LESSWRONG
LW

by Spiracular
4 min read14th Mar 20203 answers12 comments

35

Covid-19
Frontpage

The possibility of SARS-CoV-2 having Antibody-Dependent Enhancement (aka ADE) looked pretty real, to me.

If you're curious about the challenges of vaccine development for SARS-CoV-2, I recommend this article.

UPDATE: It looks like it isn't productively replicating in WBCs, but it probably is fusing with them and telling them to apoptose. Receptor uncertain, but they were checking T-cells specifically, which are exactly the WBCs that get severely depleted in severe COVID-19. They were in-vitro studies, but this mechanism matches the in-vivo results I'm seeing better, and I find the idea moderately convincing. (Article, h/t CellBioGuy's post mentioning it). SARS-2 is apparently much better at this than SARS-1.

Define ADE

Producing antibodies that are imperfect matches for one of these viruses (ex: optimized for another strain, or incompletely neutralizing) not only do not inactivate the virus, but instead get repurposed by the virus as a mechanism it can use to anchor and infect the cells that try to interact with those antibodies, most often immune cells.

Current conclusions

TL;DR: I think something else is having a larger impact than ADE on the immune system in severe COVID-19 disease. I have seen several theories, and listed some below. SARS-1 and MERS have both seen plenty of bad vaccine reactions that only show up at the animal-testing stage, and I still think vaccine development is going to be hard.

While I don't understand the upstream causes of this, Th-2 type activation has been extensively noted in both severe COVID-19 disease and bad vaccine reactions. Th-2 type immunopathology (roughly, allergy-like immune responses in lieu of virus-like immune responses) seems likely to play a large role in both complicating vaccine development, and influencing disease severity.

Something weird is going on with white blood cell counts, but I'm currently leaning towards believing it might be something else causing it. Specifically, they're seeing T-cell lymphopenia. T-cells seem to be the worst-hit WBC (hyper-activated, decreased numbers), and I'd have expected them to be close to immune to Fc-based ADE once mature; they only express the receptor while young.

SARS-1 and MERS vaccines both seem to have seen instances of bad reactions that required animal testing to become apparent. SARS-1 and SARS-2 exhibit ADE in-vitro against S-protein vaccines and not N-protein vaccines, but both S- and N-targeting vaccines sometimes had bad reactions in animal testing*. What I see as an in-vitro/in-vivo divergence boosts my impression that vaccine development will be challenging.

*For example, this mouse study with SARS-1 saw hypersensitivity reactions to several S-containing vaccine subtypes (whole virus, VLP, S-protein), while this paper tested S- or N- VRPs and found that their N-inoculation not only had had no protective effect, but made things worse when it "resulted in enhanced immunopathology... within the lungs" upon infection.

Going on the current human results and a tiny Macaque study with 4 monkeys, for SARS-CoV-2 it does appear that getting through the disease once does preclude or largely-preclude reinfection by the same strain. I don't feel I can weigh in on whether this will stay true.

A few theories trying to explain the blood results (I'm sure there's more):

  • Glucocorticoid reaction/Too much cortisol as a possible upstream cause of the lymphopenia + neutrophilia reaction seen in SARS, RSV, Ebola
  • Something is causing/reacting to the cytokine storm
    • Cytokines essentially steer the strategy of WBCs (activation/direction/activity)
    • Both severe COVID19 disease and bad vaccine reactions steer towards Th-2 type immune responses (vaguely allergy-like)
    • I found this a bit too confusing to wade into
  • Some sort of immune-cell suppression effects
    • Indirectly impacting cells at an earlier stage of blood cell development/differentiation (possibly as early as bone marrow stem cells)
    • Infecting cells via its additional receptor-binding affinities, for purely manipulation purposes (probably without viral replication)
      • It seems to clearly prefer lung and bowels for productive viral replication; we aren't seeing viral inclusion bodies in most other tissues. I don't feel this rules it out.

While I don't understand the upstream causes of this, Th-2 type activation has been extensively noted in both severe COVID-19 disease and bad vaccine reactions. Going on both the frequency with which it comes up in these contexts, and commentary by experts, Th-2 type immunopathology (roughly, allergy-like immune responses in lieu of virus-like immune responses) seems likely to play a large role in both complicating vaccine development, and influencing disease severity.

Earlier thoughts

I've repeatedly had to update in the direction of it being plausible, and I currently think it's more-likely-than-not to be a factor that will complicate vaccine development.

