About nearly ALL blood now is from people who had Covid at least once.This is lethal

[AribertDeckers]

14.1.2026
About nearly ALL blood now is from people who had Covid at least once.This is lethal/b]

Now think about blood transfers!


"Dementia timebomb warning as scientists find Alzheimer's proteins in long Covid patients' blood"
https://www.dailymail.co.uk/health/article-15462417/Dementia-timebomb-Alzheimers-long-Covid.html


https://x.com/RTHM_Health/status/2011838451771912510

[*quote*]
-----------------------
RTHM @RTHM_Health

Recent research shows blood from people with ME/CFS and Long COVID directly harms healthy muscle, reducing force, stressing mitochondria, and causing structural breakdown. Results implicate blood-borne drivers of muscle weakness, exertion intolerance, and PEM, and introduce a non-invasive lab model that mirrors push-crash dynamics.

🔗 https://doi.org/10.1088/1758-5090/adf66c

Title: Muscle tissue exposed to blood from ME/CFS & Long COVID patients leads to severe muscular and mitochondrial deterioration. Image of muscle tissue deteriorating. Site sourced at the bottom: https://doi.org/10.1088/1758-5090/adf66c Arrow pointing to swipe right



https://pbs.twimg.com/media/G-t8aMtWkAE_wjV?format=jpg&name=medium

5:30 PM · Jan 15, 2026
17.4K Views
-----------------------
[*/quote*]

[AribertDeckers]

The source #1:

https://www.sciencedirect.com/science/article/pii/S2352396425005560

[*quote*]
-----------------------
Show full outline
Figures (5)

    Fig. 1. Area under the receiver operating curve (AUC) showing the ability for...
    Fig. 2. Standardised beta coefficients establishing longitudinal rates of change after...
    Fig. 3. Prevalence of increases and decreases in biomarker levels at follow-up across...
    Fig. 4. Overlap in change in biomarkers after the onset of N-PASC
    Fig. 5. Best fitting fractional polynomial time curves stratified by the incidence of...

Tables (2)

Table 1

    Table 2

Extras (1)

    Supplemental Materials

Elsevier
eBioMedicine
Volume 123, January 2026, 106106
eBioMedicine
Articles

Increased phosphorylated tau (pTau-181) is associated with neurological post-acute sequelae of coronavirus disease in essential workers: a prospective cohort study before and after COVID-19 onset

Author links
Xiaohua Yang a, Ashley Fontana a, Sean A.P. Clouston b, Benjamin J. Luft a
Show more

Cite
https://doi.org/10.1016/j.ebiom.2025.106106
Get rights and content
Under a Creative Commons license
Open access
Summary
Background
The COVID-19 pandemic led to a spectrum of post-acute sequelae including several neurological complications including cognitive dysfunction labelled Neurological PASC (N-PASC). We hypothesised that N-PASC was associated with changes in neurological biomarkers after COVID-19.
Methods
N-PASC was established when individuals reported accepted neurological symptoms persisting for ≥3 months arising alongside validated COVID-19. Plasma samples were retrieved from before and after COVID-19 onset among all (n = 227) essential workers who developed COVID-19 with N-PASC and demographically matched with data from 227 controls who either developed COVID-19 without N-PASC (n = 124) or did not develop COVID-19 before follow-up (n = 103). We used single molecular analysis measured pTau-181, GFAP, NfL, Aβ40/42, and total Aβ burden (IAB). Risk factors for N-PASC were examined prior to COVID-19 infection. Multivariable adjusted generalised linear longitudinal modelling with random intercepts was used to examine changes in biomarkers after COVID-19 onset.
Findings
N-PASC was only associated with higher IAB before COVID-19 onset (area under the receiver-operating curve = 0.77). Longitudinal analyses revealed plasma pTau-181 levels increased by 59.3% (95% C.I. = [45.2, 73.4] P = 0.006) following COVID-19 onset in participants who developed N-PASC that were worst among participants reporting central nervous symptoms persisting ≥1.5 years. Post-COVID-19 decreased GFAP and NfL were associated with peripheral symptoms of N-PASC, but not with increased pTau-181. Having ≥20% increases in pTau-181 were associated with increased Aβ40/42 levels at follow-up, and with central neurological symptoms including lingering brain fog and loss of taste/smell.
Interpretation
N-PASC with symptoms consistent with central damage were associated with increased pTau-181 levels. Increases in pTau-181 were associated with increased risk of changes to amyloid biomarkers consistent with Alzheimer's disease in participants with N-PASC and could therefore inform N-PASC prognostication.
Funding
This study was supported in part by funding from the Centers for Disease Control and Prevention (CDC/NIOSH CDC-75D30122c15522) and the National Institutes of Health (NIH/NIA AG049953).

Previous article in issue
Next article in issue
Keywords
Phosphorylated Tau-181
Neurological biomarkers
Post-acute sequelae of COVID-19
Coronavirus disease
Research in context
Evidence before this study
A systematic literature review suggested that individuals reporting neurological post-acute sequelae (PASC) after coronavirus disease (COVID-19) have evidence of neuroinflammation and cognitive decline, while autopsy studies report evidence of vascular disease and cerebral tauopathy absent amyloidosis in non-survivors.
Added value of this study
This study examined biomarkers collected both prior to and following COVID-19 pandemic in a sample of essential workers at midlife to find that individuals with N-PASC had evidence of a 59.3% increase in pTau-181 levels from pre-COVID-19 levels that was not evident in other essential workers and appears to be worst among those whose N-PASC had persisted for more than 1.5 years.
Implications of all the available evidence
These studies imply that symptoms of N-PASC that persist for more than 1.5 years are at increased risk of developing higher than normal levels of circulating levels of pTau-181 that might portend worsened cognitive functioning as individuals age.
Introduction
The COVID-19 pandemic affected more than 775 million individuals worldwide.1 Beyond the acute respiratory disease characterising primary infection with the SARS-CoV-2 virus, accumulating evidence suggests that COVID-19 might affect the central and peripheral nervous system (CNS/PNS) early in the infection course.2,3 Consequently, headache, encephalopathy, insomnia, stroke, and seizures have been observed in as many as 84% of patients.4,5
After acute symptoms resolve, some individuals continue to experience persistent post-acute sequelae of COVID-19 (PASC).6 N-PASC symptoms include neurocognitive changes such as brain fog, forgetfulness, or diminished executive functioning in and can persist ≥3 months.7 Risk factors for N-PASC include COVID-19 severity, and pre-existing medical comorbidities including diabetes, chronic obstructive pulmonary disease, and obesity,8 and SARS-CoV-2 re-infection.9 N-PASC is relatively common in individuals who developed COVID-19 before vaccination: for example, in a prospective study examining >3000 acute COVID-19 in essential workers reported that 56% exhibited neurological symptoms, and 22% developed N-PASC.10
Seeking an explanation for persistent neurological N-PASC (hereafter, N-PASC), researchers used neuroimaging technologies to identify diffuse neuroinflammation of the cerebral parenchyma.11,12 Such inflammation, when sustained, may cause persistent effects including cortical atrophy,13 cerebral disconnection,14 and blood–brain barrier disruption.15 A review of autopsy studies reported evidence of glial activation in patients who died from COVID-19, but also revealed that glial activation in these individuals was not associated with neurological symptoms.16 Neuroimaging studies suggest that neuroinflammation and concomitant brain ageing are associated with COVID-related cognitive decline in survivors, however.17 Interestingly, a small neuropathology study showed abnormal accumulation of hyperphosphorylated Tau protein, absent concomitant with changes in cerebral amyloidosis 4–13 months after recovery from acute COVID-19 in three patients who died after recovering from SARS-CoV-2 infection.18 Reiken and colleagues19 further examined brain lysates from autopsy specimens of patients with COVID-19 infections and found increased levels of phosphorylated tau absent amyloidosis alongside evidence of posttranslational modification of the ryanodine receptor, consistent with a leaky calcium channel.20
Tauopathy is difficult to measure, but is now being monitored using phosphorylated tau-181 (pTau-181) levels in serology,21, 22, 23 where researchers have reported excellent accuracy when detecting early-stage disease in the general population.24 We hypothesised that changes in pTau-181 levels would be associated with the onset of COVID-19 among participants who developed N-PASC. Consistent with the ryanodine receptor deletion hypothesis, we hypothesised that increases in pTau-181 would not be concurrent with evidence of increases in cerebral amyloidosis but may be concurrent with contemporaneous decreases in levels of circulating GFAP.
Methods
Setting
Essential workers who participate in a pre-existing health monitoring program were recruited into a SARS-CoV-2 infection identification protocol through a comprehensive questionnaire that was delivered regularly during the COVID-19 pandemic and at annual monitoring visits thereafter25 (Supplemental Figure S1 has temporal distribution). During an essential worker monitoring program that predated the COVID-19 pandemic, participants completed a biobanking protocol including storage of plasma during annual visits occurring from 1/2019 to 5/2024. In the present study, we selected participants whose plasma samples with one observation falling before, and one after, the COVID-19 pandemic.
We over-selected individuals with N-PASC. There was evidence of sex-differences in potential controls when comparing with versus without N-PASC, so we used propensity score matching to generate a 1:1 sample matched on demographics and medical comorbidities to a group of controls (n = 227) who either reported never having COVID-19 or who developed COVID-19 but did not report any symptoms consistent with N-PASC.
Measures
N-PASC diagnosis
We followed CDC guidelines for diagnosing PASC by, first, identifying participants and collecting the date of COVID-19 symptom onset, type of acute symptom, and details about hospital admissions was collected. We verified the presence of SARS-CoV-2 infections either using polymerase chain reaction testing or antibody testing among those whose symptom onset predated test availability. A validation study examining symptom onset showed that this protocol had high accuracy to detect COVID-19-positive individuals.25 After vaccinations became available, initial vaccination date was recorded. COVID-19 severity was categorised into two groups based on clinical standards (none/asymptomatic/mild versus moderate/severe). Participants with validated COVID-19 were eligible for a diagnosis of N-PASC based on the continuation, or development, of ≥1 neurological symptom (e.g., loss of taste/smell, brain fog, dizziness, vertigo, tinnitus, headache, or balance dysregulation) that emerged <3 months after initial infection and persisted for ≥3 months. This diagnostic protocol had an accuracy of 91% to detect neuroinflammation when using a positron emission tomography protocol.12
Primary outcomes
Neurological biomarkers, including pTau-181, are now well-established and correlate well with cerebral amyloid, pTau, and neurodegeneration.26,27 Briefly, participants' plasma samples were banked in a −80 °C freezer within 30 min of blood collection and then analysed in a single batch using the ultra-sensitive SiMOA SR-X platform (Quanterix, Billerica, Massachusetts, USA). We measured the pTau-181 burden using an Advantage V2 Kit, and measured concentrations of NfL, GFAP, Aβ40, and Aβ42 in plasma using the neurology 4-plex multiplex array. Biomarker assessment succeeded in all cases for pTau-181, but failed in GFAP (n = 2), NfL (n = 4), and Aβ40/Aβ42 (n = 6). There were no statistically significant differences between individuals with successful versus unsuccessful analyses, and imputation resulted in no changes to results, so we used pairwise deletion and intent to include in all reporting.
All biomarker levels are reported as continuous variables in pg/mL. Each plate included two internal controls, and calibration covers for each biomarker.28 Quality control was performed and values falling outside assay detection limits were excluded from analysis. The most common way to use amyloid subspecies is to calculate the Aβ40/42 ratio,29 however since studies of neuroinflammatory conditions suggest that both amyloid types can diminish in serology we also calculated the inverse mean amyloid burden
.29 Higher values on all biomarkers are intended to indicate worse outcomes; to further improve comparability between biomarkers, we reported results at follow-up in terms of percent of baseline value.
We used established cutoffs to determine biomarker positivity. Chatterjee et al.27 suggest using cutoffs including pTau-181 ≥ 1.93 pg/mL, Aβ40/42 ratio ≥ 18.18, GFAP ≥ 183.63 pg/mL, and NfL ≥ 17.31 pg/mL to identify Alzheimer's Disease or a Related Dementia (ADRD). We also show results for a second cutoff for pTau-181 ≥ 1.93 pg/mL was validated that may be more useful in less severe conditions.30 There is no established cutoff for IAB, so no cutoff was reported. Since change in neuropathology in N-PASC might differ in type or severity from, we also provided results using a conservative quantitative cutoff showing the degree of increase/decrease of ≥20%, relative to baseline levels.
Demographics were measured before the COVID-19 pandemic and included sex (male versus female), age (in years) at time of the blood retrieval, educational attainment (high school diploma or less, some college, or university degree). Medical factors included a history of pre-COVID-19 diagnosis of diabetes, hypertension, heart disease, and the presence of obesity (>30 k/m2) or morbid obesity (>40 k/m2) as calculated using researcher-measured height and weight. Biomarker levels can be biased by blood volume,31 so we measured blood volume using an established protocol.32
Apolipoprotein-ε4 (APOE4) allele possession is frequently used to predict the risk of Alzheimer's disease. The APOE4 genotype was measured with the Genomics Shared Resource of Roswell Park Cancer Institute with an Infinium Global Screening Array (Illumina, San Diego, CA, USA).
Statistics
We described the sample differences when samples were retrieved before the COVID-19 pandemic. Crude differences in pre-/post-COVID-19 changes in biomarker outcomes were therefore examined with repeated measures analysis of variance (ANOVA), and P-values are reported. Multivariable adjusted changes in biomarker outcomes were measured with multilevel generalised longitudinal models.33 Since outcomes were skewed away from a biological floor, we followed prior studies in relying on a longitudinal multilevel log-Gamma model.34 Random intercepts were used in longitudinal analyses to account for individual-level variability in biomarker expression before SARS-CoV-2 infection.35 Beta coefficients, standard errors, and P-values derived from t-tests are reported. Results from multiple models were reported using bar graphs. For descriptive purposes, line plots and bar graphs described the association between time since infection and the change in biomarkers after infection, relying on ANOVA to examine differences in biomarker changes before 1.5 years, and afterwards. Since we were interested in the prevalence of large changes in biomarker values, we used a quantitative cutoff (relative change ≥20% of baseline values). Logistic regression is biased in analyses where outcomes are common (>5%), so we used robust log-Poisson regression to estimate multivariable-adjusted relative risks (aRR) and accompanying 95% confidence intervals [95% C.I.].36 We used robust log-Poisson models to examine the risk of specific N-PASC symptoms in this cohort and examined the extent to which those symptoms were associated with biomarker changes. In all models, we used a two-tailed statistical testing (α = 0.05) and reported exact values; we adjusted P-values for the false discovery rate where necessary (FDR = 0.05). Analyses were performed in Stata 17/MP [StataCorp].
A simulation power analysis (power = 0.80, α = 0.05) suggested that a longitudinal model would require a sample size ≥205 participants with N-PASC to determine a difference of ≥0.125 SDs, and ≥430 participants (half with N-PASC) to identify differences between groups over time.
Supplemental analyses
We examined the extent to which biomarkers differed between groups at baseline. We also examined the degree of agreement or association between individuals with biomarkers across different affected biomarkers. Finally, though not relevant to COVID-19 because we were examining biomarkers of ADRD, we also examined the extent to which APOE4 allele possession was associated with changes in biomarkers.
Ethics
The Committee on Research Involving Human Participants reviewed and approved this study (IRB#604113); all study procedures were completed following the protocol. All participants provided informed written consent. This report follows STROBE reporting guideline for cohort studies.37
Role of the funder
The funding agency played no role in analysis or writing and did not influence the decision to submit for publication. The sponsor was not involved in the study design or conduct, data collection or analysis, reporting, or in the decision to submit this manuscript for publication.
Results
After matching and application of inclusion/exclusion criteria (Supplemental Figure S2), a total of 227 paired samples were collected from individuals with COVID-19 and were matched to 227 participants controls, 124 of whom reported a COVID-19 infection without any persistent symptoms, and 103 reported no COVID-19 infections between two biomarker measurements both occurring prior to vaccination.
Descriptive characteristics (Table 1) showed that participants with N-PASC were in their mid-fifties (56.1 ± 7.6 years) and revealed no statistically significant differences between groups. Not shown, the average N-PASC was assessed 2.7 ± 0.7 years after N-PASC symptom onset.

