Mitochondrial-cytochrome c oxidase II promotes glutaminolysis to sustain tumor cell survival upon glucose deprivation

Abstract

Glucose deprivation, a hallmark of the tumor microenvironment, compels tumor cells to seek alternative energy sources for survival and growth. Here, we show that glucose deprivation upregulates the expression of mitochondrial-cytochrome c oxidase II (MT-CO2), a subunit essential for the respiratory chain complex IV, in facilitating glutaminolysis and sustaining tumor cell survival. Mechanistically, glucose deprivation activates Ras signaling to enhance MT-CO2 transcription and inhibits IGF2BP3, an RNA-binding protein, to stabilize MT-CO2 mRNA. Elevated MT-CO2 increases flavin adenosine dinucleotide (FAD) levels in activating lysine-specific demethylase 1 (LSD1) to epigenetically upregulate JUN transcription, consequently promoting glutaminase-1 (GLS1) and glutaminolysis for tumor cell survival. Furthermore, MT-CO2 is indispensable for oncogenic Ras-induced glutaminolysis and tumor growth, and elevated expression of MT-CO2 is associated with poor prognosis in lung cancer patients. Together, these findings reveal a role for MT-CO2 in adapting to metabolic stress and highlight MT-CO2 as a putative therapeutic target for Ras-driven cancers.

Fig. 1. MT-CO2 is essential for tumor cell survival and tumor growth upon glucose deprivation.

a–c H292 or glucose-starvation-resistant H292 (H292-V) cells (5 × 10⁵) were subcutaneously inoculated into the right scruff of the female BALB/c nude mouse (n = 5/group). The tumor volumes on indicated days after injection were examined (a). Twenty days post-injection, tumors were weighed (c), and photos were taken (b). H292 or H292-V cells were subjected to qPCR analyses (d, n  =  3 independent experiments) or were subjected to western blot analyses (e). f, g H1299 or A549 cells stably expressing shGFP, shMT-CO2-1, or shMT-CO2-2 were subjected to western blot analyses (f) or were grown in DMEM containing 2 mM glutamine in the absence of glucose (Glc−/Gln+) for 24 h. Cell viability was determined by trypan blue exclusion assay (g, n  =  3 independent experiments). h–l H1299 or A549 cells stably expressing shGFP, shMT-CO2−1, or shMT-CO2-2 were subjected to MTS analyses for cell growth (h, i, n  =  3 independent experiments). For tumor growth, indicated cells (1 × 10⁶) were subcutaneously inoculated into the right scruff of the female BALB/c nude mouse (n = 8/group). The tumor volumes on indicated days after injection were examined (j). Thirty-two days post-injection, tumors were weighed (l) and photos were taken (k). m–r H292-V cells stably expressing shGFP or shMT-CO2 were subjected to western blot analyses (m) or were grown in Glc−/Gln+ condition for 48 h followed by trypan blue exclusion assay for cell viability (n, n  =  3 independent experiments) or were subjected to MTS analyses for cell growth (o, n  =  3 independent experiments). For tumor growth, cells (5 × 10⁵) were subcutaneously inoculated into the right scruff of the female BALB/c nude mouse (n = 5/group). The tumor volumes on indicated days after injection were examined (p). Twenty-five days post-injection, tumors were weighed (r) and photos were taken (q). These experiments have been repeated three times with similar results (e, f). Data were presented as mean ± SD (d, g, h, i, n, o) or SEM (a, c, j, l, p, r). Comparisons were performed with one-way ANOVA with Tukey’s test (g, l, n, r), two-way ANOVA with Tukey’s test (a, h–j, o, p), and unpaired two-tailed Student’s t test (c, d).

