Background
Shigella is the enteroinvasive bacterium responsible for bacillary dysentery (shigellosis).
Shigella causes ~ 165,000 deaths worldwide every year, of which ~ 55,000 are in children younger than 5 years of age [
1,
2]. In recent years the treatment of shigellosis has become increasingly difficult as resistance to antibiotics has spread [
3]. Therefore, new approaches to treat and/or prevent shigellosis are highly desirable. Vaccines have proven to be an effective method to prevent various infectious diseases. Human studies have shown that a previous
Shigella infection confers up to 72% protection against subsequent disease episodes [
4‐
7]; therefore an effective
Shigella vaccine could significantly reduce the burden of this disease. However, to date, no vaccine has been licensed for
Shigella. Various promising live-attenuated oral vaccine candidates are currently under development, such as the live-attenuated vaccine candidate CVD 1208S (
S. flexneri 2a;
ΔguaBA, Δset, Δsen) [
8,
9]. Phases 1 and 2 clinical trials of this vaccine candidate have demonstrated the safety of this strain as well as its ability to induce humoral responses [
9,
10].
Shigella is an intracellular microorganism that targets macrophages and gut epithelial cells; therefore, T cell mediated immunity (T-CMI) is expected to play an important role, particularly in the resolution of the disease. Despite this, only limited information is available on the role of T cells in shigellosis. In humans, production of cytokines (e.g., IFN-γ, TNF-α, IL-6, IL-4) has been demonstrated in the supernatants of PBMC of vaccinees stimulated with soluble
Shigella antigens [
11]. Additionally, production of similar cytokines has been shown at the local level in immunohistochemical studies [
12‐
15]. However, in the case of
Shigella, the T cell subset(s) producing these cytokines have remained largely unknown. Previous studies with CVD 1208S focused on the development of humoral immune responses. Here we focused on the ability of CVD 1208S to induce T-CMI in healthy adult volunteers following vaccination with three oral doses (28 days apart). T-CMI to
Shigella IpaB, one of the immunogenic proteins that is part of the type 3 secretion system (T3SS) and used as a subunit vaccine candidate [
16,
17], was assayed 28 days after each immunization using a novel technique developed in our laboratory. CVD 1208S was able to induce cytokine production as well as upregulation of the degranulation marker CD107a in various CD8 and CD4 memory T cell subsets. CD8 T effector memory (T
EM) cells showed more pronounced multifunctional capacity than the other T cell subsets. The strongest T-CMI responses were detected after the first vaccine dose. The second and third vaccine doses induced responses mainly in volunteers that had not developed T-CMI to the previous vaccination(s). In sum, CVD 1208S is capable to induce T-CMI responses, which most likely complement the humoral responses elicited by this vaccine candidate and are likely to play an important role in combating
Shigella infections.
Discussion
Even though it is widely accepted that serotype-specific humoral responses play an important role in protection from
shigellosis [
3,
29,
34], the intracellular nature of this microorganism suggest that T-CMI might also play a significant role, particularly in ameliorating the severity and duration of symptoms. Evaluation of these immune responses in humans has been limited [
12,
35]; however, some studies have suggested a role for T-CMI in shigellosis. For example, mononuclear cells from some volunteers challenged with a Shiga toxin-deleted
S. dysenteriae 1 strain SC595 exhibited increased production of IFN-γ and IL-10 after challenge [
11]. IFN-γ was also detected in the culture supernatants of PBMC from some human volunteers vaccinated with the
S. sonnei (WRSS1) and
S. flexneri 2a (CVD 1207) candidate strains [
32,
33]. Moreover, IFN-γ production by PBMC following ex vivo stimulation with IpaB (ELISpot) was reported in recipients of the
S. flexneri 2a vaccine candidate CVD 1208S [
9]. The last report is of particular importance, since IFN-γ production was detected in 25 and 57% of volunteers receiving a single dose of either 10
8 and 10
9 CFU, respectively, of CVD 1208S. In the current study, after the first dose of CVD 1208S (10
8 CFU) IFN-γ production was detected in CD8 T
EM, T
EMRA and T
CM in 34.6, 18.2 and 45.5%, respectively, of the vaccinated volunteers. These percentages fall within the expected responses for this vaccine at this dose. Initial studies showed cytokine production by PBMC in humans after
Shigella vaccination, but the cell populations producing these cytokines remained unknown. Therefore, in the current study we further characterized the T cell subsets responsible for cytokine production after
Shigella vaccination as well as whether these cells exhibited multifunctional properties.
Our initial attempts to evaluate T-CMI to IpaB resulted in cell death of autologous B-LCL from various volunteers.
