Targeting the m⁶A RNA demethylase FTO enhances UVB-induced DNA damage repair and suppresses skin tumor growth

Abstract

Nonmelanoma skin cancers are rising in incidence and are largely driven by solar ultraviolet B radiation (UVB) exposure. A growing body of evidence suggests that m⁶A RNA methylation plays a critical role in regulating the DNA damage response to UVB. Here, we identify a novel function for the m⁶A demethylase FTO in modulating the UVB damage response and skin carcinogenesis. We show that FTO loss enhances the repair of cyclobutane pyrimidine dimers (CPD), the major DNA lesions induced by UV radiation, in a METTL14-dependent manner, at least in part by promoting protein translation of the global genome repair (GGR) factor DDB2 through increased m⁶A methylation of DDB2 mRNA. These effects were recapitulated using two small-molecule FTO inhibitors, CS1 and FB23–2. Furthermore, loss of FTO reduced tumor growth in mice and FTO expression was upregulated in cutaneous squamous cell carcinoma (cSCC) compared with normal skin. Together, these findings uncover a critical role for FTO in regulating post-transcriptional gene expression in the UVB damage response and suggest that FTO may be a therapeutic target in skin cancer.

INTRODUCTION

Nonmelanoma skin cancers (NMSC), including cutaneous squamous cell carcinoma (cSCC) and basal cell carcinoma (BCC), are among the most common malignancies worldwide.¹ In the United States, cSCC represents a major public health burden, contributing to substantial annual healthcare costs.² Although the majority of cSCC cases are curable through surgery, a subset progresses to advanced or metastatic disease that is no longer amenable to resection.³ Furthermore, therapeutic options for advanced or metastatic cSCC remain limited. The anti-PD-1 immunotherapy cemiplimab, while therapeutically beneficial, shows clinical benefit in about 50% of patients.⁴ It is estimated that cSCC contributes to ~4000–8000 deaths annually in the US.⁵ Thus, there is a need for new targeted therapies for patients with cSCC.

Given that ~90% of cSCC cases are attributed to ultraviolet (UV) radiation exposure, understanding the cellular response to UV is critical for identifying novel therapeutic targets.⁶ Ultraviolet B (UVB) is the major carcinogenic component of UV light and induces DNA damage through lesions including cyclobutane pyrimidine dimers (CPDs) and 6–4 photoproducts (6–4PPs).^7–9 These lesions are repaired through the nucleotide excision repair (NER) pathway, including global genome repair (GGR).^7–9 GGR is carried out by the xeroderma pigmentosum (XP) protein family and involves recognition of the DNA lesions by XPC and DDB2, local unwinding of the helix by XPB and XPD, and cleavage of the damage site by XPF and XPG.^7–9 Importantly, the UVB damage response is highly complex, and many of the underlying mechanisms remain incompletely understood.

Recent work has indicated that m⁶A RNA methylation and its regulatory enzymes play critical roles in carcino-genesis, including the UVB damage response and cSCC pathogenesis.¹⁰ m⁶A RNA methylation is a reversible, post-transcriptional RNA modification that occurs on various RNA species, including mRNA, tRNA, rRNA, and noncoding RNAs, and is critical for the regulation of gene expression.¹¹ m⁶A modification is dynamically regulated by writers including METTL3 and its cofactor METTL14, erasers including FTO and ALKBH5, and readers including YTHDF1 and YTHDC2, which add, remove, and detect the modification, respectively.^11,12

