Abstract P6-05-09: G protein-coupled estrogen receptor expressed in the plasma membrane is associated with worse breast cancer outcome, but does not contribute to tamoxifen resistance

Introduction - G protein-coupled estrogen receptor (GPER), or GPR30, is a membrane receptor reported to mediate rapid, non-genomic estrogen responses. Multiple studies suggest that GPR30 expression associates with breast cancer outcome, although the clinical implications appear to depend on receptor subcellular localization and level of expression. Specifically, high plasma membrane localization of GPR30 has been reported to associate with a worse prognosis. Furthermore, in vitro studies show that tamoxifen up-regulates GPR30 in breast cancer cells, suggesting that GPR30 acts as a mediator of tamoxifen resistance. To further clarify the prognostic and treatment resistance role of GPR30 in vivo, we studied the receptor in patients developing contralateral breast cancer with or without tamoxifen treatment for the first primary tumor, serving as an in vivo model of tamoxifen resistance.

Patients and methods - In a cohort of 688 patients with metachronous contralateral breast cancer, total and plasma membrane specific GPR30 expression were evaluated by immunohistochemistry. Total GPR30 was evaluated in five levels (0-4) and plasma membrane staining as positive or negative. Evaluation was successful for 559 first primary tumors and 595 contralateral tumors. In addition, matched lymph node and distant metastases were evaluated (lymph node matched to first primary tumor, n=213, matched to contralateral tumor, n=196, and distant metastasis, n=197). The relationship between GPR30 and breast cancer mortality was assessed by the Cox proportional hazards model, and illustrated by curves of cumulative incidence. The association between GPR30 and breast cancer outcome in relation to tamoxifen was assessed by comparing groups either untreated or treated with tamoxifen after the first primary tumor.

Results - GPR30 expression in the plasma membrane associated with increased risk of breast cancer mortality both when expressed in the contralateral tumor (HR=1.7, p=0.03) and matched lymph node (HR=2.0; p=0.02). Additionally, GPR30 plasma membrane expression associated with high Ki67 staining both in the first primary tumor (p<0.0001) and contralateral tumor (p<0.0001). In both the first primary tumor and contralateral tumor, GPR30 plasma membrane expression associated with estrogen receptor α (ER)-negativity (p<0.0001 and p<0.0001, respectively) and progesterone receptor (PR)-negativity (p=0.0007 and p<0.0001, respectively). Furthermore, ER and PR expression associated with total GPR30 expression in a biphasic manner in both the first primary tumor and contralateral tumor, as previously observed in three cohorts of primary breast cancer. The highest total and plasma membrane GPR30 expression was observed in triple-negative breast cancer. Total GPR30 expression of the first primary tumor and contralateral tumor did not correlate, but it was significantly lower in the matched lymph node (first primary tumor p<0.0001, contralateral tumor p<0.0001). No clear evidence was found that tamoxifen treatment during contralateral tumor development correlated with GPR30 expression, or that GPR30 expression correlated with response to tamoxifen treatment.

Conclusion - GPR30 expression in the plasma membrane associates with increased risk of breast cancer death when expressed in the contralateral tumor and match lymph node, and correlates strongly with multiple clinicopathological markers of poor outcome. On the other hand, prior tamoxifen treatment does not appear to affect GPR30 expression, suggesting that GPR30 does not mediate tamoxifen resistance. Future studies should aim to characterize the pathophysiological mechanisms and function of GPR30 when located in the plasma membrane. Here, GPR30 could be an novel target for breast cancer treatments.

Citation Format: Julia Tutzauer, Martin Sjöström, Pär-Ola Bendahl, Lisa Rydén, Mårten Fernö, Fredrik Leeb-Lundberg, Sara Alkner. G protein-coupled estrogen receptor expressed in the plasma membrane is associated with worse breast cancer outcome, but does not contribute to tamoxifen resistance [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P6-05-09.

Abstract P1-07-05: AIB1 is a new putative prognostic biomarker in the luminal A and B-like (HER2-negative) classification of invasive lobular carcinoma

Background: Estrogen receptor (ER) positive HER2-negative breast cancer comprises 75–80% of all breast cancer. This fraction is even higher (>90%) in invasive lobular carcinoma (ILC). According to the St Gallen surrogate definitions of the intrinsic subtypes, Ki67 and progesterone receptor (PgR) are used to classify these tumors as luminal A- and luminal B-like (HER2-negative). These guidelines are based on information derived from patient materials with mixed histological types, where the vast majority of the patients have invasive ductal carcinoma. The `luminal-like classification´ together with histological grade, tumor size and lymph node status is widely used in the clinic for prognostication. The aim of the present study was to investigate if the same markers are applicable for ILC, and furthermore, if additional biomarkers involved in the endocrine signaling system, e.g. Amplified in breast cancer 1 (AIB1) and the putative G protein-coupled estrogen receptor (GPER), might provide complementary prognostic information.

