CTLA-4 Blockade following Relapse of Malignancy after Allogeneic Stem Cell Transplantation is Associated with T cell Activation but not Increased Levels of T Regulatory Cells

Abstract

Cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) is a key negative regulator of T cell activation and proliferation. Ipilimumab is a human monoclonal antibody that specifically blocks the binding of CTLA-4 to its ligand. To test the hypothesis that blockade of CTLA-4 by ipilimumab could augment graft-versus-malignancy effects (GVM) without a significant impact on graft-versus-host disease (GVHD), we conducted a phase I clinical trial of ipilimumab infusion in patients with relapsed malignancy following allogeneic hematopoietic stem cell transplant (allo-HSCT). Here we report the analysis of peripheral blood T lymphocyte reconstitution, T regulatory cell (Treg) expression and T cell activation markers after a single dose of ipilimumab in 29 patients. Peripheral blood samples were collected from all patients before and after ipilimumab infusion. We analyzed lymphocyte immunophenotyes, including levels of CD4⁺CD25^high cells and T cell activation markers in all cases. Levels of CD4⁺CD25^highFoxp3⁺ cells and intracellular CTLA-4 in CD4⁺ T cells were also assessed in the last 11 cases. We found that baseline levels of CD4 and CD45RO positive T cells were lower in patients compared to normal controls. More than 50% patients had abnormally low lymphocyte counts, either CD4 or/and CD8 T cells, and some had no circulating B lymphocyte. The percentages of both CD4⁺CD25^high and CD4⁺CD25^highFoxp3⁺ T cells were significantly higher in patients prior to ipilimumab infusion than in healthy donors. 20 of 29 patients showed an elevated level of CD4⁺CD25^low activated T cells at baseline while only 3 of 26 healthy donors had such a population of activated T cells. After ipilimumab infusion, both CD4⁺ and CD8⁺ T lymphocyte counts significantly increased. There was no consistent change in absolute lymphocyte count, or in T cells expressing the activation marker CD69. However, CD4⁺CD25^low T cells in 20 of 29 patients, and CD4⁺HLA-DR⁺ T cell in the last 10 patients increased in the first 60 days following ipilimumab infusion. Although the percentages of both CD4⁺CD25^high and CD4⁺CD25^highFoxp3⁺ T cells significantly decreased during the observation period, the absolute cell counts did not change. Intracellular CTLA-4 expression in CD4⁺ CD25^lo/− T cells significantly increased after ipilimumab infusion. We conclude that CTLA-4 blockade by a single infusion of ipilimumab increased CD4⁺ and CD4⁺HLA-DR⁺ T lymphocyte counts and intracellular CTLA-4 expression at the highest dose level. There was no significant change in Treg cell numbers after ipilimumab infusion. These data show that significant changes in T cell populations occur upon exposure to a single dose of ipilimumab. Further studies with multiple doses are needed to explore this phenomenon further and to correlate changes in lymphocyte subpopulations with clinical events.

Introduction

Relapse of malignancy after allogeneic hematopoietic stem cell transplantation (allo-HSCT) remains a major obstacle to treatment success [1]. Conventional treatment of relapse following allo-HSCT is usually unsuccessful and most patients eventually succumb to their malignancy. The exact mechanism behind the failure of adoptive immunotherapy following allo-HSCT is unclear but may include the lack of specific immune activation, lack of cancer-specific antigens, poor antigen presentation to donor immune cells, and relatively few alloreactive lymphocytes compared to the numbers of proliferating cancer cells [1,2].

