HER2 Polymorphisms and Breast Cancer in Tunisian Women

Abstract

HER2 has been thought to play a critical role in both breast cancer development and progression. Any functional polymorphisms can potentially affect breast cancer risk as well as cancer phenotype and outcome. In our study, we analyzed three polymorphisms in the HER2 gene: the single-nucleotide polymorphism (SNP) HER2 Ile⁶⁵⁵Val as well as another SNP (rs903506) close to it and a new screened dinucleotide repeat H(AC)I4 in intron 4, in a sample of 148 cases and 290 controls from the Tunisian population and investigated their association with breast cancer risk. For the HER2 Ile⁶⁵⁵Val, we found similar allele frequencies between cases and controls (frequency of I allele was 0.92 and 0.91, respectively). The same was observed for the noncoding SNP (rs903506). These two SNPs also showed no association with any clinical parameters, except the association of HER2 Ile⁶⁵⁵Val with tumor size (p = 0.002). But, a significant association was found between the short tandem repeat (STR) [H(AC)I4] and breast cancer risk at both genotypic and allelic levels (p = 0.0004 and p = 0.0001, respectively). Multivariate analysis with binary logistic regression of disease status on genotypes of the three polymorphisms confirmed the association of STR with breast cancer risk (p = 0.016). Therefore, this STR seems to be a promising biomarker in breast cancer and deserves further investigation.

Introduction

HER2 (also known as ErbB2 or neu ) is a proto-oncogene located on chromosome 17 which encodes a transmembrane glycoprotein (p185) with tyrosine kinase activity (Akiyama et al., 1986; Bargmann et al., 1986). HER2 is a member of the epidermal growth factor receptor family, a family of protein tyrosine kinases involved in cell division, migration, adhesion, differentiation, and apoptosis (Yarden and Sliwkowski, 2001). P185 (kDa) plays an important role in many physiological processes such as cell growth, cell mortality factor (McKean-Cowdin et al., 2001), differentiation, and tissue development, but it is also involved in carcinogenesis and metastasis (Tommasi et al., 2003). Overexpression of HER2 has been observed in many solid tumors, including 20-30% of breast cancers and 20-59% of ovarian carcinomas (Slamon et al., 1987; Normanno et al., 2006).

Further, it has been shown that HER2 gene amplification/overexpression is correlated with worse clinicopathological parameters such as steroid hormone receptor-negative tumors, tumor size, lymph node involvement, high nuclear and histological grade, aneuploidy, proliferation index, and reduced response to chemotherapy and hormone therapy (Slamon et al., 1987; Pegram et al., 1997; Gregory et al., 2000; Witton et al., 2003; WisemAn et al., 2005). In addition to high expression level, the active form of HER2 could be induced by changes in transmembrane amino acid composition (Stern et al., 1988; Shigematsu et al., 2005).

Mutations in the transmembrane region of the rat homologue of HER2 occur in chemically induced rat neuroblastomas and lead to increased tyrosine phosphorylation (Prenzel et al., 2001). Sequence analysis of the cDNA identified an A/G transition polymorphism in the first position of codon 655 in the DNA region encoding the transmembrane domain (Papewalis et al., 1991). This variant corresponds to Ile⁶⁵⁵Val nonsynonymous substitution. Many investigators addressed the relationship between the HER2 Ile⁶⁵⁵Val polymorphism and breast cancer (Papewalis et al., 1991; Xie et al., 2000; Fleishman et al., 2002; Millikan et al., 2003; Montgomery et al., 2003). The initial report of positive association between this variation and breast cancer (Xie et al., 2000) has not been successfully replicated till now (Montgomery et al., 2003; Kamali-Sarvestani et al., 2004; An et al., 2005; Benusiglio et al., 2006).

However, subgroup analysis in several studies suggested that among women who were younger (Xie et al., 2000; Millikan et al., 2003; Montgomery et al., 2003), were physically inactive (Xie et al., 2000), had larger body mass (Xie et al., 2000; Millikan et al., 2003), or had a positive family history of breast cancer, the Val/Val genotype was associated with increased risk of breast cancer in comparison with the Ile/Ile genotype.

