Search

Figure S6. IFN-gamma induces chromatin remodeling in the GPR109A promoter regions.

Table S3. Differential expression of inflammatory genes in human colon carcinoma

Short-chain fatty acids, metabolites produced by colonic microbiota from fermentation of dietary fiber, act as anti-inflammatory agents in the intestinal tract to suppress proinflammatory diseases. GPR109A is the receptor for short-chain fatty acids. The functions of GPR109A have been the subject of extensive studies; however, the molecular mechanisms underlying GPR109A expression is largely unknown. We show that GPR109A is highly expressed in normal human colon tissues, but is silenced in human colon carcinoma cells. The GPR109A promoter DNA is methylated in human colon carcinoma. Strikingly, we observed that IFNγ, a cytokine secreted by activated T cells, activates GPR109A transcription without altering its promoter DNA methylation. Colon carcinoma grows significantly faster in IFNγ-deficient mice than in wild-type mice in an orthotopic colon cancer mouse model. A positive correlation was observed between GPR109A protein level and tumor-infiltrating T cells in human colon carcinoma specimens, and IFNγ expression level is higher in human colon carcinoma tissues than in normal colon tissues. We further demonstrated that IFNγ rapidly activates pSTAT1 that binds to the promoter of p300 to activate its transcription. p300 then binds to the GPR109A promoter to induce H3K18 hyperacetylation, resulting in chromatin remodeling in the methylated GPR109A promoter. The IFNγ-activated pSTAT1 then directly binds to the methylated but hyperacetylated GPR109 promoter to activate its transcription. Overall, our data indicate that GPR109A acts as a tumor suppressor in colon cancer, and the host immune system might use IFNγ to counteract DNA methylation–mediated GPR109A silencing as a mechanism to suppress tumor development. Cancer Immunol Res; 3(7); 795–805. ©2015 AACR.

Supplemental Data

Short-chain fatty acids, metabolites produced by colonic microbiota from fermentation of dietary fiber, act as anti-inflammatory agents in the intestinal tract to suppress proinflammatory diseases. GPR109A is the receptor for short-chain fatty acids. The functions of GPR109A have been the subject of extensive studies; however, the molecular mechanisms underlying GPR109A expression is largely unknown. We show that GPR109A is highly expressed in normal human colon tissues, but is silenced in human colon carcinoma cells. The GPR109A promoter DNA is methylated in human colon carcinoma. Strikingly, we observed that IFNγ, a cytokine secreted by activated T cells, activates GPR109A transcription without altering its promoter DNA methylation. Colon carcinoma grows significantly faster in IFNγ-deficient mice than in wild-type mice in an orthotopic colon cancer mouse model. A positive correlation was observed between GPR109A protein level and tumor-infiltrating T cells in human colon carcinoma specimens, and IFNγ expression level is higher in human colon carcinoma tissues than in normal colon tissues. We further demonstrated that IFNγ rapidly activates pSTAT1 that binds to the promoter of p300 to activate its transcription. p300 then binds to the GPR109A promoter to induce H3K18 hyperacetylation, resulting in chromatin remodeling in the methylated GPR109A promoter. The IFNγ-activated pSTAT1 then directly binds to the methylated but hyperacetylated GPR109 promoter to activate its transcription. Overall, our data indicate that GPR109A acts as a tumor suppressor in colon cancer, and the host immune system might use IFNγ to counteract DNA methylation–mediated GPR109A silencing as a mechanism to suppress tumor development. Cancer Immunol Res; 3(7); 795–805. ©2015 AACR.

Diffuse large B-cell lymphoma (DLBCL) includes two prognostically important subtypes, the germinal center B-cell (GCB) and the non-GCB types. The aim of this study was to evaluate immunohistochemical approaches for predicting the survival of patients with DLBCL following autologous hematopoietic stem cell transplantation (AHSCT). We identified 62 patients with DLBCL who either had an initial complete remission (17 patients) or received salvage chemotherapy for relapsed or refractory disease (45 patients), followed by AHSCT. Tissue microarrays were immunostained with monoclonal antibodies against GCET1, CD10, BCL6, MUM1, FOXP1 and LMO2. Using the Hans algorithm, we classified 50% of the cases as GCB type, whereas the Choi algorithm classified 58% as GCB type and LMO2 was positive in 69%. However, no significant differences were found in the 5-year overall or event-free survivals using any of these approaches. In conclusion, cell of origin fails to predict survival of DLBCL patients treated with AHSCT.

