Maternal alloimmune IgG causes anti-glomerular basement membrane disease in perinatal transgenic mice that express human laminin α5

In characterizing transgenic mice expressing human laminin α5, immunofluorescence showed that human laminin α5 is expressed exactly like mouse laminin α5 in kidney GBMs and many TBMs 12 . Specifically, when frozen kidney sections from newborn hu LAMA5 mice were sequentially immunolabeled with mouse monoclonal anti-human laminin α5 IgG (mAb 4C7) and then with anti-mouse IgG, GBMs were intensely labeled ( Fig. 1 ). Slides processed for negative controls included sections of transgenic kidneys labeled only with anti-mouse IgG, and these were negative ( Fig. 1 ). As expected, kidney sections from wildtype sibs not expressing human laminin α5 did not immunolabel with the 4C7 antibody, and control anti-mouse IgG antibodies also failed to bind ( Fig. 1 ). During the course of these experiments, however, we discovered instances in which GBMs in sections of hu LAMA5 kidneys, when labeled with anti-mouse IgG alone (no primary mouse IgG), were intensely positive ( Fig. 2 ) indicating the presence of endogenous mouse IgG within glomeruli of some transgenic pups. These unexpected results led us to analyze kidney sections from entire litters of mice, paying particular attention to how the parental crosses were made, the numbers of litters born from each dam, and the genotypes of all of the pups. When litters from wildtype females that had been mated two or more times with hu LAMA5 males were examined, GBMs in every newborn that genotyped transgenic (23 pups from 6 different parental pairs) labeled intensely with anti-human laminin α5 IgG as expected, confirming expression of human laminin α5 protein in hu LAMA5 transgenic mouse GBMs ( Fig. 2 ). When sections from these same 23 transgenic pups were labeled with anti-mouse IgG, but without primary mouse anti-human laminin α5 (either mAb 4B5 or 4C7), all GBMs were also positive ( Fig. 2 ). By contrast, kidney sections from wildtype littermates (13 pups from the same, 6 litters) were consistently negative for the presence of human laminin α5, and were also always negative when incubated with anti-mouse IgG alone ( Fig 2 ). On the other hand, when hu LAMA5 females were crossed with wildtype or with transgenic males, kidneys from their transgenic progeny also expressed human laminin α5 (6 pups from 2 litters) ( Fig. 1 ). No labeling was seen however, when these sections were incubated with anti-mouse IgG alone ( Fig. 1 ). To test whether the mouse IgG seen in glomeruli of newborn hu LAMA5 transgenics was maternally derived, we incubated maternal sera collected from two different wildtype mice bred with hu LAMA5 transgenic males on frozen sections of human kidney, and then immunolabeled these sections with anti-mouse IgG secondary antibody. For both sera, we found mouse IgG bound to human GBMs and TBMs, confirming the presence of IgG against human laminin in maternal sera ( Fig. 3 ). When sera collected from wildtype breeders before mating with transgenic males was incubated on human kidney sections, however, there was no binding of mouse IgG ( Fig. 3 .). Similarly, sera collected from hu LAMA5 females bred with wildtype male mice also failed to bind human kidney sections. Based on these results, we reasoned that wildtype females mated with hu LAMA5 males were developing an alloimmune response against fetal-derived human laminin α5, and that the mouse IgG observed within kidney sections from hu LAMA5 pups represented GBM-bound maternal alloantibody. We therefore designated wildtype females serially bred with hu LAMA5 transgenic males as “immune” females. In contrast, hu LAMA5 transgenic females mated with wildtype males failed to develop a detectable anti-human laminin α5 antibody response, probably because these mice had lifetime exposure to the human laminin α5 alloantigen (and therefore developed immune self-tolerance), and we designated these animals “non-immune” females. To confirm that the maternal alloantibody was directed specifically against laminin isoforms containing the human α5 chain, sera collected from 5 immune and 3 non-immune females were tested by ELISA on plates coated with different human laminin isoforms. Immune sera showed significant IgG binding to laminin α5β1γ1 (median OD 1.30, IQR 