What do secreted phospholipases A₂ have to offer in combat against different viruses up to SARS-CoV-2?

Abstract

Secreted phospholipases A₂ (sPLA₂s) form a widespread group of structurally-related enzymes that catalyse the hydrolysis of the sn-2 ester bond of glycerophospholipids to produce free fatty acids and lysophospholipids. In humans, nine catalytically active and two inactive sPLA₂ proteins have been identified. These enzymes play diverse biological roles, including host defence against bacteria, parasites and viruses. Several of these endogenous sPLA₂s may play a defensive role in viral infections, as they display in vitro antiviral activity by both direct and indirect mechanisms. However, endogenous sPLA₂s may also exert an offensive and negative role, dampening the antiviral response or promoting inflammation in animal models of viral infection. Similarly, several exogenous sPLA₂s, most of them from snake venoms and other animal venoms, possess in vitro antiviral activities. Thus, both endogenous and exogenous sPLA₂s may be exploited for the development of new antiviral substances or as therapeutic targets for antagonistic drugs that may promote a more robust antiviral response. In this review, the antiviral versus proviral role of both endogenous and exogenous sPLA₂s against various viruses including coronaviruses is presented. Based on the highlighted developments in this area of research, possible directions of future investigation are envisaged. One of them is also a possibility of exploiting sPLA₂s as biological markers of the severity of the Covid-19 pandemic caused by SARS-CoV-2 infection.

1. Introduction

Phospholipases A₂ (PLA₂s) comprise a superfamily of intracellular and secreted enzymes that catalyse the hydrolysis of glycerophospholipids at the sn-2 position, releasing free fatty acids and lysophospholipids [1,2]. Secreted PLA₂s (sPLA₂s) completely differ from intracellular PLA₂ enzymes in structure, enzymatic properties, tissue distribution and biological functions. sPLA₂s constitute a family of structurally-related enzymes which are widespread in nature and exert highly diverse biological functions [[3], [4], [5], [6], [7]]. Nine catalytically-active sPLA₂s and two catalytically-inactive sPLA₂-like proteins (classified into groups and subgroups) are present in humans and other mammals [3]. These human sPLA₂s are small proteins of 14–18 kDa with 6–8 disulphide bonds and a highly conserved His–Asp catalytic dyad at the active site using Ca²⁺ as a key cofactor for catalysis. They are also termed endogenous sPLA₂s and are involved in a myriad of physiological and pathological processes, from digestion, lipid metabolism, skin homeostasis, spermatogenesis or host defence to regulation of the immune response and inflammation in disease conditions such as asthma, sepsis and cardiovascular diseases. sPLA₂s participate to such processes by various mechanisms, for instance by modifying the lipid composition of cellular membranes or by producing pro- and/or anti-inflammatory lipids from both cellular and noncellular lipid components such as cell membranes, lipoproteins, microvesicles, mitochondria or foreign phospholipids from microbes and food [3,8,9].

sPLA₂s are also found in large amount in animal venoms, especially in those from snakes, scorpions and bees [[5], [6], [7]]. Venom sPLA₂s are structurally similar to endogenous mammalian sPLA₂s but display toxic effects including neurotoxicity, myotoxicity, anticoagulant activity, cardiotoxicity and haemolytic activity. They can be considered as exogenous sPLA₂s when injected to mammals, acting as weapons to kill preys. However, they are also viewed as pharmacological agents to possibly cure diseases [[10], [11], [12]]. As such, they are valuable research tools for analysing physiological and pathophysiological conditions in humans as well as promising molecules for the development of new therapeutic agents [[13], [14], [15], [16]].

Both endogenous and exogenous sPLA₂s play direct and indirect roles in host defence against various types of bacteria, parasites and viruses [[17], [18], [19], [20], [21]]. For example, an early study showed that several snake and bee venom sPLA₂s exhibit direct antiviral activity by blocking HIV-1 virus entry into host cells [21]. Later, it was shown that group III, V and X (GIII, GV and GX) endogenous sPLA₂s play antiviral roles against adenoviruses and HIV-1 by different mechanisms [[22], [23], [24]]. Other studies have shown that the levels of endogenous sPLA₂s are upregulated after infection by dengue virus, H1N1 or coronavirus [[25], [26], [27], [28]] and can modulate the antiviral response by indirect immune mechanisms [27,28].

In this review, we present the different antiviral activities of both endogenous and exogenous sPLA₂s but also their antagonistic role in worsening viral infection. It appears that sPLA₂s interfere with the life cycle of various viruses at different steps and also act indirectly by immunomodulation of the antiviral immune response. Some of their modes of antiviral action may be used to formulate strategies for the development of diagnostic and therapeutic procedures, perhaps even agents to treat viral infections.

2. Human sPLA₂s

In mammals, mainly based on information obtained from mice (m) and humans (h), 11 to 12 sPLA₂ members have been identified: GIB, GIIA, GIIC (pseudogene in humans), GIID, GIIE, GIIF, GIII, GV, GX, GXIIA, GXIIB and otoconin-90. These sPLA₂s are usually classified into three structural collections, i.e. the collection of I/II/V/X sPLA₂s, and the collections of III and XII sPLA₂s [3,4,29]. The corresponding human sPLA₂ gene names are PLA2G1B, PLA2G2A, PLA2G2C, PLA2G2D, PLA2G2E, PLA2G2F, PLA2G3, PLA2G5, PLA2G10, PLA2G12A, PLA2G12B and OC90.

