Functional regulatory variants of MCL1 contribute to enhanced promoter activity and reduced risk of lung cancer in nonsmokers: Implications for context-dependent phenotype of an antiapoptotic and antiproliferative gene in solid tumor
Article first published online: 1 SEP 2011
Copyright © 2011 American Cancer Society
Volume 118, Issue 8, pages 2085–2095, 15 April 2012
How to Cite
Jiang, Y., Wang, W., Wang, J., Lu, Y., Chen, Y., Jin, L., Lin, D., He, F. and Wang, H. (2012), Functional regulatory variants of MCL1 contribute to enhanced promoter activity and reduced risk of lung cancer in nonsmokers: Implications for context-dependent phenotype of an antiapoptotic and antiproliferative gene in solid tumor. Cancer, 118: 2085–2095. doi: 10.1002/cncr.26502
- Issue published online: 6 APR 2012
- Article first published online: 1 SEP 2011
- Manuscript Accepted: 18 JUL 2011
- Manuscript Revised: 9 JUN 2011
- Manuscript Received: 5 DEC 2010
- lung cancer;
- regulatory polymorphism;
- case-control study
Dysfunction of molecules that regulate both apoptosis and proliferation is involved in tumorigenesis. A common insertional polymorphism in promoter of MCL1, a member of BCL2 family gene with the dual regulatory functions, has been shown to be functional in leukemia, but its association with cancer predisposition and prognosis has not been well established. We hypothesized that MCL1 promoter variants may modify risk of solid cancer.
We genotyped −190 insertional polymorphism and 3 linked single nucleotide polymorphisms (SNPs) (−627A>C, −298G>C, and −235C>A) in 320 lung cancer patients and 362 controls, and analyzed their functional significance.
We confirmed that these regulatory variants correlated with enhanced promoter activity and elevated expression of both mRNA and protein in solid cancer cells and tissues. We further demonstrated that heightened expression of MCL1 resulted in decreased proliferation ability of lung cancer cells. We found a reduced cancer risk (adjusted odds ratio [OR] = 0.47; 95% confidence interval [CI] = 0.25-0.88) associated with −190 insertional genotype. Stratification analysis further showed pronounced associations in nonsmokers (OR, 0.25; 95% CI, 0.09-0.70), in females (OR, 0.22; 95% CI, 0.07-0.74), and in the histological type of adenocarcinoma (OR, 0.18; 95% CI, 0.05-0.62). Likewise, homologous diplotype of these polymorhpisms that positively affected gene expression was associated with reduced risk in nonsmokers (OR, 0.19; 95% CI, 0.06-0.58).
The present study demonstrated that common variants in MCL1 promoter correlated with increased transactivation in solid cancer cells and were associated with reduced risk of lung cancer in nonsmokers, suggesting a dominant antiproliferative function of MCL1 against its antiapoptosis effect in development of solid cancer in nonsmokers. Cancer 2012. © 2011 American Cancer Society.
Apoptosis and cell progression are 2 closely linked cellular phenotypes under strict control, perturbations of regulatory molecules with dual functions for regulation of which are implicated in tumorigenesis, as exemplified by p53,1 E2F1,2 and BCL2 family members.3 MCL1 (myeloid cell leukemia sequence 1), a well-known antiapoptotic member of the BCL2 family widely expressed in human malignant cells, especially in prostate, breast, colon, and lung epithelia,4 plays its antiapoptotic role at the early stage in a cascade leading to cytochrome c release.5 On the other hand, MCL1 also functions as a cell cycle regulator by interacting with PCNA (proliferating cell nuclear antigen), which plays a critical role in DNA replication, and inhibiting cell cycle progression,6 or by binding to and negatively regulating CDK1 (cyclin-dependent kinase 1), and arresting cell cycle in S and G2 phase.7 Overexpression of MCL1 in transgenic mice results in a high incidence of lymphoma, demonstrating that MCL1 can directly contribute to tumorigenesis.8 The biochemical functions and expression profiles of MCL1 suggest that its deregulation and dysfunction, which could partly result from mutations and polymorphisms in its coding gene, could have implications for cancer predisposition and prognosis.
