*Article* **Quinpirole-Mediated Regulation of Dopamine D2 Receptors Inhibits Glial Cell-Induced Neuroinflammation in Cortex and Striatum after Brain Injury**

**Sayed Ibrar Alam 1,†, Min Gi Jo 1,†, Tae Ju Park 2, Rahat Ullah 1, Sareer Ahmad 1, Shafiq Ur Rehman <sup>1</sup> and Myeong Ok Kim 1,\***


**Abstract:** Brain injury is a significant risk factor for chronic gliosis and neurodegenerative diseases. Currently, no treatment is available for neuroinflammation caused by the action of glial cells following brain injury. In this study, we investigated the quinpirole-mediated activation of dopamine D2 receptors (D2R) in a mouse model of traumatic brain injury (TBI). We also investigated the neuroprotective effects of quinpirole (a D2R agonist) against glial cell-induced neuroinflammation secondary to TBI in adult mice. After the brain injury, we injected quinpirole into the TBI mice at a dose of 1 mg/kg daily intraperitoneally for 7 days. Our results showed suppression of D2R expression and deregulation of downstream signaling molecules in ipsilateral cortex and striatum after TBI on day 7. Quinpirole administration regulated D2R expression and significantly reduced glial cell-induced neuroinflammation via the D2R/Akt/glycogen synthase kinase 3 beta (GSK3-β) signaling pathway after TBI. Quinpirole treatment concomitantly attenuated increase in glial cells, neuronal apoptosis, synaptic dysfunction, and regulated proteins associated with the blood–brain barrier, together with the recovery of lesion volume in the TBI mouse model. Additionally, our in vitro results confirmed that quinpirole reversed the microglial condition media complex-mediated deleterious effects and regulated D2R levels in HT22 cells. This study showed that quinpirole administration after TBI reduced secondary brain injury-induced glial cell activation and neuroinflammation via regulation of the D2R/Akt/GSK3-β signaling pathways. Our study suggests that quinpirole may be a safe therapeutic agent against TBI-induced neurodegeneration.

**Keywords:** brain injury; quinpirole; dopamine D2 receptors; glial cell; neuroinflammation; neurodegeneration

#### **1. Introduction**

Traumatic brain injury (TBI) is a global risk factor and the leading cause of neurological disability. Recent studies have reported that TBI is associated with several neurodegenerative diseases, such as Alzheimer's and Parkinson's disease [1–3]. TBI leads to a primary injury, which is followed by a secondary brain injury. Primary brain injury refers to the direct mechanical force applied at the time of the initial impact on the brain. Secondary brain injury occurs as a consequence of the initial traumatic events. It refers to the involvement of the brain vasculature as well as the blood–brain barrier (BBB) disruption, which results in significant complications in the brain [4]. Neuroinflammation is the principal hallmark of brain injury, followed by astrocyte and microglia activation and release of pro-inflammatory cytokines and chemokines, which impair the endogenous self-repair

**Citation:** Alam, S.I.; Jo, M.G.; Park, T.J.; Ullah, R.; Ahmad, S.; Rehman, S.U.; Kim, M.O. Quinpirole-Mediated Regulation of Dopamine D2 Receptors Inhibits Glial Cell-Induced Neuroinflammation in Cortex and Striatum after Brain Injury. *Biomedicines* **2021**, *9*, 47. https:// doi.org/10.3390/biomedicines9010047

Received: 19 November 2020 Accepted: 4 January 2021 Published: 7 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

ability of the brain and eventually cause neuronal apoptosis and neurodegenerative conditions [5,6]. Several studies have proved that brain injuries precipitate an inflammatory response with activation of neuroinflammatory mediators [7]. Therefore, it is important to develop neuroprotective and neurorestorative agents to treat TBI. Notably, this subject offers much scope for extensive research to treat the TBI-induced neuroinflammatory response and inflammatory cytokine release. Restoration of BBB integrity and treatment of neuroinflammation is the key therapeutic goals in patients with brain injury-induced pathological events.

Dopamine (DA) is a major neurotransmitter that controls abnormal neuronal excitotoxicity and regulates the function of the dopaminergic system in the brain [6,8,9]. Dopamine D2 receptors (D2R) belong to the class of G protein-coupled receptors that are activated by DA and participate in essential functions, including innate immunity and neuroinflammatory responses [10,11]. However, previous studies have reported a significantly increased inflammatory response in D2R-knockout (D2R–/–) mice [12]. D2R is expressed in several regions of the brain, including the cerebral cortex, hippocampus, and striatum [13]. A previous study has shown that DA receptors are expressed on glial and immune cells [14]. Several studies have reported that DA plays a vital role in humans and animals and that cortical dopaminergic dysfunction is associated with attention deficit hyperactivity disorder [15,16]. A recent study has reported that cortical D2R is involved in psychotic and mood disorders and regulates neuronal circuits [17]. Deregulation of the DA system could be a significant contributor to behavioral and cognitive deficits that are observed after TBI. A growing body of evidence suggests that D2R agonists protect against neuroinflammation and immune reactions, perhaps by inhibiting cytokine release [18,19]. An earlier study reported that quinpirole-activated D2R positively affects neuronal activity in the cingulate cortex and striatum [20]. However, limited studies have reported the role of D2R activation in the inhibition of glial cell-induced neuroinflammatory responses following brain injury.

Akt, a serine-threonine kinase, is known to play an essential role in the cell death/survival pathway. Akt phosphorylates and inhibits several substrates, including glycogen synthase kinase 3 beta (GSK3-β) [21]. A previous study investigated the regulation of the Akt pathway by stimulation of DA receptors and reported possible regulation of the Akt/GSK3 β pathway via regulation of D2R [22].

In this study, we investigated the possible regulation of D2R in the cortical region of the brain, which is the primary target of brain injury, and also explored the striatal region in a TBI mouse model. Furthermore, we investigated the therapeutic potential of post-TBI administration of quinpirole hydrochloride [19,23]. We observed that quinpirole administration at a dose of 1 mg/kg could potentially protect against brain injury-induced gliosis, neuroinflammation, neurodegeneration, lesion volume, synaptic dysfunction and and regulated proteins associated with the BBB via stimulation of D2R, particularly in the ipsilateral cortex of TBI mice. We could also confirm microglial involvement in D2R deregulation and that quinpirole at a dose of 20 μM is sufficient to stimulate D2R and regulate Akt levels in neuronal cell lines. This study highlights that quinpirole administration ipsilateral side of TBI mouse brain stimulated D2R and lead to the recovery of brain function via regulation of the Akt/GSK3-β signaling pathway and inhibition of a glial cell-induced neuroinflammatory response.

#### **2. Materials and Methods**

#### *2.1. Animals*

Male wild-type C57BL/6N mice, 7 weeks of age with 25–30 g weight, were obtained from Samtako Bio Korea. The animals were acclimatized in the animal care center at Gyeongsang National University, South Korea. The animal were maintained in the control environment with 12/12 h light/dark cycle at 23 ◦C, and 60 ± 10% humidity with free access to food and water. The mice were randomly divided into following different groups; control, TBI, and TBI + quinpirole after a week of acclimatization. The animals were handled carefully according to the guidelines of the Institutional Animal Care and Use

committee (IACUC) (5 March 2019. Approval ID: 125), Division of Life Science and Applied Life Science, Gyeongsang National University, Republic of South Korea.

#### *2.2. Quinpirole Treatment for Mice*

The treated animals were divided into the following groups:

Saline treated control group, Stab Wound Cortical Injury, Stab Wound Cortical Injury + quinpirole.

