Stages of intestinal carcinogenesis and their control by host NK cells. Implications to prevention and therapy.

DMH-induced (25mg/kg/wk) rodent intestinal carcinogenesis was re-examined using histometry under various conditions including modulation of host natural killer (NK) cell activity and mutagen administration. Early lesion was enterocyte hyperplasia all along the intestinal tract. Cell kinetic analysis showed that it was caused by a slightly altered but still functional "initiated" enterocyte population added to the normal population which remained unchanged. Primarily, initiated stem (IS) cells were added which then produced their initiated progeny which still renewed normally. NK cells countered the initiated cells selectively, either inhibiting IS cell proliferation or killing them when activated by lymphokines. Evidence was obtained that the IS cells could further transform into preneoplastic stem (PS) and then neoplastic stem (NS) cells under the influence of promoters and mutagens, respectively. Subsequent transformation of normal stem cells into IS, PS, and NS cells apparently is the basis of carcinogenesis. These produce lesions, hyperplastic, preneoplastic, and neoplastic, respectively, only when and where NK activity is inhibited. Under the influence of normal NK cells, they remain dormant (nonproliferative), compatible with normal life.

Initiated stem cells in murine intestinal carcinogenesis: prolonged survival, control by NK cells, and progression.

Weekly injections of dimethylhydrazine (DMH) (25 mg/kg), or azoxymethane (AOM) (8 mg/kg) to young adult male CDI mice for 1-2 months produced generalized intestinal crypt hyperplasia, which we measured in duodenum in terms of number of interphase and mitotic cells present in crypts. As shown earlier, the crypts expanded because of the presence of a hyperproliferative "initiated" crypt subpopulation which was also sensitive to natural killer (NK) cells. Hyperplasia was thus present as long as NK activity was suppressed by the carcinogen treatment. After interruption of the treatment for periods of 1, 2, 3, 6 and 10 months in the various groups, hyperplasia soon regressed as a result of elimination of the subpopulation by the recovering NK cells. When NK activity was once again eliminated during the terminal days of these "interruption periods" (by injections of anti-asialo GM-I antibody, alpha AGM-I), the original hyperplasia was fully reconstituted, apparently from stem cells of the subpopulation which survived up to 10 months in their crypt base location. These "initiated stem cells" represented, then, the original carcinogenic insult during the pre-cancerous period. They also appeared to be the source of the eventual neoplasia, as treating the animals with mutagens during the interruption periods produced specific changes in crypt base histology: new "crypt base basophilic" (CBB) cells appeared which produced large accumulations as well as microscopic tumors when NK activity was suppressed (by alpha AGM-I). Some of the initiated stem cells were apparently transformed into neoplastic ones which remained under NK control, the NK cells preventing the establishment of their progeny. Further experiments indicated that, although the initiated stem cells are not eliminated by normal NK activity, activated NK cells can kill them, thereby eliminating the potential source of neoplasia.

Hyperplasia of mouse duodenal crypts and its control by NK cells during the initial phase of DMH carcinogenesis.

The possible regulatory role of NK cells on early events in chemical carcinogenesis remains undefined. The present study examined whether NK cells control 1,2-dimethylhydrazine (DMH)-induced hyperplasia of the duodenal crypt in CD1 mice. Mice receiving chronic DMH treatment showed a dose-dependent hyperplasia confined to the proliferative zone, with a parallel increase in mitotic and 3H-TdR-labelled cells and significant suppression of splenic NK activity. Complete ablation of splenic NK activity with anti-asialo GM-I antibody (alpha AGM-I) treatment slightly enhanced hyperplasia. Halving of the DMH dose for 2 weeks led to regression of hyperplasia, which was totally prevented by alpha AGM-I treatment. The alpha AGM-I treatment alone did not influence crypt size in normal mice. Finally, a stimulation of NK activity with Poly I:C treatment in DMH-treated mice caused regression of the DMH-induced hyperplasia. Our results suggest that hyperplastic cells with possible genetic alterations induced by the carcinogen express target structures for NK cells, but that simultaneous carcinogen-induced suppression of NK activity hampers their containment, allowing progression of hyperplasia to neoplasia, possibly owing to additional genetic changes.

