Comparative Expression Analysis of Putative Cancer Stem Cell Markers CD44 and ALDH1A1 in Various Skin Cancer Subtypes

Abstract

Introduction

Skin cancers, particularly melanoma, are initiated and maintained by a subpopulation of tumor cells expressing stemness markers that are called cancer stem cells (CSCs). This study aimed to evaluate the expression levels and clinicopathological significance of the putative CSC markers CD44 and ALDH1A1 in patients with skin cancer.

Methods

The expression levels of CD44 and ALDH1A1 were investigated in 107 skin cancer specimens including 58 (54%) basal cell carcinomas (BCC), 37 (35%) squamous cell carcinomas (SCC), and 12 (11%) melanomas using the tissue microarray (TMA) technique. The correlation of the expression levels of these markers and clinicopathological parameters was then analyzed.

Results

The expression levels of CD44 and ALDH1A1 were significantly higher in melanoma patients than patients with SCC or BCC (p<0.001 and p = 0.002, respectively). A higher level of CD44 expression was more often found in melanoma tumor cells with a higher rate of recurrence (p = 0.029) and in SCC cases with ulceration (p = 0.01), while there was no significant correlation between ALDH1A1 expression and other clinicopathological parameters. Similarly, coexpression of CD44 and ALDH1A1 (CD44^high/ALDH1A1^high) was significantly observed in melanoma samples (p<0.001).

Conclusions

These findings suggest that a CD44^high/ALDH1A1^high phenotype in melanoma and a CD44^high phenotype in SCC can be considered candidates for targeted therapy of skin cancers aiming at CSCs.

Keywords

ALDH1A1 Cancer stem cells (CSCs)CD44 Melanoma Tissue microarray (TMA)

Introduction

Skin cancers are mainly divided into 2 major groups: non-melanoma skin cancer and melanoma. Non-melanoma skin cancer is the most common type of cancer and includes basal cell carcinoma (BCC), which comprises nearly 80% of all non-melanoma skin cancer cases, and squamous cell carcinoma (SCC), which is observed with a lower frequency (20% of all non-melanoma skin cancer cases) (1, 2). BCC and SCC are rarely fatal, whereas melanoma, which accounts for only 4% of skin cancers, is the lone culprit in 74% of all skin cancer-related deaths (1, 2). Although early-stage melanoma is curable, the prognosis of patients with advanced melanoma remains poor, with a median survival of only a few months (3, 4). The high mortality and morbidity rates of melanoma result from the high metastatic potential and the tumors’ resistance to chemo- and radiotherapy regimens (5). According to the cancer stem cell theory, only a distinct subset of cells, termed cancer stem cells (CSCs) or tumor-initiating cells (TICs), can drive tumor initiation, maintenance and progression as well as drug resistance (6, 7). CSCs have been isolated from a wide spectrum of malignancies using cell surface markers such as CD44 and ALDH1A1 (8-9-10).

CD44 is a transmembrane glycoprotein that plays an important role in growth, cell proliferation, regulation of cell adhesion and migration (11, 12). CD44, alone or in combination with other putative stem cell markers, is extensively applied for identifying CSCs in breast (13), prostate (14) and colon (15) cancers, while very limited studies have specifically investigated the differential expression of CD44 in melanoma tissue samples as a putative CSC marker (16). Although there is increasing evidence for the role of CSCs in melanoma progression (5, 17, 18), their relative importance is still unclear; for example, Quintana et al (16) showed that an average of 1 in 4 isolated cells from primary or metastatic melanoma was able to form tumors in non-obese diabetic/severe combined immunodeficiency (NOD/SCID) mice.

ALDH1A1 has been increasingly recognized in recent years as a potential CSC marker (9, 10, 19-20-21). The enzymatic activity of ALDH1A1 catabolizes the oxidation of intracellular aldehydes and thus provides resistance to some alkylating agents (22). Through the evaluation of human melanoma specimens it has been suggested that ALDH1A1-positive melanoma cells have higher tumorigenicity than ALDH1A1-negative cells, and thus ALDH1A1 may be a promising melanoma CSC marker (8-9-10); other studies, however, did not confirm these findings (23).

Given the lack of data regarding immunohistochemical expression of ALDH1A1 and CD44 as potential CSC markers in skin cancer, the present study was designed to explore the expression patterns of these markers by tissue microarray (TMA) in a well-defined series of skin cancer specimens including BCC, SCC and melanoma. We also evaluated the association between the level of expression of each marker and clinicopathological characteristics.

Material and methods

Patients and Tumor Characteristics

Archival tissue samples (107 tumor specimens including 58 BCCs, 37 SCCs and 12 melanomas) from patients with primary skin cancer diagnosed in the Shohada-e-Tajrish Hospital, a referral and teaching hospital in Tehran, Iran, were included in this study. Patients were diagnosed between 2003 and 2011 and medical records were reviewed to collect clinicopathological data including age, gender, tumor size, ulceration, tumor-infiltrating lymphocytes (TILs), Breslow thickness and Clark level (in melanoma), and histological grade (in SCC). Patients’ data were kept fully anonymous. The study protocol was approved by the Research Ethics Committee of Iran University of Medical Sciences.