Recently, the presence of insertions in the MCL1 promoter was characterized and proposed to influence clinical outcome in chronic lymphocytic leukaemia (CLL). Moshynska et al9 reported a mutation of 6 nucleotide (nt) (GGCCCC) or 18 nt (GGCTCAGGCCCC GGCCCC) insertion in the MCL1 promoter, occurring 190 bp upstream of the major transcription start site, which was found to correlate with increased RNA and protein levels in CLL cells and poor survival in a cohort of CLL patients. Several follow-up studies, however, argued against these findings and confirmed that the sequence insertions are not somatic oncogenic mutations but common hereditary polymorphisms, and they were not associated with clinical outcome of leukemia.10-16 Even though correlation between the MCL1 promoter insertion and gene expression in leukemia was subsequently validated by Saxena et al,17 its functional significance and clinical implications in solid cancer are still completely unknown.
Like other antiapoptotic BCL2 members, MCL1 is also one of the survival factors in lung cancers.18 Given the effect on gene expression of the common insertional polymorphisms in the MCL1 promoter and its presence with high frequencies in the general population as well as lung cancer patients,10 we reasoned that regulatory germline variants might act as a genetic modifier of lung cancer risk. To test this hypothesis, we examined the insertional polymorphism and 3 single nucleotide polymorphisms (SNPs) at the MCL1 promoter, and investigated their association with of lung cancer risk in a case-control study.
MATERIALS AND METHODS
Subjects and Samples
The case-control study consisted of 320 patients with lung cancer and 362 healthy controls. All subjects were unrelated ethnic Han Chinese and participants in our previous epidemiological study of lung cancer.19 The cases with primary lung carcinoma were recruited from January 1997 to November 2001 at the Cancer Hospital, Chinese Academy of Medical Sciences (Beijing). The exclusive criteria included previous cancer and previous radiotherapy or chemotherapy. All patients, without age, sex, and histology restrictions, were from Beijing City and its surrounding regions. Surgically resectable patients were mostly recruited. Population controls were accrued from a nutritional survey conducted in the same region during the period of case collection. They had no history of cancer and were frequency matched to case on age and sex. Informed consent was obtained from each subject at recruitment, and this study was approved by the institutional review board of the Chinese Academy of Medical Sciences Cancer Institute.
Genomic DNA of controls and cases were extracted from peripheral blood leukocytes. Polymorphisms in the MCL1 promoter were genotyped with polymerase chain reaction (PCR)-resequencing method using forward primer 5′-AGATGGGAGAAGCAAGCAGG-3′ and reverse primer 5′-AAGCGGAAGTGAGAAGTGGC-3′. The PCR products were subject to direct sequencing on an ABI Prism 377 DNA Sequencer.
We used chi-square tests to examine differences in demographic variables, smoking status, genotypes between cases and controls, and deviations of genotype frequencies in controls and in cases from those expected under Hardy-Weinberg equilibrium (HWE). We calculated linkage disequilibrium (LD) index (r2) with LDA software20 and predicted haplotype with PHASE.21 We estimated cancer risk associated with genotypes or diplotypes by calculating crude and adjusted odds ratio (OR) and their 95% confidence interval (95% CI) with unconditional logistic regression models. Age, sex, and pack-years smoked were included for multivariate adjustment. Information was collected on the number of cigarettes smoked per day, and the age at which the subjects started smoking and stopped smoking. Subjects who never smoked or smoked <100 cigarettes before the date of diagnosis for patients or the date of interview for controls were defined as nonsmokers. Light and heavy smokers were categorized using 50 percentile pack-year ([cigarettes per day/20] × ([years smoked]) values of controls as cutoff points (ie, ≤26 and >26 pack-years). The probability level of <.05 was used as criterion of statistical significance, and all statistical tests were 2 sided. These statistical analyses were performed by using Stata software (version 8.0).
Electrophoretic Mobility Shift Assay (EMSA)
We used the EMSA assay to analyze allele-specific binding of nuclear protein to the MCL1 promoter. Synthetic double-stranded oligonucleotides 5′-CTCAGGCCCCGGC CCCGGCC-3′ (0 nt), 5-′CTCAGGCCCCGGCCCCG GCCCCGGCC-3′ (6 nt), and 5′-CTCAGGCCCCGGC TCAGGCCCCGGCCCCGGCCCCGGCC-3′ (18 nt) corresponding to the MCL1-190 insertion polymorphisms sequence (inserted nucleotides shown in bold) were labeled with biotin. EMSA was performed by using the LightShift Chemiluminescent EMSA Kit (Pierce Rockford, Ill). Briefly, a biotin-labeled probe was combined with nuclear protein extract from an A549 cell. The unlabeled probe and an Sp1 consensus-binding site probe (5′-GCTCGCCCCGCCCCGATCGAAT-3′) were used for competition assays. After electrophoresis of reaction mixture, gel was transferred to nylon membrane and crosslink was performed with a UV crosslinker (HL-2000 Hybrolinker, UVP). Biotin-labeled DNA on the membrane was detected with stabilized Streptavidin-horseradish peroxidase conjugate.
