Combined gemcitabine and CHK1 inhibitor treatment induces apoptosis resistance in cancer stem cell-like cells enriched with tumor spheroids from a non-small cell lung cancer cell line

Abstract

Evaluating the effects of novel drugs on appropriate tumor models has become crucial for developing more effective therapies that target highly tumorigenic and drug-resistant cancer stem cell (CSC) populations. In this study, we demonstrate that a subset of cancer cells with CSC properties may be enriched into tumor spheroids under stem cell conditions from a non-small cell lung cancer cell line. Treating these CSC-like cells with gemcitabine alone and a combination of gemcitabine and the novel CHK1 inhibitor PF-00477736 revealed that PF-00477736 enhances the anti-proliferative effect of gemcitabine against both the parental and the CSC-like cell populations. However, the CSC-like cells exhibited resistance to gemcitabine-induced apoptosis. Collectively, the spheroid-forming CSC-like cells may serve as a model system for understanding the mechanism underlying the drug resistance of CSCs and for guiding the development of better therapies that can inhibit tumor growth and eradicate CSCs.

Keywords

drug resistance / cancer stem cell / checkpoint kinase 1 (CHK1) / PF-00477736 / lung cancer / tumorigenicity

Introduction

Preclinical studies on novel anticancer agents rely on human tumor-derived cell lines. Thus, conventional cancer cell lines established decades ago under serum-containing conditions are frequently used for both in vitro tests and in vivo efficacy studies [1]. These tumor models have positively affected cancer drug discovery; however, many of these models are incapable of predicting clinical outcomes [2,3]. This discrepancy between preclinical efficacy and clinical outcome may be caused by the inability of these preclinical models which simulate the cancer stem cell (CSC) compartment of tumors in clinics. More importantly, the response to standard of care agents is frequently limited to the differentiated cell population in the heterogeneous tumor cell populations, which inhibits in vitro proliferation and causes in vivo tumor shrinkage, but not the elimination of drug-resistant CSCs. For instance, the cytotoxic agent gemcitabine targets rapidly proliferating differentiated cellular compartments in pancreatic cancer, which enriches the CSC fraction after treatment [4].

Recent studies have revealed that human tumors are organized as a hierarchy that consists of heterogeneous cell populations with distinct phenotypes, within which the majority of cells may be non-tumorigenic, differentiated populations [5,6]. A small subset of the tumor cell population called tumor-initiating cells or CSCs, which have been identified in a variety of human cancers, have been shown to drive tumor initiation, progression, and metastasis [6]. Furthermore, increasing evidence demonstrates that CSCs resist standard-of-care anticancer agents [7-9].The CSC concept has opened a new avenue for drug discovery. To test whether novel anticancer agents eradicate resistant CSC populations, appropriate tumor models that replicate the CSC component of human cancers must be applied to preclinical studies. Recent identification of anti-CSC agents in a CSC-relevant model system provides evidence of the identification of anti-CSC compounds [10].

Isolating and identifying CSC populations relies primarily on the separation of tumorigenic populations from freshly isolated clinical specimens. This process involves specific cell surface markers and the sequential confirmation of their unique or enhanced tumorigenic potentials in vivo. Alternatively, CSCs isolated from primary tumors can be propagated as non-adherent 3D spheres and/or spheroids under serum-free conditions. Previous studies have demonstrated that CD133-expressing brain tumor stem cells derived from clinical samples form tumor spheroids under serum-free conditions for neural stem cells [11]. Additional studies have shown that similar serum-free conditions can be used to selectively enrich and expand CSCs in vitro as tumor spheroids from tumor tissue from melanomas [12] and ependymomas [13], as well as breast [14], colon [15,16], lung [17], and ovarian cancers [18]. In addition, the phenotype and genotype of primary human tumors are faithfully preserved in tumor spheroids cultured under stem cell culture conditions, which emphasize the advantages of such novel tumor models for drug discovery [19,20].

CSC-like populations are present in cancer cell lines based on the isolation of side populations [21,22], cell surface marker-specific cells [23,24], and aldehyde dehydrogenase-positive subsets [25]. Tumor spheroid cultures in serum-free conditions also enrich CSC-like populations from melanomas [12], pancreatic cancer [26], and gliomas [27] lines. The current lung cancer cell culture models with CSC-properties available for drug discovery are very limited. Isolating and enriching CSC-like cells from lung cancer cell lines accelerates our understanding of CSC biology and enables the evaluation of novel therapeutic agents for killing lung CSCs.

We report the isolation of CSC-like cells from a human non-small cell lung carcinoma cell line as a model for evaluating novel anticancer agents. The CSC-like cells formed non-adherent spheroids under serum-free culture conditions and exhibited essential CSC characteristics, including increased tumorigenic potential and resistance to conventional cytotoxic anticancer drugs. CSC-like cells were treated with gemcitabine alone or in combination with the novel CHK1 inhibitor PF-00477736 [28]. The CHK1 inhibitor enhanced the antiproliferative effect of gemcitabine even though these cells are less sensitive to gemcitabine. However, the CSC-like cells were highly resistant to the apoptosis induced by gemcitabine or the combination treatment. Taken together, our results show that the CHK1 inhibitor and cytotoxic agents may synergistically inhibit the proliferation of lung CSC-like cells. However, the CSC-like population appears to be resistant to apoptosis caused by the treatments. Such resistant CSC-like models provide a valuable experimental model for differentiating anti-cancer agents during preclinical discovery and for developing more effective CHK1 inhibitors.

