Costimulation blockade (CoB)–based immunosuppression offers the promise of improved transplantation outcomes with reduced drug toxicity. However, it is hampered by early acute rejections, mediated at least in part by differentiated, CoB-resistant T cells, such as CD57+PD1 CD4 T cells. In this study, we characterize these cells pretransplant, determine their fate posttransplant, and examine their proliferative capacity in vitro in humans. Our studies show that CD57+PD1 CD4 T cells are correlated with increasing age and CMV infection pretransplant, and persist for up to 1 y posttransplant. These cells are replication incompetent alone but proliferated in the presence of unsorted PBMCs in a contact-independent manner. When stimulated, cells sorted by CD57/PD1 status upregulate markers of activation with proliferation. Up to 85% of CD57+PD1 cells change expression of CD57/PD1 with stimulation, typically, upregulating PD1 and downregulating CD57. PD1 upregulation is accentuated in the presence of rapamycin but prevented by tacrolimus. These data support a general theory of CoB-resistant cells as Ag-experienced, costimulation-independent cells and suggest a mechanism for the synergy of belatacept and rapamycin, with increased expression of the activation marker PD1 potentiating exhaustion of CoB-resistant cells.

This article is featured in Top Reads, p.1407

Kidney allotransplantation is the preferred method of renal replacement therapy, with superior short- and long-term outcomes compared with dialysis (1). Optimal long-term outcomes in kidney transplantation are hampered by necessary but imperfect immunosuppression regimens. For the vast majority of patients, these regimens contain calcineurin inhibitors (CNIs) such as tacrolimus, which is nephrotoxic (2), requires drug monitoring because of differential metabolism (3), and is associated with posttransplant malignancies (4). More recently, the modified CTLA4-Ig fusion protein belatacept has been approved for maintenance immunosuppression in kidney transplantation as a CNI alternative (5, 6). Belatacept works through the mechanism of costimulation blockade (CoB) by blocking the interactions between CD28 on T cells and CD80/86 on APC (7, 8).

Although belatacept-based regimens have been shown to avoid the toxicities of CNIs, their widespread adoption has been hampered by higher rates of early acute rejection, indicating a need to pair belatacept with some adjuvant therapy, ideally one that facilitates the mechanisms of CoB (6, 9). The origins of these early CoB-resistant rejections (CoBRR) remain incompletely defined, highly relevant to the continued use of belatacept, and likely based in the mechanisms distinguishing CNI and CoB-based immunosuppressive approaches. In general, TCR signaling is required for T cell activation. As CNIs effectively inhibit this fundamental requirement, they are highly effective in preventing T cell–mediated alloimmunity. In counter-distinction, CD28–CD80/86 interactions facilitate (but do not initiate) the effects of the TCR’s primary signaling event (10). Thus, belatacept targets an important, but relatively subordinate, signaling event. Importantly, once a T cell has been activated, its dependence on CD28 engagement decreases, and in humans, CD28 expression is markedly reduced during the course of activation-induced cell differentiation, suggesting that costimulation becomes increasingly subordinate with progressive maturation (11). In counter-distinction, activated T cells become increasingly dependent on pathways governed by the mechanistic target of rapamycin (mTOR) to respond metabolically to exogenous cytokines and proliferate (12). This pathway is targeted clinically by the drug rapamycin.

Several surface molecules are known to provide evidence of a T cell’s relative maturation; among them, PD1 and CD57 have been shown to correlate with prior Ag exposure and subsequent activation-induced senescence, respectively (13). These markers have also been shown to associate with impaired viral immunity, suggesting that their acquisition has functional importance in shaping an immune response (1316). We have shown that CD57+PD1 CD4 T cells are enriched in many patients with end-stage renal disease (ESRD) and associate with the posttransplant risk for CoBRR. Moreover, these cells generally lack CD28, are present on histologic sections of rejecting grafts, and contain a transcriptional profile that is associated with allograft rejection (14, 15). Additional research has identified other subsets of mature T cells associated with CoBRR, suggesting that multiple mature cell types may contribute to this rejection phenotype and support the overarching hypothesis that repertoire maturation status influences the efficacy of CoB-based approaches (16, 17). Given these findings, the relative makeup of a transplant recipient’s T cell repertoire would be anticipated to influence the effectiveness of CNI- or belatacept-based immunosuppression.

In the current study, we have endeavored to better define the characteristics of CD57+PD1 CD4 T cells—cells known to be CoB resistant. We find that these cells have substantial CoB-independent activity and persist as a stable feature of the T cell repertoire in kidney transplant recipients despite ongoing CNI-based immunosuppression. Additionally, we show that CD57+PD1 CD4 T cells acquire PD1 with proliferation and that this acquisition is enhanced by rapamycin, but not tacrolimus.

We studied consented patients with ESRD who were awaiting transplantation (n = 39) and kidney transplant recipients (n = 46) who had available pretransplant, and when applicable, posttransplant PBMCs acquired through participation in institutional review board–approved sample acquisition and storage protocols (Pro00057497 and Pro00030485). Clinical data were abstracted from existing clinical databases and patient charts. To examine the natural history of cell populations, patients were chosen who were on CNI-based immunosuppression, did not receive lymphocyte depletion as part of their posttransplant therapy, and were not on a belatacept-based regimen. Additionally, we collected PBMC samples from healthy controls for in vitro assays enrolled under an institutional review board–approved blood acquisition protocol (Pro00062495). For all samples, peripheral blood was collected by venipuncture and isolated by Ficoll-Hypaque density gradient. Cells were then cryopreserved in FCS (Corning, Corning, NY) with 10% DMSO.

We performed multiparameter flow cytometric analysis of patient samples using our phenotypic panel that allowed for identification of CD57+PD1 T cells. We did this cross-sectionally for all patients prior to transplant and longitudinally for those patients who were transplanted. In these and subsequent flow cytometry assays, the Abs used included the following: BD Biosciences (CD2 PE CF594, 562300; CD3 Alexa 700 557943, CD4 BUV395, 563550; CD4 PE-Cy7, 557852; CD8 APC, 555369; CD11a BV421, 563936; CD14 BV510, 563079; CD20 BV510, 563067; CD28 BV711, 56313; CD28 PE-Cy5, 555730; CD57 FITC, 555619; CCR7 [CD197] PE-Cy 7, 557648; and PD1[CD279] PE, 563245; granzyme B PE-CF594, 562462), BioLegend (CD49d PE Cy7, 304314; and PD1 [CD279] PE 329906), eBioscience (CD8 APC-Alexa780, 47-0088-42; and CD40 PerCP eFluor710, 46-0409-42), Miltenyi Biotec (KLRG1 APC, 130-103-639), and Invitrogen/Thermo Fisher Scientific (CD45RA QDot 655, Q10069; Live/Dead Aqua, L34957). Data were acquired on a custom BD LSRFortessa flow cytometer and analyzed using FlowJo 9/10 (Gating Strategy Supplemental Fig. 1A, 1B; Tree Star/Becton Dickinson, Ashland, OR).

