Simple Summary
Hepatocellular carcinoma, a major type of liver cancer and a leading cause of cancer deaths worldwide, is often linked to poor outcomes due to aggressive cell growth, invasion, and spread. Researchers are exploring ways to slow or stop this aggressiveness using a special variant of the SIRT6 protein, which is found in some exceptionally long-lived people (centenarians). This variant (N308K/A313S) improves DNA repair and fights cancer in other contexts; thus, this study examined whether delivering it to liver cancer cells via a safe viral vector could reduce harmful behaviors compared to the normal SIRT6 version. The findings show that the centenarian variant more strongly curbs cell multiplication and invasion while increasing cell stiffness (a sign linked to less aggressive movement), and that it alters genes related to tissue structure and matrix remodeling. These results suggest that the variant could help suppress liver cancer progression, offering new insights for the research community into potential gene-based therapies for this deadly disease and highlighting SIRT6’s role in cancer control.
Keywords: SIRT6, transcriptomics, hepatocellular carcinoma, extracellular matrix, stiffness
Abstract
Background/Objectives: Hepatocellular carcinoma (HCC) is the sixth most prevalent cancer and a chief cause of cancer-related mortality throughout the world. SIRT6 is a fundamental sirtuin that governs several disease processes encompassing inflammation and cancer, including HCC. Longevity in centenarian Ashkenazi Jews was recently associated to novel allelic variants of SIRT6 (N308K/A313S), which ameliorate genome maintenance and DNA repair, and suppress cancer cells. It is currently unknown whether the above-mentioned SIRT6 variants display divergent or similar roles in HCC pathogenesis, compared to the wild-type (WT) counterpart. Methods: Our goal was to elucidate how these new centenarian-associated SIRT6 genetic variants may modulate HCC cell lines’ (HepG2 and Huh-7) aggressiveness and behavior, using functional and transcriptomic approaches. Results: We demonstrate that adeno-associated virus (AAV2/8)-mediated overexpression of centenarian-associated SIRT6 variants hampered HCC cell proliferation, with transcriptomic data showing the modulation of hallmark genes involved in the turnover of collagen/extracellular matrix (ECM). In addition, we found that AAV2/8-mediated overexpression of SIRT6 N308K/A313S decreased invasion and also increased stiffness in HCC cells, as measured by nanoindentation, in a more pronounced fashion compared to SIRT6 WT. Intracellular stiffness is a property of the cancer cells themselves, which, along with ECM invasiveness, plays a significant role in the progression of HCC. Conclusions: These data suggest that increased intracellular stiffening mirrors increased cell motility and invasive behavior; it can be indicative of suppressed cancer development and progression by the centenarian-associated SIRT6 N308K/A313S mutant.
1. Introduction
Hepatocellular carcinoma (HCC) is an important global health challenge, ranking as the sixth most prevalent neoplasia and a leading driver for cancer-related mortality in the world [1,2]. The incidence of HCC is increasing, not only in regions with high rates of hepatitis B and C infections, but also in areas witnessing a rise in metabolic syndrome and nonalcoholic fatty liver disease [2,3]. Despite advancements in prevention and treatment, HCC prognosis is dismal, primarily due to late diagnosis and the intricate interplay of risk factors and tumor biology [4].
The tumor microenvironment in HCC is marked by significant remodeling of the extracellular matrix (ECM), which provides structural support and actively influences tumor growth, invasion, and metastasis [5,6]. Chronic liver injury and cirrhosis lead to the accumulation and crosslinking of ECM proteins, such as collagen, resulting in increased tissue stiffness and the activation of signaling pathways that promote epithelial-to-mesenchymal transition (EMT), chemoresistance, and a more aggressive tumor phenotype [6,7]. HCC cells adapt to increased ECM stiffness by remodeling their cytoskeleton, thereby enhancing their migratory and invasive capabilities [8]. Importantly, these biomechanical alterations extend beyond the extracellular environment; intracellular stiffness and cytoskeletal tension are also influenced, affecting cell proliferation, migration, and resistance to apoptosis [8,9].
Sirtuin 6 (SIRT6), an NAD+-dependent histone deacetylase, plays a critical role in maintaining genomic stability, regulating metabolism, and modulating stress responses [10,11]. SIRT6 exerts its effects through the deacetylation of both histone and non-histone proteins, thereby influencing DNA repair, cellular metabolism, oxidative stress resistance, and inflammation in untransformed cells [11,12]. A loss of SIRT6 function is linked to genomic instability, metabolic dysfunction, and an increased risk of age-related diseases, including cancer [11,12]. In the context of HCC, SIRT6 exhibits a dual role. Its expression is often diminished in HCC tissues; this downregulation correlates with enhanced tumor growth, poor prognosis, and increased glycolytic metabolism [13,14]. Overexpression of SIRT6 in HCC cells has been shown to inhibit proliferation, induce apoptosis, and suppress tumorigenicity, primarily through the deacetylation of histone H3K9 and H3K56, as well as non-histone targets such as PKM2 [15].
However, SIRT6 can also facilitate EMT, metastasis, and resistance to apoptosis under certain conditions, underscoring the complexity of its role in HCC biology [16,17]. The regulation of SIRT6 in HCC encompasses multiple layers, including post-translational modifications, ubiquitin-mediated degradation, and microRNA-mediated repression [18,19]. Given its multifaceted functions, SIRT6 represents a promising yet complex target for therapeutic intervention in HCC. Further investigation is essential to clarify the context-specific effects of SIRT6 and to develop strategies that leverage its tumor-suppressive properties while mitigating potential pro-tumorigenic actions [20,21].