Materials and methods

Cell line, cell culture, and generation of spheroid CSC cultures

The non-small cell lung cancer cell line NCI-H1299 (ATCC; Manassas, VA) was cultured in RPMI-1640 medium containing 10% fetal bovine serum at 37 °C under a humidified 5% CO₂ atmosphere as adherent monolayers in regular cell culture flasks. To produce the serum-free CSC medium, human embryonic stem cell (hESC) medium consisting of knockout DMEM/F-12 (Invitrogen, Carlsbad, CA), 20% knockout serum replacer (Invitrogen), 200 mM l-glutamine (Invitrogen), and 10 mM MEM non-essential amino acids (Invitrogen) were conditioned with mouse embryonic fibroblasts (ATCC) for 2 days. The conditioned hESC medium was then mixed with the fresh hESC medium (1:1 ratio) and supplemented with 10 ng/ml bFGF (Invitrogen), 10 ng/ml EGF (Invitrogen), 10 μg/ml bovine insulin (Sigma-Aldrich, St. Louis, MO), 5.5 μg/ml human transferrin (Sigma-Aldrich), and 5 ng/ml sodium selenite (Sigma-Aldrich) [12,16,29]. Ultra-low attachment flasks (Corning, Corning, NY) were used to generate the CSC-like cell cultures. The CSC-like cells were induced to differentiate in the serum-containing medium used for the parental cells in regular flasks. The differentiated cells were studied after at least 4 weeks to ensure complete differentiation.

Flow cytometry analysis and BrdU incorporation assay

Standard cell surface flow cytometry was used to characterize the samples with PE-conjugated CD133 (Miltenyi, Bergisch Gladbach, Germany), APC-conjugated epithelium specific antigen EpCAM (CD326 or epithelial cell surface antigen, BD Biosciences, San Diego, CA), FITC-conjugated CD44 (BD), and the corresponding isotype-matched control antibodies (BD). BrdU incorporation and cytokeratin expression were measured simultaneously using an FITC BrdU flow kit (BD) and anti-cytokeratin antibodies (PE-conjugated cytokeratin 7/8, BD), following the manufacturers’ instructions.

RNA isolation, microarray, and gene expression data analysis

Total RNA was isolated using an RNeasy kit (Qiagen, Germantown, MD). For microarray sample processing, 2 µg of total RNA was used to synthesize cRNA using an Affymetrix IVT one-cycle labeling kit per manufacturer’s protocol. Fifteen micrograms of labeled, fragmented cRNA was hybridized to Affymetrix HGU133 plus 2.0 arrays (Affymetrix, Santa Clara, CA). Microarray data were RMA normalized using a Bioconductor Affy package (www.bioconductor.org). Principal Component Analysis (PCA) was conducted using the Matlab statistics and bioinformatics toolboxes. Differentially expressed genes were identified using Significance Analysis of Miroarray (PMID: 11309499) with a false discovery cut-off of 0.05. Hierarchical clustering was performed on the Z-transformed data using correlation similarity metrics and the centroid linkage method.

Real-time reverse transcriptase-polymerase chain reaction (RT-PCR)

Gene expression analysis of total RNA was performed via real-time RT-PCR using an ABI Prism 7900HT sequence detection system (Applied Biosystems, Foster City, CA). The primer/probe sets were ordered as inventoried in the TaqMan gene expression assays for the following genes: SFN1, YES1, JAK3, CALD1, SPP1, CD59, and GAPDH (Applied Biosystems). The reaction mixture was composed of one-step RT-PCR master mix reagents (Applied Biosystems), predesigned TaqMan gene expression assays, RNase-free water, and total RNA (450 ng) with a reaction volume of 20 µl. The RT-PCR cycling conditions were as follows: 48 °C for 30 min (reverse transcription), 95 °C for 10 min (AmpliTaq Gold activation), followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min (PCR). GAPDH expression in all samples was assessed for normalization. The reactions were run in triplicate.

Cell viability assays

The NCI-H1299 parental cells were harvested using Versene (Invitrogen) and seeded at 2000 cells per well in 100 µl of RPMI-1640 medium containing 10% fetal bovine serum in standard opaque-walled 96-well plate. The cells were cultured overnight. The medium was then replaced with 90 µl of fresh medium. On the same day, CSC-like cells were disaggregated using Versene and seeded at 2000 cells per well in 90 µl of CSC medium in an ultra-low attachment 96-well plate (Corning). Then, 10 µl of PBS containing appropriate amounts of etoposide, 5-fluorouracil (5-FU), doxorubicin, and cisplatin (Sigma-Aldrich) was added to the cells to reach the final concentrations indicated in the figures. After 72 h, the cultures were assessed using a CellTiter-Glo^® luminescent cell viability kit (Promega, Madison, WI). The cell lysates from the CSC-like cultures were transferred into opaque-walled plates prior to luminescence measurement. The results were normalized using the controls and shown as mean±SD from triplicate treatments.

