Open Full Page

Published OnlineFirst May 9, 2016; DOI: 10.1158/0008-5472.CAN-15-1524

Cancer Therapeutics, Targets, and Chemical Biology Research

Renalase Expression by Melanoma and Tumor-Associated Macrophages Promotes Tumor Growth through a STAT3-Mediated Mechanism Lindsay Hollander1,2,3, Xiaojia Guo1,2, Heino Velazquez1,2,4, John Chang1,4, Robert Saﬁrstein1,2,4, Harriet Kluger5,2, Charles Cha2,4,6, and Gary V. Desir1,2,4

Abstract

To sustain their proliferation, cancer cells overcome negative- derived inhibitory peptide decreased melanoma cell survival, and acting signals that restrain their growth and promote senescence anti-RNLS therapy blocked tumor growth in vivo in murine and cell death. Renalase (RNLS) is a secreted flavoprotein that xenograft assays. Mechanistic investigations showed that functions as a survival factor after ischemic and toxic injury, increased apoptosis in tumor cells was temporally related to signaling through the plasma calcium channel PMCA4b to acti- p38 MAPK-mediated Bax activation and that increased cell growth vate the PI3K/AKT and MAPK pathways. We show that RNLS arrest was associated with elevated expression of the cell-cycle expression is increased markedly in primary melanomas and inhibitor p21. Overall, our results established a role for the þ CD163 tumor-associated macrophages (TAM). In clinical speci- secreted flavoprotein RNLS in promoting melanoma cell growth þ mens, RNLS expression in the tumor correlated inversely with and CD163 TAM in the tumor microenvironment, with poten- disease-specific survival, suggesting a pathogenic role for RNLS. tial therapeutic implications for the management of melanoma. Attenuation of RNLS by RNAi, blocking antibodies, or an RNLS- Cancer Res; 76(13); 3884–94. 2016 AACR.

Introduction lular signal-regulated kinase (ERK), and the mitogen-activated protein kinase (p38). Renalase (RNLS), a novel secreted flavoprotein with oxido- As RNLS functions as a survival factor that engages the MAPK reductase activity (1–5), functions as a survival factor and and PI3K pathways, and because its expression is regulated by promotes cell and organ survival through a receptor-mediated STAT3 (8), we asked whether RNLS expression and signaling process that is independent of its intrinsic enzymatic activities provided a survival advantage to cancer cells. We focused on (6). Extracellular RNLS signals through the plasma membrane melanoma, a disorder in which the MAPK, PI3K, and JAK/STAT calcium ATPase PMCA4b and activates the PI3K and MAPK pathways are regulated abnormally, and for which additional pathways (7). We have identified the critical region of the RNLS therapeutic targets would be desirable. molecule that mediates its cytoprotective effects, and showed Skin cancer is a common human malignancy, and its incidence that a 20 amino acid RNLS peptide (RP-220, aa 220–239: has been increasing in developed countries (9–11). Melanoma is CIRFVSIDNKKRNIESSEIG), which is conserved in all known the deadliest form of skin cancer, with low survival rates once it isoforms but is devoid of any detectable oxidase activity, was as becomes unresectable (10). It is a molecularly heterogeneous equally effective as RNLS protein at protecting HK-2 cells and disease and some of the key alterations in signaling pathways mice against toxic and ischemic injury (6). Furthermore, RNLS that participate in disease development and progression have and RP-220 rapidly activated protein kinase B (AKT), extracel- been identified. The Ras/Raf/MEK/ERK and the PI3K/AKT signaling pathways play key roles in the pathogenesis of melanoma (9, 11, 12). Mutations in Ras, Raf, PI3K, or PTEN (PI3K 1Department of Medicine, Yale University, New Haven, Connecticut. inhibitor)canleadtothesustainedactivationofERKandAKT, 2 Yale School of Medicine, Yale University, New Haven, Connecticut. which in turn promote cell survival and proliferation. Dankort 3University of Connecticut, Farmington, Connecticut. 4VA Connecti- cut Health Care System, Yale University, New Haven, Connecticut. and colleagues demonstrated this well with conditional mela- V600E 5Department of Medical Oncology, Yale University, New Haven, Con- nocyte–specific expression of BRaf in mice, none of whom 6 necticut. Department of Surgery, Yale University, New Haven, developed melanoma, however, revealed 100% penetrance of Connecticut. melanoma development when combined with silencing of the Note: Supplementary data for this article are available at Cancer Research Pten tumor suppressor gene (13). The elucidation of these Online (http://cancerres.aacrjournals.org/). pathogenic pathways has facilitated the development of specific Corresponding Author: Gary Desir, Section of Nephrology, Department of inhibitors that target hyperactivated kinases. While these agents Medicine, Yale School of Medicine, P.O. Box 208029, New Haven, CT 06520- have proven effective in the treatment of selective groups of 8029. Phone: 203-785-4119; Fax: 203-387-8573; E-mail: [email protected] patients with metastatic melanoma, their beneficial actions are doi: 10.1158/0008-5472.CAN-15-1524 often short lived, hence the pressing need for the identification 2016 American Association for Cancer Research. of additional therapeutic targets.

