The Extinct SARS-CoV-2's in Deer Study

Alpha and Gamma variants detected in deer months after going "extinct" in human transmission.

A new pre-print helpfully shines light on SARS-CoV-2 transmission in an animal population, demonstrating that variants no longer apparent in human infection might persist in deer, at least for several months. Does this mean that the Omicron siblings came from wild animals (no), or that Delta might not be dead after all (yes)?

A minor note: Substack restored a pre-publish draft of this post a few hours after posting. Hopefully the current, correct version stays online this time.

Golden Oldies

The paper:¹

Caserta, LC. et al.

The punchline:

These findings indicate that [white-tailed deer] – the most abundant large mammal in North America – may serve as a reservoir for variant SARS-CoV-2 strains that no longer circulate in the human population.

That sounds… great…

If you anticipate deriving celestial enlightenment and subsequent detachment from all material concerns because of this post, please drop a tip now.

click button or visit https://ko-fi.com/unglossed

Background

Previous findings in a population of deer, in Ohio, showed Alpha circulating in late winter and spring of 2021.² So: Not the same as finding extinct variants in autumn of the same year, as this new paper does.

The Set-up

A total of 5,462 retropharyngeal lymph node (RPLN) samples collected from free-ranging hunter-harvested [white-tailed deer] during the hunting seasons of 2020 […] and 2021 […] were tested by SARS-CoV-2 real-time RT-PCR. SARS-CoV-2 RNA was detected in 17 samples (0.6%) from Season 1 and in 583 (21.1%) samples from Season 2.

The disparity in positivity rates might seem to be driven by background case rates in New York, where the “Wuhan” model was still “burned out” in 2020 until the winter. In summer and autumn of 2021, the Northeast, including New York, had more human cases:

The Findings

On the other hand, human case rates in the summer might not be principally relevant compared to case rates in late spring, since it seems like crossovers from that era were still transmitting in deer. This is in stark contrast to 2020, which featured an earlier end to the spring wave in New York, and resulted in very few positive PCRs in hunting season:

Caserta, LC. et al. (Fig. 1) (heavily annotated) My impression is that “September” only includes the last few days of the month.

Naturally, the collection of 69 Delta sequences between September and December, 2021, is not surprising — it is consistent with fresh human-to-deer transmission of a virus that white-tailed deer are known to be susceptible to. To collect 27 Gamma and 54 Alpha sequences in the same time-frame, however, suggests that these vanquished human variants were still circulating in this animal reservoir:

Caserta, LC. et al. (Fig. 3) (annotated)

Sequences also appear to show an accelerated rate of mutation in deer — not surprising as the virus would be expected to adapt to its new host. Here, I have added “eye-ball-software”-based mutation rates that correct for the authors’ software’s non-assumption of multiple cross-overs:

Caserta, LC. et al. Supplementary Fig. 1. Dashed orange lines added using “eye-ball-software.” Note that vertical scale is different for each image, so visual comparisons between panels are not accurate. Once again there are discrepancies in the year-position for Alpha and Delta; these appear to be data entry errors. The authors may update these images to correct some of the dots.

The difficulty in trying to judge mutation rates this way is that multiple, sequential cross-overs from humans — where the viral genome is more static — will cause the rate of mutations in deer to appear lower than it really is. And so for gamma, presumably the least recently-connected to any human cross-over, the rate appears highest (however, this might be an illusion caused by coincidentally finding the most mutated samples after the least, if transmission in deer had indeed been going on since spring).

For both Alpha and Gamma, at any rate, 2021 season deer had some impressively-mutated sequences, the kind that WHO declarations of “concern” are made of, though not out of proportion to previous sequences from North American deer (see Fig. 5 “wheel” below).

“Footprint” vs. Active Transmission?

Contra the authors, SARS-CoV-2 RNA positivity in lymph nodes might not correlate with active infection or transmission. The expected “baseline” duration of RNA viruses in either human or animal lymph nodes is surprisingly poorly-studied. As discussed in “The 60 Day RNA Mystery,” it may be the case that entire viral particles are frequently preserved in germinal centers for long periods after acute infection.

The 60 Day RNA Mystery: Spoiler / Summary

So, it could be the case that by using lymph nodes as the source material, the authors are revealing a “reservoir” that is even less poorly illuminated than they thought — and one that might not correspond to transmissible virus.

Notably, of 216 samples that were PCR positive at less than 30 cycles, only 7 were successfully cell-culture-recovered in Vero-E6 TMPRRSS2 cells; and the corresponding sequences were all for Delta variants.³

Caserta, LC. et al. (Fig.1) (annotated)

So, perhaps the reason spring-era variants (and even a few vintage B.1’s from before spring) were detected, is that RNA from spring-era deer infections was still lingering in lymph node germinal centers?

Arguing against this interpretation, however, is that the spring variants were detected in geographic and temporally related clusters. For example, sub-30-CT positives which were later sequenced as Alpha and Gamma appeared to “disperse” actively throughout nearby deer, based on collection times — suggesting active transmission, rather than the “footprint” of a spring wave.

