Toward a Shigella Vaccine: Opportunities and Challenges to Fight an Antimicrobial-Resistant Pathogen

Shigellosis causes more than 200,000 deaths worldwide and most of this burden falls on Low- and Middle-Income Countries (LMICs), with a particular incidence in children under 5 years of age. In the last decades, Shigella has become even more worrisome because of the onset of antimicrobial-resistant strains (AMR). Indeed, the WHO has listed Shigella as one of the priority pathogens for the development of new interventions. To date, there are no broadly available vaccines against shigellosis, but several candidates are being evaluated in preclinical and clinical studies, bringing to light very important data and information. With the aim to facilitate the understanding of the state-of-the-art of Shigella vaccine development, here we report what is known about Shigella epidemiology and pathogenesis with a focus on virulence factors and potential antigens for vaccine development. We discuss immunity after natural infection and immunization. In addition, we highlight the main characteristics of the different technologies that have been applied for the development of a vaccine with broad protection against Shigella.


Introduction
According to recent global estimates, diarrheal diseases cause 1.6 million deaths (all ages) and are the third leading cause of death among children younger than 5 years behind pneumonia and preterm birth complications [1]. Diarrheal diseases mainly affect lowand middle-income countries (LMICs), where accessibility to clean water, good nutrition, sustained sanitation, and healthcare are restricted and represent the leading risk factors of the disease. Among children under 5 years old, the three etiologies responsible for most of the deaths are Rotavirus, Shigella spp., and Cryptosporidium spp. [2], respectively, from viruses, bacteria, and protists ( Figure 1).
Here, we focus our attention on Shigella, the leading bacterial cause of diarrheal disease, and its virulence factors. Recently, the onset of antimicrobial resistance (AMR) has increasingly been associated with this pathogen [3], making the development of vaccines against Shigella an even greater global health priority. To date, no vaccines are widely available but different approaches are being evaluated in the clinic for the development of broadly protective vaccines. Here, we describe potential antigens that have been proposed to fight shigellosis and highlight the pros and cons of the different technologies applied for the development of a vaccine. Here, we focus our attention on Shigella, the leading bacterial cause of diarrheal disease, and its virulence factors. Recently, the onset of antimicrobial resistance (AMR) has increasingly been associated with this pathogen [3], making the development of vaccines against Shigella an even greater global health priority. To date, no vaccines are widely available but different approaches are being evaluated in the clinic for the development of broadly protective vaccines. Here, we describe potential antigens that have been proposed to fight shigellosis and highlight the pros and cons of the different technologies applied for the development of a vaccine.

Shigella Bacteria
Shigella are facultative anaerobic, non-motile, non-sporulating, rod-shaped, Gramnegative bacteria belonging to the family of Enterobacteriaceae. The bacteria are facultative intracellular pathogens that show a high specificity for the human host in which they cause shigellosis, commonly known as bacillary dysentery [4]. At the end of the 19th century, during an epidemic in Japan associated with high mortality, the microbiologist Kiyoshi Shiga isolated and characterized the bacteria causing such disease, which was later named Shigella [5].
The current classification divides the genus Shigella into four species based on serological typing: S. dysenteriae, S. boydii, S. flexneri, and S. sonnei [6]. These species are further distinguished into serotypes and subserotypes based on the specificity of the saccharide repeating units that form the O-Antigen (OAg) portion of lipopolysaccharides (LPSs). At present, the species S. dysenteriae is known to comprise 14 serotypes, S. boydii comprises 19 serotypes, S. flexneri comprises 15 serotypes and sub-serotypes, and S. sonnei comprises only one serotype [7]. Through different methodologies, it has been established that Shigella spp. belong to E. coli species [8][9][10]. Based on whole-genome sequencing, E.

Shigella Bacteria
Shigella are facultative anaerobic, non-motile, non-sporulating, rod-shaped, Gramnegative bacteria belonging to the family of Enterobacteriaceae. The bacteria are facultative intracellular pathogens that show a high specificity for the human host in which they cause shigellosis, commonly known as bacillary dysentery [4]. At the end of the 19th century, during an epidemic in Japan associated with high mortality, the microbiologist Kiyoshi Shiga isolated and characterized the bacteria causing such disease, which was later named Shigella [5].
The current classification divides the genus Shigella into four species based on serological typing: S. dysenteriae, S. boydii, S. flexneri, and S. sonnei [6]. These species are further distinguished into serotypes and subserotypes based on the specificity of the saccharide repeating units that form the O-Antigen (OAg) portion of lipopolysaccharides (LPSs). At present, the species S. dysenteriae is known to comprise 14 serotypes, S. boydii comprises 19 serotypes, S. flexneri comprises 15 serotypes and sub-serotypes, and S. sonnei comprises only one serotype [7]. Through different methodologies, it has been established that Shigella spp. belong to E. coli species [8][9][10]. Based on whole-genome sequencing, E. coli and all four Shigella species were found to share a common DNA backbone of approximately 3.9 Mb, interrupted by sequences specific to E. coli or Shigella. Comparative genomics indicate that entero-invasive E. coli (EIEC) is genetically more related to Shigella than to non-invasive E. coli [11,12]. E. coli and Shigella genomes mainly differ in the presence of many insertion sequence (IS) elements in Shigella, with a very dynamic genome [13,14]. The easy acquisition and loss of genes promote the success of Shigella as a pathogen because fast genetic adaptation, plus the acquisition of a single-copy virulence plasmid, allows bacteria to survive in different circumstances in the host [15,16].
