Preprint Restrained memory CD8+ T cell responses favors viral persistence and elevated IgG responses in patients with severe Long COVID., 2024, Rodriguez et al

[SNT Gatchaman]

Restrained memory CD8+ T cell responses favors viral persistence and elevated IgG responses in patients with severe Long COVID.
Lucie Rodriguez; Ziyang Tan; Lakshmi Kanth Tadepally; Jun Wang; Hugo Barcenilla; Zoe Swank; Fanglei Zuo; Hassan Abolhassani; Ana Jimena Pavlovitch-Bedzyk; Chunlin Wang; Laura Gonzalez; Constantin Habimana Mugabo; Anette Johnsson; Yang Chen; Anna James; Jaromir Mikes; Linn Kleberg; Christopher Sundling; Mikael Bjornson; Malin Nygren-Bonnier; Marcus Stahlberg; MIchael Runold; Sofia Bjorkander; Erik Melen; Isabelle Meyts; Johan Van Weyenbergh; Qiang Pan Hammarstrom; Mark M Davis; David R. Walt; Nils Landegren; COVID Human Genetic Effort; Alessandro Aiuti; Giorgio Casari; Jean-Laurent Casanova; MARC JAMOULLE; Judith Bruchfeld; Petter Brodin

During the COVID-19 pandemic it was widely described that certain individuals infected by SARS-CoV-2 experience persistent disease signs and symptoms, Long COVID, which in some cases is very severe with life changing consequences.

To maximize our chances of identifying the underpinnings of this illness, we have focused on 121 of the most severe cases from >1000 patients screened in specialized clinics in Sweden and Belgium. We restricted this study to subjects with objective measures of organ damage or dysfunction, >3 months following a verified, but mild- to-moderate SARS-CoV-2 infection. By performing systems-level immunological testing and comparisons to controls fully convalescent following a similar mild/moderate COVID-19 episode, we identify elevated serological responses to SARS-CoV-2 in severe Long COVID suggestive of chronic antigen stimulation.

Persistent viral reservoirs have been proposed in Long COVID and using multiple orthogonal methods for detection of SARS-CoV-2 RNA and protein in plasma we identify a subset of patients with detectable antigens, but with minimal overlap across assays, and no correlation to symptoms or immune measurements. Elevated serologic responses to SARS-CoV-2 on the other hand were inversely correlated with clonally expanded memory CD8+ T cells, indicating that restrained clonal expansion enables viral persistence, chronic antigen exposure and elevated IgG responses, even if antigen-detection in blood is not universally possible.


Link | PDF (Preprint: MedRxiv)

 

[EndME]

Cohort selection

• 121 LC patients (selected out of >1,000 subjects at at Long COVID clinics in Leuven, Belgium & Karolinska University Hospital in Stockholm, Sweden)
• mild to moderate acute infection (non-hospitalised and verified in some way) at least 3 months ago, average age=48 (min-max, 14-72), 87% females
• The average sickness duration is given in Fig. 1f and appears to have a cluster around 750 days. It's hard to infer too much directly from the graph, but all the swedish LC patients (74%) have already passed the 500 day mark.
• Different to almost all other studies they focused on patients that had "objective measures" of Long-Covid. That means at least one of the following results yielded differences to the average population microvascular dysfunction shown by magnetic resonance imaging (MRI) of the heart, endothelial dysfunction by pulsatile arterial tonometry (EndoPAT), autonomic dysfunction and postural orthostatic tachycardia syndrome (POTS), hyperventilation, pulmonary air trapping or reduced carbon monoxide diffusion capacity and other respiratory abnormalities which can be objectively measured by computer tomography (their graphs also include things such as Neuro symptoms and cognitive impairment).
  • It's unclear how objective these differences are (EndoPAT doesn't seem very objective for instance), how many differences there were on average, how these deviated from the average population etc. Going by the available graphs there seem to be patients without differences in the things they listed. However, I very much appreciate the effort they took here and would like to know how some of these things were assessed (for example cognitive impairment) and the values that were obtained. One might wander whether this will not yield a very heterogeneous study population (even though they "do not cluster in relation to organs affected or immunological states").
• Other symptoms apart from the "objective measures" are not part of this pre-print (for example PEM). They focused on the "most severe cases" in their screening, but it is unclear to me what that means. Is there an association to QoL, disability, working hours etc?

I would think some of the questions above will be addressed in follow-up research, especially given the long author list. I'd be surprised if Johan van Weyenbergh wouldn't use trancriptomics on this cohort as well (as done here).

Overall a very long paper, that seems to produce a lot of negative results for many things related to different viral persistence hypotheses.

