α Ratio Predicts Response of Pancreatic Cancer Cells to Estrogens and Phytoestrogens

Journal of Surgical Research 140, 55– 66 (2007) doi:10.1016/j.jss.2006.10.015

Estrogen Receptor ␤/␣ Ratio Predicts Response of Pancreatic Cancer Cells to Estrogens and Phytoestrogens Srivani Konduri, Ph.D., and Roderich E. Schwarz, M.D., Ph.D.1 Department of Surgery, University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School, Division of Surgical Oncology, The Cancer Institute of New Jersey, New Brunswick, New Jersey Submitted for publication August 7, 2006

reduced cytotoxicity at physiological concentrations may have clinical implications on PaCa therapy. © 2007

Background. Reports on hormone receptor expression of pancreatic cancer (PaCa) cells and treatment responses to antihormonal therapy are conflicting. We examined estrogen receptor (ER) expression in PaCa cells and investigated its function in estrogen-mediated cell proliferation. Methods. Protein levels of ER␣ and ER␤ in 8 human PaCa lines were detected by Western blot analysis. Cell proliferation was measured by sulforhodamine B analysis. ER modulators included diethylstilbestrol (DES), estradiol (E2), 4-hydroxytamoxifen (Tam), genistein (Gen), and Coumestrol (Coum). Results. ER␣ levels were detected in all eight, and ER␤ in seven cell lines. ER␤/ER␣ ratio ranged from 0.4 to 111 (median: 6.4, >5 in seven lines). Median maximal growth stimulation (in %, observed at 20 to 200 nM) was 19 (DES), 39 (E2), 20 (Tam), 22 (Gen), and ⴚ9 (Coum); median maximal inhibition (at 40 to 60 ␮M) was 59 (DES), 36 (E2), 25 (Tam), 43 (Gen), and 50 (Coum). The extent of E2 and Gen stimulatory effects correlated with the ER␤/ ER␣ ratio (Kendall’s ␶: 0.714, P ⴝ 0.024), but not ER␣ or ER␤ levels alone. Only Coum-induced inhibition correlated with the ER␤/ER␣ ratio (P ⴝ 0.006) and with ER␣ expression (r ⴝ 0.753, P ⴝ 0.03). Gemcitabine-induced PaCa cytotoxicity (at IC 40) was significantly reduced by E2, Gen, and Coum. Conclusions. PaCa proliferation in vitro is highly estrogen sensitive, and in contrast to other reports, ERs are frequently expressed. In 7/8 cell lines, ER␤ expression outweighs ER␣ expression. The impact of the ER␤/ ER␣ ratio on estrogen-mediated growth stimulation and

Elsevier Inc. All rights reserved.

Key Words: pancreatic cancer; estrogen receptor expression; estrogen receptor ␤; antiproliferative effects of ER modulators. INTRODUCTION

Pancreatic cancer (PaCa) is currently the fourth leading cause of cancer deaths in the United States, and carries a 5-y survival rate of only 4% [1]. Even after complete resection, only between 10 and 25% of patients are alive at 5 y [2, 3]. No systemic treatment has shown improved efficacy over that with the cytosine analogue gemcitabine alone [4], so that new and improved therapy options are desperately needed. Based on the apparent hormonal imbalance of PaCa incidence, with a male to female ratio of between 1.25–1.75 to 1 [5], strategies to investigate whether PaCa is a hormonally susceptible disease have been pursued. We have become interested in estrogenic regulation of PaCa cell growth after encountering strong antitumor effects of the phytoestrogenic supplement PC-Spes [6]. Earlier analyses had found evidence for estrogen binding within normal or neoplastic pancreatic tissue [7–9]. Numerous studies of estrogen receptor (ER) expression and treatment response of ER modulation in PaCa have generated conflicting results. In support of ER representing an important PaCa progression mediator or therapeutic target are reports of preclinical or in vitro treatment benefit of the ER modulator tamoxifen (Tam) [8, 10, 11], or clinical trials of Tam containing therapy [12–18]. Of note, only one of these clinical trials was a prospective, randomized trial with three treatment arm in 108 patients, and despite a median survival of 5.25 mo (Tam) versus 3 mo (controls), statistical significance was not achieved [14]. Nevertheless, other estro-

1 To whom correspondence and reprint requests should be addressed at Department of Surgery, University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School, The Cancer Institute of New Jersey, 195 Little Albany Street, New Brunswick, NJ 08901. E-mail: [email protected].

55

0022-4804/07 $32.00 © 2007 Elsevier Inc. All rights reserved.

