Emerging therapeutic targets for sepsis

ABSTRACT

Introduction: Sepsis is characterized by a dysregulated host response to infection. Sepsis-associated morbidity/mortality demands concerted research efforts toward therapeutic interventions which are reliable, broadly effective, and etiologically based. More intensive and extensive investigations on alterations in cellular signaling pathways, gene targeting as a means of modifying the characteristic hyper and/or hypo-immune responses, prevention through optimization of the microbiome, and the molecular pathways underlying the septic immune response could improve outcomes.] Areas covered: The authors discuss key experimental mammalian models and clinical trials. They provide an evaluation of evolving therapeutics in sepsis and how they have built upon past and current treatments. Relevant literature was derived from a PubMed search spanning 1987–2020.

Expert opinion: Given the complex nature of sepsis and the elicited immune response, it is not surprising that a single cure-all therapeutic intervention, which is capable of effectively and reliably improving patient outcomes has failed to emerge. Innovative approaches seek to address not only the disease process but modify underlying patient factors. A true improvement in sepsis-associated morbidity/mortality will require a combination of unique therapeutic modalities.

KEYWORDS:

• Checkpoint proteins
• epigenetics
• immunosuppression
• inflammation
• microbiome
• sepsis

1. Introduction to sepsis

Sepsis is defined as a ‘dysregulated host response to infection’ which results in ‘life-threatening organ dysfunction’ [1]. Around the world and across the age spectrum, sepsis remains a significant source of morbidity and mortality. It is estimated that nearly twenty million people develop sepsis globally each year with 1.7 million affected in the United States alone [2,3]. Sepsis contributes to one third of in-hospital deaths in the United States and global mortality is estimated to exceed 25% [2,3]. This burden of disease brings with it a significant economic cost due to extended hospital admissions requiring intensive and invasive medical care. It has had a disproportionate impact on healthcare utilization in the United States, costing more than 24 billion USD despite accounting for fewer than 4% of all admissions [4].

However, while the definition of sepsis has evolved in recent decades, the treatment approach has remained much the same. Early recognition and diagnosis are critical. Local source control may be achieved through drainage or debridement of the affected organ system when possible along with directed-antimicrobials when the microbial nature of the source is known, while the use of broad-spectrum antimicrobials acts to reduce the systemic infectious burden. Supportive measures in the form of intravenous fluids, vasopressors, and mechanical ventilation are used in response to septic shock in an attempt to prevent circulatory and respiratory collapse until the infection is adequately controlled.

1.1. Early goal-directed therapy and steroids in sepsis

In the past, an early goal-directed therapeutic (EGDT) approach was used to guide resuscitation based on static indicators of fluid status including central venous pressure and mean arterial pressure [5]. However, after three major trials – PROCESS, ProMISE and ARISE – failed to show an improvement in patient outcomes with this approach, Surviving Sepsis guidelines now advocate for early aggressive resuscitation with further titration based on fluid challenges as well as dynamic markers such as passive leg raises and arterial pulse pressure variation [6–10]. In fact, a meta-analysis including all randomized controlled trials evaluating EGDT for patients with severe sepsis or septic shock found that early lactate clearance carried a survival benefit whereas EGDT did not [11].

The use of corticosteroids in the management of sepsis has been a point of debate for decades. Their proposed benefit stemming not only from their function as a systemic immunosuppressant but from their ability to correct a relative adrenal insufficiency, which is thought to be induced by critical illness [12]. Currently, corticosteroids are only used in cases of refractory shock, defined as hypotension that requires multiple vasopressor agents despite adequate fluid resuscitation, and this remains controversial, deemed appropriate on a case-by-case basis using provider judgment and preference. Despite multiple well-powered studies and subsequent meta-analyses, while there is evidence that glucocorticoids may result in a faster resolution of shock, fewer ventilator days and a shorter length of stay in both the hospital and the intensive care unit, this does not seem to result in a reliable improvement in patient outcomes including short- or long-term mortality (Table 1) [13–18]. The notion that there is an improvement in surrogate markers like fewer ventilator days but not mortality suggests a specific patient population that would most benefit from steroids. More recently, the concept of using the combination of broad anti-oxidant/anti-inflammatory agent, high-dose Vitamin C, along with thiamine and hydrocortisone has been proposed as a potential treatment for patients in septic shock [19]. However, while having some modest beneficial effects overall, it appears to be no better than hydrocortisone alone, which as mentioned above has not been universally accepted [20].

Table 1. Overview of all included studies organized by therapeutic target of interest

