7 Oct 2013: (2013) Correction: Hypoxia-Inducible Factor-1α Polymorphisms and Risk of Cancer Metastasis: A Meta-Analysis. PLoS ONE 8(10): 10.1371/annotation/938850a0-b05c-4557-becc-9d91d3d90742. doi: 10.1371/annotation/938850a0-b05c-4557-becc-9d91d3d90742 | View correction
HIF-1α is a major regulator in tumor progression and metastasis which responds to hypoxia. Many studies have demonstrated that hypoxia-inducible factor1-α (HIF-1α) polymorphisms are significantly associated with cancer metastasis, but the results are inconsistent. We conducted a comprehensive meta-analysis to estimate the associations between HIF-1α C1772 T polymorphism and cancer metastasis.
Comprehensive searches were conducted on PubMed and EMBASE database. Fifteen studies were included in the meta-analysis. We used the OR and 95%CI to assess the associations between HIF-1α C1772T polymorphism and cancer metastasis. Heterogeneity and publication bias were also assessed by Q test, I 2, and funnel plot.
Totally, fifteen studies including 1239 cases with metastasis-positive (M+) and 2711 cases with metastasis-negative (M−) were performed in this meta-analysis. The results showed that HIF-1a C1772T polymorphism was associated with the increased risk of cancer metastasis (T allele vs. C allele, OR = 1.36, 95% CI = 1.12–1.64; TT+ TC vs. CC, OR = 1.39, 95% CI = 1.13–1.71; TT vs. TC+ CC, OR = 1.93, 95% CI = 0.86–4.36). In the subgroup analyses, the significant associations remained significant among Asians, Caucasians and other cancers in the dominant model. Publication bias was not observed in the analysis.
Our results indicate that the HIF-1αC1772T polymorphism T allele may increase the risk of cancer metastasis, which might be a potential risk factor of cancer progress.
Citation: Zhang Q, Chen Y, Zhang B, Shi B, Weng W, et al. (2013) Hypoxia-Inducible Factor-1α Polymorphisms and Risk of Cancer Metastasis: A Meta-Analysis. PLoS ONE 8(8): e70961. doi:10.1371/journal.pone.0070961
Editor: Hiromu Suzuki, Sapporo Medical University, Japan
Received: May 6, 2013; Accepted: June 24, 2013; Published: August 28, 2013
This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Funding: This study was partially supported by the Postdoctoral Science Foundation of Jiangsu Province (528), the Health Department guidance project of Jiangsu Province (Z201201), the Program for Development of Innovative Research Team in the First Affiliated Hospital of NJMU and the Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (JX10231801). 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.
Cancer metastasis is a progress that tumor cells displace from the primary site to distance site where cancer cells adapt to a tissue microenvironment and is the most important cause of death in cancer patients . The development of metastasis consists of a series of complex steps which involve in immunologic escape, angiogenesis, invasion of lymph- and blood vessels and so on . The molecular mechanisms of angiogenesis on cancer metastasis have become focal points in the past years. Recently, research of tumor metastasis has been focused on the hypoxic condition. Hypoxia is one of the most important mechanisms that induces cancer metastasis – and regulates the metastatic process by metabolism, angiogenesis, innate immunity, and stem cells induction . Hypoxia-inducible factor1α (HIF-1α) plays an important role in the growth and metastasis of tumor. Previous study has showed that increased level of HIF-1α protein was associated with lymph node metastasis and high malignant degree .
Genetic polymorphisms have been considered as the main genetic elements involved in the occurrence and development of cancer . HIF-1α C1772T (rs11549465) is a common single nucleotide polymorphism (SNP) that is located in exon 12 which results in proline to serine. The polymorphism promotes the development and progression of cancer by increasing the density of cancer microvessel . A number of studies have showed the relationship between the HIF-1α C1772T polymorphism and cancer metastasis, but the results have been disparate , –. No meta-analysis has been performed to investigate these associations so far. To better explore the associations of HIF-1α C1772T polymorphism with cancer metastasis, we conducted a meta-analysis to collect and analyze the published data.