Patients: Two hundred and thirty-three (N = 233) well-characterized patients with primary ILC, diagnosed between 1980 and 1991 were included. Forty-two percent of the patients received adjuvant endocrine treatment and 2 % received adjuvant chemotherapy. All biomarkers were analyzed immunohistochemically on tissue microarray, whereas histological grade was evaluated on whole sections according to Elston and Ellis (NHG). The primary endpoint was breast cancer mortality (BCM).

Results: In univariable analyses with 10-year follow-up, Ki67 (high vs. low), NHG (3 vs. 1+2) and AIB1 (high vs. low) were significantly associated to BCM (Hazard Ratio: 4.7, 95% CI: 2.1–10.4, p <0.001; HR: 3.1, 95% CI: 1.5–6.4, p = 0.003; HR: 3.2, 95% CI: 1.4–7.2, p = 0.005 respectively), whereas PgR (<1% vs ≥1%) and GPER (linear 0-4) were not (p = 0.25; p = 0.31 respectively). Essentially the same effect was seen after multivariable adjustment for lymph node status (+ vs. -), tumor size (>20 mm vs. <20 mm), adjuvant treatment and age (continuous). Subgrouping the tumors into luminal A- and B-like (HER2-negative) according to St Gallen surrogate definitions did not show significant prognostic differences between the two groups (p = 0.12). Patients with <20 mm, lymph node negative breast cancer and favorable tumor characteristics (low Ki67, NHG 1+2, and low AIB1) had a 10-year BCM of 4.2% (95% CI: 1.4–12%). This group constituted 34% of the patients included in the present study.

Conclusions: In contrast to other previous studies, where breast cancers of mixed histological types were included, PgR was not significantly associated to prognosis in the ER-positive HER2-negative subgroup in the present study, consisting only of ILC. The prognostic role of PgR and the clinical usefulness of the luminal A and B-like (HER2-negative) classification (using only Ki67 and PgR) in ILC is still to be further investigated. The prognostic importance of Ki67 and NHG in this subgroup was, however, confirmed also in ILC, and AIB1 might be a new putative prognostic factor. By combining Ki67, NHG, and AIB1, together with lymph node status and tumor size, a group of patients with an excellent prognosis could be identified.

Citation Format: Narbe U, Forsare C, Bendahl P-O, Lövgren K, Alkner S, Sjöström M, Rydén L, Leeb-Lundberg F, Ingvar C, Fernö M. AIB1 is a new putative prognostic biomarker in the luminal A and B-like (HER2-negative) classification of invasive lobular carcinoma [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P1-07-05.

Abstract P1-08-12: G protein-coupled estrogen receptor in the plasma membrane is prognostic in early breast cancer

Background: G protein-coupled estrogen receptor (GPER), also known as GPR30, is a novel putative estrogen receptor. Although contradictory results have been presented e.g. regarding the subcellular localization and function of the receptor, previous studies have shown a prognostic value in breast cancer and proposed treatment predictive information for tamoxifen (Tam). This study aimed at clarifying the prognostic and treatment predictive value for Tam of GPER, in different subcellular localizations, by using samples from a randomized clinical trial - the ideal population for assessing treatment prediction.

Material and Methods: GPER levels were assessed semi-quantitatively by immunohistochemistry in tissue microarrays from 742 postmenopausal breast cancer patients with no lymph node metastasis and tumor size ≤ 30mm. Patients were originally included in the STO-3 trial 1976-1990. After surgery, they were randomized to Tam treatment (40mg for 2 years or no systemic treatment), regardless of classical estrogen receptor α (ER) status. GPER staining was evaluated in carcinoma both as intensity in 5 levels regardless of subcellular localization, and in the plasma membrane in 3 levels. Due to statistical considerations regarding group size, the final analysis was made with intensity in 3 levels and plasma membrane as positive or negative. The Kaplan-Meier method and logrank test (for trend when applicable) were used for survival analysis and Cox regression analysis for obtaining hazard ratios (HR), interaction testing and multivariate modeling. Distant disease-free survival (DDFS) was used as endpoint.

Results: Analyzing all patients, we found no association between DDFS and GPER intensity. However, positive plasma membrane staining showed a strong correlation with poor prognosis (HR 1.8 p = 0.002). This was only observed in the ER+ subgroup (ER+ patients HR 2.1, p<0.001, ER- patients HR 1.1 p = 0.79). The prognostic value, in untreated patients only, was analyzed with similar results (plasma membrane staining positive vs. negative: all untreated patients HR 1.8 p = 0.008, ER+ patients HR 2.1 p = 0.003, ER- patients HR 1.1 p = 0.83).