Table 1. Acute and residual symptoms of COVID-19 in participants with N-PASC as compared to participants without N-PASC.
Characteristics    N-PASC (n = 227)    Non-PASC (n = 227)    P-value
Age, years    56.05 (7.57)    55.59 (7.48)    0.499
Sex           
Female    19 (8.37%)    12 (5.29%)    –
Male    208 (91.63%)    215 (94.71%)    0.193
Educational attainment           
High school or less    49 (21.59%)    47 (20.7%)    –
Some college    106 (46.7%)    121 (53.3%)    0.660
University degree    72 (31.72%)    59 (25.99%)    0.550
Race/Ethnicity           
Non-Hispanic white    202 (88.99%)    196 (86.34%)    –
Hispanic    21 (9.25%)    26 (11.45%)    0.530
Other    4 (1.76%)    5 (2.2%)    0.340
Medical comorbidities           
Diabetes    25 (11.01%)    34 (14.98%)    0.166
Heart disease    7 (3.08%)    7 (3.08%)    1.000
Hypertension    83 (36.56%)    75 (33.04%)    0.556
Obese    111 (48.9%)    101 (44.49%)    0.186
Morbidly obese    21 (9.25%)    15 (6.61%)    0.230
†APOE4 allele possession    11 (15.71%)    28 (28.87%)    0.047
Acute COVID-19 characteristics           
Vaccinated before infection    18 (7.93%)    31 (13.66%)    0.050
Date of diagnosis    12/6/20    12/31/20    0.344
Date of reported symptom onset    10/28/20    11/22/20    0.263
Moderate/Severe COVID-19    132 (58.2%)    56 (38.6%)    <0.001
Vaccinated at infection    0 (0%)    27 (11.89%)    <0.001
Vaccinated at follow-up    186 (81.9%)    112 (77.2%)    0.165
Hospitalised    45 (19.8%)    13 (9.0%)    0.073
Likely Variants           
Wild (Sx before 8/24/2020)    81 (37.5%)    41 (33.1%)    –
Alpha (8/25/2020–3/14/2021)    104 (48.1%)    52 (41.9%)    0.962
Delta (3/15/2021–12/4/2021)    22 (10.2%)    19 (15.3%)    0.145
Omicron (Sx after 12/5/2021)    9 (4.2%)    12 (9.7%)    0.040
Acute COVID-19 symptoms           
Fever    132 (58.1%)    47 (32.2%)    <0.001
Fatigue    159 (70%)    56 (38.8%)    <0.001
Headache    128 (56.4%)    45 (31.3%)    <0.001
Cough    142 (62.6%)    42 (28.6%)    <0.001
Chills    63 (27.8%)    18 (12.3%)    <0.001
Joint/Muscle pain    135 (59.5%)    48 (33%)    <0.001
Constipation    83 (36.6%)    19 (13.2%)    <0.001
Wheeze    78 (34.4%)    15 (10.6%)    <0.001
Sore throat    75 (33%)    24 (16.3%)    <0.001
Loss of taste or smell    129 (56.8%)    31 (21.6%)    <0.001
Gastrointestinal    118 (52%)    62 (42.7%)    0.048
Shortness of breath    125 (55.1%)    25 (17.2%)    <0.001
Nausea or vomiting    19 (8.4%)    7 (4.8%)    0.131
Dizziness/Vertigo    135 (59.5%)    55 (37.9%)    <0.001
Anxiety    23 (10.1%)    0 (0%)    <0.001
Brain fog    33 (14.5%)    1 (0.9%)    <0.001
Congestion    32 (14.1%)    16 (11%)    0.321
Weight loss    17 (7.5%)    3 (2.2%)    0.009
N-PASC symptoms           
Brain fog    143 (63%)    0 (0.0%)    <0.001
Anosmia/Ageusia    125 (55.1%)    0 (0.0%)    <0.001
Behavioural change    43 (18.9%)    0 (0.0%)    <0.001
Gastrointestinal    30 (13.2%)    0 (0.0%)    <0.001
Vertigo    25 (11%)    0 (0.0%)    <0.001
Tinnitus    11 (4.8%)    0 (0.0%)    <0.001
Dizziness    16 (7%)    0 (0.0%)    <0.001
Loss of balance    5 (2.2%)    0 (0.0%)    0.045
Note: N-PASC: post-acute sequelae of COVID-19; COVID-19 coronavirus disease 2019. P-values were estimated using Carmitage's non-parametric trend test. Note that APOE4 allele possession was only available for 167 participants.
Fig. 1 shows area under the receiver operating curve determining potential pre-COVID-19 risk factors for the development of N-PASC after COVID-19 (maximal AUC = 0.78; largest single-biomarker risk factor was IAB, AUC = 0.77). In the saturated model adjusting for demographics and biomarkers, analyses suggested that IAB (B = 4.20 [2.99–5.41] P < 0.001) and Aβ40/42 (B = −0.02 [−0.036, −0.003] P = 0.019), but not pTau-181 (P = 0.270), were elevated in participants who developed N-PASC.

    Download: Download high-res image (307KB)
    Download: Download full-size image

Fig. 1. Area under the receiver operating curve (AUC) showing the ability for pre-COVID-19 differences in each biomarker to predict the development of PASC after COVID-19 infection in the sample of individuals who developed COVID-19 between the first and second observation.
In longitudinal analyses, pTau-181 showed a statistically significant longitudinal increase (β = 0.15, P < 0.001) in the N-PASC group when compared to controls (Fig. 2) corresponding with a 0.24 pg/mL increase over baseline or a 59.3% increase (95% C.I. = [45.2, 77.34], P < 1E-06; full results in Supplemental Table S1). When compared to never-COVID-19 controls, we also found statistically significant reductions in the GFAP (β = −0.091, P = 0.007) and NfL (β = −0.077, P = 0.022) among those with N-PASC. Among participants infected with COVID-19, we also saw an increase in IAB when compared to never-infected controls (β = 0.19, p < 0.001).

    Download: Download high-res image (191KB)
    Download: Download full-size image

Fig. 2. Standardised beta coefficients establishing longitudinal rates of change after the onset of COVID-19, stratified by biomarker. Analyses were grouped to compare differences between individuals who developed COVID-19 but did not develop N-PASC (turquoise dashed bars), and those who did develop both COVID-19 and N-PASC (gold solid bars). The reference category is individuals who were followed-up twice over the same period but who did not develop COVID-19 (set at zero). Coefficients are derived from multiple biomarker-specific ln-Gamma models and adjust for age, gender, and blood volume levels, and individual differences in pre-COVID-19 proteomic regulation propensity. ∗∗FDR-Adjusted P-value < 0.05, ∗P < 0.05.
Fig. 3 illustrates the number of participants whose values on biomarkers changed from baseline, stratified by COVID-19 infection and N-PASC. Overall, 58.6% of N-PASC participants exhibited increases in pTau-181 levels ≥20% relative to pre-COVID levels. Interestingly, 25.8% of participants with N-PASC experienced ≥20% relative decreases in GFAP from pre-COVID-19 levels, while 25.6% of participants experienced ≥20% relative decreases in NfL from pre-COVID-19 levels. Though not shown in Fig. 3, having a ≥20% increases in pTau-181 levels was associated with higher risk (RR = 2.16 [1.40–3.33] P = 0.001) that pTau-181 levels were deemed abnormal (pTau-181 > 1.93, prevalence = 45.1% in participants with N-PASC), and while there was a trend linking pTau-181 increases with heightened Aβ40/42 Ratio (RR = 1.83 [0.99–3.38] P = 0.054). As shown in Supplemental Table S2, when examining crossover between biomarkers we found that 75.6% of participants with N-PASC had exhibited increased pTau-181 (57.6% showed this phenotype) with/without decreased GFAP or NfL (39.3% showed one of these phenotypes). Also not shown in Fig. 3, analyses relying on biomarkers at baseline to predict change in pTau-181 levels at follow-up revealed that PASC (β = 17.26, SE = 4.73, P < 0.001) and IAB values (β = 184.29, SE = 75.22, P = 0.014) were both associated with the degree of pTau-181 change at follow-up.

    Download: Download high-res image (926KB)
    Download: Download full-size image

Fig. 3. Prevalence of increases and decreases in biomarker levels at follow-up across COVID-19 groupings. Panel A shows ≥20% relative increases in biomarker levels from baseline. Panel B shows ≥20% relative decreases in biomarker levels from baseline, grouped by biomarker type. Estimates stratified to compare differences between individuals without COVID-19 (grey, dotted bar), those who developed COVID-19 but not N-PASC (turquoise dashed bars), and those who did develop COVID-19 and N-PASC (gold solid bars). ∗∗FDR-Adjusted P-value < 0.05, ∗P < 0.05.
Fig. 4 shows relative risks linking ≥20% increases in pTau-181 or decreases in GFAP and NfL in participants with N-PASC. These findings revealed that increases in pTau-181 was associated with a strong increase in Aβ40/42 ratios (RR = 8.16 [1.03–64.87] P = 0.048), alongside increases in the risk that IAB increased ≥20% (aRR = 1.68 [1.04–2.71] P = 0.036), alongside a concomitant decreased risk that IAB decreased by ≥ 20%. In contrast, ≥20% relative decreases in GFAP and NfL were intercorrelated (aRR = 2.85 [1.52–5.33] P = 0.001) and were associated with a very high risk that IAB increased ≥20%. Decreases in NfL were associated with ≥20% decreased Aβ40/42 ratios (Fig. 4).

    Download: Download high-res image (663KB)
    Download: Download full-size image

Fig. 4. Overlap in change in biomarkers after the onset of N-PASC. Results are stratified by colour so that ≥20% increases in biomarkers (gold) versus ≥20% decreases in biomarkers (striped, turquoise). Note that results are missing when comparing relative risks in relation to increases/decreases in the same biomarker. Panel A shows results for N-PASC-related increases in pTau-181. Panel B shows results associated with N-PASC-related decreases in GFAP. Panel C shows associations with N-PASC-related decreases in NfL. ∗∗FDR-Adjusted P-value < 0.05, ∗P < 0.05.
Table 2 shows the degree of association between biomarker changes in N-PASC and specific symptoms. These results highlight that pTau-181 changes were mostly associated with evidence of symptoms consistent with neurological changes including muscle weakness (aRR = 1.30 [1.08–1.57] P = 0.006) alongside loss of taste/smell, anxiety/depression, and brain fog. Intriguingly, reductions in GFAP were most strongly associated with evidence of muscle weakness (aRR = 1.69 [1.17–2.46] P = 0.006) alongside fatigue, and loss of taste/smell but decreases in NfL were only associated with residual shortness of breath (aRR = 1.76 [1.07–2.90] P = 0.028).

Table 2. Association between specific lingering symptoms of coronavirus disease 2019 (COVID-19) and increases in pTau-181 or decreases in GFAP/NfL.
Residual symptom    ≥20% increase in pTau-181    ≥20% decrease in GFAP    ≥20% decrease in NfL
aRR 95% C.I.    P    aRR 95% C.I.    P    aRR 95% C.I.    P
Any central nervous system    1.31 (1.09–1.58)    0.004    1.22 (0.78–1.91)    0.302    1.22 (1.22–1.22)    0.246
Any peripheral nervous system    1.28 (1.04–1.56)    0.018    1.22 (0.57–2.61)    0.015    1.23 (1.23–1.23)    0.064
Lost sense of taste/Smell    1.30 (1.07–1.59)    0.010    1.57 (1.05–2.35)    0.029    1.24 (0.80–1.92)    0.327
Anxiety or depression    1.26 (1.05–1.53)    0.016    1.21 (0.82–1.78)    0.343    1.13 (0.76–1.68)    0.556
Brain fog    1.22 (1.00–1.48)    0.047    1.21 (0.81–1.81)    0.352    1.31 (0.88–1.95)    0.185
Headache    1.13 (0.69–1.84)    0.630    1.51 (0.64–3.56)    0.346    1.41 (0.6–3.29)    0.427
Muscle weakness    1.30 (1.08–1.57)    0.006    1.69 (1.17–2.46)    0.006    1.38 (0.95–2.01)    0.090
Fatigue    1.20 (0.91–1.58)    0.191    1.67 (1.05–2.67)    0.031    1.42 (0.85–2.37)    0.179
Cough    1.12 (0.79–1.58)    0.536    1.28 (0.65–2.50)    0.472    1.28 (0.64–2.55)    0.481
Shortness of breath    1.19 (0.90–1.56)    0.217    1.46 (0.85–2.49)    0.167    1.76 (1.07–2.90)    0.028
Wheezing    1.27 (0.82–1.97)    0.282    0.90 (0.25–3.26)    0.874    1.34 (0.48–3.73)    0.574
Congestion    1.39 (0.97–2.00)    0.072    0.85 (0.24–3.01)    0.795    0.40 (0.06–2.64)    0.343
Note: Results report results after adjusting for demographics. Peripheral symptoms included numbness and tingling or pain in the extremities. Central symptoms included brain fog, loss of taste or smell, dizziness, tinnitus, loss of balance, and vertigo.
Next, we showed the mean elevation in biomarkers by time since COVID-19, stratified by infection and N-PASC status (Fig. 5). These analyses suggested that while the increase in pTau-181 levels was, on average, only 14.6% within ≤1.5 years of infection but increased 56.5% (P < 0.001) among individuals with N-PASC who were assessed >1.5 years after symptom onset. Similarly, reductions in GFAP were more pronounced among participants with N-PASC ≤1.5 years after infection (69.0% decrease, P < 0.001 versus 16.3% decrease, P = 0.33 among those within >1.5 years of infections).