3884 Cancer Res; 76(13) July 1, 2016

Treatment of Melanoma Using Inhibitors of Renalase Signaling

Materials and Methods TC10 automated cell counter. For additional studies, cell viability Reagents was determined using WST-1 reagent (Roche Diagnostics) as Human melanoma cell lines A375.S2, SkMel28, SkMel5, described previously (6). MeWo, and WM266-4 were obtained from the ATCC and maintained as recommended. Mouse metastatic melanoma cell RNA interference line B16f10 was obtained from ATCC. Cells were used within Four individual siRNAs and a siRNA SMART pool targeting six months of purchase, and a link to the method of authen- RNLS were purchased from Dharmacon. Cells were transfected tication is provided here: http://www.atcc.org/support/faqs/ with RNLS siRNA or a universal negative control siRNA (control eae27/Authenticating%20cell%20lines-249.aspx. siRNA, Dharmacon) using DharmaFECT 4 reagent (Dharmacon) fi Recombinant human RNLS was expressed, purified, concen- as instructed by the manufacturer. Knockdown ef ciency was trated, and dialyzed against PBS as described previously (14). determined by qPCR. RNLS peptides RP220 and mutated peptide RP220A were synthesized at United Peptide. Rabbit anti-RNLS mAb Mouse tumor model (ab178700), goat polyclonal anti-RNLS antibody (ab31291), Female athymic, 18–20 g nude mice (nu/nu) were obtained goat IgG, and rabbit IgG were purchased from Abcam. Anti- from Charles River Laboratories and housed in microisolator RNLS mAbs m28-RNLS and m37-RNLS were synthesized as cages, with autoclaved bedding in a specific pathogen-free facility, described previously (15). with a 12-hour light/dark cycle. Animals received water and food ad libitum, and were observed for signs of tumor growth, activity, Tissue specimens feeding, and pain, in accordance with the study protocol approved Human melanoma cDNA arrays I and II were obtained by the VA Connecticut Healthcare System Institutional Animal from OriGene Technologies. The relevant pathology reports are Care and Use Committee. available online (http://www.origene.com/qPCR/Tissue-qPCR- Xenograft tumors were established by subcutaneous injection 6 m Arrays.aspx). Human melanoma and normal skin tissue samples of A375.S2 cells (2 10 in 100 L of PBS, pH 7.6). When the – 3 obtained from US Biomax were used for IHC or immunofluo- tumors reached a volume of 50 100 mm , the mice were divided n ¼ m rescence. The relative expression levels of various genes were into a control group [ 14 treated with rabbit IgG, 40 gby m assessed by qRT-PCR, as described previously (15). intraperitoneal injection (i.p.) once weekly, and 40 g subcutaneously around the tumor site every 3 days), and an experimental group (n ¼ 14) that received m28-RNLS (40 mg i.p., once weekly, Immunohistochemical staining and Western blot analysis and 40 mg s.c., every 3 days). Tumor size was measured with digital IHC was performed as described previously (16). A detailed calipers and volume was calculated according to the formula description is available in the Supplementary Data. Western blot (length width2)xp/2. Apoptosis was quantified using the analysis was carried out as described previously (6). terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay as described previously (15). Tissue microarray Melanoma tissue microarrays were purchased from US BioMax, Isolation and polarization of bone marrow–derived Inc. and Yale tissue pathology services. This study was approved macrophages by the Human Investigation Committee of Yale University School Femur and tibial bones were isolated from 10-week-old male of Medicine (HIC protocol No. 1003006479). The Yale melano- / C57BL/6 (Harlan Sprague Dawley) or RNLS mice (21) and ma tissue microarray was constructed as described previously (17, bone marrow cells were pelleted and resuspended in Macrophage- 18). A total of 570 tissue cores representing 542 total melanoma SFM (Gibco, cat # 12065-074) supplemented with 5 ng/mL cases and a small series of controls measuring 0.6 mm were spaced mouse recombinant macrophage-colony stimulating factor 0.8 mm apart on a single glass slide. The cohort was constructed 6 (mM-CSF, ConnStem, Inc., cat # M2000) at 2 10 macro- from formalin-fixed, paraffin-embedded tissue blocks obtained phages/mL. Adherent macrophages were incubated in Macro- from the archives of the Department of Pathology at Yale Uni- phage-SFM plus mM-CSF at 30 ng/mL for additional 4 days. On versity School of Medicine. A pathologist examined each case to day 5 after isolation, fresh medium containing either 50 ng/mL select the region for inclusion in the tissue microarray. Core IL6 (R&D Systems, cat # 406-ML) or 50 ng/mL leukemia inhib- biopsies from the specimens were placed on the tissue microarray itory factor (LIF, R&D Systems, cat # 8878-LF) was added to with a Tissue Micorarrayer (Beecher Instruments). The tissue polarize macrophage to M2d tumor-associated macrophages microarrays were then cut to 5-mm sections and placed on glass (TAM). IL10 measurements were obtained by qPCR 48 hours slides with the adhesive tape transfer system (Instumedics, Inc.) later. with UV crosslinking. The specimens were all drawn from archives Statistical analysis of the data was carried out as described of tumors resected between 1959 and 1994, with a follow-up previously (15). A detailed description is available in the Sup- range of 2 months and 38 years (median follow-up time, 60 plementary Data. months). The cohort characteristics are described previously (19). The tissue microarray slide was stained as described previously (19, 20). A detailed description is available in the Sup- Results plementary Data. RNLS overexpression in melanoma To determine whether RNLS expression differed between nor- Cell viability assays mal human skin and malignant melanoma, we examined TMAs Total cell number and percentage of live cells were assessed by (Yale Tissue Microarray Facility and US Biomax, Inc.) spanning Trypan blue exclusion, and cells were counted using a Bio-Rad the progression of normal skin to benign nevus to primary and

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 3885

Hollander et al.

metastatic melanoma. The Yale TMAs contained formalin-fixed, from 1997 to 2004. One hundred nineteen patients had histo- paraffin-embedded specimens obtained from a cohort of 192 spots that were suitable for evaluation by AQUA technology. In primary melanomas collected during 1959 to 1994, a cohort of this group, we compared the outcome of patients whose tumors 246 serial primary and metastatic melanomas collected from expressed high RNLS levels (RNLS AQUA score > median AQUA 1997 to 2004, a cohort of 295 patients with benign nevi, and score 75,764.45) to those with low RNLS expression. High RNLS matched normal skin specimens from 15 patients. The demo- expression was associated with increased melanoma-specific graphics and clinical characteristics for these tissue microarrays death: 5- year and 10-year disease-specific survival rates of 55% have been described previously (22). The US Biomax array con- versus 69% and 39.7% versus 58.5%, respectively, P ¼ 0.008, (Fig. tained 74 specimens, including 35 primary melanomas, 11 met- 1D). Following multivariate analysis of this cohort, RNLS levels astatic lesions, 14 benign nevi, and 14 normal samples. Exami- were found to be independently predictive of survival in mela- nation of approximately 600 histospots for RNLS protein expres- noma (P ¼ 0.004; HR ¼ 3.130). Stage of disease at diagnosis (P ¼ sion using a quantitative, automated immunofluorescence (IF) 0.05; HR ¼ 3.940), Clark level (P ¼ 0.015; HR ¼ 1.687), and microscopy system (AQUA), revealed that progression from nor- ulceration of the primary tumor (P ¼ 0.001; HR ¼ 2.54) were also mal skin to benign nevi to primary malignant melanoma to found to independently predict survival in melanoma. The 5-year metastatic melanoma was accompanied by a significant increase survival rate for patients with tumors containing lower levels of þ in RNLS expression (P ¼ 0.009, P ¼ 0.0003, and P < 0.001, infiltrating CD163 macrophage (M2-like phenotype, tumor respectively, Fig. 1A–C). promoting) was 52% compared to 42% (P < 0.04) for those with þ We asked whether dysregulated RNLS expression and signaling tumors with higher levels of CD163 cells. These findings suggest could facilitate melanoma growth, and, therefore, serve as a that RNLS expression may serve as a useful prognostic marker in prognostic marker. We examined each primary melanoma from melanoma, and may help identify a subset of patients with a more a cohort of 246 serial primary and metastatic samples collected aggressive phenotype.

Normal skin Benign nevus Malignant melanoma 47,800 47,400 47,000 46,600 46,200 45,800 45,400 45,000

2,000

AQUA score AQUA 1,600 1,200 800 400 0 Normal skin Benign nevi Malignant melanoma C D Survival functions 1,400 1.0 1,200

1,000 0.8

Low RNLS 800 0.6

600

AQUA Score AQUA 0.4 400 High RNLS Cumulative survival 0.2 200

0 0.0 Normal skin Benign nevi Primary Metastatic melanoma melanoma 0 100 200 300 400 500 Follow up (months)

Figure 1. RNLS overexpression is associated with poor outcome in melanoma. A, RNLS expression was detected using m28-RNL, normal human skin (n ¼ 15), benign nevi (n ¼ 295), malignant melanoma (n ¼ 264); representative results, blue, nuclei; green, melanocytes; red, RNLS. B, fluorescence intensity was quantified by AQUAnalysis software, Yale; TMA: normal human skin (n ¼ 15), benign nevi (n ¼ 295), malignant melanoma (n ¼ 264). C, as in B, US Biomax TMA: human skin (n ¼ 14), benign nevi (n ¼ 14), primary melanoma (n ¼ 35), metastatic melanoma (n ¼ 11; , P ¼ 0.009; , P < 0.001). D, Kaplan–Meier survival curve, melanoma-specific death; 119 serial primary melanomas collected from 1997 to 2004, stratified into low and high RNLS expression using AQUA score ¼ 75,764.45 (, P ¼ 0.008).

