Antibody-suppressor CXCR5⁺CD8⁺ T cells are more potent regulators of humoral alloimmunity after kidney transplant in mice compared to CD4⁺ regulatory T cells

Abstract

Adoptive cell therapy (ACT), especially with CD4⁺ regulatory T cells (CD4⁺ T_regs), is an emerging therapeutic strategy to minimize immunosuppression and promote long-term allograft acceptance, though much research remains to realize its potential. In this study, we investigated the potency of novel antibody-suppressor CXCR5⁺CD8⁺ T cells (CD8⁺ T_Ab-supp) in comparison to conventional CD25^highFoxP3⁺CD4⁺ T_regs for suppression of humoral alloimmunity in a murine kidney transplant (KTx) model of antibody-mediated rejection (AMR). We examined quantity of peripheral blood, splenic and graft-infiltrating CD8⁺ T_Ab-supp and CD4⁺ T_regs in KTx recipients and found that high alloantibody-producing CCR5 KO KTx recipients have significantly fewer posttransplant peripheral blood and splenic CD8⁺ T_Ab-supp as well as fewer splenic and graft-infiltrating CD4⁺ T_regs compared to wild-type (WT) KTx recipients. ACT with alloprimed CXCR5⁺CD8⁺ T cells reduced alloantibody titer, splenic alloprimed GC B cell quantity, and improved AMR histology in CCR5 KO KTx recipients. ACT with alloprimed CD4⁺ T_reg cells improved AMR histology without significantly inhibiting alloantibody production or the quantity of splenic alloprimed GC B cells. Studies with TCR transgenic mice confirmed antigen specificity of CD8⁺ T_Ab-supp mediated effector function. In WT recipients, CD8-depletion significantly increased alloantibody titer, GC B cells, and severity of AMR pathology compared to isotype-treated controls. Anti-CD25 mAb treatment also resulted in increased but less pronounced effect on alloantibody titer, quantity of GC B cells and AMR pathology than CD8-depletion. We report for the first time that CD8⁺ T_Ab-supp cells are more potent regulators of humoral alloimmunity than CD4⁺ T_reg cells.

INTRODUCTION

Adoptive cell therapy (ACT) is an emerging therapeutic strategy for amelioration of autoimmune disease, to augment anti-tumor immunity, and to facilitate reduction or withdrawal of conventional immunosuppression while promoting long-term allograft acceptance after transplantation. Natural CD4⁺ T_regs (CD25^highFoxP3⁺CD4⁺ T cells), known to be key regulatory cells that mediate immune self-tolerance and homeostasis, have prompted pre-clinical investigations and early stage clinical trials exploring CD4⁺ T_reg therapy as a viable immunotherapeutic strategy after transplantation (1–14). While a number of cell subsets are being investigated for immunotherapeutic purposes, CD4⁺ regulatory T cells (T_regs) have substantial pre-clinical foundation, are the frontrunner in clinical trials for solid organ transplant recipients and have demonstrated feasibility and safety for ACT. However, more research is needed to determine efficacy and risk and to optimize product (including source), clinical protocols and indications.

Antibody-mediated rejection (AMR) remains a key challenge in clinical transplantation (15–17). While conventional immunosuppressants substantially reduce the incidence of T cell-mediated rejection, there are few definitive therapies to prevent, treat, or reverse AMR. In fact, current therapies for AMR carry risks associated with global non-specific immunosuppression and, in the case of co-stimulation blockade, an unexpected increased rate of acute rejection (18–21). Furthermore, the development of donor-specific alloantibody (DSA) posttransplant is associated with a higher risk for long-term graft failure after both solid organ and cellular transplants (16, 17, 22–25). The ONE study, a multicenter international coordinated effort to evaluate cell-based therapies in the same transplant population (lower risk living donor kidney transplant recipients) with similar study design, demonstrated safety and feasibility and equivalent biopsy-proven acute rejection rates between the reference and cell therapy groups. The outcomes also demonstrate that CD4⁺ T_reg cellular therapy under the clinical trial conditions had no benefit compared to the reference group in preventing or reducing the incidence of posttransplant donor specific antibody (26). Thus, there is a need to develop new therapeutic approaches to suppress humoral alloimmunity after transplant. The availability of a clinically relevant animal model for AMR provides an opportunity to investigate the relative potency of ACT with a novel antibody suppressor CXCR5⁺CD8⁺ T cell subset we discovered in mice (CD8⁺ T_Ab-supp cells) compared to conventional CD4⁺ T_reg cells for regulation of humoral alloimmunity after kidney transplant.

The rationale for investigating the potency of CD8⁺ T_Ab-supp cells compared to a reference regulatory cell subset is based on our prior work reporting their efficacy to suppress humoral immunity (including alloimmunity) in multiple preclinical rodent models (27, 28). The functionality of these CD8⁺ T_Ab-supp cells require expression of CXCR5 and IFNγ and they mediate perforin- and FasL-dependent cytotoxic killing of antibody-producing, alloprimed B cells in vitro and in vivo (27, 29, 30). Adoptive cell therapy with CD8⁺ T_Ab-supp cells into high alloantibody producing CD8 KO hepatocyte transplant recipients and CCR5 KO KTx recipient mice inhibits alloantibody production by 5 to 14-fold (27, 28), ameliorates AMR, and enhances allograft survival (28). In human kidney transplant recipients, we have identified circulating peripheral blood lymphocytes with the same phenotype as murine CD8⁺ T_Ab-supp cells. We also found that the development of de novo DSA in first time human KTx recipients inversely correlates with the quantity of peripheral blood CXCR5⁺IFNγ⁺CD8⁺ T cells (31).

While allogeneic kidney transplantation (KTx) across many fully MHC-disparate strain combinations results in spontaneous acceptance rather than acute rejection and graft loss in the absence of immunosuppression (32–36), the Fairchild lab observed enhanced alloantibody responses and rapid allograft loss in CCR5 knockout (KO) KTx recipients (37). The CCR5 KO KTx recipient generates extraordinarily high alloantibody titers (28, 37, 38), resulting in rapid induction of AMR pathology in kidney allografts by two weeks posttransplant (32, 37, 39) that recapitulates the features of human kidney AMR pathology (37). Studies have suggested that CCR5 KO mice possess a distinct immune repertoire that promotes humoral dysregulation (28, 38, 40–42). The exaggerated humoral alloimmune responses and rapid allograft rejection observed in CCR5 KO KTx recipients has been attributed to a dysfunctional CD4⁺ T_regs population (43) or a defect in quantity or trafficking of CD4⁺ T_regs to the allograft and/or lymphoid depots (44–46) associated with the genetic deficiency of the CCR5 chemokine receptor. This premise is based on data from CCR5 KO cardiac recipients reported to have fewer graft infiltrating CD4⁺ T_regs compared to WT recipients in conjunction with comparable in vitro suppressive function of WT and CCR5 KO CD4⁺ T_regs (42). However, whether or not CD4⁺ T_reg dysfunction mediates the profound dysregulation of humoral alloimmunity observed in CCR5 KO kidney transplant recipients and if ACT with CD4⁺ T_reg reverses this phenotype has not been investigated. Furthermore, based on our prior work, we hypothesized that humoral immune dysregulation in CCR5 KO mice arises from a deficit or dysfunction of the novel regulatory cell population that we described as antibody suppressor CD8⁺ T cells (CD8⁺ T_Ab-supp cells). Thus, the purpose of this study is to determine whether deficits and/or dysfunction of CD8⁺ T_Ab-supp cells, CD4⁺ T_regs, or both contribute to the immunopathologic phenotype observed in CCR5 KO KTx recipients and to ascertain the relative potency of these cell subsets as a cellular therapy to correct the dysregulated posttransplant humoral alloimmunity.

MATERIALS AND METHODS

Experimental animals.

C57BL/6 (wild-type; WT), CCR5 KO, CD8 KO, GFP, OT-I transgenic (Tg), OT-II Tg, and mOVA Tg (all H-2^b), as well as A/J (H-2^a) and FVB/N (H-2^q) mouse strains (male and female at 8–20 weeks of age; Jackson Labs) were used in this study. OT-I mouse strain contains transgenic inserts for alpha-chain and beta-chain T cell receptor that pairs with the CD8 co-receptor and recognizes ovalbumin peptide residues 257–264 (OVA_257–264) in the context of H-2K^b such that all CD8⁺ T cells recognize MHC class I presented OVA_257–264 peptide. The OT-II mouse strain expresses the alpha-chain and beta-chain T cell receptor that pairs with the CD4 co-receptor and is specific for chicken ovalbumin peptide 323–339 peptide (OVA_323–339) presented in the context of I-A^b such that all CD4⁺ T cells recognize MHC class II presented OVA_323–339 peptide. The mOVA mouse strain expresses chicken ovalbumin (Act-mOVA) on the surface of all cells. All experiments were performed in compliance with the guidelines of the IACUCs of The Ohio State University (Protocol 2019A00000124) and Nationwide Children’s Hospital (Protocol AR17–00045).

Kidney isolation and transplantation.

Murine kidney transplantation with ureteral reconstruction was performed as previously described (28, 47). In brief, the donor left kidney is isolated by ligation and division of adrenal and gonadal vessels, followed by dissection of the ureter from renal hilus to bladder. The aorta and inferior vena cava (IVC) are mobilized at the junction of the left renal vessels, the aorta ligated superiorly, and the graft perfused in situ with 0.5 cc of cold, heparinized Ringer’s lactate solution. The kidney, its associated vascular supply, ureter, and attached elliptical bladder patch were removed en bloc and placed on ice in preparation for transplant. Vascular perfusion of the kidney graft was re-established in the recipient mouse via an end-to-side anastomosis to the recipient aorta and cava using continuous 11–0 monofilament nylon suture. Ureteral reconstruction was accomplished via anastomosis of the donor bladder patch to the recipient bladder using interrupted 10–0 nylon.

Preparation and isolation of primed CXCR5⁺CD8⁺ T cells.

Lysate was prepared from allogeneic (A/J), third-party (FVB/N), or mOVA Tg kidney tissue (5 freeze thaw cycles) as previously published (27). Debris was removed by centrifugation. Lysate was administered to C57BL/6 or OT-I Tg mice by intraperitoneal injection (2 mg, IP) and primed CD8⁺ T cells were isolated from splenocytes of lysate-stimulated recipients on day 7 using negative selection magnetic beads as per the manufacturer’s recommendations (StemCell Technologies, Vancouver, Canada; purity routinely >90%). Primed CD8⁺ T cells were sorted into CXCR5⁺CD8⁺ (or CXCR5⁻CD8⁺) T cells by FACS Aria III flow cytometer (Becton Dickinson) using anti-CXCR5 monoclonal antibody (clone 2G8; Becton Dickinson). Flow-sorted CXCR5⁺CD8⁺ T cells were routinely >90% pure (Supplemental Figure 1).