Author (Citation number)		Title	Year	Target	Model	Intervention	Outcomes	Review Section
Supportive Care	Sprung et al. (13)	Hydrocortisone Therapy for Patients with Septic Shock	2008	Adrenal insufficiency	Human RCT	Corticosteroids (IV plus taper)	No difference in 28-day mortality based on corticotropin responsiveness or overall With steroids, faster resolution of shock, higher rate of superinfection vs placebo	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Fisher et al. (19)	Mechanisms of attenuation of abdominal sepsis induced acute lung injury by ascorbic acid	2012	Immunomodulation	Murine Fecal intraperitoneal injection	Vitamin C (IV infusion)	- Reduced pro-inflammatory response, maintenance of epithelial junctions, decreased alveolar permeability, and decreased coagulopathy	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Yealy et al. and the ProCESS Investigators (6)	A Randomized Trial of Protocol-Based Care for Early Septic Shock	2014	Fluid Management	Human Randomized control trial (RCT)	Protocol-based resuscitation (EGDT vs standard) vs usual care	- No difference in 90-day or 1-year mortality or need for organ support	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Peake et al. and the ARISE Investigators and the ANZICS Clinical Trials Group (7)	Goal-Directed Resuscitation for Patients with Early Septic Shock	2014	Fluid Management	Human RCT	EGDT vs usual care	- No difference in in-hospital or 90-day mortality, hospital length of stay, or need for organ support	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Mouncey, et al. and the ProMISe Investigators (8)	Trial of Early, Goal-Directed Resuscitation for Septic Shock	2015	Fluid Management	Human Pragmatic RCT	EGDT vs usual care	No difference in 90-day mortality With EGDT, longer ICU length of stay, longer duration of pressor use, longer duration of organ support	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Zhang et al. (11)	Early goal-directed therapy in the management of severe sepsis or septic shock in adults: a meta-analysis of randomized controlled trials	2015	Fluid Management	Human Meta-analysis	EGDT (standard vs modified) vs usual care	No difference in mortality (EGDT vs usual care), hospital LOS, ICU LOS, pressor use or organ support With standard EGDT, lower mortality (RR 0.84) vs modified EGDT With early lactate clearance group, reduced mortality (RR 0.66) vs EGDT	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Famous et al. (63)	Acute Respiratory Distress Syndrome Subphenotypes Respond Differently to Randomized Fluid Management Strategy	2017	NA	Human (ARDS by subphenotypes) Secondary analysis of two RCTs	Liberal vs conservative fluid resuscitation strategies	- Greater reduction in mortality with conservative strategy for those in the hyperinflammatory subphenotype	3.1.2 Targeting components of the innate immune response
	Venkatesh et al. for ADRENAL Trial Investigators and ANZICS Clinical Trials Group (14)	Adjunctive Glucocorticoid Therapy in Patients with Septic Shock	2018	Adrenal insufficiency	Human (mechanically ventilated) RCT	Corticosteroids (continuous infusion without taper)	No difference in 90-day mortality, hospital LOS or ICU LOS With steroids, faster resolution of shock and shorter duration of initial mechanical ventilation	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Annane et al. for the CRICS- TRIGGERSEP Network (15)	Hydrocortisone plus Fludrocortisone for Adults with Septic Shock	2018	Adrenal insufficiency	Human RCT	Corticosteroid + mineralocorticoid (IV without taper)	- Lower 90-day mortality (RR 0.88), duration of pressor use and duration of orgain failure	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Fang et al. (17)	Association of Corticosteroid Treatment With Outcomes in Adult Patients With Sepsis: A Systematic Review and Meta-analysis	2018	Adrenal insufficiency	Human Systematic review and meta- analysis	Corticosteroids	- Lower mortality rates (28-day (RR 0.9), ICU (RR 0.85) and in-hospital (RR 0.88)), faster resolution of shock with shorter pressor duration	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Rochwerg et al. (18)	Corticosteroids in Sepsis: An Updated Systematic Review and Meta-Analysis	2018	Adrenal insufficiency	Human Meta-analysis	Corticosteroids	- No statistical difference in mortality rates (28 days to 1 year) or LOS (ICU or hospital), faster resolution of shock, and decreased severity of organ dysfunction	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
	Fujii et al. for the VITAMINS Trial Investigators (20)	Effect of Vitamin C, Hydrocortisone, and Thiamine vs Hydrocortisone Alone on Time Alive and Free of Vasopressor Support Among Patients With Septic Shock	2020	Adrenal insufficiency, Immunomodulation	Human RCT	Corticosteroids (IV without taper), vitamin C (IV) and thiamine vs corticosteroids	- No difference in 7-day survival or pressor duration, or 90-day mortality	1.1 Early Goal-Directed Therapy and Steroids in Sepsis
Direct Pathogen Targeting	Ziegler et al. (30)	Treatment of Gram-Negative Bacteremia and Septic Shock with HA-1A Human Monoclonal Antibody against Endotoxin — A Randomized, Double-Blind, Placebo- Controlled Trial	1991	Lipopolysaccharide	Human RCT	HA-1A (LPS mAb)	Reduced mortality by 39 percent in gram-negative bacteremia No benefit for overall sepsis cohort	2.1 Targeting of LPS and TNF-a
	McCloskey et al. for CHESS Trial Study Group (31)	Treatment of septic shock with human monoclonal antibody HA-1A. A randomized, double-blind, placebo- controlled trial	1994	Lipopolysaccharide	Human RCT	HA-1A (LPS mAb)	No difference in 14-day mortality in gram-negative bacteremia No benefit for overall sepsis cohort	2.1 Targeting of LPS and TNF-a
	Beigel et al. (24)	Remdesivir for the Treatment of Covid-19 — Final Report	2020	SARS-CoV-2	Human (hospitalized with lower respiratory tract infection) RCT	Remdesivir	- Faster time to recovery (RR 1.29) and 15-day clinical improvement (OR 1.5)	1.2 The Heterogeneity of Sepsis
	Hansen et al. (25)	Studies in humanized mice and convalescent humans yield a SARS-CoV-2 antibody cocktail	2020	SARS-CoV-2	Murine and human	Neutralizing antibodies	Generation of potent and diverse antibodies using humanized mice which mimic convalescent humans Antibody cocktail for SARS-CoV-2 treatment pending results of ongoing human	1.2 The Heterogeneity of Sepsis
	Tracey et al. (29)	Anti-cachectin/TNF monoclonal antibodies prevent septic shock during lethal bacteraemia	1987	TNF	Primate E.coli infusion	Anti-TNF monoclonal antibody (mAb), pre- treatment (1 vs 2 hr before)	Prevention of shock in 1 hour group Prevention of shock, organ dysfunction, and death in 2 hour group	2.1 Targeting of LPS and TNF-a
	Fong et al. (28)	Antibodies to cachectin/tumor necrosis factor reduce interleukin 1 beta and interleukin 6 appearance during lethal bacteremia	1989	TNF	Primate, E.coli infusion	Anti-TNF monoclonal antibody (mAb), pre-treatment	No increase in IL-1 beta or IL-6	2.1 Targeting of LPS and TNF-a
	Abraham et al. for TNF-alpha MAb Sepsis Study Group (32)	Efficacy and safety of monoclonal antibody to human tumor necrosis factor alpha in patients with sepsis syndrome. A randomized, controlled, double-blind, multicenter clinical trial.	1995	TNF-alpha	Human RCT	TNF-alpha mAb	No difference in 28-day mortality for overall cohort Improved survival at 3 days for those in septic shock	2.1 Targeting of LPS and TNF-a
	Fisher et al. and the Soluble TNF Receptor Sepsis Study Group (33)	Treatment of septic shock with the tumor necrosis factor receptor:Fc fusion protein	1996	TNF-alpha	Human RCT	TNF-alpha receptor:Fc fusion protein	No improvement in 28-day mortality Increasing dose associated with increased mortality	2.1 Targeting of LPS and TNF-a
	Bernard et al. for PROWESS study group (34)	Efficacy and safety of recombinant human activated protein C for severe sepsis	2001	Activated protein C	Human RCT	Drotrecogin alfa (recombinant human activated protein C)	- Lower 28-day mortality (RR 0.81), increased bleeding complications	2.2 Activated Protein C Pathway

1.2. The Heterogeneity of Sepsis

While sepsis is, by definition, the result of a severe systemic infectious insult, in practice, it is an umbrella term used to describe a clinical syndrome, which represents the end stage of a heterogenous set of pathological processes. Sepsis stemming from pneumonia will clearly necessitate a different treatment approach with regards to the ability to achieve source control when compared with an intra-abdominal or soft tissue infection where the offending source has the potential to be physically removed. In addition, prior research has demonstrated that the responsible microbe can trigger a different underlying pathophysiology and may in fact impact patient outcomes. An extreme example of this is COVID-19 driven pneumonia/sepsis in comparison to other pandemic respiratory viruses like influenza A virus, Middle East respiratory syndrome coronavirus and severe acute respiratory syndrome coronavirus-1 [21]. A report on the efficacy of temporally-guided administration of the antiviral agent remdesivir has demonstrated a significantly shorter time to recovery and a reduction in disease severity as well as a trend toward reduced mortality by mitigating the viral pneumonia/infection and its organ-specific sequalae [22–25]. Another example, those with gram-negative bacterial sepsis were found to have significantly higher cytokine levels – tumor necrosis factor-alpha (TNF-α), IL-8, IFN-γ, IL-1, IL-4, and IL-10 – and a higher incidence of septic shock than those with a gram-positive source [26,27]. This difference in the severity and profile of the resulting cytokine storm suggests that gram-negative versus gram-positive sepsis may benefit from different immunomodulating treatment approaches. In light of this, it is understandable that, given the one-size-fits-all treatment used in sepsis currently, the mortality rate would remain high. However, in order to target patients with appropriate treatments, novel diagnostics will be required.

