Interaction between the TP63 and SHH pathways is an important determinant of epidermal homeostasis.

Deregulation of the hedgehog (HH) pathway results in overexpression of the GLI target BCL2 and is an initiating event in specific tumor types including basal cell carcinoma of the skin. Regulation of the HH pathway during keratinocyte differentiation is not well understood. We measured HH pathway activity in response to differentiation stimuli in keratinocytes. An upregulation of suppressor of fused (SUFU), a negative regulator of the HH pathway, lowered HH pathway activity and was accompanied by loss of BCL2 expression associated with keratinocyte differentiation. We used in vitro and in vivo models to demonstrate that ΔNp63α, a crucial regulator of epidermal development, activates SUFU transcription in keratinocytes. Increasing SUFU protein levels inhibited GLI-mediated gene activation in suprabasal keratinocytes and promoted differentiation. Loss of SUFU expression caused deregulation of keratinocyte differentiation and BCL2 overexpression. Using in vivo murine models, we also provide evidence of GLI-mediated regulation of the TP63 pathway. p63 expression appears essential to establish an optimally functioning HH pathway. These observations present a regulatory mechanism by which SUFU acts as an interacting node between the HH and TP63 pathways to mediate differentiation and maintain epidermal homeostasis. Disruption of this regulatory node can be an important contributor to multistep carcinogenesis.

Distinct roles of first exon variants of the tumor-suppressor Patched1 in Hedgehog signaling.

Patched1 (PTCH1) is one of the key molecules involved in the Hedgehog (HH) signaling pathway and acts as the receptor of HH ligands. Additionally, PTCH1 inhibits the positive signal transductor Smoothened (SMO). Several PTCH1 splice variants are known but the functional differences among them are not clear. Here, we demonstrate the unique biological properties of the PTCH1 isoforms generated by alternative first exon usage. All isoforms examined worked as functional receptors of both Sonic HH and Desert HH. However, the signaling upregulated isoforms PTCH1-1B and -1C inhibited SMO and the pathway transcription factors glioma 1 (GLI1) and GLI2 to a higher extent than PTCH1-1 and -1Ckid. Moreover, in situ hybridizations allowed the detection of the Ptch1 isoforms in specific structures of the developing mouse embryo. Additionally, the differences in the N-terminal tail had a dramatic influence on the steady states of the proteins, with PTCH1-1B and -1C levels being significantly higher than PTCH1-1 and -1Ckid. This implies that the pronounced signaling inhibitory properties of PTCH1-1B and -1C may be mostly due to this high-protein expression rather than to intrinsic functional differences. Thus, our study supports a role of splicing variation and promoter choice for HH signaling regulation.

Genomic organization and embryonic expression of Suppressor of Fused, a candidate gene for the split-hand/split-foot malformation type 3.

The genes for human and mouse Suppressor of Fused (SU(FU)/Su(Fu)) in the Hedgehog signaling pathway were characterized and found to contain 12 exons. Human SU(FU) localized on chromosome 10q24-25 between the markers D10S192 and AFM183XB12. We detected three additional SU(FU) isoforms, two of which have lost their ability to interact with the transcription factor GLI1. Expression analysis using whole mount in situ hybridization revealed strong expression of Su(Fu) in various mouse embryonic tissues. SU(FU) was considered a candidate gene for the split-hand/split-foot malformation type 3 (SHFM3). However, no alterations in the SU(FU) gene were found in SHFM3 patients.

Evolution of the carcinoembryonic antigen family. structures of CGM9, CGM11 and pregnancy-specific glycoprotein promoters.

