All animal and surgical procedures conformed to Directive 2010/63/EU, issued by the European Parliament. All animals were handled according to the guidelines of the TCM Animal Research Committee (TCM-LAEC20221178) of Tianjin University of Traditional Chinese Medicine. C57BL6/J mice (20 $\pm$ 20 g) were provided by Beijing Vital River Laboratory Animal Technology (Beijing, China). Lyz2Cre and VEGF-C${}^{f\mathit{}l\mathrm{/}f\mathit{}l}$ mice were provided by GemPharmatech (Jiangsu, China). PFKFB3${}^{f\mathit{}l\mathrm{/}f\mathit{}l}$ mice were provided by Shanghai Biomodel Organism Science & Technology Development (Shanghai, China). Myh6-mCherry B6 D2-tg mice were obtained from The Jackson Laboratory. Male mice were solely utilized as the animal participants in this study. The animals were housed in a specific-pathogen-free (SPF) animal facility in the experimental animal center of Tianjin University of Traditional Chinese Medicine, under controlled temperature (22 $\pm$ 2 °C) and humidity (40 $\pm$ 5%), and with a 12 hours light/dark cycle, and received a standard pellet diet with continual access to water.

Specific VEGFR3 knockdown or overexpression was achieved in the left ventricles of the mice by using adeno-associated viruses-based (AAV9) delivery vectors (Genechem, Shanghai, China). AAV9 expressing the VEGFR3 open reading frame (titer: 1.76 $\times$ 10${}^{12}$), CCR2-shRNA (titer: 1.65 $\times$ 10${}^{12}$), or a scrambled control sequence bearing no homology to known gene transcripts (titer: 1.54 $\times$ 10${}^{12}$) were injected into the left ventricle at multiple sites. After a period of 30 days, the mice were subjected to I/R surgery and subsequent treatments.

I/R models were prepared according to our previous studies [13, 14], and the left anterior descending coronary artery (LAD) was ligated through the utilization of a 6-0 silk suture. Following a duration of 60 minutes, the suture was removed to facilitate reperfusion. The experimental cohort underwent a procedure involving ligation, whereas the control group underwent a sham procedure that mimicked the experimental intervention, although without the actual ligation. In order to apply the VEGFR3 inhibitor as a treatment, mice were subjected to an intraperitoneal injection of MAZ51 at a dose of 10 mg/kg per bodyweight (#676492-10MG, sigma-Aldrich, Shanghai, China), as previously described [15, 16].

Mice echocardiography was performed using a ultra-high-resolution animal ultrasound imaging system (Vevo 2100, VisualSonic, Toronto, IL, Canada) and the left ventricular function was evaluated as previously described [16]. Long axes were observed through the utilization of M-mode imaging. Three representative cycles were captured for each animal, and measurements were performed for ejection fraction (EF%) and fractional shortening (FS%).

Bone marrow-derived macrophages (BMDM) were procured from bone marrow, utilizing the previously established method [17, 18]. The femur bones of the mice were subjected to sterilization via a 10-minute treatment with 75% ethanol, followed by flushing using Iscove’s Modified Dulbecco’s Medium (IMDM), containing 2% fetal bovine serum (FBS). After treating the red blood cells with red blood cell lysis buffer, the entire solution was centrifuged. Then, the resulting cell pellet was suspended in fresh IMDM containing 10% FBS and 20 ng/mL GM-CSF. Cells were all cultured in a humidified incubator at 37 °C and 5% CO${}_2$. Cell cultures were replenished with fresh IMDM every third day. After a period of 10 days, the macrophages were acquired. The polarization of macrophages towards M1 can be achieved by subjecting them to 1 µg/mL LPS (Sigma-Aldrich, #L4391-10X1MG) and 50 ng/mL IFN-$\gamma$ (#APA033Mu61, Cloud-Clone Corp, Wuhan, China) stimulation for a duration of 24 hours. The inhibition of macrophage glycolysis was achieved through the implementation of 3PO (10 µM/L; #HY-19824, MedChemExpress, Shanghai, China). For cell transfections, macrophages were stably transfected with VEGFR-overexpressing lentivirus (Genechem, Shanghai, China) or lentivirus containing scrambled control sequences. Briefly, macrophages were inoculated into 6-well plates and cultured at 37 °C. The infection reagent and lentivirus were added to macrophages for 72 hours, according to the manufacturer’s instructions. All cell lines were validated by STR profiling and tested negative for mycoplasma.