Regulatory T (Treg) cells are generated in the thymus and function as immunosuppressive regulators. They are best defined as a subset of CD4⁺ T cells with a phenotype of CD25⁺ and Foxp3⁺, and usually account for less than 5% of total CD4⁺ T cells in the peripheral blood [3,4]. Cytotoxic T lymphocyte antigen 4 (CTLA-4) is expressed on effector T-cells following antigen-specific activation where it functions as a key negative regulatory factor. It is also constitutively expressed on the Treg cell surface [5]. Identification of Treg cells has remained controversial due to the lack of Treg-specific markers that separate this lymphocyte subpopulation from activated T effector cells [6]. Treg cells play a critical role in maintaining immune tolerance and regulating GVHD and GVM. The mechanisms of immune suppression regulated by Treg cells have been found to require cell contact between Treg and effector cells as well as cytokines such as IL-10 and TGF-β [7]. A recent report in a mouse model found that Treg cells mediated suppression of GVHD and GVM through different mechanisms [8]. GVHD suppression did not require granzyme B, while previous studies had shown that granzyme B was involved in suppression of anti-tumor responses. Malignant cells can recruit Treg cells locally to suppress T cell function and create a favorable microenvironment for tumor cell growth [9–11]. Clinical studies have shown an increased number of Treg cells in tumor sites, peripheral blood, and tumor-infiltrated lymph nodes from both solid tumor and hematologic malignancies [12]. This increased expression of Treg cells has been associated with poor clinical outcomes [13]. CTLA-4 has become an attractive target for cancer immunotherapy since the availability of two CTLA-4 monoclonal antibodies, ipilimumab and tremelimumab. Clinical trials using ipilimumab or tremlimumab as a monotherapy or in combination with vaccines, cytokines or chemotheraputic reagents have been performed in patients with metastatic melanoma, renal cell carcinoma, non-Hodgkin’s lymphoma, prostate, colon, and ovarian cancer [14–17]. Objective response rates have been clearly related to antibody dose. The antitumor response is often associated with immune-related adverse events (IRAEs). CTLA-4 blockade-mediated immune responses are associated with tumor specific cytotoxic T cell activation and expansion. Clinical trials using CTLA-4 blockade found an increase of Th1 cytokines in patients’ plasma and an increase in HLA-DR positive CD4⁺ or CD8⁺ T cells, while the changes of CD4⁺CD25⁺ or CD4⁺Foxp3⁺ T subsets, defined as Treg cells were not consistent [18–20]. The increase of activated T cells in the circulation was highly correlated with the antibody dose in combination with GM-CSF in a trial in metastatic prostate cancer [21]. In pre-clinical studies, CTLA-4 blockade can modulate Treg cell function without reducing Treg cell numbers as well as induce both CD4 and CD8 T cell activation in murine models [22–24].

Treg cells may have several roles following allo-HSCT. The importance of Treg cells in the prevention of acute GVHD after allo-HCT has been studied and reviewed [25,26]. However, the correlation of Treg cell numbers and Foxp3 mRNA expression with chronic GVHD after allo-HSCT is not clear [27,28]. The roles Treg cells play in immune reconstitution, maintaining the balance of GVHD and GVM, and treatments such as DLI, immunosuppressive drugs, are not well understood. Clinical studies have shown that selective depletion of T lymphocytes such as CD25⁺ T cells in vitro can effectively reduce the risk of acute and chronic GVHD, but there are disadvantages such as poor engraftment, increased risk of relapse, and delayed immune reconstitution leading to serious infectious complications. Attempts have been made to separate the GVHD and GVM effects. Clinical trials in humans, however, have shown that the GVM effect and long-term survival are highly associated with the presence of GVHD [29].

We have completed a clinical trial of CTLA-4 blockade in patients with relapse of malignancy following allo-HSCT. This trial was based upon the hypothesis that augmenting the immune response to cancer cells by blocking negative regulatory signals might improve the GVM effect. This trial was facilitated by the advent of a fully human monoclonal antibody that is capable of blocking the interaction of CTLA-4 on Treg cells, and the ligands CD80/86 on the antigen presenting cells (APC). An escalating dose of ipilimumab was given as a single intravenous infusion. The clinical results of this study were previously reported [30]. Here we report the immunophenotypes of peripheral blood T cells, including T cell reconstitution, activation, and Treg expression in 29 patients before and after a single dose infusion of ipilimumab.

