Lower Frequency of PNPLA3 Polymorphism in West African Populations and Divergent Associations With Hepatocellular Carcinoma

Abstract

Background and aims: The germline variant patatin-like phospholipase domain-containing protein 3 (PNPLA3)-rs738409 (I148M) is associated with a higher risk of hepatocellular carcinoma (HCC) in European and East Asian populations, but its association with HCC in Africans has not been investigated. We examined the association between PNPLA3-rs738409 and HCC risk in Ghanaian and Nigerian populations, compared to a U.S. population.

Methods: We enrolled 363 patients with HCC and 2807 cancer-free control patients from Ghana, Nigeria, and the U.S. Germline genetic sequencing was performed, and data on PNPLA3-rs738409 were extracted for analysis. Logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals separately for each population.

Results: There was a significantly lower frequency of the PNPLA3-rs738409 risk allele [G] in the Ghanaian and Nigerian populations compared to the U.S. population. PNPLA3-rs738409 was associated with increased HCC risk in the U.S. population (OR = 3.40, P < .001), but not in the combined Ghanaian and Nigerian population (OR = 1.32, P = .30).

Conclusion: This study highlights the potential ancestral and global geographic differences in the genetic risk factors of HCC. It shows the population-specific variability in the association between PNPLA3-rs738409 and HCC risk, showing no significant association in Ghanaian and Nigerian populations, in contrast to the U.S. population. These findings emphasize the need for the inclusion of diverse populations, especially minority groups, in genetic studies for accurate and equitable assessment of heritable cancer risk.

Keywords: Disparity; Genetic Ancestry; Genetic Association Study; Polygenic Risk Score.

Figure 1

Principal component analysis (PCA) plot of United States, Ghanaian, and Nigerian samples and the 1000 Genomes Project dataset. The x-axis represents the first principal component (PC1), and the y-axis represents the second principal component (PC2). Each symbol corresponds to an individual from the study or from the 1000 Genomes Project. (A) Samples collected in this study marked by HCC cases (blue x) and controls (brown +). (B) Samples collected in this study were marked according to country of origin: Ghana (purple ◊), Nigeria (green ▽), and the United States (brown ∗). AFR, Africans; AMR, admixed Americans; EAS, East Asians; EUR, Europeans; SAS, South Asians.

Figure 2

Forest plot showing multivariable associations of PNPLA3 risk allele and other demographic and clinical characteristics with risk of HCC for the United States population. Associations were depicted using odds ratios (ORs) and 95% confidence intervals. Age OR assessed increase in risk of HCC per 10-year increase in age. Principal component ORs assess increase in risk of HCC per 1-unit increase.

Figure 3

Allele frequency distribution of PNPLA3 rs738409C >G in 3 populations. (A) Distribution of the genotypes of PNPLA3 in HCC cases and controls by population. (B) Odds ratios (ORs) and 95% confidence intervals (CIs) indicating the association of PNPLA3-rs738409 with HCC risk across populations from Nigeria, Ghana, West Africa (combined Ghana and Nigeria), and the United States. Logistic regression was adjusted for sex, hepatitis B, hepatitis C, age at diagnosis, and the first 2 ancestry-based principal components.

Figure 4

Forest plot showing multivariable associations of PNPLA3 risk allele and other demographic and clinical characteristics with risk of HCC for the Ghanaian and Nigerian populations. Associations were depicted using odds ratios (ORs) and 95% confidence intervals. Age OR assessed increase in risk of HCC per 10-year increase in age. Principal component ORs assess the increase in risk of HCC per 1-unit increase.

Figure 5

Forest plot showing multivariable associations of PNPLA3 risk allele and other demographic and clinical characteristics with risk of HCC for the Nigerian population. Associations were depicted using odds ratios (ORs) and 95% confidence intervals. Age OR assessed increase in risk of HCC per 10-year increase in age. Principal component ORs assess the increase in risk of HCC per 1-unit increase.

Figure 6

Forest plot showing multivariable associations of PNPLA3 risk allele and other demographic and clinical characteristics with risk of HCC for the Ghanaian population. Associations were depicted using odds ratios (ORs) and 95% confidence intervals. Age OR assessed increase in risk of HCC per 10-year increase in age. Principal component ORs assess the increase in risk of HCC per 1-unit increase.