Recent studies have identified specific genetic variants in the SIRT6 gene that are associated with exceptional longevity and healthy aging in humans. Two such variants, rs183444295 (resulting in alanine-to-serine substitution at position 313, A313S) and rs201141490 (asparagine-to-lysine at position 308, N308K), are notably enriched among Ashkenazi Jewish centenarians [22]. These two rare missense variants, A313S and N308K, were observed in perfect linkage and together had nearly double the allele frequency among centenarians compared with controls and represent potential candidates for promoting extended lifespan and improved healthspan [22]. The prolonged lifespan observed in Ashkenazi Jewish populations is well documented, shows strong heritability, and is accompanied by a reduced prevalence of common age-related disorders [23,24]. Hepatocellular carcinoma (HCC) is a strongly age-dependent malignancy, with both incidence and risk rising sharply with advancing age and reaching a peak after age 60 [24]. Interestingly, the SIRT6 alleles found at a higher frequency in centenarians retain robust ADP-ribosylation activity that supports genomic integrity [22].
In our previous work, we demonstrated that lentiviral overexpression of the combined SIRT6 N308K/A313S variant in mature adipocytes improved glucose handling [25] and, in a three-dimensional hepatic spheroid system, suppressed collagen accumulation and the expression of fibrosis-related genes [26]. The latter study focused on non-cancer cell types, such as hepatocytes and hepatic stellate cells [26]; whereas HCC cells are mutated hepatocytes that have lost normal function, growing uncontrollably, and developing abnormal shapes and genetics. Therefore, the present study was designed to investigate the functional impact of this longevity-associated SIRT6 N308K/A313S variant on tumorigenic properties of HCC cells, employing both phenotypic assays and transcriptome-wide analyses.
2. Materials and Methods
2.1. Cell Lines and AAV2/8
HepG2 and Huh-7 liver cancer cells were obtained from CLS-GmbH (Eppelheim, Germany) and were authenticated using STR profiling. They were cultured in high glucose (4.5 g/L) DMEM (1×) supplemented with 10% fetal bovine serum (FBS) (Sigma-Aldrich, UK), 15 mM Hepes buffer (Biowest, Nuaille, France), glutamine, 1% penicillin/streptomycin solution, and 100 μg/mL Normocin (InvivoGen, San Diego, CA, USA) at 37 °C and 5% CO2. The cell culture medium was changed every 2 days; the cells were sub-cultured using TrypLE Express when reaching 90% confluence. AAV2/8-Luc (carrying luciferase only), AAV2/8-SIRT6-WT (carrying the wild-type sequence of human SIRT6) and AAV2/8-SIRT6-N308K/A313S were produced by GenFlow Biosciences, which patent-protected their pre-clinical and clinical purposes (PCT/EP2023/084840). Information about the cloning, production, and purification of AAV2/8 vectors can be provided upon reasonable request. For AAV2/8 viral transduction, cultured cells at 70–80% confluency were serum-deprived cells for 2 h before adding the virus. AAV2/8 stocks were thawed on ice and diluted with a serum-free medium to the target multiplicity of infection (MOI) (104 to 106 viral genomes). The old medium was replaced with the virus-containing medium and cells were incubated for different amounts of time until the downstream analyses. Luminescence in cells expressing AAV2/8-Luc (carrying luciferase only) was measured using the Bright-Glo™ Luciferase Assay System (Promega, Southampton, UK).
While lentiviruses (LV) have advantages, like large genetic capacity and suitability for ex vivo applications, adeno-associated viruses (AAVs) offer advantages over lentiviruses due to their low immunogenicity, minimal pathogenicity, ability to infect non-dividing cells, and suitability for eventual in vivo gene therapy applications. Considering potential clinical applications, we therefore conducted this study using AAV2/8, which combines elements of the AAV2 and AAV8 serotypes to create a highly effective tool for delivering genetic material to target cells, particularly the liver [27]. Therefore, we used AAV2/8-Luc (carrying luciferase only), AAV2/8-SIRT6-WT (carrying the wild-type sequence of human SIRT6) and AAV2/8-SIRT6-N308K/A313S (carrying the mutations of human centenarian-associated SIRT6). Multiplicity of infection (MOI), a critical factor in gene transfer, represents the ratio of viral particles to cells and plays a role in determining gene delivery efficiency and potential toxicity. Therefore, we transduced both HepG2 and Huh-7 liver cancer cells with AAV2/8-Luc at three different MOI: 106 viral genomes (vg) (MOI 1, high); 105 vg (MOI 2, intermediate); and 104 vg (MOI 3, low) (Figure 1). No toxicity was observed; in both cell lines, MOI 1 resulted in the best transduction efficiency, nearly doubling the transfection efficiency of MOI 2 at all time points. A peak of luminescence was observed at 48–72 h post transduction in HepG2, and at 48 h in Huh-7 (Figure 1).
Figure 1.
Titration of AAV2/8 Luc in HepG2 and Huh-7 liver cancer cells. Three multiplicities of infection (MOI) doses (104, 105, 106—viral genomes) were tested in HepG2 (upper panel) and in Huh-7 liver cancer cells (lower panel) at three time points (24 h, 48 h, 72 h). Untransduced cells (CTL) were used as negative controls. Data are presented as mean ± SEM, n = 3.
An MOI of 106 was used for all subsequent experiments using AAV2/8. qPCR analysis 48 h post transduction confirmed the upregulation of SIRT6c and SIRT6 WT mRNA levels, 48 h post transduction with AAV2/8-SIRT6-N308K/A313S or AAV2/8-SIRT6-WT, respectively, in both HepG2 and Huh-7 liver cancer cells (Appendix A Figure A1A,B). SIRT6 Cent mRNA was not detected in cells transduced with AAV2/8-Luc or with AAV2/8-SIRT6-WT (Appendix A Figure A1A,B). Conversely, both exogenous SIRT6 Cent (introduced upon transduction with AAV2/8-SIRT6-N308K/A313S—orange column) and endogenous SIRT6 (WT, the one already present in the cells, not introduced with viruses—blue column) were expressed in the cells transduced with AAV2/8-SIRT6-N308K/A313S, i.e., the “SIRT6 Cent” group (Appendix A Figure A1A,B).
