- Email: [email protected]
The effect of intraperitoneal N-methyl-N-nitrosourea on hamster palatal gingiva and intermolar mucosa
Archs oral Bid. Vol. 39, No. 6, pp. 519-528, 1994 Copyright 0 1994Elsevier Science Ltd Printed in Great Britain. All rights reserved 0003-9969/94$7.00+ 0.00
Pergamon
THE EFFECT OF INTRAPERITONEAL N-METHYL-N-NITROSOUREA ON HAMSTER PALATAL GINGIVA AND INTERMOLAR MUCOSA M. D. ‘Department
MCMILLAN,’
C. J. SMITH*
and V.
RAMIREZ’
of Oral Biology and Oral Pathology, University of Otago, Dunedin, New Zealand. of Oral Pathology, University of Sheffield, Sheffield, U.K. and Wniversidad Autonoma Metropolitana-Xochimilco Mexico D.F., Mexico
*Department
(Accepted 21 December
1993)
Summary-Fifty 4- to 6-week-old male random-bred golden hamsters were injected intraperitoneally with a weight-related dose (12.5 mg/kg body weight) of N-methyl-N-nitrosourea (NMU) three times a week for 4 weeks. Groups of seven animals were killed 10, 16 and 22 weeks after the first injection. The palatal gingiva from six animals and the intermolar mucosa from 21 animals was examined. Seven male age-matched untreated control animals were killed at each period. Although all NMU-treated hamsters showed dysplastic and neoplastic changes similar to those in human oral squamous-cell carcinoma, other changes such as acantholytic dyskeratosis, invading cysts, duct-like structures and basaloid islands and cords were not. The extent and severity of the changes increased with time so that by 22 weeks there was extensive involvement of the palatal bone and marrow spaces, the molar periodontal ligament and the greater palatine neurovascular bundle by neoplastic epithelium. The invading epithelium was derived from the junctional, crevicular and palatal gingival and intermolar epithelium. The latent period for the crevicular and junctional epithelia was shorter than that for the palatal gingival and intermolar epithelium. The invasive changes from the latter epithelium were often preceded by exophytic changes such as epithelial projections, papillae and papillomas. Such changes were infrequent for the gingival, crevicular and junctional epithelia. The study shows that intraperitoneal NMU acts as a complete carcinogen on the palatal gingival and interrnolar epithehum in hamsters. Key words:
hamster,
N-methyl-N-nitrosourea,
palatal
INTRODUCTION
dimethylbenzanthracene; NMU, 4NQO,4-nitroquinoline-l-oxide.
neoplasia.
Oral lesions have been induced in a number of sites by either its topical (Milievskaja and Kiselevan, 1976; Chau and Edwards, 1984), intravenous (Herrold, 1968; Ketkar et al., 1977), subcutaneous (Herrold, 1966), intragastric (Herrold, 1968; Edwards, 1978, 1980; Edwards and Get-main, 1980a; Reibel, 1982; Suzuki, 1976; Kohgo et al., 1990) or intraperitoneal (Edwards, 1980; Edwards and Germain, 1980b; Konstantinidis, Smulow and Sonnenschein, 1982) administration in hamsters and rats. However, the results of these studies are equivocal and although the best route of administration appears to be intraperitoneal (Edwards and Germain, 1980b), a suitable dose rate able to induce intraoral neoplasms but compatible with survival of the animals has not been established. Our aim now was to determine the efficacy of intraperitoneal NMU for the induction of preneoplastic and neoplastic lesions in the intermolar platal mucosa and adjacaent gingiva in male hamsters because these areas appear to be some of the most susceptible to intraperitoneal NMU (Edwards and Get-main, 1980b).
Most experimental investigations into oral cancer and precancer use a topical carcinogen, with dimethylbenzanthracene (DMBA) application to the hamster cheek pouch being the most common (Eveson, 1981), although 4-nitroquinoline-l-oxide (4NQO) is also used in rats (Svensson and Heyden, 1982; Steidler, Rich and Reade, 1985; Rich and Reade, 1988), mice (Steidler and Reade, 1984, 1986) and hamsters (Eveson and MacDonald, 1977). However, some of the changes, especially the earlier ones, may be due, in part at least, either to the local irritant effect of the carcinogen (Eveson and MacDonald, 1978), to changes caused by the vehicle (White and AlAzzawi, 1983) or to a combination of these factors and thus are not part of the preneoplastic and neoplastic changes as such. In addition the hamster cheek pouch as a model of ‘intraoral’ carcinogenesis is criticized on a number of grounds (Eveson, 1981). The use of a systemically administered carcinogen that effects other areas would overcome these shortcomings. That N-methyl-N-nitrosourea (NMU) is an effective carcingoen for a number of anatomical sites following systemic and local administration is documented, although the number of studies is limited. Abbreviations: DMBA, methyl-N-nitrosourea;
mucosa,
MATERIALS
AND METHODS
Fifty 4-6-week-old random-bred male golden hamsters (Misocricetus auratus) were injected intraperitoneally with a weight-related dose (12.5 mg/kg body weight) of a freshly prepared O.O6rnol/l solution of
N-
519
520
M.D.
MCMILLAN
NMU in 0.1 M citrate buffer (pH 6.2) three times a week for 4 weeks. Thirty age- and sex-matched untreated control animals were used. All animals were carefully monitored and regularly weighed throughout the entire experimental period. Any animals that showed undue weight loss or appeared unwell were killed by an overdose of intraperitoneal pentobarbitone sodium followed by cervical dislocation. At 10, 16 and 22 weeks after the first injection, 11 anaesthetized animals from the experimental group and 10 from the control group had their maxillae removed. Seven animals from each time period were used in the present study. The remainder formed part of an ultrastructural study. The intermolar palatal mucosa and submucosa was surgically removed from five maxillae from each group at each time period and placed in 10% neutral-buffered formalin. The other maxillae, which contained the palatal gingiva as well as the intermolar tissues, were fixed whole. After fixation the maxillae were demineralized in a solution of formalin and formic acid. The separated mucosal specimens were then divided in half along the midline of the hard palate. The maxillae were divided in half transversely through the centre of the left and right second molar teeth. All specimens were processed routinely for light microscopy and blocked on their divided surfaces. Serial sections were cut with four sections being stained with haematoxylin and eosin every 20pm for the maxillary and every 10 pm for the mucosal specimens until the whole specimen was sectioned. The exophytic, dysplastic and invasive lesions in sections from NMU-treated animals were counted and the severity of each dysplastic lesion was scored. Care was taken not to record a lesion more than once. The score was based on how many of the following histological changes were present: loss of basal-cell polarity; multiplication of the basal-cell layer; increased nuclear-cytoplasmic ratio; drop-shaped rete ridges; irregular stratification; increased number, suprabasal or abnormal mitotic figures; cellular pleomorphism, nuclear hyperchromatism; dilated intercellular spaces; abnormal cell keratinization either singly or in groups. A number of histological levels were examined for most lesions. The score for a lesion was the sum of the different changes seen in all the sections in which it was found. For example, if in the first section there were three changes and if in the second section there were the same three changes the lesion was scored as having three changes (i.e. +). However, if two of the three changes in the second level were different to those in the first level the lesion was scored as having five changes (i.e. + + ): + 3 or 4; + + 5 or 6; + + + 7 or 8; + + + + 9 or more. No change in a given lesion was scored more than once. Limited autopsies were performed on the hamsters. Specimens for light microscopy were removed from any gross pathology and from the lungs, liver, spleen, pancreas, kidney, heart, forestomach, gastric stomach, duodenum and ileum. RESULTS
Control animals: No systemic pathology control animals.
was found
in any of the
e/ al.
The structure of the intermolar palatal mucosa and submucosa in all control specimens was similar. An orthokeratinized stratified squamous epithelium some 620 nucleated cells thick was present throughout except for a few small areas that were parakeratinized. Over the midline of the anterior two-thirds and all of the posterior one-third of the palate the epithelum was thinner with no or minimal rete ridges (Fig. 1). Rete ridges were more pronounced in the thicker epithelium elsewhere. The thickness of the stratum corneum matched that of the epithelium as a whole and usually had a smooth oral surface, although over the minor salivary glands it had a number of surface projections (Fig. 2). There were scattered mitotic figures in the basal cell layer and in the cells of the basally situated stratum spinosum. Most of the lamina propria of the intermolar palatal mucosa was a relatively dense fibrous connective tissue. That making up the rugae usually had an almost myxomatous appearance.
0.4mm Fig. 1. Control. Anterior palatal mucosa with thinner stratified squamous epithelium (E) in the midline (M). Epithelial rete ridges (arrows); regular stratum corneum (C); gingival epithelium (G); molar tooth root (R); periodontal ligament (P); lamina propria (L); greater palatine neurovascular bundle (N); palatal bone (Z). Cross-section. Haematoxylin and eosin. x40
Chemical
carcinogen&
in hamster palate
521
granulation-tissue formation bone loss, and a moderate to heavy infiltrate of neutrophilic leucocytes. NMU-Treated
animals
Initially all NMU-treated animals lost weight but by 6 or 7 weeks after the first injection most had regained this and they showed a steady weight gain for the rest of the experimental period. However, nine animals had to be killed in the first 10 weeks because of undue weight loss, with eight animals being killed later because of a reversal of their weight gain. With the exception of five of these latter animals, three of which had ischaemic necrosis of segments of the ileum and two of which had abdominal carcinomas derived from the skin, autopsy did not reveal the cause of the undue weight loss. The only constant findings in the other NMU-treated animals were inflammation of the lungs with thickening of the alveolar walls and areas of epithelial hyperplasia, hyperkeratosis and dysplasia, along with multiple papillomas of the forestomach. The severity of these changes increased with increasing age. All 33
Fig. 2. Control. Posterior palatal mucosa with epithelial surface projections (S); orthokeratinized stratified squamous epithelium (E). Minor salivary glands (A) in the lamina propria (L). Longitudinal section. Haematoxylin and eosin. x 160
The submucosa, a less fibrous, more cellular and vascular variant of the lamina propria, was minimal in amount except in the area of the greater palatine neurovascular bundle and in the posterior third of the specimens. Here it contained varying amounts of adipose tissue anteriorly and minor salivary glands posteriorly. The structure of the palatal gingival and crevicular tissues was similar in nearly all the control specimens and resembled that of the adjacent palatal mucosa, although the crevicular epithelium had only an occasional rete ridge and a thinner keratin layer. The junctional epithelium, which often extended for a short distance on to the cementum, overlay much of the crevicular epithelium and separated it from the tooth. In some sections from all specimens the intercellular spaces of the crevicular epithelium and the adjacent connective tissue contained a small number of neutrophilic leucocytes. In two specimens there was an area where foreign material was impacted in the gingiva and underlying periodontal ligament. This was associated with epithelial proliferation,
F?g. 3. Ten weeks NMU. Thickened epithelium with plug of nnainly parakeratinized cells (K). Individual cell keratinizaion (arrowheads); dilated intercellular spaces (arrows); ascular lamina propria (L). Longitudinal section. Haematoxylin and eosin. x 160