Cryopreserved PBMC from pretransplant patients were obtained from cryostorage, thawed in RPMI 1640 supplemented with penicillin–streptomycin at 100 U/ml, l-glutamine at 2 mM (all Life Technologies, Gaithersburg, MD), and 10% FCS (R10 media; Corning, Corning, NY). The cells were then labeled with Violet Proliferation Dye (VPD; BD Biosciences, San Jose, CA) using the manufacturer’s protocol. A magnet-based, pan–T cell isolation kit (Miltenyi Biotec, Auburn, CA) was used to isolate T cells from PBMC. The isolated T cells were then labeled using CD8 APC (555369; BD Biosciences), CD57 FITC (555619; BD Biosciences), and PD1 PE (560795; BD Biosciences) fluorochrome–conjugated mAbs. We sorted the CD8-negative fraction of cells into four populations: CD57+PD1, CD57+PD1+, CD57PD1, and CD57PD1+ using a MoFlo Astrios EQ cell sorter (Beckman Coulter, Indianapolis, IN). Cells were then incubated in R10 in flat-bottom plates with some wells coated with purified OKT3 CD3 Ab at 0.5 μg/ml (eBioscience, San Diego, CA). In some conditions, we cocultured the sorted, VPD-labeled T cells with unlabeled unsorted PBMC from the same cell donor (referred to as bulk cells) at a fixed ratio. We also added soluble CD28 at 2 μg/ml to some conditions (eBioscience). We added immunosuppression to some conditions using tacrolimus at 10 ng/μl (Astellas, Tokyo, Japan), rapamycin at 10 ng/μl (Pfizer, New York, NY), belatacept at 125 μg/ml (Bristol-Myers Squibb, New York, NY), adalimumab at 100 ng/ml (R&D Systems, Minneapolis, MN), or tocilizumab at 50 μg/ml (Genentech, San Francisco, CA). After 96 h of incubation at 37°C in R10, cell supernatants were reserved for cytokine detection (see below), and cells themselves were stained and analyzed by flow cytometry using a BD LSRFortessa flow cytometer (Gating Strategy, Supplemental Fig. 1C). For samples in which a permeable cell well insert was used, sorted cells were placed on the apical side of the insert and bulk PBMC below at an excess number on the basolateral side the insert in a 24-well plate (Costar, Corning, NY) with a 1.0-μm pore insert (Falcon, Corning, NY). These conditions were stimulated using CD3/28 beads (Dynabeads; Thermo Fisher Scientific, Waltham, MA).

Secreted cytokines were assessed from some supernatants using Luminex analysis. Briefly, we used a Human 25-Plex Luminex Panel (LHC0009M; Thermo Fisher Scientific; cytokines included the following: GM-CSF, IFN-α, IFN-γ, IL-1β, IL-1RA, IL-2, IL-2R, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12, IL-13, IL-15, IL-17, TNF-α, eotaxin, IP-10, MCP-1, MIG, MIP-1α, MIP-1β, and RANTES) to interrogate supernatants collected from proliferation assays described above. The supernatants were clarified by centrifugation, diluted 1:5, and analyzed on a Luminex 200 xPONENT 4.2 using standard techniques.

Patient demographics were collected, and standard summary statistics were generated. Univariable correlations of patient demographics with CD57+PD1 cell percentages were performed by linear regression. All parameters collected were included in an initial multivariable linear regression model, and a final model was created by stepwise-backward selection with a cutoff of p = 0.10. Continuous variables were compared by Wilcoxon–Mann–Whitney U tests or t tests and categorical variables by either χ2 or Fisher exact tests as appropriate.

With sorted cell subsets, we compared proliferation across different treatments using a Friedman test, with a Dunn posttest comparing stimulated conditions (with or without immunosuppression) to the unstimulated condition. Linear regression models were used to compare change in cell subsets with varying amounts of proliferation among sorted cell subsets. The baseline percentage of the sorted cell subset was defined as an individual sample’s frequency of that subset in unstimulated conditions without immunosuppression. Principal component analysis (PCA) was used to visualize the relationships of changes in cell marker expression. The PCA was developed using percentage of proliferating cells, change in CD28 expression, change in CD57 expression, and change in PD1 expression to develop the components. We also developed heatmaps to depict change in amount of proliferation and CD57/PD1 status as they were represented in the principal components (i.e., what changes were contributing to the principal components). Multivariable regression was used to determine the effect of different immunosuppressive agents on changes in cell subsets.

Differences in levels of analytes in the Luminex data were compared by a Friedman test, with a Dunn posttest comparing conditions with immunosuppression to the stimulated, nonimmunosuppressed, condition. All statistical tests were performed in STATA 15 (StataCorp, College Station, TX) and PRISM (GraphPad, San Diego, CA).

We first examined 85 patients with ESRD who were awaiting a kidney transplant. Patient demographics are summarized in Table I. Univariable linear regression showed that increasing age and CMV positivity were positively correlated with frequency of CD57+PD1 CD4 T cells prior to transplantation (p = 0.02, R2 = 0.06 for age, and p = 0.029, R2 = 0.06 for CMV; Table II). Backward-stepwise regression yielded the same result, with only age and CMV status correlated with frequency of CD57+PD1 CD4 T cells (Table II).

