Local uptake and synthesis of oestrone in normal and malignant postmenopausal breast tissues

Journal of Steroid Biochemistry & Molecular Biology 81 (2002) 57–64

Local uptake and synthesis of oestrone in normal and malignant postmenopausal breast tissues A.A. Larionov a,b , L.M. Berstein b , W.R. Miller a,∗ a

Breast Unit Research Group, Paderewski Building, WGH, Edinburgh University, Edinburgh EH4 2XU, UK b N.N. Petrov Research Institute of Oncology, St. Petersburg, Russia Received 11 April 2001; accepted 25 January 2002

Abstract Uptake and local formation of oestrone (E1) were studied in vivo by a double isotopic technique in normal and malignant breast tissues from 24 postmenopausal women with breast cancer. Active uptake of radio-labelled E1 beyond plasma was found both in normal and malignant tissue, the effect being significantly greater in non-malignant compared with cancer tissue. The presence of local E1 formation was also demonstrated in most samples. Both uptake and synthesis positively correlated with total amount of radioactive E1 found in the tissues. Uptake appeared to make a greater contribution to E1 levels within the breast than in situ synthesis, although there were marked variations between specimens from different patients and the relative proportion of synthesis to uptake was higher in tumour compared with non-malignant tissue. These results demonstrate quantitative differences in the different compartments by which postmenopausal breasts obtain oestrogen and highlight variations between individual breasts. This may be important in optimising oestrogen deprivation therapy for postmenopausal patients with hormone-dependent cancers. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Oestrogen; Postmenopausal breast; Double isotopic technique

1. Introduction Oestrogens influence the development of both the normal breast and breast cancers [12]. However, in postmenopausal women in whom most breast cancers are diagnosed, local levels of oestrogens in the breast do not simply reflect those in the blood. Thus, after the menopause, when oestrogens in circulation fall, levels in the breast do not change significantly and, as a consequence, their concentrations in the breast may be more than those in the blood [21,22]. This can only occur as a result of an active process: selective sequestration of oestrogens from the blood, or local formation by breast tissue itself. Both these mechanisms have been shown to occur. Thus, following administration of radio-labelled oestrogens, radioactivity per gram of breast tissue is often higher than per millilitre of plasma, confirming the accumulation against the concentration gradient [9,13]. Studies also confirm the ability of the breast to transform androgens into oestrogens (by the aromatase enzyme) and to inter-convert different oestrogen [15,20,22]. However, most studies of local oestrogen production and metabolism in the breast have been performed in vitro and there is a need for in vivo studies ∗

Corresponding author. Tel.: +44-131-5372501; fax: +44-131-5372449. E-mail address: [email protected] (W.R. Miller).

to clarify the relative input of active uptake and local synthesis to the oestrogenic environment in the postmenopausal breast. The present study was undertaken to investigate further in vivo oestrogen accumulation in human breast using double isotope infusion technique by means of which local E1 formation and uptake in malignant and non-malignant postmenopausal breast tissues could be compared.

2. Materials and methods 2.1. Patients Postmenopausal breast cancer patients (aged from 59 to 79) were included in the study and provided informed consent to participate. All presented to the Edinburgh Breast Unit and had a histologically confirmed diagnosis of breast cancer which was oestrogen receptor-rich (more than 20 fmol/mg cytosol protein, as measured in the biopsied tumour used for extraction of radio-labelled steroids). None had received prior endocrine treatment. Tumour, node involvement, distant metastases (as classified by the American Joint Committee on Cancer) (TNM) stage was T2 (n = 21)–T3 (n = 3), N0 (n = 20)–N1 (n = 4), M0 (n = 24) and histological grade was 2 in 18 cases

0960-0760/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 0 - 0 7 6 0 ( 0 2 ) 0 0 0 4 7 - X

