Association between ABCB1, ABCG2 carrier protein and COX-2 enzyme gene polymorphisms and breast cancer risk in a Turkish population.

AIM: Breast cancer is the most common cancer and the second leading cause of cancer-related deaths among women. Several genetic and environmental factors are known to be involved in breast cancer pathogenesis, but the exact etiology of this disease is complicated and not completely understood. We aimed to investigate whether the gene polymorphisms of ABCB1 and ABCG2 carrier proteins and COX-2 enzyme affect breast cancer risk.
METHOD: ABCG2 C421A (rs2231142), ABCB1 C3435T (rs1045642), COX-2 T8473C (rs5275) and COX-2 G306C (rs5277) were genotyped 104 breast cancer patients and 90 healthy controls using a real-time PCR for breast cancer susceptibility.
RESULTS: Patients carrying ABCG2 C421A, the CC genotype, had a higher risk of disease compared with patients carrying any A allele (OR = 3.06; 95% CI = 1.49-6.25, p = 0.0019). The other variants showed no association with breast cancer (p > 0.05). Comparing the pathological parameters with the variants, only, the frequency of C allele of ABCB1 C3435T was significantly lower in the estrogen receptor-α (ERα) (OR = 2.25; 95% CI: 0.75-6.76; p = 0.041) and progesterone receptor (PgR) (OR = 3.67; 95% CI: 1.34-10.03; p = 0.008) positive breast cancer patients.
CONCLUSION: ABCB1 C3435T and ABCG2 C421A might represent a potential risk factor for breast cancer for Turkish women.

Introduction

Breast cancer is the most frequent cancer, 30% of all new cancer diagnoses, in women, and is responsible for roughly half a million total deaths each year worldwide (Siegel et al., 2019). According to latest cancer report of Turkey Ministry of Health (2017), the rate of breast cancer is 24.9% in adult Turkish women. Some risk factors influence of developing breast cancer are menstrual history, reproductive factors, hormone use, genetics, family history, diet and exercise (Torre et al, 2017). The loss or inhibition of various ATP-binding cassette (ABC) transporters has been observed to influence tumor cell phenotypes closely associated with malignant potential, including proliferation, differentiation, migration and invasion; these observations have been made across multiple cancer types (Fletcher et al., 2016).

Multi Drug Resistance (P-glycoprotein, P-gp, ABCB1, MDR1) and Breast Cancer Resistance (BCRP, ABCG2, MXR, ABCP) ABC transporter proteins limit the intracellular concentration of the substrates via energy-dependent (active) pumping out of the cell. ABCB1 and ABCG2 protect the body against endogenous and exogenous xenobiotics with their important roles in intestinal absorption and secretion, hepato- and urinary elimination, and barrier through the placenta, testis and brain (DeGorter et al., 2012, Klaassen and Aleksunes, 2010, Robey et al., 2009). The single nucleotide polymorphism (SNP) of ABCB1, C3435T (rs1045642), occurs in exon 26, and the T allele appears to be associated with markedly lower P-gp expression compared with the C allele (Hoffmeyer et al., 2000). The SNP has been shown to be correlated with the development of various type of cancer such as colorectal (He et al., 2013), acute lymphoblastic leukemia (Yaya et al., 2014), glioma (Miller et al., 2005) and renal epithelial tumors (Haenisch et al., 2007). ABCB1 C3435T might reduce protection for cells and potentially contribute to the development of breast cancer (George et al., 2009, Wang et al., 2013). However, the results have been contradictory (Wang et al., 2012). The ABCG2 C421A (rs2231142) in exon 5 is one of the most important genetic variations and results in lower expression levels in the cellular membrane compared with the wild-type protein (Hira and Terada, 2018). BRCP is also expressed from the apical membrane of alveolar epithelial cells in breast tissue at during pregnancy and lactation and plays a role in the expulsion of accumulated toxins and carcinogens to a woman’s milk (DeGorter et al., 2012, Klaassen and Aleksunes, 2010, van Herwaarden and Schinkel, 2006). To date, studies have investigated the association between ABCG2 gene polymorphisms and susceptibility to carcinoma such as non-papillary renal cell carcinoma (Korenaga et al., 2005), B cell lymphoma (Campa et al., 2012) and prostate cancer (Hahn et al., 2006). However, the association between ABCG2 gene polymorphisms and breast carcinoma risk has been evaluated in only a few studies (Wu et al., 2015, Ghafouri et al., 2016, Li et al., 2017).

