Differential expression of tRFs and tiRNAs in PTC tissues and paired normal tissues

To determine the expression of tRFs and tiRNAs in PTC, four pairs of PTC tissues (T) and adjacent normal tissues (AT) were collected for a high-throughput sequencing experiment. The specific workflow is summarized in Fig 1A. The principal component analysis and the correlation coefficient analysis showed that there was a significantly different expression between the two groups (Fig 1B and C). A total of 140 commonly expressed tRFs and tiRNAs were detected in tumor tissues and 125 in ATs, and 110 of which were detected in both groups (Fig 1D). The component analysis revealed that tRFs were the primary down-regulated type in PTC, whereas tiRNAs were the slightly outnumbered type (Fig 1E), indicating that tRFs may play a more significant role in PTC. The volcano plot and heatmap showed that a total of 35 considerably dysregulated tRFs were detected in the tumor group. Of these, 20 were present at particularly low levels in tumor tissues compared with ATs (Fig 1F and G) (Table S1).

To further identify tRFs as prospective target molecules, the distributions of tRFs with differential expression in tumor tissues and ATs were analyzed. The result showed that tRF-5c was the main down-regulated type (Fig 1H), indicating that tRF-5c may function as a potential tumor suppressor of PTC. Considering the parameters of tRF-5c, such as P-value, FC-value, and length, tRF-30 was selected for further qRT–PCR verification in 30 pairs of tumor tissues and ATs. The result showed that tRF-30 was notably down-regulated in tumors (Fig 1I). Therefore, we hypothesized that tRF-30 might play an inhibiting role in PTC. Furthermore, the expression of tRF-30 was detected in the human PTC cell lines BCPAP, TPC1, KTC1, and NPA87, related to the Nthy-ori 3-1 normal thyroid epithelial cell line. The result showed that BCPAP and KTC1 were the lowest two cell lines in tRF-30 expression (Fig 1J). Moreover, we detected the knockdown models of tRF-30 in TPC1 and NPA87 cells by qRT–PCR assay, and the results showed that the knockdown efficiencies of the cells were unacceptably low (Fig S1A and B). Therefore, BCPAP and KTC1 cell lines were chosen for further biological function research and downstream mechanism studies.

tRF-30 inhibits proliferation and invasion in PTC cells and regulates tumor growth in a mouse model

To further explore the biological function of tRF-30, the overexpression and knockdown models were constructed in BCPAP and KTC1 cells. The transfection efficiencies were detected by qRT–PCR assay (Fig 2A and B). Cell counting kit-8 (CCK-8) assay indicated that tRF-30 overexpression significantly decreased the viability of BCPAP and KTC1 cells. In contrast, tRF-30 knockdown markedly increased cell viability (Fig 2C and D). Colony formation assay demonstrated that tRF-30 overexpression dramatically reduced colony formation in BCPAP and KTC1 cells, whereas tRF-30 knockdown considerably boosted colony formation in cells (Fig 2E and F). Transwell invasion assay revealed that tRF-30 overexpression significantly inhibited the invasion of BCPAP and KTC1 cells, whereas tRF-30 knockdown significantly promoted cell invasion (Fig 2G and H).

To further confirm these effects of tRF-30 in vivo, (1 × 10⁷, 200 μl) BCPAP cells stably transfected with tRF-30 or negative control were subcutaneously injected into the right upper back of the nude mice. Tumor volumes were measured every week and tumors were dissected after 5 wk (Fig 2I). Tumor volumes and weights were markedly decreased in the oe-tRF-30 group compared with the oe-NC group (Fig 2J and K). Moreover, subcutaneous tumors were further deployed for Ki-67 immunohistochemistry (IHC) assays. The result showed that the oe-tRF-30 group had considerably lower tumor proliferation (Fig 2L). These data revealed that tRF-30 transfection inhibited tumor growth of PTC in vivo. Collectively, both in vitro and in vivo assays demonstrated that tRF-30 acted as a suppressive factor in PTC.

tRF-30 directly binds to PC and down-regulates PC protein by decreasing its stability in PTC cells

