Electrophysiological study of coupling between cultured cells of the mouse mammary gland in five distinct physiological states.

Electrical coupling has been observed between cultured cells of the mouse mammary gland in five distinct physiological or pathological states. We have employed young primary cultures of cells dissociated from the following tissues: normal glands from young virgin or midpregnant females, hyperplastic alveolar nodules (believed to be precancerous) transplanted in gland-free mammary fat pads, and spontaneous mammary adenocarcinomas and their pulmonary metastases. All successfully impaled pairs of cells (a total of 97 pairs) were found to be ionically coupled. Furthermore, in normal and tumor cell cultures, electrical coupling was observed between dome-dome and dome-nondome cell pairs. This study correlates with electronmicroscopic studies of fresh normal, hyperplastic, and tumor samples, which show the presence of gap junctions in all three.     

Organ culture passage enhances the oncogenicity of carcinogen-induced hyperplastic mammary nodules

Summary The purpose of this study was to determine whether culturing hyperplastic mammary nodules in hormone-free medium would enhance their oncogenicity following subsequent transplantation into mammary fat pads. The underlying hypothesis is that the proliferation of transformed cells within the nodules is inhibited by hormone-dependent normal cells also present in the nodules. Accordingly, both primary hyperplastic nodules and tissues from a hyperplastic outgrowth of a primary nodule were maintained as organ cultures for varying periods in hormone-free Medium 199. The results show that whereas noncultured nodules developed mammary tumors at an incidence of only 15%, those passaged in organ culture gave rise to mammary tumors at an incidence of 40 to 43%. This threefold enhancement in the oncogenicity of mammary nodules is interpreted to be due, at least in part, to a reduction in the normal mammary cell content of nodules. Consistent with this interpretation is the observation that cultured nodules gave rise to mammary outgrowths that were predominantly hyperplastic, whereas noncultured nodules generated outgrowths with varying proportions of hyperplastic and normal ductal mammary tissue. This investigation was supported by National Cancer Institute Grant CA-17862.     

Relationship Between Organization of Mammary Tumors and the Ability of Tumor Cells to Replicate Mammary Tumor Virus and to Recognize Growth-Inhibitory Contact Signals In Vitro

Mammary tumor virus (MTV) replication was confined primarily to cells organized as acini in intact mouse mammary glands. Primary mammary tumors maintained a high degree of acinar organization and cells therein continued to replicate MTV vegetatively. Nonacinar mammary cells, derived by serial transplantation of acinar tumor cells, no longer actively replicated MTV. This suggests that phenotypic differences exist among mammary epithelial cells in their ability to support virus replication, that a fundamental relationship exists between the organization of epithelium for secretion and active virus replication, and that this relationship is not altered as a primary consequence of neoplastic transformation. Mammary epithelial cells from pregnant, non-tumor-bearing, MTV-infected BALB/cfC3H mice or from acinar mammary tumors from a number of mouse strains were grown in primary monolayer cultures. Such cell cultures under the influence of insulin and cortisol exhibited the ability to organize into discrete three-dimensional structures called "domes." MTV replication in such cultures took place primarily in cells within the organized domes. Cells cultured from nonacinar tumors did not exhibit any propensity to organize into domes, nor did they replicate MTV in primary culture. This suggests that the cell organizational requirement for MTV replication observed in vivo is conserved in primary culture. Dome formation is not an effect of virus replication, as cells from uninfected BALB/c animals organized into domes in culture without concomitant MTV replication. Growth-regulating signals, exerted between contiguous cells in cultures of non-MTV-infected mammary epithelium, were not modified by the occurrence of active virus replication nor as a direct consequence of neoplastic transformation. Cells derived from nontumor BALB/cfC3H glands and from spontaneous tumors exhibited cell growth kinetics, saturation densities, and deoxyribonucleic acid synthesis kinetics nearly identical to those of noninfected normal mammary epithelium in primary culture. Cell to cell growth regulatory signals were modified in cultures of nonalveolar tumor cells wherein evidence of overgrowth is documented.     

Fine structure of a murine mammary carcinoma cell line.