Preparation and isolation of alloprimed CD25^highCD4⁺ T_reg cells.

Isolation of CD25^highCD4⁺ T_reg cells from in vivo primed (A/J lysate primed or OVA primed) WT or OT-II hosts was performed using negative selection for CD4⁺ T cells followed by positive selection for CD25^highCD4⁺ T_reg cells (>90% purity) by magnetic beads as per manufacturer recommendations (StemCell Technologies) (48, 49). In some experimental groups CD25^highCD4⁺ T_reg cells were isolated by flow cytometric sorting (purity routinely >90%). Data generated from both isolation techniques were combined. Purified CD25^highCD4⁺ T cells expressed FoxP3 (>70% by flow cytometry; Supplemental Figure 1).

Adoptive cell transfer.

Primed CXCR5⁺CD8⁺ T cells, CXCR5⁻CD8⁺ T cells, or CD25^highCD4⁺ T cells (2×10⁶) were resuspended in serum free RPMI and adoptively transferred into CCR5 KO KTx recipients via tail vein injection.

Cellular depletion.

Groups of C57BL/6 KTx recipients were treated with either anti-CD8 (clone 53.6.7; IgG2a) or anti-CD25 (clone PC-61.5.3; IgG1) monoclonal Ab (BioXCell, New Hampshire, USA). Recipients were depleted of CD8⁺ T cells by IP injection of 100 μg of anti-CD8 mAb (day −2,−1, +7) as previously described (29). Recipients were depleted of CD25⁺CD4⁺ T_regs by IP injection of 250μg of anti-CD25 mAb (day −7, +1, +7) (50). Depletion was confirmed through flow cytometric analysis of cells in the recipient spleen and the kidney allograft. Control recipient groups were treated with rat IgG2a and IgG1 isotype control antibodies (BioXCell).

Donor-reactive alloantibody titer.

Alloantibody titer from recipient sera was quantitated using published methods (37). In brief, serum was serially diluted and incubated with allogeneic target splenocytes. Splenocytes were stained with FITC-conjugated goat anti-mouse IgG Fc (Organon Teknika, Durham, NC) and the mean fluorescence intensity (MFI) was measured by flow cytometry. The dilution associated with the MFI observed when splenocytes were incubated with a 1:4 dilution of naïve serum was divided by two and recorded as the titer.

Anti-OVA antibody ELISA.

Anti-OVA antibody production in recipient mice with mOVA KTx was quantitated by ELISA (mouse anti-OVA IgG antibody assay) per manufacturers recommendations (Chondrex, Redmond, WA).

Creation of OVA tetramers for B cells.

We followed the protocol established Taylor et al. (51). Biotin-conjugated ovalbumin (AnaSpec, Fremont, CA) was incubated at a 1:6 ratio with streptavidin-PE or streptavidin-APC (30-minute incubation at room temperature). To quantify OVA-specific B cells by flow cytometry, B cells or splenocytes were incubated with 5 nM of each tetramer for 30 minutes. Both tetramers were utilized to minimize non-specific binding, as previously described (51, 52).

Flow cytometric analysis of cell subsets.

Splenocytes and mononuclear cells from peripheral blood and the kidney allograft were isolated from A/J or mOVA kidney allograft recipients on days 0, 7, and 14 posttransplant. Single cell suspensions were incubated for 4 hours with Leukocyte Activation Cocktail (PMA, ionomycin, and Brefeldin A; Becton Dickinson), followed by extracellular and intracellular (FIX&PERM cell permeabilization kit; Thermo Fisher, Waltham, MA) staining with antibodies for flow cytometry: CD8 (clone 53–6.7), CD44 (clone IM7), CXCR5 (clone 2G8), IFNγ (clone XMG1.2), CD4 (RM4–5), FoxP3 (R16–71S), CD25 (PC61), IL4 (11B11), IL10 (JES5–16E3), IL21 (mhalx21), PD1 (J43), B220 (clone RA3–6B2), GL-7 (clone GL-7), Fas (clone Jo2), and CD138 (clone 281–2). Fluorochrome-tagged CD4-tetramer I-A^b OVA_323–339 (MBL Bio; tetramer related to OT-II CD4⁺ T cells) was utilized to identify OVA-specific CD4⁺ T cells. Fluorescence-minus-one (FMO) was used as a negative control to set the positive/negative boundaries for marker expression (53). Samples were acquired on a Becton Dickinson LSRFortessa cytometer and data were analyzsed using FlowJo. Flow cytometric analysis was performed by gating on single cell, lymphocyte populations of CD8⁺ T, CD4⁺ T, and B cells. Germinal center (GC) B cells were identified by dual staining for Fas⁺GL-7⁺ B cells (54, 55) and T follicular helper (T_FH) cells were identified as CXCR5⁺PD1⁺CD4⁺ T cells (56–58).

Cell quantification of flow analysis.

To calculate the quantity of immune cells in various tissues, we quantified the numbers of cells per volume of tissue. We used a standard section of kidney (20 mm³) for cell isolation and flow cytometry. Tissues were excised from euthanized mice and mechanically homogenized and filtered (size 70μm) followed by red blood cell lysis. The cell suspension was washed and prepared for flow cytometric analysis. We used the percentage of immune cell subsets detected by flow cytometry and multiplied by the number of total immune cells in each tissue section to report the number of cells per mm³ tissue (for kidney and other tissues in a similar fashion).

In vitro cytotoxicity.

Detection of in vitro cytotoxicity of OVA or alloprimed GC B cells (from CCR5 KO mice) was measured using a LIVE/DEAD cell-mediated cytotoxicity kit (Invitrogen, Eugene, OR), similar to previously described methods (28). In brief, alloprimed CXCR5⁺CD8⁺ T cells were isolated from C57BL/6 or OT I Tg mice 7 days after in vivo stimulation with allogeneic lysate or OVA. Target OVA-primed Tet⁺GL-7⁺Fas⁺B220⁺ GC B cells or alloprimed GL-7⁺Fas⁺B220⁺ GC B cells (isolated from CCR5 KO mice 7 days after stimulation) were stained with CFSE^high. Naïve B cells were utilized as a control target cells. CXCR5⁺CD8⁺ T cells and GC B cells were co-cultured at a 10:1 ratio for 4 hours. Propidium iodide (PI) was added to the co-cultures to assess cell death, and PI uptake in CFSE⁺ B cells was immediately analyzed by flow cytometry.

ELISPOT.

Analysis of B cells for IgG production by ELISPOT was performed using the Mouse IgG ELISpotBASIC kit (Mabtech, Nacka Strand) according to manufacturer instructions. Plates were analyzed by computer-assisted image analysis using the CTL-Immunospot S6 Universal Analyzer and associated ImmunoSpot7 software. The data are reported as the number of relative Spot Forming Cells (SFC) per 1×10⁶ splenocytes.

Histology.

On day 7 or 14 posttransplant kidney allografts or mOVA kidney transplants were excised and sectioned into halves. One half section was fixed in 10% buffered formalin (6 hours), rehydrated in 20% sucrose (18 hours), and embedded in paraffin. For histological analysis, 5μm sections were mounted on glass slides and stained with H&E. Images were captured with an Aperio Scanscope XT (Leica).

Immunofluorescence.

The other half section of kidney allografts or mOVA kidney transplants was immediately snap frozen with liquid nitrogen and OCT medium (Tissue-Tek, Torrance, CA). Frozen sections were cut at 3μm and analyzed for C4d deposition by FITC-conjugated polyclonal anti-mouse C4d antibody (59). Stained sections were imaged by Olympus BX-43 microscope (confocal microscopy).

Pathological analysis.

Kidney allografts or mOVA kidney transplants were evaluated for AMR pathology, including microvascular inflammation/peritubular capillary (PTC) margination, arteritis, and PTC C4d deposition (37, 38). Following H&E and immunofluorescent C4d staining, blinded pathological analysis of the samples was performed and scored in accordance to internationally accepted Banff criteria (grades 0–3 for each category) (60, 61). Individual scores (for PTC margination, arteritis, and PTC C4d deposition) were summed to create a composite AMR score.

Statistical analysis.

Student’s t-tests were used to test differences in continuous outcomes between two experimental groups. When more than two experimental groups were assessed, continuous outcomes measured at one time point were compared using general linear models including experimental groups as an independent variable. Continuous outcomes measured at multiple time points were compared between relevant groups using general linear models including experimental groups, time, and their interaction as independent variables. As the measurements were not conducted on the same mice over time, measurements were assumed to be independent, and no additional correlation was considered. Assumptions of normally distributed residuals were assessed graphically using q-q plots and were not considered to be violated for any of the analyses. All analyses were conducted using SAS version 9.4 (SAS Institute, Inc., Cary, NC). Results are summarized as estimated mean ± standard error. Hypothesis testing was conducted at a 5% type I error rate (alpha=0.05) and p<0.05 was considered statistically significant.

RESULTS

High alloantibody-producing CCR5 KO kidney transplant recipients have significantly reduced quantity of splenic and graft-infiltrating CD25^highCD4⁺ T_regs compared to low alloantibody-producing C57BL/6 kidney transplant recipients.

In prior studies, A/J KTx results in high alloantibody production in CCR5 KO recipients and low alloantibody production in C57BL/6 (wild type, WT) recipients that is readily detectable on D7 and peaks D14 posttransplant (28). We now find, on day 7 posttransplant, there is a significant increase in the quantity of graft-infiltrating CD4⁺ T_regs in both WT (p<0.0001) and CCR5 KO recipients (p<0.0001). After day 7 their quantity increases in both groups, however CCR5 KO recipients have 4-fold fewer quantity of graft-infiltrating CD4⁺ T_regs compared to WT recipients by day 14 (p<0.0001; Figure 1A,B). Further, quantity of splenic CD4⁺ T_regs increases at day 7 in both recipient strains but CCR5 KO recipients have 1.6-fold fewer splenic CD4⁺ T_regs (p=0.007). We did not detect differences in quantity of splenic CD4⁺ T_regs at baseline or day 14 posttransplant) (Figure 1C). No differences were observed for splenic follicular regulatory CD4⁺ T cells (day 14; Supplemental Figure 2).

Figure 1. High alloantibody-producing CCR5 KO KTx recipients have significantly reduced quantity of graft-infiltrating and splenic CD25^highCD4⁺ T_regs cells compared to low alloantibody-producing WT recipients.

