New Insight Into Hyperemesis Gravidarum and a Potential Role for GDF15

Abstract

Walker and Thompson provide a commentary on the study conducted by Fejzo et al. that identified a potential relationship between GDF15 and hyperemesis gravidarum.

Nausea and vomiting in pregnancy (NVP), otherwise known as “morning sickness,” occurs at a high frequency, but in 0.3% to 2% of pregnancies, extreme nausea and vomiting develop, resulting in hyperemesis gravidarum (HG). Given the severity of HG symptoms, such as dehydration, ketonuria, nutrient deficiency, and >5% weight loss, HG is a leading cause of hospital admission during pregnancy. Interestingly, an absence of NVP is associated with a higher risk of miscarriage, whereas having HG is associated with poor fetal outcomes ranging from preterm birth and neurodevelopmental delay. Unfortunately, there is no effective treatment regimen for HG because the cause of this disorder remains unknown. However, there is evidence that women have a significantly elevated risk of HG if their mother or sister also had HG, suggesting that there is a heritable component of HG that, if identified, may lead to the development of an effective treatment option for women with HG.

In a recent study published in Nature Communications, Fejzo et al. (1) took a particularly innovative and opportunistic approach to determine whether there is a genetic component linked to increased risk for developing HG. Normally, conducting a genomewide association study alone would not necessarily be considered innovative. However, in this instance, Fejzo et al. (1) took advantage of the vast amount of genetic and patient history data collected from customers of the privately held personal genomics and biotechnology company 23andMe. With this large data set in hand, the authors performed two separate genomewide association scans to uncover unique genetic polymorphisms in groups of unrelated women that were associated with a binary or an ordinal HG phenotype. Fejzo et al. (1) uncovered two genomewide signals in both HG phenotypes that were also independently replicated in two additional cohorts of women with a clinical diagnosis of HG. Interestingly, the locus that had the most significant association with HG contained the genes growth differentiation factor 15 (GDF15) and LRRC25, with single nucleotide polymorphisms (SNPs) found in linkage disequilibrium with GDF15. The second genomewide associated signal came from a locus near the protein coding gene IGFBP7, which has been shown to be involved in pregnancy. Although LRRC25 and IGFBP7 were identified to be highly associated with elevated risk for HG, we will focus the remainder of this commentary on GDF15 because the authors identified an SNP directly in the protein coding region of the GDF15 gene.

GDF15, also known as placental bone morphogenetic protein, macrophage inhibitory cytokine 1, and nonsteroidal anti-inflammatory drug activated gene, is a distinct member of the TGF-β family of extracellular signaling ligands that was first identified in the human placenta (2). Similar to most TGF-β family members, GDF15 is a disulfide-linked dimer that is initially synthesized as a precursor protein that requires proteolytic processing to separate its N-terminal prodomain from the C-terminal mature domain (3). However, unlike most TGF-β members, GDF15 does not require the prodomain for proper folding and also does not engage canonical TGF-β type II and type I receptors to elicit SMAD signaling (3). Other unique members of the TGF-β family such as glial-derived neurotrophic factor and artemin use a distantly related yet entirely different subclass of receptors known as glial-derived neurotrophic factor family receptor α1 and glial-derived neurotrophic factor family receptor α3, respectively. Therefore, a major breakthrough toward the understanding of GDF15-mediated signaling occurred when it was shown that GDF15 uniquely uses the orphan receptor glial-derived neurotrophic factor receptor α-like receptor (GRAFL) and requires the coreceptor RET-receptor tyrosine kinase to elicit downstream signaling (4–8). The identification of cognate receptors for GDF15 is a major step toward understanding the biological function of GDF15.