Patients and methods

Patients

29 patients with relapsed malignancy after allo-HSCT were enrolled in this trial, including 14 with Hodgkin disease (HD), 6 with multiple myeloma (MM), 2 with acute myeloid leukemia (AML), 2 with chronic lymphocytic leukemia (CLL), 2 with CML, 1 with renal cell carcinoma, 1 with breast cancer, and 1 with non-Hodgkin lymphoma (NHL). The first 17 patients were enrolled in the dose-escalation phase of the study; the later 12 patients were in the phase II portion of the study. The median time from transplant to the enrollment in the study was 21 months, ranging from 4 to 79 months. The patient characteristics were previously published [30]. Prior to the trial, 3 patients had extensive chronic GVHD, 4 patients had limited chronic GVHD, and 3 patients had grade I/II acute GVHD. Only 2 patients remained in limited cGVHD at trial entry. Immunosuppressive drugs had been discontinued for more than 6 weeks in all patients. 26 patients had progressive disease, 2 had stable disease and one was in remission after two-months of imatinib therapy. Most patients had 100% donor chimerism in both T cells and myeloid cells. Ipilimumab (Medarex, Inc., Bloomsbury, NJ) was given as a single intravenous infusion with an escalating dose from 0.1mg/kg to 3mg/kg. No other immunomodulatory therapy was given during the observation period of 60 days. Peripheral blood samples from all patients were obtained prior to (day 0) and after Ipilimumab infusion at day 7, 14, 30, and 60. Blood samples at day 1 and day 3 were obtained from the later cohorts of 14 patients.

Peripheral blood from 26 healthy individuals was obtained from the San Diego Blood Bank and used as normal controls. The peripheral blood lymphocytes from 12 donors were separated and cultured with IL-2 (200u/ml) for 3 days and were used as controls for T cell activation markers.

Lymphocyte count

The Absolute Lymphocyte Count (ALC) was calculated based on the percentage of lymphocytes in automated complete blood counts from peripheral blood. Absolute cell count was calculated as: ALC x percentage of expression of marker for T subset or B lymphocyte by flow cytometry analysis.

Antibodies

A panel of FITC, PE or PerCP labeled antibodies including CD3, CD4, CD8, CD11a, CD16, CD19, CD25, CD38, CD45RA, CD45RO, CD56, CD62L, C69, CD152 (CTLA-4), TCRαβ, TCRγδ, HLA-DR, and the isotype controls were purchased from BD Biosciences (San Jose, CA). The FITC labeled anti-GITR (glucocorticoid-induced TNF-R family related gene) was purchased from R&D system (Minneapolis, MN). The T regulatory cell staining kit including CD4 FITC, CD25 PE, Foxp3 PE-cy5 (clone PCH102), and staining buffer sets were purchased from eBioscience (San Diego, CA).

Flow cytometry

Peripheral blood mononuclear cells (PBMC) were obtained using a standard Ficoll-Hypaque density-gradient centrifugation. Cells were stained with a panel of antibodies and analyzed by FACScan and Cellquest software. For analysis of intracellular Foxp3 expression, cells were fixed after surface staining with CD4 FITC, CD25 PE, washed with a permeabilization buffer, then stained with Foxp3 PE-cy5 or isotype control. For analysis of CD4⁺CD25⁺ and CD4⁺CD25⁺Foxp3⁺ T cells, the CD4⁺ cell region was first gated using forward scatter versus FL1 (FITC) or FL3 (PerCP) in a dot plot. The CD4⁺CD25^high and CD4⁺CD25^low regions were then analyzed in a density plot. Treg cells were identified as CD25 and Foxp3 dual-positive population. The expression of T cell activation markers including HLA-DR, CD69, and CD25 were analyzed in either CD4 or CD8 positive T cell region.

Statistical analysis

For baseline values or changes from baseline, a Mann-Whitney U test was used for between-group comparisons and a Wilcoxon signed-rank test or a paired t-test was used for within-group comparisons. For data collected over several time points, mixed effects models were used to test for a significant within-group change over time, adjusted for covariates such as age at diagnosis and type of diagnosis; a random intercept was used to account for correlation of the data within each patient and conditional studentized residuals were plotted to check model fit. Pearson’s correlation tests were used to test for a non-zero correlation between two variables. Analyses used Prizm 5 (Graphpad Software, Inc) and SAS version 9.2 (the SAS Institute Inc., Cary, NC). All tests were two-sided at the 5% significance level.

