Melan-A specific CD8+ T lymphocytes after hyperthermic isolated limb perfusion: A pilot study in patients with in-transit metastases of malignant melanoma

Abstract

Purpose: Isolated limb perfusion (ILP) with hyperthermia is an effective treatment for in-transit metastases of malignant melanoma in the extremities. Preclinical studies have shown that hyperthermia may induce an immunogenic death of tumour cells. We therefore decided to study whether ILP may induce tumour-specific immune responses in the clinical setting.

Method: The number of Melan-A/Mart-1 specific CD8+ T cells, as well as other phenotypically different immune cells, was recorded in peripheral blood in 12 HLA-A2+ patients with in-transit metastases undergoing hyperthermic ILP with melphalan.

Results: All patients underwent ILP without any complication and with an overall response rate of 83%. No substantial changes in the number of circulating T-cells, B-cells, NK-cells or monocytes were observed during follow-up. Four out of 12 patients showed an elevation of Melan-A+ CD8+ T-cells 4 weeks after ILP.

Conclusion: We here report our preliminary observations that a small increase in tumour-specific T-cells could be seen in a subpopulation of patients after ILP. However, much more work is necessary to fully delineate the systemic immune response to hyperthermic ILP.

• Immunology
• isolated limb perfusion
• regional perfusion

Introduction

Lymphatic dissemination manifested as in-transit metastases occur in approximately 5–10% of patients with malignant melanoma [1]. Simple surgical excision is often possible, but is not appropriate when the interval between new lesions is short and when numerous or bulky lesions are present. In 1958 the technique of isolated limb perfusion (ILP) was developed by Creech and Krementz [2]. This technique surgically isolates the affected limb from the systemic circulation, the limb is then perfused with a high concentration of a chemotherapeutic agent. Compared to systemic administration, ILP achieves tissue concentrations of melphalan, a DNA alkylating agent, that are about 20 times higher [3]. In 1969, Stehlin combined ILP with mild hyperthermia to potentiate the effect of melphalan [4]. ILP is associated with complete response (CR) rates of about 60% and overall response rates of about 90% [5,6].

Most chemotherapeutic agents, including melphalan, induce tumour cell death by apoptosis, a process that has long been regarded as immunologically ‘silent’ [7]. However, recent evidence suggests that some anticancer drugs, such as anthracyclines and the alkylating drug cyclophosphamide, induce an immunogenic type of apoptosis that stimulates the engulfment of apoptotic bodies by dendritic cells (DC), and the activation of cytotoxic CD8+ T cells through a process known as ‘cross-priming’ [8,9]. This type of immunogenic cell death is characterised by a series of events that include preapoptotic surface translocation of calreticulin, which serves as an ‘eat me’ signal for phagocytes.

During ILP, the high local concentrations of melphalan will most certainly induce local apoptosis of most tumour cells, but would probably also negatively affect resident immune cells, including DCs. However, after subsequent reconstitution of the normal circulation, massively released tumour-derived components, including tumour-specific antigens, may get access to secondary lymphoid organs, where adaptive immune responses may take place. The perfused tumour-containing limb will further become ‘recolonised’ by new immigrating immune cells, including DCs and DC precursors from blood that may induce a tumour-specific T cell response in tumour-draining lymph nodes.

The aim of the present study was therefore to study whether treatment with hyperthermic ILP in melanoma patients with in-transit limb metastases may induce a tumour-specific immune response as determined by the number of tumour-specific CD8+ T cells in the circulation.

Patients and methods

Patients

Twelve patients with in-transit metastasis of malignant melanoma were recruited to the study from March 2008 to November 2009. There were six women and six men with a median age of 70 years at the time of perfusion. All included patients were HLA-A2 positive and all tumours were stained positive for Melan-A. Patient characteristics are summarised in Table I. The study was approved by the Human Ethics Committee of Sahlgrenska Academy, Gothenburg, Sweden.

Table I. Patient characteristics.

