*2.8. Quantitative Real-Time PCR*

Total RNA was extracted from the cell line using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's protocol. Relative quantitation by realtime PCR was performed using SYBR Green to detect PCR products in real-time using the QuantStudioTM3 system (Applied Biosystems). A melting curve analysis was performed at the end of each PCR reaction. MMP-9 gene expression was expressed as a ratio to that of GAPDH, a housekeeping gene. Oligonucleotide primer sequences were as follows: Mmp-9, forward 5'-GGACCCGAAGCGGACATTG-3' and reverse 5'- CGTCGTCGAAATGGGCATCT-3'; Gapdh, forward 5'-TGGATTTGGACGCATTGGTC-3' and reverse 5'-TTTGCACTGGTACGTGTTGAT-3'.

#### *2.9. Surface Plasmon Resonance (SPR) Spectroscopy*

Experiments were performed at 25 ◦C using a Biacore T200, and the data were analyzed using Biacore T200 evaluation software 2.0 (GE Healthcare, Stockholm, Sweden). Human

MMP-9 recombinant protein (911-MP; R&D Systems Incorporated, Minneapolis, MN, USA) was covalently coupled to a CM5 chip (GE Healthcare). All measurements were performed at 25 ◦C, using a TCNB buffer: 50 mmol·L−<sup>1</sup> Tris, 10 mmol·L−<sup>1</sup> CaCl2, 150 mmol·L−<sup>1</sup> NaCl, 0.05% Brij-35 (*w*/*v*), and pH 7.5, and metformin was injected in a two-fold dilution concentration series (range, 0.0156–15.6 <sup>μ</sup>mol·L−1). Steady-state values were calculated from the sensorgrams and plotted against concentrations. Data were fitted into a single-site binding model to calculate the KD value.

#### *2.10. Zymography*

Gelatinase activity was detected in HEK293A supernatants and recombinant human MMP-9 protein (911-MP; R&D Systems Incorporated, Minneapolis, MN, USA) after metformin incubation for 24 h. Zymography was performed according to the manufacturer's instructions (Applygen, P1700, Beijing, China). Following electrophoresis, the gels were washed twice with 2.5% Triton X-100 to remove sodium dodecyl sulfate and further washed with 50 mmol·L−<sup>1</sup> Tris–HCl pH 8.0. Gels were incubated for the following 20 h in an activation buffer (50 mmol·L−<sup>1</sup> Tris–HCl supplemented with 5 mmol·L−<sup>1</sup> CaCl2). The gels were stained with Coomassie brilliant blue R-250 and de-stained with 20% methanol and 10% acetic acid in distilled water until clear bands were visualized.

#### *2.11. Statistics*

Data are expressed as mean ± SD. All samples were independent, including those measured over time in the experiments. For parametric data, Student's *t*-test or an analysis of variance (ANOVA) was used to analyze intergroup differences for normally distributed data. For parametric data with unequal variances, ANOVA with Tukey's post hoc test was used. For non-parametric data, the Mann–Whitney U test with the exact method was used to analyze intergroup differences. A Kruskal–Wallis ANOVA combined with post hoc Tukey's multiple comparison tests was performed when more than two groups were evaluated. Data were analyzed using GraphPad Prism software (version 8.0; GraphPad Software Inc., San Diego, CA, USA), and *p* < 0.05 was considered statistically significant.