To invade, replicate, and disseminate throughout the intestinal epithelium Shigella uses a range of bacterial effector proteins that are encoded by a large 220 kb plasmid called virulence plasmid or invasion plasmid (pINV), common to all Shigella species [17]. The pINV is composed of a 30 kb pathogenicity island (PAI) that encodes the type III secretion system (T3SS) and other virulence factors involved in adhesion and actin-mediated motility, such as IcsA/VirG. In the case of S. sonnei only, the pINV also encodes the serotype-specific OAg biosynthesis locus. It has been shown that the pINV is less stable in S. sonnei than in S. flexneri, especially at environmental temperatures. The pINV maintenance depends on a toxin-antitoxin (TA) system, consisting of a toxin and a related antitoxin that blocks its activity. Such a TA system is present in pINV from S. flexneri but not from S. sonnei [18].

Transmission and Pathogenesis
The transmission route for Shigella infections is fecal-oral, usually through close contact between people or through ingestion of contaminated food or water [19].
Clinical symptoms may be mild, such as watery diarrhea, but the disease can evolve into severe inflammatory bacillary dysentery with abdominal cramps, fever, nausea, anorexia, dehydration, and stools characterized by the presence of blood and mucus. Surprisingly, the inoculum size necessary for Shigella infection is as low as 10-100 bacteria [20]. The infection is non-systemic and enterically invasive, leading to the destruction of the colonic epithelium. Following oral ingestion, Shigella, after surviving the stomach's acidic environment and the competitive gut microbiota, arrives at the terminal ileum, colon, and rectum where it permeates the mucous layer. Shigella needs to survive multiple barriers to arrive at the epithelial surface, including microbiota-mediated colonization resistance, locally-produced antimicrobial peptides, and mucinases [21]. To establish an infection, Shigella transits across the colonic epithelial layer through M cells allowing translocation of bacteria across the epithelial barrier without damaging these cells [22] (Figure 2) and efficiently invades colonic epithelial cells from the basolateral side. Bacterial spread from cell to cell is the main step towards a severe inflammatory response [23] that leads to epithelium disruption, allowing the translocation of further bacteria [24].
Upon internalization, Shigella replicates and disseminates within the mucosal epithelium and causes relevant damage to the cellular cytoskeleton. The bacillus can be engulfed by macrophages and dendritic cells, which consequently undergo pyroptosis. The consequent host innate immune response activates the recruitment of neutrophils and the production of inflammatory cytokines, causing epithelial abscesses, ulcerations, and demolition, that further enhance Shigella invasion via a disrupted epithelial barrier [7]. Shigella's ability to endure intracellularly and avoid phagocytic killing depends on T3SS effectors that reduce inflammatory responses by inhibiting the host cell's pro-inflammatory signaling pathways and cytokine production. Examples of these effectors are OspG, which impairs NF-κB activation, and IpaH9.8, which is translocated into the nucleus of epithelial cells and dampens the expression of pro-inflammatory cytokines and chemokines [25]. These mechanisms together with serotype diversity affect the development of long-term, broadly effective protection [26,27].
S. flexneri (2a and 5a) strains have been the model organisms used to characterize Shigella virulence mechanisms [4]. However, each Shigella species and subspecies appear to have different virulence mechanisms and so deserve a specific analysis [23].

Epidemiology and Medical Needs
Shigella is a top cause of Moderate to Severe Diarrhea (MSD) throughout the world. Estimates report approximately 270 million cases with 212,438 total deaths per year with 30% in children younger than 5 years, especially in LMICs (Figure 3), and 2-7 cases of Shigella per 100 child-years require clinical care [28]. The incidence of Shigella infections is very low during the first six months of life, most probably because of maternal antibody titers and the low direct interaction with the external environment. Incidence increases after this age, peaking at 12-23 months and moderately decreasing afterward [29]. Upon internalization, Shigella replicates and disseminates within the mucosal epithelium and causes relevant damage to the cellular cytoskeleton. The bacillus can be engulfed by macrophages and dendritic cells, which consequently undergo pyroptosis. The consequent host innate immune response activates the recruitment of neutrophils and the production of inflammatory cytokines, causing epithelial abscesses, ulcerations, and demolition, that further enhance Shigella invasion via a disrupted epithelial barrier [7]. Shigella's ability to endure intracellularly and avoid phagocytic killing depends on T3SS effectors that reduce inflammatory responses by inhibiting the host cell's proinflammatory signaling pathways and cytokine production. Examples of these effectors are OspG, which impairs NF-κB activation, and IpaH9.8, which is translocated into the nucleus of epithelial cells and dampens the expression of pro-inflammatory cytokines and chemokines [25]. These mechanisms together with serotype diversity affect the development of long-term, broadly effective protection [26,27].
S. flexneri (2a and 5a) strains have been the model organisms used to characterize Shigella virulence mechanisms [4]. However, each Shigella species and subspecies appear to have different virulence mechanisms and so deserve a specific analysis [23].

Epidemiology and Medical Needs
Shigella is a top cause of Moderate to Severe Diarrhea (MSD) throughout the world. Estimates report approximately 270 million cases with 212,438 total deaths per year with 30% in children younger than 5 years, especially in LMICs (Figure 3), and 2-7 cases of Shigella per 100 child-years require clinical care [28]. The incidence of Shigella infections is very low during the first six months of life, most probably because of maternal antibody titers and the low direct interaction with the external environment. Incidence increases after this age, peaking at 12-23 months and moderately decreasing afterward [29]. Shigellosis rates in high-income countries (HICs) are much lower, at around 0.01 per 100 person-years among all ages and 0.02 per 100 person-years among children [30].  Shigellosis rates in high-income countries (HICs) are much lower, at around 0.01 per 100 person-years among all ages and 0.02 per 100 person-years among children [30].
