http://orcid.org/0000-0002-2301-2392
Figures
Abstract
Tumor necrosis factor-α (TNF-α), an important factor in systematic inflammation, is reportedly involved in several cancer types. The TNF-α -308 G>A (rs1800629) polymorphism in the promoter region influences TNF-α production. The association between TNF-α -308 G>A polymorphism and colorectal cancer (CRC) is not fully understood, especially the connections between TNF-α -308 G>A polymorphism and clinical features of CRC. In this study, TNF-α -308 G>A polymorphism was genotyped in 1140 individuals with or without CRC from Southwestern China. In case-control studies, we found no association between TNF-α -308 G>A polymorphism and CRC risk. Analysis of the correlations between TNF-α -308 G>A polymorphism and clinical features of CRC revealed that TNF-α -308 A allele was associated with higher body mass index (BMI) larger tumor size, and distant tumor metastasis in all CRC patients. Notably, rectal cancer (a subtype of CRC) patients with TNF-α -308 A allele had a very high risk of distant tumor metastasis [odds ratio (OR) = 4.481; 95% confidence interval (CI): 2.072–9.693; P = 0.00025]. The association between TNF-α -308 A allele and distant tumor metastasis remained even significant after adjusting all clinical characteristics (OR = 7.099; 95% CI: 2.482–20.301; P = 0.000256) in rectal cancer patients. Our results suggested that TNF-α -308 A allele was significantly associated with distant tumor metastasis in rectal cancer patients.
Citation: Li Z, Li S-a, Sun Y, Liu Y, Li W-l, Yang L, et al. (2017) TNF-α -308 A allele is associated with an increased risk of distant metastasis in rectal cancer patients from Southwestern China. PLoS ONE 12(6): e0178218. https://doi.org/10.1371/journal.pone.0178218
Editor: Aamir Ahmad, University of South Alabama Mitchell Cancer Institute, UNITED STATES
Received: February 12, 2017; Accepted: May 9, 2017; Published: June 2, 2017
Copyright: © 2017 Li et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper.
Funding: This study was supported by the Foundation of Leading Talent Program of Health and Family Planning Commission of Yunnan Province (No. L-201205), National Natural Science Foundation of China (No. 81460424), the Scholarship Award for Excellent Doctoral Student Grant of Yunnan Province, the Foundation of Medical Discipline Leaders Program of Health and Family Planning Commission of Yunnan Province (No. D-201668), the Foundation of the Institute of Gastroenterology, Research Institutions attached to the Health and Family Planning Commission of Yunnan Province (2014NS122), the Foundation of the Institute of Breast Surgery, Research Institutions attached to the Health and Family Planning Commission of Yunnan Province (2014NS022) and the Joint Special Research Funds of Kunming Medical University (2014FZ035). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Colorectal cancer (CRC), one of the most prevalent malignant tumors, is the third most commonly diagnosed cancer and the fourth cause of cancer deaths worldwide [1]. The mortality and morbidity of CRC rapidly grows in the Chinese population. According to the latest Cancer Statistics of China [2], CRC is the fifth newly diagnosed cancer among men, and the fourth among women in China. CRC is considered to be a highly heterogeneous and polygenic disease[3, 4]. The tumorigenesis of CRC is still not completely clear and is thought to be a multistep event. Environmental factors and genetic susceptibility in different cohorts are accepted to play important roles in the tumorigenesis of CRC[5–7].
Distant metastasis adversely influences the survival of CRC patients [8] and significantly affects the treatment strategies of CRC. Inflammation and immune response play critical roles in CRC progression, including tumor metastasis and poor prognosis. The involvement of the immune system in CRC progression CRC is currently a research hotspot [9, 10]. Tumor necrosis factor-α (TNF-α) is an important pleiotropic cytokine produced by macrophages and is actively involved in immune regulations. TNF-α was thought to be associated with various malignant diseases, including cancer [11, 12]. The levels of elevated plasma TNF-α are commonly associated with poor prognosis [7, 13], such as tumor recurrence and positive lymph-node stage [14, 15]. Meanwhile, TNF-α can induce epithelial mesenchymal transformation and subsequently promotes the invasion and metastasis of colorectal cancer [16, 17]. Single nucleotide polymorphisms in TNF-α gene is reportedly associated with the cancer diagnosis, treatment outcome, and survival of cancer patients [18, 19]. For instance, TNF-α -308 polymorphism is significantly associated with the risk of gastric and hepatocellular carcinomas, triple-negative breast cancer, myeloma, and lymphoma [20–24], suggesting that TNF-α may be conducive to the early diagnosis of malignancies.