Quinpirole was dissolved in distilled water and administered daily intraperitoneally (i.p) at a dose of 1 mg/kg body weight for 7 days. For western blot (*n* = 5) and for confocal experiments (*n* = 6) mice per group were used. The chemical quinpirole was purchased from Tocris-Cookson (Bristol, UK).

#### *2.3. Stab Wound Cortical Injury*

The stab wound cortical brain injury mouse model was established as previously described with modification [24]. Briefly, the mice were anesthetized with Rompun (0.05 mL/100 g body weight) and Zoletil (0.1 mL/100 g body weight). The mice were placed on stereotaxic apparatus and the skull was exposed the by making a mid-longitudinal incision. The dental drill was used to make a circular craniotomy 4 mm in diameter (2 mm lateral to the midline and 1 mm posterior to the bregma) in the skull. For stab wound injury, a sharp edge scalpel blade was inserted (3 mm; right hemisphere) in the mouse brain and kept for 1 min in the brain and then removed slowly. The bone wax was applied to cover the rupture skull followed by stitching with a silk suture to close the wound area. Next, the animals were placed carefully by providing continuous heating with a heating lamp until fully recovered from anesthesia and proceeded for further experiments.

#### *2.4. Protein Extraction*

After the completion of the mice treatment, all the animals were first anesthetized and then sacrificed carefully. After the surgery brain were immediately collected and froze on dry ice. The ipsilateral cortex of TBI brain tissue was homogenized using PRO-PREP protein extract solution (iNtRON Biotechnology, Burlington, NJ, USA). to extract protein from tissues followed by centrifugation and stored at −80 ◦C. The samples were centrifuged at speed of 13,000× *g* rpm at 4 ◦C for 25 min. The supernatants were collected and stored at −80 ◦C for immunoblotting.

#### *2.5. Western Blot Analysis*

The western blot analysis was assessed as previously described with minor changes [25–27]. In brief, an equal volume of 20–30 μg of proteins (extracted from the ipsilateral cortex) was mixed with 2× Sample Buffer (Invitrogen). To separate the proteins, an equal volume of the proteins were run on 10% of SDS polyacrylamide gel electrophoresis and transferred to the PVDF membrane followed by blocking in 5% skim milk. The membranes were slightly washed to clear the skim milk. The primary antibody was incubated overnight at 4 ◦C 1:1000, anti-(D2R), anti-Glycogen synthase kinase 3 (p-GSK3-β) (Ser9), p-Akt (Ser473), anti-Glial fibrillary acidic protein beta (Anti-GFAP), anti-ionized calcium-binding adapter molecule 1 (anti-Iba-1), anti-phospho-c-Jun N-terminal kinase (p-JNK), anti-interleukin-1β (IL-1β), anti-caspase-3, anti-poly (ADP-ribose) polymerase-1 (Anti-PARP-1), anti-Bax, anti-Bcl-2 and anti-β-actin from Santa Cruz Biotechnology. Anti-beta actin was used as a loading control. The next day, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies diluted in 1×TBST for 1–2 h as appropriate; the immunoblots were developed using an ECL chemiluminescence system, according to the manufacturer's instructions (Amersham Pharmacia Biotech, Uppsala, Sweden).

#### *2.6. Brain Tissue Collection and Sample Preparation*

For brain tissue collection, the mice were anesthetized and transcardially perfused with saline followed by (4%) paraformaldehyde and then fixed with (4%) paraformaldehyde for 48 h. Further, the brain tissues were immersed in a 20% sucrose solution for 48 h. Next, the Brain were fixed vertically in the OCT compound medium, Sakura Finetek USA, Inc., Torrance, CA, USA). For the brain cross-section (14 μm in size) using a vibratome (Leica, Nussloch, Germany) and stored at −80 ◦C.

#### *2.7. Immunofluorescence Staining*

The tissue slides were proceeded for immunofluorescence staining as described previously with minor modification [27,28]. Initially, the slides were dried at room temperature and washed twice with PBS 0.01 M solution for 8–10 min. The tissue slides were incubated in proteinase-K (5 min) and then washed twice for 5 min in PBS solution. Next, the protein was blocked for 1 h with 5% normal serum (goat/rabbit) D2R and 0.1% Triton X-100 in 0.01 M PBS solution. The tissue slides were then incubated with primary antibodies (1:100) ratio in 0.01 M PBS solution overnight at 4 ◦C. The following antibodies were used for the immunofluorescence detection; anti-p-GSK3-β (ser9), anti-p-Akta, anti-D2R, anti-IL1-β, anti-Caspase-3, anti-PSD-95, anti-SNAP-23, anti-ZO-1 anti-CD31. The tissue slides were then incubated for 2 h in the secondary antibody (1:100) fluorescein isothiocyanate (FITC), and tetramethylrhodamine isothiocyanates (TRITC) labeled secondary antibodies (anti-goat, anti-rabbit, and anti-mouse) from Santa Cruz Biotechnology. The 4 ,6-diamidino-2-phenylindole (DAPI) was used for nucleus detection (8–10 min). The slides were covered with coverslips using with fluorescent mounting medium. Confocal laser scanning microscopy FluoViewer MPE-1000 (Olympus, Tokyo, Japan) was used to take the images and the maximum fluorescent intensity in the representative field was taken. The images were converted into Tiff format and the fluorescent intensity of the ipsilateral cortex and striatum region was measured and calculated via ImageJ win32 software (version 1.50, NIH, https://imagej.nih.gov/ij/, USA).

#### *2.8. Assessment of Brain Lesion Volume*

To measure lesion volume of the cortical area of TBI and TBI + quinpirole groups, the tissue slides were stained with cresyl violet and the images were taken with a simple light microscope and analyzed with ImageJ software. The injured areas of the TBI and TBI plus quinpirole groups were first outlined and then carefully calculated. The lesion volume was attained by multiplying the sum of the ipsilateral hemisphere area by the distance between the sections [24].

#### *2.9. Nissl Staining*

To analyze the neuronal cell death and lesion after brain injury, the Nissl staining was performed as described previously [24]. In brief, the slides were washed twice with 0.01 M PBS for 15 min followed by treatment with cresyl violet solution for another 1–15 min. The slides were washed with distilled water and dehydrated with ethanol (70%, 95%, and 100%). The tissue slides were cleared in xylene solution for 3 min and the mounting medium was added to the slides and coverslip was applied. A simple light microscope was used to examine the slides and taken images.

#### *2.10. Cell Culture and Treatment*

The Mouse hippocampal cell line HT22 and Microglial cell line BV2 were grown and maintained in Dulbecco's modified Eagle medium (DMEM) medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (Gibco, Grand Island, NY, USA). The final formulation comprises an additional 1% penicillin/streptomycin sulfate (Gibco, Grand Island, NY, USA). Cells were cultured in a humidified cell culture incubator equipped with a 5% carbon dioxide supply. Cell media was regularly replaced after every 2 days passaged. The cells were subjected to experimental procedures after confirmation of above 80% confluency.

Cell viability assay of mouse hippocampal neuronal HT22 cells was evaluated as described previously [29]. In brief, to know the effect of quinpirole the cells were cultured in 96 well plates (density of 1 × 104 cells) containing Dulbecco's modified Eagle's medium (DMEM) 100 μL. After 24 h, the attached cells were subjected to microglial conditioned media (MCM). The cells were co-treated with three different concentrations of quinpirole (10 μM, 20 μM, and 40 μM) while the control cells were cultured only in DMEM (0.01%).