Prostaglandins play a role in carcinogenesis via the suppression of immune responses, and the inhibition of apoptosis, angiogenesis, tumor cell invasion and metastasis pathways (Brasky et al., 2011, Lala et al., 2018). Prostaglandin-endoperoxide synthase 2 (COX-2) is an inducible enzyme that plays a major role in the inflammatory response by converting arachidonic acid to prostaglandins. Overexpression of COX-2 has been found in a variety of cancers; thyroid (Ucan et al., 2017); colorectal (Eberhart et al., 1994), gastric (Ristimäki et al., 1997) and breast (Liu and Rose, 1996). In recent studies, COX-2 T8473C (rs5275), G899C (rs20417) and G306C (rs5277) have been shown to cause an increase in the level of COX-2 expression (Abraham et al., 2009, Brasky et al., 2011, Yu et al., 2010, Li et al., 2009). The variants have also been investigated for their role in contributing to breast cancer risk (Li et al., 2015). However, the results have been inconclusive.

Overexpression of COX-2 can result the over-production of prostaglandins, which are substrates for P-gp and BCRP. The dysfunction or reduced function of P-gp and BCRP proteins can cause carcinogenesis via xenobiotics and the accumulation of inflammatory agents in cells (Andersen et al., 2015). Knowledge of ethnic and individual genetic differences is very important for understanding personal reactions in the case of exposure to xenobiotics/drugs (Ishikawa et al., 2012, DeGorter et al., 2012). We accordingly investigated whether the single nucleotide polymorphisms (SNPs) of ABCB1 and ABCG2 carrier proteins and COX-2 enzyme affect breast cancer risk since these genetic differences have not been clarified in Turkish population. We believe that the preliminary study could enrich the scarce literature about the polymorphisms in breast cancer susceptibility.

Materials and methods

Subjects

We evaluated the influence of ABCG2, ABCB1 and COX-2 gene polymorphisms on susceptibility to breast cancer in 104 Turkish female patients and 90 ethnic- and age-matched healthy controls between 2012 and 2015. These 104 patients had a mean age of 52 ± 12 years were operated upon at the Acibadem Maslak Hospital Breast Health Centre (Istanbul, Turkey) or admitted for follow-up after breast cancer surgery. Healthy control volunteers with a mean age of 49 ± 14 years who never had any type of cancer were in-patients with various diagnoses (e.g., eye diseases, pulmonary diseases, cardiovascular diseases and neurological disorders) at the Hospital of Istanbul University (Istanbul, Turkey). All participants provided informed consent, and the study was approved by the ethics committees of Istanbul and Acibadem Universities (2011/87-555; 2012/291). Demographic and anthropometric factors were assessed using a short questionnaire. The pathological types of the patients were categorized as invasive ductal carcinoma (IDC), invasive lobular carcinoma (ILC) or ductal carcinoma in situ (DCIS). We also evaluated the association between patient genotypes and the status of ERα, PgR and HER2.

Genotyping

Genomic DNA was extracted from whole blood using standard phenol chloroform extraction protocol and further purification was done by using High Pure PCR Product Purification Kit (Roche, Mannheim, Germany). SNP analysis was performed using a LightCycler FastStart DNA Master HybProbe (Roche, Mannheim, Germany) and custom-designed LightSNiP assay probes (Roche, Mannheim, Germany) according to the manufacturer’s instructions. ABCG2 C421A, ABCB1 C3435T, COX-2 T8473C and COX-2 G306C were genotyped using a Roche Light Cycler 480 (Roche, Mannheim, Germany) real-time PCR platform and melting curve analyses were performed by the carousel-based system PCR program. In a final volume of 20 mL reaction mix per sample, the following mixtures were added: 1X FastStart DNA Master Mix, 2 mM MgCl2, 0.2 mM LightSNP HybProbe, appropriate amount of PCR grade water and 500 ng DNA sample. The plates were sealed and centrifuged at 3000 rpm for a minute. Details of custom-designed LightSNiP assay probes were summarized in Table 1 and carousel-based system PCR program setup was given in Table 2.

Table 1

Reference sequences of custom-designed LightSNiP assay probes.