Table I.
Overall cohort demographics
Characteristicn = 85
Age [median (interquartile range)] 50 (42–61) 
Sex (female) [n (%)] 28 (33) 
Race [n (%)]  
 Black 46 (54) 
 White 35 (41) 
 Other/unknown 4 (5) 
Primary renal disease [n (%)]  
 Diabetes mellitus 27 (32) 
 Focal segmental glomerulosclerosis 6 (7) 
 Glomerulonephritis 8 (9) 
 HIV 3 (4) 
 Hypertension 33 (38) 
 IgA nephropathy 5 (6) 
 Polycystic kidney disease 5 (6) 
 Systemic lupus erythematosus 7 (8) 
History of immunosuppression [n (%)] 25 (29) 
Duration of dialysis [median (interquartile range)] 32 (7–58) 
CMV status [n (%)]a 61 (73) 
EBV status [n (%)]a 77 (98) 
Characteristicn = 85
Age [median (interquartile range)] 50 (42–61) 
Sex (female) [n (%)] 28 (33) 
Race [n (%)]  
 Black 46 (54) 
 White 35 (41) 
 Other/unknown 4 (5) 
Primary renal disease [n (%)]  
 Diabetes mellitus 27 (32) 
 Focal segmental glomerulosclerosis 6 (7) 
 Glomerulonephritis 8 (9) 
 HIV 3 (4) 
 Hypertension 33 (38) 
 IgA nephropathy 5 (6) 
 Polycystic kidney disease 5 (6) 
 Systemic lupus erythematosus 7 (8) 
History of immunosuppression [n (%)] 25 (29) 
Duration of dialysis [median (interquartile range)] 32 (7–58) 
CMV status [n (%)]a 61 (73) 
EBV status [n (%)]a 77 (98) 
a

Indicates missing data; n = 2 for CMV and n = 7 for EBV status.

Table II.
Linear regression identifying associations with CD57+PD1 CD4 T cell prevalence
VariableCoefficient (SE)p ValueR2
Univariable models    
 Age 0.093 (0.41) 0.027 0.06 
 CMV status (reference = negative) 2.78 (1.25) 0.029 0.06 
 EBV status (reference = negative) 1.16 (5.3) 0.83 0.0006 
 Prior IS use (reference = none) −0.70 (1.22) 0.57 0.004 
 Race (reference = non-Black) 1.15 (1.11) 0.30 0.01 
 Sex (reference = male) −1.02 (1.18) 0.39 0.009 
Full multivariable model    
 Age 0.10 (0.05) 0.025 0.14 
 CMV status 2.46 (1.3) 0.063 
 EBV status −2.5 (5.4) 0.62 
 Prior immunosuppression use 0.75 (1.27) 0.56 
 Race 1.40 (1.19) 0.24 
 Sex −1.55 (1.21) 0.21 
Final multivariable model    
 Age 0.09 (0.04) 0.036 0.11 
 CMV status 2.52 (1.23) 0.044 
VariableCoefficient (SE)p ValueR2
Univariable models    
 Age 0.093 (0.41) 0.027 0.06 
 CMV status (reference = negative) 2.78 (1.25) 0.029 0.06 
 EBV status (reference = negative) 1.16 (5.3) 0.83 0.0006 
 Prior IS use (reference = none) −0.70 (1.22) 0.57 0.004 
 Race (reference = non-Black) 1.15 (1.11) 0.30 0.01 
 Sex (reference = male) −1.02 (1.18) 0.39 0.009 
Full multivariable model    
 Age 0.10 (0.05) 0.025 0.14 
 CMV status 2.46 (1.3) 0.063 
 EBV status −2.5 (5.4) 0.62 
 Prior immunosuppression use 0.75 (1.27) 0.56 
 Race 1.40 (1.19) 0.24 
 Sex −1.55 (1.21) 0.21 
Final multivariable model    
 Age 0.09 (0.04) 0.036 0.11 
 CMV status 2.52 (1.23) 0.044 

Bolded p values < 0.05.

Transplanted patients were followed longitudinally after transplant, providing an opportunity to examine the cell prevalence in individuals with restored renal function. Tracking CD57+PD1 CD4 T cells over time, we saw a small, transient, but statistically significant, decrease in the proportion of these cells immediately after transplantation, perhaps related to the induction immunotherapy used in transplantation, with rebound to the pretransplant level within 3 mo (Fig. 1A). A similar pattern was seen when examining only the T effector–memory CD45RA+ (Temra) subset of CD57+ PD1 CD4 T cells (Fig. 1B) and when surveying adhesion markers known to be highly expressed on CD57+PD1 CD4 T cells—CD2, LFA1, and VLA4 (14)—with a decrease immediately after transplant and rebound to pretransplant levels within 3 mo (Fig. 1D–F). We also noted a transient decrease in the proportion of CD28CD57+ cells (Fig. 1C).

FIGURE 1.

Clinical correlates of and changes to CD57+PD1 CD4 T cells after transplantation. (A) Percentage of CD57+PD1 CD4 T cells decrease initially and rebound after transplant. (B) Although the percentage of CD57+PD1 CD4 Temra did not change after transplant, the percentage of CD28 (C), CD2+ (D), LFA1+ (E), and VLA4+ (F) CD57+ CD4 T cells all transiently decreased at 1 mo and then rebounded. Note, lines in red are in the top 20% for a given marker. *p < 0.05.

FIGURE 1.

Clinical correlates of and changes to CD57+PD1 CD4 T cells after transplantation. (A) Percentage of CD57+PD1 CD4 T cells decrease initially and rebound after transplant. (B) Although the percentage of CD57+PD1 CD4 Temra did not change after transplant, the percentage of CD28 (C), CD2+ (D), LFA1+ (E), and VLA4+ (F) CD57+ CD4 T cells all transiently decreased at 1 mo and then rebounded. Note, lines in red are in the top 20% for a given marker. *p < 0.05.

Close modal

To evaluate the proliferative capacity of CD57+PD1 T cells, we sorted T cells into four subsets based on CD57/PD1 expression: CD57+PD1, CD57+PD1+, CD57PD1, and CD57PD1+. Focusing on CD57+PD1 cells, we showed that these cells were unable to proliferate with CD3 stimulation, in contrast to CD57PD1 cells, which proliferated well. However, these sorted CD57+PD1 CD4 T cells acquired proliferative capacity when CD3 stimulation was provided in the presence of bulk, unsorted PBMC (Fig. 2). This culture arrangement provided CD57+ cells access to a more physiologic environment with monocytes and other lymphocyte lineages and their attendant cell–cell and soluble supporting factors and allowed for the direct study of stimulated, proliferating CD57+PD1 CD4 T cells. Of note, these cells also proliferated when in coculture with bulk, unsorted PBMC when a semipermeable well insert was placed between the sorted population and the bulk cells in the presence of bead stimulation (Fig. 2B).

FIGURE 2.

CD57+ PD1 CD 4 T cells do not proliferate with mAb stimulation alone, but do in the presence of bulk unsorted PBMC. (A) Proliferation of CD57+PD1- or CD57PD1-sorted CD4 T cells with CD3 stimulation alone. (B) Proliferation of CD57+PD1- or CD57PD1-sorted CD4 T cells in the presence of bulk PBMC. Transwell condition separates bulk cells from sorted cells by a 1-μm Transwell pore. Proliferation measured by VPD. Bulk cells are excluded from these plots. Proliferating cells are a percent of VPD-positive cells.