#### *2.11. Microglial Conditioned Media*

Mouse microglial cell line BV-2 was cultured to above 80% confluency were treated with Lipopolysaccharide (1 μg/mL) (Sigma-Aldrich, St. Louis, MO, USA) dissolved in Cell culturing media. After 24 h, media was aspirated and centrifuged to remove cells and debris. The clear supernatant was collected for further biochemical analysis.

#### *2.12. Statistical Analysis*

The western blot band's results were scanned and analyzed by densitometry using sigma gel software (SPSS Inc., Chicago, IL, USA). ImageJ software (National Institutes of Health, Bethesda, MD, USA) was used for immunohistological analysis, and the obtained values were calculated as the mean ± S.E.M. The data analysis was performed by using one-way ANOVA followed by a post-hoc analysis of variance for control, TBI, and treated groups comparison. The data calculation and graphs were determined by using Prism 5 software (Graph Pad Software, Inc., San Diego, CA, USA). The statistical significance values were considered as *p* < 0.05. Note: \* significantly different between control and brain injury, # significantly different between brain injury and quinpirole treated group.

#### **3. Results**

#### *3.1. Quinpirole Regulated the D2R Expression Level in the Injured Brain and HT22 Cells*

Many studies have reported that D2R agonist increases glial and neuronal cell D2R levels and suppresses the release of various inflammatory cytokines [30,31]. Studies have shown that quinpirole (a D2R agonist) activated D2R and suppressed neuroinflammation following brain injury in a mouse model of intracerebral hemorrhage (ICH) with Parkinson's disease [19]. Based on this evidence, we performed Western blot and confocal microscopy analysis to investigate the effects of quinpirole on D2R expression levels especially in the ipsilateral cortex and striatum of brain-injured mice. Our results showed decreased D2R expression levels in the TBI experimental mice group. Notably, quinpirole treatment (1 mg/kg) significantly increased D2R expression in the quinpirole-treated group compared with the non-quinpirole-treated group of TBI mice (Figure 1a).

Deregulation of GSK3β is a critical step in the development and progression of neurodegenerative diseases via activation of neuroinflammatory processes [32]. Accumulating evidence suggests that the regulation of Akt and GSK3-β attenuates neurodegeneration and neuroinflamation [33]. Research has shown that D2R activation regulates the Akt and GSK3β protein levels [22]. Therefore, we performed Western blot analysis to determine the post-TBI expression levels of p-Akt and p-GSK3β and interleukin (IL)-1β. Our results showed increased expression levels of p-GSK3-β at (Ser 9) and IL-1β and decreased expression levels of p-Akt at (Ser 473) in the ipsilateral cortex of injured mouse brains. However, quinpirole treatment significantly regulated Akt/GSK3-β phosphorylation and reduced the IL-1β expression level in ipsilateral cortex of a damaged mouse brain (Figure 1a).

Furthermore, we investigated the protective role of quinpirole in vivo by in vitro studies. We subjected the HT22 cell line to Microglial conditioned media (MCM) treatment, and the cells were collected 24 h after MCM treatment. The HT22 cells were co-treated with three different concentrations of quinpirole (10 μM, 20 μM and 40 μM). Western blot analysis revealed that MCM-induced inflammatory mediators are associated with neuronal D2R deterioration. We observed that the administration of 10 μM or 20 μM reduced the toxic effect of MCM and significantly regulated D2R, which might be associated with the regulatory activity of p-Akt (Figure 1b).

**Figure 1.** Quinpirole regulates the D2R/Akt/GSK3-β signaling pathway after brain injury. (**a**) Representative Western blot and histogram analysis of D2R, p-Akt, p-GSK3-β, and IL1-β in the ipsilateral cortex of an injured mouse brain. (**b**) Representing the Western blot analysis of D2R in HT22 cells. The β-actin was used as a loading control (*n* = 5). Western blot bands were quantified using the SigmaGel software. (**c**) Image showing results of immunofluorescence testing for D2R expression in the ipsilateral cortex of the control, brain injury, and quinpirole-treated mice groups. (**d**) D2R expression level and p-Akt co-localization in HT22 cells. (**e**) Image showing results of immunofluorescence testing for D2R expression in the ipsilateral straitum of the control, brain injury, and quinpirole-treated mice groups, with respective bar graphs (magnification ×10, *n* = 6). Data were obtained following three independent experiments. The ImageJ software was used for quantitative analysis of the confocal microscopy images and the maximum fluorescent intensity in the representative field was taken(green, FITC; red, TRITC; blue, DAPI). Values are represented as mean ± SEM. We performed the one-way ANOVA test followed by post-hoc analysis. A *p* value < 0.05 was considered statistically significant. \* significantly different between control and brain injury groups, # significantly different between the brain injury and quinpirole-treated groups. ANOVA: analysis of variance, D2R: dopamine D2 receptors, GSK3-β: glycogen synthase kinase 3 beta, IL: interleukin, SEM: standard error of mean.

These results were validated by confocal microscopy. Our results also showed that the immunoreactivity of D2R was lower in the TBI group than in the control group. In contrast, quinpirole treatment improved D2R expression levels in the ipsilateral cortex and striatum after TBI (Figure 1c,e). Co-localization analysis of Akt and D2R revealed that their expression was significantly lesser in the MCM-treated HT22 cells. In contrast, quinpirole at a dose of 20 μM activated D2R and significantly increased Akt expression in HT22 cells (Figure 1d). Next, The p-Akt expression level was analyzed with iba-1, the double Immunofluorescence test result indicated the significantly reduced expression level of p-Akt and significantly increased expression of Iba-1 in the ipsilateral cortex of TBI mouse brain. However, post-TBI quinpirole treatment reversed this effect and significantly

regulated the expression level of these markers in the ipsilateral cortex as compared to the TBI group of mice on day 7 (Figure 2a). We also evaluated Akt and IL1-β expression levels, particularly in the ipsilateral cortex, and confocal microscopy analysis revealed that post-TBI quinpirole treatment significantly regulated the expression level of these markers (Figure 2b,c). Moreover, co-localization analysis of p-GSK3β (Ser 9) and IL1-β showed increased expression levels in the ipsilateral striatum of TBI mice and also confirmed that post-TBI quinpirole treatment significantly reduced p-GSK3β and IL1-β expression levels (Figure 2d). Overall, these results confirm that post-TBI quinpirole administration may protects against neurodegenerative conditions via regulation of the D2R/Akt/GSK3β and IL-1β signaling pathways.

**Figure 2.** Quinpirole treatment reduces neuroinflammation via activation of the iba-1/p-Akt/p-GSK3-β and IL1-β signaling pathways after brain injury. (**a**) Double IF images of iba-1 and p-Akt in the ipsilateral cortex of brain-injured and quinpiroletreated mice. (**b**,**c**) Images showing results of immunofluorescence testing p-Akt (ser9) and IL1-β in the ipsilateral cortex after brain injury. (**d**) Images showing double immunofluorescence of p-GSK3-β (ser9) (FITC-label, green) and IL1-β (TRITC-label, red) (DAPI-label, blue) in the ipsilateral striatum with respective bar graphs, (magnification ×10, *n* = 6). Data were obtained after following three independent experiments. The ImageJ software was used for quantitative analysis of the confocal microscopy images and the maximum fluorescent intensity in the representative field was taken. Values are expressed as mean ± SEM. We performed the one-way ANOVA test followed by post-hoc analysis. A *p* value < 0.05 was considered statistically significant. \* significantly different between control and brain injury groups, # significantly different between the brain injury and quinpirole-treated groups. ANOVA: analysis of variance, FITC: fluorescein isothiocyanate, GSK3-β: glycogen synthase kinase 3 beta, IL: interleukin, SEM: standard error of mean, TRITC: tetramethylrhodamine-isothiocyanate.

