Amplifying colorectal cancer progression: impact of a PDIA4/SP1 positive feedback loop by circPDIA4 sponging miR-9-5p

Human specimens

Twenty paired CRC tissues and adjacent healthy tissues were sourced from consenting CRC patients at Tianjin Medical University Cancer Institute & Hospital. The study received approval from the Ethics Committee of Tianjin Medical University Cancer Institute & Hospital. All patients included in the study had not undergone any prior treatment. Immediately following extraction, the collected tissues were preserved in liquid nitrogen for subsequent research purposes.

Cell lines and culture

The CRC cell lines, including HCT116, HT29, SW480, and RKO, as well as a normal colonic epithelial cell line (NCM460), were obtained from the American Collection of Tissue Culture (ATCC, Manassas, VA, USA) and utilized in this study. Cells were cultured in Dulbecco’s Modified Eagle Medium from Gibco (New York, NY, USA). The medium for all cell lines was supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. The cells were maintained at 37°C in 5% CO₂.

Cell transfection

HCT116 and HT29 cells (2 × 10⁵ cells/well) were seeded onto 6-well plates to achieve 80% confluence, and transfected with designated plasmids using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to manufacturer’s instructions. To silence circPDIA4 or SP1 expression, cells were transfected with 25 nM of circPDIA4/SP1 shRNA (sh-circPDIA4/sh-SP1; Gene Pharma, Shanghai, China) and the negative control shRNA (sh-NC; Gene Pharma, Shanghai, China). The miRNA mimics/inhibitor for miR-9-5p and negative control (NC) mimics/inhibitor were purchased from Gene Pharma (Shanghai, China). Fulllength sequence of SP1 was inserted into pcDNA3.1 vector (Gene Pharma, Shanghai, China) for establishing the SP1 overexpressed cell line, and the empty vector was used to serve as a NC. After 48 h of transfection, the stable-expressing CRC cells were selected using puromycin. Subsequently, cells were harvested for assays.

RNA extraction and reverse transcription-quantitative polymerase chain reaction (rt-qPCR)

Total RNA was extracted using TRIzol reagent (Invitrogen), followed by reverse transcription using PrimeScript™ RT Master Mix (Takara, Dalian, China) for RNA-to-cDNA conversion. The MirX miRNA First-Strand Synthesis Kit (Takara, Shanghai, China) was used for miRNA reverse transcription. RT-qPCR was performed with SYBR Premix Ex Taq™ (Takara, Shanghai, China) on an ABI Prism 7500 system (Applied Biosystems, Foster City, CA, USA). The reaction program consisted of a 9-min pre-denaturation at 95°C, a 1-min denaturation at 95°C, a 2-min annealing at 55°C, and a 2-min extension at 72°C, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min.

The following primers were used: circPDIA4 (F: 5′-AGAGTGACCCAGCCTACCA-3′; R: 5′-GGCTGCATTACAACCAACT-3′); miR-9-5p (F: 5′-ACGCCGTCTTTGGTTATCTAGCT-3′; R: 5′-CAGTGCAGGGTCCGAGGT-3′); PDIA4 (F: 5′-GGCAGGCTGTAGACTACGAG-3′; R: 5′-TTGGTCAACACAAGCGTGACT-3′); SP1 (F: 5′-TGGCAGCAGTACCAATGGC-3′; R: 5′-CCAGGTAGTCCTGTCAGAACTT-3′); GAPDH (F: 5′-GGAGCGAGATCCCTCCAAAAT-3′; R: 5′-GGCTGTTGTCATACTTCTCATGG-3′); and U6 (F: 5′-CCATAATTCTACATCCAGCAA-3′; R: 5′-GGTATCCATCCCCTCAAGC-3′).

GAPDH served as the internal control for mRNA and circRNA. U6 served as the internal control for miRNAs.

Nuclear and cytoplasmic RNA were purified using the Ambion PARISTM Kit (Ambion, Austin, TX, USA). Following cell lysis with fractionation buffer, centrifugation (500 × g for 5 min) separated the supernatant for cytoplasmic RNA analysis and the pellet for nuclear RNA detection. Rt-qPCR was used to analyze the expression of circPDIA4, U6 (nuclear endogenous control), and GAPDH (cytoplasmic endogenous control).

