Deficiency in the mitophagy mediator Parkin accelerates murine skin allograft rejection

Abstract

Alterations in mitochondrial function and associated quality control programs, including mitochondrial-specific autophagy, termed mitophagy, are gaining increasing recognition in the context of disease. However, the role of mitophagy in organ transplant rejection remains poorly understood. Using mice deficient in Parkin, a ubiquitin-ligase which tags damaged or dysfunctional mitochondria for autophagic clearance, we assessed the impact of Parkin-dependent mitophagy on skin-graft rejection. We observed accelerated graft loss in Parkin-deficient mice across multiple-skin graft models. Immune cell distributions post-transplantation were largely unperturbed compared to wild-type; however, the CD8+ T cells of Parkin-deficient mice expressed more T-bet, IFNγ and Ki67, indicating greater priming towards effector function. This was accompanied by increased circulating levels of IL-12p70 in Parkin-deficient mice. Using a mixed leukocyte reaction, we demonstrated that naïve Parkin-deficient CD4+ and CD8+ T cells exhibit enhanced activation marker expression and proliferative responses to alloantigen, which were attenuated with administration of a pharmacological mitophagy inducer (p62-mediated mitophagy inducer), known to increase mitophagy in the absence of a functional PINK1-Parkin pathway. These findings indicate a role for Parkin-dependent mitophagy in curtailing skin-graft rejection.

1. Introduction

Organ transplantation is a lifesaving intervention for end-stage organ failure; however, long-term graft survival is limited by immune-mediated rejection¹. Understanding rejection mechanisms are therefore essential to improving transplantation outcomes. T cells are central mediators of allograft rejection, recognizing polymorphisms in donor minor- and major-histocompatibility complex (MHC) proteins^2–6. In generating effector responses, naïve T cells rapidly expand and differentiate into a pool of antigen specific clones requiring signalling, metabolic, epigenetic, and transcriptional modifications^7,8. Mitochondria are involved throughout the T cell response, including in formation of the immune synapse and regulation of intracellular Ca²⁺ concentrations^9–11, production of reactive oxygen species^12,13, ATP and biomolecule synthesis^7,14 and in determining T cell differentiation^8,15,16. Given these multifaceted roles for mitochondria, understanding the contribution of mitochondrial quality control mechanisms to allorecognition and rejection may offer new approaches to modulate the immune response.

Mitochondrial-specific autophagy, mitophagy, selectively removes damaged, dysfunctional, or superfluous mitochondria from cells via ubiquitin-dependent and independent pathways, of which the PINK1/Parkin pathway is well-characterized^17,18. Mitochondrial damage or dysfunction causes mitochondrial membrane potential to dissipate, and PINK1 stabilization and activation on the outer mitochondrial membrane^19–24. Subsequently, Parkin is recruited to damaged mitochondria, phosphorylated by PINK1, and activated to ubiquitin-tag mitochondrial membrane proteins for lysosomal recognition and degradation^{19,23,25–28}. Despite increasing recognition of perturbed mitochondrial quality control in various pathological conditions^18,29, the role of Parkin-mediated mitophagy remains unclear in organ transplant rejection. Thus, we aimed to assess the impact of Parkin-deficiency in skin-graft recipients and the role of mitophagy in immune responses to alloantigen.

In this study, utilizing Parkin knockout (PKO) mice, we observed accelerated rates of skin graft rejection conserved across major- and minor-antigen mismatch models and with co-stimulatory blockade (CSB). CD8+ T-cells demonstrated enhanced proliferative and IFNγ secreting potential in PKO mice post-transplantation. In vitro modelling of the allorecognition response revealed greater expression of antigen-specific activation markers and increased PKO CD4+ and CD8+ T cell proliferation to alloantigen stimulation which were attenuated with administration of a PINK1/Parkin independent mitophagy inducer (p62-mediated mitophagy inducer; PMI). Our findings highlight a role for mitophagy in organ transplant rejection.

2. Methods

2.1. Animals and ethics statement

C57BL/6J (strain #000664), BALB/cJ (strain #000651) and PKO mice (Prkntm1Shn/J; strain #006582) were obtained from the Jackson Laboratory. MitoQC mice were kindly provided by Professor Ian Ganley. PKO, MitoQC and PKO x MitoQC mice were bred in-house at the University of Michigan North Campus Research Complex breeding facility. Mice were age and gender matched within experiments. Experiments were approved by the Institutional Animal Care and Use Committee of the University of Michigan (PRO00010704) and in accordance with NIH guidelines.

2.2. Skin graft

Full thickness tail-skin grafts (~8mm diameter circles) were obtained from BALB/cJ or C57BL/6J donor mice. Hair was removed from the dorsal thorax of recipient mice, and graft bed prepared by excision of an ~6mm diameter circular piece of skin. Grafts were secured on the recipient’s graft bed with at least 8 interrupted 6–0 nylon sutures. On day 7, sutures were removed, and grafts were monitored daily for rejection by 2 blinded, independent assessors. Grafts were considered rejected when ≥80% of the graft area appeared necrotic. Where applicable, recipient mice were administered 250μg CTLA-4Ig (BioXCell, #BE0099) and 250μg anti-CD40L antibody (BioXCell, #BE0017–1) intraperitoneally (I.P.) at days 2, 4, 6 and 8 post-transplantation.

