COVID-19: What Now?

Written May 22, 2020. Updated Treatment section July 6.

We’re entering a different phase of the pandemic now, whether it’s safe to return to normal in your area or not. I didn’t want to make the original Q&A page too long and most of the details there still apply, but I did want more space to talk about possible tests and therapies and some immunology geekery so I’m starting this new page. The comments section is still a place where we can ask questions, post updates, commiserate, and help one another. I might update the original page from time to time. If I do, I’ll copy those updates in the comments here too so you don’t have to look in both pages. As always, I’m happy to take any suggestions.

  1. Know thy enemy: The replication cycle of SARS-CoV-2
RNA and protein made by SARS-CoV-2 become antigens for the immune system to attack and the life cycle stages can be targets for antivirals.

2. Test -> Isolate -> Contact Trace

Basic epidemiology control relies on a) knowing who is infected and who is not, b) isolating infected people, and c) contact tracing everyone who may have had contact with the infected person to isolate them too. For what to do if you need to self-isolate, refer to my original post.

3. Viral RNA Testing

To really know if a person has infectious virus, the gold standard would be to isolate virus from the patient and perform a plaque assay, in which the amount of virus is quantified by the counting the number of target cells that are killed. However, this is only allowed in special labs with higher biosafety protections and it takes multiple days.

The RT-PCR test is the most common test that is used now and will likely remain the de facto standard. This test can be performed in regular Biosafety Level 2 labs because any active virus is inactivated immediately after collection and return results in a few hours. Where the test can cause confusion is that it amplifies small fragments of RNA, which don’t necessarily mean infectious virus is present anymore, and there’s a very long tail of RNA fragments present in people who are no longer infectious. This paper shows that past day 8 of symptoms samples taken are still RNA-positive but have no infectivity.

The standard RT-PCR test for virus RNA fragments will tell you about the presence of RNA fragments, not functional virus.

4. A crash course in Immunology with a focus on antibodies

The innate immune system buys time for the adaptive immune system to ramp up.

Any pathogen that enters the body will leave traces of itself (e.g. RNA, protein), called antigens that the immune system detects and starts fighting. Cells of the innate immune system are the front line of defense. They recognize antigens that are broadly common in pathogens but not humans (e.g. flagella of bacteria, RNA sequences that look foreign). They are great at chomping up pathogens and sending out chemical signals, called cytokines, to alert the rest of the immune system that an infection is happening. But innate immune strategies can be evaded easily so we also have the adaptive immune system.

Pathogens evolve much faster than humans could ever hope to compete with. The adaptive immune system is a mini evolutionary system within the body that generates large specific immune attacks within a week AND results in immune memory for the next time the same antigen shows up. Starting with a pool of naive B or T cells, each of which expresses receptor for only one antigen on its surface, the cells only receive an activation signal when the correct antigen binds to their receptors. Only activated cells replicate so you very quickly get a large number of B and T cells that are specific for the antigen and ready to kill.

In a pool of B cells with diverse B cell receptors, antigen X only binds to the receptor that has the complementary shape for it. This triggers the cell to replicate and turn into plasma cells, which pump out large quantities of antibodies. Plasma cells die out quickly when an infection is over, so there’s another type of B cell, called memory B cell, whose role is to survive quietly until a secondary infection.

Antibodies are Y shaped proteins that are very stable. The two “top parts of the Y” are very varied to be able to bind to a great diversity of antigens. B cell receptors are antibodies that are on the surface of a B cell, which is how selecting for antigen-specific B cells also ends up selecting for antigen-specific antibodies. B cell receptors and antibodies only contact with antigens outside of a cell. Since a virus can hide inside of a cell away from antibodies, the T cells take over and search out cells containing virus to kill. T cells are fascinating and the subject of my research but they are much more complicated to deal with so they’ll never be part of large scale diagnosis tests. Designing vaccines that trigger T cells is pretty important though.

Lastly, antibodies come in several isotypes but the two most relevant ones are IgG and IgM. IgM is more of an innate immune defense and their levels go up first when an infection begins. The binding regions tend to bind loosely but because the structure is 5 antibodies stuck together, they get better binding through many contact points. After the B cells are activated and go through a process called affinity maturation, the IgG you get is the upgraded version of IgM antibody. IgG antibodies bind very strongly and specifically to the antigen and will be produced by memory B cells in a secondary infection. Presence of IgM and IgG together is strong evidence of an immune response to infection.

For an anthropomorphic anime of the body’s immune system that has pretty good accuracy, Cells At Work! (Hataraku Saibo) is pretty fun and available on Netflix.

5. Does infection actually provide immunity? Can I get reinfected?

This preprint shows that practically all patients with mild disease, produce measurable antibodies to SARS-CoV-2. This paper in Cell showed that T cells that recognized SARS-CoV-2 were detected in most recovered COVID-19 patients. Specifically, 70% had CD8+ T cells (responsible for direct killing) and 100% had CD4+ T cells (helps B cells). This is a good sign that yes, we’re seeing the products of immune memory.

When it comes to the common cold coronaviruses, people did get productively reinfected within a year with the shortest time difference being 4 weeks. There was no relationship between symptom severity and reinfection, so no obvious increase or decrease in severity. All the individuals who were completely asymptomatic during the first recorded occurrence, did not report any symptoms during subsequent infection(s) with the same coronavirus type. My interpretation is that mild infection may not provide complete immunity from reinfection but reinfection is going to also look mild. Preprint

6. Antibody / Serological testing

RT-PCR most useful before symptoms show up or in first week of symptoms. IgM and IgG only useful after week 2 or later, which unfortunately makes them useless for identifying active infections. IgG is more useful than IgM. Source: RebelEM

There’s a lot of interest in antibody testing because RT-PCR test capacity was so limited and are only useful for identifying currently infected patients. Antibody testing may tell us who is immune and serological studies may tell us how much of the population is already immune. So let’s talk about the pitfalls of antibody testing.

