Sepsis has been a major threat to global health over the past several decades. In the United States, approximately one million individuals are diagnosed with sepsis annually, with mortality estimated between 12 and 25 percent (Mayr et al., 2014; Paoli et al., 2018). An estimated 20 percent of all deaths globally were attributed to sepsis (Rudd et al., 2020). The more severe septic shock has an estimated 38 percent mortality, and half of all Americans who die in the hospital are diagnosed with sepsis (Liu et al., 2014; Vincent et al., 2019). Sepsis occurs when an infection triggers a dysregulated host immune response, leading to systemic microcirculatory and immune dysfunction. This dysregulated inflammatory response in the microvasculature leads to direct damage of cells from reactive oxygen species and other inflammatory mediators, activation of the coagulation cascade, vasodilation, and tissue hypoxia with subsequent mitochondrial dysfunction. This complex system culminates in life-threatening organ injury and metabolic derangements (Chuang et al., 2006; Robertson and Coopersmith, 2006; Galley, 2011; Angus and van der Poll, 2013; Delano and Ward, 2016; Singer et al., 2016; Prauchner, 2017). Lipids and lipid-related signaling pathways have been investigated as mediators, potentially at the blood-endothelial interface during sepsis (Amunugama et al., 2021a). Specific lipids may also have prognostic value as biomarkers in sepsis (Meyer et al., 2017; Mecatti et al., 2018; Mecatti et al., 2020; Wang et al., 2020; Amunugama et al., 2021a). Additionally, a major cause of COVID-19 mortality is sepsis-associated acute respiratory distress syndrome (ARDS). Similar to sepsis, lipids have been investigated as important mediators and biomarkers in COVID-19 (Tanner et al., 2014; Aktepe et al., 2015; Villareal et al., 2015; Jean Beltran et al., 2018; Fernández-Oliva et al., 2019; Sviridov et al., 2020; Casari et al., 2021; Mesquita et al., 2021; Theken et al., 2021).

Plasmalogens comprise a significant fraction of the lipid content in the plasma, immune cells, and endothelium (Chilton and Murphy, 1986; Chilton and Connell, 1988; Kayganich and Murphy, 1992; Murphy et al., 1992; Bräutigam et al., 1996). There is considerable diversity in plasmalogen molecular species. In general, plasmalogens contain either phosphocholine or phosphoethanolamine at the sn-3 position of the glycerol backbone. The vinyl ether aliphatic group attached to the glycerol backbone predominantly contains sixteen and eighteen carbon groups. Recently we have also shown neutrophil plasmalogens contain vinyl ether groups that are greater than twenty carbons in length (Amunugama et al., 2021b). Plasmalogens have been suggested to have important roles in biological membranes, which are due, in part, to their unique packing in membranes compared to diacyl phospholipids (Han and Gross, 1990; Han and Gross, 1991). Plasmalogens have been shown to have roles in synaptic fusion, cholesterol efflux, lipid rafts, and transmembrane protein function (Glaser and Gross, 1994; Ford and Hale, 1996; Mandel et al., 1998; Pike et al., 2002). Plasmalogens likely have key roles in inflammation at several levels. Plasmalogens are plasma-borne antioxidants and have been shown to protect endothelium from oxidative stress (Vance, 1990; Zoeller et al., 1999). The vinyl ether bond of plasmalogens is susceptible to attack by reactive species, and this propensity suggests that these lipids can protect cells by scavenging reactive oxygen species (Zoeller et al., 1988; Reiss et al., 1997; Zoeller et al., 1999; Zoeller et al., 2002; Dean and Lodhi, 2018). Additionally, plasmalogens have been shown to have a key role in macrophage phagocytosis (Rubio et al., 2018). Furthermore, plasmalogens are enriched with arachidonic acid and docosahexaenoic acid at the sn-2 position, and their metabolism by phospholipases leads to the mobilization of these fatty acids and their subsequent oxidation to bioactive eicosanoids and resolvins (Paul et al., 2019). Collectively, the roles of plasmalogens in membrane molecular dynamics, as antioxidants, and as precursors of bioactive lipids indicate they may be important in inflammation associated with disease and infection.

