Journal of
~,CancerResearch Clinical Oncology
J. Cancer Res. Clin. Oncoi. 95, 247 264 (1979)
9 Springer-Verlag1979
Early Sequential Lesions During Development of Experimental Gastric Cancer with Special Reference to Dysplasias* E. Kunze 1, A, Schauer, M. Eder, and C. Seefeldt 1 Institute of Pathology, University of G/Sttingen, Robert-Koch-Str. 40, D-M00 G6ttingen and Institute of Pathology, University of Mufiich, Thalkirchnerstr. 36, D-8000 Munich, Federal Republic of Germany
Friihver~inderungen der Carcinogenese an der Magenschleimhaut im Tierexperiment unter besonderer Beriicksichtigung der Dysplasien Zusammenfassung. Es wurde tierexperimentell die Entwicklung des Magencarcinoms in seinen Frtihphasen untersucht und gepriift, welche Rfickschlfisse auf den Cancerisierungsablauf am menschlichen Magen m6glich sind. Nach limitierter oraler Zufuhr von N-Methyl-N'-Nitro-N-Nitrosoguanidin an 174 Ratten entwickelten sich Carcinome direkt aus der sonst unver~inderten Schleimhaut fiber mehrere aufeinanderfolgende Transformationsstadien, ohne gutartig erscheinende proliferative oder neoplastische Epithell~isionen zu durchlaufen. Als erste lichtoptisch erkennbare Ver~inderung ergaben sich fokal Dysplasien G r a d I, die mit zunehmender Versuchszeit in Dysplasien G r a d II und III iibergingen. Die experimentell induzierten Dysplasien sind kausalgenetisch und in ihrem biologischen Verhalten nicht mit der Mehrzahl der am menschlichen Magen auftretenden Dysplasien gleichzusetzen. Die in der Regel im oberen Schleimhautdrittel lokalisierten Dysplasien des Menschen insbesondere des Grades I und II sind gr613tenteils entztindungsbedingt und reversibel. Die experimentellen Dysplasien im Bereich der Proliferationszone sind dagegen als potentiell pdimaligne anzusehen, da sie irreversibel waren und sich progredient weiterentwickelten. Nach den tierexperimentellen Befunden kommen beim Menschen insbesondere die Dysplasien des Grades III als m6gliche Vorl~iufer yon Carcinomen in Betracht, die sich in nicht entziindlicher Schleimhaut im Bereich der Proliferationszone entwickelt haben. Sehliisselwiirter: N-Methyl-N'-Nitro-N-Nitrosoguanidin-Carcinogenese der Magenschleimhaut-Dysplasien - Friihcarcinom * This work was supported in part by a research grant from the Buettner Foundation (G6ttingen) This work was reported in part at the Third International Symposium on Detection and Prevention of Cancer, New York, USA, 1976, and at the 62nd meeting of the Deutsche Gesellschaftfiir Pathologie, Vienna, Austria, 1978 Offprint requests to: Prof. Dr. E. Kunze (address see above) 0171- 5216/79/0095/0247/$ 3.60
248
E. Kunzeet al. Summary. The early sequential development of gastric cancer was studied with experimental animals and examined with respect to what conclusions can be drawn for understanding carcinogenesis in man. After limited oral administration of N-methyl-N'nitro-N-nitrosoguanidine to 174 rats carcinomas developed in most cases directly from the otherwise unchanged mucosa through various successive stages of transformation, without passing through a benignappearing proliferative or neoplastic epithelial lesion. Focal dysplasia grade I was the first recognizable change observed by light microscopy, followed by dysplasia grade II, and subsequently dysplasia grade III. In spite of very similar morphological characteristics, the experimentally induced dysplasias cannot be simply equated in their etiology and biological behavior with the dysplasias of the human stomach. Dysplasias of grade I and II commonly found in man are usually associated with a chronic gastritis; they are located in the upper third of the mucosa and are for the most part reversible. The experimental dysplasias occuring in the proliferative zone of an otherwise undisturbed mucosa must be considered potentially premalignant, as they are irreversible and develop progressively. This finding points out that in man dysplasias grade III within the regenerative zone of non-inflammatory mucosa should be considered particularly as possible precursors of gastric carcinomas. Key words: N-methyl-N;-nitro-N-nitrosoguanidine-Carcinogenesis of gastric mucosa-Dysplasias - Early cancer
The discovery of the carcinogenic effect of N-methyl-N'nitro-N-nitrosoguanidine (MNNG) in 1966 by Druckrey et al., Schoental et al., and Sugimura et al., made it possible to induce carcinomas selectively in the glandular part of the stomach of various laboratory animals, such as mice (Schoental and Bensted, 1970), rats (Sugimura et al., 1969; Saito et al., 1970; Schoental and Bensted, 1970; Shirai, 1974; Tahara et al., 1976), hamsters (Fujimura et al., 1970; Sugimura et al., 1969; Kogure et al., 1974), and dogs (Shimasato et al., 1971; Sugimura et al., 1971; for review of literature see Bralow, 1972; Sugimura and Kawachi, 1973; Wiebecke and Gokel, 1975). Japanese investigators predominately have shown that the experimentally induced gastric carcinomas can develop histogenetically via an adenomatous hyperplasia (Sugimura et al., 1969; Saito et al., 1970; Kohli et al., 1975; Tahara et al., 1976; Uchida et al., 1976; Takahashi et al., 1976; Kobori et al., 1976; for review of literature see Sugimura and Kawachi, 1973; Wiebecke and Gokel, 1975). In the present series of experiments an attempt has been made to clarify whether gastric cancer can also develop directly from a focal atypia of the mucosa without passing through a benign-appearing epithelial proliferation. Of particular interest was whether dysplasias, which are frequently encountered in the human stomach, are also inducible in the rat stomach. AdditionalIy, we sought to learn more about the biological behavior of dysplasias and their prospective significance in the development of gastric carcinoma, and to draw from such animal experiments conclusions regarding carcinogenesis in man.
Experimental Gastric Cancer
249
Material and Methods Two hundred and fifty adult female Wistar rats with an initial weight between 180 and 200 g received MNNG continuously for 250 days (time of exposure) in their drinking water at a dosage of 83 mg/1000 ml and were then discontinued. The aqueous carcinogenic solution was prepared fresh every 2 days. The drinking vessels were covered with black paper to protect the MNNG from degradation by light. To assure continuous observation of the carcinogenic process and to detect earliest morphological changes, the surviving 174 experimental animals were divided into 9 groups and killed after different durations of induction from experimental days 290 to 630 (Table 2). There were 20 animals in each of the first 8 groups, and 14 in the 9th. The subsequent carcinogen-free interval comprised 40 to 380 days. The rats were kept in Macrolon cages and had free access to their drinking water. They received a standardized compressed feed (Atromin) throughout the study. Imrnediately after killing, the stomachs were incised along the lesser curvature and pinned out onto a cork board. After fixation in a 10% formalin solution, a 10 mm long strip of tissue was excised for histological examination from each of six predetermined areas of the stomach: on both sides of the lesser curvature at the fundic region, at the transition of the fundic region to the antrum (both sides), at the antrum, and at the duodenum. After inbedding in paraffin at least 25 5-gm thick serial sections were cut from each specimen. A minimum of 150 sections per animal were examined. AdditionalIy, sections were cut from all tumors. The histological sections were stained routinely with hematoxylin-eosin (H. E.). Other stainings were carried out as needed (PAS, Alcian blue at pH 2.5, van Gieson).
Results
1. Dysplasias Dysplasias were the earliest carcinogenic lesions in the gastric m u c o s a recognizable by light microscopy. F o c a l changes of the m u c o s a that exhibited structural m o d i fications of the glands a n d cellular atypia were designated dysplasias. They were f o u n d exclusively in the a n t r u m a n d frequently occurred in several areas of the same s t o m a c h after longer experimental times. T h e frequency of all dysplasias within the various e x p e r i m e n t a l groups have been included in T a b l e 2. The dysplasias ocurred in a n otherwise u n d i s t u r b e d , i.e. n o n - i n f l a m m a t o r y a n d n o n fibrotic, mucosa. As a rule, dysplasias observed in h u m a n stomachs are localized in the upper third of the mucosa. However, u n d e r experimental c o n d i t i o n s in rats, the i n d u c e d dysplasias were f o u n d to develop at the t r a n s i t i o n between the m i d d l e a n d lower third of the m u c o s a w i t h i n the regenerative zone, from which they could spread in a n apical as well as in a basal direction. I n the present experimental series a total of 112 dysplasias grade I - I I I were registered in 92 affected rats (Table 2). Three grades of dysplasias were distinguished according to their histological a n d cytological features (Table 1). Table 1. Histological and cytological features of dysplasias Dysplasia I
Dysplasia II
Dysplasia III
Nuclei
Slightly irregular arrangement Slight enlargement
Pronouncedly irregular arrangement Marked hyperchromasia
Cytoplasm
Normal appearance
Mitotic activity
Low
Moderately irregular arrangement Moderate enlargement and hyperchromasia Slight to moderate basophilia Moderate
Glands
Pronouced basophilia High
250
E. Kunze et al.
Table 2. Frequency of carcinogen-induced lesions of the gastric mucosa in the different experimental groups (groups 1-8 each consisting of 20 rats, group 9 comprising 14 animals) after limited oral application of MNNG. Stepwise increase of lesions with a higher degree of atypia and malignancy, respectively, with increasing experimental time are indicated Group: Time of induction (days): Time after discontinuation (days): Type of lesion
1
2
3
4
5
6
290 340 390 440 490 40 90 140 190 240
7
8
510 530 600 260 280 350
9
630 380
Number of observed lesions/ Number of affected animals
Totals
Lesions/rats: Dysplasia grade I Dysplasia grade II Dysplasia grade III Early cancer (Mucosa-type) Early cancer (Submucosa-type) Fully developed carcinoma:
1/1 1/11 0 1/1 0 1/1 1/1 0 1/1 0
5/4 5/4 3/3 11/6 0 1/1 I 6/6 8/6 2/2 1/1 2/2 2/2 2/2 0 2/2 1/1
8/7 1/1 8/5 9/7 9/6 5/5 9/6 8/8 7/7 6/6 3/3 5/5
8/6 6/6 4/3 9/7 13/12 4/4
3/3 3/3 9/8 8/5 7/7 4/4
Total lesions :
3
18
44
44
34
0 I 2/2 3/3 0
5
10
26
34
30/26 47/36 35/30 46/36 40/39 20/20
Fig. 1. Dysplasia grade I in the lower third of the antrum mucosa, exhibiting slight cellular atypia. Magnification: ca. 600 x
Experimental Gastric Cancer
251
Fig. 2. Dysplasia grade II in the lower third of the antrum mucosa showing increased cellular atypia with enlargement and hyperchromasia of the nuclei, moderate basophilia of the cytoplasm, and increased mitotic activity. Magnification: ca, 600 x
a) Dysplasias grade I Dysplasias grade I demonstrated minimally to slightly atypically arranged glands with slight dyscaryosis and slightly increased mitotic activity (Fig. 1). The number of animals with dysplasias grade I increased with increased experimental time (Table 2). Only five animals in groups 1-4 had dysplasias grade I, 21 animals in groups 5-9. Twenty-seven of 30 dysplasias grade I did not extend to the surface of the mucosa. Atypical cells were found on the surface in only 3 dysplasias (approximately 1/1o th of all cases). A hyperplasia of the mucosa was not developed in the area of any dysplasia grade I. b) Dysptasias grade II Dysplasias grade II were distinguished from grade I on the basis of an increased glandular distortion and cellular atypia (Fig. 2). Among all exposed animals, a total of 47 dysplasias grade II in 36 affected animals were found (Table 2). They occurred most frequently in groups 5-7. Notably in group 5, several dysplasias grade II were encountered in the same animal. In 38 of the dysplasias the surface epithelium was without any cytological irregularity. In five cases a loss of cellular
