Japanese Journal of Clinical Oncology Advance Access published online on February 25, 2008
Japanese Journal of Clinical Oncology, doi:10.1093/jjco/hyn007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© The Author (2008). Published by Oxford University Press. All rights reserved
Lung Cancer Risk Associated with Thr495Pro Polymorphism of GHR in Chinese Population
1 Medical School and State Key Laboratory of Pharmaceutical Biotechnology, Life Science College, Nanjing University, Nanjing
2 Department of Medicine, Jiangsu Institute of Cancer Research, Nanjing
3 The People's Hospital of Sihong, Teaching Hospital of Xuzhou Medical College, Sihong
4 Research Center, The Second Affiliated Hospital of Nanjing Medical University, Nanjing, People's Republic of China
For reprints and all correspondence: Yayi Hou, Immunology and Reproductive Biology Laboratory, Medical School and State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Nanjing 210093, People's Republic of China. E-mail: yayihounju{at}yahoo.com.cn, yayihou{at}nju.edu.cn
Received August 19, 2007; accepted January 9, 2008
 |
Abstract |
|---|
 TOP Abstract INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONReference |
|---|
The incidence of lung cancer has been increasing over recent decades. Previous studies showed that polymorphisms of the genes involved in carcinogen-detoxication, DNA repair and cell cycle control comprise risk factors for lung cancer. Recent observations revealed that the growth hormone receptor (GHR) might play important roles in carcinogenesis and Rudd et al. found that the Thr495Pro polymorphism of GHR was strongly associated with lung cancer risk in Caucasians living in the UK (OR = 12.98, P = 0.0019, 95% CI: 1.77–
). To test whether this variant of GHR would modify the risk of lung cancer in Chinese population, we compared the polymorphism between 778 lung cancer patients and 781 healthy control subjects. Our results indicate that the frequency of 495Thr (2.8%) allele in cases was significantly higher than in controls (OR = 2.04, P = 0.006, 95% CI: 1.21–3.42) which indicated this allele might be a risk factor for lung cancer. Further analyses revealed Thr495Pro variant was associated with lung cancer in the subpopulation with higher risk for lung cancer: male subpopulation, still-smokers subpopulation and the subpopulation with familial history of cancer. In different histological types of lung cancer, Thr495Pro SNP was significantly associated with small cell and squamous cell lung cancer, but not with adenocarcinoma, which suggested a potential interaction between this polymorphism and metabolic pathways related to smoking. The potential gene–environment interaction on lung cancer risk was evaluated using MDR software. A significant redundant interaction between Thr495Pro polymorphism and smoking dose and familial history of cancer was identified and the combination of genetic factors and smoking status or familial history of cancer barely increased the cancer risk prediction accuracy. In conclusion, our results suggested that the Thr495Pro polymorphism of GHR was associated with the risk of lung cancer in a redundant interaction with smoking and familial history of cancer.
Key Words: GHR • Thr495Pro polymorphism • lung cancer • MDR • molecular epidemiology
|
INTRODUCTION |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONReference |
|---|
Lung cancer is one of the most important common diseases with complicate, multifactorial etiology, including interactions between genetic makeup and environmental factors. It has been postulated for long that individuals may differ in their susceptibility to environmental risk factors, which is supported by increasing evidence from molecular epidemiological studies on genetic risk (1). This difference of susceptibility may result from inherited polymorphisms in various genes controlling carcinogen metabolism, repair of DNA damage and cell cycle etc (2–4).
