Sugar on the cell surface isn't frosting — it's the brain's moat
The mucins on the brain's vascular endothelium are like a redwood forest; with age the leaves fall off and the blood-brain barrier gets leaky. Restore the enzyme with gene therapy and mice remember better.
The video won't play here. Listen to the audio instead:
The argument · timestamps estimated from transcript position
Glycobiology was abandoned by its tools
In the 1960s and 70s, carbohydrates were taken as seriously as proteins, but once recombinant DNA arrived, proteins could be cloned, expressed and studied — sugars could not. Sugars are not a product of the central dogma; they are the output of enzymes encoded by hundreds of genes, and molecular biology techniques cannot manipulate them directly. So the field rushed toward whatever could be accelerated, and sugar science, technically hard, was left where it stood. Bertozzi says this is not because sugars are unimportant, but because "just because you don't study an area doesn't mean it's not important."
— Carolyn BertozziShe built the flashlight before talking biology
When she started her lab at Berkeley in 1996, sugar science was missing two things proteins had long had: sequencing (knowing which sugars exist) and live imaging (seeing where sugars are and how they move). Her lab's initial goal was to fill exactly those two holes, and it took roughly a decade to develop bioorthogonal chemistry — a chemical reaction that can label sugars inside living cells and is now a general-purpose tool in the life sciences. With a flashlight, you can finally see places no one had looked before.
— Carolyn BertozziNobody had asked how sugars change with age
Tony Wyss-Coray's lab was studying transcriptomic and proteomic changes in brain endothelial cells as they age — standard practice — but nobody had asked about the glycome. Sophia, a student who had rotated in Bertozzi's lab, moved to Tony's lab and brought the sugar imaging reagents with her to compare the vasculature of old and young mice. The result was "absolutely striking" — the sugar changes were far more dramatic than any phenotype observed in the transcriptome or proteome.
— Carolyn BertozziBrain endothelial cells have mucins like a redwood forest
Mucins on the cell surface are like giant redwoods: tall, stiff, straight, projecting above everything else. The mucin forest on brain endothelial cells is far thicker than on the vasculature of heart, kidney or muscle — a landscape unique to the blood-brain barrier. With age this forest turns into winter: leaves gone, trees bare, you can see straight through. Young mice are spring, dense with foliage. Sophia found that the enzymes responsible for synthesizing these sugars drop in abundance in old mice; restoring the enzyme with a gene therapy vector restored the mucins, reduced blood-brain barrier leakage, and the mice got through the maze faster.
— Carolyn BertozziA reagent crossed the blood-brain barrier on its own
Sophia injected a mucin-binding reagent into mice and found hours later that it first bound the vasculature, then crossed the blood-brain barrier by itself into the brain parenchyma. Textbooks say proteins cannot cross the blood-brain barrier alone, and this is the biggest pain point for biologics companies working on brain drugs. The accidental finding made the team realize it could serve as a shuttle carrier — a brain shuttle — delivering cargo into the brain. Sophia now has her own lab at Harvard, and she, Bertozzi, Tony and another student are spinning out a company to do brain shuttling.
— Carolyn BertozziOnly 15 CNS biologics are approved
More than 150 biologics have been approved to treat immune disease, and a similar number for cancer, but only 15 treat central nervous system disease — and most go in by intrathecal injection: a lumbar puncture, invasive, a whole day gone, hard to scale. If those drugs could be given as ordinary intravenous injections, it would be transformative for patients. The drug classes that could get into the brain include monoclonal antibodies, enzyme replacement therapies, siRNA and antisense oligonucleotides, mRNA lipid nanoparticles, and viral and non-viral gene therapy vectors.
— Carolyn BertozziCD22 is the brake on the brain's immune cells
Tony's lab found that in microglia of aged and Alzheimer's disease mice, the sugar-binding receptor CD22 (Siglec-2) is upregulated. Bertozzi's lab had studied this receptor family in cancer: when they are upregulated they inhibit immune cells, letting cancer cells grow unchecked. Microglia are the brain's scavengers, responsible for eating protein aggregates, and their dysfunction is linked to neurodegenerative disease. Blocking CD22 with an antibody made aged mice's microglia more active and their brains cleaner. But at the time there was no brain shuttle, so it could only be injected intracranially — no path to a human therapy; now that the shuttle exists, they can go back and revisit that antibody.
— Carolyn BertozziToday's drugs rest on chemistry from a century ago
Bioorthogonal chemistry was developed in the early 2000s and is now part of the toolkit for chemists and biologists alike, with approved drugs that use the chemistry. But looking back for her Nobel lecture, Bertozzi found that the foundational chemistry they used to develop these reactions was published in 1915, in 1919, and in 1950 from Germany — all old German-language journals that needed translating. Those chemists could never have foreseen that someone a hundred years later would use it to make cancer drugs. She says most scientists will not know the true impact of their work in their lifetime, and we have to invest in basic science without knowing the impact or the timeline — otherwise the whole pipeline eventually runs dry.
— Carolyn BertozziIn their own words · checked verbatim
but not working on an area doesn't make it unimportant
Carolyn Bertozzi10:06
When a flashlight is on, that's where you look, but you don't see all the dark spaces. You need to have a different flashlight.
Carolyn Bertozzi28:21
There's a big distance between a molecular change and a behavior and understanding everything in between is kind of the ongoing challenge
Carolyn Bertozzi32:00
the textbooks would tell you that proteins on their own don't just cross the blood-brain barrier
Carolyn Bertozzi33:53
CD22 is a break type of receptor. It's inhibitory when it gets engaged with its ligands.
Carolyn Bertozzi42:27
Most scientists right now will not know the true impact of their work in their own lifetime.
Carolyn Bertozzi50:27
you have to invest in that science without knowing what its impact will be and on what timeline. You just have to make the investment. Otherwise, the whole pipeline will eventually dry up.
Carolyn Bertozzi50:27
Figures
| Total length of blood vessels in the brain | 400 miles | 23:21 |
| Approved biologics for immune disease | more than 150 | 36:43 |
| Approved biologics for central nervous system disease | 15 | 36:43 |
| When bioorthogonal chemistry was developed | about 25 years ago, in the early 2000s | 47:13 |
| Publication years of the foundational chemistry reactions behind bioorthogonal chemistry | 1915, 1919, 1950 | 48:01 |
Glossary
- bioorthogonal chemistry
- A chemical reaction that can be carried out in living cells without interfering with native biological processes, used to label and image molecules such as sugars.
- glycome
- The complete set of glycans on a cell surface, the counterpart to the concepts of genome and proteome.
- mucin
- A heavily glycosylated cell-surface protein, bottlebrush-shaped, that forms a thick layer on brain endothelial cells.
- CD22 / Siglec-2
- A sialic-acid-binding inhibitory receptor that suppresses immune cells in cancer and is upregulated in microglia with aging.
- brain shuttle
- A molecular carrier that can deliver biologic cargo across the blood-brain barrier into the brain.
How to listen
Founders and investors working on neurodegenerative disease, blood-brain barrier delivery, glycobiology or biologics; researchers interested in how basic science turns into a company.
The first 10 minutes cover the history of glycobiology and the M&M analogy — fast-forward to the aging glycome finding at 26:00.