Downregulation of LKB1/AMPK Signaling in Blood Mononuclear Cells Is Associated with the Severity of Guillain–Barre Syndrome

AMP-activated protein kinase (AMPK) is an intracellular energy sensor that regulates metabolic and immune functions mainly through the inhibition of the mechanistic target of rapamycin (mTOR)-dependent anabolic pathways and the activation of catabolic processes such as autophagy. The AMPK/mTOR signaling pathway and autophagy markers were analyzed by immunoblotting in blood mononuclear cells of 20 healthy control subjects and 23 patients with an acute demyelinating form of Guillain–Barré syndrome (GBS). The activation of the liver kinase B1 (LKB1)/AMPK/Raptor signaling axis was significantly reduced in GBS compared to control subjects. In contrast, the phosphorylated forms of mTOR activator AKT and mTOR substrate 4EBP1, as well as the levels of autophagy markers LC3-II, beclin-1, ATG5, p62/sequestosome 1, and NBR1 were similar between the two groups. The downregulation of LKB1/AMPK signaling, but not the activation status of the AKT/mTOR/4EBP1 pathway or the levels of autophagy markers, correlated with higher clinical activity and worse outcomes of GBS. A retrospective study in a diabetic cohort of GBS patients demonstrated that treatment with AMPK activator metformin was associated with milder GBS compared to insulin/sulphonylurea therapy. In conclusion, the impairment of the LKB1/AMPK pathway might contribute to the development/progression of GBS, thus representing a potential therapeutic target in this immune-mediated peripheral polyneuropathy.


Introduction
Guillain-Barré syndrome (GBS) is a heterogeneous acute autoimmune polyradiculoneuropathy with still ill-defined etiology. Although fairly rare (1-2 cases per 100,000 individuals), GBS is a considerable socioeconomic burden, since it may cause long-term disability and high mortality despite hospitalization and expensive therapy with intravenous immunoglobulins (IVIG) and/or plasma exchange [1][2][3]. The most common forms of GBS with distinct clinical presentations are acute inflammatory demyelinating polyneuropathy (AIDP), acute motor axonal neuropathy (AMAN), and acute motor-sensory axonal neuropathy (AMSAN) [4]. The pathogenesis of GBS includes the loss of immunological tolerance to myelin (AIDP) or axonal (AMAN/AMSAN) components in both T and B cell compartments, usually triggered by a gastrointestinal tract or respiratory infection and leading to inflammatory damage of peripheral nerves [5,6]. However, the molecular mechanisms underlying the initiation and progression of this pathogenic immune response are still poorly understood. healthy employees of the Clinic and their relatives. All patients included in the study fulfilled the Brighton criteria for GBS [20] and were drug-naïve at the time of sampling. The patients with axonal subtypes of GBS (AMAN/AMSAN), which are less frequent in Europe and North America [4], as well as the patients with rare GBS variants such as Miller Fisher syndrome and pharyngocervical brachial variants, were excluded from the analysis to obtain a more homogenous study group. The patients and control subjects with other systemic acute/chronic diseases or those receiving any therapy that may affect their sensory/motor function or autophagy [21] were also excluded from the study. At the time of sampling, the GBS patients' disability was evaluated using the GBS disability scale (GDS; 0-6 ranging from no disability to death) [22], wherein a score >1 indicated poor recovery on follow-up after 6 months (two patients refused control examination). Motor impairment at diagnosis was scored using the Medical Research Council sum score (MRCSS; range 0-60, with 0 denoting complete paralysis of all 12 analyzed muscles and 60 points denoting preserved muscle strength in all muscles) [23]. Blood samples (40 mL) were collected by venipuncture into anticoagulant (ethylenediaminetetraacetic acid)-containing tubes (BD, Franklin Lakes, NJ, USA). PBMCs were isolated using Lymphoprep (Axis-Shield, Oslo, Norway) gradient according to the manufacturer's instructions. The study was approved by the Ethics Committee of the Faculty of Medicine, University of Belgrade, and all subjects gave written consent to participate.