58

A.A. Larionov et al. / Journal of Steroid Biochemistry & Molecular Biology 81 (2002) 57–64

and 3 in the remainder. Matched samples of tumours and macroscopically non-malignant surrounding breast tissues (histologically mainly adipose tissue) were studied from the same patients. 2.2. Experimental procedures The double isotopic infusion method is based on simultaneous infusion of oestrogen and oestrogen precursor labelled with different isotopes: 14 C-labelled oestrone (14 C-E1) and 3 H-labelled androstendione (3 H-Ad). Infusion was as described previously [14]. In brief, 1,2,6,7-3 H-Ad (20 MBq, 85 Ci/mmol, Amersham) and 4-14 C-E1 (1 MBq, 56 mCi/mmol, Amersham) were given as an initial bolus of 20% of the total volume followed by continuous 18 h infusion of the remainder. Samples of tissues were taken immediately after the infusion and stored in liquid nitrogen until steroid extraction. Frozen tissue was pulverized and the resultant powder (∼500 mg) was suspended in water. Radio-inert E1 (500 ␮g) was added to measure procedural losses (similarly 500 ␮g of non-labelled E1 was added to plasma samples). Tissue powder suspensions and plasma samples were extracted with ethanol:acetone (1:1). The extracts were then evaporated, reconstituted in methanol (70%) and left at −20 ◦ C overnight. After centrifugation, the supernatant was dried and the residue extracted with diethyl ether and was subjected to column chromatography on lipidex 5000 [18]. The fraction corresponding to E1 was further purified by thin-layer chromatography on silica gel using cyclohexane:ethyl acetate (55:45) as solvent. Recovery of E1 was monitored by absorption at OD282 —mean values were 76% for cancer tissue (range 39–91%), 77% for normal breast tissue (72–81%) and 69% for plasma (51–81%). Radioactivity (3 H and 14 C) in E1 fractions was then counted (five times for 50 min) in NE260 liquid scintillant (Nuclear Enterprises) on a 1900 CA tricarb counter (Packard). The variation (2σ ) was not more than 5%. The potential distribution of the radioactive labels in different steroid fractions is presented in Fig. 1. Comparison of 14 C levels in E1-fractions obtained from breast and from plasma provides information about uptake of E1 from plasma and together with data on 3 H-E1 in plasma allows the calculation of expected uptake of 3 H-E1 to tissues: 3

H-E1 expected from uptake 3 H-E1 in plasma × 14 C-E1 in tissue = 14 C-E1 in plasma

The excess of observed 3 H-E1 over the expected from the uptake reflects the input of the local formation. The local oestrogen production in the tissue may be calculated as follows: 3

H-E1 expected from local synthesis = 3 H-E1 in tissue − 3 H-E1 expected from uptake

Fig. 1. Potential distribution of radioactively-labelled steroids following infusion.

Results were analysed using Wilcoxon matched pairs nonparametric test and non-parametric correlations.

3. Results Values for radioactivity in E1 fractions from plasma, normal breast and breast cancer are shown in Table 1. Levels of 14 C were significantly higher in both types of breast tissue compared with plasma (P < 0.001). Furthermore, levels of 14 C were also significantly higher in non-malignant than malignant tissue (P < 0.01). Differences were also apparent in levels of 3 H-E1 between plasma, non-malignant tissue and tumour, values being significantly higher in tissue compared with plasma and non-malignant breast compared with tumour. However, there were marked variations between individual cases. In order to demonstrate uptake of radioactivity labelled E1 into tissue in individual cases, data have been plotted as a ratio of radioactivity in tissue (per gram) to that in plasma (per millilitre) (Fig. 2). In terms of normal tissue, the tissue:plasma ratio was >1 in all specimens with a median value of about 6. These observations held for both 14 C Table 1 Levels of

14 C-

and 3 H-E1 in blood and breast tissues

Plasma Non-malignant tissue Malignant

14 C-E1

3 H-E1

33 ± 12 177 ± 70 91 ± 45

19 ± 8 117 ± 81 73 ± 45

Values present dpm per gram of tissue or per millilitre of plasma (Mean ± S.D., n = 24). For both isotopes, levels in tissues were significantly higher than in plasma (P < 0.001). Similarly levels were significantly higher in normal breast compared with malignant (P < 0.01).

A.A. Larionov et al. / Journal of Steroid Biochemistry & Molecular Biology 81 (2002) 57–64

Fig. 2. Tissue/plasma ratios of radioactivity in oestrone fractions from either non-malignant or malignant breast tissue. Horizontal lines represent median values, dashed line represents ratio of unity.

