The Antidepressant Desipramine and A2-Adrenergic Receptor Activation Promote Breast Tumor Progression in Association with Altered Collagen Structure

Cancer Prevention Research Article Research

The Antidepressant Desipramine and a2-Adrenergic Receptor Activation Promote Breast Tumor Progression in Association with Altered Collagen Structure

Mercedes J. Szpunar1, Kathleen A. Burke2, Ryan P. Dawes3, Edward B. Brown2, and Kelley S. Madden2

Abstract Emotional stress activates the sympathetic nervous system (SNS) and release of the neurotransmitter norepinephrine to promote breast tumor pathogenesis. We demonstrate here that the metastatic mammary adenocarcinoma cell line 4T1 does not express functional adrenergic receptors (AR), the receptors activated by norepinephrine, yet stimulation of adrenergic receptor in vivo altered 4T1 tumor progression in vivo. Chronic treatment with the antidepressant desipramine (DMI) to inhibit norepinephrine reuptake

increased 4T1 tumor growth but not metastasis. Treatment with a highly selective a2-adrenergic receptor agonist, dexmedetomidine (DEX), increased tumor growth and metastasis. Neither isoproterenol (ISO), a

b-AR agonist, nor phenylephrine, an a1-AR agonist, altered tumor growth or metastasis. Neither DMI- nor DEX-induced tumor growth was associated with increased angiogenesis. In DMI-treated mice, tumor VEGF, IL-6, and the prometastatic chemokines RANTES, M-CSF, and MIP-2 were reduced. Tumor collagen microstructure was examined using second harmonic generation (SHG), a nonabsorptive optical scattering process to highlight fibrillar collagen. In DMI- and DEX-treated mice, but not ISO-treated mice, tumor SHG was significantly altered without changing fibrillar collagen content, as detected by immunofluorescence.

These results demonstrate that a2-AR activation can promote tumor progression in the absence of direct sympathetic input to breast tumor cells. The results also suggest that SNS activation may regulate tumor progression through alterations in the extracellular matrix, with outcome dependent on the combination of adrenergic receptor activated. These results underscore the complexities underlying SNS regulation of breast tumor pathogenesis, and suggest that the therapeutic use of adrenergic receptor blockers, tricyclic antidepressants, and adrenergic receptor agonists must be approached cautiously in patients with breast cancer. Cancer Prev Res; 6(12); 1262–72. 2013 AACR.

Introduction tumor angiogenesis and density of tumor associated macro- In patients with cancer, chronic emotional stress or other phages (7, 8). SNS activation can also target b-AR–expres- negative psychological factors such as depression or lack of sing host cells residing in the tumor or in metastatic sites social support promote tumor growth and progression to promote tumor growth and metastasis (6, 9). These (1, 2). The sympathetic nervous system (SNS) is an impor- studies provide compelling evidence that norepinephrine tant pathway by which stress can facilitate tumor growth and AR-expressing tumor cells or host stromal cells mod- (3–6). The SNS neurotransmitters norepinephrine and ulate tumor pathogenesis. epinephrine activate a- and b-adrenergic receptors (AR). Despite progress in understanding the molecular mechan- In animal models using b-AR–expressing cancer cell lines, isms underlying sympathetic regulation of tumor progres- stressor exposure or b-AR stimulation increased tumor sion, several critical questions remain. First, the role for a-AR growth and/or metastasis by mechanisms such as increased has not been carefully examined despite the fact that in human breast cancer, a-AR expression has been linked to poor prognosis (10). Second, variation in breast cancer cell Authors' Afﬁliations: Departments of 1Pathology and 2Biomedical Engi- line adrenergic receptor expression (11, 12) is recapitulated neering; and 3Neuroscience Graduate Program, School of Medicine and Dentistry, University of Rochester, Rochester, New York in human breast tumors that display heterogeneity in a-and b-AR expression (10). The functional consequences of such Note: Supplementary data for this article are available at Cancer Prevention Research Online (http://cancerprevres.aacrjournals.org/). heterogeneity have yet to be systematically explored. It is reasonable to assume that when breast cancer cells express Corresponding Author: Kelley S. Madden, Department of Biomedical Engineering, University of Rochester Medical Center, Goergen Hall, RC no or low levels of adrenergic receptor, host stromal adren- Box 270168, Rochester, NY 14627. Phone: 585-273-5724; Fax: 585-276- ergic receptor would be the direct targets of elevated norepi- 2254; E-mail: [email protected] nephrine. Stromal cells, including cells of the immune doi: 10.1158/1940-6207.CAPR-13-0079 system, endothelial cells, and ﬁbroblasts, express a-AR and 2013 American Association for Cancer Research. b-AR normally and in tumors (8, 13, 14). We propose that

1262 Cancer Prev Res; 6(12) December 2013

Downloaded from cancerpreventionresearch.aacrjournals.org on September 23, 2021. © 2013 American Association for Cancer Research. Norepinephrine and a2-Adrenergic Receptor Activation Increase Tumorigenesis

