Search

Cell plasticity is emerging as a key regulator of tumor progression and metastasis. During carcinoma dissemination epithelial cells undergo epithelial to mesenchymal transition (EMT) processes characterized by the acquisition of migratory/invasive properties, while the reverse, mesenchymal to epithelial transition (MET) process, is also essential for metastasis outgrowth. Different transcription factors, called EMT-TFs, including Snail, bHLH and Zeb families are drivers of the EMT branch of epithelial plasticity, and can be post-transcriptionally downregulated by several miRNAs, as the miR-200 family. The specific or redundant role of different EMT-TFs and their functional interrelations are not fully understood. To study the interplay between different EMT-TFs, comprehensive gain and loss-of-function studies of Snail1, Snail2 and/or Zeb1 factors were performed in the prototypical MDCK cell model system. We here describe that Snail1 and Zeb1 are mutually required for EMT induction while continuous Snail1 and Snail2 expression, but not Zeb1, is needed for maintenance of the mesenchymal phenotype in MDCK cells. In this model system, EMT is coordinated by Snail1 and Zeb1 through transcriptional and epigenetic downregulation of the miR-200 family. Interestingly, Snail1 is involved in epigenetic CpG DNA methylation of the miR-200 loci, essential to maintain the mesenchymal phenotype. The present results thus define a novel functional interplay between Snail and Zeb EMT-TFs in miR-200 family regulation providing a molecular link to their previous involvement in the generation of EMT process in vivo.

Circulating tumor cells (CTCs), proposed as major players in cancer dissemination, have demonstrated clinical prognostic significance in several cancer types. However, their predictive value remains unclear. Here we evaluated the clinical utility of six CTC markers (tissue specific and epithelial to mesenchymal transition transcripts) both as prognostic and predictive tools in metastatic colorectal cancer (mCRC) patients. CTCs were immunoisolated from blood in 50 mCRC patients at baseline and at 4 and 16 weeks after treatment onset. Expression levels of GAPDH, VIL1, CLU, TIMP1, LOXL3 and ZEB2 were determined by qualitative polymerase chain reaction and normalized to the unspecific cell isolation marker CD45. At baseline, median progression-free survival (PFS) and overall survival (OS) for patients with high CTC markers were 6.3 and 12.7 months, respectively, versus 12.7 and 24.2 for patients with low CTC markers (PFS; p = 0.0003; OS; p = 0.044). Concerning response to therapy, PFS and OS for patients with increased CTC markers along treatment were, respectively, 6.6 and 13.1 months, compared with 12.7 and 24.3 for patients presenting CTC markers reduction (PFS; p = 0.004; OS; p = 0.007). Of note, CTC markers identified therapy-refractory patients not detected by standard image techniques. Patients with increased CTC markers along treatment, but classified as responders by computed tomography, showed significantly shorter survival times (PFS: 7.8 vs. 13.2; OS: 14.4 vs. 24.4; months). In conclusion, we have generated a CTC marker panel for prognosis evaluation and the identification of patients benefiting or not from therapy in mCRC. Our methodology efficiently classified patients earlier than routine computed tomography and from a minimally invasive liquid biopsy.

This work was supported by the Instituto de Salud Carlos III (ISCIII; Grant No. PI080971 to JP) and the Ministerio de Ciencia e Innovacion (MICINN), co-financed by the European Development Regional Fund (Grant No. RD06/0020/0013 to JP), the Junta de Andalucia (Consejeria de Salud; Grant Nos PI-0384/2007 and PI0581/2009, to JP), MICINN (Grant No. SAF-2010-21143; Consolider-Ingenio Grant No. CSD2007-00017 to AC; Grant No. SAF2010-20175 to GMB), Comunidad de Madrid (Grant No.S2010/BMD-2303, to AC and GMB), Asociacion Espanola Contra el Cancer (Grant No. AECC 2011, to GMB) and ISCIII-Ministerio de Economia y Competitividad (MINECO; Grant No. RD12/0036/0064 to JP; Grant No. RD12/0036/0007 to AC). JDM is a PhD researcher funded by the Consejeria de Salud, Junta de Andalucia (Grant No. PI0581/2009). MAC and ADL are PhD researchers funded by the ISCIII (Grant No. RD06/0020/0013 and Sara Borrell contracts, respectively).

