Associate Professor in Virology
Sydney, New South Wales
- Tweets 1,651
- Following 264
- Followers 3,820
- Likes 1,137
Images of SARS-CoV-2 cultures often tell us a thousand words. Here we have the Omicron lineage XBB.1.5 growing in two engineered cells. Both equally express ACE2 and TMPRSS2. Visually this sums up our journey into the complicated entry pathway of Omicron lineages......1/
14
158
524
212,155
Evolutionary trajectory of SARS CoV-2 from Ancestral Clade A to JN.1. What does the below mean? The virus is consolidating towards a specific pool of ACE2. Early on the ACE2 pool the virus favoured needed to be cut/cleaved. Now it heavily favours (well for JN.1) uncleave ACE2.
20
122
441
398,586
BA5 has switched tropism again to use TMPRSS2. It’s like a viral yo-yo. Here’s what Delta, BA1 and BA5 look like in cells with TMPRSS2 when exposed to positives samples and left for three days.
21
125
380
In 2021 we made tools to rapidly capture & analyse live SARS CoV2 variants. To rank them for evasion to antibodies we used antibodies pooled from thousands of convalescent/ vaccinated donors. In late November we isolated the first Omicron case in a Australia. Here is its rank
17
116
295
BA5 isn’t quite Delta yet but is on its way (sits in between Delta and BA2 in its ability to engage the Ace2-TMPRSS2 pathway). A p681r change would be very interesting to see if it can “yo yo” fully from BA2 to the tropism of Delta again.
18
111
290
This is what we use to characterise emerging variants. nature.com/articles/s41564-0…
In brief BA1 uses TMPRSS2 poorly, BA2 even more so. BA5 has now switched. A lot of this work we do in real time as the samples build in the community.
5
55
142
1- In 2021 we built a simple/powerful platform to rapidly isolate/test SARS-CoV-2 variants. The full recipe we submitted for review late last year: researchsquare.com/article/r…
Its power is speed/simplicity. Overnight the virus does its work by literally melting the cells together.
6
64
136
2/ To cut a long story very short, the difference in the two cell types and why one looks like a Delta infection (but is XBB.1.5) is the pool of ACE2 that is present in the cell with obvious replication revealed by cell-cell fusion. This pool of ACE2 is resistant TMPRSS2 action
1
17
147
10,344
11/ To really get our heads scratching. Here is the pre-print that maps the biology of this. Stefan Pohlmann's work in 2014 on SARS CoV-1 was key here! biorxiv.org/content/10.1101/…
4
23
149
14,420
Here is another example. It is effectively any culture that replicates the virus at high levels.
4
52
146
3/ Beautiful work by Stefan Polhmann's team in 2014 on SARS CoV-1 lead us to look at the dynamics between ACE2 and TMPRSS2 on SARS CoV-2 entry. pubmed.ncbi.nlm.nih.gov/2422…
1
19
139
9,819
5/ Whilst early SARS CoV-2 entered in a manner like SARS CoV-1... TMPRSS2 cleaved ACE2 facilitated entry only for SARS CoV-2 that was prior to Omicron. What defined Omicron entry, was progressive attenuation in settings where ACE2 was cleaved by TMPRSS2. It was a off switch....
2
15
119
9,076
medrxiv.org/content/10.1101/…
Our snapshot of clinically isolated BA5 virus:
1- Antibodies derived from thousands of donors observes all Omicron variants (esp. BA1, BA2 & BA5) are evasive.
2- Potency of Evusheld and Sotrovimab reduced.
3- Tropism shift back to using TMPRSS2
6
42
110
10/ So how does this explain what Omicrons are up to and how they enter.... A tissue site where the neck of ACE2 can be cleaved by TMPRSS2 (eg. Lung) turns off the ability for Omicrons to latter use TMPRSS2. This is the fundamental change from Delta to Omicron lineages to date.
2
13
119
7,216
9/ The second function of ACE2 is as a chaperone for proteins that transport amino acids. A large pool of ACE2 does this in the small intestine complexed to a protein called B0AT1/SLC6A19. In this setting we would support ACE2 to be TMPRSS2 cleavage resistant.
