Production and characterization of the nucleic acid NFs
Based on information in the saRNA database and reports by Guo and Chen [40, 41], we designed a linear DNA template (named L-T, with a phosphate group at the 5′ end) complementary to the T7 promoter primer (Table S1) with embedded sense and antisense saRNA sequences (termed VEGF-706) that can effectively activate VEGF expression (Figure 1A). Which proved that the usability of the linear template we designed and can be used to activate the miR-155 NFs.
The saRNAs were designed according to a report by Li et al33. We searched for the sequence of the host gene, MiR155HG, and 10,000 bases upstream of the 5 end of the first exon in databases such as NCBI (https://www.ncbi.nlm.nih.gov/gene/?term=) and UCSC (https://genome.ucsc.edu/), and we predicted the transcription start site to be located at base 8,912 (Figure 1B) using various online resources (http://linux1.softberry.com/cgi-bin/programs/promoter/tssp.pl and http://www.bio-soft.net/sms/cpg_island.html). Next, we predicted the regions enriched in CpG islands to ensure that the designed saRNA avoids these regions (Figure 1C). We found that the region of 7,710–9,584 was enriched in CpG islands (Figure 1C). Therefore, the saRNA sense strand region of 7,712–8,712 was selected. We screened the sense strand region of the saRNA that can activate miR-155. Finally, seven eligible saRNAs were screened out (Table 1) and embedded into the template sequence designed above (Table 2).
Table 1 Sequences in line with saRNA design principles
Gene name
|
Positive-sense strand
|
antisense strand
|
miR-155-1
|
GUC ACC UCA GCC UCC CAA A
|
CAG UGG AGU CGG AGG GUU U
|
miR-155-2
|
UAU CCC UCU UAG UCU GCU A
|
AUA GGG AGA AUC AGA CGA U
|
miR-155-3
|
GUC UGC UAG GGU UGC CAU A
|
CAG ACG AUC CCA ACG GUA U
|
miR-155-4
|
AUA GAC UGG AUG GCU GAU A
|
UAU CUG ACC UAC CGA CUA U
|
miR-155-5
|
ACA UUC UGG AGG CUA GAA A
|
UGU AAG ACC UCC GAU CUU U
|
miR-155-6
|
GGA CUC UCU UCC UGG CUU A
|
CCU GAG AGA AGG ACC GAA U
|
miR-155-7
|
UCU UCC UGG CUU ACA GGA A
|
AGA AGG ACC GAA UGU CCU U
|
Table 2 Template and primer sequence for preparing nucleic acid nanoflowers that can activate miR-155
name
|
Sequence(5’→ 3’)
|
Template-1
|
ATAGTGAGTCGTATTAACGTACCAACAAGTCACCTCAGCCTCCCAAAACTTGTTTGGGAGGCTGAGGTGACATCCCT
|
Template-2
|
ATAGTGAGTCGTATTAACGTACCAACAATATCCCTCTTAGTCTGCTAACTTGTAGCAGACTAAGAGGGATAATCCCT
|
Template-3
|
ATAGTGAGTCGTATTAACGTACCAACAAGTCTGCTAGGGTTGCCATAACTTGTATGGCAACCCTAGCAGACATCCCT
|
Template-4
|
ATAGTGAGTCGTATTAACGTACCAACAAATAGACTGGATGGCTGATAACTTGTATCAGCCATCCAGTCTATATCCCT
|
Template-5
|
ATAGTGAGTCGTATTAACGTACCAACAAACATTCTGGAGGCTAGAAAACTTGTTTCTAGCCTCCAGAATGTATCCCT
|
Template-6
|
ATAGTGAGTCGTATTAACGTACCAACAAGGACTCTCTTCCTGGCTTAACTTGTAAGCCAGGAAGAGAGTCCATCCCT
|
Template-7
|
ATAGTGAGTCGTATTAACGTACCAACAATCTTCCTGGCTTACAGGAAACTTGTTCCTGTAAGCCAGGAAGAATCCCT
|
T7
|
TAATACGACTCACTATAGGGAT
|
Note: All the above sequences (except T7) need to modify the phosphate group at the 5'end
We optimized the template concentration (Figure 2A), primer concentration ratio (Figure 2B), ligase, optimized hybridization time and ligation time (Figure 2C-E), and raw material concentration (Figure 2F). The optimized result is shown in Figure 2G-H. It was finally determined that the ratio of template concentration to primer concentration was 1:1 in subsequent experiments, and T4 DNA ligase was used for the ligation reaction. The hybridization time was 2h, the ligation time was 3h, and the rNTP concentration was 2mM. As shown in Figure 2I-J, we have successfully prepared nucleic acid nanoflowers, which are of homogenous sizes and shapes, and the size is nanoscale.
