Cardiac lymphatics retain LYVE-1-dependent macrophages during neonatal mouse heart regeneration

Abstract

In adult mice, myocardial infarction (MI) activates the cardiac lymphatics, which undergo sprouting angiogenesis (lymphangiogenesis), drain interstitial fluid and traffic macrophages to mediastinal lymph nodes (MLNs). This prevents edema and reduces inflammatory/fibrotic immune cell content to improve cardiac function. Here we investigated the role of cardiac lymphatics and macrophage clearance across the neonatal mouse regenerative window. The response to injury revealed limited lymphangiogenesis and clearance of macrophages from postnatal day 1 compared to postnatal day 7 infarcted hearts. This coincides with the maturation of lymphatic endothelial cell junctions from impermeable to permeable and with altered signaling between lymphatic endothelial cells and macrophages. Mice lacking the lymphatic endothelial receptor-1 (LYVE-1), where macrophage lymphatic trafficking is impaired in adults, experienced worse long-term outcomes after MI induced at postnatal day 1, suggesting an alternative role for LYVE-1 in macrophages. Macrophage-specific deletion of Lyve1 during neonatal heart injury impaired heart regeneration. This study demonstrates that immature cardiac lymphatics are impermeable to clearance in early neonates, ensuring retention of pro-regenerative LYVE-1-dependent macrophages.

Fig. 1. Growth of the lymphatic network from P1 to P21.

Whole-mount antibody staining for VEGFR3 confirmed that cardiac lymphatics grow and sprout extensively during postnatal development. Although little growth or sprouting was observed in the first week of postnatal life (a,b), by P14 a dense network of lymphatic vessels had formed covering the entire dorsal side of the heart (c). This network continued growing in proportion to overall heart expansion (d). Quantification of cardiac lymphatics on the dorsal side confirmed that there is a significant increase in total vessel length and total number of endpoints during the second week of life (e,f). Data are presented as mean ± s.e.m.; n = 5 for each timepoint. Scale bar, 500 μm. Source data

Fig. 2. P1 versus P7 cardiac lymphatic responses immediately following MI and 7 days after MI compared to intact P8 and P14 hearts.

Injured hearts were harvested at P1 and P7 immediately after MI surgery to visualize the initial response of the lymphatic vessels and macrophages (a,c). Macrophages were evenly distributed across all areas of the heart, suggesting that the immune response to injury had not yet fully initiated (a,c). Lymphatics extended from the base to apex in both P1 (b) and P7 (d) hearts. During ligation, the suture captured lymphatics located in the area of injury (asterisk in second panel of b and d). Whole-mount immunostaining of C57BL/6 hearts for VEGFR3, combined with light-sheet imaging (e,g). Intact P8 hearts compared to 7 days after MI at P1 revealed a limited lymphangiogenic response after injury (e). By contrast, comparing P14 intact with P7 hearts 7 days after MI revealed expanded VEGFR3⁺ lymphatic vessels covering the injury site (g). There was a significant increase in the heart size after MI at both P1 and P7 compared to the respective control stage (f,h). Asterisk indicates suture site. LA, left atria; RA, right atria. Data are presented as mean ± s.e.m. n = 2 for each timepoint (a–d); n = 12 for P8, n = 4 for P1MI7dpi in f; n = 7 for P14, n = 4 for P7MI7dpi in h. Significant differences were calculated using an unpaired, two-tailed Student’s t-test. Scale bar, 0.5 mm. Source data

Fig. 3. Lymphatic vessel expansion and macrophage accumulation at the site of injury.

