In this study, we showed that injurious MV with high Vt in a model of LPS-induced pneumonia aggravated LPS-induced myocardial dysfunction ex vivo. In contrast with our hypothesis, we could not confirm alterations in Ca2+ sensitivity or calcium handling. Additional analyses suggested that myocardial dysfunction was potentially mediated through pulmonary release of HSP70 stimulating myocardial TLR2.
We used a two-hit model with intratracheal LPS installation and MV with Vt of 19 ml/kg to create VILI in line with previous studies [20, 21]. This approach resulted in lung injury: pulmonary gas exchange deteriorated and the pulmonary wet-to-dry ratio increased, consistent with VILI. Mortality was high as in the LPS and high Vt group only one animal survived 4 hours, but this did not preclude ex vivo studies. Furthermore, myocardial function was impaired as demonstrated by a decreased CO in vivo without a fall in CVP, and a decreased LV systolic pressure, developed pressure and +dP/dtmax and increased –dP/dtmin ex vivo following LPS installation [5, 21]. Moreover, we found that the LPS-induced myocardial dysfunction in vivo and ex vivo was aggravated by high Vt ventilation and development of VILI over time as shown by a further worsening of LV function parameters. It can be surmised that the mechanisms underlying VILI and myocardial dysfunction are comparable to those in sepsis, including myocardial inflammation . In fact, MV may evoke an inflammatory response that is similar to that evoked by endotoxin, signifying a synergistic (i.e. double-hit) effect [28, 29]. Injurious MV augments lung injury caused by intra-tracheal instillation of bacterial products [30, 31]. Our results are in line with such a synergistic effect of MV in lungs and heart and we examined potential routes.
The causative mechanisms underlying the observed decrease in ex vivo myocardial function may be multifactorial. We hypothesized that high Vt ventilation caused myocardial dysfunction through a Ca2+ dependent manner. First, it can be postulated that distant organ failure after experimental injurious MV might be due to spillover of LPS from the injured lungs into the systemic circulation [31, 32] which can then decrease myocardial function by inflammatory responses decreasing myocardial Ca2+ handling and sensitivity [12, 13]. Although LPS-induced myocardial dysfunction in our study was associated with a decreased cardiac pCa50, high Vt ventilation did not affect cardiac pCa50, thereby excluding an effect on Ca2+ handling and sensitivity and rendering injurious MV-induced pulmonary release of LPS in the systemic circulation as a mediator unlikely. As we measured Ca2+ sensitivity in hearts with intact cell membranes, the Ca2+ sensitivity in this study should otherwise not be confused with myofilament sensitivity . As we cannot distinguish between myofilament sensitivity and Ca2+ handling we cannot exclude the possibility that both parameters were altered. Second, acidosis may decrease myocardial function through a Ca2+-dependent mechanism . Nonetheless, the acidosis we observed in vivo is less likely to be a cause of the observed dysfunction aggravated by injurious MV. Finally, hypoxia-induced cardiac dysfunction is also Ca2+-dependent as it effects the L-type Ca2+ channel [15–17]. However, we think that the level of hypoxia is also less likely to be cause of the observed dysfunction as there was no interaction between LPS and high Vt ventilation. Furthermore, inadequate delivery of O2 seems unlikely. In our experiments, hypoperfusion seems unlikely as a contributory factor because the coronary flow ex vivo was similar between groups. Myocardial edema may affect myocardial function but no difference in cardiac wet/dry weight ratio was found, thereby excluding edema as a causative factor .
Hence, we may conclude that the VILI-induced aggravation of myocardial dysfunction was most likely not mediated by a Ca2+-dependent pathway. In fact, our data suggest that a factor related to the severity of VILI contributed to myocardial inflammation and dysfunction, for instance via ventilator-induced lung-borne HSP70, was involved. Extracellular HSP70 can bind to myocardial TLR2 and may thereby cause an inflammatory response indicated by an increase in myocardial CXCL1 expression and Ca2+-independent myocardial dysfunction . So, this route was subsequently studied when myocardial dysfunction appeared independent of Ca2+ handling in our experiments. We found that pulmonary but not myocardial HSP70 mRNA expression increased in severe VILI, after high Vt ventilation in combination with LPS exposure, as shown before [24, 25]. We only may speculate about its involvement. In favor of our speculations is that it has been shown before that HSP70 may be actively released by cells under stress . Indeed, injuring the lung with diesel exhaust, caused increases both pulmonary and systemic HSP70 expression . Thus, HSPs may have a cytoprotective effect in VILI, but may also be involved in the activation of innate immunity when released into the systemic circulation. Explaining the high mRNA expression in the low Vt group in the absence of LPS is challenging as there is no explanation readily available. It has been discussed that the expression of HSP70 does not result from the transcription of a single gene, but is derived from what may be a complex interplay of several underlying genes . It may thus be postulated that the HSP response following MV in the absence of LPS in our model is dependent upon both the degree of lung injury, the inflammatory response and their time course, so that we may have missed upregulation of HSP70 in the high Vt control group [38, 39]. We have previously observed such a time course for TNF-α in a rat model of endotoxaemia-induced peritonitis . This aspect warrants further investigation. Myocardial HSP70 mRNA expression was not increased, so that it is conceivable that pulmonary release of HSP70 had contributed to myocardial dysfunction. The myocardium of LPS-treated animals was probably more prone to bind lung-borne extracellular HSP70 by TLR2, since TLR2 was already upregulated by LPS in agreement with others . We also found induction of CXCL1 after LPS and even more so after high Vt ventilation and CXCL1 mRNA expression inversely correlated with LV systolic pressure. Its exact role needs to be explored in further study. However, support for a Ca2+-independent role of CXC1 in myocardial dysfunction, as suggested by our study, comes from another report [26, 41] and inhibition of CXCL1 decreases right ventricular failure in a model of pulmonary embolism . In any case, myocardial expression of other inflammatory mediators was not increased, which might be due to a different time course of expression: TNF-α expression is increased 2 hours after LPS exposure, but not detectable anymore after 4 hours .
There are some limitations to the present study that may influence the interpretation of the main findings. First, mortality was high in the group of rats subjected to both high Vt ventilation and LPS. Only one animal completed the four-hour study protocol. This indicates that the time point at which the myocardial function was measured ex vivo as well as well when the organs were collected for analysis was not consistent among the animals in this group (the earliest being after 2.5 hrs of ventilation), but this does not invalidate our conclusions. Second, we were only able to measure mRNA expression levels and not the actual protein contents, but time may be needed for protein synthesis to become manifest after gene expression. Also, the concept that VILI-derived pulmonary HSP70 is directly responsible for myocardial inflammation and Ca2+-independent dysfunction warrants further investigation by mechanistic or intervention studies. We were unable to measure HSP70 in serum as its mechanistic role was not part of the primary hypothesis of our exploratory study. Our study was primarily designed as a proof of principle study to test the hypothesis that a Ca2+ dependent pathway was involved. When we refuted this hypothesis based on our findings, we explored alternative pathways in available tissue samples. Furthermore, measuring HSP70 in serum does not answer the question whether or not it is lung-borne, unless venous and arterial blood samples are taken and a pulmonary gradient can be measured.
In conclusion, our study shows that injurious ventilation with high Vt aggravates the effects of lung injury caused by intratracheal instillation of LPS on myocardial dysfunction, possibly through enhancing myocardial inflammation by lung-borne mediators, not involving Ca2+ handling and sensitivity.