59

(from administered E1) and for 3 H (E1 derived from androstenedione). With regard to malignant tissue, the majority of cases (22 of 24 for 14 C and 23 of 24 for 3 H) had ratio >1 with median excess value of about 2 for 14 C and about 4 for 3 H. However, a large range of the values between cases was evident for each of the ratios. In order to determine whether there was a relationship in oestrogen uptake between malignant and non-malignant breasts, values for uptake in paired cases are plotted in Fig. 3A. Values were greater in non-malignant tissue with the exception of a single case with a lower value and two with equal values. This difference was significant by Wilcoxon matched paired test (P < 0.001). Conversely, there was a significant association between uptake in normal and malignant tissue (P < 0.05). Furthermore, in both tumour and non-malignant tissues there was quantitative relation between uptake (as determined by tissue:plasma ratio) and total level of E1 (as determined by 3 H in E1 fraction, Fig. 4A and B). Local synthesis of oestrogen may be estimated from the 3 H levels in the purified E1 fraction. However, since the 3 H in E1 extracted from breast tissues may be derived (a) from local synthesis within the breast and (b) from the conversion of 3 H-Ad4 in extra mammary tissues, release into the bloodstream and uptake by the breast, it is necessary to subtract the latter (which can be assessed from the uptake of 14 C-oestrogen) from the total 3 H counts observed in the tissue. These data are presented for each individual patient in Fig. 5A and B and illustrate evidence for local biosynthesis in most tumours and non-malignant breast. The relationship between in situ synthesis from either tumour and non-malignant breast is shown in Fig. 3B. In general, synthesis tended to be higher in tumour than non-malignant breast. In breasts where both tissues displayed activity there was a significant quantitative

Fig. 3. Relationships between non-malignant and malignant tissue for (A) uptake of oestrone from the circulation and (B) local oestrogen formation (for tumours which display in situ synthesis). Dashed line is that on which values would be identical in both non-malignant and malignant tissue. There was a statistically significant correlation between the different types of tissue (P < 0.05 for both uptake and local formation).

60

A.A. Larionov et al. / Journal of Steroid Biochemistry & Molecular Biology 81 (2002) 57–64

Fig. 4. Relationships between oestrogen uptake and total radioactivity in (A) tumour and (B) non-malignant breast and between local oestrogen synthesis and total radioactivity in (C) tumours with evidence of biosynthesis and (D) non-malignant breast with evidence of biosynthesis.

association between the two types of tissue. Furthermore, in tissue displaying local biosynthesis, this was quantitatively related to total level of 3 H-E1 in both tumours and non-malignant breast (Fig. 4C and D). The relative contributions of uptake and synthesis in the tissues are shown in Fig. 6. There were marked variations in relative contribution of uptake and synthesis in both non-malignant and tumour tissues. However, there was a significant trend for the synthesis to make a larger contribution in tumour as compared with non-malignant breast (P = 0.05). 4. Discussion The endocrinology of the postmenopausal breast is unusual in that the profile and level of steroid hormones, in particular oestrogens, differ from those in the circulation. This suggests that local factors within the breast are important determinants of the oestrogenic environment. In order to study these, perfusion studies have been undertaken in which postmenopausal women with breast cancer have been given radioactively labelled androgens (to study conversion

to oestrogens) and oestrogens (to monitor uptake by the breast) in the period immediately before breast surgery. By extracting and purifying oestrogen fractions from plasma, non-malignant breast and tumours, the following were observed: (i) both normal and malignant breast tissue from postmenopausal women invariably accumulate oestrogens beyond levels in the circulation; (ii) there was in vivo evidence that the majority of non-malignant tissues and tumours synthesized oestrogens locally and such synthesis was related to the total level of radio-oestrogen within the tissues; (iii) the relative contribution of local uptake and synthesis varied significantly between individuals but in vivo uptake tended to be predominantly responsible for local levels of E1, particularly in non-malignant breast. The finding that postmenopausal breast tissue is able to uptake oestrogens from circulation against the concentration’s gradient has previously been reported by ourselves and others [12,21,22]. However, to our knowledge, the additional findings in the present study that (i) within individual cases, non-malignant breast invariably take up more E1 than malignant tissue and (ii) there is a significant association between uptake in non-malignant and tumour tissue, have not been previously reported.

