Molecular evidence of autophagy impairment as serum level of ATG13 is elevated in patients with myalgic encephalomyelitis and chronic fatigue syndrome: two independent case-control studies

Myalgic Encephalomyelitis and chronic fatigue syndrome is a multisystem illness characterized with extreme muscle fatigue associated with pain, neurocognitive impairment, and chronic inammation. Despite intense investigation, the molecular mechanism of this disease is still unknown. Here we report two independent case-control studies to demonstrate that autophagy proteins are strongly upregulated in the serum of ME/CFS patients indicative of severe impairment in metabolic events of autophagy. Serum samples collected from two healthy and two age-matched patients were assayed for protein aggregation, screened for autophagy-related factors via an antibody array, quantied with densitometric analyses, and nally reconrmed with ELISA analyses. Based on that double-blinded and gender-balanced study, the levels of ATG13, p62, and alpha-synuclein (α-syn) were found to be consistently elevated in the serum samples of these two ME/CFS patients. Moreover, our microglia-based oxidative stress response study and nitrite analyses indicated that serum samples of ME/CFS patients evoked the production of reactive oxygen species (ROS) and nitrite in human HMC3 microglial cells, whereas neutralization of ATG13 was shown to strongly diminish the production of ROS and nitrite demonstrating the de novo effect bloodborne autophagy factors on inducing the stress response in microglial cells. Collectively, our results indicate that the impairment of autophagy followed by upregulations of autophagy markers especially ATG13 in serum could be a pathological hallmark in ME/CFS.

Chronic in ammation and cognitive de cit with poor attention span were also common pathological hallmarks of this disease. Until now, there is little known about the molecular mechanism of this disease.
Several reports indicate that the elevated levels of autophagy-related proteins in plasma can be directly correlated with the pathogenesis of many metabolic disorders. Upscaled detection of lysosomal membrane protein LAMP-1 has been reported in the pathogenesis of lysosomal storage disease [26,27].
Plasma ATG5 level is elevated in patients with Alzheimer's disease [28]. Beclin-1 and LC3, are observed in the serum in patients with acute ischemic stroke [29]. levels of autophagic markers might be associated with coronary total occlusion and childhood cerebral palsy [30,31]. Upregulation of α-syn was demonstrated in plasma of PD patients [32]. However, it is not known if these autophagy markers are also upregulated in ME/CFS.
To explore the molecular regulations of autophagy proteins, double-blinded antibody array analyses of 20 autophagy-related proteins were performed in the serum samples of two case-control subjects that demonstrated strong elevations of ATG13, p62 and α-syn. The result was further validated with ELISA analyses recon rming that the expressions of ATG13, α-syn, and p62 were indeed strongly upregulated in the serum samples of both ME/CFS patients. Moreover, our in vitro cell culture study clearly demonstrated that the serum-derived ATG13 induced oxidative stress response and nitric oxide production in HMC3 human microglial cells. In summary, our current case study highlights that the molecular impairment of autophagy followed by elevation of autophagy markers in serum samples could be a pathological hallmark in ME/CFS.

Results
Clinical history of two subjects with ME/CFS displays signi cant muscle and cognitive impairments. In our current study, we included a 71 years old white male ME/CFS patient. Sometimes in 1988, when the subject was a 40 years old, he was initially diagnosed with acute u-like symptom. The disease eventually progressed with severe symptomology of sore throat, acute joint pain, swollen cervical lymph nodes, mental fog, memory impairment, extreme dizziness, and severe fatigue leaving the patient bedridden for 90% of the time. Although, the underlying cause behind these symptoms was not determined, but subsequent antigen testing detected positive titers for EBV, HSV1, and CMV infection. In 1991, the patient was admitted to Dr. Peterson's clinic in a cluster of patients with similar clinical symptoms. A single-photon emission computerized tomography (SPECT) scan imaging analyses revealed that there was chronic Myalgic encephalomyelitis with bilateral hypoperfusion.
