In this study, BALB/C mice were used as donors and C57BL/6J as recipients. All mice were male, aged 8–12 weeks at the time of transplant, and purchased from ENVIGO RMS SARL (Harlan Laboratories, Gannat, France). The mice were housed at the animal facility (zootechnie de Rangueil, CREFRE, INSERM US006) under SPF conditions, at a regulated temperature (20 °C), with free access to food and water. Five donor mice and five recipient mice were used in each group for transplant surgeries. The surgical procedures are illustrated in Figure 1A.

Euthanasia of the recipient mice was achieved using a double method: overdose of anesthetics combined with exsanguination by intracardiac perfusion of heparinated PBS. The graft was collected, fixed in formol for 12 h, then embedded in paraffin. Cross-sections with a thickness of 4 μm were performed using a microtome (RM2145, Leica) and put on charged slides (Superfrost, ThermoFisher Scientific). For each vessel, sections were taken every 100 μm starting from the cephalic anastomosis until the caudal anastomosis. The obtained 100 μm-spaced sections were stained with Masson Trichrome and the slides were digitized using a slide scanner (3DHISTECH) set on brightfield 40× magnification. For each obtained image, areas of the lumen (A_lum), and areas delimited by the internal and external elastic laminas (A_IEL and A_EEL) were measured using the SlideViewer software (3DHISTECH). The area of the neointima was calculated as follows: A_NI = A_IEL − A_lum; and the area of the media: A_M = A_EEL − A_IEL. The extent of intimal hyperplasia was expressed as a A_NI/A_M ratio in order to normalize the neointimal area.

Paraffin sections of aortic grafts (see Section 1.2) were deparaffinated by heating the slides at 60 °C for 10 min, soaking them in successive baths of xylene and ethanol, and then rehydrating them with water. The antigens were unmasked in citrate pH6 buffer using a pressure cooker. After cooling down, the sections were circled with a hydrophobic pen (DAKO). They were rehydrated in PBS, then PBS + Tween 0.1%. The unspecific binding was blocked by incubating the slides in PBS + BSA 3%. Then the slides were incubated at 4 °C overnight in a humid atmosphere with a solution of the primary antibody diluted in PBS + BSA 3% [primary antibodies: smooth muscle actin (SMA) rabbit polyclonal antibody, ThermoFisher, #12623297, 1:200; anti-mouse CD45 antibody, clone 30-F11, Biolegend #103101, 1:100; rabbit anti-CD3 antibody, clone SP7, Zytomed Systems #RBK024, 1:100; CD19 monoclonal antibody, clone LC1, ThermoFisher, #14-0190-82, 1:100; goat anti-mouse ICAM-1 antibody, clone M-19, Santa-Cruz Biotechnology, #sc-1511, 1:100; rabbit anti-mouse von Willebrand Factor (vWF) antibody, DAKO, #A0082, 1:100]. The slides were washed twice in PBS + Tween 0.1% before incubation with secondary antibodies (AF 488 Goat anti rabbit, Life Technologies, #A11034, 1:250; AF 555 Goat anti-mouse, Invitrogen, #A21452, 1:250; AF 555 Goat anti-rat, Invitrogen, #A214334, 1:250). After washing twice with PBS + Tween 0.1%, the slides were rinsed in distilled water and mounted with the ProLong Gold Antifade Mountant with DNA stain DAPI (ThermoFisher, #P36941). The immunofluorescence images were acquired using a confocal microscope (LSM900, Zeiss, TRI-Genotoul Imaging Facility, I2MC, Toulouse) and digitized with the ZEN software (Zeiss). The analysis was performed using ZEN and Fiji software. Smooth muscle cells, hematopoietic cells, T cells, and B cells, respectively, positive for αSMA, CD45, CD3, and CD19, were counted with the cell counter tool of Fiji. Fluorescence intensity measurements of ICAM-1 and vWF immunostaining in the luminal area were performed as follows. The process was implemented as a macro for the software the ImageJ/Fiji with the use of the MorpholibJ plugin (16). Using this plugin, we first applied a mask on the luminal layer (20 pixels width). Then, the mean intensity of the fluorescence signal (corresponding to ICAM-1 or vWF) was measured inside the mask, resulting in a quantification of the level of ICAM-1 or vWF protein content in the luminal (endothelial) layer.

