Panel of Reactive T Cells as a Measurement of Primed Cellular Alloimmunity in Kidney Transplant Candidates

Emilio D. Poggio,* Michael Clemente,† Donald E. Hricik,‡ and Peter S. Heeger† Departments of *Nephrology and Hypertension and †, Cleveland Clinic Foundation, and ‡Division of Nephrology and Hypertension, Department of Medicine, University Hospitals of Cleveland, Cleveland, Ohio

Pretransplantation panel reactive (PRA) testing assesses posttransplantation risk for antibody-mediated graft injury. It was postulated analogously that screening for effector/memory alloreactive T cells by “panel of reactive T cells” (PRT) using IFN-␥ enzyme-linked immunosorbent spot assays would evaluate independently cellular alloimmunity in transplant candi- dates. Peripheral blood from 41 hemodialysis patients who were awaiting first renal transplants were tested against a panel of allogeneic stimulator cells. Positive assays were defined arbitrarily as >25 spots/300,000 peripheral blood lymphocytes, and positive PRT was defined as when the responder reacted to 40 or 75% (PRT-75) of the stimulators. Seventeen percent of patients were PRT-75؉, whereas 32% were PRA؉. Twelve percent of the cohort was PRT-75؉/PRA؊, and only 5% of the patients were PRA؉/PRT-75؉, indicating that alloreactivity did not routinely imply sensitization and vice versa. PRT-75؉ patients were more likely to be younger (<55 yr) and black. In contrast, a positive PRA was significantly associated with female gender but not race or age. Pretransplantation screening of cellular alloimmunity by enzyme-linked immunosorbent spot–based PRT detects a subset of hemodialysis patients who differ from those that are PRA؉. Preliminary correlations with posttransplantation outcome in seven recipients suggest that PRT screening has the potential to aid in risk assessment in renal transplant candidates. J Am Soc Nephrol 17: 564–572, 2006. doi: 10.1681/ASN.2005030293

idney transplant physicians use a number of pretrans- sessment of humoral rather than cellular alloimmunity as a risk plantation tests to assess the risk for posttransplanta- factor for posttransplantation outcome. Because T cells are K tion rejection and graft loss (1,2). One such approach, known central mediators of graft injury (6–12), we hypothe- HLA matching of donor and recipient at the A, B, and DR loci, sized that enumerating the frequency and the breadth of reac- has had a marked influence on the outcome of renal trans- tivity of activated antidonor T cells in the recipient before plants. Grafts from deceased donors that are matched at all six transplantation would provide information regarding the pa- loci represent such a significant benefit to the recipient that the tient’s alloimmune repertoire that is supplemental to alloanti- United Network for Organ Sharing has adopted a national body detection. We originally described an adaptation of the organ-sharing policy to optimize the use of these organs (3). enzyme-linked immunosorbent spot (ELISPOT) assay In addition to HLA matching, renal transplant candidates all to measure activated/memory alloreactive T cells in humans are routinely screened for alloantibodies that are present in (13), and we provided strong evidence that the frequency of their serum (panel reactive [PRA]) and for antidonor these T cells pretransplantation was associated directly with alloantibodies at the time of transplantation (final cross-match) worse posttransplantation outcome (14). Other groups have (4). Newer flow cytometry-based techniques for detecting al- confirmed these findings, and pretransplantation assessments loantibody have reinforced the findings that sensitized patients of antidonor T cell are now beginning to influence with strongly positive PRA and/or positive antidonor alloan- clinical decision making in a manner analogous to final cross- tibodies are at higher risk for worse posttransplantation out- match testing for alloantibodies (11,15–17). comes. The association between a positive cross-match and hyper- As another approach to risk stratification for transplant can- acute rejection (5), as well as the readily available methods for didates, we postulated that determining the frequency of pe- measuring alloantibodies (reviewed in reference [4]), has led ripheral T cells that are reactive to a panel of alloantigens the transplant community to focus on pretransplantation as- would provide a measure of the strength and reactivity of the pre-existing alloreactive T cell repertoire analogous to how PRA testing functions as a measure of pretransplantation hu- Received March 17, 2005. Accepted November 3, 2005. moral allosensitization. If this hypothesis proves to be correct Published online ahead of print. Publication date available at www.jasn.org. and if cellular alloreactivity measured by cytokine ELISPOT detects a different population of high-risk patients from those D.E.H. and P.S.H. share senior authorship. with alloantibodies, then assessing a candidate’s cellular re- Address correspondence to: Dr. Peter S. Heeger, Department of Immunology, The Cleveland Clinic Foundation, NB30, 9500 Euclid Avenue, Cleveland, OH sponse to a panel of allogeneic stimulators (panel of reactive T 44195,. Phone: 216-445-9356; Fax: 216-444-8372; E-mail: [email protected] cells [PRT]) could be used as an adjunctive test for assessing