2.3. Quantification of peripheral cytokines

Circulating cytokine levels in plasma were quantified via cytometric multiplex bead assay using the LEGENDplex^™ mouse inflammation panel 13-plex (BioLegend, #740446) according to manufacturer’s instructions.

2.4. Tissue processing

Blood was collected via cardiac puncture and anti-coagulated with heparin (1000 IU/mL blood), centrifuged (1200g,15 minutes), plasma aspirated, and the remaining blood resuspended to the original volume with PBS. Red blood cells (RBCs) were lysed with ACK lysis buffer (Gibco, #A10492–01) for 2 minutes, quenched with PBS, centrifuged, and the lysis protocol was repeated.

Skin draining lymph nodes (LN; axillary, brachial and inguinal) and spleens were passed through 70μm cell strainers, and centrifuged. Spleens underwent a round of RBC lysis, as described above.

2.5. Flow cytometry staining

Single cell suspensions were washed with PBS, stained with live/dead fixable viability stain (ThermoFisher) and Fc blocked (BioLegend, #101320) prior to surface staining with antibodies of interest (Supplementary Table 1) in FACS wash (PBS + 2% FBS; 30 minutes, 4°C). Cells were then washed and fixed in 1% formaldehyde. For assessment of intracellular cytokines or transcription factors, cells were fixed and permeabilized with BD Cytofix/Cytoperm^™ Fixation/Permeabilization (BD Biosciences, #554714) or Transcription Factor Staining kit (BD Biosciences, #562725) before staining with antibodies targeting markers of interest.

2.6. Flow cytometric analysis

Samples were acquired on a BioRad Ze5 or BD FACSymphony A5 and data analyzed with FlowJo Software v10. To determine total cell counts, a defined volume of single cell suspension was recorded by the cytometer and adjusted to the proportion of sample stained, and total tissue count obtained at harvest using a CytoSMART Exact automated cell counter (Axion BioSystems), or volume of blood stained.

2.7. Media

Complete media: RPMI (Gibco, #11875–093) + 10% FBS (Sigma, #F2442) + 5% Penicillin/Streptomycin + 10 mM HEPES + 50 μM β-mercaptoethanol (Gibco, #15140–122, 15630–080, 21985–023). Mixed leukocyte reaction (MLR) media: complete media + 10 mM NEAA (Gibco, #11140–050) + 1 mM sodium pyruvate (Gibco, #11360–070).

2.8. Bone marrow derived dendritic cells (BMDC) generation

Marrow was flushed from femur and tibia bones of donor mice with PBS, centrifuged, and RBCs lysed, as above. Cells were seeded in 6-well plates (1 × 10⁶/mL) in complete media with 20 ng/mL GM-CSF (BioLegend, #576304) and 10 ng/mL IL-4 (BioLegend, #574304) for 7 days. Media was refreshed and cells re-stimulated with GM-CSF and IL-4 at days 3 and 6.

2.9. MLR

Irradiated (20Gy) donor BMDC (1.25 × 10⁴) from BALB/cJ or C57BL/6J mice and C57BL/6J or PKO responder splenocytes (2.5 × 10⁵) were labelled with 1μM of carboxyfluorescein diacetate succinimidyl ester (CFSE; Invitrogen, #C34554) or CellTrace Violet (CTV; Invitrogen, #C34557), respectively, for 10 minutes at room temperature. Cells were washed twice, resuspended in MLR media, and co-cultured for 72 hours before staining for flow cytometric analysis.

To assess alloantigen responses following mitophagy induction, cells were treated with 10 μM PMI (MedChemXpress, #HY-115576) or equivalent volume of DMSO (Sigma, #D2650) (vehicle).

2.10. In vitro stimulations

For cytokine assessment, splenocytes (2 × 10⁶/mL) were incubated with 25 ng/mL Phorbol 12-myristate 13-acetate (PMA; Sigma, #P8139) and 500 ng/mL Ionomycin (Io; Sigma, #I9657) for 6 hours at 37°C. 2 μM monensin (BD Biosciences, #554724) and 1 μg/mL Brefeldin A (BD Biosciences, #555029) were added after 1 hour. For T cell proliferation assessment, splenocytes were stimulated with anti-CD3/CD28 coated beads (Gibco, #11456D) for 72 hours (4.0 × 10⁴ beads per 2.5 × 10⁵ spleen cells). For in vitro comparison of MitoQC and PKO × MitoQC mitophagy, splenocytes were stimulated with either anti-CD3/CD28 beads (6.0 × 10⁴ beads per 4.0 × 10⁵ cells) for T cell responses or 25ng/mL Lipopolysaccharide (LPS; eBioscience^™, #00–4976-93) for 24 hours for assessment of Dendritic cell (DC) responses.

