High penetrance of acute intermittent porphyria in a Spanish founder mutation population and CYP2D6 genotype as a susceptibility factor

Acute intermittent porphyria (AIP, MIM # 176000) is an autosomal dominant disorder caused by a deficiency of hydroxymethylbilane synthase (HMBS; EC 2.5.1.61), which catalyzes the third step of haem biosynthesis. AIP is the most common acute porphyria, with a prevalence of 5.9 cases/million inhabitants in Europe as a whole and 6.3 cases/million inhabitants in Spain [1].

The main clinical manifestations are episodic acute neurovisceral attacks characterized by abdominal pain, vomiting, tachycardia, hypertension and dark urine.

AIP is a low-penetrant disease, and generally 10–20% of AIP carriers develop acute attacks [2‐4], known as manifest AIP (MAIP), as opposed to latent AIP carriers (LAIP) with no attacks. However, studies based on the frequency of pathogenic variants in HMBS gene in the general population suggest significantly lower penetrance (< 1%) [5, 6]. On the other hand, higher penetrance has been associated with specific mutations, such as R173W and W198X. [7].

Acute attacks are more frequent in women and rarely occur before puberty [8, 9]. Such acute attacks are often triggered by precipitating factors, which include several drugs, alcohol, steroid hormones, infection or fasting. Genetic background might be involved in variations in penetrance, although the underlying genes involved are still unknown [10].

In the liver, free haem exerts negative feedback regulation of ubiquitous 5-aminolevulinic acid synthase (ALAS1; EC 2.3.1.37), the rate-limiting enzyme in haem biosynthesis [11‐13]. Acute attacks occur when hepatic haem synthesis is overstimulated. This causes overproduction of the porphyrin precursors 5-aminolevulinic acid (ALA) and porphobilinogen (PBG), which have been associated with clinical manifestation of AIP, and increases in urinary excretion. After acute attack, ALA and PBG gradually decrease, although levels may remain high for time ranging from weeks to years [14, 15].

Drug-metabolizing cytochrome P450 enzymes (CYP) constitute the main hepatic haemoproteins [16]. CYP and ALAS1 synthesis is coordinated, while several drugs are transcriptional inducers of both genes [17, 18]. Since the turnover of CYP is a determining factor in the hepatic induction of haem synthesis, it might influence the penetrance of hepatic porphyrias.

Inter-individual differences in drug metabolism are common, in part due to inherited polymorphisms in CYP genes [19]. A high frequency of two polymorphisms in CYP1A2 and CYP1A1 genes has been reported in porphyria cutanea tarda, both polymorphisms being associated with increased enzymatic activity [20, 21]. In addition, non-functional CYP2D6*3 and *4 alleles seemed to be less frequent in AIP carriers compared to a control population [22].

We hypothesized that CYP genes might act as modifiers in AIP, and that specific alleles would constitute susceptibility factors for developing acute attacks. Taking advantage of our genetically highly homogeneous population (most AIP cases carry the founder mutation c.669_698del30 in the HMBS gene) [23], we aimed to identify AIP penetrance modifying CYP genes that act as risk factors for developing acute attack. The goal of this study was to determine AIP prevalence and penetrance in our region associated with the HMBS founder mutation and the frequency of common defective alleles that lead to a disturbance in main hepatic CYP enzymes [24], as well as to analyse their relationship with the occurrence of acute attacks in AIP. With this aim, CYP2C9*2, *3; CYP2C19*2; CYP2D *4, *5; CYP3A4*1B and CYP3A5*3 alleles were studied in a group of AIP carriers with a high prevalence of the aforementioned founder mutation.

Fifty AIP carriers aged between 16 and 77 years (44 years of mean age), 56% women, were analysed (Table 1). 78% carried the known founder pathogenic variant c.669_698del30 in HMBS gene. General penetrance was 52% (frequency of MAIP). Penetrance associated with the founder pathogenic variant was 48.7% (19/39 patients), which is similar to that associated with the pathogenic variant c.76C > T (50%), the second most frequent pathogenic variant in the studied cohort. Furthermore, based on population the census in our province (1,470,273 inhabitants) and the frequency of MAIP described in this study, the estimated symptomatic AIP prevalence was at least 17.7 cases/million inhabitants.

Allele frequencies are shown in Table 2 and genotype distributions are summarized in Table 3. All genotype frequencies met the Hardy-Weinberg equilibrium. The frequency of defective CYP2D6 alleles was 3.5 times higher in LAIP than in MAIP. MAIP was less frequent in defective allele carriers of CYP2D6 gene (CYP2D6*4 or *5 alleles), so the possibility that a patient had MAIP was reduced by 80% on average for each additional defective allele (*4 or *5) in CYP2D6 gene (OR 0.2; CI 95% 0.04–0.81; p value 0.037).

