Preprint Blood transcriptomics reveal persistent SARS-CoV-2 RNA and candidate biomarkers in Long COVID patients, 2024, Menezes et al.

Now published as a preprint. See post #7

This is an abstract from the Demystifying Long COVID International Conference 2023, converted from an image on Twitter, so it might be slightly wonky, but it looks intriguing.

Blood Transcriptomics Reveal Persistent SARS-COV-2 RNA And Candidate Clinical Biomarkers in A Belgian Long COVID Cohort
Menezes S, Jamoulle M, Carletto M, Van Holm B, Moens L, Meyts I, Maes P, Van Weyenbergh

Background: With millions of people currently suffering from Long COVID (LC), validated therapeutic options and biomarkers are direly needed to guide clinical management. In this study, we propose whole blood transcriptomics to identify non-invasive candidate biomarkers for viral persistence, disease severity and clinical evolution in a general practice-based cohort.

Materials and Methods: A cohort of LC patients (diagnosis according to WHO criteria) was followed up for 1-30 months after acute COVID. Complete clinical history and follow-up was obtained at a single general practice (MJ, Charleroi, Belgium) using electronic health records and several established clinical scales (DUSOI Duke Severity Overall Index, COOP Dartmouth Coop chart initiative) were used to quantify patient evolution. Severe neurocognitive deficits were confirmed by brain scintigraphy (SPECT), with 25/48 patients (52%) displaying vascular defects.

Whole blood samples were obtained from 48 LC patients and 12 controls (matched for age, sex, time since acute COVID-19, vaccination status and comorbidities) and analyzed by digital transcriptomic analysis (nCounter, Nanostring) to quantify a total of 800 RNAS (SARSCOV2 and innate/adaptive immunity), as previously established for critical COVID-19 (Menezes et al., Lancet Microbe 2021). Neutralizing antibodies against 10 different SARSCOV2 variants. were quantified using a sensitive electrochemoluminescent assay (MSD). Statistical analysis included non-parametric Mann- Whitney test, Spearman correlation and multivariable logistic regression.
Results: Digital transcriptomic analysis revealed a total of 212 differentially expressed genes between LC patients and matched controls.

Among 120 transcripts significantly increased in LC were several viral RNAs: Nucleocapsid, ORF7A, ORF3A, Mpro (target of Paxlovid) and antisense RNA, the latter suggesting ongoing viral replication, while Spike RNA was remarkably low/absent. In addition, several SARSCOV2-realted host genes were also increased in LC (ACE2/TMPRSS2 receptors and DPP4/FURIN proteases). Other upregulated RNAs were specific for memory B cells (CD27/IGHE/BMP8A), and platelets (CD99/PBX1/PDZK1IP1). Platelet transcripts were also positively corelated to viral load (p<10-7), providing a mechanistic link to the hypercoagulative state previously demonstrated in LC. Summarizing these 800 transcripts into biological pathways, we found significantly decreased TLR signaling (p=0.0039), lymphocyte activation (p=0.016) and immunometabolism (p=0.023) in LC patients. Moreover, immunometabolism was negatively correlated with blood viral load (R=-0.56, p<0.0001), suggesting an "exhausted" immune status in LC due to ongoing viral replication.

Using multivariable regression, we found that age and sex were not associated with "low" vs. "high" viral RNA status, whereas the number of comorbidities (1.61 95% CI [1.14-2.49], p=0.014) and the number of COVID vaccine doses (0.36 95% were CI [0.14-0.79], p=0.018) independent predictors of "low" vs. "high" status, confirming a protective role of vaccination. SPECT-positive patients could be discriminated from SPECT- negative by increased RNA levels of insulin receptor (INSR) and platelet P-selectin (SELP), further incriminating perturbated platelet
and (immuno)metabolism activation/coagulation in disease severity.

Conclusions: We used digital transcriptomics to identify non-invasive (blood) biomarkers for viral persistence, disease severity and clinical evolution, guided by SPECT imaging. Overall, increased platelet RNAs and decreased immunometabolism are significantly correlated to viral load, providing mechanistic links as well as therapeutic targets to tackle Long COVID.
Demystifying Long COVID International Conference 2023

 

A recording of the talk, as well as of every other talk can also be found on YouTube www.youtube.com/watch?v=gEsjR2y6fzs. When such "phenomenal" and very interesting preliminary results, which do significantly differ from previous results, are presented by a new team from a "smaller" lab I find it impossible to say whether it's mostly noise or indeed a groundbreaking finding. It certainly means that I'll be keeping an eye out for the paper.