Retrospective Study of Metformin Influence on GBS
To analyze the effect of metformin therapy on GBS, we used the national Registry of Rare Neuromuscular Diseases to retrieve the clinical data of 49 GBS patients with preexisting type 2 diabetes mellitus and no other acute/chronic illnesses, treated for GBS at the Neurology Clinic, University Clinical Center of Serbia (Belgrade, Serbia) in ten years from 1 January 2009 to 1 January 2019 The patients were treated for diabetes with metformin, insulin, or sulphonylurea derivatives (gliclazide, glimepiride, or glibenclamide). The severity of GBS at nadir and hospital discharge was defined as mild (GDS ≤ 3) or high (GDS ≥ 4). The Registry of Rare Neuromuscular Diseases was approved by the Ethics Committee of the Faculty of Medicine, University of Belgrade (Belgrade, Serbia), while the Institutional Review Board approved the protocol of the study.

Statistical Analysis
Because most of the patients' and control subjects' data failed to meet normality, homogeneity of variance, and/or "no outliers" assumptions, statistical analysis was performed using nonparametric tests. The two-tailed Mann-Whitney U test was used for inter-group comparisons, while Spearman's rank-order test was used to assess correlations (the linear trendlines in correlation graphs are only for visualization of correlation trends, and do not reflect the exact relationship between variables). The frequencies were compared with Fisher's exact test. The binary logistic regression (enter method) was employed for multivariate analysis after excluding multicollinearity (variance inflation factor <3), while the Box-Tidwell test was used to confirm the linearity between the independent predictors and the logit transformation of the dependent variable. The goodness of fit was assessed by the Hosmer-Lemeshow test, p > 0.05 indicating a good fit. The statistical analysis was performed using IBM SPSS software (v22), and the significance level was set to p ≤ 0.05. To reduce the risk of missing the true effects, no corrections for multiple comparisons were made [26].

Demographic and Clinical Characteristics of Study Subjects
The GBS and control groups did not significantly differ in sex distribution (male/female 14/9 vs. 11/9, p = 0.763, Fisher's exact test) or age (median/interquartile range 59/46-69 vs. 59/48-69 years, p = 0.817, Mann-Whitney U test). The sex distribution in the GBS group was consistent with the fact that men are more frequently affected than women (ratio 3:2) [1]. The clinical features of GBS patients, including GDS as a measure of overall disability, MRCSS as a measure of muscle strength, the presence of gastrointestinal/respiratory infection as a precipitating factor, mode of therapy, and death rate are presented in Table 1. All patients were treated with IVIG, and the death rate of 13% was consistent with the reported GBS mortality rates in Serbia [3]. Neither of the two disease severity measures, GDS (at admission and 6 months) or MRCSS at admission, were significantly associated with the presence of gastrointestinal/respiratory infection before disease onset (Supplementary Figure S7). While GDS values did not significantly differ between male and female patients, the MRCSS values were slightly but significantly higher in males (Supplementary Figure S7). Moreover, there was a significant positive correlation between patients' age and GDS at admission, indicating an increase in disease severity with age (Supplementary Figure S7). No correlation was observed between patients' age and MRCSS values (Supplementary Figure S7).

AMPK Signaling Pathway Is Downregulated in PBMCs of GBS Patients
To analyze the status of the AMPK signaling axis in the PBMCs of control and GBS subjects, we measured the expression and phosphorylation of AMPK, its upstream activator LKB1, and AMPK substrate Raptor. The results of immunoblot analysis revealed that the levels of phosphorylated AMPK (Figure 1a), phosphorylated and total LKB1 (Figure 1b), as well as phosphorylated and total Raptor (Figure 1c), were significantly lower in the PBMCs of GBS patients compared to control subjects. The ratio of phosphorylated and total LKB1 levels was not significantly altered in GBS PBMCs (Figure 1b), indicating that the decrease in phospho-LKB1 was mainly a consequence of reduced LKB1 expression. On the other hand, the phospho/total ratio of both AMPK ( Figure 1a) and Raptor (Figure 1c) was significantly reduced in GBS patients, thus confirming that the phosphorylation of AMPK and Raptor was impaired in GBS leukocytes. This was expected having in mind that in the AMPK signaling cascade, LKB1 phosphorylates AMPK, which then phosphorylates Raptor. These data demonstrate that the activity of the LKB1/AMPK/Raptor signaling axis in the PBMCs of GBS patients is downregulated at several levels.
phosphorylation of AMPK and Raptor was impaired in GBS leukocytes. This was expected having in mind that in the AMPK signaling cascade, LKB1 phosphorylates AMPK, which then phosphorylates Raptor. These data demonstrate that the activity of the LKB1/AMPK/Raptor signaling axis in the PBMCs of GBS patients is downregulated at several levels. The representative blots are shown, while densitometry results are presented relative to actin or as phospho/total signal ratio (horizontal lines represent median and interquartile range; * p < 0.05, twotailed Mann-Whitney U test).

