Pharmacological or genetic inhibition of iNOS prevents cachexia-mediated muscle wasting and its associated metabolism defects

Abstract

Cachexia syndrome develops in patients with diseases such as cancer and sepsis and is characterized by progressive muscle wasting. While iNOS is one of the main effectors of cachexia, its mechanism of action and whether it could be targeted for therapy remains unexplored. Here, we show that iNOS knockout mice and mice treated with the clinically tested iNOS inhibitor GW274150 are protected against muscle wasting in models of both septic and cancer cachexia. We demonstrate that iNOS triggers muscle wasting by disrupting mitochondrial content, morphology, and energy production processes such as the TCA cycle and acylcarnitine transport. Notably, iNOS inhibits oxidative phosphorylation through impairment of complexes II and IV of the electron transport chain and reduces ATP production, leading to energetic stress, activation of AMPK, suppression of mTOR, and, ultimately, muscle atrophy. Importantly, all these effects were reversed by GW274150. Therefore, our data establish how iNOS induces muscle wasting under cachectic conditions and provide a proof of principle for the repurposing of iNOS inhibitors, such as GW274150 for the treatment of cachexia.

Keywords: cachexia; cancer; iNOS; inflammation; metabolism.

Figure 1. iNOS knockout mice are resistant to LPS‐driven muscle wasting

Male C57BL/6 wild‐type (WT) and iNOS knockout (KO) mice were intraperitoneally injected with 1 mg kg⁻¹ LPS or an equivalent volume of carrier solution. Control WT, control KO, and LPS‐treated KO cohorts were pair‐fed (PF) to the WT LPS‐treated cohorts. After 18 h, mice were euthanized, and tissue samples were analyzed.

1. A

   Quadriceps from saline or LPS‐treated, WT, or iNOS KO mice were isolated and used for Western blot analysis with anti‐iNOS (n = 4). Ponceau S was used for total protein staining.

2. B

   (left) Quadriceps from saline or LPS‐treated, WT, or iNOS KO mice were isolated and used for Western blot with anti‐3NT (n = 6). Total protein levels are also shown. (right) Quantification of the 3NT‐to‐total protein ratio. Ratios are expressed relative to the saline‐treated controls.

3. C

   Representative images of tibialis anterior muscles. Scale bars represent 0.5 cm.

4. D–F

   Weight of tibialis anterior (D), quadricep (E), and soleus (F) muscle normalized to initial body weight (n = 15).

5. G

   (left) Representative photomicrographs of H&E‐stained tibialis anterior muscle sections from control and iNOS KO mice injected with or without LPS. Scale bars = 100 μm. (right) Frequency histogram showing the distribution of muscle fiber minimum Feret diameter in the tibialis anterior muscles from saline‐ or LPS‐treated (top) WT control and (bottom) iNOS KO mice (n = 4). A total of 300 fibers per muscle were used for the analysis. Statistical comparisons, mean, and standard deviation of the mean are shown in the histogram legend.

6. H

   Change in grip strength between day of injection and endpoint of experiment. (WT saline n = 12, WT LPS n = 11, iNOS KO saline n = 9, and iNOS KO LPS n = 10).

Data information: Individual data points represent values from individual mice. Error bars represent the standard deviation (SD) of the mean. (B) Statistical comparisons were made between saline‐treated controls and LPS‐treated mice of same genotype. Plotted data were relativized to saline controls of corresponding genotype. Δ indicates the difference in mean values, and P‐values were calculated with Student’s t‐test. (D–F and H) For statistical comparisons, Δ indicates the difference in mean values and P‐values were calculated with an ANOVA followed by Fisher's LSD test. (G) P‐values were calculated with a Kolmogorov–Smirnov test (***P < 0.001). Non‐statistically significant comparisons (P > 0.05) are indicated as non‐significant (ns). Source data are available online for this figure.