Fig. 2. MT-CO2 elevates GLS1 expression in facilitating glutaminolysis to promote tumor cell survival upon glucose deprivation.

a, b A549 or H1299 cells grown in DMEM containing 2 mM glutamine in the absence of glucose (Glc−/Gln+) were treated with as indicated inhibitors for 36 h (A549) or 48 h (H1299). Cell viability was determined by trypan blue exclusion assay (n  =  3 independent experiments). c A sketch depicts the process of glutaminolysis. d H1299 cells stably expressing shGFP or shMT-CO2 were cultured in DMEM containing 2 mM ¹³C₅-glutamine for 8 h. Cells were subjected to mass isotopomer analyses (n = 4 biologically independent samples per experiment). e–i H1299 or A549 cells stably expressing shGFP, shMT-CO2-1, or shMT-CO2-2 were subjected to qPCR (e, f, n  =  3 independent experiments) or western blot (g) analyses or cells were grown in Glc-/Gln+ with or without supplement of 500 μM dimethyl α-Ketoglutaric acid (DM-αKG) for 24 h. Cell viability was determined by trypan blue exclusion assay (h, i, n  =  3 independent experiments). j, k H1299-shMT-CO2-1cells stably expressing HA-GLS1 were subjected to western blot analyses (j) or cells were grown in Glc-/Gln+ condition for 24 h. Cell viability was determined by trypan blue exclusion assay (k, n  =  3 independent experiments). The samples derive from the same experiment but different gels for GLS1, MT-CO2, GAPDH, and another for HA were processed in parallel (j). l–r Cells (1 × 10⁶) were subcutaneously inoculated into the right scruff of the female BALB/c nude mouse (n = 6/group). The tumor volumes on indicated days after injection were examined (l). Thirty-two days post-injection, tumors were weighed (n) and photos were taken (m). The tumor samples were subjected to immunohistochemistry staining analyses for Ki67, MT-CO2, or GLS1 expression (o). The protein levels were quantified by average optical density (AOD) and were presented (p–r). Scale bar, 50 μm. These experiments have been repeated three times with similar results (g, j). Data were presented as mean ± SD (a, b, e, f, h, i, k) or SEM (l, n, p–r). Comparisons were performed with one-way ANOVA with Tukey’s test (a, b, e, f, k, n, p–r), two-way ANOVA with Tukey’s test (l), and unpaired two-tailed Student’s t test (h, i).

Fig. 3. MT-CO2 facilitates JUN transcription through the FAD-LSD1-H3K9me2 axis to promote GLS1 expression.

a–c Indicated cells were subjected to RNA-seq (a, n = 2 biologically independent samples per experiment), qPCR (b, n = 3 independent experiments), or western blot (c) analyses. d, e H1299-shMT-CO2 cells were transfected with c-JUN for 48 h, followed by western blot (d) or qPCR (e, n  =  3 independent experiments) analyses. The samples derive from the same experiment but different gels for GLS1, MT-CO2, GAPDH, and another for c-JUN were processed in parallel (c, d). f–i Indicated cells were subjected to measurement of the oxygen consumption rate (OCR) (f, h) or complex (II, III, IV) activity (g, i) (n  =  3 independent experiments). j Indicated cells were subjected to examine cellular FAD levels (n  =  3 independent experiments). k H1299 cells were treated with an indicated concentration of FAD for 48 h, followed by western blot analyses. l–o Indicated cells were treated with an indicated concentration FAD or riboflavin for 48 h, followed by examining the cellular FAD levels (l, n, n = 3 independent experiments) or western blot analyses (m, o). The samples derive from the same experiment but different gels for GLS1, H3K9me2, GAPDH, and another for c-JUN, MT-CO2 were processed in parallel (m, o). p H1299 cells were treated with an indicated concentration of GSK-LSD1 for 36 h, followed by western blot analyses. The samples derive from the same experiment but different gels for GLS1, H3K9me2, GAPDH, and another for c-JUN, H3K4me2 were processed in parallel (k, p). q Indicated cells were subjected to western blot analyses. The samples derive from the same experiment but different gels for LSD1, H3K9me2, GAPDH, another for EZH2, H3K27me3, another for MT-CO2, and another for H3K4me2 were processed in parallel. r, s H1299 cells were treated with 20 μM GSK-LSD1 for 36 h (r) or H1299-shMT-CO2 cells were treated with 20 μM FAD for 48 h (s), followed by CHIP-qPCR analyses (n  =  3 independent experiments). These experiments have been repeated three times with similar results (c, d, k, m, o–q). Data were presented as mean ± SD (b, e, f–j, l, n, r, s). Comparisons were performed with one-way ANOVA with Tukey’s test (b, e, j, l, n) and unpaired two-tailed Student’s t test (g, i, r, s).

