The Significance of Erythroblast Transformation Specific (Ets) Transcription Factors in Breast Cancer Progression: A Meta-Analysis

Breast cancer is the most common cause of cancer-related deaths in women all over the world (Sung et al., 2021). It is a multifactorial disease that has many environmental and genetic causes (Behravan et al., 2020). A combination of genetic abnormalities and environmental factors decides the fate of the spreading tumor. Angiogenesis plays a critical part in metastasizing cancer as blood vessels are required for the nutrient supply to the growing tumor for stabilization and maintenance (Haibe et al., 2020).

Erythroblast transformation specific (Ets) family of transcription factors has a significant role in angiogenesis and hence in cancer initiation and progression (Randi et al., 2009). Several genes from this family have atypical expression during various stages of breast cancer (Hsu et al., 2004). Ets proteins, a family of mitogen-activated protein kinase (MAPK), are highly conserved proteins with a unique winged-helix turn helix DNA binding domain (Buggy et al., 2006). It is divided into 12 subfamilies or subgroups based on Ets binding domain (EBD) sequence homology. These subgroups are ETS, ERG, PEA3, ETV-2, TCF, GABP, SPI, ELF, ERF, TEL, PDEF, and ESE (Macleod et al., 1992).

The conversion of Ets factors from normal to oncogene is important in cancer progression. Many studies show that about three Ets proteins can bind to one eukaryotic gene at a time. So instead of an individual Ets factor, a dynamic regulatory network of multiple factors is involved in cancer progression. Ets family members are related to breast carcinoma by their increased or reduced expression. This abnormality in expression leads to migration, invasion, angiogenesis, cell growth, and adhesion of tumor cells. They interact with other transcription factors like p53, N-MYC, GATA and disturb cell homeostasis (Oikawa et al., 1999). Chromosomal alterations and rearrangements like gene amplification, deletions, and translocations in Ets family members lead to abnormal expressions. Several dysregulations in breast cancer include upregulation of ETS-1, increased level of PEA-3 along with HER-2 overexpression of ESE-1 in invasive ductal carcinoma, and reduced expression of PDEF and FLI-1 in invasive breast cancer tissues (Watson et al., 2020).

The diverse results of multiple studies related to a medical question often make clinical decisions difficult. The same is true in this case as there are quite some studies with different results which make it difficult to formulate a statement about the role of Ets factors in breast cancer. Therefore, a meta-analysis was performed to collect information from different studies and determine the relationship between ETS factors and breast cancer.

The electronic search was performed in databases, PubMed and Google Scholar for identification of the studies related to Ets transcription factors expression in breast cancer. Keywords used were Ets transcription factors; Ets expression; Breast cancer; Breast carcinoma and Breast tumors. All Ets transcription family factors were also searched individually. Studies published before December 2020 were included while there was no lower date limit. Appropriate references of retrieved studies were also searched for data.

For the selection of literature following criteria were followed; (i) only original and independent studies were included for analysis, (ii) samples used in the study were of human breast tissue or human breast cell line, (iii) patients must be diagnosed originally with breast malignancy, (iv) expression of Ets transcription factors must be checked, and (v) number or percentage for positive expression of Ets or odds ratio with 95% Confidence Interval must be mentioned. Studies lacking this information, having samples taken from other than human breast tissue like mice, rabbits, or other animals, and review articles were excluded from the analysis.

A total of 250 articles were selected as a result of a preliminary search through databases. Out of these, 177 studies that were irrelevant to our research were excluded from further evaluation. 73 relevant studies were assessed at the abstract level. 58 retained studies were assessed for further full-text assessment. After full-text assessment 26 studies were selected for performing meta-analysis in which complete required information was given.

