Background
Malignant pleural mesothelioma is an aggressive tumour with a very poor prognosis [
1]. The combination of pemetrexed and cisplatin is considered the front-line regimen for this disease, yielding a response rate of 41% and a median survival of 12.1 months [
2]. Despite continuous efforts to implement new therapeutic modalities, none of these, have prolonged patient survival primarily due to chemoresistance [
1]. It has been hypothesized that tumour relapse may be associated with the drug resistance of cancer stem cells; a rare cell population with the exclusive ability to self- renew and maintain a tumour [
3]. Hence, the identification and complete elimination of these cells presents an ultimate goal in MPM therapy. Current studies have identified ALDH and CD44 as putative CSC markers which exhibit high chemoresistive properties in solid cancers; thus, rendering them as potential indicators of drug tolerance in MPM.
Aldehyde dehydrogenase enzymes are a family of intracellular enzymes that are involved in cellular proliferation, differentiation, detoxification, and drug resistance through the oxidation of cellular aldehydes [
4,
5]. Certain ALDH isozymes are upregulated in tumour cells resulting in decreased cellular sensitivity to cyclophosphamides and some oxazaphosphorines leading to decreased chemotherapeutic effect. In leukemia and lung cancer cell lines overexpression of ALDH1A2a and ALDH2 increased cell proliferation and resistance to 4-hydroperoxicyclophosphamide and doxorubicin indicating a role for the ALDH isozymes in drug resistance [
6]. The modulation of ALDH activity has been a central subject of research to improve the efficacy of conventional chemotherapeutic drugs. Studies have shown that in small cell lung cancer, lung adenocarcinoma, leukemia and lung cancer cell lines the ALDH-dependent chemoresistance can be inhibited by DEAB or siRNA that conferred sensitivity to drug treatments [
7‐
9].
CD44, a transmembrane receptor for hyaluronan is associated with aggressive tumour growth, metastasis and resistance to therapy [
10,
11]. In combination with other surface markers (e.g., ALDH and CD24), CD44 can discriminate between various cancer subsets. In solid cancers the ALDH
high/CD44
+ subpopulation has been shown to possess stem cell-like properties that convey radio- and chemoresistance [
12‐
14]. ALDH
+/CD44
+ subpopulation can be sensitized by selective inhibition of ALDH activity using DEAB or all-
trans retinoic acid, ATRA in breast cancer [
14]. As a single marker, CD44 is currently considered as a putative CSC indicator in human carcinomas including cancer of the lung. In NSCLC cell lines, sorted CD44
+ cells that bear stem cell-like properties conferred more resistance to cisplatin exhibiting lower apoptotic levels compared with CD44
- cells [
15].
Despite the current evidence linking ALDH and CD44 to drug resistance in solid tumours, the variability in the different studies still warrants further investigation to delineate the present roles of these potential CSC markers. Here, we sought to investigate whether ALDH can select for a drug-resistant subpopulation in three MPM cell lines. We also assessed whether the ALDHhigh cells were associated with CD44, thus broadening the spectrum for identification of a drug-tolerant subpopulation in MPM. The specific selection of a chemoresistant subpopulation using ALDH and CD44 may serve as a potential therapeutic target that may be employed as adjuvant therapy to the current standard treatment modalities in MPM.
Methods
Cell culture
The H28 and H2052 mesothelioma cell lines (LCD Promochem, France) were maintained in RPMI 1640 (PAA, Austria) containing 10% fetal bovine serum, FBS (PAA, Austria) and 1% penicillin/streptomycin solution (Invitrogen, Switzerland). ACC-Meso-4 cell line was purchased from Riken Cell Bank, Resource No: RBRC-RCB2293 (Ibaraki, Japan) and cultured using the above-mentioned culture medium. Cells were cultured at 37°C, 95% humidity and 5% C02. The general information issued by the providers of the three MPM cell lines does not have data on drug resistance to cisplatin.