#### **3. Results**

#### *3.1. Matrix Metalloproteinase-9 (MMP-9) Is Predicted to Bind Directly to Metformin*

We hypothesized that metformin inhibits MMP-9 activity through its direct interaction with MMP-9. Molecular modeling was performed to rationalize the activities of metformin against MMP-9. Metformin was situated in the active cavity, engaging in several interactions with MMP-9 (Figure 1a). Two hydrogen bonds were between the urea moiety and Pro-246 and Glu-227. Additionally, the protonated imine group formed an ionic bond with Glu-227. Notably, metal coordination was observed between metformin and the zinc ions, which might have strengthened the binding affinity. As shown in the protein–ligand contact histogram, the results were consistent with those of the docking study. The two hydrogen bonds formed by Pro-246 and Glu-227 were maintained at 76% and 30% of the simulation time, respectively (Figure 1b,c). A powerful coordination bond was formed between the nitrogen atom of metformin and the zinc metal ions. In addition, the amino group formed a hydrogen bond network through a water bridge with Ala-189. Further molecular dynamic (MD) simulation analysis revealed that the complex was stable during a 50 ns simulation (Figure 1d). Overall, these findings provided a better understanding of the metformin mechanisms and may facilitate a future search for optimized MMP-9 inhibitors.

#### *3.2. Metformin Directly Interacts with MMP-9 and Attenuates Its Activity*

To verify whether metformin directly binds to MMP-9, we conducted surface plasmon resonance (SPR) experiments. The findings of the SPR-based assay suggested that the binding of metformin to MMP-9 occurred with a KD of 0.6950 <sup>μ</sup>mol·L−<sup>1</sup> (Figure 2a,b). To examine the ability of metformin to inhibit MMP-9 activity, we constructed an overexpres-

sion plasmid for human MMP-9 and transfected it, or a control plasmid into HEK293A cells using lipofectamine (Figure S1). Next, we incubated the transfected cells with metformin (1 <sup>μ</sup>mol·L−1) for 24 h, and the MMP-9 activity in the cultured supernatant was detected using zymography and an MMP activity assay. Both results indicated that metformin incubation significantly attenuated the activity of MMP-9 (Figure 2c–e).

**Figure 1.** Matrix metalloproteinase-9 (MMP-9) is predicted to bind directly to metformin. (**a**) Predicted binding mode of metformin with MMP-9 (PDB id: 4wzv). (**b**,**c**) Protein–ligand contact histogram of metformin and the corresponding two-dimensional diagram predicted through MD simulations. A percentage value suggests that for X% of the simulation time, the specific interaction is maintained. (**d**) RMSD of the interaction between MMP-9 and the ligand metformin in MD simulations. MD, molecular dynamics; RMSD, root mean square deviation; MMP-9, matrix metalloproteinase-9.

**Figure 2.** Metformin directly interacts with MMP-9 and attenuates its activity. (**a**,**b**) SPR analysis of the binding between metformin and MMP-9. Recombinant human MMP-9 protein was immobilized on an activated CM5 sensor chip, and metformin was then flowed across the chip. (**c**,**d**) Representative gelatin zymogram and the quantified values of the 92 kDa MMP-9 activity in the cultured supernatant. Data are shown as mean ± SD (two-way ANOVA followed by Tukey's test, n = 5). (**e**) MMP activity in supernatant from cultured HEK293A cells was measured using a Gelatinase Assay Kit (two-way ANOVA followed by Tukey's test, n = 5). (**f**) Representative gelatin zymogram of the recombinant human MMP-9 activity after incubation with different concentrations of metformin. (**g**) The exogenous MMP-9 protein level in HEK293A cells after incubation with metformin for 24 h (both of the 2 bands were quantified, two-way ANOVA followed by Tukey's test, n = 5). (**h**) MMP-9 mRNA expression level in HEK293A cells after incubation with metformin for 24 h (n=6). (**i**,**j**) Exogenous MMP-9 degradation in metformin-treated HEK293A cells when protein synthesis was inhibited by 10 μM cycloheximide (two-way ANOVA followed by Tukey's test, n = 5).