Binding to the square part is useless if its the triangle that’s allowing viral entry into the cell

One of the main functions of an antibody is to be a physical barrier to stop the function of the antigen. So in SARS-CoV-2, the spike protein on the virus binds to ACE-2 receptor on a cell to enter the cell. Block the spike protein -> stop the virus from entering the cell. But the spike protein is still a 3D structure and your B cells have no foreknowledge of which part of the spike protein is worth blocking. Hence you can make non-neutralizing antibodies, which bind at the wrong site on the spike protein but will still trigger the B cell to replicate, versus neutralizing antibodies, which bind to the correct site and will prevent the spike protein from fulfilling its function. Non-neutralizing antibodies are kind of worthless in immunity but difficult to distinguish from neutralizing antibodies in a serological test and that can give rise to false confidence of immune protection if you get a positive antibody test. The gold standard for studying neutralizing antibodies has been to use them to block virus entry in a plaque assay and as I said before, that assay takes multiple days.

7. Real world testing error rate is not constant. It varies with the prevalence of virus or immune people in the population

Figure plagiarized from Biocentury, which has the full explanation.

8. Rapid tests that may be used at home or in a doctor’s office

With the caveats about rapid tests in mind, I suspect they are going to be available everywhere. Here’s a side by side comparison. The molecular RT-PCR test becomes a rapid test by cutting out the temperature changing steps.

The FDA has approved 3 types of rapid tests.

FDA webpage listing of approved serological tests with the caveat that they report specificity and sensitivity rate provided by the companies. When actually evaluated independently, real world specificity and sensitivity drops.

There’s also a rapid test for neutralizing antibodies from Singapore that is interesting. Instead of the gold standard test of blocking infection of cells by the virus, they’ve mimicked this interaction in a plate. The surface of a well is covered with ACE2 receptor. Add serum from patient and then wash off. Then add receptor binding domain (RBD) that has a reporter enzyme (horesradish peroxidase converts TMB to a blue color). If the patient serum has neutralizing antibodies, they bind to the ACE2 receptor and prevent RBD from binding. Ergo, a negative result means the well turns blue and a positive results means the well stays colorless. This can’t be done at home or in a doctor’s office but it may be useful in clinical labs.

Neutralizing antibody rapid test. Preprint

9. Are we reaching herd immunity yet?

List of sources

Not even close. Best seroprevalence study tracker I’ve found so far is this Google spreadsheet by Henrik Jarlov. Spain had the largest survey of the general public so far, with 60,000+ people tested. 5% of people were positive for antibodies. With this, they estimated that 90% of infections went undetected and with 27,100 deaths, an infection fatality rate of 1.1%. The NYC infection fatality rate is estimated at 1.3%.

10. Treatment options

Soooo many options. To keep track, see:

Remdesivir: Randomized control trial done with 1063 participants across 60 trial sites. Convincing statistically significant improvements in recovery time, improving it from 15 days to 11 days. Additionally, the study also showed that 8% of patients who took remdesivir died compared to 11% of patients who received the placebo, though the number of deaths were not enough to make numbers statistically significant. Remdesivir will now replace placebo in the control arm of clinical trials. Unlike hydroxychloroquine, no difference/slightly fewer adverse events reported in remdesivir arm compared to control arm. Preliminary Report at NEJM.

Dexamethasone: From Oxford RECOVERY Trial Reduced deaths by one-third in ventilated patients and by one fifth in other patients receiving oxygen only. There was no benefit among those patients who did not require respiratory support. Preprint. These are the best results we have so far.

No clinical benefit from hydroxychloroquine or lopinavir-ritanovir.

11. Vaccine options

ChadOx1 vaccine (Oxford + AstraZeneca)

Viral vector vaccine. The chimpanzee adenovirus is stripped of all the genes that allow it to package a new virus to get out of a cell. Without these genes, the adenovirus can only enter a cell once and never leave. Then the sequence for the SARS-CoV-2 spike protein was added. The result is a virus that enters a cell and tells it to make SARS-CoV-2 spike protein only. That spike protein triggers the immune system.

ChAdOx1 is running its preclinical trials and clinical trials in parallel. They just released a preprint showing preclinical results on May 13 showing that in rhesus macaque studies showing that one dose of vaccine elicited an immune response. They used a very large inoculate of virus and found that while there was no difference in nose swab RNA, vaccinated animals had much much lower RNA in the lungs and no signs of pathology. I’d rate this as results to be cautiously optimistic about. A vaccine doesn’t have to be all or nothing. We’re in a Moneyball situation where every little bit that reduces transmission and severity of symptoms adds up.

Moderna mRNA-1273 mRNA vaccine

Messenger RNA for spike protein is packaged in a lipid nanoparticle which enters a muscle cell via simple fusion. The cell automatically starts making spike protein from the mRNA.

Press release, not paper, dated May 18. After two doses all 8 initial participants evaluated to date seroconverted with binding antibody levels at or above levels seen in convalescent sera. mRNA-1273 was generally safe and well tolerated. mRNA-1273 provided full protection against viral replication in the lungs in a mouse challenge model.

12. What if we never find a vaccine? What if high temperatures in summer cause lower transmission?

What Happens Next? COVID-19 Futures, Explained With Playable Simulations by Marcel Salathe (epidemiologist) and Nicky Case (art/code).

13. Metrics for Phased Reopening A lovely site with clear metrics that will almost certainly be ignored by politicians.

14. My politicians are idiots. What can I do personally to keep myself and people around me safe?

15. Other resources