#### *3.2. Quinpirole Reduced Gloisis and Atttenates D2R/Akt Level after Brain Injury*

Gliosis plays an important role in the release of pro-inflammatory cytokines, such as IL-1β and tumor necrosis factor (TNF)-α and is a prominent feature of neurodegenerative conditions. Studies have reported that brain injury results in astrocyte and microglial activation, which precipitates further deleterious effects through the release of neuroinflammatory mediators [34–36]. Reportedly, D2R agonists are shown to significantly reduce the activation of astrocytes and the release of TNF-α in the spinal cord of a mouse model of amyotrophic lateral sclerosis and also prevent motor neuron loss (or death). Previously, studies reported the suppression of microglia following D2R activation [37]. While another study was also well suggested that Akt and GSK3β plays an essential role in glial response [38]. Based on these reports, we investigated whether quinpirole treatment could inhibit neuroinflammatory responses in our mouse model of TBI. Therefore, we evaluate glial fibrillary acidic protein (GFAP); a marker of active astrocytes, ionized calcium-binding adaptor molecule 1 (Iba-1); a marker of active microglia together with D2R and p-AKT expression level in ipsilateral or striatum after brain injury. Our double Immunofluorescence test results showed significantly increased immunoreactivity of GFAP and the expression level of D2R and p-AKT was significantly decreased in the ipsilateral cortex of TBI group as compared to saline treated group of mice. However, post-TBI quinpirole treatment reversed this effect, and significantly regulated the expression level of these markers on day 7 (Figure 3a,b). Moreover, we also checked the expression level of iba-1 in the ipsilateral striatum of TBI group of mice. Confocal microscopy result for iba-1 showed the significantly increased expression level of iba-1 in the TBI group of mice. However, the expression level of iba-1 was significantly decreased in quinpirole-treated TBI mice on day 7 (Figure 3c). Interestingly, the results of the Western blot analysis also showed increased Iba-1 and GFAP expression levels, indicating that the number of activated microglia and astrocytes was higher in TBI mice than in mice treated with saline (Figure 3d). Moreover, glial cell activation was significantly lower in the quinpirole-treated group than in the TBI group. Our results show that quinpirole treatment potentially ameliorates TBI-induced glial cell activation on day 7. This condensation of glial cells may be associated with astrocyte and microglial D2R and p-Akt modulation following quinpirole treatment.

#### *3.3. Quinpirole Reduced Neuronal Apoptosis after Brain Injury*

Many studies performed in a TBI mouse model have reported neuronal apoptosis after brain injury, particularly in the perilesional areas and striatum [39,40]. Using Western blot analysis, we investigated apoptotic markers, including Bax, Bcl-2, and PARP1 in the ipsilateral cortex (Figure 4a). We observed that compared with saline-treated mice, brain-injured mice showed a marked increase in neuronal apoptosis. Interestingly, we found significantly lower levels of p-JNK and apoptotic markers in the ipsilateral cortex in the quinpirole-treated group than in the non-quinpirole-treated group. These results were further validated by confocal microscopy. Immunofluorescence test results revealed increased expression of caspase-3 ipsilateral cortex and striatum in brain-injured mice.

Additionally, compared with the TBI group, the quinpirole-treated group showed a significant reduction in the high expression of caspase-3 in the ipsilateral cortex and striatum (Figure 4b,c). We performed Nissl staining to further assess neuronal cell death; compared with the control group, the brain-injured mice group showed a reduced number of surviving neurons in the ipsilateral cortex. Notably, quinpirole treatment reversed this effect and significantly increased the number of surviving neurons in the ipsilateral cortex of quinpirole-treated TBI mice (Figure 4d). These results suggest that the impact of brain injury extend to the ipsilateral cortex and striatum, and quinpirole treatment is known to inhibit neuronal apoptosis possibly via D2R activation in ipsilateral side of injured mouse brain.

**Figure 3.** Quinpirole reduces astrocyte and microglia activation after brain injury. (**a**) Representative confocal microscopy images showing double immunoreactivity of GFAP and D2R expression level in ipsilateral cortex of TBI mouse model. (**b**) Images of double immunoreactivity of GFAP and p-Akt expression level in ipsilateral cortex of TBI mouse model (green, FITC; red, TRITC; blue, DAPI). (**c**) Confocal images of Iba-1 in ipsilateral striatum of brain-injured and quinpirole-treated mice, with respective bar graphs, (magnification ×10, *n* = 6). (**d**) Images of Western blot and histogram analysis showing GFAP and Iba-1 expression levels in ipsilateral cortex of brain-injured and quinpirole-treated mice. The β-actin was used as a loading control (*n* = 5). The ImageJ software was used for immunohistological analysis and the number of GFAP and iba-1 cells were quantified that containing D2R and p-Akt in the representative field. Data were obtained following three independent experiments. The ImageJ software was used for quantitative analysis of the confocal microscopy images. Values are expressed as mean ± SEM. We performed the one-way ANOVA test followed by post-hoc analysis. A *p* value < 0.05 was considered statistically significant. \* significantly different between the control and brain injury groups, # significantly different between the brain injury and quinpirole-treated groups. ANOVA: analysis of variance, GFAP: glial fibrillary acidic protein, Iba-1: ionized calcium binding adaptor molecule 1, SEM: standard error of mean.

**Figure 4.** Quinpirole inhibits brain injury-induced neuronal apoptosis in mice brain. (**a**) Representative images showing results of immunoblot and histogram analysis of p-JNK, Bax, Bcl-2, and PARP-1 proteins in the ipsilateral cortex of injured mouse brain. The β-actin was used as a loading control (*n* = 5). (**b**,**c**) Immunofluorescence test images showing cl-caspase-3 immunoreactivity in the ipsilateral cortex and striatum of injured mouse brain, (green, FITC; blue, DAPI) with respective bar graphs, (magnification ×10, *n* = 6). (**d**) Nissl stain images of the ipsilateral cortex. ImageJ software was used for immunohistological analysis. Data were obtained following three independent experiments. The ImageJ software was used for quantitative analysis of the nissl images and confocal microscopy images. The integrative density of the number of caspase3 positive cells were quantified in the representative field. Values are expressed as mean ± SEM. We performed the one-way ANOVA test followed by post-hoc analysis. A *p* value < 0.05 was considered statistically significant. \* significantly different between the control and brain injury groups, # significantly different between the brain injury and quinpirole-treated groups. ANOVA: analysis of variance, SEM: standard error of mean.

#### *3.4. Quinpirole-Induced Restoration of Blood–Brain Barrier Disruption and Lesion Volume after Brain Injury*

Previous research has shown that TBI results in severe BBB disruption, which invariably leads to severe complications in the affected areas [41]. Activation of astrocytic signaling causes BBB injury through the release of cytokines or chemokines and immune cell recruitment. Therefore, we evaluated the BBB breakdown and the possible role of quinpirole in the restoration of the disrupted BBB in our TBI mouse model. On confocal microscopy, co-localization of zonula occludens-1 (ZO-1) and a cluster of differentiation 31 (CD31) proteins showed that compared with the saline-treated control mice, the braininjured mice showed significantly decreased ZO-1 expression levels in endothelial cells, and compared with brain-injured mice, the quinpirole-treated mice showed elevation of the reduced ZO-1 protein levels and a significant increase in its expression on post-TBI day 7 (Figure 5a). It is well known that brain injury immediately causes gross tissue disruption at the site of injury. Therefore, we also assessed the lesion volume on post-TBI day 7. Histopathological examination of specimens obtained from brain-injured mice showed a marked increase in the contusion and lesion volume in this group, which was significantly reduced following quinpirole treatment (Figure 5b).