Data from the Global Enteric Multicenter Study (GEMS), a 3-year, prospective, agestratified, case/control study focused on the investigation of the population-based burden, etiology, and adverse clinical consequences of acute MSD among children younger than 5 years living in sub-Saharan Africa and South Asia, established Shigella, identified by culture, as the most common etiology in children aged 24-59 months and the second most widespread etiological agent among children aged 12-23 months [31]. Among 1130 Shigella isolates, Shigella dysenteriae and S. boydii accounted for 5.0% and 5.4%, respectively, while S. flexneri accounted for 65.9% and S. sonnei accounted for 23.7%. Among all S. flexneri serotypes, five were predominant (that comprised 89.4% of S. flexneri) and included S. flexneri 2a, S. flexneri 6, S. flexneri 3a, S. flexneri 2b, and S. flexneri 1b [32].
Upon re-analysis of GEMS samples with quantitative PCR, the attributable incidence of Shigella more than doubled [29]. Shigella resulted as the major pathogen associated with dysentery (attributable fraction 63.8%) and was also the second most common pathogen associated with watery diarrhea (attributable fraction 12.9%) [29]. These findings indicate that the burden of shigellosis as defined using culture-based diagnostics is probably underestimated. The current worldwide epidemiological burden of shigellosis is mainly attributed to two species, S. flexneri and S. sonnei, which were traditionally associated with developing and developed regions, respectively [33][34][35]. Nevertheless, recent evidence indicates the emergence of S. sonnei in economically transitional states, effectively replacing S. flexneri and becoming the predominant shigellosis etiology [36]. This species replacement phenomenon has been documented several times in many countries in Asia, such as Vietnam [37], Thailand [38], and Bangladesh [39]. The reasons behind this serotype replacement are still under investigation; however, one hypothesis is that the competitive advantage of S. sonnei over S. flexneri is due to the exceptional ability of the former to acquire antimicrobial resistance genes from both commensal and pathogenic bacteria, especially in areas where the use of antimicrobials is poorly regulated. Furthermore, it has been reported that S. sonnei has an increased ability to grow successfully within Acanthamoeba present in public water supplies, which acting as host, allows the bacterium to persist in adverse environmental conditions [36,40].
The prevalence of Shigella species and serotypes changes from country to country, over time, and with economic status. Limited epidemiological data are available for S. boydii, which is uncommon in many regions of the world, with exceptions being the Indian subcontinent [41] and Latin America [42,43]. In India, a heterogeneous distribution of Shigella species and serotypes [44] has been reported between 2001 and 2003, with S. flexneri being the most common (45%) followed by S. dysenteriae (29.4%), S. boydii (14.7%), and S. sonnei (10.2%). A dramatic change was observed in the same study conducted in South India between 2003 and 2006, where S. flexneri remained the most prevalent (45%) followed by S. sonnei (31%), S. boydii (15%), and S. dysenteriae (8%). In Andaman and Nicobar Islands, the S. flexneri 2a serotype was identified to be predominant [45]. A heterogeneous distribution of Shigella serotypes was found in cases from urban and rural areas of Bangladesh [46]. S. dysenteriae, specifically S. dysenteriae 1, caused many cases of fatal dysentery epidemics since its first isolation in 18,975. However, this species is rarely isolated in current surveillance. Improvements in sanitation and anti-microbial access have been proposed to have caused a decline in illness and death related to Shigella spp., including S. dysenteriae [47,48].
A unique epidemiological niche for Shigella emerged in the 1970s as a sexually transmitted infection among men who have sex with men (MSM). In 2009, serotype S. flexneri 3a appeared in England and Wales among MSM and rapidly spread within the MSM population to regions considered at low risk for shigellosis, causing an epidemic.
Around 536 million people travel to LMICs annually and 10-40% acquire diarrhea. It has been estimated using faecal cultures that 2-9% of travelers' diarrhea is attributable to Shigella [49]. Repercussions of shigellosis during travel are the introduction of antimicrobialresistant Shigella into new populations [50] and outbreaks in tourist groups and military settings. An updated review, that includes publications between January 1990 and June 2015, reveals that travelers' diarrhea continues to be a major medical concern for deployed US military personnel and other long-term travelers, with an incidence of over 30 cases per 100 person-months [51]. Furthermore, this phenomenon increases the possibility of reinfections in areas where other serotypes predominate.
Controlling the burden of Shigella is challenging for many reasons. First, Shigella needs a low infectious dose for transmission through the fecal-oral route, and the possibility of reinfection due to the variety of Shigella species and serotypes coexisting in the same areas is high. In addition, the need for antibiotic therapy, not just oral rehydration, further complicates treatment and paves the way for the emergence of multidrug-resistant strains of Shigella. This threatens the administration of effective, affordable treatments and highlights the importance of infection prevention [52,53].

Antibiotic Resistance of Shigella
Shigella has been listed by the WHO among those pathogens for which the development of new interventions is a priority [54]. The use of antibiotics for the treatment of common bacterial infections plays an important role in reducing the burden of the disease. Nevertheless, antibiotic overuse to treat diarrhea increases antibiotic resistance [55]. Shigella spp. is resistant to most antibiotics, and drug treatment is expensive, time-consuming, and sometimes problematic, especially in areas where medical care is limited [44,56].