A.A. Larionov et al. / Journal of Steroid Biochemistry & Molecular Biology 81 (2002) 57–64

61

Fig. 5. The relative contribution of oestrogen uptake and synthesis to the total radioactivity in individual cases of (A) tumour and (B) non-malignant breast: + and − represent status of in situ oestrogen synthesis in each case.

The underlying mechanism(s) by which the postmenopausal breast sequesters oestrogen remains to be defined. Although, oestrogen receptors have been implicated, tumours with high levels of oestrogen receptors tend to have elevated levels of oestradiol [13,22], it is clear that oestrogen receptor negative tumours also accumulate oestrogen [5]. Other agents that might be involved are proteins such as SHBG [10]. In this respect, Masamura et al. [8] have

suggested that a protein with a Kd of 1 × 10−10 M may be implicated. Finally, it is possible that the lipid component of the breast may act as a sponge for steroids [4]. This last mechanism would be compatible with the finding in the present study that non-malignant breast (which has a large fat component) takes-up more E1 than tumours, which are likely to be predominantly epithelial and stromal. However, if this is the case, it should be noted that bio-availability

62

A.A. Larionov et al. / Journal of Steroid Biochemistry & Molecular Biology 81 (2002) 57–64

Fig. 6. The percentage contribution of local synthesis to total 3 H-oestrone in non-malignant tissue and tumour. Horizontal lines represent median values. Differences between the two groups was statistically significant (P < 0.05).

of oestrogens sequestrated into a lipid droplet may differ from one sequestrated into an epithelial or stromal cell compartment [4]. The positive quantitative correlation between uptake in normal and malignant tissue is intriguing and not easily explained. It suggests a field effect in which different compartments of the same breast, despite their

differing composition, have inherently similar capacities to accumulate oestrogen. Evidence for local biosynthesis of oestrogen was found in 62.5% (15 from 24) of non-malignant tissues and 79% (19 from 24) of tumours (see Fig. 5). This incidence is similar to that which has been reported for in vitro incubation of breast tissue [1,2,11]. The data are also similar to those observed in situ by Reed et al. using identical methodology on a much smaller cohort of tissues [19] and results on breast cancers published by ourselves [14,17]. Quantitatively, there was a large variation in local synthesis between different breasts and this related to total amount of tritiated E1 extracted from tissues. This would suggest that local synthesis is an important factor in determining the local environment of oestrogen within the breast in a significant proportion of breast cancer patients. It is therefore important to determine the factors which influence aromatase activity in the breast. Culture and in vitro studies suggest that growth factors, cytokines and prostaglandin have the ability to regulate oestrogen biosynthesis in peripheral tissues [16]. In this respect, it may be relevant that although level of in situ synthesis tended to be higher in tumours than in non-malignant breast (as has been observed in vitro also [7,13,22]) there was a significant association between the two tissue types. This suggests that similar factors may affect the general level of oestrogen synthesis in both tissues and that variation between patients may be associated with the local production of such factors. The increased in situ synthesis of oestrogen in breast tumours may be accounted for by multiple factors including differences in cellular composition, local regulation and availability of substrate. In a minority of tissues (six non-malignant and four tumours), because the level of observed 3 H in E1 fraction was lower than that which was expected from oestrogen uptake, a “negative” value for oestrogen synthesis was found. The same phenomenon was reported by Reed et al. [19], who sought to explain the results on the basis of experimental error. These may be procedural ones but there could also be more systematic reasons, such as the E1 fractions not being radiochemically pure, inaccurate liquid scintillation counting, further and differential metabolism of E1, compartmentalisation of locally synthesized and sequestered oestrogen (which may be differentially extracted). The calculation of in situ synthesis is derived from two components: (i) the 3 H level in tissue E1 and (ii) a correction for peripheral synthesis and uptake into tissue (which in turn is based on plasma 3 H-E1 and the ratio of 14 C-E1 in tissue and plasma). It is worth considering how these components might contribute to a negative result. For negative values to result from inaccuracies on (i), there would have to be a falsely low 3 H level in tumour E1. This would exclude the possibility of impure E1 fractions which are likely to produce falsely high 3 H levels. It is possible that imprecise dual isotope counting could allocate certain 3 H counts to 14 C but counting efficiency was rigorously monitored with long and multiple count times and calibration with known amounts of 3 H