A fine structural study was made of cells from the epithelioid MCF-8/5-2A cell line derived from a MuMTV-free D2 transplantable hyperplastic outgrowth. Electron microscopy shows the cells to be truly epithelial with many cell-to-cell junctions and microvilli. The cells are similar in many respects to normal mouse mammary gland and some of the conventional mammary tumor derived cell lines. This study supports previous observations of the absence of MuMTV in MCF-8 within the limits of morphological detection, and demonstrates the presence of numerous virus particles within, or budding into, cisternae of the endoplasmic reticulum and nuclear envelope. These intracisternal A particles have not been previously described in such abundance in mammary tumor tissue culture cells.     

Chemical carcinogen-mouse mammary tumor virus interactions in cell transformation

We have studied the process of mammary cell transformation in vitro using a single cell clone (Clone 18) from a presumptive epithelial cell line, C57MG, derived from a normal mammary gland; a mouse mammary tumor virus (MMTV) host-range variant (RIII)vp4; and the potent initiating carcinogen 7,12-dimethylbenz(a)anthracene (DMBA). After several serial subcultures, cells treated with virus and then with carcinogen exhibited an altered (transformed) morphology, a dramatic increase in anchorage independence, an increase in multinucleation after exposure to cytochalasin B, an enhanced ability to proliferate in low Ca2+ (0.01 mM) medium, and tumorigenicity when inoculated subcutaneously into athymic (nude) mice. Although some of these phenotypic alterations were observed also in cultures treated singly with MMTV or DMBA and in cultures exposed to DMBA before infection with MMTV, enhanced cytochalasin B multinucleation and tumorigenicity were properties observed only in mass cultures of cloned cells first infected with MMTV and then exposed to DMBA. This demonstrates for the first time that exposure of presumptive mammary epithelial cells to MMTV followed by DMBA, but not to either agent alone or to DMBA followed by MMTV, results in malignant transformation of these cells.     

Fine structure of a murine mammary carcinoma cell line

Summary A fine structural study was made of cells from the epithelioid MCF-8/5-2A cell line derived from a MuMTV-free, D2 transplantable hyperplastic outgrowth. Electron microscopy shows the cells to be truly epithelial with many cell-to-cell junctions and microvilli. The cells are similar in many respects to normal mouse mammary gland and some of the conventional mammary tumor derived cell lines. This study supports previous observations of the absence of MuMTV in MCF-8 within the limits of morphological detection, and demonstrates the presence of numerous virus particles within, or budding into, cisternae of the endoplasmic reticulum and nuclear envelope. These intracisternal A particles have not been previously described in such abundance in mammary tumor tissue culture cells. This study was supported by Contract NIH-NCI-E-71-2421, with the NIH and by an institutional grant to the Michigan Cancer Foundation from the United Foundation of Detroit, Michigan.     

Epithelial-Stromal Interactions in Human Breast Cancer: Effects on Adhesion,Plasma Membrane Fluidity and Migration Speed and Directness

Interactions occurring between malignant cells and the stromal microenvironment heavily influence tumor progression. We investigated whether this cross-talk affects some molecular and functional aspects specifically correlated with the invasive phenotype of breast tumor cells (i.e. adhesion molecule expression, membrane fluidity, migration) by co-culturing mammary cancer cells exhibiting different degrees of metastatic potential (MDA-MB-231>MCF-7) with fibroblasts isolated from breast healthy skin (normal fibroblasts, NFs) or from breast tumor stroma (cancer-associated fibroblasts, CAFs) in 2D or 3D (nodules) cultures. Confocal immunofluorescence analysis of the epithelial adhesion molecule E-cadherin on frozen nodule sections demonstrated that NFs and CAFs, respectively, induced or inhibited its expression in MCF-7 cells. An increase in the mesenchymal adhesion protein N-cadherin was observed in CAFs, but not in NFs, as a result of the interaction with both kinds of cancer cells. CAFs, in turn, promoted N-cadherin up-regulation in MDA-MB-231 cells and its de novo expression in MCF-7 cells. Beyond promotion of “cadherin switching”, another sign of the CAF-triggered epithelial-mesenchymal transition (EMT) was the induction of vimentin expression in MCF-7 cells. Plasma membrane labeling of monolayer cultures with the fluorescent probe Laurdan showed an enhancement of the membrane fluidity in cancer cells co-cultured with NFs or CAFs. An increase in lipid packing density of fibroblast membranes was promoted by MCF-7 cells. Time-lapsed cell tracking analysis of mammary cancer cells co-cultured with NFs or CAFs revealed an enhancement of tumor cell migration velocity, even with a marked increase in the directness induced by CAFs.Our results demonstrate a reciprocal influence of mammary cancer and fibroblasts on various adhesiveness/invasiveness features. Notably, CAFs'' ability to promote EMT, reduction of cell adhesion, increase in membrane fluidity, and migration velocity and directness in mammary cancer cells can be viewed as an overall progression- and invasion-promoting effect.     

