Fulvestrant-Mediated Attenuation of the Innate Immune Response Decreases

Author Manuscript Published OnlineFirst on August 27, 2020; DOI: 10.1158/0008-5472.CAN-20-1705 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

1 Fulvestrant-mediated attenuation of the innate immune response decreases

2 ER+ breast cancer growth in vivo more effectively than tamoxifen

3 1Annelie Abrahamsson, 1Gabriela Vazquez Rodriguez and 1Charlotta Dabrosin

4 1Department of Oncology and Department of Biomedical and Clinical Sciences, Linköping

5 University, Linköping, Sweden

6 Running title: Comparison of anti-estrogen therapies in breast cancer

7 Keywords: macrophages, neutrophils, epithelial-mesenchymal transition, microdialysis,

8 mammary gland

9 *Corresponding author:

10 Charlotta Dabrosin, MD, PhD 11 Professor of Oncology 12 Linköping University 13 Division of Oncology 14 SE-581 85 Linköping, Sweden 15 E-mail: [email protected] 16 Phone: +46 13286711 17

19 Word count: 4799

20 Number of figures: 7

22 Funding: This work was supported by grants to C.D. from the Swedish Cancer Society

23 (2018/464), the Swedish Research Council (2018-02584), LiU-Cancer, and ALF of Linköping

24 University Hospital.

26 Disclosure: None of the authors have any financial, commercial or other conflicts of interest

27 to disclose.

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2020 American Association for Cancer Research. Author Manuscript Published OnlineFirst on August 27, 2020; DOI: 10.1158/0008-5472.CAN-20-1705 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

1 Abstract

2 Although blocking estrogen-dependent signaling is a cornerstone of adjuvant treatment for

3 breast cancer, 25% of patients experience recurrent disease. Stroma events including innate

4 immune responses are key in cancer progression. How different estrogen receptor (ER)-

5 targeting therapies, including the partial agonist tamoxifen and the pure antagonist fulvestrant,

6 affect the tumor stroma has not yet been elucidated. Fulvestrant is only used in

7 postmenopausal patients and its effects in the presence of estradiol remain undetermined.

8 Here we observe that fulvestrant decreases ER+ breast cancer growth compared to tamoxifen

9 in the presence of physiological levels of estradiol in human breast cancer in nude mice and in

10 murine breast cancer in immune competent mice. Fulvestrant significantly inhibited

11 macrophage and neutrophil infiltration in both models. These effects were corroborated in a

12 zebrafish model where fulvestrant inhibited neutrophil- and macrophage-dependent cancer

13 cell dissemination more effectively than tamoxifen. A comprehensive analysis of 234 human

14 proteins released into the cancer microenvironment by the cancer cells sampled via

15 microdialysis in vivo revealed that 38 proteins were altered following both treatments; 25 of

16 these proteins were associated with immune response and were altered by fulvestrant only.

17 Compared to tamoxifen, fulvestrant significantly affected inflammatory proteins released by

18 murine stroma cells. Importantly, in vivo microdialysis of human ER+ breast cancer revealed

19 that the majority of affected proteins in murine models were upregulated in patients. Together

20 these results suggest that fulvestrant targets ER+ breast cancer more effectively than

21 tamoxifen even in presence of estradiol, mainly by attenuation of the innate immune response.

22 Statement of significance: These findings demonstrate novel effects of the pure anti-estrogen

23 fulvestrant in estrogen receptor-positive breast cancer and evaluate its effects under

24 physiological levels of estradiol, representative of premenopausal patients.

1 Introduction

2 Exposure to sex steroids such as estrogens plays an important role in the development and

3 progression of breast cancer. Up to 80% of all breast cancers are estrogen receptor positive

4 (ER+) (1). Blocking ER signaling is a cornerstone in both adjuvant and metastatic treatment

5 of breast cancer (2,3). Different strategies are used to target ER-dependent signaling,

6 including aromatase inhibitors (AIs) that reduce estrogen synthesis, selective estrogen

7 receptor modulators (SERMs), and selective estrogen receptor down-regulators (SERDs).

8 Whereas SERMs may elicit both agonistic and antagonistic effects in different organs

9 depending on their affinity to the receptor or on the ratio of ER and ER in tissues, SERDs

10 elicit no agonistic actions as ligand binding enhances the ER to be destroyed (4,5). The

11 standard-of-care adjuvant treatment is AIs for postmenopausal women and tamoxifen, a

12 SERM, for premenopausal women (6). Treatment with an AI or tamoxifen for five to ten

13 years is associated with an increase in overall survival (6). Despite the success of these

14 therapies, approximately 25% of patients will experience recurrent disease (7). The SERD

15 fulvestrant is not used in the adjuvant setting, but is used as a second-line treatment for

16 postmenopausal patients with metastatic breast cancer who develop resistance to other anti-

17 estrogen therapies, including tamoxifen (8). Although ER- breast cancer is more common in

18 premenopausal women than in postmenopausal women, the majority of premenopausal breast

19 cancers express the ER (9). Further studies of the mechanisms underlying these different

20 therapeutic approaches are needed to potentiate anti-estrogen therapy for premenopausal

21 women with ER+ breast cancer.

22 Events in the stroma, such as immune responses and angiogenesis, are hallmarks

23 of cancer and play major roles in cancer progression (10). Several studies suggest that the

24 effects of anti-estrogen therapies as well as chemotherapy are dependent on stroma-related

25 signatures (11,12). Tamoxifen may induce a microenvironment suppressive to breast cancer

1 cells, which in turn may potentiate the effects of the drug (13). We have recently shown that

2 estradiol enhances the innate immune response by increasing the infiltration of macrophages

3 into ER+ breast cancer and attracting neutrophils to the invasive margin (14,15). Additionally,

4 low metastatic ER+ cells may become highly metastatic in presence of neutrophils and

5 macrophages, an effect that is potentiated during estradiol exposure (14,15). Several

6 chemoattractants and angiogenesis stimulators also are increased during exposure to estradiol

7 (14-18).

8 The epithelial-to-mesenchymal transition (EMT) of breast cancer cells can

9 contribute to increased tumor progression and metastases, whereas a reversal of EMT to a

10 mesenchymal-to-epithelial transition (MET) reduces stemness and metastatic capacity of

11 cancer cells (19,20). An EMT can be stimulated by several signaling pathways and via cross-

12 talks with different cell types in the tumor microenvironment (21,22). In addition, it has been

13 determined that myeloid-derived immune cells such as macrophages and neutrophils support

14 EMT in several cancer forms, including breast cancer (23,24).

15 Whether different anti-estrogen therapeutic approaches can attenuate stromal

16 events, including the innate immune response and EMT, is unknown. Here, we demonstrate

17 that the SERD fulvestrant, compared to tamoxifen, enhanced tumor regression in ER+

18 experimental breast cancer, in murine immunocompetent and immunodeficient breast cancer

19 models, in presence of physiological levels of estradiol. Fulvestrant significantly reduced the

20 number of tumor-infiltrating macrophages and neutrophils on both models and reduced EMT

21 markers more effectively than tamoxifen. In the zebrafish, fulvestrant, compared to

22 tamoxifen, significantly inhibited estradiol induced ER+ cancer cell dissemination in presence

23 of neutrophils and monocytes. In a comprehensive analysis in vivo of 234 proteins released

24 into the cancer microenvironment, 38 were significantly changed after anti-estrogen therapy;

25 25 were significantly altered by fulvestrant only. The majority of the proteins were associated

1 with immune function. Of the 38 proteins found to be altered in experimental breast cancer,

2 36 were detectable in human ER+ breast cancer in vivo. Furthermore, 31 of these proteins

3 were significantly upregulated, supporting the clinical relevance of these findings. Together

4 our data suggest that the SERD fulvestrant, compared to tamoxifen, more effectively targets

5 ER+ breast cancer both by direct effects on the cancer cells and by alterations in the tumor

6 stroma.