Endometrial carcinosarcomas (ECSs) undergo a true epithelial-mesenchymal transition (EMT). The molecular determinants of the EMT in vivo are unclear, although a role for some miRNAs, mainly involving the miR-200 family, was recently suggested from in vitro cellular models. We analysed the microRNA (miRNA) signatures associated to EMT in human carcinosarcomas, and determined their relationships with EMT markers and repressors of E-cadherin transcription. The expression of E-, P- and N-cadherin, cadherin-11, p120, vimentin, SPARC, fascin and caveolin-1 was studied in a group of 76 ECS by immunohistochemistry. In addition, real-time PCR was used to measure the differences in the expression of 384 miRNAs, E-cadherin, cadherin-11, SPARC, SNAIL, ZEB1, ZEB2, TWIST-1, TCF4, TGFβ1 and TGFβ2 between the epithelial and mesenchymal components of 23 ECSs. A loss of epithelial characteristics, including cadherin switching and the acquisition of a mesenchymal phenotype, was accompanied by changes in the profile of miRNA expression and the up-regulation of all the E-cadherin repressors analysed. A greater than five-fold difference in the expression of 14 miRNAs between both neoplastic components was seen. Members of the miR-200 family were down-regulated in the mesenchymal part of the ECS. In addition, miR-23b and miR-29c, which are involved in the inhibition of mesenchymal markers, and miR-203, which is involved in the inhibition of cell stemness, were also down-regulated. Up-regulated miRNAs included miR-155, miR-369-5p, miR-370, miR-450a and miR-542-5p. These data suggest that in human ECS, the interplay between transcriptional repressors of E-cadherin and miRNAs provides a link between EMT-activation and the maintenance of stemness.

P-cadherin expression is restricted to the basal layer of stratified epithelia including that of the mammary gland. Although evidence for an important role of P-cadherin in mammary morphogenesis and tumorigenesis is increasing, the mechanisms that regulate its expression are poorly understood. We show that in basal mammary epithelial cells, beta-catenin is associated with the P-cadherin promoter and activates its expression independently of LEF/TCF in a cell-type specific manner. Down-regulation of endogenous beta-catenin levels by RNA interference technique inhibited P-cadherin promoter activity. In vivo, in skin and mammary gland of mutant mice, activation of beta-catenin signalling correlates with up-regulation of P-cadherin expression. These data suggest that beta-catenin-dependent modulation of P-cadherin expression can contribute to the establishment of the basal phenotype.

Various mutations of the hairless (hr) gene of mice result in hair loss and other integument defects. To examine the role of the hr gene in mouse development, the expression profile of hr has been determined by in situ hybridisation and correlated to the nature of genetic changes and morphological abnormalities in different mutant animals. Four variant alleles have been characterised at the molecular level. hr/hr mice produce reduced, but significant, levels of hr mRNA whereas other alleles contain mutations which would be expected to preclude the synthesis of functional product, demonstrating a correlation between allelic variation at the hr locus and phenotypic severity. hr expression was shown to be widespread and temporally regulated. It was identified in novel tissues such as cartilage, developing tooth, inner ear, retina, and colon as well as in skin and brain. Analysis of mice homozygous for the rhino allele of hairless revealed that, although no morphological defects were detectable in many tissues normally expressing hr, previously undescribed abnormalities were present in several tissues including inner ear, retina, and colon. These findings indicate that the hairless gene product plays a wider role in development than previously suspected. Dev Dyn 1999;216:113-126.

We previously isolated the 5' upstream sequences of the mouse P-cadherin gene, in which putative binding sites for several transcription factors were identified between nt-101 and +30. In the study reported here, the promoter activity of the postulated 5' cis-acting sequences of the P-cadherin promoter, and the activity of the proximal E-cadherin promoter were investigated in several murine keratinocyte cell lines showing different levels of P- and E-cadherin expression as well as different morphology and tumorigenic behavior. Cell-type specificity and optimal activity of P-cadherin expression in murine keratinocytes was conferred by 5' sequences located between nt -200 and +30, and the GC-rich region (nt -101 to +80) and a CCAAT box element (nt -65) had a major regulatory role. The cell-type specificity of the E-cadherin promoter, on the other hand, was mediated by a combination of positive regulatory elements, a GC-rich region (nt -58 to -24), and a CCAAT box (nt -65) and repressor elements inside the E-pal sequence. Interestingly, the maximum repressor effect of the E-pal element was observed in non-expressing undifferentiated spindle cells. In vitro binding studies indicated that the GC-rich region of the P-cadherin promoter was mainly recognized by Sp1-related nuclear factors, whereas both AP2- and Sp1-related factors were involved in the interaction of the GC-rich region of the E-cadherin promoter. Common factors (probably related to the CP1 family) seemed also to be involved in the recognition of the CCAAT-box element of both the E- and P-cadherin promoters, but additional specific factors participated in the interaction with the CCAAT box of the E-cadherin promoter. Our studies also support the hypothesis that loss or modification of some of the regulatory factors occurs during mouse skin tumor progression.

