The Possible Role of Smegma in Carcinoma of the Cervix^*+

^* This investigation was supported by Research Grant C-1418 from the National Cancer Institute of the National Institutes of Health, Public Health Service.
⁺ Presented at the Twentieth Annual Meeting of the South Atlantic Association of Obstetrictans and Gynecologists, Hollywood, Florida, Feb. 1 to 5, 1958.

This is an interim report of a study begun in 1950 to determine any possible role human smegma might play in cancer of the cervix. The preliminary report[1] was published elsewhere in 1956.

Several epidemiological studies of cervical cancer have recently been made,[2,3] investigating multiple social, hereditary, and environmental factors. Most evidence points to a correlation between cancer of cervix and marriage at an early age, unrepaired cervical lacerations, low economic status, and uncircumcised marital partner. Little definite statistical association between the number of pregnancies, douching with coal tar products, and diet has been obtained.

It has been reasonably well established that Jewish women have low rates of cervical cancer.[4,5] Most authorities agree that this is not due to the Mosaic laws for menstruation, as it is estimated that very few (5 per cent) Jewish women observe the laws of Nidah. Religious circumcision of the male is thought to be an important factor. Cancer of the penis has never been reported in a patient properly circumcised in infancy. This disease is also very rare in anyone circumcised during childhood.

Handley[6] has called attention to the low incidence of cancer of the cervix among Fijis. The Indians of the Fiji Islands who do not practice religious circumcision have an incidence eight times as great as that of the Fijis who do practice circumcision.

The observation that cancer very rarely occurs in nuns may indicate at least one compensation for permanent virginity. Homburger[7] feels this immunity may not be the result of abstinence from sexual intercourse but might be an indication of the need for stimulation of the cervical epithelium by pregnancy before cancer can develop.

Our results with the use of the Papanicolaou smear screening technique in one thousand consecutive clinic patients and one thousand consecutive private patients (Table I) are in agreement with those of most workers in this field. This threefold incidence of carcinoma in the clinic group over the private group supports the low-economic-group theory. We have noted that not only is the incidence of carcinoma higher, but the smears show a higher percentage of other cellular aberrations than is found among the private patients. This group is predominantly Negro (98 per cent) with coitus at an early age, unrepaired cervical lacerations, rand poor vaginal hygiene. It has been impossible to ascertain the number of uncircumcised sexual partners in this group, but it is undoubtedly high.

Table I. Papanicolaou smear screening technique in 1,000 consecutive service cases and 1,000 consecutive private cases

---

Service
	952 Negative Smears (cervical)
	1 case hyperplastic cells, negative biopsies
	47 cases questionable diagnoses
		8 cases doubtful cells
			3 doubtful, later biopsy, carcinoma
			3 doubtful, later biopsy chronic cervicitis
			2 doubtful, no follow-up
		14 cases suspicious cells
			7 suspicious, carcinoma
			4 suspicious, chronic cervicitis
			3 suspicious, no follow-up
		25 cases carcinoma cells
			22 carcinoma cells, carcinoma
			1 carcinoma cells, leukoplakia of cervix
			1 carcinoma cells, chronic cervicitis
			1 carcinoma cells, terminal cancer of breast
	Incidence, 3.2% carcinoma
Private
	984 negative cervical smears
	4 cases hyperplastic cells, negative repeat smears
	12 cases questionable diagnoses
		5 cases suspicious cells
			3 suspicious, negative follow-up
			2 suspicious, cancer established by surgery
		6 cases carcinoma cells, carcinoma
		1 case carcinoma cells, cancer of uterus
	Incidence, 0.9% carcinoma

---

Each of the preceding facts focused our attention on the role of human smegma. Most investigators[8] previously had used horse smegma because of the difficulty in obtaining the human variety. An abundant human supply «'as found in the South Carolina State Mental Hospital.

A great amount of work has been done on cancer in lower animals, the results of which are being used in attempts to interpret cancer in the human. Because a cancer was produced in one species of animal does not prove that it can be similarly produced in another. Sugar and Levy[9] stated that the belief that carcinoma of the cervix can be caused by exposure to an unproved carcinogenic agent at infrequent intervals cannot be condoned.

