Visual Abstract
Abstract
Background Patients with broad HLA sensitization have poor access to donor organs, high mortality while waiting for kidney transplant, and inferior graft survival. Although desensitization strategies permit transplantation via lowering of donor-specific antibodies, the B cell–response axis from germinal center activation to plasma cell differentiation remains intact.
Methods To investigate targeting the germinal center response and plasma cells as a desensitization strategy, we sensitized maximally MHC-mismatched rhesus pairs with two sequential skin transplants. We administered a proteasome inhibitor (carfilzomib) and costimulation blockade agent (belatacept) to six animals weekly for 1 month; four controls received no treatment. We analyzed blood, lymph node, bone marrow cells, and serum before desensitization, after desensitization, and after kidney transplantation.
Results The group receiving carfilzomib and belatacept exhibited significantly reduced levels of donor-specific antibodies (P=0.05) and bone marrow plasma cells (P=0.02) compared with controls, with a trend toward reduced lymph node T follicular helper cells (P=0.06). Compared with controls, carfilzomib- and belatacept-treated animals had significantly prolonged graft survival (P=0.02), and renal biopsy at 1 month showed significantly reduced antibody-mediated rejection scores (P=0.02). However, four of five animals with long-term graft survival showed gradual rebound of donor-specific antibodies and antibody-mediated rejection.
Conclusions Desensitization using proteasome inhibition and costimulation blockade reduces bone marrow plasma cells, disorganizes germinal center responses, reduces donor-specific antibody levels, and prolongs allograft survival in highly sensitized nonhuman primates. Most animals experienced antibody-mediated rejection with humoral-response rebound, suggesting desensitization must be maintained after transplantation using ongoing suppression of the B cell response.
- desensitization
- kidney transplantation
- carfilzomib
- belatacept
- antibody-mediated rejection
Kidney transplantation is the preferred treatment for end stage renal failure due to improved patient survival, quality of life, and healthcare costs1–3 compared with dialysis. However, patients with HLA antibodies due to prior exposure to allogeneic HLA and a heightened memory B cell repertoire comprise approximately 35% of wait-listed patients.4 Such alloantibody, reflecting B cell immunization, reduces the likelihood of receiving a suitably matched organ, lowers the rate of transplantation, and results in high wait-list mortality. Furthermore, when these patients are successfully transplanted, they have worse short- and long-term graft survival compared with patients who are not sensitized due to increased risk of immunologic rejection.5–8 Patients who are sensitized are becoming increasingly prevalent on the kidney, heart, and lung transplant wait-lists9 due to previous failed transplants, blood transfusions, or pregnancy. Thus, treatment strategies that effectively reduce the memory B cell response and thereby allow rejection-free transplantation in this challenging patient population are urgently needed.
Desensitization using intravenous Ig, plasmapheresis, rituximab, or combinations of these therapies has been used to permit transplantation across the HLA barrier.10,11 Although desensitization decreases wait times and increases the rate of transplantation, these approaches are resource intense, with increased antibody-mediated rejection (AMR) rates,12–15 and graft and patient survival have been below national averages for HLA-compatible transplantation.16 From a mechanistic perspective, existing treatments including rituximab fail to target terminally differentiated plasma cells (PCs) that reside either in the bone marrow (BM), secondary lymphoid organs, or inflammatory sites due to relatively sparse expression of surface antigens, including lack of pan–B cell markers like CD20.17
Proteasome inhibitors (PIs) that target PCs have been evaluated for treatment and prevention of AMR.17–19 This class of drugs causes PC apoptosis through the accumulation of unfolded proteins, leading to oxidative stress. Initial experience of using PIs for HLA desensitization was met with promising results.20–22 However, recent data from larger trials using PIs for desensitization have shown a lack of durable antibody reduction.23 We demonstrated in a nonhuman primate (NHP) model that the PI bortezomib transiently reduced BM-PC levels but did not affect levels of donor-specific antibodies (DSAs) as a consequence of a rapid compensatory germinal center (GC) response24 involving B cell proliferation in B cell follicles. Subsequently, we combined bortezomib with dual costimulation blockade using belatacept (modified CTLA4-Ig) and 2C10R4 (anti-CD40 mAb).25,26 Although this regimen caused PC depletion and GC collapse and prevented rejection, overall survival of transplanted recipients was not improved compared with untreated controls due to weight loss and infections. In this study, we hypothesized that the use of a less toxic, irreversible proteasome inhibitor (carfilzomib) in combination with a single costimulation blocker (belatacept) would provide less-profound suppression of protective immunity but adequately reduce the allo-MHC memory response.
