STUDIES ON EXPERIMENTAL CARCINOMA
OF THE UTERINE CERVIX
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School Of the Ohio State University
By
DANTE GIOVANNI SCARPEL.LI, B.S., M.S., M.D.
~ t ~ *x“ v vc ■■ ” “
The Ohio State University I960
pproved bys
Adviser Department of Pathology PREFACE
"We are like dwarfs seated on the shoulders of giantsj we see more things than the ancients and things more distant, but this is due neither to the sharpness of our own sight, nor to the greatness of our own stature, but because we are raised and borne aloft on that giant mass."
Bernard of Chartres, Twelfth Century
Rudolf Virchow in his classic treatise, Cellular Pathologie, published in l8£l?, presented and formulated a new concept of disease based on the premise that the elementary unit affected is the cell.
He emphasized this in the following statement, "All diseases are in the last analysis reducible to disturbances, either active or passive,
of large or small groups of living units, whose functional capacity is thus dependent on physical and chemical changes of their contents."
It is surprising to note how closely this fits present day
concepts in the study of disease. Pathology is no longer a discipline limited to pure anatomic description of disease processes. It has been amply demonstrated that in the investigation of any problem in pathology, further knowledge is dependent upon the association of physics and chemistry with morphology. Today, it is not only possible iii to describe pathologic alteration in morphologic terms but in bio chemical and biophysical ones as well.
Such an experimental approach to the study of disease necessitates the simultaneous employment of diverse morphologic, cytochemical and cytophysical techniques to elucidate the various cellular alterations that develop during the pathogenesis of a lesion. Since neither structure nor function is static, knowledge of the development of a disease process can best be gained by understand ing the relation of one to another. The present study has been approached in this manner, utilizing electron microscopy and cyto chemistry in addition to the techniques routinely employed in morpho logic pathology.
During the course of these investigations this student has been privileged to be associated with and under the guidance of
Professor Emmerich von Haam. His interest, considered suggestions and intellectual stimulation have made the researches reported herein both enjoyable and profitable. Without his assistance this work would not have been possible.
D. G. Scarpelli TABLE OF CONTENTS
Page
Introduction 1
Pathogenesis: A Correlative Cytologic AndHistologic Study 17
Pathologic Anatomy And Histology lj.2
Mitosis In Cervical Epithelium During Experimental Inflammation And Carcinogenesis 72
Quantitative Estimations Of The Deoxyribonucleic Acid Content Of Cervical Epithelial Cells During Estrus, Inflammation And Carcinogenesis 99
Cellular Ultrastructure Of Experimental Squamous Cell Carcinoma Of The Uterine Cervix 119
Cytochemistry Of Experimental Cervical Carcinoma 130
General Summary lii3
Bibliography
i V V
LIST OF FIGURES
Figure Page
1. 3,k Benzpyrene in acetone and bacteriologic wire-loop. . 9
2. Intravaginal painting procedure ...... 9
3. Necrotic squamous cells ...... 22
U. Acute cervicitis .... 22
3. Histiocytic giant cell . . 22
6. Cervical dysplasia - cytoplasmic granules ...... 23
7. Cervical dysplasia - cytoplasmic vacuoles ...... 23
8. Cervical dysplasia - micro-abscesses ...... 23
9. Precocious keratinization...... 23
10. Cervical dysplasia - precocious keratinization .... 27
11. Advanced dysplasia - atypical basal cells ...... 27
12. Advanced dysplasia - papillomatous hyperplasia . . . 27
13. Carcinoma in situ - vaginal s m e a r ...... 31
lU. Carcinoma in situ - intraepithelial nodules .... 31
13. Carcinoma in situ 31
16. Carcinoma in situ - cellular a t y p i a ...... 31
17. Vaginal smear - macronucleoli ...... 33
18. Vaginal smear - malignant fiber cells ...... 33
19. Vaginal smear - epithelial pearl ...... 33
20. Well differentiated epidermoid carcinoma ...... 33
21. Poorly'differentiated carcinoma ...... 33 vi
LIST OF FIGURES-Continued
Figure Page
22. Vaginal Smear - anaplastic carcinoma ...... 33
23. Mouse uterus ...... 1*3
2k. Squamo-columnar junction ...... 1*3
23. Pelvic lymph nodes ...... 1*3
26. Largest tumor induced ...... 1*9
27. Exophytic configuration of carcinoma , . . I4.9
28. Carcinoma high on lateral cervical wall ......
29. Extensive carcinoma in situ ...... 31
30. Large tumor of cervix ...... 3i
31. Massive tumor with necrosis ...... 3i
32. Lymph node metastasis ...... 33
33. Rectal invasion ...... 33
3h. Pulmonary metastasis ...... 33
33. Hydroureter and hydronephrosis ...... 33
36. Carcinoma in situ ...... 37
37. Carcinoma in situ - cytology ...... 37
38. Carcinoma in situ - papillomatous proliferation . . 37
39. Carcinoma in situ - pattern resembling human lesion . 37
ho. Carcinoma in situ - extensive keratinization 61 la. Carcinoma in situ - atrophic type ...... 61
ia. Early invasive carcinoma ...... 61 Vl l
LIST OF FIGURES-Continued
Figure Page
1+3 • Invasive carcinoma 6l
1+lj.. Microcarcinoma 61+
1+5. Well differentiated carcinoma - spinal type.... 61+
'1+6. Moderately well differentiated carcinoma - transitional t y p e ...... 61+
1+7• Poorly differentiated carcinoma - spindle type . . . 67
1+8. Muco-epidermoid c a r c i n o m a ...... 67
1+9. Lagging chromosomes...... 81
90. Asymmetrical d i v i s i o n ...... 81
5l. Tripolar metaphase ...... 81
92. Tripolar anaphase ...... 81
53. Tetrapolar metaphase ...... 81
91+. Hollow metaphase . 81
95'.. Micronucleus...... 81
56. Incomplete hollow metaphase ...... 81
57. Polar c h r o m o s o m e s ...... 81
58. Polar c h r o m o s o m e s ...... 81
59. Migration of polar chromosomes ...... 81
60. "Colchicine effect." ...... 81
61. Mitochondria at tumor growing edge ...... 125
62. Degenerating Mitochondria in quiescenttumor cells . 125
63. Degenerating mitochondria in tumor cells undergoing keratinization ...... 125 viii
LIST OF FIGURES-Continued
Figure Page
6I4. Microvilli ...... 127
65. Virus-like bodies ...... 127
66. Fibroblasts around a tumor nodule ...... 129
67. Fibroblasts away from tumor cells ...... 129
68. Intense basophilia of tumor buds at the growing e d g e ...... 137
69. DPN diaphorase activity in tumor cells . . . . 137
70. Glucose-6-phosphate dehydrogenase activity at the growing edge of t u m o r ...... 137
'71. Early invasive cervical carcinoma ...... 139
72. Glutamic dehydrogenase activity in the center of a keratin nodule ...... 139 ix
LIST OF TABLES
Table Page
1. Summary of Experimental Induction of Neoplasms of the Uterine Cervix and Vagina ...... 10
2. Experimental Induction of Carcinoma of the Uterine Cervix and Vagina in C^H Mice ...... 16
3. Cyto-Histologic Correlation 35
!|.. Sites of Neoplasia I4.6
5. Sites of Tumor Extension and Metastases ......
6. Metastases to Pelvic Lymph Nodes ...... 35
7. Summary of Experimental Carcinomas Induced .... 58
8 . Histologic Types of Carcinoma ...... 62
9. Incidence and Types of Normal and Abnormal Mitoses in 5,000 Dividing Cells of the Uterine Cervix . . . 75
10. Distribution of Normal and Abnormal Mitoses in Non-Malignant and Malignant Lesions ...... 76
11. Prophase-Metaphase Values and Rrophase Indices . . . 83
12. Statistical Significance . . . 8U
13. Critical PI Values ...... 87
lip. Statistical Significance of the Prophase Index in Various Cervical Lesions 90
15. Nuclear DNA of 585 Nuclei and Mitotic Index of Basal Cells in Various Stages of Estrus following Colchicine Treatment ...... 100
16. Summary of Experiments on the Nuclear DNA Content of Epithelial Cells in Various Cervical Lesions . . 103 X
LIST OF TABLES-Continued
Table Page
17. Nuclear DNA Classes and Mitotic Indices Following Insertion of Plain Thread ...... 10^
18. Nuclear DNA Classes and Mitotic Indices Following Application of Croton Oil ...... 106
19. Nuclear DNA Classes and Mitotic Indices Following Application of Crystalline 20-Methylcholanthrene ...... 108
20. Nuclear DNA Classes and Mitotic Indices Following Application of 20-Methylcholanthrene in oil . . . 110
21. Activity of Various Dehydrogenases in Normal and Malignant Cervical Epithelium ...... llj.0 hr Page Chart 1. Distribution of atypical Distributionnormal,and malignant 1. . vrg rqec fmttcsae. 78 79 abnormal Correlationmitosesand prophasebetween U. index. ofstages. mitotic Averagefrequency 3. . yormo xeietlaia o 7. 37 of Cytogramexperimentalanimal 70. No. 2. . h ula N otn neiemi acnm 112 DNAcontentincarcinoma Thenuclearepidermoid 6. vn. . Mitotic indices following endocervicalapplicationindices Mitotic . epithelial cells in various cervicalinepithelialcells various lesions. fvrossbtne. 102 substances.ofvarious LIST OF CHARTSOFLIST 36 xi INTRODUCTION
Spontaneous malignant neoplasms of the uterine cervix appear to be quite rare in both wild and domesticated mammals. No primary malignant cervical tumors were encountered in 3jh00 wild mammals autopsied at the Philadelphia Zoological Gardens (RATCLIFFE, 1933)•
However, cervical carcinoma has been reported in wild mammals; squa mous cell carcinoma in a gazelle (PETIT cited by MURRAY, 1906) and adenocarcinoma in a rhinoceros (BETHKE, 1911). More extensive systematic studd.es of diseases of domesticated mammals also show a very low incidence of cervical cancer (FELDMAN, 1932; and COTCHIN,
1936). On U,030 cases of spontaneous neoplasms of domesticated mam mals on file in the American Registry of Veterinary Pathology there are only two examples of cervical cancer, both arising, in cows (SMITH
AND JONES, 1937). Primary carcinoma of the uterine cervix is similarly rare in laboratory animals; according to SLYE (192li)and
SLYE et al. (192lj.), the incidence of spontaneous uterine tumors in mice is less than 0.06 per cent. However, in mice of the PM strain
GARDNER AND PAN (191+8) found 13 spontaneous malignant tumors of the uterine cervix, vaginal fornices and upper vagina in 36 mice, 3 epi dermoid carcinomas, 7 undifferentiated neoplasms of probable epi thelial origin, and 1 spindle cell sarcoma. Unfortunately, this
strain was lost because of a high incidence of sterility. In the 2 rat, on the other hand, spontaneous malignant epithelial neoplasms of the uterus are not rare. Although BULLOCK AND CURTIS (1930) observed
20 malignant epithelial uterine tumors in t heir series of 321 spontan eous tumors in the rat, none involved the uterine cervix.
The foregoing statistics clearly show that the incidence of spontaneous epithelial cervical cancer in mammals is by no means com parable to that observed in the human. Thus the induction of cervical cancer in laboratory animals for experimental study has necessitated the use of chemical and hormonal carcinogens. This dissertation will deal with the experimental induction, pathogenesis, pathologic anatomy, cytology and cytochemistry of these lesions.
Induction
Attempts to induce carcinoma of the uterine cervix experimen tally began shortly after the discovery of the carcinogenic properties of crude tar by YAMAGIWA AND ICHIKAWA (1918). The earliest experiments consisted of the intravaginal application of crude tar by instillation or implantation into the vaginal vault, cervix, and lower uterine seg ment. Although squamous metaplasia was readily induced, invasive carcinoma was observed in only a few cases (TEUTSCHLANDER, 1926; CIOLI,
1929; and FUSCO, 1932). The tumors induced were slow-growing, failed to metastasize, and often regressed despite continued application of tar
In view of the stimulatory effect of estrogens on the growth and development of genital tissues (ISCOVESCO, 1912; and FELLNER, 1912) and their carcinogenic action (LACASSAGNE, 1939; BONSER AND ROBSON, 19U0
HOOKER et al., 19U0; LIPSCHUTZ AND VARGA, 19Ul; and GARDNER et al.,
19UU), these substances were employed alone and in conjunction with other carcinogens and sex hormones in an effort to induce neoplasms of 3
the uterine cervix (LACASSAGNE, 1936; LOEB et al., 1936; SUNTZEFF et al.
1938; LOEB et al., 1938; GARDNER AND ALLEN, 1939; KLENITSKY, 19UO;
ALLEN AND GARDNER, 19lpL;■CROSSEN AND LOEB, 19hks PAN AND GARDNER, 19^8;
and WILLIAMS et al., 1933). A small number of lesions closely resemb
ling early epidermoid carcinoma were observed in the cervices of Macacus
rhesus monkeys that had received estrone parenterally following traum
atization of the uterine cervix (OVERHOLZER AND ALLEN, 1933). It is
interesting to note that the epithelial hyperplasia induced by estrogen
alone regressed following discontinuation of hormone treatment. However,
some of the animals which received estrogen in conjunction with cervical
trauma developed lesions closely resembling invasive carcinoma. The
true nature of these lesions remains obscure since no mention was made
by the authors of the absence or presence of either local pericervical
extension or distant metastasis. A small number of cervical carcinomas
were also observed in mice whose neck and back had been painted with
1,2,5,6 dibenzanthracene while they simultaneously received injections
of estrogen (FERRY, 1936; FERRY AND GINZTON, 1937). These results
suggest that estrogenic hormones possess a co-carcinogenic effect by
sensitizing the uterine cervix to the carcinogenic substance even though
the latter was applied at a distant site. However, owing to the small
number of animals which developed malignant lesions, it is difficult to
draw any valid conclusions from these experiments. Recently it has been
observed that estrogens may act as co^carcinogens in 20-methylcholan- threne-induced cervical carcinoma.^- Following castration only 8 of 25
E. von Haam and R. Albery (unpublished results). mice developed cervical tumors, while 17 of 23 mice receiving 1,000
mouse units of estrogen showed tumors. Mice treated with 3 units of
androgen failed to develop cervical neoplasms despite continued treat
ment with 20-methylcholanthrene. The prolonged subcutaneous adminis
tration of estrogens for periods of over one year induced cervical
carcinomas in various strains of mice; one of these metastasized widely
(GARDNER et al., 1938). Once again the percentage of tumor yield was
low and the mortality of mice from mammary carcinomas and pyometra was
high. This experiment was further complicated by the fact that the
development of mammary carcinomas occurred long before the appearance
of uterine cancer and necessitated surgical removal of the mammary
lesions to prevent premature death of the mice. More recently GARDNER
(1939) has extended these studies and reported that the latent period
of stilbestrol induced cervical and vaginal cancer was shortened by the
insertion of cholesterol pellets in the upper vagina. Additional ex
periments (GARDNER, 1939a) have shown that estrogen applied directly
to the vaginal and cervical mucosa exerts a greater carcinogenic effect
than hormone administered subcutaneously. This suggests that estrogen may act directly on the genital tissues as a carcinogen. According to
PAN AND GARDNER (191+8), testosterone propionate did not significantly
lower the incidence of estrogen induced cervical and vaginal carcinoma.
Following the pioneer chemical studies of KENNAWAY AND HEIGER
(1930) and COOK et al., (1933), experimental induction of malignant
tumors by the use of crude tar was replaced by more refined methods utilizing pure compounds possessing potent carcinogenic properties.
Purified carcinogenic substances applied directly to the uterine cervix
(D0BR0VOLOSKAIA-ZAVADSKAIA AND SIMSILEVITCH, 1933) resulted in a considerably higher yield of epidermoid carcinomas than occurred following either the application of crude tar or the injection or local application of estrogens. The epithelium of the vaginal vault and cervix of the mouse responded to carcinogenic hydrocarbons applied by painting in a manner similar to the epithelium of the skin, although the tumor yield from the vagina and cervix was significantly lower.
This may have been owing to the fact that application of carcinogenic substances to the uterine cervix was technically more difficult, or that the cervical and endometrial secretion diluted or removed the ap plied carcinogen so rapidly that it became much less effective.
VON HAAM AND MENZIES (1953) and VON HAAM AND SCARFELLI (1955) were able to induce carcinoma of the uterine cervix in a high percentage of C3H mice by intravaginal painting with a 1 per cent solution of 3>U-benz- pyrene in acetone applied by means of a wire-loop applicator which permitted accurate placement of the carcinogen to the uterine cervix and vaginal fornices. KOPROWSKA et al. (1958) induced a greater number of tumors with a shorter latent period of 10-30 weeks and fewer primary vaginal carcinomas by painting the cervix under direct observation through a speculum. MURPHY (1953) solved the problem of accurate place ment of carcinogen to the uterine cervix by inserting a string saturated with methylcholanthrene into the endocervicalc anal under direct visual ization with an infant-sized otic speculum; he was also successful in inducing a considerable number of cervical carcinomas. Utilizing a similar technique, REAGAN et al. (1955) reported a tumor incidence of
70 per cent. These results are in direct contrast to the lower yield of carcinogen-induced genital tract tumors in rats reported by McEUEN 6
(1938), VELLIOS AND GRIFFIN (1957), and GLUCKSMANN AND CHERRY (1957).
This is to be expected in view of the well known resistance of the rat to the carcinogenic effects of various polycyclic carcinogens (ICHIKAWA
AND BAUM, 19210. GLUCKSMANN AND CHERRI(1957) reported the development of sarcoma of the genital tract in 72 per cent of their experimental animals following treatment with estradiol benzoate and 9,10-dimethyl-
1,2-benzanthracene as compared to one carcinoma and one papilloma.
This observation suggests that in the rat connective tissue may be more susceptible to chemical carcinogens than squamous epithelium.
Malignant cervical and vaginal neoplasms have also been experi mentally induced by the use of pooled human smegma (PRATT-THOMAS et al.,
1956) and various non-polycyclic hydrocarbons (GARDNER, 1959; GARDNER,
1959a; and HOCH-LIGETI, 1957). PRATT-THOMAS et al. (1956) induced squamous carcinoma in the uterine cervix of mice by the application of human smegma. The long latent period of induction and low yield of tumors was felt to indicate that smegma is a weak carcinogen. These findings are in direct contrast to the negative results reported earlier by FISHMAN and his co-workers (19^1) in which no tumors were induced in mice even after 16 months of intravaginal painting with smegma. To date the active carcinogenic principle in smegma remains unknown. HOCH-LIGETI
(1957) induced a significant number of sarcomas and a single epidermoid
carcinoma in rats following continued application of various spermacidal compounds. GARDNER (1959 and 1959a) observed a large number of epiderm- moid carcinomas in the vagina of mice that had received intravaginal applications of urea, adipic acid and carboxymethylcellulose acids.
These results suggest a lack of specificity of the carcinogenic response, 7 at least insofar as vaginal epithelium is concerned, and demonstrate that prolonged exposure to unsuspected "non-carcinogenic" compounds may induce malignant neoplasms. Recently KOPROWSKA AND BOGACZ (195-9) re ported the induction of squamous carcinoma of the cervix and vagina in mice following prolonged application of crudd tobacco tar. Studies on the experimental induction of carcinomas of the uterine cervix and vagina are summarized in Table I.
As is evident above, numerous experiments have been undertaken to induce carcinoma of the uterine cervix. However, comparative studies on the efficacy of the various methods of induction have not been per formed. The present series or experiments represent such a comparative approach to the problem of the induction of cervical carcinoma in C^H mice.
MATERIALS AND METHODS
Three hundred and forty-six C^H female mice, weighing between
18 and 22 gnu, were divided into 10 groups and used for carcinoma in duction experiments, utilizing various carcinogens and methods of application as follows: Group 1 (60 animals) painted intravaginally twice weekly with a 1 per cent solution of 3}h benzpyrene (Edcan) in acetone; Group la (20 animals) served as controls and were painted in travaginally twice weekly with acetone. Painting was accomplished by means of cotton-tipped wire loops which were bent slightly to fit the curve of the genital canal (Fig. 1). The cotton-tipped wire loops were carefully placed in the vagina, which was stretched open by dorsal flexion of the tail, and inserted until firm resistance was felt (Fig.2) the cervix was then painted with 3 rotary motions and the wire loop re moved. Groups 2 (80 animals), 2a (30 control animals), 3 (86 animals), 8
Fig. 1 A 1 per cent solution of 3,14. benzpyrene in acetone and a
bacteriologic wire loop.
Fig. 2 Technic for intravaginal painting by means of a carcinogen
impregnated cotton-tipped bacteriologic wire loop. 9
•z/>/**< l*•** r * t i £ ± TABLE 1
SUMMARY OF EXPERIMENTAL INDUCTION OF NEOPLASMS OF THE UTERINE CERVIX AND VAGINA
Number, Species, Strain Induction Year And Of Experimental Mode Of'9* Time No. Reference Animals Carcinogen Application (Wks.) Tumors Remarks'3*
1926 (159) White rat, No. ? Crude tar I.V., skin 2U 1 D
1929 (22) White mice, No. ? Crude tar in xylol I.V. — — EC
1932 (k7) 50 white rats Crude tar I.V. Uo 1 EC
1933 (116) 9 Macacus rhesus Estrone and cervical S.C. 13 3 D monkeys trauma
1935 (33) White mice, No. ? Crude tar, 1,2:5:6 I.V. 22 1 EC dibenzanthracene
1936 (88) UO mice Estrone S.C. 11-29 1 EC
1936 (98) 1 mouse "Old Buffalo" Estrone S.C. 10U 1 EC (?)
1936 (126) 27 mice 1,2,5,6-dibenzan- Skin-painted 28 3 EC thracene, Estrone
1937 (127 75 mice 1,2,5,6-dibenzan- Skin-painted 26-U0 — EC thracene, Estrone
1938 (156) 235 mice - A, C57, Estrone S.C. 106 26 D D, C3H, CBA, new and Old Buffalo
H O TABLE 1— Continued
Number, Species, Strain Induction Year And Of Experimental Mode Of* Time No. Reference Animals Carcinogen Application (Wks.) Tumors Remarks
1938 (50) C3H mice, No. ? Estradiol and Estrone S.C. 38-52 19 InC benzoate EC
1938 (99) 32l* mice - A, C57, Estrone benzoate S.C. l*-80 0 M D, new and Old Buffalo, C3H, CBA
1938 (102) 12 hooded rats Estrone I.V. la 1 - EC D
1939 (51) 131* mice - C3H, CBA, Estrogens and testos S.C. 55 26 D A, C121, C57, N, terone propionate EC JK, F
19U0 (81) 170 white mice Coal tar, Tar I.V. 30-35 2 EC Folliculol S.C.
19l*l (1*) i+li mice - C57, CBA Estradiol benzoate S.C. 52 25 D
191*2 (1*1*) 10 mice - A 3 , l*-be nz pyr ene I.V. 28-56 10 EC 20 mice - A Human smegma I.V.