**Figure 5.** Quinpirole regulates the BBB-associated ZO-1 and CD31 expression levels and lesion volume after brain injury. (**a**) Representative confocal microscopy images for ZO-1 (TRITC-label, red) and CD31 (FITC-label, green) immunofluorescence reactivity in the ipsilateral cortex in injured mouse brain (green, FITC; red, TRITC; blue, DAPI). (**b**) Representative images showing TBI mouse brain after surgery, and Nissl-stained images showing the lesion volume in the brain injury and quinpirole-treated groups, with respective bar graphs, (magnification ×10, *n* = 6). Data were obtained following three independent experiments. The ImageJ software was used for quantitative analysis of the confocal microscopy images and the percentage of vessels were quantified that containing ZO-1 in the representative field. Values are expressed as mean ± SEM. We performed the one-way ANOVA test followed by post-hoc analysis. A *p* value < 0.05 was considered statistically significant. \* significantly different between the control and brain injury groups, # significantly different between the brain injury and quinpirole-treated groups. ANOVA: analysis of variance, BBB: blood–brain barrier, CD31: cluster of differentiation 31, FITC: fluorescein isothiocyanate, SEM: standard error of mean, TRITC: tetramethylrhodamineisothiocyanate, ZO-1: zonula occludens-1.

#### *3.5. Quinpirole Attenuated Synaptic Dysfunction after Brain Injury*

Previous studies have reported that brain injury causes synaptic protein loss, which leads to memory impairment [35,42]. Therefore, we evaluated the expression levels of synaptic proteins, including synaptosomal-associated protein 23 (SNAP-23) and postsynaptic density protein 95 (PSD-95) in our TBI mouse model. We performed confocal microscopy for PSD-95 and SNAP-23 in ipsilateral cortex and striatum respectively. We observed that brain injury significantly decreases synaptic protein expression, whereas quinpirole treatment significantly increased synaptic protein loss in an injured mouse brain (Figure 6a,b). Western blot analysis revealed that compared with saline-treated mice, brain-injured mice showed reduced PSD-95 and SNAP-23 expression in the ipsilateral cortex (Figure 6c). However, quinpirole treatment significantly restored synaptic proteins following brain injury.

**Figure 6.** Quinpirole regulates synaptic protein loss after brain injury. (**a**,**b**) Confocal microscopy images for PSD-95 and SNAP-23 expression in the ipsilateral cortex and striatum of an injured mouse brain (green, FITC; blue, DAPI), with respective bar graphs, the red dotted lines showing the striatum region, (magnification ×10, *n* = 6). The protein band levels were quantified using the SigmaGel software. (**c**) Images showing results of Western blot and histogram analysis for PSD-95 and sanp-23 in ipsilateral cortex of injured mouse brain. The β-actin was used as a loading control (*n* = 5). Data were obtained following three independent experiments. The ImageJ software was used for quantitative analysis of the confocal microscopy images. Values are expressed as mean ± SEM. We performed the one-way ANOVA test followed by post-hoc analysis. A *p* value < 0.05 was considered statistically significant. \* significantly different between the control and brain injury groups, # significantly different between the brain injury and quinpirole-treated groups. ANOVA: analysis of variance, PSD-95: post-synaptic density protein 95, SEM: standard error of mean, SNAP-23: synaptosomal-associated protein 23, TBI: traumatic brain injury.

#### **4. Discussion**

An optimal therapeutic approach to brain injuries is unavailable owing to the multifactorial pathogenesis of brain trauma. Brain injuries lead to cognitive dysfunction that can be prevented by DA therapies targeted at the restoration of cognitive impairment [43,44]. The most important neurotransmitters in the central nervous system: glutamate is released from multiple stores after a TBI and the activation of D2Rs could contribute in the modulation of the glutamate release both from neurons than from astrocytes. Hence, activation of D2Rs may play essential role after TBI-induce disturbance in neurotransmitters. In this study, we observed that quinpirole (a D2R agonist) plays a significant role in brain injury-induced neuroinflammation, neurodegeneration, and synaptic dysfunction. We focused on the neuroprotective effect of quinpirole following brain injury in mice. This report shows that post-TBI quinpirole administration attenuates several neuropathological events, such as glial cell activation, neuroinflammation, neuronal apoptosis, and synaptic dysfunction via the D2R/Akt GSK3β/IL-1β signaling pathways. Since the TBI-induced striatal glial activation and expression of pro-inflammatory cytokines and therapeutic potential of D2R activation in the ipsilateral striatum is mostly known previously as compared to the ipsilateral cortex; thus, still it is essential to investigate the therapeutic potential of D2R in the

cortex of TBI mouse brain [45]. Therefore, we investigated the neuroprotective effect of D2R activation mainly in the ipsilateral cortex, while we checked slightly the ipsilateral striatum of the injured mouse brain. Our results suggest that quinpirole activates D2R, which plays a crucial role in several neuropathological events particularly in ipsilateral cortex after brain injury.

Brain injury leads to neuroinflammation, which contributes to severe neurodegeneration. Studies have reported chronic neuroinflammatory responses in the cortex and hippocampus of an injured mouse brain [46,47]. In our TBI mouse model, we observed increased neuroinflammation indicated by microglial and astrocyte activation in the ipsilateral cortex and striatum. Interestingly, post-TBI quinpirole treatment significantly reduced the increased gliosis and release of pro-inflammatory markers. Our results are consistent with those reported by previous studies [12]. It is known that the regulation of p-GSK3-β via p-Akt is involved in the cell survival pathway [48]. A previous study showed that injury-induced disruption of Akt and GSK3β expression in glial cells is a major contributor to the mechanistic of glial cell adaptation as well as protection in response to cell damage. Thus Akt and GSK3β play an essential role in glial response and excitotoxic lesion outcome of injury [38]. Another study investigated the role of Akt/GSK3β pathway in acute brain injury after subarachnoid hemorrhage [49,50]. The regulation of GSK3-β and Akt via D2R could be a novel therapeutic approach following brain injury. In the present study, we investigated the protective effect of quinpirole mediated via the D2R/GSK3β/Akt signaling pathway. We found decreased expression of D2R/Akt and increased expression of p-GSK3-β and IL-1β after TBI, based on Western blot and immunofluorescence analysis. However, these levels normalized to the baseline levels in quinpirole-treated mice. A previous study also reported the anti-neuroinflammatory effect of quinpirole via D2R activation in an ICH injury model [19]. The BBB plays a central role in brain homeostasis. However, BBB disruption leads to enhanced cytokine infiltration and neuronal susceptibility.

Several tight junction proteins, including claudin, occludin, and ZO-1, are essential for the maintenance of BBB integrity [51]. BBB breakdown following brain injury is attributable to significant histopathological alterations and tissue loss in the affected areas [52]. Double immunofluorescence staining performed for ZO-1 and CD31 showed significantly low levels of these proteins in the ipsilateral cortex of an injured mouse brain. Notably, quinpirole treatment restored ZO-1 and CD31 levels in the ipsilateral cortex of an injured mouse brain. Quinpirole-regulated restoration of the disrupted BBB is attributable to reduced neuroinflammation and active gliosis. A previous study has reported that brain injury is strongly associated with deregulated tight junction proteins [53]. Brain injury is known to cause marked tissue disruption [34]. We observed increased lesion volume in an injured mouse brain, and that quinpirole treatment significantly reduced the lesion volume, which suggests that quinpirole aids in the repair of the brain after injury and restores tight junction proteins to inhibit infiltration of cytokines and other blood-borne biochemical agents. Moreover, we observed a significant increase in the contusion volume after brain injury, indicating that severe damage is associated with tissue disruption in the ipsilateral cortex of injured mouse brain. Notably, all these effects were ameliorated in brain-injured mice that received quinpirole treatment.

