The Durability and Avidity of MPXV-specific Antibodies Induced by the Two-dose MVA-BN Mpox Vaccine

Study Design

NYC OSMI investigates the immune response in adults with and without HIV who have received the MVA-BN vaccination or have recovered from mpox disease. The study is registered on ClinicalTrials.gov (NCT Number: NCT05654883). OSMI includes blood draws for serum and peripheral blood monocytes at specified intervals, as illustrated in Figure 1. Participants had the opportunity to enroll either pre- or post-mpox vaccination, with the enrollment window extending up to 365 days post-first dose or mpox symptom onset. One hundred and seventy-one (171) adult participants have given informed consent to be part of the study, as approved by the NYU Grossman School of Medicine Institutional Review Board (protocol 22-01338). Participants were recruited from city vaccination centers as well as by word of mouth. In addition to demographic and clinical information, the study includes blood draws from before vaccination, after a single dose, and multiple study visits after the second dose. In this observational study, the cohort has a range of vaccination route combinations, dosing intervals, and demographics.

• Download figure
• Open in new tab

Figure 1: The New York City Observational Study of Mpox Immunity (OSMI)

Diagram of study visits. The windows of time denoted for each visit are used throughout the paper in analyses of specific visits. Diagram generated in BioRender.

Information regarding prior smallpox vaccination was gathered from the participants and confirmed through an adjudication process. For this process, the following factors were considered: physical examination of the upper arms for presence of scar, prior Bacillus Calmette-Guerin (BCG) vaccination, prior military service or other occupational risk requiring smallpox vaccination (i.e., including laboratory work), country of birth, year of migration to the U.S., and whether the year of vaccination was consistent with active smallpox vaccine distribution in the participants country of origin. A literature review was performed to determine when individual countries ended their routine smallpox vaccination campaigns. Review of smallpox vaccination status was conducted by two infectious diseases physicians.

For people with HIV (PWH), CD4+ T cell counts were sourced from electronic medical records at the time of enrollment. The date and route of mpox vaccination were verified using electronic medical records when available or through participant’s description of the vaccination procedure.

Additional control samples and pre-OSMI enrollment samples were pulled from the New York University Langone Vaccine Center Biorepository.

MPXV virus stock propagation

All mpox virus work was conducted in a certified ABSL3 facility. MPXV clade IIb, lineage B.1, was obtained from BEI Resources (NR-58622); this virus was isolated in Massachusetts, USA during the 2022 outbreak. Working stocks of MPXV were generated by plaque purification in Vero E6 cells (ATCC #CRL-1586) followed by 3 passages of propagation in Vero E6 cells at 37°C and a multiplicity of infection (MOI) of 0.01 for 3 days each. Virus was harvested by freeze-thaw lysing scraped cells followed by clarification by centrifugation to remove cell debris. Stocks were purified at each passage over 36% sucrose in TNE buffer (Quality Biological #351-302-101) at 32,900xg for 80 min at 4°C. Purified MPXV stocks were sequence-verified.

Immunofluorescence-based MPXV microneutralization assay

On the day before MPXV infection, 1.5×10⁴ Vero E6 cells were plated in each well of a black 96-well assay plate in cell culture media (Dulbecco’s Modified Eagle Medium [DMEM] supplemented with penicillin/streptomycin, 2 mM L-glutamine, and 10% heat-inactivated fetal bovine serum [FBS]) and incubated at 37°C and 5% CO₂. Participant serum was 2-fold serially diluted in infection media (DMEM lacking sodium pyruvate supplemented with 2% heat-inactivated FBS) starting at a 2.5-fold dilution (final starting dilution is 5-fold following addition of an equal volume of virus) for 8-points. Prior to infection, serum dilutions were incubated with MPXV (sufficient for MOI ≈ 0.01) for 1 hr at 37°C. Cells were washed with 1x phosphate-buffered saline (PBS) then incubated with serum-virus mixtures for 42 hrs at 37°C and 5% CO₂. Following infection, plates were submerged in 10% formalin (Fisher Scientific #SF984) for 1 hr at RT. Fixed samples were rinsed with water then permeabilized and blocked with 3% bovine serum albumin (BSA) and 0.1% Triton X-100 in PBS for 30 min at RT. Samples were then incubated with a polyclonal rabbit anti-vaccinia virus Lister strain antibody (Abbexa #abx023200) diluted 1:1,000 in 3% BSA in PBS (blocking buffer) for 1 hr at RT. All plates were washed four times with PBS then stained with 1:2,000 dilution of donkey anti-rabbit AlexaFluor647 secondary (Thermo #A-31573) and 1.25 µg/mL DAPI in blocking buffer.