3886 Cancer Res; 76(13) July 1, 2016 Cancer Research

Treatment of Melanoma Using Inhibitors of Renalase Signaling

2 A B 6

5 ) 5 1.5 4

1 2 Relative cell viability Live cells 1 ¥ 10 1

0.5 0 BSA RNLS BSA RNLS BSA RNLS BSA RNLS No treatment BSA RNLS A375.S2 MeWo SkMel5 SkMel28

Figure 2. RNLS overexpression favors cancer cell survival. A, A375.S2, MeWo, Skmel5, and Skmel28 cells were serum-starved, treated with BSA (30 mg/mL) or rRNLS (30 mg/mL), and assessed for cell viability at 72 hours by WST-1 (n ¼ 6. , P < 0.05; , P < 0.005). B, A375.S2 cells were serum-starved for 24 hours, and then untreated or incubated with 30 mg/mL of BSA or rRNLS for 3 days; live cell number was assessed by Trypan blue assay (n ¼ 6; , P < 0.001).

RNLS overexpression favors cancer cell survival nine changed to alanine; Fig. 3D). RP220A does not mediate RNLS-mediated signaling is antiapoptotic, and protects nor- RNLS-dependent signaling, but binds to PMCA4b and antag- mal cells exposed to toxic stress from apoptotic death (23, 24). onizes the action of endogenous RNLS (7). RP-220A proved To explore whether RNLS signaling favored the survival of to be cytotoxic in increasing doses to melanoma cells in culture cancer cells, either recombinant RNLS (rRNLS) or BSA was (P < 0.005, Fig. 3D). added to serum-starved melanoma cells (A375.S2, MeWo, SkMel5, and SkMel28) in culture, and cell viability was deter- Inhibition of RNLS signaling blocks tumor growth in vivo mined. Compared with BSA, RNLS markedly increased the We injected A375.S2 (human melanoma) cells subcutane- survival of serum-starved cells, and caused an apparent increase ously into athymic nude mice to generate tumors. Once the in the proliferative rate as measured by the WST-1 assay (n ¼ 6, tumors reached a volume of approximately 50 mm3,the P < 0.05; Fig. 2A). The total cell number and percentage of live animals were then treated with either control rabbit IgG or cells of those treated with RNLS were counted to determine a RNLS-neutralizing mAb, m28-RNLS. As overall animal whether the apparent increase in proliferative rate was due an health and activity was maintained throughout the study, the increase in cell proliferation or to a decrease in the rate of antibody treatment did not appear to be toxic. Tumor size was apoptosis. As shown in Fig. 2B, treatment with RNLS showed measured every other day, and treatment with m28-RNLS increased cell counts, and increased percentage of live cells decreased tumor volume at all points tested (P < 0.05, Fig. compared with those treated with BSA, suggesting that RNLS 4A). The animals were sacrificed at day 11 due to overall tumor functions as an antiapoptotic, survival factor. size and ulceration in some animals. IHC staining of sections from the xenografted tumors with the cellular proliferation Inhibition of RNLS signaling is cytotoxic to melanoma cells in marker Ki67 revealed a significant decrease in cellular prolif- vitro eration within the tumors treated with the anti-RNLS antibody We used three approaches to determine the functional con- versus to those treated with rabbit IgG: of 35.1 2.3 positive sequences of inhibiting RNLS expression and signaling in cells/high power field in the control group versus 13.4 3.0 in melanoma. First, we evaluated the effect of decreasing RNLS the RNLS Ab-treated group (n ¼ 14, P ¼ 0.0004; Fig. 4B). To expression on cell viability. RNLS knockdown by siRNA test the efficacy of anti-RNLS therapy in immunocompetent, markedly reduced the viability of the melanoma cell lines B16f10 cells (mouse melanoma line) were subcutaneously A375.S2andSkMel28(P ¼ 0.03 and P ¼ 0.003, injected into C57BL/6 mice. Once the tumors reached a vol- respectively, Fig. 3A). Second, as the RNLS peptide RP-220 ume of approximately 500 mm3, the animals were treated with mimics the protective effect and signaling properties of rRNLS, either control rabbit IgG or a RNLS-neutralizing mAb, m28- we reasoned that it likely interacts with a critical region of the RNLS, and sacrificed at day 7 due to the very large tumor receptor for extracellular RNLS and that antibodies directed burden of the control group. As depicted in Fig. 4C, m28-RNLS against it could have inhibitory properties. Therefore, we administration caused a significant reduction in tumor volume developed a panel of mAbs against RP-220, and tested their compared with rabbit IgG. effect on cancer cell survival. Two mAbs generated against RNLS, [clones # 28-4 (m28-RNLS), 37-10 (m37-RNLS)] Inhibition of RNLS signaling blocks endogenous RNLS decreased the viability of all (total of 5) melanoma cell lines expression and STAT3 activation and induces apoptosis and tested, and representative examples are shown in Fig. 3B and C. cell-cycle arrest m28-RNLS demonstrated increasing levels of cytotoxicity in STAT3 is known to bind to the promoter region of the RNLS correlation with increasing treatment concentrations (P < gene and increase its expression, and a positive RNLS-STAT3 0.05, Fig. 3B). Third, we generated a peptide antagonist (RP- feedback loop has been suggested (8). This relationship was 220A) by decreasing the net charge of RP-220 (3 Lysine/argi- further investigated through IF tissue staining and study of

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 3887

Hollander et al.

B 1 A 1 0.9 0.9 1.2 0.8 0.8 1 0.7 0.7 0.8 0.6 0.6 0.6 0.5 0.5 0.4 0.4 Control rabbit lgG 0.4 0.2 m28-RNLS 0.3

0.3 A375.S2 Cell viability Relative 0 Relative cell viability 0.2

Relative cell viability 10 20 50 100 0.2 0.1 Antibody dose (mg/mL) 0.1 0 0 Control siRNA RNLS siRNA Control siRNA RNLS siRNA A375.S2 SkMel28 m28-RNLS m37-RNLS m28-RNLS m37-RNLS Control rbb lgG Control rbb lgG A375.S2 SkMel28 C A375.S2 SkMel28 SkMel5 D hRenalase1 (342aa) 1234 67 9

RP-220 CIRFVSIDNKKRNIESSEIG RP-220A CIRFVSIDNAAANIESSEIG

1.2

0.8

0.6

Control BSA 0.4 RP220A

Relative cell viability 0.2

0 0 50 100 150 200 Concentration (mg/mL)