The results of gel imaging of the PEI-NF complexes are shown in Figure 3A. According to lanes 1 and 2 in the figure, PEI and the nucleic acids interact via electrostatic interaction to form a positively charged complex, and the surface charge is reduced. Therefore, the target band does not migrate downward. When only PEI is present, bright bands do not appear in the absence of nucleic acids. Thus, the PEI-NF complex was successfully prepared.
PEI-NF complex and saRNA that can activate VEGF expression have no effect on cell activities
The cytotoxicity of the PEI-NF complexes towards HUVECs was evaluated by CCK-8 assays (Figure 3B). The survival rate of HUVECs treated with PEI-NFs or saRNA was >1, and there was no significant difference with the control group (P > 0.05). These results showed that neither the PEI-NFs nor the saRNA had cytotoxic effects on HUVECs. In addition, PEI-NFs and saRNA did not significantly suppress cell proliferation (P > 0.05). Thus, the PEI-NFs were further evaluated in subsequent experiments.
PEI-NFs that can activate VEGF expression successfully enter cells
Fluorescence microscopy confirmed the presence of PEI-NFs within the cells (Figure 3C). Cells stained with DAPI were imaged with an exposure time of 10 ms, and those stained with Cy3 were imaged with an exposure time of 100 ms. In the control group, no Cy3 fluorescence signal was detected, whereas in the experimental group (L-T), red fluorescence was observed, indicating that the PEI-NFs had successfully entered the cells. Flow cytometry was used to quantitatively evaluate the efficiency of PEI-NF delivery into the cells (Figure 3D, E). Compared with the control group, the efficiency of PEI-NF delivery into the cells was > 70% (P < 0.0001; Figure 3E).
PEI-NF complex and saRNA activate VEGF expression
We examined the activation of VEGF expression induced by PEI-NF complex and saRNA using RT-qPCR (Figure 3F). Compared with that in the control group, the relative expression of VEGF in the experimental group of cells treated with PEI-NFs or saRNA activated was increased by 2.5-fold (P < 0.01). This result showed that the PEI-NF complex and saRNA effectively activated VEGF expression. There was no difference in the relative VEGF expression between the PEI-NF- and saRNA-treated cells, indicating that they have similar efficiency. Thus, we confirmed that the nucleic acid NF was effective and could be used for the activation of miR-155.
Changes in the template sequences of miR-155-activating NFs have no impact on cell proliferation
In the above experiments, we screened seven saRNAs that can activate miR-155. We used the CCK-8 assay to detect the inhibitory effects of nucleic acid NFs M1–M7 prepared from seven different template sequences on cell proliferation at 12, 24, 48, and 72 h (Figure 4A). Compared with the control group, NFs M1–M7 had no inhibitory effect on cell proliferation even after 72 h (P > 0.05). Nearly all NFs had similar effects, and the cell survival rate remained nearly constant from 12 h to 24 h, slightly decreased between 24 h and 48 h, and decreased from 48 h to 72 h (to approximately 1). These findings showed that a change in the template sequence has no inhibitory effect on cell proliferation.