2.2. qPCR
Quantitative real-time PCR (qRT-PCR) was carried out following previously published protocols [28,29]. In brief, total RNA was extracted from stimulated liver cancer cells using TRIzol Reagent (Invitrogen, San Diego, CA, USA), followed by reverse transcription and cDNA amplification in a single step using the SuperScript III One-Step RT-PCR System with Platinum Taq (Invitrogen, catalog no. 10928034). Quantification was performed on a StepOnePlus Real-Time PCR System (Qiagen, Manchester, UK). Custom-designed primer pairs specific for wild-type SIRT6 (SIRT6-WT), the centenarian-enriched mutant allele (SIRT6c), and the housekeeping gene GAPDH were used. The primer sequences, expected amplicon lengths, and annealing temperatures are listed below:
SIRT6-WT forward: 5′ AGG ATG TCG GTG AAT TAC GC 3′;
SIRT6-WT reverse: 5′ TTG GCA CAT TCT TCC ACA AA 3′;
SIRT6c forward: 5′ GCC ATC CAC GCT GTT TTG AC 3′;
SIRT6c reverse: 5′ GCC TCG CCA GTT CCC ACA C 3′;
GAPDH forward: 5′ ACA GTC CAT GCC ATC ACT GCC 3′;
GAPDH reverse: 5′ GCC TGC TTC ACC ACC TTC TTG 3′.
2.3. Viability and Invasion Assays
Huh7 and HepG2 liver cancer cells were seeded into 96-well plates at 1 × 104 cells per well and incubated for 24 or 48 h, as indicated. Cell viability was assessed using the MTT assay: cells were exposed to 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide solution for 4 h, after which the medium was aspirated, the formed formazan crystals were dissolved in isopropanol, and absorbance was read at 560 nm on a plate reader. For the invasion assay, a commercial 96-well basement membrane invasion kit with an 8 µm pore size (Abcam, Cambridge, UK; catalog no. ab235697) was employed according to the manufacturer’s protocol. Invasion of HepG2 and Huh7 liver cancer cells toward a chemoattractant gradient was stimulated using the medium supplemented with 10% fetal bovine serum in the lower chamber.
2.4. Nanoindentation Approaches
Nanoindentation was performed on cells using a PIUMA nanoindenter system (Optics11, Amsterdam, The Netherlands) loaded with a spherical tip with a radius of 10 µm and a spring constant of 0.025 N/m. The cells were cultured on a suitable 25 mm plastic Petri dish and prepared for the indentation experiments. Indentations were guided using a light microscope and were performed at three distinct locations on cells that were placed approximately 3 mm apart. The nanoindentation depth was set to 500 nm, which is less than 10% of the cell thickness. To minimize the effect of the substrate, an equal approach and retraction speed of 500 nm/s were chosen, and a maximum load of 500 nN was set, allowing for controlled indentation of the cells. The Hertz model was utilized for data analysis to determine the elastic modulus of the cells. Prior to indentation, the nanoindenter system was calibrated to ensure accurate measurements. The experiments were carried out at room temperature (21 °C), with cells immersed in borate-buffered saline (BBS) and tested within 15 min after careful replacement of the culture solution with filtered BBS.
2.5. RNA-Seq and Bioinformatic Analyses
Indexed RNA-seq libraries were prepared from 250 ng of polyA-selected mRNA using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (Ipswich, MA, USA). Equimolar pools of libraries were sequenced either on an Illumina MiSeq platform (single-end 75 bp reads) or on an Illumina NovaSeq 6000 (single-end 161 bp reads). Raw read quality was inspected with FastQC. Adapter sequences and low-quality bases were removed using Trimmomatic v0.36, with the parameters CROP:250, LEADING:3, TRAILING:3, SLIDINGWINDOW:4:5, MINLEN:35, and an adapter fasta containing the Illumina universal sequence (AGATCGGAAGA). Cleaned reads were mapped to the human reference genome GRCh38 (hg38) with STAR v2.7.10a in splice-aware mode, using Gencode v43 as the comprehensive gene annotation. STAR simultaneously produced a raw gene-count matrix.
Differential expression analysis was performed in R using the DESeq2 package. The false discovery rate (FDR) was used as the primary statistical criterion for defining differential expression (upregulation/downregulation). We employed the Benjamini–Hochberg procedure to adjust p-values for multiple testing; these adjusted p-values represent the FDR. We defined upregulated genes as those with an FDR (adjusted p-value) of 0.05 or less and a log2 fold-change greater than 0.5. Downregulated genes were defined as those with an FDR (adjusted p-value) of 0.05 or less and a log2 fold-change less than −0.5. Heatmaps of clustered expression profiles were drawn with the pheatmap package; volcano plots and MA plots were generated using ggplot2 and ggpubr. Pathway and functional enrichment of differentially expressed gene sets were evaluated with clusterProfiler. Reactome pathway enrichment was tested specifically with enrichPathway. Gene ontology (Biological Process, Molecular Function, Cellular Component) and KEGG pathway analyses were also conducted. Enrichment was considered significant at a Benjamini–Hochberg-adjusted p-value < 0.05.
2.6. Statistical Analyses
All statistical tests were carried out in GraphPad Prism v8.0 (GraphPad Software, La Jolla, CA, USA). When data in all compared groups followed a normal distribution (assessed by the Shapiro–Wilk test), two-group comparisons were performed with the unpaired two-tailed Student’s t test, and multiple-group comparisons with one-way ANOVA, followed by the recommended post hoc test. In cases of non-normal distribution, the Mann–Whitney U test (two groups) or Kruskal–Wallis test (multiple groups) was applied. Results are presented as mean ± standard error of the mean (SEM), unless stated otherwise. Statistical significance is reported as * p < 0.05, ** p < 0.01, *** p < 0.001, or as otherwise specified.
3. Results
3.1. SIRT6 N308K/A313S Mutant Hampers HCC Cell Viability: Whole Transcriptome Analysis
In our earlier work, we demonstrated that lentiviral overexpression of either wild-type SIRT6 or the centenarian-associated double mutant N308K/A313S in HepG2 and Huh7 liver cancer cells caused a dramatic loss of cell viability (60–90%) within two weeks post transduction [26]. In the present study, we monitored viability kinetics at multiple time points (0, 24, 48, 72, 96, and 120 h) after transduction of HepG2 and Huh7 cells with adeno-associated viral vectors (AAV2/8) encoding luciferase (control), SIRT6-WT, or SIRT6-N308K/A313S (MOI 106). MTT assays showed that both SIRT6-WT and the centenarian mutant induced comparable, time-dependent suppression of viability in the two liver cancer cell lines; the effect became evident by 72 h and dropped below baseline levels by 120 h (Figure 2A,B). These results reinforce the notion that elevated SIRT6 activity can exert selective cytotoxic effects on malignant hepatocytes [30].