Copyright © 2006 by the American Society of Nephrology ISSN: 1046-6673/1702-0564 J Am Soc Nephrol 17: 564–572, 2006 Primed T Cell Alloreactivity in Transplant Candidates 565 risk, choosing optimal donors, deciding on immunosuppres- incubated overnight at 4°C. After an additional wash, streptavidin sion regimens, and ultimately improving outcome. horseradish peroxidase conjugate (Dako, Carpinteria, CA) was added for1hatroom temperature. After a final wash, the plates were developed with 3-amino-9-ethylcarbazole (10 mg/ml in N,N-dimethyl- Materials and Methods formamide; Sigma), prepared in 0.1 M sodium acetate buffer (pH 5.0) Study Patients ␮ mixed with H2O2 (200 l/well). The study population consisted of 41 adult hemodialysis (HD) pa- The resulting spots were counted with a Series 1 Immunospot com- tients who were on the active waiting list for renal transplantation at puter-assisted ELISPOT image analyzer (Cellular Technology, Cleve- The Cleveland Clinic Foundation or University Hospitals of Cleveland. land, OH). Results were depicted as the mean number of IFN-␥ spots Recipients of previous solid organ transplants including kidneys were per 300,000 recipient PBL on the basis of duplicate or triplicate mea- excluded from the study. The cause of ESRD in this cohort was diabetes surements in a given assay. The frequencies of alloreactive IFN-␥ (n ϭ 23), hypertension (n ϭ 7), glomerular disease (n ϭ 7), polycystic ELISPOT that were presented and used for analyses were obtained disease (n ϭ 2), and other/unknown (n ϭ 2). None of the recipients was after those that were derived from nonstimulated wells (“background”) taking immunosuppressive drugs for their primary renal disease dur- were subtracted. Previous work has demonstrated that Ͻ10 to 15 spots ing the period of the study. per 300,000 cells represent background reactivity and that Ͼ25 spots Stimulator peripheral blood lymphocytes were obtained from six per 300,000 cells functioned as a threshold that correlated with an normal healthy (based on a health questionnaire) volunteers, and stim- increased risk for graft injury posttransplantation (10,12,13,17). We ulator spleen cells were obtained from an additional two randomly therefore chose Ͼ25 spots per 300,000 PBL as an arbitrary threshold to chosen deceased donors. Three living donors and one deceased donor define a positive assay. were self-described as white, and three living donors and one deceased donor were self-described as black. Tissue typing for A, B, and DR Alloantibody Detection alleles was performed by standard molecular techniques for all study Plasma anti-HLA antibody was determined by flow cytometry using patients. All study patients were enrolled under the approved guide- HLA class I and class II -coated latex beads (FlowPRA Screening lines of the Institutional Review Boards for Human Studies at The Test; One Lambda, Inc., Canoga Park, CA) and then verified by assess- Cleveland Clinic Foundation and University Hospitals of Cleveland. ing specificity using a panel of HLA-typed specificity beads (One Lambda). The antibody testing was performed according to instruc- Sample Preparation tions supplied by the manufacturer. A positive PRA was defined as Peripheral blood samples were obtained in heparinized tubes from Ͼ10% reactivity (sensitized individuals), and highly sensitized patients HD patients just before the initiation of a dialysis treatment. Aliquots of were defined as having Ͼ50% reactivity. plasma were stored at Ϫ70°C and used for alloantibody testing (see below). Peripheral blood lymphocytes (PBL) were isolated by standard Statistical Analyses Ficoll density-gradient centrifugation, and live cells were counted using All analyses were performed using SPSS version 11.5. Values are acridine orange/ethidium bromide staining and immunofluorescence shown as mean Ϯ SD, median (range), or percentage. Categorical microscopy. The isolated PBL were used as responders in the ELISPOT variables were compared using the ␹2 test or Fisher exact test when assays as cited below. Stimulator PBL from normal volunteers and appropriate. Comparison of mean values was tested using the t test for stimulator spleen cells from deceased donors were isolated by Ficoll independent samples (two-tailed) and ANOVA, and median compari- density separation, depleted of T cells (RosetteSep; StemCell Technol- son was done using the Mann-Whitney U test. P Ͻ 0.05 was considered ogies Inc, Vancouver, BC, Canada) to eliminate their ability to produce statistically significant. IFN-␥ and frozen in aliquots to be used in ELISPOT assays as described (12,13). Results Patient Characteristics ELISPOT Assays The responder study population consisted of 41 adult pa- IFN-␥ ELISPOT assays were performed as described previously in tients who had ESRD and were on HD and awaiting renal detail (12–14). A total of 300,000 responder PBL in 200 ␮l of medium transplantation (Table 1). Mean time on dialysis was 32 mo. The (RPMI 1640 medium; Cambrex, Walkersville, MD) plus 10% human serum, with l-glutamine (Invitrogen, Grand Island, NY), penicillin/ streptomycin (Cambrex), and 50 mM 2-mercaptoethanol (Sigma, St. Louis, MO) were placed in 96-well ELISPOT plates (Millipore, Bedford, Table 1. Patient characteristicsa MA) that were precoated with capture anti–IFN-␥ antibody (BD Pharm- ingen, San Diego, CA). The cells were tested in duplicate or triplicate Total no. of patients 41 against medium alone (negative control), the panel of T cell–depleted Age (yr)b 56 Ϯ 10 (32 to 72) stimulator cells, and phytohemagglutinin (positive control) at 1 ␮g/ml Black race (n ͓%͔) 22 (53.6) medium (Sigma). Because we used unfractionated PBL, which include Female gender (n ͓%͔) 15 (36.6) antigen-presenting cells as well as T cells as responders, the assay Time on hemodialysis (mo)b 32 Ϯ 19 (1 to 84) design evaluates total alloreactivity; both the direct (recipient T cells HLA mismatches versus stimulatorsc 5(4to6) reactive to donor HLA) and indirect (recipient T cells reacting to donor No. of patients with positive PRAd 13 (33.3) processed and presented by recipient antigen-presenting cells ͓ ͔ in the ELISPOT wells) allorecognition pathways are assessed together. (n % ) All study patients were tested against the same panel of stimulators. aPRA, panel reactive antibody. The ELISPOT plates then were incubated for 18 to 24 h at 37°C. After bMean Ϯ SD (minimum to maximum). washes with PBS and PBS-Tween, a biotinylated anti–IFN-␥ detection cMedian (10th to 90th percentile). antibody (Endogen, Woburn, MA) was added and the plates were dPositive PRA was defined as Ͼ10% alloantibodies. 566 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 564–572, 2006 mean age of the cohort was 56 yr, 36.6% were female, and 53.6% were self-described as black. Characteristics of the panel of stimulator cells are shown in Table 2. The mean age of these patients was 47 yr. Two were male, and six were female. Fifty percent of the stimulators derived from black patients. The HLA alleles that were ex- pressed by the stimulator panel shown represent approxi- mately 60% of HLA antigens found in the general US popula- tion (Table 2; http://www.ashi-hla.org/publicationfiles/ archives/prepr/motomi.htm). As might be anticipated as a result of the random selection of responders and stimulators, the median number of HLA mismatches between any re- sponder and any stimulator was high (median 5; range 4 to 6).

