[CANCER RESEARCH 41, 721-734, February 1981] 0008-5472/81 /0041-OOOOS02.00 Neonatal Estrogen Treatment and Epithelial Abnormalities in the Cervicovaginal Epithelium of Adult Mice1
John-Gunnar Forsberg2 and Terje Kalland
Department of Anatomy, University of Lund, Biskopsgatan 7, S-223 62 Lund, Sweden [J-G. F.], and Institute of Anatomy, University of Bergen, Bergen, Norway IT. K.]
ABSTRACT occasionally in the uppermost part of the vagina, there were regions with HCE3 which formed glandular-like downgrowths Female NMRI mice were given injections of different doses into the stroma, resulting in adenosis (13, 19). of 17/8-estradiol, 17a-estradiol, diethylstilbestrol (DES), dienes- Later, it was demonstrated that a synthetic estrogen such as trol, frans-stilbene, progesterone, testosterone, 5a-dihydrotes- DES, injected according to the same schedule as that for tosterone, or olive oil for the first 5 days after birth. When the estradiol, resulted in the same type of changes (15); moreover, females were killed at 8 weeks after birth, all the estrogens, in old females, changes highly suggestive of malignancy de effective at different dose levels (1CT2 to 5 fig/day), had veloped within the regions of adenosis (17, 20). resulted in the display by several of the cervicovaginal prepa While HCE and adenosis were originally described in NMRI rations studied of a heterotopic columnar epithelium (HCE) in mice, the results have later been verified in other strains (51) regions where females given injections of olive oil, testoster regarding localization in the vaginal fornix and the uppermost one, 5a-dihydrotestosterone, progesterone, or frans-stilbene part of the vagina. For the uterine cervix, there is still a had only the normal squamous epithelium. The further fate of discrepancy in results from different strains. the HCE was followed at two later age stages, 36 to 52 weeks After the observation of an association between DES inges and 14 to 17 months. The HCE developed into glandular-like tion during human pregnancy and development of clear cell structures penetrating into the stroma and justifying the des adenocarcinoma in the uterine cervix and vagina of female ignation of adenosis. DES resulted in a more pronounced offspring at puberty (28, 30), as well as the occurrence of a adenosis than did 17/?-estradiol; in both cases, metaplasia was high incidence of adenosis in the vagina after such treatment, a striking component of the adenosis regions. Development of the findings from NMRI mice became attractive as a model for adenosis from HCE was dependent upon presence of the development of cancer in the human situation. The similarities ovaries. Some preparations from 44-week-old females given and dissimilarities between DES-induced changes in mice and DES injections showed signs of a beginning malignant trans humans have been discussed, and the mouse model has been formation in the adenosis regions, more evident in 17-month- pointed out to be of importance for the human situation (2, 15, old females. Among the 23 preparations in the latter group, 8 18, 41, 51 ). The cervicovaginal epithelium of a newborn mouse had changes morphologically indicating malignancy with ex is of a morphologically undifferentiated type. The develop amples of adenocarcinoma, mixed carcinoma, and squamous mental changes taking place during the first week after birth carcinoma. Because of the seemingly low aggressive nature of this malignancy, the term "pseudocarcinoma" is discussed. correspond to morphogenetic steps in the human vaginal epi thelium before the 18th to 20th week of pregnancy. In the Ten- to 12-week-old BALB/c and C57BL/6 females given DES neonatal mouse model, the mother organism and the placenta injections neonatally had HCE in the uterine cervix and vaginal do not interfere with the estrogen treatment. However, DES fornices after neonatal DES injections. Differences in extension passes the placental barrier (59), and female offspring from of HCE were observed after DES injections for three different DES-treated pregnant hamsters or mice develop epithelial 5-day periods in the neonatal and immature stages of NMRI changes in the genital tract similar to those seen after neonatal females. An interaction between different DES-sensitive param treatment (45, 57). eters to result in the pseudocarcinomas is discussed. In this study, we have analyzed the specificity of different estrogens, natural and synthetic, and other sex steroids as INTRODUCTION regards their potential for inducing epithelial aberrations of the HCE type and adenosis. For those substances active in induc The normal cervicovaginal epithelial lining in adult mice is a ing HCE, dose-response experiments were done. Strain differ stratified squamous epithelium which undergoes cyclic ences and critical periods for HCE induction and the ultimate changes related to the estrous cycle. These epithelial condi fate of HCE and adenosis in old females (oldest stage studied, tions are changed in adult females given estrogen injections 17 months) were investigated. neonatally. Thus, in 1962, it was demonstrated that such a treatment may result in persistent vaginal cornification (61, 62, MATERIALS AND METHODS 64). Forsberg (13) in 1969 described another type of perma nent epithelial aberration induced by estradiol neonatally (NMRI The majority of the mice used belong to a closed, outbred mice); mainly localized to the uterine cervix and vaginal fornix, stock of NMRI mice. In some experiments, inbred females of the BALB/c or C57BL/6 strains were used. The animals were ' This investigation was supported by grants from the Norwegian Cancer fed a standard pellet diet and given water ad libitum. Apparently Society (Landsforeningen mot Kreft) and the Norwegian Research Council for Science and Humanities. 2 To whom requests for reprints should be addressed. 3 The abbreviations used are: HCE, heterotopic columnar epithelium; DES, Received July 3, 1980; accepted October 31, 1980. diethylstilbestrol; DHT, 5a-dihydrotestosterone; CCC, common cervical canal.
FEBRUARY 1981 721
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. J-G. Forsberg and T. Kailand
pregnant mothers from breeding cages were put into separate columnar type (Fig. 1D). Interspersed in this region were small cages to give birth to their litters. Within 24 hr after birth, areas where hyperplastic nodules of cuboidal cells were seen newborn females were separated from newborn males. Fe under the columnar epithelium (Fig. 1£).Caudally in the vagina, the basal zone was hyperplastic. In the group receiving 10~2 males from several litters were pooled, and every mother animal was allowed to suckle 7 to 9 females. fig DES, both normal and affected preparations were repre In most experiments, the neonatal females were given s.c. sented. injections for 5 days, starting within 24 hr after birth, of varying The results of the mitotic rate studies are demonstrated in doses of the test substances in 0.025 ml olive oil. Pressure Fig. 1, C and D, and Chart 2. There was a progressive decrease was kept on the injection site for about 15 to 20 sec after in mitotic rate activity related to the daily DES dose. A dose as small as 10~5 fig reduced the proliferative activity. The curve withdrawal of the needle to avoid leakage. In some experi had a tendency to flatten out at 10~3 and 10~2 ¡ig,only later to ments, the injections started later after birth (Day 4 or 6) but still covered a 5-day period. Finally, in one experiment, females turn down steeply to almost zero level at 10"' and 1 fig DES. were given injections of 5 jug DES for 5 days neonatally, Histology of Cervicovaginal Epithelium from 8-Week-old ovariectomized at 8 weeks, and killed at 51 weeks. Controls NMRI Females. Epithelial abnormalities were recorded from 3 were given injections of the vehicle (olive oil) only. different regions: uppermost part of the vagina; vaginal fornix; The test substances used were DES, dienestrol, frans-stil- and CCC. "Abnormal" epithelium is defined as HCE, in direct bene, 17/î-estradiol, 17a-estradiol, progesterone, testoster contact with the basal membrane, and interspersed in areas one, and DHT. The purity of the 17a-estradiol batch used was normally lined with squamous epithelium (Fig. 2). checked in a gas chromatography system." Contamination with In Chart 3 is plotted the extension of columnar epithelium in 17/?-estradiol was found to be less than 0.5%. Stock solutions of the different substances (obtained from Sigma Chemical uterine horn Co., St. Louis, Mo.) were prepared by dissolving an appropriate amount of the test substance in a minimal amount of ethanol and adding olive oil. -JtiuL-/:./.—cervical part The animals were killed by cervical dislocation at different sq.-c.j. ages from 8 weeks to 17 months. Most test groups comprise common preparations from at least 3 different litters, no litter being cervical represented with less than 2 preparations. The whole uterine canal cervix and upper part of the vagina were dissected out in foto [cccj (Chart 1). The preparations were fixed in Bouin's solution, dehydrated in alcohol, and embedded in paraffin. The prepa rations were serially sectioned in 7-/tm-thick sections. All sec tions were stained in hematoxylin and eosin. Neonatal NMRI females were given different daily injections of DES for a 5-day period; on Day 6 they were given a s.c. injection of 1 fig colchicine in Tyrode's solution and killed 4 hr later (24 hr after the last DES injection). The test groups at this age were pooled from only 2 litters. Histological sections (7 Chart 1. A schematic illustration of the anatomy of the region studied. Only the upper part of the vagina was included in the preparations. The region between firn) were prepared from the upper part of the vagina and the the 2 dashed lines represents the normal variation for the level of the squamo- uterine cervix. The mitotic rate was calculated as the number columnar junction (sq.-c.y.). of metaphases in 1000 counted cells. Care was taken to make 6-DAY-OLD FEMALES the calculations at the same level of the cervical canal and only in its right half (as apparent in the section).