The use of gemcitabine and the CHK1 inhibitor PF-00477736 (a drug candidate in clinical trials) in cell-based assays was previously described [28]. Briefly, NCI-H1299 parental cells were plated in 96-well plates at 4000βcells/well with 100 µl of growth medium. The CSC-like cells were plated in serum-free stem cell medium in ultra-low attachment 96-well plates. The cells were treated with gemcitabine at 50 nM and PF-00477736 at 540 nM and incubated at 37 °C in 5% CO₂ for 4 days. Resazurin working solution (1 mg/ml; Sigma-Aldrich) in PBS was prepared and added (1:10 by volume) to the cells. The cells were then incubated at 37 °C under a 5% CO₂ atmosphere for 3 h to 4 h. The cell lysates from the spheroid CSC-like cells were transferred into opaque-walled plates prior to luminescence measurement. The plate was read for fluorescence at 530 nm/590 nm on a SpectraMax Gemini EM fluorescence microplate reader (Molecular Devices, Sunnyvale, CA).

Evaluation of tumorigenicity using an in vivo limiting dilution assay

Tumorigenicity in athymic female mice was assessed using a limiting dilution assay (The Jackson Laboratory, Sacramento, CA). Briefly, the viable CSC-like cells and parental cells were mixed with Matrigel (BD Biosciences) at a 1:1 volume ratio, and injected subcutaneously at 2 × 10⁵, 2 × 10⁴, and 2 × 10³ cells per animal (5 animals/group). Tumor volume was calculated as the product of length × width × height × 0.5 and presented as mean±SEM.

Western blot analysis

The parental cells were plated at 4 × 10⁵ cells in 6 cm cell culture plates in 3 ml of growth medium containing 10% FBS. The CSC-like cells were plated at 6 × 10⁵ cells in 3 ml of stem cell medium in a 6 cm ultra-low attachment dish. The cells were then treated with gemcitabine at 25 nM for parental cells and at 50 nM for the CSC-like cells, and incubated for 24 h. PF-00477736 was then added at 540 nM for an additional 24 h. The cells were harvested via trypsinization and lysed. Protein concentration was determined using a BCA assay (Thermo Scientific) and the lysates were resolved using 4% to 12% Nu-PAGE gels (Invitrogen). The proteins were transferred onto nitrocellulose membranes (Invitrogen) and the proteins were identified using antibodies specific for total CDK1 (Millipore, Billerica, MA), phosphorylated CDK1, phosphorylated CHK1 on ser317, and cleaved PARP (Cell Signaling, Danvers, MA).

Results

Establishment of spheroid CSC-like cultures from NCI-H1299 cells

The CSC-like population from a non-small cell lung cancer cell line NCI-H1299 was selectively enriched under serum-free conditions, an approach that we have previously proven to support the in vitro propagation of both normal stem cells and CSC-like cells [12,16,29]. Briefly, the parental cells were initially cultured as a conventional monolayer in serum-containing medium (Fig. 1A; left panel). After transferring into serum-free medium supplemented with bFGF and EGF for 7βdays to 10 days, the cells in the suspension were collected and plated in ultra-low attachment flasks. Within 2βweeks to 3 weeks, the floating cancer cells formed nonadherent, 3D tumor spheroids (Fig. 1A; middle panel). The tumor spheroid cells (termed CSC-like cells) were stably propagated in suspension cultures.

The CSC-like cells cultured in serum-free stem cell conditions displayed profound morphologic differences from the parental cells and proliferated at a slower rate, as determined under the BrdU incorporation assay (Fig. 1B). The fractions of proliferating cells from the CSC-like and parental cells were approximately 29.0% and 56.7%, respectively. The fraction of the cytokeratin-positive cells population in the CSC-like cells (27.4%) was smaller than that in the parental cells (44.1%; Fig. 2A). After reseeding in serum-containing medium (i.e., differentiation induction), the spheroid CSC-like cells differentiated into an adherent monolayer culture and regained an epithelial morphology within 24 h to 48 h (Fig. 1A; right panel). Compared with the CSC-like population, the differentiated cells showed a higher BrdU-incorporation rate (55.8% vs. 29%, respectively) and a greater cytokeratin-positive cell fraction (72.5% vs. 27.4%, respectively). These results suggest that the CSC-like cells represent a relatively quiescent, undifferentiated cell population with the capacity to generate more differentiated progeny that are phenotypically comparable to their parental cell population.

The original NCI-H1299 cells cultured in serum-containing media, the spheroid cells with CSC properties derived under stem-cell conditions, and the cells that differentiated from tumor spheroid cells are referred throughout this manuscript as parental cells, CSC-like cells, and differentiated cells, respectively.

