Differential peritoneal resident macrophage depletion and omental milky spot disruption after high versus low clodronate-loaded liposome dose treatment.

Margarita Ferriz, Natalia Álvarez‐Ladrón, Alejandra Gutiérrez‐González, Marta H. Fernández‐Sesma, Ana Starešinčić : Acquisition, analysis and interpretation of data. Carlos Ardavín: Conceptualization; funding acquisition; writing – original draft; supervision.

C57BL/6J and C57BL/6N mice were purchased from Charles River (L'Arbresle, France). Ccr2 −/− mice were kindly supplied by Dr F. Tacke (RW‐TH‐University Hospital Aachen, Germany) and housed at the Animal Facility of CNB/CSIC, on a 12 h/12 h light/dark cycle, with free access to food and water. Littermates of the same sex were randomly assigned to experimental groups. All the experiments were approved by the Animal Care and Use Committee of the Centro Nacional de Biotecnología‐CSIC, Madrid, under the protocols PROEX 193/19 and 086.8/24. CL‐liposome administration was performed by intraperitoneal injection of 100 μL of CL liposomes (Liposoma, Amsterdam, The Netherlands). The CL‐liposome dose was 0.5 mg/mouse and 0.025 mg/mouse for HD clodronate and LD clodronate, respectively. CL liposomes were labeled with the lipophilic carbocyanine DiO, exhibiting green fluorescence, using the Vybrant DiO cell‐labeling solution (ThermoFischer Scientific, Waltham, MA, USA), as described. 19 In brief, 1 mL CL liposomes were incubated with 10 μL Vybrant DiO dye at 1/100 for 20 min at 37°C, under gentle shaking, washed twice in PBS, and reconstituted to the original volume. Mice were injected intraperitoneally with 100 μL DiO‐labeled CL liposomes (0.5 mg/mouse) and analyzed at 90 min after injection. In the experiments shown in Supplementary figure  1d, e , CL liposome uptake was analyzed by flow cytometry in the peritoneal lavage, as described below. In the experiments shown in Supplementary figure  2d, e , CL liposome uptake was analyzed in the omentum by confocal microscopy, as described below. The peritoneal lavage was carried out by injecting into the peritoneal cavity 5 mL sterile PBS, which was subsequently harvested by performing an abdominal incision and collecting the resulting peritoneal lavage directly onto a 15 mL tube using a glass funnel, as described. 31 Cell suspensions were obtained after centrifugation of the peritoneal lavage for 5 min at 400 g. Fc receptors were blocked by incubation with an anti‐CD16/32 antibody (BD Biosciences, Franklin Lakes, NJ, USA) at 4°C for 15 min. The cells were subsequently stained for 20 min at 4°C with the fluorophore‐conjugated antibodies indicated below. Analysis of macrophage and monocytic populations present in the peritoneal fluid was performed after eight‐color immunofluorescent staining with FITC‐conjugated anti‐CD45 (BioLegend, San Diego, CA, USA), PE‐conjugated anti‐Ly6G (BD Biosciences), PE‐conjugated anti‐CD90.2 (BD Biosciences), PE‐conjugated anti‐CD19 (BD Biosciences), PECy7‐conjugated anti‐CD11b (eBioscience, San Diego, CA, USA), Brilliant Violet 421‐conjugated anti‐Tim4 (BD Biosciences), APC‐Cy7‐conjugated anti‐F4/80 (BioLegend), APC‐conjugated anti‐MHCII (eBioscience), PerCP‐Cy5.5‐conjugated anti‐Ly6C (BD Biosciences) and Pacific Blue‐conjugated anti‐B220 (BioLegend). Neutrophils, B1, B2 and T cells were analyzed using the gating strategy described in Figure  1a , after gating out CD90 − CD19 − Ly6G − CD11b + cells, as CD11b + MHCII − cells (neutrophils), CD11b − MHCII − cells (T cells), MHCII + B220 low cells (B1 cells) and MHCII + B220 high cells (B2 cells). Analysis of Gata6 expression by macrophage populations present in the peritoneal lavage cell was performed after seven‐color immunofluorescent staining with PE‐conjugated anti‐Ly6G, PE‐conjugated anti‐CD90.2, PE‐conjugated anti‐CD19, PECy7‐conjugated anti‐CD11b, Brilliant Violet 421‐conjugated anti‐Tim4, APC‐Cy7‐conjugated anti‐F4/80, FITC‐conjugated anti‐MHCII and PerCP‐Cy5.5‐conjugated anti‐Ly6C, followed by intracellular staining with PE‐conjugated anti‐Gata6 (Cell Signaling, Danvers, MA, USA), using the BD Cytofix/Cytoperm kit (BD Biosciences). Analysis of Ki67 expression by