Reporter Gene Assay
We conducted transient transfection and luciferase assays to address differential promoter activity of MCL1 regulatory variants. Four allelic reporter constructs were prepared by amplifying the MCL1 promoter region of the 907-bp length (from −944 to −37 relative to translation start site) with primers: 5′-GGTACCGAATCGTCTGA ACG-3′ (forward) and 5′-GGAAGACCCCGACTCC TTA-3′ (reserve), followed by cloning into a pGL2 basic vector. Human lung cancer cell line A549 and esophageal cancer cell line Eca109 were cotransfected with these constructs and pRL-CMV vector, as internal control, using Lipofectamine 2000 (Invitrogen, Carlsbad, Calif). The pGL2 basic vector without any insertion was used as the negative control. For each plasmid construct, 3 independent transfection experiments were performed and each were done in triplicate. Luciferase activity was measured using Dual-Luciferase Reporter Assay System (Promega, Madison, WI) on a Veritas microplate luminometer (Turner Biosystems, Madison, Wisc). Transcriptional activity for each construct was calculated as fold change of luciferase activity of each construct to that of the pGL2 basic vector.
Real-time quantitative RT-PCR
To analyze the correlation between genotype and transcript expression, we measured MCL1 mRNA in lung cancer tissues. Total RNA was isolated from lung tumor tissues with Trizol. After first-strand cDNA was synthesized, real-time quantitative RT-PCR was performed with SYBR Premix Ex Taq Kit (Takara, Japan) on Applied Biosystems 7500. cDNA was amplified with primers 5′-GAGGCTG CTTTTCTTCGC-3′ (forward) and 5′-CCGTCCGTAC TGGGTTATT-3′ (reverse). Expression of beta-actin was measured as internal control with primers 5′-CAGAG CCTCGCCTTTGCC-3′ (forward) and 5′-ATGCCGGA GCCGTTGTCG-3′ (reserve). Expression levels were normalized to the referent by relative quantification using ΔΔCT method. Expression analysis was performed in triplicate for each sample.
We measured MCL1 protein in lung cancer tissues to analyze correlation between genotype and protein expression. Formalin-fixed, paraffin-embedded lung cancer tissues sectioned at 4 μm were prepared for immunohistochemical analysis of MCL1. The endogenous peroxidase activity was blocked with 3% H2O2. The slides were incubated over night at 4°C with antibody for MCL1 (rabbit polyclonal antibody, ProteinTech Group, Chicago, Ill). Adding biotin-labeled goat antirabbit IgG for 30 minutes, peroxidase-labeled avidin-biotin for 1 hour, then DAB for enzymatic development of peroxidase. Sections were then counterstained with hematoxylin, dehydrated, mounted, and coverslipped. For each slide, mean density for images from light microscopy random fields (×200) representing the expression level of MCL1 was calculated using Image-pro plus software (Bethesda, Md).
Cell Proliferation Assay
We further analyzed the effect of MCL1 expression on proliferation of lung cancer cell line A549. Full coding sequence of MCL1 cDNA was amplified and cloned into pcDNA3.1 (called pcDNA3.1-MCL1), using forward primer: 5′-ACCCTCGAG (Xhol) TCGGGGTCTTCCC CAGTTTT-3′ and reverse primer: 5′-ACCGGATCC (BamHl) ATCTTATTAGATATGCCA AACCAGCT-3′. MCL1 siRNA and nontargeting negative control (NC) siRNA were commercially synthesized (RiboBio Co, Ltd, China). The targeting sense sequence is 5′-GGACTT TTATACCTGTTAT-3′. The NC siRNA sequence is 5′-TTCTCCGAACGTGTCACGTdTdT-3′. MCL1 and siRNA expression vectors were transiently transfected into A549 cells with Lipofectamine 2000 (Invitrogen). After 72 hours, lysates were prepared for Western blot with MCL1 antibody (Protein Tech Group, Inc, China) and β-Actin antibody (Abmart, China) as inner control. The A549 cells were counted and plated at 4 × 103 cells/well in 96-well plate. After 24 hours culture, MCL1-siRNA and NC-siRNA at 50 nM/well, pcDNA3.1-MCL1 and pcDNA3.1 at 0.2 μg/well were transfected, respectively, in 6 wells. Cell viability was detected using Cell Counting Kit-8 (Dojindo Laboratories, Japan) 1, 3, and 5 days after transfection.
