- Email: [email protected]
Stimulation of aromatase activity in breast fibroblasts by tumor necrosis factor
8cmllobawand
ELSEV’IIZR
Molecular and Cellular Endocrinology 106 (1994) 17-21
Stimulation of aromatase activity in breast fibroblasts by tumor necrosis factor a Fiona Macdiarmida, D. Wanga, Loma J. Duncan~, A. Purohita, Margaret W. Ghilchikb, M.J. Reeda nUnit ofMetabolic Medicine, St. Mary’s Hospital Medical SchooI, Imperial College of Science, Technology and Medicine, London W IPG, UK ‘The Breast Clinic, St. Mary’s Hospitul, London W2 INY, UK Received 10 June 1994;accepted31 August 1994
AhStlWt
The conversion of ~dmsten~ione to es&one, the reaction mediated by the aromatase enzyme complex, may make an important contribution to the syntbesis of estrogens in breast tissues.In the present study, the effect of the cytokine, TNFa, on ammatase activity was examined in breast tibrublasts derived from normal and malignantbreast tissue. TNFa (2.5-10.0 rig/ml), in the presence of stripped fetal calf serum and dexamethasone, significantly stimulated Iibroblast aromatase activity in a dose-dependent manner. IL-1 and IL-6 also stimulated flbroblast aromatase activity, but no marked synergism between TNFa and IL-1 or IL-6 was detected. Using a specific radioimmunoassay, significant concentrations of TNFa were detected in samples of breast cyst fluid and breast tumor cytosol, which had previously been shown to stimulate aromatase activity, but not in conditioned medium from breast tumor-derived flbroblasts. As TNFcz may be preferentially expressed and produced in the adipose tissue component of the breast, this cytokine may have au important role in regulating estrogen synthesis in normal and malignant breast tissues. Key~onis: Breast cancer; Aromatase; Estrogens; Cytokines; Tumor necrosis factor a
1. Introduction Synthesis of estrogen from androstenedione within the breast may make an important contribution to the high estrogen concentrations which are present in normal and malignant breast tissue (Bonney et al., 1983; Van Landeghem et al., 1985; Vermeulen et al., 1986; James et al., 1987). The conversion of androstenedione to estrone is mediated by the aroma&se enzyme complex and aromatase activity is detectable in 60% of breast tumors (Tilson-Mallett et al., 1983). Whether the ~orna~e complex in breast tumors can produce sufficient estrogen to exert a mitogenic effect on breast cells is controversial (Bradlow, 1982). Using a double isotopic infusion technique, local synthesis of estrone from androstenedione was found to make a major contribution to tumor estrone content in some, but not all, tumors (Reed et al., 1989). The finding of an association between aromatase activity and the presence of a tumor in breast quadrants has suggested the possibility that increased aromatase activity in a breast quadrant may favor tumor *&tins
author.Tel. 071 725 1738. Fax 071 725 1790.
0303-7207/94/$07.00 0 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 0303-7207(94)03388-A
development (Miller and G’Neill, 1987). Alternatively, it is possible that breast tumors produce factors which are able to stimulate aromatase activity in adipose tissue adjacent to the tumor. Breast cyst fluid (BCF), breast tumor cytosol (BTC) and conditioned medium (CM) from breast tumor fibroblasts have all been shown to stimulate aromatase activity in breast tissue fibroblasts (Reed et al., 1993). Several cytokines and growth factors, including interleukin-1 (IL-l), IL-6 and insulin-like growth factor-type I (IGF-I) are detectable in these prep~ations and can stimulate aroma&se activity (Reed et al., 1992, 1993). Most of the cytokines or growth factors which have so far been found to stimulate aromatase activity in breast tumor derived fibroblasts are thought to originate from the epithelial or stromal components of the breast. The major part of the breast, however, is composed of adipose tissue and there is evidence that this tissue is important for tumor growth. In mice, the ability of mammary tumors to develop is inhibited in the absence of adipose tissue (Hoshino, 1962). There is now evidence indicating that adipose tissue may be a major site of tumor necrosis factor a (TNFa) expression and production
F. ~acdia~mid
18
et al. I ~a~ecula~ and Cefluiar E~ocrin~Io~
(Hotamisligil et al., 1993; Spiegelman et al., 1993). As cytokines appear to have an important role in regulating aroma&se activity, we have examined the ability of TNFa to regulate aroma&se activity in fibroblasts derived from normal or malignant breast tissues. The possibility that TNFct might also interact with other cytokines (IL- 1, B-6) was also investigated. In addition, concentrations of TNFa were measured in samples of BCF, BTC and CM, which have previously been shown to stimulate aromatase activity, to examine whether the presence of TNFa in these preparations could account for their aromatase stimulatory activity. 2. Materials and methods 2.1. Cultw-e of~broblasts from normal and rnal~g~~t breast tissues Two samples of normal breast adipose tissue were obtained from pre-menopausal women (subjects 1 and 2) undergoing reduction mammoplasty and one tumor sample from a post-menopausal woman {subject 3). Fibroblasts were prepared from these breast tissues as previously described (Adams et al., 1988). Cells were grown until confluent in Earle’s minimum essential medium (EMEM) containing 10% fetal calf serum (FCS). When confluent, cells were washed with Earle’s balanced salt solution (5 ml) and cultured for 24 h in phenol-red free, EMEM containing 10% stripped fetal calf serum (SFCS). Cytokines (TNFa, IL-l, IL-6, Bachem Inc., UK) were added in this medium in the presence of dexamethasone (100 nmol/l, Sigma, UK). To initially demonstrate that the fibroblasts possessed aromatase activity, cells were treated with dexamethasone (100 nmoU1) and cultured in 10% stripped fetal
c
I_
0
Fig.
