30-hydroxygambogic acid activates overlapping and independent apoptotic pathways in HPV positive and HPV negative oral cancer cells

TL;DR

This study shows that 30-hydroxygambogic acid can kill both HPV-positive and HPV-negative oral cancer cells by triggering overlapping and independent apoptotic pathways.

Contribution

The study introduces GA-OH as a novel compound that activates multiple apoptotic pathways in both HPV(+) and HPV(−) oral cancer cells.

Findings

GA-OH treatment reduces the viability of both HPV(+) and HPV(−) oral cancer cells.

GA-OH induces cell death through overlapping and independent apoptotic pathways.

Gene expression changes reveal distinct and shared mechanisms of cell death in HPV(+) and HPV(−) cells.

Abstract

The global incidence of Head and Neck (HN) cancer has dramatically increased over the past few decades, primarily due to an increasing incidence of HPV infection. HPV infection desensitizes cells to apoptosis through the E6-enabled accelerated degradation of several pro-apoptotic molecules, including p53 and procaspase 8. To block this activity, we used 30-hydroxygambogic acid, GA-OH, a small molecule that binds to HPV16 E6 and inhibits the interactions of E6 with its cellular partners. We found that treatment with GA-OH affects the viability of both HPV(+) and HPV(−) oral cancer cells. Further analysis of gene expression patterns of these cell lines showed that GA-OH induces cell death through both independent and overlapping apoptotic pathways by altering gene expression in both HPV(+) and HPV(−) cancer cells.

Figures

Volcano plots generated from the DEGs found in the UMSCC19(−), SCC84(−), UMSCC47(+), and UMSCC104(+) cell lines. A) Volcano plot for the HPV(−) cell line UMSCC19 depicting the up and down-regulated DEGs. B) Volcano plot for the other HPV(−) cell line, SCC84, depicting the DEGs. C) Volcano plot for the HPV(+) cell line UMSCC47 DEGs. D) Volcano plot depicting the DEGs for the second HPV(+) cell line, UMSCC104.Fig. 1

Venn diagrams of DEGs identified in HPV(+) and HPV(−) cell lines. A) Clusters of overlapping and non-overlapping DEGs in four cell lines used in current study. B) Venn diagram showing number of 466 common genes and number of DEGs specific in HPV(+) and HPV(−) cells.Fig. 2

Network plots of GO terms enriched in oral cancer cells treated with GA-OH. A) Common DEGs, B) Only in HPV(−) cells, C) Only in HPV(+) cells. Bubble color indicates p-value, its size shows the frequency of the GO term. Highly similar GO terms are linked with the lines, their width indicates the degree of similarity.Fig. 3Table 1Common DEGs in HPV(+) and HPV(−) cells.Table 1Term%PValueGenesFold EnrichmentFDRLipid and atherosclerosis4.725.54E-07ABCA1, HSPA8, JUN, HSP90AA1, CXCL8, HSPA1L, HSPA5, NCF2, MMP1, HSPA6, NFATC2, LY96, FOS, CXCL3, CXCL2, HSPD1, ERN1, IL6, DDIT3, HSPA1B, HSPA1A, MAP2K63.6

MTT Assays from the seven cell lines. A) UMSCC47(+) cells treated with GA-OH (0-1.8 μM GA-OH); B) SCC90(+) cells treated with GA-OH (0-4 μM GA-OH); C) UMSCC104(+) cells treated with GA-OH (0-6 μM GA-OH); D) UMSCC19(−) cells treated with GA-OH (0-1.8 μM GA-OH); E) SCC84(−) cells treated with GA-OH (0-5 μM GA-OH); F) UMSCC1(−) cells treated with GA-OH (0-4 μM GA-OH); G) UDSCC2(+) cells treated with GA-OH (0-5 μM GA-OH).Fig. 4

The Caspase 3/7 assay. This graph shows the luminescence of each sample on the y-axis compared to the GA-OH treatment concentrations for each cell line on the x-axis. There were three treatments. Grey represents Control (0 μM GA-OH), White is representative of 0.5 μM GA-OH, and Black columns depict the 1 μM GA-OH treatment.Fig. 5

Flow Cytometry analysis. The left column shows the analyzed data with the Annexin V stain on the x-axis and the 7AAD stain on the y-axis. The right column depicts the graphs of the different percentages of cells in each quadrant of the treatments and cell lines in the left column. A) UMSCC47(+) analysis of 0 μM, 0.5 μM, and 1 μM GA-OH treatments from left to right respectively. B) SCC90(+) analysis of 0 μM, 0.5 μM, and 1 μM GA-OH treatments from left to right respectively. C) UMSCC104(+) analysis of 0 μM, 0.5 μM, and 1 μM GA-OH treatments from left to right respectively. D) Graph depicting the

p21 Western Blot. This figure shows the western blot run with the HPV(+) cell line UMSCC47 and the HPV(−) cell line UMSCC19. The graph displays the Relative Density Unit (RDU) compared to the GA-OH treatment concentration for the two cell lines with UMSCC47(+) in blue and UMSCC19(−) in orange.Fig. 7

Full text

Glossary

HPV –Human papilloma virusGA-OH –30-hydroxygambogic acidTRAIL –TNF-related apoptosis-inducing ligand proteinHNSCC –Head and neck squamous cell carcinomaDMEM –Dulbecco's modified eagle's mediumFBS –Fetal bovine serumEGF –Epidermal growth factorDAVID -Database for Annotation, Visualization, and Integrated DiscoveryDMSO –Dimethyl sulfoxide

Introduction

1

According to the World Health Organization, the incidence of oral cancer is expected to increase by approximately 50% by the end of 2050. Analysis of the data shows that exposure to risk factors such as tobacco, alcohol and high-risk types of HPV can lead to the development of these types of cancer. Carcinogenesis is a multifactorial process that involves genetic mutagenesis, epigenetic modifications and viral HPV infection. Human Papillomavirus (HPV) is known to infect the epithelial cells of the skin and mucous membranes such as the genital areas and the throat. The high-risk types of HPV, of which the most common are types 16 and 18, are known to be associated with cancer formation [1].

