Bridging Lipid Metabolism and Innate Host Defense

See allHide authors and affiliations

Science Translational Medicine  15 Oct 2014:
Vol. 6, Issue 258, pp. 258fs41
DOI: 10.1126/scitranslmed.3010501


Host- and pathogen-derived lipids share clearance mechanisms that pinpoint PCSK9 as a plausible target for the treatment of some patients with sepsis (Walley et al., this issue).

Throughout history, infection has been an inescapable threat. Epidemics such as the bubonic plague in the 14th century or the Spanish flu of 1918 killed tens of millions of people and altered the course of European politics. Our genes, however, reveal a deeper evolutionary history of the impact of infectious disease. Adaptive polymorphisms in the hemoglobin gene emerged in areas where malaria is endemic, and the risk of death from infection is much more strongly linked to genetic factors than the risk of death from cancer or heart disease (1). An intimate and dynamic interaction between the human and microbial worlds has driven our evolution and shaped our biology. In this issue of Science Translational Medicine, Walley and co-workers provide a fascinating window into the interface between host lipid biology and the innate immune response to bacterial pathogens (2).


The intense interplay between human and microbial biology—orchestrated over a time frame of more than 1 billion years—has favored host-pathogen responses that are symbiotic. Mitochondria are the evolutionary descendents of protobacteria that parasitized early precursors of eukaryotic cells. Viral genes that promote viral persistence have become integrated into the mammalian genome and encode apoptosis inhibitor proteins and interleukin-10. But the complexity of interactions between the host and the microorganism is probably best exemplified by lipopolysaccharide (LPS), also known as endotoxin, present in the cell wall of Gram-negative bacteria. Exposure to LPS has promoted the evolution of recognition and response mechanisms that resemble an endocrine response and involve carrier proteins such as LBP, a dedicated receptor complex that comprises Toll-like receptor 4 (TLR4), CD14, and MD2 as well as several endogenous LPS antagonists such as bactericidal permeability increasing protein (BPI) and alkaline phosphatase. As is now becoming understood, mechanisms for LPS clearance overlap with mechanisms that support endogenous lipid clearance (3).

In plasma, free LPS exists either as aggregates, in bacterial membrane fragments, or in loose association with proteins such as albumin; bound LPS exists in complexes with lipoproteins. Whereas free LPS is taken up primarily by hepatic Kupffer cells, bound LPS is cleared by hepatocytes. Until now, the mechanisms of hepatocyte-mediated LPS clearance have been unclear.

In the new work (2), the authors uncover a critical role for the proprotein convertase family member proprotein convertase subtilisin/kexin type 9 (PCSK9). Proprotein convertases facilitate the transfer of cell surface receptors into the cell for degradation in lysosomes. PCSK9 in the blood binds to low-density lipoprotein receptors (LDLRs) on the surface of hepatocytes and enhances their removal. A reduction in PCSK9 activity supports sustained expression and recycling of the LDLR and thus increases the clearance of LDLR-binding lipids from the blood. Agents that reduce serum cholesterol levels by inhibiting PCSK9 are being studied as potential treatments for heart disease.

Sepsis is a common and complex condition that results from disseminated activation of an innate host immune response to microbial infection. Impaired vital organ function is a cardinal manifestation of sepsis; in its most severe form, septic shock, tissue perfusion is threatened. The presence of endotoxin in the blood is common in sepsis; conversely, the infusion of LPS replicates many of the clinical and biochemical features of sepsis, including septic shock. Reasoning that lipids from bacterial cell walls might also be bound by plasma carrier proteins and cleared by hepatic excretion, Walley et al. hypothesized that altered PCSK9 function could modulate the host innate immune response to pathogen-derived LPSs and improve clinical outcomes of patients with sepsis (Fig. 1).

Fig. 1 LPS exploits the host’s cholesterol clearance pathway.

(A) Circulating PCSK9 protein binds to the epidermal growth factor–like repeat A (EGF-A) domain of LDLR. Once within the liver cell, the PCSK9 protein directs the LDLR to the lysosome for degradation. PCSK9 secreted from the Golgi apparatus can also bind the LDLR before it reaches the cell surface, leading to lysosomal degradation. Walley et al. demonstrated that LDLR also binds LPS, which is directed by PCSK9 for lysosomal degradation. When PCSK9 binds to an LDLR, the receptor is destroyed along with the LDL or LPS particle. (B) But if PCSK9 does not bind, the receptor can recycle to the cell surface and remove more cholesterol or LPS. Mutations or therapeutic strategies that target PCSK9 and inhibit the interaction between PCSK9 and LDLR result in increased endogenous and circulating LDLRs that can remove more LDL-C or LPS from the bloodstream and incorporate it into cells (in the case of cholesterol) or degrade it in lysosomes (LPS). [Adapted from (11)]


The authors provided support for this hypothesis in a tour de force of translational medicine. First they showed that deletion of the PCSK9 gene from mice resulted in reduced blood levels of LPS, a reduced cytokine response to systemic endotoxin challenge, and attenuation of characteristic physiological manifestations such as temperature dysregulation, hypotension, and impaired ventricular contractility. Conversely, addition of PCSK9 protein to human hepatocyte cultures caused increased degradation of LDLR, inhibiting LPS clearance. Having established preliminary proof-of-concept in a highly reductionist model, the authors proceeded to test the consequence of PCSK9 inhibition in a more relevant animal model—mice with a complex intra-abdominal bacterial infection resulting from cecal ligation and puncture (CLP)—in which antibiotic intervention was delayed 6 hours after the CLP insult. Inhibition of PCSK9 with a neutralizing antibody reduced plasma endotoxin levels, blunted the cytokine response, and doubled survival rates. Moreover, antibody-mediated inhibition of PCSK9 decreased the response to LPS in uninfected wild-type mice but had little effect in LDLR-deficient mice, suggesting that PCSK9 affects LPS clearance in an LDLR-dependent manner.