Plasma plasmalogen levels have been shown to decrease during inflammation such as during endotoxemia (Ifuku et al., 2012), Parkinson’s disease (Dragonas et al., 2009; Fabelo et al., 2011), and lupus (Hu et al., 2016). Several of these previous studies (Dragonas et al., 2009; Hu et al., 2016) have suggested the loss of plasmalogens during Parkinson’s disease and lupus is due to the associated oxidative stress. Surprisingly only one study has investigated plasmalogen loss during human sepsis, which also attributed plasmalogen loss to oxidative stress (Brosche et al., 2013). This study was limited to measuring dimethyl acetals as a measure of plasmalogen levels and was performed in a limited number of geriatric septic patients. In addition to sepsis, several investigations have emerged over the past 2 years demonstrating plasma plasmalogen levels in humans with severe COVID-19 are decreased (Schwarz et al., 2021; Snider et al., 2021). The loss of plasmalogens and other phospholipids enriched with arachidonic acid and docosahexaenoic acid as well as increased secretory phospholipase A₂ (Snider et al., 2021) in COVID-19 patients support an important role for plasmalogens as precursors of oxylipids.

We have previously shown the plasmalogen vinyl ether bond is targeted by neutrophil-derived HOCl (a product of myeloperoxidase activity) resulting in 2-chlorofatty aldehyde and 2-chlorofatty acid production (Albert et al., 2001; Thukkani et al., 2002; Anbukumar et al., 2010). Furthermore, increased 2-chlorofatty acid plasma levels associate with ARDS-caused mortality in human sepsis (Meyer et al., 2017). 2-Chlorofatty acids are also elevated in the plasma and several organs in rats subjected to cecal slurry sepsis (Pike et al., 2020). Since plasmalogens are the precursors of chlorinated lipid production during sepsis and since limited molecular detail is known about human plasma plasmalogen loss during sepsis, in the present study we have investigated plasma plasmalogen levels in human sepsis patients. Furthermore, we have employed the rat cecal slurry sepsis model to identify both plasma plasmalogen loss as well as changes in liver and kidney plasmalogen levels during sepsis. Lastly, we examined changes in plasmalogen levels in plasma and lung in mice challenged with SARS-CoV-2. Collectively, these studies show the loss of plasmalogens during sepsis and SARS-CoV-2 infection with new detail into changes in plasma molecular species, as well as changes in organs in rodent models of sepsis and COVID-19.

Plasmalogens are a lipid subclass characterized by a vinyl ether linked aliphatic group attached to the sn-1 position of glycerol, a fatty acid esterified at the sn-2 position and, in general, either phosphoethanolamine or phosphocholine at the sn-3 position. The sn-2 fatty acid of plasmalogens is enriched with arachidonic acid in many mammalian tissues and thus one role of plasmalogens has been described as a storage depot for arachidonic acid that is released during inflammation (Chilton and Connell, 1988; Ford and Gross, 1989; Braverman and Moser, 2012). The sn-1 vinyl ether is a target for reactive oxygen species leading to the release of free fatty aldehydes that subsequently can be metabolized to free fatty acids (Khaselev and Murphy, 1999; Stadelmann-Ingrand et al., 2001). The reaction of reactive oxygen species with the vinyl ether is a terminal event for ROS and thus is considered an antioxidant activity. Multiple studies have shown plasmalogens protect tissues and cells from reactive oxygen species and oxidative stress. Cells deficient in plasmalogens are susceptible to free radical-mediated toxicity (Morand et al., 1988; Zoeller et al., 1988). Furthermore, supplementing cells with precursors to plasmalogens has been shown to protect cells from reactive oxygen species including during hypoxic damage to endothelial cells (Zoeller et al., 1999). Collectively, the abundance of arachidonic acid esterified to plasmalogens that can be mobilized for eicosanoid production and the susceptibility of the vinyl ether to oxidative stress suggest plasmalogens may have important roles in infection and inflammation. Plasma plasmalogen depletion has also been demonstrated in humans with Parkinson’s disease, Alzheimer’s disease, lupus and endotoxemia (Dragonas et al., 2009; Fabelo et al., 2011; Ifuku et al., 2012; Hu et al., 2016; Su et al., 2019). In the present study we provide further support for the involvement of plasmalogens in inflammation by providing molecular detail to changes in plasmalogen levels both in plasma and in organs during sepsis as well as SARS-CoV-2 infection.