56

JOURNAL OF SURGICAL RESEARCH: VOL. 140, NO. 1, JUNE 1, 2007

genic agents such as 2-methoxyestradiol [19] and estradiol (E2) have shown antitumor benefits against PaCa in vivo [20, 21], and estrogens have been found to inhibit experimental pancreatic carcinogenesis [22– 24]. In addition, progesterones can mediate significant apoptotic antitumor effects in PaCa cells [25]. On the other hand, several attempts to link hormonal mechanisms to PaCa progression or treatment have failed to find such support. Earlier attempts to detect ERs in PaCa tissues failed, likely due to the lack of sensitive methods [26 –28]. PaCa risk was not found to be affected by the use of estrogen replacement therapy [29]. Diethylstilbestrol (DES) had no effect on pancreatic carcinogenesis in an in vivo model [30]. More importantly, Tam-based therapy failed to show any effect in a treatment model of PaCa [31] in small single-arm clinical trials [32, 33] or in two randomized clinical trials with 176 [34] or 44 patients [35]. It has become obvious that estrogenic effects on various target tissues are predominantly mediated through two ER types, ER␣ and ER␤. Agonistic or antagonistic effects depend on the combination of expression of these receptor subtypes and the different agonistic/antagonistic effects of various selective ER modulators (SERMs) [36]. In those studies that successfully identified the presence of ER in pancreatic tissues through traditional immunoabsorbance assays (equating to ER␣ detection), intratumoral levels were considered to be lower in cancer tissues than in normal pancreatic parenchyma [37, 38]. ER␣ and progesterone receptor levels in the PaCa cell line CaPan1 are higher during exponential growth, and low while stationary, indicating variability of ER expression at least in this cell line [39]. Interestingly, knowledge about ER␤ protein expression in PaCa remains insufficient to date. On the transcriptional level, ER␤ message was found to be higher in PaCa than that of (traditional) ER negative or positive breast cancer, while ER␣ RNA levels were low [40]. Because of the strong PC-Spes mediated G2/M cell cycle block and the resulting growth inhibition of PaCa cells [6], and due to the known increased ER␤ binding and transactivation activity of phytoestrogens [41, 42], evaluating the ER␤ status of PaCa and its relation to ER modulator response became of interest. We therefore examined ER␤ (and ER␣) protein expression in PaCa cells and investigated their function in estrogen-mediated cell proliferation in vitro. MATERIALS AND METHODS Cell Culture and Reagents Eight human PaCa cell lines from ATCC were studied, including BxPC3, MIA PaCa2, Panc-1, ASPC, CFPAC, HS766T, HTB 147, and CaPan2. Cells were cultured in RPMI medium supplemented with 10% FBS. Two breast cancer cell lines, MCF-7 (expressing a wild type ER ␣) and T47D (with a wild type ER ␤) were used as controls. ER modulators estradiol, distilbestrol (DES), 4-hydroxy tamoxifen

(Tam), genistein (Gen), and coumestrol (Coum) were obtained from Sigma Chemicals (St. Louis, MO). The cytotoxic agent gemcitabine (Gem) (Eli Lilly, Indianapolis, IN) was used for combination experiments.

Western Blot Analysis of ER Protein Pancreatic cancer cells were pelleted and washed twice in cold PBS and stored at ⫺70°C until further analysis. For analysis, the pellet was resuspended in Cyclin Ripa lysis buffer after adding phenylmethylsulfonylfluoride and protease inhibitors. Lysed cells were centrifuged at 12,000 rpm for 20 min and the supernatant containing the soluble protein was used for Western analysis. Protein was quantified using the Bradford method, and 20 ␮g were loaded onto a 10% polyacrylamide gel and electrophoresed at 200 V until protein bands of the molecular weight marker separated sufficiently. The proteins on the gel were transferred onto a PVDF immobilon-P membrane (Millipore, Beaford, MA). To prevent nonspecific binding, the membrane was blocked overnight in 5% nonfat dried milk at 4°C and then incubated with either ER ␣ (1:5000) or ER ␤ (1:1000) rabbit polyclonal primary antibody (Upstate, Charlottesville, VA) for 2 h at room temperature. The membranes were then incubated in a horseradish peroxidase (HRP)-conjugated goat antirabbit secondary antibody (Pierce, Rockford, IL) at 1:15,000 dilution for 1 h at room temperature. The protein bands were detected by autoradiography after chemiluminescence (Amersham, Piscataway, NJ). The intensity of each band was determined by densitometry analysis using the NIH image software program downloaded from the U.S. National Institutes of Health site, http://rsb.info.nih.gov/ nih-image/. To confirm equal loading, the membranes were then stripped and reprobed with ␣ actin mouse monoclonal primary antibody from Sigma (St. Louis, MO). Both membranes were exposed in parallel for actin and for the ER signal development to ascertain result comparability. All densitometry results were also adjusted to background signal intensity. Band intensity for each cell line was then normalized against the ␣ actin intensity, and the ER␣/␤ ratio calculated based on the normalized ER␣ to ER␤ expression intensity.

Cell Proliferation in Vitro Assay Pancreatic or breast cancer cells were plated in 96-well plates. After 6 h, E2, DES, Tam, Gen, or Coum were added to the cells in titration within a dose range from 0.02 to 80 ␮M. In parallel combination experiments, cells were exposed to Gem in addition to an ER modulator. Experiments were continued for various time periods; the standard setup was 48 h for evaluation of ER modulator dose response, but 72 h for combination experiments with gemcitabine. Some confirmatory experiments were extended to 120 h. At termination, cell growth was determined using the sulforhodamine B (SRB) assay, which involved precipitation of proteins with trichloroacetic acid. This was followed by staining them with 0.4% SRB (Sigma, St. Louis, MO) dissolved in 1% acetic acid solution. The cells were washed with freshly made 1% acetic acid to remove unbound dye followed by the addition of 10 mM unbuffered Tris. The color generated was read using a mQuant microplate reader (BioTek Instruments Inc., Winooski, VT) at an absorbance of 570 nm.

Statistical Analysis Cellular proliferation results under various conditions were standardized against untreated controls of the individual cell line. Maximum stimulation, maximum inhibition, the maximum response range (sum of maximum stimulation and inhibition), and the net maximal response (difference between maximum stimulation and inhibition) were calculated in percent, compared with control conditions. Correlations between proliferation measures and ER expression intensity were calculated via Kendall’s rank correlation (expressed in the coefficient ␶), and linear correlation (r) analysis.

KONDURI AND SCHWARZ: ER ␤/␣ RATIO IN PANCREATIC CANCER

57

FIG. 1. ER␣ and ER␤ expression. Western blot results of ER␣ and ER␤ protein in eight pancreatic cancer cell lines and two breast cancer controls are displayed, with corresponding ␣-actin bands. Significances of results were assumed at P ⬍ 0.05. Analyses were performed with StatView 5.0.1 software for Macintosh computers (SAS Institute Inc., Cary, NC).