RESULTS
Histology and Mitotic Rate Studies in Cervicovaginal Prep arations from 6-Day-old NMRI Females. In preparations from females given olive oil injections or 10~5 /ig DES, or 10~3 fig DES for the first 5 days after birth, the transformation of the fetal and neonatal pseudostratified columnar miillerian epithe lium had progressed into the vaginal fornices and into the caudal part of the CGC. As a result of the transformation process, the columnar epithelium is changed into an epithelium consisting of a superficial zone of high columnar cells and a basal zone of pale cuboidal cells (Fig. 16). After daily injections of 10"1 or 1 /¿gDESfor the same time, the conditions were different. In the upper part of the vagina O 10'5 IO"3 10"2 IO'1 and in the CCC, the epithelium was of an untransformed 1 Daily dose of DES(ug) 4 This analysis was done by Q. Plym Forshell. AB LEO Research Laboratories, Chart 2. The effect of different doses of DES (daily injections for the first 5 Helsingborg, Sweden, for which the authors are very grateful. days after birth) on the mitotic rate in the CCC of 6-day-old NMRI females.
722 CANCER RESEARCH VOL. 41
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. Neonatal Estrogen Treatment and Cervicovaginal Abnormalities
JÄVip •*M•s•••'//
20
c 40 o E E o o 60
80
_ I .- . r- alt-ald^ î*' ' Caudal tÃ
DES DIENESTROL 17J3-E-, 17oc-E2 i -Small, icattered local regie Chart 3. The extension of HCE in théCGC after various types of neonatal treatment. All females (NMRI) killed at the age of 8 weeks. Each point represents one studied preparation. The 10~3- and 5-fig dienestrol groups contain 9 studied preparations; all other groups contain 10 preparations. Te, testosterone; transstilb, frans-stilbene; 17ß-E2,17/8-estradiol; Ì7a-E2,17a-estradiol; Pr, progesterone.
the CCC in percentage of the total length of the canal (calcu Table 1 lated as number of sections of 7 fim) after various types of Hormonal specificity in induction of HCE in the CCC of NMRI females A review of the number of 8-week-old females with HCE in the 3 different neonatal treatment. In the present study, females having a regions studied after different types of treatment for the first 5 days after birth. columnar epithelium below the lowest level found in olive oil- The daily dose is indicated. injected females, i.e., the upper 35% of the length of the CCC, No. of females with HCE/no. have been classified to have HCE. This percentage is, of treated course, only an estimation of the true level of the squamocol-
umnar junction. It may be influenced by, e.g., shrinkage arti TreatmentOlive fornix0/100/100/100/100/107/108/105/100/100/108/105/106/93/101/101/100/98/100/101/108/100/100/10Uppervagina0/100/100/100/100/101/100/100/100/100/100/101/101/90/100/100/100/90/100/100/100/100/100/10 facts during tissue preparation, section artifacts, etc. However, the "percentage level" serves the practical purpose of deline oilTestosterone, figDHT, 10 ating the normal range of distribution of columnar epithelium in 5ngProgesterone, ngfrans-Stilbene.50 the CCC from regions with definite HCE in experimental fe ngDES, 10 males. The cervical HCE is characterized by a wavy appear 5(igDES, 1figDES, ance and cells with a pale-staining cytoplasm (Fig. 2). 10">gDES, Chart 3 does not give any idea of how large a part of the MistigDES, cross-sectional circumference of the cervical epithelium is of a 10"3/igDES, ngDES,1 ;ig, and progesterone, 50 columnar type. Generally speaking, as the HCE descends into ^gDienestrol,10"' (ig. and progesterone, 50 the CCC, its circumferential part of the epithelial lining is figDienestrol,5 figDienestrol,1 gradually reduced. Exceptionally, instead of being continuous, /igDienestrol,10~' HCE may occur as islands completely surrounded by squa- jigDienestrol.10~2 mous epithelium. 10~3fig1 fig17/S-Estradiol.7/8-Estradiol, 5 Columnar epithelium does not occur in the vagina and the fig1 1 vaginal fornices of normal female mice. In the vaginal fornices fig17/8-Estradiol, 10"' of experimental females, the HCE has a pronounced tendency fig17a-Estradiol, 5 fig17a-Estradiol. 1 to form downgrowths into the stroma, resulting in sac-like or 7o-Estradiol, 10~' /igCCC0/100/100/100/100/109/107/106/102/100/1010/107/104/95/108/103/100/96/103/101/108/100/100/10Vaginal cyst-like formations. The cyst-like formations sometimes gave an impression of distension because of a distended epithelial wall and secretory material in the lumen. In the few cases of different treatment groups and regions studied (for the CCC, HCE in the uppermost part of the vagina, the common finding below the upper 35% region). In all females given injections of was that of small glandular-like formations, lined by columnar olive oil, testosterone, DHT, and frans-stilbene, normal epithe epithelium and penetrating into the stroma. In the CCC, the lial conditions were found in the vagina, vaginal fornix, and the HCE forms part of the epithelial wall without any cysts or CCC. For progesterone, a daily dose of 50 fig was the highest downgrowths (Fig. 2). tolerable dose. One hundred fig progesterone daily for 5 days Table 1 summarizes results on the occurrence of HCE in resulted in a high proportion of deaths among the neonatal
FEBRUARY 1981 723
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. J-G. Forsberg and T. Kalland females; this dose was discontinued after some initial pilot on the day of birth, on Day 4, and on Day 6. The females given experiments. The 50-fig progesterone dose does not seem to the first injection on the day of birth are the same as those be completely ineffective as the spreading of the level for the described above (Chart 3; Table 1). When treatment started on squamocolumnar junction within the upper 35% region of the Day 4, 9 of 10 females had HCE in the CCC (the same incidence CCC was more pronounced than in controls (Chart 3), but in as for females treated on Days 1 to 5). However, no epithelial no case did HCE occur. Females given injections of trans- aberrations occurred in the vaginal fornices or in the vagina. A stilbene, progesterone, and DHT had a vaginal epithelium in further 2 days of delay in the start of injection (Day 6) resulted different phases of the estrous cycle while those given testos in 6 of 10 females with HCE in the CCC but not in the vaginal terone all had a cornified_cervicpvaainal epithelium. fornix or vagina proper. A postponement at the treatment Both the natural and synthetic estrogens were effective in period beyond Days 1 to 5 resulted in a slight but nonsignificant inducing HCE. The lowest effective daily dose for DES and reduction in the length of the CCC showing HCE. dienestrol was 10~2 fig; 10~1 jug 17/S-estradiol resulted in one Histology of Cervicovaginal Epithelium from C57BL/6 (10- female with HCE while only the 5-fig dose of 17a-estradiol was Week-old) and BALB/c (12-Week-old) Females Given Injec effective when HCE was defined as described above. Com tions of 5 fig DES on Days 1 to 5. Control females from both pared with olive oil-injected females, 50 jug progesterone and strains had the lowest point of the squamocolumnar junction 1 /ig 17a-estradiol resulted in a tendency to downward