Pancreatic adenocarcinoma (PAC) is the most common type of pancreatic cancer, which commonly has an unfavorable prognosis. The present study aimed to develop a novel prognostic prediction strategy for PAC patients. mRNA sequencing data of PAC (the training dataset) were extracted from The Cancer Genome Atlas database, and the validation datasets (GSE62452 and GSE79668) were acquired from the Gene Expression Omnibus database. The differentially expressed genes (DEGs) between good and poor prognosis groups were analyzed by limma package, and then prognosis‑associated genes were screened using Cox regression analysis. Subsequently, the risk score system was constructed and confirmed using Kaplan‑Meier (KM) survival analysis. After the survival associated‑clinical factors were screened using Cox regression analysis, they were performed with stratified analysis. Using DAVID tool, the DEGs correlated with risk scores were conducted with enrichment analysis. The results revealed that there were a total of 242 DEGs between the poor and good prognosis groups. Afterwards, a risk score system was constructed based on 6 prognosis‑associated genes (CXCL11, FSTL4, SEZ6L, SPRR1B, SSTR2 and TINAG), which was confirmed in both the training and validation datasets. Cox regression analysis showed that risk score, targeted molecular therapy, and new tumor (the new tumor event days after the initial treatment according to the TCGA database) were significantly related to clinical prognosis. Under the same clinical condition, 6 clinical factors (age, history of chronic pancreatitis, alcohol consumption, radiation therapy, targeted molecular therapy and new tumor (event days) had significant associations with clinical prognosis. Under the same risk condition, only targeted molecular therapy was significantly correlated with clinical prognosis. In conclusion, the 6‑gene risk score system may be a promising strategy for predicting the outcome of PAC patients.

Risk assessment and treatment choice remains a challenge in early non-small-cell lung cancer (NSCLC). The aim of this study was to identify novel genes involved in the risk of early relapse (ER) compared to no relapse (NR) in resected lung adenocarcinoma (AD) patients using a combination of high throughput technology and computational analysis. We identified 18 patients (n.13 NR and n.5 ER) with stage I AD. Frozen samples of patients in ER, NR and corresponding normal lung (NL) were subjected to Microarray technology and quantitative-PCR (Q-PCR). A gene network computational analysis was performed to select predictive genes. An independent set of 79 ADs stage I samples was used to validate selected genes by Q-PCR.From microarray analysis we selected 50 genes, using the fold change ratio of ER versus NR. They were validated both in pool and individually in patient samples (ER and NR) by Q-PCR. Fourteen increased and 25 decreased genes showed a concordance between two methods. They were used to perform a computational gene network analysis that identified 4 increased (HOXA10, CLCA2, AKR1B10, FABP3) and 6 decreased (SCGB1A1, PGC, TFF1, PSCA, SPRR1B and PRSS1) genes. Moreover, in an independent dataset of ADs samples, we showed that both high FABP3 expression and low SCGB1A1 expression was associated with a worse disease-free survival (DFS).Our results indicate that it is possible to define, through gene expression and computational analysis, a characteristic gene profiling of patients with an increased risk of relapse that may become a tool for patient selection for adjuvant therapy.

Cancer stem-like cells (CSCs)/cancer-initiating cells (CICs) are considered to be essential for tumor maintenance, recurrence and metastasis. Therefore, eradication of CSCs/CICs is essential to cure cancers. However, the molecular mechanisms of CSCs/CICs are still elusive. In this study, we investigated the molecular mechanism of the cell growth of oral CSCs/CICs. Oral CSCs/CICs were isolated as aldehyde dehydrogenase 1 bright (ALDH1(br)) cells by the ALDEFLUOR assay. Small proline-rich protein-1B (SPRR1B) gene was shown to be overexpressed in ALDH1(br) cells by a cDNA microarray and RT-PCR. SPRR1B was shown to have a role in cell growth and maintenance of ALDH1(br) cells by SPRR1B overexpression and knockdown experiments. To elucidate the molecular mechanism by which SPRR1B regulates cell growth, further cDNA microarray analysis was performed using SPRR1B-overexpressed cells and cells with SPRR1B knocked down by siRNA. Expression of the tumor suppressor gene Ras association domain family member 4 (RASSF4) was found to be suppressed in SPRR1B-overexpressed cells. On the other hand, the expression of RASSF4 was enhanced in cells in which SPRR1B expression was knocked down by SPRR1B-specific siRNA. RASSF4 has an RA (Ras association) domain, and we thus hypothesized that RASSF4 modulates the MAP kinase signal downstream of the Ras signal. MAP kinase signal was activated in SPRR1B-overexpressed cells, whereas the signal was suppressed in SPRR1B knocked down cells. Taken together, the results indicate that the expression of SPRR1B is upregulated in oral CSCs/CICs and that SPRR1B has a role in cell growth by suppression of RASSF4.

