Targeting RAGE prevents muscle wasting and prolongs survival in cancer cachexia

Abstract

Background: Cachexia, a multifactorial syndrome affecting more than 50% of patients with advanced cancer and responsible for ~20% of cancer-associated deaths, is still a poorly understood process without a standard cure available. Skeletal muscle atrophy caused by systemic inflammation is a major clinical feature of cachexia, leading to weight loss, dampening patients' quality of life, and reducing patients' response to anticancer therapy. RAGE (receptor for advanced glycation end-products) is a multiligand receptor of the immunoglobulin superfamily and a mediator of muscle regeneration, inflammation, and cancer.

Methods: By using murine models consisting in the injection of colon 26 murine adenocarcinoma (C26-ADK) or Lewis lung carcinoma (LLC) cells in BALB/c and C57BL/6 or Ager-/- (RAGE-null) mice, respectively, we investigated the involvement of RAGE signalling in the main features of cancer cachexia, including the inflammatory state. In vitro experiments were performed using myotubes derived from C2C12 myoblasts or primary myoblasts isolated from C57BL/6 wild type and Ager-/- mice treated with the RAGE ligand, S100B (S100 calcium-binding protein B), TNF (tumor necrosis factor)α±IFN (interferon) γ, and tumour cell- or masses-conditioned media to analyse hallmarks of muscle atrophy. Finally, muscles of wild type and Ager-/- mice were injected with TNFα/IFNγ or S100B in a tumour-free environment.

Results: We demonstrate that RAGE is determinant to activate signalling pathways leading to muscle protein degradation in the presence of proinflammatory cytokines and/or tumour-derived cachexia-inducing factors. We identify the RAGE ligand, S100B, as a novel factor able to induce muscle atrophy per se via a p38 MAPK (p38 mitogen-activated protein kinase)/myogenin axis and STAT3 (signal transducer and activator of transcription 3)-dependent MyoD (myoblast determination protein 1) degradation. Lastly, we found that in cancer conditions, an increase in serum levels of tumour-derived S100B and HMGB1 (high mobility group box 1) occurs leading to chronic activation/overexpression of RAGE, which induces hallmarks of cancer cachexia (i.e. muscle wasting, systemic inflammation, and release of tumour-derived pro-cachectic factors). Absence of RAGE in mice translates into reduced serum levels of cachexia-inducing factors, delayed loss of muscle mass and strength, reduced tumour progression, and increased survival.

Conclusions: RAGE is a molecular determinant in inducing the hallmarks of cancer cachexia, and molecular targeting of RAGE might represent a therapeutic strategy to prevent or counteract the cachectic syndrome.

Keywords: Cancer cachexia; Cytokines; HMGB1; Inflammation; Muscle atrophy; Myogenin; RAGE; S100B.

© 2020 The Authors. Journal of Cachexia, Sarcopenia and Muscle published by John Wiley & Sons Ltd on behalf of Society on Sarcopenia, Cachexia and Wasting Disorders.

Figures

Figure 1

Tumour presence causes a re‐expression of receptor for advanced glycation end‐products (RAGE) in muscle tissue and increases serum levels of the RAGE ligands, S100 calcium‐binding protein B (S100B) and high mobility group box 1 (HMGB1). (A–I) C57BL/6 wilde‐type (WT) mice injected with Lewis lung carcinoma (LLC‐WT) cells or vehicle (Ctrl‐WT) (n = 8 each group) were sacrificed at the indicated day post‐injection (dpi). The weights of body (A), fat (i.e. inguinal white adipose tissue, iWAT and epididymal white adipose tissue, eWAT) (B), and tibialis anterior (TA), gastrocnemius (GC) and quadriceps femoris (QF) muscles (C) were measured. (D) Reported is the average of percentage changes of TA myofiber cross‐sectional area (CSA) for LLC‐bearing mice compared with control mice (see also FigureS1A). Skeletal muscles were analysed for RAGE expression by real‐time PCR (E), western blot (WB) (F), and immunohistochemistry (IHC) (G). (H) RAGE (red) was detected by immunoflourescence (IF) in atrophic myofibers marked with atrogin‐1 (green). 4′,6‐diamidino‐2‐phenylindole (DAPI) (blue) was used to stain nuclei. Dashed lines indicated myofiber borders. (I) Serum levels of S100B and HMGB1 were measured by enzyme‐linked immunosorbent assay (ELISA). (J) S100B and HMGB1 were detected in LLC cell lysates, conditioned medium derived from LLC cells (LLC‐CM), and LLC tumour masses (LLC‐TM) by WB. M, purified S100B (5 ng), or HMGB1 (10 ng). Results are means ± standard error of the mean. Statistical analysis was conducted using the two‐tailed t‐test. *P < 0.05, **P < 0.01, and ***P < 0.001 significantly different from control mice. #P<0.05, ##P<0.01 significantly different. Scale bars (G,H), 50 μm.