Serial sections and immunostaining for PDPN and LYVE-1 in intact P8 (a–e) versus injured hearts at P1 (day 7 after MI; f–j) and in intact P14 (k–o) versus injured hearts at P7 (day 7 after MI; p–t) confirmed the limited lymphangiogenic response relative to intact P8 and P14 controls (compare a–e to f–j and k–o to p–t) after MI at P1 compared to P7 (compare g–j to q–t). There was an increased number of lymphatic vessels with dilated lumen in P7 MI samples compared to the P14 intact controls (compare b to g and l to q). hCD68–eGFP⁺ macrophages were enriched at the site of injury after MI at P1 and P7 (h and r). Suture is visible through autofluorescence in f–h and j. Quantification of cardiac lymphatic lumen (u). Quantification of macrophage density (v). b–e, magnified view of a box; g–j, magnified view of f box; l–o, magnified view of k box; q–t, magnified view of p box. Data are presented as mean ± s.e.m. n = 2 hearts for P8, P1MI7dpi and P7MI7dpi and n = 3 hearts for P14 in u. The mean lumen size per timepoint was calculated from pooled results across hearts as technical replicates. n = 8, 27, 24 and 53 for P8, P1MI7dpi, P14 and P7MI7dpi, respectively, and n = 3 for P1MI7dpi in v and n = 4 for P7MI7dpi in v. Significant differences were calculated using one-way ANOVA followed by Tukey’s multiple comparisons test. Scale bar, 0.5 mm for a, f, k and p; 0.2 mm for magnified views. Source data

Fig. 4. Adoptive transfer of splenic hCD68–GFP⁺ monocytes and imaging of CX3CR1⁺ tissue-resident macrophages reveals different levels of clearance to MLNs after MI at P1 versus P7.

Schematic of the adoptive cell transfer approach using adult hCD68–eGFP transgenic mice as splenic GFP⁺ monocyte donors, for intramyocardial delivery into recipient neonatal CD1 mice at the time of MI surgery to assess immune cell trafficking (a). Immunostaining for CD68 and endogenous GFP fluorescence in tissue sections derived from MLNs of P1 and P7 mice that underwent MI, determining the presence of cleared CD68⁺GFP⁺ macrophages (b′ white arrows and c′ white arrowheads). CD68⁺GFP⁺ macrophages were substantially reduced in MLNs after MI at P1 compared to after MI at P7 (compare b and c). Visualization of endogenous GFP⁺ macrophages in MLNs from hCD68–eGFP mice confirmed minimal clearance at P1 after MI, which appeared increased at P7, compared to the respective intact controls that contained resident MLN GFP⁺ macrophages (compare d and e and compare f and g). Similar visualization in CX3CR1–eGFP mice also confirmed minimal clearance at P1 after MI, which increased at P7 (compare h and i and compare k and j). Quantification of macrophage numbers in the MLNs validated these observations and indicated that the difference in clearance at P1 versus P7 was significant (l,m). F4/80⁺ macrophages visualized within afferent lymphatic lumens of MLNs after MI at P7 but not at P1 and evidence of macrophage drainage disruption after P7 MI in the Lyve1^−/− mutant setting (n). b′–k′ indicate magnified view of panel boxes. Data are presented as mean ± s.e.m. In l, n = 4 for P8, n = 8 for P1MI7dpi, n = 5 for P14 and n = 7 for P7MI7dpi. In m, n = 7 for P8, n = 8 for P1MI7dpi, n = 9 for P14 and n = 10 for P7MI7dpi. Magnification boxes are illustrative. Quantification was conducted across the entire MLN area within 10-μm sections. Significant differences were calculated using one-way ANOVA followed by Tukey’s multiple comparisons test. Scale bars, 50 μm for b and c; 0.5 mm for d–k; 20 μm for d′–k′; and 250 μm for n. Source data

Fig. 5. Cell–cell junctions of cardiac LECs undergo transformation during postnatal development.