A frequent chromosomal translocation in mature B-cell non-Hodgkin lymphoma affects band 3q27 and results in the deregulation of the B-cell lymphoma 6 (BCL6) gene. Two breakpoint clusters have been described thus far, the major breakpoint region (MBR) and an alternative breakpoint region (ABR) that is located 245-285 kb 5' to BCL6. Translocation at the MBR predominates in diffuse large B-cell lymphoma, whereas translocation at the ABR is reported to be frequently associated with grade 3B follicular lymphoma. However, translocation at the ABR has not been studied in a large series of follicular lymphomas, particularly t(14;18)-negative follicular lymphomas. Therefore, we studied BLC6 rearrangements at the MBR and ABR by using break-apart fluorescence in situ hybridization (FISH) probes in 142 cases of follicular lymphomas, including 63 t(14;18)-negative and 79 t(14;18)-positive cases. Conventional cytogenetic (karyotype) analysis was also performed in 58 of the 63 t(14;18)-negative cases. BCL6 rearrangement was found in 26% of t(14;18)-negative and 19% of t(14;18)-positive follicular lymphoma. t(14;18)-negative cases showed a high frequency of rearrangement at the ABR (12%) with an ABR/MBR ratio of 0.86, compared with only 5% with an ABR/MBR ratio of 0.36 in the t(14;18)-positive cases. BCL6 rearrangements were found in all grades of follicular lymphoma but were most frequent in grade 3 t(14;18)-negative follicular lymphoma (60%). FISH analysis had a higher sensitivity for detecting BCL6 rearrangements than conventional cytogenetics. In conclusion, BCL6 rearrangements occur at a similar frequency in t(14;18)-negative follicular lymphoma and diffuse large B-cell lymphoma. However, t(14;18)-negative follicular lymphoma appears to have a higher frequency of rearrangement at the ABR compared with t(14;18)-positive follicular lymphoma and diffuse large B-cell lymphoma. Therefore, it is important to perform FISH analysis with ABR to determine possible involvement of BCL6 rearrangement in follicular lymphoma, especially in t(14;18)-negative cases.

The nuclear immunoreactivity for thyroid transcription factor-1 (TTF-1) is a useful marker for identification of carcinomas of thyroid and lung origin. Our aim was to determine whether cytoplasmic staining in the liver is a result of cross-reaction of anti-TTF-1 antibody (clone 8G7G3/1, DAKO, Carpinteria, CA) or true positivity resulting from aberrant expression of TTF-1 or products of the alternatively sliced TTF-1 gene. Fresh tissue samples from liver, thyroid, and lung were obtained for HE-stained sections, TTF-1 immunostaining, and RNA and protein analyses. Western blot revealed an abundant band corresponding to an approximately 160-kd protein from liver but not either thyroid or lung tissue samples. By reverse transcriptase-polymerase chain reaction, messenger RNA of TTF-1 was not detectable in liver tissue. Our study demonstrates that TTF-1 immunoreactivity (clone 8G7G3/1) in the hepatocyte cytoplasm is due to an approximately 160-kd protein; this unique protein is not an alternative splicing product of TTF-1 and neither is it expressed in thyroid and lung tissues.

