Bird Flu, Part 2

EDIT: Nothing of what follows should be interpreted as anti-vaccine. I am pro-vaccine. I receive annual or biannual flu and covid shots. The politicization of  science is a tragedy beyond measure.

We continue from Bird Flu, Part 1.  (CDC) H5 Bird Flu: Current Situation displays

Dr. Deborah Birx thinks this is an underestimation of the threat: (CNN) Why America could fail again: Dr. Birx’s stark pandemic warning. There are two sides to this. One is the tendency of bureaucracies to prognosticate continuation of the status quo. The other is faith in vigilance –widespread testing — and effective intervention. Both are logical fallacies. Pure logic rarely helps us navigate messy real-life situations. This is one of those rare opportunities.

We assume, with good reason, that the current H5N1 bird flu will acquire the mutations required for efficient human-to-human transfer. Prevailing opinion is that this will occur in a single individual, a real Patient Zero, who is simultaneously infected by bird flu and a human-adapted strain. The details remain unsettled.

With many  viruses, evolution is dominated by the process of homologous recombination — mixing of the genomes of both viruses during replication in a cell.  Influenza viruses are negative-sense replicators, in which this is thought to be absent or rare. For pro/con, see Homologous recombination evidence in human and swine influenza A viruses  and Homologous Recombination Is Very Rare or Absent in Human Influenza A Virus 

The dominant process with flu is swapping entire segments of RNA,  known as reassortment.  Since flu viruses lack proof-reading after replication, the above is supplemented by a high frequency of random, single point mutations.

So how does Patient Zero arise? Does a rare event of homologous recombination transfer a human-adapted mutation  into the avian virus? Or does it occur, per prevailing opinion, by spontaneous point mutation in Patient Zero, with subsequent reassortment ? Does it occur in the chicken or the human? The deepest, most arduous  and thorough study is based on a single human case, a veterinarian who died in the Netherlands  in 2003. See Emergence of the Virulence-Associated PB2 E627K Substitution in a Fatal Human Case of Highly Pathogenic Avian Influenza Virus A(H7N7) Infection as Determined by Illumina Ultra-Deep Sequencing. The conclusion, with good confidence, is that it happened in the human. Human adaptation markers, PB2 E627K and HA K416R, were found in the patient, but not the chickens. Quoting,

Human adaptation markers including PB2 E627K as well as HA K416R substitutions were absent in the A(H7N7) viruses obtained from both the source and control farms.

and

Herfst et al. demonstrated that PB2 E627K was a prerequisite for the development of airborne A(H5N1) virus in ferrets (7). The accumulation of other human adaptation markers than PB2 E627K observed in avian influenza viruses from poultry and the wild bird population suggests that the introduction of this particular mutation in avian influenza viruses dramatically increases the virus pandemic potential and public health risk.

Quoting (CDC, 12/26/2024) Genetic Sequences of Highly Pathogenic Avian Influenza A(H5N1) Viruses Identified in a Person in Louisiana,

The genetic sequences of the A(H5N1) viruses from the patient in Louisiana did not have the PB2 E627K change or other changes in polymerase genes associated with adaptation to mammals and no evidence of low frequency changes at critical positions.

As with the 2003 Netherlands case,

Of note, virus sequences from poultry sampled on the patient’s property were nearly identical to the virus sequences from the patient but did not have the mixed nucleotides identified in the patient’s clinical sample, strongly suggesting that the changes emerged during infection as virus replicated in the patient.

The is the primary basis of  the CDC assertion of low risk. It also explains the phlegmatic attitude towards testing. Testing animals is not likely to find a mutation that occurs first in Patient Zero. And it’s very hard to do:

Initial attempts to sequence the virus from the patient’s clinical respiratory specimens using standard RNA extraction and multisegment-RTPCR (M-RTPCR)1 techniques yielded only partial genomic data and virus isolation was not successful…

This is not something you can pick up with a  home kit. The elaborate techniques of CDC cannot be widely applied: we could have 50,000 bird flu test-strip positives, all unadapted  for human transmission. When point mutation, or debatable recombination result in a suite that includes PB2 E627K and other virulences, Nature will outflank us, if not with the  first Patient Zero,  then by successors.

The result will be a cluster of human cases. According to proponents of vigilance,  CDC could contain the cluster, stopping the outbreak.

This neglects the root cause of flu pandemics:   a world population, antigenically naive to a new strain of flu, that without forewarning, manifests in a series of Patient Zeros, a series without limit, a game of Whac-A-Mole. Sooner or later, vigilance would fail, or be circumvented by Patient Zeros elsewhere  in the world. The event, failure of containment, will be described by CDC and others as a rapid, almost discontinuous rise in threat risk.

Suppose vigilance identifies and isolates the first Patient Zero and cluster. Is there a response that could prevent the series without limit? One option would be to produce a vaccine that would be effective against future Patient Zeros of H5N1 recombination. But this is an expensive proposition for a vaccine that might not work. Flu vaccines are  generally lousy. In a bad year, with a strain mismatch, effectiveness against a strain that is already circulating can approach zero.  What are the chances of a vaccine against a virus that has not yet developed in a future Patient Zero?

The public has been entertained, almost romantically, by sequencing of numerous genomes, the “genetic code of life.”   In organisms with stable genomes, this knowledge accretes. Exact knowledge of the genome has been successfully exploited  with human genetic diseases. In single strand reverse-sense RNA viruses such as influenza, the genome is constantly changing, by several processes:

  • Spontaneous mutations at a single point, common in the flu virus  because it has no error checking mechanism on duplicated strands. The slow change in the genome resulting from point mutation is referred to as antigenic drift.
  • Reassortment, responsible for the run-of-the-mill flu debacle.
  • A remainder, to be classified as mystery,  genetic recombination between disparate genomes, by mechanisms that have never been seen,  except by the result.

In contrast with higher organisms with relatively stable genomes, the main focus of flu virus research is mutation. Yet the rate of mutation is elusive, stymied by the shape of this strand versus that, the tendency of mutations to be nonviable,  the lack of sufficiently detailed dynamical models, and a lack of clever solutions. Perhaps neural networks will help us out of this jam. Until then a  napkin calculation, “What is the chance of bird flu outbreak this winter?” cannot have a  bottom up approach.

There are other ways. Primitive, but worth a napkin.