Paraffin sections of aortic grafts (see Section 1.2) were deparaffinated by heating the slides at 60 °C for 10 min, soaking them in successive baths of xylene and ethanol, and then rehydrating them with water. The antigens were unmasked in citrate pH6 buffer using a pressure cooker. After cooling down, the sections were circled with a hydrophobic pen (DAKO). They were rehydrated in PBS, then PBS + Tween 0.1%. Endogenous peroxidases were blocked with H₂O₂ (10 min) and the unspecific binding was blocked by incubating the slides in PBS + BSA 3%. The slides were then incubated at 4 °C overnight in a humid atmosphere with the primary antibody diluted in PBS + BSA 3% (primary antibody: monoclonal rat anti-mouse IP-10 antibody, R&D Systems, #MAB466, clone 134013, 1:50). The primary antibody was detected using an ImmPRESS HRP goat anti-rat IgG polymer detection kit (Vector Laboratories #MP-7404-50) followed by ImmPACT DAB Substrate Kit (Vector Laboratories #SK-4105). The images were acquired using a microscope (Leica DM2500) and the IP-10 signal was quantified using H&E DAB color deconvolution with Fiji software. Results are expressed as the percentage of positive pixels.

Blood samples from all mice were collected on the day of euthanasia in blood collection tubes (Sarstedt, #41.1395.005). The blood cell-free fraction was collected after centrifugation. For the analysis of donor-specific antibodies (DSA), plasma samples (diluted at 1:500, 1:1,000, and 1:10,000) were incubated for 30 min with splenocytes freshly isolated from BALB/C mice or from B6 mice (baseline negative control) and blocked with PBS + 10% fetal bovine serum in order to enable the binding of plasma antibodies directed against BALB/C epitopes. After washing, the antibody-bound splenocytes were incubated with an AF488-conjugated anti-mouse IgG antibody (Invitrogen #A1100) and a PE-Vio615 anti-mouse CD19 antibody (Miltenyi Biotech, #130-112-042) to exclude the non-specific detection of surface B-cell immunoglobulins. The data were acquired with a flow cytometer (LSRII BD Fortessa) and analyzed with the BD Diva software. The splenocyte-bound antibody levels were expressed as the mean fluorescence intensity (MFI) of AF488 on the population gated on live singlets after excluding CD19+ B cells. The MFI of B6-bound serum, considered to be an unspecific binding baseline, was subtracted to the MFI of BALB/C-bound serum to express the relative levels of DSA.

The animal protocol for aortic transplantations was approved by the CEA 122 and the French Ministry of High Education and Research (Ministère de l’Enseignement Supérieur et de la Recherche) under number #28864.

The statistical significance of the data, presented as mean ± SEM, was evaluated using non-parametric Mann–Whitney U tests. Differences were considered statistically significant at p < 0.05.

Aortic transplantation procedures using the cuff and sleeve anastomosis strategies were performed in male mice aged 8–12 weeks. A table recapitulating the ischemia times of both surgeries (Table 1) shows no statistical difference in the total and cold ischemia times between the sleeve and cuff groups but reveals a significantly longer warm ischemia time for the sleeve group. At 1 and 2 months after abdominal aortic transplantation, the aortic grafts were harvested and analyzed by histology (Figures 1A,B). Sections were taken every 100 μm starting from the cranial anastomosis and until the caudal anastomosis to provide an analysis of the full length of the graft and eventually identify anastomosis-dependent effects (Figure 1B). In both groups and at both time-points, the aortic graft cross-sections stained with Masson Trichrome displayed a large neointima composed of multiple cell layers and extracellular matrix (Figure 1C). The media appeared partially depleted of cells, indicating an immune reaction against the donor. The adventitia displayed extensive infiltrates of mononuclear cells, oftentimes organized as aggregates, suggesting an organization of immune cells in lymphoid follicles (Figure 1C). As the proximity to the anastomosis may affect the extent of intimal thickening (17, 18), we separated the graft into three zones for the quantitative analysis: cranial anastomosis; middle; and caudal anastomosis (Figures 1B,C). At 1 month post-transplantation, the proximity to the anastomosis indeed impacted the extent of intimal hyperplasia in the cuff group (cranial vs. middle, p = 0.056; caudal vs. middle, **p < 0.01). At 1 month post-transplantation, the extent of TV, expressed as the neointima/media ratio, was significantly higher in the sleeve vs. the cuff group in the cranial and middle areas (**p < 0.01, Mann–Whitney U test) (Figure 1C), reflected by significant differences in the lumen and neointima in the middle portions (Supplementary Figure S1). At 2 months, the neointima/media ratio was significantly higher in the sleeve vs. the cuff group in the cranial anastomosis portion (*p < 0.05, Mann–Whitney U test) (Figure 1C, Supplementary Figure S1). There were no major time-dependent effects of the average neointima/media ratio. The difference between 1 and 2 months in the cuff group was more noticeable than in the sleeve group (p = 0.056), suggesting that the development of TV in the cuff group was more progressive than in the sleeve group (Figure 1D).

