Annexin A2 is a multifunctional calcium- and lipid-binding protein that is expressed in nearly all human tissues and cell types. AnxA2 exists as a monomer localized to the cytoplasm, vesicle-bound, or as a heterotetrameric complex termed A2t consisting of two AnxA2 monomers bridged by an S100A10 dimer found on the inner and outer leaflet of the plasma membrane. Both AnxA2 and A2t have been implicated in a wide range of intracellular processes including membrane domain organization, membrane fusion, vesicle aggregation, cytoskeletal-membrane dynamics, epithelial cell polarity, exocytosis, endocytosis, phagocytosis, and transcriptional regulation through binding of AnxA2 to RNA (reviewed in Gerke and Moss, 2002; Rescher, 2004; Bharadwaj et al., 2013; Hitchcock et al., 2014; Hajjar, 2015; Schloer et al., 2018).

More broadly, AnxA2 has been implicated in immune function, multiple human diseases, and viral infection (Hajjar, 2015; Tanida et al., 2015; Bećarević, 2016; Schloer et al., 2018). AnxA2 expression in some cancers can promote metastasis and function as a prognostic marker of recurrence and survival (Lokman et al., 2011; Zhang et al., 2013; Xu et al., 2015). This involvement of AnxA2 in human health and disease has prompted the development of pharmacological inhibitors of AnxA2 and A2t (Reddy et al., 2011, 2012, 2014; Liu et al., 2015), and these compounds are being explored in a growing number of therapeutic contexts. One class of inhibitors, for example, has been shown to block human papillomavirus (HPV) type 16 (HPV16) infection in cervical epithelial cells (Woodham et al., 2015). It is presumed that these A2t inhibitors disrupt the function of A2t during viral infection, though this specific mechanism still needs to be verified. Importantly, HPV is just one virus in a list of at least 13 viruses with known AnxA2 associations during binding, endocytosis, and egress (Table 1).

The annexin superfamily is highly conserved across eukaryotic phyla, from unicellular organisms to complex plants and animals (Moss and Morgan, 2004; Jami et al., 2012; Einarsson et al., 2016), and annexins have been associated with both human and non-human viral pathogens (summarized in Figure 1). This review focuses on seven viruses with direct links to AnxA2 during their lifecycle. To more confidently cross-compare cellular functions, we specifically discuss the viruses that target human epithelial cells. AnxA2 is utilized by HPV, enterovirus 71 (EV71), respiratory syncytial virus (RSV), and cytomegalovirus (CMV) during cell attachment and penetration (Wright et al., 1994, 1995; Raynor et al., 1999; Malhotra et al., 2003; Derry et al., 2007; Yang et al., 2011; Woodham et al., 2012, 2015; Dziduszko and Ozbun, 2013; Taylor et al., 2018), by hepatitis C virus (HCV) and influenza A virus (IAV) during replication (LeBouder et al., 2008; Backes et al., 2010; Saxena et al., 2012; Ma et al., 2017; Solbak et al., 2017), and by measles virus (MV) during assembly and maturation (Koga et al., 2018). In some cases, AnxA2 has been implicated in multiple life cycle steps of the aforementioned viruses, underscoring the importance of a more complete approach to understanding the role of AnxA2 in viral infection.

Persistent infection with HPV can lead to the development of a variety of anogenital and oropharyngeal cancers causing significant morbidity worldwide (Watson et al., 2008; Forman et al., 2012; Spence et al., 2016). HPV is a non-enveloped double stranded DNA (dsDNA) virus that enters basal keratinocytes through a non-canonical endocytic pathway while interacting with a number of host molecules (Spoden et al., 2008; Schelhaas et al., 2012; Raff et al., 2013; Day and Schelhaas, 2014; DiGiuseppe et al., 2017). In the search to identify an HPV uptake receptor, AnxA2 and A2t were discovered as central mediators of HPV entry and intracellular trafficking. Interestingly, it has been suggested AnxA2 and A2t have independent functions in HPV attachment and intracellular trafficking (Dziduszko and Ozbun, 2013). For example, it was shown that AnxA2 and A2t colocalize with HPV at the cell surface and that antibodies against AnxA2 alter entry kinetics (Dziduszko and Ozbun, 2013), but antibodies against the S100A10 subunit (Dziduszko and Ozbun, 2013) and targeted knock-out via CRISPR/Cas9 (Taylor et al., 2018) does not affect cellular entry in vitro. Furthermore, when the full A2t complex is knocked-out HPV infection is significantly reduced as measured by reporter gene transduction. However, when S100A10 alone is knocked-out, only a moderate reduction in infection is observed (Taylor et al., 2018), emphasizing the importance of delineating the roles of monomeric AnxA2 versus heterotetrameric A2t.

