Cardiac Cell Therapy Fails to Rejuvenate the Chronically Scarred Rodent Heart

Therapeutic use of adult cells with stem or progenitor-like properties for cardiovascular disease remains controversial. Despite early pre-clinical claims of robust structural and functional recovery after myocardial infarction (MI), cell therapy trials for acute MI patients have been largely disappointing¹. Focus has shifted to other indications such as chronic ischemic cardiomyopathy and heart failure, although again, efficacy remains controversial^{2, 3}. Moreover, the mechanisms whereby cell therapy might impart benefits to a chronically infarcted heart with a stable scar remain unclear.

We recently reported that intramyocardial injection of bone marrow mononuclear cells (MNCs), cardiac-derived mesenchymal cells, non-viable cell debris, or non-cellular inflammatory molecules all equivalently improved function of the acutely infarcted rodent heart via regional macrophage induction and border zone remodeling⁴. Since this rejuvenating response occurred during early infarct maturation (one week after ischemia/reperfusion [I/R]), we also asked whether a similar effect would occur in a chronic, stable scar long after resolution of acute wound healing. This is especially relevant because most clinical trials apply cell therapy many months or years after an MI. Thus, here we attempted to nullify the hypothesis that cell therapy can revitalize hearts of mice with chronic scars (post-MI).

Male and female C57BL/6J (WT) mice were subjected to I/R injury via 120 minutes of coronary artery ligation. Eight weeks after I/R, animals were randomized to receive border zone intramyocardial (i.m.) injection of either 1.5×10⁵ mTomato-labeled MNCs or sterile saline (Figure, A). All procedures were conducted as previously described⁴. We first assessed retention of MNCs and resulting immune responses by histological analysis 3 days after i.m. injection. Small numbers of mTomato⁺ MNCs were observed in the infarct border zone, associated with greater CD68⁺ activated macrophage content versus saline controls (Figure B, C), indicating that MNCs induced acute inflammation around the stable scar. We next analyzed these mice after an additional 2 weeks by echocardiography and with histological methods. Here we also re-analyzed and compared a previously published data set from our lab where mice received i.m. injection of 1.5×10⁵ mTomato-labeled MNCs or sterile saline, but at 1 week post-I/R⁴. These two cohorts of mice were initiated concurrently with the same surgeon and experimental techniques. Mice that received cell therapy at 8 weeks post-I/R showed no improvement in contractile function as measured by fractional shortening (%FS; Figure, D) or left ventricular (LV) remodeling as measured by end-systolic volume (LVESV; Figure, E). In contrast, i.m. injection at 1 week post-I/R attenuated LV dysfunction and improved contractile performance (Figure, D, E and ref⁴). Heart rates were equivalent across all groups (Figure, F). We also assessed cardiac hypertrophy by gravimetric analysis (Figure, G) and border zone fibrotic content (Figure, H), and both were unchanged between MNC and saline treatments when delivered at 8 weeks post-I/R.

Figure: Intra-cardiac injection of Bone Marrow Mononuclear Cells Does Not Improve Cardiac Function in Mice with Chronic Ischemic Injury.

