Myengmo Kang

Clinical and molecular characterizations of novel POU3F4 mutations reveal that DFN3 is due to null function of POU3F4 protein

X-linked deafness type 3 (DFN3), the most prevalent X-linked form of hereditary deafness, is caused by mutations in the POU3F4 locus, which encodes a member of the POU family of transcription factors. Despite numerous reports on clinical evaluations and genetic analyses describing novel POU3F4 mutations, little is known about how such mutations affect normal functions of the POU3F4 protein and cause inner ear malformations and deafness. Here we describe three novel mutations of the POU3F4 gene and their clinical characterizations in three Korean families carrying deafness segregating at the DFN3 locus. The three mutations cause a substitution (p.Arg329Pro) or a deletion (p.Ser310del) of highly conserved amino acid residues in the POU homeodomain or a truncation that eliminates both DNA-binding domains (p.Ala116fs). In an attempt to better understand the molecular mechanisms underlying their inner ear defects, we examined the behavior of the normal and mutant forms of the POU3F4 protein in C3H/10T1/2 mesodermal cells. Protein modeling as well as in vitro assays demonstrated that these mutations are detrimental to the tertiary structure of the POU3F4 protein and severely affect its ability to bind DNA. All three mutated POU3F4 proteins failed to transactivate expression of a reporter gene. In addition, all three failed to inhibit the transcriptional activity of wild-type proteins when both wild-type and mutant proteins were coexpressed. Since most of the mutations reported for DFN3 thus far are associated with regions that encode the DNA binding domains of POU3F4, our results strongly suggest that the deafness in DFN3 patients is largely due to the null function of POU3F4.

Hoxc8 represses BMP-induced expression of Smad6.

Proper regulation of bone morphogenetic protein (BMP) signaling is critical for correct patterning and morphogenesis of various tissues and organs. A well known feedback mechanism is a BMP-mediated induction of Smad6, an inhibitor of BMP signaling. Hoxc8, one of the Hox family transcription factors, has also been shown to negatively regulate BMP-mediated gene expression. Here we add another level of Hoxc8 regulation on BMP signaling. Our results show that Hoxc8, when over-expressed in C3H10T1/2 or C2C12 cells, suppressed basal Smad6 promoter activity and its mRNA expression. Activation of Smad6 transcription either by BMP2 treatment or Smad1 over-expression was also abolished by Hoxc8. When chromatin was precipitated from mouse embryos with anti-Smad1 or anti-Hoxc8 antibody, Smad6 promoter sequence was enriched, suggesting that Hoxc8 proteins make complexes with Smad1 in the Smad6 promoter region. Yet, a lack of Hox binding motifs in the Smad6 promoter sequence suggests that instead of directly binding to the DNA, Hoxc8 may regulate Smad6 expression via an indirect mechanism. Our results suggest that the Smad6-mediated negative feedback mechanism on BMP signaling may be balanced by the repression of Smad6 expression by Hoxc8.

In Vitro Osteoblast Differentiation is Negatively Regulated by Hoxc8

Hoxc8 has multiple roles in normal skeletal development. In this paper, a MC3T3-E1 subclone 4 osteogenic cell differentiation model was used to examine expression of Hoxc8 at multiple stages of osteogenesis. We found that Hoxc8 expression levels do not change in the early stage but increase in the middle stage and decrease in the late stage of osteogenesis. A knockdown of Hoxc8 by small-interfering RNA transfection in C2C12 cells indicated that Hoxc8 is a negative regulator of osteogenesis. Similarly, expression of Hoxc8 in C2C12 cells decreases alkaline phosphatase levels induced by bone morphogenetic protein-2 (BMP-2). The results of this study showed that Hoxc8 is involved in BMP-2-induced osteogenesis, and osteoblast differentiation in vitro is negatively regulated by Hoxc8, suggesting that Hoxc8 regulation is essential for osteoblast differentiation.

The systemic effects of sclerostin overexpression using ΦC31 integrase in mice.

Sclerostin, encoded by the Sost gene, is mainly produced by osteocytes in bone and antagonizes the Wnt/β-catenin signaling pathway, which is a requisite for bone formation. Currently, human anti-sclerostin antibodies are being tested in phase III clinical trials. In addition, serum sclerostin levels are reported to be associated with bone mineral density and fracture risk in normal individuals; however, the correlation between serum sclerostin and bone mass remains controversial. To study the effects of the continuous exposure of exogenous sclerostin on bone, a ΦC31 integrase system, which has the characteristics of site-specificity and efficiency, was applied for the delivery of the Sost gene in this study. We injected Sost-attB plasmid with or without ΦC31 integrase plasmid into the mouse tail vein using a hydrodynamic-based method. The site-specific integration of the Sost gene into the mouse genome was confirmed by examining a pseudo-attP site on the hepatic genomic DNA. Sclerostin was expressed in the hepatocytes, secreted into the blood flow, and maintained at high concentrations in the mice with both Sost-attB plasmid and ΦC31 integrase plasmid injections, which was observed by serial measurement. Moreover, the mice with long-term high levels of serum sclerostin showed trabecular bone loss on micro-CT analysis. Peripheral B cell populations were not affected. Our results suggested that sclerostin could be expressed in the liver and sustained successfully at high levels in the blood by using the ΦC31 integrase system, leading to trabecular bone loss. These findings may help to further ascertain the effects of sclerostin introduced exogenously on the skeleton.

Membrane transducing activity of recombinant Hoxc8 protein and its possible application as a gene delivery vector.

In order to examine whether the Hoxc8 protein can deliver nucleic acid into mammalian cells, we designed several Hoxc8-derived recombinant proteins to be synthesized as glutathione S-transferase (GST) fused forms in E.coli (GST-Hoxc81-242, containing a full length of Hoxc8; GST-Hoxc8152-242, possessing a deletion of the acidic N-terminus of Hoxc8; GST-Hoxc8149-208, which contained the homeodomain only). After labeling these proteins with Oregon 488, we examined their membrane transduction ability under the fluorescence microscope and verified that all three proteins showed similar transduction efficiency. The ability of the proteins to form in vitro protein-DNA complexes was analyzed on agarose gel; both GST-Hoxc81-242 and GST-Hoxc8149-208 formed complexes. In contrast, the GST-Hoxc8152-242 protein did not form a complex. The GST-Hoxc8149-208 protein formed a complex with DNA at a mass ratio of 1 : 1 (DNA : protein), and GST-Hoxc81-242 formed a complex at a mass ratio of 1 : 5. When the DNA (pDsRed1-C1) and protein complexes were added to culture media containing mammalian cells, the cells uptook the complexes, which was indicated by red fluorescence expression under the fluorescent microscope. These results indicate that recombinant Hoxc8 derivatives that harbor a homeodomain are able to traverse the mammalian cellular membrane. DNA that is bound to the recombinant derivatives can be carried across the membrane as well. This process could be applied in the development of a useful delivery vector for gene therapy in the future.