Earlier studies have demonstrated that the genes of the human carcinoembryonic antigen (CEA) family can be divided into three subgroups, the CEA subgroup (n = 12), the pregnancy-specific glycoprotein (PSG) subgroup (n = 11), and a third subgroup (n = 6). To further characterize the CEA gene family, we have determined the genomic structures of CGM9 and CGM11, analyzed the promoter regions of all eleven PSGs, studied the CGM15-PSG13 intergenic region and the evolutionary relationships beween the CEA family genes. CGM9, a typical CEA subgroup member, was a pseudogene with the exon structure [5'UTR-L-L/N-TM-Cyt-3'UTR]. CGM11 contained a mixture of exons derived from CEA and PSG subgroup genes. The formula of the CGM11 pseudogene was [5'UTR-L-L/N-C-3'UTR]. Thus both genes lacked the IgC2-like domains typically found in CEA subfamily members. The upstream promoter regions of all eleven PSGs were characterized. All PSG promoters lacked the classical TATA and CCAAT elements, but had putative PEA3 box(es), CACCC box(es), a RARE box, and poly (dG-dT) repeats of different lengths. Five PSGs also had an SP1 site. The complete 10-kb intergenic region between CGM15 and PSG13 was sequenced. Clusters of different types of repetitive sequences were seen. The time of divergence of the CEA and PSG subfamilies was estimated to be 107.7 +/- 17.1 million years, or at about the time of human-rodent divergence. Models for the evolution of CEA and PSG and the third family subgroup genes are proposed.

A small amphipathic alpha-helical region is required for transcriptional activities and proteasome-dependent turnover of the tyrosine-phosphorylated Stat5.

Cytokines induce the tyrosine phosphorylation and associated activation of signal transducers and activators of transcription (Stat). The mechanisms by which this response is terminated are largely unknown. Among a variety of inhibitors examined, the proteasome inhibitors MG132 and lactacystin affected Stat4, Stat5 and Stat6 turnover by significantly stabilizing the tyrosine-phosphorylated form. However, these proteasome inhibitors did not affect downregulation of the tyrosine-phosphorylated Stat1, Stat2 and Stat3. With Stat5 isoforms, we have observed that tyrosine-phosphorylated carboxyl-truncated forms of Stat5 proteins were considerably more stable than phosphorylated wild-type forms of the protein. Also, the C-terminal region of Stat5 could confer proteasome-dependent downregulation to Stat1. With a series of C-terminal deletion mutants, we have defined a relatively small, potentially amphipathic alpha-helical region that is required for the rapid turnover of the phosphorylated Stat5 proteins. The region is also required for transcriptional activation, suggesting that the functions are linked. The results are consistent with a model in which the transcriptional activation domain of activated Stat5 is required for its transcriptional activity and downregulation through a proteasome-dependent pathway.

Stat5 is required for IL-2-induced cell cycle progression of peripheral T cells.

Many cytokines activate two highly homologous Stat proteins, 5a and 5b. Mice deficient in both genes lack all growth hormone and prolactin functions but retain functions associated with cytokines such as erythropoietin. Here, we demonstrate that, while lymphoid development is normal, Stat5a/b mutant peripheral T cells are profoundly deficient in proliferation and fail to undergo cell cycle progression or to express genes controlling cell cycle progression. In addition, the mice lack NK cells, develop splenomegaly, and have T cells with an activated phenotype, phenotypes seen in IL-2 receptor beta chain-deficient mice. These phenotypes are not seen in mice lacking Stat5a or Stat5b alone. The results demonstrate that the Stat5 proteins, redundantly, are essential mediators of IL-2 signaling in T cells.

Stat5a and Stat5b proteins have essential and nonessential, or redundant, roles in cytokine responses.

A variety of cytokines mediate the activation of Janus protein tyrosine kinases (Jaks). The Jaks then phosphorylate cellular substrates, including members of the signal transducers and activators of transcription (Stat) family of transcription factors. Among the Stats, the two highly related proteins, Stat5a and Stat5b, are activated by a variety of cytokines. To assess the role of the Stat5 proteins, mutant mice were derived that have the genes deleted individually or together. The phenotypes of the mice demonstrate an essential, and often redundant, role for the two Stat5 proteins in a spectrum of physiological responses associated with growth hormone and prolactin. Conversely, the responses to a variety of cytokines that activate the Stat5 proteins, including erythropoietin, are largely unaffected.

Jak2 is essential for signaling through a variety of cytokine receptors.