Caveolin-1-Deficient Mice Have An Increased Mammary Stem Cell Population with Upregulation of Wnt/?-Catenin Signaling

Here, we show that a caveolin-1 (Cav-1) deficiency leads to an amplification of the adult mammary stem cell population, both in vivo and in vitro. First, the expression of two stem cell markers, Sca-1 and Keratin 6, is dramatically increased in the hyperplastic mammary ducts of Cav-1 deficient mice, suggesting that loss of Cav-1 induces the accumulation of a progenitor cell population in the mammary gland. To independently validate these results, we reconstituted mammary acini formation in vitro via a 3D Matrigel assay system--using primary cultures of mammary epithelial cells derived from WT and Cav-1 deficient mice. We show that Cav-1 null 3D epithelial structures display an intense increase in the expression of three stem cell markers, i.e., Sca-1, keratin 6 and keratin 5. Overall, we observed a 2-to-3 fold increase in the number of Cav-1 KO acini that are positive for a given stem cell marker. Also, we show that such amplification of progenitor cells has functional consequences, as demonstrated by the abnormal presence of myoepithelial cells in the hyperplastic lesions of Cav-1 deficient mammary glands. Finally, we provide evidence that hyper-activation of Wnt/?-catenin signaling may constitute one of the down-stream mechanisms leading to mammary stem cell accumulation. The longevity and slow-dividing properties of mammary stem cells facilitates the accumulation of genetic alterations, and renders these progenitor cells the likely precursors of malignant derivatives. As such, we propose that loss of Cav-1 induces the accumulation of mammary stem cells, and that this event may be an initiating factor during mammary tumorigenesis.     

Chemical carcinogen-mouse mammary tumor virus interactions in cell transformation

Summary We have studied the process of mammary cell transformation in vitro using a single cell clone (Clone 18) from a presumptive epithelial cell line, C57MG, derived from a normal mammary gland; a mouse mammary tumor virus (MMTV) host-range variant (RIII)vp4; and the potent initiating carcinogen 7,12-dimethylbenz(a)anthracene (DMBA). After several serial subcultures, cells treated with virus and then with carcinogen exhibited an altered (transformed) morphology, a dramatic increase in anchorage independence, an increase in multinucleation after exposure to cytochalasin B, an enhanced ability to proliferate in low Ca²⁺ (0.01 mM) medium, and tumorigenicity when inoculated subcutaneously into athymic (nude) mice. Although some of these phenotypic alterations were observed also in cultures treated singly with MMTV or DMBA and in cultures exposed to DMBA before infection with MMTV, enhanced cytochalasin B multinucleation and tumorigenicity were properties observed only in mass cultures of cloned cells first infected with MMTV and then exposed to DMBA. This demonstrates for the first time that exposure of presumptive mammary epithelial cells to MMTV followed by DMBA, but not to either agent alone or to DMBA followed by MMTV, results in malignant transformation of these cells. Support for these studies was provided in part by the National Cancer Institute, Bethesda, Md., Contract N01-CP-01018.     

Two-dimensional gel analysis of polypeptides from normal,preneoplastic and neoplastic mouse mammary tissues

