The nutritional content of food in developed nations has changed dramatically over the past 75 years (1), mirroring the rise in non-communicable diseases, which account for approximately 75% of deaths in developed countries (2–4). Cardiometabolic diseases (CMD) remain the leading cause of death globally, with disparities particularly evident among various racial/ethnic minority groups in the United States (USA). Latino populations, especially Mexican Americans (MxAm), have higher rates of obesity, diabetes, hyperlipidemia, and cardiovascular diseases compared Non-Hispanic White (5–7) populations.

It is now estimated that over 70% of daily energy consumed in a modern Western diet (MWD) is from sources that were absent from our ancestors’ hunter-gatherer diet (8). A notable example of this shift is the change in dietary polyunsaturated fatty acids (PUFA) consumption, which can be traced back to recommendations from health agencies to reduce saturated fat intake in favor of PUFAs, predominantly derived from vegetable oils such as soybean, corn, and canola (9, 10). These recommendations led to a ~ 4-fold increase in the dietary consumption of the omega-6 (n-6) PUFA linoleic acid (LA) without significantly altering the consumption of dietary omega-3 (n-3) PUFAs, thereby altering the ratio of dietary n-6 to n-3 PUFAs to greater than 10:1 (1).

The dietary n-6 and n-3, 18-carbon PUFAs LA and alpha-linolenic acid [ALA, omega-3 (n-3)] are converted by the liver and other tissues to long-chain (>20 carbon), highly unsaturated fatty acids (HUFAs) utilizing a sequential set of desaturases and elongases (11, 12). Figure 1 highlights genes and enzymes key to HUFA biosynthesis and metabolism. In two parallel and competing pathways, 18-carbon n-6 or n-3 PUFAs are converted into HUFAs. In the n-6 arm of the pathway (left), ARA is synthesized from LA, utilizing 2 desaturation and 1 elongation enzymatic steps (13, 14). In the n-3 arm of the pathway (right), EPA and DHA are synthesized from dietary ALA. This second desaturation step (Δ-5 desaturase or FADS1) is rate-limiting, and genetic variation throughout the FADS gene cluster impacts the pathway efficiency via altered FADS1 transcription (15–17).

Once ARA is formed and placed into membrane phospholipids, it can be mobilized (by phospholipases) and metabolized [by cyclooxygenase (COX), lipoxygenase (LOX) and P450 pathways] into oxylipins including eicosanoids such as prostaglandins, thromboxanes and leukotrienes (Figure 1) (18–20). These then act locally to promote inflammation and thrombosis in CVD and other diseases (20). By contrast, DGLA, the n-6 HUFA proximal to FADS1, is metabolized by COX 1/2 to 1-series PGs, particularly PGE₁, and by 15-LOX to 15-HETrE (21). These two metabolites of DGLA suppress inflammation, promote vasodilation, lower blood pressure, inhibit smooth muscle cell proliferation, and exert anti-neoplastic activities (22–25). Similarly, EPA and DHA are metabolized to anti-inflammatory, anti-thrombotic and “pro-resolving” oxylipins including resolvins, protectins, and maresins (19, 26, 27). With few exceptions, DGLA, EPA, DHA and their oxylipin products exert effects that differ from or oppose ARA-derived oxylipins (Figure 1) regarding inflammation, hyperlipidaemia and disease modification. Consequently, DGLA conversion to ARA and the balance of ARA to DGLA, EPA and DHA have the potential to influence the prevalence and severity of numerous human diseases.

Numerous studies have established a connection between genetic variations in the FADS cluster and the efficiency of the HUFA biosynthetic pathway, and thus HUFA levels (28, 29). Early research indicated that high levels of ARA, as well as the ARA/LA ratio, are independent predictors for coronary artery disease (30). Furthermore, the role of FADS variants in complex lipid and inflammatory phenotypes has been extensively documented (31). Notably, the FADS gene region is recognized as the locus associated with the largest multimorbidity cluster in the entire human genome (32).

Moreover, research from various laboratories has revealed significant ancestry-based discrepancies in the frequency of FADS variants affecting HUFAs and their corresponding oxylipin metabolites, such as eicosanoids (33–36). Evolutionary pressures have shaped the allele frequencies of FADS variants across different populations, reflecting adaptations to diverse diets and environments. For example, the “derived” FADS haplotype experienced ancient positive selection in Africa, leading to its high frequency in present-day African populations (37, 38). For example, ∼80% of African Americans (AfAm), as compared to ∼45% of European Americans (EuAm), carry two copies of derived alleles associated with increased FADS1 activity, resulting in a marked increase in HUFAs, especially ARA and ARA-derived metabolites (34, 35). The introduction of agriculture is also thought to have influenced FADS allele frequencies. However, strong selection on the derived allele in European populations occurred much later than the Neolithic transition, possibly as late as the Bronze Age (39). The findings of Fumagalli et al. (40) on genetic adaptations in the Greenlandic Inuit population provided an important rationale as to why a high proportion of Indigenous American ancestry populations carry two copies of “ancestral” alleles associated with reduced FADS1 activity, resulting in a reduction in HUFAs, especially EPA production (36, 37, 41, 42). Their research demonstrated that the FADS gene cluster shows the strongest signal of genetic adaptation, reflecting evolutionary responses to a traditional marine-based diet and extreme Arctic conditions (40). Specifically, the ancestral FADS alleles were positively selected in the Inuit and remain prevalent among populations with high proportions of Indigenous American ancestry such as MxAm.