After sulfanilylguanidine was shown to be bacteriostatic against Shigellae [57] in the late 1930s (Figure 4), sulfonamides became the treatment of choice for shigellosis [58]. Favorable clinical and bacteriological results were achieved also with the use of oral chloramphenicol [59]. Tetracycline was used routinely in the 1950s and 1960s, but it was phased out because of widespread resistance and the side-effect on the discoloration of teeth in children [60]. Shigella developed resistance to all of these drugs, so the treatment shifted to ampicillin and co-trimoxazole. However, treatment recommendations were again changed to nalidixic acid because Shigella developed resistance to the former drugs. Later, when resistance capacity to nalidixic acid occurred [48], fluoroquinolones were introduced. However, fluoroquinolone-resistant strains have been isolated from various sources [61]. At the moment, the WHO recommends ceftriaxone, azithromycin, and pivmecillinam for treatment of infection by fluoroquinolone-resistant Shigella spp. [44]. Unfortunately, ceftriaxone-resistant and azithromycin-resistant strains have already been isolated in some places [62].
introduced. However, fluoroquinolone-resistant strains have been isolated from various sources [61]. At the moment, the WHO recommends ceftriaxone, azithromycin, and pivmecillinam for treatment of infection by fluoroquinolone-resistant Shigella spp. [44]. Unfortunately, ceftriaxone-resistant and azithromycin-resistant strains have already been isolated in some places [62]. About half of the Shigella strains are now resistant to several drugs [63] in many areas of the world, and this has obviously become a matter of huge concern. Bacteria can exploit several antibiotic-resistance mechanisms, among them are the use of efflux pumps to extrude the drugs, the change in membrane permeability, and the expression of enzymes able to modify and inactivate the molecules [56,[64][65][66][67]. For example, β-Lactam antibiotics, a class of antibiotics characterized by a β-lactam ring in their molecular structure, inhibit About half of the Shigella strains are now resistant to several drugs [63] in many areas of the world, and this has obviously become a matter of huge concern. Bacteria can exploit several antibiotic-resistance mechanisms, among them are the use of efflux pumps to extrude the drugs, the change in membrane permeability, and the expression of enzymes able to modify and inactivate the molecules [56,[64][65][66][67]. For example, β-Lactam antibiotics, a class of antibiotics characterized by a β-lactam ring in their molecular structure, inhibit cell wall biosynthesis by acting on penicillin-binding proteins. Furthermore, the mutation or absence of porin in the outer membrane of Gram-negative bacteria such as Shigella spp. decrease the penetration of β-lactams, such as aztreonam and dianionic moxalactam, and also the susceptibility of hydrophilic antibiotics, such as penicillin and piperacillin [64,68]. Specifically, resistance to β-lactam antibiotics is associated with modifications on OmpF and OmpC and to OmpR, which act as a transcriptional regulator [64]. In addition, LPS has been recognized as an important key player in antibiotic resistance as it is essential for the assembly of trimeric PhoE porin, confers colicin E2 resistance in S. flexneri strains [69], and has been linked with the rise in resistance toward imipenem in S. dysenteriae [70].
There is a pressing need for the development of effective strategies to fight AMR Shigella infections [71], and some therapeutic approaches have been proposed including the use of natural products, nanoparticles (NPs), and phage therapy in addition to vaccine strategies [72]. Studies have shown the efficacy of natural small molecules produced by microbes, plants, and animals for treating Shigella infections. Probiotics have been suggested as prophylaxis for antibiotic-induced diarrhea (e.g., for travelers) and are also an alternative therapeutic choice for the treatment of gastroenteritis infections [73]. NPs are also gaining more attention as they have shown a broad spectrum of antibacterial activity against pathogenic bacteria, due to their bactericidal characteristics [74]. In particular, iron oxide nanoparticles (IONPs) destroy bacterial targets by affecting membrane cell integrity and by generating free oxygen radicals. Bacteriophages kill bacteria through a lysis mechanism. In addition, phage therapy has shown to be advantageous, especially for its specificity in targeting bacterial pathogens without effecting the normal human microflora. Phages replicate at the infection site, and they have a strong bactericidal effect on antibiotic-resistant strains and fewer side effects than other therapies. However, phages are self-limiting as they persist at a shallow level on target sites [75].
Vaccines can play a major role in fighting AMR. They are used prophylactically, decreasing the number of infectious disease cases, and thus antibiotic use and the emergence and spread of AMR [76]. Different candidate vaccines have been proposed to prevent infection by Shigella spp., most of which are currently being evaluated for safety and immunogenicity in clinical trials.

Structure and Role of Shigella O-Antigens
Here, we focus on the structures and role of OAg from S. sonnei and S. flexneri, which are the most epidemiologically relevant Shigella serotypes. Gram-negative bacteria, such as Shigella, display LPS molecules on the cell surface ( Figure 5). The LPS consists of a polysaccharide chain of repeating units, called OAg, linked to a core oligosaccharide. The core is anchored to the membrane through a lipid A moiety.
In Shigella, as in all enterobacteria, the conserved inner-core part is located closer to lipid A and consists of two residues of 3-deoxy-D-manno-2-octulosonic acid (KDO) and three residues of L-glycero-D-manno-heptose (Hep). Interestingly, the Shigella inner-core is identical to that of E. coli and belongs to the so-called Salmonella type. This is characterized by phosphorylation or glycosylation by α-GlcN or α-GlcNAc of the heptose residues. Moreover, the phosphate group on Hep I is often non-stoichiometrically phosphorylated by 2-aminoethylphosphate (PEtN), forming P-PEtN. The outer-core part (connecting the inner-core to the OAg) has instead a less conserved structure. However, R1 outer-core types have been found in S. sonnei and S. flexneri 6, while R3 outer-core types have been found in all S. flexneri serotypes [77] (Figure 5).  All S. flexneri serotypes 1-5, 7, X, and Y, but not serotype 6, share a linear tetrasaccharide backbone consisting of the following OAg repeating unit: (2)-α-L-Rhap III -(1→2)-α-L-Rhap II -(1→3)-α-L-Rhap I -(1→3)-β-D-GlcpNAc-(1→) n , variably O-acetylated and/or glucosylated. Phosphorylation with PEtN has been classified into subtypes 4av, Xv, Yv, and Yv1 and described as "variant" subtypes by adding the letter "v" to the names of the corresponding PetN-negative subtypes [78]. S. flexneri serotype 6 is instead characterized by a linear polysaccharide backbone (2) The OAg biosynthesis depends on the Wzx/Wzy pathway. The inner membrane (IM) represents the site of glycan biosynthesis, where the Wzx proteins allow each single repeating unit produced by the glycosyltransferases to translocate from the cytoplasmic to the periplasmic side. After transportation of the repeating unit to the periplasm by Wzx, these are polymerized by Wzy into OAg chains, whose length is regulated by the Wzz family of proteins. The OAg is then ligated to a sugar residue of the core oligosaccharide by the WaaL ligase to form the complete LPS, which is then transported to the outer membrane (OM) by the Lpt complex. The OAg repeating units can also be polymerized into capsular polysaccharides (group 4 capsule, G4C) if, in addition to the wzx-wzy cluster, an additional G4C operon is present [82][83][84]. Formation of G4C has been reported for S. sonnei and S. flexneri 6 strains [83][84][85].