Summary High-resolution two-dimensional polyacrylamide gel electrophoresis (PAGE) was employed to reveal tumor-associated polypeptide changes, using the BALB/c C4 line mouse mammary model system, for which phenotypic and immunogenic alterations accompanying tumor progression are well defined. In the first set of experiments, polypeptide patterns from 20 µg whole tissue lysates of normal mammary gland, C4 preneoplastic hyperplatic alveolar nodule outgrowth (HAN) and spontaneous tumor from C4 HAN were compared. In order to normalize for differential cellularity and extracellular protein content in the whole tissues, our analysis included polypeptide patterns from serum, increased concentration of protein from whole normal mammary gland, and primary cultures of epithelial cells from normal gland, HAN and tumor. Using a computer-based image-analysis system, 90 polypeptides were identified in C4 tumor that were absent in C4 HAN, normal mammary gland and serum. None of the 90 polypeptides could be shown to represent a definite qualitative change in the protein composition of tumor epithelium as they were found to be either present in a higher concentration of protein from whole normal gland, or present in the primary epithelial culture from HAN, or absent in the primary epithelial culture from tumor.Conversely in the second set of experiments, when epithelial cultures were used as the starting point for comparisons to locate tumor-associated polypeptides, none of the 15 polypeptides that were present in cultures from three different tumors, and absent in the culture from normal mammary gland was specific to C4 tumor, as they were present in whole tissues of normal gland.Thus our experimental approach detected significant quantitative but no qualitative polypeptide changes in whole tumor tissue, or in tumor-derived epithelial cell cultures. This finding may reflect the limitations of the two-dimensional PAGE method, and warrants caution in the use of such gel analysis alone to identify tumor-associated proteins.Supported by NIH grant CA42522     

Telomerase activation in mouse mammary tumors: lack of detectable telomere shortening and evidence for regulation of telomerase RNA with cell proliferation.

Activation of telomerase in human cancers is thought to be necessary to overcome the progressive loss of telomeric DNA that accompanies proliferation of normal somatic cells. According to this model, telomerase provides a growth advantage to cells in which extensive terminal sequence loss threatens viability. To test these ideas, we have examined telomere dynamics and telomerase activation during mammary tumorigenesis in mice carrying a mouse mammary tumor virus long terminal repeat-driven Wnt-1 transgene. We also analyzed Wnt-1-induced mammary tumors in mice lacking p53 function. Normal mammary glands, hyperplastic mammary glands, and mammary carcinomas all had the long telomeres (20 to 50 kb) typical of Mus musculus and did not show telomere shortening during tumor development. Nevertheless, telomerase activity and the RNA component of the enzyme were consistently upregulated in Wnt-1-induced mammary tumors compared with normal and hyperplastic tissues. The upregulation of telomerase activity and RNA also occurred during tumorigenesis in p53-deficient mice. The expression of telomerase RNA correlated strongly with histone H4 mRNA in all normal tissues and tumors, indicating that the RNA component of telomerase is regulated with cell proliferation. Telomerase activity in the tumors was elevated to a greater extent than telomerase RNA, implying that the enzymatic activity of telomerase is regulated at additional levels. Our data suggest that the mechanism of telomerase activation in mouse mammary tumors is not linked to global loss of telomere function but involves multiple regulatory events including upregulation of telomerase RNA in proliferating cells.     

ω-3 and ω-6 fatty acids and PGE2 stimulate the growth of normal but not tumor mouse mammary epithelial cells: Evidence for alterations in the signaling pathways in tumor cells

The direct effect of ω-3 and ω-6 fatty acids on the proliferation of mouse mammary tumor cells (MTC) was examined in a serum-free cell culture system. While the EGF-induced proliferation of normal mammary epithelial cells was shown to be enhanced by ω-3 and ω-6 fatty acids and prostaglandins (PGs), a majority (75–80%) of primary mammary tumors were not stimulated by these agents. Compared to normal cells, some MTC cultures showed a higher susceptibility to inhibition by ω-3 fatty acids. The general lack of response of MTC cultures to PGE₂ and cyclic adenosine monophosphate (cAMP) suggests some alterations in the cAMP-mediated pathway. However, the PGE₂-induced cAMP levels and cAMP-dependent protein kinase (PKA) activities in the tumor cells were comparable to normal cells. We conclude that the proliferation of mammary tumor cells either follow a cAMP-PKA-independent pathway or have some alterations in the serine/threonine kinase mediated signaling pathway.     

New cell line. Epithelial cell line and subline established from premalignant mouse mammary tissue.

A cell line and subline with epithelial characteristics were established from mouse mammary hyperplastic alveolar nodules (HAN). The cells do not grow in suspension cultures in vitro or form tumors in vivo. The cells do produce significant amounts of C-type and A-type virus and low amounts of plasminogen activator.     

