Congenital Central Hypoventilation Syndrome: PHOX2B Mutations and Phenotype

Congenital central hypoventilation syndrome (CCHS; also known as the literary misnomer "Ondine's curse"; OMIM no. 209880) is a unique disorder of respiratory control that was first described in 1970 (1). Hirschsprung disease (HSCR) and/ or tumors of neural crest origin (neuroblastoma, ganglioneuroblastoma, and ganglioneuroma) occur in association with CCHS in approximately 20 and 6% of cases, respectively (2). Other symptoms of diffuse autonomic nervous system dysfunction/dysregulation (ANSD) are seen frequently in CCHS and include decreased heart rate variability, an attenuated heart rate response to exercise, severe constipation, esophageal dysmotility/ dysphagia, decreased perception of discomfort, pupillary abnormalities, decreased perception of anxiety, sporadic profuse sweating, and decreased basal body temperature (2).

The gene for CCHS was identified as paired-like homeobox (PHOX)2B, located on chromosome 4p12 and encoding a highly conserved transcription factor known to play a key role in the development of ANS reflex circuits in mice (3). The vast majority of individuals with CCHS are heterozygous for a polyalanine repeat expansion mutation involving the second polyalanine repeat sequence in exon 3 of PHOX2B (3-8) (Figure 1). Expansions are in-frame and range from 15- to 39-nucleotide insertions, resulting in expansion of the normal 20-repeat polyalanine tract to 25-33 repeats (4, 6, 8). Most expansion mutations occur de novo in CCHS prohands. However, in the small number of families segregating CCHS, these mutations are inherited as an autosomal dominant trait (8). Also, in about 10% of cases of CCHS, an unaffected parent shows somatic mosaicism for the expansion mutation found in his/her child (8). Polyalanine repeat expansion size has been associated with severity of autonomic dysfunction (number of ANSD symptoms) (4, 8), increased R-R interval on Holter monitoring (7), and severity of ventilatory dependence (4, 8). Four in-frame polyalanine repeat deletion polymorphisms, with 7, 13, 14, and 15 repeats, have been identified in the normal population and do not appear to cause CCHS (3-5, 8, 9).

A small number (less than 10%) of subjects with CCHS do not have a polyalanine repeat expansion but do have other mutations in PHOX2B (3-6, 8, 10) with differing mutational mechanisms and effects on the predicted protein structure. These nonpolyalanine repeat mutations are of interest in defining functionally important parts of the PHOX2B protein and the effects of different mutational mechanisms on clinical manifestations of CCHS, particularly the severity of alveolar hypoventilation, and the presence of Hirschsprung disease and tumors of neural crest origin. The goal of this project was, thus, to identify the spectrum of nonpolyalanine repeat PHOX2B mutations in our series of subjects with CCHS, to determine whether these mutations tend to be inherited or de novo, to evaluate the predicted effects of these mutations on the PHOX2B protein, and to compare clinical features of CCHS in subjects with nonpolyalanine repeat mutations with those seen in the CCHS cohort with the common polyalanine repeat expansion mutation. An abstract describing a portion of this work was presented previously (11).

METHODS

CCHS subjects were ascertained through the CCHS clinical management and research program at Rush University Medical Center (Chicago, IL), or were identified when blood or DNA was sent to the Rush Molecular Diagnostics Laboratory from referring physicians or laboratories for DNA testing for the PHOX2B polyalanine repeat expansion mutation to confirm a clinical diagnosis of CCHS. Subjects in the CCHS program were referred by physicians or through self-referral of families, and clinical information and results of genetic testing provided by physicians and families was entered into a comprehensive CCHS database. For samples referred to Molecular Diagnostics, referring professionals submitted a lab intake sheet accompanying DNA samples for clinical testing. The intake sheet contained basic information regarding presence or absence of Hirschsprung disease and neural crest tumors, and whether the patient required continuous ventilatory dependence or dependence during sleep only. When a polyalanine expansion mutation was identified, the laboratory director (E.B.-K.) contacted the referring professional to relay results and verify information on the intake form. Clinical information and genotype (but no personal identifying information) was entered into a database maintained within the lab. Subjects with clinical features typical of CCHS by physician report and by review of medical records (D.E.W.-M.), but who were not shown to have a typical expansion mutation, were offered participation in the CCHS research project so that PHOX2B gene sequencing could be obtained. All subjects (or guardians) participating in the research project for PHOX2B sequencing signed an informed consent document for CCHS gene analysis and records review. Samples in this cohort were received predominantly from the United States, with less than 10% of samples coming from abroad.

