Transmission Routes and Infection Dynamics of Salmonid Alphavirus (SAV)

Results achieved
When this project proposed in 2008 – starting 2009 – little was known about the transmission of the disease and the associated Norwegian strain of the causal agent salmon alpha virus 3 (SAV3). The project therefore aimed to reveal the relative importance of vertical and horizontal transmission, survival characteristics of the virus, and the dynamics of the infection both within and between fish populations.

These tasked were defined in three different work packages (WPs):
1. Vertical transmission
2. Shedding of SAV and virus survival in seawater
3. Models of horizontal infection dynamics and virus dispersal

WP 1: Vertical transmission
An experimental study was conducted at Veso Vikan to investigate whether SAV3 could be transmitted vertically from infected parents to their offspring. In this study the focus was on true vertical transmission which means that the virus is transmitted through infected eggs or milt directly from the parents. In this way the virus will be kept protected from any disinfection procedures after stripping/fertilization.

A group of broodstock with signs of late-stage PD and persistent RT-PCR signals for SAV were stripped and the eggs fertilized for this experiment. To evaluate potential viral contamination from the environment, fertilized ova were either not disinfected or taken through one of three different disinfection regimes. From another brood stock population tested negative for SAV/PD, fertilized eggs were routinely disinfected and used as controls.

The eggs and fry were incubated under ordinary industrial production condition and followed until smoltification. To stimulate proliferation of any latent presence of virus, all groups were exposed to external stressors at the time around start feeding. Throughout the study period selected samples were investigated by real-time RT-PCR for SAV, by histology for evidence of pathological changes consistence with PD and by serology for neutralizing antibodies against SAV. None of the samples from any of the groups showed signs of infection or presence of SAV3.

It has been concluded from this experiment that SAV 3 is not easily transmitted vertically from parents to offspring. The findings are thus in coherence with field experience that horizontal transmission is the dominating transmission pathway. As it is scientifically impossible to prove an event not to occur, a general remark would still be that proper biosecurity measures during the stripping and fertilizing procedures in the hatchery is still an important preventive measure for spreading of infectious agents (SAV3 and other agents) by potentially infected/contaminated eggs and fry.

WP 2: Shedding of SAV and virus survival in seawater
This WP concentrated on survival parameters and the monitoring over time of SAV on infected sites.

The results from WP 1 show that horizontal transmission is the most important pathway for spreading the disease. Knowledge of survival and shedding is therefore essential for implementing preventive and control measures.

Under experimental condition it was shown that SAV had a survival time at 4 and 10^oC of more than 50 days. This survival time was reduced to between 20 and 30 days at 15^oC with a quite rapid inactivation of the virus at 20^oC. This indicates that SAV may survive through most months of the year and spread through the water body surrounding the Norwegian coast. This knowledge has implications for handling solitary outbreaks in the free zone of the coast.

Using two different tank densities in an experimental cohabitation trial (additionally funded by the EU-programme NADIR), indications has been observed that the virus seem to persist longer in the higher density group. Stress indicators including the levels of cortisol, glucose, and lactase in blood were measured. The results did however not show significant differences between the groups. It is difficult to mirror experimentally the effect of densities and the stress farmed fish experience in the field, but the results may indicate an important effect of management related to disease development.

In addition, the experiment showed that infected fish shed virus through mucus and faces. This finding is consistent with previous published studies. Therefore, faces may be an important shedding route for the virus during the infectious period.

The experiment was designed for intensive sampling to monitor the rate of which the naive fish got infected. Data from these experiments give important inputs to models on how the infection behaves within a population. Such models are useful for evaluating prevention and controlling measures, i.e. the effect of manipulating population densities, vaccination and slaughtering strategy. SAV transmission parameter has yet been estimated. Due to a delayed termination of the experiment (March 2012), the analytical part is on-going. Two studies entitled “SAV3 infection dynamic study” and “Cohabitant trial for estimation of basic reproductive number (R0) of salmon alphavirus subtype3 infection in farmed Atlantic salmon in Norway” were presented in international conferences, and will be presented in peer-reviewed journals this year.

WP 3: Models of horizontal infection dynamics and virus dispersal
A stochastic dispersal model was developed to estimate to what extent a farm was at risk to infection from an outbreak site with regard to horizontal transmission. The model was fed with information on distance to neighbouring farms, disease history, local network (use of same workers, equipment, etc.) and assumptions on when a site may have been infected relative to the reporting of an outbreak. An average of 80% of the PD cases could then be explained by infection from a neighbouring infected farm and 15% by common network.

The model also indicated that infectiousness increased by the size of the infectious fish cohort and that susceptibility increases by the size of the susceptible fish cohort.

This model is useful for understanding and evaluating infectious processes and the effect of various control measures taken (e.g. vaccination, strategic slaughtering, re-location of sites).

A different approach was applied for making a quantitative risk map of PD along the coast. A Bayesian procedure was used for estimating site-specific probability of PD where the risk factors including PD history, local biomass density, and site density turn out as a predictive for PD occurrence.

Both these models clearly conclude the magnitude of biomass on sites and proximity between farms as major risk factors for PD-outbreaks.