Construction of TMA

Skin cancer TMAs were prepared as described previously (24, 25). Briefly, all hematoxylin and eosin-stained slides were reviewed to determine the best area for preparing the TMA from each specimen. Tissue arrays were then constructed by placing 0.6-mm diameter samples from 107 different tumor samples per single block, with 1-mm spacing separating TMA dots. The TMA blocks were constructed in 3 copies, each containing 1 sample from a different region of the tumor using a tissue-arraying instrument (Minicore; ALPHELYS, Plaisir, France). These TMAs were used for immunohistochemical staining of CD44 and ALDH1A1. The mean score of 3 cores was calculated as the final score.

Immunohistochemistry

Immunohistochemistry was performed on 4-µm-thick sections of skin cancer TMA slides (Superfrost Plus, Thermo Scientific, Germany) as previously described (26, 27). The sections were deparaffinized in xylene and rehydrated by passing them through graded ethanol series. The endogenous peroxidase activity was blocked by incubating in 3% hydrogen peroxide for 20 minutes at room temperature. Antigen retrieval was achieved by autoclaving slides in citrate buffer (pH = 6.0) for 10 minutes for CD44 antibody and 20 minutes microwaving for ALDH1A1 antibody. After 3 washes with Tris-buffered saline (TBS), the sections were incubated overnight at 4°C with specific antibodies against mouse monoclonal anti-human CD44 (Product code = NCL-CD44-2, Novocastra, UK) or anti-ALDH1A1 (ab52492, Abcam, Cambridge, UK) antigens with optimal dilution, 1:40 and 1:200, respectively. After washing, the TMA slides were incubated with anti-rabbit/anti-mouse EnVision reagent (Dako, Denmark) as secondary antibody for 10 minutes. The immunohistochemical staining was visualized with 3, 3’-diaminobenzidine (DAB, Dako) substrate as chromogen for 15 minutes at room temperature. The sections were lightly counterstained with hematoxylin, dehydrated through graded ethanol followed by xylene and mounted. Normal human tonsil tissue was used as positive control for CD44 antibody and normal liver was used as positive control for ALDH1A1 antibody. Incubation with only TBS was used as negative control.

Immunostaining Evaluation

Stained slides were examined by a semiquantitative scoring system on a multiheaded microscope by 2 observers (AR and ZM), and scoring was performed in a blinded manner to all the clinical data. In difficult cases, the scoring was confirmed by 2 observers and consensus was reached. Initially, the slides were scanned at 10× magnification to obtain a general impression of the overall distribution of the tumor cells. Positive cores were then assessed for localization and semiquantitatively for expression level at higher magnifications, and the final scores were given. Intensity of staining was scored as 0 (absent), 1 (weak), 2 (moderate) or 3 (strong). Also the percentage of positive tumor cells was scored as 1 (positive tumor cells <25%), 2 (positive tumor cells 25%-50%), 3 (positive tumor cells 50%-75%) and 4 (positive tumor cells >75%). Finally, the histochemical score (H-score) was calculated by multiplying the intensity of staining and the percentage of positive tumor cells.

Statistical Analysis

Statistical analyses were performed using the SPSS software version 20 (SPSS, Chicago, IL, USA). Associations between CD44 and ALDH1A1 status and clinicopathological parameters were assessed with Pearson's chi-square and Pearson's R tests. Moreover, CD44 and ALDH1A1 expression in different skin cancer types was compared using the Mann-Whitney U test. The relationships between CD44/ALDH1A1 status and clinicopathological parameters were analyzed by 1-way ANOVA and Tukey's post hoc analysis. A p value of <0.05 was considered statistically significant.

Results

Study Population

Of 107 skin cancer samples, 95 (89%) were non-melanoma skin cancer including 58 (54%) BCC and 37 (35%) SCC, whereas 12 (11%) specimens were melanoma (Tabs. I, II and III). There was a male predominance in all groups of skin cancer with different male/female ratios: 1.52 (35/23) in BCC, 3.11 (28/9) in SCC, and 3 (9/3) in melanoma. Of the 58 BCC samples, only 4 (7%) patients showed local recurrence without any metastasis. Among the SCC samples, 13 (36%) patients showed local recurrence, 12 (32%) had no local recurrence and there were no available data for 12 (32%) specimens. Nine (24%) SCC patients had metastases, 18 (49%) had no metastasis, and no data were available for 10 (27%) cases. From the data available for local recurrence in melanoma cases, 5 (42%) patients had local recurrence and 4 (33%) had no local recurrence. Dermal invasion was reported in 8 (67%) melanoma specimens, whereas 4 (33%) patients had in situ melanoma. TILs, as a prognostic factor, were categorized as absent, moderate and brisk levels (28, 29). In melanoma cases, tumor size ranged from 1 mm to 13 mm (mean = 4.1 mm); 8 (67%) had a mean size ≤4.1 mm and 4 (33%) a mean size >4.1 mm. Breslow thickness was categorized as thin melanoma (Breslow thickness ≤1 mm) and thick melanoma (Breslow thickness >1 mm) (30). Thin melanoma was found in 8 (67%) cases and thick melanoma in 4 (33%) cases. Mitotic count ≤6 and mitotic count >6 had similar frequencies (50%). The Clark levels in melanoma were also divided into 2 groups: group 1 (Clark levels I and II) and group 2 (Clark levels III through V) (28). Nine (75%) melanoma sample were categorized as group 2 and 3 (25%) as group 1 (Tab. III).

