Lysosomal/mitochondrial interaction promotes tumor growth in squamous cell carcinoma of the head and neck

Abstract

Communication between intracellular organelles including lysosomes and mitochondria has recently been shown to regulate cellular proliferation and fitness. The way lysosomes and mitochondria communicate with each other (lysosomal/mitochondrial interaction, LMI) is, emerging as a major determinant of tumor proliferation and growth. About 30% of squamous carcinomas (including squamous cell carcinoma of the head and neck, SCCHN) overexpress TMEM16A, a calcium-activated chloride channel, which promotes cellular growth and negatively correlates with patient survival. We have recently shown that TMEM16A drives lysosomal biogenesis, but its impact on mitochondrial function has not been explored. Here, we show that in the context of high TMEM16A SCCHN, (1) patients display increased mitochondrial content, specifically complex I; (2) In vitro and in vivo models uniquely depend on mitochondrial complex I activity for growth and survival; (3) NRF2 signaling is a critical linchpin that drives mitochondrial function, and (4) mitochondrial complex I and lysosomal function are codependent for proliferation. Taken together, our data demonstrate that coordinated lysosomal and mitochondrial activity and biogenesis via LMI drive tumor proliferation and facilitates a functional interaction between lysosomal and mitochondrial networks. Therefore, inhibition of LMI instauration may serve as a therapeutic strategy for patients with SCCHN.

Introduction

There are many examples of cancer cells hijacking, enhancing, or otherwise realigning the role normally played by organelles to support increased proliferation (1, 2). Alterations in the mitochondrial function can lead to an increase energy production, resistance to apoptosis, oxidative stress, and uncontrolled cell growth and division, contributing to the development and progression of cancer (3). Dysregulation of lysosomal function can enhance sequestration and decrease bioavailability of chemotherapeutic drugs, increase degradation of pro-apoptotic proteins leading to suppressed apoptosis, and enhance secretion of the lysosomal enzymes contributing to the degradation of the extracellular matrix enabling cancer cell invasion and metastasis (4).

The regulation of bioenergetics in tumor cells impacts both cellular proliferation and tumor metastasis (5). Interestingly, mitochondrial electron transport chain is required for glucose and glutamine metabolism in certain tumors (6). Similarly, inhibition of complex I rescues resistance to B-Raf inhibition in melanomas (7). These findings suggest that mitochondrial respiration is intimately linked with tumor growth and treatment resistance.

The synergistic interaction between lysosomes and mitochondria and its role in tumor growth and proliferation is a topic of recent interest (8). Prior investigations have focused on mitophagy as the primary outcome of the interaction between lysosomes and mitochondria (9). However, a newer focus has shifted to signaling and functional interactions between these organelles in pathways distinct from mitophagy.

An emerging question is whether regulation of lysosomal biogenesis and mitochondrial function (defined by either biogenesis or respiration) is coordinated (10). Lysosomal biogenesis requires the EB family of transcription factors (including TFEB and MiTF) (11). Mitochondrial biogenesis is controlled by a system of transcription factors that are activated by hypoxic signals and other inputs pertaining to redox regulation (12). While the mounting evidence suggests interaction between lysosomal and mitochondrial compartments exists (13, 14) the impact of this coordination in malignant cells has yet to be consistently explored. The central question pursued in these studies is whether and how the lysosomal and mitochondria functionally interact with each other (biogenesis and function); and whether this interaction can be exploited to target cancer cells.

SCCHN is a devastating disease with overall survival of ~50% at 5 years. In about 30% of cases, this disease is characterized by amplification of chromosomal band 11q13, associated with much poorer oncologic outcomes (15, 16). Among other genes, the 11q13 region encodes the TMEM16A calcium-activated chloride channel (15, 16). Overexpression of TMEM16A, partially driven by gene amplification, activates several signaling pathways that promote tumor growth, including lysosomal biogenesis (17). Therefore, we used this model system to interrogate the interplay between lysosomal biogenesis and mitochondrial bioenergetics.

We have previously shown upregulated lysosomal biogenesis in TMEM16A-overexpressing SCCHN models (17). Recent studies have identified the role of NRF2 pathway in the regulation of TFEB/TFE3-dependent lysosomal biogenesis (18) and mitochondrial biogenesis via NRF1 and PPARGC1alpha (19). However, these effects have not been studied in SCCHN. We posit that TMEM16A drives the mitochondrial signaling network that is promoted by NRF2. Therefore, we sought to interrogate the mechanism(s) that regulate this interaction. Using a combination of bioinformatics, human tumor analyses, and in vivo murine models, we demonstrate that lysosomal biogenesis is coordinately regulated with the expression of mitochondrial complex I. Mechanistically, we show that TMEM16A regulates mitochondrial function via activation of the oxidative stress response transcription factor NRF2. This coordinated signaling ultimately promotes cellular fitness and tumor proliferation. Pharmacologic inhibition of mitochondrial complex I is particularly effective in tumors that display this coordinate regulation of lysosomal biogenesis and complex I expression. Taken together, these data show that expression of these markers can be used to stratify patients who may benefit from tumor-specific complex I inhibition.