MCL1 Promoter Polymorphisms Are Associated with Risk of Lung Cancer
We first addressed implications of the MCL1 promoter polymorphisms in the etiology of lung cancer using a case-control study, which involved 320 cases (115 with adenocarcinoma, 143 with squamous cell carcinoma, and 62 with other histological types) and 362 controls. The distributions of sex, age, and smoking status among the subjects are summarized in Table 1. There were no significant differences between cases and controls in terms of distributions of sex and age, suggesting that the frequency matching was adequate. However, compared with controls, cases were more likely to be smokers (ever-smokers: 62.8% in cases and 33.4% in controls, P < .001). In addition, cases had a higher value of pack-years smoked than controls; 41.9% smokers among cases smoked >26 pack-years compared with 14.9% among controls (P = .000).
|(N = 320)||(N = 362)|
|Male||229 (71.6)||253 (69.9)||.632a|
|Female||91 (28.4)||109 (30.1)|
|Mean age (SD)||57.7 (9.4)||55.6 (9.5)||.098b|
|Smoking status (%)|
|Never||119 (37.2)||241 (66.6)||<.001a|
|Ever||201 (62.8)||121 (33.4)|
|≤26 pack-years||67 (20.9)||67 (18.5)||<.001a|
|>26 pack-years||134 (41.9)||54 (14.9)|
|Mean pack-years (SD)||34.2 (21.3)||26.3 (15.0)||<.001b|
|Histological type (%)|
|Squamous cell carcinoma||143 (44.7)|
We amplified and sequenced the 674-bp region upstream from first codon of MCL1, characterized 4 polymorphisms, −627A>C (rs3738486), −298G>C (rs3738485), −235C>A (rs3738484), and −190 indel (rs3831987) (upstream from the transcription start site). Their genotypic frequencies in cases and controls are shown in Table 2. Their observed genotypic distributions in controls and in cases were in HWE. The −190 indel polymorphism was in relatively weak LD (r2 = 0.28 and 0.28 in controls and cases, respectively) with the adjacent 3 SNPs, which were in completed LD with each other in both controls and cases (r2 = 1.00) (−235C>A was therefore used to represent the 3 SNPs in the following analyses). We observed significantly different genotype distribution of −190 indel between cases and controls (P = .038), which was mainly due to remarkable overpresentation of homozygous insertion genotype (ins/ins) in controls than in cases (11.9% vs 5.6%; chi-square = 6.08, P = .017).
