Thank you. Thanks for the opportunity to talk about this topic. I think it's a really interesting area. So as you know, in some patients with sepsis, hemolysis is sufficient to overwhelm the protective mechanisms and contribute to organ injury and mortality. And the thesis of my talk, following on from the prior discussion, is that hemolysis then could be a target for precision therapies in sepsis. My objectives are to review the mechanisms of hemolysis and endogenous protection, and to describe the association between hemolysis and poor outcomes in sepsis, to recognize the mechanisms of hemolysis-mediated injury, and to describe some therapeutic strategies, and discuss some of the open questions in how this pathway might be applied in precision therapies. As you're all familiar, free hemoglobin and heme are highly biologically active, and there are multiple pathways and mechanisms that have evolved to maintain the redox state of iron when it's in a red cell, and to mop up hemoglobin once it escapes from the red cell. But when hemolysis does occur, the first line of defense really is haptoglobin. So cell-free hemoglobin is bound by haptoglobin, which is cleared by macrophages and monocytes in circulation via high affinity binding with the receptor CD163. Free hemoglobin that escapes from haptoglobin breaks down, releasing hemoglobin dimers and ultimately free heme, which is an iron atom in a porphyrin ring, and that can be bound up by hemopexin, which is cleared through CD91. There is also some binding of free heme to albumin and lipoproteins in the circulation. Once in the cell, heme is broken down by a trio of enzymes, and the pathway is really dependent, or the key regulatory step is heme oxygenase 1, which has been recognized for decades. Heme is broken down to iron, which is bound up by ferritin, and to carbon monoxide and biliverdin. We know, though, that in situations of hemolysis, like in sepsis, these systems can be overwhelmed, and that can lead the free hemoglobin and the free heme to cause organ injury and increase in mortality, and this has been recognized in sepsis. So a Dantzig study in 2012 showed that free hemoglobin above the median on the day of severe sepsis diagnosis was associated with mortality in a group of 160 patients. And Jansz in 2013 studied this group of 380 patients with severe sepsis and found that any detectable cell-free hemoglobin in their assay at 24 hours was associated with mortality, and at that same 24-hour point, a higher haptoglobin was associated with survival. What's the mechanism, or the primary mechanisms, of this injury? So free hemoglobin reacts rapidly with nitric oxide to form methemoglobin and nitrate, and that local depletion of nitric oxide leads to vasoconstriction, activates the endothelium, leads to leukocyte adhesion and extravasation, and also leads to platelet activation, adhesion, and thrombosis. When in the red cell, the iron atom is kept very strictly. The redox state is maintained, but free hemoglobin escapes those mechanisms and is oxidized to the 3-plus and the 4-plus state. And those higher oxide forms can catalyze Fenton and Haber-Rice reactions and promote the production of free radicals. It can also react, free hemoglobin can also react directly with lipopolysaccharides in the membrane to cause direct damage. And once broken down, cell-free hemoglobin provides a source of free iron, which also promotes bacterial growth. When the free hemoglobin releases free heme, this is also extremely biologically active and has a range of immune modulatory effects that could be a whole discussion on its own. The molecule is hydrophobic and intercalates into the lipid membrane, where it promotes the same types of oxidative damage and blocks the effects of the proteasome that would normally help the cell deal with some of those. It also has profound immune modulating effects in neutrophils and in monocytes and macrophage populations. Initially, a lot of that is a pro-inflammatory effect, but over time, it seems to have almost an immunosuppressive effect, which we'll discuss. So if we're talking about investigating whether this pathway could be used in a precision therapy model, we have to be able to identify which of these patients will have clinically significant hemolysis. And so we need to think about what factors contribute to what is clinically significant. So part of this, presumably, is the total amount of hemolysis. So this is the mass of red cells broken down. Some of this might be from transfusions or other hemolytic insults. Another factor is the pattern of hemolysis, which we think is important. At one end, it may be someone who has a massive trauma followed by massive transfusion who gets a huge bolus of free hemoglobin. The alternate end might be a patient with sickle cell disease who has a more indolent course. There's also variation in each patient's ability to manage hemolysis products that could play into this. So we're going to try and work out which patients will have the most hemolysis to try and identify which patient populations might benefit from this precision therapy. So what are the causes of hemolysis in sepsis? Certainly there are pathogens that cause direct hemolysis, like malaria. Sepsis also leads to changes in the red cell membrane stability. This could be an LPS-mediated effect, which reduces membrane deformability, and also a pathogen-mediated effect. I included a couple of examples, but the Staph aureus alpha hemolysin, of course. Sepsis also leads to a microvascular stasis, and this can create a vicious cycle where vasoconstriction and endothelial dysfunction promote hemolysis. Hemolysis releases free hemoglobin, and that promotes further vasoconstriction and endothelial dysfunction. Sepsis complications, like DIC, can also lead to hemolysis, likely through the sharing effects of fibrin strands. There's complement-mediated red cell damage in sepsis, and also additional metabolic stress that affects red cells because of their dependence on glucose. Red cell transfusions, which may be part of the treatment for some subset of patients in sepsis, also introduce new, some hemolytic products. And I've included eryptosis, although it's not clear at this point whether this might be a protective effect or a damaging effect in these patients. So if we try and break these things down, if we're identifying a group of patients who might benefit from this type of therapy, one component is likely sepsis severity. So some portion of the sicker patients will likely have more hemolysis based on the mechanisms. One component is the development of patients who develop specific sepsis-related complications, like