v-Ha-ras transgene abrogates the initiation step in mouse skin tumorigenesis: effects of phorbol esters and retinoic acid

Transgenic mouse models that explore the multistep hypothesis of intestinal neoplasia

SV-40 T antigen (TAg), human K-rasVal12, and a dominant negative mutant of human p53 (p53Ala143) have been expressed singly and in all possible combinations in postmitotic enterocytes distributed throughout the duodenal-colonic axis of 1-12-mo-old FVB/N transgenic mice to assess the susceptibility of this lineage to gene products implicated in the pathogenesis of human gut neoplasia. SV-40 TAg produces re-entry into the cell cycle. Transgenic pedigrees that produce K-rasVal12 alone, p53Ala143 alone, or K-rasVal12 and p53Ala143 have no detectable phenotypic abnormalities. However, K-rasVal12 cooperates with SV-40 TAg to generate marked proliferative and dysplastic changes in the intestinal epithelium. These abnormalities do not progress to form adenomas or adenocarcinomas over a 9-12-mo period despite sustained expression of the transgenes. Addition of p53Ala143 to enterocytes that synthesize SV-40 TAg and K-rasVal12 does not produce any further changes in proliferation or differentiation. Mice that carry one, two, or three of these transgenes were crossed to animals that carry Min, a fully penetrant, dominant mutation of the Apc gene associated with the development of multiple small intestinal and colonic adenomas. A modest (2-5-fold) increase in tumor number was noted in animals which express SV-40 TAg alone, SV-40 TAg and K-rasVal12, or SV-40 TAg, K-rasVal12 and p53Ala143. However, the histopathologic features of the adenomas were not altered and the gut epithelium located between tumors appeared similar to the epithelium of their single transgenic, bi-transgenic, or tri-transgenic parents without Min. These results suggest that (a) the failure of the dysplastic gut epithelium of SV-40 TAg X K-rasVal12 mice to undergo further progression to adenomas or adenocarcinomas is due to the remarkable protective effect of a continuously and rapidly renewing epithelium, (b) initiation of tumorigenesis in Min mice typically occurs in crypts rather than in villus-associated epithelial cell populations, and (c) transgenic mouse models of neoplasia involving members of the enterocytic lineage may require that gene products implicated in tumorigenesis be directed to crypt stem cells or their immediate descendants. Nonetheless, directing K-rasVal12 production to proliferating and nonproliferating cells in the lower and upper half of small intestinal and colonic crypts does not result in any detectable abnormalities.

Transgenic mouse model of malignant skin melanoma

Tyr-SV40E transgenic mice are specifically susceptible to melanoma due to expression of the oncogene in pigment cells. Mice of the more susceptible lines die young of early-onset eye melanomas, when skin melanomas are still infrequent and benign. To surmount this obstacle, skin from donors of two high-susceptibility lines was grafted to Tyr-SV40E hosts of a low-susceptibility line of the same inbred strain, thereby enabling the skin to outlive the donors and continue to grow in immunocompetent but tolerant hosts. Unexpectedly, donor pigment cells in all the grafts soon selectively proliferated close to areas of greatest wound healing, forming a dense black tracery, especially at the outer rim of the grafts. These lesions slowly grew radially within the grafts, producing irregular greyish patches. Local vertical thickenings then appeared and developed into small melanomas, which soon ulcerated through the epidermis. The tumors rapidly enlarged and became deeply invasive. Discrete black nevi also arose, with many becoming larger and distinctly blue, but those not near areas of pronounced wound healing did not progress to malignancy. In this first series, malignant melanoma resulted in all the grafts from the more susceptible of two donor lines and in some grafts from the other line. Distant metastases occurred in some cases from each line. Most tumors were hypomelanotic and heterogeneous, with lobes or areas differing in melanization. The results strongly suggest that growth factors and cytokines--known to be produced in wound repair--are triggering the growth and malignant conversion of these genetically susceptible melanocytes and that in the graft situation we are merely witnessing a caricature--a usefully exaggerated manifestation of the true events underlying the genesis of melanomas. The striking resemblance to the human malignancy, the genetic uniformity and different susceptibilities of the transgenic lines, and the experimental possibilities in the grafted mice all make them an excellent model of the disease.