To interrogate potential differences in the cellular composition of the grafts between the sleeve and cuff group, we performed a quantitative analysis of immunofluorescently stained aortic grafts.

We quantified VSMCs and hematopoietic cells inside the neointima based on their positivity for αSMA and CD45, respectively (Figure 2A). The numbers of neointimal αSMA-positive VSMCs, normalized to total neointimal cells (DAPI+), were higher on average in the cuff group vs. the sleeve group, as they represented 66% vs. 60% at 1 month, and 80% vs. 64% at 2 months, respectively (Figure 2B).

The distribution of neointimal VSMCs was different between the two groups: VSMCs populated the luminal layers of the neointima in the sleeve model, while they were evenly distributed throughout the whole neointimal layer in the cuff model (Figure 2A). To summarize this observation, we normalized the VSMC numbers to the neointimal area. The resulting neointimal VSMC coverage was significantly higher in the cuff group vs. the sleeve group at 2 months post-transplantation (*p < 0.05, Mann–Whitney U test) (Figure 2C). There were no major differences in either the CD45+ hematopoietic cells infiltrated within the neointima or in the numbers of αSMA-negative and CD45-negative cells between the cuff and sleeve groups (Figures 2B,C). No differences were observed when the various cell subsets were normalized to total DAPI+ cells, meaning that the difference observed came from the distribution of VSMCs throughout the neointimal area and not their overall proportion (Supplementary Figure S2).

These data indicate that the type of anastomosis is sufficient to impact the cellular composition and distribution inside the neointima during TV.

TV, as a hallmark of chronic rejection, is characterized by the activation of the adaptive immune system against the graft. To explore the effect of the anastomosis on the types of lymphocytes found in situ, we quantified CD3+ T cells and CD19+ B cells on cross-sections of aortic grafts.

T cells were observed in all three layers (neointima, media, adventitia) in grafts from both groups (Figure 3A). No major differences were observed in the total T-cell content per section, in the neointima or in the media in the cuff vs. the sleeve models at 1 month after transplantation (Figure 3B). At 2 months, however, higher numbers of T cells infiltrated the neointima in the sleeve group compared to the cuff group (*p < 0.05, Mann–Whitney U test) (Figure 3B).

B cells were found in the adventitia in all groups (Figure 4A). They were either clustered together and intertwined with T cells, suggesting lymphoid aggregates, or disseminated along the external elastic lamina. This differential organization and variations in B-cell morphology (small round cells with a low cytosol/nucleus ratio inside the aggregates; larger cells with a higher cytosol/nucleus ratio in the disseminates fraction) suggest the coexistence of B-cell subpopulations inside the graft (as shown in Figure 4A, in the high magnification for the cuff group at 2 months). Unlike T cells, B cells were not observed in the neointima and the media (Figure 4A). Intriguingly, infiltrating B-cell numbers were significantly higher in the cuff group compared to the sleeve group at 2 months post-transplantation (**p < 0.01, Mann–Whitney U test) (Figure 4B). The differences observed in the B-cell numbers were not reflected by the quantification of systemic DSAs, whose levels did not show any significant differences between the sleeve and the cuff group (Supplementary Figure S3), suggesting and uncoupling between local intra-graft B cells and circulating antibodies.

To better understand the processes underlying the differences observed between the two methods, we performed an immunofluorescence analysis of two markers of endothelial activation (ICAM-1, vWF) and an immunohistochemical analysis of the chemokine IP-10 (interferon-gamma-inducible protein 10 kDa), associated with graft rejection in patients, in particular antibody-mediated rejection (ABMR). At 1 month after aortic transplantation, the endothelial activation marker abundance in the endothelial layer was unchanged between the two methods (Supplementary Figure S4), while, interestingly, IP-10 abundance within the grafts was significantly higher when using the cuff vs. sleeve anastomosis (*p < 0.05, Mann–Whitney U test) (Figure 5).

Overall, our data uncovered a prominent effect of the type of anastomosis on the local immune profile of aortic grafts, associated with a differential content of the chemokine IP-10.