There was also slight evidence that MAIP was reduced by each additional defective allele (*2 or 3*) in CYP2C9 gene, although not to a statistically significant extent. When a combined CYP2C9 and CYP2D6 genotype was considered (data not shown), only 16.7% of individuals carrying at least one defective allele in both genes had MAIP compared with 52.2% of carriers of defective alleles in either one of the genes and with 61.9% of normal allele carriers. Thus, MAIP frequency was reduced by 65% in individuals carrying at least one defective allele in both genes (OR 0.35; CI 95% 0.12–0.89; p value 0.0368). Other different CYP combined genotypes were not considered since there was no evidence of individual effect in acute attacks.

The urine PBG-to-creatinine ratio (Table 4) tended to be lower in individuals carrying CYP2C9*2 or *3 and CYP2D6*4 or *5 alleles compared with respective normal allele carriers. However, this reduction was not attributed to the CYP genotype by itself but was associated with the history of acute attacks (p < 0.01). There were no differences in ALA excretion according to CYP genotype.

Several clinical studies have described a penetrance of around 10–20% in AIP [2]. Other studies based in the prevalence of pathogenic variants in HMBS gene in the general populations have estimated a penetrance of below 1% [5, 6]. However, there are high-penetrant mutations such as R173W (50%) and W198X (44%) [7], which are highly frequent in Nova Scotia (Canada) [25] and Sweden [26], respectively, due to a founder effect. Here, we describe the high clinical penetrance associated with the founder mutation c.669_698del30 in HMBS gene [27], which is frequent in southern Spain. This penetrance is similar to that of the W198X mutation, which is associated with a high prevalence of AIP in Sweden (23 cases/million inhabitants). In addition, we estimated the prevalence of AIP in the Region of Murcia to be about 3 times higher than that estimated in Spain as a whole (6.3 cases/million inhabitants) [1]. All the above suggest that, while low-penetrant HMBS mutations may be frequent in a healthy population, frequent high-penetrant mutation might be responsible for MAIP in populations with a high prevalence of AIP. High-penetrant AIP mutations may determine the genetic counselling offered to a specific population and carriers should thereby benefit from specific health care planning.

It is also of great interest to recognize how the genetic background may modulate the penetrance, making personalized genetic counselling and health care more feasible. To date, a high prevalence of specific CYP alleles in some types of porphyria has been described compared to that observed in a healthy population, suggesting that they might be susceptibility factors [20‐22]. The present study provides further evidence that CYP genes may constitute penetrance-modifying factors in AIP, since CYP2D6*4 and *5 were more frequent in LAIP than in MAIP, although their allelic frequencies were similar in whole AIP carriers and the general population [27]. The differential distribution of CYP2D6*4 and *5 alleles between MAIP and LAIP supports the hypothesis that CYP2D6 may be associated with acute clinical manifestation of AIP. Accordingly, MAIP was less frequent in defective CYP2D6*4 and *5 allele carriers.

Furthermore, for the first time, the frequency of CYP2C9*2, *3; CYP2C19*2; CYP3A4*1B and CYP3A5*3 alleles has been determined in an AIP population, in which it was found to be similar to that of the general population and between MAIP and LAIP [27‐29]. However, in our opinion, we cannot dismiss a possible relationship of these genes with AIP, which would require larger studies.

Our results suggest that CYP2D6*4 and *5 defective alleles play a protective role in the clinical onset of AIP, modulating its penetrance. AIP patients with defective CYP2D6 alleles may be less susceptible to porphyrogenic xenobiotic intermediate metabolites. Another possible explanation is that defective CYP2D6 alleles might consume less haem than normal alleles, so defective CYP allele carriers might be less prone to the overstimulation of haem synthesis. This last hypothesis would be supported if baseline levels of ALA and PBG were lower in defective CYP2D6 allele carriers. However, we were unable to demonstrate that lower urinary ALA and PBG levels in defective CYP2D6 allele carriers are directly associated with CYP genotype rather than acute attack history.

The strength of this study lies in the high genetic homogeneity in terms of a causal AIP mutation in HMBS gene, which eliminates biases due to differences in penetrance associated with various mutations. However, this study is not without limitations. The low allele frequencies of the CYP genetics variants and the limited number of patients studied make it necessary to confirm these findings in larger AIP populations. Further studies which include other CYP genes, as well as other candidate genes using next generation sequencing, might provide a more comprehensive overview of their possible modulating role in AIP.

AIP prevalence in the Spanish Region of Murcia is 17.7 cases/million inhabitants, almost 3 times higher than that estimated for the rest of Spain. The penetrance is 52%, similar to other founder mutation AIP populations. CYP2D6*4 and *5 alleles may be protective factors for acute attacks, and CYP2D6 may constitute a penetrance-modifying gene. Further studies are needed to confirm these findings, which would allow a further progress in AIP clinical risk profile assessment.

We highlight the relevance of knowing the prevalence and penetrance of AIP in a given population, especially in that ones with high prevalence of specific HMBS mutations, since published data vary among populations and mutations. These results will have a positive impact on AIP carriers genetic counselling and will allow a better planning of health care resources in our population. Furthermore, we emphasize the potential of applying CYP genotyping in AIP precision medicine, enabling personalized risk stratification of acute clinical manifestation based on individual CYP allelic profile and leading to predictive personalized medicine for each AIP carrier in the future.