A variety of cytokines activate receptor-associated members of the Janus family of protein tyrosine kinases (Jaks). To assess the role of Jak2, we have derived Jak2-deficient mice. The mutation causes an embryonic lethality due to the absence of definitive erythropoiesis. Fetal liver myeloid progenitors, although present based on the expression of lineage specific markers, fail to respond to erythropoietin, thrombopoietin, interleukin-3 (IL-3), or granulocyte/macrophage colony-stimulating factor. In contrast, the response to granulocyte specific colony-stimulating factor is unaffected. Jak2-deficient fibroblasts failed to respond to interferon gamma (IFNgamma), although the responses to IFNalpha/beta and IL-6 were unaffected. Lastly, reconstitution experiments demonstrate that Jak2 is not required for the generation of lymphoid progenitors, their amplification, or functional differentiation. Therefore, Jak2 plays a critical, nonredundant role in the function of a specific group of cytokines receptors.

Signaling by the cytokine receptor superfamily.

A variety of cytokines that regulate functions of multiple lineages share the utilization of receptors that are structurally and functionally related and are referred to as the cytokine receptor superfamily. These receptors associate with one or more of the four mammalian Janus kinases (Jaks) and ligand-induced receptor aggregation results in their activation. Critical roles for Jak3 and Jak2 are demonstrated by the phenotypes of mice that lack each gene. Among the substrates of the Jaks are one or more of the seven members of the signal transducers and activators of transcription (Stats). Each Stat family member plays a critical role in the biological functions of specific cytokines as demonstrated by the phenotype of mice lacking one or more of these genes.

Intra-and inter-individual heterogeneity in exon 2 of the MDR1 gene in primary breast carcinoma and healthy individuals.

Increased expression of P-glycoprotein, encoded by the MDR1 gene, is considered to be responsible for chemotherapy failure in a number of human cancers. Although it is clear that mutations in the MDR1 gene affect substrate specificity of the transporter in multidrug-resistant cell lines, scant interest has been directed at whether mutations have a unique clinical presentation. To address this question, we studied exon 2 of the MDR1 gene in 9 patients with primary breast carcinoma and 9 healthy controls using PCR and DNA sequence analysis. In order to reduce the possibility of nucleotide misincorporations introduced by Tag polymerase, sequencing of six subclones of each DNA specimen was performed. A mutation was seen as a substitution from G to A at position -1 in two patients and one control. An A to G nucleotide substitution giving rise to an amino acid substitution (Asn-->Asp) in codon 21 at the first potential N-glycosylation site of the P-glycoprotein was seen in primary tumors from four patients and in an axillar lymph node metastases from one of these patients. This mutation was also seen in two healthy individuals, which similar to the patients, both seem to be heterozygous for this MDR1 exon 2 allele. Three other mutations were also found in the patients; a substitution of A to G at position 23 and A to G at position 52 in the same patient and in another patient, G at position 42 was changed to A. However, the last three mutations were not confirmed by repeating analysis of the original genomic sample. The results revealed different distribution of a point mutation between various parts of the same primary tumor and between a lymph node metastasis and the primary tumor tissue. Thus, demonstrating both intra-and inter-tumor heterogeneity. The results also emphasized constitutional allelic variation in the MDR1 gene. Whether this might affect sensitivity to chemotherapy has to be further evaluated.

Naturally occurring dominant negative variants of Stat5.

Stat5 was initially identified as a prolactin-induced member of the signal transducer and activator of transcription (Stat) family in sheep. However, Stat5 is also activated in the response to a variety of cytokines. In mice, and possibly in other species, there exist two Stat5 genes (Stat5a and Stat5b) that encode proteins of 92 and 94 kDa that are 95% identical. In the studies described here, we demonstrate that naturally occurring carboxyl-truncated, variant Stat5 proteins of 77 and 80 kDa exist and that these proteins are inducibly tyrosine phosphorylated in the response to several cytokines and form heterodimers with the full-length, wild-type proteins. Using expression constructs encoding truncated forms, we demonstrate that the truncated forms can be tyrosine phosphorylated and bind DNA. Surprisingly, the tyrosine phosphorylation of the carboxyl-truncated forms is considerably more stable than that of the wild-type proteins. Overexpression of a carboxyl-truncated Stat5a in cells resulted in the specific inhibition of the transcriptional activation by interleukin-3 of the genes for oncostatin M (Osm) and the cytokine-inducible, SH2 domain-containing gene (Cis), both of which have been shown to be normally regulated by Stat5. Although Stat5 dominantly suppressed the induction of these genes, no effects on cell proliferation were observed. Together, the results demonstrate the natural existence of potentially dominantly suppressive variants of Stat5 and implicate the carboxyl domain of Stats in transcriptional regulation and functions related to dephosphorylation.