Methods
Allosensitized NHP Kidney Transplantation and Desensitization Regimens
All animal care and procedures were conducted in accordance with National Institutes of Health (NIH) guidelines and were approved by the Duke University Institutional Animal Care and Use Committee (Duke IACUC 209-15-07). Male NHPs (Macaca mulatta) were obtained from Alphagenesis (Yemassee, SC). MHC class-1 and -2 mismatched pairs of NHPs were sensitized to each other with two sequential full-thickness skin transplants performed as previously described.27,28 At 8–12 weeks after the second skin transplant, NHPs were desensitized over the course of 4 weeks with once weekly treatments of the following regimens: 27 mg/m2 intravenous (iv) carfilzomib (Onyx Pharmaceuticals, San Francisco, CA) plus 20 mg/kg iv belatacept (Bristol-Myers Squibb, New York, NY), or no treatment. Two weeks after the final desensitization treatment, swapping kidney transplants with bilateral native nephrectomies were performed between the allosensitized NHP pairs as previously described.26,27 All transplanted NHPs received induction therapy with 50 mg/kg iv rhesus anti-CD4 mAb (CD4R1; NIH NHP Reagent Resource, Worcester, MA) on day −5 before transplant and 25 mg/kg iv rhesus anti-CD8 mAb (MT807R1; NIH NHP Reagent Resource) on the day of kidney transplant. Maintenance immunosuppression after kidney transplant consisted of intramuscular (im) tacrolimus (Astellas Pharma, Northbrook, IL) twice daily with the dose adjusted to maintain trough levels at 8–12 ng/ml, 30 mg/kg by mouth of mycophenolate mofetil (MMF) for oral suspension (Genentech, San Francisco, CA) twice daily, and methylprednisolone (Pfizer, New York, NY) tapered from 15 mg/kg on day of transplant and then a halved dose daily until a maintenance dose of 0.5 mg/kg was reached and continued until the end point. A subcutaneous dose of 6 mg/kg of ganciclovir (Fresenius Kabi, Lake Zurich, IL) was administered daily for rhesus cytomegalovirus (rhCMV) reactivation prophylaxis, and the dose was increased to 7.5 mg/kg for treatment when rhCMV reactivation was identified by rising viral titers measured weekly by quantitative RT-PCR. Suspected acute cellular rejection episodes (indicated by acute rises in creatinine) were treated with methylprednisolone at a dose of 125 mg/kg im daily for 3 days, 75 mg/kg im daily for 3 days, and 25 mg/kg im daily for 3 days.
Polychromatic Flow Cytometric Analysis
Lymph nodes were obtained from biopsy of axillary or inguinal lymph nodes, BM was obtained from aspiration of marrow from the iliac crest of NHPs at the time points indicated in the text. Single-cell suspensions of lymph node, BM, and PBMCs were prepared and stained with Fixable Blue viability dye according to the manufacturer’s recommendations. Cells were washed with 2% FBS in PBS and stained with the following fluorochrome-conjugated mAbs against human: CD127, CD185 (CXCR5), CD19, CD20, CD25, CD278 (inducible T-cell co-stimulator [ICOS]), CD279 (PD-1), CD28, CD3, CD38, CD4, CD8, CD95, IgD, IgG, and IgM. Cells were then fixed and permeabilized using the eBioscience Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific, Waltham, MA) and stained with fluorochrome-conjugated mAbs against human Ki67 and Foxp3 (see Supplemental Table 1 for clones and providers). The samples were washed and fixed with 1× stabilizing fixative solution (BD Bioscience, San Diego, CA). The flow cytometry was performed on a BD LSRFortessa and analyzed using FlowJo software version 9 or version 10 (Tree Star, Ashland, OR).
Detection of DSAs
DSA levels were determined by flow cytometry crossmatch using 3×105 donor PBMCs or splenocytes and recipient serum was collected from serial blood draws as previously described.28 To measure DSA titer, serum samples collected before and after desensitization were thawed from −80°C storage, spun at 14,000 × g for 10 minutes. All samples were serially diluted (twofold) from a dilution of 1:50 to 1:512,000 and compared with a neat naive sample. A mean fluorescence intensity (MFI) value lower than neat naive sample was considered negative. The final dilution with a positive signal became the titer. Briefly, donor PBMCs or splenocytes were incubated with recipient serum, washed, and stained with FITC-labeled anti-monkey IgG (KPL, Gaithersburg, MD), anti-CD20 mAb (2H7), anti-CD3 mAb (SP34–2) (both BD Bioscience), and Live/Dead Fixable Blue staining (Life Technologies, Carlsbad, CA). MFI of anti-monkey IgG on T or B cells was measured on BD LSRFortessa (BD Biosciences) and analyzed using FlowJo software version 9 or version 10 (Tree Star) and expressed as MFI or MFI change from presensitized time point.
