Biochemical transformation of bacterial lipopolysaccharides by acyloxyacyl hydrolase reduces host injury and promotes recovery Received for publication, August 4, 2020, and in revised form, October 22, 2020 Published, Papers in Press, October 26, 2020, DOI 10

Subject:BiologyPrice: Bought3

Share With

Biochemical transformation of bacterial lipopolysaccharides by acyloxyacyl hydrolase reduces host injury and promotes recovery Received for publication, August 4, 2020, and in revised form, October 22, 2020 Published, Papers in Press, October 26, 2020, DOI 10.1074/jbc.REV120.015254

Robert S. Munford1,*, Jerrold P. Weiss2, and Mingfang Lu3,* From the 1Laboratory of Clinical Immunology and Microbiology, NIAID, National Institutes of Health, Bethesda, Maryland, USA, the 2Inflammation Program, University of Iowa, Iowa City, Iowa, USA, the 3Department of Immunology and Key Laboratory of Medical Molecular Virology (MOE/NHC/CAMS), School of Basic Medical Sciences, Fudan University, Shanghai, China

Edited by Chris Whitfield

Animals can sense the presence of microbes in their tissues and mobilize their own defenses by recognizing and responding to conserved microbial structures (often called microbe-associ- ated molecular patterns (MAMPs)). Successful host defenses may kill the invaders, yet the host animal may fail to restore homeostasis if the stimulatory microbial structures are not silenced. Although mice have many mechanisms for limiting their responses to lipopolysaccharide (LPS), a major Gram-neg- ative bacterial MAMP, a highly conserved host lipase is required to extinguish LPS sensing in tissues and restore homeostasis. We review recent progress in understanding how this enzyme, acyloxyacyl hydrolase (AOAH), transforms LPS from stimulus to inhibitor, reduces tissue injury and death from infection, pre- vents prolonged post-infection immunosuppression, and keeps stimulatory LPS from entering the bloodstream. We also discuss how AOAH may increase sensitivity to pulmonary allergens. Better appreciation of how host enzymes modify LPS and other MAMPs may help prevent tissue injury and hasten recovery from infection.

Animals can recognize and respond to bacteria that invade their tissues by sensing the invaders’ MAMPs. Pro- and anti- inflammatory cytokines and other mediators are produced, neutrophils and other leukocytes are recruited to the infected site, blood flow to the tissue increases, and increased vascular endothelial permeability allows influx of antibodies, comple- ment, and other molecules from the bloodstream into the tis- sue. Both extracellular (e.g. complement and antibodies) and in- tracellular mechanisms contribute to bacterial killing. The inflammatory response may also produce tissue injury and lead to harmful immunosuppression. Although these features of the antibacterial host response

are well-understood, much less is known about how the inflam- matory process resolves, tissue injury and prolonged immuno- suppression are avoided, and normal MAMP responsiveness is restored. Recent studies have shown that lipid mediators can help resolve inflammation following trauma (1, 2), but the stud- ies reviewed here provide evidence that, if infection is the incit- ing event, it is also necessary to eliminate or inhibit the stimula- tory MAMP(s). Insight into this aspect of recovery from

MAMP-induced inflammation has come from studying how mice silence lipopolysaccharide (LPS, endotoxin), a potent Gram-negative bacterial MAMP that has long been implicated in the pathogenesis of infection-induced inflammation, sepsis, and chronic diseases (3, 4). As Lewis Thomas wrote in The Lives of a Cell (4), “The

Gram-negative bacteria . . . display lipopolysaccharide mole- cules in their walls, and these macromolecules are read by our tissues as the very worst of bad news . . . . There is nothing intrinsically poisonous about endotoxin, but it must look awful, or feel awful, when sensed by cells.” In fact, many host mole- cules can prevent cells from sensing LPS: anti-LPS antibodies, proteins that bind and sequester LPS (bactericidal permeabil- ity-increasing protein (5), lactoferrin, cathelecidin, and plasma lipoproteins), an intracellular phosphatase (6), and others (7). Remarkably, none of these molecules—alone or together—can completely silence LPS in mice, and bacteria do not usually destroy their own molecules, especially large macromolecules (8). Instead, LPS silencing is accomplished by a highly con- served host lipase, acyloxyacyl hydrolase (AOAH). Named for its chemical sites of action (9), AOAH removes from LPS the fatty acyl chains that are essential for its immunostimulatory activity and transforms this potent agonist into an effective LPS antagonist. Here we review how LPS is sensed by animal cells, the impor-