Behavior of fetal intestinal organ culture explanted onto a collagen substratum.

A model of organ culture of 18 day old fetal rat intestine (Quaroni, 1985) was modified and characterized in the present work with the purpose of developing an in vitro model for the study of intestinal epithelial cell behaviour. Fragments of this intestine were kept in suspension culture for 7 days and then explanted onto collagen (type I) matrix. Within a day, the fragments became anchored to the substratum and a circular monolayer grew out to about 1 cm diameter. In the fragments, an outer layer of absorptive epithelial cells came to enclose a stroma, which was polarized into a loose (mesenchymal) and a dense portion. The dense portion contained a mixture of smooth muscle cells and primitive stem-type epithelial cells ('p-cells'). After explantation, at the contact point with the matrix, the epithelium broke up and the mesenchyme grew into the matrix and anchored the fragment. The epithelial edges now became continuous with the developing monolayer. Radioautography with tritiated thymidine indicated a constant cell renewal in epithelium and monolayer apparently from foci of p-cells, a reserve population of which was seen to be sequestered among the smooth muscle cells. Activated stem cells could differentiate into three mature epithelial phenotypes, each differentiation pathway apparently being determined by the type of underlying stroma. Immunohistochemistry using gold- and fluorescein-labeled monoclonal antibodies indicated that adult differentiation-specific markers (e.g. brush border enzymes) were present in the fragment epithelium but not in the monolayer cells. On the other hand, the monolayer cells could be induced to express some of these markers by contact with mesenchymal cells or by co-culturing with fibroblastic cell lines. Matrigel substratum mixed with collagen (type I) supported the appearance in monolayer of strands positive for amino-peptidase and lactase. The model thus appears to be suitable for the in vitro study of epithelial renewal and differentiation, and it has already provided some results in this respect.

A gradual decrease in nucleolar size with the maturation of columnar epithelial cells in the adult rat intestine under normal and various experimental conditions.

The columnar cells, which form over 90% of the epithelium in the small intestine, undergo rapid and continuous renewal and maturation. Samples from duodenum, jejunum, upper, mid- and terminal ileum of young male rats were processed for histology. The average maximal nucleolar area was determined in 10-cell-wide bands of the basal, mid and upper levels of crypts and villi, respectively, by image analysis; in the duodenum, it was 2.8, 2.1, 1.7, 1.5, 1.3 and 0.8 (in micron2) in the respective epithelial levels from crypt base to villus top. Although villus size decreased by 68% from duodenum to terminal ileum, nucleolar area was similar at each respective epithelial level along the intestine. This indicated that nucleolar size was related to cell maturity, rather than to the size of epithelium. In other groups of rats, the duodenum was examined after administering specific inhibitors. Methotrexate (within a day) and cycloheximide (within 3 h) did not significantly affect nucleolar size, indicating that the decrease in size was not under the influence of immediate synthesis of nucleic acid or protein. On the other hand, tunicamycin (within a day) delayed the decrease and actinomycin D (within 3 h) caused a maximal decrease in all nucleoli. This implied that a glycoprotein factor and some changes in DNA were involved in the decrease in nucleolar size. The rate of protein synthesis in duodenum was then measured by grain count per cell area in autoradiographs made after 1 h of injection of [3H]leucine. From crypt base to villus base, the grain count doubled while the nucleoli decreased to nearly half of their size in the crypt base. When actinomycin D injection preceded the [3H]leucine administration, all nucleoli decreased markedly and the grain counts increased by about 30% in all epithelial levels. It thus appears that the decrease in nucleolar size stimulates protein synthesis, possibly by the release of ribosomal material or some other factor. Protein synthesis in turn has been shown to be related to cell maturation. It is concluded that the nucleolus is involved in some manner in the regulation of the maturation and renewal of the epithelial cells.

Effects of 1,2-dimethylhydrazine on the number of epithelial cells present in the villi, crypts, and mitotic pool along the rat small intestine.