191*8 (120) 35 mice - C3H Estrogens and androgens S.C. 29 13 D
1953 (166) Bale mice, No. ? Estradiol sensitized S.C. 61* 38 D Collodion particles EC
1953 (62) 80 mice - C3H 3, l*-be nzpyr ene I.V. 16-1*0 31 CiS TABLE 1— Continued
Number, Species, Strain Induction Year And . Of Experimental Mode Of* Time No. Reference Animals Carcinogen Application (Wks.) Tumors Remarks *
1953 (112) 7l* mice - A, C57 Methylcholant hrene- i.e. 30-1*0 27 InC saturated string
1951* (63) 80 mice - C3H 3, U-benzpyrene I.V. 16-1*0 36 CiS InC
1955 (11*2) 86 mice - C3H 3, lj.-benzpyrene I.V. 2l*-53 1*1* CiS InC
1955 (138) 65 mice - C3H 20-methylcholanthrene I.e. 7-16 1*6 InC
1955 (11*2) 15 mice - C3H 3, li-benzpyrene I.e. 12-38 11 CiS InC
1955 (11*2) 20 mice - C3H 20-methylcholanthrene i.e. 11-33 17 CiS InC
1956 (132) 36 mice - dba-1 Pooled human smegma I.V. 60-81 10 InC MP S
1957 (163) lj.8 rats Wistar 7,12 dimethylbenz a i.e. 16-28 13 CiS anthracene InC
1957 (71) 95 rats Wistar 8-ortho-hydroxyquino- I.V. 69-86 21 F,P line Ec,S p-dii s obutylphenoxy- polyethoxyethanol
PO TABLE 1— Continued
Number, Species, Strain Induction Year And Of Experimental Mode Of* Time No. Reference Animals Carcinogen Application (Wks.) Tumors Remarks
1958 (58) 68 rats hooded 9,10 dimethyl-l,2-benz- I.V. U7 2 Ec,P anthracene, Estradiol I.M. 2k S
1958 (85) 30 mice - C3H 3,U-benzpyrene I.V. 18-20 26 CiS InC
1959 (8U) 30 mice - C3H Crude tobacco tar I.V. 17 CiS InC 30 mice - C3H 3,U-benzpyrene I.V. 19-30 30 CiS InC
1959 Ui8) 21 mice - BC, C57 Stilbestrol I.V. U6 8 EC
1959a(U9) ko mice -BC Stilbestrol-eholesterol I.V. 37 20 EC pellets 13 mice - BC Stilbestrol-cholesterol- I.V. 36 6 EC testosterone propionate 37 mice - BC Estradiol benzoate S.C. 33 15 EC Stilbestrol 27 35 mice - BC Urea, adipic acid I.V. 70 21 InC Carboxymethylcellulos e acids
* S.C. » Subcutaneous . ** M = Metaplasia EC (?) = carcinoma-like I.V. = Intravaginal P = Polyp or papilloma InC = Invasive epidermoid carcinoma I.C. = Intracervical D = Precancer or dysplasia S = Sarcoma I.M. = Intramuscular CiS = Carcinoma in situ MP = Malignant papilloma EC = Epidermoid carcinoma F = Fibroma M Ill and 3a (29 control animals) were treated in an identical manner. In
groups U and 5 carcinogen was applied by means of a carcinogen-impreg nated string inserted endocervically according to the technique des cribed by MURPHY (1993) (2k) as follows: Group I4 (19 animals) 3,k benz- pyrene-impregnated string; ,• Group lia (3 control animals) plain string;
Group 9 (20 animals) 20-methylcholanthrene (Edcan)-impregnated string;
Group 9a (9 control animals) plain string.
All animals were kept in groups of 9 in air conditioned rooms and fed Purina Dog Checkers supplemented by daily addition of fresh bread and carrots. Vaginal smears were prepared prior to each painting period by injecting a small quantity of saline into the vagina followed by im mediate aspiration. The smears were fixed immediately and stained by the method of PAPANICOLAOU (I9i|2) and examined for the degree of estrus, the presence of leukocytes and erythrocytes, and the appearance of atypical and malignant epithelial cells.
RESULTS
One hundred and thirty-nine mice (93.2 per cent) developed either carcinoma of the uterine cervix or vagina. A considerable variation in the incidence of tumors and the latent period of induction was observed in the 9 groups treated with carcinogen. In no instance was cervical or vaginal carcinoma observed in the control mice. The percentage of animals which developed carcinoma varied from Jj.9 to 9l «6 per cent for those painted with carcinogen (Groups 1,2 and 3) to 73 and 89 per cent for those in which carcinogen-impregnated strings were placed in the endocervical canal (Groups i* and 9). The latent period of induction, which ranged from 22 to 91| weeks in the painted mice (Groups 1,2, and 3)> is
was shortened to 11 to 33 weeks when carcinogen-impregnated strings were
used, these findings are summarized in Table 2.
DISCUSSION
In view of the rapidity with which carcinogens applied by paint
ing are removed from the epithelial surface (HEIGER, 1936; AHLSTROM AND
BERG, 19U7); HEIDELBERGER AND WEISS, l?Sl), the continuous application
of carcinogen to the cervical epithelium afforded by an endocervical
carcinogen-impregnated string may be one of the factors responsible for
the differences noted between these mice and those treated by painting.
A second factor worthy of consideration is the possible co-carcinogenic
or irritative effect of the continuous low grade trauma resulting from 2 the string. KOPROWSKA ascribed the higher yield of cervical carcinoma
she obtained following intravaginal painting through an otic speculum not only to more accurate placement of carcinogen but also to the trauma
caused by the speculum. Although the effect of trauma on carcinogenesis is difficult to ascertain, as evidenced by the equivocal results obtained by various investigators (BERENBLUM, 19%k), further experiments are necessary to determine the role of trauma in cervical carcinogenesis.
KOPROWSKA, personal communication TABLE 2
EXPERIMENTAL INDUCTION OF CARCINOMA OF THE UTERINE CERVIX AND VAGINA IN C3H MICE
Method of No. of No. of Weeks of Group Induction Mice Carcinomas Percentage Development
1 P-BP 60 31 51.6 22-51*
2 P-BP 80 36 W . o 16-lj.O
3 P-BP 86 1+1+ 51.2 21+-53
1+ S-BP 15 11 73.3 12-38
5 S-MC 20 17 85.0 11-33
261 139 53.2
P-BP = painted with a 1 per cent solution of 3,1+ benzpyrene in acetone S-BP ** treated with 3,1+ benzpyrene-impregnated endocervical string S-MC = treated with 20-methylcholanthrene-impregnated endocervical string 17
PATHOGENESIS: A CORRELATIVE CYTOLOGIC
AND HISTOLOGIC STUDY
Exfoliative cytology was first employed by POUCHET (18U7) to
study ovulation. Six years later DONALDSON (185?3) described the bizarre
morphology of neoplastic cells he found in "tumor juice" as follows, "the
polygonal, more or less spherical, and ovoid cell; the caudated cell;
the fusiform cell; the concentric cell; and agglomerated nuclei connected
by amorphous tissue". This was followed by isolated reports of further
clinical applications of exfoliative cytology, notable among which is the
communication of BEALE (i860), who described the cytologic components
present in the sputum from a case of pharyngeal carcinoma. Diagnostic
cytology remained generally unknown until the classic monograph on the
diagnosis of genital cancer by vaginal smears by PAPANICOLAOU AND TRATJT
(19k3)• Today exfoliative cytology is employed widely for the diagnosis
of both neoplastic and non-neoplastic diseases in a considerable number
of organ systems.
From the experimental point of view exfoliative cytology offers
several distinct advantages in the study of the pathogenesis of disease processes affecting epithelial tissues. First and foremost, it allows
a serial study of a lesion during its development by virtue of the
steady random desquamation of cells and without disturbing it by biopsy.
As OXORN (19U8) has so well expressed it, exfoliative cytology is "a kind
of natural curettage going on without interruption, and always providing fresh and easily obtainable material for study." In addition, cytologic 18 material is easily handled and very often of sufficient amount to allow numerous studies by diverse techniques. VON HAAM AND MENZIES (195>3) were able to demonstrate malignant cells in vaginal smears of experimen tal animals sometime before the gross appearance of tumors and pointed out the advantages of this technique for studies of the pathogenesis of experimental cervical carcinoma. Since the neoplastic transformation is accompanied by a decreased mutual adhesiveness of malignant cells as shown by COHAN (I9l4l)j the rate of cellular exfoliation is increased and the vagina becomes a "sac containing free cells" providing a steady supply of malignant cells." It was felt that this method would lend itself well to comparative study of the cytologic and histologic alter ations which occur during the development of the neoplastic process.
The experimental data presented in the following sections is the result of studies on several hundred malignant tumors induced at Ohio State
University during the past 5 years (VON HAAM AND SCARFELLI, 19!?5; and
SCARPELLI AND VON HAAM, 19^7).
MATERIALS AND METHODS
Cytologic material used in this study was obtained bi-weekly by vaginal aspiration as described previously from mice of Group 3* In addition to Papanicolaou staining, additional smears were prepared for cytochemical studies to be reported in a subsequent section of this dis sertation.
As soon as abnormal cells appeared in sufficient number to be considered diagnostically significant, the animals were sacrificed, the uterus and vagina removed, and serial sections prepared for microscopic examination. Seven groups of animals were sacrificed during a period 19
from 12 to lj.0 weeks after the beginning of the experiment; this per
mitted a comparison between early and late cytologic and histologic
changes produced by the carcinogen. By means of differential counts,
quantitative "cytograms11 were constructed; although they do not reveal
the complete estrous cycles of the animals, they permitted one to
follow the cytologic pattern observed during the period of experimental
carcinogenesis.
An attempt was made in each instance, whenever possible, to
correlate the morphology of exfoliated cells with the tissues from which
they originated. This was accomplished by a terminal vaginal smear
obtained immediately prior to sacrifice.
RESULTS
Cytologic studies revealed changes in the morphology of exfoli
ated cells in various phases of carcinogen-induced carcinoma of the
uterine cervix in mice. On the basis of both cytologic and histologic findings, four distinct lesions could be differentiated: (a) acute in flammation of the cervix and vagina; (b) dysplasia of the cervical and vaginal epithelium; (c) non-invasive, intra-epithelial carcinoma (car
cinoma in situ); (d) invasive carcinoma. Except for necrosis of
squamous cells, acute inflammation was reflected mainly by the appear
ance of exudate in the vaginal smears.
Definite nuclear and cytoplasmic changes were observed in ex foliated cells during the development of cervical dysplasia and carcin
oma. Certain cellular changes preceded the appearance of neoplasraa by
a considerable time and were classified as precancerous changes. Other
changes were demonstrable only in the presence of either intra-epithelial carcinoma or invasive cancer and were considered as changes typical of 20
malignant cells.
Acute Inflammation
The early response of cervical epithelium to a twice weekly
application of 3*U benzpyrene in acetone was reflected in the vaginal
smear by large numbers of polymorphnuclear leucocytes, many of which
showed clumping around necrotic squamous cells (Fig. 3)- During the
inflammatory reaction the estrous cycle continued normally, the only
exception being the continuous presence of polymorphonuclear leuco
cytes in the vaginal smears and genital tract mucosa. The tissue
sections showed a focal polymorphonuclear infiltration of the squamous
epithelium forming small microabscesses. In more severe cases there were in addition varying numbers of polymorphonuclear leucocytes in
the submucosa accompanied by edema (Fig. U). Rarely was there focal
acute ulceration of the squamous mucosa. Although these changes are
somewhat similar to those observed during metestrus in the mouse (ALLEN,
1922), the severity of the exudative reaction and its persistence for
12 to 15 days is evidence in favor of an inflammatory response to intra- vaginal painting. Its subsidence after 2 to 3 weeks suggests a possible
adaptation of the mucous membranes to these physical and chemical irri
tants. A chronic inflammatory reaction was not observed despite the
continuation of painting, although occasionally a few multinucleated % histiocytic giant cells were observed in the vaginal smear (Fig. 5).
However, a careful search of tissue sections failed to reveal the cervix
and vagina as the source of their origin. It is possible that these
cells may have originated from the endometrium. 21
Fig. 3 . Vaginal smear. Acute cervicitis. Necrotic squamous cell
surrounded by a ring of polymorphonuclear leukocytes.
X 750.
Fig. h Acute cervicitis. Tissue section. Showing extensive edema
with numerous polymorphonuclear leukocytes in the submucosa.
X 290
Fig. 5. A multinucleated histiocytic giant ceil occasionally found
in vaginal smears after 3—1+ weeks of intravaginal painting
with 3jU benzpyrene. X 1250 • s n. 23
Epithelial Ejysplasia
Epithelial dysplasia of the uterine cervix and vagina was characterized cytologically primarily by cytoplasmic alterations in exfoliated basal and parabasal cells. These consisted of varying de grees of increased cytoplasmic basophilia, the morphology of which ranged from a fine granularity to large coarse granules (Fig. 6). In addition many basal cells contained numerous cytoplasmic vacuoles
(Fig. 7). Tissue sections showed a gross disturbance in the differ entiation and keratinization of the basal cell layer with disorderly arrangements of the cells without either mitotic figures or cell crowd ing indicative of increased proliferation. There was also some loss in the polarity of basal cells, with small focal intraepithelial accumu lation of leucocytes forming microabscesses (Fig. 8). An interesting cytoplasmic alteration which appeared to originate in the parabasal cell layer consisted of an intense eosinophilia (Fig. 9). This appeared to be keratin and was associated with an intact, though frequently damaged, nucleus. In the tissues these cells were present in small cystlike spaces in the epithelium (VON HAAM AND SCARPELLI, 1955), and in many instances masses of such "precociously keratinized" cells were seen exfoliating into the vaginal lumen (Fig 10). Many of these cells showed various stages of cell damage or were frankly necrotic. Nuclear alterations in exfoliated basal and parabasal cells consisted of decreased density of chromatin, nuclear enlargement and multinucleation.
All of the cells described above were considered atypical and were not present during the inflammatory phase.
In cases of advanced dysplasia cell necrosis disappeared and evidence of increased cell growth appeared. This was characterized by 2k
Fig. 6. Cervical dysplasia. Vaginal smear. Basal cell containing
cytoplasmic granules. X 12^0
Fig. 7. Cervical dysplasia. Vaginal smear. Basal cells containing
numerous cytoplasmic vacuoles. X 12^0
Fig. 8. Cervical dysplasia. Tissue section. There is a disorderly
arrangement of both basal and parabasal cells with numerous
intra-epithelial microabscesses. X 1^0.
Fig. 9. Cervical dysplasia. Vaginal smear showing a cluster of pre
cociously keratinized basal cells. The cytoplasm is intensely
eosinophilic. X 900 1 1 *
9
.*J4 t? 26
Fig. 10. Cervical dysplasia. Tissue section. A mass of intensely
eosinophilic "precociously-keratinized" basal cells exfoli
ating into the vaginal lumen. X 300
Fig. 11. Advanced cervical dysplasia. Vaginal smear. Plump, hyper-
chromatic atypical basal cells. X 800
Fig. 12. Advanced cervical dysplasia. Tissue section. Note the
numerous papillomatous areas of basal cell overgrowth con
sisting of moderately disordered arrangements of hyperchro-
matic basal and parabasal cells. X 370
28 the appearance of hyperchromatic basal cells which did not have the characteristics of neoplastic cells (Fig. 11). Tissue sections showed extensive papillomatous overgrowth consisting of moderately disordered arrangements of hyperchromatic basal and parabasal cells, a few of which were in mitosis (Fig. 12). The presence of papillomatosis, plump hyperchromatic cells near the mucosal surface, and the thin layer of keratin is felt to be indicative of rapid cellular growth. Eysplastic changes, with the exception of precocious cornification and papilloma tous epithelial hyperplasia, preceded the appearance of carcinoma by a considerable period of time and were also occasionally found in control animals painted with acetone alone.
Carcinoma in Situ
Cells exfoliated from this lesion showed marked variations in shape and size, large nuclei with a definite increase in the nuclear- cytoplasmic ratio (Fig. 13). The chromatin pattern was coarse and the nuclei contained multiple nucleoli. The cytoplasmic changes were usually similar to those present in epithelial dysplasia except for atypical keratinization which was found much more frequently, as evi denced in the vaginal smear by the appearance of small cells with a deep orange colored homogeneous cytoplasm. Tissue sections showed a marked proliferation of basal and parabasal cells within widened epith elial boundaries but within intact basal membranes (Fig. 14). The papillomatous overgrowth was quite marked in carcinoma in situ (Fig. lj?) and differed from the pattern present in the late stages of dysplasia
(epithelial hyperplasia) in that the papillomatous projections of cells 29
were larger and tended to crowd one another. The nuclei of the cells
were enlarged and hyperchromatic; there were increased numbers of
mitotic figures; and there was definite "cell crowding" (Fig. 16).
Because of the latter phenomenon, the atypical appearance of individu
al cells was at times not as impressive in carcinoma in situ as in
epithelial dysplasia.
Invasive Carcinoma
Vaginal smears from this lesion were generally similar to those
found in animals with intra-epithelial carcinoma with the exception
that neoplastic cells appeared in greater numbers (Fig. 17) and included fiber cells and exfoliated keratin'pearls (Fig. 18, 19) which originated from the tumor surfaces. The tumors induced were predominantely well
differentiated squamous cell carcinomas (Fig. 20). The islands of neo plastic cells showed numerous keratinized pearly bodies and were usually surrounded by severe inflammatory reactions. Exfoliated keratin pearls and fiber cells were absent in cases of pure carcinoma in situ.
In addition, vaginal smears from animals with invasive carcinoma con tained red blood cells and large numbers of polymorphonuclear leucocytes.
Red blood cells were absent in the vaginal fluid during the stages of dysplasia and present in small numbers in intraepithelial carcinoma.
Poorly differentiated anaplastic carcinomas were obtained in a few animals (Fig. 21). Vaginal smears from these animals contained large numbers of small highly undifferentiated tumor cells showing no evidence of keratinization (Fig. 22). 30
Fig. 13. Carcinoma in situ. Vaginal smear. Malignant basal cells
■ showing hyperchromatism, marked variation in size and shape,
and indistinct cytoplasmic borders. X 12^0
Fig. lit. Carcinoma in situ. Tissue section. Large areas of basal
cell overgrowth forming intraepithelial nodules within an
intact basement membrane. X 125
Fig. 13. Carcinoma in situ. Tissue section. Showing a pattern
similar to papillomatous hyperplasia but with numerous
malignant cells showing a disordered pattern of growth. X 370
Fig. 16. Carcinoma in situ. Tissue section. Photomicrograph showing
hyperchromatism, mitotic figures and marked cellular crowding.
x 1*73
32
Fig. 17. Vaginal smear. Malignant basal cells containing numerous
macronucleoli. X 12^0
Fig. 18. Vaginal smear. Cluster of elongated malignant squamous
cells (fiber cells). X 550.
Fig. 19. Vaginal smear. Exfoliated epithelial pearl consisting of
malignant squamous cells. X 890
Fig. 20. Well differentiated epidermoid carcinoma of the cervix.
Marked keratinization with pearl formation. X 2^0
Fig. 21. Poorly differentiated carcinoma . Tissue section showing
a tumor composed of bizarre, hyperchromatic tumor cells.
Note the paucity of stroma, numerous giant tumor cells and
lack of keratinization. X 115?
Fig. 22. Vaginal smear. Malignant basal cells exhibiting marked
pyknosis (anaplastic carcinoma). X 12j?0
Cytohistologic Correlation
Table 3 summarizes quantitative and qualitative data obtained
by cell counts and their correlation with the various cervical and vaginal lesions. The cytologic findings are those present in vaginal smears obtained immediately prior to sacrifice. The number of atypical cells in the vaginal smears varied from 1+ to 26 per cent in animals with epithelial dysplasia, from 8 to 38 per cent in animals with car cinoma in situ, and from 26 to 61* per cent in animals with invasive carcinoma. Malignant cells, which were absent in a 11 but one of the cases of epithelial dysplasia, were found in 3-17 per cent of the cases of carcinoma in situ and in 3-53 per cent of the cases of invasive carcinoma. The fact that one animal (No. 55) with histologically in vasive carcinoma exfoliated as few as 3 per cent recognizably malignant cells seems interesting in view of the reported false negative smears in cases of human cervical carcinoma. Chart 1 shows the average distribution of normal, atypical, and malignant epithelial cells in the vaginal smears according to the types of histologic lesions and clearly demonstrates the sharp cytologic difference between epithelial dysplasia and carcinoma in situ.
On the basis of weekly differential counts, other interesting differences were found to exist between dysplasia, carcinoma in situ and invasive carcinoma. Smears from animals with dysplasia and carcin oma in situ showed no significant changes in the normal epithelial cell population, while in animals with advanced invasive tumors a marked increase in basal cells existed. This is shown in Chart 2 which repre sents the typical cytogram obtained from the vaginal smears during the TABLE 3
CYTO-HISTOLOGIC CORRELATION
Range of Distribution Animals with Lesions of Normal Epith. Cells Range of Range of Week Carci Invasive Cornif. Atypical Malignant of Cervi- Bys- Hyper noma carci Basal Cells Lutein Cells Cells Exper. Citis plasia plasia in situ noma Cells (percent) Cells (percent) (percent)
12 2 70-75 20-25 5 k-9 0 1 65 30 5 h 0 2 35-Uo 25-30 30-ii5 6-7 0 1U 1 30 ho 30 5 0 h 10-70 15-30 10-70 6-16 0 16 5 li0-90 5-30 5-30 12-22 3-6 18 3 65-75 15-30 5-io 8-17 2—It 20 h 65-85 10-30 5-i5 11-52 3-17 1 75 15 10 h3 12 22 h 60-65 15-30 5-20 12-38 h-9 2h 1 60 20 20 10 0 6 25-80 15-30 5-ii5 16-37 2-11 2 85-95 5-l5 0 38-6U 26-1+3 ko 1 75 15 5 5 15 11 60-95 5-25 o-i5 26-i|6 3-53
Uj VJT CHART 1
CO _zl Basal cells
LlJ _ _ o 80 Cornified cells Luteinized cells \±. Atypical cells o 60 _ Malignant cells
40 _ LlI O < c r 20 Ll J \
0 - \ Cervicitis Epithelial Carcinoma Invasive dysplasia in situ carcinoma
TYPE OF LESION
VjJ O CHART 2
CYTOGRAM OF ANIMAL 70 % BASAL CELLS 80 60
40 KERAT'NIZED CELLS 20 ' I Y LUTEINIZED CELLS 20 24 28 32 36 40
80 H IS T 0P A T H 0L 0G Y : ADVANCED CARCINOMA OF CERVIX 60 PRECOC. CORN. CELLS 40
20 ATYPICAL CELLS /« / ' MALIGNANT CELLS 0 4 20 24 28 32 36 40 WEEKS
Cytogram of experimental animal No. 70 sacrificed after LO weeks. Note the change of the cyclic cellular pattern after malignant cells appear in larger numbers (at about 30 weeks). 38
pathogenesis of cervical carcinoma. The cytogram is that of animal 70 which was sacrificed i|0 weeks after the beginning of the experiment
with a large fungating carcinoma of the cervix; malignant cells had
been demonstrated constantly for the last 12 weeks. The cyclic nature
of the cellular pattern was markedly altered at about the 2Uth week,
and lutenized cells were almost absent for the last 10 weeks.