252
E. Kunze et al.
Fig. 3. Early cancer of the mucosa type (antrum) with superficial ulceration (right upper corner) Magnification: ca. 375 x
differentiation was noted with a corresponding loss of mucus production. Nine dysplasias (approx. a/s th of the cases) extended to the surface of the mucosa; two of them showed superficial erosions. A slight to moderate hyperplasia of the mucosa occured concomitantly with four of the dysplasias.
c) Dysplasias grade III Dysplasias grade III were characterized by highly distorted, atypically branched glands or tubules in back-to-back arrangement. For the most part they were lined by several layers of markedly atypical cells. The nuclei of these cells were partly hyperchromatic and partly vesicular, varied in size and shape, and frequently exhibited a single prominent nucleolus. The nucleo-cytoplasmic ratio was shifted in favor of the nucleus. The basement membrane was always intact. A total of 35 dysplasias grade III were found in 30 affected animals (Table 2). Their frequency increased with increased total experimental time. Whereas only two dysplasias grade III were found in the first four experimental groups comprising a total of 80 rats, 33 dysplasias occurred in 28 affected among 94 total animals in groups 5 9. In half of the cases, these extended up to the surface of the mucosa. Approximately one-fourth of the dysplasias showed superficial ulcerations. Six dysplasias exhibited concomitantly a moderate hyperplasia of the mucosa. Nine dysplasias which did not extend to the surface showed a loss of mucus production by the surface epithelia.
Experimental Gastric Cancer
253
Fig. 4. Highly differentiated early cancer of the incipient submucosa type, with penetration of the M. mucosae by small groups of atypical glands (antrum). Magnification: ca. 600 x
2. Early Cancer The early cancer in its mucosal and submucosal type represented an important further stage in carcinogenesis of the gastric mucosa (Fig. 3-5). Of 86 observed early cancers 81 were localized in the antrum-pyloric region and only five in the fundic region.
a) Mucosa-Type Forty-six early cancers of the mucosa-type developed in 36 affected animals (Table 2). They were predominately found in groups 6-9. Forty-one carcinomas occurred in the antrum, five in the fundic region. More than half of the carcinomas (27 cases) extended to the surface of the mucosa. Fourteen were ulcerated and corresponded to the growth pattern of type IIc. Histologically, they were welldifferentiated, adeno-tubular carcinomas (intestinal type) with only minimal to slight cell atypia.
254
E. Kunze et al.
Fig. 5. Less differentiated early cancer of the incipient submucosa type with penetration of the M. mucosae (antrum). Magnification: ca. 600 x
b) Submucosa-Type A total of 40 submucosal early carcinomas were found in 39 affected animals. They occurred most frequently in groups 6-9 (Table 2). Two-thirds of the carcinomas extended to the surface of the mucosa. Half of the cases (21) were ulcerated and belonged to growth type IIc. A few could be classified as type III. Thirty-eight were well-differentiated adenocarcinomas (intestinal type), one an undifferentiated carcinoma (anaplastic type), and one a signet-ring cell carcinoma (muco-cellular type).
3. Fully Developed Carcinomas Fully developed carcinomas infiltrating the M. propria were found in 20 cases (Table 2). Nineteen were localized in the antrum, and one in the fundic region. The majority of carcinomas developed in groups 6-9. The carcinoma which occurred in the fundic region could already be observed in the first experimental group. Histologically, there were 12 well-differentiated adenocarcinomas (pylorocardiac and intestinal cell type as well as transitional forms), three poorly
Experimental Gastric Cancer
255
Fig. 6. Circumscribedglandular proliferation (adenomatous hyperplasia) of the antrum extending into the submucosa. Proliferatedglands lined by a single layer of highly differentiated mucus-producing cylindricalcells. Magnification: ca. 240 x
differentiated adenocarcinomas (mainly solid-anaplastic growth pattern) and 1 signet-ring celt carcinoma. Four carcinomas showed a mixed pattern with well and poorly differentiated glandular structures, anaplastic components and a focal monocellular transformation. In a few carcinomas cartilage, bone and/or osteoid were found in the stroma.
4. Submucosal Glandular Proliferations (Adenomatous Hyperplasia) To be distinguished from submucosal early cancer are the sharply demarcated tumor-like foci of glandular proliferations, extending into the submucosa by penetration of the M. mucosae without any signs of an invasive growth (Fig. 6). These foci originated apparently from glands at the base of the mucosa (Figs. 7 and 8). The submucosal glandular proliferations were localized exclusively in the antrum. The glands were lined by a single layer of either columnar mucus producing or cuboidal benign-appearing epithelial cells. Occasionally, focal atypia could be established. The changes we have distinguished as submucosal glandular proliferations are identical to the "adenomatous hyperplasia" of Japanese authors.
256
E. Kunze et al.
Fig. 7. Focal glandular proliferation in the lower third of the antrum mucosa. Magnification: ca. 240 •
The number of the observed submucosal glandular proliferations was not registered in this experimental series. A special variant of the submucosal glandular proliferations were the so-called adenomatous diverticula, found in two animals. They were characterized by an extremly expansive growth in the submucosa, exhibiting centrally a deep diverticular indentation (Fig. 9). 5. Tubulo-Villous Adenomas
In the present experimental series only one tubulo-villous adenoma was found in the antrum of an animal in group 5 (Fig. 10). This tumor consisted of glands and villi with highly differentiated cylindrical cells, which were positive in the PAS and alcian blue reaction, therein being similar to foveolar and pyloric cells. The adenoma showed several loci with marked cellular atypia and increased mitotic activity and therefore must be considered to be undergoing malignant transformation.
Experimental Gastric Cancer
257
Fig. 8. Higher magnificationof the lesion shown in Fig. 7. Magnification: ca. 600 x
6. Different Lesions Pyloric gland metaplasia: Very frequently so-called pyloric gland metaplasias were found in the fundic region. These wedge-shaped mucosal lesions extended occasionally up to the M. mucosae. The specific cells of the gastric mucosa had atrophied in these areas and were replaced by undifferentiated cylindrical or cuboidal cells, some producing mucus. This lesion was designated pyloric gland metaplasia, because of the similarity with pyloric glands (Shirai et al., 1974). They occured very frequently in all groups, and were found more frequently with increased experimental time. We observed a total of 280 pyloric gland metaplasias among 122 animals. In only five cases were we able to demonstrate cellular atypia. In contrast, we encountered intestinal metaplasia rarely. It was found somewhat more frequently in the neighborhood of fully developed carcinomas. Erosions: Of a total of 174 animals 30 had developed gastric mucosal erosions. We found 15 fresh erosions in nine rats, occuring in every test group. Eleven were located in the fundic region, and five in the antrum. Healed erosions were observed in 25 cases in the antrum, again with all the test groups represented. A number of rats demonstrated simultaneously fresh and healed erosions.
258
E. Kunze et al.
Fig. 9. Adenomatous diverticulum of the antrum characterized by extensive, sharply demarcated submucosal glandular proliferations. Magnification: ca. 90 x
Fig. 10. Tubulo-villous adenoma of the antrum consisting of villi and glands lined by highly differentiated, mucus-producing cells. Magnification: ca. 90 x
Experimental GastricCancer
259
Intestinal tumors: A total of 12 tumors were found in the small intestine among
10 rats. Ten tumors were located in the duodenum, and two in the jejunum. Of the duodenal tumors, seven were well-differentiated adenocarcinomas, one an undifferentiated carcinoma and two were fibrosarcomas. The two tumors in the jejunum were an adenocarcinoma and a fibrosarcoma.
Discussion
Gastric carcinomas induced by MNNG can develop histogenetically via an adenomatous hyperplasia as well as directly from a focal atypia of the mucosa. Notably Japanese investigators were able to demonstrate that carcinomas develop from an adenomatous hyperplasia (Sugimura and Kawachi, 1973; Wiebecke and Gokel, 1975). This proliferative lesion was observed in the present work and in an earlier experimental series (Schauer et al., 1974). It seems possible that adenomatous hyperplasia also plays a role in carcinogenesis of the human gastric mucosa (Schauer and Kunze, 1978; Schauer and Kunze, 1979). The thesis of the Japanese investigators is that adenomatous hyperplasia develops via regenerative hyperplasia from erosions, which occur within 3 weeks after application of MNNG, i.e., most likely as a toxic reaction (Sugimura et al., 1969; Saito et al., 1970; Tahara, 1976; Takahashi et al., 1976; Kogura et al., 1976; for review see Sugimura and Kawachi, 1973; Wiebecke and Gokel, 1975). No conclusive evidence could be found in the present experimental series with low doses of the carcinogen for the hypothesis that erosions played a significant role in the initiation of gastric cancer. The dysplasias which we observed developed in no case from a prior existing erosion of any kind. Furthermore, after discontinuation of MNNG their number increased progressively with time, whereas the number of erosions was by comparison very small. The superficial ulcerations which we observed particularly in the areas of the dysplasias grade III were obviously of a secondary nature. As to the cause for the secondary ulceration formation, one must consider first a disturbed regeneration following a prior cell transformation, with a decreased cell replacement, as well as a destruction of the protective mucus coat normally overlying the gastric mucosa. Our investigations have clearly shown that the induced gastric carcinomas predominately developed from focal atypias or dysplasias of an otherwise undisturbed mucosa. This process took place independently from erosions, ulcers, intestinal metaplasia, and a chronic atrophic gastritis without passing through a benign-appearing proliferative or neoplastic epithelial lesion, such as adenomatous hyperplasia, hyperplasiogenous polyps, or adenomas. Our findings, which were obtained from a low-dose application of MNNG over a limited period of time, have furthermore indicated that gastric cancer develops stepwise via several successive stages of transformation with varying biological potential. An important sign for the multifocal development of gastric cancer is the appearance of preneoplastic and neoplastic changes simultaneously in various areas of the same stomach. This finding points further to various inceptions of carcinogenesis. The earliest light microscopically detectable lesions due to carcinogen application consisted of focal dysplasias, which were identical with changes described by