Caserta, LC. et al. (Fig. 7) (annotated)

While it’s not clear whether or how the authors ruled out that human hunting and collection patterns were actually driving these trends, it’s at least tentative evidence in the “transmission” column (also, sequences appeared truly “sequential,” per the text, revealing “an intricate dispersal pathway with clear connections between the sequences”).

Further, Gamma would be expected to appear more frequently outside of the cluster if long-term retention of non-transmitting viral RNA was common; it is limited to the Allegany area.⁴

Lastly, extreme mutational distance from any recorded human sequences, and narrowness of ancestry (monophylogeny relative to human sequences), is highly consistent with a single spring human-to-deer crossover that then spent a long time bouncing back and forth in deer.

As the authors put it,

All Gamma WTD sequences detected here formed a monophyletic group highly divergent from the human derived Gamma lineage sequences (Fig. 5A and C).

Caserta, LC. et al. Fig. 5. Distance from center is number of mutations from Wuhan reference genome. Black dots are deer-derived sequences from other authors. Notably, few other sources appear to have uploaded deer-derived Alpha or Gamma lineage sequences. This may once again support the notion that the Canada-Atlantic spring, 2021 wave was instrumental in creating the temporal transmission patterns observed here.

In fact, such a genetic signature would argue that long-term transmission in a zoonotic reservoir, something on the order of a half-year at least, was necessary to produce the sequences collected. So, count me on team “long term transmission reservoir” for now.⁵

Implications for Past and Future

Does Caserta, et al. vindicate the “Zoonotic Omicron Origin” theory offered to the press by experts last year? Yes and no. It is certainly a proof-of-concept that an early B.1, from spring 2020, could have transmitted for an extended time in wild animals, featuring an elevated mutation rate for some or all of that period. In fact, 4 vintage-B.1-derived genomes are found among the 154 sequences that pass quality check in this study’s autumn, 2021 samples — though they only have on the order of 20 to 50 mutations from the Wuhan reference.⁶

However, the problem of the simultaneous discovery of the B.1 and B.2 “siblings” still remains; and suggests a lab origin as the most likely explanation for how two such distantly-related strains, with apparent divergent evolutionary pressures, would emerge in the same location.

Omicron Origins: Summary / Spoiler

On the other hand, does this mean that not-so-mild Delta will one day soon be the subject of “I thought you were dead / It didn’t take” memes the world over?

In light of the analysis of the gamma cluster above, I think it is at least demonstrably possible.⁷

If you derived value from this post, please drop a few coins in your fact-barista’s tip jar.

click button or visit https://ko-fi.com/unglossed

1. Caserta, LC. et al. “White-tailed deer (Odocoileus virginianus) may serve as a wildlife reservoir for nearly extinct SARS-CoV-2 variants of concern.”

2. Hale, VL. et al. (2021.) “SARS-CoV-2 infection in free-ranging white-tailed deer.” Nature. 2022 Feb;602(7897):481-486.

This paper uses nasal swabs instead of lymph nodes.

3. This could also simply be due to the impossibility of physically separating virus-containing portions of lymph nodes from the immune cells going about their work. That viruses were in germinal centers at all, per my devil’s-advocate theory, is pure assumption. They may have been anywhere, such as the subcapsular sinus or in antigen presenting cells, due to fresh delivery from infection zones. This is in contrast to trying to culture virus from tissues where it is truly actively infectious.

4. Supplementary table 8 reports two non-cluster Gamma sequences, one of which is in Allegany and the other in next-door Steuben county.

5. One alternate possibility is a blend of lab-leak followed by in-deer transmission. As discussed in footnote 7, it is certainly likely that human labs will continue to play with old SARS-CoV-2 variants for some time to come. This could lead not just to reemergence in humans, but in animals with shorter lifespans and hence less immune memory. Such a model allows for a delayed introduction of gamma into the “cluster” population. However, once again, the extreme mutational distance from any known human sequences suggests that sustained in-deer transmission did occur.

6. From Fig. 4. However, it is difficult to tell which of the B.1 sequences are from 2021 and which from 2020; but none exceed 50 (the lavender dots adjacent to the alpha tree).

As before, black dots are deer-derived sequences from other authors. Colored dots are from this study. Two B.1.1 samples (lavender) are from 2020 and one from 2021, per Supplementary table 8, so it does not seem that any B.1.x’s from 2021 are well-spread-in-deer, highly mutated 2020 cross-overs, but rather lingering in-human strains that crossed over in spring 2021.

7. Note: Reemergence of Delta from the negaverse was predicted here in January.

Author has the thing post

Of course, it’s arguably just as likely that Delta or another pre-Omicron variant leaks from a lab studying it (by accident this time). So, the determinative factor might be whether humanity has built up enough immune debt for the original generation of the virus. For H1N1 flu, which was kept alive in labs around the world after its disappearance in 1957, lab escape back into human circulation took 20 years. See “1918, I Love You.”

own work

Since circulation in deer adds the element of active mutation, however, it might be expected to more quickly produce a (true) zoonotic re-emergence.