Regulation and expression of multidrug resistance (MDR) transcripts in the intestinal epithelium.

A paucity of information exists on the regulation of gene expression in the undifferentiated intestine. The intestinal epithelium is one of the few normal tissues expressing the multidrug resistance (MDR) genes that confer the multidrug resistant phenotype to a variety of tumours. Expression of mdr1a has been observed in the primitive rat intestinal epithelial cell line, IEC-18. It is hypothesized that characterization of MDR gene expression in IEC-18 cells will provide insight into gene regulation in undifferentiated intestinal cells. A series of hamster mdr1a promoter deletion constructs was studied in IEC-18 and a region with 12-13-fold enhancer activity was identified. This region was shown to function in an orientation- and promoter context-independent manner, specifically in IEC-18 cells. Unexpectedly, Northern probing revealed a greater expression of mdr1b than mdr1a in IEC-18 cells. A quantitative reverse transcription polymerase chain reaction assay was used to compare the relative expression of MDR genes in IEC cells, fetal intestine, and in the undifferentiated and differentiated components of adult intestinal epithelium. MDR transcript levels in IEC cells were found to resemble those of fetal intestine and small intestinal crypts, where a conversion from mixed mdr1a/mdr1b to predominantly mdr1a expression occurs as cells mature. This work describes two contributions to the field of gene regulation in the undifferentiated intestine--first, the initial characterization of a putative mdr1a enhancer region with specificity for primitive intestinal cells and secondly, the first report of mdr1b detection in the intestine and its expression in primitive cell types.

Syndecan-1 is up-regulated in ras-transformed intestinal epithelial cells.

The syndecans, a family of cell-surface heparan sulphate proteoglycans, have been proposed to mediate cellular interactions with extracellular effector molecules, such as growth factors and components of the extracellular matrix, during critical phases of development. Transcripts of all four syndecans are expressed at varying levels in the developing rat intestine and in a series of immature rat intestinal epithelial cell lines. In addition, we report the novel finding that, in the intestinal epithelial cell lines, expression of syndecan-1 transcript is up-regulated by transformation with activated H-ras. This is in contrast to other cell lines in which ras transformation is associated with a decrease in syndecan-1 levels. The observed increase in the syndecan-1 occurs as a result of increased transcription and can be correlated with the degree of transformation of the IEC-18 cells. Transformation is also associated with a decrease in apparent molecular weight and increased shedding of the proteoglycan into the culture medium. Increased shedding of syndecan-1 into the culture medium after transformation with H-ras may contribute to the disruption of proteoglycan interactions with the extracellular matrix, leading to alterations in cell adhesion and organization.

Transcriptional regulation of OCI-5/Glypican 3: elongation control of confluence-dependent induction.

OCI-5/Glypican 3, a member of the glypican family of proteoglycans, is the defective gene in the Simpson-Golabi-Behmel overgrowth syndrome. OCI-5 expression is developmentally regulated in the intestinal epithelium, and the mechanism of its regulation was studied in the rat intestinal epithelial cell line IEC-18. A large induction of OCI-5 transcript and protein was observed at high cell density. Among other glypican family members, kappa-glypican also exhibited a confluence-dependent induction in select cell types. Nuclear run-on analysis indicated that cell-density regulation of OCI-5 occurs at the level of transcription. The rat and mouse OCI-5 promoters were cloned and found to be highly conserved, located within CpG islands and contain regions of alternating purine and pyrimidine residues. No TATA-box or recognizable INR element was observed. Consensus binding sites for AP-2, SP-1, zeste and NF-1/CTF are conserved across human, mouse and rat promoters. 5' deletion mapping of the rat promoter identified regions which enhance and repress promoter activity, with no apparent confluence-dependence or tissue-specificity. Nuclear run-on analysis probing different regions of the gene suggests that elongation control plays a role in the induction of OCI-5 by confluence.

Expression of OCI-5/glypican 3 during intestinal morphogenesis: regulation by cell shape in intestinal epithelial cells.