Several types of cancers are associated with HPV infection. The two most common are cervical cancer, found in women, and head and neck cancer, also referred to as oropharyngeal cancer, found in both men and women with the highest prevalence in men [2]. Worldwide, the most common HPV-related cancer is cervical cancer [2]. However, in the US, head and neck cancer has been on the rise in recent years, becoming the most common HPV-related cancer [3]. There are around 900,000 cases of head and neck cancer per year and about 400,000 deaths per year worldwide [4]. However, HPV is not the only cause of head and neck cancer. Smoking, oral tobacco, and alcohol usage can also increase one's risk of developing head and neck cancer [3,5]. Around 90% of all head and neck cancer cases are squamous cell carcinomas (HNSCC). These cancers are found in various tissues, including those from the pharynx, nasopharynx, oral cavity, and the larynx [5].

Various preventative measures against HPV-related cancers have been employed, including the US Food and Drug Administration (FDA) HPV-approved vaccines [2]. However, available vaccines are not effective in individuals who have already been exposed to the virus. One possible way to prevent and/or reduce the effects of HPV infection is to inhibit the HPV E6 oncoprotein. This oncoprotein, by binding to and accelerating the degradation of pro-apoptotic cellular proteins, protects HPV-infected cells from agents that induce apoptosis, including chemotherapies based on TRAIL, Cisplatin, and Doxorubicin [6,7]. As a result, these chemotherapies have limited effectiveness against HPV-associated cancers.

A therapeutic strategy that targets the interaction between E6 and apoptosis-related proteins with small, drug-like molecules could potentially reconstitute cellular levels of p53 and Caspase 8, leading to the re-sensitization of cells to apoptosis and the increased efficacy of apoptosis-inducing agents against HPV(+) tumors in vivo [8]. By screening a small molecule library for target discovery and validation [9] as well as the development of assays, hit identification, and lead identification [8,10] we were able to identify 30-hydroxygambogic acid, GA-OH, as a small molecule that inhibits the interactions of E6 with its cellular partners, thereby making it a potential candidate for combinatorial treatment [11].

In this study, we demonstrate that GA-OH can adversely affect the viability of both HPV(+) and HPV(−) cancer cells and their ability to replicate, and that these two types of cells utilize both overlapping and distinct pathways in their response, including but not limited to apoptotic pathways. This data coincides with the observation that the parent compound, gambogic acid, may possess a more generalizable anti-cancer activity [12]. Overall, our data demonstrates that GA-OH may have a beneficial effect on both HPV-positive and HPV-negative oral cancers both by itself and together with other cancer treatments.

Materials & methods

2

Cell culture

2.1

Cells employed were HPV(+) or HPV(−) head and neck squamous cell carcinoma (HNSCC) cancer cell lines. The HNSCC cell lines were obtained from several sources: UM-SCC19(−), UM-SCC1(−), and UM-SCC104(+) cell lines were a gift from Dr. Thomas Carey, University of Michigan (Michigan, USA). The UM-SCC47(+), SCC90(+), and SCC84(−) cell lines were a gift from Dr. John Lee, Sanford Research (South Dakota, USA). The UD-SCC2(+) cell line originated from Drs. Thomas K. Hoffman and Henning Bier (University of Düsseldorf, Germany). They were grown in Dulbecco's modified eagle's medium (DMEM) containing 9% fetal bovine serum (FBS), 1% Penicillin-streptomycin, 0.5% Amphotericin B, and 0.2% Primocin or E-Media containing DMEM, Ham's F12, FBS, Penicillin-streptomycin, Amphotericin B, Triiodothyronine, Hydrocortisone, Cholera toxin, transferrin, insulin, EGF, and Primocin. The cells were grown at 37 °C with 5% CO_2_.

RNAseq data generation

2.2

Two HPV(+) cell lines (UMSCC47, UMSCC104) and two HPV(−) cell lines (UMSCC19, SCC84) were grown and treated with either 0 μM GA-OH (DMSO) or 0.5 μM GA-OH. Then, 24 h after treatment, the cells were collected following the Qiagen RNeasy Protect Mini Kit (74624) instructions. Once collected, the samples were prepared according to the protocol. After preparation, the concentration of the samples was determined before dilution to be 250 ng/ml. Once at the appropriate concentration, the quality of the RNA was determined using a Qubit Fluorometer. The Universal Plus™ Total RNA-Seq library preparation kit with NuQuant (Tecan) was used per manufacturer's instructions to construct RNA-Seq libraries. 100 ng of total RNA was used as input. After the first and second strand of cDNA synthesis, end-repair, adaptor index ligation, and strand selection were conducted. Barcodes with unique indices were used per sample for multiplexing. Ribosomal RNA depletion was performed by using custom InDA-C primer mixture SS5 V3 for human. Finally, libraries were amplified for 13 cycles, and purified with Agencourt XP beads (Beckman Coulter, Indianapolis, IN). The final, purified libraries were quantified using the Qubit dsDNA HS Kit on Qubit 4.0 Fluorometer (Life Technologies, Carlsbad, CA). Quality and peak size was determined with the D1000 ScreenTape on Agilent 2200 TapeStation (Agilent Technologies, Santa Clara, CA). Final libraries of different indices were diluted to 4 nM and pooled at equimolar amount for sequencing on a Illumina NextSeq 550 FlowCell with high output kit (Illumina, Inc., San Diego, CA). Single reads with 82 bp were generated.

Illumina RTA v2.4.11 software was used for basecalling and bcl2fastq v2.17.14 was used for generating fastq files.