Data extracted from selected articles included first author name, year of publication, the country where the study was conducted, name of Ets transcription factor, mean age of patients (if mentioned), several breast tissue samples, and positive expression of Ets transcription factor in both cancer and normal tissue, odds ratio with upper and lower limit with 95% confidence interval (CI), methods of expression analysis like immunohistochemistry, western blotting, qPCR, etc., relation or effect of Ets factor on other genes expression and type of breast cancer like HER2 positive, etc. Supplementary data were also retrieved if required information was not mentioned in the article.

The expression of Ets transcription factors was measured by calculating the odds ratio and standard error. For analyzing the significance of expression, a forest plot for the random effect model was used. The weight and residual of each study used in the meta-analysis were also calculated. A subgroup analysis was performed on the individual Ets factor family members. Bias in studies was checked by funnel plots followed by Egger’s regression and the Begg-Mazumdar test. Heterogeneity within the study was estimated using I-squared (Oikawa et al., 2003) and between the studies, variation was checked by the Tau2 statistics (Krishnamoorthy and Lee, 2014). All calculations were done with Comprehensive Meta-Analysis Version 3.0.

After thoroughly analyzing 250 studies finally, 26 studies were selected for performing meta-analysis in which complete required data was given (Supplementary Fig. 1). Supplementary Table S1 summarizes the characteristics of the included studies. In 10 studies the expression of the ETS factor was analyzed at the mRNA level (Benz et al., 1997; Ghadersohi and Sood, 2001; Kinoshita et al., 2002; Span et al., 2002; Tognon et al., 2002; Bièche et al., 2004; Chotteau-Lelièvre et al., 2004; Katayama et al., 2005; Buchwalter et al., 2013; Kar and Gutierrez-Hartmann, 2017), in 14 studies it was at protein level (Behrens et al., 2001; Mitas et al., 2001; Fleming, 2004; Myers et al., 2005, 2006; Xia et al., 2006; Sood et al., 2007; Turcotte et al., 2007; Sood et al., 2009; Zhang et al., 2011; Laliotis et al., 2013; Mesquita et al., 2013; Puzovic et al., 2014; Yuan et al., 2014) while in remaining 2 studies expressions was analyzed both at mRNA and protein levels (Yuan et al., 2014). Studies included in this systematic review illustrated 11 Ets transcription factors having important roles in breast cancer. These Ets transcription factors are ETS-1, ETS-2, ELK-1, ERM, ETV-4, PDEF, ELF-3, ETV-6, ETV-3, SPDEF, and ELK-4. All studies were about a single Ets transcription factor except (Myers et al., 2005; Mesquita et al., 2013) in which more than one Ets factor was evaluated. The individual study sample size ranged from 13 to 364 patients. The mean patient age was 48-64 years. In two studies odds ratio was directly calculated (Turcotte et al., 2007; Puzovic et al., 2014). In most studies, RT-PCR was used to evaluate mRNA expression (Benz et al., 1997; Ghadersohi and Sood, 2001; Kinoshita et al., 2002; Span et al., 2002; Tognon et al., 2002; Bièche et al., 2004; Chotteau-Lelièvre et al., 2004; Katayama et al., 2005; Buchwalter et al., 2013; Kar and Gutierrez-Hartmann, 2017), while for protein analysis immunohistochemistry (Behrens et al., 2001; Mitas et al., 2001; Buggy et al., 2004, 2006; Fleming, 2004; Myers et al., 2006; Xia et al., 2006; Zhang et al., 2011; Laliotis et al., 2013; Mesquita et al., 2013; Puzovic et al., 2014; Yuan et al., 2014), western blotting (Behrens et al., 2001; Buggy et al., 2004; Sood et al., 2007, 2009; Turcotte et al., 2007), ELISA (Buggy et al., 2004, 2006), and other methods (Benz et al., 1997; Myers et al., 2005) were used. The most common types of carcinoma in selected studies were ductal and lobular invasive carcinomas. One study involved a very rare subtype that is secretory breast cancer (Tognon et al., 2002). Breast cancer subtypes in these studies were triple-negative (Yuan et al., 2014), luminal (Buchwalter et al., 2013), and HER2 +ve breast cancer. In 7 studies, the subtype was not specified (Ghadersohi and Sood, 2001; Kinoshita et al., 2002; Bièche et al., 2004; Myers et al., 2005, 2006; Zhang et al., 2011; Puzovic et al., 2014).