Spontaneous cervicaI cancer in mice is a rarity. Collins[10] and his group were the first to produce carcinoma of the cervix with any regularity in the mouse with local application of carcinogens. Murphy[11] has also reported a high incidence of cervical cancer in mice using the known carcinogen, methylcholanthrene. Using single insertions of crude smegma with ligation of the vagina, Twombly[12] produced one cervical cancer in a group of 48 mice, 12 Of which survived long enough to be evaluated.

It has been hypothesized that the bacterial content of smegma, particularly Mycobacterium smegmatis, might convert its cholesterol content into a carcinogen. Sobel and Plaut[13] found that Myco. smegmatis consumed cholesterol actively when this substance was in its solid phase or in solution as its succinate.

A chemical analysis of the residue of smegma after drying to constant weight over concentrated sulfuric acid is shown in Table II. At present the fractionation of raw smegma is being done by Dr. Ray Brown of the University of Wisconsin Cancer Research Hospital and the effects of each of these fractions studied in mice by the same methods employed in this study.

Table II. chemical analysis of the residue of smegma after drying to constant weight over concentrated sulfuric acid

Constituent	%
Total lipids	28.4
Phospholipids (as lecithin)	0.8
Chlorestrol (free)	3.8
Chlorestrol esters	1.0
Neutral fats	22.8
Total fatty acids	22.4
Iodine number of fatty acids	41.3

Materials and Methods

Whole raw human smegma obtained by retraction of the foreskin and the evacuation of this material from about the glans in uncircumcised males has been of paramount importance in this experiment. The bacterial and chemical components of smegma, the bacterial flora of the female genital tract, ovarian dormoid cyst contents, and simple trauma and estrogens have also been investigated as corollary and control aspects of this problem. Two basic methods have been employed: the repeated application of these materials to the cervix and upper vagina of the mouse by means of a metal speculum and the injection of a single "dose" into the vagina followed by closure of the vaginal orifice with sutures. The metal speculum was used in conjunction with an applicator which has a cupped end holding 3 mg. of the material to be tested.

The whole raw smegma was collected in sterile normal saline, centrifuged, ground into a paste, and stored in stoppered bottles in the refrigerator.

Female mice of the dba-1 strain were used at ages ranging from 7 to 20 weeks in the different groups with the addition of one group of white mice of the Wistar strain. The dba-l strain was selected for two reasons: (1) no spontaneous cervical cancer had been reported in it, and (2) there was a known incidence of hereditary breast carcinoma in the strain. It was felt desirable to use animals which were not completely refractory to neoplastic disease.

Results

The results in the various groups are summarized in Tables III and IV. Mice were deleted from final pathological evaluation if postmortem changes were extreme, if the animals had been partially devoured by cage mates, or if death occurred within a few days after they were committed to the experiment. Animals were not sacrificed unless obviously and seriously diseased, as it was believed rational to expose them to the various agents for as long a period of time as possible.

Table III. Summary of experimental results

Group	Method	Mice Injected	Suitable for Evaluation	Tumors
1	Weekly vaginal injection of smegma	41	26	0
2	Biweekly vaginal injection of smegma	15	12	6
3	Introduction of smegma with vaginal ligation	54	25	4
4	Introduction of smegma with vaginal ligation and injection of estrogen weekly	12	12	1
5	Introduction of smegma with vaginal ligation and biweekly injection of increasing amounts of estrogen	14	4	0
6	Weekly vaginal injection of smegma and injection of estrogen weekly	10	6	0
7	Simple ligation of vagina without smegma	12	10	0
8	Introduction of sterile smegma with vaginal ligation	13	6	0
9	Introduction of ovarian dermoid-cyst contents with vaginal ligation	14	12	2
10	Biweekly vaginal injection of dermoid-cyst contents	17	11	0
11	Injection of smegma into artificial skin tunnels	15	4	0
12	Biweekly vaginal injection of culture of 100% Myco. smegmatis	14	6	0
13	Biweekly injection of mixed bacterial culture corresponding to bacteriological analysis of smegma including 10% Myco. smegmatis	35	27	1
14	Biweekly injection of 100% Myco. smegmatis in culture with 5% chlorestrol	25	16	0
15	Biweekly injection of 0.3% methylcholanthrene in coconut oil	20	16	1
16	Biweekly injection of 0.5% methylcholanthrene in lard	10	10	1
17	Insertion of clean speculum and obturator	30	21	0
18	Raw smegma biweekly vaginal injections	50	25	1
19	Human vaginal culture	23	19	0
20	White mice (Wistar strain), raw smegma biweekly	10	6	0

