Effects of the Antifertility Drug Enovid in Five Strains of Mice, With Par- ticular Regard to Carcinogenesis 1
W. E. Heston, G. Vlahakis, and B. Desmukes, Laboratory of Biology, National Cancer Institute,2 Bethesda, Maryland 20014
SUMMARY-The antifertility drug, Enovid, was tested for possible carcinogenicity in female mice of 5 specially selected strains: BALB/c, C3H, C3HIB, A, and C57BL. The drug was fed at 3 dose levels: 5 µg Enovid/g food that did not prevent reproduction, 10 ug/g that prevented some females from reproducing, and 20 ug/g that prevented all females from reproducing. The results emphasized that inbred strains of mice differed in their response to Enovid. Enovid treatment reduced the gain in weight in all strains. It had no effect on lifespan in 1 strain, but in 2 strains lifespan was lengthened and in 2 shortened because of the effect on tumors or other lesions. Cervical and vaginal lesions showing invasion of the epithelium into the stroma with varying degrees of progression were observed but limited, with few exceptions, to the BALB/c females. They occurred in control and treated females of this strain but with a higher incidence and possibly further progression in the group treated with the highest dose of Enovid. None of these lesions were observed grossly as tumors, all being found on histologic examination, and none had extended to organs and tissues beyond the vaginal wall or showed any evidence of having metastasized. Enovid treatment at these dose levels did not increase the occurrence of ovarian tumors, even in the highly susceptible strains C3H and C3HIB, or of mammary gland tumors in any strain. In fact, in the highly susceptible strain C3H mammary tumors were somewhat inhibited in the treated females. In strain C3HIB there was some inhibition of hepa- tomas by the treatment, probably related to the reduction in growth rate, and in BALB/c some inhibition of adrenocortical adenomas. Chromophobe adenomas of the hypophysis were significantly increased in old C57BL females treated with the highest dose of Enovid. Many of these were active in stimulating the mammary glands of these old females and caused the females to have to be necropsied. These results are discussed in the light of previous work on hormonally induced neoplasms in which usually relatively massive doses were given and of results of other tests for carcinogenicity of contraceptive drugs .- J Natl Cancer Inst 51: 209-224, 1973.
POSSIBLY no drugs have been administered more widely to human beings with less agreement as to their carcinogenic potential than have the anti- fertility drugs. Physicians have given the drugs with the counsel that in general they are safe, based on their own observations of women taking the “pill” and on other clinical reports (1). Yet they are forced to admit that since one can expect a long latent period for hormonally induced tu- mors, there really has not been time to gather ad- equate data from these women to be certain that neoplasms will not be induced.
The experimentalist has had too much experience with induction of tumors with hormones in labor- atory animals to share this feeling of safety [for re- view see Gardner et al. (2)]. Even a disturbance of
the hormonal balance may result in certain tumors of the endocrine system (3). This had led cautious investigators such as Hertz and Bailar (4) to plead for further clinical observations, follow-up under controlled conditions, and more experimental data.
Many reports of tests of the antifertility drugs for carcinogenicity have appeared, but the results have been suggestive rather than conclusive. Rea- soning that norethynodrel, a primary constituent of many antifertility drugs, would stimulate the hypophysis to additional secretion of prolactin, Kahn and Baker (5) injected the hormone into
1 Received January 22, 1973; accepted March 29, 1973.
2 National Institutes of Health, Public Health Service, U.S. Department of Health, Education, and Welfare.
A/J and C3H/HeJ female mice and observed an increase in occurrence of mammary tumors over that in untreated controls. However, the incidence in the C3H controls was somewhat less than might have been expected from data from other labo- ratories. Poel (6) observed hypophyseal adenomas in C57L female mice given an antifertility drug, but some old C57L females develop such tumors without treatment. Dunn (7) reported lesions of the cervix that she diagnosed as cancer in BALB/c female mice fed a liquid diet, Metrecal, in which the antifertility drug, Enovid, was suspended. The neoplasms were not large, they had not metasta- sized, and none was transplanted. Metrecal is an unusual diet for mice and at that time it contained cyclamate. However, she did not report these lesions in mice fed Metrecal alone. Lipschutz et al. (8) observed ovarian tumors in BALB/c mice treated with two 19-nor-contraceptives in choles- terol pellets implanted subcutaneously. Most of these were microtumors, though described as structurally identical with large tumors. On the other hand, Dunn did not observe ovarian tumors in BALB/c mice given Enovid orally.
It was clear that further work was needed to try to confirm or extend these observations. A problem with such worldwide importance merited more conclusive experimental data.
PROCEDURE
Oral contraceptive .- Enovid was chosen for testing, since it is one of the most widely used of the oral contraceptives. The drug was obtained through the courtesy of Dr. John K. McGowan of Searle and Company as a powder mixture of the progesten, norethynodrel, and the estrogen, mes- tranol, in the same ratio as they exist in the Enovid 10 mg tablets, i.e., 98.5:1.5.
The Enovid powder was mixed in the National Cancer Institute (NCI) diet that we have fed to our mice. The formula of this diet has been published (9). For thorough distribution, the Enovid powder was first mixed for approx- imately 16 hours in a small amount of lactose as a carrier and this in turn was combined with the NCI diet meal in a Hobart mixer for approximately 8 hours. The meal was then run through a mill that pressed it into pellets exactly like the NCI food pellets.