Materials and Methods

Cell lines

All cell lines were STR authenticated and confirmed mycoplasma-negative using the e-Myco^™ Mycoplasma PCR Detection Kit (Boca Scientific, Inc.). Human tongue squamous cell carcinoma OSC19 (RRID:CVCL_3086) were grown in high glucose DMEM supplemented with 10% FBS. OSC19 were transduced with retroviral pBABE-puromycin (RRID:Addgene_21836) control or TMEM16A plasmid as previously described here (33). HN30 (RRID:CVCL_5525) were grown in high glucose DMEM supplemented with 10% FBS, 1X MEM non-essential amino acids, 1mM sodium pyruvate and 1X MEM vitamin. Mouse oral squamous cell carcinoma MOC1 (RRID:CVCL_ZD32) cells were grown in Iscove’s Modified Dulbecco’s Medium (IMDM)/Ham’s F-12 Nutrient Mixture 2:1, 5% FBS, 1% penicillin/streptomycin mixture, 5 mg/mL insulin, 40 mg hydrocortisone, 5 mg epidermal growth factor (EGF). MOC1 cells were engineered to overexpress TMEM16A or negative using lentivirus system. After viral infection, cells were selected in puromycin for at least 2 weeks. TMEM16A expression in the stable cells was confirmed by qPCR.

siRNA Transfection

About 70% confluent cells in a 6-well plate, were transfected with either 10 μM control (QIAGEN) or esiRNA (Sigma-Aldrich) using TransIT-siQUEST reagent (Mirus Bio) following manufacturer’s protocol for 48–72h. Post transfection, these cells were used for qPCR, ATPlite assays and colony forming assays.

Confocal Microscopy

OSC19-VC and OSC-19TMEM16A cells were plated on coverslips and allowed to attach overnight. After formaldehyde fixation, cells were permeabilized in 0.1% triton (in PBS), blocked for one hour at room temperature and incubated for overnight at 4 degrees in primary antibodies against Lamp1 (Santa Cruz Biotechnology Cat# sc-20011, RRID:AB_626853) and Tomm20 (Thermo Fisher Scientific Cat# PA5-52843, RRID:AB_2648808) at 1:1000 dilution. Next day, the cells were washed three times in PBS+0.05% tween and incubated in secondary mouse (Molecular Probes Cat# A-11019, RRID:AB_143162) and rabbit antibody (Thermo Fisher Scientific Cat# A27034, RRID:AB_2536097) at 1:5000 dilution for one hour at room temperature. The cells are washed atleast 3 times in PBS+0.05% tween and stained with DAPI (Invitrogen #62248) for 10 minutes. The cells are washed and mounted in fluoro-gel (Electron Microscopy Sciences #17985-10). Confocal z-stacks were collected using an Olympus Fluoview 1000 II confocal microscope equipped with a 60x (1.4NA) optic, and analyzed using NIS-Elements (Nikon Inc., Melville NY). Co-localization was assessed by segmenting the lysosomal and mitochondrial signals based on intensity, size and shape. The percentage of total lysosomes containing mitochondria content was calculated using a binary “and” statement and expressed relative to total lysosomal count. At least 15 non-overlapping images were taken per experiment.

Cell Proliferation Assay

Cells were plated in triplicate at a confluency of 2000–4000 cells/well in a 96-well plate. Following 24 hours, cells are treated as indicated. After 72 hours of treatment, cell proliferation was measured using WST-1 Reagent (Takara Bio, Inc., Clontech Laboratories, Inc.) as per manufacturer’s instructions. The absorbance was taken at 450 nm using the Epoch Plate Reader (Agilent BioTek).

Colony Formation Assay

OSC19 were plated (in triplicate) in a 12-well plate at the confluency of 1000–2000 cells/well. After 24 hours, cells were either transfected with respective siRNA or treated with rotenone, and colonies were allowed to form until the control group reached near maximum confluency (at 7–10 days). Following treatment, cells were fixed in formalin (1:10 dilution) for 20 minutes and stained with 0.5% crystal violet for 20 minutes. Excess crystal violet was removed with tap water, and the plate was allowed to dry. The plate was then scanned and analyzed using the macro ‘colony area measurer’ of Image J (ImageJ, RRID:SCR_003070).

Western Blot

Cells grown in 10 cm dishes were collected after indicated treatments. Cell pellets were first lysed using RIPA buffer followed by sonification. For western blotting in HPV-tumors, ~30mg tumor samples were minced on dry ice in cold RIPA buffer and homogenized using a pellet pestle motor followed by sonication. Samples were kept on ice for 15 minutes followed by centrifuging at maximum speed for 20 minutes. Protein quantification was performed from lysate (supernatant) using the Bradford assay, and 80 μg protein was loaded into gel. Actin normalization was done for each gel. Immunoreactive bands were visualized and quantified using the LiCor Odyssey system.