As the HER2 gene is also an important biomarker, particularly in breast cancer, in our study we analyzed the HER2 Ile⁶⁵⁵Val polymorphism (rs1801200) as well as another single-nucleotide polymorphism (SNP) (rs903506) close to it and also a conserved dinucleotide repeat H(AC)I4 in intron 4 which we recently described (Kharrat et al., 2008) in a case-control study. We also investigated the association of these polymorphisms with HER2 protein expression and other pathological and clinical parameters.

Materials and Methods

Subject

A retrospective study was conducted on a group of 148 Tunisian women with breast cancer selected from the register of Sfax University Hospital Habib Bourguiba between the years 2001 and 2007, with a mean age of 51 ± 11.43 years (27-79). The control group contains 290 women with a mean age of 54 ± 17.24 years (20-87). All were healthy and clinically asymptomatic women randomly selected from blood donors of the Blood Transfusion Center of Sfax (Tunisia) and matched with patients for age and social conditions.

Immunohistochemistry

The status of HER2, Bcl2 protein, estrogen, and progesterone receptor was assessed by immunohistochemistry on frozen tissue. Primary antibodies used were BM5084 (HER2; ACRIS, Hiddenhausen, Germany), clone 124 (bcl2; DAKO, Carpinteria, CA), clone 1D5 (ER1; DAKO), and clone PgR636 (PR; DAKO), which were incubated for 30 min at room temperature. A biotin-labeled secondary antibody (goat anti-rabbit, 1:500; DAKO) was incubated for 15 min at room temperature. Then the avidin-biotin-horseradish peroxidase (HRP) complex was applied for 15 min as directed by (DAKO), and finally, the immunoprecipitate was visualized by treating with diaminobenzidine tetrahydrochloride (DAKO) for 30 min. The sections were counterstained with hematoxylin.

Immunostaining was considered positive if more than 5% of tumor cells were stained, and it was scored on the basis of the approximate percentage of positive tumor cells and the relative immunostaining intensity (Khabir et al., 2003).

Phenotyping

One hundred thirty-one patients had invasive ductal carcinoma, 7 patients had invasive lobular carcinoma, and 10 patients had medullar or tubular cancers. Histological grade was not available for lobular carcinomas because tubule formation is not a characteristic of the tumor. The microscopic grading of Scarff-Bloom-Richardson was used (Bloom and Richardson, 1957); among the patients with ductal carcinoma, 9 (6.8%) had grade I, 77 (58.8%) had grade II, and 45 (34.3%) had grade III tumors. The nodal status was known for 144 patients; 52 (36.1%) had no nodal involvement. Lymph node invasion was detectable in 92 cases (63.8%), of which 51 had one to three nodes involved and 41 had more than three nodes involved.

DNA extraction and polymerase chain reaction

DNA was extracted from freshly frozen breast cancer tissue of 148 patients, using a Wizard Genomic DNA purification kit (Promega, Madison, WI), and also from blood samples of all controls and patients using standard phenol-chloroform protocol.

The genotypes of the polymorphisms were determined with polymerase chain reaction (PCR) for the dinucleotide repeats H(AC)I4 in intron 4, using labeled primers, as reported by Kharrat et al. (2008), and with PCR-restriction fragment length polymorphism-based method (Pinto et al., 2004) for the two SNPs (rs1801200 and rs903506).

The primers for PCR amplification of the studied polymorphisms are given in Table 1 and Figure 1.

FIG. 1.

The position of polymorphisms studied in the HER2 gene.

Table 1.

