Tamoxifen and fenretinide in women with metastatic breast cancer

Background: Tamoxifen and fenretinide combination therapy has been shown to be an active treatment regimen in metastatic breast cancer patients. This pilot study sought to determine whether the addition of fenretinide to tamoxifen would be associated with antitumor activity in metastatic breast cancer patients who had been previously treated with tamoxifen or who had hormone receptor negative disease. The effect of this therapy on circulating plasma transforming growth factorbeta (TGF) levels and serum lipids was also examined.Patientsand Methods: Thirtyone patients were treated with tamoxifen (20mg po daily), and fenretinide (400mg po daily with a 3day drug holiday each month). Plasma TGF testing was performed using isoform specific sandwich ELISA.Results: Twenty four of the 31 patients were evaluable for an antitumor response including 14 estrogen receptor (ER) positive patients who had failed prior tamoxifen therapy, seven ERnegative patients, and three hormone therapy naive ERpositive patients. There were no objective antitumor responses; three patients had stable disease for 8, 8, and 24 months. Five patients (16%) discontinued therapy for toxicity (one for grade 3 skin rash and four for abnormal dark adaptation). There was a statistically significant decrease in total cholesterol (median change per patient of –13.5mg/dl; p=0.049, a 6.5% decrease), and an increase in HDL levels (median change per patient of +18mg/dl, p=0.0001, a 35% increase) with tamoxifen and fenretinide therapy.TGF-1 plasma levels were normal in 26 of 28 patients, and no changes in these levels post-treatment were demonstrated.Conclusions: Tamoxifen and fenretinide therapy is not an active combination in ER negative metastatic breast cancer or in patients whose disease has progressed on tamoxifen. This combination had a beneficial effect on total serum cholesterol and HDL levels with no associated rise in serum triglyceride levels. The 400mg dose of fenretinide was associated with symptomatic nyctalopia in one-third of patients making it an unsuitable dose for use in breast cancer prevention studies.     

Coexpression of c-kit and stem cell factor in breast cancer results in enhanced sensitivity to members of the EGF family of growth factors

Insulinlike growth factor (IGF)I protects many cell types from apoptosis. As a result, it is possible that IGFIresponsive cancer cells may be resistant to apoptosisinducing chemotherapies. Therefore, we examined the effects of IGFI on paclitaxel and doxorubicininduced apoptosis in the IGFIresponsive breast cancer cell line MCF7. Both drugs caused DNA laddering in a dosedependent fashion, and IGFI reduced the formation of ladders. We next examined the effects of IGFI and estradiol on cell survival following drug treatment in monolayer culture. IGFI, but not estradiol, increased survival of MCF7 cells in the presence of either drug. Cell cycle progression and counting of trypanblue stained cells showed that IGFI was inducing proliferation in paclitaxeltreated but not doxorubicintreated cells. However, IGFI decreased the fraction of apoptotic cells in doxorubicin but not paclitaxeltreated cells. Recent work has shown that mitogenactivated protein kinase (MAPK) and phosphotidylinositol3 (PI3) kinase are activated by IGFI in these cells. PI3 kinase activation has been linked to antiapoptotic functions while MAPK activation is associated with proliferation. We found that IGFI rescue of doxorubicininduced apoptosis required PI3 kinase but not MAPK function, suggesting that IGFI inhibited apoptosis. In contrast, IGFI rescue of paclitaxelinduced apoptosis required both PI3 kinase and MAPK, suggesting that IGFImediated protection was due to enhancement of proliferation. Therefore, IGFI attenuated the response of breast cancer cells to doxorubicin and paclitaxel by at least two mechanisms: induction of proliferation and inhibition of apoptosis. Thus, inhibition of IGFI action could be a useful adjuvant to cytotoxic chemotherapy in breast cancer.     