High-magnification confocal imaging of LECs within vessels stained for VE-cadherin and LYVE-1 enabled visualization of cell–cell junctions at different postnatal stages (a–i). The morphology of the junctions at P1 appeared to be continuous, resembling that of zippers (arrows in c). Zippered junctions were also observed at P7 (arrows in f), but there was also the emergence of discontinuous buttoned junctions (arrowheads in f) as well as those that were intermediate between zipper and button, indicative of a more cell-permeable endothelium. The more complete transformation to buttoned junctions was further evident by P14 (arrowheads in i), although some intermediate and zippered junctions were still evident at this stage (arrows in i). Quantification of the percent incidence of the three junction types (zippered, intermediate and buttoned) across the P1–P7–P14 timecourse (j) reveals the trend in transition from zippered (impermeable) to buttoned (permeable) during postnatal development. Macrophage morphology also transformed during this 2-week period. n = 5 for P1 and P14, n = 2 for P7; lymphatic vessel tips within the visual field were analyzed, 2–4 per heart. Significant differences were calculated using unpaired Student’s t-tests. Mean percent was plotted. Scale bars, 20 μm. Source data

Fig. 6. Expression of gene markers for macrophage and lymphatic endothelial cell populations.

Schematic of the generation of scRNA-seq datasets of control and injured P1 versus P7 CD1 mice. Hearts were harvested at 1 day after MI (P1MI1dpi and P7MI1dpi) or 7 days after MI (P1MI7dpi and P7MI7dpi). For the intact conditions, the samples were collected at either P1 or P7 (Intact P1 and Intact P7). The samples were FACS sorted using 7-AAD to isolate live cells, and libraries were prepared for sequencing using the 10x Genomics platform (a). For each timepoint, one library was generated using pooled tissues dissected from three individual animals to control for differences among individual animals, surgery and tissue dissociation variations. UMAP plot showing the different major clusters, ‘heartsClu2’, in two dimensions (b). To validate the clustering, known lymphatic-associated and macrophage-associated genes were examined using the integrated scRNA-seq dataset (c). Unbiased Gene Ontology analysis identified pathways upregulated in macrophages after injury at P7 compared to P1 (d) and genes potentially driving these pathways (e). Significant differences between Gene Ontology term enrichment were calculated using Fisher’s exact test. f, Heatmap of changes in lymphatic endothelial cell gene expression across conditions. Panel a created with BioRender.com. Clusters: 0-EC1 (endothelial cells 1), 1-FB1 (fibroblasts 1), 2-Mac, 3-FB2 (fibroblasts 2), 4-EC2 (endothelial cells 2), 5-EC3 (endothelial cells 3), 6-FB3 (fibroblasts 3), 7-Granulocytes, 8-SMC (smooth muscle cells), 9-Pericytes, 10-FB4 (fibroblasts 4), 11-EC4 (endothelial cells 4), 12-TC (T cells), 13-FB5 (fibroblasts 5), 14-EC5 (endothelial cells 5), 15-EC6, 16-BC (B cells), 17-CM (cardiomyocytes), 18-unassigned, 19-Epi (epicardium), 20-FB6 (fibroblasts 6), 21-Glial (glial cells), 22-unassigned, 23-Mo (monocytes).

Fig. 7. Functional MRI parameters of P1 versus P7 Lyve1 KO and macrophage-specific Lyve1 KO hearts 28 days after MI.

Plots from longitudinal cine MRI performed on Lyve1^+/− and Lyve1^−/− mice 28 days after MI at P1 and P7 as well as in intact control littermates at comparable P29 and P35 stages (a,b). MRI revealed significantly reduced cardiac output in P1 Lyve1^−/− mice at 28 days after MI compared to P29 intact Lyve1^−/− controls (a). The reduced cardiac output was even more significant than reductions observed for Lyve1^−/− mice injured at P7 at 28 days after MI compared to P35 intact Lyve1^−/− controls (b). Plots from MRI performed on hCD68–CreERT2;Lyve1^flox/flox mice 28 days after MI at P2 reveal impaired functional recovery across cardiac output, StV, ejection fraction and end diastolic volume consistent with a, differing only in end systolic volume (c). Data are presented as mean ± s.d. n = 6 for P29 control, n = 7 for P1MI28dpi control, n = 8 for P29 Lyve1^−/−, n = 7 for P1MI28dpi Lyve1^−/−, n = 8 for P35 control, n = 7 for P7MI28dpi control, n = 8 for P35 Lyve1^−/−, n = 10 for P7MI28dpi Lyve1^−/−, n = 7 for P2MI28dpi, n = 13 for P2MI28dpi. Significant differences were calculated using two-way ANOVA for a and b and unpaired two-tailed Student’s t-test for c. Source data

Fig. 8. Impaired vascular response in hCD68–CreERT2;Lyve1^flox/flox infarcted hearts and a population shift toward inflammatory monocytes in Lyve1 KO CD45⁺ cells at day 7 after MI.