FIGURE 2.

CD57+ PD1 CD 4 T cells do not proliferate with mAb stimulation alone, but do in the presence of bulk unsorted PBMC. (A) Proliferation of CD57+PD1- or CD57PD1-sorted CD4 T cells with CD3 stimulation alone. (B) Proliferation of CD57+PD1- or CD57PD1-sorted CD4 T cells in the presence of bulk PBMC. Transwell condition separates bulk cells from sorted cells by a 1-μm Transwell pore. Proliferation measured by VPD. Bulk cells are excluded from these plots. Proliferating cells are a percent of VPD-positive cells.

Close modal

We further examined the effect of immunosuppression on these cell subsets. Again, in the presence of our heterogenous unsorted PBMC (bulk cells), each CD57/PD1 fraction had proliferative capacity. All subsets proliferated to varying degrees with CD3 and CD3/CD28 stimulation and were inhibited variably by belatacept, rapamycin, and tacrolimus immunosuppression. Notably, the CD57PD1, CD57+PD1, and CD57PD1+ populations proliferated similarly, whereas CD57+PD1+ cells were most sensitive to immunosuppression, with even belatacept significantly inhibiting proliferation (Fig. 3A). These experiments demonstrated that many cell populations typically thought of as not having proliferative capacity are likely able to proliferate in more physiologically supportive environments.

FIGURE 3.

CD57+PD1 cells proliferate efficiently in the presence of bulk cells and are not inhibited by belatacept. This proliferation cannot be arrested by blocking of other clinically relevant cytokines. (A) Each panel shows proliferation of five unique responders, with the indicated CD57/PD1 phenotype sorted, VPD450 labeled, and returned to bulk culture. (B) Each panel shows proliferation of three unique responders, with indicated CD57/PD1 phenotype sorted, VPD450 labeled, and returned to bulk culture. n = 2 for CD57+PD1+ panel because of cell number limitations. Comparisons between individual unstimulated samples and stimulated samples (with or without immunosuppression) by Dunn posttest after a Friedman test between all groups. Significant (p < 0.05) posttests shown.

FIGURE 3.

CD57+PD1 cells proliferate efficiently in the presence of bulk cells and are not inhibited by belatacept. This proliferation cannot be arrested by blocking of other clinically relevant cytokines. (A) Each panel shows proliferation of five unique responders, with the indicated CD57/PD1 phenotype sorted, VPD450 labeled, and returned to bulk culture. (B) Each panel shows proliferation of three unique responders, with indicated CD57/PD1 phenotype sorted, VPD450 labeled, and returned to bulk culture. n = 2 for CD57+PD1+ panel because of cell number limitations. Comparisons between individual unstimulated samples and stimulated samples (with or without immunosuppression) by Dunn posttest after a Friedman test between all groups. Significant (p < 0.05) posttests shown.

Close modal

We further characterized the supernatants from these cultures and noted that several canonical proinflammatory cytokines (IFN-γ, TNF-α, IL-6, and IL-12) were present both in cultures without immunosuppression or with belatacept (data not shown). Therefore, we attempted to arrest proliferation of CD57+PD1 cells by using clinically available, therapeutic mAbs that block either TNF-α or IL-6 signaling. However, blocking these pathways either alone or in concert with belatacept did not consistently decrease the proliferative capacity of CD57+PD1 CD4 T cells (Fig. 3B)

We next turned our attention to analyzing the effects of stimulation on these sorted cell subsets, identifiable as a VPD-labeled cell fraction to distinguish them from bulk cells. Using unstimulated, nonimmunosuppressed conditions as a baseline, we calculated the absolute percent change in CD57, PD1, and CD28 expression of the sorted cells by subtracting the baseline percentage from the stimulated samples. CD57 and PD1 expression within each of the four sorted subsets changed substantially upon stimulation such that 85% of C57+PD1 cells moved to a different status after stimulation. The most frequent change in expression was the acquisition of PD1. This is consistent with what we observed longitudinally in vivo: although the frequency of CD57+PD1 cells changed early in the posttransplant course, it rebounded quickly, consistent with a cell subset with the potential to vary the expression of its surface markers. Changes in CD57 and PD1 expression were proportional to proliferation (Supplemental Fig. 2), and when controlling for immunosuppression, were greater with rapamycin in the CD57+PD1 and CD57PD1+ subsets.

We next analyzed change in CD57, PD1, and CD28 status, and proliferating cells using PCA in unstimulated and CD3 stimulated samples. We first examined the characteristics of the condition represented by PCA. As shown in Fig. 4A, cells segregated both by stimulation (with CD3-stimulated cells assigned higher values on component 1) and by CD57 status (across component 2). We next examined how changes in surface marker expression were captured by the PCA. As depicted in Fig. 4B, we saw that there was more proliferation at higher values of component 1. Fig. 4C shows that CD57+ and CD57 cells diverge with proliferation. Following stimulation, CD57+ tend to become CD57, whereas CD57 cells tend to become CD57+ (albeit at a much lower rate). Both CD57+ and CD57 cells gained PD1 at high levels of proliferation. With proliferation, CD57+ cells tended to stay CD57+, and CD57 cells tended to stay CD57. That is, both proliferation and stability of CD57 expression are positively associated with PD1 acquisition.

FIGURE 4.

PCA plots examining for change in phenotype with proliferation. (A) Cells cluster by stimulation (no stimulation versus CD3) and CD57 status. (B) PCA plot in which the coloring is varied by the percentage (%) of proliferating in each sample. As samples increase in proliferation, they move to the right on component 1. (C) Change in CD57/PD1 status is visualized on the PCA, with coloring varying with the absolute percentage change in the cell phenotype. CD57+ and CD57 cells have distinct patterns of phenotype change that become more divergent with more proliferation.

FIGURE 4.

PCA plots examining for change in phenotype with proliferation. (A) Cells cluster by stimulation (no stimulation versus CD3) and CD57 status. (B) PCA plot in which the coloring is varied by the percentage (%) of proliferating in each sample. As samples increase in proliferation, they move to the right on component 1. (C) Change in CD57/PD1 status is visualized on the PCA, with coloring varying with the absolute percentage change in the cell phenotype. CD57+ and CD57 cells have distinct patterns of phenotype change that become more divergent with more proliferation.