CCR5 KO and C57BL/6 (WT, H-2^b) mice were transplanted with allogeneic A/J (H-2^a) kidneys. On day 7 (D7) and 14 (D14), recipient spleens and kidney allografts were retrieved and processed to isolate mononuclear cells. Single cell suspensions were analysed for CD25^highCD4⁺ T cells (CD4⁺ T_regs) by flow cytometry. A) Flow cytometric gating on lymphocytes, single cells and CD25^highCD4⁺ T cells. Fluorescent minus one (FMO) was used as negative control. Representative flow panels of splenic and graft-infiltrating CD25^highCD4⁺ T_regs from D7 and D14 WT and CCR5 KO KTx recipients are shown. B) In the A/J donor kidney, a baseline quantity of CD4⁺ T_regs was detected prior to transplant (3.0±0.3 cells/mm³, n=3). By D7, both WT (204±19 cells/mm³, n=7) and CCR5 KO recipients (171±21 cells/mm³, n=9) have significantly increased quantities of graft-infiltrating CD4⁺ T_regs (*p<0.0001 for both). By day 14, transplant kidneys in WT recipients continue to have an increased quantity of graft-infiltrating CD4⁺ T_regs (144±5.2 cells/mm³, n=7; **p<0.0001) compared to naive A/J kidney (day 0). The quantity of D14 graft-infiltrating CD4⁺ T_regs in CCR5 KO recipients significantly decreases (33±5.5 cells/mm³, n=10) compared to D7 (***p=0.003) and is significantly reduced compared to D14 quantity in WT recipients (4-fold less; ^†p<0.0001). C) Naive CCR5 KO (n=3) and WT (n=4) mice have similar baseline quantity of splenic CD4⁺ T_regs (day 0; WT=28.5±6.0 vs CCR5 KO=40.3±5.8 cells/mm³; p=NS). Following transplant, the quantity of D7 splenic CD4⁺ T_regs significantly increases in WT (292±49, n=7; *p<0.0001) and also in CCR5 KO (178±34 cells/mm³, n=7; **p=0.0006) recipients. The quantity of splenic CD4⁺ T_regs was significantly less in CCR5 KO compared to WT KTx recipients on day 7 (***p=0.007) but was similar on D14 posttransplant (WT=236±21, n=7 vs CCR5 KO=152±35 cells/mm³, n=10; p=NS). Error bars designate standard error of combined data from independent experiments.

High alloantibody-producing CCR5 KO KTx recipients fail to expand peripheral blood and splenic CXCR5⁺CD8⁺ T cells posttransplant unlike low alloantibody-producing WT recipients.

There was no difference in the baseline quantity of CXCR5⁺CD8⁺ T cells detected in peripheral blood or spleen between naïve WT and CCR5 KO mice (Figure 2A–C). However, by day 7 posttransplant, the quantity of CXCR5⁺CD8⁺ T cells in WT recipients significantly increased (>10-fold) in the peripheral blood (p<0.0001) and (>4-fold) in the spleen (p<0.0001), and this increase is sustained in the spleen through D14 (p<0.01 for both). The majority (greater than 65%) of day 7 CXCR5⁺CD8⁺ T cells in both the peripheral blood and spleen of WT recipients were activated CXCR5⁺CD44⁺IFNγ⁺CD8⁺ T cells (Supplemental Figure 3).

Figure 2. High alloantibody-producing CCR5 KO KTx recipients fail to expand peripheral blood and splenic CXCR5⁺CD8⁺ T cells posttransplant unlike low alloantibody-producing WT recipients.

CCR5 KO and C57BL/6 (H-2^b) mice were transplanted with allogeneic A/J (H-2^a) kidneys. Both before (day 0) and after transplant (day 7 and 14), mononuclear cells were isolated from peripheral blood, spleens, and the A/J donor kidney. Cells were analyzed for quantity of CXCR5⁺CD8⁺ T cells by flow cytometry. A) Flow cytometric gating on lymphocytes, single cells, CD8⁺ T cells, and CXCR5 is shown. Fluorescent minus one (FMO) was used as a negative control. Representative flow panels of splenic CXCR5⁺CD8⁺ T cells from D7 WT and CCR5 KO KTx recipients compared to naïve (D0) counterparts are shown. B) On D7 and D14, a significant increase in peripheral blood CXCR5⁺CD8⁺ T cells was detected in WT recipients (D7: 117±21.8 cells/mm³, n=7 and D14: 61.7±16.9 cells/mm³, n=8) compared to naive controls (9.12±1.5 cells/mm³, n=4; *p<0.01 for both). In contrast, no significant change in the quantity of peripheral blood CXCR5⁺CD8⁺ T cells between D0 (10.9±1.3 cells/mm³, n=4), D7 (17.2±3.2 cells/mm³; n=5; p=NS), and D14 (24.1±7.7 cells/mm³; n=7; p=NS) was observed in CCR5 KO KTx recipients. Consequently, CCR5 KO recipients had significantly fewer peripheral blood CXCR5⁺CD8⁺ T cells compared to WT recipients on D7 and D14 (**p<0.03 for both). C) Compared to naive WT controls (711±135 cells/mm³, n=4) splenic CXCR5⁺CD8⁺ T cells significantly increased in WT KTx recipients on D7 posttransplant (2333±253 cells/mm³, n=6), and is sustained through D14 (1508±289 cells/mm³, n=9; *p<0.01 for both compared to naive). In contrast, no significant change in the quantity of splenic CXCR5⁺CD8⁺ T cells was observed between day 0 (445±62.0 cells/mm³, n=4) and D7 (644±51.6 cells/mm³, n=5; p=NS) through D14 (617±88.0 cells/mm³, n=8; p=NS) in CCR5 KO KTx recipients. Compared to WT, CCR5 KO KTx recipients have significantly fewer splenic CXCR5⁺CD8⁺ T cells on D7, which persists through D14 (**p<0.0009 for both). D) There is no significant change in the quantity of kidney allograft CXCR5⁺CD8⁺ T cells in either WT or CCR5 KO KTx recipients on D7 (WT: 55±29.2 vs KO: 33.8±6.5 cells/mm³, n=5; p=NS) through D14 (WT: 27.6±5.9 vs KO: 31.4± 2.9 cells/mm³, n=6–7; p=NS) compared to the quantity detected in the naive A/J donor kidney (38.4±19.8 cells/mm³, n=3). Error bars designate standard error of combined data from independent experiments.

In contrast, a comparable increase in quantity of CXCR5⁺CD8⁺ T cells in CCR5 KO recipient mice was not observed. Instead, similar quantities of CXCR5⁺CD8⁺ T cells were detected in CCR5 KO recipients on day 7 and day 14 in both peripheral blood and spleen that were unchanged from the quantity observed in naïve CCR5 KO mice (Figure 2A–C). There was no significant increase in the quantity of activated CXCR5⁺CD44⁺IFNγ⁺CD8⁺ T cells detected in either peripheral blood or spleen on day 7 or day 14 in CCR5 KO recipients (Supplemental Figure 3). Consequently, CCR5 KO KTx recipients had fewer splenic (and peripheral blood) CXCR5⁺CD8⁺ T cells and activated CXCR5⁺CD44⁺IFNγ⁺CD8⁺ T cells compared to WT recipients on day 7 (p<0.0001 for both), which persists through day 14 posttransplant (p<0.03 for both). Unlike the increase in graft-infiltrating CD4⁺ T_regs detected on day 7 after KTx in both WT and CCR5 KO recipient, the quantity of CXCR5⁺CD8⁺ T cells detected in the kidney allograft on day 7 and day 14 after KTx in both recipient strains was no different than the quantity detected in kidneys of naïve A/J mice (Figure 2D). These data indicate that while CD4⁺ T_regs traffic to the kidney, CXCR5⁺CD8⁺ T cells do not. CCR5-deficiency does not alter accumulation of CD4⁺ T_regs in the kidney allograft (day 7) but may affect the retention of CD4⁺ T_regs in the allograft (day 14; Figure 1B).

Collectively, we found that the most significant difference between high alloantibody producing CCR5 KO KTx recipients and low alloantibody producing WT KTx recipients was the failure in CCR5 KO recipients to expand CXCR5⁺CD8⁺ T cells in response to allostimulation (both in peripheral blood and spleen) and significantly fewer graft-infiltrating CD4⁺ T_regs observed on day 14.

Adoptive cell therapy with alloprimed antibody-suppressor CXCR5⁺CD8⁺ T cells (but not CD4⁺ T_regs) significantly inhibits posttransplant alloantibody production following kidney transplant.

We have recently reported that CCR5 KO kidney allograft recipients that underwent adoptive cell transfer (ACT) with alloprimed CXCR5⁺CD8⁺ T cells on D5 demonstrated significant dose-dependent and alloantigen-specific posttransplant suppression of alloantibody production (28). To investigate the optimal timing of cellular therapy for these studies we compared administration of alloprimed CXCR5⁺CD8⁺ T cells one day before transplant (D-1) or one to five days posttransplant (D1 or D5) in CCR5 KO mice. Pre-transplant ACT with alloprimed CXCR5⁺CD8⁺ T cells did not reduce alloantibody production after KTx, while both ACT D1 and D5 significantly inhibited alloantibody production compared to untreated CCR5 KO KTx recipients (D-1 and D5, not shown)(28, 62). Thus, ACT administered on D1 was pursued for the remainder of the studies.