Doses of Enovid .- Three doses of Enovid were fed. The Enovid I diet contained 5 ug Enovid/g food; Enovid II, 10 µg/g; and Enovid III, 20 µg/g. Assuming that a mouse eats 2 g of food per day, which is roughly what these mice ate of the NCI pellets, the daily doses of Enovid would be
10, 20, and 40 µg. The experiment was started with Enovid I diet but, when females fed this dose bred, Enovid II diet was introduced with double the dose. When even some fe- males fed Enovid II bred, Enovid III diet was introduced with a dose twice that in II. Except for 2 C3H females on Enovid III diet for only 1 month that became pregnant, the dose in Enovid III prevented pregnancy. This dose for the mice fed Enovid III diet was on a body-weight basis approximately that administered to women.
MOUSE STRAINS
Five strains of mice provided maximum genetic varia- tion in the test animals. They were also selected as appro- priate for the induction of specific kinds of tumors and for confirmation of earlier observations of other investigators. These strains were:
BALB/cHe .- This strain has been in our laboratory since its receipt from Dr. H. B. Andervont’s colony in 1966. It is highly susceptible to the mammary tumor virus (MTV). Originally the strain was studied in this laboratory by Deringer (10) who found that, without the introduction of MTV, breeding females had an incidence of 20% mammary tumors, most of which were adenoacanthomas. This in- cidence has continued in the colony to the present (11).
BALB/c was chosen primarily because Dunn (7) ob- served uterine cervical epithelial changes when mice of this strain were fed liquid Metrecal containing Enovid. Gardner and Allen (12) and Gardner (13) previously induced lesions of the cervix and vagina in mice of this strain with long-continued administration of estrogens. BALB/c mice are also susceptible to adrenocortical adenomas and thus suitable for testing a possible effect on this tumor. Lipschutz et al. (8) also observed ovarian tumors after treatment with 19-nor-contraceptives.
C3H/He .- This is a well-known high-mammary-tumor strain. We have used it for work in this area since 1941 when we obtained it from Andervont. It was introduced late in the present study when we realized that our line of strain A would not give an adequate test for induction of mammary tumors. Kahn and Baker (5) reported an increase in mammary tumors from norethynodrel in C3H mice. Since these mice are also susceptible to the induction of ovarian tumors, this was a good strain for comparison with the results of Lipschutz et al. (8).
C3HfB/He .- This is the original so-called mammary- tumor-agent-free strain we originated in 1945 from a C3H litter born by cesarean section and foster-nursed on a C57BL female (11). From 30 to 40% of the breeding females develop mammary tumors at an advanced age, presumably due to the NIV, a line of the MTV that appears to be transmitted through the germ cells rather than through the milk. Since the virgin females have a low incidence of mammary tumors, this strain also was especially suitable for testing for induction of these tumors. Like C3H, the C3HfB also would be susceptible to the induction of ovarian tumors.
A|He .- This strain, brought from The Jackson Labora- tory, Bar Harbor, Maine, in 1940, was started from breeders received from Dr. J. J. Bittner in 1938 (11). It has been
JOURNAL OF THE NATIONAL CANCER INSTITUTE
used extensively as a high-lung-tumor strain but was chosen because it was thought especially suitable for testing for induction of mammary tumors. Breeding A females have a medium incidence of mammary tumors, whereas the incidence in the virgins is low. Unfortunately the MTV was lost from the line about the time this study started, so that few mammary tumors occurred, even in breeding females.
C57BL|He .- This strain has been in our laboratory since 1945 when we obtained a breeding nucleus from Andervont (11). The strain is resistant to MTV and to the development of mammary tumors. It was chosen for this testing, since it is susceptible to the induction of chromophobe adenomas of the hypophysis following long-term estrogen stimulation. Spontaneous hypophyseal adenomas also occur in the strain. C57BL is related in its origin to strain C57L in which Poel (6) observed some adenomas of the hypophysis after treatment with an antifertility drug.
Outline of study .- The study was planned for each of the 3 dose levels of Enovid to be tested on 56 females of each of the 5 strains except the C3H that was used to test only the high dose. For each group of test females there was an equal number of littermate controls. At the weaning age of 4 weeks, they were set up in cages with 8 females to the cage and from that time on the test females were fed continuously the NCI pellets containing Enovid in the respective doses; the controls were fed the NCI pellets without the Enovid. Tap water was available at all times. The animal room was maintained at 78°±1º F and light was constant at 12 12 hours/day from 7 a.m. to 7 p.m.
Mice of the first two cages of each strain and each dose level, along with their controls, were weighed weekly. Sample females of the third cage were tested for fertility and all other females were kept as virgins. Animals were checked regularly and autopsied when tumors appeared or when the animals were moribund. The tumors observed grossly were taken for histologic examination, as were also the adrenal glands, ovaries, uterus, vagina, and hypophysis. The tissues were fixed in Fekete’s modification of Tellyesniczky’s solution (70% alcohol, 20 parts; formalin, 2 parts; glacial acetic acid, 1 part), sectioned, and stained with hematoxylin and eosin.
RESULTS Fertility Tests
After having been on the Enovid I diet for 2 months, to test their fertility 4 A females were placed with fertile A males. They were checked daily for vaginal plugs as an indication of having mated and later were checked for pregnancy. One showed no evidence of having mated, two were found with plugs but did not become preg- nant, and the fourth became pregnant but aborted. Four A females on the Enovid II diet for a com- parable time were also placed with fertile males.