However, there do exist viable alternative theories for a lot of what I'm seeing, some of which I couldn't rule out.

I wanted to consolidate research on this into one place, and am interested in if anyone has additional solid arguments for/against it.

I have a lot of questions about this, but lets boil it down to a few.

  1. Is this virus doing this in-vivo? What induces it? (note: it does happen in-vitro, but in-vitro is apparently easier to induce, and often not conclusive evidence that this happens in-vivo at appreciable levels)
  2. What exactly are the consequences of this ADE interaction? Can we get it down to symptoms?
  3. How does this change things?
  4. What can we do about it?

New to LessWrong?

Getting Started

FAQ

Library

Covid-19
Frontpage

35

Does SARS-CoV-2 utilize antibody-dependent enhancement?
14th Mar 2020
New Answer
New Comment

3 Answers sorted by
top scoring

Spiracular

Apr 16, 2020

130

This article largely addresses the challenges of vaccine development for SARS-CoV-2.

It was extremely clarifying for me. It seems to have consulted several experts for opinions, but is written for close-to-layman reading.

Here are a few bullet-points, but I recommend the whole thing.

  • Th-2 type immunopathology, "in which a faulty T cell response triggers allergic inflammation," plays a big role in bad vaccine response to SARS-CoV-2
    • Ralph Baric: "There is the potential for ADE, but the bigger problem is probably Th2 immunopathology"
    • Me: Th-2 responses are extensively documented in both severe COVID19 disease, and SARS-1's bad vaccine responses. This makes a lot of sense. I would like to understand what is upstream of that, sometime.
  • Another likely problem: Complement system malfunction
    • "poorly functional antibodies that form immune complexes, activating the complement system and potentially damaging the airways"
    • Me: There is most likely a scientific term for this phenomenon. I'll add it when I find it.
  • ADE is possible, but less than likely to play a major role
    • FIPV is a coronavirus where ADE is a major concern. But FIPV productively infects macrophages. I've seen no evidence so far that SARS-CoV-2 productively infects WBCs.
    • UPDATE: It looks like it isn't productively replicating in WBCs, but it probably is fusing with them and telling them to apoptose. Receptor uncertain, but they were checking T-cells specifically, which are exactly the WBCs that get severely depleted in severe COVID-19. Consider me pretty convinced. Article, h/t CellBioGuy. (P.S. SARS-2 is much better at this than SARS-1.)

For RSV* (Respiratory Syncytial Virus) and SARS-1, bad vaccine response gets blamed on both Th-2 immunopathology and antibody complexes activating the complement system. This can lead to severe disease, including the infiltration of lung tissue by neutrophils (for both) and eosinophils (for RSV?).

For SARS-1, whole-S-protein vaccines seemed more likely to produce this detrimental enhanced immune response. S-fragments containing only the receptor binding domain offered protection, but did not produce this unwanted effect.

They are hoping that the same logic and fix goes for SARS-2, but are still waiting on the test results.

*Me: Side-tangent, but both SARS and RSV seem to be nasty lung infections with some propensity to form syncytia in severe disease. Heck, one of SARS-2's novel mutations seems likely to be involved in increased syncytia formation. However, the viruses are not close relatives, and they seem to show different preferences for which particular lung cells they reshape into syncytia. I find it interesting that they ran into a similar bad reaction here.

[-]Kerry4y20

All very interesting, thank you for writing this up. Don't know enough to evaluate this, but it sounds plausible, and not very encouraging. Vaccines do not look promising, but perhaps further understanding of the disease will lead to other treatments that head off some of these complications.

