K.I. , Bossio , A. , et al. ( 2023 ). Characterization of the structure and control of the blood-nerve barrier identifies avenues for therapeutic delivery . Dev Cell 58 , 174 – 191 e178 . doi: 10.1016/j.devcel.2023.01.002 . OpenUrl CrossRef PubMed 61. ↵ Schaeffer , S. , and Iadecola , C . ( 2021 ). Revisiting the neurovascular unit . Nat Neurosci 24 , 1198 – 1209 . doi: 10.1038/s41593-021-00904-7 . OpenUrl CrossRef PubMed 62. ↵ Wadhwani , K.C. , and Rapoport , S.I . ( 1994 ). Transport properties of vertebrate blood-nerve barrier: comparison with blood-brain barrier . Prog Neurobiol 43 , 235 – 279 . doi: 10.1016/0301-0082(94)90002-7 . OpenUrl CrossRef PubMed Web of Science 63. ↵ Welch , K. , and Davson , H . ( 1972 ). The permeability of capillaries of the sciatic nerve of the rabbit to several materials . J Neurosurg 36 , 21 – 26 . doi: 10.3171/jns.1972.36.1.0021 . OpenUrl CrossRef PubMed 64. ↵ Bhat , G.P. , Maurizio , A. , Motta , A. , Podini , P. , Diprima , S. , Malpighi , C. , Brambilla , I. , Martins , L. , Badaloni , A. , Boselli , D. , et al. ( 2024 ). Structured wound angiogenesis instructs mesenchymal barrier compartments in the regenerating nerve . Neuron 112 , 209 – 229 e211 . doi: 10.1016/j.neuron.2023.10.025 . OpenUrl CrossRef PubMed 65. ↵ Cattin , A.L. , Burden , J.J. , Van Emmenis , L. , Mackenzie , F.E. , Hoving , J.J. , Garcia Calavia , N. , Guo , Y. , McLaughlin , M. , Rosenberg , L.H. , Quereda , V. , et al. ( 2015 ). Macrophage-Induced Blood Vessels Guide Schwann Cell-Mediated Regeneration of Peripheral Nerves . Cell 162 , 1127 – 1139 . doi: 10.1016/j.cell.2015.07.021 . OpenUrl CrossRef PubMed 66. ↵ Riva , N. , Chaabane , L. , Peviani , M. , Ungaro , D. , Domi , T. , Dina , G. , Bianchi , F. , Spano , G. , Cerri , F. , Podini , P. , et al. ( 2014 ). Defining peripheral nervous system dysfunction in the SOD-1G93A transgenic rat model of amyotrophic lateral sclerosis . J Neuropathol Exp Neurol 73 , 658 – 670 . doi: 10.1097/NEN.0000000000000081 . OpenUrl CrossRef PubMed 67. ↵ Kim , D.W. , Langille , B.L. , Wong , M.K. , and Gotlieb , A.I . ( 1989 ). Patterns of endothelial microfilament distribution in the rabbit aorta in situ . Circ Res 64 , 21 – 31 . doi: 10.1161/01.res.64.1.21 . OpenUrl Abstract / FREE Full Text 68. ↵ Arnold , E.S. , Ling , S.C. , Huelga , S.C. , Lagier-Tourenne , C. , Polymenidou , M. , Ditsworth , D. , Kordasiewicz , H.B. , McAlonis-Downes , M. , Platoshyn , O. , Parone , P.A. , et al. ( 2013 ). ALS-linked TDP-43 mutations produce aberrant RNA splicing and adult-onset motor neuron disease without aggregation or loss of nuclear TDP-43 . Proc Natl Acad Sci U S A 110 , E736 – 745 . doi: 10.1073/pnas.1222809110 . OpenUrl Abstract / FREE Full Text 69. ↵ Gurney , M.E. , Pu , H. , Chiu , A.Y. , Dal Canto , M.C. , Polchow , C.Y. , Alexander , D.D. , Caliendo , J. , Hentati , A. , Kwon , Y.W. , Deng , H.X. , and, et al. ( 1994 ). Motor neuron degeneration in mice that express a human Cu,Zn superoxide dismutase mutation . Science 264 , 1772 – 1775 . doi: 10.1126/science.8209258 . OpenUrl Abstract / FREE Full Text 70. ↵ Watkins , J.A. , Alix , J.J.P. , Shaw , P.J. , and Mead , R.J . ( 2021 ). Extensive phenotypic characterisation of a human TDP-43(Q331K) transgenic mouse model of amyotrophic lateral sclerosis (ALS) . Sci Rep 11 , 16659 . doi: 10.1038/s41598-021-96122-z . OpenUrl CrossRef PubMed 71. ↵ Chow , B.W. , and Gu , C . ( 2017 ). Gradual Suppression of Transcytosis Governs Functional Blood-Retinal Barrier Formation . Neuron 93 , 1325 – 1333 e1323 . doi: 10.1016/j.neuron.2017.02.043 . OpenUrl CrossRef PubMed 72. ↵ Wang , Y. , Rattner , A. , Zhou , Y. , Williams , J. , Smallwood , P.M. , and Nathans , J . ( 2012 ). Norrin/Frizzled4 signaling in retinal vascular development and blood brain barrier plasticity . Cell 151 , 1332 – 1344 . doi: 10.1016/j.cell.2012.10.042 . OpenUrl CrossRef PubMed Web of Science 73. ↵ Gustavsson , C. , Agardh , C.D. , Zetterqvist , A.V. , Nilsson , J. , Agardh , E. , and Gomez , M.F . ( 2010 ). Vascular cellular adhesion molecule-1 (VCAM-1) expression in mice retinal vessels is affected by both hyperglycemia and hyperlipidemia . PLoS One 5 , e12699 . doi: 10.1371/journal.pone.0012699 . OpenUrl CrossRef PubMed 74. ↵ Malinverno , M. , Maderna , C. , Abu Taha , A. , Corada , M. , Orsenigo , F. , Valentino , M. , Pisati , F. , Fusco , C. , Graziano , P. , Giannotta , M. , et al. ( 