Like many responsible citizens of the USA, I am hunkered down in self quarantine.  Trying to come to grips with new uncertainty in the world and how it will have impact on my life.  I find myself wanting to help. To do what I can to help stop this beast we call COVID19. But How?

Collective Behavior Modification is Part of the Answer.

What we are dealing with here mutually is a form of trauma.  We are experiencing the Kuber-Ross grief cycle. Each of us are oscillating somewhere along its spectrum.

Many of us are stuck in denial.  Although as of late, I seem to be bouncing from anger to bargaining to depression and back again (note, this entire piece for the blog is an elaborate form of bargaining!!!) 

I want to get to acceptance and start doing constructive things.

Shedding the denial by attempting to understand the size of the problem.

If you somehow have been in a timewarp, or you have been tunneling under a rock for the past few months, COVID19 is a raging pandemic that is threatening to kill our parents/grandparents. And it is causing major economic crises throughout the world. Some of our leadership say this will pass in just a few weeks from now, but when you dive into the data, that suggestion is just plain ludicrous fantasy. It will take many months of concerted efforts by everyone to tackle the COVID19 problem. If we choose to ignore what we need to do, the ramifications are immense.

Some are calling it a war, so lets look at it from that lens.

The civil war was an extremely traumatic event for my home country of the USA.  We are still dealing with its baggage to this day and its death toll was immense. World War II was a big effort, industries stopped what they were doing and channeled vast resources to the war effort. Many a dad and mum did not come home, but the concentrated effort of its population probably helped reduce loss of life.  When we look at the loss of life from World War II we have a significant 290 deaths per 100,000. Everyone knew someone who did not come home. Yet, that death rate pales in comparison to what unmitigated COVID19 could do to us in the coming months if unchecked. More than twice as many dead if we do nothing.

Is All This Stuff Real?  – Yes, ….This is Getting Real, …Real Quick

A recent article by a very respected group of researchers paints an alarming picture. Ferguson et al released a study March 16 that describes 3 scenarios each with different consequence on the health system measured in critical care beds occupied per 100,000 population.

Scenario 1. Do Nothing. 

In this scenario (black line) we get the 2.2 million dead in the USA. The surge capacity of hospital beds is overwhelmed 25x.  Many die that could have been saved, if we had the resources.

Scenario 2. Mitigation (“soft”)

In the mitigation scenario we have multiple options. We can close schools and universities. This surprisingly does not have much effect (green line). The algorithm takes into account “Household contact rates for student families increase by 50% during closure. Contacts in the community increase by 25% during closure” which offsets gains of social distancing achieved with school closure.  Case isolation by itself has a stronger effect (orange line). For Case Isolation, symptomatic cases are asked to stay home for 7 days. This reduces “non household contacts by 75% for this period. Household contacts remain unchanged. Assumes 70% of households comply with the policy.” Add to Case Isolation a Home Quarantine where family sequesters for 14 days (assume 50% comply), and we get a boost that has more than halved the lethality rate. Finally add in social distancing in the greater than 70 year olds, where this group sequesters themselves away from close contact by avoiding crowd gatherings, maintaining 6 feet distance from strangers, avoiding restaurants and the like, and we get additional decrease in hospitalization (blue line). Yet even this multi-step mitigation effort is not enough. It is too soft to have the needed impact. Even with the multiple mitigation measures in place, the capacity of the health care system is still overwhelmed by more than 8x. More severe and hard measures are needed.

Scenario 3 Suppression (“hard”)

In mitigation, we are creating a decrease in the transmission number (R0 = R “naught”), which is the average number of persons infected by a person who is actively shedding the COVID19 virus.  The do nothing R0 number is 2.4. This means an infectious person spreads COVID19 to an average of 2.4 persons. Mitigation decreases the R0 number, but does not drive it down to 1 or below. To get R0 below 1 where the infected are infecting less than one person, bigger steps need to be taken by the entire population.  To implement effective suppression, we do the Case Isolation, as seen in mitigation, but now we add in General Social Distancing. We ask everyone to avoid getting together in groups exceeding 10 persons. Each of us should maintain 6 foot distance when talking to others. We should capture a sneeze or cough in the elbow or a tissue, and wash hands way more frequently than usual. Further, we should avoid touching our hands to face as much as possible.  It is expected that “all households reduce contact outside the household, school or workplace by 75%. School contact rates are unchanged, workplace contact rates reduced by 25%. Household contact rates are assumed to increase by 25%.” Yet this his is still not enough to get to R0 down to 1, so we explore two more steps as options.

• Option 1. (Case Isolation and General Social Distancing) + Household Quarantine

• Option 2. (Case Isolation and General Social Distancing) + School and University Closure

We can see that adding in Household Quarantine, during the 5 months that suppression measures are in place, gets us just to the hospital bed surge capacity (red line). Yet in contrast, School and University Closure has an even more pronounced effect.  During the 5 month period, we are well below the stress capacity of the medical services. We will know the success of our suppression approaches when there is no capacity problem being detected.