Close modal

We also performed linear regressions to directly examine the change in CD57, PD1, and CD28 expression among the four sorted subsets (Supplemental Fig. 3). These data confirmed that in all four sorted cell populations, we saw an increase in PD1+ cells with increased proliferation. In multivariable regression, which included immunosuppression, rapamycin was independently associated with an increase in PD1+ cell frequency in both the CD57+PD1 and the CD57PD1+ subsets (Fig. 5, Supplemental Table I), meaning the frequency of PD1+ cells was greater than would be expected for a given degree of proliferation with rapamycin.

FIGURE 5.

Rapamycin enhances acquisition of PD1 among certain sorted subsets. Both CD57+PD1 (A) and CD57PD1+ (B) cells have increased acquisition of PD1 under rapamycin relative to their level of proliferation. This is shown by the rapamycin-treated samples (red markers), appearing above the linear regression line. Linear regression with 95% confidence interval (CI) plotted. Unadjusted R2 and overall p value shown.

FIGURE 5.

Rapamycin enhances acquisition of PD1 among certain sorted subsets. Both CD57+PD1 (A) and CD57PD1+ (B) cells have increased acquisition of PD1 under rapamycin relative to their level of proliferation. This is shown by the rapamycin-treated samples (red markers), appearing above the linear regression line. Linear regression with 95% confidence interval (CI) plotted. Unadjusted R2 and overall p value shown.

Close modal

CoBRR remains a significant hurdle to the widespread adoption of CoB-based immunosuppression regimens. One population of cells thought to mediate CoBRR is CD57+PD1 CD4 T cells (14). In the current study, we demonstrate that this cellular phenotype is prevalent among ESRD patients, is associated with age and CMV status, and persists over time after transplantation. This persistence is relevant to the clinical practice of late conversion from a CNI to belatacept and suggests why conversion within the first 6 mo after transplantation is often unsuccessful (18).

CD57+ CD4 T cells have been studied most extensively in CMV and HIV (19). Similar to these prior studies, we note that increasing age and CMV status are associated with an expansion of CD57+PD1 CD4 T cells. In HIV, there is a differential expansion of CD57+ CD4 T cells after viral acquisition, which correlates with disease control (20). Although transplantation is a major immunologic event, we find few differences in the frequency of the CD57+ PD1 CD4 T cell population after transplantation, except transiently during early time points of higher dose induction immunosuppression. This suggests that the factors necessary to create this pool of cells are likely established before transplantation and are not modified by events associated with transplantation, such as restoration of renal function. We posit that CD57+PD1 CD4 T cells are not dependent on uremia, dialysis, or other proinflammatory events seen in transplant recipients, but rather are exemplary of an intrinsic quality of a patient’s immune system, such as prior viral exposure.

CD57+ CD4 T cells are known to have a very low proliferative capacity in vitro (21). Indeed, we find that CD57+PD1 CD4 T cells are unable to proliferate when isolated in culture. However, we show that CD57+ PD1 CD4 T cells are able to proliferate well with anti-CD3 stimulation when that stimulation occurs in the presence of unsorted cells, as would be available in vivo. The requirement for this heterogenous population of unsorted PBMC to aid in the proliferation of CD57+ T cells suggests that the proliferative capacity of these cells has perhaps been underestimated. Additionally, we show that cell–cell contact is not necessary for this proliferative capacity and that it does not depend exclusively on certain canonical proinflammatory cytokines (TNF-α and IL-6), two pathways that could readily by inhibited in the clinic. Further investigation to determine the factors necessary for CD57+ cell stimulation is ongoing.

Although PD1 is most well known as a marker of T cell exhaustion (22), it has also been recognized as a marker of activation (23, 24). Regardless, chronically high levels of PD1 have been shown to weaken antiviral responses (22, 25) and similarly been shown to be critical in facilitating alloantigen-specific, activation-induced cell death (AICD) in experimental transplant models (26, 27). PD1 ligand is also known to be upregulated in transplanted allografts, and PD1–PD1 ligand interactions have been shown to be critical in eliminating alloantigen-specific cells from the allograft and facilitating allospecific tolerance (28, 29). That rapamycin facilitates the expression of PD1 and tacrolimus does not is of more than academic interest, as these two agents have both been used as clinical adjuvants for use with belatacept (18, 30, 31). The opposing effects of rapamycin and tacrolimus with respect to PD1 expression suggest a mechanism by which CD57+CD4 T cells are gradually eliminated from the posttransplant repertoire in the presence of rapamycin-treated (32), but not CNI-treated, patients.

Clinically, patients treated with CoB-based regimens have superior renal function to those on traditional CNI-based regimens. Indeed, this advantage is evident despite CoBRR episodes, and patients treated with CoB are less likely to develop donor-specific Ab regardless of this proclivity to early acute rejection (6, 9). In contrast, early rejection episodes under CNI-based regimens have been shown to correlate with graft loss and alloantibody formation (33). In the current study, we show that proliferating CD57+ PD1 CD4 T cells rapidly acquire PD1 with proliferation. We posit that this property makes responding T cells more likely to undergo AICD and less likely to sustain a donor-specific response. This drive toward a characteristic exhausted phenotype may, in part, explain why patients treated with CoB who experience rejection do not have worse graft outcomes and why early CoBRR is typically short lived. Indeed, previous murine experiments showed increases in AICD of allospecific T will in mice treated with CoB and rapamycin compared with CoB alone or CoB with CNI (26, 27), and preclinical studies have favored a combination of rapamycin with belatacept (34).

Additional evidence for the hypothesis that PD1 acquisition by CD57+ CD4 T cells decreases their function is found in the HIV literature. In one study, the degree of expansion of CD57+ CD4 T cells is inversely correlated with viral load in HIV elite controllers (individuals who have become infected but maintain very low viral loads for a long period of time). Moreover, these cells augment the ability of cytotoxic CD8 T cells to inhibit viral replication in vitro (20). More recently, it has been demonstrated that CD57+ CD4 T cells from elite controllers are more polyfunctional and less likely to express PD1 than those from noncontrolling, HIV-infected individuals (35). This suggests that PD1 acquisition among CD57+ CD4 T cells is deleterious to their function.