Patient	Age	Gender	Local	Vessel	N-stage
1	63	Female	Foot	Femoral	N3
2	59	Male	Lower arm	Axillary	N2c
3	92	Female	Lower leg	Femoral	N2c
4	68	Female	Foot	Femoral	N3
5	87	Male	Lower leg	Femoral	N2c
6	73	Male	Foot	Femoral	N3
7	76	Female	Lower leg	Iliacal	N2c
8	63	Male	Lower leg	Iliacal	N3
9	53	Female	Lower arm	Axillary	N3
10	81	Male	Whole leg	Femoral	N3
11	72	Male	Whole leg	Femoral	N3
12	78	Female	Whole leg	Femoral	N3

Treatment

Patients underwent ILP via the femoral (n = 8), iliac (n = 2) or axillary (n = 2) approach. The ILP technique has been described previously [10,11]. Briefly, under general anaesthesia, the blood circuit of the limb was isolated by clamping and cannulation of the major artery and vein. The cannulas were connected to an oxygenated extracorporeal circuit and the remaining collateral vessels were compressed using an inflatable tourniquet or an Esmarch bandage. Melphalan (13 mg/L in upper limb, 10 mg/L in lower limb) was administered into the perfusion circuit and the temperature was held at mild hyperthermia of 40 °C with a perfusion time of 90 min in total.

Blood samples

Blood samples were collected before ILP, at the time of discharge from the hospital (1 week) and at follow-up visit after 4 weeks. Patients were referred from all parts of Sweden, and for five of the patients living nearby we were also able to collect blood at 12 weeks after ILP.

Pentamer staining for detection of Melan-A/Mart-1 specific cytotoxic T cells

For detection of Melan-A/Mart-1 specific CD8+ cytotoxic T cells, freshly drawn EDTA blood was incubated with a phycoerythrin conjugated ELAGIGTV-pentamer (Proimmune, Oxford, UK) for 10 min at 4 °C. Following incubation the blood was washed once and incubated with CD3-FITC, CD8-PerCP and CD19-APC (all from BD Biosciences, San Jose, CA) for 20 min on ice (no in vitro expansion of the cell populations was performed). The blood was lysed, washed twice and resuspended in FACSFlow (BD Biosciences) before flow cytometry analysis. All incubation steps were performed in the dark. The cells were analysed on a BD FACSCanto and CellQuest (both from BD Biosciences; Figure 1).

Figure 1. FACS plot of Melan-A/Mart-1 specific CD8+ T-cells (Patient 1) using phycoerythrin conjugated ELAGIGTV-pentamer. The results show an increase from 1.09% before isolated limb perfusion to 6.08% after 4 weeks.

Pentamer staining for detection of glycoprotein 100 and telomerase specific cytotoxic T cells

In a pilot series including five of the 12 patients included, staining was also performed for glycoprotein 100 (gp100) and telomerase with phycoerythrin conjugated ITDQVPFSV- and ILAKFLHWL-pentamer respectively (both from Orpegen, Heidelberg, Germany). However, no measurable level was detected, and we therefore decided not to include the analyses of gp100 and telomerase in the remaining patients (data not shown).

Major lymphocyte subset in peripheral blood

Freshly drawn EDTA blood was stained with the following fluorochrome-conjugated monoclonal antibodies: CD3-PerCP, CD4-PE, CD5-FITC, CD8-PE, CD14-PE, CD45-FITC, CD45RA-FITC, CD45RO-FITC, CD56-PE and HLA-DR-PerCP (BD Biosciences). After incubation the erythrocytes were lysed and the cells were washed. For absolute quantification BD TrueCount beads (BD Biosciences) were used according to manufacturer protocol.

Response evaluation

Clinical responses are reported as the best response according to the WHO criteria [12]. CR included the disappearance of all lesions, partial response (PR) was defined as a decrease of more than 50% of the total tumour burden. Progressive disease (PD) was defined as an increase of more than 25% in existing lesions or the appearance of new lesions. No change (NC) was defined as a result where the criteria for neither CR, PR nor PD are met. Local progress is defined as the appearance of new lesions or progress of existing lesions within the treated limb, not including lymph-node metastases. Patients were followed at the discretion of the referring physician.

Statistics

When comparing Melan-A/Mart-1 specific cytotoxic T cells over time a non-parametric paired rank test was used (Wilcoxon signed rank two-tailed exact test). Comparison between lymphocyte populations over time were made using ANOVA. All statistics were calculated using SPSS® 19 Statistical Software (SPSS, Chicago, IL, USA) and at a significance level of 5%.

Results

Clinical response

All patients underwent ILP without any complications and the clinical response rate was 83% (58% CR, 25% PR and 17% NC). No patients were lost to follow-up, six patients developed local progressive disease after 19 months. Eight patients developed general metastasis after a median of 16 months. Median cancer-specific survival was 27 months with a 1-year and 2-year survival of 75% and 67% respectively.