0.76-1.55, p=0.0004) and α5β2γ1 (median OD 1.58, IQR 1.26-1.67, p<0.0001), confirming the laminin α5-specificity of the maternal alloimmune IgG ( Fig. 4A ). Non-immune sera were negative on all laminin isoforms (data not shown). Because mouse IgG2a, IgG2b, and IgG3, but not IgG1, activates complement via the classical pathway, we further analyzed the IgG subclasses of maternal anti-laminin antibodies. Sera from 3 immune females contained anti-human laminin α5β2γ1 antibodies of IgG1 subclass (OD range 1.42-2.16), as well as complement-fixing IgG2a (OD range 0.69-1.8) and IgG2b (OD range 0.43-1.14), whereas IgG3 was not detected ( Fig. 4B ). Non-immune sera were negative in this assay (OD less than 0.13). To determine if maternal anti-human laminin α5 IgG was transferred to transgenic mice before birth, kidneys from embryonic day 18.5 transgenic and wildtype mice harvested from pregnant immune females were immunolabeled with anti-mouse IgG (3 mice each genotype). Capillary loop stage and maturing glomeruli of transgenic fetal kidneys contained GBM-bound mouse IgG, but GBMs of vascular cleft and comma-shaped nephrons, which represented earlier stages of development from these same kidneys, were only faintly labeled ( Fig. 5 ). We interpreted the binding of maternal IgG to GBMs in capillary loop and maturing stage glomeruli of fetal kidneys to reflect the temporal upregulation of laminin α5 expression normally seen during glomerular development and maturation. Specifically, GBMs of vascular cleft and comma-stage glomeruli contain relatively abundant laminin α1 but little laminin α5. In early capillary loop and maturing stage glomeruli, however, laminin α5 is upregulated and becomes the sole laminin α chain found in GBMs thereafter. 4 , 5 Regardless, our findings indicated passage of maternal IgG across the placenta before birth. In contrast, GBMs in kidneys from wildtype fetuses were entirely negative at all stages of glomerular development. To test whether maternal IgG was also transferred to transgenic pups after birth, one litter of 6 mice from a non-immune breeder, which did not harbor anti-human laminin α5 IgG, was foster-nursed on one lactating immune female. Kidneys were harvested 48 hrs after the inception of foster nursing. Kidney sections from 4 hu LAMA5 pups were positive for maternal IgG in GBMs ( Fig. 6 ), indicating postnatal transfer of IgG in maternal milk. Kidneys from 2 wildtype, foster nursed siblings were negative ( Fig. 6 ). Although postnatal IgG transfer was experimentally documented in just one litter, GBMs in all glomeruli of 21 day old huLAMA5 pups contained maternal IgG, including those in subcapsular nephrons induced to form after birth. Together, these results demonstrated that hu LAMA5 fetuses and pups from immune dams were exposed to maternal anti-human laminin α5 IgG beginning in mid-gestation and continuing throughout suckling. To evaluate the development of the maternal anti-laminin α5 immune response over multiple pregnancies, sera was serially collected from two wildtype female mice bred up to 5 times with hu LAMA5 males and anti-human laminin α5 IgG titers were measured by ELISA. As shown in Figure 7A , there was little maternal anti-human laminin α5 IgG detected in sera collected after delivery of the first litter (OD less than 0.17). Beginning with the birth of the second litter, however, and continuing thereafter, significant anti-human laminin α5 titers were observed (OD range 0.96-3.0). To assess possible pathophysiological effects of maternal anti-human laminin α5 on hu LAMA5 pups, urine was collected from pups born from immune female M17 and samples were electrophoresed in polyacrylamide gels to detect urinary albumin. Urinary albumin was not detected in transgenic or wildtype pups from the first litter when maternal anti-human alloantibodies were not yet present ( Fig. 7A and B ). Although there were some variations, most transgenic pups from the second litter were albuminuric and those in the third litter excreted larger amounts of albumin ( Figure 7B ). In contrast, urinary albumin was absent from all of the wildtype pups sampled at all timepoints ( Fig. 7B ). Speculating that the proteinuria seen in transgenic pups born from immune females was caused by immunologic damage mediated by GBM-bound maternal anti-human