We also noted that the acquisition of PD1 is enhanced when cells proliferate in the presence of rapamycin. Rapamycin is an inhibitor of mTOR and has myriad effects on cellular metabolism, cell cycle, immune activation, and protein synthesis (36). Few studies have specifically examined the effects of rapamycin on PD1 expression, and little is known, especially in the context of stimulation (37). Previously, we have shown that a regimen of rapamycin and belatacept, in conjunction with induction using alemtuzumab (a CD52-specific mAb), has excellent results with little rejection and many patients able to be weaned to belatacept alone therapy (31, 38). These data support our finding that rapamycin treatment may be synergistic with CoB, either through elimination of CoB-resistant cells or rendering these cells more senescent/sensitive to apoptosis through PD1 acquisition (24).

As transplant recipients experience cancer at a higher rate than the general population experience with PD1 blockade to treat malignancy in transplant recipients is growing (39). A recent case report showed successful treatment of melanoma in a kidney transplant recipient with pembrolizumab and rapamycin as maintenance immunosuppression after a rejection episode. Similar to our results, the authors in this study also showed that rapamycin allowed for activation of CD4 T cells (as determined by CD25 expression and IFN-γ secretion), but not proliferation. Further, rapamycin prevented rejection after a single rejection episode was treated with steroids (40).

In sum, CD57+PD1 CD4 T cells may represent one of many possible phenotypes that are situated between naivete and exhaustion, allowing for costimulation independence while maintaining effector function. Moreover, these cells require assistance from other cell types for proliferation, suggesting that bystander effects from local inflammation—related to transplantation itself or infection—facilitate their allospecific activity and potentiate subsequent rejection episodes. This phenomenon has been anecdotally observed in CoB-sensitive patients off of all immunosuppression who reject in the setting of new viral infections (31).

Our study has several important limitations. First, aspects of it are performed in vitro on sorted cell populations. This may not accurately recapitulate the complex interface of the allograft and recipient immune system in vivo. However, these conditions give the greatest precision to test specific hypotheses. Additionally, our cultures of CD57+PD1 CD4 T cells with bulk unsorted PBMC do not allow analysis of the exact proliferative milieu necessary for CD57+PD1 CD4 T cells but rather demonstrate how these cells respond in a generally replete inflammatory environment. Clinical experience suggests, however, that many different inflammatory states may be of importance with regards to CoB, and therefore, it is reasonable to treat this as an exemplary condition. Moreover, this arrangement allows for the examination of the proliferative behavior of otherwise nonproliferating cell subsets. Additionally, stimulation with CD3/CD28 is different from stimulation with alloantigen. Finally, CD57+PD1 T cells are an uncommon cell population that limits their study and makes examining multiple replicates more difficult.

In conclusion, we present evidence that CD57+PD1 CD4 T cells persist over time after transplantation and that this population continues to have a vital proinflammatory phenotype inconsistent with terminal senescence. Additionally, we show that CD57+PD1 CD4 T cells are replication incompetent in the absence of bulk unsorted PBMC but, with help, are able to proliferate in the presence of belatacept. Finally, we show that CD57+PD1 CD4 T cells acquire PD1 with proliferation and that this acquisition is enhanced with rapamycin treatment. These data provide a mechanistic basis for the clinical benefit observed with the coadministration of belatacept and rapamycin.

We thank the patients for generously donated samples (Duke University) for this study, the Substrate Services Core Research Support facility for processing samples, and the Surgical Center for Outcomes Research (Duke University) for supplying clinical data

This work was supported by a grant from the U.S. Food and Drug Administration, National Institutes of Health R01 FD003539 and an award from Bristol-Myers Squibb.

The online version of this article contains supplemental material.

Abbreviations used in this article:

AICD

activation-induced cell death

CNI

calcineurin inhibitor

CoB

costimulation blockade

CoBRR

CoB-resistant rejection

ESRD

end-stage renal disease

PCA

principal component analysis

VPD

Violet Proliferation Dye.