move within the same cranial 35% area of the CCC as did the NMRI ment of the level of the squamocolumnar junction within the females (Chart 4). In the C57BL/6 strain, neonatal DES treat upper 35% of the CCC (Chart 3). When 50 ¡igprogesterone ment resulted in induction of areas with HCE in the CCC in 7 of were combined with 10~' or 1 fig DES, the effect of DES was 10 preparations; in 6 cases, HCE and cysts were in the vaginal potentiated; both the number of females with HCE and the fornices; and, finally, in 2 cases, small sacs with columnar extension of HCE increased compared with the situation in epithelium were seen in the uppermost part of the vagina. To females given injections of the same doses of DES alone. judge from the sample of 9 preparations, BALB/c females Only the highest dose (5 ¡ig)of 1/«-estradici and 17/J- seem to be more resistant to induction of HCE. Only one female estradiol caused HCE in females throughout the CCC (Chart demonstrated HCE in the CCC, 4 had cysts in the vaginal 3). A single example of the same effect was seen in each of the fornices, and no aberrations were seen in the vagina. groups given injections of 10"' fig dienestrol or DES (for DES, Level of the Squamocolumnar Function at Different Times 97% of the CCC). For higher doses of DES, there was a in Life. The lowest level of the squamocolumnar junction in the tendency to an increased number of females with HCE through CCC of control females (given injections of olive oil neonatally) out the CCC (Chart 3). For both the estradiol epimers and DES, was studied at 3 different ages: 8 weeks, 40 weeks, and 17 a dose-response effect was evident, as regards both the num months after birth. A comparison of the conditions at the 3 ber of females with HCE and the extension of HCE in the CCC. different ages does not indicate any age-related change; at all For dienestrol, the dose-response relationship was more diffi 3 ages, the lowest level of columnar epithelium occurred in the cult to evaluate. The same estrogens that induced HCE in the CCC also Cranial induced HCE in the vaginal fornix, while no HCE occurred in the upper part of the vagina after injections with 17a- or 17ß- 0| estradiol. In the latter region, foci of HCE were rarely seen after injections of 5 fig DES or dienestrol as well as after the combined treatment with progesterone and DES. Because of technical problems, no attempt was made to quantify the ex 20 tension of HCE in the vaginal fornix in a way similar to that o"5ue done for the CCC. From Chart 3 and Table 1, it is evident that among the substances tested only estrogens resulted in HCE. For all the 40Ô-Ct1 estrogens used except dienestrol, there is an approximate resemblance in incidence of HCE in the CCC and vaginal fornix at all the dose levels studied. For dienestrol, the concordance is more questionable; in the vaginal fornix, a relationship in the 60 dose-response incidence figures seems to exist while this is i01II not the case for the CCC. The exceptional findings of HCE in the uppermost part of the vagina are evident. Females given injections of less than 1 /ug 17/8-estradiol or less than 5 jug 17a- 60100Cauda estradiol all had a vaginal epithelium in different phases of the estrous cycle. Besides the HCE aberration described above, females given neonatal injections of the highest estrogen dose (5 jug) had a deviating cervical structure. The gross structure of the epithelial Controls 5 pg Controls 5 pg DESC DES lining was simpler than in controls, and the lumen had much end-----»•tT•n«10...:.•n.e9t*n.9»:•1 57BI/6BALB/cIO reduced folding. weeks 12 weeks The effect of injections of 5 fig DES during a 5-day period Chart 4. Extension of HCE in the CCC of C57BL/6 and BALB/c mice given starting at 3 different times was compared. Treatment started injections of 5 ¿igDESfor the first 5 days of life. Ten- to 12-week-old females.
724 CANCER RESEARCH VOL. 41
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. Neonatal Estrogen Treatment and Cervicovaginal Abnormalities
upper 35% of the CGC. Depending on the stage of the estrous such a grouping involves subjective factors. Its validity for the cycle, the border was more or less distinct. In the diestrus 3 groups, HCE, local adenosis, and general adenosis, was stage, the squamous epithelial type was represented by a basal checked in a blind test. Because of the pronounced differ layer of cuboidal cells and a superficial layer of columnar cells. ences, it was no problem to obtain the same result from the Through the proestrus and estrus stages, this epithelium was blind test as from the preliminary grouping. dramatically changed into a highly typical squamous epithelium In the groups of females given injections of low daily DES doses [10~3 and 10~2 jug (Chart 5, B and Q] only one case with an abrupt border to the columnar epithelium. Histology of Cervicovaginal Preparations from Neonatally each of HCE and local cervical adenosis occurred. No epithelial Estrogen-injected Females Killed at 36 to 52 Weeks after lesions were seen in the upper part of the vagina or in the vaginal fornix. Increasing the dose to 10~1 and 1 ¡IQDES(Chart Birth. Control preparations from 26 females all had a normal Cervicovaginal squamous epithelium in different phases of the 5, D and E) resulted in examples of both local and general estrous cycle. All epithelial lesions to be described were lo adenosis, one case in each treatment group being of the latter cated below the upper 35% of the CGC. type. Irrespective of the presence of local or general adenosis, The lesions were grouped under the 4 headings used in many of the glandular-like ducts had signs of ongoing metapla Chart 5. Adenosis was defined as glandular-like downgrowths sia; i.e., the squamous epithelium seemed to undermine the into the stroma, lined with a regular columnar epithelium. In the columnar epithelium which was shed. Sometimes the columnar "local adenosis" group were included cases with more or less epithelium was completely replaced by a squamous epithelium. circumscript regions of adenosis penetrating only a short dis From Chart 5, where the number of preparations with meta- tance into the stroma (Fig. 3C). In the more advanced "general plastic regions is indicated with an asterisk, it is evident that adenosis" cases, the downgrowths formed a prominent tissue metaplasia was a common phenomenon in combination with contribution along the CCC, penetrating deeply into the stroma adenosis. Only one preparation with local adenosis from a (Fig. 3D). The high correlation between vertical and horizontal female given an injection of 1 fig DES had no sign of metaplasia. extension of adenosis simplified the classification. However, Metaplasia may be a process reducing the extension of HCE in the epithelial wall of the CCC. Interspersed in the morphologi B D cally normal squamous epithelium were regions, often in direct continuity with HCE, with an abrupt transition into a low squa MalignancyGeneral 0 0T2*2201*K*2k mous epithelium with less maturation than the normal-looking one (Fig. 3F). This low squamous epithelium was interpreted adenosisLocal 0 as a metaplastic epithelium in a region earlier occupied by HCE.