2. Revisiting past therapeutic trials

2.1. Targeting of LPS and TNF-α

Past research has attempted to primarily modulate the innate/pro-inflammatory immune response directly in an effort to lessen the subsequent dysfunction and ultimately improve patient outcomes. Based on a number of promising animal sepsis models which characterized factors involved in the mammalian septic immune response, human trials were initiated more than twenty years ago targeting lipopolysaccharide (LPS) and TNF-α as possible drivers of septic morbidity and mortality (Table 1) [28,29]. While LPS is a unique component of the cell wall found in gram-negative bacteria and known to drive the acute inflammatory response, TNF-α is a potent pro-inflammatory cytokine, the concentration of which has been found to be inversely related to patient outcomes in sepsis.

An initial exploratory trial using a human monoclonal antibody known as HA-1A, designed to target a LPS-specific binding domain, demonstrated a promising mortality benefit for septic patients with gram-negative bacteremia (Table 1) [30]. However, use of HA-1A within the context of a pragmatic trial did not demonstrate any improvement in patient outcomes [31]. Similarly, two large-scale randomized, controlled clinical trials targeting TNF-α – one through use of a monoclonal antibody and the other through a fusion protein directed at its unique receptor – failed to show a survival benefit in those with both gram-positive and gram-negative sepsis [32,33].

2.2. Activated protein C pathway

Drotrecogin alfa is a recombinant form of human activated protein C (APC) and was FDA-approved in 2001 for use in sepsis after the PROWESS study demonstrated a mortality benefit [34]. However, further study including the follow up PROWESS-SHOCK trial failed to confirm these findings even in the highest risk septic patients and the drug was ultimately removed from the market in 2011(Table 1) [35].

Renewed exploration of the therapeutic utility of this pathway has focused on thrombomodulin, an endogenous anticoagulant protein that is an important cofactor in the activation of protein C. Commonly expressed on endothelial cells, it is often downregulated in certain disease states including disseminated intravascular coagulation (DIC) and sepsis. Recombinant human soluble thrombomodulin (rhTM) has already been successfully used in the treatment of DIC, with phase III clinical trials showing superiority over heparin in terms of resolution rates and safety profile (Table 1) [36].

A spectrum of coagulopathies has been found in patients with sepsis, ranging from mild derangements to DIC, and are believed to be fundamental to the development of end-organ dysfunction and ultimately mortality. Yamakawa et al. found that in a small group of patients with sepsis-induced DIC administration of rhTM resulted in significant improvement in organ dysfunction as marked by Sequential Organ Failure Assessment (SOFA) scores and a significant reduction in 28-day mortality versus controls [37]. A subsequent study by Kato et al. focused on septic patients with mild versus severe coagulopathy and, once again, found an improvement in SOFA scores, especially the respiratory component, and 90-day mortality amongst those in the severe group [38]. However, the SCARLET trial, which was a randomized controlled trial involving over 800 patients with sepsis-associated coagulopathy, failed to show a benefit in terms of 28-day mortality (Table 1) [39]. Despite this, a recent study by Yoshihiro et al. focused on sepsis patients with severe respiratory failure rather than a coagulopathy and found a reduction in both ICU and hospital mortality rates in comparison to those who did not receive rhTM [40].

The use of the APC pathway as a therapeutic target reinforces two important lessons: the cumulative immune response is complex and a given pathway should not be rejected due to the failure of one mechanistic approach; and an individual’s septic phenotype (e.g. DIC, respiratory failure) plays a significant role in the efficacy of a therapeutic target.

2.3. Lessons learned and the path forward

Following these failed attempts to integrate animal and human therapeutic targets in sepsis, disappointment bred multiple theories that attempted to explain why the survival benefit seen in experimental models failed to carry over into human patients. Animal models come with innate limitations given their use of young animals, typical of an inbred strain, raised in highly-controlled living conditions, as well as the experimental difficulties associated with recreating a realistic timeline of human sepsis [41]. For example, while mice will frequently be dosed with the therapeutic agent of choice shortly after initiation of sepsis, humans are likely to present in a delayed fashion with a more advanced immune cascade underway, resulting in a less efficacious intervention at that particular timepoint of septic presentation.

Along these lines, it is also argued that sepsis in humans is a disease of ‘complexity’, resulting from patient history, experiences, developmental status and co-morbidities, which further challenges animal modeling [42]. However, cancer, which is also considered a complex condition that represents a constellation of different diseases and pathologies, which many felt could not be appropriately modeled in experimental animals, has clinically benefitted substantially from therapeutic approaches derived from reductionist animal modeling including checkpoint proteins, designer T-cells, to name a few [43–45].

Thus, while we need to better delineate the nature of the ‘complexity’ of sepsis in humans and in animals. Animal models will still be needed if we are ever to translate new knowledge of unique human septic phenotypes into a mechanistic understanding of the pathological processes under-pinning them, which in turn should drive discovery of novel or better diagnostics and therapeutics for this condition [46,47]. Importantly, this will require not only continued development, modification, and application of the animal models used to emulate them, but standardization of such models to better match the select septic human endo- or phenotypes to which we want to apply these novel diagnostics and therapeutics [48–50].

3. Emerging therapeutics

3.1. Host response

After encountering an infectious agent, the innate immune system promptly initiates a wide-reaching signaling cascade which attempts to stimulate a variety of immune cells while also simultaneously keeping their actions in check. As has been previously discussed in detail in the review by Chun et al., the innate and adaptive immune responses are dependent on the particular infectious agent – where it resides and how it interacts with the host body, for example – and their efficacy is dependent on interactions between numerous cell types and signaling proteins [51]. This robust cascade response provides a plethora of promising therapeutic targets in combating sepsis.

3.1.1. PD-1/PD-L1 and the role of checkpoint proteins

While the initial immune response in sepsis is robust and broadly pro-inflammatory, if left unchecked, an uncontrolled hyperinflammatory response becomes overwhelming and induces its own organ dysfunction and mortality [52]. As a result, a compensatory anti-inflammatory response occurs simultaneously, utilizing multiple signaling pathways including checkpoint proteins like programmed cell death-ligand 1 (PD-L1) and its receptor, PD-1 [53]. Though PD-1 and PD-L1 are both expressed on a range of immune cells, PD-L1 is also expressed on nonimmune cells and organ tissues, representing an essential link between the inflammatory response and end-organ damage. In an effort to restore the immune response to homeostasis, both function through a coinhibitory mechanism to encourage immune cell anergy in the periphery and to shift the overall cytokine profile in an anti-inflammatory direction [54].