To verify whether the inhibition of MMP-9 activity by metformin is a direct binding effect, we conducted a test tube experiment. The results showed that the activity of MMP-9 was not affected by metformin binding directly to MMP-9 (Figure 2f). However, Western blotting results suggested that metformin treatment could decrease the protein level of MMP-9 (Figure 2g). Western blot analysis of MMP-9 in the total cell lysate consistently revealed two bands of apparent molecular masses of 85 and 92 kDa. It was previously shown that the 85 kDa band represents an underglycosylated precursor form of MMP-9 found intracellularly, whereas the 92 kDa band represents a fully glycosylated mature form that is secreted into the extracellular space [21]. Further, we detected the transcription level of MMP-9 by polymerase chain reaction and found that metformin did not change the mRNA level of MMP-9 (Figure 2h).

Accordingly, we became interested in establishing whether metformin downregulated the MMP-9 protein level by driving its degradation. To this end, we used eukaryotic inhibitor cycloheximide to inhibit protein synthesis in HEK293A cells to study the degradation of MMP-9 with or without metformin. We found that the exogenous MMP-9 protein was continuously degraded from 1 to 3 h, and metformin treatment effectively decreased MMP-9 protein expression by accelerating its degradation (Figure 2i,j).

#### *3.3. Metformin Inhibits Local Plaque and Circulation MMP-9 Activity in ApoE-/- Mice*

To further confirm whether metformin inhibits MMP-9 activity in vivo, we constructed a carotid artery plaque model in ApoE-/- mice (Figures S2 and S3) [19]. After a consecutive 14-day metformin treatment (200 mg·kg<sup>−</sup>1) by intragastric gavage (Figures 3a and S4), we found that active MMP-9 and MMP-9 expression decreased in the plaque by immunofluorescence staining. However, metformin did not affect MMP-2/12 expression, which was reported to be related to plaque instability (Figure 3b,c). Moreover, the serum MMP-9 activity was detected using an MMP activity assay (Figure 3d). The results showed that metformin treatment inhibited local plaque and circulating MMP-9 activity.

**Figure 3.** Metformin inhibits local plaque and circulating MMP-9 activity in ApoE-/- mice. (**a**) Flowchart illustrating the experimental procedure for actuating metformin treatment in a collar-induced carotid

atherosclerotic plaque model. (**b**) Representative images of immunofluorescence staining for activematrix metalloproteinase (MMP)-9, MMP-2, and MMP-12 in plaque after metformin treatment. Scale bars represent 20 μm. (**c**) Quantification of immunofluorescence staining for MMP family in plaque after metformin treatment. Unpaired Student's *t*-test, n = 6 per group. (**d**) A Gelatinase Assay Kit was used to detect relative MMP activity in serum. Unpaired Student's *t*-test, n = 6 per group.

#### *3.4. Metformin Improves Atherosclerotic Plaque Stability in ApoE-/- Mice*

To determine the protective effects of metformin on atherosclerosis, we assessed the vulnerability index (VI) of the RCCA plaque using histology. The composition of plaques, including macrophages, collagen, lipids, and smooth muscle cells (SMCs) was demonstrated by CD-68, Sirius red staining, oil red O staining, and α-SMA immunostaining, respectively (Figure 4a,b). Sirius red staining results showed that the collagen content was preserved by the metformin treatment. Oil red O staining, α-SMA, and CD-68 immunofluorescence results suggested that there were no significant differences in lipid, SMCs, and macrophage content after metformin treatment. As each feature alone is insufficient for identifying high-risk plaques, the ratio between stable and unstable plaque components is often used to calculate the VI (macrophage content + lipid core content)/(SMC content + collagen content) in experimental studies [22]. The results showed that with the metformin treatment, plaque VI was significantly decreased, indicating that metformin had a beneficial effect on plaque stability (Figure 4c).

**Figure 4.** Metformin improves atherosclerotic plaque stability in ApoE-/- mice. (**a**) Representative images of Sirius red staining for plaque collagen, immunostaining for the macrophage marker CD-68, smooth muscle cell marker α-SMA, and oil red O staining for intimal lipid in plaque within the right common carotid artery. Scale bars for Sirius red staining and oil red O staining represent 200 μm and 20 μm for immunostaining. (**b**) Quantification of the positive area as a percentage of the whole plaque area. Unpaired Student's *t*-test, n = 6 per group. (**c**) The vulnerability index is calculated by dividing the area of macrophage+lipid by that of smooth muscle cells+collagen. Unpaired Student's *t*-test, n = 6 per group. Data are presented as the mean ± SD. HFD, high-fat diet; Chol, cholesterol.

