Plague, caused by the gram-negative bacterium Yersinia pestis, is notorious for its involvement in three of the seven deadliest pandemics recorded in global history, including the recent COVID-19 pandemic. The three historic plague pandemics, the most infamous of which was the Black Death of the Middle Ages, collectively caused an estimated 200 million deaths1,2. Unfortunately, plague is still an endemic disease in parts of the world, with outbreaks being reported to the WHO from over 33 countries including Madagascar, Democratic Republic of the Congo (DRC), India, China, Peru, and occasionally, the south-western USA3. In these regions, disease is maintained by the existence of infected animal (mostly rodent) reservoirs of Y. pestis4,5.

Transmission to humans is predominantly promoted by flea bite; those fleas having fed on infected rodents (Fig. 1). However, Y. pestis is an obligate parasite and even if the rat population is reduced, the organism can infect mice, prairie dogs, rabbits, and members of the cat family, including the domestic cat4,5,6.

The figure depicts various routes for the flea-vectored transmission of plague to man. The figure is reproduced from Williamson and Westlake (2019)8 with permission (License 5753521304382, Oxford University Press).

Infection through flea bite causes bubonic plague, which if undiagnosed, can develop into a septicemic infection or secondary pneumonic plague. Pneumonic plague is highly contagious requiring prompt antibiotic therapy for survival, as the mortality rate approaches 100% if untreated6,7,8. In pneumonic plague, Y. pestis can be transmitted to healthy individuals who are in close contact by respiratory droplets, establishing further cases of primary pneumonic plague, leading to disease outbreaks which may transform into epidemics and pandemics6.

With this epidemiology, poor living conditions augment the endemicity of plague, which requires close contact with a rodent population. However, in endemic regions such as Madagascar, the lack of an approved vaccine means that outbreaks have to be controlled by antibiotic therapy, administered to the patients and those in immediate contact with the infected individuals. Whilst timely antibiotic therapy is effective in treating the infection, case fatality rates still reached up to 8.6% during the 2017 Madagascar outbreak despite aggressive antibiotic therapy9. Additionally, there is a demonstrable risk of the development of antibiotic resistance. Indeed, antimicrobial resistant (AMR) Y. pestis strains have been identified in Peru and Madagascar10,11. Therefore, there is a clear need for a safe, effective, and licensed vaccine for use in endemic regions to control or prevent infection, as well as to protect military and civilians at large from potential biothreat attacks.

Y. pestis has evolved from the relatively mild gastrointestinal Y. pseudotuberculosis (notably serotype 1b) between 1500 and 20,000 years ago12, although archaeological evidence has suggested that the plague-causing bacterium existed long before previous estimates13.

The evolution of Y. pestis has resulted in the inactivation of genes required for an enteric lifestyle and by the acquisition of plasmids encoding new virulence factor-encoding genes14. In common with other pathogenic yersiniae (e.g., Y. pseudotuberculosis and Y. enterocolitica), Y. pestis possesses a 70-kilobase (kb) virulence plasmid designated as pYV/pCD1 that carries a Type III secretion system (TTSS) operon15,16. However, Y. pestis has acquired two additional unique plasmids, including a 9.5-kb pPCP1/pPla/pPst encoding a bacterial surface-bound protease (plasminogen activator, Pla), which has potent fibrinolytic activity1. In addition, this plasmid possesses pesticin and coagulase encoding genes which enable bacterial transmission from the flea17. The other 100–110 kb pFra/pMT1 plasmid18 codes for two important proteins, Fraction 1 (F1) antigen and a phospholipase D known as murine toxin. The F1 antigen forms a polymeric anti-phagocytic capsule around the bacteria18 whilst murine toxin has a role in preserving Y. pestis in the flea gut19. During its evolution from enteric to flea-vectored pathogen, Y. pestis has lost intestinal adhesin and invasin genes, but has retained the heme locus and possesses a number of chromosomal-encoded genes such as the ph6/psa fimbrial and attachment-invasion locus (ail) which promote colonization to the host cells19,20,21.

In the process of acquiring a new mechanism of infection, Y. pestis has also activated genes which enable the pathogen to evade the defenses of its successive hosts. In purified or recombinant forms, some of these encoded gene products have provided vaccine targets and are therefore summarized here.

Y. pestis can survive and grow in the flea’s (most notably the rat flea Xenopsylla cheopis) foregut, leading to ‘blockage’ of the flea. The proper functioning of the bacterial hemin storage system is thought to play an important role in the formation of this blockage19, which during the flea bite, results in the regurgitation of a dense bolus of bacteria5 into a new host. Y. pestis expresses other genes in the flea gut such as a ‘murine’ toxin with phospholipase D activity20 and a lipopolysaccharide (LPS) core modification locus, which together are required for biofilm formation and blockage of the flea20,21,22. However, transcriptional analysis of Y. pestis in the flea gut has identified a wide range of additional genes, such as insecticidal-like toxin genes, which are differentially regulated such that bacteria regurgitated into a new host have increased resistance to innate immune effectors23.

