#115–David Watkins, PhD: Immunology, monoclonal antibodies, & vaccine strategies for COVID-19

Attia describes this as a 'stark wake-up call' analogous to measuring LDL cholesterol in isolation without knowing apoB: the test tells you a molecule is present, not whether it is doing the dangerous (or protective) job. The practical implication for patients is that a positive antibody test is not a green light to relax precautions, particularly for high-risk individuals.

Watkins uses this cascade throughout the episode as a teaching device: interferons from the innate response signal the adaptive system to activate; B cells in germinal centers go through affinity maturation with help from CD4 T cells; CD8 cells simultaneously kill virus factories; the plasma cells that settle in bone marrow represent long-term humoral memory. Understanding which arm of this cascade a pathogen attacks — HIV targets CD4 cells, the general of the army — immediately predicts why a given vaccine strategy will fail.

Watkins explains MHC diversity using the skin graft analogy: you could not receive a skin graft from him because his T cells would recognize your MHC molecules as foreign and reject the graft — MHC is the most polymorphic locus in the human genome. In the HIV context this means that the viral escape mutant that outruns the CD8 response in patient A may be the dominant peptide presented in patient B. Designing a universal T cell vaccine requires identifying conserved epitopes presented across the most common global HLA types that the virus cannot mutate without losing fitness.

Watkins notes: 'You can do a prime boost with a different vaccine to boost your immune responses.' He cites the yellow fever vaccine study as the gold-standard comparator: 17D generates near-universal 1:5,000 neutralizing titers because it replicates transiently and provides prolonged antigen exposure — the closest thing to a real infection that any vaccine has achieved. The mRNA platforms for COVID cannot replicate, so antigen exposure is brief; a boost dose re-primes the germinal center reaction and has been observed to increase both titer and breadth of neutralization.

Watkins frames this as a layered strategy borrowed from the HIV playbook: we need drugs, social distancing, and vaccines. Monoclonal antibodies occupy the 'drug for prevention' niche, analogous to PrEP (Truvada) for HIV. He notes the existing Humira example: tens of millions of people already receive monoclonal antibody injections chronically — the distribution infrastructure is proven.

Watkins: 'We have to be a bit smarter now — this is not the first SARS virus we've seen in the last 20 years so we need to try to anticipate the next one.' The HIV bNAb field pioneered the cohort-based isolation strategy: tens of thousands of patients followed for a decade, approximately 1% of whom develop broad neutralization. Those rare individuals' antibodies now sit in GMP cell banks. The same strategy, applied to endemic OC43 and 229E coronavirus cohorts, might yield pan-coronavirus bNAbs years before the next pandemic requires them.

Watkins is explicit about the three unknowns still applying to any COVID vaccine: we do not know what levels of neutralizing antibody responses will be sufficient to prevent infection, the frequency of people who will develop them, or the duration they will last. He frames duration as especially critical because if antibody levels decay over months, re-vaccination frequency becomes a population-level logistics challenge.

Watkins explains the distinction: neutralizing antibodies bind the specific region of the spike protein that docks to the ACE2 receptor on human cells, physically blocking infection. Non-neutralizing antibodies bind elsewhere on the spike and have no proven protective effect. The neutralization assay requires incubating a patient's serum with live virus over cell cultures for several days — it is technically demanding and cannot be done at point-of-care scale. Watkins notes the data also showed one or two 'super-responders' whose serum neutralized the virus at dilutions of 1:5,000 — the same individuals who would become the donors for a monoclonal antibody program.

Watkins describes the architecture: B cells enter germinal centers in lymph nodes where T helper cells provide the survival signal for cells showing improved binding. Cells with tighter-binding variants win, proliferate, and continue mutating. After 7–14 days the winners exit as either circulating memory B cells or large plasma cells that home to bone marrow, where they become permanent antibody factories. A bone marrow plasma cell can secrete neutralizing antibodies for years — this is why the measles vaccine confers lifelong protection from a single course. The HIV exception is that the virus mutates its envelope faster than the B cell can affinity-mature against it, leading to a perpetual arms race that in rare individuals takes 5–10 years to resolve into broadly neutralizing antibodies.

Watkins was at Harvard Primate Center when his lab first observed the two-week escape phenomenon. A student brought him sequencing data showing the infecting virus was gone and a new virus had appeared with a single amino-acid change in the CD8 epitope. Watkins initially told the student there must be a sequencing error. The broader implication: every infected HIV individual is not fighting 'HIV' — they are fighting a swarm of distinct quasi-species, many of which have already escaped whatever immune response was generated against the founder virus. This is the core reason a T cell vaccine targeting any single epitope will always fail.

Watkins explains the two compounding problems: first, the envelope trimer is present at very low density on the virion surface compared to, say, the influenza hemagglutinin. Second, the variable loops of the envelope are covered in host-derived sugars — a form of molecular camouflage that makes the epitopes look 'self' to the immune system. The few conserved neutralization-sensitive sites are recessed in a narrow canyon accessible only to a very long antibody loop (the CDR H3) that most human antibodies simply do not possess. The rare individuals who eventually make broadly neutralizing antibodies have undergone years of co-evolution between virus and immune system that is not replicable by any current vaccine strategy.

Watkins describes how the bNAb isolation process works: cohorts of thousands of HIV patients are followed longitudinally; their serum is tested against panels of diverse HIV isolates. Individuals whose serum neutralizes 50+ strains are identified as 'elite neutralizers.' Their memory B cells are then sorted, their antibody genes cloned, and candidate antibodies expressed recombinantly and screened. The lead candidates enter animal models and eventually human safety and prevention trials. Watkins's own lab is applying this same workflow to both the current SARS-CoV-2 outbreak and to prepare for future novel viruses.

Watkins describes the Ebola precedent: a monoclonal antibody injection during the 2019 DRC Ebola outbreak dropped case fatality from ~50% to ~15%. He argues this is the template for all future hemorrhagic fever and novel coronavirus outbreaks. The cost advantage of a cheap vaccine makes vaccination the population-level priority for coronaviruses, but monoclonal antibodies fill the gap for the elderly (who don't mount robust vaccine responses), for healthcare workers needing immediate protection before vaccine immunity matures, and as a treatment bridge. He notes the approach is 'a logical extension of a vaccine — we're simply taking from the best responders the best antibodies and distributing that to everybody.'