CircRNA stability assay

For RNase R treatment, total RNA (10 μg) was incubated with or without 20 U/mL of RNase R(Lucigen, Middleton, WI, USA) for 15 min at 37°C. For blocking transcription, 5 μg/mL actinomycin D (Sigma, St. Louis, MO, USA) or NC DMSO (Sigma, St. Louis, MO, USA) was added into cell culture medium. After treatment with actinomycin D and RNase R, RT-qPCR was carried out to determine the expression levels of PDIA4 and circPDIA4 in HCT116 and HT29 cells.

RNA immunoprecipitation (RIP) experiment

To determine the interaction between circPDIA4 and miR-9-5p, the Magna RIP™ RNA-binding protein immunoprecipitation kit (Millipore, Burlington, MA, USA) was used. Tumor cells were lysed using lysis buffer and the extracted cells were then incubated at 4°C with magnetic beads conjugated with antibodies against IgG or Argonaute 2 (AGO2) from Millipore. RNA was purified using TRIzol reagent (Invitrogen) and subsequently analyzed by rt-qPCR.

Chromatin immunoprecipitation (ChIP) PCR

ChIP assay was performed using a ChIP kit (Millipore) following the manufacturer’s protocol. First, after being cross-linked with 1% formaldehyde, HCT116 or HT29 cells (1 × 10⁷) were sonicated at 30% maximum power for 8 min (5 s pulse after every 10 s). Supernatants were transferred into a new tube for immunoprecipitation with 1 μg of anti-SP1 (ab124804; Abcam, Cambridge, MA, USA) or IgG antibody after centrifugation at 15,000 × g for 10 min. The target protein and the binding DNA complex was sedimented using prewashed agarose beads (cat# 17–0963-03; GE Healthcare, Chicago, IL, USA). After elution and purification, DNA was analyzed by rt-qPCR. The following primers were used in ChIP-PCR: PDIA4 (F: 5′-TGTTCTTACTGCCAAGTTCGG-3′; R: 5′-GGCGCTATTGGCAATCGGT-3′).

Western blot

Cell lysis was achieved using radioimmunoprecipitation assay buffer (Beyotime, Shanghai, China). The protein concentration was determined using the bicinchoninic acid method. Proteins were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membranes (Beyotime). The membranes were then blocked with 5% skimmed milk. The blots were incubated at 4°C overnight with primary antibodies against SP1 (ab124804, 1:1,000, Abcam) and β-actin (ab8226, 1:1,000, Abcam). The membranes were subsequently incubated with an HRP-labelled secondary antibody for an additional 1 h. Signal visualization was achieved using an enhanced chemiluminescence kit (Millipore).

Cell counting kit 8 (CCK-8) assay

Cell viability was assessed using the CCK-8 assay (Dojindo Laboratories, Kumamoto, Japan) following the provided instructions. CRC cells (1 × 10⁴ cells/well) were seeded in a 96-well culture plate. At 0, 24, 48, 72, and 96 h, 10 μL of CCK-8 solution was added to each well and incubated for an additional 4 h. The absorbance at 450 nm was measured using a spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).

Transwell assay

Transwell assays were performed to evaluate the migratory and invasive capabilities of cells. Chambers with 8-μm polycarbonate membranes (Millipore) were utilized. For invasion assessments, filters were pre-coated with 500 ng/mL of Matrigel solution (BD Biosciences, San Jose, CA, USA) at 37°C for 4 h. The upper chamber was seeded with 200 μL of serum-free DMEM medium containing 1 × 10⁵ cells, while the lower chamber was filled with 500 μL of DMEM medium with 20% FBS. After 18 h, cells on the upper surface were removed using cotton swabs and cells that had migrated or invaded to the bottom of the membrane were fixed and stained with 0.5% crystal violet (Beyotime). Cell counting was performed under a light microscope (Olympus, Tokyo, Japan) in five random fields. Similar procedures were followed for the Transwell migration assays, excluding the Matrigel coating step.

Luciferase reporter assay

To investigate the interaction between miR-9-5P and circPDIA4 or SP1, sequences of circPDIA4 and SP1, along with the corresponding mutated sequences, were designed. These sequences were then inserted into the luciferase reporter, pmirGLO (Promega, Madison, WI, USA). The resulting plasmids were co-transfected with miR-NC or miR-9-5P mimic into HCT116 and HT29 cells. The Dual Luciferase Assay Kit (Promega) was used to measure the relative luciferase activity 48 h post-transfection.