We previously showed that in response to UVB exposure, the m⁶A writer METTL14, in coordination with reader YTHDF1, facilitates GGR in keratinocytes and suppresses UVB-induced skin tumorigenesis.¹³ Consistent with a protective role, METTL14 expression is decreased in UVB-induced skin and further decreased in cSCC skin compared with UVB-protected skin in patient samples.¹⁴ While these findings have established a tumor-suppressive role for the m⁶A methyltransferase METTL14 in the skin, less is known about the function of m⁶A demethylases in the UVB response or cSCC development. In particular, FTO is an RNA demethylase that primarily targets m⁶A modifications for demethylation.^11,12 Previous work has shown that FTO localizes to DNA damage sites after UVA laser microirradiation, along with METTL3 and METTL14.¹⁵ Although one study showed that FTO was decreased in HaCaT human keratinocyte cells and mouse skin after UVB damage,¹⁶ recently, we demonstrated that FTO expression is increased by both UVB irradiation and arsenic in mouse skin to enhance arsenic/UVB-induced skin tumorigenesis in mice.¹⁷ However, the function of FTO in the UVB-induced DNA damage response remains unknown. Notably, several recent studies have identified potent and selective small-molecule inhibitors of FTO. Huang et al.¹⁸ developed FB23–2 and reported that it suppresses acute myeloid leukemia (AML) progression in xenograft models. Su et al.¹⁹ later developed FTO inhibitors CS1 and CS2 with ~10-fold greater potency and demonstrated antileukemic activity in vivo. Motivated by these developments, we investigated here whether genetic or pharmacologic inhibition of FTO modulates the UVB-induced DNA damage response and skin carcinogenesis.

MATERIALS AND METHODS

Cell culture and UVB radiation

HaCaT (human keratinocyte, generously provided by Dr. Fusenig), A431 (human skin squamous carcinoma), and HEK293T (human embryonic kidney) cell lines were maintained in Dulbecco’s modified Eagle medium (DMEM, Invitrogen, Carlsbad, CA, USA). Wild-type and FTO knockout (KO) cells were generated using CRISPR as described previously.¹⁷ The culture medium was supplemented with 10% fetal bovine serum (Gibco), 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen, Carlsbad, CA, USA). Normal human epidermal keratinocytes (NHEK) cells were purchased from Lonza and maintained in KGM Gold keratinocyte growth basal medium (Lonza, 00192151) and KGM Gold keratinocyte growth medium supplements and growth factors (Lonza, 00192152). For UVB irradiation, the cells were washed twice with phosphate-buffered saline (1× PBS buffer, Invitrogen) and treated with UVB (20 or 30 mJ/cm², unless otherwise specified) using a UV Stratalinker 2400 equipped with UVB bulbs (Stratagene). Control samples underwent mock irradiation. The UVB dosage was consistently measured using a Goldilux UV meter fitted with a UVB sensor (Oriel Instruments).

Lentivirus generation and infection

Negative control (Sigma, SHC005) and METTL14 shRNA lentivirus plasmids were obtained from Sigma. CRISPR-Cas9 nuclease, sgRNAs targeting human FTO, and human sgRNA scramble controls were generated as described previously.¹⁷ Lentivirus particles were generated by co-transfecting HEK-293T cells with lentiviral vectors, the pCMVΔ8.2 packaging vector, and the pVSV-G envelope vector using X-tremeGENE^™ 9 DNA Transfection Reagent (MilliporeSigma). Supernatants containing virus particles were harvested at 24- and 48-h posttransfection and used to infect recipient cells. Target cells were infected in the presence of polybrene (8 μg/mL; Sigma-Aldrich) and subsequently selected with geneticin (G418; Sigma) or puromycin (Santa Cruz Biotechnology, 1 μg/mL) for a minimum of 6 days. The plasmids were validated by sequencing at the Comprehensive Cancer Center DNA Sequencing Facility, University of Chicago.

siRNA transfection

siRNAs targeting FTO, METTL14, DDB2, as well as a control siRNA, were obtained from Santa Cruz Biotechnology (sc-75002, sc-89054, sc-37799, sc-37007). Cells were transfected with these siRNAs using the Nucleofector device (Amaxa, Gaithersburg, MD, USA) following the protocol provided by the manufacturer.

Quantitative real-time PCR (qPCR)

Quantitative real-time PCR assays were conducted using a CFX Connect real-time system (Bio-Rad, Hercules, CA, USA) equipped with Bio-Rad iQ SYBR Green Supermix (Bio-Rad, Hercules, CA, USA). The threshold cycle number (CQ) for each sample was determined in triplicate and subsequently normalized to the housekeeping control for relative mRNA expression. All primer sequence information is provided below.

