C4B null alleles are not associated with genetic polymorphisms in the adjacent gene CYP21A2in autism

Autism is a severe neurodevelopmental disorder that is approximately four times more common in males than females. The current prevalence for the disorder is approximately 1 in 152 children [1]. Although there are various mechanisms that can lead to this behaviorally defined condition, in most cases the etiology remains unknown. A strong genetic component clearly exists [2]; however, consistent detection of disease associated genetic variants has rarely been reported. Whole genome scans using microarray technology may better detect the genetic contributions to autism susceptibility [3].

Genes located in the RCCX module found on chromosome 6 in the human leukocyte antigen (HLA) locus are associated with various disease states [4, 5]. This module contains the genes RP, C4, CYP21, and TNX (abbreviated RCCX) in a contiguous sequence. Different variants of these four genes can exist in the RCCX module including RP1 or RP2, C4A or C4B (long or short), CYP21A2 or CYP21A1P, and TNXA or TNXB. RP2 and TNXA are gene fragments while CYP21A1P is a pseudogene. A single chromosome usually contains one, two, three RCCX modules in tandem, but rare cases can have four modules.

Because of the diversity in the number and size of the RCCX modules misalignments and unequal crossovers occur during meiosis resulting frequently in deletions, conversions, duplications along with the acquisition of mutations from nearby pseudogenes or gene segments [6]. One such mutation in this module is the missing/nonfunctional C4A or C4B gene (C4 null allele). Therefore, the C4 containing complex is an extraordinarily complex region of the human genome [7] (Figure 1).

The RCCX module may play a significant role in the genetic underpinnings of autism. Several studies have shown that the frequency of C4B null alleles is increased in individuals with autism [8–10], the most recent of which found that 42.4% of autistic subjects carry a C4B null allele compared to 14.5% of controls [10]. The CYP21A2 gene is located approximately 3 kb downstream of C4 and the concurrent deletion of C4B with portions of the CYP21A2 has been described [11, 12]. Therefore, the aim of the present research was to determine if the C4B null allele, found frequently in subjects with autism, is associated with CYP21A2 mutations. As well, the overall frequency of CYP21A2 mutations in autistic verses control subjects was determined.

C4 genes encode innate immune C4 proteins that are important in the complement cascade. CYP21A2 encodes an enzyme, 21-hydroxylase, which is important in the synthesis of cortisol and in maintaining proper androgen levels.

Over 2,250 genetic typings of CYP21A2 mutations were completed in 80 autistic and 60 controls subjects that had previously been typed for C4B null alleles [10]. Fifteen total CYP21A2 mutations where detected; however, only one individual (autistic) carried a chromosome containing both a C4B null allele and CYP21A2 mutations. Therefore, in these subjects it does not appear that C4B null alleles are associated with the CYP21A2 mutations studied.

Overall, the number of CYP21A2 mutations did not differ between the autism and control groups and no group differences were found in frequencies of individual CYP21A2 mutations. The frequency of CYP21A2 mutations seen in this research is similar to general population frequencies reported by others [19]. Five individuals with autism (6.25%) and five controls (8.33%) carried a CYP21A2 mutation. Two individuals with autism and one control subject carried two or three mutations. In these cases the mutations all typed to individual parents. Thus, the multiple mutations were found on single chromosomes. When a mutation was found in the child's DNA it was also present in DNA from one of the parents, thus confirming both accurate typing and absence of de novo mutations in the study children.

HLA extended haplotyes known to contain C4B null alleles are increased in frequency in autistic individuals (i.e. HLA extended haplotypes 35.2 (n = 4), 35.3 (n = 2), 44.1 (n = 8), and 58.1 (n = 1)). Extended haplotypes 35.3 and 58.1 have monomodular RCCX structures (C4A: C4B null allele) and extended haplotypes 35.2 and 44.1 are bimodular (C4A, C4A:C4B null allele). No CYP21A2 mutations were observed in these haplotypes.

The HLA 8.1 is present in 10% of the Caucasian population and represents the most common extended haplotype. It has a monomodular RCCX structure with a C4A null allele and a normal C4B gene (C4A null allele: C4B). This extended haplotype, referred to as COX in the MHC Haplotype Project [20], has been completely sequenced [21]. Of the eighteen subjects with an 8.1 extended haplotype one had the rare SNP creating the TTG codon that encodes for leucine at position 281 while 17 subjects had the common codon GTG that encodes for valine at position 281. This observation is in agreement with DNA sequencing data that shows low level SNP diversity in 8.1 extended haplotypes from different individuals [22].

The specific CYP21A2 mutations were chosen because they are the most common mutations that result in significant decreases in 21-hydroxylase activity. One of the mutations investigated was a 30 kb deletion/conversion that deletes C4B and converts CYP21A2 into a nonfunctional CYP21A1P/CYP21A2 hybrid with its pseudogene (Figure 2). This mutation was of particular interest because it provides a direct link between a C4B null allele and a CYP21A2 mutation. One autistic subject and two controls carried this mutation. Again, these three polymorphisms were inherited and did not involve de novo mutations.

It has been reported that two of the mutations, P30L and V281L, partially inhibit 21-hydroxylase activity (30–60% of normal), whereas the other mutations analyzed cause either complete or nearly complete enzyme inhibition [23]. Four subjects with autism (5%) carried a partially inhibiting mutation (P30L or V281L) compared to two control subjects (3.33%). This difference is not statistically significant.

21-hydroxylase deficiency is the most common cause of congenital adrenal hyperplasia. Some evidence supports the idea of 21-hydroxylase being involved in autism [24], which could result in the excessive androgen production seen in some cases [25] and thereby contribute to disease etiology [26, 27]. The present data does not provide genetic support for 21-hydroxylase involvement in autism.

This study examined both mono and bimodular RCCX genetic modules that contain C4B null alleles for mutations in the adjoining CYP21A2 gene. The C4B null alleles seen in autism are not associated with the CYP21A2 genetic mutations examined in this study. The frequency of CYP21A2 mutations was similar between the autism and control groups. Based on family typings no de novo mutations of C4B or CYP21A2 were apparent in study subjects. Therefore, the CYP21A2 mutations studied do not appear to contribute to the etiology of autism. However, a role for CYP21A2 in autism cannot be ruled out as other factors affecting CYP21A2 gene expression such as promoter polymorphisms or epigenetic variation were not studied and may be relevant [28]. As well, a weak association may be beyond the statistical power of the present study to detect.