Fig. 4. Glucose deprivation activates Ras signaling to promote MT-CO2 expression critical for cancer cell survival and tumor growth.

a H292 or glucose-starvation-resistant H292 (H292-V) cells were subjected to western blot analyses. b–d H292 or H1975 cells expressing KRas^G12V were subjected to qPCR (b, n  =  3 independent experiments) or western blot (c) analyses or were grown in DMEM containing 2 mM glutamine in the absence of glucose (Glc−/Gln+) for 24 h, followed by examining cell viability (d, n  =  3 independent experiments). e–h A549 or H292-V cells expressing shGFP, shKRas-1, or shKRas-2 were subjected to western blot analyses (e, g) or cells were grown in Glc−/Gln+ condition for 24 h (A549) or 48 h (H292-V), followed by examining cell viability (f, h, n  =  3 independent experiments). i, j H292-KRas^G12V or H1975-KRas^G12V cells expressing shMT-CO2 were subjected to western blot analyses (i) or cells were grown in Glc−/Gln+ condition for 24 h, followed by examining cell viability (j, n  =  3 independent experiments). k–q H292-KRas^G12V cells stably expressing shMT-CO2 were subjected to MTS analyses (k, n  =  3 independent experiments). For tumor growth, cells (8 ×10⁵) were subcutaneously inoculated into the right scruff of the female BALB/c nude mouse (n = 7/group). The tumor volumes on indicated days after injection were examined (l). Twenty days post-injection, tumors were weighed (n) and photos were taken (m). Immunohistochemistry staining analyses were performed to examine Ki67 and MT-CO2 expression (o–q). Scale bar, 50 μm. r, s The transgenic mice (Rosa26-LSL-KRas^G12D, C57BL/6) were used to assess the effects of MT-CO2 on KRas^G12D-induced lung tumor growth (n = 5/group), as described in Materials and Methods. Lungs were stained by H&E for histological examination (r). The numbers of observable nodules in the lung were presented (s). The samples derive from the same experiment but different gels for pERK, Ras, GAPDH, and another for ERK, MT-CO2 were processed in parallel (a, c, e, g, i). These experiments have been repeated three times with similar results (a, c, e, g, i). Data were presented as mean ± SD (b, d, f, h, j, k) or SEM (l, n, p, q, s). Comparisons were performed with one-way ANOVA with Tukey’s test (f, h, j, n, p, q), two-way ANOVA with Tukey’s test (k, l), and unpaired two-tailed Student’s t test (b, d, s).

Fig. 5. Glucose starvation specifically facilitates MT-CO2 expression via downregulation of IGF2BP3.

a m6A RIP-qPCR was performed to analyze the levels of m6A-modificated mitochondrial genes in H1299 cells (n = 3 independent experiments). b, c H292 or glucose-starvation-resistant H292 (H292-V) cells were subjected to RNA-seq analyses (b, n = 2 biologically independent samples per experiment) or were subjected to western blot analyses (c). d, e H1299 cells expressing shGFP, shIGF2BP3-1, or IGF2BP3-2 were subjected to western blot (d) or qPCR (e, n  =  3 independent experiments) analyses. f, g H1299 cells expressing IGF2BP3 or vector control (Vec) were subjected to western blot (f) or qPCR (g, n  =  3 independent experiments) analyses. The samples derive from the same experiment but different gels for IGF2BP3, MT-CO2, GAPDH, another for MT-ND1, and another for MT-CO1, MT-ATP6 were processed in parallel (d, f). h, i H1299 cells expressing shGFP, shIGF2BP3-1, or IGF2BP3-2 (h) or H1299 cells expressing IGF2BP3 or Vec (i) were treated with actinomycin D (10 μM) for an indicated time point. Cells were subjected to qPCR analyses. The half-life of the MT-CO2 mRNA levels was shown (n  =  3 independent experiments). j IGF2BP3 RIP-qPCR was performed to analyze the interaction between IGF2BP3 and mitochondrial mRNAs in H1299 cells (n  =  3 independent experiments). k H292-V cells expressing IGF2BP3 or Vec were subjected to western blot analyses. These experiments have been repeated three times with similar results (c, d, f, k). Data were presented as mean ± SD (a, e, g, h–j). Comparisons were performed with two-way ANOVA with Tukey’s (h) or Bonferroni’s (i) test and unpaired two-tailed Student’s t test (a, e, g, j).