Following staining, cells were washed four times with 1x PBS with 100 µL of PBS left in each well for imaging and quantification using the BioTek Cytation 7 Cell Imaging Multi-Mode Reader and Gen5 Image Prime software. MPXV neutralizing titers (ID₅₀ values) were calculated from percent inhibition using GraphPad Prism 10.0.03 non-linear regression (variable slope with four parameters) with top and bottom constraints (100 and 0, respectively).

Multiplexed immunoassay for binding antibodies and avidity

Binding and avidity of MPXV-specific IgG in serum were measured using the Luminex platform for a 12-plex assay. Eight different MPXV proteins were included: A29 (Sino Biological #40891-V08E), A30 (Cell Sciences #YVV15001A), A35 (Sino Biological #40886-V08H), B16 (Cell Sciences #YVV17401A), B21 (Cell Sciences #YVV16301A), E8 (Cell Sciences #YVV13201A), H3 (Sino Biological #40893-V08H1), and L1 (Sino Biological #40889-V07E). In addition, quality control beads were included: rubella virus (RUBV) E1/E2 (ACROBiosystems #GL2-R5583), BSA (Thermo Scientific #J65097.22), human recombinant IgG1 (Sigma-Aldrich #I5154), and IC45 (Luminex #MRP1-045-01). Differing bead regions were coupled to antigen using the xMAP Antibody Coupling Kit (Luminex #40-50016) at the following concentrations: 10 pmol/1×10⁶ beads for H3, A35, RUBV E1/E2, and human IgG1; 50 pmol/1×10⁶ beads for E8 and A29; and 100 pmol/1×10⁶ beads for A30, B21, L1, and BSA following prior optimization. Beads were multiplexed with 1,000 beads per region for each sample. All MPXV antigens are His-tagged and coupling was confirmed using an anti-His antibody (Abcam #ab27025).

Samples were analyzed by first incubating equal volumes of complement-inactivated serum (heated for 30 min at 56°C) and multiplexed beads for 1 hr in the dark on a shaker at RT. All sera were analyzed at final dilutions of 1:100 and 1:2,000, prepared in technical duplicate in 1x PBS (-Ca²⁺ -Mg²⁺) with 0.01% BSA and 0.02% Tween-20 (PBS-TB). The beads were then washed twice with PBS-TB before being incubated with either PBS (-Ca²⁺ -Mg²⁺) or the chaotropic agent, 2M ammonium thiocyanate (NH₄SCN) for 30 min in the dark on a shaker at RT; one technical replicate of each dilution was used for each condition. Beads were again washed and then incubated for 30 min as before with rabbit anti-human IgG H&L detection antibody (Abcam #ab97158) diluted to 2 µg/mL in PBS-TB. Finally, the beads were washed before a final incubation with streptavidin-PE (BioLegend #405204) diluted to 4 µg/mL in PBS-TB for 30 min as before. The beads were washed a final time and then run on a Luminex 200 instrument with sample volume set to 50 µL. Instrument settings were as follows: 50 bead minimum per region and DD gate set 6,000 to 17,000. Binding titers were calculated by measuring the area under the curve (AUC) of the median fluorescence intensity of the two dilutions that received PBS, not NH₄SCN. To calculate avidity, the binding AUC from the NH₄SCN condition was divided by the AUC from the PBS condition to determine a ratio that is reported as the avidity index. Samples were normalized by the inclusion of a positive control pool in each batch generated from pre-2022 controls with prior smallpox vaccination.