Treatment with BMP-7 causes a shift in the differentiation pathway from myoblastic to osteoblastic in C2C12 mouse myoblast precursor cells in vitro. The underlying molecular mechanism is largely unknown. BMP-7 at 200 ng/ml completely inhibited myotube formation in C2C12 cells and dramatically induced alkaline phosphatase activity up to 20-fold when compared to untreated cells by day 12 in culture. The level of Runx2/Cbfa1 mRNA, a bone-specific transcription factor, was also stimulated up to 6-fold by BMP-7 with a peak at 24 h. In addition BMP-7 treatment stimulated a 55-fold increase in osteocalcin mRNA as early as 24 h after treatment. A novel finding was that the expression of the chondrocyte markers Sox9 and type II collagen was increased as well. Runx2/Cbfa1 is a molecular switch for osteoblast differentiation. To initiate the study of modulators of Runx2/Cbfa1, such as kinases and cofactors, during osteoblastic differentiation of C2C12 cells treated by BMP-7 in vitro, microarray analyses of gene expressions were performed. Microarray data suggested that a total of 882 transcripts were either up- or downregulated at least 2-fold. Cluster analyses revealed 76 genes (including ESTs) with expression patterns that paralleled Runx2/Cbfa1. Thirteen of these 76 genes were initially selected as potential transcription modulators for further study; including CCAAT/enhancer binding protein delta, distal- less homeobox 1, forkhead box F2, insulin-like growth factor binding protein 4, an ortholog of human osteoclast stimulating factor 1 and p300/CBP-associated factor. Some transcription modulators have been associated with osteoblastic differentiation or interacted with Runx2/Cbfa1. Most of them have not been extensively studied in osteoblastic differentiation and in relationship to Runx2/Cbfa1. Thus, these studies identify potential regulators for Runx2/Cbfa1 and osteoblast differentiation. In addition, our data revealed for the first time that BMP-7 not only induced the expression of osteoblastic differentiation markers but also stimulated the expression of chondroblastic markers in C2C12 cells.

Strategies for localized skeletal regeneration using either cell or ex vivo gene transfer typically utilize preosteoblasts from bone or bone marrow biopsies. Our studies have demonstrated than ex vivo bone morphogenetic proteins (BMP) transduced fibroblasts convert to osteoblasts and form bone in vivo [1, 2]. In addition when suspended in a variety of thermoset hydrogels, these cells are capable of ectopic or orthotopic bone formation. To study the mechanism of the phenotypic conversion of BMP transduced fibroblasts, we have initiated characterization of three-dimensional (3D) cultures derived from these cell/collagen hydrogel composites. These data reveal that the BMP, but not control transduced fibroblasts, secrete BMP-7 and acquire some chondroblastic traits in 3D cultures. These traits include altered cell shape and changes in the extracellular matrix such as accumulation of cartilage proteoglycan, type II collagen, and mineral deposition by 3 weeks. These studies suggest that this culture system may ...

Experimental cell or ex vivo gene therapy for localized bone formation typically uses osteoprogenitor cells propagated from periosteum or bone marrow. Both require bone or marrow biopsies to obtain cells. We have demonstrated that implantation of gingival or dermal fibroblasts transduced with BMP ex vivo, using a recombinant adenovirus (AdCMVBMP) attached to porous biodegradable scaffolds, form bone in vivo. Here we show that BMP-7-transduced fibroblasts suspended in injectable thermoset hydrogels form complete ossicles on subcutaneous injection and repair segmental defects in rat femurs. Bone formation was preceded by an intermediate cartilage stage. To determine the fate of the implanted transduced cells, thermoset hydrogel suspensions of ex vivo BMP-7-transduced or nontransduced fibroblasts were placed in diffusion chambers and implanted to allow development in vivo without direct contact with host cells. Only the BMP-transduced fibroblasts formed bone within the diffusion chambers in vivo, revealing that BMP transduction induces osteoblastic conversion of these cells.

Recombinant human BMP-7 (bone morphogenetic protein-7, osteogenic protein-1) is osteogenic, dentinogenic and cementogenic when implanted into the appropriate tissue in vivo. However, most studies characterizing the induction of these tissues have implanted BMP-7 into freshly surgerized, clinically healthy tissues. To determine if BMP-7 is dentinogenic in inflamed dental pulps, we applied BMP-7 to inflamed ferret pulps. A single application of 5 microg of a commercial preparation of lipopolysaccharide (LPS) from Salmonella typhimurium directly to the coronal pulp induced a reversible mixed inflammatory exudate of moderate intensity within 3 d. Treatment with a single application of 2.5, 7.5 or 25 microg recombinant human BMP-7/mg collagen (2 mg total mass/tooth) induced reparative dentinogenesis in controls but not LPS treated dental pulps. These data reveal that a single application of up to 50 microg/tooth of exogenous recombinant BMP-7 is insufficient to induce reparative dentinogenesis in ferret teeth with reversible pulpitis. Given that pulp cells in the inflamed tissues likely retain the capacity to respond to exogenous BMP-7, it is possible that insufficient active recombinant protein is available to induce tissue formation in experimentally inflamed dental pulps.