 

[SNT Gatchaman]

Though they offer indirect evidence for viral persistence, with direct evidence in plasma in only 10%, suggesting possibly viral persistence elsewhere such as marrow or gut.

results imply that either the assays used for detection of persistent antigens in plasma are not sensitive enough to detect all plasma antigens present, or that viral reservoirs confined to tissues might not leak antigens into plasma, or that viral persistence is not a universal feature in patients with Long COVID although this latter point is at odds with the persistently elevated IgG responses to SARS-CoV-2 seen in our cohorts and in other

In summary they say —

Direct analyses of SARS-CoV-2 RNA and protein antigens in plasma using multiple orthogonal methods reveal subsets of antigen positive individuals, but minimal overlap between assays used, suggesting that viral persistence is not uniformly detectable in plasma and that elevated serological responses is a more sensitive immunological marker of Long COVID. [...] The frequency of clonally expanded SARS-CoV-2 specific memory CD8+ T cells on the other hand are inversely correlated with elevated IgG responses in Long COVID, which add additional indirect support for the hypothesis of viral persistence, and a possible mechanism thereof in the form of restrained, rather than exhausted cytotoxic CD8 + T cell responses in patients with severe Long COVID.

Since standard ELISA assays are not as sensitive as single-molecule array (SIMOA) assays, also for detecting SARS-CoV-2 specific IgGs, we performed SIMOA assays and compared absolute IgG antibody concentrations (pg/ml) in relation to the time from first known SARS-CoV-2 infection (Fig. 1f). Results showed an even more pronounced difference between patients with severe Long COVID and convalescent controls

these Long COVID cases were sampled long after their acute COVID-19, after which IgG responses are expected to decline with time, but instead showed the opposite

the N antigen was detectable in 10/100 Long COVID cases, but only 1/50 convalescent controls (Fig. 2c). [...] These findings imply that persistent SARS-CoV-2 antigens are detectable in a subset of ~10% of patients with severe Long COVID.

The presence of two or more detectable antigens was only found in Long COVID patients, representing about 10% of the cohort. There was no obvious link between detectable antigen levels in plasma and symptom groups.

Expansion of CD33-expressing monocytes over CD33negative monocytes in patients with Long COVID and elevated antibody responses (Fig. 3a-b), indicates an attempt at dampening a persistent innate immune response to persistent antigen.

The most highly correlated proteins were circulating CD40, PD-L1, TRAIL and TNFa supporting an ongoing immune response

Autoantibodies to type-I IFN have been associated with life-threatening COVID-19 pneumonia due to impaired IFN-I-mediated inhibition of viral replication. Such autoantibodies increase in frequency with age, are more common in males than females for unknown reason [...] We investigated levels of autoantibodies to four different IFNa subtypes and IL1RN but found no evidence of these autoantibodies in patients with Long COVID

The fraction of T cell clusters identified through GLIPH2 which mapped to likely SARS-CoV-2 specificity was 604 in Long COVID and 232 in convalescent controls. This expansion of T cells specific for SARS-CoV-2 in Long COVID is in line with elevated IgG responses and further suggests antigen persistence.

an inverse relationship was found between the expansion of such cells and plasma anti-SARS-CoV-2 spike IgG levels (Fig. 4e). This result suggests that individuals who fail to mount a clonally expanded memory CD8+ T cell response to SARS-CoV-2 develop a viral reservoir with persistent antigen that drive up an elevated anti-SARS-CoV-2 spike IgG response with time. This contrast with the normal tapering of IgG titers over time following infections in convalescent individuals.

Patients with higher anti-spike IgG levels and possible viral persistence had a higher expression of KIR3DL1 (Fig. 4f), an NK cell receptor recently shown to be a marker of regulatory CD8 + T cells. The function of these cells is to kill CD4+ T cells and limit autoimmune responses following infections

Our results imply that elevated SARS-CoV-2-specific IgG responses to spike (RBD), as measured by sensitive SIMOA assays, is a sensitive marker of severe Long COVID, and a likely result from chronic antigen stimulation and viral persistence. This proxy measure correlates with both innate inflammatory responses, and a restrained, unexpanded memory CD8+ T cell response towards SARS-CoV-2, as characterized by upregulation of negative regulators (KIR3DL1) and reduced expression of cytotoxic molecules (GZMB, GZMH), but is not associated with exhausted phenotypes.

A strong initial adaptive response might increase the chance of viral clearance and reduce the risk of Long COVID, while a sustained and elevated long-term response to SARS CoV-2 with elevated titers occur once a viral reservoir has been established leading to chronic antigen stimulation.

 

[Amw66]

I have no biology background so this may be non sensical / irrelevant. Apologies in advance

How does this compare to HIV?

 

[SNT Gatchaman]

10 min talk by Petter Brodin on this preprint at the Polybio Spring 2024 Symposium.