Steroid Hormone Receptors and Melanoma



Steroid Hormone Receptors and Melanoma
0022-202X/ 80/ 7406-0379$02.00/ 0
Co pyri ght © 1980 by Th e Williams & Wilkins Co.
VoL 74, No. G
Printed in U. S.A.
Steroid Hormone Receptors and Melanoma
Division of S w gical Oncology and M CV- VCU Cancer Center, M edical College of Vi1ginia (JN), Richmond, Vi1ginia, and Medical Breast
Cancer Service, Medicine Branch, National Cancer Institute (ML), Bethesda, Mw )•land, U.S.A.
temperature dependent activation step whi ch permits nuclear
translocation a nd presumed binding to specific nucleru· sites.
Transcription of specific DNA segm ents ensues with appru·en t
regula tion at least at t h e level of t he primary mRNA tra nscript.
The mRNA produced is processed, transported to t he cyto plasm , and event ually t ranslated into specific proteins, th e end
product of steroid action. These protein products ar e felt to be
responsibl e for a ll of t h e phenotypic effects of steroids. Thu s,
although t h e presence of a cytoplas mi c receptor appears to be
a necessru·y prerequisite for t he steroid to exer t its specific
m echa nism of action, t he presence of th e receptor need not
necessarily signify hormonal responsiveness. This h as been
clearly demonstrated in patients with metastatic breast cancer,
where only 60 to 70 % of patients whose tumors contain estrogen
r ecep tor activity r espond to endocrine manipulation [7].
The possible dependence of huma n mela noma on endocrine
influences has been speculated upon for ma ny year s. Clinical
data suggest that females h ave a better prognosis t ha n mal es;
t h e improved survival in women is irrespective of site of the
primary lesion or clinical stage [1]. Numerous case r eports h ave
demonstrated an exacerbation of pre-existing melanomas during pregna ncy with regression followin g cessation of pregnancy.
Recently, Shiu et al [2] h ave shown in a retrospective review of
fe male melanoma patients that stage I patients wit h "activatio n" of the lesion during pregnancy h ad the sam e prognosis as
women wi t hou t any such association, bu t stage II pa tients
(those with regional lymph node m etastases) had a significantly
worse prognosis if the mela noma appeared or flar ed during
pregna ncy when compared to age-match ed nulliparous or paro us stage II patients wit hout any such "activation ." It t hus
a ppears from this clinical informat ion th at a subset of patients
with melanoma may h ave a hormonally responsive tumor.
Hormonal th er apy of disseminated mela noma h as been used
infrequentl y a nd is generally regarded as ineffective. However ,
o ne study has noted that a drug wi th glucocorticoid and a nt iestrogenic activities, N SC- 17256 (6a-methylpregn-4-ene3,11,20t rione) produced objective tumor regressions in 5 of 44 patients
[ 3]. Others have reported 2 cases of objective tumor regressions
fo llowing hypophysectomy, one patient who responded to testosterone, a nd one who responded to ethinylestr adiol [ 4]. So me
propor tion of patients may, therefore, h ave tumors which are
hormonally de pendent a nd ar e sensitive to endocrine manipula tions.
Several animal models h ave been studied for possible endocrine influences on melanoma . A transplantable mela noma in
h a msters h as been shown to possess a faster initial grow th rate
in females than males, bu t when ovariectom y was performed at
t h e time of t umor transplantation this difference was no longer
a pparent [5]. Lipkin h as shown that in vitro gr owth of a no th er
transpla nta ble h amster melanoma was inhibited by pharma cologically relevant concentrations of either test osterone or, to
a lesser extent, estradiol; in vivo, th e tumor grew more ra pidly
and res ulted in shorter survival in females t han in males [6].
Thus, these experimental findings may bear some relationship
to th e above noted human da ta and suggest the possibility t hat
some melanoma may be hormonally sensit ive.
