Regulatory T Cells and T Cell Depletion: Role of Immunosuppressive Drugs

Patients

Twenty-one white primary kidney transplant recipients were enrolled under a protocol that was approved by the Ethics Committee of the Renal Transplant Center of Bergamo. Recipients were selected according to age 18 to 70 yr, current panel reactive antibodies <10%, non-HLA identical to the donor and a negative cross-match test, and random allocation on a 1:1 basis to either low-dosage sirolimus or low-dosage CsA added to low-dosage mycophenolate mofetil (MMF) and induction therapy with Campath-1H. Baseline characteristics are shown in Table 1. Schematic representation of the study design is outlined in Figure 1. Informed written consent was obtained from all participating patients according to the Declaration of Helsinki.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Schematic representation of study design. Twenty-one white primary kidney transplant recipients were randomly allocated on a 1:1 basis to either low-dosage sirolimus (SRL group; n = 11) or low-dosage cyclosporin A (CsA group; n = 10) added on to low-dosage mycophenolate mofetil (MMF) and induction therapy with Campath-1H (C1H). Immunosuppressive treatments are indicated by arrows. At surgery, patients were given 500 mg of methylprednisolone intravenously, followed by 30 mg of Campath-1H intravenously. Methylprednisolone was also infused on days 1 (250 mg) and 2 (125 mg). Sirolimus (low SRL) was started on day 1 (4 mg/d orally, then adjusted to expected target trough levels). Intravenous CsA (low CsA) was started soon after surgery (1 to 2 mg/kg per d); on day 1, CsA was shifted to oral CsA (2 mg/kg twice daily) and adjusted according to trough blood concentrations. In both groups, patients were also given MMF at the oral dosage of 500 mg twice daily starting on day 1 (low MMF). Phenotypic and functional assessments are outlined on the left. Timing of each evaluation is marked with a cross.

Immunosuppressive Therapy

At surgery, patients were given methylprednisolone (500 mg intravenously), followed by a 30-mg infusion of Campath-1H. Methylprednisolone was also infused on days 1 (250 mg) and 2 (125 mg) after transplantation. Low-dosage sirolimus (Rapamune, Wyeth-Lederle, Rome, Italy) was started on day 1 (4 mg/d orally, then adjusted to target trough levels of 5 to 10 ng/ml, by HPLC [21]). CsA was given intravenously (1 to 2 mg/kg per d) on day 1, then shifted to oral CsA (2 mg/kg twice daily) and adjusted to trough blood concentrations of 120 to 220 ng/ml (by HPLC [22]) in the first month and of 70 to 120 ng/ml thereafter (Table 2). In both groups, patients were also given MMF (Cell Cept; Roche, Milan, Italy) at the oral dosage of 500 mg twice daily from day 1.

Follow-Up and Patient Outcome

Patients were followed up for 24 mo. Mean serum creatinine concentration progressively decreased during the first 2 mo and remained stable in the two study groups up to study end. GFR, measured by plasma clearance of iohexol (23), showed a stable graft function at 6, 12, and 24 mo (Figure 2) in both groups. In sirolimus-treated patients, however, GFR was numerically higher than in the CsA group at all time points. The adverse events were comparable in the two study groups (Table 3). Three patients, one on sirolimus and two on CsA, had an acute rejection episode at 14, 9, and 210 d after transplantation, respectively, and fully recovered with intravenous methylprednisolone. One patient who was on sirolimus died from sepsis 18 mo after transplantation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Kidney graft function. GFR at 6, 12, and 24 mo after C1H induction in SRL-treated (n = 11, n = 10 at 24 mo; □) and CsA-treated (n = 10 ▪) groups. Data are means ± SEM.

Peripheral Leukocyte Count and Lymphocyte Phenotype

Peripheral blood mononuclear cells (PBMC) were monitored by flow cytometry using TruCount tubes and analyzed by FACSCalibur cytometer (BD Bioscience, San Jose, CA). For T lymphocyte subset immunophenotyping, fresh blood samples and PBMC were incubated with fluorochrome-conjugated mAb against human CD3, CD4, CD8, CD25, CD69, CD45RO, and Ki-67 antigens (BD Bioscience). The samples were analyzed by FACSort or by FACSAria cytometers (BD Bioscience). For each marker, blank samples with isotype-matched control antibodies were analyzed.