2.11. Regulatory T cell (Treg) suppressor assay

Tregs and responder (CD4+CD25−) T cells were isolated from naïve C57BL/6J or PKO spleens using the Regulatory T Cell Isolation Kit (Miltenyi, #130–091-041), and labelled with CFSE or CTV, respectively. 1 × 10⁵ responder cells were co-cultured for 72 hr, 37°C with: 1.25 × 10⁴ – 5.0 × 10⁵ Tregs, 1 × 10⁵ autologous, CFSE-labelled, irradiated (20Gy) splenocytes, and 1 μg/mL anti-CD3 antibody (BioLegend, #100340). Cells were then stained for flow cytometric analysis. Percent suppression was calculated relative to maximum proliferation (no Treg controls) for each strain.

2.12. Statistics

Statistical significance was calculated in GraphPad Prism 9.4.1 by Mann-Whitney U test, with corrections for multiple tests using the Holm-Sidak method where applicable. Survival was determined by log-rank analysis and necrotic score over time data assessed by a 2-way ANOVA with repeated measures. P<0.05 was considered statistically significant.

3. Results

3.1. Accelerated skin allograft rejection in PKO recipients

To investigate the impact of mitophagy in organ transplant rejection, we first utilized a model of skin allograft rejection in which donors and recipients are MHC-disparate. Tail-skin was transplanted from BALB/cJ donor mice (H-2^d) to C57BL/6J (H-2^b; Wild-type (WT)) or whole-body PKO (C57BL/6 genetic background) recipient mice, which globally lack functional Parkin and hence Parkin-dependent mitophagy. To temper rejection rates, recipient mice were administered CTLA-4Ig and anti-CD40L antibody as CSB (Figure 1A) and monitored for necrosis. We observed earlier onset of graft necrosis (median 12 vs 14 days post-transplantation, p=0.0195; Figure 1C) and an increased necrotic score (percent area necrotic) over-time (p=0.0146; Figure 1D) in PKO recipients compared to WT controls. We also observed a modest, but significant increase in the kinetics of graft rejection in PKO recipients (median survival times (MST) of 15 vs 17 days post-transplantation; p=0.0438; Figures 1E–F).

Figure 1:

Skin allografts are rejected more rapidly in PKO than WT recipients. (A) Schematic of experimental design. PKO (n=12) or WT (n=11) recipient mice received tail skin allografts from BALB/cJ donor mice. Recipient mice were administered CSB with CTLA-4Ig (250ug) and anti-CD40L (250ug) I.P. at days 2, 4, 6 and 8 post-transplantation. (B) Skin grafts were monitored daily for signs of rejection and scored based on the total area of the graft necrotic. (C) The day post-transplantation which grafts achieved a necrotic score of 1 or greater and (D) the necrotic score over time was assessed in graft recipients. (E) Grafts were considered rejected when at least 80% of the total graft area appeared necrotic (necrotic score=5). (F) Representative images of graft appearance in WT and PKO recipient mice post-transplantation. Images are not to scale. Significance of graft survival was assessed using the log-rank test, time to necrosis by Mann-Whitney U-test and necrotic score over-time using a 2-way ANOVA with repeated measures, with p-value reported for the comparison between strains over time (row × column factor). P<0.05 was considered statistically significant. Error bars represent standard error of the mean (SEM). Schematic in panel A created with BioRender.com

To determine whether accelerated graft rejection was specific to the CSB model, we assessed graft survival in two alternate skin-transplantation models: 1) MHC-disparate strain combination without CSB; and 2) H-Y minor antigen mismatch model. We also noted a modest, albeit significant increase in the rate PKO mice rejected BALB/cJ allografts without CSB (MST 8.5 vs 9.5 days post-transplantation, p=0.0171; Supplementary Figure 1A), and greater necrotic-score overtime (p=0.0019; Supplementary Figures 1B–C), and a more substantial difference in graft-survival with minor-antigen mismatch (MST 19 vs 26 days post-transplantation, p=0.0061; Supplementary Figure 1D). These skin graft models suggest Parkin deficiency accelerates allograft rejection.

3.2. Unaltered peripheral immune cell distributions between strains at day 13 post-transplantation

To determine if skin-graft rejection rate differences were associated with alterations in immune cell distributions between the strains, we characterized frequency and number of major leukocyte populations in transplant recipients which received CSB. Our analysis focused on day 13 post-transplantation where the greatest difference in mean necrotic score was observed (WT=1, PKO=2.82; Figure 1D). Flow cytometric analysis (Supplementary Figure 2) of the spleen (Figure 2A), blood (Figure 2B) and skin draining LN (Figure 2C) revealed no significant perturbations in frequency or total number of B cells, total, CD4+, or CD8+ T cells, NK cells, monocytes/macrophages, or neutrophils between WT or PKO mice. Differences in total CD45+ live cell counts were not observed between WT and PKO spleen or blood (Figure 2A–B).

