Introduction

Neutralizing antibody against measles is an established correlate of protection with a threshold protective titer of 120 mIU/mL1. However, recent meta-analyses have revealed that vaccine-induced, measles-specific antibodies wane in a time-dependent manner in recipients of two doses of measles-mumps-rubella (MMR) vaccine, with an initial period of greater annual waning (121.8 mIU/mL/year) in the first 5 years post-vaccination, followed by a slower (24.1 mIU/mL/year) at later timepoints post-vaccination2,3. Live MMR vaccine has additional limitations such as: vaccine hesitancy4,5,6, the requirement of a cold chain, and contraindications for use in immunocompromised and pregnant individuals7. In addition, pre-existing maternal antibodies suppress the infant’s antibody response to measles vaccine, hence limited seroconversion rates are observed in young children8. Meanwhile, measles-specific T cell responses can be elicited in infants as young as 6 months old9. Accumulated evidence also supports the role of T cells in the clearance of measles-infected cells10,11. Therefore, eliciting T cell-mediated immunity is a complementary approach in the development of a next-generation measles vaccine12.

Identifying immunogenic antigens is one of several critical steps in the development of T cell-mediated vaccines. Pathogen-derived peptides represent an excellent antigen source due to their intrinsic properties such as safety profile, chemical stability, and low production cost13,14. In an ongoing effort to better understand cellular immunity to measles, we previously identified 13 measles virus (MV)-derived, naturally-processed and presented peptides, bound to class-II HLA-DRB1*03:01 molecules, in MV-infected Priess (EBV-transformed lymphoblastoid) cells15. These peptides stimulated the production of IFN-γ, IL-416, and the proliferation of T cells isolated from healthy recipients of measles-containing vaccine17, demonstrating that they are immunogenic epitopes. However, the immunogenicity/protective capacity of these peptides in small animal models remains to be characterized. Herein, we evaluated the immunogenicity of these 13 previously identified peptides in a mouse model. Of these 13 peptides, the top four immunogenic peptides were incorporated into a vaccine formulation with incomplete Freund’s adjuvant and CpG. This peptide-based vaccine significantly reduced morbidity-associated weight loss in immunized mice after measles virus challenge compared to placebo.

Results

Characteristics of previously identified peptides

Thirteen (13) naturally processed class-II HLA-DRB1*03:01 bound measles peptides were previously identified from MV-infected EBV-transformed Priess (B) cells using two-dimensional liquid chromatography combined with gas phase fractionation tandem mass spectrometry15. The length of these peptides spanned from 11 to 25 amino acids (aa) with their molecular weight (Mw) ranging from 1159.3 to 2799.2 (g/mol) (Table 1). These peptides were derived from three measles structural proteins, including 7 peptides from phosphoprotein, 5 peptides from nucleocapsid, and 1 peptide from hemagglutinin (Table 1). Of the 7 peptides from the phosphoprotein, 4 peptides (peptides #1–4) shared the same 17-aa core sequence. Two other peptides from the phosphoprotein (peptides #5 and #6) shared a 12-aa sequence, while peptides #8 and #9 from nucleoprotein have an identical 14-aa sequence.

Table 1 A list of measles-derived peptides used in this study.
Full size table

Four of thirteen peptides induced positive IFN-γ responses

We arbitrarily divided the 13 peptides into two pools: pool #1 (peptides #1–6) and pool #2 (peptides #7–13) and formulated these pooled peptides with the Incomplete Freund’s adjuvant (IFA) and CpG 1826 for mouse immunization (Fig. 1A). Mice were immunized with either one (Fig. 1B,C) or two doses (Fig. 1D,E) of peptide vaccine with a two-week interval between doses. Two weeks after the last dose of vaccine, single cell suspensions from spleen and lymph nodes of euthanized mice (thereafter referred to as ‘lymphocytes’) were used to evaluate the immunogenicity of these peptides using an IFN-γ ELISPOT assay. We defined a positive response as an IFN-γ ELISPOT response with a stimulation index (SI) ≥ 2 and a significance level of p < 0.05 (see Methods). In mice immunized with one dose of the peptide vaccine, only peptide #4 from pool #1 induced a positive IFN-γ response with a SI of 22.44 (p < 0.001) (Fig. 1B). Other peptides from pool #1 failed to induce a significant cell response (i.e., SI of ~ 1). In pool #2, peptides #12, #11, and #7 induced significant IFN-γ responses with SI of 28.43, 3.55, and 2.05, respectively (Fig. 1C). Interestingly, lymphocytes from mice immunized with either peptide pool also recognized measles virus (MV) with a SI of 11, confirming that one or more of the measles-derived peptides in each pool can be processed from whole virus particles and presented by murine APCs (Fig. 1B,C).