macrophage populations present in the peritoneal lavage cell was performed after seven‐color immunofluorescent staining with PE‐conjugated anti‐Ly6G, PE‐conjugated anti‐CD90.2, PE‐conjugated anti‐CD19, PECy7‐conjugated anti‐CD11b, Brilliant Violet 421‐conjugated anti‐Tim4, APC‐Cy7‐conjugated anti‐F4/80, FITC‐conjugated anti‐MHCII and PerCP‐Cy5.5‐conjugated anti‐Ly6C, followed by intracellular staining with Alexa Fluor 647‐conjugated anti‐Ki67(BD Biosciences), using the BD Cytofix/Cytoperm kit (BD Biosciences). Analysis of the uptake of DiO‐labeled CL liposomes by macrophage populations present in the peritoneal lavage cell was performed as indicated above, after seven‐color immunofluorescent staining with PerCP‐conjugated anti‐CD45 (BioLegend), PE‐conjugated anti‐Ly6G, PE‐conjugated anti‐CD90.2, PE‐conjugated anti‐CD19, PECy7‐conjugated anti‐CD11b, Brilliant Violet 421‐conjugated anti‐Tim4, APC‐Cy7‐conjugated anti‐F4/80, APC‐conjugated anti‐MHCII and Pacific Blue‐conjugated anti‐B220. Antibodies anti‐Ly6G, CD90 and CD19 were used together as PE conjugates with the purpose of gating out neutrophils, T cells and B cells, respectively. Data were acquired on an LSRII cytometer (BD Biosciences, San José, CA, USA) and analyzed using FlowJo X software (Tree Star, Ashland, OR, USA). Omentum samples were fixed in 4% paraformaldehyde for 5 min at RT. Blocking of unspecific binding of primary antibodies was performed by incubation in 2% BSA in PBS for 30 min at 4°C. Samples were then incubated for 2 h at 4°C with DAPI (Sigma–Aldrich, St. Louis, MO, USA) and the following fluorophore‐conjugated antibodies. Analysis of omental milky spot vascular network was performed after immunofluorescent staining with Alexa Fluor 488‐conjugated anti‐CD31 (BioLegend) and Alexa Fluor 647‐conjugated anti‐CD19 (BioLegend). Analysis of omental milky spot resident macrophages was performed after immunofluorescent staining with Alexa Fluor 488‐conjugated anti‐Tim4 (Invitrogen‐Thermo Fischer Scientific) and Alexa Fluor 647‐conjugated anti‐CD19. Analysis of CL‐liposome uptake in the omentum was performed after immunofluorescent staining with Alexa Fluor 594‐conjugated anti‐CD31 (Biolegend) and Alexa Fluor 647‐conjugated anti‐Tim4 (Biolegend). Whole omentum samples were mounted with Mowiol (Hoechst, Frankfurt, Germany) on μ‐Dish 35 mm dishes with No. 1.5 ibidi Polymer Coverslips (ibidi, Gräfelfing, Germany), as described. 31 Images were acquired on a multispectral Leica Stellaris 5 confocal microscope (Leica Microsystems, Wetzlar, Germany) and analyzed using the ImageJ software (NIH, MD, USA). Quantification of CD19 and Tim4 staining in omentum milky spots, shown in Supplementary figures  1h, i , was performed in eight representative omental fields per condition (five of which per condition are shown in Supplementary figure  1h ), selected from three mice per condition, using the ImageJ software, after conversion into binary images. Assessment of milky spot vascular network area, shown in Supplementary figures  2a, b and 3a, b , was performed by quantifying CD31 staining in eight (Supplementary figure  2b ) or five (Supplementary figure  3b ) representative omental fields per condition (five of which per condition are shown in Supplementary figures  2a and 3a ), selected from three mice per condition, using the ImageJ software after conversion into binary images. Milky spot vasculature fractal dimension, shown in Supplementary figures  2c and 3c , was calculated based on CD31 staining of the same representative omental fields per condition used to assess the milky spot vascular network area, after conversion into binary images, using Fractal Box Count in ImageJ, as described. 32 Data are presented as mean ± SD. Statistical analyses were performed using Microsoft Excel or GraphPad PRISM 9 software. Statistical significance was determined using a two‐tailed Student's t ‐test (when comparing two datasets) or one‐way ANOVA followed by post hoc Tukey's multiple comparisons test (when comparing three or more datasets). Statistical significance is indicated by asterisks (* P  < 0.05, ** P  < 0.01, *** P  < 0.001, **** P  < 0.0001).