Figure 3. Inhibition of RNLS signaling is cytotoxic to melanoma cells. A, transient transfection of A375.S2 and SK-Mel-28 with RNLS-speciﬁc siRNA, or control siRNA, and cell viability assessed at 72 hours by WST-1 (n ¼ 6; , P ¼ 0.03; , P ¼ 0.003). B, cells treated with indicated antibodies for 72 hours and cell viability assessed by WST-1; left, m28-RNLS and m37-RNLS: RNLS monoclonals; right, A375.S2 treated with m28-RNLS for 72 hours (n ¼ 6; , P < 0.05; , P < 0.005). C, representative photos of A375.S2, SkMel28, SkMel5 after 72-hour incubation with rabbit IgG or m28-RNLS. D, amino acid (AA) sequence of RNLS peptide antagonist (RP220A). A375.S2 cells treated as indicated, cell viability assessed by WST-1 at 72 hours (n ¼ 6; , P < 0.005).

the cell lysates from the xenografted tumors treated with suggesting that tumor-infiltrating cells play a key role in RNLS control IgG and m28-RNLS. Significant coexpression of RNLS production and secretion. with phosphorylated and total STAT3 was noted in the tumor In addition, we noted increased expression of the cell-cycle samples (Fig. 5A) as assessed by IF. Treatment with m28-RNLS inhibitor p21. Antibody treatment markedly increased the expres- caused a dramatic reduction in RNLS protein expression and in sion of the cell-cycle regulator p21 in the tumor samples: 24.2 both total and phosphorylated STAT3 (Fig. 5A). Changes in 2.4 positive cells in the Ab-treated group vs. 12.2 1.0 in the protein expression were confirmed by Western blot analysis as control group (n ¼ 14, P ¼ 0.009; Fig. 5E). TUNEL staining showninFig.5BandC.Intumorstreatedwithm28-RNLS, revealed a significant increase in the average number of cells STAT3 phosphorylation at tyrosine 705 (p-Y705-STAT3) and undergoing apoptosis in the Ab-treated tumors over the control total STAT3 were significantly decreased (n ¼ 8, P < 0.005; Fig. group with an average of 13.3 0.6 positive cells versus 4.3 0.2 5B and C). (n ¼ 14, P < 0.001; Fig. 5E). The increase in apoptosis was To test whether the significant decrease in RNLS expression was temporally related to phopshorylation of p38 MAPK, and sub- primarily occurring in the melanoma cells, we used human and sequent activation of the B-cell lymphoma 2–related protein Bax mouse-specific primers to amplify tumor (human) and endoge- (Fig. 5F). These data indicate that treatment with anti-RNLS nous (mouse) RNLS in the tumor mass. As depicted in Fig. 5D, antibody causes a marked reduction in total and phosphorylated treatment with m28-RNLS causes a significant reduction in mouse STAT3, decreases cell proliferation, and increases apoptosis in RNLS expression, without affecting human (tumor) expression, tumor cells.

3888 Cancer Res; 76(13) July 1, 2016 Cancer Research

Treatment of Melanoma Using Inhibitors of Renalase Signaling

A 500 B

Rabbit lgG Control: rabbit IgG Treated: m28-RNLS 400 ) 3

300

200 Tumor volume (mm Tumor 100 m28-RNLS

0 02486 10 12 Time (days)

Figure 4. C 6,000 Inhibition of RNLS signaling blocks melanoma growth in vivo. A, nude mice xenografted with A375.S2 cells; tumor size measured 5,000 )

3 prior to treatment every 3 days with 2 mg/kg of rabbit IgG or n ¼ P < 4,000 m28-RNLS ( 14 per group; , 0.05). B, representative IHC stains for Ki67 (brown) in A375.S2-xenografted 3,000 tumors (n ¼ 14 each) treated with m28-RNLS or rabbit IgG. C, C57BL/6 mice were xenografted with B16-F10 2,000 cells and treated every 3 days with 2 mg/kg rabbit IgG n ¼ n ¼ ﬁ

Tumor volume (mm Tumor ( 5) or m28-RNLS ( 5); animals were sacri ced 1,000 at day 7 due to large tumor burden in the control group (, P < 0.025). 0 Day 0 Day 7 Day 0 Day 7 Rabbit IgG m28-RNLS

Plasma membrane Ca ATPase PMCA4b mediates RNLS- ERK activation by RNLS was most notable at the 10-minute time dependent STAT3 and ERK1/2 phosphorylation in B16-F10 point. PMCA4b-targeting siRNAs decreased gene expression by melanoma cells 79% 2% (n ¼ 4), and reduced RNLS-dependent ERK1/2 We have previously identified PMCA4b as a receptor for phosphorylation by approximately 95% (Fig. 6F, right). extracellular RNLS (7). PMCA4b is overexpressed in colon and In contrast to what has been observed in B16F10 (Fig. 6E), breast cancer, and its catalytic domain interacts with the proa- Panc1 pancreatic cancer cells (15), and the human kidney cell HK- poptotic tumor suppressor Ras-associated factor 1 (RASSF1) 2 (28), RNLS does not activate STAT3 in Raw 264.7 macrophage to inhibit EGF-dependent ERK phosphorylation (25–27). cells (Fig. 6F). To test whether RNLS signaled via STAT3 in PMCA4b expression in human melanoma cell lines was con- nontransformed macrophages, we treated bone marrow–derived firmed by Western blotting and qPCR (Fig. 6A and B). To study (BMDM) nonpolarized macrophages (M) with rRNLS. As the interaction of PMCA4b and RNLS in melanoma, we tested shown in Fig. 6G, RNLS-mediated STAT3 phosphorylation was for coexpression of both proteins in B16-F10 xenografts. Data not detected in BMDMs, M suggesting that in contrast to other obtained from control animals (Fig. 6C, top) indicate high cells, RNLS does not activate the STAT3 pathway in macrophages. expression of PMCA1/4, along with significant coexpression þ with RNLS. In animals treated with m28-RNLS, the expression Inhibition of RNLS signaling increases the ratio of CD86 to þ of both proteins is diminished, as is the extent of coexpression CD163 TAMs (Fig. 6C, bottom). B16-F10 cells grown in serum-free medium The melanocytes did not appear to be the main source of the express and secrete RNLS (Fig. 6D). RNLS in the melanoma histospots, as there was minimal overlap To examine the role of PMCA4b in RNLS-dependent signaling noted between RNLS and melanocyte staining (Fig. 1A). Mela- in B16-F10 cells, we decreased its expression using siRNA and nomas often have significant infiltration of immune cells, includ- compared RNLS-mediated ERK and STAT3 signaling. In control ing macrophages. The infiltrating macrophages appeared to con- studies, nontargeting siRNAs affected neither PMCA4b gene tribute the majority of the tumoral RNLS as a substantial com- expression (assessed by qPCR), nor RNLS-mediated ERK or STAT3 ponent of the RNLS staining noted in each histospot overlapped phosphorylation as evidenced by the expected transient activation significantly with the pan-macrophage marker CD68 (Fig. 7A, most notable at the 5-minute time point (Fig. 6E, left). In contrast, top). Upon further investigation, we determined that RNLS was þ PMCA4b-targeting siRNAs decreased gene expression by 85% coexpressed predominantly with CD163 (M2-like) TAMs (Fig. þ 4% (n ¼ 3), and reduced RNLS-dependent ERK and STAT3 7A, middle). Coexpression of RNLS with CD86 (M1-like) þ phosphorylation by approximately 90% (Fig. 6E, right). macrophages was minimal (Fig. 7A, bottom). M2-like (CD163 ) To examine the role of PMCA4b in RNLS-mediated signaling in macrophages are associated with immune escape and shown to þ macrophages, we used siRNA to decrease its expression in Raw promote cancer development and spread, while M1-like (CD86 ) 264.7, a macrophage cell line derived from mouse blood. In macrophages are typically proinflammatory and inhibit tumor control studies, nontargeting siRNAs affected neither PMCA4b growth (29, 30). gene expression (assessed by qPCR), nor RNLS-mediated ERK We have examined RNLS expression in macrophages infiltrat- phosphorylation. As shown in Fig. 6F (left), the expected transient ing tumor xenografts in nude mice. Fig. 7B is a representative