Recently, several observations revealed growth hormone-insulin like growth factor (GH–IGF) axis might be associated with development of cancer (5). Besides its action as a pituitary hormone, growth hormone 1 (GH1) is also locally produced in mammary epithelial cells where it can act in an autocrine/paracrine manner (6). Recent data showed that autocrine GH1 production resulted in hyperproliferation of mammary carcinoma cells and enhanced transcriptional activation mediated by growth hormone receptor (GHR) (7). Insulin like growth factor-1 (IGF1), which is up-regulated by GH, regulates cellular proliferation and apoptosis and has been shown to increase tumor growth (8). Elevated levels of circulating IGF-1 have been shown to confer an increased risk to various tumors including breast (9), colorectal (10), lung (11) and prostate cancers (12). Therefore, GH1 and its main effector, IGF-1, are mitogens, which combined with GHR not only stimulate the normal mammary development but also play an important role in the development of cancer. Because of the importance, the relations between GHR gene and the development of cancers were investigated by several groups (11,13,14). The human GHR consists of nine coding exons out of which exons 3–7 encode the extracellular domain. Four cSNPs including Gly186Gly, Cys440Phe, Thr495Pro and Ile544Leu and a deletion polymorphism (GHRd3) lacking 22 residues encoded by exon 3 have been reported (15,16). However, the studies have not provided consistent evidence for the association between the polymorphisms and cancer risk. Wagner et al. did not observe any effect of the four polymorphisms (Gly186Gly, Cys440Phe, Ile544Leu and GHRd3) on breast cancer risk (11), while Rudd et al. found that Thr495Pro was strongly associated with lung cancer risk (13). Rudd et al. investigated 64 associated SNPs mapping to genes encoding pivotal components of the GH-IGF1 pathway including three non-synonymous cSNPs in GHR. Their result suggested that 11 of the 64 SNPs contributed to lung cancer susceptibility. Among the three non-synonymous cSNPs of GHR, Cys440Phe and Ile544Leu were not related with cancer risk, while people having Thr allele of Thr495Pro had a 12-fold increased risk for lung caner in Caucasians (13).
In this study, we investigated the GHR Thr495Pro polymorphism for an association with lung cancer risk in Chinese population. The results indicated Thr495Pro polymorphism was strongly associated with lung cancer risk in Chinese population.
|
PATIENTS AND METHODS |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONReference |
|---|
Patient Recruitment and Sample Collection
This study included 778 lung cancer patients and 781 cancer-free control subjects. All subjects were genetically unrelated ethnic Han Chinese and were from Nanjing City and its surrounding regions in southeast China. Patients with histopathologically confirmed incident lung cancer were recruited between July 2002 and November 2004. Cancer-free control subjects were selected from a pool of healthy volunteers during the same period. These subjects had no history of cancer and were frequency-matched to the cases on age (±5 years), sex and residential area (urban or rural areas). Each participant was scheduled for an interview after written informed consent was obtained, and a structured questionnaire was administered by interviewers to collect information on demographic data and environmental exposure history including tobacco smoking. Those who had smoked less than one cigarette per day and shorter than 1 year in their lifetime were defined as non-smokers. Otherwise, they were considered as ever-smokers. Those smokers who quit for more than 1 year were considered as former-smokers. Pack-years smoked [(cigarettes per day/20) x years smoked] were calculated to indicate the cumulative smoking dose. Family history of cancer was defined as any self-reported cancer in first-degree relatives (parents, siblings or children). After interview,
5-ml venous blood sample was collected from each participant.
Genotyping Assays
GHR Thr495Pro polymorphism was genotyped using MALDI-TOF mass spectrometry of primer extension products (Sequenom iPlexsystem, Hamburg, Germany). Amplification primers, including tag sequences were as follows: 5'-TCGACTTTTATGCCCAGGTG-3' and 5'-CTGCCTTATTCTTTTGGCCC-3'; Extension primer was as follow: 5'-CTTACCACACTACCTGCTG-3'; final volume of 5 µl according to manufacturer's conditions. Typing of 10% of the samples was repeated blind to confirm assay fidelity.
Statistical Analysis
Chi-square tests were used to examine deviation of genotype frequencies from the Hardy–Weinberg equilibrium (HWE) among controls, as well as to test differences in the distributions of genotypes between cases and controls in dominant, additive and recessive genetic model.
With the statistical software of spss11.0, adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated as a measure of association with the risk of lung cancer by using unconditional logistic regression adjusted for age, sex, smoking status and familial history of cancer (17,18). All statistical tests were two-sided and the significance level P < 0.05 was used for all tests.
Analysis of Gene–Environment Interaction
Gene–environment interactions were evaluated by using the non-parametric multifactor dimensionality reduction (MDR) software (19–21) with Thr495Pro polymorphism, gender, age, trichotomized cumulative smoking dose and family history of cancer. With MDR, genotype and environmental factors were pooled into high- and low-risk groups, effectively reducing the multifactor predictors from n dimensions to 1 dimension. The new one-dimensional multifactor variable was evaluated for its ability to classify and predict disease status through cross-validation and permutation testing.
