Human liver cancer cell line HepG2 (HB-8065, ATCC, Manassas, VA, USA) were cultured in DMEM (11965092, Thermo Fisher Scientific, Waltham, MA, USA) containing 10% Fetal Bovine Serum (FBS, 10091, Thermo Fisher, USA), 1% Penicillin-Streptomycin Solution (15140-122, Thermo Fisher, USA) in a humidified atmosphere with 5% CO₂ at 37 °C, and validated by short tandem repeat (STR) profiling and tested negative for mycoplasma contamination using MycoSensor PCR assay kit (Agilent, Hangzhou, China).

Human shPDGF-D (NM_025208.5, target sequence: 5^′-ATCAAGAACGAACCAAATTAA-3^′) lentiviral vector and its negative control (NC, target sequence: 5^′-TTCTCCGAACGTGTCACGTCC-3^′) were purchased from VectorBuilder (Guangzhou, China). In short, the vector was transfected into 293T cells with Lipofectamine™ 3000 (L3000150, Thermo Fisher, USA), and the transfection efficiency was assessed using a fluorescence microscope (BX53, Olympus, Tokyo, Japan) 24 hours after transfection. The cell supernatant was collected and filtered, and the filtered virus was then infected with HepG2 cells. After 48 hours, the expression of PDGF-D was detected by quantitative real-time polymerase chain reaction (qRT-PCR) and Western blotting.

Briefly, total RNA was extracted by Trizol reagent (B511311, Sangon Biotech, Shanghai, China). Next, the PrimeScript RT reagent kit (Takara, Shiga, Japan) was used to synthesize the cDNA of each sample. The expression of PDGF-D, $\alpha$-SMA, and FAP in cells was detected by a Real-Time PCR system (7500, ThermoFisher, USA) using Probe qPCR Mix (D7271, Beyotime, Shanghai, China). GAPDH was used as the internal reference. The relative expression levels were quantified by the 2^{-Δ⁢Δ⁢CT} approach. The sequences of the primers used are displayed in Table 1.

Total protein was extracted using RIPA Lysis Solution (P0013K, Beyotime, China). The extracted protein has gone through the process of electrophoresis. After incubating with the primary antibody overnight at 4 °C, the membrane was washed with TBST buffer and incubated with the secondary antibody for 2 hours at room temperature. The protein bands were visualized by ECL Western Blotting Substrate (PE0010, Solarbio, Beiijng, China) and quantified using the Image J software (1.52v, NIH, Bethesda, MD, USA), with GAPDH serving as the internal control. The primary and secondary antibodies were as follows: PDGF-D (DF12690, 1:1000, Affinity, West Bridgford, UK); CD63 (1:1000, ab217345, Abcam, Fremont, CA, USA); CD81 (1:1000, ab109201, Abcam); CD9 (1:1000, ab92726, Abcam); $\alpha$-SMA (1:1000, AF1032, Affinity), FAP (1:2000, AF0739, Affinity); matrix metallopeptidase-9 (MMP-9, AF5228, 1:2000, Affinity); vascular endothelial growth factor (VEGF) (1:1000, AF5131, Affinity); GAPDH (1:10,000, ab8245, Abcam); goat anti-rabbit antibody (1:5000, ab205718, Abcam); goat anti-mouse antibody (1:5000, ab6789, Abcam).

The separation of exosomes is described previously by sequential centrifugation [13]. In short, the HepG2 cells were cultured in a DMEM medium containing 10% exosomal-free FBS. After 48 hours, the supernatants from cells were centrifuged at 300 g for 10 minutes to discard cells, and then the supernatant was centrifuged at 16,500 g for 20 minutes to remove cell debris. The supernatant was filtrated using a 0.22 µm syringe filter and then centrifuged at 120,000 g (Beckman Type 90 Ti) for 70 minutes to precipitate exosomes. Finally, the exosome pellet was resuspended in PBS or lysis buffer before further analysis. In order to verify the exosomes, a nanoparticle tracking analysis (NTA, ZetaView PMX 120) was carried out to estimate the size distribution and nanoparticle concentration of the exosomes, the morphology of the exosomes was observed by using a transmission electron microscope (TEM, JEM-1400, JEOL, Akishima, Japan), and the specific markers of the exosomes (CD63, CD81 and CD9) were detected by western blot.