Figure 2:

Immune cellular distribution is similar between PKO and WT mice post-transplantation. WT (n=9) and PKO (n=8) mice received BALB/cJ tail skin grafts with CTLA-4Ig and anti-CD40L CSB. At day 13 post-transplantation, cell counts and frequency were assessed by flow cytometry in the spleen (A), blood (B) and skin-draining LN (C). Statistics assessed by Mann-Whitney U test and corrected for multiple comparisons with the Holm-Sidak method. P<0.05 was considered statistically significant. Error bars represent SEM.

3.3. Assessment of peripheral cytokine concentrations between PKO and WT mice

Since immune cell distributions at day 13 post-transplantation in CSB recipients were similar, we investigated peripheral cytokine profiles by multiplex cytometric bead assay as a potential explanation for differences noted in MST noted between strains. As shown in Figure 3, we observed significantly more IL-12p70 (p=0.0277) within the plasma of PKO mice, a cytokine known to drive differentiation of Th1 cells and promote IFNγ production in CD8+ T cells and NK cells³⁰. We also observed non-significant trends towards increased IFNγ (p=0.0825) and GM-CSF (p=0.0553) within PKO mouse plasma. In naïve mice, these same differences were not present, although we did observe mildly decreased TNFα (p=0.0173) and IL-23 (p=0.0281) in PKO animals at baseline compared to WT. Following transplantation, we observed significant increases in IL-12p70 (WT p=0.0041; PKO p=0.0258) and IL-17A (WT p=0.0195; PKO p=0.0351) in both strains relative to their naïve counterparts.

Figure 3:

Analysis of peripheral cytokines in WT and PKO naïve and allogeneic skin graft recipients. Plasma was obtained from naïve WT (n=6) or PKO (n=6), and WT (n=14) or PKO (n=13) mice which received BALB/cJ tail skin grafts with CTLA-4Ig and anti-CD40L CSB at day 13 post-transplantation. Inflammatory cytokines were quantified by flow-cytometry based multiplex immunoassay. Statistics assessed by Mann-Whitney U test. P<0.05 was considered statistically significant. Error bars represent SEM.

3.4. Altered proliferative potential and cytokine production in PKO T cells post-transplantation

Given the differences in peripheral cytokines, and importance of T cells in rejection responses, we next investigated if there were strain-based phenotypic differences in T cells post-transplantation. We observed no difference in expression of the Th1 master transcriptional regulator, T-bet amongst CD4+ T cells; however, significant increases in the Th17-regulator RORγT (p=0.0493) and FoxP3+CD25+ Treg frequency (p=0.0055) were noted within the PKO CD4+ compartment (Figure 4A). Contrastingly, T-bet expression was elevated in PKO CD8+ T cells, consistent with greater priming towards cytotoxic effector function (p=0.0213; Figure 4B).

Figure 4:

PKO CD8+ T cells have altered effector phenotype following allograft transplantation. On day 13 post-transplantation, single cell suspensions were prepared from the spleens of PKO or WT mice which received BALB/cJ skin allografts and CSB with CTLA-4Ig and anti-CD40L. (A) CD4+ T cells were intracellularly stained the expression of T-bet, RORγT and FoxP3/CD25. (B) CD8+ T cells were characterized for the expression of T-bet by intracellular staining. The frequency of IFNγ, TNFα, IL-2 and IL-17A producing cells amongst the bulk CD4+ (C) and CD8+ (D) populations were analyzed in response to PMA/Io stimulation for 6 hours. Additional analysis of the cytokine producing frequencies amongst CD44^hi CD4+ (E) and CD44^hi CD8+ (F) populations, as well as the dual IFNγ+TNFα+producing CD8+populations (G). CD4+ and CD8+ T cells were also assessed for Ki67 expression ex vivo. Representative staining for Ki67 expression on bulk CD4+ and CD8+ T cell populations (I). (J) Summary of Ki67 expression data assessed by Mean Fluorescence Intensity (MFI) and frequency amongst indicated T cell populations in WT (n=14) and PKO (n=13) mice. (K) Ki67 MFI and frequency assessment amongst CD44^hiCD4+ and CD44^hiCD8+ populations. Significance was assessed by Mann-Whitney U test, with p<0.05 considered significant. Error bars represent SEM.

To assess whether transcriptional profile differences translated to functional differences between PKO and WT T cells, maximal cytokine production potential was determined by stimulating day 13 post-transplantation splenocytes with PMA + Io. No difference was observed in IFNγ, TNFα, IL-17A, or IL-2 producing frequency of WT or PKO CD4+ T cells, (Figure 4C), however, PKO recipients had increased splenic CD8+ T cell IFNγ+ frequency and trend towards increased TNFα frequency (p=0.0206 and 0.1139, respectively; Figure 4D). As a surrogate for antigen-activated cells, we interrogated CD44^hiCD4+ and CD44^hiCD8+ T cells for cytokine producing frequency and observed an additional trend towards increased IL-17A producing frequency amongst CD44^hiCD4+ PKO T cells (p=0.0745; Figure 4E), while the increased frequency of PKO IFNγ+ producing CD8+ T cells was conserved amongst the CD44TNFα+IFNγ+ producing cells amongst (bulk p=0.0464, CD8+ CD44^hi p=0.0152; Figure 4G). Maximal cytokine production did not differ between the strains prior to transplantation (Supplementary Figure 3A–B), suggesting transplantation-induced expansion of pro-inflammatory cytokine producing CD8+ T cells.

