Changing TFAM Expression Modulates HNC Malignancy
To examine the roles of TFAM and its targeting mtDNA during HNC development, multiple in vitro analyses in HNC cells, with different origins in response to TFAM changes, were performed. The results showed that TFAM mRNA is reduced in tested HNC cells in contrast to normal human oral fibroblasts (OF) (Fig. 1A). Next, HNC cells deficient (Fig. 1B) or enforced expressing for TFAM (Fig. 1C) were utilized to better determine the role of TFAM in controlling HNC tumorigenicity. In addition to protein expression, decreased copy number of mtDNA encoded ETC subunits (potential TFAM targets) including ND1, ND2, ND3, ND4, ND4L, ND5, ND6 (Complex I, Fig. 1D), CytB (Complex III, Fig. 1E), Cox I, Cox II, Cox III (Complex IV, Fig.1F) as well as ATP8, ATP6 (Complex V, Fig. 1G) were detected by real-time PCR analysis in TFAM-silencing HNC cells when compared with control cells. At the translational level, while no significant changes for nuclear encoded ETC proteins (NDUFB8, SDHB, UQCRC2, ATP5A) were found in TFAM knockdown HNC cells, mtDNA encoded COXII protein was greatly down-regulated in TFAM-silencing SAS, OECM1 and HSC3 cells (Fig. 1H), highlighting the knockdown efficiency in our experimental setting. In short, shRNA mediated TFAM silencing could functionally reduce TFAM and its targeting cues in HNC cells.
Next, multifaceted cellular assays were performed in TFAM deficient and overexpressing HNC cells. By using Trypan blue exclusion and MTT assays, it was shown that TFAM knockdown led to increased HNC cell proliferation in vitro (Fig. 2A). In contrast, cell growth was downregulated in TFAM overexpressing HNC cells when compared with control cells (Fig. 2B). Similar to in vitro growth, larger HNC-bearing xenografic tumors were detected in vivo in most tested HNC cells (Fig. 2C) implying that TFAM expression is negatively associated with HNC cell growth. Further analysis found that TFAM deficiency resulted in reduced G0/G1 phase but increased distribution in both S and G2M phases (Fig. 2D). Meanwhile, decreased apoptotic rate was detected in TFAM-silencing HNC cells (Fig. 2E), indicating that TFAM mediated regulation for HNC cell growth is due to modulation of both cell cycling and cell apoptosis.
As enhanced metastatic activity and drug resistance are also hallmarks of cancer cells [27, 28], cell motility and chemodrug treatment, sensitivity was next assessed in HNC cells that are deficient for TFAM expression. By using Transwell-based migration assays, TFAM loss significantly promoted HNC cell migration when compared with control cells (Fig. 2F). As for the changes of therapeutic sensitivity in response to reduced TFAM expression in HNC cells, the half maximal inhibitory concentrations (IC50) of CDDP, 5-FU and TAXOL, the most common chemodrugs for HNC in clinic, were determined. A greater half maximal inhibitory concentration (IC50) for CDDP, 5-FU and Taxol was detected in TFAM deficient HNC cells when compared with control cells (Fig. 2G). Taken together, TFAM negatively correlated with malignancy index in HNC cells, confirming that TFAM acts as a tumor suppressing factor in HNCs.
TFAM Deficiency Modulates Metabolic Plasticity in HNC cells
Metabolic plasticity was recently demonstrated in numerous studies showing that cancers could evolve to adapt environmental stresses allowing cell survival during progression [29]. Our previous findings demonstrated that manipulations for pyruvate metabolic molecules LDHA and PDHA1 led to a metabolic shift between a wide spectrum of metabolic pathways and further analysis confirmed that this metabolic reprogramming is essential for LDAH/PDHA1 mediated malignant changes in HNC cells [26]. We herein tested whether a metabolic shift could also be responsible for TFAM mediated cellular changes. Multiple bioenergetic readouts including glucose uptake, intracellular pyruvate level, extracellular lactate production, PDH activity and intracellular ATP level in TFAM-silencing HNC cells were examined. No significant difference for glucose uptake activity and intracellular pyruvate content between TFAM-silencing and control HNC cells was detected (Fig. 3A, B) whereas increasing ATP level is detected in response to TFAM loss (Fig. 3C). Interestingly, lactate secretion was elevated (Fig. 3D) while decreased PDH activity, using either Western blot analysis for phosphorylation status (Ser293) of PDHA1 or colorimetric assay for PDH activity was detected (Fig. 3E, F). Mitochondrial analysis using Mitotracker red staining and Seahorse Mito-stress analyzer showed that the mitochondrial membrane potential (Fig. 3G) as well as basal and maximal respiration (Fig. 3H), was down-regulated in TFAM-silencing HNC cells when compared with control cells, showing that TFAM loss could induce Warburg phenotype. Interestingly, Reactive Oxygen Species (ROS), a byproduct of Oxidative Phosphorylation (OxPhos) reaction, was increased in most TFAM-silencing HNC cells compared with control counterparts whereas mtDNA encoded COXII was negatively correlated with ROS level (Fig. 1H & Fig. 3I). This observation revealed a possibility that TFAM loss might trigger ETC dysfunction thereby resulting in greater ROS leak and eventually increased oncogenicity.