Table I

Association of CD44 and ALDH1A1 expression with clinicopathological parameters in basal cell carcinoma

Clinicopathological characteristics	Total number (%)	CD44 expression (mean H-score = 21)		P value	ALDH1A1 expression (mean H-score = 10)		P value
		Low	High		Low	High
Mean age ± SD (60 ± 14.32 years)
≥60	27 (47)	21 (78)	6 (22)	0.55	23 (85)	4 (15)	0.12
<60	31 (53)	26 (84)	5 (16)		21 (68)	10 (32)
Gender
Male	35 (60)	26 (74)	9 (26)	0.1	26 (74)	9 (26)	0.72
Female	23 (40)	21 (91)	2 (9)		18 (78)	5 (22)
History of burn
Yes	1 (2)	1 (100)	0 (0)	0.62	1 (100)	0 (0)	0.56
No	57 (98)	46 (81)	11 (19)		43 (75)	14 (25)
Ulceration
Yes	31 (53)	24 (77)	7 (23)	0.45	24 (77)	7 (23)	0.76
No	27 (47)	23 (85)	4 (15)		20 (74)	7 (26)
Histological subtypes
Solid	30 (52)	22 (73)	8 (27)	0.58	21 (70)	9 (30)	0.34
Micronodular	9 (16)	8 (89)	1 (11)		8 (89)	1 (11)
Adenoid	6 (10)	6 (100)	0 (0)		5 (83)	1 (17)
Pigmented	5 (9)	4 (80)	1 (20)		5 (100)	0 (0)
Superficial	3 (5)	3 (100)	0 (0)		2 (67)	1 (33)
Keratotic	2 (3)	2 (100)	0 (0)		1 (50)	1 (50)
Infiltrative	2 (3)	1 (50)	1 (50)		2 (100)	0 (0)
Superficial spreading	1 (2)	1 (100)	0 (0)		0 (0)	1 (100)
Tumor-infiltrating lymphocytes
Absent	34 (59)	26 (77)	8 (23)	0.54	29 (85)	5 (15)	0.09
Moderate	23 (40)	20 (87)	3 (13)		14 (61)	8 (39)
Brisk	1 (1)	1 (100)	0 (0)		1 (100)	0 (0)
Severity of inflammatory infiltration
No infiltration	8 (14)	6 (75)	2 (25)	0.11	6 (75)	2 (25)	0.94
Weak	26 (45)	18 (69)	8 (31)		20 (77)	6 (23)
Moderate	23 (40)	22 (96)	1 (4)		17 (74)	6 (26)
Severe	1 (1)	1 (100)	0 (0)		1 (100)	0 (0)
Margin involvement
Yes	33 (57)	24 (73)	9 (27)	0.06	26 (79)	7 (21)	0.55
No	25 (43)	23 (92)	2 (8)		18 (72)	7 (28)
Local recurrence
Yes	4 (7)	4 (100)	0 (0)	0.31	4 (100)	0 (0)	0.24
No	54 (93)	43 (80)	11 (20)		40 (74)	14 (26)

Table II

Association of CD44 and ALDH1A1 expression with clinicopathological parameters in squamous cell carcinoma

Clinicopathological characteristics	Total number (%)	CD44 expression (mean H-score = 40)		P value	ALDH1A1 expression (mean H-score = 10)		P value
		Low	High		Low	High
Mean age ± SD (63 ± 17.28 years)
≤63	17 (46)	9 (53)	8 (47)	0.66	15 (88)	2 (12)	0.86
>63	20 (54)	12 (60)	8 (40)		18 (90)	2 (10)
Gender
Male	28 (76)	15 (54)	13 (46)	0.49	26 (93)	2 (7)
Female	9 (24)	6 (67)	3 (33)		7 (78)	2 (22)	0.2
Histological grade
Well	12 (33)	7 (58)	5 (42)	0.42	11 (92)	1 (8)	0.8
Moderate	23 (62)	12 (52)	11 (48)		20 (87)	3 (13)
Poor	2 (5)	2 (100)	0 (0)		2 (100)	0 (0)
Metastasis
Yes	9 (24)	6 (67)	3 (33)	0.33	8 (89)	1 (11)	0.49
No	18 (49)	8 (44)	10 (56)		17 (94)	1 (6)
No data	10 (27)	7 (70)	3 (30)		8 (80)	2 (20)
Local recurrence
Yes	13 (36)	10 (77)	3 (23)	0.08	12 (92)	1 (8)	0.72
No	12 (32)	4 (33)	8 (67)		11 (92)	1 (8)
No data	12 (32)	7 (58)	5 (42)		10 (83)	2 (17)
History of burn
Yes	5 (13)	3 (60)	2 (40)	0.9	4 (80)	1 (20)	0.73
No	24 (65)	13 (54)	11 (46)		22 (92)	2 (8)
No data	8 (22)	5 (63)	3 (37)		7 (88)	1 (12)
Ulceration
Yes	27 (73)	15 (56)	12 (44)	0.43	24 (89)	3 (11)	0.12
No	8 (22)	4 (50)	4 (50)		8 (100)	0 (0)
No data	2 (5)	2 (100)	0 (0)		1 (50)	1 (50)
Tumor-infiltrating lymphocytes
Absent	13 (35)	10 (77)	3 (23)	0.18	10 (77)	3 (23)	0.19
Moderate	18 (49)	8 (44)	10 (56)		17 (94)	1 (6)
Brisk	6 (16)	3 (50)	3 (50)		6 (100)	0 (0)
Severity of inflammatory infiltration
Weak	13 (35)	10 (77)	3 (23)	0.18	10 (77)	3 (23)	0.19
Moderate	18 (49)	8 (44)	10 (56)		17 (94)	1 (6)
Severe	6 (16)	3 (50)	3 (50)		6 (100)	0 (0)
Margin involvement
Yes	23 (62)	11 (48)	12 (52)	0.31	22 (96)	1 (4)	0.20
No	13 (35)	9 (69)	4 (31)		10 (77)	3 (23)
No data	1 (3)	1 (100)	0 (0)		1 (100)	0 (0)

