We expect clade 3c2.a viruses to further predominate over 3c3.a and 3c3.b viruses. We do not observe the appearance of additional fast growing clades indicative of the emergence of new antigenic variants.
There were 4 clades of H3N2 viruses circulating in 2015 (3c2.a, 3c3.a, 3c3.b and ancestral 3c3). 3c2.a viruses have predominated with 78% of 2015 viruses belonging to clade 3c2.a, while 12% of viruses belonged to clade 3c3.a and 6% of viruses belonged to clade 3c3.b. Clades 3c2.a, 3c3.a and 3c3.b can easily be distinguished by the residue at HA1 site 159. In 3c2.a this is tyrosine (Y), in 3c3.a this is serine (S) and in 3c3.b this is phenylalanine (F).
3c2.a has been increasing in frequency, going from 50% globally in October 2014 to 80% of post-July viruses. 3c3.a has been steadily decreasing in frequency, going from a peak of 54% in August 2014 to a present low of 10% of post-July viruses. The first 3c3.b viruses appear in the database in August 2014. They have been slow to take off; just 2% of post-July viruses belong to clade 3c3.b. In fact, 3c3.b viruses appear to have peaked in frequency in May/June at 8%. For perspective, 3c2.a viruses first appear in the database in November 2013 and reach 10% frequency two months later in January 2014 and 50% eleven months later. 3c3.b viruses show no where near this level of success. Barring additional antigenic changes, it appears that 3c2.a viruses will take over the virus population and that 3c3.a and 3c3.b will soon be extinct.
These patterns of clade growth and decline extend to the regional level. 3c2.a viruses now dominate throughout the world. 3c3.b viruses briefly increased in frequency in Europe, but have declined in frequency in recent samples.
The continued dominance of 3c2.a is corroborated with a large local branching index and comparatively low mutational load as measured by the number of non-epitope substitutions since the root of the tree. Within 3c2.a, one clade (the uppermost clade with in 3c2.a) is starting to dominate but this clade is not characterized by any amino acid differences within HA. We observe no rapidly expanding subclades within 3c2.a or 3c3.b.
Summarizing data from recent reports by the WHO CC London
, we observe approximately a 2.8-fold drop in titer between serum raised against vaccine strain A/Switzerland/9715293/2013 and 3c2.a viruses. This corresponds to a 1.5 antigenic unit distance between clades. This can be seen on the left in the raw titer data (viruses are colored based on titer drop relative to A/Switzerland/9715293/2013 serum) and also on the right in a statistical model of titer drop. The statistical model finds that 3c2.a and 3c3.a both appear homogeneous in terms of titer measurements within the clade, further supporting the absence of new adaptive variants. Note however, that this analysis is averaging across the serum NIBF13/14 raised against cell-culture virus and the serum F25/14 raised against egg-culture virus.
Here we compare titer measurements between cell-culture based serum NIBF13/14 on the left and the egg-culture based serum F25/14 on the right. Averaged across measurements, titer drop between cell-culture serum and egg-culture serum is similar with an average antigenic distance of 1.5 for cell-culture measurements and an average antigenic distance of 1.5 for egg-culture measurements. Note however, that absolute titers against F25/14 are higher than against NIBF13/14. This suggests higher "serum potency" in the egg-culture serum, although we do not observe a difference in titer drop. This suggests that egg-adaptation during production of serum F25/14 did not result in antigenic change relative to 3c2.a viruses.
Thus, 3c2.a and 3c3.a viruses, although not antigenically identical appear quite similar. Observed rapid displacement of 3c3.a viruses by 3c2.a viruses makes sense in this context, i.e. we expect strong immunological competition between clades.