Figure 1
figure 1

Immunogenicity of 13 measles-derived peptides in mice. (A) Thirteen peptides were divided into two pools: pool #1 (peptides #1–6, data shown in B, D) and pool #2 (peptides #7–13, data shown in C, E). Mice were immunized with one dose (data shown in B, C) or two doses (D, E) of peptide pool #1 or #2 formulated with incomplete Freund’s adjuvant (IFA). Two weeks after the last immunization, mice were euthanized, and mesenteric lymph nodes and spleen were harvested for the IFN-γ ELISPOT assay. Lymphocytes were incubated with each of peptide in peptide pool #1 (B, D) or pool #2 (C, E). Inactivated measles virus (MV) and phytohemagglutinin (PHA) were used as positive stimulants. As a negative control (Unstim), lymphocytes were incubated in cell culture media at the same experimental conditions.

Full size image

We observed similar results in mice immunized with 2 doses of the peptide vaccine, in which four peptides #4, #12, #11, and #7 induced positive IFN-γ responses (Fig. 1D,E). We also observed a nonsignificant difference in IFN-γ responses against each peptide in lymphocytes from mice immunized with 1 and 2 doses of peptide vaccine (Fig. 1B–E). Based on these results, four immunogenic peptides (peptides #4, #7, #11, and #12) were down-selected to formulate with different adjuvants for further experiments.

Four down-selected peptides induced positive responses when formulated with Th1-biased adjuvants

Four down-selected peptides (peptides #4, #7, #11, and #12) were formulated with different commercially available vaccine adjuvants (Table 2). Mice were immunized with two doses of peptide vaccine with a four-week interval between doses. Two weeks after the second dose, lymphocytes were isolated from euthanized mice and recall immune responses were quantified by IFN-γ ELISPOT and cytokine ELISA (Fig. 2A). Of the eight adjuvants used, lymphocytes from mice immunized with the Th1-biased adjuvants, CpG 1826, R848, and poly(I:C), in combination with IFA, produced statistically significant IFN-γ responses (Fig. 2B). Of these three adjuvants, CpG 1826 and poly(I:C) induced comparable IFN-γ responses, while R848 induced a weaker IFN-γ response (Fig. 2B). In each of the three adjuvants, the highest response was found against peptide #4 while negligible responses were detected against peptide #7. Other adjuvants, including IFA and MPLS, IFA and Pam3CSK4, AddaVax, Quil A, and N-Glycol, failed to induce IFN-γ responses above background levels observed in the PBS negative control (Fig. 2B).

Table 2 A list of adjuvants used to formulate four down-selected peptides.
Full size table
Figure 2
figure 2

Immunogenicity of four down-selected peptides formulated with different adjuvants in mice. (A) Mice were immunized with two doses (with four-week interval) of four down-selected peptides (peptides #4, 7, 11, 12) formulated with different adjuvants. Two weeks after the second dose, mice were euthanized; mesenteric lymph nodes and spleen were harvested for IFN-γ ELISPOT assay and cytokine ELISA. (B) Lymphocytes were incubated with individual peptide or MV and the numbers of IFN-γ-secreting cells were assessed by the IFN-γ ELISPOT assay. In addition, lymphocytes were incubated with either individual peptide, MV, or PHA overnight and the level of a panel of Th1/Th2 cytokines was quantified in the supernatant by multiplex ELISA (C, D, E).