Table III

Association of CD44 and ALDH1A1 expression with clinicopathological parameters in melanoma

Clinicopathological characteristics	Total number (%)	CD44 expression (mean H-score = 156)		P value	ALDH1A1 expression (mean H-score = 58)		P value
		Low	High		Low	High
Mean age ± SD (59 ± 13.89 years)
≤59	8 (67)	3 (37)	5 (63)	0.67	4 (50)	4 (50)	0.08
>59	4 (33)	2 (50)	2 (50)		4 (100)	0 (0)
Gender
Male	9 (75)	4 (44)	5 (56)	0.73	8 (89)	1 (11)	0.005
Female	3 (25)	1 (33)	2 (67)		0 (0)	3 (100)
Tumor size
≤4.1 mm	8 (67)	5 (63)	3 (37)	0.038	6 (75)	2 (25)	0.38
>4.1 mm	4 (33)	0 (0)	4 (100)		2 (50)	2 (50)
Dermal invasion
Yes	8 (67)	3 (37)	5 (63)	0.67	6 (75)	2 (25)	0.38
No (in situ melanoma)	4 (33)	2 (50)	2 (50)		2 (50)	2 (50)
Metastasis
Yes	11 (92)	5 (45)	6 (54)	0.37	7 (64)	4 (36)	0.46
No	1 (8)	0 (0)	1 (100)		1 (100)	0 (0)
Local recurrence
Yes	5 (42)	2 (40)	3 (60)	0.029	4 (80)	1 (20)	0.63
No	4 (33)	0 (0)	4 (100)		2 (50)	2 (50)
No data	3 (25)	3 (100)	0 (0)		2 (67)	1 (33)
Tumor-infiltrating lymphocytes
Absent	10 (84)	4 (40)	6 (60)	0.34	7 (70)	3 (30)	0.27
Moderate	1 (8)	0 (0)	1 (100)		1 (100)	0 (0)
Brisk	1 (8)	1 (100)	0 (0)		0 (0)	1 (100)
Severity of inflammatory infiltration
No infiltration	6 (50)	1 (17)	5 (83)	0.13	3 (50)	3 (50)	0.15
Weak	4 (34)	3 (75)	1 (25)		4 (100)	0 (0)
Moderate	1 (8)	0 (0)	1 (100)		1 (100)	0 (0)
Severe	1 (8)	1 (100)	0 (0)		0 (0)	1 (100)
Margin involvement
Yes	2 (17)	1 (50)	1 (50)	0.79	1 (50)	1 (50)	0.58
No	10 (83)	4 (40)	6 (60)		7 (70)	3 (30)
Mitoses (per 10 HPF)
≤6	6 (50)	2 (33)	4 (67)	0.55	3 (50)	3 (50)	0.22
>6	6 (50)	3 (50)	3 (50)		5 (83)	1 (17)
Breslow thickness
Thin (≤1 mm)	8 (67)	4 (50)	4 (50)	0.4	4 (50)	4 (50)	0.08
Thick (>1 mm)	4 (33)	1 (25)	3 (75)		4 (100)	0 (0)
Clark level
Group 1 (I, II)	3 (25)	1 (33)	2 (67)	0.73	3 (100)	0 (0)	0.15
Group 2 (III-V)	9 (75)	4 (44)	5 (56)		5 (56)	4 (44)
Lymph node involvement
Yes	8 (67)	3 (37)	5 (63)	0.67	5 (63)	3 (37)	0.66
No	4 (33)	2 (50)	2 (50)		3 (75)	1 (25)

The values in bold are statistically significant.