OCI-5, the rat homologue of human glypican 3 (GPC3), is believed to be involved in morphogenesis and growth control during development. The finding that GPC3 is mutated in patients with the Simpson-Golabi-Behmel overgrowth syndrome is consistent with this idea. In this report, using RNA in situ hybridization, expression of OCI-5 in the developing intestine is detected in both endoderm- and mesenchyme-derived cells in a phased manner related to age and proximal/distal position. To investigate the mechanism of its regulation during intestinal development, OCI-5 expression was studied in the primitive rat intestinal epithelial cell line IEC-18. The expression of the OCI-5 transcript is increased in IEC-18 cells at confluence, in low calcium media, and during spheroid culture, all conditions which result in the cells acquiring a more rounded cell shape. In contrast, cytoskeletal disruption with colchicine causes cells to flatten and spread and abolishes both the confluence- and the low calcium-dependent induction of OCI-5. Treatment with vanadate, a phosphatase inhibitor, causes cells to acquire a spindle-shaped morphology and prevents OCI-5 induction in all situations. Nuclear run-on analysis demonstrates that the rate of OCI-5 transcription is increased at confluence, in low calcium media, and during spheroid culture of IEC-18, and decreased by treatment of cells with colchicine. Together, these data suggest that OCI-5 expression is regulated in IEC-18 by cell shape. The pattern of expression of OCI-5 in the developing intestine is consistent with it playing a role in epithelial-mesenchymal interactions during intestinal morphogenesis, when cell shape changes are likely to occur.

Regulation of transforming growth factor beta receptors in H-ras oncogene-transformed rat intestinal epithelial cells.

We have reported that over expression of the H-ras oncogene causes resistance to growth inhibition by transforming growth factor beta 1 (TGF-beta 1) and a time-dependant switch of type II to type I TGF-beta receptor expression in the rat intestinal epithelial cell line IEC-18 (J. Filmus, J. Zhao, and R. N. Buick, Oncogene, 7: 521-526, 1992). Here, we investigate the possible mechanisms involved in H-ras-mediated regulation of TGF-beta receptors in an IEC-18 cell clone expressing H-ras, conditional on the activity of a dexamethasone-sensitive promoter. The switch from type II to type I receptor expression in response to H-ras expression has a requirement for de novo RNA synthesis. In addition, accumulation of TGF-beta receptor type II mRNA is approximately 5-fold lower in ras-expressing cells compared to control cells. Nuclear run-on experiments suggest that the down-regulation of type II receptor mRNA by H-ras oncogene is based, at least in part, on reduced transcription. We have also analyzed the consequences of H-ras expression on the properties of the TGF-beta receptors. Type I and II in IEC-18 cells and type I receptors in ras-transformed cells have similar characteristics in terms of binding affinities for TGF-beta 1 (or TGF-beta 2) turnover rates and glycosylation states. Notably, the type I receptors in ras-transformed cells are not capable of ligand-induced internalization. Although H-ras expression in IEC-18 cells causes resistance to TGF-beta-mediated growth inhibition, the cells remain responsive to TGF-beta 1 stimulation of fibronectin expression. These results are discussed in the context of the knowledge of TGF-beta receptor complexity and signal transduction, and with reference to the potential role for loss of TGF-beta-mediated negative growth regulation in malignant transformation.

Mouse Paneth cell defensins: primary structures and antibacterial activities of numerous cryptdin isoforms.

Cryptdins are antimicrobial peptides of the defensin family that are produced by intestinal Paneth cells. mRNAs encoding 17 cryptdin isoforms have been characterized from a cDNA library generated from a single jejunal crypt. Six cryptdin cDNAs correspond to known peptides, and the remainder predict 11 novel Paneth cell defensins. Most cryptdin cDNAs have > or = 93% nucleotide sequence identity overall, except for cryptdin 4 and 5 cDNAs, whose respective mature peptide-encoding regions are only 74 and 78% identical to that of cryptdin 1. Cryptdin cDNAs differ at a small number of nucleotide positions: frequent substitutions were found in codons 38 and 52 of the propiece and in codons 68, 73, 76, 87, and 89 of the deduced peptides; cDNA clones with changes in codons 74, 83, and 88 were found, but there were fewer of these. The antimicrobial activities of cryptdins 1 to 6 were tested against Escherichia coli ML35 in two assays. In an agar diffusion assay, the potencies of cryptdins 1 to 3, 5, and 6 were approximately equivalent to that of rabbit neutrophil defensin NP-1 but cryptdin 4 was 30 times more active than NP-1. In a bactericidal assay system, cryptdins 1 and 3 to 6 were equally active at 10 micrograms/ml but cryptdin 2 and rabbit NP-1 were not active at this concentration. Since cryptdins 2 and 3 differ only at residue 10 (Thr and Lys, respectively), this amino acid appears to function in bactericidal interaction with E. coli. The demonstration that Paneth cells express a diverse population of microbicidal defensins further implicates cryptdins in restricting colonization or invasion of small intestinal epithelium by bacteria.

Activated ras and src induce CD44 overexpression in rat intestinal epithelial cells.