One showed a plug but did not become pregnant, and the other three showed no evidence of having mated. It was, therefore, assumed without testing that the strain A females on the Enovid III diet were not fertile.
All of 4 BALB/c females on Enovid 1 diet for 132 months when placed with fertile BALB/c males became pregnant. They gave birth to litters of 7, 7, 7, and 8 offspring and weaned 6, 6, 5, and 7. Of 4 BALB/c females on Enovid Il diet for a comparable time when placed with fertile males, 2 showed no evidence of having mated, 1 showed a vaginal plug but did not become pregnant, and the fourth became pregnant twice. With the first pregnancy she gave birth to 4 offspring, none of which lived to weaning age, and with the second pregnancy she gave birth to 1 that lived to weaning age. Of the 4 BALB/c females fed the Enovid III diet for 1}2 months and placed with fertile males, 3 showed no evidence of mating and the fourth had a vaginal plug but did not become pregnant.
All 4 C3HfB females placed with fertile C3HfB males after 112 months on Enovid I diet became pregnant. Each of 2 had litters of 4 and weaned 4, one had a litter of 7 and weaned 7, and 1 had 2 stillborn. Four C3HfB females were placed with fertile males after 3 months on the Enovid II diet. Three showed no evidence of mating and one became pregnant, but the young were born prematurely and eaten. It was thus evident without testing that C3HfB females on Enovid III diet probably would not breed.
Four C57BL females on Enovid I diet for 2 months were placed with fertile C57BL males. One did not evidence mating, two had vaginal plugs but did not become pregnant, and one became pregnant and gave birth to four stillborn. After these results on the lowest dose we did not try to mate any C57BL females on Enovid II or Enovid III diets.
Four C3H females on Enovid III diet for only 1 month were placed with fertile C3H males. Two showed no evidence of having mated, and the other two became pregnant. One of these died while aborting and the other aborted dead fetuses. Four more C3H females were placed with fertile males after they had been on the Enovid III diet for 2 months, but three showed no evidence of
having mated and the fourth had a vaginal plug but did not become pregnant.
Thus, although there was some strain variation, this dose range gave us a low dose on which most of the females still would breed, a medium dose with which some of the females bred, and a high dose that prevented breeding in all strains at least after 2 months on the treatment.
The outstanding detrimental effect on breeding was that those females becoming pregnant while on Enovid treatment, particularly on diet II, tended to abort. This might have been prevented, however, had the Enovid been discontinued when the females were found to be pregnant. Preliminary breeding tests were carried out on offspring of Enovid-treated females, but no abnormalities in their breeding behavior or their offspring were observed.
Growth
From the weight data of the groups of 16 mice of each of the 5 strains on each of the 3 diets and their controls that were weighed weekly, the average growth curves shown in text-figure 1 were constructed. Mice of all groups gained weight. Those mice fed Enovid III diet gained the least, those fed Enovid II more than those fed Enovid III, and those fed Enovid I more than those fed Enovid II. The mice fed the control diet gained the most weight. This would indicate it was the Enovid itself that decreased the weight gain rather than some other quality of the diet, such as possible variation in hardness of the pellets that might have occurred, since the Enovid diets were pelleted in our mill. Strains C3H, C3HfB, and A that gained a large amount on the control diet showed greater difference in weight gain between the controls and the treated groups than did strains BALB/c and C57BL that gained less on the control diet. Except for strain A females fed Enovid II and III diets that lost some weight during the Ist and 2d weeks, all groups gained steadily while fed the diets containing Enovid.
Lifespan
The relationship between treatment with Enovid and lifespan of the animals varied among the strains.
30
BALB/c
O HHH
25
20
15
C
40
C3 H
35
II
30
AVERAGE WEIGHT - GRAMS
25
C
20
I
15
30
C3 HfB
II
25
II
20
15
c
30
A
25
I
JI
20
15
OHHE
20
C 57 BL
15
10
5
10
15
20
25
30
35
TIME FED ENOVID - WEEKS
There was no effect of Enovid on the lifespan of BALB/c females (table 1). Some treated BALB/c females, especially those on the highest dose of Enovid, showed evidence of polyuria. They drank an excessive amount of water and their cages were wet soon after being cleaned. The polyuria ap- peared soon after treatment began, but after 4-6 weeks the females recovered. In some females, stones were deposited in the urinary bladder, but they also occur in untreated, old BALB/c females. Aside from the polyuria and the stones in the bladder, the treated females remained in good health until their terminal poor condition. All groups lived to an old age averaging from 616 to 655 days, with the average age of each treated group approximately that of the respective control group. There was some reduction of age of the groups with higher doses of Enovid, which it would seem had been due to the polyuria in these
JOURNAL OF THE NATIONAL CANCER INSTITUTE
| Group | Total No. of mice | Av age at death (days) | Epithelial lesions of cervix and vagina | Adreno- cortical adenomas | Ovarian tumors | Mammary gland tumors | Lung tumors | Neoplasms of retic- ular tissue | Other neoplasms |
|---|---|---|---|---|---|---|---|---|---|
| Enovid I | 52 | 655 | 10 | 14 | 2 | 6 | 13 | 1 Islet cell tumor of pancreas. 1 Epidermoid carcinoma of lip. 3 Hemangioendothelioma. | |
| Control I. | 55 | 651 | 18 | 22 | 1 | 1 | 18 | 13 | 1 Adenocarcinoma of stomach. 1 Carcinoma of intestine. 3 Sarcoma of uterus. 1 Fibroblastic tumor of vagina. 2 Hemangioendothelioma. 1 Harderian gland tumor. |
| Enovid II | 50 | 640 | 17 | 7 | 2 | 2 | 11 | 7 | 1 Hemangioendothelioma. 2 Harderian gland tumor. |
| Control II | 53 | 648 | 16 | 11 | 2 | 15 | 6 | 1 Fibroblastic tumor of uterus. 2 Sarcoma of uterus. 5 Hemangioendothelioma. 1 Adenoma of uterus. 1 Papilloma on tail. | |
| Enovid III | 55 | 621 | 32 | 3 | 5 | 4 | 3 | 3 | 1 Islet cell tumor of pancreas. 2 Harderian gland tumor. |
| Control III | 55 | 616 | 18 | 10 | 3 | 3 | 12 | 8 | 2 Harderian gland tumor. 4 Hemangioendothelioma. |
TESTING ENOVID FOR CARCINOGENICITY IN MICE
groups, had there not been a similar reduction in their respective control groups.
