A Systematic Approach towards Optimizing a Cohabitation Challenge Model for Infectious Pancreatic Necrosis Virus in Atlantic Salmon (Salmo salar L.)

Abstract

A cohabitation challenge model was developed for use in evaluating the efficacy of vaccines developed against infectious pancreatic necrosis virus (IPNV) in Atlantic salmon (Salmo salar L) using a stepwise approach. The study involved identifying a set of input variables that were optimized before inclusion in the model. Input variables identified included the highly virulent Norwegian Sp strain NVI015-TA encoding the T217A221 motif having the ability to cause >90% mortality and a hazard risk ratio of 490.18 (p<0.000) for use as challenge virus. The challenge dose was estimated at 1x10(7) TCID50/mL per fish while the proportion of virus shedders was estimated at 12.5% of the total number of fish per tank. The model was designed based on a three parallel tank system in which the Cox hazard proportional regression model was used to estimate the minimum number of fish required to show significant differences between the vaccinated and control fish in each tank. All input variables were optimized to generate mortality >75% in the unvaccinated fish in order to attain a high discriminatory capacity (DC) between the vaccinated and control fish as a measure of vaccine efficacy. The model shows the importance of using highly susceptible fish to IPNV in the optimization of challenge models by showing that highly susceptible fish had a better DC of differentiating vaccine protected fish from the unvaccinated control fish than the less susceptible fish. Once all input variables were optimized, the model was tested for its reproducibility by generating similar results from three independent cohabitation challenge trials using the same input variables. Overall, data presented here show that the cohabitation challenge model developed in this study is reproducible and that it can reliably be used to evaluate the efficacy of vaccines developed against IPNV in Atlantic salmon. We envision that the approach used here will open new avenues for developing optimal challenge models for use in evaluating the efficacy of different vaccines used in aquaculture.

Fig 1. High and low virus challenge dose.

Kaplan Meyer’s (KM) survival analysis of Atlantic salmon challenged using a high and low challenge dose carried out using strain NVI015-TA (Study II). (A) KM survival analysis for the high challenge dose (HC_dose), 1x10⁷ TCID₅₀/fish. (B) KM survival analysis for the low challenge dose (LC_dose), 1x10⁵ TCID₅₀/fish. Fig 1A shows a wider discriminatory capacity (DC>58%) compared to Fig 1B (DC = 40%) between the vaccinated and unvaccinated control fish.

Fig 2. Proportion of virus shedders.

KM survival analysis for cohabitees challenged with different (12.5%, 20% and 30%) proportions of virus shedders of the total number of fish per tank (N = 90; Study III). All virus shedders were injected with 1x10⁷ TCID₅₀/fish with strain NVI015-TA. Mortality of shedders in the 10% and 30% proportion groups started on day 18 after challenge while for the 20% group day 21 after challenge was the first day of mortality. Overlaps in the KM survival curves for the cohabitees show that there was no statistical difference between the 10% and 20% (p>0.368), as well as between the 10% and 30% (p>0.1051) proportion of virus shedders of the total number of fish per tank.

Fig 3. Effect of genetic susceptibility.

KM survival analysis for different strains of Atlantic salmon having different susceptibility to IPNV infection, challenged with strain NVI015-TA at 1x10⁷TCID₅₀/fish. (A) KM survival analysis for the highly susceptible strain (HS_strain). (B) KM survival analysis for the less susceptible strain (LS_strain), both including vaccinated and unvaccinated control fish. Note that the discriminatory capacity (DC) between the vaccinated and control fish for the HS_strain shown in Fig 3A (DC = 58.89%) was more than twice that of the LS_strain shown in Fig 3B (DC = 24.22%).

Fig 4. Salmon production cycle.

The production cycle of Atlantic salmon depicting the timing of IPN vaccination, challenge and convalescence. A indicates the immune induction period being the period between vaccination and challenge. Note that vaccination is carried out at the parr stage in freshwater before smoltification. B indicates the time of challenge at the post-smolt stage soon after transfer to seawater, which conforms to the time when most outbreaks occur during the Atlantic salmon production cycle. C shows the period after lethal challenge when the survivors become carriers or persistently infected with post challenge virus during convalescence.

Fig 5. Experimental timing.

Sampling time-point established for the cohabitation challenge model segmented into A and B. (A) shows the immune induction period (I and II), between vaccination and challenge, depicting the progression of antibody responses for fish immunized using a commercial vaccine (Pharmaq AS). Note that antibody responses increased by threefold from (I) during the early antibody responses when levels were at 0.326 OD₄₉₀ at 330 degree days (dd) to (II) pre-challenge antibody responses when levels were at 1.221 OD₄₉₀ at 672 dd. (B) shows the post challenge period depicting the progression of cumulative mortality in the vaccinated and control fish. Cumulative mortality for both the vaccinated and control fish progressed in three stages namely (III) lag-phase, (IV) exponential phase, and (V) plateau phase. Blue arrows below show the sampling time-points during immune induction and the post challenge period.

Fig 6. Cohabitation challenge model design.

Shows a schematic design of a cohabitation challenge model based on a three parallel tank system with the number of fish estimated using the Cox PH regression model in which the statistical power function for sample size estimate was set at 80%, 95% CI and hazard risk (HR) ratio set 0.5. In each tank, 36 fish were vaccinated using a commercial vaccine (PHARMAQ AS, Oslo) and another 36 were injected with phosphate buffered saline (PBS) to serve as unvaccinated controls. After vaccination fish were kept at 12°C and left for smoltification. The study design was divided into immune induction, challenge period and convalescence stages. After smoltification, challenge was carried out by adding 9 virus shedders per tanks, which is 12.5% of the total number of fish per tank, and the virus shedders are injected with 1 x 10⁷ TCID₅₀/fish of strain NVI015-TA. The immune induction period was for evaluating immune responses prior to challenge while the challenge period enables evaluation of PCSP/RPS. The convalescence period covers the period after acute infection when fish stopped dying and enabled evaluation of post challenge recovery as well as assessing the number of virus carriers linked to persistent infections.

Fig 7. Mortality in non-vaccinated controls.

KM survival analysis for comparison of PCSP induced by strain NVI015-TA in the control fish challenged using a three parallel tank system described in Fig 6. Mortality in Tanks 1, 2 and 3 started at 19, 24, and 21 days post challenge, respectively. Overlaps in the KM survival curves for control fish from Tanks 1, 2 and 3 show that there was no statistical difference (p = 0.8223) in PCSPs for the three tanks.

Fig 8. Immunohistochemistry of target organs.

Immunohistochemistry (IHC) of fish infected with strain NVI015-TA. Red stain shows presence of IPNV in the infected tissues. (A) Exocrine pancreas with condensed and pyknotic nuclei (A,B,C) are indicative of cellular degeneration and necrosis. (B) Liver with vacuolation (A) and cells with apoptotic nuclei (B) and necrotic liver cells (C).