522
M. D. MCMILLAN et al.
remaining animals had lesions in the palatal gingiva and intermolar mucosa when killed. Grossly, the mucosa of the hard palate in all IO-week NMU animals was more irregular than in the controls, although the form of the rugae was still evident. This irregularity was caused by variable numbers of small surface projections, depressions, and the occasional distinct papilloma, and by an increased amount of epithelial desquamation. There was no regular pattern to the distribution of these changes. Similar changes were found in all the 16and 22-week animals but with increasing age they were more pronounced, involved a greater proportion of the mucosa and the papillomas were more frequently seen. At 22-weeks the mucosal surface was often so irregular than the normal rugal pattern was lost. Although the surface of the palatal gingiva in all animals was more irregular than in the controls, this was due to the presence of depressions and epithelial desquamation. Surface projections were found only infrequently.
0.1 mm Fig. 5. Ten weeks NMU. Irregular surface epithelium (E,) covering an inflamed Iamina propria (L) that contains deeper epithelial islands (E,). Dilated thin-walled vessels (W), one containing epithelium (arrowhead). Nautrophilic leucocytes (arrows). Longitudinal section. Haematoxylin and eosin. x 160
Fig. 4. Ten weeks NMU. Inflamed lamina propria (L) covered by an irregular, thickened epithelium (E). Indistinct epithelialeonnective tissue junction (arrowheads); neutrophihc Ieucocytes (arrows). Longitudinal section. Haematoxylin and eosin. x 160
Histological changes seen in animals treated with NMU for 10, 16 and 22 weeks are illustrated in Figs 3-5, 6-8 and 9-14, respectively. All animals showed varying numbers of exophytic (Figs 6, 7 and 14) dysplastic (Figs 46 and 8) and invasive (Figs 5,9-14) lesions. These were derived from either the gingival or intermolar epithelium (Table 1). The incidence and severity of the dysplastic lesions for each time period are listed in Table 2. As well as differences in the number of exophytic, dysplastic and invasive lesions present at each time period, there were also differences in the depth of invasion for invading lesions, and in the amount of surface epithelium involved. Both of these features increased with increasing age, with the latter feature being confined to lesions derived from the intermolar mucosa. Also, in the lo- and 16-week specimens, invasive lesions derived from the crevicular and junctional epithelium had penetrated to a greater depth than those derived from the palatal gingival and intermolar mucosa. In the
523
Chemical carcinogenesis in hamster palate
that at times had a basaloid appearance (Fig. 11). At 10 weeks the majority of the invading epithelium was confined to the more laterally situated lamina propria and submucosa along with the more superficial periodontal tissues. Some of the invading epithelium had involved the alveolar and palatal bone, the greater palatine neurovascular bundle and the palatal marrow spaces in the 16-week specimens. Also, in some instances involvement of the periodontal ligament extended to the root apices of the molar teeth. By 22 weeks the involvement of these deeper structures was more extensive (Figs 9-l 1, 13 and 14). In the more posterior areas of the specimens, although some of the invading tissue was close to the salivary glands, the glands themselves were normal. The associated connective tissue in all NMU specimens were usually more cellular and vascular than in control specimens. It contained numerous plump fibroblasts and varying numbers of inflammatory cells (Figs 4, 5 and 12). Dilated vessels were often in close association with the surface and invading
tion
0.2mm Fig. 6. Sixteen weeks NMU. Area of markedly thickened irregular surface epithelium (E) that has numerous reteridges extending into the underlying lamina propria. Surface projections (arrows). Longitudinal section. Haematoxylin and eosin. x 80
22-week specimens this difference was not as marked. The invasive lesions had a similar range of histological appearances, irrespective of either their origin or whether they were in the lo-, 16- or 22-week specimens. There were cords, strands, islands and single cells or small groups of epithelium within the connective tissue (Figs 5, 6 and 9-14). Although a number of the epithelial islands appeared to have undergone cystic change (Figs 9 and 14), examination of a number of histological levels showed that they were, in fact, deep extensions from the surface epithelium that had a patent lumen and formed duct-like structures (Figs 9, 10 and 14). In the 16- and 22-week specimens some of the invading epithelium appeared at least partly degenerate. The intercellular spaces of the degenerate epithelium usually contained neutrophilic leucocytes (Fig. 9). The degree of differentiation of the invading epithelium varied from well-differentiated stratified squamous epithelium showing regular otthokeratinization, to less-differentiated epithelium lacking any evidence of keratiniza-
0.4mm Fig. 7. Sixteen weeks NMU. Area of more exophytic surface epithelium (E) having a very irregular surface. Gingival epithelium (G); fibrous lamina propria (L); greater palatine neurovascular bundle (N); palatal bone (Z). Cross-section. Haematoxylin and eosin. x 55
524
M. D. MCMILLAN rr al.
extended down to the basal cell layer. Although the epithehum below some of the more superficial keratin plugs was normal, in many instances it was dysplastic (Figs 3 and 8). A striking feature of all NMU specimens was the individual cell keratinization, which was found at all levels in the epithelium (Figs 3 and 8). At times only one or a small group of keratinized cells were present in the stratum spinosum, while elsewhere vertical columns of these cells extended throughout part or all of the epithelium. Some of these columns were only one cell wide while others were more extensive. Cells showing individual keratinization were round in outline, had very eosinophilic cytoplasm, were surrounded by a much wider intercellular space than elsewhere in the epithehum and often appeared to lack contact with adjacent cells. At times the dilated intercelhtlar space extended from the surface of the stratum corneum down through the basal cell layer to the epithelialLconnective tissue junction. Many of the nuclei of the keratinized cells appeared similar to
;-2.
,I
0.1 mm Fig.
8. Sixteen
weeks
NMLJ.
Very irregular
surface
epi-
thelium showing dilated intercellular spaces (arrow). vidual stratum
indicell keratinization (arrowheads) and thickened corneum (C). Longitudinal section. Haematoxylin and eosin. x 160
epithehum (Figs 5 and 12) and at times appeared to contain epithehal cells (Fig 5). The surface epithelium in all NMU specimens was often thickened and showed acanthosis and usually hyperorthokeratosis, although there were also areas of hyperparakeratosis. The amount of epithelium involved and its thickness increased with increasing age of the specimens and by 22 weeks the connective tissue component of the rete ridges and the lamina propria was at times replaced by hyperplastic epithehum. Although some of this acanthotic epithehum was dysplastic, some was not. The exposed oral surface of the intermolar mucosa and palatal gingiva in all NMU specimens WdS irregular and this irregularity increased with increasing age of the specimens. In part this irregularity was caused by the exophytic lesions (Table I) (Figs 6, 7 and 14) and in part by the oral openings of the duct-like structures (Fig. 14) irregular desquamation and the presence of keratin plugs (Figs 3 and 8). The number and size of the keratin plugs also increased with increasing age of the specimens and at times they
Fig. 9. Twenty-two weeks NMU. Extensive epithelial invasion (arrows) involving the palatal bone (Z) and periodontal ligament (P). Degenerate epithelium (arrowheads); cystic change (D); greater palatine neurovascular bundle (N); molar tooth root (R). Cross-section. Haematoxylin and eosin. x 60
Chemical
carcinogenesis
Fig. IO. Twenty-two weeks NMU. Lateral area with markedly thickened surface epithelium (E) and apparent keratin pearls (K). Invading gingival epithelium with lumen (arrows); area of bone resorption (arrowhead); periodontal ligament (P). Cross-section. Haematoxylin and eosin. x 45
in hamster
palate
525
Fig. 12. Twenty-two weeks NMU. Invading epithelium (arrows) deep in lamina propria. Blood vessels (V). Longi-
tudinal section. Haematoxylin and eosin. x 250
NMU in age- and sex-matched hamsters to induce preneoplastic and neoplastic change in the palatal those of the surrounding epithelium but others were gingival and intermolar epithelium. It also provides a either pyknotic or contained aggregations of chrodetailed description of the histological changes inmatin. duced over an extended period and confirms that The epithelium in NMU specimens did, at times, intraperitoneal NMU is not associated with the high also show some of the following: acantholysis; columincidence of carcinoma of the forestomach, and connar basal cells with very eosinophilic cytoplasm; sequent high death rate, that the intragastric route of aggregations of sebaceous-like cells in the stratum administration is (Edwards and Germain, 1980a, b). spinosum. The reasons for the deterioration in health of 12 of the 17 hamsters that had to be killed prematurely were not from the This could DISCUSSION Fig. apparent 12. Twenty-two weeksautopsies. NMU. Invading epithelium Fig. IO. Twenty-two weeks NMU. Lateral area with only be attributed to ill-defined chronic effects (V). of Longi(arrows) deep in lamina propria. Blood vessels markedly thickened surface epithelium (E) and apparent The keratin structure theInvading intermolar mucosa, NMU cytotoxicity (Edwards and Get-main, 1980b), tudinal section. Haematoxylin and eosin. x 250 pearls of(K). gingivalpalatal epithelium with lumen submucosa and area palatal gingiva in the (arrowhead); control animals some of the earlier problems could have (arrows); of bone resorption periodontalalthough was similar of the rats (McMillan, ligamentto that (P). Cross-section. Haematoxylin 1979). and The eosin. xbeen 45 caused immunosuppression NMU byin the age-transient and sex-matched hamsters that to induce reason for the localized areas of parakeratinized NMU preneoplastic induces (Herrold, 1966; Leaver, Swann and palatal and neoplastic change in the epithelium but they could be thebut result of were Magee, gingival 1969; Edwards and Germain, those isofunknown the surrounding epithelium others and intermolar epithelium.1980b). It also provides a previouseither trauma. examination of crossand histological longitudinal changes secpyknotic or contained aggregations of chro- The detailed description of the inThe matin. present study is the first to use a standardized tions atduced regularover intervals through period the palatal an extended and gingival confirms that intraperitoneal dose of in the The epithelium NMUcomplete specimenscarcingoen did, at times, intraperitoneal NMU is not associated with the high also show some of the following: acantholysis; columincidence of carcinoma of the forestomach, and connar basal cells with very eosinophilic cytoplasm; sequent high death rate, that the intragastric route of aggregations of sebaceous-like cells in the stratum administration is (Edwards and Germain, 1980a, b). spinosum. The reasons for the deterioration in health of 12 of the 17 hamsters that had to be killed prematurely were not apparent from the autopsies. This could DISCUSSION only be attributed to ill-defined chronic effects of The structure of the intermolar palatal mucosa, NMU cytotoxicity (Edwards and Get-main, 1980b), submucosa and palatal gingiva in the control animals although some of the earlier problems could have was similar to that of the rats (McMillan, 1979). The been caused by the transient immunosuppression that reason for the localized areas of parakeratinized NMU induces (Herrold, 1966; Leaver, Swann and epithelium is unknown but they could be the result of Magee, 1969; Edwards and Germain, 1980b). previous trauma. The examination of cross- and longitudinal secThe present study is the first to use a standardized tions at regular intervals through the palatal gingival intraperitoneal dose of the complete carcingoen
Fig. I I. Twenty-two weeks NMU. Epithelial deep within the palatal bone(Z). Cross-section. lin and eosin. x 150
invasion (E) Haematoxy-
Fig. I I. Twenty-two weeks NMU. Epithelial deep within the palatal bone(Z). Cross-section. lin and eosin. x 150
Fig. 13. Twenty-two weeks NMU. Invading epithelium (arrows) deep in lamina propria associated with the greater palatine nerve (N) and blood vessel (V); epithelial cyst (D). Longitudinal section. Haematoxylin and eosin. x 160
invasion (E) Haematoxy-
Fig. 13. Twenty-two weeks NMU. Invading epithelium (arrows) deep in lamina propria associated with the greater palatine nerve (N) and blood vessel (V); epithelial cyst (D). Longitudinal section. Haematoxylin and eosin. x 160
M. D. MCMILLAN
526
et al.