Previous studies showed the 16:0 dimethyl acetal derivative of plasmalogens containing a sixteen-carbon vinyl ether aliphatic group bound to the glycerol backbone are decreased 55% in plasma of twenty geriatric septic patients compared to age-matched healthy subjects (Brosche et al., 2013). In this previous study, data for 18:0 dimethyl acetals were not reported and changes in 16:0 dimethyl acetal were from patient plasma collected within 24 h of severe sepsis diagnosis. Human plasma pPC levels are ∼8–10 fold greater than pPE levels, and pPC is highly enriched in molecular species containing a sixteen-carbon vinyl ether aliphatic group bound to the glycerol backbone, suggesting the plasma plasmalogens that decreased in geriatric sepsis patients (Brosche et al., 2013) are from pPC pools. In contrast to this previous study, our findings from the MESSI cohort were from patient plasma collected 7d following ICU admission for sepsis. This difference in time for plasma collection prevents direct comparisons to the previously reported study (Brosche et al., 2013). However, in the present studies pPE molecular species containing 16:0 vinyl ether groups, as well as 16:0 pLPC, were decreased in the human sepsis cohort. Plasma pPE species containing 18:0 vinyl ether groups were also significantly decreased in septic subjects investigated in our study. Future studies should be directed at determining details of plasmalogen loss at 24 h and examine longitudinal changes in plasmalogen loss. It will also be interesting to compare changes in human plasmalogen molecular species at 24 h to the changes we observed in the rat plasma plasmalogen molecular species that changed 20 h post CS injection. Interestingly with rat sepsis, plasma plasmalogen loss at 20 h decreased in several pPC and pPE species as well as pLPC. A summary of levels of plasmalogen and diacyl species shows a general downward trend in plasmalogen levels in sepsis, excluding livers from of septic rats ( Figure 7 ). In particular, this summary highlights the many differences in changes elicited during sepsis between plasmalogen and diacyl phospholipid levels depending on the tissue and particular phospholipid class. One of the more striking observations is the loss of pPE in plasma in contrast to increases in diacyl PE during sepsis in both humans and rats.

The mechanisms responsible for plasma pLPC and pPE loss during sepsis are not known, but several mechanisms seem likely. One mechanism is that loss of plasmalogen is due to oxidative stress during sepsis. We have previously shown plasma 2-chlorofatty acid levels are elevated in human sepsis (Meyer et al., 2017; Amunugama et al., 2021b). Furthermore, in this rat sepsis model there are increased levels of 2-chlorofatty acid levels (Pike et al., 2020), which is derived from plasmalogens (Albert et al., 2001; Thukkani et al., 2002; Amunugama et al., 2021b). During sepsis the tissue plasmalogen pool or the specific plasmalogen molecular species targeted by HOCl has not been determined. In this respect it could be speculated that the impressive loss of plasma pLPC, which is overall a small pool of the total plasmalogen, could be responsible for the nanomolar levels of 2-chlorofatty acid observed during sepsis. It is also possible that the loss of plasmalogens is due to the activation of phospholipases. It has been suggested that phospholipase A₂-mediated release of arachidonic acid from plasmalogens is important in the production of oxylipids in COVID-19 (Schwarz et al., 2021; Snider et al., 2021). The phospholipase A₂ mechanisms may be directly responsible for pPE loss. It is also possible pLPC loss is due to either accelerated use as an acceptor by acyltransferases leading to conserved levels of pPC despite putative oxidative loss or tissue uptake during sepsis. Another possibility is pPE and pLPC decrease as a result of reduced release from the liver and vascular endothelium. In human sepsis, HDL-cholesterol decreases (Vavrova et al., 2016; Tanaka et al., 2019), which may also be due to decreased secretion from the liver. Decreased plasma plasmalogens and increased liver plasmalogens during sepsis are similar to plasmalogen changes in H-Lrpprc mice, a mouse model of the monogenic form of the mitochondrial disease, Leigh syndrome (Ruiz et al., 2019). In H-Lrpprc mice, hepatic Far1 and Agps are also elevated suggesting decreased plasma plasmalogen levels mediate a feedback system to increase liver plasmalogen biosynthesis. Such a feedback system may also be responsible for elevated liver plasmalogen levels in livers during sepsis. It will be interesting in future studies to examine Agps and Far1 as well as differences in the levels of the plasmalogen precursors, alkyl ether lipids, in the livers from septic and control rats.