Different classes of steroid hormones are thought to exe rt
th eir speciJic effects on cells th1·ough similar mec hanisms. Cell
membran es are fr eely permeable to steroid hormones. Once
inside the cell the steroid binds to an intrace llular protein (th e
" receptor") . The receptor-steroid complex t hen undergoes a
Reprint requ ests to: James P. Ne ifeld, M.D., P.O. Box ll , MCV
Station, R ichm ond, Virgin ia 23298.
The clinical and experimental data described above led to
our preliminary investigations reported in 1976 [8.9]. Using a
dextr a n-coated ch ar coal assay, separate estrogen, progesteron e,
glu cocorticoid, a nd androgen binding activi ties we re evalu ated
in a series of 35 patients. Estrogen binding activity was detected
in 45 %, progesterone binding activity in 22 %, androgen binding
activity in 17%, a nd glucocorticoid binding activity in 19%. The
amount of estrogen binding activity was generally low (w hen
compared to breast cancer ) with th e highest value being 91
fem to moles/mg of cytosol protein. Nevertheless, m any of th e
dissociation constants obtained by Scatchard analysis were in
the ra nge previously observed in breast cancer a nd in other
endocrine responsive tissues. In 2 cases with estrogen binding
activity where enough tumor tissue was available to p erform
specificity studies, the binding of 3 H-estra diol was competed for
only by compounds with estrogeni c or anti-estroo-enic activities·
thus, the binding of estradiol noted in melano~a a ppeared t~
be similar in specificity to the estrogen receptor activity noted
in breast can cer. Unfortunately, sucrose density gradient analyses were no t performed to compar e sedimentation ch aracteristics to those found in other estrogen target tissues. We concluded from this s mall series t hat melanoma appeared to bind
steroid hormones in a m a nner similar to that fo und in endocrine-responsive tissues and tumors a nd we hypothesized t h at
th ese investigations migh t allow selection of a subset of patients
whose t umors migh t r espond to endocrine therapy.
Since our preliminary observations we have extended our
series to 111 patien ts a nd oth ers have also investigated hormone
binding in patients wit h melanoma.
' Estrogen binding activi ty has been a nalyzed in all 111 patients. This assay has been validated both by its predictive
accuracy in metastatic breast cancer [10] and by co mparison
with r esul ts obta ined with reference powders. Twenty-fo ur
patients had estrogen binding activity with valu es ranging fro m
4 to 91 fmo ls/ mg of cytosol protein; onl y 9 patients had values
Vol. 74, No. 6
greater than 20 and only 1 greater th an 50. Dissociation constants (K 0 ) ranged from 0.147 to 8.15 nM; only 3 were greater
than 3 nM . Specific binding of th e labeled estradiol was inhibited
by 17,8-estradiol, diethylstilbestrol, and tamoxifen, all compounds which bind to the estrogen r eceptors demonstrated in
other tissues; significant competition was no t observed at up to
2000-fold molar excesses of 17a-estradiol, dihydrotestosterone,
progesterone, testostero ne, or cortisol. Mi gration on sucrose
density grad ients in 2 cases showed 4S peaks, bu t we have been
unable to detect a defini te 8S peak. The presence of binding
activity did not correlate with sex, site, or stage of disease.
Several other groups have assayed smaller numbers of patients for specific estrogen binding activi ty . Creagan eta! (personal communication) , at the Mayo Clinic, assayed 38 melanoma specimens in 34 patients; 4 patients had estrogen binding
activity noted but the values were only 3,3,4 and 5 fm ol/mg of
cytosol protein. The patient with a value of 5 fmoljmg had 2
subseq uent biopsies that were negative. Kokoschka a nd colleagues found that 15% of 41 melanomas were estrogen receptor
positive. He also noted t hat 81% were a ndrogen receptor positive and claimed all to be progestin rece ptor positive [11].