For single-cell suspensions, the Multi-Tissue Dissociation Kit 2 (#130-110-203, Miltenyi, Bergisch Gladbach, Germany) and gentleMACS Dissociator device (Miltenyi) were used to dissociate cells from mice hearts. The collected cells were incubated with antibodies at 37 °C for 20 minutes. F4/80${}^{+}$ macrophages were isolated and purified using a FACSAria II (BD, Franklin Lakes, NJ, USA) flow cytometer. Flow cytometry was performed using a Flow Cytometer (NovoCyte, Agilent, Santa Clara, CA, USA). Data were analyzed using FlowJo software (Version V10.8; TreeStar, Ashland, OR, USA) and NovoExpress (Version 15.0; Agilent). The antibodies used in this study were anti-mouse CD16/32 (TruStain FcX™, Biolegend, San Diego, CA, USA), CD45 antibody, anti-mouse, APC (#130-123-784, Miltenyi); CD11b antibody, anti-human/mouse, FITC (#130-113-796, Miltenyi); PerCP anti-mouse F4/80 antibody (#123126, Biolegend); Ly6G antibody, PE, eBioscience™, REAfinity™ (#12-9668-82, Thermo Fisher, Waltham, MA, USA); APC/cyanine7 anti-mouse Ly-6C antibody (#128026, Biolegend); FITC anti-mouse CD64 antibody (#161008, Biolegend).

Extracellular acidification rates (ECARs) were assessed using a Seahorse analyzer (XF96, Agilent), as previously described [19]. Briefly, macrophages were cultured in 96-well plates followed by the sequential addition of 20 mM glucose, 2 µM oligomycin, and 80 mM 2-DG. ECARs were calculated using the preloaded Seahorse analyzer software.

Following the collection process, plasma was subjected to centrifugation at a speed of 3000 rotations per minute (rpm) for a duration of 15 minutes to facilitate the estimation of biochemical parameters. ANP and BNP were detected in the plasma using an enzyme-linked immunosorbent assay (ELISA) kit (ANP: #SEA225Mu; BNP: #SEA541Mu, Cloud-Clone Corp), as per the instructions provided by Cloud-Clone Corp.

After being immersed in 4% paraformaldehyde for a duration of 72 hours, the heart tissues were fixed. Subsequently, the tissues were embedded in paraffin and sectioned into 4 µm thick pieces. Then, these sections were subjected to Masson staining. For immunofluorescence (IF) analysis, sections were incubated with LYVE1 (#ab218535, Abcam, Cambridge, MA, USA), Prox1 (#ab199359, Abcam), and CD68 (#ab283654, Abcam) at 4 °C for 12 hours, and subsequently incubated with anti-Rabbit HRP (Alexa Fluor 647) and anti-Mouse HRP (Alexa Fluor 488) at 37 °C for 2 hours. After washing with PBS, the sections were counterstained with DAPI. The sections were examined using a fluorescence microscope and a digital camera (3DHistech, Budapest, Hungary).

