Introduction

The development of multidrug resistance (MDR) followed by the failure of chemotherapy response is a critical problem in breast cancer (BC), but the underlying molecular mechanisms are incompletely understood. Numerous studies have aimed to establish the role of ATP-dependent efflux pumps, also named ATP-binding cassette (ABC) transporters, in MDR and to link their expression to the chemotherapy response [1–5].

ABC transporters are encoded by a large family of genes, called MDR genes, of which the major ones are represented by the members of the ABCB, ABCC, and ABCG subfamilies [5]. Although a number of clinical studies have reported that a high level of tumor ABC transporters is associated with cancer progression, a clear link between expression level and chemotherapy sensitivity or disease outcome has not been identified [5]. Heterogeneity in the results is a common feature of studies evaluating associations between ABC transporter expression and response to chemotherapy, particularly neoadjuvant or preoperative therapy [3–6]; one of the main reasons for this is limitations of the assays that have been used [7].

Neoadjuvant chemotherapy is considered a suitable model for direct assessment of treatment response depending on MDR gene expression before and after therapy administration. Considering this fact, we focused on the key MDR genes (ABC transporters) expressed in breast tumors: ABCB1 (MDR1), ABCC1 (MRP1), ABCC2 (MRP2), ABCC3 (MRP3), ABCC5 (MRP5), ABCG1 (BCRP1), and ABCG2 (BCRP), together with the glutathione S-transferase pi (GSTP1) gene, are involved in cellular phase II metabolism, detoxification and protection of tumor cells from genotoxic agents including chemotherapeutic drugs [8], and MVP (LRP1) operates as a cytoplasmic and/or nuclear membrane-associated drug transporter, perhaps in conjunction with ABC transporters [9]. The aim of this study was to investigate whether alterations in the levels of tumor expression of MDR genes (before and after NAC) in BC are associated with the response to NAC. Assessments of expression were carried out by TaqMan-based quantitative reverse transcriptase PCR (qRT-PCR), known as a superior method for MDR gene expression profiling [7]. It should be noted that in most studies mentioned above, MDR gene expression was evaluated just once before starting chemotherapy; however, it seems reasonable to assume that in clinical situations, not only the initial level of expression but also the active process of MDR formation (acquired or adaptive MDR) are of great significance.

Materials and methods

Patients and tumors

Patients (n = 84) with clinical stage IIA to IIIC (T 1-4 N 0−3 M 0 ) BC in the age range 28–61 years (mean age 46.95 ± 0.73) treated in the Cancer Research Institute (Tomsk, Russia) between 2006 and 2011 were included. The procedures followed in this study were in accordance with the Helsinki Declaration (1964, amended in 1975 and 1983). This study was approved by the institutional review board, and all patients signed an informed consent for voluntary participation. The histological diagnosis of invasive cancer, including estrogen (ER), progesterone (PR), and HER2 receptors were determined. The basic clinicopathological parameters of the patients are shown in Table 1.

Table 1 The clinicopathological parameters of BC patients
Full size table

All patients were primarily treated with NAC in accordance with "Consensus Conference on Neoadjuvant Chemotherapy (NAC) in Carcinoma of the Breast, April 26–28, 2003, Philadelphia, Pennsylvania" [10]. They received two to four preoperative cycles of FAC regimen (5-fluorouracil 600 mg/m2, adriamycin 50 mg/m2, and cyclophosphamide 600 mg/m2 at intervals of 3 weeks), CAX (cyclophosphamide 100 mg/m2 intramuscular injection, adriamycin 30 mg/m2 intravenous injection, xeloda 1200 mg/m2 oral administration), or taxotere in standard doses (100 mg/m2 1-h infusion of 1 day) on the basis of body surface area, administered as a monotherapy.

Physical examination was performed before NAC and was repeated after 2 cycles of NAC and before surgery to determine clinical response. Imaging of the primary breast lesion was performed with mammography and/or ultrasonography. Clinical and imaging responses were categorized according to International Union Against Cancer criteria [11]. A complete response (CR) was defined as complete disappearance of primary tumor and lymph node metastasis. A partial response (PR) was determined as a tumor reduction >50 % and stable disease (SD) as a tumor reduction ≤50 % or a tumor size increase of <25 %. Progressive disease (PD) was described as an increase of >25 % in tumor size. In this way, patients were grouped into clinical responders (CR and PR) and non-responders (SD and PD).