tance of LPS acylation in determining the agonistic potency of LPS, the structure and enzymatic properties of AOAH, and evi- dence that AOAH plays significant roles in both moderating and terminating host inflammatory responses to LPS and Gram-negative bacteria. Recent findings have pointed to im- portant roles for AOAH in limiting inflammatory responses to Gram-negative bacteria, restoring normal immune responsive- ness after LPS exposure, preventing stimulatory LPS from entering the bloodstream, and reducing the agonistic activity of intestinal LPS. We then describe human genetic evidence that AOAH may be an “essential” gene and studies that have identi- fied a unique trans (cross-chromosomal) regulatory mecha- nism and potential associations with colitis and asthma.

Sensing LPS

The LPS produced by most Gram-negative bacteria has a polysaccharide chain of variable length that is anchored to the

*For correspondence: Robert S. Munford, munfordrs@niaid.nih.gov; Min- gfang Lu, mingfanglu@fudan.edu.cn.

17842 J. Biol. Chem. (2020) 295(51) 17842–17851 Published in the U. S. A.

REVIEWS

This is an Open Access article under the CC BY license.

 

mailto:munfordrs@niaid.nih.gov

mailto:mingfanglu@fudan.edu.cn

http://crossmark.crossref.org/dialog/?doi=10.1074/jbc.REV120.015254&domain=pdf&date_stamp=2020-10-26

http://creativecommons.org/licenses/by/4.0/

 

bacterial outer membrane by the lipid A moiety (10) (Fig. 1). Animal cells sense extracellular LPS through the receptor known as MD2-TLR4 (myeloid differentiation factor 2–Toll- like receptor 4). Other host proteins (mainly LPS-binding pro- tein (LBP), phospholipid transfer protein (11, 12), soluble and/ or membrane CD14 (sCD14 and/or mCD14), and albumin (13)) can extract individual LPS molecules from bacterial mem- branes and transfer them to cell surface (or endosomal) MD-2– TLR4. (Fig. 2A) Activation occurs most potently (i.e. at picomo- lar concentrations (14)) when LPS molecules with hexaacyl lipid A (Fig. 1) form ternary complexes with MD-2–TLR4. These ternary complexes form stable dimers that promote as- sembly of multiprotein intracellular signaling complexes and initiate intracellular signal/transduction (15). MyD88-depen- dent signaling is initiated by activated receptor complexes at the cell surface, whereas MyD88-independent (TRIF-depend- ent) signaling is triggered by activated LPS–MD-2–TLR4 re- ceptor complexes within endosomes (16). Together, these sig- nals initiate inflammatory cellular and humoral responses to kill the invading bacteria.

From stimulus to inhibitor: the importance of LPS acylation

Most of the aerobic Gram-negative bacteria that inhabit mu- cosal surfaces and/or commonly cause disease in animals pro- duce LPS that has hexaacyl lipid A (3). In Escherichia coli LPS, for example, myristoyl and lauroyl chains are attached via acy- loxyacyl linkages to the hydroxyl functions of two of the four diglucosamine-linked 3-hydroxymyristoyl chains (Fig. 1). Some other Gram-negative aerobes and most Gram-negative anae- robes (e.g. Bacteroides species), produce pentaacyl LPS, a weaker MD-2/TLR4 agonist (17) that may inhibit hexaacyl LPS signaling (18). The fatty acid composition can vary both within and between species (19); in Vibrio cholerae and Pseudomonas

aeruginosa, for example, secondary acyl chains can be hydroxy- lated (20, 21). AOAH cleaves only the acyloxyacyl linkages, removing the

secondary (or “piggyback”) acyl chains to produce pentaacyl or tetraacyl lipid A (Fig. 1). An early study found that human AOAH preferentially removed shorter (laurate . myristate), and nonhydroxylated (laurate . 2-OH–laurate) secondary chains (22). The resulting tetraacyl LPS still binds the MD-2– TLR4 receptor with comparable affinity, but the probability that the tetraacyl LPS–MD-2–TLR4 complex will form a stim- ulatory ternary complex is markedly reduced (23–25). This may reflect ligand acylation-dependent differences in the ability of the resulting ternary complexes to form the stable higher- order structures required for signal transduction (15). The extraction and transfer of LPS monomers from LPS-rich