Detection of Total IgG, Anti-rhCMV Glycoprotein B, and Anti-Tetanus Antibodies
To measure anti–glycoprotein B (anti-gB) and anti–tetanus toxoid (anti-TT)–specific serum IgG, ELISA plates were coated with 30 ng RhCMV gB (courtesy of Dr. Mark Walters at the University of Alabama at Birmingham) or 30 ng TT (Creative Diagnostics) per well. After blocking with assay diluent (1× PBS containing 4% whey, 15% normal goat serum, and 0.5% Tween 20), threefold dilutions of serum samples were added and later detected with a horseradish peroxidase–conjugated mouse anti-monkey IgG (Southern Biotech), and then developed using SureBlue Reserve tetramethylbenzidine peroxidase substrate (KPL). Anti-TT antibodies are reported as log10 area-under-the-curve values, which indicate the magnitude of the area under the sigmoidal dilution versus OD curve. Area under the curve was chosen because full sigmoidal binding curves were not obtained, complicating the determination of the median effective dose.
Kidney Allograft Survival, Histology, Immunohiostologic Analysis, and Pathologic Grading
Peripheral blood was drawn weekly for complete blood counts, serum chemistries, rhCMV viral titers, and tacrolimus trough levels. Allograft survival end points were defined by persistent elevations in serum creatinine of >4 mg/dl. Ultrasound-guided core needle biopsies of the kidney allografts were performed at 4 and 8 weeks post-transplant in those that survived to these time points. Formalin-fixed paraffin Schiff (Periodic acid–Schiff) or polyclonal anti-human C4d (American Research Products, Waltham, MA) was used for histologic evaluation. Histology specimens were evaluated in a blinded fashion by a transplant pathologist and scored according to the current Banff criteria of renal allograft pathology.29–32 Biopsy or necropsy findings were analyzed both individually and in functional clusters.33–36
Statistical Analyses
Statistical analyses were performed using Prism 7.0 and 8.0 (GraphPad Software, San Diego, CA). Values of P<0.05 were considered to be statistically significant. Survival analysis was performed using the Kaplan–Meier method and log-rank test. Normally distributed data within the same treatment group but at different time points were evaluated using a two-tailed paired t test. The ratio t test was performed for the DSA titer. Statistical comparisons between different groups were performed with two-tailed unpaired t test for normally distributed data or the Mann–Whitney U test for categorical data.
Results
Desensitization with Carfilzomib and Belatacept Significantly Reduces BM PCs, Circulating Follicular Helper T Cells, and Class-Switched IgG+ B Cells in Allosensitized NHPs
Maximally MHC-mismatched donor-recipient pairs of NHP subjects were sensitized by two sequential skin transplants resulting in high DSA levels. The allosensitized pairs were then treated with no desensitization (control), or desensitization with carfilzomib and belatacept (CFZ-Bela) (Figure 1A). One month of treatment with CFZ-Bela was tolerated well. As shown in Supplemental Figure 1, liver enzymes alanine aminotransferase and alkaline phosphatase remained stable. Platelet count and leukocyte populations were not significantly affected by the treatment; specifically, the absolute counts of CD4+ and CD8+ T cells and CD20+ B cells in peripheral blood remained stable (Supplemental Figure 2). As expected, the frequency of PCs (CD20−CD3−CD19+CD38+ cells) in BM significantly declined with CFZ-Bela treatment (P=0.019; Figure 1B). Notably, class-switched (IgD−IgG+CD20+) B cells were reduced (Figure 1D). Both ICOS+PD1+ and CXCR5+ICOS+PD1+ CD4 T cells were greatly reduced in circulation (Figure 1D). However, we also observed the decline of both regulatory T (Treg) cell (CD4+CD25+CD127lo) and follicular regulatory T (Tfr) cell (CD4+CXCR5+CD25+CD127lo) populations in peripheral blood (Figure 1E). Therefore, we also evaluated Treg/follicular helper T cell (Tfh) and Tfr/Tfh ratios to assess immune deviation; however, the ratios showed only a trend toward immune regulation (Supplemental Figure 3A).
Desensitization with CFZ-Bela significantly reduced BM PCs, circulating Tfh cells, and isotype-switched IgG+ B cells. (A) Schematic representations of sensitizations, desensitizations, and kidney transplantation. (B–E) Representative flow plots of PCs in BM, isotype-switched B cells in blood, Tfh and CXCR5+ Tfh cells in blood, Treg and CXCR5+ Treg cells in blood harvested pre- and postdesensitization with CFZ-Bela. BM PCs (CD19+CD38+), circulating isotype-switched B cells (IgD−CD38+), circulating Tfh (ICOS+PD1+) and CXCR5+ Tfh cells (CXCR5+ICOS+PD1+), and circulating Treg (CD25+CD127lo) and Tfr (CXCR5+CD25+CD127lo) cells were significantly reduced after CFZ-Bela treatment. Bx, biopsy; LN, lymph node.
