MicroRNA-18a prevents senescence of mesenchymal stem cells by targeting CTDSPL

Senescence impaired the biological functions of mesenchymal stem cells

Long-term expansion of UCMSCs resulted in increased population double time (PDT) (Figure 1A), indicating the growth arrest of long-term cultured UCMSCs. Late passaged UCMSCs also showed a larger cell size than early passaged UCMSCs (Figure 1B). To describe the physiological changes of UCMSCs upon prolonged culture, we first detected the immunophenotypes expression of early passaged (P3-P6) and late passaged (P11-P14) UCMSCs. Early passaged UCMSCs expressed the typical immunophenotypes of MSCs, which showed negative of CD34, CD45, HLA-DR and positive of CD29, CD73, CD90, and CD105. Comparatively, all negative markers, especially CD45 and HLA-DR were up-regulated in late-passaged UCMSCs. A slight down-regulation of the positive marker CD105 (from 100% decreased to 86.2%), as well as an enhanced expression of CD90 (from 92.7% increased to 100%) in late passaged UCMSCs were also detected (Figure 1C). Despite these changes, the majority of the late-passaged UCMSCs maintained expression of the typical MSC markers. Therefore, spontaneous differentiation was not the most pressing factor that influenced UCMSC’s function.

Figure 1. Late passaged UCMSCs showed senescence-associated changes and impaired biological functions. (A) Population double time (PDT) of UCMSCs during long-term expansion. (B) The cell size of early and late passaged UCMSCs was measured using forward scatter (FSC) and side scatter (SSC) by flow cytometry. (C) Immunophenotypes characterization of early and late passaged UCMSCs. (D) SA-β-gal staining of early and late passaged UCMSCs. (E) Quantification of the SA-β-gal staining positive cells (n = 3). (F) Western blot analysis of p16 expression in UCMSCs of passages 4, 8, and 11. (G) Early and late passaged UCMSCs were induced to osteogenic (Ost) or adipogenic (Ado) differentiation. Osteogenic efficiency was evaluated by alkaline phosphatase (ALP) staining; adipogenic efficiency was evaluated by oil red O staining. (H) The differentiation rate was quantized. (I) Wound healing assay was performed for early and late passaged UCMSCs and (J) migrated cell numbers were quantified in 0 hours, 6 hours, 12 hours and 24 hours (n = 3). ^***p < 0.001.

We then detected expressions of senescence-associated markers. Early passaged UCMSCs showed a low positive rate (5%) of SA-β-gal stain, while more than half of the late passaged UCMSCs showed positive SA-β-gal stain (Figure 1D, 1E). The other senescence-associated marker, P16 also showed up-regulated upon passages (Figure 1F).

To evaluate the functions of UCMSCs, we analyzed the differentiation potential of UCMSCs by inducing them to osteogenic or adipogenic differentiation. The late-passaged UCMSCs showed remarkably reduced rates of osteogenic and adipogenic differentiation than the early-passaged UCMSCs. In the case of adipogenic differentiation, a massive of late-passaged cells lost during the differentiation process. This result implied that the majority of the late-passaged UCMSCs lost their differentiation potentials (Figure 1G, which was quantized in Figure 1H). We also performed a wound-healing assay. The early-passaged UCMSCs showed higher migration abilities than the late-passaged UCMSCs (Figure 1I, 1J).

Collectively, these results showed that the long-term culture-induced senescence of UCMSCs, which impaired their biological functions.

CTDSPL up-regulation induced premature senescence of mesenchymal stem cells

The up-regulation of miR-26a and miR-26b in late-passaged UCMSCs has been identified in our previous study [20]. Here we first verified the expression of miR-26a/b in UCMSCs. Significant up-regulation of both miR-26a and miR-26b in long-term cultured UCMSCs (Figure 2A) was verified by quantitative RT-PCR assay. The genomic loci of miR-26a/b harbor the intron site of the CTDSP family genes (CTDSP1, CTDSP2, CTDSPL) and the expression of miR-26a/b has been closely linked to that of their host genes (Figure 2B). Therefore, the up-regulation of miR-26a/b in old UCMSCs inspired us to analyze whether their host gene CTDSPL, which was previously identified as a tumor-suppressor gene, was up-regulated in the late passage of UCMSCs. Consistent with our hypothesis, the Western blot assay showed an up-regulation of CTDSPL in late-passaged UCMSCs (Figure 2C). CTDSPL and P16 expression results were further quantized in Figure 2D, 2E.