Four general steps were involved in implementing the MDR method for case–control studies. In Step 1, a set of n genetic and/or discrete environmental factors was selected from the pool of all factors. In Step 2, the n factors and their possible multifactor cells were represented in n-dimensional space. Then, the ratio of the number of cases to the number of controls was estimated within each multifactor class. In Step 3, each multifactor cell in n-dimensional space was labeled either as ‘high-risk’, if the cases: controls ratio 
1, or as ‘low-risk’. In this way, a model for both cases and controls was formed by pooling high-risk cells into one group and low-risk cells into the other group. This reduced the n-dimensional model to a one-dimensional model. In Step 4, the prediction error of each model was estimated by 100-fold cross-validation. This was used to reduce the chance of type I error due to multiple testing (22). Here, the data were randomly divided into 100 equal parts. The MDR model was developed for each possible 99/100 (training set) of the subjects and then was used to make predictions about the disease status of each possible 1/100 (testing set) of the subjects excluded. The proportion of subjects for whom an incorrect prediction had been made was an estimation of the prediction error. To reduce the possibility of poor estimates of the prediction error that were due to chance divisions of the data set, the 100-fold cross-validation was repeated 100 times, and the prediction errors were averaged. The four steps of the MDR method were repeated for each possible combination. Among this set of best multifactor models, the combination of loci and/or discrete environmental factors that minimized the prediction error was selected.
Further, the fitness or value of an MDR model was assessed by estimating accuracy in the training set and the testing set. Accuracy is a function of the percentage of true positives (TP), true negatives (TN), false positives (FP) and false negatives (FN) and is defined as (TP + TN)/(TP + TN + FP + FN). This process is repeated for all 100 pieces of the data and the 100 testing accuracies were averaged to provide an estimate of predictive ability or generalization (21). Finally, all the variables in the best model were combined and dichotomized according to the MDR software and their ORs and 95% CIs in relation to lung cancer risk were calculated in logistic regression models.
|
RESULTS |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONReference |
|---|
We investigated the GHR Thr495Pro polymorphism in a cohort of 778 Chinese lung cancer cases and 781 matched controls. The 778 lung cancer cases and 781controls appeared to be adequately matched on age and sex. The distributions of age, gender, sex, smoking status and family history characterization among the subjects were summarized in Table 1. There were no significant differences between case and control subjects in terms of distributions of age, gender and sex. Compared with control subjects, the cases were more likely to be smokers (OR = 2.04, P < 0.001, 95% CI: 1.66–2.51) (ever-smokers: 69.4% in cases and 52.6% in controls) and cigarette smoking was associated with a 2.30-fold (OR = 2.30, P < 0.001, 95% CI: 1.72–3.09) increased risk for lung cancer among the former-smokers and a 1.95-fold (OR = 1.95, P < 0.001, 95% CI: 1.56–2.43) increased risk among still-smokers. The risk also depended on the dosage of smoke. The heavy-smokers (OR = 2.82, P < 0.001, 95% CI: 2.21–3.59) had much higher risk than light-smokers (OR = 1.40, P = 0.009, 95% CI: 1.09–1.79) had. Further, a family history of cancer in first-degree relatives results in a 1.60-fold (OR = 1.60, P < 0.001, 95% CI: 1.26–2.03) increased risk for lung caner. Of the 778 lung cancer cases, 736 (94.6%) were classified as non-small cell lung cancer (274 adenocarcinoma, 202 squamous cell carcinoma, and 260 large cell, mixed cell carcinomas or undifferentiated carcinoma), and only 42 (5.4%) as small cell lung cancer. Further analysis revealed that there were 197 smokers (69 light-smokers and 128 heavy-smokers) in small cell and squamous cell lung cancer patients, accounting for 80.7%, much higher than that of adenocarcinoma patients (62.8%, including 69 light-smokers and 93 heavy-smokers).
