8–12-week-old male wildtype BALB/c, C57BL/6, and B6.SJL-Ptprca Pepcb/BoyJ (CD45.1⁺ for donor/recipient discrimination) mice were purchased from Charles River Laboratories (Charles River, Cologne, Germany) and kept under standard laboratory animal conditions, receiving human care in compliance with the “Principles of Laboratory Animal Care” prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH Publication No. 86–23, revised 1985). All animal experiments were approved by the Landesamt für Gesundheit und Soziales Berlin, Germany (G 0089/16).

Anti-mouse ATG was prepared by purifying the IgG fraction from serum of rabbits immunized with pooled thymocytes prepared from NOD, C3H/He, DBA/2, and C57BL/6 mice (Sanofi). ATG was administered to C57BL/6 mice intra-peritoneally at 25 mg/kg bodyweight at day −2 and −1. Control animals received purified IgG from unimmunized rabbits. Animals were sacrificed on day 0 for direct cellular analysis or kidney procurement for transplantation.

Allogeneic renal transplantations were performed as previously described [5, 18, 19]. Briefly, the left donor kidney was flushed with saline containing heparin (100 U/mL, Panpharma, Trittau, Germany) and procured. End-to-side anastomoses between the renal donor vessels and the recipient’s abdominal aorta and inferior vena cava were performed following a knotless technique. For urinary tract reconstruction, the ureter was directly anastomosed into the bladder. The duration of cold and warm ischemia of allografts was maintained at 30 min, respectively. Animals were sacrificed on day 7 without receiving immunosuppression.

Serum samples were stored at −20°C until creatinine and urea were measured using the CREP2 Creatinine Plus version 2 and Urea/BUN assays, respectively, on a Roche/Hitachi Cobas C 701/702 system (Roche, Basel, Switzerland).

For isolating renal mononuclear cells (MNCs), kidney tissue was mechanically dissociated and digested in RPMI medium (Corning, Manassas, VA, United States) supplemented with collagenases II and IV (Gibco/Invitrogen, Worthington) and DNase I (Roche Diagnostics) for 45 min at 37°C. Afterwards, leukocytes were enriched using CD45 Microbeads over MACS LS columns (both Miltenyi Biotec, Bergisch Gladbach, Germany). Mechanically dissociated lung tissue was digested with collagenase II and DNAse I; leukocytes were retrived after red blood cell lysis using ACK buffer. MNCs from spleen, lymph nodes and peripheral blood were isolated by density gradient centrifugation.

Typically, 1 × 10⁶ cells were surface stained with the respective antibodies listed in Supplementary Figures S1, S3. For intracellular staining, cells were fixed, permeabilized (FoxP3/transcription factor staining buffer set; Thermo Fisher, Darmstadt, Germany) and stained with the respective anti-transcription factor antibodies (Supplementary Figure S3). Data was acquired on a FACS Fortessa X20 (BD Biosciences, Heidelberg, Germany) and analyzed using FlowJo software 10 (BD Biosciences). A gating strategy for identification of lymphoid and myeloid cells is depicted in Supplementary Figure S2. As the predominant population, CD11b⁺F4/80⁺ macrophages were quantified in kidney, whereas Ly6C⁺MHC-II^- monocytes were analyzed in all other organs. Innate lymphoid cells were identified as illustrated in Supplementary Figure S4. For generation of t-Distributed Stochastic Neighbor Embedding (t-SNE) plots, data from samples of interest were concatenated, followed by t-SNE analysis in FlowJo. Graphical illustrations were designed with Biorender.

Statistical analysis was performed using GraphPad Prism 8.4.3 (GraphPad, Boston, MA, United States). Distribution of values was assessed by Kolmogorov-Smirnov normality test. Depending on distribution, comparisons were conducted using either T- or Mann–Whitney U test. Statistical significance was considered for p values ≤0.05.

First, we aimed to address how application of ATG to a potential solid organ donor impacts the lymphocyte composition in various organs. Naïve C57/BL6 mice were therefore treated with ATG before organ harvesting as summarized in Figure 1A. Control Ig treated mice of the same age and sex served as controls. For assessing all major lymphoid and myeloid cell subsets, mononuclear cells isolated from kidney, lung, spleen, lymph nodes and peripheral blood were surface stained using a FACS panel covering major cell lineage markers as depicted in Supplementary Figure S1, followed by acquisition on an BD Fortessa X20 flow cytometer.

To identify major differences induced by treatment, we first followed an unbiased analysis approach employing the t-SNE (t-Distributed Stochastic Neighbor Embedding) algorithm embedded in FlowJo. For that, FACS data derived from organs of controls and ATG treated animals were tagged according to sample type, followed by concatenation into one single data file. Thereafter, t-SNE analysis was conducted, allowing unbiased assessment of dominant differences. Overlay of t-SNE plots for each organ type consistently identified clusters present in controls, but absent in treated animals (Figure 1B, left column). For identification of cell types driving this different clustering, expression levels of all markers used for t-SNE were overlaid, pointing to a reduction of CD4⁺ and CD8⁺ T cells in all organs, as exemplarily depicted in Figure 2B (right columns).

We further analyzed all datasets after manual gating (depicted in Supplementary Figure S2A), enabling quantification and statistical examination of differences as illustrated in Figure 2A, where frequencies of the indicated cell types are presented as percentage of the total CD45⁺ leukocyte population across all organs. Overall, lymphocytes were strongly diminished by ATG treatment, whereas this effect was not observed for myeloid cell subsets. In detail, CD4⁺ and CD8⁺ T cells were significantly reduced in lymphoid organs (lymph nodes and spleen) and almost completely depleted in blood, lung and kidney. With respect to the latter, raw data presented in Supplementary Figure S2B allows a more direct estimation of renal T cell numbers remaining after ATG administration.