Full size image

The collected lymphocytes were further incubated with either individual peptides or controls overnight and cell culture supernatants were collected and tested for cytokines by ELISA. We observed detectable levels of IFN-γ, IL-2, and TNF-α in cell culture supernatants (Fig. 2C–E). Consistent with the IFN-γ ELISPOT results, these cytokines were detected in lymphocytes from mice immunized with peptides formulated with CpG 1826, poly(I:C), and R848 (Fig. 2C–E). Other cytokines, including pro-inflammatory IL-1β, IL-6, and CXCL1; Th2-biased IL-4 and IL-5, were also measured, but no increase in secretion was detected upon restimulation with peptides (data not shown).

Selected peptides mitigated weight loss in immunized mice after viral challenge

We further evaluated the immunogenicity of these four down-selected peptides (peptides #4, #7, #11, and #12) in MV-challenged mice. It is well-known that mice are not a natural host for MV and the virus is rapidly cleared due to the innate antiviral responses mediated by IFN signaling pathways. Viral infection requires not only expression of a viral receptor, but also disruption of the IFN pathways, or the use of neuro-adapted virus strains. For our experiment, we crossed HLA-DR3 transgenic mice with human CD46 transgenic mice. The F1 progeny expressed both the necessary viral receptor and the appropriate HLA molecule. Furthermore, we blocked IFNA signaling pathways in these mice with an initial bolus of 2.5 mg of IFNAR inhibitor MAR1-5A3 one day before viral challenge and maintained the blockage with weekly doses of 1 mg of MAR1-5A3 throughout the observation period (Fig. 3A). Mice were monitored for general health and weight loss up to 36 days after viral challenge. We observed that unvaccinated mice experienced a significant weight loss of ~ 7% four days after viral challenge and continued to lose up to 11% of their weight at day 31 after viral challenge (Fig. 3B). Meanwhile, vaccinated mice maintained ~ 96–98% of their weight before viral challenge throughout the study period. The weight loss was significantly different between vaccinated and unvaccinated mice on day 12 after viral challenge, continuing until the end of study period at day 36 (Fig. 3B).

Figure 3
figure 3

Challenge study. (A) Mice were immunized with four peptides (peptides #4, 7, 11, and 12) formulated with IFA and CpG 1826. Twenty-eight days after immunization, mice were challenged with 5 × 106 pfu of MV. One day before viral challenge, mice were treated with a bolus of 2.5 mg of IFNAR inhibitor MAR1-5A3 to block IFN signaling pathways. The blockage of IFN signaling pathways was maintained after viral challenge by the injection of 1 mg of MAR1-5A3 weekly. Mice’s weight was monitored up to 36 days after viral challenge, then mice were euthanized to harvest spleen and lymph nodes for the IFN-γ ELISPOT assay. (B) Weight loss, as compared to the weight at Day 28. (C) Lymphocytes were incubated with either individual peptide or a mixture of four peptides or MV (positive control) and the number of IFN-γ-secreting cells were assessed by the IFN-γ ELISPOT assay. Asterisk “*”: p < 0.05; “ns”: nonsignificant difference.

Full size image

Thirty-six days after viral challenge, mice were euthanized, and lymphocytes were isolated for use in the IFN-γ ELISPOT assay (Fig. 3A). The lymphocytes were incubated with either individual peptides, a mixture of all four peptides, or MV. IFN-γ responses to peptide #4 were significantly greater in vaccinated mice, with a SI of ~ 8, compared to those in unvaccinated mice. While a trend toward higher immune responses was observed in the vaccinated group for the other subdominant epitopes (peptide #7, #11, and #12), none of these differences reached statistical significance (Fig. 3C). As expected, challenged animals mounted an immune response to whole virus, with the peptide-vaccinated group displaying a slightly higher immune response in some animals, although this increased response was not statistically significant at the group level.

Discussion

MV-specific T cell responses are essential for recovery from measles infection. While humoral immunity, specifically neutralizing antibody response, has been established as a correlate of protection against infection1, ample evidence supports the role of T cell-mediated immunity in MV clearance after infection10,11. MV infection is accompanied and followed by the generation of MV-specific, IFN-γ-producing CD4+ and CD8+ T cells. Their frequencies in circulation as well as polyfunctionality have been shown to correlate with viral control and viral RNA level reduction in rhesus macaques11,18,19. Depletion of CD8+ T cells led to uncontrolled measles viremia in macaques20, and while insufficient to protect from symptomatic infection, pre-existing MV-specific CD4+ and CD8+ T cell responses resulted in a two-fold reduction in the duration of viremia19. The clinical importance of T cell immunity in measles has been established by comparing differential measles outcomes between children with congenital antibody deficiencies versus those with defects in T cell-mediated immunity; children with agammaglobulinemia clear natural measles infection uneventfully, while those with cellular immunodeficiencies are prone to progressive and sometimes fatal measles infection10,21,22.

