Integrative Analysis Prioritizes CRIP2 as a Candidate Associated with Myocardial Copper-Handling Responses After Myocardial Infarction

Figures

• 

  Figure 1. Multi-cohort transcriptomic integration and proteomic Mendelian randomization nominally prioritize CRIP2 as a candidate downregulated in myocardial infarction. (A) Intersection of differentially expressed genes across infarcted myocardium (IM), peri-infarct zone (PZ), and non-infarcted myocardium (NI) in GSE775, yielding 174 shared genes. (B) Intersection of the 174 GSE775 genes with the in-house mouse MI RNA-sequencing cohort, yielding 82 shared genes. (C) Two-sample MR estimates for circulating candidate protein levels and MI risk. IVW estimates were CRIP2: OR 0.831 (95% CI 0.735–0.939), p = 0.00306, 12 variants; CCDC80: OR 0.798 (95% CI 0.677–0.940), p = 0.00702, 19 variants; and GLIPR2: OR 0.916 (95% CI 0.840–0.999), p = 0.0461, 20 variants. These associations were nominal (p < 0.05), and none met the Bonferroni threshold (p < 0.00156). (D) Quantitative real-time PCR validation in the local mouse LAD-ligation MI model at four weeks after MI. Data are mean ± SEM (n = 5 mice per group for each gene; exact p values are displayed; two-tailed unpaired Student’s t-test). Panel colors and symbols are defined as follows: colors in (A,B) correspond to the labeled comparisons or…

• 

  Figure 2. Single-cell transcriptomic profiling maps Crip2 expression across cardiac cell populations after myocardial infarction. (A) Dot plot of canonical marker genes used to annotate five major cardiac cell populations: cardiomyocytes (CM), fibroblasts (FB), endothelial cells (EC), macrophages (Mac), and smooth muscle cells/pericytes (SMC/Pericyte). Dot size and color indicate the percentage of expressing cells and average expression level, respectively. (B) UMAP visualization of cardiac single-cell transcriptomes split by condition (MI vs. Sham). (C) Feature plots map Crip2 expression density and show a descriptively lower signal after MI. (D) Dot plot comparing Crip2 expression across cell populations and conditions. Because two nominal MI raw matrices were byte-identical, the unique-heart pseudobulk sensitivity analysis included three sham and two MI hearts, and did not detect a significant condition-level Crip2 difference (exact permutation p = 0.40); the atlas is therefore interpreted as cell-type localization evidence, rather than replicated proof of post-MI downregulation. In (B), colors denote annotated cell types; in (C), red and gray indicate higher and lower Crip2 expression,…

• 

  Figure 3. Exploratory in silico modeling of CRIP2 perturbation and downstream copper-handling network responses after myocardial infarction. (A) Geneformer embedding shifts (cosine distance) among copper-handling and mitochondrial respiratory genes after virtual CRIP2 perturbation in MI and sham contexts. (B) Comparison of virtual CRIP2 knockout (KO) and overexpression (OE) effects; KO effects were often, but not invariably, larger than OE effects. (C) scTenifoldKnk topology Z-scores after virtual CRIP2 knockout in MI and sham contexts; no adjusted-significant topology effects were identified. (D) Exploratory comparison of Geneformer KO_MI embedding shifts with scTenifoldKnk Z_MI scores across 13 genes (Spearman ρ = −0.313, 95% bootstrap CI −0.751 to 0.262, p = 0.297; 100,000 resamples). The non-significant association does not support cross-method concordance. Cox17 was not included because it did not pass the Geneformer top-2048-gene input threshold. Blue and red denote sham and MI in (A,C), whereas blue and red denote overexpression and knockout in (B). Red points in (D) represent individual genes, and the dashed line is the fitted trend. FDR, false discovery rate.

• 

  Figure 4. CRIP2 perturbation alters selected copper-handling transcripts in H9c2 cardiomyoblasts. (A) Bubble heatmap of GTEx co-expression correlations between CRIP2 and 14 copper-handling genes across human cardiovascular tissues; circle size represents |r|. * p < 0.05, ** p < 0.01, *** p < 0.001. (B) Western blots confirming CRIP2 knockdown and overexpression; GAPDH was the loading control. (C,D) Quantitative PCR analysis after Crip2 knockdown (C) and overexpression (D). Knockdown reduced Slc31a1 and increased Slc25a3, whereas reduced Slc25a3 after overexpression was observed in the Crip2-OE#2 condition. Data are mean ± SD from three independent biological experiments (n = 3; one-way ANOVA with Dunnett’s post hoc test). These assays quantify transcript abundance, and do not directly measure copper flux or cuproptosis. In (A), red and blue indicate positive and negative correlations, respectively; color intensity reflects correlation magnitude, and |r| is the absolute Spearman correlation coefficient. In (C,D), blue, tan, and gray identify the control and two perturbation groups, and points denote independent biological experiments. Abbreviations: GTEx, Genotype-Tissue Expression; qPCR,…

• 

  Figure 5. CRIP2 perturbation is associated with relative metabolic activity during hypoxia/reoxygenation in H9c2 cardiomyoblasts. (A) Knockdown module: siNC and siCRIP2 under normoxia or hypoxia/reoxygenation (H/R). (B) Overexpression module: empty vector and CRIP2-OE under normoxia or H/R. Bars show mean ± SEM, and open symbols show three independent biological experiments (n = 3); three technical wells were averaged within each experiment. Values were normalized within each experiment to the corresponding module-specific normoxic control. Displayed pairwise p values are from two-sided paired t-tests on blank-corrected mean OD450 values. Interaction p values are from blocked two-factor linear models containing CRIP2 treatment, oxygen condition, their interaction, and experimental batch as a block; knockdown and overexpression modules were analyzed separately. CCK-8 measures tetrazolium-reducing metabolic activity/relative viability, and does not directly measure cell death or cuproptosis. Blue and orange denote the module-specific control and CRIP2 perturbation, respectively; light and dark shades denote normoxia and H/R. CCK-8, Cell Counting Kit-8.

• 

  Figure 6. CRIP2 perturbation is associated with context-dependent copper-handling and oxidative-stress transcriptional responses during hypoxia/reoxygenation in H9c2 cardiomyoblasts. (A–E) Knockdown module: siNC and siCRIP2 under normoxia (N) or hypoxia/reoxygenation (H/R). (F–J) Overexpression module: empty vector and CRIP2-OE under normoxia or H/R. Crip2, Slc31a1, Slc25a3, Atp7a, and Hmox1 were quantified relative to Gapdh using the 2−ΔΔCt method. Bars show mean ± SEM, and open symbols show three independent biological experiments (n = 3); three technical wells were averaged within each experiment. Interaction p values are from blocked two-factor linear models on ΔCt values containing CRIP2 treatment, oxygen condition, their interaction, and biological batch as a block. Interaction q values are Benjamini–Hochberg-adjusted across five genes within each module. p(H/R treatment) denotes the descriptive two-sided paired comparison between CRIP2 perturbation and the corresponding control under H/R, and is not multiplicity-adjusted. Blue bars denote module-specific controls, red bars denote siCRIP2, and green bars denote CRIP2-OE; light and dark shades denote normoxia and H/R, respectively.…