RESULTS ER Protein Expression

Based on Western blot analysis, measurable ER␣ levels were each detected in all eight PaCa cell lines. Visible bands corresponding with ER␤ protein were identified in seven PaCa cell lines, but not for the cell line Panc-1 (Fig. 1). Median relative expression compared with the two positive breast cancer controls was 44% (ER␣, comparison cell line MCF7) and 102% (ER␤, comparison cell line T47D). Based on densitometry analysis, the ER␤/ER␣ ratio was determined for each PaCa line. This ratio ranged from 0.4 to 111, with a median of 6.4. A calculated ER␤/ER␣ ratio of greater than 5 was observed in seven of eight PaCa cell lines (Fig. 2).

Cell Proliferation in Vitro

PaCa cell proliferation in vitro over 48 h was measured in the presence of ER modulating agents at a dose range of 0.02 to 80 ␮M. All seven lines tested responded with varying degrees of growth stimulation or inhibition as a result of exposure to DES, E2, Tam, or Gen. Virtually all cell lines responded with increased proliferation to the lower concentration of ER modulators, while maximal growth inhibition was seen at the highest doses. Medians of the maximum measurable growth stimulation among all cell lines (in % above controls, always observed at 20 to 200 nM) were 19 (DES), 39 (E2), 20 (Tam), 22 (Gen), and ⫺9 (Coum). The ranges for this maximum growth stimulation among all cell lines were (in % above controls): ⫺3 to 120 for DES, ⫺6 to 88 for E2, ⫺5 to 101 for Tam, ⫺46 to 109 for Gen, and ⫺16 to 12 for Coum. The highest stimulation observed was 120% above control, for cell line HTB147 exposed to DES at 200 nM. ER modulator dose-dependent proliferation results for this and two

FIG. 2. ER␤ to ER␣ ratio, by cell line. The ratio of ER␤ to ER␣ protein expression, based on densitometry results, is shown for eight pancreatic cancer cell lines. (Color version of figure is available online.)

58

JOURNAL OF SURGICAL RESEARCH: VOL. 140, NO. 1, JUNE 1, 2007

DES

250

200

150

% Proliferation

150

HTB147 Panc-1 CFPAC

100

100

50 Con 0.02 0.2

5

20

50 0

40 60 (µM)

Con 0.02 0.2

150

HTB147 Panc-1 CFPAC

100 50 Con 0.02 0.2

5

20

40 60 (µM)

20

Gen

250

200

0

5

Tam

250

% Proliferation

HTB147 Panc-1 CFPAC

% Proliferation

% Proliferation

200

0

E2

250

200 150 100 50 0

40 60 (µM)

HTB147 Panc-1 CFPAC

Con 0.02 0.2

5

20

40 60 (µM)

FIG. 3. Cell proliferation with ER modulators. in vitro pancreatic cancer cell proliferation of three representative cell lines in the presence of four different estrogen receptor modulators is shown. HTB147: high ER␤ to ER␣ ratio; CFPAC: intermediate ER␤ to ER␣ ratio; Panc-1: low ER␤ to ER␣ ratio. DES ⫽ diethylstilbestrol; E2 ⫽ estradiol; Tam ⫽ 4-hydroxytamoxifen; Gen ⫽ genistein. (Color version of figure is available online.)

other representative cell lines are displayed in Fig. 3. Growth inhibition of ⬎25% occurred for most ER modulators and cell lines only at the highest concentration points of ⬎40 ␮M, which appear pharmacologically difficult to realize in vivo. There were two notable exceptions pertaining to the phytoestrogens Gen and Coum: Panc-1 cells displayed strong growth inhibition

at Gen does as low as 20 nM (Fig. 3), and all cell lines responded with milder growth inhibition, but no significant stimulation, to Coum. Table 1 summarizes data relevant to inhibition of proliferation for all cell lines. The medians of the greatest proliferation inhibition among all lines (in % below controls, seen at 40 to 60 ⬍1,13⬎M) were 59 (DES), 36 (E2), 25 (Tam), 43

TABLE 1 Inhibition of PaCa Cell Line Proliferation by ER Modulators*

Cell line

ER␤/ER␣ ratio

DES dose range for ⬎25% inhibition (␮M)

E2 dose range for ⬎25% inhibition (␮M)

Tam dose range for ⬎25% inhibition (␮M)

Gen dose range for ⬎25% inhibition (␮M)

Coum dose range for ⬎25% inhibition (␮M)

Panc-1 CFPAC CaPan2 ASPC BxPC-3 MIAPaCa2 HS766T HTB147

0.38 5.49 5.83 5.98 6.87 23.7 31.4 111.5

ⱖ20 60 ND ⱖ60 ⱖ40 ⱖ40 60 ⱖ40

5 60 ND ⱖ40 ⱖ40 ⱖ40 60 ⱖ40

NR 60 ND NR ⱖ60 ⱖ60 60 NR

ⱖ0.02 60 ND NR 80 80 60 60

ⱖ0.25 ⱖ20 ⱖ20 ⱖ10 ⱖ10 ⱖ0.25 ⱖ5 ⱖ10

NR ⫽ not reached; ND ⫽ not done; DES ⫽ diethylstilbestrol; E2 ⫽ estradiol; Tam ⫽ 4-hydroxytamoxifen; Gen ⫽ genistein; Coum ⫽ coumestrol. * Results reflect cell proliferation in 48-h assays compared with controls. Greatest responses observed in two to five experiments per cell line are charted.

KONDURI AND SCHWARZ: ER ␤/␣ RATIO IN PANCREATIC CANCER

59

(Gen), and 50 (Coum). The ranges for this maximum growth inhibition among all cell lines were (in % below controls): 40 to 84 for DES, 25 to 71 for E2, 12 to 43 for Tam, 20 to 54 for Gen, and 35 to 82 for Coum. In cell lines ASPC and Panc-1, exposure to ER modulating agents was extended to 72 and 120 h, but no increase in growth inhibition could be observed for any agent.

alone and growth stimulation response. The cell line pattern of maximal proliferation increase appeared comparable after E2 and Gen exposure, as depicted in Fig. 5.