Detection of IFN-␥–Secreting PBL Reactive to a Panel of Allogeneic Stimulator Cells PBL from each responder were tested in IFN-␥ ELISPOT assays against the same panel of stimulators. An example of one representative assay is shown in Figure 1A, where PBL from five individual HD patients were tested in duplicate against the eight different stimulators. PBL plus medium alone served as a negative control (Figure 1A, top), and phytohemag- glutinin stimulation was used as a positive control (Figure 1A, bottom). On the basis of published studies that defined the coefficient of variance of the ELISPOT assay at 20 to 30% (13) and on the basis of published correlates between frequencies of alloreactive PBL with poor outcome (12), we defined a response of Ͼ25 ELISPOT per 300,000 cells as a positive result. A summary of the detected results from all samples that responded to each individual stimulator, sorted by stimulator, Figure 1. Panel of reactive T cells (PRT). (A) Representative is shown in Figure 1B. Each stimulator induced a wide range of enzyme-linked immunosorbent spot (ELISPOT) plate showing responses. PBL from some patients exhibited positive responses duplicate results of five responders against eight stimulators. to a given stimulator, ranging from 26 to Ͼ150 IFN-␥ ELISPOT The first row (medium) shows spontaneous secretion of IFN-␥ per 300,000 cells, whereas PBL from other patients did not and served as “background.” The frequencies of IFN-␥ used for respond to the same stimulator population. The heterogeneity analyses were obtained by subtracting the “background” count. of the results for each stimulator suggests that the detected (B) Quantified summary of all results sorted by stimulator. frequencies of IFN-␥ producers were dependent on character- Dotted line represents cutoff of 25 ELISPOT per 300,000 re- sponder peripheral blood lymphocytes (PBL). (C) Quantified istics of the responder, rather than the trivial possibility that summary of all results sorted by responder. Dotted line repre- some stimulators were prepared better (and thus more viable sents cutoff of 25 ELISPOT per 300,000 responder PBL. Arrows with better stimulating capacity) than others. (triangle symbols) are representative patients highlighted in the The same data are replotted with each responder on the x text. axis to illustrate the variable reactivity of a given responder to