Some studies have focused on the association between TNF-α -308 polymorphism and CRC risk in different ethnicities and geographical regions worldwide [7, 24, 25]. However, results are inconsistent. Notably, research has been performed on associations between TNF-α -308 polymorphism and clinical features of CRC, especially in highly aggressive CRC with distant metastasis. In the present work, we conducted a case-control study to explore the role of TNF-α -308 polymorphism in CRC susceptibility, as well as the relationship between genotypes and clinicopathological characteristics.
Patients and methods
Ethics statement
Our study protocol was approved by the Institutional Review Board of Kunming Medical University. All study subjects provided written informed consent to participate in the study. The study conformed to the tenets of the Declaration of Helsinki. All methods applied were carried out in accordance with the approved guidelines.
Study subjects
This case-control study included 570 patients with constitutive CRC and 570 cancer-free controls. All subjects were from southwestern China. Patients were recruited from January 2010 to December 2015 in the First Affiliated Hospital of Kunming Medical University & Yunnan Tumor Hospital and had been diagnosed with histologically confirmed CRC. We classified the BC subtypes as colon and rectal cancer. Controls were selected based on a physical examination in the same region during the same period as patient recruitment. The selection criteria included no history of cancer and frequency matching to cases by age and gender. At recruitment, demographic information and clinical characteristics of each participant were collected.
Genotyping
DNA was extracted from leukocytes using a Wizard® Genomic DNA Purification Kit (Promega Corporation, FL, USA) according to the manufacturer’s instructions. Genotyping of TNF-α -308 polymorphism (rs1800629) was performed using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Primer sequences are available upon request.
Statistical analysis
Χ2 tests was used to examine the deviation in genotype frequencies from Hardy–Weinberg equilibrium (HWE) in controls and was also used to analyze differences in demographic variables and genotype distributions for different clinical features of CRC. The association of TNF-α- 308 polymorphism with CRC incidence was calculated using adjusted odds ratios (AORs) and 95% confidence intervals (CIs) from multivariate logistic regression. The analysis was performed using SPSS 17.0 (SPSSInc., Chicago, IL, USA). P value less than 0.05 was considered statistically significant.
Results
Cases and controls
We had a total of 570 CRC cases, in which 250 were colon cancer, 316 were rectum cancer, and 4 were both with colon and rectal cancer (Table 1). The demographic characteristics of cases and controls are presented in Table 1.
In this case-control study, the distributions of age and gender were consistently investigated among all cases and controls (P = 0.46 and 0.77, respectively). The genotype frequency of TNF-α-308 G and A alleles in the controls and cases concorded with HWE (P = 0.85 and 0.67, respectively). Considering that CRC is a complicated disease, many risk factors (such as age) are associated with the occurrence and progression of CRC [26, 27]. Therefore, we conducted stratified analysis on the data according to age, gender, tumor site, tumor size, tumor node metastasis (TNM) stage, presence of hypertension, diabetes mellitus, and levels of body Therefore, we conducted a stratified analysis on data according to age, gender, tumor site, tumor size, tumor-node-metastasis (TNM) stage, presence of hypertension, diabetes mellitus, and levels of body mass index (BMI) to determine whether the TNF-α-308 G>A polymorphism was associated with CRC incidence in specific subsets of the study population.
TNF-α -308 G>A polymorphism and CRC
The frequency of TNF-α -308 A allele in CRC patients and relevant normal individuals was 0.070 and 0.067, respectively (Table 2). No significant difference in TNF-α -308 A and G alleles was observed between CRC cases and controls.
Considering the few number of TNF-α -308 AA genotype, the dominant genetic model (GG/GA+AA) was used to analyze the genotype distribution of TNF-α -308 polymorphism. In the dominant genetic model, no significant (Table 3) difference between CRC cases and controls was found in TNF-α -308 polymorphism (OR = 0.884; 95% CI: 0.624–1.251; P = 0.535). TNF-α -308 polymorphism was not associated with the risk of colon cancer (OR = 0.895; 95% CI: 0.586–1.367; P = 0.661) or rectal cancer (OR = 1.386; 95% CI: 0.895–2.147; P = 0.165).