Correlation of ER Expression and ER Modulation Proliferative Effects

Coum-induced inhibition, but not that of any other ER modulator tested, correlated with the ER␤/ER␣ ratio (r ⫽ 0.86, P ⫽ 0.006), and with ER␣ expression (r ⫽ 0.753, P ⫽ 0.03), but not with ER␤ levels. A maximum response range was calculated in the attempt to measure the bidirectional sensitivity and responsiveness of each cell line to any ER modulator. This was compared to a net maximum response as a measure of predominant stimulatory or inhibitory responses by cell lines. Examples for Gen exposure are shown in Fig. 6. Both measures in this case correlate significantly with the ER␤/ER␣ ratio; in addition, ER␤ levels correlated with the net maximum response after Gen exposure (r ⫽ 0.773, P ⫽ 0.04). The maximum response range of seven PaCa cell lines to four ER modulators with significant ER␤/ER␣ correlation is depicted in Fig. 7. In case of Coum, the net maximum response correlated with ER␣ (r ⫽ 0.91, P ⫽ 0.003), and the maximum response range with ER␤/ER␣ ratio (r ⫽ 0.874, P ⫽ 0.005).

To examine the tendency by individual cell lines to respond to ER modulation with both increased and decreased proliferation as dose-dependent phenomenon, we attempted to correlate these phenomena with our results from the ER␣ and ER␤ protein expression analysis. Because of the heterogeneity of the ER␣ and ER␤ densitometry results between cell lines, nonparametric rank correlations with predefined aspects of ER modulator-influence on proliferation and ER expression were performed. The extent of E2 and Gen stimulatory effects correlated with the ER␤/ER␣ expression ratio (Kendall’s ␶: 0.714, P ⫽ 0.024), but not ER␣ or ER␤ levels alone. For DES and Tam, group proliferation stimulatory results did not correlate. The overall range of proliferative impact, a measure of general bidirectional responsiveness to the ER modulation, confirmed a correlation between Gen and the ER␤/ER␣ expression ratio (P ⫽ 0.05). More complete information on Kendall correlation statistics is listed in Table 2. Linear correlation statistics showed a strong association for three of the ER modulating agents (DES, E2, Gen) between proliferation stimulation and the ER␤/ ER␣ ratio (Fig. 4). There was no significant correlation between stimulatory growth response and ER␤/ER␣ ratio after Coum exposure. There was no significant correlation between ER␣ or ER␤ expression levels

Correlation of ER Expression and ER Modulation Inhibitory Effects

ER Modulation and Gemcitabine Cytotoxicity

The effect of all five ER modulators on gemcitabinemediated PaCa cell inhibition was analyzed via 72 h assays in the two cell lines ASPC and Panc-1, that were among those more likely ones to respond to ER modulators with growth inhibition. Gemcitabine-induced Panc-1 cytotoxicity (at the IC 40 of 25 ␮M) was reduced

TABLE 2 Cell Response Based on ER Expression Data: Kendall Correlation Statistics* ER␣

Stimulation DES E2 Tam Gen Inhibition DES E2 Tam Gen

ER␤

␶

P

␶

0.48 ⫺0.429 ⫺0.333 ⫺0.238

NS NS NS NS

0.333 0.238 0.333 0.429

⫺0.143 ⫺0.048 0.238 ⫺0.143

NS NS NS NS

⫺0.048 0.238 0.143 ⫺0.619

ER␤/␣ P

NS NS NS NS NS NS NS 0.05

␶

P

0.048 0.714 0.429 0.714

NS 0.02 NS 0.02

0.048 ⫺0.148 ⫺0.143 ⫺0.333

NS NS NS NS

␶ ⫽ Kendall’s tau statistic; P ⫽ P-value; DES ⫽ diethylstilbestrol; E2 ⫽ estradiol; Tam ⫽ tamoxifen; Gen ⫽ genistein; NS ⫽ not significant (⬎0.05). * In vitro growth stimulation and inhibition of eight PaCa cell lines after exposure to various ER modulators were correlated with ER expression status.

60

JOURNAL OF SURGICAL RESEARCH: VOL. 140, NO. 1, JUNE 1, 2007

FIG. 4. Linear correlation: growth stimulation and ER␤ to ER␣ ratio, by ER modulator. The maximum observed in vitro growth stimulation over controls by four different estrogen receptor modulators is shown for seven pancreatic cancer cell lines. r ⫽ correlation coefficient. DES ⫽ diethylstilbestrol; E2 ⫽ estradiol; Tam ⫽ 4-hydroxytamoxifen; Gen ⫽ genistein. (Color version of figure is available online.)

by 38% (E2), 7% (DES), 6% (Tam), 68% (Gen), or 48% (Coum), each at the 10 nM concentration. Similar combination experiments with ASPC cells yielded reduction of gemcitabine-induced toxicity by 144% (E2), 32% (DES), 14% (Tam), and 24% (Gen). No ER modulator concentration up to 40 ␮M was able to mediate any further inhibition synergistic to the gemcitabine effect. Mild additive effects were seen at 20 to 40 ␮M of Tam. The results of these combination experiments for the Panc-1 line are shown in Fig. 8. DISCUSSION

Our results show that pancreatic cancer cells express functional estrogen receptors, generally with a high ER␤ to ER␣ ratio. This is the first report confirming

ER␤ protein expression in PaCa cells, lending support to previous observations of high ER␤ RNA levels in PaCa [40]. In vitro proliferative responses are observed in all cell lines, and stimulation and inhibition was a dose-dependent response to all SERMs evaluated. Cell lines with higher ER␤ to ER␣ ratio tended to show greater responsiveness to all agents tested. Unfortunately, it appears that physiologically obtainable doses of estrogens predominantly promote growth, and may impact negatively on cytotoxic therapy with gemcitabine. Thus, no immediate therapeutic applications of those ER modulating agents tested can be derived from our results. A possible exception from this observation may rest with the phytoestrogen coumestrol, which tended to cause less proliferative and more yet mild inhibitory effects in all cell lines. Nevertheless, the

FIG. 5. Maximal proliferation increase, by cell line. Maximal in vitro proliferative response increase over controls as a result of exposure to two different estrogen receptor modulators is shown for seven pancreatic cancer cell lines. (Color version of figure is available online.)