BALB/C mice (20–30 g) were acquired from Vital River (Beijing, China) and kept in a constant temperature (21 $\pm$ 0.5 °C) room with a natural circadian alternation and 50% humidity. After the mice were anesthetized with 1% sodium pentobarbital (50 mg/kg, P-010, Merck, German), the liver tissues of the mice were collected, cut into small pieces, and digested with type I collagenase (SCR103, Merck, German) at 37 °C for 2 hours. After centrifugation, the precipitate was resuspended with a complete medium and inoculated into 10 cm dishes.

All animal experiments were permitted by the Animal Experimentation Ethics Committee of Hangzhou Eyong Pharmaceutical Research and Development Center (Approval number: EYOUNG-20201030-02) and in accordance with the guidelines of the National Institutes of Health Laboratory Animal Care and Use Guidelines.

The fibroblasts were placed in a 6-well plate at a density of 1 $\times$ 10⁵ per well supplemented with 10% FBS-containing DMEM. After culturing for 24 hours, 50 µL of HepG2 shNC-exosome or HepG2 shPDGF-D-exosome medium with a concentration of 400 µg/µL was added to the 6-well plate for co-culture for 48 hours. For exosome uptake assays, the extracted exosomes were labeled using PKH67 (green) in line with the manufacturer’s methods and co-cultured with fibroblasts at a concentration of 200 µg/mL for 48 h. Visualization of exosome uptake was observed using a laser scanning confocal microscope (LSM880, Zeiss, Jena, Germany).

Fibroblasts (2 $\times$ 10³ per well) were added into 96-well plates, followed by the co-culture treatment with the exosomes (50 µg/mL) obtained from HepG2 with PDGF-D knockdown. After co-culture for 24, 48, 72, and 96 hours, 10 µL CCK-8 reagent (C0037, Beyotime, China) was added to each well and cultured for 2 h at 37 °C. The absorbance was measured by a microplate reader (ThermoFisher) at a wavelength of 450 nm.

The proliferating fibroblasts were determined by the EdU Kit (C0075, Beyotime, China). In short, the prepared EdU working solution was added to the above-mentioned 6-well plate for 48 hours of co-cultivation and incubated for 2 hours, then the culture solution was removed and 1 mL of fixing solution was added for fixation for 15 minutes. Then, after removing the fixing solution, cells were reacted with the permeating solution for 15 minutes and with a click reaction solution for 30 minutes. Then Hoechst 33342 solution was added to stain the nucleus. Finally, the results were observed under a fluorescence microscope (BX63, Olympus, Japan).

To detect the expression of E-cadherin, vimentin, $\alpha$-SMA, and FAP, cells or tissues were fixed with 4% paraformaldehyde (E672002, Sangon, China), permeabilized with Triton-X-100 (A110694, Sangon), followed by blocking. Subsequently, cells were reacted with the primary antibody E-cadherin (1:100, ab231303, Abcam, UK), vimentin (1:200, ab92547, Abcam, UK), $\alpha$-SMA (1:200, AF1032, Affinity, USA), FAP (1:200, AF0739, Affinity, USA) and CD31 (1:500, AF6191, Affinity, USA) overnight at 4 °C. The next day, cells were reacted with goat anti-rabbit antibody Alexa Fluor® 647 (ab150079, Abcam, UK) or Alexa Fluor® 488 (ab150077, Abcam, UK). The nucleus was stained with DAPI (E607303, Sangon) for 15 minutes. Finally, the images were captured using the fluorescence microscope and analyzed by using ImageJ software v2.0.0 (NIH, Bethesda, MD, USA).

Cell Cycle Kit (C1052, Beyotime, China) was used to detect the cell cycle. In short, the co-cultured cells were collected and centrifuged at 1000 g for 5 minutes. After removing the supernatant, the cells were resuspended with PBS (E607008, Sangon, China) and transferred to the centrifuge tube. Precooled 70% ethanol was added and incubated at 4 °C for 30 minutes. After washing, the configured propidium iodide (PI) staining solution was added and incubated in the dark for 30 minutes. Finally, the cell cycle was detected by flow cytometry.

A long scrape was created by a pipette, and PBS was used to remove detached cells on the surface. Afterward, the cells were incubated with DMEM in the incubator for 24 hours. The wounds were photographed at 0 and 48 hours by a microscope (BX53M, Olympus, Japan).

