Development of a virulent O'nyong'nyong challenge model to determine heterologous protection mediated by a hydrogen peroxide-inactivated chikungunya virus vaccine

Abstract

O'nyong-nyong virus (ONNV) is a mosquito-transmitted alphavirus identified in Uganda in 1959. The virus has potential for enzootic and urban transmission cycles, and in humans, ONNV infection manifests as fever, rash, and joint/muscle pain that can persist. There are currently no specific vaccines or antiviral treatments for ONNV. Since highly passaged alphaviruses often lose pathogenic features, we constructed an infectious clone for ONNV-UVRI0804 (ONNV0804), a 2017 isolate from a febrile patient in Uganda. Viral replication for ONNV0804 was compared to the highly passaged strain, ONNVUgMP30, and ONNVUgMP30 replicated to higher levels in human dermal fibroblasts and Vero cells, but both viruses replicated similarly in C6/36 and mouse embryonic fibroblast cells. We performed a head-to-head comparison of in vivo virulence in both immunocompetent C57BL/6 mice and interferon deficient AG129 mice. In both mouse strains, ONNV0804 was substantially more pathogenic than ONNVUgMP30. Unlike ONNVUgMP30, ONNV0804 caused significant footpad swelling and broader tissue distribution with higher vRNA loads at both 5- and 43-days post-infection (dpi) relative to ONNVUgMP30. This finding indicates that ONNV can persist in joint and muscle tissues for long periods of time, which has been associated with chronic arthritogenic human disease. In AG129 mice, ONNV0804 caused a more rapid onset of disease, higher viremia, and a >800-fold increase in virulence. Previous studies have shown that CHIKV infection or vaccination can provide cross-reactive immunity to ONNV. To determine if a CHIKV vaccine can protect against the more virulent ONNV0804 strain, we vaccinated mice with a hydrogen peroxide-inactivated CHIKV vaccine, HydroVax-CHIKV. Neutralizing antibody titers were determined against ONNV0804 and CHIKV and animals were challenged with ONNV0804. An optimized two-dose vaccination regimen of HydroVax-CHIKV protected against lethal infection and reduced virus-associated arthritogenic disease. These data indicate that we have developed new and robust models for studying severe ONNV disease and that HydroVax-CHIKV vaccination can protect against infection with a highly pathogenic contemporary strain of ONNV.

Fig 1. ONNV strain genetic comparison, growth characteristics in four cell lines, and antiviral inhibition.

(A) Maximum likelihood phylogenetic tree constructed in MEGA software using the Dayhoff model and structural protein (C/E3/E2/6K/E1) amino acid sequences from all ONNV strains with available complete genomes and selected related alphaviruses of the Semliki Forest virus complex. ONNV₀₈₀₄ and ONNV_UgMP30 strains are indicated in red outlined boxes. (B) Summary of amino acid differences between ONNV Gulu UgMP30 and UVRI0804 strains. The nsP3 opal stop codon is in bold lettering. Growth kinetics of ONNV_UgMP30 (red) and ONNV₀₈₀₄ (blue) strains in (C) C6/36, (D) MEF, (E) Vero, and (F) NHDF cell lines. In three individual experiments, cells were infected at a multiplicity of infection equal to 0.5 in triplicate wells. Viral supernatants were collected at the indicated timepoints in hours post-infection (hpi) and titered by plaque assays. Titers are reported in plaque forming units (PFU) per mL of viral supernatant. The dotted line represents the limit of detection at 33.3 PFU/mL. Mean values from the three experiments and error bars with standard deviation are plotted and analyzed by multiple paired t tests with Holm-Šídák’s multiple comparisons where not significant (ns) P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P < 0.0001. (G) Antiviral effects of CHIKV nsp2 protease inhibitor, RA-00002034, against ONNV₀₈₀₄ and ONNV_UgMP30 strains. NHDFs were treated with DMSO control or 2-fold serial dilutions of RA-00002034 ranging from 40-0.019 µM for 1 hour prior to infection with ONNV₀₈₀₄ or ONNV_UgMP30 (MOI = 0.5 PFU/cell). At 2 hpi, cells were washed, and fresh medium was added containing compound. At 24 hpi, supernatants were collected, and infectious virus was quantified by plaque assay. Limit of detection was plotted as 10 PFU/mL.