2019 ). Endothelial cell clonal expansion in the development of cerebral cavernous malformations . Nat Commun 10 , 2761 . doi: 10.1038/s41467-019-10707-x . OpenUrl CrossRef PubMed 75. ↵ Dominov , J.A. , Madigan , L.A. , Whitt , J.P. , Rademacher , K.L. , Webster , K.M. , Zhang , H. , Banno , H. , Tang , S. , Zhang , Y. , Wightman , N. , et al. ( 2023 ). Up-regulation of cholesterol synthesis pathways and limited neurodegeneration in a knock-in Sod1 mutant mouse model of ALS . bioRxiv . doi: 10.1101/2023.05.05.539444 . OpenUrl Abstract / FREE Full Text 76. ↵ Pober , J.S. , and Sessa , W.C . ( 2007 ). Evolving functions of endothelial cells in inflammation . Nat Rev Immunol 7 , 803 – 815 . doi: 10.1038/nri2171 . OpenUrl CrossRef PubMed Web of Science 77. ↵ Kalucka , J. , de Rooij , L. , Goveia , J. , Rohlenova , K. , Dumas , S.J. , Meta , E. , Conchinha , N.V. , Taverna , F. , Teuwen , L.A. , Veys , K. , et al. ( 2020 ). Single-Cell Transcriptome Atlas of Murine Endothelial Cells . Cell 180 , 764 – 779 e720 . doi: 10.1016/j.cell.2020.01.015 . OpenUrl CrossRef PubMed 78. ↵ Pober , J.S . ( 1988 ). Warner-Lambert/Parke-Davis award lecture. Cytokine-mediated activation of vascular endothelium. Physiology and pathology . Am J Pathol 133 , 426 – 433 . OpenUrl PubMed Web of Science 79. ↵ Denzer , L. , Muranyi , W. , Schroten , H. , and Schwerk , C . ( 2023 ). The role of PLVAP in endothelial cells . Cell Tissue Res 392 , 393 – 412 . doi: 10.1007/s00441-023-03741-1 . OpenUrl CrossRef PubMed 80. ↵ Milici , A.J. , Watrous , N.E. , Stukenbrok , H. , and Palade , G.E . ( 1987 ). Transcytosis of albumin in capillary endothelium . J Cell Biol 105 , 2603 – 2612 . doi: 10.1083/jcb.105.6.2603 . OpenUrl Abstract / FREE Full Text 81. Reese , T.S. , and Karnovsky , M.J . ( 1967 ). Fine structural localization of a blood-brain barrier to exogenous peroxidase . J Cell Biol 34 , 207 – 217 . doi: 10.1083/jcb.34.1.207 . OpenUrl Abstract / FREE Full Text 82. ↵ Yang , A.C. , Stevens , M.Y. , Chen , M.B. , Lee , D.P. , Stahli , D. , Gate , D. , Contrepois , K. , Chen , W. , Iram , T. , Zhang , L. , et al. ( 2020 ). Physiological blood-brain transport is impaired with age by a shift in transcytosis . Nature 583 , 425 – 430 . doi: 10.1038/s41586-020-2453-z . OpenUrl CrossRef 83. ↵ Tuma , P. , and Hubbard , A.L . ( 2003 ). Transcytosis: crossing cellular barriers . Physiol Rev 83 , 871 – 932 . doi: 10.1152/physrev.00001.2003 . OpenUrl CrossRef PubMed Web of Science 84. ↵ Razani , B. , Engelman , J.A. , Wang , X.B. , Schubert , W. , Zhang , X.L. , Marks , C.B. , Macaluso , F. , Russell , R.G. , Li , M. , Pestell , R.G. , et al. ( 2001 ). Caveolin-1 null mice are viable but show evidence of hyperproliferative and vascular abnormalities . J Biol Chem 276 , 38121 – 38138 . doi: 10.1074/jbc.M105408200 . OpenUrl Abstract / FREE Full Text 85. ↵ Zimnicka , A.M. , Husain , Y.S. , Shajahan , A.N. , Sverdlov , M. , Chaga , O. , Chen , Z. , Toth , P.T. , Klomp , J. , Karginov , A.V. , Tiruppathi , C. , et al. ( 2016 ). Src-dependent phosphorylation of caveolin-1 Tyr-14 promotes swelling and release of caveolae . Mol Biol Cell 27 , 2090 – 2106 . doi: 10.1091/mbc.E15-11-0756 . OpenUrl Abstract / FREE Full Text 86. ↵ Krolak , T. , Chan , K.Y. , Kaplan , L. , Huang , Q. , Wu , J. , Zheng , Q. , Kozareva , V. , Beddow , T. , Tobey , I.G. , Pacouret , S. , et al. ( 2022 ). A High-Efficiency AAV for Endothelial Cell Transduction Throughout the Central Nervous System . Nat Cardiovasc Res 1 , 389 – 400 . doi: 10.1038/s44161-022-00046-4 . OpenUrl CrossRef PubMed 87. ↵ Andreone , B.J. , Chow , B.W. , Tata , A. , Lacoste , B. , Ben-Zvi , A. , Bullock , K. , Deik , A.A. , Ginty , D.D. , Clish , C.B. , and Gu , C . ( 2017 ). Blood-Brain Barrier Permeability Is Regulated by Lipid Transport-Dependent Suppression of Caveolae-Mediated Transcytosis . Neuron 94 , 581 – 594 e585 . doi: 10.1016/j.neuron.2017.03.043 . OpenUrl CrossRef PubMed 88. ↵ Ben-Zvi , A. , Lacoste , B. , Kur , E. , Andreone , B.J. , Mayshar , Y. , Yan , H. , and Gu , C . ( 2014 ). Mfsd2a is critical for the formation and function of the blood-brain barrier . Nature 509 , 507 – 511 . doi: 10.1038/nature13324 . OpenUrl CrossRef PubMed Web of Science 89. ↵ Gerber , D. , Pereira , J.A. , Gerber , J. , Tan , G. , Dimitrieva , S. , Yanguez , E. , and Suter , U . ( 2021 ). Transcriptional profiling of mouse peripheral nerves to the single-cell