Cha udhur i and his colleagues (personal communi cation) , at the
University of Illinois, found estrogen binding activity in 10 of
27 (37 %) patients; their values also te nded to be low and their
dissociation constan ts were in the range of 0.1 to 1 nM . On
sucrose density gradients only 4S activity has been detected. It
is of interest that they also investigated benign nevi; no nevi
from 15 patients without melanoma contained estrogen binding
activity, but 9 of 22 benign nevi from patients with melanoma
did contain estrogen binding activity [1 2]. The significance of
this is unclear, but it is possible t hat melanocytes that contain
estrogen binding activity may be "activated" and may be premalignant.
We have also investigated binding activity for other steroid
hormones in melanoma. Progesterone binding was detected in
18 of 85 (21 %) patients; th e values ranged from 11 to 320 fmol/
mg of cytosol protein (3 values were greater t han 50 fmol/mg)
with dissociation constants of 0.45 to 18 nM (only 2 values were
larger t han 6 nM). Androgen binding activity was present in 11
of 71 (15 %) patients; concentrations were very low, ranging
from 4.7 to 57.9 fmol/mg of cytosol protein, only 4 greater tha n
15 fmol/mg. Dissociation constants ranged from 0.3 to 30 nM.
Glucocorticoid binding activity was found in 18 of 65 (28%)
patients with values from 10 to 300 fmol/mg of cytosol protein
and dissociation constants of about 2 to 5 nM. When t hese
values are compared to the values usually seen in breast cancer
or other r eceptor containing, endocrine-responsive tissues the
binding capacities are quite low bu t the dissociation constants
are similar. Clearly, high affinity binding data are not equivalent
to demonstration of true receptors. Detailed specificity studies,
sucrose density gradient behavior, nuclear translocation, and
association with hormonal response would all tend to suggest
that these binding activities might be true receptors. The
limited specificity studies for estrogen binding activity are
similar to those no ted in endocrine responsive tissues, bu t other
specificity data are not yet available. Finally, sucrose density
gradient analysis both in our la boratories and by others has no t
yet demonstrated the patterns observed in breast cancer, uterine tissue, or oth er steroid-hormone responsive tissues. It does
appear, however, that high affinity, limited capacity binding for
4 classes of steroid hormones may be fo und in human melanoma.
Human melanoma cells being grown in long term tissue
cul ture may offer an opportunity to correlate the presence or
abse nce of steroid hormone binding ac tivi ty with responsiveness
to endocrine th erapy. We have investigated 21 melanoma cell
lines and have been unable to detect estrogen binding activity;
of th e 5 cell lines tested to date for estrogen responsiveness,
none have demonstrated a ny effect of estrogenic or anti-estrogenic compounds on aH -thymidine or 14 C-leucine incorporation.
Chaudhuri et al [13] have reported th at 4 of 6 mela noma cell
lines contained estTOgen binding activity with val ues ra nging
from 62.7 to 238 fmol/mg cytosol protein and dissociatio n
constan ts from 0.3 to 4 nM . Progesterone binding activity was
present in 3 of 6 cell lines and no a ndrogen binding activity was
detected. Cell lines that contained specific binding appeared to
be responsive (as measured by increase in celt numbers) to
physiologic concentrations of the a ppropriate steroid ; cell lines
lac king such binding showed no r esponse to steroid hormone
administration. These data have not yet been confir med with
other cell lines (Neifeld JP, unpublished observations).
Beattie et a! [14] have investigated the effect of sex on the
growth of melanoma in nude mice. They compared one cell line
that contained estrogen binding activi ty (ER+) to one that did
not (ER-) for growth after transplantation into nude mice. The
ER- line exhibited increased latency when compared to t he
ER+ line, but there was no difference in tumor incidence or
growth in males. The ER- cell line grew faster in females than
did th e ER+ cell lin e. In other work from th e sam e labo ratory
[1 5], estradiol decreased tumor ta ke in the estrogen binding cell
line and late ncy was increased, although there was no effect of
estradiol on tumor growth. They interpreted these observations
as being consistent with estrogen-inhibiting growth of th e ER+
cell lines. They felt this was consistent with the more favorable
prognosis of melanoma in women [16]. This work has yet to be
confirm ed by other investigators.