Analysis of CD44 Expression and Its Correlation with Clinicopathological Features

Each tumor type was divided into either lower (≤ mean of H-scores) or higher (> mean of H-scores) CD44 expression. The mean H-scores in BCC and SCC samples were 21 and 40, respectively, whereas the mean H-score in melanoma cases was 156. Forty-seven of 58 BCC samples (81%) expressed lower levels of CD44, while 11 (19%) cases expressed higher levels (Fig. 1). Of the 37 SCC samples, 21 (57%) had low expression and 16 (43%) had high expression of CD44 (Fig. 1). Low expression of CD44 was seen in 2 (17%) of 12 melanoma samples, while high expression was found in 10 (83%) cases (Fig. 1). The expression of CD44 (in terms of intensity staining, percentage of positive cells and H-score) showed a highly significant difference between BCC, SCC and melanoma cases (all p<0.001). Moreover, the Mann-Whitney U test indicated a significant difference in CD44 expression between BCC and SCC (p = 0.001 for percentage of positive cells and p = 0.003 for H-score) and melanomas (all p<0.001). Also the expression of CD44 was significantly different between SCC and melanoma (p = 0.003 for intensity of staining, p = 0.001 for percentage of positive cells, and p = 0.017 for H-score). In BCC samples, there was a significant correlation between CD44 expression and ulceration, indicating that lower expression levels of CD44 were mainly found in BCC cases with ulceration (p = 0.05 for percentage of positive cells). A trend was evident between CD44 expression and margin involvement (p = 0.06), and also between CD44 expression and TILs (p = 0.08), while no significant correlation was seen between CD44 expression and other clinicopathological parameters including local recurrence (Tab. I). In SCC specimens, we found a significant positive correlation between overexpression of CD44 and ulceration (p = 0.01 for intensity of staining). No significant association was observed between CD44 expression and histological grade and metastasis (Tab. II). In melanoma cases we found that larger tumors expressed higher levels of CD44 (p = 0.038). Furthermore, a higher level of CD44 expression in melanoma samples was correlated with a higher rate of recurrence (p = 0.029). There was no significant association between the levels of CD44 expression and Clark levels, metastasis, and lymph node involvement (p = 0.67) (Tab. III).

Fig. 1

Expression of CD44 in different skin cancer subtypes. In each type of tumor, intensity of staining was graded as (A) 0 (absent), (B) 1 (weak), (C) 2 (moderate) or (D) 3 (strong). The strongest staining intensity was observed in melanoma patients.

Analysis of ALDH1A1 Expression and its Correlation with Clinicopathological Parameters

We also performed immunohistochemical analysis to investigate the expression of another putative CSC marker, ALDH1A1, in the same series of skin tumors. The pattern of ALDH1A1 staining was cytoplasmic. The mean H-score was the same (=10) in both BCC and SCC samples, while the mean H-score in melanoma cases was 58. Lower reactivity of ALDH1A1 was seen in 44 (76%) BCC cases, whereas higher ALDH1A1 reactivity was found in 14 (24%) cases (Fig. 2). Thirty-three (89%) of 37 SCC patients displayed lower expression of ALDH1A1 and higher expression was found in only 4 (11%) patients (Fig. 2). Of 12 melanoma specimens, 8 (67%) had lower ALDH1A1 expression levels and 4 (33%) had higher levels (Fig. 2). ALDH1A1 expression was significantly different between the studied skin cancer subtypes (p = 0.002 for the percentage of positive cells). Using the Mann-Whitney U test, we found a significant difference between the expression of ALDH1A1 in melanoma compared with SCC (p = 0.048 for staining intensity, p = 0.003 for percentage of positive cells, and p = 0.018 for H-score) and BCC (p = 0.002 for percentage of positive cells). Box plot diagram of CD44 and ALDH1A1 expression have been shown in (Fig. 3). In BCC cases, higher levels of ALDH1A1 expression were significantly correlated with ulceration (p = 0.05 for the percentage of positive cells) and TILs (p = 0.05), while there was no significant association of ALDH1A1 expression with histological type and local recurrence (Tab. I). In SCC cases, we found no significant correlation between the level of ALDH1A1 expression and clinicopathological factors including metastasis, local recurrence, and margin involvement (Tab. II). Although ALDH1A1 expression showed a trend with Breslow thickness (p = 0.08 for H-score), there was no significant association with other clinicopathological factors including depth of invasion, local recurrence, metastasis, Clark level and lymph node involvement (Tab. III).

Fig. 2

Expression of ALDH1A1 in different skin cancer subtypes. In each type of tumor, staining intensity was graded as (A) 0 (absent), (B) 1 (weak), (C) 2 (moderate) or (D) 3 (strong).

Fig. 3

Box-plot diagram of CD44 and ALDH1A1 expression in skin tumor subtypes. In each panel the vertical axis shows the total immunostaining score (H-score) obtained as described in the Methods section. Panels A and B depict the expression of CD44 and ALDH1A1 in basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and melanoma.

Combined Analysis of CD44 and ALDH1A1 Expression

Evaluation of the relation between the expression of CD44 and ALDH1A1 showed a significant reciprocal correlation in our series of skin tumors (p<0.001). In the current study, a CD44^high/ALDH1A1^high phenotype was designated as group 1, a CD44^low/ALDH1A1^high or CD44^high/ALDH1A1^lowphenotype was designated as group 2, and a CD44^low/ALDH1A1^low phenotype was designated as group 3. The CD44^low/ALDH1A1^lowphenotype was observed in 58 (54%) cases, whereas 10 (9%) samples showed the CD44^high/ALDH1A1^high phenotype. Expression of any marker (CD44^low/ALDH1A1^high or CD44^high/ALDH1A1^low) was found in 39 (37%) skin cancer specimens. The majority of BCC samples (67%) displayed the CD44^low/ALDH1A1^lowphenotype, while coexpression of the markers (CD44^high/ALDH1A1^high) was seen in 5 (9%) BCC cases (Fig. 4). Although expression of any marker (CD44^low/ALDH1A1^high or CD44^high/ALDH1A1^lowphenotype) and the CD44^low/ALDH1A1^low phenotype were observed in 20 (54%) and 17 (46%) of SCC cases, respectively, the CD44^high/ALDH1A1^highphenotype was not found in SCC samples (0%) (Fig. 4). In melanoma, the phenotype groups 1 and 2 had similar frequencies (42%) (Fig. 4).