|Genotype||Controls (n = 362)||Cases (n = 320)||P valuea||Crude OR (95% CI)b||Adjusted OR (95% CI)b|
|−235 C>A (rs3738484)|
|C/C||143||39.5||108||33.8||0.259||1.00 (Ref)||1.00 (Ref)|
|C/A||164||45.3||154||48.1||1.24 (0.88-1.76)||1.61 (1.11-2.32)|
|A/A||55||15.2||58||18.1||1.40 (0.87-2.23)||1.52 (0.94-2.45)|
|−190 indel (rs3831987)|
|0/0||178||49.2||155||48.5||0.038||1.00 (Ref)||1.00 (Ref)|
|0/ins||141||38.9||147||45.9||0.009||1.20 (0.86-1.66)||1.12 (0.80-1.57)|
|ins/ins||43||11.9||18||5.6||0.48 (0.25-0.89)||0.47 (0.25-0.88)|
By using logistic regression analysis, we assessed association between cancer risk and genotypes of −235C>A and −190 indel (Tables 2 and 3). We found significant increase in cancer risk among subjects carrying −235C>A C/A (OR, 1.61; 95% CI, 1.11-2.32) compared with wild C/C genotype. On the other hand, homozygous insertion genotype of −190 indel (ins/ins) was associated with substantial reduction in cancer risk compared with wild 0/0 genotype (OR, 0.47; 95% CI, 0.25-0.88). In further stratification analysis by sex, age, smoking status, and histology type, we found that −235C>A A/A genotype was associated in dominant genetic mode with increased cancer risk in females (OR, 3.10; 95% CI, 1.20-7.99) and nonsmokers (OR, 2.59; 95% CI, 1.28-5.23). In like manner, we observed pronounced association in recessive genetic mode of −190 indel insertion genotype (ins/ins) with reduced cancer risk in females (OR, 0.22; 95% CI, 0.07-0.74), nonsmokers (OR, 0.25; 95% CI, 0.09-0.70), and in the histological type of adenocarcinoma (OR, 0.18; 95% CI, 0.05-0.62). We further separately estimated haplotype frequencies in cases and controls, and observed balanced distribution of 4 haplotypes between cases and controls (data not shown). As an exploratory effort, we reconstructed diplotypes for each subjects and found 1 diplotype (C-ins/C-ins), compared with the referent diplotype (A-0/A-0), was associated with significantly reduced risk in overall population (OR, 0.42; 95% CI, 0.21-0.84), especially in nonsmokers (OR, 0.19; 95% CI, 0.06-0.58) and in females (OR, 0.12; 95% CI, 0.03-0.55). Taken together, this case-control study suggested that wild or heterozygous −235 C>A and homozygous −190 insertion were associated with reduced risk of lung cancer in nonsmokers and females.
|Variables||−235 C>A (Cases/Controls)||−190 Indel Polymorphism (Cases/Controls)|
|C/C||C/A||OR (95% CI)a||A/A||OR (95% CI)a||0/0||0/Ins||OR (95% CI)a||Ins/Ins||OR (95% CI)a|
|Male||79/93||106/116||1.39 (0.89-2.17)||44/44||1.22 (0.70-2.12)||112/136||103/91||1.31 (0.88-1.97)||14/26||0.67 (0.32-1.41)|
|Female||29/50||48/48||1.97 (1.03-3.77)||14/11||3.10 (1.20-7.99)||43/42||44/50||0.77 (0.42-1.43)||4/17||0.22 (0.07-0.74)|
|≤55 y old||34/67||55/88||1.18 (0.63-2.21)||25/27||1.61 (0.65-3.98)||62/86||44/72||0.79 (0.46-1.35)||8/24||0.44 (0.18-1.12)|
|>55 y old||74/76||99/76||1.64 (1.02-2.63)||33/28||1.25 (0.66-2.36)||93/92||103/69||1.48 (0.95-2.31)||10/19||0.52 (0.22-1.23)|
|Nonsmokers||32/100||64/111||1.94 (1.15-3.26)||23/30||2.59 (1.28-5.23)||62/113||52/97||0.88 (0.55-1.42)||5/31||0.25 (0.09-0.70)|
|Smokers||76/43||90/53||1.18 (0.69-2.01)||35/25||0.91 (0.47-1.77)||93/65||95/44||1.40 (0.86-2.29)||13/12||0.76 (0.32-1.84)|
|≤26 pack-years||21/23||34/36||0.94 (0.43-2.04)||12/8||1.58 (0.52-4.38)||28/34||34/25||1.63 (0.79-3.38)||5/8||0.81 (0.22-2.92)|
|>26 pack-years||55/20||56/17||1.37 (0.62-3.01)||23/17||0.69 (0.30-1.61)||65/31||61/19||1.33 (0.67-2.65)||8/4||0.93 (0.25-3.40)|
|Adenocarcinoma||37/143||60/164||1.35 (0.79-2.28)||18/54||1.37 (0.69-2.72)||61/178||51/141||0.96 (0.61-1.52)||3/43||0.18 (0.05-0.62)|
|SSC||51/143||67/164||1.57 (0.94-2.62)||25/54||1.26 (0.65-2.41)||64/178||68/141||1.19 (0.85-1.66)||11/43||0.85 (0.39-1.88)|
|Others||20/143||27/164||1.27 (0.67-2.37)||15/54||1.95 (0.91-4.18)||30/178||28/141||1.31 (0.88-1.97)||4/43||0.53 (0.17-1.62)|
MCL1 Promoter −190 Insertions Increase Binding Affinity for Nuclear Protein
Using transcription factor binding site predicting software (Alibaba2), we found that sequences surrounding −190 indel site in the MCL1 promoter might be a cis-element potentially targeted by Sp1, a ubiquitous zinc-finger transcription factor (Fig. 1A). Interestingly, the presence of insertion alleles (6 or 18 nt) extends length of the putative Sp1 binding sites, suggesting that −190 insertions may increase the affinity of Sp1 to this region. EMSA in A549 cells confirmed binding of the cis-element with Sp1 and different affinity of deletion and insertion alleles. All of the 3 biotin-labeled oligonucleotide probes, corresponding to the 0-, 6-, or 18-nt insertion alleles, generated a specific DNA/nuclear protein complex (Band I), but the insertion alleles bands were obviously thicker than that of no insertion allele. The shift band was completely abolished either by 200-fold relative unlabeled allele probes or by 200-fold unlabeled Sp1 consensus probe (Fig. 1B). These results showed that −190 insertion alleles increased the affinity of Sp1 to this region in the MCL1 promoter.
