To the Editor: Angiogenesis and atherogenesis share a number of pathways suggesting that interrelated tradeoffs might be inherent to therapies designed to enhance collateral formation and cardiac repair (1). This “Janus-like” effect on atherosclerosis progression was observed in several experimental models exposed to therapy with bone marrow stem cells (BMSCs) (1- 4). A recent clinical study reported high rates of in-stent restenosis after BMSC mobilization (5). In our study (6), CD133⁺ enriched BMSC exerted beneficial effects on cardiac recovery in patients with reperfused myocardial infarction, but the restenosis rates were also higher as expected after intracoronary BMSC therapy (7- 11). Hence, we investigated the effects of intracoronary CD133⁺ enriched BMSC on in-stent neointimal proliferation and distal atherosclerosis progression in patients with recent myocardial infarction treated with angioplasty using bare-metal stents.

Thirty-eight patients with acute myocardial infarction due to occlusion of the proximal left anterior descending coronary artery (LAD) were studied. The cell group consisted of 21 patients receiving intracoronary injection of CD133⁺ enriched BMSC and follow-up catheterization with quantitative coronary angiography (QCA) and coronary functional assessment with the pressure-derived fractional flow reserve (FFR) (12- 13). The control group consisted of 17 patients matched for ejection fraction, infarction size, and location with control catheterization between 4 and 8 months after the infarction. Segmental QCA analysis was performed for the stented segment and non-stented portion of the mid LAD and distal LAD. Data are shown as mean ± SEM. Paired t test and unpaired t test were used as appropriate.

There were no differences in clinical and demographic characteristics between groups (not shown). Baseline angiographic and functional characteristics are shown in (Table 1). At follow-up, no significant changes were noted in luminal diameters of the contralateral artery (CLA) of either group (from 2.66 ± 0.09 mm to 2.65 ± 0.10 mm in the cell group, and from 2.71 ± 0.14 mm to 2.78 ± 0.17 mm in the control group, both p = NS). In contrast, higher loss index of the stented segment in the cell group (−0.42 ± 0.07 vs. −0.22 ± 0.06, p < 0.05) was paralleled by a greater leftward shift in the cumulative distribution of the minimal luminal diameter (MLD) as compared with the control group (Figure 1A). In non-stented segments, cumulative luminal loss of the reference diameter (RD) and MLD at the mid and distal segment of the infarct-related artery (IRA) were higher in the cell group (Figure 1B) or in the subgroup of patients without significant in-stent restenosis, as compared with relevant control subjects (Figure 1C). Significant distal de novo stenosis (>50% stenosis) was seen in two patients in the cell group and none in control subjects. In a subset of the cell group patients undergoing serial intravascular ultrasound study (n = 5), plaque burden significantly increased in the IRA (from 45.9 ± 4.3% to 56.3 ± 3.9%, p < 0.05) but remained unchanged in the CLA of the same patients (from 40.5 ± 1.3% to 40.3 ± 1.2%, p = NS).

Luminal changes by QCA were associated with a significant decrease in FFR of the IRA in the cell group, with no change in the control group (−0.21 ± 0.05 vs. 0.01 ± 0.04, p < 0.05). In patients without in-stent restenosis, the cell group showed a higher decrease in FFR as compared with relevant control subjects (−0.13 ± 0.05 vs. 0.06 ± 0.03, p < 0.05). The cell group patients receiving ≥9 × 10⁶ CD133⁺ cells showed larger luminal loss in the mid non-stented segment of the IRA as compared with patients receiving a lower number of cells (−0.87 ± 0.19 mm vs. −0.25 ± 0.21 mm, p < 0.05). The cell group patients with in-stent restenosis or a de novo stenosis showed lower systemic interleukin-10 levels (1.95 ± 0.96 ng/ml vs. 5.01 ± 1.82 ng/ml, p = 0.07) and higher vascular endothelial growth factor (VEGF)-A levels (215.5 ± 42.9 pg/ml vs. 86.3 ± 29.0 pg/ml, p < 0.05) as compared with the cell group patients without new lesion or restenosis. No differences in plasma interleukin-6, monocyte chemoattractant protein-1, and basic fibroblast growth factor levels between groups were observed (not shown).

Present data indicate that intracoronary injection of enriched BMSC is associated with greater in-stent proliferation and larger luminal loss in non-stented IRA segments that result in a significant decrease in pressure-derived FFR. These changes are consistent with decreased epicardial conductance of the IRA, owing to diffuse luminal loss and increased plaque burden. Our study suggests several mechanisms for the “Janus-like” effect. First, the total number of CD34⁺ or CD133⁺ cells in the cell suspension was higher than in previous studies (5- 11), and patients with a higher number of injected CD133⁺ cells showed a larger luminal loss of distal non-stented segments. Hence, dose-dependent increase in local concentrations of pro-angiogenic molecules might have exacerbated pro-atherogenic effects. Second, cell-treated patients with in-stent restenosis or a de novo stenosis showed lower levels of interleukin-10 and higher levels of serum VEGF-A, suggesting disequilibrium between pro-atherogenic and anti-atherogenic factors as a predisposing factor. Nevertheless, local concentration of inflammatory cytokines or biological activity of CD133⁺ could provide further insights into these hypotheses. As an alternative mechanism, in-stent proliferation might be related to repetitive balloon occlusion at the time of the cell injection, albeit at low inflation pressure, by impairing ongoing re-endothelialization. Note that earlier studies (6,14) demonstrated immunological safety of immunomagnetic isolation with absence of human anti-mouse antibodies after stem cell enrichment in bone marrow transplant patients, which argues against immunological response as the underlying mechanism.

In conclusion, our findings of reduced luminal diameters and epicardial conductance of the IRA, as assessed from the pressure-derived FFR, seem to be consistent with a higher risk for atherosclerosis progression after intracoronary administration of enriched BMSC. Yet, these data were obtained from retrospective analysis; they lack randomized controls and systematic use of intravascular ultrasound imaging to track changes in the vascular wall. Future and ongoing randomized studies using intracoronary injection of enriched and unfractionated BMSC should define the risk and mechanisms of potential “Janus-like” effects on the epicardial coronary circulation.