DSPC, DSPE-PEG-2000, and DSPE-PEG2000-biotin (Avanti Polar Lipids, Alabaster, AL) were collected to be liposomes at the ratio of 9:.5:.5 mg. Next, 10 mg of lipid powder, HMME and LND solution (at the concentration ratio of 2:1) was mixed in chloroform. We evaporated the mixture on the rotary evaporator at 45°C–50°C until the dry mixed lipid thin film on the bottom of the bottle was formulated. Then, the dry lipid film was dissolved in 10 ml phosphate-buffered saline (PBS) and vortexed. The mini-extruders (Avanti Polar Lipids, Alabaster, AL, United States) were used to prepare nano-level mixture lipid NBs. The HMME-LND mixed lipid solution was transferred into a sealed bottle in which air was replaced with C₃F₈ (Research Institute of Physical and Chemical Engineering of Nuclear Industry, Tianjin, China). After oscillation, the HMME-LND vesicles (HMME-LND@C₃F₈-NBs) were ready for characterization. HMME @C₃F₈-NBs were prepared in the same method, and the only difference is the absence of LND in the complex.

Scanning electron microscopy (SEM, Hitachi SU5000, Japan) and transmission electron microscopy (TEM, Hitachi TEM system, Japan) were applied to observe the HMME-LND@C₃F₈-NBs after fabrication. Dynamic light scattering (DLS, Zetasizer Nano ZS90, Malvern Instruments, United Kingdom) measured the size distribution and the zeta potential. Absorption spectra were recorded by the let–visible spectrophotometer (UV-Visible Spectrophotometer, Thermo Evolution 201, America).

The Institute of Cancer Research affiliated with the Harbin Medical University provided the HCC cell lines (Huh7 and HepG2), which the Ethics Committee approved. Dulbecco’s Modified Eagle Medium (DMEM, Hyclone, Logan, UT, United States) added 12% fetal bovine serum (FBS, Gibco, Carlsbad, CA) to cultivate HepG2 and Huh7 cells. The cells were placed in the incubator with 5% CO₂ at 37°C.

For the assay, 1 × 10⁴ cells of HepG2 and Huh7 were seeded into 96-well plates incubated for 24 h separately. Afterward, the fresh DMEM mixed with different concentrations of LND, HMME, HMME@C₃F₈-NBs, and HMME-LND@C₃F₈-NBs replaced the culture medium. US (1 MHz, 3.5 W/cm²) irradiated the cells incubated with HMME, HMME@C₃F₈-NBs, and HMME-LND@C₃F₈-NBs for 30 s. The cell counting kit 8 (CCK–8) assessed the cell viability of HepG2 and Huh7 after 24 h. The microplate reader (Promega Corp, Madison, WI, United States) detected the absorption of the plates at 450 nm wavelength. The combination index (CI) values were calculated using CompuSyn software19.

After the treatment with different groups combined with US (1 MHz, 3.5 W/cm²), HepG2 cells and Huh7 cells were incubated with DCFH–DA (Applygen Technologies Inc., Beijing, PR China) for 20 min, the concentration of DCFH–DA was 10 mmol/L.

The mitochondrial membrane potential for early apoptosis was detected with a JC-1 fluorescence probe (Beyotime, Jiangsu, China). First, JC-1 liquid was added in HepG2 and Huh7 cells for 20 min without light, and then, JC-1 staining buffer was used to wash the cells twice before taking photos with the fluorescence microscope. ImageJ software (National Institutes of Health, Bethesda, MD, United States) calculated the average intensity of fluorescence.

The annexin V-FITC Apoptosis Detection Kit (Beyotime, Jiangsu, China) tested the apoptosis of HepG2 and Huh7 cells; 8 × 10⁴ cells, which were washed with cold PBS, were counted and resuspended in 195 μl combination liquid, and then, those cells were stained with 5 μl Annexin V-FITC and 10 μl propidium iodide (PI). The apoptosis rates were measured with flow cytometry (BD Biosciences, United States) immediately. Annexin V-FITC+/PI− confirmed the apoptotic cells, and Annexin V-FITC+/PI + stained the necrotic cells.

The total RNAs before (HepG2 cells named Group A and Huh7 cells named Group C) and after (HepG2 cells named Group B and Huh7 cells named Group D) treated with HMME-LND@C₃F₈-NBs were isolated and tested. The following sequencing of genes was implemented by Novogene Bioinformatics Technology Cooperation (Beijing, China).

We selected some differentially expressed lncRNAs to reveal further clarifying mechanisms in the treatment of HCC on the gene level because of the complexity of lncRNAs. The differentially expressed genes between the two parallel experimental groups (FDR < .01, the FC value was in the same direction, and the absolute value was more significant than two) were screened to verify the accuracy of RNA-seq. Next, we compared the differential lncRNAs with the Lnc2Cancer (Gao et al., 2021) and LncRNADisease (Bao et al., 2019) database, The lncRNAs consistent with RNA-seq and associated with HCC from both databases were selected. Finally, the selected lncRNAs were validated with a molecular biology experiment.