Gene name	Primer direction	Primer sequence
Human DDB2	Forward	5′-AGGACGCGATGGCTCCCAAGA-3′
	Reverse	5′-TTCTGCTAGGACCGGAGCCCT-3′
Human GAPDH	Forward	5′-AATCCCATCACCATCTTCCA-3′
	Reverse	5′-TGGACTCCACGACGTACTCA-3′

m⁶A dot blot

Total RNA was extracted using a RNeasy plus Mini Kit (QIAGEN, Hilden, Germany) following the manufacturer’s protocol. Subsequently, the RNA was denatured by heating at 98°C for 10 min and then applied onto an Amersham Hybond-N+ membrane (GE Healthcare, Chicago, IL, USA), which underwent UV-cross-linking twice. After drying, the membrane was blocked with 5% BSA (in 1× PBST) for 1 h and incubated overnight at 4°C with a specific anti-m⁶A antibody (Synaptic Systems, 202003; dilution: 1:2500). The membrane was treated with HRP-conjugated anti-rabbit IgG (Cell Signaling Technology; dilution range: 1:2000; Beverly MA, USA) for 1 h at room temperature and visualized using Thermo ECL SuperSignal Western Blotting Detection Reagent (Thermo Fisher Scientific; Waltham, MA, USA).

Cell proliferation

Cell proliferation was evaluated using the Cell Counting Kit-8 (CCK-8) (Sigma-Aldrich, St. Louis, MO, USA), following the instructions provided by the manufacturer.

Immunoblotting

To obtain protein extracts, cells were first washed with cold PBS and subsequently lysed using RIPA buffer (Thermo Fisher Scientific, Waltham, MA, USA) containing a combination of protease and phosphatase inhibitors (Thermo Fisher Scientific, Waltham, MA, USA). The lysates were subjected to sonication and then centrifuged for 15 min at 4°C. Protein concentrations were determined using the Pierce BCA assay (Thermo Fisher Scientific, Waltham, MA, USA), after which the samples were heated at 70°C for 10 min. The levels of protein expression were evaluated by SDS-polyacrylamide gel electrophoresis followed by immunoblot analysis. The antibodies utilized in this study were as follows: anti-DDB2 (Santa Cruz; sc-81246; diluted at 1:500), anti-FTO (Santa Cruz; SC-271713; diluted at 1:200), anti-GAPDH (Santa Cruz; sc-47724; diluted at 1:5000), anti-m⁶A (Synaptic Systems; 202003; diluted at 1:2500), and anti-Cyclobutane Pyrimidine Dimers (Cosmo Bio USA; CAC-NM-DND001; diluted at 1:4000).

m⁶A RNA immunoprecipitation (m⁶A-RIP)

Total RNA in the amount of 100–150 μg was isolated from the cells using TRIzol according to the manufacturer’s instructions. The mRNA was subsequently purified with the Dynabeads mRNA DIRECT Kit (Thermo Fisher, 61012). One microgram of mRNA was sheared to ~200 nucleotides using a Bioruptor^® Pico Sonication System, and 5% of the fragmented mRNA was retained as input material. Fragments of mRNA containing m⁶A modifications were enriched using the EpiMark N⁶-Methyladenosine Enrichment Kit (NEB, E1610S) and then purified with the RNA Clean and Concentrator kit (Zymo Research).

CPD slot blotting

Cells were harvested at different time points after UVB exposure, and genomic DNA was isolated using the QIAamp DNA Mini Kit (51306, QIAGEN, Hilden, Germany) according to the manufacturer’s protocol. The levels of CPDs were quantified through immunoblot analysis using an anti-CPD antibody (TDM-2, Cosmo Bio Co., Koto-Ku, Tokyo, Japan; CAC-NM-DND-001; diluted at 1:4000). Methylene blue staining served as a loading control for total DNA content. Repair dynamics were evaluated by calculating the percentage of CPD repair based on the ratio of the average optical density at specific points compared with 0 h. At 0 h, the CPD damage level corresponds 100% prior to any repair following UVB irradiation.