Anoxia, wound healing, VL30 elements, and the molecular basis of malignant conversion

Although VL30 retrotransposable elements have been associated with certain cancers for nearly twenty years, because of their expression in rodent malignancies and recombination into murine sarcoma viruses, their causative role, if any, in cancer has been uncertain and enigmatic. Recent findings suggest loss of normal transcriptional control of specific VL30 element expression may make a critical contribution to tumor progression at a step associated with malignant conversion, by bringing into play a cellular program normally involved in wound healing. This program, the fibroblast anoxic response system, includes an adaptation to glycolytic metabolism, secretion of metalloproteinases, and activation of an endonuclease. While appropriate for facilitating debris removal during wound healing, loss of control of this program in a cell which has already progressed to the benign neoplastic state has the potential to simultaneously produce the invasiveness and genomic instability characteristic of malignancy. Examination of tumors and tumor derived cell lines has confirmed that key aspects of this system are in fact activated in cancer.

Multiple tumor types appear in a transgenic mouse with the ras oncogene

A transgenic mouse strain with the zeta-globin promoter and the vHa-ras oncogene develops an array of mesenchymal and epithelial neoplasms described here. The predominate mesenchymal tumors were dermal spindle cell tumors, which resembled malignant fibrous histiocytomas found in humans. They were associated with hepatosplenomegaly and developed beneath squamous papillomas. The hepatosplenomegaly was associated with infiltrates of cells that tended toward myelocytic or monocytic differentiation. Other epithelial tumors included keratoacanthomas and squamous cell carcinomas. Squamous cysts, some with squamous cell carcinomas, of the salivary glands and mammary carcinomas were also found. Odontogenic tumors, which sometimes differentiated into ameloblastomas, were one of the more unusual tumor types observed. Other, less frequent tumors were also noted. The tumors described here are a potentially valuable experimental resource that may lead to an understanding of malignant fibrous histiocytoma-like lesions, odontogenic tumors, and tumor progression.

Targeting of the rasT24 oncogene to the proximal convoluted tubules in transgenic mice results in hyperplasia and polycystic kidneys

Five families of transgenic mice were derived from one-cell-stage embryos injected with gamma GT-rasT24, a fusion gene consisting of the gamma-glutamyl transpeptidase (gamma GT) 5' flanking region containing promoter I linked to a mutated (codon 12) human H-ras oncogene. The transgene was expressed selectively in the kidneys, eyes, and brains of all families as determined by reverse transcription-polymerase chain reaction, nuclease protection assays, and in situ hybridization. In two of five families, kidney lesions consisting of proximal tubular hyperplasia, renal cysts, and microadenomas developed in male animals; males also expressed higher levels of gamma GT/rasT24 RNA. Early lesions consisted of proximal tubular hyperplasia as defined by alkaline phosphatase histochemistry, gamma GT immunohistochemistry, and electron microscopy and could be correlated with the presence of rasT24 RNA within the cystic proximal tubular epithelium by in situ hybridization. Advanced lesions also involved other segments of the nephron and consisted of cysts lined by a flattened unicellular layer of attenuated epithelium. No rasT24 could be identified within cystic lesions of the distal nephron and collecting tubules by in situ hybridization, and they most likely arise by external compression. Animals from the two transgenic strains exhibiting cystic lesions die of renal failure beginning at 8 months of age. No difference in cell-cycle parameters or DNA ploidy between transgenic and control kidneys was identified by flow cytometric analysis. No renal carcinomas developed. The primary renal effects of the H-rasT24 oncogene in this model system consist of proximal tubular hyperplasia and polycystic kidneys. This model appears to provide a useful in vivo system for the study of ras oncogene function and control of renal cell proliferation.