The studies for which meta-analysis was performed, the expression of Ets transcription factors was closely related to breast cancer. In 19 studies, the increased expression of Ets factors was observed in breast cancer patients (Benz et al., 1997; Kinoshita et al., 2002; Mitas et al., 2001; Tognon et al., 2002; Bièche et al., 2004; Chotteau-Lelièvre et al., 2004; Fleming, 2004; Katayama et al., 2005; Buggy et al., 2006; Myers et al., 2006; Sood et al., 2007; Buchwalter et al., 2013; Laliotis et al., 2013; Mesquita et al., 2013; Puzovic et al., 2014; Yuan et al., 2014; Kar and Gutierrez-Hartmann, 2017). In 9 studies the expression was almost equal in both cases (Ghadersohi and Sood, 2001; Span et al., 2002; Buggy et al., 2004; Myers et al., 2005; Turcotte et al., 2007; Sood et al., 2009; Zhang et al., 2011) while in 2 studies expressions of ETS factor in breast cancer samples were less than that of normal samples (Buggy et al., 2004; Xia et al., 2006).

A random effect model was chosen to calculate the odds ratio and heterogeneity. OR (Odds Ratio) value of individual studies and forest plot is illustrated in Supplementary Fig. 2. Combined odds ratio was OR = 3.185, 95% CI (2.161-4.693) and P < 0.001. This shows a statistically significant relationship between ETS factors overexpression and breast cancer risk. The I-square statistics to measure heterogeneity between studies showed I2= 86% meaning that 86% of the observed variance between studies is due to the real difference in effect size and only 14% of observed variance should be expected to base on a random error. The tau2 value to measure the variance among studies was 0.841. It was observed that this heterogeneity was due to studies (Behrens et al., 2001; Span et al., 2002; Tognon et al., 2002; Katayama et al., 2005; Xia et al., 2006; Mesquita et al., 2013), therefore a meta-analysis was performed excluding these studies. The combined OR after excluding these studies is OR = 2.564, 95% CI (1.88-3.48) and P < 0.001. After excluding these studies the I2 and tau2 values reduced to 76% and 0.369, respectively (Fig. 1).

Bias in studies was calculated by funnel plot for standard error of random effect model as shown in Figure 2 and Supplementary Figure 2. The symmetrical plot indicates that there are no biases in studies included in the meta-analysis. The Egger’s regression test and Begg and Mazumdar test details are shown on funnel plots. Sub-group analysis was performed for the individual Ets- family factors to minimize heterogeneity among the included studies. In each specific group, the effect of individual Ets factor was evaluated on breast cancer. The forest plots for Ets-1 (OR = 2.149, 95% CI = 1.141 – 4.048, P = 0.018), Ets-2 (OR = 3.06, 95% CI = 1.226 – 7.648, P = 0.017) and ETV-4 (OR = 2.885, 95% CI = 0.779 – 10.684, P = 0.113) are shown in Figure 4a, b, c. The funnel plots for publication bias were also determined and shown in Figures 3 and 4. All reported p values were two-sided and p values < 0.05 were regarded as statistically significant.

The progression of breast cancer is rather unpredictable due to the greater variation and heterogeneity in the underlying causes (Kar et al., 2017). It is, therefore, particularly important to identify the biomarkers that can predict the progression of the disease. Several molecular targets have been identified which can serve as potential biomarkers for breast cancer but still there is a need to find better targets for metastatic breast cancer (Polyak, 2011). Ets family of transcription factors have emerged to play a significant role in the progression of breast malignancy. In the present study, a meta-analysis of 26 studies was performed after scrutiny to determine the relationship between Ets factors overexpression and breast cancer progression. A thorough literature survey could not find any study which has comprehensively analyzed and reviewed the Ets factors and breast cancer occurrence.