Desensitization with CFZ-Bela Reduces DSA Levels while Preserving Protective Humoral Immunity in Allosensitized NHPs
DSA levels in serum measured by T and B cell flow crossmatch reflected those changes of BM PCs and lymph node Tfh cells and declined significantly after 1 month of treatment with CFZ-Bela (Figure 2A). We also evaluated the DSA titer from serum samples collected before and after desensitization. As shown in Figure 2B, the DSA titer measured from serial dilution was significantly reduced after CFZ-Bela treatment (n=6). However, the DSA titer was unchanged in the serum samples from control animals (n=4). Given that PIs lead to apoptosis of PCs, we sought to address whether protective antibodies decline with such treatment. Our prior studies using proteasome inhibition with dual costimulation blockade resulted in a prominent reactivation of rhCMV, limiting long-term animal survival.26 We therefore measured total IgG, anti-gB (CMV) antibody, and anti-TT antibody in serum before and after treatment. All animals were vaccinated with tetanus at least 4 months before allosensitization. Interestingly, the levels of total IgG, gB-specific, and TT-specific antibodies remained unchanged during treatment (Figure 2C).
Desensitization with CFZ-Bela significantly reduced DSA levels but total IgG, anti-CMV, and anti-TT antibodies were not affected. (A) DSA levels measured by T cell flow crossmatch and B cell flow crossmatch before allosensitization (naive, white bars), after allosensitization but before desensitization (pre, blue bars), and after desensitization (post, orange bars) are expressed as MFI in control- (n=4) and CFZ-Bela–treated (n=6) animals. (B) Evaluated DSA titer based on twofold serial dilution showed a significant reduction after desensitization with CFZ-Bela. DSA titer from date-matched control animals was not changed over time. (C) Total IgG, antibody against CMV (gB), and antibody against tetanus (TT) were not significantly changed by desensitization with CFZ-Bela treatment. *P<0.05, **P<0.005.
Belatacept with Carfilzomib Promotes Immune Modulation in Lymph Node–Resident Cells
To assess the effect of CFZ-Bela treatment on secondary lymphoid tissues, we analyzed lymph nodes before and after treatment. PD-1hi CD4 T cells were noted only in the lymph node and may represent bona fide GC Tfh cells. Similar to our previous reports with dual costimulation blockade and bortezomib,25,26 CFZ-Bela also showed a strong trend toward reduced Tfh cells in lymph nodes (P=0.067; Figure 3A). One animal (H34V) showed increased frequency of Tfh and CXCR5+ Tfh cells in the lymph node. The Treg population of lymph node showed a significant reduction (Figure 3B), whereas the frequencies of CD4+ T cells, CD8+ T cells, and CD20+ B cells in lymph node did not change significantly with CFZ-Bela (data not shown). Isotype-switched B cells homogenously declined in the lymph node (Figure 3C, Supplemental Figure 4). We observed collapsed GC morphology after CFZ-Bela with in situ GC staining for CD20 and Ki67 (Figure 3D). Lymph node Treg/Tfh and Tfr/Tfh ratios did not show significant immune deviation (Supplemental Figure 3B). Because Tfh and Treg cells are not the only cell populations affected by the treatment, we analyzed cell populations using unbiased t-distributed stochastic neighbor embedding plots (Figure 3E, Supplemental Figure 5) that show modified CD8 populations by CFZ-Bela in the probed area. ICOS expression on CD8 T cells was greatly reduced such that CD8+ICOS+PD1+ cells significantly declined (Figure 3F). These data demonstrate a broad effect of CFZ-Bela resulting in strong reduction of Tfh cells (CD4+ICOS+PD1high) and Treg cells but also a reduced activated CD8 T cell population in lymph nodes.
Desensitization with CFZ-Bela suppressed already established cells in the lymph node. (A) Representative flow plots of Tfh and CXCR5+ Tfh cells in lymph nodes before and after CFZ-Bela treatment. Lymph node (LN) Tfh (ICOS+PD1+) cells showed strong trend of reduction, whereas no difference was observed on CXCR5+ Tfh cells in the lymph node. (B) Representative flow plots of Treg and Tfr cells in the lymph nodes before and after CFZ-Bela treatment. Lymph node Treg cells were also significantly decreased, whereas no difference was observed on Tfr cells in the lymph node. (C) Isotype-switched B cells in the lymph node were significantly decreased. (D) Representative immunohistochemistry of lymph nodes including B cell follicles and GC staining for Ki67 (red) and CD20 (brown) before and after CFZ-Bela treatment. (E) t-distributed stochastic neighbor embedding (t-SNE) plot shows changes of PD1 and ICOS expression on CD8 T cell population (dotted area). (F) Desensitization with CFZ-Bela decreased PD-1 expression on CD8 T cells and significantly reduced ICOS+PD1+CD8 T cells.
