The Relationship of ABCB1/MDR1 and CYP1A1 Variants with the Risk of Disease Development and Shortening of Overall Survival in Patients with Multiple Myeloma

(1) Background: The aim of our study was to analyze the possible relationship of ABCB1 and CYP1A1 gene variants with susceptibility and outcome of multiple myeloma (MM); (2) Methods: Genomic DNA samples from 110 newly-diagnosed MM patients and 100 healthy blood donors were analyzed by methods-PCR-RFLP (for ABCB1 3435C > T, CYP1A1 6235T > C—m1), automated DNA sequencing (for ABCB1 1236C > T, 2677G > T/A) and allele-specific PCR (for CYP1A1 4889A > G—m2); (3) Results: The genotypic frequencies of CYP1A1 4889A > G variant were not in Hardy-Weinberg equilibrium for MM patients. The presence of m1 and m2 CYP1A1 alleles decreased the risk of MM—OR = 0.49 (p = 0.011) and OR = 0.27 (p = 0.0003), respectively. In turn, TT genotype (ABCB1 2677G > T/A) increased the risk of this disease (p = 0.007). In the multivariate Cox analysis CT + TT genotypes (ABCB1 3435C > T) were associated with decreased risk of death (HR = 0.29, p = 0.04). In log-rank test in patients with CT genotype (ABCB1 3435C > T) was observed association of overall survival with the type of treatment; (4) Conclusions: Our findings suggest that T-alleles of ABCB1 2677G > T/A and m1/m2 alleles of CYP1A1 affected the susceptibility of MM. Moreover, T-allele of ABCB1 3435C > T might be independent positive prognostic factor in MM.


Introduction
Multiple myeloma (MM) is a malignant disease characterized by clonal expansion of plasma cells in the bone marrow [1]. Multi-drug resistance (MDR) is one of the major problems, which occurs during clinical therapy [2]. Among many mechanisms of MDR, the most prominent role is played by multidrug resistance-associated protein-1 (MRP1), also known as permeability-glycoprotein (P-GP) or ATP-binding cassette sub-family B member 1 (ABCB1) [3]. It acts as a transmembrane efflux pump that transfers drugs and toxins from cytosol to the extracellular matrix [4]. The ABCB1 protein is involved in the body's defense against endogenous and exogenous toxic compounds [5]. It is encoded by ABCB1 gene also known as MDR1 or P-GP (locus 7q21). A number of single nucleotide variants were identified in the ABCB1 gene, of which the most common studied are variants 3435C > T (in exon 26, rs1045642), 1236C > T (in exon 12, rs1128503) and 2677G > T/A (in exon 21, rs2032582) [6,7]. These variants may affect protein expression [8]. The variant 2677G > T/A affects the protein sequence (p.Ala893Ser/Thr) and function [6]. In contrast, variants present at nucleotides 3435 [ATC > ATT] and 1236 [GGC > GGT] are caused by synonymous substitution-Ile1145Ile and Gly412Gly, respectively [8]. The presence

Patients and Samples
The study group consisted of 110 newly-diagnosed patients with MM (Caucasian population), who were hospitalized ( Table 1. The control samples (Caucasian population) were obtained from 100 healthy blood donors (50 males and 50 females, with mean age 34.4 years) attending the regional Blood Donation and Blood Treatment Center in Kielce, Poland. All participants of the study provided written informed consent.
We followed the methods of Zmorzynski et al. 2020 [36]. The therapeutic induction regimens consisted of thalidomide and/or bortezomib combined with steroids and/or cyclophosphamide. Response to treatment was evaluated according to the International Myeloma Working Group guidelines, as described elsewhere [37,38]. Overall survival (OS) encompassed time from diagnosis until relapse, progression, death due to tumor effect or last follow-up, and time from diagnosis until death by any cause or last follow-up, respectively. Progression free survival (PFS) was estimated as the time elapsed between treatment initiation and tumor progression or death from any cause [36,39].
Peripheral blood (from healthy blood donors) and bone marrow aspirates (from MM patients) were used for DNA isolation and the determination of the ABCB1 and CYP1A1 variants.
Cell cultures from bone marrow aspirates were established to carry out the research associated with bortezomib (n = 50), as described by Zmorzynski et al. 2020 [36].