|
A base substitution from C to A on 1526nt of cDNA of GHR resulted an amino acid changes from Pro to Thr. Genotyping of the Thr495Pro polymorphism was performed by using the MALDI-TOF mass spectrometry of primer extension products (Sequenom iPlexsystem, Hamburg, Germany). Out of 781controls and 778 lung cancer patients, 778 controls and 774 patients were successfully analysed. Genotyping rate was 99.6%. In controls, 756 (97.2%) were homozygous for the C allele (CC), 22(2.8%) were heterozygous (CA) and none homozygous for the A allele (AA) was found, hence the frequency of C allele was 98.6%. The Hardy–Weinberg equilibrium was first tested. The genotype distributions in the control group did not deviate from the Hardy–Weinberg equilibrium. In cases, 731 (94.4%) were homozygous for the C allele (CC), 42(5.4%) were heterozygous (CA) and 1 was homozygous for the A allele (AA). The frequency of A (2.8%) allele in cases was significantly higher than in controls (OR = 2.04, P = 0.006, 95% CI: 1.21–3.42) which indicated the A allele might be a risk factor for lung cancer.
The genotype distribution and allele frequencies of Thr495Pro for the 774 lung cancer patients and 778 control subjects were summarized in Table 2. This table also shows lung cancer risk related to them. When the overall lung cancer cases were compared with the controls, the distributions of the genotype were tested. The genotype frequency of cases was found significantly differed from that of the controls. The people with heterozygous and homozygous for the A allele at Thr495Pro polymorphism had higher risk for lung cancer (OR = 2.02, P = 0.007, 95% CI: 1.20–3.41).
|
Further, we compared Thr495Pro allelic and genotype distribution between cases and controls with stratification by age, sex, smoking dose and status, familial cancer history and histological type of cancer. A allele of Thr495Pro variant was found significantly associated with lung cancer risk in several subpopulations. In the male population, A allele (OR = 2.18, P = 0.007, 95% CI: 1.22–3.92) and CA genotype (OR = 2.09, P = 0.013, 95% CI: 1.15–3.80) significantly increased the risk of lung cancer. A allele (OR = 3.42, P = 0.009, 95% CI: 1.28–9.13) and CA genotype (OR = 3.15, P = 0.018, 95% CI: 1.16–8.54) also increased the susceptibility in still-smoker population. And in those who with familial history of cancer in first relations, A allele (OR = 3.04, P = 0.037, 95% CI: 1.01–9.14) and CA genotype (OR = 3.13, P = 0.035, 95% CI: 1.03–9.51) were associated with the increase of the risk for lung cancer. But there were no differences found in other subpopulation. In different histological types of lung cancer, Thr495Pro polymorphism seemed to have different effects. A allele (OR = 2.52, P = 0.007, 95% CI: 1.33–4.80) and CA genotype (OR = 2.59, P = 0.006, 95% CI: 1.35–4.95) significantly increased the risk of small cell and squamous cell lung cancer, but the effect on adenocarcinoma was not definite.
We then analysed the potential gene–environmental interactions on lung cancer risk using MDR software. Analysis was performed with the Thr495Pro variant, age, gender, the trichotomized cumulative smoking dose and the family history of cancer. After the variables in the best model combined and dichotomized according to the MDR software, smoking dose was found to be the strongest single-factor with accuracy 0.60 (OR = 2.4849, P < 0.0001, 95% CI: 1.9987–3.0895), which was more accurate than genetic single-factor (accuracy = 0.51, OR = 2.0214, P = 0.0073, 95% CI: 1.1973–3.4127). However, two factor model including smoking dose and Thr495Pro polymorphism only increased accuracy and OR a little (accuracy = 0.60, OR = 2.5147, P < 0.0001, 95% CI: 2.0315–3.1128), which indicated that these two factors provided redundant information. All other models combined Thr495Pro variants with any other factors were listed in Table 3. The strongest lung cancer prediction model consisted of all five factors, which resulting in an OR 3.6222 with a P value under 0.0001. (OR = 3.6222, P < 0.0001, 95% CI: 2.9266–4.4831)
|
Figure 1 was a graphical representation of the interactions between five attributes (Thr495Pro variant, age, gender, the trichotomized cumulative smoking dose and the family history of cancer) from the MDR analysis using an ‘interaction dendrogram’. The result indicated that the interaction between age and gender was in a synergetic manner, whereas the interaction of Thr495Pro variant, smoking status and family history of cancer was in a redundant manner.
|
|
DISCUSSION |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONReference |
|---|
The growth hormone receptor gene (GHR) belongs to the super family of cytokine receptors. GHR protein is potentially involved in cancer development via its role in initiating GH-dependent signal transduction pathways and stimulating insulin-like growth factor I production. Growth hormone and IGF-I were thought to play important roles in carcinogenesis (5).