Results

Flow cytometry analysis

Lymphocyte subset analysis

Eighteen patients had lymphocytopenia defined as an absolute lymphocyte count (ALC) less than 1,500/µl in peripheral blood at the baseline. Among them, six patients had both CD4⁺ and CD8⁺ T cells <200µl, ten patients had either CD4⁺ or CD8⁺ <200/µl. Five patients had no detectable CD19⁺ B cells. Two patients with relapsed CLL had ALC >10,000/µl, and 70–90% of these cells were CD19⁺ monoclonal B cells. Two patients with relapsed AML had a small percentage of circulating blasts. The phenotype of lymphocyte subsets from patients at the baseline and healthy individuals is shown in Figure 1A. Although the overall pattern of lymphocyte subsets appeared similar between the patients and the normal controls, the percentage of CD3⁺, CD4⁺, and CD45RO⁺ T cells was significantly lower in patients. The ratio of CD4 to CD8 T cells was close to 1:1 in the patients, significantly lower than the ratio of 2:1 in the healthy individuals (p=0.01). We analyzed the absolute cell count of lymphocyte subsets at day 0 and at day 60 in patients, as shown in Figure 1B. The cell numbers of all the T cell subsets were significantly increased at day 60. Nineteen (66%) patients had CD4⁺ T cell count and 23 patients (79%) had CD8⁺ T cell count increased from day 0 to 60; among 16 patients with baseline either CD4 or/and CD8 cell counts <200/µl, 7 out of 10 (70%) had an increase of CD4⁺ T cell count and 9 out of 11 (82%) had an increase of CD8⁺ T cell count. There was no consistent change of CD19⁺ B cells and CD56⁺ CD16⁺CD3^neg NK cells from day 0 to 60 after the ipilimumab infusion.

Figure 1.

(A). Lymphocyte phenotypes of patients at baseline and normal controls. The percentages of CD3, CD4, and CD45RO positive T cells are significantly lower in patients than in the healthy individuals (p=0.036, p<0.0001, and p=0.032). (B). Absolute numbers of lymphocyte subsets in patients before and after ipilimumab infusion at day 60. CD19⁺ B cells were excluded in 2 cases of CLL to reduce the huge variation this would cause. The cell counts in all T cell subsets significantly increased from day 0 to day 60 (Wilcoxon signed rank test, p<0.01 for the T subsets). The results are presented as Mean ± SD.

We compared the change in counts of ALC, CD4 and CD8 positive T cells between two groups of patients; those who received less than 3mg/kg of ipilimumab (n=14) and those who received 3mg/kg (n=15). The cell counts of CD4⁺ and CD8⁺ T cells from day 0, 7, 14, 30 and 60 in both groups are shown in Figure 2. We found a significant increase of CD4⁺ T cells from day 0 to 60 in patients who received the higher dose. The median change was 95 cells/µL (Q1–Q3: 0–170) in the lower dose group, and 260 cells/µL (Q1–Q3: 0–360) in the higher dose group (p=0.049). There was no consistent change of the ALC and CD8⁺ T cells between two groups.

Figure 2.

Comparison of CD4 and CD8 positive T cells changes between two groups of patients before and after ipilimumab. A. At day 0, although most patients in the lower dose group had lower CD4 positive T cell counts than the patients that had 3mg/kg, there was no significant difference between the two groups, p=0.09. After the antibody infusion, the CD4⁺CD3⁺ T cell count was significantly increased in the higher dose group from day 0 to day 60, p=0.049. B. At day 0, the absolute cell count of CD8⁺CD3⁺ T cells showed no difference between the two groups of patients, p=0.52. There was no significant change in both groups after the antibody infusion.

CD4⁺CD25^high Treg cells

We initially analyzed CD4⁺CD25^high T cells as Treg cells in all 29 patients and 26 healthy donors by the classic method [31]. Within the CD4⁺ T cell region, there were three cell populations based on CD25 expression by density plot analysis as shown in Figure 3A. They were the CD25^neg region, the CD25^low region, and a tail-like region with low mean fluorescence intensity (MFI) for CD4 expression defined as CD25^high region. At baseline, the percentage of CD4⁺CD25^high T cells was significantly higher in patients (median 7.2%, range 1.5–22%) than healthy donors (median 3.2%, range 0.7–6.5%), p<0.0001. Out of 29 patients, eleven patients had CD4⁺ T cell counts less than 200/µl; for these the percentage of CD4⁺CD25^high T cells was significantly higher (median of 9.1%, range 1.2–22%) than for the other patients (median of 6.7%, range 1.8–14%) at baseline, p=0.04. The absolute cell count of CD4⁺CD25^high T cells in patients with a lower CD4⁺ T cells was 15/µl (range 1–39/µl), compared to 28/µl (range 5–116/µl) in the patients with a higher CD4⁺ T cells. The expression of CD4⁺CD25^high T cells in total CD4 positive cells and the absolute cell counts from day 0, 7, 14, 30 and 60 are shown in Figure 3B. After ipilimumab infusion, using a mixed effects model adjusted for dose group, age and diagnosis, the percentage of CD4⁺CD25^high T cells significantly decreased over time (change per month: −0.66%, 95%CI=−1.26%, −0.03%, p=0.04), but the absolute CD4⁺CD25^high cell count did not change significantly among all 29 patients (p=0.1). The model for comparing counts used percent change from baseline as the outcome.