laminin α5 IgG, we immunolabeled kidney sections with anti-mouse IgG2a to detect presence of this complement-activating IgG isoform. We also evaluated glomerular deposition of mouse complement C3c and C4. As shown in Figure 8 , GBM-bound maternal IgG2a was seen in all transgenic pups examined (25 pups from 9 separate litters), and this was also almost always accompanied by glomerular C3 deposition (24/25 pups). Immunofluorescence labeling of sections from 3 of these IgG2a+, C3+ animals showed that transgenic glomeruli were also positive for C3c and C4 ( Fig. 8 ), which indicates complement activation via the classical pathway. Sections from wildtype sibs (18 pups from the same 9 litters) were entirely negative for IgG2a and C3, and one of these control animals was also negative for C3c and C4 ( Fig. 8 ). Finally, 21 day-old hu LAMA5 glomeruli contained more CD45+ leukocyte infiltrates than wildtype littermates born from the same immune female ( Fig. 9 ). The CD45+ cell counts averaged ~1.5 cells/glomerular cross section (cgs) in hu LAMA5 glomeruli vs. ~0.9 cells/gcs in wildtype, which approximates those seen in other models of glomerulonephritis in mice. 13 The most common features observed by electron microscopy were the presence of neutrophil and mononuclear cell infiltrates within glomerular capillaries, and podocyte foot process effacement ( Fig. 10 , representative from samples from 11 animals), which are hallmarks of immunologically-mediated glomerular damage. There were also broad areas of thickened and split GBMs, however, which is not commonly seen in most cases of anti-GBM disease ( Fig. 10 ). All wildtype animals examined (5) had normal glomerular morphology.

All animal experiments were approved by the University of Kansas Medical Center Institutional Animal Care and Use Committee. Mice expressing human laminin α5 were originally engineered as described 12 . Wildtype (129X1/SvJ) mice were obtained from Jackson Laboratories, Bar Harbor, ME. For most experiments, hu LAMA5 transgenic males were crossed with wildtype females, and progeny were genotyped by PCR using primers designed to hybridize specifically with human LAMA5 12 In some cases, hu LAMA5 transgenic females (nonimmune mice) were crossed with wildtype males. For cross-fostering studies, newborn pups from these matings were nursed on wildtype dams (immune females) that had previously been crossed with hu LAMA5 transgenic males. For immunofluorescence, mouse kidneys were frozen in isopentane chilled in a dry ice-acetone bath. Cryosections were fixed with 100% methanol and immunolabeled. Antibodies used in this work were: Alexa Fluor 488-conjugated goat anti-mouse IgG, Alexa Fluor 594-conjugated goat anti-mouse IgG2a, Alexa Fluor 488-conjugated anti-rat IgG (all from Invitrogen, Carlsbad, CA), Alexa Fluor 488-conjugated rat monoclonal anti-CD45 (Novus Biologicals, Centennial, CO), mouse monoclonal antihuman laminin α5 (clones 4C7 and 4B5; characterized as previously described) 53 , 54 , rat monoclonal IgG against mouse C4 (Hycult Biotech, Wayne, PA), fluorescein-conjugated goat anti-mouse C3 (MP Cappel/Fisher Bioscience), and FITC-conjugated goat antimouse C3c and C4 (Nordic-MUbio, Susteren, Netherlands). For some experiments, mouse anti-human laminin α5 clone 4B5 was directly conjugated to Alexa Fluor 488 using the Zenon mouse IgG1 kit (Invitrogen). In other experiments, serum samples from immune and non-immune mice were incubated on cryosections of human kidney, which was obtained from the PKD Biomaterials Core at KU Medical Center. Fluorescence images were obtained with a Leica TCS SPE confocal microscope. For electron microscopy, kidneys were fixed in situ with Karnovsky’s fixative and postfixed in Palade’s osmium, dehydrated and embedded in Polybed 812. Ultrathin sections were stained with lead citrate and uranyl acetate and examined with a JEOL JEM 1400 electron microscope. At various dates after birth, drops of urine gently expressed from newborn mice were collected in microfuge tubes, and 6μl urine samples, alongside lanes containing 5μg and 0ĝ bovine serum albumin (BSA), were electrophoresed in 12% polyacrylamide gels. Gels were stained with Coomassie Blue. Recombinant human laminins α1β1γ1, α1β2γ1, α2β1γ1, α2β2γ1, α4β1γ1, α4β2γ1, α5β1γ1, and α5β2γ1, were purchased from Biolamina (Sundbyberg, Sweden). ELISA was carried out in Nunc Maxisorp 96-well polystyrene plates (Thermo Fisher) coated overnight with 100ng laminin per well in 50mM carbonate buffer, pH 9.6, and blocked with 1% bovine serum albumin (BSA). After incubation with mouse sera (diluted 1/500-1/1000), binding of mouse IgG or IgG subclasses to laminin-coated wells was detected using alkaline phosphatase-conjugated goat anti-mouse IgG (Rockland Immunochemicals, Gilbertsville, PA) or horseradish peroxidase-conjugated goat antimouse IgG1, IgG2a, or IgG2b (Bethyl, Montgomery, TX). After the addition of colorimetric substrate, absorbance was read with an ELISA plate reader. Statistical analyses were performed using GraphPad Prism software (San Diego, CA). IgG binding to laminin was expressed as optical density (OD) values and summarized as medians and range (or interquartile range). Significance of differences between two groups was evaluated by Mann-Whitney test. To gauge anti-laminin specificity further, IgG binding to each laminin isoform was assessed using Friedman test followed by Dunn’s post-hoc tests ( Fig 4B ). CD45 staining of hu LAMA5 vs wildtype glomeruli was expressed as the number of CD45+ cells present per glomerular crosssection, and significance was evaluated by unpaired t test ( Fig. 9C ). A value of P < 05, adjusted for multiple comparisons where applicable, was considered statistically significant.

Our results from this novel in vivo transgenic system show that an allovariant protein (human laminin α5), which is expressed by fetal transgenic mice in its proper anatomical microdomain (basement membranes), induced an alloimmune humoral response in wildtype pregnant mice. This resulted in transmission of maternal anti-human laminin α5 IgG to fetal and neonatal transgenic progeny, where the maternal IgG behaved as an autoantibody and caused glomerulonephritis in the newborn. The immunologically-mediated glomerular disease bore features typical of anti-GBM nephritis: linear, ribbon-like pattern of GBM-bound IgG, leukocyte infiltration, and proteinuria. Further, production of complement-fixing IgG2a and IgG2b anti-laminin was associated with glomerular deposition of complement C3c and C4. Binding of these complement components may be essential for development of anti-GBM nephritis, since mice with kidney-bound anti-GBM antibodies restricted to the IgG1 subclass, which show no glomerular complement activation, exhibit normal kidney function and morphology. 14 , 15 Additionally, maternal alloimmunization was first detected during the second pregnancy, and the severity of proteinuria in neonates worsened with serial pregnancies and correlated with increased titers of maternal anti-human laminin α5 IgG. This unique system may provide a useful experimental model for studying antenatal glomerular disease and may shed more light on fetomaternal alloimmunization processes. Pregnancy represents an immunological paradox in which the mother routinely mounts an immune response against infectious pathogens, thereby protecting herself and her baby, and yet she also usually tolerates her semi-allogeneic fetus to preserve the pregnancy. In humans, maternal-fetal antibody transfer occurs predominantly before birth across the placenta but, in many mammalian species including rodents, also continues after birth across the intestine through suckling. The transfer is selective for the IgG class of immunoglobulins, occurs by transcytosis across placental syncytiotrophoblasts (and yolk sac endoderm in mice) 16 before birth and small intestinal epithelium postnatally. In mice, the distribution of IgG subclasses in newborn pups closely resembles that in maternal sera (IgG1>IgG2a=IgG2b=IgG3). 17 IgG selectivity and transcytosis through epithelial cells is mediated by the neonatal IgG Fc receptor (FcRn), which diverts internalized, receptor-bound IgG from lysosomal degradation intracellularly and allows for transcellular transport and release from basolateral plasma membranes. 18 , 19 Although maternal IgG transfer is generally beneficial, there are cases where maternal autoantibodies and alloantibodies are transferred to the fetus/infant, causing neonatal diseases such as anti-phospholipid syndrome, myasthenia gravis, bullous pemphigoid, systemic lupus erythematous, and perinatal membranous glomerulopathy (discussed further below). 