Surgery (radical resection or sectoral resection or mastectomy) was performed within 1 to 2 weeks after the last administration of chemotherapy in responsive patients. After surgery, adjuvant chemotherapy or hormonal therapy was given.

Fresh BC tissues were obtained during the initial diagnostic biopsy (~10 mm3) before NAC and in the course of tumor resection after NAC (~60–70 mm3). The obtained tissue samples were stored in RNAlater solution (Ambion, USA) at −80 °C per the manufacturer’s instructions until further use. Histological diagnosis was confirmed for all samples.

RNA isolation and cDNA synthesis

Total RNA was extracted from 84 samples of pre- and 75 post-NAC tumor tissues using the RNeasy mini kit plus DNase I digestion (Qiagen, Germany); a total of 9 patients showed CR, rendering it impossible to obtain further tumor samples. Ribolock RNase inhibitor (Fermentas, Lithuania) was added to the isolated RNA.

RNA concentration and quality were measured with a NanoDrop-2000 spectrophotometer (Thermo Scientific, USA). The concentration of RNA ranged from 80 to 250 ng/μl. The optical density ratios at 260/280 and 260/230 to examine RNA quality were in the range of 1.95–2.05 and 1.90–2.31, respectively. RNA integrity was assessed by visualization of the 28S and 18S ribosomal RNA in 1.5 % agarose gels followed by 0.02 % ethidium bromide staining.

RNA was reverse transcribed into cDNA using the RevertAid kit with random hexanucleotide primers (Fermentas, Lithuania) following the manufacturer’s instructions.

Quantitative real-time PCR

The expression levels of the ABCB1, ABCC1, ABCC2, ABCC3, ABCC5, ABCG1, ABCG2, GSTP1, and MVP genes were measured by qRT-PCR based on TaqMan technology using a Rotor-Gene-6000 instrument (Corbett Research, Australia). In addition, one internal control gene—GAPDH (glyceraldehyde-3-phosphate dehydrogenase)—was used to normalize expression levels of the MDR genes.

qRT-PCR was performed in triplicate reactions in a volume of 15 μl containing 250 μM dNTPs (Sibenzyme, Russia), 300 nM forward and reverse primers, 200 nM probe, 2.5 mM MgCl2, 1× SE buffer (67 mM Tris–HCl pH 8.8 at 25 °C, 16.6 mM (NH4)2SO4, 0.01 % Tween-20), 2.5U Hot Start Taq polymerase (Sibenzyme, Russia), and 50 ng of template cDNA. Samples were heated for 10 min at 95 °C, followed by 40 cycles of amplification for 10 s at 95 °C and 20 s at 60 °C.

Primer and probe (FAM-BHQ1) sequences were designed using Vector NTI Advance 11.5, Oligo 7.5, and the NCBI Nucleotide Database (http://www.ncbi.nlm.nih.gov/nuccore) (Table 2). Primers/probes were synthesized by the DNA-Synthesis Company (Russia). PCR products were visualized by 1.5 % agarose gel electrophoresis with 0.02 % ethidium bromide.

Table 2 Sequence of the primers and probes used in the study
Full size table

The mean expression level of each target gene was calculated for tumor tissue normalized to GAPDH. The average Ct (cycle threshold) was estimated for both the gene of interest and GAPDH. Relative expression was evaluated using the Pfaffl method, and the following formula was used to determine the expression ratio between the sample and the calibrator [12]:

$$ Ratio = \frac{{(E_{target} )^{\Updelta Ct,target\;(calibrator - test)} }}{{(E_{ref} )^{\Updelta Ct,ref\;(calibrator - test)} }} $$

The relative expression level was also normalized to a calibrator consisting of a pool of normal breast tissue specimens. For this purpose, specimens of adjacent normal breast tissue from 10 BC patients (NAC free) were used as a source of normal RNA. The results were articulated as n-fold differences in ABCB1, ABCC1, ABCC2, ABCC3, ABCC5, ABCG1, ABCG2, GSTP1, and MVP gene expression relative to GAPDH and normal breast tissue.