Trizol reagent (Thermo Fisher Scientific, Carlsbad, United States) was used to obtain the RNAs from HepG2 and Huh7 cells. Then, the researcher reverse transcribed the RNA to cDNA (Transcriptor First Strand cDNA Synthesis Kit, Roche, Switzerland, Germany). The qRT-PCR was performed by the SYBR Green real-time detection kit (CWBio, Beijing, China) with the CFX96 Detection System (Bio-Rad, California, United States). The relative expression of LINC00601 and ZFAS1 was calculated using the 2-ΔΔCt method and repeated three times. The following primers were used: LINC00601 (Forward: 5′-GAG​CTG​CAC​TGA​CCA​GTA​GG-3′, Reverse: 5′GTG​CTG​GCA​GAT​GGA​TCA​CT-3′); ZFAS1: (Forward: 5′-ACG​TGC​AGA​CAT​CTA​CAA​CCT-3′, Reverse: 5′-TAC​TTC​CAA​CAC​CCG​CAT-3′); GAPDH (Forward: 5′-CAT​GAG​AAG​TAT​GAC​AAC​AGC​CT-3′, Reverse: 5′-AGT​CCT​TCC​ACG​ATA​CCA​AAG​T-3′).

GraphPad Prism 8 software and Origin 2019 software performed statistical analysis. The experiments of HCC cells were repeated three times. The data were presented as an average of three replicates ±standard deviation; the p < .05 was considered statistically significant between control and samples values.

The appearance of pure NBs was shown as a uniform white emulsion (right of Figure 1A), and the HMME-LND@C₃F₈-NBs were rendered as a uniform pink emulsion (left of Figure 1A). Figures 1B,C displayed that HMME-LND@C₃F₈-NBs are shown as uniformly round under SEM and TEM. The size of the NBs was 284.76 ± 83.7 nm (Figure 1E). The zeta potential value of the complexes was –7.78 ± 6.72 mV (Figure 1F), and this showed that the HMME-LND@C₃F₈-NBs complexes were negatively charged. The absorbance peaks of HMME and LND were observed for the HMME-LND@C₃F₈-NBs (Figure 1D), indicating the joint of HMME and LND into NBs.

The viability of HepG2, Huh7 cells was examined after interference with LND, HMME, HMME @C₃F₈-NBs, and HMME-LND@C₃F₈-NBs for 24 h. The results show that all the groups reduced the cell viability in a dose-dependent manner (Figures 2A,B, all p < .05). Figures 2A,B also show the half-maximal inhibitory concentration (IC50) of the four groups. The IC50 (shown as HMME/LND) of HMME-LND@C₃F₈-NBs in HepG2 cells and Huh7 cells were 7.01/3.50 μg/ml (CI value: .713) and 2.99/1.50 μg/ml (CI value: .264). The groups of HepG2 cells (Figure 2C) and Huh7 cells (Figure 2D) treated with HMME-LND@C₃F₈-NBs show more significant cytotoxicity than the other groups (all p < .05); the results reveal that NBs combined with HMME and LND decreased cell viability more significantly than LND alone, HMME alone, and NBs combined with HMME.

The HMME-LND@C₃F₈-NBs groups in HepG2 and Huh7 cells exhibited more green fluorescence than the other three groups as shown in Figures 3A,C, which confirms that the HMME-LND@C₃F₈-NBs could be a more effective sonosensitizer generating ROS with LFUS. The fluorescence intensity of HMME @C₃F₈-NBs with LFUS was higher than HMME at the corresponding concentrations. Compared with the HMME @C₃F₈-NBs groups, LND improved the ability of the NBs complex to produce ROS. The quantification of fluorescence intensity in different groups of HepG2 and Huh7 cells was shown in Figures 3B,D. The fluorescence intensity of each group was statistically significant.

Fluorescence microscopic imaging was used to detect the mitochondrial membrane potential in HepG2 (Figure 4A) and Huh7 cells (Figure 4C). The mitochondrial membrane potential decreased after treatment with HMME-LND@C₃F₈-NBs more significantly than the control group, which showed conspicuous green fluorescence. NBs mediated with both HMME and LND were more evident than the NBs only mediated with HMME. The ratio of red to green fluorescence decreased memorably, quantifying fluorescence intensity as shown in Figures 4B,D. JC-1 aggregate converted to a monomeric form in the HepG2 cells, and Huh7 cells after SDT mediated with NBs complex, namely, early apoptosis occurred in the treatment group.

HepG2 and Huh7 cells were stained with HMME, HMME @C₃F₈-NBs, and HMME-LND@C₃F₈-NBs combined with LFUS for 24 h. The results show that the proportion of apoptotic cells (Q4) and necrotic cells (Q2) in the HMME-LND@C₃F₈-NBs group was higher than the others (Figure 5). That is, the NBs mediated with HMME induced apoptosis in HepG2 and Huh7 cells, and the addition of LND amplified the effect significantly.

