Influence of CYP2B6 and CYP3A4 polymorphisms on the virologic and immunological responses of patients treated with efavirenz containing regimen

Abstract

Objectives

The main objective of this study was to evaluate the effect of CYP2B6 and CYP3A4 polymorphisms on the virological and immunologic responses of HIV patients. A total of 153 HIV-positive patients were enlisted for the study.

Patients and methods

Viral load and median CD4 T cell counts were evaluated at baseline and month 6 (M6). Samples were identified using TaqMan genotyping assays.

Results

The AG in CYP2B6 rs2279343 was associated with VLS compared to homozygous AA. In the dominant model, the AG/GG genotypes were associated with VLS compared to the AA genotype. Moreover, in overdominant model, the AG genotype was associated with VLS compared to AA/GG. Regarding immunological response, only the AG in SNP rs2279343 CYP2B6 was associated with an increase in CD4 cell count between baseline and M6. In CYP2B6 rs3745274, the CD4 cell count at M6 was higher than that of baseline for GG carriers and for GT carriers. In CYP3A4 rs2740574, the TC carriers showed a higher median CD4 count at M6 compared to that of the baseline count, as well as for CC carriers. The best genotypes combination associated with CD4 cell count improvement were AA/AG in SNP rs2279343 and GG/GT in SNP rs3745274.

Conclusion

Our findings support the fact that CYP2B6 rs2279343 could help in the prediction of VLS and both SNPs rs3745274 and rs2279343 in CYP2B6 and CYP3A4 rs2740574 were associated with immune recovery in Malian HIV-positive patients.

Introduction

Human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS) is and continues to be a major public health issue around the world, particularly in developing countries. Globally, the number of people living with the disease was close to 37.7 million in 2020 [1]. Progress noted at the clinical, diagnostic, and biological levels has made it possible to transform AIDS now into a chronic disease for patients observing treatment [2,3]. In Sub-Saharan African countries, including Mali, AIDS is treated with efavirenz (EFV), a non-nucleoside reverse transcriptase inhibitor that is commonly used in HIV-positive patients in combination with nucleoside reverse transcriptase inhibitors [4,5]. Furthermore, the EFV-containing regimen can be given to any patient, even when paired with rifampicin in patients co-infected with tuberculosis [6,7]. Therefore, EFV is a drug of choice for the management of HIV-positive patients in developing countries, such as Mali. However, some patients on ARV treatment often experience serious side effects such as central nervous system-related toxicity associated with an inadequate clinical response [8,9], leading some to discontinue the medication. This interindividual variability in ARVs treatment outcomes can be explained in part by the physiological state of the individual, the existence of comorbidities, or environmental factors [10–12]. Differences in treatment response, on the other hand, could be explained by mutations in the viral genome [13] as well as interindividual genetic variations in genes involved in ARVs metabolism, such as cytochrome P450 genes [14–16]. It is well established that EFV metabolism is primarily mediated by the CYP2B6 gene which converts EFV to 8-hydroxy-efavirenz, but other isoforms such as CYP1A2, CYP3A4, and CYP3A5 also play a role [17]. Individual genetic variations, such as single nucleotide polymorphisms (SNPs) associated with functional loss or gain, can have a significant impact on the intra- and inter-patient plasma concentrations of EFV [18–20]. The Clinical Implementation of Pharmacogenetics Consortium (CIPC) published a guideline for CYP2B6 and antiretroviral therapy with efavirenz in 2019 [21,22]. Several single nucleotide variants in CYP2B6 gene, including rs3745274, rs2279343, and rs28399499, extensively investigated in EFV pharmacogenetic studies are included in the CIPC guidelines. The CYP2B6 rs3745274, which results in a change from G to T at position 516 in exon 4, increases the plasma concentration of EFV with consequences for the central nervous system. However, at exon 5, the SNP rs2279343 which results in the change from A to G at position 785 leads to a decrease in the catalytic activity of the gene [7,15,16,18]. As a result, the presence of these SNPs may be positively or negatively correlated with the EFV plasma concentrations [23,24], which would explain the occurrence of side effects and delay in virological suppression. This usually leads some patients to abandon treatment and constrains clinicians to change treatment lines. In a previous study on the healthy Malian population, we found the allele frequencies were 37% for CYP2B6 rs3745274, 38% for CYP2B6 rs2279343, and 75% for CYP3A4 rs2740574 [25]. Several studies around the world have shown that SNPs in cytochrome P450 genes are involved in EFV treatment outcomes. However, by carefully checking the literature review, these data seem to be lacking for the Malian population. Therefore, the present study is carried out to assess the relationship between the rs3745274; rs2279343 in the CYP2B6 gene and the rs2740574 in the CYP3A4 gene on EFV treatment outcomes in a sample of the Malian population. Results from this study will provide one the first EFV pharmacogenetic data in our population and will facilitate the evidence-based management of HIV-positive patients.