In strain C3H the average lifespan of the females treated with Enovid was higher than that of their littermate controls (table 2). This, however, was because the death of most females in this strain is caused by mammary tumors and they occurred more frequently and earlier in the controls than in the treated females.
C3HfB like BALB/c is a long-lived strain, but in C3HfB there was a slight reduction in lifespan in the Enovid groups (table 3), the significant difference being between the group treated with the highest dose of Enovid and its control group (01=±90, 02=±70, t=2.158; 0.02<P<0.05). There is no apparent explanation of how the Enovid treatment has caused this reduction.
In strain A an effect of Enovid treatment on lifespan opposite that in C3HfB was noted. The two groups on the lower doses of Enovid lived longer than their respective controls (table 4; Enovid I vs. control I, 1=±73, 02=±80, t= 3.821; P<0.01, Enovid II vs. control II, 1=±66, 02=±72, t=2.812; P<0.01). Nephritis with its preceding amyloidosis is the primary cause of death in strain A, especially in the absence of mammary tumors, and it can be delayed or reduced by factors such as calorie restriction that reduce growth. The reduced growth of these two groups of treated mice was, therefore, probably responsible for increasing their lifespan by delaying the onset of nephritis. The group treated with the highest dose of Enovid had a greatly reduced lifespan, but since the lifespan of their controls was reduced in the same amount, this must have been due to some factor other than the hormonal treatment.
C57BL is also a long-lived strain. None of the treated groups of this strain, however, lived as long on the average as their respective controls (table 5), but, the difference between the group treated with the highest dose of Enovid and its
control group is not as highlysignificant (ơ1=±106, 02=±137, t=2.092; 0.02<P<0.05) as the differ- ence between the group treated with the lowest dose of Enovid and its control (1=±147, 02= ±115, t=3.384; P<0.01). The difference be- tween the group treated with the intermediate dose and its control group is not statistically significant. Any real reduction in lifespan in this strain associated with the hormone treatment was probably because of the higher incidence and/or earlier development of hypophyseal adenomas described later in the treated groups.
Neoplasms
Strain BALB/c .- Primary interest was focused on the epithelial lesions of the uterine cervix and vagina because of Dunn’s earlier report (7) of induction of carcinoma of the uterine cervix in this strain with Enovid fed in Metrecal. Similar epithelial lesions were observed in a rather high number of the females of this study. As Suntzeff et al. (14) observed many years ago, these lesions displayed various gradations from simple down- growth of the epithelium into the stroma, usually with atrophy of the surface epithelium, through those having some dyskeratosis with keratin pearls to those lesions that appeared as early carcinomas, but there were not clear lines of distinction between these gradations. None of these lesions was ob- served grossly as tumors. All were found during the histologic examination of the sections, as Dunn had apparently found those in her study.
Table 1 shows that the lesions occurred in the controls as well as the treated groups, with inci- dences not significantly different from those in the two groups treated with the lower doses of Enovid. In the group treated with the highest dose of Enovid, however, the incidence of these lesions was significantly higher than that in the control group (x2=7.187; P<0.01).