Materials and Methods
Literature source and search strategy
The published case-control studies that investigated the associations between the HIF-1α C1772T polymorphism and cancer metastasis were searched on PubMed and EMBASE database (between January 1, 2005 and December 1, 2012). The keywords and terms used for this search were “HIF-1 OR hypoxia-inducible factor-1”, “polymorphism” AND “cancer”. In addition, references of retrieved publications were also screened by hand-searched. Studies included in the current meta-analysis have to meet the following inclusion criteria: (i) independent case-control design, (ii) evaluation of the associations between HIF-1α C1772T polymorphism and cancer metastasis, and (iii) provide available genotype frequency.
Two investigators independently extracted data. If the data was different, the third reviewer was asked to check until the data was right. For each eligible study, we collected the following information: first author's name, the year of publication, country of origin, cancer type, ethnicity of study population, the number of metastasis-positive (M+) and metastasis-negative (M−) cases and genotypes. The criteria of metastasis-positive (M+) and metastasis-negative (M−) depended on TNM . The patients were assigned to metastasis-positive (M+) depended on the presence of detectable lymph nodes metastasis or distant metastasis at the time of diagnose or follow-up, or assigned to metastasis- negative (M−). In one study, if lymph nodes metastasis and distant metastasis were both investigated, we selected lymph nodes metastasis as the criteria in the current meta-analysis.
The strength of the association between HIF-1α C1772 T and cancer metastasis was measured by odds ratio (OR) and 95% confident interval (CI). The influence of study size of evaluated studies on the results was assessed by the Weight. The statistical significance of the pooled OR was determined using the Z-test (P <0.05 was considered significant). The pooled OR was first evaluated on allele (T allele vs. C allele), and then on the dominant model (TT+CT vs. CC) and recessive model (TT vs. CT+CC). Subgroup analysis were also conducted by ethnicity, cancer types (if one cancer type contained one individual study, it was combined into the “other cancers” group).
Heterogeneity between studies was assessed by the χ2-based Q-test and I2. If P>0.05 of the Q -test indicated a lack of heterogeneity across eligible studies, a fixed-effects model (the Mantel-Haenszel method) was used for meta-analysis. Otherwise, the random-effects model was used (DerSimonian and Laird method). Funnel plot and Egger's linear regression test were applied to evaluate the potential publication bias. All statistical analyses were carried out with Stata software (version 11.0, USA), using two-sided P-values.
Characteristics of studies
27 studies were obtained to evaluate the relationship between HIF-1α C1772 T and cancer metastasis. Twelve of them were excluded (8 studies were excluded for lacking of accurate staging, 4 studies without genotype data). Finally, 15 articles , – were included in our meta-analysis (Figure 1). The detail characteristics of eligible studies were summarized in Table 1. Among these studies, there were 10 studies of Asians and 5 of Caucasians. All studies were case-control studies which contained four studies with breast cancer, two with colorectal cancer and other cancers group. As shown in Table 1, 1239 (M+) cases and 2711 (M−) cases were included in the study.
Figure 1. The detailed process of identifying eligible studies.doi:10.1371/journal.pone.0070961.g001
Table 1. Characteristics of HIF-1a polymorphisms Included in the Meta-analysis.doi:10.1371/journal.pone.0070961.t001
Quantitative data synthesis
We assessed the associations between the HIF-1α C1772T polymorphism and cancer metastasis. Overall, when all the eligible studies were pooled into the meta-analysis, variant T allele significantly increased the risk of cancer metastasis, compared with the wild-type C allele (OR = 1.36, 95% CI = 1.12–1.64, P = 0.002, P heterogeneity = 0.17. I2 = 25.9; Figure 2. A), the dominant mode (TT+CT vs. CC) showed that there were significant associations between HIF-1α C1772 T and cancer metastasis (OR = 1.39, 95% CI = 1.13–1.71, P = 0.002, P heterogeneity = 0.27. I2 = 16.8%; Figure 2. B). Significant associations were not observed under the recessive model (TT vs. CC+CT) (OR = 1.93, 95% CI = 0.86–4.36, P = 0.11, P heterogeneity = 0.13. I2 = 44%; Figure 2. C). The pooled OR for C vs. T and TT/CT vs. CC indicated that HIF-1a C1772T was significantly associated with a increased risk of cancer metastasis.
Figure 2. Forest plot of cancer metastasis associated with the HIF-1a C1772 T polymorphisms.