surfaces (e.g. the bacterial outer membrane) to CD14 is neces- sary not only for efficient delivery of activating (hexaacyl) LPS to MD-2–TLR4 but also to present the LPS as a substrate for AOAH (26) (Fig. 2B). The apparent “Km” is 100-fold lower for transfer of the LPS monomer from sCD14 to MD-2–TLR4 than it is for deacylation by AOAH of LPS bound to sCD14 (26). However, the higher levels of sCD14 versus MD-2 often found in tissues suggest that monomeric LPS:sCD14 intermedi- ates could accumulate, followed by AOAH action and delivery of much less stimulatory (“?stimulatory”) AOAH-deacylated LPS (dLPS) to MD-2–TLR4 (Fig. 2B). Taken together, these properties seem most compatible with the ability of LPS at picomolar concentrations to activate potent immune responses acutely followed by a subsequent dLPS-assisted return to “rest- ing” conditions. Deacylation by AOAH reduces the stimulatory potency of

hexaacyl LPS molecules by as much as 50–100-fold (27–30). Moreover, dLPS can competitively reduce binding of hexacyl LPS to MD-2–TLR4 and thereby inhibit formation of the hex- aacyl LPS–MD-2–TLR4 ternary complexes needed to activate

Figure 1. AOAH removes secondary acyl chains from LPS. E. coli LPS and lipid A are shown. The diglucosamine backbone is phosphorylated at 1 and 49, and four primary 3-hydroxy fatty acyl chains are attached to the backbone in ester or amide linkage. Secondary acyl chains (red), usually myristate or laurate, are attached via acyloxyacyl linkage to two of the primary chains. Six acyl chains and both phosphates are required for optimal recognition by the MD-2–TLR4 receptor on animal cells. AOAH cleaves the acyloxyacyl linkages (arrows), converting stimulatory hexaacyl LPS into antagonistic tetraacyl LPS.

JBC REVIEWS: AOAH transforms LPS

J. Biol. Chem. (2020) 295(51) 17842–17851 17843

 

 

signaling (Fig. 2B). dLPS inhibition of LPS stimulation has been demonstrated in vitro using human umbilical vein endothelial cells (28, 31), human neutrophils (30), and murine macro- phages and splenocytes (29). Even substoichiometric concen- trations of dLPS may inhibit LPS signaling (32). In addition, dLPS can compete with LPS at each of the preceding steps involved in presenting LPS to MD-2–TLR4 (33) (i.e. by engag- ing LBP and/or CD14) (34). Teghanemt et al. (35) and Gioan- nini et al. (25, 26) demonstrated directly that dLPS interacts with MD-2 to form a complex with markedly reduced ability to activate TLR4, therefore acting as an MD-2–TLR4 antagonist. Activation of LPS-sensing cytosolic caspases by LPS also appears to depend upon the acylation state of LPS (36–38). An effect of AOAH-dependent LPS deacylation on regulating the activation of this cytosolic recognition/response system has not been reported. Independent evidence for the importance of lipid A acylation

in host–Gram-negative bacteria interactions has come from Gram-negative pathogens that can make hypoacylated (,6 acyl chains) LPS (39). Intracellular Shigella flexneri bacteria can avoid activating an inflammatory response by reducing LPS acylation (40), Salmonella typhimurium may produce a pen- taacyl LPS with low stimulatory activity (41), Francisella spp. make tetraacyl LPS (42), and Yersinia pestis switches produc- tion from hexaacyl to tetraacyl LPS at mammalian body tem- peratures and prevents mobilization of TLR4-dependent host defenses (43, 44). For these intracellular bacteria, producing hypoacyl lipid A/LPS is an immune evasion mechanism. As will be discussed below, AOAH-mediated transformation

of LPS from hexaacyl to pentaacyl or tetraacyl can be anti- inflammatory (blunting LPS signaling via TLR4) as well as re-

storative (reestablishing normal host responsiveness to infec- tion by inactivating stimulatory LPS). First, we summarize some of the enzyme’s important features.