DIC, might be targets. And also specific pathogens seem to be important, whether that's pathogens that cause direct damage or those that have relevant toxins. The timing question is also interesting, and I don't think we have an answer to that. But what creates that peak of hemolysis is relevant. This is a nice experiment that Bala and the group did in 1993 that showed that initial exposure of the vascular endothelium to heme products promotes production of inflammatory signals and sensitizes the endothelium to oxidative injury. But actually, a prolonged exposure or an exposure followed by a delay is actually protective. So the chart at the bottom is two groups of cultured endothelial cells that were exposed in the dark line to hemine or methemoglobin or to PBS for an hour. The cultured cells are then left for 15 hours and then exposed again to heme and to peroxide. And the group or the group of endothelial cells that had been previously sensitized by exposure to heme are protected from further injury, likely through induction of heme oxidase 1 and some increase in the ferritin levels. So if we can identify which group of patients will have the most hemolysis, the next thing is which group of patients are the most susceptible to hemolytic injury, because this also varies. One thing that's come up in the last 10 or so years is the variation in human haptoglobin. So haptoglobin is highly abundant. The expression of haptoglobin, which is an acute phase reactant, of course, is induced by catecholamines, endotoxin, IL-6, IL-1. There are actually two human alleles of haptoglobin, which lead to three major phenotypes. And the three phenotypes have different activities when it comes to clearing free hemoglobin. It turns out that the haptoglobin 2 phenotype is actually less effective at blocking the inflammatory and oxidative effects of free hemoglobin. And Kirchberger in 2019 published a nice study showing that patients who were homozygous for haptoglobin 2 who had sepsis and detectable free hemoglobin, so already narrowing the population down, had increased risk of developing ARDS. The other patient-to-patient variable that's been identified that might be relevant is actually changes in heme oxygenase 1 polymorphisms. This is a nice study from Xu in 2009. So they demonstrated that polymorphic GT repeats in the heme oxygenase promoter regulated expression of the protein. And actually, they found that patients in the ICU with risk factors for ARDS who had longer GT repeat sequences had reduced ARDS risk. So they identified a subpopulation who were protected from hemolytic injury. So if we can identify who is going to be at risk of a lot of hemolysis and who is going to be susceptible to hemolytic injury, how would we go about treating them? Ideally, we would prevent the hemolysis. And some of that is the non-precision part that we've been talking about, identifying the source of sepsis, treating source control, those kind of things. If we can identify a population who it might benefit, we could supplement endogenous hemoglobin or heme binding proteins, namely haptoglobin and hemopexin. There's some chance, or there's some early studies showing that traditional antioxidant drugs can help, niacin, glutamine, ascorbate. There's some studies on Tylenol that can block some of the redox effects of these pathways. Several groups have tried to improve nitric oxide hemostasis. For example, with inhaled nitric oxide. And then it is possible to augment the expression of some of the clearance mechanisms. And the most notable one is that CD163, which is involved in haptoglobin uptake, is actually expressed, or the expression is increased, in response to steroids. This is actually some of Dr. Remy's experiments studying the therapeutic effect of haptoglobin in dogs who have Staph aureus pneumonia and who received exchange transfusion or not. On the left is a chart with, so the dogs have a Staph aureus pneumonia and have received exchange transfusion. And the purpose of the exchange transfusion is largely to introduce those hemolytic products and are treated with haptoglobin in the dotted line or without haptoglobin in the solid line, showing improved survival with haptoglobin treatment. One of the interesting things, or most interesting things, in my mind about the series of experiments is the chart on the right. And so these dogs had the same Staph aureus pneumonia without exchange transfusion. So no exogenous source of free hemoglobin. And while there wasn't a statistical difference in survival, there was still a large numerical difference in survival between those dogs that were treated with haptoglobin and those treated with an albumin control. This is a nice study in mice looking at the therapeutic potential of hemopexin. And so this is, the chart shows survival days after cecal ligation and puncture in mice who are treated with intraperitoneal PBS in the clear boxes, IgG in gray, or hemopexin in black, showing, again, a significant improvement in survival in mice treated with hemopexin. This is a randomized trial performed by Jans in the group and published in 2015 that took patients with severe sepsis and randomized them to receive just a standard dosing of enteral Tylenol or placebo for three days, so one gram Q6 for three days in patients with sepsis, and showed a reduction, significant reduction, in the markers of oxidation and a reduction in the day three creatinine, suggesting that even some therapies that are pretty familiar to us and common could affect these pathways. So what are the open questions? I think we still need to really nail down how we would identify which patients with sepsis have clinically significant hemolysis. And then another open question is, when would we treat? Ideally, as we talked about, we would treat everyone before the fact. We'd boost everyone's haptoglobin right before they got their sepsis, but that's obviously not feasible. So we would probably want to treat as early as possible and may want to give additional therapy at times of anticipated additional hemolysis, be that invasive procedures or transfusions. So in conclusion, hemolysis is a clinically relevant mechanism of injury in sepsis, and there are pretty good therapies and leads for other therapies that can block some of these pathways. There is variation in both the degree of hemolysis and the susceptibility to hemolytic injury across patients with sepsis that could lend itself to a personalization of treatment. And so it could be another avenue for this type of personalized therapy that we've been talking about. Thank you.

hemolysis

sepsis

organ injury

precision therapies

haptoglobin