New paper on @biorxivpreprint:
"Landscape and Selection of Vaccine Epitopes in SARS-CoV-2"
(biorxiv.org/content/10.110…)
Collaboration with @ChristofCSmith1 @bgvincentlab @JuliaKodysh @timodonnell @glycam @Olivercgrant @PEPperCHIP @erikgarrison & others.
Thread:
(1/)
"Landscape and Selection of Vaccine Epitopes in SARS-CoV-2"
(biorxiv.org/content/10.110…)
Collaboration with @ChristofCSmith1 @bgvincentlab @JuliaKodysh @timodonnell @glycam @Olivercgrant @PEPperCHIP @erikgarrison & others.
Thread:
(1/)
There are a zoo of vaccine approaches to SARS-CoV-2:
Front-runners: inactivated virus (SinoVac), mRNA (BioNTech, Moderna), electroporated DNA (Inovio), adenovirus (CanSino, Oxford), rVSV (IAVI), rS (Novavax) &c
Missing from the list: peptides
2/
Front-runners: inactivated virus (SinoVac), mRNA (BioNTech, Moderna), electroporated DNA (Inovio), adenovirus (CanSino, Oxford), rVSV (IAVI), rS (Novavax) &c
Missing from the list: peptides
2/
Up until I joined @CompMedUNC / @UNC_Lineberger in March I worked on peptide vaccines targeting mutations in cancer (first in @hammer_lab and then @OpenVax)
Peptide vaccines aren't often used for infectious disease because they can only target linear B-cell epitopes.
3/
Peptide vaccines aren't often used for infectious disease because they can only target linear B-cell epitopes.
3/
However, peptide vaccines do at least provide control over which epitopes you elicit an immune responses against, which could be important if dominant approaches to vaccination result in antibody dependent enhancement (ADE).
nature.com/articles/s4157…
4/
nature.com/articles/s4157…
4/
Also, they're fairly effective effective at eliciting T-cell responses & T cells seem to play a large role in both SARS and SARS-CoV-2:
5/
5/
It's possible that vaccination with peptides, regardless of adjuvant, might be insufficiently immunogenic for protection. There are many potential tricks to overcome that, such as conjugation to a carrier protein: nature.com/articles/s4159…
6/
6/
But, more importantly, it seems like "just peptides" become a stronger immunogen when you co-locate multiple epitopes on the same peptide ()
Which brings me to the subject of our preprint...
7/
Which brings me to the subject of our preprint...
7/
Designing a set of vaccine peptides simultaneously targeting linear B-cell epitopes of SARS-CoV-2 along with both CD4+ and CD8+ T cell epitopes, with T cell epitopes bound to MHC alleles achieving high population coverage.
8/
8/
Summary:
T-cell: MHC binding, immunogenicity, exclude polymorphic sites, abundant proteins (M, N, S).
B-cell: use linear B-cell epitopes from patient serum, filter polymorphic sites, glycosites, inaccessible regions.
Vaccine: pick long peptides with multiple epitopes
9/
T-cell: MHC binding, immunogenicity, exclude polymorphic sites, abundant proteins (M, N, S).
B-cell: use linear B-cell epitopes from patient serum, filter polymorphic sites, glycosites, inaccessible regions.
Vaccine: pick long peptides with multiple epitopes
9/
Details:
To predict T cell epitopes we started by predicting MHC binding (using NetMHCpan 4.0-BA & NetMHCIIpan 3.2, but also evaluated other models)
Adjusted nM cutoffs to account for predictor accuracy on measured MHC affinities for SARS & viral peptides in IEDB
10/
To predict T cell epitopes we started by predicting MHC binding (using NetMHCpan 4.0-BA & NetMHCIIpan 3.2, but also evaluated other models)
Adjusted nM cutoffs to account for predictor accuracy on measured MHC affinities for SARS & viral peptides in IEDB
10/
Let's say that our MHC binding predictions are great -- those SARS-CoV-2 peptides still just MHC ligands and only a subset can become true T cell epitopes.
To further refine the set we made a model of p(immunogenic|MHC ligand) from IEDB's viral pMHC tetramer results
11/
To further refine the set we made a model of p(immunogenic|MHC ligand) from IEDB's viral pMHC tetramer results
11/
Aside: Tetramer (or dextramer, &c) data is one of the few kinds of "immunogenicity" sources in IEDB that's unlikely to be tainted by computational assignment of MHC allele (as opposed to ELISpot, which is dominated by unknown alleles being guessed using NetMHC)
12/
12/
CD4+/CD8+ immunogenicity models were logistic regressions with forward feature selection. Features were MHC affinity, mass spec presentation score, MHCflurry processing (for CD8+) and amino acid properties from journals.plos.org/ploscompbiol/a…
(e.g. smaller AAs = less likely)
13/
(e.g. smaller AAs = less likely)
13/
After dropping the bottom 50% of peptides by immunogenicity scores, we further filtered by removing polymorphic sites and keeping only the top 3 proteins (M, N, S) from semi-quantitative mass spec data.