SNS activation can promote tumor pathogenesis by acting the University of Rochester Committee on Animal Res- on stromal cells to alter the tumor extracellular matrix. ources. The University of Rochester Animal Resource is fully To test this hypothesis, we have used multiphoton laser accredited by AAALAC International. scanning microscopy and second harmonic generation (SHG) to visualize a component of the tumor stroma, Cell lines and tissue culture fibrillar collagen. SHG is an endogenous optical signal 4T1 (mammary adenocarcinoma; CRL-2539), MDA-MB- produced when 2 excitation photons combine to produce 231 (MB-231; HTB-26), and human foreskin fibroblast-1 one emission photon, "catalyzed" by a non-centrosymmet- cells (HFF; SCRC-1041) were acquired from American Type ric structure, such as ordered collagen triple helices (15). Culture Collection (ATCC) within the last 2 years. ATCC Tumor collagen fiber microstructure, as revealed by SHG, is authenticates cell lines using short tandem repeat analyses. of great interest because several studies have suggested that Upon acquisition, cells were expanded for no more than it influences tumor progression, specifically tumor metas- 3 passages and frozen. All cell lines were used within 3 tasis. It is important to note that not all collagen fibers months of thawing. For passaging, 4T1 was grown in RPMI produce detectable SHG (16), and that tumor cells can containing penicillin/streptomycin and 10% fetal calf þ migrate toward blood vessels via SHG fibers, locomoting serum. MB-231 and HFF were grown in Dulbecco’s mod- along such fibers more efficiently than cells moving inde- ified Eagle medium containing L-glutamine and supple- pendently. Interestingly, the extent of SHG-associated mented with penicillin/streptomycin and 10% fetal calf tumor cell motility is correlated with metastatic ability serum. All media and media components were purchased (17, 18). In patients with breast cancer biopsies, we dem- from Gibco (Invitrogen Inc.). All cells were regularly tested onstrated shifts in SHG emission patterns associated with for mycoplasma contamination. progression to more metastatic disease (19), and lymph In vitro culture of 4T1 cells to measure proliferation and node metastasis was associated with increased breast tumor cytokine production was described previously (11). Pro- þ SHG collagen I density (20). Finally, SHG-based tumor- liferation was assessed using CyQuant NF Proliferation associated collagen signatures are prognostic factors for Assay kit (Invitrogen) following the manufacturer’s ins- disease-free survival, independent of tumor grade, size, and tructions. Fluorescence was detected with a 485 nm hormonal receptor status (21). Thus, SHG imaging repre- excitation filter and 530 nm emission filter using a multi- sents a novel imaging modality to assess the impact of SNS well plate reader (Biotek). For cytokines, cell-free super- activation on tumor extracellular matrix and explore stro- natants were harvested after 72 hours in culture and mal pathways that may be activated by norepinephrine to stored at 80C. influence tumor progression. We demonstrate here that 4T1, a metastatic mammary Drug treatment adenocarcinoma (22), lacks functional a- and b-AR, and is Drug treatments commenced 2 days before tumor cell unable to respond to norepinephrine in vitro. Therefore, we injection and continued for the duration of the experiment. propose 4T1 as an excellent model for investigating effects For DMI treatment, continuous release pellets (Innovative of SNS activation on tumors in the absence of direct sym- Research of America) were implanted subcutaneously pathetic input to the tumor cells themselves. To explore the under ketamine/xylazine anesthesia (90/9 mg/kg). The impact of elevated norepinephrine in the 4T1 tumor model, adrenergic receptor agonists ISO (Sigma-Aldrich), phenyl- mice were treated with the norepinephrine reuptake inhib- ephrine (Sigma-Aldrich), and DEX (Pfizer) were dissolved itor, desipramine (DMI), a drug used clinically as an anti- in sterile saline. Mice were injected intraperitoneally (i.p.) depressant (23, 24). We demonstrate that chronic treatment daily for the duration of the experiment. ISO and phenyl- with DMI increased orthotopic 4T1 tumor growth, whereas ephrine doses were chosen based on ISO-induced increased highly selective a2-AR activation by treatment with dexme- tumor pathogenesis reported in other murine tumor mod- detomidine (DEX) increased both tumor growth and metas- els (7, 8) and on pilot toxicity studies. Two doses of DEX tasis. DMI and DEX treatments were associated with altered were tested: 10 mg/kg elicited no apparent sedative effects SHG-emitting tumor collagen. These results provide evi- and 25 mg/kg elicited mild and transient anesthetic effects dence that activation of the SNS can promote tumor pro- (slowed movements following injection that were no longer gression and metastasis through interactions with the apparent 1 hour after injection). DMI, ISO, and phenyl- tumor stroma, and can do so through activation of a2-AR. ephrine treatment elicited early and transient decreases in body weight (10%) that recovered over time.

Materials and Methods Tumor implantation, growth, and tissue harvest Mice 4T1 cells (1 105 in sterile saline) were injected into the Female BALB/cByJ mice (6–8 weeks of age; The Jackson third mammary fat pad (MFP) under ketamine/xylazine Laboratories) were housed 3 to 4 per cage with food and anesthesia. Tumors were measured with calipers every 2 to 3 water ad libitum, and used experimentally 2 weeks after days without knowledge of experimental group. Mice were arrival. All animal procedures were carried out in accor- sacriﬁced by pentobarbital overdose (200 mg/kg, i.p.) fol- dance with the National Institutes of Health Guide for the lowed by cervical dislocation to harvest tumor, spleen, and Care and Use of Laboratory Animals and were approved by lungs. Tumors and spleens were weighed and divided. For

www.aacrjournals.org Cancer Prev Res; 6(12) December 2013 1263

catecholamine and cytokine analyses, tissue was immedi- conditions in each sample, including excitation wavelength ately placed on dry ice and stored at 80C. (810 nm), laser power (8 mW at the sample), and photo- Tumor volume (V) was calculated using the equation V ¼ multiplier tube voltages. To detect immunofluorescence, 1/2 length width2. Tumor growth is presented either as fluorophore emission was collected using bandpass filters the raw tumor volume over time or as normalized tumor 520/40 (for FITC) and 635/30 (for Alexa-Fluor 594). SHG growth. Normalized tumor growth was calculated by divid- signal was separated from fluorescence by a 475 nm long- ing an individual’s tumor volume at a given time point by its pass dichroic (Semrock) and detected through a bandpass volume at the earliest time point all tumors were detected 405/30 emission filter. (day 3 or 5 post-4T1). All images were analyzed by personnel blinded to group using custom algorithms in ImageJ (NIH Freeware). To Norepinephrine/normetanephrine and cytokine quantify CD31 and Ki67 immunofluorescence, a threshold analyses that excluded autofluorescence was determined in adjoin- For norepinephrine and normetanephrine, tissue was ing tissue sections stained with secondary antibody alone, homogenized in 0.01 M HCl at 10% volume (mL) by and the percentage of pixels above threshold was calculated. tissue weight. Norepinephrine and normetanephrine Average blood vessel area was calculated and normalized to were determined by ELISA (Rocky Mountain Diagnostics) cell density based on 40,6-diamino-2-phenylindole nuclear following the manufacturer’s instructions. For cytokines, staining. To quantify SHG and anti-collagen immunostain- tissue was homogenized in radioimmunoprecipitation ing, a common threshold was determined for all samples by assay buffer containing protease inhibitors (Pierce). Cyto- determining background pixels averaged from 2 tissue-free kines were measured using mouse-specific ELISAs (R&D images from each channel. The average background inten- Systems) following the manufacturer’s instructions. sity was subtracted from all images, then common SHG and For multiple analyte analysis, a Milliplex mouse cytokine/ immunostained thresholds were applied to distinguish chemokine magnetic bead panel kit (Millipore; catalogue SHG or immunostained pixels from background pixels. #MCYTOMAG-70K) was used following the manufacturer’s Two calculations were used to represent SHG or immuno- instructions. A Luminex 200 plate reader equipped with histochemical collagen signal: the percentage of pixels xPOnent software (University of Rochester Flow Cytometry above threshold in each image and the average intensity Core) was used to determine median fluorescence intensity of those pixels above threshold. The SHG and collagen for each analyte. The concentration of each analyte was immunostaining values from 5 regions of interest from calculated using the corresponding standard curve fit to a 5- each tumor section were averaged for each animal. parameter logistic equation. For all ELISAs, absorption was measured at 450 nm using Lung metastases a multiwell plate reader (Synergy HT; Biotek Instruments Lungs were fixed in 10% formalin and paraffin embed- Inc.). Curve fitting and sample concentration calculations ded. Five-mm-thick sections were collected every 100 mm were conducted with Gen5 software (Biotek). Concentra- through the entire lung. Tissue sections were stained using tions were normalized to total protein in homogenates as standard hematoxylin and eosin (H&E) techniques. Metas- determined with a BCA protein assay (Pierce). tases were visualized using a 4 objective lens and counted in each tissue section by a blinded observer. Immunohistochemistry, SHG, and image analysis Tumors were fixed in 4% paraformaldehyde for 72 hours, b-AR Expression and Intracellular Cyclic AMP followed by incubation in 10% sucrose and 30% sucrose for A standard radioligand-binding assay was used to deter- 24 hours each. Three adjacent tumor sections (20-mm-thick) mine specific binding of 125I-cyanopindolol (NEN Radio- were collected every 100 mm. Standard immunohistochem- chemicals) to whole cells to quantify b-AR expression, as ical techniques were used to detect blood vessels using rat- described previously (11). The procedure to measure intra- þ anti CD31 antibody (diluted 1:20; Abcam), proliferating cellular cAMP was described previously (11). Cyclic AMP cells using polyclonal rabbit-anti Ki67 (1:500; Abcam) or content was measured by ELISA (R&D Systems) following collagen I using polyclonal rabbit anti-collagen type I the manufacturer’s instructions. antibody (1:200; Abcam). Species-appropriate Alexa-Flu- or-594-conjugated secondary antibodies (Invitrogen) were Statistical analyses used to detect the primary antibodies. Fluorescein isothio- Statistical analyses were conducted with GraphPad þ cyanate (FITC)-conjugated anti-F4/80 antibody (Abcam) PRISM software with P < 0.05 considered statistically sig- þ was used to detect F4/80 macrophages. nificant. When 2 groups were compared, an F test for For immunofluorescent and SHG imaging of collagen, variance was conducted to compare variance. If variance tumor sections were imaged using multiphoton microsco- was similar, an unpaired two-tailed Student t test was used. py. Five random fields of view per tumor section were If variance differed, group comparisons were conducted imaged by a blinded observer using a 0.8 NA 20 water using the nonparametric Mann–Whitney U test. To com- immersion objective lens and electronic zoom at 1, one pare more than 2 groups, one- or two-way ANOVA was section per tumor. SHG and immunofluorescence emission used. When variance differed significantly between groups, was collected simultaneously under constant imaging the nonparametric Kruskal–Wallis test was used, with post

