key: cord-102977-yci9kq6x
authors: Liu, Haiming; Luo, Jiaohua; Guillory, Bobby; Chen, Ji-an; Zang, Pu; Yoeli, Jordan K.; Hernandez, Yamileth; Lee, Ian (In-gi); Anderson, Barbara; Storie, Mackenzie; Tewnion, Alison; Garcia, Jose M.
title: GHSR-1a is not Required for Ghrelin’s Anti-inflammatory and Fat-sparing Effects in Cancer Cachexia
date: 2019-12-06
journal: bioRxiv
DOI: 10.1101/866376
sha: 
doc_id: 102977
cord_uid: yci9kq6x

Adipose tissue (AT) atrophy is a hallmark of cancer cachexia contributing to increased morbidity/mortality. Ghrelin has been proposed as a treatment for cancer cachexia partly by preventing AT atrophy. However, the mechanisms mediating ghrelin’s effects are incompletely understood, including the extent to which its only known receptor, GHSR-1a, is required for these effects. This study characterizes the pathways involved in AT atrophy in the Lewis Lung Carcinoma (LLC)-induced cachexia model and those mediating the effects of ghrelin in Ghsr+/+ and Ghsr−/− mice. We show that LLC causes AT atrophy by inducing anorexia, and increasing AT inflammation, thermogenesis and energy expenditure. These changes were greater in Ghsr−/−. Ghrelin administration prevented LLC-induced anorexia only in Ghsr+/+, but prevented WAT inflammation and atrophy in both genotypes, although its effects were greater in Ghsr+/+. LLC-induced increases in BAT inflammation, WAT and BAT thermogenesis, and energy expenditure were not affected by ghrelin. In conclusion, ghrelin ameliorates WAT inflammation, fat atrophy and anorexia in LLC-induced cachexia. GHSR-1a is required for ghrelin’s orexigenic effect but not for its anti-inflammatory or fat-sparing effects.