Pretransplant Desensitization with Belatacept and Carfilzomib Reduced Early AMR and Prolonged Graft Survival
To evaluate the overall effect and durability of desensitization induced by CFZ-Bela, swapping life-sustaining renal transplants were performed between the allosensitized pairs 2 weeks after the final CFZ-Bela infusion. All recipients received anti-CD4 and anti-CD8 mAb induction therapy and maintenance immunosuppression with tacrolimus, MMF, and methylprednisolone as shown in Figure 1A. Survival after kidney transplantation was significantly prolonged in subjects desensitized with CFZ-Bela (119±75 days, P<0.05; Figure 4A) compared with controls (5±3 days) and subjects desensitized with carfilzomib alone (6±1 days; Supplemental Figure 6). No animal showed significant weight loss (Supplemental Figure 7). Five out of six treated animals showed long-term graft survival. Evidence of allograft injury by AMR score (g+ptc+c4d) was significantly lower in the first month after kidney transplant in the CFZ-Bela group compared with controls (Figure 4, B and C). C4d staining was present regardless of treatment which may represent the effect of residual DSA after desensitization. It is notable that kidney biopsies interpreted according to Banff criteria strongly suggested less antibody-mediated injury at 1 month post-transplantation (Supplemental Figure 8).
Desensitization with CFZ-Bela reduced early antibody-mediated injury and prolonged allograft survival in sensitized NHPs. (A) Percentage survival of sensitized NHP subjects after life-sustaining kidney transplant from the same donor to which they were previously sensitized. Before kidney transplant, recipients received either no desensitization (controls, black) or desensitization (once weekly infusions for 4 weeks) treatment with CFZ-Bela (blue). (B) Representative hematoxylin and eosin (H&E) (top panels) and c4d (bottom panels) staining of grafts from control at euthanasia (postoperative day 9 [POD9]) and CFZ-Bela–treated animals at 1 month biopsy (POD27). Asterisk (*) indicates interstitial hemorrhage, which can be from an AMR or severe cellular rejection within arteritis. (C) AMR scores calculated based on Banff grading (measuring antibody-mediated injury, g+ptc+c4d). AMR score was significantly reduced in CFZ-Bela–treated animals compared with controls. g, allograft glomerulitis; ptc, peritubular capillaritis.
CFZ-Bela–Based Desensitization Did Not Prevent Late Rebound DSA with Allograft Injury in Allosensitized Recipients
We continuously monitored animals treated with CFZ-Bela desensitization. Unfortunately, the perioperative CFZ-Bela desensitization did not completely prevent a post-transplant humoral response. As shown in Figure 5A, animals showed elevated levels of DSA post-transplant at varying post-transplant time points. Interestingly, the post-transplant serum creatinine level was reflecting the elevation of DSA (Supplemental Figure 7). It is notable that animals with limited graft survival showed earlier DSA rebound at post-transplant week 1 (Figure 5B). We evaluated post-transplant PC populations in BM and lymph node together with Tfh cells and IgG B cells (Figure 5C). We observed that PCs (CD19+CD38+) were not fully recovered by 1 month post-transplantation in the BM, whereas the PC population significantly increased in lymph node. This may reflect more rapid PC differentiation in lymph node rather than repopulation in BM after kidney transplantation in the sensitized animals. Despite the significant survival advantage of monkeys desensitized with CFZ-Bela, allograft histology even in long-term survivors demonstrated evidence of both cellular-mediated rejection and AMR (Figure 6). Individual Banff scores illustrate that glomerulitis, inflammation in the scarred cortex, and tubulitis contributed significantly to the pathology of late AMR when comparing early and late kidney biopsies. Additionally, clustered Banff scores demonstrate that microcirculation inflammation, antibody-mediated injury, cell-mediated rejection, chronic allograft damage, and microcirculation lesions were prominent in late kidney biopsies because peri-transplant desensitization failed to prevent late AMR. Graft histology showed low-level acute rejection grades (Table 1). It is notable that one animal (H34V) shows elevation of only class-2 but not class-1 