Features of the Polymorphisms and Primer Sequences Used for Genotyping

Polymorphisms	Type	Primers	References
rs1801200	A/G	F: 5′-AGAGCGCCAGCCCTCTGACGTCCAT-3′ R: 5′-TCCGTTTCCTGCAGCAGTCTCCGCA-3′	Pinto et al. (2004)
rs903506	A/G	F: 5′-TCCGGAAGTACACGATGCGG-3′ R: 5′-TCAGCTCCGTCTCTTTCAGG-3′	www.gdb.org
H(AC)I4	(AC)_n	F: 5′-FAM TTCAGTCCTCTGACCCCAAG-3′ R: 5′-GGACAGGGCTCTAGCTTCC-3′	Kharrat et al. (2008)

SNP genotyping

For Ile⁶⁵⁵Val, PCR products were digested by restriction enzyme Alw26 for 2 h at 37°C and subsequently separated by electrophoresis in 3% Nuseiv/agarose (2:1, wt:wt) (digestion of the G allele at codon 655 results in two fragments of 116 and 32 bp, whereas the A allele is undigested and gives a fragment of 148 bp). For the SNP rs903506, PCR products were digested by restriction enzyme NciI for 2 h at 37°C and also separated by electrophoresis in 3% Nuseiv/agarose (2:1, wt:wt) (digestion of the G allele results in two fragments of 106 and 91 bp).

The two SNPs are located at 175 bp apart; Ile⁶⁵⁵Val (rs1801200) is located in exon 17 of the gene HER2 (encoding the transmembrane domain), whereas the A23271G polymorphism (rs903506) is within intron 17.

The dinucleotide repeat H(AC)I4 genotyping

Assessment of the dinucleotide repeat H(AC)I4 in intron 4 was performed using Super Mixte Applied Biosystems (2 × ) in a 6.25 μL reaction volume containing 100 ng DNA and 10 μM of each primer. The PCR reactions were performed using a GenAmp PCR system 9600 thermocycler (PerkinElmer, Massy, France). Cycling conditions for screening the microsatellite were as follows: initial heating at 95°C for 5 min, followed by 30 cycles of denaturation at 94°C for 11 min, annealing at 56°C for 1 min, and extension at 72°C for 1 min. About 0.5 μL of the PCR products was mixed with 9.5 μL of formamide containing 0.3 μL of ROX400 molecular weight standard (Applied Biosystems, Foster City, CA). The denatured PCR products were separated on the ABI PRISM 310 DNA Analyzer (Applied Biosystems). Evaluation of the collected data was done with the GeneScan Analysis software (Applied Biosystems) and the fragment lengths were determined. To confirm the number of repeats, direct sequencing of the PCR products was carried out.

Statistical analysis

Genotypic and allelic frequencies were calculated by simple counting using Microsoft Excel. The exact test for the Hardy-Weinberg equilibrium (HWE) was performed using the Genetic Data Analyses program (version 1.1) (Hauptmann et al., 2003). Association of the studied polymorphisms with breast cancer susceptibility was assessed using the case-control chi-square test with p ≤ 0.05 taken as the cutoff for statistical significance. Odds ratios (OR) and their 95% confidence intervals (CIs) were estimated using programs from the linkage utility package (http://linkage.rockefeller.edu).

The association of the polymorphisms with pathological and clinical parameters was evaluated by two-way analysis of variance, Spearman correlation, and Student's t-test. A multivariate logistic regression of the disease status on the polymorphisms and age was performed to test associations with risk corrected for age. These analyses were computed using appropriate functions from the R language (www.r-project.org).

Results

HER2 expression and clinical parameters

In this sample, overexpression was observed for HER2, bcl2, ER1, and PR in 62%, 69%, 65%, and 58% of tumors, respectively. HER2 was positively correlated with tumor size (r_s = 0.318; p = 0.003) and bcl2 status (r_s = 0.270; p = 0.013) but not with Scarff-Bloom-Richardson grade (p = 0.25). There was no overall significant association between the expression of HER2 and that of hormonal receptors.

HER2 polymorphisms

Genotype frequencies in control samples were in HWE for both SNPs (p = 0.14 and p = 0.33). For the noncoding SNP (rs903506), the frequencies of the wild-type (G) allele were 75% and 66% in cases and controls, respectively, but this difference is not statistically significant (Table 2). No association was found between SNP rs903506 and clinical parameters in our study.

Table 2.