Mammalian sPLA₂s are differentially expressed in tissues and cell types, and play distinct, mostly non-redundant, roles that may be dependent or independent of their enzymatic activity. All but two mammalian sPLA₂s are enzymatically active, with a wide range of specific activities depending on the nature of glycerophospholipids, and in vitro versus in vivo settings [30,31]. GXIIB sPLA₂ is a catalytically-inactive sPLA₂-like protein due to substitution of the highly conserved histidine by a leucine residue in the active site [29]. Despite its obvious lack of enzymatic activity, the protein has been shown to play a role in hepatic triglyceride metabolism by an unknown molecular mechanism [32]. Otoconin-90 is another sPLA₂-like protein found in the inner ear and involved in balance regulation [[33], [34], [35]]. It possesses a tandem repeat of two sPLA₂-like domains [33] with substitutions in key amino acids involved in enzymatic activity, rendering it catalytically inactive. It can however still bind Ca²⁺ ion, which is important for its binding properties within the inner ear extracellular matrix called otoconia [[33], [34], [35]]. The proposed biological functions of endogenous sPLA₂s are largely based on observations from transgenic and/or knock-out mouse lines [8,36,37]. Although a large number of important sPLA₂ functions have been revealed by using this approach, it should be emphasized that the results obtained in mice and humans are not always comparable. For example, GIIC sPLA₂, which is abundantly expressed in meiotic cells in the mouse testis, is found as a pseudogene in the human genome [38,39]. Similarly, GIIE sPLA₂ may play various roles related to inflammation and obesity in the mouse [40] but is barely expressed in most human cells [41,42]. GX sPLA₂ is present in the acrosome of mouse spermatozoa while it seems that GIID sPLA₂ and/or other sPLA₂s are present in human spermatozoa [[43], [44], [45]]. Furthermore, most studies were performed with sPLA₂ knock-out mouse lines in the C57Bl/6 background, where the GIIA sPLA₂ is naturally absent due to a frameshift mutation [46], while this sPLA₂ is highly expressed in other mouse strains like Balb/C [46] or in human tissues where it participates to inflammation and associated diseases, cancer and antibacterial defence among other functions [3,[47], [48], [49]]. This raises issues for possible redundancy, overlooked or overestimated roles of certain sPLA₂s, and the translational impact of the findings from mice to humans. Therefore, with the aim to better understand the biological roles of the different human sPLA₂s in diseases, which may be useful for diagnosis and treatment, it is reasonable to combine the key observations obtained from mouse models with those directly derived from human studies, from tissue distribution and function of human sPLA₂s in healthy and disease conditions up to analyses of sPLA₂ gene polymorphisms and mutations [[50], [51], [52], [53], [54], [55], [56]].

Finally, it should be mentioned that endogenous sPLA₂s exhibit complex enzymatic properties depending on the lipid membranes that they will encounter once secreted. Indeed, they can act on a large variety of lipid substrates to produce multiple lipid mediators or modify the biochemical or biophysical properties of cell membranes or extracellular microvesicles, making their action both complex and unique to the microenvironment in which each of the different sPLA₂s operate. sPLA₂s act primarily on different extracellular or noncellular phospholipids such as those of the plasma membrane, lipoproteins, lipid microparticles, lung surfactant, skin lipid bodies, food and bacterial membranes, mitochondria, but also on intracellular membranes during their secretion [8,19,30,31,50], [[57], [58], [59], [60], [61], [62], [63]]. Products of sPLA₂ phospholipolysis, free fatty acids and lysophospholipids, are involved in a number of cellular pathways by generation of multiple bioactive lipid mediators, promotion of membrane remodelling, modification of extracellular noncellular lipid components, such as microparticles and lipoproteins, lung surfactant, or degradation of foreign phospholipids from microbes, microbiota and dietary components in response to given microenvironmental cues.

Some of the major biological effects of endogenous sPLA₂s and their potential involvement in various human disorders are listed in Table 1 .

Table 1.

Key biological roles of endogenous mammalian sPLA₂s, including their potential antiviral and/or proviral biological roles (underlined). The table is largely summarized from Refs. [9,40]; with additional information on GIB [56,64], GIII [23], GV [22,23], GX [22,23,27], GXIIA [[65], [66], [67], [68]] and otoconin-90 [[33], [34], [35]].