The possibility that pLPC is a circulating precursor to enrich plasmalogens in endothelium is intriguing. Plasmalogen enhancement in isolated cell studies protects cells from oxidative stress (Zoeller et al., 1999). Additionally, several studies have investigated plasmalogen precursors as a potential treatment in inflammatory diseases (Bozelli and Epand, 2021; Paul et al., 2021). Enhancing plasmalogen levels is difficult since dietary consumption of plasmalogens is reduced due to the acidic environment of the gastrointestinal tract. Using acid-stable precursors such as alkyl ether lipids will raise plasmalogen levels over time following desaturation of the alkyl ether bond to the vinyl ether. However, under acute conditions such as sepsis, the conversion of an alkyl ether to plasmalogens likely will be very slow. On the other hand, circulating pLPC already has the vinyl ether bond and lysolipids are rapidly incorporated into cells. It will be important in the future to determine the source of circulating pLPC under physiological conditions as well as during sepsis. It could be envisaged that pLPC is a product of lipoprotein-associated pPC hydrolysis by either secretory phospholipase A₂ or lipoprotein lipase. During sepsis pLPC levels potentially are dependent on a combination of oxidation of pPC or pLPC and pPC hydrolysis. Finally, the role of pLPC during sepsis needs to be further considered as a biomarker of outcomes. Similarly, the role of other plasmalogens, as well as the relationship of plasma plasmalogen levels with changes in plasma 2-chlorofatty acid levels, need to be considered as outcome predictors. The relationship of plasmalogen and chlorinated lipid levels may also allow distinction of changes in these lipids with greater specificity to infection compared to other disease states associated with only decreased plasma plasmalogen levels with the exception of lupus (Dragonas et al., 2009; Fabelo et al., 2011; Ifuku et al., 2012; Mahieu et al., 2014; Hu et al., 2016; Paul et al., 2019; Su et al., 2019).

The studies herein show plasmalogen loss during sepsis. However, there are several limitations to these studies. In the human studies we analyzed differences between septic humans and healthy control humans. Our healthy cohort average age was thirty-eight while the sepsis group was sixty. To overcome this limitation, we selected the oldest individuals (n = 7) in the healthy group and assessed differences in this control subset compared to the larger group of septic subjects (Figures 1D–F, 2; Table 1). These additional analyses indicated plasma pLPC and pPE levels were reduced in the sepsis cohort when compared to this age-aligned control subgroup. Another limitation is that we have no data on the sex of individuals in our healthy cohort, while our sepsis cohort was comprised of 40% females. Our rat studies focused on changes occurring only in male rats and 20 h following cecal slurry injection. Thus, comparisons of rat specimens to human specimens were collected at different times and sex differences were not a parameter in the rat studies. It should also be appreciated that plasma levels of plasmalogens were considerably different in healthy controls due to the inherent differences in plasmalogen levels in man versus rat. Nevertheless, both human and rat sepsis led to decreases in plasma plasmalogen levels, and the rat studies afforded the opportunity to investigate changes in plasmalogen levels in the liver and kidney during sepsis. There were also limitations to the SARS-CoV-2 infection studies when comparisons are made to the rat and human sepsis studies. The SARS-CoV-2 infection studies were a viral infection elicited by airway inoculation to transgenic mice expressing the hACE2 receptor in all epithelial cells. Humans do not express ACE2 in all epithelial cells. Furthermore, these studies were performed only in female mice due to availability of genotyped mice for this study. Future studies are needed to consider sex as a parameter in both SARS-CoV-2 infected mice and rat cecal slurry sepsis. Compared to the unknown time for human sepsis beginning and the known time for CS injection, mouse infections with SARS-CoV-2 leading to pulmonary inflammation require time for viral replication to elicit injury which is typically 3–5 days. While our human and rat sepsis studies involved systemic infection, SARS-CoV-2 infection of K18 mice initially was primarily localized to infection of the respiratory tree. Infection led to robust increases in the expression of pro-inflammatory cytokines. The loss of plasmalogen in the lung during SARS-CoV-2 infection likely is the result of oxidative stress. We did not observe a loss in DPPC in the lung of infected mice. The chemical makeup of plasmalogens compared to DPPC provides a contrast in susceptibility to oxidative stress. The plasmalogen vinyl ether bond is a target for oxidation while the saturated fatty acids of DPPC are very stable under oxidative stress. Similar to findings with severe COVID-19 patients (Schwarz et al., 2021; Snider et al., 2021) we detected decreases in plasma plasmalogens in infected K18 mice.

This is the first demonstration of the loss of plasmalogens at a molecular species level in human sepsis. Furthermore, we show pLPC loss in both human and rodent sepsis. It is possible that plasma pLPC is a critical lipid to maintain endothelial plasmalogen levels under oxidative stress associated with sepsis. The demonstration of plasmalogen loss during SARS-CoV-2 further highlights the nature of plasmalogen loss during oxidative stress associated with infectious disease. The role of plasmalogens as biomarkers of outcomes in sepsis and COVID-19 need to be explored as well as the potential protective role of plasmalogens during infectious disease.