Compared to other CD4+ T cells, Tregs have a greater dependence on mitochondrial-based energy production which may regulate their suppressive functions^15,31,32. Assessment of naïve splenic frequencies suggested that the PKO Treg population is also elevated in frequency at baseline (p=0.0043; Supplementary Figure 3C–D), and we postulated that this may serve to compensate for decreased suppressive capacity. However, a Treg suppressor assay showed no difference between WT and PKO Tregs in their ability to suppress T cell proliferation (Supplementary Figure 3E).

Interestingly, both PKO CD4+ and CD8+ T cells were more frequently Ki67^hi (CD4+ p=0.0321; CD8+ p=0.0150) and had greater total Ki67 expression than WT post-transplantation (CD4+ p=0.0439, CD8+ p=0.0255; Figure 4I–J). Further assessment amongst the CD44^hi populations echoed this phenotype (Figure 4K), suggesting greater proliferative potential in PKO T cells post-transplantation.

3.5. Naïve PKO T cells have altered responsiveness to alloantigen

As we observed accelerated graft rejection and enhanced proliferative potential post-transplantation, we hypothesized PKO mice would have altered basal responses to alloantigen. We performed a MLR to monitor T cell proliferative responses in naïve PKO and WT mice and assessed activation induced marker (AIM) expression as a surrogate measure of antigen-specific responses. AIM expression has been utilized to capture the breadth of antigen-specific T cell responses irrespective of functional phenotype and includes assessment of TCR-inducible activation marker expression³³. We observed enhanced proliferation of PKO CD4+ and CD8+ T cells (p=0.0022, Figure 5A–B; p=0.0043, Figure 5E–F, respectively) in response to alloantigen stimulation and AIM expression at 72-hour post-stimulation (p=0.0411, Figures 5C–D; p=0.0152; Figure 5G–H, respectively); however, no difference was observed in response to syngeneic stimulation or without stimulation (Figure 5A–D). Analysis of single activation markers revealed an overall profile of enhanced T cell alloantigen activation in PKO splenocytes with increased CD69 (p=0.0087), OX-40 (p=0.0411) and CD25 (p=0.0087) expression amongst the CD4+ compartment (Figure 5I), and CD69 (p=0.0411) and OX-40 (p=0.0260) within the CD8+ subset (Figure 5J). Overall, our findings suggest PKO CD4+ and CD8+ T cells mount more substantial, antigen-specific proliferative responses to alloantigen than WT T cells pre-transplantation. To assess if these differences in PKO and WT CD4+ and CD8+ proliferative responses were observed independently of alloantigen stimulation, we stimulated splenocytes from the same WT and PKO responders with anti-CD3/CD28 for 72 hours and observed no significant differences between the strains (Supplementary Figures 4A–B). We also assessed the capacity of WT or PKO BMDCs to stimulate BALB/cJ T cell responses but observed no difference in CD4+ or CD8+ proliferation or AIM expression (Supplementary Figure 4C–F).

Figure 5:

Alloantigen stimulates greater CD4 and CD8 T cell activation and proliferative responses in Parkin KO mice. Splenocytes from naïve Parkin KO and WT mice were stimulated with irradiated BALB/cJ or C57BL/6 (B6) BMDC for 72 hours. (A) CD4 T cell proliferation was assessed via CTV dye dilution. (B) Representative CTV histograms for BALB/cJ BMDC stimulated CD4 T cells. BMDC stimulated histograms are overlayed in red. (C) Quantification of AIM expression on CD4 T cells. (D) Representative flow cytometry plots of AIM expression on CD4 T cells stimulated with BALB/cJ BMDC. B6 stimulated flow cytometry plots are overlayed in grey. (E) CD8 T cell proliferation was assessed via CTV dye dilution. (F) Representative histograms of CD8 T cell proliferation in response to BALB/cJ BMDC stimulation, overlayed with B6 BMDC stimulated control histograms (red). (G) Quantification of AIM expression on BALB/cJ and B6 BMDC stimulated CD8 T cells and (H) representative flow cytometry plots of BALB/cJ stimulated CD8 T cells with B6 stimulated flow cytometry plots overlayed in grey. Quantification of activation marker expression in alloantigen stimulated CD4 (I) and CD8 T cells (J). Error bars represent SEM, n=6 per group. Significance was assessed by Mann-Whitney U test. P<0.05 considered significant.