The random effect model was selected for the meta-analysis, as this model allowed examining the true variation in effect size (odds ratio) among individual studies. Effect size can be slightly higher or lower according to the characteristics or condition of subjects in a study. In combined effect size, the random effect model gives the mean effect size for all studies and more precise effect size from studies having a large number of samples or patients as compared to small sample size studies can be obtained (Li et al., 2002; Walker et al., 2008). The publications finalized in our meta-analysis had variable Ets factors for breast cancer patients, so the random effect model was found to be more appropriate. In contrast, the fixed-effect model was not selected because it assumes the same effect size for all studies.

The meta-analysis results showed that Ets transcription factors family overexpression increases the odd ratios of breast cancer in a significant way (combined OR= 3.262). However, the Ets family factor Ets-1 was found more closely related to breast cancer occurrence (OR= 2.149) compared to others.

Multiple pieces of evidence suggested the regulation of breast cancer metastasis by the combined action of several Ets factors affecting various pathways. Ets factors are also associated with the poor prognosis of breast cancer. The possible mechanism of their action on breast cancer progression is maybe through the angiogenesis pathway (Folkman, 1995). Among the selected 26 studies the increased expression of Ets transcription factors was observed in all types of breast cancer cases as compared to respective controls. However, there was one study (Xia et al., 2006) showing the downregulation of the Ets factor, PEA3 in HER-2/neu breast metastasis and was not associated with clinicopathological features.

Ets family of transcription factors regulates the expression of numerous signaling molecules and regulators of tumor progression. The pathways of the tumor microenvironment and their interactions are also under influence of Ets factors (Haidich, 2010). It is evident from the literature that some Ets factors have coordinated functions and control tissue homeostasis for the tumor microenvironment. Ets factors altered expression in breast tumorigenesis is determined in several studies. Some Ets factors are overexpressed while others are down-regulated during breast tumorigenesis thus acting as both activators and suppressors of the process.

Our literature search and meta-analysis results have demonstrated that Ets-1 is the most investigated factor among all Ets proteins. The subgroup analysis was performed with Ets-1, Ets-2, and ETV-4 only because for other factors number of studies was not enough (3 or less). Among these factors, the Ets-1 was most significantly related to breast cancer progression with minimum heterogeneity i.e. tau2 was 0.54 as compared to 0.67 and 2.66 for Ets-2 and ETV-4, respectively.

Ets-1 has been found closely related to angiogenic pathways where it is involved in inducing the expression of pro-angiogenic factors (Randi et al., 2009). By enhancing the angiogenic mechanism Ets-1 can contribute to invasiveness and progression of breast cancer. Several findings have shown that Ets-1 expression is related to aggressive angiogenesis and invasive phenotypes (Furlan et al., 2019; Ehrenfeld et al., 2019). Ets-1 has a significant correlation with several important molecules like uPA (Urokinase activator) (Madunić, 2018) and HER2/neu (a proto-oncogene) (Nazir et al., 2019), therefore, its over-expression can be associated with breast tumor progression. Moreover, there is also a significant correlation between Ets-1 expression and VEGF and PAI-1 (Yuan et al., 2014). Based on these observations Ets-1 can be used as a prognostic factor to determine breast cancer progression.

From this meta-analysis; we conclude that Ets family members can serve as the biomarkers for the progression of breast cancer. However further validation needs to be done. It is worth mentioning that instead of causative factors Ets proteins are more involved in the development of metastasis in breast cancer. Since Ets proteins appeared at various stages of breast cancer, therefore, their expression can predict the disease progression.

The authors are thankful to the Department of Statistics, GC University, Lahore, Pakistan for their support. No funding of any type was received for this project.

None to declare

There is supplementary material associated with this article. Access the material online at: https://dx.doi.org/10.17582/journal.pjz/20220119060146

The authors have declared no conflict of interest.