As has been widely reported, tRFs contribute to tumor progression by interacting with target RNAs or proteins (Xie et al, 2020; Yu et al, 2021), so we speculated that tRF-30 would exert a suppressive effect by associating with target proteins. RNA-pulldown assay was performed to verify the hypothesis (Fig 3A). The biotin-labeled tRF-30 and its antisense probes were synthesized and incubated with protein lysates of BCPAP cells to identify distinct protein bands. The results from two independent tRF-30 pulldown assays detected a 130-kD band on the silver staining gel (Fig 3B). Furthermore, 208 specific binding proteins were selected by mass spectrometry because of their presence in the band containing the tRF-30 sequence, but not that containing the antisense sequence (Table S2). Among them, PC was verified by three independent pulldown experiments and Western blot analysis (Fig 3C). In addition, RNA immunoprecipitation (RIP) assays also confirmed that tRF-30 was enriched when an antibody against PC was used in BCPAP and KTC1 cell lysates, compared with the IgG antibody group (Fig 3D).

The above mentioned studies have proved that tRF-30 could specifically interact with PC, but more research into the interaction’s molecular processes is required. The qRT–PCR assay results showed that tRF-30 overexpression and knockdown had no significant effect on the PC mRNA expression in BCPAP and KTC1 cells (Fig 3E). Moreover, the Western blot assay indicated that tRF-30 adversely regulated PC protein levels in cells (Fig 3F). Therefore, we speculated that tRF-30 might regulate PC protein stability in PTC cells. Cycloheximide (CHX, 50 μg/ml) was used to stop protein (including PC) synthesis in BCPAP and KTC1 cells. The results showed that tRF-30 transfection enhanced the degradation of PC protein, whereas sh-tRF-30 transfection prolonged the half-life of PC protein in cells. GAPDH was referred as a control (Fig 3G and H). The protein degradation mediated by the ubiquitin–proteasome pathway is the primary mechanism for the regulation of intracellular protein levels, which contributes to the degradation of more than 80% of intracellular protein (Wang & Maldonado, 2006). The results of the in vitro ubiquitination assay in Fig S2 displayed that tRF-30 modulated the degradation of PC protein in BCPAP and KTC1 cells via the ubiquitination pathway. Collectively, these above mentioned outcomes demonstrated that tRF-30 directly bound to PC, regulated ubiquitin/proteasome-dependent degradation, and decreased the stability of PC protein in PTC cells.

PC promotes the proliferation and invasion of PTC cells and is up-regulated in PTC tissues

It has been confirmed in previous research that PC expression is dysregulated in several tumor types, including lung cancer, breast cancer, PTC (Wang et al, 2021; Liu et al, 2022b; Chen et al, 2022), etc. As shown in Fig 4A, PC was significantly elevated in PTC tissues. Then, we confirmed that expression levels of PC mRNA (Fig 4B) and protein (Fig 4C) were much higher in BCPAP cells than in KTC1 and TPC1 cells, and lowest in Nthy-ori 3-1 cells, suggesting the potential pro-tumorigenic role for PC in PTC. We therefore selected BCPAP and KTC1 cells to investigate the carcinogenic effect of PC. The overexpression and knockdown efficiencies were detected by qRT–PCR and Western blot (Fig 4D and E). The viability of BCPAP and KTC1 cells was significantly higher than that of vector mock-transfected control cells after PC overexpression, as shown by the CCK-8 assay. In contrast, PC knockdown markedly reduced cell viability (Fig 4F and G). In addition, PC overexpression significantly increased colony formation in BCPAP and KTC1 cells, whereas PC knockdown greatly reduced colony formation in cells (Fig 4H and I). Transwell invasion assay showed that the invasion ability of PC-overexpressed BCPAP and KTC1 cells was considerably increased compared with the control groups, whereas PC knockdown markedly reduced the invasion ability of cells (Fig 4J and K). These results suggested that PC could be an oncogenic driver in PTC, which was inconsistent with the impact of tRF-30 on PTC.

