Hook
Scientists engineered an antibody that completely blocks Epstein-Barr virus from entering human cells. It targets gp42, a protein on the viral surface. Mice that received the antibody before exposure showed zero infection.
EBV has infected 95% of adults worldwide. Most of us carry it for life. The immune system already sees it, already responds to it, already reached a truce.
Why did the body miss this pattern in the first place?
Immune Evasion
EBV infects B cells — the white blood cells responsible for making antibodies. That’s like a burglar hiding in the security office.
When the virus first enters the body, it replicates aggressively. B cells detect it, mount a response, produce antibodies. The infection looks like it’s under control.
Then EBV goes dormant. It inserts its genetic material into the B cell’s DNA and stops replicating. The cell doesn’t die. It doesn’t show obvious signs of infection. It just becomes part of the body’s long-term immune memory — with a virus inside it.
Natural antibodies are good at recognizing active virus: proteins on the surface of replicating particles, fragments released during cell destruction. But they’re not optimized to detect dormant virus hiding inside healthy-looking cells.
The immune system faces a pattern-recognition problem under constraint. It has to distinguish ‘threat’ from ‘self.’ EBV sits in the gray zone: it’s foreign genetic material, but it’s inside a cell the body needs. The immune system could destroy every infected B cell — but B cells are part of immune memory. Destroying them means losing the ability to respond to other pathogens.
So the body chooses a truce. It keeps EBV in check, kills cells when the virus tries to reactivate, and tolerates low-level infection. For most people, that truce holds for life.
For some — people with weakened immune systems, people who develop certain cancers — the truce breaks down. The virus reactivates. The body can’t suppress it. Disease follows.
The gap between ‘infected’ and ‘sick’ is the gap between pattern recognition that’s good enough and pattern recognition that’s complete.
Antibody Design
The Fred Hutch team engineered mice with human antibody genes. When exposed to EBV, these mice produced ten human antibodies: two targeting gp350, eight targeting gp42.
The gp350 antibodies provided partial protection. One gp42 antibody blocked infection completely.
Both proteins sit on the viral surface. Both help EBV attach to and enter B cells. But gp42 does something gp350 doesn’t: it’s required for the virus to fuse with the cell membrane. Block gp42, and the virus can’t get inside.
Natural antibodies target both proteins. But they don’t block gp42 effectively enough. The engineered antibody does. In mice given the gp42 antibody before exposure, zero viral DNA appeared in their blood afterward. In untreated mice, infection was universal.
This is remedial education for the immune system. The body already learned to recognize gp42. But it learned a weak version of the pattern — one that left gaps. The engineered antibody shows the immune system a stronger version: same target, better grip.
The immune system didn’t fail. It just optimized for a different problem.
Immune Memory
Evolution doesn’t optimize for perfect immunity. It optimizes for survival and reproduction.
An immune response that completely eliminated EBV would require destroying every infected B cell. B cells are expensive to make and slow to replace. They carry memory of every pathogen the body has encountered. Destroying them would leave the immune system temporarily blind to other threats.
The cost of perfect EBV clearance — losing immune memory to dozens of other pathogens — outweighs the benefit for most people. So natural selection favored immune systems that tolerate low-level EBV infection.
This is a trade-off built into every learning system. Complete coverage costs more than partial coverage. The body chose partial.
Antibody engineering isn’t constrained by the same trade-offs. A synthetic antibody can be more aggressive because it’s not competing with other immune functions for resources. It can target a viral protein more precisely because it doesn’t have to be produced by B cells that might themselves be infected.
The engineered antibody doesn’t replace natural immunity. It supplements it. Natural immunity keeps EBV dormant most of the time. The antibody steps in when dormancy breaks down — or when the body’s natural truce was never enough in the first place.
This reveals the shape of the gap between evolved systems and engineered ones. Evolution works with constraints: limited resources, competing priorities, genetic variation that can’t be too tightly controlled. Engineering works under different constraints: synthesis cost, delivery mechanisms, immune reactions to foreign proteins.
Natural immunity reached its limit not because it failed, but because it solved a different optimization problem than the one we wanted solved.
Close
The gp42 antibody doesn’t clear EBV from 95% of carriers. It prevents infection in people who haven’t been exposed yet. Your immune system learned the version that let you survive, not the version that made you safe.