Another patient in our current study is currently a 67 years old female who had similar initial u-like symptoms in 1988, which rapidly progressed to severe sinus infection with methicillin-resistant staphylococcus aureus (MRSA) infection. Eventually the patient received 6 turbinate surgeries. Although, the subject did not have any sinus problems, but has a history of depression, severe di culty in sleeping (staying asleep speci cally) and experiences fatigue, pain (general), IBS symptomology, severe memory impairment, and major brain fog. Recently the patient experienced extremely high blood pressure, with no noted cardiac dysfunction.
Both patients are classic cases of ME/CFS with extreme muscle fatigue and severe cognitive impairment. They routinely visit Dr. Peterson's clinic for their treatment regimen. Blood collection was performed with their consent and based on the guideline of approved IRB protocol.
Autophagy-driven degradation of metabolically inactive protein is an important quality control process of cellular metabolism that directly regulates the function of muscle tissue [33] and also coordinates cognitive function [34]. Interestingly, both of these biological processes are signi cantly damaged in chronic fatigue syndrome, which intrigued us to study the impairment of autophagy in ME/CFS. Since autophagy impairment can be evaluated with the upscaled expressions of autophagy proteins in serum, next, we wanted to study the expressions of autophagy proteins in the serum of ME/CFS patients.
Estimation of protein aggregation propensities in serum samples of ME/CFS patients. Severe impairment in autophagy might result the increased tendency of protein aggregation in serum [35][36][37]. Therefore, rst, we performed a thio avin T (ThT)-based protein aggregation study in patient serum. The analysis is a uorimetric tracking analysis that generates a sigmoidal evolution of protein from its liner structure to aggregated morphology. This type of kinetic study was previously done[38] in a cell-free system to analyze the kinetics of amyloid-β protein aggregation [39]. In our present study, we adopted similar uorimetric tracking study to explore if the serum proteins of ME/CFS patients have higher propensities to form aggregates. The accelerated rate of aggregation does not necessarily re ect the impairment of physiological mechanism of protein degradation; however, it might indicate that there is a higher tendency to form aberrant protein aggregates in serum samples of ME/CFS patients. Accordingly, serum samples of both 74 years old healthy control (Fig. 1A) and a 71 years old male ME/CFS (Fig. 1B) patient displayed increasing but non-liner pattern of aggregation as indicated with nonlinear tted curves of Florescence intensities (486 nm/450 nm) with the function of increasing time from 0 to 90 minutes of ThT addition. The plotting equation is with goodness of t was determined as R 2 = 0.9780 (~ 1 for liner) for healthy control and 0.9570 for patient. Interestingly, we observed that serum proteins of patient displayed signi cantly faster rate of protein aggregation compared to healthy control as indicated with steeper slope of 70.70 ± 1.956 in patient compared to relatively at slope of 17.72 ± 0.3806 in control (Fig. 1E). Moreover, the signi cance analysis revealed that there is a strong difference of protein aggregation rate in ME/CFS patient [F 29,29 =15.99 ( *** p < 0.0001)] compared to its gender-matched healthy control (Supplementary Table 1). Next, we performed a similar comparison study between a 68 years old healthy (Fig. 1C) female and a 67 years old female ME/CFS patient (Fig. 1D) that resulted a similar outcome of protein aggregation upon ThT addition. In that case-control study, we also observed a signi cantly faster aggregation of serum proteins in female patient compared to healthy control (Supplementary Table 1) [F 29,29 =15.99 ( *** p < 0.0001)]. The non-liner tted curve generated a steeper slope of 59.68 ± 3.005 in patient versus 19.36 ± 1.599 in healthy control (Fig. 1F). Taken together, our ThT assay revealed that serum samples of two ME/CFS patients have higher and faster propensity of protein aggregation compared to their age-matched and gender-balanced control.