DSA (Supplemental Figure 9). Interestingly, this animal showed an increase in Tfh and CXCR5+ Tfh cells in the lymph nodes during desensitization. All long-surviving animals developed tertiary lymphoid structures in the graft (Figure 5D, Supplemental Figure 10). In contrast to our previous experience of desensitization with bortezomib/belatacept/anti-CD40 mAb,26 no CFZ-Bela–treated animals experienced infectious complications including clinically significant reactivation of rhCMV or other toxicity attributable to the desensitization regimen (data not shown). Here, we demonstrate that short-term pretransplant desensitization with CFZ-Bela markedly prolongs allograft survival in MHC-mismatched recipients, but post-transplant—while on conventional tacrolimus/MMF/steroid immunosuppression—there is histologic evidence of AMR associated with the rebound of antibody-mediated injury. This suggests that targeting the B cell axis with costimulation blockade needs to continue after transplantation. Belatacept-based maintenance immunosuppression after kidney transplantation may achieve more sustained control over the GC.37
Desensitization with CFZ-Bela did not prevent rebound DSA and later AMR. (A) Normalized post-transplant DSA kinetics to pretransplant time point. Serum DSA level was evaluated by B cell flow crossmatch. (B) Animals with shorter graft survival (red line) showed significantly elevated DSA 3–4 days after kidney transplantation (Kd3-4). (C) Post-transplant T cells, B cells, and PCs in the BM and lymph nodes (LN). Lymph node PCs, Tfh cells, and IgG B cells were significantly increased after kidney transplantation, whereas BM PCs were not significantly increased. (D) Lymphoid neogenesis in the graft of desensitized animals. Representative hematoxylin and eosin (H&E), CD3, CD20, and CD68 staining from CFZ-Bela–treated animal at euthanasia. Images were adapted from whole slide scan. Mon, month; Rx, treatment; Tx, transplantation.
Histologic analysis illustrates peri-transplant desensitization did not prevent late AMR. (A) The selected individual Banff histologic scores. (B) The selected Banff histologic scores as functional clusters (microcirculation inflammation is considered AMR with a score of more than two). Data are presented as mean+SD, with analysis by the paired t test. *P<0.05, **P<0.01, ***P<0.001, NA, not applicable. ah, arteriolar hyalinosis; CADI, Chronic Allograft Damage Index score 2; cg, allograft glomerulopathy; cg+mm, glomerulopathy and microcirculation deterioration; ci, interstitial fibrosis; ct, cortical tubular atrophy; ci+ct, tubulointerstitial scarring; cv, vascular fibrous intimal thickening; g, allograft glomerulitis; g+ptc, microcirculation inflammation; g+ptc+c4d, antibody-mediated injury; i, interstitial inflammation; i-IFTA, inflammation in scarred cortex; mm, mesangial matrix; ptc, peritubular capillary inflammation; ptc+g+cg, microcirculation lesions; t, tubulitis; ti, total inflammation; v, arteritis; v+t+i, cell-mediated rejection.
Banff allograft pathology quantitative scores and other pathologic characteristics
Discussion
In humans, DSA levels above threshold render patients unacceptable for kidney transplantation due to risk of AMR. In this study, we created an arduous immunologic barrier for successful kidney transplantation in which sensitization between maximally MHC-mismatched animals through two sequential skin transplants created a high level of donor-specific immunity. Despite the challenge of this model, we demonstrate a significant reduction of DSA and prolonged allograft survival with dual PC and GC targeting using the commercially available, Food and Drug Administration (FDA)–approved drugs CFZ-Bela.
No recipients had any noted off-target drug effects attributable to this regimen or clinically significant reactivation of rhCMV, despite aggressive cytolytic induction immunotherapy at the time of renal transplantation. These findings suggest this desensitization regimen, in combination with current standard-of-care therapies that effectively target circulating alloantibodies, holds promise for translation into desensitization protocols for human solid organ transplantation, especially given the formidable donor-specific NHP sensitization (two skin grafts and full MHC mismatch) compared with humans sensitized to third party HLA antigens.