Murine models have demonstrated a survival benefit amongst both PD-1 and PD-L1 knockout mice, and human studies evaluating septic patients have shown that non-survivors as well as those who developed a secondary infection had significantly higher rates of both PD-1 and PD-L1 expression on blood leukocytes (Table 1) [54–56]. Following development of a human monoclonal antibody targeting PD-L1, a recent trial was conducted to assess for the efficacy and safety of this antibody for septic patients. This phase 1b, randomized, placebo-controlled trial used escalating doses of anti-PD-L1 antibody in comparison to placebo in 24 patients with sepsis-associated immunosuppression to evaluate for mortality, adverse effects, and markers of immune recovery (Table 1) [57]. While there was no significant change in mortality or rates of adverse events between groups, increased expression of HLA-DR – a marker of restored immune function – was associated with higher doses of the antibody. Larger trials are needed to more thoroughly study the role of the PD-1/PD-L1 pathway in immune dysregulation, a hallmark feature of the pathophysiology of sepsis.

Overall, checkpoint proteins play an essential role in transitioning from a hyper- to hypo-inflammatory response, making them an ideal target to reduce secondary infections and overall mortality. However, it remains important to develop the means to easily and cost-effectively identify an individual’s inflammatory response to determine if this intervention is appropriate for their case. In addition, trials of these medications in cancer benefit from treatment based upon expression, a tactic that would be beneficial in sepsis.

3.1.2. Targeting components of the innate immune response

As previously described, a variety of cellular pathways are activated simultaneously following an infectious insult. This redundancy has led some to theorize that more than one component may need to be targeted to effectively reduce immune dysfunction and improve outcomes in sepsis. While cytokines were previously categorized as being either pro- or anti-inflammatory in nature, it is now clear that their function is far more complex, affected by their location, the surrounding milieu of immune/nonimmune cells and other cytokines. However, given their central role in regulating the immune response and the fact that an inappropriate expression can result in organ and tissue damage, they remain a popular target for therapeutics [58].

As was discussed previously, despite promising murine models of TNF-α neutralization, human studies in sepsis failed to show a clinical benefit. However, given the complexity of the signaling network in which they function, more recent studies have focused on the utility of targeting two cytokines simultaneously. In Berghe et al., they were able to demonstrate that the concurrent use of an IL-1β receptor antagonist and anti-IL-18 antibody in three lethal models of murine sepsis resulted in a significant survival benefit and was protective in the case of TNF or LPS administration [59].

Similarly, the combination of immune-stimulatory cytokines, like IL-7, IL-15, or interferon-gamma, and antibodies that antagonize checkpoint protein ligation have been proposed for the treatment of sepsis [60]. An example of this is compassionate use of combined anti-PD-L1 and interferon-gamma to treat a patient with fungal sepsis following trauma who had failed all conventional therapy as was described by Grimaldi et al (Table 1) [61]. Given that this patient had evidence of immunosuppression affecting both the innate and adaptive pathways, the combination therapy was administered in an effort to address the full spectrum of dysfunction rather than just a single element, and to respond to the specific septic phenotype this patient was exhibiting.

Activation of complement and toll-like receptors (TLR) by pathogen-associated molecular patterns including LPS occurs following infection, representing potent components of the innate immune response. Given that they work in parallel, but through distinct pathways, Huber-Lang et al. hypothesized that dual blockade of C5a and CD14 would, therefore, improve survival in murine models of polymicrobial sepsis. In comparison to sham and single blockade groups, the use of both agents resulted in a less robust cytokine response as well as improvement in both morbidity and mortality following CLP [62].

The complicated network of cell signaling in sepsis provides a rich source of potential therapeutic targets. However, as early trials have demonstrated, elimination of a single element will likely be inadequate due to the redundancy built into this multifaceted system. As the last few murine studies demonstrate, improved efficacy may result from concurrent blockade of either more than one aspect of the signaling cascade or more than one of its pathways. However, to do this it will be critical to know what septic patients stand to benefit from such therapies, not only from a temporal perspective, but based on the expressions of markers that are perceived to be mechanistically effected by the given novel treatments being applied and not simply provided to individuals based on broad sepsis defined symptoms as they exist. The potential value of such septic patient sub-phenotyping has recently been documented by Famous et al. and Dahmer et al. where identification of specific inflammatory phenotypes in acute respiratory distress syndrome (ARDS) dictated fluid management and where genetic variants in IL-1 affected clinical outcomes including mortality, respectively (Table 1) [63,64]. To add to this complexity, Ottinger et al demonstrated that when compared to young patients, geriatric patients display a distinctly dampened cytokine response to critical illness and infection [65]. Further, if a geriatric patient exhibited an inflammatory response akin to younger patients, then mortality was markedly increased compared with the typical inflammatory response.

In summary, while the innate immune system is rich in potential therapeutic targets and remains a promising area of research, as has been demonstrated by the study of cytokines in this setting, the difficulty in transitioning to effective interventions lies in the fact that their actions are not uniform and can be influenced by where they are and what point in the inflammatory timeline they are expressed. Additionally, the built-in redundancy within these signaling cascades will require concurrent targeting of more than one element to induce a meaningful alteration in the inflammatory response which will change clinical outcomes. The ultimate goal in this approach will be to allow for the immune system to attack the pathogen but resolve the inflammation prior to host cellular damage.

3.2. Microbiome

Targeted molecular therapeutics geared toward modification of the gut microbiome can be classified as a vital area for continued study and consideration, as recent work has highlighted the beneficial role of host flora in modifying the immune response during sepsis. In both experimental and clinical sepsis, the prevalence of certain bacterial populations has been found to change following the infectious insult and appear to impact host response. Interestingly, these populations seem to be conserved across species, illustrating the importance of this microbial response and how it may benefit the host during periods of immune dysregulation.

Current research includes use of the microbiome as a phenotyping tool to more easily identify sepsis, identification of gut flora alterations resulting from supplementation with pre- and probiotics, impact of fecal transplantation for dysbiosis, and use of the lung as an endpoint or diagnostic tool in assessing sepsis.

Sepsis results in periods of hyperinflammation and suppression. As currently understood, the characteristic hypo-inflammatory phase of clinical sepsis in latter stages leads to exhaustion of the effector response of the immune system [66]. This then allows for immune suppression and susceptibility to later infections. As septic patients typically undergo antibiotic treatment, a considerable impact on the gut microbial communities is induced throughout treatment. Potential contributions from opportunistic pathogens surviving antibiotic use in critical care is mentioned as a significant challenge in providing adequate treatment [67]. Further, not only do specific microbes induce distinct immune cell phenotypes, it has been proposed that probiotic therapies may become more targeted and selective based on the specific immune dysfunction identified within certain patients [68].

Understanding the role and importance of the gut during critical illness has been a driving factor in the field. Others have mentioned gut disruptions as a commonly associated occurrence in critically ill individuals, and have subsequently linked this to mortality increases [69]. Sepsis deaths have mostly been related to drug-resistant infections in the later stages of hospitalization [67]. T_H17 cells, which travel through the walls of the intestines, were suggested as a contributor to immunosuppression partly due to their interactions with gut microbes. These cells are also capable of ramping up the immune response. Either of these inflammatory profiles pertaining to sepsis resolution would benefit from increased biomarker evaluation, and this dictates a considerable portion of recent work.