Increasing evidence has shown neuronal apoptosis after brain injury [54]. Our results showed increased expression of neuronal apoptotic markers, including Bax and PARP1, and decreased expression of Bcl-2, an anti-apoptotic protein. Quinpirole treatment significantly reduced the increased levels of pro-apoptotic and increase the reduced level of anti-apoptotic markers in the ipsilateral cortex of brain-injured mice. A previous study supports the protective role of D2/D3 receptor agonist ropinirole protects against apoptosis-induced neurodegeneration via a JNK-dependent pathway [11]. Our results are consistent with those of a previous study in which D2R agonists were shown to reduce neuronal apoptosis [55]. In accordance with the caspase3 result, the nissl staining results also showed the significantly increase number of apoptotic and degenerated neurons in the ipsilateral side of TBI mouse brain as compared to saline- treated control group of

mice. However, the number of damage and degenerated neuronal cells were significantly reduced in quinpirole-treated group of mice after TBI on day 7.

Synaptic protein loss is associated with brain injury and leads to cognitive deficits and impaired neurotransmission. A study has reported that DR activation protects against amyloid-β oligomer-mediated synaptic dysfunction. Therefore, we evaluated synaptic protein markers after brain injury. The results of Western blot and immunofluorescence testing showed that quinpirole treatment reversed the deregulated levels of synaptic protein markers, including PSD-95 and SNAP-23 in TBI mice. Moreover, other studies have also reported that D2R is vital for several brain functions, including learning and working memory. We concluded that quinpirole could be a potentially useful therapeutic agent to restore synaptic function after brain injury and to improve the cognitive performance of brain-injured mice.

These results suggest that brain injury may cause D2R suppression, which consequently activates deleterious signaling pathways at a later stage after brain injury, and that quinpirole-mediated D2R activation produces a neuroprotective effect in the ipsilateral cortex and striatum of injured brains.

#### **5. Conclusions**

This study highlights the significant role of D2R in neurodegenerative conditions affecting the ipsilateral cortex after brain injury and that D2R regulation might be an effective therapeutic strategy to inhibit glial cell-induced neuroinflammation in a mouse model of brain injury (Figure 7). In this study, we discuss the role of quinpirole (a D2R agonist), that can potentially attenuate several neuropathological processes via D2R/Akt/GSK3-β/IL-1β signaling in the ipsilateral cortex and striatum of an injured mouse brain. Further studies are warranted to gain a deeper understanding of the molecular mechanisms contributing to the neuroprotective effects of quinpirole via D2R activation in neuropathological events associated with brain injury.

**Figure 7.** Schematic representation of the proposed mechanism of neuroprotection of quinpirole against brain injury-induced neuroinflammation, BBB disruption and neurodegeneration via D2R and Akt/GSK3-β/IL-1β signaling in the injured mouse brains.

**Author Contributions:** S.I.A. designed the study, performed the experimental work, and wrote the manuscript. M.G.J. reviewed the manunsript and edited the figures. R.U. and T.J.P. review and edited the manuscript. S.U.R. help in writing the manuscript. S.A. conducted experiment. M.O.K. is a corresponding author, reviewed and approved the manuscript, and holds all the responsibilities related to this manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the Neurological Disorder Research Program of the National Research Foundation (NRF) funded by the Korean Government (MSIT) (2020M3E5D9080660).

**Institutional Review Board Statement:** This study was carried out in animals in accordance with approved guidelines (Approval ID: 125) by the animal ethics committee (IACUC) of the Division of Applied Life Science, Gyeongsang National University, South Korea.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The authors hereby declares that the data presented in this study will be presented upon request from the corresponding author.

**Acknowledgments:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Review* **Relevance of Autophagy and Mitophagy Dynamics and Markers in Neurodegenerative Diseases**

**Carlotta Giorgi 1, Esmaa Bouhamida 1, Alberto Danese 1, Maurizio Previati 2, Paolo Pinton <sup>1</sup> and Simone Patergnani 1,\***


**Abstract:** During the past few decades, considerable efforts have been made to discover and validate new molecular mechanisms and biomarkers of neurodegenerative diseases. Recent discoveries have demonstrated how autophagy and its specialized form mitophagy are extensively associated with the development, maintenance, and progression of several neurodegenerative diseases. These mechanisms play a pivotal role in the homeostasis of neural cells and are responsible for the clearance of intracellular aggregates and misfolded proteins and the turnover of organelles, in particular, mitochondria. In this review, we summarize recent advances describing the importance of autophagy and mitophagy in neurodegenerative diseases, with particular attention given to multiple sclerosis, Parkinson's disease, and Alzheimer's disease. We also review how elements involved in autophagy and mitophagy may represent potential biomarkers for these common neurodegenerative diseases. Finally, we examine the possibility that the modulation of autophagic and mitophagic mechanisms may be an innovative strategy for overcoming neurodegenerative conditions. A deeper knowledge of autophagic and mitophagic mechanisms could facilitate diagnosis and prognostication as well as accelerate the development of therapeutic strategies for neurodegenerative diseases.

**Keywords:** autophagy; mitophagy; neurodegeneration; multiple sclerosis; Alzheimer's disease; Parkinson's disease; biomarker; therapy

#### **1. Introduction**

Neurodegenerative disorders refer to a large group of pathological conditions in which components of the nervous system lose their structure and function. These diseases are primarily classified according to the clinical features (dementia, tremor, rigidity, and bradykinesia) but especially according to the anatomic distribution of the neurodegenerative lesions [1]. The aggregation of misfolded brain proteins represents the main cause of neuronal damage in hereditary and sporadic neurodegenerative disorders (Figure 1) [2]. Alzheimer's disease (AD) is a late-onset form of dementia and is considered the most frequent type of neurodegeneration worldwide: nearly 50 million people live with AD and related dementia. AD is characterized by a progressive loss of neurons determined by the aberrant accumulation of tau protein and beta-amyloid protein (Aβ protein) [3]. Human brains express six isoforms of tau, whose cellular function is to stabilize the interactions of microtubules with other proteins. To exert this function, tau protein must be phosphorylated. However, in AD, tau is hyperphosphorylated, a condition that induces conformational changes and the aggregation of the tau protein [3]. The second most common neurodegenerative disease is Parkinson's disease (PD). Persons affected by this disease present some common symptoms (such as anxiety, depression, rigidity, and

**Citation:** Giorgi, C.; Bouhamida, E.; Danese, A.; Previati, M.; Pinton, P.; Patergnani, S. Relevance of Autophagy and Mitophagy Dynamics and Markers in Neurodegenerative Diseases. *Biomedicines* **2021**, *9*, 149. https://doi.org/10.3390/ biomedicines9020149