We previously identified 13 MV-derived peptides in MV-infected cells15, and herein examined the immunogenicity of these peptides in a mouse model using our approach as previously established23. Immunization with one immunodominant (peptide #4) and three subdominant peptides (peptides #7, #11, and #12) significantly decreased weight loss after viral challenge compared to mock immunization. Our results also indicated a consistent trend towards greater T cell reactivity in vaccinated animals. In the case of peptide #4, there was an almost tenfold stronger response in peptide-vaccinated animals. Our working hypothesis is that the increased T cell response (capable of recognizing infected cells—see Figs. 2B, 3B) is a contributing factor to the reduction in weight loss after infection.

Of 13 identified peptides, four peptides induced positive IFN-γ responses (Fig. 1). However, results from our viral challenge study suggested that the IFN-γ responses in vaccinated mice were driven primarily by a single peptide (peptide #4) (Fig. 3C), suggesting that this peptide could be formulated in a peptide-based measles vaccine. Although a concern may be raised whether a single peptide-based vaccine is sufficient for the protection, a previous study has shown that a single-epitope T cell-based vaccine sufficiently protected immunized mice from SARS-CoV-224. However, a vaccine with optimal immunogenicity in the human population is likely to consist of several peptides capable of broad binding across multiple HLA alleles, necessitating work on identifying additional T cell epitopes in the context of other peptides. Although we did not formulate each peptide individually and evaluate its protective capacity in this study, our results from the viral challenge study suggest that the protection was largely conferred by peptide #4 (Fig. 3C). Further investigation will be necessary to confirm this finding and to identify additional epitopes that may also contribute to or enhance the level of T cell-mediated protection.

It is worth noting that in our previous study, peptide #1 demonstrated the capacity to elicit the proliferation of T cells isolated from human peripheral blood17. However, in this study peptide #1 induced an IFN-γ response only at background levels in the mouse model used in this study (Fig. 1B,C). Interestingly, peptide #4 and peptide #1 share the same sequence of 19 aa, but peptide #4 is 3-aa longer than peptide #1 (Table 1); and peptide #4 was the strongest IFN-γ inducer in our animal model (Fig. 1B,C). The contribution of the additional three amino acids to enhanced immunogenicity in our animal model is not yet known, but could provide insights into the potential presence of T cell epitope(s) and the immunological relevance of the studied viral peptides. Similarly, peptides #2 and #3 share the same sequence of 17 aa and 19 aa with peptide #4, respectively, but none of them induced IFN-γ production (Fig. 1). Similarly, peptide #8 induced the production of IFN-γ and IL-4 in 15.3% and 23.1%, respectively, of PBMC samples isolated from 281 subjects previously vaccinated with MMR vaccine16; however, it produced only background levels of IFN-γ response in our animal model (Fig. 1C,E). Despite using HLA-DR3 transgenic animals, not all of the peptides were highly immunogenic. The animals also expressed endogenous MHC alleles, which may have affected antigen presentation. It is possible that the additional flanking residues for peptide #4 enhanced MHC presentation in additional to the HLA-DR3-mediated antigen presentation25. Together, these results suggest that the immunogenicity of peptides needs to be validated before formulating them into a vaccine and reinforces the fact that care must be taken in extrapolating in vitro experimental results in humans to animals and vice versa.

Of note, four down-selected peptides (peptide #4, #7, #11, and #12) only exhibited optimal immunogenicity when formulated with adjuvants known to induce a Th1 bias in the resulting immune response: CpG 1826, R848, poly(I:C) (Fig. 2)26. As expected, following immunization, only cytokines that have been produced from activated Th1 cells (IFN-γ, IL-2, and TNF-α) were detected (Fig. 2) while Th2-biased cytokines (IL-4 and IL-5) were not detectable.