Almost 361.8 million clean reads were isolated. The classification of raw reads is shown in Supplementary Figure S1. Almost 7,219 lncRNA transcripts were filtered as novel lncRNAs (Figure 6A). Those lncRNAs included 44.8% long intergenic noncoding RNAs (lincRNAs), 13.6% antisense lncRNAs, and 41.6% sense overlapping lncRNAs (Figure 6B). The distribution of expression levels of genes is shown with boxplots (Figure 6C).

The level of mRNA, lncRNA, and circRNA were standardized based on the FPKM method. Differentially expressed mRNA, lncRNA, and circRNA were identified by edgeR software, the BH correction was performed on the obtained p values, and p < .05 were selected as differentially expressed mRNA, lncRNA, and circRNA. In total, 5,035 mRNA genes were differentially expressed between groups B and A, including 3,013 upregulated and 2022 downregulated genes (Figure 7A left); meanwhile, 4,578 mRNA genes were detected between groups D and C, including 2,348 upregulated and 2,230 downregulated transcripts (Figure 7A right). The expression of lncRNA in the groups also showed differences: 3,693 lncRNA (including 2,187 upregulated and 1506 downregulated genes) genes expressed differentially in group B vs. A (Figure 7B left); meanwhile, 3,664 lncRNA genes were found in group D vs. C, including 2029 upregulated and 1635 downregulated genes (Figure 7B right). Then, a total of 4,321 circRNA genes were expressed differentially between groups B and A, including 3,324 upregulated and 997 downregulated genes (Figure 7C left). Meanwhile, 3,868 circRNAs were differentially expressed between groups D and C, including 2,307 upregulated and 1561 downregulated genes (Figure 7C right). The above result of groups B/A and D/C were compared and shown with Venn pictures, 2,390 mRNA genes, 1556 lncRNA genes, and 1456 circRNA genes are the same part of the comparison result of the two samples (Supplementary Figure S2). In other words, those genes were differentially expressed in the two HCC cell lines after SDT; we call them stable differential genes of two parallel groups. The differentially expressed mRNAs, lncRNAs, and circRNAs were performed with cluster analysis, and the heat map (red and blue indicate high and low expression genes separately) shows the result according to log10 (FPKM+ 1) value (Figure 8). Groups A and B were similar to groups C and D separately, indicating that these genes were differentially expressed in HepG2 and Huh7 cells and could be used as biomarkers to predict their status.

In this analysis, 578 differential mRNA and 770 differential lncRNA target GO terms were enriched significantly between groups B and A (Supplementary Tables S1, S3, p < .01), whereas 638 mRNA GO and 693 lncRNA targets terms showed apparent enrichment between groups D and C (Supplementary Tables S2, S4, p < .01) based on the survey. In addition, 566 and 519 differentially expressed circRNAs GO terms enriched significantly in group B vs. A (Supplementary Table S5, p < .01) and group D vs. C (Supplementary Table S6, p < .01), respectively. The representative selected enriched terms of differential genes are shown in Tables 1–3. The detailed classification of GO terms of mRNA, lncRNA, and circRNA in different groups are shown in Supplementary Figure S3. A directed acyclic graph (DAG, Supplementary Figure S4) displays the hierarchical enrichment relations. Each node shows the name and the adjusted p-value of the terms. The NCBI’s Gene Expression Omnibus could be used to get the KEGG enrichment pathway. KEGG enrichment analysis of the differentially expressed mRNAs and lncRNA targets in the two parallel experimental groups did not simultaneously enrich into the same statistically significant pathway. Therefore, KEGG enrichment analysis of the stable differential genes in the two experimental groups was conducted. Some statistically significant pathways of mRNAs and lncRNA targets are listed in Supplementary Tables S7, S8, and the complete results are listed in Supplementary Tables S9, S10, respectively. Table 4 shows the top five pathway enrichments of differential circRNAs in the two groups. The results of pathway enrichments of circRNAs are shown in Figure 9.

In the course of the current experiment, we intersected the differential lncRNAs screened in the two parallel experimental groups with the HCC-related genes that had been verified in the Lnc2Cancer and LncRNA Disease database, 12 meaningful lncRNAs were screened out, including seven upregulated lncRNAs and five downregulated lncRNAs (Table 5). LINC00601 and ZFAS1 were selected for validation in qRT-PCR due to their being expressed the same in the two databases and having significant differences in RNA-seq results.

The expression of LINC00601 and ZFAS1 in HepG2 and Huh7 cells after SDT was measured through qRT-PCR. The results confirmed that the expression of LINC00601 was reduced in HepG2 and Huh7 cells after SDT. Meanwhile, the expression of ZFAS1 was upregulated after SDT in HCC cell lines (Figure 10). These results were consistent with RNA sequencing and related lncRNA databases, so we selected LINC00601 and ZFAS1 as representatives for further research.