1264 Cancer Prev Res; 6(12) December 2013 Cancer Prevention Research

hoc analysis by Dunn’s multiple comparison test. Tumor and other b-AR–expressing breast cancer cell lines (11). growth over time was analyzed using repeated measures Furthermore, specific binding of the b-AR ligand 125I-cya- two-way ANOVA. Significant interactions or main effects nopindolol was readily detectable with MB-231 cells, but not were analyzed by simple effects or by Holm–Sidak multiple 4T1 cells (Fig. 1C), and isoproterenol (ISO), a b-AR agonist, comparison test. elevated intracellular cAMP in MB-231 cells but not in 4T1 cells (Fig. 1D). Together, these results demonstrate that 4T1 cells do not possess cell surface b-AR. In terms of a-AR, the Results a1-agonist phenylephrine did not significantly alter 4T1 4T1 tumor cells do not respond to norepinephrine proliferation or VEGF production (Fig. 1E and F). The a2- in vitro or signal via adrenergic receptor agonist DEX had no effect except at the highest concentration To determine if norepinephrine directly regulates 4T1 tested (42 mmol/L), where 4T1 proliferation was reduced and tumor cell functional responses, norepinephrine or selective VEGF production was increased (Fig. 1G and H; Supplemen- adrenergic receptor agonists were incubated with 4T1 cells tary Fig. S1A and S1B). However, yohimbine, an a2-AR in vitro. Norepinephrine did not alter 4T1 proliferation (Fig. antagonist, did not block these DEX-induced effects (Sup- 1A) or VEGF production (Fig. 1B) under in vitro conditions plementary Fig. S1A and S1B), and there was no evidence that altered VEGF production and proliferation in MB-231 for a2-AR signaling by cAMP inhibition (Supplementary

Figure 1. 4T1 cells do not express functional adrenergic receptor. Norepinephrine (NE) does not alter 4T1 cell proliferation (A) or VEGF production (B). 4T1 cells were incubated with NE at the concentrations indicated. Results are expressed as mean SD of triplicate wells from a representative experiment of 2 experimental repetitions. Statistical analyses: A, no main effects of NE, time, or NE time interaction; B, no main effects of NE or NE time interaction, main effect of time, P < 0.0001; C, b-AR expression as measured by specific 125I-CYP ligand binding of 4T1 (stars) and the b-AR- expressing line MB-231 (diamonds). D, b-AR agonist ISO induced cAMP response in MB- 231 compared with 4T1. E, F, 4T1 incubated with the a1-AR agonist phenylephrine (PE). Kruskal–Wallis test, no significant effects of PE. G, H, 4T1 incubated with the a2-AR agonist DEX. G, Kruskal–Wallis test, effect of DEX (P ¼ 0.03, indicated by asterisk), with no significant differences versus 0 drug by Dunn's multiple comparisons test; H, Kruskal– Wallis test, DEX treatment, P ¼ 0.053 with no differences versus 0 drug by Dunn's multiple comparisons test.