    Download: Download high-res image (554KB)
    Download: Download full-size image

Fig. 5. Best fitting fractional polynomial time curves stratified by the incidence of COVID-19 and the presence of N-PASC. Three groups include individuals who developed a validated case of acute COVID-19 and developed subsequent N-PASC (gold solid line), individuals who developed a validated case of acute COVID-19 but who did not develop any N-PASC (navy dashed line), and post-pandemic values for individuals who had not yet developed COVID-19 (charcoal long-dashed line). 95% confidence intervals are shown in transparent grey boxes. Panels A–E show the temporal trajectories of each biomarker, while Panel F shows the estimated differences before/after 1.5 years after COVID-19. Note reports the average difference in levels of change from normal among individuals whose N-PASC has lasted more, or less, than 1.5 years.
Supplemental analyses examined whether APOE4 genotype was associated with increases in pTau-181 levels in N-PASC. In 167 participants who had agreed to genotyping, we found that APOE4 possession was not associated with change in any of the biomarkers among individuals with N-PASC (P-values > 0.10), but was generally associated with higher pTau-181 levels (P = 0.002 at baseline, P = 0.006 at follow-up in participants with N-PASC).
Discussion
In this moderately-sized prospective study of patients who developed N-PASC we reported that, when compared to pre-COVID-19 information, levels of pTau-181 (a biomarker of neurodegenerative diseases) increased more among participants with N-PASC than among similar individuals who, at the time of data collection, had not developed COVID-19 or had developed an acute case of COVID-19 that lacked residual symptoms. Results are consistent with the view that changes in pTau-181 inconsistent with normal ageing may be common among participants with N-PASC. Indeed, more than half of participants with N-PASC experienced a ≥20% increase in absolute levels of pTau-181 relative to pre-COVID-19 levels. Among participants with N-PASC who exhibited pTau-181 increases ≥20% relative to pre-COVID levels, 45.1% expressed pTau-181 levels above an established cutoff to identify ADRD.30 Examining progression, increases in pTau-181 levels were higher among those with N-PASC duration periods >1.5 years since infection among those with N-PASC, consistent with a heightened potential for longitudinal progression, and increases were also associated with increased risk of abnormal AB40/42 ratios consistent with Alzheimer's disease.27
N-PASC was associated with changes in pTau-181 that exceeded cutoffs used in studies of ADRD. Several studies have noted evidence of change to cognition after SARS-CoV-2 infections emerging among those with N-PASC.38 Evidence from biomarker studies is clear in indicating that increases in the Aβ40/42 ratio alongside increases in pTau-181 are phenotypic of sporadic ADRD,27 and may portend a worsening prognosis in a subset of individuals. If prognosis is poorer, then researchers may expect that studies reporting COVID-19 related cognitive decline may find worsening functioning over longer follow-up periods.38, 39, 40 Further research is warranted to determine the cognitive implications of biomarker dysregulation in N-PASC.
The findings of persistent elevation of plasma pTau-181 support the conclusion that pTau-181 may identify a role of sustained tau pathology after infection. The fact that average elevation in pTau-181 grew over the observational period may, if replicated longitudinally, suggest a temporal lag between COVID-19 onset and increases in circulating levels of tau phosphorylation. These findings support pTau-181 as a valuable longitudinal biomarker for N-PASC and highlight the potential need for early, tau-targeted interventions to mitigate progressive cognitive decline. However, these results also require replication in neuroimaging to both clarify whether elevations for pTau-181 indicate the presence of cerebral tauopathy and, if so, what type while also determine the extent to which stability in pTau-181 indicate the absence of cerebral tauopathy N-PASC rather than biomarker insensitivity or random error over time.
Our study found that patients who develop N-PASC after COVID-19 might share certain clinicopathological features with AD. Indeed, while Tau functions adaptively to stabilise neuronal microtubules under normal physiologic conditions23 and can also be dysregulated and spread by reactive glia,41 especially in the context of chronic inflammation.42 However, the prognostic implications of increases in circulating pTau-181 absent concurrent amyloidosis are unknown. Usually, β-amyloid peptide plays a central role in triggering Tau phosphorylation in ADRD,43 a process that leads to subsequent microtubule destabilisation resulting in threads that coalesce into neuronal tangles.44 Intriguingly, in this study we found that evidence of increased pTau-181 was associated with increases in AB40/42 ratios consistent with a pTau-mediated ADRD. Further studies are needed to determine whether the increased levels of plasma pTau-181 correlate with evidence of cerebral Tauopathy and, if results are replicated, pTau-181 might aid in diagnosis and might serve as an important monitoring and therapeutic target.45
We found that individuals who developed N-PASC had higher Aβ40/42 ratios (AUC = 0.63, P = 0.007), NfL (AUC = 0.59, P = 0.002), and IAB values (AUC = 0.74, P < 0.001) before developing COVID-19. Higher values suggest that N-PASC might be more likely in those individuals who have heightened vulnerability to neurological disease. Amyloidosis often requires a secondary neuropathology to elicit the most severe symptomatology. If these findings indicate that cerebral amyloidosis is present, even in its mildest forms, then the post-COVID-19 increase in pTau-181 may correspond to the onset of pathological Alzheimer's disease.
One potentially paradoxical finding was that higher pre-COVID-19 NfL was weakly associated with higher risk of N-PASC but that individuals with N-PASC then saw decreased NfL and GFAP following COVID-19 onset appears paradoxical. We did not find that this was associated with evidence of increased pTau-181 but, instead, found that the decrease in NfL was coupled with reduced GFAP. One possibility is that reductions in GFAP and NfL indicate that the neuroimmune system is repairing itself and utilising these proteins to aid in neurogenesis, but if so the presence of persistent symptoms seems unlikely and we would expect to see glial activation in autopsy studies, a result that has not been identified.18,19 An alternative explanation may be that in some individuals, COVID-19-related neuroinflammation restricted glymphatic clearance causing increased aggregation of larger proteins (like NfL or GFAP: width ∼10 nm) but potentially allowing smaller proteins like pTau-181 (5 nm) to pass through unimpeded. Further studies are needed that seek to determine the impact of N-PASC on glymphatic clearance to determine if it, and not reductions in NfL and GFAP, are also associated with peripheral nervous and cerebrovascular systems.
Evidence suggests that COVID-19 involves the neuroimmune system in a heretofore unrecognised way and as the immune response systematically evolves, the infection resolves.46,47 However, perhaps because of the immunologically privileged nature of the central nervous system, infections may persist and give rise to indolent subacute encephalitis with concomitant neuroinflammation.48 Prior neuroimaging studies have demonstrated that patients with N-PASC show diffuse changes in white and grey matter connectivity, neuroinflammation, and cerebral atrophy Molecular imaging studies have revealed diffuse microglial activation.11,12 Yet, microglial activation is known to facilitate the spread of pre-existing cerebral tau across neurons, so perhaps activated microglia release inflammatory cytokines to trigger kinases responsible for tau phosphorylation and accelerate progression of latent neuropathology.49
Limitations
This study nevertheless has several important limitations. First, this study used information on plasma distribution occurring before the COVID-19 pandemic to help develop biomarker-related results and provide evidence of mild to severe increases in pTau-181 after N-PASC development. Studies relying solely on post-COVID information might have diminished effect sizes, because they cannot ensure normal pTau-181 levels in participants before infection.
Second, this study relied on a sample of essential workers who participated in an occupational monitoring study. Since N-PASC in this population was diagnosed prospectively as individuals experienced a pandemic, and COVID-19 diagnoses were verified by medical charts, this study is likely to provide more reliable and sensitive results than other studies.
Third, because the study was limited to several types of essential workers, relatively few women were included. Despite the fact that women make up the majority of participants in studies of N-PASC, in our study male participants tend to have more severe COVID-19,50 but clinical studies have noted that women are about 90% more likely to report having ≥3 N-PASC symptoms and carry a higher levels of post-COVID fatigue.51 Since women are also known to have higher burden of cerebral tau,52 more work is needed that specifically focuses on sex differences in N-PASC.
Fourth, while examining several cognitively active phenotypes in these data, we did not examine cognition in this study. Future work is needed not only to examine whether concurrent changes in biomarkers mediate the established relationship between N-PASC and cognition but also whether changes in biomarkers portend a pattern of progressive decline consistent with ADRD as was implicated by concurrent elevations in pTau-181 and the Aβ40/42 ratio.
Fifth, while results for GFAP and NfL appear to go in an incorrect direction, it is worth noting that GFAP and NfL may be expressed by other tissues including in the reproductive organs (GFAP, NfL) and eyes (NfL) and so decrements in the blood may not directly reflect changes in the brain. Future research is needed that specifically examines whether changes in the blood are correlated with similar changes evident in the brains of individuals with N-PASC.
Finally, while we examined a biomarker for cerebral tauopathy, this study did not test whether tau in N-PASC was associated with cerebral tau burden. It is possible that changes in blood do not always reflect changes in the brain, so follow-up research is needed that specifically examines the implications of this work to cerebral tauopathy.
Clinical implications
Since elevated pTau-181 levels often predict cognitive decline, this study might imply that the long-term prognosis for participants with N-PASC may be poorer in patients with increased pTau-181. Long-term studies are required to determine whether pTau-181 levels will continue to rise or will stabilise over time in participants with N-PASC. Additionally, prognostic studies are necessary that determine if pTau-181 increases are predictive of subsequent cognitive decline and impairment in N-PASC. However, the finding that changes in pTau-181 are independent of changes in NfL and GFAP might support the view that N-PASC is a heterogeneous condition with different symptoms that might indicate independent neuropathological processes. Thus, if cerebral tau is present, then results could highlight the potential for neuroprotective strategies including anti-inflammatory or anti-Tau therapies to help mitigate COVID-19-related cognitive decline.53 Future research is warranted that elucidates the mechanisms through which COVID-19 influences Tau phosphorylation and assesses the long-term prognostic implications of pTau-181 elevation in at-risk populations.
Contributors
Clouston, Yang, and Luft had full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis. Study concept and design: Yang, Clouston, Luft. Acquisition, analysis, and interpretation of data: Yang, Clouston, Fontana. Drafting of the manuscript: Clouston, Yang. Critical revision of the manuscript for important intellectual content: All Authors. Statistical analysis: Clouston. Obtained funding: Clouston, Luft. Administrative, technical, or material support: Fontana, Yang. Study supervision: Luft. All authors read and approved of the final version of the manuscript.
Data sharing statement
These data represent private health information that also include potentially identifiable longitudinal data. After publication, a limited dataset can be shared with bona fide researchers upon receipt of a written request by email and after executing a data use agreement. Analytic code relies on standard packages available from Stata MP V.17, and detailed analyses will be shared with researchers via the open science framework (osf.io).
Declaration of interests
The authors have no conflicts of interest to disclose.
Acknowledgements
This study was supported in part by funding from the Centers for Disease Control and Prevention (CDC/NIOSH CDC-75D30122c15522) and the National Institutes of Health (NIH/NIA AG049953).
Appendix A. Supplementary data
Download: Download Word document (135KB)

Supplemental Materials.
References

    1
    World Health Organization
    COVID-19 weekly epidemiological update
    WHO Special Edition (2024)
    Google Scholar