Table 1. Number of exophytic (E), dysplastic (D) and invasive (I) lesions in NMU-treated animals Palatal Gingiva Time (weeks) 10 16 22
Number of animals/ (sections) 2 (368) 2 (392) 2 (376)
E D
Intermolar Mucosa
I
2 2 5 2 4 7 - - 6
Number of animals/ (sections) I (388) 7 (395) 7 (406)
E D
I
21 21 9 45 33 19 19 20 27
Exophytic lesions were either conical projections of the stratum corneum, larger papillae-like structures that had a core of nucleated epithelial cells or papillomas with a central connective tissue core. Invasive lesions showed frank invasion of the connective tissue by the epithelium. in the present study in that what in some sections from NMU-treated hamsters appeared to be isolated islands and cords of invading epithelium were in fact in continuity with the surface epithelium in sections taken at a different level. This also enabled the origin of the invading epithelium to be determined. Similarly, what in some sections appeared as deeply situated cysts, which have also been described in oral mucosa following NMU administration (Suzuki, 1976; Edwards and Germain, 1980a; Kohgo et al., 1990) were usually found to be duct-like structures that opened into the mouth. Cysts or duct-like structures are not a feature of human oral squamous-cell carcinoma, or of experimental oral carcinoma in animals induced by either 4-NQO (Svensson and Heyden, 1982; Philipsen and Fisker, 1983; Rich and Reade, 1988) or DMBA (Salley, 1957; Mori et al., 1962; Skhlar et al., 1980; White, Gohari and Smith, 1981). Individual cell keratinization is found in dysplasia and carcinoma in human oral epithelium. However, the extensive areas of individual cell keratinization that often resembled the focal acantholytic dyskeratosis of Darier’s disease (Philipsen and Fisker, 1983) seen in this and other studies employing NMU (Suzuki, 1976; Edwards, 1978; Edwards and Germain, 1980a, b; Chau and Edwards, 1984) are not. Similar changes are found in 4-NQO; (Svensson and Heyden, 1982; Philipsen and Fisker, 1983) and occasionally DMBA- (Eveson and MacDonald, 1977. 1981) treated oral mucosa. None of these studies provided evidence that these changes were related to the subsequent development of malignancy. Initially, multiple exophytic or invasive lesions involved the palatal gingival and intermolar tissues. Before 22 weeks, however, lesions associated with the former tissues were mostly invasive, and were derived from the junctional and crevicular epithelium, while demonstrated
Fig. 14. Twenty-two weeks NMU. Exophytic surface projection (E,); keratin-filled lumen (arrowheads); deeper invading epithelium (EJ; cystic change (D); greater palatine neurovascular bundle (N); individual cell keratinization (arrows). Longitudinal section. Haematoxylin and eosin. x 50 and intermolar tissues, as undertaken in this study, is essential when a mucosa of variable structure is used in investigations into experimental preneoplasia and neoplasia. This is because normal structural variations may be mistaken for pathological change even in an apparently homogeneous mucosa (McMillan and Kerr, 1990). In addition, the presence and true organization of the pathological changes may not be apparent from either macroscopic examination or a small number of histological sections. This is amply
Table 2. Incidence and severity of change in the dysplastic lesions in NMU-treated Palatal Gingiva ____ Time (weeks) 10 16 22
Number of animals/ (sections) 2 (368) 2 (392) 2 (376)
+
++
+++
-
2
--
1
‘)
I
Intermolar Mucosa
~~~~
++++
animals (see p. 520)
Number of animals/ (sections)
+
7 (388) 7 (395) 7 (406)
15 20 8
++
+++
++++
4 8 4
2 3 8
2
Chemical carcinogenesis in hamster for the palate exophytic lesions predominated. Similar findings for exophytic and invasive lesions following NMU administration are described for the gingiva (Suzuki, 1976) and hard palate (Edwards and Germain, 1980a, b) but Kohgo et al. (1990) describe only invasive gingival carcinomas. A similar disparity in the latent period before invasive lesions predominated between the gingiva and hard palate, as found in the present study, has also been reported (Edwards, 1978; Edwards and Germain, 1980a) as has the fact that the initial changes in the gingiva involve the junctional and crevicular epithelium (Suzuki, 1976; Kohgo et al., 1990) following intragastric NMU. It is known that stimulated proliferation may play a part in the initiation, promotion and propagation of carcinogenesis (Craddock, 1976; Kunze and Pacha, 1985; Kunze et al., 1989) and that continuous, non-specific irritation may serve as a locus for experimental tumour formation in skin (Argyris, 1980), the gastrointestinal tract (Pozharisski, 1975; Barthold and Beck, 1980) and buccal mucosa (Konstantinidis et al., 1982). Although it has been suggested that impaction of hair, bedding and food in the gingival crevice of NMU-treated animals may be a cocarcinogenic factor (Edwards, 1978), this was only rarely observed in the present study, which confirms the observtions of Suzuki (1976) that this form of local irritation is not a factor. Much of the invading epithelium was in the form of distinct cords, islands and cysts or duct-like structures that were composed of well-differentiated, malignant squamous epithelium. This was irrespective of whether they were derived from junctional, crevicular, gingival or palatal epithelium, which confirms the findings of Herrold (1968); Suzuki (1976); Edwards and Germain (1980a, b) and Kohgo et al. (1990). However, by 22 weeks some of the larger epithelial islands and cords were less well differentiated and occasionally basaloid in appearance. This again was irrespective of their origin. This latter appearance is also noted by Edwards and Germain (1980a, b). Kohgo et al. (1990), in their study of NMU-induced changes in the gingiva and endentulous alveolar ridge, describe invasion by nests of relatively poorly differentiated epithelium. Similar nests and even apparently individual epithelial cells were present not only in the connective tissue of the hard palate, gingiva and periodontal ligament in the present study, but also within the lumen on thin-walled vessels. These features are not described by Herrold (1968), Suzuki (1976) or Edwards and Germain (1980a, b). Whether the vessels containing the epithelial cells were blood or lymphatic could not be determined but none contained red blood cells. The presence of epithelial cells in vessels in experimental oral carcinogenesis is only rarely reported (Salley and Kreshover, 1959; Herrold, 1968; Eveson and MacDonald, 1981; Shingaki, Otake and Nakajima, 1987) as is regional lymph-node involvement (Salley, 1954; Rwomushana, Polliack and Levij, 1970; Eveson and MacDonald, 1981; Safour et al., 1984; Shingaki et al., 1987). The potential of NMU-induced oral carcinomas to spread to distant sites requires further investigation. Although some of the invading epithelium that was associated with an inflammatory response appeared
527
palate
to be becoming degenerate, this was not common. Such changes could indicate that this epithelium was being destroyed. However, whether with time such a process would involve all of the invading epithelium is problematical. However, inflammation was also associated with the changed but viable surface epithelium in NMU-treated animals and with the viable invading epithelium. Whether the inflammatory response was directed towards the changed epithelium or not is unknown. What is known is that at 22-weeks viable invading epithelium was present deep in the periodontal ligament and maxillary bone and marrow and was associated with the intrabony part of the greater palatine neurovascular bundle. The increased vascularity often associated with the altered surface and invading epithelium is probably indicative of the angiogenic response characteristics of carcinogenesis (Folkman, 1978, 1986; Ferguson and Smillie, 1979; White and Al-Azzawi, 1983; Polverini, Shimizu and Solt, 1988), although part at least may have been related to the inflammatory response. We have clearly shown that our experimental regimen resulted in preneoplastic and neoplastic change in the stratified squamous epithelium of the maxillary intermolar mucosa and adjacent gingiva in hamsters. Although changes such as acantholytic dyskeratosis, cysts and duct-like structures are not typical of human oral squamous-cell carcinoma, our protocol provides a suitable model for the investigation of oral carcinogenesis. It also allows investigation of invasion of bone by malignant oral epithelium, a feature that is not seen in experimental
hamster cheek-pouch carcinogenesis. The use of an experimental system employing a systemic carcinogen rather than one of the two commonly used topical ones (i.e. DMBA and 4-NQO) is important because it eliminates any local irritant effect of the carcinogen or vehicle. In addition, it allows the comparison of changes induced by three rather than two oral carcinogens. Such comparison is necessary to determine that any changes induced are indicative of the preneoplastic and neoplastic process and are not just peculiar to a particular carcinogen. In addition, the NMU model described is simple, does not require extended carcinogen treatment, and may also provide a useful tool to investigate aberrations in keratinization. AcknowledRemenrs-This study was in part undertaken while one of the authors (M. D. McMillan) held a Commonwealth Medical Fellowship in the Department of Oral Pathology, University of Sheffield, while V. Ramirez was a graduate student in this department. D. Thompson, H. Walker and D. Teale of Sheffield and A. Wiles and J. Thomson of Dunedin are thanked for their excellent technical assistance. Mrs P. K. Cosgriff is thanked for typing the manuscript.