According to the guidelines provided by the manufacturer (Roche, Basle, Switzerland), mRNA reverse transcription was conducted using the mRNA reverse transcription kit to generate cDNA. RT-PCR was conducted by a CFX96TM PCR detection system (BioRad, Redmond, WA, USA) using the SYBR Green PCR master mix (Roche). GAPDH and U87 were employed as internal reference genes for normalization purposes. All used primer sequences are provided below. GAPDH forward: TGTGTCCGTCGTGGATCTGA; reverse: CCTGCTTCACCACCTTCTTG. U87 forward: ACAATGATGACTTATGTTTTT; reverse: GCTCAGTCTTAAGATTCTCT. IL-1$\beta$ forward: TGTGCTCTGCTTGTGAGGTGCTG; reverse: GCCGTGGTTGGAGAGATAGG. IL-6 forward: TAGTCCTTCCTACCCCAATTTCC; reverse: TTGGTCCTTAGCCACTCCTTC. TNF-$\alpha$ forward: GGTGATCGGTCCCCAAAGGGATGA; reverse: TGGTTTGCTACGACGTGGGCT. PKM2 forward: GTGGAGATGCTGAAGGAGATG; reverse: CAACAGGACGGTAGAGAATGG. GLUT1 forward: CGTCGTTGGCATCCTTATTG; reverse: CTTCTTCAGCACACTCTTGG. PFKFB3 forward: TGGTTGTATCAGTGGAAGGAAGG; reverse: GGAAGGAAGGAAGGAAGGAAGG. PFKFB2 forward: GGCAACATCCTCGTTATCTC; reverse: GTCTACAGCATCCACATTCAG. HK2 forward: TGCGAATATGGTTGCCTCATCTTG; reverse: CTCCTCCTCCTCCTCCTCTTCC. VEGF-C forward: CCGCTGTGTCCCATCGTATTG; reverse: AGTCCTCTCCCGCAGTAATCC. VEGF-D forward: AGTTATAGATGAAGAATGGCAGAG; reverse: CTTGAAGAATGTGTTGGTTGTC. Pdpn forward: CACCTCAGCAACCTCAGACC; reverse: TAACAAGACGCCAACTATGATTCC. CcL21a forward: CTGCTTCAACCATCACATCTG; reverse: GCTGTCTCCTTCCTCATTCC.

The data were analyzed using SAS statistical software (v9.4, SAS Institute, Cary, NC, USA). The mean $\pm$ standard deviation (SD) was used to present all quantitative data in this study. Pairwise comparisons were conducted using two-tailed Student’s t-tests. To compare differences across multiple groups with a single variable, a one-way analysis of variance (ANOVA) was performed. A p value 0.05 was considered statistically significant. The figure legends contain all numerical data pertaining to the experiments conducted.

In this research, we utilized intramyocardial injections of adeno-associated viruses (AAV9) to inhibit (VEGFR3 shRNA) or promote the expression of VEGFR3 (VEGFR3 ORF) in the heart. Moreover, to serve as the control groups, AAV9-empty vectors were injected into the myocardium of the sham, I/R, and MAZ51 (a VEGFR3 tyrosine kinase inhibitor that selectively hinders the activation of VEGFR3 caused by VEGF-C) groups. To establish a model of ischemic heart failure (HF), the approach of subjecting the heart to 1 hour of cardiac ischemia followed by 14 days of reperfusion (I/R) was employed (Fig. 1A). Cardiac ultrasound was performed to assess the cardiac function of each experimental cohort (Fig. 1B–D). The M-mode pattern of the cardiac ultrasound, which served as a representative, indicated that the state of the left ventricular wall motion was considerably diminished in both the MAZ51 intervention and the VEGFR3 shRNA groups, in contrast to the I/R group. Conversely, the VEGFR3 ORF group significantly improved the state of the left ventricular wall motion following I/R (Fig. 1B). The measurement outcomes of the left ventricular ejection fraction (Fig. 1C, EF%) and short axis shortening rate (Fig. 1D, FS%) demonstrated a consistent trend; notwithstanding the reduction in MAZ51 and VEGFR3 shRNA, where no significant difference was observed, while VEGFR3 ORF displayed a noteworthy augmentation. In contrast to the I/R group, the degree of cardiac fibrosis in mice subjected to I/R and treated with MAZ51 and VEGFR3 shRNA exhibited a significant increase. Conversely, VEGFR3 ORF was found to notably ameliorate cardiac fibrosis in mice (Fig. 1E,F). Therefore, to evaluate lymphatic neogenesis in the heart of I/R mice in diverse experimental groups, immunofluorescence staining of lymphatic markers, such as lymphatic endothelial hyaluronic acid receptor 1 (Lyve1) and homeobox transcription facter1 (Prox1), was employed. The results demonstrated that the interference of MAZ51 and VEGFR3 shRNA had an adverse impact on the development of cardiac lymphatic vessels in mice that underwent I/R, whereas the administration of VEGFR3 ORF demonstrated a marked improvement and facilitation of cardiac lymphatic neogenesis (Fig. 1G–I). The process of phagocytosis, in which macrophages engulf deceased cells, results in the transportation of cardiac antigens to the mediastinal lymph nodes (MLNs) [20]. In these nodes, the antigens may be presented to T cells, which are already present in the area, by a process known as cross-presentation [21]. Thus, to assess the impact of lymphangiogenesis on the drainage of myocardial cell debris towards the MLNs, an analysis was conducted on the level of cardiogenic (Myh6-mCherry) antigen present in the MLNs. The flow cytometry findings revealed that the number of cardiac mCherry antigens in the drained MLNs was significantly reduced upon inhibition of lymphangiogenesis (MAZ51 and VEGFR3 shRNA), whereas promoting lymphangiogenesis significantly increased the aforementioned number (Fig. 1J,K). The mRNA expression of genes linked with lymphatic neogenesis, including VEGFR3, VEGF-D, podoplanin (Pdpn), and Ccl21, displayed coherence with the preceding results (Fig. 1L). Furthermore, the serum levels of ANP and BNP (Fig. 1M), which are markers of heart failure, indicated that hindering lymphangiogenesis through the use of MAZ51 and VEGFR3 shRNA could considerably accelerate the advancement of the disease. Conversely, promoting lymphangiogenesis yielded contrasting outcomes. In conclusion, the promotion of lymphangiogenesis proved to be advantageous in the restoration of the heart following I/R.