HER2 Genotype Distribution of Study Subjects

	Genotypes			Alleles (%)
Polymorphism	AA (%)	AG (%)	GG (%)	A	G
rs1801200
Cases (n = 148)	130 (88.83)	14 (10)	4 (2.7)	274	22
Controls (n = 290)	240 (83)	50 (17.5)	0 (0)	530	50
Genotypic association		p = 0.0017
Allelic association				p = 0.54
rs903506
Cases (n = 60)	33 (55)	24 (40)	3 (5)	80	30
Controls (n = 200)	89 (44.5)	87 (43.5)	24 (12)
Genotypic association		p = 0.18
Allelic association				p = 0.07

For SNP rs1801200, no significant difference was found in allele frequencies between cases and controls (frequency of I allele was 0.92 and 0.91, respectively). In contrast, the homozygous V genotype was present in 2.7% of patients but absent in the control group. When genotype frequencies are considered, we found a significant difference (Fisher's exact test, p = 0.0017) between controls and patients (Table 2). However, this association was not detected when we compared the [Val/Val + Ile/Val] versus [Ile/Ile] (p = 0.16).

Moreover, we looked for a potential association between the rs1800200 and clinicopathological parameters in the patient's sample. Associations between genotype and major parameters studied were nonsignificant (Table 3). However, when we compare genotype frequencies within cases stratified by age group (> or ≤ 45 years) and hormonal status (ER+) for the rs1801200, a weak association was found (p [ ≤ 45] = 0.048). Also, association of the rs1801200 polymorphism with clinical parameters assessed by two-way analysis of variance showed a significant result with tumor size (p = 0.002).

Table 3.

Association Between the rs1801200 Genotype and the Short Tandem Repeat H(AC)I4 and Clinical Parameters

	Genotype counts (Ile⁶⁵⁵Val)				STR H(AC)I4
Parameters	AA	AG	GG	Chi-square p-value	SS	SL	LL	Chi-square p-value
Tumor size
Mean	4.86	8.02	2.75	6.36	5.02	5.74	4.90	0.652
± Standard deviation	3.016	5.710	1.55	0.002^a	2.899	3.803	3.50	0.523
Lymph node
Negative (n = 52)	44	6	2	1.021	10	15	27	3.182
Positive (n = 92)	83	7	2	0.600	20	15	57	0.204
Tumor grade
SBR I/II (n = 86)	77	7	2	2.26	20	17	49	1.829
SBR III (n = 45)	36	7	2	0.32	6	10	29	0.401
Estrogen receptor
Negative (n = 43)	37	5	1	1.609	12	8	23	2.309
Positive (n = 90)	82	5	3	0.447	15	18	57	0.315
Progesterone receptor
Negative (n = 55)	47	7	1	2.21	17	9	29	6.865
Positive (n = 84)	76	5	3	0.330	11	21	52	0.032
Bcl2 protein
Negative (n = 26)	24	2	0	2.273	8	2	16	0.87
Positive (n = 58)	56	1	1	0.321	13	7	38	0.648
HER2 protein
Negative (n = 40)	34	4	2	4.789	8	6	26	0.57
Positive (n = 57)	55	2	0	0.091	15	7	35	0.750

p-value by Fisher's one-way analysis of variance test.

STR, short tandem repeat; SBR, Scarff-Bloom-Richardson.

To investigate ethnic specificities, we compared the genotype frequencies of the rs1801200 in our population with published data available for Chinese (Xie et al., 2000), African (Ghanaian) (Ameyaw et al., 2000), Afro-American (Millikan et al., 2003), British (Benusiglio et al., 2005), Portuguese (Pinto et al., 2004), Turkey (Akisiki and Dalay, 2004), Iranian (Kamali-Sarvestani et al., 2004), Slovak (Zubor et al., 2006), Korean (An et al., 2005), and American (Nelson et al., 2005).

Figure 2 shows the frequency of the Ile/Ile and Val/Ile + Val/Val genotype; we observed that we can distinguish three ethnic groups in the control sample (Fig. 2a): Caucasian, Asiatic, and African. The same pattern was found for the patient subset (Fig. 2b).

FIG. 2.

Allele frequencies of the Ile⁶⁵⁵Val polymorphism in (a) controls and (b) breast cancer patients.

The dinucleotide repeats H(AC)I4 and breast cancer

The number of AC repeats ranged from 16 to 35 among the Tunisian samples (Fig. 3). Genotype frequencies in control samples were in HWE for the [H(AC)I4] (corrected p = 0.44).