No obvious difference in tamoxifen response was observed across plasma membrane or intensity groups, and tests for interaction were not significant.

A multivariate model including GPER in plasma membrane, ER, histological grade, HER2, tamoxifen and tumor size showed that GPER was an independent prognostic factor (HR 1.6 p = 0.01). Finally we created a group with ER+, progesterone receptor (PR) + patients treated with Tam, as this group today is treated with Tam and thought to have a good response. GPER in the plasma membrane significantly separated this group into an excellent prognosis group and a poor prognosis group (HR 3.3, p = 0.01). The excellent prognosis group, which constitutes more than half of ER+ patients, had a 20 year DDFS of 91% (95% CI 84-95).

Conclusion: We found no treatment predictive value of GPER for Tam. However, GPER expressed in the plasma membrane was a strong independent prognostic factor for a poor prognosis in ER+ breast cancer. Used in ER+, PR+, tamoxifen treated patients, it can distinguish patients with an excellent prognosis from patients with a poor outcome that may benefit from additional treatment.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-08-12.

New topics in bradykinin research

Bradykinin has been implicated to contribute to allergic inflammation and the pathogenesis of allergic conditions. It binds to endothelial B(1) and B(2) receptors and exerts potent pharmacological and physiological effects, notably, decreased blood pressure, increased vascular permeability and the promotion of classical symptoms of inflammation such as vasodilation, hyperthermia, oedema and pain. Towards potential clinical benefit, bradykinin has also been shown to exert potent antithrombogenic, antiproliferative and antifibrogenic effects. The development of pharmacologically active substances, such as bradykinin receptor blockers, opens up new therapeutic options that require further research into bradykinin. This review presents current understanding surrounding the role of bradykinin in nonallergic angioedema and other conditions seen by allergists and emergency physicians, and its potential role as a therapeutic target.

The functional interaction of EGF and PDGF with bradykinin in the proliferation of human gingival fibroblasts

Epidermal growth factor (EGF) and platelet-derived growth factor (PDGF)-BB are both involved in periodontal wound healing. Each of these growth factors exerts a positive proliferative effect on cells of the periodontium in vitro. However, in vivo the peptide bradykinin is one of a complex array of mediators present in addition to these growth factors. The purposes of this investigation were to: 1) evaluate bradykinin interactions with EGF and PDGF-BB altering cell proliferation in cultured human gingival fibroblasts (HGF), periodontal ligament cells (HPDL), and cells derived from alveolar bone (HOB); and 2) determine at the signal transduction level the mechanism of interaction between EGF and bradykinin in HGF. EGF and PDGF-BB stimulated DNA synthesis in a concentration-dependent manner, as measured by [3H] thymidine incorporation. Bradykinin alone did not alter significantly based DNA synthesis values; however, bradykinin in combination with EGF reduced DNA synthesis to nearly basal levels and bradykinin in combination with PDGF reduced the DNA synthesis over 50%. Examination of the interactions between bradykinin and EGF signal transduction pathways revealed that PGE2 release was increased in the presence of bradykinin and EGF (167 +/- 33% to 317 +/- 29%). The bradykinin-stimulated PGE2 release was completely abolished by indomethacin. Indomethacin also was found to block the bradykinin inhibition of EGF-induced DNA synthesis.(ABSTRACT TRUNCATED AT 250 WORDS)

Bradykinin inhibition of EGF- and PDGF-induced DNA synthesis in human fibroblasts

Bradykinin exhibits proliferative influences in several types of cells; however, in the present study, bradykinin did not promote DNA synthesis but actually inhibited the DNA synthesis induced by epidermal growth factor (EGF) and platelet-derived growth factor (PDGF) in human gingival fibroblasts (HGF). This dose-dependent inhibitory effect was a specific intracellular interaction in that increasing concentrations of EGF did not counteract the inhibitory actions of bradykinin when added at 100 nM. The phosphoinositide-calcium signaling cascade is a likely point of interaction for the inhibitory influences of bradykinin; however, no interactions between bradykinin and EGF were observed with the generation of inositol phosphates or intracellular calcium fluxes. The inhibitory influences of bradykinin do not appear to be the result of a transmodulation of the EGF receptor, since EGF-mediated autophosphorylation was not negatively affected by bradykinin. Bradykinin-stimulated prostaglandin E2 (PGE2) release was potentiated by EGF, and, in the presence of indomethacin, the inhibition of the EGF-induced DNA synthesis by bradykinin was minimized. The results presented demonstrate that bradykinin can inhibit EGF- and PDGF-induced DNA synthesis and suggest that PGE2 synthesis is responsible for the observed bradykinin inhibition of EGF-induced DNA synthesis.

