We appreciate the interest of Dr. Jaeschke and colleagues in our previous work1 reporting that following acetaminophen (APAP) overdose, cell death products attract neutrophils to the liver, where they act as key protagonists of liver injury amplification. We expanded previous findings2, 3 by demonstrating that necrosis-derived products and inflammatory mediators, in particular mitochondrial DNA (mitDNA) and CXCR2-chemokines, reached systemic circulation following acute liver necrosis and could be responsible for the remote lung injury observed in mice treated with APAP.1 Corroborating this finding, blood samples from acute liver failure (ALF) patients presented increased circulating levels of mitDNA and CXCL8 compared with healthy volunteers.
APAP overdose causes intracellular accumulation of the reactive metabolite NAPQI (N-acetyl-p-benzoquinoneimine), which leads to intracellular oxidative stress, mitochondria dysfunction, DNA damage, and further cell death by way of oncotic necrosis.4 Following necrosis, bona fide intracellular contents are released into the interstitium and activate the immune system via several receptors, including Toll-like, cytokines, chemokines and formyl peptide receptors (FPR1).5 In fact, the administration of cell-derived contents or necrotic cells per se triggers robust neutrophil recruitment and inflammation in several experimental models.3, 6, 7 Considering these facts, the contribution of neutrophils would be expected in necrotic diseases in general. Accordingly, previous studies have shown the detrimental role of neutrophil infiltration in liver injury models other than APAP-dependent necrosis, including ischemia-reperfusion (I/R) and carbon tetrachloride (CCl4)-induced hepatotoxicity.8-11 These data together reinforce that, independent of the cause of sterile liver necrosis, the release of cellular products will attract and activate potentially harmful neutrophils and this may be a cornerstone of injury amplification.
Our group previously reported that neutrophils are attracted to sites of liver necrosis following an intravascular gradient of CXCR2-chemokines and mitochondria-derived formyl peptides.12 Nevertheless, it remained unclear whether the consequence of liver neutrophil recruitment during sterile inflammation is resolution or additional damage. Hence, we used a number of different research tools to reduce the intrinsic technical limitations. Neutrophil depletion by anti-Gr1 antibody treatment is widely accepted in the literature,13-15 despite a few reports regarding the restrictions of its use as a method to study the role of neutrophils in liver injury.16, 17 To circumvent this issue, we also used pharmacological blockage of different neutrophil chemotactic receptors (CXCR2 and FPR1 antagonists) to inhibit liver neutrophil migration and activation and the findings were consistent: the reduction in neutrophil infiltration culminated in less liver damage.1 Both the antagonists used in our study were vehicle-controlled and there are no reports in the literature of any intrinsic activity of these drugs on APAP metabolism. Furthermore, it has been previously reported that both CXCR2 and FPR1 antagonists also have potent antiinflammatory properties in non-APAP related experimental models.18, 19
The challenge was to determine whether reduced neutrophil infiltration was a cause or a consequence of milder liver injury using only in vivo approaches. For this, we attempted to reproduce the encounter of neutrophils with hepatocytes in vitro using HepG2 coculture and then blocked the CXCR2 and FPR1 receptors. In agreement with previously established optimal conditions for coincubation experiments,20 neutrophils were directly cytotoxic to HepG2 cells, and both HepG2-separation by transwell inserts and CXCR2-FPR1 antagonism significantly reverted neutrophil killing behavior, which could be the result of reduced reactive oxygen species (ROS) production.1 Likewise, in the simultaneous presence of both APAP and neutrophils, the resulting HepG2 necrosis was even greater. Despite the extensively debated limitations for translating in vitro data to an in vivo scenario, our in vitro strategy confirmed findings from the murine model. Thus, combined with an in vivo model, neutrophils-HepG2 coculture (or alternative hepatic lineages) may provide mechanistic opportunities for cytotoxicity studies.
In conclusion, liver injury following APAP overdose is a result of both immune system-dependent cytotoxicity and direct APAP metabolism hepatotoxicity (Fig. 1). Limiting the reactive immune response to cell death products diminished the progression of ALF in mice.1 The concerns highlighted by Dr. Jaeschke and colleagues reveal the need for further studies in developing novel and more consistent neutrophil inhibition/depletion strategies, since promising new therapeutic approaches rely on the modulation of neutrophil emigration and activation within sites of sterile inflammation.