Statistical analysis

Comparisons between qualitative attributes (e.g., ER+ and ER− or N0 and N1−3) were made using logistic regression, and the difference in expression level of each gene between the pre- and post-NAC condition was determined using the Wilcoxon signed-rank test. Relationships between changes in MDR gene expression in pre- and post-NAC samples, the clinicopathological characteristics of the patients, and different clinical responses to FAC, CAX, and taxotere chemotherapy were evaluated with the chi-square test. Statistical analyses were performed using STATISTICA 8.0 software (StatSoft, Tulsa, OK, USA). Differences were significant if the p value was <0.05.

Results

According to clinical criteria, after NAC, 35/60 (58.3 %) of patients treated with FAC or CAX had objective tumor responses, whereas response failure was detected in 25/60 (41.7 %) patients. Patients who had been treated with taxotere showed tumor responses in 13/24 (54.2 %) cases versus 11/24 (45.8 %) with no response. No significant association was found between tumor response and any clinicopathological parameter (Table 3).

Table 3 Clinicopathological parameters and NAC response
Full size table

To determine whether MDR gene expression correlated with patient and tumor characteristics, we screened pre-NAC samples of breast tumor from 84 patients using qRT-PCR (table in Online Resource 1). With one notable exception, no significant differences in expression level were observed according to patient age, menopausal status, tumor size, histological type (ductal or lobular), histological form (unicentric or multicentric), ER, PR, or nodal status. However, ABCB1 expression appeared to be higher in T3−4 patients versus T1−2 patients (p = 0.037), while ABCC5 and MVP levels depended on lymph node involvement (p = 0.025) and histological type (p = 0.029), respectively. The mean levels of expression of the studied genes before NAC did not differ between patients with responding or resistant tumors as shown in the table (Online Resource 1). In addition, no significant differences in average expression levels were observed between pre- and post-NAC samples when the total group was analyzed, including both responders and non-responders (data not shown).

We studied the expression of genes in paired pre- and post-operative breast tumor specimens (n = 75) and compared changes in their expression in response to NAC. Expression data were available for 75/84 tumors because post-operative tumor tissue from patients who had complete responses (9 patients) was not available to be screened. When alterations in gene expression between paired pre- and post-NAC specimens were calculated, they revealed cases showing either increases or decreases in expression level as well as some without expression. We did not observe any association between clinicopathological parameters and numbers of patients showing increases or decreases in post-NAC expression of MDR genes. As an exception, we noted that expression of the ABCB1 and MVP genes was frequently increased in unicentric versus multicentric patients (p = 0.046 and p = 0.038, respectively), while alterations in ABCC3 activity depended on ER status (p = 0.015; table in Online Resource 2).

However, there was a significant correlation between tumor response to NAC and number of cases displaying expression changes in MDR genes after NAC. In the group of patients with different chemotherapy regimens, reductions in expression of MDR genes in post-NAC tumor samples compared with pre-NAC specimens were found in the majority of responders (63–90 % cases). On the contrary, gene upregulation was detected in 58–94 % of cases showing no objective response to chemotherapy (10−13  < p < 0.02; table in Online Resource 2). The mean values for gene expression in pre- and post-NAC tumor samples are presented for responders and non-responders in Fig. 1. The data illustrated here confirmed once again the results of table (Online Resource 2) and showed that expression level of the most studied genes, such as ABCB1, ABCC1, ABCC2, ABCG1, ABCG2, MVP, and GSTP1, was lower after NAC compared with before chemotherapy in responsive cases, whereas patients with resistance to chemotherapy displayed increases in the activities of the majority of the above-mentioned genes in post-chemotherapy tumor tissue in comparison with pre-NAC samples (10−6 < p < 0.04). Interestingly, the expression of ABCG1 gene, detected as decreased in cases with response to NAC, did not significantly change between before and after chemotherapy in non-responsive subjects (p = 0.20; Fig. 1).