Characterization of cDNA encoding novel pregnancy-specific glycoprotein variants.

The human pregnancy-specific glycoprotein (PSG) family consists of eleven closely related molecules mainly synthesized by placental syncytiotrophoblasts and whose function(s) are unknown. They belong to the carcinoembryonic antigen (CEA) family. As a step toward understanding PSG function, we have analysed 84 PSG cDNA clones from a fetal liver library with respect to domain arrangement and PSG identity. Four novel PSG cDNAs derived from the PSG4, PSG7, PSG11, and PSG13 genes were characterized. The PSG11 and PSG13 cDNAs had novel domain arrangements: L-N-B2-C (named type III) and L-A1-B2-C (named type IV), respectively. These splice variants were also demonstrated in placenta. PSG4 cDNA had a type IIa (L-N-A1-B2-C) and PSG7 cDNA a type I (L-N-A1-A2-B2-C) domain arrangement. PSG1, PSG4, PSG5 were found at highest frequency while PSG8 and PSG12 cDNA clones were not detected.

Gene organization of the pregnancy-specific glycoprotein region on human chromosome 19: assembly and analysis of a 700-kb cosmid contig spanning the region.

The pregnancy-specific glycoprotein (PSG) gene family consists of 11 closely related genes that form a subgroup of the carcinoembryonic antigen (CEA) gene family on 19q13.2. Using a high-resolution restriction fragment fingerprinting technique, we have assembled 256 cosmids from the PSG region into a single 700-kb contig. Fluorescence in situ hybridization to sperm pronuclei and cosmid walking experiments indicated that this PSG contig was directly telomeric of CGM8 at the telomeric end of the CEA subgroup gene cluster. Detailed restriction mapping and hybridization with gene-specific probes indicated that the order of the 11 Previously identified PSG genes is cen-PSG3-PSG8-PSG12-PSG1-PSG6-PSG7-+ ++PSG13-PSG2-PSG5-PSG4-PSG11-tel. The CEA subgroup gene CGM11 is located at the telomeric end of the PSG gene cluster. The PSG genes are all oriented in tandem with the 5'-3' direction of transcription from telomere to centromere. The detailed map also led to the identification of seven new CEA family genes in this region. One of these (CGM12), located between CGM8 and PSG3, is a member of the CEA subgroup. The remaining six (CGM13 through CGM18) are interspersed among the PSG genes and appear to form a third distinct subgroup within the CEA gene family.

The pregnancy-specific glycoprotein (PSG) gene cluster on human chromosome 19: fine structure of the 11 PSG genes and identification of 6 new genes forming a third subgroup within the carcinoembryonic antigen (CEA) family.

The human pregnancy-specific glycoprotein (PSG) genes belong to the carcinoembryonic antigen (CEA) family, which in turn is a member of the immunoglobulin superfamily. We have analyzed a 700-kb cosmid contig spanning the PSG region on chromosome 19q13.2. The region contains 11 closely related PSG genes organized in tandem with a highly conserved structure and organization. Seven novel genes (CGM12 to CGM18) were found in the PSG region. CGM12 belongs to the CEA subgroup and appears to be a pseudogene. CGM13 to CGM18 forms a third new subgroup within the CEA gene family. The members of this new subgroup show 94-99% identity to each other but only 70-80% to other members of either the CEA or the PSG subgroups. They are composed of exons encoding two IgC-like domains and short hydrophilic carboxyl terminals similar to those of the PSGs. Unlike any of the known CEA family genes, however, they seem to lack the exon for an IgV-like N-terminal domain.

Cytokine profile and ultrastructure of intraepithelial gamma delta T cells in chronically inflamed human gingiva suggest a cytotoxic effector function.