I.
1
I.1
1.
-
2 PRFM
3
Ic 1.2’ Dw
3
f
lWnMf+SFCS
Aromatase activity in fibroblasts derived from normal
(I
and 2) or
malignimt breast tissue (3). Activity in fibroblusts cultured in phenol-red free medium containing
10% SFCS (PRFM)
was very low, but detect-
able. Culture of fibroblasts in stripped fetal calf serum (WCS) presence
of
dexomethasone
(Dex.,
IOOnM)
resulted
in the
in significant
(P < 0.001) stimulation of aromutase activity (mean f SEM, n = 3).
106 fl994) 17-21
calf serum, a combination known to stimulate aromatase activity (Simpson et al., 1981). 2.2. Aro~tase act~vi~ Aromatase activity was assayed in intact cell monolayers by measuring the release of 3Hz0 from [1@-3H]androstenedione [Du Pont (UK) Ltd.] over a 24-h period, as described previously (Reed et al., 1992, 1993) with the exception that fibroblasts were incubated with substrate for 24 h. Aroma~e activity was linear over this period and there was a significant correlation (r = 0.9, P < 0.01) between aromatase activity and incubation time. Results are expressed as femtomoles of product produced per million cells per 24 h, mean + SEM. 2.3. R~ioimm~~ass~y of TiVFa Concentrations of TNFa in BCF, BTC and CM were measured by a specific radioimmunoassay (Amersham Int., UK). Cross-reaction of other cytokines in this assay was
F. Macdiannid et al. I Molecular and Cellular Endocrinology 106 (1994) 17-21
T 1 L
C
2.5
5.0
10.0
C
2.5 5.0
1
10.0
i C
2.5
2
5.0
10.0
3
TNF(ng/ml)
Fig. 2. Effect of TNFa on aromatase activity in tibroblasts derived from normal (1 and 2) or malignant breast tissue (3). Treatments were added in phenol-ted free medium containing 10% SFCS and cells were cultured in this medium + TNFa in the presence of dexamethasone (100 nmolll). Addition of TNFa resulted in a significant stimulation of aromatase activity (P < 0.001) at all doses tested. Significant correlations were found between the concentration of TNFa added to cells and aromatase activity (mean f SEM, n = 3).
activity was examined (Fig. 2). TNFa stimulated fibroblast aromatase activity in a dose-dependent manner over the range 2.5-10.0 ng/ml. Analysis of the relationship between aromatase activity and the dose of TNFa tested using Spearman’s rank correlation revealed significant (P <
19
0.001) correlations: 1, p = 0.79; 2, p = 0.93; 3, P = 0.81. The effect of the lowest concentrations of TNFa tested appeared to be greater in the fibroblasts derived from malignant breast tissue, but further studies will be required to confirm the significance of this observation. Having shown that TNFa stimulated aromatase activity, its effectiveness was compared with other cytokines (IL-l and IL-6) which are known to stimulate aromatase activity in the presence of dexamethasone. As found previously, both IL-l and IL-6 stimulated aromatase activity in tumorderived fibroblasts with IL-6 having a greater stimulatory effect than IL-l (Fig. 3). In fibroblasts derived from malignant breast tissue, IL-6 (50 ng/ml) stimulated aromatase activity to a similar extent to TNFa (10 ng/ml). Higher concentrations of TNFa were not tested because of the potential anti-proliferative effect of this cytokine. In fibroblasts derived from adipose tissue from premenopausal women, aromatase activity was stimulated by IL-l and IL-6 (Fig. 3). However, the effect of IL-l and TNFa on aromatase activity was greater in fibroblasts derived from subject 2, with IL-6 having only a small effect on aromatase activity in these fibroblasts. For fibroblast derived from adipose tissue, maximal stimulation of aromatase activity occurred after treatment of fibroblasts from subject 2 with IL-l. The possibility that some interaction between the different cytokines might occur in their ability to stimulate fibroblast aromatase activity was also examined. For this, the effect of treating cells with IL-l (50 ng/ml) plus IL-6 (50 ng/ml), IL-l (50 ng/ml) plus TNFa (10 ng/ml) or IL-6
1
=Ooo I T
T
h P1 3ooo Q
+l
sf
!