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 3889

Hollander et al.

Figure 5. Inhibition of RNLS signaling blocks RNLS expression and STAT3 activation, induces apoptosis, and cell-cycle arrest. A, xenografts were treated with rabbit IgG or m28-RNLS and probed by IF; representative results, blue, nuclei; green, RNLS; and red, phospho STAT3 (left) or total STAT3 (right). B, as in A, tumor cell lysates probed for RNLS, pSTAT3, totalSTAT3, p21 by Western blot analysis (representative results). C, quantiﬁcation of STAT3 protein expression in samples shown in B; p-Y705-STAT3 signals were normalized to total STAT3, and total STAT3 signals were normalized to protein loading measurements (n ¼ 3; , P < 0.05; , P < 0.005). D, xenograft tumors (n ¼ 14 each) treated with rabbit IgG or m28-RNLS, probed for human and mouse RNLS expression by qPCR (, P < 0.05). E, IHC images of A375.S2–xenografted tumors (n ¼ 14 each) for TUNEL assay or p21 (brown) treated with m28-RNLS or rabbit IgG. F, A375.S2 cells treated with anti-RNLS antibody or goat IgG; time course of p38 phosphorylation and Bax expression by Western blot analysis; phospho STAT3, p-Y705-STAT3; p-p38, phosphorylated p38; Bax, bcl-2 like protein 4.

þ picture of RNLS expression in CD68 cells (pan-macrophage polarized macrophages (M). As shown in Fig. 7D, the addition marker) in the tumor mass (n ¼ 4) of control animals treated of B16f10 CM stimulated an 8-fold increase RNLS expression in with rabbit IgG. Manual count of cells expressing RNLS, CD68, or M compared with control, indicating that factor(s) secreted by both indicates that 60% of cells expressing RNLS also express melanoma markedly upregulate RNLS in macrophages. As shown þ CD68, and that 25% of CD68 cells express RNLS. Treatment of in Fig. 6F, our data suggest that RNLS activates ERK1/2, but not the xenografts with m28-RNLS antibody led to a marked decrease STAT3, in an autocrine fashion in M. To examine the potential þ in the number of CD163 TMAs, and the remaining cells did not phenotypic consequence of RNLS expression in M, we com- express detectable levels of RNLS (Fig. 7C). pared the response of WT- and RNLS-deficient M to two stimuli Although catecholamines and nicotine have been shown to [IL6 and Leukemia inhibitory factor (LIF)] reported to polarize upregulate RNLS gene expression in epithelial cells (8, 31), we M toward a tumor-promoting M2d phenotype (32). We find have no knowledge of the signals that affect RNLS expression in that IL6 and LIF stimulated IL10 gene expression significantly macrophages. We hypothesized that tumor cells could modulate more in WT than in KO macrophages (Fig. 7E). RNLS expression by macrophage, and that RNLS could facilitate macrophage polarization toward a M2d-like, tumor-promoting Discussion phenotype. This was tested by examining the effect of conditioned media (CM) obtained from cultured murine melanoma cells RNLS expression is markedly increased in melanoma cell lines (B16f10) on RNLS expression by bone marrow–derived, non- and tumor samples. In patients with metastatic melanoma, RNLS

3890 Cancer Res; 76(13) July 1, 2016 Cancer Research

Treatment of Melanoma Using Inhibitors of Renalase Signaling

A BC 120 MeWo A375.S2 HK-2 WM-266-4 SkMel28 100 217 K- PMCA4b

80 PMCA4b 123 K- 60

40 (normalized to 18S)

Relative RNA expression Relative 20

48 K- 0 A375.S2 SkMel28 MeWo WM266-4

µL C.M. of B16-F10 in DMEM

10 µL DMEM-10%10 µL FBSDMEM0.1 µL C.M. 1.3of B16-F10 13 in µ LDMEM C.M. of B16-F10 in DMEM

Control siRNA PMCA4b siRNA 75 K- -RNLS F pERK1/2 705 50 K- P-Y -STAT3

-RNLS 37 K- GAPDH RNLS + +++ Time (min) 10 60 10 60

15 K-

EGP-Y705-STAT3 Control siRNA PMCA4b siRNA TOTAL STAT3

pERK1/2 BSA – ++– ––– RNLS – ––++ –– 705 P-Y -STAT3 IL-6 –– ––– ++ Time (min) 0110 600 60 10 60 GAPDH RNLS ––+++ + ++ Time (min) 55305 120 530120

Figure 6. Extracellular RNLS signals via PMCA4b and phosphorylates ERK1/2 and STAT3 in B16-F10 melanoma cells. A, lysates from melanoma cell lines A375.S2, SkMel28, MeWo, WM266-4, and human kidney 2 positive control (HK-2) cells probed for PMCA4b by Western blot analysis; representative results (n ¼ 3). B, lysates from melanoma cell lines A375.S2, SkMel28, MeWo, WM266-4 probed for PMCA4b by qPCR and mRNA levels normalized to 18S expression (n ¼ 3). C, B16-F10 xenograft tumors treated with rabbit IgG (top) or m28-RNLS (bottom), probed for RNLS and PMCA4 (blue, nuclei; green, PMCA4; red, RNLS); representative result, n ¼ 5. D, samples of DMEM with FBS, serum-free DMEM, and serum-free medium after incubation with B16-F10 cells for 72 hours were collected and probed for RNLS by Western blot analysis; representative results. E, B16-F10 cells were transfected with control or PMCA4b siRNA, treated with BSA (30mg/mL) or rRNLS (30mg/mL) for the indicated time, and lysates probed by Western blot analysis; representative results; n ¼ 3. F, raw 267.4 blood macrophage cells were transfected with control or PMCA4b siRNA, treated with rRNLS (30 mg/mL) for indicated time, and cell lysates probed as in E; representative study, n ¼ 3. G. mouse BMDMs were treated with BSA (30 mg/mL), rRNLS (30 mg/mL), or IL6 (50 ng/mL) and cell lysates probed by Western blot analysis; representative study, n ¼ 3. CM, conditioned medium; pERK1/2, phosphorylated ERK; P-Y705-STAT3, phosphorylated STAT3 at tyrosine 705.