 

Table IV. Summary of raw smegma groups

---

Total No. of dba-1 mice used for injection and ligation			160
No. suitable for final evaluation			88
Results:
	Epidermoid carcinoma	8
	Sarcoma	2
	Malignant papilloma	1
	Marked hyperplasia	22

---

After the impressive results found in Group 2 (the biweekly injection of raw smegma), this experiment was repeated. It was begun in May, 1956, on 50 animals (Group 18), of which only 6 remain alive at present (19 died early in study of "pneumonia"). Of the 25 available for study, there was no epithelial change in 7, mild to marked hyperplasia in 17, and one cancer of the cervix (Fig. 1). This tumor was in a mouse after 13³/₄ months of exposure to 118 injections.

In the groups of 88 mice in which whole raw smegma has been used by either the injection or ligation techniques, there have been 8 epidermoid carcinomas or an incidence of 10.2 per cent. If hyperplastic changes are included there are significant changes in 35.2 per cent (exclusive of the sarcomas). It is probable that many of the lesions placed in the hyperplastic category could have justifiably been classified as intraepithelial carcinoma (Figs. 2 and 3). In a detailed histological study of induced cervical carcinoma in the mouse, Scarpelli and von Haam[14] produced 139 malignant lesions in 261 C₃H mice, using either 3,4 benzpyrene or 20-methylcholanthrene. They classified these neoplasms as invasive in 72 animals and noninvasive in 67. It appears from study of their description and illustrations that we produced many lesions comparable to their noninvasive carcinomas. Indeed we have been impressed with the difficulty in differentiating between various stages of atypical hyperplasia and carcinoma, but in an experiment dealing with the testing of a material of unknown carcinogenic potency, we choose to err on the conservative side and to classify borderline states as hyperplasia.

Fig. 1. -- Squamous-cell carcinoma produced by applications of smegma biweekly for 14 months. (Hematoxylin and eosin. X250; reduced ¹/₅)

Fig. 2. -- Marked epithelial hyperplasia in mouse receiving applications of 100 per cent Myco. smegmatis with 5 Per cent cholesterol biweekly for 9 months. (Hematoxylin and eosin. X85; reduced ¹/₅)

Fig. 3. -- Illustration of the type of epithelial hyperplasia which was very difficult if not impossible to distinguish from intraepithelial carcinoma. (Hematoxylin and eosin; X85; reduced ¹/₅)

In Group 17, in which the clean speculum and obturator were simply repeatedly introduced into the vagina, no carcinoma developed. Two mice had marked hyperplasia of the cervicovaginal mucosa after 20 months (163 introductions).

Fig. 4. -- Marked epithelial hyperplasia with convoluted and reduplicated nests of cells. (Hematoxylin and eosin. X250; reduced ¹/₅)

The various groups (12, 13, 11, 19) in whieh bacteria were used supplied interesting results in all cases in which Myco. smegmatis was a component.

In Group 12, with 100 per cent Myco. smegmatis culture used biweekly, no invasive carcinomas were encountered, but 8 mice died almost immediately and 2 of the remaining animals exhibited intense epithelial hyperplasia (Fig. 4) after 10 months (63 injections) and 11 months (70 injections), respectively. These changes most probably represented carcinomatous transformation, but as they could not definitely be proved they were not designated as such.

In Group 13, in which mixed bacterial cultures were used corresponding to analysis of smegma and including 10 per cent Myco. smegmatis, there were even more pronounced results. One invasive carcinoma was found after 21¹/₂ months (141 injections). Two other mice showed intense hyperplasia, one with papilloma after 14³/₄ months (95 injections). These hyperplastic states were indistinguishable from intraepithelial carcinoma.

No carcinoma was found in Group 14 but in one case profound epithelial hyperplasia occurred at the end of 14 months (114 injections).