Figure EV1. Characterization of the LPS‐induced immune response in WT and iNOS KO mice

Male C57BL/6 wild‐type (WT) and iNOS knockout (KO) mice were intraperitoneally injected with 1 mg kg⁻¹ LPS or an equivalent volume of carrier solution. Control WT, control KO, and LPS‐treated KO cohorts were pair‐fed (PF) to the WT LPS‐treated cohorts. After 18 h, mice were euthanized, and tissue samples were analyzed.

1. A

   Spleen weight normalized to initial body weight (n = 15).

2. B

   Gating strategy for flow cytometry analysis of spleen M1 and M2 macrophages.

3. C

   Percentage of CD86⁺ M1 splenic macrophages relative to the total F4/80+Ly6C‐ macrophage population (WT saline n = 6, WT LPS n = 6, iNOS KO saline n = 4, and iNOS KO LPS n = 4).

4. D

   Percentage of CD206⁺ M2 splenic macrophages relative to the total F4/80+Ly6C‐ macrophage population (WT saline n = 6, WT LPS n = 6, iNOS KO saline n = 4, and iNOS KO LPS n = 4).

5. E

   Quadriceps from saline or LPS‐treated, WT or iNOS KO mice were isolated and used for RT–qPCR analysis for F4/80 mRNA expression. The – Δ Δ CT was plotted relative to the saline conditions of each background and normalized to GAPDH mRNA levels (n = 5).

6. F–I

   Serum from saline or LPS‐treated, WT or iNOS KO mice was collected. Fold change in IL‐1α (F), IL‐1β (G), IL‐6 (H), and TNF‐α (I) levels in LPS‐treated mice was plotted relative to saline condition (WT saline n = 6, WT LPS n = 6, iNOS KO saline n = 4, and iNOS KO LPS n = 4).

Data information: Individual data points represent values from individual mice. Error bars represent the standard deviation (SD) of the mean. (A, C, D) For statistical comparisons, Δ indicates the difference in mean values and P‐values were calculated with an ANOVA followed by Fisher's LSD test. (E‐I) Statistical comparisons were made between saline‐treated controls and LPS‐treated mice of same genotype. Plotted concentration data were relativized to saline controls of corresponding genotype. Δ indicates the difference in mean values, and P‐values were calculated with Student’s t‐test.

Figure EV2. Characterization of the LPS‐induced effects on muscle integrity/composition in WT and iNOS KO mice

Male C57BL/6 wild‐type (WT) and iNOS knockout (KO) mice were intraperitoneally injected with 1 mg kg⁻¹ LPS or an equivalent volume of carrier solution. Control WT, control KO, and LPS‐treated KO cohorts were pair‐fed (PF) to the WT LPS‐treated cohorts. After 18 h, mice were euthanized, and tissue samples were analyzed.

1. Percent body weight change from time of injection to endpoint of experiment. (n = 15).

2. Gastrocnemius weight normalized to initial body weight (n = 15).

3. Frequency histogram showing the distribution of muscle fiber CSA in the tibialis anterior muscles from (left) WT control and (right) iNOS KO mice (n = 4). A total of 300 fibers per muscle were used for the CSA analysis. Statistical comparisons, mean, and standard deviation of the mean are shown in the histogram legend.

4. (left) Photomicrographs of immunofluorescence images of tibialis anterior muscle sections from WT control and iNOS KO mice injected with or without LPS. Sections were stained against MyHC I (Blue), MyHC IIa (Green), and laminin (Red). Unstained fibers were classified as MyHC IIb/x. Scale bars = 50 μm. (right) Percentage of type I, type IIa, and type IIb/IIx fibers in the whole tibialis anterior muscles from control and iNOS KO mice (n = 2).

Data information: Individual data points represent values from individual mice. Error bars represent the standard deviation (SD) of the mean. (C) P‐values were calculated with a Kolmogorov–Smirnov test (**P < 0.01). Non‐statistically significant comparisons (P > 0.05) are indicated as non‐significant (ns).

Figure EV3. Effect of GW on the immune response as well as muscle integrity/composition in C26‐tumor‐bearing mice

Male BALB/C mice were injected subcutaneously with C26 cells (1.25 × 10⁶ cells) or an equivalent volume of saline. After 5 days, and everyday thereafter, saline‐ and C26‐injected mice were treated with or without GW (5 mg kg⁻¹). After 16 days, mice were euthanized and tissue samples were analyzed.