Figure 3.

A. Flow cytometry analysis of CD4⁺CD25^high Treg cells from two patients (A and B) are shown in the density plots of the CD4⁺ T cell region. Patient A: two regions in CD4⁺ T cells were identified, CD25^neg and a tail-like CD25^high, without a clear-cut region of CD25^low. This pattern was identified in 9/29 patients as well as in 23/26 normal controls. Patient B: three regions in CD4⁺ T cells were identified, CD25^neg, CD25^low, and a tail-like CD25^high, based on the cell density. MFI indicated the CD4 expression for each gated cell region, which was found lower in CD4⁺CD25^high T cell than CD25^low and CD25^neg cell regions. This pattern was identified in 20/29 patients.

B. The expression of CD4⁺CD25^high T subset in CD4 positive cells and the absolute cell counts in 29 patients before and after antibody infusion. There was a significant decrease in the percentage expression of the CD4⁺CD25^high T subset from the baseline to day 60, p=0.04, but the absolute cell count did not change.

Detection of Foxp3 in CD4⁺ T cells

When the antibody for Foxp3 became available for flow cytometry studies, we analyzed intracellular levels of Foxp3 and CTLA-4 in CD4⁺ T cells for the last 11 patients who received higher dose of ipilimumab and from 12 healthy donors. A representative flow cytometry analysis of a patient before and at day 7 after the ipilimumab infusion is shown in Figure 4. At baseline, the percentage of CD4⁺CD25^highFoxp3⁺Treg was significantly higher in patients than normal controls. It accounted for a median of 3.0% (range 0.4–7.1%) of CD4⁺ T cells in patients, compared to a median of 0.6% (range 0.2–3.9%) of CD4⁺ T cells in the controls, p=0.004. The absolute cell count of CD4⁺CD25^highFoxp3⁺ T cells (median of 11/µl, range 2–65/µl) was similar to CD4⁺CD25^high T cells (median of 15/µl, range 6–67/µl) at baseline. Foxp3 positive cells were also found in CD4⁺CD25^lo/− regions as shown in Figure 4. The percentage of CD4⁺Foxp3⁺ T cells accounted for a median of 4.4% of CD4⁺ T cells (range 1.7–10.7%) in patients which was significantly higher than in the controls (median of 1.4%, range 0.2–3.9%), p=0.002. The absolute cell count of CD4⁺Foxp3⁺ T cells accounted for a median of 10/µl, range 5–96/µl at the baseline. The expression of CD4⁺CD25^highFoxp3 T cells in total CD4 positive cells and the absolute cell counts from day 0, 7, 14, 30 and 60 are shown in Figure 5.

Figure 4.

Flow cytometry analysis of intracellular Foxp3 and CTLA-4 expression in CD4⁺ T cells. The density plot analysis gated on CD4⁺ T cells from one patient prior to and on day 7 after ipilimumab infusion. The percentage of CD4⁺CD25^highFoxp3⁺ cells was similar to the percentages of CD4⁺CD25^high and CD4⁺CD25^highCTLA-4⁺ T cells and did not change after treatment. Foxp3 was also expressed in CD4⁺CD25^lo/− cells, which were slightly increased in this case after treatment. CTLA-4 expression in CD4⁺CD25^lo/− T cells increased 3-fold at day 7 from the baseline.

Figure 5.

The expression of CD4⁺CD25^highFoxp3⁺ T subset in CD4 positive cells and the absolute cell count from later 11 patients before and after the antibody infusion. There was a significant decrease in the percentage expression in CD4⁺CD25^highFoxp3⁺ T subset, p=0.02, but the absolute cell count did not change.