Every year, over 1,500,000 individuals in the US are diagnosed with cancer. Cachexia (involuntary 63 loss of muscle and adipose tissue) is present in up to 80% of cancer patients, is strongly associated 64 with higher morbidity and mortality, and is reported as the direct cause of death in 20-40% of these 65 patients (Dewys, Begg et al., 1980 , Fearon, Strasser et al., 2011 . Adipose tissue, once considered 66 only a high-energy fuel reserve, has emerged recently as an active metabolic organ modulating 67 inflammation, energy expenditure and food intake in non-cancer settings (You & Nicklas, 2006) . 68 Accelerated loss of adipose tissue plays an important role in cancer cachexia contributing 69 significantly to the increased morbidity and mortality seen in this setting (Fouladiun, Korner et al., 70 2005) . 71 72 Increased inflammation is common in the setting of cancer (Garcia, Garcia-Touza et al., 2005) and 73 is associated with adipose tissue wasting in human studies (Lerner, Hayes et al., 2015) . White 74 adipose tissue (WAT) is a significant source of inflammatory cytokines accounting for more than 30% 75 of circulating interleukin (IL)-6 (Michaud, Boulet et al., 2014) and this and other inflammatory 76 cytokines have been linked to WAT atrophy in the setting of cancer (Petruzzelli, Schweiger et al., 77 2014, Tsoli & Robertson, 2013 , Tsoli, Swarbrick et al., 2016 . Also, a phenotypic switch from WAT to 78 brown adipose tissue (BAT) known as "browning" is thought to contribute to the overall increase in 79 energy expenditure and WAT atrophy seen in cancer cachexia (Petruzzelli et al., 2014) . 80 Nevertheless, the mechanisms regulating adipose tissue atrophy and dysfunction in this setting are 81 incompletely understood. 82 9 EE levels in response to LLC tumor implantation when compared to Ghsr +/+ . Animals 174 co-administered ghrelin were not statistically different from vehicle-treated, tumor-bearing animals. 175 Tumor implantation also decreased spontaneous locomotor activity in both genotypes and ghrelin 176 administration did not prevent these changes . The respiratory quotient (RQ), was 177 significantly decreased by tumor implantation and was not affected by genotype or ghrelin 178 administration (Fig 5 G-I) . Adipose tissue atrophy is a central component of the cancer anorexia and cachexia syndrome 182 (CACS) leading to increased morbidity and mortality (Das, Eder et al., 2011) . Recently, emerging 183 roles for inflammation, WAT browning and increased BAT thermogenesis have been demonstrated 184 in this setting (Daas, Rizeq et al., 2018 , Dalal, 2019 , Han, Meng et al., 2018 , Kir, White et al., 2014 Kliewer, Ke et al., 2015 , Petruzzelli et al., 2014 , Rohm, Schafer et al., 2016 2019, Wang, Zhu et al., 2019) ; however, the pathways involved and their potential as therapeutic 187 targets are not well-known. Ghrelin and agonists of its only known receptor, show 188 potential to ameliorate CACS at least in part by preventing fat atrophy, but the specific mechanisms 189 mediating these effects have not been fully characterized. Given that there are no treatments for cancer cachexia and that several clinical trials targeting this pathway have failed to 191 meet their primary endpoints (Garcia et al., 2015 , Temel, Abernethy et al., 2016 , there is a pressing 192 need to improve our understanding of the mechanisms of action of ghrelin in this setting. In this 193 study we show that ghrelin prevents LLC tumor-induced weight loss, fat atrophy and WAT 194 inflammation without affecting tumor-induced BAT inflammation, WAT browning, and increased BAT 195 uncoupling and whole-body energy expenditure. We confirmed that its orexigenic effects are 196 GHSR-1a-dependent, and also show that other novel GHSR-1a-independent mechanisms are 197 involved given the partial improvements in fat atrophy and WAT inflammation seen in ghrelin-treated, 198 Ghsr -/animals. Also, this is the first report of macrophages as the source of WAT and BAT in the setting of CACS. Weight loss and survival rates are correlated with IL-6 levels in cancer patients (Garcia et al., 2005, 202 Moses, Maingay et al., 2009 , Scott, McMillan et al., 1996 . These observations and several 203 mechanistic studies support the premise that inflammation plays a central role in CACS. Increases 204 11 in IL-1β and TNF contribute to anorexia (Baracos, Martin et al., 2018 , Braun, Zhu et al., 2011 , Khatib, 205 Gaidhane et al., 2018 , and TNF and IL-6 promote lipolysis and inhibit lipogenesis in WAT leading to 206 weight loss (Fearon, Glass et al., 2012 , Han et al., 2018 , Jeanson, Carriere et al., 2015 2014, Ruan, Hacohen et al., 2002) . In