1 Materials and Methods

2 Cell line

3 The ER+ cell lines MCF-7 (ATCC Cat# HTB-22, RRID:CVCL_0031) re-authenticated using

4 the short tandem repeat profiling at the Uppsala Genome Center were maintained in DMEM

5 (Gibco Cat# 11880-028) with 10% FBS (Gibco Cat# 26140-079), 2 mM glutamine (Gibco

6 Cat# 25030-024) and 50 IU/ml/50 µg/ml Penicillin-G/Streptomycin (Gibco Cat#15070-063).

7 Breast cancer models

8 The Institutional Animal Ethics Committee at Linköping University approved

9 this study, which conformed to regulatory standards of animal care. Oophorectomized

10 athymic mice and FVB/N mice (Balb/C-nu/nu, 6-8 weeks old, Scanbur, Sweden) were housed

11 at Linköping University in ventilated cages with a light/dark cycle of 12/12 hours with rodent

12 chow and water available ad libitum. Mice were anesthetized via intraperitoneal (i.p.)

13 injection of ketamine/xylazine and implanted with a subcutaneous (s.c.) 3-mm pellet

14 containing either 17β-estradiol (0.18 mg/60-day release, Innovative Research of America,

15 Sarasota, FL, USA) or placebo. The active pellet releases serum concentrations of 150-250

16 pM estradiol (25).

17 Seven days after surgery, 5 × 106 MCF-7 cells or 1 × 106 or tumor cells derived from a

18 transgenic mouse strain expressing polyoma middle T (PyMT) antigen under the control of

19 the mouse mammary tumor virus (MMTV) long terminal repeat (26) were injected into the

20 dorsal mammary fat pads in 200 µl PBS. Because MCF-7 cells require estrogen for tumor

21 formation and growth, a non-estrogen control group was not possible in this in vivo model.

22 The PyMT mice develop spontaneous adenocarcinomas of all mammary epithelium by 8 to

23 10 weeks of age. These tumors were excised, dissociated in a collagenase/dispase solution

24 (100 ml PBS with 25 mg collagenase/250 mg dispase; Roche, Nutley, USA) to generate a

25 single-cell suspension and cultured until confluence. This procedure generates tumor cells

1 with maintained expression of ER at significant levels as early carcinoma stages still express

2 the receptor whereas ER expression decreases at later stages of tumor progression as

3 previously described (25,27,28). Mice were treated with fulvestrant (5 mg/mouse twice per

4 week, s.c.) or tamoxifen (1 mg/mouse every second day, s.c.) in addition to the estradiol

5 exposure. Tumor areas were calculated using the formula length/2 x width/2 x .

6 Microdialysis in mice

7 Prior to the microdialysis experiment, tumor-bearing mice with size-matched tumors were

8 anesthetized with i.p. injections of ketamine/xylazine. An anesthetic state was maintained as

9 needed by repeated s.c. injections of ketamine/xylazine. Body temperatures were maintained

10 using a heat lamp. Microdialysis probes with 4-mm membranes (CMA 20, 100-kDa cutoff;

11 CMA Microdialysis AB, Kista, Sweden) were inserted into tumor tissue and connected to a

12 microdialysis pump (CMA 102; CMA Microdialysis AB) perfused at 0.6 μl/min with 154

13 mmol/L NaCl and 60g/L hydroxyethyl starch (Voluven®; Fresenius Kabi, Uppsala, Sweden),

14 as previously described (16,17). After a 60-minute equilibrium period, outgoing perfusates

15 (i.e., microdialysates) were collected and stored at -80°C for subsequent analysis.

16 Zebrafish tumor xenograft model

17 MCF-7 cells were treated with E2 1 nM 48 hours before experiments and labeled with Fast

18 DiI™ oil red dye (ThermoFisher Scientific Cat# D3899), 4 µg/mL in PBS, 24 hours before

19 injections.

20 Human neutrophils and monocytes were isolated from venous blood from a healthy female

21 donor. Peripheral blood mononuclear cells were obtained by gradient separation with Ficoll-

22 Paque (GE Healthcare Cat#17-1440-02) and monocytes were separated by negative isolation

23 by using the Dynabeads™ Untouched™ Human Monocytes kit (ThermoFisher Scientific

24 Cat# 11350D) by following the provider’s instructions. Neutrophils were isolated as described

25 previously (PMID:30105032).

1 The animal ethics committee at Linköping University approved all zebrafish experiments.

2 Transgenic Tg(fli1:EGFP)y1 zebrafish embryos were collected and maintained in E3 embryo

3 medium with 0.2 mM 1-phenyl-2-thiourea (PTU) at 28°C. Dil-labeled MCF-7 cells were

4 injected with 50% neutrophils or with 10% monocytes into the perivitelline space of 2 days

5 old zebrafish embryos. Correctly injected embryos were selected under fluorescence and

6 incubated in E3/PTU + E2 1nM ± tamoxifen 1µM ± fulvestrant 1µM at 28°C during 1 or 3

7 days where indicated. After incubation, anesthetized zebrafish embryos were assessed for

8 cancer cell dissemination in the tail region under fluorescence. Images of disseminated cancer

9 cells were acquired with the Olympus CellSens Imaging software version 1.16 (Olympus

10 cellSens Software, RRID:SCR_016238) by using an Olympus BX43 light/fluorescence

11 microscope (10X/0.30 magnification) with excitation filters BP460-495 and BP530-550, and

12 Olympus DP72 CCD camera.

13 Immunohistochemistry

14 Formalin-fixed tumors were paraffin-embedded and cut in 4-μm sections, de-paraffinized, and

15 exposed to rat anti-mouse F4/80 (Abcam Cat# ab6640, RRID:AB_1140040), rat anti-mouse

16 Ly6G (BD Biosciences Cat# 551459, RRID:AB_394206), rat on mouse HRP Polymer Kit

17 (BioCare Medical Cat# RT517), rabbit anti-human von Willebrand factor (Agilent Cat#

18 A0082, RRID:AB_2315602), mouse anti-human Ki67 (Agilent Cat# M7240,

19 RRID:AB_2142367), rabbit anti-human N-cadherin (clone EPR19654, Abcam Cat#

20 ab207608), mouse anti-human E-cadherin (Novus Biologicals Cat# NBP2-47827), mouse

21 anti-human CD68 (Agilent Cat# GA60961-2), rabbit anti-human Slug (Abcam Cat# ab27568,

22 RRID:AB_777968), rabbit anti-human Snail (Novus Biologicals Cat# NBP2-27293), rabbit

23 anti-mouse Fibroblast activation protein (FAP) (Abcam Cat# ab218164), mouse anti-human

24 estrogen receptor α (Agilent Cat# M7047, RRID:AB_2101946), rabbit anti-human vimentin

25 (Abcam Cat# ab16700, RRID:AB_443435), Dako EnVision+System-HRP Labelled Polymer

1 anti-rabbit (Dako Cat# K4002) and anti-mouse (Dako Cat# K4000). Sections were

2 counterstained with Mayer’s hematoxylin. Negative controls did not show staining. Images of

3 10 areas of each tumor section from 3-4 mice in each treatment group, and three tumor

4 sections and three normal breast tissue sections from each patient were acquired on an

5 Olympus BX43 microscope (×40/0.75 magnification) and digitally analyzed and quantified

6 using ImageJ software version 1.52n (ImageJ, RRID:SCR_003070).