Autophagy Markers and AKT/mTORC1 Pathway Are Not Affected in PBMCs of GBS Patients
We next assessed the status of autophagy in the PBMCs of GBS patients by examining the expression of autophagy genes, the protein levels of pro-autophagic regulators beclin-1 and ATG5, the conversion of LC3-I (a mammalian homolog of ATG8) to autophagosome associated LC3-II, and autophagy-selective degradation of cargo receptors SQSTM1 and NBR1. The RT-qPCR analysis revealed that out of seventeen autophagy regulators tested, the levels of mRNA encoding autophagy transcription factor FOXO1 and four ATG molecules (ATG4B, ATG13, ATG14, and ULK1, a mammalian homolog of ATG1), were significantly reduced in the PBMCs of GBS patients compared to control subjects ( Figure 2a and Supplementary Figure S8). However, the immunoblot analysis demonstrated that the protein levels of autophagy markers LC3-II, ATG5, beclin-1, SQSTM1, and NBR1 did not significantly differ between the PBMCs of GBS patients and healthy controls (Figure 2b). Similarly, the activation of the autophagy-regulating AKT/mTORC1 signaling axis, The representative blots are shown, while densitometry results are presented relative to actin or as phospho/total signal ratio (horizontal lines represent median and interquartile range; * p < 0.05, two-tailed Mann-Whitney U test).

Autophagy Markers and AKT/mTORC1 Pathway Are Not Affected in PBMCs of GBS Patients
We next assessed the status of autophagy in the PBMCs of GBS patients by examining the expression of autophagy genes, the protein levels of pro-autophagic regulators beclin-1 and ATG5, the conversion of LC3-I (a mammalian homolog of ATG8) to autophagosome associated LC3-II, and autophagy-selective degradation of cargo receptors SQSTM1 and NBR1. The RT-qPCR analysis revealed that out of seventeen autophagy regulators tested, the levels of mRNA encoding autophagy transcription factor FOXO1 and four ATG molecules (ATG4B, ATG13, ATG14, and ULK1, a mammalian homolog of ATG1), were significantly reduced in the PBMCs of GBS patients compared to control subjects (Figure 2a and Supplementary Figure S8). However, the immunoblot analysis demonstrated that the protein levels of autophagy markers LC3-II, ATG5, beclin-1, SQSTM1, and NBR1 did not significantly differ between the PBMCs of GBS patients and healthy controls (Figure 2b). Similarly, the activation of the autophagy-regulating AKT/mTORC1 signaling axis, assessed by measuring the phosphorylation of AKT and mTORC1 substrate 4EBP1, was not altered in GBS leukocytes (Figure 2b). Therefore, despite the partial suppression of ATG transcription, the expression of autophagy markers and the activity of the AKT/mTORC1 pathway were mainly unaffected in GBS leukocytes. assessed by measuring the phosphorylation of AKT and mTORC1 substrate 4EBP1, was not altered in GBS leukocytes (Figure 2b). Therefore, despite the partial suppression of ATG transcription, the expression of autophagy markers and the activity of the AKT/mTORC1 pathway were mainly unaffected in GBS leukocytes.