 

Thanks @EndME! Here's a link to the start of this specific talk.

 

I've made a thread for this conference: day 1 here.

 

A small group of patients were followed before and after first and second line treatments. GP-based "real world" cohort. Observational not an interventional trial.

First line treatment is targeting platelets and the microcirculation: aspirin + clopidogrel + piracetam
Second line treatment is targeting viral persistence: paxlovid (off-label)

Transcriptomics (nanoString): viral RNA

The patients having paxlovid had samples at day 0 and day 15 of treatment. In the Q&A he said the patient reported outcomes showed symptom return following ceasing paxlovid.

In LC the volcano plot showed upregulation of viral, platelet and B cell (noting IgE heavy chain) transcripts.

Reducing the plot to a simplified diagram shows heterogeneity in patients. Eg for "immunometabolism" there is a homogeneous response in HCs but heterogenous in LC with both up and down on the pathway scores.

 

However that heterogeneity in the immunometabolism score correlated with viral RNA load.

 

Merged thread

Blood transcriptomics reveal persistent SARS-CoV-2 RNA and candidate biomarkers in Long COVID patients
Soraya Maria MENEZES; MARC JAMOULLE; Maria P Carletto; Bram Van Holm; Leen Moens; Isabelle Meyts; Piet Maes; Johan Van Weyenbergh

With an estimated 65 million individuals suffering from Long COVID, validated therapeutic strategies as well as non-invasive biomarkers are direly needed to guide clinical management.

We used blood digital transcriptomics in search of viral persistence and Long COVID diagnostic biomarkers in a real-world, general practice-based setting with a long clinical follow-up. We demonstrate systemic SARS-CoV-2 persistence for more than 2 years after acute COVID-19 infection. A 2-gene biomarker, including SARS-CoV-2 antisense RNA, correctly classifies Long COVID with 93.8% sensitivity and 91.7% specificity.

Specific immune transcripts and immunometabolism score correlate to systemic viral load and patient-reported anxiety/depression, providing mechanistic links as well as therapeutic targets to tackle Long COVID.


Link | PDF (Preprint: MedRxiv)

 

Details matter and I have not seen the paper, but if this pans out its potentially huge.

 

I've reformatted to add links to GeneCards. From supplementary materials, in order the 70 (FDR) DEGs are —

Spoiler: 70 DEGs

FYN
GATA3
BMP8A
PDZK1IP1
MAPK13
COASY
PTGS2
CD40LG
ETS1
CD99
IL28A
ORF7a
FGFR1
CD276
SOD1
CCL23
CSF3R
VSIR
MIF
IL1B
TLR9
TTK
CCR3
ADGRE1
IGHE
CX3CL1
CLEC7A
CDH4
IL3RA
Nucleocapsid
TMPRSS2
LGALS3
HIST1H1C
CDH1
PBX1
HOXD4
CLC
CCR4
CCR7
SERPINE3
TGM2
CCL18
ALOX15
ARF6
RAF1
TOP2A
FOSB
SARS-CoV-2_orf1ab_REV
BTG2
ITGAX
CSF2RA
CXCL13
ADGRE5
CCL24
ICOS
TLR6
CALR
IL5RA
MYC
PLAU
PTGDR2
DUSP1
CXCL14
TRAC
NFYC
CD28
MAP2K3
ELL2
ABCC8
S1PR1

 

For the top DEG, from GeneCards: FYN — "This gene is a member of the protein-tyrosine kinase oncogene family. It encodes a membrane-associated tyrosine kinase that has been implicated in the control of cell growth. The protein associates with the p85 subunit of phosphatidylinositol 3-kinase and interacts with the fyn-binding protein."

Note previous comments relating to WASF3 and metformin —

 

Now published
https://www.thelancet.com/journals/lanmic/article/PIIS2666-5247(24)00055-7/fulltext