Dihydropicrotoxinin binding sites in mammalian brain: interaction with convulsant and depressant benzodiazepines

The specific binding of [3H] alpha-dihydropicrotoxinin to rat brain membranes was inhibited competitively and potently (IC50 congruent to 100 nM) by a convulsant benzodiazepine drug, RO5-3663. This compound did not inhibit high affinity flunitrazepam binding to the same tissue under similar conditions, and its reported pharmacological activity as an antagonist of GABAergic synaptic transmission, which resembles that of picrotoxinin, appears to involve the picrotoxinin binding sites. Other benzodiazepines such as diazepam, in micromolar concentrations, inhibited picrotoxinin binding in a stereospecific and chemically specific manner. However, the order of potency of a series of depressant benzodiazepines did not correlate well with pharmacological activities nor with reported activities for displacement of high affinity benzodiazepine 'receptor' binding sites (although heterogeneity of both picrotoxinin and benzodiazepine binding site populations may make difficult such comparisons). A comparison of benzodiazepine-displaceable benzodiazepine binding and benzodiazepine-displaceable picrotoxinin binding for different brain regions and subcellular fractions revealed a very similar though not identical distribution of these two classes of drug receptor, again suggesting that the two are not identical. Both classes of drug binding site also showed a very similar distribution to sodium-independent GABA receptor binding sites, which is consistent with other evidence that at least part of these 3 receptor types may be found at least sometimes coupled together in the postsynaptic membrane GABA receptor-ionophore complex.

Interaction of anticonvulsants with the barbiturate-benzodiazepine-GABA receptor complex

Unlike the anesthetic barbiturate pentobarbital and the anxiolytic pyrazolopyridine etazolate, which enhance [3H]diazepam binding to rat brain membranes, the anticonvulsant barbiturates phenobarbital and metharbital, and also chlormethiazole, at therapeutic concentrations (10-1000 muM), do not stimulate [3H]diazepam binding, but instead block the enhancement by both pentobarbital and etazolate. The same anticonvulsants at similar concentrations inhibit [3H]alpha-dihydropicrotoxinin (DHP) binding suggesting that these anticonvulsants compete for the same receptor sites as pentobarbital and etazolate, designated the barbiturate-picrotoxinin receptor component of the benzodiazepine-GABA receptor complex.

Barbiturate receptor sites are coupled to benzodiazepine receptors

Barbiturates enhance the binding of [3H]diazepam to benzodiazepine receptor sites in rat brain. This effect occurs at pharmacologically relevant concentrations of barbiturates, and the relative activity of a series of compounds correlates highly with anesthetic activity of the barbiturates and with their ability to enhance postsynaptic inhibitory responses to the neurotransmitter gamma-aminobutyric acid. Barbiturate enhancement of benzodiazepine binding is stereospecific, with the more active anesthetic isomers of N1-methylbarbiturates being also more active than their stereoisomers in enhancing benzodiazepine binding. The active barbiturates produce a reversible enhancement in the affinity of specific benzodiazepine binding with no effect on the number of binding sites. The barbiturate enhancement, but not the baseline benzodiazepine binding, is competitively inhibited by the convulsant picrotoxinin (at 1-10 microM), a drug that has been shown to label barbiturate-sensitive brain membrane sites related to the gamma-aminobutyric acid receptor-ionophore complex. The barbiturate effect is also dependent upon the presence of certain anions, and only those anions, that penetrate the chloride channels regulated by gamma-aminobutyric acid receptors. These results suggest that picrotoxin-sensitive barbiturate binding sites are coupled to benzodiazepine receptors in the gamma-aminobutyric acid receptor-ionophore complex, and that these binding sites have the properties of pharmacologically relevant receptors that mediate at least part of the action of various nervous system depressant and excitatory drugs.

Endogenous inhibitors of picrotoxinin-convulsant binding sites in rat brain

The binding of [3H]-alpha-dihydropicrotoxinin (DHP) to rat brain membranes was inhibited by aqueous extracts of the same tissue, as well as by several naturally occurring substances such as pyrimidines, purines, and their metabolites. The 200 000 X g supernatant fraction of cortex homogenates inhibited DHP binding competitively, reversibly, without preincubation, at 0 degrees C, with 40 microgram protein inhibiting 50%. Compounds such as adenine, hypoxanthine, inosine, cytosine, allantoin and uric acid at about 0.1 mM also inhibited DHP binding competitively, suggesting a relationship between DHP binding sites, benzodiazepine binding sites, and GABA receptor binding sites.