| Group | Total No. of mice | Av age at death (days) | Mammary gland tumors | Ovarian tumor | Adreno- cortical adenoma | Hypophy- seal tumor | Hepatomas |
|---|---|---|---|---|---|---|---|
| Enovid III. | 53 | 274 | 39 | 1 | |||
| Control III | 55 | 244 | 55 | 1 | 1 | 2 |
| Group | Total No. of mice | Av age at death (days) | Ovarian tumors | Mam- mary gland tumors | Epithelial lesions of cervix and vagina | Adreno- cortical adenomas | Hypo- physeal tumors | Lumg tumors | Hepatomas | Other neoplasms | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Enovid I | 56 | 626 | 15 | 6 | 6 | 13 | 6 1 1 1 | Fibrosarcoma. Sarcoma of vagina. Harderian gland tumor. Reticulum cell neoplasm. | |||
| Control I | 54 | 641 | 13 | 6 | 2 | 1 | 1 | 5 | 46 | 1 Sarcoma of uterine cervix. 1 Adenoma of uterus. 1 Papilloma. 2 Hemangioendothelioma. 5 Sarcoma. 2 Lymphocytic leukemia. 2 Adrenomedullary adenoma. | |
| Enovid II | 53 | 626 | 11 | 3 | 1 | 4 | 1 2 1 1 1 | Sarcoma of vagina. Reticulum cell neoplasm. Adrenomedullary tumor. Hemangioendothelioma. Harderian gland tumor. | |||
| Control II | 52 | 637 | 12 | 4 | 3 | 4 | 1 | 3 | 34 | 4 1 1 | Sarcoma. Hemangioendothelioma. Reticulum cell neoplasm. |
| Enovid III | 52 | 600 | 15 | 3 | 3 | 1 | 4 | 3 | 1 2 1 | Adrenomedullary tumor. Hemangioendothelioma. Lymphocytic leukemia. | |
| Control III | 50 | 635 | 18 | 7 | 2 | 6 | 37 | 4 Sarcoma. 2 Hemangioendothelioma. 2 Harderian gland tumor. | |||
HESTON, VLAHAKIS, AND DESMUKES
| Group | Total No. of mice | Av age at death (days) | Mam- mary gland tumors | Ovarian tumors | Epithe- lial lesions of cervix and vagina | Adreno- cortical adenoma | Hypo- physeal tumor | Lung tumors | Other neoplasms |
|---|---|---|---|---|---|---|---|---|---|
| Enovid I | 57 | 585 | 2 | 29 | 1 Sarcoma of uterus. 2 Reticulum cell neoplasm. 1 Hepatoma. | ||||
| Control I | 61 | 532 | 1 | 1 | 1 | 28 | 1 Reticulum cell neoplasm. 1 Myoepithelioma. 1 Lymphocytic leukemia. 1 Tumor of the pars intermedia of the hypophysis. | ||
| Enovid II | 63 | 557 | 3 | 1 | 30 | ||||
| Control II | 59 | 522 | 1 | 21 | 1 Reticulum cell neoplasm. 1 hepatoma. | ||||
| Enovid III | 54 | 477 | 1 | 1 | 1 | 16 | 4 Hemangioendothelioma. 1 Adenomatous polyp of uterus. | ||
| Control III | 49 | 478 | 17 | 1 Lymphocytic leukemia. |
| Group | Total No. of mice | Av age at death (days) | Hypo- physeal tumors | Neo- plasms of reticular tissue | Ovarian tumors | Epithelial lesion of cervix and vagina | Other neoplasms | |
|---|---|---|---|---|---|---|---|---|
| Enovid I | 47 | 602 | 21 | 15 | 2 | 1 Lung tumor. 1 Adrenomedullary tumor. 1 Hemangioendothelioma. | ||
| Control I. | 50 | 691 | 17 | 11 | 1 | 6 Hemangioendothelioma. 1 Adrenomedullary tumor. 1 Harderian gland tumor. | ||
| Enovid II | 53 | 607 | 32 | 5 | 1 | 1 | 1 Adrenomedullary tumor. 1 Harderian gland tumor. | |
| Control II | 50 | 633 | 17 | 12 | 4 | Hemangioendothelioma. | ||
| Enovid III | 49 | 567 | 36 | 4 | 1 | 1 1 | Mammary gland tumor. Plasma cell tumor. | |
| Control III | 51 | 617 | 15 | 10 | 1 Papilloma. | |||
| 1 Hepatoma. | ||||||||
| 1 Hemangioendothelioma. 1 Tumor of the pars intermedia of the hypophysis. | ||||||||
To ascertain whether Enovid treatment had influenced the progression of these epithelial lesions, we classified all in grades 1 through 3, much as Gardner and Allen (12) did, illustrating the various grades by photomicrographs. Grade 1 included those showing early downgrowths of epithelium into the stroma; grade 2, further epithelial prolif- eration with areas of dyskeratosis and pearl forma- tion; and grade 3 with even more extensive proliferation of the epithelium, though the lesion was still confined to the vaginal wall (figs. 1-4). None of these lesions extended into tissues and organs beyond the vagina showing definite evi- dence of malignancy, as Gardner and Allen observed in theirs classified as grade 4 or as others have observed after treatment with higher doses of estrogen, with carcinogenic hydrocarbons or a combination of these treatments. There was no evidence that any of these lesions had metastasized, and, since they were observed only in sections, none was transplanted. The listing in table 6 shows that lesions of all three grades occurred in the con- trols as well as in the treated mice, with some indication that treatment with the highest dose of Enovid had slightly increased the progression.
Thus these epithelial lesions of the cervix and vagina occur in old BALB/c females. Treatment
| Group | Total No. of lesions | Grade | ||
|---|---|---|---|---|
| 1 (number) | (number) 2 | (number) 3 | ||
| Enovid I | 10 | 7 | 3 | |
| Control I | 18 | 11 | 4 | 3 |
| Enovid II | 17 | 11 | 6 | |
| Control II | 16 | 9 | 7 | |
| Enovid III | 32 | 14 | 12 | 6 |
| Control III | 18 | 12 | 4 | 2 |
with a sterilizing dose of Enovid significantly increased the occurrence of the lesions and may have advanced their progression. In none of the control or treated females had the lesions progressed to a high degree of malignancy as evidenced by extension beyond the vaginal wall or metastasis. Nevertheless, the lesions are of interest and deserve further study to define more clearly the level of treatment at which they can be induced to form malignant, metastasizing fneoplasms. The study of the lesions in untreated emales might contribute toward a better understanding of carcinoma in situ in women. Furthermore, this strain of mouse might provide a model system for extending into experi- mental animals the observations of Herbst and
co-workers (15) of the appearance of adeno- carcinoma of the vagina of young women whose mothers had been treated with stilbestrol started during the first trimester to maintain pregnancy. There is a difference, however, in that the neo- plasms in these young women apparently arose from rudimental müllerian duct tissue that had been caused to persist in the vaginal wall, whereas the lesions in the BALB/c mice arose from the epithelium of the vagina and cervix.