Tobacco smoking is the leading cause of lung cancer worldwide. Gene expression in surgically resected and microdissected samples of non-small-cell lung cancers (18 squamous cell carcinomas and nine adenocarcinomas), matched normal bronchial epithelium, and peripheral lung tissue from both smokers (n = 22) and non-smokers (n = 5) was studied using the Affymetrix U133A array. A subset of 15 differentially regulated genes was validated by real-time PCR or immunohistochemistry. Hierarchical cluster analysis clearly distinguished between benign and malignant tissue and between squamous cell carcinomas and adenocarcinomas. The bronchial epithelium and adenocarcinomas could be divided into the two subgroups of smokers and non-smokers. By comparison of the gene expression profiles in the bronchial epithelium of non-smokers, smokers, and matched cancer tissues, it was possible to identify a signature of 23 differentially expressed genes, which might reflect early cigarette smoke-induced and cancer-relevant molecular lesions in the central bronchial epithelium of smokers. Ten of these genes are involved in xenobiotic metabolism and redox stress (eg AKR1B10, AKR1C1, and MT1K). One gene is a tumour suppressor gene (HLF); two genes act as oncogenes (FGFR3 and LMO3); two genes are involved in matrix degradation (MMP12 and PTHLH); three genes are related to cell differentiation (SPRR1B, RTN1, and MUC7); and five genes have not been well characterized to date. By comparison of the tobacco-exposed peripheral alveolar lung tissue of smokers with non-smokers and with adenocarcinomas from smokers, it was possible to identify a signature of 27 other differentially expressed genes. These genes are involved in the metabolism of xenobiotics (eg GPX2 and FMO3) and may represent cigarette smoke-induced, cancer-related molecular targets that may be utilized to identify smokers with increased risk for lung cancer.

Overexpression of SPRR1B in bronchial epithelial cells is a marker for early metaplastic changes induced by various toxicants/carcinogens. Previously, we have shown that the transcriptional stimulation of SPRR1B expression by phorbol 12-myristate 13-acetate (PMA) is mainly mediated by a -150/-94 bp enhancer harboring two critical 12-O-tetradecanoylphorbol-13-acetate-responsive elements (TREs) and by Jun.Fra-1 dimers. Here, we show that a region between -54 and -39 bp containing an ETS-binding site (EBS) and a GC box is essential for both basal and PMA-inducible SPRR1B transcription. In vivo footprinting demonstrated binding of transcription factors to these elements. However, unlike enhancer TREs, exposure of cells to PMA did not significantly alter the footprinting pattern at these elements. Mutations that crippled both the EBS and GC box suppressed both basal and PMA-inducible SPRR1B transcription. Consistent with this, overexpression of EBS-binding proteins ESE-1 and ESE-3 significantly stimulated SPRR1B promoter activity. Furthermore, preceding SPRR1B transcription, PMA up-regulated mRNA expression of ETS family members such as ESE-1 and ESE-3. Although ESE-1 synergistically activated c-Jun- and PMA-enhanced SPRR1B transcription, coexpression of Sp1 and ESE-1 showed no synergistic or additive effect on promoter activity, indicating an obligatory role for AP-1 proteins in such regulation. In support of this notion, deletion or mutation of two functional TREs inhibited ESE-1- and Sp1-enhanced promoter activation. Thus, the interaction between ESE-1 and Sp1, and AP-1 proteins that bind to the proximal and distal promoter regions, respectively, play a critical role in the induction of squamous differentiation marker expression in bronchial epithelial cells.

Exposure of distal bronchiolar region to various toxicants and pollutants suppresses Clara cell differentiation marker expression and greatly enhances the induction of squamous cell differentiation (SCD). Here, we demonstrate for the first time phorbol 13-myristate 12-acetate (PMA)-inducible expression of SCD markers, SPRRs, in Clara-like H441 cells. The transcriptional stimulation of human SPRR1B expression is mainly mediated by a -150- to -84-bp region that harbors two critical activator protein (AP)-1 sites. In unstimulated cells, the -150- to -84-bp region is weakly bound by AP-1 proteins, mainly JunD and Fra1. However, PMA prominently induced the binding of JunB and Fra1. Consistent with this, overexpression of wild-type Jun proteins upregulated the SPRR1B promoter activity. Conversely, a c-jun mutant suppressed both basal and PMA-inducible reporter gene expression. Intriguingly, overexpression of fra2 suppressed PMA-inducible reporter activity, whereas fra1 significantly enhanced basal level activity, indicating an opposing role for these proteins in SPRR1B expression in a manner similar to that observed in proximal tracheobronchial epithelial cells (BEAS-2B clone S6). Interestingly, unlike in S6 cells, a catalytically inactive c-Jun NH(2)-terminal kinase (JNK) 1 mutant significantly reduced the PMA-inducible SPRR1B promoter activity in H441 cells. Thus either temporal expression and/or spatial activation of AP-1 proteins by JNK1 might contribute to the induction of SCD in Clara cells.