PC participates in the regulation of tRF-30 on PTC

We then confirmed PC expression levels in PTC tissues compared with ATs by IHC assay. The results showed that the PTC group had obviously higher PC levels than the AT group (Fig 5A), which was consistent with the previous studies (Fig 4A). To further validate whether tRF-30 regulates the biological behaviors of PTC via its association with PC, we performed a “rescue” experiment by transfecting tRF-30 with or without PC overexpression in BCPAP and KTC1 cells. After combined transfection, cell proliferation and invasion were detected, and CCK-8 assay results showed that PC transfection enhanced cell viability. Relative to the oe-tRF-30 group, cell viability in the oe-tRF-30 + oe-PC group was increased, which presented that PC transfection weakened the influence of tRF-30 overexpression on cell viability (Fig 5B). Similar results were found for the cell invasion experiment. Fig 5C and D illustrated that PC transfection elevated cell invasion, and alleviated the oe-tRF-30–resulted reduction of the invasion ability. Moreover, in vivo assays showed that PC expression in the oe-tRF-30 group was significantly decreased compared with that in the oe-NC group (Fig 5E). In summary, we speculated that the effects of tRF-30 on the proliferation and invasion of PTC cells were dependent on PC, the expression of which was regulated by tRF-30.

tRF-30 mediates PC to decrease TCA cycle metabolites and regulate metabolic reprogramming in PTC cells

PC is a biotin-dependent enzyme, which is important for replenishing TCA cycle intermediates and mediating metabolic reprogramming in eukaryotes (Valle, 2017). It would be intriguing to investigate whether tRF-30 affects the TCA cycle and mediates metabolic reprogramming when it binds to PC in PTC cells. To further verify the metabolic regulation of tRF-30, a metabolomic analysis using liquid chromatography–tandem mass spectrometry (LC–MS/MS) was performed in BCPAP cells with or without tRF-30 transfection. The specific workflow is summarized in Fig 6A. The principal component analysis results indicated that there was an ideal significantly different expression between the two groups (Fig 6B). A total of 34 commonly expressed metabolites were clustered by sample groups and metabolin classes (Fig 6D). Notably, the volcano plot and heatmap revealed that tRF-30 considerably decreased the levels of citrate, oxaloacetate (OAA), the primary intermediates of the TCA cycle, and ornithine, the product of TCA-related urea cycle, in BCPAP cells. In addition, tRF-30 remarkably increased the levels of pyruvate, the main substrate of PC carboxylation, and guanosine, the precursor of TCA-related GTP synthesis, in BCPAP cells (Fig 6C and E). These modulated metabolites were involved in metabolic signaling pathways, including the TCA cycle, pyruvate metabolism, amino acid metabolism, and 2-oxocarboxylic acid metabolism (Fig 6F). Dysfunction of PC blocks the transition of pyruvate to OAA, which was consistent with our results that overexpression of tRF-30 resulted in the decrease of OAA and other TCA cycle-associated metabolites (Fig 6G). These results suggested that tRF-30 combination with PC could inhibit the replenishment of TCA cycle intermediates and change the metabolic reprogramming in PTC cells.

To confirm whether tRF-30 regulates TCA cycle intermediates via its association with PC, we transfected tRF-30 with or without PC overexpression in BCPAP and KTC1 cells. After tRF-30 and/or PC transfection, cell citrate and pyruvate levels were detected by ELISA assays. The results showed that exogenous expression of PC could partially weaken the decreased levels of citrate and the increased levels of pyruvate in tRF-30 transfection cells (Fig 7A and B). These results revealed that inhibition of PC by tRF-30 could interfere with TCA cycle intermediate anaplerosis and dysregulated metabolic intermediates to suppress the development of PTC via altering metabolic reprogramming (Fig 7C).

Carcinoma cells show preferential use of lactate-generating glycolysis over the common route of oxidative phosphorylation. This altered metabolism, named the “Warburg effect,” has been considered for a long time to be the major metabolic reprogramming in cancer. The changes in glycolytic pathways have been associated directly or indirectly with the downstream targets of many different ncRNAs (Mirzaei & Hamblin, 2020). Therefore, we further explored whether tRF regulated the Warburg effect. The glucose uptake and lactate production assay results showed that tRF-30 overexpression and knockdown had no significant effect on the glycolysis in BCPAP and KTC1 cells (Fig S3A and B). Therefore, we speculated that tRF-30 had no significant effect on the Warburg effect in PTC cells.