The polyalanine repeat coding sequence in exon 3 of PHOX2B was analyzed for all subjects as described previously (8). If no expansion was seen, the entire coding region and intron-exon boundaries of PHOX2B were sequenced after polymerase chain reaction amplification of each exon followed by column purification of polymerase chain reaction products and automated cycle sequencing (8).

RESULTS

A total of 184 samples from individuals with the characteristic phenotype of CCHS were analyzed. A subgroup of 67 of these cases was presented previously (8). Of that group, 65 were reported to have a PHOX2B common polyalanine repeat mutation, 1 was found to have a nonpolyalanine repeat mutation, and 1 presented a normal coding for 20 polyalanine repeats in both alleles of the PHOX2B gene. We subsequently identified a labeling error at the laboratory of origin for the sample from the subject with the normal result. A new sample from the same subject demonstrated that the subject was heterozygous for a PHOX2B polyalanine repeat expansion mutation with 30 repeats. In the current expanded cohort, of 184 total samples analyzed at Rush University Medical Center with typical features of CCHS, 170 (92%) were heterozygous for the polyalanine repeat expansion mutation in PHOX2B, with the size of the repeat tract ranging from 25 to 33 repeats. A deletion polymorphism of 15 repeats was also found in one individual (1%) heterozygous for a repeat expansion mutation, similar to the frequency of this variant in the normal population. Fourteen probands with typical CCHS but not possessing a repeat expansion mutation were heterozygous for other mutations in PHOX2B. Missense mutations occurred in four probands (two mutations), one proband had a nonsense mutation, and nine probands (four mutations) had frameshift mutations, all located at the end of exon 2 or within exon 3 (Figure 1). Mutations and the resultant predicted effect on the PHOX2B protein are shown in Table 1. Results of PHOX2B analyses performed on the accessible parents and siblings of these 14 probands, if available, are also shown in Table 1. Six of nine mutations for which both parental samples were available and negative were apparently de novo, but the missense mutation in proband 2 was also present in the reportedly asymptomatic mother and one frameshift mutation (proband 6) was present in the reportedly asymptomatic mother and a sister with Hirschsprung disease but not CCHS, suggesting variable penetrance for this particular PHOX2B mutation. Further, the mother of proband 9 showed low-level somatic mosaicism for the 35-bp deletion. Although only the mother's sample was available for proband 1, the mother was a carrier of the missense mutation (same mutation as for proband 2).

Other nonpolyalanine repeat expansion mutations with significant effect on the PHOX2B protein have been reported as associated with CCHS by groups in Italy (4, 10), Japan (5), and France (6). A summary of all mutations identified by other groups is shown in Table 2. Data from subjects in this report (Table 1), and those in Table 2, reveal that 32 individuals with CCHS and nonpolyalanine repeat mutations in PHOX2B have been described worldwide, and mutations include mostly frameshift mutations (24 of 32, 75%), but also a nonsense mutation (1 of 32, 3%), missense mutations (6 of 32, 19%), and a missense mutation with stop codon alteration (1 of 32, 3%). The 38- and 35-bp deletions begin in the first codon of the polyalanine repeat, and these are recurring mutations (found in seven and three probands in unrelated families, respectively) identified by investigators in different countries. The G422A and A428G mutations have also each been found in several unrelated probands, and these are the only two missense mutations yet identified in CCHS probands. All CCHS-associated mutations at our center and elsewhere are found at the end of exon 2 or in exon 3 (Figure 1). Interestingly, the Italian 618delC mutation, located in the same area of PHOX2B as the single base-pair deletion in our 577delG family, was inherited from the asymptomatic mother (4), consistent with our finding that some frameshift mutations can be variably penetrant. It should be noted, however, that comprehensive physiologic testing of these family members has not been reported.

Review of available clinical information from our cohort and from subjects reported worldwide (4-6, 10) (Figure 2) with PHOX2B mutations revealed that HSCR was reported in 87% of cases (26/30) with a non-polyalanine repeat mutation, and 17% of cases (60/345) with a polyalanine repeat mutation. Continuous ventilatory dependence was reported in 79% of cases (15/19) with a non-polyalanine repeat mutation and 34% of cases (66/193) with a polyalanine repeat mutation. Neural crest tumors were reported in 50% of cases (8/16) known to be over one year of age with a non-polyalanine repeat mutation, and 1 % of cases (4/279) with a polyalanine repeat mutation. All polyalanine repeat probands with tumors were heterozygous for large (either 31 or 33 repeat) expansion mutations. Thus, a much higher rate of HSCR, higher frequency of continuous ventilatory dependence, and more frequent neural crest tumors is identified in the non-polyalanine repeat mutation group compared to those with a polyalanine repeat expansion, both in our series of patients and when published data from all other groups globally were combined with our data (4-6, 10) (Figure 2).