Anal cancer originates from a peculiar histological region and provides a useful model for investigating alterations in proliferation and/or differentiation of neoplastic keratinocytes. Epidermal differentiation complex (EDC) genes, which form one of the major gene clusters in the human genome, are involved in the terminal differentiation of epithelial cells and in many instances have been implicated in epithelial tumours. We constructed a DNA macroarray capable of characterising the expression profiles of the entire EDC gene complex in normal mucosa and anal cancer biopsies of seven unrelated patients. Brain tissue and cultured keratinocytes were used as controls. All anal cancer samples showed expression profiles in which none of the EDC genes was silent, as evaluated by phosphor-imager analysis. Variance analysis showed significantly lower expression of SPRR2 with respect to SPRR1 or SPRR3, and significantly higher expression of S100A8 than of other S100A subfamily members. At hierarchical clustering analysis, the four basaloid anal cancer cases conglomerated in the top five positions. The macroarray method used by us provides the first demonstration of the expression profile of the EDC gene family in anal cancer, and is capable of producing significant information on the subgrouping of epithelial tumours such as anal cancer.

The overexpression of SPRR1B in bronchial epithelium is a marker for early metaplastic changes and the loss of its expression is associated with an irreversible malignant transformation. In the present study, we have used a model system consisting of normal and malignant bronchial epithelial (BE) cells to elucidate the differential transcriptional control of SPRR1B. SPRR1B expression is either detectable or PMA (phorbol 13-myristate 12-acetate) -inducible in several malignant BE cells including squamous, adeno, small and large cell carcinomas. Loss of SPRR1B expression is correlated well with the lack of strong in vivo protein-DNA interactions at the -152 bp promoter, which contains two functional TRE sites. Even though the basal level AP-1 protein DNA binding pattern is different between normal and malignant cells, PMA significantly enhances Jun and Fos binding to the consensus TRE site in both cell types. Intriguingly, the composition of AP-1 protein binding to the -152 to -86 bp SPRR1B promoter is quite different. In untreated cells, SPRR1B promoter is predominantly occupied by JunD and Fra2. PMA significantly induced binding of JunB and Fra1 in normal cells, while JunB and Fra2 bound to TREs in the malignant cells. Overexpression of fra1 in malignant cells significantly enhanced SPRR1B promoter activity. In contrast, overexpression of fra2, but not fra1, strongly reduced both basal and PMA-inducible promoter activities in normal cells. Together, these results indicate that either temporal expression and/or differential activation of AP-1 proteins, especially Fra1 and Fra2, might contribute to the dysregulation of terminal differentiation marker, SPRR1B, expression in various BE cells.

Scirrhous gastric cancer is often accompanied by metastasis to the peritoneum and/or lymph nodes, resulting in the highest mortality rate among gastric cancers. Mechanisms involved in gastric cancer metastasis are not fully clarified because metastasis involves multiple steps and requires the accumulation of altered expression of many different genes. Thus, independent analysis of any single gene would be insufficient to understand all of the aspects of gastric cancer metastasis. In this study, we performed global analysis on differential gene expression of a scirrhous gastric cancer cell line (OCUM-2M) and its derivative sublines with high potential for metastasis to the peritoneal cavity (OCUM-2MD3) and lymph nodes (OCUM-2MLN) in a nude mice model. By applying a high-density oligonucleotide array method, expression of approximately 6800 genes was analyzed, and selected genes were confirmed by the Northern blot method. In our observations in OCUM-2MD3 cells, 12 genes were up-regulated, and 20 genes were down-regulated. In OCUM-2MLN cells, five genes were up-regulated, and five genes were down-regulated. The analysis revealed two functional gene clusters with altered expression: (a) down-regulation of a cluster of squamous cell differentiation marker genes such as small proline-rich proteins [SPRRs (SPRR1A, SPRR1B, and SPRR2A], annexin A1, epithelial membrane protein 1, cellular retinoic acid-binding protein 2, and mesothelin in OCUM-2MD3 cells; and (b) up-regulation of a cluster of antigen-presenting genes such as MHC class II (DP, DR, and DM) and invariant chain (II) in OCUM-2MLN cells through up-regulation of CIITA (MHC class II transactivator). We then analyzed six gastric cancer cell lines by Northern blot and observed preferential up-regulation of trefoil factor 1, alpha-1-antitrypsin, and galectin 4 and down-regulation of cytidine deaminase in cells prone to peritoneal dissemination. Genes highly correlated with invasion or peritoneal dissemination of gastric cancer, such as E-cadherin or integrin beta4, were down-regulated in both of the derivative cell lines analyzed in this study. This is the first demonstration of global gene expression analysis of gastric cancer cells with different metastatic potentials, and these results provide a new insight in the study of human gastric cancer metastasis.