It was worth noting that IFN-γ responses were comparably detected in mice that received one dose or two doses of our peptide vaccine (Fig. 1). These results suggested that at the concentration of peptides used in our peptide vaccine, one dose may be sufficient to induce T cell responses as we did not observe a boosting effect of the second dose, where one might be present at lower antigen doses or with different adjuvant combinations. Future development of a peptide-based vaccine for measles needs to optimize peptide concentration for vaccine formulations.

There are several strengths to our study, including the use of a small defined set of antigens in the vaccine formulation, the focus on T cell responses in viral clearance (an understudied immune effector function), and the use of transgenic mice expressing human HLA and CD46—that are both susceptible to measles infection and are capable of presenting viral epitopes known to elicit recall responses in humans. The study also has several limitations. First, we did not evaluate the antibody responses in immunized mice in this study and cannot comment on the possible contributions of humoral immunity to the observed protection (as a result of coexisting B cell epitopes). However, as our peptides were originally identified as T cell epitopes, the IFN-γ ELISPOT assay was appropriate for the evaluation of immunogenicity of these peptides27,28. Second, although we successfully used a transgenic mouse model expressing both HLA-DR3 and human CD46, this mouse model has some intrinsic limitations, such as the requirement to block IFN type I response during the viral challenge. Since measles virus is human pathogen, a highly related animal model, such as non-human primates29,30 would be an appropriate model to conduct follow-up evaluation the immunogenicity of measles vaccines, although feasibility and high cost limit their use. Third, the immunogenicity induced by our peptide vaccine is likely due to activation of T cells and destruction of virally infected cells, which in turn would reduce the viral load in infected mice. Consistent with this hypothesis, a previous study established a correlation between weigh loss in CD46 transgenic mice after viral challenge and viral replication.31 As we did not quantify viral load in our study, future experiments should quantify viral load in infected mice to confirm this hypothesis .

In summary, we conducted a pilot study to evaluate the immunogenicity of 13 MV-derived peptides previously identified in human cells infected with MV in order to determine if a peptide-based vaccine eliciting a T cell response would provide protection against MV viral challenge. Our data confirmed the Th1-biased immune properties of these peptides, and four down-selected peptides demonstrated their protective capabilities in preventing infection-associated weight loss in immunized mice. Our results highlight the potential of a peptide-based vaccine as a platform for next-generation candidate measles vaccines.

Methods

The described methods are similar or identical to the ones in our previously published studies23.

Ethics statement

The Institutional Animal Care and Use Committee (IACUC) at Mayo Clinic reviewed and approved all animal experiments. All procedures of animal experiments were conducted in accordance with the approved guidelines for animal experimentation at Mayo Clinic and the ARRIVE reporting guidelines.

Peptides

Thirteen previously identified HLA-DRB1-restricted peptides (Table 1) were synthesized by Mimotopes Pty Ltd (Melbourne, Australia). The purity of synthesized peptides was > 90%, as analyzed by mass spectrometry. A stock peptide solution of 2 mg/mL was prepared in DMSO and diluted to required concentrations in either PBS (for vaccination) or cell culture medium (for in vitro assay).

Development of a transgenic mouse model

Transgenic mice expressing HLA-DR3 (courtesy of Dr. Chella David, Mayo Clinic) were bred with transgenic mice expressing human CD46 (courtesy of Dr. Roberto Cattaneo, Mayo Clinic). Expression of both human genes in animals from the F1 generation was verified by flow cytometry using fluorochrome-labeled antibodies (BD Biosciences).

Peptide immunization

The thirteen peptides were divided into two pools: pool #1 (peptide #1–6, as in Table 1) and pool #2 (peptides #7–13, Table 1). Mice (n = 3) were subcutaneously immunized with 100 µL of peptide vaccine formulation, containing i) 50 µL of Incomplete Freund’s Adjuvant (IFA), ii) 50 µL of solution (in PBS) containing 100 µg of CpG 1826 and 20 µg of each peptide in pool #1 or pool #2. In a two-dose experiment, mice received an identical booster two weeks after the first dose (Fig. 1A). As a control group, mice were immunized with 50 µL of PBS containing 1 × 106 pfu of live measles virus (Edmonston strain). Mice were euthanized two weeks after the last vaccination. Mesenteric lymph nodes and spleens were harvested for in vitro assays.