18 The risks to the fetus can be severe, causing spontaneous abortion, stillbirth, preterm birth, low birth weight, and developmental and neurological abnormalities. 20 , 21 Further, if fetomaternal hemorrhage occurs, pregnant women can readily develop alloantibodies to paternally-derived, incompatible alloantigens on fetal red blood cells, platelets, and leukocytes, causing life-threatening fetal hemolytic disease. 22 Exactly how fetal human laminin α5 stimulated an immune response in pregnant wildtype mice is unclear. Laminins are evolutionarily ancient extracellular matrix proteins, 23 and there is more than 78% amino acid sequence identity between human and mouse laminin α5 (UniProt Consortium; https://www.uniprot.org/ ). Nonetheless, mice develop antibodies targeting human (but not murine) laminin α5 chain after immunization with human placenta laminin or with human laminin α5β1γ1 in adjuvant. 24 , 25 Several laminin protein isoforms, including those containing laminin α5, have been localized previously in human 26 and mouse 27 placentae. Conceivably, during formation and growth of the hemochorial arrangement of the labyrinthine placenta, there may be fragmentation or shedding of placental basement membranes, which could subsequently be phagocytosed and processed by maternal antigen presenting cells in the decidua, uterus-draining lymph nodes, or spleen. These cells, in turn, may then prime fetal antigen-reactive T cells in distal lymphoid tissues to establish a maternal humoral response against the extrinsic laminin α5. 28 Alternatively, there may be fetomaternal immune sensitization during parturition itself. Either possibility might explain the higher maternal anti-human laminin α5 titers observed with serial pregnancies, but much more work needs to be carried out to clarify how the hu LAMA5 transgenic fetuses immunized wildtype females. We also need to identify which domains of human laminin α5 stimulated the mouse immune response and elicited pathogenic antibodies. Anti-GBM glomerulonephritis is exceptionally rare in the pediatric population. 29 A recent literature review documented 27 cases of anti-GBM disease in children ranging in age from 19 months to 16 years old, often with poor outcomes, 30 but we have been unable to find any reported cases of definitive anti-GBM disease in newborn, infant humans. Another review reported 8 women with proven anti-GBM disease while pregnant. 31 Despite presence of circulating anti-GBM antibodies in at least some of the neonates, there were no kidney or lung diseases seen in the newborns, although most of the deliveries were premature. 31 Anti-GBM disease is more common in older individuals, where autoantibodies directed against the NC1 domain of the α3 chain of type IV collagen bind to the GBM and can initiate a rapidly progressive glomerulonephritis. 32 – 34 This is often accompanied by pulmonary hemorrhage due to the binding of autoantibodies to alveolar basement membranes as well (Goodpasture syndrome). Anti-type IV collagen autoimmunity seems to be triggered by exposure to endogenous or exogenous oxidants, environmental hydrocarbons, and/or tobacco smoke, but the precise mechanisms are obscure. 29 , 35 Treatment includes plasmapheresis to remove anti-GBM autoantibodies and immunosuppressive therapies, which can predispose patients to recurrent infections. Even with treatment, some patients develop end-stage renal failure and require dialysis or transplantation. 36 , 37 Patients with Alport syndrome, who have mutations in one or more of the COL4A3, COL4A4 , or COL4A5 genes, are unable to assemble and maintain a stable α3α4α5 collagen(IV) network in their GBMs, and usually progress to renal failure. If transplanted, some Alport patients develop an alloimmune response against the NCI domains of the α3, α4, and/or α5(IV) collagen present in the kidney graft, leading to anti-GBM nephritis and graft rejection. 32 , 38 , 39 In both Goodpasture or Alport post-transplant nephritis, however, the underlying auto- or alloantibodies arise directly from the affected individual, and not, as in our case, through maternal-fetal IgG transfer. Another immunologically mediated glomerular disease in humans is membranous nephropathy (MN), which frequently results in renal failure. Approximately 80% of MN patients present with primary (formerly “idiopathic”) MN, a disease mediated by autoantibodies forming subepithelial immune complexes with podocyte antigens, while 20% of MN cases are secondary to infections, cancers, or certain autoimmune diseases such as lupus. 