E
11
I
2000
122
gK E
nglml
IL-1
IL-6
TNF
L-l
IL-6
TNF
60
50
10
50
50
10
2
_r IL-1
IL-6
50
50
1000
-
0
TNF 10
JY 55
IL-1
IL-1
1:6
T+NF T+NF 1
IL.6
IL-1
IL-1
IL-6
IL:
T;F
T+NF
2
IL-1 IL:
IL-1
IL-6
T+NF
T+NF
3
3
Fig. 3. Comparison of the effects of IL-l, IL-6 or TNFa on aromatase activity in fibroblasts derived from normal (I and 2) or malignant breast tissue (3) (mean 2 SEM, n = 3). Treatments were added in phenol-red free medium containing 10% SFCS + cytokines and were cultured in the presence of dexamethasone (100 nmolll).
Fig. 4. The effect of IL-l plus IL-6 or TNFa plus IL-1 or IL-6 on aromatase activity in fibroblasts derived from normal (I and 2) or malignant breast tissue (3). Additive values, derived from the effect of each cytokine alone on aromatase activity, are indicated at the top of each column (mean & SEM, n = 3). Treatments were added as indicated in the legends to the other figures.
20
F. Mucdiarmid et al. f Moleculur and Ceflulur Endocrinology 106 (1994) I741
(50 ng/ml) plus TNFG (10 nglml) was compared with the expected additive effect of treating fibroblasts with each of these cytokines (Fig. 4). No evidence of any synergistic interaction of the cytokines in their ability to stimulate aromatase activity was detected. In two of the three fibroblast cell lines, IL-l plus TNFa, when added together, produced an increase in aromatase activity close to that expected due to an additive effect (99% and 116%, values shown at top of the respective columns for fibroblasts from subjects 2 and 3 in Fig. 4). The addition of IL-I and IL-6 tended to reduce the effect expected from each cytokine (49-82%). In fibroblasts derived from adipose tissue, IL-6 plus TNFa increased the stimulatory effect when each cytokine was tested alone (122% and 140%), whereas for the fibroblasts derived from malignant tissue, their combined effect was reduced (55%). 3.1. Effect of cytokines on cell growth Fibroblasts only grow at a relatively slow rate and the present investigation was not primarily carried out to establish the effect of IL-l, IL-6 and TNFa on cell growth. Fibroblasts were only treated with cytokines for 48 h and during this time no significant effect on cell growth was detected.
of T’Fa in breast cyst fluid, breast tumor cytosol and conditioned medium from tumor derived fibroblasts 3.2. Concentrations
Significant concentrations of TNFa were detectable in samples of BCF previously shown to stimulate aromatase activity in cultured breast fibroblasts (820-l 120 pg/ml, mean + SD = 957 + 199 pg/ml, n = 7). No difference was detected between TNFa concentrations in BCF with high or low electrolyte ratios. TNFa was also measured in a BTC and CM previously shown to stimulate aromatase activity. TNFa was present in BTC (630 pg/ml) whereas TNFa was not detectable in CM from tumor-derived fibroblasts. 4. Discussion Results obtained from this study have shown that TNFa in the presence of SFCS and dexamethasone is a potent stimulator of aromatase activity in fibroblasts derived from both normal and malignant breast tissue and that it stimulates activity in a dose-dependent manner. As previously reported for IL-1 and IL-6, TNFa was only able to stimulate aromatase activity in the presence of dexamethasone. However, in the absence of serum and dexamethasone, TNF has previously been reported to inhibit aromatase expression in cyclic AMP stimulated adipose stromal cells (Simpson et al., 1989). It is not yet clear whether dexamethasone, as used in the present study, has a permissive role in allowing aromatase activity to be induced or whether it also influences cytokine receptor concentrations in fibroblasts. The ability of IL-l and IL-6 to stimulate
aromatase activity in fibrobI~ts derived from malignant breast tissue was confirmed in the present study. IL-6, however, appears to have a somewhat smaller stimulatory effect on aromatase activity in fibroblasts derived from breast adipose tissue. IL-6 is secreted by breast tumor fibroblasts (Adams et al., 1991; Reed et al., 1993), and it is likely that this cytokine may act in an autocrine manner in regulating aromatase activity in stromal cells within breast tumors. IL-l is also secreted by breast tumor fibroblasts although at a tower level than IL-6 (Duncan and Reed, 1994). Stimulation of aromatase activity by TNFa was similar to that of IL-6 in the fibroblasts derived from the malignant breast tissue but lower (markedly for subject 2) for fibroblasts derived from normal breast tissue. As cytokines can interact to stimulate the production of other cytokines or cytokine receptors, the effect of combining the different cytokines tested was also examined. No evidence was obtained from the present investigation that IL-l, IL-6 or TNFa interact to produce a synergistic