Over the past decade, GDF15 has gained recognition as a staunch circulating biomarker associated with or predictive of adverse clinical outcomes in a number of disease states such as heart failure, diabetes, kidney disease, and cancer cachexia (3). Whether elevated GDF15 or polymorphisms in GDF15 are causal or augment symptoms of these diseases remains unknown. Transgenic overexpression studies in mice showed a direct link between circulating GDF15 levels and improved metabolic phenotypes in diabetic mouse models and invoked anorexialike phenotypes in healthy mice (3). Thus, GDF15 has emerged as a potentially powerful antiobesity biologic (4–8). A number of independent groups have convincingly demonstrated that exogenous administration of recombinant GDF15 or Fc-GDF15 (Fc fusion) curbs appetite, resulting in weight loss in nonhuman primates, and that this effect is mediated through a direct interaction between GDF15 and its cognate receptor glial-derived neurotrophic factor family receptor α-like (GFRAL) (4–8). Interestingly, GFRAL expression appears to be exclusive to the area postrema of the brainstem in mice, rats, and monkey and in human medulla (4–8), an area that is unprotected by the blood-brain barrier and strongly tied to appetite, nausea, and vomiting. Indeed, these findings provide a potential explanation for why elevated GDF15 may be involved in promotion of NVP and HG. The limited expression profile of GRAFL makes it difficult to predict how GDF15 may function in other aspects of biology, such as placentation or maintenance of pregnancy, where GDF15 is thought to suppress production of proinflammatory cytokines.

The NVP- and HG-associated SNP in GDF15 identified by Fejzo et al. (1) is located in the GDF15 protein coding region and results in a histidine to aspartate missense mutation at amino acid position 6 (H6D) in the mature GDF15 growth factor. Consistent with previous studies (9), Fejzo et al. (1) provide data to suggest that the H6D polymorphism results in elevated levels of GDF15, although it is unclear how or why this polymorphism would cause this change. One possibility may be that this polymorphism alters the biosynthesis (e.g., protein folding, dimerization) or processing of the prodomain to liberate the mature growth domain. It is known that perturbations in the processing of the GDF15 precursor or removing the prodomain altogether can dramatically alter the serum levels of GDF15 (10). Another possibility is that the H6D mutation alters the native interaction between the prodomain and the mature domain, thereby affecting function. Similarly, it is well known that some ligands in the TGF-β family intimately interact with their respective prodomains as a way to control ligand function and activity. Based on analysis of the GDF15:GRAFL cocrystal structure (7), it does not appear that H6D would affect the interaction between GDF15 and GRAFL, but the possibility exists that it may alter the formation of the ternary receptor signaling complex between GDF15, GRAFL, and RET receptor kinase or an interaction with an unknown regulatory binding partner. Completion of experiments that directly address how this polymorphism affects the biological function of GDF15 will be very helpful toward understanding the impact of GDF15 in these diseases.

The study by Fejzo et al. (1) represents an exciting scientific step toward the use of the ever-growing number of genetic databases generated by privately held companies and their customers to identify genetic associations with disease. Here, the authors identified a potentially novel genetic association between a polymorphism in GDF15 and expectant mothers who had NVP or HG. As the authors also stated, future research, perhaps through gain of function and loss of function studies, should be directed toward determining whether this polymorphism causes the disease or whether it increases the severity of NVP and HG. Certainly, this study provides the impetus to determine whether the serum levels of GDF15 are altered in patients with HG; however, one should ultimately correlate these results with signaling activity because it may be a better indicator of the functional impact of GDF15 misregulation. Provided a causal relationship is established and given that blockade of GDF15 has been shown to restore appetite and promote weight gain, blocking GDF15 therapeutically might be a powerful treatment option for expectant mothers with HG.

Glossary

Abbreviations:

GDF15

growth differentiation factor 15

GFRAL

glial-derived neurotrophic factor family receptor α-like

GRAFL

glial-derived neurotrophic factor receptor α-like receptor

H6D

histidine to aspartate missense mutation at amino acid position 6

HG

hyperemesis gravidarum

NVP

nausea and vomiting in pregnancy

SNP

single nucleotide polymorphism