Figure 2.
MTT viability assays: MTT assay on HepG2 (A); and Huh-7 liver cancer cell line (B) transduced with AAV2/8-Luc, AAV2/8-SIRT6-WT and AAV2/8-SIRT6-N308K/A313S for up to five days (0 h, 24 h, 48 h, 72 h, 96 h, 120 h). Cell survivals were significantly lower in AAV2/8-SIRT6-WT and AAV2/8-SIRT6-N308K/A313S-transduced cells from AAV2-8-Luc at the indicated time points, with a p < 0.05 (*), p < 0.01 (**) or p < 0.001 (***). Data are presented as mean ± SEM, n = 3.
Given that the longevity-associated N308K/A313S variant might produce unique epigenetic signatures capable of altering transcriptional programs, we investigated whether AAV2/8-mediated overexpression of SIRT6-WT versus the centenarian mutant elicited distinct gene-expression profiles in HepG2 and Huh7 liver cancer cells, and how each differed from the luciferase control. For each cell line, three experimental conditions were established: MOI 106: AAV2/8-Luc (Luc); AAV2/8-SIRT6-WT (WT); and AAV2/8-SIRT6-N308K/A313S (Cent). To capture transcriptional changes before the onset of widespread cell death (Figure 2), total RNA was harvested at 72 h post transduction for RNA sequencing. Principal component analysis (PCA) of the transcriptome data revealed clear separation among the three conditions in both cell lines (Figure 3A,B), although one replicate from the Cent group in Huh7 clustered separately (Figure 3B).
Figure 3.
Global transcriptomic profiling in hepatocellular carcinoma cells: two-dimensional principal component analysis (PCA) plots derived from variance-stabilized expression values (DESeq2) for HepG2 (A); and Huh7 (B) liver cancer cells. Each point represents an individual biological replicate, illustrating clear separation of samples according to experimental condition (Luc control, SIRT6-WT, and SIRT6-Cent), reflecting distinct overall gene-expression signatures.
Pairwise differential expression analysis (adjusted p-value ≤ 0.05 and |log2 fold change| ≥ 0.5) yielded the following numbers of differentially expressed genes (DEGs) in HepG2 cells:
WT vs. Luc: 950 DEGs (330 up, 620 down);
Cent vs. Luc: 810 DEGs (244 up, 566 down);
Cent vs. WT: 545 DEGs (256 up, 289 down) (Figure 4A).
Figure 4.
Differential expression and pathway analysis in HepG2 liver cancer cells: (A) volcano plots displaying differentially expressed genes across the three pairwise comparisons (SIRT6-WT vs. Luc, SIRT6-Cent vs. Luc, and SIRT6-Cent vs. SIRT6-WT). The horizontal axis shows log2 fold-change, and the vertical axis indicates −log10(adjusted p-value). Upregulated genes are shown in blue, downregulated genes in red; and (B) reactome pathway enrichment results for the same comparisons. The bubble size corresponds to the number of differentially expressed genes contributing to each pathway, while color intensity reflects the degree of statistical enrichment.
Reactome pathway enrichment performed on these DEG sets in HepG2 cells highlighted distinct biological themes: steroid metabolism was the top-enriched pathway for WT vs. Luc, extracellular matrix proteoglycans dominated the Cent vs. Luc comparison, and general extracellular matrix organization emerged as the leading pathway, followed by regulation of insulin-like growth factor 1 (IGF-1), when comparing the centenarian mutant directly to wild-type SIRT6 (Figure 4B).
Next, we focused on the Huh-7 liver cancer cells. With the same thresholds used for the HepG2 liver cancer cell line, the comparison between WT versus Luc identified only 48 DEGs, of which 38 were up-regulated and 10 down-regulated (Figure 5A, left); comparing Cent versus Luc, we identified 611 DEGs, of which 218 were up-regulated and 393down-regulated (Figure 5A, center); comparing Cent versus WT identified 456 DEGs, of which 146 were up-regulated and 310 down-regulated (Figure 5A, right). Interestingly, enrichment analysis using the Reactome Pathway Database with the same filters as those used for HepG2 liver cancer cells, identified the same top pathway involved in all comparisons (WT versus Luc, Cent versus Luc, Cent versus WT): Extracellular matrix organization, followed by similar processes (extracellular structure organization, ECM proteoglycans, degradation of the extracellular matrix) (Figure 5B). The lists of DEGs for HepG2 and Huh-7 liver cancer cells, relative to the volcano plots in Figure 4A and Figure 5A, are reported in Tables S1 and S2, respectively. The lists of DEGs for HepG2 and Huh-7 liver cancer cells, relative to the reactome pathway analyses (three top pathways; Figure 4B and Figure 5B), are reported in Tables S3 and S4, respectively. Overall, the transcriptomic data suggest that SIRT6, WT or centenarian-associated proteins, may play a role in liver cancer cells viability and turnover of collagen/ECM, which in turn are hallmarks of invasion and aggressiveness.
Figure 5.
Differential expression and pathway analysis in Huh7 liver cancer cells: (A) volcano plots of differentially expressed genes for the pairwise comparisons (SIRT6-WT vs. Luc, SIRT6-Cent vs. Luc, and SIRT6-Cent vs. SIRT6-WT), plotted as log2 fold-change versus −log10(adjusted p-value). Upregulated and downregulated genes are colored blue and red, respectively; and (B) top enriched reactome pathways identified for each comparison. Bubble size indicates gene count; color gradient represents enrichment significance.