KONDURI AND SCHWARZ: ER ␤/␣ RATIO IN PANCREATIC CANCER

61

FIG. 6. Genistein response range, by cell line. Maximum response range (sum of maximum stimulation and inhibition), and the net maximal response (difference between maximum stimulation and inhibition) after exposure to genistein is shown for seven pancreatic cancer cell lines. (Color version of figure is available online.)

unique pattern of ER expression in PaCa, and its implications for cancer progression and therapy deserve additional comments. Most information on ER␤ expression in cancer is derived from studies on breast, colon, or prostate cancer. In normal mammary tissue, the ER␤ to ER␣ ratio is high, and decreases with increasing carcinoma progression [43]. For most breast cancers, the ER␤ message is most likely lower than the ER␣ message amount; when expressed, ER␤ appears to be able to influence ER␣ behavior [44]. ER␤-positive breast cancers are usually also ER␣-positive, but ER␤ correlates reversely with progesterone receptor (PR) status [45]. ER␤ message was found to be lowest in ER␣⫹/PR⫹ breast cancers [46]. Other studies found ER␤ to be associated with N0, low grade, and low S-phase fraction [47], and a trend toward higher ER␤2 and 5 subtypes over ER␤1 in

breast cancers of increasing grade [48]. Interestingly, ER␤ was upregulated in Tam-resistant breast cancer [49]. In colon cancers, ER␤ levels were also lower than in normal mucosa [50], although equal amounts of RNA levels suggest a posttranscriptional mechanism for this phenomenon [51]. Although equal concentrations of estrogens had different effects on colon cancer cell lines [52], ER␤⫹/ER␣⫺ colon cancer cells were growth inhibited by Gen and Tam at 10 ␮M [53]. Prostate cancers appear to carry a lower ER␤ amount that normal prostate tissue, too [54]. Higher ER␤ expression was found in untreated prostate cancer [55], or in cancer metastatic to lymph nodes or bone, in comparison to localized, treated disease [56]. ER␤ was considered important in mediating estrogen and antiestrogen inhibition of prostate cancer growth [57]. In ovarian epithelium, ER␤ levels are unaffected by the

FIG. 7. Maximum response range, by cell line. Maximum response range (sum of maximum stimulation and inhibition) after exposure to four different estrogen receptor modulators is shown for seven pancreatic cancer cell lines. r ⫽ correlation coefficient. DES ⫽ diethylstilbestrol; E2 ⫽ estradiol; Tam ⫽ 4-hydroxytamoxifen; Gen ⫽ genistein. (Color version of figure is available online.)

62

JOURNAL OF SURGICAL RESEARCH: VOL. 140, NO. 1, JUNE 1, 2007

FIG. 8. Estrogen effect on in vitro proliferation in combination with gemcitabine. Effects on cell proliferation by five different estrogen receptor modulators in the presence or absence of gemcitabine are shown for the cell line Panc-1 (low ER␤ to ER␣ ratio). (A) DES ⫽ diethylstilbestrol; (B) E2 ⫽ estradiol; (C) Tam ⫽ 4-hydroxytamoxifen; (D) Gen ⫽ genistein; (E) Coum ⫽ coumestrol. (Color version of figure is available online.)

malignant state of epithelial cells [58], and ER␣ is the dominant ER in ovarian cancers [59]. Similarly, endometrial cancer shows ER␤ message in only 35% of specimens, typically coexpressing ER␣ [60]. ER␤ expression was not found in normal lung tissues or lung adenocarcinomas [61]. In comparison with most of

these other organ sites, pancreatic ER␤ expression appears to persist or resurge in most invasive cancers, and to correspond to measurable yet lower levels of ER␣ protein. This unique pattern of ER regulation in PaCa cells may suggest some special growth promoting characteristics, which may still maintain a rationale for hor-

KONDURI AND SCHWARZ: ER ␤/␣ RATIO IN PANCREATIC CANCER

FIG. 8.

monally targeted therapy in this disease. While it appears not possible to draw firm conclusions from the effects associated with the ER␤/␣ ratio in our results, the findings certainly suggest an importance to both ER types for hormonal, antiestrogenic PaCa therapy. For this goal, a better understanding of the interactions between ER␣ and ER␤ appears desirable. Four distinct ER signaling pathways have been described [62], and evidence exists for a spectrum of ER ␣ and ␤ interactions from distinct pathway signaling activity [63] over heterodimerization [64] to ER␤-induced modulation of ER␣ transcriptional activity [65]. ER␤ transactivation itself is strongly dependent on the receptor ligand: ER␤-mediated E2 effects at AP1 sites were found to inhibit, those of antiestrogens were found to activate transcription [66]. Some designer SERMs create more active ER/AP1 conformations through ER␤ than through ER␣, and would therefore be of interest for the evaluation of PaCa cell growth inhibition. Overall, biological effects of ER modulators are not alone determined by the ER subtype involved, but as importantly by the ligand structure, the nature of the responsive gene promoter, the character and balance of coactivators and corepressors, and the diversity of tar-