Unlike the significant inhibition of alloantibody production observed after ACT with 2×10⁶ alloprimed CXCR5⁺CD8⁺ T cells into CCR5 KO KTx recipients (4.7-fold reduction; p<0.0001; Figure 3A), ACT with alloprimed WT CD25^highCD4⁺ T_regs (isolated from in vivo alloprimed mice) had no effect on posttransplant alloantibody production after KTx in CCR5 KO mice (p=NS, Figure 3A). ACT with CD25^highCD4⁺ T_reg from naïve mice also did not reduce alloantibody production in CCR5 KO recipient mice (6083±521 vs. 6012±549; p=NS, not shown). While CCR5 KO CXCR5⁺CD8⁺ T cells slightly inhibited alloantibody production (p=0.03; but less than WT CXCR5⁺CD8⁺ T cells, p<0.0001), ACT with other comparison CD8⁺ T cell subsets [including third-party (FVB/N) primed CXCR5⁺CD8⁺ T cells isolated from WT mice or alloprimed CXCR5⁻CD8⁺ T cells isolated from WT mice] did not significantly inhibit alloantibody production in CCR5 KO recipients consistent with prior results (28). Likewise, ACT with CD4⁺ T_regs isolated from alloprimed CCR5 KO mice did not significantly inhibit alloantibody production in CCR5 KO recipients (Figure 3A). We next quantified the number of CXCR5⁺CD8⁺ T cells and CD25^highCD4⁺ T_reg cells in the spleen and kidney allograft following ACT (Figure 3B–G). ACT of 2×10⁶ CXCR5⁺CD8⁺ T cells increased the quantity of CXCR5⁺CD8⁺ T cells detected in the spleen (D14, p<0.0001; Figure 3B,D), but did not impact the quantity detected in the kidney allograft (Figure 3B,E). ACT of 2×10⁶ CD25^highCD4⁺ T_regs into CCR5 KO KTx recipients increased the quantity of CD25^highCD4⁺ T_regs detected in the spleen (p<0.0001; Figure 3C,F) and kidney allograft (p<0.0001; Figure 3C,G). Of note, CD25^highCD4⁺ T_reg cell ACT was associated with an increase in the quantity of endogenous splenic CXCR5⁺CD8⁺ T cells compared to untreated controls (p=0.06; Figure 3D). To clarify if the observed increase in quantity of splenic CXCR5⁺CD8⁺ T cells was primarily due to adoptively transferred cells, we performed ACT studies using GFP⁺CXCR5⁺CD8⁺ T cells or GFP⁺CD25^highCD4⁺ T_regs. Results revealed that exogenously transferred GFP⁺CXCR5⁺CD8⁺ T cells comprised the dominant source of increased CXCR5⁺CD8⁺ T cell quantity detected in the spleen (Supplemental Figure 4). Furthermore, adoptively transferred CD8⁺ T cells maintain CXCR5 expression. Similarly, exogenously transferred CD25^highCD4⁺ T_reg cells are the dominant source of increased quantity of CD25^highCD4⁺ T_regs detected in the kidney allograft. Of note, exogenously transferred CD25^highCD4⁺ T_regs have reduced CD25 expression 2 weeks posttransplant.

Figure 3. Adoptive cell therapy with alloprimed antibody-suppressor CXCR5⁺CD8⁺ T cells (but not CD4⁺ T_regs), significantly inhibits alloantibody production following kidney transplant.

CCR5 KO (H-2^b) mice were transplanted with allogeneic A/J (H-2^a) kidneys. CCR5 KO KTx recipient groups underwent adoptive cell transfer (ACT) with either 2×10⁶ alloprimed CXCR5⁺CD8⁺ T_Ab-supp cells or CD4⁺ T_regs via tail vein injection on D1. Tissues were harvested on D14 and single cell suspensions analyzed for quantity of CXCR5⁺CD8⁺ T cells or CD25⁺FoxP3⁺CD4⁺ T_reg cells. A) On D14, sera were collected and analyzed for alloantibody titer. CCR5 KO KTx recipients produced significantly lower alloantibody titer following ACT with alloprimed CXCR5⁺CD8⁺ T cell (1259±436, n=6; *p<0.0001) compared to untreated controls (5951±404, n=7). ACT with CD4⁺ T_reg did not impact alloantibody titer (5688±436, n=6; p=NS) compared to untreated controls. Alloantibody titer in the group that received ACT with CD4⁺ T_regs was significantly higher than the group that received ACT with CXCR5⁺CD8⁺ T cells (*p<0.0001). ACT with alloprimed CCR5-deficient CXCR5⁺CD8⁺ T cells reduced alloantibody titer (4215±95, n=4, **p=0.03) but to a lesser extent than ACT with WT CXCR5⁺CD8⁺ T cells (*p<0.0001). ACT with alloprimed CCR5-deficient T_reg (4877±168, n=4), third party-primed CXCR5⁺CD8⁺ T cells (5060±372, n=5), or alloprimed CXCR5⁻CD8⁺ T cells (5105±469, n=5) did not reduce alloantibody titer (p=NS for all). Dashed line represents naïve control sera. B,C) Representative flow cytometric analysis of splenic and graft-infiltrating CXCR5⁺CD8⁺ T cells (gated on single cell lymphocytes) and CD25⁺FoxP3⁺CD4⁺ T_regs (gated on single cell CD4⁺ T cells) are shown, comparing cell quantities in untreated control CCR5 KO KTx recipients (CCR5 KO) compared to CCR5 KO KTx recipients that received adoptive cell transfer of 2×10⁶ CXCR5⁺CD8⁺ T cells (CCR5 KO + CD8 ACT) or CD25⁺CD4⁺ T_regs (CCR5 KO + T_reg ACT). Gating on single cells and lymphocytes (for CD8⁺ T cells, B) or CD4⁺ T cells (for T_reg cells, C) was based on gating in Figure 1A and Figure 2A. Fluorescent minus one (FMO) was used as negative controls. D) CCR5 KO KTx recipients that received CXCR5⁺CD8⁺ T cell ACT had significantly increased quantity of splenic CXCR5⁺CD8⁺ T cells (2961±317 cells/mm³, n=11) 14 days posttransplant, compared to untreated controls (593±102 cells/mm³, n=8; *p<0.0001). CD4⁺ T_reg ACT was associated with an increase in the quantity of endogenous splenic CXCR5⁺CD8⁺ T cells (1259±146, n=10; **p=0.06). E) CCR5 KO KTx recipients that underwent CXCR5⁺CD8⁺ T cell ACT demonstrated no difference in the quantity of graft-infiltrating CXCR5⁺CD8⁺ T cells (41.0±9.0 cells/mm³, n=9) 14 days posttransplant, compared to untreated controls (31.4±2.9 cells/mm³, n=7) or when compared to recipients that received ACT with CD4⁺ T_regs (38.6±4.7 cells/mm³, n=9; p=NS for both). F) ACT of CD4⁺ T_reg cells into CCR5 KO KTx recipients significantly enhanced the quantity of splenic CD4⁺ T_reg cells (1197±301 cells/mm³, n=6) 14 days posttransplant compared to untreated controls (168±34.7 cells/mm³, n=10; * p<0.0001) and compared to recipients that received ACT with CXCR5⁺CD8⁺ T cells (112±16.3 cells/mm³, n=6; **p<0.0001). G) Similarly, ACT with CD4⁺ T_regs, was associated with significantly increased quantity of CD25^highFoxP3⁺CD4⁺ T_regs (167±27.8 cells/mm³, n=6) in the kidney allograft 14 days posttransplant compared to untreated controls (33.5±5.5 cells/mm³, n=10; * p<0.0001) and compared to recipients that received ACT with CXCR5⁺CD8⁺ T cells (21.5±3.3 cells/mm³, n=9; ** p<0.0001). CXCR5⁺CD8⁺ T cell ACT did not impact the quantity of graft-infiltrating CD4⁺ T_regs (p=NS). Error bars designate standard error of combined data from independent experiments.

High alloantibody-producing CCR5 KO KTx recipients have an increased quantity of splenic germinal center B cells, IgG-producing B cells, plasma cells, and CD4⁺ T_FH cells compared to low alloantibody-producing WT recipients.

Considering the marked enhanced alloantibody titer detected posttransplant in CCR5 KO recipients compared to WT controls, we expected a corresponding difference in the quantity of germinal center (GC) B cells and T_FH cells. A significant increase in splenic GC B cell quantity from baseline was detected in CCR5 KO recipients on D7 (p<0.0001), which was sustained through D14 (p<0.0001; Figure 4A,B). WT KTx recipients manifested an increase in splenic GC B cell quantity D7 posttransplant (p=0.06; Figure 4A,B). By D14, CCR5 KO recipients had 8.7-fold more splenic GC B cells than WT counterparts (p<0.0001). Naïve CCR5 KO and WT mice have comparable baseline quantities of splenic GC B (GL-7⁺Fas⁺B220⁺) cells (Figure 4B). Similarly, ELISPOT analysis shows a 3-fold higher proportion of antibody-producing B cells relative to total B cells in D7 CCR5 KO recipients compared to WT counterparts (p=0.003; Figure 4C). These differences in B cell populations could not be attributed to strain-associated differences in total splenic B cell quantities (Supplemental Figure 5A). Likewise, CCR5 KO recipients had significantly greater quantities of splenic plasma cells (CD138⁺B220⁻) compared to WT recipients (p<0.0001; Figure 4D), consistent with previous reports (28).

Figure 4. High alloantibody-producing CCR5 KO KTx recipients have an increased quantity of splenic germinal center B cells, IgG-producing B cells, plasma cells, and CD4⁺ T_FH cells compared to low alloantibody-producing WT recipients.

CCR5 KO and C57BL/6 (H-2^b) mice were transplanted with allogeneic A/J (H-2^a) kidneys. On day 7 and 14, recipient spleens and kidney allografts were retrieved and processed for mononuclear cells. Single cell suspensions were analyzed for germinal center (GC) B (GC; GL-7⁺Fas⁺B220⁺) and IL-21⁺CD4⁺ T_FH (IL-21⁺CXCR5⁺PD1⁺CD4⁺) cells by flow cytometry. A) Representative flow cytometric gating on lymphocytes, single cells, B220⁺, GL-7⁺ and Fas⁺ cells in posttransplant WT and CCR5 KO KTx recipients are shown for GC B cells. Fluorescent minus one (FMO) was used as a negative control. B) Naive CCR5 KO (n=3) and WT (n=4) mice have similar baseline quantity of splenic GC B cells (day 0; WT=1017±258 vs CCR5 KO=423±87.5 cells/mm³; p=NS). By D7 posttransplant, CCR5 KO recipients had significantly increased quantity of GC B cells compared to baseline (*p<0.0001) and compared to WT counterparts (WT=2287±367, n=5 vs CCR5 KO=4733±410 cells/mm³, n=5; **p=0.0007); this difference was more pronounced on D14 (WT=1105±221 cells/mm³, n=8 vs CCR5 KO=5081±589 cells/mm³, n=8; ***p<0.0001). The GC B cell quantity in D7 WT recipients increased (^†p=0.06) and returned to pre-transplant levels by D14 (p=NS). C) By D7 posttransplant, CCR5 KO KTx recipients have more IgG-producing B cells (38.0±8.3 per 10k total B cells) relative to WT counterparts (17.1±3.6 per 10k total B cells, n=9 for both; *p=0.003) by ELISPOT analysis. D) CCR5 KO recipients have more plasma cells compared to WT counterparts on D14 (1083±113 vs 464±69.8 cells/mm³; *p<0.0001). E) Representative flow cytometric gating on lymphocytes, single cells, CD4⁺, CXCR5⁺, PD1⁺, and IL-21⁺ shown for T_FH cells. F) By D7, CCR5 KO recipients had significantly increased quantity of splenic IL-21⁺CD4⁺ T_FH cells (355±61.9 cells/mm³, n=5) compared to baseline quantities in the naive control (day 0, 69.7±10.0 cells/mm³, n=3), which was sustained through D14 (451±69.3 cells/mm³, n=8; *p<0.002 for both) and significantly elevated compared to the quantity of IL-21⁺CD4⁺ T_FH cells in WT counterparts on both D7 (119±41.3 cells/mm³, n=5) and D14 (61.5±14.9 cells/mm³, n=8; **p<0.006 for both). In contrast, WT recipients had no significant change in the quantity of IL-21⁺CD4⁺ T_FH cells detected in the spleen on D7 or D14 (compared to baseline; 46.3±37.8 cells/mm³, n=4; p=NS). Error bars designate standard error of combined data from independent experiments.