www.aacrjournals.org Cancer Prev Res; 6(12) December 2013 1265

Fig. S1C; ref. 25), as demonstrated here in a human fibroblast mg/kg ISO (nonselective b-AR agonist), 10 mg/kg phenyl- line (Supplementary Fig. S1D). We conclude that 4T1 cells ephrine (a1-AR), or 10 and 25 mg/kg DEX (a2-AR) begin- do not express functional a1-, a2-, or b-AR and therefore ning 2 days before 4T1 injection and continuing until norepinephrine cannot directly affect 4T1 function. We next sacrifice. Neither ISO nor phenylephrine treatment altered tested if elevated synaptic norepinephrine altered 4T1 tumor normalized tumor growth, tumor weight, or lung metasta- pathogenesis in vivo. ses (Fig. 3A–D). Tumor VEGF and IL-6 did not differ between phenylephrine or ISO treatment and saline con- DMI, a norepinephrine reuptake inhibitor, increases trols at sacrifice (Fig. 3E and F). Daily treatment with the 4T1 tumor growth, but not metastasis highly selective a2-AR agonist DEX (10 and 25 mg/kg) DMI inhibits norepinephrine reuptake through the nor- increased the rate of tumor growth and the number of epinephrine transporter and thereby increases synaptic metastasis in the lung compared with saline controls (Fig. norepinephrine (23, 24). To assess alterations in synaptic 3G and H). Immunohistochemistry using anti-Ki67 to norepinephrine in the periphery, norepinephrine and its detect proliferating cells revealed an increase in proliferating metabolite normetanephrine were measured in the densely cells in the 10 mg/kg DEX group compared with saline and innervated spleen. Norepinephrine concentration primarily 25 mg/kg groups (Fig. 3I). DEX treatment did not alter tumor represents norepinephrine synthesized and stored intraneu- VEGF (Fig. 3J) or IL-6 (Fig. 3K) at sacrifice. ronally, but normetanephrine is produced by an extraneur- onal enzyme (catechol-O-methyltransferase), and reflects Potential cytokine/chemokine mechanisms underlying norepinephrine released and metabolized in the synapse DMI- and DEX-induced tumor pathogenesis (26). In mice implanted subcutaneously with 21-day con- To further probe the mechanisms underlying differences tinuous release pellets containing DMI (10 mg) or placebo, in DMI versus DEX-induced tumor progression, additional DMI increased splenic norepinephrine 2-fold and normeta- tumor cytokines and chemokines were measured by multi- nephrine 4-fold relative to placebo in association with plex analysis. In DMI-treated mice, the proinflammatory reduced spleen mass 3 days after pellet implantation (Sup- cytokine TNF-a was nonsignificantly increased (Fig. 4A; plementary Fig. S2A–S2C). The increased normetanephrine Mann–Whitney, P ¼ 0.07). A similar trend was detected in is evidence of effective norepinephrine uptake blockade, mice treated with 10 mg/kg DEX (Fig. 4B; ANOVA, P ¼ 0.052) and elevated normetanephrine (but not norepinephrine) but not in phenylephrine- or ISO-treated mice (data not was detected in the spleen up to 7 days after DMI implan- shown). TNF-a added directly to 4T1 cells did not alter tation (data not shown). In comparison to daily intraperi- proliferation in vitro (Supplementary Fig. S4), indicating that toneal injection of DMI, subcutaneous pellets elicited greater elevated TNF-a cannot directly increase 4T1 tumor growth. magnitude and longer lasting elevation in splenic norepi- Furthermore, neither DMI nor DEX treatment altered the nephrine and normetanephrine (data not shown). proinflammatory cytokine IL-1b, the T-cell-associated cyto- To assess tumor growth, mice were implanted with pellets kines IL-2 and IFN-g, or the anti-inflammatory cytokine IL- containing DMI or placebo 2 days before 4T1 injection, a 10 (data not shown). However, several tumor chemokines treatment regimen similar to the chronic pharmacological that promote tumor metastasis and regulate macrophage b-AR activation that elicited increased tumor growth/meta- activity including RANTES (CCL-5), M-CSF (CSF-1), and stasis in b-AR–expressing tumor models (7–9). Using pellets MIP-2 (CCL-2) were decreased by DMI treatment (Fig. containing 5, 7.5, or 10 mg DMI in pilot studies (data not 4A), but not altered by DEX treatment (Fig. 4B). Neither shown), we determined the 10 mg dose increased tumor DMI nor DEX treatment significantly altered the density of þ growth most effectively. DMI treatment (10 mg) increased F4/80 tumor macrophages (Supplementary Fig. S5). tumor volume (mm3; Fig. 2A; see figure legends for statistical analyses) and growth rate (volume normalized; Fig. 2B), in DMI- and DEX-induced tumor growth is associated association with significantly increased tumor weight by day with altered SHG-producing tumor collagen 14 post-4T1 injection (Fig. 2C). Despite the increase in tumor Structural alterations in fibrillar collagen, uniquely visi- growth, metastasis to the lungs was not altered in DMI-treated ble via SHG imaging, are associated with tumor cell prolif- mice (Fig. 2D and E). DMI treatment significantly reduced eration and motility (18, 27). Our laboratory has demon- tumor VEGF, a key proangiogenic cytokine (Fig. 2F), and strated that stromal TNF-a knockout, or depletion of transiently decreased tumor IL-6, a proinflammatory cytokine macrophages, reduced tumor growth and metastasis and with proangiogenic activity (Fig. 2G); however, DMI treat- was associated with alterations in SHG (28). These findings, þ ment did not alter CD31 blood vessel density (Supplemen- combined with the association between DMI- and DEX- tary Fig. S3). Tumor norepinephrine was not altered by DMI induced tumor growth and trends toward increased tumor treatment (Fig. 2H). A transient increase in tumor normeta- TNF-a, led us to explore a matrix-based mechanism under- nephrine was detected at day 12 post-4T1 injection (Fig. 2I). lying the increased tumor growth. We analyzed collagen in 4T1 tumors from DMI-, DEX-, a2-AR activation increases breast tumor growth and and ISO-treated mice using 2 methods: standard immu- metastasis nohistochemical analysis and SHG imaging. Figure 5A To determine if selective activation of adrenergic receptor shows 2 representative images from a 4T1 tumor with can increase tumor growth, mice were injected daily with 5 SHG-producing collagen (in blue) and collagen type I

1266 Cancer Prev Res; 6(12) December 2013 Cancer Prevention Research

A Tumor volume BCTumor growth rate Tumor weight 600 40 400 Placebo Placebo Placebo ) *** 3 * 10 mg DMI 10 mg DMI 10 mg DMI 30 300 400 *** * 20 200

200 10 100 Tumor volume (mm Tumor weight (mg) weight Tumor Normalized tumor volume 0 0 0 141210753 14121075 Day 7 Day 12 Day 14 Time (post-4T1 injection) Time (post-4T1 injection) Time (post-4T1 injection)

D Lung metastasis EFTumor VEGF G 20 H&E lung Tumor IL-6 0.05 Placebo Placebo 0.3 10 mg DMI 15 10 mg DMI * 0.04

0.2

10 µ g protein) µ g protein) 0.03 *

0.1 5 0.01 No. of metastases IL-6 (pg/ VEGF (pg/ VEGF

0 0.0 0.00 Placebo 10 mg DMI Day 7 Day 12 Day 14 Day 7 Day 12 Day 14 Time (post-4T1 injection) Time (post-4T1 injection)

HITumor NE Tumor NMN 2.0 6 Placebo Placebo 10 mg DMI DMI 1.5 * 4

1.0

Normalized NE NE Normalized 0.5 Normalized NMN NMN Normalized

0.0 0 Day 7 Day 12 Day 14 Day 7 Day 12 Day 14 Time (post-4T1 injection) Time (post-4T1 injection)

Figure 2. DMI treatment increased tumor growth but not metastasis. Mice were implanted with 10 mg DMI or placebo continuous release pellets 2 days before 4T1 inoculation. Tumor growth expressed as volume (A) or normalized (B) and as tumor weight (C). Lung metastasis (D) were measured 14 days post- 4T1 injection. E, representative H&E-stained lung with metastatic lesion indicated by arrow. 4magnification; scale bar ¼ 200 mm. A, D, results represent 1 of 2 experimental replications at each time point. For (H) and (I) results from 2 experimental repetitions were normalized relative to the respective placebo controls. Results are expressed as mean SEM, n ¼ 6–7 mice per group for day 7 and 12, n ¼ 9–10 per group day 14 post-4T1 injection. Statistical analyses: (A) main effect of DMI treatment, P ¼ 0.047; treatment time interaction, P < 0.0001; main effect of time, P < 0.0001; B, main effect of DMI treatment, P ¼ 0.0005; treatment time interaction, P ¼ 0.0001, main effect of time, P < 0.0001; C, tumor weight: main effect of DMI, P ¼ 0.02; DMI time interaction, P ¼ 0.04, main effect of time, P < 0.0001; D, lung metastasis: Student t test, P ¼ 0.6; F, VEGF: DMI treatment, P ¼ 0.0006, no interaction or effect of time; G, IL-6: main effect of DMI treatment, P ¼ 0.004, no DMI time interaction, main effect of time, P ¼ 0.0001; H, tumor normetanephrine (NMN): DMI time interaction, P ¼ 0.02). Asterisks indicate significant differences versus corresponding placebo control group by Holm–Sidak's multiple comparisons test (P < 0.05). detected by immunohistochemistry (in green) in the same indicate changes in tumor collagen microstructure with section. Image analysis revealed that in 4T1 tumors from DMI, DEX, and ISO treatment, and each treatment unique- DMI-treated mice, the SHG-emitting pixel intensity was ly correlates with augmentation of primary tumor growth increased (Fig. 5B) without a change in the density of SHG- (DMI), tumor growth and metastasis (DEX), or no alter- emitting collagen as determined by the number of SHG ation (ISO; summarized in Table 1). pixels above a common threshold (Fig. 5C). No change in total collagen was detected by anti-collagen I immunohistochemistry (Fig. 5D). However, DEX treatment did not Discussion alter the intensity of SHG pixels above threshold (Fig. 5E), Adrenergic receptor activation promotes tumor growth or but it increased the number of SHG pixels above threshold metastasis in several animal models of cancer in which the (Fig. 5F) with no change in total collagen (Fig. 5G). The tumor cells express functional adrenergic receptor (7, 8, 29). b-agonist ISO produced no change in SHG intensity (Fig. We demonstrate here that the mammary adenocarcinoma 5H) and a nonsignificant increase in SHG pixel number cell line 4T1 lacks a-AR and b-AR expression and signaling (Fig. 5I, P ¼ 0.07). By immunohistochemistry, ISO also capacity and is unable to directly respond to norepineph- produced a nonsignificant reduction in total collagen pixel rine. To our knowledge, no studies have investigated the in intensity (Fig. 5J, P ¼ 0.15). These alterations in SHG vivo impact of sympathetic activation and norepinephrine