adenosisHŒ In all the preparations included in the groups described thus J_"0" J far, the glandular-like downgrowths in the regions of adenosis had an irregular and tortuous course, but the epithelial cells had a quite normal appearance. Scattered mitoses were seen. Normal000026 Infiltration with round cells was often striking. Neonatal treatment ¡Mg/day.treatment Controls lO~3DES-5 lO'2DES-5 Iff'DES-5 l DES •5 5DES-S Females given 5 /ig DES for the first 5 days of life (Chart 5F) period m days) had dramatic epithelial abnormalities in the regions studied. Age in weeks or 39-40\ 36-52 \ 36 w The lesions were most prominent in the lower part of the CCC months and in the vaginal fornices but also engaged the upper part of
weeksr0r2+C27517 months the vagina. Four of the 9 preparations studied presented a »J10522K000010r1 picture compatible with severe general adenosis in the CCC H0rri125 and vaginal fornices. The stroma was massive penetrated by 3m!l irregular glandular-like ducts, in some cases almost reaching Malignancy 4a¡: O Is21 the mesothelial surface of the cervix. The ducts were lined by General a regular high columnar epithelium. Peripherally, the latter adenosis o often changed into a low cuboidal epithelium. Again, epithelial Local cords and tubules containing squamous epithelium formed a adenosis prominent part of the picture. The normal gross morphology of O the epithelial CGC was lost in these regions with massive epithelial penetration of the stroma. In contrast to the conditions in the CCC, the uterine epithelium did not demonstrate any
Neonatal treatment 5 Q£S.5 170- hyperplastic changes; only some dilation of uterine glands was (pg/day -treatment ox"8weeks 5DES-IDES- E2-536 DES-5 Controls period m days! seen. Age in In a further 2 preparations with the same massive general weeks or Sii ¿0-52w3652 -52 wM,X 1-. months 17months 17months months adenosis, regions occurred in which the columnar cells of the Chart 5. Late epithelial effects in the CCC after various neonatal treatments. irregular and highly invasive glandular-like structures had an The number of preparations with different types of lesions as well as the age of abnormal appearance. The nuclei were hyperchromatic and the females (NMRI) at the time of killing is indicated. For explanations of the severity grades "local adenosis," "general adenosis," and "malignancy," see polymorphic, and the cells were densely packed. The number text. Capita! letters heading the columns, different treatment groups (see text). of mitotic figures indicated a high proliferative activity. Because Asterisk, metaplasia in glandular ducts; w, weeks; m, mixed carcinomas; a, the general impression was that of a transformation into ade- adenocarcinomas; s, squamous carcinomas; ox, ovariectomy; 17ß-Ez,17/i-es- nocarcinoma, the 2 preparations were included in the "malig- tradiol.
FEBRUARY 1981 725
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. J-G. Forsberg and T. Kalland
nancy" group (Chart 5F; Fig. 48 ). One of the preparations had injected, may be of the ovary-dependent or ovary-independent pronounced changes in the upper part of the vagina, the whole type (61 ). The latter is a result of a direct estrogen effect on vaginal wall being penetrated by adenosis and metaplastic the neonatal vaginal cells (64). The ovary-dependent persistent epithelium of a benign type. There was a striking difference vaginal cornification results from lower estrogen doses; in this between the histology of these preparations and that from case, the primary effect of estrogen is in the hypothalamic- females given an identical neonatal treatment but ovariecto- pituitary control system, resulting in no luteinizing hormone mized at the age of 8 weeks. None of the preparations from surge (24). Testosterone injected neonatally has the same the latter females had changes beyond local adenosis (Chart effect as do estrogens, which has been ascribed to an aroma- 5G). tization in the central nervous system (54, 61). The results in A single injection of 5 /ig DES on the day of birth resulted in this study are in line with these well-known facts. DES and epithelial abnormalities in the uterine cervix, but the incidence dienestrol are more potent in the induction of persistent corni of these lesions was lower than in females given the same dose fication compared with estradici (41), the 17/?-epimer being for 5 days. Two preparations from females given the single more potent than the 17a-epimer. Because daily vaginal dose (Chart 5H) were classified as general adenosis, but this smears were not included as part of this study, conclusions on was not as impressive by far as in females treated for 5 days. persistent vaginal cornification are based only on findings of all Among 16 females given injections of 5 /¿g17/S-estradiol (Chart females in a treatment group with a cornified vaginal epithelium. 51), 6 had local adenosis and one had general adenosis, the In the present study, no attempt was made to separate ovary- latter being of a more restricted type than in the females dependent cornification from the ovary-independent type. included in Chart 5F. In addition to the permanently changed character of the Common for all the cervices exposed to estrogen was the squamous vaginal epithelium, neonatal estrogen treatment of reduced size and simple morphology of the epithelial part of NMRI mice results in the occurrence of HCE in the uppermost the organ compared with controls. part of the vagina, in the vaginal fornix, and in the CCC (13, Histology of Preparations from 14- to 17-Month-old Fe 19). The HCE has been interpreted as a remaining columnar males Given Injections of Olive Oil or DES Neonatally. All the miillerian epithelium, which has been blocked in its normal olive oil-injected control preparations had a normal cervicovag- transformation into a squamous epithelium, a process which inal epithelium in different phases of the estrous cycle. takes place in the neonatal period of mice (19). Among the 23 preparations from females given injections of To speak about HCE in the uterine cervix, it is essential first 5 /ig DES for 5 days after birth, 6 exhibited completely normal to know the normal distribution of columnar epithelium in this epithelial conditions, 7 preparations were classified as having region. In studies on the squamocolumnar junction and its local adenosis, and 2 were classified as having general aden cyclic variations, the junction has been localized at or just osis. Again, metaplasia occurred within the adenosis regions. above the bifurcation of the CCC (13, 19, 25, 43). However, Eight preparations were included in the "malignancy" group columnar epithelium may extend downwards for a short dis (Chart 5L). Among these, 4 contained regions compatible with tance into the lateral walls of the CCC, and the squamous adenocarcinoma, one was considered to represent a squamous epithelium may extend somewhat cranially into part of the carcinoma, and, in 3 cases, regions with a morphology repre epithelial lining of the bifurcated cervical canals. In the 8-week- sentative of adenocarcinoma and squamous carcinoma oc old olive oil-injected females studied in this investigation, 2 curred in one and the same preparation, which is why the females were exceptions from earlier observations (13) be designation mixed carcinoma is used (Fig. 4). The term "car cause columnar epithelium formed part of the epithelial lining cinoma" is here adopted in a purely descriptive sense to in the upper 35% region of the length of the CCC. Based on a indicate a histological picture characterized by nuclear pleo- study of 40-week-old and 17-month-old controls, no evidence morphism, irregular and massively invading epithelial struc was found for an age-related change in distribution of the tures, and loss of maturation. A pronounced inflammatory cervical columnar epithelium. reaction was seen in the stroma with leukocytes invading the Both the synthetic estrogens used and the 2 epimers of epithelium. All the carcinomas were of a well-differentiated estradiol induced HCE. The incidence of females with HCE as type, and they were all localized to the cervical region with no well as the extension of the latter increased with increasing ingrowth into adjacent organs; no métastaseswere seen. doses. A possible