Upon infection of a new mammalian host, the plague bacilli are vulnerable to phagocytosis by polymorphonuclear leukocytes (PMNs or neutrophils) and/or monocytes. The bacteria may be killed within PMNs, but can persist within monocytes and express various virulence determinants, allowing Y. pestis growth and eventual release from the monocytes24. The fibrillar adhesin pH6 antigen is induced by low phagosomal pH (4.5)25 and promotes bacterial adhesion to host cells, thereby enhancing resistance to phagocytosis26. Secretion of the F1 antigen with capsule formation is triggered by a temperature shift from 28 °C in the flea to 37 °C in humans or other mammals. The F1 capsule also plays a key role in avoiding phagocytosis27. However, non-capsulated Y. pestis retains its full capability to cause pneumonic infection in animals, while having reduced virulence during bubonic infection28.

The dominant anti-host effects are due to a temperature shift induction of the TTSS carried on the virulence plasmid pYV/pCD1. TTSS effectors, historically called Yersinia outer membrane proteins (Yops), have cytotoxic and phagocyte regulatory effects, are secreted through an injectosome after Y. pestis makes contact with the host cell, and are delivered into target cells15. The function of many of the Yops has been delineated for this well-characterized secretion system, and serves as a paradigm for other bacterial TTSS’s15. For example, the YopE protein is a cytotoxin and the YopH protein is a tyrosine phosphatase with anti-phagocytic activity29. The V (or Low calcium response V, LcrV) antigen plays a pivotal role by orchestrating intracellular Yop low calcium response protein G (LcrG) elaboration of the injectosome and then itself being delivered through this needle-like structure to be assembled as a pentamer at the tip30. Additionally, V antigen secreted from Y. pestis exerts a local immunomodulatory effect in the host by down-regulating the production of interferon-γ (IFNγ) and tumor necrosis factor-α (TNFα)31,32.

Plasminogen activator (Pla) is another major virulence factor in Y. pestis. Pla is an outer membrane-located protease, which breaks down the physical barriers of connective tissue in the host, thus promoting the systemic dissemination of the bolus of Y. pestis injected by the flea. The requirement for Pla has driven the selection in Y. pestis of the “rough” phenotype of LPS, which lacks an O antigen33,34, a rare phenomenon amongst gram-negative bacteria which has possibly resulted from the bacterium’s transmission through the flea, but which is necessary for Pla to be functional35,36. Inactivation of the O-antigens on Y. pestis LPS exposes the LPS core, so that Y. pestis can interact with C-type lectin receptors on host macrophages, promoting its uptake, and thus accelerating bacterial dissemination in the host37. Our study has also shown that the Δpla mutant is unable to survive efficiently in murine and human macrophages, unlike the wild-type Y. pestis38.

The bacteria disseminate from the site of primary infection into draining regional lymph nodes. Within the lymph node, further growth of the bacteria accompanied by a massive inflammatory reaction leads to lymphadenopathy and the formation of buboes, typically in the groin or axillae. In the bubo, bacteria are predominantly extracellular, mainly due to the TTSS which is highly expressed in the lymph node39. An ability to proliferate in the bubo40 is enabled by the efficient and abundant iron acquisition systems possessed by Y. pestis41.

Eventually, the bacteria are disseminated by the lymphatic system, gain access to the blood stream, and colonize pulmonary tissues, which may lead to development of the pneumonic form of the disease. When left untreated, pneumonic plague induces an overwhelming septicemia which triggers septic shock in the host. However, the precise mechanisms that lead to the death of the host have not been identified but involve multi-organ failure, during which the systemic induction of nitric oxide synthase may contribute, as seen with other gram-negative septicemias42.

Whilst pigmentation (pgm)-negative strains of Y. pestis are usually avirulent and attenuated, the risk of reversion to virulence was highlighted by the fatal case of a laboratory worker who was unknowingly suffering from hemochromatosis and was exposed to the attenuated pgm-Y. pestis laboratory strain KIM. This individual developed plague and died, presumably due to his hemochromatosis-induced iron overload condition providing the infecting KIM strain, attenuated through defects in its iron acquisition ability, with sufficient iron to render it virulent43.

The early use of an inactivated whole cell vaccine for plague by Haffkine between 1897 and 1935 successfully curtailed plague outbreaks in India. This was the first demonstration that components of Y. pestis, even when inactivated, could be immunogenic. Haffkine’s heat-killed whole cell (KWC) vaccine was administered to the human population in an estimated 24 million doses44 (Table 1).