www.aacrjournals.org Cancer Prev Res; 6(12) December 2013 1267

PE/ISO treatment DEX treatment A B G H 10 40 Saline 40 Saline 40 Saline 5 mg/kg ISO 10 mg/kg PE 10 µg/kg DEX 8 30 30 30 25 µg/kg DEX * * ** 6 ** 20 20 20 4

10 10 10 2 No. of metastases Normalized tumor volume Normalized tumor volume 0 Normalized tumor volume 0 0 0 19171412107 19171412107 19171412107 19171412107 Saline 10 µg/kg 25 µg/kg Time (days post-4T1 injection) Time (days post-4T1 injection) Time (days post-4T1 injection) Dexmedetomidine dose CD I 15 500 2.5 SAL SAL * * 400 2.0 10 * 300 1.5

200 1.0 5 100 % Ki67+ Pixels 0.5 * No. of metastases No. of Tumor weight(mg) * 0.0 * 0 0 SALINE 10 µg/kg 25 µg/kg Saline PE ISO Saline PE ISO Dexmedetomidine dose DEX * DEX EF J 0.5 0.08 0.5

0.4 0.4 0.06 0.3 0.3 µ g protein)

µ g protein) 0.04 * 0.2 0.2 * * 0.02 0.1 0.1 IL-6 (pg/ VEGF (pg/ µ g protein) VEGF (pg/ 0.0 0.00 0.0 Saline PE ISO Saline PE ISO Saline 10 µg/kg 25 µg/kg Dexmedetomidine dose K 0.06

0.04 µ g protein)

0.02 IL-6 (pg/

0.00 Saline 10 µg/kg 25 µg/kg Dexmedetomidine dose

Figure 3. 4T1 tumor growth and metastasis in mice treated with ISO, PE, or DEX. Treatment was initiated 2 days before 4T1 injection and continued daily until sacriﬁce on day 19 post-4T1 injection. ISO and PE treatment were part of the same experiment and shared a saline control group, but for ease of comparison, the normalized tumor growth for ISO and PE are graphed separately in (A) and (B). A–F, PE and ISO treatment, n ¼ 6 per group. G–K, DEX treatment, n ¼ 9 þ per group. I, representative images of Ki67 proliferating cells in 2 tumors each from saline and DEX-treated (10 mg/kg) mice; 10 magniﬁcation; scale bar ¼ 200 mm. Asterisks and arrows indicate 2 morphologies apparent with Ki67 staining in 4T1 tumors. All results are expressed as mean SEM. Statistical analyses: A, ISO, normalized tumor volume: no main effect of treatment (P ¼ 0.14) or interaction by time (P ¼ 0.22), main effect of time (P < 0.0001); B, PE, normalized tumor volume: no treatment (P ¼ 0.4) or interaction (P ¼ 0.4), main effect of time (P < 0.0001); C–F, PE or ISO versus saline, Student t test, P > 0.05. G, DEX, normalized tumor volume, main effect of treatment, P ¼ 0.009; interaction, P < 0.0001; time, P < 0.0001; Simple effects analysis of main effect, P < 0.0001 versus saline; H, metastasis, P ¼ 0.03; Holm–Sidak analysis, P < 0.05 versus saline; I, Ki67 analysis: Kruskal–Wallis, P ¼ 0.04; J, VEGF, Kruskal–Wallis, P ¼ 0.6; K, IL-6, Kruskal–Wallis, P ¼ 0.3.

stimulation on tumor pathogenesis when the tumor cells vation may modify the tumor extracellular matrix to regu- cannot directly respond to norepinephrine. Doing so late pathogenesis. Together, these results reveal novel path- removes the inﬂuence of tumor adrenergic receptors and ways by which SNS activation can drive tumor growth and thus allows the investigation of adrenergic receptor stimu- metastasis despite the inability of the tumor cells themselves lation of adrenergic receptor-expressing host stromal cells, to respond to norepinephrine. effects that may be masked when tumor cells express functional adrenergic receptor. Under these conditions, we A role for a2-AR and host stromal cells in tumor growth demonstrated that inhibition of norepinephrine reuptake and metastasis to elevate synaptic norepinephrine promotes 4T1 tumor The ﬁnding that DMI treatment elevated norepinephrine growth, and a2-AR activation can drive tumor growth and and increased 4T1 tumor growth is consistent with other metastasis. The distinct alterations in SHG emission from reports demonstrating augmentation of tumor growth and/ tumor collagen accompanying elevated norepinephrine or metastasis associated with sympathetic innervation and and adrenergic receptor stimulation suggests that SNS acti- b-AR–expressing tumor or stromal cells (6–8). However, our

1268 Cancer Prev Res; 6(12) December 2013 Cancer Prevention Research

Figure 4. Multiplex analysis of tumor chemokines and cytokines in mice treated with (A) DMI or (B) DEX. DMI elicited reductions in chemokines that were not observed with DEX treatment. Results are expressed as mean SEM, n ¼ 7 to 8 mice per group. A, Student t test or nonparametric Mann–Whitney (M-W) P-values are indicated for each chemokine/cytokine. B, P-values represent analysis by one-way ANOVA. indicates signiﬁcant differences versus corresponding control group (P < 0.05).