260
E. Kunze et al.
other investigators variously as "glandular atypisms" (Bralow et al.; 1970), "atypical gland" (Mori et al., 1969; Shirai et al., 1974), "glandular dysplasias or glandular dysplastic lesions" (Justrabo et al., 1975),."dysplasias" (Uchida et al., 1976; Borchard et al., 1976; Rumpf et al., 1978), and "immature tubules" (Nagayo et al., 1972). In spite of their great morphological similarity one cannot equate simply the experimentally induced dysplasias with those found in the human stomach (Krauspe and Gusek, 1961; Schade, 1961; Nakamura et al., 1966; Nagayo, 1971; Sugano et al., 1971 ; Cain und Kraus, 1973; Grundmann, 1975; Oehlert et al., 1975; Fevre et al., 1976; Grundmann, 1978; Oehlert et al., 1978), with respect to their natural history and biological behavior. Three fundamental differences can be seen: 1. In the vast majority of cases dysplasias in the human stomach are combined with varying degree of chronic gastritis. 2. Human dysplasias occur as a rule in the upper third of the gastric mucosa. 3. Dysplasias in man are for the most part reversible. Oehlert et al. (1975, 1978) consider an increased cell loss with a following accelerated cell regeneration and a simultaneous loss of differentiation to be the essential process leading to the development of dysplasias in man. Extensive biopsy material of such proliferative atypical lesions has indicated that most dysplasias grade I and II appear reversible. Accordingly, all the dysplasias grade I and over 50% of the dysplasias grade II were reversible in a control period of 1 month to 2 years. Of the dysplasias grade II, 30% remained unchanged. The situation was different with dysplasias grade III. Only 22% of these cases were reversible, the other 78% remaining unchanged during 4 years of subsequent observation (Oehlert et al., 1978). In four cases with dysplasias grade III, early cancers developed at a later date. These preliminary results permit no definitive conclusions as to the biological behavior or the prospective significance of dysplasias grade III with regard to development of carcinomas in the human stomach, as the time of observation is too short, and nothing being known about the further fate of the persisting dysplasias. The experimentally induced dysplasias developed in a mucosa which showed no prior inflammation or fibrosis. Furthermore, they occured in the all-l~mportant regenerative zone at the transition of the middle to the lower third of the antrum mucosa. Similar findings were obtained in animal experiments (Borchard et al., 1976) and in the stomach of man (Cain and Kraus, 1973; Nagayo, 1975; Grundmann, 1975). This phenomenon would seem to be explained by the fact that as a rule, proliferating, DNA-synthesizing cells are more easily predisposed to malignant transformation than dividing cells in the resting period (Go-phase). Autoradiographic studies have pointed out that the physiological mode of proliferation has already been disturbed within dysplasias (Kunze et al., 1976; Oehlert et al., 1978). This pertains to the balance of cells which are dividing and remain in the proliferative zone and to the cells which have lost their mitotic potential and move toward the lumen. The basic phenomenon in the early phase of carcinogenesis consisted in a widening of the proliferative zone. This occured in two different ways. The first way consisted in the expansion of the proliferative zone toward the surface, during which the superficial cells, having lost their mitotic potential, were replaced one after the other with atypical cells. The dysplasias
Experimental Gastric Cancer
261
demonstrated an increased replacement of normal surface epithelium with proliferating atypical cells the higher their grade. As such, dysplasias grade I extended only in 1/1o th, dysplasias grade II in 1/5 th, and dysplasias grade III in 1/2 of the cases to the surface of the mucosa. The second way is characterized by a shift of proliferating cells away from the lumen in a basal direction. It is not clear why the normal cell movement toward the lumen is reversed to the opposite direction. Perhaps a disturbance in the position of the mitotic spindle plays a role. The widening of the proliferative zone in either an apical or basal direction explains in many ways the development of the various growth types of early gastric cancer. The first proliferative type with an upward movement of proliferating atypical cells leads to the development of early cancer types I, IIa, and IIb. The second proliferative type with a shift of the cells in a basal direction leads to the development of early cancers types IIc and III. An important feature of the second proliferative type consisted of an initially unaltered surface mucosa, which is understandable as being due to the basal shift of the proliferation. This finding seems very important for the gastroenterologist, who would not be able to detect such early changes. The second proliferative type could probably be explained by the supposition that within the proliferative zone only the lower portion of cells were cancerized, whereas the cells located apically were not. The cellular migration toward the lumen would remain intact, guaranteed by cell replacement through unaffected proliferating cells. The dysplasias induced by M N N G were irreversible by contrast to the dysplasias usually found in the human stomach. They were observed a long time after discontinuation of the carcinogen and their number increased with increasing experimental time. These data agree with the view that the carcinogenic cell transformation has occured long before the appearance of morphological changes. Consequently, the latent period up to the manifestation of the carcinogenic effect is extremely long. The experimental data presented provide strong evidence that a considerable number of dysplasias do not remain stable in their morphological behavior, but rather progress by gradual change to a more malignant morphology. In favor of a progressive development is the gradual transition between dysplasias grade I, II, III, as well as their sequential appearance in the various experimental groups, specifically an increase of the dysplasias grade II from group 3 on and dysplasias grade III from group 5 on (Table 2). It appears that the experimentally induced dysplasias must be variously evaluated as regards their prospective significance for carcinogenesis. The large number of dysplasias grade I and II lead us to surmise that these advance into a higher grade of dysplasia only after a long induction time or retain their biological character indefinitely. As such, one must assess the malignant potential of the dysplasias grade I and II as less than that of dysplasias grade III. From group 6 on, early cancers developed rather quickly from dysplasias grade III or via changes in the character of carcinomas in situ, as described in the human stomach (Mallory, 1940; Bocian and Geschke, 1958; Kuhlencordt, 1959; Schade, 1961; Nakamura et al., 1966; Nakamura et al., 1968; Nagayo, 1971; Schade, 1974; Serck-Hanssen et al., 1974; Mason, 1974; Russo et al., 1976). The different growth types of the induced early cancers, and their histological types, as well as their localization within the antrum agree for the most part with
262
E. Kunze et al.
the changes in the human stomach (Nagayo, 1965; Nakamura et al., 1968; Cain and Kraus, 1973; Nagayo, 1975; Elster et al., 1975; Grundmann, 1975; Grundmann; 1978; Elster and Thom/tsko, 1978; Johansen, 1976). The histological spectrum of the fully developed carcinomas also corresponded to that of the human stomach (Mulligan and Rember, 1954; Nagayo and Komagoe, 1961; Lauren, 1965; Nagayo, 1975; Nakamura et al., 1968; Mulligan, 1972; Nagayo, 1975; Teglbjaerg and Vetner, 1977; Johansen, 1976). Before drawing any conclusion from the present animal experiments regarding human carcinogenesis, it must be said, that the situation in man is more complex due to the almost omnipresent chronic gastritis. This is particularly the case with dysplasias, which must be evaluated in some ways individually in human and rat stomach. As a rule, the dysplasias grade I and II in the human stomach which occur in the upper third of the gastric mucosa, are reversible and therefore cannot be considered for the most part as precancerous. As regards the biological behavior of dysplasias grade III, there is a great divergence of opinion. Oehlert et al. have shown in their investigations that this form of dysplasias frequently persists even after 4 years and not infrequently occurs in the vicinity of fully developed carcinomas. It would appear that much longer observations are necessary to evaluate the prospective significance of dysplasias grade llI for carcinogenesis in man. The experimentally induced dysplasias of all grades are clearly irreversible when they occur in the lower third of the mucosa and without simultaneous chronic gastritis. They undergo a progressive developmental process, which ultimately leads step by step to a carcinoma growth. Therefore, they must be considered potentially premalignant. The findings obtained from the animal experiments suggest that dysplasias must be considered possible precursors of carcinomas in the human stomach when they develop in the lower third of non-inflammatory mucosa within the regenerative zone. Such dysplasias deserve more concerted attention because of their potential significance for carcinogenesis in man. Acknowledgement. The authors thank Mrs. U. Geistert for her technical assistance.