LightSNiP	Reference Sequence	Melting temperature
ABCG2 C421A (rs2231142)	GCACTCTgACggTgAgAgAAAACTTA [A/C] AgTTCTCAgCAgCTCTTCggCTTgC	57.74 °C for allele [C]61.92 °C for allele [A]
ABCB1 C3435T (rs1045642)	AgCCgggTggTgTCACAggAAgAgAT [C/T] gTgAgggCAgCAAAggAggCCAACA	55.76 °C for allele [C]63.05 °C for allele [T]
COX-2 G306C (rs5277)	TTCgAAATgCAATTATgAgTTATgT [C/G] TTgACATgTAAgTACAAgTgTCTTT	53.64 °C for allele [G]62.30 °C for allele [C]
COX-2 T8473C (rs5275)	TTTgAAATTTTAAAgTACTTTTggT [C/T] ATTTTTCTgTCATCAAACAAAAACA	52.94 °C for allele [T]61.12 °C for allele [C]

rs: reference SNP number; alleles in the square brackets indicates the polymorphisms.

Table 2

Carousel-based system PCR program setup.

Program Name	Cycles	Analysis Mode	Target (°C)	Acquisition Mode	Hold (sec)
Pre-Incubation	1	None	95	None	600
Amplification	45	Quantification	95	None	10
			60	Single	10
			72	None	15
Melting Curve	1	Melting Curve	95	None	30
			40	None	120
			75	Continuous	–
Cooling	1	None	40	None	30

Genotyping was performed by scientists blinded to the patients’ case control status. A 10% random sample was genotyped twice for quality assurance. Also, to confirm the genotyping results of the variants, the selected PCR amplified DNA samples (n = 2, for each genotype in the cases and controls) were examined with DNA sequencing. The results were 100% concordant.

Statistical analysis

The sample size was calculated by an online sample size estimator (http://osse.bii.a-star.edu.sg). Hardy-Weinberg equilibrium (HWE) analysis was performed using the Chi-square (χ2) test. For the analysis of genotype frequencies, the wild-type category (chosen either as the most common wild-type frequency or arbitrarily if the two alleles exhibited similar frequencies) was used as the reference group. Data comparisons were done by using Fisher’s exact test. To evaluate the association between the genotype frequencies and breast cancer, odds ratios (ORs) and 95% confidence intervals (CIs) were estimated. All of the statistical analyses were performed using Statistical Package for Social Sciences (SPSS) software (Version 17; SPSS Inc, Chicago, IL, USA). A two-sided p value less than 0.05 was considered to be statistically significant.

Results and discussion

Although the correlation between the SNPs of the ABCG2, ABCB1 and COX-2 genes with breast cancer risk has been reported in some studies, no meaningful relationship has been demonstrated thus far. The substrate specificities of ABCG2 and ABCB1 are quite similar, and ABCG2 and ABCB1 are involved in the transport of COX-2 mediated inflammatory agents (Klaassen and Aleksunes, 2010, Yu et al., 2010). Therefore, we evaluated the association between functional and common variants (ABCG2 C421A, ABCB1 C3435T, COX-2 T8473C and COX-2 G306C), and susceptibility to breast cancer in a cohort of Turkish women.

Firstly, we determined that no significant differences in age (52.4 ± 12.5 vs. 49.4 ± 14.2 years) or BMI (27.9 ± 5.3 vs. 24.5 ± 4.5 kg/cm2) between the breast cancer and control groups, respectively, suggesting that the matching based on these two variables was adequate. Secondly, the genotype distributions did not significantly deviate from the HWE in either the case or control groups for any of the examined SNPs (p > 0.05). After determining a lack of bias among the study populations, the differences between cases and controls in the distributions of ABCG2 C421A, ABCB1 C3435T, COX-2 T8473C and COX-2 G306C genotypes were analyzed (Table 3).

Table 3

Genotype distributions and features of the studied SNPs.

				Frequencies
SNPs	Amino acid change	Variant allele	Genotypes	Cases (n = 104, %)	Controls (n = 90, %)	OR (95% CI)	p value
ABCG2 C421A (rs2231142)			CC	90 (86.5)	61 (67.7)	CC vs. any A3.06 (1.49–6.25)	0.002*
	Q141K	C	CA	14 (13.5)	25 (27.7)
			AA	0 (0.0)	4 (4.4)
MAF				0.072	0.183
ABCB1 C3435T (rs1045642)			CC	25 (24.2)	16 (18.1)	CC vs. any T1.24 (0.61–2.53)	0.361
	I1145I	C	CT	37 (35.9)	40 (45.4)
			TT	41 (39.8)	32 (36.3)
MAF				0.422	0.418
COX-2 G306C (rs5277)			GG	46 (44.6)	39 (43.3)	GG vs. any C1.06 (0.59–1.86)	0.853
	V102V	G	,GC	47 (45.6)	36 (40.0)
			CC	10 (9.7)	15 (16.7)
MAF				0.325	0.366
COX-2 T8473C (rs5275)			TT	72 (69.2)	58 (64.4)	TT vs. any C1.24 (0.68–2.26)	0.479
	–	T	TC	28 (26.9)	28 (31.1)
			CC	4 (3.8)	4 (4.4)
MAF				0.173	0.200

SNPs, single nucleotide polymorphisms; rs, reference SNP number; MAF, minor allele frequency; OR, odds ratio; 95% CI, 95% confidence intervals. *p < 0.05 indicates statistical significance.