Fig. 1
figure 1

The expression level of MDR genes in breast tumors collected before and after neoadjuvant chemotherapy (NAC). The expression level, as mean and standard error (M ± SE), was given for patients with partial response (a, n = 38) and stabilization/progression of disease (b, n = 27) after NAC. Statistical analysis: p value—Wilcoxon signed ranks test. Significant differences were shown for ABCB1 (p = 5.0 × 10−6), ABCC1 (0.0349), ABCC2 (8.0 × 10−6), ABCG1 (7.8 × 10−5), ABCG2 (4.9 × 10−5), GSTP1 (0.0018), and MVP (0.0003) genes in the group with partial response and for ABCB1 (2.0 × 10−6), ABCC1 (0.0179), ABCC2 (0.0007), ABCG2 (5.0 × 10−5), GSTP1 (0.0179), and MVP (0.0376) genes in the group with stabilization/progression of disease

Full size image

Alterations in gene expression following NAC were also assessed separately in the FAC-CAX and taxotere-treated patient groups. In the case of FAC-CAX regimens, comparison of pre- and post-NAC specimens showed that a high percentage of cases (from 67 to 93 %) with decreased expression of ABCB1, ABCC1, ABCC2, ABCC5, ABCG1, ABCG2, MVP, and GSTP1 genes was in the group of responders, while non-responders (55–96 % of cases) displayed high expression of these genes after chemotherapy. In both responsive and resistant group, the high significance of differences was between cases with decrease and increase in expression of only ABCB1, ABCC1, ABCC2, and ABCG2 genes (10−9 <p < 10−5; Fig. 2). In taxotere, 85–92 % of responsive cases showed decrease in ABCB1, ABCC2, ABCG1, ABCG2, and MVP levels in post-NAC tumor samples relative to pre-chemotherapy specimens. Interestingly, 67–91 % of cases with response failure to taxotere had increased expression of the same genes (10−5 < p < 0.03; Fig. 2). In both the FAC-CAX- and taxotere-treated groups, the significant differences were shown between patients, which had alterations only in expression of ABCB1, ABCC2, and ABCG2 genes during NAC (Fig. 2). Moreover, using logistic regression, we demonstrated strong associations between alterations in expression of these three genes and NAC response (OR = 78.9 [95 % CI: 14.87–508.27]; p < 10−14).

Fig. 2
figure 2

The percentage of cases with alteration in MDR gene expression depending on different neoadjuvant chemotherapy (NAC) regimen and response. Patients treated in cyclophosphamide–adriamycin–xeloda (CAX) and 5-fluorouracil–adriamycin–cyclophosphamide (FAC) regimens are presented in Fig. 2a, patients with taxotere—in Fig. 2b. PR designates partial response, SD stable disease, PD progressive disease. n1 means the number of patients with decrease in gene expression or without expression after neoadjuvant chemotherapy, n2—the number of patients with increase in gene expression after neoadjuvant chemotherapy. Statistical analysis: p value—Chi-squared (df = 1) test. Significant differences were shown for ABCB1 (p = 2.2 × 10−9), ABCC1 (6.9 × 10−5), ABCC2 (4.4 × 10−6), ABCC5 (0.0096), ABCG1 (0.0003), ABCG2 (1.2 × 10−7), GSTP1 (0.0003), and MVP (0.0013) genes in the group with CAX-FAC and for ABCB1 (8.9 × 10−5), ABCC2 (0.0006), ABCG1 (0.0131), ABCG2 (0.0031), and MVP (0.0260) genes in the group with taxotere

Full size image

Discussion

Neoadjuvant chemotherapy (NAC) is an integral part of a multimodality approach in the management of locally advanced BC [15], expanding surgical options, improving cosmetic results, and allowing oncologists to assess tumor response to therapy. Theoretically, it may also provide early control of micrometastatic disease [16, 17]. NAC is highly effective, with a clinical response rate ranging from 50 to 90 %, although with a much lower pathological complete response (pCR) rate, ranging from 2 to 27 % [18, 19]. The ability to achieve pCR is considered a key marker for adjuvant chemotherapy response and disease-free and overall survival [15, 16, 20].