7 For immunofluorescence, sections were exposed to rat anti-mouse Ly6G,

8 incubated with conjugated donkey anti-rat antibody (Alexa flour 594, Thermo Fisher

9 Scientific Cat# A-21209, RRID:AB_2535795) and mounted using SlowFade Gold containing

10 4’,6-diamidino-2-phenylindole (DAPI) (Invitrogen Cat# S36938). Samples were visualized

11 with an Olympus BX43 light/fluorescence microscope (×40/0.75 magnification) with

12 excitation filters BP360-370 and BP530-550 using an Olympus DP72 CCD camera, analyzed

13 by CellSens Imaging software, and converted to RGB images with the same threshold using

14 ImageJ.

15 Human study

16 The Regional Ethical Review Board of Linköping, Sweden approved the study, which was

17 carried out in accordance with the Declaration of Helsinki. All subjects gave written informed

18 consent. Ten women with ongoing early ER+ breast cancer underwent microdialysis the day

19 before their scheduled surgery. During microdialysis, one catheter was inserted into the

20 cancer tissue and a second catheter was inserted into normal adjacent breast tissue. The

21 microdialysis catheters (M Dialysis AB, Stockholm, Sweden) consisted of a 10-mm long

22 tubular dialysis membrane (diameter 0.52mm, 100,000 atomic mass cut-off) glued to the end

23 of a double-lumen tube. Catheters were inserted via a splitable introducer (M Dialysis AB),

24 connected to a microinfusion pump (M Dialysis AB), and perfused with 154 mmol/L NaCl

25 and 60g/L hydroxyethyl starch (Voluven®; Fresenius Kabi, Uppsala, Sweden) at 0.5 µL/min.

1 Prior to the insertion of each catheter, each area was treated with 0.5 ml lidocaine (10 mg/ml)

2 intracutaneously. After a 60-min equilibration period, the outgoing perfusate was stored at -

3 80°C for subsequent analysis.

4 Human monocytes, isolated as described above, were cultured 24 hours in 96 well plates in

5 DMEM with 2 mM glutamine and 10% heat inactivated FBS, containing E2 1nM 1µM ±

6 tamoxifen or 1µM ± fulvestrant. Culture media was analyzing for ENA78, FGF basic, G-CSF,

7 GM-CSF, INF, IL-1, IL-1, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-17, CCL2, CCL3,

8 CCL4, CCL5, TNF using Magnetic Luminex High Performance Assay Kit (R&D Systems

9 Cat# LUHM000) analyzed with a multiplex bead reader FlexMap 3D (Luminex Corporation,

10 Austin, US).

11 Olink Proteomics

12 Samples were submitted on 96-well PCR plates to Olink Proteomics (Uppsala, Sweden) for

13 analyses. In brief, 1 μl of undiluted sample was incubated with proximity antibody pairs

14 tagged with DNA reporter molecules. After binding of the antibodies to their corresponding

15 antigens, the respective DNA tails formed an amplicon by proximity extension, which was

16 quantified by high-throughput real-time PCR (BioMark™ HD System; Fluidigm Corporation,

17 San Francisco, CA, USA). The generated fluorescent signal directly correlated with protein

18 abundance. The output from the Proseek Multiplex protocol was correlated in quantitation

19 cycles (Cq) produced by the BioMark Real-Time PCR Software. To minimize variation

20 within and between runs, the data were normalized using both an internal control (extension

21 control) and an interplate control, and transformed using a predetermined correction factor.

22 The pre-processed data were provided in the arbitrary unit, normalized protein expression

NPX 23 (NPX), on a log2 scale, which then were linearized using the formula 2 . A high NPX value

24 corresponded to a high protein concentration. Values represented a relative quantification,

25 which means that no comparison of absolute levels between different proteins could be made.

1 Statistics

2 The Wilcoxon signed-rank test was used to compare paired observations. The Student’s t-test

3 was used for quantitative data. For all tests, P<0.05 was considered significant. The

4 proteomics data was analyzed using the Student’s t-test followed by the Benjamini, Krieger,

5 and Yekutieli procedure to assess the false discovery rate (FDR), which was set to 5%. Only

6 proteins above the limit of detection (LOD) in ≥50% of the samples were included in the

7 analyses. Data were expressed as the mean ± standard deviation (SD). All analyses were

8 performed using Prism 7.0 (GraphPad Software, San Diego, CA, USA).

1 Results

2 Fulvestrant inhibited the growth of experimental ER+ breast cancer in vivo more

3 effectively than tamoxifen

4 ER+ breast cancers were established orthotopically in the mammary fat pad in

5 oophorectomized nude mice. Physiological levels of estradiol were maintained in all animals.

6 At similar tumor sizes, treatment with either tamoxifen or fulvestrant was initiated in presence

7 of estradiol and continued for 24 days. As demonstrated in Fig. 1A, fulvestrant treatment

8 resulted in significantly decreased tumor growth compared to tamoxifen as measured both

9 during tumor growth in vivo by tumor area and at the end of treatment by tumor weight. Both

10 anti-estrogen treatments decreased cell proliferation as compared to untreated tumors, and

11 fulvestrant exhibited significantly decreased proliferation as compared to tamoxifen (Fig. 1B).

12 As expected, fulvestrant induced downregulation of ER in the tumor cells (Fig. 1C).

13 Molecular characterization of the extracellular microenvironment in vivo in experimental

14 ER+ breast cancer during anti-estrogen therapy

15 Next, we wanted to examine whether the treatments differentially affected the release of

16 proteins into the extracellular space, thereby altering the intercellular cross-talk. Tumors from

17 the different treatment groups were subjected to microdialysis to sample the extracellular

18 proteins in vivo. In total, 234 individual human proteins (listed in Supplementary Table 1)

19 were detected in the microdialysates of the breast cancer tumors from the mice. After FDR

20 correction, we observed significant alterations in 38 proteins from the ER+ tumors of mice

21 treated with fulvestrant; of these, 4 were significantly upregulated and 34 were significantly

22 downregulated (Fig. 2A). In ER+ tumors from tamoxifen-treated animals, 12 proteins were

23 significantly downregulated and 1 protein was significantly upregulated (Fig. 2A). All 13

24 proteins that were significantly changed in the tamoxifen-treated group also were significantly

25 altered in in the fulvestrant-treated group (Fig. 2B). Thus, 25 proteins were significantly

1 changed in the fulvestrant-treated group that were not changed in the tamoxifen-treated group,

2 whereas no proteins were changed only in the tamoxifen group (Fig. 2B).

3 The majority of proteins released from human cancer cells were associated with

4 immune function. Therefore, we next determined whether each treatment affected

5 inflammatory proteins released by stroma cells of murine origin. A total of nine proteins

6 could be detected in all samples. As shown in Fig. 2C, fulvestrant significantly increased the

7 levels of murine IL-6 and murine IL-10 and decreased the levels of murine CCL2, whereas

8 murine CCL5 was unaffected. No differences were detected in murine TGF-, CCL3,

9 CCL20, CXCL1, or CXCL9.

10 To determine whether the hormone treatments could affect the secretome of macrophages per

11 se we set up cultures of freshly isolated human monocytes and exposed them to estradiol

12 alone and in combination with tamoxifen or fulvestrant. Indeed, we found that fulvestrant, in

13 presence of estradiol, significantly decreased the levels of ENA78, IL-, IL-8, CCL2, CCL3,

14 CCL4, and TNFFig. 2D). Tamoxifen added to estradiol did not change the levels of any of

15 the detected cytokines. FGF basic, G-CSF, GM-CSF, INF, IL-1, IL-2, IL-4, IL-5, IL-6, IL-

16 10, IL-17 were below the detection limit of the assay.