Patient tissue samples

A total of 50 pairs of human PTC tissues and adjacent normal tissues were collected from patients who underwent thyroidectomy in the Department of Breast and Thyroid Surgery at the Jinhua Municipal Central Hospital Medical Group between 2020 and 2022 and preserved in liquid nitrogen. All samples were pathologically diagnosed with PTC, and no patients received preoperative treatment. This study was approved by the Ethical Review Committee of Jinhua Municipal Central Hospital Medical Group and all specimens were collected with written informed consent obtained from patients.

Human PTC cell lines

The human PTC cell lines (BCPAP, TPC1, KTC1, NPA87) and the normal thyroid epithelial cells (Nthy-ori 3-1) were purchased from the iCell Bioscience Inc. All cell lines were identified by STR analysis and removed from mycoplasma contamination. Cells were cultured in RPMI 1640 medium (Sigma-Aldrich) supplemented with 10% FBS (Gibco) and 1% penicillin–streptomycin (Gibco) at 37°C under 5% CO₂.

Tissue tRF and tiRNA sequencing and analysis

To identify tRFs and tiRNAs in human PTC tissues, total RNAs were extracted from four pairs of PTC tissue and paired normal tissues using TRIpure Total RNA Extraction Reagent (ELK Biotechnology). The small RNA-sequencing libraries were constructed as the method mentioned previously (Mo et al, 2019). Briefly, total RNA samples were pretreated by 3′-aminoacyl (charged) deacylation, 5′-OH (hydroxyl group) phosphorylation, m¹A demethylation, and m³C demethylation. Next, the pretreated total RNA samples were used for library construction and small RNA-sequencing. For tRF and tiRNA sequencing, 15–40 nt RNA biotypes were sequenced with the NextSeq 500/550 V2 system (#FC-404-2005; Illumina) according to the manufacturer’s protocol. The tRFs and tiRNAs sequencing analysis were performed by the Arraystar tRF and tiRNA-seq data package. The above sequencing data are accessible through accession number GSE197438 in the NCBI Gene Expression Omnibus (GEO) database.

qRT–PCR

The total RNAs of human PTC tissues and PTC cell lines were extracted using TRIpure Total RNA Extraction Reagent (ELK Biotechnology). The Bulge-loop miRNA qRT–PCR Starter Kit (RiboBio Co) was used for tRF and tiRNA reverse transcription in a 10 μl reaction volume. Then, the cDNA was amplified by qRT–PCR using EnTurbo SYBR Green PCR SuperMix (ELK Biotechnology) with the QuantStudio 6 Flex Real-Time PCR System (Life Technologies). Actin was used as the endogenous control for mRNA, and U6 was used as the endogenous control for tRFs and tiRNAs. The expression of genes was calculated using the 2^−ΔΔCt method. The primer sequences were listed in Table S3.

Lentivirus preparation and cell transfection

The tRF-30 overexpression model (oe-tRF-30) and the PC overexpression model (oe-PC) were constructed by transfection with the mimic sequence. The tRF-30 knockdown model (sh-tRF-30) and the PC knockdown model (sh-PC) were constructed by transfection with the shRNA lentivirus. All the lentivirus models were designed and synthesized by QIAGEN. The PTC cells were infected with lentiviruses by using Lipofectamine 3000 reagent (Invitrogen) according to the manufacturer’s protocol. All the mimic and shRNA sequences are listed in Table S4.

CCK-8 assay

The CCK-8 assay (C0038; Beyotime) was used to detect cell proliferation. Briefly, cells were plated in the 96-well plates at a density of 1,000 cells per well, at fixed intervals of 0, 24, 48, 72, and 96 h; 10 μl of CCK-8 reagent was added to each well and incubated for 2 h under 37°C. The OD values at 450 nm were measured by the microplate reader (Thermo Fisher Scientific). Each assay was repeated three times.

Colony formation assay

The plate cloning assay was used to detect cell colony formation. Briefly, Cells were seeded in the six-well plates at a density of 600–800 cells per well and cultured for seven consecutive days at 37°C under 5% CO₂. Then, the cells of each well were collected, fixed in 4% PFA for 30 min, stained with crystal violet for 30 min, and washed with PBS. Finally, the clones of each well were photographed under the light microscope (Olympus) and counted by the ImageJ software (NIH, Bethesda, Maryland, USA). Each assay was repeated three times.