A.A. Larionov et al. / Journal of Steroid Biochemistry & Molecular Biology 81 (2002) 57–64

and 14 C being assessed alone and in combination. Furthermore, a similar proportion of low negative in situ synthesis cases are apparent after infusion of 14 C androstenedione and 3 H E1 (unpublished results). The most likely reason for “artefactually” low 3 H-E1 is that the synthesized 3 H-E1 is further metabolized and sequestered in cellular compartments which render the extraction incomplete. Negative values could also result from false high component (ii), but given that 14 C-E1 is one of the infused steroids it is unlikely that purified 14 C-E1 fractions will be impure and there is no reason to believe 3 H plasma E1 should be contaminated. Random purified E1 fractions from both tumour and plasmas were subject to chemical derivative formation characterization and were shown to have consistent 3 H and 14 C specific radioactivities and 3 H/14 C ratios. It should be noted that, if further metabolism of sequestration of E1 is responsible for the apparent negative in situ synthesis, differential effects on locally-synthesized 3 H-E1 would need to occur. It is worth emphasising that, although all the results in the present study are based on a single measurement, in a control experiment on a limited number of large tumours replicate extractions yielded only consistent positive or negative values suggesting negative results are not simply inaccurate measurements or a summation of random errors. Similarly, our published data from studies in which the same tumour has been studied before and after systemic endocrine therapy, indicates that tumours which at the outset of treatment have no evidence for in situ synthesis never become positive with therapy [17]. For these reasons, when assessing the proportion of tissues with in situ aromatase, any positive value was regarded as evidence of activity (Fig. 5). Because of the importance of local oestrogens in the continued growth of certain hormone-dependent tumours [6,12] and the availability of drugs which block either oestrogen synthesis or mechanism of action [3], it was interesting to assess the relative contribution of uptake and synthesis to the oestrogen pool within tissues. These calculations suggest that the relative contribution varies enormously between different individuals both in terms of normal and malignant breast tissue. This meant that in certain tissues uptake alone could account for the local oestrogen levels whereas in other tissues local synthesis was more influential. In contrast to the study of Reed at al. [19], the present findings suggest that generally (but not individually) uptake outweighed synthesis. However, the relative proportion of synthesis to uptake tended to be higher in malignant as compared with non-malignant tissue. This probably reflects tissue composition in which the higher proportion of fat in non-malignant tissue might result in enhanced oestrogen uptake, whereas increased production of oestrogen in synthesizing cells and locally inducing factors may predominate in tumours. In conclusion, the present study illustrates that the postmenopausal breast and its tumours have an extraordinary ability to obtain oestrogens and may utilize active mechanisms such as uptake against a concentration gradient and/or local biosynthesis. Different breasts may use these mech-

63

anisms to varying extent but the factors influencing these processes are largely undefined. However, given the key role that oestrogens play in both the genesis and growth of hormone-dependent tumours, it is essential that further detailed research is conducted, with hope that agents are developed which block either oestrogen uptake or biosynthesis within the breast.

Acknowledgements The study was supported by grant of the Royal Society/NATO (99A).