3.2.1. Immunophenotype identification and biomarker potential of the gut

In effort to determine whether septic patient outcomes were linked to the presence of particular groups of microorganisms, studies have sought to identify the potential link between disease prognosis and specific bacterial communities (Table 1) [70]. For example, peri-rectal swabs of intensive care unit patients were sequenced and identified elevated levels of Bilophila, Fusobacterium, and Parabacteroides – all of which have previously been linked to inflammation. Moreover, septic patients who succumbed to their disease had increased prevalence of these disease-linked communities than those who survived. Agudelo-Ochoa et al. posit that important species like these may function as biomarkers for sepsis progression, allowing for categorization of patient gut flora.

Identifying the source of an infection is an important step in sepsis management. However, inaccuracy may lead to ineffective treatment and profound patient outcomes. Despite the fact that bacteremia can result from a multiplicity of sources, Tamburini et al. sought out a more precise means of identification to provide an improvement over the current available methods which can, at times, be based on supposition (Table 1) [71]. Through novel bioinformatics approaches, Tamburini et al. utilized the tool StrainSifter to help identify the origin of bloodstream infections. Stem cell transplant patients presenting with bloodstream levels of both Klebsiella pneumoniae and Eschericia coli microbes also exhibited increased intestinal levels of these communities. This data provided a basis for the assertion that the gut may be a reservoir of both infectious populations.

Despite the apparent risk that colonization by gut microbes present in septic patients, it has been shown that commensal organisms and their activity offer key protections during disease progression. These normal members of the microbiota are part of a natural symbiosis with the individual in which they reside, even helping to increase secretion of certain molecules relevant in the immune response to infection. After induction of sepsis, Wilmore et al. found that mice gavaged with proteobacteria were able to upregulate immunoglobulin A (IgA) in their serum, conferring protection against polymicrobial sepsis [72]. Furthermore, in a cohousing study which evaluated mortality rates following sepsis in mice ordered from different sources, these groups displayed differences in their gut microbial communities which is thought to have contributed to their differing mortality rates [73]. This conclusion was supported by the fact that, upon cohousing of these two groups, survival rates following sepsis improved when compared with non-cohoused subjects. Furthermore, despite vendor source, mortality rates among cohoused subjects were similar. While this does not represent a means of sepsis treatment, the use of the microbiome as a diagnostic tool through identification of immunophenotype might significantly benefit both clinicians and patients to guide treatment and improve outcomes. Importantly, it also represents an aspect of complex, nascent gut microbial diversity, that may need to be considered on a more general basis when considering/maximizing the actions/efficacies of novel therapeutics mentioned earlier, which are largely initially delineated/vetted in inbred animals that lack diverse microbial histories based on their housing and other issues [74,75].

3.2.2. Probiotics, synbiotics and sepsis relief

Given that the microbiome is a modifiable system, offering a variety of benefits to the living system in which it resides, diet supplementation through pre- and probiotics could aid in the expansion of key gut community members that may impact the host as they respond to certain disease processes including sepsis. Probiotics are microbes that do not contribute to disease but whose presence benefits the host [76]. Prebiotics refers to indigestible fibers that spur the function and growth of specific colonic microbial communities. Investigators have worked to demonstrate the potential effectiveness of probiotic treatment in septic mice, again through the use of the murine CLP technique to model sepsis [77]. Four weeks before induction, mice were pretreated with a salt solution as a control, or, were given the probiotic Lactobacillus rhamnosus GG (LGG). Benefits of probiotic treatment prior to experimental sepsis included reduced mortality as well as a decrease in inflammatory cytokines IL-2 and IL-22 when compared to saline treated septic mice, as assessed through serum ELISA. However, results for probiotic-treated mice were still significantly higher than sham controls. Treated mice displayed improved diversity and richness of their gut flora. In addition, these mice featured improvements in tight junctions between colonic epithelial cells-similar to sham surgery results-which were conversely diminished in saline treated septic mice. Probiotic treated subjects also exhibited levels of colonic epithelial apoptosis and cellular proliferation within the colon more similar to sham subjects, as improved over saline controls.

Beyond these animal results, a recent clinical trial was conducted to understand the outcome of Lactobacillus and Bifidobacteria probiotic supplementation in colorectal cancer patients not receiving antibiotics following surgical intervention (Table 1) [78]. Investigators noted a significant decrease in many cytokines more traditionally associated with inflammation like TNF-α, IL-6, and IL-22 in the probiotic-treated compared to baseline. This result supports probiotic use following surgery to prevent some of the inflammatory conditions present during the septic immune response, as in intra-abdominal sepsis following a colorectal procedure.

Not all probiotic-focused therapies for sepsis have been entirely successful, with several trials producing mixed results. A randomized clinical trial was conducted in premature infants presenting with late onset sepsis, where probiotics were given to assess their effectiveness in treatment (Table 1) [79]. The probiotic consisted of Bifidobacterium infantis and Bifidobacterium lactis along with Streptococcus thermophiles. A statistically significant difference between placebo enrollees and probiotic treated was found for Bell Stage 2 necrotizing enterocolitis. However, there was no difference between mortality or late-onset sepsis incidence.

In another example of microbiome remodeling via probiotics, results of a clinical trial exhibited improvements in preterm neonatal deficiencies, but did not improve sepsis rates, incidence of NEC or the frequency of local infections (Table 1) [80]. The probiotic Lactobacillus reuteri was utilized in a randomized clinical trial to examine if it could help improve on preterm-associated deficiencies like growth, but also on its ability to help obviate infection. While this treatment improved subjects’ ability to feed, grow and shortened the duration of hospitalization, neither infection frequency, sepsis rates nor the development of NEC were significantly different.

Similarly mixed results have been found in research on septic adult patients as well. Early in sepsis, a patient’s gut microbiome is typically less diverse and contains higher levels of pathogen associated molecular patterns (PAMPs) like peptidoglycans and endotoxin, which are proposed to serve as proxy for microbial translocation [81]. In a randomized clinical trial, subjects were offered a probiotic featuring Lactobacillus and Bifidobacterium species, along with Enterococcus feacium, that increased microbiota diversity. As Stadlbauer et al. argue the importance of the gut in the development of sepsis, it may be through solving issues related to dysbiosis and diversity that headway can be made. During the study, certain PAMPs like peptidoglycans were brought close to healthy control levels. However, endotoxin remained significantly elevated, though there was an eventual trend toward decreasing levels. The origin or source of this endotoxin could not be established. Despite these results, interestingly, markers for gut permeability did not change in septic patients.

In mechanically ventilated septic patients, a study was conducted in which the effectiveness of synbiotics were gauged in their ability to modify the gut microbiome (Table 1) [82]. Synbiotics are the combination of both pre- and probiotics [76]. Benefits to the patient including the attenuation of septic complications were assessed and linked to the intervention with synbiotics. Shimizu et al. argued that the supportive treatment of sepsis, typically provided (i.e. antibiotics administration, use of inotropic medications and blood transfusions) may have a significant negative impact on the gut microbiome which could be assuaged with synbiotic use. In this study, synbiotic treatment consisted of both Lactobacillus casei and Bifidobacterium breve, along with galacto-oligosaccharides. While there was no change in rates of bacteremia or mortality, there were significantly fewer instances of enteritis and ventilator-associated pneumonia with synbiotic treatment.