3.2. SIRT6 N308K/A313S Mutant Decreases Invasion and Increases Stiffness in HCC Cells
We then investigated how overexpression of wild-type SIRT6 (WT) or the centenarian-associated double mutant N308K/A313S (Cent) affected the invasive capacity of HepG2 and Huh7 liver cancer cells, using AAV2/8 vectors with luciferase (Luc) as the control. Invasion was assessed in a basement membrane-coated Boyden chamber system, in which cells degrade the extracellular matrix, migrate across an 8-µm pore membrane, and move toward a chemoattractant gradient. Following established protocols [31,32], 10% fetal bovine serum (FBS) served as the chemoattractant for both cell lines. As illustrated in Figure 6, FBS-induced invasion was markedly suppressed in cells overexpressing either SIRT6-WT or SIRT6-Cent compared with the Luc control after 24 h.
Figure 6.
Invasive capacity of HCC cells after SIRT6 overexpression: invasion through basement membrane-coated inserts was evaluated in HepG2 (A); and Huh7 (B) liver cancer cells 48 h after transduction with AAV2/8-Luc (control), AAV2/8-SIRT6-WT, or AAV2/8-SIRT6-N308K/A313S, followed by a 24 h migration period toward 10% FBS as chemoattractant. White bars: no chemoattractant (negative control); black bars: migration inducer provided with the kit (positive control). Both SIRT6-WT and the centenarian-associated SIRT6 variant dramatically reduced cell invasion compared with the Luc control. Data are presented as mean ± SEM, n = 3. Data are presented as mean ± SEM, n = 3. Statistical significance: * p < 0.05, ** p < 0.01, *** p < 0.001 versus AAV2/8-Luc; $$$ p < 0.001 versus AAV2/8-SIRT6-WT.
Notably, the centenarian variant (Cent) produced a significantly stronger anti-invasive effect than SIRT6-WT in both HepG2 (A) and Huh7 (B) liver cancer cells. Because the biomechanical properties of cancer cells themselves contribute critically to HCC progression—where softer tumor cells typically exhibit enhanced motility and invasiveness [33]—we next quantified intracellular mechanical stiffness using a fiber-optic nanoindentation platform. Measurements performed on HepG2 (Figure 7A) and Huh7 (Figure 7B) liver cancer cells revealed that overexpression of the centenarian-associated SIRT6 mutant dramatically increased cellular stiffness relative to both the Luc control and SIRT6-WT conditions.
Figure 7.
Nanoindentation assays in: HepG2 (A); and Huh-7 liver cancer cell lines (B) transduced with AAV2/8-Luc, AAV2/8-SIRT6-WT and AAV2/8-SIRT6-N308K/A313S for two days prior to incubation in the Boyden chamber for a further 24 h; 10% FBS was used as chemoattractant. Cell stiffness (measured in Pascal, Pa) was significantly higher in AAV2/8- AAV2/8-SIRT6-N308K/A313S-transduced cells compared to AAV2/8-Luc and to AAV2/8-SIRT6-WT. p < 0.05 (*), p < 0.01 (**) versus AAV2/8-SIRT6-WT.
These findings indicate that the pronounced stiffening induced by the longevity-linked N308K/A313S variant correlates with its superior suppression of migratory and invasive phenotypes, supporting a potential protective role against malignant progression.
4. Discussion
The present study elucidates the anti-tumorigenic effects of overexpressing the recently described centenarian-associated SIRT6 variant (N308K/A313S) in liver cancer cell lines, HepG2 and Huh-7, using AAV2/8-mediated transduction. Our findings demonstrate that both wild-type (WT) SIRT6 and the N308K/A313S mutant significantly impair cell proliferation, reduce invasive capacity, and enhance intracellular stiffness, with the mutant exhibiting more pronounced effects in invasion and stiffness assays. Transcriptomic analyses further reveal differential gene expression profiles, particularly in pathways related to ECM organization and metabolism, underscoring the potential of this variant to modulate HCC aggressiveness. ECM components related to the inhibition of HCC aggressiveness include, among others, decorin, laminin-332, collagen IV and fibronectin [34]; which ECM component is specifically under the control of SIRT6 enzymatic or genomic activities remains to be elucidated. These results align with the established tumor-suppressive functions of SIRT6 in HCC while highlighting the enhanced efficacy of longevity-associated variants, offering insights into their therapeutic potential for age-related cancers such as HCC. Our proliferation assays showed a time-dependent reduction in cell viability upon SIRT6 overexpression, which became evident at 72 h post transduction and reached sub-basal levels by 120 h. This is consistent with prior reports, where SIRT6 overexpression inhibits HCC cell growth through mechanisms such as apoptosis induction and cell cycle arrest [13,15]. The selective toxicity observed in cancer cells, as previously noted [30], may stem from SIRT6’s role in regulating glycolytic metabolism via deacetylation of targets, such as PKM2, shifting cells away from Warburg-effect dependency [21]. Notably, the N308K/A313S variant, enriched in centenarians and linked to enhanced ADP-ribosylation activity [22], did not differ significantly from the WT in proliferation inhibition but outperformed it in other phenotypes. This suggests that, while both forms suppress growth, the variant’s mutations may confer additional epigenetic or enzymatic advantages, potentially amplifying genome stability and stress resistance in HCC contexts [11]. Transcriptomic profiling via RNA-Seq revealed distinct clustering of overexpression groups in PCA, with hundreds of differentially expressed genes (DEGs) identified across comparisons. In HepG2 liver cancer cells, WT overexpression versus control yielded 950 DEGs that were predominantly downregulated, enriching the steroid metabolism pathway. In contrast, the N308K/A313S variant enriched for ECM proteoglycans and organization mirrored a pattern in Huh-7 liver cancer cells, where ECM-related pathways dominated all comparisons. These findings implicate SIRT6 in ECM remodeling, a critical driver of HCC invasion and metastasis [5,6]. Downregulation of ECM genes may counteract the stiff, fibrotic microenvironment typical of HCC, which promotes epithelial-to-mesenchymal transition (EMT) and chemoresistance [7,8]. The variant’s stronger impact on ECM pathways aligns with our prior work showing its antifibrotic effects in hepatic models [26], suggesting that it may disrupt collagen deposition and tumor-stroma interactions more effectively than WT. This is particularly relevant given HCC’s association with cirrhosis and metabolic syndrome, where ECM stiffness exacerbates tumor progression [2,3].