Next, we evaluated the quantity of CD4⁺ T_FH cells in CCR5 KO and WT KTx recipients. We observed a significant increase in the quantity of splenic IL-21⁺CD4⁺ T_FH (IL-21⁺CXCR5⁺PD1⁺CD4⁺) cells in CCR5 KO recipients on D7 and D14 compared to baseline levels (p<0.002 for both; Figure 4E,F). In contrast, no significant change in IL-21⁺CD4⁺ T_FH cell quantity was observed in WT recipients posttransplant (Figure 4F). By D14, CCR5 KO recipients had 7.2-fold more splenic IL-21⁺CD4⁺ T_FH cells than WT counterparts (p<0.006). Of note, CCR5 KO mice have a significantly greater pretransplant quantity of CD4⁺ T_FH cells (p<0.006) compared to WT counterparts that is sustained on both day 7 and 14 posttransplant (p<0.006 for both D7 and D14) (Supplemental Figure 5C).

ACT with alloprimed CXCR5⁺CD8⁺ T cells (but not with CD4⁺ T_regs) significantly downregulates posttransplant expansion of GC B cells and IL-21⁺CD4⁺ T_FH cells in CCR5 KO KTx recipients.

ACT with 2×10⁶ alloprimed CXCR5⁺CD8⁺ T cells into CCR5 KO recipients resulted in a significant reduction in the quantity of splenic GC B cells (D14, p=0.02; Figure 5A,B), without impacting total B cell populations (data not shown). Similarly, CXCR5⁺CD8⁺ T cell ACT resulted in a ~5-fold reduction in the quantity of IL-21⁺CD4⁺ T_FH cells (p=0.0002; Figure 5C) and total CD4⁺ T_FH (p=0.03), without impacting total CD4⁺ T cell population (data not shown). In contrast, alloprimed CD4⁺ T_reg ACT into CCR5 KO recipients did not significantly impact either GC B cell quantity (Figure 5B) or IL-21⁺CD4⁺ T_FH cell quantity (Figure 5C) when compared to untreated controls. Consequently, quantities of both GC B cells and IL-21⁺CD4⁺ T_FH cells were significantly higher in CCR5 KO recipients that received CD4⁺ T_reg ACT compared to CXCR5⁺CD8⁺ T cell ACT (p<0.004 for both; Figure 5B, 5C).

Figure 5. ACT with alloprimed CXCR5⁺CD8⁺ T cells but not with CD4⁺ T_regs significantly downregulates posttransplant expansion of GC B cells and IL-21⁺CD4⁺ T_FH cells and both ACTs improve AMR pathology in CCR5 KO KTx recipients.

CCR5 KO (H-2^b) mice were transplanted with allogeneic A/J (H-2^a) kidneys. Groups underwent ACT of either 2×10⁶ alloprimed CXCR5⁺CD8⁺ T_Ab-supp cells or CD4⁺ T_regs via tail vein injection on D1. Tissues were harvested on D14 and single cell suspensions analyzed for quantity of GC B and IL-21⁺CD4⁺ T_FH cells. A) Representative flow cytometric gating on GC B cells (GL-7⁺Fas⁺) and IL-21⁺CD4⁺ T_FH (IL-21⁺CXCR5⁺PD1⁺CD4⁺) cells are shown, comparing cell quantities in untreated control CCR5 KO KTx recipient mice (CCR5 KO) to CCR5 KO KTx recipient mice that received adoptive cell transfer of 2×10⁶ alloprimed CXCR5⁺CD8⁺ T cells (CCR5 KO + CD8 ACT) or CD25⁺CD4⁺ T_regs (CCR5 KO + Treg ACT). Gating on single cells and lymphocytes as well as B220 (for GC B cells) or CD4⁺ T cells (for T_FH cells) was based on gating in Figure 4A, Figure 4E, and respective fluorescent minus one controls. B) CCR5 KO KTx recipients that underwent ACT of CXCR5⁺CD8⁺ T cells had significantly fewer splenic GC B cells (3435±530 cells/mm³, n=7) on D14 compared to untreated controls (5581±446 cells/mm³, n=8; *p=0.02) and compared to recipients that received CD4⁺ T_reg ACT (**p=0.03). CD4⁺ T_reg ACT did not significantly impact GC B cell quantity compared to untreated controls (5423±664 cells/mm³, n=6; p=NS). C) CCR5 KO KTx recipients that underwent ACT of CXCR5⁺CD8⁺ T cells had significantly fewer IL-21⁺CD4⁺ T_FH cells (75.1±18.1 cells/mm³, n=9) compared to both untreated controls (451±69.3 cells/mm³, n=8; *p=0.0002) and recipients that underwent CD4⁺ T_reg ACT (357±74.3 cells/mm³, n=9; **p=0.002). CD4⁺ T_reg ACT did not significantly impact the quantity of IL-21⁺CD4⁺ T_FH cell quantity compared to untreated controls (p=NS). D) Allografts from recipient mice were scored according to Banff criteria, with a composite and individual parameter AMR pathology scores for peritubular capillary (PTC) margination, arteritis, and C4d deposition. Composite AMR score in recipients that received ACT with 2×10⁶ CXCR5⁺CD8⁺ T cells (3.8±0.3, n=10; *p<0.0001) or CD25⁺CD4⁺ T_regs (5.3±0.2, n=6; **p=0.005) was significantly reduced, indicating substantial amelioration of AMR pathology compared to untreated CCR5 KO recipients (7.5±0.3, n=7). Composite AMR pathology score was lower in recipients that received ACT with CXCR5⁺CD8⁺ T cells compared to those that received ACT with CD4⁺ T_regs (***p=0.03). Error bars designate standard error of combined data from independent experiments.

ACT with alloprimed antibody-suppressor CXCR5⁺CD8⁺ T cells ameliorates day 14 AMR pathology to a greater degree than ACT with CD4⁺ T_regs.

We previously reported that ACT with CXCR5⁺CD8⁺ T cells, as late as D5 posttransplant when humoral alloimmune responses have been activated, significantly reduces D14 alloantibody titer and improves day 14 AMR pathology in CCR5 KO KTx recipients (28). For these studies in which ACT with CXCR5⁺CD8⁺ T cell occurred on D1 posttransplant, we also found significantly improved D14 composite and individual parameter AMR pathology scores compared to untreated controls (p<0.005 for all; Figure 5D). Following the ACT of control alloprimed CXCR5⁻CD8⁺ T cells or third party-primed CXCR5⁺CD8⁺ T cells, no amelioration of AMR pathology was observed [not shown and as previously described (28) respectively]. While kidney allografts in CCR5 KO recipients that underwent ACT with an equivalent quantity of CD4⁺ T_regs also had a significantly lower composite AMR score (p=0.005), individual parameter scores for PTC margination and PTC C4d deposition were not significantly reduced. The improvement in composite score associated with CD4⁺ T_reg cell ACT was unexpected given the high level of alloantibody in these recipients and suggests that its beneficial impact on AMR score may occur by a different mechanism. Collectively, these data indicate that ACT of WT alloprimed CXCR5⁺CD8⁺ T cells into CCR5 KO KTx recipients restores a deficit in endogenous circulating and splenic CXCR5⁺CD8⁺ T cells and results in significant reduction of posttransplant alloantibody titer, antibody-producing GC B cells, IL-21⁺CD4⁺ T_FH cells, and significantly improves kidney allograft pathology scores. ACT with CD4⁺ T_reg enhances accumulation of graft-infiltrating CD4⁺ T_reg and also improves kidney allograft composite pathology score.

Adoptive cell therapy with primed OVA-specific antibody-suppressor CXCR5⁺CD8⁺ OT-I cells (but not primed OVA-specific OT-II CD4⁺ T_regs) significantly inhibits posttransplant antibody production following mOVA kidney transplant.

In preceding studies, we investigated the outcome of ACT with polyclonal alloprimed CXCR5⁺CD8⁺ T cells and CD25^highCD4⁺ T_regs cell subsets upon posttransplant humoral alloimmunity. To investigate the outcome of ACT with antigen specific CXCR5⁺CD8⁺ T cell (OT-I) and CD25^highCD4⁺ T_reg cell (OT-II) subsets upon humoral immunity, we utilized a model of mOVA KTx into CCR5 KO recipient mice. Following mOVA KTx to CCR5 KO mice, high titer anti-OVA antibody was produced (Figure 6A). Anti-OVA antibody production was significantly inhibited (5.5-fold reduction; p=0.0001) in CCR5 KO recipients that received mOVA KTx and ACT with 2×10⁶ OVA-primed OT-I CXCR5⁺CD8⁺ T cells. In contrast, ACT with OVA-primed OT-II CD25^highCD4⁺ T_regs did not inhibit anti-OVA IgG antibody production. As observed with the polyclonal responses, ACT with OVA-primed OT-I CXCR5⁺CD8⁺ T cells resulted in significantly reduced quantity of splenic, OVA-tetramer positive (Tet⁺) GC B cells and Tet⁺ IL21⁺T_FH cells (p<0.009 for both) while ACT with OVA-primed OT-II CD25^highCD4⁺ T_regs did not (Figure 6B–D). Furthermore, OVA-specific OT-I CXCR5⁺CD8⁺ T cells mediate specific in vitro cytotoxicity towards Tet⁺ GC B cells (but not alloprimed GC B cells) (p<0.0001; Figure 6E). In contrast, alloprimed WT CXCR5⁺CD8⁺ T cells mediate cytotoxic killing of alloprimed GC B cells (p<0.0001) but not Tet⁺ GC B cells. These data demonstrate the in vivo and in vitro antigen specific effector function of antibody-suppressor CXCR5⁺CD8⁺ T cells.

Figure 6. Adoptive cellular therapy with OVA-primed OT-I CXCR5⁺CD8⁺ T cells suppresses anti-OVA IgG production, quantity of GC B cell, and CD4 T_FH after mOVA KTx in CCR5 KO recipients.