In most of the animals with epithelial dysplasia or non-
invasive carcinoma of the cervix, the cyclic cellular pattern was only
slightly disturbed, and luteinized cells could be found in considerable
numbers at set intervals. In animals with invasive carcinoma, keratin
ized and luteinized cells decreased sharply, and basal cells of the parabasal and intermediate types appeared to predominate. Red blood
cells were generally absent during the stages of dysplasia and non-
invasive carcinoma. They were always a prominent feature in animals with invasive carcinoma.
DISCUSSION
The pathogenesis of 3,U benzpyrene-induced cervical and vaginal
carcinoma in C^H mice was characterized by the following sequences of
cytologic and histologic changes: 1. acute inflammation; 2. epith
elial dysplasia including papillomatous epithelial hyperplasia;
3« intraepithelial carcinoma or carcinoma in situ; and U. invasive
carcinoma. Qualitative and quantitative changes in cell morphology
and population demonstrated by exfoliated cells, paralleled the histo logic alterations present in the tissues during the development of
carcinoma. Similar findings have been reported by REAGAN et al. (1955)
and KOPROWSKA et al. (1958). These cytologic changes can be classified 39 into three categories: 1. non-specific inflammatory reaction; 2. pre- cancerous changes; and 3* malignant alterations. The non-specific inflammatory reaction is probably the response of genital epithelium to the trauma of painting or the irritative effect of the endocervical string and the noxious properties of the carcinogen and solvent. Its subsidence after 2 to 3 weeks suggests a possible adaptation of the mucous membranes to these physical and chemical irritants. Dysplastic changes, with the exception of precocious cornification and papillom atous epithelial hyperplasia, preceded the appearance of carcinoma by a considerable period of time and were also occasionally found in control animals painted with acetone alone. The hyperchromatic, plump, polygonal basal cells exfoliated from animals with papillomatous epi thelial hyperplasia bear a striking resemblance to the "dyskaryotic" cells described in human material by various investigators (LAPID AND
GOLDBERGER, and REAGAN et al., 1933)* A relation appears to exist between impairment of normal growth and maturation of vaginal and cervical squamous epithelium, and the appearance of precociously corni- fied basal cells. Whether this cytologic alteration is an abnormal response of carcinogen-treated basal cells to estrogen, or merely a direct toxic effect of the carcinogen on the basal cells, remains at present unknown.
Malignant transformation was primarily typified by alterations of the nucleus and nucleolus, i.s., hyperchromasia, nuclear enlargement and multiple macronucleoli. Nuclear pyknosis, cytoplasmic basophilia, precociously cornified basal cells, leucocytes, and red blood cells were so constantly found in animals with invasive or non-invasive ko
carcinoma that they could be spoken of as representing a "malignant pattern", although none of these changes alone was diagnostic of malig nancy. This observation is identical to that made by VON HAAM (19$h) following the examination of a large number of vaginal smears in women.
On the basis of purely cytologic criteria applied to a study of exfoli ated cells3 a differentiation could not be made between invasive and non-invasive carcinoma. This finding is similar to that reported by
REAGAN AND HAMONIC (19^6) in a cytologic evaluation of carcinoma in situ and invasive cervical cancer in women. As stated, however, quantitative differences appeared to exist as evidenced by earlier and greater numbers of neoplastic and r,ed blood cells present in vaginal smears obtained from-mice with invasive carcinoma. These two lesions were differentiable only on the basis of histopathologic criteria, i.e., non-invasive carcinoma is a lesion contained within an intact basement membrane.
Recently REAGAN AND WENTZ (19^9) have reported on the changes antedating the development of carcinoma and failed to find an "in situ" stage. In their study, dysplasia was followed by epithelial infiltra tion indicative of early invasive carcinoma. It is significant, how ever, that they induced cancer by means of a carcinogen-impregnated string. By virtue of the high doses of carcinogen applied and the short latent period of cancer induction by the string method, a direct comparison with painting experiments is not possible. It may well be that the in situ stage of 20-methylcholanthrene string induced-carcinoma is of such brief duration, that short continuous exposure periods of several days duration are necessary for its development. Other Investigators employing painting techniques(VELLIOS AND GRIFFIN, 190:7; and KOPROWSKA et. al., 1 9 % ) and utilizing weak carcinogens, such as, human smegma (PRATT-THOMAS et al., 19f>6) or crude tobacco tar (KOPROW
SKA AND BOGACZ, 19>9) also induced a considerable number of non- invasive carcinomas. Thus, the duration of the intra-epithelial phase appears to depend on the potency of the carcinogen as well as the dosage and method of application.
It is interesting that the normal cyclic cellular pattern of the animals continued during the first two stages of neoplastic development and was only disturbed in animals with invasive carcinoma. At present it is not known whether this is a failure of vaginal epithelium to re spond to estrogen or merely a reflection of debilitation on the hormonal physiology of the animals. Epithelial dysplasia preceded the appearance of non-invasive carcinoma by a considerable length of time. The fact that it was only very occasionally observed in our control animals suggests that it may belong to the group of precancerous lesions, although additional experiments are needed to support this concept. PATHOLOGIC ANATOMY AND HISTOLOGY
Although squamous cell carcinoma of the uterine cervix has
been induced experimentally in mice by numerous investigators, the path
ologic anatomy of these tumors has not been described. This section
deals -with a detailed description of the gross and microscopic anatomy
of 139 experimental tumors of the genital tract induced in the previous
study.
MATERIALS AND METHODS
When cytologic evidence of carcinoma was observed, the animals were sacrificed at various periods so that both early and advanced
lesions could be studied. The uterus and vagina were removed; if there was gross evidence of extension of tumor or metastases, the involved
lymph nodes and organs were removed also, fixed in 10 per cent buffered
formalin, and embedded in paraffin. Serial sections were stained with
hematoxylin and eosin and examined microscopically. Selected sections
from each tumor were stained by the Regaud modification of Heidenhain's
iron hematoxylin method for demonstration of tonofibrils. The tumors were classified by the criteria suggested by MARTZLOFF (1923).
RESULTS
Prior to the presentation of experimental data, the gross and microscopic anatomy of the adult mouse uterus will be briefly described for purposes of orientation. The uterus is composed of two horns,
3-10 mm. in length, which fuse in a Y fashion to form an undivided
caudal segment, the corpus uteri (Fig. 23). The lumina of the uterine
h2 h3 horns are lined by columnar cells in simple branched uterine glands.
The lamina propria consists of reticular tissue. The mucosa is called the endometrium. The columnar epithelium gradually changes to one lined by low cuboidal cells with a transition to stratified squamous epithelium high in the corpus uteri (Fig. 2k). The myometrium surround ing the endometrium consists of circularly arranged smooth muscle fibers.
The outer surface of the myometrium is covered by a layer of loose con nective tissue containing large blood and lymphatic vessels called the stratum vasculosum. This in turn is covered by longitudinal smooth mucle fibers. A serous membrane covers the uterine horns and is con tinuous with the broad ligaments. At the point of fusion of the uterine horns, the longitudinal muscle layer and stratum vasculosum disappear.
Laterally on each side of the corpus uteri the lumen of the vagina forms a high and deep fornixj the dorsal and ventral walls of the corpus are fused with the vagina. The vagina is covered in its entirety by squamous epithelium.
The lymph nodes which drain the uterus, broad ligament and asso ciated structures in the mouse consist of three groups (DUNN, 19kk) distributed as follows: 1. lumbar-paired lymph nodes lying above the bifurcation of the aorta, 2. caudal - a solitary node lying below the bifurcation of the aorta, 3* sciatic-paired lymph nodes lying beneath the gluteal muscles at the site of appearance of the sciatic nerve.
These are shown diagrammatically in Fig. 2k.
Distribution of Lesions
Malignant lesions restricted to the uterine cervix were found in kl animals (lj.1.0 per cent), lesions involving both the uterine n
Fig. 23. A low power microphotograph of a mouse uterus. Note the
two horns which fuse in a Y fashion to form an undivided
V.V caudal segment. The cervix can be seen projecting into
the vaginal lumen. X 10,5
Fig. 2U. The squamo-columnar junction of the mouse uterus located
high in the corpus uteri. X 35
Fig. 25. Diagrammatic representation of pelvic lymph nodes in the
mouse.
A - Kidney B - Lumbar group C - Sciatic group D - Caudal lymph node
b6
cervix and vagina were present in 55 animals (39*5 per cent), and
carcinomas limited to the vaginal mucosa were observed in 27 animals
(19.5 per cent). These are summarized in Table i±. In the series in which an endocervical string was used, a greater percentage of tumors was localized solely in the uterine cervix, in contrast to the con
siderable number of vaginal carcinomas and lesions involving both sites when tumors were induced by painting (Groups 1,2 and 3). Animals in which either separate carcinomas were found in the uterine cervix and vagina or in which the lesions were so extensive that the primary sites were not ascertainable were interpreted as having developed carcinoma
at both sites.
TABLE k
SITES OF NEOPLASIA
Groups Vagina Cervix Both Total
1 8 11 12 31
2 9 11 16 36
3 9 13 22 UU
h 1 8 2 11
5 0 lh 3 17
Total 27 57 55 139
Per cent 19.5 Ui.o 39.5 100 h7
Pathologic Anatomy
The earliest malignant lesions observed by gross examination of the cervix and vagina appeared as irregular reddish granular areas which bled readily when touched or gently rubbed with a cotton-tipped appli cator* The majority of invasive tumors were of the exophytic type and varied considerably in size. The largest tumor measured 2.7 x 2.2 x
2.it cm. (Fig. 26), while the smallest lesion was demonstrable only under the microscope. Most of the cervical lesions were located either near the external os at the portio vaginalis (Fig. 27), op high on the lateral cervical walls (Fig. 28). No lesions were found at the squamo- columnar junction. Early vaginal lesions were generally localized in two areas: either low near the vaginal orifice or at the vaginal for- nices. A considerable number of animals in Groups 1, 2 and 3 showed neoplastic changes involving large areas of both vaginal and cervical epithelium (Fig. 29).
In advanced lesiohs the vagina was filled with a large mass which surrounded the uterine cervix, showed foci of ulceration and hemorrhage, and obliterated the vaginal fornices (Fig. 30). All animals with advanced lesions showed evidence of vaginal bleeding. On section many of the larger tumors showed extensive necrosis with superimposed infection (Fig. 31)* A considerable number of invasive carcinomas showed extension and metastases into the pelvic lymph nodes (Fig. 32), lateral pelvic wall, broad ligaments, urinary bladder, rectum (Fig. 33),
Kidney, liver, and lung (Fig. 3k), in a manner strikingly similar to that seen in human cervical carcinoma. Ureteral obstruction with accom panying hydronephrosis and pyelonephritis (Fig. 35) was observed in 19 U8
Fig. 26. C^H mouse with an extensive cervical carcinoma filling
the entire pelvis.
Fig. 27. Epidermoid carcinoma of papillary exophytic type near
the external os at the portio vaginalis. X
Fig. 28. Squamous carcinoma of the uterine cervix. Tissue section.
Note the focus of neoplastic transformation located high on
the lateral mucosa of the cervix. X 7> \ m s ' v j • i •?>,^?<^~ '' ,•; ijf* > m4 •Sx*.i-!t!iiMteeia^ m i m , ;i'i \.
>f‘*r
,t ' . 7, > t o* •i;Vv*i* 'U A w i i t t ’M)*' :^ 3 ? V J ^ v" V n i / w
W j-oV. 9 o
Fig. 29. A low poxrer microphotograph showing extensive areas of
carcinoma in situ involving vaginal and cervical epithelium.
X 10.9
Fig. 30. A low power microphotograph showing a large necrotic tumor
surrounding the uterine cervix and obliteration of the
vaginal fornices. X 8
Fig. 31* A low power microphotograph showing a massive tumor with
extensive necrosis. X 12
52
Fig. 32. Pelvic lymph node containing metastatic squamous ceil
carcinoma. X 135
Fig. 33. Rectum showing tumor invasion of the muscularis and sub
mucosa. X 80
Fig. 3U. Lung containing numerous nodules of metastatic squamous
cell carcinoma. X 130
Fig. 35. Ureteral obstruction with hydronephrosis, hydroureter,
and distension of the urinary bladder. 53
W m ' M ^ m
& S & 3 0 , Sh
animals. These results are summarized in Table £.
TABLE 5
SITES OF TUMOR EXTENSION AND METASTASES
Site No. of Mice
Lymph n o d e s ...... 32
Lateral pelvic wa l l ...... 28
Urinary b l a d d e r ...... 22
U r e t e r s ...... 19
R e c t u m ...... 1 $
K i d n e y ...... lit
L i v e r ...... 3
L u n g ...... 2
In 32 animals the distribution of metastases observed in
various lymph nodes draining the pelvic region were as follows. The paired lumbar lymph nodes were the most frequently involved, metastatic
carcinoma was present in 18 mice. The caudal lymph node lying beneath
the posterior end of the rectum was involved in 7 animals. It is of
interest that in each instance the rectum was also invaded by carcinoma,
Since the caudal lymph node drains the lymphatics of the tail and rec
tum, metastases in this node probably are the result of extension of tumor to the rectum. The renal lymph nodes were involved in Jj. animals in which large tumors were present which filled the pelvis. The
sciatic lymph nodes showed the presence of tumor metastases i n '3 anim als. These results are summarized in Table 6. TABLE 6
METASTASES TO PELVIC LIMPH NODES
Lymph Nodes No. of Mice
Lumbar 18
Caudal 7
Renal k
Sciatic 3
Total 32
Hlstopathology
Both intra-epithelial, or non-invasive, and invasive carcinomas were induced as shown in Table 7. In the majority of instances, intra- epithelial carcinoma in the mouse was characterized by the following morphologic changes: 1. Intra-epithelial nodules of cellular prolifer ation suggestive of independent growth (Fig. 36), and 2. squamous cells showing nuclear enlargement, hyperchromatism, cell crowding, and increa sed mitoses (Fig. 37). Microscopically, non-invasive carcinomas of two distinct morphologic types were distinguishable: diffuse and focal.
Sixty-three animals (91;.1 per cent) showed intra-epithelial lesions of the diffuse type (Fig, 38) with a marked diffuse papillomatous prolif eration of basal and parabasal cells within widened epithelial boundaries and with intact basement membranes. In certain areas the neoplastic transformation was not accompanied by papillomatous basal cell overgrowth. The pattern more closely resembled that commonly ob served in intra-epithelial carcinoma in the human (Fig. 39). The cells were predominantly of intermediate size although considerable cytologic variation was observed. In addition there were nuclear enlargement, I
%
Fig. 36. Carcinoma in situ showing numerous intraepithelial nodules
of increased basal cell growth. X 310
Fig. 37. Carcinoma in situ. Malignant squamous cells showing nu
clear enlargement, hyperchromatism, cell crowding and
increased mitoses. X U30
Fig. 38. Carcinoma in situ showing a diffuse papillomatous pro
liferation of basal and parabasal cells with cellular
crowding, hyperchromatism and numerous mitoses. X 220
Fig. 39. Carcinoma in situ. A variant resembling somewhat the
pattern of intraepithelial carcinoma commonly observed in
the human. X 290 -'?» ^ b ..r ■■ JKBSi
* \ 0 £
M
' * W j t ©
«G
*m r (v v yfri sv.vyaffi'»'' ?*\*zxi9s*}
tmjK&vOi ^^pr^rivFvSwVi*'1^® :^>rr7. :* •**' ti'v;i./v " • I •7*’ * ti>'‘, ,/i
f : 0 M »
■
* 1 I TABLE 7
SUMMARY OF EXPERIMENTAL CARCINOMAS INDUCED
No. of Noninvasive Invasive Total Group Mice Carcinoma Percentage Carcinoma Percentage Percentage
1 60 19 31.6 12 20.0 31.6
2 80 21 26.2 13 18.8 U3.0
3 86 17 19.3 27 31.7 31.2
h 13 3 20.0 8 3 3.3 73.3
3 20 7 33.0 10 30.0 83.0
261 67 23.6 72 27.6 33.2
vn CO hyperchromasia, cell crowding, and numerous mitotic figures, many of which were atypical. In the more advanced cases of intra-epithelial carcinoma, there were accumulations of numerous inflammatory cells in the submucosa. These were predominantly lymphocytes, monocytes with occasional plasma cells. Considerable disturbance and, in some instances, complete absence of normal stratification and differenti ation of superficial epithelial cells were observed. It was not uncommon to find keratinization at the surface of intra-epithelial
lesions. In some instances, this was considerable, forming small surface papillomata (Fig. 1;0). In Ii. animals (5.9 per cent) intra epithelial lesions of the focal type were observed in low, atrophic appearing epithelium with discrete areas of papillomatous basal cell growth containing small groups of malignant cells (Fig. l;l). The most outstanding feature of this lesion was the absence of florid epithelial cell overgrowth and crowding so characteristic of the diffuse type. No evidence of transition from focal to diffuse lesions was seen. Multi- centric areas of non-invasive malignant transformation were observed in
21 animals.
Invasive carcinomas showed a disruption of the basement mem brane by malignant squamous cells invading the underlying stroma. This varied from thin strands of malignant cells (Fig. 1+2) indicative of early invasion to wide sheets of neoplastic cells seen in the more ad vanced lesions (Fig. ii3). Invasive lesions varied in size from wide fields of malignant transformation to small microscopic lesions (micro- carcinomas) which were largelj'- limited to the cervical mucosa but showed morphologic evidence of early invasion as shown in Figure 14;. 60
Fig. 1*0. Carcinoma in situ, showing extensive keratinization forming
a small surface papilloma. X 120
Fig. 1*1. Atrophic type of carcinoma in situ showing small foci of
malignant basal cells. Note absence of keratinization.
X 323
Fig. 1*2. Early invasive carcinoma showing thin strands of neoplastic
cells infiltrating beyond the basement membrane. Note the
florid inflammatory reaction in the stroma. X 160
Fig. 1*3. Invasive carcinoma of the uterine cervix showing disruption
of the basement membrane by sheets of neoplastic cells. X 100 61
f ||
J 62
Some of these early lesions were exophytic and showed extensive surface keratinization suggestive of originating from intra-epithelial lesions of similar type described above. In 18 animals various stages of pro gressive malignant transformation were present simultaneously. Multi- centric lesions were demonstrable in 33 animals.
Although the majority of invasive tumors induced were of the squamous cell type, marked cytologic variation was observed. On this basis four distinctly different cytologic variants could be differenti ated: well differentiated (spinal cell) type, moderately well differ entiated (transitional cell) type, undifferentiated (spindle cell) type, and muco-epidermoid type. The incidence of these lesions is summarized in Table 8.
TABLE 8
HISTOLOGIC TYPES OF CARCINOMA
Vagina Cervix Both Total Percentage
Noninvasive carcinoma Diffuse type IS? 2f? 23 63 9k.l Focal type 1 2 1 1* $.9 -37 100.0
Invasive carcinoma Spinal cell type h 7 10 21 29.2 Transitional cell type 6 17 13 36 50.0 Spindle cell type 1 8 1U 19.U Muco-epidermoid type 1 1 1.1* 72 100.0
The cytologic characteristics of spinal cells as defined by
MARTZLOFF (1923) consist of polyhedral cells with ample eosinophilic cytoplasm and well defined cell membranes (Fig. 1|9). The nuclei are 63
Fig. UU. Microcarcinoma largely limited to the cervical mucosa and
showing limited early invasion. X 230
Fig. US. Well differentiated epidermoid carcinoma showing cells of
the spinal type with small nuclei, ample cytoplasm, and
well defined cell membranes. X 260
Fig. U6. Moderately well differentiated epidermoid carcinoma of
transitional cell type showing a decrease in cytoplasm and
numerous mitotic figures. X 260 6k
“'4- '•.. ' S S ' a K '• • *» S&m%)\d?m<%&.
«■..». j. ,.i%V«i_'f«fn®
r/i..' j^ w k 'T*. 65 moderately hyperchromatic and occasionally show macronucleoli. In addition, there were well developed tonofibrils in 73 per cent of the epidermoid tumors of the spinal cell type. There was also a tendency for the cells to form wide sheets. Keratinization with epithelial pearl formation was prominent, while mitotic figures were scarce.
These well differentiated cells closely resembled the normal cells of the upper stratum spinosum of cervical epithelium in the mouse during estrus. Tumors consisting predominantly of well differentiated spinal cells were found in 21 animals (29.2 per cent).
Transitional cells were more rounded and showed a decrease in the amount of cytoplasm, with large hyperchromatic nuclei, frequent macronucleoli, and mitotic figures (Fig. I4.6). Tonofibrils were demon strable in 22 per cent of the squamous cell carcinomas of the trans itional cell type. Keratinization and epithelial pearl formation were seen only occasionally. Foci of necrosis were more frequent than in the spinal cell group. These cells resemble those found in the lower portions of the stratum spinosum of cervical epithelium during estrus.
In three instances small foci of malignant squamous cells which appeared to be secreting mucus were found interspersed among sheets of tumor cells. The mucus was PAS-positive and not salivary-araylase labile.
Tumors of the moderately well differentiated transitional cell variety were observed in 36 animals (50 per cent).
Spindle cells appeared as elongated cells with scant, pale, eosinophilic cytoplasm, large hyperchromatic nuclei, and often multiple macronucleoli (Fig. 1;7). Numerous mitotic figures were present also.
Keratinization and epithelial pearl formation were absent. Although 66
Fig. J4.7. Poorly differentiated epidermoid carcinoma of the spindle
cell variety. X 210
Fig. i|8. Muco-epidermoid carcinoma showing nests of moderately well
differentiated squamous cells with transition to mucus-
containing cells. Several acinus-like structures may be
seen. X 3^0 67
f i m 68 tonofibrils were absent, these cells resembled the spindle-shaped hyperchromatic cells contained in the germinal layer of cervical epith elium. Necrosis with secondary bacterial infection was a common find ing. Malignant lesions consisting primarily of undifferentiated spindle
shaped cells were present in ll* animals (19.U per cent).
In a single instance (l.ij. per cent) an invasive muco-epidermoid tumor was observed. Morphologically, it was characterized by nests of moderately well differentiated squamous cells among which were inter spersed many acinus-like structures. These were lined by squamous cells which contained varying amounts of pale blue staining cytoplasmic material and showed transition to plump, rounded cells filled with mucus (Fig. 1;8) which was PAS-positive and resistant to salivary- amylase digestion. Epithelial pearl formation was absent.
DISCUSSION
Experimentally induced carcinoma of the uterine cervix and vagina in C^H mice showed a striking biologic and pathologic resemblance to human cervical cancer. Invasive lesions extended locally into sur rounding tissues, metastasized to pelvic lymph nodes, produced obstruc tive uropathy, and in several instances metastasized widely. The experimental lesions differed from their human counterparts largely in two respects. The majority of experimental lesions in the mouse appear ed to originate at the portio vaginalis in contrast to the spontaneous carcinomas in women which originate at the squamo-columnar junction.