The ABCG2 C421A variant has been shown to be associated with a reduction of BCRP protein expression, and therefore decreased carrying capacity (Hira and Terada, 2018). The protein expression in subjects with the C421A mutant variant is reduced compared with that of patients carrying the wild-type variant (Mizuarai et al., 2004). While the incidence of the ABCG2 C421A polymorphism is 30% in Far Easterners, it has been reported to be approximately 10% and 13% in Caucasians and Middle Easterners, respectively (Kim et al., 2010). Ghafouri et al. (2016) found that the most frequent genotype in patient groups was the AA genotype; its frequency was significantly different from that of the control subjects (p = 0.04). In the present study, CC genotype was the most frequent genotype in both our case and control groups, unlike to Kurdish populations in Sanadaj-Iran in comparison with Ghafouri et al., 2016. Wu et al. (2015) investigated the correlation between the ABCG2 C421A polymorphism and breast cancer susceptibility in 1169 patients with breast cancer and 1244 healthy controls. The authors showed that the ABCG2 C421A AA genotype was significantly associated with an increased risk for developing breast carcinoma (p = 0.033). According to our results, ABCG2 C421A was significantly associated with an increased risk of breast cancer (p = 0.0019). However, the patients carrying the CC genotype interestingly had a higher risk of disease compared with the patients carrying any A allele (OR = 3.06; 95% CI = 1.49–6.25) (Table 3). We indicate the results should be confirmed with the larger group because the frequency of the patients carrying AA genotype was ≤ 4.4%. The other variants, CA and AA genotypes, might be not associated with breast cancer (p ≥ 0.632) (Table 3).

Many epidemiological studies have examined the relationship between ABCB1 polymorphisms and breast cancer. Wang et al. (2013), in a meta-analysis, evaluated 10 case-control studies that encompassed 5282 breast cancer patients and 7703 healthy controls. It was suggested that ABCB1 C3435T polymorphism might contribute to individual susceptibility to breast cancer. The recent studies have demonstrated a significant difference in the distribution of C and T alleles between breast cancer patients and healthy controls (Abuhaliema et al., 2016, Gutierrez-Rubio et al., 2015, Macías-Gómez et al., 2014). However, some studies failed to find a significant relationship (Li et al., 2017, Ghafouri et al., 2016). Turgut et al. (2007) assessed the relationship between breast cancer risk and the ABCB1 C3435T variant in 50 healthy volunteers and 57 breast cancer patients in Turkish population. They found the polymorphism was associated with a 1.5-fold increase in breast cancer risk. However, we found no relationship between the ABCB1 C3435T variant and breast cancer development (OR = 1.242, 95% CI: 0.61–2.53; p = 0.36) (Table 3).

COX-2 T8473C is a common polymorphism associated with several cancers in different ethnic populations (Jiang et al., 2014). Thus far, more than 20 COX-2 genetic variations have been investigated for the risk of breast cancer (Dai et al., 2014). However, Li et al., 2015 have been assessed the results as conflicting. In a recent meta-analysis, Jiang et al. (2014) found that the COX-2 T8473C gene polymorphism might not be a risk factor for breast cancer among Caucasians. Three studies focused on COX-2 G306C and the risk of breast cancer, however, failed to find any significant relationship between polymorphism and breast cancer risk (Dossus et al., 2010, Abraham et al., 2009). Similarly, we did not find any significant differences in COX-2 gene variants between breast cancer patients and the control group (Table 3).

As it is well known, ERα plays an important role in the progression of breast cancer. Patients with high ERα or PgR expression have a better prognosis. It has been shown that E2-mediated downregulation of P-gp expression occurs in ERα-positive human breast cancer cells and possibly also in normal breast tissues (Hua et al., 2018). In addition, estrogenic activity has been shown to downregulate P-gp expression via post-transcriptional processes in ERα-positive cell lines, leading to increased cellular uptake of P-gp substrates (Mutoh et al., 2006).