40 Patients with MN exhibit glomerular IgG in granular patterns, activation of complement, widening of the GBM and proteinuria. In a search for the pathomechanisms causing primary MN in a subgroup of infants, neutral endopeptidase (NEP) was identified on their podocytes that was targeted by maternal antibodies. 41 Further analysis showed that the mothers had MME mutations (which encores NEP), and had developed alloantibodies to paternally-derived NEP, which is also expressed by placental syncitiotrophoblasts. 42 These anti-NEP antibodies crossed the placenta, formed subepithelial immune deposits on basal aspects of fetal podocytes, and caused antenatal MN. 41 , 42 Although the pathogenesis of anti-NEP MN is generally similar to our results with human laminin α5 alloimmunization, the histopathologic features we observed generally resembles anti-GBM disease and not MN. This is due to the localization of the laminin α5 alloantigen specifically within the GBM and not on the podocyte plasma membrane, as is the case for NEP. Unlike classic anti-GBM disease, however, we also observed splitting and lamination of the GBM. This may have been due to the deposition of maternal anti-laminin α5 into the GBM at a time when it is still undergoing active assembly, which may have interfered with laminin polymerization and/or its binding to integrins. 25 Antibody binding may have also disrupted the normal GBM fusion and splicing processes that occur during the course of glomerulogenesis. 4 , 43 We are currently designing new experiments to address these possibilities. We are also evaluating long term consequences of maternal IgG deposition into GBMs during infancy, and our preliminary findings show variable outcomes, ranging from premature death to survival with mild proteinuria. We suspect that this variability is due to maternal titers, and new studies are underway to clarify this likelihood. We are unaware of any previous reports regarding laminin α5 alloimmunization during pregnancy in humans. Very recently, however, a number of mutations have been identified in the human LAMA5 gene, making alloimmunization a feasible possibility. One report identified a dominant mutation in LAMA5 that causes multiple internal and external abnormalities affecting the eyes, muscles, gastrointestinal and endocrine systems, and skin. 44 Mutations in LAMA5 have also been found in three unrelated families with pediatric nephrotic syndrome, 45 and LAMA5 mutations have been implicated in cases of adult-onset focal glomeruloscerosis (FSGS). 46 , 47 Another recent paper reported findings that mutations to both COL4A5 and LAMA5 contribute to a complex, Alport-like spectrum including hematuria, proteinuria, FSGS, and renal cysts. 48 Taking all of the results together, the maternal anti-human laminin a5 alloimmune response described here may represent a new model for investigating perinatal glomerular disease as well as the immunopathology of placental basement membranes. Along these lines, we are currently addressing whether maternal anti-human laminin α5 binds to trophoblast or endothelial basement membranes of hu LAMA5 transgenic placenta. Interestingly, anti-laminin antibodies have previously been identified in sera of monkeys with histories of reproductive failure, 49 and circulating anti-laminin IgGs have also been found in infertile women with endometriosis. 50 Additionally, serum levels for circulating laminin protein have been shown to be significantly elevated in preeclamptic women compared to those with normal pregnancies, 51 although placental origin of the circulating laminin in these cases has not been proven. Finally, our results may represent a cautionary tale. Transgenic mice expressing human proteins (and those from other species) have become increasingly popular experimental tools for investigating a host of clinical problems including infectious disease, cancer biology and therapy, and transplantation. 52 Some of these models may have an overlooked, alloimmune component that could profoundly affect outcomes and interpretations.