Polysome profiling

Cells were grown until they reached approximately 80% confluency. A sucrose gradient was prepared immediately before use. Prior to collection, cycloheximide (CHX) was added to the culture medium at a concentration of 50 μg/mL for 7 min. Four 15 cm² plates of cells from each group were collected and washed with cold PBS containing 50 μg/mL CHX. The lysis buffer consisted of 20 mM HEPES (pH 7.6), 100 mM KCl, 5 mM MgCl₂, 100 μg/mL CHX, and 1% Triton X-100, supplemented with freshly added Complete, Mini, EDTA-free Protease Inhibitor Cocktail and SUPERNase In. The cells were disrupted and subsequently treated with Turbo DNase, with the loading quantity standardized based on the OD260 readings. Twenty percent of the sample was reserved as input material, while the remaining portion was placed at the top of the sucrose gradient and subjected to ultra-centrifugation. Following this, the sample was divided into 30 fractions (each 0.5 mL in volume) and analyzed using a Gradient Station (BioCamp), which included an ECONOUV monitor (Bio-Rad, Hercules, CA) and a Gilson FC203B fraction collector (Mandel Scientific, Guelph, Canada). RNA was extracted from fractions 5 through 18 and then assessed via qPCR analysis. The expression levels of DDB2 in each fraction were adjusted relative to GAPDH and the saved input.

Human skin/tumor samples

The use of de-identified normal human skin and human skin cancer was carried out in compliance with approved IRB protocols from the University of Chicago Institutional Review Board. The human SCC array was obtained from US Biomax (SK801b, Derwood, MD).

Immunofluorescence

Immunofluorescence was performed as described previously.¹⁷ Briefly, the human SCC tissue array slide was pretreated by antigen retrieval and incubated with blocking solution. The slide was incubated at 4°C with primary anti-FTO (Abcam, ab92821, 1:100) or anti-cytokeratin (ORIGENE, BP5069, 1:200) antibodies. After removing the primary antibodies, the slide was washed and then incubated at room temperature with Alexa Fluor 594-conjugated secondary rabbit IgG (Jackson ImmunoResearch, 711–585–152, 1:100), Alexa Fluor 488-conjugated secondary mouse IgG (Jackson ImmunoResearch, 715–545–150, 1:100), or DyLight 405 AffiniPure secondary Guinea Pig IgG (Jackson ImmunoResearch, 706–475–148, 1:200) for 1 h. After washing, the slide was mounted with Fluoromount Mounting Medium (Sigma-Aldrich, F4680) followed by analysis using a fluorescence microscope (Olympus IX71, Olympus Life Science, Japan). Two investigators independently scored the immunofluorescence/immunohistochemistry intensity and staining blindly, as 3 (strong), 2 (medium), 1 (weak), and 0 (negative).

Mouse studies

All animal procedures were authorized by the Institutional Animal Care and Use Committee (IACUC) at the University of Chicago. Athymic nude mice were sourced from Harlan-Envigo. Female mice, aged between 6 and 8 weeks, were subcutaneously inoculated with 1 × 10⁶ A431 human squamous cell carcinoma cells suspended in PBS. Injections included cells both with and without FTO knockout. Tumor progression was tracked over time and assessed using a caliper. The tumor volume was estimated using the equation: (mm³) = d² × D/2, wherein “d” denotes the smallest diameter and “D” indicates the largest diameter.

Statistical analyses

The statistical analysis was performed using GraphPad Prism software version 9.0 (GraphPad Software, Inc., CA). All data are expressed as the mean with standard error of the mean (SE), unless otherwise indicated. To evaluate statistical differences among experimental groups, Student’s t-test was used unless otherwise indicated. A p-value of less than 0.05 was considered as statistically significant.

RESULTS

FTO deletion promotes UVB-induced DNA damage repair

To investigate the role of FTO in the UVB damage response, we knocked out FTO using CRISPR-Cas9 in HaCaT cells, an immortalized human keratinocyte cell line (Figure 1A). Dot blot analysis revealed a drastic increase in global RNA m⁶A levels in FTO knockout (FTO KO) cells relative to wild-type (WT) cells, consistent with the known role of FTO as an m⁶A eraser (Figure 1B). To assess the functional impact of FTO loss on UVB-induced damage, we performed slot blot analysis of CPDs in HaCaT cells exposed to 20 mJ/cm² UVB irradiation. Interestingly, FTO KO cells led to enhanced CPD repair at 6-, 9-, and 12-h post-UVB irradiation compared with WT cells (Figure 1C,D). These results suggest that FTO loss improves UVB-induced DNA damage repair.