Upregulations of autophagy markers in the serum of ME/CFS patients. Upregulations of autophagy markers in serum have been documented in many neurodegenerative and metabolic disorders. To evaluate these markers in ME/CFS, we performed a human autophagy antibody array analysis in the serum of two case-control subjects. Total protein concentration was assessed by Bradford procedure followed by loading of equal amount of proteins (1.4 mg per sample) in array blots encoded with 20 autophagy-related proteins ( Supplementary Fig. 1C). After that, the signal was detected with IRDye800conjugated streptavidin and the image was captured in Li-Cor Odyssey Sa imager as discussed in method section. Surprisingly, we observed that expressions of ATG5, ATG7, ATG13, p62, Rheb, and α-syn were strongly upregulated in serum sample of 71-year-old ME/CFS patient ( Fig. 2B; Supplementary  Fig. 1B) compared to age-matched healthy subject ( Fig. 2A; Supplementary Fig. 1A). The result was further con rmed with a densitometric analyses (Fig. 2C, 2D, 2E, 2F, 2G, and 2H are ATG 5, 7, 13, p62, Rheb, and α-syn respectively) with the signi cance of means calculated based on three independent experiments at ***p < 0.0001. To further con rm the result, we performed ELISA analyses of ATG5, ATG13, p62 and α-syn in serum sample of male ME/CFS patient with commercially available kits as described in method section. The detection sensitivity of these ELISA kits varies from lot to lot and also based on serum concentrations of proteins. Therefore, to demonstrate the accurate detection of these proteins, we adopted a dilution series of serum samples for ELISA-based detection of ATG5, ATG13, p62, and α-syn. Accordingly, a dose responsive non-liner tting curve analysis followed by measuring X intercept demonstrate that 1:4 dilution of serum provides most e cient detection of these serum-derived factors (Supplementary Fig . Although, ELISA is most reliable technology to quantify serum proteins ,sometimes low signal-to-noise ratio of HRP detection method limits the accurate quanti cation [40]. In our manuscript, we introduced a novel near infrared-based ELISA method, which produces extremely sensitive detection of low abundance protein in serum mainly due to the high signalto-noise ratio of near-infrared probe [41]. Upon binding to biotinylated secondary antibody, IRDye-800 conjugated streptavidin produces extremely sharp signal at the end point that was detected in Li-Cor Odyssey Sa plate reader. Evidently, strong green signals representing upscaled abundance of ATG5 ( were observed in serum sample of 71-yearold ME/CFS male compared to low green signals indicative of low levels of these proteins in agematched control. The speci city of the signal was validated with no signal in 2 • antibody control whereas equal loading was veri ed with total albumin in respective serum samples. The result was further quanti ed with a densitometric analysis relative to total albumin concentrations ( Supplementary Fig. 3A). Collectively, these results suggest that serum sample of 71-year-old male ME/CFS patient displayed upscaled expressions of ATG5, ATG13, p62, and α-syn, with maximum difference in ATG13.
Next, we performed similar antibody array analysis in the serum sample of 67-years-old female patient, which indicated signi cant elevations of LC3a, LC3b, p62, α-syn, and ATG-13 proteins ( Fig. 3B: Supplementary Fig. 1D) compared to age-matched healthy control ( Fig. 3A; Supplementary Fig. 1C). The result was further quanti ed with densitometric analyses (Fig. 3C-3G). To nullify the possibility of unequal loading of protein in array, we estimated the total protein concentration with Bradford method [42] and 1.5 mg total protein was loaded on array membrane for each sample. Since antibody array is a semi-quantitative method to estimate serum protein, next we performed a quantitative ELISA analysis of LC3a, ATG5, ATG13, p62 and α-syn. Although, LC3a ELISA analyses (Fig. 3H) did not display any signi cance, we observed strong elevations of ATG13 (Fig. 3I), p62 (Fig. 3J), and α-syn ( Fig. 3K) in 67-years-old female ME/CFS patient compared to healthy control with maximum difference in ATG13 level (***p < 0.001). Results were con rmed after ve independent experiments.
To further con rm, we performed near-infrared ELISA assay for LC3a (Fig. 3L), ATG13 (Fig. 3M), p62 (Fig. 3N), and α-syn ( Fig. 3O) with albumin as loading control. Interestingly, similar to our quantitative ELISA method, no difference in LC3a expression was observed ( Fig. 3L; Supplementary Fig. 3B). However, expressions of ATG13, p62, and α-syn were consistently higher in patient compared to healthy control ( Fig. 3M-3O: Supplementary Fig. 3B). In both case-control studies, we observed that serum levels of p62, α-syn and ATG13 were consistently elevated suggesting that these patients might have signi cant impairment in the formation of autophagosomes.