Fig. 6. Elevated MT-CO2 expression is linked to oncogenic Ras and is positively correlated with the expression of c-JUN and GLS1, and with poor clinical outcomes in human lung cancer.

a–d Human lung cancer tissue microarrays consisting of 27 pairs of tumors and adjacent tissues were subjected to immunohistochemistry analyses for MT-CO2 and GLS1 expression (a) with quantitative analyses using average optical density (AOD) (b, c). The Pearson correlation between MT-CO2 and GLS1 in protein levels was analyzed (d). Scale bar, 200 μm. e–g Human lung tumor specimens harboring wild-type KRAS alleles (KRAS^WT, n = 11) or a KRAS mutant allele (KRAS^MT, n = 11) were qPCR analyzed for MT-CO2 (e), GLS1(f), and c-JUN (g) mRNA levels. h–j Human lung tumor samples (n = 23) were qPCR analyzed for MT-CO2, GLS1, and c-JUN mRNA levels. The Pearson correlation between MT-CO2 and GLS1 (h) or c-JUN (i) in mRNA levels was analyzed. The correlation between MT-CO2 mRNA levels and the overall survival (OS) in lung cancer patients was analyzed (j). k The correlation between GLS1 mRNA levels and overall survival (OS) in lung cancer patients was analyzed using the Kaplan–Meier Plotter database. Data were presented as mean ± SEM (b, c, e, f, g). Statistical analyses were performed with unpaired two-tailed Student’s t test (b, c, e, f, g), Log-rank (Mantel–Cox) test (j, k), and two-tailed Pearson correlation analysis (d, h, i).

Fig. 7. Berberine inhibits the MT-CO2-GLS1 axis to suppress oncogenic Ras-induced glutaminolysis and tumor growth.

a, b Indicated cells were treated with an indicated concentration of berberine for 24 h prior to western blot analyses (a) or were treated with berberine (5 μM, and thereafter) in the presence of 500 μM dimethyl α-Ketoglutaric acid (DM-αKG) upon Glc−/Gln+ condition for 24 h, followed by examining cell viability (b, n = 3 independent experiments). c, d H292-KRas^G12V cells were treated with berberine for 24 h prior to western blot analyses (c) or were treated with berberine in the presence of 500 μM DM-αKG upon Glc-/Gln+ condition for 24 h, followed by examining cell viability (d, n  =  3 independent experiments). The samples derive from the same experiment but different gels for MT-CO2, GLS1, GAPDH, another for pERK, Ras, and another for ERK were processed in parallel (c). e–l H292-KRas^G12V (8 ×10⁵, e–h) or A549 (1 ×10⁶, i–l) cells were subcutaneously inoculated into the right scruff of the female BALB/c nude mouse (n = 6/group, H292; n = 8/group, A549). On day 14 (H292) or 22 (A549) after inoculation, mice were given berberine (40 mg/kg) by oral gavage daily. The tumor volumes on indicated days after gavage were examined (e, i). On day 16 (H292) or 12 (A549) after gavage, tumors were weighed (f, j). Immunohistochemistry staining was performed to examine MT-CO2 and GLS1 expression (g, h, k, l). m–o Lewis lung carcinoma (LLC, 5 ×10⁵) cells were subcutaneously inoculated into the right scruff of each C57BL/6 mice (n = 7/group). On day 7 after inoculation, mice were given berberine (40 mg/kg) by oral gavage daily. The tumors volume on indicated days after gavage were examined (m). On day 16 after gavage, tumors were weighed (n). The mice’s overall survival (Tumor volume >1500 mm³ is defined as ethical death) was examined (n = 7/group) (o). p A model depicting the role of MT-CO2 in tumor cell survival upon glucose deprivation. These experiments have been repeated three times with similar results (a, c). Data were presented as mean ± SD (b, d) or SEM (e–n). Comparisons were performed with one-way ANOVA with Tukey’s test (b, d, f–h), two-way ANOVA with Tukey’s (e) or Bonferroni’s (i, m) test, unpaired two-tailed Student’s t test (j–l, n), and Log-rank (Mantel–Cox) test (o).