2
12
111
6,602
Fold ranking of antibody neutralisation drops of the usual suspects in circulation (primary clinical isolates here). Testing is done with approximately 14,000 pooled US donors from April. Details as per figure 6J here: nature.com/articles/s41564-0…
3
38
109
8/ The body has two pools of ACE2 that do a very different function. The first is embedded in the well known angiotensin system. Here the neck of ACE2 needs to be exposed to be regulated by either TMPRSS2 or ADAM17 cleavage. One signals ACE2 to stay, the other to be shed.
1
13
109
6,755
4/ In this work, Stefan's team observed SARS-CoV-1 entry was facilitated by TMPRSS2 cleaving not only the Spike glycoprotein but also the receptor ACE2. When we looked at this for SARS CoV-2..... across all isolates that covered the pandemic, we saw something interesting.
1
8
107
8,661
7/ So what does this mean? It is well established that Omicrons are attenuated in the Lung but do well in other tissues. Why and how would the above observations fit with this?
1
11
104
10,617
6/ For latter TMPRSS2 activation of Spike. Omicron's simply did not like pools of ACE2 that were cleaved.
1
6
93
7,456
Work in progress here:
biorxiv.org/content/10.1101/…
1
19
88
11,948
The switch hypothesis that to a large extent highlights the basic mechanism of this evolution.
2
16
83
20,924
bloomberg.com/news/articles/…
1- Discussions maybe early with respect to an emerging variant with a competitive advantage over the others but they are still important. We still don’t know a lot about how and why the virus targets parts of our bodies and where it is consolidating to.
3
15
69
17,201
nationalgeographic.com/magaz…
Some scientists have names;)…. They even put their name in the gear/mask sometimes.
4
14
62
Spikes are very similar, so you would expect the same for BA4. The question now is if animal models/TMPRSS2 use predicts severity with various degrees of immunity (unvaccinated, convalescent (+\-vaccinated), vaccinated and vaccinated plus convalescent. Lots of variables now.
3
6
73
Here is the TMPRSS2 rankings. Of note, better TMPRSS2 use can dictate better infection across the respiratory tract. Methods and approaches are in the link below. Points are independent experiments with clinical isolates. ncbi.nlm.nih.gov/pmc/article…
2
29
68
2- The primary concern I see now is apathy and lack of discussion. Not everyone is right but discussion and enquiry are still paramount when we need to know more. Knowledge always brings power and it is very clear at the moment that there is no interest in investing in it.
1
11
66
2,785
Turns out it was growing already. Early days, but the cultures look more like pre-Omicron. BA.2.3.20 looks similar. Lots of cell-cell fusion.
10
23
65
More time lapse showing high rates of viral replication. When it gets rolling, it is very destructive. Best way to explain this is cells just “melt”.
3
27
60
6- Using antibodies from thousands of plasma donors with two doses of their vaccine, we ranked Omicron against every variant in 2021. The fold reduction in how this could block the virus compared to earlier variants enable us to see Omicron's potential in vaccinated populations.
5
24
63
5- Then Omicron arrived late November. It also did a great job in melting cells. The virus did cause extensive syncytia (when cells melt together). It did so in a manner that was not dependent on TMPRSS2 though (Nafamostat was no longer working).
4
21
59
Seeing lots of these ("Covid balls" when the virus grows) but waiting on sequencing. Still a lot of BA.5 and lineages thereof (no BQ.1.1). Head to head the parent BA.2.75 is not as evasive or fit as BA.5. Slowly accumulating both convergent sublineage flavours now to test.
5
5
56
8- Whilst ourselves and others will initially report on data for peak responses, duration and maturation of a groups response will take time to document (months).
On a final note, whilst many are moving on from Covid, many scientists still and working day and night to help here.
3
4
50
2,828
Cell sheet collapsing into two large viral syncytial balls. Video is 22 hours total. Images and videos highlight the ability of viral glycoproteins to dissolve cellular membranes. The conditions favour the virus here. It is replicating faster than cells can divide.
2
19
51
7,831
Here is Omicron versus the sum of thousands of pooled donors who are either recovering from infection and/vaccinated (Poly-IgG). On average 16.8 fold drop in titers. This material gives a great snapshot at a population level of how a variant will evade.
3
30
54
1- 2022 was a confusing year for tracking variants. Soup or swarm, but for us each week it was alphabet spaghetti. Rather than drill down on one we followed them as clusters. The BQ clan and the BA.2.75s (of which XBB.1 is part of...well a good recombinant chunk).