References Bocian, J.J., Geschke, A. E. : Carcinoma in situ of the stomach. Arch. Pathol. 65, 6-12 (1958) Borchard, F., Rumpf, P., Schacht, U., Palomba, P.: Formale Pathogenese des chemisch induzierten Magenkarzinoms der Ratte nach Vagotomie und Gastroenterostomie. Verh. Dtsch. Ges. Pathol. 60, 452 (1976) Bralow, S. P., Gruenstein, M., Meranze, D.R., Bonak-dar-pour, A., Shimkin, M. B.: Adenocarcinomas of glandular stomach and duodenum in Wistar rats ingesting N-methyl-N'-nitro-N-nitrosoguanidine, histopathology, and associated secretory changes. Cancer Res. 30, 1215-1222 (1970) Bralow, S.P.: Experimental gastric carcinogenesis. Digestion 5, 290-310 (1972) Cain, H., Kraus, B.: Frtihkarzinom des Magens. Morphologische Beobachtungen und Probleme. Dtsch. Med. Wochenschr. 98, 1591-1596 (1973) Druckrey, H., Preussmann, R., Ivankovic, S.: Zur Erzeugung subcutaner Sarkome an Ratten, Carcinogene Wirkung von Hydrazodicarbons~iure-bis(methyl-nitrosamid), N-Nitroso-N-n-butylharnstoff, N-Methyl-N-nitroso-nitroguanidin und N-Nitroso-imidazolidon. Z. Krebsforsch, 68, 87~02 (1966) Elster, K., Kolaczek, F., Shimamoto, K., Freitag, H. : Early gastric cancer - Experience in Germany. Endoscopy 7, 5 10 (1975) Elster, K., Thomasko, A.: Klinische Wertung der histologischen Typen des Magenfriihkarzinoms Eine Analyse von 300 F~illen. Leber Magen Darm 8, 319-327 (1978)
Experimental Gastric Cancer
263
Fevre, D.I., Green, P. H. R, Barrat, P.J., Nagy, G.S.: Review of five cases of early gastric carcinoma. Gut, 17, 4147 (1976) Fujimura, S., Kogure, K., Oboshi, S., Sugimura, T.: Production of tumors in glandular stomach of hamsters by N-methyl-N'-nitro-N-nitrosoguanidine. Cancer Res. 30, 1444-1448 (1970) Grundmann, E.: Histologic types and possible initial stages in early gastric carcinoma. Beitr. Pathol. 154, 256 280 (1975) Grundmann, E.: Early gastric cancer today. Pathol. Res. Pract. 162, 347 360 (1978) Johansen, A. A. : Early gastric cancer. Curr. Top. Pathol. 63, 147 (1976) Justrabo, E., Martin, M. S., Michiels, R., Martin, F., Bastien, H., Knopf, F., Cabanne, F.: Experimental gastro-intestinal neoplasms and hepatic cysts in rats induced by N-methyl-N'-nil.ro-N-nitrosoguanidine. Pathol. Europ. 10, 61-71 (1975) Kobori, O., Gedigk, P., Totovic, V.: Early changes of glandular stomach in Wistar rats ingesting N-methyl-N'-nitro-N-nitrosoguanidine (MNNG): with special reference to light microscopic, electron microscopic, and enzyme histochemical study of the regenerating epithelium induced by MNNG. Z. Krebsforsch. 87, 127-138 (1976) Kogure, K., Sasadaira, H., Kawachi, T., Shimosato, Y., Tokunaya, A., Fujimura, S., Sugimura, T.: Further studies on induction of stomach cancer in hamsters by N-methyl-N'-nitro-N-nitrosoguanidine. Br. J. Cancer 29, 132-142 (1974) Kohli, Y., Hashimoto, J., Takeda, S., Kawai, K.: Cell kinetics of gastric carcinoma and other gastric lesions in rats by N-methyl-N'-nitro-N-nitrosoguanidine with or without TWEEN 60. Gann 66, 133-140 (1975) Krauspe, C., Gusek, W.: Ober Fri.ihstadien des Magenkarzinoms. Verh. Dtsch. Ges. Pathol. 45, 183-185 (1961) Kuhlencordt, F.: Das Carcinoma in situ des Magens und der kleine Magenkrebs. Dtsch. Med. Wochenschr. 84, 2111-2115 (1959) Kunze, E., Schauer, A., Eder, M., Seefeldt, C.: Evidence of various proliferative types during development of early gastric cancer. Third International Symposium on Detection and Prevention of Cancer, New York 1976, Abstract No. 688 Laurdn, P.: The two histological main types c~f ~astric carcinoma: diffuse and so-called intestinaltype carcinoma. Acta Pathol. Microbiol.. Scand. 64, 3149 (1965) Mallory, T.B.: Carcinoma in situ of the stomach and its bearing on the histogenesis of malignant ulcers. Arch. Pathol. 30, 348 362 (1940) Mason, M.K.: Surface carcinoma of the stomach. In: Early gastric cancer. Grundmann, E., Grunze, H., Witte, S., eds., pp 3944. Berlin, Heidelberg, New York. Springer 1974 Mori, K., Ohta, A., Murakami, T., Tamura, M., Kondo, M., Ichimura, H.: Carcinomas of the glandular stomach and other organs of rats by 4-hydroxyaminoquinoline-l-oxide hydroehloride. Gann 60, 627~530 (1969) Mulligan, R.M., Rember, R.R.: Histogenesis and biologic behavior of gastric carcinoma. Arch. Pathol. 58, 1-25 (1954) Mulligan, R.M.: Histogenesis and biologic behavior of gastric carcinoma. Ann. Pathol. 7, 340415 (1972) Nagayo, T., Komagoe, T.: Histological studies of gastric mucosal cancer with special reference to relationship of histological pictures between the mucosal cancer and the cancer-bearing gastric mucosa. Gann 51, 109-119 (1961) Nagayo, T.: Histological diagnosis of biopsied gastric mucosa with special reference to that of borderline lesions. Gann Monograph Res. 11, 245-256 (1971) Nagayo, T., Ito, M., Yamada, S.: Histogenesis and autoradiography of adenocarcinoma of the glandular stomach in rats induced by oral aministration of N,N'-2,7-Fluorenylenebisacetamide combined with irradiation to the stomach region. Gann 63, 143-151 (1972) Nagayo, T.: Microscopical cancer of the stomach - a study on histogenesis of gastric carcinoma. Int. J. Cancer 16, 52-60 (1975) Nakamura, K., Sugano, H., Takagi, K., Fuchigami, A.: HistopathologicaI study on early carcinoma of the stomach: criteria for diagnosis of atypical epithelium. Gann 57, 613~520 (1966) Nakamura, K., Sugano, H., Takagi, K. : Carcinoma of the stomach in incipient phase: its histogenesis and histological appearances. Gann 59, 251-258 (1968) Oehlert, W., Keller, P., Henke, M., Strauch, M.: Die Dysplasien der Magenschleimhaut. Dtsch. Med. Wochenschr. 100, 1950-1965 (1975) Oehlert, W., Henke, M., Strauch, M., Keller, P.: Die Dysplasien der Magenschleimhaut, ihre Pathogenese und ihre klinische Bedeutung. Therapiewoche 28, 49664975 (1978)
264
E. Kunze et al.
Rumpf, P., Schacht, M., Palomba, P., Kramer, K., Schmitz, H., Borchard, F.: Chemisch induzierte Karzinogenese des Rattenmagens nach Vagotomie und Resektion. Z. Gastroenterol. 16, 85-94 (1978) Russo, A., Grasso, G., Sanfilippo, G., Giannone, G.: Criteri applicativi della biopsia mirata ed aspetti istologici nella diagnosi del carcinoma gastrico in situ. Tumori 62, 39-36 (1976) Saito, T., Inokuchi, L., Takayama, S., Sugimura, T.: Sequential morphological changes in N-methyiN-nitro-N-nitrosoguanidine carcinogenesis in the glandular stomach of rats. J. Natl. Cancer Inst. 44, 769-783 (1970) Serck-Hanssen, A., Marcussen, J., Liavag, L.: Endoscopic brush cytology in the diagnosis of gastric disease. Comparison with other diagnostic measures. In: Early gastric cancer. Grundmann, E., Grunze, H., Witte, S., eds., pp. 139-147. Berlin, Heidelberg, New York: Springer 1974 Schade, R.O.K.: Cancerisierung und Frfihkarzinom der Magenschleimhaut. Verh. Dtsch. Ges. Pathol. 45, 179-183 (1961) Schade, R.O.K.: The borderline between benign and malignant lesions in the stomach. In: Early gastric cancer. Grundmann, E., Grunze, H., Witte, S., eds., pp 45-53. Berlin, Heidelberg, New York: Springer 1974 Schauer, A., Kunze, E., Ober, M.: Early carcinoma des Magens im Tierexperiment. Deutscher Krebskongrel3 1974, Abstract Kongrel3berichte, S. 19. Miinchen: Demeter Verlag 1974 Schauer, A., Kunze, E.: Beziehungen der adenomat6sen Hyperplasie der Magenschleimhaut zur Cancerogenese. Akt. Gastrologie 7, 4177429 (1978) Schauer, A., Kunze, E.: Relation of adenomatous hyperplasia of the gastric mucosa to carcinogenesis. Path. Res. Pract. 164, 238-248 (1979) Schoental, R.: Carcinogenic activity of N-methyl-N-nitroso-N-nitroguanidine. Nature (Lond.) 209, 726-727 (1966) Schoental, R., Bensted, J.P.M. : Gastro-intestinal turnouts in rats and mice following various routes of administration of N-methyl-N-nitroso-N'-nitroguanidine and N-ethyl-N-nitroso-N'-nitroguanidine. Br. J. Cancer 23, 757-764 (1970) Shimosato, Y., Tanaka, N., Kogure, K., Fujimura, S., Kawachi, T., Sugimura, T.: Histopathology of tumors of canine alimentary tract produced by N-methyl-N'-nitro-N-nitrosoguanidine, with particular reference to gastric carcinoms. J. Natl. Cancer Inst. 47, 1053-1070 (1971) Shirai, T.: Periodical observations of the glandular stomach in rats treated with N-methyl-N'-nitroN-nitrosoguanidine. Nagoya Med. J. 19, 155-167 (1974) Sugani, H., Nakamura, K., Takagi, K. : An atypical epithelium of the stomach. Gann Monograph on Cancer Res. 11, 257 269 (1971) Sugimura, T., Nagao, M., Okuda, Y.: Carcinogenic action of N-methyl-N-nitro-N-nitrosoguanidine. Nature (Lond.) 210, 962-963 (1966) Sugimura, T., Fujimura, S., Kogure, K., Baba, T., Saito, T., Nagao, M., Hosoi, H., Shimosata, Y., Yokoshima, T.: Production of adenocarcinomas in glandular stomach of experimental animals by N-methyl-N-nitro-N-nitrosoguanidine. Gann Monograph 8, 157-176 (1969) Sugimura, T., Tanaka, N., Kawachi, T., Kogure, K., Fujimura, S., Shirnosato, Y.: Production of stomach cancer in dogs by N-methyl-N'-nitro-N-nitrosoguanidine. Gann 62, 67 (1971) Sugimura, T., Kawachi, T.: Experimental stomach cancer. Methods Cancer Res. 7, 245-308 (1973) Tahara, E., Haisuka, S., Tokuoka, S.: An ultrastrnctural study on precancerous and cancerous lesions of the glandular stomach in rats induced by N-methyl-N'-nitro-N-nitrosoguanidine. Acta Pathol. Jap. 26, 1-15 (1976) Takahashi, M., Shirai, T., Fukushima, S., Hahanouchi, M., Hirose, M., Ito, N.: Effect of fundic ulcers induced by iodoacetamide on development of gastric tumors in rats treated with N-methyl-N'nitro-N-nitrosoguanidine. Gann 67, 47-54 (1976) Teglbjaerg, P.S., Vetner, M.: Gastric carcinoma I. Acta Path. Microbiol. Scan& Sect. A 85, 519-527 (1977) Uchida, Y., Schlake, W., Roessner, A., Riihland, D., Themann, H., Grundmann, E.: Development of tumors in the glandular stomach of rats after oral administration of carcinogens. Z. Krebsforsch. 87, 199-212 (1976) Wiebecke, B., Gokel, J.M.: Experimentelle Cancerogenese des Magen-Darm-Kanals. Handbuch der Allgemeinen Pathologie, 6. Band. Geschwiilste. 7. Tell, S. 693-730. Berlin, Heidelberg, New York: Springer 1975 Received June 5, 1979/Accepted October 3, 1979
~,CancerResearch Clinical Oncology
J. Cancer Res. Clin. Oncoi. 95, 247 264 (1979)
9 Springer-Verlag1979
Early Sequential Lesions During Development of Experimental Gastric Cancer with Special Reference to Dysplasias* E. Kunze 1, A, Schauer, M. Eder, and C. Seefeldt 1 Institute of Pathology, University of G/Sttingen, Robert-Koch-Str. 40, D-M00 G6ttingen and Institute of Pathology, University of Mufiich, Thalkirchnerstr. 36, D-8000 Munich, Federal Republic of Germany