It is clear that there are different molecular genetic markers associated with the presence or absence of pCR after pre-operative chemotherapy, which help tumors escape the toxicity induced by an ineffective chemotherapeutic regimen in non-responsive patients, and could assist planning an alternative course of therapy. Currently, clinical tests for predicting cancer response are not available, and individual markers have shown little predictive value [21, 22]; however, the latest studies show great promise in developing tests to predict chemotherapy response [23–25].

It is well established that many breast tumors that initially respond to chemotherapy subsequently develop resistance to a broad range of drugs. In most cases, drug resistance is the outcome of a variety of cellular and pharmacological processes, of which the most significant is the activity of ABC transporters ejecting cytostatic agents from tumor cells against the concentration gradient with ATP energy consumption [5]. In the present study, we have focused on the key MDR genes that encode ABC transporters: ABCB1 (MDR1), ABCC1 (MRP1), ABCC2 (MRP2), ABCC3 (MRP3), ABCC5 (MRP5), ABCG1 (BCRP1), and ABCG2 (BCRP), the increased expression of which may provoke resistance to neoadjuvant chemotherapy in BC. In addition, taking into account the major role of GSTP1 in metabolite detoxification [8] and MVP (LRP1), which acts as a cytoplasmic and/or nuclear membrane-associated drug transporter, perhaps in conjunction with ABC transporters [9], these genes have also been included in our study.

The role played by MDR genes in clinical treatment is a subject of debate. A number of studies have attempted to assess the relationship between MDR gene expression and type of chemotherapy response as well as disease outcome in BC patients; however, there are many contradictions in the data [3–6]. In the present study, we did not demonstrate a link between the pre-NAC levels of MDR gene expression with any clinicopathological parameter of BC except some casual observations (table in Online Resource 1). Our observations, which are in accordance with previous data [26–29], most likely explain the weak association of clinicopathological parameters with NAC response in BC.

In most studies, MDR gene expression was evaluated once before starting chemotherapy; however, it is known that chemotherapy may modify gene expression, and it seems reasonable to assume that not only the initial level of expression but also the active process of MDR formation are of great significance. Based on the above reasoning, we analyzed whether chemotherapy induced changes in MDR gene expression correlate with immediate clinical response to NAC. The expression of the MDR genes listed above was evaluated in paired breast tumor samples collected before and after NAC. In our study, MDR gene activity was increased in non-responsive patients and decreased in responders after NAC. These data are in agreement with results of Linn et al. [30] reporting changes in expression of ABCC1, ABCB1, and MVP genes in opposing directions in NAC, and observations of Chevillard and colleagues [31] demonstrating an association between up- and downregulation of ABCB1 gene expression after chemotherapy with resistance and sensitivity to NAC, respectively. A more recent study [32] also showed that BC patients in whom expression of the ABCB1 gene was induced during the process of NAC did not respond to chemotherapy, whereas cases without gene upregulation displayed high rates of successful treatment.

In accordance with the above-mentioned studies and other studies [33–36], the changing expression of MDR genes during NAC is a well-known phenomenon; however, there is heterogeneity in establishing a clear link between a unidirectional change in MDR gene activity (increase/decrease) with chemotherapy efficiency. In particular, Singh and coauthors [35] demonstrated increase in ABCB1 gene expression after NAC in both responsive and non-responsive BC cases. Faneyte et al. [33, 34] also failed to show significant differences in post-NAC expression levels of ABCB1 and ABCC1-3 in responders and non-responders with BC.

The alterations of MDR gene activity during NAC is a well-known phenomenon related to regulation of their expression and/or change in the total number of tumor cells expressing them, as has been shown by immunohistochemistry more than once [2]. Genetic elements and processes such as DNA methylation, histone deacetylation, transacting proteins (transcription factors), DNA sequence variants, and microRNAs (miRNAs) could be involved in the modulation of MDR gene expression at either the transcriptional or translational levels [37–39].