Tumour necrosis factor‐α and interleukin‐1 and ‐6 in fibrocystic breast disease

The risk of developing breast cancer is higher in women presenting gross cystic disease (cysts > 3mm in diameter) of the breast with intracystic K+/Na+ > 3 as compared with K+/Na+ < 3. The present study reports the levels of tumour necrosis factor (TNF), interleukin1 (IL1), and interleukin6 (IL6) in the breast cyst fluid of women with gross cystic disease and analyses the relationship between the intracystic concentration of these cytokines, sex steroid hormones, and the K+/Na+ ratio. The concentration of these cytokines, estradiol, testosterone, dehydroepiandrosterone sulfate (DHEAS), and 17OHprogesterone were determined in the breast cyst fluid of 54 women with gross cystic disease. No significant differences were found in the cystic levels of IL1 between cysts with intracystic K+/Na+ < 3 and > 3. However, in cysts with intracystic K+/Na+ > 3 we found a lower concentration of IL6 and TNF than in those with intracystic K+/Na+ < 3.Stepwise multiple linear regression analysis demonstrated that the concentration of IL6 in breast cyst fluid was predicted statistically by a negative regression coefficient for the concentration of estradiol and DHEAS, and by a positive regression coefficient for the concentration of TNF. The concentration of TNF in breast cyst fluid was predicted statistically by a positive regression coefficient for the concentration of IL6, and by a negative regression coefficient for the concentration of estradiol. No candidate variable was included in the model to predict concentrations of IL1 in breast cyst fluid. Our results indicate that IL6 and TNF could have a local protector role in gross cystic disease, and that they could be used as a marker to identify cyst type.     

Unfavourable change in mammographic patterns and the breast cancer risk factors

The purpose of the study was to estimate the incidence of unfavourable mammographic pattern by risk factors of breast cancers. Data consisted of 1890 Finnish women with mammographic pattern of either N1 or P1 (Wolfe's classification) at the initial screening. The screening was repeated every second year from 1982 to 1990 and at each screening round the mammographic pattern was assessed. The incidence rate of P2,DY pattern was 1.9/100 woman years. The incidence of P2,DY pattern was significantly related to age. The ageadjusted odds ratio (OR) (based on logistic regression) was 2.0 (95 CI 1.0–3.9) among women with hormonal replacement therapy (HRT), 0.6 (95 CI 0.4–0.9) among postmenopausal women, 0.2 (95 CI 0.1–0.4) among women with large breasts and 0.2 (95 CI 0.1–0.3) among women with large body mass index (BMI). After multivariate adjustment by logistic regression only the effect of BMI remained statistically significant, odds ratio of P2,DY pattern for women with BMI 25 or more was 0.2 (95 CI 0.1–0.6) compared to women with BMI less than 20.     

Invasion marker PAI‐1 remains a strong prognostic factor after long‐term follow‐up both for primary breast cancer and following first relapse

In 1991, our group was the first to report the prognostic strength of plasminogen activator inhibitor type 1 (PAI1) in primary breast cancer. The prognostic impact of invasion markers PAI1 and urokinasetype plasminogen activator (uPA) on diseasefree survival (DFS) and overall survival (OS) in breast cancer has since been independently confirmed. We now report on the prognostic impact of PAI1 and uPA after longterm median followup of 77 months for our cohort (n=316).Levels of uPA, PAI1, and cathepsin D were determined in tumor tissue extracts by immunoenzymatic methods. Sphase fraction (SPF) was measured flowcytometrically in paraffin sections. Using logrank statistics, optimized cutoffs were found for PAI1 (14ng/mg), uPA (3ng/mg), cathepsin D (41pmol/mg), and SPF (6%). In all patients, various factors (PAI1, uPA, nodal status, SPF, cathepsin D, grading, tumor size, hormone receptor status) showed significant univariate impact on DFS. In Cox analysis, only nodal status (p < 0.001, RR: 3.1) and PAI1 (p < 0.001, RR: 2.7) remained significant. In nodenegative patients (n = 147), PAI1, uPA, and SPF had significant univariate impact on DFS, whereas in Cox analysis, only PAI1 was significant. PAI1 was also significant for DFS within subgroups defined by established factors. In CART analysis, uPA enhanced the prognostic value of PAI1 and nodal status for determination of a verylowrisk subgroup. For OS, only lymph node status and PAI1 were significant in multivariate analysis. PAI1 levels in the primary tumor were also a significant prognostic marker for survival after first relapse in both univariate and multivariate analysis.     