An unresolved question is if the variation in FADS in Hispanic populations that induces very low levels of n-3 HUFAs and imbalances in n-6 to n-3 HUFA ratios influences the risk of cardiometabolic diseases often observed in these populations. To address this important question, we investigated the relationship between variation the in genes (FADS1,2,3 and ELOVL2,5) that code for HUFA biosynthetic enzymes on a robust set of cardiometabolic markers and traits in adult Latino participants from the Arizona Insulin Resistance registry (AIR) (43).

Diabetes and hyperlipidemia are significant health concerns in Mexican and MxAm populations, who have higher prevalence rates compared to Non-Hispanic White (NHW) populations (60, 61). For example, MxAm are 1.7 times more likely to be diagnosed with diabetes than NHW, and over 50% of Hispanic adults (including MxAm) are predicted to develop type 2 diabetes in their lifetime (62). MxAm with diabetes also tend to have worse glycemic control and higher rates of diabetes-related complications, such as kidney disease. Additionally, MxAm are more likely to have higher triglyceride levels and lower HDL cholesterol levels compared to NHW (61, 62). Non-alcoholic fatty liver disease (NAFLD) affects about 25% of the general population, but the prevalence is higher in Hispanic populations, with MxAm having the highest rate at 42.8% (63).

Socioeconomic factors, including lower income, decreased access to healthcare, and language barriers, clearly contribute to the higher prevalence of these conditions in MxAm populations (61, 64). Genetic variants also play a role. For instance, the SLC16A11 risk variants, more common in people of Mexican ancestry, are associated with increased diabetes risk (64–66). Variations in the PNPLA3 gene, which predisposes people to store extra fat in liver cells, are found more often in those of Hispanic origin and are also associated with greater severity of NAFLD (67, 68).

In the current study of Latino adults of MxAm descent from the Southwest US, we observed high frequencies of alleles linked to the ancestral FADS haplotype, comparable to those observed in MxAm populations from the US West Coast. In contrast, these alleles are much less frequent in African and European ancestry populations and are also associated with a reduced capacity to synthesize n-6 and n-3 HUFAs, largely due to their impact on the rate-limiting FADS1 enzymatic step (15). This observation was confirmed in our study and mirrored findings in other Latino populations (36, 41, 42), where a 33 and 38% reduction in circulating levels of ARA and EPA, respectively, was seen when individuals with the TT genotype at rs174537 were compared to those with the GG genotype.

Within the TT genotype at rs174537, mean ARA levels (56 ng/mL) were 18-fold and 5-fold higher than mean EPA (3 ng/mL) and mean DHA (12 ng/mL) levels, respectively. Moreover, individual ARA/EPA ratios display a high level of variability (coefficient of variation ~109%) with a strong positive skew, leading to a mean ARA/EPA ratio of ~30 to 1 for the TT group. Although this ratio does not differ significantly between genotypes, the extremely low concentrations of EPA in TT individuals raise an important question as to whether there is an EPA deficiency with the potential to impact cardiometabolic disease. This question is especially relevant given findings from recent notable n-3 HUFA supplementation randomized controlled trials such as ASCEND, VITAL, and REDUCE-IT (69–71). Notably, only REDUCE-IT provided higher doses of EPA to patients at high risk of CVD and demonstrated a reduction in composite CVD morbidity and mortality (including CVD mortality, non-fatal MI, non-fatal stroke, cardiovascular revascularization or unstable angina) in the treatment group (69).

While the current study corroborated the frequencies of the ancestral haplotype and its impact on n-6 and n-3 HUFA levels as has been observed in other African, European and Indigenous American ancestry cohorts, the major objective here was to determine whether FADS/ELOVL variants impact important cardiometabolic traits, including triglycerides, HDL, VLDL, fat mass, and insulin resistance markers. Previously, our laboratory examined the impact of the ancestral FADS haplotype on clinical phenotypes in six Hispanic subgroups in the Multi-Ethic Study of Atherosclerosis (MESA) (42). Results from this study suggested that FADS variation was associated with circulating levels of triglycerides, HDL-C, E-selectin, s-ICAM, waist-hip ratio, height, and weight. However, only associations for triglycerides and height remained significant when the regression model included the principal components of ancestry. That study underscored the need to investigate the clinical impact of FADS variation in a more homogeneous cohort.