Materials and Methods

Study participants

In the current study, a total of 153 HIV-positive patients with a median age of 35 years range (18-71 years) were enrolled. The clinical follow-up of patients was conducted at the infectious and tropical diseases department of the university hospital of Point G, Bamako, and at the Center of Listening, Care, Animation, and Counseling of people living with HIV, Bamako, Mali during the period 2018 to 2021. The study was approved by the Ethics Committee of the Faculty of Medicine and Odontostomatology (FMOS) /Faculty of Pharmacy (FAPH), University of Sciences, Techniques and Technologies of Bamako (USTTB) under the number 2018/113/CE/FMPOS. Following a detailed explanation of the study protocol, informed consent was provided by all participants who satisfied the eligibility criteria. Patients were treated with a regimen containing EFV 600 mg and two nucleoside reverse transcriptase inhibitors (tenofovir 300 mg and lamivudine 300 mg) and followed for 6 months. An operating sheet was established for the collection of socio-demographic, clinical, and biological data from patients by clinicians on D1, D15, M1, M3, and M6. Viral load and CD4 count were measured at D1 and M6. Five milliliters of peripheral venous blood from each participant were drawn into an EDTA tube and kept at minus 20 °C until DNA extraction.

Identification of CYP2B6 (rs3745274; rs2279343) and CYP3A4 (rs2740574) genetic profiles

The Gentra Puregene Blood Kit was used to extract genomic DNA from white blood cells. Patients’ genetic profiles were identified by TaqMan drug metabolism genotyping assays for allelic discrimination (Applied Biosystems Genotyping Assays). The assays number used were C_7817765_60 rs3745274 for CYP2B6, C_1837671_50 rs2740574 for CYP3A4 rs2740574, we performed a custom designed TaqMan assay for CYP2B6 rs2279343. The 7500 Fast Real-Time PCR System (Applied Biosystems, 850 Lincoln Centre Drive, Foster City, California 94404, USA) has been used to identify the different SNPs. The composition of the PCR mixture, as well as the PCR run method, have been previously described in detail by Kassogue et al [25].

Statistical analysis

Statistical package SPSS version 16 (SPSS Inc., Chicago, IL, USA) was used to analyze the data. The Chi-square (X²) test or Fisher’s exact test was used to estimate the difference between the distribution of genotypes and EFV treatment outcome. Odds ratio (OR) with confidence interval (CI) at 95% was calculated to measure the strength of association of an individual’s genetic profile with viral load suppression. Mann-Whitney U tests were used to compare the baseline CD4 T cell count to that of M6 for each genotype as well as for each genetic model. SNPStats was used to test Hardy-Weinberg equilibrium between observed and expected alleles to estimate the haplotype distribution and finally to assess their association with ARVs treatment outcomes [26]. A p < 0.05 (two-sided) was considered as significant.

Results

Baseline characteristics

Clinical and demographic characteristics of study participants are summarized in Table 1. Women were more represented than men with 60.1% versus 39.9%. We noted that 51% of our participants had a median age ≤ 36 years and a median weight ≤ 58 kg. We observed that 56.2% of our patients had a baseline viral load < 10,000 copies/mL, 19.6% between 10,000-100,000 copies/mL and 24.2% a viral load > 100,000 copies/mL. Regarding the baseline CD4 count, 45.1% had a rate < 200 Cell/mm³ followed by 19% with a CD4 rate between 200-350 Cell/mm³ and 35.9% with a CD4 rate > 350 Cell/mm³.

Table 1. Demographic and clinical features of study participants at baseline and at month 6.