Figure 2. miR-26a/b and CTDSPL expression were up-regulated following senescence of UCMSCs. (A) Quantitative RT-PCR assay was performed to detect miR-26a and miR-26b expression in UCMSCs. (B) Gene structure of CTDSPL. (C) Western blot assay was performed to detect CTDSPL expression in UCMSCs. (D) The expression of P16 was quantized. (E) CTDSPL was also quantized. ^*p < 0.05; ^**p < 0.01.

To determine whether the up-regulation of CTDSPL correlated with the senescence of MSCs, we performed a CTDSPL over-expression assay. The complete coding sequence of CTDSPL was inserted into the downstream of the GFP fluorescent protein sequence in the pEGFP-C1 plasmid. Compared to the control group, ectopic expression of CTDSPL significantly increased the rate of SA-β-gal positive cells in UCMSCs (Figure 3A, 3B). By fluorescence, we also observed that over-expression of CTDSPL resulted in an enlarged morphology change of UCMSCs (Figure 3C). As revealed by F-actin labeling (Figure 3D), this morphology change was associated with the dramatic reorganization of cytoskeletal networks. Western blot assay further showed that over-expression of CTDSPL induced the up-regulation of p16 (Figure 3E) and reduced the total and phosphorylated RB proteins (Figure 3F). Collectively, these results indicated that CTDSPL is capable of inducing senescence in UCMSCs.

Figure 3. Over-expression of CTDSPL induced premature senescence of UCMSCs. (A) SA-β-gal staining of control and CTDSPL over-expressed UCMSCs. (B) SA-β-gal positive cells were quantified (n = 3). (C) Morphology changes were quantified by flow cytometry. (D) Representative confocal microscopy images of GFP tagged control or CTDSPL transfected UCMSCs. F-actin was labeled with Alexa Fluor 633 conjugated phalloidin; nuclei were stained with Hoechst 33342. (E) Western blot assay of p16 expression after transfected with control or CTDSPL plasmid. (F) Western blot assay of RB and pRB expression after transfected with control or CTDSPL plasmid. ^***p < 0.001.

Internal microRNA-18a inhibits the senescence pathway in mesenchymal stem cells by directly repressing the expression of CTDSPL

Next, we further investigated whether CTDPL expression inhibition may repress the senescence process during long-term culture of MSCs. First, we screen the internal miRNAs that may regulate the expression of CTDSPL. The putative miRNAs that target CTDSPL were analyzed through the public miRNAs-targets database miRanda [21], Pictar2 [22], and TargetScan [23]. Among them, miR-18a-5p has been selected with the consideration of abundant expression in early-passaged UCMSCs and significant down-regulation in late-passaged UCMSCs, as shown in our previous study. To further verify this result, we measured the expression level of miR-18a-5p in different passaged UCMSCs by quantitative RT-PCR. Compared to early passaged UCMSCs, the prolonged culture of UCMSCs significantly down-regulated miR-18a-5p (Figure 4A). Furthermore, when transfected with a synthetic miR-18a-5p inhibitor, the expression of CTDSPL in UCMSCs was significantly up-regulated compared to the miR-NC control group (Figure 4B). Meanwhile, compared to UCMSCs transfected with NC inhibitor, UCMSCs transfected with hsa-miR-18a-5p inhibitor showed a larger proportion of SA-β-gal positive cells (Figure 4C, 4D) and an up-regulation of P16 expression (Figure 4E). These data suggested that hsa-miR-18a-5p played an important role in preventing the senescence of UCMSCs.

Figure 4. Inhibition of miR-18a-5p induced premature senescence of UCMSCs. (A) Relative expression of miR-18a-5p following passaging was analyzed by quantitate RT-PCR. (B) Western blot assay of CTDSPL expression after miR-NC inhibitor or miR-18a-5p inhibitor transfection. (C) SA-β-gal staining of UCMSCs transfected with miR-NC inhibitor or miR-18a-5p inhibitor. (D) SA-β-gal positive cells were quantified (n = 3). (E) Western blot assay of p16 expression in UCMSCs after miR-NC or miR-18a-5p inhibitor transfection. (F) Schematic representation of the reporter plasmids psiCHECK2-CTDSPL-3UTR-Wild, psiCHECK2-CTDSPL-3UTR-Mutation and psiCHECK2-CTDSPL-3UTR-Delation. (G) Luciferase reporter assay was performed to verify the direct repression of CTDSPL by miR-18a-5p. ^*p < 0.05; ^**p < 0.01.