FIGURE 1.

FTO knockout enhances UVB-induced DNA damage repair. (A) Immunoblot analysis of FTO and GAPDH in HaCaT cells with or without FTO knockout. (B) Dot blot analysis of global m⁶A in total RNA in WT or FTO KO HaCaT cells. (C) Slot blot analysis of UVB-induced CPDs in WT or FTO knockout HaCaT cells at different time points post-UVB irradiation (20 mJ/cm²). (D) Quantification of (C). Mean ± SE (n = 3). ****p < 0.0001; Student’s t-test. Methylene blue (MB) was used as an equal loading control for total RNA (B) and DNA (C).

FTO loss enhances UVB-induced DNA damage repair in a METTL14-dependent manner

Given that FTO KO increased m⁶A levels in HaCaT cells, we next asked whether FTO modulates UVB-induced DNA repair through an m⁶A-dependent pathway involving the m⁶A writer METTL14. Previous work has shown that knockdown of METTL14 impairs CPD repair through an m⁶A-dependent mechanism.¹³ To determine whether FTO and METTL14 exert their effects on UVB-induced DNA repair via regulation of a shared pathway, we knocked down METTL14 or FTO using shRNA and siRNA, respectively, in HaCaT cells and assessed CPD repair (Figure 2A–C). As expected, FTO knockdown cells exhibited enhanced CPD repair compared with control cells (Figure 2B,C). However, this improvement was abolished by concurrent METTL14 depletion, resulting in CPD levels comparable to those observed in shNC cells (Figure 2B,C). To evaluate these phenotypes in primary cells, we performed similar experiments in NHEK cells. Consistently, siRNA-mediated FTO knockdown in NHEK cells enhanced UVB-induced CPD repair (30 mJ/cm²), whereas METTL14 silencing had the opposite effect (Figure 2D–F). These results suggest that FTO and METTL14 act in a shared m⁶A-dependent pathway to regulate UVB-induced CPD repair.

FIGURE 2.

Effect of FTO knockout on DNA damage repair depends on METTL14. (A) Immunoblot analysis confirming METTL14 knockdown and FTO knockdown in HaCaT cells. (B) Slot blot analysis of UVB-induced CPDs in HaCaT cells with or without METTL14 knockdown and/or FTO knockdown at different time points post-UVB irradiation (20 mJ/cm²). (C) Quantification of B. Mean ± SE (n = 3). (D) Immunoblot analysis confirming METTL14 knockdown and FTO knockdown in normal human epidermal keratinocytes (NHEK cells). (E) Slot blot analysis of UVB-induced CPDs in NHEK cells with or without METTL14 knockdown or FTO knockdown at different time points post-UVB irradiation (30 mJ/cm²) (right). (F) Quantification of E. Mean ± SE (n = 3). Methylene blue (MB) was used as an equal loading control for total DNA (B and E). ***p < 0.001; ****p < 0.0001; Student’s t-test (C, F).

FTO knockout increases DDB2 expression via m⁶A-dependent translational regulation to affect CPD repair

Given the observed interaction of FTO and METTL14 in UVB-induced DNA repair, we hypothesized that FTO might target the GGR factor DDB2, a key METTL14 target identified in our previous study.¹³ Indeed, immunoblot analysis revealed a marked increase in DDB2 protein levels in FTO KO HaCaT cells compared with WT cells (Figure 3A). While FTO KO increased m⁶A modification of DDB2 mRNA transcripts as shown by m⁶A-RIP-qPCR (Figure 3B), DDB2 mRNA levels remained unchanged (Figure 3C), suggesting that regulation occurs at the protein level. Prior studies have shown that m⁶A modification can modulate gene expression via control of mRNA translation.²⁰ To determine whether the increase in DDB2 protein was a result of enhanced translation, we performed polysome profiling to assess ribosome occupancy on DDB2 mRNA. Polysome profiles demonstrated a pronounced shift of DDB2 transcripts toward heavy polysome fractions in FTO KO cells, indicative of heightened translational efficiency (Figure 3D). qPCR analysis of ribosome-bound DDB2 mRNA across gradient fractions confirmed an increased enrichment in translationally active fractions (Figure 3E). These findings suggest that FTO loss elevates intracellular m⁶A levels, which in turn promotes ribosome loading and translational upregulation of DDB2 mRNA. Finally, to test whether the enhanced CPD repair observed upon FTO loss is mediated by DDB2, we knocked down DDB2 in FTO KO HaCaT cells. Indeed, DDB2 knockdown prevented the effect of FTO knockdown of CPD repair (Figure 3F–H), suggesting that DDB2 may be a major functional target of FTO in the UVB-induced repair phenotype.