The initial reports describing steroid hormon e receptors in
mela nom a (vide supra) and suggesting their presence migh t
indicate endocrine responsiveness have led to several clinical
studies. At the N atio nal Cancer Instit ute 35 patients with
measurable disease have been treated with diethylstilbestrol
(15 mg/ day). There were 2 objective responses (5.7 %), but
neither patient's tumor con tained estrogen binding. M eyskens
(personal communication), at the U niversity of Arizona, h as
treated 11 patients with T amoxifen 20 mg/day. One patient
had a complete response and 3 had partial responses (>50%
regression of tumor). Estrogen binding analyses were no t performed. Creagan et al (personal communication) have treated
25 patients with measurable, advanced disease by administering
T amoxifen 40 mg/day. No patient had an objective 1·esponse
although 2 patients did have less than 50% regressions of tumor.
The median time to progression was about 30 days and median
sw·vival was about 100 days. In a recently reported trial, one of
15 evaluated patients achieved a partial regression of soft tissue
disease with medroxyprogesterone acetate (500 m g/day) [1 7].
Steroid binding assays were not reported. Thus, these stud ies
suggest that few patients with metastatic melanoma, even if
t heir tumors might contain steroid hormone binding activity,
will be responsive to endocrine manipulation.
The a pparent discrepa ncy between t he presence of specific
steroid hormone binding in human m elanoma and lack of
substantial response rates to endocrine therapy is intriguing.
There are several possible explanations which require fu ture
study. First, as suggested above, the presence of steroid hormone receptors does no t necessarily signify steroid hormone
responsiveness. There may be defects in translocation of receptor, transcription, translation, or, in fact, th e growth and r eplication of melanoma cells may be independent of an intact
steroid hormone pathway.
A second possibility is t hat the steroid binding represents
binding to either an inactive receptor or another protein present
in some melanomas. Although binding constants a nd s pecifici-
June 1980
t ies appear similar to those noted in endocrin e r esponsive
t umors a nd normal tissues, t he pattern of migration on sucrose
d e nsity gr adients has not yet been shown to contain a n 8S peak.
S in ce the basic genetic " machinery" in a ll cells is the same,
d e r epression of the chromat in segm ent that codes for steroid
h o rmon e binding activity could result in synthesis of a receptor
(or portion of a rece ptor) t ha t has no biological significance.
The obj ective response t o steroid hormones by tumors that
lack s pecific binding is also of gr eat inte rest. This m ay be due
to tec hnical errors in the assay. It could also be due to an
indirect effect of t he steroid on tumor growt h. For example,
estrogens ar e known to affect the activity of MSH (as well as
g ona dotrophins, prolactin, glucocorticoids, etc. ), a nd thJ·ough
s u ch an effect the a pparent inhibition (or stimulation) of tumor
growth could occm. Another possibili ty migh t be that the
hormone effec t could be mediated through some intracellular ,
n o nTeceptor mechanism. Currentl y, experiments in our and
oth er la boratories ar e investigating t hese possibiMies.
Although several te ntative conclusions may be te ndered regarding steroid hormone receptors and melanoma , it must be
re m embered t hat only a small number of pa tients have been
investigated and tha t these statem ents may change as more
d a ta become availabl e:
1. R etrospective studies suggest a small subset of patients
with mela noma m ay have t umors responsive to steroid hormones.
2. The da ta from several centers suggest that a small perce ntage of patients have tum ors t hat contain specific high
a ffini ty, limited capacity binding for steroid hormones.