Although this point is of interest, it must be remembered that the chemical induction of cancer is highly artificial and that a direct comparison with naturally occurring tumors is difficult if not impossible. The complete absence of lesions at the squamocolumnar
junction in the mouse is no doubt due to its considerable distance from
the portio vaginalis. The endocervix in the mouse is lined by squamous
epithelium with epithelial transition occurring at the bifurcation of
the uterus. Thus neither vaginal nor low endocervical application of
carcinogen would reach this area. The shortened latent period and
higher yield of tumors following the endocervical application of carcin
ogen-impregnated strings may be the result of higher concentrations of
carcinogen in continuous contact with the epithelium. Another point
of difference between the experimental and naturally occurring lesions
is the histologic configuration of most of the intra-epithelial lesions.
Experimental non-invasive carcinoma showed considerable surface kerat
inization in contradistinction to the non-keratinized lesions described
in the human.
Non-invasive and early invasive lesions showed a tendency to
separate from the underlying stroma in a manner similar to that reported
in human cases of non-invasive cancer by REAGAN et a l ., (1955). This
may be due to the progressive loss of tonofibrils during carcinogenesis
as shown by VON ALBERTINI (1953). However, the demonstration of tono
fibrils in well differentiated tumors suggests a relation between cellu
lar differentiation and tonofibril formation.
The presence of small foci of malignant basal cells localized
solely in the stratum germinativum in the focal type of non-invasive
carcinoma strongly suggests that at least in some instances the neoplas
tic change may originate there. Furthermore, the multiple foci and
extensive diffuse fields of cervical mucosa which showed malignant 70 changes are evidence in favor of a multicentric pattern of origin.
This finding is in accord with the view held by WILLIS (19UU) that the carcinogenic stimulus acts on large groups of cells and extensive areas of epithelium rather than on a single cell.
The cytologic variants of squamous cell carcinoma, as MARTZLOFF
(1923) pointed out, closely resemble the cells of the various layers in mature squamous epithelium and represent varying degrees of cellular differentiation.
It is interesting that a muco-epidermoid tumor and mucus-secre ting foci should be observed in tumors arising from squamous epithelium devoid of mucus-secreting cells. However, since it has been observed that squamous epithelium of the genital tract in mice has the potential of mucus production and is capable of mucification under a variety of physiologic and hormonal stimuli (HISAW et al., 1927; MEYER AMD ALLEN,
1932, 1933), mucus production in malignant squamous cells must be inter preted as a metaplastic reaction. Similar tumors have been induced by intravaginal painting with crude tobacco tar by K0PR0WSKA AND BOGACZ
(1959).
SUMMARY
Attempts at the experimental induction of carcinoma of the uterine cervix in 261 C^H mice resulted in 139 malignant lesions (53*2 per cent). Non-invasive carcinoma was observed in 67 animals, while
72 mice developed invasive carcinoma. Invasive lesions showed intra- pelvic extension, metastases, and obstructive uropathy in a manner similar to that observed in analogous tumors in man.
Histopathologic studies showed the majority of lesions to 71 consist of moderately well differentiated squamous cells. In li* animals (19.U per cent) poorly differentiated squamous cell carcinoma was observed, and in a single animal a muco-epidermoid carcinoma was seen. MITOSIS IN CERVICAL EPITHELIUM DURING EXPERIMENTAL
INFLAMMATION AND CARCINOGENESIS
Since the early descriptions of abnormal mitosis in malignant cells by VON HANSEMANN (1890) and PIANESE (1896), numerous detailed cytological studies of this phenomenon have appeared. The majority of these investigations deal with the classification of abnormal mitotic types and the possible mechanism involved (KOLLER, 19^7)• Although abnormal mitoses were also studied during inflammation by KEMP (1930) and more recently in carcinoma in situ by PARMENTIER AND DUSTIN (19>l), and HAMPERL et ai. (195U), no information is available on mitotic ab normalities during the pathogenesis of cancer. This section deals with a comparative study of the incidence and types of mitotic abnormalities in epithelial cells of the uterine cervix in C^H mice during experimen tal inflammation and the induction of cervical carcinoma.
MATERIALS AND METHODS
In total, 30 C^H female mice were selected for the present study. These consisted of four groups of ten animals each, showing the following lesions of the uterine cervix: inflammation, dysplasia, car cinoma in situ, and invasive cancer. A fifth group consisting of ten normal C^H mice was used as controls. Inflammation was produced by painting mice intravaginally with a J4 per cent solution of croton oil in acetone. Epithelial dysplasia and cervical carcinoma were produced by painting the animals intravaginally with a 1 per cent solution of
3,U benzpyrene in acetone. The technic of painting is similar to that
72 73
described in detail earlier. Smears were prepared previous to each painting procedure and were stained by the method of PAPANICOLAOU
(I9l|.2). As soon as a definite diagnosis of inflammation, dysplasia, or
carcinoma was made, the animals were sacrificed. The uterus and vagina were removed, fixed in 10 per cent buffered formalin, and serial sec
tions (5 and 10 p.) were prepared for microscopic examination. In addi
tion to the routine hematoxylin and eosin stain, selected sections were
subjected to the Feulgen reaction (FEULGEN AND ROSSENBECK, 192lj.) for the demonstration of chromosomes. Tissue sections were hydrolyzed in
N HCL at 60°C. for 12 minutes, followed by staining in leukobasic fuchsin for li? hours. In order to more clearly demonstrate the spindle mechanism, sections were stained by the technique of VAN GIESON (1889).
Serial sections of uterine cervix were examined with an oil- immersion objective at a magnification of 970 diameters. The stage and type of mitosis were determined in 100 dividing epithelial cells in each animal. To insure a more representative sampling of mitoses, num erous regions of the cervical epithelium were examined. Morphologic criteria characteristic of the various stages of mitosis were followed closely. Cells in which the mitotic stage was not readily ascertainable were not considered in this study. To determine the relative duration of prophase and metaphase, the prophase index was calculated for each animal by dividing the number of metaphases by the number of prophases according to the method of TIMONEN AND THERMAN (19^0). These data were subjected to statistical analysis by the Statistics Laboratory of the Ohio State University. 7k
RESULTS
The mitotic aberrations found in ^000 dividing cells in normal cervical epithelium, during inflammation, and in the various stages of carcinogenesis are shown in Table 9. A marked difference in the number of abnormal mitoses was found between inflammatory and precancerous cervical lesions and cervical cancer. A combination of the data into non-malignant and malignant groups is shown in Table 10. Among the non-malignant cells showing mitotic abnormality, the proportion was i*.!?!? with a standard deviation of 0.U7 per cent. Among the malignant cells, the proportion showing mitotic abnormality was 31.1^ with a standard deviation of l.Olj. per cent. The two proportions were signifi cantly different, with a probability of error less than .001.
Normal Epithelial Cells
No instances of abnormal mitosis, such as, multipolarity, chromosomal lagging, hollow spindles, asymmetrical divisions, or polar, chromosomes were encountered in a study of 1000 dividing cells in the cervical epithelium of normal C^H mice. The number of cells in prophase was roughly equal to those in metaphase. The prophase index varied from
.73 to 1.23 with a mean value of .96.
Inflammation
Twenty-eight cells with evidence of abnormal mitosis were found in 1000 dividing epithelial cells in inflammatory lesions of the uterine cervix. Twenty-one of these showed lagging chromosomes with numerous examples of bridge formation (Fig, i|.9). Seven cells showing asymmetri cal mitosis were also encountered. The prophase index varied from 0.82 TABLE 9
INCIDENCE AND TYPES OF NORMAL AND ABNORMAL MITOSES IN 5,000 DIVIDING CELLS OF THE UTERINE CERVIX
(Cervical Lesions Normal Inflam Carcinoma Invasive Mitoses Epithelium mation Dysplasia in situ Carcinoma
Multipolar 0 0 0 29 52
Hollow metaphase 0 0 7 23 55
Asymmetric division 0 7 U kl 6i
Lagging chromosomes 0 21 52 112 137
Polar chromosomes 0 0 0 58 h9
Normal 1000 972 937 731 62+6
Totals 1000 1000 1000 1000 1000 76
to 1.6U with a mean value of 1.U2
Dysplasia
Sixty-three cells showing abnormal mitosis were observed in 1000
dividing cells in cervical dysplasia. Fifty-two examples of lagging
chromosomes, seven cells showing hollow metaphase plates (ring forms),
and four instances of asymmetric mitosis were found (Fig. £0). The pro
phase index varied from 1.19 to 2.03, with a mean value of 1.77.
TABLE 10
DISTRIBUTION OF NORMAL AND ABNORMAL MITOSES IN NON-MALIGNANT AND MALIGNANT LESIONS
Non-Malignant Malignant Total
Mitotic Abnormality 91 623 71k
Normal 1909 1377 3286
Total 2000 2000 Uooo
Carcinoma in Situ
Two hundred and sixty-nine cells showing abnormal mitosis were found in 1000 dividing cells in carcinoma in situ of the uterine cervix.
One hundred and twelve cells showed, lagging chromosomes, 29 showed multi
polar mitoses (Figs. 9l, 92, 93)., 23 mitotic figures showed hollow raetaphase plates (Figs. 9U, 99)> four of which were incomplete ring
forms (Fig. 96), and If/ instances of asymmetrical mitosis were found.
In addition, 98 cells showed the presence of chromosomes distributed
about both centrosomes (polar chromosomes) (Fig. 97)* The prophase
index varied from 1.89 to 9 .6 with a mean value of 7.3 * Invasive Carcinoma
Three hundred and fifty-four abnormal mitoses were encountered
in 1000 dividing cells in invasive carcinoma of the uterine cervix. One
hundred and thirty-seven cells showed lagging chromosomes, 55 cells
showed hollow metaphase plates, eleven of which showed incomplete ring
formation, 52 showed multipolar mitoses, k7 showed asymmetric mitosis,
and h9 cells showing polar chromosomes were found (Fig. 58). In several
instances polar chromosomes appeared to be composed of two distinct
chromosomal masses (Fig. 59). Occasional cancer cells showed widely
scattered chromosomes, the so-called "colchicine effect" (Fig. 60).
The prophase index varied from 1.79 to 2k, with a mean value of 11.2.
A progressive decrease in the relative frequency of 'the pro phase stage of mitosis was evident during the pathogenesis of cervical
cancer. The relative frequencies of the various mitotic stages in normal animals and in i| types of cervical lesions are summarized in
Chart 3. An increase in the frequencies of metaphases and a decrease in the frequencies of prophases occurred during the various stages of carcinogenesis. The frequencies of anaphase and telophase did not vary
appreciably. A direct relation appears to exist between the number of
abnormal mitoses and the prophase index, as shown in Chart b, where the logarithm of the prophase index was plotted against the number of abnor mal mitoses. An increase in both the prophase index and number of abnormal mitoses occurred in malignant lesions, which sharply differ entiated them from non-malignant lesions.
Although the prophase indices of prophase index (PI) in normal mice approximated a ratio of 1, while in inflammation, dysplasia, CHART 3
AVERAGE FREQUENCY OF MITOTIC STAGES AND ABNORMAL MITOSES IN VARIOUS CERVICAL LESIONS
■ I LAGGING P - PROPHASE 80 m MULTIPOLAR M - METAPHASE □ HOLLOW METAPHASE A - ANAPHASE 70-i H ASYMMETRICAL T - TELOPHASE EH3 POLAR CHROMOSOMES 60
50
4 0
30-|
20
HR 10 | 0 P M A T P M A T P M A T P M A T P M A T NORMAL INFLAMM. DYSPLASIA CA. IN SITU INVASIVE CA.
The average frequency of mitotic stages in normal animals and the four types of cervical lesions prophase index.prophase Scattergram showing a correlation between abnormal mitosis and abnormalthe mitosis acorrelation showingbetween Scattergram
NUMBER OF ABNORMAL MITOSES 20 60 30- 0 5 40+ 40+
9 00 0 . 06 . 08 . 10 . 12 . 1.4 1.3 1.2 l.l 1.0 0.9 0.8 0.7 0.6 0.5 4 0 3 0 2 0 01 0 0 ~9 s T a SITU IN CARCINOMA ■ + INFLAMMATION + ODYSPLASIA NORMAL . IVSV CANCER INVASIVE O O POHS INDEX PROPHASE OF LOG
CHARTL A ■■
80
Figures - 6 0'
Fig. it9. Bipolar mitosis showing lagging chromosomes.
Fig. 50. Bipolar anaphase showing asymmetrical division of chromosomes.
Fig. 5l. Tripolar metaphase.
Fig. 52. Tripolar anaphase.
Fig. 53• Tetrapolar metaphase showing numerous spindles (van Gieson
stain).
Fig. 5U. Hollow metaphase plate.
Fig. 55. Micronucleus showing a hollow metaphase mitosis.
Fig. 56. Incomplete hollow metaphase plate
Fig. 57. Metaphase showing polar chromosomes.
Fig. 58. Metaphase showing polar chromosomes. Two chromosome masses
are present at the upper pole (Feulgen reaction).
Fig. 59. Metaphase showing polar migration of a chromosomal mass.
Fig. 60. Malignant squamous cell in division showing the so-called
"colchicine effect." Note the widely scattered and con
tracted chromosomes.
-''Taken at a magnification of 1300 diameters. 81 82
carcinoma in situ, and invasive carcinoma the PI values were always
greater than 1 (Table 11). Since the PI in normal mice tended to
approach 1, PI data from all the groups were subjected to statistical
analysis by testing at various significance levels the hypothesis that
the prophase index = 1 (i.e., p = J, where p = the conditional prob
ability that a cell is in metaphase) (Table 12). Critical PI values were constructed where the critical PI value was that value above which
further increases were of diagnostic significance at various levels of probability. These data are summarized in Table 13. The results
showed that in a sample as small as 6U cells, composed of cells either
in prophase or metaphase, a critical PI value of 2.£6 or greater was
diagnostic of malignancy, with a probability of error less than .0005.
The number of animals with PI values greater than critical PI values
at various levels of significance are shown in Table lU. At a low level of significance (P <".05), seven of ten animals with dysplasia had significant PI values; however, when the level of significance was increased to a probability of error less than .0005, none of these values was significant. On the other hand, eighteen of twenty animals with either carcinoma in situ or invasive carcinoma showed significantly increased PI values with a probability of error less than .0005.
DISCUSSION
Cytological analysis of mitotic abnormalities in cervical epith
elial cells during inflammation and the pathogenesis of 3,U-benzpyrene- induced cervical carcinoma in C^H mice revealed a statistically signifi
cant increase in the number of abnormal mitoses as the malignant state was approached. Lagging chromosomes constituted the most common TABLE 11
PROPHASE - METAPHASE VALUES AND PROPHASE INDICES
Group I - Normal Group II - Inflam. Group III - Efysplasia P M Total P.I. P M Total P.I. PM Total P.I.
1*7 33 80 0.70 33 5o- 83 1.52 31 56 87 1.81 h9 1*8 97 0.98 ia 1*0 81 0.98 27 55 82 2.0i* kO 1*? 85 1.12 31* 5o 81* 1.1*7 30 60 90 2.00 39 37 76 0.95 37 52 81* 1.62 31 59 90 1.90 la 37 78 0.90 30 1*8 78 1.60 1*0 5i 91 1.28 hh 38 82 0.86 31 5i 82 1.61* 1*2 5o 92 1.19 1*0 36 76 0.90 35 51* 89 1.51* 30 61 91 2.03 1*0 la 81 1.02 1*0 33 73 0.82 33 53 86 1.61 39 35 71* 0.90 33 50 83 1.52 32 60 92 1.88 1*1 50 91 1.22 35 57 92 1.63 29 58 87 2.00
Group I? - Ca in Situ Group V - Invasive Carcinoma P M Total P.I. P M Total P.I. 13 65 78 5.00 1* 80 81* 20.00 9 72 81 8.00 17 60 77 3.53 9 70 79 7.78 12 61+ 76 5.33 7 65 72 9.29 30 51* 81* 1.80 33 60 93 1.82 7 83 90 11.86 8 71 79 8.88 9 75 81* 8.33 9 80 89 8.89 5 60 65 12.00 8 75 83 9.38 12 58 70 1+.83 7 65 72 9.29 3 72 75 21*. 00
7 61 68 8.71 ' 1* 87 91 21.75
CD TABLE 12
STATISTICAL SIGNIFICANCE*
n .95 .99______.999 11 1 10 0 11 0 11 12 2 10 0 11 0 12 13 2 11 1 12 0 13 1U 2 12 1 13 0 li; 15 3 12 2 13 1 lit
16 3 13 2 Ik 1 i5 17 k 13 2 15 1 16 18 k 1U 3 15 1 17 19 k 15 3 16 2 17 20 5 15 3 17 2 18
21 5 16 k 17 2 19 22 5 17 k 18 3 19 23 5 17 k 19 3 20 2k 6 18 5 19 3 21 25 •7 18 t; 20 U 21
26 7 19 6 20 U 22 27 7 20 6 21 h 23 28 8 20 6 22 5 23 29 8 21 7 22 5 2k 30 9 21 7 23 5 25
31 9 22 7 2k 6 25 32 9 23 8 2h 6 26 33 10 23 8 25 6 27 3k 10 2k 9 25 7 27 35 11 2h 9 26 7 28 CD 4=" — ......
TABLE 12— Continued
n .95 • 99 •999
36 11 25 9 27 7 29 37 12 25 10 27 8 29 38 12 26 10 28 8 30 39 12 27 11 28 8 31 l+o 13 27 11 29 9 31 ill 13 28 11 30 9 32 1+2 11+ 28 12 30 10 32 1+3 11+ 29 12 31 10 33 1+1+ 15 29 13 31 10 31+ 1+5 15 30 13 32 11 31+
1+6 15 31 13 33 11 35 1+7 16 31 11+ 33 11 36 1+8 16 32 11+ 31+ 12 36 1+9 17 32 15 31+ 12 37 50 17 33 15 35 13 37
52 18 31+ 16 36 13 39 51+ 19 35 17 37 11+ 1+0 56 20 36 17 39 15 la 58 21 37 18 1+0 16 1+2 60 21 39 19 ia 16 1+1+
62 22 1+0 20 1+2 17 1+5 61+ 23 ia 21 1+3 18 1+6 66 21+ 1+2 22 1+1+ 19 1+7 68 25 1+3 22 1+6 20 1+8 70 26 1+1+ 23 1+7 20 50 TABLE 12— Continued
n .95 .99 .999 72 27 k9 2k U8 21 5l 7ii 28 U6 25 k9 22 52 76 28 U8 26 50 23 53 78 29 k9 27 5l 2U 9k 80 30 50 28 52 2U 56
82 31 51 28 9k 25 57 Qk 32 52 29 >> 26 58 86 33 53 30 56 27 59 88 3k 9k 31 57 28 60 90 35 get 32 58 29 61
92 36 56 33 59 29 63 9k 37 57 3U 60 30 6k 96 37 59 3ii 62 31 65 98 38 60 35 63 32 66 100 39 61 36 6U 33 67
* X is the number of successes in n trials, p = 1/2. The tabulated numbers a and b are determined so that P( a X b ) .95, .99, .999. TABLE 13
CRITICAL PI VALUES*
n 2 1/256 1/256 .03$
11 10.0 12 3.o 13 3.30 12.00 Ik 6.00 13.00 13 U.oo 6.30 1U.00
16 U.33 7.00 13.00 17 3.29 7.30 16.00 18 3.90 3.oo 17.00 19 3.73 9.33 8.30 20 3.0 3.67 9.00
21 3.20 U.23 9.30 22 3.ko i*.5o 6.33 23 3.1*0 i*.73 6.67 2h 3.00 3.80 7.00 29 2.37 U.00 3.23
26 2.71 3.33 3.3o 27 2.86 3.30 3.73 28 2.30 3.67 1*.60 29 2.62 3.1k 2*.80 30 2.33 3.29 3.oo
31 2.UU 3.1*3 1*.17 32 2.36 3.00 k.33 33 2.30 3.12 i*.3o 3k 2.1*0 2.78 3.86 39 2.18 2.89 i*.00 TABLE 13— Continued
n 2 1/256 1/2 % .05$ 36 2.21 3.00 1+.11+ 37 2.08 2.70 3.62 38 2.17 2.80 3.75 39 2.25 2.55 3.88 l+o 2.08 2.61+ 3.1+1+
ia 2.15 2.73 3.56 1+2 2.00 2.50 3.20 U3 2.07 2.58 3.30 1+1+ 1.93 2.38 3.1+0 k$ 2.00 2.1+6 3.09
1+6 2.07 2.51+ 3.18 hi 1.9U 2.36 3.27 1+8 2.00 2.1+3 3.00 h9 1.88 2.21 3.08 90 1.91+ 2.33 2.85
52 1.89 2.25 3.00 51+ 1.81+ 2.18 2.86 56 1.80 2.29 2.73 58 1.76 2.22 2.62 60 1.86 2.16 2.75
62 1.82 2.10 2.65 -X-X-6I4 1.78 2.05 2.56 66 1.75 2.00 2.1+7 68 1.72 2.09 2.1+0 70 1.69 2.Oil 2.50
CO CO TABLE 13— Continued
n 2 1/2# 1/2# .05# 72 1.67 2.00 2.1*3 7l* 1.61* 1.96 2.36 76 1.71 1.92 2.30 78 1.69 1.89 2.25 80 1.67 1.86 2.33
82 1.61* 1.93 2.28 81* 1.62 1.90 2.23 86 1.61 1.87 2.18 88 1.59 1.81* 2.11* 90 1.57 1.81 2.10
92 1.56 1.79 2.17 91* 1.51* 1.76 2.13 96 1.59 1.82 2.10 98 1.58 1.80 2.06 100 1.56 1.78 2.03
-Critical values of PI in samples haying a total of n metaphase and prophase cells for different significance levels.
■^Example cited in text. 90
TABLE lit
STATISTICAL SIGNIFICANCE OF THE PROPHASE INDEX IN VARIOUS CERVICAL LESIONS
No. Animals With Prophase Indices (Pi) Greater Than Critical PI Values at Lesions Specified Significance Levels
P <.0£ P <.009 P <.ooo£
Normal (controls) 0 0 0
Inflammation 0 0 0
E^splasia 7 6 0
Carcinoma in situ 10 10 9
Invasive Carcinoma 10 9 9 91 abnormality found and were present in inflammation, dysplasia, carcin oma in situ and invasive carcinoma. Lagging has been attributed by
DARLINGTON,(19U2; and ROLLER,(19^3) to an increased "stickiness" of chromosomes, which delays and modifies chromosomal alignment and separ ation.
Although numerous examples of asymmetrical division were observed, direct cytological evidence of nondisjunction was obtained only in a few instances of extremely unequal division. The occasional presence of micrometaphase figures suggests that some of the asymmetri cal mitoses result in cells which may be viable and capable of further division.
Multipolar mitoses constituted the most frequent multipolar anomaly, although both quadripolar and pentapolar forms were also present. MOORHEAD AND HSU (1996) have shown by cinematographic studies of HeEa cells in tissue culture that multipolar spindle formation is the result of incomplete fusion of two or more separate spindles.
TIMONEN AND THEKMAN (1990) explain multipolarity on the basis of desyn chronization of the various processes that constitute the mitotic cycle.