FIGURE 3.

FTO knockout increases translation of DDB2 in an m⁶A-dependent manner to promote DNA damage repair. (A) Immunoblot analysis of DDB2 and GAPDH in WT/FTO KO-1/FTO KO-2 HaCaT cells. (B) qPCR analysis of m⁶A levels on DDB2 mRNA following m⁶A IP. Mean ± SD (n = 3). (C) qPCR analysis of DDB2 mRNA levels. Mean ± SD (n = 3). (D) Polysome profiling assays in fractionations of WT/FTO KO-1/FTO KO-2 HaCaT cell lysates. (E) qPCR analysis of DDB2 mRNA levels in different fractions of ribosomes (n = 3). (F) Immunoblot analysis confirming DDB2 knockdown and FTO knockout in HaCaT cells. (G) Slot blot analysis of UVB-induced CPDs in HaCaT cells with or without DDB2 knockdown and/or FTO knockout at different time points post-UVB irradiation (20 mJ/cm²). (H) Quantification of G. Mean ± SE (n = 3). Methylene blue (MB) was used as an equal loading control for total DNA (G). *p < 0.05; ***p < 0.001; ns, not significant; Student’s t-test (B, C, H).

Pharmacologic inhibition of FTO recapitulates genetic knockout effects on UVB-induced DNA damage repair

To evaluate the potential of targeting FTO to protect against UVB-induced DNA damage, we tested two small-molecule FTO inhibitors, CS1 and FB23–2,^18,19 for their ability to modulate UVB-induced DNA repair. First, treatment of HaCaT cells with CS1 (200 nM, 72 h) revealed an increase in DDB2 expression (Figure 4A). Global m⁶A levels, as measured by dot blot, were elevated in CS1-treated cells (Figure 4B). Consistent with the phenotype observed upon FTO knockout, CS1-treated cells demonstrated a robust increase in m⁶A enrichment on DDB2 mRNA, while DDB2 mRNA levels remained unchanged (Figure 4C,D). Functionally, CS1-treated HaCaT cells exhibited enhanced CPD repair (Figure 4E,F). Rescue experiments in FTO KO HaCaT cells showed that CS1 treatment did not further enhance CPD repair (Figure 4G,H), indicating that the DNA repair-promoting effects of CS1 are mediated though FTO inhibition.

FIGURE 4.

FTO inhibitor CS1 increases global m⁶A levels and promotes DNA damage repair in human keratinocytes. (A) Immunoblot analysis of DDB2, FTO, and GAPDH in HaCaT cells treated with or without CS1 (200 nM) for 72 h. (B) Dot blot analysis of global m⁶A levels in HaCaT cells treated with or without CS1 (200 nM) for 72 h. (C) qPCR analysis of DDB2 mRNA levels in HaCaT cells treated with or without CS1 (200 nM) for 72 h. Mean ± SE (n = 3). (D) qPCR analysis of the m⁶A levels of the DDB2 transcript following m⁶A IP in HaCaT cells treated with or without CS1 (200 nM) for 72 h. Mean ± SD (n = 3). (E) Slot blot analysis of UVB-induced CPDs in vehicle or CS1 pretreated HaCaT cells at different time points post-UVB irradiation (20 mJ/cm²). (F) Quantification of E. Mean ± SE (n = 3). (G) Slot blot analysis of UVB-induced CPDs in vehicle or CS1 pretreated WT/FTO KO-1 HaCaT cells at different time points post-UVB irradiation (20 mJ/cm²). (H) Quantification of G. Mean ± SE (n = 3). Methylene blue (MB) was used as an equal loading control for total RNA (B) and DNA (E and G). ***p < 0.001; ****p < 0.0001; ns, not significant; Student’s t-test, differences from vehicle (C, D) or indicated groups (F, H).