Evaluation of lysosomal function in ME/CFS patients. Upon enclosure in autophagosomes, cellular components are directed and fused to lysosomes for hydrolytic degradation [43]. Therefore, next, we wanted to evaluate if there is any lysosomal impairment in these patients. Lysosome is the home of cathepsin, a class of acid proteases that plays an essential role in the degradation of autophagic material and maintaining the cellular homeostasis of metabolism [44]. Therefore, lysosomal impairment can be evaluated by measuring the level of different cathepsins and other proteases. We adopted a human protease proteome pro ler array, which can quantitatively detect 40 different blood-borne proteases (layout was shown in supplementary Fig. 4) including Cathepsins, Kallikreins, and matrix metalloproteinases. Interestingly, we did not observe any difference in the expression of lysosomal cathepsins between healthy control (Fig. 4A) and CFS male patient (Fig. 4B) as well as between female control (Fig. 4C) and female ME/CFS patient (Fig. 4D). Although, strong signals were observed for lysosomal proteases such as Cathepsin A (Fig. 4E), D (Fig. 4F), and X ( Fig. 4G) in both case-control group, no signi cant difference was observed. We observed upregulation of cathepsins in male patient, however that difference did not pass t-test for the statistical signi cance. In contrast, we observed signi cant upregulation of extra-lysosomal proteases such as MMP-9 ( Fig. 4H) and proteinase-3 ( Fig. 4K) in male ME/CFS, but not female ME/CFS patient. Collectively, these results suggest that although there was signi cant impairment in the formation of autophagosomes, there was no abnormality in the expression of lysosomal proteases and a majority of other extra-lysosomal proteases.
Exploring the effect of autophagy marker ATG13 in the microglial stress response. Since we observed strong upregulations of ATG13 in serum samples of both patients, next we wanted to see if ATG13 plays any role in the pathogenesis of ME/CFS. Increase ROS production has been implicated in the pathogenesis of ME/CFS [45] and several other neuroin ammatory diseases such as MS[46], PD [47], and AD [48]. In order to explore the effect of ATG13 on microglial ROS production, we created an in vitro cell culture model (Fig. 5A), in which HMC3 human microglial cells were treated with serum-constituted DMEM media (1:1 v/v) followed by measuring ROS load in different time periods starting from 30 mins to 2 hrs with DCFDA staining procedure [46]. As a control, healthy serum-constituted media was added on HMC3 cells and evaluated for ROS production. Interestingly, our immuno uorescence analyses (Fig. 5B) revealed that patient, but not healthy serum, time-dependently increased the ROS production. The DCFDA-labelled ROS signal appeared as early as 30 minutes and reached maximum at 120 min. This semiquantitative analyses was further con rmed with a real-time kinetic study (Fig. 5C), in which ROS production was measured at a ratio of 485/535 nm as a function of increasing time. Interestingly, we observed that serum sample of male patient increased microglial ROS production with increasing time, whereas healthy serum did not evoke any ROS production. Next, we performed similar uorescence imaging (Fig. 5D) and kinetic (Fig. 5E) analyses with serum samples of female ME/CFS patient and agematched healthy control. Accordingly, we observed that patient serum elevated ROS production with increasing time. Taken together, these results established a new cell culture model to evaluate the ROSinducing capacity in serum samples of ME/CFS patients. With the aid of this model, we were able to quantify ROS-production in HMC3 human microglial cells upon stimulation with ME/CFS serum. Next, we wanted to study if ATG13 in ME/CFS serum enhances ROS production in HMC3 microglial cells. To study that, we adopted a neutralization technique ( Fig. 5A; lower panel), in which patient sera were added on a 96 well