#### **4. Discussion**

In this study, we demonstrated that metformin directly binds to MMP-9 and accelerates its degradation. Furthermore, we proved that metformin improved atherosclerotic plaque stability by inhibiting local plaque and circulating MMP-9 in ApoE-/- mice (Figure 5).

**Figure 5.** Schematic showing metformin directly binding to MMP-9 to improve plaque stability. MMP matrix metalloproteinase-9, ECM extracellular matrix.

Collagens are most abundant in the extracellular matrix, joined by elastin that confers elastic recoil to the artery [23]. Loss of collagen, which normally provides the main tensile strength of the artery wall, is an important cause of atherosclerotic plaque rupture, which underlies most cases of ACS [24]. MMPs have specific proteolytic activity against the ECM, which can result in the thinning of the fibrous cap and plaque instability [11,25]. MMP-9, also known as gelatinase B, is a widely investigated member of the MMP family. Studies have shown a strong relationship between MMP-9 and plaque instability [26,27], which indicates that MMP-9 may be a therapeutic target for preventing plaque instability. Currently, inflammatory pathways are the main therapeutic targets for plaque instability, such as the monoclonal antibody inhibiting interleukin-1β (called canakinumab) [28] and PCSK-9 inhibitors [29]. Both canakinumab and PCSK-9 inhibitors have anti-inflammatory effects. Moreover, PCSK-9 inhibitors also have an inhibitory effect on MMP-2, but cannot inhibit MMP-9 [30]. So, the mechanism by which canakinumab and PCSK-9 inhibitors stabilize plaques may be different from metformin. In addition, there are few plaquestabilizing drugs targeting MMP-9. Metformin interferes with the pathophysiology of multiple cancers and diabetes by reducing MMP-9 expression [31–33]. However, there are still many studies showing that metformin can increase MMP-9 expression [34], including some clinical trials [35,36]. Whether metformin stabilizes plaque by modulating MMP-9 activity and expression remains unknown. Our results indicated that metformin directly binds to MMP-9, and significantly downregulated MMP-9 expression/activity levels in local plaque and circulation, which may explain the role of metformin in improving plaque stability.

It is generally accepted that metformin inhibits pro-inflammatory cytokine release, such as IL-1β, IL-6, and TNF-α, to have anti-inflammatory effects [37–41]. Destabilization of the atherosclerotic plaque is associated with increased inflammatory cytokine production [42,43]. To investigate whether metformin protects plaque stability by inhibiting inflammation, we measured plaque IL-1β, IL-6, and TNF-α levels. The immunofluorescence staining results suggested that metformin treatment did not affect the levels of IL-1β, IL-6, and TNF-α in plaque (Figure S5). H. Wu et al. found that macrophage infiltration was significantly reduced after 16 weeks of metformin treatment [44]. However, our immunofluorescence staining results suggested that as short as two weeks of metformin treatment had no significant anti-inflammatory effect. This may have been due to the short

treatment time in our animal model. Additionally, metformin has been reported to promote macrophage cholesterol efflux, thus decreasing the lipid content of atherosclerotic plaques and increasing plaque stability [44]. In this study, after consecutive 14-day metformin treatment (200 mg·kg−1) by intragastric gavage, we found that only the collagen content of the plaque was preserved, whereas intimal lipids, macrophages, and SMCs showed no significant difference, indicating that metformin improved plaque stability by reducing ECM degradation.