Friihver~inderungen der Carcinogenese an der Magenschleimhaut im Tierexperiment unter besonderer Beriicksichtigung der Dysplasien Zusammenfassung. Es wurde tierexperimentell die Entwicklung des Magencarcinoms in seinen Frtihphasen untersucht und gepriift, welche Rfickschlfisse auf den Cancerisierungsablauf am menschlichen Magen m6glich sind. Nach limitierter oraler Zufuhr von N-Methyl-N'-Nitro-N-Nitrosoguanidin an 174 Ratten entwickelten sich Carcinome direkt aus der sonst unver~inderten Schleimhaut fiber mehrere aufeinanderfolgende Transformationsstadien, ohne gutartig erscheinende proliferative oder neoplastische Epithell~isionen zu durchlaufen. Als erste lichtoptisch erkennbare Ver~inderung ergaben sich fokal Dysplasien G r a d I, die mit zunehmender Versuchszeit in Dysplasien G r a d II und III iibergingen. Die experimentell induzierten Dysplasien sind kausalgenetisch und in ihrem biologischen Verhalten nicht mit der Mehrzahl der am menschlichen Magen auftretenden Dysplasien gleichzusetzen. Die in der Regel im oberen Schleimhautdrittel lokalisierten Dysplasien des Menschen insbesondere des Grades I und II sind gr613tenteils entztindungsbedingt und reversibel. Die experimentellen Dysplasien im Bereich der Proliferationszone sind dagegen als potentiell pdimaligne anzusehen, da sie irreversibel waren und sich progredient weiterentwickelten. Nach den tierexperimentellen Befunden kommen beim Menschen insbesondere die Dysplasien des Grades III als m6gliche Vorl~iufer yon Carcinomen in Betracht, die sich in nicht entziindlicher Schleimhaut im Bereich der Proliferationszone entwickelt haben. Sehliisselwiirter: N-Methyl-N'-Nitro-N-Nitrosoguanidin-Carcinogenese der Magenschleimhaut-Dysplasien - Friihcarcinom * This work was supported in part by a research grant from the Buettner Foundation (G6ttingen) This work was reported in part at the Third International Symposium on Detection and Prevention of Cancer, New York, USA, 1976, and at the 62nd meeting of the Deutsche Gesellschaftfiir Pathologie, Vienna, Austria, 1978 Offprint requests to: Prof. Dr. E. Kunze (address see above) 0171- 5216/79/0095/0247/$ 3.60
248
E. Kunzeet al. Summary. The early sequential development of gastric cancer was studied with experimental animals and examined with respect to what conclusions can be drawn for understanding carcinogenesis in man. After limited oral administration of N-methyl-N'nitro-N-nitrosoguanidine to 174 rats carcinomas developed in most cases directly from the otherwise unchanged mucosa through various successive stages of transformation, without passing through a benignappearing proliferative or neoplastic epithelial lesion. Focal dysplasia grade I was the first recognizable change observed by light microscopy, followed by dysplasia grade II, and subsequently dysplasia grade III. In spite of very similar morphological characteristics, the experimentally induced dysplasias cannot be simply equated in their etiology and biological behavior with the dysplasias of the human stomach. Dysplasias of grade I and II commonly found in man are usually associated with a chronic gastritis; they are located in the upper third of the mucosa and are for the most part reversible. The experimental dysplasias occuring in the proliferative zone of an otherwise undisturbed mucosa must be considered potentially premalignant, as they are irreversible and develop progressively. This finding points out that in man dysplasias grade III within the regenerative zone of non-inflammatory mucosa should be considered particularly as possible precursors of gastric carcinomas. Key words: N-methyl-N;-nitro-N-nitrosoguanidine-Carcinogenesis of gastric mucosa-Dysplasias - Early cancer
The discovery of the carcinogenic effect of N-methyl-N'nitro-N-nitrosoguanidine (MNNG) in 1966 by Druckrey et al., Schoental et al., and Sugimura et al., made it possible to induce carcinomas selectively in the glandular part of the stomach of various laboratory animals, such as mice (Schoental and Bensted, 1970), rats (Sugimura et al., 1969; Saito et al., 1970; Schoental and Bensted, 1970; Shirai, 1974; Tahara et al., 1976), hamsters (Fujimura et al., 1970; Sugimura et al., 1969; Kogure et al., 1974), and dogs (Shimasato et al., 1971; Sugimura et al., 1971; for review of literature see Bralow, 1972; Sugimura and Kawachi, 1973; Wiebecke and Gokel, 1975). Japanese investigators predominately have shown that the experimentally induced gastric carcinomas can develop histogenetically via an adenomatous hyperplasia (Sugimura et al., 1969; Saito et al., 1970; Kohli et al., 1975; Tahara et al., 1976; Uchida et al., 1976; Takahashi et al., 1976; Kobori et al., 1976; for review of literature see Sugimura and Kawachi, 1973; Wiebecke and Gokel, 1975). In the present series of experiments an attempt has been made to clarify whether gastric cancer can also develop directly from a focal atypia of the mucosa without passing through a benign-appearing epithelial proliferation. Of particular interest was whether dysplasias, which are frequently encountered in the human stomach, are also inducible in the rat stomach. AdditionalIy, we sought to learn more about the biological behavior of dysplasias and their prospective significance in the development of gastric carcinoma, and to draw from such animal experiments conclusions regarding carcinogenesis in man.
Experimental Gastric Cancer
249
Material and Methods Two hundred and fifty adult female Wistar rats with an initial weight between 180 and 200 g received MNNG continuously for 250 days (time of exposure) in their drinking water at a dosage of 83 mg/1000 ml and were then discontinued. The aqueous carcinogenic solution was prepared fresh every 2 days. The drinking vessels were covered with black paper to protect the MNNG from degradation by light. To assure continuous observation of the carcinogenic process and to detect earliest morphological changes, the surviving 174 experimental animals were divided into 9 groups and killed after different durations of induction from experimental days 290 to 630 (Table 2). There were 20 animals in each of the first 8 groups, and 14 in the 9th. The subsequent carcinogen-free interval comprised 40 to 380 days. The rats were kept in Macrolon cages and had free access to their drinking water. They received a standardized compressed feed (Atromin) throughout the study. Imrnediately after killing, the stomachs were incised along the lesser curvature and pinned out onto a cork board. After fixation in a 10% formalin solution, a 10 mm long strip of tissue was excised for histological examination from each of six predetermined areas of the stomach: on both sides of the lesser curvature at the fundic region, at the transition of the fundic region to the antrum (both sides), at the antrum, and at the duodenum. After inbedding in paraffin at least 25 5-gm thick serial sections were cut from each specimen. A minimum of 150 sections per animal were examined. AdditionalIy, sections were cut from all tumors. The histological sections were stained routinely with hematoxylin-eosin (H. E.). Other stainings were carried out as needed (PAS, Alcian blue at pH 2.5, van Gieson).
Results
1. Dysplasias Dysplasias were the earliest carcinogenic lesions in the gastric m u c o s a recognizable by light microscopy. F o c a l changes of the m u c o s a that exhibited structural m o d i fications of the glands a n d cellular atypia were designated dysplasias. They were f o u n d exclusively in the a n t r u m a n d frequently occurred in several areas of the same s t o m a c h after longer experimental times. T h e frequency of all dysplasias within the various e x p e r i m e n t a l groups have been included in T a b l e 2. The dysplasias ocurred in a n otherwise u n d i s t u r b e d , i.e. n o n - i n f l a m m a t o r y a n d n o n fibrotic, mucosa. As a rule, dysplasias observed in h u m a n stomachs are localized in the upper third of the mucosa. However, u n d e r experimental c o n d i t i o n s in rats, the i n d u c e d dysplasias were f o u n d to develop at the t r a n s i t i o n between the m i d d l e a n d lower third of the m u c o s a w i t h i n the regenerative zone, from which they could spread in a n apical as well as in a basal direction. I n the present experimental series a total of 112 dysplasias grade I - I I I were registered in 92 affected rats (Table 2). Three grades of dysplasias were distinguished according to their histological a n d cytological features (Table 1). Table 1. Histological and cytological features of dysplasias Dysplasia I
Dysplasia II
Dysplasia III
Nuclei
Slightly irregular arrangement Slight enlargement
Pronouncedly irregular arrangement Marked hyperchromasia
Cytoplasm
Normal appearance
Mitotic activity
Low
Moderately irregular arrangement Moderate enlargement and hyperchromasia Slight to moderate basophilia Moderate
Glands
Pronouced basophilia High
250
E. Kunze et al.
Table 2. Frequency of carcinogen-induced lesions of the gastric mucosa in the different experimental groups (groups 1-8 each consisting of 20 rats, group 9 comprising 14 animals) after limited oral application of MNNG. Stepwise increase of lesions with a higher degree of atypia and malignancy, respectively, with increasing experimental time are indicated Group: Time of induction (days): Time after discontinuation (days): Type of lesion
1
2
3
4
5
6
290 340 390 440 490 40 90 140 190 240
7
8
510 530 600 260 280 350
9
630 380
Number of observed lesions/ Number of affected animals
Totals
Lesions/rats: Dysplasia grade I Dysplasia grade II Dysplasia grade III Early cancer (Mucosa-type) Early cancer (Submucosa-type) Fully developed carcinoma:
1/1 1/11 0 1/1 0 1/1 1/1 0 1/1 0
5/4 5/4 3/3 11/6 0 1/1 I 6/6 8/6 2/2 1/1 2/2 2/2 2/2 0 2/2 1/1
8/7 1/1 8/5 9/7 9/6 5/5 9/6 8/8 7/7 6/6 3/3 5/5
8/6 6/6 4/3 9/7 13/12 4/4
3/3 3/3 9/8 8/5 7/7 4/4
Total lesions :
3
18
44
44
34
0 I 2/2 3/3 0
5
10
26
34
30/26 47/36 35/30 46/36 40/39 20/20
Fig. 1. Dysplasia grade I in the lower third of the antrum mucosa, exhibiting slight cellular atypia. Magnification: ca. 600 x
Experimental Gastric Cancer
251
Fig. 2. Dysplasia grade II in the lower third of the antrum mucosa showing increased cellular atypia with enlargement and hyperchromasia of the nuclei, moderate basophilia of the cytoplasm, and increased mitotic activity. Magnification: ca, 600 x
a) Dysplasias grade I Dysplasias grade I demonstrated minimally to slightly atypically arranged glands with slight dyscaryosis and slightly increased mitotic activity (Fig. 1). The number of animals with dysplasias grade I increased with increased experimental time (Table 2). Only five animals in groups 1-4 had dysplasias grade I, 21 animals in groups 5-9. Twenty-seven of 30 dysplasias grade I did not extend to the surface of the mucosa. Atypical cells were found on the surface in only 3 dysplasias (approximately 1/1o th of all cases). A hyperplasia of the mucosa was not developed in the area of any dysplasia grade I. b) Dysptasias grade II Dysplasias grade II were distinguished from grade I on the basis of an increased glandular distortion and cellular atypia (Fig. 2). Among all exposed animals, a total of 47 dysplasias grade II in 36 affected animals were found (Table 2). They occurred most frequently in groups 5-7. Notably in group 5, several dysplasias grade II were encountered in the same animal. In 38 of the dysplasias the surface epithelium was without any cytological irregularity. In five cases a loss of cellular