Endogenous mammalian sPLA2	Tissue and cell distribution	Physiological role	Pathophysiological role
GIB	pancreatic acinar cells, duodenum, ileum, stomach, lung, blood, spermatozoa	digestion of dietary phospholipids in the small intestine, immune response	obesity, lung inflammation, anergy of T cells in HIV infection (i.e.proviral role), anthelmintic
GIIA	intestine, Paneth cells, platelets, leukocytes, synoviocytes, skin, biological fluids (blood, tears, seminal and synovial fluid, lung alveolar space)	regulation of intestinal microflora, inflammation and lipid metabolism, role in host defence and in innate and immune responses, direct killing of Gram-positive bacteria and of commensal bacteria	amplification of inflammation in many inflammatory diseases and cardiovascular diseases, role in sepsis and various bacterial and viral infections, role in cancer and metastasis
GIID	immune system, lymphoid and dendritic cells, M2 macrophages, regulatory T cells	anti-inflammatory, resolution of Th1 immunity, candidate as an immune checkpoint	decrease of immune response, anti-tumour response, antiviral immunity in H1N1 and SARS-CoV infections(i.e.proviral role)
GIIE	adipocytes, hair follicles, uterus	fat deposition in adipose tissue and liver, hair follicle homeostasis, lipid metabolism	obesity (age-induced anti-obesity?), skin disorders
GIIF	skin, suprabasal keratinocytes	skin homeostasis, lipid metabolism	skin disorders such as cancer and psoriasis
GIII	intestine, liver, epididymal epithelial cells, mast cells, colonic epithelial cells, aorta, skin, brain	male fertility, mast cell maturation	anaphylaxis, colon cancer and colitis, atherosclerosis, skin inflammation, antiviral (adenovirus) role
GV	bronchial epithelial cells, endothelial cells, macrophages, dendritic cells, hematopoietic cells, cardiomyocytes, aorta, hypertrophic adipocytes, pancreatic β-cells, gut epithelium, brain	host defence, surfactant degradation, M2 macrophage polarization, Th2 immunity, phagocytosis of microorganisms, phagocytosis of immune complexes, apoptosis of injured myocardial cells, reduced adipose tissue inflammation by unsaturated fatty acids, anti-arthritis, anti-obesity	airway lung injury, asthma, atherosclerosis, aortic inflammation, myocardial infarction, aneurysm, antiviral (adenovirus) role
GX	airway epithelium, intestinal mucosa, stomach, immune system, macrophages, adrenal glands, dorsal root ganglia, hematopoietic cells, neutrophils, adipocytes, sperm acrosome, hair follicles	enhanced lipid accumulation and TLR4 signalling in macrophages, reduced corticosteroid synthesis by downregulating adrenal steroidogenic acute regulatory protein, neuritogenesis and pain transmission, macrophage function, reduced Th1 immunity and atherosclerotic plaque formation, suppression of insulin secretion, phospholipid digestion in the gastrointestinal lumen, repression of adipogenesis by inhibiting liver X receptor activation, anti-atherosclerosis, anti-obesity, sperm fertility, hair homeostasis	airway inflammation, asthma, decreased tissue damage by neutrophils, hypercorticosteronemia, pain, aneurysm, myocardial infarction, diabetes, adiposity, alopecia, antiviral (adenovirus) role, antiviral immunity in H1N1 infection (i.e.proviral role)
GXIIA	ubiquitous, neurons, epithelial cells of gastrointestinal tract	neurogenesis and cognition, host defence, killing of Gram-positive bacteria in vitro, poorly understood molecular mechanisms	lipid metabolism disorders?
GXIIB	liver, intestine	triglyceride metabolism by unknown molecular mechanisms, anti-steatosis effect	lipid metabolism disorders?
Otoconin-90	inner ear	maintenance of inner ear balance system	imbalance

3. Antiviral activity of endogenous mammalian sPLA₂s

Mammalian sPLA₂s, especially GIIA sPLA₂, are involved in direct and indirect host defence against various pathogens, including bacteria, parasites and viruses, and as such represent an important component of the innate immune system [17,18,24,27,28,47,56,64], [[69], [70], [71], [72]]. For example, studies as early as back in 1979 have shown that GIIA sPLA₂ can exert direct killing of Gram-positive bacteria [73,74]. It was a result of the ability of this enzyme, due to its highly positive net charge, to penetrate the bacterial peptidoglycan envelope, to reach the bacterial plasma membrane and to hydrolyse it [18,47]. In addition, the same sPLA₂ is responsible for an indirect host response triggering innate immune response, possibly via hydrolysis of lipid microvesicles or mitochondria released by activated platelets (or other cells) to release potent lipid mediators, or when certain bacterial strains manipulate the host defence activity within a bacterial niche [57,61,75]. Multiple studies have shown that GIIA sPLA₂ has in vitro and in vivo antibacterial activity against various bacterial strains [17,18,47,69,70].

The antiviral activity of endogenous mammalian sPLA₂s is much less documented than their antibacterial roles. It was first reported in vitro for two potent plasma membrane-hydrolysing sPLA₂s, hGV and hGX sPLA₂s, which could prevent host cells from being infected with adenoviruses [22,23] (Table 1). These two sPLA₂s are expressed in human airway epithelium and macrophages, and the expression of hGV sPLA₂ is upregulated by virus-related stimuli of these cells. Adenovirus particles were insensitive to exogenously added hGV and hGX sPLA₂s. However, addition of recombinant hGV and hGX sPLA₂s, but not hGIIA, to HEK293 cells suppressed the number and size of adenovirus plaque formation. In contrast to the main antibacterial action of sPLA₂s, which depends on their direct hydrolytic action on bacterial membranes [70], the antiviral action of GV and GX sPLA₂s was indirect and dependent on their ability to act on host cell membranes (Fig. 1 ), leading to a decrease in adenovirus entry by clathrin-mediated endocytosis. The observation that suppression of adenoviral infection by hGV and hGX sPLA₂s depends on catalytic activity was confirmed with catalytically-inactive sPLA₂ mutants and supported by the finding that lysophosphatidylcholine, an sPLA₂-hydrolytic product of phosphatidylcholine enriched in the outer leaflet of the plasma membrane, also suppresses adenoviral infection. It was also suggested that the innate immune action of these sPLA₂s may be directed not only against adenoviruses but also against other respiratory viruses [22]. Besides endogenous GV and GX sPLA₂s, hGIII sPLA₂ was also able to inhibit adenovirus entry into host cells [23]. It was shown that, in addition to the catalytically-active sPLA₂ domain, the N-terminal domain unique to this enzyme is required for the anti-adenoviral effect observed in human bronchial epithelial cells. This suggests a slightly different antiviral mechanism in the case of hGIII sPLA₂. By lipidomics analysis, the authors provided evidence that hGIII, hGV and hGX sPLA₂s target different phospholipids in host cell membranes, all leading to production of lysophosphatidylcholine, which was proposed to be responsible for the antiviral action. However, the in vivo role of these sPLA₂s against adenovirus infection remains to be demonstrated.