Other neoplasms in the BALB/c groups are listed in table 1. The small adrenocortical adenomas to which this strain is susceptible were most frequent, but there was no evidence that these adenomas were induced by the Enovid treatment. In fact, they were observed more frequently in the control than in the treated groups. There was also no evidence that the occurrence of any of the other neoplasms, including the few ovarian and mammary tumors, was influenced by Enovid treatment up to the sterilizing dose. These mammary tumors were of the same types as previously observed in the strain (10), including adenocarcinomas of types B and C and adenoacanthomas.
Hyalinization of the uterine endometrial stroma, considered a characteristic of estrogen-treated mice, was observed in the BALB/c females on the highest dose of Enovid, and in some of those on the medium dose, but not in those on the lowest dose. In general this was also true for A and C57BL females, but C3H and C3HfB females were resistant to such hyalinization with these doses.
Strain C3H .- Tumors in C3H females treated with Enovid and in their controls are listed in table 2. In contrast with the observation by Kahn and Baker (5) that norethynodrel increased mam- mary tumors in C3H, treatment with a sterilizing dose of Enovid reduced their occurrence. All the control C3H females developed mammary tumors at the expected average young age of 244 days, whereas only 39 of the 53 treated females developed mammary tumors and they arose at a somewhat older average age of 274 days. These results may have been related to growth. As noted in text-figure 1, while the treated females grew after the treatment began, they did not grow nearly as much as did the control females. Many factors reducing growth reduce occurrence of various
kinds of tumors, including mammary tumors [for discussion of this subject see (16)].
Since most females of these groups were autop- sied with mammary tumors at a relatively young age, few other tumors were observed and all of these except an hypophyseal adenoma were in the controls. Although the hypophyses were not weighed, all those of the treated and control females of this high-mammary-tumor strain C3H and also of strain C3HfB were noticeably larger than the nontumorous ones of BALB/c, A, and C57BL. Yet they were not tumorous except for the three so listed.
Strain C3HfB .- Since these females did not get as many mammary tumors as did the C3H females and thus lived much longer, they displayed a greater variety of tumors (table 3).
This was the best strain for study of the possible effect of Enovid on the occurrence of ovarian tumors, since the strain is relatively susceptible to both spontaneous ovarian tumors and those induced with stilbestrol and radiation. It also provided data that can be compared with those of Lipschutz et al. (8) who reported an increase in ovarian tumors in BALB/c mice with antifertility drugs. Results in table 3, however, show that while in the present study ovarian tumors, most of which were tubular adenomas with an occasional granulosa cell tumor, occurred in a relatively high number of the females receiving Enovid, they were not any more fre- quent in those on the highest dose than in those on the lower doses or any more frequent in the treated groups than in the control groups.
This responsive strain is also appropriate for testing for any effect of Enovid on the occurrence of mammary tumors. Hormonal stimulation from having litters increases the occurrence of mammary tumors in the females (17). Furthermore, estrogen- cholesterol pellets induce mammary tumors in the males of the strain (18). However, treatment with Enovid did not increase the occurrence of mam- mary tumors in this strain (table 3). In fact, these tumors were in fewer females of each treated group than in their respective controls. The mammary tumors were of the types usually found in this strain.
Of the various other tumors in C3HfB, none showed any significant difference between the treated groups and the controls except for the hepatomas. Many more females had hepatomas
among the control groups than among the re- spective treated groups. This difference may have been due in part to the fact that the controls lived a little longer and in part to the greater average weight of the controls, since almost anything that decreases growth decreases occurrence of hepatomas in mice (16, 19). Those females receiving the highest dose of Enovid with the least gain in weight also had fewer hepatomas than those treated with the lowest dose. Whether there was any direct effect of Enovid in reducing the hepatomas in this strain is problematic. Estrogen-cholesterol pellets have been shown to decrease the occurrence of hepatomas in males of this strain to more nearly the incidence in females (18).
Strain A .- Tumors in this strain are listed in table 4. Strain A was included because it was expected to be particularly suitable to test for an increase in mammary tumors with Enovid treat- ment. Virgin females have an incidence of mam- mary tumors of less than 5%, but with added hormonal stimulation, such as that from breeding, this is raised to a relatively high incidence. How- ever, soon after the beginning of this test it was noted that most of our strain A colony had lost the MTV. Only one control female in all three groups had a mammary tumor and it was an adenoacanth- oma. Three in the Enovid II group had mammary tumors. Two were type-B adenocarcinomas and arose in two of the four females placed with males to test whether or not they would breed. These two, however, had no litters. The third tumor in this group was an adenoacanthoma. One female in the Enovid III group had an adenoacanthoma of the mammary gland. Thus a few more treated females had mammary tumors than the controls, but there is not conclusive evidence that they were due to the Enovid treatment. Some of these females with mammary tumors may have had the MTV. In no group were there more mammary tumors than could have been expected in virgin females of this strain with the virus.
