HR during CPB is a serious clinical issue that can seriously affect patient safety by prolonging operative times or leading to discontinuation of the procedure. While clinical practice guidelines have been established for AT-mediated HR, none exist for non-AT-mediated HR. Therefore, definitive treatment strategies for non-AT-mediated HR have not been clearly defined. Recently, case reports have shown that AT improves non-AT-mediated HR (2, 18); therefore, we aimed to experimentally confirm the effect of AT on AT-mediated HR. To the best of our knowledge, this is the first experimental report demonstrating that AT improved non-AT-mediated HR.
Many mechanisms may potentially lead to non-AT-mediated HR (e.g., increased heparin binding protein, high platelet count, low albumin concentrations, preoperative relative hypovolemia, and certain medications) (1, 17, 19–21), and we hypothesize that various factors are involved in the induction of non-AT-mediated HR. In the present study, we focused on PF4, one of the heparin binding proteins, to develop non-AT-mediated HR models. PF4, a chemokine of the CXC family present in the α-granules of platelets, is known as a risk factor for non-AT-mediated HR because it is easily released upon platelet activation (22, 23) and is a potent heparin inhibitor. Furthermore, since rPF4 does not cause hematological or hemodynamic abnormalities when administered to animals or humans (24–26), we developed a model of non-AT-mediated HR using rPF4 in this study.
First, we established that rPF4 induces non-AT-mediated HR by inhibiting the anticoagulant effect of heparin without affecting AT activity. The normal blood concentration of PF4 is < 20 ng/mL. However, PF4 has been reported to be released in concentrations > 2 µM (< 12 µg/mL) in the vicinity of injured vessel walls upon platelet activation (27). Thus, we selected the rPF4 treatment dose based on that condition (in vitro study: 1–25 µg/mL; in vivo study: the blood concentration in mice at a dose of 100–400 µg/kg is approximately 1.4–5.5 µg/mL). Our in vitro study showed that 25 µg/mL rPF4 completely neutralized 0.5 U/mL heparin. This result is consistent with that in a previous report (28) and shows the validity of the current experimental model. We measured APTT instead of activated clotting time (ACT) for the experimental murine model because ACT is unsuitable as a rodent coagulation measurement assay due to its brevity and great variability (29).
Next, we experimentally demonstrated that AT ameliorated non-AT-mediated HR. Our in vitro and in vivo studies showed that pAT improved the rPF4-induced HR in a dose-dependent manner. Our in vitro study showed no difference in APTT between the rPF4 5 µg/mL + pAT 30 U/mL group and the pAT 30 U/mL group, suggesting no interaction between rPF4 and pAT. In addition, rPF4 5 µg/mL + pAT 30 U/mL slightly prolonged APTT, whereas rPF4 5 µg/mL + heparin + pAT 30 U/mL markedly prolonged it, suggesting that the improvement effect of AT on non-AT-mediated HR was mediated by heparin.
In addition, we performed a ROTEM analysis using human whole blood to validate these results under near-clinical conditions. Therein, rPF4 was confirmed to inhibit heparin-induced CT prolongation, and pAT improved rPF4-induced HR, corroborating the in vitro and in vivo results. Of note, AT suppresses fibrin clot formation by inhibiting thrombin. This may explain the changes in maximum clot firmness observed at 30 U/mL pAT in our study. Importantly, our results suggest that high doses of AT may affect clot firmness.
We also investigated the onset mechanisms of non-AT-mediated HR by rPF4 and how pAT ameliorates non-AT-mediated HR. Lane et al. (28) speculate that PF4 inhibits heparin through the following mechanisms: (a) PF4 sterically inhibits the binding of heparin to AT or (b) a PF4-heparin-AT complex is formed, and the PF4 section of the complex sterically inhibits the thrombin/Xa binding site of AT. In our in vitro experiments, the anticoagulant effect of AT was maintained even when AT was added to a sample with heparin completely neutralized by rPF4. Thus, the PF4-heparin-AT complex is unlikely to have formed; that is, the heparin inhibitory mechanism of PF4 is (a) above. Furthermore, we confirmed that rPF4 has a higher binding affinity for heparin than pAT, consistent with a previous report (30). Therefore, we concluded that cationic PF4 induces HR by preferentially binding to polyanionic heparin over endogenous AT, reducing the amount of heparin that can bind to AT. Conversely, our in vitro study showed that the improvement effect of AT on HR is mediated by heparin. Therefore, we speculate that the non-AT-mediated HR amelioration mechanism of AT involves an increase in the absolute number of AT-heparin complexes.
The present study focused only on PF4, one of the potential risk factors of non-AT-mediated HR. However, we speculate that various factors may contribute to the onset of non-AT-mediated HR. Indeed, some patients with non-AT-mediated HR have elevated blood levels of other heparin-binding proteins (e.g., fibrinogen and FVIII) (2, 19, 31, 32). Since the non-AT-mediated HR amelioration mechanism of AT inferred in this study is via binding of AT to heparin, we suggest that elevated blood AT levels, including concentrated human AT administration, may alleviate non-AT-mediated HR caused by any heparin-binding protein.
Lastly, we examined the causes of elevated blood PF4 levels using our rat models. Heparin treatment has been reported to minimally activate platelets (33, 34) and was presumed to release only minimal amounts of PF4. Our in vivo and ex vivo results supported this hypothesis. In contrast, LPS administration in rats greatly increased blood PF4 levels to the same extent as treatment levels in our in vitro and in vivo studies. This may be because platelets were activated by inflammation and thrombotic/hemostatic reactions following the cytokine storm induced by LPS. Thus, we suggest that inflammation and thrombotic/hemostatic reactions are involved in elevated blood PF4 levels. In the condition requiring extracorporeal circulation (i.e., myocardial infarction or aortic dissection), those reactions are assumed to be occurring and PF4 levels in the blood may be elevated. In addition, our findings that blood PF4 levels increased with LPS administration imply that infections may also induce HR. Recently, ECMO use has increased rapidly owing to the coronavirus disease 2019 (COVID-19) pandemic (35). Some patients with COVID-19 have been reported to have non-AT-mediated HR and often have elevated blood fibrinogen and FVIII levels, erythrocyte sedimentation rate, and C-reactive protein, indicating a severe inflammatory condition (36–38). The relationship between HR in COVID-19 and PF4 remains unclear; however, we speculate that blood PF4 levels in patients with moderate and severe COVID-19 are likely elevated because anti-PF4 polyanion antibodies (i.e., anti-PF4 antibodies), such as heparin complex antibodies, are detected in 95% of hospitalized patients with COVID-19, irrespective of previous heparin treatment (39). Therefore, AT administration may be effective against HR in patients with COVID-19.
A limitation of our study is that it is an in vitro and in vivo study, which differs from the conditions in the human body. Further clinical studies are required to confirm these findings in humans.