Increased tumorigenic potential of CSC-like cells

The tumorigenic potential of the CSC-like cells and parental cells was determined using in vivo limiting dilution assays. Briefly, various numbers of dissociated single cells were implanted subcutaneously into athymic female mice, and the tumor volumes were calculated at regular intervals for 42 days. Except for the groups of mice inoculated with 2 × 10³ cells/animal (both CSC-like and parental cells), the xenograft tumors were consistently developed in all other groups (Table 1). Greater tumor incidence was observed in the CSC-like cell group injected with 2 × 10⁴ cells than in the parental cell group. On Day 34, the volume of the CSC-like cell-derived tumors in the mice inoculated with 2 × 10⁴ cells was greater than that of the parental cell-derived tumors (Fig. 2B; P<0.05). Similarly, on Day 42, the average volume of the CSC-like cell-derived tumors was greater than that of the parental cell-derived tumors in the mice inoculated with 2 × 10⁴ (P<0.01) and 2 × 10⁵ (P<0.01) cells (Fig. 2B). These results suggest that tumorigenic potentials are significantly different between CSC-like cells and the parental cells in the groups inoculated with 2 × 10⁴ cells per animal to 2 × 10⁵ cells per animal parental cells. The CSC-like cells exhibited higher tumorigenic potential than the parental cells.

Self-renewal capability of CSC-like cells

The spheroid-forming ability of the CSC-like cells was determined using in vitro self-renewal assays, as previously described [11,16]. Briefly, the CSC-like cells were dissociated and plated in ultra-low attachment 96-well plates at different densities. The spheroids in each well were counted manually after 7 days. Approximately 50% of the CSC-like cells demonstrated the capacity to reform spheroids, implying their self-renewal ability (Fig. 3A). Additionally, a fraction of the primary cells dissociated from the CSC-like cell-derived xenograft tumors consistently reformed tumor spheroids under stem cell conditions (Fig. 3B). When injected into athymic mice, 2 × 10⁴ cells of the 2nd generation of cultured spheroid cells formed palpable tumors within 20 days (data not shown). These results suggest that the self-renewal potential of the spheroid CSC-like cells persists after in vivo transplantation.

Gene expression profiles of cultured parental and CSC-like cells and their xenograft tumors

The gene expression profiles of the parental cells, CSC-like cells, and xenograft tumors derived from both cell types were examined using Affymetrix array analysis (n = 3). The expression data was deposited at GEO (www.ncbi.nlm.nih.gov/geo/) with accession number GSE21612. TaqMan quantitative RT-PCR of these RNA samples was also performed for SFN1, YES1, JAK3, CALD1, SPP1, CD59, and the house keeping gene GAPDH to validate the array data. The results demonstrate that, consistent with the array data and using quantitative RT-PCR, overexpression of these genes was detected in the cultured CSC-like cells and in the xenograft tumors derived from the CSC-like population relative to their parental counterparts (Fig. 3C).

Unsupervised hierarchical cluster analysis of the gene expression profiles of all in vitro cultures and in vivo xenograft samples revealed two distinct clusters: one consisted only of cultured parental cells, and the other consisted of cultured CSC-like cells and xenografts derived from the CSC-like and parental cells (Fig. 4A and Fig.4B). PCA on genome-wide mRNA profiles of all samples illustrates their relative distribution (Fig. 4C). Clearly, the gene signatures of both clusters are distinct. The gene expression profile of parental cells in culture was markedly different from that of their derived xenografts. By contrast, the gene expression profile of the CSC-like cells in culture showed substantial similarities to that of their derived xenografts. The gene expression profile of the CSC-like cells was markedly different from that of the cultured parental cells. In the CSC-like cells, 46 genes were upregulated whereas 94 genes were downregulated (three fold cut off; Fig. 4B; Table 2). Notably, CD59 was consistently overexpressed in both the cultured CSC-like cells and their derived xenograft tumors (Fig. 4B).

Phenotypic characterization of CSC-like cells

The expression of stem cell markers CD44 and CD133 combined with EpCAM was determined using flow cytometry. In general, all three cell populations were negative for EpCAM (Fig. 5). The percentages of CD44 expressing cells in parental, CSC-like, and differentiated cells were 99.85%, 77.44%, and 99.65%, respectively. By contrast, the percentages of CD133 expressing cell population were extremely low in parental (0.08%), CSC-like (0.14%), and differentiated cells (0.037%). This finding demonstrates that transition of CSC-like cells into the differentiated progeny is accompanied by the acquisition of certain phenotypic characteristics of parental cells (i.e., CD44 expression). However, both CD44 and CD133 may not be appropriate markers for the CSC fraction in NCI-H1299-derived CSC-like population.