Differentially expressed genes (DEGs) identification

2.3

The RNA-Seq raw fastq files were first trimmed using Trim Galore (v0.6.5), then the trimmed reads were aligned to the human reference genome GRCh38 using STAR [13] with default parameter settings. The aligned bam files were processed using HTSeq-Count [14] for gene quantification. Genes with CPM ≥1 in all samples were used for differential gene expression analyzed using R package edgeR [15], and the differentially expressed genes (DEG) were identified with FDR <0.01, and |log2(FC)| > 1.

DEGs and pathway enrichment analysis of DEGs

2.4

Venn diagrams for identified DEGs were built using Venny2.1 online software [16]. To identify pathways enriched with DEGs, we used the DAVID database that provides annotation information of biological functions for genes [17]. The KEGG, Kyoto Encyclopedia of Genes and Genomes, database was used as a source of pathways analyzed with DAVID online software. Gene Ontology (GO) analysis was performed with GOrilla online software [18].

Supplemental gene set enrichment analysis [19] was performed in R (version 4.1.1) with the “fgsea” library (version 1.18.0), using the Hallmark and Reactome gene sets for Homo Sapiens from the Molecular Signatures Database [20]. Genes were ranked according to sign (log Fold Change) ∗ -log (P-value) quantities from DEG analyses (treatment vs. control). Enriched gene sets were considered “common” if they had Adjusted P-value (padj) < 0.01 in all four cell lines, and in that case normalized enrichment scores (NESs) and padj values were averaged across all four cell lines for summarization. Gene sets unique to HPV(−) cell lines had padj <0.01 in at least one HPV(−) cell line, but in neither HPV(+) cell lines, and in that case, NESs and padj values were averaged between the two HPV(−) cell lines for summarization. Similarly, gene sets unique to HPV(+) cell lines had padj <0.01 in at least one HPV(+) cell line, but in neither HPV(−) cell lines, and in that case, NESs and padj values were averaged between the two HPV(+) cell lines for summarization.

GA-OH treatment and MTT assay

2.5

The MTT assay was carried out on the individual cell lines seeded in 96 well plates with either four technical replicates (UMSCC47(+), UMSCC19(−)) or eight technical replicates (SCC90(+), UMSCC104(+), UDSCC2(+), UMSCC1(−), SCC84(−)). Each cell line was tested using a minimum of three biological replicates. Once seeded, the cells were allowed to incubate for 24 h, then the plates were treated with various concentrations of GA-OH ranging from 0 to 6 μM. Twenty-four hours after treatment, MTT analysis was performed as described [7].

Caspase 3/7 Glo assay

2.6

The Caspase 3/7 Glo Assay (Promega, G8091) was carried out according to the manufacturer's protocol. Each of the cell lines was assayed in duplicate for each of the three treatment concentrations (0 μM GA-OH (DMSO), 0.5 μM GA-OH, and 1 μM GA-OH) and in triplicate for the media blanks. The cells were seeded in a white-walled 96-well plate with approximately 20,000 cells per well. Twenty-four hours after seeding, the cells were treated according to their type: blank wells – 10 μL DMSO in media without cells, negative control – 10 μL DMSO in media with cells, 0.5 μM experimental sample – 10 μL 0.5 μM GA-OH, 1 μM experimental sample – 10 μL 1 μM GA-OH. Then, 24 h after treatment, 100 μL of the caspase buffer solution was added to each well and the plate was allowed to incubate at room temperature for about 3 h before the luminescence was read by a plate reader.

Flow cytometry

2.7

Cells were maintained as previously described and treated with either DMSO (0 μM GA-OH), 0.5 μM GA-OH, or 1 μM GA-OH 24 h before collection. Once collected, the cells were counted for an ending total of 200,000–300,000 cells per well. They were washed twice with Cell Staining Buffer (Biolegend, cat. 420201) before following the staining protocol for Annexin V (Biolegend, cat. 640918) and 7AAD (Biolegend, cat. 420403). Each cell line contained ten individual samples: non-stained, single-stained Annexin V, single-stained 7AAD, double-stained Annexin V + 7AAD, 0 μM GA-OH Annexin V, 0.5 μM GA-OH Annexin V, 1 μM GA-OH Annexin V, 0 μM GA-OH Annexin V + 7AAD, 0.5 μM GA-OH Annexin V + 7AAD, 1 μM GA-OH Annexin V + 7AAD. Samples were then run on a MACSquant Analyzer 10 Flow Cytometer (Miltenyi Biotec), and the data was analyzed using FlowJo analysis software.

p53 ELISA

2.8

The p53 ELISA kit was run according to the protocol (Invitrogen, cat. BMS256). The seven cell lines were treated with either 0 μM GA-OH (DMSO), 0.5 μM GA-OH, or 1 μM GA-OH and were left to incubate in the incubator for 24 h. Then, the media was collected and centrifuged at 1500 rpm for 5 min. As indicated by the kit instructions, 50 μL of each sample was added to the appropriate wells in triplicate. The protocol was followed according to the kit instructions and the Substrate Solution was allowed to incubate for approximately 15 min before the Stop Solution was added. Then the absorbances were read using a plate-reader spectrophotometer.

Western blot

2.9

Cells were grown as previously described, then treated with 0 μM GA-OH (DMSO), 0.5 μM GA-OH, or 1 μM GA-OH. Twenty-four hours after treatment, the cells were first washed with PBS twice before being collected using 250 μL sterile water, 250 μL Loading Buffer, 25 μL 2-mercapto-ethanol and 2 μL Bromophenol blue. After sonication, samples were incubated at 90 °C for 10 min and applied to the gel (Thermo Fisher). The gels were run at 100 V for 132 min, before being transferred to a membrane using the iBlot Transfer Device (Invitrogen). Once transferred, the membrane was transferred to a black box and allowed to incubate at room temperature on the hula mixer for 1 h in Blocking Buffer (Li-Cor). After the incubation, the blocking buffer was removed, and the membrane was allowed to incubate overnight (ON) in the primary antibody of choice at 4 °C on a rocker. The next morning, the primary antibody was collected, and the membrane was washed with TBS-T three times for 10 min each, then it was rinsed with PBS once before the secondary antibody was added. With the secondary antibody, the membrane was incubated at room temperature for 30 min to 3 h. After the secondary antibody incubation, the membrane was washed with TBS-T three to five times for 10 min each before being rinsed with PBS 3 times. Next, the membrane was then scanned using the Li-Cor Odyssey CLx Imager and analyzed in ImageStudio. To reprobe with another primary antibody, the PBS was removed, the new primary antibody was added, and the membrane was once again incubated ON at 4 °C on a rocker and the steps were repeated.