results differ in several important ways from these reports. bility is that norepinephrine stimulation of b-AR or a1-AR We found no evidence that increased tumor growth was pathways may oppose a2-AR signaling in the tumor or associated with b-AR activation or with increased angiogen- elsewhere in the periphery. This is supported by the obser- esis or proangiogenic cytokines. Furthermore, in this adren- vation that although DEX ligation of a2-AR did not affect ergic receptor-negative tumor model, b-AR activation with VEGF, IL-6, RANTES, and other prometastatic cytokines, ISO did not increase lung metastasis or elevate M-CSF (also DMI-induced elevation of synaptic norepinephrine reduced known as CSF-1; data not shown), as reported by Sloan and them, presumably via adrenergic receptors other than a2, colleagues. We contend a b-AR–induced reduction in CSF-1 thus providing an antimetastatic counter to the prometa- or other stromal-derived cytokines/chemokines may be static effects of a2-AR stimulation by elevated norepineph- obscured if tumor cytokine production (such as CSF-1) is rine. One way to test this possibility is to selectively block increased by b-AR activation. In this scenario, whether or not adrenergic receptor in DMI-treated mice, but we found the the tumor cells can respond to b-AR stimulation may dictate combination of DMI and b-blocker treatment was associ- directionally opposite tumor outcomes. ated with a high level of mortality; we therefore tested direct Our results are consistent with a small number of reports stimulation of b-AR and a1-AR. Although we did not detect of increased proliferation by a2-AR–expressing breast signiﬁcant alterations in 4T1 tumor growth and metastasis tumor cell in vitro and increased tumor growth in vivo with ISO or phenylephrine treatment, in ISO-treated mice, a (13, 29, 30), but we have been unable to detect functional trend toward reduced tumor growth (Fig. 3A) and chemo- a2-AR in several breast cancer cell lines (data not shown). In kine production was noted (data not shown), and we are fact, DEX at supra micromolar concentrations may act on further exploring b-AR activation in this tumor model. imidazoline receptors (31) to reduce 4T1 cell proliferation We have yet to identify the stromal targets of norepine- and increase VEGF production in vitro (Fig. 1G and H). This phrine and a2-AR stimulation or their location. DMI only concentration of DEX, orders of magnitude above the Ki for transiently increased synaptic norepinephrine in the tumor, a2-AR (1.08 nmol/L; ref. 32), is unlikely to be achieved in as indicated by the norepinephrine metabolite normeta- vivo at the doses used here. Instead, we propose that DEX nephrine. It should be noted that at no time point post- activation of host stromal a2-AR increased tumor growth DMI treatment was splenic or tumor norepinephrine or and metastasis. Intriguingly, in human breast cancer, the normetanephrine content reduced, indicative of sympathe- a2A-AR gene was one of 26 tumor stroma genes that together tic nerve depletion because of chronic elevation of norepi- predicted poor outcome (33). One question raised by our nephrine. Splenic norepinephrine and normetanephrine results is why DMI treatment—and increased synaptic nor- were dramatically increased early after DMI implantation, epinephrine—did not activate a2-AR to affect metastatic an effect that subsided but was still apparent as the tumor outcome along with increased tumor growth. One possi- developed (data not shown). The DMI-induced increase in

www.aacrjournals.org Cancer Prev Res; 6(12) December 2013 1269

Figure 5. Tumor collagen structure is differentially altered by DMI, DEX, and ISO treatment. Tumor slices were stained for collagen by standard immunohistochemical techniques and imaged to detect immunoﬂuorescent and SHG emission by multiphoton microscopy. Image analysis was conducted with Image J as described in materials and methods. A, 2 representative pseudo-colored images of SHG (blue) versus collagen type I (green) from a 4T1 tumor. Scale bars ¼ 100 mm. B, E, H, SHG pixel intensity above threshold; C, F, I, percentage of SHG pixels above threshold; D, G, J, anticollagen immunohistochemical analysis. Results shown are mean SEM, n ¼ 8 to 9 mice per group for both DMI and DEX experiments; n ¼ 6 per groups for ISO experiment. Asterisk indicates signiﬁcant differences based on Student t test, P < 0.05 versus placebo or saline by Student t test; M-W ¼ Mann–Whitney nonparametric U test.

splenic norepinephrine points to the potential targeting of involving the mild sedation (central nervous system effects) adrenergic receptor–expressing cells in extratumoral organs, observed with 25 mg/kg DEX. Nonetheless, our results dem- such as spleen and bone marrow, that play a role in 4T1 onstrate that the elevated tumor growth associated with DEX pathogenesis (34). Similarly, DEX treatment may target a2- and DMI treatment is driven by changes apparent within the expressing cells within the tumor and the lung to promote tumor, including the extracellular matrix. metastasis. Furthermore, we cannot rule out these drugs acting at the level of the central nervous system. For example, A novel mechanism for SNS regulation of tumor þ the DEX-induced increase in tumor Ki67 proliferating cells progression: collagen microstructure and SHG imaging (Fig. 3I) and TNF-a (Fig. 4B) at 10 mg/kg, but not 25 mg/kg, Tumor stromal cells, including macrophages and ﬁbro- may indicate distinct mechanisms underlying DEX treatment blasts, regulate tumor collagen structure. The detected SHG

Table 1. Summary of DMI, DEX, and ISO-induced tumor matrix alterations

Treatment SHGþ content SHGþ ﬁbers Total collagen (AR-selectivity) Tumor growth Metastasis (intensity) (% pixels) (IF) DMI (mixed AR) Increased No change Increased No change No change DEX (a2) Increased Increased No change Increased No change ISO (b-AR) No change No change No change No change Decreased