1. A

   iNOS and tubulin (loading control) protein levels in the gastrocnemius muscle were assessed by Western blot (n = 3).

2. B

   Spleen weight normalized to initial body weight.

3. C

   Quadriceps from control and C26 mice treated with or without GW were isolated and used for RT–qPCR analysis for F4/80 mRNA expression. The – Δ Δ CT is plotted relative to the saline condition and normalized to GAPDH mRNA levels (n = 5).

4. D–G

   Serum from saline and C26‐tumor‐bearing mice treated with or without GW was collected. Fold change in IL‐1α (D), IL‐1β (E), IL‐6 (F), and TNF‐α (G, n = 4 for C26 + GW) levels was plotted relative to saline condition.

5. E

   Inguinal fat pad weight normalized to initial body weight.

6. F

   Frequency histogram showing the distribution of muscle fiber CSA in the gastrocnemius muscles from control and C26 mice treated with or without GW (saline n = 4, C26 n = 4, GW 5 mg/kg n = 3, and C26 GW 5 mg/kg n = 4). A total of 600–700 fibers per muscle were used for the CSA analysis. Statistical comparisons, mean, and standard deviation of the mean are shown in the histogram legend.

7. G

   (left) Photomicrographs of immunofluorescence images of gastrocnemius muscle sections from control and C26 mice treated with or without GW. Sections were stained against MyHC I (Blue), MyHC IIa (Green), and laminin (Red). Unstained fibers were classified as MyHC IIb/x. Scale bars = 50 μm. (right) Percentage of type I, type IIa, and type IIb/IIx fibers in the whole gastrocnemius muscles from control and C26 mice treated with or without GW (n = 2).

Data information: Individual data points represent values from individual mice, with a total of six mice per cohort (n = 6) unless stated otherwise. Error bars represent the standard deviation (SD) of the mean. (B–H) For statistical comparisons, Δ indicates the difference in mean values and P‐values were calculated with an ANOVA followed by Fisher's LSD test. (I) P‐values were calculated with a Kolmogorov–Smirnov test (***P < 0.001). Source data are available online for this figure.

Figure EV4. Cytokine‐mediated loss of complex II and IV integrity is reversed with aminoguanidine

C2C12 myotubes were treated with or without IFNγ (100 U/ml) and TNF‐α (20 ng/ml) and the indicated doses of aminoguanidine (AMG). Protein content was extracted 24 h after treatment.

1. A

   Western blot analysis for ETC protein complex subunits.

2. B–F

   Quantification of complex subunits normalized to VDAC (outer mitochondrial membrane; OMM) and relative to untreated control. (B) NDUFB8 (complex I; CI), (C) SDHB (complex II; CII), (D) UQCRC2 (complex III; CIII), (E) MTCO1 (complex IV; CIV), and (F) ATP5A (complex V; CV).

Data information: Individual data points are from three independent experimental replicates (n = 3). Error bars represent the standard deviation (SD) of the mean. For statistical comparisons, Δ indicates the difference in mean values and P‐values were calculated with an ANOVA followed by Fisher's LSD test. Non‐statistically significant comparisons (P > 0.05) are indicated as non‐significant (ns). Source data are available online for this figure.

Figure 2. Genetic ablation of iNOS prevents LPS‐driven deregulation of the TCA cycle and energy production

Male C57BL/6 wild‐type (WT) and iNOS knockout (KO) mice were intraperitoneally injected with 1 mg kg⁻¹ LPS or an equivalent volume of carrier solution. Control WT, control KO, and LPS‐treated KO cohorts were pair‐fed (PF) to the WT LPS‐treated cohorts. After 18 h, mice were euthanized and the metabolome of tibialis anterior muscles was analyzed through LC‐MS/MS.

1. Scores scatter plot of partial least square discriminant analysis (PLS‐DA) model of metabolites from WT and iNOS KO mice treated with or without LPS. Metabolomic data were range‐scaled and mean‐centered.

