- Top of page
Post-transplant monitoring of cellular immunity might be useful in predicting long-term outcomes of kidney transplant recipients. We used an enzyme linked immunoabsorbent spot (ELISPOT) assay to serially measure the frequency of peripheral blood lymphocytes producing interferon-gamma in response to stimulator cells from donors or third parties in 55 primary kidney transplant recipients. Mean frequencies measured during the first 6 months after transplantation correlated significantly with the serum creatinine concentration at both 6 and 12 months following transplantation. The mean frequencies were higher in patients with acute rejection than in those without acute rejection. Multiple regression analyses indicated that the correlations between the early ELISPOT measurements of interferon-gamma and serum creatinine were independent of acute rejection, delayed graft function, or the presence of panel reactive antibodies before transplantation. Patients with low mean frequencies of interferon-producing cells in the early post-transplant period were generally free from acute rejection and exhibited excellent renal function at 6 and 12 months post-transplant. In conclusion, using the ELISPOT assay, we show an independent correlation between early cellular alloreactivity and long-term renal function. Increased levels of early alloreactivity measured with this assay may serve as a surrogate for chronic allograft dysfunction.
- Top of page
Both immune and nonimmune mechanisms contribute to the pathogenesis of chronic allograft nephropathy, which remains a leading cause of late renal allograft failure (1). Evidence for an immune etiology derives largely from registry data or multivariate analyses that demonstrate a correlation between episodes of clinically evident acute rejection and long-term allograft dysfunction (1–4). Because T lymphocytes are thought to be central mediators of both acute and chronic allograft rejection (5), there is considerable interest in developing methods for monitoring T-cell function that can serve as surrogate markers for poor long-term outcomes.
To this end, our laboratory has developed an enzyme-linked immunoabsorbent spot (ELISPOT) assay capable of detecting cytokine secretion by individual, antigen-reactive T cells (6–8). Our previous data suggested that this assay detects activated or memory T cells before transplantation, and showed that the method holds some promise as a predictor of post-transplant outcomes (6–8). Post-transplant monitoring of cellular immunity might be useful in assessing the degree of immunosuppression provided by drug therapy and in predicting long-term outcomes. In an effort to test this hypothesis, we used the ELISPOT approach to serially measure the frequency of peripheral blood lymphocytes (PBLs) producing interferon-gamma (IFN-γ) in response to stimulator cells from donors or third parties following kidney transplantation. In the study reported herein, we retrospectively correlated early measures of these frequencies (IFN-ELISPOTS) with long-term renal function and other clinical parameters.
- Top of page
Clinical characteristics of the patients are shown in Table 1. The mean age of the recipients was 48, 42% were female and 44% were African-American. Delayed graft function occurred in 14.5% of the recipients and the acute rejection rate within the first 6 months post-transplant was 9% (five of 55 patients). The mean number of HLA matches and mismatches (A, B, DR) between the donors and the recipients were 2.0 and 3.6, respectively. As shown in Table 1, the degree of HLA matching and mismatching between the recipients and the third parties was similar. However, the third parties were not HLA-similar to the real donors, having an average of 3.8 HLA mismatches and only 2.0 HLA matches. Recognizing that the study was limited to primary transplant recipients, only 16 of the 55 patients (29%) had a positive panel of reactive antibodies (PRAs) detected by flow cytometry before transplantation, and only 18% had PRAs greater than 20%.