Figure 2

Depletion of receptor for advanced glycation end‐products (RAGE) prevents cancer‐induced cachexia, prolongs survival and maintains spleen morphology in Lewis lung carcinoma (LLC) tumour‐bearing mice. (A–I) Wild type (WT) and Ager−/− mice (n = 15 each group) injected with LLC cells were monitored for tumour growth (A), changes in body weight (B), and survival rates (C) until 40 dpi and sacrificed at the indicated dpi. (D) The number of lung metastases/mouse was determined (see also Figure S2C) over time in LLC‐WT and LLC‐Ager−/− mice (n = 8 each group). (E) Skeletal muscles were excised and weighed. (F) Reported is the average of percentage changes of tibialis anterior (TA) of the cross‐sectional area (CSA) for each mouse model compared to control WT mice (see also Figure S3A,B). (G) Plotted are the means ± standard error of the mean (SEM) of time latencies to fall in Kondziela's inverted screen test, in which each point represents an individual mouse. Spleens were weighed (H), and spleen morphology was analysed by haematoxylin/eosin (H&E) staining (I). Dashed lines indicate the boundaries of white and red pulps. Results are means ± SEM (A–F). Statistical analysis was conducted using the two‐tailed t‐test (A,B,E–H) or Mann–Whitney test (D). Significance (P) is indicated for each time‐point starting from 22 dpi (B). *P < 0.05, **P < 0.01, and *** P<0.001, significantly different from internal control mice (D,E,G,H). $P < 0.05 and $$P < 0.01, significantly (D). §P < 0.05 and §§P < 0.01, significantly different from WT (F). #P < 0.05, ## P<0.01, and ###P < 0.001 significantly different. Scale bars in (I), 200 μm.

Figure 3

Receptor for advanced glycation end‐products (RAGE) activity is critical for Lewis lung carcinoma (LLC)‐induced muscle wasting. (A–D) Gastrocnemius muscles from LLC‐wild type (WT) and LLC‐Ager−/− mice were analysed compared with internal control mice (Ctrl) at the indicated dpi. (A) myoblast determination protein 1 (MyoD), myogenin and the adult fast [myosin heavy chain (MyHC)‐II] and slow (MyHC‐I) myosin heavy chain isoforms were detected by western blot (WB). Reported are the relative densities. (B) Expression of developmental MyHC (dMyHC) at 25 dpi was analysed by western blot (WB). α‐Actinin or GAPDH were used for loading control (A,B). (C) Levels of Fbxo32, Trim63 and Myog were analysed by real‐time PCR. (D) Myogenin (red) and atrogin‐1 (green) were detected by immunofluorescence (IF). 4′,6‐diamidino‐2‐phenylindole (DAPI) (blue) was used to stain nuclei. See also Figure S5. Reported are high‐magnification insets with myofibers defined by dashed lines. Results are means ± standard error of the mean. Statistical analysis was conducted using the two‐tailed t‐test. *P < 0.05, **P < 0.01, and *** P<0.001, significantly different from internal control mice. #P < 0.05, ##P < 0.01 and ###P < 0.001 significantly different. $P < 0.05 and $$P < 0.01 LLC‐Ager−/− vs. LLC‐WT mice significantly different (C). Scale bars in (D), 25 μm.

Figure 4

Receptor for advanced glycation end‐products (RAGE) signalling is required for tumour necrosis factor (TNF)α ± interferon (IFN)γ‐induced reduction of myotube size. (A–E, G–M) Myotubes obtained by culturing C2C12 myoblasts in differentiation medium (DM) for 4 days were added with TNFα (20 ng/ml) ± IFNγ (100 U/ml) in absence and presence of a RAGE blocking antibody (Ab‐RAGE) (10 μg/ml) for the indicated time. (A) Myosin heavy chain (MyHC)‐II expression was analysed by immunofluorescence (IF). Reported are the percentages of myotubes diameters compared with control. (B) RAGE expression was evaluated by western blot (WB). (C) The conditioned media of TNFα ± IFNγ‐treated myotubes were processed for detection of released S100 calcium‐binding protein B (S100B) and high mobility group box 1 (HMGB1) by WB. The tubulin is relative to cell lysates from which the media are derived. (D,E) levels of MyHC‐II protein and messenger RNA (mRNA) were evaluated by WB (D) and real‐time PCR (E). (F) Myotubes obtained from primary myoblasts from wild type (WT) or Ager−/− mice were treated with TNFα ± IFNγ to analyse MyHC‐II expression. (G) Levels of Ager, Myog, Fbxo32, and Trim63 were analysed by real‐time PCR. (H) Total and phosphorylated protein kinase B (Akt), p38 mitogen‐activated protein kinase (MAPK), and p65 levels were analysed by western blot (WB). (I) Levels of Myog and Fbxo32 were evaluated by real‐time PCR. (J,K) Myogenin (J) and myoblast determination protein 1 (MyoD) (K) expression was evaluated by WB. (L,M) the expression and localization of total and phosphorylated signal transducer and activator of transcription 3 (STAT3) were studied by WB (L) and IF (M). Reported are the relative densities with respect to tubulin or total form of phosphorylated protein (D,H,J–L). Results are means ± standard error of the mean (SEM) (A,F) or standard deviation (SD) (D,E,G–L). Statistical analysis was conducted using the two‐tailed t‐test. *P < 0.05, **P < 0.01, and ***P<0.001, significantly different from internal control. #P < 0.05 and ##P < 0.01, significantly different. Scale bars (A,F,M), 100 μm. See also Figure S6,7.