non-cancer settings, one third of the circulating IL-6 is 208 produced by WAT (Mohamed-Ali, Goodrick et al., 1997) and most of this WAT-derived IL-6 comes 209 from the stroma-vascular fraction composed of endothelial cells, monocytes/macrophages, 210 myocytes, and fibroblasts (Fain, Madan et al., 2004) , although it can also be derived from 211 adipocytes (Fain, 2006) . Macrophages in WAT are known to be the source of proinflammatory 212 cytokines in conditions leading to AT hypertrophy including obesity (Di Gregorio, Yao-Borengasser 213 et al., 2005 , Divoux, Tordjman et al., 2010 , Lumeng, Deyoung et al., 2007 ) but this has not been 214 previously shown in CACS. Here we show that LLC tumor implantation induces an increase in 215 inflammatory cytokines in circulation as well as in BAT and WAT. Moreover, these AT cytokines 216 appear to be derived exclusively from macrophages residing in these tissues. Adipose tissue 217 atrophy in cancer patients with CACS has been associated with an increase in subcutaneous AT 218 macrophages (Batista, Henriques et al., 2016 , de Matos-Neto, Lima et al., 2015 , Henriques, Sertie 219 et al., 2017 and tissue inflammation (Batista, Olivan et al., 2013 , de Matos-Neto et al., 2015 Henriques et al., 2017). Although, macrophage infiltration has also been described in WAT from 221 tumor-bearing rodents (Henriques et al., 2017 , Machado, Costa Rosa et al., 2004 2014), to our knowledge this is the first report of macrophages as the source of pro-inflammatory 223 cytokines in adipose tissue in CACS. These findings may explain why AT remains an important 224 source of pro-inflammatory cytokines even when the adipocyte mass is significantly reduced in this 225 setting. Also, this may be clinically relevant to cancer patients since knowing the source of 226 inflammation may allow us to target these pathways more effectively (Henriques, Lopes et al., 227 2018). Previously, we have shown that activation of GHSR-1a by ghrelin or GHSR-1a agonists (GHS) 230 increases food intake and body weight (13, 39, 40) . Our group and others also have shown that 231 ghrelin reduces fat oxidation and lipolysis and increases lipogenesis and adiposity in a rodent model 232 of cisplatin-induced cachexia by a combination of food intake-dependent and independent 233 mechanisms (Chen et al., 2015 , Garcia et al., 2013b , Porporato, Filigheddu et al., 2013 . Ghrelin is 234 thought to have anti-inflammatory effects in other settings (Deboer, Zhu et al., 2008 , Dixit, Schaffer 235 et al., 2004 , Tsubouchi, Yanagi et al., 2014 but this is not yet clear in CACS. Some reports suggest 236 an anti-inflammatory effect of native ghrelin administration, but this was not confirmed in other 237 studies using GHSR-1a agonists (Chen et al., 2015 , Garcia, Friend et al., 2013a . In the current 238 study, we report that ghrelin modulates inflammation in a tissue-specific manner. Ghrelin did not 239 prevent tumor-induced increases in circulating inflammatory cytokines or in BAT IL-1β or MCP-1 240 protein levels. However, it mitigated LLC-induced inflammation in WAT. This effect was seen in both 241 genotypes although it was clearer in wild type animals partly because Ghsr -/mice appear to be 242 resistant to tumor-induced inflammation. GHSR-1a is not expressed in adipocytes (Sun, Garcia et 243 al., 2007) but is present in macrophages (Ma, Lin et al., 2013) and our findings are consistent with a 244 previous report showing that old, non-tumor-bearing Ghsr -/mice have reduced macrophage 245 infiltration, a shift on macrophage differentiation towards a more anti-inflammatory phenotype, and 246 decreased inflammation in adipose tissue (Lin, Lee et al., 2016) . However, a GHSR-1a-independent 247 effect of ghrelin on macrophages is also possible as it has been proposed in other settings (Avallone, 248 Demers et al., 2006 , Bulgarelli, Tamiazzo et al., 2009 , Lucchi, Costa et al., 2017 . Taken together, 249 our data is consistent with a WAT-specific, anti-inflammatory effect of ghrelin that is partly GHSR-1a 250 dependent. This is clinically relevant as GHSR-1a agonists are in clinical development for CACS 251 and their effect on these GHSR-1a independent pathways is not known (Garcia et al., 2015) . Also, 252 13 the differences we report between serum, WAT and BAT levels underscore the limitations of relying 253 exclusively on circulating cytokine levels when trying to determine the potential role of inflammation 254 in other tissues. 