17 Fulvestrant decreased the infiltration of innate immune cells and angiogenesis more

18 potently than tamoxifen in experimental ER+ breast cancer

19 Several of the proteins that were selectively changed in the fulvestrant group were associated

20 with immune function and angiogenesis. Therefore, we investigated whether there were any

21 differences between the treatment groups in the number of infiltrating innate immune cells

22 into the tumors or vessel area. Neutrophils were predominantly found in the invasive margin

23 of the tumors, but almost no neutrophils were detected within the tumor tissue, as previously

24 described (15). Therefore, we quantified the number of neutrophils in the invasive margin.

25 Both treatments resulted in a decreased number of neutrophils compared to the control group

1 (Fig. 3A). In addition, fulvestrant significantly reduced the number of neutrophils compared

2 to tamoxifen (Fig. 3A).

3 In contrast to neutrophils, macrophages were detected within the tumor stroma.

4 Therefore, we quantified the percentage of area that was positively stained for the

5 macrophage marker F4/80. Both treatments resulted in a significantly decreased area of

6 infiltrating macrophages compared to the control group (Fig. 3B). Fulvestrant also

7 significantly decreased the area of infiltrating macrophages compared to tamoxifen (Fig. 3B).

8 Angiogenesis was measured as the percent vessel area. Both treatments resulted

9 in decreased angiogenesis compared to the control group (Fig. 3C). In addition, fulvestrant

10 was more efficient in decreasing angiogenesis compared to tamoxifen (Fig. 3C). No

11 differences were found in the amount of tumor associated fibroblasts between the different

12 treatments (Fig. 3D).

13 Changes in EMT markers after fulvestrant and tamoxifen treatment of experimental ER+

14 breast cancer

15 The cross-talk between cancer cells and immune cells has been shown to affect the EMT of

16 cancer cells (29). We investigated whether the reduced number of innate immune cells in the

17 tumors of tamoxifen- or fulvestrant-treated mice resulted in changes in EMT markers. During

18 the EMT, E-cadherin is downregulated and N-cadherin is upregulated. E-cadherin was

19 expressed in all tumors with no detectable changes between the three groups (Fig. 4A). A

20 stronger expression of N-cadherin was detected across all groups, and both tamoxifen and

21 fulvestrant induced a significant downregulation of N-cadherin compared to tumors in the

22 control group (Fig. 4B). In addition, fulvestrant significantly decreased N-cadherin compared

23 to tamoxifen (Fig. 4B). The transcription factor Snail, which promote mesenchymal

24 transition, were significantly downregulated by both treatments compared to control tumors

25 (Fig. 4C). The transcription factor Slug was also downregulated by both treatments and

1 significantly decreased in the fulvestrant treated tumors compared to the tamoxifen group

2 (Fig. 4D). The mesenchymal protein vimentin was not detectable in any tumors.

3 Fulvestrant decreased the infiltration of innate immune cells more potently than tamoxifen

4 in experimental ER+ breast cancer in immunocompetent mice

5 The advantages of using the nude mouse model of cancer growth is that human cancer cells

6 can be investigated and as the stroma is of murine origin it is possible to distinguish events

7 originating from the cancer cells (human proteins) or from the stroma (murine proteins). By

8 using the nude Balb/C-nu/nu mouse, investigations of the innate immune system are possible

9 as these mice have an intact innate immunity, including B cells and NK cells, but lack T cells.

10 However, to further understand the effects on innate immune cells in an immunocompetent

11 host we next set up mice with ER+ breast cancer from the PyMT transgenic mouse model.

12 Fulvestrant reduced tumor growth significantly compared to tamoxifen (Fig. 5A). At the end

13 of the experiment tumor weights were; 508 mg±43 (SD) in the E2 alone group, 460±112 in

14 the E2+Tam group, and 310±42 in the E2+Fulv group leading to E2 vs. E2+Tam, not

15 significant, E+Tam vs. E2+Fulv, P<0.01, and E2 vs. E2+Fulv, P<0.001 n=7 in each group.

16 Additionally, similar to the results in nude mice, fulvestrant significantly decreased the

17 number of neutrophils in the invasive margin of the tumors and infiltration of macrophages

18 into the tumor tissue (Fig. 5B-C).

19 Fulvestrant decreased ER+ cancer cell dissemination more potently than tamoxifen

20 As metastasis is the major cause of death in breast cancer patients we next investigated if

21 there were any differences between the treatments of the capacity of cancer cells to

22 disseminate. In the zebrafish model of cancer cell metastases, fulvestrant inhibited estradiol

23 dependent dissemination of ER+ breast cancer cells in presence of neutrophils and monocytes

24 more effectively than tamoxifen, (Fig. 5 D-E).

25 Upregulation of proteins in human ER+ breast cancer in vivo

1 To elucidate whether human breast cancer exhibit increased infiltration of innate immune

2 cells and if the affected proteins in experimental ER+ breast cancer were present―and

3 therefore druggable―in human breast cancer, we stained human breast cancers for CD68 and

4 performed microdialysis in women with ER+ breast cancer prior to their surgery. Human

5 breast cancers exhibited increased numbers of innate immune cells compared to normal

6 adjacent breast tissue (Fig. 6A). Of the 38 proteins that were significantly up- or

7 downregulated in experimental breast cancer samples, 36 were detectable in the human

8 samples; only UMOD (uromodulin) and IL-20A were not detectable. Of the 36 detected

9 proteins, 31 were significantly up-regulated in ER+ breast cancers compared to normal

10 adjacent breast tissue (Figs. 6B and 7A-C).

1 Discussion

2 Today, approximately 25% of patients with ER+ breast cancer will have a recurrence of their

3 disease despite adjuvant anti-estrogen therapies. Thus, there is a need to potentiate ER

4 targeting therapies against breast cancer. Previous studies have demonstrated that events in

5 the stroma are determinants for the efficiency of cancer therapies. How different ER targeting

6 therapies may affect the tumor stroma has not yet been elucidated. In our current study, the

7 ER antagonist fulvestrant was identified as a more efficient therapy against ER+ breast cancer

8 than the SERM tamoxifen in presence of physiological levels of estradiol. In addition to

9 decreased growth and proliferation of ER+ breast cancer tumors in mice, fulvestrant induced

10 profound effects in the tumor stroma by significantly reducing the infiltration of macrophages

11 and neutrophils as compared to tamoxifen in both immunodeficient and immunocompetent

12 mice. Additionally, in presence of estradiol, fulvestrant was more effective than tamoxifen in

13 inhibiting cancer cell dissemination mediated by neutrophils and monocytes. In vivo sampling

14 of 234 extracellular proteins identified 38 proteins related to inflammation that were

15 significantly altered after fulvestrant and tamoxifen therapy; 25 of these proteins were

16 significantly changed in fulvestrant-treated animals only. Of these 38 proteins, 36 were

17 detected in ER+ human breast cancer in vivo and 31 of these were upregulated. Thus, the

18 proteins that were targeted by fulvestrant in the animal model were also upregulated in

19 humans. This suggests clinical relevance of the identified immune mediated mechanism of

20 action by anti-estrogen therapy.