CD44 is an adhesion molecule that is involved in the progression of several tumor types, including those originating in the intestine. There are several alternatively spliced forms of CD44. Here we show that intestinal epithelial cells express the standard form of CD44 (CD44s). The same form of CD44 is found in IEC-18, a cell line derived from normal rat intestinal crypts. Upon transfection of IEC-18 cells with ras or src, two oncogenes that are frequently activated in intestinal tumors, a significant induction of CD44s is observed. A causal role for ras in this induction is shown by using IEC clones transfected with an inducible ras expression vector. The oncogene-transformed IEC clones display a high degree of hyaluronic acid-dependent cell-cell adhesion that is not observed in the parental IEC-18 cells suggesting that ras- and src-induced overexpression of CD44 can alter the adhesion properties of intestinal cells.

In vitro cDNA amplification from individual intestinal crypts: a novel approach to the study of differential gene expression along the crypt-villus axis.

The mouse small intestine is lined with a monolayer of continuously renewing epithelial cells. Cells of four distinct epithelial lineages are derived clonally from the stem cell zone, located near the crypt base, from which cells differentiate and migrate to the villus tip. The kinetics of the multilineage process are well understood. However, the molecular mechanisms underlying gene expression during lineage commitment and cell proliferation and differentiation remain obscure. A novel approach to the problem is presented here. Single intact epithelial crypts were isolated by incubation in ethylenediaminetetraacetic acid and mechanical vibration of everted mouse intestinal or colonic segments. Crypts isolated in this manner were suitable for mRNA-directed polymerase chain reaction, thus generating crypt epithelium-specific cDNA. The fidelity of transcript amplification was confirmed by Southern blot hybridization with cloned intestinal transcripts. To demonstrate the potential utility of crypt-specific cDNA, the amplified transcripts from a single jejunal crypt were used to construct a cDNA library, characterization of which revealed a high representation of cryptdin-1-related transcripts. This study presents a technique which will facilitate comprehensive analyses of gene expression in the differentiating mammalian intestine.

Relationship of levels and kinetics of H-ras expression to transformed phenotype and loss of TGF-beta 1-mediated growth regulation in intestinal epithelial cells.

Immortalized, nontumorigenic rat intestinal epithelial cells (IEC-18) can be transformed to tumorigenicity by expression of an activated human H-ras gene. Here we describe the characteristics of an IEC-18 cell line in which the activated human H-ras gene has been introduced under the control of the steroid-sensitive promoter of the mouse mammary tumor virus long-terminal repeat. The clonal cell line (IEC-18 C125) is phenotypically normal in the absence of the transcription inducer, dexamethasone, and transformed when treated with high levels of inducer. Transformed morphology and growth characteristics are dependent on levels of H-ras expression. IEC-18 C125 cells have been used to demonstrate a general relationship (dose and kinetic) between H-ras expression and loss of TGF-beta 1-mediated growth regulation. This effect occurs concomitantly with a ras-dependent change in the profile of TGF-beta 1 binding proteins.

Overexpression of H-ras oncogene induces resistance to the growth-inhibitory action of transforming growth factor beta-1 (TGF-beta 1) and alters the number and type of TGF-beta 1 receptors in rat intestinal epithelial cell clones.

In this report, we utilize rat intestinal cell (IEC-18) clones expressing an activated human H-ras gene to investigate the relationship between malignant transformation and growth control by transforming growth factor beta (TGF-beta). We demonstrate that clones expressing high levels of H-ras oncogene lose sensitivity to the growth inhibitory action of TGF-beta. The loss of sensitivity is related to the degree of H-ras expression and is shown to be a direct consequence of H-ras expression through the use of a clonal cell line with inducible expression of activated H-ras. Co-incident with the loss of growth inhibition, ras-expressing clones display an altered expression of TGF-beta-binding proteins as detectable by [125I]TGF-beta cross-linking. While IEC-18 cells express type II (92 kDa) binding protein predominantly, H-ras expression induces a shift to predominantly type I (69 kDa) binding protein expression.

Clonogenic cell assay for anchorage-dependent squamous carcinoma cell lines using limiting dilution.

