Introduction

Breast cancer (BC) is second leading cause of cancer deaths worldwide and approximately 1.7 million new cases are being diagnosed every year and 521,900 deaths occurred in 2012 alone globally1. It has been estimated that 252,710 new cases of invasive breast cancer will be diagnosed in 2017 among women in US alone. Although, breast cancer is most common in females it also rarely diagnosed in male individuals and 2,470 males are estimated to be diagnosed with breast cancer in 2017 in United States2. Among the overall cancer deaths worldwide, approximately 60% of deaths occur in developing countries including India. In Indian woman, majority of cancer related deaths are due to breast cancer1.

BC is highly heterogeneous and ~60–70% is of estrogen receptor positive which responds to anti-hormone therapy3. Estrogen receptor (ER) plays an important role in breast cancer progression and treatment. Approximately 20–30% breast cancers are of Human epidermal growth factor receptor2 (Her2) positive and are highly aggressive in nature4. High levels of Her2 expression was also observed in tamoxifen resistant breast cancers. Human epidermal growth factor receptor family members are a group of molecules having tyrosine kinase activity with no natural ligand found till date. Heterodimerization among the family members leads to autophosphorylation of cytoplasmic domain which leads to cell proliferation5,6,7. Her2 is highly expressed in various cancers types viz. breast, endometrial, ovarian, colon, lung, prostrate and cervical cancers. Role of ERBB2/Her2 in physiological processes (cell growth, differentiation and tissue development) as well as in carcinogenesis and metastasis has been well investigated8,9,10. Her2 plays major role in the regulation of several pathways such as Raf/Ras/MAPK and PI3K/AKT pathways11. Receptor mediated signaling pathways has pivotal role in the regulation of normal cell function, growth and division. However disruption of these pathways might lead to several cancers12,13,14,15. Her2 positive breast cancers show poor survival rate, treatment with tyrosine inhibitors showing promising results in harboring these aggressive tumors16. Trastuzumab a monoclonal antibody specifically binds to Her2 and disrupts the downstream pathways of Her2 and it is effectively used for the treatment of Her2 positive breast cancers17,18,19. However several patients developed resistance to trastuzumab over a period of time20. Recent studies suggested that Her2 Ile655Val polymorphism is associated with cardiac toxicity. Moreover, it has been identified that both the Her2 Ala1170Pro polymorphisms also responsible for increasing the risk of cardiac toxicity in women administrated with trastuzumab21,22.

Genetic epidemiological studies indicated association between single nucleotide polymorphisms and different cancers23,24,25. Cell cycle regulatory role of Her2 and its importance in prognosis of breast cancer clearly indicates that polymorphism in coding region of Her2 might be associated with either cancer susceptibility risk or resistance. One such single nucleotide transition mutation in transmembrane domain coding region of Her2 at codon 655 [Isoleucine (Ile) to Valine (Val) mutation, Her2 Ile655Val] was well investigated in different populations in relation to risk of breast cancer26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61. Milikan et al.40 reported the association of Valine allele at Her2 655 codon with breast cancer risk. Whereas, Baxter et al.29 and Xie et al.60 found no association with breast cancer risk in women aged <40 years, post menopausal respectively. However, few researchers performed meta-analysis and tried to conclude the possible correlation of Her2 polymorphism with breast cancer risk. Tao et al.56 showed no association in overall analysis, however mild association of Her2 polymorphism with susceptibility to breast cancer in Asian ethnic group was suggested. Another meta-analysis by Chen et al.62 including 32 case control studies revealed comparable distribution of Her2 Ile655Val variants among cases and controls in Caucasian, American and European population. Interestingly, Asian ethnic group showed significant association of breast cancer risk with Her2 Ile655Val polymorphism. In the present meta-analysis, total of 35 case-control studies were analyzed and investigated for possible association of Her2 Ile655Val polymorphism with development of breast cancer. Furthermore, we subgrouped included reports according to ethnicity and the association was analyzed .