Anti-MPXV H3 Enzyme-Linked Immunosorbent Assay (ELISA)

Binding antibodies against MPXV H3 protein were measured via direct ELISA of participant serum as previously described⁷. Briefly, ninety-six well plates were coated with 0.5 µg/mL MPXV H3 protein (Sino Biological Inc., 40893-V08H1) diluted in PBS. Plates were blocked with PBS +0.05% Tween 20 (Thermo Fisher Scientific) (PBS-T) +5% non-fat milk (blocking buffer) for 1 hr at RT. Heat-inactivated serum samples were serially diluted in blocking buffer and incubated in wells for 2 hr at RT. Horseradish peroxidase (HRP)-conjugated goat-anti-human IgG (Southern BioTech, 2040-05) was diluted in blocking buffer (1:2,000) and added to each well to incubate for 1 hr at RT. Finally, plates were developed for 5 min with 3,3’,5,5’-Tetramethylbenzidine Peroxidase Substrate (Thermo Fisher Scientific) followed by 1N hydrochloric acid to halt the assay. Absorbance at 450 nm was measured on a Synergy 4 (BioTek) plate reader. Endpoint titers were determined for each sample by taking the reciprocal of the highest dilution that gives a positive reaction. Reaction positivity was determined by calculating 2 times the standard deviation plus the mean of 16 pre-2022 control samples with no prior history of smallpox vaccination. Endpoint titers were calculated by interpolating the cutoff values from a dilution curve with non-linear fit in GraphPad Prism v.9.5.1. Normalization for batch effect was addressed by the inclusion of a positive control pool generated from pre-2022 controls with prior smallpox vaccination. Sera with titers below the limit of detection (50) were scored as 25.

Statistical Methods

Paired measurements in figures are analyzed by Wilcoxon matched-pairs signed rank test. Comparisons of multiple sample groups in figures are conducted by Kruskal-Wallis test with Dunn’s method for multiple comparisons. Linear mixed-effects regressions (LMER) were used to model neutralizing titer durability with repeated measurements. The lme4 package (v.1.1-35.1) in R v.4.3.2 was used for all LMER calculations. Models were fit using the natural log transformation of the neutralizing titers. Both random intercepts only and random intercepts and slopes models were tested, but random intercepts only models were used as random intercepts and slopes models either failed to converge or had insufficient data for model fitting. Model predictions were generated using the ggeffects package (v.1.3.2). Half-lives were calculated by rearrangement of the LMER to its exponential decay form and used the predicted slope for days post dose 2. LMER model parameters for each instance are detailed in the Supplementary Materials. Correlation testing, paired measurements, and multiple comparison testing were all conducted in GraphPad Prism v.10.0.3. For all figures, p-value <0.05 is *, <0.01 is **, and <0.001 is ***.

The New York City Observational Study of Mpox Immunity

We report here on 159 MVA-BN vaccinees with or without a prior history of smallpox vaccination (Table 1) and nine mpox convalescent individuals (Supplementary Table S1). The remaining study participants are excluded due to missing data or visits. Study visits are described in Figure 1; participants could enroll at any time point. Time points include a baseline (V1), a visit between vaccine doses (V2), and three visits following receipt of the second vaccine dose: median visit dates of ∼3 weeks post dose two (V3), ∼3 months post dose two (V4), and ∼9-10 months post dose two (V5). The majority of participants were younger than 50 years old and identified as male and a member of the lesbian, gay, bisexual, transgender, and queer+ community (LGBTQ+), in line with the groups recommended to receive vaccination as outlined by the CDC¹⁹. Of note, the median age for vaccinees with a history of smallpox vaccination was 62, as compared to 33 for those without prior smallpox vaccination. Nearly a third of the participants had received a prior smallpox vaccination (referred to herein as experienced participants). Approximately half of those experienced participants were PWH, while only ∼15% of naïve participants were PWH. Due to the EUA for ID administration, OSMI includes a range of dosing combinations: SC-SC, SC-ID, ID-SC, and ID-ID. SC-ID was the most common in the study (∼33% of experienced participants and ∼46% of naïve participants).