1
Axelrod
,
D. A.
,
M. A.
Schnitzler
,
H.
Xiao
,
W.
Irish
,
E.
Tuttle-Newhall
,
S.-H.
Chang
,
B. L.
Kasiske
,
T.
Alhamad
,
K. L.
Lentine
.
2018
.
An economic assessment of contemporary kidney transplant practice.
Am. J. Transplant.
18
:
1168
1176
.
2
Naesens
,
M.
,
D. R. J.
Kuypers
,
M.
Sarwal
.
2009
.
Calcineurin inhibitor nephrotoxicity.
Clin. J. Am. Soc. Nephrol.
4
:
481
508
.
3
Tavira
,
B.
,
C.
Díaz-Corte
,
D.
Coronel
,
F.
Ortega
,
E.
Coto
.
2014
.
[Pharmacogenetics of tacrolimus: from bench to bedside?].
Nefrologia
34
:
11
17
.
4
Sprangers
,
B.
,
V.
Nair
,
V.
Launay-Vacher
,
L. V.
Riella
,
K. D.
Jhaveri
.
2018
.
Risk factors associated with post-kidney transplant malignancies: an article from the Cancer-Kidney International Network.
Clin. Kidney J.
11
:
315
329
.
5
Vincenti
,
F.
,
C.
Larsen
,
A.
Durrbach
,
T.
Wekerle
,
B.
Nashan
,
G.
Blancho
,
P.
Lang
,
J.
Grinyo
,
P. F.
Halloran
,
K.
Solez
, et al
Belatacept Study Group
.
2005
.
Costimulation blockade with belatacept in renal transplantation.
N. Engl. J. Med.
353
:
770
781
.
6
Vincenti
,
F.
,
L.
Rostaing
,
J.
Grinyo
,
K.
Rice
,
S.
Steinberg
,
L.
Gaite
,
M.-C.
Moal
,
G. A.
Mondragon-Ramirez
,
J.
Kothari
,
M. S.
Polinsky
, et al
.
2016
.
Belatacept and long-term outcomes in kidney transplantation.
N. Engl. J. Med.
374
:
333
343
.
7
Linsley
,
P. S.
,
P. M.
Wallace
,
J.
Johnson
,
M. G.
Gibson
,
J. L.
Greene
,
J. A.
Ledbetter
,
C.
Singh
,
M. A.
Tepper
.
1992
.
Immunosuppression in vivo by a soluble form of the CTLA-4 T cell activation molecule.
Science
257
:
792
795
.
8
Larsen
,
C. P.
,
T. C.
Pearson
,
A. B.
Adams
,
P.
Tso
,
N.
Shirasugi
,
E.
Strobert
,
D.
Anderson
,
S.
Cowan
,
K.
Price
,
J.
Naemura
, et al
.
2005
.
Rational development of LEA29Y (belatacept), a high-affinity variant of CTLA4-Ig with potent immunosuppressive properties.
Am. J. Transplant.
5
:
443
453
.
9
Vincenti
,
F.
,
B.
Charpentier
,
Y.
Vanrenterghem
,
L.
Rostaing
,
B.
Bresnahan
,
P.
Darji
,
P.
Massari
,
G. A.
Mondragon-Ramirez
,
M.
Agarwal
,
G.
Di Russo
, et al
.
2010
.
A phase III study of belatacept-based immunosuppression regimens versus cyclosporine in renal transplant recipients (BENEFIT study).
Am. J. Transplant.
10
:
535
546
.
10
Chen
,
L.
,
D. B.
Flies
.
2013
.
Molecular mechanisms of T cell co-stimulation and co-inhibition [Published erratum appears in 2013 Nat. Rev. Immunol. 13: 542.].
Nat. Rev. Immunol.
13
:
227
242
.
11
Mou
,
D.
,
J.
Espinosa
,
D. J.
Lo
,
A. D.
Kirk
.
2014
.
CD28 negative T cells: is their loss our gain?
Am. J. Transplant.
14
:
2460
2466
.
12
Zeng
,
H.
,
H.
Chi
.
2017
.
mTOR signaling in the differentiation and function of regulatory and effector T cells.
Curr. Opin. Immunol.
46
:
103
111
.
13
Larbi
,
A.
,
T.
Fulop
.
2014
.
From “truly naïve” to “exhausted senescent” T cells: when markers predict functionality.
Cytometry A
85
:
25
35
.
14
Espinosa
,
J.
,
F.
Herr
,
G.
Tharp
,
S.
Bosinger
,
M.
Song
,
A. B.
Farris
III
,
R.
George
,
J.
Cheeseman
,
L.
Stempora
,
R.
Townsend
, et al
.
2016
.
CD57(+) CD4 T cells underlie belatacept-resistant allograft rejection.
Am. J. Transplant.
16
:
1102
1112
.
15
George
,
R. P.
,
A. K.
Mehta
,
S. D.
Perez
,
P.
Winterberg
,
J.
Cheeseman
,
B.
Johnson
,
J.
Kwun
,
S.
Monday
,
L.
Stempora
,
B.
Warshaw
,
A. D.
Kirk
.
2017
.
Premature T cell senescence in pediatric CKD.
J. Am. Soc. Nephrol.
28
:
359
367
.
16
Mathews
,
D. V.
,
W. C.
Wakwe
,
S. C.
Kim
,
M. C.
Lowe
,
C.
Breeden
,
M. E.
Roberts
,
A. B.
Farris
,
E. A.
Strobert
,
J. B.
Jenkins
,
C. P.
Larsen
, et al
.
2017
.
Belatacept-resistant rejection is associated with CD28+ memory CD8 T cells.
Am. J. Transplant.
17
:
2285
2299
.
17
Cortes-Cerisuelo
,
M.
,
S. J.
Laurie
,
D. V.
Mathews
,
P. D.
Winterberg
,
C. P.
Larsen
,
A. B.
Adams
,
M. L.
Ford
.
2017
.
Increased pretransplant frequency of CD28+ CD4+ TEM predicts belatacept-resistant rejection in human renal transplant recipients.
Am. J. Transplant.
17
:
2350
2362
.
18
Adams
,
A. B.
,
J.
Goldstein
,
C.
Garrett
,
R.
Zhang
,
R. E.
Patzer
,
K. A.
Newell
,
N. A.
Turgeon
,
A. S.
Chami
,
A.
Guasch
,
A. D.
Kirk
, et al
.
2017
.
Belatacept combined with transient calcineurin inhibitor therapy prevents rejection and promotes improved long-term renal allograft function.
Am. J. Transplant.
17
:
2922
2936
.
19
Booiman
,
T.
,
F. W.
Wit
,
A. F.
Girigorie
,
I.
Maurer
,
D.
De Francesco
,
C. A.
Sabin
,
A. M.
Harskamp
,
M.
Prins
,
C.
Franceschi
,
S. G.
Deeks
, et al
Co-morBidity in Relation to Aids (COBRA) Collaboration
.
2017
.
Terminal differentiation of T cells is strongly associated with CMV infection and increased in HIV-positive individuals on ART and lifestyle matched controls.
PLoS One
12
: e0183357.
20
Johnson
,
S.
,
M.
Eller
,
J. E.
Teigler
,
S. M.
Maloveste
,
B. T.
Schultz
,
D. Z.
Soghoian
,
R.
Lu
,
A. F.
Oster
,
A.-L.
Chenine
,
G.
Alter
, et al
.
2015
.
Cooperativity of HIV-specific cytolytic CD4 T cells and CD8 T cells in control of HIV viremia.
J. Virol.
89
:
7494
7505
.
21
Palmer
,
B. E.
,
N.
Blyveis
,
A. P.
Fontenot
,
C. C.
Wilson
.
2005
.
Functional and phenotypic characterization of CD57+CD4+ T cells and their association with HIV-1-induced T cell dysfunction.