2
A.D. Proal, M.B. VanElzakker
Long COVID or Post-acute Sequelae of COVID-19 (PASC): an overview of biological factors that may contribute to persistent symptoms
Front Microbiol, 12 (2021), Article 698169
View in Scopus
Google Scholar3
E. Panagea, L. Messinis, M.C. Petri, et al.
Neurocognitive impairment in long COVID: a systematic review
Arch Clin Neuropsychol, 40 (1) (2025), pp. 125-149
Crossref
View in ScopusGoogle Scholar4
H.E. Davis, L. McCorkell, J.M. Vogel, E.J. Topol
Long COVID: major findings, mechanisms and recommendations
Nat Rev Microbiol, 21 (3) (2023), pp. 133-146
Crossref
View in ScopusGoogle Scholar5
S.R. Stein, S.C. Ramelli, A. Grazioli, et al.
SARS-CoV-2 infection and persistence in the human body and brain at autopsy
Nature, 612 (7941) (2022), pp. 758-763
Crossref
View in ScopusGoogle Scholar6
S.J. Ahmad, C.M. Feigen, J.P. Vazquez, A.J. Kobets, D.J. Altschul
Neurological sequelae of COVID-19
J Integr Neurosci, 21 (3) (2022), p. 77
View PDFView articleGoogle Scholar7
E.L. Graham, J.R. Clark, Z.S. Orban, et al.
Persistent neurologic symptoms and cognitive dysfunction in non-hospitalized Covid-19 "long haulers"
Ann Clin Transl Neurol, 8 (5) (2021), pp. 1073-1085
Crossref
View in ScopusGoogle Scholar8
E.L. Hill, H.B. Mehta, S. Sharma, et al.
Risk factors associated with post-acute sequelae of SARS-CoV-2: an N3C and NIH RECOVER study
BMC Public Health, 23 (1) (2023), p. 2103
View in Scopus
Google Scholar9
T.K. Babalola, S.A.P. Clouston, Z. Sekendiz, et al.
SARS-COV-2 re-infection and incidence of post-acute sequelae of COVID-19 (PASC) among essential workers in New York: a retrospective cohort study
Lancet Reg Health Am, 42 (2025), Article 100984
View PDFView articleView in ScopusGoogle Scholar10
E. Lhuillier, Y. Yang, O. Morozova, et al.
The impact of world trade center related medical conditions on the severity of COVID-19 disease and its long-term sequelae
Int J Environ Res Public Health, 19 (12) (2022), Article 6963
Crossref
View in ScopusGoogle Scholar11
J. Braga, M. Lepra, S.J. Kish, et al.
Neuroinflammation after COVID-19 with persistent depressive and cognitive symptoms
JAMA Psychiatry, 80 (8) (2023), pp. 787-795
Crossref
View in ScopusGoogle Scholar12
S.A. Clouston, P. Vaska, T. Babalola, et al.
Glial activation among individuals with neurological post-acute sequelae of Coronavirus disease 2019: a brain fog positron emission tomography study of [18F]-FEPPA
Brain Behav Immun Health, 44 (2025), Article 100945
View PDFView articleView in ScopusGoogle Scholar13
S. Capelli, A. Arrigoni, A. Napolitano, et al.
MRI evidence of gray matter loss in COVID-19 patients with cognitive and olfactory disorders
Ann Clin Transl Neurol, 11 (9) (2024), pp. 2457-2472
Crossref
View in ScopusGoogle Scholar14
S. Huang, Z. Zhou, D. Yang, et al.
Persistent white matter changes in recovered COVID-19 patients at the 1-year follow-up
Brain, 145 (5) (2022), pp. 1830-1838
Crossref
View in ScopusGoogle Scholar15
C. Greene, R. Connolly, D. Brennan, et al.
Blood–brain barrier disruption and sustained systemic inflammation in individuals with long COVID-associated cognitive impairment
Nat Neurosci, 27 (3) (2024), pp. 421-432
Crossref
View in ScopusGoogle Scholar16
G. Cosentino, M. Todisco, N. Hota, et al.
Neuropathological findings from COVID-19 patients with neurological symptoms argue against a direct brain invasion of SARS-CoV-2: a critical systematic review
Eur J Neurol, 28 (11) (2021), pp. 3856-3865
Crossref
View in ScopusGoogle Scholar17
B. Besteher, T. Rocktäschel, A.P. Garza, et al.
Cortical thickness alterations and systemic inflammation define long-COVID patients with cognitive impairment
Brain Behav Immun, 116 (2024), pp. 175-184
View PDFView articleView in ScopusGoogle Scholar18
X. Qi, S. Yuan, J. Ding, et al.
Emerging signs of Alzheimer-like tau hyperphosphorylation and neuroinflammation in the brain post recovery from COVID-19
Aging Cell, 23 (11) (2024), Article e14352
View in Scopus
Google Scholar19
S. Reiken, L. Sittenfeld, H. Dridi, Y. Liu, X. Liu, A.R. Marks
Alzheimer's-like signaling in brains of COVID-19 patients
Alzheimers Dement, 18 (5) (2022), pp. 955-965
Crossref
View in ScopusGoogle Scholar20
H. Ma, F. Yang, M.R. Butler, J. Rapp, Y.-Z. Le, X.-Q. Ding
Ryanodine receptor 2 contributes to impaired protein localization in cyclic nucleotide-gated channel deficiency
eNeuro, 6 (3) (2019), Article ENEURO.0119-19.2019
View in Scopus
Google Scholar21
C. Tissot, A. L Benedet, J. Therriault, et al.
Plasma pTau181 predicts cortical brain atrophy in aging and Alzheimer's disease
Alzheimers Res Ther, 13 (1) (2021), p. 69
View in Scopus
Google Scholar22
E.H. Thijssen, R. La Joie, A. Wolf, et al.
Diagnostic value of plasma phosphorylated tau181 in Alzheimer's disease and frontotemporal lobar degeneration
Nat Med, 26 (3) (2020), pp. 387-397
Crossref
View in ScopusGoogle Scholar23
M.M. Mielke, C.E. Hagen, J. Xu, et al.
Plasma phospho-tau181 increases with Alzheimer's disease clinical severity and is associated with tau- and amyloid-positron emission tomography
Alzheimers Dement, 14 (8) (2018), pp. 989-997
View PDFView articleCrossrefView in ScopusGoogle Scholar24
K.M. Quispialaya, J. Therriault, A. Aliaga, et al.
Plasma phosphorylated tau181 outperforms [18F] fluorodeoxyglucose positron emission tomography in the identification of early Alzheimer disease
Eur J Neurol, 31 (12) (2024), Article e16255
View in Scopus
Google Scholar25
O. Morozova, S.A.P. Clouston, J. Valentine, A. Newman, M. Carr, B.J. Luft
COVID-19 cumulative incidence, asymptomatic infections, and fatality in Long Island, NY, January-August 2020: a cohort of World Trade Center responders
PLoS One, 16 (7) (2021), Article e0254713
Crossref
View in ScopusGoogle Scholar26
A. Cano, M. Capdevila, R. Puerta, et al.
Clinical value of plasma pTau181 to predict Alzheimer's disease pathology in a large real-world cohort of a memory clinic
eBioMedicine, 108 (2024), Article 105345
View PDFView articleView in ScopusGoogle Scholar27
P. Chatterjee, S. Pedrini, J.D. Doecke, et al.
Plasma Aβ42/40 ratio, p-tau181, GFAP, and NfL across the Alzheimer's disease continuum: a cross-sectional and longitudinal study in the AIBL cohort
Alzheimers Dement, 19 (4) (2023), pp. 1117-1134
Crossref
View in ScopusGoogle Scholar28
M. Kritikos, E.D. Diminich, J. Meliker, et al.
Plasma amyloid beta 40/42, phosphorylated tau 181, and neurofilament light are associated with cognitive impairment and neuropathological changes among World Trade Center responders: a prospective cohort study of exposures and cognitive aging at midlife
Alzheimers Dement (Amst), 15 (1) (2023), Article e12409
View in Scopus
Google Scholar29
S.A.P. Clouston, Y. Deri, E. Diminich, et al.
Posttraumatic stress disorder and total amyloid burden and amyloid-beta 42/40 ratios in plasma: results from a pilot study of World Trade Center responders
Alzheimers Dement (Amst), 11 (2019), pp. 216-220
View PDFView articleCrossrefView in ScopusGoogle Scholar30
A.J. Thomas, C.A. Hamilton, A. Heslegrave, et al.
A longitudinal study of plasma pTau181 in mild cognitive impairment with lewy bodies and Alzheimer's disease
Mov Disord, 37 (7) (2022), pp. 1495-1504
Crossref
View in ScopusGoogle Scholar31
A. Manouchehrinia, F. Piehl, J. Hillert, et al.
Confounding effect of blood volume and body mass index on blood neurofilament light chain levels
Ann Clin Transl Neurol, 7 (1) (2020), pp. 139-143
Crossref
View in ScopusGoogle Scholar32
H.J. Lemmens, D.P. Bernstein, J.B. Brodsky
Estimating blood volume in obese and morbidly obese patients
Obes Surg, 16 (6) (2006), pp. 773-776
Crossref
View in ScopusGoogle Scholar33
S. Rabe-Hesketh, A. Skrondal, A. Pickles
Generalized multilevel structural equation modeling
Psychometrika, 69 (2) (2004), pp. 167-190
View in Scopus
Google Scholar34
P. Zhang, D. Luo, P. Li, L. Sharpsten, F.A. Medeiros
Log-gamma linear-mixed effects models for multiple outcomes with application to a longitudinal glaucoma study
Biom J, 57 (5) (2015), pp. 766-776
Crossref
View in ScopusGoogle Scholar
35
S. Rabe-Hesketh, A. Skrondal
Multilevel and longitudinal modeling using Stata
STATA press (2008)
Google Scholar
36
W. Chen, L. Qian, J. Shi, M. Franklin
Comparing performance between log-binomial and robust Poisson regression models for estimating risk ratios under model misspecification
BMC Med Res Methodol, 18 (1) (2018), p. 63
Google Scholar
37
E. von Elm, D.G. Altman, M. Egger, et al.
The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies
Ann Intern Med, 147 (8) (2007), pp. 573-577
Crossref
View in ScopusGoogle Scholar38
Z. Sekendiz, O. Morozova, M.A. Carr, et al.
Characterization of change in cognition before and after COVID-19 infection in essential workers at midlife
Am J Med Open, 12 (2024), Article 100076
View PDFView articleView in ScopusGoogle Scholar39
R.J. Caselli, Y. Chen, K. Chen, R.J. Bauer, D.E.C. Locke, B.K. Woodruff
Cognition before and after COVID-19 disease in older adults: an exploratory study
J Alzheimers Dis, 91 (2023), pp. 1049-1058
Crossref
View in ScopusGoogle Scholar40
L. Crivelli, K. Palmer, I. Calandri, et al.
Changes in cognitive functioning after COVID-19: a systematic review and meta-analysis
Alzheimers Dement, 18 (5) (2022), pp. 1047-1066
Crossref
View in ScopusGoogle Scholar41
N. Maphis, G. Xu, O.N. Kokiko-Cochran, et al.
Reactive microglia drive tau pathology and contribute to the spreading of pathological tau in the brain
Brain, 138 (Pt 6) (2015), pp. 1738-1755
Crossref
View in ScopusGoogle Scholar42
G.S. Bloom
Amyloid-β and tau: the trigger and bullet in Alzheimer disease pathogenesis
JAMA Neurol, 71 (4) (2014), pp. 505-508
Crossref
View in ScopusGoogle Scholar43
C.R. Jack Jr., D.A. Bennett, K. Blennow, et al.
NIA-AA research framework: toward a biological definition of Alzheimer's disease
Alzheimers Dement, 14 (4) (2018), pp. 535-562
View PDFView articleCrossrefView in ScopusGoogle Scholar44
M.D. Weingarten, A.H. Lockwood, S.-Y. Hwo, M.W. Kirschner
A protein factor essential for microtubule assembly
Proc Natl Acad Sci U S A, 72 (5) (1975), pp. 1858-1862
Crossref
View in ScopusGoogle Scholar
45
N. Basheer, T. Smolek, I. Hassan, et al.
Does modulation of tau hyperphosphorylation represent a reasonable therapeutic strategy for Alzheimer's disease? From preclinical studies to the clinical trials
Mol Psychiatry, 28 (6) (2023), pp. 2197-2214
Crossref
View in ScopusGoogle Scholar46
S. Furman, K. Green, T.E. Lane
COVID-19 and the impact on Alzheimer's disease pathology
J Neurochem, 168 (10) (2024), pp. 3415-3429
View in Scopus
Google Scholar47
J.A. Mapunda, H. Tibar, W. Regragui, B. Engelhardt
How does the immune system enter the brain?
Front Immunol, 13 (2022), Article 805657
View in Scopus
Google Scholar48
F. Gao, V. Mallajosyula, P.S. Arunachalam, et al.
Spheromers reveal robust T cell responses to the Pfizer/BioNTech vaccine and attenuated peripheral CD8+ T cell responses post SARS-CoV-2 infection
Immunity, 56 (4) (2023), pp. 864-878.e4
View PDFView articleView in ScopusGoogle Scholar49
C. Langworth-Green, S. Patel, Z. Jaunmuktane, et al.
Chronic effects of inflammation on tauopathies
Lancet Neurol, 22 (5) (2023), pp. 430-442
View PDFView articleView in ScopusGoogle Scholar50
J.-M. Jin, P. Bai, W. He, et al.
Gender differences in patients with COVID-19: focus on severity and mortality
Front Public Health, 8 (2020), Article 152
View PDFView articleGoogle Scholar51
C. Fernández-de-las-Peñas, J.D. Martín-Guerrero, Ó.J. Pellicer-Valero, et al.
Female sex is a risk factor associated with long-term post-COVID related-symptoms but not with COVID-19 symptoms: the LONG-COVID-EXP-CM multicenter study
J Clin Med, 11 (2) (2022), p. 413
Crossref
View in ScopusGoogle Scholar
52
Y. Zhang, J. Lu, M. Wang, C. Zuo, J. Jiang
Influence of gender on tau precipitation in Alzheimer's disease according to ATN research framework
Phenomics, 3 (6) (2023), pp. 565-575
Google Scholar
53
G.A. Harris, L.R. Hirschfeld, M.I. Gonzalez, M.C. Pritchard, P.C. May
Revisiting the therapeutic landscape of tauopathies: assessing the current pipeline and clinical trials
Alzheimers Res Ther, 17 (1) (2025), p. 129
View in Scopus
Google Scholar
Cited by (0)
© 2025 The Author(s). Published by Elsevier B.V.
Recommended articles

    Predicting political beliefs with polygenic scores for cognitive performance and educational attainment
    Intelligence, Volume 104, 2024, Article 101831
    Tobias Edwards, ..., James J. Lee
    Intravenous SARS-CoV-2 Spike protein induces neuroinflammation and alpha-Synuclein accumulation in brain regions relevant to Parkinson's disease
    Brain, Behavior, and Immunity, Volume 129, 2025, pp. 102-123
    Cara Sophie Schreiber, ..., Franziska Richter

All content on this site: Copyright © 2026 Elsevier B.V., its licensors, and contributors. All rights are reserved, including those for text and data mining, AI training, and similar technologies. For all open access content, the relevant licensing terms apply.
-----------------------
[*/quote*]

[AribertDeckers]

Source #2:


There is a truck-load of graphics and images in the original. To view these, do read the original!

The PDF of the article:

https://iopscience.iop.org/article/10.1088/1758-5090/adf66c/pdf


https://doi.org/10.1088/1758-5090/adf66c
==>
https://iopscience.iop.org/article/10.1088/1758-5090/adf66c

[*quote*]
-----------------------
Paper • The following article is Open access

Metabolic adaptation and fragility in healthy 3D in vitro skeletal muscle tissues exposed to chronic fatigue syndrome and Long COVID-19 sera

Sheeza Mughal*, Félix Andújar-Sánchez, Maria Sabater-Arcis, Glória Garrabou, Joaquim Fernández-Solà, Jose Alegre-Martin, Ramon Sanmartin-Sentañes, Jesús Castro-Marrero, Anna Esteve-Codina, Eloi Casals, Juan M Fernández-Costa* and Javier Ramón-Azcón*

Published 8 August 2025 • © 2025 The Author(s). Published by IOP Publishing Ltd
Biofabrication, Volume 17, Number 4
Complex Human Model Systems: From Development through Translation in Pharma
Citation Sheeza Mughal et al 2025 Biofabrication 17 045006DOI 10.1088/1758-5090/adf66c
Download Article PDF
Download PDF
Article metrics

10095 Total downloads
Article has an altmetric score of 583
Submit
Submit to this Journal
Share this article

Abstract

Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) and Long Covid-19 (LC-19) are complex conditions with no diagnostic markers or consensus on disease progression. Despite extensive research, no in vitro model exists to study skeletal muscle wasting, peripheral weakness, or potential therapies. We developed 3D in vitro skeletal muscle tissues to map muscle adaptations to patient sera over time. Short exposures (48 H) to patient sera led to a significant reduction in muscle contractile strength. Transcriptomic analysis revealed the upregulation of protein translation, glycolytic enzymes, disturbances in calcium homeostasis, hypertrophy, and mitochondrial hyperfusion. Structural analyses confirmed myotube hypertrophy and elevated mitochondrial oxygen consumption In ME/CFS. While muscles initially adapted by increasing glycolysis, prolonged exposure (96–144 H) caused muscle fragility and weakness, with mitochondria fragmenting into a toroidal conformation. We propose that skeletal muscle tissue in ME/CFS and LC-19 progresses through a hypermetabolic state, leading to severe muscular and mitochondrial deterioration. This is the first study to suggest such transient metabolic adaptation.