Young adult male rats received 1,2-dimethylhydrazine (25 mg/kg) twice weekly for 2 months and once a week thereafter for up to 6 months. Histological samples of duodenum, jejunum, and upper, mid-, and terminal ileum were prepared from groups killed at each month. Using cell counts, the average number of epithelial cells was determined per representative section of villi and crypts and was used as an index of villus size and crypt size, respectively. The average number of mitotic figures in representative crypt sections was also determined. All three parameters increased during treatment, and the increments showed a specific pattern in relation to time and intestinal region. Mitotic number showed a consistent change all along the small intestine: close to 20% rise by 3 months; decrease to near control level by the fourth month; and a rise thereafter. Probably, a systemic stimulation of mitotic activity by 1,2-dimethylhydrazine took place. The crypt size index changed similarly, showing highly significant correlation with mitotic number. This correlation indicated that average mitotic time and cell cycle time remained unchanged and the number of divisions increased in progenitor cells. Calculations showed that only a fraction of the progenitor cells was involved. These were probably "initiated" cells. There was a net increase of initiated cell numbers with time, but a sharp drop at 4 months indicated that there is a mechanism inducing a regression of the initiated cell population. Villus size increased linearly in the duodenum and jejunum. In the ileum, there was also a net increase but with some initial fluctuation. In general, villus size seemed to increase so as to maintain a fairly stable turnover time. This would mean that the increased mitotic activity was balanced by increasing differentiation.

Radioautographic visualization of differences in the pattern of [3H]uridine and [3H]orotic acid incorporation into the RNA of migrating columnar cells in the rat small intestine.

The epithelium of rat small intestine was radioautographed to examine whether RNA is synthesized by the salvage pathway as shown after [3H]uridine injection or by the de novo pathway as shown after [3H]orotic acid injection. The two modes of RNA synthesis were thus investigated during the migration of columnar cells from crypt base to villus top, and the rate of synthesis was assessed by counting silver grains over the nucleolus and nucleoplasm at six levels along the duodenal epithelium--that is, in the base, mid, and top regions of the crypts and in the base, mid, and top regions of the villi. Concomitant biochemical analyses established that, after injection of either [5-3H]uridine or [5-3H]orotic acid: (a) buffered glutaraldehyde fixative was as effective as perchloric acid or trichloracetic acid in insolubilizing the nucleic acids of rat small intestine; (b) a major fraction of the nucleic acid label was in RNA, that is, 91% after [3H]uridine and 72% after [3H]orotic acid, with the rest in DNA; and (c) a substantial fraction of the RNA label was in poly A+ RNA (presumed to be messenger RNA). In radioautographs of duodenum prepared after [3H] uridine injection, the count of silver grains was high over nucleolus and nucleoplasm in crypt base cells and gradually decreased at the upper levels up to the villus base. In the rest of the villus, the grain count over the nucleolus was negligible, while over the nucleoplasm it was low but significant. After [3H]-orotic acid injection, the number of silver grains over the nucleolus was negligible at all levels, whereas over the nucleoplasm the number was low in crypt cells, but high in villus cells with a peak in mid villus. The interpretation is that, except for a small amount of label incorporated into DNA from either precursor by crypt cells, the bulk of the label is incorporated into RNA as follows. In the crypts, cells make almost exclusive use of uridine, that is, of the salvage pathway, for the synthesis of ribosomal RNA in the nucleolus and of messenger and transfer RNA in the nucleoplasm. However, when cells pass from crypt to villus, they mainly utilize orotic acid--i.e., the de novo pathway--for the synthesis of messenger and transfer RNA within the nucleoplasm.

Morphometric study of the rat duodenal epithelium during the initial six months of 1,2-dimethylhydrazine carcinogenesis.

