Guest post by Rud Istvan,
Rumination #5 developed possible explanations for two emerging observations:
-Why more CVN71 sailors did not test positive for the Wuhan coronavirus because their adaptive immune systems were not naïve;
-Why hypertension, diabetes, and obesity are the main NYC comorbidities in the serious/critically ill under age 65 CoViD-19 patient cohorts.
This guest post extends those two ideas to new observational information about antibody testing, and a very recent mainstream media (MSM) observation that an abnormally high percentage of ICU CoViD-19 patients have various forms of thrombosis. As with guest post #5, there is a high reliance on the medical literature searchable at www.ncbi.nlm.nih.gov. The most relevant literature is referenced in what follows by PCM number. For those interested in further personal research, this post also provides key words and concepts for Google exploration.
Naiveté and serologic antibody testing
CVN71 sailors were tested for Wuhan virus using PCR, which looks for the active presence of Wuhan RNA. About 15% were positive; the remainder had no evidence for the virus despite the extreme ‘lack of social distancing’ that concerned the Captain so much he was removed from command. The naiveté hypothesis speculates that (recent) exposure to the four common cold coronas may have conferred a degree of Wuhan immunity; much like vaccination with cowpox provides immunity to smallpox. This would be especially true for the two ‘beta’ serotypes OC43 and HKU1, as Wuhan is also a beta serotype. This hypothesis would also explain positive but permanently asymptomatic cases (20% in South Korea); a weaker (older common corona cold?) but still a sufficient timely response by the alerted adaptive immune system.
BUT, if true, then this hypothesis also presents a potentially seriously confounding factor for the serologic antibody tests now being developed–so far with little to no regulatory oversight or rigorous independent evaluation. To the point, Roche’s CEO said publically last week that many of these ‘would be of no or little value’. To understand why requires three background science ‘facts’.
First, the antibodies being tested for come in two flavors. IgMs form early and decline ‘quickly’. They say you recently had the virus, but do not confer lasting immunity. IgGs form starting about 10 days into a viral infection, peak at about 21 days after infection, and are what confers any degree of lasting immunity.
Most of the antibody tests being developed are qualitative ‘color changing’ lateral flow tests for IgM, IgG, and a control strip (which if it does not change color means the test failed). They are the same in operating principle as home pregnancy test kits.
Second, these antibodies can work in two ways. The ‘minority’ way is as a ‘neutralizing’ antibody that binds to and directly inactivates a virion in some fashion. This is the hope behind the convalescent plasma treatments now in clinical trials. For coronaviruses, there are three targets that potentially neutralize:
- The N protein (the nucleocapsid coat) which, when interfered with by a neutralizing antibody, prevents the encapsulated RNA from ‘extracting itself’ to enable intracellular replication. As a sub-result of a separate pulmonary study concerning influenza, 105 adults with COPD were tested for N antibodies to the four common cold viruses. 104 tested positive to 229E, 105 to OC43, 103 to NL63, and 96 to HKU1. VERY HIGH. The Wuhan N is likely to be similarly highly specific, a good thing for serologic antibody testing as explained below.
- The S (spike) protein, which has two neutralizing antibody binding sites, S1 (the ACE2 receptor), and S2, which varies by serotype. S1 lets the virus latch onto a cell, after which S2 enables the virus to enter the cell. The problem is that Wuhan uses the same S1 for ACE2 as the common cold coronas—hence the naiveté hypothesis.
The ‘majority’ way antibodies work is just ‘binding’ somewhere to some protein fragment. Corona has M (membrane), E (envelope), and S other than S1 and S2 as possibilities. These do not inactivate the virus, but do signal adaptive immune system cells to eventually come clean up as they multiply in response. There has been a fair amount of work on the original SARS, showing that alveoli macrophages can ‘ingest and get infected’, but that the virus cannot replicate inside the macrophage. A ‘cleanup needed on isle 3’ announcement mechanism. What is not yet available in the literature is how similar/dissimilar these binding antibodies might be between the common cold coronas and Wuhan. The more similar, the more likely both IgMs and IgGs will be insufficiently specific.
Third is the sensitivity and specificity of the resulting antibody test. Sensitivity is how many true positives are missed. Specificity is how many false positives are generated. As of now, only a few manufactures have offered their internal (unverified by the FDA) estimates based on small samples of people. As one would expect, results vary. Becton Dickenson says their new test is 88% sensitive and 90% specific. The importance of this requires using the formal probability mathematics of Bayes Theorem. As an example from a very new clinical lab article, if you assume the true prevalence of Wuhan is 5%, then a 90% specific antibody test (BD as advertised) will produce 70% false positives and be essentially useless for CFR denominator and herd immunity purposes. (Evaluation.com, an on line journal for clinical labs, Cairn article from 4/22/20)
Unknown is how sensitive and specific these tests can be made by careful antibody selection. The widely cited new Stanford paper on 3300 Santa Clara residents said it used a lateral flow test of both IgM and IgG from Premier Biotech in Minneapolis that it had independently validated in a small sample as a combined (different methods) 80% sensitive and ~99% specific– without separately checking for common cold corona N. So I checked Premier. Turns out they are only distributing a test from Hangzhou Biotest Biotech, Ltd. And Biotest’s Chinese website doesn’t provide specificity and sensitivity estimates at all; the specifications page is blank.
The Roche CEO is correct—we have a ways to go yet with antibody testing.
This deadly ‘new’ complication has been much in the news in recent days after Broadway star Nick Cordero, age 41, had to have his right leg amputated as a result of CoViD-19 induced deep vein thrombosis. Media reports are that the incidence is between 20-25% of all ICU patients. This could certainly be a further explanation for the number of cardiovascular deaths from ‘heart attack without atherosclerosis’ pointed out in #5.
One possibility is that propensity to clot is known to be disproportionately high in those with diabetes and hypertension, so just another sequelae of obesity. But Cordero was not obese and reportedly healthy, without hypertension or diabetes.
There is another. PMC3809294 points out that platelets must have roles in addition to clotting. The paper estimates that thrombosis requires about 10E+9 platelets/liter of blood, while typically there are about 250E+9 per liter. To initiate clotting, platelets must be ‘activated’. PMC6048695 says that it is known that interactions between platelets and viral pathogens results in activation via plasma proteins called kinocidins. And finally, newish PMC4270245 has a detailed mapping of platelet activating receptors. It turns out that common cold coronaviruses directly active platelets via receptor GPV1, and that IgG antibody/antigen pairs directly activate platelets via the FcyR11 receptor.
As a result of this new knowledge, it would appear that adding an anticoagulant such as clopidrogel (Plavix) should become standard prevention therapy in serious/critical hospital cases. There is however as of today (4/23/20) no medical consensus on that, any more than about hydroxychloroquine plus zinc as a treatment therapy.