Changes in peritoneal cavity macrophage (MØ) populations induced by intraperitoneal administration of a high (0.5 mg/mouse) CL‐liposome dose (HD clodronate), corresponding to the dose used by most research groups with the aim of depleting peritoneal macrophages, 7 , 8 have been analyzed by flow cytometry. Analyses were performed at days 1, 5, 10 and 20 postinjection (days 1, 5, 10 and 20). Based on their differential expression of the resident macrophage marker F4/80 and the peritoneal resident macrophage marker Tim4, macrophage and monocytic populations present in the peritoneal fluid, included, in the steady state, Tim4 + F4/80 + MHCII − MØs, Tim4 − F4/80 + MHCII − MØs and Tim4 − F4/80 lo MHCII + MØs (Figure  1a ). Both Tim4 + F4/80 + MHCII − MØs and Tim4 − F4/80 + MHCII − MØs expressed the transcription factor Gata6 (Supplementary figure  1a, b ), known to be crucial for resident peritoneal MØ‐specific gene expression, proliferation and survival. 25 Therefore, both Tim4 + F4/80 + MHCII − MØ and Tim4 − F4/80 + MHCII − MØ populations correspond to resident peritoneal MØs (resMØs), classically known, in the steady state, as large peritoneal macrophages. They will be hereafter termed Tim4 + resMØs and Tim4 − resMØs, respectively. Tim4 + resMØs are generated during embryonic life, essentially from yolk sac macrophages and fetal liver monocytes, whereas Tim4 − resMØs derive from adult bone marrow Ly6C hi monocytes. 25 In the steady state, Tim4 − F4/80 lo MHCII + MØs have been classically described as small peritoneal macrophages. They are derived from Ly6C hi monocytes 25 and accordingly, they did not express Gata6 (Figure  1a and Supplementary figure  1a, b ). They will be hereafter termed MHCII + moMØs (for MHCII + monocyte‐derived MØs). Effect of high‐dose clodronate versus low‐dose clodronate on peritoneal cavity immune cells. (a) FACS gating strategy allowing the characterization of macrophage and monocytic populations present in the peritoneal fluid in the steady state. (b, c) FACS analysis (b) and absolute number per mouse (c) of peritoneal macrophage and monocytic populations at the indicated times after HD clodronate administration. (d, e) FACS analysis (d) and absolute number per mouse (e) of peritoneal macrophage and monocytic populations at the indicated times after LD clodronate administration. In (a) , (b) and (d) , colors used to label the cell populations above the dot plots correspond to those used to establish the gates defining those cell populations. In (c) and (e) , data are expressed as mean ± SD of 6 mice/condition. Similar results were obtained in three independent experiments. * P  < 0.05, ** P  < 0.01, *** P  < 0.001, **** P  < 0.0001. HD clodronate treatment led, at day 1, to a complete depletion of both Tim4 + resMØs and Tim4 − resMØs (Figure  1b, c ). Depletion of Tim4 + resMØs and Tim4 − resMØs persisted for 3 weeks after HD clodronate administration (Figure  1c and Supplementary figure  1c ). Interestingly, partial recovery of both Tim4 + and Tim4 − resMØs was noticeable from week 8 (Supplementary figure  1c ). The number of MHCII + moMØs was markedly reduced, although a small number was detectable, most likely reflecting that MHCII + moMØs had a lower liposome uptake capacity than Tim4 + resMØs and Tim4 − resMØs, as assessed by measuring the uptake of DiO‐labeled CL liposomes by flow cytometry, which confirmed that Tim4 + resMØs and Tim4 − resMØs efficiently internalized CL liposomes (Supplementary figures  1d, e ). At day 5, a high number of Ly6C hi monocytes and neutrophils were recruited to the peritoneal cavity, which were no longer detectable from day 10 onwards, reflecting that HD clodronate caused a transient inflammatory response promoting their recruitment (Figure  1b, c ), in line with previous reports describing that CL liposome treatment, using different routes of administration, can promote local or systemic inflammation, 20 , 21 , 22 as pointed out in the Introduction section. A significant increase in MHCII + moMØs occurred at day 5, most likely reflecting that new MHCII + moMØs were formed from recruited Ly6C hi monocytes, through an intermediate population of Tim4 − Gata6 − F4/80 − Ly6C + MHCII + monocyte‐derived cells (hereafter Ly6C + MHCII + moCs), which displayed the same