More specifically, they attribute multipolarity to a continued division of centrosomes, while the chromosomes divide only once.
Polar chromosomes identical with the "metaphase a trois groupes" described in human cases of cervical cancer by PARMENTIER AND DUSTIN
(1991) were found in experimental cervical carcinoma in mice. Contrary to their findings, this peculiar mitotic anomaly was found with equal frequency both in non-invasive and invasive carcinoma. VON MOELLENDORFF
(1939) reported similar polar chromosomal aggregations in tissue cultures 92
of fibroblasts following treatment with either polycyclic carcinogens, estrogenic, or androgenic hormones. Although the exact mechanism of polar chromosome formation is not known, experimental studies by DUSTIN
AND PARMENTIER (1953) have shown that it is not related to chromosome rupture and is probably associated with disturbances of the spindle mechanism. The exclusive presence of polar chromosomes in dividing malignant epithelial cells suggests a possible relationship between this mitotic anomaly and the malignant state.
The PI values found in normal and neoplastic cervical tissues of the mouse are of the same magnitude and range as those observed by
TIMONEN AND THERMAN (1950) in their study of human cervical cancer cells.
In the mouse, PI values greater than 2.17 were found in experimentally induced carcinoma of the uterine cervix. Then these data were subjected to rigid statistical analysis, PI values of this magnitude were found to be highly significant. This critical PI value is higher than that of
1.5 reported for human cervical cancer by TIMONEN AND THERMAN (1950) and is no doubt due to differences in the duration of the mitotic cycle of the mouse.
The increase in the PI values during the development of cervical carcinoma suggests a correlation between a shortening of the prophase and/or a prolongation of the metaphase and malignancy. However, owing to the fact that the present data were obtained from fixed tissues, it was impossible to determine exactly where the change occurred in the mitotic cycle. The observations of LAMBERT (1913) and LEWIS AND LEWIS
(1932) that anomalous mitoses require more time for completion than do normal ones may explain the direct relation between the frequency of 93
abnormal mitoses and increased PI values.
In view of recent reports by HSU (195U) and MOORHEAD AND HSU
(1956) that early prophase is not readily recognizable in fixed and
stained cells, prophase indices determined from fixed tissues may be in
error. HSU (195U) also showed that a discrepancy exists between the
prophase coefficients (prophases divided by metaphase) obtained from a
study of fixed cells and those obtained from living cells. His results
serve to emphasize not only the difficulties encountered when a dynamic
process, such as cell division, is studied by static methods utilizing
dead cells, but also the need for a re-evaluation of mitotic phenomena
by cinematographic studies of living cells.
Previous observations (VON HAAM AND SCARPELLI, 1955) indicated that, although both qualitative and quantitative cytological changes
occurred during the malignant transformation of squamous cells, they
were gradual and roughly paralleled the histological alterations seen
in the various stages of carcinogenesis. The present studies show that disturbances in the metaphase/prophase ratio during the pathogene
sis of cancer is not gradual and sharply differentiates malignant from non-malignant lesions. The similarity between carcinoma in situ and invasive carcinoma with reference to mitotic abnormalities and prophase indices lends further support to the concept that both lesions are
closely related and that carcinoma in situ probably represents a malig nant lesion.
It should be emphasized, however, that this data are applicable
only for experimental carcinoma in situ and invasive carcinoma in the mouse cervix. 9k
SUMMARY
Although a few mitotic abnormalities were found in inflammatory
cervical lesions, ever increasing numbers were observed during the
pathogenesis of experimentally induced cervical cancer.
Lagging chromosomes constituted the most frequent mitotic abnor mality. Polar chromosomes were found only in malignant cervical lesions.
Changes in the frequency and/or length of prophase and metaphase were observed which sharply differentiated benign from malignant lesions
during the pathogenesis of cancer.
The critical prophase indices which were diagnostic of malig nancy ranged from 2.15? to 2.£6 or above, depending on the number of either prophase or metaphase mitoses counted. QUANTITATIVE ESTIMATIONS OF THE DEOXYRIBONUCLEIC ACID
CONTENT OF CERVICAL EPITHELIAL CELLS DURING
ESTRUS, INFLAMMATION AND CARCINOGENESIS
Studies on the deoxyribonucleic acid (DNA) content of normal
non-dividing plant and animal nuclei have shown that it is quite stable
in interphase, showing little variation from cell to cell (BOIVIN et al.,
I9U83 MIRSKI AND RIS, 19^9). This is in sharp contrast to the wide variation and generally larger amounts of DNA found in tumor cells
(STOWELL, 19U75 MELLORS et al., 1932; BADER, 1993; and LEUCHTENBERGER
et al., 195k) and normal cells during periods of rapid growth (SWIFT,
1950). Preliminary experiments utilizing endocervical strings contain ing methylcholanthrene resulted in the prompt appearance in vaginal
smears of cells containing increased amounts of nuclear DNA. The present studies were undertaken to determine the DNA content of exfoli
ated cervical and vaginal epithelial cells during the estrous cycle and in experimental inflammation and chemically induced cervical carcinoma.
MATERIALS AND METHODS
Sixty-six C^H female mice, 2 to 3 months of age and weighing 18 to 22 g. were divided into 12 groups. All animals were kept in an air-
conditioned room (68 - 72° F), fed Purina Dog Checkers supplemented by
daily additions of fresh bread and carrots. Groups 1 through 6 were used for a study of normal estrus.
Groups 1 and 2 consisted of 3 animals each. Animals of Groups
3 and 6 were ovariectoraized 10 days before start of experiment. Daily
93 96
vaginal smears were taken and animals sacrificed at various stages of
estrous cycle. Animals of Group 6 were injected subcutaneously with 29
RU of theelin in peanut oil and sacrificed 10 and 20 hours later.
Animals of Group 9 served as ovariectomized controls.
Fifty-two C^H female mice were used to study experimental inflam
mation and carcinogenesis. The early effects of croton oil and methyl-
cholanthrene were studied in U2 ovariectomized mice 2 - 3 months of age.
Ten C^H mice 7-12 months of age with cytologically and histologically
proved early or advanced cervical carcinoma were also studied. Endo-
cervical threads were placed in five groups of mice according to the
method of MURPHY (1993)* Sterile nylon 0 threads 9 cm. in length were
inserted into the cervical os through the vagina by means of a straight,
blunt-tipped needle, so that the end of the thread was just inside the
cervix. The thread was carried through the uterine wall and fixed by a
simple knot. Group 7, which served as negative controls, consisted of
seven mice with plain sterile nylon thread; Group 8 consisted of four
teen mice with threads dipped in a I; per cent solution of croton oil in
olive oil; Group 9 consisted of fourteen mice with threads coated with
methylcholanthrene crystals; and Group 10 consisted of seven mice with
threads dipped in a i; per cent solution of methylcholanthrene in olive
oil. Group 11 consisted of five mice with cervical cancer produced by
methylcholanthrene endocervical threads, and Group 12 of five mice with
cervical carcinoma produced by intravaginal painting twice weekly with
a 1 per cent solution of 3}U-benzpyrene in acetone. Smears were collect
ed from Groups 9 and 6 twice weekly after a diagnosis of malignancy was made. 97
Animals of Groups 1, 8, 9 and 10 were sacrificed 1, 2, 3, 9, 8,
11, and ll; days following operation. Groups 11 and 12 were killed at various times so as to include animals with early and advanced cervical cancer. Mitoses in the vaginal and cervical epithelium of all animals were arrested in metaphase by the colchicine method of BULLOUGH (191+9) •
Since BLOCH (1953) has shown that increases in nuclear DNA due to re constitution of nuclei arrested at metaphase by colchicine first begin to appear 10 hours following treatment, the period of treatment in this study was limited to 5 hours. Each animal received 0.1 mg. of colchi cine in 0.25 cc. of normal saline subcutaneously at 11 A.M. and was sacrificed by cervical dislocation at U P.M. Vaginal smears were
collected previous to the injection of colchicine and prior to sacri fice by the method previously described.
The smears were fixed immediately and stained by PAPANICOLAOU'S method (19U2). The uterus and vagina were removed in toto, fixed in
10 per cent buffered formalin, embedded in paraffin, and serial sections were cut alternately at 7 and 12 p. Sections cut at 7 p were used for mitotic counts. The cervical and vaginal epithelium at the vaginal fornix was examined with an oil immersion objective at a magnification of 970 diameters. All the nuclei in one field (between 65 and 75) were counted. Adjacent fields were examined until 1000 nuclei had been counted. This represented an average of fifteen oil-immersion fields.
In each animal 7000 cells were counted, and the mitotic index was expres sed as the number of dividing cells per 1000 cells. Sections cut at 12 p. were used for the demonstration of DNA. They were subjected to the
FEULGEN reaction (1921+), which was carried out on smears and tissue 98
sections after hydrolysis with n HCL at 60° C. for exactly 12 minutes.
Thirty to 90 intact interphase basal and parabasal cell nuclei selected
at random in each vaginal smear were subjected to the Feulgen reaction
and were measured by a microspectrophotometer similar to that described
by POLLISTSR (1992). A wave-length of 960 ra ji was used, which was
supplied by a Bausch & Lomb grating monochromator. This wave-length
represents the maximum absorption peak of the reddish purple dye. To minimize light scattering, slides of smears were mounted in oil with a
refractive index of 1.972, which matched that of unstained cervical
mucus. Relative amounts of DNA of individual nuclei in the smears were
obtained by multiplying the extinction values by the nuclear area,
since all the nuclei were flattened (SWIFT, 1999). The major (a) and minor (b) axes of nuclei were measured with an optical micrometer, and
the area was computed by the formula A = |a x gb. Interphase basal
cell nuclei were measured in tissue sections of the uterus at the cervi
cal and vaginal portions of the fornix. Care was taken to select uncut nuclei with no overlap; this limited the number of nuclei measured in the tissue sections. Relative amounts of nuclear DNA in the tissue sections were obtained by the method of SWIFT (1990). DNA values of approximately 1.00 (1.02 - 1.22) were obtained for nuclei containing
DNA corresponding to a 2n chromosome number; these were classified as diploid. Those containing DNA corresponding to more than 2n and less than a ipi chromosome number were classified as intermediates. Those containing twice the amount of DNA, equivalent to a Un chromosome number, were classified as tetraploids, DNA values equivalent to more than 1+n chromosomes constituted the polyploid class. Mean relative DNA 99 values obtained on more than ten nuclei were subjected to a statistical
analysis, and the standard error was calculated.
OBSERVATIONS
The results of DNA measurements in 990 basal and parabasal cell
nuclei, 20 keratinized cells containing pyknotic nuclei, and 1$ kera
tinized cells showing advanced pyknosis, karyolysis or karyorrhexis are
presented in Table 15. Representative histograms were prepared showing
the DNA content of basal, parabasal and keratinized cells during various
stages of estrus. Basal cells containing intermediate and tetraploid
amounts of nuclear DNA were in vaginal smears and tissue sections during
early proestrus, where keratinization was not of sufficient thickness to impair basal cell exfoliation. During estrus basal cells containing increased nuclear DNA, i.e. intermediates and tetraploids, were absent in vaginal smears while a considerable number were present in the tissues. Twenty hours following subcutaneous injection of theelin in ovariectomized mice the number of exfoliated basal cells containing in creased nuclear DNA in smears had decreased, while basal cells contain ing increased nuclear DNA in the tissues had increased. Appearance of cells showing increased amounts of DNA paralleled the increased mitotic activity of vaginal and cervical basal cells during proestrus and
estrus. During periods of quiescence (raetestrus, diestrus and cast rates), when mitotic activity was low, a constant diploid value of
nuclear DNA was present. Even though some of the superficial cells exfoliated during estrus showed a marked degree of nuclear enlargement and early keratinization, the DNA content remained at a diploid value.
A significant decrease in nuclear DNA occurred only in keratinized TABLE
NUCLEAR DNA OF 585 NUCLEI AND MITOTIC INDEX OF BASAL CELLS IN VARIOUS STAGES OF ESTRUS FOLLOWING COLCHICINE TREATMENT, 5 HR.
Total Diploid Cells % Inter Tetraploid Cells No. of Stage of Nuclei DNA, mediate DNA, Mitotic Group Animals Estrus Counted % Rel. Units Cells % Rel. Units Index
1 3 Proestrus 150 82 1.03+.09 12.6 5.3 1.98+.21 1*0 .2+ .la
2 3 Estrus 85 67.3 1.07+.11 22.2 10.5 2.0I4.+ .30 223.8+10.2 20* 100 1.01+.27 0 0 15 .53+.31- 0
3 2 Metestrus 80 100 1.11+.22 0 0 9.5+1.11
1* 2 Biestrus 75 98.7 1.10+.16 1 0 7.3± .92
>£ 2 Castrates 75 100 1.08+.13 0 0 11.8+1.23
6 1 Castrate + 1*5 71 1.05+.17 22.2 6.7 2.10+.18 28.7+3.2 theelin, 10 hr 1 Castrate + 50 61* 1.06+.11 26 1 0. 1.90+.22 l*8.3l6.1* theelin, 20 hr
* Pyknotic nuclei. Advanced pyknotic nuclei. 101 cells showing advanced pyknosis, keryolysis or karyorrhexis.
Following the previously described endocervical placements of variously treated threads, three distinct lesions were found: (a) acute ulcerative cervicitis, (b) cervical basal-cell hyperplasia, and (c) epi dermoid carcinoma. The results are summarized in Table 16. Although nuclear DNA was generally increased in epidermoid carcinoma as compared with hyperplasia, only a slight increase in cells containing intermedi ate amounts of nuclear DNA was noted. This was in contrast to a three fold increase in cells containing nuclear DNA corresponding to a polyploid number of chromosomes.
One day after endocervical thread placement, increases in the mitotic frequency of basal and parabasal cells were noted. The intensi ty and duration of the mitotic response appeared to be relative to the material applied (Chart £). This was accompanied by the appearance of cells containing increased amounts of nuclear DNA in the vaginal smears and tissues. A direct relation appeared to exist between the height of mitotic activity and the number of basal cells containing increased amounts of nuclear DNA.
Sterile Thread
Two days following plain thread insertion, the mitotic frequency of cervical and vaginal basal and parabasal cells had risen to an aver age of 80 mitoses per 1000 cells. This gradually decreased on the l4th day to 36 mitoses per 1000 cells (Chart $). During the peak of mitotic frequency, microspectrophotometric analysis of smears and tissues
showed a decrease in diploid nuclei from 90.4 to 82 per cent and an increase in tetraploid nuclei from 2.4 to 7.4 per cent. On the 14th Mitotic indices following endocervical application of various substancesofapplication various endocervical following indices Mitotic
MITOSES PER IOOO CELLS 300 250 200 100 150 - 0 5 - - EHLHLNHEE N OIL IN METHYLCHOLANTHRENE METHYLCHOLANTHRENE LI THREAD PLAIN CROTON OIL AS OLWN ISRIN F THREAD FOLLOWINGDAYS OF INSERTION
CHART 5 CHART 102 TAHLE 16 SUMMAH7 OF EXPERIMENTS ON THE NUCLEAR DNA CONTENT OF EPITHELIAL CELLS IN VARIOUS CERVICAL LESIONS
■P 5 . u # I s S h ^ £l* a M a s s No. Nuclei No. Measured Calls Diploid Main Cant Main Bar Calls Polyploid Group Animal Experimental Procedure Lesion * 3 5 £S3 Cent Mean Par Tetraploid Cent Mean Per Celle 7 7 1-lU days after insertion of plain thread 1.02*0.06 81*0 2.12*0.15 86.7 9.5 3.8 1-lU days after insertion Acute ulcera 1.12+0.13 8 11* of thread with croton tive cerri- 2.10*0.10 oil citis 3.27" 851 1*.52 77.3 lit. 9 7.6 1-lU days after insertion 1.08*0.08 9 11* of thread coated with Cervical hy 861* 2.10*0.19 67.1 20.9 10.1* methylcholanthrene perplasia 1*.17~ crystals 6.29 1-lU days after insertion 10 7 of thread dipped in li Cervical hy 1.22*0.12 per cent mathylcholan- perplasia 660 2.1*3+0.23 66.1* 21.0 10.9 threne in oil l*.6l“ 30-U5 weeks after inser 1.13+0.09 11 s tion of thread ooated Epidermoid 2.32+0.30 with mettylcholan- carcinoma 1*12 1*.68+0.51* 1*1.1* 26.9 27 threne crystals 6.62“ 13.27 32-I>0 weeks after paint 1.21*0.17 12 s ing intravaginally with Epidermoid 2.1*6*0.32 1 per cent 3jU-benspy- carcinoma 527 L.S8+0.L8 56.9 15.9 23.3 ene 6.89“ 10J+ day following treatment a total of U.6 per cent intermediate and 1.8 per cent tetraploid nuclei were found. These results are summarized in
Table 17. By the 3rd day histological studies showed a moderate degree of vaginal and cervical cornification and a mild inflammatory response.
Croton Oil Thread
Peaks of mitotic frequency following croton oil treatment occurred on the 2nd and 8th daysj the mitotic counts were 125? and 12U per 1000 cells, respectively (Chart £). The percentage of diploid nu clei decreased from 92.1 to 71.3 by the 2nd day and had risen to 80.9 by the lUth day. On the 2nd day 18.9 per cent intermediate and 9.8 per cent tetraploid nuclei were present in the vaginal smears and cervical epithelium. On the 8th day the number of epithelial cells containing increased amounts of nuclear DNA remained at 19.£ per cent intermediate cells with a decrease to 7.6 per cent tetraploid cells (Table 18). In a vaginal smear from one animal two polyploid nuclei containing increa sed amounts of DNA corresponding to a 6n and 8n chromosome number were found on the 11th day following treatment. In this group of mice as a whole a very profound inflammatory response was present by the 2nd day, with a diffuse leukocytic infiltration and vesicle formation in the mucosa, submucosal edema, and congestion. Focal mucosal denudation with necrosis and shallow ulcers was seen on the 3rd day. Throughout the period of inflammation a slight degree of keratinization occurred and did not appreciably increase up to the ll+th day. The epithelium in creased from two to four cell layers in the castrate animal prior to croton oil treatment to ten to eighteen cell layers by the li^th day. TABLE 17
NUCLEAR DNA CLASSES AND MITOTIC INDICES FOLLOWING INSERTION OF PLAIN THREAD
Days Total Following Cells Per Inter Per Tetra Per Mitotic Application Counted Diploid Cent mediate Cent ploid Cent Index*
1 125 113 90.U 9 7.2 3 2.k 29
2 122 100 82.0 10.6 9 7.U 80
3 126 105 83.3 15 11.9 6 U.8 5U
126 110 87.3 13 10.3 3 2.k 63
8 128 107 83.6 16 12.5 5 3.9 U8
11 10k 90 86.5 10 9.6 h 3.9 Uo
1U 10 9 102 93.6 5 U.6 2 1.8 36
•^Number of dividing cells/1000 cells. TABLE 18
NUCLEAR DNA CLASSES AND MITOTIC INDICES FOLLOWING APPLICATION OF CROTON OIL
Days Total Following Cells Per Inter Per Tetra Per Poly Per Mitotic Application Counted Diploid Cent mediate Cent ploid Cent ploid Cent Index
1 126 116 92.1 10 7.9 0 0 0 0 60
2 122 87 71.3 23 18.9 12 9.8 0 0 12$
3 123 91 7U.0 21 17.1 11 8.9 0 0 106
3 122 89 73.0 22 18.0 11 9.0 0 0 ' 119
8 118 86 72.9 23 19.$ 9 7.6 0 0 12b
11 12$ 96 76.8 19 1$.2 8 6.I4. 2 1.6 88
1U 11$ 93 80.9 9 7.8 13 11.3 0 0 li8 106 107
Crystalline Methylcholanthrene Thread
Two days after endocervical application of crystalline methyl cholanthrene the average mitotic index was 160 per 1000 cells. By the ll+th day the mitotic index had risen to 2^8 (Chart £)• Larger numbers of nuclei containing increased amounts of DNA were found in the vaginal smears and cervical epithelium as the mitotic frequency increased. Dy the lUth day 29.1 per cent intermediate, 11.3 per cent tetraploid, and
I4..8 per cent polyploid nuclei were present (Table 19). Fourteen highly polyploid nuclei, corresponding to a 6n, 8n, and 12n chromosome number, were encountered throughout the period of treatment. Occasional cells showed marked nuclear enlargement and contained nuclear DNA correspon ding to a diploid number of chromosomes. Inflammation and superficial mucosal necrosis were much less severe than that seen following the application of croton oil. By the liith day the height of the epithelium had increased to fifteen to twenty cell layers with cell crowding, vari ation in cell size and shape, and marked nuclear hyperchromasia.
Histologically, these lesions closely resembled carcinoma in situ.
Keratinization was slight throughout methylcholanthrene treatment.
Methylcholanthrene in Olive Oil
Methylcholanthrene in olive oil caused a more rapid though less intense increase in the mitotic frequency of vaginal and cervical epith elium than did crystalline methylcholanthrene (Chart £)• The highest percentage of cells containing increased amounts of nuclear DNA was found on the lirth day, when 27.5- per cent intermediate, l£.8 per cent tetraploid, and 3.3 per cent polyploid nuclei (6n and 8n) were found in TABLE 19
NUCLEAR DNA CLASSES AND MITOTIC INDICES FOLLOWING APPLICATION OF CRYSTALLINE 20-METHYLCH0LANTHRENE
Days Total Following Cells Per Inter Per Tetra Per Poly Per Mitotic Application Counted Diploid Cent mediate Cent ploid Cent ploid Cent Index
1 120 93 77.5 22 18.3 5 k .2 0 0 86
2 123 83 67.5 26 21.1 Ik 11.k 0 0 160
3 125 96 76.8 16 12.8 13 10.k 0 0 125
3 123 78 63.k 29 23.6 15 12.2 1 0.8 177
8 121 8k 69.k 21 17. k 13 10.7 3 2.5 150
11 128 77 60.2 31 2k.2 16 12.5 k 3.1 229
Ik 12k 68 5k.8 36 29.1 lk 11.3 6 k .8 258
H COO 1 0 9
the vaginal smears and cervical epithelium (Table 20). Many macro
phages containing oil droplets in the cytoplasm appeared in the smears by the 3rd and 5th days.
The nuclear DNA content of epithelial cells in the various
cervical lesion summarized earlier in Table 16 is evidence that the
application of croton oil or carcinogen resulted in the appearance of
cells containing increased amounts of DNA and showing a considerable variation in DNA content.
Methylcholanthrene String-induced Cervical Carcinomas
The mitotic frequency of methylcholanthrene string-induced
cervical cancer tissue varied between 232 and U08 per 1000 cells. No significant difference in mitotic frequency of tumor tissue was noted between early and advanced cervical cancer, although a lower mitotic in dex was present in tumors showing extensive central necrosis and infec tion. Abnormal cell division, such as lagging chromosomes, hollow metaphase plates, and occasional tri-and tetrapolar mitotic figures, was frequently observed. Malignant basal cells containing intermediate, tetraploid, and highly polyploid amounts of nuclear DNA were present in considerable numbers both in smears and tumor tissue (Chart 6).