REFERENCES
Argyris T. S. (1980) Tumor promotion by abrasion induced epidermal hyperplasia in the skin of mice. J. invest. Derm. 75, 36&362. Barthold S. W. and Beck D. (1980) Modification or early dimethylhydrazine carcinogenesis by colonic mucosal hyperplasia. Cancer Res. 40, 44514455.
528
M. D. MCMILLAN et al.
Chau K. K. and Edwards M. B. (1984) Topical carcinogenesis by N-methyl-N-nitrosourea in Syrian hamster cheekpouch and oral mucosa. Archs oral Biol. 29, 185-190. Craddock V. M. (1976) Cell proliferation and experimental liver cancer. In Liver Cancer (Eds Cameron H. M. and Linsell C. A.), pp. 153-201. Elsevier, Amsterdam. Edwards M. B. (1978) Some effects of N-methylnitrosourea on the oral and odontogenic tissues of the Syrian golden hamster. Archs oral Biol. 23, 515-524. Edwards M. B. (1980) Effects of N-methylnitrosourea on oral and dental tissues in the inbred Lewis rat. Archs oral Biol. 25, 5965. Edwards M. B. and Germain J. P. (1980a) Oral carcinogenesis by intragastric N-methyl-N-nitrosourea in the Syrian hamster. Archs oral Biol. 25, 623429. Edwards M. B. and Germain J. P. (1980b) Oral carcinogenesis by intraperitoneal N-methyl-N-nitrosourea in the Syrian hamster. Archs oral Biol. 25, 631-633. Eveson J. W. (1981) Animal models of intra-oral chemical carcinogenesis: a review. J. oral Parh. 10, 129-146. Eveson J. W. and MacDonald D. G. (1977) Effects of the water-soluble carcinogen 4-nitroquinoline N-oxide on hamster lingual mucosa. Oral Surg. 44, 60&605. Eveson J. W. and MacDonald D. G. (1978) Quantitative histological changes during early experimental carcinogenesis in the hamster cheek pouch. Br. J. Derm. 98, 639644. Eveson J. W. and MacDonald D. G. (1981) Hamster tongue carcinogenesis. I. Characteristics of the experimental model. J. oral Path. 10, 322-331. Ferguson J. W. and Smillie A. C. (1979) Vascularization of premalignant lesions in carcinogen-treated hamster cheek pouch. J. natn. Cancer Inst. 63, 138331392. Folkman J. (1978) Tumour angiogenesis and tumour immunity. In immunological Aspects of Cancer (Ed. Castro J. E.), pp. 267-282. University Park Press, Baltimore, MD. Folkman J. (I 986) How is blood vessel growth regulated in normal and neoplastic tissue? Cancer Res. 46, 467473. Herrold K. M. (1966) Carcinogenic effect of N-methyl-Nnitrosourea administered subcutaneously to Syrian hamsters. J. path. Bact. 92, 35-41. Herrold K.. M. (1968) Odontogenic tumors and epidermoid carcinomas of the oral cavitv. Oral Surp. 25, 262-272. Ketkar M., Reznik G., Haas H., HilfrichJ. and Mohr U. (1977) Tumors of the heart and stomach induced in European hamsters by intravenous administration of N-methyl-N-nitrosourea. J. natn. Cancer Inst. 58, 1695-1697. Kohgo T., Uy G. H., Iizuka T., Shindoh M. and Amemiya A. (1990) Histopathologic study of squamous cell carcinoma of the gingiva in edentulous and dentulous jaws of hamsters treated with N-methylnitrosourea, J. oral Path. 19, 202-207. Konstantinidis A., Smulow J. B. and Sonnenschein C. (1982) Tumorigenesis at a predetermined oral site after one injection of N-nitroso-N-methylurea. Science 211, 1235-1237. Kunze E. and Pacha H. (1985) Modifikation der urothelkarzinogenese durch proliferationsstimulation. In Experimentelle Urologic (Eds Harzmann R., Jacobi G. H. and WeiBach L.). pp. 419429. Springer, Heidelberg. Kunze E., Graewe T., Scherber S., Weber J. and Gellhar P. (1989) Cell cycle dependence of N-methyl-N-nitrosoureainduced tumour development in the proliferating, partially resected rat urinary bladder. Br. J. exp. Path. 70, 1255142. Leaver D. D., Swann P. F. and Magee P. N. (1969) The induction of tumours in the rat by a single oral dose of N-nitrosomethylurea. Br. J. Cancer 23, 1777187. McMillan M. D. (1979) The surface structure of the completely and incompletely orthokeratinized oral epithelium in the rat: a light, scanning and transmission electron microscope study. Am. J. Anat. 156, 337-351.
McMillan M. D. and Kerr M. A. (1990) A light and scanning electron microscope study of epithelial thickenings and rete ridges in the adult hamster cheek pouch. Archs oral Biol. 35, 235-240. Milievskaja I. L. and Kiseleva N. S. (1976) Comparative study of the carcinogenic activities of nas and some chemical carcinogens when introduced into the buccal pouch of the Syrian hamster. Bull. Wld Hlth Organ 54, 607414. Mori M., Miyaji T., Murata I. and Nagasuna H. (1962) Histochemical observations on enzymatic processes of experimental carcinogenesis in hamster cheek pouch. Cancer Res. 22, 1323-1326. Philipsen H. P. and Fisker A. V. (1983) Focal acantholytic dyskeratosis in experimental oral carcinogenesis in rats. J. oral Path. 12, 3(r36. Polverini P. J., Shimizu K. and Solt D. B. (1988) Control of angiogenic activity in carcinogen-initiated and neoplastic hamster pouch keratinocytes and their hybrid cells. J. oral Path. 17, 522-527. Pozharisski K. M. (1975) The significance of nonspecific injury for colon carcinogenesis in rats. Cancer Res. 35, 38243830. Reibel J. (1982)Carcinogenesis in forestomach and changes in oral mucosa of rats induced by intragastric N-methylnitrosourea. &and. J. dent. Res. 90, 382-389. Rich A. M. and Reade P. C. (1988) Histomorphometric analysis of epithelial changes in chemically induced oral mucosal carcinogenesis in rats. J. oral Path. 17, 528-533. Rwomushana J. W., Polliack A. and Levij I. S. (1970) Cervical lymph node metastasis of hamster cheek pouch carcinoma induced with DMBA. J. dent. Res. 49. 184. Safour 1. M., Wood N. K., Tsiklakis K., Doemling D. B. and Joseph G. (1984) Incisional biopsy and seeding in hamster cheek pouch carcinoma. J. denf. Res. 63, 1 I161 120. Salley J. J. (1954) Experimental carcinogenesis in the cheek oouch of the Svrian hamster. J. dent. Res. 33. 253-261. Sailey J. J. (1957)Histologic changes in the hamster cheek pouch during early hydrocarbon carcinogenesis. J. dent. Res. 36, 48-55. Salley J. J. and Kreshover S. J. (1959) The effect of topical application of carcinogens on the palatal mucosa of the hamster. Oral Surg. 12, 501-508. Shingaki S., Otake K. and Nakajima T. (1987) Regional lymph node metastasis in carcinoma of the hamster tongue. Oral Surg. 64, 190-196. Shklar G., Schwartz J., Grau B. A., Trickier D. P. and Wallace K. D. (1980) Inhibition of hamster buccal pouch carcinogenesis by 13-cis-retinoic acid. Oral Surg. 50, 45552. Steidler N. E. and Reade P. C. (1984) Experimental induction of oral squamous cell carcinomas in mice with 4-nitroauinolinel-oxide. Oral Surp. 57, 524-53 1. Steidler A. E. and Reade P. C. T1986) Initiation and promotion of experimental oral mucosal carcinogenesis in mice. J. oral Path. 15, 4347. Steidler N. E., Rich A. M. and Reade P. C. (1985) Experimental induction of preneoplastic and neoplastic changes in the lineual mucosa of rats. J. Biol. buccale 13,339-346. Suzuki M. (1976) Histological studies of early changes of the hamster gingival epithelium by N-nitrosomethylurea. Bull. Tokyo med. dent. Univ. 23, 115-134. Svensson S. and Heyden G. (1982) Experimental induction of irreversible precancerous changes in the palatal epithelium of the rat. Int. J. oral. Sura. 11, 52258. White F. H. and Al-Azzawi B. (1983)‘Vascularity in experimental oral neoplasia: a stereological approach. Acta Sterrol. 2/l, 119-126. White F. H., Gohari K. and Smith C. J. (1981) Histological and ultrastructural morphology of 7, 12, dimethylbenz(a)-anthracene carcinogenesis in hamster cheek pouch epithelium. Diagn. Histopath. 4, 307-333
Pergamon
THE EFFECT OF INTRAPERITONEAL N-METHYL-N-NITROSOUREA ON HAMSTER PALATAL GINGIVA AND INTERMOLAR MUCOSA M. D. ‘Department
MCMILLAN,’
C. J. SMITH*
and V.