Results

3

GA-OH activates both overlapping and distinct pathways in HPV( + ) and HPV(−) cell lines

3.1

To identify changes in the transcriptomes of HPV-positive and HPV-negative HNSCC oral cancer cells in response to GA-OH treatment, we treated two HPV(+) and two HPV(−) cell lines with 0.5 μM GA-OH for 24 h followed by RNA isolation. These samples were analyzed by RNAseq to identify differentially expressed genes (DEGs) based on treatment and HPV status, as illustrated in Fig. 1. Our results identified multiple DEGs across the HPV(+) cell lines: UMSCC47 and UMSCC104, and the HPV(−) cell lines: UMSCC19 and UMSCC84. In addition, we noted that the HPV(+) UMSCC 47 cell line displayed fewer DEG, possibly due to technical issues. The identified DEGs in these four cell lines are listed in Supplemental Tables 1, 2, 3, and 4 correspondingly.Fig. 1. Volcano plots generated from the DEGs found in the UMSCC19(−), SCC84(−), UMSCC47(+), and UMSCC104(+) cell lines. A) Volcano plot for the HPV(−) cell line UMSCC19 depicting the up and down-regulated DEGs. B) Volcano plot for the other HPV(−) cell line, SCC84, depicting the DEGs. C) Volcano plot for the HPV(+) cell line UMSCC47 DEGs. D) Volcano plot depicting the DEGs for the second HPV(+) cell line, UMSCC104.Fig. 1

Following the identification of the DEGs, we performed Gene Ontology (GO) and pathway analyses using the Gorilla and DAVID website analysis tools to determine the most significantly enriched GO terms and particular pathways, correspondingly. To facilitate this, we created a Venn diagram comparing all cell lines (Fig. 2A) as well as one comparing the HPV+ and HPV- cells (Fig. 2B). As noted above, the cell line UMSCC47 had a significantly lower number of identified DEGs compared to the other three, prompting us to exclude it from further analysis to avoid bias.Fig. 2. Venn diagrams of DEGs identified in HPV(+) and HPV(−) cell lines. A) Clusters of overlapping and non-overlapping DEGs in four cell lines used in current study. B) Venn diagram showing number of 466 common genes and number of DEGs specific in HPV(+) and HPV(−) cells.Fig. 2