During the 1990s, there were several commercial suppliers of the KWC vaccine against plague. Subsequently, plague vaccine USP (United States Pharmacopeia; 1939–1999), containing formaldehyde-killed bacteria, was manufactured by Cutter Laboratories, USA. In 1994, the manufacturing was transferred to Greer Laboratories Inc., USA. In 1999, the production of this vaccine was discontinued largely because of severe side effects and its protection against bubonic but limited efficacy against pneumonic plague45,46,47,48,49,50,51,52. An alternative heat-killed (KWC) vaccine was also manufactured by the Commonwealth Serum Laboratories (CSL, Australia)53 and until November 2005, was licensed for clinical use in Australia. Additionally, a Y. pestis isolate (EV76-NIIEG Y. pestis)54, which is attenuated due to deletion of the pigmentation locus (pgm), has been used as a vaccine for many years and is licensed for use in China and Russia specifically55 where plague is endemic. The vaccine can be administered by various routes; however, the vaccine is fully virulent under iron-overload conditions, i.e., in individuals with hemochromatosis56,57 (Table 1).

The seminal observation in 1956 by Bacon and Burrows that Pasteurella pestis (now Y. pestis) could be anti-phagocytic in the absence of capsule, led to the identification of a new virulence antigen, which they named the V antigen58. This paved the way for subsequent research to the present day on the immunogenic and protective potential of this and other virulence factors of Y. pestis59,60. Building on the observation that the F1 antigen-containing Cutter KWC vaccine needed the addition of a recombinant V (rV) antigen to fully protect mice against pneumonic plague61, Williamson, et al. demonstrated the synergistic effect of F1 and V in combination. Whilst vaccines lacking the V antigen may protect against bubonic plague, several groups showed that the inclusion of the V antigen was an essential requirement for protection against pneumonic plague61,62.

Much work has been carried out to determine the protective potential of other antigens derived from Y. pestis in native or recombinant form in addition to F1 and V, such as Pla, a protein constituent of the injectisome known as Yersinia secretory factor F (YscF), and a range of other Yops, and their various combinations63,64,65,66,67. Whilst some of these imparted partial protective efficacy and are useful adjuncts in some vaccine formulations (see below), to date F1 and V remain the key proteins which individually have protective efficacy, but which in combination, are consistently synergistic and, therefore, form the core building blocks of most vaccine approaches.

Currently, there are more than 21 candidate vaccines in the preclinical phase3. Below, we have reviewed the pre-clinical candidates (Tables 1–5) and subsequently those that are in early clinical development with a timeline (Fig. 2). The pre-clinical candidates can be broadly categorised as subunit, live attenuated, vectored (bacterial or viral), DNA, or messenger RNA (mRNA).

*Adjuvant not specified. Ages of study participants ranges from 18 to 55 years. All vaccines were given in 2–3 doses intramuscularly over a range of 6 months. The EV 76 NIIEG vaccine was given 1–4 times at intervals of 1–3 months. ?data not published; !data not conclusive.

Many groups have now shown that immunization with the F1 and V subunit antigens provides a high degree of protection against infection caused by Y. pestis in a range of animal models68,69,70,71,72,73,74,75,76,77,78. The use of F1 and V in combination (F1 + V) or as a genetic fusion (F1-V) has the advantage that protection can be maintained against acapsular (F1-negative) Y. pestis strains which still retain virulence79. Many different formulations have been researched with a view to finding one that provides comprehensive protection with the least number of doses, is stable, and maintains immunogenicity when escalated up the species from mice to humans. These candidate vaccines which have been studied in much more detail are summarized in Table 2.

Adjuvants used in preclinical rF1V vaccine development studies include the toll-like receptor 5 (TLR5) ligand flagellin and protollin73,74,75. Packaged in a polyanhydride nanoparticle with cyclic dinucleotide and delivered as a single dose intranasally, rF1V protected mice against pneumonic plague76. A truncated form of rV (rV10) has also been shown to be protective, as has rV10 in manganese silicate nanoparticles71,77. Similarly, the peptidoglycan-free outer membrane vesicles (OMV) with a phage lytic system has been demonstrated to be efficacious78, as have microvesicles derived from human commensal gut bacteria for immunization with V antigen80 or OMV’s from Y. pseudotuberculosis81. A dry formulation of rF1+rV delivered on calcium phosphate-decorated microparticles demonstrated enhanced immunogenicity and efficacy82.

A study has shown that the polymeric form of F1 led to rapid protective humoral immune response by activating innate-like B1b cells