63

Continued

get genes [67, 68]. Unfortunately, most of these elements remain unstudied in PaCa. It thus remains unclear which ER ligand would be able to consistently induce growth inhibition of PaCa at pharmacologically obtainable concentrations, and if any such response could even be consistently elicited in different, heterogeneous PaCas. Phytoestrogens deserve special mentioning when discussing ER␤-mediated cell growth phenomena. Phytoestrogens have been shown to be ER␤-specific agonists for transcriptional activity, in contrast to for instance E2 [42]. Gen and Coum were found to have the strongest ER␤ binding affinity out of several plantderived estrogens [41]. At least in breast cancer cells, Gen or quercetin, but not other phytoestrogens were inducing apoptosis, suggesting variable ER transactivation or a variability in cellular targets [69]. Also, in MCF7 breast cancer cells, the transcriptional activity of phytoestrogens was weaker than ER binding affinity would suggest [70]. We had observed strong antitumor effects of the phytoestrogenic herbal mix PC-Spes [6], but were now unable in this analysis to elicit a similar PaCa cell response to Gen or Coum at nanomolar concentrations, with the exception of the low ER␤-expressing

64

JOURNAL OF SURGICAL RESEARCH: VOL. 140, NO. 1, JUNE 1, 2007

line Panc-1 for Gen toxicity, and a broader range of cell line susceptibility to Coum at higher concentrations. In this context it appears simplistic to attribute the mixed PaCa response to phytoestrogens to the gender origin of the cancer cell line [71]. Much more complex regulators of ER-mediated cell growth or arrest signals are suspected to be operational, aside from the ER-related functional constellation in PaCa likely being unique. One also needs to consider that other cancer-promoting mechanisms may also be important targets for ERregulated activities, such as VEGF induction [72] or coupling with EGF receptor signaling [73]. In this case, lack of cell death or antiproliferative responses to ER modulation in vitro would not equate to the lack of any growth inhibition in vivo, and beneficial effects of targeting ER could be obtainable through inhibition of secondary mechanisms. Thus, our results with phytoestrogens in vitro should not discourage from in vivo treatment evaluations. In summary, PaCa proliferation in vitro is highly estrogen sensitive, and in contrast to other reports, ERs are frequently expressed. In seven out of eight cell lines, ER␤ protein expression outweighs ER␣ expression. The impact of the ER␤/ER␣ ratio on estrogenmediated growth stimulation and reduced cytotoxicity at physiological concentrations may have clinical implications on PaCa therapy. The role of high ER ␤/␣ ratio for tumor progression of PaCa deserves verification in pathology specimens, and further investigation regarding therapeutic targeting options. For this, selective ER␤ antagonists would appear desirable agents in the ER-directed treatment of PaCa. REFERENCES 1. 2. 3.

4.

5.

6.

7.

8.

9.

Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2006. CACancer J Clin 2006;56:106. Cameron JL, Riall TS, Coleman J, et al. One thousand consecutive pancreaticoduodenectomies. Ann Surg 2006;244:10. Fong Y, Gonen M, Rubin D, et al. Long-term survival is superior after resection for cancer in high-volume centers. Ann Surg 2005;242:540. Burris HA III, Moore MJ, Andersen J, et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997;15:2403. Andren-Sandberg A, Hoem D, Backman PL. Other risk factors for pancreatic cancer: hormonal aspects. Ann Oncol 1999; 10(Suppl 4):131. Schwarz RE, Donohue CA, Sadava D, et al. Pancreatic cancer in vitro toxicity mediated by Chinese herbs SPES and PC-SPES: Implications for monotherapy and combination treatment. Cancer Lett 2003;189:59. Greenway B, Iqbal MJ, Johnson PJ, et al. Oestrogen receptor proteins in malignant and fetal pancreas. Br Med J (Clin Res Ed) 1981;283:751. Pousette A, Carlstrom K, Skoldefors H, et al. Purification and partial characterization of a 17 ␤-estradiol-binding macromolecule in the human pancreas. Cancer Res 1982;42:633. Benz C, Hollander C, Miller B. Endocrine-responsive pancreatic

10.

11.

12.

13. 14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

26.

27.

28.

29.