In vitro sensitivity of CSC-like cells to chemotherapeutics

In vitro sensitivity of parental and CSC-like cells was determined by cell viability assay. Compared to parental cells, the CSC-like cells were less sensitive to doxorubicin, etoposide, 5-fluorouracil, and cisplatin (Fig. 6A). The IC₅₀ values in CSC-like cells were higher than parental cells for doxorubicin (1.95 µM vs. 0.17 µM), etoposide (3.11 µM vs. 1.37 µM), 5-FU (20.23 µM vs. 12.00 µM), and cisplatin (4.14 µM vs. 0.18 µM), demonstrating chemoresistance of CSC-like cells. These results are in agreement with earlier studies showing the resistance of CSCs isolated from various types of human tumors [7-9].

PF-00477736 enhances the antiproliferative activity of gemcitabine in both parental and CSC-like populations

CHK1 inhibitors were developed to treat cancers in combination with standard-of-care agents to boost the effectiveness of the cytotoxic agent [28]. The effectiveness of gemcitabine against non-small cell lung cancer has been recently shown in a neoadjuvant setting [30]. To evaluate whether combined gemcitabine and CHK1 inhibitor therapy enhances the effectiveness of gemcitabine in lung CSC-like populations, we investigated the efficacy of PF-0477736 in combination with gemcitabine. PF-00477736 treatment alone (540 nM) resulted in 17% growth inhibition of the parental cells and 13% growth inhibition of the CSC-like cells (Fig. 6B). Although gemcitabine treatment alone (50 nM) resulted in 23.2% growth inhibition of parental cells, it was less effective against the CSC-like cells (6.5%). Co-administration of PF-00477736 and gemcitabine resulted in an approximately twofold increase in growth inhibition against both parental cells and CSC-like cells. These data suggest that the combination treatment enhances the antiproliferative effect of gemcitabine in both cell populations regardless of the differential responses of the two cell populations to the treatment.

Resistance of CSC-like cells to gemcitabline-induced cellular apoptosis

Although CHK1 inhibitor enhances the antiproliferative effect of gemcitabine against both CSC-like and parental cells, CSC-like cells seemed more resistant to the gemcitabine treatment and the combined gemcitabine and PF-00477736 treatment (Fig. 6B). To understand the mechanisms underlying the intrinsic resistance of CSC-like cells, we evaluated the changes in the DNA damage response pathway after the treatments. As expected, PF-00477736 alone had no effect on the cellular levels of the phosphorylated CHK1 (pCHK1) and CDK1 (pCDK1) proteins in both cell types compared with the vehicle controls (Fig. 6C). By contrast, gemcitabine treatment alone for 24 h and for 48 h upregulated the pCHK1 and pCDK1 levels in both the parental and the CSC-like cells. These results suggest that gemcitabine triggered the activation of the CHK1-mediated S-phase checkpoint in both cell types. However, the combined gemcitabine and PF-00477736 treatment prevented such increase in pCDK1 induced by gemcitabine, implicating checkpoint abrogation and cell entry into mitosis.

Treatment of parental cells with gemcitabine for 48 h induced poly(ADP-ribose) polymerase (PARP) cleavage, a apoptotic cell marker (Fig. 6C). Gemcitabine-induced apoptosis was enhanced when the treatment was combined with PF-00477736. These data suggest that inhibiting CHK1 in the presence of gemcitabine induces checkpoint abrogation, followed by mitotic catastrophe and cell death in the parental cells. By contrast, PARP cleavage was not detected in CSC-like cells after the gemcitabine treatment or the combination treatment, which suggests that the treatments do not activate the PARP-mediated apoptotic pathway. These results imply that although the combination treatment triggers the abrogation of the gemcitabine-induced S-phase checkpoint in both parental and CSC-like cells, CSC-like cells escape the apoptotic process that normally follows checkpoint abrogation and mitosis through alternative mechanisms.

Discussion

Recent progress in cancer biology emphasizes that eradication of cancer requires targeting and eliminating CSCs. However, the response of CSCs to novel therapeutic agents during development remains largely unknown. To this end, establishing CSC relevant models is crucial for testing our compounds in preclinical and clinical development.

Ideally, CSCs could be isolated and propagated from freshly resected clinical samples. However, as we have described in our previous study [16], such an approach is limited by the availability of fresh samples, fairly low success rates, and its time-consuming process. Increasing evidence suggests that CSCs may persist in some cancer cell lines [12,21-27], which provides a feasible approach for generating CSC-relevant models. Moreover, the biological differences between traditional cancer cell lines and their CSC populations remain largely unknown, particularly when these models are applied to routine drug discovery. Using NCI-H1299 cell line as a model system, we isolated a subpopulation with CSC properties from the parental cell population. Our data demonstrate the substantial differences between the responses to anticancer treatments of the two cell populations.

Consistent with the notion that CSC are relatively quiescent as reported previously [6,16], our study demonstrated that CSC-like cells propagated at a lower proliferation rate in vitro. Regardless of their relatively quiescent state in vitro, the CSC-like cells exhibited higher tumorigenic potential and the ability to self-renew when inoculated into mice. Moreover, CSC-like cells are resistant to multiple conventional cytotoxic drugs. These data confirm the CSC properties of isolated tumor spheroid populations. Spheroid formation under stem cell culture conditions has been successfully utilized to isolate CSCs from a variety of cancer cell lines [12,26,27,31]. However, the resulting tumor spheroids require further investigation to confirm their CSC characteristics, particularly through functional assays such as tumorigenicity, drug resistance, and self-renewal. In our studies, we found that tumor spheroids derived from some cancer cell lines are less tumorigenic in mice or did not show substantial resistance to chemotherapeutic agents (data not shown). Therefore, tumor spheroid formation may not be a generic approach for isolating CSC-like cells from any cancer cell lines.