The extent to which biological responses observed in relatively simplistic murine models reflect human biology is controversial (4, 5). What is not controversial, however, is that none of the many targets that have shown promise in mice have been translated into effective human therapies for sepsis, despite more than one hundred phase 2 and 3 clinical trials costing $10 billion or more (6). A striking deficiency in the experimental portfolio that has preceded such trials has been work that established biological activity of a therapeutic agent in humans with sepsis.

The Vancouver group (2) addresses this deficiency in several ways. First, they stratify subjects from two separate cohorts of patients with sepsis on the basis of the presence of either loss-of-function or gain-of-function polymorphisms in the PCSK9 gene. In both populations, loss-of-function polymorphisms are associated with better survival and reduced cytokine levels, while gain-of-function polymorphisms had the opposite effect. And in a separate cohort of healthy volunteers who received intravenous endotoxin, the authors found that the loss-of-function polymorphisms were associated with reduced cytokine levels. Last, they show that patients who are homozygous for a minor allele of the LDLR gene that inhibits binding of the LDLR by PCSK9 were unaffected by the PCSK9 genotype. In aggregate, the findings—deriving from complementary and elegant studies in both mice and people—both provide new insight into the regulation of LPS clearance and point to a target that is amenable to therapeutic manipulation in sepsis patients.

The new work makes a compelling biological case that PCSK9 contributes to adverse outcome in some patients with sepsis. But it would be simplistic to conclude that PCSK9 is a promising target in sepsis—not because it might not play a role but rather because the dominant clinical research model of sepsis is inadequate to identify who may or may not benefit from intervention or who might even be harmed by such a treatment. Indeed, the potential for harm is greatest for those patients with active infection, because the innate immune response—the mechanism of illness in sepsis—is also critical to the successful clearance of infection.

For more than 30 years, studies of immunomodulatory therapy in sepsis have recruited patients who manifested a pattern of physiological derangements—fever (or hypothermia), tachycardia, and tachypnea—occurring in association with suspected or confirmed infection and new onset of the dysfunction of one or more organ systems. This constellation of abnormalities has been termed “sepsis syndrome;” but the assumption that these symptoms delineate a homogeneous patient population is demonstrably wrong (6).

Patients with clinical sepsis vary with respect to the site, bacteriology, and even presence of infection; the presence of co-morbidities; and the extent and nature of underlying organ dysfunction. Striking differences are also evident in the levels of circulating inflammatory mediators (7). Moreover, the consequences of manipulating a particular therapeutic target can vary with both the intervention and the insult. For example, preclinical studies in which TNF is neutralized show benefit in models of endotoxemia or systemic Gram-negative infection but harm when the challenge organism is Streptococcus pneumoniae or an intracellular pathogen such as Listeria or Mycobacterium tuberculosis (8); neutralization of LPS also has been shown to increase the mortality of patients with Gram-positive infection (9).

A conceptual shift is needed if we are to translate promising findings from studies of biology into effective treatments for patients with complex life-threatening disorders such as sepsis. Increasing attention is being turned toward the approach used in oncology. Oncologists long ago abandoned the idea that there is a cure for cancer, replacing it with the more nuanced concept that there are many effective treatments for specific cancers but that their use must be guided by pathological staging systems and the use of biomarkers. This approach has made adjuvant chemotherapy a reality and has given us trastuzumab (a monoclonal antibody that targets the Her2/neu receptor and prolongs survival for women with breast cancers that overexpress the Her2/Neu receptor) and the tyrosine kinase inhibitor imatinib (which is effective in patients with chronic myelogenous leukemia that expresses the activated BCR/Abl tyrosine kinase) (10).

So what should the next steps be in evaluating inhibition of PCSK9 in patients with sepsis? The findings of Walley et al. suggest a plausible way forward. First, because PCSK9 loss-of-function mutations are associated with improved outcome, initial clinical trials should only recruit patients with the PCSK9 gain-of-function variant rs505151. Second, because the proposed benefit of PCSK9 inhibition arises from a reduction in pathogen lipid levels, the intervention should be used only in patients in whom these levels are elevated—rapid assay of endotoxin could readily define such a patient. Finally, because the intervention is complex, unproven, and likely costly, its use should be restricted to patients whose risk of death—readily quantified using a severity of illness measure—is substantial. This population would include some patients with sepsis syndrome but would also likely include patients who fail to meet those criteria, having diagnoses such as severe acute pancreatitis or a ruptured abdominal aortic aneurysm.

Translational research in sepsis has been ill served by overly simplistic assumptions regarding observations in preclinical models and overly optimistic estimates regarding the number of patients who might benefit from specific therapies. The estimate that there are 750,000 cases of severe sepsis in the United States each year has suggested to the pharmaceutical industry that an effective therapy could generate more than $1 billion in annual sales; thus the industry has increased the pressure to move rapidly from preclinical studies to phase 3 trials to maximize the commercial life of the drug. This approach has been uniformly unsuccessful. New translational strategies are needed. The systematic approach used by Walley and colleagues (2) is a welcome step forward in that process, but one that needs to be matched by equally thoughtful and data-informed approaches to clinical evaluation.


  1. Competing interests: J.C.M. is a member of the Steering Committee for the EUPHRATES Trial (Spectral Diagnostics), which is evaluating a technology to remove LPS in sepsis, and of the Data Safety Monitoring Board of a study of recombinant thrombomodulin in sepsis (Asahi Kasei Pharma). C.d.S. declares that she has no competing interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article