Fig. 1.

The antiviral action of endogenous sPLA₂s in in vitro cell systems. Some sPLA₂s, for example human, may act on viruses either directly, by degrading their lipid envelope (virucidal effect?), such as in the case of HIV-1, or indirectly, by acting on cell lipid membranes to prevent effective production of viral particles, such as in the case of non-enveloped adenoviruses. Via enzymatic activity, sPLA₂s hydrolyse phospholipids (PL) from the virus or infected cellular membranes to produce lysophospholipids (LPL) and free fatty acids (FA) that can disturb some key steps of the viral life cycle or lipid metabolism of the host cell. The enveloped virus (e.g. HIV-1), with a lipid membrane derived from host plasma membrane, and the non-enveloped (no lipid) virus (e.g. adenovirus) are shown in violet and blue, respectively. The dashed line separates direct versus indirect sPLA₂ actions on cells. CAR, coxsackievirus and adenovirus receptor.

In a subsequent study, it was found that endogenous hGX sPLA₂ displays antiviral activity against different enveloped lentiviruses including HIV-1, while the enzyme was found unable to prevent adenovirus infection, which contrasts with the above study [24]. The enzymatic activity of hGX sPLA₂ was necessary for the observed antiviral effect (Fig. 1). Other endogenous sPLA₂s such as hGIIA, hGIID, hGIII, hGV and hGXIIA appeared to be ineffective, yet this was not analysed using pure recombinant enzymes. The antiviral activity was proposed to be due to a direct virucidal effect, i.e. direct hydrolysis and degradation of the viral membrane [24]. It was assumed that the lipid envelope of HIV-1 may be susceptible to the enzymatic action of hGX sPLA₂, since it originates from the host cell membranes and is rich in its outer leaflet in zwitterionic phospholipids including phosphatidylcholine [76], while hGX sPLA₂ has been shown to hydrolyse phosphatidylcholine from the plasma membrane of live cells [77,78]. However, more detailed lipidomics studies have shown that the lipid composition of the HIV virus particles is unique and distinct from the cell plasma membrane [[79], [80], [81]], leaving unknown the enzymatic properties of this sPLA₂ on the surface of HIV particles and whether it has a real capacity to directly hydrolyse and kill the virus (i.e. via a virucidal effect). Furthermore, the concentration of recombinant hGX sPLA₂ that inhibits HIV-1 infection in cellular assays versus the one that would produce the direct virucidal activity were not determined. Interestingly, the antiviral activity was observed despite the resistance of virus preparations to lysis by antibody-mediated complement activation, indicating that the sPLA₂ may act independently of the adaptive immune response or when the complement pathway is ineffective. As hGX sPLA₂ is highly expressed in the intestinal mucosa [82,83], the primary site of HIV-1 replication on natural infection, the authors suggested further studies to explore the role of hGX sPLA₂ in the innate immunity against HIV-1 infection in the gastrointestinal system. Application of the recombinant protein and/or upregulation of hGX sPLA₂ expression may thus limit viral replication and reduce the incidence of productive replication at the primary infection sites. However, as above for other sPLA₂s, the in vivo role of hGX sPLA₂ against HIV-1 infection remains to be demonstrated.

In a mouse model of infection by H1N1 influenza virus, mGX sPLA₂ (Pla2g10 gene) has been involved in host antiviral response, but in a negative and likely indirect manner [27]. Survival after infection was greater in Pla2g10-deficient mice than in wild-type littermates. mGX sPLA₂ was induced by H1N1 infection in bronchial epithelial cells and inflammatory cells, and was involved in the production of potent proinflammatory lipid mediators, likely contributing to acute lung injury. On the other hand, B and T cell antiviral responses were stronger in Pla2g10-deficient mice, suggesting that Pla2g10 promotes excessive lung inflammation and inhibits an efficient adaptive immune response, leading to poor survival. Collectively, these in vivo findings suggest that GX sPLA₂ is a “proviral sPLA₂” which contrasts with its above in vitro antiviral activity (Table 1). Based on data in this mouse model, GX sPLA₂ may be a potential therapeutic target that should be inhibited during influenza infection. However, its in vivo antiviral or proviral role after infection with adenovirus or HIV-1 (with mouse adapted strains and related lentiviruses) remains unknown and should be tested in relevant animal models.