FOXP3 Expression in CD4⁺ Subset

After gating for CD3, CD4⁻, CD4⁺, CD4⁺CD25⁻, and CD4⁺CD25^low, CD4⁺CD25^high T cells were isolated by cell sorting (FACSAria; BD Bioscience), and purity was >99%. Total RNA was reverse-transcribed to cDNA, and quantitative real-time PCR was performed on a TaqMan ABI Prism 5700 Sequence Detection System (Applied Biosystems, Foster City, CA) with SYBR Green PCR Core reagents and the following primers: FOXP3, forward (300 nM) 5′-AGC CAT GGA AAC AGC ACA TTC-3′ and reverse (300 nM) 5′-GAG CGT GGC GTA GGT GAA A-3′. β2-Microglobulin served as housekeeping gene. The ΔΔCt equation was used to compare the FOXP3 gene expression in each sample with the expression in a pool of CD3⁺CD4⁺ cells from healthy subjects (calibrator). Results were expressed as arbitrary units taking the expression in the calibrator as 1.

Measurement of Lymphocyte Alloreactivity

Statistical Analyses

Results are mean ± SEM. Baseline characteristics of patients were compared by χ² test or by t test as appropriate. Variations in peripheral blood count, phenotype, and functional data were assessed by ANOVA. The statistical significance level was defined as P < 0.05. The Bonferroni correction for multiple comparison was applied.

Phenotype of Peripheral Blood Leukocytes in Renal Transplant Patients after Campath-1H

Campath-1H profoundly depleted total B cells, NK cells, monocytes, and T cells and CD4⁺ and CD8⁺ T cell subsets (Figure 3, A and B). Confirming previous reports (24), the depleted CD4⁺ T cell pool, however, was enriched by memory CD45RO⁺ cells (Figure 3C). Campath-1H–induced cell depletion was not influenced by the type of maintenance immunosuppression therapy (Figure 3, D through I). CD3⁺CD4⁺ T cells recovered very slowly: At month 24 after transplantation, their count was approximately one third of baseline values in both sirolimus- and CsA-treated patients (Figure 3H). In the sirolimus group, CD3⁺CD8⁺ T cell counts remained lower than before transplantation until month 24 after surgery, whereas in the CsA group, they returned to preoperative values within 4 to 5 mo (Figure 3I). B cells, NK cells, and monocytes recovered faster than CD4⁺ and CD8⁺ T cells, reaching pretransplantation values within 6 (B cells) and 3 mo (NK cells and monocytes) after transplantation in both groups (Figure 3, D through F).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Effect of C1H on circulating leukocyte subsets in renal transplant patients. Using flow cytometry analysis, absolute numbers of total CD3⁻CD19⁺ B cells and CD3⁻CD16⁺CD56⁺ NK cells (A) and of total CD3⁺ T cells and of CD3⁺CD4⁺ (CD4⁺) and CD3⁺CD8⁺ (CD8⁺) subsets (B) were calculated for healthy subjects (n = 11) and patients at baseline (before transplantation, n = 21) and at 14 d after transplantation and C1H induction (n = 21; data from SRL- and CsA-treated patients were cumulated because no difference was recorded between the two groups at this time). Absolute numbers of monocytes were obtained from a complete blood count that was done on the same day (A). (C) Percentage of memory CD45RO⁺ cells within CD3⁺CD4⁺ from healthy subjects and from patients at baseline and at 14 d after transplantation and C1H induction (cumulative data from the SRL and the CsA group). (D through I) Kinetics of absolute numbers of repopulating total CD3⁻CD19⁺ B cells, CD3⁻CD16⁺CD56⁺ NK cells, monocytes, CD3⁺ T cells, and CD3⁺CD4⁺ and CD3⁺CD8⁺ T cell subsets in the peripheral blood of C1H-treated renal transplant patients from baseline (time 0) to 24 mo after transplantation. ○, SRL recipients (n = 11, n = 10 at 24 mo); ▪, CsA recipients (n = 10). Data are means ± SEM. *P < 0.05 versus time 0; §P < 0.05 versus SRL-treated group at the same time point.