EV71 is a causative agent of hand, foot, and mouth disease (HFMD), a common infection in infants and children that can sometimes lead to severe illness and long term neurological conditions (Chan and AbuBakar, 2004). EV71 is a non-enveloped single-stranded RNA (ssRNA) virus that enters cells through an unknown dynamin-independent pathway (Yuan et al., 2018). In an effort to understand initial host–virus interactions, AnxA2 was identified as a cell surface attachment factor through anti-EV71 immunoprecipitation and mass spectrometric analysis of infected cells in vitro (Yang et al., 2011). Using immunofluorescence microscopy, these authors also demonstrated that AnxA2 and EV71 colocalize at the cell surface, and that pretreatment with recombinant AnxA2 (rAnxA2) or antibodies against AnxA2 yields reduced infectivity. Results from this work showed that AnxA2 and EV71 colocalize at the cell surface, but they did not address if the reduction in infectivity was due to reduced binding, entry, or replication. Furthermore, their yeast two-hybrid experiments showed that EV71 capsid protein VP1 interacted with the C-terminus of AnxA2, which may also implicate A2t as serving a functional role. Future studies investigating AnxA2 or A2t in EV71 endocytosis could yield interesting results given the varied implications of AnxA2 in this process.

Infants and elderly can develop severe lower respiratory disease from infection by RSV (Nair et al., 2010) – an enveloped ssRNA virus. The mechanism of RSV cellular entry is disputed, with independent reports suggesting plasma membrane fusion, clathrin-dependent endocytosis, and macropinocytic mechanisms (Kolokoltsov et al., 2007; Collins and Graham, 2008; Gutiérrez-Ortega et al., 2008; Krzyzaniak et al., 2013). Fucoidan is a polysaccharide that inhibits RSV infection in vitro and in vivo. Because it is assumed that fucoidan works by binding to RSV receptors, Malhotra et al. (2003) employed solid-phase-immobilized fucoidan as an affinity matrix to isolate potential RSV-binding partners on epithelial cells and identified AnxA2 using mass spectrometry. The authors show that treatment with rAnxA2 reduced RSV infection as measured by fluorescent focus assay 24h post-infection, but similar to Yang et al. they did not investigate the mechanism of infection reduction beyond cell surface interactions. An independent study did, however, demonstrate that AnxA2 is not involved in virus assembly (Shaikh et al., 2012). As was the case with EV71, a more detailed analysis of RSV endocytosis has the potential advance our understanding of AnxA2-mediated endocytosis.

Cytomegalovirus infection in immunocompromised individuals or through congenital transmission can lead to serious diseases including pneumonia and hearing loss (Fowler and Boppana, 2018). CMV is an enveloped dsDNA virus that is able to establish life-long persistence, and multiple CMV entry mechanisms have been described. Interestingly, it has been hypothesized that the viral entry route may actually influence the outcome of infection (Murray et al., 2018). Early work first discovered AnxA2 on the surface of CMV particles isolated from human fibroblasts, and found that rabbit antiserum against AnxA2 inhibited CMV infection in vitro (Wright et al., 1994, 1995). Using synthetic membrane systems and rAnxA2, Raynor et al. (1999) demonstrated enhanced binding of CMV if rAnxA2 was present and attributed fusion events to A2t. Follow up studies elucidated that AnxA2 is not essential for CMV entry (Pietropaolo and Compton, 1999; Esclatine et al., 2001), however, viral gene expression and completion of viral life cycle are dependent on AnxA2 and A2t (Derry et al., 2007) and progeny virions have been shown to contain both forms of AnxA2 on viral envelopes (Wright et al., 1994). These findings together suggest multiple roles of both AnxA2 and A2t for CMV trafficking and progeny egress.