Clonogenic assays under either anchorage-dependent or -independent conditions are very useful for testing the sensitivity of tumor cells to cytotoxic drugs and radiation. These assays have not been widely used with squamous-cell carcinomas (SCC) because of poor tumor-cell viability and poor cloning efficiency, especially in semi-solid media. To find a clonogenic assay suitable for use with human squamous cancers we tested SCC lines, derived in our laboratory from patients with head and neck cancer, for the capacity to form colonies in soft agar and in 96-well plates. Of 13 UM-SCC lines tested for colony formation in agarose, only UM-SCC-11A was capable of growth in conventional semi-solid media. One other line, UM-SCC-14C, produced colonies in agarose only in the presence of epidermal growth factor. In contrast, all 17 of the SCC lines tested exhibited colony formation in adherent cell culture using limiting dilution in 96-well plates. The plating efficiencies of the SCC lines in the 96-well plate assay ranged from 0.02 to 0.52 colonies (wells)/cell whereas the PE values in soft agar were lower, ranging from 0.0055 to 0.0086 colonies/cell. The 96-well plate assay is not affected by cell migration, a problem encountered with some cell lines when clonogenic assays are performed in Petri dishes. UM-SCC-11A was tested for radiation sensitivity both in soft agar and in the 96-well plate assay. Comparable results were obtained. In summary, the majority of SCC cell lines did not form viable colonies in soft agar but the 96-well plate assay was applicable to a broad spectrum of anchorage-dependent human SCC cell lines and provides an efficient method for evaluating clonogenic cell survival.

Activation of the Na+/H+ antiport is not required for epidermal growth factor-dependent gene expression, growth inhibition or proliferation in human breast cancer cells.

Mitogen interaction with specific receptors in many cell types leads to activation of the Na+/H+ antiport and a resultant cytoplasmic alkalinization. Since amiloride inhibits both Na+/H+ exchange and cell proliferation, it has been hypothesized that activation of the antiport is an obligatory requirement for mitogenesis. However, concentrations of amiloride which inhibit the antiport also inhibit other cellular processes, including protein synthesis and phosphorylation. We have used an epidermal growth factor (EGF) receptor gene-amplified human breast cancer cell line, the growth of which is inhibited by high levels of EGF in culture (MDA-468) and a variant, the growth of which is stimulated by EGF (MDA-468-S4), along with two potent amiloride analogues to examine whether activation of the Na+/H+ antiport and cytoplasmic alkalinization is necessary for both EGF-dependent effects to occur. At concentrations of the amiloride analogues which block Na+/H+ exchange in both cell types by 76-98%, the EGF-dependent alterations in [3H]thymidine incorporation or induction in c-myc or c-fos gene transcription were unaltered. These results were confirmed by a lack of effect of the amiloride analogues on both the growth-stimulatory and growth-inhibitory effects on EGF in an anchorage-independent growth assay. Similarly, in pH-altered media that prevented normal cytoplasmic alkalinization, the response of both MDA-468 and MDA-468-S4 to EGF activation was unaltered. In addition, activation of the Na+/H+ antiport alone was not sufficient to induce c-myc and c-fos transcription in either cell type. Taken together, these data suggest that neither the Na+/H+ antiport nor cytoplasmic alkalinization are necessary or sufficient for either EGF-dependent growth stimulation or growth inhibition in MDA-468 human breast cancer cells.

Human ovarian carcinoma: evidence for patient-related differences in susceptibility to cytotoxic effectors that attack different cellular subpopulations within a tumour.

Human ovarian carcinoma cells obtained from ascites were tested for susceptibility to lysis by peripheral blood NK cells, alpha-interferon-activated NK cells, and interleukin 2-activated killer cells. Cryopreserved tumour cell preparations were used to allow repeated testing of the same target, and the tumour cells were fractionated using albumin density gradients to determine if fractions containing clonogenic (stem) cells were killed. Four tumour cell donors were studied and each showed a different pattern of susceptibility of unfractionated tumour to lysis by different effector cells. Using fractionated tumour cells, we found that NK and interferon-activated NK cells did not always lyse cells in the clonogenic fractions and that interferon activation could in some cases shift killing away from the clonogenic fractions and towards the peak of proliferating (but not self-renewing) colony forming cells. Interleukin 2-activated killer cells (LAK) however, killed the fractions containing clonogenic cells in all 4 cases. The magnitude of killing seen when fractions of the original tumour were tested were often striking when compared to lysis of the unfractionated cells. Apparent heterogeneity between patients and stem cell susceptibility to effector cells may be important determinants of the efficacy of treatment of patients with biologic response modifiers such as interferon and interleukin 2.

Isolation of a cDNA corresponding to a developmentally regulated transcript in rat intestine.