1270 Cancer Prev Res; 6(12) December 2013 Cancer Prevention Research

signal from a collagen fiber is sensitive to the amount of treat cancer pain, adding urgency to delve further into the collagen (15), as well as the diameter of the fibrils that form mechanisms underlying chronic DEX treatment and fibers (35, 36), their spacing (36), and the order versus increased tumor growth and metastasis (43, 44). Finally, disorder in fibril packing (37). Here, we define a change in our results with DMI and elevated synaptic norepineph- one or more of the latter 3 parameters (fibril diameter, fibril rine imply a balance between prometastatic effects of a2- spacing, and order) as a change in collagen fiber "micro- AR and antimetastatic b-AR with increased norepineph- structure." Hence, the signal produced by SHG differs from rine release. If true, caution should be applied in the immunofluorescent detection of collagen, which is sensitive clinical adjuvant use of b-blockers, as proposed by others to epitope concentration and reports primarily the amount (3, 45). of collagen in a given region of interest. The 2 readouts [SHG In summary, our results strongly implicate a2-AR activa- and immunofluorescence (IF)] from a given collagen fiber tion as a promoter of tumor pathogenesis in the absence of can be compared with gain insight into the extent of changes direct sympathetic input to the tumor cells. The results in fibril microstructure (primarily alters SHG) versus suggest a unique, matrix-based mechanism whereby nor- changes in collagen content (alters both SHG and IF; see epinephrine can facilitate breast tumor growth through refs. 28, 38, and 39). regulation of the tumor extracellular matrix, specifically As revealed by image analysis, tumor SHG-emitting col- collagen microstructure. Further investigation into this lagen was increased by both DMI and DEX treatment in mechanism is critical in part because of the possibility of subtly different ways. In DMI-treated mice, tumor SHG using SHG imaging to detect SNS-induced alterations in the pixel intensity above threshold increased without an tumor matrix as a marker of a more aggressive tumor increase in the percentage of pixels above threshold, repre- phenotype. Our results suggest that norepinephrine can þ senting a change in SHG fiber microstructure (Fig. 5B and elicit directionally opposing effects within the tumor that C). However, increased SHG emission from tumors from must be carefully investigated to understand the impact of DEX-treated mice was because of an increase in the per- stress-induced SNS activation in a disease as molecularly centage of pixels above threshold without an increase in the heterogeneous as breast cancer. intensity of those pixels above threshold (Fig. 5E and F), þ suggesting an increase in SHG fiber content relative to Disclosure of Potential Conflicts of Interest control tumors. Neither of these types of increased SHG was No potential conflicts of interest were disclosed. associated with altered total collagen as measured by IF (Fig. Authors' Contributions 5D and G), indicative of an alteration in the structure of the Conception and design: M.J. Szpunar, E.B. Brown, K.S. Madden tumor collagen and not collagen deposition. Interestingly, Development of methodology: M.J. Szpunar, E.B. Brown ISO elicited a trend toward decreased total collagen as Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.J. Szpunar, K.A. Burke, R.P. Dawes, E.B. Brown, measured by IF (Fig. 5J), and yielded no change in tumor K.S. Madden growth or metastasis despite the trend toward increased Analysis and interpretation of data (e.g., statistical analysis, biosta- þ SHG fiber content (Fig. 5I). Based on the evidence that tistics, computational analysis): M.J. Szpunar, K.A. Burke, R.P. Dawes, E.B. Brown, K.S. Madden microstructural changes in tumor collagen, as detected by Writing, review, and/or revision of the manuscript: M.J Szpunar, K.A. SHG, may drive tumor cell proliferation, local invasion, and Burke, R.P. Dawes, E.B. Brown, K.S. Madden Administrative, technical, or material support (i.e., reporting or orga- metastasis (21, 40, 41), we contend that the treatment- nizing data, constructing databases): E.B. Brown dependent changes in SHG-emitting collagen shown here Study supervision: E.B. Brown, K.S. Madden reflect a stromal-based mechanism by which adrenergic receptor activation may promote tumor progression. Stud- Acknowledgments ies are underway to understand the adrenergic receptor The authors thank K. Liverpool, D. Byun, T. Bubel, G. Arcuri, T. Wolfgang, þ and P. Stastka for their excellent technical assistance. mechanisms that alter SHG collagen microstructure and to demonstrate that such changes lead to DEX- or DMI- Grant Support induced tumor progression. These results illustrate the M.J. Szpunar was supported by Department of Defense Predoctoral potential power of using SHG imaging in the 4T1 model Training Award (W81XWH-10-1-0058); Medical Scientist Training Pro- gram (NIH T32 GM07356); National Center for Research Resources (TL1 to distinguish overlapping and opposing effects of increased RR024135). K.A. Burke was supported by Department of Defense Era norepinephrine with sympathetic activation consistent of Hope Scholar Research Award (W81XWH-09-1-0405) to E.B. Brown. with our proposal that norepinephrine elicits effects via R.P. Dawes was supported by NIH Training Grant in Neuroscience. E.B. Brown was supported by Pew Scholar in the Biomedical Sciences mixed adrenergic receptor signaling that can oppose each Award; Department of Defense Era of Hope Scholar Research Award other. (W81XWH-09-1-0405); National Institutes of Health Director’s New Innovator Award (1 DP2 OD006501-01). K.S. Madden was supported by Department of Defense IDEA Award (W81XWH-10-01-008), National Clinical implications Institutes of Health (1 R21 CA152777-01). AnimportantaspectofthisworkisthatDMIandDEX The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked are used clinically. A retrospective study examining the advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate clinical use of antidepressants and breast cancer devel- this fact. opment found an association between DMI and increased breast cancer risk (42), consistent with the protumor Received March 2, 2013; revised October 13, 2013; accepted October 15, growth effect shown here. DEX is used as a sedative to 2013; published online December 5, 2013.