Male Wistar rats weighing 100 g received 1,2-dimethylhydrazine (25 mg/kg) s.c.) twice a week for 2 months and once a week thereafter for an additional 4 months. Groups of four to six rats were sacrificed monthly. Paraffin sections of duodenum were prepared and stained with periodic acid-Schiff and hematoxylin for cell counts and with toluidine blue for measuring nucleolar area. As an index of villus size and crypt size, the mean number of epithelial cells per representative sections of villi and crypts were used. Mitotic activity was assessed by counting the mean number of mitotic figures per representative crypt section. Nucleolar area was assessed from the analysis of drawn (camera lucida) images of nucleoli of columnar cells at six levels of the epithelium: lower, mid, and upper parts of the crypts and villi. Villus size increased progressively during the 6-month treatment, from 272 +/- 2 (S.E.) to 349 +/- 8. Crypt size increased from 118 +/- 2 in a wavy fashion, showing maximum (139 +/- 5, 143 +/- 2) at 3 and 6 months and minimum (123 +/- 3, 127 +/- 1) at 1 and 4 months. Mitotic number displayed a similar pattern of increase so that the percentage of mitotic figures in the crypts (mitotic index) remained constant (about 5%) in control and experimental animals. Nucleolar area in the controls decreased with age from 4.2 +/- 0.08 sq micron in lower crypt at 1 month to 2.8 +/- 0.04 sq micron at 6 months. During 1,2-dimethylhydrazine treatment, lower crypt nucleoli increased to 4.5 +/- 0.12 sq micron after 3 months and decreased slightly thereafter, reaching 4.0 +/- 0.14 sq micron by the sixth month. The nucleoli furthermore decreased gradually along epithelium (nucleolar compaction) by an average of 0.23% per cell position in control as well as treated animals. It appeared, then, that the main effect of 1,2-dimethylhydrazine was the enlarging of the four parameters measured. This effect seemed to relate in some manner to tumor formation as all the enlargements were attenuated in intestinal tissue adjacent to tumors.

Morphological observations on mucus-secreting nongoblet cells in the deep crypts of the rat ascending colon.

In the ascending colon of adult male rats, large accumulations of differentiated, mucous-type cells were observed which formed the deep one-half to one-third of the crypts almost exclusively. Mitotic activity was localized to the midcrypt, to columnar-type cells. The mucous-type cells were designated as "deep crypt secretory" (DCS), and their histological study was initiated. In the light microscope, the apical cytoplasm was distended with mucous vacuoles which stained differently from goblet cells after using periodic acid-Schiff (PAS), alcian blue, or a Masson stain. In electron-micrographs, a well-developed Golgi complex and rough edoplasmic reticulum in addition to a large number of electron-opaque vacuoles indicated active production of a glycoprotein. The midcrypt mitotic activity was further demonstrated by using the colchicine technique and 3H-thymidine radioautography. A gradual increase of mucus content and a decrease of nuclear and nucleolar size in the DCS cells from midcrypt to crypt base indicated that these cells originate in midcrypt and mature as they progress toward the crypt base. Cell counts showed that the number of cells constituting the crypts was about the same all along the colon, but the percentage of DCS cells varied: 33% in ascending and 21% in transverse colon. Only 5% and 8% deep-crypt mucous-type cells were found in descending colon and cecum, respectively; but these cells differed in appearance from the mature DCS cells. In conclusion, the presence of a large number of nonmitotic end cells with intense secretory activity has been ascertained for the deep crypts of the ascending colon. The exact relation of these cells to the vacuolated or other epithelial cell types remains for further study.

Changes in the size and structure of the nucleolus of columnar cells during their migration from crypt base to villus top in rat jejunum.

An image analyser was used to measure the area of the nucleolus and its component parts in columnar cells at six levels of the jejunal epithelium, corresponding to stages in cell migration from crypt base to villus top. In columnar cells of crypt base, which function as stem cells for the epithelium, the nucleolus is large (3.1 micron2), irregular and reticulated. As cells migrate up the crypt, divide and differentiate, the nucleolus decreases in size (1.7 micron2) and becomes spherical, but remains reticulated. In the fully differentiated cells of the midvillus, however, the nucleolus becomes small (0.9 micron2) and compact. At the villus top, as the cells display early signs of degeneration, the nucleolus is further compacted (0.5 micron2). Most nucleolar components also decrease in size. Pars fibrosa (about 19% of the nucleolar area in crypt base) and pars granulosa (about 70%) decrease in proportion to the rest of the nucleolus, except in mid-villus and villus top where loss of pars granulosa predominates. In contrast, the total area of fibrillar centres remains constant (about 0.1 micron2), even though individual centres are small and numerous in crypt base, larger and fewer at higher levels, and they coalesce into a single structure in villus top. The other nucleolar components are also segregated into distinct, but adjacent, areas at this level. The changes in size and structure of the nucleolus taking place during the migration of columnar cells can be correlated with the maturation of the cells and the loss of their ability to synthesize ribosomal RNA.