Export citation and abstractBibTeXRIS
Previous article in issue
Next article in issue

Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 license. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Supplementary data
1. Introduction

Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), also known as systemic exertion intolerance disease, is a heterogeneous, multisystemic, chronic condition which induces physical and cognitive impairment. Despite the condition affecting approximately 17–24 million people worldwide, its underlying pathomechanism remains poorly understood [1, 2]. ME/CFS predominantly affects women and is characterized by exercise intolerance, post-exertion (mental, emotional, and physical) symptom exacerbation (PESE), or post-exertional malaise (PEM) [3–5]. Some also report heightened sensitivity to sound, light, and chemicals. Minor activities such as brushing teeth or the anticipation of a hospital visit may trigger severe symptoms, forcing long recuperation times. ME/CFS patients (labeled as CFS in figures), often present with a history of viral or bacterial infection or a predisposition to cancer [6]. Due to non-specific symptoms, patient to patient variability and an overlap of symptomology with other conditions such as multiple sclerosis (MS), irritable bowel syndrome, fibromyalgia and some types of cancer, misdiagnosis is common. Clinical heterogeneity, is therefore, significant and diagnosis is ultimately dependent on exclusion based on patient-self reporting. At least 20 different diagnoses criteria exist which change frequently and are not unanimously approved till date [7]. Patients diagnosed with COVID-19 also report similar symptoms such as impaired exercise tolerance, extreme fatigue and PESE/PEM after both physical and mental exertion. This has been termed as the post-acute sequelae of COVID-19 or Long COVID-19 (LC-19). Recent studies confirm the various overlapping features between ME/CFS and LC-19 [4, 8, 9].

Peripheral fatigue, a decline in muscle's ability to contract and generate force due to internal causative factors and independent of changes in the brain or spinal cord, is one of the many hallmarks of ME/CFS and LC-19. Patients report high fatiguability which is the tendency of their muscles to tire and lose strength quickly, exercise intolerance and an overall reduced exercise capacity confirmed by clinical studies [3, 4]. Multiple theories exist in contemporary literature attempting to offer various explanations. The Energy Envelope Theory proposes that patients have a limited energy capacity threshold, which when exceeded leads to a worsening of symptoms. This explanation implicates impaired mitochondrial function in ME/CFS patients which compromises their capability for aerobic respiration [10]. The alternative, anaerobic respiration, is less efficient, not only yielding less ATP per glucose molecule but also leading to an accumulation of lactic acid. This results in a decrease in physiological pH, potentially impairing muscle function and inducing fatigue [11]. The push-crash cycle explanation suggests that CFS patients experience a burst of adrenaline and energy causing them to overexert and eventually crash [12]. Studies have also implicated hyperactivation of inflammatory processes and cytokine storms to explain PESE/PEM [2, 8, 13]. Dysfunctional mitochondrial oxidative phosphorylation (OXPHOS) and metabolic disturbances are till date the most widely agreed upon explanations for peripheral fatigue. Other explanations include changes in the composition of systemic factors and the presence of autoantibodies [14–16]. It is evident that investigating metabolic alterations can help in identifying misallocation of energy or metabolic plasticity to manage against stress which could potentially lead to fatigue and oxidative stress [17].

In the last few years, studies on these two idiopathic conditions are predominantly based on blood, muscle biopsies, and cardiopulmonary exercise (ergometry testing) for moderate to severely ill patients [7, 18]. While important, these studies offer limited to no insight on the mechanism of disease advancement but instead, focus on presenting the clinical picture at a timepoint. The major drawback of these studies is, therefore, an inability to address the issue of progressive muscle wasting which is often wrongfully attributed to only physical inactivity rather than a direct effect of the disease. In patients, understanding this decline in skeletal muscle performance can be difficult due to heavy reliance of current testing methods on biopsies, graded exercise testing and the high probability of false positives at an early stage of disease onset. These approaches also threaten patient safety, comfort, and risk the development of PEM/PESE for those being frequently tested.

[AribertDeckers]

Conventional muscle models, such as monolayer cultures are not representative of the complex in vivo skeletal muscle structural and functional features, making them untranslatable [19]. Similarly, data from rodent muscle models has limited reproducibility in humans, with many of the drugs tested in rodent models failing clinical trials [20]. This lack of suitable in vitro models to evaluate and reproduce PEM/PESE presents a critical gap in biomedical research on ME/CFS and LC-19. Recent advances in 3D bioengineered in vitro skeletal micro-physiological systems replicate patient-specific physiological responses and enable the evaluation of pathophysiological responses to circulating systemic factors, such as autoantibodies, circulating cytokines and metabolic or redox toxins present in patient sera. These systems offer a non-invasive testing platform to accurately model conditions without bringing any discomfort to the patient.

Our work aims to understand the pathomechanism of two idiopathic conditions: ME/CFS and LC-19 by using 3D skeletal muscle tissues developed from immortalized human muscle progenitor cells. These mature and well-differentiated tissues were then exposed to sera from ME/CFS, LC-19 and healthy donors for short and long exposures. Exposing tissues to patient sera allowed us to deliver actual systemic insults to muscles in a controlled environment without confounding factors such as physical deconditioning. The contractile profile of the muscle was recorded through electric pulse stimulation (EPS) and transcriptome of treated tissues was analyzed through total RNA sequencing. Subsequently, we quantified mitochondrial morphology and function. Our findings suggest that exposure of 3D skeletal muscle tissues to sera from ME/CFS and LC-19 patients induces significant contractile dysfunction. Both conditions demonstrate mitochondrial dysfunction and transcriptionally adapt distinct responses against sera exposure. ME/CFS muscles, in particular, show elevated oxygen consumption, proton-leak and ATP-linked respiration. Metabolically, both diseases shift muscle energy production towards alternative processes as compensation for mitochondrial defects. Overall, skeletal muscle degeneration in these conditions follows a multi-phasic progression from compensatory adaptation to structural and metabolic collapse driven by mitochondrial impairment and disrupted protein turnover.
2. Results
2.1. Tissues exhibit contractile weakness at short exposure to patient sera

Impaired power output of skeletal muscles and extreme fatigue is frequently reported by ME/CFS and LC-19 patients [21, 22]. Among the multiple theories existing in contemporary literature, the role of systemic factors in inducing multi-organ complications is also widely accepted. To this end, we fabricated 3D in vitro skeletal muscle tissues from healthy immortalized human myogenic progenitor cells as discussed before [23] to evaluate their contractile performance in response to patient and control sera at short exposure (figure 1(A)). All tissues were well-differentiated with long, multinucleated myotubes aligned perpendicular to the two pillars prior treatment. The contraction regime included progressively increasing stimulation frequencies to induce both twitch and tetanic contractions (figure 1(B)). The voltage was kept constant at 10 VPP (peak to peak voltage) with 10 s of relaxation after each 10 s stimulation period.

Figure 1. Comparative contractile dynamics post sera treatment for 48 h. (A) Schematic of 3D in vitro skeletal muscle tissue maturation and treatment. (B) Stimulation regimen and profile. (C) Maximum contractile force of tissues post treatment. (D) Retention time during maximum contraction, time in peak (TIP). (E) Contractile velocity. (F) Power of contraction (G) twitch spectrum during 1 Hz stimulation. (H) Maximum force during 1 Hz stimulation. (I) Tetanic spectrum during 50 Hz stimulation. (J) Maximum force at 50 Hz. (K) Sample distribution with peak forces at 25 and 50 Hz (L) time taken for the force to drop to 50% of peak value during sustained tetanic stimulation at 50 Hz (T50%). (M) Parameters from tetanic contractions to measure contractile velocity. Statistical analyses: biological replicates: CFS and control, n = 4; LC-19 n = 5 sera from patients or donors. Technical replicates: n = 3–7 tissues per serum. Data show the mean ± s.d. Statistical analysis: one-way ANOVA with Tukey's post hoc test *P ⩽ 0.05; **P ⩽ 0.01.

Download figure:
Standard image
High-resolution image

Post 48 h treatment, the tissues exposed to patient sera, henceforth named as diseased tissues, exhibited a significant drop in maximum contractile force (figure 1(C)). Although statistically insignificant, the CFS tissues had a much lower contractile force compared to LC-19 diseased tissues. time in peak (TIP) is the ability of a tissue to maintain peak performance during stimulation. The TIP for diseased treatment groups is significantly shorter compared to the healthy controls (figure 1(D)). Power and velocity of contractions help interpret the ability of a muscle to execute work (figures 1(E) and (F)). These parameters were significantly diminished for all the diseased tissues, albeit more for CFS samples. Consequently, at a higher frequency of 50 Hz, these tissues demonstrated significantly compromised force of contraction as indicated by a much shorter time to drop to 50% relaxation (figure 1(L)). A functional muscle is capable to sustain an increase in its force of contraction proportionally with the stimulus until a certain limit is reached, beyond which depletion of the energy reserves or changes in structural integrity induce impairment [24]. For LC-19 diseased tissues, this limit was reached at a lower stimulating frequency compared to the control and CFS samples. At 25 Hz, a majority of LC-19 tissues appeared to show peak performance following which they could no longer maintain their maximum force against the stimulus (figure 1(k)).

These findings indicate that patient sera not only lowered the overall ability of the muscle to execute work but post peak performance (figure 1(E), (F) and (M)), the tissues experienced weakness and inability to reach the same performance again, particularly for LC-19 diseased tissues. The model, at short exposure, therefore, replicates muscle weakness and compromised contractile performance frequently experienced by patients. It also hints at the applicability of the push-crash cycle theory on diseased tissues.
2.2. Transcriptomics reveal metabolic plasticity, drop in mitochondrial fission and altered calcium homeostasis at short exposure

To investigate the underlying mechanistic changes in diseased tissues that caused contractile impairment, transcriptomic analyses of the short exposure tissue samples was performed using total RNA sequencing. Interestingly, the multidimensional scaling (MDS) analyses indicated that both ME/CFS and LC-19 treatments clustered together with no significant differentially expressed genes (DEGs) (figure 2(A)). There were, however, several DEGs when ME/CFS and LC-19 tissues were compared to Controls (figures 2(B) and (C)). We further performed gene ontology (GO) enrichment analyses between both patient groups and controls, to identify dysregulated pathways and gene set enrichment anlaysis (GSEA) between ME/CFS and LC-19 to identify any underlying differential biological tendencies. Genes of interest were filtered based on Fold Change and adjusted p-values for both up and downregulated genes (FC > ±1, and adj p-value < 0.05). Expression of genes with maximum fold change identified by total RNA sequencing was then validated by qRT-PCR. The findings from GO enrichment analyses (figure 6) indicated key changes in muscle structure and contractile function along with cellular metabolism, particularly protein synthesis and translation in ME/CFS while LC-19 samples had an upregulation in mitochondrial structure and function as well as protein translation. The GO enrichment analysis bubble plots provide a comprehensive profile of differential gene expression and regulation patterns (figure 3).

Figure 2. RNA-Seq-MDS and differential expression analyses (A) a multidimensional scaling map for samples. (B) Volcano plot of differentially expressed genes (DEGs) between ME/CFS and control. (C) Volcano plot of differentially expressed genes (DEGs) between LC-19 and control. (D) Volcano plot of differentially expressed genes (DEGs) between ME/CFS and LC-19.

Download figure:
Standard image
High-resolution image

Figure 3. Gene ontology enrichment analysis. (A) CFS vs CNT GO: bioprocesses (B) CFS vs CNT GO: cellular components (C) CFS vs CNT GO: molecular function (D) LC-19 vs CNT GO: bioprocesses (E) LC-19 vs CNT GO: cellular components (F) LC-19 vs CNT GO: molecular function. N = 3 (3 biological replicates and 3 technical replicates per biological replicate were used). Threshold: log2FC ⩾ 1 and adjusted p value ⩽ 0.05.

Download figure:
Standard image
High-resolution image

ME/CFS vs control: GO biological processes, cellular components and molecular functions show significant enrichment for upregulated processes particularly in translation, contraction and developmental processes (figures 3(A)–(C)). Prominent upregulated cellular components included ribosomal structures (cytosolic ribosome, large ribosomal subunit) and muscle-related structures such as sarcomere, myofibril, I band, and contractile muscle fiber (figure 3(B)). Upregulated molecular functions in CFS showed a distinct pattern focused on structural components, particularly extracellular matrix (ECM) elements. These include ECM structural constituent conferring tensile strength with high enrichment (FE = 9.48), ECM structural constituent, structural constituent of ribosome, and structural molecule activity (figure 3(C)). According to this GO profile, in ME/CFS skeletal muscle samples there is an upregulation of genes involved in protein translation, ECM, and developmental processes, while genes involved in basic metabolic, transcriptional, and organelle functions are generally downregulated. This could indicate an environment characterized by chronic stress, impaired metabolic activity, and ongoing or maladaptive tissue remodeling.

LC-19 vs control: similar to ME/CFS vs control comparison, a broad downregulation was observed for genes associated with cellular and nuclear functions. In contrast, upregulated terms included mitochondrial structures and ribosomal subunits, such as the mitochondrial inner membrane, mitochondrial envelope, and cytosolic ribosome (figures 3(D) and (E)). The molecular function (GO:MF) results further supported this tendency. Downregulated terms include binding activities, such as protein, nucleotide, and anion binding, as well as structural molecule activity with moderate enrichments (figure 3(F)). This pattern suggested a general reduction in the molecular interactions and structural roles. In contrast, the upregulated molecular functions included long-chain fatty acyl-CoA dehydrogenase activity and electron transfer activity. The structural constituent of ribosome and oxidoreduction-driven active transmembrane transporter activity are also upregulated. These findings indicate a pronounced shift in LC-19 tissue toward enhanced mitochondrial fatty acid metabolism, electron transport, and protein synthesis processes.