DNA Isolation
DNA was isolated with the use of a commercial kit (Qiagen, Hilden, Germany) according to manufacturer's procedure. The concentration and quality of DNA was checked using NanoDrop device (Thermo Fisher Scientific, Waltham, MA, USA).

ABCB1 Genotyping
The ABCB1 variants 1236C > T, 2677G > T/A were analyzed using automated DNA sequencing. The genotyping of 3435C > T variant was performed by polymerase chain reaction restriction fragment length polymorphism (PCR-RFLP).

Automated DNA Sequencing
The following primers were used to amplify 346 bp (containing 1236 nucleotide) and 223 bp (containing 2677 nucleotide), respectively.

PCR-RFLP
For analysis of 3435C > T variant PCR-RFLP method was applied according to validated protocol of Turgut et al., 2006 [40]. ABCB1 gene fragment length of 238 bp was amplified by PCR using primers: The PCR product was digested with MboI (Thermo Fisher Scientific, Waltham, MA, USA) according to manufacturer's procedure. Restriction enzyme produced two fragments of 172 bp and 60 bp or one fragment of 238 bp for presence of C or T allele, respectively. The fragments were analyzed on 3% agarose gel, stained with SimplySafe (Eurx, Gdańsk, Poland) and visualized in G:Box (Syngene, Cambridge, UK) ( Figure 2A). For each sample was carried out an independent PCR analysis.
For PCR product was divided into two parts (10 µL each, called P57 and P58) and amplified fragments were reamplified. The reaction mixture was the same as that for the first step except the primer mixture-reverse primer was used in two separate reactions. In the first reaction (so called P57), wild type primer (P57, WT, CYP1A1*1) 5 -GAA GTG TAT CGG TGA GAC CG -3 was used. The second reaction was carried out with the variant primer (P58, m2, CYP1A1*2C) 5 -GAA GTG TAT CGG TGA GAC CG-3 . The mixture was heated 95
For 6235T > C a total of 25 μL reaction mixture consisted of 100 ng genomic DNA, 10μM of each primer, 0.25 mM dNTPs mixture and 0.31 U of HD polymerase (Clontech, Takara Bio USA, San Jose, CA, USA) with 1× PCR reaction buffer (Clontech, Takara Bio USA, San Jose, CA, USA). The primers for PCR reaction were following:

Bortezomib In Vitro Treatment
Bone marrow aspirates (n = 50) (mean number of plasma cells-31.31% ± 20.69) were used with established cell cultures as described by Zmorzynski et al. 2019 [46]. Bortezomib (LC Laboratories, Woburn, MA, USA, 200 mg/mL) was dissolved in DMSO and added to cell cultures. The final concentration of DMSO in culture medium was lower than 0.1%. The cell cultures without bortezomib were used as a control. The cultures were grown without G-CSF and were routinely terminated. To determine the number of apoptotic, necrotic and viable cells Annexin V-Cy3 Apoptosis Detection Kit (Sigma-Aldrich, St. Louis, MO, USA), as well as fluorescence microscope were used. The viable cells were stained with 6-CF (6-carboxyfluorescein)-green dye, and necrotic cells showed red staining due to the presence of AnnCy3 (Annexin V Cy3.18). Cells starting apoptotic process were stained with green and red dyes. Plasma cells were analyzed according to Carter's et al. [47].

Statistical Analysis
Laboratory/clinical values were compared with studied genetic variants using an independent t-test and Chi-square test for continuous variables and categorical variables, respectively. The association of analyzed variants with clinical data was evaluated using Chi-square test or Fisher's exact test. The quantitative data was shown as frequency or percentage. Deviation of genotype frequencies in controls (healthy blood donors) and cases (MM patients) from Hardy-Weinberg equilibrium (HWE) was assessed by Chi-squared test [48]. The Cox proportional hazard model was used for univariate and multivariate analysis of OS and PFS. The Kaplan-Meier method and the log-rank test were used for survival analysis. We assumed a 5% error of inference and a related level of significance p < 0.05 pointing to the existence of statistically significant differences. Statistical analyzes were performed using the Statistica ver. 12.5 (StatSoft) software.