MCL1 Promoter Variants Enhance Transcriptional Activity and Correlate with Increased Expression of Both mRNA and Protein
To directly determine allele-specific effects of MCL1 promoter polymorphisms (−627A>C, −298G>C, −235C>A, and −190 indel, the 3 SNPs are in complete LD) on native promoter activity, 4 luciferase reporter gene constructs (pAGC-0, pAGC-18, pAGC-6, and pCCA-6) were generated to transiently transfect A549 and Eca109 cells (Fig. 2A). As to −190 indel, reporter gene expressions driven by plasmids containing insertion alleles (pAGC-18 and pAGC-6) were significantly higher than those driven by no insertion allele plasmid pAGC-0 (P < .05) (Fig. 2B). As to the 3 linked SNPs, in the same background of wild-type of no insertion at −190 site, activities of plasmid pCCA-0 was significant higher than that of pAGC-0 (P < .05). These results suggested that MCL1 promoter variants differed in their capacity of transactivation.
To verify the effects of regulatory variants on MCL1 transcription in vivo, we measured its mRNA level in lung cancer tissues from 28 surgically resectable incident patients, and analyzed their correlation with MCL1 promoter genotypes and diplotypes (Fig. 3). We set risk-prone genotypes (−235A/A and −190indel 0/0) and diplotype (A-0/A-0) as respective reference groups. mRNA expression in group −235A/C was significantly higher than that in referent A/A group (median 0.40 vs 0.32, P = .014). mRNA expression in 0/18 heterozygotes was significantly higher than that in the referent 0/0 (median, 0.60 vs 0.29, P < .001). Furthermore, when combining genotypes of the 2 polymorphisms into diplotypes, we found that 2 diplotypes (A-0/C-18, median = 0.56, P = .018; C-0/C-18, median = 0.57, P = .034), which contain both the 2 protective alleles (−235C and −190 insertion), correlated with significantly heightened mRNA expression compared with the reference (A-0/A-0, median = 0.32). We further addressed correlation between MCL1 promoter variations and protein expression by immunohistochemically measuring MCL1 protein in lung cancer tissues from 16 patients. Consistently, MCL1 promoter genotypes or diplotypes containing protective alleles (−235C and −190 insertion) were associated with significantly heightened protein expression in lung cancer tissues compared with respective references (Fig. 4). These data further clearly demonstrated that protective genotypes of MCL1 promoter correlated with enhanced transcriptional activity and heightened gene expression in lung cancer cells.
Overexpression of MCL1 Correlates with Decreased Cell Proliferation
Because the regulatory variants of MCL1 were shown to be associated with enhanced gene expression but reduced risk of lung cancer, we analyzed the effects of expression of MCL1 on cell proliferation ability of lung cancer cell line A549. Western blot assay confirmed the knocked-down expression of MCL1 in cells treated MCL1-siRNA compared with NC siRNA, and heightened expression of MCL1 in cells transfected with pcDNA3.1-MCL1 compared with control vector (Fig. 5A). In the cell viability assay, the knocked-down expression of MCL1 correlated with significantly higher cell viability. Likewise, we also observed statistically significant correlation between overexpression of MCL1 and lower cell viability (Fig. 5B). These results revealed a possible mechanism that the association between reduced lung cancer susceptibility and MCL1 regulatory variants with enhanced promoter activity might be partly ascribed to inhibition of cell proliferation resulted from heightened expression of MCL1.