(A) C allele vs. T allele. (B) The dominant model. (C) The recessive model. A fixed-effects model was used. The squares and horizontal lines correspond to the study-specific OR and 95% CI. The area of the squares reflects the weight (inverse of the variance). The diamond represents the summary OR and 95% CI.doi:10.1371/journal.pone.0070961.g002
Under the dominant model, subgroup analysis based on cancer type was performed, significant associations were not found in colorectal cancer and breast cancer, but a significant association in other cancer was observed. In the stratified analysis by ethnicity, significant associations were observed among Asians and in Caucasians (Table 2).
Table 2. Stratified analysis of HIF-1a polymorphisms on cancer metastasis.doi:10.1371/journal.pone.0070961.t002
Tests for heterogeneity and Sensitivity
Totally, no significant heterogeneity was observed among studies for the associations between HIF-1α C1772 T polymorphism and cancer metastasis in the pooled analysis and stratified analysis of dominant model. The fixed-effects model was performed in the meta-analysis. Any single study was not found to change the pooled OR qualitatively by sensitivity analysis indicated that this meta-analysis is stable.
Begg's funnel plot and Egger's test were performed to assess the publication bias of literatures. As shown in Figure 3, the shape of the funnel plot did not reveal any evidence of obvious asymmetry in all comparison models. Also, the results of Egger's test did not show any evidence of publication bias. (T vs. C, t = −0.13, P = 0.900, 95%CI = −1.43–1.27; TT/TC vs. CC, t = 0.02, P = 0.983, 95%CI = −1.23–1.26; TT vs. TC/CC, t = −0.73, P = 0.519, 95%CI = −7.60–4.77).
The meta-analysis investigates the associations between HIF-1α C1772 T polymorphism and cancer metastasis. We found that variant T allele significantly increased the risk of cancer metastasis.
Hypoxia-inducible factor1α (HIF-1α) is a transcription factor was first found as a regulator of renal production of erythropoietin (Epo) . It is a helix-loop-helix transcription factor that consists of α and β subunits. HIF-1α subunit is regulated by oxygen pressure and HIF-1 β is constitutively expressed . In normoxic conditions, HIF-1α is hydroxylated at specific proline residues. Under hypoxic conditions, HIF-1α is induced and combines with the β subunit, then removes to the nucleus and initiates gene transcription , . The HIF-1α protein contains five functional domains: basic helix-loop-helix (bHLH), Per/Arnt/Sim (PAS), N-terminal (N-TAD), C-terminal (C-TAD), and oxygen-dependent degradation (ODD) .
HIF-1α is overexpressed in regional or distant metastases and is also expressed higher in preneoplastic and premalignant lesions, indicating that overexpression of HIF-1α can occur very early in carcinogenesis which may become a potential biomarker of predicting tumor progress  and a good target for the detecting of tumor metastasis. HIF-1 activates the transcription of a large number of genes that for erythropoietin, vascular endothelial growth factor, endothelin-1, nitric oxide synthase, heme oxygenase-1, and insulin-like growth factor-2. However, the expression of vascular endothelial growth factor (VEGF) is an essential molecular event for tumor development and metastasis . It is known that HIF-1 induces erythropoiesis and angiogenesis, and is also involved in the regulation of both vascular tone and glucose metabolism . And HIF-1 transcriptional activity is regulated by the HIF-1α subunit. There were evidences indicate that HIF-1 plays an important role in cancer progression and metastasis , . Higher expression levels of HIF-1α have been reported in human malignancies including colon, breast. HIF-1α polymorphism C1772T in human were initially identified in renal cell carcinoma patients which cause amino acid substitutions within the N-TAD, however, the difference in genotype distribution was not indicated between renal cell carcinoma cases and controls . It has been reported that the C1772T polymorphism T allele in HIF-1α represent higher transcriptional activity than that of wild-type C allele under both normoxic and hypoxic conditions . Therefore, the presence of this polymorphism might be associated with in cancer risk and cancer metastasis. However, the studies had controversial conclusions. A meta-analysis has proved that the HIF-1α C1772 T polymorphism is significantly associated with higher cancer risk . The aim of this study was to investigate the association between HIF-1α C1772 T polymorphism and cancer metastasis. In our analysis, we found that variant T allele significantly increased the risk of metastasis. Further, the associations were very stable, which did not change apparently when the sensitivity analyses were performed. The results indicated that T allele is a potential risk factor in cancer metastasis. There was a significant association between HIF-1α C1772 T polymorphism and cancer metastasis under the dominant model, while no association was found in the recessive model. One possible explanation is that the number of studies which investigate the TT genotype separately is too small. The results may also indicate that heterozygous T has a stronger effect on an individual's phenotype than homozygous T. It is mean that individuals with CT genotype may have a higher risk of metastasis than those with TT genotype. Further larger sample size and well-designed studies should be performed to testimony our results.