Demographic and baseline clinical features of study participants

Parameters		N (%)
Gender
Female		92 (60.1)
Male		61 (39.9)
Age (years)
≤ 36		78 (51)
> 36		75 (49)
Median age (IQR)	36 (30-44)
Weight (Kg)
≤ 58		78 (51)
> 58		75 (49)
Median weight (IQR) at baseline	58 ((51-58)
Median weight (IQR) at M6	63 (56-72.25)
HIV-1 RNA copies ml-1
< 10000		86 (56.2)
10000-100000		30 (19.6)
> 100000		37 24.2)
Median RNA copies at baseline	7300 (493-130010)
Median RNA copies at M6	49 (49-278.50)
CD4 T Cell/mm3
< 200		69 (45.1)
200-350		29 (19)
> 350		55 (35.9)
Median CD4 T Cell/mm3 at baseline	246 (102-469.75)
Median CD4 T Cell/mm3 at M6	393 (203.75-574.25)

N : number of patients, IQR : interquartile range

Virologic response

The distribution of CYP2B6 (rs3745274; rs2279343) and CYP3A4 (rs2740574) regarding viral load suppression (VLS) status of patients at month 6 are summarized in Table 2. Allele distribution did not deviate from Hardy-Weinberg equilibrium neither in women nor in men. We found that the distribution of the different genotypes of the SNP rs3745274 in the CYP2B6 gene was statistically similar between patients who achieved VLS and those who did not achieve it, p > 0.05. The same trend was observed in the different genetic models, including recessive, dominant, and overdominant models, p > 0.05. Interestingly in CYP2B6 rs2279343, the heterozygous AG was significantly associated with VLS achievement compared to the homozygous AA wild-type genotype (OR 2.6; 95% CI 1.3–5.3; p = 0.01). The homozygous GG mutant genotype compared to AA wild-type as well as the recessive model were not found to be associated with VLS, p > 0.05. However, individuals harboring AG/GG genotypes were more likely to achieve VLS compared to AA wild-type genotype carriers in the dominant genetic model (OR 2.5; 95% CI 1.3–5; p = 0.01). Moreover, in the overdominant model, we noted that the heterozygous AG was significantly associated with VLS when compared to individuals harboring AA/GG genotype (OR 2.05; 95% CI 1.3–2.9; p = 0.03). Looking at the distribution of rs2740574 in the CYP3A4 gene, we did not observe any trend. The frequencies of the different genotypes in the different genetic models were comparable between the two groups of patients living with HIV (see table 2). In Table 3, the distribution of haplotypes showed no trend in terms of viral load suppression, p > 0.05.

Table 2. Distribution of alleles and genotypes according to viral suppression status at 6 months of ARVs treatment.

SNPs	VLS N (%)	No VLS N (%)	OR (95% CI)	P value
rs3745274 CYP2B6 (N = 153)
GG	30 (33.7)	25 (39.1)	Ref.
GT	49 (55.1)	30 (46.9)	1.4 (0.6-2.7)	0.5
TT	10 (11.2)	9 (14.1)	0.9 (0.3-2.6)	1
GG/GT a	79 (88.8)	55 (85.9)	Ref.
TT	10 (11.2)	9 (14.1)	0.7 (0.3-2)	0.6
GG b	30 (33.7)	25 (39.1)	Ref.
GT/TT	59 (66.3)	39 60.9)	1.3 (0.6-2.4)	0.5
GG/TT c	40 (54.1)	34 (45.9)	Ref.
GT	49 (62)	30 (38)	1.38 (0.7-2.6)	0.33
HWE: p-value	0.13	1
rs2279343 CYP2B6 (N = 153)
AA	24 (27)	31 (48.4)	Ref.
AG	52 (58.4)	26 (40.6)	2.6 (1.3-5.3)	0.01
GG	13 (14.6)	7 (11)	2.4 (0.8-6.9)	0.12
AA/AG a	76 (85.4)	57 (89.1)	Ref.
GG	13 (14.6)	7 (10.9)	1.2 (0.5-3.1)	0.6
AA b	24 (27)	31 (48.4)	Ref.
AG/GG	65 (73)	33 (51.6)	2.5 (1.3-5)	0.01
AA/GG c	37 (49.3)	38 (50.7)	Ref.
AG	52 (66.7)	26 (33.3)	2.05 (1.3-2.9)	0.03
HWE: p-value	0.08	0.3
rs2740574 CYP3A4 (N = 151)
TT	4 (4.5)	2 (3.2)	Ref.
CT	28 (31.8)	19 (30.2)	0.7 (0.1-4.4)	1
CC	56 (63.6)	42 (66.7)	0.6 (0.1-3.8)	1
TT/TC a	32 (36.4)	21 (33.3)	Ref.
CC	56 (63.6)	42 (66.7)	0.8 (0.4-1.7)	0.7
TT b	4 (4.5)	2 (3.2)	Ref.
CT/CC	84 (95.5)	61 (96.8)	0.7 (0.1-3.8)	1
TT/CCc	60 (57.7)	44 (42.3)	Ref.
TC	28 (59.6)	19 (40.4)	1.1 (0.5-2.2)	0.8
HWE: p-value	0.8	0.8