The invasion assays demonstrated a marked reduction in migratory capacity upon SIRT6 overexpression, with the N308K/A313S variant inducing significantly greater inhibition compared to WT in both cell lines. This anti-invasive effect corroborates studies where SIRT6 suppresses EMT and metastasis in HCC, potentially via the deacetylation of histone H3K9/56 or non-histone targets [16,17]. However, SIRT6’s role, tumor-suppressive in some contexts but protumorigenic in others (e.g., facilitating EMT under stress), highlights the need for cell type-specific modulation [12,20]. Our data suggest that the centenarian variant tips the balance toward suppression, possibly due to the enhanced regulatory functions observed in longevity cohorts [23,35]. Complementing this, nanoindentation revealed increased intracellular stiffness with variant overexpression, inverting the typical softening associated with invasive HCC cells [33]. Stiffer cells may resist the deformation required for migration, linking biomechanical changes to reduced aggressiveness [8,9]. This increase in stiffness could reflect cytoskeletal reorganization, aligning with SIRT6’s influence on genomic stability and metabolism [10].
5. Limitations
Study limitations include the in vitro focus; in vivo validation in animal models is needed to assess the systemic effects and immune responses. Additionally, while RNA-Seq provided comprehensive insights, the validation of key DEGs and an exploration of posttranslational mechanisms [18,19] would strengthen the conclusions. The context-dependent role of SIRT6 also warrants caution, as overexpression might exacerbate protumor effects in certain microenvironments [6].
6. Conclusions
Centenarian-associated SIRT6 N308K/A313S inhibited the aggressiveness and invasion of HCC cells. Clinically, these results position the N308K/A313S variant as a promising gene therapy target for HCC, especially given AAV2/8’s liver tropism and low immunogenicity [30]. Age-related incidence of HCC [25] and poor prognosis [1] underscore the value of longevity-inspired interventions. Our prior findings on the benefits of this variant in glucose sensitivity and fibrosis [26] extend to antitumor effects, potentially addressing metabolic drivers of HCC [14].
Future studies should investigate the variant’s impact on HCC stemness, chemoresistance, and patient-derived models, potentially integrating it with existing therapies. Targeting SIRT6 regulators, like microRNAs [12], or developing variant-specific activators could also enhance precision. Overall, this work advances understanding of SIRT6’s context-dependent roles in HCC and supports longevity-associated variants as novel therapeutic avenues to combat this malignancy.
Acknowledgments
We thank Bruna Menegassi and Sant’Agata for their support.
Abbreviations
The following abbreviations are used in this manuscript:
| AAV | Adeno-Associated Virus |
| AAV2/8 | Adeno-Associated Virus serotype 2/8 |
| ADP | Adenosine Diphosphate |
| ANOVA | Analysis of Variance |
| BBS | Borate Buffered Saline |
| cDNA | Complementary DNA |
| DEG | Differentially Expressed Gene |
| DMEM | Dulbecco’s Modified Eagle Medium |
| ECM | Extracellular Matrix |
| EMT | Epithelial-to-Mesenchymal Transition |
| FBS | Fetal Bovine Serum |
| GO | Gene Ontology |
| HCC | Hepatocellular Carcinoma |
| KEGG | Kyoto Encyclopedia of Genes and Genomes |
| LV | Lentivirus |
| MA | Mean Average (plot type) |
| mRNA | Messenger RNA |
| MOI | Multiplicity of Infection |
| MTT | 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide |
| NAD+ | Nicotinamide Adenine Dinucleotide (oxidized form) |
| OE | Overexpression |
| PCA | Principal Component Analysis |
| PKM2 | Pyruvate Kinase M2 |
| qPCR | Quantitative Polymerase Chain Reaction |
| RNA-Seq | RNA Sequencing |
| RT-PCR | Reverse Transcription Polymerase Chain Reaction |
| SEM | Standard Error of the Mean |
| SIRT6 | Sirtuin 6 |
| STAR | Spliced Transcripts Alignment to a Reference |
| vg | Viral Genome |
| WT | Wild Type |
Supplementary Materials
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cancers18050812/s1, Table S1. The lists of DEGs for HepG2 cells (as in Figure 4A); Table S2. The lists of DEGs for Huh-7 cells (as in Figure 5A); Table S3. The lists of DEGs for HepG2 cells (as in Figure 4B); Table S4. The lists of DEGs for Huh-7 cells (as in Figure 5B).
Appendix A
Figure A1.
mRNA expression of SIRT6 WT or SIRT6 Cent in HepG2 (A) and in Huh-7 (B) liver cancer cells. Cells were transduced with AAV2/8-Luc, AAV2/8-SIRT6-WT and AAV2/8-SIRT6-N308K/A313S for 48 h. mRNA levels of SIRT6 WT and SIRT6c were assessed by qPCR. Data are presented as mean ± SEM, n = 3. p < 0.001 (***) versus AAV2/8-Luc and AAV2/8-SIRT6-N308K/A313S-transduced cells.
Author Contributions
M.V. (Manlio Vinciguerra) designed the study; M.V. (Maanya Vittal) and A.K. performed the research; M.V. (Maanya Vittal) and A.K. collected data; M.V. (Maanya Vittal), N.L., A.K. and M.V. (Manlio Vinciguerra) analyzed data; E.L., R.A., T.M. and M.V. (Manlio Vinciguerra) provided supervision and intellectual input; M.V. (Manlio Vinciguerra) wrote the paper. All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request. The RNA-seq data have been deposited in Gene Expression Omnibus (GEO) with the dataset identifier GSE314110.
Conflicts of Interest
Author Eric Leire was employed by the companies GenFlow Biosciences SRL Biopark Gosselies and Clinique 135, Clinic of Plastic and Esthetic Surgery. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Funding Statement
This work was supported by GenFlow Bioscience, and by the European Commission Horizon 2020 Framework Program (Project 856871—TRANSTEM, to MVin).