After ipilimumab infusion, there was a significant decrease from day 0 to day 60 in the percentage of CD4⁺CD25^highFoxp3⁺ Treg cells (p=0.02) and CD4⁺Foxp3⁺ T cells (p=0.02) in the 11 patients analyzed, while both absolute cell counts did not change significantly, p ≥0.9. The model for comparing counts used a percent change from baseline as the outcomes.

Detection of CTLA-4 in CD4⁺ T cells

In the last 11 patients analyzed, intracellular CTLA-4 expression was detected in both CD25^high and CD25^lo/− cells, accounting for 3.4% (range 0.5–8.6%) of CD4⁺CD25^high T cells and 5.8% (range 2.3–12%) of CD4⁺CD25^lo/− T cells at the baseline. As shown in Figure 6, the expression of CD25, Foxp3 and CTLA-4 in CD4⁺ T cells in 11 patients were all significantly higher than in the normal controls. The absolute cell count of CD4⁺CD25^highCTLA-4⁺ T cells was comparable to the CD4⁺CD25^highFoxp3⁺ T cell count at baseline (median of 17/µl, range 1–65/µl). The CD4⁺CD25^lo/−CTLA-4⁺ T cell count was comparable to the CD4⁺Foxp3⁺ T cell count (median of 27/µl, range 5–99/µl).

Figure 6.

(A). Comparison of the expression of Foxp3 and CTLA-4 in CD4⁺ T cells in patients and normal controls. The percentage of Foxp3 and CTLA-4 in both CD25^lo/− and CD25^high T cells was significantly higher in 11 patients than 12 normal controls (p<0.005). (B). The absolute cell counts of Foxp3⁺ and CTLA-4⁺ in CD4⁺ CD25^high T cells did not change at day 60, but CD4⁺CD25^lo/−CTLA-4⁺ T cells were significantly increased at day 60 (p=0.001).

The intracellular CTLA-4 expression in CD4⁺CD25^lo/− T cells was significantly increased in the 11 patients after ipilimumab infusion (change per month: 6.3%, 95%CI=(2.7%, 9.9%)(p=0.001) as shown in Figure 6. This increase was observed in almost all 11 patients at day 7 and sustained to day 60 (data not shown).

Expression of T cell activation markers

We analyzed CD69 expression on both CD4⁺ and CD8⁺ T cells before and after ipilimumab infusion in 29 patients. At baseline, CD69⁺ T cells accounted for a median of 0.8% (range 0–24%) of total CD4⁺ T cells in patients compared to median of 0.3% (range 0–5%) in normal controls (n=12; p=0.07); for CD8⁺ T cells the median was 2.1% (range 0.3–28%) of CD69⁺ T cells in patients, compared to 2.0% (range 0.5–4.3%) in normal controls (p=0.6). After ipilimumab infusion, overall there was no significant change of CD69 expression in the 29 patients. However, the expression of CD69 on either CD4⁺ or CD8⁺ T cells increased 3 to 60-fold in 8 patients lasting from day 7 to day 60. Overall, there was no significant change of CD69 expression in 29 patients.

The expression of activation marker CD25 was analyzed on both CD4⁺ and CD8⁺ T cells. CD4⁺CD25^low activated T cells were detectable in 20 patients (n=29) at the baseline, while these cells were only found in 3 of 26 healthy individuals studied. Among these 20 patients, these cells accounted for a median of 43% (range 21–64%) of CD4⁺ T cells at baseline. The percentage of CD8⁺CD25⁺ T cells was elevated to 10–22% of CD8⁺ T cells in 6 patients, but only 1 of 12 normal controls at baseline. After ipilimumab infusion, there was no significant change of CD25 expression on CD8⁺ T cells. Although the expression of CD4⁺CD25^low T cells remained stable in all patients at day 60, the cell count significantly increased from day 0 to day 60 (from median of 134/µl, range 23–482/µl at baseline to median of 237/µl, range 46–1197/µl at day 60), p=0.01 (daily increase: 2.52, 95%CI: 1.26–3.78, p=0.002).