kinetics as Ly6C hi monocytes. The number of MHCII + moMØs progressively decreased from day 5 to day 20, probably reflecting that no monocyte recruitment occurred from day 5 onwards. Since, as stated above, both Tim4 − resMØs and MHCII + moMØs derive from Ly6C hi monocytes, the observation that MHCII + moMØs, but not Tim4 − resMØs, were formed from day 5 to day 20 after HD clodronate treatment most likely reflects their differing microenvironmental requirements. Indeed, the development and function of peritoneal resMØs, unlike MHCII + moMØs, are controlled by the transcription factor Gata6, which is activated by retinoic acid, claimed to be primarily produced by omental fibroblastic reticular cells. 25 Consequently, given that omental milky spot integrity is severely compromised following HD clodronate administration, as described below (Figure  3 ), omental disorganization could hinder the generation of new Tim4 − resMØs from Ly6C hi monocytes recruited after HD clodronate treatment. The low CL‐liposome dose (0.025 mg/mouse) used in this study was determined, based on a titration experiment (not shown), as the minimal dose ensuring an efficient peritoneal macrophage depletion, without compromising omental integrity. Injection of a low CL‐liposome dose (LD clodronate) resulted in a complete depletion of both Tim4 + resMØs and Tim4 − resMØs, at days 1 and 5 (Figure  1d, e ). Ly6C hi monocytes and neutrophils recruited to the peritoneal cavity were detectable at day 1 after LD clodronate, although their number was significantly lower (around sevenfold in the experiment shown in Figure  1d, e ) than after HD clodronate‐induced recruitment occurring at day 5, indicating that the inflammatory reaction induced by HD clodronate was significantly stronger than that induced by LD clodronate, in agreement with a previous study assessing the effect of different intraperitoneal CL‐liposome doses on peritoneal neutrophilia. 23 Ly6C hi monocytes and neutrophils were barely detectable from day 5 onward after LD clodronate, indicating that LD clodronate caused a weak and transient inflammatory response, resulting in a single wave of Ly6C hi monocyte and neutrophil recruitment. Recruitment of Ly6C hi monocytes was paralleled, at day 1, by an increased number of Ly6C + MHCII + moCs, but not of MHCII + moMØs. This most likely reflects that although new MHCII + moMØs were formed from Ly6C hi monocytes at day 1, their number did not increase due to the depletion of preexisting MHCII + moMØs by LD liposome administration. Nevertheless, the number of MHCII + moMØs increased at day 5, as a result of their generation from Ly6C hi monocytes recruited at day 1, and progressively decreased onwards (Figure  1d, e ). Interestingly, the Tim4 + resMØ population was gradually restored from day 10, so that at day 20, the number of Tim4 + resMØs was around 2/3 of the number found in the steady state (Figure  1e ). This was paralleled by the generation of a substantial number of new Tim4 − resMØs, so this population underwent a fivefold increase by day 20. Consequently, the total number of resMØs (i.e. the sum of Tim4 + and Tim4 − resMØs) was comparable to their number in the steady state. As expected, newly formed Tim4 + and Tim4 − resMØs, but not MHCII + moMØs, expressed Gata6 (Supplementary figures  1f, g ). The analysis of Ccr2 −/− mice, deficient in the chemokine receptor CCR2, controlling Ly6C hi monocyte egress from the bone marrow and recruitment to inflamed organs, 26 revealed that in these mice Tim4 + resMØs were barely detectable at day 10 after LD clodronate and at day 20, their number was lower than in wild‐type mice (Figure  2a, b ). As expected, Ly6C hi monocyte recruitment to the peritoneal cavity induced by LD clodronate was significantly impaired in Ccr2 −/− mice at day 1, when Ly6C hi monocyte recruitment occurred (Figure  2c ), and also at days 10 and 20, although only low numbers of Ly6C hi monocytes were detectable at these time points, even in wild‐type mice (Figure  2b ). Impact of CCR2 deficiency on the recovery of peritoneal macrophage populations after low‐dose clodronate treatment. (a, b) FACS analysis (a) and absolute number per mouse (b) of peritoneal macrophage and monocytic populations, at the indicated times after LD clodronate