Frequencies of Alleles and Genotypes
Baseline characteristics for 110 MM patients included in this analysis are shown in Table 1. Genotyping of studied genetic variants was successfully performed for MM patients and healthy blood donors. The HWE test showed that the genotypic frequencies of CYP1A1 4889A > G variant were not in HWE for MM patients, because the p value was lower than 0.05 and χ 2 was higher than 3.84 (Table 2).

Studied Variants and the Risk of MM Development
The CYP1A1 6235T > C and ABCB1 gene variants were balanced. The differences in genotypic and allelic frequencies of ABCB1 2677G > T/A (TT genotype and T allele), as well as CYP1A1 variants-6235T > C (WT/m1 genotype and m1 allele) and CYP1A1 4889A > G (m2/m2 genotype and m2 allele), between study and control groups were statistically significant ( Table 3).
The studied ABCB1 and CYP1A1 variants influenced the risk of MM. The presence of TT genotype of ABCB1 2677G > T/A polymorphism increased MM risk 2.87-fold (p = 0.007) ( Table 3). The presence of CYP1A1*2A (m1) and CYP1A1*2C (m2) alleles decreased the risk of the disease (Table 3).

ABCB1 and CYP1A1 Variants as a Risk Factors of Death or MM Progression
Rare homozygotes were analyzed together with heterozygotes due to small sample size. A univariate Cox analysis revealed that patients at stage III according to ISS and without auto-HSCT had a 3.03-fold and 6.05-fold increased risk of death (Table 4). Similar findings were observed in the case of disease relapse or progression in MM patients at stage III (HR = 3.12, p = 0.001) and without auto-HSCT (HR = 3.07, p = 0.001) ( Table 4). Moreover, the univariate Cox analysis showed decreased risk of death in MM patients with CT + TT genotypes of ABCB1 3435C > T variant (HR = 0.34, p = 0.04) ( Table 4). The analysis of response rate in MM showed that patients at stage III or without auto-HSCT had an increased chance of progressive disease-OR = 7.66 (p < 0.001) or OR = 5.59 (p = 0.001), respectively. The studied variants of ABCB1 and CYP1A1 genes did not affect the response to treatment of MM patients.

Correlation between Analyzed Polymorphisms and MM Clinical Features
Potential relationships between clinical data and selected genotypes were analyzed. We found, that wt/m2 + m2/m2 genotypes in comparison to wt/wt genotype of CYP1A1 4889A > G were associated with higher free light chain ratio (618.74 vs. 177.09, p < 0.001), β2-microglobulin concentration (mg/L) (7.68 vs. 5.66, p = 0.03) and creatinine concentration (mg/dL) (2.53 vs. 1.38, p = 0.003) ( Table 6). Moreover the presence of CYP1A1*2A allele (of 6235T > C variant) was associated with higher number of platelets (K/µL) (196.4 vs. 241.1, p = 0.01) ( Table 6). We did not observe relationships of the studied variants with high-risk chromosomal abnormalities, smoking status, exposure to carcinogens, or family history of cancer. The multivariate Cox regression analysis confirmed that that patients without auto-HSCT had increased risk of death (Table 5). Moreover, it showed decreased risk of death in the group of MM patients with CT + TT genotypes of ABCB1 3435C > T variant -HR = 0.34, p = 0.04 (Table 5).