TNF-α -308 G>A polymorphism and clinical risk of CRC
To understand the association between TNF-α -308 G>A polymorphism and clinical features of CRC, logistic regression analysis was used to evaluate the correlations between TNF-α-308 G>A polymorphism and clinical characteristics (gender, age at diagnosis, BMI, tumor stage, tumor size, lymph-node metastasis, distant metastasis, history of polyp, and family history of cancer by CRC tumor-site classification), as shown in Table 4. In all CRC patients, the GA and AA genotypes were associated with higher BMI index (OR = 2.056; 95% CI: 1.197–3.530; P = 0.01) and larger tumor size (OR = 1.81; 95% CI: 1.029–3.183; P = 0.042). Notably, TNF-α -308 GA and AA genotypes showed strong association with distant tumor metastasis (OR = 2.911; 95% CI: 1.697–4.992; P = 0.00017). Furthermore, rectal cancer patients with TNF-α -308 A allele had a very high risk of distant tumor metastasis (OR = 4.481; 95% CI: 2.072–9.693; P = 0.00025), as shown in Table 4. After adjusting tumor stage, pathologic status, tumor size, and lymph-node metastasis status, the association became more significant (OR = 5.549; 95% CI: 2.437–12.632; P = 0.000045). The association between TNF-α -308 A allele and distant tumor metastasis remained even significant after adjusting all clinical characteristics (OR = 7.099; 95% CI: 2.482–20.301; P = 0.000256) in rectal cancer patients. By contrast, no association between TNF-α -308 GA and AA genotypes and distant metastasis was observed in colon-cancer patients (OR = 1.685; 95% CI: 0.592–4.796; P = 0.328). TNF-α -308 G>A was apparently not associated with other clinical features of CRC in our study.
The total number of individuals may not be the same because of censored data.
TNF-α -308 G>A polymorphism and physicochemical characteristics of CRC
TNF-α is an important cytokine related to immune regulation and inflammation. Therefore, the correlation of TNF-α -308 polymorphism with inflammatory, physiological, and biochemical indices was analyzed in CRC patients. The levels of neutrophil, leukocyte, cholesterol, triglyceride, albumin, high-density lipoprotein cholesterol (HDL-c), and low-density lipoprotein cholesterol (LDL-c) were used to access the relation with TNF-α -308 polymorphism in CRCs patients (Table 5). In all CRC patients, GA and AA genotypes were more frequent in patients with neutrophil percentage exceeding 75% than those with lower neutrophil percentage (OR = 4.764; 95% CI: 2.39–9.495; P = 0.000012). However, the association between TNF-α -308 G>A polymorphism and neutrophil percent disappeared in patients with rectal or colon cancer. TNF-α -308 G>A polymorphism was associated with the number of neutrophils, which may have resulted from inflammation caused by CRC. In addition to neutrophil percentage, other clinical feathers as listed in Table 5 were not associated with TNF-α -308 GA and AA genotypes.
The total number of individuals may not be the same because of censored data.
Discussion
TNF-α, an important inflammatory cytokine, is mainly produced by activated macrophages [28]. Recent studies have shown that TNF-α is actively involved in cellular apoptosis [29–31] and immune system[32–34]. TNF-α is also considered to be a very important connector between inflammation and cancer progression [35, 36]. These results suggested that TNF-α played a vital role in the etiology and progression of cancer. TNF-α -308 G>A polymorphism located in the promoter region reportedly influences TNF-α production [37]. Some studies have investigated the association between TNF-α -308 G>A polymorphism and CRC risk, but results are inconsistent [7]. Moreover, the association between TNF-α -308 G>A polymorphism and the clinical features of CRC is rarely studied. Therefore, from the point of view of population genetics, the role of TNF-α -308 G>A polymorphism in colorectal cancer requires further study.