A third approach was the development of 3D-hydrophysical model for current simulations and between-site dispersal of SAV on a 150 m grid resolution. The area for this model was Nordfjord (Sogn og Fjordane county). The model included tidal water, wind exposure and the effluent from rivers draining into the fjord. A water-contact matrix was developed for all the farms in the region showing which sites were in contact with each other by receiving water from other sites. This information may be used to organize sites into zones or strategically remove a site to break a transmission link between farms. Furthermore, it is useful in a risk based surveillance situation to identify (or rank) farms at risk of infection.

The hydrodynamic model on horizontal transmission was tested in Romsdalsfjorden using SAV 3. This is a semi-closed fjord system with significant variation in water transport that will affect the water contact pattern and also create a time dependent increase in the risk of getting infected. The model nicely showed how the PD situation in the fjord system could be explained in time by water contact patterns. A further conclusion of this work was that an increase of one month staying in a current network with a PD location increased the estimated risk of obtaining PD by 35%. This raises a question whether the time period a site is exposed for virus is more important than the virus load and how quickly interventions measures should be implemented.

Through a similar approach the hydrodynamic model on horizontal transmission was tested in Romsdalsfjorden, using SAV 3 and output from current simulations being available for the area. This is a semi-closed fjord system with significant variation in water transport that will affect the water contact pattern and also create a time dependent increase in the risk of getting infected. The model nicely showed how the PD situation in the fjord system could be explained in time by water contact patterns. Also, the risk for PD increased the shorter the time from output to sea until experiencing contact network with another PD location. Field data from two locations showing the cage-to-cage dynamic of PD (through regular sampling and PCR screening) were further collected as part of the TRISAV-project in 2011. These unique field data will provide the basis for a following-up project/paper on cage-to-cage dynamic. Most of this work has developed within a close collaboration with a PhD work at Ålesund University College, and papers will be submitted in 2012.

The models mentioned above have all focused on transmission between sites. By using field mortality data and experimental data from the challenge study (see WP 2) disease transmission parameters are estimated to develop a model describing the disease development within a population (a cage and/or a site). By establishing knowledge on the spreading capacity, intervention measures can be tested with regard to their efficiency. Two papers from this work will be submitted in 2012.

Models are usually data hungry calling for a need to do research for parameter estimation. In this project various minor studies have been performed to generate such knowledge. Survival temperature was tested experimentally. Field data indicate that within the survival temperature window, the most significant effect of temperature is the change in temperature either increasing or decreasing. This may be reflecting the occurrence of PD being more frequent in spring and autumn through the summer months. The study indicates that if a site is exposed to infection when the temperature is increasing, the PD develops faster compared to periods when the temperature is decreasing. These results will be presented at ISVEE 2012 and as separate papers.

SAV is excreted through mucus and faces – through the lateral line and gills. This is seen both in experiments and in the field. Besides, SAV has been isolated from salmon fat, including the tiny fat layer on the water surface. There is so far no registration of SAV3 in sediments or been isolated from wild non-salmon fish or other invertebrates at infected sites. The conclusion is that SAV 3 primarily seems to be associated to the salmon itself and substances of salmon nature at marine farming sites.

Conclusions
The project has concluded through various models that the major component for SAV3 transmission is through horizontal pathways. The most important risk factor for a farm to be infected is contact through water current to a neighbouring infectious farm (“down stream”). The effect of such contact is influenced of local biomass density and site density. High cage density indicates an increase in the persistence of an SAV-infection. Adding local contact network (workers and equipment), these risk factors explains almost 100% of the risk.

The study failed to demonstrate any true vertical transmission and believe it to be of minor importance for the spreading of PD.

The project shows that SAV3 is shed through mucus and faces. Its potential survival time is more than 50 days at a temperature interval that probably makes it viable through most of the year in the coastal water of Norway.

The project has generated new knowledge on important input factors used in dynamic disease modeling. The models have all shown the importance of horizontal spreading between sites. These models together with models for within site dynamics are useful tools for evaluating interventions to prevent and control the spread of the disease and may be used in the future to help deciding upon infrastructure and predicting sites at risk.

The project has shown its relevance as being an important input source to the evaluation of the PD strategy just requested by the Norwegian Food Safety. And even more so, as one now see an emerge of the SAV2-like subtype epidemic in Norway, the project has just in time created tools that may be of help in this new situation to give more specific management advice to industry and authorities.

Papers submitted /in preparation for submission by 2012
• Risk map and spatial determinants of pancreas disease in the marine phase of Norwegian salmonid farming sites, submitted to Veterinary Research.
• Integration of hydrodynamics into aquaculture management, validation of practice and assessment (in preperation).
• Using mathematical models to evaluate preventive and control strategies for spreading of salmon alphavirus (SAV) subtype 3 within a farm of Norwegian Atlantic salmon (in preperation).
• Basic reproductive number (R0) and a spread of salmon alpha virus (SAV) subtype 3 within cages of farmed Atlantic salmon in Norway (in preperation).
• Effect of contact networks on transmission of pancreas disease between marine farming sites for Atlantic salmon (in preperation).
• Temperature dynamics influences dynamics of Pancreas disease at marine production sites for salmon (Salmo salar L.) (in preperation).
• Transmission dynamics of Pancreas disease at marine farming sites for salmon (Salmo salar L.) (in preperation).

The project has been co-financed through the Norwegian Research Council (NRC). NRC has been responsible for ensuring research quality and for administrative project co-ordination. For further details and publications, see also the project page at NRC (NRC project no.: 190484).