FIG. 3.

Allelic distribution of the H(AC)I4 repeats.

We subdivided the alleles into two classes: short (S) (≤25 repeats) and long (L) (>25 repeats). A highly significant association was found between the short tandem repeat [H(AC)I4] screened and breast cancer risk at genotypic and allelic levels (p = 0.0004 and p = 0.0001, respectively) (Table 4). Particularly, we found that the subjects having two copies of the H(AC)I4 allele with >25 repeats showed 2.44-fold increased risk for breast cancer (OR = 2.44; p = 0.00017; 95% CI: 1.53, 3.91) compared with subjects having one or both alleles with ≤25 repeats. The risk also tended to increase when we compared the subject having two copies of the H(AC)I4 allele with >25 repeats and those having one copy of this polymorphism (OR = 2.93; p = 0.00013; 95% CI: 1.67, 5.14) (Table 4).

Table 4.

Allelic and Genotypic Distribution for the H(AC)I4 Microsatellite Marker

	Cases		Controls
Genotypes	n	%	n	%	OR	95% CI
H(AC)I4
Both alleles >25	88	59.4	54	37.5	2.44^a	1.53-3.91
LL
One allele >25	30	20.3	54	37.5	2.93^b	1.67-5.14
SL
Both alleles ≤25	30	20.3	36	25	1.31^c	0.76-2.27
SS
Genotypic association			p = 0.0004
Both alleles >25	88		54		2.44^d	1.53-3.91
Any alleles ≤25	60		90
Allelic association			p = 0.0001

[LL] versus [SL + SS].

[LL] versus [SL].

[LL + SL] versus [SS].

[L] versus [S].

OR, odds ratio; CI, confidence interval.

Further, binary logistic regression of disease status on these three polymorphisms confirmed the association of the microsatellite polymorphism with the risk of breast cancer (p = 0.016). However, no association was found between H(AC)I4 and any of the clinical parameters considered in our study (after correction for multiple testing) (Table 3).

Discussion

HER2 is the second member of the EGFR family which has emerged over the past 20 years to be one of the most important proto-oncogenes in invasive breast cancer. In fact, breast cancer is the most widespread malignant disease affecting women. Its incidence has a slight growing potential. Given the determining role of HER2 overexpression in breast cancer pathogenesis and potential tumorigenicity of point mutation in the transmembrane-encoding region of HER2 gene in animal model, any polymorphism in this gene region might conceivably be a candidate gene for human breast cancer.

The most studied polymorphism of such type has been the Ile⁶⁵⁵Val polymorphism (rs1801200). However, until now no convincing evidence for any association between this polymorphism in HER2 and breast cancer risk has been reported.

A study by Fleishman et al. (2002) stated that the presence of isoleucine in codon 655 destabilizes the formation of active HER2 heterodimers, even in the presence of receptor overexpression. In contrast, the presence of valine can render the receptor in an active state, with slower endocytosis and rapid receptor recycling. But, Pinto et al. (2005) hypothesized that GG (Val/Val) genotype encodes receptors with higher dimerization capacities and possibly constitutive activation. However, use of bioinformatics tools for prediction of SNP effect on protein function failed to find any functional impact of this polymorphism (Choura and Rebai, 2009).

Several investigators have examined the relationship of this polymorphism with breast cancer risk (Xie et al., 2000; Montgomery et al., 2003; Kamali-Sarvestani et al., 2004; An et al., 2005; Benusiglio et al., 2005; Kalemi et al., 2004; Nelson et al., 2005; Benusiglio et al., 2006; Zubor et al., 2006). There have been inconsistent findings from one study to another. One population-based case-control study among Chinese women suggested that Val/Val homozygosity at erbB-2 codon 655 was associated with increased risk of breast cancer among younger women less than 45 years of age (Xie et al., 2000). Other studies (British women, Caucasian subjects, African-American, Korean, and Iranian), however, stated that the valine allele was not a breast cancer risk factor (Montgomery et al., 2003; Kamali-Sarvestani et al., 2004; An et al., 2005; Benusiglio et al., 2006). Our study also showed no association between HER2 Ile⁶⁵⁵Val polymorphism and risk of breast cancer development. This is in agreement with most published studies (Montgomery et al., 2003; Kamali-Sarvestani et al., 2004; An et al., 2005; Benusiglio et al., 2005; Benusiglio et al., 2006; Zubor et al., 2006). However, our study showed that the frequency of Val/Ile in Tunisian controls and patients is similar to that in African-American women and Turkish, but higher than that in Caucasian (Slovak, American, British, White, and Portuguese). Therefore, this difference in distribution can be explained by a different genetic background.