Table 1. Clinical characteristics of the 55 kidney transplant recipients
|Age (years)||48 ± 13 (range 25–74)|
|Acute rejection in first 6 months||9%|
|Delayed graft function||14.5%|
|Mean donor-stimulated||28.6 ± 39 (range 0–178)|
|IFN-ELISPOTS (spots per 300 000 cells)||15.2 ± 31 (range 0–156)|
|mean third party-stimulated|| |
|IFN-ELISPOTS (spots per 300 000 cells)||21 ± 27 (range 0–103)|
|mean combined donor and third-party-stimulated|| |
|IFN-ELISPOTS (spots per 300 000 cells)|| |
|Serum creatinine (mg/dL) at 6 months||1.45 ± 0.06 (range 0.6–3.5)|
|Serum creatinine (mg/dL) at 12 months||1.62 ± 0.7 (range 0.7–4.6)|
|Donor HLA match (A,B,DR)||2.0 ± 1.6 (range (0–6)|
|Donor HLA mismatch (A,B,DR)||3.6 ± 1.7 (range 0–6)|
|Third-party HLA match (A,B,DR)||1.7 ± 0.2|
|Third-party HLA mismatch (A,B,DR)||4.3 ± 0.6|
|Class 1 PRAs (%)||14 + 30 (range 0–99)|
|Class 2 PRAs (%)||7.5 + 23 (range 0–97) |
As previously shown by our group, we took advantage of the high resolution and rapid production of IFN-γ by ELISPOT to define the frequency of primed/memory alloreactive T cells in the peripheral blood of each transplant recipient (6,7). Importantly, as the responder cells used in this assay were PBLs (consisting of both recipient T cells and recipient APCs capable of processing donor antigen in the ELISPOT plates), the detected ELISPOTs likely represented cytokines produced by individual, primed, donor-reactive T cells responding through both the direct and indirect allorecognition pathways at a given time point.
Thirty-seven of the 55 patients, including the five patients who ultimately experienced acute rejection, had donor-stimulated IFN-ELISPOTS measured before transplantation (i.e. before administration of immunosuppression). The mean frequency of pretransplant IFN-ELISPOTS was significantly higher in patients with subsequent acute rejection than in those without clinically evident rejection episodes (79 ± 69 vs. 30 ± 44 spots per 300 000 cells; p = 0.039), consistent with previously published studies by our laboratory (6,7). Interestingly, there was a negative correlation between pretransplant IFN-ELISPOTS and the number of HLA matches (R = −0.407; p = 0.03) and a statistically insignificant trend toward a positive correlation between pretransplant IFN-ELISPOTS and the number of HLA mismatches (R = 0.356; p = 0.063). However, there was no significant correlation between pretransplant IFN-ELISPOT values and serum creatinine levels at 6 months (R = −0.014, p = NS) or 12 months (R = 0.114, p = NS) post-transplant.
The mean frequency of post-transplant donor-reactive and third-party-reactive IFN-ELISPOTs, shown in Figure 1, was higher in rejectors than in nonrejectors, although the difference only reached statistical significance for third-party responses. Importantly, we noted a weak but statistically significant correlation between the donor-stimulated and third-party-stimulated IFN-ELISPOTS (R = 0.312; p < 0.05). Thus, for the remainder of the analyses, the donor and third-party responses were analyzed both separately and by using a combined average of donor and third-party responses.
Figure 1. Comparison of average post-transplant frequencies of recipient cells producing interferon-gamma (IFN) in patients with and without acute rejection. AVGDIFN = mean frequency of donor-stimulated spots per 300 000 cells in the first 6 months post-transplant; AVG3PIFN = mean frequency of third-party-stimulated spots per 300 000 cells in the first 6 months post-transplant; PBL = peripheral blood lymphocytes. *p = 0.16; **p = 0.001.
Download figure to PowerPoint
Table 2 depicts the bivariate correlations between several relevant continuous variables and the serum creatinine concentration at 6 and 12 months post-transplant. Notably, both donor-stimulated and third-party-stimulated IFN-ELISPOTS correlated positively and significantly with the serum creatinine concentration at both 6 and 12 months. The statistical strength of this association was even stronger when donor and third-party responses were used in combination (R = 0.586, p < 0.001 for creatinine at 6 months; R = 0.532, p < 0.001 for creatinine at 12 months). When the correlation analysis was performed only in patients without acute rejection, a statistically significant association remained between average IFN-ELISPOT and serum creatinine at 6 months (R = 0.426, p = 0.01) and 12 months (R = 0.376, p = 0.05). Table 3 shows differences in serum creatinine concentrations at 6 and 12 months based on gender, ethnicity (African-American vs. White), delayed graft function, and the presence or absence of acute rejection in the first 6 months. As indicated in these univariate analyses, at both 6 and 12 months, the patients with delayed graft function or acute rejection exhibited significantly higher serum creatinine concentrations than those without these complications.