Soon after, it has been shown that mGIID sPLA₂ (Pla2g2d gene) is negatively involved in the adaptive immune response in a mouse model of severe acute respiratory syndrome coronavirus (SARS-CoV) infection [28]. mGIID sPLA₂ is a basic protein which was cloned from spleen and other lymphoid tissues and show a modest in vitro enzymatic activity as compared to its close paralog GIIA sPLA₂ [30,84]. However, in vivo, mGIID sPLA₂ was shown to preferentially cleave phosphatidylethanolamine esterified with various polyunsaturated fatty acids at the sn-2 position [85], releasing not only omega-6 arachidonic acid but also omega-3 eicosapentaenoic and docosahexaenoic acids, which are precursors of potent anti-inflammatory lipid mediators (resolvins) involved in resolution of inflammation during infection [86]. In mice and humans, GIID sPLA₂ is expressed mainly in dendritic cells and M2 macrophages of secondary lymphoid organs, such as spleen and lymph nodes, is downregulated by pro-inflammatory stimuli, and is likely playing an anti-inflammatory and pro-resolving role [40]. For these reasons, it was termed a “resolving sPLA₂” that ameliorates inflammation through mobilisation of pro-resolving lipid mediators [40,85]. A dual role for GIID sPLA₂ as an immunosuppressive endogenous sPLA₂ in inflammation and cancer was proposed in which it exerts an anti-inflammatory effect but decreases the anti-tumour effect, thereby promoting tumour development [87]. An immunosuppressive function of GIID sPLA₂ was proposed even earlier, when it was cloned from lymphoid tissues of lymphotoxin-deficient mice [88], and shown to be highly expressed in regulatory T cells and to mediate the immunosuppressive function of these cells [89]. Regarding its role during viral infection, a critical role of mGIID sPLA₂ in age-related susceptibility to SARS-CoV infection was observed [28], revealing a proviral role of this sPLA₂. The authors carried out most of the experiments in mice, but the conclusions they made may be applicable to humans. In agreement with this, oxidative stress was found to induce GIID sPLA₂ expression in mice and in human monocyte-derived macrophages. The study showed that GIID sPLA₂, whose expression in lung dendritic cells increased with age in response to oxidative stress and also during coronavirus infection, contributed to worse outcome in mice infected with SARS-CoV. These results were supported by the observation that GIID sPLA₂-deficient mice, in comparison with wild-type mice, showed a much higher rate of survival after infection with not only SARS-CoV but also influenza A viruses. On the contrary, the attenuated antiviral immunity in transgenic mice overexpressing GIID sPLA₂ leads to severe lung inflammation and early death. The proposed molecular mechanism is that GIID sPLA₂, expressed in lung dendritic cells, produces prostaglandin D₂ and other lipid mediators that prevent the migration of dendritic cells to lymph nodes, thereby suppressing T-cell activation. This results in a decreased antiviral immunity and, hence, increased viral infection [90]. Accordingly, the specific inhibition of GIID sPLA₂ represents a potentially therapeutic approach for the treatment of patients with severe respiratory infection due to influenza viruses or coronaviruses [40]. Very recently, it has been reported on the unexpected role of mGIID sPLA₂ in the lung, serving as an important modulator of respiratory dendritic cell activation, with protective and pathogenic effects in respiratory CoV infections and immunization [91]. Together, these findings on a proviral role of GIID sPLA₂ during infection with coronavirus and influenza viruses are similar to those observed for GX sPLA₂ and H1N1 infection, suggesting that both sPLA₂s may share some common molecular mechanisms of action and may be both targeted by specific inhibitors for the treatment of infections with respiratory viruses.

Finally, recent studies suggest that hGIB sPLA₂ may negatively contribute to infection by HIV-1 by an indirect mechanism leading to anergy of CD4+ T cells, which may be at the basis of the chronic immunodeficiency syndrome observed in HIV patients [56]. This sPLA₂ may act catalytically at the surface of CD4⁺ T cells in synergy with enzymatic cofactors such as the HIV gp41 envelope protein or degradation products. Importantly, the in vitro effects of the sPLA₂ on CD4⁺ T cell anergy is dependent on its enzymatic activity and can be blocked by a specific neutralizing antibody, which may represent a therapeutic molecule. However, the effect of GIB sPLA₂ on T cell anergy should be further demonstrated in in vivo models of HIV infection or T cell immune dysfunction. The cellular source of hGIB sPLA₂ that is found in the plasma of HIV patients also remains to be determined.