Peak MPXV neutralization titers are comparable for naïve and experienced MVA-BN vaccinees but differ significantly in durability

To measure the neutralizing titers of OSMI participants, a fluorescence-based microneutralization assay was developed using authentic clade IIb MPXV. Pre-outbreak samples from VACV-experienced individuals have significantly higher neutralizing titers at baseline (Fig. 2A, pre-2022 controls), in line with prior work showing long-lasting durability for orthopoxvirus immunity^8–12. Aggregate data show no statistically significant difference between naïve and experienced participants at V3 (geometric mean titers [GMTs] of 95 versus 127, respectively), highlighting similar peak neutralizing titers (Fig. 2A). However, by V4, there is a significant difference between the two groups (GMTs of 45 for naïve versus 111 for experienced). This significant difference continues at V5 with 81% of VACV-naïve vaccinees below the positivity threshold, compared to only 14% of VACV-experienced vaccinees (GMTs of 24 versus 85, respectively). Nine mpox-convalescent individuals are included as another control group, some of whom had received at least one dose of MVA-BN (Supplementary Table S1). At ∼1 year after symptom onset, these participants have GMTs greater than or equal to peak titers for vaccinees, suggesting a more durable neutralizing antibody response post MPXV infection as compared to MVA-BN vaccination. These temporal patterns of neutralizing antibodies over one year are also observed for anti-MPXV H3 IgG binding antibody titers by ELISA, although a lower proportion of VACV-naïve MVA-BN vaccinees (50%) are the below the limit of detection for one-year H3-binding antibody (Supplementary Fig. S1A).

• Download figure
• Open in new tab

Figure 2: The MPXV neutralizing titers of naïve individuals wane at a significantly faster rate than those of experienced individuals

A) Aggregate mpox virus (MPXV) neutralizing titer data for each study visit sorted by prior smallpox vaccination status. Colored/black bars on each group indicate the geometric mean titer. Serum neutralizing titers were measured by a fluorescence-based microneutralization assay using authentic clade IIb MPXV. Pre-2022 controls come from the New York University Langone Vaccine Center Biorepository. Convalescent (conv) participants are sorted by whether they received MVA-BN. B) Longitudinal neutralizing titers from naïve participants across all five study visits. Bolded line indicates the mean neutralizing titer for participants at each time point. C) Longitudinal neutralizing titers from experienced participants across all five study visits. Bolded line indicates the mean neutralizing titer for participants at each time point. Horizontal, black dashed line in all panels indicates the positivity threshold for neutralization (ID₅₀ = 33.7) as based on the pre-2022 negative controls (mean of controls plus two times the standard deviation). Statistical testing conducted by Kruskal-Wallis test with Dunn’s method for multiple comparisons for panel A and Wilcoxon matched-pairs signed rank test for panels B and C. Exp, experienced vaccinees (prior history of smallpox vaccination); % neg, percent of samples below the positivity threshold.

HIV status has no significant effect on responses to MVA-BN in people with CD4 counts over 250 cells/mm³

One of the original study objectives was to determine in PWH whether the EUA dose-sparing ID route was immunogenically comparable to the licensed full-dose SC route, as prior studies of MVA-BN vaccination in PWH only examined the approved SC route^20,21. All of our participants with HIV had CD4+ T cell counts over 250 and most were over 500 (Supplementary Table S2). Combining all routes of vaccination, neutralizing titers are similar between HIV-uninfected participants and PWH in both the naïve and experienced groups (Supplementary Fig. S2). When focusing only on those that received at least one ID dose, a similar result is found (Supplementary Fig. S3). We additionally examined the effect of HIV status in a linear mixed-effects regression and similarly found no significant effect on titers in VACV-naïve (Supplementary Table S3) or VACV-experienced participants (Supplementary Table S4). Lastly, using a linear mixed-effects model, we tested whether CD4 counts had a significant effect on titers and found no such effect in naïve (Supplementary Table S5) or experienced participants (Supplementary Table S6).