effect of progesterone cannot be excluded. The reduction in size of the CGC was in some preparations The conclusion is that estrogen injected into neonatal NMRI so pronounced that the lumen was almost obliterated. The females induced HCE both in the CCC and in the vaginal fornix, same preparations had no signs of adenosis (Fig. 36). as well as occasionally in the uppermost part of the vagina. Among the 10 preparations from 14-month-old females given By starting the 5-day period of DES treatment on different injections of 10"' fig DES for 5 days neonatally (Chart 5J), one days after birth (Days 1, 4, or 6), it was possible to influence had a well-differentiated squamous carcinoma localized to the the development of HCE in a time-dependent way. When treat vaginal fornix and lower part of the CCC, and infiltrating to the ment started on Day 4 or 6, no HCE occurred in the upper mesothelial surface (Fig. 4/4). Local adenosis was seen in 5 vagina or vaginal fornix, and the incidence of HCE in the CCC preparations, and HCE was found only in 2 preparations. was reduced. These results are in line with the cranially pro gressing differentiation of the cervicovaginal epithelium (16). The progressive development of HCE to glandular-like down- DISCUSSION growths into the stroma, resulting in adenosis, is a postpubertal Estrogens injected into neonatal mice affect the genital epi event, controlled by ovarian factors (13). This was supported thelium in different ways. The vaginal epithelium shows a by the results from the present study. After ovariectomy at 8 persistent cornification which, depending on dose of estrogen weeks, no general adenosis was present. In 3 preparations
726 CANCER RESEARCH VOL. 41
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. Neonatal Estrogen Treatment and Cervicovaginal Abnormalities 10~1 or 1 fig DES for the first 5 days after birth, there were with local adenosis, only a few scattered glandular elements occurred. The difference between these females and identically substantial areas in the upper part of the vagina with an treated but nonovariectomized females is striking. There are untransformed epithelium, in contrast to the controls. Only autoradiographic evidences for the occurrence of specific es rarely was HCE found in the corresponding region of 8-week- trogen receptors in the adenosis epithelium (37). old females. The scattered, hyperplastic nodules of cuboidal Recently, development of HCE and adenosis-like lesions was cells under the columnar cells are interesting in this connection. demonstrated in the vaginal fornix after neonatal DES or 17/8- Similar nodules were never seen in the HCE areas of 8-week- estradiol treatment of females from 3 inbred strains and one old females. The nodules may be thought of to expand under outbred stock of mice (51 ). Thus, as regards localization to the the surrounding columnar epithelium, finally growing together vaginal fornix, these estrogen-induced lesions are not strain and forming a complete basal zone under the nontransformed specific. In the inbred BALB/cCrgl, C3H/Crgl, and C57BL/ epithelium, the latter finally being sloughed (60, 63). This is a Crgl strains and the noninbred NCS stock, no signs of HCE or possible explanation for the more restricted untransformed adenosis-like lesions could be demonstrated in the CCC of epithelial area in 8-week-old females compared with 6-day-old females given injections of DES or 17/8-estradiol 3-benzoate females. (51). A detailed analysis of the epithelial conditions in the CCC In the mouse, metaplasia was a prominent feature in the of NCS and BALB/cCrgl mice did not reveal any difference in glandular-like downgrowths, both in cases of local and general distribution of columnar and squamous epithelium between adenosis. In human females with vaginal adenosis, after fetal control females and those given injections of estrogen neo- exposure to DES, metaplasia is commonly encountered and natally. Instead, even in controls, a considerable proportion of has been referred to as a reparative phenomenon (3, 29, 50, the CCC contained columnar epithelium. In the BALB/cCrgl 55, 56). The histological morphology sometimes pointed to a strain, there was a tendency for cervical glands to occur in complete replacement of HCE and early adenosis in the vaginal more caudal parts of the CCC after estrogen treatment com fornix and CCC with a metaplastic squamous epithelium, even pared with controls. Strain and hybrid differences in sensitivity of a dysplastic type. Until the ultimate fate of the metaplastic squamous epithelium is known, the work "reparative" should to DES have been reported (26, 27). In our hands, C57BL/6 mice developed HCE after neonatal DES treatment, while only be avoided and used only in the sense of describing a replace 1 of 9 comparably treated BALB/c mice showed HCE. ment of adenosis columnar epithelium with squamous epithe lium. The risk of a DES-exposed woman's developing squa No cases of vaginal or cervical adenosis were seen in ICR/ Jcl mice given a single injection of 50 fig DES per g body mous neoplasia in vagina and cervix is still controversial (23, weight within 12 hr after birth (48), nor did prenatal DES 55). treatment result in adenosis in the female offspring, in spite of Besides the general resemblance in adenosis between mice DES being recovered from the fetal compartment (46, 59). The and humans, there are also more detailed similarities. The negative result with the high neonatal dose was explained by process of metaplasia has already been discussed. In both the authors to be due to the very high dose used, which killed humans and mice, an inflammatory reaction usually accompa affected newborns. By injecting 10 or 100 fig DES per kg body nies adenosis (29, 56). weight into pregnant CD-1 mice (Days 9 through 16 of gesta The difference in doses of synthetic estrogens and estradiol tion), an exceptional case of glandular features resembling that would induce HCE may be explained by differences in adenosis was found in the vaginal fornix of an 11-month-old bindings to a-fetoprotein, "an estrogen-buffering mechanism" female offspring5. Additional glandular elements were found in fetal and neonatal rodent tissues, controlling the availability with higher incidence in the vaginal walls, but these glands of active estradiol (9, 51, 65). DES was about 100-fold as originated from the urethra. In the uterine cervix, epithelial potent as was 17/î-estradioland about 500-fold as potent as abnormalities ranged from mild cellular atypia to an epidermoid was 17a-estradiol. tumor, but no adenosis. The gross enlargement of the cervix The similarity in HCE induction by the 2 epimers of estradiol was evident. In the present material, the enlargement is as is interesting. The small contamination with 17/}-estradiol (less cribed to stromal changes (increase in number of cells, edema, than 0.5%) in the 17a-estradiol batch used cannot explain the sometimes hyalinization), while the cervical lumen and its epi striking effect from experiments using 5 fig 17a-estradiol. How thelial lining are often reduced. ever, a local metabolic conversion of the 17a epimer to 17ß- The 2 studies referred to above thus argue that prenatal DES estradiol should not be excluded. 17a-Estradiol has a low treatment does not induce cervicovaginal adenosis, while the affinity to the estrogen receptor, a low biological activity, and present study and that by Plapinger and Bern (51 ) demonstrate a very slight effect on the induction of induced protein (1, 5, its occurrence after neonatal treatment (localization in vagina 39, 40, 52). Roughly speaking, to induce HCE, the 17a-epimer and vaginal fornix only). The failure to induce a high incidence was about one-fifth as active as was 17/8-estradiol. This small of adenosis using prenatal estrogen treatment points to the difference raises the question of whether induction of HCE is differentiate level of the cervicovaginal epithelium as being of related to the estrogen-receptor complex. The common denom crucial importance. inator for the 4 active estrogens may be some type of active The work ' 'adenosis' ' has sometimes been used too liberally. metabolite. Target tissue peroxidase or cytochrome P-450 may Thus, the vaginal changes described as adenosis after neonatal oxidize DES and dienestrol to p-quinone and the 2 epimers of antiestrogen treatment in rats correspond to the present au estradiol to catechol estrogens, eventually epoxides or qui thors' HCE (4). ñones (49); both 2-hydroxy derivatives of estradiol and p- In preparations from 6-day-old females given injections of quinone bind covalently to nucleic acids (46, 47). This implies that we must postulate some minor formation of similar reactive 5 J. A. McLachlan, R. R. Newbold. and B. C. Bullock, personal communication. metabolites from exogenous progesterone (31) or this steroid