The phenotype and genotype of primary human tumors are better preserved in cancer cells and tumor spheroids cultured under serum-free conditions [19,20]. In our study, we compared the gene expression profiles of CSC-like cells isolated from the NCI-H1299 cancer cell line. Surprisingly, the global gene expression profile of the cultured CSC-like cells was substantially similar to that of their derived xenografts. The similar expression profiles were observed among xenograft tumors derived from the parental cells, xenograft tumors from CSC-like cells, and cultured CSC-like cells. These data imply the persistence of a small subpopulation of CSC-like cells in the parental cell line. This subpopulation produces the spheroid CSC-like cells and drives the tumor formation in mice. Therefore, the expression profiles of the resulting xenograft tumors derived from parental cells and CSC-like cells share more similarities than differences. Based on these results, we logically assumed that the preclinical use of CSC-like cells and their xenografts may provide better information for the prediction of the clinical outcome of novel agents in both in vitro and in vivo assays.

Our study found that the expression of CD133, a stem cell marker of lung tumors [17,32] was fairly low in both the parental cells and the CSC-like cells. These data suggest that CD133 may not be an appropriate marker for the CSC-like cells isolated from the NCI-H1299 cell line. The expression of CD44, another potential marker for lung CSCs [33], was also lower in the NCI-H1299 CSC-like cells. Increased CD59 expression was observed in an Affymetrix array and real-time RT-PCR analysis of the CSC-like cells compared with the parental cells. The increased CD59 expression in the CSC-like cells may be crucial for their survival from complement-dependent and antibody-mediated killing [34]. Whether CD59 is suitable for distinguishing CSC subpopulations from the parental cells requires further investigation.

CHK1 kinase is a target for developing novel therapy. As a key element in the DNA damage response pathway, CHK1 kinase is activated in response to agents that cause DNA damage. In tumor cells, CHK1 activation induces the S-phase checkpoint and causes cell cycle arrest, allowing time for DNA repair, which impairs the effectiveness of many chemotherapeutic agents. Therefore, inhibiting CHK1 creates a “synthetic lethal” response by overriding the checkpoint defense of tumor cells against the lethal damage induced by DNA-directed chemotherapeutic agents, thereby boosting the therapeutic effects of chemotherapy.

PF-00477736 is a potent, selective ATP-competitive small molecule inhibitor that efficiently inhibits CHK1 [28]. The radioresistance of glioma stem cells has been attributed to the activation of the DNA damage checkpoint response and the increase in DNA repair capacity [36]. However, whether CHK1 inhibitors in combination with cytotoxic agents can overcome drug resistance in a CSC population remains unclear. In this study, we investigated whether PF-00477736 prevents the activation of checkpoint induced by chemotherapeutic agent gemcitabine in lung CSC-like population. We found that the combination of the cytotoxic drug gemcitabine and PF-00477736 results in an enhanced antiproliferative effect in CSC-like cells, which indicates the benefit of using the combination treatment. CSC-like cells were significantly less sensitive to treatment with gemcitabine alone as well as to the combination treatment with PF-004477736. Furthermore, we found that compared with their parental cells, CSC-like cells are highly resistant to apoptosis induced by gemcitabine alone and by the combination treatment. These results suggest that the combined gemcitabine and CHK1 inhibitor therapy enhances the antiproliferative effect of gemcitabine in CSC-like cells even though the treatment incompletely depletes CSC populations in lung cancers. Further understanding drug resistance mechanisms using the CSC-relevant models becomes important for developing more effective therapeutic strategies.

In summary, tumor spheroid cells derived from the NCI-H1299 cancer cell line exhibit the important functional properties of lung cancer CSCs. The CSC-like population is highly tumorigenic and drug resistant. They are capable of differentiation and self-renewal, fulfilling the criteria for CSCs. Moreover, compared with their parental cells, the CSC-like cells are highly similar to their derived xenograft tumors at the transcriptional level. Furthermore, the novel CHK1 inhibitor PF-00477736 enhances the antiproliferative effect of gemcitabine on the CSC-like cells. However, the CSC-like cells seem to be resistant to gemcitabine-induced apoptosis. Overall, CSC-like cells isolated from cancer cell lines may provide a better model system for preclinical studies on anticancer agents.

Compliance with ethical guidelines

All of the authors are/were employees of Pfizer, Inc. All institutional and international guidelines for the care and use of laboratory animals were followed.