We then analyzed the DEGs common to both HPV(+) and HPV(−) cells, as well as those found exclusively in each type of cells. Fig. 3 shows Gene Ontology (GO) terms enriched in both cell lines (3 A), negative (3B) or positive (3C) cells treated with GA-OH. Common to both types of cells are DEGs enriched in two major clusters: response to oxidative stress, and immune response including regulation of extrinsic apoptotic signaling (Fig. 3A). Apoptotic mitochondrial changes are also enriched among this group of DEGs. Enriched GO terms linked to chromatin organization are found in DEGs of HPV(−) cells, whereas for HPV(+) cells are noted DEGs involved in protein folding (Fig. 3B and C). Further analysis of the common DEGs using the DAVID web service revealed 16 KEGG pathways with an FDR <0.05, which are detailed in Table 1. These pathways include genes involved in the MAPK and NF-kappa B pathways, histone-encoding genes, and those related to cytokine responses. Conversely, the DEGs exclusive to HPV(−) cells showed less significant pathway enrichment, with only six pathways having an FDR <0.05. Table 2 lists 15 pathways with a P value less than 0.05, highlighting that histone-encoding genes are particularly overrepresented, indicating significant DNA structural changes as a response to GA-OH treatment. Additionally, the p53 pathway and cell cycle regulation appear to be affected by GA-OH treatment. DAVID analysis of genes found exclusively in HPV(+) cells revealed limited pathway enrichment, identifying only five pathways with a P value less than 0.05, including the PI3K-Akt and MAPK signaling pathways (Table 3).Fig. 3. Network plots of GO terms enriched in oral cancer cells treated with GA-OH. A) Common DEGs, B) Only in HPV(−) cells, C) Only in HPV(+) cells. Bubble color indicates p-value, its size shows the frequency of the GO term. Highly similar GO terms are linked with the lines, their width indicates the degree of similarity.Fig. 3. Table 1Common DEGs in HPV(+) and HPV(−) cells.Table 1. Term%PValueGenesFold EnrichmentFDRLipid and atherosclerosis4.725.54E-07ABCA1, HSPA8, JUN, HSP90AA1, CXCL8, HSPA1L, HSPA5, NCF2, MMP1, HSPA6, NFATC2, LY96, FOS, CXCL3, CXCL2, HSPD1, ERN1, IL6, DDIT3, HSPA1B, HSPA1A, MAP2K63.650.0002Legionellosis2.362.68E-06CLK1, HSPA8, IL6, CXCL8, HSPA1L, HSPA6, CXCL3, CXCL2, HSPA1B, HSPD1, HSPA1A7.030.0004Alcoholism3.861.81E-05H2BC9, H3C8, H2AC8, H2BC7, H2AC4, H2BC4, H2AC16, H2BC17, GNGT1, H2AC11, H2AC12, H2BC13, H3C11, H2BC10, H3C12, GNB4, FOSB, CREB53.430.0012Protein processing in endoplasmic reticulum3.651.95E-05PPP1R15 A, HSPA8, ERO1B, HSP90AA1, HSPA1L, HSPA5, HSPA4L, HSPA6, HERPUD1, ERN1, DNAJB2, DNAJB1, HSPH1, DDIT3, CRYAB, HSPA1B, HSPA1A3.580.0012Systemic lupus erythematosus3.223.05E-05H2BC9, H3C8, H2AC8, H2BC7, C1S, H2AC4, H2BC4, H2AC16, H2BC17, H2AC11, H2AC12, H2BC13, H3C11, H2BC10, H3C123.860.0012MAPK signaling pathway4.943.09E-05HSPA8, DUSP5, JUN, PLA2G4D, HSPA1L, GADD45A, DUSP1, HSPA6, PLA2G4C, FOS, DUSP8, PGF, CACNA1H, GADD45G, VEGFA, FGF19, DDIT3, STMN1, MAP3K6, FGFR3, HSPA1B, HSPA1A, MAP2K62.740.0012Kaposi sarcoma-associated herpesvirus infection3.863.10E-05CDKN1A, JUN, TCF7L1, CXCL8, CSF2, NFATC2, FOS, CXCL3, CXCL2, VEGFA, GNGT1, RCAN1, IL6, MAP1LC3B, UBC, GNB4, E2F2, MAP2K63.290.0012IL-17 signaling pathway2.586.14E-05FOSL1, IL6, HSP90AA1, JUN, CXCL8, CSF2, MMP1, FOSB, FOS, CXCL3, CXCL2, MUC5AC4.520.0021Rheumatoid arthritis2.362.69E-04IL11, IL6, JUN, CXCL8, CSF2, MMP1, FOS, CXCL3, CXCL2, ATP6V0A1, VEGFA4.190.0083Estrogen signaling pathway2.794.81E-04HSPA8, JUN, HSP90AA1, HSPA1L, HSPA6, FOS, KRT9, KRT33B, KRT17, HSPA1B, HSPA1A, HBEGF, CREB53.350.0133Bladder cancer1.508.68E-04CDKN1A, CXCL8, MMP1, E2F2, FGFR3, VEGFA, HBEGF6.110.0195Neutrophil extracellular trap formation3.229.19E-04H2BC9, H3C8, H2AC8, H2BC7, NCF2, H2AC4, H2BC4, H2AC16, H2BC17, H2AC11, H2AC12, H2BC13, H3C11, H2BC10, H3C122.800.0195Transcriptional misregulation in cancer3.229.65E-04H3C8, CDKN1A, CXCL8, CSF2, GADD45A, ETV5, PBX1, GADD45G, CCNA1, IL6, NR4A3, H3C11, DDIT3, H3C12, NFKBIZ2.780.0195Ferroptosis1.509.89E-04MAP1LC3B, HMOX1, SLC3A2, SLC7A11, GCLM, SAT1, FTL5.960.0195Fluid shear stress and atherosclerosis2.580.00188251NPPC, HSP90AA1, JUN, NCF2, DUSP1, CAV1, HMOX1, FOS, ARHGEF2, SQSTM1, MAP2K6, VEGFA3.050.0346Table 2HPV- exclusive DEGs.Table 2. Term%PValueGenesFold EnrichmentFDRSystemic lupus erythematosus3.786.61E-14H2AC15, H2AC13, H2AC14, H2BC26, H3C13, H2BC21, H3C1, H3C10, H3C3, H4C2, H3C2, H4C1, H4C3, H3C7, H4C5, H4C8, H2AZ2, H2BC8, H2AC6, H2BC6, H2BC3, H2BC18, H2BC12, H2BC14, H2BC15, H4C16, H2AC20, H4C13, H2AC21, H2BC115.501.05E-11Neutrophil extracellular trap formation4.416.96E-14HMGB1, H2AC15, HDAC9, ACTG1, H2AC13, H2AC14, H2BC26, H3C13, CLEC7A, H2BC21, CASP4, H3C1, H3C10, H3C3, H4C2, H3C2, H4C1, H4C3, H3C7, H4C5, H4C8, H2AZ2, H2BC8, H2AC6, H2BC6, H2BC3, H2BC18, H2BC12, H2BC14, H2BC15, H4C16, H2AC20, H4C13, H2AC21, H2BC114.641.05E-11Alcoholism4.292.15E-13SHC4, H2AC15, HDAC9, H2AC13, H2AC14, H2BC26, H3C13, H2BC21, H3C1, H3C10, H3C3, H4C2, H3C2, H4C1, H4C3, H3C7, H4C5, H4C8, H2AZ2, H2BC8, H2AC6, H2BC6, H2BC3, H2BC18, GNAO1, H2BC12, H2BC14, H2BC15, H4C16, H2AC20, H4C13, H2AC21, H2BC11, ATF44.612.16E-11Viral carcinogenesis3.401.02E-07HDAC9, CDC20, H2BC26, H2BC21, H4C2, H4C1, SKP2, H4C3, YWHAH, H4C5, H4C8, H2BC8, H2BC6, H2BC3, TRAF1, H2BC18, H2BC12, H2BC14, H2BC15, H4C16, H4C13, H2BC11, CDK2, CDK1, IL6ST, ATP6V0D2, ATF43.367.66E-06Cell cycle2.404.30E-05PLK1, PPP2R5B, BUB1B, TTK, ESCO2, KNL1, PKMYT1, CDC25C, AURKB, TICRR, SGO1, CDC20, CCNB2, CHEK2, CDK2, CDK1, SKP2, MCM2, YWHAH3.060.00259116Steroid biosynthesis0.888.37E-05SQLE, EBP, MSMO1, DHCR24, DHCR7, LSS, FDFT18.920.00419991Biosynthesis of amino acids1.262.50E-03PC, PSAT1, SHMT2, GPT2, SHMT1, IDH2, PHGDH, NAGS, BCAT1, ASS13.400.10731185Glycine, serine and threonine metabolism0.884.86E-03DMGDH, ALAS1, AOC2, PSAT1, SHMT2, SHMT1, PHGDH4.350.18275623One carbon pool by folate0.636.79E-03DHFR, SHMT2, MTHFD2, SHMT1, ALDH1L26.370.22720873p53 signaling pathway1.138.72E-03CCNB2, STEAP3, CHEK2, CDK2, CDK1, MDM4, THBS1, GTSE1, DDB23.060.26256982Oocyte meiosis1.512.03E-02SGO1, CDC20, CCNB2, CPEB1, PLK1, CDK2, PPP2R5B, CDK1, ITPR1, PKMYT1, CDC25C, YWHAH2.200.55472964Cytokine-cytokine receptor interaction2.522.47E-02IL32, CCL24, TSLP, CSF1, IL10RA, TNFRSF9, IL31RA, IL20RB, IL16, NGF, OSMR, IL22RA1, IL36B, CLCF1, TNFSF9, TNFRSF14, IL36RN, IL6ST, IL18R1, TNFRSF211.710.6191096Proteasome0.763.29E-02PSMB7, PSMD12, PSMA1, PSMD13, PSMB8, PSMB93.320.76145971Metabolic pathways9.463.96E-02DMGDH, ALAS1, ST6GALNAC2, PANK1, MSMO1, ALDH1L2, HK2, SMPD3, HYAL2, NAMPT, ENPP2, AOX1, PHGDH, TK1, PCYT1B, GPT2, MARS1, AMPD3, LSS, DCK, ITPKC, MTHFD2, PIGA, B3GNT4, ADPRM, BCAT1, ATP6V0D2, MGLL, ACOT4, PTGES, IDI1, PFKFB3, MVK, SHMT2, SHMT1, GCNT1, NDUFA4L2, ALOX12B, HSD17B11, FUT1, FUT3, PTGS1, ALDH3B2, EBP, CYP11A1, RDH16, B4GALNT2, B4GALNT4, FDFT1, GSTM3, OLAH, FDPS, DUT, AOC2, GK, SRD5A2, GALNT2, NOS3, IDH2, SCD5, GFPT1, DHCR24, ASS1, DHFR, LACC1, SQLE, PDE10A, PC, PSAT1, SCD, ALPP, MGAT4A, DHCR7, NAGS, GALK11.220.7955768ATP-dependent chromatin remodeling1.263.96E-02CECR2, H2AZ2, H2AC13, H2AC14, H2AC6, H2AC20, ACTR5, H2AC21, H2AC15, ACTG12.180.7955768Table 3HPV + exclusive DEGs.Table 3KEGG Pathway%PValueGenesFold EnrichmentFDRPI3K-Akt signaling pathway3.740.003971IFNAR2, PHLPP1, HSP90AB1, LAMA1, LAMA4, TNC, LAMB1, PRLR, AREG, NR4A1, ERBB3, ARTN, DDIT4, MYB, ITGB8, IL7R, FGFR12.220.99Transcriptional misregulation in cancer2.200.020441CCNA2, MEF2C, ZEB1, HPGD, GADD45B, PAX8, TLX3, IGFBP3, TFE3, ETV42.451.00MAPK signaling pathway2.860.024445MAP4K1, MEF2C, GADD45B, PLA2G4E, HSPB1, AREG, CACNA1G, IL1A, NR4A1, ERBB3, ARTN, MAPT, FGFR12.051.00ECM-receptor interaction1.320.039282SV2A, LAMA1, LAMA4, TNC, ITGB8, LAMB13.191.00Human papilloma virus infection2.860.048336HES7, IFNAR2, NOTCH1, LAMA1, LAMA4, MX2, TNC, LAMB1, OASL, CCNA2, WNT6, HEY1, ITGB81.851.00Influenza A1.760.069927IFNAR2, IL1A, RSAD2, OAS1, OAS2, CCL5, MX2, TNFSF102.201.00Progesterone-mediated oocyte maturation1.320.084559CCNA2, HSP90AB1, CDC26, CPEB3, MAD1L1, CPEB42.561.00Cytokine-cytokine receptor interaction2.420.097837EDAR, IFNAR2, IL1A, CCL20, CCL5, LEPR, TNFSF10, CXCL1, INHBA, IL7R, PRLR1.741.00