To further investigate the therapeutic potential of FTO inhibition, we next treated HaCaT cells with the FTO inhibitor FB23–2 (2.5 μM for 72 h).¹⁸ As expected, FB23–2 treatment upregulated DDB2 protein levels (Figure 5A). FB23–2 also increased global m⁶A levels compared with vehicle-treated controls in both HaCaT and NHEK cells (Figure 5B). Interestingly, FB23–2 treatment restored DDB2 protein expression in METTL14-depleted cells, reversing the effect caused by METTL14 loss (Figure 5C). FB23–2 pretreatment accelerated CPD clearance in METTL14-deficient cells, compared with vehicle-treated shMETTL14 (Figure 5D,E). These results demonstrate that FB23–2-mediated FTO inhibition reverses the effect of METTL14 deficiency, reactivating GGR-mediated DNA damage repair. Taken together, these data indicate that small-molecule inhibition of FTO with either CS1 or FB23–2 recapitulates the phenotypic effects of genetic FTO knockout in enhancing UVB-induced DNA damage repair. These findings highlight the potential of targeting the m⁶A-FTO axis to counteract deficiencies in the m⁶A writer complex and promote genome stability in response to UVB-induced damage.

FIGURE 5.

FTO inhibitor FB23–2 promotes DNA damage repair in a METTL14-dependent manner in human keratinocytes. (A) Immunoblot analysis of DDB2 and GAPDH in HaCaT cells treated with or without FB23–2 (2.5 μM) for 72 h. (B) Dot blot analysis of global m⁶A levels in HaCaT cells (left) and NHEK cells (right) treated with or without FB23–2 as in A. (C) Immunoblot analysis of DDB2 and GAPDH in shNC/shMETTL14–1/shMETTL14–2 HaCaT cells treated with or without FB23–2. (D) Slot blot analysis of UVB-induced CPDs in vehicle or FB23–2-treated shNC/shMETTL14–1/shMETTL14–2 HaCaT cells at different time points post-UVB irradiation (20 mJ/cm²). (E) Quantification of D. Mean ± SE (n = 3). **p < 0.01; ***p < 0.001; Student’s t-test. Methylene blue (MB) was used as an equal loading control for total RNA (B) and DNA (D).

Loss of FTO suppresses skin tumor growth

Finally, we investigated the role of FTO in cSCC. In vitro proliferation assays under low-serum (2% FBS) conditions revealed reduced proliferation in FTO KO A431 cells at 48 h post-seeding (Figure 6A,B). Aligning with the observed beneficial role of FTO KO on cell proliferation in vitro, subcutaneous injection of FTO KO A431 cells into nude mice led to a drastic reduction in both tumor volume (Figure 6C) and tumor weight (Figure 6D) compared with WT controls. To investigate FTO expression in human samples, we performed immunofluorescence staining of human cSCC tissues. We observed robust FTO expression co-localizing with keratin-positive tumor regions, while normal skin showed minimal FTO signal (Figure 6E). Accordingly, quantification of staining intensity in tumor samples revealed an increase in FTO levels in tumor tissues compared with normal controls (Figure 6F). Together, these findings emphasize a critical role for FTO in promoting cSCC tumor growth and support further exploration of FTO inhibition as a preventive and/or therapeutic strategy in this cancer.

FIGURE 6.

FTO knockout suppresses cSCC cell proliferation and tumorigenicity. (A) Immunoblot analysis confirming FTO knockout in A431 human squamous carcinoma cells. (B) Cell proliferation (CCK-8) of A431 cells with or without FTO knockout cultured in 2% FBS. (C) Tumor volume in nude mice following subcutaneous inoculation of WT/FTO KO-1/FTO KO-2 A431 cells. (D) Tumor weight in nude mice following subcutaneous inoculation of WT/FTO KO-1/FTO KO-2 A431 cells. (E) Representative immunofluorescence staining of FTO and keratin in normal human skin (n = 6) and human cutaneous squamous cell carcinoma (cSCC) samples (n = 56). (F) Percentage of tumors (in stacked column format) for each score of FTO. 0 (Negative), 1 (Weak), 2 (Medium), and 3 (Strong). Results are shown as Mean ± SE (n = 8 for B; n = 5 for C and D). ***p < 0.001; ****p < 0.0001; (24 and 48 h for B; day 11, 13, and 15 for C and D), compared with the WT group (B–D); Student’s t-test. **p < 0.01; Mann–Whitney U test (F).