Mdm2 sensitizes MCF7 breast cancer cells to cisplatin or carboplatin

Overexpression of Mdm2 in cancer cells with otherwise wildtype p53 is believed to be an alternative mechanism for p53 inactivation during carcinogenesis. Because a number of genetic alterations that inactivate p53, including mutation, homozygous deletion, or viral oncoprotein expression (e.g. HPV16E6), inhibit DNA repair, we tested the hypothesis that Mdm2 would likewise inhibit DNA repair. Repair of cisplatininduced DNA damage was reduced in MCF7 cells overexpressing Mdm2, compared to MCF7 cells in which wildtype p53 function was intact. MCF7Mdm2 cells exhibited preferential sensitivity to cisplatin and carboplatin. MCF7Mdm2 cells showed a pronounced Sphase arrest after cisplatin treatment, similar to that observed in mutantp53 cells in the present and prior studies. MCF7 cells with intact wildtype p53, on the other hand, arrested primarily in G2/M phase after cisplatin treatment. These findings indicate that Mdm2 overexpression can recapitulate the effect of p53 mutations on DNA repair of cisplatin lesions.     

Synergistic action of apoptosis induced by eicosapentaenoic acid and TNP‐470 on human breast cancer cells

The effects of eicosapentaenoic acid (EPA) and an angiogenesis inhibitor (TNP470) on the suppression of breast cancer cell growth were examined in five human breast cancer cell lines (MDAMB231, T47D, MCF7, KPL1, and MKLF). In all five cell lines, EPA and TNP470 alone both showed tumor growth inhibition in a time and dosedependent manner, and in combination, a synergistic effect was seen at high concentrations. EPA plus TNP470 treatment evoked apoptosis as confirmed by the appearance of sub G1 populations, by DNA fragmentation, and by cell morphology. With the combination, the expression of Bax and Bc1x_S, the apoptosisenhancing proteins, was more upregulated and that of Bcl2 and Bclx_L, the apoptosissuppressing proteins, was more downregulated compared to the use of EPA or TNP470 alone, suggesting that their synergistic effect was due to an acceleration of apoptosis.     

Involvement of nuclear steroid/thyroid/retinoid receptors and of protein kinases in the regulation of growth and of c‐erbB and retinoic acid receptor expression in MCF‐7 breast cancer cells

Nuclear steroid/thyroid/retinoid receptors and cerbB membrane receptor tyrosine kinases control epithelial growth and differentiation. Retinoid receptors can dimerize with the vitamin D receptor, the glucocorticoid receptor or the thyroid receptor. Furthermore, multiple cerbB receptor dimers have been identified. It has been shown that some of these receptor pathways communicate with each other via crossconnected regulatory networks. Molecular interactions between retinoid receptors or estrogen receptors (ER) and cerbB2, and between ER and retinoic acid receptor(RAR) have been reported. Here, we demonstrate the effects of steroids/thyroids/retinoids and of activators of protein kinase A (forskolin, Forsk) and C (12Otetradecanoylphorbol13acetate, TPA), on growth and expression of cerbB and RARs in MCF7 breast cancer cells, which contain high levels of RAR and , and which express significant amounts of cerbB2 and 3. All transretinoic acid (tRA), the antiestrogen ICI 182 780 (ICI), Forsk and TPA reduced, whereas triiodothyronine and 17estradiol (E₂) stimulated cell growth. Flow cytometry revealed that tRA and E₂ reduced cerbB2 and 3, whereas tamoxifen, Forsk and TPA upregulatedcerbB2. cerbB3 was coregulated with cerbB2. Northern analysis demonstrated that RAR was downregulated by dexamethasone, ICI, and TPA, whereas vitamin D₃ and E₂ upregulated RAR. RAR expression was less responsive to such treatment, being reduced only by ICI and Forsk. These data indicate that nuclear receptor and protein kinase signaling communicate with each other and control the expression of RARs and cerbB receptors. Efficient growth control requires the coordinated interplay of both receptor systems.     