A. Summary diagram for studies using C57BL/6J mice that received ischemia/reperfusion (I/R) injury via 120 min of left coronary artery reversible ligation. Eight weeks after I/R injury, mice received direct intramyocardial (i.m.) injection of 150,000 bone marrow mononuclear cells (MNCs) divided over two sites flanking the infarct border zone as previously described⁴. MNCs were isolated from Rosa26-mTomato mice on the C57BL/6J background. Male and female mice were used in all experiments and for MNC isolations. Diagram created with Biorender.com. B, C. Representative confocal immunofluorescence histological micrographs from hearts subjected to I/R injury followed by sterile saline (B) or mTomato⁺ (red) MNC (C) i. m. injection at 8 weeks post-I/R, then sacrificed 3 days after injection for analysis. Anti-CD68 antibody showed activated macrophages (white) and DAPI showed nuclei (blue). Scale bars in B, C = 100 μm. D–F. Experimental groups as described in A were assessed by 2D M-mode echocardiography, 2 weeks after cell therapy (10 weeks post-I/R). All mice received I/R injury except for Sham. A pre-therapy echo (pre-i.m.) was also performed on these mice at 8 weeks post-I/R, one day prior to cell or saline injection. For comparison, previously published data⁴ from mice that received either sham or cell therapy at 1 week post-I/R were re-analyzed and shown again here (denoted with hashed lines through box and whisker plots) for comparison. These studies were performed concurrently using the same methodological approaches. Left ventricular fractional shortening (%FS; D) and end-systolic volumes (LVESV; E) were similar between mice that received i.m. saline or MNCs at 8 weeks post-I/R. In contrast, mice that received i.m. MNCs at 1 week post-I/R showed significantly higher %FS and reduced LVESV after an additional 2 weeks. *p<0.05 versus Sham; #p<0.05 versus Sal. (1 week I/R) by one-way ANOVA with Tukey’s multiple comparisons test. F. Heart rates during echocardiographic assessment were not different across all groups. G. Cardiac hypertrophy measured by heart weight/body weight (HW/BW) was not altered in mice receiving i.m. MNCs at 8 weeks post-I/R. H. Fibrosis at the infarct border zone was quantified from histological sections stained with Masson’s trichrome as the percentage of fibrotic area over total tissue area analyzed. Mice receiving i.m. MNCs at 8 weeks of I/R showed no difference in local tissue fibrosis at the infarct border zone, versus saline-injected controls. All data in D–H are shown as box-and-whisker plots with the median value (black bar inside box), 25th and 75th percentiles (bottom and top of box, respectively), and minimum and maximum values (bottom and top whisker, respectively) indicated. The number (n) of animals is shown below the plot. I. Representative confocal immunofluorescence micrographs from hearts of mice subjected to I/R injury followed by sterile saline or mTomato⁺ MNC i. m. injection at either 1 week post-I/R (I) or 8 weeks post-I/R, then sacrificed 2 weeks for later for COMP/Thbs5 antibody staining (green). Scale bars = 100 μm. J. Quantitation of COMP immunostaining from groups shown in I. Confocal micrographs were analyzed in a blinded and automated fashion using N.I.S. Elements software from n=4 (Sal. 1w post-I/R), n=6 (MNC 1w post-I/R), n=9 (Sal. 8w post-I/R) or n=11 (MNC 8w post-I/R) mice. Data are shown as box-and-whisker plots with the median value (black bar inside box), 25th and 75th percentiles (bottom and top of box, respectively), and minimum and maximum values (bottom and top whisker, respectively) indicated. Groups treated at 1 week post-I/R are denoted with dashed line boxes and groups treated at 8 weeks post-I/R are denoted with solid boxes as in all other panels. *p<0.05 versus Sal. 1w post-I/R by Kruskal-Wallis test with Dunn’s multiple analysis test.

We also analyzed deposition of cartilage oligomeric matrix protein (COMP, also known as thrombospondin-5) at the infarct border zone (Figure, I). COMP is a matricellular protein highly expressed in chronic infarct scars, along with persistent accumulation of stably differentiated cardiac fibroblasts (termed matrifibrocytes)⁵. Mice that received i.m. MNC or saline injection at 1 week post-I/R showed only modest COMP deposition by 3 weeks. In contrast, when I/R injury was allowed to stabilize for 8 weeks prior to i.m. injection, both MNC and saline-injected mice showed significantly greater accumulation of COMP (Figure, J), with an elaborated array of dense fibers reminiscent of a mature infarct scar⁵. These observations suggest a distinct scar microenvironment in early (1–3 week post-I/R) versus late-stage (8–10 week post-I/R) infarcts, which could account for the inability of MNCs to alter fibrotic characteristics and attenuate dysfunction with late administration. Alternatively, the tissue environment of the chronic scar could impart differential effects on recruited or resident immune cells such that inflammation-based infarct healing was not achieved.

Overall, our data demonstrate a “point of no return” for rejuvenating the infarcted rodent heart via intra-cardiac injection of MNCs. Indeed, MNCs delivered into the chronic scar did not improve function despite inducing localized inflammation. This suggests cell therapy must be applied within the short time window of acute healing after I/R injury, where macrophage subsets can differentially modulate fibroblasts and extracellular matrix properties to alter functional characteristics of the early forming scar⁴. These protective effects may not be possible once the infarct scar has reached a mature state with matrifibrocyte accumulation. As our study utilized only MNCs, future studies are warranted using other cell-based modalities. Overall, a more complete mechanistic picture should inform current and future use of cardiac cell therapy for chronic ischemic cardiomyopathies or heart failure, where patients will most often present with stable, long-lived scars.

Requests of materials, datasets, and protocols used in this study should be directed to the corresponding author and will be made available to investigators upon reasonable request. No human subjects or materials were used in this study. All animal procedures were conducted in accordance with institutional guidelines and were approved by the Institutional Animal Care and Use Committee of Cincinnati Children’s Hospital.