plate, which is coated with ATG13 neutralization antibody. After 30 minutes, ATG13-depleted sera were harvested from the top and then proceeded with a similar sets of experiments to evaluate microglial ROS production. Interestingly, ATG13 depletion from the sera of both male (Fig. 5F) and female patients (Fig. 5G), signi cantly attenuated the production of ROS in HMC3 cells suggesting a direct role of serumderived ATG13 in stimulating oxidative stress in microglial cells. Microglial production of nitric oxide (NO) also plays an important role in the pathogenesis of neuroin ammatory diseases [46,49]. Next, we wanted to study if ATG13 could directly stimulate nitrite production in microglia. Surprisingly, Stimulation of HMC3 microglial cells with increasing doses (1, 2, 5, and 10 ng/mL) of ATG13 protein alone induced nitrite production (Fig. 6A). The nitrite production was compared with positive control LPS. Although, ATG13-stimulated nitrite production was almost 2-times lower than that of LPS stimulation, combination of 1 ng/mL ATG13 with 0.5 µg/mL LPS stimulated the production of nitrite on a par with 1 µg/mL LPS. Accordingly, 1:4 (25 µL serum + 75 µL serum-free DMEM/F12 media) and 1:2 (50 µL serum + 50 µL serum-free DMEM/F12 media) dilution (v/v) of serum from male ME/VFS patient dose-dependently stimulated nitrite production in HMC3 microglial cells. On the other hand, we observed that depletion of ATG13 signi cantly ablated the ability of the patient's serum to stimulate nitrite production in HMC3 cells. Taken together, these results suggest that elevated ATG13 in the serum of ME/CFS patients directly stimulates oxidative stress and nitrite production in microglial cells.
Regulation of ATG13 in the serum samples of ME/CFS patients. Next, we wanted to measure ATG13 level and levels of other autophagy-related proteins such as ATG5 and p62 in the sera of 12 ME/CFS patients and 12 healthy controls. The cohort of 24 samples from Dr. Peterson's clinic was randomly selected and displayed with their age, ethnicity, and gender in a table (Table 1). Brie y, serum samples were diluted (1:4 dilution = 25 µL serum + 75 µL sample diluent) and then assayed with competitive ELISA technique for ATG13 (Fig. 7A), ATG5 (Fig. 7B), LC3a (Fig. 7C), and p62 (Fig. 7D). Quanti cation was performed based on the standard curve equation as shown in supplementary Fig. 5. Interestingly, we observed that ATG13 levels are higher in 10 patients (out of 12) compared to 10 healthy controls (total 12). No difference was observed while comparing the levels of ATG5, LC3a, and p62. Taken together, our results suggest that Table 1 Gender, ethnicity, age, disease status, and serum ATG13 (± SD) levels in a cohort of 24 subjects (12 healthy and 12 ME/CFS). High or low was determined with reference to mean (= 4.8030). Expression above mean is considered to be high and below mean is considered to be low. Red

Discussion
Several lines in our current report highlighted that there could be a strong impairment of autophagy function in ME/CFS patients. First, a ThT-based protein aggregation study revealed that serum samples of ME/CFS patients had signi cantly higher propensity to form aberrant protein aggregation indicative of potential autophagy impairment, which is an essential pathway for these protein to be targeted and degraded in lysosomes. Second, our dot blot antibody array analyses indicated that serum levels of, ATG5, ATG7, ATG13, p62, and α-syn were strongly elevated in 71 years old male ME/CFS patient. Third, similar antibody array analysis in a 67 years old female patient detected signi cantly high levels of ATG13, LC3a, LC3b, p62, and αsyn. Fourth, our immunoblot analyses further con rmed that there were strong upregulations of p62, α-syn, and LC3 levels in serum of these patients. The result was further corroborated with densitometric analyses. All these results combinedly suggest that there is a substantial impairment of autophagy in these two cases.