Phase II of doxorubicin/taxol in metastatic breast cancer

Purpose. To assess the response rate, survival, and toxicity of Taxol®(paclitaxel) as 1h infusion plus doxorubicin as firstline treatment for patients with metastatic breast cancer (MBC).Patients and methods. Seventysix patients with untreated MBC were recruited. All of them had measurable disease and were evaluable for toxicity. Fiftyfive percent of the patients had visceral involvement. The dose of doxorubicin was fixed at 50mg/m² as a short intravenous infusion, followed by 200mg/m² of Taxol as a 1h intravenous infusion. Doxorubicin was administered during the first seven cycles, continuing with Taxol only up to a maximum of ten cycles.Results. Neutropenia was the most important toxicity: 30% grade 3 and 18% grade 4. Only 2 patients showed a decrease in the left ventricular ejection fraction (LVEF) which caused discontinuing the treatment. No clinical congestive heart failure (CHF) was observed. Seventyfour patients were eligible for response evaluation: 10 (14%) achieved complete response (CR) and 46 (62%) achieved partial response (PR). The mean duration of response was 13.47± 1.35 months (95% confidence interval (CI): 10.82; 16.12) and the mean survival was 21.50± 1.42 months (95% CI: 18.72; 24.29).Conclusion. The overall response (OR) rate was 76%. No CHF was assessed and 2 patients stopped treatment due to LVEF decrease. Although doxorubicin 50mg/m² followed by Taxol 200mg/m² in 1h intravenous infusion presents a toxicity profile which demands a close followup, it represents a convenient outpatient schedule with similar activity rate compared to longer Taxol infusions.     

Adrenal androgens and human breast cancer: A new appraisal

A clearer picture of the role of adrenal androgens in the etiology of breast cancer is beginning to emerge. Women who develop breast cancer in premenopausal years tend to have subnormal serum levels of adrenal androgens, while subjects who develop the disease in postmenopausal years have supranormal levels of these hormones. Androgens, by acting via the androgen receptor, oppose estrogen-stimulated cell growth in premenopausal years. In postmenopausal women, elevated adrenal androgen levels stimulate cell growth by the action of the unique adrenal androgen 5-androstene-3,17-diol, also termed hermaphrodiol, via its combination with the estrogen receptor in a hormone milieu lacking, or having low concentrations of, the classical estrogen 17-estradiol.     

Genetic polymorphisms of platelet adhesive molecules: association with breast cancer risk and clinical presentation

The main platelet adhesive receptors integrin 21, integrin IIb3 and glycoprotein (GP) Ib are also expressed in breast carcinoma cells. They play a key role in tumor cell-induced platelet aggregation and in adhesive interactions necessary for tumoral invasion and metastasis. Several polymorphisms affecting these molecules, two in integrin 2 (C807T and G1648A), one in integrin 3 (T1565C) and one in GP Ib (VNTR), influencing their levels, structure, and possibly their function, have been previously described and associated with cardiovascular diseases. In this study, we investigated the association of these polymorphisms with breast cancer risk or clinical presentation. We studied 101 patients with invasive breast cancer. The main prognostic variables were recorded, and genomic PCR analysis of these polymorphisms was performed. A group of 101 control subjects matched on age and sex was studied and compared with patients. No association was found between VNTR (GP Ib) polymorphism and breast cancer risk or presentation. Genotype and allele frequencies of C807T and G1648A polymorphisms of integrin 2 were not statistically different in breast cancer patients and controls, although we found an association between the 1648G/G genotype and higher disease stages (III and IV) (p = 0.02). Breast cancer risk was higher in carriers of 3 integrin T/T genotype (OR = 2.08, 95% CI = 1.04–4.16, p = 0.04). Furthermore, genotype 1565T/T was also associated with axillary nodal metastasis (p = 0.017) and with tumoral diameter greater than 2 cm (p = 0.02). Although confirmatory studies are needed, our results suggest that polymorphic genetic variation of integrins expressed in platelets and epithelial breast cells could modify the risk and the biological aggressiveness of breast carcinomas.     

The pure antiestrogen ICI 182780 is more effective in the induction of apoptosis and down regulation of BCL‐2 than tamoxifen in MCF‐7 cells

There is increasing evidence that induction of apoptosis by antihormones is an important mechanism in regard to their growth inhibitory action on hormone dependent tumors. In this report we have compared the efficiency of tamoxifen (Tam) and the pure antiestrogen ICI 182780 (ZM) to induce apoptosis in the estrogen dependent breast cancer cell line MCF7. Clear evidence for induction of apoptosis could be demonstrated after treatment with both antiestrogens. Application of the pure antiestrogen ZM led to a significantly higher induction of apoptosis compared to the partial agonistic compound Tam. The ability of the two compounds to induce apoptosis correlated with their growth inhibitory action. On the molecular level administration of ZM led to a time dependent steady decrease of BCL2 mRNA and protein. Administration of Tam also initially decreased the expression of BCL2. In contrast to ZM treatment, BCL2 expression increased again after 8h of incubation with Tam. After 96h Tam treated cells expressed BCL2 levels nearly as high as untreated cells. In general, ZM decreased BCL2 levels more effectively than Tam. Our results demonstrate that ZM and Tam possess quantitative and qualitative differences in their ability to down regulate BCL2 expression. The higher ability of the pure antiestrogen to down regulate BCL2 expression may explain the superiority of the pure antiestrogen to induce apoptosis and to inhibit the growth of MCF7 cells.     