Table 2. Normal healthy volunteer characteristics

Total no. of stimulators 8 Normal volunteers/deceased donors 6/2 Age (yr)a 47 Ϯ 13 (35 to 76) Black race 4/4 Male/female gender 2/6 Alleles present in the stimulator panelb HLA A A2, A3, A11, A24, A28, A30, A31, A33, A34 HLA B B18, B35, B51, B8, B13, B14, B44, B45, B50, B53, B60, B62 HLA DR DR7, DR1, DR4, DR11, DR13, DR3, DR12, DR16, DR17, DR18

aMean Ϯ SD (minimum to maximum). bHLA antigens are listed in order of frequencies expressed in the panel. J Am Soc Nephrol 17: 564–572, 2006 Primed T Cell Alloreactivity in Transplant Candidates 567 each member of the stimulator panel (Figure 1C). Again, note the heterogeneity of the detected frequencies of IFN-␥ produc- ers for each responder. PBL from some patients did not respond to any stimulator (e.g., patients 5 and 21), suggesting that they were not reactive to alloantigens. Others responded to some but not all stimulators (e.g., patient 9). PBL from still other patients responded to all stimulators (e.g., patient 37), suggesting a highly alloreactive state.

PRT To facilitate our analysis of cellular alloimmunity to this panel of alloantigens, we defined a PRT for each responder as the percentage of stimulators (0 to 100%) that induced a posi- tive assay result. As shown in Figure 2A, 46% of the patients had a positive PRT-40 (reacted positively to Ͼ40% of the panel, e.g., three of eight stimulators tested), and 17% of the patients had a positive PRT-75 (reacted positively to Ͼ75% of the panel, e.g., more than six of eight stimulators tested). In comparison, 32% of the patients in this cohort had a PRA of Ͼ10%, whereas 22% of the patients were highly sensitized, as defined by a PRA of Ͼ50%. Whereas there was overlap between those patients with a positive PRT and those with a positive PRA, 12 to 34% of study patients (depending on whether we defined a positive as re- sponding to Ͼ75 versus Ͼ40% of the stimulators) were PRT positive but PRA negative (Figure 2, B and C). Similarly, 17 to 20% of the study patients exhibited a positive PRA but a neg- ative PRT. Only 5 to 12% of the patients were PRT positive and PRA positive.

Demographic Correlates of a Positive PRT We next asked whether the presence of a positive PRT cor- related with known epidemiologic risk factors, some of which have been associated with an elevated risk for posttransplan- tation graft dysfunction. As shown in Figure 3, patients who were younger than 55 yr were more likely to exhibit a positive PRT-40 or PRT-75 than older patients (P Ͻ 0.05 for each), and Figure 2. Relative frequencies of PRT and panel reactive anti- there was a strong, albeit not statistically significant, trend for bodies (PRA) suggest that cellular reactivity does not necessar- black responders to exhibit a positive PRT-40 (54 versus 38%; ily suggest humoral sensitization and vice versa. (A) Percentage ϭ Figure 3, middle). This trend became more evident in “strongly of patients in cohort (n 41) with a positive PRT-40, PRT-75, or Ͼ Ͼ alloreactive,” PRT-75ϩ patients (27 versus 5%; P ϭ 0.09 by PRA 10% or 50%. (B) Pie chart showing the makeup of the cohort using PRT-40 and PRA Ͼ10%. (C) Pie chart showing the Fisher test). There was no statistically significant correlation makeup of the cohort using PRT-75 and PRA Ͼ50%. between the prevalence of positive PRT and male versus female gender (Figure 3, bottom). The incidence of positive PRT-75 was not significantly different in those with a history of mul- tiple blood transfusions (29%) versus those without a history of history of multiple blood transfusions (50 versus 25%; P ϭ 0.15) (30%; P ϭ 0.81; data not shown). Similarly, and in women with a history of pregnancies (100 versus 67%; there was no difference in PRT-75 responses in women with or P ϭ 0.14 by Fisher exact test). Finally, as shown in Table 3, a without history of pregnancies (100 versus 84.6%, respectively; positive PRT-75 was significantly more prevalent in the popu- P ϭ 1.0). In contrast, a positive PRA (Ͼ10%) was statistically lation of younger black patients versus older nonblack patients. more prevalent in women versus men (P Ͻ 0.01; Figure 4, In contrast, there was no difference between the prevalence of bottom), and this difference remained significant in highly a positive PRA (at either 10 or 50%) between these two sub- sensitized patients (PRA of Ͼ50%; P Ͻ 0.01). No significant groups (Table 3). All correlations were independent of HLA correlations between a positive PRA and either younger versus matching (data not shown), although the high number of HLA older age or black versus nonblack race were detected (Figure 4, mismatches in all combinations (four to six; Table 2) precluded top and middle). A strong nonstatistically significant trend to our ability to assess this variable fully. exhibit a positive PRA-50 test was seen in patients with a Because studies have shown that the strength of PRA tests 568 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 564–572, 2006