RAMIREZ’
of Oral Biology and Oral Pathology, University of Otago, Dunedin, New Zealand. of Oral Pathology, University of Sheffield, Sheffield, U.K. and Wniversidad Autonoma Metropolitana-Xochimilco Mexico D.F., Mexico
*Department
(Accepted 21 December
1993)
Summary-Fifty 4- to 6-week-old male random-bred golden hamsters were injected intraperitoneally with a weight-related dose (12.5 mg/kg body weight) of N-methyl-N-nitrosourea (NMU) three times a week for 4 weeks. Groups of seven animals were killed 10, 16 and 22 weeks after the first injection. The palatal gingiva from six animals and the intermolar mucosa from 21 animals was examined. Seven male age-matched untreated control animals were killed at each period. Although all NMU-treated hamsters showed dysplastic and neoplastic changes similar to those in human oral squamous-cell carcinoma, other changes such as acantholytic dyskeratosis, invading cysts, duct-like structures and basaloid islands and cords were not. The extent and severity of the changes increased with time so that by 22 weeks there was extensive involvement of the palatal bone and marrow spaces, the molar periodontal ligament and the greater palatine neurovascular bundle by neoplastic epithelium. The invading epithelium was derived from the junctional, crevicular and palatal gingival and intermolar epithelium. The latent period for the crevicular and junctional epithelia was shorter than that for the palatal gingival and intermolar epithelium. The invasive changes from the latter epithelium were often preceded by exophytic changes such as epithelial projections, papillae and papillomas. Such changes were infrequent for the gingival, crevicular and junctional epithelia. The study shows that intraperitoneal NMU acts as a complete carcinogen on the palatal gingival and interrnolar epithehum in hamsters. Key words:
hamster,
N-methyl-N-nitrosourea,
palatal
INTRODUCTION
dimethylbenzanthracene; NMU, 4NQO,4-nitroquinoline-l-oxide.
neoplasia.
Oral lesions have been induced in a number of sites by either its topical (Milievskaja and Kiselevan, 1976; Chau and Edwards, 1984), intravenous (Herrold, 1968; Ketkar et al., 1977), subcutaneous (Herrold, 1966), intragastric (Herrold, 1968; Edwards, 1978, 1980; Edwards and Get-main, 1980a; Reibel, 1982; Suzuki, 1976; Kohgo et al., 1990) or intraperitoneal (Edwards, 1980; Edwards and Germain, 1980b; Konstantinidis, Smulow and Sonnenschein, 1982) administration in hamsters and rats. However, the results of these studies are equivocal and although the best route of administration appears to be intraperitoneal (Edwards and Germain, 1980b), a suitable dose rate able to induce intraoral neoplasms but compatible with survival of the animals has not been established. Our aim now was to determine the efficacy of intraperitoneal NMU for the induction of preneoplastic and neoplastic lesions in the intermolar platal mucosa and adjacaent gingiva in male hamsters because these areas appear to be some of the most susceptible to intraperitoneal NMU (Edwards and Get-main, 1980b).
Most experimental investigations into oral cancer and precancer use a topical carcinogen, with dimethylbenzanthracene (DMBA) application to the hamster cheek pouch being the most common (Eveson, 1981), although 4-nitroquinoline-l-oxide (4NQO) is also used in rats (Svensson and Heyden, 1982; Steidler, Rich and Reade, 1985; Rich and Reade, 1988), mice (Steidler and Reade, 1984, 1986) and hamsters (Eveson and MacDonald, 1977). However, some of the changes, especially the earlier ones, may be due, in part at least, either to the local irritant effect of the carcinogen (Eveson and MacDonald, 1978), to changes caused by the vehicle (White and AlAzzawi, 1983) or to a combination of these factors and thus are not part of the preneoplastic and neoplastic changes as such. In addition the hamster cheek pouch as a model of ‘intraoral’ carcinogenesis is criticized on a number of grounds (Eveson, 1981). The use of a systemically administered carcinogen that effects other areas would overcome these shortcomings. That N-methyl-N-nitrosourea (NMU) is an effective carcingoen for a number of anatomical sites following systemic and local administration is documented, although the number of studies is limited. Abbreviations: DMBA, methyl-N-nitrosourea;
mucosa,
MATERIALS
AND METHODS
Fifty 4-6-week-old random-bred male golden hamsters (Misocricetus auratus) were injected intraperitoneally with a weight-related dose (12.5 mg/kg body weight) of a freshly prepared O.O6rnol/l solution of
N-
519
520
M.D.
MCMILLAN
NMU in 0.1 M citrate buffer (pH 6.2) three times a week for 4 weeks. Thirty age- and sex-matched untreated control animals were used. All animals were carefully monitored and regularly weighed throughout the entire experimental period. Any animals that showed undue weight loss or appeared unwell were killed by an overdose of intraperitoneal pentobarbitone sodium followed by cervical dislocation. At 10, 16 and 22 weeks after the first injection, 11 anaesthetized animals from the experimental group and 10 from the control group had their maxillae removed. Seven animals from each time period were used in the present study. The remainder formed part of an ultrastructural study. The intermolar palatal mucosa and submucosa was surgically removed from five maxillae from each group at each time period and placed in 10% neutral-buffered formalin. The other maxillae, which contained the palatal gingiva as well as the intermolar tissues, were fixed whole. After fixation the maxillae were demineralized in a solution of formalin and formic acid. The separated mucosal specimens were then divided in half along the midline of the hard palate. The maxillae were divided in half transversely through the centre of the left and right second molar teeth. All specimens were processed routinely for light microscopy and blocked on their divided surfaces. Serial sections were cut with four sections being stained with haematoxylin and eosin every 20pm for the maxillary and every 10 pm for the mucosal specimens until the whole specimen was sectioned. The exophytic, dysplastic and invasive lesions in sections from NMU-treated animals were counted and the severity of each dysplastic lesion was scored. Care was taken not to record a lesion more than once. The score was based on how many of the following histological changes were present: loss of basal-cell polarity; multiplication of the basal-cell layer; increased nuclear-cytoplasmic ratio; drop-shaped rete ridges; irregular stratification; increased number, suprabasal or abnormal mitotic figures; cellular pleomorphism, nuclear hyperchromatism; dilated intercellular spaces; abnormal cell keratinization either singly or in groups. A number of histological levels were examined for most lesions. The score for a lesion was the sum of the different changes seen in all the sections in which it was found. For example, if in the first section there were three changes and if in the second section there were the same three changes the lesion was scored as having three changes (i.e. +). However, if two of the three changes in the second level were different to those in the first level the lesion was scored as having five changes (i.e. + + ): + 3 or 4; + + 5 or 6; + + + 7 or 8; + + + + 9 or more. No change in a given lesion was scored more than once. Limited autopsies were performed on the hamsters. Specimens for light microscopy were removed from any gross pathology and from the lungs, liver, spleen, pancreas, kidney, heart, forestomach, gastric stomach, duodenum and ileum. RESULTS
Control animals: No systemic pathology control animals.
was found
in any of the
e/ al.
The structure of the intermolar palatal mucosa and submucosa in all control specimens was similar. An orthokeratinized stratified squamous epithelium some 620 nucleated cells thick was present throughout except for a few small areas that were parakeratinized. Over the midline of the anterior two-thirds and all of the posterior one-third of the palate the epithelum was thinner with no or minimal rete ridges (Fig. 1). Rete ridges were more pronounced in the thicker epithelium elsewhere. The thickness of the stratum corneum matched that of the epithelium as a whole and usually had a smooth oral surface, although over the minor salivary glands it had a number of surface projections (Fig. 2). There were scattered mitotic figures in the basal cell layer and in the cells of the basally situated stratum spinosum. Most of the lamina propria of the intermolar palatal mucosa was a relatively dense fibrous connective tissue. That making up the rugae usually had an almost myxomatous appearance.
0.4mm Fig. 1. Control. Anterior palatal mucosa with thinner stratified squamous epithelium (E) in the midline (M). Epithelial rete ridges (arrows); regular stratum corneum (C); gingival epithelium (G); molar tooth root (R); periodontal ligament (P); lamina propria (L); greater palatine neurovascular bundle (N); palatal bone (Z). Cross-section. Haematoxylin and eosin. x40
Chemical
carcinogen&
in hamster palate
521
granulation-tissue formation bone loss, and a moderate to heavy infiltrate of neutrophilic leucocytes. NMU-Treated
animals
Initially all NMU-treated animals lost weight but by 6 or 7 weeks after the first injection most had regained this and they showed a steady weight gain for the rest of the experimental period. However, nine animals had to be killed in the first 10 weeks because of undue weight loss, with eight animals being killed later because of a reversal of their weight gain. With the exception of five of these latter animals, three of which had ischaemic necrosis of segments of the ileum and two of which had abdominal carcinomas derived from the skin, autopsy did not reveal the cause of the undue weight loss. The only constant findings in the other NMU-treated animals were inflammation of the lungs with thickening of the alveolar walls and areas of epithelial hyperplasia, hyperkeratosis and dysplasia, along with multiple papillomas of the forestomach. The severity of these changes increased with increasing age. All 33
Fig. 2. Control. Posterior palatal mucosa with epithelial surface projections (S); orthokeratinized stratified squamous epithelium (E). Minor salivary glands (A) in the lamina propria (L). Longitudinal section. Haematoxylin and eosin. x 160
The submucosa, a less fibrous, more cellular and vascular variant of the lamina propria, was minimal in amount except in the area of the greater palatine neurovascular bundle and in the posterior third of the specimens. Here it contained varying amounts of adipose tissue anteriorly and minor salivary glands posteriorly. The structure of the palatal gingival and crevicular tissues was similar in nearly all the control specimens and resembled that of the adjacent palatal mucosa, although the crevicular epithelium had only an occasional rete ridge and a thinner keratin layer. The junctional epithelium, which often extended for a short distance on to the cementum, overlay much of the crevicular epithelium and separated it from the tooth. In some sections from all specimens the intercellular spaces of the crevicular epithelium and the adjacent connective tissue contained a small number of neutrophilic leucocytes. In two specimens there was an area where foreign material was impacted in the gingiva and underlying periodontal ligament. This was associated with epithelial proliferation,
F?g. 3. Ten weeks NMU. Thickened epithelium with plug of nnainly parakeratinized cells (K). Individual cell keratinizaion (arrowheads); dilated intercellular spaces (arrows); ascular lamina propria (L). Longitudinal section. Haematoxylin and eosin. x 160