VLS: viral load suppression (HIV-1 RNA copies ml^-1less than 50 copies), No VLS (HIV-1 RNA copies ml^-1higher than 50 copies),

^arecessive model,

^bdominant model,

^coverdominant model

Table 3. Haplotype distribution of CYP2B6 polymorphism according to viral suppression status from SNPs rs3745274 and rs2279343.

Haplotypes	VLS (%)	No VLS (%)	OR (95% CI)	P-value
GA	53.69	58.06	Ref.	-
TG	36.28	26.81	1.52 (0.87-2.67)	0.15
GG	7.54	4.44	1.91 (0.68-5.38)	0.22
TA	2.49	10.69	0.44 (0.16-1.19)	0.11

Global haplotype association p-value: 0.032; VLS: (HIV-1 RNA copies ml^-1less than 50 copies), No VLS (HIV-1 RNA copies ml^-1higher than 50 copies).

Immunologic response

As shown in Figure 1, we tested the immunological response by comparing the median baseline CD4 cell count to that at month 6 for each SNP. We noted that the median CD4 count at month 6 (M6) was not statistically different from the baseline CD4 count for individuals carrying the wild-type AA genotype of CYP2B6 rs2279343 (312 vs. 192.5; U = 1225, p = 0.06). The same trend was observed with the mutant GG genotype showed a similar trend (342.5 vs. 306; U = 203, p = 0.1). In contrast, we noted that the M6 median CD4 count was statistically higher than that of the baseline for the heterozygous AG (433.5 vs. 306, U = 2278.5, p = 0.01). Regarding CYP2B6 rs3745274, a significant statistical difference was observed between CD4 count at M6 and baseline CD4 count for the GG wild-type genotype (333 vs. 199, U = 1238, p = 0.04) and for the heterozygous GT genotype (412 vs. 277, U = 2372, p = 0.013). In contrast, the TT mutant genotype showed no trend in terms of immunological improvement (382 vs. 314, U = 180.5, p = 0.3). In CYP3A4 rs2740574, individuals with the heterozygous TC genotype at month 6 showed a higher median CD4 count compared to baseline CD4 count (539 vs. 341, U = 844.5, p = 0.024). Individuals carrying the homozygous mutant CC genotype showed a similar trend (358 vs. 206, U = 8757, p = 0.004). However, for those with TT wild-type genotype, the median CD4 counts at M6 and at baseline were statistically comparable (398 vs. 189, U = 22, p = 0.8). As shown in Figure 2, individuals carrying the AA/AG genotypes of SNP rs2279343 showed at month 6 a higher median CD4 count compared to baseline CD4 count (394 vs. 266.5, U = 6936, p = 0.003). This trend was observed within individuals harboring GG/GT genotype of SNP rs3745274 (394 vs. 246, U = 7079.5, p = 0.002). However, individuals carrying TT/TC genotypes SNP rs2740574 did not show a statistical difference between the median CD4 count at month 6 and baseline (462.5 vs. 324, U = 1178, p = 0.046). In addition, the haplotypes resulting from the two SNPs in the CYP2B6 gene were not associated with immune recovery.

Figure 1.

Median (interquartile range) evolution in CD4 T cell count from baseline to M6 in patients receiving efavirenz-containing therapy according to the different genotype profiles in the three SNPs. The light gray boxplot represents the baseline CD4 T cell count, while the dark gray boxplot represents the median CD4 T cell count at M6. The horizontal lines represent the median and the lower and upper limits of the boxplot represent 25% percentile and 75% percentile. Mann–Whitney test were used to compare CD4 T cell count from baseline to that of M6.

Figure 2.

Median CD4 T cell count evolution from baseline to M6 in patients regarding the different combinations of wild-type homozygous/heterozygous in the three SNP. The light gray boxplot represents the baseline CD4 T cell count, while the dark gray boxplot represents the median CD4 T cell count at M6. The horizontal lines represent the median and the lower and upper limits of the boxplot represent 25% percentile and 75% percentile. Mann–Whitney test were used to compare CD4 T cell count from baseline to that of M6.