GO and pathway analyses revealed that GA-OH treatment triggered pro-apoptotic cell death gene responses in both HPV(+) and HPV(−) cell lines. At the same time, their response differs, with a more pronounced involvement of histone genes expression changes in HPV(−) cells.

GA-OH affects survival in both HPV( + ) and HPV(−) HNSCC cell lines

3.2

Data from the RNAseq analysis predicted that treatment with GA-OH would affect viability (Table 1, Table 2, Table 3). To test this prediction, we examined the effect of GA-OH on the viability of HPV(+) and HPV(−) HNSCC cell lines by conducting MTT assays on seven cell lines (four HPV(+) and three HPV(−) cell lines). These cells were treated with GA-OH, and 24 h later, the MTT assay was conducted. We found that GA-OH was effective in inducing cell death in both HPV(+) and HPV(−) HNSCC cell lines, with IC_50_ values between approximately 1 and 4 μM GA-OH (Table 4, Fig. 4). Each cell line was assayed multiple times to determine the most appropriate range of concentrations for that specific cell line. The cell lines that were more resistant to treatment with GA-OH were assayed at higher concentrations of GA-OH, while those that were less resistant to treatment with GA-OH were assayed with lower concentrations of GA-OH (Fig. 4).Table 4. The IC_50_ table lists all calculated IC_50_ values for the cell lines and their HPV status.Table 4. Cell LineHPVIC_50_ GA-OH (μM)UMSCC47+0.99SCC90+2.59UMSCC104+2.08UDSCC2+4.19UMSCC19-1.06SCC84-2.69UMSCC1-1.87Fig. 4MTT Assays from the seven cell lines. A) UMSCC47(+) cells treated with GA-OH (0-1.8 μM GA-OH); B) SCC90(+) cells treated with GA-OH (0-4 μM GA-OH); C) UMSCC104(+) cells treated with GA-OH (0-6 μM GA-OH); D) UMSCC19(−) cells treated with GA-OH (0-1.8 μM GA-OH); E) SCC84(−) cells treated with GA-OH (0-5 μM GA-OH); F) UMSCC1(−) cells treated with GA-OH (0-4 μM GA-OH); G) UDSCC2(+) cells treated with GA-OH (0-5 μM GA-OH).Fig. 4