522
M. D. MCMILLAN et al.
remaining animals had lesions in the palatal gingiva and intermolar mucosa when killed. Grossly, the mucosa of the hard palate in all IO-week NMU animals was more irregular than in the controls, although the form of the rugae was still evident. This irregularity was caused by variable numbers of small surface projections, depressions, and the occasional distinct papilloma, and by an increased amount of epithelial desquamation. There was no regular pattern to the distribution of these changes. Similar changes were found in all the 16and 22-week animals but with increasing age they were more pronounced, involved a greater proportion of the mucosa and the papillomas were more frequently seen. At 22-weeks the mucosal surface was often so irregular than the normal rugal pattern was lost. Although the surface of the palatal gingiva in all animals was more irregular than in the controls, this was due to the presence of depressions and epithelial desquamation. Surface projections were found only infrequently.
0.1 mm Fig. 5. Ten weeks NMU. Irregular surface epithelium (E,) covering an inflamed Iamina propria (L) that contains deeper epithelial islands (E,). Dilated thin-walled vessels (W), one containing epithelium (arrowhead). Nautrophilic leucocytes (arrows). Longitudinal section. Haematoxylin and eosin. x 160
Fig. 4. Ten weeks NMU. Inflamed lamina propria (L) covered by an irregular, thickened epithelium (E). Indistinct epithelialeonnective tissue junction (arrowheads); neutrophihc Ieucocytes (arrows). Longitudinal section. Haematoxylin and eosin. x 160
Histological changes seen in animals treated with NMU for 10, 16 and 22 weeks are illustrated in Figs 3-5, 6-8 and 9-14, respectively. All animals showed varying numbers of exophytic (Figs 6, 7 and 14) dysplastic (Figs 46 and 8) and invasive (Figs 5,9-14) lesions. These were derived from either the gingival or intermolar epithelium (Table 1). The incidence and severity of the dysplastic lesions for each time period are listed in Table 2. As well as differences in the number of exophytic, dysplastic and invasive lesions present at each time period, there were also differences in the depth of invasion for invading lesions, and in the amount of surface epithelium involved. Both of these features increased with increasing age, with the latter feature being confined to lesions derived from the intermolar mucosa. Also, in the lo- and 16-week specimens, invasive lesions derived from the crevicular and junctional epithelium had penetrated to a greater depth than those derived from the palatal gingival and intermolar mucosa. In the
523
Chemical carcinogenesis in hamster palate
that at times had a basaloid appearance (Fig. 11). At 10 weeks the majority of the invading epithelium was confined to the more laterally situated lamina propria and submucosa along with the more superficial periodontal tissues. Some of the invading epithelium had involved the alveolar and palatal bone, the greater palatine neurovascular bundle and the palatal marrow spaces in the 16-week specimens. Also, in some instances involvement of the periodontal ligament extended to the root apices of the molar teeth. By 22 weeks the involvement of these deeper structures was more extensive (Figs 9-l 1, 13 and 14). In the more posterior areas of the specimens, although some of the invading tissue was close to the salivary glands, the glands themselves were normal. The associated connective tissue in all NMU specimens were usually more cellular and vascular than in control specimens. It contained numerous plump fibroblasts and varying numbers of inflammatory cells (Figs 4, 5 and 12). Dilated vessels were often in close association with the surface and invading
tion
0.2mm Fig. 6. Sixteen weeks NMU. Area of markedly thickened irregular surface epithelium (E) that has numerous reteridges extending into the underlying lamina propria. Surface projections (arrows). Longitudinal section. Haematoxylin and eosin. x 80
22-week specimens this difference was not as marked. The invasive lesions had a similar range of histological appearances, irrespective of either their origin or whether they were in the lo-, 16- or 22-week specimens. There were cords, strands, islands and single cells or small groups of epithelium within the connective tissue (Figs 5, 6 and 9-14). Although a number of the epithelial islands appeared to have undergone cystic change (Figs 9 and 14), examination of a number of histological levels showed that they were, in fact, deep extensions from the surface epithelium that had a patent lumen and formed duct-like structures (Figs 9, 10 and 14). In the 16- and 22-week specimens some of the invading epithelium appeared at least partly degenerate. The intercellular spaces of the degenerate epithelium usually contained neutrophilic leucocytes (Fig. 9). The degree of differentiation of the invading epithelium varied from well-differentiated stratified squamous epithelium showing regular otthokeratinization, to less-differentiated epithelium lacking any evidence of keratiniza-
0.4mm Fig. 7. Sixteen weeks NMU. Area of more exophytic surface epithelium (E) having a very irregular surface. Gingival epithelium (G); fibrous lamina propria (L); greater palatine neurovascular bundle (N); palatal bone (Z). Cross-section. Haematoxylin and eosin. x 55
524
M. D. MCMILLAN rr al.
extended down to the basal cell layer. Although the epithehum below some of the more superficial keratin plugs was normal, in many instances it was dysplastic (Figs 3 and 8). A striking feature of all NMU specimens was the individual cell keratinization, which was found at all levels in the epithelium (Figs 3 and 8). At times only one or a small group of keratinized cells were present in the stratum spinosum, while elsewhere vertical columns of these cells extended throughout part or all of the epithelium. Some of these columns were only one cell wide while others were more extensive. Cells showing individual keratinization were round in outline, had very eosinophilic cytoplasm, were surrounded by a much wider intercellular space than elsewhere in the epithehum and often appeared to lack contact with adjacent cells. At times the dilated intercelhtlar space extended from the surface of the stratum corneum down through the basal cell layer to the epithelialLconnective tissue junction. Many of the nuclei of the keratinized cells appeared similar to
;-2.
,I
0.1 mm Fig.
8. Sixteen
weeks
NMLJ.
Very irregular
surface
epi-
thelium showing dilated intercellular spaces (arrow). vidual stratum
indicell keratinization (arrowheads) and thickened corneum (C). Longitudinal section. Haematoxylin and eosin. x 160
epithehum (Figs 5 and 12) and at times appeared to contain epithehal cells (Fig 5). The surface epithelium in all NMU specimens was often thickened and showed acanthosis and usually hyperorthokeratosis, although there were also areas of hyperparakeratosis. The amount of epithelium involved and its thickness increased with increasing age of the specimens and by 22 weeks the connective tissue component of the rete ridges and the lamina propria was at times replaced by hyperplastic epithehum. Although some of this acanthotic epithehum was dysplastic, some was not. The exposed oral surface of the intermolar mucosa and palatal gingiva in all NMU specimens WdS irregular and this irregularity increased with increasing age of the specimens. In part this irregularity was caused by the exophytic lesions (Table I) (Figs 6, 7 and 14) and in part by the oral openings of the duct-like structures (Fig. 14) irregular desquamation and the presence of keratin plugs (Figs 3 and 8). The number and size of the keratin plugs also increased with increasing age of the specimens and at times they
Fig. 9. Twenty-two weeks NMU. Extensive epithelial invasion (arrows) involving the palatal bone (Z) and periodontal ligament (P). Degenerate epithelium (arrowheads); cystic change (D); greater palatine neurovascular bundle (N); molar tooth root (R). Cross-section. Haematoxylin and eosin. x 60
Chemical
carcinogenesis
Fig. IO. Twenty-two weeks NMU. Lateral area with markedly thickened surface epithelium (E) and apparent keratin pearls (K). Invading gingival epithelium with lumen (arrows); area of bone resorption (arrowhead); periodontal ligament (P). Cross-section. Haematoxylin and eosin. x 45
in hamster
palate
525
Fig. 12. Twenty-two weeks NMU. Invading epithelium (arrows) deep in lamina propria. Blood vessels (V). Longi-
tudinal section. Haematoxylin and eosin. x 250
NMU in age- and sex-matched hamsters to induce preneoplastic and neoplastic change in the palatal those of the surrounding epithelium but others were gingival and intermolar epithelium. It also provides a either pyknotic or contained aggregations of chrodetailed description of the histological changes inmatin. duced over an extended period and confirms that The epithelium in NMU specimens did, at times, intraperitoneal NMU is not associated with the high also show some of the following: acantholysis; columincidence of carcinoma of the forestomach, and connar basal cells with very eosinophilic cytoplasm; sequent high death rate, that the intragastric route of aggregations of sebaceous-like cells in the stratum administration is (Edwards and Germain, 1980a, b). spinosum. The reasons for the deterioration in health of 12 of the 17 hamsters that had to be killed prematurely were not from the This could DISCUSSION Fig. apparent 12. Twenty-two weeksautopsies. NMU. Invading epithelium Fig. IO. Twenty-two weeks NMU. Lateral area with only be attributed to ill-defined chronic effects (V). of Longi(arrows) deep in lamina propria. Blood vessels markedly thickened surface epithelium (E) and apparent The keratin structure theInvading intermolar mucosa, NMU cytotoxicity (Edwards and Get-main, 1980b), tudinal section. Haematoxylin and eosin. x 250 pearls of(K). gingivalpalatal epithelium with lumen submucosa and area palatal gingiva in the (arrowhead); control animals some of the earlier problems could have (arrows); of bone resorption periodontalalthough was similar of the rats (McMillan, ligamentto that (P). Cross-section. Haematoxylin 1979). and The eosin. xbeen 45 caused immunosuppression NMU byin the age-transient and sex-matched hamsters that to induce reason for the localized areas of parakeratinized NMU preneoplastic induces (Herrold, 1966; Leaver, Swann and palatal and neoplastic change in the epithelium but they could be thebut result of were Magee, gingival 1969; Edwards and Germain, those isofunknown the surrounding epithelium others and intermolar epithelium.1980b). It also provides a previouseither trauma. examination of crossand histological longitudinal changes secpyknotic or contained aggregations of chro- The detailed description of the inThe matin. present study is the first to use a standardized tions atduced regularover intervals through period the palatal an extended and gingival confirms that intraperitoneal dose of in the The epithelium NMUcomplete specimenscarcingoen did, at times, intraperitoneal NMU is not associated with the high also show some of the following: acantholysis; columincidence of carcinoma of the forestomach, and connar basal cells with very eosinophilic cytoplasm; sequent high death rate, that the intragastric route of aggregations of sebaceous-like cells in the stratum administration is (Edwards and Germain, 1980a, b). spinosum. The reasons for the deterioration in health of 12 of the 17 hamsters that had to be killed prematurely were not apparent from the autopsies. This could DISCUSSION only be attributed to ill-defined chronic effects of The structure of the intermolar palatal mucosa, NMU cytotoxicity (Edwards and Get-main, 1980b), submucosa and palatal gingiva in the control animals although some of the earlier problems could have was similar to that of the rats (McMillan, 1979). The been caused by the transient immunosuppression that reason for the localized areas of parakeratinized NMU induces (Herrold, 1966; Leaver, Swann and epithelium is unknown but they could be the result of Magee, 1969; Edwards and Germain, 1980b). previous trauma. The examination of cross- and longitudinal secThe present study is the first to use a standardized tions at regular intervals through the palatal gingival intraperitoneal dose of the complete carcingoen
Fig. I I. Twenty-two weeks NMU. Epithelial deep within the palatal bone(Z). Cross-section. lin and eosin. x 150
invasion (E) Haematoxy-
Fig. I I. Twenty-two weeks NMU. Epithelial deep within the palatal bone(Z). Cross-section. lin and eosin. x 150
Fig. 13. Twenty-two weeks NMU. Invading epithelium (arrows) deep in lamina propria associated with the greater palatine nerve (N) and blood vessel (V); epithelial cyst (D). Longitudinal section. Haematoxylin and eosin. x 160
invasion (E) Haematoxy-
Fig. 13. Twenty-two weeks NMU. Invading epithelium (arrows) deep in lamina propria associated with the greater palatine nerve (N) and blood vessel (V); epithelial cyst (D). Longitudinal section. Haematoxylin and eosin. x 160
M. D. MCMILLAN
526
et al.