Single-cell preparations of parental and ALDH-sorted MPM cell lines were resuspended in an appropriate amount of sphere-forming medium (RPMI1640 supplemented with 20 ng/ml EGF and bFGF, [Invitrogen, Switzerland]; 4 μg/ml insulin, [Sigma, Germany]; 1 ml B27, [Invitrogen, Switzerland] and 1% penicillin/streptomycin solution). For all cell lines, 5 x 10
3 cells/ml/well were seeded onto a 24-well ultra-low adherent plate (Costar, USA). Cells were incubated at 37°C, 95% humidity and 5% C0
2 for 7–14 days. The documentation of images and evaluation of sphere-forming efficiency were performed on day 7. Sphere-forming efficiency (%) was determined by dividing the number of spheres formed by the original number of seeded cells. The quotient was then multiplied by 100 [
16]. Images were taken with Leica DMI 4000B at 5x magnification.
Drug treatment
Drug resistance to cisplatin of mesothelioma cells were assessed by exposure to the IC50 values obtained for the non-sorted and ALDH-sorted cells for each of the three MPM cell lines. For the determination of IC50, a dilution series of 2-fold increments of cisplatin (0–256 μM Cisplatin, CDDP, Bristol Myers Squibb, Switzerland) were prepared in RPMI 1640 supplemented with 10% FBS and 1% penicillin/streptomycin. Cells at a density of 5 x 103cells/100 μl/well in 96-well plates were incubated in media with or without the addition of cisplatin. Following a 48- and 72-hr incubation periods, culture media was aspirated, then replenished with XTT cell proliferation assay (Roche Chemicals, Switzerland) reagents. After a 30-min incubation at 37°C, formazan production was measured spectrophotometrically at 450 nm. Three independent experiments in triplicate were performed.
For cisplatin treatment, cells were cultured at 5 x 10
4 cells/well in a 6-well plate (in three replicates) 48 hours prior to the addition of the previously determined IC
50 of cisplatin for each cell line in RPMI 1640 containing 10% FBS and 1% penicillin/streptomycin solution. Following the 48- and 72-h hour treatments at 37°C, cells were washed with PBS and harvested to perform the following: mRNA isolation, sphere formation assay and cell viability. Pre-treatment of cells with 100 μM of ALDH inhibitor, DEAB (Sigma, Germany) was done for 48 h prior to cisplatin treatment [
6,
14].
Aldefluor assay and flow cytometry
The Aldefluor kit (Stem Cell Technologies, Canada) was used to identify the cells expressing ALDH activity. Cells (0.5 – 1.0 x 106) were incubated in assay buffer containing ALDH substrate BODIPY-aminoacetaldehyde (BAAA). As negative control, an aliquot of the ALDH substrated-treated cells were immediately quenched with specific ALDH inhibitor, DEAB. Both tubes were incubated for 45 min at 37°C. After incubation, cells were centrifuged and the pellets were resuspended with 500 μl of assay buffer prior to data acquisition using the green fluorescence channel of the LSR II flow cytometer (Becton Dickinson). This flow cytometry-based assay is termed as FACS analysis in this manuscript. DEAB-treated cells served as control to set the ALDHhigh regions. The same staining procedure was applied before sorting the cells with FACS Aria using FACS Diva software (Becton Dickinson). Cell purity was determined at each independent sorting. After sorting of the ALDHhigh and ALDHlow fractions, these were immediately used in all experiments. ALDHhighCD44+ phenotype was assessed by immediately re-staining the freshly-sorted ALDHhigh and ALDHlow cells with mouse anti-human CD44 APC-H7 (clone G44-26, BD Pharmingen, USA) and appropriate isotype control (mouse IgG2bK) for 30 min on ice in the dark. After washing with FACS buffer, data acquisition followed immediately. Propidium iodide (PI) staining was used to exclude the non-viable cells. FACS data were analysed with Flowjo software 7.2.5 (Treestar, Oregon, USA).
RNA extraction and real-time quantitative PCR
Cell cultures were collected in an appropriate amount of RNAprotect™ cell reagent (Qiagen, Germany) followed by total RNA extraction using RNeasy Kit ( Qiagen) according to manufacturer’s instructions. Complementary DNA (cDNA) was synthesized by using the High capacity cDNA reverse transcription kit (Applied Biosystems, Rotkreuz, Switzerland) following the manufacturer’s instructions. The mRNA levels of the housekeeping gene β2-microglobulin, β2M, and the target genes ALDH1A1, ALDH1A2, ALDH1A3, ALDH2 and CD44 were quantified with the commercially available TaqMan “Assay on Demand” primer/probes (β2M – Hs 99999903_m1; ALDH1A1 –Hs 00167445_m1; ALDH1A2- Hs 0018025_m1; ALDH1A3 –Hs 00167476_m1; ALDH2 – Hs 010007998_ml; CD44 –Hs01075861_m1) (Applied Biosystems; Rotkreuz, Switzerland). Twenty nanograms of resulting cDNAs were subjected to quantitative RT-PCR, in a 10 μl final reaction volume and analyzed in triplicates. Gene expression was detected using ABI 7900 sequence detection system. The gene expression level of each target gene was normalized by the endogeneous gene, β2M and compared among cells by the ΔΔCT method. Baseline and threshold for Ct calculation were set automatically with ABI Prism SDS 2.1 software.