CCR5 KO (H-2^b) mice were transplanted with kidneys from mOVA transgenic donor mice (H-2^b). Groups of CCR5 KO KTx recipients underwent adoptive cell transfer (ACT) with 2×10⁶ OVA-primed OT-I CXCR5⁺CD8⁺ T cells or OVA-primed OT-II CD25^highCD4⁺ T_reg on day 1 posttransplant. A) On day 14, sera was collected from recipient mice and analyzed for anti-OVA IgG antibody titer by ELISA. ACT with OVA-primed OT-I CXCR5⁺CD8⁺ T cells significantly reduced anti-OVA IgG antibody production after KTx (1.6±0.4 mg/mL, n=4; *p=0.0001) compared to untreated CCR5 KO KTx recipients (8.8±0.8 mg/mL; n=4; *p=0.0001) and after ACT with OVA-primed OT-II CD25^highCD4⁺ T_regs (7.1±0.7 mg/mL, n=4; **p=0.0006). OVA-primed OT-II CD25^highCD4⁺ T_regs did not suppress anti-OVA IgG antibody production compared to untreated CCR5 KO KTx recipients (p=NS). B) Splenocytes were isolated on day 14 to analyze GC B cell and CD4⁺ T_FH cell quantity. Representative flow cytometric gating on OVA tetramer-positive GC B cells (Tet⁺GL-7⁺Fas⁺) and OVA tetramer-positive IL-21⁺CD4⁺ T_FH (Tet⁺IL-21⁺CXCR5⁺PD1⁺CD4⁺) cells are shown, comparing cell quantities in untreated control CCR5 KO KTx recipient mice (CCR5 KO) to CCR5 KO KTx recipient mice that received adoptive cell transfer of 2×10⁶ OVA-primed OT-I CXCR5⁺CD8⁺ T cells (CCR5 KO + OT-I CXCR5⁺CD8) or OVA-primed OT-II CD25^highCD4⁺ T_regs (CCR5 KO + OT-II CD4⁺T_reg). Gating on single cells and lymphocytes as well as B220 (for GC B cells) or CD4⁺ T cells (for T_FH cells) is based on gating in Figure 4A, Figure 4E, and respective fluorescent minus one (FMO) controls. OVA-tetramer positive GC B cells were detected with dual staining using OVA-FITC and OVA-APC with respective FMO controls. OVA-specific CD4⁺ T_FH cells were detected with I-A OVA 323–339 tetramer staining with FMO control. C) CCR5 KO KTx recipients that underwent ACT with OVA-primed OT-I CXCR5⁺CD8⁺ T cells had significantly fewer splenic Tet⁺GC B cells (36±3 cells/mm³, n=4) on D14 compared to untreated controls (496±38 cells/mm³, n=4; *p<0.0001) and compared to recipients that received OVA-primed OT-II CD25^highCD4⁺ T_regs (424±26 cells/mm³, n=4; **p<0.0001). ACT with OVA-primed OT-II CD4⁺ T_reg did not significantly impact GC B cell quantity compared to untreated controls (p=NS). D) CCR5 KO KTx recipients that underwent ACT with OVA-primed OT-I CXCR5⁺CD8⁺ T cells had significantly fewer Tet⁺IL-21⁺CD4⁺ T_FH cells (75±7 cells/mm³, n=4) compared to both untreated controls (377±55 cells/mm³, n=4; *p=0.009) and recipients that underwent ACT with OVA-primed OT-II CD25^highCD4⁺ T_regs (582±97 cells/mm³, n=4; **p=0.0004). ACT with OVA-primed OT-II CD4⁺ T_reg did not significantly impact the quantity of Tet⁺IL-21⁺CD4⁺ T_FH cell quantity compared to untreated controls (p=NS). E) In an in vitro cytotoxicity assay, OVA-primed OT-I CXCR5⁺CD8⁺ T cells, OVA-primed wild-type (WT) CXCR5⁺CD8⁺ T cells, and A/J-alloprimed CXCR5⁺CD8⁺ T cells were co-cultured with naive B cells (B220⁺), flow-sorted OVA-primed OVA-tetramer positive GC B cells (Tet⁺GL-7⁺Fas⁺B220⁺), or A/J-alloprimed GC B cells (GL-7⁺Fas⁺B220⁺). Primed CXCR5⁺CD8⁺ T cells and B cells were co-cultured at a 10:1 ratio for 4 hours and analyze for cytotoxicity (propidium iodide uptake). OVA-primed OT-I CXCR5⁺CD8⁺ T cells mediated significantly higher in vitro cytotoxicity towards OVA-primed Tet⁺GC B cells (8.6±1.1%; n=4) compared to A/J-primed GC B cells (1.9±0.2%; n=4) or naive B cells (1.2±0.1%; n=3; *p<0.0001 for both). Similarly, OVA-primed WT CXCR5⁺CD8⁺ T cells mediated significantly higher in vitro cytotoxicity towards OVA-primed Tet⁺GC B cells (6.8±0.6%; n=4) compared to A/J-alloprimed GC B cells (1.7±0.1%; n=4) or naive B cells (1.2±0.1%; n=3; **p<0.0009 for both). In contrast, A/J-alloprimed CXCR5⁺CD8⁺ T cells mediated significantly higher in vitro cytotoxicity towards A/J-alloprimed GC B cells (7.4±0.2%; n=4) compared to OVA-primed Tet⁺GC B cells (2.8±0.4%; n=4) or naive B cells (0.5±0.1%; n=3; ***p<0.0001 for both). Error bars designate standard error of combined data from independent experiments.

Dysregulation of humoral alloimmunity and severity of AMR pathology in WT KTx recipients is greater after CD8⁺ T cell depletion than after CD4⁺ T_reg depletion.

To compare the effect of CD8⁺ T cell depletion with CD4⁺ T_reg depletion upon humoral alloimmunity after KTx in WT C57BL/6 mice, we tested the impact of anti-CD8 and anti-CD25 mAb treatment upon alloantibody titer, quantity of GC B cells, and IL-21⁺CD4⁺ T_FH cells. We first confirmed that administration of anti-CD8 mAb significantly depletes splenic CXCR5⁺CD8⁺ T cells and that anti-CD25 mAb treatment depletes graft-infiltrating (and splenic; not shown) CD4⁺ T_reg cells compared to recipient mice untreated or treated with isotype controls (p<0.002 for CXCR5⁺CD8⁺ T cells; p<0.0001 for CD4⁺ T_reg cells; Figure 7A–C). We found that respective cell subset depletion persists through D14 at the time of KTx pathology examination. CD8⁺ T cell depletion did not impact the quantity of graft-infiltrating CD4⁺ T_regs (Figure 7A,C). Likewise, CD25⁺ T cell depletion did not impact the quantity of splenic CXCR5⁺CD8⁺ T cells compared to untreated and isotype treated controls (Figure 7A,B). The reduction of splenic CXCR5⁺CD8⁺ T cells in CD8-depleted WT KTx mice correlated with significantly higher alloantibody titer compared to untreated and isotype treated controls (p<0.0001 for both; Figure 7D). The depletion of CD4⁺ T_regs was also associated with increased alloantibody titer compared to untreated and isotype treated controls (p<0.01 for both). However, the magnitude of increase in alloantibody titer was greater following CD8-depletion than after CD4⁺ T_reg depletion (p<0.0001) (Figure 7D).

Figure 7. Alloantibody titer after kidney transplant in WT recipients is higher after CD8⁺ T cell-depletion than after CD4⁺ T_reg depletion.

C57BL/6 (WT, H-2^b) mice were transplanted with allogeneic A/J kidneys (H-2^a). A group of recipients underwent CD8⁺ T cell-depletion pre-transplant day-1 and on D7 via I.P. injection of 200ul anti-CD8 or isotype control mAb. Another group of recipients underwent CD4⁺ T_reg depletion pre-transplant d-7 and on D1 and 7 via I.P. injection of 250ul anti-CD25 or isotype control mAb. Tissues were harvested on D14 and mononuclear cells analyzed for efficacy of cell depletion. A) Representative flow cytometric gating on CXCR5⁺CD8⁺ T cells and CD25^highFoxP3⁺CD4⁺ T_regs are shown, comparing cell quantities in untreated or isotype control treated mice to mice that underwent CD8⁺ T cell or CD4⁺ T_reg-depletion, respectively. Gating on single cells and lymphocytes (for CD8⁺ T cells) or CD4⁺ T cells (for T_reg cells) is based on gating in Figure 1A, Figure 2A, and respective fluorescent minus one controls. B) CD8⁺ T cell-depletion resulted in significant reduction in quantity of splenic CXCR5⁺CD8⁺ T cells (222±63.4, n=4) compared to the quantity in untreated WT recipients (2058±239, n=6) and isotype (IgG2a) control treated recipients (2095±223, n=4) that persisted through D14 (*p<0.002). CD4⁺ T_reg depletion (2693±624, n=4) and isotype (IgG1) control treated recipients (2385±381, n=4) did not impact the quantity of endogenous CXCR5⁺CD8⁺ T cell population (p=NS for both) compared to untreated WT recipients (2693±624, n=4). C) CD4⁺ T_reg-depletion significantly reduced the quantity of CD4⁺ T_regs (7.3±2.0 cells/mm³, n=5) detected in the allograft on D14 compared to untreated controls (144±5.2 cells/mm³) and isotype (IgG2a) control treated recipients (117±10.0, n=4; *p<0.0001 for both), while CD8-depletion had no effect on the quantity of graft-infiltrating CD4⁺ T_regs (117±9.9 cells/mm³, n=4; p=NS) compared to untreated controls and isotype control treated recipients (IgG2a, 124±16.2, n=4). D) On D14, sera were harvested and analyzed for alloantibody titer. WT KTx recipients produced significantly higher alloantibody titer following CD8⁺ T cell-depletion (6078±627, n=6) compared to untreated (1508±145, n=10) and isotype (IgG2a) control treated WT recipients (1536±117, n=4; *p<0.0001 for both). CD4⁺ T_reg depletion also resulted in significantly increased alloantibody titer (3180±195, n=6) compared to untreated and isotype (IgG1) control treated WT recipients (1582±212, n=4; **p<0.01 for both). However, alloantibody titer in CD8-depleted KTx recipients was significantly higher than in CD4⁺ T_reg-depleted KTx recipients (***p<0.0001). Dashed line represents naïve control sera. Error bars designate standard error of combined data from independent experiments.

In CD8-depleted WT KTx recipients, there was a significant increase in the quantity of splenic GC B cells (p<0.0001; Figure 8A,B) and an increased quantity of IL-21⁺CD4⁺ T_FH cells (p<0.03; Figure 8A,C) corresponding to increases in alloantibody titer. In CD4⁺ T_regs depleted WT KTx recipients, there was an increase in the quantity of GC B cells (p<0.0001; Figure 8B) and IL-21⁺CD4⁺ T_FH cells (p<0.009; Figure 8C) compared to untreated and isotype treated controls. However, the magnitude of increase in quantity of GC B cells was higher following CD8-depletion than after CD4⁺ T_reg depletion (p<0.0001; Figure 8C).