To further understand the association between TNF-α -308 G>A polymorphism and CRC, 1140 individuals with or without CRC were enrolled in the present study. We found that TNF-α -308 G>A polymorphism showed no association with CRC risk in Han Chinese residing in Southwestern China, as shown in Tables 2 and 3. This result was generally consistent with meta-analysis results [32] indicating that TNF-α -308 G>A polymorphism does not confer susceptibility to CRC risk in Eastern populations. Notably, in Western populations, TNF-α -308 G>A polymorphism is thought to be a risk factor for CRC [32]. Therefore, our data further supported the notion that the genetic predisposition of TNF-α-308 G>A polymorphism to CRC was related to the substantial population heterogeneity.
The association between TNF-α -308 G>A polymorphism and the clinical features of CRC remains unclear. Recently, Li et al [24] suggested that the TNF-α-308 A allele is a risk factor of distant metastasis in patients with triple-negative breast cancer. This finding indicates that TNF-α -308 G>A polymorphism may be involved in cancer progression. In the present work, TNF-α -308 A allele was associated with higher BMI index, larger tumor size, and distant metastasis in CRC patients, as shown in Table 4. Our data showed that TNF-α -308 G>A polymorphism may be involved in cancer progression. Meanwhile, the correlation between TNF-α -308 A polymorphism and distant metastasis remained in rectal cancer patients (Table 4), indicating that TNF-α -308 G>A polymorphism may be an important risk factor affecting rectal-cancer patients. The potential role of TNF-α 308 A allele in CRC metastasis requires further verification through animal experiments.
TNF-α -308 G>A polymorphism was associated with the number of neutrophils in all CRC patients (Table 5), indicating that inflammation in CRC patients may be related to TNF-α -308 G>A polymorphism. However, the association between TNF-α -308 G>A polymorphism and the number of neutrophils cannot be further detected in CRC subtypes (rectal and colon cancers), as shown in Table 5. The association between TNF-α-308 G>A polymorphism and neutrophils may be unrelated to neoplasm location (colon or rectum), although further confirmation is necessary.
There are some limitations in this study. First, the sample size is modest in the present study. It could not fully detect the susceptibility of TNF-α -308 G>A polymorphism in each subtypes of CRC. Second, the results of the present study did not make a replicated study in another cohort of CRC patients. It should be further verified. Finally, the functional study was not performed to determine the role of TNF-α -308 A allele in distant metastasis of CRC subtypes.
Generally, in Han Chinese of southwestern China, TNF-α -308 G>A polymorphism may not predispose to CRC risk, but it may significantly increase risk of distant metastasis in rectal cancer patients.
Supporting information
S1 File. Genotype and clinical information for CRC patients in the present study.
https://doi.org/10.1371/journal.pone.0178218.s001
(ZIP)
Acknowledgments
We deeply appreciate Dr. Ceshi Chen, Dr. Rong Liu (Key Laboratory of Animal Models and Human Disease Mechanisms of Chinese Academy of Sciences & Yunnan Province, Kunming Institute of Zoology) and Dr. Chenggang Zou (Life science college, Yunnan University) for providing revision suggestions.
Author Contributions
- Conceptualization: KHW YLL TNZ.
- Data curation: SAL YLL.
- Formal analysis: ZL YS YL.
- Funding acquisition: ZL YS YL.
- Investigation: ZL YS.
- Methodology: YLL SAL ZL YS YL.
- Project administration: KHW SAL YLL.
- Resources: WLL LY YD HG.
- Writing – original draft: ZL.
- Writing – review & editing: ZL YLL JYL YL.
References
- 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA: a cancer journal for clinicians. 2016;66(1):7–30. pmid:26742998.
- 2. Chen W, Zheng R, Baade PD, Zhang S, Zeng H, Bray F, et al. Cancer statistics in China, 2015. CA: a cancer journal for clinicians. 2016;66(2):115–32. pmid:26808342.
- 3. Dowty JG, Win AK, Buchanan DD, Lindor NM, Macrae FA, Clendenning M, et al. Cancer risks for MLH1 and MSH2 mutation carriers. Human mutation. 2013;34(3):490–7. pmid:23255516.
- 4. Bredberg A. Cancer: more of polygenic disease and less of multiple mutations? A quantitative viewpoint. Cancer. 2011;117(3):440–5. pmid:20862743.