To validate the interaction between SP1 and the PDIA4 promoter, the PDIA4 promoter was cloned into the pGL3 luciferase reporter vector. Subsequently, it was co-transfected into HCT116 and HT29 cells along with pcDNA3.1/SP1 or vector.

Animal experiment

Six-week-old male BALB/c nude mice (n = 6) were obtained from Vital River Company (Beijing, China) and were utilized in subcutaneous injection experiments with 5 × 10⁶ HCT116 cells into the right flanks. This study received approval from the Animal Care and Use Committee of Tianjin Medical University Cancer Institute & Hospital, adhering to the guidelines of the National Institutes of Health. The mice were randomly divided into two groups, each comprised of three mice. After 1 month, the mice were euthanized and measurements of tumor weight, volume, and size were obtained to determine the impact of the study interventions on tumor growth and development.

Immunohistochemistry (IHC)

Tumor tissues were fixed in 10% formalin and sectioned into 4-μm slices. The sections were blocked with rabbit serum and incubated overnight at 4°C with primary antibodies against Ki67 (ab16667, 1:200, Abcam) and CD31 (ab9498, 1:2,000, Abcam). Subsequently, the sections were incubated with the secondary antibody [goat anti-rabbit IgG H&L (HRP)] (ab7090, 1:1,000, Abcam), followed by diaminobenzidine (DAB) staining and hematoxylin re-dyeing. Slide observations were conducted using a microscope (Nikon, Tokyo, Japan).

Statistical analysis

Statistical analysis was performed using GraphPad Prism (version 8.0; GraphPad Software, La Jolla, CA, USA). Data are presented as the mean ± standard deviation. Pearson analysis was used for correlation analysis. Student’s t-test was used for comparing two groups and one-way analysis of variance was applied for comparisons between multiple groups. A P < 0.05 was considered statistically significant.

High PDIA4 mRNA expression facilitates circRNA-mediated pathologic functions in CRC through an ER-independent pathway

To discern the involvement of PDI genes in CRC, we analyzed PDIA1-6 mRNA expression within the TCGA-COADREAD dataset. The analysis indicated that in addition to PDIA2, PDIA mRNAs were predominantly upregulated in CRC tissues (Figure 1A). To determine if this overexpression translated to enhanced protein folding and disulfide isomerase activity, single-sample gene set enrichment analysis (ssGSEA) was executed to examine the enrichment of pertinent gene sets. This revealed a pattern of substantial enrichment for gene signatures linked to protein folding and disulfide isomerase activity in CRC (Figure 1A) with mild associations for PDIA1, 2, and 6 (Figure 1B and Figure S1A, B, F), while other PDIA mRNAs did not significantly correlate with these gene sets (Figure 1B and Figure S1C-E). Further examination into protein levels indicated that there was no marked overexpression of PDIA proteins in CRC except PDIA3 (Figure 1C). This suggests that the overexpressed PDIA mRNAs may not translate into functional proteins or be involved in ER-associated protein folding processes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

PDIA mRNA and circPDIA4 expression patterns in CRC tissues and cells. (A) Heatmap illustrating PDIA1-6 mRNA expression and enrichment of protein folding-related signatures (GOBP_PROTEIN_FOLDING and GOMF_PROTEIN_DISULFIDE_ISOMERASE_ACTIVITY, https://www.gsea-msigdb.org) detected through single-sample GSEA in the TCGA-COADREAD dataset (https://portal.gdc.cancer.gov/). (B) Radar chart depicting the correlation values between PDIA1-6 mRNA expression and the enrichment score of protein folding-related signatures in the TCGA-COADREAD dataset. (C) Level of PDIA1-6 protein expression in normal (n = 3) and CRC tissues (n = 10). The foundational tissue microarray data and representative IHC images were obtained from the Human Protein Atlas database (https://www.proteinatlas.org/). Quantification of the bottom graph was performed using Image J software. Data are presented as the mean ± SD. *P < 0.05. (D) Volcano plot showing the differentially expressed circRNAs in CRC tissues compared to matched adjacent tissues in the GSE205643 dataset. (Red: upregulated genes in CRC; blue: downregulated genes in CRC; yellow: circPDIA4). (E) Venn diagram illustrating the overlap between differentially expressed circRNAs in the GSE205643 dataset and circRNAs derived from PDIA1-6 transcripts (http://www.circbase.org/). CircPDIA4 is back-spliced by PDIA4 exons 2, 3, 4, 5, 6, and 7. (F) CircPDIA4 expression levels were detected by rt-qPCR in freshly collected CRC tissues (n = 20) and matched normal tissues (n = 20). Data are presented as the mean ± SD. ***P < 0.001. (G) CircPDIA4 expression was detected in NCM460, HCT116, HT29, SW480, and RKO cells through rt-qPCR. Data are presented as the mean ± SD. ***P < 0.001.