3
15
51
17,195
These are now what we call Covid balls. Here we have genetically engineeered a cell that loves to catch SARS CoV2. In the upper left is the day it saw a positive swab. The next panel is the next day- The big balls are infected cells. The bottom panel is 48hrs
11
48
Here is BA2.75 growing. The balls that look like “smarties” are syncytia that form in a TMPRSS2 dependent manner and very similar to what we see with BA5. So it doesn’t need l452r change to do this. Would be interesting to watch what l452r does & p681r
3
14
49
1- @Mike_Honey & @JosetteSchoenma have been working a lot in that space. Mike communicates this regularly. Josette actually flagged our own BR.2 to be designated BR2.1. @BenjMurrell is helping with beautiful figures like this.
4
6
42
8,396
thelancet.com/journals/ebiom…
-> Our work now is published online. A big thanks to the VIIM cohort, ADAPT cohort, CSL Behring and of course everyone who work so hard to get datasets out with new variants appearing weekly at times.
To summarise this work in 2022, see thread below
1
14
46
17,473
They all cause syncytia. VOCs do it better…b117, b1351, P1, b1617.2. The earlier variants also, but take a little longer to get there.
2
15
41
biorxiv.org/content/10.1101/… Efficient TMPRSS2 use is associated with better infection in animal models and in vitro “lung” cellular models.
12
45
9- Please be as respectful as you can to people that contribute in this space. We talk datasets and not drama.
If I can take a line from one of my favourite kids movies... "When you need me but don't want me I must stay".
3
2
45
2,765
Here is BR.2.1 growing. Enters cell more like BA.2 than its relatives from the BA.5 clan… BQ.1.x etc. When the data comes in, we’ll give a snapshot of antibody evasion.
2
11
42
Four cold-causing coronaviruses may provide clues to COVID’s future | Science |…. Interesting debate from both sides. There has been a trajectory for SARS CoV2… now we need to understand if it continues and what it means in vivo. science.org/content/article/…
4
14
40
5,660
The power of merging functional genomics with virology. A great collaboration with the Neely lab. These approaches are also great in determining what emerging viruses may also need and want. @KirbyInstitute @Neely_Lab
sydney.edu.au/news-opinion/n…
2
13
41
6,286
1- Often when you give a virus what it’s needs to enter and strip away the internal restrictions, cell-cell fusion will take place. For this cell clone it does it so well we use it for rapid high content screens. See nature.com/articles/s41564-0…
3
10
42
2- My team isolates hundreds of variants circulating in Australia. We test them in a manner that is real time. One of the first data on primary XXB.1 isolates and neutralisations is in a pre-print that is up online here: medrxiv.org/content/10.1101/…
4
6
34
2,160
….I think we need to consider a “route 3” for entry. And it can form syncytia. It just needs cells that it can grow in well. Cells with TMPRSS2 it clearly no longer prefers. Finding what it likes will determine is change in tropism
1
11
41
7- Pre-print is here. medrxiv.org/content/10.1101/…
Take home message is we need to continue to monitor the mix carefully. The less efficient TMPRSS2 and lower fusogenicity with XXBs is a positive sign. Disease severity in the clinic is another story though.
2
12
36
4,147
And the team is part of a network. We do not work in isolation. We are nothing with diagnostic crews taking and literally testing 100,000 of samples. Then surveillance crews sequencing thousands of whole genomes that enable samples to reach us for viral isolation and study.
1
5
31
1,470
Viral CPE “yo-yo”. If you look closely you will see that are still tethered to each other by membrane “string”.
3
8
36
1- The conversation here has many complexities and I will do my best to cover this in a thread.
Many of us globally are studying antibody responses that block the virus (neutralising antibodies). They come after vaccination and/or infection. Generous participants help here.
2
8
33
3,824
4- When we used the platform on positive swabs, we could also detect viral infectivity at very low levels and also rank viral fitness. In short, viral loads do correlate very well with viral infectivity if the sample is taken at the acute stages of infection.
1
8
38
Here is BQ.1.2. The big hole is generated as it is more efficient in entering this cell type. The efficiency here is primarily generated by a protein called TMPRSS2. At the moment the variants are converging to evade antibodies but at times they diverge in how they enter cells.