expression is inversely correlated with disease-specific survival. isms that underlie the antitumor effects observed with inhibition Examination of the pattern of expression of RNLS in melanoma of RNLS signaling. Inhibition of RNLS signaling by the RNLS þ þ suggests that upregulation predominantly occurs in the cellular monoclonal m28-RNLS increases the ratio of CD86 to CD163 þ þ components of tumor-associated stroma, specifically in CD163 TAMs, and decreases RNLS secretion by CD163 TAMs. RNLS is macrophages. Experimental data indicate that alternatively acti- expressed in other cells and tissues and circulates in plasma. m28- þ vated macrophages (M2-like, CD163 ) recruited into the tumors RNLS has a direct effect on melanoma cells by inhibiting RNLS suppress the immune response against the tumor, increase angio- signaling. The net result is a dramatic fall in total and phosphor- genesis, and facilitate tumor cell migration, invasion, and dis- ylated STAT3, leading to apoptosis of melanoma cells. semination (33–35). TAMs account for a significant percentage of The regulatory promoter elements and transcription factors the tumor mass in human melanoma, and also in the xenograft that regulate RNLS gene expression have been recently investi- model described in this work. We find that RNLS is preferentially gated (8) and these data point to a key role for STAT3. Our results þ expressed in CD163 TAMs, suggesting that M2-like TAMs could suggest the existence of a feedforward loop between RNLS and facilitate tumor progression by secreting RNLS, which in turn STAT3, in which signals that upregulate STAT3 increase RNLS gene would act in a paracrine fashion to promote tumor growth. Figure expression, which in turn increases STAT3 activity. The existence 7F illustrates a working model that incorporates the key mechan- of such an interaction between RNLS and STAT3 has important

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 3891

Hollander et al.

ACControl: rabbit lgG Treated: m28-RNLS

D 9 8 7 6 5 4 B 3 2 1 (Fold, normalized to 18S rRNA) (Fold, Relative RNLS mRNA expression Relative 0 BMDB BMDB + B16f10 CM

E 2.5 KO F CD163+ TAM CD86+ TAM WT 2

1.5 m28-RNLS

RNLS 1 CD86+/CD163+ RNLS RNLS m28-RNLS 0.5 m28-RNLS t-STAT3 (normalized to GAPDH)

Melanoma IL10 mRNA expression Relative 0 IL6 LIF Stimulus Apoptosis

Figure 7. RNLS expressed in CD163þ TAMs in melanoma. A, human melanoma examined by IF for coexpression of RNLS and pan-macrophage marker CD68 (top); blue, nuclei; green, RNLS; and red, all macrophages; DAPI, nuclear stain; RNLS-CD68, merged RNLS and CD68 stains. Middle, melanoma examined by IF for RNLS and M2 marker CD163; blue, nuclei; green, RNLS; red, M2 macrophages; DAPI, nuclear stain; RNLS-CD163, merged RNLS, CD163 stains. Bottom, IF of melanoma for RNLS and M1 marker CD86; blue, nuclei; green, RNLS; red, M1 macrophages; DAPI, nuclear stain; RNLS-CD163, merged RNLS, CD86. B, xenograft in nude mice treated with rabbit IgG and probed for RNLS and macrophages (CD68þ cells) by IF (n ¼ 4); representative results, blue, nuclei; green, macrophages; red, RNLS. C, xenograft treated with rabbit IgG or m28-RNLS, probed for RNLS and M2 TAMs (CD163þ cells) by IF; representative result; green, M2 macrophages; red, RNLS. D, relative RNLS mRNA level by qPCR in BMDMs with or without conditioned medium from B16f10 (BMDM and BMDM þ B16f10 CM, respectively; n ¼ 4; , P < 0.0001). E, relative mRNA levels of IL10 by qPCR in macrophages from WT (n ¼ 4) or RNLS KO mice (n ¼ 4) with indicated stimuli ( and , P < 0.02, 0.0001, respectively). F, proposed mechanism of action of m28-RNLS. Inhibition of RNLS signaling increases CD86þ to CD163þ TAM ratio, decreases RNLS secretion from CD163þ TAMs, and inhibits RNLS signaling in melanoma, leading to signiﬁcant reduction in total and phosphorylated STAT3 and apoptosis. TAM, tumor- associated macrophages; CD163, alternatively activated macrophage (M2) marker; CD86, classically activated macrophage (M1) marker; m28-RNLS, antirenalase mAb; t-STAT3, total STAT3; p-STAT3, phosphorylated STAT3.

implications regarding the role of RNLS signaling in the patho- ment. A STAT3 feed-forward loop between cancer cells and non- genesis of cancer. Indeed, there are extensive data pointing to a key transformed and stromal cells has been documented in cancer role for the STAT family proteins, particularly STAT3, in the (37–39). For instance, STAT3 is constitutively activated in mul- induction and maintenance of an inﬂammatory microenviron- tiple myeloma patients. In the IL6-dependent human myeloma ment that facilitates malignant transformation and cancer pro- cell line U266, IL6 signals through Janus kinases to activate STAT3, gression (36). STAT3 signaling is often persistently activated in which in turn upregulates antiapoptotic factors, and promotes malignant cells, and such activation not only drives tumor cell the survival of tumor cells (37). Through various mechanisms, proliferation, but also increases the production of a large number STAT3 has also been found to be constitutively activated in a of genes that sustain inﬂammation in the tumor microenviron- majority of melanomas leading to increases in tumor cell survival,