We report the isolation of a cDNA clone corresponding to a transcript that is accumulated differentially in rat intestine during development. Clone OCI-5 was selected from the rat intestinal cell line IEC-18, which represents primitive intestinal epithelial crypt cells. Expression was high in rat fetal intestine between 15 and 19 days of development and thereafter was progressively down regulated, becoming undetectable after weaning. Clone OCI-5 detected homologous sequences in human and murine cells. In particular, a high level of expression was detected in CaCo-2, a human colon carcinoma cell line, which is known to express molecules characteristic of fetal small intestinal cells. Expression of a homologous gene was also detected in F9 murine teratocarcinoma cells when they were induced to differentiate into parietal or visceral endodermlike cells. When IEC-18 cells were transformed by activated H-ras or v-src genes, expression of clone OCI-5 was suppressed; the degree of down-regulation correlated with the extent of morphological change induced in the transformed IEC-18 cells. The sequence of clone OCI-5 showed an open reading frame that was capable of encoding a protein of 597 amino acids, but no strong homology was found with any of the proteins registered in the protein sequence data base.

Malignant properties of sublines selected from a human bladder cancer cell line that contains an activated c-Ha-ras oncogene.

The human bladder cancer cell line MGH-U1 (also designated T-24 or EJ) contains an activated c-Ha-ras oncogene, which is amplified as compared to normal human fibroblasts. We have generated sublines from the MGH-U1 cell line: the MGH-U1/OCI subline was generated by dissociating spheroids formed from MGH-U1 cells; the U1-m/F1 and OCI-m/F1 were generated by in vivo passage of experimental lung metastases formed after i.v. injection of MGH-U1 and MGH-U1/OCI lines into immune-deprived mice; the U1/t subline was generated by in vivo passage of i.m. tumors formed from MGH-U1 cells. All sublines formed tumors in immune-deprived mice from smaller i.m. inocula than the parent line, and the U1-m/F1 subline generated more spontaneous metastases in lungs. Lung colony forming efficiency after i.v. injections of cells into similar mice was also greater for the sublines than for the parent MGH-U1 cells. The U1-m/F1 and OCI-m/F1 were the most tumorigenic lines. Early passages of the MGH-U1/OCI subline showed the presence of double minute chromosomes, and amplification and increased expression of the c-Ha-ras oncogene as compared to the parental cell line. These changes were not present in later cultures of MGH-U1/OCI cells, and no consistent difference in the levels of gene amplification or expression between the parent line and the sublines was found. Thus the content and expression of the activated c-Ha-ras oncogene does not correlate with malignant properties of the sublines.

Sodium butyrate induces differentiation in breast cancer cell lines expressing the estrogen receptor.

Addition of sodium butyrate (NaB) to 6 cultured human breast carcinoma cell lines results in a dose and time-dependent growth inhibition. Kinetic evidence, related to the growth of a minority cell population which decreases in size with time of exposure, is presented to indicate that the NaB effect is reversible. In those cell lines that express the estrogen receptor (ER), growth inhibition is accompanied by a more differentiated phenotype, which is characterized by increased accumulation of lipid and milk-fat globule membrane glycoproteins. The potential for differentiation is not blocked by tamoxifen, indicating that the relationship to ER expression is likely secondary to the association of ER expression with a particular stage of secretory cell differentiation that is susceptible to NaB induction. Of the 3 lines shown to respond in this way (MCF-7, ZR-75-1, and MDA-134), ZR-75-1 is an extreme example that may serve as a model for studies of gene expression during human mammary epithelial cell differentiation.

G-protein-mediated epidermal growth factor signal transduction in a human breast cancer cell line. Evidence for two intracellular pathways distinguishable by pertussis toxin.

The MDA-468 human breast cancer cell line has an amplified epidermal growth factor (EGF) receptor gene (20 x) and correspondingly overexpresses the EGF receptor. Since this cell line is growth inhibited by supra-physiological levels of EGF in tissue culture, it has been possible to select variant cells which have lost the chromosome bearing the amplified EGF receptor domain and which are capable of growing in high levels of EGF. One such cell line (MDA-468-S4) shows an absolute requirement for EGF for growth in anchorage-independent tissue culture conditions. We have utilized MDA-468 and MDA-468-S4 to examine the intracellular transduction of EGF signals leading to growth inhibition and proliferation, respectively. We report that in anchorage-independent conditions, pertussis toxin can abrogate both the EGF-dependent growth inhibition in MDA-468 cells and the EGF-dependent cell proliferation in MDA-468-S4 cells. This inhibition is paralleled by the ADP-ribosylation of an endogenous 41,000-dalton membrane protein in both MDA-468 and MDA-468-S4 cells. In contrast, the toxin does not prevent the transient, augmented expression of c-myc and c-fos mRNA seen in response to EGF in both cell types. These data suggest 1) the notion of more than one simultaneous, parallel, intracellular EGF-dependent signal transduction pathway and 2) G-protein involvement in at least one pathway mandatory for the growth modulating responses to EGF in anchorage-independent conditions, but distinct from that inducing c-myc and c-fos mRNA expression.