Basement membranes are sheets of extracellular matrices that underlie all epithelial cells, including the vascular endothelium, and surround peripheral nerves, muscle cells, and adipocytes. They are composed of polymerized meshworks of laminin heterotrimers and type IV collagen heterotrimers, and also include nidogen, proteoglycans, and other glycoproteins. Assembly of basement membranes containing polymers of the laminin α1β1γ1 isoform are essential for early embryogenesis, and mice with genetic deletions of either the Lama1, Lamb1 , or Lamc1 gene (encoding laminin α1, β1, and γ1, respectively) die at embryonic day 7 or earlier. 1 , 2 Embryos lacking the collagen α1α2α1(IV) network fare slightly better, but survive only until embryonic day 11.5. 3 During kidney development, the glomerular basement membrane (GBM), a component of the plasma filtration barrier, undergoes laminin and collagen IV isoform substitutions, which appear necessary for acquisition of glomerular permselective barrier properties. Specifically, GBMs of the earliest glomeruli contain laminin α1β1γ1 and collagen α1α2α1(IV), whereas those of fully mature glomeruli contain laminin α5β2γ1 and collagen α3α4α5(IV). 4 Additionally, laminin and collagen IV isoform substitutions occur on different developmental timetables, with laminin transitioning occurring early in glomerular development, and type IV collagen exchange occurring later. 5 Mice with deletions of Lama5 die before birth with neural tube closure defects, placental vascular abnormalities, 6 and avascular glomeruli. 7 Mice with a hypomorphic mutation of Lama5 survive birth but have greatly reduced protein expression and die at 3-4 weeks of age with proteinuria and numerous kidney cysts. 8 Additionally, there is overexpression of laminin α5 in humans with Alport disease, possibly to compensate for an absence of collagen α3α4α5(IV) in GBMs of these patients, and in mouse and dog models of this disorder. 9 Mice with deletions of Lamb2 are born normal, but podocyte foot processes efface and mice die of renal failure a few weeks after birth. 10 Mutations to the LAMB2 gene cause Pierson syndrome in humans, which results in congenital nephrosis, ocular abnormalities and neonatal respiratory distress. 11 To explore GBM laminin biology further, we engineered transgenic mice that overexpress the human laminin α5 polypeptide chain. 12 Immunoprecipitation studies show that the human laminin α5 chain heterotrimerizes with cognate mouse laminin β and γ chains. 12 The hu LAMA5 transgenic mice deposit apparently large amounts of laminin heterotrimers containing human laminin α5 in the same basement membranes that contain the native, mouse laminin α5 chain. Importantly, human laminin α5 is also expressed at the correct developmental stages in glomeruli. Although these transgenic animals appear normal, there is suppression of transcription of the native mouse Lama5 gene and decreased deposition of mouse laminin α5 protein in GBMs. 12 Here we report that wildtype females mated with hu LAMA5 males developed a maternal humoral response against the paternally-derived human laminin α5 protein in offspring. Our findings show that maternal anti-human laminin α5 IgG was transferred to transgenic progeny before and after birth, bound to their GBMs, activated complement, and caused perinatal anti-GBM glomerulonephritis and proteinuria.