increase in aromatase activity. Whereas IL-l plus TNFa generally acted to produce an additive effect, suggesting independent pathways of action, the addition of IL-l and IL-6 together resulted in a lower than additive response, suggesting common pathway of action in these fibroblasts. Some difference in the effect of IL-6 plus TNFa was noted between normal and tumor derived fibroblasts, but further studies are required to confirm these observations. Significant concentrations of TNFa were detected in samples of BCF and BTC, but not CM, previously found to stimulate aromatase activity (Reed et al., 1993). Although only a small number of BCF samples have been examined so far, no difference in TNFa concentrations was detected for BCF with high or low Na+/K+ ratios. This is in contrast to IL-6 concentrations in BCF which are significantly higher in BCF with high electrolyte ratios, but similar to IL-1 where no difference between the two cyst types was detected (Reed et al., 1992). Concentrations of TNFa in BCF and BTC were approximately 1 ngfml. While this concentration is somewhat lower than that used to stimulate fibroblast aromatase activity, albumin or albuminoid-like proteins, which are present in BCF and BTC, have been shown to potentiate the effect of cytokines and growth factors, such as IGF-I, in their ability to stimulate aromatase activity (Singh et al., 1992; Reed et al., 1993). TNFa was not detected in a sample of CM which had previously been shown to stimulate fibroblast aromatase activity, although high concen~ations of IL-6 are present in this CM (Reed et al., 1993). It is possible therefore that breast stromal cells are not a major site of TNFa production, and there is now evidence that adipose tissue is a major site of TNFa expression and production (Spiegelman et al., 1993; Hotamisligil et al., 1993). Previous investigations have revealed an association between the presence of a breast tumor and aromatase activity in breast quadrants in which the tumor is located (Miller and O’Neill, 1987). The finding that TNFa is able to stimulate aromatase activity
F. Macdiarmid et al. I Molecular and Cellular Endocrinology 106 (1994) 17-21
and evidence that adipose tissue is able to express and produce TNFa suggests the possibility that increased production of TNFa in a particular breast quadrant could lead to an increase in aromatase activity and estrogen synthesis sufficient to support tumor development. In summary, TNFa has been found to be a potent factor involved in the regulation of aromatase activity in fibroblasts derived from both normal and malignant breast tissues. Significant concentrations of TNFa were detected in BCF and BTC, previously shown to stimulate aromatase activity. As TNFa may be preferentially expressed and produced by the adipose tissue component of the breast, this cytokine may have an important role in regulating estrogen formation and the production of tumors within the breast. References Adams, E.F., Newton, C.J., Tait, G.H., Braunsberg, H., Reed, M.J. and James, V.H.T. (1988) Int. J. Cancer 42, 119-122. Bonney, R.C., Reed, M.J., Davidson, K., Beranek, P. and James, V.H.T. (1983) Clin. Endoctinol. 19.727-739. Bradlow, H.L. (1982) Cancer Res. 42,3382s-3386s. Duncan, L.J. and Reed, M.J. (1994) J. Steroid Biochem. Mol. Biol. 49, 63-68.
21
Hoshino, K. (1962) J. Natl. Cancer Inst. 29.835-843. Hotamisligil, G.S., Shargill, N.S. and Spiegelman, B.M. (1993) Science 259,87-91 James, V.H.T., McNeill, J.M., Lai, L.C., Newton, C.J., Ghilchik, M.W. and Reed, M.J. (1987) Steroids 50.269-279. Miller, W.R. and O’Neill, J.S. (1987) Steroids 50,537-547. Reed, M.J., Coldham, N.G., Pate& S.R., Ghilchik, M.W. and James, V.H.T. (1992) J. Endocrinol. 132, R5-R8. Reed, M.J., Owen, A.M., Lai, L.C., Coldham, N.G., Shaikh, N.A. and James, V.H.T. (1989) Int. J. Cancer 44.233-237. Reed, M.J., Topping, L., Coldham, N.G., Purohit, A., Ghilchik, M.W. and James, V.H.T. (1993) J. Steroid Biochem. Mol. Biol. 44, 589596. Simpson, E.R., Ackerman, G.E., Smith, M.E. and Mendelson, C.R. (1981) Proc. Natl. Acad. Sci. USA 78,5690-5694. Simpson, E.R., Merrill, J.C., Hollub, A.J., Graham-Lorence, S. and Mendelson, C.R. (1989) Endocr. Rev. 10, 136-148. Singh, A., Blench, I., Morris, H.R., Savoy, L.-A. and Reed, M.J. (1992) Mol. Cell. Endocrinol. 85, 165-173. Spiegelman, B.M., Choy, L., Hotamisligil, G.S., Graves, R.A. and Tontonoz, P. (1993) J. Biol. Chem. 268.68236826. Tilson-Mallett, N., Santner, S.J., Feil, P.D. and Santen, R.J. (1983) J. Clin. Endocrinol. Metab. 57. 1125-l 128. Van Landeghem, A.A.J., Poortman, J., Nabuurs, M. and Thijssen, J.H.H. (1985) Cancer Res. 45,2900-2904. Vermeulen, A., Deslypere, J.P., Paridaens, R., Leclercq, C., Roy, F. and Henson, J.C. (1986) Eur. J. Cancer Clin. Oncol. 22.515-525.