We analyzed CD4⁺HLA-DR⁺ activated T cells in the last 10 patients enrolled. At baseline, these accounted for a median of 11% (range 3–31%) of CD4⁺ T cells in patients, compared to a median of 4.9% (range 1–28%) in normal controls (n=16), p=0.02. After the ipilimumab infusion, both the percentage expression and the absolute cell count of CD4⁺HLA-DR⁺ T cells were significantly increased in these 10 patients. The absolute cell count in patients increased from a median of 46/µl (range 6–134/µl) at baseline to 87/ul (range 41–683/µl) at day 60, p=0.004.

The absolute cell counts of CD4⁺CD69⁺ and CD8⁺CD69⁺ T cells, CD4⁺CD25^low and CD4⁺HLA-DR⁺ T cells in patients who received 3 mg/kg dosing before and after antibody infusion are shown in Figure 7.

Figure 7.

Analysis of activated T cells in patients receiving higher doses of ipilimumab. The absolute numbers of CD4⁺CD69⁺, CD8⁺CD69⁺, CD4⁺CD25^low T cells in the last 15 patients, and CD4⁺HLA-DR⁺ in the last 10 patients are shown here before and after the antibody infusion. There were significant increases in CD4⁺HLA-DR⁺ and CD4⁺CD25^low T cell counts from the baseline to day 60, p=0.04 and p=0.002. There was no significant change in CD69⁺ T cell counts.

Correlation with clinical findings

Clinical response was evaluated monthly by physical examination, CT scans, cytogenetics, PCR or FISH analysis of bone marrow samples. Three patients had objective response after ipilimumab infusion [30]. Comparing the responding patients with others, there was no significant difference in the ALC and CD4⁺CD25^high T cell counts, One week after ipilimumab infusion, both CD4⁺ and CD8⁺ T cells increased in all responding patients. The activated T cell counts increased in 2 of the 3 patients.

Clinical GVHD was evaluated monthly in all patients after the ipilimumab infusion. There was no significant change in GVHD status in 2 patients with limited cGVHD and 8 patients with a prior history of GVHD. Three patients with a history of aGVHD had low ALC <1000/µl at baseline. Eight patients had a high percentage of CD4⁺CD25^high T cells, ranging from 5.4 to 15.7% of the total CD4⁺ T cells. Nine patients received DLI after the ipilimumab infusion and only one patient developed grade I acute GVHD.

Discussion

Allo-HSCT can cure a number of malignancies through the GVM effect. This effect is even more important when reduced intensity conditioning is used. However, donor-derived alloreactive T cells can also cause life-threatening GVHD. Activation of T cells requires the recognition of specific antigens, as well as co-stimulatory molecules such as CD80 and CD86 expressed on antigen-presenting cells. Upon activation, CTLA-4 is up-regulated and expressed on the T cell surface to provide negative feedback to activated T cells. CTLA-4 is constitutionally expressed on Treg cells. CTLA-4 blockade has been shown to induce anti-tumor effects in humans with melanoma and certain other malignancies [14–17, 32–34].

We have studied Treg cells and T cell activation markers in a unique group of patients who underwent ipilimumab therapy for relapsed malignancy after allo-HSCT. The patients received a single dose of ipilimumab infusion in an attempt to augment GVM and induce regression of their malignancy. The median time from allo-HSCT to the ipilimumab infusion was 21 months. Immune reconstitution was incomplete in more than 50% of the patients, based on low CD4⁺ and CD8⁺ T cell counts, inverted CD4 to CD8 cell ratios, and undetectable circulating CD19⁺ B cells. A number of factors have significant impact on immune reconstitution after allo-HSCT, including prior chemotherapy, underlying malignancy, conditioning regimen, status of GVHD, immunosuppressive agents, as well as the number of Treg cells and their functions. Our clinical trial is the first attempt to study the effect of CTLA-4 blockade on the expression of Treg cells and T cell activation markers in this unique patient population.

