Two human ovarian cancer cell lines, ES-2 and SKOV3, were cultured as previously described (Bai et al., 2020).

ES-2 and SKOV3 cells were seeded in a six-well plate overnight and infected with lentiviruses expressing LV-miR-194-5p or LV-miR-194-5p-NC, LV-miR-194-5p-inhibition or LV-miR-194-5p-inhibition-NC, LV-IGF1R-RNAi or LV-IGF1R-RNAi-NC (GeneChem, Shanghai, China). After 72 h of infection, the cells were selected with 2 μg/mL puromycin for two weeks to establish stable cell lines.

Total RNA was extracted using RNA Easy Fast Tissue/Cell Kit (TianGen, Beijing, China), followed by reverse transcription into cDNA using PrimeScript™ RT reagent Kit (Takara, Dalian, China). TB Green® Premix Ex Taq™ II (Takara, Dalian, China) was used to run qPCR. The primer sequences (listed in Table 1) were designed and synthesized by Sangon Biotech (Shanghai, China). The expression of miRNA or mRNA was quantified by the 2^–ΔΔCT method, using U6 as an internal reference gene for microRNA assays and β-actin as an internal reference gene for gene assays.

The cells were lysed and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The protein was transferred to a polyvinylidene fluoride membrane and blocked using a blocking buffer. Then, the membranes were incubated with monoclonal primary antibodies respectively, including anti-insulin-like growth factor1 receptor (IGF1R; CST, USA), anti-phosphorylation of serine/threonine kinase (p-AKT; S473, CST, USA), anti-AKT (CST, USA) or β-actin (Bioss, Beijing, China). For the detection of the key genes or enzymes involved in glycolysis, we used polyclonal antibodies, including anti-pyruvate kinase M2 (anti-PKM2), anti-lactate dehydrogenase A (anti-LDHA), anti-glucose transporter 1 and 3 (anti-GLUT1, and GLUT3) (Proteintech, Wuhan, China). After washing, the membranes were incubated with horseradish peroxidase-conjugated secondary antibody (Bioss, Beijing, China) and developed using the enhanced chemiluminescence kit (NCM Biotech, Suzhou, China). The signals were captured by Amersham Imager 600 (USA). The intensity of reference gene β-actin served as the loading control.

The paraffin sections of ovarian cancer tissue were soaked in a series of xylene and graded alcohols and incubated in the boiling buffer containing ethylenediaminetetraacetic acid for antigen retrieval. To inhibit endogenous peroxidase, the sections were incubated in 3% H₂O₂ solution for 8 min and blocked with goat serum. Then the sections were incubated with a primary antibody against IGF1R (Bioss, Beijing, China), followed by incubation with a secondary antibody (Bioss, Beijing, China). After hematoxylin staining, tissue sections were soaked in an alcohol gradient, followed by soaking in xylene. The images were captured by a microscope, and the target signals were quantified by the Image-J software (USA).

The concentration of ATP was examined using the Enhanced ATP Assay Kit (Beyotime, Shanghai, China). Each well of a twelve-well plate was seeded with 8 × 10⁴ cells and lysed for 42 h. The 96-well plate was preincubated with 100 μL ATP assay buffer for 5 min, and 10 μL of supernatants were added for the determination of ATP concentration. The absorbance was read by a multifunctional microplate reader (PerkinElmer, USA) and calculated by substituting the assay values into the standard curve. The linear range of the assay was between 0.1 nM to 10 µM.

Glucose uptake was detected by Glucose Uptake Colorimetric Assay Kit (Biovision, USA). The cells were cultured overnight in a growth medium without serum and 100 μL of Krebs-Ringer-Phosphate-Hepes buffer containing 2% bovine serum albumin for 40 min. The cells were treated with 0.1 mg/mL insulin to activate the glucose transporter and 10 mM 2-deoxy-D-glucose (2-DG) for 20 min separately. To degrade endogenous NAD(P) and denature enzymes, cells were lysed in 90 μL extraction buffer at 90°C for 40 min and cooled down on the ice for 5 min. Then, neutralization buffer (12 mL) was mixed with 38 μl of Mix B and added to each well, and the absorbance was measured at 412 nm at 37°C by the multifunctional microplate reader. The standard glucose uptake curve was plotted using the standard well absorbance value, and then the glucose uptake level was calculated.