Survival of MM Patients and Studied Variants
We analyzed the association between studied genotypes and survival of MM patients. In the log rank test (Figures 3 and 4), without taking into account the type of treatment, the difference in OS between CC vs. CT + TT genotypes of ABCB1 3435C > T was observed (Figure 3).
It was performed a log rank analysis taking into account studied variants and the type of treatment (thalidomide vs. bortezomib vs. both-thalidomide and bortezomib). We found association of CT genotype of ABCB1 1236C > T with the treatment type and OS (p = 0.014) or PFS (p = 0.021) ( Figure 5). Taking into account the type of treatment and OS similar results were observed in the case of GT genotype of ABCB1 2677G > T/A variant (p = 0.013) and CT genotype of ABCB1 3435C > T variant (p = 0.005). Moreover, in the analysis of treatment and PFS we found association with CT genotype of ABCB1 1236C > T (p = 0.021), CT genotype of ABCB1 3435C > T (p = 0.033), wt/wt genotype of CYP1A1 6235T > C (p = 0.016) and wt/wt genotype of CYP1A1 4889A > G (p = 0.021) ( Figure 5).

In Vitro Study with Bortezomib
In in vitro studies, bortezomib increased the number of apoptotic and necrotic cells in all studied genotypes. The higher number of apoptotic cells was observed at 1 nM and 2 nM of bortezomib in patients with wt/wt genotype (of CYP1A1 4889A > G) and CC genotype (of ABCB1 3435 C > T) in comparison to those with wt/m2 + m2/m2 genotypes (20.67% vs. 13.06%, p = 0.02) and CT + TT genotypes (21.90% vs. 16.60%, p = 0.03), respectively ( Figure 6A,B). Higher number of viable cells was found at 1 nM and 12 nM of bortezomib in cells with wt/m2 + m2/m2 (of CYP1A1 4889A > G) and GT + TT genotypes (of ABCB1 2677G > T/A) in comparison to wt/wt and GG genotypes, respectively ( Figure 6C,D). In the case of 3435C > T variant was observed tendency for lower number of viable cells with CC genotype in comparison to CT + TT genotypes at 4 nM and 12 nM of bortezomib-58.8% vs. 67.65% and 48.87% vs. 54.41%, respectively ( Figure 6E). It was performed a log rank analysis taking into account studied variants and the type of treatment (thalidomide vs. bortezomib vs. both-thalidomide and bortezomib). We found association of CT genotype of ABCB1 1236C > T with the treatment type and OS (p = 0.014) or PFS (p = 0.021) ( Figure 5). Taking into account the type of treatment and OS similar results were observed in the case of GT genotype of ABCB1 2677G > T/A variant (p = 0.013) and CT genotype of ABCB1 3435C > T variant (p = 0.005). Moreover, in the analysis of treatment and PFS we found association with CT genotype of ABCB1 1236C > T (p = 0.021), CT genotype of ABCB1 3435C > T (p = 0.033), wt/wt genotype of CYP1A1 6235T > C (p = 0.016) and wt/wt genotype of CYP1A1 4889A > G (p = 0.021) ( Figure 5).

In Vitro Study with Bortezomib
In in vitro studies, bortezomib increased the number of apoptotic and necrotic cells in all studied genotypes. The higher number of apoptotic cells was observed at 1 nM and 2 nM of bortezomib in patients with wt/wt genotype (of CYP1A1 4889A > G) and CC genotype (of ABCB1 3435 C > T) in comparison to those with wt/m2 + m2/m2 genotypes (20.67% vs. 13.06%, p = 0.02) and CT + TT genotypes (21.90% vs. 16.60%, p = 0.03), respectively ( Figure 6A,B). Higher number of viable cells was found at 1 nM and 12 nM of bortezomib in cells with wt/m2 + m2/m2 (of CYP1A1 4889A > G) and GT + TT genotypes (of ABCB1 2677G > T/A) in comparison to wt/wt and GG genotypes, respectively (Figure 6C,D). In the case of 3435C > T variant was observed tendency for lower number of viable cells with CC genotype in comparison to CT + TT genotypes at 4 nM and 12 nM of bortezomib-58.8% vs. 67.65% and 48.87% vs. 54.41%, respectively ( Figure 6E).