Screening of saRNAs that can activate miR-155 expression
To screen for saRNAs that can activate miR-155, we evaluated the NFs M1–M7 for their capacity to activate MiR155HG expression (Figure 4B–D) and miR-155 expression (Figure 4E–G) in HUVECs after treatment for 12, 24, 48, or 72 h. Compared with that in the control group, relative MiR155HG gene expression in the experimental groups treated with M1–M7 significantly increased within 48 h (P < 0.01) and decreased between 48 h and 72 h. miR-155 expression increased over time. Based on the relative expression levels of miR-155, all seven saRNAs were found to activate miR-155 expression. M5 had the best effect; it enhanced miR-155expression level by 11.5-fold (P < 0.001) (Figure 4H). Therefore, this saRNA was used to activate miR-155 expression in further experiments.
NFs that can activate miR-155 expression successfully enter cells
Using fluorescence microscopy, we confirmed the presence of NFs that can activate miR-155 within cells (Figure 5A). Cells stained with DAPI were imaged with an exposure time of 10 ms, and those stained with Cy3 were imaged with an exposure time of 100 ms. In the control group, no Cy3 fluorescence signal was detected, whereas in the experimental group (M5), red fluorescence was observed, indicating that the M5 NFs had successfully entered the cells.
Overexpression of miR-155 promotes the migration of HUVECs
Cell migration was detected by a scratch assay at 0 h,12 h, 24 h, 48 h, and 72 h after transfection of the cells (Figure 5B). All cells migrated over time, but compared with the control group, cells with activated miR-155 expression migrated faster, indicating that miR-155 overexpression promoted cell migration as indicated by scratch closure. This result preliminarily indicated that miR-155 overexpression may cause inflammation.
Activation of miR-155 affects the expression levels of key signaling molecules and inflammatory factors
In a preliminary experiment, we treated HUVECs with different concentrations of LPS and found that 0.1 μg/mL LPS effectively stimulated HUVECs; miR-155 expression first increased and then decreased, and the highest expression level was approximately 4-fold higher than that in the control group (Figure S2A). Treatment of cells with a high concentration of LPS caused apoptosis (Figure S2B–H). Next, we studied whether activated miR-155 can trigger an inflammatory response. We treated HUVECs with the miR-155-activating NFs for 72 h and used RT-qPCR to detect changes in the expression of inflammation-related effectors and signaling pathway genes (Figure 6A). After miR-155 activation, the relative expression levels of inflammation-related genes changed. Gene expression of the anti-inflammatory factor SHIP1 was significantly reduced (P < 0.01). Gene expression of the pro-inflammatory factors TNF-ɑ, IFN-γ, IL-1β, IL-6, and FOXO3A was significantly increased (P < 0.05). IKKɛ activity is related not only to inflammatory diseases, but also to cancer onset. IKKɛ may act as an oncogene promoting malignant transformation and tumor progression. Our research showed that after miR-155 was activated, IKKɛ gene expression was significantly increased (P < 0.001), suggesting that miR-155 activation may be related to cancer. The PI3K/AKT signaling pathway regulates multiple biological processes [46] and is closely related with tumor development and metastasis [47]. We found that PI3K/AKT gene expression increased after miR-155 activation (P < 0.01), which may be related to tumor development. NF-κB is a key transcription factor involved in inflammatory signaling pathways and responsible for the initiation of transcription of downstream inflammatory factors [48]. Activated miR-155 increased the expression of NF-κB (P < 0.05) as well as that of pro-inflammatory factors (P < 0.05), indicating the induction of an inflammatory response. Together, these results suggested that miR-155 overexpression is related to inflammation and tumorigenesis.
Activated miR-155 significantly upregulates IL-1β protein expression and downregulates SHIP1 protein expression
To verify that activation of miR-155 can induce inflammation, we treated HUVECs with miR-155-activating NFs for 72 h and then measured the expression of pro-inflammatory and anti-inflammatory proteins by western blotting. Compared with the control group, HUVECs treated with LPS or miR-155-activating NFs showed upregulated IL-1β protein expression and downregulated SHIP1 protein expression. However, compared with LPS, miR-155 activation had a significantly stronger promotive effect on IL-1β protein expression (P < 0.05, Figure 6B, C) and suppressive effect on SHIP1 protein expression (P < 0.05, Figure 6D, E). Thus, NF-M5 significantly induced the expression of the inflammatory factor IL-1β and reduced that of the negative regulator of inflammation, SHIP1. These results indicated that in the absence of exogenous inflammatory factors, NFs can directly activate miR-155 expression in cells and induce cell inflammation.