There was no significant difference between treated and control groups of this strain with respect to any other tumors observed. The strain is susceptible to pulmonary adenomas, but there was no evidence that they were increased by the treat- ment. Fewer of the females in the Enovid III group
and their controls had pulmonary tumors than of the other groups, but this was because of their unexplained reduced lifespan. The hypophyseal tumor in control I group was unusual in being of the pars intermedia. The one in the Enovid III group was a chromophobe adenoma of the anterior lobe like the type observed in other strains in this study.
Strain C57BL .- As previously stated, the prin- cipal purpose for using strain C57BL was to test whether Enovid administered in doses up to the sterilizing dose would increase the occurrence of tumors of the hypophysis. C57BL mice are rela- tively susceptible to chromophobe adenomas of the anterior lobe of the hypohysis, the usual tumor of this organ in the mouse. As noted in table 5, these chromophobe adenomas occurred in approximately one-third of the controls, but treatment with Enovid increased their occurrence with increase in dose. The incidence in the females treated with the highest dose of Enovid was statistically sig- nificantly greater than that in their controls (x2=22.471; P<0.001) as was the incidence in the total treated females greater than that in the total controls (x2=22.471; P<0.001).
These chromophobe adenomas varied from small adenomas to large ones that were killing the animal. On the average they were more advanced in the treated females where many were causing stimula- tion of the mammary gland in these old females.
The other rather frequent neoplasms of this strain were those of the reticular tissue. The in- cidence of these neoplasms was lower in the groups treated with the two higher doses of Enovid than in their controls, but this was probably due to the more advanced hypophyseal tumors in these groups that reduced the average lifespan.
Other tumors in the C57BL females are listed, but there is no evidence that the Enovid treatment had any influence on their occurrence. Of par- ticular interest was the rare tumor of the pars intermedia of the hypophysis, such as the one found in a strain A female. There was a mammary tumor in one treated female, whereas these tumors are rare in this strain. It was a fibrous carcinoma of low grade malignancy.
DISCUSSION
The effect of a carcinogen must be expressed in probabilities rather than absolutes. No one will argue that, with adequate doses, certain hormones will increase the probability that mice of specific strains will develop certain tumors of the reproduc- tive and endocrine systems. The question here was whether in these selected sensitive mouse strains the antifertility drug, Enovid, fed in doses up to and including that dose which prevents reproduc- tion, would prove to be carcinogenic, i.e., would significantly increase the occurrence of any neo- plasm over that in the controls. The results were not very frightening.
It is significant that within this dose range, Enovid treatment did not increase the occurrence of ovarian tumors especially in C3HfB females that should be very sensitive to this test. The fact that we got no increase in this strain or in BALB/c, whereas Lipschutz et al. (8) reported induction of ovarian tumors in BALB/c with norethindrone and norethynodrel, may have been due in part to the fact that in their study the hormone was ad- ministered as a subcutaneous pellet and the ef- fective dose may have been higher. Also in their study, it was the norethindrone that appeared to be significantly effective. Furthermore, both were given without estrogens. However, in other studies estrogen has induced ovarian tumors.
It is also significant that within this dose range Enovid treatment did not increase the occurrence of mammary tumors in any strain, and in the highly susceptible strain C3H these tumors were actually reduced in the treated females. This ap- pears to be in contrast with the observations published by Kahn and Baker (5) of an increase in mammary tumors in A and C3H virgin females treated with norethynodrel. However, whereas we fed Enovid, they injected norethynodrel without the mestranol. Furthermore, some of their doses were higher, though it would be difficult actually to attempt to compare effective doses.
Also noteworthy was the reduction in occurrence of corticoadenomas of the adrenal gland in strain BALB/c and hepatomas in strain C3HfB with Enovid treatment. While the reduction in hepato- mas was undoubtedly related to the reduction in
body weight and possibly to the estrogen in the Enovid, the reduction in the corticoadenomas of the adrenal glands may also have been the result of the estrogen (20).
Our results regarding the epithelial lesions of the cervix and vagina in the BALB/c females seem to temper Dunn’s report (7) of induction of cancer of the uterine cervix in BALB/c mice fed Enovid. While we observed the lesions she de- scribed, we were unable to draw a clear line of distinction between them and those in the controls. While the highest dose of Enovid significantly increased the number of these lesions and possibly caused them to progress a little beyond those in the controls, never did they appear as grossly identi- fiable tumors, none had invaded beyond the vaginal wall, none had metastasized, and since they were observed only in the histologic section, none was transplanted.
We were able to confirm Poel’s earlier report (6) of an increase in tumors of the hypophysis with Enovid treatment. The increase was noted only in the susceptible strain C57BL and the tumors were in very old animals. Poel’s observation also had been limited to the related susceptible strain C57L. The question of malignancy of these chromophobe adenomas is sometimes raised. However, in this study the larger ones were active in stimulating the mammary glands in these old C57BL females and were the cause for the females to have to be necropsied.
Results from experimental animals should be applied to man with care, especially in this area where the use of antifertility drugs is so greatly needed in controlling probably the world’s greatest problem, overpopulation. Such results are of greatest value in directing our attention to certain areas for investigation in human beings.
While some concern over the possibility of inducing cancer of the cervix and vagina in women on Enovid should not be dismissed by investigators, in not one mouse in this study did the Enovid treatment induce a carcinoma of the cervix or vagina that killed the animal. However, these epithelial lesions of the cervix and vagina of BALB/c females are interesting, especially since they also occur spontaneously. These lesions are of
JOURNAL OF THE NATIONAL CANCER INSTITUTE
value in other studies of the effect of antifertility drugs in combination with other known carcinogens (21), and continuation of such studies should be encouraged, but they may have even greater value as models in other areas of investigation.