Discussion
In the current study, we analyzed the correlation of ABCB1 and CYP1A1 variants with the risk and the outcome of MM, as well as response to bortezomib treatment under in vitro conditions. We found relationships between the ABCB1 2677G > T/A, CYP1A1 6235T > C and 4889A > G variants with the risk of MM. Moreover, we found that the ABCB1 3435C > T variant affected the OS.
The mechanisms associated with ABC transporters are responsible for the development of the MDR phenotype in MM patients [8]. The tumor cells were able to overcome the drugs cytotoxicity. This was the main cause of treatment failure in MM [8]. The identification of new prognostic and predictor factors may help to define subgroups

Discussion
In the current study, we analyzed the correlation of ABCB1 and CYP1A1 variants with the risk and the outcome of MM, as well as response to bortezomib treatment under in vitro conditions. We found relationships between the ABCB1 2677G > T/A, CYP1A1 6235T > C and 4889A > G variants with the risk of MM. Moreover, we found that the ABCB1 3435C > T variant affected the OS.
The mechanisms associated with ABC transporters are responsible for the development of the MDR phenotype in MM patients [8]. The tumor cells were able to overcome the drugs cytotoxicity. This was the main cause of treatment failure in MM [8]. The identification of new prognostic and predictor factors may help to define subgroups of patients that most likely can benefit from chemotherapy [49]. Among analyzed variants of the ABCB1 gene, including 3435C > T (synonymous variant), 1236C > T (synonymous variant) and 2677G > T/A (non-synonymous variant), the most important seems to be the one located at nucleotide 3435 [50]. It leads to changes at the mRNA level and affects protein folding [51]. Some studies have confirmed the relationship of the ABCB1 3435 C > T variant with MM risk. However, published data are still inconclusive [50]. The meta-analysis performed by Razi and co-workers did not find an association of ABCB1 3435C > T with MM susceptibility [50]. In contrast, we found the relationship of TT genotypes of 3435C > T with an almost threefold increased risk of MM development. It is possible that the ABCB1 3435C > T variant is in linkage disequilibrium with some alleles, which may impact MM development [6]. Similarly, the presence of the T-allele of ABCB1 2677G > T/A variant was associated with a higher MM risk. Drain et al. in a study of 134 MM patients, found an association between the CC genotype of ABCB1 3435C > T variant with shorter OS, which is in line with our observations [8]. Similar results were obtained by Drain et al. in another study [52]. Moreover, they did not find a relationship between the ABCB1 1236C > T and 2677G > T/A variants with OS [8]. In our study, we observed that allele frequencies and distributions of analyzed ABCB1 variants were comparable in MM patients and healthy blood donors, which is in agreement with previous reports concerning healthy Caucasian populations [5,6]. Silent variants at the 3435 nucleotide of the ABCB1 gene can alter ABCB1 protein conformation, which may change the specificity of binding with substrates [53]. ABCB1 3435C > T variant is present in the gene coding sequence (exon 26), and it is likely linked to other regions that regulate gene expression, for example the promoter or enhancer sequence or regions associated with mRNA processing [49]. It is suggested that the ABCB1 3435C > T variant may be significant only in individuals with specific carcinogen exposure [6]. However, in our study we did not find an association of the analyzed variants with carcinogen exposure, including smoking. Our findings suggest that the T-alleles (CT and TT genotypes) of ABCB1 3435C > T variant might be associated with a lower risk of death in MM patients. Univariate and multivariate Cox analysis suggested that the presence of T (at 3435 nucleotide) might be a positive prognostic factor. The T-allele of this variant can be associated with lower ABCB1 gene expression in comparison to C-allele [54]. Moreover, we observed longer OS for patients with the CT genotype of ABCB1 3435C > T treated with thalidomide and bortezomib in comparison to those with CT genotype treated with thalidomide or bortezomib. This effect might be not observed in TT genotypes due to the small sample size. In the case of second analyzed synonymous variant of ABCB1 gene-1236C > T, we found only association of CT genotype with treatment type and OS/PFS. A more detailed analysis on a larger cohort would be recommended.