Figure 8. CD8⁺ T cell-depletion after kidney transplant in WT recipients results in a more significant expansion of GC B cells and more severe AMR pathology than after CD4⁺ T_reg cell depletion.

C57BL/6 (WT, H-2^b) mice were transplanted with allogeneic A/J kidneys (H-2^a). A group of recipients underwent CD8⁺ T cell depletion pre-transplant day-1 and on D7 via I.P. injection of 200μg anti-CD8 or isotype (IgG2a) control mAb. Another group of recipients underwent CD4⁺ T_reg depletion pre-transplant d-7 and on D1 and 7 via I.P. injection of 250μg anti-CD25 or isotype (IgG1) control mAb. Kidney allograft and spleen tissue were retrieved 14 days after transplant for analysis. A) Representative flow cytometric gating on GC B cells (GL-7⁺Fas⁺) and IL-21⁺CD4⁺ T_FH (IL-21⁺CXCR5⁺PD1⁺CD4⁺) cells are shown, comparing cell quantities in untreated control WT KTx recipients to WT KTx recipients that underwent CD8⁺ T cell or CD4⁺ T_reg depletion, respectively. Gating on single cells and lymphocytes as well as B220 (for GC B cells) or CD4⁺ T cells (for T_FH cells) is based on gating in Figure 4A, Figure 4E, and respective fluorescent minus one controls. B) Compared to untreated WT recipients (1105±221, n=8) and isotype (IgG2a) control treated recipients (628±121, n=4), CD8⁺ T cell depletion resulted in a significant increase of splenic GC B cell quantity (7761±431, n=4; *p<0.0001 for both). CD4⁺ T_reg depletion resulted in a significant increase of splenic GC B cells quantity (4911±429, n=4) compared to untreated WT recipients and isotype (IgG1) control treated recipients (1486±342, n=4; **p<0.0001 for both) but not as great an increase in splenic GC B cells compared to CD8⁺ T cell depletion (***p<0.0001). C) CD8-depletion increased the quantity of splenic IL-21⁺CD4⁺ T_FH (163±62.9, n=5) compared to untreated controls (61.5±14.9, n=8) and isotype (IgG2a) control treated recipients (49.9±7.2, n=4; *p<0.03 for both). CD4⁺ T_reg depletion resulted in significantly increased splenic IL-21⁺CD4⁺ T_FH cell quantity (210±38.0, n=5) compared to untreated and isotype (IgG1) control treated recipients (74±27.1, n=4; **p<0.009; for both). D) Kidney transplant tissue from untreated, CD4⁺ T_reg or CD8-depleted recipient mice were analyzed for severity of AMR pathology and scored according to Banff criteria, with a composite score and individual parameter histologic scores for PTC margination, arteritis, and C4d deposition. CD8⁺ T cell-depletion significantly increased the D14 composite AMR pathology score (6.8±0.4, n=5; *p<0.0001) indicating more severe immune injury compared to untreated WT recipients (2.4±0.2, n=5). CD25⁺CD4⁺ T_reg cell-depletion did not significantly impact AMR score (3.6±0.7, n=5; p=NS) compared to untreated WT recipients and was significantly less than CD8-depleted recipient mice (**p=0.0007). Error bars designate standard error of combined data from independent experiments.

Furthermore, while CD8-depletion in WT KTx recipients was associated with increased severity of AMR histopathology in kidney allografts (p<0.0001), CD4⁺ T_reg cell-depletion was not (Figure 8D). Depletion of CD4⁺ T_regs in WT recipients did not impact either composite or individual parameter AMR pathology scores (p=NS for all) while CD8-depletion resulted in an increase in both individual parameters and composite AMR pathology scores (p<0.004, Figure 8D). Altogether, these studies with ACT into high alloantibody-producing CCR5 KO KTx recipients and subset depletion in low alloantibody-producing WT KTx recipients demonstrate the higher potency of alloprimed CXCR5⁺CD8⁺ T cells compared to CD4⁺ T_regs for regulation of humoral alloimmunity.

DISCUSSION

The deleterious consequences of DSA upon both short and long-term allograft outcomes after transplantation coupled with the absence of specific immunotherapeutic strategies to suppress DSA after solid organ and cell transplantation necessitate investigations to identify key drivers of dysregulated humoral immunity and to test therapeutic strategies to restore immune regulation. The availability of a clinically relevant animal model for AMR after KTx provided an opportunity to determine the extent to which deficits and/or dysfunction of antibody suppressor CD8⁺ T cells or CD4⁺ T regulatory cells contribute to the dysregulated humoral alloimmunity observed in CCR5 KO KTx recipients and to test the efficacy of adoptive cellular therapy to regulate humoral alloimmunity following kidney transplantation.

We previously published the remarkable efficacy of ACT with alloprimed CXCR5⁺CD8⁺ T cells to suppress humoral alloimmunity after transplant (27, 28). The current studies uncovered a marked deficit of alloprimed CXCR5⁺CD8⁺ T cells in the spleen and peripheral circulation at day 7 and 14 after KTx in CCR5 KO recipient mice. The fact that alloprimed CXCR5⁺CD8⁺ T cells do not expand after KTx in CCR5 KO recipients and the reduced efficacy of ACT with alloprimed CCR5-deficient CXCR5⁺CD8⁺ T cells to suppress alloantibody production in CCR5 KO KTx recipients imply that CCR5 is critical to the development, activation, expansion, effector function and/or trafficking of antibody-suppressor CXCR5⁺CD8⁺ T cells. However, additional studies clarified that CCR5-deficiency does not affect proliferation or trafficking since after adoptive transfer CCR5-deficient CXCR5⁺CD8⁺ T cell proliferation (Supplemental Figure 6) and trafficking to lymphoid tissue (Supplemental Figure 7) are not significantly different than after ACT with WT CXCR5⁺CD8⁺ T cells. Interestingly, CCR5-deficiency did significantly impact the trafficking of CXCR3⁺CD8⁺ T cells to the kidney allograft. We published that the CXCR3⁺CD8⁺ T cell subset (but not the CXCR5⁺CD8⁺ T cell subset) mediates transplant rejection (27, 63). Thus, CCR5 appears to differentially impact the trafficking pattern of alloprimed CXCR3⁺ but not CXCR5⁺ CD8⁺ T cell subsets. CCR5-deficiency does not block CXCR5⁺CD8⁺ T cell subset activation since the percentage of CD44⁺IFNγ⁺ CXCR5⁺CD8⁺ T cells is similar between C57BL/6 and CCR5 KO recipients [approximately 70% of splenic alloprimed CXCR5⁺CD8⁺ T cells are CD44⁺IFNγ⁺ in WT and in CCR5 KO recipients; Supplemental Figure 3, with similar IFNγ MFI for CD44⁺CXCR5⁺CD8⁺ T cells of WT and CCR5 KO mice (not shown)]. In accordance with our findings that total quantity of CD44⁺IFNγ⁺ CXCR5⁺CD8⁺ T cells is reduced in CCR5 KO compared to WT recipients, it has been reported by another group that CCR5 KO transplant recipients have 60% of the quantity of IFN-γ⁺CD8⁺ T cells compared to WT transplant recipients (64). When we analyzed our data, we also found that CCR5 KO KTx recipients have reduced quantity of IFN-γ⁺CD8⁺ T cells (65%) compared to WT KTx recipients (not shown). We have previously reported using IFNγ KO CD8⁺ T cells that CD8⁺ T_Ab-supp cells require IFNγ for in vitro and in vivo function. However, it is not clear if this was because IFNγ was required for maturation and/or effector function of CD8⁺ T_Ab-supp cells (29, 65). Alloprimed CCR5-deficient CXCR5⁺CD8⁺ T cells mediate reduced in vitro cytotoxicity against alloprimed IgG⁺ B cells compared to WT CXCR5⁺CD8⁺ T cells (28). Thus, both CCR5 and IFNγ may be required for optimal development of antibody-suppressor CXCR5⁺CD8⁺ T cells posttransplant. Overall, these data suggest that the adverse effect of CCR5-deficiency on development and/or function of CXCR5⁺CD8⁺ T cells in the lymphoid compartment accounts for the observed dysregulated humoral alloimmunity in CCR5 KO Tx recipients.

CCR5 KO KTx recipients had significantly reduced quantities of graft-infiltrating CD4⁺ T_regs (~70% IL-10⁺FoxP3⁺) on day 14 posttransplant compared to WT KTx recipients. Our data has some differences and some similarities to results reported for bm12 cardiac allografts transplanted into WT or CCR5 KO mice that are likely attributed to strain, MHC mismatch and transplant organ differences (42). In the cardiac transplant studies a reduced quantity of graft-infiltrating CD4⁺ T_regs was observed on day 10 in CCR5 KO compared to WT recipients, similar to our findings in the KTx model. However, Nozaki et al. reports that ACT with CD4⁺ T_regs (1.2×10⁶ cells) results in long-term acceptance of bm12 cardiac allografts. In the current studies we observed mild to moderate improved KTx pathology (despite no effect on alloantibody titer) in recipients that received ACT with CD4⁺ T_regs. The conclusion from the murine cardiac transplant studies was that acute rejection of bm12 cardiac allografts in CCR5 KO mice (compared to prolonged acceptance in WT mice) arises from a deficit in quantity or trafficking of CD4⁺ T_regs in CCR5 KO recipients needed to control donor-reactive effector T cells in the allograft. It is important to note that the full MHC mismatch KTx model in the current studies results in high alloantibody production and AMR pathology whereas the bm12 cardiac transplant studies skewed to T cell-mediated rejection and intense CD4⁺ T cell infiltration as these recipients had undetectable alloantibody titer and absence of C3d deposition in the allografts (42).

When we tested the relative efficacy of equal quantities of alloprimed antibody-suppressor CXCR5⁺CD8⁺ T cells compared to alloprimed CD4⁺ T_regs for downregulation of alloantibody in CCR5 KO KTx recipients, we found significant differences in effector function and trafficking pattern. Only ACT with CXCR5⁺CD8⁺ T cells downregulated humoral alloimmunity after KTx. ACT with CXCR5⁺CD8⁺ T cells resulted in a nearly 5-fold reduction in alloantibody titer, significantly decreased the quantity of recipient splenic GC B cells and T_FH cells (but not T_FR cells), and significantly reduced the severity of AMR pathology. The reduction of IgG alloantibody production was broad as ACT with CXCR5⁺CD8⁺ T cells inhibited production of all IgG isotypes (Supplemental Figure 8). Further, ACT with alloprimed CXCR5⁺CD8⁺ T cells also suppressed IgM alloantibody titer in CCR5 KO KTx recipients (1071±71 vs 3168±341 in control CCR5 KO recipients; p=0.01, not shown). This suggests that alloprimed CXCR5⁺CD8⁺ T cells target both IgM⁺ and IgG⁺ GC B cells for cytotoxic clearance. We also found differences in trafficking pattern of CXCR5⁺CD8⁺ T cells and CD4⁺ T_regs. ACT of CXCR5⁺CD8⁺ T cells was associated with an increase in quantity of these cells primarily in lymphoid tissue (the spleen) and not the kidney allograft whereas ACT with CD4⁺ T_regs was associated with an increase in CD25⁺FoxP3⁺CD4⁺ T_regs quantity in both the spleen and the kidney allograft. ACT studies using GFP⁺ cell subsets confirmed that CXCR5⁺CD8⁺ T cells traffick to lymphoid tissue whereas CD4⁺ T_regs traffick to both lymphoid tissue and the allograft. Interestingly, CCR5-deficiency did not impair CXCR5⁺CD8⁺ T cell trafficking to lymphoid tissue but abrogated trafficking of CXCR3⁺CD8⁺ T cells to the allograft. In addition, CCR5 deficiency impaired trafficking and/or retention of CD4⁺ T_regs in the allograft.