administration, in wild type (WT) and Ccr2 −/− mice. (c) Absolute number per mouse of peritoneal macrophage and monocytic populations at d0 and d1 after LD clodronate administration, in WT and Ccr2 −/− mice. (d, e) FACS analysis (d) and quantification (e) of Ki67 expression by Tim4 + resMØs, Tim4 − resMØs and MHCII + moMØs, at the indicated times after LD clodronate administration, in WT and Ccr2 −/− mice. In (a) , colors used to label the cell populations above the dot plots correspond to those used to outline the gates defining those cell populations. In (b) , (c) and (e) , data are expressed as mean ± SD of 6 mice/condition. In (d) , the percentage of Ki67 + cells is indicated; light gray profiles correspond to isotype control antibody staining. Similar results were obtained in two independent experiments. * P  < 0.05, ** P  < 0.01, *** P  < 0.001, **** P  < 0.0001, ns nonsignificant. The analysis of the proliferation marker Ki67 revealed that Tim4 + resMØs displayed a high cell division rate at day 10 that decreased at day 20 (Figure  2d, e ), both in wild‐type and Ccr2 −/− mice, supporting that their recovery occurred by the expansion through cell division of remaining Tim4 + resMØs or an undefined Tim4 + resMØ precursor population. Given that Tim4 + resMØs are of embryonic origin and not derived from Ly6C hi monocytes, their reduced number in Ccr2 −/− mice could reflect that their recovery was delayed compared to wild‐type mice. Accordingly, confocal microscopy analysis of omental milky spots at day 10 after LD clodronate administration revealed that milky spot B cells and omental Tim4 + resMØs were significantly reduced in Ccr2 −/− mice, leading to more pronounced disruption of milky spot organization (Supplementary figure  1h, i ). Therefore, since, as stated above, formation of peritoneal resMØs could be affected by omental milky spot dysfunction, the more pronounced milky spot alterations displayed by Ccr2 −/− mice at day 10 after LD clodronate administration could lead to a delayed recovery of Tim4 + resMØs. Besides, since Tim4 − resMØs are of Ly6C hi monocytic origin, their recovery after LD clodronate in Ccr2 −/− mice, in which Ly6C hi monocyte recruitment was markedly reduced (Figure  2b, c ), most likely resulted from the expansion by cell division of a limited number of Tim4 − resMØs generated from the reduced number of recruited Ly6C hi monocytes. Indeed, in Ccr2 −/− mice Tim4 − resMØs exhibited a high proliferation rate at day 10 that decreased at day 20 (Figure  2d, e ). In addition, the fact that the proliferation rate of Tim4 − resMØs was comparable in wild‐type and Ccr2 −/− mice supports that, also in wild‐type mice, recovery of Tim4 − resMØs primarily relied on cell proliferation (Figure  2d, e ). Of note, MHCII + moMØs that, as stated above, were generated in wild‐type mice from Ly6C hi monocytes recruited following LD clodronate, displayed a significantly higher proliferation rate in Ccr2 −/− mice at day 10 (Figure  2d, e ), indicating a compensation process to restore MHCII + moMØs, whereas Ly6C hi monocyte recruitment was significantly impaired. Recovery of Tim4 + and Tim4 − resMØ populations after LD clodronate‐mediated depletion by cell division‐driven expansion is somehow unexpected, because, while self‐renewal contributes to a great extent to resMØ maintenance in homeostasis, resMØ replacement during disease‐induced resMØ depletion has been reported to rely essentially on differentiation of resMØs from recruited monocytes in nonhomeostatic inflammatory conditions that likely prevent self‐renewal‐mediated resMØ replacement. 27 Therefore, it can be hypothesized that the resolution of the weak inflammatory process associated to resMØ depletion by LD clodronate treatment led to a particular, homeostatic‐like condition allowing resMØ recovery by self‐renewal. Interestingly, CSF1‐dependent resMØ proliferation contributed to the late stages of re‐establishment of peritoneal resMØs after zymosan‐induced acute inflammation, 28 and replacement of lung resMØs after influenza virus‐ or diphtheria toxin‐mediated depletion. 