14/
14/
In total we went from 3730 predicted MHC-I ligands and 5045 predicted MHC-II ligands to 292 predicted CD8+ T cell epitopes and 616 predicted CD4+ T cell epitopes.
15/
15/
On the B-cell side, we started with 5 data sources for linear epitopes of anti-SARS-CoV-2 antibodies.
One unpublished peptide array study from @PEPperCHIP (which helped us get started), 3 published peptide mapping studies, and one PhIPseq paper: medrxiv.org/content/10.110…
16/
One unpublished peptide array study from @PEPperCHIP (which helped us get started), 3 published peptide mapping studies, and one PhIPseq paper: medrxiv.org/content/10.110…
16/
Across 5 sources, we found 58 linear B-cell epitopes on S, some with significant overlap.
We filtered these epitope regions by accessibility (from ), removed glyco- and polymorphic sites & only kept 4+ AAs near functional regions (RBD, FP, HR1/2)
17/
We filtered these epitope regions by accessibility (from ), removed glyco- and polymorphic sites & only kept 4+ AAs near functional regions (RBD, FP, HR1/2)
17/
The three targetable linear B-cell epitope regions are:
S456-473 (RBM loop)
S580-583 (near RBD)
S809-812 (near FP)
The last two were the targets of neutralizing antibodies in biorxiv.org/content/10.110…
& the last B-cell epitope region occurred in all 5 sources!
18/
S456-473 (RBM loop)
S580-583 (near RBD)
S809-812 (near FP)
The last two were the targets of neutralizing antibodies in biorxiv.org/content/10.110…
& the last B-cell epitope region occurred in all 5 sources!
18/
To combine the B-cell and T-cell epitope predictions, we created many small peptide sets to achieve high population coverage of Class I / Class II alleles and/or the B-cell epitopes. Since we want to use these peptides experimentally, we also included +/- H2b & H2d ligands.
19/
19/
Many of "minimal" vaccine peptide sets were redundant (contained some of the same peptides), so we collapsed them down to a table of 22 unique peptide sequences.
These are each annotated with whether they contain H2b/d murine MHC ligands, to suit your experiments.
20/
These are each annotated with whether they contain H2b/d murine MHC ligands, to suit your experiments.
20/
This computational work has been going on in parallel with mouse vaccine experiments being done by both the @bgvincentlab and @faeznm. From a smaller set of preliminary peptides we have already seen some immunogenicity. Now we're waiting for the 22 new peptides...
21/
21/
We also have recombinant RBD from @MarkBlenner, rTTHC to use a carrier protein from Peter Kwong's group at the VRC, and other reagents/ingredients coming in from different groups.
I'm really excited for the next round of experiments...
22/
I'm really excited for the next round of experiments...
22/
Which peptides and/or adjuvants will work? Is a carrier protein necessary? Will the T cell responses be enough to suppress SARS-CoV-2? Will our antibody responses be neutralizing and/or protective?
Tune in for the next paper...
23.
Tune in for the next paper...
23.
Extra thought:
We started with measured B-cell epitopes (albeit of unknown function) but the T-cell epitopes were purely computational predictions. Since we did these analyses, 2 preprints came out which identify T-cell epitopes in recovered SARS-CoV-2 patients...
24/
We started with measured B-cell epitopes (albeit of unknown function) but the T-cell epitopes were purely computational predictions. Since we did these analyses, 2 preprints came out which identify T-cell epitopes in recovered SARS-CoV-2 patients...
24/
"Different pattern of pre-existing SARS-COV-2 specific T cell immunity in SARS-recovered and uninfected individuals" (biorxiv.org/content/10.110…) found 2 recurrent epitope regions (N101-120, N321-340), both of which overlap our vaccine peptides (Figure 5B, peptides 4-8)
25/
25/
"SARS-CoV-2 epitopes are recognized by a public and diverse repertoire of human T-cell receptors"
(medrxiv.org/content/10.110…)
also found 2 recurrent epitopes (S269-277, S1000-1008). These are also contained in our vaccine peptide set
(5B, peptides 11 & 15), So far so good!
26.
(medrxiv.org/content/10.110…)
also found 2 recurrent epitopes (S269-277, S1000-1008). These are also contained in our vaccine peptide set
(5B, peptides 11 & 15), So far so good!
26.