We observed that levels of ATG5, ATG7, and ATG13 were strongly elevated in a 71-years old male ME/CFS patient, whereas only ATG13 was signi cantly upregulated in 67 years old female ME/CFS patient A recent report suggests that upscaled detection of ATG proteins, speci cally ATG5 and ATG12, in serum potentially indicates progressive neuronal damage in brain of AD patients. Aberrant aggregation of Amyloid-β in AD brain stimulates the impairment of autophagy resulting the release of ATG5 and ATG12 in CSF and blood serum. Therefore, a strong impairment of autophagy and subsequent release of autophagy marker proteins such as ATG5 in blood could be a reliable indicator of cognitive impairment in ME/CFS patient. In fact, 71 years old ME/CFS patient has signi cant cognitive impairment and memory loss. Additionally, we observed a strong upregulation of α-syn in serum of that male patient indicating a potential relationship of alpha-synucleinopathy [50][51][52] and Lewybody dementia [52] with ME/CFS.
Although the detection of ATG13 in blood serum was unexpected, we observed a strong upregulation of ATG13 in 67 years old female ME/CFS patient. ATG13 is an important factor for cardiac development as atg13 de cient mouse displays growth retardation and severe de cit in cardiac development [53]. A recent study [54] indicates that during myocardial infarction plasma level of atg13 rises and selective inhibition of circulating ATG13 bene ts the protection against myocardial infarction and other cardiovascular impairments. Interestingly, that female patient in our present study suffers from cardiovascular abnormalities with exceptionally high blood pressure. Therefore, high level of circulating ATG3 might account for the high blood pressure in that patient. Interestingly, ATG13 also regulates the expression of antiviral interferon β to restrict viral infection and plays an essential role in innate immune response to provide antiviral immunity [55]. Therefore, functional de ciency of atg13 gene increases the susceptibility to viral infection and viral reactivation. Both are very common in ME/CFS patients. Moreover, our results further demonstrated that ATG13 alone stimulated the oxidative and nitosylative stress response in microglial cells, whereas selective ablation of ATG13 in serum of ME/CFS patient ameliorated oxidative stress response in microglial cells.
Further characterization of serum ATG13 is mandated. ATG13 participates in the early stage of autophagy vesicle formation process via forming a complex of ULK1 (unc-51 like autophagy activating kinase 1), FIP200 [56], and ATG101 [57]. Activation of mammalian target of rapamycin (mTOR) directly causes hyperphosphorylation of ATG13 in multiple serine residues [58], inhibits the formation of ULK1 complex, and subsequently aborts the autophagy process resulting the accumulation of aberrant proteins and depolarized mitochondria in the biologically active cell. Therefore, characterization of hyperphosphorylated (or nitrosylated?) form of ATG13 is mandated in the serum of ME/CFS patients. ME/CFS patients are highly heterogeneous in terms of pathophysiological outcomes and serum marker proteins. Although, serum levels of p62 and ATG5 were found to be elevated in both of our case studies, we did not see any detectable amount of ATG5 and LC3 in a cohort of serum samples of 10 ME/CFS and 10 healthy subjects. Taken together, our current case study provides a molecular evidence of autophagy impairment in ME/CFS.

Materials And Methods
Reagents Antibody array: RayBio ® C-series Human Autophagy Array was performed as described in manufacturer's protocol. Array membrane was engraved with 20 autophagy related proteins in duplicate in 8X8 format. Brie y, each membrane was soaked with blocking buffer for 1 hr, incubated with 1:5 diluted serum samples (1.4 mg protein as measured with Bradford protein detection method) for 5 hrs at 37ºC, washed 5 times with 1X wash buffer, incubated with biotin-conjugated secondary antibody for 2 hrs at 37ºC followed by wash (5X) and detection with IRDye800-conjugated streptavidin at Licor Odyssey Sa infrared scanner. All necessary reagents and antibodies were supplied with the kit except IRDye800conjugated streptavidin as the application of streptavidin conjugate was devised for the detection of membrane in Sa scanner. Scanning process was done at resolution of 200 µm with focus offset at 3.0 mm. Density of individual protein was measured in ImageJ software, subtracted from blank, normalized with positive control, and then plotted at GraphPad Prism 8 software.  Figure 1 Thio avin T assay to measure the protein aggregation tendency of serum proteins in ME/CFS patients.