NK cells showed a similar pattern with the exception of lymph nodes where NK cells were already rare in controls. Interestingly, B cells, identified according to B220 expression, showed a relative increase within the CD45⁺ leukocyte population in both lymphoid organs. Within the myeloid cell compartment, we observed only few changes following ATG injection. Of note, most prominent alterations were confined to the kidney and encompassed a moderate relative increase in frequencies of CD11c⁺MHC-II⁺ dendritic cells and CD11b⁺F4/80⁺ macrophages. In all other organs, Ly6C⁺MHC-II^- monocytes instead of macrophages were quantified as the dominating population that was only modestly, but significantly increased in blood of ATG treated animals.

Provided that lymphocytes represented the cell population most strongly affected by ATG treatment, we further aimed at analyzing innate lymphoid cells (ILCs), a multifaceted group of lymphocytes that does not express characteristic T-, B- or dendritic cell associated lineage markers [20]. Their transcriptional hallmarks broadly mirror the discrimination among T helper (Th) cell type 1, 2 and 17 subsets; accordingly, group 1 ILC, that include conventional NK cells, express T-bet, group 2 ILC are GATA-3⁺, whereas group 3 members are RORγt⁺. ILC type 2 cells have already been demonstrated to protect murine kidneys from glomerulosclerosis [21] and renal ischemia reperfusion injury [22]. Given the potential importance of ILCs in transplantation, we therefore particularly focused on the kidney for ILC analysis. The flow cytometric marker panel for ILC subset identification is depicted in Supplementary Figure S3 with the gating strategy being summarized in Supplementary Figure S4. Of note, ILC frequencies were calculated within CD45⁺lineage⁻ cells; therefore, they could not directly be compared to the leukocyte percentages depicted in Figure 2A. Following ATG treatment, we detected a significant reduction in ILC type 1 and 2 subsets compared to controls, whereas portions of the ILC3 subpopulation were not affected (Figure 2B).

As a summary, all major lymphocyte subpopulations with the exception of B cells and ILC3 were significantly diminished as a consequence of ATG treatment in both non-lymphoid (kidney, lung) and lymphoid (lymph node, spleen) organs as well as in peripheral blood.

The persistence of donor-derived, intra-graft leukocytes after transplantation has been already shown to impact transplant survival and -function in clinical studies as well as in experimental models [8, 23]. We therefore addressed whether depletion of donor cells by ATG affects allogenic kidney transplantation outcome both on the cellular and functional level. As depicted in Figure 3A, kidneys from control or treated C57BL/6 mice were transplanted into BALB/c mice, representing a fully MHC-mismatched donor:recipient combination. Cellular and functional readout was performed on day 7, representing a typical timepoint in this acute rejection model, as recently published by us [5].

In an identical approach as for the analysis of ATG-induced alterations in naïve mice, we employed t-SNE for unbiased identification of pre-treatment effects on kidney transplantation outcome. For that, MNCs from the kidney graft were surface stained as depicted in Supplementary Figure S1 and data was acquired by FACS. Overlay of t-SNE plots indicated only few dominant differences in cellular composition (Figure 3B, left column). For identification of cell types driving slightly different clustering, expression levels of candidate markers were overlaid, pointing to a moderate reduction of graft CD4⁺ T and NK cells and an increase in neutrophils (Figure 3B, right columns).

We further analyzed all datasets after manual gating (Supplementary Figure S2), enabling quantification and statistical examination of differences as illustrated in Figure 4A. Of note, we did not detect additional intra-graft differences than those identified by t-SNE that did only reach statistical significance for neutrophils and a trend for CD4⁺ T cells. Analysis of MNCs from other organs showed a slight, but significant decrease in lung B cells in case of ATG donor-pre-treatment, whereas neutrophils were increased in spleen and showed a trend in lung.

Based on the finding that donor ATG-pretreatment resulted in moderately diminished frequencies of intra-graft CD4⁺ T cells, we conducted subanalyses, assessing quantities of effector-memory (T_em) and central-memory- (T_cm) type T cells identified according to CD44 and CD62L expression (Figure 4B). Interestingly, we noted a trend towards reduced portions of intra-graft CD4⁺ T_cm after ATG pre-treatment, whereas T_em remained unchanged (Figure 4C). The same applied to the CD8⁺ T cell compartment where the drop in T_cm frequencies reached significance (Figure 4D).

Importantly, as already demonstrated earlier using mice expressing the congenic markers CD45.1 and CD45.2 for donor/recipient discrimination [5], CD45⁺ donor leukocytes were gradually depleted in untreated animals until day 7 in our fully mismatched transplantation model, highlighting that more than 99% of analyzed cells at this timepoint were recipient-derived (Supplementary Figure S5A). Subset analysis revealed that a considerable number of renal TCRβ⁺ donor T cells is still detectable at day 3 post transplantation (Supplementary Figure S5B).

For deciphering the consequences of donor ATG-pretreatment on renal function, creatinine and urea levels were determined in serum on day 7 after transplantation. Importantly, we did not detect significant differences between graft recipients of control or ATG pre-treated donor organs (Figure 5).

In summary, ATG pre-treatment of the organ donor did only result in a minor cellular re-composition of the allograft in our transplantation model. In line with the aforementioned, we did not note relevant changes in graft function in association with this type of therapeutical intervention.