4. Antiviral activity of exogenous sPLA₂s

More than two decades ago, before studies with endogenous sPLA₂s on viruses, it was shown that several venom sPLA₂s exhibit a potent antiviral activity against HIV-1 in vitro [21] (Table 2 and Fig. 2 ). In this study, the bee venom sPLA₂ (bvPLA₂) and three snake venom sPLA₂s were found to be potent inhibitors of HIV-1 replication (IC₅₀ < 1 nM). hGIB and hGIIA sPLA₂s did not show such inhibitory effects while other human sPLA₂s could not be tested in those days, because they were either unknown or not available as recombinant pure proteins [4,5]. Four venom sPLA₂s among 11 enzymes — taipoxin from the Australian coastal taipan (Oxyuranus scutellatus scutellatus), the basic sPLA₂ (CM-III) from the Mozambique spitting cobra (Naja mossambica mossambica), nigexine from the black-necked spitting cobra (Naja nigricollis), and bvPLA₂ from honey bee (Apis mellifera) — were able to protect human primary blood leukocytes from infection by various HIV-1 strains [21]. The observed inhibition resulted neither from a direct virucidal effect nor from a cytotoxic effect on host cells. Time-of-addition experiments monitoring the antiviral effect of sPLA₂ after the virus was added to host cells showed that the inhibitory sPLA₂s did not interfere with physical binding of the virus to the cell surface but blocked virus entry, possibly by interfering with the complex fusion step between viral envelope and plasma membrane. Accordingly, the inhibitory sPLA₂s were effective when added to cells before the virion uncoating step and dissociation of the reverse transcriptase complex from host cell membranes, but were ineffective after these steps (Fig. 2). In a follow-up study, a variant of HIV-1 was selected from a mixture of HIV-1 particles present in primary HIV isolates for its resistance to bvPLA₂ inhibition [103]. It turned out that this variant, HIV_RBV-3, could escape from inhibition of cell entry to increasing doses of bvPLA₂ by evolving or selecting a particular envelope glycoprotein mutated in the N-terminal region and the V1–V2 loop extensions [103]. These mutations allowed the HIV-1 variant to infect cells by a completely different clathrin endosomal-dependent pathway and to become resistant to inhibition by bvPLA₂ by about 100-fold, when compared to the original virus. In an effort to identify the region of bvPLA₂ involved in its inhibitory effect, a series of bvPLA₂ peptides were screened, and an N-terminal 15-amino acid peptide was found to block entry of T-tropic HIV-1 strains into host cells, obviously without requiring enzyme catalytic activity. However, surprisingly, the mechanism of action of this peptide was different from that of the whole bvPLA₂ molecule, as the peptide selectively inhibited T- but not M-tropic HIV-1 viruses by interfering with binding of HIV-1 to the CXCR4 chemokine coreceptor on target cells [93]. Conversely, bvPLA₂ was active on both T- and M-tropic HIV-1 strains and did not bind to the CXCR4 co-receptor. Finally, additional structure-function studies using the neurotoxic snake venom sPLA₂ OS2 (from Oxyuranus scutellatus scutellatus venom), which was also effective at inhibiting HIV-1, showed that the N-terminal region of OS2 and its enzymatic activity were both important to inhibit HIV-1 infection [92].

Table 2.

Antiviral activity of exogenous venom sPLA₂s. Most of them are monomeric sPLA₂s homologous to mammalian endogenous sPLA₂s. They belong to groups IA, IB, IIA or III according to the historical classification of sPLA₂s [1]. Heterodimeric crotoxin is composed of a basic GIIA sPLA₂ (CB) and an acidic (CA) enzymatically-inactive subunit of three peptides derived from a GIIA sPLA₂ precursor. Taipoxin is a heterotrimeric protein whose α and β subunits are GIA sPLA₂s, and with the γ subunit similar to a GIB pro-sPLA₂. D49 and K49 denotes enzymatically-active and -inactive sPLA₂ variants, respectively, due to a substitution of the conserved aspartic acid residue at position 49 in the Calcium-binding loop (binding of the Ca²⁺ cofactor is needed for catalytic activity) of enzymatically-active sPLA₂s with a lysine residue.

Exogenous sPLA2	Venom source	Type of viruses tested	Proposed mechanism of antiviral activity	Dependence on enzymatic activity	Reference
bvPLA2, CM-III, nigexine, taipoxin and OS2	Apis mellifera, Naja mossambica mossambica, Naja nigricollis, and Oxyuranus scutellatus scutellatus (honey bee and three snake species)	Retroviridae (lentiviruses)	inhibition of HIV-1 entry step, inhibition of both M− and T-tropic strains	dependent	[21,93,103]
p3bv (15-aa bvPLA2 peptide)	Apis mellifera (honey bee)	Retroviridae	inhibition of the interaction between the CXCR4 chemokine co-receptor and HIV-1 T- but not M-tropic strains, thereby preventing virus binding to host cells	independent	[93]
AP-sPLA2	Acanthaster planci (starfish)	Retroviridae	unknown, against HIV infection	not determined	[94]
Bl-D49 and K49 sPLA2s	Bothrops leucurus (snake)	Flaviviridae	preventing cells from being infected with dengue virus	not determined	[95]
crotoxin, crotoxin B subunit (CB PLA2), and recombinant CB1 and CB2 isoforms	Crotalus durissus terrificus (snake)	Flaviviridae, Peribunyaviridae, Togaviridae, Picornaviridae	possible virucidal activity against dengue, yellow fever and Zika viruses (all Flaviviridae) and other enveloped viruses from Peribunyaviridae and Togaviridae families by enzymatic cleavage of the virus lipid bilayer envelope that would cause a destabilization of the E proteins on the virus surface, leading to its inactivation	likely dependent	[[96], [97], [98], [99]]
crotoxin and its CB PLA2 subunit	Crotalus durissus terrificus (snake)	Flaviviridae	possible virucidal activity against hepatitis C virus (Flaviviridae) and/or interference with virus entry into host cells	not determined	[100]
CM-II	Naja mossambica mossambica (snake)	Flaviviridae, Retroviridae, Coronaviridae, Togaviridae, Herpesviridae, Orthomyxoviridae, Paramyxoviridae, Picornaviridae, Rhabdoviridae	possible specific virucidal activity against hepatitis C, dengue and Japanese encephalitis viruses (all Flaviviridae) and other viruses whose envelope is derived from the endoplasmic reticulum membrane by attacking their membranes; but also virucidal against HIV-1 whose membrane buds from the plasma membrane	likely dependent	[101]
Mt-I (D49) and Mt-II (K49)	Bothrops asper (snake)	Flaviviridae, Herpesviridae, Orthomyxoviridae, Picornaviridae, Rhabdoviridae	possible virucidal activity against dengue and yellow fever viruses (both Flaviviridae)	dependent	[102]

Fig. 2.