Evidence points to untranslated mRNAs potentially playing roles beyond coding by forming circular RNAs (circRNAs) via alternative splicing¹⁵. To determine if untranslated PDIA transcripts generate circRNAs implicated in CRC pathology, a circRNA expression dataset (GSE205643) derived from normal and CRC tissues was analyzed, which showed that among 448 significantly different circRNAs in CRC (P < 0.05; Figure 1D), only circPDIA4 originated from PDIA1-6 genes (Figure 1E). To substantiate these results, 20 pairs of matched normal and CRC tissue samples were collected from our facility and rt-qPCR was used to quantify circPDIA4 expression in these samples and in four CRC cell lines (HT29, HCT116, SW480, and RKO) and the normal colonic mucosal cell line, NCM460. CircPDIA4 was significantly upregulated in CRC tissues (Figure 1F) and cell lines (Figure 1G). Moreover, circPDIA4 displayed typical circRNA characteristics, such as enhanced stability (Figure S2A) and resistance to RNase R digestion (Figure S2B), unlike PDIA4 mRNA.

To clarify the biological impact of circPDIA4 on CRC, shRNA-mediated knockdown was used in HCT116 and HT29 cell lines (Figure S3), which resulted in decreased cell proliferation (Figure 2A, B), migration (Figure 2C), and invasion (Figure 2D). These results implied that elevated PDIA4 mRNA, irrespective of the role in protein folding within the ER, fosters circPDIA4 formation, thereby facilitating CRC malignant evolution.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Cellular functions of circPDIA4 in CRC cells. The impact of circPDIA4 depletion on the cell growth of HCT116 (A) and HT29 (B) cells was assessed using the CCK-8 method. The influence of circPDIA4 depletion on CRC cell migration (C) and invasion (D) was evaluated through Transwell assays without or with Matrigel coating. The cells were photographed and counted in five random distinct fields. Data are presented as the mean ± SD. Scale bar = 50 μm; ***P < 0.001.

