Tuesday, February 14, 2012

Approved at the last minute (literally) that I can go to Geneva and was granted an Export Control License for me, Barbara Jasny (representing Bruce Alberts of Science), Paul Keim (Chair, NSABB), and Yoshi Kawaoka (author of Nature article) Feb. 16-17, 2012 – Geneva – Astounding!! During Ron’s presentation of the data, I asked the “transforming” question that exposed the confusion that has been propagated regarding the study. It relates to the uncoupling of the issue of aerosol transmission from deadly pathogenesis. Bottom line is that the aerosol-transmitted virus when transmitted to uninfected ferret, not only does not kill them, it does not even cause significant disease. At worst, some mild flu like sxs. See below for my detailed description of the data gathered from the papers, the Geneva presentation together with discussions during the Geneva meeting including a dinner at Keiji’s home with Ron, AB Osterhaus, Keiji Fukuda, and Yoshi on the evening of Feb. 15 (more on that later) as well as extensive discussion by phone with Ron after I returned to USA on Feb. 19 and 20, 2012. Recommendation (consensus) of the WHO-convened group was to 1) respect the PIP agreement; 2) extend the voluntary pause (“moratorium”) on H5N1 transmissibility/pathogenicity research; 3) Delay publication of the 2 manuscripts until the safety and security issues of such research could be solidified for others. General agreement that the experiments of Yoshi and Ron were carried with safety and security that met or exceeded the guidelines; 4) After meeting conditions in #3, then publish the manuscripts in full; 5) discussion that the publishing of redacted manuscripts with mechanism for vetting and distributing to those with a need to know was completely impractical. This point was fortified by the almost unbelievable and counterintuitive requirements of Export Control Licenses. There was a unanimous recommendation for everything except the recommendation to publish. Nancy Cox did not comment, but Paul Keim and I stood by the recommendations of the NSABB. In the roll-out after the meeting in Geneva, Keiji at first messed up and said that everything was “unanimous consensus”, which is an oxymoron. Christ Feig is a former CNN producer and old friend of mine and is the communications director for WHO. I noticed the slip-up and she got Keiji to correct while still on the TV press conference. Emails were going back and forth from me to HHS (Bill Hall) to some of the NSS people when he made this mistake, but everyone calmed down after he corrected it. I was in the unusual position of officially sticking with the NSABB recommendation while being the one who actually led the WHO discussion to come to this point. The smart reporters like Jon Cohen and Helen Branswell picked this up and were sympathetic. The decision was written up on the front page of the NY Times by Denise Grady on Sunday. It was interesting how she quoted me with an articulate explanation of why the WHO group came to this decision at the same time that she mentioned that I stuck by the NSABB recommendations. Anyone with insight could figure out what I did. Of note, Paul Keim did not realize that the aerosol-transmitted H5N1 did not make the animals ill, much less kill them!!! There was a media storm after the decision was announced. See below for summary of meeting: Sunday, February 19, 2012 Summary of discussion at the WHO Geneva Meeting on Feb. 16-17, 2012 plus additional discussion with Ron Fouchier and Yoshi Kawaoka. Anthony S. Fauci, M.D. Fouchier Experiments (Science) and related information: Fouchier worked with a high pathogenicity (HP) H5N1 influenza virus obtained from a patient in Indonesia (A/Indonesia/5/05). He had been working for 4 to 5 years on trying to study the transmissibility of H5N1 for the purposes of determining the molecular aspects of increased transmissibility. His attempts at developing an H5N1 with increased transmissibility in a mammalian model have failed up to now. Recently, he used the already existing literature to identify the mutations that have been reported to be associated with the emergence of previous pandemics (1918 H1N1, 1957 H2N2, 1968 H3N2, 2009 H1N1 and other non-pandemic outbreaks, e.g. 1977 H1N1). He analyzed these and determined the 3 mutations that statistically seemed to be associated most closely with emergence of pathogenic pandemic viruses. Again, these mutations have already been reported in many published papers. With regard to the “technique” of passaging in ferrets, there are many papers that have appeared in the literature for decades (since the 1930s) describing the ferret passage technique to increase transmissibility. This is similar to passage techniques widely used to adapt a virus to a particular species. It has been done in mice, monkeys, etc. Usually (not always), when viruses are passaged in a particular species, as they gain transmissibility, i.e. species-specific adaptation, they develop decreased pathogenicity, i.e. attenuation. Fouchier took these 3 mutations and inserted them into a wild type (WT) HPH5N1 (see above). This technique is widely used in molecular virology. He then passaged the virus 10 times in ferrets. “Passaged” means inserting the virus into the nose of the