TFAM mediated metabolic shift was further evident by a LC-MS based metabolomics analysis for glycolytic and TCA cycle metabolites. The analysis shows that only metabolites in the “payoff” phase of glycolysis including 1,3 bisphosphoglycaerate, 2-/3-phosphoglycerate, and phosphoenoalpyruvate (PEP) were decreased in TFAM-silencing HNC cells while other intracellular metabolites remain unchanged (Fig. 3J and Fig.S1-2). This result could be explained as a consequence of higher glycolytic metabolism in response to TFAM loss, thereby leading to a greater consumption of glyceraldehyde-3-phosphate (G-3-P) in order to meet energetic demand. As new glucose input was not significantly altered, we further suspected that there might be alternative changes of other cellular metabolic cues, such as amino acid metabolism, that could potentially compensate deregulated mitochondrial activity in a condition of TFAM knockdown to support tumorigenic activity in HNC cells. Nevertheless, no significant changes for amino acid level were found in TFAM-silencing HNC cells when compared with control cells (Fig.S3), implying that biomolecules might be not limiting for TFAM mediated HNC oncogenic regulation. Taken together, our results demonstrated that TFAM loss facilitated HNC cell malignancy likely via intrinsic metabolic reprogramming away from mitochondrial metabolism towards to aerobic glycolysis without altering external nutrition uptake and nitrogen metabolic pathway.
Oncogenic Akt and ERK Signaling Pathways Regulate TFAM Mediated HNC Oncogenicity
In addition to metabolic changes, it is widely accepted that activation of various oncogenic pathways is essential for cancer development [30, 31]. Numerous studies have shown that Akt-mTORC and EGFR-ERK1/2 signaling pathways are highly expressed in HNC cells and crucial for HNC carcinogenic identity, both in vivo and in vitro [32-36]. Therefore, we investigated if Akt/ERK signals are key regulators for TFAM mediated oncogenic changes in HNC cells. Western blot analysis showed that phosphorylated Akt (Ser473), phosphorylated p44/42 MAPK ERK1/2 (Thr202/Tyr204) and mTORC pathway effector phosphorylated S6 Ribosomal Protein (Ser235/Ser236) were all upregulated in TFAM-silencing HNC cells compared with control cells (Fig. 4A). At the cellular level, IFA further confirmed that HNC cells, with reduced TFAM expression, expressed a higher amount of phosphorylated Akt/ERK/S6 proteins, indicating that TFAM mediated malignant changes could be made through the modulations of ERK1/2 and Akt-mTORC-S6 signaling pathways (Fig. 4B). To further define the significance of Akt/ERK signaling pathways in regulating TFAM-mediated neoplastic characteristics, ERK1/2 inhibitor PD98059 and PKB/Akt inhibitor MK2206 were applied in TFAM-silencing HNC cells and cell proliferation was examined. The results showed that efficient inhibition of ERK1/2 and AKT activity (Fig.S4-5) significantly abolished increased cell growth in TFAM-silencing HNC cells. Strikingly, a combinational treatment of PD98059 and MK2206 exhibited a dose-dependent synergetic effect in controlling TFAM mediated HNC cell growth (Fig. 4C). These results confirmed a novel notion that mtDNA loss in response to TFAM deficiency could trigger cytosolic signaling alteration. Interestingly, previous studies have reported that the recruitment of Akt protein to mitochondria could inactivate PDC thus resulting in downregulation of OxPhos pathway under hypoxic condition [37], further supporting a potential crosstalk between Akt signal and mitochondrial metabolism. Collectively, these findings provided an alternative scheme for development of TFAM/Akt-ERK combinational anti-cancer therapeutic strategy for HNCs.
Decreased TFAM and Its Downstream Genes in Human HNCs
As previous studies found that TFAM and mtDNA levels were positively correlated with colorectal cancer prevalence but negatively associated with liver cancer frequency [38, 39], the association between TFAM expression and HNC oncogenicity needs further determination through additional clinical trials. To this end, the expression of TFAM and its target genes in HNC and their corresponding normal tissues was analyzed using The Cancer Genome Atlas (TCGA) based databases. The results showed a positive correlation between TFAM mRNA expression and mtDNA encoded ETC I/II/IV/V transcripts (Fig. 5A). Moreover, the TFAM mRNA level was positively correlated with mRNA expression for OxPhos factors (PDHA1, PGC1α, PPARGC1β) but negatively associated with glycolytic enzymes (HK2, PFKM, PGK1) (Fig. 5B). The prognostic significance of TFAM in HNC patients was also defined and it was found that patients bearing HNCs with greater TFAM expression tend to have a better Overall Survival (OS) rate (Fig. 5C). Interestingly, HNC patients with TFAM genetic alteration exhibited worse OS rate than HNC patients without TFAM alteration (Median Survival: 22.19 months vs. 56.94 months) indicating that maintenance of TFAM integrity could be essential for better prognosis in HNC patients (Fig. 5D) The expression of TFAM and its downstream targets was further examined in IRB approved paired adjacent normal (N) and tumor (T) clinical tissues from HNC patients (N=18; Table S1). In agreement with database analysis, TFAM mRNA expression and copy numbers of all mtDNA encoded ETC genes were significantly downregulated in tumors compared with corresponding normal tissues, suggesting that decreased TFAM levels might be important for HNC development (Fig. 5E). Further analysis to define a potential association of TFAM/mtDNA encoded ETC gene expression and disease progression stratified by TNM scaling was next determined. Even though the data showed no statistical significance, all mtDNA encoded ETC genes showed a decreasing trend over a T (from T1 to T4) and N stages (from N0 to N2) (Fig.6-7), suggesting that mtDNA encoded ETC genes reversely associated with disease severity. To draw a more definite conclusion, a larger number HNC samples could be required.