We have shown that gamma delta T cells in human gingiva have an intraepithelial location and, that in the chronic inflammatory disease periodontitis, the expression of CD45RO and CD8 or CD4 is induced on gamma delta T cells. To study the role of gamma delta T cells in local antibacterial responses, we determined the cytokine profiles of isolated human gingival cells. Different T cell subpopulations, isolated by positive selection with mAb-coated magnetic beads and macrophages, as well as epithelial cells, were analyzed for expression of mRNA for 15 cytokines by reverse transcriptase-PCR. The ultrastructure of gingival gamma delta T cells was also studied. The gamma delta T cells expressed mRNA for IFN-gamma, TNF-alpha, TGF-beta 1, and IL-6. Expression of IFN-gamma was a consequence of inflammation. CD4+ gamma delta T cells expressed IFN-gamma only, whereas CD8+ gamma delta T cells expressed all four cytokines. CD8+ cells expressing IFN-gamma, TNF-alpha, and IL-6 in combination suggest a cytotoxic effector function. Gingival gamma delta T cells contained cytoplasmic electron-dense membrane-bound granules and multivesicular bodies that are ultrastructural characteristics of cytotoxic cells. Epithelial cells from inflamed gingiva expressed HLA-DR, CD1a, CD1c, and heat shock protein 60 on the cell surface. They also expressed mRNA for IL-1 beta, IL-6, IL-8, TNF-alpha, and TGF-beta 1. Thus, epithelial cells may function as accessory cells in immune activation and, at the same time, be target cells for CD8+ gamma delta T cells reactive with CD1 Ag or heat shock protein. These results suggest that gamma delta T cells constitute a first line of defense in gingiva, preventing entrance of pathogens by cytotoxicity against infected and stressed epithelial cells, and by control of epithelial cell growth through secretion of regulatory cytokines.

Identification of three new genes and estimation of the size of the carcinoembryonic antigen family.

Using carcinoembryonic antigen (CEA) subgroup-specific degenerate PCR primers, we have identified three new CEA gene family member L/N exons (CGM9, CGM10, and CGM11) and all previously reported L/N exons of the CEA subgroup (CEA, BGP, NCA, CGM1, CGM2, CGM6, CGM7, and CGM8). This suggests that the CEA subgroup contains 11 genes. CGM9, CGM10, and CGM11 seem to be pseudogenes. A deletion of an asparagine in CGM9 results in loss of a glycosylation site, which is conserved throughout the CEA gene family. We have previously suggested the number of genes in the pregnancy-specific glycoprotein (PSG) subgroup to be 11, which together with this study indicates that the CEA gene family contains 22 genes in all. Parsimony analysis of the CEA subgroup interrelationships suggests that CGM7 occupies the most primitive position within the CEA subgroup, being a sister group to the rest. CEA, BGP, NCA, and CGM1 form a fairly well-supported group within the CEA subgroup.

The pregnancy-specific glycoprotein family of the immunoglobulin superfamily: identification of new members and estimation of family size.

The members of the carcinoembryonic antigen (CEA)/pregnancy-specific glycoprotein (PSG) gene family have a characteristic N-terminal domain that is homologous to the immunoglobulin variable region. We have estimated the size of the PSG subfamily by identification of N-domain exons from isolated genomic clones and from total genomic DNA through PCR amplification and DNA sequence determination. The PSG subfamily contains at least 11 different genes. For 7 of these, two DNA sequences differing from each other in 1 to 4 nucleotides were detected. Most likely, they represent different alleles. They are PSG1, PSG2, PSG3, PSG4, PSG5, PSG6, PSG7, PSG8, PSG11, PSG12, and PSG13. Six of the N-domain sequences described here are new. All of the PSGs except PSG1, PSG4, and PSG8 contained the arginine-glycine-aspartic acid sequence at position 93-95 corresponding to the complementarity determining region 3 of immunoglobulin. Parsimony analysis of 24 CEA and PSG sequences using 12 members of the immunoglobulin gene superfamily as outgroups to root the family tree shows that the N-domain of the CEA group genes evolved in one major branch and the PSG group genes in the other.