255 256 Energy expenditure is an important mechanism in the regulation of body weight and is increased in 257 CACS (Garcia et al., 2013a , Kir, Komaba et al., 2016 , Rohm et al., 2019 . Factors contributing to EE 258 include physical activity and resting EE (REE) (Silver, Dietrich et al., 2007, Vazeille, Jouinot et al., 259 2017) and adipose tissue can lead to an increase in REE by uncoupling oxidative phosphorylation in 260 mitochondria thereby releasing heat through activation of a proton leak (Nicholls, 1976, 261 Okamatsu-Ogura, Kitao et al., 2007) . In WAT, browning has been noted in multiple cancer cachexia 262 models with adipocytes showing an upregulation of the main regulator of thermogenesis, UCP1 263 (Dong, Lin et al., 2018 , Vaitkus & Celi, 2017 . In BAT, increased thermogenesis has been reported in 264 cachectic animals (Kir et al., 2014) independently of decreased food intake or their ability to 265 maintain their body temperature (Tsoli, Moore et al., 2012) . Proinflammatory cytokines have been 266 suggested as key drivers of WAT browning (Han et al., 2018 , Petruzzelli et al., 2014 and of BAT 267 thermogenesis through activation of sympathetic nervous system or targeting BAT directly (Arruda, 268 Milanski et al., 2010 , Dascombe, Rothwell et al., 1989 , Li, Klein et al., 2002 , Tsoli et al., 2012 . Here 269 we show that LLC-tumor implantation led to an increase in total EE in spite of a significant decrease 270 in physical activity, suggesting an increase in REE. This was associated with an increase in UCP-1 271 expression in WAT (browning) and in BAT. Moreover, these effects were more marked in Ghsr -/mice 272 suggesting a protective role of GHSR-1a in this setting. These results agree with previous reports in 273 aged, non-tumor-bearing Ghsr -/showing higher levels of thermogenesis and energy expenditure 274 when compared to aged-matched, wild-type mice (Lin, Saha et al., 2011) . The effect of ghrelin or 275 GHSR1a agonists on energy expenditure is unclear with some studies showing a decrease in EE 276 14 (Borner, Loi et al., 2016 , Villars, Pietra et al., 2017 while others showed no effect (Adachi, Takiguchi 277 et al., 2010 , Tschop, Smiley et al., 2000 , Vestergaard, Djurhuus et al., 2008 . In this study, we did 278 not see a significant effect of ghrelin on preventing LLC-induced fat browning, BAT thermogenesis, 279 increased REE or decreased physical activity in the setting of CACS despite the fact that ghrelin 280 prevented fat and weight loss and anorexia. We hypothesize that differences in the models, route of 281 administration and treatment regimen and agents used (LLC mice vs. C26 mice or hepatoma model 282 in rats, administration via s.q. vs. oral gavage vs. osmotic mini pump, ghrelin vs. GHSR1a agonists) 283 could account for these discrepancies. More studies will be needed to test this hypothesis. Calderon-Dominguez, Mir et al., 2016) . In CACS the aforementioned tumor-induced inflammation is 289 thought to play an important role in BAT thermogenesis (Petruzzelli et al., 2014 , Tsoli et al., 2012 ; 290 however, the source of inflammation in BAT is not known. Similar to WAT, we found that BAT IL-6 291 and TNF come exclusively from macrophages in the setting of cachexia. However, their expression 292 in BAT were lower than in WAT and no significant changes were found in response to tumor 293 implantation or ghrelin. We found a significant tumor-effect on increasing IL-1β levels in BAT 294 although ghrelin did not prevent this increase, suggesting tissue-specific differences in inflammation 295 between BAT and WAT in response to tumor and ghrelin. Taken together, these results are 296 important because they show that tumor-induced WAT browning and BAT thermogenesis are 297 associated with significant increases in REE and appear to be independent of inflammation given 298 that downregulating inflammation does not prevent uncoupling in WAT and that BAT IL6 and TNF 299 levels were not upregulated upon tumor implantation. In addition, our data suggests that WAT is a 300 15 significant source of inflammatory cytokines, which express the highest levels of IL-1β, IL-6, and 301 TNF when compared to BAT and circulating levels. 302 303 There were limitations to our approach. This study was not set up to establish the safety of ghrelin 304 administration in the setting of cancer. Nevertheless, none of the studies published to date using 305 ghrelin or GHSR-1a agonists in mice or humans have shown an increase in tumor progression 306 (Sever, White et al., 2016) . Also, the experiments were not designed to characterize other 307 mechanisms contributing to the protective role of GHSR-1a in this setting. Lastly, our data suggest 308 that there is an alternative receptor for ghrelin although identification of this receptor remains elusive 309 and is the focus of other studies. 310