Interplay between AMPK, ERK, and PKC Signaling Pathways in PBMCs of GBS Patients
To account for the apparent absence of autophagy suppression in the face of AMPK downregulation in GBS PBMCs, we assessed the activation of other autophagy-activating signaling pathways, namely ERK, p38 MAPK, and PKC. The phosphorylation of ERK and PKC substrate GSK3β, but not p38 MAPK, was significantly increased in the PBMCs of GBS patients compared to control subjects (Figure 3a). We next assessed if the observed increase in PKC and/or ERK activation might be connected to AMPK downregulation in the PBMCs of GBS patients. The phosphorylation of AMPK and LKB1, as well as the total levels of LKB1, were inversely correlated with those of phosphorylated GSK3β, while a positive correlation was observed between the phosphorylated forms of ERK and GSK3β (Figure 3b). No significant correlation was found between the phosphorylation of p38 MAPK and GSK3β, nor between the levels of phospho-ERK or phospho-p38 and phospho-AMPK, phospho-LKB1, or total LKB1 levels (Supplementary Figure S9). Collectively, these data indicate that the activation of ERK and PKC might counterbalance AMPK inhibitionmediated autophagy suppression in GBS leukocytes, with PKC possibly being connected with both AMPK downregulation and ERK activation.
To account for the apparent absence of autophagy suppression in the face of AMPK downregulation in GBS PBMCs, we assessed the activation of other autophagy-activating signaling pathways, namely ERK, p38 MAPK, and PKC. The phosphorylation of ERK and PKC substrate GSK3β, but not p38 MAPK, was significantly increased in the PBMCs of GBS patients compared to control subjects (Figure 3a). We next assessed if the observed increase in PKC and/or ERK activation might be connected to AMPK downregulation in the PBMCs of GBS patients. The phosphorylation of AMPK and LKB1, as well as the total levels of LKB1, were inversely correlated with those of phosphorylated GSK3β, while a positive correlation was observed between the phosphorylated forms of ERK and GSK3β (Figure 3b). No significant correlation was found between the phosphorylation of p38 MAPK and GSK3β, nor between the levels of phospho-ERK or phospho-p38 and phospho-AMPK, phospho-LKB1, or total LKB1 levels (Supplementary Figure S9). Collectively, these data indicate that the activation of ERK and PKC might counterbalance AMPK inhibition-mediated autophagy suppression in GBS leukocytes, with PKC possibly being connected with both AMPK downregulation and ERK activation.

Downregulation of LKB1/AMPK Pathway Correlates with GBS Severity and Poor Outcome
To explore the possible role of reduced leukocyte AMPK signaling in GBS, we correlated the activity of the AMPK signaling axis with GDS and MRCSS as the measures of GBS patients' disability and muscle strength. The total levels of LKB1, phospho/total AMPK ratio, and total levels of Raptor displayed a significant inverse correlation with GDS at hospital admission (Figure 4a). The total levels of LKB1 were also positively correlated with MRCSS values at hospital admission, while a trend toward a positive association of the phospho/total AMPK ratio and total levels of Raptor with MRCSS failed to reach statistical significance (Figure 4b). The observed correlations were not due to the confounding effect of age or sex, as they did not influence the expression/activation of AMPK signaling molecules associated with disease severity (total LKB1, phospho/total , and the levels of phosphorylated ERK, p38 MAPK, and GSK3β were assessed by immunoblotting. The representative blots and densitometry data (relative to actin) are shown in (a), with horizontal lines representing median values and interquartile range (* p < 0.05, two-tailed Mann-Whitney U test). The correlations between GSK3β and phospho-AMPK, phospho-LKB1, or phospho-ERK are presented in (b) (r s -Spearman's correlation coefficient; * p < 0.05, Spearman's rank-order test).