Figure 5

Dual trophic and atrophying effect of receptor for advanced glycation end‐products (RAGE) signalling in myotubes in the absence of cytokines. (A–N) C2C12 myotubes were cultured with S100 calcium‐binding protein B (S100B) (0–20 μg/ml) in the absence or presence of either Ab‐RAGE or SB203580, or with high mobility group box 1 (HMGB1) (0–30 μg/ml). Levels of MyHC‐II was evaluated by western blot (WB) (A), immunofluorescence (IF) (B) and real‐time PCR (C). Reported are the percentages of myotubes diameters relative to control (B). (D,E) myoblast determination protein 1 (MyoD) (D) and Ager (E) expression was analysed by western blot (WB) and real‐time PCR, respectively. (F) Levels of RAGE, myogenin, MyoD, and total and phosphorylated protein kinase B (Akt), p38 mitogen‐activated protein kinase (MAPK), and signal transducer and activator of transcription 3 (STAT3) were evaluated by WB. (G) Levels of Ager, Myog, Fbxo32, and Trim63 were analysed by real‐time PCR. (H) the localization of pho‐STAT3 was evaluated by IF. (I) The expression of myosin heavy chain (MyHC)‐II was evaluated by WB, and the percentage of myotubes diameters respect to control was determined. (J,K) Levels of Ager (J) and Myog, Fbxo32, and Trim63 (J,K) were analysed by real‐time PCR. (L‐N) Levels of MyoD protein (L,M) were analysed by WB, and total and phosphorylated STAT3 levels were analysed by WB (N). (O,P) Gastrocnemius muscles of WT and Ager−/− mice (n = 5 each group) were injected daily for three consecutive days with S100B (50 ng/muscle) or vehicle [phosphate‐buffered saline(PBS)] and analysed for the expression of Ager (O), and Fbxo32 and Myog (P) by real‐time PCR. Results are means ± standard error of the mean (B,I) or standard deviation (C,E,G,J–P). Statistical analysis was conducted using the two‐tailed t‐test. *P < 0.05 and **P < 0.01, significantly different from internal control. #P < 0.05 and ##P < 0.01, significantly different. Scale bars (B,H,I), 100 μm.

Figure 6

Reducing receptor for advanced glycation end‐products (RAGE) activity prevents reduction of myotube size induced by tumour‐derived factors and reduces the release of cachexia‐inducing factors from Lewis lung carcinoma (LLC) tumour. (A–D) C2C12 myotubes cultured with conditioned medium derived from LLC cells (LLC‐CM) or C26 cells (C26‐CM) in the absence or presence of Ab‐RAGE (10 μg/ml). (A) Ager levels were analysed by real‐time PCR. (B) myosin heavy chain (MyHC)‐II expression was analysed by immunofluorescence (IF), and the percent changes in myotube diameter relative to the control were determined. Trim63 and total and phosphorylated p38 mitogen‐activated protein kinase (MAPK) were analysed by real‐time PCR (C) and western blot (WB) (D), respectively. (E) C2C12 myotubes were cultured with medium conditioned by LLC masses (LLC‐TM) derived from wild type (WT) or Ager−/− (n = 5 each group) at various day post injection. Measurements of myotube diameters after MyHC‐II immunocytochemistry (ICC) analysis are shown as percentage change relative to control. (F) LLC‐TM derived from WT or Ager−/− were analysed for S100 calcium‐binding protein B (S100B) and high mobility group box 1 (HMGB1) content by WB. GAPDH is relative to tumour lysates from which the media are derived. Results are means ± standard error of the mean (B,E) or standard deviation (A,C,D). Statistical analysis was conducted using the two‐tailed t‐test. *P < 0.05 and **P < 0.01, significantly different from internal control. #P < 0.05 and ##P < 0.01, significantly different. Scale bars (B,E), 100 μm. See also Figure S9.