carcinoma: Steroid binding and cytotoxicity studies in human tumor cell lines. Cancer Res 1986;46:2276. Poston GJ, Townsend CM, Jr., Rajaraman S, et al. Effect of somatostatin and tamoxifen on the growth of human pancreatic cancers in nude mice. Pancreas 1990;5:151. Selvan RS, Metzgar RS, Petrow V. Growth modulatory effects of some 6-methylenic steroids on human and hamster pancreatic adenocarcinoma cells in vitro. Drug Des Discov 1992;9:119. Theve NO, Pousette A, Carlstrom K. Adenocarcinoma of the pancreas—a hormone sensitive tumor? A preliminary report on Nolvadex treatment. Clin Oncol 1983;9:193. Wong A, Chan A, Arthur K. Tamoxifen therapy in unresectable adenocarcinoma of the pancreas. Cancer Treat Rep 1987;71:749. Keating JJ, Johnson PJ, Cochrane AM, et al. A prospective randomized controlled trial of tamoxifen and cyproterone acetate in pancreatic carcinoma. Br J Cancer 1989;60:789. Wong A, Chan A. Survival benefit of tamoxifen therapy in adenocarcinoma of pancreas. A case-control study. Cancer 1993;71:2200. Rosenberg L, Barkun AN, Denis MH, et al. Low dose octreotide and tamoxifen in the treatment of adenocarcinoma of the pancreas. Cancer 1995;75:23. Horimi T, Takasaki M, Toki A, et al. The beneficial effect of tamoxifen therapy in patients with resected adenocarcinoma of the pancreas. Hepatogastroenterology 1996;43:1225. Wenger FA, Jacobi CA, Siderow A, et al. Hormone therapy of postoperative recurrent pancreatic carcinoma with octreotide and tamoxifen. Chirurg 1999;70:694. Schumacher G, Kataoka M, Roth JA, et al. Potent antitumor activity of 2-methoxyestradiol in human pancreatic cancer cell lines. Clin Cancer Res 1999;5:493. Sumi C, Brinck-Johnsen T, Longnecker DS. Inhibition of a transplantable pancreatic carcinoma by castration and estradiol administration in rats. Cancer Res 1989;49:6687. Kuramoto M, Yamashita J, Ogawa M. Tissue-type plasminogen activator predicts endocrine responsiveness of human pancreatic carcinoma cells. Cancer 1995;75:1263. Lhoste EF, Roebuck BD, Brinck-Johnsen T, et al. Effect of castration and hormone replacement on azaserine-induced pancreatic carcinogenesis in male and female Fischer rats. Carcinogenesis 1987;8:699. Sumi C, Longnecker DS, Roebuck BD, et al. Inhibitory effects of estrogen and castration on the early stage of pancreatic carcinogenesis in Fischer rats treated with azaserine. Cancer Res 1989;49:2332. Longnecker DS, Sumi C. Effects of sex steroid hormones on pancreatic cancer in the rat. Int J Pancreatol 1990;7:159. Abe M, Yamashita J, Ogawa M. Medroxyprogesterone acetate inhibits human pancreatic carcinoma cell growth by inducing apoptosis in association with Bcl-2 phosphorylation. Cancer 2000;88:2000. Targarona EM, Pons MD, Gonzalez G, et al. Is exocrine pancreatic cancer a hormone-dependent tumor? A study of the existence of sex hormone receptors in normal and neoplastic pancreas. Hepatogastroenterology 1991;38:165. Taylor OM, Teasdale J, Cowen PN, et al. Classical oestrogen receptor is not detectable in pancreatic adenocarcinoma. Br J Cancer 1992;66:503. Ollayos CW, Riordan GP, Rushin JM. Estrogen receptor detection in paraffin sections of adenocarcinoma of the colon, pancreas, and lung. Arch Pathol Lab Med 1994;118:630. Adami HO, Persson I, Hoover R, et al. Risk of cancer in women receiving hormone replacement therapy. Int J Cancer 1989; 44:833.

KONDURI AND SCHWARZ: ER ␤/␣ RATIO IN PANCREATIC CANCER 30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

Baum A, Kamino K, Son WC, et al. Lack of influence of diethylstilbestrol on induced pancreatic ductal carcinomas in Syrian hamsters. Exp Toxicol Pathol 1995;47:117. Wenger FA, Kilian M, Mautsch I, et al. Influence of octreotide and tamoxifen on tumor growth and liver metastasis in N-nitroso-bis(2-oxopropyl)amine-induced pancreatic cancer in Syrian hamsters. Horm Res 2000;54:74. Crowson MC, Dorrell A, Rolfe EB, et al. A phase II study to evaluate tamoxifen in pancreatic adenocarcinoma. Eur J Surg Oncol 1986;12:335. Swarovsky B, Wolf M, Havemann K, et al. Tamoxifen or cyproterone acetate in combination with buserelin are ineffective in patients with pancreatic adenocarcinoma. Oncology 1993;50: 226. Bakkevold KE, Pettersen A, Arnesjo B, et al. Tamoxifen therapy in unresectable adenocarcinoma of the pancreas and the papilla of Vater. Br J Surg 1990;77:725. Taylor OM, Benson EA, McMahon MJ. Clinical trial of tamoxifen in patients with irresectable pancreatic adenocarcinoma. The Yorkshire Gastrointestinal Tumor Group. Br J Surg 1993; 80:384. Katzenellenbogen BS, Katzenellenbogen JA. Estrogen receptor transcription and transactivation: Estrogen receptor ␣ and estrogen receptor ␤: Regulation by selective estrogen receptor modulators and importance in breast cancer. Breast Cancer Res 2000;2:335. Friess H, Buchler M, Kiesel L, et al. LH-RH receptors in the human pancreas. Basis for antihormonal treatment in ductal carcinoma of the pancreas. Int J Pancreatol 1991;10:151. Singh S, Baker PR, Poulsom R, et al. Expression of estrogen receptor and estrogen-inducible genes in pancreatic cancer. Br J Surg 1997;84:1085. Hollande E, Fanjul M, Houti N, et al. Expression of estrogen receptors during growth of human pancreatic adenocarcinoma cells (Capan-1)-relationship with differentiation. In Vitro Cell Dev Biol Anim 1998;34:593. Iwao K, Miyoshi Y, Ooka M, et al. Quantitative analysis of estrogen receptor-␣ and -␤ messenger RNA expression in human pancreatic cancers by real-time polymerase chain reaction. Cancer Lett 2001;170:91. Kuiper GG, Lemmen JG, Carlsson B, et al. Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor ␤. Endocrinology 1998;139:4252. An J, Tzagarakis-Foster C, Scharschmidt TC, et al. Estrogen receptor ␤-selective transcriptional activity and recruitment of coregulators by phytoestrogens. J Biol Chem 2001;276:17808. Roger P, Sahla ME, Makela S, et al. Decreased expression of estrogen receptor ␤ protein in proliferative preinvasive mammary tumors. Cancer Res 2001;61:2537. Knowlden JM, Gee JM, Robertson JF, et al. A possible divergent role for the estrogen receptor ␣ and ␤ subtypes in clinical breast cancer. Int J Cancer 2000;89:209. Cullen R, Maguire TM, McDermott EW, et al. Studies on estrogen receptor-␣ and -␤ mRNA in breast cancer. Eur J Cancer 2001;37:1118. Dotzlaw H, Leygue E, Watson PH, et al. Estrogen receptor-␤ messenger RNA expression in human breast tumor biopsies: Relationship to steroid receptor status and regulation by progestins. Cancer Res 1999;59:529. Jarvinen TA, Pelto-Huikko M, Holli K, et al. Estrogen receptor ␤ is coexpressed with ER␣ and PR and associated with nodal status, grade, and proliferation rate in breast cancer. Am J Pathol 2000;156(1):29 –35. Leygue E, Dotzlaw H, Watson PH, et al. Expression of estrogen

65

receptor ␤1, ␤2, and ␤5 messenger RNAs in human breast tissue. Cancer Res, 1999;59:1175. 49.