Downregulation of LKB1/AMPK Pathway Correlates with GBS Severity and Poor Outcome
To explore the possible role of reduced leukocyte AMPK signaling in GBS, we correlated the activity of the AMPK signaling axis with GDS and MRCSS as the measures of GBS patients' disability and muscle strength. The total levels of LKB1, phospho/total AMPK ratio, and total levels of Raptor displayed a significant inverse correlation with GDS at hospital admission (Figure 4a). The total levels of LKB1 were also positively correlated with MRCSS values at hospital admission, while a trend toward a positive association of the phospho/total AMPK ratio and total levels of Raptor with MRCSS failed to reach statistical significance (Figure 4b). The observed correlations were not due to the confounding effect of age or sex, as they did not influence the expression/activation of AMPK signaling molecules associated with disease severity (total LKB1, phospho/total AMPK, total Raptor) (Supplementary Figure S10). The levels of phospho-LKB1, phospho-AMPK, total AMPK, and phospho-Raptor, as well as phospho/total LKB1 and phospho/total Raptor ratio, were not significantly correlated with either GDS or MRCSS, although a tendency toward negative correlation with GDS was evident in the case of phospho-Raptor and phospho/total Raptor ratio (Supplementary Figure S11). The lower levels of phospho-LKB1 ( Figure 5a) and phospho-Raptor (Figure 5c) at admission were significantly associated with poor recovery (GDS > 1) at 6 months ( Figure 5). A similar trend was observed with the phospho/total LKB1 ratio (Figure 5a), phospho-AMPK, and total AMPK levels (Figure 5b), and phospho/total Raptor ratio (Figure 5c), but the associations did not reach statistical significance. On the other hand, no significant associations were observed between GDS (at admission or after six months) and autophagy markers (LC3-II, ATG5, beclin-1, SQSTM1, and NBR1) or phosphorylated forms of 4EBP1, AKT, ERK, p38 MAPK, and GSK3β (Supplementary Figures S12 and S13). Therefore, the downregulation of the leukocyte AMPK signaling axis correlated with increased severity and poor outcome of GBS independently of age, sex, and the expression of autophagy markers or the activation of autophagy-related signaling pathways.  Figure S10). The levels of phospho-LKB1, phospho-AMPK, total AMPK, and phospho-Raptor, as well as phospho/total LKB1 and phospho/total Raptor ratio, were not significantly correlated with either GDS or MRCSS, although a tendency toward negative correlation with GDS was evident in the case of phospho-Raptor and phospho/total Raptor ratio (Supplementary Figure S11). The lower levels of phospho-LKB1 ( Figure 5a) and phospho-Raptor (Figure 5c) at admission were significantly associated with poor recovery (GDS > 1) at 6 months ( Figure 5). A similar trend was observed with the phospho/total LKB1 ratio (Figure 5a), phospho-AMPK, and total AMPK levels (Figure 5b), and phospho/total Raptor ratio (Figure 5c), but the associations did not reach statistical significance. On the other hand, no significant associations were observed between GDS (at admission or after six months) and autophagy markers (LC3-II, ATG5, beclin-1, SQSTM1, and NBR1) or phosphorylated forms of 4EBP1, AKT, ERK, p38 MAPK, and GSK3β ( Supplementary Figures S12 and S13). Therefore, the downregulation of the leukocyte AMPK signaling axis correlated with increased severity and poor outcome of GBS independently of age, sex, and the expression of autophagy markers or the activation of autophagy-related signaling pathways.

Treatment with Metformin Is Associated with Better GBS Outcome in Diabetic Patients
Finally, to assess the potential therapeutic benefit of counteracting AMPK signaling deficit in GBS, we compared the disease severity in GBS patients with type 2 diabetes treated with AMPK activator metformin and those receiving insulin and/or sulphonylurea derivatives without metformin. The two groups did not significantly differ in sex distribution, age, fasting blood glucose level at the time of admission, the presence of respiratory or gastrointestinal infection before GBS onset, or the proportion of patients who received IVIG/plasma exchange therapy (Table 2). On the other hand, the proportion of patients with mild disability (GDS ≤ 3) at disease nadir and hospital discharge was significantly higher in the metformin group (Table 2). To examine the prognostic value of metformin therapy in diabetic GBS patients, a simple binary logistic regression model for predicting GBS outcome at hospital discharge was created, incorporating patients' age, Figure 5. Association of leukocyte AMPK signaling with GBS outcome. The levels of phosphorylated, total, and phospho/total ratio of LKB1 (a), AMPK (b), and Raptor (c) in PBMCs of GBS patients (n = 21) were determined by immunoblotting and compared between patients with good (GDS ≤ 1, n = 12) and poor (GDS > 1, n = 9) disease outcome at 6 months (horizontal lines represent median and interquartile range; * p < 0.05, two-tailed Mann-Whitney U test).

Treatment with Metformin Is Associated with Better GBS Outcome in Diabetic Patients
Finally, to assess the potential therapeutic benefit of counteracting AMPK signaling deficit in GBS, we compared the disease severity in GBS patients with type 2 diabetes treated with AMPK activator metformin and those receiving insulin and/or sulphonylurea derivatives without metformin. The two groups did not significantly differ in sex distribution, age, fasting blood glucose level at the time of admission, the presence of respiratory or gastrointestinal infection before GBS onset, or the proportion of patients who received IVIG/plasma exchange therapy (Table 2). On the other hand, the proportion of patients with mild disability (GDS ≤ 3) at disease nadir and hospital discharge was significantly higher in the metformin group (Table 2). To examine the prognostic value of metformin therapy in diabetic GBS patients, a simple binary logistic regression model for predicting GBS outcome at hospital discharge was created, incorporating patients' age, preceding gastrointestinal infection, and the mode of diabetes treatment as independent variables. The model was significant (chi-square = 16.203, p = 0.001), explaining 38.9% (Nagelkerke R 2 = 0.389) of the variance in GBS outcome, correctly classifying 75.5% of cases, and presenting a good fit to the data (Hosmer-Lemeshow p = 0.689). In addition to the absence of antecedent gastrointestinal infection, metformin therapy was a significant independent predictor of good recovery, with metformin-treated patients being almost six times more likely to have a favorable outcome ( Table 3). The inclusion of patients' sex and glycemia at admission, as a measure of diabetes control, did not further improve the performance of the model (Nagelkerke R 2 = 0.399, correct prediction 75.5%) (Supplementary Table S1). These data indicate that diabetic patients receiving metformin are expected to experience milder GBS.