To summarize, many of the inherent draw-backs in probiotic use described here include the issues with the translational value of pretreatments in animal models. However, the results detailed from probiotic clinical trials were favorable and are considered more impactful than those derived from modeling. Given these mixed results overall, further research must be done to determine the ideal role for pro-, pre- or synbiotic administration to generate the most successful outcomes in septic patients.

3.2.3. Fecal microbiota transplantation in sepsis resolution

Fecal microbiota transplant (FMT) has been utilized in rodent experimental models of sepsis, as well as in clinical settings to treat or determine the potential to correct dysbiosis and confer host protection to infection. This method has proven to be effective in treating patients with Clostridium difficile infections, where the introduction of donor material allowed for the repopulation of healthy gut flora in sick patients (Table 1) [83]. A similar approach is taken in sepsis FMT studies whereby resolution of dysbiosis or upregulation of the immune response are achieved through the introduction of microbial members whose metabolic activity benefits the host. Through interactions with T_H17 cells, which were mentioned previously as having the ability to travel through the gut lining, microbes from FMT in septic individuals may interact and influence periods of immune suppression or hyperinflammation [65].

In rat models of sepsis utilizing LPS to induce endotoxemia, seven-day old rats were pretreated daily for 15 days with a variety of pro- and prebiotics to assess the potential benefits of such supplementation [84]. Either L. casei or L. rhamnosus administration best provided protection to the host, as illustrated by inflammatory and oxidative stress markers. In another experiment, rats were pretreated for 15 days with either of these microbes. Following euthanasia, fecal contents were diluted and administered to 21-day old rats, which were then subjected to either LPS or Zymosan to induce sterile inflammation- a model of SIRS. Results for FMT probiotic treatment of endotoxemia compared to LPS alone included reduction in oxidative damage markers including protein carbonyl levels (except in L. rhamnosus probiotic FMT), nitrate-nitrate and thiobarbituric acid reactive substance (TBARS) levels, as well as a reduction in inflammatory markers including IL-1β, TNF-α, IL-6 and MPO. Following sterile inflammation, the same markers of oxidative stress and inflammation were reduced when compared to zymosan alone in both probiotic FMT treatments, as well as in control FMT. Authors argue that a healthy gut microbiota-not the effects of colonization specifically by Lactobacillus-may have produced such results. This work sheds additional light on FMT methods, displaying its potential as an emerging treatment option for septic patients, particularly in pediatric cases, in addition to its established role in treating gut dysbiosis.

Another experiment utilized stool from a patient who succumbed to late-onset sepsis which was found to contain multi-drug resistant (MDR) Klebsiella oxytoca, MDR Serratia marcescens, and tetracycline-resistant Enterococcus faecalis as well as the fungal species Candida albicans [85]. Using these four pathogenic members of the patient stool sample, mice were inoculated via the gut after undergoing a partial hepatectomy and being deprived of food prior to surgery, in an effort to better mimic the selective pressures on these microbes found in clinical settings where surgical trauma and intervention have occurred. Additionally, mice were given antibiotics. FMT utilizing cecal material from healthy littermates was used to determine its impact on host susceptibility to sepsis and mortality. This produced a > 70% survival benefit in live vs. either autoclaved transplant material or no FMT intervention – illustrating the necessity of live microbes in conferring protection. Additional results included lowering the systemic burden of these pathogenic community members following live FMT treatment. Blood S. marcescens and liver/spleen C. albicans levels were significantly lower than autoclaved FMT treatments. In an intraperitoneal sepsis model utilizing the same 4 pathogens for inoculation, Kim et al. found a significant decrease in levels of cecal butyrate, a short-chain fatty acid byproduct of microbial metabolism, which when produced by beneficial microbes, plays a role in ridding the system of pathogenic members. However, after sepsis induction and subsequent FMT, butyrate levels were brought back to those seen in un-inoculated subjects. This result further emphasizes the importance of specific host-microbial community members and the protective effects of their activity within the gut. By utilizing patient-derived strains, Kim et al. provided a fascinating result in mice that would ideally translate to the clinic through similar methods of FMT.

In a rat experimental sepsis model utilizing LPS administration, Li et al. studied the impact of FMT on sepsis-associated encephalopathy [86]. Results again illustrated the interplay between host microbial communities and the immune response. Similar to results from others, they found that opportunistic Proteobacteria decreased, and there was an increase in commensal Firmicutes in septic rats who had received an FMT vs those who did not. FMT after sepsis induction also led to decreased levels of pro-inflammatory cytokines TNF-α, IL-1β, and IL-6 in the hippocampus, when compared to non-FMT mice. However, the concept of ‘pro’ and ‘anti’ inflammatory roles for cytokines is understandably no longer an acceptable way of defining their activity. In addition to its use in experimental sepsis models, FMT has also been utilized clinically to treat individuals with severe sepsis. In two patients who presented with severe diarrhea and multi-organ dysfunction syndrome (MODS), FMT was utilized as an effective means of treatment for dysbiosis of the intestinal microbiota, as identified through sequencing (Table 1) [87]. Healthy donor pathogen-free FMT was utilized to correct this and resulted in reduction of opportunistic members of the phylum Proteobacteria with an increase in commensal members of phylum Firmicutes. Diminishing levels of pro-inflammatory marker IL-6 as well as several other key diagnostic endpoints for inflammation were identified. Clinically, patient body temperatures were brought back to within normal range, and bowel habits normalized as well.

The results of FMT-based microbiome modifications indicate that this treatment might play as a broader role in the treatment of sepsis in the future, potentially helping to improve patient outcomes and resolve dysbiosis.

3.2.4. Gut microbial diversity and its effect on the lung

A common co-related disease endpoint of sepsis that often acts as an indicator of poor patient outcome and death is indirect acute lung injury or ARDS [88]. As a result of the dysregulated immune response during sepsis, MODS, particularly of the lungs, remains a challenge to effective treatment by clinicians. In a study comparing the lung microbiota of sepsis-induced ARDS patients who had recently undergone major abdominal surgery to non-septic controls who underwent esophageal resection, blood and BAL fluid were taken at various time points to measure both community diversity via sequencing and culture assays and to determine the impact on clinical outcomes (Table 1) [89]. Results indicated that both reduced time spent on a ventilator and in the ICU for ARDS subjects correlated with high alpha diversity. Dysbiosis was evident in ARDS compared to control data, where lung alpha diversity was found to be significantly lower. The most frequently occurring sequence variant, (read-outs which allow for microbial identification), in ARDS samples made up a significantly greater proportion of reads than the most frequent variant seen in controls. This is described by authors as dominance, and such an outcome further illustrates this state of dysbiosis.