Light and scanning electron microscopic observations of the effects of sublethal doses of methotrexate on the rat small intestine.

The small intestines of adult rats were examined by light and scanning-electron microscopy after sublethal doses of methotrexate were injected at 5, 3 and 1 mg, respectively, per rat per day, for three days. Methotrexate inhibited mitosis and thereby disrupted the steady state system of the epithelium. Villi and crypts progressively diminished up to about four and one-half days after the initial injection. Thereafter, recovery began and, by day 7, relatively normal morphology was restored. In the degenerative phase, the loss of crypt-villus continuum was frequently observed, the former crypts forming cyst-like structures. The columnar cells became flat and pleomorphic but epithelial continuity was maintained. Goblet cells apparently decreased in number. Paneth cells, especially in the ileum, appreciably increased in size and number. During the recovery phase, the cystic crypts apparently re-established continuity with the villus epithelium. Size and proportion of all epithelial cell types returned to normal. Scanning electron microscopy showed villus fusion and the cellular pleomorphism and loss of microvilli during the degenerative phase. During recovery of the villi some alteration in orientation and shape remained as shown by scanning electron microscopy.

A region of mitochondrial division in the epithelium of the small intestine of the rat.

Electron microscopic examination of samples from various regions of the rat small intestine was carried out. The number of mitochondria in the epithelial cells was estimated. The counts were made in sections of cells cut along their longitudinal central plane. The errors involved in extrapolating these counts to the whole cells were also estimated. The average mitochondrial number per cell section was 21 in the lower third of the crypts, it gradually increased in the mid and upper thirds and reached about double, 42, at the villus base. The known forms of dividing mitochondria were identified in the mid and upper third of the crypts. The counts remained around 42 along the epithelium of the villi. Crypt cells are continually produced in the lower crypt; these cells migrate to the villi while differentiating into nonproliferative absorptive cells. After inhibiting mitosis by methotrexate, this migration continued (Altmann, '74) and mitochondrial division persisted. In segments of the jejunum isolated surgically from the functional intestine for three weeks, mitosis and cell migration continued, but no evidence of mitochondrial duplication was found. Each mitochondrion probably undergoes a division as the crypt cells migrate from the mid-crypts to the villus. As a result, the villus epithelial cells contain double numbers of mitochondria. It appears that the mitochondrial division is not directly related to mitosis and is elicited by a stimulus present only in the functional intestine.

Morphological effects of a large single dose of cycloheximide on the intestinal epithelium of the rat.

Adult male rats received 15 mg/kg cycloheximide and the subsequent morphological effects at three and six hours after injection were evaluated using histometry, light and electron microscopy, histological demonstration of terminal web and acid phosphatase, and radioautography with tritiated thymidine. Rapid atrophy of the villi took place, progressing from the villus tip by premature exfoliation of epithelial cells. The crypts also diminished by random exfoliation of many crypt cells and by partial or complete disintegration. Mitosis and epithelial cell migration were absent. By six hours, the area occupied by the villi and the crypts per unit length of histological section was decreased by about 70-90% in most of the small intestine but only by about 40-60% in the duodenum and the terminal ileum. In the upper half of the villi, the epithelium was strongly positive for acid phosphatase and contained large numbers of round bodies resembling primary lysosomes. In the lower half, the microvillous border and terminal web were found to be disrupted. Animals receiving only 5 mg/kg cycloheximide also showed the atrophy of villi and crypts, and the round bodies resembling lysosomes. Evidence from several sources has indicated that protein synthesis in normal villus epithelial cells subsides toward the villus tip and becomes minimal at exfoliation. At exfoliation, proteins responsible for epithelial cohesion probably fail because they are no longer replenished. Cycloheximide appears to accelerate this process.