3892 Cancer Res; 76(13) July 1, 2016 Cancer Research

Treatment of Melanoma Using Inhibitors of Renalase Signaling

proliferation, metastasis, angiogenesis, and decreases in tumor framework to further investigate the use of anti-RNLS therapy for immune response (11, 40–42). the treatment of malignant melanoma, alone or in conjunction RNLS protects against ischemic renal and cardiac injury (21, 43, with other TAM- or melanoma-inhibiting drugs, such as CSF-1R 44). The promoter region of the RNLS gene contains four poten- inhibitors or MAPK pathway inhibitors, respectively. Because tial hypoxia-inducible factor-1 alpha (HIF-1a) response ele- there are multiple mechanisms for regulating MAPK and PI3K ments, and has been shown to represent novel target gene of and JAK/STAT3 and as there is crosstalk between pathways, cell HIF-1a (44). These data raise the intriguing possibility that certain fate depends on the dynamic balance and integration of multiple physiologic and pathophysiologic effects of HIF-1a in normal signals, and our data suggest that RNLS inhibition will tilt the cells may be mediated by RNLS, and this could also be relevant to balance toward cancer cell death. cancer pathogenesis and progression. RNLS mediates cytoprotection by increasing the antiapoptotic Disclosure of Potential Conflicts of Interest factor Bcl2, and preventing the activation of effector caspases (6). G. Desir has ownership interest (including patents) in Personal Therapeutics, As RNLS is an NADH oxidase (1), it is possible that its protective Bessor Pharma, issued patents, Yale University and is a consultant/advisory action is in part mediated by its enzymatic properties (6, 21, 24). board member for Personal Therapeutics and Bessor Pharma. No potential Inhibition of RNLS signaling in A375.S2 cells is associated with conflicts of interest were disclosed by the other authors. sustained activation of p38 MAPK, followed by activation of the apoptotic factor Bax, and apoptosis. The MAPK p38 is a stress- Authors' Contributions activated protein kinase that has been implicated in inflam- Conception and design: L. Hollander, X. Guo, H. Velazquez, R. Safirstein, mation, cell differentiation, cell-cycle regulation, and apoptosis C. Cha, G. Desir (45). For instance, nerve growth factor withdrawal was shown Development of methodology: L. Hollander, X. Guo, C. Cha, G. Desir Acquisition of data (provided animals, acquired and managed patients, to cause apoptosis following sustained activation of JNK and provided facilities, etc.): L. Hollander, X. Guo, H. Velazquez, H.M. Kluger, p38, and downregulation of ERK (46). However, as under G. Desir certain conditions inhibition of p38 can block apoptosis Analysis and interpretation of data (e.g., statistical analysis, biostatistics, (45), p38's role in apoptosis is clearly context dependent. Our computational analysis): L. Hollander, H. Velazquez, J. Chang, R. Safirstein, data suggest that in A375.S2 cells, RNLS-dependent activation H.M. Kluger, C. Cha, G. Desir of p38 causes apoptosis. Writing, review, and/or revision of the manuscript: L. Hollander, J. Chang, R. Safirstein, H.M. Kluger, C. Cha, G. Desir Inhibition of RNLS signaling markedly decreases the expression Administrative, technical, or material support (i.e., reporting or organizing of Ki-67 in xenographs of melanoma, and we interpret the data as data, constructing databases): L. Hollander, C. Cha indicating that RNLS signaling is a key driver of tumor prolifer- Study supervision: L. Hollander, R. Safirstein, C. Cha, G. Desir ation, and that RNLS inhibition decreases the proliferative rate of tumors. Many of the key factors that determine cell-cycle progres- Acknowledgments fi sion have been identi ed, and include a set of cyclin-dependent The authors thank Dr. Ruth Halaban for helpful comments and suggestions. kinases (CDK) along with two classes of CDK inhibitors (47). The expression of p21, a CDK inhibitor belonging to the CIP/KIP family, is regulated by RNLS signaling. Inhibition of RNLS sig- Grant Support naling is associated with a marked increase in p21 expression. p21 This work was supported in part by VA Connecticut (H. Velazquez, is a negative regulator of cell cycle that can maintain cells in G0, R. Safirstein, and G.V. Desir), NIH grants RC1DK086465, RC1DK086402, block G1–S transition, and cause G1 or inter-S-phase arrest (47). and R01DK081037 (G.V. Desir), National Cancer Institute research grant Therefore, the increase in p21 expression could account for the CA-16359, and the Lampman surgical award from Yale University Department decrease in cell proliferation observed in tumors treated with an of Surgery (C. Cha). The costs of publication of this article were defrayed in part by the payment of anti RNLS antibody. In addition, p38 has also been shown to page charges. This article must therefore be hereby marked advertisement in affect cell-cycle progression (45), and activation of p38 by anti- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. RNLS treatment could also contribute to cell-cycle arrest. Our findings identify RNLS as a secreted protein that can Received June 3, 2015; revised March 16, 2016; accepted April 5, 2016; promote the survival and growth of cancer cells, and provide a published online July 1, 2016.

References 1. Farzaneh-Far R, Desir GV, Na B, Schiller NB, Whooley MA. A functional 5. Li G, Xu J, Wang P, Velazquez H, Li Y, Wu Y, et al. Catecholamines regulate polymorphism in renalase (Glu37Asp) is associated with cardiac hyper- the activity, secretion, and synthesis of renalase. Circulation 2008;117: trophy, dysfunction, and ischemia: data from the heart and soul study. 1277–82. PLoS ONE 2010;5:e13496. 6. Wang L, Velazquez H, Moeckel G, Chang J, Ham A, Lee HT, et al. Renalase 2. Desir GV, Tang L, Wang P, Li G, Sampaio Maia B, Quelhas-Santos J, et al. prevents AKI independent of amine oxidase activity. J Am Soc Nephrol Renalase lowers ambulatory blood pressure by metabolizing circulating 2014;25:1226–35. catecholamines. J Am Heart Assoc 2012;1:e002634 7. Wang L, Velazquez H, Chang J, Saﬁrstein R, Desir GV. Identiﬁcation 3. Desir GV, Wang L, Peixoto AJ. Human renalase: a review of its biology, of a receptor for extracellular renalase. PLoS ONE 2015;10: function, and implications for hypertension. J Am Soc Hypertens 2012;6: e0122932. 417–26. 8. Sonawane PJ, Gupta V, Sasi BK, Kalyani A, Natarajan B, Khan AA, et al. 4. Xu J, Li G, Wang P, Velazquez H, Yao X, Li Y, et al. Renalase is a novel, Transcriptional regulation of the novel monoamine oxidase renalase: soluble monoamine oxidase that regulates cardiac function and blood crucial roles of transcription factors Sp1, STAT3 and ZBP89. Biochemistry pressure. J Clin Invest 2005;115:1275–80. 2014;53:6878–92.

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 3893

Hollander et al.