level to build a sciatic nerve ATlas (SNAT) . Elife 10 . doi: 10.7554/eLife.58591 . OpenUrl CrossRef 90. Toma , J.S. , Karamboulas , K. , Carr , M.J. , Kolaj , A. , Yuzwa , S.A. , Mahmud , N. , Storer , M.A. , Kaplan , D.R. , and Miller , F.D . ( 2020 ). Peripheral Nerve Single-Cell Analysis Identifies Mesenchymal Ligands that Promote Axonal Growth . eNeuro 7 . doi: 10.1523/ENEURO.0066-20.2020 . OpenUrl Abstract / FREE Full Text 91. Wolbert , J. , Li , X. , Heming , M. , Mausberg , A.K. , Akkermann , D. , Frydrychowicz , C. , Fledrich , R. , Groeneweg , L. , Schulz , C. , Stettner , M. , et al. ( 2020 ). Redefining the heterogeneity of peripheral nerve cells in health and autoimmunity . Proc Natl Acad Sci U S A 117 , 9466 – 9476 . doi: 10.1073/pnas.1912139117 . OpenUrl Abstract / FREE Full Text 92. ↵ Zhao , X.F. , Huffman , L.D. , Hafner , H. , Athaiya , M. , Finneran , M.C. , Kalinski , A.L. , Kohen , R. , Flynn , C. , Passino , R. , Johnson , C.N. , et al. ( 2022 ). The injured sciatic nerve atlas (iSNAT), insights into the cellular and molecular basis of neural tissue degeneration and regeneration . Elife 11 . doi: 10.7554/eLife.80881 . OpenUrl CrossRef 93. ↵ Vanlandewijck , M. , He , L. , Mae , M.A. , Andrae , J. , Ando , K. , Del Gaudio , F. , Nahar , K. , Lebouvier , T. , Lavina , B. , Gouveia , L. , et al. ( 2018 ). A molecular atlas of cell types and zonation in the brain vasculature . Nature 554 , 475 – 480 . doi: 10.1038/nature25739 . OpenUrl CrossRef PubMed 94. ↵ Shahriar , S. , Biswas , S. , Zhao , K. , Akcan , U. , Tuohy , M.C. , Glendinning , M.D. , Kurt , A. , Wayne , C.R. , Prochilo , G. , Price , M.Z. , et al. ( 2024 ). VEGF-A-mediated venous endothelial cell proliferation results in neoangiogenesis during neuroinflammation . Nat Neurosci 27 , 1904 – 1917 . doi: 10.1038/s41593-024-01746-9 . OpenUrl CrossRef PubMed 95. ↵ Hofstetter , H.H. , Shive , C.L. , and Forsthuber , T.G . ( 2002 ). Pertussis toxin modulates the immune response to neuroantigens injected in incomplete Freund’s adjuvant: induction of Th1 cells and experimental autoimmune encephalomyelitis in the presence of high frequencies of Th2 cells . J Immunol 169 , 117 – 125 . doi: 10.4049/jimmunol.169.1.117 . OpenUrl Abstract / FREE Full Text 96. Linthicum , D.S. , Munoz , J.J. , and Blaskett , A . ( 1982 ). Acute experimental autoimmune encephalomyelitis in mice. I. Adjuvant action of Bordetella pertussis is due to vasoactive amine sensitization and increased vascular permeability of the central nervous system . Cell Immunol 73 , 299 – 310 . doi: 10.1016/0008-8749(82)90457-9 . OpenUrl CrossRef PubMed Web of Science 97. ↵ Lutz , S.E. , Smith , J.R. , Kim , D.H. , Olson , C.V.L. , Ellefsen , K. , Bates , J.M. , Gandhi , S.P. , and Agalliu , D . ( 2017 ). Caveolin1 Is Required for Th1 Cell Infiltration, but Not Tight Junction Remodeling, at the Blood-Brain Barrier in Autoimmune Neuroinflammation . Cell Rep 21 , 2104 – 2117 . doi: 10.1016/j.celrep.2017.10.094 . OpenUrl CrossRef PubMed 98. ↵ Medzhitov , R . ( 2008 ). Origin and physiological roles of inflammation . Nature 454 , 428 – 435 . doi: 10.1038/nature07201 . OpenUrl CrossRef PubMed Web of Science 99. ↵ Butovsky , O. , Siddiqui , S. , Gabriely , G. , Lanser , A.J. , Dake , B. , Murugaiyan , G. , Doykan , C.E. , Wu , P.M. , Gali , R.R. , Iyer , L.K. , et al. ( 2012 ). Modulating inflammatory monocytes with a unique microRNA gene signature ameliorates murine ALS . J Clin Invest 122 , 3063 – 3087 . doi: 10.1172/JCI62636 . OpenUrl CrossRef PubMed Web of Science 100. ↵ Quek , H. , Cuni-Lopez , C. , Stewart , R. , Colletti , T. , Notaro , A. , Nguyen , T.H. , Sun , Y. , Guo , C.C. , Lupton , M.K. , Roberts , T.L. , et al. ( 2022 ). ALS monocyte-derived microglia-like cells reveal cytoplasmic TDP-43 accumulation, DNA damage, and cell-specific impairment of phagocytosis associated with disease progression . J Neuroinflammation 19 , 58 . doi: 10.1186/s12974-022-02421-1 . OpenUrl CrossRef 101. ↵ Dal-Secco , D. , Wang , J. , Zeng , Z. , Kolaczkowska , E. , Wong , C.H. , Petri , B. , Ransohoff , R.M. , Charo , I.F. , Jenne , C.N. , and Kubes , P . ( 2015 ). A dynamic spectrum of monocytes arising from the in situ reprogramming of CCR2+ monocytes at a site of sterile injury . J Exp Med 212 , 447 – 456 . doi: 10.1084/jem.20141539 . OpenUrl Abstract / FREE Full Text 102. ↵ Girgis , N.M. , Gundra , U.M. , Ward , L.N. , Cabrera , M. , Frevert , U. , and Loke , P . ( 2014 ). Ly6C(high) monocytes become alternatively activated macrophages in schistosome granulomas with help from CD4+ cells . PLoS Pathog 10 , e1004080 . doi: 10.1371/journal.ppat.1004080 . OpenUrl CrossRef PubMed 103. ↵ Boring , L. , Gosling , J. , Chensue , S.W. , Kunkel , S.L. , Farese , R.V. , Jr. , Broxmeyer , H.E. , and Charo , I.F . ( 1997 ). Impaired monocyte migration and reduced type 1 (Th1) cytokine responses in C-C chemokine receptor 2 knockout mice . J Clin Invest 100 , 2552 – 2561 . doi: 10.1172/JCI119798 . OpenUrl CrossRef PubMed Web of Science 104. ↵ Woodfin , A. , Voisin , M.B. , Imhof , B.A. , Dejana , E. , Engelhardt , B. , and Nourshargh , S . ( 2009 ). Endothelial cell activation leads to neutrophil transmigration as supported by the sequential roles of ICAM-2, JAM-A, and PECAM-1 . Blood 113 , 6246 – 6257 . doi: 10.1182/blood-2008-11-188375 . OpenUrl Abstract / FREE Full Text 105. ↵ Zarbock , A. , and Ley , K . ( 2008 ). Mechanisms and consequences of neutrophil interaction with the endothelium . Am J Pathol 172 , 1 – 7 . doi: 10.2353/ajpath.2008.070502 . OpenUrl CrossRef PubMed Web of Science 106. ↵ DiStasi , M.R. , and Ley , K . ( 2009 ). Opening the flood-gates: how neutrophil-endothelial interactions regulate permeability . Trends Immunol 30 , 547 – 556 . doi: 10.1016/j.it.2009.07.012 . OpenUrl CrossRef PubMed 107. ↵ Santos-Lima , B. , Pietronigro , E.C. , Terrabuio , E. , Zenaro , E. , and Constantin , G . ( 2022 ). The role of neutrophils in the dysfunction of central nervous system barriers . Front Aging Neurosci 14 , 965169 . doi: 10.3389/fnagi.2022.965169 . OpenUrl CrossRef PubMed 108. ↵ Borchelt , D.R. , Davis , J. , Fischer , M. , Lee , M.K. , Slunt , H.H. , Ratovitsky , T. , Regard , J. , Copeland , N.G. , Jenkins , N.A. , Sisodia , S.S. , and Price , D.L . ( 1996 ). A vector for expressing foreign genes in the brains and hearts of transgenic mice . Genet Anal 13 , 159 – 163 . doi: 10.1016/s1050-3862(96)00167-2 . OpenUrl CrossRef PubMed Web of Science 109. ↵ Wang , J. , Xu , G. , Slunt , H.H. , Gonzales , V. , Coonfield , M. , Fromholt , D. , Copeland , N.G. , Jenkins , N.A. , and Borchelt , D.R . ( 2005 ). Coincident thresholds of mutant protein for paralytic disease and protein aggregation caused by restrictively expressed superoxide dismutase cDNA . Neurobiol Dis 20 , 943 – 952 . doi: 10.1016/j.nbd.2005.06.005 . OpenUrl CrossRef PubMed Web of Science 110. ↵ Wang , Y. , Nakayama , M. , Pitulescu , M.E. , Schmidt , T.S. , Bochenek , M.L. , Sakakibara , A. , Adams , S. , Davy , A. , Deutsch , U. , Luthi , U. , et al. ( 2010 ). Ephrin-B2 controls VEGF-induced angiogenesis and lymphangiogenesis . Nature 465 , 483 – 486 . doi: 10.1038/nature09002 . OpenUrl CrossRef PubMed Web of Science 111. ↵ Chiot , A. , Lobsiger , C.S. , and Boillee , S . ( 2019 ). New insights on the disease contribution of neuroinflammation in amyotrophic lateral sclerosis . Curr Opin Neurol 32 , 764 – 770 . doi: 10.1097/WCO.0000000000000729 . OpenUrl CrossRef PubMed 112. Gentile , F. , Scarlino , S. , Falzone , Y.M. , Lunetta , C. , Tremolizzo , L. , Quattrini , A. , and Riva , N . ( 2019 ). The Peripheral Nervous System in Amyotrophic Lateral Sclerosis: Opportunities for Translational Research . Front Neurosci 13 , 601 . doi: 10.3389/fnins.2019.00601 . OpenUrl CrossRef 113. ↵ Ionescu , A. , Altman , T. , and Perlson , E . ( 2023 ). Looking for answers far away from the soma-the (un)known axonal functions of TDP-43, and their contribution to early NMJ disruption in ALS . Mol Neurodegener 18 , 35 . doi: 10.1186/s13024-023-00623-6 . OpenUrl CrossRef 114. ↵ Chen , M.B. , Yang , A.C. , Yousef , H. , Lee , D. , Chen , W. , Schaum , N. , Lehallier , B. , Quake , S.R. , and Wyss-Coray , T . ( 2020 ). Brain Endothelial Cells Are Exquisite Sensors of Age-Related Circulatory Cues . Cell Rep 30 , 4418 – 4432 e4414 . doi: 10.1016/j.celrep.2020.03.012 . OpenUrl CrossRef PubMed 115. ↵ Zlokovic , B.V . ( 2008 ). The blood-brain barrier in health and chronic neurodegenerative disorders . Neuron 57 , 178 – 201 . doi: 10.1016/j.neuron.2008.01.003 . OpenUrl CrossRef PubMed Web of Science 116. ↵ Riva , N. , Iannaccone , S. , Corbo , M. , Casellato , C. , Sferrazza , B. , Lazzerini , A. , Scarlato , M. , Cerri , F. , Previtali , S.C. , Nobile-Orazio , E. , et al. ( 2011 ). Motor nerve biopsy: clinical usefulness and