Results

Characteristics of eligible studies

To understand association of Her2 Ile655Val polymorphism with breast cancer risk, we have performed meta-analysis using 35 case-control eligible studies including 19, 220 cases and 22, 306 controls. Genotype and allele frequency for case and control of each eligible study was extracted and the characteristics of each study are shown in Table 1. For subgroup analysis the identified studies were categorized based on their ethnicity viz. Caucasian, American, Afro-American, African, European and Asian respectively.

Table 1 Characteristics and distribution of Her2 polymorphism in each study involved in meta-analysis.

Heterogeneity test

To evaluate the heterogeneity among the studies Q test with I2 statistics were used. I2 more than 50 (I2 > 50) with significant p-value (p < 0.05) considered to be presence of heterogeneity among included studies. Among the models tested, heterogeneity was observed in allele comparison, heterozygous and dominant genetic models. However, other genetic comparison models such as recessive and homozygous were homogeneous. Observations of heterogeneity Q test and I2 statistics of each model are shown in Table 2. Based on results of heterogeneity test, fixed or random effect model was used for meta-analysis.

Table 2 Statistics for heterogeneity analysis and publication bias.

Publication bias

Begg’s funnel plot and egger’s regression test was performed to assess the publication bias within the studies included in meta-analysis. Results are imputed in Table 2. We observed significant publication bias in all genetic models tested and were resolved by “trim and fill” technique (Supplementary Fig. 1).

Statistical analysis

In the present study 35 case-control studies were included and cumulative analysis demonstrated the association of Her2 polymorphism with increased risk of breast cancer. The overall allele model revealed association between Her2 polymorphism and breast cancer risk with Odds ratio (OR) = 1.130, 95% confidence interval (CI) = 1.051–1.216, p = 0.001. Furthermore, both dominant and heterozygous models showed significant association of Her2 Ile655Val polymorphism with increased risk of breast cancer (Dominant model Val-Val + Ile-Val vs Ile-Ile: OR = 1.127, 95% CI = 1.038–1.223, p = 0.004; Heterozygous Ile-Val vs Ile-Ile: OR = 1.100, 95% CI = 1.016–1.192, p = 0.019). However, comparison of genotypes in other genetics models didn’t show significant association (Homozygous Val-Val vs Ile-Ile: OR = 1.034, 95% CI = 0.937–1.142, p = 0.503; Recessive Val-Val vs Ile-Ile + Ile-Val: OR = 1.041, 95% CI = 0.945–1.147, p = 0.418) (Figs 13). Furthermore, Studies were grouped based on the techniques used for the detection of polymorphism and were analyzed for the association with breast cancer. Studies which used RFLP method as genotypic detection were showing significant association with breast cancer risk in all the models (Overall allele Val vs Ile: OR = 1.236, 95% CI = 1.091–1.400, p = 0.001; Homozygous Val-Val vs Ile-Ile: OR = 1.177, 95% CI = 0.017–1.362, p = 0.028; Heterozygous Ile-Val vs Ile-Ile: OR = 1.183, 95% CI = 1.026–1.364, p = 0.021; Recessive Val-Val vs Ile-Ile + Ile-Val: OR = 1.192, 95% CI = 1.033–1.375, p = 0.016; Dominant model Val-Val + Ile-Val vs Ile-Ile: OR = 1.233, 95% CI = 1.066–1.424, p = 0.005) (Figs 4 and 5). However, the studies in which Taqman used as detection method showed no association with increased risk of breast cancer (Fig. 6).

Figure 1
figure 1

Forest plot: Overall allele and genotypic (Homozygous: Val-Val vs. Ile-Ile) analysis of Her2 Ile655Val (rs1136201) gene polymorphism and validation of it’s association with breast cancer risk. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR).

Figure 2
figure 2

Forest plot: Genotypic (Heterozygous: Ile-Val vs Ile-Ile and Recessive: Val-Val vs. Ile-Ile + Ile-Val) analysis of Her2 Ile655Val (rs1136201) gene polymorphism and investigation of it’s association with breast cancer risk using OR with 95% CI.

