Phage bet-hedging via 10,000x reversion, and the first humanized EBV antibodies
Phages use the 10,000-fold reversion rate of SSRs to hedge their bets, humanized EBV antibodies stop tumors in vivo, and vaccine work advances on several fronts.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Phages deliberately let a fraction of their offspring come out broken
Jolene walks through a Nature Microbiology paper: homonucleotide repeats (SSRs) in phage genes slip easily during replication, producing insertions or deletions that shift the reading frame. That means some fraction of the progeny can never make the protein properly. This randomness is exactly what bet-hedging rests on: most sequences are adapted to the current environment, but a persistent low-level pool of variants can become the survivors the moment conditions change sharply. The population trades a small cost for long-term survivability.
— Jolene RamseyNo need to adapt in advance: reversion runs 10,000 times faster
The paper measured the reversion frequency in the SSR regions and found it roughly 10,000 times higher than ordinary point mutation. An order of magnitude like that means a phage can switch a gene's function on or off within very few generations, instead of waiting for a rare point mutation. Jolene stresses that this mechanism spares the population from having to pre-adapt to environmental change; it continuously generates diversity and waits to be selected.
— Jolene RamseyT7 carries no SSRs at all, so bet-hedging is not standard equipment
In the Basel phage collection, the T7 genome has no SSRs. Rich comments that T7 is ‘businesslike’ — straight to the point, without any of these fancy mechanisms. That contrast shows SSRs are not standard equipment across phages but an evolutionary strategy of certain lineages in particular ecological niches. Even lytic phages on the same host can organize their genomes in completely different ways.
— Rich ConditThe hard part of humanization is the constant region, not the variable region
When a mouse antibody is engineered into a humanized one, the variable and constant regions conflict, and someone has to spend years adjusting the constant region so it is compatible with the variable region before a therapeutic antibody can be made. Brienne says this was her aha moment; she had not understood why mice were necessary in the first place. Vincent adds that going straight to human antibodies may fail to turn up the antibody you want, whereas the advantage of the mouse platform is that you can screen large numbers of clones and engineer the antigen, getting the full process of B cell clonal diversification — after which you still need classical hybridoma technology to obtain monoclonal antibodies.
— Brienne, VincentAfter all the hybridoma screening, only 10 antibodies were left
The group immunized mice with recombinant GP350 and GP42, made hybridomas, plated 384-well plates, and ran high-throughput flow screening with antigen-coupled magnetic beads, testing both binding and the ability to inhibit infection. Positive clones were subcloned and their heavy- and light-chain genes sequenced, then expressed as recombinant human IgG1 antibodies. The final yield was 2 GP350-binding antibodies and 8 GP42-neutralizing antibodies. The whole workflow lays out the complete path from animal immunization to recombinant antibody.
— VincentThe antibodies also blocked EBV challenge in living animals
Two strains of immunodeficient mice were engrafted with human CD34+ hematopoietic stem cells, given 500 micrograms of antibody intraperitoneally (ATX42-2, ATX350-2, MO1 and others), then challenged intravenously with EBV and followed for 11 weeks. The antibodies prevented tumor formation, suppressed viral DNA replication and blocked splenomegaly. In mice treated with ATX42-2 and MO1, no EBER-positive cells were detectable in the spleen. That means human antibodies protect against EBV challenge in a living animal, not merely neutralize in vitro.
— VincentAlum-adjuvanted subunit vaccines are dead; nanoparticle vaccines are still alive
A GP350 subunit vaccine with alum adjuvant failed to generate protective immunity in trials, and GP42's conformation is too variable to reliably elicit the right antibodies. But a GP350 ferritin nanoparticle vaccine has already shown safety and immunogenicity in phase one, with seropositive participants remaining free of EBV infection out to 540 days. A GH/GP42 ferritin nanoparticle and Moderna's mRNA vaccine (containing four antigens) are also in development. Vincent expresses disappointment that the paper did not try combining the GP350 and GP42 antibodies.
— VincentAnti-vaxxers may not lack information so much as the willingness to think
Vincent again recommends “Thinking, Fast and Slow”, quoting Joshua Reynolds: ‘There is no expedient to which a man will not resort to avoid the real labor of thinking.’ He thinks cognitive economy may partly explain anti-science and anti-vaccine attitudes: deep thinking burns energy, and people gravitate to low-effort type one thinking. He immediately adds, ‘that's laziness, cognitive laziness.’ The argument traces the difficulty of science communication back to a default human cognitive mode rather than to a simple shortage of information.
— VincentIn their own words · checked verbatim
the frequencies were quite different. I think they said it was 10,000fold different reversion frequency.
Jolene Ramsey21:19
Oh, and what I was disappointed that they didn't try the combination of the two, 350 and 42.
Vincent1:08:08
it's the first time that we've got these kinds of antibodies, human antibodies, right?
Vincent1:12:13
There is no expedient to which a man will not resort to avoid the real labor of thinking.
Vincent1:43:45
Figures
| Difference in reversion frequency | 10,000-fold | 21:19 |
| New EBV-attributable cancer cases per year | 358,000 | 42:36 |
| EBV-attributable deaths per year | 209,000 | 42:36 |
| GP350-binding antibodies obtained | 2 | 55:47 |
| GP42-neutralizing antibodies obtained | 8 | 55:47 |
| Antibody injection dose | 500 micrograms | 1:05:01 |
| Weeks of observation | 11 weeks | 1:05:01 |
| Cassini launch year | 1997 | 1:38:41 |
| Time Cassini took to reach Saturn | 7 years | 1:38:41 |
| Year Cassini was crashed | 2017 | 1:38:41 |
| Number of Saturn's moons | 260 | 1:39:41 |
Glossary
- SSR / simple sequence repeat
- A repeated nucleotide sequence within a gene that slips easily during replication, causing insertions or deletions.
- bet-hedging
- An evolutionary strategy in which a population maintains multiple phenotypes at once to cope with environmental change.
- contingency locus
- A genomic region that generates variation across a population through a high mutation rate.
- hybridoma
- An immortalized cell line made by fusing a myeloma cell with a B cell, used to produce monoclonal antibodies.
How to listen
Researchers working on phages or antibodies, investors tracking EBV vaccine progress, and engineers who want to understand why science communication keeps failing.
The phage details in the first half hour are fairly basic; if you only care about antibody development, start at 50:44.