histopathological criteria . Ann Neurol 69 , 197 – 201 . doi: 10.1002/ana.22110 . OpenUrl CrossRef PubMed 117. ↵ Beachy , N. , Satkowiak , K. , and Gwathmey , K.G . ( 2019 ). Vasculitic Neuropathies . Semin Neurol 39 , 608 – 619 . doi: 10.1055/s-0039-1688990 . OpenUrl CrossRef PubMed 118. ↵ Collins , M.P. , and Hadden , R.D . ( 2017 ). The nonsystemic vasculitic neuropathies . Nat Rev Neurol 13 , 302 – 316 . doi: 10.1038/nrneurol.2017.42 . OpenUrl CrossRef PubMed 119. ↵ Taib , S. , Durand , J. , Dehais , V. , Boulay , A.C. , Martin , S. , Blugeon , C. , Jourdren , L. , Freydier , R. , Cohen-Salmon , M. , Hazan , J. , and Brunet , I . ( 2025 ). Vascular dysfunction is at the onset of oxaliplatin-induced peripheral neuropathy symptoms in mice . Life Sci Alliance 8 . doi: 10.26508/lsa.202402791 . OpenUrl Abstract / FREE Full Text 120. ↵ Akassoglou , K. , Yu , W.M. , Akpinar , P. , and Strickland , S . ( 2002 ). Fibrin inhibits peripheral nerve remyelination by regulating Schwann cell differentiation . Neuron 33 , 861 – 875 . doi: 10.1016/s0896-6273(02)00617-7 . OpenUrl CrossRef PubMed Web of Science 121. ↵ Dean , T. , Mendiola , A.S. , Yan , Z. , Meza-Acevedo , R. , Cabriga , B. , Akassoglou , K. , and Ryu , J.K . ( 2024 ). Fibrin promotes oxidative stress and neuronal loss in traumatic brain injury via innate immune activation . J Neuroinflammation 21 , 94 . doi: 10.1186/s12974-024-03092-w . OpenUrl CrossRef PubMed 122. ↵ Ryu , J.K. , Yan , Z. , Montano , M. , Sozmen , E.G. , Dixit , K. , Suryawanshi , R.K. , Matsui , Y. , Helmy , E. , Kaushal , P. , Makanani , S.K. , et al. ( 2024 ). Fibrin drives thromboinflammation and neuropathology in COVID-19 . Nature 633 , 905 – 913 . doi: 10.1038/s41586-024-07873-4 . OpenUrl CrossRef PubMed 123. ↵ Yoshikawa , M. , Aizawa , S. , Oppenheim , R.W. , and Milligan , C . ( 2022 ). Neurovascular unit pathology is observed very early in disease progression in the mutant SOD1(G93A) mouse model of amyotrophic lateral sclerosis . Exp Neurol 353 , 114084 . doi: 10.1016/j.expneurol.2022.114084 . OpenUrl CrossRef PubMed 124. ↵ Palomo , M. , Moreno-Castano , A.B. , Salas , M.Q. , Escribano-Serrat , S. , Rovira , M. , Guillen-Olmos , E. , Fernandez , S. , Ventosa-Capell , H. , Youssef , L. , Crispi , F. , et al. ( 2023 ). Endothelial activation and damage as a common pathological substrate in different pathologies and cell therapy complications . Front Med (Lausanne ) 10 , 1285898 . doi: 10.3389/fmed.2023.1285898 . OpenUrl CrossRef PubMed 125. ↵ Zhang , J . ( 2022 ). Biomarkers of endothelial activation and dysfunction in cardiovascular diseases . Rev Cardiovasc Med 23 , 73 . doi: 10.31083/j.rcm2302073 . OpenUrl CrossRef PubMed 126. ↵ De Caterina , R. , Libby , P. , Peng , H.B. , Thannickal , V.J. , Rajavashisth , T.B. , Gimbrone , M.A. , Jr. , Shin , W.S. , and Liao , J.K. ( 1995 ). Nitric oxide decreases cytokine-induced endothelial activation. Nitric oxide selectively reduces endothelial expression of adhesion molecules and proinflammatory cytokines . J Clin Invest 96 , 60 – 68 . doi: 10.1172/JCI118074 . OpenUrl CrossRef PubMed Web of Science 127. ↵ Kempe , S. , Kestler , H. , Lasar , A. , and Wirth , T . ( 2005 ). NF-kappaB controls the global pro-inflammatory response in endothelial cells: evidence for the regulation of a pro-atherogenic program . Nucleic Acids Res 33 , 5308 – 5319 . doi: 10.1093/nar/gki836 . OpenUrl CrossRef PubMed Web of Science 128. ↵ Sehnert , B. , Burkhardt , H. , Wessels , J.T. , Schroder , A. , May , M.J. , Vestweber , D. , Zwerina , J. , Warnatz , K. , Nimmerjahn , F. , Schett , G. , et al. ( 2013 ). NF-kappaB inhibitor targeted to activated endothelium demonstrates a critical role of endothelial NF-kappaB in immune-mediated diseases . Proc Natl Acad Sci U S A 110 , 16556 – 16561 . doi: 10.1073/pnas.1218219110 . OpenUrl Abstract / FREE Full Text 129. ↵ Walchli , T. , Ghobrial , M. , Schwab , M. , Takada , S. , Zhong , H. , Suntharalingham , S. , Vetiska , S. , Gonzalez , D.R. , Wu , R. , Rehrauer , H. , et al. ( 2024 ). Single-cell atlas of the human brain vasculature across development, adulthood and disease . Nature 632 , 603 – 613 . doi: 10.1038/s41586-024-07493-y . OpenUrl CrossRef 130. ↵ Orsenigo , F. , Conze , L.L. , Jauhiainen , S. , Corada , M. , Lazzaroni , F. , Malinverno , M. , Sundell , V. , Cunha , S.I. , Brannstrom , J. , Globisch , M.A. , et al. ( 2020 ). Mapping endothelial-cell