Finally, a luciferase reporter assay was performed to confirm whether the regulation effect of miR-18a-5p on CTDSPL by direct targeting mechanism. The 3′ untranslated region (UTR) sequence that included the putative miR-18a-5p binding site was inserted into the psiCheck2 dual luciferase reporter plasmid. We also constructed a mutation and a deletion plasmid as control (Figure 4F). Compared to co-transfecting miR-NC with the wide-type plasmid, the relative luciferase intensity was significantly repressed when co-transfecting miR-18a-5p with the wild-type plasmid, while co-transfecting miR-18a-5p with the mutation or deletion plasmid reversed this effect (Figure 4G). These results implied that miR-18a-5p repressed the expression of CTDSPL by directly targeting its 3′UTR site.

Stable expression of microRNA-18a-5p attenuated senescence and improved self-renewal of mesenchymal stem cells

To evaluate whether continuous overexpression of miR-18a-5p could attenuate senescence and improve self-renewal of MSCs, we over-expressed miR-18a-5p by lentivirus. Since the accumulation of reactive oxygen species (ROS) impaired homeostasis and is taken as the first inducer of senescence, we used a florigenic dye DCF-DA to measure the total ROS activity within the cell. The diffused DCF-DA is first deacetylated by cellular esterases to a non-fluorescent compound, and then can be oxidized by ROS into DCF, which is a highly fluorescent compound. We found the total ROS level in the late passaged (passage 11) UCMSCs was significantly higher than the early passaged (passage 4) UCMSCs (Figure 5A). Considering mitochondria are the main source of cellular ROS, we then evaluated the mitochondrial ROS level in UCMSCs. Using the MitoTracker Green FM probe to label the mitochondria, we found the mitochondria mass increased with the senescence of UCMSCs. We then detected the mitochondrial ROS level by using the MitoSOX Red probe, which can be oxidized by mitochondrial superoxide resulting in the emission of red fluorescence. The late-passaged UCMSCs showed a higher level of mitochondrial ROS than the early-passaged UCMSCs (Figure 5A, 5B). Therefore, the senescence of UCMSCs is accompanied by an increased level of cellular ROS. However, 48 hours after overexpressing miR-18a-5p in UCMSCs, we detected reduced levels of total ROS, mitochondrial mass and mitochondrial ROS (Figure 5C, 5D). These results implied that increased expression of miR-18a-5p attenuated the senescence of UCMSCs.

Figure 5. miR-18a-5p overexpression reduced ROS levels of late passaged UCMSCs (passage 11). (A, B) Total ROS, mitochondrial mass and mitochondrial ROS levels of early and late passaged UCMSCs were detected by flow cytometry (A) and confocal microscope (B). (C) Flow cytometry analysis of the total ROS, mitochondrial mass and mitochondrial ROS levels of miR-18a-5p and control lentivirus vector transduced UCMSCs and (D) relative fluorescence intensity of miR-18a-5p overexpressing groups relative to control groups were quantified (n = 3). ^*p < 0.05; ^**p < 0.01.

To investigate whether increased expression of miR-18a-5p contributed to the self-renewal of UCMSCs, we detected the expression of Oct4 and Nanog, which are well-known stem cell markers that control self-renewal. Western blot showed that 48 hours after overexpressing miR-18a-5p, the expression of Oct4 and Nanog was up-regulated while the expression of P16 was down-regulated (Figure 6A).