Identification of scar area at 7 dpi using wheat germ agglutinin (WGA) and Picrosirius red fibrosis stain. Visualization of vasculature and macrophages with CD31 (PECAM1) and IBA1, respectively. Representative images of littermate control (a) and hCD68–CreERT2;Lyve1^flox/flox (b) sections illustrating reduced neovascular response in Cre⁺ sections. Quantification of scar area revealed no significant difference between conditions (c). Quantification of discrete PECAM1 signal within the infarct zone relative to area as an indicator of vascular response revealed significantly reduced PECAM1-stained vasculature in Cre⁺ hearts (d). Quantification of macrophages by IBA1 stain revealed no significant difference between conditions (e). scRNA-seq was conducted, comparing CD45⁺ enriched cells from neonatal Lyve1 KO versus WT hearts at P2MI7dpi. The samples were FACS sorted using 7-AAD and CD45 to isolate live CD45⁺ cells, and libraries were prepared for sequencing using the 10x Genomics platform. UMAP plot of grouped WT and Lyve1 KO CD45⁺ cells (f). Comparison of CD45⁺ cell subset proportions between WT and Lyve1 KO conditions (g), including statistical analyses of quantified differences between subsets after deconvolution of individual heart samples by Vireo5 (ref. ) (h). Macrophage subset clustering in WT (i) and Lyve1 KO (j). Lyve1 gene expression within subclusters in WT (k) and Lyve1 KO (l). Dot plot illustrating relative expression of key pro-angiogenic, pro-inflammatory and pro-fibrotic genes in each macrophage subset and between WT and Lyve1 KO (m). Cumulative apoptotic marker expression scores between conditions and macrophage subsets (n). Representative LYVE1⁺ macrophage possessing an HA glycocalyx (o). Differential HA glycocalyx staining between WT and Lyve1^−/− macrophages (p). n = 4 control hearts, n = 3 hCD68–CreERT2;Lyve1^flox/flox hearts; four sections per heart for a–e, pooled samples from n = 5; five hearts for f–m. The box center in violin plots in n indicates the median; the lower and upper hinges correspond to the first and third quartiles; and the whiskers extend to values with a distance from the hinges that is at most the interquartile range multiplied by 1.5. Box plot parameters, including cell counts, are available in the Source Data. Unpaired Student’s t-tests were used to determine significance in c–e. Bonferroni-corrected pairwise Wilcoxon rank-sum test was used to determine significance in n. Qualitative observations in o and p were repeated across the scar and in a second infarcted heart. Scale bars, 200 μm for a, b and p; 50 μm for o. FDR, false discovery rate; HA, hyaluronic acid; macro, macrophage; NS, not significant; T_reg, regulatory T cell. Source data

Extended Data Fig. 1. Heart growth during postnatal development.

Hearts were harvested to quantify the length and width size across the ventral surface at different postnatal stages (A). The heart was found to grow from base to apex in proportion to increasing width and area (B-D). Four weeks after birth, the area of the heart had increased four-fold on the ventral side. Data are presented as mean ± SEM. n = 5 for P1; n = 3 for P3, P5, P9; n = 4 for P8, P11, P16, P21, P28; n = 10 for P14. Significant differences were calculated using 1-way ANOVA followed by Tukey’s multiple comparisons test. Scale bar: 1 mm. Source data

Extended Data Fig. 2. The cardiac lymphatic network undergoes expansion until late postnatal development.