The antiviral action of exogenous sPLA₂s in in vitro cell systems. Direct action: certain venom sPLA₂s may exert an antiviral activity by direct hydrolysis of the lipid viral envelope (virucidal effect?). The surface of various enveloped viruses, with the lipid membrane budding from different host cell membranes, is shown by different colours (orange, originating from the endoplasmic reticulum; blue-green, from the Golgi apparatus; green, from the plasma membrane; violet, from the plasma membrane (e.g. HIV-1) or endoplasmic reticulum). Indirect action: venom sPLA₂s may exert antiviral activity by binding and hydrolysis of host cell membranes, thereby inhibiting one or several of the first steps of viral infection: from binding of viruses to entry or virus-driven lipid metabolism. The dashed line separates both direct and indirect sPLA₂ actions on cells. A red curved arrow denotes the unsuccessful attack of a virus on the host cell as a result of the sPLA₂ action. AP-sPLA₂, a catalytically-active sPLA₂ from Acanthaster planci starfish; CB, crotoxin B subunit; CM-II and CM-III, catalytically-active sPLA₂s from Naja mossambica mossambica snake venom; Mt-I, a catalytically-active sPLA₂ from Bothrops asper snake venom. Other abbreviations used are the same as those in Fig. 1 and Table 2.

During the last decade, several other studies have been carried out, showing a broad antiviral activity of different snake venom sPLA₂s (Table 2 and Fig. 2). From these reports, it is difficult to unequivocally conclude that enzymatic activity of an sPLA₂ is absolutely necessary for the antiviral effect observed. For example, both catalytically-active (D49) and catalytically-inactive (K49) sPLA₂s, isolated from the venom of the Bahia lancehead (Bothrops leucurus), seemed to show similar antiviral activity in a cellular model of infection by the dengue virus [95]. However, since no dose-response curves to measure the relative IC₅₀ values between D49 and K49 sPLA₂s were performed, it is difficult to compare the relative activity of the two types of sPLA₂s and exclude the possibility of a minor contamination of the K49 sPLA₂ preparation with the D49 active sPLA₂ purified from the same venom. Furthermore, some reports claimed that sPLA₂s exert a direct virucidal activity but this was not conclusively demonstrated by either measuring the direct hydrolytic activity of the sPLA₂ on viral particles or performing lipidomics analyses, or by washing out the sPLA₂ after the incubation of viruses with this latter, or by addition of an sPLA₂ inhibitor that would prevent the subsequent action of the sPLA₂ on cells at the time of addition of the virus–sPLA₂ mixture to host cells.

It was reported that crotoxin from the South American rattlesnake (Crotalus durissus terrificus) exhibits antiviral activity against dengue and yellow fever viruses [96]. In addition to preventing the cells from viral infections, the authors observed a pronounced effect of crotoxin when preincubated with the virus particles, suggesting a virucidal effect. However, the sPLA₂–virus mixture was not separated before addition to cells, leaving open the possibility that the sPLA₂ acts in a subsequent step of infection, for instance by interfering with the cell machinery during virus binding or entry. Crotoxin B (CB), the catalytically-active subunit of heterodimeric crotoxin, appeared to lose antiviral activity when its enzymatic activity is inhibited [97,98]. The virucidal action of crotoxin may be the result of hydrolysis of glycerophospholipids on the virus envelope, leading to a disruption of the lipid bilayer and destabilization of envelope (E) proteins on the viral surface. Molecular dynamics simulations suggested that crotoxin may gain access to the dengue virus lipid bilayer through the pores found on each of the twenty 3-fold vertices in the E protein shell on the virus surface [98]. However, definitive experimental evidence supporting this scenario is lacking. Crotoxin and its CB subunit had antiviral activity not only against Flaviviridae viruses (such as dengue, yellow fever and Zika) but also against Peribunyaviridae and Togaviridae viruses whose envelopes are different, originating from Golgi and plasma membranes, respectively. For example, two recombinant enzymatically active CB variants, CB1 and CB2, showed antiviral effects, not only against dengue, yellow fever and Zika viruses, but also against chikungunya virus (Togaviridae) [99]. In a separate study [100], crotoxin and its CB subunit were effective against in vitro infection with hepatitis C virus (Flaviviridae). The results indicated that CB blocks viral infection by action on the virus particle and host cells [100]. Thus, crotoxin and its CB subunit exert antiviral effects by possibly acting at different stages of the virus life cycle.

Mt-I, a myotoxic sPLA₂ from terciopelo (Bothrops asper) snake venom was shown to exert antiviral activity against the enveloped viruses of Flaviviridae family, similar to that of crotoxin and CB [102]. Interestingly, the results showed that Mt-I, which is catalytically active (D49), was approximately 1,000-fold more potent than its catalytically inactive variant Mt-II (K49), suggesting a critical role of the sPLA₂ enzymatic activity in the antiviral effect of Mt-I sPLA₂.

