Ehlers-Danlos syndrome (EDS) is a heterogeneous group of heritable connective-tissue disorders, generally affecting skin, joints, and blood vessels. The most recent classification recognizes six subtypes (Beighton et al. 1998), of which the hypermobility type (HT-EDS [formerly EDS type III] [MIM 130020]) is the most common. This type of EDS is similar to benign joint hypermobility syndrome (BJHS), and both are often considered to represent the same hyperlaxity syndrome, since no clear clinical distinction can be made (Grahame 1999). Although various causative genes have been found in all other types of EDS, the genetic basis of HT-EDS or BJHS remains unexplained (Steinmann et al. 2002). One family has been described that has a missense mutation in COL3A1 (Narcisi et al. 1994), resulting in a phenotype that resembles HT-EDS, without obvious vascular complications. Mutations in COL3A1 generally result in the severe vascular type of EDS (MIM 130050). To our knowledge, no other cases of COL3A1 mutations in HT-EDS have been reported.

Recently, we showed that deficiency of the extracellular-matrix protein tenascin-X (TNX), encoded by the TNXB gene, causes a new type of recessively inherited EDS (Schalkwijk et al. 2001). Patients with complete deficiency of TNX showed marked joint hypermobility, skin hyperextensibility, and easy bruising. The absence of atrophic scars and recessive inheritance distinguishes TNX deficiency from the classical type of EDS. In our initial report (Schalkwijk et al. 2001), only a few heterozygous family members were available for examination. Here, we have examined all 20 heterozygous family members (individuals from families A–D in table 1) who were available for further study, regardless of clinical symptoms; in all of these individuals, we have found significantly reduced serum TNX levels (56% ± 6% vs. 100% ± 14% in the control population; P<.001, by Student's t test) (fig. 1f), and, in 17 of them, we have confirmed heterozygosity for a truncating TNXB mutation (table 1). Clinical examination revealed generalized joint hypermobility in nine family members (45%), using the Beighton score (Beighton et al. 1973), for HT-EDS, or the Brighton criteria (Grahame et al. 2000), for BJHS (table 1 and fig. 1e). Skin hyperextensibility and easy bruising, frequently seen in the individuals with complete TNX deficiency, were absent. A number of patients with haploinsufficiency had recurring joint dislocations and chronic joint pain, as are seen in HT-EDS and BJHS. Only four family members carrying two normal TNXB alleles were available for study, of whom none had hypermobility. The local medical ethics committee (CMO Regio Arnhem-Nijmegen) approved the study protocol, and informed consent was obtained from all patients.

A striking finding is that 0 of the 6 males with haploinsufficiency fulfilled the clinical criteria for HT-EDS or BJHS, whereas 9 of 14 (64%) females were positive. This finding is in accordance with previous population-based studies that show a female preponderance in joint hypermobility syndromes (Larsson et al. 1987; Rikken-Bultman et al. 1997). In a control group of 30 unaffected females of the same age as the females with haploinsufficiency in the present study, we found no individuals with a Beighton score >4. This indicates that the prevalence of generalized joint hypermobility in a population of females with haploinsufficiency is significantly higher than in a control population (P<.001, by χ² test). No sex differences in serum TNX levels in unaffected individuals and individuals with haploinsufficiency were found (not shown).

Because our observations in families carrying previously described TNXB mutations suggested an association between TNXB haploinsufficiency and joint hypermobility, we wondered about the prevalence of TNXB haploinsufficiency in patients with HT-EDS. We measured serum TNX levels (by ELISA) in an unselected cohort of 80 patients with HT-EDS who were recruited through the Dutch organization for patients with EDS. All patients were diagnosed with HT-EDS by a medical specialist, and ~90% were female. Although the mean serum TNX level was not different in the cohort with HT-EDS overall (99.4%±19.7%) (fig. 1f), six of these patients (7.5% [all female]) had serum TNX levels >2.5 SDs (65%) below the mean for unaffected individuals. On the basis of the normal distribution of serum TNX levels, only 0.6% of individuals would be expected to have such low serum TNX levels, which is significantly less than the frequency found in the population with HT-EDS described in the present study (P<.001, by Fisher's exact test).

Clinically, patients with reduced TNX levels showed hypermobile joints, often associated with joint subluxations and chronic musculoskeletal pain (table 1). The clinical findings in these patients differ from those with complete TNX deficiency. Patients with haploinsufficiency do not have skin hyperextensibility and lack the easy bruising seen in patients with TNX deficiency. In addition, TNXB haploinsufficiency is expected to be an autosomal dominant trait, which is in accordance with the observed mode of inheritance of HT-EDS and BJHS.

On screening for the presence of a 30-kb deletion described previously (Burch et al. 1997; Schalkwijk et al. 2001), we found that this deletion was present in one of these six patients. The 30-kb deletion creates a fusion gene of TNXB and XA, a partial duplicate of TNXB. The XA gene has an internal deletion that truncates its ORF, rendering XA and the fusion gene nonfunctional (Gitelman et al. 1992). The deleted allele also lacks CYP21, so this individual is also a carrier for congenital adrenal hyperplasia. Subsequently, we PCR amplified and directly sequenced the coding regions and the intron-exon boundaries of TNXB in the other five patients presumed to have haploinsufficiency (for primers used, see Schalkwijk et al. 2001). One patient (individual F in table 1) was heterozygous for a 2-bp deletion, [AA56063] del, in exon 8, resulting in a premature stop codon at the position of amino acid 1231. In the other four patients, we were unable to identify mutations in TNXB. These patients may have mutations, in regulatory sequences or in exons of the TNXB gene, that have not yet been identified, or they may represent the extreme in normal variation of TNX expression.