In summary, ghrelin prevents LLC tumor-induced body weight and fat loss by a combination of 312 GHSR-1a-dependent mechanisms including preventing anorexia, and other mechanisms that are 313 partly GHSR-1a-independent. The increase in inflammation in AT induced by tumor implantation is 314 prevented by ghrelin only in WAT; however, tumor-induced WAT browning, and increased BAT 315 inflammation, uncoupling and whole body energy expenditure are not prevented by ghrelin even 316 when the presence of GHSR-1a appears to contribute to maintaining energy balance in this setting. 317

Tumor-induced WAT browning and BAT thermogenesis are associated with significant increases in 318 REE and these seem to be independent of inflammation given that downregulating it does not 319 prevent these changes. These results are clinically relevant because they show that ghrelin Five to seven-month-old male C57BL/6J growth hormone (GH) secretagogue receptor wild type 329 (Ghsr +/+ ) and knockout (Ghsr -/-) congenic mice were used for all experiments. Briefly the Ghsr +/+ and 330

Ghsr -/mice were originally from Dr. Roy G. Smith Ph.D's laboratory (Sun, Butte et al., 2008) 

The procedures of tumor implantation (TI) and ghrelin intervention were described previously (Chen 342 et al., 2015) . In brief, mice were injected subcutaneously (s.q.) with Lewis lung carcinoma (LLC) 343 cells (1 × 10 6 cells, CRL1642, American Type Culture Collection, Manassas, VA) into the right flank 344 or with equal volume and number of heat-killed LLC cells (HK). Approximately 7 days after tumor 345 implantation (TI), when the tumor was palpable (~1cm in diameter), the tumor-bearing mice were 346 treated with either acylated ghrelin (AS-24160, Anaspect, Fremont, CA) at a dose of 0.8 mg/kg or 347 vehicle (0.9% sodium chloride, 8881570121, COVIDIEN, Dublin, Ireland), s.q., twice daily, while 348 mice in HK group received vehicle (saline, same volume), s.q., twice daily for two weeks. 

The Comprehensive Laboratory Animal Monitoring System (CLAMS™, Columbus Instruments, 363 Columbus, OH) was used to identify metabolic parameters of the animals as we previously 364 described (Guillory, Chen et al., 2017) . Ghsr +/+ and Ghsr -/mice were individually housed in CLAMS 365 cages for 96 hours before TI as well as at the endpoint (see the Supplemental Fig. 5 , timeline for the 366 study). The first 12 hours of CLAMS was considered as the acclimation phase and the data for the 367 next 72 hours were analyzed. Oxygen consumption (VO 2 ) (mL/h), carbon dioxide production (VCO 2 ) 368 (mL/h), and locomotor activity (infrared beam-break counts) were recorded automatically by the 369 CLAMS system every 20 min. The respiratory exchange ratio (RQ) and energy expenditure (EE, or 370 heat generation) were calculated from VO 2 and VCO 2 gas exchange data as follows: RQ = 371 VCO 2 /VO 2 and EE = (3.815 + 1.232 × RQ) × VO 2 , respectively. Energy expenditure was then 372 18 normalized to LBM for statistical analysis using two-way analysis of variance (ANOVA). Alternatively, 373

we also analyzed EE value by ANCOVA with LBM as a covariate. Locomotor activity was measured 374 on x-and z-axes by the counts of beam-breaks during the recording period. The data shown in the 375 results was summarized as the mean of every 6 hours in a 72-hour-period. For iWAT and BAT samples, 150ug of the protein lysate was diluted with Diluent 41 and loaded onto 385 each well. The plate was incubated at room temperature (RT) with shaking for 2h followed by 3 386 times of wash in phosphate buffered saline with .05% Tween 20 (PBS/T). Sulfo-tag labeled 387 detection antibody was then added to plates and incubated for 2.5h. After another 3 washes in 388 PBS/T, Read Buffer T(2x) was added and the plate was read on MSD Sector Imager (MSD). 389 390 Immunohistochemistry 391