2. Pathway analysis using MetaboAnalyst 4.0 Software comparing significantly altered pathways from WT saline to LPS as well as iNOS KO saline to LPS. Pathways are ranked by their significance and filtered based on a Pathway Impact Score > 0.1. Metabolomic data were range‐scaled and mean‐centered. P‐values were obtained using GlobalTest, and the −log(P‐value) corresponding to a P‐value of 0.05 is indicated by the red dashed line.

3. (left) Western blot analysis of pThr172‐AMPK (pAMPK) and total AMPK (AMPK) in quadriceps muscle. (right) Quantification of the pAMPK‐to‐AMPK ratio. Ratios are expressed relative to the saline‐treated controls.

4. Schematic of glycolysis and the TCA cycle. White boxes denote detected metabolites. Gray boxes denote undetected metabolites. Dashed arrows indicate the presence of multiple reactions between metabolites.

5. Relative concentrations of metabolites involved in glycolysis and TCA cycle.

6. Relative concentration of arginine.

7. Schematic of acylcarnitine metabolism. Gray circles denote key enzymes involved in metabolic processes highlighted.

8. Relative estimated activity of CPT1, CPT2, LCADH, and β‐oxidation.

Data information: Individual data points represent values from individual mice, with a total of six mice per cohort (n = 6). (B) P‐values were obtained using GlobalTest. (C–H) Error bars represent the standard deviation (SD) of the mean. Statistical comparisons were made between saline‐treated controls and LPS‐treated mice of same genotype. Plotted concentration data were relativized to saline controls of corresponding genotype. P‐values were calculated with Student’s t‐test (*P < 0.05; **P < 0.01). (C) Δ indicates the difference in mean values. Source data are available online for this figure.

Figure 3. GW274150 treatment reduces muscle wasting in the C26 model

Male BALB/C mice were injected subcutaneously with C26 cells (1.25 × 10⁶ cells) or an equivalent volume of saline. After 5 days, and everyday thereafter, saline and C26 injected mice were treated with or without GW (5 mg kg⁻¹). After 16 days, mice were euthanized and tissue samples were analyzed.

1. A

   Effect of GW on tumor weight.

2. B

   (left) Representative image of Western blot analysis of 3NT staining and stain free imaged total protein levels from quadriceps muscle. (right) Quantification of the 3NT‐to‐total protein ratio (saline n = 5, C26 n = 4, GW 5 mg/kg n = 5, and C26 GW 5 mg/kg n = 5).

3. C

   Percent body weight change from day 0 to day 16.

4. D–G

   Weight of tibialis anterior (D), gastrocnemius (E), quadricep (F), and soleus (G) muscle normalized to initial body weight.

5. H

   (left) Representative photomicrographs of gastrocnemius muscle sections from control and C26 mice treated with or without GW taken after H&E staining. Scale bars = 100 μm. (right) Frequency histogram showing the distribution of muscle fiber minimum Feret diameter in the gastrocnemius muscles from control and C26 mice treated with or without GW (saline n = 4, C26 n = 4, GW 5 mg/kg n = 3, and C26 GW 5 mg/kg n = 4). A total of 600‐700 fibers per muscle were used for the analysis. Statistical comparisons, mean, and standard deviation of the mean are shown in the histogram legend.

6. I

   Change in grip strength from before tumor cell injection (day 0) and before endpoint collection (day 16).

Data information: Individual data points represent values from individual mice, with a total of six mice per cohort (n = 6) unless stated otherwise. Error bars represent the standard deviation (SD) of the mean. (A) Δ indicates the difference in mean values, and P‐values were calculated with Student’s t‐test. (B–G, I) For statistical comparisons, Δ indicates the difference in mean values and P‐values were calculated with an ANOVA followed by Fisher's LSD test. (H) P‐values were calculated with a Kolmogorov–Smirnov test (***P < 0.001). Source data are available online for this figure.