Some studies have been reported that no statistically significant association between ABCB1 C3435T and ABCG2 C421A polymorphisms and tissue expression of ERα, PgR, HER2/neu, and Ki67 (Ghafouri et al., 2016, Li et al., 2017). Comparing the pathological parameters with the variants, we found that the ABCB1 C3435T C allele was less frequent in breast cancer patients with ERα-positive (OR = 2.25; 95% CI: 0.75–6.76; p = 0.041) and PgR (OR = 3.67; 95% CI: 1.34–10.03; p = 0.008) status. No significant differences in HER2 status, triple negative status (none of ERα-, PgR and HER2 expressed) or pathological type were found with the variants (Table 4).

Table 4

Relations between polymorphisms and clinicopathological characteristics.

Variables	No. of patients (%)	ABCG2 C421A (rs2231142)			p value	ABCB1 C3435T (rs1045642)			p value	COX-2 G306C (rs5277)			p value	COX-2 T8473C (rs5275)			p value
		CC	CA	AA		CC	CT	TT		GG	GC	CC		TT	TC	CC
Pathological type					0.657				0.975				0.056				0.080
IDC	61 (58.7)	53	8	0		16	22	23		45	14	2		25	32	4
ILC	8 (7.7)	7	1	0		2	3	3		7	1	0		3	2	3
DCIS	8 (7.7)	6	2	0		1	3	3		6	0	2		2	5	1
unknown	27 (25.9)
ERα status					0.367				0.041*				0.062				0.896
positive	57 (54.8)	49	8	0		12	26	19		40	3	14		23	29	5
negative	17 (16.3)	16	1	0		6	2	9		15	2	0		6	9	2
unknown	30 (28.9)
PgR status					0.904				0.008*				0.449				0.896
positive	42 (40.4)	42	6	0		11	24	13		35	9	3		29	23	4
negative	31 (29.8)	23	3	0		7	4	15		19	5	0		9	15	2
unknown	31 (29.8)
HER2 status					0.108				0.878				0.927				0.399
positive	16 (15.4)	12	4	0		4	5	7		12	3	1		4	10	2
negative	56 (53.8)	51	5	0		14	21	21		40	13	3		24	25	7
unknown	32 (30.8)
Triple negative					0.987				0.317				0.596				0.675
yes	8 (7.7)	7	1	0		3	1	4		7	1	0		2	5	1
no	63 (60.5)	55	8	0		15	25	23		45	15	3		26	31	6
unknown	33 (31.7)

IDC, Invasive ductal carcinoma; ILC, Invasive Lobular Carcinoma; DCIS, Ductal carcinoma in situ; ERα, estrogen receptor-α; PR, progesterone receptor; HER2, human epidermal growth factor receptor 2. *p < 0.05 indicates statistical significance.

Based on the results reported by Ghafouri et al. (2016), the A allele of the ABCG2 C421A variant was associated with IDC (p < 0.05). Wu et al. (2015) found that an association between this polymorphism and ER and PgR status in breast cancer patients. However, we not find any relationship between the clinicopathological status of breast cancer patients and the ABCG2 C421A polymorphism (Table 4). Similarly, Li et al. (2017) did not find a significant difference in the distribution of the ABCG2 C421A variant between breast cancer patients and healthy controls; there was also no difference based on the clinical status of the patients (p > 0.05).

Like other studies (Leo et al., 2006, Ranger et al., 2004), we did not find a significant correlation between COX-2 SNPs and the clinical status of breast cancer patients (Table 4).

Conclusion

We observed that ABCG2 C421A was significantly associated with an increased risk of breast cancer (p = 0.0019). No studies have demonstrated the prevalence of the ABCG2 C421A polymorphism in Turkish society and its relationship to breast cancer; our results are the first data. ABCB1 C3435T did not exhibit any relationship with breast cancer development. However, we found significant differences between ABCB1 C3435T variant and ERα and PgR status with breast cancer in our population of Turkish women. ABCB1 C3435T might be associated with a potential risk for breast cancer in Turkish women. These data might be useful for identifying individuals at risk of developing breast cancer. However, our results were obtained with a limited sample size; we were accordingly only able to draw preliminary conclusions at this time. Future studies based on larger, stratified case-control populations are still necessary to clarify the different effects of the ABCB1, ABCG2 and COX-2 polymorphisms on cancer risk. Larger sample sizes and functional assays will be required to confirm our findings.