As a result of the dual desensitization treatment with CFZ-Bela, we noted changes in several subpopulations of immune cells despite unchanged absolute counts and proportions of CD4+, CD8+, and CD20+ cells. First, we demonstrate a significant reduction in putative antibody-secreting PCs (CD19+CD38+) in the BM of dual-treated animals who achieved long-term allograft survival. The rationale for use of carfilzomib for PC inhibition comes from experience with this drug in the field of oncology. Carfilzomib is FDA approved for patients with relapsed or refractory multiple myeloma, a PC malignancy, and was shown to improve median progression-free survival compared with bortezomib, with significantly lower rates of peripheral neuropathy in a large randomized controlled trial.38 Similarly, several studies have shown that proteasome inhibition effectively induces apoptosis of non-neoplastic PCs.17,39–43 Interestingly, CD28 expression has been postulated to be critical to the survival and expansion of multiple myeloma cells in humans.44–46 Additionally, ligand binding to CD80 and CD86 may directly affect B cell and antibody production by signaling through these receptors.47–49 Most recently, belatacept was shown in vitro to reduce plasmablast differentiation, antibody production, and Blimp1 production in a T cell–independent manner.50 These findings raise the possibility that both drugs concomitantly impair PC survival. However, PI monotherapy, bortezomib,25 or carfilzomib in the sensitized NHP model did not show a profound reduction of DSA. This observation was also made in the randomized placebo-controlled trial of proteasome inhibition in AMR (BORTEJECT); namely, that bortezomib monotherapy did not affect de novo DSA.51 It was speculated that there is a compensatory mechanism of GC activation after PC depletion. Therefore, we hypothesize that blocking the GC response in combination with PC depletion would reduce DSA in sensitized NHPs. Fittingly, in the setting of PC inhibition, DSA serum levels were also significantly decreased by dual treatment as measured by both B and T cell crossmatch, whereas the DSA of nondesensitized controls showed no significant reduction over time.
Concurrently, we demonstrate a significant reduction in the proportion of activated Tfh cells (CD4+ICOS+PD1high) in the lymph nodes of dual-treated animals. We presume this to be an effect of belatacept, as treatment with bortezomib24 and carfilzomib monotherapy tended to induce this population of cells that direct development of naive B cells into plasma and memory B cells. Belatacept is known to block T cell activation because binding CD80 and CD86 on antigen-presenting cells prevents delivery of a costimulatory second signal by inhibiting the interaction of these ligands with CD28 expressed on T cells.52 Indeed, our recent work using a NHP model of AMR demonstrated that belatacept may inhibit GC Tfh cells.53 Other authors have confirmed these findings and have shown that belatacept blocks CD28-mediated activation of Tfh cells, suggesting this impaired crosstalk may be a mechanism for the low rate of de novo DSA development in patients who have received a kidney transplant and are maintained on belatacept-based immunosuppression.50 Furthermore, we found a reduction in GC response as well as isotype-switched B cells in the lymph nodes of dual desensitized animals. Ultimately, it appears that proteasome inhibition with single costimulation blockade is sufficient to block the rebound humoral response induced by PI-mediated PC apoptosis.
To evaluate the effect of treatment on allograft survival, we performed kidney transplants after desensitization. Control animals that were not treated (n=4) or treated with carfilzomib monotherapy (n=3) succumbed to rapid AMR within 1 week post-transplantation. Six animals treated with dual targeting (CFZ-Bela) also received kidney transplantation from their maximally MHC-mismatched skin donor. In the dual-treated animals, five of six animals exhibited markedly prolonged allograft survival (log-rank P<0.05 versus controls and versus carfilzomib monotherapy), and two animals reached the study end point of 6 months post-transplant with good renal function. In contrast, sensitized controls and carfilzomib-monotherapy animals showed severe antibody-mediated injury including glomerulitis, peritubular capillaritis, and C4d deposition at the time of early necropsy; whereas less leukocyte infiltration was observed in the dual-treated animals in protocol biopsies performed at later time points.
Given that PIs lead to apoptosis of PCs, a major concern is whether all PCs are killed equally, which could lead to a loss of protective immunity. Indeed, in our prior studies using proteasome inhibition with dual costimulation blockade, prominent reactivation of rhCMV was a major hindrance to long-term graft survival.26 However, although DSAs were reduced, levels of gB- and TT-specific antibodies remained unchanged during treatment. The difference in loss of MHC antibody versus unaltered anti-tetanus antibody could be due to the timing of vaccination and sensitization because the vaccinations were performed before the sensitization. This would mean the efficiency of desensitization is less when the time elapsed since the immunizing event increases. However, the time difference is approximately 2 months at most between the vaccination and the first skin graft, and such a length of time may reflect a different maturity of immune response. The differential effect of CFZ-Bela on anti-viral versus anti-donor humoral responses also may depend on the rate of IgG production by PCs. Resting tetanus-specific PCs have been reported to be in a low state of antibody production and endoplasmic reticulum stress, and may therefore be relatively resistant to proteasome inhibition.54 These findings suggest that this dual therapy is somewhat selective, likely owing to a higher state of activation of DSA-producing PCs in the presence of repeated challenge with donor HLA. In aggregate, these findings suggest that dual treatment with CFZ-Bela may balance reduction in alloantibody responses while preserving protective immunity against opportunistic pathogens.