A reduction in MDR gene activity during NAC could be the result of expression inhibition and/or a lower percentage of tumor cells expressing the genes owing to cell death during chemotherapy [2, 34]. Our data indicate that along with the influence of reduction of number of tumor cells, a decrease in MDR gene expression is possible caused by strong repression of transcription during chemotherapy. In particular, the study of level of MDR genes in biopsy samples, taken from four patients before NAC and after 7 days from start of CAX chemotherapy, showed a twofold decrease in expression in 3/4 (75.0 %) patients, although therapy response was still non-significant. Further, all these patients displayed partial regression of tumor (own unpublished data). The well-known fact in the regulation of MDR gene transcription is strong binding promoter of the ABCB1 gene, one of the most important drug transporters, by the wild-type p53 protein and further repression of transcription [40, 41]. It is not excluded that p53, high level of which is induced by chemotherapeutic agents due to DNA damage, can downregulate ABCB1 transcription. However, mutation analysis is needed to clarify p53 status, because it is known that mutant p53 upregulates ABCB1 expression [42]. In addition, because surgery was performed one to 2 weeks after the end of NAC, MDR gene expression in post-NAC tumor samples could be restored to normal by natural processes.

An increase in MDR gene activity after NAC is evidently related with expression induction and/or selection of resistant clones with high expression of drug transporters and further increase in their number. So, there are many data demonstrating that chemotherapy agents decrease DNA methylation and inhibit activity of histone deacetylases [43–45] that can make MDR gene promoters accessible and competent for subsequent transcriptional activation. Different mechanisms of induction of DNA hypomethylation by chemotherapeutic drugs have been described in the recent review [45]. It is also known that deregulation of Raf/MEK/ERK, MAPK, and JNK pathways, which could be caused by chemotherapy, induces the expression of drug transporters [46, 47]. Again, it should not be forgotten that mutant p53 induces MDR gene expression, as it was already mentioned above.

It must be taken into account that in many cases, the expression of MDR genes is inversely correlated with the levels of several miRNAs. In turn, miRNA activity is greatly modulated by chemotherapeutic drugs most likely via DNA damage, DNA demethylation, and histone deacetylase inhibition, which were found to be related to extensive and rapid alteration of microRNA levels [48, 49].

Therefore, in clinical practice, as opposed to in vitro, there is a chemotherapy-linked process of MDR formation controlled both by the organism and by the tumor factors. The study of MDR development appears to be a key factor for understanding mechanisms of chemoresistance. Therefore, to predict chemotherapy efficiency, it is necessary to understand why, with the same treatment regimen, some patients display a decrease in MDR gene expression and a good response, while others display gene upregulation and resistance to therapy. The answer to the question most likely lies in individual features of the patients and the tumor. Aside from the chemotherapy-related factors modifying MDR gene expression, gene polymorphism is one of the major keys to understanding the development of MDR gene expression and drug resistance in individual patients, a point that was partially demonstrated in our previous papers [50, 51]. It should be noted that intratumoral morphological heterogeneity in BC or the presence of five different morphological types of infiltrating components (morphological structures), mirroring the architectural arrangements of tumor cells [52], results in diverse “portrait” of MDR gene expression [53] and different response to NAC [54] within one tumor.

Taken together, the data in this study suggest that changes in the expression vector of ABCB1 (MDR1), ABCC1 (MRP1), ABCC2 (MRP2), ABCC3 (MRP3), ABCC5 (MRP5), ABCG1 (BCRP1), and ABCG2 (BCRP) genes during the chemotherapy process or the development of adaptive MDR, but not the mRNA levels of these genes per se, are associated with NAC efficiency. Increases in MDR gene expression after NAC resulted in poor responses, whereas decreases were related to high chemotherapy efficiency. Once again, we would like to note that future studies should focus on the molecular basis of how the expression of these transporters is regulated in normal breast cells and in their malignant counterparts, as previously suggested by Prof. M. Tien Kuo. These studies may lead to novel strategies of controlling MDR through gene regulation [4].