Upon screening 46 European American BC cases from the AHCC for rare non-synonymous variants in HCAR1, HCAR2, and HCAR3, a total of four variants were identified in four different BC cases. These variants were exclusively identified in HCAR1 and HCAR3, which is notable considering their suggested oncogenic role and requirement for BC proliferation and survival, compared to the demonstrated tumor suppressor properties of HCAR2 (4–6).
The HCAR1 protein is known as the lactate receptor, and, upon HCAR1 binding, lactate inhibits lipolysis (1, 2, 31). Liu et al. identified ligand-binding pockets in HCAR1, specifically demonstrating that particular missense variants in transmembrane domains three (p.Arg99Ala), six (p.Tyr233Ala and p.Arg240Ala), and seven (p.Thr267Ala) diminished the response of HCAR1 to L-lactate (Fig. 1A) (31). It is important to note the proximity of p.Arg240Ala to one of the missense variants detected in this study, p.Leu241Phe, within the sixth transmembrane domain (Fig. 1A). The vital function of Arg240 in lactate binding, as well as the identification of other critical residues directly located at a transmembrane and extracellular (Arg71) (32) or cytoplasmic (Ala110) (33) loop junction (Fig. 1A), hints towards the importance of the highly conserved Leu241 and the potential damaging effects of an amino acid substitution at that location. Coincidentally, the other HCAR1 missense variant detected in a BC affected-individual in this study, p.Pro20Ala, is also near a transmembrane/extracellular domain junction and specifically located in the extracellular N-terminal domain (Fig. 1A). Another key, extracellular residue with similar junction proximity, Cys88, abolishes receptor activity when substituted with an Ala or Ser. Furthermore, two other critical cysteine residues, Cys6 and Cys7, are located in the N-terminus (Fig. 1A) (32); thus, the highly conserved Pro20 may also be vital for protein function. The need to further study p.Pro20Ala is compounded by the fact that the N-terminus, despite being highly variable between different GPCRs and not extensively studied, is important for ligand binding, dimerization, signaling, and surface expression (32, 34–36); additionally, p.Pro20Ala was determined to be genetically associated with BC.
Considering that several of the prediction software suggested both HCAR1 p.Pro20Ala and p.Leu241Phe were pathogenic, determining their functional involvement in BC is critical. To date, all functionally assessed HCAR1 variants have been deemed loss-of-function (highlighted (in color) in Fig. 1A) (31–33). However, with HCAR1 typically being regarded as critical for BC proliferation and survival by controlling lipid/fatty acid metabolism(5, 6), a loss-of-function mutation would presumably result in BC cell death. Interestingly, knocking down HCAR1 has different effects on different BC molecular subtypes (5, 6). For instance, knocking down HCAR1 in a HER2-enriched BC cell line, HCC1954 (ER-, PR-, HER2+), and triple negative BC cell line, HCC38 (ER-, PR-, HER2-), resulted in a significant decrease of cell viability within 48 hours of transfection. However, there was no significant change in viability regarding the luminal B cell line, BT-474 (ER+, PR+, HER2+), similar to the non-tumorigenic epithelial breast cell line, MCF12A (5). Furthermore, when HCAR1 was knocked down in the luminal A BC cell lines, MCF-7 and T47D, cell viability decreased (6). In our study, HCAR1 p.Pro20Ala and p.Leu241Phe were detected in individuals diagnosed with luminal subtypes. Specifically, the individual with p.Pro20Ala had luminal A BC (ER + and HER2-), which according to Lee et al. requires HCAR1 to proliferate (6). The individual with p.Leu241Phe was ER + and PR + but HER2 status was unknown; thus, the subtype could not be confirmed as luminal A or B. If it was luminal B, HCAR1 expression would not be not required for survival (5), whereas it would be required for luminal A (6). In additional to knock down studies, Lee et al. investigated HCAR1 expression levels in different BC molecular subtypes and noted that ER + BC cell lines expressed HCAR1 at a higher level (6); this ER + BC-association has also been reported for another GPCR, GPR30 (37). Thus, it is important to determine if HCAR1 p.Pro20Ala and p.Leu241Phe are specifically associated with ER + BC, as well as if they are loss- or gain-of-function mutations.