COMPARISON OF THE GROWTH AND DIFFERENTIATION OF EPITHELIAL CELLS FROM NORMAL AND HYPERPLASTIC LENSES OF THE CHICK: STUDIES OF IN VITRO CELL CULTURES

Epithelial cells from hyperplastic lenses of a strain of chicks (Hy-1) selected for high growth rate were dissociated and cultured in vitro and compared with lens epithelial cells from a normal strain (N) in similar conditions. The hyperplastic lens cells showed remarkable motility and adhesiveness after dissociation and formed cell aggregates of various sizes before attaching to the substrate, giving a rather low plating efficiency. The lens structures (lentoid bodies) developed in partially confluent cultures of Hy-1 cells at least three days earlier than those in the cultures from normal control cells, in which the lens structures developed only after the cultures reached confluence. The results of culture at low cell density showed that the Hy-1 cell population consisted of at least two cell types different from each other in growth capacity. These striking differences in in vitro behaviour of dissociated cells from normal and hyperplastic lens epithelia and the results of clonal culture are discussed in relation to the possible mechanisms of abnormal morphogenesis and growth which are likely to be involved in the development of the hyperplastic lens in situ .     

The role of tumor-derived cytokines on the immune system of mice bearing a mammary adenocarcinoma. I. Induction of regulatory macrophages in normal mice by the in vivo administration of rGM-CSF

Using a dimethylbenzanthracene-induced immunogenic nonmetastatic murine mammary adenocarcinoma in BALB/c mice, our previous work has shown that splenocytes from tumor bearers have reduced responses to both mitogens and Ag including tumor-associated Ag. NK and cytotoxic T cell activities are also reduced in splenocytes of tumor bearers. Mac-1+2+ macrophages induced in mammary tumor bearers are capable of down-regulating lymphocyte responses to mitogens and tumor-associated Ag by cell to cell contact interaction and increased PGE2 production. We have found that the tumor constitutively releases a granulocyte-macrophage (GM)-CSF-like factor in vivo and in vitro, which may be responsible for the systemic increase in cells of the macrophage lineage in tumor-bearing mice. A tumor cell line established from the in vivo tumor expresses and releases GM-CSF as shown by Northern and Western blot analyses. Daily i.p. injections for 3 wk of 10,000 U of rGM-CSF into normal mice induces hemopoietic and immunologic alterations similar to those observed in tumor bearers. Mac-1+ and/or Mac-2+ macrophages can also be detected in the spleens and bone marrow of the mice treated with rGM-CSF. Additionally, splenocytes from rGM-CSF-treated mice have reduced responses to mitogens and their peritoneal exudate cells can cause in vitro down-regulation of proliferative responses of lymphocytes from normal mice. The suppression can be partially reversed by the addition of indomethacin to the cultures suggesting that PGE2 may contribute to the effect. rGM-CSF enhances the in vitro release of PGE2 by the spleen, bone marrow, and peritoneal cells of normal mice. These data indicate that the high levels of GM-CSF constitutively produced by the tumor may be responsible for the hemopoietic changes and immunologic alterations observed in tumor-bearing mice.     

Transforming c-Ki-ras mutation is a preneoplastic event in mouse mammary carcinogenesis induced in vitro by N-methyl-N-nitrosourea.

Mouse mammary epithelial cells can be transformed in primary cultures to preneoplastic and neoplastic states when treated with N-methyl-N-nitrosourea (MNU). Mammary carcinomas arising from MNU-induced hyperplastic alveolar nodules (a type of mouse mammary preneoplastic lesion) contained transforming c-Ki-ras genes when examined by the NIH 3T3 focus assay. Hybridization of allele-specific oligonucleotides to c-Ki-ras sequences amplified by the polymerase chain reaction demonstrated the presence of a specific G-35----A-35 point mutation in codon 12 in each of the NIH 3T3 foci as well as the mammary carcinomas. This mutation resulted in the substitution of the normal glycine with an aspartic acid. Furthermore, this mutation in the c-Ki-ras proto-oncogenes was also detected in 9 of 10 hyperplastic alveolar nodules. These results demonstrate that the specific c-Ki-ras mutation is a preneoplastic event in MNU-induced mouse mammary carcinogenesis.     