Figure 3
figure 3

Forest plot: Genotypic (Dominant: Val-Val + Ile-Val vs Ile-Ile) analysis of Her2 Ile655Val (rs1136201) gene polymorphism and evaluation of its association with increased risk of breast cancer. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR).

Figure 4
figure 4

Forest plot: Overall allele and genotypic (Homozygous: Val-Val vs. Ile-Ile and Heterozygous: Ile-Val vs Ile-Ile) analysis of studies in which RFLP used as detection method for Her2 Ile655Val (rs1136201) gene polymorphism and evaluation of its association with breast cancer risk. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR).

Figure 5
figure 5

Forest plot: Genotypic (Recessive: Val-Val vs. Ile-Ile + Ile-Val; Dominant: Val-Val + Ile-Val vs Ile-Ile) analysis of studies in which RFLP used as detection method for Her2 Ile655Val (rs1136201) gene polymorphism and evaluation of its association with increased risk of breast cancer. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR).

Figure 6
figure 6

Forest plot: Overall allele and genotypic analysis of studies in which Taqman used as detection method for Her2 Ile655Val (rs1136201) gene polymorphism and evaluation of its association with breast cancer risk. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR). Overall allele Val vs Ile: OR = 0.990, 95% CI = 0.944–1.039, p = 0.695; Homozygous Val-Val vs Ile-Ile: OR = 0.927, 95% CI = 0.811–1.060, p = 0.270; Heterozygous Ile-Val vs Ile-Ile: OR = 1.004, 95% CI = 0.946–1.066, p = 0.890; Recessive Val-Val vs Ile-Ile + Ile-Val: OR = 0.928, 95% CI = 0.813–1.059, p = 0.266; Dominant model Val-Val + Ile-Val vs Ile-Ile: OR = 0.997, 95% CI = 0.941–1.055, p = 0.904.

Subgroup analysis

As the previous meta-analysis presented association of Her2 gene polymorphism with susceptibility to breast cancer in Asian population only, in the present analysis we re-accessed possible link of Her2 polymorphism with BC in different ethnic groups. In our study subgroup analysis with 15 case-control studies identified the association with increased risk of breast cancer in Asian ethnicity in overall allele and dominant models. Similarly, African group with 3 successful included case-control studies also showed association with breast cancer risk in recessive and homozygous models. However, 5 case control studies from Caucasian, 4 from American subgroup, 2 limited studies from Afro-American ethnic group and 8 studies from European ethnicity showed no association of Her2 polymorphism with breast cancer risk in all the models (Table 3) (Figs 712).

Table 3 Subgroup analysis of Her2 Ile 655 Val polymorphism and its association with breast cancer risk.
Figure 7
figure 7

Forest plot: Analysis of Her2 Ile655Val (rs1136201) gene polymorphism data from Caucasian ethnic group and validation of its correlation with breast cancer susceptibility using OR with 95% CI. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR).

Figure 8
figure 8

Forest plot: Her2 Ile655Val (rs1136201) gene polymorphism data from American ethnic group showing OR and 95% CI for analyzing its association with breast cancer risk. Squares represents OR and horizontal line represents 95% Confidence Interval (CI) of odds ratio (OR).

Figure 9
figure 9

Forest plot: Her2 Ile655Val (rs1136201) gene polymorphism data from Afro-American sub group population showing OR and 95% CI for validating its association with breast cancer risk. Black square represents OR and horizontal line representing 95% CI.

Figure 10
figure 10

Forest plot: Overall analysis of African ethnic group Her2 Ile655Val (rs1136201) gene polymorphism data for evaluation of its association with breast cancer susceptibility. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR).

Figure 11
figure 11

Forest plot: Overall analysis of Her2 Ile655Val (rs1136201) gene polymorphism from European subgroup with OR and 95% CI for investigating the association with breast cancer risk. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR).

Figure 12
figure 12

Forest plot: Overall analysis of Her2 Ile655Val (rs1136201) gene polymorphism data from Asian ethnic group for the evaluation of association with breast cancer susceptibility. Black squares represent the value of OR and horizontal line indicates 95% Confidence Interval (CI) of odds ratio (OR).