Table 2. Bivariate correlations between the clinical variables and the serum creatinine concentration
| ||Serum creatinine at 6 months||Serum creatinine at 12 months|
|Mean donor-stimulated IFN-ELISPOTS||0.421||0.002||0.389||0.006|
|Mean third-party-stimulated IFN-ELISPOTS||0.467||<0.001||0.397||0.005|
|Mean combined donor and third-party-stimulated IFN-ELISPOTS||0.586||<0.001||0.532||<0.001|
Table 3. Influence of the additional clinical variables on the serum creatinine concentration
| ||Serum creatinine (mg/dL)|
| ||At 6 months|| || ||At 12 months|| || |
|Delayed graft function||2.1 ± 0.5||1.3 ± 0.5||0.001||2.1 ± 0.5||1.5 ± 0.7||0.015|
|African-American||1.6 ± 0.7||1.3 ± 0.4||NS||1.8 ± 0.8||1.6 ± 0.6||NS|
|Gender: male/female||1.6 ± 0.5/1.6 ± 0.8|| ||NS||1.8 ± 0.4/ 1.6 ± 0.9|| ||NS|
|Acute rejection within 6 months||2.3 ± 0.9||1.4 ± 0.5||0.003||2.7 ± 1.3||1.5 ± 0.5||0.003|
Figure 2 depicts the relationship between the mean (combined-donor-reactive plus third-party-reactive) IFN-ELISPOTS in the first 6 months post-transplant, arbitrarily classified into four quantitative categories, and the serum creatinine concentration at 6 months. It is noteworthy that 26 of the 55 patients (47%) had mean IFN-ELISPOTS of zero to 10 spots per 300 000 cells. For this group of patients, the mean serum creatinine concentration was 1.17 ± 0.3 mg/dL. Notably, no patient in this category experienced a clinically detected acute rejection episode. With increasing mean IFN-ELISPOTS, the serum creatinine concentration increased significantly (p < 0.005). Patients with >30 spots per 300 000 cells (n = 8) had a mean serum creatinine concentration at 6 months post-transplant of 2.1 ± 0.7 mg/dL.
Figure 2. Scatterplots of serum creatinine concentrations by category of post-transplant alloreactivity measured by the combined mean frequency of donor-stimulated plus third-party-stimulated IFNγ-ELISPOTs per 300 000 cells over the first 6 months. PBL = peripheral blood lymphocytes.
Download figure to PowerPoint
Multiple linear regression was performed (Table 4) to assess the independent correlations of age, gender, weight, HLA match and mismatch, pretransplant PRA, delayed graft function, acute rejection in the first 6 months, and mean IFN-ELISPOTS in the first 6 months on the serum creatinine concentration at both 6 and 12 months post-transplant. The 12-month analysis was limited to 48 of the 55 patients with sufficient follow up. At 6 months post-transplant, significant independent correlates of serum creatinine concentration were delayed graft function (p < 0.0001), mean IFN-ELISPOTS (p = 0.001), pretransplant PRA to class 2 antigens (p = 0.007), pretransplant PRA to class 1 antigens (p = 0.041), and acute rejection (p = 0.048). At 12 months post-transplant, significant independent correlates of serum creatinine concentration were acute rejection (p = 0.001), African-American ethnicity (p = 0.015), mean IFN-ELISPOTS (p = 0.03), and male gender (p = 0.03). In an effort to assess the impact of calcineurin inhibitor toxicity on serum creatinine concentration, the regression analyses were repeated for patients receiving tacrolimus (n = 52) and using trough tacrolimus at 6 months post-transplant as an additional independent variable. There was no independent correlation between the tacrolimus level and serum creatinine at either 6 or 12 months (p = 0.54 and 0.63, respectively). Moreover, inclusion of the tacrolimus levels did not alter the list of statistically significant independent correlates shown in Table 4.