We have made several interesting observations from the trial. First, almost all patients had a significantly higher percentage of Treg cells (both CD4⁺CD25^high and CD4⁺CD25^highFoxp3⁺ Tregs) than the normal controls despite their lymphocytopenia and low CD4⁺ T cell count. We found that the patients with CD4⁺ T cells <200/µl had a higher proportion of CD4⁺CD25^high Treg cells than the patients with CD4⁺ T cells >200/µl. Treg cells have been shown to suppress the proliferation, differentiation and cytokine production of T effector cells. The high percentage of Treg cells may offer a potential explanation for the immune deficiency and the relapse of malignancy. Blockade of CTLA-4 on Treg cells may contribute to the increased number of T cells in the most patients after a single dose of ipilimumab. Clinical studies of myeloma patients after allo-HSCT have shown that Treg cell reconstitution occurs earlier and faster than conventional CD4⁺ T cell reconstitution [35]. These CD4⁺CD25⁺Foxp3⁺ Treg cells are donor-derived memory-type T cells, and expand primarily in bone marrow. Treg cell reconstitution can occur as early as day 30 post-transplant in patients who received CD25-depleted allo-HSCT. These Treg cells are believed to derive from CD4⁺CD25⁻ naïve-type T cells by their intracellular Foxp3 expression [36]. It is debatable whether there is a correlation between Treg cell frequency and the status of chronic GVHD, as well as on long-term immune reconstitution. Without functional analysis, it is difficult to determine the potency of the suppressive function of these Treg cells on GVHD and GVM. A recent report from a multicenter clinical trial with ipilimumab monotherapy in patients (n=284) with pretreated advanced melanoma showed a dose-dependent efficacy. At 10mg/kg patients had a significantly higher response rate, although IRAEs were found in all dose ranges but a large increase of ALC was associated with the highest antibody dose at 10mg/kg [37]. The increase of not only CD4⁺ T cells but CD4⁺Foxp3⁺ Treg cells was reported in a clinical trial using ipilimumab in patients with progressive metastatic hormone-refractory prostate cancer [19]. The expansion of Treg cells and T effector cells were antibody dose dependent. It should be noted that Foxp3 can be induced in activated effector T cells [38].

Secondly, the presence of a high percentage of CD4⁺CD25^low activated T cells in 20 out of 29 patients certainly raises speculation that these donor-derived T effector cells could be functionally activated to mount a GVM attack against the original malignancy. We have observed some evidence of a dose-dependent T cell expansion and an increased expression of T cell activation markers such HLA-DR in 10 patients after ipilimumab infusion. The patients who received 3mg/kg of ipilimumab had a significant increase of CD4⁺ T cell counts compared to those at the lower doses, and there was a significant increase of CD4⁺HLA-DR⁺ activated T cell counts over time in the last 10 patients studied. This finding is similar to other clinical studies of CTLA-4 blockade [39, 40]. There was no significant change of CD4⁺CD25^high and CD4⁺CD25^highFoxp3⁺ Treg cells. Despite the evidence of T cell expansion and activation, there was no clinically significant immune reaction after ipilimumab infusion, even in three patients who had objective responses. These data lead us to believe it may be possible to selectively activate the GVM effect without precipitating clinically significant GVHD. The current trial showed that ipilimumab at 3mg/kg was safe to administer in this patient population. This is a relatively low dose compared with the doses used in phase III melanoma trials which is 10mg/kg given every 2 weeks. Since dose may be important for optimal clinical effects, a higher dose and multiple administrations may provide a more potent and long-lasting GVM.

Studies have shown that T cells activated by cytokines such as IL-2 and IFNγ had increased levels of both surface and intracellular CTLA-4 expression where CTLA-4 plays an immunosuppressive role [41]. CTLA-4 blockade may not only affect Treg cell function but also affect all activated T cells, leading to anti-tumor activities and autoimmune adverse events [42, 43]. In 11 patients studied in our trial, we found that intracellular CTLA-4 expression in CD4⁺CD25^lo/− T cells increased as early as day 1 after ipilimumab infusion and lasted for two months, while intracellular Foxp3 expression did not change significantly. At the same time, we observed the appearance of newly activated T cells. We speculate that the increase of intracellular CTLA-4 levels may be associated with T cell activation.

Treatment for relapsed malignancy after allogeneic stem cell transplantation remains a major challenge. Methods to separate GVHD and GVM will be the intense focus of research and clinical trials in the future. Anti-CTLA-4 targeted therapy is one of the options to selectively activate the immune system and this may lead to augmented GVM activity. We are planning a subsequent study of multiple dosing of ipilimumab in this patient population when the antibody becomes available for clinical trials.