BAL fluid from both ARDS patients and experimental sepsis models has shown an increase in gut-associated microbes within the lung through the use of sequencing (Table 1) [90]. At the time of this publication, there had not been confirmation via culture of bacterial translocation from gut to lung, necessitating additional or alternative means of analysis. Sequencing results dictated that Proteobacteria were enriched in ARDS patients. Additionally, there was an abundance of a Bacteroides strain in ARDS-patient BAL fluid, sharing perfect sequence alignment with several gut-associated microbes. Before experimental induction of sepsis, a related and abundant strain in the lower GI of murine subjects was detected. The same population was found to be most abundant in the lungs post sepsis, giving authors an indication of origin or source. There was a positive correlation between patient serum TNF-α and the level of Bacteroides, but not with levels of alveolar TNF-α. A negative correlation existed between alveolar TNF-α and phylum Bacteroidetes, but phylum Proteobacteria exhibited positive correlation for the same pro-inflammatory marker in ARDS patients. It is important to define or resolve the interactions taking place between host and pathogen, characterizing these changes within the microenvironment they occur. As shown in a murine infection model utilizing several unique strains of Streptococcus pneumoniae, both host and microbial gene expression were assessed in various organs [91]. Among several tissues assayed were both the lung and nasopharynx. Interestingly, S. pneumoniae within the nasopharynx does not normally cause injury to or damage host tissue. Through RNA-sequencing and principal component analysis, clustering of bacterial datapoints helped identify profiles of these strains specific to roles in either host colonization, illustrated by defined nasopharyngeal clustering away from other organs, or pathogenesis, occurring within the other tissues surveyed. This result outlines the importance of understanding host-microbial environmental factors as they pertain to infection and resolution. Organ specific contributions to pathology, such as those mentioned here, should be considered in the context of mechanistic sepsis therapeutics.

As detailed by these changes in community frequencies within the lung microbiome of septic and ARDS patients, options for resolution may necessitate the use of targeted therapies, as in FMT and probiotic treatment, to correct a variety of associated dysbiotic events and disease-related immunophenotypes. As additional studies surveying the lung microbiome in septic and/or ARDS patients are completed, emphasis can be placed on identifying patterns in microbial communities present during disease. Recognizing conserved changes within the microbial flora of the lung during sepsis illustrates its role as a future biomarker in sepsis diagnosis. Authors characterize a potential link between these organs during the progression of sepsis and ARDS, which they posit as a possible indicator of related disease mechanisms. Results identified potential instances of gut microbial translocation to the lung. As improvements are made in next generation sequencing techniques, the timely identification of a conserved septic/ARDS patient microbial phenotype unique to these diseases may offer improvements in clinical diagnosis and treatment.

3.3. Epigenetics in sepsis relief

As the technical aspects associated with next generation sequencing have improved and become more cost effective, these read-outs have been more heavily relied on in patient studies. This is especially the case in sepsis, where epigenetic modifications and the immune response have been further characterized as a result of these improvements. Utilizing a murine model of acute lung injury (ALI)-induced sepsis (ALI-sepsis), Bomsztyk et al. were interested in determining the importance of transcriptional regulation within organs implicated in sepsis progression and multi organ failure (e.g., the liver, kidney and lung) [92]. They specifically focused on vascular endothelial growth factor (VEGFr/VEGF) and Tie2/angiopoietin (Tie2/Ang), factors associated with endothelial growth and development. Pairing the experimental ALI-sepsis model and with multiplex chromatin immunoprecipitation platform (Matrix ChIP) methods, it was determined that at these genes the density of RNA polymerase II molecules was reduced, a change that was most evident in the lung. This result corresponded with reduced expression of endothelial factor-related genes across each organ surveyed, as assessed via reverse transcription-polymerase chain reaction (RT-PCR). As they also measured transcription permissive and repressive histone modifications following experimental modeling, it was found that in each organ permissive H3 lysine acetylation markers decreased at these genes. Minor changes in repressive marks H4K20m3, H3K9m2, H3K9m3, and H3K27m3 were detected in these organs, with the lung being the only tissue source in which there was a reduction in transcription permissive marks H3K4m3 and H3Km2. Bomsztyk et al. argued that these epigenetic changes occur systemically as opposed to being confined to an isolated organ like the lungs.

In other work meant to further elucidate the relationship between genetic modifications made in the septic individual, Davenport et al. utilized leukocytes from peripheral blood of patients with pneumonia-derived sepsis and symptoms of organ dysfunction to survey gene expression (Table 1) [93]. Cluster analysis of expression resulted in two defined groups, and between them, there was no difference in the expression of IL-1β, TNF-α or IL-6 genes related to pro-inflammatory responses. Through pathway analysis of genes differentially expressed, they identified variances between groups in key processes like cytotoxicity, apoptosis, and the activation of T-cells. The group found to have higher mortality also demonstrated elevated expression of proteins like TOLLIP and IRAK3 which act to down-regulate toll-like receptor (TLR) signaling. In the same group, human leukocyte antigen (HLA) class II-related genes displayed reduced expression and a similar pattern was evident for a variety of other genes associated with the activation of T-cells.

A collection of seven different genes, which predicted group classification within this cohort, were used to classify individuals of another cohort also presenting with pneumonia-derived sepsis. Similarly, higher mortality was found in the same group as the previous cohort. Pathway analysis for both cohorts revealed a functional enhancement of genes associated with T-cell exhaustion, among other key processes. Use of genetic variant markers conferring alterations in gene expression were identified, impacting regulation of sepsis-associated genes like IL18RAP, NLRC5 and PADI4. Davenport et al. argue that understanding the unique pattern of gene expression in septic individuals may improve selection of clinical trial enrollees and provide more targeted pathophysiological data to guide treatment.

In the whole blood of septic patients, elevated expression of the gene AQP5 was found to be associated with sepsis-related mortality (Table 1) [94]. Non-survivors exhibited increased CpG methylation at a site within the gene’s promoter region. DNA methylation is most closely associated with gene suppression. The transcription factor NF-κB, Rump et al. posit, may act to inhibit expression of AQP5 and, indeed, they confirmed in another experiment that NF-κB binds to this promoter site. They believed that the increased methylation seen in fatal cases and the inability of NF-κB to bind and inhibit expression of AQP5 resulted in reduced expression in these individuals. This understanding may provide another therapeutic target, which could influence patient outcomes during septic challenge.

In another study, the methods of ChIP-seq were again utilized to assess DNA modifications of CD14⁺⁺CD16⁻ monocytes in two septic patients versus four healthy individuals (Table 1) [95]. While patients varied in terms of the etiology of their sepsis, there were similarities in the patterns of histone modifications for both. Investigators were able to identify genes displaying differences in promoter histone marks. H3K9Ac and H3K4me3 are active marks, which were downregulated in a group of genes from septic patients. These genes also exhibited an increase in the inactive H3K27me3 mark, indicating that these genes are likely turned off as a result of a septic challenge. In another collection of genes displaying a reduction in H3K27me3 and an increase in active marks, the indication is that these are active genes. Through gene ontology term analysis, investigators found immune response genes were overrepresented in both collections of silenced and active genes. Weiterer et al. posit that the intricate nature of the septic immune response and the uncertain nature in which the associated genes are categorized as being either beneficial or detrimental to disease progression as possible reasons for such findings [95]. Such studies will require much larger cohort assessments to clearly establish such correlations, let alone to begin to speak to causal relationships.