AOAH biosynthesis, structure

AOAH is a glycoprotein with Mr = 52,000–60,000 (45, 46). Its two potential subunits are joined by a single disulfide bond before the precursor peptide is cleaved to form the mature pro- tein (46). The smaller subunit is a member of a protein family that includes several proteins that act at lipid-water or lipid-air interfaces or as cofactors for lysosomal hydrolases (47). The larger subunit (;50,000 Da) is a GDSL lipase; the best-charac- terized member of this family is platelet-activating factor ace- tylhydrolase (48). When AOAH cDNA was expressed in cul- tured fibroblasts, much of the precursor protein was secreted, internalized by other cells, and cleaved within an acidic endoso- mal-lysosomal compartment to form the mature enzyme (49), which was more active toward LPS than was the precursor (48, 49). Analysis of the crystal structure led Gorelik et al. (48) to con-

clude that “LPS binds to AOAH with its fatty acid tails covered by the hydrophobic pocket formed by the saposin and catalytic domains and a secondary (acyloxyacyl) chain buried in the hydrophobic tunnel at the active site.” A space-filling model resembles the “fingers in glove” appearance of lipid A inserted into MD-2 (50). The DNA sequence that encodes AOAH has been found in

all vertebrates studied except fish, which produce a TLR4 that is not activated by LPS (51), and also in many invertebrates (52). It is likely that AOAH was the LPS esterase originally

Figure 2. A, soluble cofactors enable LPS signaling. A single Gram-negative bacterium or outer membrane vesicle can contain .106 LPS molecules, all present in the outer leaflet of the outer membrane. 1, LBP binds to membrane particles containing hexaacyl (stimulatory) LPS and initiates the extraction and transfer of LPS monomers to sCD14 or mCD14. 2, LPS can be exchanged between sCD14 and mCD14. Either sCD14 or mCD14 can transfer the LPS monomer to MD2- TLR4 (3), which initiates TLR4 signaling following formation and dimerization of LPS-MD2-TLR4 ternary complexes (see text for more details) (4). B, AOAH deacylates LPS. 5, sCD14 can transfer LPS to mCD14 or to extracellular AOAH, which removes the two secondary acyl chains (red) to produce tetraacyl (?stimulatory, antagonistic) dLPS. 6, dLPS is transferred back to sCD14 (no protein-free form of either hexaacyl or tetraacyl LPS is detected). sCD14 can trans- fer the tetraacyl dLPS to MD2-TLR4, but this ternary complex, unlike that of hexaacyl LPS-MD-2–TLR4 (Fig. 2A), does not form a stable dimerized ternary com- plex that can initiate signaling. 7, dLPS may also compete with hexaacyl LPS for binding MD2-TLR4. 8, mCD14 can promote uptake of LPS into an endosomal compartment, where it can be deacylated by AOAH.

JBC REVIEWS: AOAH transforms LPS

17844 J. Biol. Chem. (2020) 295(51) 17842–17851

 

 

described in the slime mold Dictyostelium discoideum (53), which also has lipid A–deacylating amidases and uses Gram- negative bacteria as a foodstuff (54). Although AOAH can act on other lipids, including phospholipids (55) and some bacte- rial lipoproteins (48), its activity toward lipid A/LPS is unique.

Cellular sources of AOAH

Although AOAH was intially found in neutrophils (9), cellu- lar levels of AOAH are often lower in circulating and exudate neutrophils than in monocytes/macrophages, hepatic Kupffer cells (56), immature dendritic cells (57, 58), and NK cells. AOAH expression has also been found in mucosal ILC1 cells, which are closely related to NK cells but lack granulysin and perforin and have a genomic “superenhancer” region that includes part of the AOAH gene (59). Naive T cells just released from the thymus express AOAH mRNA along with mRNAs for complement receptors, IL-8, and TLR1 (60), consistent with a role in antimicrobial defense. Macrophages internalize LPS into an endosomal compart-