Figure 4. Pharmacological inhibition of iNOS reduces C26‐induced derangement of amino acids and impairment of energy production

Male BALB/C mice were injected subcutaneously with C26 cells (1.25 × 10⁶ cells) or an equivalent volume of saline. After 5 days and everyday thereafter, the C26 tumor‐bearing mice were injected with either saline or GW 5 mg kg⁻¹. After 16 days, mice were euthanized and the metabolome of tibialis anterior muscles was analyzed through LC‐MS/MS.

1. Scores scatter plot of partial least square discriminant analysis (PLS‐DA) model of metabolites from saline, C26, and C26 + GW‐treated mice. Metabolomic data were range‐scaled and mean‐centered.

2. Pathway analysis using MetaboAnalyst 4.0 Software comparing significantly altered pathways from saline to C26 as well as C26 to C26 + GW. Pathways are ranked by their significance and filtered based on a Pathway Impact Score > 0.1. Metabolomic data were range‐scaled and mean‐centered. P‐values were obtained using GlobalTest, and the −log(P‐value) corresponding to a P‐value of 0.05 is indicated by the red dashed line.

3. (left) Western blot analysis of pThr172‐AMPK (pAMPK) and total AMPK (AMPK) in quadriceps muscle (n = 3). (right) Quantification of the pAMPK‐to‐AMPK ratio relative to the saline‐treated control (n = 3).

4. Relative concentrations of metabolites involved in glycolysis.

5. Relative concentrations of amino acids.

6. Relative concentrations of odd‐chain acylcarnitines.

7. Relative estimated activity of CPT1, CPT2, LCADH, and β‐oxidation.

Data information: Individual data points represent values from individual mice, with a total of six mice per cohort (n = 6) unless otherwise stated. (B) P‐values were obtained using GlobalTest. (C–G) Error bars represent the standard deviation (SD) of the mean. P‐values were calculated with an ANOVA followed by Fisher's LSD test (*P < 0.05; **P < 0.01; ***P < 0.001). (C) Δ indicates the difference in mean values. Source data are available online for this figure.

Figure 5. Cytokine treatment of C2C12 myotubes alters the levels of TCA cycle intermediates and activates AMPK in an iNOS‐dependent manner

C2C12 myotubes were treated with or without IFNγ (100 U/ml) and TNF‐α (20 ng/ml) and the indicated doses of GW. Protein content and metabolites were extracted from cells 24 h after treatment and analyzed for AMPK phosphorylation, ACC phosphorylation, and total iNOS levels as well as GC‐MS, respectively. Myotube integrity and widths as well as phosphorylation of S6 and S6K were assessed 48 h after treatment.

1. Representative immunofluorescence imaging for myoglobin and myosin heavy chain (MyHC) in not treated (NT) controls and IFNγ/TNF‐α (IT) samples treated with or without GW. Nuclei were visualized with DAPI staining (n = 4).

2. Quantification of mean fiber widths (n = 4).

3. (top) Western blot analysis for iNOS and tubulin (n = 3). (bottom) Media nitrite levels (n = 3).

4. Heatmap visualizing mean concentration corresponding to metabolites of IFNγ/TNF‐α (IT)‐treated samples relative to not treated (NT) controls (n = 3). Red and green indicate an increase or decrease in metabolite levels, respectively.

5. Western blot analysis for pThr172‐AMPK (pAMPK), total AMPK (AMPK), pSer79‐ACC (pACC), and total ACC (ACC).

6. Quantification of the pAMPK‐to‐AMPK ratio (left) and the pACC‐to‐ACC ratio (right) relative to the untreated control (n = 3).

7. Cellular ATP content quantified as a percentage of the untreated control (n = 3).

8. Western blot analysis for pThr389‐S6K (pS6K), total S6K (S6K), pSer235/236‐S6 (pS6), and total S6 (S6) (n = 5).

9. Quantification of the (left) pS6K‐to‐S6K ratio and the (right) pS6‐to‐S6 ratio relative to the untreated control (n = 5).