CircPDIA4 upregulates SP1 expression by sequestering miR-9-5p

CircRNAs display diverse cellular distributions and molecular functions. In the nucleus circRNAs modulate gene expression by interacting with RNA polymerase II, altering splicing patterns, or regulating transcription^16,17. In the cytoplasm circRNAs act as microRNA (miRNA) sponges, affect mRNA stability, regulate translation, or modify protein functions through direct interactions^17,18. To ascertain the cellular localization of circPDIA4 in CRC cells, nuclear and cytoplasmic RNA were isolated from two CRC cell lines (HCT116 and HT29) and measured circPDIA4 levels (Figure S4). Our findings indicated that > 77% of circPDIA4 is localized in the cytoplasm (Figure 3A). To probe the molecular role of circPDIA4 in CRC cell cytoplasm, RNA immunoprecipitation (RIP) was used to investigate the association with AGO2, a central component of the RNA-induced silencing complex (RISC), suggesting circPDIA4 function as a miRNA sponge¹⁹. The majority of circPDIA4 was shown to be associated with AGO2 proteins (Figure 3B). Computational simulations using databases, such as circBank, starBase, and circAtlas, revealed that circPDIA4 likely regulates three miRNAs (miR-9-5p, miR-942-5p, and miR-3944-3p) (Figure 3C). Comparative base pairing analysis showed that miR-9-5p forms a robust ‘8 mer’ complementarity with circPDIA4, suggesting a stronger binding and regulatory potential compared to miR-942-5p and miR-3944-3p, both of which exhibit a ‘7 mer-m8’ pattern with a mismatch at the 8th position. Consequently, we are now focusing corollary investigations on miR-9-5p due to its stronger and more stable binding affinity, which forms an 8 mer binding site^20,21. Experimental evidence corroborated this, showing that circPDIA4 knockdown results in a notable increase in miR-9-5p expression (Figure 3D). Next, to further validate the specific interaction between circPDIA4 and miR-9-5p, a luciferase reporter gene assay system was constructed containing both wild-type and mutated miRNA binding sites of circPDIA4 (Figure 3E), which were designed based on the binding site indicated in Figure 3B. The results showed that mutating the binding site effectively prevented the reduction in luciferase activity typically caused by miR-9-5p overexpression (Figure 3F). Contrary to the circPDIA4 expression pattern in CRC tissues and cells (Figure 1F, G), miR-9-5p was significantly downregulated in CRC tissues and cell lines (Figure 3G, H). These observations implied that miR-9-5p, targeted by circPDIA4, may have a critical role in CRC malignancy.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Identification of circPDIA4-targeting miRNAs. (A) Schematic illustrating the intracellular distribution of circPDIA4 in CRC cells (HCT116 and HT29). Green, cytoplasm; red, nucleus. (B) The RNA bound to the AGO2 protein, obtained through RIP, was used as a template for specific circPDIA4 rt-qPCR to verify its interaction with the AGO2 protein. (C) A Venn diagram depicts the common targets of circPDIA4 identified across circBank, starBase, and circAtlas databases. Additionally, the predicted binding sequence of circPDIA4 with miRNA, as per the starBase database, is provided. (D) The level of miR-9-5p expression in circPDIA4-depleted and control CRC cells was evaluated using rt-qPCR. (E) Schematic illustrating a luciferase reporter system containing either wild-type (WT) or mutated (MUT) miR-9-5p binding sites on circPDIA4. (F) The luciferase reporter experiment assessed the binding interaction between miR-9-5p and circPDIA4 in CRC cells by calculating the relative fluorescence intensity. (G) Level of miR-9-5p expression was detected by rt-qPCR in freshly collected CRC tissues (n = 20) and matched normal tissues (n = 20). (H) miR-9-5p expression was measured in NCM460, HCT116, HT29, SW480, and RKO cells through rt-qPCR. Data are presented as the mean ± SD. ***P < 0.001.