Shigella-induced cell death due to necrosis and apoptosis has been reported by various groups in several cell types (e.g., macrophages, neutrophils, B cells) [
36‐
38]. IpaB-induced apoptosis in macrophages has been described and the mechanism identified [
39]. Induction of necrosis by
Shigella in human neutrophils has also been reported. This process was dependent on IpaB and IpaC and involved actin polymerization [
36]. However, the mechanism by which IpaB induces necrosis in B-LCL remains unknown. A possible molecule to consider is CD44, the receptor for hyaluronic acid and a widely distributed cell surface glycoprotein that exist in a variety of isoforms. CD44 has been implicated in a number of cellular adhesion processes, as well as cell growth, differentiation, modulation of apoptosis and necrosis [
40]. It was recently reported that the binding of the inflammatory mediator ultra-low-molecular-weight hyaluronan (ULMW-HA) to CD44 in B-precursors leukemia cells resulted in necrosis [
37]. The degree of necrosis induction depended on the level of CD44 expression; cells that expressed high levels of CD44 had higher rates of necrosis. Importantly, CD44 has been reported as a receptor for
Shigella IpaB in epithelial cells [
41,
42]. Moreover, CD44 is expressed on B cells [
43‐
45]. It is reasonable to speculate that if CD44 is involved in induction of necrosis in B-LCL, incorporation of IpaB into LPS-nanoparticles is likely to mask the CD44-biding site of IpaB. Another aspect to consider is that bacterial LPS is composed of large repetitive polysaccharide chains which facilitate simultaneous binding of multiple B cell receptors (BCRs). For this reason, LPS is considered a T-cell-independent type 1 (TI-1) antigen [
46], capable of eliciting strong polyclonal B cell activation due to the repetitive cross-linking of multiple BCRs. Incorporation of IpaB into LPS-nanoparticles might facilitate uptake of IpaB by B-LCL, using the TI-1 route while bypassing CD44-binding and therefore limiting necrosis. The use of fluorescently labeled IpaB allowed verification of its uptake by B-LCL. In sum, the approach to deliver IpaB in the context of LPS-nanoparticles to B-LCL reduced necrosis and facilitated their activation (Fig.
1) enabling the use of these cells as stimulators/targets in T-CMI assays.
Once this method to produce stimulator/target cells was developed and optimized, we evaluated the induction of CD4 and CD8 T-CMI by determining cytokine induction (IFN-γ, IL-2, IL-17A, TNF-α) or CD107a upregulation in various T memory subsets (Fig.
2). Overall, CVD 1208S induced T-CMI responses in all the CD8 and CD4 T cell memory subsets assessed. The first vaccine dose induced the strongest and most diverse responses in both CD8 and CD4 T cell subsets (Figs.
3,
4,
6). The second and third doses also induced T-CMI responses; however, these were mainly in volunteers who had not shown T-CMI responses to the previous doses. No booster effect was noted. Overall, after receiving the three vaccine doses, production of at least 1 cytokine was reported in 8–10 of 11 (72.2–90.9%) vaccinated volunteers in the CD8 and CD4 T cell subsets (Figs.
3c,
6c, d). The reason for the lack of booster responses after the second and third vaccine doses is unclear. However, it can be speculated that for this particular vaccine, the window of time between the doses was not optimal. In non-human primate studies involving malaria vaccine candidates, it was shown that extending the time between vaccine doses from 4 to 8 weeks boosted the malaria-specific CD8 T cell responses that were associated with protection from wild-type challenge [
47]. Similarly, in hepatitis B vaccine studies, the immunity to this vaccine was enhanced in several volunteers receiving a delayed third vaccine dose [
48,
49]. Therefore, spacing of the vaccine doses might provide an additional tool to enhance the response to CVD 1208S.
CD8 T
EM and CD8 T
CM cell showed the strongest and most diverse responses to CVD 1208S after the first dose. However, only CD8 T
EM cells showed a higher frequency of MF than S+ cells. Interestingly, in CD8 T
EM cells TNF-α was detected only in cells that produced another cytokine or expressed CD107a; therefore, TNF-α+ cells were mainly MF (Fig.
5c). Moreover, 4 of the 5 predominant MF populations expressed TNF-α, providing more evidence of the importance of this cytokine (Fig.
5d). It is also important to note that previous studies identified IFN-γ as one of the main cytokines induced by live-attenuated
Shigella vaccines [
11,
32,
33]. In the current study, we confirmed IFN-γ induction by CVD 1208S and showed that this cytokine was increased in 4 of the 5 predominant MF populations. TNF-α and IFN-γ are inflammatory cytokines and the latter plays an important role in the resistance to infection by
Shigella in the mouse model [
50]. One of the mechanisms of resistance involves inhibition of
Shigella replication via the cytoplasmic RNA sensor retinoic acid-inducible gene I (RIG-I) [
51]. The mechanism by which TNF-α limits
Shigella infection is still unclear; however, this cytokine is known to play a critical role controlling other intestinal infectious bacteria in the mouse model [
52]. MF cells appear to be of particular importance for control of disease progression in HIV and other diseases [
53‐
56]. Whether the same principle applies to
Shigella remains to be explored, but if we assume that this is the case, the induction of CD8 T
EM cells expressing TNF-α and/or IFN-γ along with other cytokines suggest that the MF immunity induced by CVD 1208S could aid in controlling infection of
Shigella. Pioneering experiments in rectal biopsies from convalescent patients showed cells producing cytokines [
14,
15]; demonstrating the induction of T-CMI at the intestinal level. Since CVD 1208S is a live-attenuated oral vaccine strain able to invade epithelial cells, we expect this vaccine to induce robust local T-CMI responses mimicking those induced by wild-type
Shigella.