29 Therefore, HD clodronate treatment led to a long‐term depletion of both Tim4 + resMØs and Tim4 − resMØs and a transient reduction of MHCII + moMØs, paralleled by a strong inflammatory response leading to the recruitment to the peritoneal cavity of high numbers of Ly6C hi monocytes and neutrophils. In contrast, LD clodronate led to a transient depletion of both Tim4 + and Tim4 − resMØs, concurrent with a weak inflammatory response that allowed the progressive recovery of Tim4 + and Tim4 − resMØ populations, essentially through cell proliferation, while the MHCII + moMØ population was maintained. The omentum fulfills essential functions related to peritoneal immunity, resMØ homeostasis and resMØ replacement during disease‐driven inflammatory conditions. Omental immune cells are organized in functional units, harboring a specialized vascular system, called milky spots. 24 Importantly, B1 cell homing to the omentum and milky spot development are controlled by the chemokine CXCL13, 30 whose production within the peritoneal cavity largely relies on peritoneal resMØs. 25 Therefore, taking into account the functional interdependence of peritoneal resMØs and the omentum, we next sought to assess the differential effects of HD clodronate versus LD clodronate treatment on omental milky spot integrity. Imaging of the omentum was performed using a whole mount immunofluorescence combined with a confocal microscopy protocol optimized in our laboratory, 31 after staining with antibodies against the endothelial cell marker CD31, the B‐cell marker CD19, and the peritoneal resMØ marker Tim4. In nontreated mice, milky spots, which, as previously noted, can be readily visualized by the B‐cell expression of CD19, harbored a developed vascular network as revealed by the expression of CD31 by endothelial cells (Figure  3a ). Tim4 staining revealed the presence of a large number of resMØs associated with and surrounding milky spot B‐cell areas (Figure  3b ). HD clodronate treatment led from day 5 to day 20 to a progressive disorganization of the milky spot vascular network which, at day 20, appeared severely damaged (Figure  3c ). At day 5, omental Tim4 + resMØs associated with milky spots were not detectable either within or surrounding B‐cell areas; absence of omental Tim4 + resMØs persisted at days 10 and 20 (Figure  3d ). In contrast, LD clodronate treatment did not significantly affect the milky spot vascular network, which is maintained from day 5 to day 20 (Figure  3e ). Quantification of the omental milky spot vascular network area, performed on the basis of the expression of the endothelial marker CD31, in multiple confocal microscopy fields showing the milky spot vasculature, in control mice and at day 20 after LD or HD clodronate treatment, confirmed an important decrease in milky spot vascular area in HD clodronate‐treated mice, that correlated with a pronounced reduction in the milky spot B‐cell compartment. In contrast, no significant alterations in the milky spot vascular network area were found after LD clodronate treatment (Supplementary figure  2a, b ). Consistent with this result, the analysis of milky spot vasculature fractal dimension, a metric used to quantify the structural and branching complexity of a vascular network, 32 , 33 revealed that the fractal dimension was significantly lower after HD clodronate treatment than in nontreated mice, whereas the fractal dimension of LD clodronate‐treated mice was comparable to that of nontreated mice (Supplementary figure  2c ). In accordance with our flow cytometry analysis of resMØ depletion and recovery in the peritoneal cavity after LD clodronate (Figure  1d, e ), omental Tim4 + resMØs associated with milky spots in LD clodronate‐treated mice were barely detectable at day 5 but were progressively restored from day 10 (Figure  3f ). Differential effect of high‐dose versus low‐dose clodronate on omentum milky spot organization. (a) Imaging of the omentum by whole mount immunofluorescence and confocal microscopy, in the steady state, after immunofluorescent staining with antibodies against the endothelial marker CD31 and the B‐cell marker CD19. (b) Whole‐mount immunofluorescence and confocal microscopy image of an omental milky spot, in the steady state, after immunofluorescent staining with antibodies against the resMØ marker Tim4 and the B‐cell marker CD19. (c) Imaging of the omentum by whole mount immunofluorescence and confocal microscopy, at the indicated