DISCUSSION

Emerging evidence suggests a pivotal role for m⁶A RNA methylation in the response to UVB-induced DNA damage. Recently, we have shown that the m⁶A writer METTL14 promotes the expression of the GGR factor DDB2 via increased mRNA stability to enhance GGR in keratinocytes, in coordination with the reader YTHDF1.¹³ In addition, the m⁶A reader YTHDC2 has been shown to inhibit the GGR response and is upregulated in cSCC tissue compared with normal skin, further implicating the m⁶A regulatory machinery in cSCC development.²¹ However, the potential role of m⁶A demethylases has remained elusive. Here, we extend these findings by demonstrating that the m⁶A eraser FTO plays a role in suppressing UVB-induced DNA damage repair. Specifically, genetic knockout or pharmacological inhibition of FTO led to accelerated CPD repair following UVB irradiation. Furthermore, the knockdown of METTL14 reversed the effect of FTO loss, suggesting that (1) the mechanism is m⁶A-modification dependent and (2) FTO and METTL14 likely regulate shared targets in the UVB-induced DNA damage response.

Given the potential overlap in targets, we examined whether DDB2, a downstream target of METTL14 in the UVB damage response, might also be regulated by FTO. Indeed, we observed that DDB2 knockdown diminished the effect of FTO loss on CPD repair, suggesting that FTO’s effect on DNA repair is mediated at least in part by regulating DDB2 expression. Further investigation into the underlying mechanism revealed that FTO regulates DDB2 expression at the translational level through m⁶A modification of DDB2 mRNA. While m⁶A is implicated in both mRNA stability and translation, our data that FTO loss increases DDB2 protein levels without affecting mRNA abundance points to a translational mode of control, which we subsequently confirmed through polysome profiling. These results expand our understanding of how m⁶A modification regulates the UVB-induced DNA damage response at the translational level.

Finally, we observed that FTO expression was higher in cSCC tissues compared with normal skin, and FTO deletion decreased tumor growth in mice. Taken together, our findings suggest that FTO may represent a targetable enzyme for UVB-associated skin malignancies, such as cSCC. This work opens several avenues for future investigation. First, additional studies could identify the m⁶A reader that mediates the translational effects of FTO loss in the context of UVB exposure. Potential candidates include YTHDF1, which is known to regulate m⁶A-dependent translation and coordinates with METTL14-mediated methylation to regulate DDB2 mRNA stability, and YTHDC2, which we have previously shown to affect UVB-induced GGR and has also been reported to regulate mRNA translation by binding m⁶A within coding regions.^13,21 Second, future studies may identify additional targets of FTO after UVB irradiation through sequencing approaches, as FTO may regulate other repair factors not investigated in the present study. Third, further preclinical studies are warranted to assess the efficacy of FTO inhibitors, such as CS1 and FB23–2, on UVB-induced skin tumorigenesis in vivo using UVB-induced cSCC mouse models, including whether systemic or local administration is sufficient to inhibit UVB-induced tumor initiation or reduce the size of established tumors.

Abbreviations:

6–4PPs

6–4 photoproducts

AML

acute myelocytic leukemia

BCC

basal cell carcinoma

CPD

cyclobutane pyrimidine dimers

cSCC

cutaneous squamous cell carcinoma

GGR

global genome repair

MB

methylene blue

NER

nucleotide excision repair

NHEKs

normal human epidermal keratinocytes

NMSC

nonmelanoma skin cancers

UVB

ultraviolet B

XP

xeroderma pigmentosum

DATA AVAILABILITY STATEMENT

All data needed to evaluate the conclusions in the paper are present in the paper.