Whole-mount imaging of Vegfr3^+/LacZ hearts showed extensive growth and sprouting of the cardiac lymphatic vasculature on both the dorsal and ventral surfaces of the heart during postnatal stages (A). Isolated LECs were present on the dorsal side of the heart during early postnatal stages (grey arrows at P1, P3) and by P11 and P14 sprouts were detected arising from a large dorsal lymphatic vessel (white arrows). These sprouts produced a dense plexus of vessels near the apex at P16 (black arrow at P16 - P28). Areas depleted of lymphatics were present in stereotypical locations of the heart, such as the apex on the ventral side (black arrowheads at P28) and the right ventricle on the dorsal side (black arrowheads at P28). Whole mount hearts were cleared using iDISCO at P1, P7, and P14 and stained for VEGFR3 before imaging with Lightsheet microscopy (B-J). Initial lymphatics were identified morphologically (C, C’, F, F’, I, I’). Representative sections (D, G, J) from the 3-dimensional images (B, E, H) illustrate lymphatics are largely absent at depth. n = 3-5 for each time point in A. Scale bars: 0.5 mm for A, B, E, H; 200μm for C, F, I; 50μm for magnified views.

Extended Data Fig. 3. Imaging-based assessment and quantification of the post-natal cardiac lymphatic plexus.

Light-sheet and wholemount confocal imaging of Prox1-tdTomato (A-C) and Vegfr3^+/lacZ (D) intact hearts, at post-natal days 7 and 14 respectively, enable visualisation of the cardiac lymphatic plexus which can be segmented using Image J (E) and quantified in terms of parameters such as vessel length, calibre, junctions, branch points etc using AngioTool^TM (F). The total vessel length and number of end points remained stable until P14 and then significantly increased on both sides of the heart via lymphangiogenesis (G). By P16 the lymphatics were fully formed on the dorsal side, while the ventral lymphatics were delayed, reaching full expansion by P21 (H). Data are presented as mean ± SEM. n = 5 for P1, P16, P21; n = 3 for P3, P5, P9; n = 4 for P8, P11, P28. Significant differences were calculated using 2-way ANOVA. Scale bars 0.5 mm, A and 1 mm, D. Source data

Extended Data Fig. 4. Expression of lymphatic markers across the post-natal period.

Whole-heart samples across post-natal stages P0-P28 were analysed by qPCR for lymphatic markers. Expression levels of genes implicated in lymphangiogenesis, such as Vegfr3 (A), Vegfc (B), Vegfd (C), Prox1 (D) and Nrp2 (E) revealed biphasic increases, with the first between P2 and P3 and a second around P9. Pdpn increased only between P9 and P11 (F). Genes implicated in lymphatic function, such as Ccl21 (G) and Lyve1 (H) displayed an expression pattern where levels fluctuated during the first week, before becoming increasing significantly at P21 and thereafter remained stable. Data are presented as mean ± SEM; n = 3 for each time point. Significant differences were calculated using 1-way ANOVA followed by Tukey’s multiple comparisons test. Source data

Extended Data Fig. 5. Successful engraftment of CD68 + GFP+ monocytes in the heart after adoptive cell transfer.

GFP and CD68 immunostaining of cardiac tissue sections revealed appropriate engraftment of CD68⁺GFP⁺ monocytes within the injury area of the heart 7 days after adoptive transfer concurrent with MI at P1 and P7 (white arrows) (A-D). Two representative hearts for each condition shown. n = 4 hearts analysed across each stage. Scale bar: 50 μm.

Extended Data Fig. 6. Proliferation of macrophages in the MLNs is restricted to the subcapsular sinus in both injured and intact post-natal mice.