Carotenoids,retinol, and vitamin E and risk of proliferative benign breast disease and breast cancer

We investigated the relationship between serum levels of retinol, -carotene, -carotene, lycopene, -tocopherol, and -tocopherol as well as intakes of retinol, carotene, and vitamin E and the risks of breast cancer and proliferative benign breast disease (BBD) in a case-control study of postmenopausal women in the Boston, MA (United States) area. Serum nutrient data were available for 377 women with newly diagnosed stage I or II breast cancer and 173 women with proliferative BBD. Controls were 403 women who were evaluated at the same institutions but did not require a breast biopsy or whose biopsy revealed nonproliferative BBD. We observed no significant associations between serum levels of these micronutrients and risk of proliferative BBD or breast cancer. The risk of breast cancer was decreased among women in the highest quintile of intake of vitamin E from food sources only (odds ratio [OR] for the highest quintile = 0.4,95 percent confidence interval [CI]=0.2–0.9; P, trend across quintiles = 0.02) but less so for total vitamin E intake including supplements (OR=0.7, CI=0.4–1.3; P, trend = 0.07).This project was supported by research grant CA 40429 from the National Cancer Institute. Dr London was supported in part by an Institutional National Research Service Award (CA 09001) from the National Cancer Institute.     

Estrone sulfate: A potential source of estradiol in human breast cancer tissues

Summary Local formation of estradiol in human breast tumors could provide a more important source of estrogen than is delivered from plasma. Prior studies have suggested that estrone is primarily synthesized from estrone sulfate. The enzyme 17-hydroxysteroid dehydrogenase (HSD) would be required to convert estrone to estradiol.This study characterized HSD in 1000 × g supernatants from human breast tumors. Estradiol synthesis was linearly related to tissue concentration or time over the range studied. Cofactor requirements varied with estrone concentration. High and low affinity sites were found in 50% of tissues studied, while the remainder contained only low affinity sites. Screen assays showed measurable activity in all 42 samples tested. This activity ranged from 0.73–>100 nmol estrone synthesized/g protein/hr, with a median activity of 5.9 nmol/g/hr.We evaluated the biological relevance of the sulfatase-HSD pathway by testing the ability of estrone sulfate to stimulate colony formation in soft agar cultures of nitrosomethylurea-induced rat mammary tumors. The maximally effective concentration ranged from 10^–7 to 10^–4 M. Significant stimulation of colony formation was observed in 7 of 8 experiments. The estrone sulfate stimulation pattern was similar to that previously observed with estradiol. Of the³H-estrone sulfate added to the dishes, 20–98% was recovered as estrone and 0.2–6% as estradiol. These studies suggest that the requisite enzymes are present in human breast tumors for conversion of estrone sulfate to estradiol, and that this pathway may be biologically significant.This work presented in part at the International Congress on Endocrinology of the Breast, September 19–22, 1984, Torino, Italy.     

Phase II study of cisplatin and 5‐fluorouracil in previously treated metastatic breast cancer: an Eastern Cooperative Oncology Group study (PA 185)

Purpose: The present study was conducted to investigate the efficacy and toxicity of a cisplatin and 5fluorouracil (5FU) combination in previously treated advanced breast cancer.Methods: Thirtysix women with recurrent metastatic breast cancer were entered on a phase II study of 5FU 1000mg/m²/day given intravenously as a continuous infusion on days 1–3 and cisplatin 30mg/m²/day given intravenously over 1h on days 2–4, repeated every 21 days. All subjects had received one previous chemotherapy regimen for metastatic disease and either progressed during treatment or relapsed after responding to previous chemotherapy. Fourteen patients had also received previous adjuvant chemotherapy, 17 patients had previous radiation therapy, and 29 patients had previous hormonal therapy.Results: Among 32 responseevaluable patients, there were 10 partial remissions (31%) and 1 complete remission (3%), with an overall objective response rate of 34%. Median duration of response was 4 months. Median survival was 10.5 months for responders and 9.5 months for the entire group. Toxicity was mild to moderate in most patients. Overall twelve patients experienced grade 3 toxicity (10 hematologic, 1 mucositis, and 2 nausea). There were no grade 4 or 5 toxicities.Conclusion: Infusional cisplatin and 5FU is a well tolerated and active regimen in women with previously treated advanced breast cancer.     