Table 4. Results of the stepwise multiple linear regression using the serum creatinine concentration as the dependent variable (includes all patients with and without acute rejection)
|Serum creatinine at 6 months Variable||Beta coefficient||p||Serum creatinine at 12 months Variable||Beta coefficient||p|
|Delayed graft function||0.570||<0.001||Acute rejection in first 6 months||0.437||0.001|
|PRAs (class 2)||0.516||0.007||Average IFN-ELISPOTS||0.287||0.03|
|PRAs (class 1)||0.388||0.041||Male Gender||0.287||0.03|
|Acute rejection in first 6 months||0.268||0.048|| || || |
|Age, gender, HLA match, HLA match, HLA mismatch, ethnicity, body weight|| ||NS||Age, HLA match, HLA mismatch, body weight, delayed graft function|| ||NS|
The multiple linear regression analyses were also repeated after eliminating patients who experienced acute rejection in the first 6 months after transplantation. As shown in Table 5, at 6 months post-transplant, significant independent correlates of serum creatinine concentration were delayed graft function (p < 0.001), mean IFN-ELISPOTS (p = 0.004), and pretransplant PRA to class 2 antigens (p = 0.024). At 12 months, the only statistically significant independent correlate of serum creatinine concentration was delayed graft function (p < 0.001), while correlations with PRA to class 2 antigens (p = 0.062), mean IFN-ELISPOTS (p = 0.065), and male gender (p = 0.075) did not quite reach statistical significance.
Table 5. Results of the stepwise multiple linear regression using the serum creatinine concentration as the dependent variable (includes only patients without acute rejection)
|Serum creatinine at 6 months Variable||Beta coefficient||p||Serum creatinine at 12 months Variable||Beta coefficient||p|
|Delayed Graft Function||0.687||<0.001||Delayed graft function||0.530||<0.001|
|Average IFN-ELISPOTS||0.425||0.004||PRAs (class 2)||0.286||0.062|
|PRAs (class 2)||0.353||0.024||Average IFN-ELISPOTS||0.396||0.065|
| || || ||Male gender||0.318||0.075|
|Age, gender, HLA match, HLA mismatch, ethnicity, body weight, PRAs (class 1)|| ||>0.10||Age, HLA match, HLA mismatch, body weight, delayed graft function, PRAs (class 1), ethnicity|| ||>0.10|
- Top of page
Defining reliable surrogate markers for the outcomes of renal transplant recipients has become a priority for the transplant community (9–11). Despite a dramatic decrease in the incidence of acute rejection over the last decade, chronic allograft dysfunction remains a dominant cause of late graft loss. An ability to segregate allograft recipients into high- vs. low-risk for ultimate graft failure based on early post-transplant surrogate markers could provide the clinician with an opportunity for early intervention in high-risk patients while possibly sparing low-risk recipients from excessive immunosuppression.
Although many factors contribute to chronic graft failure [including the quality of the donor organ, effects of drugs, recurrent disease, and hypertension (4,12–14)], persistent cellular immune reactivity directed at graft antigens remains a prevailing mechanism. Results of the present study suggest that early post-transplant cellular alloreactivity, as assessed by serial ELISPOT measurements of IFN-γ-producing PBLs, is predictive of subsequent renal function independent of clinical variables known to impact upon serum creatinine concentrations following kidney transplantation. Indeed, mean IFN-ELISPOTs measured in the first 6 months post-transplant directly correlated more significantly with subsequent serum creatinine concentrations than recipient age, gender, or weight: the classic demographic variables known to influence serum creatinine levels. The findings show that measurement of IFN-ELISPOTS in the early post-transplant period can serve as a surrogate marker for subsequent renal allograft dysfunction, independent of delayed graft function, acute rejection, and pretransplant sensitization.
Cross-sectional analyses of human transplant recipients with or without chronic allograft dysfunction have shown a strong correlation between poor graft function and T-cell reactivity to donor HLA-derived peptides (indirect pathway) (15–17). In addition to detecting indirect reactivity in patients with chronic graft dysfunction, Lechler and colleagues found that T cells responding through the direct pathway (but not the indirectly pathway) were preferentially inhibited in this population, further implicating indirect rather than direct reactivity as mediating chronic allograft dysfunction (15,18). While such results may seem to contradict our finding that high frequencies of PBLs reactive to intact donor cells correlate with poor outcome, it should be emphasized that this ELISPOT assay is a measure of the total alloimmune response; we did not specifically determine whether the IFN-γ-producing PBLs were responding through the direct vs. indirect pathway. In addition, our findings focus on immune reactivity within the first 6 months of transplantation (a time during which direct reactivity may dominate the alloimmune repertoire) rather than evaluating immune reactivity at much later time points (when indirect reactivity may dominate the alloimmune repertoire) as was carried out in the other reports in which indirect alloreactivity was implicated as a surrogate marker.