The iWAT and BAT were mounted with OCT (VWR 25608-930, VWR, Radnor, PA) and flash frozen 392 in liquid nitrogen-chilled isopentane immediately after tissue collection. The OCT-mounted iWAT 393 and BAT blocks were sliced at 14μm using a Cryostat (Leica CM3050S, Nussloch, Germany) at 394 -40 o C. Before the process of staining, slides were dehydrated at RT for 30 minutes followed by 395 incubating in methanol for 15 minutes at -20 o C. To identify the colocalization of F4/80 and IL-6 or 396 TNFα in iWAT and BAT, slides were blocked with 10% donkey serum for 1 hour at RT and followed 397 by incubating in primary antibodies (F4/80 Monoclonal Antibody 1:100, MF48000, Thermo Fisher 398 Scientific; Anti-IL-6 antibody 1:100, ab6672, Abcam; TNF alpha monoclonal antibody, FITC, 399 eBioscience™ 1: 200, Thermo Fisher Scientific) Signaling). The stained slides were dehydrated by 70%, 90%, 100% ethanol, and 100% xylene 411 sequentially and mounted with coverslips by using Permount (SP15-100, Thermo Fisher Scientific). 412

All stained slides were imaged by Nikon NiE microscope at 20x (iWAT) or 40x (BAT). The positive 413 cells (immunofluorescence) or positive area (DAB stain) in the section were quantified and 414 normalized to the total area of the section (mm 2 ) using ImageJ analysis software (National Institutes 415 of Health, http://rsb.info.nih.gov/ij/). Two-way ANOVA was performed to identify differences between genotypes (Ghsr +/+ vs. Ghsr -/-) 419 across treatments (HK, TV, and TG) followed by Fisher's LSD post hoc test. For inflammatory 420 20 cytokines, Kruskal-Wallis test was performed to identify the differences between groups. For energy 421 expenditure, ANCOVA was also used for analysis in addition to ANOVA with LBM as a covariate to 422 identify differences between genotypes across treatments followed by Fisher's LSD post hoc test. 423 Values are presented in mean ± SEM. All statistical testing was performed using IBM SPSS version 424 Adachi S, Takiguchi S, Okada K, Yamamoto K, Yamasaki M, Miyata H, Nakajima K, Fujiwara Y, 446 Hosoda H, Kangawa K, Mori M, Doki Y (2010) Effects of ghrelin administration after total 447 gastrectomy: a prospective, randomized, placebo-controlled phase II study. 

and thermogenic responses to IL-1beta in mice

Ghrelin-induced adiposity is independent of orexigenic effects

A switch from white to brown 617 fat increases energy expenditure in cancer-associated cachexia

Acylated and unacylated ghrelin impair skeletal muscle atrophy in 621 mice

Coactivator of Nuclear Receptors Linked to Adaptive Thermogenesis

Diet-induced obesity causes insulin resistance in 625 mouse brown adipose tissue

An AMP-activated protein kinase-stabilizing peptide 629 ameliorates adipose tissue wasting in cancer cachexia in mice

Energy metabolism in cachexia

What we talk about when we talk about fat

Tumor necrosis factor-alpha 633 suppresses adipocyte-specific genes and activates expression of preadipocyte genes in 3T3-L1 634 adipocytes: nuclear factor-kappaB activation by TNF-alpha is obligatory

The relationship between weight loss 636 31 and interleukin 6 in non-small-cell lung cancer

Is there an effect of ghrelin/ghrelin analogs on cancer? A 638 systematic review

Changes in body mass, energy balance, physical function, 640 and inflammatory state in patients with locally advanced head and neck cancer treated with 641 concurrent chemoradiation after low-dose induction chemotherapy