In this study, we found that common variants in MCL1 promoter influenced risk of lung cancer. Females and nonsmokers carrying homozygous genotypes for −190 insertion and wild variant of the 3 SNPs were at reduced cancer risk. We further examined functional impact of these regulatory variants and demonstrated that −190 insertional variants enhanced the binding affinity of transcription factor Sp1 to the cis-element, correlated with increased transcriptional activity and heightened gene expression in solid cancer cells. Moreover, we preliminarily demonstrated that decreased cell proliferation ability resulted from heightened MCL1 expression may partly be the underlying mechanism for the genetic association. These findings confirmed functional significance of regulatory variations in the MCL1 promoter, suggested for the first time that MCL1 polymorphism is a putative genetic modifier of risk of lung cancer.
Functional analyses of regulatory polymorphisms in the MCL1 promoter suggested that their association with risk of lung cancer might be ascribed to enhanced gene expression. The promoter with no −190 insertion (+0) contains 5 (GGCCCC) repeats and 1 (GCTCA) motif, the +6-promoter has an extra repeat and +18-promoter has 2 extra repeats and 1 extra motif, which are typically binding motifs for Sp1 as confirmed in ChIP analysis by Saxena et al.17 Our EMSA assays, luciferase reporter gene assays in solid cancer cell lines in vitro, superimposed with measurement of mRNA and protein expression in lung cancer tissues, further illustrated that −190 insertional polymorphisms enhanced MCL1 promoter activity and resulted in heightened gene expression. These results are in line with observations in previous reports.9, 17 In hematopoietic and epithelial cell lines, 6- and 18-nt insertions have positive effect on transcriptional activity of MCL1 compared with the wild-type or with an experimentally created deletion.17, 22 However, controversial results between different studies also existed. Freeman et al11 observed that insertional variants displayed decreased promoter activity. Dicker et al10 reported no significant difference in MCL1 transcript levels between samples of CLL patients containing the insertional polymorphism and the wild type. Several explanations may account for this inconsistency. First, different settings, such as tumor chemotherapy regime, genotype status of cell lines, and stimulating method for transfection, could affect MCL1 expression. Second, tissue-specific trans-acting factors other than Sp1 may complicate the effect of regulatory variants. Third, the promoter region of MCL1 being evaluated and length of the transfected segment, which may vary between different studies, are likely to influence promoter activity in vitro. In fact, we screened 3 linked common SNPs closely upstream of −190 indel polymorphism. Although we did not cover them for EMSA analysis due to they are not located in putative binding sites of any well-known transcription factor in computer prediction, they indeed correlated with differential expression of mRNA and protein in cancer tissues and with various transactivation capacity in reporter gene assays.
Further mechanistic investigation on cellular phenotype for heightened expression of MCL1 revealed that inhibition of proliferation may partly account for the association between the regulatory variants in the MCL1 promoter and decreased cancer predisposition. MCL1 is well known as a pro-survival BCL2 family member. MCL1 is frequently overexpressed in cancer cells from a variety of tumors and manifest its antiapoptotic effect in response to various stresses such as chemotherapy.4, 5 On the other hand, however, the antiproliferative phenotype of overexpression of MCL1 is thus far relatively less characterized. In this study, we used siRNA-knockdown and overexpression strategies and showed that strengthened expression of MCL1 resulted in substantial reduction in cell proliferation ability of lung cancer cells. Therefore, even though we unexpectedly observed substantial reduction in risk of lung cancer, especially in nonsmokers and females associated with functional regulatory variants of MCL1 that contribute to enhanced gene expression, these apparently paradoxical results could plausibly be explained with the context-dependent cellular phenotype of MCL1 in solid tumor.