Fig. 4

Distribution of CD44/ALDH1A1 phenotypes in skin cancer subtypes including basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and melanoma. Group 1: CD44^high/ALDH1A1^high phenotype; group 2: CD44^low/ALDH1A1^high or CD44^high/ALDH1A1^low phenotype; group 3: CD44^low/ALDH1A1^low phenotype.

The correlation of CD44/ALDH1A1 phenotypes with clinicopathological parameters was investigated by 1-way ANOVA and Tukey's post hoc analysis. We found that CD44/ALDH1A1 status significantly correlated with tumor type (p<0.001). Also among the different subtypes of skin tumors, the CD44/ALDH1A1 status was significantly different between melanoma versus BCC (p<0.001) and SCC (p = 0.003) (Fig. 3). No significant association was found between CD44/ALDH1A1 phenotypes and other clinicopathological parameters in BCC and melanoma.

Discussion

Among various markers for identification of CSCs, ALDH1A1 has been reported to be a melanocytic stem cell marker (8-9-10). ALDH1A1 is an enzyme that accelerates oxidation of aldehydes to their corresponding carboxylic acids (31). ALDH1A1 has been widely used for the identification and isolation of CSCs from many types of cancers including melanoma (9, 19, 20, 23, 32, 33). For the first time (34), this study showed that ALDH1A1 had higher expression levels in melanoma than in BCC and SCC. In line with our data, previous studies showed that high expression of ALDH1A1 correlated strongly with aggressive tumor behavior and poor prognosis in breast, colon and lung cancer (19, 20, 33), whereas for melanoma there is no consensus about the biological role(s) of ALDH1A1 in CSCs (8, 9, 23, 34). There is some evidence that ALDH1-positive melanoma cells fulfill the CSC criteria based on in vivo tumorigenicity as compared to the parental melanoma cells or ALDH1-negative melanoma cells (8-9-10), although another study demonstrated that ALDH1-positive melanoma cells could convert to ALDH1-negative cells and both of these subpopulations had comparable tumorigenic potential (23). In a pioneer study, targeting ALDH1A1 in melanoma cells enriched for CSCs using siRNA or shRNA resulted in a significant decrease in cell growth, induced sensitization to chemotherapeutic drugs, and reduced tumorigenicity (9). Supplementation of chemotherapeutic agents with ALDH1 inhibitors in a human melanoma xenograft model reduced tumor growth in vivo and led to significantly decreased numbers of residual tumorigenic cells (10). In agreement with a previous report, knockdown of ALDH1A1 in ALDH1-positive melanoma cells resulted in significant decreases in tumorigenic potential, tumor growth, and metastasis (10). Considering the mounting evidence for the role of ALDH1A1 in conferring aggressive tumor behavior on breast, colon, lung and ovarian cancers (19, 20, 32, 33, 35), more comprehensive studies are warranted to more precisely define the role of ALDH1A1 in melanoma progression.

The CD44 antigen is a cell-surface molecule involved in cell proliferation and survival, cell-cell interactions, cell adhesion and migration (11, 36). There is still insufficient data regarding the role of CD44 as a promising CSC marker in melanoma (11, 13, 26, 37-38-39). We found that the highest percentages of CD44-positive tumor cells were detected in melanoma samples, followed by SCC and lastly BCC samples. Our study confirmed previous data showing low expression of CD44 in BCC samples and intermediate expression levels in SCC specimens (40). In the current study, higher CD44 expression was more often found in melanoma samples with larger sizes. Additionally, a positive correlation was observed between the expression of CD44 and recurrence in melanoma cases. These findings may suggest that high CD44 expression confers aggressiveness and metastatic potential on melanoma cells. In several previous reports, the role of CD44 as a potential CSC marker was explored in melanoma samples using flow cytometry; the results showed that melanoma cells expressing CD44 and those that were CD44 negative had similar CSC-like properties in in vitro and in vivo assessments (16, 41). Despite these findings, a more recent report has shown that high expression of CD44v6 correlates with brain metastasis from melanoma (42). Interestingly, Todaro et al found that colorectal CSCs expressing CD44v6 could form colorectal tumors that were more aggressive and more frequently metastasized (43).

We also evaluated the coexpression of ALDH1A1 and CD44 in the examined skin cancer specimens. The CD44^high/ALDH1A1^high expression pattern was predominantly found in melanoma cases compared to the other types of skin cancer, which might account for the fact that melanoma is the most aggressive skin tumor.