In 2007, the study by Rudd et al. provided evidence that inherited predisposition to lung cancer was in part mediated through low-penetrance alleles and specifically identifies variants in genes comprising the GH-IGF pathway as susceptibility alleles. In their study, GHR Thr495Pro polymorphism was strongly associated with lung cancer risk (OR = 12.98, P = 0.0019, 95% CI: 1.77–) (13). In our study, we confirmed the result about the relation between GHR Thr495Pro polymorphism and lung cancer in Chinese population. Compared with British Caucasians, Chinese had a higher MAF (minor allele frequency) at GHR Thr495Pro locus. The A allele frequency was 1.4% in Chinese controls, whereas the frequency was only 0.1% in British Caucasians. However, the OR for lung cancer risk in Caucasians was much bigger than in Chinese (OR = 2.04, P = 0.006, 95% CI: 1.21–3.42).
The relations between GHR variants and cancer risk were studied by several groups in recent years (11,13,14). After the report by Rudd et al., Wagner et al. investigated the polymorphisms exon3deletion, Gly186Gly, Cys440Phe and Ile544Leu for their contribution to the risk of breast cancer (11). However, they did not observe any effect of the GHR polymorphisms on breast cancer risk in their study. In 2007, Mckey et al. performed a haplotype based analysis with the SNPs of GHR in prostate cancer. Their results suggested that whereas genetic variation in the GHR gene did not seem to play a major role in prostate cancer etiology, one haplotype in the 3' region might be potentially relevant to cases with later onset of prostate cancer (14). GHR is located on chromosome 5p13.1-p12 and is 87 Kb long, with nine coding exons encoding 620 amino acids. Exon3deletion polymorphism is due to a deletion between two retroviral sequences flanking exon 3 that mimic alternative splicing (15,16). It has been hypothesized that a loss or retention of exon 3 could affect the receptor function. However, the binding affinity of GH1 to each receptor isoform was similar and there was no evidence that the d3 variant was associated with the cancer risk (23). The Gly186Gly SNP is, according to the UniProtKB/Swiss-Prot database [Swiss-Prot: P10912 [GenBank] ], located at a turn position in the protein structure but the SNP does not change the amino acid. The Cys440Phe, Thr495Pro and Pro579Thr SNP are both located in the 3' region of the gene. Both of these amino acid substitutions were predicted to be possibly damaging according to the PolyPhen database (http://tux.embl-heidelberg.de/ramensky/polyphen.cgi). In hapmap database, these three SNPs were in one LD block, which covers all the 3' region of GHR. However, the MAF of these variants was low.
Tobacco smoking has been acknowledged to be the major risk factor for lung cancer, contributing to a 10-fold increase in risk in long-term smokers compared with non-smokers (24). In our study, heavy-smokers got a 2.82-fold (OR = 2.82, P < 0.001, 95% CI: 2.21–3.59) increased risk for lung cancer. Further analysis indicated that the smoking had different effects on different histological types of lung cancer. Consistent with former epidemiological studies (25,26), we found that small cell and squamous cell types of lung cancer were highly associated with habitual smoking, whereas adenocarcinoma was weakly associated. Familial history of cancer was also a risk factor for lung cancer.
We found that the A allele of Thr495Pro variant of GHR was a heretic risk factor in Chinese. Further, our results indicated Thr495Pro variant was associated with lung cancer in male, still-smokers and with familial history of cancer subpopulation, which all with higher risk for lung cancer. In the analysis in different histological types of cancer, Thr495Pro variant was found to influence the risk of small cell and squamous cell types of lung cancer, which was believed to be more associated with smoking. The result suggested that the Thr495Pro SNP might have interaction with smoking carcinogensis process in lung cancer. Therefore, it was meaningful to analyse the genotype of Thr495Pro polymorphism in subpopulation with higher lung cancer risk in order to prevent the disease more efficiently.