Benzpyrene-induced Cervical Carcinomas
The frequency of mitoses in cervical cancer tissue induced by painting with benzpyrene varied between 267 and lfL9 per 1000 cells.
Less tumor necrosis and infection were present than in the tumors in duced by the endocervical string technic. The incidence of abnormal mitoses and basal cells containing intermediate, tetraploid, and TABLE 20
NUCLEAR DNA CLASSES AND MITOTIC INDICES FOLLOWING APPLICATION OF 2O-METHYLC HOLANTHRENE IN OIL
Days Total Following Cells Per Inter Per Tetra Per Poly Per Mitotic Application Counted Diploid Cent mediate Cent ploid Cent ploid Cent Index
1 127 97 76.li 21 16.9 9 7.1 0 0 97
2 131 96 73.3 22 16.8 12 9.2 1 0.7 190
3 120 86 71.7 19 19.8 13 10.8 2 1.7 127
✓£ 128 88 68.8 27 21.1 12 9. li 1 0.7 lli8
8 118 7k 62.6 28 23.7 13 11.0 3 2.9 197
11 116 68 98.6 30 29.9 19 12.9 3 2 .6 193
1U 120 61; 93.1; 33 27.9 19 19.8 li 3.3 179 110 Ill polyploid amounts of nuclear DNA was essentially similar to that ob served in the metbylcholanthrene-induced carcinomas (Chart 6).
DISCUSSION
The stimulating effect of estrogenic hormones on growth of genital epithelium is well established, although the bulk of uterine growth following estrogen treatment is largely due to increase in cell volume rather than to cell multiplication (SALVATORE, 191*8). ALLEN et al., (1937) utilizing colchicine t.o arrest mitoses have shown that estrogens actually markedly increase the rate of mitosis in vaginal and cervical epithelium of the mouse. The presence of numerous mitoses and interphase basal cells containing intermediate and tetraploid values of nuclear DNA in vaginal and cervical tissue during proestrus and estrus is in accord with this observation. This is in contrast to previous reports of constancy of nuclear DNA in hormonally stimulated uteri
(GELFANT AND CLEMMONS, 1955). However, several reasons may account for this difference. First, cells of vaginal and cervical epithelium where highest mitotic activity is seen, were not studied; secondly, due to the fact that vaginal and cervical epithelium represents but a small fraction of total uterine weights, a chemical determination of nucleic acids of the total uterus does not reflect the high degree of DNA synthesis occurring there. BLOCH 91953) found that although colchicine arrests ceils in mitosis it does not interfere with DNA synthesis and that nuclei falling into higher DNA classes began to appear 10 hours after colchicine treatment. The period of colchicine treatment used in this stuc^y was of insufficient length to impair the rate of mitosis or to result in polyploidy due to DNA synthesis without subsequent mitotic CHART 6
NUCLEAR DNA CONTENT OF MALIGNANT SQUAMOUS CELLS
20-METHYCH0LANTRENE i n d u c e d t u m o r s 3,4-BENZPYRENE INDUCED TUMORS
s m e a r CZD intermediate nuclei
tissue ESS intermediate nuclei
EZ3 tetraploid nuclei
POLYPLOID NUCLEI
X EARLY INVASIVE CARCINOMA m
INDIVIDUAL ANIMALS STUDIED
The D M content of nuclei in vaginal smears and cervical tissues in epidermoid carcinoma. 112 113 division. The increase in nuclear DNA is probably due to a premitotic synthesis of DNA, as has been shown to occur in rapidly growing tissues by PATAU AND SWIFT (1953). Furthermore, since HOWARD AND FELC (1951) found that synthesis of DNA in the interphase nucleus occurs within a relatively short time and is completed some time prior to mitosis, it is not surprising to find resting basal cells with increased amounts of nuclear DNA during periods of augmented growth. The preponderance of basal cells showing intermediate DNA values, and the relatively small tetraploid population, may indicate that once DNA synthesis has been completed division quickly ensues. This may explain the low mitotic index of uterine tissue following hormonal stimulation reported by
DRASHER AND S ALVATORE (19U8).
Basal cells containing increased amounts of DNA were present in vaginal smears taken during proestrus. The paucity of basal cells in smears during late proestrus and throughout estrus is due to the thick layer of keratin which prevents their exfoliation and is not a true re flection of the high degree of DNA synthesis present in the tissues during these periods.
Although cells exfoliated into the vaginal pool are more prone to cellular death and pyknosis, rnicrospectrophotometric studies revealed that the majority of nuclei contained the normal diploid amount of DNA.
This was true even though nuclear size was increased and early keratin ized cells showing advanced pyknosis. This is in agreement with
LEUCHTENBERGER'S report (1950) on nuclear DNA changes 'during pyknosis.
It would appear from these experiments that the DNA content of exfoliated basal and parabasal cells in the vaginal pool is quite stable Ill* and within limits, corresponds to the DNA content of these cells in the cervical and vaginal mucosa during various physiological states. These findings indicate that except during high estrus it is possible to
study DNA synthesis of epithelial cells during the estrous cycle by microspectrophotometric analysis of representative vaginal smears.
An increased mitotic frequency and growth rate of cells follow ing methylcholanthrene treatment have been observed by numerous workers
on a wide variety of biological material (HAMMETT AND REIMANN, 193!?;
OWENS et al., 1939j SPENCER AND MELROY, 19Ulj COOPER AND RELLER, 19U2).
In the present study a similar effect was noted soon after the appli
cation of croton oil to the cervical and vaginal epithelium of mice.
However, by the 8th day the number of mitoses decreased. When methyl cholanthrene was used the mitotic frequency continued to increase.
These results seem to indicate that distinct differences exist between
croton oil and methylcholanthrene. However, this difference cannot be further qualified until the mitogenic effect of croton oil and methyl
cholanthrene on the genital epithelium is studied at various dose levels. In view of recent reports (ROE, 19^6 and BOUTWELL et al., 19!?7) the mitogenic variations observed may be related to the weak carcino
genic properties of croton oil. Although numerous other factors may
also be responsible for the differences, the binding of methylcholan
threne to cytoplasmic particulates and cell proteins (SIMPSON AND
CRAMER, 19k$ and HEIDELBERGER AND MOLDENHAUER, 195^) may account for
its more prolonged and pronounced stimulation of mitoses. A similar protein-binding phenomenon for croton oil has not been reported.
The rapid rise in mitotic frequency observed in all the animals 115 shortly following thread placement occurred regardless of the material applied and strongly suggests that this is a nonspecific response by the vaginal and cervical mucosa to mechanical trauma of the endocervical procedure. This may also explain the deviation from the observations of
IVERSON AND EDELSTIEN (1952), who reported a temporary depression of epidermal mitoses following application of carcinogens to mouse skin.
These findings tend to corroborate the report of ROGERS AND ALLEN (1937) on the augmented mitoses and growth of vaginal epithelium following intravaginal swabbing. Although the increased numbers of cells in meta phase probably represent an augmented mitotic rate, the lengthening of mitosis by polycyclic carcinogens may also be a contributing factor
(LUDFORD, 1953).
After li; days of treatment with crystalline raetbylcholanthrene, a severe hyperplasia of the vaginal and cervical epithelium resulted which bore a close histological resemblance to carcinoma in situ.
Since these lesions were not allowed to proceed to invasive cancer, no conclusion as to their malignant potentialities can be reached.
Differences were also noted between,the inflammatory reaction incited by croton oil and by methyicholanthrene on the genital epithel ium. Increased necrosis and large numbers of leukocytes and phagocytes appeared when croton oil was used, while with methyicholanthrene a mild to moderate inflammatory reaction resulted.
The physical state of the material applied appeared to influence the intensity and duration of the mitotic response, as indicated by the slight drop in mitotic frequency on the lUth day when a solution of methyicholanthrene in olive oil was used, as compared with the continued 116 increase when crystalline methyicholanthrene was applied. The large numbers of oil-laden macrophages in smears from animals to which oil was applied suggest that the decrease in mitotic activity may be due to re moval by phagocytes of the material from the epithelial surfaces.
The prompt appearance in the tissues and smears of cells con taining increased amounts of nuclear DNA during periods of heightened mitotic activity indicates a close relation between DNA synthesis and mitosis. This is in accordance with the results of HOWARD AND PELC
(19>1)j who found that DNA synthesis in the interphase nucleus occurs rapidly and is completed before mitosis begins. The presence of basal cells containing intermediate (aneuploid) amounts of nuclear DNA in vaginal smears and tissues during the application of either methyicho lanthrene or croton oil made the demonstration of DNA classes difficult.
In animals with carcinoma the degree of aneuploidy made impossible the establishment of classes. This was in sharp contrast to the distinct
DNA classes found in normal castrate and estrogen-stimulated mice.
Occasional cells containing larger amounts of nuclear DNA, cor responding to a 6n and 8n chromosome number, made their appearance early in the course of either croton oil or methyicholanthrene treatment. In creased amounts of nuclear DNA may be due to either internal reduplica tion of DNA chromonema (polyteny) or reduplication of chromosomes (poly ploidy). It is impossible from the present data to differentiate be tween these mechanisms. Internal reduplication of DNA chromonema in epidermal cells of mice was noted by BIESELE AND COWDRI (19UU) as soon as 1-g- days after the application of methyicholanthrene. Mitotic abnor malities have been reported by KEMP (1930) to occur in inflammation and tissue regeneration. Moreover, similar 117 variations in chromosome number and mitosis have been observed by
TIMONEN (19>0) in normal human endometrium. These results all serve to emphasize the -ease with which the delicate processes of chromosomal reduplication and mitosis may be impaired. It would appear that poly ploidy and polyteny do not constitute specific attributes of the malig nant cell. Since smears and tissues in early and invasive cancer consistently showed a higher tetraploid and polyploid nuclear population than in inflammation, polyploidy appears to assume a quantitative rather than a qualitative relation to the malignant state. Although some de gree of the marked polyploidy found in the malignant nuclei is attribu table to increased mitotic frequency, the numerous abnormal mitoses encountered in these cells would also contribute to the disturbances in nuclear DNA content observed. Various abnormalities of chromosomal re duplication and spindle formation in malignant cells have been reported
(LEVAN AND HAUSCHKA, 19^3 and FANKHAUSER, 19$k).
SUMMARY
Increased amounts of nuclear DNA were found in basal cells during periods of augmented growth in the estrous cycle. These changes in nuclear DNA parallel the mitotic activity of the basal cells. Exfoli ated basal cells containing intermediate and tetraploid amounts of DNA were found only during early proestrus. The paucity of a tetraploid nu clear population during estrogenic stimulation may indicate rapid cellu lar division once DNA synthesis has occurred.
Microspectrophotometric analysis of nuclear DNA and mitotic counts revealed that a correlation exists between increased nuclear DNA and augmented mitoses, or mitotic rate. Endocervical application of 118 croton oil or 20-methylcholanthrene resulted in augmented mitoses and increased exfoliation of epithelial cells containing greater amounts of nuclear DNA. Distinct differences appear to exist between the mito- genic effect of croton oil and methyicholanthrene on the genital epi thelium of the mouse. Marked increases in the mitotic rate and nuclear
DNA (polyploidy) were observed in many malignant cells. CELLULAR ULTRASTRUCTURE OF EXPERIMENTAL SQUAMOUS
CELL CARCINOMA OF THE UTERINE CERVIX
The morphology of experimental squamous cell carcinoma has been largely studied by means of the light microscope. As in other tumors the most distinguishing characteristics of malignancy appear to be
localized in the nucleus. Observations by HCWATSON AND HAM (195k) suggest that various cytoplasmic components of neoplastic cells differ in certain respects from their normal counterparts. The present inves tigation was undertaken to examine the fine structure of chemically induced squamous cell carcinoma of the uterine cervix with special emphasis on the cytoplasm of the neoplastic cells.
MATERIALS AND METHODS
Squamous cell carcinoma of the uterine cervix was induced in
15 C^H female mice by the'technic of MURPHY (1953)- Eight to ten small pieces of tumor tissue were removed from various regions including the peripheral growing edge and quiescent central areas. Care was taken not to include necrotic tissue. Adjacent bits of tumor tissue were fixed in buffered formalin, embedded in paraffin, sectioned, and stained with
hematoxylin and eosin for purposes of orientation as to region and de gree of viability and growth activity. For electron microscopy the tissue was immediately immersed in several drops of 1 per cent osmium tetroxide buffered with veronal-acetate to a pH of 7.3 (PALADE, 1952).
These were then trimmed and immersed in 2 ml of the fixative and fixed for 3 hours at 3°C. Dehydration in graded alcohols and embedding in
119 120 methacrylate were carried out as outlined by PALADE (19^2) with the i substitution of either benzoyl peroxide or a-a-azodi-iso-butyronitrile
as catalysts instead of 2,i|.-dichlorobenzoyl peroxide. Sections were
cut with a Porter-Blum microtome at thicknesses of 0.03 - 0.06 p utilizing a diamond knife. The sections were floated on 2$ per cent
solution of acetone in water and mounted on standard electron micro
scope copper grids. They were examined in an Akashi, model TRS-!?0,
electron microscope fitted with a $0 y. objective aperture.
OBSERVATIONS
At low magnifications neoplastic cells from induced cervical
carcinomas showed the morphological characteristics of cancer cells
described earlier. Although there were considerable morphologic vari
ations from cell to cell, the following generalizations regarding the fine structure of their cytoplasm can be drawn. The mitochondria
showed considerable variation in both size and morphology. In cells at the apparent growing edge of tumors mitochondria were most numerous; occasionally elongate forms were also seen (Fig. 61). The cristae did not differ appreciably from those seen in normal mitochondria. In some tumor cells swollen mitochondria were seen interspersed among undamaged ones. These organelles were frequently globose containing vacuoles of varying size and often surrounded by densely osmiophilic material. In some instances remnants of cristae could be seen in the vacuoles. This was most prevalent in more quiescent, centrally located areas of tumor.
Cells in these areas generally contained fewer mitochondria. Malignant squamous cells undergoing keratiniazation showed extensive mitochondrial degeneration (Figs. 62, 62). These organelles were absent in fully 121 keratinized cells. The ergastoplasra in malignant cells consisted of o large numbers of granules measuring 80 to 300 A in diameter with only- occasional short filaments of endoplasmic reticulum. Numerous ergasto- plasmic vacuoles were also present. In cells undergoing keratinization the ergastoplasm contained many tightly packed granules; keratin fibrils first appeared near the nuclear membrane. Cell surfaces showed many microvilli (Fig. 6h) many of which were intimately entangled with simi lar structures of adjacent cells in contrast to the orderly interdigit- ations of non-neoplastic squamous cells. Desmosomes were generally absent or poorly formed in neoplastic cells. Not infrequently malignant squamous cells were separeted from each other by large lacuna-like inter-: cellular spaces lined by microvilli. In occasional cells ovoid bodies o 900 -1200 A in diameter were seen in the cytoplasm which somewhat re sembled virus particles. These consisted of a dense osmiophilic outer membrane and an inner clear zone, the center of which contained a small dense body (Fig. 69). BER0M et al. (19>9) also found virus-like bodies in chemically-induced experimental cervical carcinomas; however, these were smaller and usually concentrated in large intercellular spaces.
Fibroblasts immediately adjacent to tumor cells had large nuclei, prominent nucleoli and exhibited a well organized ergastoplasm consisting of numerous a-cytomembranes and granules with large bundles of extracellular collagen fibrils (Fig. 66). This was in contrast to nuclear and cytoplasmic morphology of fibroblasts not immediately sur rounding tumor cells. These fibroblasts exhibited large masses of extracellular collagen fibrils but lacked the large nuclei, nucleoli and highly organized endoplasmic reticulum (Fig. 67). DISCUSSION
These results suggest that there are no striking qualitative
ultrastructural differences between normal and neoplastic squamous
cells and are in essential agreement with the findings of DALTON AND
FELIX (19!?U) and DEROM et al. (19!j9). There were, however, certain
quantitative differences such as the increased ergastoplasmic granules
and paucity of cCcytomembranes in neoplastic cells. Similar findings
have been described by PALADE (19!jU) in embryonic and rapidly prolif
erating adult cells. However, this point is by no means settled, since
HOWATSON AND HAM (19>£) have identified well-formed cytomembranes in
the ergastoplasm of embryonic rat liver. According to these investi
gators, organized ergastoplasmic structures are related to the differen
tiated stage, being associated with specialized synthesis rather than
cell growth.
The quantitative and qualitative mitochondrial alterations
observed merit further investigation. The variations in mitochondrial population between the apparent growing edges and quiescent areas of
tumor are of considerable interest. Are these differences merely the
result of increased oxygen tension and supply of nutrients at the grow
ing edge in contrast to the hypoxia and malnutrition in the more central
areas of tumor? or is there a relation between the mitochondrial population of a tumor cell and its biological behavior? This important
question remains to be answered. Deficient desmosomes and curious inter
twined and twisted microvilli were found consistently in tumor cells which were frequently separated by extensive intercellular spaces. This
is morphologic evidence of the decreased mutual cohesiveness of 123 epithelial tumor cells described by COMAN (19UU)•
It is difficult to assess the significance of the virus-like cytoplasmic particles we observed in some of the tumor cells, since it has been well established that neoplastic cells are more susceptible to viral infection (BAND AND GEY, 1952; MOORE, 195Uj and KOPROWSKA AND
KOPROWSKI, 1957). Thus it may be that these particles, if indeed they are viruses, simply represent secondary viral infection of neoplastic cells. On the other hand, viruses may play a role in chemical carcino genesis. In any event caution must be exercised, in the interpretation of these findings, as stressed by BERNHARD (1958), especially since the time viral nature of these particles remains to be established. 12U
Fig. 61. Electron micrograph of an experimental carcinoma of the cervix.
The section was taken at the apparent growing edge showing
numerous mitochondria^ some of which are short rodlefs. Note
the absence of highly organized ergastoplasm. x 301:0
Fig. 62. Electron micrograph of the experimental cervical carcinoma
shown in Fig. fiL. This section was taken from a quiescent
area of tumor. In addition to the paucity of mitochondria
there is evidence of extensive mitochondrial damage. XJ^IOO
Fig. 63. Electron micrograph of an experimental cervical carcinoma.
Showing numerous degenerating mitochondria in a keratinized
tumor cell. X I4.OOO 125
vt>;/ & >- .. -JB
'<*; ''A 126
Fig. 6I4., Electron micrograph of an experimental cervical carcinoma.
The cytoplasm of the cell is filled with numerous small
ergastoplasmic granules (EG). Only a few short membranes are
present. The mitochondria (M) show considerable variation in
diameter. An intensely osmiophilic, degenerating mitochondria
(DM) can also be seen. The nuclear membrane (NM) consists of
a double membrane with numerous small pores. The cell mem
branes show many small microvilli (MV). Large intercellular
spaces (IGS) separate the tumor cells. X 16,000
Fig. 65. Electron micrograph of an experimental cervical carcinoma.
Numerous keratin fibrils (K) can be seen near the nuclear
membrane of a tumor cell in an early phase of keratinization.
The ergastoplasmic granules are dense and appear more differ
entiated than in the ergastoplasm of adjacent cells. The
mitochondria (M) are densely packed and show indistinct mem
branes. A virus-like particle (VP) with a central dense
body is present in the cytoplasm. X 15,000 127 128
Fig. 66. Electron micrograph. A fibroblast showing a well-developed
ergastoplasmic reticulum with large masses of collagen
fibrils. Note the tumor cell immediately nearby. X 2^00
Fig. 67. Electron micrograph. A fibroblast in an area distant from
tumor. Note the poorly developed endoplasmic reticulum^fc).
X 2^00
CYTOCHEMISTRY OF EXPERIMENTAL CERVICAL CARCINOMA
It has become increasingly apparent that the morphologic and
biologic differences between normal and neoplastic cells may have their
basis in fundamental biophysical and biochemical differences at the
cellular level. In recent years many new methods and instruments have
been developed for the study of cellular biophysics and biochemistry.
Excellent reviews and texts describing and evaluating these have been written by GLICK (191*9), ERANKO (1999), OSTER AND POLLISTER (1996),
MELLORS (1999), and PEARSE (1999) •
Although the cytochemical study of tissue sections is limited by
the rigors of tissue preparation and chemical reaction and is often dif ficult to quantitate, it offers the distinct advantage of supplying both morphological and chemical data on individual cells in a tissue. This
is in contrast to the quantitative data obtained by biochemical studies utilizing tissue slices or tissue homogenization and differential cen
trifugation, since these methods are complicated by the cellular heter
ogeneity of tissues and thus fail to give information on individual
cells.
MATERIALS AND METHODS
Cervical carcinoma was induced in sixty female C^H mice by the
MURPHY technic (1993). Twenty C^H mice served as controls; they were
sacrificed during various stages of the estrous cycle to determine the variations secondary to estrogen stimulation. Vaginal smears were pre pared weekly and stained by the method of PAPANICOLAOU (19^2). When
130 1 3 1 cytologic evidence of carcinoma was observed, the animals were sacri ficed. The uterus and vagina were removed as rapidly as possible, hemisected sagittally, one-half was mounted on metal chucks and frozen within 20 seconds by dipping the base of the chuck in an acetone-solid
COg mixture at 70° Cj these were employed for the demonstration of various enzymes. The second half was fixed in 10 per cent buffered formalin, embedded in paraffin in the routine manner for morphologic studies and the histochemical demonstration of glycogen by the periodic acid-Schiff (PAS) reaction (McMANUS, 19^8), utilizing salivary amylase controls and ribonucleoprotein (RNA) by the methyl green-pyronine stain (KURNICK, 1952). For the demonstrations of various oxidative enzymes sections were cut at ij. and 8 ji mounted on coverslips and incu bated the appropriate medium containing 7.5 per cent polyvinylpyrroli done (FVP) and the appropriate substrates and chemicals. In the case of beta-glucuronidase tissue sections were fixed in formalin-chloral hydrate fixative and the reaction performed according to the method of
FISHMAN AND BAKER (1956). Acid phosphatase activity was demonstrated by the azo dye coupling method utilizing sodium a-naphthyl phosphate
(SELIGMAN AND MANHEIMER, 19U9). To demonstrate diphosphopyridine (DPN) and triphosphopyridine diaphorase, succinic, B-hydroxybutyric and glucose-6-phosphate dehydrogenases, sections were incubated in 3,5 di phenyl-2 (U, 5 dimethylthiazol-2-6l) tetrazolium bromide (lmg/ml) and the appropriate substrate, coenzyme and either sodium azide or sodium amytal to block the respiratory chain (SCARPELLI, HESS AND PEARSE, 1958,
HESS, SCARPELLI AND PEARSE, 1958). To demonstrate malic, isocitric, glutamic, alcohol, lactic, and ^-glycerophosphate dehydrogenases 132 incubation was carried out in a medium containing Nitro BT (lmg/ml)
(TSOV, CHENG, NACHLAS and SELIGMAN, 1956, NACHLAS, WALKER and SELIGMAN,
1958) with the appropriate substrates, diphosphopyridine nucleotide, and sodium cyanide as the respiratory chain inhibitor. Incubation was carried out at 37°C and was limited to 15 minutes of normal cervical epitheliumj weak enyzmatic activity encountered in the carcinomas re quired an incubation up to 30 minutes. The reaction was stopped and the sections fixed by immersion in 10 per cent buffered formalin and mounted in glycerin jelly.