For the stratification analyses based on ethnicity and cancer types. There was an evidence to indicate that the HIF-1α C1772T polymorphism was significantly associated with increased risk of cancer metastasis among Asians and Caucasians only for dominant genetic model. However, no significant association was found between the C1772T polymorphism and cancer metastasis in colorectal cancer and breast cancer under dominant model. Further studies using larger sample size are needed to validate. Kuwai et al.  indicated that the C1772T polymorphism in HIF-1α is not association with the progression and metastasis of colorectal cancer, while the study of patients with ESCC has a contrary result . It may because C1772T has different values in different kinds of tumors. The T/T genotype is rare in our study, thus, the absence of association between T/T and cancer metastasis may be due to chance. Extended epidemiological studies would be needed to determine if this genotype is associated with cancer metastasis. There is a previous founding reported that HIF-1α genetic variant increases risk of breast Lymph node metastasis , which is consistent with our subgroup analysis. It is likely that variations of sample sizes may account for contradictory results.
There are some limitations and potential bias that must be acknowledged in our meta-analysis. First, only published studies were included in our studies, many unpublished data have been ignored in the analysis. Therefore, potentially publication bias will be existed in our results, although the statistical data did not reflect it. Second, because of lacking of detailed analysis about age, gender, smoking, drinking and so on, those potential factors may influence our results. Third, the number of the included studies was not large enough. So the statistical power is weak to evaluate the association between HIF-1a polymorphisms and metastasis, especially in stratified analyses.
In summary, our meta-analysis reveal that the HIF-1α C1772T polymorphism can increase the risk of cancer metastasis. Although some results of the analysis are limited by the small number of studies. Our results suggest that the polymorphism is a potential risk factor. Large sample size and well-designed studies are needed to evaluate our finding.
Conceived and designed the experiments: LJZ YBH. Performed the experiments: ZPC NNG. Analyzed the data: QZ YC BZ BS WJW ZPC NNG. Contributed reagents/materials/analysis tools: BS WJW ZPC. Wrote the paper: QZ YC BZ.
- 1. Gupta GP, Massague J (2006) Cancer metastasis: building a framework. Cell 127: 679–695. doi: 10.1016/j.cell.2006.11.001
- 2. Bohle AS, Kalthoff H (1999) Molecular mechanisms of tumor metastasis and angiogenesis. Langenbecks Arch Surg 384: 133–140. doi: 10.1007/s004230050183
- 3. Tsai YP, Wu KJ (2012) Hypoxia-regulated target genes implicated in tumor metastasis. J Biomed Sci 19: 102. doi: 10.1186/1423-0127-19-102
- 4. Majmundar AJ, Wong WJ, Simon MC (2010) Hypoxia-inducible factors and the response to hypoxic stress. Mol Cell 40: 294–309.5. doi: 10.1016/j.molcel.2010.09.022
- 5. Semenza GL (2012) Hypoxia-inducible factors: mediators of cancer progression and targets for cancer therapy. Trends Pharmacol Sci 33: 207–214. doi: 10.1016/j.tips.2012.01.005
- 6. Lu X, Kang Y (2010) Hypoxia and hypoxia-inducible factors: master regulators of metastasis. Clin Cancer Res 16: 5928–5935. doi: 10.1158/1078-0432.ccr-10-1360
- 7. Naidu R, Har YC, Taib NA (2009) Associations between hypoxia-inducible factor-1alpha (HIF-1alpha) gene polymorphisms and risk of developing breast cancer. Neoplasma 56: 441–447. doi: 10.4149/neo_2009_05_441