The presented study showed that CYP1A1*2A and CYP1A1*2C alleles decreased the risk of MM development. Similar results were obtained by Kang et al., 2008 [30]. However, their study was conducted among an Asian population [30]. The previous report by Lincz et al. suggested no association of CYP1A1 variants with MM in Caucasians [55]. In our study, the CYP1A1 4889A > G variant was not in HWE. The CYP1A1*2A (m1) and CYP1A1*2C (m2) alleles may enhance the enzyme activity and therefore the metabolism of corresponding substrates may be at a higher level [23]. CYP1A1 6235T > C variant raises the risk for solid tumors, for example lung cancer, cervical cancer, prostate cancer and laryngeal cancer [56][57][58][59]. In contrast, negligible relations between CYP1A1 6235T > C variant and gastric cancer, colorectal cancer, breast cancer and esophageal cancer risks were found [60][61][62][63]. Therefore this variant may play a diverse role in different cancers.
The clinical stage of MM can be measured by serum markers, including β2-microglobulin, free light chain ratio, creatinine and C-reactive protein levels followed by confirmation with invasive biopsy of bone marrow [64]. In our study, we found an association of the CYP1A1*2C (m2) allele with a higher free light chain ratio, and higher β2-microglobulin and creatinine concentrations, which are negative prognostic factors. In contrast, the presence of the CYP1A1*2C allele decreased the risk of MM. These results are not consistent and require studies on a larger cohort. Moreover, CYP1A1 4889A > G genotypes were not in HWE, which suggest their role in disease susceptibility. However, due to low sample size this result could be obtained by error sampling.
Interactions between MM cells and the bone marrow microenvironment play a critical role in the development of MDR [65]. We observed the relationship of the ABCB1 and CYP1A1 variants with response to bortezomib treatment under in vitro conditions. In the case of T-allele of ABCB1 2677G > T/A, we noted a statistically significant higher number of viable cells at 12nM of bortezomib. However, we did not observe significant changes in the level of apoptotic and necrotic cells, taking into account the presence of the T-allele of ABCB1 2677G > T/A. It is possible that this result was obtained accidentally or that higher doses of bortezomib should be applied. Similarly, the presence of a T-allele of ABCB1 3435 C > T variant was associated with a lower number of apoptotic cells, but only at 2 nM of bortezomib. The CYP1A1*2C (m2) allele affected the number of viable cells and apoptotic cells only at 1nM of bortezomib. More reliable results would be associated with a significant change in the number of cells with increasing bortezomib dose. It would be recommended to repeat the in-vitro experiment with modified conditions, including the replacement of RPMI with AIM-V media. RMPI and 10% FCS induce apoptosis of primary B-cells [66].
There are some limitations of our study regarding the number of MM patients, as well as apoptosis detection in the in vitro experiment. The number of participants in the study group was relatively small, which was due to the low incidence of MM. However, the number of 110 MM patients was enough for most analyzes. Fluorescent microscopy was used to evaluate apoptosis the in in-vitro study. A more accurate method for apoptosis detection is FACS (flow cytometry-based apoptosis detection). Unfortunately, during the experiment time, FACS was not available to us. The set used for apoptosis analysis was dedicated and validated to fluorescent microscopy.
Several large scale genome-wide association studies (GWASs) have failed to identify ABCB1 and CYP1A1 polymorphisms in association with MM [67][68][69]. Although GWAS was made in the United States and in European ancestry populations, it is possible that in the Polish population the studied variants may have a prognostic significance [69][70][71]. In the case of ABCB1 variants, a similar study in the Polish population was made by Jamroziak and coworkers [6]. They did not find whether common ABCB1 variants affect predisposition to MM [6]. However, we found an association of the ABCB1 2677G > T/A variant with an increased risk of MM. This result might be due to error sampling or due to the fact that the study included the southwestern Polish population. Further analysis in this field would be recommended.

Conclusions
In spite of this study's limitations and the need for prospective studies with larger sample sizes, our findings suggest that the T-allele of ABCB1 2677G > T/A increased the risk of MM. In contrast, CYP1A1*2A (m1) and CYP1A1*2C (m2) alleles decreased the susceptibility of this disease in the Caucasian population in southeastern Poland. Moreover, the C-allele of ABCB1 3435C > T was associated with shorter OS. Further analysis on a larger cohort could help to better understand the significance of studied variants in the development and outcome of MM.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The clinical data used to support the findings of this study are available from the corresponding author upon request.