An ex vivo gene therapy strategy was used to achieve localized skeletal regeneration in vivo. When an adenovirus vector engineered to express bone morphogenetic protein 7 transduced human gingival fibroblasts or rat dermal fibroblasts, these nonosteogenic tissues formed bone and supported the development of hematopoietic tissue when transplanted into immunocompromised mice. Transduced gingival fibroblasts formed marrow-containing ossicles in 100% of transplants after 1-2 weeks in vivo (n = 30). Immunostaining with murine and human-specific antisera raised against osteonectin and in situ hybridization of human-specific Alu genomic sequence demonstrated that the newly formed bone organ was a chimera of both the human donor and the mouse recipient cells. In experiments of greater clinical relevance, AdCMVBMP-7-transduced dermal fibroblasts repaired critical size skeletal defects in rat calvariae (n = 6). The results of this study suggest a bifunctional role of BMP-7-transduced fibroblasts. The transduced, nonosteogenic cells not only secreted biologically active BMP-7 in vitro and in vivo, but also differentiated into bone-forming cells in vivo. This model exploits the use of an easily biopsied, self-regenerating tissue such as gingiva or skin and suggests that local regeneration of tissues by ex vivo gene therapy may not require that autogenous cells be cultured from the tissue that is to be regenerated.

Dentin sialoprotein (DSP) and dentin phosphoprotein (DPP; phosphophoryn) are two principal dentin-specific non-collagenous proteins. DPP is extremely acidic and is rich in aspartic acid and serine. By virtue of this structure, DPP may bind large amounts of calcium and may facilitate initial mineralization of dentin matrix collagen as well as regulate the size and shape of the crystals. The function of DSP is not known. DSP and DPP are encoded by a single gene in both rat and mouse, and are uniquely expressed in odontoblasts and transiently in pre-ameloblasts. Because DSP and DPP are isolated from dentin as distinct proteins and appear to be present in different amounts, the nascent dentin sialophosphoprotein (DSPP) is likely cleaved to yield DSP and DPP. However, when, where and how the DSPP is cleaved into DSP and DPP is not clear. To further elucidate the structure and function of human DSP and DPP, we have cloned DPP and DSP cDNA by reverse transcriptase-polymerase chain reaction (RT-PCR) strategies, and then cloned and initiated characterization of a human dentin sialophosphoprotein gene. The genomic organization of human DSPP is very similar to that of mouse, containing five exons and four introns, suggesting it is a homologue of mouse dentin sialophosphoprotein (DSPP). Exons 1-4 encode for DSP, while exon 5 encodes for the C-terminus of DSP and the whole DPP. A 4.6-kb RNA transcript was detected on Northern blot analyses of total RNA extracted from immature (open root apices) human teeth using either a human DPP or DSP probe.

Autologous tissue grafting for the restoration of oral tissues is limited by several factors, including the availability of sufficient donor tissue. One solution to this problem may be to develop substitute tissue grafts by attaching disaggregated autologous cells propagated in vitro to scaffolds composed of natural or synthetic polymers. We have earlier demonstrated that human dental pulp and gingival fibroblasts (HPF, HGF) adhere to non-woven polyglycolic acid (PGA) scaffolds, proliferate and produce extracellular matrix in vitro. We now report that such HPF and HGF adhered to PGA scaffolds survive when implanted into subcutaneous sites in immuno-compromised mice. The transplanted cells synthesize and secrete type I collagen, cellular fibronectin and may express genes implicated in transducing bone morphogenetic protein (BMP) signals. Messenger RNA for BMP-2, -4, -7 (OP-1), the BMP type I receptors Act RI, BMPR-1A and 1B, the type II receptor BMPR-II, and type I collagen were detected by reverse transcription-polymerase chain reaction (RT-PCR). These data revealed that three adult human dental pulp and gingival cell populations, each from individual donors, attached to PGA scaffolds and cultured for 24 h in vitro, survive implantation and express genes indicative of a capacity to produce extracellular matrix. The implanted cells may also express genes associated with responsiveness to BMP-mediated tissue inductive signals.