Two-hundred μg protein containing serum samples of (A) 74 years old healthy male, (B) 71 years old male ME/CFS patient, (C) 68 years old healthy female, and (D) 67 years old ME/CFS female were assayed for a Thio avin T(ThT)-based protein aggregation assay. ThT Fluorescence absorbance [486/450 nm] was measured from 0 to 90 mins at 3 minutes of interval after ThT addition through autoinjector. Results are mean ± SD of three independent experiments. (E &F) Comparison of slope values between healthy controls (green) and patients (red). (E) represents liner slope value difference between 74 years healthy (green) and 71 years ME/CFS (red) males, whereas (F) represents the slope difference between 68 years healthy (green) and 67 years old ME/CFS females suggesting that ME/CFS patients might have potential abnormality in clearance of proteinopathic aggregates. Rheb[***p<0.005(=0.00334)], and (H) α-syn [***p<0.005(=0.00172)] were measured in ImageJ software (NIH, USA) after subtraction with blank followed by normalization with positive control. Results were mean ± SD and con rmed after three independent experiments. ELISA analyses of (I) ATG5, (J) ATG13, (K) p62, and (L) α-syn were performed in healthy (green) and patient (red). Results were con rmed and considered to be signi cant after 5 independent experiments and indicated with 5 dots in a dotted histogram. Near -infrared ELISA assay for (M) ATG5, (N) ATG13, (O) p62, (P) α-syn. A paired t-test was done to assess the signi cance of mean. Antibody control was assessed without sample addition and then probing each well with only 2 º antibody. Albumin was run for both samples as a loading control. ROS production was assayed with DCFDA staining procedure. (C) Realtime ROS production was assayed in HMC3 microglia after stimulation with male healthy (green) and ME/CFS (red) subjects. A timesensitive recording was performed at VictorX3 multichannel plate reader using 485 nm/535 nm wavelength (Ex/Em) protocol. A time-dependent ROS production assay was recorded at 485 nm/535 nm (Ex/Em) wavelength (D) HMC3 microglial cells were stimulated with sera (1:2 diln) of female healthy and ME/CFS patient for 60 and 120 mins followed by probing with DCFDA. Images were captured in FITClter of Hund uorescence microscope. (E) Realtime ROS production was assayed at different time interval in HMC3 microglial cells once stimulated with sera of female healthy (green) and ME/CFS (red) subjects. (F) Similar ROS assay was performed in HMC3 cells after stimulation with the serum of male ME/CFS patient (red) and ATG13-depleted serum (green). (G) Similar ROS assay was performed in presence of female ME/CFS serum and ATG13-depleted serum. Result was mean ± SD of three independent experiments.

Figure 6
Effect of serum-derived ATG13 on the production of nitric oxide (nitrite) in HMC3 microglial cells. (A) HMC3 microglial cells were stimulated with recombinant ATG13 proteins (1-10 ng/mL), LPS (0.5-1 μg/mL), and ATG13 (1-5 ng/mL) + LPS (0.5 μg/mL) under serum free condition for 24 hrs. After that, colorimetric Griess reaction was performed in supernatant and OD was measured at 575 nm. (B) Serum sample of 71-year old male ME/CFS patient was neutralized with ATG13 and applied on HMC3 microglia at 1:2 and 1:4 dilution (diluted with serum-free DMEM media) for 24 hrs. After 24 hrs nitric oxide production was measured by Griess method. The effect was compared with equivalent dilutions of serum without ATG13 neutralization. Results are mean ± SD of three independent experiments. Signi cance of mean was tested with unpaired t-test and resultant p values were presented underneath. Analyzing expressions of ATG13, ATG5, LC3a, and p62 in a cohort of 10 ME/CFS and 10 healthy serum samples. A cohort of 10 healthy controls and 10 ME/CFS patients were analyzed for the expressions of (A) ATG13, (B) ATG5, (C) LC3a, and (D) p62 in serum samples. Results are mean ± SD of three independent experiments. Signi cance of mean was tested with unpaired t-test and resultant p values were presented underneath.