LPS molecules on the bacterial surface have a significant role in the pathogenesis of Shigella, including the protection of the bacteria from the lytic action of serum complement and the promotion of adhesion and internalization of bacteria to intestinal epithelial cells [86]. The diversity of the S. flexneri OAg is considered an important virulence factor that enhances the survival of pathogens. In addition, modifications such as serotype-specific glucosylation promote invasion of S. flexneri into host cells [87].

Structure and Role of T3SS Proteins
The weaponry that Shigella employs for tissue invasion and an intracellular lifestyle is encoded by a single DNA fragment of approximately 30 kb placed in the virulence plasmid [88]. The central element of this machinery is the type III secretion system (T3SS) [14] that forms a macromolecular needle-like structure across the membrane of the target eukaryotic cell and enables the bacteria to translocate at least a set of different effector proteins from their bactplasm directly into the eukaryotic host cells. Once injected, these virulence factors affect host cell function, support infection, and help the bacteria to evade host cell immune responses [4,21,89,90].
The complexity of the T3SS structure and composition has been deeply investigated over the years [91]. Briefly, in the structure of T3SS, it is possible to identify three elements: a cytosolic complex that forms the so-called "sorting platform", the transmembrane basal body crossing the inner and the outer membranes, and the extracellular needle with the tip complex that protrudes on the bacterial surface. At the distal end, the tip is composed of IpaB, IpaC, and IpaD [92]. IpaD facilitates the assembly of IpaB and IpaC into the needle, and IpaB and IpaC are hydrophobic proteins able to form a pore in eukaryotic membranes, thus allowing for effectors delivery [92,93]. Membrane insertion and T3SS activity are promoted by the interaction of IpaB with cholesterol in the host membrane [94,95]. The needle has a helical symmetry resulting from the homopolymerization of approximately 100 MxiH subunits, and it has been proposed to have a role in propagating secretion signals [96]. Indeed, specific amino acid changes in MxiH are implicated in controlling the amount and type of the secreted substrates. However, it is still not clear how such signals would be received at the tip of the MxiH assembly.
The needle is inserted in the basal body consisting of an outer ring (composed of the secretin MxiD) that contacts the inner ring (composed of MxiG). Just below the basal body, there is the sorting platform with its major component Spa33 [97]. Spa33 interacts with the hub of the sorting platform that is formed by the stalk Spa13 [98,99] and Spa47, and this complex has been suggested to bind chaperones to help guide the secretion cargo, including IpaB/IpaC [100]. IpaB has been shown to be the first translocator protein recruited to the needle tip.
Multiple physiological factors (such as cell contact, media composition, serum, temperature, and pH) as well as artificial inducers (such as the Congo Red dye) have been shown to regulate the expression of T3SS genes through transcriptional activators belonging to the AraC family [101,102].
The T3SS has uniquely evolved to upregulate Shigella virulence in response to environmental levels of the bile salt deoxycholate (DOC), essential for proper dietary fat solubilization and intestinal absorption [103]. As other enteric pathogens, Shigella encounter high concentrations of DOC in the small intestine and in the colon. Bile enhances the expression of the Shigella T3SS effectors ospE1/ospE2 as adhesins [104]. Furthermore, the T3SS tip protein IpaD binds DOC and undergoes conformational changes supporting the recruitment of IpaB [105][106][107]. Upon contact with host membrane components, IpaC is recruited to the tip of the T3SS, completing the assembly of the translocon pore in the host cell membrane and promoting injection of effectors into the host cell and pathogen entry [95].

Role of Other Virulence Factors
The ability to adhere to host cells is extremely important for intracellular pathogens, and adhesins are a diverse class of virulence factors [108]. Few studies have examined the molecular mechanisms of Shigella adhesion. Shigella can spread intercellularly using actin-based motility (ABM), a process in which the host cell cytoskeleton is used by the bacterium to spread from cell to cell and to avoid extracellular immune defenses [109]. ABM is mediated by the autotransporter protein IcsA (VirG), whose secretion is independent of T3SS. IcsA is located in the outer membrane with asymmetric conformation through the action of several proteins such as DegP, Skp, and SurA [110,111], and newly formed IcsA is preferentially incorporated into the old pole of the bacterium. Polar localization of IcsA is critical to control the unidirectional movement of the bacterium and is maintained by IcsP, OmpA, the chaperon DnaK, and the enzymes YidC and Apyrase [112][113][114][115][116]. In particular, IcsP is an outer membrane protease that cleaves IcsA, and both of them are encoded by the pINV. S. flexneri IcsP mutants show reduced amounts of secreted IcsA passenger domain in culture supernatants and display IcsA across the entire cell surface. The expression of IcsP is finely tuned by the Fur/RyhB regulatory pathway in response to the low concentrations of free iron, a condition encountered within colonic epithelial cells [117].