Neutralizing titers decay significantly faster in naïve vaccinees compared to experienced vaccinees

To further leverage the longitudinal nature of the OSMI study, we separately examined participants with samples from consecutive visits (Fig. 2B and C). Following the first dose of MVA-BN, naïve participants already had significantly higher neutralizing titers (Fig. 2B). This significant rise continued after the second dose, with an observed peak at V3. Subsequent visits each showed significant declines. For experienced individuals (Fig. 2C), there were insufficient paired samples to gauge the effects of the first dose, but following dose two, titers did not begin to decline until after V4. A similar pattern was seen with anti-MPXV H3 IgG titers (Supplementary Fig. S1B and C). To statistically evaluate the rate of change in each group, paired fold changes in MPXV neutralizing titers for V3 versus V4 and V4 versus V5 were compared for naïve versus experienced individuals. In both instances, naïve participants have a significantly greater loss of titers from one visit to the next (Supplementary Fig. S4).

This was further borne out by modeling titers over time using a linear mixed-effects regression (Supplementary Tables S3 and S4). For naïve participants, neutralizing titers were significantly dependent on dosing interval and days post dose two (model parameters can be found in Supplementary Table S7), while neutralizing titers in experienced participants only significantly depended upon days post dose two (Supplementary Table S8). Using these models, the half-life of neutralizing antibodies is predicted to be 168 days (95% C.I.: 147 to 197 days) in naïve individuals, regardless of dosing interval. Participants that received the licensed 1-month dosing interval lost titer positivity by 160 days post dose two (Fig. 3A). The time to negativity increased by ∼1 month for each additional month between MVA-BN doses (Fig. 3A inset table). This is likely driven, in part, by the significant positive correlation of peak MPXV neutralizing titers with dosing interval (Supplementary Fig. 5). Antibody decline is far slower in VACV-experienced MVA-BN vaccinees (Fig. 3B) with a predicted half-life of 387 days (95% C.I.: 289 to 600 days). The predicted time to cross the neutralizing titer positivity threshold is >350 days, i.e., beyond the available data. Together these data highlight the difference in the longevity of serum neutralization activity for the two-dose MVA-BN regimen in VACV-naïve people versus VACV-experienced people.

• Download figure
• Open in new tab

Figure 3: Modeling of neutralizing titer durability shows naïve individuals revert to negative titers in under a year, while experienced individuals remain positive beyond a year

A) Linear mixed-effects regression of naïve participants using all data points after the second dose of MVA-BN. Regressions of natural log-transformed neutralizing titers use days post dose 2 and dosing interval as fixed effects and account for random subject effects by allowing for random intercepts. Colored regression lines with transparent ribbons indicate the predicted titers with 95% confidence interval. Model parameters are listed in Supplementary Table S7. B) Linear mixed-effects regression of experienced participants using all data points after the second dose of MVA-BN. Regression of natural log-transformed neutralizing titers uses days post dose 2 as a fixed effect and accounts for random subject effects by allowing for random intercepts. The solid black regression line with transparent ribbon indicates the predicted titers with 95% confidence interval. Model parameters are listed in Supplementary Table S8. Horizontal, black dashed line in each panel indicates the positivity threshold for neutralization (Fig. 2).

IgG responses to MPXV A35 and H3 are immunodominant regardless of smallpox vaccination history, but have lower avidity in naïve vaccinees

To further characterize the serological responses of the naïve and experienced groups, a subset of participants with multiple study visits from each group was created. The sera subsets have neutralizing titers similar to the larger study groups (Supplementary Fig. S6). Using a multiplexed immunoassay, IgG binding titers and avidity for the following eight MPXV antigens (Table 2) were measured: A29, A30, A35, B16, B21, E8, H3, and L1. These surface antigens were selected as known protective targets^22–29 or potential neutralizing targets based on previous reports^30–32. IgG avidity was measured in this assay using an additional incubation step with 2M ammonium thiocyanate following serum incubation. We found that the H3 and A35 viral proteins were immunodominant in both naïve and experienced individuals across all time points post-dose two (Supplementary Fig. S7A). Binding titers for H3 and A35 were also significantly correlated with MPXV neutralization in both groups across multiple time points (Supplementary Fig. S7B). As the VACV homologs of H3 and A35 have previously been shown to be protective antibody targets in orthopoxvirus infections^22–24, we decided to focus on these two antigens for our analysis.