- 5. Favoriti P, Carbone G, Greco M, Pirozzi F, Pirozzi RE, Corcione F. Worldwide burden of colorectal cancer: a review. Updates Surg. 2016;68(1):7–11. pmid:27067591.
- 6. Negrini S, Gorgoulis VG, Halazonetis TD. Genomic instability—an evolving hallmark of cancer. Nature reviews Molecular cell biology. 2010;11(3):220–8. pmid:20177397.
- 7. Min L, Chen D, Qu L, Shou C. Tumor necrosis factor-a polymorphisms and colorectal cancer risk: a meta-analysis. PloS one. 2014;9(1):e85187. pmid:24404201.
- 8. Ronnekleiv-Kelly SM, Burkhart RA, Pawlik TM. Molecular markers of prognosis and therapeutic targets in metastatic colorectal cancer. Surgical oncology. 2016;25(3):190–9. pmid:27566022.
- 9. Garagnani P, Pirazzini C, Franceschi C. Colorectal cancer microenvironment: among nutrition, gut microbiota, inflammation and epigenetics. Current pharmaceutical design. 2013;19(4):765–78. pmid:23016865.
- 10. Nearchou A, Pentheroudakis G. The significance of tumor-associated immune response in molecular taxonomy, prognosis and therapy of colorectal cancer patients. Annals of translational medicine. 2016;4(14):271. pmid:27563658.
- 11. Bazzoni F, Beutler B. The tumor necrosis factor ligand and receptor families. The New England journal of medicine. 1996;334(26):1717–25. pmid:8637518.
- 12. Jovanovic DV, Di Battista JA, Martel-Pelletier J, Jolicoeur FC, He Y, Zhang M, et al. IL-17 stimulates the production and expression of proinflammatory cytokines, IL-beta and TNF-alpha, by human macrophages. Journal of immunology (Baltimore, Md: 1950). 1998;160(7):3513–21. pmid:9531313.
- 13. Szlosarek PW, Balkwill FR. Tumour necrosis factor alpha: a potential target for the therapy of solid tumours. The Lancet Oncology. 2003;4(9):565–73. pmid:12965278.
- 14. Grimm M, Lazariotou M, Kircher S, Hofelmayr A, Germer CT, von Rahden BH, et al. Tumor necrosis factor-alpha is associated with positive lymph node status in patients with recurrence of colorectal cancer-indications for anti-TNF-alpha agents in cancer treatment. Cellular oncology (Dordrecht). 2011;34(4):315–26. pmid:21573932.
- 15. van Horssen R, Ten Hagen TL, Eggermont AM. TNF-alpha in cancer treatment: molecular insights, antitumor effects, and clinical utility. Oncologist. 2006;11(4):397–408. pmid:16614236.
- 16. Wu Y, Zhou BP. TNF-alpha/NF-kappaB/Snail pathway in cancer cell migration and invasion. British journal of cancer. 2010;102(4):639–44. pmid:20087353.
- 17. Balkwill F. TNF-alpha in promotion and progression of cancer. Cancer metastasis reviews. 2006;25(3):409–16. pmid:16951987.
- 18. Hunter K. Host genetics influence tumour metastasis. Nature reviews Cancer. 2006;6(2):141–6. pmid:16491073.
- 19. Cariaso M, Lennon G. SNPedia: a wiki supporting personal genome annotation, interpretation and analysis. Nucleic Acids Res. 2012;40(Database issue):D1308–12. pmid:22140107.
- 20. Garrity-Park M, Loftus EV Jr., Sandborn WJ, Smyrk TC. Myeloperoxidase immunohistochemistry as a measure of disease activity in ulcerative colitis: association with ulcerative colitis-colorectal cancer, tumor necrosis factor polymorphism and RUNX3 methylation. Inflammatory bowel diseases. 2012;18(2):275–83. pmid:21425209.
- 21. Zambon CF, Basso D, Navaglia F, Falda A, Belluco C, Fogar P, et al. Increased risk of noncardia gastric cancer associated with proinflammatory cytokine gene polymorphisms. Gastroenterology. 2004;126(1):382–4. pmid:14753224.
- 22. Guo XF, Wang J, Yu SJ, Song J, Ji MY, Cao Z, et al. TNF-alpha-308 polymorphism and risk of digestive system cancers: a meta-analysis. World journal of gastroenterology. 2013;19(48):9461–71. pmid:24409077.