CTLA-4 (CD152) is encoded by the CTLA4 gene and expressed on the surface of T cells, which are involved in the cellular immune attack on antigens (44). Ipilimumab (MDX-010, Yervoy; Bristol-Myers Squibb), a fully human monoclonal antibody against CTLA-4, was approved by the US Food and Drug Administration for the treatment of both early- and late-phase melanoma (45, 46). Additionally, interferon-alpha and other immune modulators such as interleukin 2 have been widely used in melanoma treatment (47). These novel modalities, beside other options such as surgery, often transiently eliminate the bulk of melanoma tumors, but a CSC subpopulation can survive, resulting in drug resistance, tumor progression, and relapse (48). Complete regression and efficient treatment of melanoma can be achieved by targeting melanoma CSCs (and not normal stem cells) in addition to destroying the bulk of the tumor (49, 50).

In the present study, overexpression of ALDH1A1 and CD44 was found in melanoma samples compared to non-melanoma skin cancer samples including BCC and SCC. Increased expression of ALDH1A1 and CD44 appears to confer aggressive behavior on melanoma, and these markers may therefore be considered potential candidates for targeting melanoma CSCs.

Conclusions

The distinct differential expression of the potential CSC markers CD44 and ALDH1A1 in melanoma compared with SCC and BCC showed that expression of CD44/ALDH1A1 is associated with aggressive tumor behavior and metastasis. The CD44^high/ALDH1A1^high phenotype in melanoma samples and CD44^high cells in SCC can be considered as reliable targets for diagnosis and therapy of skin cancers.

Footnotes

Financial support: This study was supported by a grant from Iran University of Medical Sciences (Grant #12417-12-04-89).

Conflict of interest: The authors declare that they have no conflict of interest.

References

Cakir

BÖ

Adamson

Cingi

Epidemiology and economic burden of nonmelanoma skin cancer. Facial Plast Surg Clin North Am 2012 20 4 419–422

Dubas

Ingraffea

Nonmelanoma skin cancer. Facial Plast Surg Clin North Am 2013 21 1 43–53

Rigel

Epidemiology of melanoma. Semin Cutan Med Surg 2010 29 4 204–209

Rigel

Russak

Friedman

The evolution of melanoma diagnosis: 25 years beyond the ABCDs. CA Cancer J Clin 2010 60 5 301–316

Lee

Barthel

Schatton

Melanoma stem cells and metastasis: mimicking hematopoietic cell trafficking?. Lab Invest 2014 94 1 13–30

Han

Shi

Gong

Zhang

Sun

Cancer stem cells: therapeutic implications and perspectives in cancer therapy. Acta Pharm Sin B 2013 3 2 65–75

Jordan

Guzman

Noble

Cancer stem cells. N Engl J Med 2006 355 12 1253–1261

Boonyaratanakornkit

Yue

Strachan

. Selection of tumorigenic melanoma cells using ALDH. J Invest Dermatol 2010 130 12 2799–2808

Luo

Dallaglio

Chen

, ALDH1A isozymes are markers of human melanoma stem cells and potential therapeutic targets. Stem Cells 2012 30 10 2100–2113

10.

Yue

Huang

Fong

. Targeting ALDH1 to decrease tumorigenicity, growth and metastasis of human melanoma. Melanoma Res 2015 25 2 138–148

11.

Jaggupilli

Elkord

Significance of CD44 and CD24 as cancer stem cell markers: an enduring ambiguity. Clin Dev Immunol 2012 2012708036

12.

Naor

Wallach-Dayan

Zahalka

Sionov

Involvement of CD44, a molecule with a thousand faces, in cancer dissemination. Semin Cancer Biol 2008 18 4 260–267

13.

Al-Hajj

Wicha

Benito-Hernandez

Morrison

Clarke

Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci USA 2003 100 7 3983–3988

14.

Collins

Berry

Hyde

Stower

Maitland

Prospective identification of tumorigenic prostate cancer stem cells. Cancer Res 2005 65 23 10946–10951

15.

Dalerba

Dylla

Park

. Phenotypic characterization of human colorectal cancer stem cells. Proc Natl Acad Sci USA 2007 104 24 10158–10163

16.

Quintana

Shackleton

Sabel

Fullen

Johnson

Morrison

Efficient tumour formation by single human melanoma cells. Nature 2008 456 7222 593–598

17.

Vlashi

Pajonk

Cancer stem cells, cancer cell plasticity and radiation therapy. Semin Cancer Biol 2015 3128–35

18.

Pattabiraman

Weinberg

Tackling the cancer stem cells-what challenges do they pose?

Nat Rev Drug Discov 2014 13 7 497–512

19.

Ginestier

Hur

Charafe-Jauffret

. ALDH1 is a marker of normal and malignant human mammary stem cells and a predictor of poor clinical outcome. Cell Stem Cell 2007 1 5 555–567

20.

Jiang

Qiu

Khanna

. Aldehyde dehydrogenase 1 is a tumor stem cell-associated marker in lung cancer. Mol Cancer Res 2009 7 3 330–338

21.

Chen

Hsu

. Aldehyde dehydrogenase 1 is a putative marker for cancer stem cells in head and neck squamous cancer. Biochem Biophys Res Commun 2009 385 3 307–13

22.

Jackson

Brocker

Thompson

. Update on the aldehyde dehydrogenase gene (ALDH) superfamily. Hum Genomics2011 5 4 283–303

23.

Prasmickaite

Engesaeter

BØ

Skrbo

. Aldehyde dehydrogenase (ALDH) activity does not select for cells with enhanced aggressive properties in malignant melanoma. PLoS One 2010 5 5 e10731

24.