3.6. Pharmacological induction of mitophagy attenuates PKO T cell responsiveness to alloantigen

To investigate if mitophagy was altered during allograft rejection, we performed skin-graft experiments using MitoQC reporter mice. MitoQC mice have a mitochondrial GFP-mCherry tag; upon delivery to acidic lysosomes, GFP fluorescence is quenched, allowing only mCherry to be detected³⁴. We assessed total mitophagy by the frequency of GFP^lowmCherry+ (‘lysosomal frequency’) splenic T cell subsets at day 4 (before rejection), day 8 (during graft rejection) and day 14 (post-rejection) of allogeneic and syngeneic skin-graft recipients without CSB. At day 8 post-transplantation, allografted splenic CD4+ T cells had significantly greater mitophagy than syngeneic controls (p=0.0368, Supplementary Figure 5A–B) and a similar trend was observed in CD8+ T cells (p=0.0645, Supplementary Figure 5A–B), suggesting an overall increase in T cell mitophagy that was not present at day 4 and 12 post-transplant (Supplementary Figure 5C). Concurrent assessment of splenic DCs revealed no difference in total mitophagy between allogeneic or syngeneic graft recipients at any of the assessed timepoints (Supplementary Figure 5D–E).

As validation of altered mitophagy with Parkin deficiency, we assessed lysosomal frequencies amongst T cell subsets between MitoQC and PKO × MitoQC mice. With CD3 + CD28 stimulation, we observed significantly lower lysosomal frequencies in PKO x MitoQC CD3+ (p=0.0196; Supplementary Figure 6A–B), CD4+ (p=0.0275; Supplementary Figure 6C–D), and CD8+ (p=0.0231; Supplementary Figure 6E–F) splenic T cells compared to MitoQC mice, suggesting that total T cell mitophagy is altered in PKO mice. However, we were not able to resolve differences in DCs (Supplementary Figure 6G–H).

To understand if altered T cell responses observed to alloantigen in vitro related to Parkin’s role in mitophagy, we included an established mitophagy inducer, PMI, which induces mitophagy in the presence and absence of a functional PINK1/Parkin pathway^35,36 in our MLR. PMI treated PKO CD4+ T cells displayed decreased proliferation (Figures 6A–B; p=0.0022) and AIM expression (Figures 6C–D; p=0.0022). Mitophagy induction also reduced proliferation (Figures 6E–F; p=0.0022) and AIM expression (Figures 6G–H; p=0.0022) in PKO CD8+ T cells. Consistent with PMI inducing mitophagy in both the presence and absence of functional Parkin, mitophagy induction with PMI similarly affected CD4+ and CD8+ proliferation and AIM marker expression in WT and PKO T cells. Collectively, these data suggest the accelerated alloantigen responses in PKO CD4+ and CD8+ T cells is due in part to a deficiency in Parkin-dependent mitophagy.

Figure 6:

Mitophagy induction attenuates T cell proliferation and activation in vitro. PKO and WT splenocytes were treated with 10μM PMI or vehicle control and stimulated with BALB/cJ BMDC for 72 hours. (A) Proliferation was assessed via CTV dye dilution as depicted in representative histograms (B) for PKO vehicle and PMI treated samples. (C) AIM expression was quantified for CD4+ T cells. (D) Representative flow cytometry plots for AIM expression in PKO vehicle and PMI treated CD4+ T cells. (E) Quantification of proliferation via CTV dye dilution on WT and PKO CD8+ T cells. (F) Representative histograms for PKO CD8+ T cell proliferation in vehicle and PMI treated samples. (G) AIM expression was assessed with PMI treatment in PKO and WT CD8+ T cells. (H) Representative flow cytometry plots for PMI and vehicle treated PKO CD8+ T cells. Data shown is representative of two-independent experiments; error bars signify SEM. Statistics assessed by Mann-Whitney U test. P<0.05 considered significant.

4. Discussion

Mitochondrial specific autophagy, known as mitophagy, remains understudied in the context of allorejection. To reconcile this deficit in knowledge, we evaluated the contribution of Parkin-dependent mitophagy to allograft rejection using multiple skin-graft models. With PKO recipient mice, we show accelerated rates of skin-graft rejection across models of major- and minor-antigen mismatch, and in the presence of CSB (Figure 1 and Supplementary Figure 1). During allograft rejection, PKO CD4+ and CD8+ T cells displayed increased proliferative capacity, and IFNγ secreting potential in CD8+ T cells (Figure 4). Prior to transplantation, PKO CD4+ and CD8+ T cells mounted greater proliferative and antigen-specific AIM responses to alloantigen (Figure 5) that was attenuated with a Parkin-independent mitophagy inducer (Figure 6)³⁵. Collectively, our results indicate a role for Parkin-dependent mitophagy in tempering alloantigen responses in skin-graft recipients.

There are increasing associations between mitophagy perturbations and human disease; however, studies investigating mitophagy effectors in the adaptive immune system are limited. Utilizing viral infection models, Gupta et al., and Franco et al. independently implicate mitophagy-mediators NIX^37,38 and Parkin³⁸, in CD8+ T cell memory formation. Mitophagy deficiency has also been reported to exacerbate mitochondrial dysfunction and exhaustion in tumor infiltrating T cells chronically exposed to antigen³⁹. While these studies highlight importance for mitophagy mediators late in T cell responses or with persistent antigen exposure, the role of mitochondrial quality control programs early in primary T cell responses warrants further investigation. Our study demonstrates enhanced alloantigen responsiveness in naïve PKO T cells within 72 hours of antigen stimulation suggesting that mitophagy may assist in regulating these early responses.