Caspase 3/7 is activated in both HPV( + ) and HPV(−) cell lines

3.3

The Caspase 3/7 Glo assay was employed to validate activation of the apoptotic pathway. Each of 7 cell lines (3 HPV(−) lines, UMSCC1, UMSCC19, and SCC84; and 4 HPV(+) lines, UMSCC104, SCC90, UMSCC47 and UDSCC2) was treated with either DMSO (0 μM GA-OH), 0.5 μM GA-OH, or 1 μM GA-OH for 24 h before the assay was conducted according to the protocol. Once the caspase reagent was added, the plate was incubated at room temperature for about 3 h before the luminescence was determined for each sample using a plate reader. Results of this Caspase 3/7 analysis clearly show that all tested cell lines significantly increase the activity of these proteins following treatment (Fig. 5).Fig. 5. The Caspase 3/7 assay. This graph shows the luminescence of each sample on the y-axis compared to the GA-OH treatment concentrations for each cell line on the x-axis. There were three treatments. Grey represents Control (0 μM GA-OH), White is representative of 0.5 μM GA-OH, and Black columns depict the 1 μM GA-OH treatment.Fig. 5

GA-OH induces cell death in HPV( + )HNSCC cell lines via apoptosis

3.4

To confirm the mechanism of cell death initiated by treatment with GA-OH for the HPV(+) cells we employed flow cytometry in addition to the caspase assay. The three HPV(+) cell lines UMSCC47, SCC90 and UMSCC104 were treated with either DMSO (0 μM GA-OH), 0.5 μM GA-OH, or 1 μM GA-OH 24 h before collection. Then, the cells were prepared according to the staining protocols for Annexin V to detect apoptotic cells and 7AAD as a cell viability stain to determine which cells were alive or dead. The results demonstrated that GA-OH was able to kill the three HPV(+) HNSCC cell lines via apoptosis, as there was a clear dose-dependent increase in the number of cells in either early or late-stage apoptosis following treatment (Fig. 6).Fig. 6. Flow Cytometry analysis. The left column shows the analyzed data with the Annexin V stain on the x-axis and the 7AAD stain on the y-axis. The right column depicts the graphs of the different percentages of cells in each quadrant of the treatments and cell lines in the left column. A) UMSCC47(+) analysis of 0 μM, 0.5 μM, and 1 μM GA-OH treatments from left to right respectively. B) SCC90(+) analysis of 0 μM, 0.5 μM, and 1 μM GA-OH treatments from left to right respectively. C) UMSCC104(+) analysis of 0 μM, 0.5 μM, and 1 μM GA-OH treatments from left to right respectively. D) Graph depicting the percentages of cells in each quadrant of the UMSCC47(+) cell line with the 0 μM GA-OH on the left, 0.5 μM GA-OH in the middle, and 1 μM GA-OH on the right. E) Same as D but with the SCC90(+) cell line. F) Same as D and E but with the UMSCC104(+) cell line.Fig. 6

Protein expression

3.5

Data from the RNAseq analysis predicted an increase in the expression of apoptotic genes such as p21. The p21 protein is a well-known CDK inhibitor that is responsible for inhibiting the cell cycle progression through the G1 and S phases, thus promoting apoptosis [21]. To test this prediction, we assessed expression of the p21 protein using western blot analysis in two cell lines, UMSCC47 (+) and UMSCC 19 (−) (Fig. 7). Each cell line was treated with either DMSO (0 μM GA-OH), 0.25 μM GA-OH, 0.5 μM GA-OH, or 1 μM GA-OH 24 h before collection. Fig. 7 shows the western blot run with the HPV(+) cell line UMSCC47 and the HPV(−) cell line UMSCC19. The graph displays the Relative Density Unit (RDU) compared to the GA-OH treatment concentration for the two cell lines with UMSCC47(+) in blue and UMSCC19(−) in orange. Both cell lines show increase of p21 expression during GA-OH treatment, especially in HPV(−) cell line.Fig. 7p21 Western Blot. This figure shows the western blot run with the HPV(+) cell line UMSCC47 and the HPV(−) cell line UMSCC19. The graph displays the Relative Density Unit (RDU) compared to the GA-OH treatment concentration for the two cell lines with UMSCC47(+) in blue and UMSCC19(−) in orange.Fig. 7

Discussion

4

In recent years, head and neck cancer has been on the rise and has become the most common HPV-associated cancer within the US [2,5]. With the negative side effects of the current treatment options, there is clearly a need to explore more effective alternatives to improve patients' quality of life [6,7]. Our research investigated the use of GA-OH as a small molecule inhibitor of E6 to restore cellular levels of p53, caspase 8, and other molecules that can induce apoptosis in cancer cells during cancer treatment.

In this study, we investigated the molecular mechanisms of action for GA-OH, which occur through activation of apoptotic pathways in both HPV(+) and HPV(−) cells. It is known that the original compound from which GA-OH was derived, Gambogic acid (GA), possesses anticancer activity [22], and our current findings confirm that the derivative, GA-OH, does as well. To better understand the pathways involved, we queried the mechanisms through which cells react to this treatment.