Kononen

Bubendorf

Kallioniemi

. Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 1998 4 7 844–847

25.

Sabet

Rakhshan

Erfani

Madjd

Co-Expression of Putative Cancer Stem Cell Markers, CD133 and Nestin, in Skin Tumors. Asian Pac J Cancer Prev 2014 15 19 8161–8169

26.

Roudi

Madjd

Korourian

. Clinical significance of putative cancer stem cell marker CD44 in different histological subtypes of lung cancer. Cancer Biomark 2014 14 6 457–467

27.

Roudi

Korourian

Shariftabrizi

Madjd

Differential expression of cancer stem cell markers ALDH1 and CD133 in various lung cancer subtypes. Cancer Invest 2015 33 7 294–302

28.

Piras

Perra

Murtas

. The stem cell marker nestin predicts poor prognosis in human melanoma. Oncol Rep 2010 23 1 17–24

29.

Clemente

Mihm

Jr Bufalino

Zurrida

Collini

Cascinelli

Prognostic value of tumor infiltrating lymphocytes in the vertical growth phase of primary cutaneous melanoma. Cancer1996 77 7 1303–1310

30.

Balch

Gershenwald

Soong

Final version of 2009 AJCC melanoma staging and classification. J Clin Oncol 2009 27 36 6199–6206

31.

Yoshida

Rzhetsky

Hsu

Chang

Human aldehyde dehydrogenase gene family. Eur J Biochem 1998 251 3 549–557

32.

Mizuno

Suzuki

Makino

. Cancer stem-like cells in ovarian clear cell carcinoma are enriched in ALDH-high population associated with the accelerated scavenging system of reactive oxygen species. Gynecol Oncol 2015 137 2 299–305

33.

Huang

Hynes

Zhang

. Aldehyde dehydrogenase 1 is a marker for normal and malignant human colonic stem cells (SC) and tracks SC overpopulation during colon tumorigenesis. Cancer Res 2009 69 8 3382–3389

34.

La Porta

Zapperi

Human breast and melanoma cancer stem cells biomarkers. Cancer Lett 2013 338 1 69–73

35.

Raha

Wilson

Peng

. The cancer stem cell marker aldehyde dehydrogenase is required to maintain a drug-tolerant tumor cell subpopulation. Cancer Res 2014 74 13 3579–90

36.

Marhaba

Zöller

CD44 in cancer progression: adhesion, migration and growth regulation. J Mol Histol 2004 35 3 211–31

37.

Hurt

Kawasaki

Klarmann

Thomas

Farrar

CD44+ CD24− prostate cells are early cancer progenitor/stem cells that provide a model for patients with poor prognosis. Br J Cancer 2008 98 4 756–765

38.

Murphy

Wilson

Girouard

Frank

Stem cells and targeted approaches to melanoma cure. Mol Aspects Med 2014 3933–49

39.

Wang

Gao

Suo

. Identification and characterization of cells with cancer stem cell properties in human primary lung cancer cell lines. PLoS One 2013 8 3 e57020

40.

Seelentag

Günthert

Saremaslani

. CD44 standard and variant isoform expression in human epidermal skin tumors is not correlated with tumor aggressiveness but down‐regulated during proliferation and tumor de‐differentiation. Int J Cancer 1996 69 3 218–224

41.

Quintana

Shackleton

Foster

. Phenotypic heterogeneity among tumorigenic melanoma cells from patients that is reversible and not hierarchically organized. Cancer Cell 2010 18 5 510–523

42.

Marzese

Liu

Huynh

. Brain metastasis is predetermined in early stages of cutaneous melanoma by CD44v6 expression through epigenetic regulation of the spliceosome. Pigment Cell Melanoma Res 2015 28 1 82–93

43.

Todaro

Gaggianesi

Catalano

. CD44v6 is a marker of constitutive and reprogrammed cancer stem cells driving colon cancer metastasis. Cell Stem Cell 2014 14 3 342–56

44.

Carreno

Bennett

Chau

. CTLA-4 (CD152) can inhibit T cell activation by two different mechanisms depending on its level of cell surface expression. J Immunol 2000 165 3 1352–1356

45.

Maker

Yang

Sherry

. Intrapatient dose escalation of anti–CTLA-4 antibody in patients with metastatic melanoma. J Immunother 2006 29 4 455–463

46.

Lipson

Drake

Ipilimumab: an anti-CTLA-4 antibody for metastatic melanoma. Clin Cancer Res2011 17 22 6958–6962

47.

Atkins

Kunkel

Sznol

Rosenberg

High-dose recombinant interleukin-2 therapy in patients with metastatic melanoma: long-term survival update. Cancer J Sci Am 2000 6S11–4

48.

Frank

Schatton

Frank

The therapeutic promise of the cancer stem cell concept. J Clin Invest 2010 120 1 41–50

49.

Wong

Liu

Ridky

Cassarino

Segal

Chang

Module map of stem cell genes guides creation of epithelial cancer stem cells. Cell Stem Cell 2008 2 4 333–344

50.

Nakanishi

Seno

Fukuoka

. Dclk1 distinguishes between tumor and normal stem cells in the intestine. Nat Genet 2013 45 1 98–103