No correlation was found between HER2 polymorphisms analyzed in our study with histological grade or lymph node status. By contrast, Papadopoulou et al. (2007) reported a significant association in Greek samples with these parameters. However, the tumor size was found significantly associated with this polymorphism in our study. The same result was observed in gastric cancer Japanese patients (Kuraoka et al., 2003).

No significant association was found between HER2, Bcl2 protein, and estrogen/progesterone receptor and the SNPs studied in our populations. This polymorphism does not seem to alert or increase the expression of proteins implicated in breast cancer mechanisms. The same results were reported by Kamali-Sarvestani et al. (2004) in an Iranian population. Conversely, in the African-American population, Millikan et al. (2003) observed an insignificant trend for HER2 overexpression in women with Val/Val genotype. Puputti et al. (2006) suggested that in some HER2-positive breast tumors, the Val allele is overrepresented and could be potentially advantageous in the progression of breast cancer. These inconsistent results cannot be clearly explained, but it may be possibly due to a difference in the extent of the disequilibrium among different ethnic groups (Hauptmann et al., 2003).

It has been shown that the first intron in several genes, including EGFR, plays an important regulatory role (Bornstein et al., 1988; Sica et al., 1992), particularly the AC repeat (Maekawa et al., 1989). In fact, polymorphic dinucleotide repeat are rarely found in the coding regions of genes and are very often located in intergenic noncoding regions or in introns (Akai et al., 1999). Further, it was suggested that these repeats have some physiological functions, such as regulation of gene expression (Kashi et al., 1997; Sharma et al., 2007). The dinucleotide repeats H(AC)I4 is the only AC repeat found by homology in the HER2 gene in intron 4 (Kharrat et al., 2008). In our study, this polymorphism proved a significant association with breast cancer. This polymorphic region may be involved in conformational changes after transcription factor binding, as reported by Kharrat et al. (2008). In fact, it was observed that (TG/CA)_n repeats of length n ≥12 units exert a downregulatory effect on transcription. Moreover, this negative modulatory effect increases with increasing length of repeat. Indeed, we observed that genes with long introns harbor more repeats compared with genes with short introns, suggesting a strong role of intron length as controlling factor for abundance of repeats (Sharma et al., 2007). These preliminary results deserve further confirmation.

Conclusion

Our results suggest that the well-studied Ile⁶⁵⁵Val genetic polymorphism of HER2 is not associated with an increased risk of breast cancer in Tunisian women. Also, we did not find an association between this polymorphism and major pathological and clinical features such as tumor grades, node involved, ER, PR, and bcl-2, and HER2 status of the disease. However, a new microsatellite in intron 4 showed promising results with a high association with breast cancer risk. Further investigations are needed to replicate this initial finding and assess its potential as a risk predictor.

Footnotes

Acknowledgments

This work was supported by the Ministry of Higher education, Scientific Research and Technology, Tunisia, and by the Kuwait University/Research Administration (grant no. YN01/05).

Disclosure Statement

No competing financial interests exist.

References

Akai

, Kimura

, Hata

. 1999. Transcriptional regulation of the human type I collagen alpha2 (COL1A2) gene by the combination of two dinucleotide repeats. Gene, 239:65-73.

Akisiki

, Dalay

. 2004. Estrogen receptor codon 594 and HER2 codon 655 polymorphisms and breast cancer risk. Exp Mol Pathol, 76:260-263.

Akiyama

, Sudo

, Ogawara

et al. 1986. The product of the human c-erbB-2 gene: a 185-kilodalton glycoprotein with tyrosine kinase activity. Science, 232:1644-1646.