animal, getting the animal to get infected, then taking virus from the nasal passages of the infected animal and inserting it into the nasal passage of the next animal in the sequence, and so on. The endpoint that is aimed at is to get a virus that is then able to be transmitted by aerosol (i.e. not relying on direct insertion into the nasal passage). Since ferrets sneeze when they are infected, one determines “aerosol transmission” by housing an infected ferret in a cage next to a separate cage of an uninfected ferret that is constructed such that an aerosol can get from one cage to the other due to “holes” in the barriers between cages. The virus resulting from the passaging was able to be transmitted by aerosol and was a “quasi-species”, i.e. it contained different “versions” of the mutated virus. The viruses that were transmitted by aerosol were then sequenced and all of these viruses had 5 mutations in common: the 3 mutations that were inserted by reverse genetics and 2 additional ones. In addition, the viruses had 4 to 7 variable mutations (this latter sequencing information was not ready for reporting in the original manuscript). This virus isolate was used in the pathogenesis studies reported in the manuscript. It was the isolate with the minimal number of mutations (9 mutations - 5 common mutations + 4 variable mutations). After submission of the manuscript the authors took a wild-type H5N1 clone and inserted this minimal number of mutations to develop a virus that was transmissible without the need for passage. The pathogenesis studies with this cloned virus were identical to those of the isolate reported in the manuscript (see below). It is EXTREMELY IMPORTANT to point out that the animals that got infected by aerosolized virus described above did not die nor did they get sick. They were followed for 14 days. A few got minor flu-like symptoms, all recovered, they then were sacrificed since logistically the author could not house them indefinitely. Thus, the “engineered” virus that was made transmissible, did not kill or make the animals sick when exposed in the manner that flu in transmitted in humans. PATHOGENICITY. The classic way that influenza virus pathogenicity is established (welldescribed in the literature) for the ferret model is by direct insertion of the virus into the nasal cavity and/or directly into the trachea. Fouchier did both nasal and tracheal insertion of the viruses. In the nasal insertion studies, 106 (1 million) infectious doses of virus is inserted into the nose. This is a dose that far exceeds a dose that one gets through aerosolized virus (i.e. sneezing). When this dose of the pandemic 2009 H1N1 virus (not considered a particularly pathogenic virus, particularly in people who have been exposed to H1N1 over years by infection or vaccination - children who are naïve to this virus have had more problems) is inserted into the ferret, the animals develop flu-like symptoms but do not die. When HPH5N1 is inserted the animals essentially all get neurovirulence, they recover, none die. When the modified (aerosoltransmitted) H5N1 virus is inserted, there is some neurovirulence, but actually less neurovirulence than with the wild typeH5N1. Again, all recover and none die. In the tracheal insertion studies, 106 virus is inserted directly into the trachea. Again, this is a dose and a situation that would be extremely difficult, if not impossible, to achieve by aerosolized transmission. In these experiments, when this amount of wild type (unmodified) HPH5N1 is inserted into the trachea, all animals die on day 3. Similarly, when the modified H5N1 is inserted, all animals die on day 3. However, and importantly, when 2009H1N1 that was shown to be a mild pathogenicity virus was inserted under the same conditions, it killed some, but not all the animals. Although, this was a qualitative and not a quantitative study, it strongly suggests that in this model, the modified H5N1 is not dramatically different from the 2009 H1N1 in pathogenicity. Also, it should be pointed out that in the ferret model, 2009 H1N1 is much more transmissible (100% transmission on day one) than is the modified H5N1. Specifically, transmission with the modified H5N1 does not occur until day 3 to 4 and even then it is only 75% transmissible. This suggests that the R0 may be less than 1. This is important with regard to concerns over the potential catastrophic effects of lab accident and the known conditions that are needed for the start of an epidemic/pandemic. In this regard, there should be serious discussion of how pandemics occur. Nancy Cox can elaborate. Ferrets are naïve to influenza virus, i.e. they have no pre-existing immunity. Humans do have pre-existing immunity to influenza viruses due to repeated exposures (except in young children) and/or vaccinations. It is EXTREMELY IMPORTANT to point out that if you vaccinate the ferrets with seasonal H1N1, they are not protected against death from tracheal insertion of high dose (106 infectious doses) modified or wild type H5N1. Fouchier did not test the effect on transmission yet (moratorium). However, if you pre-infect ferrets with seasonal H1N1, let them recover and then insert high dose H5N1 into their trachea, this fully protects against disease and no ferrets