1. Naturally Occurring Temperature-Sensitive WT Influenza Viruses
An important characteristic of any virus is its
non-ts phenotype—an ability for replication at the elevated temperatures of 38–40 °C, which exceed the upper limit of optimal values. In the past, it was thought that the typical WT virus is always
non-ts and that this property determines viral virulence. In those times, primary screening of reassortant LAIV candidates was based on
ts/ca attenuation markers
[1][2]. Reassortants that did not contain suitable laboratory markers of attenuation (
ts/ca phenotype) were screened out; only then, analyses of the genome composition, which at that time was quite complex, started. The
ts-phenotype of the reassortant LAIV candidate is critical because
ts viruses cannot multiply at the temperature of the lower respiratory tract.
The first mention of natural
ts WT influenza viruses can be found in publications from the 1980s
[3][4][5][6][7]. Later, it was found that changes in the
ts/non-ts phenotype have a regular wave-like nature
[8][9][10]. At the beginning of each influenza pandemic/epidemic cycle, the circulation of
non-ts viruses was detected. Further evolution is leading to the change in
non-ts with
ts variants. The prevalence of
ts strains in circulation indirectly indicates that novel
non-ts viruses are expected to appear in circulation.
Thus, the permanent circulation of
ts viruses can be considered as a precursor for the appearance of an antigenically distinct virus, seasonal or even pandemic. This assumption is supported by the fact that just before the 2009 pandemic, a kind of “calm before the storm” was noticed: only
ts influenza A(H1N1), A(H3N2), and B viruses were detected in circulation
[8].
The
ts phenotype of the WT virus does not influence the efficiency of reassortment but interferes with the efficacy of primary screening of the egg-derived LAIV candidate, since one of the laboratory selective markers of attenuation, the
ts marker, is lost. The problem of the existence of
ts viruses lies elsewhere—since
ts viruses are at the end of the
ts wave of circulating viruses, they may be less immunogenic than their
non-ts counterparts, whose circulation started this wave. It has been suggested that
non-ts WT parent viruses may enhance the immunogenicity of LAIV and vice versa; LAIVs based on
ts WT parent viruses were of low immunogenicity. Unfortunately, this study only tested a limited number of vaccines
[10].
It seems reasonable that new potential candidates for vaccine strains should be evaluated not only in terms of the novelty of surface antigens but also taking into account temperature sensitivity in their replication.
2. Naturally Occurring Cold-Adapted WT Influenza Viruses
In nature, not only natural temperature-sensitive viruses circulate, which are numerous, but sometimes
ca viruses also appear, which usually possess the
non-ts phenotype
[11]. Unlike natural
ts viruses, there are so few that it is not possible to make any assumptions about the reasons and regularities in their appearance. In fact, the term “
ca” (cold-adapted) is not quite appropriate in this case, since these viruses were not adapted to low temperatures by laboratory manipulations. It would be more accurate to talk about WT viruses that sufficiently replicate at low temperatures or WT viruses that are naturally resistant to low temperatures. The role of cold-resistance for replication of some natural isolates has not been studied yet; however, it can be assumed that
non-ts/ca WT viruses, which can reproduce in a very wide temperature range, from 25 °C to 40 °C, can effectively infect both the upper and lower respiratory tract.
Unlike ts WT viruses whose ts phenotype does not influence the efficiency of reassortment, the ca phenotype of the WT parent dramatically disturbs the first steps in the reassortment process that are carried out at a low temperature of 25–26 °C. A loss of the ca-selective factor may lead to a significant increase in the total number of ca reassortants, but the overall number of reassortants with the desired 6:2 genome composition is decreasing.
3. Sensitivity of WT Viruses to Nonspecific Thermostable Serum γ-Inhibitors
WT influenza viruses exhibit marked differences in their sensitivity to nonspecific thermostable γ-inhibitors due to their distinguishable receptor specificity. H3N2 viruses, which preferentially bind the α-2,6 receptors, are very sensitive to serum thermostable γ-inhibitors, while H1N1 strains with α-2,3 or mixed α-2,3/α-2,6 specificity exhibit an inhibitor-resistant phenotype
[12][13][14]. Before the 1970s–1980s, the majority of influenza B viruses possessed an inhibitor-resistant phenotype. In the 1980s, they diverged into two distinct genetic lineages, B/Victoria and B/Yamagata
[15]. Since then, there have been two parallel evolutionary pathways of influenza type B in the human population
[16]. After separation, the B/Victoria lineage viruses retained a high level of inhibitor resistance of past strains. Contrarily, viruses of the B/Yamagata lineage acquired high inhibitor sensitivity
[17].
The standard scheme for the preparation of vaccine strains by the method of classical reassortment includes the use of anti-MDV serum
[18][19][20]. This provides a selective advantage for reassortants in inheriting HA and NA from an antigenically relevant WT virus. However, the selection of 6:2 reassortants, based on inhibitor-susceptible WT viruses, can be complicated by nonspecific binding of their HA by γ-inhibitors, which are presented in the anti-serum against MDV
[14].
The data presented in
[14] were used for drawing
Figure 1. Analysis of genome composition of 883 reassortants, obtained by classical reassortment in eggs of MDVs with 40 WT influenza viruses, which possessed a different degree of sensitivity to nonspecific γ-inhibitors, revealed the following consistent pattern: all reassortants inherited WT HA; nevertheless, the belonging of the remaining genes to WT or MDV parents varied
[14]. The majority of reassortants based on inhibitor-resistant WT viruses (88.7%) inherited NA from the WT parent; also, the highest percentage of 6:2 reassortants (31.4%) was achieved (
Figure 1, left panel).