Quantitative Real time PCR

Tissue samples from patients with SCCHN were collected after obtaining informed consent. All patients were subjected to biopsy before the initiation of any treatment. The clinical characteristics of all patient samples used for the study is listed in Supplementary Table ST4. The tumor samples were transferred in RPMI media supplemented with antibiotics and stored in RNAse later until RNA isolation using the RNeasy Mini Kit (QIAGEN). Briefly, ~25 mg tissue was disrupted in cold RLT buffer and homogenized through QIA shredder spin column. The lysate was again centrifuged before washing and eluting in water. RNA was quantified using the Gen5 Take 3 Module (Agilent BioTek) and assessed for quality with 260/230 absorbance ratio. For cell lines, cells were scraped into a pellet and RNA was isolated following Qiagen protocol. iScript Reverse Transcription (Bio-Rad) was used for cDNA synthesis. Quantitative PCR (qPCR) was conducted using SYBR green master mix using ACTIN and GAPDH as housekeeping controls. Relative gene expression was calculated using the 2^−ΔΔCq method. The primer sequences are listed in Supplementary Table ST6. PCR conditions used are 40 cycles at 15 sec for denaturation at 95°C, 30 sec annealing at 60°C, 30 sec extension at 72°C.

Mitotracker Green

Cells are plated in triplicate in 6-well plate and allowed to reach 70% confluency. After trypsinizing and washing once with PBS, ~30,000 cells are suspended in 100nM mitotracker green (Invitrogen #M7515) diluted in cold Krebs buffer. A tube containing FCCP (5 μM) + mitotracker green is used as technical control for the experiment. Cells are allowed to stain on ice for 30 minutes. Tubes are centrifuged, washed once and resuspended in 250μl cold Krebs buffer and analyzed by flow cytometery on Beckman Coulter Cytoflex analyzer on 488nm laser. Non-viable cells were excluded from analysis using PI stain. Median fluorescence intensity (MFI) values were taken for unstained, mitotracker green stained and FCCP + mitotracker green stained groups.

Complex I Assay

Cell pellets were frozen on dry ice after collection and immediately submitted to University of Pittsburgh Center for Metabolism and Mitochondrial Medicine (C3M) for measurement of complex activity using spectrophotometric kinetic assay system. Briefly, complex I activity was measured by spectrophotometrically (340 nm) monitoring the oxidation of 100 μM NADH in the presence of 10 μM coenzyme Q₁ in the presence and absence of 25 μM rotenone as previously described (41).

ATP Production

At least 10,000 cells/well were plated in 96-well plate, allowed to attach overnight, and treated with indicated treatments. ATPlite Luminescence Assay System (Perkin Elmer Inc.) kit was used to measure ATP production. A standard curve of ATP was run and normalized using the crystal violet. Luminescence was read using the Synergy H1 Hybrid microplate reader (BioTek).

Tetramethylrhodamine, ethyl ester (TMRE)

OSC19-VC/OE cells were plated at a confluency of 15000–20000 cells/well in a 96-well black plate. Following plating of cells, the quantification of changes in mitochondrial membrane potential was determined following the protocol of the TMRE-Mitochondrial Membrane Potential Assay Kit (Abcam). Briefly, live cells were first either stained with carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) or left unstained, then incubated with TMRE, and finally washed with 1X Phosphate Buffered Saline (PBS) and 0.2% Bovine Serum Albumin (BSA). The 96-black well plate was analyzed using the Epoch Microplate Spectrophotometer (BioTek) at excitation and emission values of 549 nm and 575 nm, respectively.

In vivo Xenograft study

All the in vivo studies follow the ARRIVE guidelines (Animal Research: Reporting of In Vivo Experiments) available at www.ARRIVEguidelines.org. The primary outcome of animal studies is to study the differential effect of complex I inhibitor IACS-10759in high TMEM16A expressing tumors as compared to Vector Control. For MOC1 xenograft experiments, 5-week-old female C57BL/6J mice (The Jackson Laboratory Cat#000664, RRID:IMSR JAX:000664) were injected in bilateral flanks subcutaneously with 3 million MOC1-VC and MOC1-TMEM16A cells (in 50% Matrigel). After tumors became palpable, mice were randomized into four treatment groups: MOC1-VC Control, MOC1-VC IACS-10759, MOC1-TMEM16A Control, MOC1-TMEM16A IACS-10759. Each group had 9–10 mice. IACS-10759 dissolved in 0.5% sodium carboxymethyl cellulose (Na-CMC), was administered to mice once a day via oral gavage for eight consecutive days (8 total treatments) at a dose of 10 mg/kg. Animals were handled and euthanized in accordance with IACUC guidelines. The MOC1 xenograft experiment was done twice, and combined data from both experiments is shown. Tumors were harvested after 8 treatments of IACS-10759 was done (n=17–19 tumors). We selected a moderate sample size considering the number of groups in the study. All mice included in the study were healthy with tumor burden within the specified limits of IACUC protocol. There was no data exclusion. Two investigators were involved in the study: one administered the treatment, as labelled in the record notebook. The second investigator only measured the tumors in the mice, which were recorded by the first investigator. At the time of recording, outcome measures assessed were volume of the tumors using the following formula: (l*w*w)/2 (mm³) and weight of the mice (g).

In vivo PDX study

After surgical resection, the tumor sample was transported in RPMI media supplemented with Pen/Strep and anti-anti. The tumor was implanted into the flank of 6-week-old female NOD SCID (The Jackson Laboratory Cat#001303 RRID:IMSR_JAX:001303) mice and passaged once before being used for the study. The mice were randomized when average tumor volume reached an average of 100–150mm³. Mice were randomized into 2 groups of untreated and IACS treatment. The differences in relative tumor volume for each day was calculated using two-way ANOVA with Sidak’s multiple comparison test.

Electron Microscopy of Xenograft tumors

Female nude mice were bilaterally injected with 3 million OSC19-VC and OSC19-TMEM16A cells until tumors reached 75–100 mm3. After sacrificing the mice, tumors were surgically removed and fixed in Karnovsky’s fixing buffer. The tumors were then sent to the Center for Biological Imaging at the University of Pittsburgh for further processing. Images were captured at 40000X. Image J software was used to analyze the mitochondria perimeter and area of the images.

Statistics

All experiments are done at least 3 times in triplicate, and combined data of N=3 is shown throughout the manuscript figures, unless otherwise stated. All analyses were performed blinded. All data is represented as Mean ± S.E.M. Significance is defined as follows: *p<0.05, **p<0.001 and ***p<0.0001. The test used is listed in each respective figure legend. Unless otherwise noted in each figure, a p value of less than 0.05 is used to denote significance. Students t test is used to compare differences between 2 groups. Where indicated, significance is calculated after comparison between groups as indicated by the lines connected to bars using one-way ANOVA with Tukey’s multiple comparison.

Study Approval

HPV-tumors were obtained from the University of Pittsburgh Medical Center in accordance with established University of Pittsburgh IRB guidelines. A written informed consent was obtained from all the patients before inclusion in the study. All data generated from patient samples collected was used in the study, i.e., there was no attrition. The clinical characteristics of all patient samples used for the study is listed in Supplementary Table ST4 and ST5. All animal studies were done in accordance with approved IACUC protocol.

Data Availability Statement

The data generated in this study are available within the article and its supplementary data files

Results

TMEM16A overexpressing SCCHN exhibit coordinated regulation of lysosomal biogenesis and mitochondrial complex I.

The current study focuses on HPV-negative SCCHN models, that bear largely different mutational signature from HPV-positive SCCHN tumors and display unfavorable outcomes including poor survival. Unless otherwise stated, human oral squamous cell carcinoma derived from a tongue tumor: OSC19 cell line is used as in vitro model across the data presented in this study. TMEM16A upregulation is a hallmark and a driver of proliferation in many malignancies, including SCCHN (20). We previously showed an upregulation of lysosomal flux comprising both lysosomal biogenesis and exocytosis in human cancers, particularly in the context of TMEM16A upregulation(17). Given the emerging evidence of functional interplay between lysosomes and mitochondria, we sought to determine whether TMEM16A drives a synergistic relationship between lysosomes and mitochondria in malignant cells.

We use a combination of evolutionary rate covariation (ERC) and gene expression analysis by employing a publicly available cancer cell line database of gene expression (21) to provide an unbiased prediction of the functional content of lysosome-mitochondria interaction in TMEM16A-upregulating SCCHN models. ERC is a computational approach that presumes shared evolutionary pressure and therefore arranges a collective evolutionary history of proteins that are involved in functional interactions (22). Such histories can be tracked and correlated by calculating the divergence of amino acid or gene sequence pairs in several organisms in a clade or a set of clades. Proteins whose evolutionary histories correlate are candidates for functional interaction, and these correlations do not rely on prior empiric data.

We used a publicly available ERC portal https://csb.pitt.edu/erc_analysis/ to study the coevolution between known mitochondrial and lysosomal proteins in mammals. The analysis was based on the Mitocarta 3 database (1136 proteins) (23) and a published lysosomal proteome (417 proteins) (24). The mitochondrial proteins were divided into two groups based on their location in the organelle, outer membrane and intermembrane. Similarly, we segregated the lysosomal proteins into those containing and not containing transmembrane domains (TMD+ and TMD− respectively). The coevolution between proteins was calculated as the correlation between their percent identity values in 49 species of mammals (Supplementary Table ST1 and ST2). Supplementary Table ST3 lists the lysosomal and mitochondrial protein types and specific coevolving proteins using coevolution with 10% of the corresponding organelle’s proteome as a threshold.

A detailed analysis of these protein classes identifies several potential functional complexes shared by the outer mitochondrial membrane and lysosomal proteins, including proteins involved in ion exchange, mitochondrial positioning, and fusion/fission. This coevolution indicates a close functional (but not necessarily physical) engagement between lysosomes and mitochondria. Next, we sought to understand if this interaction is enriched in tumors that are putatively more aggressive. We used TMEM16A expression as a surrogate of tumor aggressiveness (25). Therefore, we parsed Cancer Dependency Map database, Expression 22Q2 dataset (26), focusing on cancer cell lines of upper aerodigestive origins (58 lines). This approach yielded a set of genes that co-evolve and whose expression is co-regulated in the context of TMEM16A expression (Fig. 1A). Interestingly, these analyses revealed a specific co-regulatory effect on genes coding for the mitochondrial complex I components.

Figure 1. TMEM16A overexpressing SCCHN exhibit coordinated regulation of lysosomal biogenesis and mitochondrial complex I.

ERC analysis reveals several genes coding for lysosomal proteins and mitochondrial complex I proteins that co-evolve and are coregulated in squamous cell carcinoma of the head and neck cell lines (A). The expression of several mitochondrial and lysosomal genes and TMEM16A expression is co-regulated in a larger cohort of human SCCHN tumors (N=24). Correlation coefficients (p < 0.05) of regression analyses of various pairs of genes are shown in (B). Immunoblots of human tumor tissue show that these proteins are co-expressed in tumors that display overexpression of TMEM16A (C). Immunofluorescence analyses in OSC19 cells demonstrates co-localization of lysosomes with mitochondria in the context of TMEM16A overexpression (VC=Vector Control). The significance shown is calculated using an unpaired t test (D).

We next sought to answer whether the expression of these genes correlate in human SCCHN tumor tissues. We used qPCR to analyze the expression of genes coding for the lysosomal (SLC11A2, VAMP8 and TPCN2) and mitochondrial (NDUFS2 and NDUFB8) proteins implicated in coevolution and coregulation in the context of TMEM16A-upregulating SCCHN models described above in a cohort of human SCCHN tumors, followed by a correlation analysis. The clinical characteristics of patient samples used in the cohort are listed in SupplementaryTable ST4. The correlation coefficients resulting from these analyses are shown in Fig. 1B and specific plots are in Supplementary Fig. S1A, B. Interestingly, western blotting (Fig. 1C) shows a) upregulation and b) substantial correlation between the protein expression of genes coding for lysosomal (Lamp1, SLC11A2 and Vamp8) and mitochondrial complex I protein (NDUFB8), in the context of TMEM16A overexpression in human tumor samples, as listed in Supplementary Table ST5.

In line with the data from human tumors, we observed a particularly robust induction of complex I genes upon TMEM16A expression in OSC19 cells (Supplementary Fig. S1C, D; VC is vector control). Interestingly, there was a lack of correlation between TMEM16A, and genes related to complex III (UQCRH and UQCRC) and IV (COX4i1 and COX1); suggesting that the link between lysosomal biogenesis and mitochondrial function is specific to complex I (Supplementary Fig. S1E–H).

We further expanded our investigations to determine whether the observed co-expression translated to a potential physical interaction between lysosomes and mitochondria. We used mitochondrial marker Tomm20 (Translocase of Outer Mitochondrial Membrane 20), and Lamp1 (lysosomal-associated membrane protein 1), to flag the lysosomes in SCCHN cell line OSC19, designed to overexpress TMEM16A or vector control (VC). Expectedly, lysosomal biogenesis (triggered by TMEM16A overexpression) leads to increased co-localization between these organelles (Fig. 1D).

We have previously shown that TMEM16A-induced lysosomal biogenesis causes a collateral increase in lysosomal exocytosis (17). Therefore, we sought to determine if other lysosomal functions like autophagy are also increased. Using electron microscopy, we identified increased autophagic vacuoles (defined as double membraned intracellular vacuoles) in TMEM16A overexpressing xenograft tumors (Supplementary Fig. S1I).

Mitophagy is a process by which effete mitochondria are degraded. Given the observed co-localization between lysosomes and mitochondria, we determined the effect of mitophagy in our cells. PINK1 (PTEN-induced kinase 1) mediated mitophagy is a well-identified mechanism to eliminate dysfunctional mitochondria via autophagy in mammalian cells (27). Upon treatment with the electron transport chain uncoupler FCCP (positive control), we noted an induction of PINK1 expression (indicative of mitophagy), while the endogenous PINK1 expression was detectable. This suggests that TMEM16A does not directly impact mitophagy (Supplementary Fig. S1J), suggesting that the observed co-localization is not due to increased mitophagy. Taken together, these data demonstrate that lysosomal biogenesis is coordinately regulated with the expression of mitochondrial complex I proteins, independently of mitophagy.

TMEM16A promotes mitochondrial mass and biogenesis

The data described above indicate coordinated biogenesis of mitochondria and lysosomes in a TMEM16A-high patient-derived SCCHN tumor samples. Our prior studies show that TMEM16A promotes lysosomal biogenesis in SCCHN models (17). Here, we sought to determine if TMEM16A promotes mitochondrial biogenesis. Mitotracker staining in cells engineered to overexpress TMEM16A evidenced increased mitochondrial mass (Fig. 2A), indicating increased mitochondrial content in these cells.

Figure 2: TMEM16A promotes upregulation of mitochondrial mass and biogenesis.

OSC19 cells were engineered to stably overexpress TMEM16A. These cells were subjected to flow cytometry with Mitotracker dye to measure mitochondrial mass. TMEM16A overexpression increased Mitotracker staining. VC is vector control. (A). Gene expression analyses revealed a profound overexpression of genes associated with mitochondrial biogenesis in these cells (B). Consistent with prior data, the activity of mitochondrial complex I was increased upon TMEM16A overexpression (C). Cells were treated with diluent or rotenone and assessed for cell viability using colony formation assays (D). Similarly, VC or TMEM16A overexpressing cells were treated with diluent or rotenone for 2–4h and assessed for cell viability using WST assays (E). Tumor xenografts were generated by inoculating nude mice with the respective cells. Tumors were harvested and subjected to EM analyses to measure mitochondrial perimeter and area (F) For all panels, significance is calculated using an unpaired t test.

We find that several transcription factors that have a role in mitochondrial biogenesis, redox balance, and defense against antioxidant damage, including PGC1α and NRF2 (28) are amplified in TMEM16A-overexpressing OSC19 (Fig. 2B). This is in concert with the prior evidence of higher baseline levels of reactive oxygen species (ROS) in TMEM16A-overexpressing SCCHN models (17, 29). TMEM16A-promoted mitochondrial biogenesis is associated with increased expression of genes linked with mitochondrial antioxidants (SOD1, SOD2 and GPX) and regulators of mitochondrial respiration (cytochrome oxidase and citric synthase) (Fig. 2B). Proteomic and biochemical analyses reveal that the expression of complex I proteins is increased upon TMEM16A expression, which is correlated with an increase in complex I activity, providing physiological evidence for the effects of TMEM16A expression on mitochondrial function (Fig. 2C).

We extended these findings to determine if genetic knock-down of TMEM16A in the SCCHN cell line with endogenously high expression levels of the HN30 gene led to a notable decrease in mitochondrial biogenesis, mass and complex I activity (Supplementary Fig. S2A–D). The upregulation of mitochondrial biogenesis in TMEM16A-overexpressing OSC19 cells is reflected in the pronounced dependence of these cells’ proliferation on mitochondrial function.

Suppressing the mitochondrial function using rotenone (Fig. 2D, E) consequentially decreased the proliferation of specifically TMEM16A-overexpressing OSC19 cell line. Similarly, using OSC1-VC and -TMEM16A xenograft model in nude mice, we find that TMEM16A overexpression leads to increased mitochondrial mass, as measured by mitochondrial area and perimeter on electron microscopic images (Fig. 2F).

NRF2 activation in TMEM16A is pivotal in regulating mitochondrial function

NRF2 is regulated by its proteolytic degradation under normal conditions, which is suppressed under oxidative stress leading to NRF2 buildup and transcriptional activation of NRF2 dependent genes (30). Western blot in Fig 3A shows that TMEM16A-overpexressing cells display increased NRF2 levels. Using small interfering RNA (siRNA) targeting NRF2 (validated in Fig. 3A), we show that the changes in ATP production and mitochondrial mass in response to NRF2 suppression are more pronounced in the TMEM16A overexpressing cells (Fig. 3B, C). From a mechanistic perspective, we noted that treatment with siNRF2 remarkably reduced the expression of transcription factors that drive mitochondrial complexes and antioxidants (Fig. 3D, E). These data indicate that TMEM16A drives mitochondrial function via NRF2 (Fig. 3F).

Figure 3: NRF2 activation in TMEM16A is pivotal in regulating mitochondrial function.

Western blot shows short-interfering RNA (siRNA) targeted at NRF2 knocks down NRF2 protein (A). The effect of siNRF2 on ATP production was measured using ATPlite assay (B) mitochondrial mass (C) mitochondrial complex (D) and antioxidants transcript (E). For (D) and (E), bars are compared to fold change of VC + siControl. The putative model for TMEM16A induced mitochondrial function (F). For (B) and (C), statistics is done using one-way ANOVA with Tukey’s multiple comparisons test. For (D) and (E), unpaired t test is used.

Interaction between lysosomes and mitochondrial complex I regulate cancer cell viability

Based on the data described above, we postulated that a coordinated signaling between lysosomes and mitochondria regulates cellular biogenetics (model in Fig. 4A). We first tested the hypothesis that inhibition of mitochondrial complex I disrupts lysosomal biogenesis. Treatment of SCCHN with the complex I inhibitor IACS-10759 led to a substantial decrease in lysosomal biogenesis, as assessed by the expression of integral lysosomal genes (Fig. 4B). Interestingly, treatment with inhibitors of complex IV (ADDA5, a cytochrome oxidase inhibitor), complex III (Atovoquone) or oxidative phosphorylation uncoupler FCCP did not significantly impact lysosomal biogenesis (Supplementary Fig. S3). These data, along with our bioinformatic analysis strongly suggest the interaction between lysosomes and mitochondria is driven by complex I. A corollary of this observation as determined from ERC and co-expression analysis, is that inhibition of the integral lysosomal genes involved in the lysosome-mitochondria axis in TMEM16A overexpressing SCCHN should impact mitochondrial oxidative phosphorylation. Transient knockdown of lysosomal genes TPCN2, SLC11A2 and VAMP8 reduced ATP production (Fig. 4C, D) and impacted the cell growth in colony forming assay specifically in TMEM16A overexpressing cells (Fig. 4E). Vamp8 knockdown was confirmed by western blot (Supplementary Fig. S4A), however, on account of lack of suitable antibody for TPCN2 and endogenous expression for SLC11A2 in OSC19, knockdown has been confirmed via qPCR (Supplementary Fig. S4B, C).

Figure 4: Interaction between lysosomes and mitochondrial complex I regulate cancer cell viability.

A proposed model for the interaction is shown in (A). Treatment with the complex I inhibitor, IACS-10759 reduces lysosomal biogenesis (B). Knock-down of lysosomal and ERC genes leads to a reduction in ATP production (C, D) and cell proliferation, as measured by colony formation (E). For (B), statistics is calculated using unpaired t test, for (C-E), one-way ANOVA with Tukey’s multiple comparisons test is used.

TMEM16A-driven tumor xenograft models are sensitive to complex I inhibition

Further evidence of mitochondrial function repression in the context of TMEM16A was demonstrated by complex I inhibitor IACS-10759 as TMEM16A overexpressing OSC19 line is significantly more sensitive to this inhibitor than the vector control (Fig. 5A). Supplementary Fig. S2E and F show Suppp-10759 effects in endogeneously high TMEM16A expressing HN30 cell line, which is in accordance with the OSC19 data. Fig. 5B, C show that IACS-10759 inhibited the enhanced ATP production and mitochondrial membrane potential in the TMEM16A-overexpressing OSC19 cell line. This drug is in current clinical trials for several cancer types (31),(32) although its effects in the SCCHN context have not been established. The next set of experiments seeks to further investigate the complex I inhibitor IACS-10759 in SCCHN models.

Figure 5: Complex I inhibition sensitizes TMEM16A-driven tumor xenograft models.

Control and TMEM16A overexpressing OSC19 cells were evaluated for sensitivity to the complex I inhibitor, IACS-10759 (A). The effect of IACS-10759 on ATP production and TMRE fluorescence is shown in (B, C). The effect of IACS-10759 on subcutaneous tumors generated from the syngeneic murine SCCHN cell line (MOC1) is shown in (D). A patient-derived xenograft (PDX) was treated with IACS-10759 to demonstrate the anti-tumor activity (E). Representative tumor pictures and tumor weights after necropsy are shown in (F, G). For (B) and (C), one-way ANOVA with Tukey’s multiple comparison test is used, for (D) and (E), two-way ANOVA with Tukey’s multiple comparison test is used, unpaired t test is used for (G).

To generate a mouse in vivo SCCHN model, we subcutaneously injected C57 BL/6J mice with control (VC; Vector Control) and TMEM16A-overexpressing syngeneic MOC1 (squamous cell carcinoma of the mouse oral cavity) cell lines (Supplementary Fig. S5A) and documented the development of tumors. Mice were treated with 10mg/kg body weight IACS-10759 for 8 consecutive days. The tumor volume data in Fig. 5D and Supplementary Fig. S5B show enhanced suppression of tumor growth by IACS-10759, especially in the TMEM16A overexpressing MOC1 tumors, with the effect of IACS-10759 observed as early as 4^th treatment. Reduced tumor burden is evidenced by decreased Ki67 staining in the TMEM16A-high IACS-10759 treated tumors (Supplementary Fig. S5C, D). The mice maintained non-significant body weight change throughout the duration of IACS-10759 treatment. (Supplementary Fig. S5E). The authors acknowledge that TMEM16A overexpressing cells and xenografts have historically grown faster compared to low TMEM16A expressing models (33). In vitro, MOC1 cells are aggressive at baseline, and we observed that there was no additional amplification on cell proliferation after TMEM16A overexpression in MOC1 cells. We see a similar trend in vivo, and further experiments would need to be done to characterize the mechanism. Additional evidence indicating TMEM16A overexpression is highly sensitive to IACS-10759 is given in Supplementary Fig. S6.

We also corroborated the anti-tumor effects of IACS-10759 using high-TMEM16A expressing patient-derived xenograft (PDX) implanted into immunocompetent NOD SCID mice (Fig. 5E–G). Insert shows confirmation of TMEM16A expression in the PDX used for the study. Clinical characteristics of the patient (to form PDX) have been tabulated in Supplementary Fig. S5F. Taken together, these data show the unique dependence of TMEM16A-overexpressing SCCHN models on mitochondrial function.

Since TMEM16A drives cancer cell proliferation (25), we wanted to test if these changes could be replicated in conditions of increased cellular proliferation that are independent of TMEM16A (34, 35). To test this, we used a transiently overexpressed mutant version of PI3KCA (PI3KCA-H1047R), an oncogenic mutation that is frequently encountered in SCCHN. Interestingly, overexpression of PI3KCA-H1047R did not consistently induce mitochondrial biogenesis, although we noted a remarkable (p<0.0001) increase in the expression of the transcription factor TFAM (Supplementary Fig S7A, B). Similarly, we did not observe any increase in lysosomal biogenesis (Supplementary Fig S7C). We confirmed these data by overexpressing EGFR, a known SCCHN oncogene. EGFR overexpression did not promote mitochondrial or lysosomal biogenesis (Supplementary Fig S7D–F). Taken together, these data suggest that LMI is specifically driven by TMEM16A overexpression.

Discussion

Many of the mechanism(s) that regulate tumor cell growth and proliferation remain elusive (36). In addition to canonical signaling pathways that regulate cellular proliferation, intracellular organelles also impact cell growth and proliferation (37). Indeed, several important signaling molecules like mTORC1, AMPK, GSK3 localize to the lysosomal membrane (38). We and others have studied the role of lysosomes in regulating cell proliferation since downstream signaling from the lysosomal membrane promotes tumor cell proliferation (1, 17, 39, 40).

Lysosomal biogenesis is regulated by the CLEAR network of genes. However, the upstream signaling nodes that regulate the CLEAR network may also have a redundant role in regulating cell proliferation. An important question in the field is how do lysosomes synergize with other organelles in the context of tumor proliferation? Our data provide evidence that lysosomal-mitochondrial synergy mediates cell growth and proliferation in SCCHN. This synergy is facilitated by proto-oncogene TMEM16A. Upon overexpression of TMEM16A, lysosomal and mitochondrial biogenesis are upregulated in a coordinated fashion leading to improved cellular fitness. Specifically, genes that encode components of mitochondrial complex I are enriched in our co-expression analyses. This suggests that there may be a unique survival advantage upon overexpression of mitochondrial complex I proteins. Interestingly, TMEM16A expression seems to promote dependence on mitochondrial respiration in our model. However, lysosomal biogenesis appears to serve a more general role. Mechanistically, TMEM16A overexpression drives NRF2, which in turn regulates mitochondrial function. The effect of NRF2 activation on mitochondrial function has not been previously reported in SCCHN.

This interaction between lysosomes and mitochondria promotes tumor cell proliferation. Interestingly, knock-down of specific lysosomal genes reduces ATP production and cell viability in a complex I dependent manner. Furthermore, treatment with a complex I inhibitor did not induce an additional reduction of ATP, suggesting that the LMI is specific to mitochondrial complex I. In concert with this postulate, complex III nor complex V inhibition did not impact lysosomal biogenesis. To our knowledge, this is the first mechanistic description of how lysosomes affect mitochondrial respiration.

The expression of constitutively active PI3KCA mutant (PI3KCA-H1047R) or overexpression of EGFR did not promote lysosomal or mitochondrial biogenesis, providing further evidence that the TMEM16A specifically promotes LMI. However, we cannot completely exclude the possibility that other signaling pathways that promote tumor cell proliferation may also enhance LMI, and more work is required to fully elucidate the molecular networks that promote/regulate LMI.

Our data highlight the importance of lysosomal-mitochondrial interaction and how it promotes tumor growth. This novel idea calls attention to how intra-cellular communication between organelles is promoted through intracellular signaling pathways. From a translational perspective, our data implicate complex I as a potential molecular target in carcinomas that overexpress TMEM16A. The anti-tumor activity of IACS-10759 in our pre-clinical models provides a strong rationale upon which clinical trials can be initiated to improve treatment for patients with malignancies that are driven by TMEM16A. This could create a new paradigm to treat these patients.

Implications statement.

Intervention of lysosome-mitochondria interaction may serve as a therapeutic approach for patients with high TMEM16A expressing SCCHN.