Table 1. Number of exophytic (E), dysplastic (D) and invasive (I) lesions in NMU-treated animals Palatal Gingiva Time (weeks) 10 16 22
Number of animals/ (sections) 2 (368) 2 (392) 2 (376)
E D
Intermolar Mucosa
I
2 2 5 2 4 7 - - 6
Number of animals/ (sections) I (388) 7 (395) 7 (406)
E D
I
21 21 9 45 33 19 19 20 27
Exophytic lesions were either conical projections of the stratum corneum, larger papillae-like structures that had a core of nucleated epithelial cells or papillomas with a central connective tissue core. Invasive lesions showed frank invasion of the connective tissue by the epithelium. in the present study in that what in some sections from NMU-treated hamsters appeared to be isolated islands and cords of invading epithelium were in fact in continuity with the surface epithelium in sections taken at a different level. This also enabled the origin of the invading epithelium to be determined. Similarly, what in some sections appeared as deeply situated cysts, which have also been described in oral mucosa following NMU administration (Suzuki, 1976; Edwards and Germain, 1980a; Kohgo et al., 1990) were usually found to be duct-like structures that opened into the mouth. Cysts or duct-like structures are not a feature of human oral squamous-cell carcinoma, or of experimental oral carcinoma in animals induced by either 4-NQO (Svensson and Heyden, 1982; Philipsen and Fisker, 1983; Rich and Reade, 1988) or DMBA (Salley, 1957; Mori et al., 1962; Skhlar et al., 1980; White, Gohari and Smith, 1981). Individual cell keratinization is found in dysplasia and carcinoma in human oral epithelium. However, the extensive areas of individual cell keratinization that often resembled the focal acantholytic dyskeratosis of Darier’s disease (Philipsen and Fisker, 1983) seen in this and other studies employing NMU (Suzuki, 1976; Edwards, 1978; Edwards and Germain, 1980a, b; Chau and Edwards, 1984) are not. Similar changes are found in 4-NQO; (Svensson and Heyden, 1982; Philipsen and Fisker, 1983) and occasionally DMBA- (Eveson and MacDonald, 1977. 1981) treated oral mucosa. None of these studies provided evidence that these changes were related to the subsequent development of malignancy. Initially, multiple exophytic or invasive lesions involved the palatal gingival and intermolar tissues. Before 22 weeks, however, lesions associated with the former tissues were mostly invasive, and were derived from the junctional and crevicular epithelium, while demonstrated
Fig. 14. Twenty-two weeks NMU. Exophytic surface projection (E,); keratin-filled lumen (arrowheads); deeper invading epithelium (EJ; cystic change (D); greater palatine neurovascular bundle (N); individual cell keratinization (arrows). Longitudinal section. Haematoxylin and eosin. x 50 and intermolar tissues, as undertaken in this study, is essential when a mucosa of variable structure is used in investigations into experimental preneoplasia and neoplasia. This is because normal structural variations may be mistaken for pathological change even in an apparently homogeneous mucosa (McMillan and Kerr, 1990). In addition, the presence and true organization of the pathological changes may not be apparent from either macroscopic examination or a small number of histological sections. This is amply
Table 2. Incidence and severity of change in the dysplastic lesions in NMU-treated Palatal Gingiva ____ Time (weeks) 10 16 22
Number of animals/ (sections) 2 (368) 2 (392) 2 (376)
+
++
+++
-
2
--
1
‘)
I
Intermolar Mucosa
~~~~
++++
animals (see p. 520)
Number of animals/ (sections)
+
7 (388) 7 (395) 7 (406)
15 20 8
++
+++
++++
4 8 4
2 3 8
2
Chemical carcinogenesis in hamster for the palate exophytic lesions predominated. Similar findings for exophytic and invasive lesions following NMU administration are described for the gingiva (Suzuki, 1976) and hard palate (Edwards and Germain, 1980a, b) but Kohgo et al. (1990) describe only invasive gingival carcinomas. A similar disparity in the latent period before invasive lesions predominated between the gingiva and hard palate, as found in the present study, has also been reported (Edwards, 1978; Edwards and Germain, 1980a) as has the fact that the initial changes in the gingiva involve the junctional and crevicular epithelium (Suzuki, 1976; Kohgo et al., 1990) following intragastric NMU. It is known that stimulated proliferation may play a part in the initiation, promotion and propagation of carcinogenesis (Craddock, 1976; Kunze and Pacha, 1985; Kunze et al., 1989) and that continuous, non-specific irritation may serve as a locus for experimental tumour formation in skin (Argyris, 1980), the gastrointestinal tract (Pozharisski, 1975; Barthold and Beck, 1980) and buccal mucosa (Konstantinidis et al., 1982). Although it has been suggested that impaction of hair, bedding and food in the gingival crevice of NMU-treated animals may be a cocarcinogenic factor (Edwards, 1978), this was only rarely observed in the present study, which confirms the observtions of Suzuki (1976) that this form of local irritation is not a factor. Much of the invading epithelium was in the form of distinct cords, islands and cysts or duct-like structures that were composed of well-differentiated, malignant squamous epithelium. This was irrespective of whether they were derived from junctional, crevicular, gingival or palatal epithelium, which confirms the findings of Herrold (1968); Suzuki (1976); Edwards and Germain (1980a, b) and Kohgo et al. (1990). However, by 22 weeks some of the larger epithelial islands and cords were less well differentiated and occasionally basaloid in appearance. This again was irrespective of their origin. This latter appearance is also noted by Edwards and Germain (1980a, b). Kohgo et al. (1990), in their study of NMU-induced changes in the gingiva and endentulous alveolar ridge, describe invasion by nests of relatively poorly differentiated epithelium. Similar nests and even apparently individual epithelial cells were present not only in the connective tissue of the hard palate, gingiva and periodontal ligament in the present study, but also within the lumen on thin-walled vessels. These features are not described by Herrold (1968), Suzuki (1976) or Edwards and Germain (1980a, b). Whether the vessels containing the epithelial cells were blood or lymphatic could not be determined but none contained red blood cells. The presence of epithelial cells in vessels in experimental oral carcinogenesis is only rarely reported (Salley and Kreshover, 1959; Herrold, 1968; Eveson and MacDonald, 1981; Shingaki, Otake and Nakajima, 1987) as is regional lymph-node involvement (Salley, 1954; Rwomushana, Polliack and Levij, 1970; Eveson and MacDonald, 1981; Safour et al., 1984; Shingaki et al., 1987). The potential of NMU-induced oral carcinomas to spread to distant sites requires further investigation. Although some of the invading epithelium that was associated with an inflammatory response appeared
527
palate
to be becoming degenerate, this was not common. Such changes could indicate that this epithelium was being destroyed. However, whether with time such a process would involve all of the invading epithelium is problematical. However, inflammation was also associated with the changed but viable surface epithelium in NMU-treated animals and with the viable invading epithelium. Whether the inflammatory response was directed towards the changed epithelium or not is unknown. What is known is that at 22-weeks viable invading epithelium was present deep in the periodontal ligament and maxillary bone and marrow and was associated with the intrabony part of the greater palatine neurovascular bundle. The increased vascularity often associated with the altered surface and invading epithelium is probably indicative of the angiogenic response characteristics of carcinogenesis (Folkman, 1978, 1986; Ferguson and Smillie, 1979; White and Al-Azzawi, 1983; Polverini, Shimizu and Solt, 1988), although part at least may have been related to the inflammatory response. We have clearly shown that our experimental regimen resulted in preneoplastic and neoplastic change in the stratified squamous epithelium of the maxillary intermolar mucosa and adjacent gingiva in hamsters. Although changes such as acantholytic dyskeratosis, cysts and duct-like structures are not typical of human oral squamous-cell carcinoma, our protocol provides a suitable model for the investigation of oral carcinogenesis. It also allows investigation of invasion of bone by malignant oral epithelium, a feature that is not seen in experimental
hamster cheek-pouch carcinogenesis. The use of an experimental system employing a systemic carcinogen rather than one of the two commonly used topical ones (i.e. DMBA and 4-NQO) is important because it eliminates any local irritant effect of the carcinogen or vehicle. In addition, it allows the comparison of changes induced by three rather than two oral carcinogens. Such comparison is necessary to determine that any changes induced are indicative of the preneoplastic and neoplastic process and are not just peculiar to a particular carcinogen. In addition, the NMU model described is simple, does not require extended carcinogen treatment, and may also provide a useful tool to investigate aberrations in keratinization. AcknowledRemenrs-This study was in part undertaken while one of the authors (M. D. McMillan) held a Commonwealth Medical Fellowship in the Department of Oral Pathology, University of Sheffield, while V. Ramirez was a graduate student in this department. D. Thompson, H. Walker and D. Teale of Sheffield and A. Wiles and J. Thomson of Dunedin are thanked for their excellent technical assistance. Mrs P. K. Cosgriff is thanked for typing the manuscript.