Figure 3. Demographic correlates with PRTϩ status. Univariate analysis of the prevalence of PRT-40ϩ patients (A) or PRT-75ϩ patients (B) with age (top), race (middle), and gender (bottom).

can vary in a given patient when tested at different time points episodes, whereas one of the two patients with a positive (4), we tested whether cellular alloreactivity as assessed by PRT PRT-75 had biopsy-confirmed acute cellular rejection of the varies over time. We performed repeat PRT testing in 17 of the transplanted kidney within the initial 8 mo posttransplantation. 41 patients 6 to 20 mo after the initial assay and using the same Notably, all seven patients were PRA negative. panel of stimulator cells. As shown in Figure 5, PRT reactivity did change in a number of the tested patients between the two time points, although only one of 14 patients became newly Discussion highly reactive (new ϩ PRT-75). In a manner analogous to PRA testing to detect alloantibody- based sensitization in transplant candidates, ELISPOT PRT test- Correlation between Positive PRT and Posttransplantation ing provides information regarding the strength of the effector/ Clinical Outcome memory cellular alloimmune repertoire. Because cellular Although the correlations between PRT, PRA, and epidemi- alloimmunity, specifically the frequency of effector/memory ologic risk factors for posttransplantation graft injury are in- alloreactive T cells, has been implicated as an important medi- triguing, the clinically relevant issue is whether ELISPOT PRT ator of graft injury (6,9,10,17), this new ability to quantify such values correlate independently with posttransplantation graft T cells in the peripheral blood of transplant candidates has the injury. Correlations between PRT results and clinical outcomes potential to function as an adjunctive tool for clinical risk for the seven patients who underwent kidney transplantation assessment. The assay is technically straightforward and can be are shown in Table 4 (all seven of these patients underwent accomplished in Ͻ24 h, making it feasible for large-scale clin- single PRT testing within 6 to 12 mo of transplantation). Al- ical application. Our analysis of ELISPOT PRT assay results in though the numbers are small, the results show that the five this cohort of 41 transplant candidates provides several new patients with a negative PRT-75 remain free of renal rejection pieces of information. J Am Soc Nephrol 17: 564–572, 2006 Primed T Cell Alloreactivity in Transplant Candidates 569

Figure 4. Demographic correlates with PRAϩ status. Univariate analysis of the prevalence of PRAϩ (10%) patients (A) or highly sensitized PRAϩ (50%) patients (B) with age (top), race (middle), and gender (bottom).

Table 3. Demographic correlates of positive PRTa and PRA

Black Patients Ͻ55 Years Nonblack Patients Ն55 Years P Valueb

PRT-40 (n ͓%͔) 7/11 (63.6) 4/15 (26.7) 0.11 PRT-75 (n ͓%͔) 5/11 (45.5) 0/15 (0) Ͻ0.01 PRA-10 (n ͓%͔) 4/11 (36.4) 4/15 (26.7) 0.68 PRA-50 (n ͓%͔) 2/11 (18.2) 4/15 (26.7) 1.0

aPRT, panel of reactive T cells. bBy Fisher exact test.

First, we found that human reactivity to alloantigens can be plementary information regarding the nature of the primed, manifest through expression of alloantibodies or alloreactive T human alloimmune repertoire. cells or both cellular and , a finding that Although it would be of interest to define and compare extends previously published work by our group (10). We specificities of the alloreactive T cells and alloantibodies in this found that only 12% of our cohort exhibited a positive PRT and cohort, the relatively small number of responders that were a positive PRA, whereas 34% of patients had a positive PRT but both PRA and PRT positive (12%), the limited number of stim- a negative PRA, and 20% exhibited a positive PRA but a neg- ulators (eight), and the highly polymorphic nature of HLA ative PRT (Figure 2). Thus, the information obtained from molecules prevented us from being able to make definitive simultaneously testing PRA and ELISPOT PRT provides com- conclusions on this issue. One long-term goal is to expand the 570 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 564–572, 2006