Herd Immunity – Resisting the Urge to Celebrate Too Soon.

Unfortunately, when the suppression measures are lifted, the COVID19 is due to come roaring back this fall. The advantage of option 1 (Household Quarantine) is it allows for better development of Herd Immunity.  From wikipedia we have the definition of Herd Immunity as occurring “when a large percentage of a population has become immune to an infection, whether through previous infections or vaccination, thereby providing a measure of protection for individuals who are not immune”  With an R0 of 2.4 for COVID19, roughly half of a population needs to be exposed, either by recovering from infection or use of effective vaccine, and R0 is driven down to 1 or less.

Pharmaceutical Interventions – what can we do?

Chloroquine. Currently there are no approved treatments as a vaccine or drug therapy against COVID19. Yet, if we had them in place, we might be able to more quickly get control of this infectious disease and be spared the estimated 12 to 18 months of various mitigation and suppression techniques – approaches that slowly build herd immunity at the least amount of deaths. Scientists and the pharmaceutical industry are rallying quickly get pharmaceutical interventions in place.  And they have made some interesting findings. Hydroxychloroquine, less toxic than Chloroquine, has undergone a small clinical trial on COVID19 patients with very encouraging results. Gautret et al observed a dramatic 2-3 fold faster clearance of SARS-CoV-2 (the “official name” of the COVID19 virus) relative to untreated controls. The results became even more dramatic for another small group of patients that received both hydroxychloroquine and azithromycin – complete clearance in all patients was observed at day 4 of the 6 day observation window.  Yet the numbers tested are tiny. So repeating this in larger populations will tell us if they are onto something. Nevertheless, it is a promising start. The current hypothesis is that this antimalarial drug blocs viral envelope fusion by altering the pH of the endosome and thereby slowing down the activity of the acid proteases present (cathepsins or possibly TMPRSS2).  Yet it is important to consult a doctor first, because self medication has resulted in unnecessary death.

The action of chloroquine may be multimodal.  In 2005, it was demonstrated chloroquine in a SARS-CoV infection of a cell line caused incomplete glycosylation of ACE2 and that it can have “an antiviral effect during pre- and post-infection conditions suggest that it is likely to have both prophylactic and therapeutic advantages.”

Camostat mesilate. In a drug targeting approach, Hoffmann et al monitored the classic endosomal-lysosomal entry for coronaviruses with an endosomal fusion assay. They found entry into the cytoplasm to be mediated by the activity of TMPRSS2 and cathepsin proteases. First, these authors made a comparison between SARS-CoV (the coronavirus causing the pandemic of 2012) and the SARS-CoV-2 (the coronavirus causing the current COVID19 pandemic) and found S protein In COVID19 is more prone to cleavage at the S1/S2 site. Next, they looked at inhibitors of cathepsin (E-64d) and TMPRSS2 (camostat) and found that, depending on cell type, inhibition by either protease could interfere with early fusion. When they looked at a lung cell line (Calu-3), they found that camostat could strongly inhibit early fusion. The site of cleavage for TMPRSS2 protease is the same site for furin and loss of this ability to cleave S-protein is critical for viral entry into the cell. Further Camostat is nearly established for efficacy and safety in humans for treating pancreatitis.  In summary, these researchers found “SARS-CoV-2 can use TMPRSS2 for S protein priming and camostat mesylate, an inhibitor of TMPRSS2, blocks SARS-CoV-2 infection of lung cells.”   So we have two very good candidate molecules for use in suppressing COVID19.

Smoking Gun – But Where is the Expression?

Although camostat can have a dramatic impact on early fusion and it appears to be acting on TMPRSS2 serine protease, what has puzzled me is the tissue specificity. If the primary mode of infection is the lung, then it stands to reason that lung tissue should have high expression of the protease. Yet when one looks at the tissue-specific expression profiles on the Human Protein Atlas (HPA), the expression of TMPRSS2 is absent in the lung.

TMPRSS2 is related to TMPRSS11D 55% sequence similarity and 40% identity). The  Pharos Drug Database indicates the TMPRSS11D gene may also be involved in coronavirus fusion in the cell.  Like TMPRSS2, the Pharos database indicates TMPRSS11D protease “plays a role in the proteolytic processing of ACE2.” Intriguingly, the TMPRSS11D gene has a signature for expression for being in lung tissue via Human Protein Atlas (HPA) (dark green bar). Further, examining ligands for these two genes via the Pharos Drug Database indicates both proteins share two ligands (compound 5 and its derivative CHEMBL1809251). Since these shared molecules have similar binding affinities between the proteins, it may be the  topography of their active sites is similar. Although they both can cleave ACE2, it remains to be shown if they also have similar activities on the S protein. Yet if true, then we have two enzymatic targets for therapeutic development.