Reply
4Spiracular4y
Vaccines are still our best shot in the long-term. I wouldn't phrase it as "vaccines do not look promising," but more as "SARS is relatively hard to vaccinate well." I do think we'll have a vaccine that works reliably, eventually. No other antiviral method has their price-to-effectiveness ratio. We were able to find fixes to the problems with some SARS-1vaccines, and I think we'll be able to route around these problems for SARS-2 as well. This just means that I don't expect vaccine development to be quite as fast as it would be for viruses without these known problems. Additionally, I suspect animal-testing could be crucial to the development of a safe vaccine, unless we're willing to risk a few human lives in their stead (which, maybe we are). And speaking personally, until the clinical trial results are in, I'm inclined to be cautious about taking vaccines that use large swathes of the viral S-protein, although I suspect some with smaller fragments will turn out to be fine.
2Kerry4y
Very delayed response, sorry. I suspect that by the time we have a vaccine ready to go on a mass scale, it won't make a huge difference. People will return to life before then, for the most part. Not sure if the most vulnerable are able to get vaccines or if that is dangerous--if they can, it will make a difference for them. I don't think it will eradicate the disease because not everyone will choose to get it (especially as it seems dangerous side effects could be a thing with this vaccine, due to the autoimmune response, and being comfortable about this will take years, and it isn't clear how dangerous it is for most people), it would be a huge and imperfect effort even if we mandated it, and presumably the disease will change over time, requiring new vaccinations. So I don't think a vaccine is going to be what changes things here. Obviously, it is still playing out, and the data about risks that comes out regarding both the disease and the vaccine, along with other practical issues, will affect the final outcome.
3Spiracular4y
I did specify long-term, which for me meant time-frames of around a year to a decade out. Honestly, I suspect you're largely right about the short-term. Well, except I might be more optimistic about vaccination efforts. Effective vaccination pushes in the past give me some hope. Also, the mutation rate is a good bit lower than the seasonal flu. SARS-CoV-2's point-mutations per year is around 28 substitutions, which is about 1/2 as many as the flu. Or around 1/3 the rate, at ~1.1e-3 subs per site per year, compared to flu's 3.3 subs per site per year. (Different genome lengths, hence the different answers.)
1CraigMichael3y
What would be the difference between a large swathe and smaller fragments here?

Spiracular

Mar 14, 2020

120

Update 2020/04/02 : See second answer

V1.3

Here's a pass at my answers at present (2020/03/15)

  1. Is this virus doing this in-vivo? What induces it?
    • At the moment, I'm leaning towards yes
    • Some other coronaviruses absolutely have it (ex: FIP, MERS)
    • in-vitro, they could get it to happen using SARS-1 vaccines containing S-protein
    • There are some white blood cell (WBC) anomalies, but what people are noticing is mostly T-cell reductions, a variety that shouldn't be directly impacted by ADE (wrong receptors)
      • Possible alternative explanation for some of the white blood cell reductions: HSCs (the stem cells that make blood) or the thymus are being impacted in some way that reduces the rate of immune cell production. Some similar downstream effects, so you would probably have to test distal causes to differentiate this theory from ADE.
  2. What exactly are the consequences of this ADE interaction? Can we pin it down to symptoms?
    • It's not dengue (no hemorrhagic fevers here). So what are the symptoms of this? (Patients who previously had SARS-1 are probably the example to look at here.)
    • One preprint proposed that the current wide variance in severity is already caused by the presence/absence of this reaction with the antibodies an individual produces
    • Are there diagnostic tests that search for an ADE interaction? Are there high-specificity tests for ADE that are easier or faster than RNA-testing for SARS-2 in some cases?
      • Low T-cell counts are common in severe cases, but not specific to this illness and unlikely to be caused by ADE. High inflammation is even less specific.
  3. How does this change things?
    • For vaccines, it probably means the most basic naiive S-protein-heavy vaccine types need to be thoroughly tested in animals and may not work as intended.
    • If this enters circulation as a recurring cold, ADE could leave me deeply concerned about multiple distinct subtypes evolving and existing over time. But I also find the proposal that this already is happening and is what causes variance in illness-severity pretty convincing, and in that case it may add up to the numbers we're already seeing? Or maybe the post-SARS Chinese numbers? I'm torn.
    • Concrete prediction: People who had the SARS-1 or MERS vaccine previously (esp. if vs. S-protein, which most include) will tend to get a more severe case with SARS-2.
  4. What can we do about it?
    • Vaccines <del>targeting just N-protein</del> (ETA: N-protein vaccines seem to still have in-vivo second-exposure issues) (or another viral protein) or antibody-based therapies might be able to route around this issue.
[-]Spiracular4y20

Here's some of the papers I've looked at, and my interpretation (warning: this is messy).

Some (but not all) other coronaviruses, ex: FIP, have had vaccines that presented with this problem (imperfect antibodies against the vaccine resulted in increased severity of illness compared to baseline).

An in-vitro experiment suggesting that nCOV could use imperfect antibodies as a viable "anchor" for infecting white blood cells. Was tested using previous SARS-1 vaccines.