Figure 1. Genome composition (%) of reassortants derived by classical reassortment of
ca MDVs with resistant or sensitive to nonspecific thermostable γ-inhibitors A(H1N1), A(H2N2), A(H3N2), A(H5N1), B/Victoria lineage, and B/Yamagata lineage WT influenza viruses
[14]). Left panel—328 reassortants of
ca MDVs with 20 WT viruses that are resistant to nonspecific thermostable γ-inhibitors were analyzed; 555 reassortants of
ca MDVs with 20 WT viruses that are sensitive to nonspecific thermostable γ-inhibitors were analyzed. 6:2 genome composition—HA and NA are inherited from the WT parent, and 6 internal genes are inherited from MDV; 5:3 genome composition—HA, NA and one of the internal genes are inherited from the WT parent, the other five internal genes are inherited from MDV; 7:1 genome composition—HA is inherited from WT parent, all internal genes and NA are inherited from MDV.
In contrast, the efficiency of obtaining 6:2 reassortants was much lower (7.2%) if the WT parent virus possessed a high degree of sensitivity to nonspecific thermostable γ-inhibitors; clones with the 7:1 genotype (25%) prevailed among the obtained reassortants. Corruption of the constellation of genes encoding HA and NA was observed—only a quarter of all reassortants inherited both WT HA and WT NA and three-quarters had WT HA + MDV NA
[14]) (
Figure 1, right panel).
Thus, the inhibitor sensitivity of WT viruses becomes an obstacle to the effective preparation of vaccine reassortants for LAIV by classical reassortment, since immune serum against MDV is involved in the selection process/screening of vaccine candidates. Contrarily, the inhibitor resistance guarantees a faster and more stable result in the preparation of vaccine strains. The development of LAIV based on the classical reassortment method would benefit from the recommendation of viruses with a high level of resistance to inhibitors. On the other hand, inhibitor-sensitive viruses retain a preference for α-2,6-linked residues. For now, the question, “what should be the best vaccine strain—inhibitor-sensitive or inhibitor-resistant?”, remains open.
4. Infectivity of WT Viruses and LAIV Candidates
One of the key indicators of the quality of reassortant candidates for IIV is their high HA titer. There has been up to a 512-fold increase in HA titers of PR8-based vaccine reassortant observed as compared to the respective WT parent virus
[21]. Reassortants that produce high HA titers do not always have a high yield of infectious viruses
[22] but infectious viral titers of reassortant candidates are not so critical for IIV. For example, reassortants prepared on a high-yielding PR8 donor, NIBRG-23 (H5N1) and VN/PR/CDC-RG (H5N1), displayed rather low infectious viral titers, which did not exceed 6.2 log
10 EID
50/mL and 7.7 log
10 EID
50/mL, correspondingly
[23][24].
On the contrary, infectivity is critical for LAIV. Whereas the WT parent virus typically has relatively low infectious titers (6.2–7.7 log
10 EID
50/mL
[24]), the titers of LAIV candidates on the backbone of
ca MDV are usually 8.7–10.2 log
10 EID
50/mL
[18][23][24].
As for the viruses to be recommended, typically, a reference strain and a few reference strain-like viruses that are similar in antigenic properties to the reference virus are recommended. Sometimes reference strain-like viruses appear to be less or more effective in the development of reassortant vaccine candidates than reference strains. For instance, based on our experience, the reassortant LAIV candidate of A/Leningrad/134/17/57 MDV with A/Brisbane/34/2018 (H3N2) WT parent (A/Kansas/14/2017-like virus recommended for use in 2019-2020 Northern Hemisphere influenza season) displayed ~ 1 log10 EID50/mL higher infectious activity than the reassortant candidate based on the A/Kansas/14/2017 (H3N2) reference strain, respectively. In contrast, the reassortant LAIV candidate of A/Leningrad/134/17/57 MDV with the A/Michigan/173/2020 (H3N2) WT parent (A/Darwin/9/2021-like virus recommended for use in 2022-2023 Northern Hemisphere influenza season) displayed ~ 1.0–1.5 lg10 EID50/mL lower infectious activity than the reassortant candidate based on the A/Darwin/9/2021 (H3N2) reference strain, respectively (I. Kiseleva, E. Bazhenova, E. Stepanova, N. Larionova and L. Rudenko. Personal communications).
Interestingly, the reassortment of
ca MDV with PR8-based vaccine strains for IIV of relatively low infection titers led to a dramatic increase in infectivity of the resulting reassortants
[23][24]. Unfortunately, there are cases when the presence of genes from an attenuated ca MDV does not significantly increase the infectious viral titers of reassortants. This has been observed in recent years for A(H3N2) influenza viruses and may be related to the receptor specificity of these viruses. If A(H1N1) and A(H1N1)pdm09 influenza viruses possess α-2,3 or α-2,3/α-2,6 specificity, due to which they multiply well in eggs without prior adaptation, then A(H3N2) influenza viruses retaining a preference for α-2,6 specificity
[12] have always been a problem for reproduction in eggs, becoming even more serious recently. Sometimes, national influenza centers that conduct year-round surveillance for influenza were not able to isolate A(H3N2) viruses in eggs to be recommended for seasonal vaccines in a timely manner.