Transwell invasion assay

The invasion assay was carried out using the transwell chambers (Corning) with Matrigel mix (3 mg/ml) (Becton Dickinson and Co.[BD]) coated inserts. Cells were starved on day 1 by culture in the FBS-free medium. On day 2, 3.0 × 10⁵ cells were resuspended in 500 μl of serum-free medium and inoculated into the upper chamber, and 700 μl of RPMI 1640 medium containing 20% FBS without cells was added to the lower chamber as the chemoattractant. The transwell chambers were incubated at 37°C for 48 h under 5% CO₂. Then, the cells on the surface of the lower chamber were fixed with 4% PFA for 30 min, stained with crystal violet for 30 min, and washed with PBS. Finally, the chambers were photographed under the light microscope and the cells were counted by the ImageJ software. Each assay was repeated three times.

RNA-pulldown assay and mass spectrometry analysis

Biotin-labeled tRF-30 and antisense probes were synthesized by GenScript. The Pierce Magnetic RNA-Protein Pull-Down Kit (20164; Thermo Fisher Scientific) was used to perform the RNA-pulldown assay according to the manufacturer’s protocol. The bead–RNA–protein complexes were diluted in SDS buffer, developed with a Fast Silver Stain Kit (P0017S; Beyotime), and the retrieved proteins were sent for mass spectrometry analysis.

RIP assay

The RIP assay was carried out using the RIP Kit (17-701; Millipore). Briefly, rabbit anti-PC antibody (PA5-50101; Thermo Fisher Scientific) or rabbit anti-IgG antibody (PP64B; Millipore) was integrated by the magnetic beads and incubated with the PTC cells lysate. Finally, the protein-binding RNA was retrieved and subjected to qRT–PCR analysis.

In vitro ubiquitination assay

Briefly, after different transfections, BCPAP and KTC1 cells were treated with 10 μM MG132 for 5 h (S1748; Beyotime) to inhibit degradation of ubiquitinated proteins. RIPA lysis and extraction buffer (Solarbio) were used to split cells. The lysate supernatant was collected for co-immunoprecipitation with anti-PC antibody (PA5-50101; Thermo Fisher Scientific) or anti-IgG antibody and Protein A/G agarose (Thermo Fisher Scientific) with gentle rocking at 4°C overnight. After centrifugation and precipitation, the supernatant was subjected to Western blotting. Antibodies against the following were used for Western blotting: mouse anti-PC antibody (sc-271493; Santa Cruz Biotechnology), mouse anti-ubiquitin antibody (3936; CST). Rabbit anti-GAPDH antibody (5174; CST) was used as the endogenous control.

Western blot analysis

Total cell protein was extracted using RIPA lysis and extraction buffer (Solarbio) supplemented with a protease inhibitor cocktail (P8340; Roche). After quantification of the protein concentration by the BCA protein detection kit (Beyotime), 20 μg protein lysates were loaded and separated with 10% SDS–PAGE (Epizyme), and then the protein bands were transferred onto a PVDF membrane (Millipore). The membrane was blocked with 5% skim milk for 1 h, and then incubated at 4°C overnight with mouse anti-PC antibody (sc-271493; Santa Cruz Biotechnology). Rabbit anti-GAPDH antibody (5174; CST) was used as the endogenous control. Then, the membrane was washed with Tris-buffered saline supplemented with Tween 20 (TBST) three times and incubated with the corresponding secondary antibody (Proteintech) for 2 h under room temperature and washed three times with TBST. Finally, the membrane was visualized with ECL reaction reagents (AS1059; Aspen) and detected by the Western Lightning Gel Imaging System (GE). Quantitative analysis of the individual band was measured by ImageJ software.

IHC staining

The IHC staining assay was carried out using the IHC Kit (Zsgb Bio) according to the manufacturer’s protocol. Briefly, the PFA-fixed and paraffin-embedded tissues were sectioned. The sections were deparaffinized in xylene, hydrated in ethanol with descending concentrations, blocked the endogenous peroxidase activity with serum, and incubated with primary antibodies at 4°C overnight and secondary antibodies for 1 h under room temperature. Finally, the sections were stained by DAB reagent and H&E staining (DAB substrate chromogen system; Dako); the slides were sealed and images were obtained with the microscope. Antibodies were used against the following: rabbit anti-Ki67 antibody (NBP2-22112; Novus) and mouse anti-PC antibody (sc-271493; Santa Cruz Biotechnology).