Figure 6. Stable expression of miR-18a-5p improved self-renewal of UCMSCs. (A) Western blot assay of Oct4, Nanog and P16 in miR-18a-5p and control lentivirus vector transduced UCMSCs. (B, C) Competitive growth assay of miR-18a-5p and control transduced UCMSCs. GFP ratios were measured by flow cytometry before and after two passaging cultures. (D) GFP ratio of UCMSCs in competitive growth assay before and after culture. The efficiency of osteogenic differentiation (E) was increased in UCMSCs stably expressing miR-18a-5p (F), and similarly, the potential for adipogenic differentiation (G) was enhanced in the group with stable miR-18a-5p transduction. Osteogenic efficiency was evaluated by ALP staining; adipogenic efficiency was evaluated by oil red O staining. The results were quantized in Figure F and Figure H. ^*p < 0.05; ^**p < 0.05.

To measure the effect of continuous overexpression of miR-18a-5p on cell growth, we performed a competitive cell growth assay. The GFP labeled miR-18a-5p or control lentivirus vector transduced UCMSCs were mixed with untransduced UCMSCs with a 1:1 ratio. Then the mixed cells were cultured for another two passages and the GFP+/GFP- ratio was measured by flow cytometry (Figure 6B). Before passaging, the percentage of GFP+ cells were 42.3% for control and 45.0% for miR-18a-5p over-expressing cells. After two passages, the percentage of GFP+ cells in the control group dropped to 28.7%, while the percentage of GFP+ cells expressing miR-18a-5p maintained a similar ratio (40.6%) (Figure 6C, 6D). These results indicated that stable expression of miR-18a-5p improved the growth activity of UCMSCs.

Finally, to evaluate the differentiation potentials of UCMSCs stably expressing miR-18a-5p, we performed an osteogenic and adipogenic differentiation assay. The results showed that stable transduction of miR-18a-5p enhanced both the osteogenic and adipogenic differentiation ability of UCMSCs (Figure 6E, 6G and quantized in Figure 6F, 6H respectively).

Cell culture

Human umbilical cord tissue slices were kindly donated from Shandong Cell-Tissue Bank (Jinan, Shandong, China). The isolation of UCMSCs was performed as previously described [15]. Briefly, arteries and veins were first removed and the remaining tissues were chopped into small pieces. All pieces were placed in a DMEM-low glucose medium supplemented with 5% fetal bovine serum (Hyclone, South Logan, UT, USA) and 10 ng/ml basic fibroblast growth factor (Peprotech, London, UK) and cultured in 37°C at 5% CO₂. In total, UCMSCs from five samples were used in this study. HEK293T cells were purchased from Applied Biological Materials China (Nanjing, Jiangsu, China) and maintained with DMEM-high glucose medium supplemented with 10% fetal calf serum (Hyclone) in 37°C at 5% CO₂.

Osteogenic and adipogenic differentiation

For osteogenic differentiation, confluent UCMSCs growing in six-well plates were subjected to osteogenic medium composed of DMEM with low glucose supplemented with 10% FBS (Hyclone), 1% penicillin/streptomycin (Hyclone), 100 nM dexamethasone (Sangon Biotech, Shanghai, China), 10 mM β-glycerophosphate (Sigma, St. Louis, MO, USA) and 50 μM ascorbic acid (Sigma). At the end of the three weeks, cells were fixed with 4% paraformaldehyde for 30 minutes at room temperature and stained with BCIP/NBT (Amresco, Solon, OH, USA) for the visualization of alkaline phosphatase (ALP) expression. For adipogenic differentiation, confluent UCMSCs growing in six-well plates were treated with DMEM with low glucose supplemented with 10% fetal bovine serum (Hyclone), 1% penicillin/streptomycin (Hyclone), 1 μM dexamethasone (Sangon Biotech), 500 μM isobutylmethylxanthine (Sigma), 200 μM indomethacin (Sigma) and 10 μg/ml insulin (BasalMedia, Shanghai, China). At the end of the three weeks, cells were fixed with 10% buffered formalin for 30 minutes at room temperature and stained with oil red O for the visualization of oil droplets.

Plasmids construction

For the CTDSPL overexpression plasmid, the coding sequence of CTDSPL was inserted between the HindIII and BamHI sites in the pEGFP-C1 plasmid. For the wild-type luciferase reporter plasmid, the part sequence of CTDSPL 3′ untranslated region (UTR) that contains the putative miR-18a-5p binding site was inserted between the XhoI and NotI sites in psiCHEKC2 vector (Promega, Madison, WI, USA). For the 3′UTR mutated and deleted reporter plasmids, the mutation or deletion of specific sites in wild-type reporter plasmid was generated using a Mut Express II Fast Mutagenesis Kit (Vazyme, Nanjing, Jiangsu, China) according to the manufacturer’s procedure. To construct the lentivirus transduced miR-18a-5p overexpression plasmid, the primary miR-18a-5p sequence was inserted between the BamHI and XbaI sites in the pLVX-AcGFP-N1 vector. To construct a GFP-labeled lentivirus transduced miR-18a-5p overexpression plasmid, the primary miR-18a-5p sequence was inserted into the XbaI site in the pLVX-AcGFP-N1 vector.