J. Immunol.
175
:
8415
8423
.
22
Barber
,
D. L.
,
E. J.
Wherry
,
D.
Masopust
,
B.
Zhu
,
J. P.
Allison
,
A. H.
Sharpe
,
G. J.
Freeman
,
R.
Ahmed
.
2006
.
Restoring function in exhausted CD8 T cells during chronic viral infection.
Nature
439
:
682
687
.
23
Hong
,
J. J.
,
P. K.
Amancha
,
K.
Rogers
,
A. A.
Ansari
,
F.
Villinger
.
2013
.
Re-evaluation of PD-1 expression by T cells as a marker for immune exhaustion during SIV infection.
PLoS One
8
: e60186.
24
Petrovas
,
C.
,
J. P.
Casazza
,
J. M.
Brenchley
,
D. A.
Price
,
E.
Gostick
,
W. C.
Adams
,
M. L.
Precopio
,
T.
Schacker
,
M.
Roederer
,
D. C.
Douek
,
R. A.
Koup
.
2006
.
PD-1 is a regulator of virus-specific CD8+ T cell survival in HIV infection.
J. Exp. Med.
203
:
2281
2292
.
25
Radziewicz
,
H.
,
C. C.
Ibegbu
,
M. L.
Fernandez
,
K. A.
Workowski
,
K.
Obideen
,
M.
Wehbi
,
H. L.
Hanson
,
J. P.
Steinberg
,
D.
Masopust
,
E. J.
Wherry
, et al
.
2007
.
Liver-infiltrating lymphocytes in chronic human hepatitis C virus infection display an exhausted phenotype with high levels of PD-1 and low levels of CD127 expression.
J. Virol.
81
:
2545
2553
.
26
Li
,
Y.
,
X. C.
Li
,
X. X.
Zheng
,
A. D.
Wells
,
L. A.
Turka
,
T. B.
Strom
.
1999
.
Blocking both signal 1 and signal 2 of T-cell activation prevents apoptosis of alloreactive T cells and induction of peripheral allograft tolerance.
Nat. Med.
5
:
1298
1302
.
27
Wells
,
A. D.
,
X. C.
Li
,
Y.
Li
,
M. C.
Walsh
,
X. X.
Zheng
,
Z.
Wu
,
G.
Nuñez
,
A.
Tang
,
M.
Sayegh
,
W. W.
Hancock
, et al
.
1999
.
Requirement for T-cell apoptosis in the induction of peripheral transplantation tolerance.
Nat. Med.
5
:
1303
1307
.
28
Wang
,
L.
,
R.
Han
,
W. W.
Hancock
.
2007
.
Programmed cell death 1 (PD-1) and its ligand PD-L1 are required for allograft tolerance.
Eur. J. Immunol.
37
:
2983
2990
.
29
Keir
,
M. E.
,
S. C.
Liang
,
I.
Guleria
,
Y. E.
Latchman
,
A.
Qipo
,
L. A.
Albacker
,
M.
Koulmanda
,
G. J.
Freeman
,
M. H.
Sayegh
,
A. H.
Sharpe
.
2006
.
Tissue expression of PD-L1 mediates peripheral T cell tolerance.
J. Exp. Med.
203
:
883
895
.
30
Ferguson
,
R.
,
J.
Grinyó
,
F.
Vincenti
,
D. B.
Kaufman
,
E. S.
Woodle
,
B. A.
Marder
,
F.
Citterio
,
W. H.
Marks
,
M.
Agarwal
,
D.
Wu
, et al
.
2011
.
Immunosuppression with belatacept-based, corticosteroid-avoiding regimens in de novo kidney transplant recipients.
Am. J. Transplant.
11
:
66
76
.
31
Kirk
,
A. D.
,
A.
Guasch
,
H.
Xu
,
J.
Cheeseman
,
S. I.
Mead
,
A.
Ghali
,
A. K.
Mehta
,
D.
Wu
,
H.
Gebel
,
R.
Bray
, et al
.
2014
.
Renal transplantation using belatacept without maintenance steroids or calcineurin inhibitors.
Am. J. Transplant.
14
:
1142
1151
.
32
Xu
,
H.
,
V. A.
Bendersky
,
T. V.
Brennan
,
J. R.
Espinosa
,
A. D.
Kirk
.
2018
.
IL-7 receptor heterogeneity as a mechanism for repertoire change during postdepletional homeostatic proliferation and its relation to costimulation blockade-resistant rejection.
Am. J. Transplant.
18
:
720
730
.
33
Loupy
,
A.
,
D.
Vernerey
,
C.
Tinel
,
O.
Aubert
,
J.-P.
Duong van Huyen
,
M.
Rabant
,
J.
Verine
,
D.
Nochy
,
J.-P.
Empana
,
F.
Martinez
, et al
.
2015
.
Subclinical rejection phenotypes at 1 year post-transplant and outcome of kidney allografts.
J. Am. Soc. Nephrol.
26
:
1721
1731
.
34
Lo
,
D. J.
,
D. J.
Anderson
,
T. A.
Weaver
,
F.
Leopardi
,
M.
Song
,
A. B.
Farris
,
E. A.
Strobert
,
J.
Jenkins
,
N. A.
Turgeon
,
A. K.
Mehta
, et al
.
2013
.
Belatacept and sirolimus prolong nonhuman primate renal allograft survival without a requirement for memory T cell depletion.
Am. J. Transplant.
13
:
320
328
.
35
Phetsouphanh
,
C.
,
D.
Aldridge
,
E.
Marchi
,
C. M. L.
Munier
,
J.
Meyerowitz
,
L.
Murray
,
C.
Van Vuuren
,
D.
Goedhals
,
S.
Fidler
,
A.
Kelleher
, et al
.
2019
.
Maintenance of functional CD57+ cytolytic CD4+ T cells in HIV+ elite controllers.
Front. Immunol.
10
:
1844
.
36
Zaza
,
G.
,
S.
Granata
,
C.
Caletti
,
L.
Signorini
,
G.
Stallone
,
A.
Lupo
.
2018
.
mTOR inhibition role in cellular mechanisms.
Transplantation
102
(
Suppl. 1
):
S3
S16
.
37
Hurez
,
V.
,
V.
Dao
,
A.
Liu
,
S.
Pandeswara
,
J.
Gelfond
,
L.
Sun
,
M.
Bergman
,
C. J.
Orihuela
,
V.
Galvan
,
Á.
Padrón
, et al
.
2015
.
Chronic mTOR inhibition in mice with rapamycin alters T, B, myeloid, and innate lymphoid cells and gut flora and prolongs life of immune-deficient mice.
Aging Cell
14
:
945
956
.
38
Schmitz
,
R.
,
Z. W.
Fitch
,
H.
Xu
,
A.
Ghali
,
A. K.
Mehta
,
A.
Guasch
,
A. D.
Kirk
.
2020
.
Kidney transplantation using alemtuzumab, belatacept, and sirolimus: five-year follow-up.
Am. J. Transplant.
20
:
3609
3619
.
39
Kumar
,
V.
,
A. B.
Shinagare
,
H. G.
Rennke
,
S.
Ghai
,
J. H.
Lorch
,
P. A.
Ott
,
O. E.
Rahma
.
2020
.
The safety and efficacy of checkpoint inhibitors in transplant recipients: a case series and systematic review of literature.
Oncologist
25
:
505
514
.
40
Esfahani
,
K.
,
T.-A.
Al-Aubodah
,
P.
Thebault
,
R.
Lapointe
,
M.
Hudson
,
N. A.
Johnson
,
D.
Baran
,
N.
Bhulaiga
,
T.
Takano
,
J.-F.
Cailhier
, et al
.
2019
.
Targeting the mTOR pathway uncouples the efficacy and toxicity of PD-1 blockade in renal transplantation.
Nat. Commun.
10
:
4712
4719
.

The authors have no financial conflicts of interest.

Supplementary data