The colorimetric lactate kit (Solarbio, Beijing, China) was used to measure lactate concentration. Cells (5 × 10⁶ cells) were lysed in 1 mL of extraction buffer I through ultrasonic disruption and then centrifuged. Supernatants (700 μL) and extraction buffer II (150 μL) were mixed, followed by centrifugation at 10,000 g at 4°C for 10 min. Then, supernatants (10 μL) were mixed with assay buffer for 20 min at 37°C. The absorbance at 570 nm was read by a multifunctional microplate reader. The standard lactate solution curve was obtained by diluting 2.5, 1.25, 0.625, 0.3125, 0.15625, and 0.078 µmol/mL of the lactate standard into the 96-well plate. The linear range of the assay was between 0.05 mM to 20 mM.

Seahorse XF Glycolysis Stress Test Kit (Agilent, USA) was used to examine ECAR. Then, in the XFe24 microplate, (2 × 10⁴) cells were seeded. The cartridge sensor was hydrated overnight in a CO₂-free incubator. Before detection with seahorse, the cells were treated with 2 mM glutamine for 1 h in a CO₂-free incubator. Glucose (100 mM), oligomycin (100 µM), and 2-DG (500 mM) were added sequentially into each well and detected through Seahorse XFe24 Analyzer (Agilent, USA).

The tumor cells have a higher glucose uptake rate and produce more ATP and lactate than normal cells because of the Warburg effect. To detect the role of miR-194-5p in the Warburg effect in ovarian cancer cells, we established ovarian cancer cells with stable overexpression of miR-194-5p or its knockdown by lentivirus infection (Fig. 1A). Further measurements showed that miR-194-5p overexpression decreases ATP generation (Fig. 1B), glucose uptake (Fig. 1C), and lactate production (Fig. 1D), whereas knocked down of miR-194-5p expression exhibited the opposite effect. Consistently, examination of glycolytic metabolism showed that ECAR and maximum glycolytic capacity are significantly decreased in the cells with miR-194-5p overexpression, while downregulation of miR-194-5p expression increased ECAR and maximum glycolytic capacity (Figs. 1E and 1F). These data thus indicated that miR-194-5p suppresses the Warburg effect in ovarian cancer cells.

Many key enzymes or genes of glycolytic metabolism are altered by the Warburg effect. Therefore, we assessed the mRNA and protein expression of PKM2, LDHA, GLUT1, and GLUT3 in ovarian cancer cells when miR-194-5p was overexpressed or its expression was knocked down. PKM2 determines whether glucose is channeled into the lactate-producing pathway. LDHA is a step-controlling enzyme that controls the last step of glycolysis by mediating the interconversion of pyruvate and lactate. GLUT1 and GLUT3 are two important glucose transporters involved in glycolysis in tumor cells (Vaupel and Multhoff, 2021). We observed that miR-194-5p overexpression reduces the mRNA (Figs. 2A–2D) and protein (Fig. 2E) expression of PKM2, LDHA, GLUT1, and GLUT3, whereas knockdown of miR-194-5p expression gives the opposite results.

In our previous studies, we showed that IGF1R is the direct target of miR-194-5p (Bai et al., 2020). Immunohistochemical analysis revealed that compared with normal ovarian tissues, IGF1R expression increased in ovarian cancer tissues (Fig. 3A). To investigate the role of IGF1R in the Warburg effect, we knocked down IGF1R expression in ovarian cancer cells (Fig. 3B) and found a decrease in ATP generation (Fig. 3C), glucose uptake (Fig. 3D), lactate production (Fig. 3E) and ECAR level (Figs. 3F and 3G). In agreement with that, ovarian cancer cells with reduced IGF1R expression showed decreased expression of PKM2, LDHA, GLUT1, and GLUT3 mRNA and proteins (Figs. 3H and 3I). These results indicate that IGF1R promotes the Warburg effect in ovarian cancer cells.

IGF1R and its downstream PI3K-AKT pathway serve as important regulators in the energy metabolism in tumor cells (Amutha and Rajkumar, 2017). To clarify whether the function of miR-194-5p on the Warburg effect is related to the IGF1R/PI3K/AKT signaling pathway, we tested AKT activation when IGF1R expression was knocked-down in ovarian cancer cells. The level of p-AKT decreased in the cells with reduced expression of IGF1R (Fig. 4A). We further examined whether miR-194-5p contributes to the IGF1R/PI3K/AKT pathway and found that IGF1R expression and AKT activation decreased with miR-194-5p overexpression. In contrast, the knockdown of miR-194-5p expression increased the expression of these proteins (Fig. 4B). These results further suggest that miR-194-5p mediates the PI3K/AKT pathway by targeting IGF1R in ovarian cancer cells. Taken together, our result reveals that miR-194-5p suppresses the Warburg effect in ovarian cancer cells through the PI3K/AKT signaling pathway by targeting IGF1R (Fig. 4C).