PDZD2, a candidate for brachydactyly type A1, encodes a secreted protein that negatively modulates hedgehog signaling

Abstract

Hedgehog (Hh) is an important morphogen that dictates tissue patterning during embryonic development. Recent studies showed that mutation in Indian Hedgehog(IHH)resulted in Brachydactyly type A1(BDA1);however, the disease pathogenesis in patients without IHH mutation remained unknown. PDZD2 is a multi-PDZ domain-containing protein of unknown function in early development. Human PDZD2 is mapped to chromosome 5p13.2, which co-localizes with the disease-associated gene in a family of BDA1 patients, suggesting involvement of PDZD2 in limb development. In situ hybridization revealed that Pdzd2 was expressed in the distal mesenchyme partially overlapping with Shh expressionin mouse limb bud. During digit patterning, Pdzd2 was expressed in the interzone regions that flanked the Ihh/Gli1-expressing phalanx condensation. Moreover,Pdzd2 was expressed in the paraxial mesoderm adjacent to the differentiating neural tube. Pdzd2expression in various Hh-active tissues in mouse and chicken suggested an evolutionary conserved role of PDZD2 in modulating general Hh signaling during early development. Interestingly, PDZD2 protein was detected at the neural tube away from its site of synthesis, suggesting a non-cell autonomous role of PDZD2 possibly via its secreted form, sPDZD2.

Functional studies showed that overexpression of sPDZD2 in the chicken neural tube leads to down-regulation ofNKX2.2andOLIG2expression.sPDZD2 was shown to counteract the ectopic NKX2.2 expression induced by long-range signaling of ectopic HH. Consistently, sPDZD2exhibited an inhibitory effect on SHH-induced reporter activity in a Gli-luciferase cell line. For in vivo analysis, a transgenic mouse line carrying a floxed Pdzd2 allele (Pdzd2fl) was generated to ablate Pdzd2 expression.Pdzd2+/fl mice were crossed with Protamine-cre to generate the null allele (Pdzd2-). However, heterozygous intercross yielded no homozygous mutant as early as E9.5, suggesting early embryonic lethality. Thus, conditional Pdzd2 knock-out in the limb was pursuedusingPrx1-cre.However, no significant perturbation of Hh signaling was observed in Pdzd2fl/fl:Prx1-cre limb buds, which might be due to incomplete knock-out of Pdzd2, or functional redundancy among Hh modulators. To study the relevance of Pdzd2in the development of BDA1, Pdzd2-/fl mouse was crossed with the BDA1 mouseIhhE95K/E95Kto study the effect of reducing Pdzd2expression under defective Hh signaling. Preliminary analysis of the Pdzd2+/-, Ihh+/E95K compound mutants showed more severe phenotypes comparing with IhhE95K/E95K. These included delayed limb development and further diffusion ofGli1expression in the digits, suggestive of a direct involvement of Pdzd2in BDA1. It was speculated that Pdzd2negatively modulated Ihh signaling and restricted Hh signals from entering the interzone, which was required for normal digit patterning. Depletion of Pdzd2might result in an expansion of Ihh signaling into the interzone, leading to the BDA1phenotypessimilar to the current BDA1 disease model proposed forIhhE95Kmutation. Taken together, my study revealed the novel expression of Pdzd2in close proximity to multiple Hh-active tissues during early development and provided the first evidence that PDZD2/sPDZD2 is a negative modulator of general Hh signaling. These data strongly supported PDZD2as a disease-associated locus in the family of BDA1 patients that do not carry IHHmutations.