Cultures of the vagina were obtained from 25 patients and the bacterial flora analyzed as to the type of organism and per cent. An artificial culture was then recreated which had the following composition: Staphylococcus albus 40 per cent, Staphylococcus aureus 10 per cent, Escherichia coli 10 per cent, Aerobacter aerogenes 10 per cent, diptheroids 10 per cent, and Lactobacilli sp. 20 per cent. No carcinomas resulted and the hyperplastic changes were not as marked as those noted in the other groups. There was marked hyperplasia in one animal after 9¹/₂ months and 73 injections. This material proved to be quite lethal for the mice and it was difficult to keep them alive for a satisfactory length of time.

Biweekly injection of raw smegma into another strain of mouse in Group 20 did not produce carcinoma. There was one example of marked hyperplasia after 12¹/₂ months (104 injections). One animal is still alive.

Comment

The clinical and sociological facts and observations concerning smegma strongly indicate that in some way it is implicated in the genesis of penile and cervical carcinoma. Whole raw smegma as well as some of its components is stimulatory to the cervico-vaginal epithelium of mice and invasive carcinoma will eventuate if the stimulus persists. It must be realized that the epithelium of the lower genital tract of mice is labile and probably reacts proliferatively to a wide variety of stimuli. Whole raw smegma, however, exclusive of known carcinogens, has proved to be the most effective stimulus in this experiment, followed by either some of its components or closely allied substances.

If lack of circumcision should prove to be of significance in the development of cervical cancer, a gradual reduction of this type of cancer may be expected in the United States. Wynder[2] has noted the circumcision rate in groups of hospitals studied to be 8O to 85 per cent of the non-Jewish males.

As Younge[15] has recently stressed, celibacy and late marriage will never be popular methods of the control of cervical cancer. He enthusiastically feels that invasive cancer of the cervix can be prevented by widespread advocation of circumcision, care of the abnormal-appearing cervix, and early detection of intraepithelial cancer.

Epidermoid cancer of the cervix has been noted in women exposed only to circumcised males and in virgins. Other etiological factors than those involving coitus and lack of circumcision must therefore exist.

Conclusions

1. The circumstanial evidence of the relationship between carcinoma of the cervix in women and smegma has been discussed.

2. In dba-1 strain mice, caieer of the cervix can be produced by human smegma, if this stimulus is continued for 14 months or more.

The objections to the use of lower animals in reference to human cancer are obvious. A similar study with the use of an experimental animal of primate level is contemplated.

The technical assistance of Miss Ann Clark and Mr. Frank Wyman is gratefully acknowledged.

We also wish to thank Frank Cordle of the Department of Microbiology and Robert Brown and Ethel MciVIinn of the Department of Medical Illustration for their cooperation.

References

1. Pratt-Thomas, H. R., Heins, H. C., Latham, E., Dennis, E. J., and McIver, F. A.: Cancer 9: 671, 1956.
2. Wynder, E. L., Cornfield, J., Schroff, P. D., and Doraiswami, K. R.: Am. J. Obst. & Gynec. 68: 1016, 1954.
3. Lombard, H. L., and Potter, E, A.: Cancer 3: 960, 1950.
4. Weiner, L., Burke, L., and Goldberger, M. A.: Am. J. Obst. & Gynec. 61: 418, 1951.
5. Kennaway, E. L.: Brit. J. Cancer 2: 177, 1948.
6. Handley, W. S.: Lancet 1: 987, 1936.
7. Homburger, F.: The Biologic Basis of Cancer Management, New York, 1957, Hoeber-Harper.
8. Plaut, A., and Kohn-Speyer, A. S.: Science 105: 391, 1947.
9. Sugar, M., and Levy, W. E.: New Orleans M. & S. J. 103: 424, 1951.
10. Collins, V. J., Gardner, W. U., and Strong, L. C.: Cancer Res. 3: 29, 1943.
11. Murphy, A.: Am. J. Path. 29: 608, 1953.
12. Twombly, G. H.: Personal communications.
13. Sobel, H., and Plaut, A.: J. Bact. 57: 377, 1949.
14. Scarpelli, D. G., and von Haam, E.: Am. J. Path. 33: 1059, 1957.
15. Younge, P. A.: Obst. & Gynec. 10: 469, 1957.