Peptidomimetic regulation of growth hormone secretion

Characterization of adult ghrelin and ghrelin receptor 646 knockout mice under positive and negative energy balance

Ghrelin and growth hormone secretagogue receptor expression 648 in mice during aging

Anamorelin in 650 patients with non-small-cell lung cancer and cachexia (ROMANA 1 and ROMANA 2): results from 651 two randomised, double-blind, phase 3 trials

Ghrelin induces adiposity in rodents

A guide to analysis of mouse energy 656 metabolism

Activation of thermogenesis in brown adipose tissue and 659 dysregulated lipid metabolism associated with cancer cachexia in mice

Cancer cachexia: malignant inflammation, tumorkines, and metabolic 661 mayhem

Lipolytic and thermogenic depletion of adipose tissue 663 in cancer cachexia

Ghrelin relieves 665 cancer cachexia associated with the development of lung adenocarcinoma in mice

The role of adipose tissue in cancer-associated cachexia

Immune Modulation 670 of Brown(ing) Adipose Tissue in Obesity

Relation between hypermetabolism, cachexia, and survival in cancer 673 patients: a prospective study in 390 cancer patients before initiation of anticancer therapy

Acute effects of ghrelin administration on glucose and lipid metabolism

Receptor Agonist HM01 Attenuates Cachexia in Mice Bearing Colon-26 (C26) Tumors

Multiple roles of adipose tissue in cancer formation and progression

Beige adipocytes are a distinct type of thermogenic fat cell in mouse and 686 human

Chronic inflammation: role of adipose tissue and modulation by weight loss

HK: 692 heat-killed + vehicle; TV: tumor + vehicle; TG: tumor + ghrelin. Changes in (A) body weight (carcass 693 weight, n = 8-10) and (B) fat body mass by NMR expressed as % change from baseline

Average cumulative 695 food intake (FI) normalized to baseline FI (g/g, black areas represent food intake in the nighttime, 696 and the bottom areas in the bars represent food intake in the daytime, n = 4-6). * p < 0.05 compared 697 to HK within the same genotype. # p <0.05 compared to TV within the same genotype

HK: 703 heat-killed + vehicle; TV: tumor + vehicle; TG: tumor + ghrelin. Protein levels of inflammatory 704 markers (A)IL-1β, (B) IL-6, and (C) TNF; and (D) macrophage marker MCP-1 in iWAT (pg/mg)

01 compared to HK within the same genotype. # p < 0.05 compared to TV within the 706 same genotype. No genotype difference was detected. Data are shown as mean ± SE. n = 707 6-7/group. (E-F) Colocalization of inflammation and macrophages in iWAT. (E) Representative 708 images of colocalization of inflammatory marker IL-6 and macrophage marker F4

F4/80 in FITC green; nuclei in DAPI blue). (F) Representative images of colocalization of 710 inflammatory marker TNF and macrophage marker F4/80 in iWAT (TNF in FITC green

Positively stained inflammatory markers and colocalizations with 712 macrophages are indicated by the white arrows. Scale bars

HK: 715 heat-killed + vehicle; TV: tumor + vehicle; TG: tumor + ghrelin. Protein levels of inflammatory 716 markers (A)IL-1β, (B) IL-6, and (C) TNF; and (D) macrophage marker MCP-1 in iWAT (pg/mg)

001 compared to HK within the same genotype. # p < 0.05; ### p < 0.001 718 compared to TV within the same genotype. No genotype difference was detected. Data are shown 719 as mean ± SE. n = 6-7/group. (E-F) Colocalization of inflammation and macrophages in BAT

Representative images of colocalization of inflammatory marker IL-6 and macrophage marker F4/80

F4/80 in FITC green; nuclei in DAPI blue). (F) Representative images of 722 colocalization of inflammatory marker TNF and macrophage marker F4/80 in BAT

