Intermittent pacing therapy favorably modulates infarct remodeling

In four chronically instrumented swine we assessed safety and proper functioning of the pacemakers by investigating the acute hemodynamic responses to intermittent dyssynchronous pacing of the left ventricle (IPTVVI) in comparison with intermittent synchronous atrial pacing (IPTAAI). See Supplemental Methods and Results for detailed protocol and results.

All data are expressed as mean ± SEM. Data were analyzed using two-way (time × treatment) RM-ANOVA followed by post hoc testing, when appropriate, or within treatment group differences with paired t testing, and between treatment group differences with unpaired t testing (SPSS 15.0, IBM, Armonk, NY, USA). One-way ANOVA testing with Fisher’s least significant difference correction (LSD) was used to analyze gene expression data. P ≤ 0.05 (two-tailed) was considered statistically significant.

Safety and proper functioning of ventricular pacing protocols were ascertained in four pigs as evidenced by the absence of ventricular arrhythmias and a programmed increase in heart rate of 30 beats/min (see Supplemental Methods and Results and Supplemental Figure 1 for details).

Table 1 summarizes global LV anatomy, function and infarct geometry data at baseline and at follow-up. In MI control animals, two hours of occlusion of the LCx resulted in AMI that comprised 31 ± 2% of the LV. Already at 3-day post-AMI, LV end-diastolic volume (EDV, 80 ± 3 ml) and end-systolic volume (ESV, 54 ± 3 ml) were larger as compared to EDV (68 ± 3 ml) and ESV (30 ± 2 ml) in a historical control group of healthy adolescent swine of similar age [30]. MI resulted in a depressed ejection fraction (EF) of 33 ± 2%, compared to 55 ± 3% in normal healthy swine [30]. During the 5 weeks follow-up, total LV mass did not change, as the increase in mass of the remote myocardium was balanced by an equivalent decrease in infarct mass (Table 1). EDV and ESV further increased to 113 ± 5 and 68 ± 3 ml, respectively, with EF tending to slightly recover from 33 ± 2 to 40 ± 2% (P = 0.06). The increase in EDV over time in swine with MI was in part attributable to physiological growth of the heart, as EDV in healthy swine showed a modest increase from 68 ± 3 to 91 ± 6 ml [30]. In contrast to the increase in ESV over time in swine with MI (from 54 ± 3 to 68 ± 3 ml), there was a negligible increase in ESV in healthy swine (from 30 ± 2 to 32 ± 3 ml), associated with an increase in EF from 55 ± 3 to 64 ± 3% [30]. These findings indicate that MI resulted in LV remodeling and dysfunction that was already observed early at 3 days and was sustained throughout the 5 weeks follow-up period.

Global LV and infarct geometry parameters showed no significant differences between IPTVVI and MI control (Table 1) at the 3-day post-AMI baseline. Moreover, the changes in global LV anatomy and function during 5 weeks of follow-up were not different between IPTVVI and MI control. However, infarct geometry responded markedly to IPTVVI (Table 1; Fig. 1). In MI control animals, infarct mass decreased over time as a result of a significant decrease in infarct thickness with no significant changes in infarct length or circumference, and no changes in the total number of infarcted segments. In contrast, in IPTVVI animals, the reduction in infarct mass was primarily due to a decrease in total number of infarcted segments (comprising reductions in both infarct length and circumference), with no significant decrease in infarct thickness. The latter could have played a role in preventing the reduction in infarct mass from reaching statistical significance compared to MI control group (P = 0.07; Fig. 1).

Several studies in swine [21] and rabbits [45, 46] have shown that pretreatment with ventricular, but not atrial [27], pacing is capable of limiting myocardial infarct size. Subsequent studies have demonstrated that not only preconditioning [21, 45, 46], but also postconditioning with brief periods of ventricular pacing in the early reperfusion phase [1‐3, 45] can limit myocardial infarct size. Moreover, the effects of this early protection against myocardial necrosis were sustained over a six week period, resulting in a trend towards blunted LV remodeling and improved LV function [2]. The present study is the first to investigate the effects of prolonged IPT on infarct remodeling, independent of its protection against acute myocardial necrosis in a preclinical animal model. The results of this study clearly demonstrate that IPTVVI started 3 days after reperfusion, at a time when necrosis can no longer be affected, significantly influenced remodeling of the infarct region.

Early studies in humans in the pre-thrombolysis era, reported disproportionate thinning and stretching of the infarcted segment [18, 48, 49]. Recent post-thrombolysis studies in humans with reperfused AMI confirm that regional myocardial wall thinning represents (transmural) myocardial infarction [35]. Furthermore, limited scar burden is associated with improved contractility [35] and blunted remodeling [28], whereas rupture-prone cardiac aneurysms are the consequence of continued ventricular wall thinning [22]. Also, late dilatation of the LV after primary percutaneous intervention remains of clinical significance [37] and may represent a potential target of IPTVVI. Thus, the IPTVVI-induced alteration of infarct geometry may mitigate the sequence of events leading to ventricular wall thinning and limit LV remodeling. Although significant changes in infarct geometry and composition were observed in the present study, these alterations did not yet translate into favorable global LV remodeling at 5 weeks follow-up. This finding, which is similar to previous observations with cell-therapy studies in swine [29], as well as humans [19], with AMI, suggests that more pronounced changes in infarct geometry are required to translate into benefits at the global LV level. This may also explain why a recent clinical study on peri-infarct zone pacing after AMI (PRomPT Trial) reported no difference in LV function or geometry at 18-month follow-up [39]. A limitation of the present study is that we investigated only a single algorithm of pacing therapy, so that we cannot exclude that other algorithms or pacing protocols may produce larger regional effects that do translate into global LV improvements [3]. Clearly, future studies, are necessary to optimize the onset, timing, duration and mode of pacing therapy.