Whereas using HLA-derived synthetic peptides can be used to assess indirect responses after graft injury is manifest (9,10), there is no evidence that PBL from patients who have not received a transplant respond to indirectly presented HLA- derived peptides. The extensive polymorphic nature of the HLA complex makes peptide-based screening impractical, and preliminary studies with isolated T cells from two individuals showed that the reactivity occurred through the direct pathway (data not shown). More important, studies that provided strong correlates between pretransplantation immunity and posttrans- plantation outcomes used unfractionated PBL as responders (10,12–14,17). Future work to assess pretransplantation reactiv- ity to indirectly presented peptides as well as pretransplanta- tion PRT and correlating results with posttransplantation out- come need to be performed to address the importance of pretransplantation indirect reactivity in more detail. We found several intriguing correlations between the PRT positivity and other epidemiologic factors associated with post- transplantation graft injury. Black transplant candidates, in par- ticular, younger black candidates, had a higher incidence of positive PRT than the other tested subgroups (Figure 3, Table 3). From epidemiologic studies, it is well known that this group of transplant candidates is at high risk for posttransplantation acute rejection and graft failure, raising the possibility that pretransplantation effector/memory alloreactive T cells con- tribute to the high-risk state (18–22). The additional observa- tion that PRT positivity is less frequent in older transplant candidates (Figure 3) is interesting. Published studies suggest inherently diminished immune reactivity as patients age (23,24). Specifically, in the transplant setting, epidemiologic data reveal a lower incidence of acute rejection in older trans- plant recipients (23–25). The lower incidence of PRT positivity in this Ͼ55 yr old population suggests a functional basis for these previously identified clinical associations. Importantly, we did not detect a correlation between the prevalence of alloantibody sensitization (ϩ PRA) and age or race, providing further support for the speculation that the effector/memory cellular alloimmune repertoire contributes to the higher risk for posttransplantation graft injury. Pregnancy and multiple blood transfusions were associated with alloanti- body sensitization (but not with PRTϩ), verifying previous associations found in the literature (4). Our data provide preliminary evidence from seven patients implicating a relationship between PRT results and posttrans- Figure 5. PRT variability over time. (A) Results of repeated PRT plantation outcome, independent of the presence of alloanti- testing using PRT-75 and PRT-40 in 17 of the 41 patients who bodies. We and others have shown that the frequency of pre- were tested between 6 and 20 mo apart. (B) Example of a ␥ repeated PRT in a responder with no change (top), decrease transplantation IFN- ELISPOT that are reactive to donor cells (middle), and increase (bottom) responses. correlates with early acute rejection episodes and posttrans- plantation renal function independent of other recipient and donor factors that are known to influence these outcomes stimulator panel to include a large number of HLA haplotypes (11,14,17). If additional prospective analyses show that the PRT that ultimately would permit us to define the specificity of the proves to have similar predictive value, then the relationship cellular alloimmune response, an approach that is beyond the between the final pretransplantation donor-specific IFN-␥ scope of this work. ELISPOT assay and PRT as measures of T cell reactivity may Our approach using PBL as responders detects the total prove to be analogous to the relationship between the “final alloresponse but does not differentiate T cells that respond cross-match” and PRA in assessing alloantibody sensitization. through the direct versus the indirect allorecognition pathways. The data presented herein suggest that a high frequency of J Am Soc Nephrol 17: 564–572, 2006 Primed T Cell Alloreactivity in Transplant Candidates 571

Table 4. Correlation of PRT results with posttransplantation clinical outcomea

h/o SCr6Mo Patient Age Gender Race Type of Month Acute Posttransplantation Immuno- PRT-75 PRA Transplant Posttransplantation Rejection (mg/dl) suppression Results Results

1 48 M White CADb 18 No 1.7 S, MMF, P ϪϪ 2 66 M Black CAD 11 Yes 2.4 FK, MMF, P ϩϪ 3 32 F White CAD 10 No 1.1 FK, MMF, P ϪϪ 4 58 M White CAD 10 No 1.0 FK, MMF, P ϪϪ 5 52 F Black LD 7 No 1.0 FK, MMF, P ϪϪ 6 53 M Black CAD 8 No 1.4 S, MMF, P ϩϪ 7 61 F White CAD 8 No 1.4 S, MMF, P ϪϪ

ah/o, history of; SCr, serum creatinine; M, male; F, female; CAD, deceased donor; LD, living donor; SPK, simultaneous pancreas kidney transplant; S, sirolimus; FK, tacrolimus; MMF, mycophenolate mofetil; P, prednisone. bExtended donor criteria organ.