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
References
- 1.Konyn P., Ahmed A., Kim D. Current epidemiology in hepatocellular carcinoma. Expert Rev. Gastroenterol. Hepatol. 2021;15:1295–1307. doi: 10.1080/17474124.2021.1991792. [DOI] [PubMed] [Google Scholar]
- 2.Jiang Z., Zeng G., Dai H., Bian Y., Wang L., Cao W., Yang J. Global, regional and national burden of liver cancer 1990-2021: A systematic analysis of the global burden of disease study 2021. BMC Public Health. 2025;25:931. doi: 10.1186/s12889-025-22026-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ogasawara S., Koroki K., Kanzaki H., Kobayashi K., Kiyono S., Nakamura M., Kanogawa N., Saito T., Kondo T., Nakagawa R., et al. Changes in therapeutic options for hepatocellular carcinoma in Asia. Liver Int. 2022;42:2055–2066. doi: 10.1111/liv.15101. [DOI] [PubMed] [Google Scholar]
- 4.Zhou T.C., Qin D., Yan P., Lan B., Wang Q., Tan J. Primary liver cancer burden and its association with health development in the Western Pacific, 1990–2021. Front. Oncol. 2025;15:1627330. doi: 10.3389/fonc.2025.1627330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Galasso L., Cerrito L., Maccauro V., Termite F., Ainora M.E., Gasbarrini A., Zocco M.A. Hepatocellular Carcinoma and the Multifaceted Relationship with Its Microenvironment: Attacking the Hepatocellular Carcinoma Defensive Fortress. Cancers. 2024;16:1837. doi: 10.3390/cancers16101837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Calitz C., Rosenquist J., Degerstedt O., Khaled J., Kopsida M., Fryknäs M., Lennernäs H., Samanta A., Heindryckx F. Influence of extracellular matrix composition on tumour cell behaviour in a biomimetic in vitro model for hepatocellular carcinoma. Sci. Rep. 2023;13:748. doi: 10.1038/s41598-023-27997-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Miyauchi Y., Yasuchika K., Fukumitsu K., Ishii T., Ogiso S., Minami T., Kojima H., Yamaoka R., Katayama H., Kawai T., et al. A novel three-dimensional culture system maintaining the physiological extracellular matrix of fibrotic model livers accelerates progression of hepatocellular carcinoma cells. Sci. Rep. 2017;7:9827. doi: 10.1038/s41598-017-09391-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wang Z., Wang W., Luo Q., Song G. High matrix stiffness accelerates migration of hepatocellular carcinoma cells through the integrin β1-Plectin-F-actin axis. BMC Biol. 2025;23:8. doi: 10.1186/s12915-025-02113-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Schrader J., Gordon-Walker T.T., Aucott R.L., Van Deemter M., Quaas A., Walsh S., Benten D., Forbes S.J., Wells R.G., Iredale J.P. Matrix stiffness modulates proliferation, chemotherapeutic response, and dormancy in hepatocellular carcinoma cells. Hepatology. 2011;53:1192–1205. doi: 10.1002/hep.24108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Collins J.A., Kim C.J., Coleman A., Little A., Perez M.M., Clarke E.J., Diekman B., Peffers M.J., Chubinskaya S., Tomlinson R.E., et al. Cartilage-specific Sirt6 deficiency represses IGF-1 and enhances osteoarthritis severity in mice. Ann. Rheum. Dis. 2023;82:1464–1473. doi: 10.1136/ard-2023-224385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Watroba M., Szukiewicz D. Sirtuins at the Service of Healthy Longevity. Front. Physiol. 2021;12:724506. doi: 10.3389/fphys.2021.724506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kim H.G., Huang M., Xin Y., Zhang Y., Zhang X., Wang G., Liu S., Wan J., Ahmadi A.R., Sun Z., et al. The epigenetic regulator SIRT6 protects the liver from alcohol-induced tissue injury by reducing oxidative stress in mice. J. Hepatol. 2019;71:960–969. doi: 10.1016/j.jhep.2019.06.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zhang Z.G., Qin C.Y. Sirt6 suppresses hepatocellular carcinoma cell growth via inhibiting the extracellular signal-regulated kinase signaling pathway. Mol. Med. Rep. 2014;9:882–888. doi: 10.3892/mmr.2013.1879. [DOI] [PubMed] [Google Scholar]
- 14.de Céu Teixeira M., Sanchez-Lopez E., Espina M., Garcia M.L., Durazzo A., Lucarini M., Novellino E., Souto S.B., Santini A., Souto E.B. Sirtuins and SIRT6 in Carcinogenesis and in Diet. Int. J. Mol. Sci. 2019;20:4945. doi: 10.3390/ijms20194945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bhardwaj A., Das S. SIRT6 deacetylates PKM2 to suppress its nuclear localization and oncogenic functions. Proc. Natl. Acad. Sci. USA. 2016;113:E538–E547. doi: 10.1073/pnas.1520045113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Han L.L., Jia L., Wu F., Huang C. Sirtuin6 (SIRT6) Promotes the EMT of Hepatocellular Carcinoma by Stimulating Autophagic Degradation of E-Cadherin. Mol. Cancer Res. 2019;17:2267–2280. doi: 10.1158/1541-7786.MCR-19-0321. [DOI] [PubMed] [Google Scholar]
- 17.Zhang C., Yu Y., Huang Q., Tang K. SIRT6 regulates the proliferation and apoptosis of hepatocellular carcinoma via the ERK1/2 signaling pathway. Mol. Med. Rep. 2019;20:1575–1582. doi: 10.3892/mmr.2019.10398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Elhanati S., Ben-Hamo R., Kanfi Y., Varvak A., Glazz R., Lerrer B., Efroni S., Cohen H.Y. Reciprocal Regulation between SIRT6 and miR-122 Controls Liver Metabolism and Predicts Hepatocarcinoma Prognosis. Cell Rep. 2016;14:234–242. doi: 10.1016/j.celrep.2015.12.023. [DOI] [PubMed] [Google Scholar]