Statistical analysis
Statistical analyses were performed using Graph Pad Prism 5.0 software© (San Diego, Ca). Two-tailed Student’s t-test was used to compare 2 groups. One- or two-way ANOVA with Bonferroni post hoc test as appropriate was performed to compare the values of >2 groups. The statistical significance was set at p <0.05.
Discussion
The identification of a chemoresistant population offers the possibility of cell-specific therapeutic approaches that may augment the current treatment modalities in malignant pleural mesothelioma. Here we show that ALDHhighCD44+ cells are involved in conferring resistance to cisplatin in the three MPM cell lines; H28, H2052 and Meso4. We also demonstrate that ALDH as a single marker is not sufficient to define chemoresistant, sphere-forming cell populations in the tested MPM cell lines.
Several reports have supported the utility of ALDH to identify CSC-like populations in different malignancies and its potential as a therapeutic target [
5,
8,
20,
25,
26]. We found that all MPM cell lines are capable of producing spheres for three consecutive generations thus sustaining the property of self-renewal, a stem cell feature considered as a key discriminating difference between CSCs and non-CSCs [
27]. The presence of ALDH activity in H28, H2052 and Meso4 further supports the existence of putative CSC populations in these cell lines. Unexpectedly, the ALDH
high- and ALDH
low-sorted cells which were supposed to demarcate a CSC- from non-CSC-like cells showed sphere-forming ability and generate ALDH activity although the ALDH
low cells were less efficient. They also both showed an association with CD44. Wang et al. [
12] and Prasmickaite et al. [
22] also found that ALDH
br/+and ALDH
low/- cells repopulated stem cell heterogeneity, formed spheroids and generated tumours in epithelial and malignant melanoma. In the lung adenocarcinoma cell line, SPC-A1 both ALDH
lo and ALDH
high-sorted cells formed colonies and developed tumours although ALDH
lo was less efficient [
28]. In the MPM cell line, MSTO211H and Ewing’s sarcoma cell lines only purified ALDH
bright cells generated both ALDH
bright/ALDH
low cells, but ALDH
low did not repopulate ALDH
bright cells [
21,
29]. These studies suggest that the type of tumour and heterogeneity within a cancer [
30] are at least in part, responsible for the differential behavior of ALDH in the malignant setting. Moreover, ALDH activity may be dependent on whether the tumour conforms to the cancer stem cell model. The CSC model proposes the presence of a cellular hierarchy in the tumour, and that only a subset of tumour cells possess the ability of self-renewal and to generate the different phenotypes that comprise the neoplasm [
31]. A conversion of ALDH
low cells into ALDH
high cells in the presence of an appropriate environment during the short-term, in vitro culture is a possible explanation. This assumption follows the stemness phenotype model (SPM) proposing that all cancer cells possess stem cell properties, and that stemness is modulated by the environment such that CSCs and non-CSCs can interconvert into each other when changes in the environment favours this conversion [
32]. In support of this hypothesis, it has been shown that CD44
- Du145 prostate cells produced CD44
+ cells in vitro [
33], and in the MCF-7 breast cancer cell line, the sorted non-SP cells gave rise to SP cells [
34]. The question of whether MPM follows the CSC or SMP model warrants further investigation.