To elucidate the molecular mechanisms underlying the progression of CRC malignancy mediated by miR-9-5p, a comprehensive analysis of the downstream targets was performed using various databases, including starBase, miRDIP, miRMap, and miRWalk. Our predictions identified seven downstream target genes (Figure 4A), including KIF1B, MTMR3, IGF1, and SP1, all of which have previously been implicated in tumor progression^22–25. Notably, potential targets of circPDIA4, such as miR-942-5p and miR-3944-3p (Figure 3C), were also predicted to target SP1 mRNA (Figure S5A). Upon analysing the convergence of the targets for these three miRNAs, SP1 was shown to be the only common target (Figure S5B). Therefore, SP1 was further investigated as a downstream candidate of the circPDIA4/miRNA-9-5p interaction. Furthermore, an AGO2-based RIP experiment in CRC cells supported the enrichment of miR-9-5p and SP1 mRNA within the RISC (Figure 4B), suggesting a potential downregulation of SP1 expression. Indeed, treatment with a miR-9-5p mimic significantly reduced SP1 mRNA (Figure 4C) and protein levels (Figure 4D and Figure S5B) in CRC cells, a finding that was consistent with the depletion of circPDIA4 (Figure S5C, D). To validate the direct silencing effect of miR-9-5p on SP1, a luciferase reporter system containing the wild-type or mutant miR-9-5p targeting site within the SP1 3′UTR was established (Figure 4E). The results demonstrated that treatment with the miR-9-5p mimic significantly reduced luciferase activity based on the wild-type SP1 3′UTR, whereas the mutant type showed no such effect (Figure 4F). Additionally, in agreement with the circPDIA4 expression pattern (Figure 1F, G), SP1 displayed significant upregulation in both CRC tissues (Figure 4G) and cells (Figure 4H). To validate the pathologic function of the circPDIA4/miR-9-5p/SP1 regulatory axis in CRC cells, rescue experiments involving miR-9-5p/SP1 on circPDIA4-depleted CRC cells were performed. Both miR-9-5p antagomir treatment and overexpression of SP1 rescued the inhibitory effects on cell growth (Figure 5A, B), migration (Figure 5C), and invasion (Figure 5D) caused by circPDIA4 depletion. These findings strengthen our understanding of the circPDIA4/miR-9-5p/SP1 regulatory axis as a crucial modulator in CRC progression.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Identification of miR-9-5p-targeting mRNAs. (A) Venn diagram illustrating the shared targets of miR-9-5p identified across the miRDIP, miRMap, miRWalk, and starBase databases. (B) The RNA bound to the AGO2 protein, obtained through RIP, was used as a template for specific rt-qPCR targeting miR-9-5p and SP1 mRNA to verify their interaction with the AGO2 protein. (C) The depletion of SP1 mRNA expression by miR-9-5p in CRC cells was evaluated using rt-qPCR. Cellular mRNA treated with the miR-9-5p mimic or a non-specific control (NC) mimic served as the template. (D) The downregulation of SP1 protein by miR-9-5p in CRC cells was assessed using western blot. (E) Schematic illustrating a luciferase reporter system containing the wild-type (WT) or mutant (MUT) miR-9-5p targeting site within the SP1 3′UTR. (F) The luciferase reporter experiment assessed the binding interaction between miR-9-5p and SP1 3′UTR in CRC cells by calculating the relative fluorescence intensity. (G) SP mRNA expression was detected by rt-qPCR in freshly collected CRC tissues (n = 20) and matched normal tissues (n = 20). (H) SP mRNA expression was measured in NCM460, HCT116, HT29, SW480, and RKO cells through rt-qPCR. Data are presented as the mean ± SD. ***P < 0.001.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Validation of cellular functions of circPDIA4/miR-9-5p/SP1 axis in CRC cells. We investigated the effects of the circPDIA4/miR-9-5p/SP1 regulatory axis on the growth of colorectal cancer (CRC) cell lines, including HCT116 (A) and HT29 (B). Using the CCK-8 assay, we measured cell growth under several conditions: in control cells; in cells depleted of circPDIA4; in circPDIA4-depleted cells treated with a miR-9-5p antagonist (referred to as miR-9-5p-I); and in circPDIA4-depleted cells with SP1 overexpression. Additionally, we evaluated the role of the circPDIA4/miR-9-5p/SP1 axis in CRC cell migration using Transwell assays (C), and in cell invasion using Matrigel-coated Transwell assays (D). The cells were photographed and counted in five random distinct fields. Data are presented as the mean ± SD. Scale bar = 50 μm; ***P < 0.001.