OBSERVATIONS
Ribonucleoprotein
Nuclear ribonucleic acid (RNA) stained either by the methyl greenpyronine technique (von HAAM AND SCARPELLI, 1955) or azure F*" appeared increased in neoplastic squamous cells when compared to normal basal and parabasal cells. Neoplastic cells in the growing portions of tumors generally displayed larger nucleoli which stained more intensely for RNA and showed more cytoplasmic basophilia than did tumor cells in more quiescent areas (Fig. 68). Quantitative microspectrophotometric estimations of these cells^ showed increases of RNA per unit area of cytoplasm as high as 28 per cent. However, when these data were calcu lated on a per cell basis, no significant differences were found between the total cellular RNA content of quiescent and growing tumor cells,
owing largely to the smaller cell size of the latter.
1 Kleinfeld, R., personal communication. 133 Glycogen
During high estrus normal squamous epithelium contained large amounts of glycogen. In diestrus no glycogen was present. There was a pregressive loss of glycogen during dysplasia. Glycogen was generally absent in malignant squamous cells. Occasionally small amounts of glycogen were found in a few highly differentiated malignant squamous cells.
Beta-Glucuronidase
In normal cervical and vaginal epithelium there was an increase in enzyme activity of basal and parabasal cells which reached a maximum during high estrus. In dysplasia increased enzyme activity was observed.
Malignant epithelium showed intense enzyme activity which was distribu ted evenly throughout viable areas of tumor. Decreased, and in some instances, no enzyme activity was found in nonviable, necrotic areas
Acid Phosphatase
Acid phosphatase activity in normal cervical and vaginal epi thelium was limited to the basal cell layer. Slightly increased activity was found during high estrus. Moderate to markedly increased enzyme activity was present in the cytoplasm of dysplastic and neo plastic squamous cells. Highest activity was found in rapidly growing anaplastic tumor cells. Acid phosphatase activity in either normal or neoplastic cells was localized solely in the cytoplasm.
Oxidative Enzymes - Normal Cervical Epithelium
The dehydrogenase systems were localized in mitochondria as rows of spherical formazan deposits measuring between 0.3-0.5 y in 13k diameter. These were separated from each other by a distance of 0.3 Ji.
The majority of mitochondria in the squamous cells were short 1-2 p rods containing 1-3 formazan deposits. These were distributed in a perinu clear fashion with no particular directional orientation. The highest activities were obtained for succinic, isocitric and glutamic dehydro genases ; these were localized in the basal and parabasal cell layers with weak to negative reactions in the cornified cell layer. Little variation was observed in the intensity and distribution of dehydro genase activity within any given cell layer in the cervical epithelium.
Variations in enzyme activity of squamous epithelium were observed during the estrous cycle. There was a gradual increase in enzyme activity beginning during proestrus and reaching a maximum at midcycle during high estrus •, thereafter there was a decrease in intensity of en zyme activity, reaching a minimum activity at diestrus.
Oxidative Enzymes - Malignant Squamous Epithelium
Tissue for enzyme studies was selected from non-necrotic areas.
With several exceptions malignant tissues showed consistently weaker dehydrogenase activities than normal cervical epithelium. Marked vari ations in the number and distribution of mitochondria containing the various dehydrogenases were observed in malignant squamous cells. This varied from cells containing scant mitochondria in the perinuclear cyto plasm to cells in which the cytoplasm was filled with enzymatically active mitochondria (Fig. 69). At the level of the light microscope no consistent differences were noted between the fine localization of enzymatically active sites in mitochondria in malignant cells and those in normal squamous epithelium. A decrease in activity of succinic, 13$ malic, DPN-linked isocitric, y#-hydroxybutyric, and lactic dehydrogen ases was found in malignant epithelium as compared to normal cervical epithelium. On the other hand malignant tissues exhibited an increased glucose-6-phosphate dehydrogenase activity which was highest at the periphery of the growing edge of the tumors, while tumor cells near the more central portions of the tumor showed weak enzyme activity (Fig. 70).
In an animal with early invasive carcinoma increased glucose-6-phosphate activity was observed in tumor cells at the site of basement membrane disruption (Fig. 71). Occasional malignant cells showing a high oC-glycerophosphate activity were scattered throughout the tumors.
Malignant squamous cells undergoing keratinization showed an increased glutamic dehydrogenase activity which was most marked in desquamated cells in the center of keratin pearls (Fig. 72). Necrotic portions of the tumors showed little to no activity for the various dehydrogenases.
No differences were noted between normal and malignant squamous cells when alcohol or reduced DPN and TPN were used as substrates. These findings are summarized in Table 21.
DISCUSSION
High Beta-glucuronidase activity was observed in neoplastic cells and tissues by FISHMAN et al. (19U7), CAMPBELL (19U9), SELIGMAN et al (l9h9) and Anlyan et al. (19$0). Although the function of beta- glucuronidase is intimately related to the function and metabolism of steroid hormones, the significance of its high activity in neoplastic tissues remains at present obscure. The cytoplasmic localization of alkaline phosphatase is in contrast to the nuclear localization observed in tumor cells by LEMON AND WISSEMAN (I9ii9). Since a cytoplasmic 1 3 6
Fig. 68. Invasive carcinoma showing the intense basophilia in tumor
buds at the apparent growing edge. The edge of a more
quiescent area of tumor is evident at the lower left. Note
the weak basophilia of these tumor cells. Azure blue B.
X 3lS
Fig. 69. Invasive carcinoma showing the individual variation of tumor
cells with respect to DPN-diaphorase enzyme activity. X U90
Fig. 70. Invasive carcinoma showing an intense glucose-6-phosphate
dehydrogenase at the growing edge in contrast to the weak
reaction in the more central areas of tumor. X 60 137 138
Fig. 71• Early invasive cervical carcinoma showing the intense glucose-
6-phosphate dehydrogenase activity at the apparent growing
edge. X £0
Fig. 72. Epidermoid carcinoma showing intense glutamic dehydrogenase
activity exhibited by some tumor cells near the center of a
keratin nodule. X £20 139
i © - ' •' jvii
r. g^-j 1
h #
S x\ * t ** fesa^sr'! IliO
TABLE 21
ACTIVITY* OF VARIOUS DEHYDROGENASES IN NORMAL AND MALIGNANT CERVICAL EPITHELIUM
Cervical Epithelium Normal Malignant Dehydrogenase 15 Min. l£ Min. 30 Min.
Succinic 2 0 1 Isocitric 2 0 1 Malic 1 0 1 Glutamic 1 1 2 /3 -hydroxybutyric 1 0 0 Glucose-6-phosphate 1 2 3 eC -glycerophosphate 1' 2 3 Lactic 1 0 1 Alcohol 0 0 0 DPN-diaphorase 2 2 3 TPN-diaphorase 1 1 2
*Enzyme activities were graded visually using an arbitrary 0—U range to evaluate relative intensities of the reactions.
localization is in agreement with current biochemical studies, it is probable that nuclear localization is the result of spurious diffusion
of reaction products.
The weak activity exhibited by most of the dehydrogenase enzyme
systems in the squamous cell carcinoma is in accord both with WARBURG'S
(1930) classic biochemical studies on tumor tissues of nearly 30 years
ago and more recent work (GREENSTEIN, 19^h and WEINHOUSE, ALLEN and
MILLINGTON (1953)* The most obvious advantage offered by the present
study is, of course, the intracellular localization of enzyme activity
at the mitochondrial level.
Although ALLARD and his co-workers (1952) have shown by their
counting techniques that neoplastic liver cells contain fewer raito- lUi
chondria than normal ones, their results represent an average value of
mitochondria per cell and do not give data on individual cells. Our
results show that malignant cells show as marked an individual cellular
variation in oxidative enzyme activity as they do for the nuclear con
tent of deoxyribonucleoprotein (SCARPELLI AND von HAAM, 1958). Whether
this is due to a variation in the number of mitochondria or to quali
tative and quantitative differences in the mitochondria remains at
present undetermined.
Cytochemical variations in enzyme activity in tumor tissue must
be interpreted with caution since they do not necessarily bear a re
lation to changes in the concentration of enzymes but may be due to
factors such as: l) competition for essential metabolites between tumor
and other tissues2) the presence of enzyme inhibitors produced by the
malignant tissues (HARGREAVES AND DEUTSCH, 1952), similar enzyme inhibi
tors having also been described in embryonic tissues (GUSTAFSON, 195U)j
and 3) mitochondrial swelling to due to cell damage, since this results
in increased enzyme activity (SCARPELLI AND PEARSE, 1958). This may
explain the high glutamic dehydrogenase activity observed in desquamated
malignant squamous cells in the center of keratin pearls. These cells
are no doubt subjected to varying degrees of hypoxia, malnutrition,
senescence and trauma following exfoliation.
The increased glucose-6-phosphate dehydrogenase activity exhibi
ted by soma of the tumor cells is of considerable interest since this
enzyme functions in the hexosemonophosphate cycle (HMP) which is the metabolic pathway responsible for the direct oxidation of glucose. In
addition this is the source of ribose-5-phosphate which is utilized in the synthesis of ribonucleic acid. GLOCK AND McLEAN (19$h) have shown that embryonic, rapidly proliferating, malignant and such hormone-depen dent tissues as ovary, adrenal, and lactating mammary gland exhibit a highly active HMP cycle. These tissues all synthesize large amounts of
RNA. These workers also found a close correlation between the RNA con tent and levels of the various HMP cycle enzymes of liver tissue from rats under various conditions of nutrition and hormone balance.
The high -glycerophosphate dehydrogenase activity observed in some of the tumor cells probably represents the augmented glycolysis characteristic of neoplastic tissues. In view of COHEN'S {I9$h) experi ments, which have shown, that the glycolytic pathway is capable of pro viding precursors for the synthesis of DNA, it is tempting to relate these differences in the dehydrogenase activities of various tumor cells to the quantitative differences in the nucleoprotein content of the so- called A and B type tumor cells described by CASPERSSON AND SANTESSON
(19l|.2) at the peripheral growing edge and the more central areas of tumors. Such a correlation, however, must await further investigations.
At present it is not known whether these biochemical variations in the tumor cell population are associated with differences in their biologi cal behavior. GENERAL SUMMARY
The genesis of a pathological process is a concept of its development based on serial observations which have been arranged in a particular sequence so as to give the most reasonable view of its evo lution. Such an attempt has been made in the interpretation of the foregoing experiments. The pathogenesis of experimental squamous cell carcinoma of the uterine cervix in mice appears to be a gradual process which is invariably characterized by the following stages: (a) acute inflammation, (b) dysplasia, (c) non-invasive, intraepithelial carcin oma (carcinoma in situ), and (d) invasive carcinoma. This sequence was found so consistently that there is little doubt that every invasive carcinoma induced experimentally was preceded by a carcinoma in situ, and that this in turn was preceded by epithelial dysplasia. At least part of the acute inflammatory phase is probably a non-specific response of the genital mucosa to the physical and chemical irritation of either the painting or string technique. It is included in the pathogenetic se quence until its true relation to carcinogenesis is more fully under stood.
Experimentally induced carcinoma of the uterine cervix and
vagina showed a striking biologic and pathologic resemblance to cervi cal cancer in women. Invasive lesions extended locally into surrounding tissues, metastasized to pelvic lymph nodes, produced obstructive uro- pathy, and in several instances, metastasized widely. The experimental
1U3 lesions differed from their human counterparts in their localization within the cervix due to the anatomical differences between the mouse and human uterus. Correlative cytohistologic studies showed that the nature of a cervical lesion is reflected in the cells desquamated from its mucosal surface. Cytologic changes were observed both in the nucle us and cytoplasm of cervical squamous cells during the pathogenesis of cancer. Those cellular changes which preceded the appearance of neo plasms by a considerable period of time were classified as precancerous.
Those demonstrable only in the presence of either in situ or invasive cancer were considered as changes typical of malignancy. These cellular alterations were both of a morphologic and chemical nature. Dysplastic cells contained enlarged, hyperchromatic nuclei and were arranged in a disorderly fashion in the cervical epithelium. Many dysplastic basal cells showed evidence of premature keratinization. Cytochemically these cells showed evidence of abnormal nucleoprotein metabolism. Neoplastic cells showed more morphologic variations, many bizarre mitoses and marked disturbances both in deoxyribonucleic and ribonucleic acids.
Enzymatic studies showed, with a few exceptions, an increased activity of various enzymes in dysplasia in contrast to lower activities observed in carcinoma.
Malignant cells show as marked an individual cellular variation in oxidative enzyme activity as they do for the nuclear content of DNA.
Increased activity of glucose-6-phosphate and a glycerophosphate dehy drogenases was observed in malignant cells, the former enzyme being highest at the growing edge of tumor. Electron microscopic studies suggest that these variations may be due to variations both in number iU5 and viability of mitochondria.
These studies lend support to the concept that the morphologic and biologic differences between normal and neoplastic cells may have their basis in fundamental biophysical and biochemical differences at the cellular level. BIBLIOGRAPHY
1. ALBERTINI, A. von: Electron microscopic study of epidermal carcinoma induced by methylcholanthrene in the mouse. J. Nat. Cancer Inst. 13: 11:73-1 1*99? 1953-
2. ALLARD, C., MATHIEU, R., DeLAMIRANDE, G., and CANTERO, A.: Mito chondrial population in mammalian cells 1. Description of a counting technic and preliminary results on rat liver in different physiological and pathological states. Cancer Research 12: 1:07-1*12? 1992.
3. ALLEN, E.: The oestrous cycle in the mouse. Am. J. Anat. 30: 297- 372, 1922.
U. ALLEN, E., and GARDNER, W.U.: Cancer of the cervix of the uterus in hybrid mice following long-continued administration of estrogen. Cancer Research 1: 3!?9-366, 19lil •
9. ANLYAN, A.J., GAMBLE, J., and HOSTER, H.A.: -Glucuronidase activity of the white blood cells in human leukemias and Hodgkin’s dis ease. Cancer 3: 116-123, 1990.
6. BADER, S.: Distribution of desoxyribose nucleic acid in tumor nuclei. Proc. Soc. Exp. Biol. & Med. 82: 312-319, 1993.
7. BANG, F.B. and GEY, G.O.: Comparative susceptibility of cultured cell strains to the virus of eastern equine encephalomyelitis. Bull Johns Hopkins Hosp. 91: 1:27-1*61, 1992.
8. BEALS, L.S.: Examination of sputum from a case of cancer of the pharynx and the adjacent parts. Arch. Med. (Lond.) 2: I4J4—ij.8 , i860.
9. BERENBLUM, I.: Carcinogenesis and tumor pathogenesis. Advances in Cancer Research 2: 129-179? 1991:.
10. BERNHARD, W . : Electron microscopy of tumor cells and tumor viruses. Cancer Research 18: 1:91-909? 1998.
11. BETHKE, R.: Multiple tumoren im nasenhorn. Frankfurt 2. Path. 6: 19-26, 1911.
12. BIESELE, J. J., and COWDRY, E.V.: Chromosomal changes in epidermal carcinogenesis. J. Nat. Cancer Inst. 1:: 373-381:, 19l*l*.
11:6 11*7
13. BIRBECK, M.S.C. and MERCER, E.H.: The electron microscopy of the human hair follicle. Part 1. Introduction and the hair cortex. J. Biophysic. Biochem. Cytol. 3'- 203-21U, 1957.
lU- BLOCH, D.P.: Effect of colchicine on synthesis of desoxyribonu cleic acid in tissue cultured rat fibroblasts. Proc. Soc. Exp. Biol. & Med. 8h: 31*1-31*5, 1953.
15. BOIVIN, A., VENDRELY, R., and VENDRELY, C.: L'acide desoxyribonu- cleique du noyau cellulaire, depositaire des caracteres hereditaries; arguments d'ordre analytique. Compt. rend. Acad Sc. 226: 1061-1063, 191*8.
16. BONSER, G.M., and ROBSON, J.M.: The effects of prolonged estrogen administration upon male mice of various strains: Development of testi cular tumors in the Strong A strain. J. Path. Bact. 5l: 9-22, 19l*0.
17. BOUTWELL, R.K., BOSCH, D., and RUSCH, H.P.: On the role of croton oil in tumor formation. Cancer Research 17: 71-75, 1957.
18. BULLOCK, F.D. and CURTIS, M.R.: Spontaneous tumors of the rat. Amer. J. Cancer li*: 1-115, 1930.
19. BULLOUGH, W.S.: The action of colchicine in arresting epidermal mitosis. J. Exp. Biol. 26: 287-291, 19l*9.
20. CAMPBELL, J.G.: The intracellular localization of -glucuronidase. Brit. J. Exp. Path. 30: 51*8-551*, 19l*9.
21. CASPERSSON, T.C., and SANTESSON, L.: Studies on protein metabolism in the cells of epithelial tumours. Acta Radiological Suppl. I4.6, 19l*2.
22. CIOLI, R.: Tumori sperimentali da catrame dell utero e loro svil- luppo in rapporto alia funzione ovarica. Boll, e atti d r Accad. med. di Roma 55: 291-296, 1929.
23. COHEN, P.P.: Chemical pathways of metabolism. (Greenberg Ed.) Vol. 1., 173-233, 1951*.
2l*. COMAN, D.R.: Decreased mutual adhesiveness a property of cells from squamous cell carcinoma. Cancer Research 1*: 625-629, 19l*l*.
25. COOK, J.W., HEWETT, C.L., and HEIGER, I.: The isolation of a cancer-producing hydrocarbon from coal tar II. Isolation of 1:2- and i*:5-benzpyrene, perylene, and 1:2-benzanthracene. J. Chem. Soc., Part I 396-398, 1933.
26. COOPER, Z.K., and RELLER, H.C.: Mitotic frequency in methylcho- lanthrene epidermal carcinogenesis in mice. J. Nat. Cancer Inst. 2: 335-31*1*, 191*2. l l j . 8
27. COTCHIN, E.: Neoplasms of domesticated mammals. A Review. Commonwealth Agricultural Bureaux, Farnham Royal, Bucks, England. p. 100, 19*6 .
28. CROSSEN, R.J., and LOEB, L.: Effect of long-continued adminis tration of an estrogen on the sex organs of mice which have passed the reproductive period. Arch. Path. 37: 303-312, 19l;lw
29. DALTON, A.J., and FELIX, M.D.: The electron microscopy of normal and malignant cells. Ann. N.Y. Acad. Sci. 63: 1117-llUO, 1936.
30. DARLINGTON, C.D.: Chromosome chemistry and gene action. Nature 1U9: 66-69, 19U2.
31. DeROM, F., THIERY, M., SEBRUYNS, M., DeGROODT, M., and LAGASSE, A.: L ’ultrastructure du cancer experimental du col de l'uterus. Gynecol. et Obstet. 38: 182-190, 1939.
32. DiSTEFANO, H.S.: A cytochemical study of the Feulgen Nucleal Re action. Chromosoma 3: 282-301, 19^8.
33. DOBROVOLOSKAIA-ZAVADSKAIA, N., and SIMSILEVITCH, F.: Au sujet de ' la reaction des organes genitaux de la souris aux instillations du 1:2: 3:6, dibenzanthracene dans le vagin. In libro de Oro Angel H. Roffo, Buenos Aires, Argentina. A. Guidi Buffarini 1910-1933; 293-301.
3U. DONALDSON, F.: The practical application of the microscope to the diagnosis of cancer. Am. J. Med. Sci. N.S. 23: U3-70, 1833•
33. DRASHER, M.L.: Morphological and chemical observations on the mouse uterus during the estrous cycle, and under hormonal treatment. J. Exp. Zool. 119: 333-333* 1932.
3 6. DUNHAM, L.V.: A survey of transplantable and transmissible animal tumors. J. Nat. Cancer Inst. 13: 1299-1377* 1933*
37. DUNN, T.: Normal and pathologic anatomy of the reticular tissue in laboratory mice with a classification and discussion of neoplasms. J. Nat. Cancer Inst, li;: 128l-ll;33, 193U.
38. DUSTIN, P., and PARMENTIER, R.: Donnees experimentales sur la nature des mitoses anarmales observees dans certains epitheliomas due col uterin. Gynecol, et Obstet. 32: 238-263, 1933.
• • # * 39* ERANKO, 0.: Quantitative methods in histology and microscopic histochemistry. Little, Brown and Co., Boston, 1933* ij.0. FANKHAUSER, G. In: E.J. Boell (ed.), Dynamics of Growth Processes, pp. 68-91;. Princeton, J.J.: Princeton University Press, 1931;.
1(1. FELDMAN, W.H.: Neoplasms of domesticated animals. Saunders, Philadelphia, 1932. 11*9
1*2. FELLNER, D.O.: Experimentell erzeugte wachstrumsverander ungen am weiblichen genitale der kaninchen. Centralbl. F. allg Path u path Anat. 23: 673-875, 1912.
1*3• FEULGEN, R., AND ROSSENBECK, H.: Mikroskopisch chemischer nach- weis einer nucleinsaure von typus der thymonucleinsaure und die darauf beruhende elektive Farbung von Zellkernen in mikroskopischen Praparaten Hoppe-Seyler's Ztscher f. physiol. Chem. 135: 203-21*8, 1921*.
1*1*. FISHMAN, M., SHEAR, M.F., FRIEDMAN, H.F., and STEWART, H.L.: Studies in carcinogenesis. XVII. Local effect of repeated application of 3,U-benzpyrene and of human smegma to the vagina and cervix of mice. J. Nat. Cancer Inst. 2: 361-367, 19l*l.
1*5. FISHMAN, W.H., ANLYAN, A.J., and GORDON, E.: -Glucuronidase activity in human tissues. Some correlations with processes of malig nant growth and with the physiology of reproduction. Cancer Research 7: 808-818, 19l*7.
1*6. FISHMAN, W.H., and BAKER, J.R.: Cellular localization of -Glu curonidase in rat tissues. J. Histochem. Cytochem. 1*: 570-587, 1956.
1*7. FUSCO, G.: Cancro sperimentale dell utero e della vagina. Arch. Ostet. Ginec. 19: 315—3UU, 1*67-1*69, 1932.
1*8. GARDNER, W.U.: Carcinoma of the uterine cervix and upper vagina; Induction under experimental conditions in mice. Ann. New York Acad. Sci. 75: 51*3-561*, 1959.
1*9. GARDNER, W.U.: Experimental induction of uterine cervical and vaginal cancer in mice. Cancer Research 19: 170-176, 1959a.
50. GARDNER, W.U.: ALLEN, E., SMITH, G.M., and STRONG, L.C.: Carcinoma of the cervix of mice receiving estrogens. J. Amer. Med. Ass. 110: 1182-1183, 1938.
51. GARDNER, W.U., and ALLEN, E.: Malignant and non-malignant uterine and vaginal lesions in mice receiving estrogens and estrogens and andro gens simultaneously. Yale J. Biol. Med. 12: 213-23U, 1939.
52. GARDNER, W.U., DOUGHERTY, T.F., and WILLIAMS, W.L.: Lymphoid tumors in mice receiving steroid hormones. Cancer Research 1*: 73-87, 191*1*.