ELSEV’IIZR
Molecular and Cellular Endocrinology 106 (1994) 17-21
Stimulation of aromatase activity in breast fibroblasts by tumor necrosis factor a Fiona Macdiarmida, D. Wanga, Loma J. Duncan~, A. Purohita, Margaret W. Ghilchikb, M.J. Reeda nUnit ofMetabolic Medicine, St. Mary’s Hospital Medical SchooI, Imperial College of Science, Technology and Medicine, London W IPG, UK ‘The Breast Clinic, St. Mary’s Hospitul, London W2 INY, UK Received 10 June 1994;accepted31 August 1994
AhStlWt
The conversion of ~dmsten~ione to es&one, the reaction mediated by the aromatase enzyme complex, may make an important contribution to the syntbesis of estrogens in breast tissues.In the present study, the effect of the cytokine, TNFa, on ammatase activity was examined in breast tibrublasts derived from normal and malignantbreast tissue. TNFa (2.5-10.0 rig/ml), in the presence of stripped fetal calf serum and dexamethasone, significantly stimulated Iibroblast aromatase activity in a dose-dependent manner. IL-1 and IL-6 also stimulated flbroblast aromatase activity, but no marked synergism between TNFa and IL-1 or IL-6 was detected. Using a specific radioimmunoassay, significant concentrations of TNFa were detected in samples of breast cyst fluid and breast tumor cytosol, which had previously been shown to stimulate aromatase activity, but not in conditioned medium from breast tumor-derived flbroblasts. As TNFcz may be preferentially expressed and produced in the adipose tissue component of the breast, this cytokine may have au important role in regulating estrogen synthesis in normal and malignant breast tissues. Key~onis: Breast cancer; Aromatase; Estrogens; Cytokines; Tumor necrosis factor a
1. Introduction Synthesis of estrogen from androstenedione within the breast may make an important contribution to the high estrogen concentrations which are present in normal and malignant breast tissue (Bonney et al., 1983; Van Landeghem et al., 1985; Vermeulen et al., 1986; James et al., 1987). The conversion of androstenedione to estrone is mediated by the aroma&se enzyme complex and aromatase activity is detectable in 60% of breast tumors (Tilson-Mallett et al., 1983). Whether the ~orna~e complex in breast tumors can produce sufficient estrogen to exert a mitogenic effect on breast cells is controversial (Bradlow, 1982). Using a double isotopic infusion technique, local synthesis of estrone from androstenedione was found to make a major contribution to tumor estrone content in some, but not all, tumors (Reed et al., 1989). The finding of an association between aromatase activity and the presence of a tumor in breast quadrants has suggested the possibility that increased aromatase activity in a breast quadrant may favor tumor *&tins
author.Tel. 071 725 1738. Fax 071 725 1790.
0303-7207/94/$07.00 0 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 0303-7207(94)03388-A
development (Miller and G’Neill, 1987). Alternatively, it is possible that breast tumors produce factors which are able to stimulate aromatase activity in adipose tissue adjacent to the tumor. Breast cyst fluid (BCF), breast tumor cytosol (BTC) and conditioned medium (CM) from breast tumor fibroblasts have all been shown to stimulate aromatase activity in breast tissue fibroblasts (Reed et al., 1993). Several cytokines and growth factors, including interleukin-1 (IL-l), IL-6 and insulin-like growth factor-type I (IGF-I) are detectable in these prep~ations and can stimulate aroma&se activity (Reed et al., 1992, 1993). Most of the cytokines or growth factors which have so far been found to stimulate aromatase activity in breast tumor derived fibroblasts are thought to originate from the epithelial or stromal components of the breast. The major part of the breast, however, is composed of adipose tissue and there is evidence that this tissue is important for tumor growth. In mice, the ability of mammary tumors to develop is inhibited in the absence of adipose tissue (Hoshino, 1962). There is now evidence indicating that adipose tissue may be a major site of tumor necrosis factor a (TNFa) expression and production
F. ~acdia~mid
18
et al. I ~a~ecula~ and Cefluiar E~ocrin~Io~
(Hotamisligil et al., 1993; Spiegelman et al., 1993). As cytokines appear to have an important role in regulating aroma&se activity, we have examined the ability of TNFa to regulate aroma&se activity in fibroblasts derived from normal or malignant breast tissues. The possibility that TNFct might also interact with other cytokines (IL- 1, B-6) was also investigated. In addition, concentrations of TNFa were measured in samples of BCF, BTC and CM, which have previously been shown to stimulate aromatase activity, to examine whether the presence of TNFa in these preparations could account for their aromatase stimulatory activity. 2. Materials and methods 2.1. Cultw-e of~broblasts from normal and rnal~g~~t breast tissues Two samples of normal breast adipose tissue were obtained from pre-menopausal women (subjects 1 and 2) undergoing reduction mammoplasty and one tumor sample from a post-menopausal woman {subject 3). Fibroblasts were prepared from these breast tissues as previously described (Adams et al., 1988). Cells were grown until confluent in Earle’s minimum essential medium (EMEM) containing 10% fetal calf serum (FCS). When confluent, cells were washed with Earle’s balanced salt solution (5 ml) and cultured for 24 h in phenol-red free, EMEM containing 10% stripped fetal calf serum (SFCS). Cytokines (TNFa, IL-l, IL-6, Bachem Inc., UK) were added in this medium in the presence of dexamethasone (100 nmol/l, Sigma, UK). To initially demonstrate that the fibroblasts possessed aromatase activity, cells were treated with dexamethasone (100 nmoU1) and cultured in 10% stripped fetal
c
I_
0
Fig.