www.aacrjournals.org Cancer Prev Res; 6(12) December 2013 1271

References 1. Reiche EM, Nunes SO, Morimoto HK. Stress, depression, the immune 24. Glowinski J, Axelrod J. Inhibition of uptake of tritiated-noradrenaline in system, and cancer. Lancet Oncol 2004;5:617–25. the intact rat brain by imipramine and structurally related compounds. 2. Armaiz-Pena GN, Lutgendorf SK, Cole SW, Sood AK. Neuroendocrine Nature 1964;204:1318–9. modulation of cancer progression. Brain Behav Immun 2009;23:10–5. 25. Marjamaki A, Ala-Uotila S, Luomala K, Perala M, Jansson C, Jalkanen 3. Cole SW, Sood AK. Molecular pathways: b-adrenergic signaling in M, et al. Stable expression of recombinant human a2-adrenoceptor cancer. Clin Cancer Res 2012;18:1201–6. subtypes in two mammalian cell lines: characterization with 4. Antoni MH, Lutgendorf SK, Cole SW, Dhabhar FS, Sephton SE, [3H]rauwolscine binding, inhibition of adenylate cyclase and RNase McDonald PG, et al. The influence of bio-behavioural factors on protection assay. Biochim Biophys Acta 1992;1134:169–77. tumour biology: pathways and mechanisms. Nat Rev Cancer 2006; 26. Eisenhofer G, Kopin IJ, Goldstein DS. Catecholamine metabolism: a 6:240–8. contemporary view with implications for physiology and medicine. 5. Lutgendorf SK, Sood AK. Biobehavioral factors and cancer progres- Pharmacol Rev 2004;56:331–49. sion: physiological pathways and mechanisms. Psychosom Med 27. Egeblad M, Rasch MG, Weaver VM. Dynamic interplay between 2011;73:724–30. the collagen scaffold and tumor evolution. Curr Opin Cell Biol 6. Magnon C, Hall SJ, Lin J, Xue X, Gerber L, Freedland SJ, et al. 2010;22:697–706. Autonomic nerve development contributes to prostate cancer pro- 28. Burke RM, Madden KS, Perry SW, Zettel ML, Brown EB. Tumor-asso- gression. Science 2013;341:1236361. ciated macrophages and stromal TNF-alpha regulate collagen structure in 7. Thaker PH, Han LY, Kamat AA, Arevalo JM, Takahashi R, Lu C, et al. breast tumor models as visualize by second harmonic generation. Chronic stress promotes tumor growth and angiogenesis in a mouse J Biomed Opt 2013;18:86003. model of ovarian carcinoma. Nat Med 2006;12:939–44. 29. Bruzzone A, Pinero CP, Castillo LF, Sarappa MG, Rojas P, Lanari C, 8. Sloan EK, Priceman SJ, Cox BF, Yu S, Pimentel MA, Tangkanangnukul et al. a2-Adrenoceptor action on cell proliferation and mammary V, et al. The sympathetic nervous system induces a metastatic switch tumour growth in mice. Br J Pharmacol 2008;155:494–504. in primary breast cancer. Cancer Res 2010;70:7042–52. 30. Vazquez SM, Mladovan AG, Perez C, Bruzzone A, Baldi A, Luthy IA. 9. Campbell JP, Karolak MR, Ma Y, Perrien DS, Masood-Campbell SK, Human breast cell lines exhibit functional a2-adrenoceptors. Cancer Penner NL, et al. Stimulation of host bone marrow stromal cells by Chemother Pharmacol 2006;58:50–61. sympathetic nerves promotes breast cancer bone metastasis in mice. 31. Khan ZP, Ferguson CN, Jones RM. a2 and imidazoline receptor PLoS Biol 2012;10:e1001363. agonists. Their pharmacology and therapeutic role. Anaesthesia 10. Powe DG, Voss MJ, Habashy HO, Zanker KS, Green AR, Ellis IO, et al. 1999;54:146–65. a- and b-adrenergic receptor (AR) protein expression is associated 32. Piletz JE, Zhu H, Chikkala DN. Comparison of ligand binding affinities with poor clinical outcome in breast cancer: an immunohistochemical at human I1-imidazoline binding sites and the high affinity state of a2 study. Breast Cancer Res Treat 2011;130:457–63. adrenoceptor subtypes. J Pharmacol Exp Ther 1996;279:694–702. 11. Madden KS, Szpunar MJ, Brown EB. b-Adrenergic receptors (b-AR) 33. Finak G, Bertos N, Pepin F, Sadekova S, Souleimanova M, Zhao H, regulate VEGF and IL-6 production by divergent pathways in high et al. Stromal gene expression predicts clinical outcome in breast b-AR-expressing breast cancer cell lines. Breast Cancer Res Treat cancer. Nat Med 2008;14:518–27. 2011;130:747–58. 34. DuPre SA, Hunter KW Jr. Murine mammary carcinoma 4T1 induces a 12. Vandewalle B, Revillion F, Lefebvre J. Functional b-adrenergic recep- leukemoid reaction with splenomegaly: association with tumor- tors in breast cancer cells. J Cancer Res Clin Oncol 1990;116:303–6. derived growth factors. Exp Mol Pathol 2007;82:12–24. 13. Bruzzone A, Pinero CP, Rojas P, Romanato M, Gass H, Lanari C, et al. 35. Han X, Burke RM, Zettel ML, Tang P, Brown EB. Second harmonic a(2)-Adrenoceptors enhance cell proliferation and mammary tumor properties of tumor collagen: determining the structural relationship growth acting through both the stroma and the tumor cells. Curr between reactive stroma and healthy stroma. Opt Express 2008;16: Cancer Drug Targets 2011;11:763–74. 1846–59. 14. Nance DM, Sanders VM. Autonomic innervation and regulation of the 36. Williams RM, Zipfel WR, Webb WW. Interpreting second-harmonic immune system (1987–2007). Brain Behav Immun 2007;21:736–45. generation images of collagen I fibrils. Biophys J 2005;88:1377–86. 15. Moreaux L, Sandre O, Charpak S, Blanchard-Desce M, Mertz J. 37. Lacomb R, Nadiarnykh O, Townsend SS, Campagnola PJ. Phase Coherent scattering in multi-harmonic light microscopy. Biophys J matching considerations in second harmonic generation from tissues: 2001;80:1568–74. effects on emission directionality, conversion efficiency and observed 16. Sidani M, Wyckoff J, Xue C, Segall JE, Condeelis J. Probing the morphology. Opt Commun 2008;281:1823–32. microenvironment of mammary tumors using multiphoton microsco- 38. Diop-Frimpong B, Chauhan VP, Krane S, Boucher Y, Jain RK. Losartan py. J Mammary Gland Biol Neoplasia 2006;11:151–63. inhibits collagen I synthesis and improves the distribution and efficacy 17. Wang W, Wyckoff JB, Frohlich VC, Oleynikov Y, Huttelmaier S, Zavadil of nanotherapeutics in tumors. Proc Natl Acad Sci U S A 2011;108: J, et al. Single cell behavior in metastatic primary mammary tumors 2909–14. correlated with gene expression patterns revealed by molecular pro- 39. Ingman WV, Wyckoff J, Gouon-Evans V, Condeelis J, Pollard JW. filing. Cancer Res 2002;62:6278–88. Macrophages promote collagen fibrillogenesis around terminal end 18. Condeelis J, Segall JE. Intravital imaging of cell movement in tumours. buds of the developing mammary gland. Dev Dyn 2006;235:3222–9. Nat Rev Cancer 2003;3:921–30. 40. Provenzano PP, Eliceiri KW, Campbell JM, Inman DR, White JG, Keely 19. Burke K, Tang P, Brown E. Second harmonic generation reveals matrix PJ. Collagen reorganization at the tumor-stromal interface facilitates alterations during breast tumor progression. J Biomed Opt 2013;18: local invasion. BMC Med 2006;4:38. 31106. 41. Provenzano PP, Inman DR, Eliceiri KW, Knittel JG, Yan L, Rueden CT, 20. Kakkad SM, Solaiyappan M, Argani P, Sukumar S, Jacobs LK, Leibfritz et al. Collagen density promotes mammary tumor initiation and pro- D, et al. Collagen I fiber density increases in lymph node positive breast gression. BMC Med 2008;6:11. cancers: pilot study. J Biomed Opt 2012;17:116017. 42. Sharpe CR, Collet JP, Belzile E, Hanley JA, Boivin JF. The effects of 21. Conklin MW, Eickhoff JC, Riching KM, Pehlke CA, Eliceiri KW, Pro- tricyclic antidepressants on breast cancer risk. Br J Cancer 2002;86:92–7. venzano PP, et al. Aligned collagen is a prognostic signature for 43. Roberts SB, Wozencraft CP, Coyne PJ, Smith TJ. Dexmedetomidine survival in human breast carcinoma. Am J Pathol 2011;178:1221–32. as an adjuvant analgesic for intractable cancer pain. J Palliat Med 22. Aslakson CJ, Miller FR. Selective events in the metastatic process 2011;14:371–3. defined by analysis of the sequential dissemination of subpopulations 44. Ugur F, Gulcu N, Boyaci A. Intrathecal infusion therapy with dexme- of a mouse mammary tumor. Cancer Res 1992;52:1399–405. detomidine-supplemented morphine in cancer pain. Acta Anaesthe- 23. Schildkraut JJ, Dodge GA, Logue MA. Effects of tricyclic antidepres- siol Scand 2007;51:388. sants on the uptake and metabolism of intracisternally administered 45. Powe DG, Entschladen F. Targeted therapies: using b-blockers to norepinephrine-H3 in rat brain. J Psychiatr Res 1969;7:29–34. inhibit breast cancer progression. Nat Rev Clin Oncol 2011;8:511–2.

1272 Cancer Prev Res; 6(12) December 2013 Cancer Prevention Research

Downloaded from cancerpreventionresearch.aacrjournals.org on September 23, 2021. © 2013 American Association for Cancer Research. The Antidepressant Desipramine and α2-Adrenergic Receptor Activation Promote Breast Tumor Progression in Association with Altered Collagen Structure

Mercedes J. Szpunar, Kathleen A. Burke, Ryan P. Dawes, et al.

Cancer Prev Res 2013;6:1262-1272.

Updated version Access the most recent version of this article at: http://cancerpreventionresearch.aacrjournals.org/content/6/12/1262

Supplementary Access the most recent supplemental material at: Material http://cancerpreventionresearch.aacrjournals.org/content/suppl/2014/07/23/1940-6207.CAPR-13-0079.DC 1

Cited articles This article cites 45 articles, 9 of which you can access for free at: http://cancerpreventionresearch.aacrjournals.org/content/6/12/1262.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://cancerpreventionresearch.aacrjournals.org/content/6/12/1262.full#related-urls

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 Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerpreventionresearch.aacrjournals.org/content/6/12/1262. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.