Interestingly, Zychlinsky et al. [118] used S. flexneri ∆ipaB and ∆ipaD mutants as a tool to study adhesion processes and found that these mutants hyperadhere to a variety of host cells, surprisingly, using one pole in contrast with the wild-type that contacted host cells using the long axis. They demonstrated that this hyperadhesion phenotype was due to IcsA, which is localized in a polar manner. Moreover, treatment with the bile salt DOC induced IcsA-dependent adhesion in a T3SS-dependent manner, showing that IcsA becomes adhesive in response to physiological stimuli.
In addition to IcsA, the multivalent adhesion molecule (MAM) has also been identified as an important adhesin during the earliest stages of infection [119]. MAM is non-redundant with IcsA, so both adhesins may be required during the attachment, and MAM-mediated adherence is a prerequisite for T3SS-mediated invasion of host epithelial and phagocytic cells and necessary for Shigella pathogenicity in vivo [119].
Mahmoud et al. reported that DOC negatively regulates MAM and IcsA in S. sonnei, in contrast with what was reported for S. flexneri. The author speculated that the differential effect of DOC on S. flexneri and S. sonnei may explain the lower virulence of S. sonnei, which indeed requires a higher infectious dose to cause disease [20].

Immunity to Shigella
Observational and challenge studies suggest that immunity to Shigella spp. is serotypespecific [120][121][122][123], thus supporting the development of multivalent OAg-based vaccines. In particular, phase 3 studies with the NIH S. sonnei glycoconjugate vaccine found a strong association between serum OAg IgG and protective efficacy [124][125][126], and a recent analysis performed by Cohen and colleagues confirmed anti-OAg IgG as potential correlate of protection against shigellosis [127]. After oral administration of live attenuated Shigella vaccines and natural infection with wild-type Shigella, fecal OAg IgA levels seem instead to correlate with protection, very likely as a result of mucosal immunity in contrast to systemic immunity induced by parenteral vaccines. There are also data suggesting that surface protein antigens, in particular the Ipa proteins, can play a role in protection despite being common to multiple serotypes [128]. Recent data from Shigella CHIM studies suggest that anti-IpaB IgG could also be a correlate of protection [129]. It seems that multiple factors can contribute to protection and different immune mechanisms could be required to elicit protective immunity depending on the Shigella serotype [130].
In the absence of a correlate of protection, it is important to investigate not only the absolute level of antibodies elicited, but also their functionality. In particular, serum bactericidal assay (SBA) has been proposed for this scope [131,132]. A recent study found a strong association of specific SBA titers against S. flexneri 2a in adults with a reduction in disease post-challenge with wild-type organisms [133], supporting the value of this assay to predict vaccine efficacy. Furthermore, bactericidal activity has been detected in sera from individuals naturally exposed to Shigella and antibody-dependent complement-mediated killing has been associated with protection against other bacterial pathogens [134].
Some studies have also looked at cellular markers of immunity, suggesting that antibody secreting cells, memory B cells, and CD4 + and CD8 + T cells could be additional readouts to be considered [135][136][137].
Since Shigella is a human-restricted pathogen, animal infection models currently developed to assess the potential protective efficacy of candidate vaccines, including a murine pulmonary model [138], a cynomolgus monkey S. dysenteriae 1 model [139], a guinea pig keratoconjunctivitis model [140], and a guinea pig and piglet oral and intrarectal challenge model [141] do not mirror human infections well. For this reason, testing candidate Shigella vaccines in humans seem to be the most valuable way to assess their efficacy. Aligned with this, Shigella CHIMs have been established, using either the S. sonnei strain 53G or the S. flexneri 2a strain 2457T, to provide preliminary estimates of vaccine efficacy in humans [142][143][144]. However, CHIM trials are conducted in adults in high-income countries and the appropriateness to evaluate vaccine efficacy in infants in LMICs based on them is controversial. Furthermore, the conditions of a challenge study are artificial, with higher doses of the challenge agent administered following neutralization of gastric acidity as compared with natural infection [145]. However, CHIMs also provide the opportunity to conduct in-depth immune response characterizations pre-and post-vaccination or pre-and post-infection. Many vaccines based on different technologies are currently in the clinic and moving them to proof of concept phase 2b/phase 3 trials could be important to obtain more data and to better understand key contributors to protective efficacy. These results will also help understand how preclinical immune data can be predictive of human data, supporting the development of current and future vaccines.

Status of Vaccines in Development
The heterogeneous distribution of Shigella serotypes across countries implies that multivalent cross-protective vaccines are required to address the burden of shigellosis. According to GEMS data, a 4-component vaccine consisting of S. sonnei and S. flexneri 2a, 3a, and 6 OAg could confer direct protection of 64% globally, which could increase to > 85% because of cross-protection against heterologous S. flexneri serotypes [32,46].
The need for multivalency adds complexity to the development of a vaccine [146] and, at present, there are no licensed vaccines widely available against Shigella [147].
Only two Shigella vaccines are on the market, but these are limited to Russia (Shigellvak, solution of LPS extracted from S. sonnei bacteria) and China (FS, a live attenuated, oral bivalent S. flexneri 2a and S. sonnei vaccine developed at the Lanzhou Institute of Biological Products). Over the years, potential candidates have included killed, live attenuated, and subunit vaccines.
