Abbasi A;  Harbor-UCLA Medical Center, 1124 West Carson Street, Torrance, CA, USA.
Hansen N; Palade J; Paredes D; Meechoovet B; Van Keuren-Jensen
K; Pirrotte P; Stringer WW

Scientific Reports. 16(1):3469, 2026 Jan 26.
VI 1

Headings placed by Dr Paul Older not the authors

Background The Persistence of SARS-CoV-2 in tissues has been proposed as a driver of
prolonged symptoms in long COVID. Pulmonary rehabilitation with exercise
training is a well-established intervention for improving symptoms,
functional capacity, and inflammation in chronic cardiorespiratory
diseases. To investigate whether long COVID is associated with persistent
viral or immune-related signals, we analyzed the long RNA profile of
circulating extracellular vesicles (EVs) to determine the presence of
virus-related transcripts and assess changes in response to exercise
training.
Methods Fourteen adults with long COVID participated in this
single-center pilot clinical trial and completed a 10-week aerobic
exercise training program (twenty 1.5 h sessions). Serum-derived EV RNA
profiles were analyzed via sequencing at rest (T0) and peak
cardiopulmonary exercise testing (T1), before (V2) and after (V24)
exercise training.
Results Differentially expressed genes (DEGs) were identified
(q < 0.05), and pathway activation analysis was performed. Serum EVs
carried diverse RNA species, including protein-coding RNAs, long
non-coding RNAs, short non-coding RNAs, and pseudogenes, with no
virus-related RNAs detected. No significant DEGs were identified at rest
between pre- and post-training, nor in response to acute exercise at
pre-training. However, following training, 53 DEGs were found at peak
exercise (V24T1) compared to rest (V24T0), including three upregulated
genes (ANK3, FTO, FCN1) and 50 downregulated genes (TOP 5: MYL9, NRGN,
H2AC6, MAP3K7CL, B2M). These genes were primarily involved in inflammation
and metabolism. Pathway analysis revealed significant regulation of 100
pathways at post-training compared to pre training, predominantly
inactivated, including pathways involved in inflammation (STAT3 signaling)
and metabolism (O-linked glycosylation).
Conclusions Acute exercise and exercise
training modulated EV-associated gene expression in long COVID, primarily
through transcriptional downregulation. Suppression of inflammation- and
immune-related genes post-training highlights potential molecular
mechanisms underlying symptom improvement and identifies candidate
biomarkers of recovery biology in long COVID. Importantly, while exercise
training did not substantially alter EV RNA content at rest, it enhanced
the body’s ability to mount a dynamic EV-mediated molecular response
during exertion, reflecting improved physiological adaptability.