Despite these encouraging results, there are limitations to this regimen that deserve careful consideration. One animal (G25H) with a high DSA level after desensitization showed allograft rejection in the first week post-transplant. This suggests that antibody removal (for example by plasmapheresis) may need to be combined with a CFZ-Bela regimen to prevent early graft loss when antibody levels are high. One animal (H34V) did not appear to respond to CFZ-Bela, with paradoxic increase in Tfh and CXCR5+ Tfh cells in the lymph nodes. This suggests that a phenotype resistant to this dual targeting strategy may exist. Interestingly, this animal showed elevation of only class-2 but not class-1 DSA. Another limitation was the rebound of DSA, which ultimately promoted AMR. At this point, it is unclear that growing lymphoid structures in the graft actually related to the AMR. A belatacept-based maintenance immunosuppression after kidney transplantation may show potential benefit controlling the GC response.37 We are currently investigating belatacept with or without calcineurin inhibitors as a maintenance immunosuppression in this setting. Despite these limitations, combined proteasome inhibition and costimulation blockade warrants further examination in the broad context of antibody-mediated diseases, including desensitization, AMR, and autoimmunity.
Disclosures
Dr. Knechtle reports common stocks of Bristol-Myers Squibb, during the conduct of the study. Ms. Olaso reports grants from American Society of Transplant Surgeons, during the conduct of the study. Dr. Permar reports grants and personal fees from Merck, grants and personal fees from Moderna, personal fees from Pfizer, and personal fees from Sanofi, outside the submitted work. Dr. Saccoccio reports grants from Thrasher Research Foundation, outside the submitted work. All of the remaining authors have nothing to disclose.
Funding
This NHP study was supported by part of the NIH NHP Transplantation Tolerance Cooperative Study Group sponsored by the National Institute of Allergy and Infectious Diseases, grant U19AI131471 (awarded to Dr. Knechtle).
Supplemental Material
This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2019030304/-/DCSupplemental.
Supplemental Figure 1. Desensitization with carfilzomib and belatacept does not significantly change liver enzymes, platelet counts, and circulating leukocyte populations.
Supplemental Figure 2. The frequency and absolute number of circulating T and B cell populations are not changed by desensitization with carfilzomib and belatacept.
Supplemental Figure 3. Lymph node and circulating Tfr/Tfh ratios are not significantly changed after desensitization with carfilzomib and belatacept.
Supplemental Figure 4. Preservation of IgD+IgM+ cells in the blood, lymph nodes and bone marrow.
Supplemental Figure 5. Identification of immune cell subsets using flow cytometry.
Supplemental Figure 6. Single targeting desensitization with carfilzomib alone fails to prolong allograft survival, reduce early antibody-mediated injury and DSA in allosensitized NHP primates.
Supplemental Figure 7. Post-transplant weight and serum creatinine changes.
Supplemental Figure 8. Desensitization with CFZ-Bela reduced the individual Banff histologic score and AMR related score (g+ptc).
Supplemental Figure 9. Post-transplant DSA kinetics in CFZ-Bela treated animals.
Supplemental Figure 10. Development of ectopic germinal center in CFZ-Bela desensitized animals.
Acknowledgments
We would like to gratefully acknowledge the Duke Laboratory Animal Resources staff and the expert assistance of Dr. Kyha Williams and Dr. Felicita Smith for animal care. We would also like to acknowledge the Substrate Services Core Research Support for weekly viral monitoring and histology support from Dr. Mingqing Song. We greatly appreciate Dr. Maragatha Kuchibhatla (Duke Department of Biostatistics and Bioinformatics) for the statistical support.
Anti-CD4 and anti-CD8 mAb used in this study were provided by the NIH Nonhuman Primate Reagent Resource (R24 OD010976, U24 AI126683).
Data were presented as a plenary presentation at the American Transplant Congress on June 3, 2018 in Seattle, WA.
Dr. Kwun and Dr. Knechtle conceived the study; Dr. Ezekian, Dr. Schroder, Dr. Kwun, and Dr. Knechtle performed NHP experiments and analyzed data; Dr. Mulvihill, Dr. Barbas, Dr. Collins, and Dr. Freischlag participated in NHP surgery; Mr. Yoon and Dr. Yi performed in vitro experiments; Dr. Smith and Dr. Permar performed assays evaluating protective immunity; Ms. Olaso analyzed data; Dr. Farris interpreted data (pathologist); Dr. Ezekian, Dr. Schroder, Dr. Farris, and Dr. Knechtle wrote the manuscript with input from all coauthors.
Footnotes
B.E. and P.M.S. contributed equally to this work.
J.K. and S.J.K. contributed equally to this work.
Published online ahead of print. Publication date available at www.jasn.org.
- Copyright © 2019 by the American Society of Nephrology