Divergent B-cell repertoire remodelling by mRNA, DNA and live attenuated vaccines in fish

Abstract

Vaccination is critical for the future of aquaculture, and nucleic acid vaccines have a major potential for fighting emerging fish infectious diseases, yet their mechanisms remain poorly understood. We compared B-cell responses induced by an mRNA, a DNA, and an attenuated vaccine, all encoding the same antigen against a fish rhabdovirus. Rainbow trout IgHμ repertoires were examined to investigate how vaccines reshape clonal composition and complexity of the B-cell repertoire. The attenuated virus drove protection through a small number of highly shared public clonotypes encoding neutralizing antibodies. The mRNA vaccine profoundly remodelled the repertoire in some individuals and induces low, but still protective, neutralising Ab titers without public expansions. The DNA vaccine induced high neutralizing Ab titers, providing full protection with minimal impact on B-cell repertoire. Clustering analysis revealed partial sharing of private responses between fish. These findings highlight profound divergences between fish B-cell responses to nucleic acid and attenuated vaccines whilst all of three vaccines induce protective responses.

Fig. 1. Characteristics of the clonotype composition three months after vaccination with nucleic acid vaccines.

A IGHV and J gene usage for each vaccine type. B Frequency distribution of IgHμ clonotypes, categorised into rank ranges, is shown for each fish. The repertoire corresponding to the top 10 most frequent clonotypes is shown in black. C Measures of clonotype diversity between different vaccination types using the Shannon Diversity Index.

Fig. 2. Distinct private responses of the LNP (RNAgVHSV) vaccine.

A UpSet plot illustrating highly shared clonotypes, defined as those detected in four or five fish per group based on MID counts. The left panel shows the number of clonotypes meeting this shared criterion within each vaccination group. The right panel presents the number of clonotypes shared between groups (top) and the selected pairwise comparisons (bottom). B UpSet plot showing the top 50 most abundantly expressed clonotypes for each vaccination group. The left panel shows the number of clonotypes in the top clonotype lists within each vaccination group. The right panel presents the number of clonotypes shared between groups (top) and the selected pairwise comparisons (bottom). Both UpSet plots are based on data from a subsampling of 100,000 sequences per condition. C The top 25 clonotypes identified in each RNAgVHSV-vaccinated fish, highlighting the unique private immune responses elicited by the vaccine. The heat map colouring represents the MID count, with red indicating 800 or more reads and grey indicating the absence of a clonotype. D Probability of generation of the top 5 clonotypes from each RNA vaccinated fish compared to the top 5 clonotypes across all 5 fish in the DNA, PBS and attVHSV vaccinated groups (upper panel). The lower panel shows pgen distribution of clonotypes from PBS samples. * denotes significant difference from the PBS distribution where p < 0.01 by a one or two-sided Kolmogorov–Smirnov test.

Fig. 3. Public clonotypes against VHSV.

A Bubble plot displaying the public response to VHSV in each vaccination group. Bubble size corresponds to the average frequency of MID counts across fish in each vaccination group, while bubble colour indicates the number of fish expressing each clonotype. All 8 public clonotypes belong to the IGHV1-18*01 and IGHJ3 families with their respective CDR3 regions being shown. B Bar chart showing the average sum of MID counts relating to all 8 public clonotypes. ^a and ^b denote significant differences after one-way ANOVA where p < 0.05. Data are based on a subsampling of 100,000 MID counts per fish.

Fig. 4. Highly shared clonotypes within IGHV subgroups.

Barplots showing cumulative expression of the top 50 clonotypes shared by n individuals (i.e., present in 1, in 2, in 3 or in 4 fish), for a given IGHV subgroup, within each vaccinated group after three months post initial vaccination. Individual barplots show total expression and sharing of the TCL clonotypes for that condition compared to other conditions with PBS (in black), AttVHSV (in orange), DNAgVHSV (in blue) and RNAgVHSV (in green). X axis (1, 2, 3 and 4) demonstrates the level of sharing within each condition. Y axis represents cumulative expression based on the average MID count of 10 subsamplings of 10,000 per individual. To improve visualization of sharing at the VH subgroup level, we excluded the fish with the lowest MID counts for certain VH subgroups in each vaccination group from this part of the analysis, this was also done at the VH gene level. See Fig. S3 for IGHV subgroups where subsampling was possible.

Fig. 5. Highly shared clonotypes at the IGHV gene level.

A Bar plots showing cumulative expression of the top 50 clonotypes shared by n individuals (i.e., present in 1, in 2, in 3 or in 4 fish), for a given IGHV gene, within each vaccinated group after three months post initial vaccination. Individual bar plots show total expression and sharing of the TCL clonotypes for that condition compared to other conditions with PBS (in black), AttVHSV (in orange), DNAgVHSV (in blue) and RNAgVHSV (in green). X axis (1, 2, 3 and 4) demonstrates the level of sharing within each condition. Cumulative expression is based on the average MID count of 10 subsamplings of 1000. B Dotplot showing the distribution of clonotype frequency at different levels of sharing (n = in 1, in 2, in 3 or in 4 fish) between the top 50 clonotypes for each vaccinated group with the average MID count also shown. The data are calculated based on the average MID count of 10 subsamplings of 1000 per individual. See Fig. S4 for all IGHV gene bar plots where subsampling was possible.

Fig. 6. Classes of clonotype frequencies.

A Schematic representation of the top expressed private clonotypes (empty circles), and public VH1-18-JH3 clonotypes (grey circles). Circle size indicates the frequency class of a given clonotype, with the number next to each circle representing the number of clonotypes within that class. B Bar plot showing the distribution of clonotype frequency classes (both public and private). Frequencies are based on MID counts, summing both public and other responses. Clonotype classification is derived from nucleotide sequence data. Frequency classes are arbitrarily defined for visualization purposes. Data are based on a subsampling of 20,000 MID counts per fish.

Fig. 7. Clonal clustering analysis.

A Bar chart showing average number of clusters found per vaccine condition. B Venn diagram showing the top expressed clusters (>2000 MID sum in at least one fish per vaccination group). C Bar chart showing average frequency of clonotypes within each cluster of interest per vaccination group. D Dot plot showing the average per group of “cluster pgen”. We define the cluster pgen as the sum of pgen associated to all clonotypes detected within each individual for a cluster of interest.

Fig. 8. Schematic diagram of cluster 9352 (IGHV1-18*01 F_IGHJ3,5D) relating to the public response.

Outer Venn’s depict sharing of clonotypes found in cluster 9352 between fish within groups (subsampling of 20,000 MIDs), whilst the size of the circle represents the sum of expression of clonotypes within cluster 9352 for each fish (subsampling of 100,000 MIDs). Amino acid logos are given for each group. The central Venn depicts sharing of clonotypes with cluster 9352 between groups.

Fig. 9. Schematic diagram of cluster 75229 (IGHV6-4*02 F_IGHJ3,5D).

Outer Venn’s depict sharing of clonotypes found in cluster 75229 between fish within groups (subsampling of 20,000 MIDs), whilst the size of the circle represents the sum of expression of clonotypes within cluster 75229 for each fish (subsampling of 100,000 MIDs). Amino acid logos are given for each group. The central Venn depicts sharing of clonotypes with cluster 75229 between groups.