93. GARDNER, W.U., and PAN, S.C.: Malignant tumors of the uterus and vagina in untreated mice of the P M stock. Cancer Research 8: 2l*l- 256, 191*8.
51*. GELFANT, S., and CLEMMONS, J.J.: The influence of estrogen on nuclear volumes and chemical composition. J. Cell and Comp. Physiol 1*6: 529-539, 1955. i9o 99. GIESON, I. van: Laboratory notes of technical methods for the nervous system. New York M.J., 90: 97-60, 1889.
96. GLICK, D.: Techniques of histo-and cytochemistry. (Inter science Publishers, New York, 19k9)•
97. GLOCK, G.E., and McLEAN, P.: Levels of enzymes of the direct oxidative pathway of carbohydrate metabolism in mammalian tissues and tumors. Biochem. J. 96: 171-179, 199U.
98. GLUCKSMANN, A., and CHERRY, C.P.: The role of the ovary in the in duction of tumors by the local application of 9,10-dimethyl-1,2-benzan thracene to the genital tract of rats. Brit. J. Cancer 12: 32-U3, 1997.
99. GREENSTEIN, J.P.: The biochemistry of cancer. (Academic Press, New York, 195b)•
60. GUSTAFSON, T.: Enzymatic aspects of embryonic differentiation. Int. Rev. Cytol. 3: 277-327, 195k.
61. HAAM, E. von: Some observations in the field of exfoliative cytology. Am. J. Clin. Path. 2 k ‘ 692-662, 199U.
62. HAAM, E. von; and MENZIES, P.: Experimental studies in exfoliative cytology. Proc. Amer. Assoc. Cancer Research 1:22, 1993.
63. HAAM, E. von, and SCARPELLI, D.G.: Experimental carcinoma of the cervix: A comparative cytologic and histologic study. Cancer Research, 19: kk9-k55, 1999.
6U. HAMMETT, F.S., and REIMANN, S.P.: Effect of methylcholanthrene on developmental growth of obelia geniculata. Am. J. Cancer, 29: 807-808, 1939.
69. HAMPERL, H., KAUFMANN, C., and OBER, K.G.: Histologische unter suchungen an der cervix schwang erer fraven. Die erosion und das car cinoma in situ. Arch Gynak. i8U: 181-280, 195k •
66. HANSEMANN, D. von: Ueber asymmetrische zelltheilung in epithel- krebsen und deren biologische bedentung. Virchow's Arch. F. path, anat. 119: 299-326, 1890.
67. HARGREAVES, A.B. and DEUTSCH, H.F.: The in vitro inhibition of catalase by a tumor factor. Cancer Research 12: 720-726, 1992.
68. HEIDELBERGER, C., and MOLDEMHAUER, M.G.: Quantitative studies on the binding of various hydrocarbons to the proteins of the skin of mice. Proc. Am. Assoc. Cancer Research 2(1): 2k, 1995. l5l
69. HESS, R., SCARPELLI, D.G., and PEARSE, A.G.E.: The cytochemical localization of oxidative enzymes II ftyridine nucleotide - linked de hydrogenases. J. Biophys. and Biochem. Cytol. 1*: 753-760, 1958.
70. HISAW, F.L., Meyer, R.K., and WEICHERT, C.K.: Inhibition of ovulation and associated histological changes. Proc. Soc. Exper. Bio. and Med. 25: 75U-756, 1927-28.
71. HOCH-LIGETI, C.: Effect of prolonged administration of spermici dal contraceptives on rats kept on low-protein or on full diet. J. Nat. Cancer Inst. 18: 661-685, 1957.
72. HOOKER, C.W., GARDNER, W.U. and PFEIFFER, C.O.: Testicular tumors in mice receiving estrogens. J.A.M.A. 115: l*l*3-l*l*5, 19l*0.
73. HOWARD, A., and PELC, S.R.: Nuclear incorporation of P ^ as demon strated by autoradiographs. Exp. Cell Research 2: 178-187, 1951.
7U. H0WATS0N, A.F. and HAM, A.W.: Electron microscope study of sections of two rat liver tumors. Cancer Research l5: 62-79, 1955. .
75. HSU, T.C.: Cytological studies on HeLa, a strain of human cervi cal carcinoma. L. Observations on mitosis and chromosomes. Texas Rep, Biol, & Med. 12: 833-81*6, 1951*.
76. ICHIKAWA, K., and BAUM, S.M.: Reaction locale chez les animaux resistant a la production du cancer par badigeonnage au goudron de houille (Rats et Cobayes). Bull. Assoc. Franc. Etude Cancer 13: 386- 395, 1921*.
77. ISC0VESC0, H.: Les lipoides de L'ovaire. Compt rend. Soc. de biol. 73: 16-18, 1912.
78. IVERSON S., and EDELSTEIN, J.M.: On the mechanism of experimental carcinogenesis. 1*. The early mitotic effect of 9,10-dimethyl-1,2- benzanthracene on epidermal cells of the mouse ear. Acta Path, et Micro biol. Scandinav. 30: 213-222, 1952.
79. KEMP, T.: Ueber die somatischen mitosen bei menschen und warrablu- tigen tieren unter normalen und pathologischen verhaltnissen. Ztschr. F. zellforsch. u. mikr. Anat. 11: 1*29-1*1*1*, 1930.
80. KENNAWAY, E.L. and HEIGER, I. Carcinogenic substances and their fluorescence spectra. Brit. Med. J. 1: 10l*l*-10l*6, 1930.
81. KLENITSKY, J.S.: On experimental cancer of the uterine cervix. Bull Biol, et med exper. U.R.S.S. 9: 3-6, 19l*0.
82. ROLLER, P.O.: Origin of malignant tumor cells. Nature l5l: 2l*l*- 21*6, 191*3. 152 83. KOLLER, P.C.: Abnormal mitosis in tumors. Brit J. Cancer 1: 38- kl, 19U7.
8U. KOPROWSKA, I., and BOGACZ, J.: A cytopathologic study of tobacco tar-induced lesions of uterine cervix of mouse. J. Nat Cancer Inst. 23: 1-19, 1959.
85. KOPROWSKA, I., BOGACZ, J., PENTIKAS, C., and STYPULKOWSKI, ¥.: Induced cervical carcinoma of mouse. A quantitative cytologic method for evaluation of the neoplastic process. Cancer Research 18: 1186- 1190, 1958.
86. KOPROWSKA, I., and KOPROWSKI, H.: Enhancement of susceptibility to virus infection in the course of a neoplastic process. Ann N.Y. Acad. Sci. 68: k0b-bl8, 1957.
87. KURNICK, N.B.: Histological staining with methyl green-pyronin. Stain. Technol. 27: 233-21:2, 1952.
88. LACASSAGNE, A.: Tumeurs malignes apparues au cours d 'un traite- ment hormonal combine, chez des souris appartenant a des lignees re- fractaires au cancer spontane. Compt rend soc. de biol. 121: 606- 609, 1936.
89 . LACASSAGNE, A.: Relationship of hormones and mammary adenocarcino ma in the mouse. Am. J. Cancer 37: l)lU-U2U, 1939
90. LAMBERT, R.A.: Comparative studies upon cancer cells. II. The character of growth in vitro with special reference to cell division. J. Exper Med. 17: 1:99-510, 1913.
91. LAPID, L.S., and GOLDBERGER, M.A.: Exfoliative dyskaryotic cells associated with atypical cervical lesions. Am. J. Obstet Gynec. 6l: 132U-1328, 1951.
92. LEMON, H.M., and WISSEMAN, C.L.: Acid phosphomonoesterase activity of human neoplastic tissue. Science 109: 233-235, 19U9.
93. LEUCHTENBERGER, C.: A cytochemical study of pycnotic nuclear de generation. Chromosoma 3: Ut9-li73, 1950.
9k. LEUCHTENBERGER, C.; LEUCHTENBERGER, R., and DAVIS, A.M.: A micro- spectrophotometric study of the desoxyribose nucleic acid (DNA) content in cells of normal and malignant human tissues. Am. J. Path. 30: 65- 85, 195U.
95. LEVAN, A., and HAUSCHKA, T.S.: Endomitotic reduplication mechanisms in ascites tumors of the mouse. J. Nat. Cancer Inst. lU: 1-U3, 1953.
96. LEV/IS, M.R., and LEWIS, W.H.: Malignant cells of walker rat sarco ma no. 338. Am J. Cancer 16: 1153-1183, 1932. 153
97* LIPSCHUTZ, A., and VARGA, L., Jr.: Structure and origin of uterine and extragenital fibroids induced experimentally in the guinea pig by prolonged administration of estrogens. Cancer Research 1: 236- 2h9, 191(1.
98. LOEB, L., BURNS, E.L., SUNTZEFF, V., and MOSKOP, M.: Carcinoma like proliferations in vagina, cervix and uterus of mouse treated with estrogenic hormones. Proc. Soc. Exp. Biol. & Med. 35: 320-322, 1936.
99• LOEB, L., SUNTZEFF, V., and BURNS, E.L.: Growth processes in duced by estrogenic hormones in the uterus of the mouse. Amer. J. Cancer 3U: la 3-1(27, 1938.
100. LUDFORD, R.J.: Chemically induced derangements of cell division. J. Roy. Microscop. Soc. 73: 1023, 1953-
101. MARTZL0FF, K.H.: Carcinoma of the cervix uteri: A pathological and clinical study with particular reference to the relative malignancy of the neoplastic process as indicated by the predominant type of cancer cell. Bull, John Hopkins Hosp. 3h: lla-ll(9, 18U-195* 1923*
102. McEUEN, C.S.: Occurrence of cancer in rats treated with Oestrone. Am. J. Cancer 3k: 18U-195, 1938.
103. McMANUS, J.F.A.: Histological and histochemical uses of peri odic acid. Stain Technol. 23: 99-108, 19U8.
10l|. MELL0RS, R.C. (ed.): Analytical cytology, 2nd ed. McGraw-Hill, New York, 1959-
105. MELL0RS, R.C., KEANE, J.F., Jr., and PAPANICOLAOU, G.N.: Nucleic acid content of the squamous cancer cell. Science 116: 265-269, 1952.
106. MEYER, R.K., and ALLEN, W.M.: Production of mucification of vag inal epithelium of rodents by oestrous hormone. Science 75: 111-112, 1932.
107. MEYER, R.K., and ALLEN, W.M.: The production of mucified cells in the vaginal epithelium of certain rodents by oestrin and by corpus luteum extracts. Anat. Rec. 56: 321-31(3? 1933.
108. MIRSKY, A.E., and RIS, H.: Variable and constant components of chromosomes. Nature 163: 666-667, 19l(9.
109. M0ELLEND0RFF, ¥. von: Zur kenntnis der mitose. VIII. Zur an alyse des pathologischen wachstums hervorgerufen durch chloral hydrat, geschlechtshormone und cancerogene kohlenwasserstoffe. Ztschr. f. Zellforsch. u. mikr. Anat. 29: 706-71(9? 1939.
110. MOORE, A.E.: The effect of viruses on tumors. Ann. Rev. Microbiol. 8: 393-UlO, 195U. i5U
111. MOORHEAD, P.S., and HSU, T.C.: Cytologic studies of HeLa, a strain of human cervical carcinoma. III. Durations and characteristics of the mitotic phases. J. Nat. Cancer Inst. 16: 10U7-1066, 1956.
112. MURPHY, E.D.: Studies on carcinogen-induced carcinoma of the cervix in mice. Amer. J. Path. 29: 608, 1953*
113. MURRAY, J.A.: The zoological distribution of cancer. Imp. Cancer Res. Fund Sci. rep. 3: Ul-60, 1908.
11U. NACHLAS, M. M., WALKER, D.G., and SELIGMAN, A.M.: The histocherai- cal localization of diphosphopyridine nucleotide diaphorase. J. Biophys and Biochem. Cytol. U: 29-38, 1958.
115. OSTER, G. and POLLISTER, A.W.: (ed.): Physical techniques in biological research. Vol. III. Cells and tissues. (Academic Press, New York, 1956).
116. 0VERH0LZER, M.D. and ALLEN, E.: Ovarian hormone and traumatic stimulation of monkey's cervix to a condition resembling early cancer. Proc. Soc.. Exp. Biol, and Med. 30: 1322-1326, 1933.
117. OWENS, S.E., WEISS, H.A., and PRINCE, L.H.: Carcinogens and Planarian tissue regeneration. Am. J. Cancer 35: I4.2i4.-U26, 1939.
118. OXORN, H.: Cervical cytology. Key to diagnosis of early uterine cancer. Surg. Gynec. Obstet. 87: 197-205, 19U8.
119. PALADE, G. E.: A small particulate component of the cytoplasm. Anat Rec. 118: U32-U33, 195U-
120. PAN, S.C., and GARDNER, W.U.: Carcinomas of the uterine cervix and vagina in estrogen-and androgen-treated hybrid mice. Cancer Re search 8: 337-3U5, 19U8.
121. PAPANICOLAOU, G. N.: A new procedure for staining vaginal smears. Science 95: U38-U39, 19U2.
122. PAPANICOLAOU, G. N., and TRAUT, H.F.: Diagnosis of cancer by vaginal smears, New York, The Commonwealth Fund, 19U3.
123. PARMENTIER, R., and DUSTIN, P., Jr.: Reproduction experimentale d' une anamalie particuliere de la metaphase des cellules malignes (metaphase 'a' trois groupes). Caryologia U: 98-109, 1951.
12U. PATAU, K., and SWIFT, H.: The DNA-content (Feulgen) of nuclei during mitosis in a root tip of onion. Chromosoma 6: 1U9-169, 1953*
125. PEARSE, A. G. E.: Histochemistry: Theoretical and Applied: 2nd edition (Churchill, London, 1959). 155
126. PERRY, I.H.: Production of carcinoma of the uterus in mice. Proc. Soc. Exp. Biol, and Med. 35s 325-326, 1936.
127. PERRY, I.H., and GINZTON, L.L.: The development of tumors in female mice treated with 1:2:5:6-dibenzanthracene and theelin. Am. J. Cancer 29: 680-701;, 1937.
128. PIANESE, G.: Beitrag zur hitologie und aetiologie des car- cinoms. Ziegler's Beitrage s. path. Anat., Supp. 1: 193, 1896.
129. POLLISTER, A.W.: Photomultiplier apparatus for microspectro photometry of cells. Lab. Invest. 1: 106-111;, 1952.
130. PORTER, K.R.: Observations on the submicroscopic structure of animal epidermis. Anat. Rec. 118: i;33, 1951;.
131. POUCHET, F.A.: Theorie positive de 1'ovulation spontanee et de la fecondation des maramiferes et de l'espece humaine, basee sur 1'ob servation de toute la serie animale. Paris, Balliere, 181;7•
132. PRATT-THOMAS, H.R., HEINS, H.C., LATHAM, E., DENNIS, E.J., and McIVER, F.A.: The carcinogenic effect of human smegma; An experi mental study 1. Preliminary report. Cancer 9: 671-680, 1956.
133. RATCLIFFE, H.L.: Incidence and nature of tumors in captive wild mammals and birds. Amer. J. Cancer 17: 116-135, 1933•
13U. REAGAN, J.W., and HAMONIC, K.J.: The cellular pathology in car cinoma in situ. Cancer 9: 385-l;02, 1956.
135. REAGAN, J.W., HICKS, D.J., and SCOTT, R.B.: Atypical hyperplasia of uterine cervix. Cancer 8: l;2-52, 1955a.
136. REAGAN, J.W., SEIDEMANN, I.L., and SARACUSA, Y.: The cellular morphology of carcinoma in situ and dysplasia or atypical hyperplasia of the uterine cervix. Cancer 6: 22l;-235, 1953*
137. REAGAN, J.W. and WENTZ, W.B.: Changes in the mouse cervix antedating induced cancer. Cancer 12: 389-395, 1959.
138. REAGAN, J.W., WENTZ, W.B. and MACHICAO, N.: Induced cancer of the cervix uteri in the mouse. Arch. Path. 60: U5l-U6l, 1955.
139. ROE, F.J.C.: The development of malignant tumors of mouse skin after "initiating" and "promoting" stimuli. III. The carcinogenic action of croton oil. Prit. J. Cancer 10: 72-78, 1956. lUO. ROGERS, P.V., and ALLEN, E.: Epithelial growth caused by stimu lation with various smear methods as demonstrated by mitotic stasis with colchicine. Endocrinology 21: 629-632, 1937* 156
11*1. SALVATORE, C. A.: A cyfcological examination of uterine growth during the estrous cycle and artifically induced estrus. Rev. Brasil Biol. 8: 505-523, 19k8.
ll*2. SCARPELLI, D.G., and von HAAM, E.: Experimental carcinoma of the uterine cervix in the mouse. Am. J. Path. 33? 1059-1073, 1957.
11*3. SCARPELLI, D.G., HESS, R., and PEARSE, A.G.E.: The cytochemical localization of oxidative enzymes. 1. Diphosphopyridine nucleotide di- aphorase and triphosphopyridine nucleotide diaphorase. J. Biophys. and Biochem. Cytol. 1*: 753-760, 1958.
11*1*. SCARPELLI, D.G., and PEARSE, A.G.E.: Physical and chemical pro tection of cell constituents and the precise localization of enzymes. J. Histochem. Cytochem. 6: 369-376, 1958. ll*5. SCARPELLI, D.G., and FEARSE, A.G.E.: A cytochemical study of various oxidative enzymes in experimental carcinoma of the uterine cervix. Trans. Inter-Society Cytol. Council 2-8, 1956.
11*6. SELBY, C.C.: An electron microscope study of the epidermis of mammalian skin in thin sections. 1. Dermo-epidermal junction and basal cell layer. J. Biophysic. Biochem. Cytol. 1: 1*29-111*, 1955.
11*7. SELIGMAN, A.M., NACHLAS, M.M., MANHEIMER, L.H., FRIEDMAN, O.M., and WOLF, G.: Development of new methods for the histochemical demon stration of hydrolytic intracellular enzymes in a program of cancer research. Ann. Surg. 130: 333-311, 19l9.
118. SIMPSON, W.L., and CRAMER, W.: Fluorescence studies of carcin ogens in skin. II. Mouse skin single and multiple applications of 20- methylcholanthrene. Cancer Research 5: 119-163, 1915.
119. SLYE, M.: The fundamental harmonies and the fundamental differ ences between spontaneous neoplasms and all experimentally produced tumorsstudies in incidence and heritability of spontaneous tumors in mice. J, Cancer Res. 8: 210-273, 1921.
150. SLYE, M.j HOLMES, H.F., WELIS, H.G.: Primary spontaneous tumors of the uterus in mice with a review of comparative pathology of uterine neoplasmj studies on the incidence and the heritability of spontaneous tumors in mice. J. Cancer Res. 8: 96-118, 1921.
151. SMITH, A., and JONES, T.C.: Veterinary Pathology. (Lea & Febiger, Philadelphia, 1957)•
152. SOLOMON, J.B.: Glutamic dehydrogenase in the developing chick embryo. Biochem J. 66: 261-270, 1957
153. SPENCER, R.R., and MELROY, N.B.: Effect of carcinogens on small organisms. III. Cell-division rate and population levels of methylcho- lanthrene-adapted paramecia. J. Nat. Cancer Inst. 2: 185-191, 19ll. 157 l5U. STOCKARD, C.R., and PAPANICOLAOU, G.N.: The existence of a typical oestrous cycle in the guinea-pig-with a study of its hsbologi- cal and physiological changes. Am. J. Anat. 22: 225-283, 1917.
15$. STOWELL, R.E.: Histochemical observations on nucleic acids in homologous normal and neoplastic tissues. Symp. Soc. Exper. Biol. 1: 190-205, 19k7.
156. SUNTZSFF, V., BURNS, E.L.. MOSKOP, M., and LOEB, L.: On the pro liferative changes taking place in the epithelium of the vagina and cervix of mice with advancing age and under the influence of experimen tally administered estrogenic hormones. Am. J. Cancer 32: 256-289, 1938.
157. SWIFT, H.H.: The desoxyribose nucleic acid content of animal nuclei. Physiol. Zool. 23: 169-198, 1950.
158. SWIFT, H.H.: Cytochemical techniques for nucleic acids. In: E. Chargaff and J. N. Davidson (eds.) The nucleic acids. New York: Aca demic Press, Inc. 1955.
159. TEUTSCHLANDER, K.: Experimentelle erzengung von "cholesteatom" und cancroid in uterus der ratte. Ztschr f. Krebsforsch. 23: 161-176, 1926.
160. TIMONEN, S.: Mitosis in normal endometrium and genital cancer. Acta Obst. & Gynec. Scandinav. 31: 6-88, 1950.
161. TIMONEN, S., and THERMAN, E.: The changes in the mitotic mech anism of human cancer cells. Cancer Research 10: U31-U39, 1950.
162. TSOU, K.C., CHENG, C-S., NACHLAS, M.M., and SELIGMAN, A.M.: Syn thesis of some p-nitrophenyl substituted tetrazolum salts as electron acceptors for the demonstration of dehydrogenases. J. Am. Chem. Soc. 78: 6139-61^, 1956.
163. VELLIOS, F., and GRIFFIN, J.: The pathogenesis of dimethyl- benzanthracene induced carcinoma of the cervix of rats. Cancer Research 17: 36U-366, 1957.
161*. WARBURG, 0. Metabolism of tumors. (Constable & Co., London, 1930)
165. WEINHOUSE, S., ALLEN, A., and MILLINGTON, R.H.: Metabolism of neoplastic tissue. V. Fatty acid oxidation in slices of transplanted tumors. Cancer Research 13: 367-371, 1953*
166. WILLIAMS, W.L., PAULLUS, K., and ERICKSON, C.C.: Proliferative epithelial changes and carcinoma of the cervix and vagina in mice with prolonged administration of alpha estradiol-sensitized collodion particles. Proc. Am. Assoc. Cancer Research 1: 60, 1953. 198
167. WILLIS, R.A.: The mode of origin of tumors. Solitary localized squamous cell growth of the skin. Cancer Research ij.: 630-61*i|., 19kh-
168. YAMAGIWA, K., and ICHIKAWA, K.: Experimental study of the patho genesis of carcinoma. Am. J . Cancer 3s 1-29* 1918. 159
AUTOBIOGRAPHY
I, Dante Giovanni Scarpelli, was born in Padua, Italy,
February 5, 192?. I received my secondary school education in the public schools of Cleveland, Ohio, and my undergraduate training at
Baldwin-Wallace College, which granted me the Bachelor of Science degree in 1990. From the Ohio State University, I received the
Master of Science degree in 1953. During this period my graduate work was carried out under the direction of Professor Ralph A. Knouff in the Department of Anatomy. In 195U I received the degree Doctor of Medicine from the Ohio State University. After serving a year of internship at the University Hospital I became an assistant to Professor
Emmerich von Haam in the Department of Pathology at Ohio State Univer sity. I was appointed as a post-doctoral fellow in Pathology by the
National Science Foundation for the period extending from 1955 - 1957*
I was then appointed a Senior Research Fellow in Pathology by the
United States National Institutes of Health for five years. While holding this appointment I have specialized in general and experimental pathology and have completed the requirements for the degree Doctor of
Philosophy.