3. The presence (or absence) of steroid hormone binding
activity in melanoma has no t been shown to correlate with in
v ivo (and possibly in vitro) responses to endocrine manipulation.
l. Whi te LP: Studi es on mela noma: II. Sex a nd surviva l in huma n
melanoma. N Eng! J Med 260:789-797, 1959
2. S hiu MH, Schottenfeld D, MacClean B, Fort ner J G: Adv erse effect
of preg na ncy on mela noma : A rea ppra isal. Ca ncer 37:181- 187,
3. Jo hnso n R O, B isel H , An dJ"ews N, W ilso n W, Rochlin D, Segaloff
A, Krementz E, Aust J , Ansfield F: P hase I clinical st udy of 6a·methylpregn-4-ene-3, 11 ,20-t io ne (NSC- 17256). Ca nce r Ch emoth er R ep 50:67 1-673, 1966
4. Bodenh a m D C, H a le B: Maligna nt m ela noma, E nd oc rin e T herapy
in Maligna nt Disease. Ed ited by B A Stoll. Lo ndon, W.B. S a un ders Co., Ltd., 1972, pp 377-383
5. Rosenberg J C, Assimaco poulos C, Rosenberg S A: T he ma li gna nt
mela noma of ha msters: III. E ffects of sex a nd castra ti on on t he
growt h of the transplan ted t um or. Ann NY Acad Sci 100:297304, 1963
6. Lipkin G: Sex facto rs in growt h of ma ligna nt mela noma in ha ms ters:
In vivo and in vit ro co rrelations. Ca ncer 11es 30:1 928- 1930, 1970
7. McG uire WL, Carb one PP, Sears ME , Escher GC: Estrogen receptors in huma n breast ca ncer: An overview, Estroge n Receptors in
Huma n Breast Ca ncer . Edi ted by WL McG uire, PP Ca rbone, EP
Vollmer. New York , R ave n P ress, 1975, pp 1-7
8. Fisher RI, Neifeld J P Lippma n ME: Oestroge n rece pto rs in hum a n
malignant mela nom a. La ncet ii:337- 339, 1976
9. Neife ld JP, L ippma n ME, Fish er RI: R eceptors fo r ster oid ho rmones in hum an mela noma. S urgical Fo rum 27:108-1 10, 1976
10. Allegra J C, Lippma n ME, Thompso n EB, S imon R, Barlock A,
Green L, Huff KK , D o HMT, Ai t ken S, Warren R: Steroid
horm one recepto rs in huma n breast ca nce r. P roc Am Assoc
Can ce r Res 19:336, 1978
11. Kokoschka E M, S pona J , Bieglmaye r Ch , Sch m idt J : H orm one
recepto r a na lysis in mali gna nt mela noma patients. Ca nce r Treat
R ep 63: 1200, 1979
12. Chaudhuri PK, Walker MJ , K ee hn D, B eattie CW, D asG up ta T K:
Estrogen recepto r in hum an benign nevi a nd skin. S urgical Foru m
30:128-129, 1979
13. Cha udhuri PK, Wa lker MJ , Beattie CW , D asG upLa TK: E ndoc rine
co rrelates of hum a n ma li gna nt mela noma. J S urg R es 26:214219, 1979
14. Beattie CW , Chaudhuri PK, Walker MJ , D asG up ta T K: InOu ence
of sex o n t he rate of growth of hum an mali gnant mela noma . P roc
Am Assoc Ca ncer Res 20: 130, 1979
15. Cha udhuri P K , Walker MJ, Bea tt ie CW , D asG up ta T J<: E ndocrine
influences of t he growt h of hum a n mela noma in vivo. Prese nted
at t he 32 nd Annu al M eeting of t he Society of S urgical Oncology,
Apnl17-21, 1979
16. Beattie CW, Cha udhuri PK , Wa lk er MJ , DasG upt.a T K: H ormo na l
regulatiOn of t he growth a nd meta tasis of huma n m ela noma .
Cancer Treat Rep 63: 11 99, 1979
17. Beretta G, T a biadon D, Fossati P: Clinical evalu atio n of medroxyprogeste rone acetate (M AP) in m ali gnant melanoma . Cancer
T reat Rep 63: 1200, 1979

Similar documents