Sensitivity analysis

We analyzed the influence of each individual study on the pooled OR by sensitivity analysis. One study was excluded each time and meta-analysis was performed. The results showed no individual study affected the pooled OR significantly, suggesting this meta-analysis is relatively credible, stable and not dependent on any individual study (Figs 1315).

Figure 13
figure 13

Sensitivity analysis: Sensitivity analysis showing no effect of single study on odds ratio (OR) in overall allele and genotypic (Homozygous: Val-Val vs Ile-Ile) analysis models.

Figure 14
figure 14

Sensitivity analysis: Sensitivity analysis with each study removal showing no effect on odds ratio (OR) in genotypic (Heterozygous: Ile-Val vs Ile-Ile and Recessive: Val-Val vs Ile-Ile + Ile-Val) analysis models of Her2 Ile655Val (rs1136201) gene polymorphism.

Figure 15
figure 15

Sensitivity analysis: Sensitivity analysis showing no effect of single study on OR of genotypic (Dominant: Val-Val + Ile-Val vs Ile-Ile) analysis model of Her2 Ile655Val (rs1136201) gene polymorphism.

Discussion

Human epidermal growth factor family members are a group of receptors with tyrosine kinase activity which affects cell proliferation and survival63,64. Dimerization of Her family members leads to autophosphorylation of tyrosine residues in the cytoplasmic domain and leads to cell proliferation and tumorigenesis5,6,7. Although Her family members lack natural ligand for signaling, various synthetic ligands have been developed and they are demonstrated to be efficientive in terms of drug delivery. Among the all-family members, Her2 is an important molecule and expression of Her2 is elevated in various cancers8,9,10. Approximately 20–30% breast cancers show 40–100 fold elevated levels of Her2, whereas other cancer types such as ovarian, endometrial, gastric and esophageal cancers were also detected with over-expressed Her2 protein levels65,66,67,68,69,70,71. Single nucleotide polymorphisms (SNPs) are playing an important role in various cancer types and are capable of serving as diagnostic tools23,24,25. One such single nucleotide polymorphism with substitution of isoleucine with valine at codon 655 in transmembrane region of Her2 has been found to be playing an important role in development of cancer72. The transmembrane domain region of Her2 with valine at 655 domain region stabilizes the formation of protein dimer and thus predisposing to an auto-activity of the receptor73. The hydrophobicity and conformational stability of the hydrophobic domains such as transmembrane domains may alter due to Isoleucine to valine change74. Her2 Ile655Val polymorphism was well studied for association with breast cancer risk, whereas other polymorphism at 1170 codon of Her2 (Pro1170Ala) was correlated with cardiotoxicity75. Her2 Ile655Val polymorphism is not only associated with breast cancer risk but also associated with other cancers such as ovarian and endometrial cancers76,77. However, these results are inconsistent and a stringent and powerful analysis is required to conclude the association with breast cancer. In the present study we have analyzed the association of Her2 Ile655Val polymorphism with increased breast cancer susceptibility using powerful tool comprehensive meta-analysis (CMA). Overall allele comparison genetic model results suggest that valine allele in Her2 655 codon favors the development of breast cancer in worldwide population. Heterozygous, dominant models also prove that Her2 polymorphism is associated with increased risk of breast cancer. Whereas, subgroup analysis showing different results for different ethnic population. Earlier reports by Wang et al.58 and Chen et al.62 demonstrated association of Her2 valine allele with breast cancer risk in Caucasian population. In contrast, our study failed to show such link. The present study has several advantages over earlier reports. We have included more number of studies in the current meta-analysis including larger number of cases and controls.