- 8. Rannala B (2001) Finding genes influencing susceptibility to complex diseases in the post-genome era. Am J Pharmacogenomics 1: 203–221. doi: 10.2165/00129785-200101030-00005
- 9. Huang LE, Gu J, Schau M, Bunn HF (1998) Regulation of hypoxia-inducible factor 1alpha is mediated by an O2-dependent degradation domain via the ubiquitin-proteasome pathway. Proc Natl Acad Sci U S A 95: 7987–7992. doi: 10.1073/pnas.95.14.7987
- 10. Ling TS, Shi RH, Zhang GX, Zhu H, Yu LZ, et al. (2005) Common single nucleotide polymorphism of hypoxia-inducible factor-1alpha and its impact on the clinicopathological features of esophageal squamous cell carcinoma. Chin J Dig Dis 6: 155–158. doi: 10.1111/j.1443-9573.2005.00223.x
- 11. Orr-Urtreger A, Bar-Shira A, Matzkin H, Mabjeesh NJ (2007) The homozygous P582S mutation in the oxygen-dependent degradation domain of HIF-1 alpha is associated with increased risk for prostate cancer. Prostate 67: 8–13. doi: 10.1002/pros.20433
- 12. Lee JY, Choi JY, Lee KM, Park SK, Han SH, et al. (2008) Rare variant of hypoxia-inducible factor-1alpha (HIF-1A) and breast cancer risk in Korean women. Clin Chim Acta 389: 167–170. doi: 10.1016/j.cca.2008.02.001
- 13. Apaydin I, Konac E, Onen HI, Akbaba M, Tekin E, et al. (2008) Single nucleotide polymorphisms in the hypoxia-inducible factor-1alpha (HIF-1alpha) gene in human sporadic breast cancer. Arch Med Res 39: 338–345. doi: 10.1016/j.arcmed.2007.11.012
- 14. Kim HO, Jo YH, Lee J, Lee SS, Yoon KS (2008) The C1772T genetic polymorphism in human HIF-1alpha gene associates with expression of HIF-1alpha protein in breast cancer. Oncol Rep 20: 1181–1187. doi: 10.3892/or_00000127
- 15. Knechtel G, Szkandera J, Stotz M, Hofmann G, Langsenlehner U, et al. (2010) Single nucleotide polymorphisms in the hypoxia-inducible factor-1 gene and colorectal cancer risk. Mol Carcinog 49: 805–809. doi: 10.1002/mc.20655
- 16. Hsiao PC, Chen MK, Su SC, Ueng KC, Chen YC, et al. (2010) Hypoxia inducible factor-1alpha gene polymorphism G1790A and its interaction with tobacco and alcohol consumptions increase susceptibility to hepatocellular carcinoma. J Surg Oncol 102: 163–169. doi: 10.1002/jso.21539
- 17. Shieh TM, Chang KW, Tu HF, Shih YH, Ko SY, et al. (2010) Association between the polymorphisms in exon 12 of hypoxia-inducible factor-1alpha and the clinicopathological features of oral squamous cell carcinoma. Oral Oncol 46: e47–53. doi: 10.1016/j.oraloncology.2010.04.009
- 18. Kang MJ, Jung SA, Jung JM, Kim SE, Jung HK, et al. (2011) Associations between single nucleotide polymorphisms of MMP2, VEGF, and HIF1A genes and the risk of developing colorectal cancer. Anticancer Res 31: 575–584.
- 19. Wang X, Liu Y, Ren H, Yuan Z, Li S, et al. (2011) Polymorphisms in the hypoxia-inducible factor-1alpha gene confer susceptibility to pancreatic cancer. Cancer Biol Ther 12: 383–387. doi: 10.4161/cbt.12.5.15982
- 20. Qin C, Cao Q, Ju X, Wang M, Meng X, et al. (2012) The polymorphisms in the VHL and HIF1A genes are associated with the prognosis but not the development of renal cell carcinoma. Ann Oncol 23: 981–989. doi: 10.1093/annonc/mdr325
- 21. Fraga CA, de Oliveira MV, de Oliveira ES, Barros LO, Santos FB, et al. (2012) A high HIF-1alpha expression genotype is associated with poor prognosis of upper aerodigestive tract carcinoma patients. Oral Oncol 48: 130–135. doi: 10.1016/j.oraloncology.2011.08.023
- 22. Mera-Menendez F, Hinojar-Gutierrez A, Guijarro Rojas M, de Gregorio JG, Mera-Menendez E, et al.. (2012) Polymorphisms in HIF-1alpha affect presence of lymph node metastasis and can influence tumor size in squamous-cell carcinoma of the glottic larynx. Clin Transl Oncol.