Fig. 1.

Promoting lymphatic regeneration is beneficial for heart repair after ischemia-reperfusion (I/R). (A) Schematic overview of experimental procedure. (B) Representative image of M-mode in a cardiac ultrasound. Quantification of (C) EF% and (D) FS%, 14 days after I/R, data are presented as mean $\pm$ SD of n = 8–10 mice. (E–F) Representative Masson staining of heart transverse sections (scale bar = 1000 µm) and quantified IS/IV (infarct size: IS; left ventricular rate: LV, data are presented as mean $\pm$ SD of n = 3. (G–I) Representative immunostaining of Lyve1 and Prox1 in heart transverse sections (scale bar = 20 µm) and quantified Lyve1 and Prox1 area%, data are presented as mean $\pm$ SD of n = 3. (J–K) MLNs were harvested, and flow cytometric analysis of CD45${}^{+}$ CD11b${}^{+}$ CD64${}^{+}$ mCherry${}^{+}$ cells was performed, data are presented as mean $\pm$ SD of n = 3 mice. (I) The mRNA levels of VEGF-C, VEGF-D, podoplanin (Pdpn), and Ccl21 in heart, data are presented as mean $\pm$ SD of n = 3 mice. (L) The mRNA expression of VEGF-C, VEGF-D, podoplanin (Pdpn), and Ccl21, data are presented as mean $\pm$ SD of n = 6 mice. (M) The levels of ANP and BNP in serum, data are presented as mean $\pm$ SD of n = 6 mice. Statistical significance is shown as: **p 0.01, compared to the Sham group; ${}^{\mathrm{\#}}$p 0.05 and ${}^{\mathrm{\#}\mathrm{\#}}$p 0.01, compared to the I/R group.