Prognostic significance of Ki‐67 and topoisomerase IIα expression in infiltrating ductal carcinoma of the breast

To evaluate the prognostic relevance of Ki67 and topoisomerase II expression in relation to tumor stage, grade, and hormone receptor content, 942 ductal infiltrating carcinomas of the breast were examined by means of the monoclonal antibodies KiS11 (Ki67) and KiS4 (topoisomerase II). pS2, cerbB2, and p53 were additionally considered as prognostic variables. The median followup time was 149 months. Eighthundredandsixtythree tumors reacted with KiS11 and KiS the labeling indices of the two antigens were closely associated (r=0.93). Both correlated positively with the tumor size, cerbB2, and p53 expression, and negatively with patient age, hormone receptor content, and pS2 immunostaining. In the univariate analysis, KiS11 and KiS4 scores, nodal status, tumor size, tumor grade, and progesterone receptor content strongly predicted both overall and metastasisfree survival (p <0.00001). Estrogen receptor status, p53, and cerbB2 were of minor significance. Concerning overall survival, multivariate Cox regression analysis selected a KiS4 score >25% (p < 0.00001) next to the nodal status, and before tumor size, progesterone receptor content, and patient age. Independent predictors of the occurrence of distant metastases were nodal status, KiS4, tumor size, grade 1, and progesterone receptor negativity, in that order. The KiS11 score was of independent prognostic significance only if examined as a continuous variable. We conclude that topoisomerase II expression as assessed by monoclonal antibody KiS4 may add valuable information to current prognostic models for breast cancer. Its predictive value appears to be essentially related to the proliferative activity of tumor cells.     

Assessment of quality of life in women undergoing hormonal therapy for breast cancer: validation of an endocrine symptom subscale for the FACT‐B

Existing quality of life instruments do not include adequate items to measure the side effects and putative benefits of hormonal treatments given in breast cancer. We report the development and validation of an 18 item endocrine subscale (ES) to accompany a standardised breast cancer quality of life measure, the Functional Assessment of Cancer Therapy (FACTB) [1]. The FACTES (FACTB plus ES) was tested initially on 268 women with breast cancer receiving endocrine treatments. Alpha coefficients for all subscales demonstrated good internal consistency (range = 0.65–0.87). Testretest reliability of the ES indicated good stability (r = 0.93, p < 0.001). Advanced breast cancer patients' quality of life was high, showing the efficacy of endocrine therapy, but women with primary disease reported better physical, social, and functional wellbeing and fewer breast cancer concerns. Most frequently reported symptoms were loss of sexual interest (31%), weight gain (25%), and hot flushes (24%). Significant differences were found between treatment groups for hot flushes and vaginal dryness. Two assessments of the instrument's responsiveness to change were made; 32 women in a clinical trial of endocrine therapy and 18 women without breast cancer taking HRT completed the FACTES at baseline, 4, 8, and 12 weeks. Trial patients reported significantly more symptoms at 8 and 12 weeks than at baseline. Women taking HRT reported significantly fewer or less severe symptoms than at baseline. In conclusion the FACTES has acceptable validity and reliability and is sensitive to clinically significant change, making it suitable for clinical trials of endocrine therapy.     

Autocrine and paracrine growth regulation of human breast cancer

Summary We consider the hypothesis that estrogen control of hormone dependent breast cancer is mediated by autocrine and paracrine growth factors secreted by the breast cancer cells themselves. Though we show direct, unmediated effects of estrogen on specific cell functions, we also provide evidence that human breast cancer cells secrete a collection of growth factors (IGF-I, TGF, TGF, a PDGF-like competency factor, and at least one new epithelial colony stimulating factor). Some of these are estrogen-regulated in hormone dependent cells, and are constitutively increased in cells which acquire independence either spontaneously or byras transfection. Collectively, the secreted growth factors are capable of promoting tumor formation by MCF-7 cells in nude mice, though not to the same extent as estrogens. There would seem to be potential for clinical intervention in the autocrine and paracrine control of breast cancer cells, including some cells which are no longer dependent on estrogens.