1
11
30
6- what’s the take home message? Cell-cell fusion here is a sensor. It’s simply telling us it is using entry factors like TMPRSS2 more efficiently. Viruses generally do this to be fitter and this can indeed be related to a shift in tropism (better entry into other cell types).
2
5
29
See details of the supplementary movie 1 here. In brief, cells are being infected so fast that the outside of the cells become like a virion and after that they fuse/melt together. Movie is around 3 min per frame. Total duration is around 20 hours.
nature.com/articles/s41564-0…
1
15
34
XBB and BR2.1 less than BA5s. That is clear. BQ1.xs are a bit of a mixed bunch. BQ.1.2 is actually higher than BA5. The others lower so far but we are keeping an eye on those with wthe 666 Spike polymorphism (that literal the spot) that BQ.1.2 has.
3
8
32
1- Mechanistically it goes towards explaining why Omicrons are attenuated in tissue settings like the lung. Early on SARS CoV-2 entered cells in a manner similar to SARS CoV-1. With the Omicron switch, it simply means we need to consider the pathogenic consequences in other ways.
1
2
38
1,795
2- Whilst the "melting" reaction looks catastrophic, we did see order in this reaction. By simply adding a dye that detected cell nuclei on the day the cells saw the virus, we could rapidly quantify the level of viral destruction:
2
6
33
5- One pool of viruses looks more aggressive in growth than BA.5. Another pool looks more Like BA.2. Early genomic linkage supports BA.5 sublineages as the former and BA.2.75 as the latter. BA.5 sublineages love TMPRSS2 & BA.2.75s not so much.
2
6
34
Sotrovimab is still active. It didn’t do so well with BQ.1.1 but with XBB.1 and XBB.1.5 it is still retains good activity. Evusheld unfortunately is gone with these emerging variants.
2
13
30
4,055
‘Immunological unicorn’: the Australian lab growing coronavirus – and its startling discovery theguardian.com/australia-ne…
5
9
27
At the end of this movie, you will see a cell escapes from the syncytial “balloons”. This will give you an idea of how big the structures are relative to a single cell.
1
8
29
SARS CoV2 is an enveloped virus and albeit very infectious per particle has not changed in terms of its stablility in the environment. Variants are more “fit” due to subtle changes in the Spike. This has increased infectivity to particle ratios….what is that then?
1
9
31
Thanks Mike. Our studies are in the lab and this work importantly relates to efficacy in the real world. Here they conclude: “The overall CVE for COVID-19 mortality for the bivalent BA.4-5 vaccine was 77.8% and for the BA.1 booster vaccine was 80.1%. “.
4
2
30
4,787
We are looking into it now. It’s based primarily on donations and the generous people who volunteer to give blood for us to look at this. Bulk of donations are obviously Pfizer and Moderna at the moment.
2
2
28
2,732
3- As it is simple and "dirt cheap", this enabled testing at scale. Add serum or therapeutic "X" or "Y" and you could test hundreds of samples in a day and get results overnight:
1
1
27
Thanks for your support. The heavy lifting is done by diagnostic units and surveillance units at NSW Health and Units like Sydpath in our state. For tracking variants testing is changing here and globally. Surveillance needs support in this setting. Without it we fly blind.
1
3
25
2,205
Movie from today’s talk. A “Beta” infection overnight. Each ball is a cell. As the cell rapidly becomes infected, the Spike from the virus turns up on the cell surface. Now with the outside of the cell/balls looking like a virus, what we see is a melting/fusion reaction.
2
12
27
Omicron ranking against all variants of concern in convalescent plus convalescent plus vaccinated. Tracking at 17 to 22 fold reduction. Two dose vaccine (AZ or Pfizer) not reaching titers in donors tested so far. Will post what we see with large polyclonal IgGs batches next
1
12
21
1- Closing off of the 2021 variant chapter from the team's perspective. We only had so many hands and we needed to find simple and cost-effective solutions where we could scale
microbiologycommunity.nature… #behindthepaper @StuartTurville @KirbyInstitute @UNSWMedicine @UNSW
1
11
25
1. Below is a Wall Street Journal article really covering an interesting aspect of science communication. Whilst Iw as briefly quoted, here is the context as a scientist working in this particular space and how it actually happens in the real world.....
wsj.com/health/covid-variant…
2
4
27
5,352
7- So whats the take home message. All vaccines that are presently available will give a response that will generate antibodies even towards the most evasive variants we know. Over time we will learn if BA.5 based vaccines maybe better than others in Australian cohorts.