Ameyaw

, Thornton

, McLeod

. 2000. Re: population-based, case-control study of HER2 genetic polymorphism and breast cancer risk. J Natl Cancer Inst, 92:1947

, Kim

, Oh

et al. 2005. Her2 genotype and breast cancer progression in Korean women. Pathol Int, 55:48-52.

Bargmann

, Hung

, Weinberg

. 1986. The neu oncogene encodes an epidermal growth factor receptor-related protein. Nature, 319:226-230.

Benusiglio

, Lesueur

, Luccarini

et al. 2005. Common ERBB2 polymorphisms and risk of breast cancer in a white British population: a case-control study. Breast Cancer Res, 7:204-209.

Benusiglio

, Pharoah

, Smith

et al. 2006. HapMap-based study of the 17q21 ERBB2 amplicon in susceptibility to breast cancer. Br J Cancer, 95:1689-1695.

Bloom

, Richardson

. 1957. Histological grading and prognosis in breast cancer: a study of 1409 cases of which 359 have been followed for 15 years. Br J Cancer, 11:359-377.

10.

Bornstein

, McKay

, Liska

et al. 1988. Interactions between the promoter and first intron are involved in transcriptional control of alpha 1(I) collagen gene expression. Mol Cell Biol, 8:4851-4857.

11.

Fleishman

, Schlessinger

, Ben-Tal

. 2002. A putative molecular-activation switch in the transmembrane domain of erbB2. Proc Natl Acad Sci USA, 99:15937-15940.

12.

Gregory

, Powles

, Salter

et al. 2000. Prognostic relevance of cerbB2 expression following neoadjuvant chemotherapy in patients in a randomised trial of neoadjuvant versus adjuvant chemoendocrine therapy. Breast Cancer Res Treat, 59:171-175.

13.

Hauptmann

, Sigurdson

, Chatterjee

et al. 2003. Re: population-based, case-control study of HER2 genetic polymorphism and breast cancer risk. J Natl Cancer Inst, 95:1251-1252.

14.

Kalemi

, Lambropoulos

, Gueorguiev

et al. 2004. The association of p53 mutations and p53 codon 72, Her 2 codon 655 and MTHFR C677T polymorphisms with breast cancer in Northern Greece. Cancer Lett, 222:57-65.

15.

Kamali-Sarvestani

, Talei

, Merat

. 2004. Ile to Val polymorphism at codon 655 of HER-2 gene and breast cancer risk in Iranian women. Cancer Lett, 215:83-87.

16.

Kashi

, King

, Soller

. 1997. Simple sequence repeats as a source of quantitative genetic variation. Trends Genet, 13:74-78.

17.

Khabir

, Ghorbel

, Daoud

et al. 2003. Similar BCL-X but different BCL-2 levels in the two age groups of North African nasopharyngeal carcinomas. Cancer Detect Prev, 27:250-255.

18.

Kharrat

, Al'Fadhli

, Rebaï

. 2008. Screening and validation of dinucleotide repeats in intron1 of human EGFR gene and its paralog in Her2 gene. J Recept Signal Transduct, 28:475-483.

19.

Kuraoka

, Matsumura

, Hamai

et al. 2003. A single nucleotide polymorphism in the transmembrane domain coding region of HER-2 is associated with development and malignant phenotype of gastric cancer. Int J Cancer, 107:593-596.

20.

Maekawa

, Imamoto

, Merlino

et al. 1989. Cooperative function of two separate enhancers of the human epidermal growth factor receptor proto-oncogene. J Biol Chem, 264:5488-5494.

21.

McKean-Cowdin

, Kolonel

, Press

et al. 2001. Germ-line HER-2 variant and breast cancer risk by stage of disease. Cancer Res, 61:8393-8394.

22.

Millikan

, Eaton

, Worley

et al. 2003. HER2 codon 655 polymorphism and risk of breast cancer in African Americans and whites. Breast Cancer Res Treat, 79:355-364.

23.

Montgomery

, Gertig

, Baxter

et al. 2003. The HER2 I655V polymorphism and risk of breast cancer in women <age 40 years. Cancer Epidemiol Biomarkers Prev, 12:1109-1111.