The ability to produce VEGF-C has been observed in cells that exhibit macrophage-like characteristics [7]. The current study employed Lyz2${}^{\text{𝐶𝑟𝑒}}$ VEGF-C ${}^{f\mathit{}l\mathrm{/}f\mathit{}l}$ mice to abolish the expression of VEGF-C in macrophages, where Lyz2 expression is evident in myeloid cell lines, such as monocytes and fully developed macrophages. The outcomes of immunofluorescence staining for CD68 and Lyve1 revealed that the ablation of VEGF-C expression in macrophages led to a significant decrease in the number of lymphatic vessels following I/R (Fig. 2A,B). The results obtained by flow cytometry analysis revealed that the ablation of VEGF-C expression in macrophages led to a marked increase in the number of macrophages present in the heart (Fig. 2C,D). Furthermore, it caused a significant elevation in the mRNA expression of inflammatory factors, such as IL-1$\beta$, IL-6, and TNF-$\alpha$, in the heart (Fig. 2E). It has been established that the suppression of glycolysis could effectively impede the inflammatory reaction of macrophages [22, 23]. However, the correlation between glycolysis inhibition and the production of VEGF-C by macrophages remains ambiguous. Notably, following I/R, there was an upregulation in the expression of glycolytic-associated genes, including PKM2, GLUT1, PFKFB3, and HK2, within the cardiac macrophages, obtained by flow cytometry (FACS) (Fig. 2F). The glycolytic pressure test revealed a marked elevation in extracellular acidification rates (ECARs) post-I/R, indicating heightened glycolytic activity in cardiac macrophages following this event (Fig. 2G). In order to assess the impact of glycolysis inhibition on lymphatic neogenesis, we employed LPS to provoke inflammation and heighten glycolysis in bone marrow-derived macrophages (BMDM) of I/R mice [24]. Then, we utilized the PFKFB3 inhibitor 3PO to intervene, as well as using adenovirus overexpression of PFKFB3. The outcomes demonstrated that the inhibition of PFKFB3 had a significant impact on the glycolytic activity of BMDM, causing a decrease in its level (Fig. 2H). Moreover, the inhibition of PFKFB3 led to a significant increase in the expression of VEGF-C and VEGF-D in BMDM (Fig. 2I). In addition, 3PO could reduce the glycolysis level of inactive macrophages (Fig. 2H) yet have no considerable influence on the expression of VEGF-C and VEGF-D (Fig. 2I). These results suggested that inhibiting the level of glycolysis in an activated state of macrophages can lead to an increase in the expression of VEGF-C and VEGF-D. Conversely, the overexpression of PFKFB3 resulted in a significant increase in the glycolytic activity of BMDM (Fig. 2J), which was accompanied by a notable decrease in the expression of VEGF-C and VEGF-D (Fig. 2K). To summarize, the expression of PFKFB3 in macrophages possesses a strong correlation with the production of VEGF-C. Moreover, it has been observed that hindering the function of PFKFB3 could stimulate the generation of VEGF-C.

Fig. 2.

PFKFB3 regulates the production of VEGF-C by macrophages. (A–B) Representative immunostaining of Lyve1 and CD68 in heart transverse sections (scale bar = 20 µm) and quantified Lyve1 and CD68 area%, data are presented as mean $\pm$ SD of n = 3. (C–D) Flow cytometric analysis of CD11b${}^{+}$ F4/80${}^{+}$ cells in the heart, data are presented as mean $\pm$ SD of n = 3 mice. (E) The mRNA levels of IL-1$\beta$, IL-6, and TNF-$\alpha$ in the heart, data are presented as mean $\pm$ SD of n = 3 mice. (F) The mRNA levels of PKM2, GLUT1, PFKFB3, PFKFB2, and HK2 in F4/80${}^{+}$ macrophages in the heart obtained by flow cytometry (FACS), data are presented as mean $\pm$ SD of n = 3 mice. (G) ECAR measurements in heart F4/80${}^{+}$ macrophages, as measured by Seahorse Bioscience XF96 analyzer, n = 3. (H,J) ECAR measurements in BMDMs of I/R mice, as measured by Seahorse Bioscience XF96 analyzer, n = 3. (I,K) The mRNA levels of VEGF-C and VEGF-D in BMDMs of I/R mice, data are presented as mean $\pm$ SD of n = 3 mice. Statistical significance is shown as: *p 0.05 and **p 0.01, compared to the sham group; ${}^{\mathrm{\#}\mathrm{\#}}$p 0.01, compared to the I/R group.