I.
1
I.1
1.
-
2 PRFM
3
Ic 1.2’ Dw
3
f
lWnMf+SFCS
Aromatase activity in fibroblasts derived from normal
(I
and 2) or
malignimt breast tissue (3). Activity in fibroblusts cultured in phenol-red free medium containing
10% SFCS (PRFM)
was very low, but detect-
able. Culture of fibroblasts in stripped fetal calf serum (WCS) presence
of
dexomethasone
(Dex.,
IOOnM)
resulted
in the
in significant
(P < 0.001) stimulation of aromutase activity (mean f SEM, n = 3).
106 fl994) 17-21
calf serum, a combination known to stimulate aromatase activity (Simpson et al., 1981). 2.2. Aro~tase act~vi~ Aromatase activity was assayed in intact cell monolayers by measuring the release of 3Hz0 from [1@-3H]androstenedione [Du Pont (UK) Ltd.] over a 24-h period, as described previously (Reed et al., 1992, 1993) with the exception that fibroblasts were incubated with substrate for 24 h. Aroma~e activity was linear over this period and there was a significant correlation (r = 0.9, P < 0.01) between aromatase activity and incubation time. Results are expressed as femtomoles of product produced per million cells per 24 h, mean + SEM. 2.3. R~ioimm~~ass~y of TiVFa Concentrations of TNFa in BCF, BTC and CM were measured by a specific radioimmunoassay (Amersham Int., UK). Cross-reaction of other cytokines in this assay was
F. Macdiannid et al. I Molecular and Cellular Endocrinology 106 (1994) 17-21
T 1 L
C
2.5
5.0
10.0
C
2.5 5.0
1
10.0
i C
2.5
2
5.0
10.0
3
TNF(ng/ml)
Fig. 2. Effect of TNFa on aromatase activity in tibroblasts derived from normal (1 and 2) or malignant breast tissue (3). Treatments were added in phenol-ted free medium containing 10% SFCS and cells were cultured in this medium + TNFa in the presence of dexamethasone (100 nmolll). Addition of TNFa resulted in a significant stimulation of aromatase activity (P < 0.001) at all doses tested. Significant correlations were found between the concentration of TNFa added to cells and aromatase activity (mean f SEM, n = 3).
activity was examined (Fig. 2). TNFa stimulated fibroblast aromatase activity in a dose-dependent manner over the range 2.5-10.0 ng/ml. Analysis of the relationship between aromatase activity and the dose of TNFa tested using Spearman’s rank correlation revealed significant (P <
19
0.001) correlations: 1, p = 0.79; 2, p = 0.93; 3, P = 0.81. The effect of the lowest concentrations of TNFa tested appeared to be greater in the fibroblasts derived from malignant breast tissue, but further studies will be required to confirm the significance of this observation. Having shown that TNFa stimulated aromatase activity, its effectiveness was compared with other cytokines (IL-l and IL-6) which are known to stimulate aromatase activity in the presence of dexamethasone. As found previously, both IL-l and IL-6 stimulated aromatase activity in tumorderived fibroblasts with IL-6 having a greater stimulatory effect than IL-l (Fig. 3). In fibroblasts derived from malignant breast tissue, IL-6 (50 ng/ml) stimulated aromatase activity to a similar extent to TNFa (10 ng/ml). Higher concentrations of TNFa were not tested because of the potential anti-proliferative effect of this cytokine. In fibroblasts derived from adipose tissue from premenopausal women, aromatase activity was stimulated by IL-l and IL-6 (Fig. 3). However, the effect of IL-l and TNFa on aromatase activity was greater in fibroblasts derived from subject 2, with IL-6 having only a small effect on aromatase activity in these fibroblasts. For fibroblast derived from adipose tissue, maximal stimulation of aromatase activity occurred after treatment of fibroblasts from subject 2 with IL-l. The possibility that some interaction between the different cytokines might occur in their ability to stimulate fibroblast aromatase activity was also examined. For this, the effect of treating cells with IL-l (50 ng/ml) plus IL-6 (50 ng/ml), IL-l (50 ng/ml) plus TNFa (10 ng/ml) or IL-6
1
=Ooo I T
T
h P1 3ooo Q
+l
sf
!