ment that contains AOAH (61); mCD14 can enable this uptake (62) (Fig. 2B). LPS deacylation occurs over several hours and can be inhibited by agents that prevent endosomal acidification (61), in keeping with the enzyme’s low pH optimum for LPS deacylation in vitro. In general, cellular and secreted levels of AOAH increase for 1 or 2 days after the cells are stimulated with LPS or other agonists. Murine peritoneal macrophages greatly increased AOAH synthesis when they were stimulated in vitro with interferon-g or LPS (63). Murine alveolar macro- phages also constitutively express little AOAH, but 18 h follow- ing stimulation with LPS, either in vivo or ex vivo, they increased AOAH mRNA expression 40–100-fold (64). Accord- ing to a recent comparison of mRNA abundance in murine and human brain cells, AOAH mRNA is much more highly expressed in human than in murine microglia (65). AOAH mRNA abundance increased in embryonic murine microglia, provided that the mother had received a standard (not germ- free) diet (66), pointing to a possible role for AOAH or translo- cated LPS (see below) in embryonic development. Janelsins et al. (67) reported that resident colonic dendritic

cells (DCs) express more AOAH than do DCs in other murine organs; AOAH mRNA abundance decreased when antibiotics were added to the drinking water and was barely detectable in colonic DCs from TLR42/2 mice. AOAH activity also increased when murine marrow–derived DCs were treated with LPS (57). Like peritoneal macrophages, the marrow-derived DCs deacyl- ated the LPS in phagocytosed E. coli (57). Murray et al. (58) recently reported that circulating myeloid precursor DCs increased AOAH mRNA abundance 4-fold in people living with (latent) HIV. There is an important exception to the generalization that

AOAH is mainly produced by myeloid or lymphoid cells. When radiolabeled AOAH cDNA was used to perform a Northern blot analysis of mouse tissues, the greatest signal came from the kidney (68). In situ hybridization revealed that AOAH mRNA was localized to proximal tubule cells; further study found that mice and humans secrete mature AOAH into the urine. The AOAH could be taken up by bladder cells and deacylate LPS. A

protective role for AOAH seems likely in the urinary tract, where E. coli and other aerobic Gram-negative bacteria are the most common pathogens. AOAH can also deacylate LPS extracellularly (Fig. 2B). In ex

vivo studies of a sterile inflammatory exudate induced in rab- bits, Weinrauch et al. (69) showed the presence of AOAH in both inflammatory cells (macrophages . neutrophils) and in cell-free inflammatory fluid. Conversion of LPS to dLPS was demonstrable using either purified LPS or intact E. coli contain- ing metabolically prelabeled LPS; it was greatest when macro- phages were present (70). In summary, although AOAH has been studied most often in

phagocytes that respond to LPS and contribute to innate anti- bacterial defense—macrophages, DCs, and neutrophils—it is also produced by Kupffer cells, NK cells, ILC1 cells, recent thy- mic immigrant T cells, microglia, and renal proximal tubule cells. Cell stimulation is generally followed by gradual increases in AOAH abundance from low constitutive levels to maximal production within a day or two.

AOAH prevents, moderates, and terminates responses to LPS

Evidence that AOAH influences how animals respond to LPS and Gram-negative bacteria has come from studies in transgenic AOAH-producing (71) and Aoah2/2 (57) mice. Sig- nificant roles have been found for preventing hexaacyl LPS from entering the bloodstream from a tissue site or the intes- tine, moderating and shortening the inflammatory response to Gram-negative bacteremia and pulmonary challenge, and ter- minating the immunosuppressive period of cellular reprogram- ming (“tolerance”) that follows exposure to LPS.

AOAH prevents stimulatory LPS from entering the bloodstream

The pathological consequences of Gram-negative bactere- mia and endotoxemia have ranged from metabolic diseases to septic shock and death. Like Gram-negative bacteria, LPS moves via lymphatics from a subcutaneous injection site to draining lymph nodes and continues via lymphatics into the bloodstream. AOAH-dependent deacylation, largely carried out by macrophages, can inactivate most of the LPS before it reaches the circulation. When Aoah2/2 and Aoah1/1 mice were compared, subcutaneously injected LPS induced more ro- bust proliferation of B cells and plasmablasts in draining lymph nodes of Aoah2/2 mice and increased their blood IgM and IgG3 levels for weeks (72, 73), confirming that AOAH partici- pates in limiting responses to LPS even before the LPS mole- cules reach draining lymph nodes and the bloodstream. Many investigators have studied the role of microbiota-

derived LPS (endotoxemia) in the pathogenesis of diseases such as atherosclerosis, diabetes mellitus, metabolic syndrome, and others (74–77). There is evidence for translocation of stimula- tory LPS into the bloodstream from the colon (78), where intes- tinal AOAH is most abundant (79). Qian et al. (79) obtained indirect evidence for AOAH-dependent LPS inactivation by measuring TLR4-stimulating activity using a TLR4 reporter cell line (80). In stool, mesenteric lymph nodes, plasma, and