- 23. Chai D, Chen YL, Zheng A, Liu YY, Chu YX, et al. (2010) [Relationship between polymorphism of hypoxia inducible factor-1alpha and cervical cancer in Han population in Sichuan Province of China]. Sichuan Da Xue Xue Bao Yi Xue Ban 41: 674–677.
- 24. Sobin LH, Fleming ID (1997) TNM Classification of Malignant Tumors, fifth edition (1997). Union Internationale Contre le Cancer and the American Joint Committee on Cancer. Cancer 80: 1803–1804. doi: 10.1002/(sici)1097-0142(19971101)80:9<1803::aid-cncr16>3.0.co;2-9
- 25. Smaldone MC, Maranchie JK (2009) Clinical implications of hypoxia inducible factor in renal cell carcinoma. Urol Oncol 27: 238–245. doi: 10.1016/j.urolonc.2007.12.001
- 26. Kizaka-Kondoh S, Inoue M, Harada H, Hiraoka M (2003) Tumor hypoxia: a target for selective cancer therapy. Cancer Sci 94: 1021–1028. doi: 10.1111/j.1349-7006.2003.tb01395.x
- 27. Semenza GL, Wang GL (1992) A nuclear factor induced by hypoxia via de novo protein synthesis binds to the human erythropoietin gene enhancer at a site required for transcriptional activation. Mol Cell Biol 12: 5447–5454.
- 28. Nagy G, Kovacs-Nagy R, Kereszturi E, Somogyi A, Szekely A, et al. (2009) Association of hypoxia inducible factor-1 alpha gene polymorphism with both type 1 and type 2 diabetes in a Caucasian (Hungarian) sample. BMC Med Genet 10: 79. doi: 10.1186/1471-2350-10-79
- 29. Zhong H, De Marzo AM, Laughner E, Lim M, Hilton DA, et al. (1999) Overexpression of hypoxia-inducible factor 1alpha in common human cancers and their metastases. Cancer Res 59: 5830–5835.
- 30. Zheng ZL, Hwang YH, Kim SK, Kim S, Son MJ, et al. (2009) Genetic polymorphisms of hypoxia-inducible factor-1 alpha and cardiovascular disease in hemodialysis patients. Nephron Clin Pract 113: c104–111. doi: 10.1159/000228542
- 31. Kizaka-Kondoh S, Inoue M, Harada H, Hiraoka M (2003) Tumor hypoxia: a target for selective cancer therapy. Cancer Sci 94: 1021–1028. doi: 10.1111/j.1349-7006.2003.tb01395.x
- 32. Clifford SC, Astuti D, Hooper L, Maxwell PH, Ratcliffe PJ, et al. (2001) The pVHL-associated SCF ubiquitin ligase complex: molecular genetic analysis of elongin B and C, Rbx1 and HIF-1alpha in renal cell carcinoma. Oncogene 20: 5067–5074. doi: 10.1038/sj.onc.1204602
- 33. Tanimoto K, Yoshiga K, Eguchi H, Kaneyasu M, Ukon K, et al. (2003) Hypoxia-inducible factor-1alpha polymorphisms associated with enhanced transactivation capacity, implying clinical significance. Carcinogenesis 24: 1779–1783. doi: 10.1093/carcin/bgg132
- 34. Zhao T, Lv J, Zhao J, Nzekebaloudou M (2009) Hypoxia-inducible factor-1alpha gene polymorphisms and cancer risk: a meta-analysis. J Exp Clin Cancer Res 28: 159. doi: 10.1186/1756-9966-28-159
- 35. Kuwai T, Kitadai Y, Tanaka S, Kuroda T, Ochiumi T, et al. (2004) Single nucleotide polymorphism in the hypoxia-inducible factor-1alpha gene in colorectal carcinoma. Oncol Rep 12: 1033–1037. doi: 10.3892/or.12.5.1033