Sirolimus- but Not CsA-Based Maintenance Immunosuppression Is Associated with In Vivo Expansion of CD4⁺CD25^high Cells after Campath-1H

The number and the proportion of circulating CD4⁺ cells that comprised CD25^high cells were comparable in participants, analyzed at baseline, and in healthy individuals (n = 11; mean age 36 yr; range 32 to 55; Figure 4, A and B). Campath-1H induced a profound and unselective depletion of CD4⁺CD25⁻, CD4⁺CD25^low (effector T cells) (25) and CD4⁺CD25^high subsets (Figure 4, A through C) in transplant patients.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

SRL maintenance therapy favors expansion of CD4⁺CD25^high subset after C1H induction. (A) Using flow-based frequency enumeration and absolute CD3⁺CD4⁺ cell counts that were obtained on the same day, absolute numbers of CD3⁺CD4⁺CD25⁻, CD3⁺CD4⁺CD25^low, and CD3⁺CD4⁺CD25^high cells were calculated in healthy subjects (n = 11) and in patients at baseline (n = 21) and at 14 d after transplant (after C1H, n = 21). (B) Percentage of CD25^high within the CD3⁺CD4⁺ subset from healthy subjects (n = 11) and from patients at baseline (n = 21), at 14 d (n = 21, data from the SRL and CsA groups were cumulated because no difference was recorded between the two groups at this time), and at 12 mo (data from SRL group [n = 11; □] and CsA group [n = 10; ▪] are presented separately) after transplantation and C1H induction. Data are means ± SEM. °P < 0.05 versus healthy subjects, patient baseline, and CsA. (C) Representative FACS plots of CD25 expression on CD3⁺CD4⁺ cells from healthy subjects, patients at baseline and at 14 d after transplantation and C1H induction, and in patients who were on either SRL or CsA therapy and studied 12 mo after transplantation. CD3⁺CD4⁺ cells were designated CD25⁻ when CD25 expression fell within the background staining that was observed using an isotype control, CD25^high when CD25 expression exceeded that seen in the CD4⁻ population, and CD25^low when CD25 expression fell between these regions (26). (D through F) Using the gating shown in C, posttransplantation longitudinal analysis of the percentage of CD3⁺CD4⁺ that were CD25⁻, CD25^low, or CD25^high are shown. ○, SRL recipients (n = 11, n = 10 at 24 mo); ▪, CsA recipients (n = 10). Data are means ± SEM. *P < 0.05 versus patient baseline (time 0); §P < 0.05 versus the SRL-treated group at the same time point.

From month 1 after transplantation, the percentage of CD25^high cell subset in total CD3⁺CD4⁺ T cells progressively increased over baseline in sirolimus-treated patients (Figure 4, D through F), reaching values that were significantly higher than before transplantation from months 4 to 24 postoperatively. In the CsA group, the trend of CD4⁺CD25^high cell percentage to increase was milder: Values that were significantly higher than before transplantation were recorded only at month 6 (Figure 4, D through F). As a consequence, the Treg/effector T cells ratios were significantly higher (P < 0.05) in the sirolimus than in the CsA group from months 2 through 24 (data not shown).

For assessment of whether enrichment in CD4⁺CD25^high T cells was associated with cell proliferation, staining for Ki-67 (a nuclear cell proliferation–associated antigen that is expressed in all active stages of cell cycle) (26) was performed in cells that were taken from patients at baseline and at the time of maximal CD4⁺CD25^high cell expansion (3 to 5 mo after transplantation). Higher cycling was observed in the three CD4⁺CD25 cell subsets from all patients who were studied 3 to 5 mo after transplantation (P < 0.05) than in cells from patients at baseline and from healthy individuals (Figure 5C), which is consistent with CD4⁺ T cell recovery after Campath-1H–induced lymphopenia (Figure 3H). Of note, the proportion of CD4⁺CD25^highKi-67⁺ cells was significantly (P < 0.05) higher in the sirolimus- than in the CsA-treated group, whereas no difference in the percentages of both CD4⁺CD25^lowKi-67⁺ and CD4⁺CD25⁻Ki-67⁺ cells was found between the two groups (Figure 5C).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