GHR was involved in the GH/IGF axis signal pathway, which was not sensitive to the environmental factor including smoking according nowadays knowledge. However, in our gene–environmental analysis by MDR software, it was revealed that there had redundant interaction between Thr495Pro polymorphism and smoking dose, also with familial history of cancer. We also performed the analysis of the interaction by the method of stratification. As shown in Fig. 2, because of the limit of the sample numbers and the low MAF of the polymorphism, several cells had too small amount of samples to evaluate the relation of the three risk factors and lung cancer. Comparing to using the conventional method with the nonparametric and genetic model-free MDR approach, multilocus genotypes/diplotypes and environmental factors can be evaluated together to fit the best prediction model and further be pooled into high- and low-risk groups, effectively reducing the dimensionality from n dimensions to one dimension (22). Figure 2 represented the example of the analysis of the combination of Thr495Pro variants, trichotomized cumulative smoking dose and the family history of cancer. High- and low-risk groups were generated by MDR and the OR value was 2.58 (95% CI: 2.08–3.20), much smaller than the product of the OR values of each risk factors (8.07 = 2.02 x 1.61 x 2.48), which indicated the combination of the three factors provided redundant information. The results suggested that consistently with the former analysis in small cell and squamous cell types of lung cancer, there might be a potential interaction between the GH/IGF pathway and smoking metabolism pathway, which need more researches to discover in larger study population. The redundant information between Thr495Pro and familial history of cancer was easy to understand because Thr495Pro polymorphism might be an important basis of the familial hereditary risk for lung cancer.
|
|
CONCLUSION |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONReference |
|---|
In this study, we investigated the GHR Thr495Pro polymorphism for an association with lung cancer risk in Chinese population. The results suggested that Thr495Pro polymorphism of GHR might modulate the risk of lung cancer in a redundant interaction with smoking and familial history of cancer. Further researches are necessary to confirm the result and to investigate the relation between GHR and smoking.
Conflict of interest statement None declared.
|
Reference |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONCONCLUSIONReference |
|---|
1 Lee GY, Jang JS, Lee SY, Jeon HS, Kim KM, Choi JE, et al. XPC polymorphisms and lung cancer risk. Int J Cancer (2005) 115:807–13.[CrossRef][Web of Science][Medline]
2 Shields PG, Harris CC. Cancer risk and low-penetrance susceptibility genes in gene–environment interactions. J Clin Oncol (2000) 18:2309–15.
3 Mohrenweiser HW, Jones IM. Variation in DNA repair is a factor in cancer susceptibility: a paradigm for the promise and perils of individuals and population risk estimation? Mutat Res (1998) 400:15–24.[Web of Science][Medline]
4 Goode EL, Ulrich CM, Potter JD. Polymorphisms in DNA repair genes and associations with cancer risk. Cancer Epidemiol Biomarkers Prev (2002) 11:1513–30.
5 Wang X, He K, Gerhart M, Jiang J, Paxton RJ, Menon RK, et al. Reduced proteolysis of rabbit growth hormone (GH) receptor substituted with mouse GH receptor cleavage site. Mol Endocrinol (2003) 17(10):1931–43.
6 Wagner K, Hemminki K, Grzybowska E, Bermejo JL, Butkiewicz D, Pamula J, et al. Polymorphisms in the growth hormone receptor: a case-control study in breast cancer. Int J Cancer (2006) 118(11):2903–6.[CrossRef][Medline]
7 Kaulsay KK, Zhu T, Bennett W, Lee KO, Lobie PE. The effects of autocrine human growth hormone (hGH) on human mammary carcinoma cell behavior are mediated via the hGH receptor. Endocrinology (2001) 142:767–77.
8 Khandwala HM, McCutcheon IE, Flyvbjerg A, Friend KE. The effects of insulin-like growth factors on tumorigenesis and neoplastic growth. Endocr Rev (2000) 21:215–44.