ME/CFS vs LC-19: although there was a clear absence of significant DEGs between ME/CFS vs LC-19, we were interested to investigate if there were any subtle underlying biological trends or tendencies. Gene set enrichment analysis comparing ME/CFS and LC-19 revealed some interesting differences (figure S1). In ME/CFS, there was significant upregulation of pathways related to ECM organization, including integrin cell surface interactions, collagen degradation, and elastic fiber formation, indicating tissue remodeling and structural changes. Alternatively, compared to LC-19, ME/CFS also showed strong downregulation of mitochondrial energy production pathways, such as the citric acid (TCA) cycle, respiratory electron transport, and ATP synthesis by chemiosmotic coupling, as well as pathways involved in DNA synthesis and neuronal signaling. These findings highlight a pattern of impaired cellular energy metabolism and enhanced ECM remodeling in ME/CFS compared to LC-19. Our gene set enrichment analysis, further suggested that all compensatory pathways between ME/CFS and LC-19 were upregulated in the latter compared to the former except for cellular response to starvation (figure S2).

Mitochondrial complexes and dynamics:

Gene expression data from RNA sequencing indicated that transcript levels corresponding to Mitochondrial Respiratory Chain complexes were dysregulated in diseased groups, highlighting a compensatory response (figure 4(I)). The upregulation was comparatively more for the LC-19 samples compared to ME/CFS. The same trend was observed for mitochondria encoded genes and those encoded in the nuclear genome. Mitochondrial dynamin related protein 1 (DNM1L) transcript levels were downregulated indicating a drop in mitochondrial fission (figure 3(I)). This observation in conjunction with the upregulation of Mitofusin-2 (MFN2) and SET and MYND domain containing 1 (SMYD1) signposts towards adaptive changes in mitochondrial performance in favor of mitochondrial network fusion. Furthermore, upregulation in TCA cycle and glycolytic gene expression for ME/CFS and LC-19 compared to controls indicates cellular adaptation against increased energy demands. In the RNA sequencing analysis, we also observed indication of mitochondrial apoptosis through the upregulated expression of AIFM1 (apoptosis-inducing factor, mitochondria associated 1) and ENDOG (endonuclease G). AIFM1 is involved in caspase-independent apoptosis and translocated from the mitochondria to the nucleus upon apoptotic stimulation. It has also been implicated in maintaining mitochondrial OXPHOS [25]. Similar to AIFM1, ENDOG is also a pro-apoptotic mitochondrial protein. We also observed a downregulation in high temperature requirement A2 which suggested compromised mitochondrial protein quality control and accumulation of damaged or misfolded proteins [26].

Figure 4. qRT-PCR relative gene expressions of 48 h tissue cohort. (A) SMYD1 (B) ATP2A1( C) FHL1 (D) ENO3 (E) PDK4 (F) PDK3 (G) NOG and (H) ESR2. (I) Heatmap of differentially expressed genes from RNASeq. Statistical analyses: biological replicates: ME/CFS and control, n = 3 sera per condition. Technical replicates: n = 4 tissues per serum for all analyses. Data show the mean ± s.d. Statistical analysis: one-way ANOVA with Tukey's post hoc test *P ⩽ 0.05; **P ⩽ 0.01.

Download figure:
Standard image
High-resolution image

Validation of DEG expression

A lysine methyl transferase specific for striated muscles, SMYD1 plays an important role in muscle differentiation and mitochondrial bioenergetics including stabilization of respiratory complexes and cristae formation [27, 28]. It was found to be upregulated in ME/CFS and LC-19 tissues compared to the controls (figures 4(A) and (I)). SMYD1 gain-of-function has been previously associated with upregulation of mitochondrial respiration as a protective mechanism against injury [27]. Furthermore, the levels of ATP2A1 which codes for SERCA1 Ca2+-ATPase were also upregulated (figures 4(B) and (I)). Predominantly present in type II fast twitch muscle fibers, this pump is a key regulator of relaxation dynamics of a striated muscle. Found in the sarcoplasmic reticulum (SR) of muscle cells it pumps the calcium ions from the cytoplasm into the SR, allowing muscle relaxation post contraction and restocking ions for the next contraction [29]. Downregulation in ATP2B4 (calcium ATPase isoform 4) observed in RNA-Seq, (responsible for removing intracellular calcium ions against the large gradients), indicates high calcium sequestering (figure 4(I)). Coupled with an upregulation in ATP2A1 levels, this indicates a disturbance in calcium homeostasis. Increase of calcium sequestering by ATP2A1 in the SR has been related to a disturbance in mitochondrial function and an inducer of fatigue. Increased mitochondrial calcium can trigger the production of ROS through the electron transport chain, particularly the levels of superoxide radicals [30, 31]. Downregulation of SOD2 (superoxide dismutase 2) was observed in diseased tissues, confirming the previously reported data [8]. Coupled with an increase in calcium sequestering, a decrease in SOD2 levels could signal oxidative stress (figure 4(I)).

Four-and-a-half LIM protein 1 (FHL1) has been implicated in inducing myotube hypertrophy [32]. It was found to be overexpressed in the diseased tissues (figures 4(C) and (I)). We also observed an upregulation in transcript levels of glycolytic enzymes such as AGL which codes for glycogen debranching enzyme and ENO3 (beta Enolase) (figures 4(D) and (I)). Both are responsible for regulating the glycolytic pathway and controlling the levels of glucose. Pyruvate dehydrogenase kinase 4 (PDK4) is known to be an important metabolic regulator which increases fatty acid oxidation (figures 4(E) and (I)) [33, 34]. A shift towards fasting-type energy metabolism has been associated with an overexpression of PDK4 [35]. ME/CFS and LC-19 diseased tissues showed an increase in PDK4 gene levels indicating metabolic plasticity under stress conditions warranting a high energy demand. The levels of PDK3, however, were downregulated favoring glucose oxidation (figures 4(F) and (I)). Aberrant PDK levels have previously been 1ssociated with metabolic adaptation in ME/CFS patients [36]. Downregulation of the NOG (noggin) gene has been associated with an increase in bone morphogenetic protein signaling pathway implicated in inducing skeletal muscle hypertrophy which we observed in both LC-19 and ME/CFS treatments (figures 4(G) and (I)) [37]. Estrogen receptor type 2 or ESR2 is known to regulate skeletal muscle growth, regeneration through activation of satellite cells, activating anabolic pathway, and metabolic homeostasis. Loss of ESR2 expression negatively impacts all these processes [38–40] (figures 4(H) and (I)).
2.3. Structural analyses indicates hypertrophy, mitochondrial hyperfusion and elevated oxygen consumption capacity

To further investigate our transcriptomic findings, we undertook structural analyses of myotubes from our short exposure sample cohort and the mitochondria therein. Quantification of myotube diameter was performed by calculating Feret's diameter for individual, transversely cut tubes. The diameter appeared to be enlarged compared to the controls indicating hypertrophic tendencies in diseased ME/CFS and LC-19 tissues (figures 5(A) and (B)). Hypertrophy has been evidenced to be a compensatory response against stress which include self-repair processes [41]. Moreover, quantification of mitochondrial networks showed hyperfusion evidenced by increased mitochondrial branching and mean branch length (figures 3(C)–(E)). Mitochondria had a high aspect ratio and appeared to be hyperbranched in the cytoplasmic space across the length of a myotube as well as close to the nuclei (figure 5(E)). Fusion has been considered to be a positive response related to mitochondrial health, but evidence suggests that above control levels, excess fusion equates to an increased stress response [42–44]. This coupled with an upregulation in pathways of mitochondrial protein elongation, OXPHOS and energy metabolism observed for ME/CFS and LC-19 in RNA Sequencing (figure 3) confirmed the mitochondrial compensatory response against stress.

Figure 5. Structural and functional phenotypic validation. (A) Representative confocal images of transverse cross sections stained with an antibody against the mature muscle marker sarcomeric actinin (SAA), phalloidin (F-actin) and DAPI (nuclei). Scale bar: 40 µm. (B) Quantification of myotube diameters n = 3. (C) Quantification of mitochondrial branch length and (D) average network branches n = 3. (E) Representative confocal images of mitochondrial networking stained with inner mitochondrial membrane marker TOMM20 and DAPI. Scale bar = 10 µm (F) profile for oxygen consumption rate (OCR) from the MitoStress test. (G) Extracellular acidification rate (ECAR) profile. (H) Basal oxygen consumption rate (I) proton leak OCR (inset EACR) (J) maximum oxygen consumption rate (inset EACR at maximum respiration). Statistical analyses: biological replicates: CFS and control, n= 3 patients or donors. Technical replicates: n = 4 per serum. Data show the mean ± s.d. Statistical analysis: one-way ANOVA with Tukey's post hoc test *P ⩽ 0.05; **P ⩽ 0.01.

Download figure:
Standard image
High-resolution image

We were further intrigued to check mitochondrial functional capacity by the MitoStress Test and observed an increase in the overall oxygen consumption rate (OCR) by the diseased cells compared to the controls (figure 5(F)). The extracellular acidification rate (ECAR) measures the glycolytic process in the cells as a response to treatments. Both the OCR and the EACR were the highest in ME/CFS patients compared to the other two groups (figure 5(G)).

Using this data, we further quantified mitochondrial functions to get deeper insights into respiration and glycolysis after 48 h of patient serum exposure. The basal respiration for ME/CFS samples was substantially higher as was the proton leak post Oligomycin induced blockage of complex V (figures 5(H) and (I)). A corresponding increase in ECAR suggested a significantly higher increase in Glycolysis compared to LC-19 and Control groups (figure 5(I)). A similar trend was observed during the introduction of FCCP, an uncoupler that allows to quantify the maximum OCR that the mitochondria are capable of achieving (figure 5(J)).

Other parameters like ATP-linked respiration, non-mitochondrial respiration and reserve capacity were also slightly upregulated for ME/CFS samples (figures S4(A)–(C)). Non-mitochondrial respiration is attributed to oxygen consumed by non-mitochondrial cellular enzymes and processes, predominantly those that originate from pro-oxidant and pro-inflammatory enzymes such as cyclo-oxygenases, cytochrome P450s or NADPH oxidases [45]. To investigate this further we calculated the proton production rate (PPR) and the rate of coupled oxygen consumption (oxygen consumption coupled to ATP production) according to the protocol outlined by Mookerjee et al [46], at 48 h of serum treatment, there is an increase in total oxygen consumption, of which the coupled OCR is the highest in tissues treated with ME/CFS (figure S4D) samples followed by the controls and LC-19. To check the dynamics within each treatment we normalized the rates of all mitochondrial processes to basal respirations and confirmed that in LC-19 samples, nonmitochondrial respiration is consuming most of the oxygen (figure S4(F)). However, for ME/CFS samples all processes appear to be upregulated signaling not only a heavy energy burden but also perhaps an attempt to maintain muscle contractile force.
Tissue weakness and fragility increases with exposure along with mitochondrial fragmentation

At 48 h, diseased tissues exhibited a high compensatory response against the systemic stress factors. Based on this data we decided to conduct some initial tests to determine how the damage manifests and progresses phenotypically during prolonged exposure to patient sera. For this experiment, we used tissues obtained from the same batch of encapsulation to avoid variability in handling and exposed them to LC-19 and control sera for 48, 96 and 144 h. The diseased tissues were weaker as evidenced by a lower T50% compared to the controls (figure 6(A)). Tissue survival decreased sharply with time for both diseased and control groups, but the decline was two-fold higher for the former than the latter. Brightfield images of the tissues showed distorted LC-19 tissues compared to the controls (figure B). Furthermore, quantification of myotube diameters showed progressive atrophy, with significantly decreased diameters over time (figures 6(D) and (E)). We then analyzed mitochondrial morphology and quantified mitochondrial networks for each time point. Our data indicated a decline in mitochondrial branching and mean branch length. The mitochondria not only assumed the familiar globular geometry observed during fragmentation but also toroidal conformations indicating changes in mitochondrial membrane potential at 144 h of patient serum exposure (figures 6(F) and (G)). The toroidal conformations are typically observed with FCCP administration at high dosages due to depolarization of the mitochondrial membrane, resulting in a drop in mitochondrial membrane potential and uncoupling of mitochondrial OXPHOS and ATP synthesis [47]. These findings confirmed mitochondrial stress induced by systemic stress factors coupled with myotube atrophy at longer exposures. Our results of prolonged exposure to patient sera further warrant investigation due to small sample size but signpost at progressive deterioration of muscle structure, function and mitochondrial energy production, mimicking the conditions observed in patients.

Figure 6. Structural changes over long exposures (A) relative absolute force at 50 Hz for control. The dotted line indicates the time taken for the force to drop to 50% of its peak value under sustained tetanic stimulation of 50 Hz (T50%) (B) relative absolute force at 50 Hz for LC-19 tissues over time. (C) Brightfield images of progressive muscle exposure to control and LC-19 sera. (D) Representative confocal images of transverse cross sections stained with an antibody against the mature muscle marker sarcomeric actinin (SAA), phalloidin (F-actin) and DAPI (nuclei). Scale bar: 20 µm (10 µm inset) (E) quantification of myotube diameters overtime (F) quantification of mitochondrial branch length overtime (G) representative confocal images of mitochondrial networking stained with inner mitochondrial membrane marker TOMM20 and DAPI. Scale bar: 10 µm. Statistical analyses: biological replicates: CFS and control, n = 1. Technical replicates: n = 4 per serum. Data show the mean ± s.d. Statistical analysis: one-way ANOVA with Tukey's post hoc test *P ⩽ 0.05; **P ⩽ 0.01.

Download figure:
Standard image
High-resolution image
3. Discussion

Research on ME/CFS began when a series of outbreaks of unknown etiology were recorded in 1934. The term 'benign myalgic encephalomyelitis' was first used for patients presenting with malaise, swollen lymph nodes, severe muscular pain, and throat problems. The prefix 'benign' was used to indicate that the condition did not cause mortality. In 1970, however, these reports were declared to be psychosocial phenomena caused by mass hysteria because the patients were commonly female, and the disease lacked physical signs. Over time, research revealed that although it did not cause mortality, the disease was severely disabling and the prefix 'benign' was eventually dropped and ME/CFS was adopted as an umbrella term [48].

Since then, extensive research has revealed physiological and metabolic disturbances that induce multisymptomatic complications in the patients. Contemporary research has implicated the immune system, mitochondrial performance, inflammation induced parenchymal damage of the nervous system, disturbances in amino acid metabolism, local hypoxia, and oxidative damage to induce disease onset [5, 6, 18, 22, 48]. The causative factors were speculated to be either a history of bacterial or viral infections, certain therapies, heightened immune response, a history of cancer and circulating systemic factors. The COVID-19 pandemic brought forward a cohort of LC-19 patients who reported symptoms similar to ME/CFS, warranting a thorough investigation into the overlapping features between the two. In summary, a small fraction of the scientific community is now beginning to appreciate and understand that ME/CFS is an illness, but patients often must wait for years before they can receive a formal diagnosis due to a lack of targeted diagnostic tests. The treatment process is also symptomatic and not definitive.