Fig 2. ONNV pathogenesis, disease, and viral persistence in immunocompetent mice.

(A) Overview of study design. C57BL/6 mice (n = 5/group) were inoculated with a low or high dose of either ONNV strain in the right footpad (s.c.) then (B) body weight and (C) footpad swelling were monitored for 14 days. An additional group of mice (n = 5/group) were inoculated with 10³ PFU of either ONNV strain or PBS, tissues were collected at 7 dpi and perfused with 4% PFA for (D) H&E histological staining with ellipses designating areas undergoing inflammation and (E) inflammation grading on a scale of 0 to 10 with 0 indicating no inflammation and 10 indicating the most severe inflammation (see methods). Mean and SEM are plotted and analyzed by two-way ANOVA with Tukey’s multiple comparisons. (F) For comparison of viral replication in tissues, mice (n = 5/group) were inoculated with 10³ PFU of either ONNV strain and ankles, calf muscles, quadricep muscles, spleen, heart, and brain were collected for vRNA quantification by qRT-PCR. Data in (F) are log-transformed, the mean and standard error are plotted, and data are analyzed by two-way ANOVA with Šídák’s multiple comparisons where ns P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P < 0.0001. Only significant comparisons are shown.

Fig 3. ONNV RNA persistence at 43 dpi and the development of neutralizing antibodies in immunocompetent mice.

C57BL/6 mice (n = 5/group) were inoculated in the right footpad (s.c.) with a low (10³) or high (10⁵ or 10⁷ PFU) dose of ONNV₀₈₀₄ or ONNV_UgMP30. The animals were humanely euthanized at 43 dpi for the detection of vRNA in tissues from mice challenged with each ONNV dose (A). Total RNA was processed from tissue lysates and vRNA data are log-transformed; the mean and standard error are plotted. Neutralizing antibodies against ONNV and related alphaviruses by 50% plaque reduction neutralization test (PRNT₅₀) in serum from mice challenged with the higher challenge doses only (n = 5 per group except in assays intervals against MAYV (n = 3) due to limited serum volume) (B). Geometric mean titers (GMT) of neutralizing antibody from mice infected with 10³ PFU of ONNV₀₈₀₄ or 10⁷ PFU of ONNV_UgMP30 are shown with error bars that represent 95% confidence intervals. The neutralization titers are analyzed by mixed-effects analysis two-way ANOVA with Šídák’s multiple comparisons. The vRNA persistence data are analyzed by two-way ANOVA with Šídák’s multiple comparisons. Only significant comparisons are shown in the figure (ns P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P < 0.0001).

Fig 4. ONNV₀₈₀₄ is more virulent than ONNV_UgMp30 in AG129 immunodeficient mice.

AG129 mice were infected with a range of doses of ONNV_UgMP30 (n = 3–5/dose except n = 1 at 1 PFU) or ONNV₀₈₀₄ (n = 3–9/dose) and monitored for 14 days after challenge. Kaplan-Meier survival curves for (A) ONNV_UgMP30 and (B) ONNV₀₈₀₄ with the calculated 50% humane endpoint dose (HE₅₀). The humane endpoint was defined as loss of 25% of total body weight or observance of severe lethargy. Changes in body weight over 14 days for (C) ONNV_UgMP30 and (D) ONNV₀₈₀₄. Changes in footpad swelling over 14 days for (E) ONNV_UgMP30 and (F) ONNV₀₈₀₄. (G) Comparison of time to humane endpoint compared by two-way ANOVA with Šídák’s multiple comparisons. Animals that survived infection are plotted at 15 days to humane endpoint. (H) Serum collected at 3 dpi was titered in triplicate by plaque assays and mean values for each mouse are plotted. The X’s in panels (G, H) indicate doses that were not evaluated. The LOD for this assay is 33.3 PFU/mL. Serum titers are log-transformed and analyzed by two-way ANOVA with Šídák’s multiple comparisons where ns P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P < 0.0001.

