The Matrix Biochemistry Section focuses its research on the functions of five major noncollagenous proteins first found associated with the mineralized matrix of bones and teeth. These include: bone sialoprotein (BSP);osteopontin (OPN);dentin matrix protein 1 (DMP1);dentin sialophosphoprotein (DSPP);and matrix extracellular phosphoglycoprotein (MEPE). We have made a strong case for the genetic relatedness of these seemingly different proteins and there is increasing acceptance of the SIBLING (Small Integrin-Binding LIgand, N-linked Glycoprotein) family concept. The genes encoding these proteins are all clustered in a tandem fashion within a short (400,000 basepair) region of human chromosome 4. Because the same genes are clustered together in all mammals studied, they are likely to be the result of gene duplications and subsequent divergence more than 180 million years ago. One of the conserved motifs among the different SIBLING members (and among animal species) is the integrin-binding tripeptide, arginine-glycine-aspartate (RGD), used to bind the SIBLINGs to the cell surface. In previous work with a former postdoctoral fellow, Dr. Kalu Ogbureke, we showed the up-regulation of three SIBLING proteins (BSP, DSPP and OPN) in nearly all cases of oral cancer. Oral cancer (like colon and cervical cancer) can often be discovered in a precancerous state by health care professionals during routine screenings. A major problem is that while some precancerous lesions may progress to cancer, most others that look the same do not. The ability to identify those that are most likely to progress can be of obvious benefit to the patient. Extensive surgery as well as chemo or radiation therapies may then be selected for only those cases where progression to frank cancers is most likely. In a 2010 Cancer publication we showed that: 1) there was no significant correlation between degree of dysplasia in oral premalignant lesions and transition to cancer;2) 78% of the patients whose precancerous lesions expressed DSPP (but not BSP) progressed to cancer;and 3) none of those that expressed BSP (without DSPP) progressed. Because the average rate of progression within our cohort was 20%, the BSP-positive state appears to be protective while a DSPP-positive status indicates a significant risk over presenting with dysplasia alone. In our collaboration with Dr. Neal Fedarko (John Hopkins Univ.), we continued to develop reagents to perfect the detection of SIBLING proteins in serum of cancer patients. The most recent research project was in response to a 2006 publication we had with our long-term collaborators in Belgium (Bellahcne &Castronovo) showing the strong up-regulation of DSPP in prostate cancer tumors. Earlier papers with Dr. Fedarko using BSP, OPN, and DMP1 serum assays (NIH holds our patent for the special processing that is necessary for detecting the SIBLINGs in serum) showed good statistical separation of normal vs. cancer patient groups but there was always concern that individual patients could have elevated levels of one or more of these proteins and yet not have one of the indicated cancers. (Bone turnover, for example, can elevate BSP and OPN while virtually any inflammation event will elevate OPN.) We have now shown that DSPP is highly elevated in the serum of prostate cancer patients (Clinical Cancer Res). The validation group analysis showed that segregating samples as a function of both PSA and DSPP removed any overlap between the normal and cancer group, thereby increasing sensitivity (90%) and specificity (100%) of the assay (giving a ROC of 98%). DSPP, possibly the most acidic and hydrophilic protein made by humans, is the most abundant noncollagenous protein in dentin. We analyzed 36 mammalian species (one or more from each major branch of the mammalian family tree) as well as a reptile and chicken for the DPP portion of DSPP (BMC Evol Biol). While the integrin-binding motif, RGD, was present in about half of the species, only vestigial remnants of this tripeptide were identified in others. This suggests that unlike other SIBLINGs, the retention of the ability of DSPP to bind to integrins through this motif is being lost with time. The number of tandem repeats of the nominal SerSerAsp phosphorylation motif in toothed mammals (including baleen whale and platypus which have embryonic teeth), ranged from 75 (elephant) to >230 (human). The only two truly toothless species we were able to analyze (anteaters) independently acquired nonsense codons immediately C-terminally to their BMP1-susceptible motifs. This suggests that while the cleavage may be important for processing of the DSPP product in other tissues, the C-terminal DPP domain itself may be required only in dentin. From our studies we have proposed that DSPP is the result of the most recent duplication of a SIBLING gene family. The DMP1 gene of a common ancestor of the reptiles and mammals duplicated and then the mammals and reptiles used different copies to separately acquire a DSPP-like protein by convergent evolution. In 2010 we also finished two related projects involving the processing of secreted proteins by the matrix metalloproteinase, BMP1. First we expanded our earlier work on the cleavage of DMP1 by showing that a second SIBLING, DSPP, was also differentially cleaved by three different isoforms of BMP1 into the two most abundant noncollagenous proteins found in dentin (Matrix Biology 2010). Furthermore, two proteins (PCOLCE and sFRP2) reported to enhance the digestion of some BMP1 substrates were tested for their effects on the processing of both DSPP and DMP1. Our results showed that neither enhancer helped any of the three isoforms of BMP1 digest DSPP or DMP1 suggesting that either these SIBLING proteins do not require such enhancement for efficient cleavage or that another, currently unknown protein, serves this function. Transduced fibroblasts process DMP1 in vivo much more efficiently than our three recombinant BMP1 isoforms in vitro suggesting that the condition for processing is in some way much better in cells. We found that BMP1 is more stable in normal saline but is a much better protease in low salt. Therefore, in vivo biochemical conditions and an unidentified enhancing protein both remain reasonable hypotheses at this time. In the same vein, another of our 2010 papers showed that decorin, a small proteoglycan thought to be involved in both collagen fibril assembly and matrix sequestration of members of the TGF-beta superfamily, has its conserved propeptide removed by the same three isoforms of BMP1. As a result of our control experiments on sFRP2 noted above, we saw a previously unreported biological activity of this important protein. Secreted Frizzled-related proteins are involved in embryonic development as well as pathological conditions such as bone and myocardial disorders as well as many cancers. Because of their sequence homology with the Wnt-binding domain of Frizzled, sFRPs have generally been considered antagonists of canonical Wnt signaling. Using human embryonic kidney cells (HEK293A), we published that sFRP2 enhanced Wnt3a-dependent phosphorylation of LRP6 as well as elevated both total beta-catenin levels and its nuclear translocation. While addition of recombinant sFRP2 had no activity by itself, Top/Fop luciferase reporter assays showed a dose-dependent increase of Wnt3a-mediated transcriptional activity by sFRP2. This sFRP2 enhancement of Wnt3a signaling was abolished by treatment with the Wnt antagonist, Dickkopf-1 (DKK1). These results support sFRP2s role as an enhancer of Wnt/beta-catenin signaling, a result with biological impact for both normal development and diverse pathologies such as tumorigenesis.
Showing the most recent 10 out of 17 publications