alloreactive T cells as assessed by PRT will be predictive of clinical tool for pretransplantation risk assessment. It is antici- posttransplantation outcome. However, because the threshold pated that through clearer identification of high-risk patients for defining a positive response in this cohort was chosen on the basis of functional measures of immune reactivity, phy- arbitrarily, it is possible that different/higher thresholds will sicians ultimately will be able to make individualized therapeu- improve the predictive value of the PRT. Evaluating frequen- tic decisions to optimize outcome and avoid the potential pit- cies of T cell alloreactivity in large numbers of normal volun- falls of using epidemiologic data to define treatment regimens teers may also aid in better delineating clinically relevant empirically. thresholds for predicting outcome. Regardless, these findings support the contention that ELISPOT screening for PRT status will provide different information than screening tests that are Acknowledgments focused on detection of alloantibodies. Prospective studies that E.D.P. was a recipient of a fellowship grant from the National Kidney analyze larger numbers of patients with posttransplantation Foundation during 2003 to 2004. The studies were supported by in part follow-up will be required to make definitive conclusions re- by National Institutes of Health contract N01-AI-05410 and by the Leonard Rosenberg Foundation. garding these hypotheses. The variability in detected PRT responses that are found in some patients (Figure 5) is not surprising. From a technical standpoint, this test has a known coefficient of variability of References approximately 20% (13), and because we used only eight dis- 1. Pascual M, Theruvath T, Kawai T, Tolkoff-Rubin N, Co- tinct stimulators, a change in reactivity to a single stimulator simi AB: Strategies to improve long-term outcomes after could alter a patient’s status from negative to positive. These renal transplantation. N Engl J Med 346: 580–590, 2002 technical limitations will be overcome as we (1) expand our 2. Sayegh MH, Carpenter CB: Transplantation 50 years lat- stimulator population and (2) develop well-characterized and er—Progress, challenges, and promises. N Engl J Med 351: 2761–2766, 2004 uniform populations of in vitro–expanded B cell stimulators 3. Takemoto SK, Terasaki PI, Gjertson DW, Cecka JM: Twelve (26) (rather than T cell–depleted PBL or spleen cells) for use in years’ experience with national sharing of HLA-matched large-scale clinical trials. In addition, alterations in PRT reac- cadaveric kidneys for transplantation. N Engl J Med 343: tivity may reflect environmental influence such as undiagnosed 1078–1084, 2000 intercurrent infections, alterations in medications, recent im- 4. Gebel HM, Bray RA, Nickerson P: Pre-transplant assess- munizations, etc., that can influence significantly the cellular ment of donor-reactive, HLA-specific antibodies in renal alloimmune response (27). None of the patients with significant transplantation: Contraindication vs. risk. Am J Transplant alterations in PRT from our cohort has received a transplant, so 3: 1488–1500, 2003 we cannot comment on whether the detected alterations in PRT 5. Patel R, Terasaki PI: Significance of the positive crossmatch are clinically important. This issue is being addressed specifi- test in kidney transplantation. N Engl J Med 280: 735–739, cally in ongoing clinical trials. 1969 6. Sayegh MH: Why do we reject a graft? Role of indirect allorecognition in graft rejection. Kidney Int 56: 1967–1979, Conclusion 1999 We have defined a novel approach for evaluating and screen- 7. Russell PS, Chase CM, Colvin RB: Alloantibody- and T ing cellular alloimmune reactivity in transplant candidates. If cell–mediated immunity in the pathogenesis of transplant these findings are verified to correlate with posttransplantation arteriosclerosis: Lack of progression to sclerotic lesions in B outcomes, independent of PRA and HLA matching, then pre- cell-deficient mice. Transplantation 64: 1531–1536, 1997 transplantation PRT testing may become a useful adjunctive 8. Steele DJ, Laufer TM, Smiley ST, Ando Y, Grusby MJ, 572 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 564–572, 2006