The aspect of PESE or PEM as mentioned before, is crippling for the patients. Controversial experiments on exercise-based interventions and repetitive biopsy extractions are, therefore, dangerous as they risk exacerbating patient symptoms. Taking advantage of the significant leaps in 3D organ and tissue culture technologies, we decided to investigate the impact of systemic factors on the functional performance and structural integrity of skeletal muscles. To this end, we employed immortalized human muscle progenitor cells to fabricate tissues capable of executing contraction dynamics upon electric stimulation. This not only allowed us to evaluate disease manifestation and progression as close as possible to the in vivo environment but also without causing any distress to the patients. Through this experimental plan, we also sought to answer whether the observed muscular weakness is due to deconditioning or whether certain systemic stress factors in the patient serum are responsible for inducing diminution of physical strength.

To investigate this problem, our tissues were treated with sera from ME/CFS, LC-19 and healthy donors as controls for 48 h initially. Prior to treatment however, we checked cellular viability against the sera at 0.5% and 5% (v/v) concentrations and observed no changes in cellular viability in both 2D and 3D cultures. In our 3D tissues, post-treatment we observed that the maximum contractile strength of ME/CFS and LC-19 treated tissues was severely compromised compared to the controls. A distinguishing feature of the LC-19 tissues was that they reached peak performance at lower frequencies (25 Hz) of our stimulation regimen and could not sustain the same strength at a subsequent higher frequency (50 Hz). This observation coupled with a shorter T50% at 50 Hz confirmed contractile weakness. Previous studies have also shown a lower peak power output in LC-19 patients observed through an exercise test on a cycle ergometer, which our study validated through 3D in vitro skeletal muscle tissues [4, 49]. These findings provide strong evidence against the hypothesis that inactivity is the only cause of muscle wasting in ME/CFS and LC-19 patients. The contractile performance of a skeletal muscle can be diminished due to structural damage, lack of creatine reserves, local hypoxia or insufficient production of ATP due to disturbances in mitochondrial function [50]. The process of OXPHOS produces the maximum amount of ATP by reducing oxygen to sustain cellular bioenergetics [51]. In skeletal muscles, a large proportion of this energy is expended to form Actin–Myosin cross-bridges necessary for contraction. The remainder is used for protein synthesis, glycolysis, ionic balance maintenance and calcium homeostasis. A disturbance or burden on any of these processes may result in an energy imbalance [52]. The observed alteration in contractile force warranted a thorough investigation on the underlying metabolic and bioenergetic processes.

Through RNA sequencing we sought to map the comparative transcriptomic profiles of all the treatment groups. We observed a significant increase in protein translation processes both cytoplasmic and mitochondrial, in ME/CFS and LC-19 cohorts compared to the control. The ME/CFS samples also had an upregulation of extracellular remodeling pathways. Alternatively, LC-19 samples had an upregulation of mitochondrial inner membrane components, ETC, TCA and mitochondrial fatty acid metabolism. These patterns indicated fundamentally different anti-stress mechanisms between two similar conditions of ME/CFS and LC-19.

There was an overall increase in mitochondrial OCR and EACR in the diseased cohort indicating an increased energy demand with a high proton-leak. At 48 h of exposure, it appears that the ME/CFS and LC-19 tissues are at the intersection of mitochondrial damage and unbalanced protein turnover. Furthermore, there was also an increase in non-mitochondrial oxygen consumption, which meant that of all the oxygen consumed, a significant portion was probably being used by pro-inflammatory or compensatory processes that levied a high energy burden. ATP also serves as a signaling molecule to recruit cytokines and activate inflammatory processes by binding on purinergic receptors on the immune cells [53]. Multiple studies have reported the infiltration and accumulation of immune cells in skeletal muscles in LC-19 and post-viral ME/CFS [6, 8, 44, 54]. The activation of pro-inflammatory processes can also be strengthened by a high transcriptional expression of antioxidant genes such as SOD1, present in the cytosol and nucleus and SOD3 present in the ECM [55–57]. An intriguing observation, however, was the downregulation of SOD2 which is present in the mitochondria and manages mitochondrial inflammation [8]. The same trend was observed for ME/CFS group as well, indicating oxidative stress as a common denominator between the two as can be observed in figure 4(I). Furthermore, elevated levels of ATP2A1 and downregulated ATP2B4 indicating high sarcoplasmic calcium levels were observed for both diseased groups.

Colosio et al, attributed compromised power output observed in patients to peripheral determinants such as mitochondrial dysfunction [49]. A recent study by Shin et al, evidenced that the SARS-CoV-2 virus enhances mitochondrial metabolism to boost viral propagation [44]. This is accompanied by an increase in mitochondrial fusion. After observing an elevated expression of DNM1 (DRP1) and MFN2 in our RNASeq data, we quantified mitochondrial morphology and detected an increase in mitochondrial fusion despite the large innate heterogeneity in biological replicates. Furthermore, an elevated expression of MTIF2 evidenced an upregulation of mitochondrial protein synthesis and assembly of the OXPHOS complexes. This compensatory upregulation in the expression of mitochondrial genes evidences the presence of a heavy energy burden which required a compensatory metabolic response [58]. The rate of oxygen consumption coupled to ATP synthesis was the lowest in LC-19 cohort, as was the PPR. This indicates that a large part of the consumed oxygen is not being used to maintain contractile strength but instead being invested in an early compensatory adaptation against stress. Mitochondria are double-membered central energy sources that utilize oxygen to produce ATP to sustain cellular bioenergetics. Through the MitoStress test we observed a significant increase in basal mitochondrial respiration as well, indicating a heavy energy burden after 48 h of treatment with sera.

ME/CFS and LC-19 have multiple overlapping features as reported by several studies and confirmed by our RNA Sequencing analyses as well. In the mitochondrial paradigm, however, there do appear to be different trends between the two. Our gene set enrichment analysis suggested that all compensatory pathways between ME/CFS and LC-19 were upregulated in the latter compared to the former except for cellular response to starvation. For ME/CFS pathways concerning ECM remodeling appeared to be upregulated compared to LC-19. At 48 h of exposure, while sera from ME/CFS and LC-19 appears to condition mitochondria towards hyper consumption of oxygen and metabolism, the ATP produced is not sufficient to maintain contractile performance. We hypothesize this adaptive metabolic reprogramming to not be a complete metabolic switch but instead of a transient nature allowing the cell some buffer time before giving way to atrophy.

Continuing LC-19 serum exposure beyond 48 h induced progressive myotube atrophy evidenced by a reduction in myotube diameter. The tissues contractile strength was compromised and could not maintain strength. A majority of the tissues fractured before they were electrically stimulated to monitor their contractile profiles. Moreover, the mitochondria assumed not only the signature globular geometry observed during fragmentation but also adopted a toroidal conformation indicating altered fusion dynamics, dissipation of mitochondrial membrane potential and cytoskeletal detachment. The toroidal or donut formation precedes fragmentation, which we also observed progressively with time [47].

ME/CFS and LC-19 are complicated multi-symptomatic conditions which have severe physical manifestations with a dysregulated energy balance at their core (figure 7). Within a single cell or tissue type various metabolic and homeostatic processes appear to be altered (hyperregulated during short exposures), we speculate that at short exposures, there comes a point in which the biosystem attempts to resist against stress but when the stress is prolonged, the resistance gives in, allowing rapid deterioration. Our research presents the first 3D in vitro model to study performance profiles of ME/CFS and LC-19 patients without distressing them directly. It also provides evidence against two popular misconceptions of hypochondriasis and muscle wasting due to inactivity. While our investigative insights provide pivotal findings regarding disease progression, we do acknowledge the small sample size. This work not only presents a new avenue for ME/CFS and LC-19 research but also posits an important question regarding the nature of systemic factors that are responsible for inducing this stress response. It further highlights the necessity to probe deeper the mechanistic alterations in both the diseases conditions with the minimum amount of confounding variables. Considering the heterogeneity observed in ME/CFS and LC-19 patients (in symptoms, severity, and underlying triggers), it is important to have investigative tools that can reflect this complexity. Traditional approaches often struggle with high variability, especially in smaller patient cohorts, making it hard to detect meaningful trends. Our 3D muscle model offers a way to explore these differences more directly. While we acknowledge the small sample size, the model not only helps understand how the patient serum affects muscle tissue but also presents an opportunity to investigate the personalized effect of patient serum on otherwise healthy tissues.

Figure 7. Graphical representation of key findings: (A) baseline structure of a myofiber (B) key changes after 48 h of exposure to ME/CFS and LC-19 sera. (C) Progressive atrophy during long exposure to patient sera.

Download figure:
Standard image
High-resolution image
4. Materials and methods
4.1. Study population

This pilot cohort study was conducted involving 4 individuals with ME/CFS (mean age: 49.1 ± 3.1 years), 5 long COVID patients (mean age: 48.3 ± 2.4 years) with ongoing post-exertional fatigue for at least 12 months and 4 age-, sex and BMI-matched non-fatigued sedentary controls (mean age: 46.7 ± 1.8 years) recruited consecutively between September 2021 and December 2022 from the largest outpatient tertiary referral center in Spain (ME/CFS Clinical Unit, Vall d'Hebron University Hospital, Barcelona, Spain) and Hospital Clinic, Barcelona, Spain. After receiving verbal and written information on the study protocol, all subjects signed informed consent before participation. The study protocol was reviewed and approved by the IRB of Vall d'Hebron University Hospital (reference number: PR/AG 201/2016). All procedures were conducted in accordance with the ethical standards of the board and the 1964 declaration of Helsinki for human research including its later amendments.
4.2. Eligibility criteria

ME/CFS patients were potentially eligible if they were female, aged ⩾ 18 years and had a confirmed diagnosis for ME/CFS by a specialist physician based on international consensus criteria (2011 ICC), the recommended diagnostic criteria for ME/CFS research purposes [59]. Long COVID patients were eligible for enrollment if they were female, aged ⩾ 18 years and, following a confirmed diagnosis of acute COVID-19 infection based on a positive SARS-CoV-2 RT-PCR test on nasal swab during the COVID-19 pandemic, were still suffering from persistent fatigue, unexplained PEM and other core symptoms/signs ⩾ 3 months after the acute COVID-19 infection, and met the same clinical case criteria (2011 ICC) for ME/CFS [59]. No-fatigued sedentary donors were eligible if they had neither experienced any post-exertional fatigue/malaise and post-acute infection illnesses-associated symptoms/signs nor had been in recent contact with anyone who was infected with SARS-CoV-2 (COVID-19) within 90 d prior to the study, no significant neurological, cardiac, endocrine, or neuroimmune disorders, no alcohol or drug dependence, and did not use daily prescribed medications. All no-fatigued sedentary subjects were recruited through word-of-mouth from the local community and did not meet the case criteria for either ME/CFS or long COVID at the time of enrollment.

All participants were of Caucasian descent, from the same geographical area, and had a sedentary lifestyle at the time of the study. They were subject to stringent exclusion criteria, as previously described [60]. The major exclusion criteria were a relevant previous or current diagnosis of autoimmune disorder, MS, psychosis, major depression disorder, heart disease, hematological disorders, infectious diseases, sleep apnea or metabolic disorders; pregnancy or breast-feeding; smoking habit; strong hormone-related drugs; and preexisting fatigue-associated symptoms or evidence of multiorgan failure that did not meet the case criteria for ME/CFS and long COVID used in this study.
4.3. Blood collection and processing

After a 12 h overnight fasting, 6 ml of peripheral whole blood were collected from each participant between 8:00 a.m. and 10:00 a.m. by a trained phlebotomy nurse (venipuncture from an antecubital vein with a 19-gauge needle) into SST-tubes for serum isolation (BD Vacutainer, Becton Dickenson, Sarstedt, Barcelona, Spain) in the USIC outpatient clinical unit, Vall d'Hebron University Hospital, and Hospital Clinic, Barcelona Spain. One tube was transported and delivered to the local core laboratory within 2 h of collection for routine blood analyses, following standard and recommended procedures. The other blood tube was immediately centrifuged at 2500 rpm for 15 min at 4 °C (Thermo Scientific, Waltham, MA, USA), after which serum specimens was collected and stored in aliquots at −80 °C until further assays. An aliquot of the serum samples was shipped on dry ice to IBEC laboratory for further studies in the 3D 'in vitro' human skeletal muscle model. No serum sample was thawed more than twice.
4.4. Study design

To understand the pathomechanism and develop an in vitro model of ME/CFS and LC-19 disease manifestation in a skeletal muscle we adopted a sequential experimentation design. This type of study design helps in understanding idiopathic diseases, where each experiment is planned based on the inferences obtained from a previous experiment. In summary, we made 3D in vitro skeletal muscle tissues from immortalized human muscle progenitor cells. After 6 d of differentiation, we treated the mature tissues with 5% patient or control sera for 48, 96 and 144 h and then proceeded with subsequent analyses.
4.5. Fabrication of PDMS platforms

The 3D polydimethylsiloxane (PDMS) platforms were fabricated as published before [61]. Briefly, Master molds were 3D printed using SolusProto (Reify3D, CA, USA) with a Direct Light Projection 3D printer (Solus DLP 3D Printer, Reify3D) at low resolution (80 × 45 mm with a pixel distance of 45 microns). The design was converted to a standard tri-angle language format for printing. After removing all traces of uncured resin by thorough washing and cleaning with ResinAway, the thermo-cured 3D master molds were activated with ozone plasma for 30 s and silanized in a vacuum desiccator with a few drops of trichloro (1H,1H,2H,2H-perfluorooctyl)-silane (PFOTS, Sigma-Aldrich). These 3D printed molds were used to prepare negative molds with PDMS (5:1), which were then silanized after thorough washing. Post silanization, the negative molds were used to fabricate the final casting molds with PDMS (10:1). After fabrication, all the casting molds were thoroughly washed and dried before using for cell encapsulation [62].