Positively stained inflammatory markers and 724 colocalizations with macrophages are indicated by the white arrows. Scale bars

HK: heat-killed + vehicle; TV: tumor + vehicle; TG: 727 tumor + ghrelin. (A) Representative IHC images of UCP-1 in iWAT. (B) UCP-1 positive area is 728 expressed as % of the total analyzed area in iWAT

UCP-1 in BAT. (D) UCP-1 positive area is expressed as % of the total analyzed area in BAT

001 compared to HK within the same genotype. Genotype effects 731 are shown as p-values above the corresponding figures (p < .05). Data are shown as mean ± SE. 732 Scale bars

HK: heat-killed + vehicle; TV: tumor + 735 vehicle; TG: tumor + ghrelin. (A-C) Energy expenditure adjusted by LBM is expressed (A) compared 736 to the baseline; (B) every 6 hours; and (C) average of every 6 hours. (D-F) Ambulatory activity is 737 expressed (D) compared to baseline

G-I) Respiratory Quotient (RQ) is expressed (G) compared to baseline

every 6 hours; and (I) average of every 6 hours. *p<0.05 compared to HK within the same genotype

Genotype effects are shown in p-values above the corresponding figures (p < 0.05). N = 4 for HK 741 groups and N = 6 for the rest of the groups. Data are shown as mean ± SE

Data 744 is expressed as box-and-whisker plot showing the median (middle line), mean (middle cross), upper 745 and lower quartiles (box), maximum and minimum (whiskers)

Supplemental Fig. 2. High resolution images of immunohistochemistry staining in iWAT

Representative images of colocalization of inflammatory marker IL-6 and macrophage marker F4/80

F4/80 in FITC green; nuclei in DAPI blue). (B) Representative images of 752 colocalization of inflammatory marker TNF and macrophage marker F4/80 in iWAT

Positively stained inflammatory markers and 754 colocalizations with macrophages are indicated by the white arrows. Scale bars

Supplemental Fig. 3. High resolution images of immunohistochemistry staining in BAT

Representative images of colocalization of inflammatory marker IL-6 and macrophage marker F4/80

F4/80 in FITC green; nuclei in DAPI blue). (B) Representative images of 759 colocalization of inflammatory marker TNF and macrophage marker F4/80 in BAT

Positively stained inflammatory markers and 761 colocalizations with macrophages are indicated by the white arrows. Scale bars

Effects of ghrelin on LLC-induced protein-level changes in inflammation (IL-1β

MCP-1) in plasma (pg/mg, n = 11-14). *, **: different than HK 765 within the same genotype (*: p < .05; **: p < .01). Genotype effects are shown in p-values above the 766 corresponding figures

Timeline of current study. Ghsr +/+ and -/-mice were injected with LLC (T, 1 × 769 106 cells, s.q.) into the right flank or with equal volume and number of heat-killed LLC cells (HK)

the 771 tumor-bearing mice were treated with either acylated ghrelin, 0.8 mg/kg (TG) or vehicle (0.9% 772 sodium chloride, TV), s.q., twice daily, while mice in HK group received vehicle (saline, same 773 volume), s.q., twice daily for two weeks

days before tumor noted, baseline) and weekly till the endpoint. All the mice were 775 individually housed in CLAMS cages for 96 hours before TI (11-7 days before tumor noted, baseline) 776 as well as at the endpoint

Borner T, Loi L, Pietra C, Giuliano C, Lutz TA, Riediger T (2016) The ghrelin receptor agonist HM01 469 mimics the neuronal effects of ghrelin in the arcuate nucleus and attenuates anorexia-cachexia 470 syndrome in tumor-bearing rats. Am J Physiol Regul Integr Comp Physiol 311: R89-96 471 Braun TP, Zhu X, Szumowski M, Scott GD, Grossberg AJ, Levasseur PR, Graham K, Khan S, 472 Damaraju S, Colmers WF, Baracos