diversity in cerebral cavernous malformations at single-cell resolution . Elife 9 . doi: 10.7554/eLife.61413 . OpenUrl CrossRef PubMed 131. ↵ Yau , A.C.Y. , Globisch , M.A. , Onyeogaziri , F.C. , Conze , L.L. , Smith , R. , Jauhiainen , S. , Corada , M. , Orsenigo , F. , Huang , H. , Herre , M. , et al. ( 2022 ). Inflammation and neutrophil extracellular traps in cerebral cavernous malformation . Cell Mol Life Sci 79 , 206 . doi: 10.1007/s00018-022-04224-2 . OpenUrl CrossRef 132. ↵ Fournier , A.P. , Tastet , O. , Charabati , M. , Hoornaert , C. , Bourbonniere , L. , Klement , W. , Larouche , S. , Tea , F. , Wang , Y.C. , Larochelle , C. , et al. ( 2023 ). Single-Cell Transcriptomics Identifies Brain Endothelium Inflammatory Networks in Experimental Autoimmune Encephalomyelitis . Neurol Neuroimmunol Neuroinflamm 10 . doi: 10.1212/NXI.0000000000200046 . OpenUrl Abstract / FREE Full Text 133. ↵ Buchtele , N. , Schwameis , M. , Gilbert , J.C. , Schorgenhofer , C. , and Jilma , B . ( 2018 ). Targeting von Willebrand Factor in Ischaemic Stroke: Focus on Clinical Evidence . Thromb Haemost 118 , 959 – 978 . doi: 10.1055/s-0038-1648251 . OpenUrl CrossRef PubMed 134. De Meyer , S.F. , Stoll , G. , Wagner , D.D. , and Kleinschnitz , C. ( 2012 ). von Willebrand factor: an emerging target in stroke therapy . Stroke 43 , 599 – 606 . doi: 10.1161/STROKEAHA.111.628867 . OpenUrl Abstract / FREE Full Text 135. ↵ Gragnano , F. , Sperlongano , S. , Golia , E. , Natale , F. , Bianchi , R. , Crisci , M. , Fimiani , F. , Pariggiano , I. , Diana , V. , Carbone , A. , et al. ( 2017 ). The Role of von Willebrand Factor in Vascular Inflammation: From Pathogenesis to Targeted Therapy . Mediators Inflamm 2017 , 5620314 . doi: 10.1155/2017/5620314 . OpenUrl CrossRef PubMed 136. ↵ Singh , R.B. , Naderi , A. , Cho , W. , Ortiz , G. , Musayeva , A. , Dohlman , T.H. , Chen , Y. , Ferrari , G. , and Dana , R . ( 2022 ). Modulating the tachykinin: Role of substance P and neurokinin receptor expression in ocular surface disorders . Ocul Surf 25 , 142 – 153 . doi: 10.1016/j.jtos.2022.06.007 . OpenUrl CrossRef PubMed 137. ↵ Camilli , C. , Hoeh , A.E. , De Rossi , G. , Moss , S.E. , and Greenwood , J. ( 2022 ). LRG1: an emerging player in disease pathogenesis . J Biomed Sci 29 , 6 . doi: 10.1186/s12929-022-00790-6 . OpenUrl CrossRef PubMed 138. ↵ Knowland , D. , Arac , A. , Sekiguchi , K.J. , Hsu , M. , Lutz , S.E. , Perrino , J. , Steinberg , G.K. , Barres , B.A. , Nimmerjahn , A. , and Agalliu , D . ( 2014 ). Stepwise recruitment of transcellular and paracellular pathways underlies blood-brain barrier breakdown in stroke . Neuron 82 , 603 – 617 . doi: 10.1016/j.neuron.2014.03.003 . OpenUrl CrossRef PubMed Web of Science 139. ↵ Liebner , S. , Corada , M. , Bangsow , T. , Babbage , J. , Taddei , A. , Czupalla , C.J. , Reis , M. , Felici , A. , Wolburg , H. , Fruttiger , M. , et al. ( 2008 ). Wnt/beta-catenin signaling controls development of the blood-brain barrier . J Cell Biol 183 , 409 – 417 . doi: 10.1083/jcb.200806024 . OpenUrl Abstract / FREE Full Text 140. ↵ Daneman , R. , Agalliu , D. , Zhou , L. , Kuhnert , F. , Kuo , C.J. , and Barres , B.A . ( 2009 ). Wnt/beta-catenin signaling is required for CNS, but not non-CNS, angiogenesis . Proc Natl Acad Sci U S A 106 , 641 – 646 . doi: 10.1073/pnas.0805165106 . OpenUrl Abstract / FREE Full Text 141. ↵ Eser Ocak , P. , Ocak , U. , Sherchan , P. , Gamdzyk , M. , Tang , J. , and Zhang , J.H. ( 2020 ). Overexpression of Mfsd2a attenuates blood brain barrier dysfunction via Cav-1/Keap-1/Nrf-2/HO-1 pathway in a rat model of surgical brain injury . Exp Neurol 326 , 113203 . doi: 10.1016/j.expneurol.2020.113203 . OpenUrl CrossRef PubMed 142. Park , M.H. , Lee , J.Y. , Park , K.H. , Jung , I.K. , Kim , K.T. , Lee , Y.S. , Ryu , H.H. , Jeong , Y. , Kang , M. , Schwaninger , M. , et al. ( 2018 ). Vascular and Neurogenic Rejuvenation in Aging Mice by Modulation of ASM . Neuron 100 , 167 – 182 e169 . doi: 10.1016/j.neuron.2018.09.010 . OpenUrl CrossRef PubMed 143. ↵ Yang , J.M. , Park , C.S. , Kim , S.H. , Noh , T.W. , Kim , J.H. , Park , S. , Lee , J. , Park , J.R. , Yoo , D. , Jung , H.H. , et al. ( 2020 ). Dll4 Suppresses Transcytosis for Arterial Blood-Retinal Barrier Homeostasis . Circ Res 126 , 767 – 783 . doi: 10.1161/CIRCRESAHA.119.316476 . OpenUrl CrossRef PubMed 144. ↵ Akassoglou , K. , Davalos , D. , Mendiola , A.S. , Petersen , M.A. , Ryu , J.K. , Schachtrup , C. , and Yan , Z . ( 2024 ). Pioneering discovery and therapeutics at the brain-vascular-immune interface . Cell 187 , 5871 – 5876 . doi: 10.1016/j.cell.2024.09.018 . OpenUrl CrossRef PubMed 145. ↵ Mastorakos , P. , and McGavern , D . ( 2019 ). The anatomy and immunology of vasculature in the central nervous system . Sci Immunol 4 . doi: 10.1126/sciimmunol.aav0492 . OpenUrl Abstract / FREE Full Text 146. ↵ Theofilis , P. , Sagris , M. , Oikonomou , E. , Antonopoulos , A.S. , Siasos , G. , Tsioufis , C. , and Tousoulis , D . ( 2021 ). Inflammatory Mechanisms Contributing to Endothelial Dysfunction . Biomedicines 9 . doi: 10.3390/biomedicines9070781 . OpenUrl CrossRef PubMed 147. ↵ Lopez-Lopez , A. , Gamez , J. , Syriani , E. , Morales , M. , Salvado , M. , Rodriguez , M.J. , Mahy , N. , and Vidal-Taboada , J.M . ( 2014 ). CX3CR1 is a modifying gene of survival and progression in amyotrophic lateral sclerosis . PLoS One 9 , e96528 . doi: 10.1371/journal.pone.0096528 . OpenUrl CrossRef PubMed 148. ↵ Beard , R.S. , Jr. , Hoettels , B.A. , McAllister , J.M. , Meegan , J.E. , Wertz , T.S. , Self , D.A. , Hrkach , D.E. , Greiner , D. , Chapman , K. , Villalba , N. , et al. ( 2025 ). Progression of experimental autoimmune encephalomyelitis in mice and neutrophil-mediated blood-brain barrier dysfunction requires non-muscle myosin light chain kinase . J Cereb Blood Flow Metab 45 , 1203 – 1220 . doi: 10.1177/0271678X251318620 . OpenUrl CrossRef PubMed 149. ↵ Mora , J.S. , Bradley , W.G. , Chaverri , D. , Hernandez-Barral , M. , Mascias , J. , Gamez , J. , Gargiulo-Monachelli , G.M. , Moussy , A. , Mansfield , C.D. , Hermine , O. , and Ludolph , A.C . ( 2021 ). Long-term survival analysis of masitinib in amyotrophic lateral sclerosis . Ther Adv Neurol Disord 14 , 17562864211030365 . doi: 10.1177/17562864211030365 . OpenUrl CrossRef PubMed 150. ↵ Zamudio , F. , Loon , A.R. , Smeltzer , S. , Benyamine , K. , Navalpur Shanmugam , N.K. , Stewart , N.J.F. , Lee , D.C. , Nash , K. , and Selenica , M.B . ( 2020 ). TDP-43 mediated blood-brain barrier permeability and leukocyte infiltration promote neurodegeneration in a low-grade systemic inflammation mouse model . J Neuroinflammation 17 , 283 . doi: 10.1186/s12974-020-01952-9 . OpenUrl CrossRef PubMed 151. ↵ Hatzipetros , T. , Kidd , J.D. , Moreno , A.J. , Thompson , K. , Gill , A. , and Vieira , F.G . ( 2015 ). A Quick Phenotypic Neurological Scoring System for Evaluating Disease Progression in the SOD1-G93A Mouse Model of ALS . J Vis Exp . doi: 10.3791/53257 . OpenUrl CrossRef PubMed 152. ↵ Madisen , L. , Zwingman , T.A. , Sunkin , S.M. , Oh , S.W. , Zariwala , H.A. , Gu , H. , Ng , L.L. , Palmiter , R.D. , Hawrylycz , M.J. , Jones , A.R. , et al. ( 2010 ). A robust and high-throughput Cre reporting and characterization system for the whole mouse brain . Nat Neurosci 13 , 133 – 140 . doi: 10.1038/nn.2467 . OpenUrl CrossRef PubMed Web of Science 153. ↵ Saederup , N. , Cardona , A.E. , Croft , K. , Mizutani , M. , Cotleur , A.C. , Tsou , C.L. , Ransohoff , R.M. , and Charo , I.F . ( 2010 ). Selective chemokine receptor usage by central nervous system myeloid cells in CCR2-red fluorescent protein knock-in mice . PLoS One 5 , e13693 . doi: 10.1371/journal.pone.0013693 . OpenUrl CrossRef PubMed 154. ↵ Bolger , A.M. , Lohse , M. , and Usadel , B . ( 2014 ). Trimmomatic: a flexible trimmer for Illumina sequence data . Bioinformatics 30 , 2114 – 2120 . doi: 10.1093/bioinformatics/btu170 . OpenUrl CrossRef PubMed Web of Science 155. ↵ Dobin , A. , Davis , C.A. , Schlesinger , F. , Drenkow , J. , Zaleski , C. , Jha , S. , Batut , P. , Chaisson , M. , and Gingeras , T.R . ( 2013 ). STAR: ultrafast universal RNA-seq aligner . Bioinformatics 29 , 15 – 21 . doi: 10.1093/bioinformatics/bts635 . OpenUrl CrossRef PubMed Web of Science 156. ↵ Liao , Y. , Smyth , G.K. , and Shi , W . ( 2014 ). featureCounts: an efficient general purpose program for assigning sequence reads to genomic features . Bioinformatics 30 , 923 – 930 . doi: 10.1093/bioinformatics/btt656 . OpenUrl CrossRef PubMed Web of Science 157. ↵ Love , M.I. , Huber , W. , and Anders , S . ( 2014 ). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 . Genome Biol 15 , 550 . doi: 10.1186/s13059-014-0550-8 . OpenUrl CrossRef PubMed 158. ↵ Chen , E.Y. , Tan , C.M. , Kou , Y. , Duan , Q. , Wang , Z. , Meirelles , G.V. , Clark , N.R. , and Ma’ayan , A . ( 2013 ). Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool . BMC Bioinformatics 14 , 128 . doi: 10.1186/1471-2105-14-128 . OpenUrl CrossRef PubMed 159. ↵ Yu , G. , Wang , L.G. , Han , Y. , and He , Q.Y . ( 2012 ). clusterProfiler: an R package for comparing biological themes among gene clusters . OMICS 16 , 284 – 287 . doi: 10.1089/omi.2011.0118 . OpenUrl CrossRef PubMed Web of Science 160. ↵ Skene , N.G. , and Grant , S.G . ( 2016 ). Identification of Vulnerable Cell Types in Major Brain Disorders Using Single Cell Transcriptomes and Expression Weighted Cell Type Enrichment . Front Neurosci 10 , 16 . doi: 10.3389/fnins.2016.00016 . OpenUrl CrossRef 161. ↵ Newman , A.M. , Steen , C.B. , Liu , C.L. , Gentles , A.J. , Chaudhuri , A.A. , Scherer , F. , Khodadoust , M.S. , Esfahani , M.S. , Luca , B.A. , Steiner , D. , et al. ( 2019 ). Determining cell type abundance and expression from bulk tissues with digital cytometry . Nat Biotechnol 37 , 773 – 782 . doi: 10.1038/s41587-019-0114-2 . OpenUrl CrossRef PubMed 162. ↵ Zheng , G.X. , Terry , J.M. , Belgrader , P. , Ryvkin , P. , Bent , Z.W. , Wilson , R. , Ziraldo , S.B. , Wheeler , T.D. , McDermott , G.P. , Zhu , J. , et al. ( 2017 ). Massively parallel digital transcriptional profiling of single cells . Nat Commun 8 , 14049 . doi: 10.1038/ncomms14049 . OpenUrl CrossRef PubMed 163. ↵ Hao , Y. , Hao , S. , Andersen-Nissen , E. , Mauck , W.M ., 3rd . , Zheng , S. , Butler , A. , Lee , M.J. , Wilk , A.J. , Darby , C. , Zager , M. , et al. ( 2021 ). Integrated analysis of multimodal single-cell data . Cell 184 , 3573 – 3587 e3529 . doi: 10.1016/j.cell.2021.04.048 . OpenUrl CrossRef PubMed 164. ↵ Korsunsky , I. , Millard , N. , Fan , J. , Slowikowski , K. , Zhang , F. , Wei , K. , Baglaenko , Y. , Brenner , M. , Loh , P.R. , and Raychaudhuri , S . ( 2019 ). Fast, sensitive and accurate integration of single-cell data with Harmony . Nat Methods 16 , 1289 – 1296 . doi: 10.1038/s41592-019-0619-0 . OpenUrl CrossRef PubMed 165. ↵ Hafemeister , C. , and Satija , R . ( 2019 ). Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression . Genome Biol 20 , 296 . doi: 10.1186/s13059-019-1874-1 . OpenUrl CrossRef PubMed 166. ↵ Brunson , J.C . ( 2020 ). ggalluvial: Layered Grammar for Alluvial Plots . J Open Source Softw 5 . doi: 10.21105/joss.02017 . OpenUrl CrossRef PubMed 167. ↵ R, K. ( 2025 ). pheatmap: Pretty Heatmaps . R package version 1.0.13 . https://github.com/raivokolde/pheatmap . View the discussion thread. Back to top Previous Next Posted September 04, 2025. Download PDF Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Early endothelial activation at the blood–nerve barrier defines a hallmark of ALS Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Early endothelial activation at the blood–nerve barrier defines a hallmark of ALS Ganesh Parameshwar Bhat , Santo Diprima , Cristian G Beccaria , Paola Podini , Aurora Badaloni , Alessia Motta , Ilaria Brambilla , Luis Martins , Daniela Boselli , Stefano de Pretis , Giovanni Nardo , Caterina Bendotti , Nilo Riva , Matteo Iannacone , Angelo Quattrini , Dario Bonanomi bioRxiv 2025.08.31.673224; doi: https://doi.org/10.1101/2025.08.31.673224 Share This Article: Copy Citation Tools Early endothelial activation at the blood–nerve barrier defines a hallmark of ALS Ganesh Parameshwar Bhat , Santo Diprima , Cristian G Beccaria , Paola Podini , Aurora Badaloni , Alessia Motta , Ilaria Brambilla , Luis Martins , Daniela Boselli , Stefano de Pretis , Giovanni Nardo , Caterina Bendotti , Nilo Riva , Matteo Iannacone , Angelo Quattrini , Dario Bonanomi bioRxiv 2025.08.31.673224; doi: https://doi.org/10.1101/2025.08.31.673224 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Neuroscience Subject Areas All Articles Animal Behavior and Cognition (7622) Biochemistry (17645) Bioengineering (13867) Bioinformatics (41873) Biophysics (21420) Cancer Biology (18550) Cell Biology (25447) Clinical Trials (138) Developmental Biology (13361) Ecology (19866) Epidemiology (2067) Evolutionary Biology (24289) Genetics (15587) Genomics (22473) Immunology (17707) Microbiology (40322) Molecular Biology (17144) Neuroscience (88457) Paleontology (666) Pathology (2826) Pharmacology and Toxicology (4815) Physiology (7634) Plant Biology (15111) Scientific Communication and Education (2042) Synthetic Biology (4285) Systems Biology (9813) Zoology (2268)