Surface of the final casting platforms was treated with filtered 5% Pluronic F-127 (Sigma-Aldrich) in PBS for atleast 2 h at 4 °C to facilitate hydrogel detachment. Pluronic was removed using a filter paper without grazing the bottom of the well. The platforms were stored at 4 °C until tissue fabrication.
4.6. Cell culture

Human immortalized muscle precursor cells were obtained from the Institute NeuroMyoGene, Lyon, France [63]. Specifically, cell line control 2-E4 (CNT2-E4) was used with a doubling time of 3.55 d and 99.40% of CD56+ cells. The cells were cultured in skeletal muscle basal growth medium (SMC-GM, C23060, PromoCell) with supplemental mix (C23060, PromoCell), 1% penicillin-streptomycin (10 000 U ml−1, 15140-122, Thermo Fisher Scientific) and 10% fetal bovine serum (10270-106, Thermo Fisher Scientific). For differentiation, the cells were cultured in Dulbecco's modified Eagle medium (DMEM), high glucose, GlutaMAX (Gibco, Thermo Fisher Scientific), 1% v/v penicillin–streptomycin–glutamine (P/S-G, 100×, Gibco, Thermo Fisher Scientific) and 1% v/v insulin–transferrin–selenium–ethanolamine supplement (ITS-X, 100×, Gibco, Thermo Fisher Scientific).
4.7. Tissue fabrication

An established protocol in our group was adopted to fabricate functional skeletal muscle tissues [23]. Briefly, for encapsulation, the human muscle precursor cells were trypsinized and resuspended in skeletal muscle growth medium. The cells were encapsulated at a density of 2.5 × 107 cells ml−1 in 30% v/v Matrigel Growth Factor Reduced basement membrane matrix (Corning), 2 U ml−1 thrombin from human plasma (Sigma-Aldrich) and 4 mg ml−1 fibrinogen from human plasma (Sigma-Aldrich). During hydrogel casting, care was taken to avoid bubbles and cold plasticware was used to prevent polymerization of Matrigel. The mixture was spread as homogenously as possible between the pillars without grazing the surface. After hydrogel introduction all tissues were incubated at 37 °C for 30 min to allow for matrix compaction before adding skeletal muscle growth medium supplemented with 1 mg ml−1 of 6-amino-caproic acid (ACA, SigmaAldrich). The hence formed tissues were allowed to grow for 2 d after which differentiation was initiated for another 6 d by replacing growth medium with differentiation medium (DM) supplemented with 1 mg ml−1 ACA (DM/ACA). Subsequently, half of the DM/ACA was replenished every 2 d to maintain tissue survival.
4.8. Sera treatment on cell monolayers and 3D in vitro skeletal muscle tissues

Muscle constructs with visible structural abnormalities, such as failure to display hydrogel compaction were excluded. Experiments were performed in triplicate on different days and/or in separate batches to ensure reproducibility of our findings. For construct treatment experiments, serum biosamples were randomly assigned to the control or treatment groups. Investigators performing data analysis were blinded to the study hypotheses. After 6 d of differentiation, tissues were treated with 5% (v/v) of serum supplemented in DM/ACA for 48, 96 and 144 h. The serum was diluted from the stock in DM/ACA and added to the respective cultures.
4.9. EPS and force analyses

Post treatment, the bioengineered skeletal muscle tissues were placed in a 12-well plate with custom-built graphite electrodes connected to a pulse generator (Multifunction Generator WF1974, NF Corporation). An XL S1 cell incubator maintained at 37 °C and 5% CO2 was used to acquire brightfield images of the plate on top of a Zeiss Axio Observer Z1/7 microscope. The tissues were electrically stimulated to induce both twitch and tetanic contractions. Square monophasic waves of 1–50 Hz were used to generate an electrical field strength of 1 V mm−1 and applied for 10 s. The duty cycle was adjusted as needed to ensure all pulses were 1 ms long. Post analyses the cells were fixed and stored.

Videos were recorded for both pillars at 10X and 39 frames per second during stimulation. Contractile parameters were calculated using pillar displacement from origin during deflection. Force, power and contraction velocity were calculated using the following formulae:

where: F is force exerted by the tissue; E is the Young's modulus of the PDMS (calculated to be 1.6 ± 0.1 MPa.); D is diameter of the post; a is location of the tissue on the pillar, L is the height of the post; and d is displacement of the pillar. Spring constant k was calculated using the above parameters to be 3.54 N m−1. V(max) is the maximum contractile velocity [64], TTP is the time in seconds to reach maximum displacement, T(delay) is the delay time in seconds between each stimulation and P is power in µW calculated as shown in figure 1(M).

T50% is the time it takes for the force to drop to 50% of its peak value during sustained tetanic stimulation of 50 Hz. It is calculated as shown in figure 1(M). TIP is the ability of a tissue to maintain peak performance during stimulation or in other words TIP is the duration for which the muscle tissue maintains at least 95% of peak force during a sustained tetanic contraction. TIP has been derived from the force-time plot (figure 1(M))
4.10. RNA extraction, mRNA library preparation and sequencing

Post treatment tissues were flash frozen in liquid nitrogen. Sterilized pestles were used to homogenize the tissues in QIAzol reagent (QIAGEN), and the total RNA was extracted using the miRNeasy Micro Kit (QIAGEN) according to the manufacturer's instructions. Total RNA content was assessed using Nanodrop ND-1000 full spectrum spectrophotometer. Technical replicates were pooled, and the total RNA concentration was adjusted to 100 ng µl−1, approximately 1.3 µg of total RNA per biological replicate or serum. The quantity and quality of the total RNA sample were assessed using the Qubit RNA BR Assay kit (Thermo Fisher Scientific) and the RNA 6000 Nano Bioanalyzer 2100 Assay (Agilent).

RNA-Seq libraries were prepared with the Illumina Stranded Total RNA Prep with Ribo-Zero Plus (Illumina), following the manufacturer's recommendations, starting with 0.5 μg of total RNA as the input material. The final library was validated on an Agilent 2100 Bioanalyzer using the DNA 7500 assay.

The libraries were sequenced on the NovaSeq 6000 (Illumina) with a read length of 2 × 51 bp following the manufacturer's protocol for dual indexing. Image analysis, base calling, and quality scoring of the run were performed using the manufacturer's software real time analysis (RTA v3.4.4), followed by the generation of FASTQ sequence files
4.11. RNA-Seq data processing and analysis

RNA-Seq reads were mapped against the human reference genome (GRCh38) with STAR 2.7.8a [65] (ref 1) using ENCODE parameters. Genes were quantified with RSEM 1.3.0 [66] using the gencode 44 annotation. Differential expression was performed with limma [67] using the voom transformation of the counts [68]. For the comparisons LC vs CTL and CF vs CTL the DEGs between untreated-CTL were removed. MDS plots was performed with the limma function plotMDS. Functional enrichment for the up and down-regulated genes was performed separately with g:Profiler [69]. Gene set enrichment analysis was done with the pre-ranked list of genes by the limma t moderated statistic, using FGSEA 1.12.0 [70].
4.12. qRT-PCR validation

qRT-PCR was performed as described before [62]. Briefly, 1 μg of RNA was digested with DNaseI (Invitrogen) and retrotranscribed with SuperScriptII (Invitrogen) using random hexanucleotides. For each biological replicate, qRT–PCR reactions from 10 ng of cDNA were carried out in triplicate using TaqMan probes (applied biosystems). GAPDH, RPLP0 and B2M were used as the endogenous control after comparing their stable expression across different treatment regimens. Thermal cycling was performed using the QuantStudio 5 RT-PCR system (applied biosystems). Relative expression to the endogenous gene and the control group was obtained by the 2 − ΔΔCt method. Pairs of samples were compared using a two-tailed unpaired t-test (α = 0.05), applying Welch's correction when necessary.
4.13. Cryosectioning and immunoflourescence

After treatments or EPS, the tissues were fixed in 10% formalin (approximately 4% formaldehyde) (Sigma-Aldrich). For transversal sections, tissues were incubated in 30% sucrose solution for 48 h to preserve cellular structure against crystallization. The tissues were then embedded in optimal cutting temperature compound (PolyFreeze, Sigma-Aldrich) in plastic Cryomolds® (VWR, PA, USA) using chilled isopentane. Transverse sections (15 µm) were obtained by sectioning with a Leica CM1900 cryostat. The sections were transferred to SuperFrost Plus Adhesion slides (Fisher Scientific) and stored at −20 °C.

A PAP pen (ImmEdge, Vector laboratories, CA, USA) was used to encircle tissue sections to minimize waste of antibodies. Sections were rehydrated using PBS and permeabilized with 0.1% Triton X-100 in PBS (PBS-T) for 15 min. Blocking was performed using 0.3% Triton X-100 with 3% Donkey serum in PBS for 30–40 min at room temperature. The tissue sections were then incubated with primary antibodies (see table S1 for details) in blocking buffer at 4 °C overnight.

After overnight incubation, the sections were washed three times with PBS-T and incubated with fluorophore-tagged secondary antibodies and phalloidin (table S1) in blocking buffer for 60 min at room temperature. Post secondary antibody labeling, three PBS-T washes were done, and sections were mounted using VECTASHIELD plus mounting medium with DAPI (Palex). Sections were secured with coverslips and sealed using a transparent enamel.
4.14. Imaging and quantitative morphological analyses

Sections were analyzed using brightfield and confocal imaging with ZEISS Axio Observer Z1/7 and ZEISS LSM800 microscopes, respectively. 20X and higher magnifications were used to procure images from the confocal microscope. After acquisition, images were processed and analyzed using Fiji/ImageJ. To calculate myotube diameter, SAA or phalloidin and DAPI channels were used. Images were binarized and the free hand tool was used to mark the boundary for obtaining Ferret's diameter (n ⩾ 50 per image, at least three images were procured per tissue).

Mitochondrial network analysis in 2D monolayer cultures was performed using mitochondrial network analyzer (StuartLab) plugin in Fiji/ ImageJ. Care was taken to ensure same scale parameters for quantifying images obtained from different magnifications. For each condition, three technical replicates were imaged. Quantification was then performed using atleast three images from each condition relative to the number of nuclei. To evaluate mitochondrial morphology, mitochondrial morphometry macro was used in Fiji/ImageJ (S4).
4.15. MitoStress test—seahorse analysis

Immortalized human muscle progenitor cells were seeded at a density of 4000 cells per well in Seahorse XF HS Mini cell culture plates. Cells were allowed to proliferate until the formation of a uniform monolayer with little to no empty spaces. Once confluent, the growth medium was replaced with DM for 6 d after which 5% serum treated was initiated for 48 h. Once the 48 h treatment was completed, the media was removed, and cells were washed with PBS thrice. The standard assay media was prepared by supplementing Seahorse XF DMEM Media with 10 mM Glucose, 1 mM pyruvate and 2 mM L-Glutamine. Subsequent analyses and treatment steps were followed from the Seahorse XF Cell Mito Stress Test Kit user guide. The concentrations for Oligomycin (1 uM and 2 uM) and FCCP (1.5, 1.8 and 2 uM) were optimized on untreated cells. Final concentrations of the treatments were 1.5 uM of Oligomycin, 1.5 uM of FCCP and 0.5 uM of Rotenone/antimycin A. Post experiment, the cells were lysed in RIPA Buffer for protein quantification and normalization. Detailed equations are presented below (table 1) and a figure schematic explaining drug injections has been shown in figure S6:

Table 1. Equations used in the seahorse analysis.
Parameter    Equation
Non-mitochondrial    Minimum rate measurement after Rotenone/antimycin A injection
Basal respiration    (Last rate measurement before first injection)—(non-mitochondrial respiration rate)
Maximal respiration    (Maximum rate measurement after FCCP injection)—(non-mitochondrial respiration)
H+ (Proton) leak    (Minimum rate measurement after Oligomycin injection)—(non-mitochondrial respiration)
ATP-linked respiration    (Last rate measurement before Oligomycin injection)—(minimum rate measurement after Oligomycin injection)
Spare respiratory capacity    (Maximal respiration)—(Basal respiration)
4.16. Statistical analyses

Statistical analyses were performed using Prism 9.5 software (GraphPad). All data were tested for normality by using the Shapiro–Wilk test (P ⩾ 0.05). The comparisons between more than two groups were performed by one-way ANOVA. For two groups, unpaired two-tailed parametric Student's t-test with Welch's correction was applied, whereas for non-normal data sets, unpaired two-tailed non-parametric Student's t-test with Mann–Whitney U test was performed. In the case of ANOVA, if a significant trend was observed, Tukey's post hoc tests with 95% confidence interval were applied (P < 0.05)
Acknowledgments

We extend gratitude to Dr Benedicte Chazaud from the Institut NeuroMyoGène, Lyon, France, for providing us the human immortalized muscle precursor cells and to the MicroFabSpace and Microscopy Characterization Facility, Unit 7 of Unique Scientific and Technical Infrastructures (ICTS) 'NANBIOSIS' from Networking Biomedical Research Centre-Bioengineering, Biomaterials and Nanomedicine (CIBER-BBN) at the Institute for Bioengineering of Catalonia (IBEC) for their technical support. We would also like to acknowledge the thorough critique and suggestions on this manuscript presented by Dr Adolfo López de Munain Arregui from the Clinical Neurosciences Unit of Policlínica Gipuzkoa, San Sebastian, Spain and Professor Dr Karl Morten, University of Oxford John Radcliff Hospital, The United Kingdom. Finally, we would also like to express our appreciation to the entire team of Biosensors for Bioengineering group at IBEC for their valuable feedback during the manuscript preparation process.
Data availability statement

All data that support the findings of this study are included within the article (and any supplementary files).
Funding

This work received financial support from the Ministerio de Ciencia e Innovación (Spain; grant PID2022-136833OB-C22 to J.R.-A.).

S.M. was supported by the FPI predoctoral fellowship from Ministerio de Economía y Competitividad (PRE2020-092676, ayudas para contratos predoctorales para la formación de doctores).

Open Access funding provided by Institute for Bioengineering of Catalonia.

The Spanish Ministry of Science, Innovation and Universities through the 'Severo Ochoa' Program for Centres of Excellence in R&D (CEX2023-001282-S), the CERCA Programme/Generalitat de Catalunya (2021-SGR-01495).

Institutional support to CNAG was provided by the Spanish Ministry of Science and Innovation through the Instituto de Salud Carlos III, and by the Generalitat de Catalunya through the Departament de Salut and the Departament de Recerca i Universitats.
Author contributions

Conceptualization: JR-A, JF-C, SM, GG, JC-M, AE-C.

Methodology: SM, FA-S, AE-C, EC.

Investigation: SM, FA-S, GG.

Funding acquisition: JR-A.

Project administration: JR-A.

Supervision: JR-A, JF-C. JF-S, JA-M, RS-S.

Writing—original draft: SM.

Writing—review & editing: JR-A, JF-C, GG, JC-M, AE-C.
Conflict of interest

All authors declare no competing interests.
-----------------------
[*/quote*]