Identifying the molecular and cellular signature of cardiac dilation following myocardial infarction

Abstract

Establishing molecular and cellular indicators that reflect the extent of dilation of the left ventricle (LV) after myocardial infarction (MI) may improve diagnostic and prognostic capabilities. We queried the Mouse Heart Attack Research Tool (mHART) 1.0 for day 7 post-MI mice (age 3-9 months, untreated males and females) with serial echocardiographic data at days 0, 1, and 7 (n = 51). Mice were classified into two subgroups determined by a median fold change of 1.6 in end-diastolic dimensions (EDD) normalized to pre-MI values; n = 26 fell below (moderate; mean of 1.42 ± 0.01) and n = 25 fell above this cut-off (extreme; mean of 1.79 ± 0.01; p < 0.001 vs. moderate). Plasma proteomic profiling of 34 analytes measured at day 7 post-MI from male mice (n = 12 moderate and 12 extreme) were evaluated as the test dataset, and receiver operating curve (ROC) analysis was used to assess strength of biomarkers. Females (n = 6 moderate and 9 extreme) were used as the validation dataset. Both by t-test and characteristic (ROC) curve analysis, lower macrophage inflammatory protein-1 gamma (MIP-1γ), lymphotactin, and granulocyte chemotactic protein-2 (GCP-2) were identified as plasma indicators for dilation status (p < 0.05 for all). Macrophage numbers were decreased and complement C5, laminin 1, and Ccr8 gene levels were significantly higher in the LV infarcts of the extreme dilation group (p < 0.05 for all). A composite panel including plasma MIP-1γ, lymphotactin, and GCP-2, and LV infarct Ccr8 and macrophage numbers strongly mirrored LV dilation status (AUC = 0.92; p < 0.0001). Using the mHART 1.0 database, we determined that a failure to mount sufficient macrophage-mediated inflammation was indicative of exacerbated LV dilation.

Keywords: Big data; Cardiovascular disease; Inflammation; LV dilation; Macrophage; Myocardial infarction; Proteomics.

Figure 1.. Two distinct dilation response groups were identified.

(A) Using mHART a median change in end diastolic dimension (EDD) of 1.6 fold, mice were classified into moderate and extreme groups. (B) Partial least squares discriminant analysis (PLS-DA) of 34 plasma analytes from mHART revealed separate clustering of the two groups. (C) By pathway analysis, chemokine signaling was downregulated (blue) and toll-like receptor, cell adhesion, and complement and coagulation were upregulated (orange) in the extreme group compared to the moderate group. No differences was observed in cytokine-cytokine receptor interaction (gray). Z-score was calculated based on the mean and standard deviation of extreme compared to moderate.

Figure 2.. Analysis of plasma proteins identified potential indicators that correlate with dilation outcome.

(A) Heat map of the top 15 differentially expressed plasma analytes (mHART plasma proteomic dataset) showed strong clustering of groups based on protein profiles; 3 were significantly downregulated in the extreme group compared to the moderate group (◄). (B) Macrophage inflammatory protein (MIP)1γ, (C) lymphotactin, and (D) granulocyte chemotactic protein (GCP)2 were significantly lower in the extreme dilation group. n=12 male mice/group.

Figure 3.. MIP1γ, GCP-2, and lymphotactin were strong indicators of moderate LV dilation in males.

By ROC analysis, (A) high circulating MIP-1γ at day 7 post-MI was a strong indicator of moderate LV dilation. (B) Granulocyte chemotactic protein (GCP)2, and (C) lymphotactin were also strong indicators of moderate LV dilation. (D) Myeloperoxidase (MPO) was not a strong indicator of day 7 post-MI LV dilation. Plasma proteomic data collected from males within mHART was used for analysis. n=12 male mice/group.

Figure 4.. Validation using a separate MI cohort of females also identifies MIP1γ and GCP-2 as indicators of moderate LV dilation.

(A) Using a separate cohort of female day 7 post-MI mice (n=6 moderate; n=9 extreme), MIP-1γ was significantly reduced in the extreme dilation group. (B) MIP-1γ negatively correlated with end diastolic dimension (EDD) and ROC analysis indicated (C) MIP1γ and (D) GCP-2 were strong indicator of LV dilation. Plasma proteomic data collected from females within mHART was used for analysis.

Figure 5.. Extreme dilator group had fewer macrophages without differences in neutrophil numbers.

(A) No difference in neutrophil numbers at day 7 post-MI between groups (arrows depict positive staining). (B) ROC analysis indicated neutrophil numbers did not reflect LV dilation status. (C) Macrophage numbers were significantly lower in the extreme group compared to the moderate group (arrows depict positive staining) and (D) ROC analysis indicated macrophage numbers reflected LV dilation state. Immuonistochemistry data collected from mHART was used for analysis. Scale bar, 60 μM.

Figure 6.. Elevated Ccr8 gene expression in the day 7 infarct predicts extreme dilation.

(A) Assessment of 184 genes collected from the mHART database indicated C5, lamanin a1 (Lama1), and Ccr8 were significantly higher in the day 7 infarct of extreme dilators compared to the moderate group. MIP-1γ gene expression in the infarct was not significantly different between groups indicating plasma differences were post-transcriptional. (B) Ccr8 negatively correlated with macrophages in the infarct area and (C) positively correlated with EDD. (D) ROC analysis substantiated Ccr8 as an indicator of LV dilation.

Figure 7.. Macrophages express increased levels of macrophage inflammatory protein (MIP)1-γ early post-myocardial infarction and correlate to spherical index.

(A) Sequencing of macrophages isolated from the infarct at post-MI days 0, 1, 3, and 7 showed MIP-1γ expression peaked at post-MI day 1 and decreased by day 7 to levels lower than day 0 macrophages. (B) Macrophage numbers negatively correlated with spherical index. (C) Macrophage-specific expression of MIP-1γ had a non-linear relationship with macrophage numbers indicating too much or too little MIP-1γ may be a negative regulator of macrophage recruitment. Dashed lines represent confidence interval. Data was generated from a separate cohort of mice.[21] n=81 mice total with n=4 pooled biological replicates/day post-MI. *p<0.05 vs day 0. FPKM= fragments per kilobase million.

Figure 8.. Multi-marker strategy with key independent biomarkers improves indication of LV dilation status.

(A) Single marker ROC analysis of MIP-1γ, lymphotactin, GCP-2, Ccr8, and macrophage numbers using males and females (n=18 moderate; n=21 extreme). (B) Using a five panel test improved determination of LV dilation state. All data was collected from the mHART database.