The earliest vaccines were based on the whole-cell formalin or heat killed approach [148], but these suffered from high levels of reactogenicity. More recently, the Walter Reed Army Institute for Research (WRAIR) developed formalin-inactivated S. sonnei (SsWc) [149] and S. flexneri 2a (Sf2aWC) [150] monovalent whole-cell vaccine candidates, which were well tolerated in a phase 1 trial, though immune responses were variable. A trivalent version was then developed by PATH and WRAIR [151] containing S. flexneri 2a and 3a as well as S. sonnei, but these were never tested in clinical trials. A Heat Killed Multi Serotype Shigella (HKMS) vaccine is being developed by India's National Institute of Cholera and Enteric Diseases (NICED) [152] and contains S. flexneri 2a, S. flexneri 3a, S. flexneri 6, S. sonnei, S. dysenteriae 1, and S. boydii 4. This kind of approach has now been largely abandoned.
In the 1960s, Mel and colleagues of the Yugoslav Army developed a live attenuated vaccine that showed protection in military recruits and children aged 2-8 years [153,154]. However, the major challenges of this technology were the need for a high number of doses, relatively short-lived protection, and manufacturing issues, so these vaccines were never licensed. More recently, progresses in whole genome sequencing allowed the development of well-defined live attenuated vaccines with targeted genetic mutations, but balancing between acceptable levels of reactogenicity and sufficient immunogenicity remained a challenge. Among the different vaccine candidates developed, the Center for Vaccine Development (CVD) at the University of Maryland developed a number of guaBA-based live attenuated candidates using the S. flexneri 2a 2457T strain. In particular, CVD 1208S moved to a phase 2 challenge trial, but after the recruitment of 20 subjects, the study was stopped because of reactogenicity issues [155]. SC602, developed by the Pasteur Institute using virG and iuc deletions from S. flexneri 2a 494 wild-type strain, was mildly reactogenic and immunogenic in US adults and conferred protection against fever, dysentery, and severe symptoms after challenge [156], but it induced a low immune response when tested in adults and children in Bangladesh [157]. WRAIR developed a series of virG-based live attenuated vaccines using the wild-type Moseley S. sonnei strain (having the plasmid encoding for the OAg relatively stable). The first-generation WRSS1 vaccine resulted as being immunogenic but reactogenic when tested among US and Israeli adults in phase 1 and phase 2 studies [158,159]. WRSS1 was better tolerated in a phase 1 study in Bangladesh, but immune responses were modest, requiring multiple doses and short duration [160]. Moreover, attenuating mutations were introduced to overcome the safety issues resulting in the WRSS2 and WRSS3 candidate vaccines. While both have deleted enterotoxin genes senA and senB, WRSS3 has msbB deleted, resulting in a less acylated and therefore less reactogenic lipid A [161]. Both these vaccines were well tolerated and immunogenic in a phase 1 trial in the US [162] (Table 1).  [174] Several candidates are in preclinical development [175], and recent studies are evaluating Shigella genetically modified bacteria, either retaining one OAg repeating unit on the bacterial surface (∆wzy; Truncated Shigella, International Vaccine Institute, IVI and PATH) [163] or not displaying LPS-OAg at all (∆rfbF; Eveliqure's ShigETEC) [164] (Table 1). Both approaches result in increased exposure of well-conserved outer membrane proteins, with the potential to provide broader coverage against multiple Shigella serotypes. An innovative vaccine approach, under development by Protein Potential, uses the Ty21a typhoid vaccine displaying Shigella LPS; unfortunately, inconsistency in production caused variability in immunogenicity and protection in controlled human infection models (CHIM) studies [165,176].
Parenteral subunit approaches target instead specific Shigella antigens. Many of these candidate vaccines are OAg-based. Proof of efficacy was demonstrated 25 years ago with an S. sonnei OAg chemically conjugated to a recombinant exotoxin of Pseudomonas aeruginosa (rEPA) developed by John Robbins at the U.S. National Institutes of Health (NIH) [125,177,178]. Indeed, the vaccine conferred 74% protection in Israeli adults after a single dose [124]. In 2010, protection was also demonstrated in Israeli children aged 3-4 years, but the vaccine failed to protect the younger population [125].
More recently, recombinant glycoconjugates produced in genetically engineered E. coli have been proposed, with a bioconjugate against S. flexneri 2a, developed by LimmaTech Biologics, proving to be immunogenic in phase 1 clinical trial [179]. In a following CHIM study in US adults, the vaccine did not meet the primary endpoint of protection against all forms of shigellosis but resulted in being protective against more severe shigellosis in a post hoc analysis, and the protection was associated with anti-OAg specific IgG response. Such results allowed further development of a 4-component formulation, S4V-EPA, made of rEPA bioconjugates of S. sonnei and S. flexneri 2a, 3a, and 6, that is currently completing an age-descending dose-finding phase 2 trial in Kenya [166] ( Table 1).
The Institute Pasteur has instead developed a well-defined synthetic glycoconjugate vaccine made of synthetically produced S. flexneri 2a oligosaccharides chemically linked to tetanus toxoid (TT) carrier protein [177], which was demonstrated to be safe and immunogenic in a phase 1 study in Israeli adults even after a single dose [180]. Sf2a-TT synthetic conjugate is now being tested in an age-descending dose-finding phase 2 trial in Kenya and in a CHIM trial at the CDC [167] ( Table 1).
The traditional glycoconjugate approach has been used by Beijing Zhifei Lvzhu Biopharmaceuticals for the development of a bivalent vaccine, ZF0901, made of S. sonnei and S. flexneri 2a OAg conjugated to TT. After promising results in phase 1 [181] and phase 2 trials, ZF0901 is currently being tested in a phase 3 study [168] (Table 1).