JBC REVIEWS: AOAH transforms LPS

J. Biol. Chem. (2020) 295(51) 17842–17851 17845

 

 

lung, TLR4 stimulation was greater in Aoah2/2 mice than in Aoah1/1 littermate controls. The stimulatory activity de- creased when polymyxin B was added to inhibit LPS or when antibiotics that kill aerobic Gram-negative bacteria were added to the drinking water to reduce hexaacyl LPS abun- dance in the intestine (79). Protection from systemic stimulation by translocated in-

testinal LPS has often been attributed to the ability of intesti- nal alkaline phosphatase to inactivate LPS by removing one or both phosphates from lipid A (Fig. 1) (81, 82). With the recent discovery that certain primary (glucosamine-linked) acyl chains must be removed from lipid A before intestinal alkaline phosphatase can cleave either of the lipid A phos- phates (Fig. 1) (83), AOAH became the most likely host mechanism for reducing the stimulatory potency of LPS mol- ecules in the intestine by modifying lipid A structure. As noted above, its production by colonic DCs increases with hexaacyl LPS exposure. As reported by d’Hennezel et al. (84), the net impact of the

abundant Bacteroides (largely pentaacyl and monophos- phoryl (85)) LPS in the human colon is to decrease TLR4 sig- naling by competing with TLR4-activating hexaacyl LPS (18, 86). Although Bacteroides genus tetraacyl lipid A was also found in the colon, the extent to which AOAH contributes to reducing levels of stimulatory intestinal LPS and generating increased levels of competing, weakly agonistic dLPS is not known. Colonic AOAH may help prevent excessive LPS- induced tolerance in DCs (see below), inflammation stimu- lated by excess hexaacyl LPS (as may accumulate during colitis (80)), and translocation of hexaacyl LPS into the blood- stream (79).

AOAH decreases LPS-induced inflammation in tissues

Although AOAH may be synthesized slowly after the onset of infection or LPS challenge, the enzyme can play a significant role in preventing potentially harmful consequences of the inflammatory response. For example, AOAH shortened the du- ration of LPS– or Klebsiella pneumoniae–induced pulmonary inflammation in models of experimental pneumonia (64). In the Aoah2/2 animals, persistent LPS stimulated alveolar mac- rophages and epithelial cells to recruit more neutrophils to the lung, greatly delaying recovery and decreasing survival. AOAH has also improved survival from experimental Gram- negative bacteremia (71), hastened recovery in bacteremia sur- vivors (87), and moderated LPS-induced hepatic inflammation (56, 71, 88). In addition to the tissue injury that can be produced by ex-

cessive host responses to infection, tissue damage can be caused when host cells produce potentially toxic mediators, such as oxidized phospholipids and lysophospholipids. For example, dead cells can release oxidized phosphatidylcholine that mCD14 can deliver into DCs and activate an inflamma- tory response (89). Although AOAH can deacylate phospho- lipids and lysophospholipids in vitro (48, 55), a role for the enzyme in inactivating these or other lipids in vivo has not been established.