Tao et al.56 reported the association of Her2 polymorphism with breast cancer risk in Asian population whereas later Wang et al.58 and Chen et al.62 showed no such association with breast cancer. In this present meta-analysis, we have performed subgroup analysis and demonstrated that valine allele is associated with breast cancer risk in Asian population. In addition to that Val-Val + Ile-Val vs Ile-Ile model also prove the susceptibility of Her2 polymorphism with breast cancer. We also observed that subjects with valine/valine genotype are susceptible for the development of breast cancer in African population. These results are in agreement with the study demonstrated by Wang et al.58; however other studies failed to show such association with breast cancer susceptibility. Ethnic groups such as American, European and Afro-American are not showing such association with breast cancer risk. Our present meta-analysis includes all the studies in which either Taqman or RFLP used as detection method. Frank et al. suggested the biasness in the methods used for the detection of polymorphism and suggested that Taqman method is capable of producing false results78. We excluded the studies in which Taqman method used as detection method and performed the analysis. Studies which used other than Taqman method for the detection of polymorphism showed significant association with breast cancer risk in all models.

In conclusion, our present meta-analysis demonstrated that valine allele is susceptible in overall worldwide population and Asian ethnic group. Her2 Ile655Val polymorphism is associated with breast cancer risk in Asian, African population but not in other ethnic groups such as Caucasian, European, American and Afro-American. These results suggest that Her2 Ile655Val polymorphism could be considered as possible susceptible bio marker for the detection of breast cancer.

Materials and Methods

Literature search and identification of relevant studies

A systematic extensive search was performed to extract the appropriate published reports using online databases i.e., Pubmed, EMBASE and Google scholar. The publication search was performed by three independent authors (BMK, SC& DRM) using either single or combination of given keywords i.e., “Her2 Ile655Val polymorphism”, “Herceptin receptor polymorphism”, “rs1136201” and breast cancer. In addition to the preliminary online database search we have checked the cross references for the potential publications, those possibly missed in preliminary search. Our present study includes recently published (earliest by 2017) 35 case-control studies with 19, 220 cases and 22, 306 controls for Her2 Ile655Val polymorphism (Supplementary Fig. 2).

Inclusion and exclusion of studies

The studies which met all the criteria given below have been included in the present meta-analysis: (a) studies published in English, (b) must have case-control or cohort design, (c) have available genotype frequency of both the cases and controls or have odds ratio (OR) and 95% confidence interval (CI) values, (d) evaluating the association of Her2 Ile655Val polymorphism with breast cancer risk and (e) studies representing original data. The studies excluded based on the criteria given below: (a) studies published in other languages except English, (b) studies having only case samples, (c) representing risk of other cancers, (d) without genotypic distribution and allele frequency data and (e) reviews and abstracts.

Data extraction

The data extraction was performed by three independent authors (BMK, SC & DRM) independently and the disagreement about the studies between the authors was resolved and came to a conclusion by conducting a group discussion within the authors. We followed previously established data form to extract the data from the studies and the following data was extracted from each article: first author’s name, year of publication, country, ethnicity, number of case and control samples, genotype distribution, allele frequency for each case and control.

Meta-analysis

The current meta-analysis was performed using comprehensive meta-analysis version 3 software (CMA v3) https://www.meta-analysis.com/pages/comparisons.php. CMA v3 is a powerful tool to analyze and has several advantages over other software available for computational meta-analysis. Combined odds ratio with 95% CI was calculated and was taken into consideration to apprise the association of Her2 polymorphism with breast cancer risk. Chi-Squared based Q test was performed to analyze the heterogeneity and p-value < 0.05 was considered as significant. In case of no significant heterogeneity fixed effect model was used to assess the combined OR. In contrast, Random effect model was considered to calculate the combined odds ratio with 95% CI among the studies. I2 statistics was used to quantify inter study variability, greater I2 value depicts greater degree of heterogeneity. Publication bias was examined using Begg’s funnel plot. Egger’s linear regression test was employed to analyze and measure the asymmetry of Begg’s funnel plot and the significance of intercept was assessed by t-test. Intercept considering p-value < 0.05 was considered as significant and the publication bias was reduced using “trim and fill” method.

Availability of data and materials

All those named as authors confirmed the availability of data and materials.