E
11
I
2000
122
gK E
nglml
IL-1
IL-6
TNF
L-l
IL-6
TNF
60
50
10
50
50
10
2
_r IL-1
IL-6
50
50
1000
-
0
TNF 10
JY 55
IL-1
IL-1
1:6
T+NF T+NF 1
IL.6
IL-1
IL-1
IL-6
IL:
T;F
T+NF
2
IL-1 IL:
IL-1
IL-6
T+NF
T+NF
3
3
Fig. 3. Comparison of the effects of IL-l, IL-6 or TNFa on aromatase activity in fibroblasts derived from normal (I and 2) or malignant breast tissue (3) (mean 2 SEM, n = 3). Treatments were added in phenol-red free medium containing 10% SFCS + cytokines and were cultured in the presence of dexamethasone (100 nmolll).
Fig. 4. The effect of IL-l plus IL-6 or TNFa plus IL-1 or IL-6 on aromatase activity in fibroblasts derived from normal (I and 2) or malignant breast tissue (3). Additive values, derived from the effect of each cytokine alone on aromatase activity, are indicated at the top of each column (mean & SEM, n = 3). Treatments were added as indicated in the legends to the other figures.
20
F. Mucdiarmid et al. f Moleculur and Ceflulur Endocrinology 106 (1994) I741
(50 ng/ml) plus TNFG (10 nglml) was compared with the expected additive effect of treating fibroblasts with each of these cytokines (Fig. 4). No evidence of any synergistic interaction of the cytokines in their ability to stimulate aromatase activity was detected. In two of the three fibroblast cell lines, IL-l plus TNFa, when added together, produced an increase in aromatase activity close to that expected due to an additive effect (99% and 116%, values shown at top of the respective columns for fibroblasts from subjects 2 and 3 in Fig. 4). The addition of IL-I and IL-6 tended to reduce the effect expected from each cytokine (49-82%). In fibroblasts derived from adipose tissue, IL-6 plus TNFa increased the stimulatory effect when each cytokine was tested alone (122% and 140%), whereas for the fibroblasts derived from malignant tissue, their combined effect was reduced (55%). 3.1. Effect of cytokines on cell growth Fibroblasts only grow at a relatively slow rate and the present investigation was not primarily carried out to establish the effect of IL-l, IL-6 and TNFa on cell growth. Fibroblasts were only treated with cytokines for 48 h and during this time no significant effect on cell growth was detected.
of T’Fa in breast cyst fluid, breast tumor cytosol and conditioned medium from tumor derived fibroblasts 3.2. Concentrations
Significant concentrations of TNFa were detectable in samples of BCF previously shown to stimulate aromatase activity in cultured breast fibroblasts (820-l 120 pg/ml, mean + SD = 957 + 199 pg/ml, n = 7). No difference was detected between TNFa concentrations in BCF with high or low electrolyte ratios. TNFa was also measured in a BTC and CM previously shown to stimulate aromatase activity. TNFa was present in BTC (630 pg/ml) whereas TNFa was not detectable in CM from tumor-derived fibroblasts. 4. Discussion Results obtained from this study have shown that TNFa in the presence of SFCS and dexamethasone is a potent stimulator of aromatase activity in fibroblasts derived from both normal and malignant breast tissue and that it stimulates activity in a dose-dependent manner. As previously reported for IL-1 and IL-6, TNFa was only able to stimulate aromatase activity in the presence of dexamethasone. However, in the absence of serum and dexamethasone, TNF has previously been reported to inhibit aromatase expression in cyclic AMP stimulated adipose stromal cells (Simpson et al., 1989). It is not yet clear whether dexamethasone, as used in the present study, has a permissive role in allowing aromatase activity to be induced or whether it also influences cytokine receptor concentrations in fibroblasts. The ability of IL-l and IL-6 to stimulate
aromatase activity in fibrobI~ts derived from malignant breast tissue was confirmed in the present study. IL-6, however, appears to have a somewhat smaller stimulatory effect on aromatase activity in fibroblasts derived from breast adipose tissue. IL-6 is secreted by breast tumor fibroblasts (Adams et al., 1991; Reed et al., 1993), and it is likely that this cytokine may act in an autocrine manner in regulating aromatase activity in stromal cells within breast tumors. IL-l is also secreted by breast tumor fibroblasts although at a tower level than IL-6 (Duncan and Reed, 1994). Stimulation of aromatase activity by TNFa was similar to that of IL-6 in the fibroblasts derived from the malignant breast tissue but lower (markedly for subject 2) for fibroblasts derived from normal breast tissue. As cytokines can interact to stimulate the production of other cytokines or cytokine receptors, the effect of combining the different cytokines tested was also examined. No evidence was obtained from the present investigation that IL-l, IL-6 or TNFa interact to produce a synergistic