One of the greatest challenges in vaccine development is to optimize antigen dose in order to generate a threshold of antibody titer that can be used as a measure of protective immunity. Unlike studies done in humans where challenge models are often not applicable for ethical reasons , in this study we show that the antigen dose used in vaccine formulation corresponded with PCSP in vaccinated fish after challenge with IPNV. This demonstrates that it is possible to establish a threshold of antigen dose required to protect fish from mortality that would serve as a measure of vaccine efficacy. In addition, antibody levels obtained after vaccination but prior to challenge corresponded with PCSP, further augmenting the possibility of determining a threshold end point of antibody titer that can serve as a signature of protective immunity. This was in line with assertions made by Pulendran et al.  that when a protective threshold of antibody titer is achieved or exceeded, vaccination is assumed to have reached a signature of protective immunity. In higher vertebrates, these tests have been standardized so that signatures of protective immunity based on thresholds of neutralizing antibodies have been determined for different pathogens . However, no similar threshold units of antibody titers have been established in teleosts. In the present study, an antigen dose of 2 × 1010 TCID50/mL yielding a mean antibody titer ≥ 1.4 OD490 at challenge attained a PCSP ≥ 90% after challenge with the highly virulent Norwegian Sp strain NVI-015 in the HiAg group. Therefore these findings demonstrate that at these threshold units of antigen dose (2 × 1010 TCID50/mL) and antibody titer (≥ 1.4 OD490) vaccination against IPNV infection attained high protection (PCSP > 90%) in vaccinated fish. However, it is interesting to note that reducing the threshold unit of antigen dose by one log (101), lowered the antibody titer by threefold and decreased protection to less than half (PCSP ≤ 42%) as shown in the LoAg group that was vaccinated with an antigen dose one log lower than the HiAg group. Overall, these findings demonstrate that in vaccinology, a suboptimal antigen dose one log (101) lower than the protective threshold unit can lead to significant differences in vaccine protection.
Progression of antibody responses for fish used in this study can be divided into two stages namely the immune induction and post challenge periods. Immune induction covered the period between vaccination up to the time of challenge at 8 wpv in which fish developed antibodies in response to vaccination. At challenge (8 wpv), antibody levels in the HiAg were threefold higher than levels in the LoAg suggesting that post vaccination antibody responses were antigen dose dependent. The post challenge period covered the incubation and acute stages of infections. During the incubation period, antibody levels in both the HiAg and LoAg groups declined, which could be attributed to their utilization in virus neutralization after challenge. Neutralization of virus by antibodies could have played a major role in delaying the onset of post challenge mortality in vaccinated fish unlike the control group in which no antibodies were present to neutralize replicating virus; fish therefore started dying by 19 dpc. It is likely that mortality in the LoAg group was delayed until after depletion of antibodies which gave way to an increase in virus replication by 21 dpc. High antibody levels in the HiAg group could have played a major role in suppressing the increase in virus replication during the acute stage of infection leading to delayed onset and reduction of mortality in this group. While antibodies in the vaccinated groups declined due to their usage in virus neutralization, control fish that had no antibodies prior to challenge developed antibody responses that increased to higher levels during the acute stage of infection as an indication of a primary response of exposure to the challenge virus. This could explain reasons why the LoAg group had lower antibody levels than the control group during acute infection (21–56 dpc). The major difference between the vaccinated and control fish is that antibodies from vaccinated fish were generated in response to vaccination from inactivated vaccines prior to challenge while antibodies from control fish were generated in response to the challenge virus. Hence, antibodies from control fish could not have played a major role in post challenge protection because they were generated after exposure to the challenge virus. On the contrary, antibodies from vaccinated fish could have played a major role in post challenge protection because they were generated prior to challenge from inactivated vaccines for protection against the challenge virus.
One of the prime objectives in vaccination is to prevent establishment of infection soon after exposure to the virus. Overall infection rates of the current study were higher in control fish during the incubation period, followed by the LoAg group that had low antibody levels with the lowest infection rates being in the HiAg that had the highest antibody levels. This was consistent with our earlier observations that fish vaccinated with less potent vaccines get infected much earlier with higher infection rates than fish vaccinated with more potent vaccines . Given that increase in antibody levels corresponded with decrease in overall infection rates, it is likely that circulating antibody levels played an important role in preventing dissemination and seeding of virus in the predilection sites in the HiAg group. This was supported by virus re-isolation data in which the number of fish having viraemia was the lowest in the HiAg group and was consistent with our previous study in which we showed that high antibody levels were correlated with the absence of viraemia, unlike the less protective vaccines that had early onset of viraemia and linked to high infection rates . Another important observation made from this study was that although the pancreas and liver are target organs prone to tissue damage, the seeding of virus in the liver was delayed until the acute stage of infection for most fish while infections in the pancreas occurred early during the incubation period. Furthermore, these data show that for fish that were only infected in the pancreas without involvement of the liver, infection did not proceed to cause pathological disease. These findings point to the importance of involvement of the liver during acute infection. This was clearly demonstrated in the HiAg group that had low infection rates in the liver and yet the pancreas had 100% infection rates, but because the liver was not infected in most fish, no pathological disease developed. Consequently, the seeding of virus coupled with tissue damage in the liver corresponded with an increase in post challenge mortality in the LoAg and control groups during acute infection. Conversely, the presence of high antibody levels in the HiAg group corresponded with low infection rates in the liver. Although we did not determine the role of the cellular mediated immune response in eliminating virus infected cells, these observations suggest that high antibody levels could play an important role in preventing the seeding of virus in the liver as a protective mechanism against tissue damage. This was in line with observations made elsewhere that preventing deposition of virus in target organs prevents the establishment of clinical disease [36, 37]. However, it is important to point out that it is possible that other arms of the immune system not evaluated in the present study could have contributed to reducing post challenge mortality. This was demonstrated in the LoAg group that had lower antibody levels and higher viral copy numbers at 21 dpc than the control group and yet this group had less mortality than the control group.