In the absence of immunosuppression, A/J KTx into CCR5 KO recipients results in severe AMR pathology by two weeks posttransplant which corresponds with graft loss in all recipients within three weeks posttransplant (28). In this study, we evaluated the impact of ACT on the 2-week KTx pathology as an indicator of allograft health. An improvement in KTx pathology was observed after both ACT with CXCR5⁺CD8⁺ T cells and after ACT with CD4⁺ T_regs into CCR5 KO KTx recipients. However, the improvement was more substantial with CXCR5⁺CD8⁺ T cells ACT. ACT with CD4⁺ T_regs moderately improved AMR pathology in the absence of any reduction in alloantibody titer or quantities of GC B cells, or CD4⁺ T_FH cells. These data suggest that the immunoregulatory effects of CD4⁺ T_regs and CXCR5⁺CD8⁺ T cells on KTx pathology predominantly target different pathways. Indeed, the bm12 cardiac transplant studies in CCR5 KO recipients skews towards T cell-mediated acute rejection that was attributed to a defect in quantity or trafficking of CD4⁺ T_regs to the cardiac allograft. Transfer of WT CD4⁺ T_regs cells into CCR5 KO cardiac recipients significantly reduced CD4⁺ T cell graft infiltration and inflammation and restored the allograft acceptance phenotype observed in WT recipients (42). CD4⁺ T_regs cells could also downregulate inflammation and tissue injury mediated by innate immune cells such as NK cells (66, 67) and myeloid cells (68) that have been implicated in the pathology of AMR and rejection in CCR5 KO KTx recipients. Our results taken together with published studies suggest that CD4⁺ T_regs predominantly regulate cell-mediated injury within the allograft whereas CXCR5⁺CD8⁺ T cells regulate humoral immunity through functions that constrain germinal center responses in lymphoid tissue. The localization of CD8⁺ T_Ab-supp effector function to germinal centers is consistent with the requirement for CXCR5 expression and the association of CXCR5⁺CD8⁺ T cell-mediated inhibition of alloantibody and AMR with direct cytotoxic killing of alloprimed IgG⁺ B cells in vitro (27, 28) and in vivo (28). Thus, while both CXCR5⁺CD8⁺ T cells and CD4⁺ T_regs mediate immune regulation, these subsets have distinct trafficking patterns and effector function.

ACT studies using monoclonal OVA-primed OT-I CXCR5⁺CD8⁺ T cells and OVA-primed OT-II CD4⁺ T_regs mirrored results with polyclonal T cell subsets. Only ACT with OVA-primed OT-I CXCR5⁺CD8⁺ T cells (and not OVA-primed OT-II CD4⁺ T_regs) reduced anti-OVA IgG titer produced in CCR5 KO mice transplanted with mOVA KTx. This correlated with reduced quantity of OVA-tetramer positive GC B cells. In vitro cytotoxicity studies demonstrated OVA-primed OT-I CXCR5⁺CD8⁺ T cells kill OVA-primed tetramer positive GC B cells but not alloprimed GC B cells (isolated from OVA-primed or alloprimed CCR5 KO mice respectively). These studies provide more in-depth mechanistic insight about the antigen specificity of CD8⁺T_Ab-supp mediated downregulation of humoral immunity that occurs through cytotoxic killing of antigen-specific GC B cells.

Depletion studies in WT KTx recipients indicate that depletion of either CXCR5⁺CD8⁺ T cells or CD4⁺ T_regs dysregulates humoral alloimmune responses. However, the impact of CD8⁺ T cell depletion was more pronounced. CD8-depleted WT KTx recipients had 4-fold higher alloantibody titer than control untreated recipients and 2-fold higher alloantibody titer than CD4⁺ T_reg depleted KTx recipients. Further, the depletion of CD8⁺ T cells increased the production of all alloantibody IgG isotypes, while CD25-depletion only resulted in enhanced IgG1 and IgG2b alloantibody production (Supplemental Figure 8). Limitations of interpretation of the depletion studies include the global depletion of CD8⁺ T cells rather than the specific CXCR5⁺CD8⁺ T cell subset. Nevertheless, the heightened humoral alloimmune responses are consistent with prior studies showing that CD8⁺ T cell depletion results in higher antibody levels in response to transplantation, platelets, allergens, bacteria, and viral infection (69–77). Depletion of CD4⁺ T_regs in this study was achieved using anti-CD25 mAb as reported by other groups (78, 79).

In the ACT studies, only a single dose of CD4⁺ T_reg cells was tested. While it is possible that a more substantial decrease of alloantibody titer could have been observed with ACT at higher quantities of CD4⁺ T_regs, ACT of 2×10⁶ cells restores the recipient population to double that found in WT KTx recipients. Furthermore, we did observe biologic effects of CD4⁺ T_reg ACT at this dose on improved kidney allograft pathology. In addition, while CD25⁺CD4⁺ T_regs are typically thought to inhibit (pro-inflammatory) CD8⁺ T cell responses (80, 81), we found that ACT of WT CD4⁺ T_regs into CCR5 KO recipients was associated with a significant increase in the quantity of endogenous splenic CXCR5⁺CD8⁺ T cells. Interestingly, this increase in splenic CXCR5⁺CD8⁺ T cells was not accompanied by a concomitant downregulation of humoral alloimmunity consistent with the interpretation that alloprimed CCR5 KO mice have a functional deficit of endogenous (CCR5-deficient) CXCR5⁺CD8⁺ T cells (28). We also observed that CD4⁺ T_reg ACT did not significantly increase the quantity of activated CD44⁺IFNγ⁺CXCR5⁺CD8⁺ T cells (data not shown). Despite the observation that ACT with CD4⁺ T_reg increased the quantity of splenic CXCR5⁺CD8⁺ T cells in CCR5 KO KTx recipients, we also found that CD4⁺ T_reg depletion did not reduce the quantity of endogenous CXCR5⁺CD8⁺ T cells in WT KTx recipients. Further studies are needed to understand how CD4⁺ T_regs impact splenic CXCR5⁺CD8⁺ T cells. Altogether these studies demonstrate the unique and potent regulatory function of primed CXCR5⁺CD8⁺ T cells upon humoral alloimmunity. The persistence of heightened GC B cell and T_FH responses on day 14 posttransplant in CCR5 KO mice compared to WT mice suggests that CD8⁺ T_Ab-supp cells are critical for humoral immune homeostasis.

There is growing interest in harnessing cellular-based therapy in transplantation to prevent immunologic allograft loss and/or minimize immunosuppression therapy. Enthusiasm for CD4⁺ T_reg cell therapy derives from effectiveness demonstrated in animal transplant models (82, 83) and has led to phase I clinical trials in human living-donor KTx recipients (13, 26). Likewise, the translational potential of antibody-suppressor CXCR5⁺CD8⁺ T cells as a cell therapy, to suppress alloantibody production and prevent or treat AMR is supported by pre-clinical data in mouse models and our findings in primary human KTx recipients that the quantity of peripheral blood CD8⁺ T cells with the antibody-suppressor phenotype (CXCR5⁺IFNγ⁺CD8⁺ T cells) inversely correlates with the development of de novo DSA (31). In the current studies, we found that impairment of antibody-suppressor CXCR5⁺CD8⁺ T cells is the critical immune mechanism driving high alloantibody production and AMR pathology in CCR5 KO recipients. Furthermore, ACT studies demonstrate that alloprimed antibody-suppressor CXCR5⁺CD8⁺ T cells are more potent regulators of alloantibody production and AMR pathology in CCR5 KO KTx recipients when compared to CD4⁺ T_regs (see paradigm, Figure 9). Our data also suggest that while these two cell subsets likely target different immune locales and mechanisms impacting acute rejection after KTx, there may be crosstalk between antibody-suppressor splenic CXCR5⁺CD8⁺ T cells and CD4⁺ T_regs. Altogether, our studies support the need for further investigation to determine the role of key molecular signals that guide CD8⁺ T_Ab-supp cell activation, expansion/contraction, trafficking, target cell engagement, and effector function. These studies are necessary to advance CD8⁺ T_Ab-supp cells as a potential cellular therapy to prevent or treat AMR in KTx recipients and/or as a biomarker to assess risk of developing de novo DSA posttransplant.

Figure 9. Paradigm for regulation of humoral alloimmunity by antibody-suppressor CXCR5⁺CD8⁺ T cells and CD4⁺ T regulatory cells.

Antibody-suppressor CXCR5⁺CD8⁺ T cells downregulate posttransplant alloantibody production in part by FasL- and perforin-dependent killing of alloprimed self-IgG1⁺ germinal center (GC) B cells (thick red line), and by downregulation of CD4⁺ T_FH cells (thick red line) either by direct suppression, or indirectly via abrogation of CD4⁺ T_FH and B cell interactions. CD4⁺ T_regs downregulate CD4⁺ T_FH cells and their downstream contribution to alloantibody production (thin red line). CD4⁺ T_regs may also expand the quantity of antibody-suppressor CXCR5⁺CD8⁺ T cells (thin dashed red line).

Key points:

1. CXCR5⁺CD8⁺ T cells are more potent inhibitors of AMR than CD4⁺ T_reg cells.

2. High alloantibody producing transplant recipients are CXCR5⁺CD8⁺ T cell-deficient.

3. CXCR5⁺CD8⁺ T cell mediated suppression of humoral immunity is antigen specific.

Nonstandard Abbreviation

ACT

adoptive cell transfer

GC

germinal center

KTx

kidney transplant

POD

postoperative day

SOT

solid organ transplant

T_Ab-supp

antibody-suppressor T cells

T_regs

regulatory T cells

T_FH

follicular helper CD4⁺ T cell

WT

wild-type