Academic Editor: Arnab Ghosh Received: 30 December 2020 Accepted: 1 February 2021 Published: 4 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

tremor) that worsen over time [4]. PD belongs to a class of neurodegenerative diseases named "synucleinopathies", since the protein aggregations typical of the disease, Lewy bodies (LBs), are formed by different types of proteins, and α-synuclein (α-syn) is the major constituent. Several studies have demonstrated that the aggregation and oligomerization of α-syn are determined by alternative splicing events and by post-translation modifications, such as ubiquitination, oxidative nitration, truncation, and phosphorylation [5]. Furthermore, genetic studies on the familial form of PD have unveiled that specific mutations in the encoding gene, SNCA, accelerate the production of insoluble aggregates and oligomers [4]. Neuronal damage, neuronal loss, and a reduction in brain volume are also important factors inducing long-term disability in patients affected by multiple sclerosis (MS), a complex and multifactorial disorder leading to severe physical or cognitive disabilities and neurological defects [6]. Unlike other types of neurodegenerative disorders, MS is not due to the excessive accumulation of misfolded proteins but is the result of a state of persistent inflammation and adverse immune-mediated processes that activates a cascade of molecular events that provoke demyelination in the white as well as gray matter and subsequent axonal and neuronal damage [7]. Despite the existence of fundamental differences among these neurodegenerative conditions, a growing body of evidence demonstrates that they share important pathogenic mechanisms, of which mitochondrial (dys)function, inflammation, infection, and immune responses are the most frequent [8–10]. In addition, in recent years, the autophagy process has also been found to be particularly associated with neurodegeneration [11]. Autophagy is a cellular catabolic pathway in which cytosolic components, bacteria, viruses, macromolecules, and whole organelles are transported to lysosomes for degradation. To exert these multiple functions, specialized forms of autophagy also exist, the most studied being mitophagy (the selective removal of damaged mitochondria) [12]. Autophagy itself and its specialized forms play important roles in physiological as well as pathological conditions. Under normal conditions, autophagy removes unnecessary material, regulates the physiological turnover of organelles, and meets energetic demand. In pathological conditions, autophagy may have both favorable and deleterious roles. As demonstrated, a loss/gain of function in the autophagic process and increase/decrease in the expression of crucial autophagic mediators have been associated with diverse human disorders, particularly cancer [13,14] and neurodegeneration [15]. Most importantly, different studies have not only confirmed the importance of autophagic dynamics during neurodegeneration but also suggested that several proteins involved in this catabolic process may be considered potential markers for predicting neurodegenerative conditions [16,17]. Analysis of the distribution of autophagy partners and regulators along the different pathologic steps of neurodegenerative disorders may improve the knowledge of the contribution of autophagic processes to neurodegenerative conditions. In addition, we may develop innovative neuroprotective therapies and unveil new potential biomarkers for the early diagnosis and clinical management of these diseases. In this review, we discuss the roles of autophagy and its specialized form mitophagy in different neurological disorders. We explore the possibility of using molecular partners of autophagy and mitophagy as biomarkers for neurodegenerative disease status. Finally, the pharmacological modulation of these processes is discussed as a potential strategy for building new therapeutic approaches against neurodegeneration.

**Figure 1.** Causes and risk factors of Alzheimer's disease (AD), Parkinson's disease (PD), and multiple sclerosis (MS). AD and PD are the most common neurodegenerative diseases, characterized by an abnormal aggregation of protein inclusions in the brain. While in AD, the progressive loss of neurons is caused by an aberrant accumulation of beta-amyloid and tau protein, PD displays inclusions named Lewy bodies, where α-synuclein (α-syn) is the major constituent. MS is the most frequent cause of disability among young adults after traumatic brain injury. The neurodegenerative condition of MS is not due to an excessive accumulation of misfolded protein but to adverse immune-mediated processes and chronic inflammation that provoke demyelinating and neurodegenerative processes during the entire life of the patient. Among the different molecular mechanisms and risk factors involved in these neurodegenerative conditions, it has been demonstrated that autophagy and mitophagy play an important role.

#### **2. A General Overview of Autophagy**

The word autophagy was introduced in late 1963 by the biochemist Christian de Duve [18] and defines a self-degradative cellular pathway whose intent is to degrade and recycle cellular contents. Autophagy exists in three forms that are classified according to their mechanisms and cellular functions: macroautophagy, microautophagy, and chaperone-mediated autophagy (CMA). During microautophagy, the cytosolic material is wrapped and transported directly into the lumen of lysosomes. The main function of microautophagy (mA) is to control cell survival and organellar turnover upon nitrogen restriction. Unfortunately, due to the lack of specific methods for measuring mA (apart from electron microscopy), the effective contributions of mA in mammalian cells remain little studied, and most studies about mA molecular processes are carried out in yeast [19]. Despite this, different investigations suggest that the molecular dynamics of mA existing in yeast may be conserved in mammalian mA. Consistently with this, it has been demonstrated that the endosomal sorting complex required for transport (ESCRT) system is involved in mammalian [20] and yeast mA [21]. Furthermore, a prolonged starvation condition [22,23] as well as cellular treatments with the macrolide compound rapamycin activates mA in both mammalian and yeast cells [21,24,25].

CMA has an important role in protein quality control (QC) and is responsible for degrading a specific subset of oxidized and damaged proteins. The selectivity of CMA is conferred by the existence of a specific pentapeptide motif (KFERQ), which is present in the amino acid sequences of all CMA substrates. This motif is identified by the cytosolic chaperone heat shock-cognate protein of 70 kDa (hSC70), which brings the protein target directly to the lysosome surface [26]. In the last decade, several advances have been made in understanding the molecular mechanisms of CMA. These findings suggest an important contribution of CMA to diverse human diseases, including neurodegeneration [26]. Undoubtedly, the best-characterized and most prevalent form of autophagy in mammalian cells is macroautophagy (hereafter referred to as autophagy), whose multistep process and contribution to the pathophysiology of diverse neurodegenerative conditions will be discussed throughout this review.

Autophagy, a complex intracellular process that is very ancient and has been strongly conserved during evolution, exists to identify and capture a wide group of intracellular components, ranging from low-dimensional biological macromolecules to whole organelles, and bring them to the lysosomal compartment. Its physiological value rests on two main activities. On the one hand, autophagy acts as a QC mechanism that reshapes the cell, ensuring the removal of damaged proteins and organelles [27]. Selective forms of autophagy can specifically target mitochondria (mitophagy), the endoplasmic reticulum (ER; reticulophagy), peroxisomes (pexophagy), and lipid droplets (lipophagy). In addition, autophagy participates in the struggle against invading pathogens (xenophagy), inducing cell defense [12].

On the other hand, lysosomal degradation represents an important source of amino acids and lipids for the de novo synthesis of proteins and lipids. This is of particular importance during starvation, which limits amino acid availability. The limited availability of amino acids affects protein synthesis, which can be performed only in the presence of all the necessary building blocks, in particular, essential amino acids. Under shortage conditions, amino acid pool completeness can be fulfilled only through the degradation of cellular proteins. In such a way, autophagy represents a fundamental survival mechanism, particularly during stress conditions originating from hypoxia or pathogen invasion [27].

Thus, it is not surprising that energy availability can regulate or trigger autophagy and, in particular, that a large number of stimuli converge on metabolic energy sensors, such as mammalian target of rapamycin (mTOR) and 5 adenosine monophosphate-activated protein kinase (AMPK), which, in turn, regulate autophagy [28].