An interesting alternative subunit vaccine approach has been proposed by WRAIR [169] combining Shigella LPS to Ipa proteins. Native Invaplex (InvaplexNAT) contained a complex of IpaB, IpaC, IpaD, and LPS extracted from wild type S. flexneri 2a. Intranasal immunization resulted to be well tolerated and immunogenic in phase 1 studies [182,183] but failed to protect in a CHIM trial [184]. Artificial Invaplex (InvaplexAR) was then developed using recombinant IpaB and IpaC produced in E. coli and LPS extracted from S. flexneri 2a: it was well-tolerated but not consistently immunogenic in a phase 1 study [169]. A new version, the artificial detoxified Invaplex (InvaplexAR-Detox), contains LPS extracted from an S. flexneri 2a strain with deleted msbB genes. This detoxified form of LPS has allowed its parenteral administration with good safety and immunogenicity results in a recent phase 1 study [169] (Table 1).
GSK Vaccines Institute for Global health (GVGH) has proposed GMMA, outer membrane vesicles (OMVs) from bacteria genetically mutated to increase yields and reduce LPS reactogenicity, for OAg delivery [185]. GMMA resemble the outside of bacteria, where protective antigens, including proteins, are normally found. Moreover, GMMA contain pathogen-associated molecular patterns conferring self-adjuvanticity. The S. sonnei monovalent GMMA candidate 1790GAHB has been tested in phase 1 studies in France and the UK [186] and has subsequently been tested in Kenyan adults [187], showing to be well-tolerated and immunogenic and able to boost bactericidal anti-OAg IgG response in naturally exposed or pre-vaccinated subjects [186,188,189]. However, the vaccine failed to elicit protection in a CHIM study on US adults [190]. An improved version of S. sonnei GMMA, characterized by higher OAg density, has then been developed and combined with GMMA from S. flexneri 1b, 2a, and 3a in a four-component formulation (altSonflex1-2-3) currently being tested in phase 1-2 trials [170] (Table 1).
In addition, the University of Navarra in Spain is developing an acellular Shigella vaccine candidate based on OMV encapsulated in nanoparticles, which is currently being tested at the preclinical level [171].
Protein-based subunit vaccine candidates are also under development. They may offer broad protection against all major serotypes but have only been tested in animals so far. Among these, the DB Fusion, developed by PATH, consists of a genetic fusion of the T3SS proteins IpaB and IpaD [172] co-administered intradermally with double mutant Heat Labile Toxin (dmLT) from ETEC with the aim to induce both mucosal and systemic immunity. A 34 kDa OmpA outer membrane protein developed by NICED could provide protection according to preclinical data [173]. IcsP, also named pan-Shigella surface protein 1 (PSSP-1), was also suggested as a promising antigen for a broadly protective vaccine against Shigella [174].

Conclusions and Future Perspectives
Vaccination represents an unquestionable solution against infectious diseases, having prevented 700 million cases and more than 150 million deaths during the last century [191]. The Expanded Program on Immunization (EPI) was introduced in 1974 [192], targeting all children throughout the world against different diseases such as diphtheria, tetanus, pertussis, measles, poliomyelitis, tuberculosis, Haemophilus influenzae type b, rotavirus, and pneumococcus. In addition, since 2000 the Global Alliance for Vaccines and Immunization (GAVI) has helped to increase childhood immunization coverage in poor countries by supporting the introduction of new vaccines [193]. The latest Weekly Epidemiological Report from the WHO Strategic Advisory Group of Experts on Immunization (SAGE) reported the successful introduction of 128 vaccines since 2010 in 86 LMICs [194]. Furthermore, vaccination appears as one of the most promising interventions to fight multidrug-resistant bacterial pathogens effectively [76].
Despite the many years that vaccines have been developed and tested in the clinic against Shigella, no licensed vaccine is widely available. Shigella is the major bacterial cause of childhood diarrheal death globally, and lack of proper hygiene and sanitation results in repeated diarrheal cases often associated with cognitive impairment and stunting [195,196]. Diagnostic tools and antibiotics are expensive, limited, or not available in LMICs, and according to the current WHO diarrhea treatment guidelines, antibiotics can be given only in case of dysentery, meaning that many children in LMICs with shigellosis do not get antibiotics. Furthermore, as highlighted by the WHO and CDC, the treatment of Shigella is increasingly challenging due to the spread of antimicrobial-resistant circulating Shigella strains [54,197,198]. All of these factors make the development of a vaccine against Shigella a high priority. Because Shigella immunity appears to be serotype-specific, the OAg has been proposed as the main vaccine target. The need for multivalency also makes the development of a vaccine against Shigella more complex.
Evidence of efficacy was obtained by Mel and colleagues with a live attenuated vaccine, and for many years, efforts have been concentrated on this approach, but live attenuated vaccine candidates suffer from reactogenicity and/or insufficient immunogenicity. Furthermore, these kinds of vaccines require multiple doses and are characterized by a short-lived duration of protection. Later on, efficacy was demonstrated with an S. sonnei OAg glycoconjugate vaccine, but the vaccine failed in children below 3 years of age that are the main target for the Shigella vaccination.
Recent improvements in the manufacturing of well-defined glycoconjugate vaccines (including bioconjugation and synthetic approach), use of alternative delivery systems (as it is the use of GMMA), and combination of LPS with potentially broad protective protein antigens (e.g., Invaplex) could change the scenario and open the way to the licensure of a Shigella vaccine in the near future. Vaccine candidates are now approaching phase 3 clinical trials in infants in endemic countries, and the comparison of different technologies could help identify a correlate of protection. Field efficacy [199] trials should be feasible as the incidence of Shigella infection is relatively high in endemic regions, and a quantitative PCR assay has been developed to be more sensitive than traditional culture methods. Discussions with regulatory authorities will be beneficial to maximize the success of these trials and accelerate vaccine implementation. These results will also help to understand how preclinical immune data can be predictive of human data, supporting the development of current and future vaccines.