AOAH shortens endotoxin tolerance/immunosuppression

As if to prevent damage from “friendly fire,” the initial inflammatory response to LPS is typically followed by a state of tolerance (cellular reprogramming) (90–92) during which many of the host animal’s pro-inflammatory responses to sensing LPS are diminished while some anti-inflammatory responses increase. Host defenses are typically weaker during this period of relative immunosuppression. Homeostasis is usually restored within a few days, and the animal’s responses to subsequent LPS exposure (or microbial invasion) return to normal. When Lu et al. gave a small intraperitoneal dose of LPS to

Aoah2/2 mice, however, the animals remained tolerant for many weeks and were more likely than Aoah1/1 mice to die from live E. coli challenge (93). Further study found that stimu- latory LPS molecules persisted for many weeks in the peritoneal cavities of Aoah2/2 mice and continued to tolerize macro- phages there (94). AOAH was required to end the tolerant state (Fig. 3). It seems unlikely that long-term epigenetic changes prolonged tolerance because the tolerant peritoneal macro- phages quickly regained responsiveness when they were trans- ferred to a nontolerant, Aoah1/1 animal (94). In keeping with these findings in living mice, Mages et al.

(95) found large AOAH increases in LPS-tolerant marrow mac- rophages in vitro. In addition, stimulating cells with one TLR agonist may invoke tolerance to others (“cross-tolerance”), plausibly a general mechanism for reducing cell activation dur- ing recovery from infection. Lu et al. (57) found that AOAH ac- tivity in murine DCs increased when the cells were treated not only with LPS (TLR4) but also with TLR agonists CpG oligonu- cleotides (TLR9) and Micrococcus luteus (TLR2)—but not with inflammatory cytokines. AOAH production may increase dur- ing tolerance elicited by many different bacterial MAMPs. Another role for AOAH in the regulation of LPS-induced tol-

erance was reported by Janelsins et al. (67), who found that CD1031CD11b1ALDH2 colonic dendritic cells express much more AOAH than do DCs in other organs. Aoah2/2 colonic DCs expressed more features of tolerance—less IL-6 produc- tion, less Th17 polarization, and greater Treg cell induction— than did colonic DCs from Aoah1/1 mice. There is also evidence that intestinal AOAH can influence

tolerance to pulmonary allergens (79). Previous studies had found that gut-derived or intratracheally administered hexaacyl LPS and other TLR agonists can decrease allergic TH2 re- sponses linked to the development of asthma, whereas gut- derived pentaacyl LPS may permit them (96–99); in other stud- ies, low LPS doses elicited TH2 allergic responses (e.g. IL-4, IL-5, and IL-10) when mice were challenged with house dust mite (HDM) extract, whereas high doses did not (100, 101). In keeping with these results, Qian et al. (79) found that LPS trans- locating from the intestine to the lungs reduced pulmonary epi- thelial cell TH2 responses to HDM, whereas translocating dLPS enhanced allergen sensitivity (Fig. 4). Taken together, the in vivo studies have shown that AOAH

may play important roles in terminating LPS- and Gram-nega- tive bacteria-induced tolerance, restoring innate immune

JBC REVIEWS: AOAH transforms LPS

17846 J. Biol. Chem. (2020) 295(51) 17842–17851

 

 

responsiveness, and influencing host susceptibility to pulmo- nary allergens.

An essential gene? Disease associations?

If AOAH is required to silence LPS in vivo in mice, does it play a role in human disease causation? Although intronic SNPs in the AOAH gene have been associated with chronic rhi- nosinusitis (102), asthma (103), and early onset Alzheimer’s dis- ease (104), how these mutations might alter AOAH abundance or function is not understood. Taking a different approach, Gorelik et al. determined the three-dimensional structure of AOAH, characterized its active site and catalytic mechanism, and used this information to derive a list of mutations that should alter the enzyme’s function (48). A comparison of this list with the mutations listed in the human gnomAD browser (Broad Institute, 2020) did not identify a homozygote (missense or loss of function mutation) at any of the amino acid positions predicted to alter the enzyme’s activity. This observation and the gene’s conservation during evolution suggest that AOAH

might be an “essential” gene. Although an essential role has not been apparent in AOAH-deficient C57Bl/6J or C3H/HeN labo- ratory mice, their ability to recover from natural infections and many other challenges has not been tested. It is also possible that AOAH production differs from indi-

vidual to individual, and an interesting regulatory mechanism has been suggested by human gene-disease association studies. Several research groups have found highly significant associa- tions between polymorphisms in the same intron in the HLA- DR region on chromosome 6 (6p21) and AOAH expression on chromosome 7 (7p14.2). In one study, the polymorphism was also associated with reduced risk of asthma (105), in another study it was associated with lung cancer (106), and in several other studies (107–109) a highly significant association was found with colitis. The most detailed analysis was by Fairfax et al. (109), who found that the minor C allele at an intronic SNP (rs28366298) was specific to the HLA-DRB1*04, HLA- DRB1*07, and HLA-DRB1*09 alleles and that in each instance, AOAH mRNA abundance in peripheral blood monocytes was reduced and colitis risk was increased. In both mice and humans, AOAH deficiency has been associ-