Discussion
To the best of our knowledge, the presented results suggest for the first time that the downregulation of leukocyte AMPK signaling might be associated with the severity/outcome of the AIDP form of GBS. Moreover, the possible therapeutic benefit of counteracting AMPK suppression in GBS is indicated by a retrospective analysis showing a milder disease in diabetic GBS patients treated with the AMPK activator metformin.
The main finding of the present study is that the activity of the LKB1/AMPK/Raptor signaling axis was significantly reduced in the PBMCs of AIDP patients compared to age/sex-matched healthy control subjects. The initial defect was presumably the decrease in the expression of LKB1, which subsequently caused the suppression of the downstream signaling molecules AMPK and Raptor. However, it is also possible that AMPK and/or Raptor were downregulated at least in part independently of LKB1, as actually indicated by the decrease in the total levels of Raptor in AIDP leukocytes. Similar to our results, AMPK downregulation in T cells and peripheral blood lymphocytes has previously been reported in a mouse model of multiple sclerosis and human hereditary metabolic/neuroinflammatory disease X-linked adrenoleukodystrophy, respectively [27,28]. These previous studies and our data support the possible involvement of dysregulated leukocyte AMPK signaling in immune-mediated central and peripheral demyelinating disorders. Indeed, AMPK has been increasingly recognized as an important regulator of the metabolism, maintenance, expansion, and function of both T and B cell lineages [29][30][31]. While the exact mechanisms of AMPK involvement in the neuroinflammatory damage in AIDP remain to be investigated, its protective role has been suggested by an association between the downregulation of different AMPK signaling members and increased disability and/or motor impairment in GBS. Moreover, the lower activation of the LKB1/AMPK/Raptor pathway in peripheral blood leukocytes was associated with poor recovery from AIDP, indicating a potential prognostic value of leukocyte AMPK signaling status in this disease. However, this association, although significant, was rather weak, thus remaining to be confirmed in larger studies. Also, while the present study was limited to the analysis of the whole PBMC population, we plan to use cell sorting and immunocytochemistry to further evaluate AMPK signaling in specific PBMC subpopulations.
Having in mind the well-known role of AMPK in relieving autophagy from mTORC1 suppression [32], it is somewhat unexpected that the expression of autophagy markers and mTORC1 activity in the PBMCs of GBS patients remained unchanged. This is even more surprising given that AMPK downregulation in GBS leukocytes was associated with reduced mRNA levels of autophagy transcription factor FOXO1 and four ATG molecules involved in autophagy induction (ULK1 and ATG13), the initiation of autophagosome formation (ATG14), and the cleavage of LC3 to generate LC3-II precursor LC3-I (ATG4B) [10]. While this agrees with the involvement of AMPK in ATG transcription [33], the observed AMPK dissociation from mTORC1 signaling and expression of autophagy markers could stem from the counterbalancing activation of other signaling pathways known to modulate mTORC1/autophagy, such as ERK, PKC, or p38 MAPK [13][14][15][16]. Indeed, while p38 MAPK activation was not altered in GBS leukocytes, the phosphorylation of ERK and PKC substrate GSK3β, reflecting PKC activity, was significantly enhanced. Accordingly, the previous analysis of the transcriptional profile of GBS leukocytes revealed the upregulation of both ERK and PKC pathways [34]. Moreover, our findings that PKC activity was correlated inversely with AMPK and positively with ERK phosphorylation, indicate that PKC could contribute to both AMPK inhibition and ERK activation in GBS leukocytes. This is consistent with the ability of PKC to inhibit AMPK [35][36][37][38] and activate ERK [39][40][41][42] in various experimental settings. Despite the increased activity, neither PKC nor ERK was associated with GBS disability or motor impairment scores, thus further emphasizing the potential value of AMPK signaling status as a marker of GBS severity and outcome. AMPK phosphorylates a variety of metabolic targets [43], which might affect immune cells independently of mTORC1 and/or autophagy. However, it should be noted that we analyzed the steady-state levels of an autophagosome marker LC3-II and autophagy-selective targets SQSTM1 and NBR1 as a proxy for autophagic status, without directly evaluating autophagic turnover (flux). We are currently seeking to confirm the absence of autophagy modulation in GBS by using a recently developed protocol for measuring autophagic flux in whole blood, thus avoiding cell cultivation in nutrient-rich media that affect mTOR signaling and autophagy [44].
To assess the possible usefulness of AMPK as a therapeutic target in GBS, we exploited the fact that the anti-diabetic drug metformin achieves its insulin-sensitizing effect primarily through AMPK activation [45]. The milder GBS symptoms observed in metformin-treated, compared to insulin/sulphonylurea-treated GBS patients with type 2 diabetes, indicates that pharmacological reversal of AMPK signaling block might be associated with reduced GBS severity. Moreover, while extending to type 2 diabetes the previous findings on preceding diarrhea as a predictive factor of poor GBS outcome [46], we demonstrate that metformin treatment was a significant independent predictor of favorable GBS outcome at hospital discharge. While the effects of metformin in immune-mediated damage of peripheral nerves are yet to be explored, our data are consistent with its ability to reduce inflammatory CNS demyelination in animal models of multiple sclerosis via an AMPK-dependent mechanism [47,48]. However, we and others have previously shown that diabetes worsens the prognosis of GBS [49][50][51], possibly due to the presence of pre-existing nerve injury, reduced capacity for nerve regeneration, and/or systemic inflammation [52][53][54]. The non-metformin group in our study mainly included patients treated with insulin, a second-line medication for type 2 diabetes patients in which hyperglycemia could not be controlled with metformin [55]. Hence, it is possible that the observed association between metformin treatment and better GBS outcome was actually due to inherently milder diabetes in the metformin group. On the other hand, the absence of association between fasting glycemia at admission and metformin treatment or GBS outcome indicates that the last two variables might be connected independently of diabetes control. Nevertheless, more extensive studies are required to corroborate the proposed beneficial effect of metformin in GBS.