Cancer may be understood as the net effect of differences in gene expression between normal and transformed cells. In a novel direct approach applying this principle, complete genes expressed at altered mRNA levels in malignant versus normal esophageal epithelium were identified and isolated from cDNA libraries. One clone was expressed in normal esophageal mucosae but absent in esophageal carcinomas. By in situ hybridization, Northern blotting, and immunohistochemistry, expression of this gene was restricted to normal esophageal mucosa; it is designated esophagin. Esophagin expression was greatest in the superficial, most mature layers of esophageal squamous mucosa and was restricted to this organ, being undetectable in other squamous epithelia. A genomic clone localized esophagin to chromosomal region 1q21-q22. The expressed protein contains multiple direct repeats of an 8-amino acid motif rich in proline, with significant homology to the cornifin, pig 20K, monkey MT5, and human small proline-rich genes spri and spril. Esophagin constitutes the newest and largest member of this small proline-rich gene family and is associated with differentiation and the benign phenotype of the human esophageal epithelial cell.

A final event in the terminal differentiation of stratified squamous epithelia is the formation of a cornified cell envelope, which is a complex of several proteins cross-linked together by transglutaminases. One set of proteins is the family of small proline rich (SPR) proteins. In human foreskin epidermal cell envelopes, SPRs serve as cross-bridging proteins among the more abundant loricrin. In order to study further their evolution and expression, we have isolated and sequenced cDNAs encoding two mouse SPR1 proteins, SPR1a and SPR1b Comparative sequence analysis showed the preservation of the overall structure of mammalian SPR1 proteins with highly conserved termini and a central peptide domain repeated 13 (SPE1a) or seven (SPR1b) times. Tissues obtained from mouse fetal, newborn, and adult skin were tested by Northern blot analyses, in situ hybridization and immunohistochemistry using an antibody raised to a synthetic peptide corresponding to the C terminus of the SPR1a protein. Skin expression was first detected in fetal periderm in anagen hair follicles of newborn and older mice, and in the thickened epidermis of the lip and footpad, but no signal was detected in interfollicular trunk epidermis. High levels of SPR1a expression were found in epithelia from the forestomach and penis, and in benign squamous papillomas. Other epithelia expressing SPR1a include the tongue, esophagus, and vagina. Whenever detected, SPR1a positive staining was present in the spinous and granular layers. In the forestomach and papillomas, the periphery of cells in the cornified layer was also stained. Our results suggest that SPR1a participates widely in the construction of cell envelopes in cornifying epithelia characterized by either increased thickness or a requirement for extreme flexibility. Based on its likely function as a cross-bridging protein in cell envelopes, we conclude that the mechanical attributes of cell envelopes may be determined in part by the SPR1 content, in accordance with the specific function of the epithelium.

The small proline-rich protein 1 (SPRR1) gene encodes a precursor of the keratinocyte cornified envelope. To understand SPRR1 regulation we investigated its expression and modulation in keratinocytes in vivo and in vitro. SPRR1 was strongly expressed in suprabasal layers of the epidermis in newborn skin but only weakly expressed in adult skin. Both in vivo and in vitro, SPRR1 was not expressed in undifferentiated cells of basal or squamous carcinomas. However, within the same tumors and in premalignant lesions of squamous cell origin, cells with histologic evidence of differentiation showed a relative increase in SPRR1 transcript level. Within 24 h physiologic doses of uv irradiation induced SPRR1 mRNA in vivo. To investigate the possibility that SPRR1 expression is regulated by uv-induced cytokines, keratinocytes were stimulated with interleukin-1 (IL-1) and interleukin-3 (IL-3). Both significantly induced SPRR1 mRNA, while TGF-beta, known to lower IL-1 receptor in keratinocytes, down-regulated it. Moreover, proximity to inflammatory cells in vivo was associated with SPRR1 induction in anaplastic tumor cells. Our data suggest that SPRR1 is induced early in differentiation of normal keratinocytes but is not expressed in anaplastic cells of keratinocyte origin. Further, its regulation in skin appears to be modulated at least in part through cytokine release.

The present study examines interferon-gamma (IFN gamma)-induced changes in the expression of immunomodulatory genes, proliferation-associated genes, and squamous-specific genes in primary cultures of human bronchial epithelial cells and fibroblasts. IFN gamma induced the expression of guanylate binding protein (GBP or p67) and the MHC class II antigen, HLADR alpha, in both epithelial cells and fibroblasts. In contrast, the expression of complement component C3 was induced in bronchial epithelial cells but not in fibroblasts. Similarly, IFN gamma induced growth arrest (EC50 approximately 50 U/ml) only in bronchial epithelial cells. This growth arrest was accompanied by a down-regulation of cdc2, E2F-1, and p53 mRNA levels and was associated with expression of the squamous-specific marker genes, transglutaminase type I and cornifin. These findings are consistent with IFN gamma inducing squamous differentiation in bronchial epithelial cells. In contrast, several lung carcinoma cell lines did not respond to IFN gamma with respect to the down-regulation of proliferation-associated genes or the induction of squamous-specific genes. However, GBP expression was induced in all the cell lines in response to IFN gamma. The present study demonstrates that cultured human bronchial epithelial cells are sensitive to the immunomodulatory, growth-inhibitory, and differentiation-inducing properties of IFN gamma. In contrast, several lung carcinoma cell lines are insensitive to the growth-inhibitory and differentiation-inducing actions of IFN gamma, suggesting they may have acquired defects in certain IFN gamma signaling pathways. Although the growth of human bronchial fibroblasts is not altered, expression of certain immunomodulatory genes is induced by IFN gamma.(ABSTRACT TRUNCATED AT 250 WORDS)

Interferon gamma (IFN-gamma) is a potent inducer of squamous differentiation in normal human epidermal keratinocytes. This induction is characterized by a > or = 95% decrease in the mRNA level of two growth regulatory genes, cdc2 and E2F-1, and a 7-15-fold increase in the expression of two squamous cell-specific genes, transglutaminase type I and cornifin. In contrast to the decrease in cdc2 and E2F-1 expression, the increase in transglutaminase type I and cornifin mRNAs by IFN-gamma occurs after a lagtime of more than 12 h. These results are consistent with the hypothesis that in normal human epidermal keratinocyte cells irreversible growth arrest precedes the expression of the squamous-differentiated phenotype. The action of IFN-gamma on the expression of squamous cell-specific genes is antagonized by retinoic acid and transforming growth factor beta 1. Both factors are potent suppressors of the induction of transglutaminase type I and cornifin; however, they do not prevent the commitment to irreversible growth arrest. Several squamous cell carcinoma cell lines do not show a detectable decrease in cdc2 or increase in transglutaminase type I mRNA levels after IFN-gamma treatment and appear to be altered in their control of squamous differentiation.

The expression of cornifin, a putative cross-linked envelope precursor, was investigated in several squamous differentiating tissues by in situ hybridization and immunohistochemical analysis. Cornifin mRNA and protein, which are absent in the normal mucociliary tracheal epithelium, are induced in the suprabasal layers of the squamous metaplastic tracheal epithelium of vitamin A-deficient hamsters. Similar to the induction of squamous metaplasia in vivo, culture of rabbit tracheal cells in the absence of retinoids results in squamous differentiation and expression of cornifin. This induction of cornifin expression is suppressed by retinoic acid and several of its analogs. Cornifin mRNA and protein are also detected in the suprabasal layers of the squamous epithelium of rabbit esophagus and tongue. The distribution of cornifin in human epidermis was compared with that of two other crosslinked envelope precursor proteins, involucrin and loricrin. The localization of cornifin and involucrin is very similar. Both are induced in the spinous layer and appear at an earlier stage during epidermal differentiation than loricrin. The expression of cornifin is greatly increased in psoriatic skin. Cornifin mRNA is barely detectable in normal epidermis, whereas it is present at relatively high levels in the suprabasal layers of psoriatic epidermis. Topical treatment with RA results in thickening of the skin and increases the level of cornifin mRNA and protein in the upper spinous layers of mouse skin. Cornifin expression correlates generally with squamous differentiation in a variety of tissues and is abnormally regulated in psoriatic skin and in skin treated topically with retinoic acid.