Rodent models for targeted oncogenesis of the prostate gland

Currently, prostate cancer ranks as the most frequent non-skin malignancy detected in males. Yet, of the major human cancers, it remains one of the least understood in terms of its molecular and genetic basis. Research on prostate cancer has been limited by the paucity of tissues available for study. Much of the tissue obtained through surgery for localized prostate cancer will be required for pathological staging and grading. The more aggressive forms of prostate cancer are usually detected subsequent to metastatic involvement at which point there is little reason to surgically remove the prostate tumor(s). A final complication is the propensity of prostate cancer to metastasize to the bone, a site extremely difficult to obtain suitable biopsies for study. Further hindering research efforts on prostate cancer is the lack of suitable animal models for study. In contrast to its frequent occurrence in humans, prostate cancer is a rare event in most other mammalian species, particularly laboratory rodents. Therefore, in order to make this disease more amenable for study, there is a growing effort to identify or develop a means to target oncogenesis to the prostate gland of rodents. As will be reviewed here, this goal is being approached with the use of 3 different methods; one that takes advantage of the unique androgenic hormone requirement for prostate growth to exaggerate the effects of carcinogens at that site and two methods (recombinant retrovirus transduction prior to organ reconstitution and transgenic targeting) that allow direct genetic manipulation of cells in the prostate gland leading to the development of prostatic malignancy.

ras protein p21 processing enzyme farnesyltransferase in chemical carcinogen-induced murine skin tumors

Farnesylation of ras protein p21 is crucial for the protein's membrane localization, which is essential for its cell-transforming activity, which in turn is thought to be critical for the ultimate induction of cancer. The cytosolic enzyme farnesyltransferase plays a major role in posttranslational modification of p21, but the level of farnesyltransferase activity in mammalian tumors and its relationship to the processing of cytosolic p21 that leads to tumorigenesis are unknown. We report here that farnesyltransferase activity was significantly higher in chemical carcinogen-induced benign skin papillomas in SENCAR mice than in the epidermises of control animals. The enzyme is primarily epidermal in origin, and kinetic studies with cytosol from epidermis and papillomas showed that the reaction was linear with respect to time, substrate concentration, and protein content. Skin papillomas showed significantly elevated levels of both cytosolic and membrane-bound Ha-ras p21, whereas far lesser cytosolic and almost negligible amounts of membrane-bound p21 were present in the epidermis of control mice. There was a positive correlation between increased enzyme activity in papilloma cytosol and the processing of overexpressed cytosolic Ha-ras p21 for its localization to membrane.

Midkine (MK), a retinoic acid (RA)-inducible gene product, produced in E. coli acts on neuronal and HL60 leukemia cells

We have shown previously that (i) retinoic acid (RA), an anti-neoplastic agent, activates the midkine (MK) gene in mammalian embryonic carcinoma cells, and that (ii) the MK of 118 amino acids, purified from L cells, induces neurite outgrowth of mammalian embryonic brain cells. In this paper, we describe an unconventional strategy for the purification of a fully active MK from E. coli with a high yield. The MK was overproduced in E. coli as a glutathione S-transferase (GST) fusion protein. The MK fusion protein extracted from the bacterial inclusion bodies with guanidine-HCl was renatured, refolded slowly and cleaved by thrombin at the site where the GST links to the MK. The purified free MK, like RA, induced neurite outgrowth from central neurons of the mouse spinal cord, and suppressed the growth of human HL60 leukemia cells in vitro. Unlike RA, however, the MK did not induce granulocytic differentiation of HL60 cells. Furthermore, the MK supported the survival of an NGF-insensitive sensory neuron subpopulation(s) from chicken embryo dorsal root ganglion. Thus, the actions of the MK and leukemia inhibitory factor (LIF) are surprisingly similar. There is no sequence similarity between MK and LIF, however, and unlike MK, LIF production does not appear to be RA-inducible.