Alternate roles for Parkin beyond mitophagy have been suggested, including as a tumor-suppressor, owing to reports of Parkin mutation and loss of expression in multiple human cancers and elevated rates of tumorigenesis in Parkin-deficient mice^40–47. These cancer phenotypes have been linked to increased proliferative capacity of Parkin-deficient cells and those bearing the mutations observed in human malignancy, which are abrogated with Parkin-overexpression^{41–43,46,48,49}. In our study, we observed increased PKO CD4+ and CD8+ T cell proliferation in response to alloantigen stimulation (Figure 5) which was attenuated with pharmacological mitophagy induction (Figure 6). Although our studies suggest a link between Parkin’s role in mitophagy and decreased proliferation, alternate mechanisms may also be involved. Parkin’s ubiquitin ligase activity has been implicated in cell cycle arrest and mitosis regulation; however, rigorous assessment of Parkin-dependent mitophagy is limited in these studies^41,49–52.

While the PINK1/Parkin mitophagy pathway is well-characterized, alternate, and potentially overlapping mitophagy pathways also exist. Proteins including NIX and BNIP3 can be expressed on the outer mitochondrial membrane and facilitate receptor-mediated mitophagy by directly engaging autophagosomal machinery^53–56. Although regulation of receptor- and ubiquitin-mediated mitophagy appears different^54,57–59, potential interactions exist between these pathways^60,61. Moreover, alternate PINK1-dependent E3-ubiquitin ligases also exist, suggesting that PINK1 recruitment of autophagic machinery can occur without Parkin^62–64. Despite redundancies in mitophagy pathways, we observed accelerated graft-rejection in the absence of Parkin across multiple skin-graft models (Figure 1 and Supplementary Figure 1).

Supporting a relationship between inadequate mitophagy and allograft rejection, RNA sequencing of human kidney allografts revealed downregulation of mitophagy pathways and transcripts related to PINK1 and Parkin in biopsies categorized as acutely rejecting compared to those classified as normal⁶⁵. Further understanding of the interplay, regulation, and cues of mitochondrial quality control pathways in vivo and across tissue-types will be paramount to informing relevance of Parkin in models of transplantation rejection.

With a pharmacological mitophagy inducer, we decreased proliferation and AIM expression in response to alloantigen in CD4+ and CD8+ T cells in PKO mice (Figure 6), suggesting mitophagy induction in transplant recipients may prolong allograft survival. However, since both donor and responder cells were treated with the mitophagy inducer, it is unclear whether this phenotype is a direct consequence of PMI’s effects on responder cells and/or the donor cells. Multiple groups have investigated the impact of modulating bulk autophagy in various organ transplant models with conflicting results^66–70. Clinically used mTOR inhibitors are known to induce autophagy; however, the degree to which autophagy induction contributes to immune suppression is unclear^71,72. mTOR signaling may also regulate PINK1/Parkin mitophagy and mTOR inhibition may similarly induce mitophagy^73,74. Our study provides rationale to further investigate the role of mitophagy, and impact of clinically used immunosuppression on mitophagy, in future pre-clinical and clinical cohorts.

Our analysis primarily focused on the impact of mitophagy deficiency in the recipient’s immune response as we believe this offered the most practical opportunity for therapeutic intervention. However, the influence of mitophagy modulation exclusively in the donor or graft were not assessed in this study. Damage-associated molecular patterns (DAMPs), accumulate with mitochondrial stress, damage, and dysfunction, and can activate innate immune response pathways^75,76. Ischemic reperfusion is one such source of these DAMPs during transplantation, which may contribute to graft damage and dysfunction⁷⁷. Whether inducing mitophagy could alleviate the source of DAMPs and attenuate alloantigen responses remains unknown and requires further investigation.

Collectively, our study demonstrates accelerated skin-graft rejection in the absence of Parkin expression, which was associated with enhanced T cell activation and proliferative responses to alloantigen. Both proliferation and activation were attenuated by exposure to a mitophagy inducer in vitro, demonstrating a previously unreported role of mitophagy in T cell activation in the context of transplantation. In addition to highlighting a potential role for Parkin-dependent mitophagy in murine skin allograft models, this study provides rationale for further investigation of the role of mitophagy in alternate transplantation models, and in the donor. Understanding the role of mutations in Parkin and altered Parkin-dependent mitophagy in human graft rejection would be of particular interest in future studies.

Supplementary Material

supplement

Supplementary figure 1: Accelerated graft rejection in PKO mice recapitulates over multiple skin-transplantation models. Graft survival (A), necrotic score (B) and representative images of grafts over-time (C) in PKO and WT mice transplanted with BALB/cJ tail skin without administration of CSB. N=12 per group. Images are not to scale. (D) Male WT skin grafts were transplanted to female WT or PKO recipients and graft survival assessed N=9 per group. Significance of graft survival curves was assessed by log-rank analysis. Necrotic score over-time was analyzed by a 2-way ANOVA with repeated measures and reported p-value represents comparison between strains over time (row × column factor). P<0.05 considered significant.