21 The intercellular cross-talk between cancer cells and stroma cells in the tissue

22 microenvironment is a key determinant for carcinogenesis and the efficacy of cancer therapy

23 (12,13). Innate immune cells such as macrophages and neutrophils constitute a major

24 component of the stroma in human breast cancer (30,31). All immune cells express ERs, and

25 estrogen governs several signaling pathways within these cells (32). Whether estrogen

1 exposure results in a pro- or anti-inflammatory state is dependent on hormone levels. At

2 normal physiological levels, as observed during the menstrual cycle, estradiol promotes the

3 production of pro-inflammatory cytokines. In contrast, at very high levels (e.g., as observed

4 during pregnancy), estradiol promotes an anti-inflammatory response (32). In the present

5 study, we administered menstrual cycle-related physiological levels of estradiol, thus

6 reflecting premenopausal breast cancer conditions. Fulvestrant has previously been shown to

7 be more effective than tamoxifen after estrogen withdrawal (33). Our data suggest that

8 fulvestrant also exerts increased efficacy compared to tamoxifen in the presence of estradiol,

9 suggesting that fulvestrant would also be effective in premenopausal women. Based on the

10 types of extracellular proteins released by the cancer cells as identified during microdialysis,

11 the major difference between fulvestrant and tamoxifen therapy may be in the mediation of

12 the inflammatory response and angiogenesis. Several inflammatory proteins were

13 significantly affected by the two treatments. Whereas both treatments increased IL-1RA,

14 which is anti-tumorigenic, (16,34,35), fulvestrant but not tamoxifen significantly decreased

15 IL-2RA levels. As an inhibitor of IL-2, IL-2RA has been associated with cancer progression

16 (36). The immune regulatory proteins ITGAM (Integrin Alpha M), LILBRs (leukocyte

17 immunoglobulin-like receptor subfamily B), TIMD4 (T-cell immunoglobulin and mucin

18 domain containing 4), and CCL18, which were more effectively decreased by fulvestrant

19 compared to tamoxifen, have been shown to be up-regulated in cancer and associated with

20 enhanced tumor growth and poor survival (37-40). These inflammatory proteins may affect

21 macrophage and neutrophil infiltration into tumors. Reducing the number of infiltrating innate

22 immune cells into ER+ breast cancer may be critical, as it has been shown previously that

23 these cells may turn non-metastatic ER+ breast cancer cells into metastatic cells (14,15).

24 Additionally, innate immune cells may be activated by various cues in the tumor

25 microenvironment into tumor-promoting or anti-tumorigenic phenotypes depending on the

1 type and amounts of secreted factors. Our data show that while fulvestrant decreased the

2 number of innate immune cells, the treatment also increased stroma-derived inflammatory

3 cytokines. This may be explained by phenotypic changes of the immune cells by fulvestrant

4 as the different treatments had no effect on tumor associated fibroblasts. These results were

5 corroborated in vitro where fulvestrant affected the secreted levels of several cytokines

6 without affecting the number of cells. Inflammation and angiogenesis are inter-twined

7 processes and several proteins regulate both of these processes. Two of the most potent

8 proteins in these events are VEGF and IL-8, which were affected by both treatments in line

9 with previous published data (17,18,41,42). Fulvestrant but not tamoxifen also affected

10 additional potent pro-angiogenic proteins; ANGPTL3 (angiopoietin-related protein 3) and

11 AP-N (aminopeptidase N), both of which have been implicated in breast cancer progression,

12 were significantly decreased (43,44). In line with the decreased levels of pro-angiogenic

13 proteins, fulvestrant decreased tumoral angiogenesis more potently than tamoxifen. The

14 clinical relevance of our experimental findings was confirmed in our clinical human samples

15 from ER+ breast cancer, where the majority of the affected proteins also were upregulated.

16 Fulvestrant also exerted increased effects in the tumor stroma as murine IL-6

17 and murine IL-10, which have been linked with a good prognosis of breast cancer, were

18 significantly increased compared to tamoxifen, whereas murine CCL2 was similarly

19 decreased by both fulvestrant and tamoxifen (45). These results support that stromal cells

20 expressing ERs respond to anti-estrogen therapy and are consistent with prior studies

21 (14,15,46).

22 In addition to the inflammatory mediators, a number of growth factors and

23 proteins involved in metabolism also were affected by the different treatments. In particular,

24 some growth factors and proteins responded to fulvestrant only including the growth factors

25 TNXB (tenacin-X), PLTP (phospholipid transfer protein), SERPINA7 (thyroxine-binding

1 globulin) and the tumors suppressor proteins AXIN1 and Dkk-1 (Dickkopf-related protein 1)

2 corroborating previous data showing that the downregulation of AXIN1 by estrogen plays an

3 important role in ER+ breast cancer (47).

4 Metastatic disease is the major cause of breast cancer-associated mortality

5 including ER+ breast cancer. The majority of ER-expressing primary breast cancers maintain

6 the ER at the metastatic site and almost a third of initially ER negative primary tumors gain

7 ER expression in the metastatic lesion (48). Thus, investigating metastatic mechanisms of

8 ER+ breast cancer is key for improved treatments. Hitherto, there are no experimental breast

9 cancer models in which ER+ primary breast cancer spontaneously metastasizes with

10 maintained ER expression. However, the zebrafish metastases model allows for such

11 investigations (49). Our present data clearly showed that fulvestrant inhibited ER+ primary

12 tumor growth as well as cancer cell dissemination more efficiently than tamoxifen.

13 Fulvestrant also affected EMT features of the cancer cells by down-regulation of N-cadherin

14 expression, which empower cells from primary tumors to become metastatic. Interestingly,

15 similar results have previously been shown in estrogen-dependent lung cancer growth (46).

16 The EMT program is activated by several pathways, among which signaling between cancer

17 cells and the microenvironment in general and immune cells in particular play important roles

18 (50). Thus, therapeutic effects directly on tumor cells as well as indirect effects via infiltrated

19 immune cells, may result in changes of the EMT program.

20 Targeting the ER in ER+ breast cancer is the current gold standard for both

21 adjuvant and metastatic treatment. We have shown here that direct targeting of the ER with

22 the pure anti-estrogen fulvestrant is a more efficient approach in treating ER+ breast cancer

23 than direct targeting with the partial agonist tamoxifen in presence of estradiol. Our findings

24 suggest that the effectiveness of fulvestrant is due to both direct effects on the cancer cells and

25 on profound effects on the tumor stroma.

1 Currently, fulvestrant is the only approved SERD in clinical practice. Due to its

2 low bioavailability, the only administrative route for therapy is intramuscular injections,

3 which limits its clinical use. Several early phase studies are currently ongoing to test oral

4 SERDs compared to fulvestrant in patients with metastatic cancer (https://clinicaltrials.gov).

5 If oral SERDs are as efficient as fulvestrant, then more patients may be able to benefit from

6 this therapy. Moreover, as fulvestrant had major effects on intercellular cross-talk and

7 immune function, combinations of fulvestrant and immune-mediated therapies may be

8 interesting to explore. Finally, although fulvestrant is used currently only in postmenopausal

9 patients, our data support that fulvestrant may be effective in premenopausal patients.