The goal of our studies was to develop a simple and reliable early surrogate marker of late transplant outcome rather than to assess mechanisms of chronic graft dysfunction in humans. It was not our intention to determine whether the ELISPOT assay is a surrogate for either acute rejection or chronic allograft nephropathy. Instead, our intent was to correlate average ELISPOT values over time with renal function, and not a histologic lesion. It is important to emphasize that identification of a surrogate marker does not define a cause-and-effect relationship. It is tempting to speculate that the ELISPOT assay for IFN-γ reflects active alloreactivity resulting from relative under-immunosuppression. However, surveillance biopsies to assess whether histologic evidence of subclinical rejection correlated with this measure of alloreactivity were not performed and we did not test the hypothesis that increasing immunosuppression could decrease alloreactivity and thereby improve subsequent renal function. It is also possible that the increased ELISPOT frequencies are a marker for T-cell help required to induce the production of alloantibodies that may play an important role in the pathogenesis of chronic allograft nephropathy (19,20). Finally, it is plausible that this surrogate marker is a nonspecific measure of overall immune reactivity, perhaps only a correlate of certain cytokine gene polymorphisms that have been associated with poor outcome (21,22). Further studies in larger numbers of patients are needed to better discern these mechanistic links. The identification of the post-transplant ELISPOT approach as a surrogate marker for long-term renal function nevertheless represents a significant step forward, and could have important implications for the clinical care of transplant recipients.
The attraction of a donor-specific assay of cellular immunity is the theoretical possibility of identifying patients with immunologic ‘tolerance’ manifested by donor-specific hyporesponsiveness and intact responses to third parties (23). In the absence of true tolerance, however, the high frequency and polyclonality of donor-reactive T-cell alloimmunity may be dominated by responses directed towards certain individual HLA molecules shared by donor and third parties (24) and can result in cross-reactive responses towards a variety of different allogeneic MHC molecules (25). Thus, the fact that reactivity to third-party antigens also contributes to the prediction of outcome as shown in these studies, is not particularly surprising, is consistent with previous studies suggesting that nonspecific immunity can correlate with rejection (26,27), and could represent a practical advantage. The fact that third-party stimulator cells provide predictive information potentially obviates the requirement for obtaining and storing donor cells, making the approach more practical for post-transplant immune monitoring on a large scale.
Potential limitations of our study include the relatively small number of patients, the low incidence of acute rejection, and the relatively short duration of follow up. In addition, we did not independently assess the influence of several nonimmune factors (e.g. systemic hypertension, hyperlipidemia, and viral infection) that may play a role in the pathogenesis of chronic allograft nephropathy. Finally, we acknowledge that serum creatinine concentration may be an inaccurate measure of renal function in kidney transplant recipients. However, recent studies have emphasized the predictive value of serum creatinine itself in predicting renal failure in both renal transplant recipients (28) and the general population (29). Although we did not measure the glomerular filtration rate directly, we were careful to incorporate the most common clinical variables influencing the glomerular filtration rate and serum creatinine concentration (i.e. age, gender, and body weight) into our multivariate analyses.
Despite these limitations, the data convincingly show that post-transplant monitoring of peripheral blood by ELISPOT is a clinically feasible approach for categorizing renal allograft recipients into low-risk and high-risk subgroups. Patients with a low frequency of alloreactive IFN-γ producing PBLs as assessed by ELISPOT are generally free from acute rejection and have excellent renal function at 6–12 months. Higher frequencies of IFN-γ producers are associated with acute rejection and, independently, with chronic allograft dysfunction. Further studies with longer follow up are needed to determine whether the ELISPOT assay can be used to guide the tapering of immunosuppression in patients with low levels of alloreactivity or to guide increased immunosuppression in those with high levels of alloreactivity.