FEBRUARY 1981 727
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. J-G. Forsberg and T. Kalland influences the activity of the enzymes responsible for formation Vaginal and cervical tumors induced by neonatal estrogen of active estrogen metabolites. The target tissue for the estro have most often been classified as being of squamous type (2, gen effect leading to HCE, the untransformed cervicovaginal 8, 64) even though glandular elements may occur (2, 41). columnar miillerian epithelium, responds to estrogen differently Typical for the present material was the dominance of glandular than the epithelium in the same region after the transformation structures with 3 adenocarcinomas and 6 mixed tumors being and also differently than uterine epithelium (11, 14, 42). An represented while only one was classified as being of squa estrogen interference with stromal-epithelial interactions in the mous type. All tumors were of a well-differentiated type. cervicovaginal region must be taken into consideration (6, 7). Accepting the line of development from HCE to malignancy Under continuous estrogen treatment of neonatal mice (Days in mouse, nonepithelial factors may interact with aberrant epi 1 to 5 after birth), the mitotic rate increases initially in the thelial cells to result in final overt malignancy. Central in this uterine epithelium and later declines (14). In contrast to this, discussion is the role of, e.g., the endogenous hormonal milieu the mitotic rate is reduced in the cervical epithelium, from 18 and the state of the immune system. In a series of papers, hr after the first injection and later to approach a level just permanent changes in the immune system and lymphocyte above zero. We have no evidence for a colchicine-like effect of populations have been described from perinatally DES-injected estrogens in the cervical epithelium (14, 53, 58). The results mice6 (32-36, 38, 44). Especially interesting is the reduced from the cervical mitotic rate studies are supported by those activity of natural killer cells (34). A daily neonatal dose of 5 from cell kinetic analyses in the same regions and studies on /ig 17/S-estradiol did not interfere with the future lymphocyte DMA synthesis in autoradiograms (10, 11). The depressed mitogen response to concanavalin A or lipopolysaccharide mitotic rate has been linked to development of HCE (16) while 5 fig DES did (38). The 52-week-old females given because the normal transformation process is paralleled by a injections of 17/î-estradioldid not show any signs of malig high mitotic activity (12). In this study, the DES mitotic rate- nancy in the cervix of vagina. The importance of the disturbed inhibiting effect in the cervical epithelium of 6-day-old females immune functions for a more general tumor development in was demonstrated to be dose dependent. The curve for mitotic DES-treated females is under study in our laboratory. rate inhibition had a plateau at 1CT3 to 10~2 jug DES which Human endometrial carcinomas, associated with estrogen corresponds to a critical level for HCE induction. Another intake, are usually described to be of an early-stage and well- consequence of the same estrogen effect is the typical simple differentiated type but without aggressive behavior (66). With structure and reduced size of the CCC which is independent of respect to these characteristics, the similarity to the experi HCE and generally occurs in NMRI mice treated neonatally mental system described in this paper is obvious. For both with estrogen. Similar observations were made by Plapinger types of lesions, a more appropriate designation could be "pseudocancer" as proposed for the human lesions by Weiss and Bern (51). From 8 to 36 weeks and later, there was a progressive and Sayretz (66). The human vaginal and cervical clear cell development of HCE into adenosis of different extent, classified adenocarcinoma is of quite another and more aggressive type as local or general adenosis. Rather unexpectedly, a single (29, 56). dose of 5 /ig of DES on the day of birth could result in general adenosis. A probable explanation for this is a prolonged ré REFERENCES sorption of above-threshold dose of DES from the oil solution 1. Anstall. H. B. Steroidal hormonal analogues. In: H. V. Richenberg (ed.). depot. Within general adenosis, a further progression occurred Biochemistry of Hormones. Vol. 8, Ser. 1. pp. 2-282. London: Butterworths. into different types of malignancy with invasion of the whole 1974. 2. Bern, H. A.. Jones. L. A., Mills, K. T., Kohrman, A., and Mori, T. Use of the cervical wall to the mesothelial surface. The first examples of neonatal mouse in studying long-term effects of early exposure to hormones malignancy were seen in preparations from 44-week-old fe and other agents. J. Toxicol. Environ. Health., ) (Suppl. 1): 103-116, 1976. males given injections of 5 /¿gofDES neonatally; at 17 months, 3. Bibbo. M., Gill, W. B., Ozizi, F., Blough, R., Fang, V. S., Rosenfield, R. L., the incidence was about 35%. Even a daily dose of 10"' DES Schumacher, G. F. B., Sleeper, K., Sonek, M. G., and Wied. G. L. Follow-up study of male and female offspring of DES-exposed mothers. Obstet. Gyn- resulted in malignancy as seen in a 14-month-old female. No ecol.. 49: 1-8, 1977. examples of carcinoma or suspected malignancy were found 4. Chamness. G. C., Bannayan, G. A., Landy, L. A., Jr., Sheridan, P. J., and McGuire, W. L. Abnormal reproductive development in rats after neonatally after neonatal treatment with 17/?-estradiol, but general aden administered antiestrogen (tamoxifen). Biol. Reprod.. 21: 1087-1090, osis occurred. In the human DES-exposed population, there is 1979. as yet no definite proof for a direct transition of adenosis into 5. Chernayaev, G. A.. Barkova, T. I., Egorova, V. V., Sorokina, I. B., Anan- chenko. S. N.. Mataradse, G. D., Sokolova, N. A., and Rosen, W. B. A series clear cell adenocarcinoma even though the 2 lesions occur of optical structural and isomerie analogs of estradiol: a comparative study together (56). of the biological activity and affinity to cytosol receptor of rabbit uterus. J. Steroid. Biochem., 6: 1483-1488, 1975. When the terms carcinoma or malignancy are used in this 6. Cunha, G. R. Epithelial-stromal interactions in development of the urogenital connection, they refer to the morphological appearance of the tract. Int. Rev. Cytol., 47: 137-194. 1976. lesions: cellular pleomorphism; cellular proliferation; and signs 7. Cunha, G. R., Chung. L. W. K., Shannon, J. M., and Reese, B. A. Stromal- epithelial interactions in sex differentiation. Biol. Reprod., 22. 19-42, 1980. of invasion of cervical stroma. Definite proof for malignancy in 8. Dunn. T.. and Green. A. W. Cysts of the epididymis, cancer of the cervix, our mouse model is lacking. We have not observed métastases. glandular cell myoblastoma and other lesions after estrogen injection in The latter finding is in agreement with our experience from newborn mice. J. Nati. Cancer Inst., 31: 425-455, 1963. 9. Doskeland. S-O., Kalland, T., and Forsberg, J-G. A study of the estradiol- studies of methylcholanthrene-induced cervical carcinomas induced augmentation of a specific cell product in the vaginal epithelium of (21, 22) which can grow to an enormous size locally but will the neonatal mouse. Dev. Growth Differ., 21: 111-118. 1979. 10. Eide. A. The effect of oestradiol on the DNA synthesis in neonatal mouse only rarely metastasize. However, cervical and vaginal tumors uterus and cervix. Cell Tissue Res.. 156: 551-555, 1975. induced by sex steroids in the neonatal period have been maintained in transplant and in serial passage in host animals 6 S. C. Ways, P. B. Blair, H. A. Bern, and M. O. Staskawicz. personal (41). communication.