Supplementary figure 2: Representative gating strategy for the identification of immune cell subsets in the spleen, blood and skin-draining LNs. Following identification of CD45+ events, cells were identified by FSC-A vs SSC-A, followed by exclusion of doublets (FSC-A vs FSC-H) and dead cells. T cells were identified at CD3+, and further discriminated into CD4+ CD8− (CD4+ T cells) and CD4− CD8+ (CD8+ T cells). Monocytes/macrophages were identified as Siglec F-, Ly6G-, NK1.1- and CD11b^hiCD115+. NK cells were defined as CD3-CD19-NK1.1+, B cells as CD3-NK1.1-CD19+ and neutrophils as CD3-CD19-NK1.1- cells which were CD11b^hiLy6G+.

Supplementary figure 3: Functional characterization of naïve WT and PKO T cells. Single cell suspensions were obtained from the spleens of naïve WT (n=6) and PKO (n=6) mice and stimulated for 6 hours with PMA + Io. Following stimulation, CD4+ (A) and CD8+ (B) T cells were stained for TNFα, IFNγ, IL-2 and IL-17A. Splenic Treg frequency was assessed by intracellular staining in naïve PKO (n=6) and WT mice (n=6) (C). Representative staining of Treg frequency in the spleens of naïve WT and PKO mice (D). Tregs were isolated from the spleens of naïve WT and PKO mice and co-cultured with autologous Treg-depleted CD4+ T cells, irradiated autologous splenocytes and anti-CD3 antibody for 72 hours. Treg suppressive capacity was determined by the ability of the CTV labelled autologous responder CD4+T cells (T responder) to proliferate in the presence of various Treg : T responder (E)(n=3, assayed in triplicate). Error bars represent SEM. Statistics assessed by Mann-Whitney U test. P<0.05 considered significant.

Supplementary Figure 4: Assessment of naïve proliferative responses and PKO vs WT BMDCs capacity to stimulate alloantigen responses. The proliferative capacity of naïve WT and PKO CD4+ and CD8+ T cells in response to anti-CD3 + anti-CD28 stimulation for 72 hours (A) was assessed by CTV dye dilution (n=6 per group; assayed in duplicate). (B) Representative histograms of CTV dye dilution in indicated populations of WT and PKO T cells. The capacity of WT and PKO BMDCs to stimulate BALB/cJ CD4+T cell AIM expression (C) and proliferative responses (D) in the presence of alloantigen was assessed after 72 hr co-culture. AIM expression (E) and proliferative responses (F) were also measured in WT CD8+ T cells. Error bars represent SEM. Significance was assessed by Mann-Whitney U test. P<0.05 considered significant.

Supplementary Figure 5: Mitophagy is altered during allograft rejection. MitoQC mice were transplanted with allogenic (‘allo’;BALB/cJ) major-antigen mismatch or syngeneic (‘syn’; WT) skin grafts. Total mitophagy was assessed in CD4+ and CD8+ T cell subsets prior to rejection (A; day 4 post-transplantation), around the time of rejection (B; day 8 post-transplantation) or following complete graft rejection (C; day 14). Classic DCs, identified as F4/80-ve, CD11c+ MHC II+, CD19- Ly6G- CD3- CD49b-, B220-, CD45+ live cells (D) were also assessed for total mitophagy responses at day 4, 8 and 14 post-transplantation (E). Day 4 and 8, n=6 per group; Day 14 n=3 per group. Error bars represent SEM. Significance was assessed by Mann-Whitney U test. P<0.05 considered significant.

Supplementary Figure 6: Assessment of total mitophagy responses in MitoQC vs PKO MitoQC splenic T cells and DCs. MitoQC (n=12) and PKO × MitoQC (n=12) splenic single cell suspensions were stimulated with anti-CD3/anti-CD28 stimulation for 24 hours and total mitophagy response assessed by quantification of the lysosomal frequency of bulk CD3+ (A-B), CD4+ (C–D) and CD8+ (E–F) T cells. Lysosomal frequencies were also assessed in response to LPS stimulation for 24 hours (G-H), in MitoQC (n=8) and PKO × MitoQC (n=7) splenic single cell suspensions. Error bars represent SEM. Significance was assessed by Mann-Whitney U test. P<0.05 considered significant.

Abbreviations:

AIM

activation induced marker

BMDC

bone marrow derived dendritic cells

B6

C57BL/6J

CSB

Co-stimulatory blockade

CFSE

carboxyfluorescein diacetate succinimidyl ester

CTV

CellTrace Violet

DAMP

damage-associated molecular pattern

DC

Dendritc Cell

DMSO

dimethylsulfoxide

IP

intraperitoneal

Io

ionomycin

LN

lymph node

MFI

mean fluorescence intensity

MST

median survival time

MHC

major histocompatibility

MLR

mixed leukocyte reaction

PMI

P62-mediated mitophagy inducer

PMA

phorbol 12-myristate 13-acetate

PKO

Parkin knock-out

RBC

red blood cells

WT

wild-type

Data availability statement

No high complexity data sets are contained within this manuscript.