- 23. Jin Y-J, Lee D, Chung Y-H, Kim JA, Kim SE, Lee Y-S, et al. Tumor Necrosis Factor-Alpha Gene Polymorphism Associated With Development of Hepatitis B Virus–associated Hepatocellular Carcinoma. Journal of Clinical Gastroenterology. 2015;49(8):e76–e81. pmid:25319734
- 24. Li HH, Zhu H, Liu LS, Huang Y, Guo J, Li J, et al. Tumour Necrosis Factor-alpha Gene Polymorphism Is Associated with Metastasis in Patients with Triple Negative Breast Cancer. Scientific reports. 2015;5:10244. pmid:26165253.
- 25. Zhang H, Wang M, Shi T, Shen L, Liang L, Deng Y, et al. TNF rs1799964 as a Predictive Factor of Acute Toxicities in Chinese Rectal Cancer Patients Treated With Chemoradiotherapy. Medicine. 2015;94(45):e1955. pmid:26559268.
- 26. Jeon YJ, Kim JW, Park HM, Kim JO, Jang HG, Oh J, et al. Genetic variants in 3'-UTRs of methylenetetrahydrofolate reductase (MTHFR) predict colorectal cancer susceptibility in Koreans. Scientific reports. 2015;5:11006. pmid:26046315.
- 27. Brenner H, Kloor M, Pox CP. Colorectal cancer. Lancet. 2014;383(9927):1490–502. pmid:24225001.
- 28. Spriggs DR, Deutsch S, Kufe DW. Genomic structure, induction, and production of TNF-alpha. Immunology series. 1992;56:3–34. pmid:1550865.
- 29. Lu J, Li Y, Shen Z, Lu C, Lu L. TNF-alpha is involved in apoptosis triggered by grass carp reovirus infection in vitro. Fish Shellfish Immunol. 2016;55:559–67. pmid:27346157.
- 30. Wang GH, Jiang FQ, Duan YH, Zeng ZP, Chen F, Dai Y, et al. Targeting truncated retinoid X receptor-alpha by CF31 induces TNF-alpha-dependent apoptosis. Cancer Res. 2013;73(1):307–18. pmid:23151904.
- 31. Liu Y, Zhu L, Liang S, Yao S, Li R, Liu S, et al. Galactose protects hepatocytes against TNF-alpha-induced apoptosis by promoting activation of the NF-kappaB signaling pathway in acute liver failure. Laboratory investigation; a journal of technical methods and pathology. 2015;95(5):504–14. pmid:25751739.
- 32. Lim DS, Yawata N, Selva KJ, Li N, Tsai CY, Yeong LH, et al. The combination of type I IFN, TNF-alpha, and cell surface receptor engagement with dendritic cells enables NK cells to overcome immune evasion by dengue virus. Journal of immunology (Baltimore, Md: 1950). 2014;193(10):5065–75. pmid:25320280.
- 33. Loubaki L, Chabot D, Bazin R. Involvement of the TNF-alpha/TGF-beta/IDO axis in IVIg-induced immune tolerance. Cytokine. 2015;71(2):181–7. pmid:25461397.
- 34. Wang N, Li GN, Wang XB, Liang T, Hu L. TNF-alpha promoter single nucleotide polymorphisms and haplotypes associate with susceptibility of immune thrombocytopenia in Chinese adults. Hum Immunol. 2014;75(9):980–5. pmid:25158149.
- 35. Xu G, Zhang J. Suppression of FasL expression in tumor cells and preventing TNF-induced apoptosis was better for immune cells survival. Journal of cancer research and clinical oncology. 2008;134(10):1043–9. pmid:18461366.
- 36. Hojilla CV, Jackson HW, Khokha R. TIMP3 regulates mammary epithelial apoptosis with immune cell recruitment through differential TNF dependence. PloS one. 2011;6(10):e26718. pmid:22053204.
- 37. Wilson AG, Symons JA, McDowell TL, McDevitt HO, Duff GW. Effects of a polymorphism in the human tumor necrosis factor alpha promoter on transcriptional activation. Proceedings of the National Academy of Sciences of the United States of America. 1997;94(7):3195–9. pmid:9096369.