times post‐HD clodronate administration, after immunofluorescent staining with antibodies against the endothelial marker CD31 and the B‐cell marker CD19. (d) Whole mount immunofluorescence and confocal microscopy images of omental milky spots, at the indicated times post‐HD clodronate administration, after immunofluorescent staining with antibodies against the resMØ marker Tim4 and the B‐cell marker CD19. (e) Imaging of the omentum by whole mount immunofluorescence and confocal microscopy, at the indicated times post‐LD clodronate administration, after immunofluorescent staining with antibodies against the endothelial marker CD31 and the B‐cell marker CD19. (f) Whole mount immunofluorescence and confocal microscopy images of omental milky spots, at the indicated times post‐LD clodronate administration, after immunofluorescent staining with antibodies against the resMØ marker Tim4 and the B‐cell marker CD19. Arrows point to enlargements of the areas defined by a white‐lined square. Similar results were obtained in two independent experiments. In (b) , (d) and (f) , white dashed lines indicate the milky spot limit. In order to address whether the severe alterations in omental milky spot vasculature caused by HD clodronate administration could be due, at least in part, to the uptake of CL liposomes by endothelial cells that might critically compromise their viability, DiO‐labeled CL liposomes were injected intraperitoneally and subsequently detected within omental milky spots by confocal microscopy (Supplementary figure  2d ). Sequential high magnification optical sections, obtained at different levels of the Z‐axis, revealed that liposomes can be readily observed in the cytoplasm of omental Tim4 + resMØs but were not detectable inside milky spot endothelial cells (Supplementary figure  2e ), supporting that vascular damage caused by HD clodronate was not the consequence of endothelial cell death resulting from CL liposome uptake. The C57BL/6J mouse line used in this study is the most widely employed C57BL/6 substrain in biomedical research over the last few decades. However, since C57BL/6J mice were found to carry the nicotinamide nucleotide transhydrogenase (Nnt) null mutation, 34 leading to glucose intolerance, reduced insulin secretion and impaired mitochondrial function, 35 the C57BL/6N substrain is becoming a standard background strain and has been chosen by major international initiatives, such as the International Mouse Phenotyping Consortium, because of its greater ES cell efficiency for homologous recombination. In order to explore whether the metabolic alterations associated with the Nnt null mutation displayed by C57BL/6J mice could contribute to the omental milky spot disorganization, caused by HD clodronate administration, the effect of HD clodronate was compared in C57BL/6J and C57BL/6N mice. Our data, shown in Supplementary figure  3 , revealed that the milky spot vascular network area and the fractal dimension at day 20 after HD clodronate treatment and, consequently, the milky spot vascular network disorganization, were comparable in both substrains, and therefore not linked to the metabolic alterations associated with C57BL/6J mice. In conclusion, HD clodronate treatment led to a long‐lasting resMØ depletion but induced peritoneal inflammation and caused a severe and persistent omental milky spot disorganization that prevented peritoneal resMØ replacement and might severely impair the homeostatic and immune defense functions of the omentum. Therefore, resMØ depletion, inflammation and omental dysfunction all contributed to the final outcome of experiments based on HD clodronate administration and, consequently, no straightforward conclusions might be drawn on the function of peritoneal resMØs from experiments relying on their depletion by HD clodronate, that is, using the CL‐liposome dose used by most research groups. 7 , 8 In contrast, LD clodronate led to a transient depletion of resMØs that were progressively restored from day 10, but did not promote significant milky spot alterations or peritoneal inflammation, therefore ensuring an efficient peritoneal macrophage depletion in the absence of significant inflammation and omental structural and vascular alterations, yet during a limited time window.

