Frequencies of circulating MDSC correlate with clinical outcome of melanoma patients treated with ipilimumab

Abstract

Metastatic melanoma has a poor prognosis with high resistance to chemotherapy and radiation. Recently, the anti-CTLA-4 antibody ipilimumab has demonstrated clinical efficacy, being the first agent to significantly prolong the overall survival of inoperable stage III/IV melanoma patients. A major aim of patient immune monitoring is the identification of biomarkers that predict clinical outcome. We studied circulating myeloid-derived suppressor cells (MDSC) in ipilimumab-treated patients to detect alterations in the myeloid cell compartment and possible correlations with clinical outcome. Lin(-) CD14(+) HLA-DR(-) monocytic MDSC were enriched in peripheral blood of melanoma patients compared to healthy donors (HD). Tumor resection did not significantly alter MDSC frequencies. During ipilimumab treatment, MDSC frequencies did not change significantly compared to baseline levels. We observed high inter-patient differences. MDSC frequencies in ipilimumab-treated patients were independent of baseline serum lactate dehydrogenase levels but tended to increase in patients with severe metastatic disease (M1c) compared to patients with metastases in skin or lymph nodes only (M1a), who had frequencies comparable to HD. Interestingly, clinical responders to ipilimumab therapy showed significantly less lin(-) CD14(+) HLA-DR(-) cells as compared to non-responders. The data suggest that the frequency of monocytic MDSC may be used as predictive marker of response, as low frequencies identify patients more likely benefitting from ipilimumab treatment. Prospective clinical trials assessing MDSC frequencies as potential biomarkers are warranted to validate these observations.

Fig. 1

MDSC frequencies are increased in patients with melanoma. PBMC from healthy donors and melanoma patients were assessed by flow cytometry. a % lin⁻ CD14⁺ HLA-DR⁻ cells (monocytic MDSC) in healthy donors (HD; 0.69 ± 0.67 %, N = 15) and melanoma patients (2.57 ± 2.97 %, N = 49). Symbols in the figure correspond to either 1 blood sample per patient or the average frequency from several blood samples obtained per patient. Melanoma treatments received: tumor surgery only (circles), ipilimumab (squares), vemurafenib (triangles), ipilimumab followed by vemurafenib (diamonds). b Representative dot plots for one healthy donor and c one melanoma patient. *p < 0.05

Fig. 2

MDSC in untreated patients with and without active diseases. PBMC from healthy donors and untreated melanoma patients were assessed by flow cytometry. a % lin⁻ CD14⁺ HLA-DR⁻ cells in HD (0.69 ± 0.67 %, N = 15) and non-treated melanoma patients (1.86 ± 2.56 %, N = 23). b % lin⁻ CD14⁺ HLA-DR⁻ cells in untreated patients after tumor resection (1.36 ± 1.10 %, N = 6) and untreated patients with tumor burden (2.03 ± 2.92 %, N = 17). Symbols in the figure correspond to 1 blood sample per patient

Fig. 3

MDSC in ipilimumab-treated patients and in patients with different disease stages. PBMC from melanoma patients receiving ipilimumab treatment were assessed by flow cytometry. a % lin⁻ CD14⁺ HLA-DR⁻ cells in melanoma patients receiving ipilimumab treatment. MDSC frequencies are displayed for each day of blood withdrawal: baseline 3.02 ± 3.53 % (N = 8), t1d3 1.07 ± 0.75 % (N = 7), t1d7 1.69 ± 1.90 % (N = 8), before t3 1.00 ± 0.75 % (N = 5), t3d7 2.10 ± 1.77 % (N = 6) and t4d20 2.09 ± 2.50 % (N = 3). b MDSC frequencies are displayed at baseline (3.02 ± 3.53 %, N = 8) and during/after treatment (1.46 ± 1.17 %, N = 8). c % lin⁻ CD14⁺ HLA-DR⁻ cells in ipilimumab-treated patients with normal baseline serum LDH level (1.73 ± 1.42 %, N = 5) and elevated baseline serum LDH level (2.31 ± 1.55 %, N = 9). d % lin⁻ CD14⁺ HLA-DR⁻ cells in HD (0.69 ± 0.67 %, N = 15) and ipilimumab-treated patients with metastatic staging M1a (0.78 ± 0.45 %, N = 4), M1b (N = 2) and M1c (2.58 ± 1.69 %, N = 8). Symbols in the figure correspond to either 1 blood sample per patient or the average frequency from several blood samples obtained per patient. Clinical responder (triangle upward) and clinical non-responder (triangle downward). ***p < 0.001. LDH serum lactate dehydrogenase

Fig. 4

MDSC in patients responding or non-responding to ipilimumab treatments. PBMC from melanoma patients receiving ipilimumab treatment were assessed by flow cytometry. a % lin⁻ CD14⁺ HLA-DR⁻ cells in ipilimumab-treated patients responding (1.24 ± 1.11 %, N = 8) or non-responding (3.06 ± 1.51 %, N = 6) to the treatment. b Baseline levels of % lin⁻ CD14⁺ HLA-DR⁻ cells for clinical responders (0.93 ± 0.75 %, N = 3) and clinical non-responders (4.85 ± 4.42 %, N = 4) and frequencies during treatment for clinical responders (1.00 ± 0.97 %, N = 3) and non-responders (1.87 ± 1.45 %, N = 4). c % lin⁻ CD14⁺ HLA-DR⁻ cells in ipilimumab-treated patients with distant metastasis (M1c) responding (1.39 ± 1.94 %, N = 3) or non-responding (3.77 ± 0.59 %, N = 4) to the treatment. Symbols in the figure correspond to either 1 blood sample per patient or the average frequency from several blood samples obtained per patient. *p < 0.05