Similar to MPXV neutralizing titers, both anti-H3 (Fig. 4A) and anti-A35 (Fig. 4B) IgG reach comparable peak titers in naïve and experienced individuals. IgG titers against each antigen also decline faster in naïve participants as compared to experienced vaccinees. Anti-H3 IgG begins to wane after V3 in both groups (Fig. 4C and 4D), while anti-A35 IgG begin to wane after V3 in naïve individuals (Fig. 4E), but only after V4 in experienced individuals (Fig. 4F). The timing of A35 waning more closely matches the waning of neutralizing titers in experienced vaccinees. We also measured avidity for anti-H3 and -A35 IgG. Aggregate data show significantly higher IgG avidity against H3 (Fig. 5A) and A35 (Fig. 5B) in the VACV-experienced group across all time points post dose two. Surprisingly, longitudinal samples had relatively limited changes in anti-H3 IgG avidity for VACV-naïve vaccinees (Fig. 5C). However, anti-H3 IgG avidity significantly declines from V3 to V4 and V4 to V5 in the experienced group (Fig. 5D), indicating the generation of de novo low avidity responses that reduce the bulk avidity in the sera. Changes in the avidity of anti-A35 IgG are more muted but of a similar direction (Fig. 5E and 5F). Taken together these data suggest similar B cell specificities (immunodominance) for naïve and experienced MVA-BN vaccinees, but a dramatic difference with avidity being much higher in VACV-experienced MVA-BN vaccinees. We observed no clear evidence of ongoing affinity maturation in samples post dose two in either naïve or experienced individuals.

• Download figure
• Open in new tab

Figure 4: Anti-MPXV H3 and -A35 IgG reach comparable peak titers in naïve and experienced participants, but wane more quickly in naïve participants

A) Aggregate anti-MPXV H3 IgG titers as measured by multiplexed immunoassay. Titers are reported as AUC from two serum dilutions. B) Aggregate anti-MPXV A35 IgG titers as measured by multiplexed immunoassay. Titers are reported as AUC from two serum dilutions. C) Longitudinal anti-MPXV H3 IgG titers for naïve participants at V1 and V3-5 as measured by multiplexed immunoassay. Titers are expressed as AUC from two serum dilutions. D) Same as C, but for experienced participants across V3-5. E) Longitudinal anti-MPXV A35 IgG titers for naïve participants at V1 and V3-5 as measured by multiplexed immunoassay. Titers are expressed as AUC from two serum dilutions. F) Same as E, but for experienced participants across V3-5. In panels A and B, colored/black bars indicate the geometric mean. For panels C-F, bolded line indicates the mean IgG titer for participants at each time point. Black, dashed line in all panels is positivity threshold for binding as determined by pre-2022 negative controls (same samples and method as in Fig 2). Statistical testing conducted by Kruskal-Wallis test with Dunn’s method for multiple comparisons for panels A and B, and by Wilcoxon matched-pairs signed rank test for panels C-F.

• Download figure
• Open in new tab

Figure 5: Experienced vaccinees generate higher avidity IgG against MPXV H3 and A35

A) Aggregate avidity indices for anti-MPXV H3 IgG. B) Aggregate avidity indices for anti-MPXV A35 IgG. C) Longitudinal avidity indices for anti-MPXV H3 IgG antibodies for V3-5 of naïve participants. D) Same as C, but for experienced participants. E) Longitudinal avidity indices for anti-MPXV A35 IgG antibodies for V3-5 of naïve participants. F) Same as E, but for experienced participants. Avidity was measured in a multiplexed immunoassay using a 2M ammonium thiocyanate (NH₄SCN) wash and expressed as the ratio of the AUC from the NH₄SCN condition over the AUC from a PBS control condition. In panels A and B, colored/black bars indicate the mean. Statistical testing conducted by Kruskal-Wallis test with Dunn’s method for multiple comparisons for panels A and B, and by Wilcoxon matched-pairs signed rank test for panels C-F. Exp, experienced vaccinees (prior history of smallpox vaccination).