die. This strongly suggests that pre-existing immunity due to prior infection with seasonal influenza A protects against serious disease with this modified virus (heterosubtypic immunity). This could be a possible explanation for why so few people get infected with H5N1 in the wild. Yoshi Kawaoka Experiments (Nature). The studies of this investigator are very similar to those of Fouchier except that Kawaoka took the HA from the HPH5N1 and did random mutations from which he selected those that enhanced transmissibility when this HA was reassorted with the 2009 H1N1to make a hybrid of the H5N1 and the 2009 H1N1. This is something that we are concerned could happen in the wild. He then did similar passages as described above. Kawaoka’s results differed from the Fouchier study in that although his virus was transmissible, it was not any more pathogenic than the 2009 H1N1 influenza. EPIDEMIOLOGICAL DATA: With regard to the importance of these data for the real and present danger of the evolution of H5N1 in the wild, as well as for the impact on surveillance, Kawaoka and Fouchier have done some epidemiological “mapping” of their modified virus with what is going on in the wild. Viruses were studied that were obtained from chickens in Japan, Mongolia, Nepal, and Egypt as well as viruses from people infected with H5N1 in Egypt. Some of the mutations that Fouchier and Kawaoka described involving the PB2 gene (627 K) and the receptor binding genes (HA) were present in 100% of the H5N1 viruses that infected humans and only in 30% of the chicken viruses. This suggests that the viruses that accumulate the mutations in question are the ones that transmit from chickens to humans. In addition, many of the viruses circulating in chickens are only 2 to 3 mutations away from the mutations described by Fouchier in their modified viruses. People will argue that knowing mutations cannot help predict or alert public health officials to what will emerge as a pandemic. They state that since viruses have several routes to get to where they want to go, i.e. to be able to be transmitted, the studies in question should not even have been performed. However, historically, after pandemic viruses emerge and infect people and then you go back and trace what was evolving in birds, or (more importantly) in mammals you see a pattern where you could have actually traced and connected the dots pointing to the emergence of a pandemic. This was clearly the case where the pandemic 2009 H1N1 virus was evolving in pigs for at least 10 years and was “working its way” towards human transmissibility. If there had been better surveillance in both pigs and humans as to what was evolving, then as Tom Frieden has said, we could have had a major head-start on the 2009 H1N1 virus. What the Fouchier and Kawaoka viruses tell us is that the virus in the wild is creeping towards a virus that they have shown is capable of transmission in mammals, presumably including humans. Whether the virus ultimately uses that precise pathway or those precise mutations is not clear, but clearly these studies are critical to understand the process of evolution and adaptability of viruses that are a real threat of working their way to transmissibility in humans. Not making these data freely available to scientists only slows down the necessary work and the incentive to get more scientists involved, and does not really provide a serious terrorist with any additional advantage since these techniques have been widely available for years. An uninformed terrorist would not be able to do this. An informed virologist-terrorist just needs to go to PubMed to easily figure out what was done and how it was done. Not publishing the work may do more harm than good by discouraging legitimate scientists from getting involved. Clearly, there needs to be assurance of biosafety and biosecurity in the laboratories that work with transmissibility of H5N1. The laboratories in question (Fouchier and Kawaoka) have met or exceeded our standards for these types of studies. Redacted versus Unredacted: The impracticality of redacted versus unredacted manuscripts was clear to everyone at the Geneva meeting and the likelihood of the unredacted manuscripts ultimately reaching the internet were important factors in the consensus at WHO to publish in full after a pause to assure biosafety and biosecurity and to have a better articulation of the benefits of the research as described above. The way forward. People are making incendiary statements without really having discussed with the investigators the precise details of the experiments as described above. The interaction of the NSABB with the investigators (Fouchier and Kawaoka) was on a relatively brief conference call. Much additional information that was available during the intensive 2 day discussion in Geneva was not available to the NSABB. My recommendation is to reassemble the working group of the NSABB for a face-to-face meeting with the investigators, security people, international public health people (similar to the WHO meeting) and a larger number of influenza experts with hands on expertise in such studies. Post- Geneva: Have spent the past few days explaining the above summary to HHS, White House NSS X 2. Could not comment too much to the press.