Pharmacogenetic predictors of nausea and vomiting of pregnancy severity and response to antiemetic therapy: a pilot study

Nausea and vomiting of pregnancy (NVP) affects up to 80% of all pregnant women [1–3]. Although NVP is most common during the first trimester of pregnancy, as many as 20% of pregnant women are affected beyond 12 weeks of gestation and throughout the day (vs. isolated to the morning hours). Of the women affected, 1 to 3% experience a severe form of NVP called hyperemesis gravidarum (HG) which includes weight loss, dehydration, and electrolyte imbalances. In mild cases, NVP causes discomfort and inconvenience; however, HG can pose significant risks to the health of the pregnant woman and the fetus—often necessitating hospitalization [4]. Even the milder cases of NVP can have a significant impact on the quality of a woman’s life and contribute significantly to health care costs and time lost from work [2, 3, 5]. Termination of otherwise wanted pregnancies has been reported in women suffering from severe, prolonged NVP [6]. However, because NVP is rarely life-threatening, the goal of pharmacologic treatment is typically improvement in the quality of life for affected pregnant women [7].

Several dopamine antagonist antiemetics commonly used in pregnancy also act as weak serotonin receptor antagonists. These include promethazine, prochlorperazine, and metoclopramide. Additionally, the 5-HT₃ receptor antagonist ondansetron is an often used antiemetic in pregnancy. These antiemetics have been studied extensively and all seem to be safe for use in pregnancy [7–10]. However, response to these medications in the treatment of NVP is highly variable [7, 11–13]. While pharmacologic intervention may ameliorate symptoms in many women, some women improve dramatically while others continue to experience severe NVP despite treatment. As such, pharmacogenetic polymorphisms become an important consideration in evaluating variability in responsiveness to pharmacologic intervention. Many antiemetics are substrates for serotonin receptors and transporters—which are known to have clinically relevant genetic polymorphisms. Previous studies have shown that variations in the HTR3B gene predict the efficacy of antiemetics in cancer patients [14] and the occurrence of selective serotonin reuptake inhibitor (SSRI)-induced nausea [15]. However, there have been limited studies which have investigated the effects of the serotonin receptor gene polymorphisms on the efficacy of antiemetics used in the treatment of NVP [16].

The objective of this study was to evaluate the role of serotonin receptor subunit HTR3A and HTR3B genotype with NVP severity and antiemetic efficacy.

The serotonin receptor genotypes were obtained for all included subjects. For rs1176744, subjects were categorized as having the genotype AA (wild-type, 6, 21.4%), AC (heterozygous, n = 16, 57.1%), or CC (homozygous variant, n = 6, 21.4%). For rs3782025, subjects were homozygous wild type (AA, n = 3, 11.5%) or heterozygous (AG, n = 14, 23, 88.5%). For rs3831455, 22 subjects (84.6%) were homozygous wild type and 4 (15.4%) carried the deletion allele. For rs1062613 SNP, 10 (41.7%) were homozygous wild type (CC), 9 (37.5%) were heterozygous (CT), and 5 (20.8%) were homozygous variant alleles (TT).

Figures 1 and 2 display the polymorphisms’ impact on PUQE scores (Figure 1) and QOL scores (Figure 2). For the rs3782025 polymorphism, the initial and final PUQE scores were significantly better for carriers of a variant allele (p = 0.02 for both comparisons). The final PUQE score was significantly worse for carriers of the rs1062613 variant (p = 0.01). There were trends toward significant differences for rs3782025 final QOL scores (p = 0.08), rs3831455 final PUQE scores (p = 0.07), and rs1062613 initial QOL scores (p = 0.05).

Although the complete pathogenesis of NVP is unknown, serotonin is a key mediator in inducing nausea and vomiting through a mechanism in which 5-HT acts on its receptor in the gastrointestinal tract, 5-HT₃. This receptor relays information to the central nervous system through depolarization of vagal afferent nerves and the vagus nerve. This signal is thought to initiate the vomiting reflex [19]. Excess serotonin is released by small intestinal enterochromaffin cells in response to drugs such as chemotherapeutic agents [20]. 5-HT₃ antagonists are a class of medications that act at the 5-HT₃ receptor, preventing serotonin from binding and thereby preventing vagal afferent stimulation and induction of the vomiting reflex. Many 5-HT₃ antagonists are antiemetics and are used in the prevention and treatment of nausea and vomiting.

The ligand-gated ion channel serotonin receptor type 3 (5-HT₃) is found in the autonomic, enteric, and sensory nervous systems [21]. Although five subunits of 5-HT₃ have been identified in humans (5-HT_3A-3E), 5-HT_3A is the only subunit that is capable of forming homo-oligomeric receptor complexes that are functional. All other subunits are able to form only heteropentameric receptor complexes with 5-HT_3A [22]. Therefore, 5-HT_3A is a consistently important subunit and 5-HT_3B has previously been shown to be a major factor in HTR3 function [23]. Additionally, 5-HT_3A and 5-HT_3B are best characterized and both functional and clinically relevant genetic polymorphisms exist in the genes of these subunits so we focused on these relevant polymorphism. To our knowledge, only one other study has examined the relationship between NVP and 5-HT₃ receptor genotype. This study focused largely on a different set of genetic polymorphisms and found a weak association between NVP and two HTR3C variants but did not find any association when examining two HTR3A variants, including one that was also tested in this study [16].