ated provisionally with less severe asthma and greater risk of colonic inflammation. Discovering how AOAH expression on chromosome 7 is regulated by an intronic sequence on chro- mosome 6 is a fascinating challenge. Studies to date have not found evidence for regulation by cis-mediation, noncoding RNA, or an intermediary gene product.

Conclusions

As Elsbach noted decades ago (7), “Effective host defense . . . requires that the inflammation-generating foreign materials be removed and the signals turned off.” We now know that many host molecules can transiently inhibit LPSs without perma- nently silencing them, whereas AOAH irreversibly transforms stimulatory LPS into a weak agonist that can competitively in- hibit LPS signaling. Understanding how AOAH contributes to silencing LPS and possibly other lipid-containing MAMPs (such as bacterial outer membrane lipoprotein, a potent TLR2 agonist (48)) may yield insights applicable to MAMPs from

Figure 3. AOAH promotes recovery from tolerance. 1, LPS is injected into the peritoneal cavity of a naive Aoah1/1 or Aoah2/2 mouse. 2, the resident and recruited peritoneal macrophages are stimulated by LPS and may internalize it. 3, after the initial inflammatory responses subside, the macrophages become tolerant; bioactive LPS (orange stars) remains. 4, in the Aoah1/1 mouse, AOAH transforms LPS to dLPS (green stars), and the macrophages recover from toler- ance. 5, if AOAH is absent, LPS cannot be inactivated (orange stars). Bioactive LPS in extracellular fluid keeps stimulating macrophages locally, preventing them from recovering from tolerance. These mice are more likely to respond slowly and die quickly after they are challenged with live E. coli on day 10.

Figure 4. Intestinal AOAH modifies pulmonary immune responses. Colo- nic AOAH reduces translocation of bioactive LPS to the lungs, decreasing induction of tolerance in pulmonary epithelial cells to Th2 allergens, such as HDM.

JBC REVIEWS: AOAH transforms LPS

J. Biol. Chem. (2020) 295(51) 17842–17851 17847

 

 

other microbes, some of which may also require biochemical modification for sustained inactivation (110, 111). Just as “source control” (e.g. killing the bacteria, draining the ab- scess) is a critical step in managing infected patients who de- velop sepsis (112, 113), permanently silencing the patients’ inciting MAMPs may be needed to prevent prolonged infec- tion sequelae (114) and coordinate with other resolving mechanisms (1, 2) to clear the infection battlefield and restore homeostasis.

Acknowledgments—We dedicate this review to Emil C. Gotschlich, whose suggestion prompted the search for the LPS lipase; to Paul D. Rick, whose Salmonella mutant enabled its discovery and assay; and to the memory of Peter Elsbach (1924–2020), whose pioneering studies on the mechanisms, regulation, and possible roles of bacte- rial digestion by host enzymes set the stage for the studies described here. We also thank Wei Jiang and Luciana Giono for contributing to the figures.

Funding and additional information—This work was supported by National Natural Science Foundation of China Grants 31570910, 31770993, and 91742104 (to M. L.) (Fudan), National Institutes of Health Grants R01 AI18188 (to R. S. M.) and R01 AI059372 (to J. P. W.), the Division of Intramural Research, NIAID, National Institutes of Health (to R. S. M.), and Veterans Affairs Grant I01 BX000949/BX/BLRD (to J. P. W.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conflict of interest—The authors declare that they have no conflicts of interest with the contents of this article.

Abbreviations—The abbreviations used are: MAMP, microbe-asso- ciated molecular pattern; LPS, lipopolysaccharide; AOAH, acylox- yacyl hydrolase; LBP, LPS-binding protein; sCD14, soluble CD14; mCD14, membrane CD14; dLPS, AOAH-deacylated LPS; NK, nat- ural killer; DC, dendritic cell; IL, interleukin; HDM, house dust mite.