Gu K,* Chang SR,* Slaven MS, Clarkson BH, Rutherford RB, Ritchie HH:Human dentin phosphophoryn nucleotide and amino acid sequence. Eur J OralSci 1998; 106: 1043–1047. # Eur J Oral Sci, 1998Dentin sialoprotein (DSP) and phosphophoryns (DPP) are major dentin-specificnon-collagenous proteins and are synthesized by odontoblasts. DPP areextremely acidic, rich in aspartic acid and serine, possess a high affinity forcalcium and collagen, and are believed to function in dentin mineralization.Whereas DSP and DPP are the products of a single gene in mouse and rat, ananalogous human gene has not been described. Using RT-PCR based cloningstrategies, we have cloned human DPP cDNA from immature molar root totalRNA. The open reading frame of this human DPP cDNA comprises 2364bpencoding 788 amino acids rich in serine (58%), aspartic acid (26%) andasparagine (9%). These are mostly arranged as (DSS)

Mesenteric lipodystrophy is a rare condition characterized by tumor-like expansion of the mesocolon. The etiology remains obscure, but autoimmunity, abdominal trauma, abdominal surgery, and ischemic injury have all been postulated. To our knowledge, there have been no previous reports of synchronous mesenteric lipodystrophy and angiodysplasia. Whether these are independent or associate entities remains unknown. We present the clinical, radiological, and pathological findings of such a case.

Summary Osteopontin ( SPP1 ) is reportedly the most up-regulated gene in primary central nervous system lymphoma (PCNSL). Our objective was to confirm immunoexpression of osteopontin and determine if CD44v6 and CD44H played a significant role as receptors for osteopontin in PCNSL. Twenty PCNSL, 12 nodal diffuse large B-cell lymphoma (N-DLBCL), and 17 extra-nodal DLBCL (EN-DLBCL) archival pathology cases were examined. Osteopontin nuclear positivity was observed in 20 (100%) of 20 PCNSL cases, 16 (95 %) of 17 EN-DLBCL, and 3 of 12 (25%) N-DLBCL. The immunohistochemical score of osteopontin in PCNSL (7.0 ± 3.5) and EN-DLBCL (4.4 ± 4.1) was significantly higher than N-DLBCL (0.3 ± 0.6). Sixteen cases were positive for CD44v6 (33%), including 6 PCNSL, and 5 each EN-DLBCL and N-DLBCL; no statistical difference was observed. CD44H was positive in all cases except one PCNSL but without any significant differences across the 3 groups. CD44H expression was significantly higher in non–germinal center B-cell (GCB) (score 12 ± 1.5) as compared to the GCB group (9.5 ± 3.1), and in non-GCB PCNSL (7.9 ± 4.2) as compared to non-GCB non-CNS lymphoma (2.8 ± 4.0) ( P = .009); the differences were insignificant for osteopontin and CD44v6. Neither CD44H nor CD44v6 scores correlated with the osteopontin expression score or Ki-67 index. Osteopontin immunoexpression was highest in PCNSL, suggesting its probable role in its pathogenesis. However, its lack of correlation with CD44v6 excludes the latter as the likely osteopontin receptor in PCNSL. The significantly higher CD44H expression in the non-GCB than GCB group may contribute to the aggressiveness of the non-GCB DLBCL. Further studies are needed to elucidate the pathway and the prognostic/predictive role of osteopontin in PCNSL.

Diffuse large B-cell lymphoma (DLBCL) includes two prognostically-important subtypes, the germinal center B-cell (GCB) and the non-GCB types. The aim of this study was to evaluate immunohistochemical approaches for predicting the survival of patients with DLBCL following autologous hematopoietic stem cell transplantation (AHSCT). We identified 62 patients with DLBCL who either had an initial complete remission (17 patients) or received salvage chemotherapy for relapsed or refractory disease (45 patients), followed by AHSCT. Tissue microarrays were immunostained with monoclonal antibodies against GCET1, CD10, BCL6, MUM1, FOXP1 and LMO2. Using the Hans algorithm, we classified 50% of the cases as GCB type, whereas the Choi algorithm classified 58% as GCB type and LMO2 was positive in 69%. However, no significant differences were found in the five-year overall or event-free survivals using any of these approaches. In conclusion, cell-of-origin fails to predict survival of DLBCL patients treated with AHSCT.