252
E. Kunze et al.
Fig. 3. Early cancer of the mucosa type (antrum) with superficial ulceration (right upper corner) Magnification: ca. 375 x
differentiation was noted with a corresponding loss of mucus production. Nine dysplasias (approx. a/s th of the cases) extended to the surface of the mucosa; two of them showed superficial erosions. A slight to moderate hyperplasia of the mucosa occured concomitantly with four of the dysplasias.
c) Dysplasias grade III Dysplasias grade III were characterized by highly distorted, atypically branched glands or tubules in back-to-back arrangement. For the most part they were lined by several layers of markedly atypical cells. The nuclei of these cells were partly hyperchromatic and partly vesicular, varied in size and shape, and frequently exhibited a single prominent nucleolus. The nucleo-cytoplasmic ratio was shifted in favor of the nucleus. The basement membrane was always intact. A total of 35 dysplasias grade III were found in 30 affected animals (Table 2). Their frequency increased with increased total experimental time. Whereas only two dysplasias grade III were found in the first four experimental groups comprising a total of 80 rats, 33 dysplasias occurred in 28 affected among 94 total animals in groups 5 9. In half of the cases, these extended up to the surface of the mucosa. Approximately one-fourth of the dysplasias showed superficial ulcerations. Six dysplasias exhibited concomitantly a moderate hyperplasia of the mucosa. Nine dysplasias which did not extend to the surface showed a loss of mucus production by the surface epithelia.
Experimental Gastric Cancer
253
Fig. 4. Highly differentiated early cancer of the incipient submucosa type, with penetration of the M. mucosae by small groups of atypical glands (antrum). Magnification: ca. 600 x
2. Early Cancer The early cancer in its mucosal and submucosal type represented an important further stage in carcinogenesis of the gastric mucosa (Fig. 3-5). Of 86 observed early cancers 81 were localized in the antrum-pyloric region and only five in the fundic region.
a) Mucosa-Type Forty-six early cancers of the mucosa-type developed in 36 affected animals (Table 2). They were predominately found in groups 6-9. Forty-one carcinomas occurred in the antrum, five in the fundic region. More than half of the carcinomas (27 cases) extended to the surface of the mucosa. Fourteen were ulcerated and corresponded to the growth pattern of type IIc. Histologically, they were welldifferentiated, adeno-tubular carcinomas (intestinal type) with only minimal to slight cell atypia.
254
E. Kunze et al.
Fig. 5. Less differentiated early cancer of the incipient submucosa type with penetration of the M. mucosae (antrum). Magnification: ca. 600 x
b) Submucosa-Type A total of 40 submucosal early carcinomas were found in 39 affected animals. They occurred most frequently in groups 6-9 (Table 2). Two-thirds of the carcinomas extended to the surface of the mucosa. Half of the cases (21) were ulcerated and belonged to growth type IIc. A few could be classified as type III. Thirty-eight were well-differentiated adenocarcinomas (intestinal type), one an undifferentiated carcinoma (anaplastic type), and one a signet-ring cell carcinoma (muco-cellular type).
3. Fully Developed Carcinomas Fully developed carcinomas infiltrating the M. propria were found in 20 cases (Table 2). Nineteen were localized in the antrum, and one in the fundic region. The majority of carcinomas developed in groups 6-9. The carcinoma which occurred in the fundic region could already be observed in the first experimental group. Histologically, there were 12 well-differentiated adenocarcinomas (pylorocardiac and intestinal cell type as well as transitional forms), three poorly
Experimental Gastric Cancer
255
Fig. 6. Circumscribedglandular proliferation (adenomatous hyperplasia) of the antrum extending into the submucosa. Proliferatedglands lined by a single layer of highly differentiated mucus-producing cylindricalcells. Magnification: ca. 240 x
differentiated adenocarcinomas (mainly solid-anaplastic growth pattern) and 1 signet-ring celt carcinoma. Four carcinomas showed a mixed pattern with well and poorly differentiated glandular structures, anaplastic components and a focal monocellular transformation. In a few carcinomas cartilage, bone and/or osteoid were found in the stroma.
4. Submucosal Glandular Proliferations (Adenomatous Hyperplasia) To be distinguished from submucosal early cancer are the sharply demarcated tumor-like foci of glandular proliferations, extending into the submucosa by penetration of the M. mucosae without any signs of an invasive growth (Fig. 6). These foci originated apparently from glands at the base of the mucosa (Figs. 7 and 8). The submucosal glandular proliferations were localized exclusively in the antrum. The glands were lined by a single layer of either columnar mucus producing or cuboidal benign-appearing epithelial cells. Occasionally, focal atypia could be established. The changes we have distinguished as submucosal glandular proliferations are identical to the "adenomatous hyperplasia" of Japanese authors.
256
E. Kunze et al.
Fig. 7. Focal glandular proliferation in the lower third of the antrum mucosa. Magnification: ca. 240 •
The number of the observed submucosal glandular proliferations was not registered in this experimental series. A special variant of the submucosal glandular proliferations were the so-called adenomatous diverticula, found in two animals. They were characterized by an extremly expansive growth in the submucosa, exhibiting centrally a deep diverticular indentation (Fig. 9). 5. Tubulo-Villous Adenomas
In the present experimental series only one tubulo-villous adenoma was found in the antrum of an animal in group 5 (Fig. 10). This tumor consisted of glands and villi with highly differentiated cylindrical cells, which were positive in the PAS and alcian blue reaction, therein being similar to foveolar and pyloric cells. The adenoma showed several loci with marked cellular atypia and increased mitotic activity and therefore must be considered to be undergoing malignant transformation.
Experimental Gastric Cancer
257
Fig. 8. Higher magnificationof the lesion shown in Fig. 7. Magnification: ca. 600 x
6. Different Lesions Pyloric gland metaplasia: Very frequently so-called pyloric gland metaplasias were found in the fundic region. These wedge-shaped mucosal lesions extended occasionally up to the M. mucosae. The specific cells of the gastric mucosa had atrophied in these areas and were replaced by undifferentiated cylindrical or cuboidal cells, some producing mucus. This lesion was designated pyloric gland metaplasia, because of the similarity with pyloric glands (Shirai et al., 1974). They occured very frequently in all groups, and were found more frequently with increased experimental time. We observed a total of 280 pyloric gland metaplasias among 122 animals. In only five cases were we able to demonstrate cellular atypia. In contrast, we encountered intestinal metaplasia rarely. It was found somewhat more frequently in the neighborhood of fully developed carcinomas. Erosions: Of a total of 174 animals 30 had developed gastric mucosal erosions. We found 15 fresh erosions in nine rats, occuring in every test group. Eleven were located in the fundic region, and five in the antrum. Healed erosions were observed in 25 cases in the antrum, again with all the test groups represented. A number of rats demonstrated simultaneously fresh and healed erosions.
258
E. Kunze et al.
Fig. 9. Adenomatous diverticulum of the antrum characterized by extensive, sharply demarcated submucosal glandular proliferations. Magnification: ca. 90 x
Fig. 10. Tubulo-villous adenoma of the antrum consisting of villi and glands lined by highly differentiated, mucus-producing cells. Magnification: ca. 90 x
Experimental GastricCancer
259
Intestinal tumors: A total of 12 tumors were found in the small intestine among
10 rats. Ten tumors were located in the duodenum, and two in the jejunum. Of the duodenal tumors, seven were well-differentiated adenocarcinomas, one an undifferentiated carcinoma and two were fibrosarcomas. The two tumors in the jejunum were an adenocarcinoma and a fibrosarcoma.
Discussion
Gastric carcinomas induced by MNNG can develop histogenetically via an adenomatous hyperplasia as well as directly from a focal atypia of the mucosa. Notably Japanese investigators were able to demonstrate that carcinomas develop from an adenomatous hyperplasia (Sugimura and Kawachi, 1973; Wiebecke and Gokel, 1975). This proliferative lesion was observed in the present work and in an earlier experimental series (Schauer et al., 1974). It seems possible that adenomatous hyperplasia also plays a role in carcinogenesis of the human gastric mucosa (Schauer and Kunze, 1978; Schauer and Kunze, 1979). The thesis of the Japanese investigators is that adenomatous hyperplasia develops via regenerative hyperplasia from erosions, which occur within 3 weeks after application of MNNG, i.e., most likely as a toxic reaction (Sugimura et al., 1969; Saito et al., 1970; Tahara, 1976; Takahashi et al., 1976; Kogura et al., 1976; for review see Sugimura and Kawachi, 1973; Wiebecke and Gokel, 1975). No conclusive evidence could be found in the present experimental series with low doses of the carcinogen for the hypothesis that erosions played a significant role in the initiation of gastric cancer. The dysplasias which we observed developed in no case from a prior existing erosion of any kind. Furthermore, after discontinuation of MNNG their number increased progressively with time, whereas the number of erosions was by comparison very small. The superficial ulcerations which we observed particularly in the areas of the dysplasias grade III were obviously of a secondary nature. As to the cause for the secondary ulceration formation, one must consider first a disturbed regeneration following a prior cell transformation, with a decreased cell replacement, as well as a destruction of the protective mucus coat normally overlying the gastric mucosa. Our investigations have clearly shown that the induced gastric carcinomas predominately developed from focal atypias or dysplasias of an otherwise undisturbed mucosa. This process took place independently from erosions, ulcers, intestinal metaplasia, and a chronic atrophic gastritis without passing through a benign-appearing proliferative or neoplastic epithelial lesion, such as adenomatous hyperplasia, hyperplasiogenous polyps, or adenomas. Our findings, which were obtained from a low-dose application of MNNG over a limited period of time, have furthermore indicated that gastric cancer develops stepwise via several successive stages of transformation with varying biological potential. An important sign for the multifocal development of gastric cancer is the appearance of preneoplastic and neoplastic changes simultaneously in various areas of the same stomach. This finding points further to various inceptions of carcinogenesis. The earliest light microscopically detectable lesions due to carcinogen application consisted of focal dysplasias, which were identical with changes described by