The key findings of the current study demonstrate that the ancestral haplotype is strongly associated with several cardiometabolic endpoints. For example, at rs174455, the homozygous ancestral FADS genotype (compared to the homozygous derived genotype) was linked to a ~40% increase in fasting insulin and HOMA-IR, a ~30% increase in triglycerides and VLDL, and modest decreases in HDL and female anthropometric measures. Interestingly, ELOVL2/5 variants were also associated with increased glucose, insulin, and HOMA-IR, indicating variation in the HUFA biosynthetic pathway outside of the FADS cluster may also affect cardiometabolic endpoints. The underlying molecular mechanisms that link FADS, circulating HUFAs, and cardiometabolic endpoints are not yet clearly understood. However, n-3 HUFA supplementation has been shown to impact insulin-stimulated glucose disposal, potentially improving insulin sensitivity (72, 73). Additionally, the ratio of EPA to ARA has been shown to be a promising indicator of glycemic control (74).

Interestingly, alleles associated with the ancestral haplotype are often strongly linked to significantly higher triglyceride levels (36, 38, 75), whereas n-3 HUFA supplementation is well-documented to reduce circulating triglyceride levels (42, 76, 77). Mechanistically, n-3 HUFAs have been shown to: (1) inhibit hepatic synthesis and secretion of VLDL particles, the primary carriers of triglycerides in the bloodstream; (2) enhance the activity of lipoprotein lipase, an enzyme that hydrolyzes triglycerides in VLDL and chylomicrons, accelerating their clearance from circulation; (3) inhibit acyl-CoA:1,2-diacylglycerol acyltransferase, a key enzyme in triglyceride biosynthesis; and (4) increase peroxisomal β-oxidation, thereby promoting fatty acid metabolism (75, 76). These findings support our hypothesis that n-3 HUFA deficiency, particularly of EPA, that results from FADS gene-by-diet PUFA interactions in MxAm individuals of Indigenous American ancestry, may contribute to insulin resistance, elevated triglyceride levels, and an increased risk of diabetes and NAFLD. Notably, our study suggests that these metabolic conditions, when linked to n-3 HUFA deficiency, may be specifically mitigated through targeted n-3 HUFA supplementation.

A limitation of our study design is the lack of dietary data in our cohort, which prevents a complete assessment of gene-by-dietary PUFA interactions. However, we recently published data from a similar cohort of MxAm from the same region in Arizona that provide relevant context (77). In that cohort, approximately 35% of total caloric intake came from fat, with an average total fat intake of 59.6 g/day. PUFAs contributed 5–10% of energy, primarily from the n-6 PUFA, LA, with a mean intake of 12 ± 6.5 g/day (6.9 ± 2.2% of total energy). The mean n-6 to n-3 PUFA ratio in this population was 8.8:1, closely resembling the dietary patterns of other populations consuming a modern Western diet. Another limitation of our study is the lack of data on lifestyle factors and supplementation (e.g., PUFA supplements), which could influence the interpretation of our findings.

While the study’s focus on self-reported Latinos of Mexican origin from the AIR registry provided a targeted dataset, it may limit the generalizability of our findings to other Latino populations due to potential underrepresentation of broader genetic diversity. However, our analysis of the MESA cohort corroborates these findings, revealing strong negative associations between Indigenous American genetic ancestry and HUFA levels. Furthermore, the FADS rs174537 SNP accounted for a substantial portion of the ancestry-related effects on n-3 HUFAs, including EPA, and showed strong associations with various metabolic, inflammatory, and anthropometric traits, particularly circulating triglycerides.

Taken together, the current study suggests that gene-by-diet interactions associated with the quantities and ratios of dietary PUFAs, combined with the FADS ancestral haplotype, may play an important role in risk, complications and cardiometabolic disease disparities observed in MxAm populations. This study highlights the complex interplay between genetic variants, particularly in the FADS genes, and cardiometabolic traits in MxAm populations. The findings also underscore the importance of considering both genetic predispositions and environmental factors, such as diet, in understanding the health disparities that disproportionately affect these populations. Importantly, the gene-by-diet interactions involving PUFAs, HUFAs, and their downstream metabolic effects point to potential avenues for personalized nutrition and therapeutic strategies to mitigate cardiometabolic risk in this population.

Future research should further explore how dietary interventions, particularly those aimed at optimizing PUFA intake or supplementing with n-3 HUFAs, could modify the effects of FADS variants. This is particularly relevant when interactions between the FADS gene and dietary PUFAs result in an increased risk of n-3 HUFA deficiency, which would be expected in the 40–50% of MxAm homozygous for the ancestral haplotype. Given the high prevalence of hyperlipidemia in this population and the biological mechanisms linking n-3 HUFAs to elevated lipids, individuals in this group may especially benefit from n-3 HUFA supplementation. Supplementation with doses ranging from 2 to 4 grams per day has been shown to be most effective in lowering triglyceride levels, and the beneficial effects are typically observed within 6 weeks and can continue to improve over a period of up to 6 months in most studies (78). Additionally, expanding studies to other Latino populations and controlling for other socioeconomic and lifestyle factors will be critical in developing targeted prevention and treatment programs. Ultimately, addressing these gene-by-diet interactions could help reduce the burden of diabetes, cardiovascular disease, and related conditions in MxAm, Mexican and other Latino communities.