As pointed out by Plotkin  the correlate of clinical disease is not necessarily the same correlate that predicts the cut-off against infection. In the present study, infection was detected in tissues with mean viral copy numbers ≤ 102.6 which corresponded with the expression of innate immune genes. However, these viral copy numbers were inadequate to cause tissue damage in target organs. It is only when viral copy numbers exceeded 107.0 that tissue damage was detected, implying that this cutoff could serve as a correlate of pathology in target organs. By determining the quantity of viral copy numbers required to establish pathological disease, immunization can be optimized to obtain antibody levels able to prevent tissue damage. For example, in the case of measles vaccines , antibody titers > 200 mIU/mL were protective against infection, whereas titers between 120–200 mIU/mL were protective against clinical disease but not infection. Titers <120 mIU/mL were not protective at all. Although we did not determine the cut-off that prevents establishment of infection, our findings show that at 1.4 OD490, antibody levels corresponded with the absence of tissue damage in the HiAg group. In terms of tissue susceptibility, data presented here show that limiting virus replication in the liver and pancreas corresponded with failure to establish pathological disease during acute infection. This was shown in the HiAg group that had low viral copy numbers linked to the absence of tissue damage. Conversely, copy numbers ≥ 107.0 in the liver and pancreas of the LoAg and control groups were linked to tissue damage and high mortality. Although we did not rule out the possible involvement of the cellular mediated immune response in eliminating virus infected cells, these data suggest that antibodies could play a pivotal role in suppressing virus replication to levels below the correlate of pathology as a mode of protection against tissue damage and mortality.
Innate immune genes like IFN and IFN-inducible genes (IGS) are known to be the first line of defense against viral infections [39–41]. Expression of IFNα and Mx correlated with establishment of infection in the control group during the incubation period indicating that upon infection, IPNV induces the expression of early immune genes. Expression of these genes did not only correlate with infection at predilection sites, but was also highly correlated with an increase in viral copy numbers suggesting that these genes could serve as biomarkers of tissue tropism as well as indicators of increase in virus replication. This entails that, instead of tracking virus distribution and monitoring increase in replication capacity, expression of these genes could serve as alternative indicators of tissue tropism and an increase in viral copy numbers. For example, high viral loads in the liver during acute infection correlated with high expression levels of both genes while low infection rates during the incubation period were characterized by low expression levels. In vaccinated fish, up-regulation of these genes was indicative of low efficacy or vaccine failure as shown in the LoAg group in which an increase in viral copy number during acute infection was marked with high levels of IFNα and Mx. Conversely, down regulation of these genes in the HiAg was indicative of high protection linked to low viral copy numbers. Here, we demonstrate for the time, that the kinetics of IFNα and Mx expression follow the distribution and replication of IPNV making these genes potential biomarkers of infection progression as well as indicators of vaccine efficacy.
In summary, this study has shown that the mechanism of vaccine protection against IPNV infection is a stepwise process in which vaccine induced protection reduces post challenge infection rates as a first step. This implies that some fish do not get infected and for fish that get infected, the next mode of protection is to prevent seeding of virus in target organs, particularly in the liver. Finally, for fish that get infected in target organs, protection is by reducing virus replication to levels below the correlate of pathology. Determining the correlate of pathology based on viral copy numbers at which tissue damage is established provides a benchmark against which vaccine protection can be measured. Use of innate immune genes as indicators of tissue tropism and increase in virus replication provide the first methodological evidence that gene signatures can be used as biomarkers of IPNV infection progression in salmonids. We demonstrate that antigen dose in vaccine formulation corresponds with vaccine efficacy and that antibody levels can be used as a signature of protective immunity against pathological disease and mortality. Overall, these findings show that by streamlining testing procedures, immunological thresholds that reliably protect fish from disease can be established in IPNV vaccinology.