Clodronate‐loaded liposomes (CL liposomes) were originally designed in Dr van Rooijen's lab with the aim of depleting mouse splenic macrophages to explore their function. 1 After CL liposomes are engulfed by phagocytic cells, clodronate is released to the cytosol and metabolized by Class II aminoacyl‐tRNA synthetases to a nonhydrolyzable ATP analog, adenosine 5′‐(β,γ‐dichloromethylene) triphosphate, 2 that binds and inhibits the ATP/ADP translocase, leading to the loss of mitochondrial inner membrane integrity and cell death by apoptosis. 3 CL liposomes have been widely employed, using different routes of administration, as an alternative to other strategies relying on blocking CSF1R activity with anti‐CSF1R antibodies 4 , 5 or CSF1R kinase inhibitors, 6 to deplete diverse macrophage subpopulations in different organs, that is, liver Kupffer cells, peritoneal macrophages, brain microglia or lymph node subcapsular macrophages, with the aim of assessing their specific role in homeostatic, repair, regulatory, defense, autoimmune and inflammatory processes. 7 , 8 Besides, CL liposome administration has allowed us to investigate the tumor‐promoting function of tumor‐associated macrophages in different mouse cancer models, including teratocarcinoma, 9 adenocarcinoma 10 and peritoneal ovarian cancer metastasis. 11 Experimental evidence accumulated over the last decades has revealed that administration of CL liposomes had a long‐term anti‐inflammatory effect in several mouse models of diseases associated with macrophage‐induced inflammation, such as rheumatoid arthritis, 12 , 13 , 14 liver fibrosis, 15 endometriosis 16 or chronic obstructive pulmonary disease (COPD). 17 In this regard, the CL‐liposome anti‐inflammatory effect has been classically linked to the depletion of macrophages with an inflammatory function. 18 However, this concept has been recently challenged by a recent report demonstrating, using a mouse model of rheumatoid arthritis, that the dampening of inflammation observed after intravenous CL‐liposome administration was not due to the depletion of synovial macrophages, but to the functional inactivation of inflammatory neutrophils induced by the uptake of CL liposomes, which led to a reduced release of ROS, cytokines and neutrophil extracellular traps. 19 In contrast, CL liposome treatment was reported to cause inflammation in different experimental models. In a study aiming at exploring how inhaled particulates induced airway allergy, intratracheal instillation of CL liposomes triggered IL‐1α release by dying alveolar macrophages, contributing to lung inflammation, paralleled by the formation of inducible bronchus‐associated lymphoid tissue (iBALT). 20 Intraperitoneal CL liposome administration‐mediated depletion of adipose tissue macrophages was performed to assess its potential as an effective therapeutic strategy to reduce obesity‐driven inflammation and metabolic dysfunction, causing adipose tissue neutrophilia and increased IL‐6 plasma levels. 21 Moreover, in a recent report exploring the relevance of histone release in sepsis, intravenous CL‐liposome injection induced the release from dying macrophages of systemic chromatin, known to promote inflammation and immune dysfunction during systemic candidiasis. 22 Taken together, these studies reveal that the collateral effects of CL‐liposome treatment have to be explored in‐depth when designing and interpreting experiments involving macrophage depletion by administration of CL liposomes. Interestingly, in a recent study, exploring the potential of peritoneal macrophages to colonize the omentum, the effect of different CL‐liposome doses after intraperitoneal administration was assessed. 23 Whereas depletion of peritoneal macrophages induced by administration of a standard 0.5 mg/mouse CL‐liposome dose, that is, the dose recommended by the manufacturer and used in numerous studies, 8 caused a marked peritoneal neutrophilia, lowering the CL‐liposome dose to 0.03 mg/mouse ensured an efficient macrophage depletion that was not paralleled by a detectable neutrophil recruitment, 23 supporting that a low CL‐liposome dose might reduce the inflammation induced by a standard CL‐liposome dose. Based on these observations, with the aim of exploring the collateral effects of peritoneal macrophage depletion by CL liposome administration, the effect of an intraperitoneal standard high (0.5 mg/mouse) versus low (0.025 mg/mouse) CL‐liposome dose on the kinetics of peritoneal macrophage subpopulations and structural integrity of the omentum, an organ fulfilling a crucial function in peritoneal immunity, 24 has been investigated.

The authors declare no commercial or financial conflicts of interest.

Supplementary figure 1. Supplementary figure 2 . Supplementary figure 3 .