Metformin has protective effects by activating AMPK in intact cells and in vivo [45]. AMPK confers benefits in chronic inflammatory diseases, such as atherosclerosis, independent of its ability to normalize blood glucose levels. There was evidence that metformin inhibited TNF-α-induced MMP-9 upregulation in neutrophils, which might have been mediated via an AMPK-dependent pathway [46]. Metformin administration suppressed MMP-9/MMP-2 and mTOR expression and increased Akt and AMPK expression, indicating that metformin reduced the expression of MMPs via the AMPK signaling pathway [47]. In this study, we first found that metformin binds to MMP-9. The MMP-9 binding regions of metformin are situated in the active cavity and engage in several interactions with MMP-9. Moreover, the combination of metformin and MMP-9 significantly accelerated MMP-9 protein degradation, which may also account for the effect of metformin downregulating MMP-9 expression level and improving plaque stability.

Protein homeostasis is responsible for basic cellular functions, such as the regulation of the level of key enzymes and the removal of abnormal proteins [48]. Our results suggested that the combination of metformin and MMP-9 significantly accelerated MMP-9 protein degradation. Chang Y et al. reported that cells treated with MG-132, a proteasome inhibitor, exhibited a significant MMP-9 protein accumulation compared to its accumulation in the untreated controls, indicating that the degradation of the MMP-9 protein is in a proteasomedependent manner. Moreover, SMURF1, an E3 ubiquitin ligase, binds MMP-9 to promote its degradation [49]. In this study, we first found that metformin binds to MMP-9. The MMP-9-binding regions of metformin are situated in the active cavity and engage in several interactions with MMP-9. Further, MMP-9 was shown to have two N-glycosylation sites, which seems to be important for MMP-9 protein structure stabilization and secretion, on asparagine residues at position 38 in the propeptide domain and in the catalytic domain at position 120 [50–52]. In subsequent research, we have two directions to further explore the potential mechanism of metformin regulation of MMP-9: (1) metformin affects the binding of MMP-9 to SMURF1, thus promoting MMP-9 ubiquitination and accelerating its degradation; (2) metformin affects the role of N-glycosylation in MMP-9 and decreases MMP-9 protein structure stabilization.

In conclusion, we have demonstrated that metformin directly binds to MMP-9 and accelerates its degradation, thus preserving the collagen content of plaque and improving atherosclerotic plaque stability. Further, these findings could significantly impact the development of the search for new drugs and pleiotropic mechanisms.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/jcdd10020054/s1, Figure S1: MMP-9 was successfully overexpressed in HEK293A cells. Figure S2: Serum triglyceride and total cholesterol levels increased in animal models. Figure S3: The carotid plaque model was successfully constructed. Figure S4: Metformin successfully activated AMPK. Figure S5: Metformin treatment had no significant anti-inflammatory effect in our model.

**Author Contributions:** Conceptualization, H.X., Y.Z., X.C. and S.W.; methodology and software, X.C. and S.W.; validation, formal analysis and data curation, X.C.; resources and technical supports, S.W., W.X. and M.Z.; writing—original draft preparation, X.C.; writing—review and editing, H.X. and Y.Z.; funding acquisition and project administration, H.X. and Y.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Natural Science Foundation of China [81830009 to Y.Z., 82030072 to H.X.], Michigan Medicine-PKUHSC Joint Institute for Translational and Clinical Research [BMU2019JI007 to Y.Z.], CAMS Innovation Fund for Medical Sciences to [No. 2021-I2M-5-003 to H.X.] and the Key Clinical Projects of Peking University Third Hospital [BYSYZD2019022 to H.X.].

**Institutional Review Board Statement:** The investigations conformed to the US National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996). Animal experiments were approved by the Committee of Peking University on Ethics of Animal Experiments (LA 2018-112) and conducted in accordance with the Guidelines for Animal Experiments, Peking University Health Science Center.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available in this article.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Journal of Cardiovascular Development and Disease* Editorial Office E-mail: jcdd@mdpi.com www.mdpi.com/journal/jcdd

Academic Open Access Publishing

www.mdpi.com ISBN 978-3-0365-8117-0