Discussion

CYP P450 plays a central role in the biotransformation of a wide range of drugs, including anticancer drugs, steroid hormones, and anti-infectives. [27–31]. Moreover, genes encoding CYP P450 enzymes are highly polymorphic. Thus, a group of individuals suffering from a given disease and taking the same drug may display varied therapeutic responses depending on their genetic background. Indeed, the therapeutic response of an individual is closely related to his genetic profile. In the present study, we noted that the frequencies of the SNPs investigated were comparable to the frequencies reported in our previous study carried out on the Malian healthy population [25]. The SNPs explored in the present study were in Hardy-Weinberg equilibrium. We found that the frequencies of the different alleles and genotypes of the SNP rs3745274 in exon 4 of the CYP2B6 gene were statistically similar between patients who reached viral load suppression (VLS) at 6 months and those who did not. Chang et al. reported similar results in HIV-positive patients in a study in Uganda [32]. Another study conducted in Port-au-Prince corroborated our findings [33]. Contrary to our results, Ribaudo et al. found that black patients with 516/983 CYP2B6 poor metabolizer profiles had the lowest cumulative rate of virologic failure [34]. However, patients with the AG genotype, which corresponds to intermediate metabolizer, were 2.6 times more likely to reach VLS than those with the other genotypes for the SNP CYP2B6 rs2279343 in exon 5. This SNP has been linked to a loss-of-function, resulting in decreased protein synthesis [35]. Therefore, having an A allele that contributes 50% of protein synthesis could be beneficial in achieving VLS. Patients carrying the AG/GG genotype in the dominant genetic model were 2.5 times more likely to reach VLS at 6 months. Furthermore, the AG genotype was found to be associated with VLS with an OR of 2.05 when compared to the AA/GG carriers. Conversely, the SNP rs2740574 in the CYP3A4 gene was not associated with VLS in any of the genetic models tested. On the other hand, haplotypes deriving from SNPs rs2279343 and rs3745274 were not associated with the VLS at M6. In addition, it should be noted that the current study only examined two SNPs in the highly polymorphic CYP2B6 gene. Therefore, results for other SNPs could be masked as reported in the study by Rotger et al. [36]. Interestingly, patients with the AG genotype showed a statistically significant improvement in immune response when compared to other genotypes of the CYP2B6 rs2279343. This shows that having the heterozygous genotype of this SNP enhances immune recovery and so protects patients from opportunistic infections. Regarding the CYP2B6 rs3745274, we noted that only patients with the mutant TT genotype did not show a significant improvement in immunological response at M6. This could be explained by the fact that these so-called poor metabolizers are unable to properly control viral replication. As a result, they will be exposed to greater EFV plasma concentrations, which could lead to central nervous system-related side effects, noncompliance, or even cessation of treatment. Contrary to our findings, Oluka et al. observed that the CYP2B6 rs3745274 was not associated with immunological recovery in patients treated with a nevirapine-containing regimen in Kenyan women based-study [37]. Regarding the CYP3A4 rs2740574, conversely to the other profiles, the TT genotype showed a statically comparable CD4 count at baseline and M6. This observation could be explained by the small number of patients with TT genotype. Overall, the effect of CYP3A4 rs2740574 on the gene's expression is still debatable [38,39]. Except for the combined TT/TC genotype in the CYP3A4 gene, the genotypes AA/AG (rs2279343) and GG/GT (rs3745274) in the CYP2B6 gene exhibited a significant improvement in the immune response when considering the combination of the genotypes "wild-type homozygous + heterozygous".

Conclusion

The present study examined three pharmacogenetically relevant SNPs in the CYP 450 gene, including CYP2B6 and CYP3A4. Interestingly, we noted that heterozygous AG (rs2279343) was associated with VLS at M6. Moreover, for the same SNP, both dominant and overdominant genetics models were found to be associated with VLS. In terms of immunological recovery, substantial improvement was observed in carriers of heterozygous genotypes for the two SNPs in the CYP2B6 gene as well as those with wild-type homozygous-heterozygous combined genetic profiles. These findings could aid in the implementation of pharmacogenetics in our country, particularly for HIV-positive patients on an efavirenz-based regimen.