References

[1] Sharma SV, Haber DA, Settleman J. Cell line-based platforms to evaluate the therapeutic efficacy of candidate anticancer agents. Nat Rev Cancer2010; 10(4): 241–253

[2] Johnson JI, Decker S, Zaharevitz D, Rubinstein LV, Venditti JM, Schepartz S, Kalyandrug S, Christian M, Arbuck S, Hollingshead M, Sausville EA. Relationships between drug activity in NCI preclinical in vitro and in vivo models and early clinical trials. Br J Cancer2001; 84(10): 1424–1431

[3] Voskoglou-Nomikos T, Pater JL, Seymour L. Clinical predictive value of the in vitro cell line, human xenograft, and mouse allograft preclinical cancer models. Clin Cancer Res2003; 9(11): 4227–4239

[4] Jimeno A, Feldmann G, Suárez-Gauthier A, Rasheed Z, Solomon A, Zou GM, Rubio-Viqueira B, García-García E, López-Ríos F, Matsui W, Maitra A, Hidalgo M. A direct pancreatic cancer xenograft model as a platform for cancer stem cell therapeutic development. Mol Cancer Ther2009; 8(2): 310–314

[5] Stingl J, Caldas C. Molecular heterogeneity of breast carcinomas and the cancer stem cell hypothesis. Nat Rev Cancer2007; 7(10): 791–799

[6] Clarke MF, Dick JE, Dirks PB, Eaves CJ, Jamieson CH, Jones DL, Visvader J, Weissman IL, Wahl GM. Cancer stem cells—perspectives on current status and future directions: AACR Workshop on cancer stem cells. Cancer Res2006; 66(19): 9339–9344

[7] Todaro M, Alea MP, Di Stefano AB, Cammareri P, Vermeulen L, Iovino F, Tripodo C, Russo A, Gulotta G, Medema JP, Stassi G. Colon cancer stem cells dictate tumor growth and resist cell death by production of interleukin-4. Cell Stem Cell2007; 1(4): 389–402

[8] Ma S, Lee TK, Zheng BJ, Chan KW, Guan XY. CD133⁺ HCC cancer stem cells confer chemoresistance by preferential expression of the Akt/PKB survival pathway. Oncogene2008; 27(12): 1749–1758

[9] Salmaggi A, Boiardi A, Gelati M, Russo A, Calatozzolo C, Ciusani E, Sciacca FL, Ottolina A, Parati EA, La Porta C, Alessandri G, Marras C, Croci D, De Rossi M. Glioblastoma-derived tumorospheres identify a population of tumor stem-like cells with angiogenic potential and enhanced multidrug resistance phenotype. Glia2006; 54(8): 850–860

[10] Gupta PB, Onder TT, Jiang G, Tao K, Kuperwasser C, Weinberg RA, Lander ES. Identification of selective inhibitors of cancer stem cells by high-throughput screening. Cell2009; 138(4): 645–659

[11] Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, Dirks PB. Identification of a cancer stem cell in human brain tumors. Cancer Res2003; 63(18): 5821–5828

[12] Fang D, Nguyen TK, Leishear K, Finko R, Kulp AN, Hotz S, Van Belle PA, Xu X, Elder DE, Herlyn M. A tumorigenic subpopulation with stem cell properties in melanomas. Cancer Res2005; 65(20): 9328–9337

[13] Taylor MD, Poppleton H, Fuller C, Su X, Liu Y, Jensen P, Magdaleno S, Dalton J, Calabrese C, Board J, Macdonald T, Rutka J, Guha A, Gajjar A, Curran T, Gilbertson RJ. Radial glia cells are candidate stem cells of ependymoma. Cancer Cell2005; 8(4): 323–335

[14] Ponti D, Costa A, Zaffaroni N, Pratesi G, Petrangolini G, Coradini D, Pilotti S, Pierotti MA, Daidone MG. Isolation and in vitro propagation of tumorigenic breast cancer cells with stem/progenitor cell properties. Cancer Res2005; 65(13): 5506–5511

[15] Ricci-Vitiani L, Lombardi DG, Pilozzi E, Biffoni M, Todaro M, Peschle C, De Maria R. Identification and expansion of human colon-cancer-initiating cells. Nature2007; 445(7123): 111–115

[16] Fang DD, Kim YJ, Lee CN, Aggarwal S, McKinnon K, Mesmer D, Norton J, Birse CE, He T, Ruben SM, Moore PA. Expansion of CD133(+) colon cancer cultures retaining stem cell properties to enable cancer stem cell target discovery. Br J Cancer2010; 102(8): 1265–1275

[17] Eramo A, Lotti F, Sette G, Pilozzi E, Biffoni M, Di Virgilio A, Conticello C, Ruco L, Peschle C, De Maria R. Identification and expansion of the tumorigenic lung cancer stem cell population. Cell Death Differ2008; 15(3): 504–514

[18] Zhang S, Balch C, Chan MW, Lai HC, Matei D, Schilder JM, Yan PS, Huang TH, Nephew KP. Identification and characterization of ovarian cancer-initiating cells from primary human tumors. Cancer Res2008; 68(11): 4311–4320