Glimcher LH, Auchincloss H Jr: Two levels of help for B 17. Augustine JJ, Siu DS, Clemente MJ, Schulak JA, Heeger PS, cell alloantibody production. J Exp Med 183: 699–703, 1996 Hricik DE: Pre-Transplant IFN-gamma ELISPOTs are as- 9. Najafian N, Salama AD, Fedoseyeva EV, Benichou G, sociated with post-transplant renal function in African Sayegh MH: Enzyme-linked immunosorbent spot assay American renal transplant recipients. Am J Transplant 5: analysis of peripheral blood reactivity to do- 1971–1975, 2005 nor HLA-DR peptides: Potential novel assay for prediction 18. Katznelson S, Gjertson DW, Cecka JM: The effect of race of outcomes for renal transplant recipients. J Am Soc Neph- and ethnicity on kidney allograft outcome. Clin Transpl rol 13: 252–259, 2002 379–394, 1995 10. Poggio ED, Clemente M, Riley J, Roddy M, Greenspan NS, 19. Young CJ, Gaston RS: Renal transplantation in black Amer- Dejelo C, Najafian N, Sayegh MH, Hricik DE, Heeger PS: icans. N Engl J Med 343: 1545–1552, 2000 Alloreactivity in renal transplant recipients with and with- 20. Hutchings A, Purcell WM, Benfield MR: Increased co- out chronic allograft nephropathy. J Am Soc Nephrol 15: stimulatory responses in African-American kidney allo- 1952–1960, 2004 graft recipients. Transplantation 71: 692–695, 2001 11. Nickel P, Presber F, Bold G, Biti D, Schonemann C, Tullius 21. Kerman RH, Kimball PM, Van Buren CT, Lewis RM, Ka- SG, Volk HD, Reinke P: Enzyme-linked immunosorbent han BD: Possible contribution of pretransplant immune spot assay for donor-reactive interferon-gamma-producing responder status to renal allograft survival differences of cells identifies T-cell presensitization and correlates with black versus white recipients. Transplantation 51: 338–342, graft function at 6 and 12 months in renal-transplant re- 1991 cipients. Transplantation 78: 1640–1646, 2004 22. Kerman RH, Kimball PM, Van Buren CT, Lewis RM, Ka- 12. Hricik DE, Rodriguez V, Riley J, Bryan K, Tary-Lehmann han BD: Stronger immune responsiveness of blacks vs M, Greenspan N, Dejelo C, Schulak JA, Heeger PS: Enzyme whites may account for renal allograft survival differences. linked immunosorbent spot (ELISPOT) assay for interfer- Transplant Proc 23: 380–382, 1991 on-gamma independently predicts renal function in kid- 23. Friedman AL, Goker O, Kalish MA, Basadonna GP, Kliger ney transplant recipients. Am J Transplant 3: 878–884, 2003 AS, Bia MJ, Lorber MI: Renal transplant recipients over 13. Gebauer BS, Hricik DE, Atallah A, Bryan K, Riley J, Tary- aged 60 have diminished immune activity and a low risk of Lehmann M, Greenspan NS, Dejelo C, Boehm BO, Hering rejection. Int Urol Nephrol 36: 451–456, 2004 BJ, Heeger PS: Evolution of the enzyme-linked immu- 24. Moreso F, Ortega F, Mendiluce A: Recipient age as a de- nosorbent spot assay for post-transplant alloreactivity as a terminant factor of patient and graft survival. Nephrol Dial potentially useful immune monitoring tool. Am J Trans- Transplant 19[Suppl 3]: iii16–iii20, 2004 plant 2: 857–866, 2002 25. Pallardo Mateu LM, Sancho Calabuig A, Capdevila Plaza 14. Heeger PS, Greenspan NS, Kuhlenschmidt S, Dejelo C, L, Franco Esteve A: Acute rejection and late renal trans- Hricik DE, Schulak JA, Tary-Lehmann M: Pretransplant plant failure: Risk factors and prognosis. Nephrol Dial frequency of donor-specific, IFN-gamma-producing lym- Transplant 19[Suppl 3]: iii38–iii42, 2004 phocytes is a manifestation of immunologic memory and 26. Zand MS, Bose A, Vo T, Coppage M, Pellegrin T, Arend L, correlates with the risk of posttransplant rejection epi- Lee FE, Bozorgzadeh A, Leong N: A renewable source of sodes. J Immunol 163: 2267–2275, 1999 donor cells for repetitive monitoring of T- and B-cell allo- 15. Hricik DE, Heeger PS: Minimization of immunosuppres- reactivity. Am J Transplant 5: 76–86, 2005 sion in kidney transplantation. The need for immune mon- 27. Roddy M, Clemente M, Poggio ED, Bukowski R, Thakkar itoring. Transplantation 72: S32–35, 2001 S, Waxenecker G, Loibner H, Himmler G, Hricik DE, Hee- 16. Heeger PS, Hricik D: Immune monitoring in kidney trans- ger PS: Heterogeneous alterations in human alloimmunity plant recipients revisited. J Am Soc Nephrol 13: 288–290, associated with . Transplantation 80: 297–302, 2002 2005

See related editorial, “Measuring T Cell Alloreactivity to Predict Kidney Transplant Outcomes: Are We There Yet?” on pages 328–330.

Access to UpToDate on-line is available for additional clinical information at http://www.jasn.org/