Earlier work speaking to potential directions for translational sepsis epigenetics work was completed on the activity of sirtuin 1 (SIRT1). Septic patient leukocytes in a model of endotoxin tolerance displayed the deacetylase activity of SIRT1 with co-factor NAD+ to ultimately inhibit transcriptional regulation by NFκB [96]. This occurs at the promoter regions of TNF-⍺ and IL-1β, and alters their chromatin state from being responsive to TLR-4 signaling to silent. In other work, the hypoinflammatory phase of obese septic (late stage) mice was found to be directed by SIRT2 [97]. Through SIRT2, but interestingly not SIRT1 inhibition, subject mortality was benefited. Appropriate cellular activation and inflammation could occur following SIRT2 inhibition. While the goal of achieving a mechanistic therapy has yet to be accomplished, work such as this provides valuable insight on potential options.

4. Conclusion

Sepsis remains a considerable challenge to global public health based on its broad impact and the significant financial burden it imposes on both patients, providers, and society in general. Sepsis presents acute complications that must be addressed promptly to ensure survival. However, the potential for long-term and chronic effects create additional problems for a patient long after stabilization occurs. Current treatment options include aggressive resuscitation, antimicrobials, and, at times, the administration of steroids to support the patient during this insult. However, the heterogeneous nature of the disease etiology and multitude of systems affected during a septic response presents significant obstacles to adequate treatment. Despite the previous failures of clinical trials, ongoing research efforts continue to produce interesting results in several areas detailed here that may bare therapeutic fruit.

A broad collection of findings has been presented, including results from patients with sepsis as well as mammalian experimental models of sepsis, from which the most promising therapeutic leads may be gleaned. It is thought that through an improved understanding of the intracellular signaling pathways involved in the septic immune response as dictated by circulating cytokines and other key effectors that we can significantly improve the capabilities of clinicians. Targeting these pathways and tailoring the immune response such that infection is resolved without the effects of cytokine release syndrome, for example, represents the most desirable outcome of current research efforts. Immune signaling networks have importantly been shown to be impacted by resident gastrointestinal flora.

As the gut microbiome displays evidence of dysbiosis upon sepsis, recent efforts have focused on understanding how it may be leveraged as a septic diagnostic tool through immunophenotyping. Alterations made to the microbiome through fecal transplant or the use of pro- and prebiotics have been shown to alleviate certain characteristic effects of immune dysregulation. While the microbiota provides a viable option for therapeutic intervention, other methods have broadened the focus of such efforts. Next generation sequencing techniques have become more practical with time, and this has allowed for a better understanding of gene expression in septic individuals. Work has focused on identifying changes in gene expression following sepsis and understanding how patterns of epigenetic modifications might predict outcome and better dictate care. These efforts may in time provide clinicians with therapeutic gene targets for the treatment of septic patients.

5. Expert opinion

Recent work in sepsis therapeutics has relied on insights gleaned from altering the host response to infection. The immune dysregulation characteristic to sepsis, and the heterogeneity by which it is clinically exhibited create challenges in formulating an appropriate cure-all. A focus of preclinical research in sepsis has been resolution of translatable drug targets found in murine septic models. With regard to the landscape of mechanistic approaches in correcting the aberrant immune response during sepsis, work done on the deacetylase activity of SIRT1 and 2 remains ripe with potential for legitimate therapeutic leads. The epigenetic modifications that occur during sepsis related to immune dysregulation characterized through the use of next generation sequencing techniques, is currently providing vital information in our understanding of expression changes and their impact on patient outcomes.

Patient characteristics have been used to further delineate and categorize the data generated from these methods, revealing patterns in expression seemingly linked to patient survival and other cellular mediated processes associated with sepsis. These and other modifications, as in DNA methylation pertaining to gene suppression, have been identified in specific septic patient genes and linked to mortality. Information such as this is providing the necessary background for gene-focused therapeutics in sepsis treatment. These techniques have more recently revealed common molecular pathways resultant from sepsis, which not only provide novel molecular targets, but also a conserved pattern of gene expression which may be used for diagnostic purposes.

The goal of these efforts is ultimately to resolve an effective treatment for the varying sources of the septic immune response. A therapeutic such as this might offer added benefits in the line of ARDS and multiple organ failure, as these are commonly associated co-morbidities, particularly in traumatic injury. Indirectly achieved would be an alleviation of financial burden on patients, paying attention to the considerably high global incidence of sepsis. Unfortunately, the many challenges facing us in achieving these goals are significant in the context of evaluating translatability of pre-clinical data. Mammalian systems have provided an acceptable means of asking biological questions, but we seem to lack the correct understanding of how to appropriately apply the answers. Many of these obstacles have proven to be substantial, as evidenced by prior clinical trial failures. The issues associated with antimicrobial resistance present an added hurdle to sepsis treatment, and must also be kept in context.

The gut and lung microbiome remain particularly interesting research areas, as characterization of resident populations during sepsis pathology has implicated contributions from specific members. The use of probiotics in animal models nicely illustrates how the influences from resident communities can be modified to benefit the host septic response. Other work in humans has focused on preventing mortality from the aberrant immune response to infection. Importantly, probiotic clinical trial data have given reason to consider these supplements concurrent to any sepsis treatment a vital option.

We believe the field will be guided by improved targeting of multiple approaches- concurrently and away- from the broadly prescribed and often ineffective one-size fits all treatments presently available. The unique approaches had taken to address the current deficiencies in sepsis therapy represent promising pathways to future resolution. Presumably, over the next few years, therapeutics will likely take greater account of specific patient factors, more heavily based on molecular and phenotypic characterization for the customized treatment of those individuals most likely to respond. The culminated results from such thorough and impactful work have transformed expectations and provided realistic endpoints for broadly beneficial but well-defined options in clinical sepsis management.

Article highlights

• Sepsis is a heterogeneous process and clinical outcomes are affected by a number of factors including patient genetics underlying key components of the immune system, the individual microbiome profile present at baseline, the location/type of infectious insult, and the nature of the initial host response.

• Animal models of sepsis will continue to play an essential role in identifying new therapeutic targets; however, their ability to mirror human septic presentations must be improved.

• Sepsis clearly induces epigenetic changes which affect crucial aspects of the immune response, indicating that future research should focus on reversal of these changes as a potential therapeutic target and the impact that may have on septic outcomes.

• Improvement of clinical outcomes will depend on development of efficient and cost-effective methods of identifying individual patient sepsis phenotypes at the bedside through characterization of cytokine profiles, leukocyte protein expression patterns, and use of sequencing technology to identify genetic variants and alterations.

• The use of pre-, pro- or synbiotics, and fecal transplantation could modify the individual microbiome to affect clinical outcomes related to sepsis through primary prevention, severity reduction, and prevention of secondary infections like pneumonia.

• Successful therapeutic interventions will not be reliant on a single element of the individual immune response but will require a multi-pronged approach to modify the host environment, identify the individual inflammatory phenotype, and rebalance the host response to infection.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose

Additional information

Funding

The research of the authors was supported by NIH T32-GM065085 (E.W.T), NIH T32-HL134625 (B.E.A.), NIH P20-GM103652 (S.F.M.), NIH K08-GM110495 (DSH) as well as NIH R35-GM118097 (A.A.).