HCAR3, which is only found in higher primates, is the receptor for 3-hydroxylated β-oxidation intermediates, particularly 3-hydroxy-octanoate (1, 2). When activated, HCAR3 inhibits free fatty acid release from cells, providing a negative feedback mechanism to offset stimuli that promote lipolysis and fatty acid oxidation. Knocking down HCAR3 in BC cell lines BT-474, HCC1954, and HCC38 induced cell death, suggesting that HCAR3 has oncogenic properties. Introducing fatty acid oxidation inhibitors mitigated the knock down effects, confirming that uncontrolled up-regulation of fatty acid oxidation promotes BC cell death; thus, HCAR3 plays an vital role in controlling fatty acid metabolism in BC cells (5). Accordingly, one can presume that the two BC-associated HCAR3 variants identified in this study, p.Arg187Gln and p.Gln373Lysfs*82, have gain-of-function effects. However, it is important to note that HCAR3 knock down effects have not been assessed in luminal A BC cell lines, which is the molecular subtype reported in the two BC-affected individuals with each HCAR3 variant. Numerous HCAR3 genetic variants have been reported in publically available databases (13, 16), as well as through polymerase chain reaction (PCR)-based techniques (38), and while their pathogenic effects have not been functionally assessed, our study suggests that rare non-synonymous variants in HCAR3 may enhance the receptor’s ability to control fatty acid metabolism. Nonetheless, HCAR3 p.Arg187Gln, located in the third extracellular topological domain (Fig. 1), was unanimously predicted to be benign. Even though prediction software have been shown to misclassify known pathogenic variants (39), it is important to note that HCAR2, which shares 95% sequence identity with HCAR3, has a glutamine at that overlapping position. Though we have confirmed HCAR3 p.Arg187Gln through nested PCR (Supplementary information: Figure S4), it is unknown if this change would affect the function of HCAR3. That being said, with such slight differences between the two proteins, perhaps each alteration is key to protein function. On another note, this variant was detected in an early onset BC case also determined to harbor a clinically significant frameshift mutation in NBN (12, 30). The interaction of these two variants and their combined ability to promote BC is unknown, but, intriguingly, expression of both HCAR3 and NBN have been reported to be dysregulated in oocytes of older women, when investigating why aneuploidy pregnancies occur in women of older ages. Overall, this observation suggests that these genes may play a role in proper chromosome segregation and maintaining genomic integrity, which is a phenomenon also disrupted in cancer (40, 41).
The HCAR3 frameshift mutation, p.Gln373Lysfs*82, significantly extends the C-terminal, cytoplasmic tail of the mutant HCAR3 protein and changes the secondary and tertiary protein structure (Fig. 2). Again, based on the suggested oncogenic role of HCAR3 in BC, HCAR3 p.Gln373Lysfs*82 may potentially result in a gain-of-function. Interestingly, distinct mutation profiles, corresponding to clusters of nonsense and frameshift mutations in the C-termini of GPCRs, GPR34, CCR6, and CCR4, have been reported in mucosa-associated lymphoid tissue (MALT) lymphoma and adult T cell leukemia/lymphoma (ATLL) as gain-of-function mutations (42–44). Even though the nonsense and frameshift mutations reported in GPR34, CCR6, and CCR4 truncate the encoded proteins, PSORT predicted that HCAR3 p.Gln373Lysfs*82 abolishes a prenylation motif in a manner similar to how the GPR34 mutations eliminate a key phosphorylation motif and ultimately dysregulate the receptor’s desensitization process (42, 43). Additionally, the mutant HCAR3 protein gains an ER Membrane Retention Signal, potentially affecting internalization patterns, which is also disrupted with CCR4 gain-of-function mutations (42, 44). Contrarily, the addition of a predicted peroxisomal targeting signal to the mutant HCAR3 hints toward protein degradation, a loss-of-function mechanism, and, on a similar note, read-through mutations that result in mutant proteins with C-terminal extensions in PNPO and HSD3B2 cause hereditary disorders through protein degradation (45). Nonetheless, GATA3 frameshift mutations that extend the C-terminus are the most common somatic mutation identified in The Cancer Genome Atlas (TCGA) BC patients and display gain-of-function activity (46). Loss-of-function GATA3 mutations were also identified, demonstrating that both loss- and gain-of-function mutations can be identified in the same gene and associated with BC. Similarly, TP53, a clinically valid BC susceptibility gene, has both tumor suppressor and oncogenic properties (47–49). Thus, the exact functional consequences of HCAR3 p.Gln373Lysfs*82 may be complex but are important to elucidate, especially considering the extensive homology between HCAR3 and HCAR2, and that HCAR2 has domains in the cytoplasmic, C-terminus vital for receptor export, internalization, constitutive activity, and desensitization (50).