HPV infection can reduce the sensitivity of cells to apoptosis through the E6-enabled accelerated degradation of p53, procaspase 8 and other apoptotic molecules. Consistent with this working model, we found that there was increased caspase activity in all cell lines treated with GA-OH regardless of their HPV status (Fig. 5). It is known that the E6 and E7 oncoproteins are also able to target the Notch and Wnt signaling pathways to inhibit Notch signaling while simultaneously maintaining the activation of the Wnt pathway to both limit cellular differentiation and increase cellular proliferation [23].

RNAseq data determined that there is an abundance of DEGs in both the HPV(+) and HPV(−) cell lines (Fig. 1), with the number of DEGs identified in HPV(+) cells clearly fewer than those found in the HPV(−) cell lines. It is not currently clear if this reflects an actual difference in the number of genes that change their expression patterns during treatment, a difference in the magnitudes of gene expression change, or if it may be a technically induced artifact. The reduced number of DEGs found in HPV(+) cells, particularly UMSCC47, could suggest that they have a more stable gene expression profile under the treatment conditions when compared to the HPV(−) cells, or that they employ a different mechanism of response that involves fewer genes. A major limitation of our analysis is that it depends on arbitrary cutoffs for the selection of DEGs. To address this limitation, we supplemented our DEG analysis with gene set enrichment analysis [19]. The results of this analysis for each cell line, using the full set of DEGs comparing treatment vs. control, are given in Supplemental Tables 5, 6, 7, and 8. Some themes unique to HPV(−) cells in our initial analysis, such as cell cycle, showed up as being common to both HPV(+) and HPV(−) cells (Supplemental Table 9). This suggests that pathways such as cell cycle are indeed affected in both HPV(+) and HPV(−) cells, with a smaller magnitude change in HPV(+) cells either due to a diminished response or to technical artifacts in the data. Many pathways were unique to either HPV(−) cells or HPV(+) cells in the supplemental analysis, as well (Supplemental Tables 10 and 11, respectively).

It could be beneficial to further analyze these activated pathways to determine whether they are more efficient or are able to rely on fewer genes to achieve a similar outcome to their HPV(−) counterparts. In addition, it would be useful to explore whether the observed differences in gene expression translate to functional differences as a response to treatment with GA-OH through qPCR or various functional assays. While the original findings show a difference in the quantity of DEGs between HPV(+) and HPV(−) cell lines, more research is needed to elucidate the underlying reasons and potential implications of these differences. Nevertheless, analysis of DEGs identified in both types of cells showed that in reaction to GA-OH treatment they activated both overlapping and unique mechanisms.

Gene Ontology analysis showed that the DEGs involved in the response varied between two cell types. While HPV(−) cells reacted mostly through DEGs changes involved in chromosome reorganization (Fig. 3B), HPV(+) cells showed increased DEGs enrichment in response to protein folding (Fig. 3C). Comparison of enriched pathways found five common pathways that are affected in both cell types of cells (Table 1). These include the p53 signaling pathway, transcriptional misregulation in cancer, and cell cycle regulation, in addition to alcoholism and systemic lupus erythematosus. Genes that are affected in lupus represent histones: H2A, H2B, H3, and H4. Alcoholism pathways also include histone genes adding genes of two transcription factors, CREB and D FOB B proto-oncogene. Changes in histone genes expression lead to changes in chromatin structure to make damaged DNA more accessible initiating DNA damage response. Activation of pathways in response to DNA damage suggests that GA-OH is acting as a genotoxic agent. In the UMSCC104(+) cell line, we identified the MAPK signaling pathways as one of the major pathways activated during GA-OH treatment (Table 3). Within the SCC84(−) cell line, we identified the p53 pathway as one of the major pathways activated when treated with GA-OH (Table 2).

Increased expression of p21 protein in GA-OH treated cells confirmed our observation on transcriptome level. Western blot analysis shows in Fig. 7 expression of p21 in the UMSC47 and UMSCC19 cells demonstrating that the expression of p21 was increased according to the dose of GA-OH. The graph shows the Relative Density Unit (RDU), and the various concentrations of GA-OH used for both of the cell lines analyzed. There is a clear upward trend of p21 with an increase in GA-OH concentration.

In summary, we determined that GA-OH causes effects consistent with genotoxic activity against both HPV(+) and HPV(−) HNSCC cell lines, and that it induces cell death in both cell types via apoptosis (Fig. 4). We also determined that GA-OH activates both overlapping and differential pathways in the HPV(+) and HPV(−) cell lines. We anticipate that GA-OH will be primarily useful as paired with other cancer therapeutics, as it would sensitize HPV + cells to these other treatments due to its effects on various pathways. These various affected pathways could result in more effective treatment, as overlapping pathways could allow for more ways to achieve the same goal, while the differential pathways could allow the compound more ways to circumvent the cancer cell's survival methods. These findings suggest the potential of compounds related to GA and GA-OH as a treatment strategy for a wider range of patient populations. Understanding the mechanistic effects of GA-OH in HPV(+) and HPV(−) cell lines is the first step of many to potentially unlock an effective and novel combinational therapy for HNSCC patients.

CRediT authorship contribution statement

Danelle Grubbs: Writing – review & editing, Writing – original draft, Visualization, Investigation, Formal analysis, Data curation. Sonia Whang: Writing – review & editing, Visualization, Formal analysis, Data curation. Valeria Rodarte: Writing – review & editing, Investigation, Formal analysis, Data curation. Briza Martinez: Writing – review & editing, Investigation, Formal analysis, Data curation. Valeri Filippov: Writing – review & editing, Writing – original draft, Visualization, Validation, Formal analysis, Data curation. John Chen: Supervision, Software, Resources. Julia Unternaehrer: Software, Resources. Isaac Kremsky: Writing – review & editing, Formal analysis. Brigitte Vazquez: Investigation, Data curation. Penelope J. Duerksen-Hughes: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Funding acquisition, Conceptualization.

Funding

This project has been funded in whole or part by NIH: CA095461, NS073059, CA221479, CA209829.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.