Low-grade follicular lymphoma (LGFL) is known to have a good prognosis. However, the clinical relevance of the presence and extent of diffuse growth areas within LGFL is not clear. Therefore, we studied 457 patients with nodal LGFL seen over a 20-year period (1982-2002). Relevant clinical and survival data were obtained and the pathology specimens were subclassified into predominantly follicular LGFL (or=80% follicular areas), significantly follicular LGFL (30-70% follicular areas), significantly diffuse LGFL (30% follicular areas) and pure diffuse follicle centre lymphoma (DFCL). The majority of the patients were treated with anthracycline-based regimens. Patients with DFCL were slightly older (median age, 67 years), more likely to have bulky disease (5 cm), and often had suboptimal biopsies. However, no association was observed between the extent of diffuse areas and survival across the four subgroups. We conclude that the subclassification of LGFL based on the extent of diffuse areas does not have clinical relevance.

The t(14;18)(q32;q21) translocation is the genetic hallmark of follicular lymphoma. Detection of this translocation can facilitate the diagnosis of follicular lymphoma and can be used to monitor response to therapy and level of residual disease. We herein review and compare practical techniques for detecting t(14;18)(q32;q21), including conventional cytogenetics, fluorescence in situ hybridization, Southern blot analysis, and polymerase chain reaction–based assay. Emphasis is placed on fluorescence in situ hybridization and polymerase chain reaction–based assay, given the applicability of these techniques to fixed, paraffin-embedded tissue.

Mesenteric lipodystrophy is a rare condition characterized by tumor-like expansion of the mesocolon. The etiology remains obscure, but autoimmunity, abdominal trauma, abdominal surgery, and ischemic injury have all been postulated. To our knowledge, there have been no previous reports of synchronous mesenteric lipodystrophy and angiodysplasia. Whether these are independent or associate entities remains unknown. We present the clinical, radiological, and pathological findings of such a case.

Advanced breast cancers frequently metastasize to bone, resulting in osteolytic lesions, yet the underlying mechanisms are poorly understood. Here we report that nuclear factor-kappaB (NF-kappaB) plays a crucial role in the osteolytic bone metastasis of breast cancer by stimulating osteoclastogenesis. Using an in vivo bone metastasis model, we found that constitutive NF-kappaB activity in breast cancer cells is crucial for the bone resorption characteristic of osteolytic bone metastasis. We identified the gene encoding granulocyte macrophage-colony stimulating factor (GM-CSF) as a key target of NF-kappaB and found that it mediates osteolytic bone metastasis of breast cancer by stimulating osteoclast development. Moreover, we observed that the expression of GM-CSF correlated with NF-kappaB activation in bone-metastatic tumor tissues from individuals with breast cancer. These results uncover a new and specific role of NF-kappaB in osteolytic bone metastasis through GM-CSF induction, suggesting that NF-kappaB is a potential target for the treatment of breast cancer and the prevention of skeletal metastasis.

Strategies for localized skeletal regeneration using either cell or ex vivo gene transfer typically utilize preosteoblasts from bone or bone marrow biopsies. Our studies have demonstrated than ex vivo bone morphogenetic proteins (BMP) transduced fibroblasts convert to osteoblasts and form bone in vivo. In addition when suspended in a variety of thermoset hydrogels, these cells are capable of ectopic or orthotopic bone formation. To study the mechanism of the phenotypic conversion of BMP transduced fibroblasts, we have initiated characterization of three-dimensional (3D) cultures derived from these cell/collagen hydrogel composites. These data reveal that the BMP, but not control transduced fibroblasts, secrete BMP-7 and acquire some chondroblastic traits in 3D cultures. These traits include altered cell shape and changes in the extracellular matrix such as accumulation of cartilage proteoglycan, type II collagen, and mineral deposition by 3 weeks. These studies suggest that this culture system may be useful for elucidating early mechanistic events in the BMP-induced conversion of fibroblasts to osteoblasts.

Bone morphogenetic proteins (BMP) have been shown to induce reparative dentine formation experimentally but the cells responsible, which respond to BMPs, have not been identified. The BMP signal is probably mediated by interaction of type I and II BMP receptors (R). Here, the RNA of human adult dental pulp and pulp cells in culture was examined by reverse transcription (RT) polymerase chain reaction (PCR) for evidence of mRNA for BMPs. mRNAs for BMP-2, -4, osteogenic protein-1, ActR-l (activin-like kinase receptor), BMPR-IA, -IB and -II were detected by RT-PCR. The 698-bp PCR fragment for BMPR-IB was used to probe pulp cells for expression of that receptor. Cell expression of BMPR-IB was detected by the hybridization probe. The findings suggest that resident pulp cells may be able to respond to BMPs to initiate tissue formation.