SP1 induction of PDIA4 transcription regulates circPDIA4 expression

As a transcription factor, SP1 is integral in modulating genes responsible for cell growth, differentiation, and apoptosis²⁶. High SP1 expression in CRC is correlated with increased tumor proliferation, invasion, and metastasis²⁷. To explore SP1 downstream targets, the Sidra-LUMC AC-ICAM CRC patient transcriptomic database was analyzed²⁸. Our analysis revealed a significant positive correlation between SP1 and genes linked to CRC malignancy, such as TUG1, CCR5, YY1, and IGF1R^24,29–31 (Figure 6A), as well as PDIA4 (Figure 6A, B). This finding suggested a potential amplifying loop in which SP1 and circPDIA4 synergistically activated oncogenic gene expression, thereby promoting CRC progression. Moreover, a Jaspar database analysis (https://jaspar.elixir.no/) identified a specific SP1 binding motif within the 1,629 base pairs upstream of the PDIA4 transcription start site (Figure 6C). This motif functionality was confirmed via SP1 antibody-based ChIP-PCR in CRC cells (Figure 6D). To examine PDIA4 transcriptional activation by SP1, a luciferase reporter assay was performed with the PDIA4 promoter and the wild type or a mutated SP1 binding site (Figure 6E). SP1 overexpression significantly enhanced luciferase activity for the wild-type promoter but not for the mutant (Figure 6F), confirming SP1 transcriptional regulation of PDIA4 through the binding site. Interestingly, as shown in Figure 6A, when analysing genes co-expressed with SP1, a positive correlation was also demonstrated between the expression of QKI and SP1, albeit not a significant correlation. Previous research has reported QKI as a key catalytic gene promoting the formation of circPDIA4 from PDIA4¹³. This discovery was an incentive to further investigate the relationship between SP1 and circPDIA4. To determine whether SP1-induced PDIA4 transcription leads to circPDIA4 formation, SP1 and circPDIA4 expression was measured in 20 CRC tissue samples. A strong positive correlation (r = 0.82) was observed between the circPDIA4 and SP1 levels (Figure 6G). Additionally, SP1 knockdown markedly reduced circPDIA4 levels in CRC cells (Figure 6H). These findings indicated that SP1 not only modulates PDIA4 transcription but may also influence the biogenesis of circPDIA4 biogenesis, contributing to the oncogenic pathway in CRC.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Elucidation of positive feedback regulation of PDIA/cirPDIA4 by SP1. (A) Ranking graph displaying mRNAs co-expressed with SP1 mRNA in the CRC database (Reference: Sidra-LUMC AC-ICAM). (B) Dot plot illustrating the correlation between SP1 and PDIA4 mRNA expression in the CRC database (Sidra-LUMC AC-ICAM). (C) The predicted SP1 binding site on the PDIA4 promoter, as identified by the JASPAR database, was shown. (D) The binding capability of the SP1 protein to the promoter region of PDIA4 was validated by ChIP-PCR. (E) Schematic illustrating a luciferase reporter system containing either wild-type (WT) or mutated (MUT) SP1 binding sites on the PDIA4 promoter. (F) The luciferase reporter assay was conducted to evaluate the impact of SP1 on the transcriptional activity of the PDIA4 promoter in CRC cells by calculating the relative fluorescence intensity. (G) Dot plot illustrating the correlation between SP1 mRNA and circPDIA4 expression in the CRC by rt-qPCR. (H) RT-qPCR was used to assess the expression of circPDIA4 following SP1 depletion in CRC cells. Data are presented as the mean ± SD. ***P < 0.001.

Blocking the circPDIA4/miR-9-5p/SP1 regulatory axis repressed tumorigenicity of CRC cells

To elucidate the role of the circPDIA4/miR-9-5p/SP1 axis in CRC progression, an in vivo study was conducted. CRC cells with suppressed circPDIA4 expression were injected subcutaneously into nude mice. After a 1-month observation period, the mice were euthanized and the resulting tumors were analyzed for size, volume, and weight. The results, as depicted in Figure 7A-C, revealed that depletion of circPDIA4 led to a significant decrease in tumor volume and weight. Further molecular analysis of the tumor tissue indicated an increase in miR-9-5p levels and a decrease in SP1 expression due to the absence of circPDIA4 (Figure 7D). Additionally, a reduction in the proliferative marker, Ki67, was observed in the tumors (Figure 7E). These findings from in vivo experiments support the conclusion that the circPDIA4/miR-9-5p/SP1 regulatory axis has a crucial role not only at the cellular level but also in the actual progression of CRC in a living organism. Moreover, pan-cancer analysis showed that PDIA4 and SP1 were highly expressed in 22 of 33 cancer types compared with normal tissues (Figure S6). Therefore, we can conclude that the regulatory effect of circPDIA4/miR-9-5p/SP1 axis is pan-carcinogenic.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

Depletion of circPDIA4 in CRC cells repressed the in vivo tumorigenicity by increasing miR-9-5p expression and suppressing SP1 expression. (A-C) Representative images (A) of tumors were obtained 1 month post-injection of HCT116 cells—either circPDIA4-depleted or control—into the subcutaneous tissue of nude mice. Subsequently, tumor volumes (B) and weights (C) were quantified. Tumor volume calculation formula: V = π/6 × L × W². (D) RT-qPCR was used to assess the levels of circPDIA4, miR-9-5p, and SP1 expression in tumor tissues. (E) Immunohistochemistry (IHC) was performed to detect proliferating cells (Ki67-positive) in tumor tissues. Scale bar = 80 μm. (F) Schematic diagram illustrating the PDIA4/SP1 positive feedback loop involving circPDIA4/miR-9-5p in CRC. Data are presented as the mean ± SD. ***P < 0.001; **P < 0.01.