In our study, we employed Lyz2${}^{\text{𝐶𝑟𝑒}}$ PFKFB3${}^{f\mathit{}l\mathrm{/}f\mathit{}l}$ mice to ablate the expression of PFKFB3 in macrophages with the aim of investigating the impact of PFKFB3 on lymphatic neogenesis in the context of I/R. The findings of the cardiac ultrasound indicate that the inhibition of PFKFB3 expression in macrophages could enhance the left ventricular wall motion in mice following I/R (Fig. 3A), leading to a notable improvement in ejection function (EF%) in mice with I/R (Fig. 3B). According to Masson’s research, the inhibition of PFKFB33 expression in macrophages could decrease the degree of cardiac fibrosis in I/R mice (Fig. 3C,D). Additionally, immunofluorescence staining of Lyve1 and Prox1 confirmed that this intervention resulted in an improvement in lymphatic neogenesis (Fig. 3E–G). The mRNA expressions of VEGF-C, VEGF-D, Pdpn, and Ccl21 were significantly increased following the deletion of PFKFB3 in mice subjected to I/R (Fig. 3H). Flow cytometry was employed to separate cardiac macrophages and qPCR was used to detect the mRNA expression of VEGF-C, IL-1$\beta$, IL-6, and TNF$\alpha$ in cardiac macrophages. The results indicated that the expression of VEGF-C in PFKFB3${}^{\mathrm{-}\mathit{}\mathrm{/}\mathrm{-}}$ cardiac macrophages was significantly higher than the PFKFB3${}^{f\mathit{}l\mathrm{/}f\mathit{}l}$ cardiac macrophages (Fig. 3I), and the mRNA expression of IL-1$\beta$, IL-6, and TNF$\alpha$ in PFKFB3${}^{\mathrm{-}\mathit{}\mathrm{/}\mathrm{-}}$ cardiac macrophages was significantly lower than the PFKFB3${}^{f\mathit{}l\mathrm{/}f\mathit{}l}$ cardiac macrophages (Fig. 3J). The evidence gathered suggests that PFKFB3 may have a role in modulating the lymphangiogenesis function during I/R injury.

Fig. 3.

Elimination of PFKFB3-enhanced lymphangiogenesis following I/R. (A) Representative image of M-mode in a cardiac ultrasound of Lyz2${}^{\text{𝐶𝑟𝑒}}$ PFKFB3${}^{f\mathit{}l\mathrm{/}f\mathit{}l}$ mice. Quantification of (B) EF% after I/R, data are presented as mean $\pm$ SD of n = 6 mice. (C,D) Representative Masson staining of heart transverse sections (scale bar = 1000 µm) and quantified IS/IV (infarct size: IS; left ventricular rate: LV, data are presented as mean $\pm$ SD of n = 3. (E–G) Representative immunostaining of Lyve1 and Prox1 in heart transverse sections (scale bar = 20 µm) and quantified LYVE1 and Prox1 area%, data are presented as mean $\pm$ SD of n = 3. (H) The mRNA levels of heart VEGF-C, VEGF-D, podoplanin (Pdpn), and Ccl21, data are presented as mean $\pm$ SD of n = 3 mice. (I) The mRNA levels of VEGF-C in cardiac macrophages, data are presented as mean $\pm$ SD of n = 3 mice. (J) The mRNA levels of IL-1$\beta$, IL-6, and TNF$\alpha$ in cardiac macrophages, data are presented as mean $\pm$ SD of n = 3 mice. Statistical significance is shown as: *p 0.05 and **p 0.01, compared to the Sham group.

The quantification of neutrophils, monocytes, and macrophages in the heart following I/R was assessed through the utilization of flow cytometry (Fig. 4A–D). The findings indicated that the elimination of PFKFB3 resulted in a substantial decrease in the population of neutrophils (Fig. 4B,E), monocytes (Fig. 4C,F), and macrophages (Fig. 4D,G) within the heart of mice undergoing I/R. Additionally, there was a significant reduction in the expression of inflammatory factors (IL-1$\beta$, IL-6, and TNF$\alpha$) in the heart of Lyz2${}^{\text{𝐶𝑟𝑒}}$ PFKFB3${}^{f\mathit{}l\mathrm{/}f\mathit{}l}$ mice (Fig. 4H). In brief, our study has substantiated that the elimination of PFKFB3 expression in macrophages could considerably augment I/R lymphangiogenesis, bolster the robustness of cardiac immune cells, and curtail the inflammatory reaction in the cardiac tissue.

Fig. 4.

Absence of PFKFB3 in macrophages results in improved stability of cardiac immune cells post I/R. (A) Flow cytometric analysis of (B,E) neutrophils (CD11b${}^{+}$ Ly6G${}^{+}$), (C,F) monocytes (CD11b${}^{+}$ Ly6C${}^{\text{high/low}}$), and (D,G) macrophages in the heart (CD11b${}^{+}$ F4/80${}^{+}$), data are presented as mean $\pm$ SD of n = 3 mice. (H) The heart mRNA levels of IL-1$\beta$, IL-6, and TNF$\alpha$, data are presented as mean $\pm$ SD of n = 3 mice. Statistical significance is shown as: **p 0.01, compared to the Sham group.