- 19.Kohli S., Bhardwaj A., Kumari R., Das S. SIRT6 Is a Target of Regulation by UBE3A That Contributes to Liver Tumorigenesis in an ANXA2-Dependent Manner. Cancer Res. 2018;78:645–658. doi: 10.1158/0008-5472.CAN-17-1692. [DOI] [PubMed] [Google Scholar]
- 20.Lee N., Ryu H.G., Kwon J.-H., Kim D.-K., Kim S.R., Wang H.J., Kim K.-T., Choi K.Y. SIRT6 Depletion Suppresses Tumor Growth by Promoting Cellular Senescence Induced by DNA Damage in HCC. PLoS ONE. 2016;11:e0165835. doi: 10.1371/journal.pone.0165835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Huang Z., Zhao J., Deng W., Chen Y., Shang J., Song K., Zhang L., Wang C., Lu S., Yang X., et al. Identification of a cellularly active SIRT6 allosteric activator. Nat. Chem. Biol. 2018;14:1118–1126. doi: 10.1038/s41589-018-0150-0. [DOI] [PubMed] [Google Scholar]
- 22.Simon M., Yang J., Gigas J., Earley E.J., Hillpot E., Zhang L., Zagorulya M., Tombline G., Gilbert M., Yuen S.L., et al. A rare human centenarian variant of SIRT6 enhances genome stability and interaction with Lamin A. EMBO J. 2022;41:EMBJ2021110393. doi: 10.15252/embj.2021110393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Atzmon G., Schechter C., Greiner W., Davidson D., Rennert G., Barzilai N. Clinical phenotype of families with longevity. J. Am. Geriatr. Soc. 2004;52:274–277. doi: 10.1111/j.1532-5415.2004.52068.x. [DOI] [PubMed] [Google Scholar]
- 24.Sheedfar F., Di Biase S., Koonen D., Vinciguerra M. Liver diseases and aging: Friends or foes? Aging Cell. 2013;12:950–954. doi: 10.1111/acel.12128. [DOI] [PubMed] [Google Scholar]
- 25.Frohlich J., Liorni N., Mangoni M., Lochmanová G., Pírek P., Kaštánková N., Pata P., Kucera J., Chaldakov G.N., Tonchev A.B., et al. Epigenetic and transcriptional control of adipocyte function by centenarian-associated SIRT6 N308K/A313S mutant. Clin. Epigenetics. 2024;16:96. doi: 10.1186/s13148-024-01710-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Frohlich J., Raffaele M., Skalova H., Leire E., Pata I., Pata P., Gorbunova V., Vinciguerra M. Human centenarian-associated SIRT6 mutants modulate hepatocyte metabolism and collagen deposition in multilineage hepatic 3D spheroids. Geroscience. 2023;45:1177–1196. doi: 10.1007/s11357-022-00713-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ferla R., Alliegro M., Marteau J.-B., Dell’aNno M., Nusco E., Pouillot S., Galimberti S., Valsecchi M.G., Zuliani V., Auricchio A. Non-clinical Safety and Efficacy of an AAV2/8 Vector Administered Intravenously for Treatment of Mucopolysaccharidosis Type VI. Mol. Ther. Methods Clin. Dev. 2017;6:143–158. doi: 10.1016/j.omtm.2017.07.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Demolli S., Doebele C., Doddaballapur A., Lang V., Fisslthaler B., Chavakis E., Vinciguerra M., Sciacca S., Henschler R., Hecker M., et al. MicroRNA-30 mediates anti-inflammatory effects of shear stress and KLF2 via repression of angiopoietin 2. J. Mol. Cell Cardiol. 2015;88:111–119. doi: 10.1016/j.yjmcc.2015.10.009. [DOI] [PubMed] [Google Scholar]
- 29.McKee C., Sigala B., Soeda J., Mouralidarane A., Morgan M., Mazzoccoli G., Rappa F., Cappello F., Cabibi D., Pazienza V., et al. Amphiregulin activates human hepatic stellate cells and is upregulated in non alcoholic steatohepatitis. Sci. Rep. 2015;5:8812. doi: 10.1038/srep08812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Van Meter M., Mao Z., Gorbunova V., Seluanov A. SIRT6 overexpression induces massive apoptosis in cancer cells but not in normal cells. Cell Cycle. 2011;10:3153–3158. doi: 10.4161/cc.10.18.17435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Xia H., Cao J., Li Q., Lv Y., Jia W., Ren W., Cheng Q., Song X., Xu G. Hepatocellular Carcinoma-propagating Cells are Detectable by Side Population Analysis and Possess an Expression Profile Reflective of a Primitive Origin. Sci. Rep. 2016;6:34856. doi: 10.1038/srep34856. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Gao H., Yin F., Guan D., Feng Y., Zheng Q., Wang X., Zhu M., Zhang X., Cheng S., Chen T., et al. Liver cancer: WISP3 suppresses hepatocellular carcinoma progression by negative regulation of β-catenin/TCF/LEF signalling. Cell Prolif. 2019;52:e12583. doi: 10.1111/cpr.12583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wang J., Zhang B., Chen X., Xin Y., Li K., Zhang C., Tang K., Tan Y. Cell mechanics regulate the migration and invasion of hepatocellular carcinoma cells via JNK signaling. Acta Biomater. 2024;176:321–333. doi: 10.1016/j.actbio.2024.01.024. [DOI] [PubMed] [Google Scholar]
- 34.Roy A.M., Iyer R., Chakraborty S. The extracellular matrix in hepatocellular carcinoma: Mechanisms and therapeutic vulnerability. Cell Rep. Med. 2023;4:101170. doi: 10.1016/j.xcrm.2023.101170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Atzmon G., Cho M., Cawthon R.M., Budagov T., Katz M., Yang X., Siegel G., Bergman A., Huffman D.M., Schechter C.B., et al. Genetic variation in human telomerase is associated with telomere length in Ashkenazi centenarians. Proc. Natl. Acad. Sci. USA. 2010;107:1710–1717. doi: 10.1073/pnas.0906191106. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request. The RNA-seq data have been deposited in Gene Expression Omnibus (GEO) with the dataset identifier GSE314110.