Results of this study cause one to appraise with caution the value of preliminary reports of mam- mary tumors in experimental animals treated with antifertility drugs or reports of mammary tumors in single or few animals so treated. In comparison with our results, Rudali et al. (22) observed no acceleration or increase in frequency of mammary carcinogenesis in C3H, RIII, or (C3H X RIII)F1 female mice with three contraceptives and their gestagens. Some evidence from their observations as well as those of the present study points more toward a protective effect.
Yet the present study and the earlier observations of Poel (6) do point to the possibility of induction of tumors of the hypophysis in older susceptible individuals treated with antifertility drugs and suggest the advisability of collecting hypophyses post mortem from women who have taken these drugs and sending them to a central pathology laboratory to see if there are changes that could be associated with the treatment.
In making any comparison with the effect of these drugs on women it should be emphasized that this study was not planned to give information on vascular accidents, the greatest hazard so far recognized in women, nor did it yield any informa- tion in this area. The study also would not yield information on possible other changes that estrogen might induce in species other than the mouse.
REFERENCES
(1) BARBER HR, GRABER EA, O’ROURKE JJ: Are the Pills Safe? Springfield, Ill., Charles C Thomas, 1969
(2) GARDNER WU, PFEIFFER CA, TRENTIN JJ: Hormonal factors in experimental carcinogenesis. In Physio- pathology of Cancer (Homburger F, ed.), 2d ed. New York, Hoeber-Harper, chapt. 6, 1959, pp 152-237
(3) WOOLLEY GW, FEKETE E, LITTLE CC: Gonadectomy and adrenal neoplasms. Science 97:291, 1943
(4) HERTZ R, BAILAR JC III: Estrogen-progestogen combinations for contraception. JAMA 198:1000- 1006, 1966
(5) KAHN RH, BAKER BL: Effect of long-term treatment with norethynodrel on A/J and C3H/HeJ mice. Endrocrinology 84:661-668, 1969
(6) POEL WE: Pituitary tumors in mice after prolonged feeding of synthetic progestins. Science 154:402- 403, 1966
(7) DUNN TB: Cancer of the uterine cervix in mice fed a liquid diet containing an antifertility drug. J Natl Cancer Inst 43:671-692, 1969
(8) LIPSCHUTZ A, IGLESIAS R, PANASEVICH VI, et al: Ovarian tumours and other ovarian changes in- duced in mice by two 19-nor-contraceptives. Br J Cancer 21:153-159, 1967
(9) HESTON WE, VLAHAKIS G, DERINGER MK: High incidence of spontaneous hepatomas and the in- crease of this incidence with urethan in C3H, C3Hf, and C3He male mice. J Natl Cancer Inst 24:425-435, 1960
(10) DERINGER MK: Occurrence of mammary tumors, reticular neoplasms, and pulmonary tumors in strain BALB/cAnDe breeding female mice. J Natl Cancer Inst 35:1047-1052, 1965
(11) HESTON WE, VLAHAKIS G: Mammary tumors, plaques, and hyperplastic alveolar nodules in various combinations of mouse inbred strains and the different lines of the mammary tumor virus. Int J Cancer 7:141-148, 1971
(12) GARDNER WU, ALLEN E: Malignant and non- malignant uterine and vaginal lesions in mice receiving estrogens and estrogens and androgens simultaneously. Yale J Biol Med 12:213-234, 1939
(13) GARDNER WU: Carcinoma of the uterine cervix and upper vagina: Induction under experimental con- ditions in mice. Ann NY Acad Sci 75:543-564, 1959
(14) SUNTZEFF V, BURNS EL, MOSKOP M, et al: On the proliferative changes taking place in the epithelium of vagina and cervix of mice with advancing age and under the influence of experimentally ad- ministered estrogenic hormones. Am J Cancer 32:256-289, 1938
(15) HERBST AL, ULFELDER H, POSKANZER DC: Adeno- carcinoma of the vagina. Association of maternal stilbestrol therapy with tumor appearance in young women. N Engl J Med 284:878-881, 1971
(16) HESTON WE, VLAHAKIS G: Influence of the Av gene on mammary-gland tumors, hepatomas, and normal growth in mice. J Natl Cancer Inst 26:969-983, 1961
(17) HESTON WE: Mammary tumors in agent-free mice. Ann NY Acad Sci 71:931-942, 1958
(18) HESTON WE, DERINGER MK: Occurrence of tumors in agent-free strain C3Hf male mice implanted with estrogen-cholesterol pellets. Proc Soc Exp Biol Med 82:731-734, 1953
(19) HESTON WE, VLAHAKIS G: Factors in the causation of spontaneous hepatomas in mice. J Natl Cancer Inst 37:839-843, 1966
(20) WOOLLEY GW, LITTLE CC: Prevention of adrenal cortical carcinoma by diethylstilbestrol. Proc Natl Acad Sci USA 32:239-240, 1946
(21) MYHRE E, BJÖRO K: Hormones and Cervical Cancer. Oslo, Universitetsforlaget, 1971
(22) RUDALI G, COEZY E, CHEMAMA R: Mammary car- cinogenesis in female and male mice receiving contraceptives or gestagens. J Natl Cancer Inst 49:813-819, 1972
1
5
O
2
2
3
4