1 References

2 1. Grann VR, Troxel AB, Zojwalla NJ, Jacobson JS, Hershman D, Neugut AI. Hormone 3 receptor status and survival in a population-based cohort of patients with breast 4 carcinoma. Cancer 2005;103:2241-51 5 2. Goldhirsch A, Ingle JN, Gelber RD, Coates AS, Thurlimann B, Senn HJ. Thresholds for 6 therapies: highlights of the St Gallen International Expert Consensus on the primary 7 therapy of early breast cancer 2009. Ann Oncol 2009;20:1319-29 8 3. Lewis JS, Jordan VC. Selective estrogen receptor modulators (SERMs): mechanisms of 9 anticarcinogenesis and drug resistance. Mutat Res 2005;591:247-63 10 4. Jordan VC. Antiestrogens and selective estrogen receptor modulators as 11 multifunctional medicines. 2. Clinical considerations and new agents. Journal of 12 medicinal chemistry 2003;46:1081-111 13 5. Jordan VC. Antiestrogens and selective estrogen receptor modulators as 14 multifunctional medicines. 1. Receptor interactions. Journal of medicinal chemistry 15 2003;46:883-908 16 6. Burstein HJ, Curigliano G, Loibl S, Dubsky P, Gnant M, Poortmans P, et al. Estimating 17 the benefits of therapy for early-stage breast cancer: the St. Gallen International 18 Consensus Guidelines for the primary therapy of early breast cancer 2019. Ann Oncol 19 2019;30:1541-57 20 7. Ruhstaller T, Giobbie-Hurder A, Colleoni M, Jensen MB, Ejlertsen B, de Azambuja E, et 21 al. Adjuvant Letrozole and Tamoxifen Alone or Sequentially for Postmenopausal 22 Women With Hormone Receptor-Positive Breast Cancer: Long-Term Follow-Up of the 23 BIG 1-98 Trial. J Clin Oncol 2019;37:105-14 24 8. Robertson JF, Come SE, Jones SE, Beex L, Kaufmann M, Makris A, et al. Endocrine 25 treatment options for advanced breast cancer--the role of fulvestrant. Eur J Cancer 26 2005;41:346-56 27 9. Clarke CA, Keegan TH, Yang J, Press DJ, Kurian AW, Patel AH, et al. Age-specific 28 incidence of breast cancer subtypes: understanding the black-white crossover. J Natl 29 Cancer Inst 2012;104:1094-101 30 10. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 31 2011;144:646-74 32 11. Busch S, Ryden L, Stal O, Jirstrom K, Landberg G. Low ERK phosphorylation in cancer- 33 associated fibroblasts is associated with tamoxifen resistance in pre-menopausal 34 breast cancer. PLoS One 2012;7:e45669 35 12. Farmer P, Bonnefoi H, Anderle P, Cameron D, Wirapati P, Becette V, et al. A stroma- 36 related gene signature predicts resistance to neoadjuvant chemotherapy in breast 37 cancer. Nat Med 2009;15:68-74 38 13. Hattar R, Maller O, McDaniel S, Hansen KC, Hedman KJ, Lyons TR, et al. Tamoxifen 39 induces pleiotrophic changes in mammary stroma resulting in extracellular matrix 40 that suppresses transformed phenotypes. Breast Cancer Res 2009;11:R5 41 14. Svensson S, Abrahamsson A, Rodriguez GV, Olsson AK, Jensen L, Cao Y, et al. CCL2 42 and CCL5 Are Novel Therapeutic Targets for Estrogen-Dependent Breast Cancer. Clin 43 Cancer Res 2015;21:3794-805 44 15. Vazquez Rodriguez G, Abrahamsson A, Jensen LD, Dabrosin C. Estradiol Promotes 45 Breast Cancer Cell Migration via Recruitment and Activation of Neutrophils. Cancer 46 Immunol Res 2017;5:234-47

1 16. Lindahl G, Saarinen N, Abrahamsson A, Dabrosin C. Tamoxifen, flaxseed, and the 2 lignan enterolactone increase stroma- and cancer cell-derived IL-1Ra and decrease 3 tumor angiogenesis in estrogen-dependent breast cancer. Cancer Res 2011;71:51-60 4 17. Garvin S, Dabrosin C. Tamoxifen inhibits secretion of vascular endothelial growth 5 factor in breast cancer in vivo. Cancer Res 2003;63:8742-8 6 18. Aberg UW, Saarinen N, Abrahamsson A, Nurmi T, Engblom S, Dabrosin C. Tamoxifen 7 and flaxseed alter angiogenesis regulators in normal human breast tissue in vivo. 8 PLoS One 2011;6:e25720 9 19. Mani SA, Guo W, Liao MJ, Eaton EN, Ayyanan A, Zhou AY, et al. The epithelial- 10 mesenchymal transition generates cells with properties of stem cells. Cell 11 2008;133:704-15 12 20. Polyak K, Weinberg RA. Transitions between epithelial and mesenchymal states: 13 acquisition of malignant and stem cell traits. Nat Rev Cancer 2009;9:265-73 14 21. Ricciardi M, Zanotto M, Malpeli G, Bassi G, Perbellini O, Chilosi M, et al. Epithelial-to- 15 mesenchymal transition (EMT) induced by inflammatory priming elicits mesenchymal 16 stromal cell-like immune-modulatory properties in cancer cells. Br J Cancer 17 2015;112:1067-75 18 22. Yang J, Weinberg RA. Epithelial-mesenchymal transition: at the crossroads of 19 development and tumor metastasis. Developmental cell 2008;14:818-29 20 23. Toh B, Wang X, Keeble J, Sim WJ, Khoo K, Wong WC, et al. Mesenchymal transition 21 and dissemination of cancer cells is driven by myeloid-derived suppressor cells 22 infiltrating the primary tumor. PLoS Biol 2011;9:e1001162 23 24. Sangaletti S, Tripodo C, Santangelo A, Castioni N, Portararo P, Gulino A, et al. 24 Mesenchymal Transition of High-Grade Breast Carcinomas Depends on Extracellular 25 Matrix Control of Myeloid Suppressor Cell Activity. Cell Rep 2016;17:233-48 26 25. Dabrosin C, Margetts PJ, Gauldie J. Estradiol increases extracellular levels of vascular 27 endothelial growth factor in vivo in murine mammary cancer. Int J Cancer 28 2003;107:535-40 29 26. Guy CT, Cardiff RD, Muller WJ. Induction of mammary tumors by expression of 30 polyomavirus middle T oncogene: a transgenic mouse model for metastatic disease. 31 Mol Cell Biol 1992;12:954-61. 32 27. Lin EY, Jones JG, Li P, Zhu L, Whitney KD, Muller WJ, et al. Progression to malignancy 33 in the polyoma middle T oncoprotein mouse breast cancer model provides a reliable 34 model for human diseases. Am J Pathol 2003;163:2113-26 35 28. Dabrosin C, Palmer K, Muller WJ, Gauldie J. Estradiol promotes growth and 36 angiogenesis in polyoma middle T transgenic mouse mammary tumor explants. 37 Breast Cancer Res Treat 2003;78:1-6 38 29. Marcucci F, Bellone M, Caserta CA, Corti A. Pushing tumor cells towards a malignant 39 phenotype: stimuli from the microenvironment, intercellular communications and 40 alternative roads. Int J Cancer 2014;135:1265-76 41 30. Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature 42 2008;454:436-44 43 31. Whiteside TL. The tumor microenvironment and its role in promoting tumor growth. 44 Oncogene 2008;27:5904-12 45 32. Kovats S. Estrogen receptors regulate innate immune cells and signaling pathways. 46 Cell Immunol 2015;294:63-9