Chronic infection with HCV can lead to the development of liver cirrhosis and hepatocellular carcinoma. HCV is an enveloped ssRNA virus that enters the host cell through endocytosis, replicates in ER-replication complexes, and exits via exocytosis (Farquhar et al., 2012; Lindenbach and Rice, 2014; Benedicto et al., 2015). Many viruses express non-structural (NS) proteins to aid in efficient and successful infection; accordingly, NS proteins often function within viral replication complexes. Lai et al. (2008) investigated host factors that might interact with a specific NS protein complex of HCV (NS3/NS4A) known to interact with actin filaments in kidney epithelial cells. NS3/NS4A expression and co-immunoprecipitation followed by mass spectrometry identified AnxA2 as an interacting host factor (Lai et al., 2008). Given that lipid rafts were demonstrated to be involved in the formations of HCV RC complexes and because AnxA2 is associated with both lipid rafts and interacted with NS4A (Gokhale et al., 2005), the authors published a follow-up study asking if AnxA2 aids in the formation of HCV replication complexes (Saxena et al., 2012). Their report details the localization of AnxA2 at HCV replication complexes via immunofluorescence and immuno-electron microscopy, a reduction in the number of these structures following AnxA2 siRNA silencing, and a reduction in HCV RNA synthesis. The reduction in RNA synthesis, however, was measured via HCV replicase activity although there was no observed change in relative mRNA levels. An independent report conclusively demonstrated that although monomeric AnxA2 colocalizes with HCV NS proteins, AnxA2 silencing has no direct effect on HCV RNA replication but causes a significant reduction in intra- and extracellular virus titers (Backes et al., 2010). Based on these findings, the authors conclude that AnxA2 is involved in viral assembly as opposed to replication. Interestingly, overexpression of AnxA2 led to an enrichment of HCV NS proteins at replication complex sites (Saxena et al., 2012), a mechanism that may in fact promote virus assembly and support the claim of these authors.

Of the four types of influenza viruses, A, B, C, and D, influenza A viruses (IAV) and influenza B viruses cause epidemics of seasonal disease and respiratory infections. IAV type H1N1 and zoonotic avian IAV type H5N1 have both been associated with AnxA2 (LeBouder et al., 2008; Ma et al., 2017). IAV is an enveloped ssRNA virus that enters host cells via endocytosis, replicates in the nucleus, and buds from the plasma membrane for release (Salomon and Webster, 2009). It has been shown that AnxA2 and A2t are present on IAV H1N1 viral envelopes and that A2t, a plasminogen receptor, is responsible for the conversion of plasminogen to plasmin, a process involved in IAV replication (LeBouder et al., 2008). The authors demonstrated reduced viral titer after inhibiting plasminogen activation but did not tease out the precise involvement of AnxA2. An independent report investigated the role of AnxA2 in IAV H5N1 replication and found that silencing AnxA2 via siRNA inhibited viral protein expression and reduced progeny titer and proposed a mechanism by which AnxA2 bridged the gap between NS1 and p53, extending the amount of time cells could produce new virions (Ma et al., 2017). These data support the hypothesis that AnxA2 is involved in viral replication or assembly but do not preclude AnxA2 involvement during the preceding steps.

Measles is a highly contagious respiratory infection that is caused by MV – an enveloped ssRNA virus. After MV fuses with the host cell membrane, genome replication occurs in the cytoplasm and the virus is released by budding at the plasma membrane (Jiang et al., 2016). Knockdown of AnxA2 via shRNA in cervical epithelial cells (HeLa) caused reduced MV progeny virus generation 24 h post-infection, but did not affect MV entry and RNA replication (Koga et al., 2018). Normally, MV matrix protein (M protein) aids in connecting the viral capsid to the viral envelope and localizes to the plasma membrane where MV particles will form. Koga et al. (2018) went on to reveal that in the absence of AnxA2, M protein expression is decreased and mis-localized from the plasma membrane to the perinuclear space. Finally, the authors found that the observed M trafficking effect was due to monomeric AnxA2 versus A2t (Koga et al., 2018).