The infarct zone is increasingly being appreciated as an area with relevant biological activity and therapeutic potential [8, 40]. Cardiac fibroblasts, including the active collagen-secreting myofibroblasts, are the dominant cell type in the infarct region and are recognized as essential in infarct remodeling [7, 20, 41, 43]. Myofibroblasts typically appear in the infarct area at 4–5 days after AMI, reach a peak at 1–2 weeks and continue to reside up to at least 4 weeks [10] and possibly months to years [43]. In the present study, IPTVVI, started at three days post-AMI, increased myofibroblast numbers in the infarct zone significantly. Myofibroblasts could have contributed to the geometric changes in the infarct region produced by IPTVVI in several ways. First, in the infarcted myocardium, myofibroblasts are responsible for collagen turnover thus contributing to the delicate balance between ECM synthesis and degradation [43]. Consistent with a role of myofibroblasts in ECM synthesis, Col1a1 expression was increased by IPTVVI. Moreover, a trend towards increased expression of TIMP-1 was found in infarcts of IPTVVI animals (P = 0.09) compared to MI control, which is consistent with the observations of Mukherjee et al. [32] that TIMP-1 co-localizes with myofibroblasts within the infarct zone. Although the increased Col1a1 expression and the trend towards an increase in TIMP-1 did not translate into an increase in collagen as measured with histology, it is plausible that myofibroblasts, a rich source of bioactive molecules [34], modulated infarct geometry by affecting ECM turnover. Second, myofibroblasts are capable of tonic contraction, and could therefore improve structural integrity of the scar and increase mechanical strength in the (sub)acute phase [43]. Particularly, the latter is in concordance with the geometrical changes observed in the present study.

The exact mechanism by which IPTVVI influenced myofibroblast presence was not determined in the present study. However, it is well known that LV pacing results in considerable changes in LV contraction patterns, even in the peri-infarct zone [25, 45], resulting in alterations in regional stretch and loading conditions [17, 33, 36, 46]. Since mechanical tension is an important stimulus for cardiac fibroblast to myofibroblast differentiation [11, 15, 16, 26], it is likely that regional alterations in myocardial stretch produced by IPTVVI stimulated resident fibroblasts to differentiate into myofibroblasts.

No change in circulating arterial plasma levels of the inflammatory marker TNFα was found in IPTVVI vs. MI control swine, at either 1 or 5 weeks after AMI, indicating that the effect of pacing on release of these proteins in pigs with infarcts was not discernible in the systemic circulation even as early as 1 week post-AMI. Hence, altered myofibroblast presence within the infarcted area of IPTVVI treated swine was likely the result of local effects on myofibroblast recruitment and/or differentiation. As myofibroblast migration and differentiation are initiated in the early phase after MI, expression studies at 5-week follow-up can by definition not identify which molecular mechanisms underlie the higher myofibroblast presence. Interestingly, the expression of the Fzd2 receptor, involved in myofibroblast homeostasis [24], was increased in both MI control and IPTVVI treated animals, which may have served to maintain the myofibroblasts in the infarcted region. It has previously been shown that activation of the TGFβ signaling pathway increases myofibroblast presence [23], and therefore it is possible that this pathway was only activated in the early phase of IPTVVI. This is also suggested by a study in swine, in which continuous low dose electrical stimulation within the infarct region, not only resulted in higher numbers of myofibroblasts, but also in elevated TGFβ-R1 activity, when measured as early as one week after onset of stimulation [32]. Expression of TGFβ-R1 and PAI-1, a measure for TGFβ-R1 signaling were downregulated in both the MI control and IPTVVI group as compared to healthy control animals underlining the transient nature of the local inflammatory responses at the molecular level to regional electrical stimulation. Although in agreement with a previous study from our laboratory [42], TGFβ3 was upregulated in the infarcted tissue, this was not altered by IPTVVI, making it unlikely that activation of the TGFβ-pathway at five weeks follow-up contributed to the increased presence of the myofibroblasts in the IPTVVI treated animals.

In conclusion, the present study shows that 5 weeks of IPTVVI, a regionally targeted non-pharmaceutical approach that was safe (no arrhythmias and maintained cardiac output), favorably influenced the infarct remodeling process, likely by increasing myofibroblast content in the infarct region. Thus, whereas MI control swine showed a reduction in infarct mass over the 5-week follow-up period, which was principally due to infarct thinning, IPTVVI resulted in a reduction in infarct mass that was principally due to a decrease in the number of infarcted LV segments while infarct thinning was prevented. Histological assessment revealed increased numbers of myofibroblasts in the infarct zone. Taken together, these findings suggest that IPT in the peri-infarct zone represents a novel adjunctive therapy to favorably modulate infarct healing in patients with acute myocardial infarction.