3
6
23
940
These are early warning systems that look at viral “potential”. This is looking at changes in viral entry, antibody binding and the ability for the virus to navigate innate viral restrictions. How this plays out with respect to disease severity in the clinic is the next step.
2
3
26
There are lots of variables in what determines waste water viral loads. And yes often the most simple answers are the ones that are right. But the question of what it is up to in the gut are still very important and something that needs to be resolved at many levels.
1
24
847
4- TMPRSS2 use and fusogenicity of the virus when this protein was present was a continuum. BQs at one end and BA.2.75s at the other. Initially BQ.1.2 was our biggest concern as the fusogenicty was high an above that of BA.5 (see image in panel C: BQ.1.2 vs D: BR.2.1).
1
5
20
4,106
2- So what are we seeing now during the Omicron waves. In primary swabs BA.1 and BA.2 cell- cell fusion was poor. So they were either attenuated or they like something else at the membrane. Looking at primary cultures it wasn’t attenuation but a tropism shift.
1
3
25
4- So what are we seeing post BA.2. In this clonal line you can see overnight the difference between a BA.2 and BA.5. It’s night and day see Fig. 2 below. Now we are seeing two phenotypes pop up in this convergence of variants. thelancet.com/journals/ebiom…
1
8
24
It’s a weird one. Everyone talks convergence of variants, but this is one that stands alone in doing something different in entering cells. Defined by spike I666V, it’s awfully close to furin cleavage site. I joked it was a 666 polymorphism, but it is defined as such.
4
24
6- Now it appears we are in the XB era. XBB.1, XBB.1.5 and in Australia XBF. From what we have learnt from the parents and the relatives, the replication will be similar to BA.2. Fusogenicity and efficient TMPRSS2 use on the lower side.
3
2
18
1,902
@JosetteSchoenma has looked at this very closely too. The BA.2.75 is a BR.2 with a unique change in the ORF8 gene. It’s now designated/called BR.2.1 thanks to her hard work in tracking it. It’s interesting as it’s dominating in amongst a real mix/swarm of variants.
2
4
21
5- But how the virus transmits is another manner. What we saw was that high fusogenicity didn't track with prevalence in the community. The winners were all derived from BA.2.75 or recombinants thereof.
1
5
20
3,116
…..”when we are told it’s over but someone forgot to tell the virus”. Yes we need to think this through. Thank you Peter.
3
21
We have only looked at JN.1. In that setting it is heavily attenuated for TMPRSS2 use when WT ACE2 is co-expressed. Sigal lab has observed a similar observation. It ramps up TMPRSS2 use when cleavage resistant ACE2 is expressed. It’s very fit in that setting.
1
3
22
3,444
We have. Sits low in neutralisation assays alongside XBB.1.22.1 and those with F456L. Not a big shift from XBB.1.5 though but consistent with Ben’s ranks. Tropism is in line with XBBs in general.
3
2
21
11,247
To the mitsubishi driver heading north on the pacific highway…. You’ve summed up the feeling of the planet at the moment
3
18
It’s pretty elusive here in Australia at the moment. That and the changes in testing have led to our surveillance efforts becoming more of a trickle now. Happy to collaborate with teams, if they’d be willing to ship these to us. Assays to look at it closely are very fast.
21
463
2- Pooled IgGs from over 420,000 donors from the US taught us a lot about the emerging variants and to a certain extent a trajectory of antibody based immunity. In the spaghetti, the variants clustered. There were layers of evasiveness. BQs/XXB/XBF/BR2.1 won the day.
1
4
19
1,422
3. From that you can calculate infectivity to particle ratios. So that tells you the minimum number of particles that can start an infection. Right now, fitter variants like Delta can start infections with smaller viral particle numbers then the earlier variants.
2
7
22
Important question & a tough one to answer. I’ll post this figure and walk people through it. It’s very important but convoluted at many levels.
1. Firstly, this is the US. Each point is a around 30,000 donor antibodies pooled. Each color is a different variant.
1
5
16
5,214