In vivo tumorigenesis model

The 5-wk-old female BALB/c nude mice were chosen for in vivo tumorigenesis to study the effect of tRF-30 on tumor growth. The nude mice were obtained from Vital River Laboratories and fed in the Experimental Animal Center of Jinhua Municipal Central Hospital Medical Group. The nude mice were randomly divided into each group (N = 10 mice per group). 1 × 10⁷ BCPAP cells stably transfected with tRF-30 or control sequence were subcutaneously injected into the right upper back skin area of the nude mice. The tumor volumes were measured every week. After 5 wk, the tumors were dissected for weight measurement, IHC staining, and WB analysis. All animal experiments were approved by the Ethical Review Committee of Jinhua Municipal Central Hospital Medical Group.

Cell pyruvate assay

The pyruvate assay kit (ab65342; Abcam) was used to measure pyruvate concentration in cells according to the manufacturer’s protocol. Briefly, 5 × 10⁶ BCPAP cells or KTC1 cells were homogenized in the pyruvate assay buffer and de-proteinized using a 10-kD spin filter (Amicon Ultra-0.5 ml Centrifugal Filter) with centrifugation at 14,000g for 10 min under 4°C. Then, the filtrate was added by the reaction mix and the plate was incubated at RT for 30 min under the dark. Finally, the mixture was measured at 587 nm with a microplate reader (model 680; Bio-Rad) and the concentration of pyruvate was interpolated from the standard curve. Treatment groups were calculated relative to the control group with a value of 1.

Cell citrate assay

The citrate assay kit (ab83396; Abcam) was used to assess citrate levels in cell lysates according to the manufacturer’s protocol. Briefly, 5 × 10⁶ BCPAP cells or KTC1 cells were resuspended in 100 μl of assay buffer and centrifuged at 16,000g for 5 min under 4°C. Then, the supernatant was de-proteinized using a 10 kD spin filter (Amicon Ultra-0.5 ml Centrifugal Filter) with centrifugation at 14,000g for 10 min under 4°C. Then, the filtrate was added by the reaction mix and the plate was incubated at RT for 30 min under the dark. Finally, absorbance at 570 nm was detected by a microplate reader (model 680; Bio-Rad) and the concentration of citrate was interpolated from the standard curve. Treatment groups were calculated relative to the control group with a value of 1.

Lactate production and glucose uptake measurement

3.0 × 10⁵ cells were inoculated onto six-well plates and incubated at 37°C for 24 h. Afterwards, 5 μl of the cell culture supernatant were collected in 96-well plates and mixed with 200 μl of the Glucose Assay Reagent (Sigma-Aldrich) or 100 μl of Lactate Assay Regent (Sigma-Aldrich). After incubation at 37°C for 20 min, the mixture was read calorimetric at the specific wavelength and the concentration of glucose and lactate was interpolated from the standard curve.

Targeted metabolomic analysis

The cell metabolites were extracted by the manufacturer’s protocol. In brief, adding precooled methanol extraction solvent (Merk), vortexing for 3 min, freeze–thawing with liquid nitrogen, and an ice water bath for 10 min. The cell suspension was centrifuged at 12,000g for 10 min under 4°C and the supernatant was filtered through the 3 kD cutoff filter (Amicon Ultra, Merk KGaA); finally, the filtrate was centrifugally concentrated to LC–MS/MS analysis. The metabolites of the TCA cycle, glycolysis, and other intracellular metabolites were determined using LC–MS on a tandem mass spectrometer coupled to ultra-performance liquid chromatography (Ex-ion LC AD). Sample quality was guaranteed by multiple reaction monitoring.

Statistical analysis

All the results were presented as mean ± SE. Statistical analysis was carried out using SPSS 26.0 (IBM). t test, Mann–Whitney U test, and analysis of variance (ANOVA) were performed to compare the statistical significance between two or more groups. A paired t test was performed to compare the expression of tRF-30 in 30 PTC tissues and paired normal tissues. For all data, a P-value < 0.05 was statistically significant (*P < 0.05, **P < 0.01, and ***P < 0.001).