24.

Nelson

, Gould

, Hampton

et al. 2005. A case-control study of the HER2 Ile655Val polymorphism in relation to risk of invasive breast cancer. Breast Cancer Res, 7:357-364.

25.

Normanno

, De Luca

, Bianco

et al. 2006. Epidermal growth factor receptor (EGFR) signaling in cancer. Gene, 366:2-16.

26.

Papadopoulou

, Simopoulos

, Tripsianis

et al. 2007. Allelic imbalance of HER-2 codon 655 polymorphism among different religious/ethnic populations of northern Greece and its association with the development and the malignant phenotype of breast cancer. Neoplasma, 54:365-373.

27.

Papewalis

, Nikitin

AYu

, Rajewsky

. 1991. G to A polymorphism at amino acid codon 655 of the human erbB-2/HER2 gene. Nucleic Acids Res, 19:5452

28.

Pegram

, Finn

, Arzoo

et al. 1997. The effect of HER-2/neu overexpression on chemotherapeutic drug sensitivity in human breast and ovarian cancer cells. Oncogene, 15:537-547.

29.

Pinto

, Pereira

, Portela

et al. 2005. The influence of HER2 genotypes as molecular markers in ovarian cancer outcome. Biochem Biophys Res Commun, 335:1173-1178.

30.

Pinto

, Vasconcelos

, Costa

et al. 2004. HER2 polymorphism and breast cancer risk in Portugal. Eur J Cancer Prev, 13:177-181.

31.

Prenzel

, Fischer

, Streit

et al. 2001. The epidermal growth factor receptor family as a central element for cellular signal transduction and diversification. Endocr Relat Cancer, 8:11-31.

32.

Puputti

, Sihto

, Isola

et al. 2006. Allelic imbalance of HER2 variant in sporadic breast and ovarian cancer. Cancer Genet Cytogenet, 167:32-38.

33.

Sharma

, Kumar

, Brahmachari

et al. 2007. Abundance of dinucleotide repeats and gene expression are inversely correlated: a role for gene function in addition to intron length. Physiol Genomics, 31:96-103.

34.

Shigematsu

, Takahashi

, Nomura

et al. 2005. Somatic mutations of the HER2 kinase domain in lung adenocarcinomas. Cancer Res, 65:1642-1646.

35.

Sica

, Tan

, Rice

et al. 1992. The c-rel protooncogene product c-Rel but not NF-kappa B binds to the intronic region of the human interferongamma gene at a site related to an interferon-stimulable response element. Proc Natl Acad Sci USA, 89:1740-1744.

36.

Slamon

, Clark

, Wong

et al. 1987. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science, 235:177-182.

37.

Stern

, Kamps

, Cao

. 1988. Oncogenic activation of p185neu stimulates tyrosine phosphorylation in vivo. Mol Cell Biol, 8:3969-3973.

38.

Tommasi

, Fedele

, Crapolicchio

et al. 2003. ErbB2 and the antimetastatic nm23/NDP kinase in regulating serum induced breast cancer invasion. Int J Mol Med, 12:131-134.

39.

Wiseman

, Makretsov

, Nielsen

et al. 2005. Coexpression of the type 1 growth factor receptor family members HER-1, HER-2, and HER-3 has a synergistic negative prognostic effect on breast carcinoma survival. Cancer, 103:1770-1777.

40.

Witton

, Reeves

, Going

et al. 2003. Expression of the HER1-4 family of receptor tyrosine kinases in breast cancer. J Pathol, 200:290-297.

41.

Xie

, Shu

, Deng

et al. 2000. Population-based, case-control study of HER2 genetic polymorphism and breast cancer risk. J Natl Cancer Inst, 92:412-417.

42.

Yarden

, Sliwkowski

. 2001. Untangling the ErbB signalling network. Nat Rev Mol Cell Biol, 2:127-137.

43.

Zubor

, Vojvodova

, Danko

et al. 2006. HER-2 [Ile655Val] polymorphism in association with breast cancer risk: a population-based case-control study in Slovakia. Neoplasma, 53:49-55.