9 Toniolo P, Bruning PF, Akhmedkhanov A, Bonfrer JM, Koenig KL, Lukanova A, et al. Serum insulin-like growth factor-I and breast cancer. Int J Cancer (2000) 88:828–32.[CrossRef][Web of Science][Medline]
10 Ma J, Pollak MN, Giovannucci E, Chan JM, Tao Y, Hennekens CH, et al. Prospective study of colorectal cancer risk in men and plasma levels of insulin-like growth factor (IGF)-I and IGF-binding protein-3. J Natl Cancer Inst (1999) 91:620–5.
11 Yu H, Spitz MR, Mistry J, Gu J, Hong WK, Wu X. Plasma levels of insulin-like growth factor-I and lung cancer risk: a case-control analysis. J Natl Cancer Inst (1999) 91:151–6.
12 Chan JM, Stampfer MJ, Giovannucci E, Gann PH, Ma J, Wilkinson P, et al. Plasma insulin-like growth factor-I and prostate cancer risk: A prospective study. Science (1998) 279:563–6.
13 Rudd MF, Webb EL, Matakidou A, Sellick GS, Williams RD, Bridle H, GELCAPS Consortium, et al. Variants in the GH-IGF axis confer susceptibility to lung cancer. Genome Res (2006) 16(6):693–701.
14 McKay JD, Kaaks R, Johansson M, Biessy C, Wiklund F, Bälter K, et al. Haplotype-based analysis of common variation in the growth hormone receptor gene and prostate cancer risk. Cancer Epidemiol Biomarkers Prev (2007) 16(1):169–73.
15 Pantel J, Machinis K, Sobrier ML, Duquesnoy P, Goossens M, Amselem S. Species-specific alternative splice mimicry at the growth hormone receptor locus revealed by the lineage of retroelements during primate evolution. J Biol Chem (2000) 275:18664–9.
16 Stallings-Mann ML, Ludwiczak RL, Klinger KW, Rottman F. Alternative splicing of exon 3 of the human growth hormone receptor is the result of an unusual genetic polymorphism. Proc Natl Acad Sci USA (1996) 93:12394–9.
17 Sham PC, Curtis D. Monte Carlo tests for associations between disease and alleles at highly polymorphic loci. Ann Hum Genet (1995) 59:97–105.[Web of Science][Medline]
18 Wallenstein S, Hodge SE, Weston A. Logistic regression model for analyzing extended haplotype data. Genet Epidemiol (1998) 15:173–81.[CrossRef][Web of Science][Medline]
19 Ritchie MD, Hahn LW, Roodi N, Bailey LR, Dupont WD, Parl FF, et al. Multifactor-dimensionality reduction reveals high-order interactions among estrogen-metabolism genes in sporadic breast cancer. Am J Hum Genet (2001) 69(1):138–47.[CrossRef][Web of Science][Medline]
20 Ritchie MD, Hahn LW, Moore JH. Power of multifactor dimensionality reduction for detecting gene-gene interactions in the presence of genotyping error, missing data, phenocopy, and genetic heterogeneity. Genet Epidemiol (2003) 24(2):150–7.[CrossRef][Web of Science][Medline]
21 Hahn LW, Ritchie MD, Moore JH. Multifactor dimensionality reduction software for detecting gene-gene and gene-environment interactions. Bioinformatics (2003) 19(3):376–82.
22 Coffey CS, Hebert PR, Ritchie MD, Krumholz HM, Gaziano JM, Ridker PM, et al. An application of conditional logistic regression and multifactor dimensionality reduction for detecting gene-gene interactions on risk of myocardial infarction: the importance of model validation. BMC Bioinform (2004) 5:49.[CrossRef][Medline]
23 Dos Santos C, Essioux L, Teinturier C, Tauber M, Goffin V, Bougneres P. A common polymorphism of the growth hormone receptor is associated with increased responsiveness to growth hormone. Nat Genet (2004) 36:720–4.[CrossRef][Web of Science][Medline]
24 Doll R, Peto R. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst (1981) 66(6):1191–308.[Web of Science][Medline]
25 Morabia A, Wynder EL. Cigarette smoking and lung cancer cell types. Cancer (1991) 68(9):2074–8.[CrossRef][Web of Science][Medline]
26 Schoenberg JB, Wilcox HB, Mason TJ, Bill J, Stemhagen A. Variation in smoking-related lung cancer risk among New Jersey women. Am J Epidemiol (1989) 130(4):688–95.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||