9. Gray-Schopfer V, Wellbrock C, Marais R. Melanoma biology and new 27. Brini M, Cal T, Ottolini D, Carafoli E. The plasma membrane calcium targeted therapy. Nature 2007;445:851–57. pump in health and disease. FEBS J 2013;280:5385–97. 10. Lowe GC, Saavedra A, Reed KB, Velazquez AI, Dronca RS, Markovic SN, 28. Guo X, Hollander L, MacPherson D, Wang L, Velazquez H, Chang J, et al. et al. Increasing incidence of melanoma among middle-aged adults: an Inhibition of renalase expression and signaling has antitumor activity in epidemiologic study in Olmsted County, Minnesota. Mayo Clinic Proc pancreatic cancer. Sci Rep 2016;6:22996. 2014;89:52–59. 29. Biswas SK, Mantovani A. Macrophage plasticity and interaction with 11. Lesinski GB.The potential for targeting the STAT3 pathway as a novel lymphocyte subsets: cancer as a paradigm. Nat Immunol 2010;11:889–96. therapy for melanoma. Future Oncol 2013;9:925–7. 30. Mantovani A, Sozzani S, Locati M, Allavena P, Sica A. Macrophage polar- 12. Yajima I, Kumasaka MY, Thang ND, Goto Y, Takeda K, Yamanoshita O, ization: tumor-associated macrophages as a paradigm for polarized M2 et al. RAS/RAF/MEK/ERK and PI3K/PTEN/AKT signaling in malignant mononuclear phagocytes. Trends Immunol 2002;23:549–55. melanoma progression and therapy. Dermatol Res Pract 2012;2012: 31. Li G, Xu J, Wang P, Velazquez H, Li Y, Wu Y, et al. Catecholamines regulate 354191. the activity, secretion and synthesis of renalase. Circulation 2008;117: 13. Dankort D, Curley DP, Cartlidge RA, Nelson B, Karnezis AN, Damsky WEJr, 1277–82. et al. Braf(V600E) cooperates with Pten loss to induce metastatic mela- 32. Duluc D, Delneste Y, Tan F, Moles M-P, Grimaud L, Lenoir J, et al. Tumor- noma. Nat Genet 2009;41:544–52. associated leukemia inhibitory factor and IL-6 skew monocyte differentiation 14. Desir GV, Tang L, Wang P, Li G, Sampaio-Maia B, Quelhas-Santos J, et al. into tumor-associated macrophage-like cells. Blood 2007;110:4319–30. Renalase lowers ambulatory blood pressure by metabolizing circulating 33. Ruhrberg C, De Palma M. A double agent in cancer: Deciphering macro- adrenaline. J Am Heart Assoc 2012;1:e002634. phage roles in human tumors. Nat Med 2010;16:861–62. 15. Guo X, Hollander L, MacPherson D, Wang L, Velazquez H, Chang J, et al. 34. Pollard JW.Tumour-educated macrophages promote tumour progression Inhibition of renalase expression and signaling has antitumor activity in and metastasis. Nat Rev Cancer 2004;4:71–78. pancreatic cancer. Sci Rep 2016;6:22996. 35. Hao N-B, Lu€ M-H, Fan Y-H, Cao Y-L, Zhang Z-R, Yang S-M. Macrophages in 16. Guo X, Schmitz JC, Kenney BC, Uchio EM, Kulkarni S, Cha CH. Intermedin tumor microenvironments and the progression of tumors. Clin Develop is overexpressed in hepatocellular carcinoma and regulates cell prolifera- Immunol 2012;2012:11. tion and survival. Cancer Sci 2012;103:1474–80. 36. Yu H, Pardoll D, Jove R. STATs in cancer inflammation and immunity: a 17. Berger AJ, Kluger HM, Li N, Kielhorn E, Halaban R, Ronai Z, et al. leading role for STAT3. Nat Rev Cancer 2009;9:798–809. Subcellular localization of activating transcription factor 2 in melanoma 37. Catlett-Falcone R, Landowski TH, Oshiro MM, Turkson J, Levitzki A, Savino specimens predicts patient survival. Cancer Res 2003;63:8103–7. R, et al. Constitutive activation of Stat3 signaling confers resistance to 18. Rimm DL, Camp RL, Charette LA, Costa J, Olsen DA, Reiss M. Tissue apoptosis in human U266 myeloma cells. Immunity 1999;10:105–15. microarray: a new technology for amplification of tissue resources. Cancer J 38. Yu H, Kortylewski M, Pardoll D. Crosstalk between cancer and immune 2001;7:24–31. cells: role of STAT3 in the tumour microenvironment. Nat Rev Immunol 19. Berger AJ, Camp RL, Divito KA, Kluger HM, Halaban R, Rimm DL. 2007;7:41–51. Automated quantitative analysis of HDM2 expression in malignant mel- 39. Ara T, Song L, Shimada H, Keshelava N, Russell HV, Metelitsa LS, et al. anoma shows association with early-stage disease and improved outcome. Interleukin-6 in the bone marrow microenvironment promotes the growth Cancer Res 2004;64:8767–72. and survival of neuroblastoma cells. Cancer Res 2009;69:329–37. 20. Nicholson AD, Guo X, Sullivan CA, Cha CH. Automated quantitative 40. Kortylewski M, Jove R, Yu H. Targeting STAT3 affects melanoma on analysis of tissue microarray of 443 patients with colorectal adenocarci- multiple fronts. Cancer Metastasis Rev 2005;24:315–27. noma: low expression of Bcl-2 predicts poor survival. J Am Coll Surg 41. Emeagi PU, Maenhout S, Dang N, Heirman C, Thielemans K, Breckpot K. 2014;219:977–87. Downregulation of Stat3 in melanoma: reprogramming the immune 21. Wu Y, Xu J, Velazquez H, Wang P, Li G, Liu D, et al. Renalase deficiency microenvironment as an anticancer therapeutic strategy. Gene Ther 2013; aggravates ischemic myocardial damage. Kidney Int 2011;79:853–60. 20:1085–92. 22. Gould Rothberg BE, Berger AJ, Molinaro AM, Subtil A, Krauthammer MO, 42. Yang CH, Fan M, Slominski AT, Yue J, Pfeffer LM. The role of constitutively Camp RL, et al. Melanoma prognostic model using tissue microarrays and activated STAT3 in B16 melanoma cells. Int J Interferon Cytokine Mediat genetic algorithms. J Clin Oncol 2009;27:5772–80. Res 2010;2010:1–7. 23. Wang L, Velazquez H, Moeckel G, Chang J, Ham A, Lee HT, et al. Renalase 43. Guo X, Wang L, Velazquez H, Safirstein R, Desir GV. Renalase: its role as a prevents AKI independent of amine oxidase activity. J Am Soc Nephrol cytokine, and an update on its association with type 1 diabetes and 2014;25:1226–35. ischemic stroke. Curr Opin Nephrol Hypertens 2014;23:513–8. 24. Lee HT, Kim JY, Kim M, Wang P, Tang L, Baroni S, et al. Renalase protects 44. Du M, Huang K, Huang D, Yang L, Gao L, Wang X, et al. Renalase is a novel against ischemic AKI. J Am Soc Nephrol 2013;24:445–55. target gene of hypoxia-inducible factor-1 in protection against cardiac 25. Armesilla AL, Williams JC, Buch MH, Pickard A, Emerson M, Cartwright EJ, ischaemia–reperfusion injury. Cardiovasc Res 2014;105:182–91. et al. Novel functional interaction between the plasma membrane Ca2þ 45. Ono K, Han J. The p38 signal transduction pathway activation and pump 4b and the proapoptotic tumor suppressor Ras-associated Factor 1 function. Cell Signal 2000;12:1–13. (RASSF1). J Biol Chem 2004;279:31318–28. 46. Xia Z, Dickens M, Raingeaud J, Davis RJ, Greenberg ME. Opposing effects of 26. Agathanggelou A, Cooper WN, Latif F. Role of the ras-association domain ERK and JNK-p38 MAP kinases on apoptosis. Science 1995;270:1326–31. family 1 tumor suppressor gene in human cancers. Cancer Res 2005;65: 47. Jung Y-S, Qian Y, Chen X. Examination of the expanding pathways for the 3497–508. regulation of p21 expression and activity. Cell Signal 2010;22:1003–12.

3894 Cancer Res; 76(13) July 1, 2016 Cancer Research

Renalase Expression by Melanoma and Tumor-Associated Macrophages Promotes Tumor Growth through a STAT3-Mediated Mechanism

Lindsay Hollander, Xiaojia Guo, Heino Velazquez, et al.

Cancer Res 2016;76:3884-3894. Published OnlineFirst May 9, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-15-1524

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2016/06/21/0008-5472.CAN-15-1524.DC1

Cited articles This article cites 47 articles, 15 of which you can access for free at: http://cancerres.aacrjournals.org/content/76/13/3884.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/76/13/3884.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://cancerres.aacrjournals.org/content/76/13/3884. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.