Despite the fact that CD8 TCM cells also showed strong and diverse responses, MF cells in this compartment were less frequent and no cytokine showed predominance as TNF-α did in CD8 TEM cells. CD4 T-CMI responses in all the T memory subsets analyzed (TEM, TEMRA and TCM) followed the same pattern than CD8 TEM cells responses, but showed neither MF dominance, nor the same degree of importance for TNF-α production. The diverse T-CMI detected in the different CD8 and CD4 memory subsets reflect the complexity of the responses to Shigella and probably also demonstrate the different kinetics of these cell populations. The current study provides evidence of the MF nature of T-CMI systemic responses. Moreover, since CD4 T-CMI responses, which are critical for the development of humoral responses are present, it is likely that CVD 1208S induces long term memory.
Of note, we also evaluated expression of CD107a on the surface of the cells. CD107a and b are normally present inside the cell in the membrane of cytotoxic granules. Mobilization of CD107a and b to the surface of the cells is dependent on degranulation and therefore an indication of cytotoxic activity (CTL) [
57,
58]. CTL activity has been reported in both CD8 and CD4 T cells [
59]. In the current study we showed induction of CTL activity (CD107a expression on the surface of the cell) by CVD 1208S in all the subsets of CD8 and CD4 cells (T
EM, T
EMRA and T
CM subsets). As expected CD8 cells showed stronger CTL activity than CD4 T cells and among the subsets, CD8 T
CM cells showed a higher number of volunteers with CTL activity [6 of 11; 55% (Fig.
6d)] after the first vaccine dose, suggesting that in the case of
Shigella, this subset could play an important role limiting disease progression. Additionally, in CD8 T
EM cells, CTL was identified in 2 of the 5 predominant MF groups; suggesting that CTL activity in combination with other cytokines (e.g., IFN-γ and TNF-α) could have an important role limiting shigellosis. This is, to our knowledge, the first report showing induction of CTL activity by a live-attenuated oral vaccine to
Shigella and suggest that CTL could be important in eliminating
Shigella-infected cells.
During the development of immune responses, primed cells migrate to secondary lymphoid organs. Some of these cells are identified in the peripheral blood. Therefore, PBMC provide a window to explore some of the responses that are being developed at the local (gut) level. Nevertheless, the responses in the periphery are expected to be of smaller intensity, and perhaps different in their characteristics, than at the local level. Of particular importance as a first step in studying this phenomena is the measurement of the expression of gut homing molecules, such as integrin α4β7, which suggests that the immune cells measured systemically are in the process of migrating to the gut. Importantly, we demonstrated that CD8 T
EM cells expressing integrin α4β7 showed enhanced cytokine production (TNF-α and IFN-γ) than those cells that were integrin α4β7− (Fig.
8). This was detected in three of four individuals (75%) that had increased expression these cytokines 28 days after the first vaccine dose. This clearly suggested that IpaB specific CD8 T
EM cells continued to migrate to the gut 4 weeks after vaccination. In the individual that showed the higher cytokine production of cytokines in integrin α4β7− cells, it is likely that most of the IpaB specific cells had already migrated to the gut. This would be in line with previous studies in individuals challenged with wild-type
Salmonella Typhi (another enteric pathogen), who showed a decrease the percentage of CD8 T
EM cells expressing integrin α4β7 6–13 days after challenge [
30] suggesting an early migration of these cells to the gut. Importantly, decrease in integrin α4β7+ cells was noted only in those individuals that developed disease after challenge. Homing of immune cells is a very dynamic process and in the current study, we did not collect earlier time points that have allowed a more in depth analysis of the migration process of these cells. Future studies will address this issue in more detail.
The current study describes the T-CMI responses to a vaccine candidate. Therefore, it is not possible to identify immunological correlates of protection. Future studies in which volunteers will be orally immunized and subsequently exposed to wild-type Shigella flexneri 2a hold the potential to identify defined T-CMI responses associated with protection from shigellosis. The novel technique presented in this manuscript will be of importance in this endeavor.
In sum, we have developed a novel method for measuring T-CMI to IpaB, which is an immunogenic protein present in all
Shigella species and a component of some current subunit vaccine candidates [
16,
17]. We used this technique to demonstrate the induction of T-CMI, and characterized in depth, for the first time, these responses, including the T memory subsets elicited and their homing potential, following exposure to CVD 1208S, a leading attenuated
S. flexneri 2a vaccine candidate. Upcoming studies involving vaccination and/or challenge with attenuated and wild-type
S. flexneri organisms will be essential in elucidating the role of T-CMI in controlling
Shigella infection.