[19] Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM, Pastorino S, Purow BW, Christopher N, Zhang W, Park JK, Fine HA. Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serum-cultured cell lines. Cancer Cell 2006; 9(5): 391–403

[20] De Witt Hamer PC, Van Tilborg AA, Eijk PP, Sminia P, Troost D, Van Noorden CJ, Ylstra B, Leenstra S. The genomic profile of human malignant glioma is altered early in primary cell culture and preserved in spheroids. Oncogene2008; 27(14): 2091–2096

[21] Kondo T, Setoguchi T, Taga T. Persistence of a small subpopulation of cancer stem-like cells in the C6 glioma cell line. Proc Natl Acad Sci U S A2004; 101(3): 781–786

[22] Patrawala L, Calhoun T, Schneider-Broussard R, Zhou J, Claypool K, Tang DG. Side population is enriched in tumorigenic, stem-like cancer cells, whereas ABCGŽ2+ and ABCG2- cancer cells are similarly tumorigenic. Cancer Res2005; 65(14): 6207–6219

[23] Hermann PC, Huber SL, Herrler T, Aicher A, Ellwart JW, Guba M, Bruns CJ, Heeschen C. Distinct populations of cancer stem cells determine tumor growth and metastatic activity in human pancreatic cancer. Cell Stem Cell2007; 1(3): 313–323

[24] Huang P, Wang CY, Gou SM, Wu HS, Liu T, Xiong JX. Isolation and biological analysis of tumor stem cells from pancreatic adenocarcinoma. World J Gastroenterol2008; 14(24): 3903–3907

[25] Charafe-Jauffret E, Ginestier C, Iovino F, Wicinski J, Cervera N, Finetti P, Hur MH, Diebel ME, Monville F, Dutcher J, Brown M, Viens P, Xerri L, Bertucci F, Stassi G, Dontu G, Birnbaum D, Wicha MS. Breast cancer cell lines contain functional cancer stem cells with metastatic capacity and a distinct molecular signature. Cancer Res2009; 69(4): 1302–1313

[26] Gou S, Liu T, Wang C, Yin T, Li K, Yang M, Zhou J. Establishment of clonal colony-forming assay for propagation of pancreatic cancer cells with stem cell properties. Pancreas2007; 34(4): 429–435

[27] Zhou ZH, Ping YF, Yu SC, Yi L, Yao XH, Chen JH, Cui YH, Bian XW. A novel approach to the identification and enrichment of cancer stem cells from a cultured human glioma cell line. Cancer Lett2009; 281(1): 92–99

[28] Blasina A, Hallin J, Chen E, Arango ME, Kraynov E, Register J, Grant S, Ninkovic S, Chen P, Nichols T, O’Connor P, Anderes K. Breaching the DNA damage checkpoint via PF-00477736, a novel small-molecule inhibitor of checkpoint kinase 1. Mol Cancer Ther2008; 7(8): 2394–2404

[29] Yu H, Fang D, Kumar SM, Li L, Nguyen TK, Acs G, Herlyn M, Xu X. Isolation of a novel population of multipotent adult stem cells from human hair follicles. Am J Pathol 2006; 168(6): 1879–1888

[30] Dudek AZ, Lesniewski-Kmak K, Larson T, Dragnev K, Isaksson R, Gupta V, Maddaus MA, Kratzke RA. Phase II trial of neoadjuvant therapy with carboplatin, gemcitabine plus thalidomide for stages IIB and III non-small cell lung cancer. J Thorac Oncol2009; 4(8): 969–975

[31] Rappa G, Mercapide J, Anzanello F, Prasmickaite L, Xi Y, Ju J, Fodstad O, Lorico A. Growth of cancer cell lines under stem cell-like conditions has the potential to unveil therapeutic targets. Exp Cell Res2008; 314(10): 2110–2122

[32] Tirino V, Camerlingo R, Franco R, Malanga D, La Rocca A, Viglietto G, Rocco G, Pirozzi G. The role of CD133 in the identification and characterisation of tumour-initiating cells in non-small-cell lung cancer. Eur J Cardiothorac Surg2009; 36(3): 446–453

[33] Basak SK, Veena MS, Oh S, Huang G, Srivatsan E, Huang M, Sharma S, Batra RK. The malignant pleural effusion as a model to investigate intratumoral heterogeneity in lung cancer. PLoS ONE2009; 4(6): e5884

[34] Fishelson Z, Donin N, Zell S, Schultz S, Kirschfink M. Obstacles to cancer immunotherapy: expression of membrane complement regulatory proteins (mCRPs) in tumors. Mol Immunol2003; 40(2–4): 109–123

[35] Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, Dewhirst MW, Bigner DD, Rich JN. Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature2006; 444(7120): 756–760

[36] Lundholm L, Hååg P, Zong D, Juntti T, Mörk B, Lewensohn R, Viktorsson K. Resistance to DNA-damaging treatment in non-small cell lung cancer tumor-initiating cells involves reduced DNA-PK/ATM activation and diminished cell cycle arrest. Cell Death Dis2013; 4: e478

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg