TLR4 and SARM1 modulate survival and chemoresistance in an HPV-positive cervical cancer cell line

Human Papillomavirus is responsible for a wide range of mucosal lesions and tumors. The immune system participate in tumorigenesis in different ways. For example, signaling pathways triggered by Toll-like receptors (TLR) play a role in chemotherapy resistance in several tumor types and are candidates for contributing to the development of HPV-induced tumors. Here, we studied the receptor TLR4 and the adaptor molecule SARM1 in HeLa cells, an HPV-positive cervical cancer cell line. Knocking out of these genes individually proved to be important for maintaining cell viability and proliferation. TLR4 knock out cells were more sensitive to cisplatin treatment, which was illustrated by an increased frequency of apoptotic cells. Furthermore, TLR4 and SARM1 modulated ROS production, which was induced by cell death in response to cisplatin. In conclusion, TLR4 and SARM1 are important for therapy resistance and cervical cancer cell viability and may be relevant clinical targets.

www.nature.com/scientificreports/ We have previously reported that cervical cancer cell lines expressing E6 and E7 from high-risk HPV types have several alterations on the TLR pathway, such as increased expression of HMGB1, SARM1 and TLR4 15 . Here, we sought to further dissect the function of the TLR pathway in a cervical cancer cell line focusing on the roles of TLR4 and SARM1 on cell survival and cisplatin resistance.

Knock out of either TLR4 or SARM1 reduces viability and proliferation of HeLa cells.
To study the role of TLR4 and SARM1 in the malignant phenotype of cervical cancer cells, we knocked out these genes individually in HeLa cells using CRISPR/Cas9 and isolated individual clones of the edited cells. Lack of protein expression was confirmed by western blot (Fig. 1a). One clone of each of the TLR4 and SARM1 knockout cell lines was then randomly chosen and selected for further analysis. From here on, they will be referred to as TLR4KO (clone 5) and SARM1KO (clone 3).
First, we compared cell viability of TLR4KO and SARM1KO cells to the unedited parental HeLa cells with or without exposure to Lipopolysaccharide (LPS) treatment. Depletion of TLR4 or SARM1 induced approximately 25% reduction of cell viability when compared to the parental HeLa cells (Fig. 1b). Specifically in the case of TLR4KO, loss of viability was further increased by 15% by the LPS treatment (Fig. 1b). In addition, both TLR4KO and SARM1KO cells proliferated less than the parental HeLa cells, with SARM1KO cells showing a reduction of 80% and TLR4KO cells of 45% in proliferation (Fig. 1c).

Lack of TLR4 and SARM1 reduces the clonogenic and migration potential of HeLa cells.
We also analyzed the effect of knocking out TLR4 and SARM1 on the clonogenic and migratory potential of HeLa cells. We observed that depletion of SARM1 and TLR4 reduced the ability of HeLa cells to form colonies when seeded at low density (Fig. 2a). Furthermore, SARM1KO colonies were significantly smaller, consistent with its pronounced reduction of proliferation (Fig. 1c).
To evaluate the migration potential of the cells, we performed a wound healing test. Cells were previously treated with mitomycin-C to rule out that any observed effect was due to proliferation differences between the parental cells and TLR4KO and SARM1KO cells ( Supplementary Fig. S2). Interestingly, SARM1KO cells showed a significant reduction in migration after 48 h (Fig. 2b, c). On the other hand, TLR4KO cells displayed a significant increase in migration compared to parental cells, noticeable after 48 h of the generation of the wound (Fig. 2b, c). Our results confirm that lack of TLR4 and SARM1 interfere with important tumor-related characteristics such as clonogenic and migration potentials.

TLR4 and SARM1 knockouts sensitize HeLa cells to chemotherapy. To determine if TLR4
and SARM1 impact the response of cervical cancer cells to chemotherapy, we treated parental, TLR4KO and SARM1KO cells with different concentrations of cisplatin. Figure 3a shows that TLR4KO and SARM1KO cells had higher sensitivity to cisplatin than parental HeLa cells. This enhanced effect is also illustrated by the calculation of the IC50 (concentration at which we observe a 50% reduction in cellular viability): around 4.8 µM for TLR4KO, 3.9 µM for SARM1KO, and 17 µM for parental cells.
TLR4KO induces apoptosis after cisplatin treatment. We then investigated whether knocking out either TLR4 or SARM1 induces cell death. In the case of TLR4KO cells, we observed that cisplatin treatment significantly increased apoptosis when compared to parental HeLa cells, and this effect was enhanced when cisplatin was combined with LPS (Figs. 3b and 4a). Interestingly, more SARM1KO cells went through apoptosis only when treated with both cisplatin and LPS, as compared to parental cells treated with cisplatin alone. We observed no differences in the induction of necrosis between cells (KO or parental), or between LPS or cisplatin treatments conditions. Nevertheless, parental and KO cells responded to cisplatin with and without LPS, showing a significant increase in necrosis rate, despite individual differences in susceptibility (Fig. 4b).
To confirm that apoptosis was being triggered, we measured the amount of activated caspase 3 by flow cytometry. Although cisplatin induced caspase 3 activation in both parental and KO cells, the effect was more pronounced in TLR4KO cells compared to unedited HeLa cells ( Fig. 4c and Supplementary Fig. S3). Even though SARM1KO cells also showed higher levels of activated caspase 3 after cisplatin treatment, this increase was not significantly higher than the levels observed in parental cells ( Fig. 4c and Supplementary Fig. S3). In addition, when using cisplatin combined with LPS, both TLR4KO and SARM1KO cells showed an increase in active caspase 3 levels compared to parental cells ( Fig. 4c and Supplementary Fig. S3). LPS alone increased apoptosis levels in TLR4KO cells compared to parental cells (Fig. 4c). Taken together, our data shows that TLR4 depletion sensitizes HeLa cells to chemotherapy-induced apoptosis independent of LPS exposure whereas deleting SARM1 sensitizes cells to chemotherapy only in the presence of LPS.
TLR4 and SARM1 modulate ROS production. Cisplatin and the TLR pathway modulate production of reactive oxygen species (ROS), which induces cell damage and stimulates apoptosis. Therefore, we measured ROS production in parental, TLR4KO and SARM1KO HeLa cells exposed to the same treatments performed before (Fig. 5a). We noticed that, compared to parental HeLa cells, basal ROS levels were lower in SARM1KO and higher in TLR4KO. Cisplatin increased ROS levels in edited and unedited cells, whereas LPS lowered ROS levels only in unedited HeLa cells and had no significant effect on ROS levels in TLR4KO and SARM1KO cells. When cells were treated with cisplatin and LPS, both TLR4KO and SARM1KO cells showed an increase in ROS levels similar to that obtained when cells were treated with cisplatin alone. Both TLR4KO and SARM1KO cells modulated ROS levels although LPS decreased ROS production only in parental HeLa cells. www.nature.com/scientificreports/ Since cisplatin can induce apoptosis through ROS production, we exposed cells to a ROS inhibitor, N acetyl L cysteine (NAC) to determine if the chemotherapy sensitivity of knockouts were related to ROS levels (Supplementary Fig. S4). We observed that NAC did not alter cisplatin-induced apoptosis of SARM1KO or parental cells. On the other hand, apoptosis levels were significantly lower in NAC-treated TLR4KO cells exposed to cisplatin (Fig. 5b). No effect in necrosis was observed (Fig. 5c). Therefore, our results suggest that the apoptosis observed in TLR4KO cells in response to cisplatin is related to an increase of cellular ROS levels. www.nature.com/scientificreports/ www.nature.com/scientificreports/ www.nature.com/scientificreports/ Expression of TLR4 and SARM1 alter expression of genes involved in tumorigenic and inflammatory processes. To better understand the effect of knocking out TLR4 and SARM1, we evaluated the expression profile of genes potentially modulated by TLR and related to several steps in the tumorigenic and inflammatory processes. We checked the expression levels of IL6 and IL18 (which are related to TLR triggering and inflammation), Superoxide Dismutase 2 (SOD2, responsible for ROS and oxidative stress modulation), Cyclins D1 and E1 (CCND1 and CCNE1, respectively, which are involved in cell cycle progression), and Ki-67 (a marker of cell proliferation) in HeLa cells depleted of TLR4 or SARM1, in comparison to the parental HeLa cells (Fig. 6).
In both TLR4KO and SARM1KO cells, we observed lower levels of IL18 compared to parental cells. On the other hand, while TLR4KO cells showed decreased levels of IL6, we observed increased levels of IL6 in SARM1KO cells, as compared to the unedited HeLa cells. SOD2 expression was higher in SARM1KO cells, but not TLR4KO cells, compared to parental cells. CCND1 expression levels increased in both SARM1KO and TLR4KO cells, while CCNE1 levels decreased significantly only in TLR4KO cells. In addition, KI-67 expression levels were lower in TLR4KO and SARM1KO cells compared to parental HeLa cells (Fig. 6).
In conclusion, lack of TLR4 and SARM1 in HeLa cells changed the expression of genes related to inflammatory response, cell cycle and proliferation.

Discussion
TLR4 has been implicated in both triggering innate immune response against bacteria and colon cancer development 26 . In addition, in ovarian tumors, its activation leads to increased cell proliferation and cytokines and chemokines production 27 . Nevertheless, TLR4 has different roles depending on tumor type. In hepatocellular carcinoma, for example, its absence leads to more DNA damage, inhibition of senescence and tumor progression 28 . In pancreatic tumors, LPS stimulates tumor development and increases invasiveness in vivo through activation of NF-κB via TLR4 signaling 29 . LPS treatment has also been implicated in the establishment of cervical tumors in a mouse model, while TLR4 silencing led to the formation of smaller tumors 30 .
Several studies also showed a role for TLR4 in tumorigenesis and therapy resistance. There is no consensus of a single set of genes involved in these processes, but some studies suggest an association between TLR4 and epithelial-to-mesenchymal transition through STAT3 activation 31 as well as a role in proliferation through the activation of pathways such as MAPK 32 . Studies of chemotherapy resistance related to TLR4 activation also revealed changes in the expression of genes related to apoptosis and cell cycle 33 .
Few studies have shown an association between increased TLR4 expression in HPV-positive cervical tumor samples and resistance to apoptosis 15,34,35 . We have previously reported that TLR4 had higher expression levels in HPV-positive tumor cells and in normal cells expressing HPV16 E6 and E7 oncoproteins 15 . After further investigation, our results confirmed the role of TLR4 in cell survival, proliferation and response to cisplatin in HeLa cells. Lack of TLR4 in this cell line led to increased ROS production and consequently increased apoptosis after treatment.
Another poorly understood protein related to TLR4 is SARM1, a negative regulator of TLR signaling 36 . Other functions have been described for SARM1 in neuronal cell models, such as modulating mitophagy 37 , increasing resistance to TNF treatment-induced apoptosis 38 , and triggering sarmoptosis, a recently uncovered form of programmed cell death, after ROS production 39 .
However, no relationship between SARM1 and cancer or HPV infection has been reported until now. This is the first study to suggest its possible role in HPV tumorigenesis. Initially, we described the increased expression of its gene in cervical tumor cell lines 15 ; here we are reporting its role in cell survival and response to chemotherapy.
Also, we describe changes in the expression of a set of genes after knockout of SARM1 or TLR4. One of the genes is SOD2, a known regulator of oxidative stress and ROS levels. SOD2 has been previously associated with cervical cancer and high grade lesions of the cervix 40 . In addition, higher SOD2 expression is associated with presence of lymph node metastasis in penile cancer 41 . SOD2 is involved in resistance to TNF treatment in keratinocytes 42 and a relationship between LPS treatment and SOD2 expression and activity has also been demonstrated 43 . Additionally, we are the first group to report SOD2 and ROS regulation by SARM1 expression.
TLR pathway is also involved in interleukin expression. We analyzed the expression of IL18 and IL6, interleukins linked to inflammation and tumorigenesis 44 . High risk HPV promote upregulation of IL6 and TLR4, resulting in an inflammatory environment that can contribute to tumor development 45,46 . IL6 can also lead to STAT3 activation, and this pathway contributes to cell proliferation and survival of cervical cancer cells 47 . Here we show that the expression of IL6 and IL18 is modulated by TLR4 and SARM1 in cervical cancer cells.
We also observed a decrease in proliferation after knocking out SARM1 or TLR4, which prompted us to check CCND1, CCNE1 and Ki67. The proliferation marker Ki-67 is associated with high risk HPV and lesion progression 48 , and is widely used as a biomarker, together with p16, to identify HPV infected tissues 49 . Cyclins D1 and E1 (encoded by CCND1 and CCNE1, respectively) are responsible for cell cycle progression. The oncoprotein E6 of HPV can activate the transcription factor eIF4E, and its inhibition in cervical cancer cell lines leads to CCND1 downregulation 50 . Interestingly, cyclin D1 leads to the release of E2F by inducing phosphorylation of pRB, a protein that is already targeted by HPV oncoprotein E7 51,52 . Despite the role of CCNE1 in the proliferation of normal cell lines, it has been reported as a biomarker for chemotherapy resistance in ovarian cancer and for prognosis in breast cancer 53,54 . Both CCND1 and CCNE1 are upregulated in squamous cervical carcinomas 55 . Our data showed modulation of all three genes, Ki67, CCND1 and CCNE1, by TLR4 and SARM1.
In conclusion, we demonstrate here that, in HeLa cells, knock out of SARM1 drastically reduced cell viability and proliferation while lack of TLR4 significantly increased apoptosis and ROS production after cisplatin treatment. Further studies are warranted to explore the potential use of these findings in the clinic. www.nature.com/scientificreports/  www.nature.com/scientificreports/
HeLa cells were purchased from American Type Culture Collection (ATCC® CCL-2™, Lot Number: 59681574) in December 2012. ATCC certified cells through several tests, including: mycoplasma contamination, COI assay, STR analysis, Sterility and Human pathogenic virus presence. Upon receipt HeLa cells were divided, frozen and kept in low passage for all subsequent experiments. Cells were grown in MEM culture medium supplemented with 10% fetal bovine serum (M10 medium) at 37ºC with 5% CO 2 and transfected using the FuGENE HD reagent (Promega, Madison, WI) according to the manufacturer's protocol at a 3:1 ratio between reagent and plasmids, followed by selection with 2 µg/ml of puromycin. Individual clones were isolated by serial dilution of selected cells in 96-well plates. Alterations were confirmed by Sanger sequencing (Supplementary Fig. S1). Knockout cells were obtained before passage 10.
Cell viability assay. Cell viability analysis was performed using the PrestoBlue ™ Cell Viability Reagent kit (ThermoFisher Scientific, Waltham, MA). Cells were seeded in 96-well plates, and 100 μl of complete medium plus 10 μl of PrestoBlue reagent was added to each well. The plates were incubated for 1 h at 37 °C protected from light. Absorbance was measured at 570 and 600 nm. For viability analysis, cells were previously treated with Lipopolysaccharide (LPS) (InvivoGen, San Diego, CA) at 10 µg/mL with and without cisplatin at 4.7 µM (Accord Farmacêutica, São Paulo, Brazil) and analyzed after 72 h of treatment.
Clonogenic, proliferation and migration assays. For clonogenic assays, cells were seeded at low density (1000 cells per well) in 6-well plates. After 2 weeks, cells were fixed with formalin and stained with crystal violet (0.05% w/v). The number of colonies formed was counted automatically using the ImageJ software (National Institutes of Health, Bethesda, MD). For proliferation assays, 5000 cells were seeded per well in 24-well plates and counted every day in triplicates. For migration assays, 3 × 10 6 cells were seeded in 6-well plates; 24 h later, cells were treated with 10 µg/ml of mitomycin C (Sigma-Aldrich, St. Louis, MO) for 2 h at 37ºC. A wound healing assay was performed to observe cell migration, after the mitomycin C treatment. Photos were taken at 0 h, 24 h and 48 h, and the area not occupied by cells was measured using the ImageJ software.
Cell death assays and ROS production measurement. Apoptosis was determined using the CellEv-ent™ Caspase-3/7 Green Detection Reagent kit (ThermoFisher Scientific, Waltham, MA) following the manufacturer's protocol. Apoptosis and necrosis were also determined using double staining with the eBioscience™ 7-AAD Viability Staining Solution and eBioscience™ Annexin V-FITC Apoptosis Detection Kit (ThermoFisher Scientific, Waltham, MA) and analyzed by flow cytometry.
For reactive oxygen species (ROS) determination, cells were cultured in 6-well plates with 10 µg/mL of LPS, 4.7 µM of cisplatin and 100 µM N-acetyl-l-cysteine (NAC) for 72 h, followed by 30 min of incubation with the DCFDA/H2DCFDA reagent (Cellular ROS Assay Kit, ThermoFisher Scientific, Waltham, MA) and analysis by flow cytometry.
For cell cycle analysis, 10 5 cells were seeded per well in 6-well plates and, after 24 h, cells were treated with 10 µg/ml mitomycin for 2 h. Cells were fixed with 70% ethanol, washed with PBS and treated with 5 µl of RNAse (10 mg/ml, ThermoFisher Scientific, Waltham, MA), 0.5 µl of 7-AAD and 0.5 µl of Triton X-100 (Sigma-Aldrich, St. Louis, MO) diluted in 500 µl of PBS for 30 min at 37 °C. Cells were then analyzed by flow cytometry.
All experiments were performed on a BD Accuri™ C6 cytometer (BD Biosciences, CA) and the results analyzed with FlowJo software (BD, Franklin Lakes, NJ, version V10).