increase in aromatase activity. Whereas IL-l plus TNFa generally acted to produce an additive effect, suggesting independent pathways of action, the addition of IL-l and IL-6 together resulted in a lower than additive response, suggesting common pathway of action in these fibroblasts. Some difference in the effect of IL-6 plus TNFa was noted between normal and tumor derived fibroblasts, but further studies are required to confirm these observations. Significant concentrations of TNFa were detected in samples of BCF and BTC, but not CM, previously found to stimulate aromatase activity (Reed et al., 1993). Although only a small number of BCF samples have been examined so far, no difference in TNFa concentrations was detected for BCF with high or low Na+/K+ ratios. This is in contrast to IL-6 concentrations in BCF which are significantly higher in BCF with high electrolyte ratios, but similar to IL-1 where no difference between the two cyst types was detected (Reed et al., 1992). Concentrations of TNFa in BCF and BTC were approximately 1 ngfml. While this concentration is somewhat lower than that used to stimulate fibroblast aromatase activity, albumin or albuminoid-like proteins, which are present in BCF and BTC, have been shown to potentiate the effect of cytokines and growth factors, such as IGF-I, in their ability to stimulate aromatase activity (Singh et al., 1992; Reed et al., 1993). TNFa was not detected in a sample of CM which had previously been shown to stimulate fibroblast aromatase activity, although high concen~ations of IL-6 are present in this CM (Reed et al., 1993). It is possible therefore that breast stromal cells are not a major site of TNFa production, and there is now evidence that adipose tissue is a major site of TNFa expression and production (Spiegelman et al., 1993; Hotamisligil et al., 1993). Previous investigations have revealed an association between the presence of a breast tumor and aromatase activity in breast quadrants in which the tumor is located (Miller and O’Neill, 1987). The finding that TNFa is able to stimulate aromatase activity
F. Macdiarmid et al. I Molecular and Cellular Endocrinology 106 (1994) 17-21
and evidence that adipose tissue is able to express and produce TNFa suggests the possibility that increased production of TNFa in a particular breast quadrant could lead to an increase in aromatase activity and estrogen synthesis sufficient to support tumor development. In summary, TNFa has been found to be a potent factor involved in the regulation of aromatase activity in fibroblasts derived from both normal and malignant breast tissues. Significant concentrations of TNFa were detected in BCF and BTC, previously shown to stimulate aromatase activity. As TNFa may be preferentially expressed and produced by the adipose tissue component of the breast, this cytokine may have an important role in regulating estrogen formation and the production of tumors within the breast. References Adams, E.F., Newton, C.J., Tait, G.H., Braunsberg, H., Reed, M.J. and James, V.H.T. (1988) Int. J. Cancer 42, 119-122. Bonney, R.C., Reed, M.J., Davidson, K., Beranek, P. and James, V.H.T. (1983) Clin. Endoctinol. 19.727-739. Bradlow, H.L. (1982) Cancer Res. 42,3382s-3386s. Duncan, L.J. and Reed, M.J. (1994) J. Steroid Biochem. Mol. Biol. 49, 63-68.
21
Hoshino, K. (1962) J. Natl. Cancer Inst. 29.835-843. Hotamisligil, G.S., Shargill, N.S. and Spiegelman, B.M. (1993) Science 259,87-91 James, V.H.T., McNeill, J.M., Lai, L.C., Newton, C.J., Ghilchik, M.W. and Reed, M.J. (1987) Steroids 50.269-279. Miller, W.R. and O’Neill, J.S. (1987) Steroids 50,537-547. Reed, M.J., Coldham, N.G., Pate& S.R., Ghilchik, M.W. and James, V.H.T. (1992) J. Endocrinol. 132, R5-R8. Reed, M.J., Owen, A.M., Lai, L.C., Coldham, N.G., Shaikh, N.A. and James, V.H.T. (1989) Int. J. Cancer 44.233-237. Reed, M.J., Topping, L., Coldham, N.G., Purohit, A., Ghilchik, M.W. and James, V.H.T. (1993) J. Steroid Biochem. Mol. Biol. 44, 589596. Simpson, E.R., Ackerman, G.E., Smith, M.E. and Mendelson, C.R. (1981) Proc. Natl. Acad. Sci. USA 78,5690-5694. Simpson, E.R., Merrill, J.C., Hollub, A.J., Graham-Lorence, S. and Mendelson, C.R. (1989) Endocr. Rev. 10, 136-148. Singh, A., Blench, I., Morris, H.R., Savoy, L.-A. and Reed, M.J. (1992) Mol. Cell. Endocrinol. 85, 165-173. Spiegelman, B.M., Choy, L., Hotamisligil, G.S., Graves, R.A. and Tontonoz, P. (1993) J. Biol. Chem. 268.68236826. Tilson-Mallett, N., Santner, S.J., Feil, P.D. and Santen, R.J. (1983) J. Clin. Endocrinol. Metab. 57. 1125-l 128. Van Landeghem, A.A.J., Poortman, J., Nabuurs, M. and Thijssen, J.H.H. (1985) Cancer Res. 45,2900-2904. Vermeulen, A., Deslypere, J.P., Paridaens, R., Leclercq, C., Roy, F. and Henson, J.C. (1986) Eur. J. Cancer Clin. Oncol. 22.515-525.