260
E. Kunze et al.
other investigators variously as "glandular atypisms" (Bralow et al.; 1970), "atypical gland" (Mori et al., 1969; Shirai et al., 1974), "glandular dysplasias or glandular dysplastic lesions" (Justrabo et al., 1975),."dysplasias" (Uchida et al., 1976; Borchard et al., 1976; Rumpf et al., 1978), and "immature tubules" (Nagayo et al., 1972). In spite of their great morphological similarity one cannot equate simply the experimentally induced dysplasias with those found in the human stomach (Krauspe and Gusek, 1961; Schade, 1961; Nakamura et al., 1966; Nagayo, 1971; Sugano et al., 1971 ; Cain und Kraus, 1973; Grundmann, 1975; Oehlert et al., 1975; Fevre et al., 1976; Grundmann, 1978; Oehlert et al., 1978), with respect to their natural history and biological behavior. Three fundamental differences can be seen: 1. In the vast majority of cases dysplasias in the human stomach are combined with varying degree of chronic gastritis. 2. Human dysplasias occur as a rule in the upper third of the gastric mucosa. 3. Dysplasias in man are for the most part reversible. Oehlert et al. (1975, 1978) consider an increased cell loss with a following accelerated cell regeneration and a simultaneous loss of differentiation to be the essential process leading to the development of dysplasias in man. Extensive biopsy material of such proliferative atypical lesions has indicated that most dysplasias grade I and II appear reversible. Accordingly, all the dysplasias grade I and over 50% of the dysplasias grade II were reversible in a control period of 1 month to 2 years. Of the dysplasias grade II, 30% remained unchanged. The situation was different with dysplasias grade III. Only 22% of these cases were reversible, the other 78% remaining unchanged during 4 years of subsequent observation (Oehlert et al., 1978). In four cases with dysplasias grade III, early cancers developed at a later date. These preliminary results permit no definitive conclusions as to the biological behavior or the prospective significance of dysplasias grade III with regard to development of carcinomas in the human stomach, as the time of observation is too short, and nothing being known about the further fate of the persisting dysplasias. The experimentally induced dysplasias developed in a mucosa which showed no prior inflammation or fibrosis. Furthermore, they occured in the all-l~mportant regenerative zone at the transition of the middle to the lower third of the antrum mucosa. Similar findings were obtained in animal experiments (Borchard et al., 1976) and in the stomach of man (Cain and Kraus, 1973; Nagayo, 1975; Grundmann, 1975). This phenomenon would seem to be explained by the fact that as a rule, proliferating, DNA-synthesizing cells are more easily predisposed to malignant transformation than dividing cells in the resting period (Go-phase). Autoradiographic studies have pointed out that the physiological mode of proliferation has already been disturbed within dysplasias (Kunze et al., 1976; Oehlert et al., 1978). This pertains to the balance of cells which are dividing and remain in the proliferative zone and to the cells which have lost their mitotic potential and move toward the lumen. The basic phenomenon in the early phase of carcinogenesis consisted in a widening of the proliferative zone. This occured in two different ways. The first way consisted in the expansion of the proliferative zone toward the surface, during which the superficial cells, having lost their mitotic potential, were replaced one after the other with atypical cells. The dysplasias
Experimental Gastric Cancer
261
demonstrated an increased replacement of normal surface epithelium with proliferating atypical cells the higher their grade. As such, dysplasias grade I extended only in 1/1o th, dysplasias grade II in 1/5 th, and dysplasias grade III in 1/2 of the cases to the surface of the mucosa. The second way is characterized by a shift of proliferating cells away from the lumen in a basal direction. It is not clear why the normal cell movement toward the lumen is reversed to the opposite direction. Perhaps a disturbance in the position of the mitotic spindle plays a role. The widening of the proliferative zone in either an apical or basal direction explains in many ways the development of the various growth types of early gastric cancer. The first proliferative type with an upward movement of proliferating atypical cells leads to the development of early cancer types I, IIa, and IIb. The second proliferative type with a shift of the cells in a basal direction leads to the development of early cancers types IIc and III. An important feature of the second proliferative type consisted of an initially unaltered surface mucosa, which is understandable as being due to the basal shift of the proliferation. This finding seems very important for the gastroenterologist, who would not be able to detect such early changes. The second proliferative type could probably be explained by the supposition that within the proliferative zone only the lower portion of cells were cancerized, whereas the cells located apically were not. The cellular migration toward the lumen would remain intact, guaranteed by cell replacement through unaffected proliferating cells. The dysplasias induced by M N N G were irreversible by contrast to the dysplasias usually found in the human stomach. They were observed a long time after discontinuation of the carcinogen and their number increased with increasing experimental time. These data agree with the view that the carcinogenic cell transformation has occured long before the appearance of morphological changes. Consequently, the latent period up to the manifestation of the carcinogenic effect is extremely long. The experimental data presented provide strong evidence that a considerable number of dysplasias do not remain stable in their morphological behavior, but rather progress by gradual change to a more malignant morphology. In favor of a progressive development is the gradual transition between dysplasias grade I, II, III, as well as their sequential appearance in the various experimental groups, specifically an increase of the dysplasias grade II from group 3 on and dysplasias grade III from group 5 on (Table 2). It appears that the experimentally induced dysplasias must be variously evaluated as regards their prospective significance for carcinogenesis. The large number of dysplasias grade I and II lead us to surmise that these advance into a higher grade of dysplasia only after a long induction time or retain their biological character indefinitely. As such, one must assess the malignant potential of the dysplasias grade I and II as less than that of dysplasias grade III. From group 6 on, early cancers developed rather quickly from dysplasias grade III or via changes in the character of carcinomas in situ, as described in the human stomach (Mallory, 1940; Bocian and Geschke, 1958; Kuhlencordt, 1959; Schade, 1961; Nakamura et al., 1966; Nakamura et al., 1968; Nagayo, 1971; Schade, 1974; Serck-Hanssen et al., 1974; Mason, 1974; Russo et al., 1976). The different growth types of the induced early cancers, and their histological types, as well as their localization within the antrum agree for the most part with
262
E. Kunze et al.
the changes in the human stomach (Nagayo, 1965; Nakamura et al., 1968; Cain and Kraus, 1973; Nagayo, 1975; Elster et al., 1975; Grundmann, 1975; Grundmann; 1978; Elster and Thom/tsko, 1978; Johansen, 1976). The histological spectrum of the fully developed carcinomas also corresponded to that of the human stomach (Mulligan and Rember, 1954; Nagayo and Komagoe, 1961; Lauren, 1965; Nagayo, 1975; Nakamura et al., 1968; Mulligan, 1972; Nagayo, 1975; Teglbjaerg and Vetner, 1977; Johansen, 1976). Before drawing any conclusion from the present animal experiments regarding human carcinogenesis, it must be said, that the situation in man is more complex due to the almost omnipresent chronic gastritis. This is particularly the case with dysplasias, which must be evaluated in some ways individually in human and rat stomach. As a rule, the dysplasias grade I and II in the human stomach which occur in the upper third of the gastric mucosa, are reversible and therefore cannot be considered for the most part as precancerous. As regards the biological behavior of dysplasias grade III, there is a great divergence of opinion. Oehlert et al. have shown in their investigations that this form of dysplasias frequently persists even after 4 years and not infrequently occurs in the vicinity of fully developed carcinomas. It would appear that much longer observations are necessary to evaluate the prospective significance of dysplasias grade llI for carcinogenesis in man. The experimentally induced dysplasias of all grades are clearly irreversible when they occur in the lower third of the mucosa and without simultaneous chronic gastritis. They undergo a progressive developmental process, which ultimately leads step by step to a carcinoma growth. Therefore, they must be considered potentially premalignant. The findings obtained from the animal experiments suggest that dysplasias must be considered possible precursors of carcinomas in the human stomach when they develop in the lower third of non-inflammatory mucosa within the regenerative zone. Such dysplasias deserve more concerted attention because of their potential significance for carcinogenesis in man. Acknowledgement. The authors thank Mrs. U. Geistert for her technical assistance.