REFERENCES
Argyris T. S. (1980) Tumor promotion by abrasion induced epidermal hyperplasia in the skin of mice. J. invest. Derm. 75, 36&362. Barthold S. W. and Beck D. (1980) Modification or early dimethylhydrazine carcinogenesis by colonic mucosal hyperplasia. Cancer Res. 40, 44514455.
528
M. D. MCMILLAN et al.
Chau K. K. and Edwards M. B. (1984) Topical carcinogenesis by N-methyl-N-nitrosourea in Syrian hamster cheekpouch and oral mucosa. Archs oral Biol. 29, 185-190. Craddock V. M. (1976) Cell proliferation and experimental liver cancer. In Liver Cancer (Eds Cameron H. M. and Linsell C. A.), pp. 153-201. Elsevier, Amsterdam. Edwards M. B. (1978) Some effects of N-methylnitrosourea on the oral and odontogenic tissues of the Syrian golden hamster. Archs oral Biol. 23, 515-524. Edwards M. B. (1980) Effects of N-methylnitrosourea on oral and dental tissues in the inbred Lewis rat. Archs oral Biol. 25, 5965. Edwards M. B. and Germain J. P. (1980a) Oral carcinogenesis by intragastric N-methyl-N-nitrosourea in the Syrian hamster. Archs oral Biol. 25, 623429. Edwards M. B. and Germain J. P. (1980b) Oral carcinogenesis by intraperitoneal N-methyl-N-nitrosourea in the Syrian hamster. Archs oral Biol. 25, 631-633. Eveson J. W. (1981) Animal models of intra-oral chemical carcinogenesis: a review. J. oral Parh. 10, 129-146. Eveson J. W. and MacDonald D. G. (1977) Effects of the water-soluble carcinogen 4-nitroquinoline N-oxide on hamster lingual mucosa. Oral Surg. 44, 60&605. Eveson J. W. and MacDonald D. G. (1978) Quantitative histological changes during early experimental carcinogenesis in the hamster cheek pouch. Br. J. Derm. 98, 639644. Eveson J. W. and MacDonald D. G. (1981) Hamster tongue carcinogenesis. I. Characteristics of the experimental model. J. oral Path. 10, 322-331. Ferguson J. W. and Smillie A. C. (1979) Vascularization of premalignant lesions in carcinogen-treated hamster cheek pouch. J. natn. Cancer Inst. 63, 138331392. Folkman J. (1978) Tumour angiogenesis and tumour immunity. In immunological Aspects of Cancer (Ed. Castro J. E.), pp. 267-282. University Park Press, Baltimore, MD. Folkman J. (I 986) How is blood vessel growth regulated in normal and neoplastic tissue? Cancer Res. 46, 467473. Herrold K. M. (1966) Carcinogenic effect of N-methyl-Nnitrosourea administered subcutaneously to Syrian hamsters. J. path. Bact. 92, 35-41. Herrold K.. M. (1968) Odontogenic tumors and epidermoid carcinomas of the oral cavitv. Oral Surp. 25, 262-272. Ketkar M., Reznik G., Haas H., HilfrichJ. and Mohr U. (1977) Tumors of the heart and stomach induced in European hamsters by intravenous administration of N-methyl-N-nitrosourea. J. natn. Cancer Inst. 58, 1695-1697. Kohgo T., Uy G. H., Iizuka T., Shindoh M. and Amemiya A. (1990) Histopathologic study of squamous cell carcinoma of the gingiva in edentulous and dentulous jaws of hamsters treated with N-methylnitrosourea, J. oral Path. 19, 202-207. Konstantinidis A., Smulow J. B. and Sonnenschein C. (1982) Tumorigenesis at a predetermined oral site after one injection of N-nitroso-N-methylurea. Science 211, 1235-1237. Kunze E. and Pacha H. (1985) Modifikation der urothelkarzinogenese durch proliferationsstimulation. In Experimentelle Urologic (Eds Harzmann R., Jacobi G. H. and WeiBach L.). pp. 419429. Springer, Heidelberg. Kunze E., Graewe T., Scherber S., Weber J. and Gellhar P. (1989) Cell cycle dependence of N-methyl-N-nitrosoureainduced tumour development in the proliferating, partially resected rat urinary bladder. Br. J. exp. Path. 70, 1255142. Leaver D. D., Swann P. F. and Magee P. N. (1969) The induction of tumours in the rat by a single oral dose of N-nitrosomethylurea. Br. J. Cancer 23, 1777187. McMillan M. D. (1979) The surface structure of the completely and incompletely orthokeratinized oral epithelium in the rat: a light, scanning and transmission electron microscope study. Am. J. Anat. 156, 337-351.
McMillan M. D. and Kerr M. A. (1990) A light and scanning electron microscope study of epithelial thickenings and rete ridges in the adult hamster cheek pouch. Archs oral Biol. 35, 235-240. Milievskaja I. L. and Kiseleva N. S. (1976) Comparative study of the carcinogenic activities of nas and some chemical carcinogens when introduced into the buccal pouch of the Syrian hamster. Bull. Wld Hlth Organ 54, 607414. Mori M., Miyaji T., Murata I. and Nagasuna H. (1962) Histochemical observations on enzymatic processes of experimental carcinogenesis in hamster cheek pouch. Cancer Res. 22, 1323-1326. Philipsen H. P. and Fisker A. V. (1983) Focal acantholytic dyskeratosis in experimental oral carcinogenesis in rats. J. oral Path. 12, 3(r36. Polverini P. J., Shimizu K. and Solt D. B. (1988) Control of angiogenic activity in carcinogen-initiated and neoplastic hamster pouch keratinocytes and their hybrid cells. J. oral Path. 17, 522-527. Pozharisski K. M. (1975) The significance of nonspecific injury for colon carcinogenesis in rats. Cancer Res. 35, 38243830. Reibel J. (1982)Carcinogenesis in forestomach and changes in oral mucosa of rats induced by intragastric N-methylnitrosourea. &and. J. dent. Res. 90, 382-389. Rich A. M. and Reade P. C. (1988) Histomorphometric analysis of epithelial changes in chemically induced oral mucosal carcinogenesis in rats. J. oral Path. 17, 528-533. Rwomushana J. W., Polliack A. and Levij I. S. (1970) Cervical lymph node metastasis of hamster cheek pouch carcinoma induced with DMBA. J. dent. Res. 49. 184. Safour 1. M., Wood N. K., Tsiklakis K., Doemling D. B. and Joseph G. (1984) Incisional biopsy and seeding in hamster cheek pouch carcinoma. J. denf. Res. 63, 1 I161 120. Salley J. J. (1954) Experimental carcinogenesis in the cheek oouch of the Svrian hamster. J. dent. Res. 33. 253-261. Sailey J. J. (1957)Histologic changes in the hamster cheek pouch during early hydrocarbon carcinogenesis. J. dent. Res. 36, 48-55. Salley J. J. and Kreshover S. J. (1959) The effect of topical application of carcinogens on the palatal mucosa of the hamster. Oral Surg. 12, 501-508. Shingaki S., Otake K. and Nakajima T. (1987) Regional lymph node metastasis in carcinoma of the hamster tongue. Oral Surg. 64, 190-196. Shklar G., Schwartz J., Grau B. A., Trickier D. P. and Wallace K. D. (1980) Inhibition of hamster buccal pouch carcinogenesis by 13-cis-retinoic acid. Oral Surg. 50, 45552. Steidler N. E. and Reade P. C. (1984) Experimental induction of oral squamous cell carcinomas in mice with 4-nitroauinolinel-oxide. Oral Surp. 57, 524-53 1. Steidler A. E. and Reade P. C. T1986) Initiation and promotion of experimental oral mucosal carcinogenesis in mice. J. oral Path. 15, 4347. Steidler N. E., Rich A. M. and Reade P. C. (1985) Experimental induction of preneoplastic and neoplastic changes in the lineual mucosa of rats. J. Biol. buccale 13,339-346. Suzuki M. (1976) Histological studies of early changes of the hamster gingival epithelium by N-nitrosomethylurea. Bull. Tokyo med. dent. Univ. 23, 115-134. Svensson S. and Heyden G. (1982) Experimental induction of irreversible precancerous changes in the palatal epithelium of the rat. Int. J. oral. Sura. 11, 52258. White F. H. and Al-Azzawi B. (1983)‘Vascularity in experimental oral neoplasia: a stereological approach. Acta Sterrol. 2/l, 119-126. White F. H., Gohari K. and Smith C. J. (1981) Histological and ultrastructural morphology of 7, 12, dimethylbenz(a)-anthracene carcinogenesis in hamster cheek pouch epithelium. Diagn. Histopath. 4, 307-333