Speirs V, Malone C, Walton DS, et al. Increased expression of estrogen receptor ␤ mRNA in tamoxifen-resistant breast cancer patients. Cancer Res 1999;59:5421.

50.

Campbell-Thompson M, Lynch IJ, Bhardwaj B. Expression of estrogen receptor (ER) subtypes and ER␤ isoforms in colon cancer. Cancer Res 2001;61:632.

51.

Foley EF, Jazaeri AA, Shupnik MA, et al. Selective loss of estrogen receptor ␤ in malignant human colon. Cancer Res 2000;60:245.

52.

Fiorelli G, Picariello L, Martineti V, et al. Functional estrogen receptor ␤ in colon cancer cells. Biochem Biophys Res Commun 1999;261:521.

53.

Arai N, Strom A, Rafter JJ, et al. Estrogen receptor ␤ mRNA in colon cancer cells: Growth effects of estrogen and genistein. Biochem Biophys Res Commun 2000;270:425.

54.

Latil A, Bieche I, Vidaud D, et al. Evaluation of androgen, estrogen (ER ␣ and ER ␤), and progesterone receptor expression in human prostate cancer by real-time quantitative reverse transcription-polymerase chain reaction assays. Cancer Res 2001;61:1919.

55.

Ito T, Tachibana M, Yamamoto S, et al. Expression of estrogen receptor (ER-␣ and ER-␤) mRNA in human prostate cancer. Eur Urol 2001;40:557.

56.

Leav I, Lau KM, Adams JY, et al. Comparative studies of the estrogen receptors ␤ and ␣ and the androgen receptor in normal human prostate glands, dysplasia, and in primary and metastatic carcinoma. Am J Pathol 2001;159:79.

57.

Lau KM, LaSpina M, Long J, et al. Expression of estrogen receptor (ER)-␣ and ER-␤ in normal and malignant prostatic epithelial cells: Regulation by methylation and involvement in growth regulation. Cancer Res 2000;60:3175.

58.

Lau KM, Mok SC, Ho SM. Expression of human estrogen receptor-␣ and -␤, progesterone receptor, and androgen receptor mRNA in normal and malignant ovarian epithelial cells. Proc Natl Acad Sci U.S.A. 1999;96:5722.

59.

Pujol P, Rey JM, Nirde P, et al. Differential expression of estrogen receptor-␣ and -␤ messenger RNAs as a potential marker of ovarian carcinogenesis. Cancer Res 1998;58:5367.

60.

Utsunomiya H, Suzuki T, Harada N, et al. Analysis of estrogen receptor ␣ and ␤ in endometrial carcinomas: correlation with ER ␤ and clinicopathologic findings in 45 cases. Int J Gynecol Pathol 2000;19:335.

61.

Omoto Y, Kobayashi Y, Nishida K, et al. Expression, function, and clinical implications of the estrogen receptor ␤ in human lung cancers. Biochem Biophys Res Commun 2001;285:340.

62.

Hall JM, Couse JF, Korach KS. The multifaceted mechanisms of estradiol and estrogen receptor signaling. J Biol Chem 2001; 17:17.

63.

Jones PS, Parrott E, White IN. Activation of transcription by estrogen receptor ␣ and ␤ is cell type- and promoter-dependent. J Biol Chem 1999;274:32008.

64.

Pettersson K, Delaunay F, Gustafsson JA. Estrogen receptor ␤ acts as a dominant regulator of estrogen signaling. Oncogene, 2000;19:4970.

65.

Hall JM, McDonnell DP. The estrogen receptor ␤-isoform (ER␤) of the human estrogen receptor modulates ER␣ transcriptional activity and is a key regulator of the cellular response to estrogens and antiestrogens. Endocrinology 1999;140:5566.

66.

Paech K, Webb P, Kuiper GG, et al. Differential ligand activation of estrogen receptors ER␣ and ER␤ at AP1 sites. Science 1997;277:1508.

66 67.

JOURNAL OF SURGICAL RESEARCH: VOL. 140, NO. 1, JUNE 1, 2007

Katzenellenbogen BS, Choi I, Delage-Mourroux R, et al. Molecular mechanisms of estrogen action: selective ligands and receptor pharmacology. J Steroid Biochem Mol Biol 2000;74:279. 68. Dutertre M, Smith CL. Molecular mechanisms of selective estrogen receptor modulator (SERM) action. J Pharmacol Exp Ther 2000;295:431. 69. Balabhadrapathruni S, Thomas TJ, Yurkow EJ, et al. Effects of genistein and structurally related phytoestrogens on cell cycle kinetics and apoptosis in MDA-MB-468 human breast cancer cells. Oncol Rep 2000;7:3. 70. Morito K, Hirose T, Kinjo J, et al. Interaction of phytoestrogens with estrogen receptors ␣ and ␤. Biol Pharm Bull 2001; 24:351.

71.

Lyn-Cook BD, Stottman HL, Yan Y, et al. The effects of phytoestrogens on human pancreatic tumor cells in vitro. Cancer Lett 1999;142:111.

72.

Bausero P, Ben-Mahdi M, Mazucatelli J, et al.. Vascular endothelial growth factor is modulated in vascular muscle cells by estradiol, tamoxifen, and hypoxia. Am J Physiol Heart Circ Physiol 2000;279:H2033.

73.

Curtis SW, Washburn T, Sewall C, et al. Physiological coupling of growth factor and steroid receptor signaling pathways: Estrogen receptor knockout mice lack estrogen-like response to epidermal growth factor. Proc Natl Acad Sci U.S.A. 1996;93: 12626.