Data information: Individual data points represent independent experimental replicates. Error bars represent the standard deviation (SD) of the mean. For statistical comparisons, Δ indicates the difference in mean values and P‐values were calculated with an ANOVA followed by Fisher's LSD test. Source data are available online for this figure.

Figure 6. GW274150 prevents a cytokine‐induced shift to aerobic glycolysis in C2C12

C2C12 myotubes were treated with or without IFNγ (100 U/ml) and TNF‐α (20 ng/ml) and the indicated doses of GW. ATP production rates (J‐ATP) from oxidative phosphorylation (oxidative) and glycolysis (glycolytic) were determined from measurements of extracellular flux 24 h after treatment.

1. Bioenergetic profiles. Highlighted squares are defined by the theoretical maximal J‐ATP_oxidative and J‐ATP_glycolytic rates.

2. (left) Basal J‐ATP_glycolytic and J‐ATP_oxidative rates. (right) Total basal J‐ATP rate.

3. Glycolytic index of basal metabolism.

4. Total bioenergetic capacity.

Data information: Individual data points represent three independent experiments (n = 3). The data points for each experiment are calculated from the average of technical triplicates. Error bars represent the standard deviation (SD) of the mean. For statistical comparisons, Δ indicates the difference in mean values and P‐values were calculated with an ANOVA followed by Fisher's LSD test.

Figure 7. Inflammation‐mediated loss of complex II and IV integrity are reversed with GW274150

C2C12 myotubes were treated with or without IFNγ (100 U/ml) and TNF‐α (20 ng/ml) and the indicated doses of GW. Protein content was extracted 24 h after treatment.

1. Western blot analysis for ETC protein complex subunits.

2. (left) Quantification of SDHB (complex II; CII) normalized to VDAC (outer mitochondrial membrane; OMM) and relative to untreated control. (right) Quantification of MTCO1 (complex IV; CIV) normalized to VDAC and relative to untreated control.

Data information: Individual data points are from four independent experimental replicates (n = 4). Error bars represent the standard deviation (SD) of the mean. For statistical comparisons, Δ indicates the difference in mean values and P‐values were calculated with an ANOVA followed by Fisher's LSD test. Source data are available online for this figure.

Figure 8. Inflammation‐mediated loss of mitochondrial integrity is reversed with genetic and pharmacological inhibition of iNOS

1. Male C57BL/6 wild‐type (WT) and iNOS knockout (KO) mice were intraperitoneally injected with 1 mg kg⁻¹ LPS or an equivalent volume of carrier solution. Control WT, control KO, and LPS‐treated KO cohorts were pair‐fed (PF) to the WT LPS‐treated cohorts. After 18 h, mice were euthanized, and tibialis anterior muscles were imaged by transmission electron microscopy. (left) Representative micrograph of tibialis anterior muscle (n = 2). Scale bar = 1 µm. (right) Zoomed section of representative image to highlight mitochondria. Scale bar = 0.5 µm.

2. Male BALB/C mice were injected subcutaneously with C26 cells (1.25 × 10⁶ cells) or an equivalent volume of saline. After 5 days, and everyday thereafter, saline‐ and C26‐injected mice were treated with or without GW (5 mg kg⁻¹⁾. After 16 days, mice were euthanized and gastrocnemius muscles were imaged by transmission electron microscopy. (left) Representative micrograph of gastrocnemius muscle (n = 2). Scale bar = 1 µm. (right) Zoomed section of representative image to highlight mitochondria. Scale bar = 0.5 µm.

Data information: Images representative of 2 individual mice.

Figure 9. Schematic depicting the role of the iNOS/NO pathway in promoting energy crisis during cachexia‐induced muscle wasting

(outlined in black) Upon activation by underlying diseases such as cancer, the iNOS/NO pathway is activated. In turn, iNOS inhibits mitochondrial energy production in skeletal muscles reducing ATP production which leads to muscle wasting. As a result of these effects, anabolic signaling is reduced and muscle loss is induced. (highlighted in red) On the hand, pharmacological inhibition of iNOS activity with drugs such as GW274150 (GW) reverses these effects, preventing muscle wasting.