Visualisation of endogenous hCD68-eGFP and phospho-histone H3 (PH3) in MLNs from control (intact) P14 MLNs compared to MLNs from injured P7 mice at day 7 post-MI, revealed similar macrophage proliferation levels and distribution in both conditions (compare A to D). Macrophages were mostly negative for PH3 in the medullary sinus (B and E). Increased levels of PH3+ positive mitotic macrophages were found in the subcapsular sinus (C-F) in MLNs from both control and injured mice. Visualisation of endogenous CX3CR1-eGFP and PH3 in MLNs from uninjured mice at P8 and P14 compared to those from injured P1 and P7 mice 7 days post-MI (D-N). CD68+ and CX3CR1+ macrophages were similarly distributed and there was no significant difference in PH3+ macrophages between conditions (O). Data are presented as mean ± SEM; n = 7 for P8, n = 8 for P1MI7dpi, n = 9 for P14, 10 for P7MI7dpi. Significant differences were calculated using 1-way ANOVA followed by Tukey’s multiple comparisons test. Scale bars: 50 μm. Source data

Extended Data Fig. 7. P1 versus P7 macrophage and monocyte populations have distinct proliferation profiles and differentially interact with lymphatic endothelial cells.

Pooled UMAP plot showing the different major clusters (A). Stacked violin plots showing expression of marker genes for each cluster (B). UMAP of the macrophage and monocyte clusters separated based on timepoint and condition (P1 intact, P1MI1dpi, P1MI7dpi, P7 intact, P7MI1dpi and P7MI7dpi) (C). The percentage of cell cycle genes for each macrophage and monocyte cluster confirmed the existence of a proliferating population, with approximately half of the Mf2 cluster at phase S of cell division (D). The percentage of mitochondrial RNA (E) and the number of molecules (F) detected within cells confirmed the existence of an apoptotic population. Stacked violin plots showing expression of marker genes for each cluster suggests at least two tissue-resident macrophage populations, designated macrophage 1 (Mf1; Lyve1⁺;Ccr2⁻;Arg1⁻) and macrophage 2 (Mf2; Lyve1⁻;Ccr2⁺;Arg1⁺) (G). The relative percentage of each cell population at different timepoints and conditions as represented in a proportional bar chart (H). The unbiased differential expression analysis identified genes that were enriched in each cluster by comparing across all the other clusters (I). Comparison of total incoming path weights and total outgoing path weights across populations (J). Hierarchical network diagram of significant cell-cell interaction pathways, with arrows and edge colour indicating signalling direction ligand:receptor (K). Summed ligand weights across ligand and receptor target paths for top ligands in LECs (L) and macrophages (Mf) (M). Subclustering of endothelial cells to identify LECs (N). Unbiased GO term analysis showed pathways enriched at P7MI7dpi, vs P1MI7dpi but not between P1 and P7 (O). Genes implicated in enriched pathways (P). Relative abundance of macrophage subsets grouped by individual hearts following de-multiplexing of pooled samples (Q). Cumulative expression scores of cell cycle markers showed a decrease in Lyve1^−/− CCR2- macrophage proliferation, but no change in CCR2+ or monocytes (R). In vivo validation of increased apoptosis; CC3 co-expression with F4/80 within the infarct zone 7 days following MI at P2 was significantly increased in the Lyve1^−/− context (S). Violin plot indicating expression levels of CD44 were significantly reduced in the CCR2- macrophage cluster, but using vireo5 were unchanged following loss of Lyve1 (T). CM = cardiomyocytes, Art EC = arterial endothelial cells, VEC = venous endothelial cells, Endo = endocardium, Prol VEC = proliferating VEC, LEC = lymphatic endothelial cells, Fb = fibroblasts, Prol Fb = proliferating Fb, Epic = epicardium, SMC = smooth muscle cells, Peri = pericytes, Mf = macrophages, DC-like = dendritic cell-like, Gran = granulocytes. Data are presented as mean ± SD. Significant differences between Enriched Pathways in O and P were calculated using Fisher’s exact test. The box centre in violin plots in R and T indicates the median, the lower and upper hinges correspond to the first and third quartiles, and the whiskers extend to values with a distance from the hinges that is at most the inter-quartile range (IQR) multiplied by 1.5. Box plot parameters, including cell counts are available in Source Data. Bonferroni-corrected pairwise Wilcoxon Rank Sum test was used to determine significance in R and T. Unpaired Student’s t tests were used to determine significance in S. n = 4, 4 for S. Scale bar 200μm. Source data

Extended Data Fig. 8. Network analysis showing LEC-macrophage signalling.