CD3⁺CD4⁺CD25^high cells express FOXP3 and undergo homeostatic peripheral expansion in renal transplant patients after C1H induction. Both events are favored by SRL maintenance therapy. (A) FOXP3 mRNA expression in CD3⁺CD4⁺CD25⁻, CD3⁺CD4⁺CD25^low, and CD3⁺CD4⁺CD25^high cells in healthy subjects (n = 11), patients at baseline (n = 21), and SRL-treated patients (n = 10; ○) and CsA-treated patients (n = 10; •) 24 mo after transplantation. (B) FOXP3 mRNA expression in positively selected CD3⁺CD4⁺ cells in healthy subjects (n = 11) and in patients at baseline (n = 21) and at 24 mo after transplantation (SRL [n = 10; □]; CsA [n = 10; ▪]). Results are expressed as arbitrary units, taking the expression in CD3⁺CD4⁺ cells from a pool of healthy subjects as calibrator (arbitrary units = 1). Data are means ± SEM. *P < 0.05 versus healthy subjects, patients at baseline, and CsA-treated patients at the same time point. (C) Expression of the proliferation marker Ki-67 in CD3⁺CD4⁺CD25⁻, CD3⁺CD4⁺CD25^low, and CD3⁺CD4⁺CD25^high cells from healthy subjects and from patients at baseline and at 3 to 5 mo after transplantation (at the time of maximal expansion of CD3⁺CD4⁺CD25^high cells; n = 6 for each group). Data are means ± SEM. °P < 0.05 versus the same CD25 subset from healthy subjects and patients at baseline; *P < 0.05 versus CsA-treated patients at the same time point.

Expanding CD4⁺CD25^high Cells Express the Treg Hallmark FOXP3

The previous results document that Campath-1H–induced T cell depletion favors the emergence of CD4⁺CD25^high T cells in patients who receive sirolimus maintenance therapy. However, CD25 cannot be regarded as a specific hallmark of Treg, because also effector/memory T cells express this molecule on their surface (27). To ascertain whether the high levels of CD25 expression in regenerating CD4⁺ cells upon Campath-1H induction reflected a regulatory phenotype, we evaluated the mRNA expression level of FOXP3, a gene that encodes a transcription factor that is required for Treg development and function (28,29). Using electronically sorted CD3⁺CD4⁺ subpopulations, we saw the highest levels of FOXP3 expression in CD25^high subset both in patients and in healthy individuals, with intermediate and low levels of FOXP3 expression in the CD25^low and CD25⁻ subsets, respectively (Figure 5A). FACS analysis showed that the large majority (99.0%; range 98.5 to 99.5%) of CD4⁺CD25^high cells from all groups were CD69⁻, excluding that they were activated cells (30). There was a significantly (P < 0.05) increased FOXP3 expression in CD4⁺CD25^high cells that were isolated 24 mo after transplantation from sirolimus- compared with CsA-treated patients or from participants at baseline and healthy individuals (Figure 5A).

To quantify Treg in the total CD3⁺CD4⁺ cell population, we further evaluated FOXP3 expression in positively selected CD3⁺CD4⁺ T cells. Cells that were isolated 24 mo after transplantation from sirolimus-treated patients had higher (P < 0.05) FOXP3 expression compared with CD3⁺CD4⁺ cells from the same patients at baseline or from healthy individuals (Figure 5B), which is consistent with both CD4⁺CD25^high T cell expansion that was observed through flow cytometry and increased FOXP3 expression in this cell subset. Notably, in CsA-treated patients, FOXP3 expression in CD3⁺CD4⁺ T cells was lower (P < 0.05) than in the sirolimus group (Figure 5B). Altogether, these results indicate that, after lymphocyte depletion by Campath-1H induction, sirolimus but not CsA increased the pool of FOXP3-expressing CD4⁺CD25^high cells. We saw no evidence in any group of significant FOXP3 expression in CD4⁻ cells (FOXP3 expression <0.01 arbitrary units).

CD4⁺CD25^high T Cells from Sirolimus-Treated Patients Suppress T Cell Alloreactivity Ex Vivo