A variant form of the 5-HT_3B receptor (rs1176744 variants) has been shown to exhibit a substantially increased maximum response (7-fold increased mean open time) to serotonin compared to the wild type form [24]. While Tremblay et al. [14] found that this variant was not predictive of nausea and vomiting incidence in cancer patients, Sugai et al. [15] found this variant to be predictive of higher nausea incidence in a paroxetine-induced nausea setting. Our results indicate that this variant appears to contribute to decreased response to antiemetic therapy for NVP as homozygous variant subjects were prescribed more antiemetic medications to control symptoms than those with at least one wild type allele. These results suggest that women with this particular homozygous mutation in HTR3B may have more recalcitrant NVP. As such, homozygous variant patients may not be receiving high enough dosing of antiemetic therapy to inhibit serotonin binding at the 5-HT₃ receptor target due to the increased affinity of the receptor for serotonin in these subjects.

The 5’-UTR HTR3A variant, rs1062613, has also been shown to have functional significance. This nonsense mutation occurs in the upstream open reading frame of HTR3A. Luciferase reporter constructs showed that presence of the variant allele results in significant increases in expression [25]. Despite this, Kaiser et al. [26], Sugai at al. [15], and Goeke et al. [16] found that HTR3A polymorphisms were not good predictors of antiemetic treatment in cancer patients, SSRI-induced nausea, or NVP, respectively. However, our data showed that despite statistically similar initial PUQE scores, carriers of a variant rs1062613 allele had significantly worse final PUQE scores. We also found that there is a trend toward significant differences in initial QOL scores with variant carriers having worse initial QOL compared to other patients. This discrepancy could be caused by medication and setting differences amongst studies. Our results suggest that patients who carry at least one variant allele may not be receiving a sufficient dose of antiemetics to antagonize the increased receptors that may be expressed.

Previous work has found that the 3 base pair insertion/deletion polymorphism, rs3831455, in the promoter region of the HTR3B gene was found to significantly associate with incidence of nausea and vomiting caused by chemotherapy [14]. This HTR3B promoter deletion, or variant, was also found to change the mRNA structure and was suggested to have an effect on translation efficiency [27] and nuclear protein binding [28]. Our results indicated a trend for this insertion/deletion polymorphism with worse final PUQE scores in patients carrying a variant allele. This could be due to increased translation and expression in those carrying the variant which could lead to insufficient antiemetic dosing to control symptoms.

Although the functional consequences of the HTR3B variant rs3782025 are unclear this polymorphism has been associated with vulnerability to alcohol use disorders (AUD), another condition that has been suggested to be affected by 5-HT dysfunction [29]. We found this polymorphism to be significantly associated with both initial and final PUQE scores, with variant carriers having better scores. Carriers of this variant also trended towards better final QOL scores. These data suggest that carrying a variant rs3782025 allele may be protective against NVP, as baseline severity and outcome are improved in these women.

For patients with significant emesis this may be an important step in improving treatment strategies. If these findings are confirmed in larger studies, it is possible that women with certain serotonin receptor genotypes may be better suited for one antiemetic drug versus another or may require a higher starting dose of drug. Alternate therapy may be prescribed based on genotype or patients may start therapy on multiple drugs rather than monotherapy. This could provide significant time and money savings as well as provide much needed symptom relief and may improve safety for both the mother and the fetus.

Our study was limited by small sample size. This pilot study was not powered to find differences in secondary outcomes, however the trends seen here will inform samples size for follow-up studies. There was also a potential for recall bias, however validated measures were used, increasing the validity of the results. Another limitation is that the study did not mandate a specific NVP treatment algorithm and did not direct physician’s prescribing patterns in any way. As such, while the number of antiemetic medications given to each subject was used as a surrogate for severity of NVP, it is confounded by variability in provider practice. Because standard practice is to start with one drug and add more to obtain symptom control we believe that this is a reasonable surrogate of symptom severity. We also did not measure treatment adherence. These results need to be validated in a larger cohort. Additionally, ability to perform subgroup analyses of treatment groups by increasing sample size may improve the ability to build a pharmacogenetic treatment model, which could be used to guide clinical practice. A prospective treatment trial based on genotype and examining an expanded serotonin receptor pathway genetic analysis is currently being planned.

In conclusion, this pilot study of pharmacogenetic predictors of NVP therapy response demonstrates that genotype for serotonin receptor subunits 5-HT_3A and 5-HT_3B may play a role in disease severity and response. There is a need for more pharmacogenetic studies in pregnancy, particularly for conditions necessitating drug therapy such as NVP. As pharmacogenetic studies emerge in obstetric conditions such as NVP, the potential for individualized pharmacotherapy at the outset of treatment may be realized.