Retinoic acid receptor and retinoid X receptor expression in retinoic acid-resistant human tumor cell lines

Retinoic acid (RA) has profound effects on cell proliferation and differentiation both in vitro and in vivo. Many human cell lines are known to be sensitive to the growth-inhibitory action of RA. We analyzed established human solid tumor-derived cell lines for their RA sensitivity. Growth inhibition by RA in monolayer was examined by [3H]thymidine incorporation and cell proliferation. Here we report that 11 widely used human cell lines were RA resistant. The majority are carcinoma derived (A-431, BT-20, C-41, ACHN, HCT116, 293, A549, and PA-1); two are sarcoma derived (Saos-2 and A673); and one is a melanoma cell line (A-375). Since nuclear retinoid receptors are implicated in the biological effects of RA, we examined the expression of retinoic acid receptors (RARs) RAR alpha, RAR beta, RAR gamma, and the retinoid X receptors (RXRs) RXR alpha, RXR beta, and RXR gamma in the RA-resistant cell lines by northern blotting and by RNase protection analysis for RAR beta. RAR alpha transcripts were constitutively expressed in all cell lines. By contrast, RAR beta was expressed in only seven RA-resistant cell lines (Saos-2, ACHN, 293, A549, A-375, A673, and PA-1), and its level was enhanced by RA in some cases. In most cell lines, RAR gamma expression was low and was not affected by RA. The RXR genes showed a very distinct expression pattern in the group of selected cell lines. In general, RXR alpha was the most abundantly expressed subtype, RXR beta was expressed at low levels, and RXR gamma could not be detected. In none of the RA-resistant cell lines was RXR expression modulated by RA. The results presented here indicate that the resistance of these human tumor cell lines to RA cannot be simply correlated with expression of RAR or RXR or both.

Strategy of research for cancer—chemoprevention

It has been reported that environmental chemicals are important factors in terms of both development and prevention of human cancer. For the latter, detection of early stages is an essential first step followed by clinical trials for surveying populations at risk. Thus a great deal of attention has been focused on these areas. However, investigations of possibilities for active prevention of cancer development itself form another major project. Chemoprevention of carcinogenesis, which means prevention of carcinogenesis by exogenous chemical compounds, has been investigated extensively in a variety of organs in animal models. Usually attention is concentrated on only one organ. However, antioxidants, such as BHA, exert very different effects on different organs, suggesting the necessity of whole body approaches to the question of chemoprevention. Furthermore, the mechanisms of chemoprevention, including the step of carcinogenesis, i.e., initiation, promotion, progression or whole carcinogenesis steps, in which exogenous compounds exert their protective effects, should be considered. A medium-term bioassay system and a multi-organ carcinogenesis system, which can be used for investigation of potential for cancer chemoprevention, have been developed in our laboratory. Dose dependent inhibitory effects were established for both BHA and alpha-tocopherol in the medium-term bioassay system, and inhibition of small intestinal carcinogenesis by catechins in green tea has also been investigated in our multi-organ carcinogenesis protocol. It is extremely important for prevention of human cancer that we find new candidates for chemopreventive agents using animal studies. This paper reviews published reports on chemoprevention, taking into account effective stages, and proposes suitable experimental animal models for future investigations in this increasingly important area.

Transgenic models of tumor development

Numerous cancer-prone strains of mice have been created by the introduction of candidate tumor-promoting genes into fertilized eggs. Each transgenic strain is predisposed to develop specific types of tumors, but they usually arise stochastically because of the need for spontaneous mutation of genes that collaborate with the introduced oncogene. These mice are providing insights into the effects of individual oncogenes on cellular proliferation, differentiation, and viability, as well as on oncogene cooperativity. Their predisposed state imposes sensitivity to viral and chemical carcinogenesis, and the mice should prove valuable in tests of potential carcinogens, therapies, and preventive measures.