ACE2 expression also puzzled me. Its protein expression overview in Human Protein Atlas (HPA) shows no lung expression but expression in the gut is off-the-charts. Auguring in on two papers indicates that its expression does occur in the lung, but only in a small subset of lung cells . These references indicate many other tissues have high expression of of ACE2 (gut and throat). Further examination of tissue localization for SARS-CoV-2 indicates the following tissues exhibit high infection. We have our expected lung (alveolar epithelial cells) but we also have gut (mucosal enterocytes of the intestine, stomach, trachea/bronchus, distal convoluted renal tubule, sweat gland, parathyroid, pituitary, pancreas, adrenal gland, liver and cerebrum) and many of these tissues also express ACE2.

Using Molecular Dynamics Modeling to Dock Candidate Compounds

If an enzyme has a good crystal structure, one can do simulated docking of compounds to come up with categories of interacting molecules that can be used as hits for exploring their capacity to block viral entry into target cells. Regrettably crystal structures are lacking for both TMPRSS2 and TMPRSS11D. Not true for ACE2 gene, since the finding of its association to SARS in 2003, multiple crystal structures are available (6LZG, 6M0J, 6M17, 6VW1). In one recent structure, we can see the binding interface between CoV-2 and ACE2. One approach might be to design an interference molecule that “cloaks” the ACE2 molecule. Ideally, it would interfere with S protein binding but not block normal enzymatic activity. 

Another possible target for drug development is the main protease (“Mpro”). The main protease (also called “3CLpro”) is a protease that helps process the long protein polymer made from the viral genome into functional fragments.  The main protease as been derived for its structure at the molecular level. A peptidomimetic α-ketoamides as broad-spectrum inhibitors has been designed and if safety profiling indicates it has low toxicity against human proteases, screening other drug candidates for interaction could prove to be therapeutically useful.

Time is of the essence – the evolving landscape

Finding new therapeutic approaches quickly is important because as COVID19 spreads, the virus is able to explore diversity through mutation. Increasing levels of heterogeneity can be expect for a single stranded RNA virus – errors in the genome during the viral replication cycle will accumulate.  A recent study was made available on the web that looks at genetic diversity among COVID19 strains.  Especially disconcerting is that as the virus spreads, different strains are arising with different mutational lineages.  Mutations in the S-peptide binding to ACE2 or in the S1/S2 cleavage site could render a new lineage that is more virulent.  This is important because the binding affinity (Kd) of SARS-CoV-2 is nearly 5x more than Sars-CoV.

Fight Viral Diversity with Human Diversity

The natural diversity of human populations might offer some defense against viral evolution.  The Gnomad database is a good resource for examining the diversity in humans. If we look at the binding interface between S-peptide and human ACE2, we cans see if there are humans with variations at the binding interface which may disrupt COVID19 infections. There are multiple EM structures to reference for examining the binding interface (6LZG, 6M0J, 6M17, 6VW1). Using the 6M17 structure, Yan et al showed the binding interface is in close proximity to many residues in ACE2 (Gln24, Asp30, His34, Tyr41, Gln42, Met82, Lys353, and Arg357).

Many gene variants seen in human populations exist at this binding interface. One intriguing residue is Lys26Arg. This genetic variation exist in 600 for every 100,000 persons. Although the change may not seem to be too drastic – a positive charged amino acid is substituted with a similarly charged amino acid.  Yet we know that it is arginine, and according to a prior blog post, the arginine amino acid hold special privilege for its involvement in population variation analysis.

We can float two hypothesis regarding this Lys26Arg variant  

Hypothesis A – Resistance: Persons with the Lys26Arg human variant might have an ACE2 protein with disrupted interaction to the S-protein spike. The S-protein becomes highly compromised for tricking this ACE2 protein into helping it get inside the cell.  

Hypothesis B – Sensitivity: Persons with the Lys26Arg human variant might have an ACE2 protein exhibiting stronger interaction to the S-protein spike. The S-protein can use this ACE2 protein to get inside the cell more efficiently.

The Problem of Heterogeneity

We are diploid organisms which means we have two copies for every gene.  In regards to hypothesis A, this mean most of the persons carrying the Lys26Arg have only one copy.  The other copy is the common natural variant (“wildtype”). Because they harbor a wild type variant, these persons at best would about 50% less susceptible. Thus hypothesis A for COVID19 resistance would be a subtle effect.  If on the other hand, the variant had 10x has more binding to S-protein, then persons carrying this variant could be more susceptible at greater than 2 fold effect. Systems that could measure these binding effects in diploid animal formats would be elucidating for which hypothesis dominates for a given variant.

In summary, there are promising targets for developing therapeutics and vaccines. One target is the interaction of SARS-CoV-2 with ACE2. Another target is the activity of the human proteases (TMPRSS2, cathepsins and possibly TMPRSS11D) that process and cleave the S2 fragment of the spike (S-protein) allowing it to have easier access into the cell. And finally, the third target is the main protease, the enzyme that processes the polypeptide made from the mRNA transcript of the COVID19 genome.