Cell transfection

To overexpress CTDSPL in MSCs, the Neon electroporation system (Invitrogen, Eugene, OR, USA) was used according to the manufacturer’s guidelines. The empty pEGFP-C1 plasmid was used as a control. The transfected MSCs were seeded into six-well plates and cultured for 72 hours. To inhibit miR-18a-5p in MSCs, the synthetic miR-18a-5p inhibitors (GenePharma, Shanghai, China) were transfected into MSCs using a Neon electroporation system (Invitrogen). The electroporation condition was 1350V and one pulse with pulse width 30 ms. The transfected cells were cultured for 72 hours. The non-sense inhibitors served as a control. For the luciferase reporter assay, HEK293T cells were seeded into 96-well plates and cultured until 80% confluence. The synthetic miR-18a-5p mimics (GenePharma) and luciferase reporter plasmids were co-transfected into HEK293T cells using Lipofectamine 2000 (Invitrogen) following the manufacturer’s procedure.

Immunophenotypes assay

To characterize the immunophenotypes of UCMSCs, early passaged (passage 3 to passage 6) or late passaged (passage 11 to passage 14) UCMSCs were collected and incubated with fluorescence conjugated antibodies against CD34, CD45, HLA-DR, CD73, CD105 (Miltenyi Biotec, Bergisch Gladbach, Germany), and CD29, CD90 (BD Bioscience, San Jose, CA, USA). Fluorescence intensity was detected using the Accuri C6 flow cytometer (BD Bioscience). Analysis was performed using FlowJo software (Tree Star, Inc., San Carlos, CA, USA).

Cell migration assay

UCMSCs were cultured in six-well plates until 100% confluence. Then the cells were scratched with 100 μl pipette tips, washed with PBS and incubated for another 24 hours. Pictures were obtained under an inverted microscope (Nikon Eclipse Ti-S, Tokyo, Japan) at different time points. The results were analyzed using ImageJ software (developed at the National Institutes of Health).

Quantitative RT-PCR assay

A stem-loop RT-PCR method was used to quantify the expression of mature miRNAs [48]. Total RNAs were isolated using RNAiso plus (Takara, Dalian, Liaoning, China) following standard procedure. First-strand cDNA was generated with miRNA-specific RT primers using M-MLV reverse transcriptase (Promega, Madison, WI, USA). Quantitative RT-PCR (qRT-PCR) was performed with EvaGreen (Biotium, Hayward, CA, USA) using ABI 7500 Real-time PCR instrument (Applied Biosystems, Foster City, CA, USA). U6 was used as a housekeeping gene. And 2^−ΔΔCT method was used to analyze the data. The primers used in this study were as follows:

hsa-miR-26a-5p, 5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACagccta-3′ (RT primers), 5′-CCGCCGTTCAAGTAATCCAG-3′ (forward); has-miR-26b-5p, 5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACacctat-3′ (RT primers), 5′-CGCCGCTTCAAGTAATTCAGGAT-3′ (forward); hsa-miR-18a-5p, 5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACCTATCT-3′ (RT primers), 5′-CACGCGTAAGGTGCATCTAGT-3′ (forward); universal reverse primers, 5′-CCAGTGCAGGGTCCGAGGTA-3′; U6, 5′-CTCGCTTCGGCAGCACA-3′ (forward), 5′-AACGCTTCACGAATTTGCGT-3′ (reverse); CTDSPL, 5′-cataagcttacatggacggcccggccatcat-3′ (forward), 5′-gacggatccctacctattgcagagtctgtg-3′ (reverse).