In a separate experiment, down-selected peptides were formulated with different adjuvants (Table 2). Mice (n = 3) were immunized with an initial dose of 100 µL of vaccine (containing adjuvant and 20 µg of each peptide) and were boosted with identical vaccine dose one month later (Fig. 2A). Mice were euthanized two weeks after the second dose and mesenteric lymph nodes and spleens were harvested for in vitro assays (Fig. 2A).

Intranasal measles virus challenge study

Measles infection in transgenic mice is short-lived and the virus is rapidly cleared unless innate antiviral responses mediated through IFNAR signaling pathway are inhibited. Therefore, for this challenge experiment, we treated mice (expressing both HLA-DR3 and CD46) with an initial bolus of 2.5 mg anti-IFNAR blocking antibody MAR1-5A3 one day before viral challenge and with maintenance doses of 1 mg MAR1-5A3 weekly throughout the course of the experiment (Fig. 3A). Mice (n = 5) were immunized with 100 µL of peptide vaccine formulation, containing (i) 50 µL of Incomplete Freund’s Adjuvant (IFA), (ii) 50 µL of solution (in PBS) containing 100 µg of CpG 1826 and 20 µg of each of four peptides (peptides #4, #7, #11, and #12). One month after peptide vaccine immunization, mice were intraperitoneally challenged with 100 µL of PBS, containing 5 × 106 pfu of measles virus (Edmonston strain). Mice were monitored for health (signs of infection) and weight for 36 days after viral challenge, then were euthanized for the isolation of spleens and mesenteric lymph nodes for in vitro assays (Fig. 3A).

IFN-γ ELISPOT assay

A single cell suspension derived from the mesenteric lymph nodes and spleen, hereafter referred to as lymphocytes, was prepared and used in an IFN-γ ELISPOT assay (catalog no. BD 551083, BD Biosciences) using our previously published protocol23. Briefly, 200,000 cells (suspended in 100 µL of cell culture medium RPMI) per well in a 96-well ELISPOT plate were incubated with 100 µL of RPMI medium containing 5 µg/mL (final concentration for each peptide) of either peptide pools or individual peptides in quadruplicate. After incubation for 18 h in a 37 ℃/5% CO2 incubator, plates were washed, developed, and allowed to dry in accordance with the manufacturer’s protocol. The dried plates were read on a CTL ELISPOT reader (Cellular Technology Limited). DMSO (at equal volume to the peptides) and inactivated measles virus (Edmonston strain, multiplicity of infection/MOI of 2) in RPMI medium were used as negative and positive controls, respectively. PHA (50 ng/mL) was used as an additional positive control.

Multiplex cytokine ELISA

Lymphocytes (1,000,000 cells) were incubated with either individual peptides or peptide pools in 200 µL of RPMI medium containing 5 µg/mL (final concentration for each peptide). After 18 h-incubation in a 37 ℃/5% CO2 incubator, the cell culture supernatant was collected and subjected to multiple MSD ELISA assay, following the manufacturer’s instruction. Inactivated measles virus (Edmonston strain, MOI = 2) and DMSO (at equal volume to the peptides) in RPMI medium were used as positive and negative controls, respectively. PHA (50 ng/mL) was used as an additional positive control.

Statistical analyses

The statistical analyses were performed similarly to our previously published method23. Briefly, the average of IFN-γ spots (calculated as spots per 200,000 cells—the number present in each well) from experimental groups were compared to that of negative control well (DMSO in cell culture RPMI). A two-tailed t test was used to compare the difference in the average of IFN-γ spots between groups. The stimulation index was defined as the average spot-forming cells in wells stimulated with peptides divided by the average spot-forming cells in wells stimulated with DMSO in cell culture RPMI (background). Positive responses in mice were defined as those with the stimulation index of ≥ 2 and a significance level of p < 0.05. All figures were plotted with GraphPad Prism (version 9).