Our findings showed that ALDH activity is resistant to cisplatin treatment in the ALDH-sorted fractions of the three MPM cell lines. This consolidates with the biological function of ALDH in its ability to detoxify anticancer drugs such as oxazaphosphorines, cyclophosphamides and taxanes thus conferring drug resistance [
5,
7,
14]. Consistent with our findings, chemoresistance as an attribute of ALDH
+ cells has been documented in different solid tumours including lung cancer and primary MPM specimens [
8,
21,
26,
29,
35]. The presence of ALDH
highCD44
+ subpopulations in the ALDH-sorted fractions highly indicates that in addition to ALDH, CD44 may also contribute to cisplatin resistance. The current consensus posits that CD44
+ subfractions in many human cancers are highly malignant and drug resistant. In non-small cell lung cancer, CD44
+-sorted cells with stem cell-like properties were found to be more resistant to cisplatin than CD44
- cells [
15]. Wang et al. [
12] proposed that the combination of ALDH1 and CD44 stringently defined ovarian cancer stem cells, which showed chemoresistance and poor clinical clinical outcome. This is strengthened by the recent evidence that ALDH
hiCD44
hi tumour-initiating cells maintained lung tumorigenicity and drug resistance in patient-derived lung cancer cells [
36].
Our results demonstrated that sequential DEAB and cisplatin treatments remarkably diminished cell viability in the ALDH
high-sorted cells, a short-term elimination of spheres, but not a complete inhibition of sphere formation in all of the ALDH
high and ALDH
low fractions of MPM cell lines. The profound effect of DEAB in the cell viability of ALDH
high fractions is supported by the findings which showed that the downregulation of ALDH isozymes in A549 lung cancer cell line altered cell proliferation and motility, whereas an analogous experiment in cell lines devoid of ALDH-expressing cells had an insignificantly less inhibitory effect on cell proliferation demonstrating a functional role of ALDH in the regulation of cell growth [
9]. Croker and Allan [
14] also observed a significant reduction in cell viability but not a complete inhibition of a long-term re-growth of colony-forming ability of chemo- or radiation-treated ALDH
hiCD44
+ cells pre-treated with DEAB in breast cancer. A plausible explanation for the re-emergence of spheres may be attributed to the survival of putative CSCs which escaped the cytotoxic effect of cisplatin, and were tolerant to DEAB treatment. We speculate that the tested MPM cell lines have a heterogeneous cancer cell population. The more differentiated cells within the hierarchy might be efficiently killed by chemotherapy, whereas the less differentiated cells bearing CSC phenotypes survive and give rise to new transit-amplifying cells with the capacity to regenerate the culture [
37]. The possibility of cells existing in a dormant quiescent state with the capacity to regrow when the environmental cues are appropriate could also be an attribute [
38]. If the observed cisplatin resistance is unique to ALDH
high cells only, then a DEAB-mediated sensitization process should have prevented the re-growth of spheres. The failure of DEAB to sensitize both the ALDH-sorted fractions to cisplatin strongly supports our assumption that the ALDH
highCD44
+ cells are crucial players in conveying drug tolerance. Hence, specific targeting of both phenotypes may offer a more effective chemotherapy.
Studies have shown that ALDH activity may reflect other ALDH isozymes in addition to the prevalent ALDH1A1 which are important in the regulation of several biological activities including drug resistance [
39,
40]. Hence, the identification of specific isozymes contributing to ALDH activity is a critical factor. We observed an upregulation of ALDH1A2 in H28 and ALDH1A3 in H2052 and Meso4 in the cisplatin + DEAB-resistant cells of the ALDH-sorted fractions of the three MPM cell lines, indicating the implication of these two isozymes in conveying resistance to cisplatin and DEAB. Other groups [
6,
8,
28] have detected ALDH1A1, ALDH1A2, ALDH3A1, and ALDH2A1 in lung cancers which were shown to have an association with chemoresistance. CD44 was likewise increased after an analogous treatment with cisplatin + DEAB highly suggestive of an essential role in conferring drug tolerance. Our data indicate that at the transcript level, ALDH and CD44 are important players in the observed resistance to cisplatin and DEAB. Notably, inhibition of ALDH activity cannot sensitize the ALDH
hghCD44
+ cells to drug treatment.
Competing interests
The authors declare that they have no conflict of interest.
Authors’ contribution
LCD conceived and designed the study, analyzed the data and wrote the manuscript. LF and RB performed cellular and RT-qPCR assays, flow cytometry-based analyses, collected and graphed experimental data. RAS gave suggestions to improve the impact of the study and approved the final version of the manuscript for publication. GK made significant contributions to the design of the study, analysis of data and critically improved the manuscript for intellectual content. All authors read and approved the final manuscript.