Western blot assay

Whole-cell lysates of MSCs were extracted using a total protein extraction kit (Sangon Biotech, Shanghai, China). Protein samples were resolved by 12% SDS-PAGE and transferred to PVDF membranes (Millipore, Billerica, MA, USA). Membranes were blocked with 5% non-fat milk or bovine serum albumin (BSA) for 1 hour at room temperature and subsequently incubated overnight at 4°C with diluted primary antibodies. The primary antibodies used in this study included: CTDSPL (Santa Cruz Biotech, Santa Cruz, CA, USA), p16 (Santa Cruz Biotech), pRB (Cell Signaling Technology, Beverly, MA, USA), RB (Cell Signaling Technology), GAPDH (Santa Cruz Biotech) and Tubulin (Transgene Biotech, Beijing, China). To detect the signals, the membranes were then incubated with the appropriate horseradish peroxidase-conjugated secondary antibody (CWBio, Beijing, China) for 1 hour at room temperature. Finally, the signals were visualized using an enhanced chemiluminescence substrate (ECL; CWBio).

SA-β-gal activity assay

To examine senescence-associated β-galactosidase (SA-β-gal) activity, MSCs were seeded into six-well plates. Senescence-associated SA-β-gal staining was performed using an SA-β-gal staining kit (Beyotime, Shanghai, China). Cells were observed using an inverted microscope (Nikon Eclipse Ti-S). Three random fields per well were selected to count SA-β-gal positive cells.

Cytoskeleton labeling and imaging

PEGFP-CTDSPL or empty vector-transfected MSCs were fixed with 4% paraformaldehyde (PFA) for 20 minutes, permeabilized with 0.1% Triton X-100 for 5 minutes and blocked with 1% bovine serum albumin (BSA) for one hour. Then, the cells were stained with Alexa Fluor 633 Phalloidin (Invitrogen) for half an hour and Hoechst 33342 for 10 minutes. All procedures were performed at room temperature. Fluorescence was detected using a spinning disk confocal microscope system (Andor Technology, Belfast, Northern Ireland) on an inverted microscope (Nikon Eclipse Ti-E, Tokyo, Japan).

Lentivirus transduction

For preparation of lentivirus, 293T cells were transfected with a mixture of plasmids containing lentiviral vector (control and miR-18a-5p) and packaging plasmids (pMDL, pVSVG and pREV) using CaPO₄ precipitation according to a previously described protocol [49]. Media containing lentiviruses were collected at 24 hours and 48 hours after transfection. Then the media were filtered through a 0.45 μm pore size filter and centrifuged at 100,000 × g for 70 min at 4°C. After pouring off the supernatant, the pellet was resuspended in 100 μl PBS and stored at −80°C. For the transduction of MSCs, cells were seeded in a six-well plate. The next day, lentiviral supernatants supplemented with 10 μg/ml polybrene (Sigma) were added. Cells were incubated for another 24 hours before changing to fresh media.

Reactive oxygen species assay

MSCs cultured in six-well plates were incubated with 5 μM DCFH-DA (Beyotime) for 20 minutes, or 100 nM MitoTracker Green FM (Invitrogen) for 15 minutes, or 5 μM MitoSOX Red (Invitrogen) for 10 min at 37°C protected from light. For flow cytometry assay, cells were washed with fresh media three times, and detached with trypsin. For confocal microscope imaging, cells were counter-stained with Hoechst33342 to visualize the nuclei.

Luciferase reporter assay

To test the binding of miR-18a-5p to CTDSPL, HEK293T cells were pre-seeded into a 96-well plate one day before transfection. Then the cells were co-transfected with: 1) non-sense oligonucleotides control and wild type CTDSPL-3′UTR plasmid; 2) miR-18a-5p mimics and wild type CTDSPL-3′UTR plasmid; 3) miR-18a-5p mimics and mutated CTDSPL-3′UTR plasmid; 4) miR-18a-5p mimics and deleted CTDSPL-3′UTR plasmid. After 48 hours post-transfection, Renilla and firefly luciferase activities were measured using the Dual-Glo Luciferase Assay System (Promega, Madison, WI, USA) according to the manufacturer’s instructions.

Statistical analysis

All data were presented as means ± SEM. Statistical analysis was performed using Graphpad Prism (Graphpad Software. San Diego, CA, USA). A two-tail student’s t-test was performed to evaluate the significance level of the two groups. A p-value of less than 0.05 was considered statistically significant.