Fig 5. HydroVax-CHIKV immunization elicits antibodies that cross-neutralize ONNV₀₈₀₄ and cross-protect against lethal arthritogenic disease in AG129 mice.

(A) Study schematic. AG129 mice aged 18–21 weeks old were immunized in the left leg with 0.1% Alum alone (Group 1, vehicle control, n = 8) or 5 μg of HydroVax-CHIKV adjuvanted in 0.1% Alum (Group 2, Prime-only, n = 9) at the indicated timepoints. Mice in Group 3 (Prime-boost, n = 9) were boosted with the same dose 28 days later. Serum was isolated two days prior to viral challenge for assessment of neutralization titers. Animals were challenged in the right footpad with 1,700 HE₅₀ (10 PFU) of ONNV₀₈₀₄ then blood samples were drawn for quantification of viremia at 3 dpi by plaque assay. Animals were monitored daily for changes in footpad swelling and body weight with the humane endpoint defined as loss of 25% of total body weight or observance of severe lethargy. (B) Homotypic CHIKV_181/25 neutralization and (C) heterotypic ONNV₀₈₀₄ neutralization titers with geometric mean shown, by 50% plaque reduction neutralization test (PRNT₅₀). (D) Kaplan-Meier survival curve with log rank Mantel-Cox (****P < 0.0001). Changes in (E-H) footpad swelling and (I-L) body weight for each group up to 14 dpi. Compiled footpad and body weight mean with SEM are plotted in (H) and (L), respectively. Surviving animals were humanely euthanized at 35 dpi and serum was collected to determine (B, C) neutralizing antibody titers. Data in (B, C) are log-transformed and are analyzed by two-way ANOVA with Tukey’s multiple comparisons where ns P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P < 0.0001. Data in panels (H) and (L) were analyzed by mixed-effect analysis with Dunnett’s multiple comparisons and only significant comparisons compared to group 1 controls are shown.

Fig 6. HydroVax-CHIKV vaccination partially cross-protects against 170,000 HE₅₀ of ONNV₀₈₀₄ in AG129 mice.

(A) Study schematic. AG129 mice aged 18–21 weeks old were immunized in the left leg with 0.1% Alum alone (Group 4, Vehicle control, n = 5) or 5 μg of HydroVax-CHIKV adjuvanted in 0.1% Alum (Groups 5 and 6, Prime-only, n = 6) at the indicated timepoints. Mice in Group 7 (Prime/boost, n = 6) were boosted with the same dose 30 days later. Serum was isolated two days prior to viral challenge for assessment of (B) homotypic CHIKV_181/25 neutralization and (C) heterotypic ONNV₀₈₀₄ neutralization titers by PRNT₅₀, geometric mean titers (GMT) are shown. Animals were challenged in the right footpad with 170,000 HE₅₀ (10³ PFU) of ONNV₀₈₀₄ then blood samples were drawn for quantification of (D) viremia at 3 dpi by plaque assay. In B-D, GMT is shown. (E) Kaplan-Meier survival curve with log rank Mantel-Cox (**P = 0.0018, ****P < 0.0001). Animals were monitored daily for changes in (F-J) body weight and (K-O) footpad swelling with the humane endpoint defined as loss of 25% of total body weight or severe lethargy. For (B-C), data are analyzed by Kruskal-Wallis test with Dunn’s multiple comparisons and only significant comparisons are shown. In (D), data are analyzed by ANOVA with Tukey’s multiple comparisons. In (J, O), data were analyzed by mixed-effect analysis with Dunnett’s multiple comparisons and only significant comparisons compared to group 4 controls are shown. Only significant comparisons are shown: ns P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P < 0.0001.