Interpretation: Assuming it's the same case among SARS subtypes, antibodies against the spike-pr

... (read more)
Reply
1jmh4y
Does Italy fit that line of thought?
1Spiracular4y
I'm generally skeptical of the theory in that paper being provable at all unless they're very careful. There are just so many possible confounders. (I had actually only skimmed it until now. Its theory seems to be that some low-symptom coronavirus cold did the priming? Ugh, given our current viral monitoring habits (bad), that sounds like a hopeless non-starter for study.) I haven't compared the exact death-rates. Off the top of my head, I remember Italy generating the impression that hospital overflow can make a difference that changes 0.5% to 4% in death rates, but I didn't question or dissect it. I don't think Italy has ever seen SARS before.
3Lukas_Gloor4y
About the death rates I feel pretty informed (unlike with ADE), and I think there's nothing that needs to be explained that isn't already likely covered by varying degrees of hospital crowding and (in Italy's case) demographics such as the early outbreak starting primarily in hospitals (as opposed to Germany where it started disproportionally in travellers who visited affected places). Edit: It would be concerning to see unusually high death rates in countries that don't seem to have hospital crowding yet, but I'm not aware of any such examples. The difference between death rates in France and Italy (high) and Germany (very low) are possibly worth looking into, but even that seem to me like it can be explained well through other factors (Germany had really good testing, and again demographics – it just makes a huge difference if the virus ever hits an entire hospital or nursing home or Church with elderly demographic).
1Lukas_Gloor4y
Even though I don't know enough biology to understand what you two are discussing, it sounds pretty concerning! :/ I'm just commenting because I have a suggestion about a possible data point to check (and maybe it's a total non-starter). Singapore doctors seem to have noticed that quick tests for Dengue virus test "positive" (for Dengue) on some Covid-19 patients. The doctors say it's because the viruses are similar also biologically. Maybe looking at what these tests target could be useful here? And personally I'd be curious about more context on the implications of the above hypothesis – whether there's anything you can say about risk factors for a particularly severe disease runout. For instance, is it good or bad if someone had especially many or especially few colds in the last 2 years? Or lifetime? Etc. (Of course, no need to reply, esp. because my questions may not make much sense..)
2Spiracular4y
The idea of "if you've had a wide variety of other colds in a past, maybe you'll get it worse" actually sounds like a fairly good one! But also, there's probably some other reason they are getting more and worse illnesses in the first place? Would be hard to separate out the causes.
2Spiracular4y
Flipped through the article, my guess is that it was a case of "Person X was infected with both (or used to have Dengue), and they tested Dengue first so they dismissed them without even considering that they had both!" (Antibody-tests are usually highly-specific, and Dengue and Coronavirus are pretty different viruses, so I'm inclined to trust that the "Dengue +" result was correct and just how they handled that information was wrong.)

Spiracular

Apr 02, 2020

100

Something weird is going on with white blood cell (WBC) counts, but I'm currently leaning towards believing something else is causing it.

Specifically, they're seeing T-cell lymphopenia & neutrophilia during acute infection (too few lymphocytes, too many neutrophils), along with blood-clot markers (high D-dimer, a thrombosis indicator, presages death).

T-cells seem to be the worst-hit WBC (hyper-activated, decreased numbers), and I'd have expected them to be close to immune to Fc-based ADE once mature; they only express the receptor while young.

This points to some other factor being at play, and leans me against ADE being a major determinant of individual disease severity. I still expect vaccine development to be challenging, and to come with risks of bad reactions if there is not extensive animal testing (since SARS-1 and MERS vaccines repeatedly ran into issues, including for a vaccine against N-protein that shouldn't trigger ADE in-vitro).

A few theories I've seen floating around for the blood results (I'm sure there's more out there):

  • This old (2008) SARS-1 paper posits that glucocorticoid reaction (too much cortisol) could be an upstream cause of the lymphopenia + neutrophilia reaction seen in SARS, RSV, Ebola.
  • Something is causing/reacting to the cytokine storm
    • Both severe COVID19 disease and bad vaccine reactions steer towards Th-2 type immune responses (vaguely allergy-like responses)
    • I haven't deep-dived this.
    • If I want to follow up on this, from a paper-skim it seems to be a Th2-type cytokine reaction. Confused about IL-10 results.
  • Some sort of immune-cell suppression effects, either infecting or indirectly impacting cells at an earlier stage of blood cell development/differentiation.
    • We've found some additional receptor-binding affinities for SARS-CoV-2, so it's not just ACE2-binding. But by far the most productively-infected tissues seem to be the lungs and bowels.
    • I have the impression that it's not productively replicating in white blood cells, but also feel that doesn't totally rule out interference with them.
Moderation Log