Melan-A specific CD8+ CTLs

Four weeks following ILP there was a statistically significant (p = 0.024) increase in Melan-A/MART-1 specific CD8+ CTLs (percentage of the total number of T-cells). However, the absolute numbers are very small with a median (IQR) of 0.17% (0.98) and a mean (SD) of 0.86% (1.55) respectively. There was a trend (p = 0.055) towards a decrease in Melan-A specific CTLs after 1 week with a median (IQR) of 0.10% (0.21). When comparing 1 week with 4 weeks there was a significant increase (p = 0.016) with a median (IQR) of 0.09% (0.73). Yet again, the median and mean absolute numbers are small, which are explained by the fact that out of the 12 included patients there are only four patients who could be regarded as immunological responders. These four patients show an increase between 0.84–4.99% in Melan-A CD8+ specific CTLs after 4 weeks, where the remaining patients show no real difference (−0.13–0.30%). At 12 weeks after ILP, the amount of Melan-A specific CTLs seems to remain at baseline; however, samples were only available for five patients and only for two of the four patients considered as immunological responders.

Major lymphocyte populations

No statistical significant changes in major lymphocyte subpopulations were observed during the follow-up of the individual patients (Figure 2). When looking at predictive factors for clinical response, CD3+8+ cells (1.06 ± 0.63 versus 0.32 ± 0.17; p = 0.01), CD3+8+45RA+ cells (0.69 ± 0.42 versus 0.11 ± 0.08; p = 0.003) and CD3+DR+ (0.61 ± 0.44 versus 0.10 ± 0.05; p = 0.006) were significantly higher preoperatively in the group of patients with a CR after ILP (n = 6) compared to patients without CR (NC+PR, n = 6) (Figure 3). There were no significant predictive factors in the major lymphocyte populations for the patients regarded as immunological responders (n = 4) compared to the non-responding patients (n = 8).

Figure 2. Levels (mean ± SD) of different lymphocyte subpopulations before and after (1 week, 4 weeks and 3 months) isolated limb perfusion. No significant changes were observed.

Figure 3. (a) Of the analyzed subpopulations in Figure 2, the levels of CD3+8+, CD3+8+45RA+ and CD3+DR+ cells were significant for clinical response. (b) Levels of Melan-A specific CD8+ T-cells during 12 weeks follow-up after isolated limb perfusion.

Discussion

The rationale for this study was recent preclinical data indicating that hyperthermia as well as apoptosis may be immunogenic. In line with the hypothesis that part of the anti-tumour effect of conventional chemotherapy is caused by direct killing of tumour cells and part by indirect killing via stimulation of anti-tumour immunity [13], we tried to investigate the effect of hyperthermic ILP on the tumour specific immune response measured as the number of Melan-A specific CD8+ T cells.

High doses of some chemotherapeutic compounds are immunosuppressive, while low doses stimulate the immune response by, for example, increasing the expression of HLA class I antigens on the surface of tumour cells, increasing cross-presentation of tumour antigens to CD8+ T cells and decreasing immunosuppressive effects by myeloid derived suppressor cells and regulatory T cells. It is still unknown if the DNA alkylating agent melphalan is able to induce a full-blown immunogenic cell death and what its effects are on the immune response at lower doses. However, cyclophosphamide, another DNA alkylating agent, has recently been shown to induce tumour apoptosis with strong immunogenic features, including surface translocation of calreticulin [9].

Our main finding was that no significant increase in major lymphocyte subpopulations or Melan-A specific CD8+ CTLs was induced by ILP. However, in four of the 12 patients (33%), a small but significant increase of Melan-A specific CD8+ CTLs was found 4 weeks after ILP. Some major limitations of the present study are that only 12 patients were included and that we were able to follow them for only 4 weeks. Moreover, only Melan-A specific T cells were recorded and our results have therefore to be seen as preliminary observations that have to be interpreted with caution; we cannot exclude long-term immunogenic effects or the induction of CTLs against epitopes other than Melan-A.

Much more work is therefore necessary to delineate fully the systemic immune response to hyperthermic ILP. Such work has to be performed on larger patient groups, and include markers that detect circulating CTLs against tumour-associated epitopes other than Melan-A, together with immunohistochemical analysis of both tumour and draining lymph-node infiltrating CTLs [14–16].

The finding that CD3+8+ cells, CD3+8+45RA+ cells and CD3+DR+ were higher preoperatively in patients with CR after ILP has to be seen as very preliminary and certainly needs validation. We are currently following patients prospectively to address this finding in a larger series.

With new immunologically active cancer treatments, such as ipilimumab or anti-PD-1 antibodies, a transient increase in tumour specific T-cells after ILP could be synergistic leading to an increased immunological response.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.