Cumulating evidences show that members of BCL2 family not only have antiapoptotic activity but also are antiproliferative and which of the dual functions predominates is lineage specific and context dependent.3 The present study revealed that MCL1 promoter variants were associated with enhanced expression but with decreased risk of lung cancer in nonsmokers, which is apparently inconsistent with its antiapoptotic function but instead suggests its possible dominant tumor-suppressive roles in development of lung cancer in nonsmokers. The antiproliferative function of MCL1 was previously reported to be linked to its functional interaction with PCNA6 and CDK1.7 Consistently, in a recent report profiling somatic copy-number alteration in cancers, MCL1 was in an amplification peak in cancers across multiple tissue types including lung and breast cancers, and the authors observed a more pronounced reduction in proliferation rates among MCL1-amplified cell lines compared with MCL1-unamplified control cell lines.23 These reports are in line with our observations in this study that heightened expression of MCL1 resulted in decreased cell proliferation ability of lung cancer cells. Another possible mechanism for the genetic association might be linked to interaction between MCL1 and other genetic alterations in c-Myc and p53, which are common in lung cancer.18 A recent paper has reported that MCL1 overexpression correlates with poor patient survival, but only when accompanied by overexpression of the Myc protein, suggesting that MCL1 overexpression mediates a shift in the threshold of Myc activity that is required to initiate the apoptotic and ARF/p53 intrinsic tumor surveillance pathways.18 In nonsmokers, lung tissues are least frequently exposed to carcinogens in cigarette smoke. One reasonable hypothesis would be that solid tumors are characterized by a proliferative pretumor phase during which the antiproliferative effect of MCL1 could be more consequential than its antiapoptotic activity, and up-regulated MCL1 may arrest cell cycle progression, thus preventing cells from replicating altered DNAs. In support of this scenario, we also observed that −190 insertion genotype was associated with decreased risk of lung adenocarcinoma, which is the most common histological type of lung cancer in women, lifetime nonsmokers.24 On the other hand, when cells are exposed to carcinogens or DNA damaging compounds, MCL1 can regulate survival, and enhanced levels of MCL1 may protect lung cancer cells from death induced by a variety of pro-apoptotic stimuli.25 Many reports verified that overexpressed MCL1 in multiple cancer tissues displayed drug resistance and poor prognosis. In smoking status, lung cells are incessantly stimulated by tobacco carcinogens, among which nicotine could induce MCL1 phosphorylation and significantly enhance its antiapoptotic function.26 In this instance, the dual roles of MCL1 could counteract with each other, which could partly account for null association between MCL1 promoter variants and lung cancer risk in smokers.
Other members of BCL2 family also play both antiapoptotic and antiproliferative roles in tumorigenesis in a context-dependent way. In Myc-induced lymphoma, the antiapoptotic function of BCL2 is clearly dominant over its antiproliferative effect.27 In contrast, in some solid tissues such as breast, liver, and colon, multiple studies show that the effect of BCL2 on carcinogenesis is consistent with its ability to interfere with cell proliferation and cause a delay rather than an increase in tumorigenesis, suggesting the antiproliferative function of BCL2 translates into tumor suppression.3 Of note, variant allele (A) of a common SNP, −938 C>A at BCL2 promoter, was demonstrated to be an unfavorable genetic marker of CLL, displaying increased transcriptional activity, increased BCL2 protein expression in B cells from CLL patients, and poor prognosis.28 In solid tumor, however, the −938 C>A variant genotypes is associated with decreased risk of squamous cell carcinoma of the head and neck in a recent case-control study.29 These observations suggest that the balance between antiapoptotic and antiproliferative functions of BCL2 family members could be influenced by tumor pathology and genetic context.
In conclusion, the present study confirmed that −190 insertional polymorphism and neighboring linked SNPs in MCL1 promoter positively affect gene expression, revealed that overexpression of MCL1 could result in reduced cell proliferation, and provided first evidence that these functional regulatory variants were associated with significant reduction in risk of lung cancer in nonsmoking population. Because this is a case-control study and the findings of subgroups are limited due to small sample sizes in the strata, our results need to be validated by studies in other populations with larger sample size. This molecular epidemiological study implies a dominant antiproliferative function of MCL1 against its antiapoptosis effect in development of solid cancer in nonsmokers. Further functional studies are warranted to reveal complex roles of MCL1 in DNA damage and cell cycle control, and context-dependent relationship between its dual functions in tumorigenesis.
H. Wang gratefully acknowledges the Hong Kong Qiu Shi Science & Technologies Foundation membership at the Institute for Advanced Study. This study was supported by grants from the National Natural Science Foundation of China (30971593, 81172093, and 30890034), Program for New Century Excellent Talents in University (NCET-07-0204) and Shanghai Rising-Star Program (07QA14006).
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.