The HaCa4 cell line, derived from a mouse skin carcinoma induced by Harvey murine sarcoma virus, is highly tumorigenic when injected into nude mice and produces multiple metastases in the lungs. HaCa4 cells express high levels of viral Ha-ras oncogene products, anomalously synthesize the embryonic/simple epithelial keratin K8, and have lost the expression of the cell-cell adhesion receptor E-cadherin (E-CD). E-CD(+) cell clones (E62 and E24), obtained by transfection of an exogenous E-CD cDNA into HaCa4 cells, had a decreased ability to migrate through type IV collagen matrices. However, the E-CD (+) E62 clone remained as metastatic as the parental cell line, whereas the E24 clone, which does not take up the exogenous cDNA but spontaneously switches on the endogenous E-CD gene, suppressed the metastatic phenotype although it maintained its tumorigenicity. E24 cells had fivefold to sixfold lower levels of viral Ha-ras mRNA and p21 protein than the other cell lines. In addition, they did not synthesize K8 but rather switched on keratin K19. The comparison of E-CD proteins synthesized by E62 and E24 cell lines revealed no structural or functional differences because both localized at cell-cell contacts and associated with alpha-catenin, beta-catenin, and plakoglobin. Furthermore, E-CD was still expressed in metastatic lung nodules produced by E62 cells. These results suggest that suppression of the metastatic phenotype in E24 cells occurs independently of E-CD expression and correlates with decreased levels of the oncogenic ras p21 protein.

Expression of the cell adhesion molecules E-cadherin, P-cadherin and alpha 6 beta 4 integrin and of the keratin K13 has been analyzed in chemically induced benign skin papillomas with genetically pre-determined risks for malignant conversion. It has been previously shown that papillomas induced in mice lacking both alleles of the p53 gene have a much higher rate of malignant conversion than those induced in wild-type and heterozygous p53 mice. Alterations in the expression pattern of the E-cadherin molecule, including focal loss at cell-cell contacts and heterogeneous distribution in the differentiated layers, were found in about 70% of the p53 null papillomas. In contrast, all of the wild-type and over 85% of the heterozygous p53 papillomas exhibited an expression pattern of E-cadherin indistinguishable from that of normal epidermis. Alterations in P-cadherin expression were also detected in the p53 null papillomas: aberrant suprabasal localization and heterogeneous distribution were observed more frequently than in heterozygous and wild-type p53 papillomas. The alpha 6 beta 4 integrin showed suprabasal expression in more than 70% of the papillomas derived from either wild-type, heterozygous or homozygous p53 null mice. Surprisingly, the extent of the suprabasal localization of alpha 6 beta 4 decreased in the p53 null papillomas. Aberrant keratin K13 expression was also detected in the majority of cases of all p53 genotypes, but again there was a clear decrease in expression levels in the p53 null papillomas. These alterations were also associated with keratinocytic atypia, which increased significantly in the p53 null papillomas. Changes in these parameters were particularly evident during malignant conversion in invasive regions of one progressing p53 null papilloma. Our results indicate the existence of dynamic changes in the expression pattern of the 3 cell adhesion molecules analyzed and identify down-regulation of E-cadherin as an early step in malignant conversion.

Antibodies against pure discoidin I have been used as a tool to ascertain the role of this lectin in aggregation of Dictyostelium discoideum. Discoidin I is widely expressed over the cell surface of aggregation-competent AX-2 cells, as ascertained by indirect immunofluorescence with specific (antidiscoidin I) antibodies. Univalent antidiscoidin I antibodies (Fab fragments) inhibit the aggregation-specific intercellular adhesion of D discoideum AX-2 cells in an in vitro assay. This inhibition depends on antibody concentration and cell density; a 50% inhibition of cell aggregation was obtained at antidiscoidin I Fab concentration of 4.5 mg/ml and 1 X 10(6) cells/ml. Aggregation and morphogenesis on solid support is also effectively inhibited when AX-2 cells are starved in the presence of antidiscoidin I Fab fragments. The inhibition of morphogenesis is also dose dependent and more effective than in the in vitro assay. No inhibition of aggregation either in the in vitro assay or on morphogenesis on solid support was observed with preimmune Fab fragments at any of the concentrations tested (up to 9.6 mg/ml).