DISCUSSION

This article describes PHOX2B mutation analysis in a large cohort of 184 subjects referred to Rush University Medical Center with the classical CCHS phenotype. Ninety-two percent of subjects with CCHS are heterozygous for the polyalanine repeat expansion mutation in exon 3. However, in this series, all 14 subjects presenting with typical features of CCHS who did not have a polyalanine repeat expansion mutation in PHOX2B had a different mutation in the PHOX2B gene. The 100% mutation detection rate in this cohort most likely has to do with the careful clinical review and restriction of inclusion to subjects who met strict criteria for CCHS (2). It is possible that occasional additional subjects with nonpolyalanine repeat mutations in PHOX2B would be identified if nontypical subjects with CCHS-spectrum symptoms were analyzed.

Mutations described here and worldwide include missense, nonsense, and frameshift mutations, but all are located at the 3' end of the PHOX2B gene from the last 6 bp of exon 2 to the end of exon 3. Both of the missense mutations in our series, A428G and G422A, occurred in two unrelated CCHS probands. These same mutations were identified in the French series (6), suggesting these mutations have arisen several times on independent backgrounds. Furthermore, it is worth noting that these two mutations are the only described missense mutations identified in PHOX2B and occurring in cases of CCHS. Both mutations alter a highly conserved amino acid in the PHOX2B sequence; however, they also both alter residues predicted to be important in the splice donor consensus sequence for splicing of exons 2 and 3 and may actually exert their effect through a splicing defect. Altered splicing would produce an abnormal PHOX2B C terminus, similar to the abnormal proteins observed with the frameshift mutations. Further work is needed to analyze RNA splicing with these mutations in a tissue expressing PHOX2B. There is another multiply recurrent mutation in our series, also reported in both French (6) and Italian (4) cohorts, that involves deletion of 35 or 38 bp from the beginning of the polyalanine repeat stretch, which may be generated through a specific mutational mechanism during replication of the polyalanine repeat.

Most nonpolyalanine repeat mutations arise de novo, although they can be inherited and variably penetrant in families. Family members who have HSCR but not CCHS may carry these mutations, and some carriers may be asymptomatic as seen in families of probands 1, 2, and 6. Also, low-level somatic mosaicism for a large deletion was observed in the mother of proband 9. Taken together, this information suggests that recurrence risk can exist in families in which the CCHS proband has a nonpolyalanine repeat mutation, and emphasizes the importance of screening parents once a mutation has been identified in their child. Furthermore, apparently asymptomatic parents or siblings of a CCHS child who carry the child's mutation in PHOX2B may be at risk for developing hypoventilation in sleep or a later, adult-onset atypical CCHS phenotype (13). Identification of mutation carriers in these families allows for appropriate clinical monitoring and early implementation of treatment, if required.

As was also observed in the CCHS cohort from France (6), nonpolyalanine repeat mutations appear virtually always to result in Hirschsprung disease and to have a high rate of neural crest tumors in patients who live past one year of age, suggesting that many of these mutations are more disruptive to the function of PHOX2B than the polyalanine repeat expansions. Individuals with CCHS and a nonpolyalanine repeat mutation, thus need to be monitored closely for emergence of a neural crest tumor, specifically neuroblastoma. Nonpolyalanine repeat mutations appear to be divided into two groups, according to clinical presentation. The majority of these mutations appear to produce very severe disease with continuous full ventilatory support and HSCR with extensive gut involvement, and in those over one year of age an increased tumor risk (11 patients in our series including all large deletions and one other +1 frameshift, one nonsense mutation and one missense mutation). There is a minority however which also has a very high incidence of HSCR but milder disease, and incomplete penetrance in the families of three patients with a particular missense mutation or frameshift. Thus, presence of HSCR and a CCHS phenotype is a strong predictor of a nonpolyalanine repeat mutation in PHOX2B when polyalanine repeat expansion screening is negative.