Treg are anergic T cells that respond poorly to allogeneic stimuli and are also capable of inhibiting the alloreactive response of effector T cells (31). Therefore, we performed functional assays to address whether the emergence of Treg was associated with host T cell hyporesponsiveness against donor antigens. Both the proliferative response of T cells to alloantigens in MLR and the frequency of previously activated/memory T cells by ELISPOT for IFN-γ after overnight exposure to alloantigens were analyzed (32). Samples were taken before transplantation and at two different time intervals after T cells had repopulated the peripheral blood at an adequate amount (at months 12 and 24 after transplantation). In sirolimus-treated patients, the anti-donor T cell proliferative response and the frequencies of IFN-γ–producing donor-reactive cells were significantly (P < 0.05) reduced at both posttransplantation points as compared with before transplantation (Figure 6, A and B). Posttransplantation anti–third-party alloreactivity was significantly lower than pretransplantation as well (Figure 6, A and B). These results could not be attributed to incomplete recovery of T cell count or to a state of general immunosuppression caused by maintenance therapy, because T cells that were isolated at the same time points responded normally to a polyclonal T cell stimulus with PHA (Figure 6B). In CsA-treated patients, T cells that were studied at the same time points showed donor-specific hyporesponsiveness assayed with MLR and ELISPOT, compared with before transplantation (Figure 6, A and B). T cell response to PHA was normal (Figure 6B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Functional assays on repopulating T cells after C1H induction in renal transplant patients. (A) MLR proliferative responses (stimulation index) to donor (□, SRL; ▪, CsA) and third-party (□) antigens of T cells from transplant patients who received SRL (n = 6) or CsA (n = 6) and for whom donor cells were available. Each patient was studied at baseline and at 12 and 24 mo after transplantation. (B) Enzyme-linked immunosorbent spot frequency of IFN-γ–producing activated/memory T cells (spots/300,000 peripheral blood mononuclear cells [PBMC]) to donor (□, SRL; ▪, CsA) and third-party (□) antigens from transplant patients who received SRL (n = 6) or CsA (n = 6). Each patient was studied at baseline and at 12 and 24 mo after transplantation. The responses to phytohemagglutinin (PHA), used as positive controls, are shown as ○ (SRL) or ▪ (CsA) with respective Y-scale shown on the right. (C) Effect of CD4⁺CD25^high depletion by FACS sorting on the anti-donor (□, SRL; ▪, CsA) and anti third-party (□) frequencies of IFN-γ–producing cells. The number of spots/300,000 PBMC measured before (−) and after (+) CD4⁺CD25^high depletion in PBMC from renal transplant patients studied 24 mo after surgery (SRL, n = 6; CsA, n = 6) are shown. (D) Effect of the addition of autologous CD4⁺CD25^high to CD4⁺CD25^high-depleted PBMC that were isolated at baseline on the frequencies of anti-donor (□) and anti third-party (□) IFN-γ–producing T cells in SRL-treated patients. CD4⁺CD25^high cells were isolated by sorting from SRL patients at 24 mo after transplantation (n = 3), and 5000 cells were added to 300,000 pretransplantation CD4⁺CD25^high-depleted PBMC from the same patients. (E) The frequencies of IFN-γ–producing cells in response to donor (▪) and to third-party antigens (□) in PBMC from CsA-treated patients studied 24 mo after transplantation (n = 3) in the absence (0) or in the presence of 10 or 50 U/ml IL-2. Data are means ± SEM. *P < 0.05 versus pretransplantation; °P < 0.05 versus predepletion (−); #P < 0.05 versus CD25^high-depleted PBMC (+/−).

To clarify the role of Treg in these phenomena, we repeated ELISPOT assays at 24 mo after transplant and after CD4⁺CD25^high cell depletion by sorting. As shown in Figure 6C, in the sirolimus group, depletion of CD4⁺CD25^high cells was associated with a statistically significant (P < 0.05) increase in the frequency of IFN-γ–producing effector/memory cells to both donor and third-party antigens. The suppression ratios, defined as frequency after depletion minus frequency before depletion divided by frequency after depletion (33), were comparable for both anti-donor (0.43 ± 0.06) and anti–third-party (0.45 ± 0.06) response. To confirm that the CD4⁺CD25^high subset in sirolimus-treated patients had regulatory activity and to exclude any possible overlapping inhibitory effect of the concomitant immunosuppressive therapy on the function of effector T cells, we added sorted CD4⁺CD25^high cells, taken from these patients at 24 mo after transplantation, to CD4⁺CD25^high-depleted PBMC from the same patients before surgery. Addition of CD4⁺CD25^high cells consistently reduced the IFN-γ frequencies to both donor and third-party antigens (Figure 6D). The amount of added Treg did correspond to the percentage of circulating CD4⁺CD25^high cells in patients before transplantation, which supports a potential clinical significance of our findings.

Conversely, in patients who were on CsA, CD4⁺CD25^high cell depletion had no effect on the frequencies of anti-donor IFN-γ–producing T cells that were taken at 24 mo after transplantation (Figure 6C). These results indicate that Treg did not play a significant role in the hyporesponsiveness to donor alloantigens in CsA-treated patients. By contrast, addition of IL-2 to the ELISPOT assay increased the IFN-γ frequencies against donor antigens to reach anti–third-party values (Figure 6E). Thus, anergy rather than regulation contributed to donor-specific hyporesponsiveness in CsA-treated patients (34).