1 33. Osborne CK, Coronado-Heinsohn EB, Hilsenbeck SG, McCue BL, Wakeling AE, 2 McClelland RA, et al. Comparison of the effects of a pure steroidal antiestrogen with 3 those of tamoxifen in a model of human breast cancer. J Natl Cancer Inst 4 1995;87:746-50 5 34. Bar D, Apte RN, Voronov E, Dinarello CA, Cohen S. A continuous delivery system of IL- 6 1 receptor antagonist reduces angiogenesis and inhibits tumor development. Faseb J 7 2004;18:161-3 8 35. Elaraj DM, Weinreich DM, Varghese S, Puhlmann M, Hewitt SM, Carroll NM, et al. 9 The role of interleukin 1 in growth and metastasis of human cancer xenografts. Clin 10 Cancer Res 2006;12:1088-96 11 36. Jiang T, Zhou C, Ren S. Role of IL-2 in cancer immunotherapy. Oncoimmunology 12 2016;5:e1163462 13 37. Kang X, Kim J, Deng M, John S, Chen H, Wu G, et al. Inhibitory leukocyte 14 immunoglobulin-like receptors: Immune checkpoint proteins and tumor sustaining 15 factors. Cell Cycle 2016;15:25-40 16 38. Gabrilovich DI, Nagaraj S. Myeloid-derived suppressor cells as regulators of the 17 immune system. Nature reviews Immunology 2009;9:162-74 18 39. Tan X, Zhang Z, Yao H, Shen L. Tim-4 promotes the growth of colorectal cancer by 19 activating angiogenesis and recruiting tumor-associated macrophages via the 20 PI3K/AKT/mTOR signaling pathway. Cancer Lett 2018;436:119-28 21 40. Su S, Liu Q, Chen J, Chen J, Chen F, He C, et al. A positive feedback loop between 22 mesenchymal-like cancer cells and macrophages is essential to breast cancer 23 metastasis. Cancer Cell 2014;25:605-20 24 41. Bendrik C, Dabrosin C. Estradiol increases IL-8 secretion of normal human breast 25 tissue and breast cancer in vivo. J Immunol 2009;182:371-8 26 42. Dabrosin C. Variability of Vascular Endothelial Growth Factor in Normal Human 27 Breast Tissue in Vivo during the Menstrual Cycle. J Clin Endocrinol Metab 28 2003;88:2695-8 29 43. Carbone C, Piro G, Merz V, Simionato F, Santoro R, Zecchetto C, et al. Angiopoietin- 30 Like Proteins in Angiogenesis, Inflammation and Cancer. Int J Mol Sci 2018;19 31 44. Wickstrom M, Larsson R, Nygren P, Gullbo J. Aminopeptidase N (CD13) as a target for 32 cancer chemotherapy. Cancer Sci 2011;102:501-8 33 45. Ahmad N, Ammar A, Storr SJ, Green AR, Rakha E, Ellis IO, et al. IL-6 and IL-10 are 34 associated with good prognosis in early stage invasive breast cancer patients. Cancer 35 Immunol Immunother 2018;67:537-49 36 46. Hamilton DH, Griner LM, Keller JM, Hu X, Southall N, Marugan J, et al. Targeting 37 Estrogen Receptor Signaling with Fulvestrant Enhances Immune and Chemotherapy- 38 Mediated Cytotoxicity of Human Lung Cancer. Clin Cancer Res 2016;22:6204-16 39 47. Chimge NO, Little GH, Baniwal SK, Adisetiyo H, Xie Y, Zhang T, et al. RUNX1 prevents 40 oestrogen-mediated AXIN1 suppression and beta-catenin activation in ER-positive 41 breast cancer. Nature communications 2016;7:10751 42 48. Lindstrom LS, Karlsson E, Wilking UM, Johansson U, Hartman J, Lidbrink EK, et al. 43 Clinically used breast cancer markers such as estrogen receptor, progesterone 44 receptor, and human epidermal growth factor receptor 2 are unstable throughout 45 tumor progression. J Clin Oncol 2012;30:2601-8 46 49. Rouhi P, Jensen LD, Cao Z, Hosaka K, Lanne T, Wahlberg E, et al. Hypoxia-induced 47 metastasis model in embryonic zebrafish. Nature protocols 2010;5:1911-8

1 50. Chaffer CL, Weinberg RA. A perspective on cancer cell metastasis. Science 2 2011;331:1559-64 3

1 Figure Legends

2 Figure 1. Fulvestrant inhibited the growth of experimental estrogen receptor positive (ER+)

3 breast cancer more effectively than tamoxifen in vivo.

4 Oophorectomized athymic mice supplemented with physiological levels of estradiol (E2)

5 were injected with MCF-7 into the dorsal mammary fat pads. At similar tumor sizes, mice

6 either continued with E2 or were additionally treated with tamoxifen (E2+Tam) (1 mg/mouse

7 every second day, subcutaneously [s.c.]) or fulvestrant (E2+Fulv) (5 mg/mouse every 3 days,

8 s.c.).

9 A. Tumor growth. **P<0.01; ***P<0.001 vs. control (E2) and ##P<0.01 vs. E2+Tam. Tumor

10 weight at the end of the experiment. **P<0.01, ***P<0.001. Data are presented as the mean ±

11 standard deviation (SD).

12 B. Tumor sections from each treatment group were stained for proliferation (Ki67) and

13 quantified as the percentage of area with positive staining in the different treatment groups.

14 *P<0.05, **P<0.01, ***P<0.001. Data presented as the mean ± SD.

15 C. Tumor sections from each treatment group were stained for ERα and quantified as the

16 percentage of cells with positive staining in the different treatment groups. ***P<0.001. Data

17 presented as the mean ± SD.

19 Figure 2. Molecular characterization of the extracellular microenvironment in vivo in

20 experimental estrogen receptor positive (ER+) breast cancer during anti-estrogen therapy.

21 Oophorectomized athymic mice supplemented with physiological levels of estradiol (E2)

22 were injected with MCF-7 into the dorsal mammary fat pads. At similar tumor sizes, mice

23 either continued with E2 or were additionally treated with tamoxifen (E2+Tam) (1 mg/mouse

24 every second day, subcutaneously [s.c.]) or fulvestrant (E2+Fulv) (5 mg/mouse every 3 days,

1 s.c.). Size-matched tumors from the different treatment groups underwent microdialysis for

2 sampling of extracellular proteins in vivo.

3 A. Volcano plots illustrating the log10 statistical significance (false discovery rate [FDR]

4 adjusted p-value) in relation to the log2 fold change of 234 human proteins released by the

5 cancer cells treated with E2+Tam or E2+Fulv vs. control (E2). Proteins that passed the FDR-

6 adjusted P<0.05 significance threshold and were downregulated (log2 fold change <1.0) are

7 marked in blue. Proteins that passed the FDR-adjusted P<0.05 significance threshold and

8 were upregulated (log2 fold change >1.0) are marked in red.

9 B. A Venn diagram illustrating significantly altered proteins after FDR correction for

10 E2+Tam or E2+Fulv vs. control (E2). Proteins that were downregulated are marked in blue

11 and proteins that were upregulated are marked in red.

12 C. Murine immune-regulating proteins released by the stroma. *P<0.05, **P<0.01,

13 ***P<0.001. Data are presented as the mean ± standard deviation (SD).

14 D. Freshly isolated human monocytes were cultured for 24 hours in presence of estradiol

15 (E2), E2 + tamoxifen (Tam) or E2 + fulvestrant (Fulv) and secreted cytokines were quantified

16 as described in the materials and methods section. *P<0.05, **P<0.01. Data are presented as

17 the mean ± standard deviation (SD).