Role of tumor-derived cytokines on the immune system of mice bearing a mammary adenocarcinoma. II. Down-regulation of macrophage-mediated cytotoxicity by tumor-derived granulocyte-macrophage colony-stimulating factor

Peritoneal elicited macrophages (PEM) from mammary tumor-bearing mice have a decreased capacity to become cytotoxic against syngeneic, allogeneic, and xenogeneic target cells upon in vitro stimulation with LPS, as compared with PEM of normal mice. A regulatory mechanism other than PG release is suggested because the addition of both indomethacin and LPS to macrophage cultures from tumor-bearing mice caused no changes in their cytotoxic capability. Because tumor products have been implicated in the down-regulation of immune responses, we investigated whether pretreatment with supernatants from the tumor cell line DA-3, derived from the in vivo mammary adenocarcinoma D1-DMBA-3, affects the cytolytic capacity of macrophages. This treatment inhibits, in a dose-dependent fashion, the ability of stimulated normal PEM to kill target cells. Partial purification of DA-3 cell line supernatant showed that most of the inhibitory activity was exerted by factors with a molecular mass greater than 10 kDa and less than 30 kDa. However, slight inhibition could also be observed with fractions containing molecules less than 10 kDa. The data suggest that more than one factor released by the mammary tumor cells may be involved in the down-regulation of macrophage-mediated cytotoxicity. Because the DA-3 cells constitutively produce granulocyte-macrophage CSF (GM-CSF), which has a molecular mass of 27 kDa, we pretreated PEM from normal mice in vitro with rGM-CSF for 24 h. This resulted in a dose-dependent decrease in their capacity to kill tumor target cells upon LPS stimulation. Furthermore, PEM from normal mice injected with rGM-CSF for 25 days displayed a profound decrease in their cytolytic ability against DA-3 targets upon in vitro stimulation with increasing amounts of LPS. The pretreatment of PEM from normal mice with a combination of DA-3 cell supernatants and specific anti-GM-CSF partially neutralized the inhibitory effect of the DA-3 supernatant on macrophage tumoricidal capability. These results indicate that tumor-derived GM-CSF is an important factor involved in the decreased macrophage cytotoxicity during mammary adenocarcinoma progression.     

Lipid-like agent from human neoplastic cells suppresses cell-mediated immunity

Summary A lipid-like factor (LLF) of a phospholipid nature was isolated from spent cell-free media of explants and cell cultures from lipids collected from human mammary carcinoma and malignant melanoma cells. LLF modulates macrophage properties and inhibits macrophage chemotactic activity, spreading, lipopolysaccharide-induced tumoricidal activities, and macrophage migration. LLF is unique to tumor cells and is not present in detectable quantities in normal mammary epithelial cells or skin fibroblasts. LLF also inhibits human normal lymphocytes' response to mitogenic stimulation. Partial purification of LLF from human mammary carcinoma is attained by a combination of chloroform extraction and filtration through Amicon molecular membranes. LLF activity is not sensitive to trypsin, pronase, bovine spleen phosphodiesterase II, alkaline phosphatase, or ribonuclease, but it is completely inactivated with phospholipase and lipoprotein lipase.     

Epithelial induction of stromal tenascin in the mouse mammary gland: from embryogenesis to carcinogenesis

The distribution of the extracellular matrix glycoprotein tenascin was studied by immunofluorescence in the developmental history of the mouse mammary gland from embryogenesis to carcinogenesis. Tenascin appeared only in the mesenchyme immediately surrounding the epithelia just starting morphogenesis, that is, in embryonic mammary glands from 13th to 16th day of gestation, in mammary endbuds which are a characteristic structure starting development during maturation of the mammary gland, and in the stroma of malignant mammary tumors. However, tenascin was absent in the elongating ducts of embryonic, adult, proliferating, and involuting mammary glands and preneoplastic hyperplastic alveolar nodules. The transplantation of embryonic submandibular mesenchyme into adult mammary glands induces the development of duct-alveolus nodules, which morphologically resemble developing endbuds. Tenascin reappeared around those nodules during the initial stages of their development. Tenascin expression could be induced experimentally in several ways. First, tenascin was detected at the site where the first mammary tumor cells GMT-L metastasized. Second, tenascin was detected in the connective tissue in the tumors derived from the injected C3H mammary tumor cell line CMT315 into Balb/c nude mouse. Cross-strain marker anti-CSA antiserum clearly showed that the tenascin-positive fibroblasts were of Balb/c origin. Third, when embryonic mammary epithelium was explanted on to embryonic mammary fat pad cultures, the mesenchymal cells condensed immediately surrounding the epithelium. Tenascin was detected in these condensed cells. From these three observations we conclude that both embryonic and neoplastic epithelium induced tenascin synthesis in their surrounding mesenchyme.