References Bocian, J.J., Geschke, A. E. : Carcinoma in situ of the stomach. Arch. Pathol. 65, 6-12 (1958) Borchard, F., Rumpf, P., Schacht, U., Palomba, P.: Formale Pathogenese des chemisch induzierten Magenkarzinoms der Ratte nach Vagotomie und Gastroenterostomie. Verh. Dtsch. Ges. Pathol. 60, 452 (1976) Bralow, S. P., Gruenstein, M., Meranze, D.R., Bonak-dar-pour, A., Shimkin, M. B.: Adenocarcinomas of glandular stomach and duodenum in Wistar rats ingesting N-methyl-N'-nitro-N-nitrosoguanidine, histopathology, and associated secretory changes. Cancer Res. 30, 1215-1222 (1970) Bralow, S.P.: Experimental gastric carcinogenesis. Digestion 5, 290-310 (1972) Cain, H., Kraus, B.: Frtihkarzinom des Magens. Morphologische Beobachtungen und Probleme. Dtsch. Med. Wochenschr. 98, 1591-1596 (1973) Druckrey, H., Preussmann, R., Ivankovic, S.: Zur Erzeugung subcutaner Sarkome an Ratten, Carcinogene Wirkung von Hydrazodicarbons~iure-bis(methyl-nitrosamid), N-Nitroso-N-n-butylharnstoff, N-Methyl-N-nitroso-nitroguanidin und N-Nitroso-imidazolidon. Z. Krebsforsch, 68, 87~02 (1966) Elster, K., Kolaczek, F., Shimamoto, K., Freitag, H. : Early gastric cancer - Experience in Germany. Endoscopy 7, 5 10 (1975) Elster, K., Thomasko, A.: Klinische Wertung der histologischen Typen des Magenfriihkarzinoms Eine Analyse von 300 F~illen. Leber Magen Darm 8, 319-327 (1978)
Experimental Gastric Cancer
263
Fevre, D.I., Green, P. H. R, Barrat, P.J., Nagy, G.S.: Review of five cases of early gastric carcinoma. Gut, 17, 4147 (1976) Fujimura, S., Kogure, K., Oboshi, S., Sugimura, T.: Production of tumors in glandular stomach of hamsters by N-methyl-N'-nitro-N-nitrosoguanidine. Cancer Res. 30, 1444-1448 (1970) Grundmann, E.: Histologic types and possible initial stages in early gastric carcinoma. Beitr. Pathol. 154, 256 280 (1975) Grundmann, E.: Early gastric cancer today. Pathol. Res. Pract. 162, 347 360 (1978) Johansen, A. A. : Early gastric cancer. Curr. Top. Pathol. 63, 147 (1976) Justrabo, E., Martin, M. S., Michiels, R., Martin, F., Bastien, H., Knopf, F., Cabanne, F.: Experimental gastro-intestinal neoplasms and hepatic cysts in rats induced by N-methyl-N'-nil.ro-N-nitrosoguanidine. Pathol. Europ. 10, 61-71 (1975) Kobori, O., Gedigk, P., Totovic, V.: Early changes of glandular stomach in Wistar rats ingesting N-methyl-N'-nitro-N-nitrosoguanidine (MNNG): with special reference to light microscopic, electron microscopic, and enzyme histochemical study of the regenerating epithelium induced by MNNG. Z. Krebsforsch. 87, 127-138 (1976) Kogure, K., Sasadaira, H., Kawachi, T., Shimosato, Y., Tokunaya, A., Fujimura, S., Sugimura, T.: Further studies on induction of stomach cancer in hamsters by N-methyl-N'-nitro-N-nitrosoguanidine. Br. J. Cancer 29, 132-142 (1974) Kohli, Y., Hashimoto, J., Takeda, S., Kawai, K.: Cell kinetics of gastric carcinoma and other gastric lesions in rats by N-methyl-N'-nitro-N-nitrosoguanidine with or without TWEEN 60. Gann 66, 133-140 (1975) Krauspe, C., Gusek, W.: Ober Fri.ihstadien des Magenkarzinoms. Verh. Dtsch. Ges. Pathol. 45, 183-185 (1961) Kuhlencordt, F.: Das Carcinoma in situ des Magens und der kleine Magenkrebs. Dtsch. Med. Wochenschr. 84, 2111-2115 (1959) Kunze, E., Schauer, A., Eder, M., Seefeldt, C.: Evidence of various proliferative types during development of early gastric cancer. Third International Symposium on Detection and Prevention of Cancer, New York 1976, Abstract No. 688 Laurdn, P.: The two histological main types c~f ~astric carcinoma: diffuse and so-called intestinaltype carcinoma. Acta Pathol. Microbiol.. Scand. 64, 3149 (1965) Mallory, T.B.: Carcinoma in situ of the stomach and its bearing on the histogenesis of malignant ulcers. Arch. Pathol. 30, 348 362 (1940) Mason, M.K.: Surface carcinoma of the stomach. In: Early gastric cancer. Grundmann, E., Grunze, H., Witte, S., eds., pp 3944. Berlin, Heidelberg, New York. Springer 1974 Mori, K., Ohta, A., Murakami, T., Tamura, M., Kondo, M., Ichimura, H.: Carcinomas of the glandular stomach and other organs of rats by 4-hydroxyaminoquinoline-l-oxide hydroehloride. Gann 60, 627~530 (1969) Mulligan, R.M., Rember, R.R.: Histogenesis and biologic behavior of gastric carcinoma. Arch. Pathol. 58, 1-25 (1954) Mulligan, R.M.: Histogenesis and biologic behavior of gastric carcinoma. Ann. Pathol. 7, 340415 (1972) Nagayo, T., Komagoe, T.: Histological studies of gastric mucosal cancer with special reference to relationship of histological pictures between the mucosal cancer and the cancer-bearing gastric mucosa. Gann 51, 109-119 (1961) Nagayo, T.: Histological diagnosis of biopsied gastric mucosa with special reference to that of borderline lesions. Gann Monograph Res. 11, 245-256 (1971) Nagayo, T., Ito, M., Yamada, S.: Histogenesis and autoradiography of adenocarcinoma of the glandular stomach in rats induced by oral aministration of N,N'-2,7-Fluorenylenebisacetamide combined with irradiation to the stomach region. Gann 63, 143-151 (1972) Nagayo, T.: Microscopical cancer of the stomach - a study on histogenesis of gastric carcinoma. Int. J. Cancer 16, 52-60 (1975) Nakamura, K., Sugano, H., Takagi, K., Fuchigami, A.: HistopathologicaI study on early carcinoma of the stomach: criteria for diagnosis of atypical epithelium. Gann 57, 613~520 (1966) Nakamura, K., Sugano, H., Takagi, K. : Carcinoma of the stomach in incipient phase: its histogenesis and histological appearances. Gann 59, 251-258 (1968) Oehlert, W., Keller, P., Henke, M., Strauch, M.: Die Dysplasien der Magenschleimhaut. Dtsch. Med. Wochenschr. 100, 1950-1965 (1975) Oehlert, W., Henke, M., Strauch, M., Keller, P.: Die Dysplasien der Magenschleimhaut, ihre Pathogenese und ihre klinische Bedeutung. Therapiewoche 28, 49664975 (1978)
264
E. Kunze et al.
Rumpf, P., Schacht, M., Palomba, P., Kramer, K., Schmitz, H., Borchard, F.: Chemisch induzierte Karzinogenese des Rattenmagens nach Vagotomie und Resektion. Z. Gastroenterol. 16, 85-94 (1978) Russo, A., Grasso, G., Sanfilippo, G., Giannone, G.: Criteri applicativi della biopsia mirata ed aspetti istologici nella diagnosi del carcinoma gastrico in situ. Tumori 62, 39-36 (1976) Saito, T., Inokuchi, L., Takayama, S., Sugimura, T.: Sequential morphological changes in N-methyiN-nitro-N-nitrosoguanidine carcinogenesis in the glandular stomach of rats. J. Natl. Cancer Inst. 44, 769-783 (1970) Serck-Hanssen, A., Marcussen, J., Liavag, L.: Endoscopic brush cytology in the diagnosis of gastric disease. Comparison with other diagnostic measures. In: Early gastric cancer. Grundmann, E., Grunze, H., Witte, S., eds., pp. 139-147. Berlin, Heidelberg, New York: Springer 1974 Schade, R.O.K.: Cancerisierung und Frfihkarzinom der Magenschleimhaut. Verh. Dtsch. Ges. Pathol. 45, 179-183 (1961) Schade, R.O.K.: The borderline between benign and malignant lesions in the stomach. In: Early gastric cancer. Grundmann, E., Grunze, H., Witte, S., eds., pp 45-53. Berlin, Heidelberg, New York: Springer 1974 Schauer, A., Kunze, E., Ober, M.: Early carcinoma des Magens im Tierexperiment. Deutscher Krebskongrel3 1974, Abstract Kongrel3berichte, S. 19. Miinchen: Demeter Verlag 1974 Schauer, A., Kunze, E.: Beziehungen der adenomat6sen Hyperplasie der Magenschleimhaut zur Cancerogenese. Akt. Gastrologie 7, 4177429 (1978) Schauer, A., Kunze, E.: Relation of adenomatous hyperplasia of the gastric mucosa to carcinogenesis. Path. Res. Pract. 164, 238-248 (1979) Schoental, R.: Carcinogenic activity of N-methyl-N-nitroso-N-nitroguanidine. Nature (Lond.) 209, 726-727 (1966) Schoental, R., Bensted, J.P.M. : Gastro-intestinal turnouts in rats and mice following various routes of administration of N-methyl-N-nitroso-N'-nitroguanidine and N-ethyl-N-nitroso-N'-nitroguanidine. Br. J. Cancer 23, 757-764 (1970) Shimosato, Y., Tanaka, N., Kogure, K., Fujimura, S., Kawachi, T., Sugimura, T.: Histopathology of tumors of canine alimentary tract produced by N-methyl-N'-nitro-N-nitrosoguanidine, with particular reference to gastric carcinoms. J. Natl. Cancer Inst. 47, 1053-1070 (1971) Shirai, T.: Periodical observations of the glandular stomach in rats treated with N-methyl-N'-nitroN-nitrosoguanidine. Nagoya Med. J. 19, 155-167 (1974) Sugani, H., Nakamura, K., Takagi, K. : An atypical epithelium of the stomach. Gann Monograph on Cancer Res. 11, 257 269 (1971) Sugimura, T., Nagao, M., Okuda, Y.: Carcinogenic action of N-methyl-N-nitro-N-nitrosoguanidine. Nature (Lond.) 210, 962-963 (1966) Sugimura, T., Fujimura, S., Kogure, K., Baba, T., Saito, T., Nagao, M., Hosoi, H., Shimosata, Y., Yokoshima, T.: Production of adenocarcinomas in glandular stomach of experimental animals by N-methyl-N-nitro-N-nitrosoguanidine. Gann Monograph 8, 157-176 (1969) Sugimura, T., Tanaka, N., Kawachi, T., Kogure, K., Fujimura, S., Shirnosato, Y.: Production of stomach cancer in dogs by N-methyl-N'-nitro-N-nitrosoguanidine. Gann 62, 67 (1971) Sugimura, T., Kawachi, T.: Experimental stomach cancer. Methods Cancer Res. 7, 245-308 (1973) Tahara, E., Haisuka, S., Tokuoka, S.: An ultrastrnctural study on precancerous and cancerous lesions of the glandular stomach in rats induced by N-methyl-N'-nitro-N-nitrosoguanidine. Acta Pathol. Jap. 26, 1-15 (1976) Takahashi, M., Shirai, T., Fukushima, S., Hahanouchi, M., Hirose, M., Ito, N.: Effect of fundic ulcers induced by iodoacetamide on development of gastric tumors in rats treated with N-methyl-N'nitro-N-nitrosoguanidine. Gann 67, 47-54 (1976) Teglbjaerg, P.S., Vetner, M.: Gastric carcinoma I. Acta Path. Microbiol. Scan& Sect. A 85, 519-527 (1977) Uchida, Y., Schlake, W., Roessner, A., Riihland, D., Themann, H., Grundmann, E.: Development of tumors in the glandular stomach of rats after oral administration of carcinogens. Z. Krebsforsch. 87, 199-212 (1976) Wiebecke, B., Gokel, J.M.: Experimentelle Cancerogenese des Magen-Darm-Kanals. Handbuch der Allgemeinen Pathologie, 6. Band. Geschwiilste. 7. Tell, S. 693-730. Berlin, Heidelberg, New York: Springer 1975 Received June 5, 1979/Accepted October 3, 1979