Conclusions
The present report indicates that the impairment of the AMPK pathway in PBMCs might contribute to the development and/or progression of the AIDP form of GBS, thus representing a possible therapeutic target in this autoimmune disorder. Accordingly, restoring the defective leukocyte AMPK signaling with AMPK-activating drugs, such as clinically approved antidiabetic metformin, might be worth considering as a novel add-on therapeutic strategy in GBS. Larger studies are required to confirm these observations and extend them to other forms of GBS and/or specific leukocyte subpopulations, as well as to validate the prognostic/therapeutic potential of assessing/manipulating the leukocyte AMPK signaling axis in this immune-mediated neuropathy.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells11182897/s1, Figure S1: Original immunoblots of phospho/total AMPK and phospho/total LKB1; Figure S2: Original immunoblots of phospho/total Raptor and LC3-I/II; Figure S3: Original immunoblots of ATG5, beclin-1, and SQSTM1; Figure S4: Original immunoblots of phospho-4EBP1, phospho-ERK, and phospho-GSK3β; Figure S5: Original immunoblots of phospho-AKT and phospho-p38 MAPK; Figure S6: Original immunoblots of NBR1; Figure S7: Association between GBS severity and sex/age or prior infection; Figure S8: Expression of autophagy genes in PBMCs of GBS patients and control subjects; Figure S9: Correlation between AMPK, ERK, and PKC activation in PBMCs of GBS patients; Figure S10: Correlation between leukocyte AMPK signaling and GBS patients' age; Figure S11: Correlation between leukocyte AMPK signaling and GBS severity; Figure S12: Correlation between autophagy markers/regulators and GDS at admission; Figure S13: Association between leukocyte autophagy markers/regulators and GBS outcome; Table S1: Addition of sex and fasting glycemia to a predictive model for good GBS outcome in diabetic patients.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available in the article and Supplementary Materials.