Lack of identification of mutations in exon 1 and most of exon 2 (Figure 1) may suggest one of two possibilities, given that disruption of the 5' end of the gene is more likely to lead to PHOX2B haploinsufficiency caused by lack of production of any PHOX2B protein due to nonsense-mediated decay of mRNA. One possible interpretation would be that haploinsufficiency may produce a more severe reduction of activity than that which accompanies a marginally functional 3'-mutated protein. Consequently, haploinsufficiency induced by mutations at the 5' end of the gene would be incompatible with life. Alternatively, and more likely, haploinsufficiency is insufficient to cause the phenotype and therefore individuals with 5' mutations in PHOX2B do not have CCHS and are not ascertained. In this scenario, the mutations in distal exon 2 and exon 3 of PHOX2B result in a dominant negative or gain-of-function effect. Dominant negative effects have been proposed as a mutational mechanism for other transcription factors in which polyalanine expansion mutations cause disease, such as HOXD13 mutations in synpolydactyly syndrome (14). These effects may occur because of the mutant PHOX2B protein directly inhibiting the function of the normal protein during the normal dimerization process required for the transcriptional regulation function of PHOX2B, or as a result of sequestering of the normal protein in nuclear or cytoplasmic aggregates, as has been demonstrated for other autosomal transcription factors (HOXD13, FOXL2, and ZIC2) with associated polyalanine repeat expansions (14-17).

Support for the dominant negative/gain-of-function hypotheses over the haploinsufficiency model for PHOX2B mutations in CCHS comes from prior reports of two individuals with interstitial deletions of 4p12 containing the PHOX2B region (18, 19) who therefore have haploinsufficiency for PHOX2B but do not have CCHS. Also, mice heterozygous for a targeted phox2b deletion show only a mild respiratory phenotype, limited to the newborn period (20). Humans with a similar phenotype would be unlikely to be ascertained clinically, although they might have an increased risk of sudden infant death syndrome or sudden unexplained death in childhood. Further support for the dominant negative hypothesis comes from the observation of sequestering of the normal PHOX2B protein in nuclear aggregates from cells coexpressing normal and mutant PHOX2B (10). However, one study found that a 10-fold excess of the mutant protein was required before significant amounts of the normal protein were observed in aggregates (21). Thus, the presence of the aggregates may indicate a gain-of-function mechanism producing cellular toxicity or dysfunction.

Both polyalanine repeat mutations and nonpolyalanine repeat mutations in PHOX2B impair transcriptional function, and transcriptional impairment increases with mutation size for polyalanine repeat mutations (10, 21). For nonpolyalanine repeat mutations, the least impairment of transcriptional function, comparable to activity from the smallest (five-repeat) polyalanine expansion, was seen with the 618delC mutation (10), which results in a mutated protein similar to our 577delG mutant, both -1 frameshifts occurring in the same area of the protein. This is consistent with the finding that the 618delC (4), 577delG (this article), and five-repeat polyalanine expansion mutations can all be associated with incomplete penetrance in families of the CCHS proband or adult-onset disease (13). Thus, specific types and locations of mutations may be more prone to present with variable penetrance, based on a milder effect on PHOX2B-mediated transcription. The G422A mutation also shows incomplete penetrance, but the effect on PHOX2B function remains to be tested.

Polyalanine expansions resulted in cytoplasmic sequestration of the mutant protein, whereas this effect was not seen with frameshifts (10, 21). Frameshift mutations in fact resulted in enhanced PHOX2A transcription, whereas polyalanine repeat mutations reduced PHOX2A transcription (10). The different pattern of transcriptional dysregulation and mutant protein aggregation observed with frameshifts suggests that these mutations exert effects on PHOX2B target promoters through a somewhat different mechanism than the more prevalent polyalanine repeat mutations. Excessive activation of PHOX2A was proposed to be a potential factor in the high incidence of neural crest tumors in CCHS subjects with nonpolyalanine repeat mutations. In any case, given the more severe hypoventilation observed in most subjects with nonpolyalanine repeat mutations, and near universal presence of HSCR, many of these mutations seem to result in cellular disease mechanisms distinct from or in addition to those operant in the presence of a polyalanine repeat mutation, thus producing a larger and more generalized impact on PHOX2B function.

Continued identification of nonpolyalanine repeat mutations in PHOX2B in individuals with symptoms in the CCHS/ANSD spectrum will yield further information about the importance of different domains in the PHOX2B protein and will help elucidate the genetic mechanisms through which mutations in PHOX2B result in disease.

Note added in proof: Subsequent to the acceptance of this work, 17 more CCHS subjects were identified for a total of 201 PHOX2B mutation-confirmed cases of CCHS at Rush University Medical Center. These new cases included 15 with polyalanine repeat mutations and 2 with nonpolyalanine repeat mutations (722del38 and 689dup8).

Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript.

© 2006 American Thoracic Society Provided by ProQuest LLC. All Rights Reserved.