Prediction of miR-155 target genes and analysis of gene co-expression
We predicted the human target genes of miR-155 using four miR-155 target gene prediction tools and analyzed the enriched regions of the target genes involved in the inflammatory signaling. The predicted target genes are listed in Table S3. Although the numbers of target genes yielded by the different tools differed, there were a large number of common genes and only a few genes were predicted by only one tool. In total, 64 target genes were predicted by all four tools and thus had a high confidence (Figure 7A). Therefore, these genes were selected as the final miR-155 target genes. Next, we used the DAVID tool to analyze the enrichment of these genes in signaling pathways, which revealed that the miR-155 target genes showed a distinct enrichment pattern. Forty-seven target genes were involved in 32 signaling pathways. There were eight groups of target genes involved in different inflammatory signaling pathways (Table S4), suggesting that miR-155 regulates inflammatory signal transmission by targeting these genes to ultimately regulate the onset and development of inflammation. The PI3K-AKT pathway was enriched in miR-155 target genes. Our previous studies showed that the expression of the effectors SHIP1 and FOXO3A changed significantly after the activation of miR-155 expression (Figure S2B). This indicates that miR-155 may promote PI3K-AKT signaling by inhibiting the expression of SHIP1, thereby causing inflammation.
Co-expression of miR-155 with its target SHIP1 cancels out their individual effects, suppressing inflammation
In the saRNA database, we did not find an saRNA that can activate SHIP1 expression. Therefore, according to the saRNA design principles reported by Li et al33, we searched the 5,000 bases upstream of the 5 end of the SHIP1 gene, INPP5D, for the transcription start site (Figure 7B) and a region enriched in CpG islands. The transcription start site was at 1,911, and there was no region enriched in CpG islands. Therefore, the saRNA sense strand region (1–1,711) was selected. We screened out eight saRNAs that met all the conditions (Table S5), and embedded the selected sequences into the template sequence (Table S6). We used the same methods as those used above to prepare NFs that can activate SHIP1 expression. The CCK-8 assay was used to evaluate the inhibitory effects of nucleic acid NFs S1–S8 prepared from the eight template sequences on cell proliferation at 12 h, 24 h, 48 h, and 72 h (Figure 7C). Nucleic acid NFs S1–S8 had no inhibitory effect on cell proliferation in the first 48 h, but they did inhibit growth after 48 h (P < 0.05). All nucleic acid NFs had similar efficacy. This finding indicated that a change in SHIP1 expression affects cell proliferation and that SHIP1 expression is activated within 48 h to 72 h, inhibiting cell proliferation via the PI3K-AKT pathway.
Nucleic acid NFs S1–S8 activated SHIP1 expression after 72 h, as shown in Figure 7D. S7 was the most effective. We evaluated cell morphology after the activation of miR-155 expression, SHIP1 expression, and both in HUVECs (Figure 7E–H). In the presence of serum, the cell growth rate was high, and the cells were in good state and displayed long spindle-like morphology and tight connections (Figure 7E). After miR-155 activation, the cell growth rate increased compared to that of control cells, the cells were in a normal state, cell density increased, and cells were slightly overlapping (Figure 7F). Upon SHIP1 activation, the cell growth rate decreased, cell density was obviously reduced, and the cells showed a scattered distribution (Figure 7G), suggesting that SHIP1 expression enhanced apoptosis in HUVECs. When miR-155 and SHIP1 expression was activated simultaneously, the cell growth rate was comparable to that of control cells and the cell density was only slightly decreased, indicating a normal state (Figure 7H). Further, upon simultaneous activation of miR-155 and SHIP1, the expression of both genes did not change (Figure 7I). These findings indicated that upon simultaneous activation of miR-155 and SHIP1, their individual effects are canceled out, and inflammation would not be promoted.