An interesting finding resulted from the study of the antiviral action of the CM-II sPLA₂ isolated from the venom of the Mozambique spitting cobra (Naja mossambica mossambica) against different types of viruses [101]. CM-II possesses potent antiviral activity (IC₅₀ values of 0.03–1.3 ng/ml, i.e. less than 1 nM) against hepatitis C, dengue and Japanese encephalitis viruses, all from the Flaviviridae family, whose viral particles bud from the endoplasmic reticulum. In contrast, CM-II was virtually ineffective against viruses that bud either from the plasma membrane (Sindbis, influenza and Sendai viruses) or the trans-Golgi network (herpes simplex virus). One exception was HIV-1, which was inhibited by CM-II with an IC₅₀ value of 5.4 ng/ml, a finding in line with the IC₅₀ value measured for the analogous CM-III sPLA₂ in an earlier study [21]. The potent antiviral activity of CM-II against hepatitis C and dengue viruses was inhibited by the broadly specific sPLA₂ inhibitor manoalide, suggesting that enzymatic activity is necessary. Thus, based on differences in the physicochemical properties of the phospholipid bilayers of viruses and host cells, broad-spectrum antiviral sPLA₂s may be developed to specifically target viral envelope lipid bilayers derived from the endoplasmic reticulum without targeting the lipid bilayers of host cell membranes.

Most recently, AP-sPLA₂, isolated from the crown-of-thorns starfish (Acanthaster planci) was shown to exert in vitro activity against HIV infection of peripheral blood mononuclear cells [94]. The authors hypothesized but did not demonstrate that AP-sPLA₂, like endogenous GX sPLA₂ [24], directly hydrolyses the HIV phospholipid bilayer envelope which causes the virus to become inactive.

Overall, these studies have shown that several but not all venom exogenous sPLA₂s can exert potent and broadly specific antiviral effects against different viruses by multiple mechanisms, from a possible direct virucidal action to more complex interplay between sPLA₂, virus and host cells, from primary binding to the plasma membrane and specific receptors to physical virus entry, membrane fusion, internalization, multiplication and budding steps. In most cases, the exact mechanism of action, which may even be miscellaneous for a single sPLA₂ acting on different viruses, remain to be discovered or at least ascertained by well-designed experiments. Furthermore, a major limitation is that all of these studies were performed only in vitro in cellular models of viral infection. Their therapeutic potential should be confirmed by performing in vivo experiments to provide a pathophysiological relevance of the findings.

5. Conclusion and future perspectives

sPLA₂s constitute an important family of both enzymatically active and inactive proteins with a wide spectrum of activities whose biological functions, including antiviral action, have not been completely elucidated. Endogenous sPLA₂s play an important but complex role in the innate immune system, sometimes defensive, sometimes offensive [8,9,104]. These yin-yang roles, for instance the proviral or antiviral effects of sPLA₂s, are sometimes observed when a particular biological system is analysed in in vitro versus in vivo conditions, and as such is difficult to understand as a whole. Furthermore, it is difficult to definitely conclude from the current studies whether a certain sPLA₂ is directly or indirectly influencing the virus life cycle, from direct virucidal effect to indirect effect on the binding, entry, or even budding steps of viruses to host cells (Fig. 1, Fig. 2). As a great deal of studies presented above has reported on the antiviral effects of sPLA₂s by using in vitro cellular models of viral infections, there is a need to perform the corresponding in vivo experiments. Only in vivo results will expose the real antiviral potential of sPLA₂s, acting either as therapeutic molecules or as “offensive” molecules whose activity would need to be hampered by active site-specific inhibitors or inhibitory antibodies. Most intensively, efforts should be concentrated on the most potent sPLA₂ representatives acting on the virus life cycle, such as GIII, GV and GX sPLA₂s, which may act on both host cell membranes and the viral envelope that has a lipid membrane derived from host cells, yet with subtle differences in their membrane lipid composition and biophysics. sPLA₂s may also play an indirect role in the antiviral adaptive immune response where, interestingly, some sPLA₂s like GX but also GIID sPLA₂ have a negative (proviral) effect on the viral infection by H1N1 and SARS-coronavirus adapted to mouse strains [27,28]. In addition, it has been recently shown that GIB sPLA₂ is a new player in lentiviral infections, where it may exert a proviral role by inducing CD4+ T cell anergy and thereby participate to the pathophysiology of HIV infection [56]. Here, the specific or combined inhibition of involved sPLA₂s may be beneficial to prevent viral infection. Exogenous sPLA₂s, mostly but not exclusively of snake venom origin, also display antiviral effect that may be mediated by their enzymatic activity to either directly attack the host-derived membranes of certain enveloped viruses or interfere with the viral cell cycle at different steps. Whether these sPLA₂s may be used as antiviral agents in vivo remains to be demonstrated, considering of course their toxicities.

Future research on the antiviral potential of sPLA₂s may thus proceed in four major directions. (1) Confirmation or disproval of the promising in vitro antiviral effects of certain venom and endogenous sPLA₂s by conducting appropriate in vivo studies. (2) Description of the exact molecular mechanism(s) of direct or indirect antiviral activity of the various sPLA₂s, as a starting point for development of new antiviral sPLA₂-based therapies. This would also include unravelling of the indirect mechanisms by which GX and GIID sPLA₂s (or possibly other isoforms) exacerbate H1N1 and coronavirus infections to pave the way for new therapeutic avenues. (3) Further description of the broadly specific antiviral role of endogenous and exogenous sPLA₂s against various types of viruses, up to infection by SARS-CoV-2 virus, the etiological agent of the Covid-19 pandemic. (4) Exploration of the possibility that various sPLA₂s such as GX or GIID may represent new biomarkers of severity in viral infections, as recently shown for the inflammatory-type GIIA sPLA₂ in Covid-19 patients with severe disease [105].

Authors’ contributions

J.P. wrote the first drafts of the manuscript, figures and tables, F.B. and G.L. significantly upgraded and polished the main text, figures and tables, and I.K. checked the facts and reviewed the manuscript.