728 CANCER RESEARCH VOL. 41
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. Neonatal Estrogen Treatment and Cervicovaginal Abnormalities
11. Eide, A. The effect of estradici on the cell kinetics in the uterine and cervical Butterworths, 1974. epithelium of neonatal mice. Cell Tissue Kinet., 8: 249-257, 1975. 41. Kohrman, A. F. The newborn mouse as a model for study of effects of 12. Forsberg, J-G. Mitotic rate and autoradiographic studies on the derivation hormonal steroids in the young. Pediatrics, 62: 1143-1149. 1978. and differentiation of the epithelium in the mouse vaginal anläge Acta An.il 42. Kvinnsland. S.. and Forsberg. J-G. Adenylate cyclase activity in mouse 62: 266-282, 1965. vaginal epithelium. Age related changes in basal activity and estradiol 13. Forsberg. J-G. The development of atypical epithelium in the mouse uterine sensitivity. Dev. Growth Differ.. 79: 71-75, 1977. cervix and vaginal fornix after neonatal estradiol treatment. Br. J. Exp. 43. Lamb, J. C., IV, Stumpf, W. E.. Newbold, R. R., and McLachlan, J. A. Pathol., 50: 187-195, 1969. Transitional changes in the surface epithelium of the cycling mouse vagina, 14. Forsberg, J-G. An estradiol mitotic rate inhibiting effect in the müllerian cervix, and uterus: scanning electron microscopic studies. Biol. Reprod.. epithelium in neonatal mice. J. Exp. Zool., / 75. 369-374, 1970. 19: 701-711, 1978. 15. Forsberg, J-G. Estrogen, vaginal cancer and vaginal development. Am. J. 44. Luster, M. I., Faith, R. E.. McLachlan, J. A., and Clark, G. C. Effect of in Obstet. Gynecol., 1 13: 83-87, 1972. utero exposure to diethylstilbestrol and the immune response in mice. 16. Forsberg, J-G. Cervicovaginal epithelium: its origin and development. Am. Toxicol. Appi. Pharmacol.. 47: 279-285. 1979. J. Obstet. Gynecol., 775: 1025-1041, 1973. 45. McLachlan, J. A. Prenatal exposure to diethylstilbestrol in mice: toxicological 17. Forsberg, J-G. Late effects in the vaginal and cervical epithelia after injec studies. J. Toxicol. Environ. Health. 2: 527-537. 1977. tions of diethylstilbestrol into neonatal mice. Am. J. Obstet. Gynecol., 121: 46. Metzler, M., and McLachlan, J. A. Oxidative metabolites of diethylstilbestrol 101-104, 1975. in the fetal, neonatal and adult mouse. Biochem. Pharmacol., 27: 1087- 18. Forsberg, J-G. Adenosis and clear-cell carcinomas of vagina and cervix. 1094, 1978. Am. J. Pathol., 84: 669-672. 1976. 47. Metzler, M., and McLachlan. J. A. Peroxidase-mediated binding of diethyl 19. Forsberg, J-G. Developmental mechanism of estrogen-induced irreversible stilbestrol and catechol estrogens to DNA: a potential factor for their carci- changes in the mouse Cervicovaginal epithelium. Nati. Cancer Inst. Monogr, nogenicity. Cancer Treat. Res., 63: 1181, 1979. 51: 41-56, 1979. 48. Nomura. T., and Kanzaki, T. Induction of urogenital anomalies and some 20. Forsberg, J-G. Physiological mechanisms of diethylstilbestrol organotropic tumors in the progeny of mice receiving diethylstilbestrol during pregnancy. carcinogenesis. Arch. Toxicol. Suppl., 2. 263-274, 1979. Cancer Res., 37: 1099-1104. 1977. 21. Forsberg, J-G.. and Stray Breistein, L. A synergistic effect of oestradiol and 49. Paul, S. M., and Axelrod, J. Catechol estrogens: localization, formation and prolactin influencing the incidence of 3-methylcholanthrene induced cervical actions in brain and other estrogen-sensitive tissues. In: E. Genazzani, F. D. carcinomas in mice. Acta Pathol. Microbiol. Scand. Sect. A, 84: 384-390, Carlo, and W. I. P. Mainwaring (eds.). Pharmacological Modulation of Steroid 1976. Action, pp. 1-10. New York: Raven Press, 1980. 22. Forsberg, J-G., and Stray Breistein, L. Prolactin and 3-methylcholanthrene 50. Pienkowski. M. M., Mann, L. C., Rosloniec, E. F., and Welsch, C. W. induced cervical carcinoma. Effect of bromocriptine. Acta Pathol. Microbiol. Accretion of biopsy specimens of vaginal adenosis from patients exposed in Scand. Sect. A, 87. 151-156, 1979. utero to diethylstilbestrol, when transplanted to athymic nude mice. J. Nati. 23. Fowler, W., and Edelman, D. A. In utero exposure to DES. Evaluation and Cancer Inst., 62: 521-523, 1979. followup of 199 women. Obstet. Gynecol., 51: 459-463, 1978. 51. Plapinger, L., and Bern, H. A. Adenosis-like lesions and other Cervicovaginal 24. Gorski, R. A., HarÃan, R. E., and Christensen. L. W. Perinatal hormone abnormalities in mice treated perinatally with estrogen. J. Nati. Cancer Inst.. exposure and the development of neuroendocrine regulatory processes. J. 63:507-518, 1979. Toxicol. Environ. Health, 3. 97-121, 1977. 52. Puca, G. A., and Bresciani. F. Interactions of 6,7-3H-17/j-estradiol with 25. Graham. C. E. Cyclic changes in the squamo-columnar junction of the mouse mammary gland and other organs of the C3H mouse in vivo. Endocrinology. cervix uteri. Anat. Ree., 155: 251-260, 1966. 85: 1-10, 1969. 26. Greenman, D. L., Dooley, K., Breeden, C. R., and Gross, G. H. Strain 53. Rao, P. N., and Engelberg, J. Structural specificity of estrogens in the differences in the response of the mouse to diethylstilbestrol. J. Toxicol. induction of mitotic chromatid non-disjunction in HeLa cells. Exp. Cell Res., Environ. Health. 3: 589-597, 1977. 48: 71-81, 1967. 27. Greenman, D. L., Delongchamp, R. R., and Highman. B. Variability of 54. Reddy, V. V., Naftolin. F., and Ryan, K. J. Aromatization in the central response to diethylstilbestrol: a comparison of inbred with hybrid mice. J. nervous system of rabbits: effects of castration and hormone treatment. Toxicol. Environ. Health. 5. 131-143, 1979. Endocrinology. 92: 589-594, 1973. 28. Herbst, A. L., Cole, P., Colton. T., Robboy, S. J., and Scully, R. E. Age- 55. Robboy, S. J., Keh, P C., Nicherson, R. J., Helmanis, E. K., Prat, J., incidence and risk of diethylstilbestrol-related clear cell adenocarcinoma of Szyfelbein, W. M.. Taft. P. D.. Barnes, A. B., Scully, R. E., and Welch, W. R. the vagina and cervix. Am. J. Obstet. Gynecol., 128: 43-48, 1977. Squamous cell dysplasia and carcinoma in situ of the cervix and vagina after 29. Herbst, A. L., Scully. R. E., Robboy, S. J., and Welch, W. R. Complications prenatal exposure to diethylstilbestrol. Obstet. Gynecol.. 57: 528-535. of prenatal therapy with diethylstilbestrol. Pediatrics, 62 (Suppl.): 1151- 1978. 1159. 1978. 56. Robboy. S. J.. Scully. R. E., Welch, W. R., and Herbst. A. L. Intrauterine 30. Herbst, A. L., Ulfelder, H., and Poskanser, D. C. Adenocarcinoma of the diethylstilbestrol exposure and its consequences. Arch. Pathol. Lab. Med.. vagina: association of maternal stilbestrol therapy with tumor appearance in 707: 1-5, 1977. young women. N. Engl. J. Med., 284. 878-881, 1971. 57. Rustia, M., and Shubik, P. Transplacental effects of diethylstilbestrol on the 31. Kadis, B. Steroid epoxides in biological systems: a review. J. Steroid genital tract of hamster offspring. Cancer Lett., 7: 139-146, 1976. Biochem., 9: 75-81, 1978. 58. Sawada, M., and Ishidate, M. Colchicine-like effect of diethylstilbestrol (DES) 32. Kailand. T. Decreased and disproportionate T-cell population in adult mice on mammalian cells in vitro. Mutât.Res., 57: 175-182, 1977. after neonatal exposure to diethylstilbestrol. Cell. Immunol., 57: 55-63. 59. Shah. H. C., and McLachlan, J. A. The fate of diethylstilbestrol in the 1980. pregnant mouse. J. Pharmacol. Exp. Ther., 797: 687-696, 1976. 33. Kailand. T. Alterations of antibody response in female mice after neonatal 60. Takasugi, N. Morphogenesis of estrogen-independent proliferation and cor- exposure to diethylstilbestrol. J. Immunol., 724: 194-198, 1980. nification of the vaginal epithelium in neonatally estrogenized mice. Proc. 34. Kalland. T. Reduced natural killer activity in female mice after neonatal Jpn. Acad.. 47: 193-198. 