References [1] L.M. Berstein, A.A. Larionov, A. Kyshtoobaeva, K.M. Pozharisski, V.F. Semiglazov, O.A. Ivanova, Aromatase in breast cancer tissue— localization and relationship with reproductive status of patients, J. Cancer Res. Clin. Oncol. 122 (8) (1996) 495–498. [2] P. Bolufer, E. Ricart, A. Lluch, C. Vazquez, A. Rodriguez, A. Ruiz, F. Llopis, J. Garcia-Conde, R. Romero, Aromatase activity and estradiol in human breast cancer: its relationship to estradiol and epidermal growth factor receptors and to tumor-node-metastasis staging, J. Clin. Oncol. 10 (3) (1992) 438–446. [3] A.M. Brodie, V.C. Njar, Aromatase inhibitors and their application in breast cancer treatment, Steroids 65 (4) (2000) 171–179. [4] W.H. Cleland, C.R. Mendelson, E.R. Simpson, Aromatase activity of membrane fractions of human adipose tissue stromal cells and adipocytes, Endocrinology 113 (6) (1983) 2155–2160. [5] J. Fishman, J.S. Nisselbaum, C.J. Menendez-Botet, M.K. Schwartz, Estrone and estradiol content in human breast tumors: relationship to estradiol receptors, J. Steroid Biochem. 8 (8) (1977) 893–896. [6] B.E. Henderson, H.S. Feigelson, Hormonal carcinogenesis, Carcinogenesis 21 (3) (2000) 427–433. [7] V.H. James, J.M. McNeill, L.C. Lai, C.J. Newton, M.W. Ghilchik, M.J. Reed, Aromatase activity in normal breast and breast tumor tissues: in vivo and in vitro studies, Steroids 50 (1–3) (1987) 269– 279. [8] S. Masamura, S.J. Santner, P. Gimotty, J. George, R.J. Santen, Mechanism for maintenance of high breast tumor estradiol concentrations in the absence of ovarian function: role of very high affinity tissue uptake, Breast Cancer Res. Treat. 42 (3) (1997) 215– 226. [9] J.M. McNeill, M.J. Reed, P.A. Beranek, R.C. Bonney, M.W. Ghilchik, D.J. Robinson, V.H. James, A comparison of the in vivo uptake and metabolism of 3 H-oestrone and 3 H-oestradiol by normal breast and breast tumour tissues in post-menopausal women, Int. J. Cancer 38 (2) (1986) 193–196. [10] S. Meyer, C. Brumm, H.E. Stegner, G.H. Sinnecker, Intracellular sex hormone-binding globulin (SHBG) in normal and neoplastic breast tissue—an additional marker for hormone dependency? Exp. Clin. Endocrinol. 102 (4) (1994) 334–340. [11] W.R. Miller, Aromatase activity in breast tissue, J. Steroid Biochem. Mol. Biol. 39 (5B) (1991) 783–790. [12] W.R. Miller, Oestrogen and breast cancer, R.G. Landes Co., Austin, 1996, 207 pp. [13] W.R. Miller, Uptake and synthesis of steroid hormones by the breast, Endocr. Relat. Cancer 4 (1997) 307–311. [14] W.R. Miller, Biology of aromatase inhibitors: pharmacology/ endocrinology within the breast, Endocr. Relat. Cancer 6 (2) (1999) 187–195.

64

A.A. Larionov et al. / Journal of Steroid Biochemistry & Molecular Biology 81 (2002) 57–64

[15] W.R. Miller, A.P. Forrest, Oestradiol synthesis by a human breast carcinoma, Lancet 2 (7885) (1974) 866–868. [16] W.R. Miller, P. Mullen, P. Sourdaine, C. Watson, J.M. Dixon, J. Telford, Regulation of aromatase activity within the breast, J. Steroid Biochem. Mol. Biol. 61 (3–6) (1997) 193–202. [17] W.R. Miller, J. Telford, C. Love, R.C.F. Leonard, S. Hillier, H. Gundacker, H. Smith, J.M. Dixon, Effects of letrozole as primary medical therapy on in situ oestrogen synthesis and endogenous oestrogen levels within the breast, Breast 7 (1988) 273– 276. [18] M.J. Reed, P.A. Beranek, M.W. Ghilchik, V.H. James, Oestrogen production and metabolism in normal postmenopausal women and postmenopausal women with breast or endometrial cancer, Eur. J. Cancer. Clin. Oncol. 22 (11) (1986) 1395–1400.

[19] M.J. Reed, A.M. Owen, L.C. Lai, N.G. Coldham, M.W. Ghilchik, N.A. Shaikh, V.H. James, In situ oestrone synthesis in normal breast and breast tumour tissues: effect of treatment with 4-hydroxyandrostenedione, Int. J. Cancer 44 (2) (1989) 233–237. [20] T. Utsumi, N. Yoshimura, M. Maruta, S. Takeuchi, J. Ando, K. Maeda, N. Harada, Significance of steroid sulfatase expression in human breast cancer, Breast Cancer 6 (4) (1999) 298–300. [21] A.A. van Landeghem, J. Poortman, M. Nabuurs, J.H. Thijssen, Endogenous concentration and subcellular distribution of oestrogens in normal and malignant human breast tissue, Cancer Res. 45 (6) (1985) 2900–2906. [22] A. Vermeulen, J.P. Deslypere, R. Paridaens, Steroid dynamics in the normal and carcinomatous mammary gland, J. Steroid Biochem. 25 (5B) (1986) 799–802.