BALB/C nude mice (20–30 g) were also purchased from Vital River, and housed in a constant temperature (21 $\pm$ 0.5 °C) room with a natural circadian alternation and 50% humidity. Mice can eat and drink freely, and begin the experiment about 7 days after acclimatization. Nude mice were randomly divided into 2 groups, three in each group. Fibroblasts (2 $\times$ 10⁶) co-cultured with HepG2 shNC-exosome or HepG2 shPDGF-D-exosome and HepG2 cells (3 $\times$ 10⁶) were mixed in 200 µL PBS and then injected subcutaneously into mice. The volume of the subcutaneous tumor was measured with a caliper every three days and calculated using the formula: length $\times$ width²/2. Four weeks later, the mice were euthanized by intraperitoneal injection of 250 mg/kg sodium pentobarbital [14]. The tumors were collected for follow-up immunofluorescence staining, qRT-PCR, and western blot.

Statistical analysis was performed by GraphPad Prism 8.0 (GraphPad Software, Inc., San Diego, CA, USA). The results were expressed by mean $\pm$ standard deviation (SD). The comparisons between the two groups were analyzed using a student’s t-test. Differences with p 0.05 were considered statistically significant.

To evaluate the effect of PDGF-D, we transfected HepG2 cells with shPDGF-D, and the transfection efficiency was shown in Fig. 1A,B (p 0.01). We determined the diameter and concentration of exosomes through nanoparticle tracking analysis (NTA) (Fig. 1C) and observed the morphology of exosomes through a TEM (Fig. 1D). The NTA showed that the peak diameter of exosomes was 110 nm (Fig. 1C). At the same time, the expression of characteristic proteins of exosomes (CD63, CD81, and CD9) was confirmed by western blot. Silencing PDGF-D did not affect the expression level of characteristic proteins of exosomes (Fig. 1E). The green fluorescence in exosomes was found in the perinuclear region of fibroblasts, suggesting that PKH67-labeled exosomes could enter into the cytoplasm of fibroblasts (Fig. 1F).

Fig. 1.

Identification of exosomes in HepG2 cells with silencing of PDGF-D. (A,B) Relative expression of PDGF-D in HepG2 cells transfected with specific shRNAs against PDGF-D. Data are represented as mean $\pm$ SD (n = 3), **p 0.01, compared with the NC shRNA group. (C) nanoparticle size distribution. (D) Morphology of exosomes measured by transmission electron microscope (TEM). Scale bar, 100 nm. (E) Western blot was used to analyze the specific exosome surface markers (CD63, CD81, CD9). (F) A diagram of the uptake of exosomes in HepG2 cells with silencing of PDGF-D by fibroblasts. Scale bar, 10 µm.

The exosomes of HepG2 cells with PDGF-D knockdown were co-cultured with mouse primary cultured fibroblasts, and the effects of exosomes derived from HepG2 cells on the biological behavior of fibroblasts were evaluated through a series of functional experiments. Compared with the Exos-NC shRNA group, exosomes derived from HepG2 cells transfected with shPDGF-D inhibited the cell viability and proliferation of fibroblasts (Fig. 2A,B, p 0.01), induced the prolongation of the G0/G1 phase and the shortening of S phase (Fig. 2C, p 0.01), and reduced the migration (Fig. 3A, p 0.01).

Fig. 2.

Knockdown of PDGF-D in HepG2-derived exosomes inhibited the proliferation of fibroblasts. (A) Fibroblasts were incubated with exosomes in HepG2 cells with silencing of PDGF-D for 24, 48, 72, and 96 h, respectively. Cell viability was analyzed by cell counting kit-8 (CCK-8) assay. (B) EdU assays of Fibroblasts incubated with exosomes in HepG2 cells with silencing of PDGF-D were performed to evaluate cell proliferative ability. Scale bar, 50 μm. (C) Cell cycle distributions of fibroblasts in PDGF-D knockdown HepG2-derived exosomes were presented by flow cytometry. Results are represented as mean $\pm$ SD (n = 3). *p 0.05, **p 0.01 compared with Exos-NC shRNA group.

Fig. 3.