1971. exposure to diethylstilbestrol. J. Immunol., 724 1297-1300, 1980. 61. Takasugi, N. Cytological basis for permanent vaginal changes in mice treated 35. Kalland, T., and Forsberg, J-G. Delayed hypersensitivity response to oxa- neonatally with steroid hormones. Int. Rev. Cytol., 44: 193-224. 1976. zolone in neonatally estrogenized mice. Cancer Lett., 4: 141-146. 1978. 62. Takasugi, N., Bern, H. A., and De Ome, K. B. Persistent vaginal cornification 36. Kalland. T., and Forsberg, J-G. Permanent inhibition of capping of spleen in mice. Science (Wash. D.C.), 738: 438-439, 1962. lymphocytes from neonatally oestrogen-treated female mice. Immunology, 63. Takasugi, N., and Kamishima. Y. Development of vaginal epithelium showing 39:281-284, 1980. irreversible proliferation and cornification in neonatally estrogenized mice: 37. Kalland, T., Fossberg, T. M., and Forsberg, J-G. Localization of 3H-estradiol- an electron microscope study. Dev. Growth Differ., 75: 127-140, 1973. 17/? in diethylstilbestrol-induced adenosis. Obstet. Gynecol., 57: 464-467, 64. Takasugi, N., Kimura, T., and Mori, T. Irreversible changes in mouse vaginal 1978. epithelium induced by early postnatal treatment with steroid hormones. In: 38. Kalland, T., Strand, 0., and Forsberg, J-G. Long-term effects of neonatal S. Kazda and H. Denenberg (eds.). The Postnatal Development of Pheno- estrogen treatment on mitogen responsiveness of mouse spleen lympho type, pp. 229-249. Prague: Academia, 1970. cytes. J. Nati. Cancer Inst., 63: 413-421, 1979. 65. Verseé,V., and Barel, A. O. Rat o-fetoprotein. Purification, physicochemical 39. Katzenellenbogen, B. S., and Gorski, J. Estrogen actions on syntheses of characterization, oestrogen-binding properties and chemical modification of macromolecules in target cells. In: G. Litwack (ed.), Biochemical Actions of the thiol group. Biochem. J., 775: 73-81. 1978. Hormones, pp. 188-245. New York: Academic Press, Inc., 1975. 66. Weiss, N. S., and Sayvetz, T. A. Incidence of endometrial cancer in relation 40. King. R. J. B.. and Mainwaring, W. I. P. Steroid-cell interactions. London: to the use of oral contraceptive. N. Engl. J. Med., 302: 551 -554, 1980.
FEBRUARY 1981 729
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. J-G. Forsberg and T. Kailand
Fig. 1. A. cross-section through the CCC and the laterally bordering vaginal fornices (vf). First day of life. Note the high pseudostratified columnar epithelium, x 210. B, section from the upper part of the vagina after the epithelial transformation has taken place. The epithelium has a superficial zone of high columnar cells and a basal zone of cuboidal and more pale-staining cells, x 520. C, cross-section through the CCC from a female given an olive oil injection for the first 5 days after birth and 1 fig colchicine 4 hr before death on Day 6. Note the great number of metaphases superficially in the epithelium, x 210. D, section corresponding to that demonstrated in C; the female was given 1 /ig DES instead of olive oil. No metaphases are seen in this section from the CCC. In the bordering vaginal fornix (vl), both an area with untransformed columnar epithelium and a region with an estrogen-stimulated basal zone occur, x 210. E, High-power magnification to demonstrate a hyperplastic nodule of pale-staining cells below the high columnar cells. The latter belongs to the bordering untransformed columnar epithelium the cells of which show signs of estrogen stimulation, but no proliferation, x 520. Fig. 2. Sections from 8-week-old NMRI females given olive oil injections (A) or 5 ¿igDESfor the first 5 days after birth. A, normal configuration of the CCC and vaginal fornices (vt) The DES treatment results in a simpler cervical gross structure (B); the epithelial lining is of both a stratified squamous type and HCE. D, higher magnification to demonstrate the appearance of HCE. C, 2 typical cysts, lined with HCE, from the vaginal fornix. A, x 40; B, C, x 130; D. x 210. Fig. 3. Histological sections from the CCC and vaginal fornix from preparations most of which are included in the material shown in Chart 5. A, section through the CCC from a 40-week-old control female. In some of the females given injections 5 fig DES in the neonatal period there was a very pronounced reduction in size of the CCC. B. 17-month-old female. C. section through the CCC demonstrating HCE in the epithelial wall as well as downgrowth of 2 glandular-like elements, one with metaplasia; this represents a case of "local adenosis" (44-week-old female given 5 /ig DES on the day of birth) D, general adenosis" (36-week-old female given an injection of 1 /ig DES for 5 days after birth). E, regular arrangement of the columnar epithelium in the glandular-like elements of general adenosis (44-week-old female receiving 5 /ig DES for the first 5 days after birth). F, vaginal fornix from a 36-week-old female given 10 2 >ig DES for 5 days after birth. Note the occurrence of 2 types of squamous epithelium with an easily recognizable border. A. B, x 30; C. x 110; D, x 40; E, F, x 200. Fig. 4. A. section through the fornicate region from a female given injections of 10 ' fig DES for 5 days after birth and killed at the age of 14 months. The lesion was diagnosed as a well-differentiated squamous carcinoma. B, section from the uterine cervix from a female given injections of 5 fig of DES for 5 days after birth and killed at the age of 44 weeks. The lesion contained regions highly suggestive of adenocarcinoma. C, adenocarcinoma in a female given injections of 5 ng DES for 5 days neonatally and killed at the age of 17 months. Note nuclear atypia and formation of "glands within glands." D, section from the uterine cervix of a 17-month-old female given injections of 5 fig DES neonatally. The lesion was diagnosed as mixed carcinoma. A, x 35; B, x 45; C, x 225; D, x 285.
730 CANCER RESEARCH VOL. 41
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. Neonatal Estrogen Treatment and Cervicovaginal Abnormalities
!>Uk -JW*
ms•:',* ' ' m ••'••..'',r*aïirVK-i- il 1C "-''•' -w^'*^
FEBRUARY 1981 731
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. ***-3»v: '.. J ' •*8 » «.:'•.. .- •-'... :.i '^^^B
B?Ä1äa; •" " if^ÃŽV'V;/;?/ N • %^ÃÃ^}<:^ ••-'• Õj9^.¿T*S •UkJl^
v «Ã:
^-,-A '
' • *, • 2C
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. ••.-vç^r-•:-.-*r^v
', -3A 3B
•:" • -
.
- - .'. . .
•
V " ' •*;•'«'•\/'
riSMB'*•'^-
;3C 3D
:" :3F
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. .
~;'.¿4
>~
t •*Vo ÕSÕ v: -
' AA
— ..-.«—^»^E^ÕirV'^ S~5' ' , *'' '"<>'^^'¿f • ¿a.!fe¿SsÃ^%Ã^
¿%«ii>-*^r*W?^ - •" . •'i*1 .'-^ •--,i Ve ' ,Ã- '" ÕlÕS^teÕ^S^ÕAÕEnV»»tîCT&*T-^^ - v*'«*. V ».¿V-'V.- ' ^ L-' 734 CÕNCER RESEARCH VOL. 41 Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research. Neonatal Estrogen Treatment and Epithelial Abnormalities in the Cervicovaginal Epithelium of Adult Mice John-Gunnar Forsberg and Terje Kalland Cancer Res 1981;41:721-734. Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/41/2/721 E-mail alerts Sign up to receive free email-alerts related to this article or journal. Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected]. Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/41/2/721. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site. Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1981 American Association for Cancer Research.