20 Figure 3. Fulvestrant reduced innate immune cells in experimental estrogen receptor

21 positive (ER+) breast cancer more effectively than tamoxifen in vivo.

22 Oophorectomized athymic mice supplemented with physiological levels of estradiol (E2),

23 were injected with MCF-7 into the dorsal mammary fat pads. At similar tumor sizes, mice

24 either continued with E2 or were additionally treated for 24 days with tamoxifen (E2+Tam) (1

1 mg/mouse every second day, subcutaneously [s.c.]) or fulvestrant (E2+Fulv) (5 mg/mouse

2 every 3 days, s.c.).

3 A. Ten hot spot areas of each tumor were selected to quantify the number of neutrophils

4 (Ly6G; red) infiltrating the invasive margin; scale bar = 20 µm, *P<0.05, **P<0.01,

5 ***P<0.001. Data are presented as the mean ± standard deviation (SD).

6 B. Tumor sections stained for infiltrating macrophages were quantified as the percentage of

7 area with positive staining for the macrophage marker F4/80; scale bar = 20 µm, ***P<0.001.

8 Data are presented as the mean ± SD.

9 C. Ten hot spot areas of each tumor were selected for quantification of the vessel area stained

10 with von Willebrand factor; scale bar = 100 µm, *P<0.05, ***P<0.001. Data are presented as

11 the mean ± SD.

12 D. Ten hot spot areas of each tumor were selected for quantification of tumor associated

13 fibroblast stained with fibroblasts activation protein (FAP); scale bar = 20 µm. Data are

14 presented as the mean ± SD.

17 Figure 4. Epithelial-to-mesenchymal transition (EMT) expression in experimental estrogen

18 receptor positive (ER+) breast cancer treated with fulvestrant or tamoxifen.

19 Oophorectomized athymic mice supplemented with physiological levels of estradiol (E2)

20 were injected with MCF-7 into the dorsal mammary fat pads. At similar tumor sizes, mice

21 either continued with E2 or were additionally treated for 24 days with tamoxifen (E2+Tam) (1

22 mg/mouse every second day, subcutaneously [s.c.]) or fulvestrant (E2+Fulv) (5 mg/mouse

23 every 3 days, s.c.).

1 A. Tumor sections stained for E-cadherin were quantified as the percentage of area with

2 positive staining; scale bar = 20 µm. Data are presented as the mean ± standard deviation

3 (SD).

4 B. Tumor sections stained for the EMT marker N-cadherin were quantified as the percentage

5 of area with positive staining; scale bar = 20 µm, *P<0.05, ***P<0.001. Data are presented as

6 the mean ± SD

7 C. Tumor sections stained for the EMT marker Snail were quantified as the percentage of

8 cells with intense staining; scale bar = 20 µm, ***P<0.001. Data are presented as the mean ±

9 SD

10 D. Tumor sections stained for the EMT marker Slug were quantified as the percentage of cells

11 with intense staining; scale bar = 20 µm, *P<0.05, ***P<0.001. Data are presented as the

12 mean ± SD

15 Figure 5. Fulvestrant reduced innate immune cell infiltration in estrogen receptor positive

16 (ER+) breast cancer in immune competent mice and decreased ER+ breast cancer cell

17 dissemination in zebrafish

18 Oophorectomized FVB/N mice were supplemented with a physiological level of estradiol

19 (E2) and injected with PyMT tumor cells in the mammary fat pad and treated for 11 days with

20 tamoxifen (E2+Tam) (1 mg/mouse every second day, subcutaneously [s.c.]) or fulvestrant

21 (E2+Fulv) (5 mg/mouse every 3 days, s.c.) or left untreated.

22 A. Tumor growth. *P<0.05 **P<0.01; ###P<0.001 vs. E2. Data are presented as the mean ±

23 standard deviation (SD).

1 B. Ten hot spot areas of each tumor were selected to quantify the number of neutrophils

2 (Ly6G; red) infiltrating the invasive margin; scale bar = 20 µm, **P<0.01, ***P<0.001. Data

3 are presented as the mean ± standard deviation (SD).

4 C. Tumor sections stained for infiltrating macrophages were quantified as the percentage of

5 area with positive staining for the macrophage marker F4/80; scale bar = 20 µm, *P<0.05,

6 ***P<0.001. Data are presented as the mean ± SD.

7 D. Zebrafish embryos, with green blood vessels, were co-injected with MCF-7 + 50% human

8 neutrophils (Neu). All embryos were exposed to estradiol (E2) and treated with tamoxifen

9 (Tam) or fulvestrant (Fulv), as described in the materials and methods section. The number of

10 disseminated cells was counted 24 hours after injection. scale bar = 100 µm, (n=20-22 in each

11 group). *P<0.05, **P<0.01. Data are presented as mean ± SD.

12 E. Zebrafish embryos were co-injected with MCF-7 + 10% human monocytes (MQ). All

13 embryos were exposed to estradiol (E2) and treated with tamoxifen (Tam) or fulvestrant

14 (Fulv), as described in the materials and methods section. The number of disseminated cells

15 was counted 3 days after injection. scale bar = 100 µm, (n=20-22 in each group). *P<0.05,

16 **P<0.01. Data are presented as mean ± SD.

17 BV=Blood vessels. Arrowheads show disseminated cancer cells.

20 Figure 6. Extracellular levels of proteins with immune modulating functions in human

21 estrogen receptor positive (ER+) breast cancer in vivo.

22 A. Sections from breast cancers and normal adjacent breast tissue from patients were stained

23 with CD68, a marker for innate immune cells, and % area was quantified as described in the

24 materials and methods section, n=30 in each group (cancer and normal breast tissue

1 respectively), **P<0.01. Representative sections are depicted. Data are presented as the mean

2 ± standard deviation (SD), scale bar = 20 µm.

3 B. Patients with breast cancer underwent microdialysis one day prior to their surgery. One

4 catheter was inserted into the breast cancer (red bars) and another catheter was inserted into

5 normal adjacent breast tissue (white bars) for in vivo collection of extracellular proteins.

6 The multiplex proximity extension assay was used for protein quantification. The data

7 represent protein abundance in linear values (2NPX as described in Material and Methods);

8 **P<0.01, *P<0.05. Data are presented as the mean ± standard deviation (SD).

9 Figure 7. Extracellular levels of angiogenic factors, growth factors, and metabolic proteins

10 in human estrogen receptor positive (ER+) breast cancer in vivo.

11 Patients with breast cancer underwent microdialysis one day prior to their surgery. One

12 catheter was inserted into the breast cancer (red bars) and another catheter was inserted into

13 normal adjacent breast tissue (white bars) for in vivo collection of extracellular proteins.

14 The multiplex proximity extension assay was used for protein quantification. The data

15 represent protein abundance in linear values (2NPX as described in Material and Methods)

16 A. Extracellular levels of angiogenic factors.

17 B. Extracellular levels of growth factors.

18 C. Extracellular levels of metabolic proteins.

19 *P<0.05, **P<0.01. Data are presented as the mean ± standard deviation (SD).

Fulvestrant-mediated attenuation of the innate immune response decreases ER+ breast cancer growth in vivo more effectively than tamoxifen

Annelie Abrahamsson, Gabriela Vazquez Rodriguez and Charlotta Dabrosin

Cancer Res Published OnlineFirst August 27, 2020.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-20-1705

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2020/08/27/0008-5472.CAN-20-1705.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet Manuscript been edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2020/08/27/0008-5472.CAN-20-1705. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.