Knockdown of PDGF-D in HepG2-derived exosomes inhibited the migration of fibroblasts. (A) Wound healing assays were performed in fibroblasts treated with exosomes in HepG2 cells with silencing of PDGF-D. Scale bar, 200 μm. (B) Immunofluorescent analysis for E-cadherin and Vimentin was detected in fibroblasts. Nuclei were stained blue (DAPI), and E-cadherin and Vimentin were stained red, Data are shown as mean $\pm$ SD (n = 3). Scale bar, 50 μm. *p 0.05, **p 0.01 compared with the Exos-NC shRNA group.

The conversion of fibroblasts to CAFs is related to the EMT process, and activated fibroblasts have the ability to promote invasion and angiogenesis. Therefore, we evaluated the expression levels of EMT-related proteins, CAF markers, MMP-9, and VEGF. As shown in Fig. 3B, exosomes secreted by HepG2 cells transfected with shPDGF-D could significantly increase the expression of E-cadherin in fibroblasts (p 0.01), but inhibit the expression of vimentin (p 0.05). We confirmed by qRT-PCR, western blot, and immunofluorescence that exosomes secreted by HepG2 cells transfected with shPDGF-D can reduce the expression levels of $\alpha$-SMA and FAP in fibroblasts when compared with the Exos-NC shRNA group (Fig. 4A–C, p 0.01). In addition, the levels of MMP-9 and VEGF proteins were also inhibited by exosomes secreted by HepG2 cells transfected with shPDGF-D (Fig. 4B, p 0.05).

Fig. 4.

Knockdown of PDGF-D in HepG2-derived exosomes regulates $\alpha$-SMA, FAP, MMP 9, and VEGF expression in fibroblasts. (A) $\alpha$-SMA and FAP expression levels were detected after being treated with exosomes in HepG2 cells with silencing of PDGF-D by real-time polymerase chain reaction (qRT-PCR). (B) Levels of the $\alpha$-SMA, FAP, MMP 9, and VEGF protein in fibroblasts, as determined by Western blotting. (C) Immunofluorescence staining of $\alpha$-SMA and FAP in fibroblasts treated with exosomes in HepG2 cells with silencing of PDGF-D. Scale bar: 50 µm. Data are shown as mean $\pm$ SD (n = 3). *p 0.05, **p 0.01 compared with the Exos-NC shRNA group. MMP, matrix metalloproteinases; VEGF, vascular endothelial growth factor.

To further verify our conclusions, we carried out in vivo experiments. The results showed that compared with the Exos-NC group, the tumor size and volume of mice injected with HepG2 cells and fibroblasts (co-cultured with exosomes secreted by HepG2 cells transfected with shPDGF-D) were significantly reduced (Fig. 5A–C, p 0.01). Not only that, fibroblasts (co-cultured with exosomes secreted by HepG2 cells transfected with shPDGF-D) reduced the expression of CD31, vimentin, $\alpha$-SMA, FAP, MMP-9, and VEGF in tumor tissues, but elevated the levels of E-cadherin (Fig. 6A–C, p 0.01).

Fig. 5.

Knockdown of PDGF-D in HepG2-derived exosomes suppresses the growth of Liver cancer tumor in vivo. (A) Representation images of tumor tissues of nude mice. Compared with the vector group, the (B) tumor weight and (C) volume significantly reduced in Knockdown of PDGF-D in HepG2-derived exosomes-treated fibroblasts and HepG2 cells in nude mice. Data are shown as mean $\pm$ SD (n = 3). **p 0.01 compared with the Fibroblasts (Exos-NC)+Hepg2 group.

Fig. 6.

Knockdown of PDGF-D in HepG2-derived exosomes regulates CD31, E-cadherin, Vimentin, $\alpha$-SMA, FAP, MMP 9, and VEGF expression in tumor tissues. (A) Immunofluorescence staining of CD31, E-cadherin, and Vimentin in tumor tissues (magnification: $\times$400, scale bar = 20 µm). (B) $\alpha$-SMA and FAP expression levels were detected in tumor tissues by qRT-PCR. (C) Levels of the $\alpha$-SMA, FAP, MMP9, and VEGF protein in tumor tissues were detected by Western blotting. Data are shown as mean $\pm$ SD (n = 3). **p 0.01 compared with the Fibroblasts (Exos-NC)+Hepg2 group.