Prognostic impact of the putative cancer stem cell markers ABCG 2, CD 133, ALDH 1A1 and CD 44V7/8 in metastatic melanoma

This is the authors' abstract. We don't add key points for this paper.

Dear Editor, Despite recent therapeutic advances, a diagnosis of metastatic melanoma still foreshadows a grim prognosis for the majority of patients.1 2 So‐called cancer stem cells (CSCs) are believed to be associated with the failure of current conventional cancer therapies.3 CSCs share common features with corresponding tissue stem cells, such as self‐renewal capacity and the ability to give rise to progeny with the potential to proliferate and differentiate. Thus, therapeutically targeting CSCs may eliminate the root cause of the cancer; however, this would require the accurate identification of markers distinguishing CSCs from normal cells. The expression of candidate CSC markers is associated with a poor prognosis in a number of cancer types,4,5,6,7,8 but their clinical significance remains unclear, and to the best of our knowledge there have been no clinical studies in melanoma. Here, we aimed to determine the clinical significance of four molecules identifying putative CSCs: CD133, ABCG2, ALDH1A1 and CD44v7/8. Using fluorescence microscopy we determined the expression of CD133, ABCG2, ALDH1A1 and CD44v7/8 in formalin‐fixed, paraffin‐embedded metastatic lesions from 40 patients [24 men, 16 women; 15 patients with stage III melanoma, 25 patients with stage IV melanoma; mean age 63 years (range 39–89); survival time < 5 years (see Table 1 for a detailed description of the patient cohort)] treated at Tübingen University Hospital's dermatology department (ethics approval number 017/2016BO2). The following antibodies were used to detect CSC markers in melanoma tissues: ALDH1A1 rabbit monoclonal [lot: GR41450‐6; clone: EP1933Y (Abcam, Cambridge, U.K.)], CD133 rabbit polyclonal (lot: X13030523; Fitzgerald, Acton, MA, U.S.A.), ABCG2 mouse monoclonal [lot: D15KF02234; clone: BXP‐21 (Biolegend, San Diego, CA, U.S.A.)] and CD44v7/8 mouse monoclonal [lot: 051114; clone: VFF‐17 (Bio‐Rad, Hercules, CA, U.S.A.)]. When comparing expression of these proteins between melanoma and stromal cells, the Melan A mouse monoclonal antibody [lot: 00060204; clone: A103 (Dako, Hamburg, Germany)] was used to confirm the identity of melanoma cells in a subset of samples (n = 8, data not shown). The secondary antibodies Alexa Fluor 488 donkey antirabbit IgG (H+L) and Cy3 donkey antimouse IgG (H+L) (both from Jackson ImmunoResearch Laboratories, West Grove, PA, U.S.A.) were used for all experiments. Optimum staining conditions were determined for all antibodies individually. Tissues were stained according to a previously described protocol but using the following antigen retrieval method.9 10 Fluorescence intensity of each tissue was assessed by comparison with a control tissue (mounted to the same slide and stained with secondary antibody only) to give a fluorescence index (fold‐increase in stained tissue over control tissue). An average of 12 images per tissue covering the entire tumour were captured at ×20 magnification. Statistical analyses were performed using GraphPad Prism version 6 (GraphPad Inc., La Jolla, CA, U.S.A.) and SPSS version 24 (IBM, Ehningen, Germany). Clinical characteristics of the patient cohort Data are n (%). NR, not relevant; NA, not available. Clinical characteristics of the patient cohort Data are n (%). NR, not relevant; NA, not available. We observed that the majority of tissue specimens were positive for ALDH1A1 (n = 28), ABCG2 (n = 32) and CD133 (n = 38), while CD44v7/8 was expressed less frequently (n = 15) (Fig. 1a). In general, we observed similar expression of these proteins by all tumour cells within a tissue. Representative tissues stained for each protein are shown in Figure 1(b). Of note was that compared with stromal cells, melanoma cells expressed higher levels of all four proteins (Fig. 1c, selected results shown). Furthermore, we found that expression of the CSC marker ABCG2 correlated with CD133 (P < 0·0001) and CD44v7/8 (P = 0·021) in melanoma tissues (Fig. 1d). We confirmed coexpression of ALDH1A1 with CD44v7/8 and ABCG2 with CD133 (data not shown) (owing to the experimental set‐up other combinations of coexpression were not possible to test directly). We then evaluated expression levels of ALDH1A1, ABCG2, CD133 and CD44v7/8 for association with clinical features, including stage, M category, progression time (stage III–IV and diagnosis to stage IV), age and sex. We found that patients with stage IV disease had tumours with higher levels of ABCG2 than those with stage III melanoma (P = 0·003) (Fig. 1e), while no other clinical correlations were identified. We then sought to determine if these proteins were relevant for patient survival. Using the Kaplan–Meier method, univariate analysis showed that ABCG2 (log‐rank P < 0·001) and CD133 (log‐rank P = 0·021) but not ALDH1A1 or CD44v7/8 correlated with survival of these patients with metastatic melanoma (Fig. 1f). Multivariate survival analysis considering patient age, sex, disease stage and M category showed ABCG2 (P = 0·017) to be an independent prognostic factor. Cancer stem cell (CSC) markers are widely expressed in melanoma tissues and correlate with patient clinical features. (a) Melanoma tissues commonly express CD133, ABCG2, ALDH1A1 and CD44v7/8. (b) Representative images (×40 magnification) of tissues positive for each of the CSC markers investigated. (c) Melanoma cells show higher expression of CSC markers than surrounding nonmelanoma cells (ALDH1A1 and CD44v7/8 not shown) (****P < 0·0001). (d) ABCG2 expression correlates with the levels of CD133 and CD44v7/8. (e) Higher levels of ABCG2 were found in patients with stage IV melanoma than in those with stage III melanoma. (f) Kaplan–Meier survival analysis shows that patients with high levels of ABCG2 and CD133 experience worse prognosis [cohort divided into ‘high’ and ‘low’ based on the median fluorescence index (MFI)]. The concept of CSCs implies that successful cancer therapy will depend on eliminating these cells, but current therapies mostly remain ineffective at achieving this. ABCG2 and CD133, found here to correlate with patient survival, may make useful therapeutic targets that show preferential activity against tumour cells; they were expressed by essentially every melanoma cell in the tumour and were found at higher levels in melanoma cells than in surrounding nonmelanoma cells. To the best of our knowledge this is the first report to demonstrate that the CSC markers ABCG2 and CD133 are negative prognostic factors in melanoma, which may set the stage for their validation as novel therapeutic targets. We are grateful to Christof Zanke of the University Hospital Tübingen for producing software used to assess fluorescence microscopy images.