Tree plot showing outgoing connections from the LECs to Mf (A) and from Mf to LECs (B). The single outgoing LEC signal was identified as Reelin (Reln), a previously described lymphangiocrine factor, interacting with ITGβ▢ on Mf (A). Top node refers to source population, second layer to ligands, third layer to receptors and bottom node representing the target population (A, B). Reln expression was shown by qPCR to decrease between P1 (n = 6) and P7 (n = 5), while Itgb1 expression was unchanged (C). Immunostaining for ITGFB1 combined with IBA1 (D) and REELIN combined with LYVE-1 (E-H) in resident macrophages and LECs, respectively, revealed elevated expression of Reelin at P2MI5dpi compared to P7MI5dpi and sham controls across equivalent timepoints (Figure E-H). Data are presented as mean ± SD. Qualitative observations in D-H were seen across tissue sections and in 1 additional heart. Scale bar: 25 μm for D; 150 μm for E-H. Source data

Extended Data Fig. 9. Additional functional MRI parameters of Lyve1 KO hearts 28-days after MI at P1 and P7.

Plots from longitudinal cine MRI performed on Lyve1^+/− and Lyve1^−/− mice 28 days after MI at P1 and P7, as well as in intact control littermates at comparable P29 and P35 stages (A,B). MRI revealed significantly reduced end diastolic mass (EDM), end systolic mass (ESM), and heart weight (hwt) in P1 Lyve1^−/− mice at 28 days post-MI compared to Lyve^+/− MI controls at the equivalent timepoint (A). Conversely, EDM, ESM, and hwt were all significantly increased in Lyve1^−/− mice injured at P7 at 28-days post-MI compared to Lyve1^+/− 28dpi controls (B). Data are presented as mean ± SD; n = 6 for P29 control, n = 7 for P1MI28dpi control, n = 8 for P29 Lyve1^−/− n = 7 for P1MI28dpi Lyve1^−/−. n = 8 for P35 control, n = 7 for P7MI28dpi control, n = 8 for P35 Lyve1^−/− n = 10 for P7MI28dpi Lyve1^−/−. Significant differences were calculated using 2-way ANOVA. Source data

Extended Data Fig. 10. hCD68-CreERT2;R26R-TdTomato reports LYVE-1 positive macrophages in P2 hearts 7 days post-MI and MRI cine images of hCD68-CreERT2;Lyve1^fl°^x/fl°^x hearts 28-days post-MI with corresponding picrosirius red staining.

Macrophages expressing TdTomato in P2 infarcted hearts at 7dpi. LYVE-1 + /TdTomato+ macrophages confirmed by co-staining with CD68 and LYVE-1 (A). Quantification of co-staining in sections (n = 3) (B). Tamoxifen injected at P1, MI surgery at P2 to avoid inducing respiratory failure. Longitudinal and transverse cine MRI representative images from control and hCD68-CreERT2+ animals at P2MI28dpi accompanied by transverse picrosirius sections matched to the corresponding level and illustrated by schematic (C, D). Picrosirius red staining revealed no significant difference in relative scar area between control and hCD68-CreERT2+ animals at 28 days post-MI (E). Asterisk marks suture site. RA = right atrium, LA = left atrium, RV = right ventricle, LV = left ventricle. n = 7 for control, n = 13 for hCD68Cre^ERT2;Lyve1^fl/fl hearts. Panel C created in BioRender; Chapman, B. (2025) https://BioRender.com/z1nstjj. Unpaired Student’s t test was used to determine significance. Data are presented as mean ± SD. In A, scale bar 350μm; zoomed 150μm. In D, scale bar 5 mm in MRI; 1 mm in picrosirius sections. Source data