In cells, mTOR exists in two complexes, mTORC1 and mTORC2, which not only are composed of different protein-binding partners but also regulate different pathways. The primary role of mTORC2 is to regulate cell survival and cytoskeletal organization, while its role in autophagy remains poorly understood. Recent work has shed light on this obscure point. Indeed, the transforming growth factor beta (TGFB)/INHB/activin signaling pathway has been recently identified as an upstream regulator of mTORC2. TGFB-INHB/activin mediates mTORC2 inhibition and regulates the autophagic flux and the cardiac functions in a *Drosophila* cardiac-specific knockdown of TGFB-INHB/activin model [29]. Another investigation recently confirmed the importance of mTORC2 for autophagy. In this case, it has been demonstrated that mTORC2 exists on a molecular axis with the serum- and glucocorticoid-inducible kinase 1 (SGK-1) and, in this state, controls autophagy and mitophagy induction. Consistently, mTORC2- or SGK-1 deficient *C. elegans* models present a perturbed mitochondrial homeostasis and aberrant ROS production, which trigger autophagy and mitophagy. Excessive autophagic and mitophagic fluxes, in turn, result in developmental and reproductive deficits in mTORC2- or SGK-1-deficient animals [30]. Oppositely, the primary role of mTORC1 is to play a pivotal role in cellular catabolic pathways, particularly autophagy [31]. To exert its function, mTORC1 integrates different stimuli, including hormonal stimulation, nutrient availability, and the oxygen level. In the presence of normal levels of energy and amino acids, mTOR inhibits autophagy through specific unc-51-like autophagy-activating kinase 1 (ULK1) serine phosphorylation at the phosphorylation site Ser 757. By contrast, in response to nutritional deprivation, oxygen unavailability, and mitochondrial dysfunction, AMPK activates autophagy through

the phosphorylation of ULK1 at Ser 317 and Ser 777 [32]. Interestingly, another research group demonstrated that AMPK may phosphorylate ULK in additional sites. Indeed, by employing a bioinformatic approach, it has been found that ULK1 contains a further four potential AMPK sites [33]. Three of them (Ser 555, Ser 637, and Thr 574) were also identified by mass spectrometry in cells pretreated with an AMPK activator, while the site Ser 467 was confirmed by immunoblotting with phosphospecific antibodies [33]. Unfortunately, this work lacks an analysis of the effect of the different phosphorylations on autophagy. By using SILAC (stable isotope labeling with amino acids) technology, other work mapped 13 new phosphorylation sites of ULK1 [34]. All of them were dependent on nutrient availability, but only Ser 638 and Ser 758 displayed the most significant changes. In addition, time course experiments investigating the response to nutrient availability demonstrated that these phosphorylations were differentially regulated and that mTOR mediated both phosphorylations. Intriguingly, the authors also demonstrated that the phosphorylation at Ser 638 was also mediated by AMPK [34]. Altogether, these findings demonstrate that ULK1 is the key regulator of autophagy, and the occurrence of different protein phosphorylation events is crucial for regulating its activity. Furthermore, the concurrent existence of at least two opposite regulatory pathways that converge on ULK1 signaling (mediated by MTOR and AMPK) allows the cell to better adapt to extracellular and intracellular variations but also affects several pathological conditions.

In the cells, ULK1 forms a complex with autophagy-related (ATG) 13/200-kDa focal adhesion kinase family-interacting protein (FIP200) and ATG101. As reported above, ULK1 activity is mainly regulated by phosphorylation/desphosphorylation events mediated by AMPK and mTOR. In addition, it has been demonstrated that ULK1 is able to phosphorylate itself at Thr 180 [35] and FIP200, ATG13, and ATG101 [36,37] and that the phosphorylation events are regulated by protein phosphatase. Protein phosphatase 2A (PP2A) and protein phosphatase 1D magnesium-dependent delta isoform (PPM1D) regulate the ULK1 phosphorylation [38,39]. PP2C phosphatases (Ptc2 and Ptc3) mediate the dephosphorylation of ATG13 30655342. The ULK1/ATG13/FIP200/ATG101 molecular axis represents the most upstream regulatory complex related to double-membrane vacuole (autophagosome) formation [28]. Autophagosomes symbolize the starting moment of the whole autophagic process, which begins with the formation of double-membrane lined vesicles that fuse together to engulf portions of the cytoplasm. The resulting double-membrane vacuoles are autophagosomes, which can fuse with vesicles of the endocytic pathway at different stages of maturation or directly with lysosomes, becoming autolysosomes. In autolysosomes, acidic hydrolases break down macromolecules into smaller constituents that are released back to the cytosol by lysosomal transporters and permeases. Once activated, the ULK1/ATG13/FIP200/ATG101 molecular axis also phosphorylates and activates coiled-coil, moesin-like BCL2 interacting protein (BECN1) [40,41]. BECN1 can be part of a complex including class III phosphatidylinositol 3-kinase (PI3K) and its regulatory proteins vacuolar protein sorting 34 (Vsp34), p150, and ATG14L. Upon activation, this complex is involved in the nucleation and elongation of autophagosomes. The first step occurs on the surface of the membranes of the ER, mitochondria, Golgi complex, endosomes, or plasma membrane [42] and consists of the phosphorylation of phosphatidylinositol to form phosphatidylinositol-3-phosphate (PI3P). This phosphoinositide behaves as a positive regulator of autophagy. In fact, the presence of PI3P at the source membrane triggers the docking of several adaptor proteins, which, in turn, induce and sustain the elongation of the sack-like, omega-shaped structure, which grows, binds, and surrounds the material intended to be digested.

Another interaction of BECN1 can exert an inhibitory effect on autophagy [43]. BECN1 has been reported to bind B-cell lymphoma (BCL)-2, BCL-XL, and other members of the BCL-2 family through the BCL-2-homology-3 (BH3) domain. The consequence of this interaction is a diminution of the interaction between BECN1 and the class III PI3K complex, which prevents the formation of phagophores [43]. Accordingly, BCL-2 phosphorylation can reverse BECN1 sequestration and restore autophagy stimulation [43].

The other two systems, ATG12–ATG5–ATG16L1 and microtubule-associated protein 1A/1B-light chain 3 (LC3)–phosphatidylethanolamine (PE) complexes, seem to play an important role in the elongation and closure of autophagosomes, although the underlying mechanism has not yet been clarified. A key process during autophagosome elongation and closure is the lipidation of the LC3 protein, which is joined to the membrane PE. Once inserted into the autophagosomal membrane, the lipidated complex can further recruit other adaptor proteins. This allows autophagosomes to recognize cargo material, and elongate and close the vesicle. The fusion of the autophagosomes with the lysosome is the subsequent step, which, in a normally operating lysosome, is followed by lysosomal compartment acidification, the degradation of macromolecules by hydrolases and lipases, and the recycling of the base constituents (Figure 2).

**Figure 2.** Molecular mechanisms of autophagy and mitophagy. The mammalian target of rapamycin (mTOR) and the 5 adenosine monophosphate-activated protein kinase (AMPK) are the main negative and positive regulators of autophagy, respectively. One of the primary targets of the action of mTOR and AMPK is the unc-51-like autophagy-activating kinase 1 (ULK1)/autophagy-related (ATG) 13/FIP200 (200-kDa focal adhesion kinase family-interacting protein) complex, which is the main regulator of autophagosomal formation. Other important proteins that participate in this molecular process are the coiled-coil, moesin-like BCL-2 interacting protein (BECN1), class III phosphatidylinositol 3-kinase (PI3K), vacuolar protein sorting 34 (Vsp34), ATG14L, p150, and IMPase. The activity of BECN1 in regulating the autophagy process is also mediated by the interaction with BCL-2. During the elongation of the autophagosome, a series of autophagyrelated (ATG) proteins are involved. In particular, two specific complexes were found to be essential for completing autophagosomal formation: (ATG)12–ATG5–ATG16L1 and microtubule-associated protein 1A/1B-light chain 3 (LC3)– phosphatidylethanolamine (PE) complexes. Mitochondria are particularly vulnerable to stress signals, such as ROS, which, in turn, can cause severe mitochondrial dysfunction and activate the mitophagic process. PINK1 senses this mitochondrial damage and phosphorylates and recruits Parkin to the outer mitochondrial membrane of the mitochondria. Phosphorylation converts Parkin to an active ubiquitin (Ub)-dependent enzyme and mediates the phosphorylation of different mitochondrial proteins. During this process intervene different Ub-binding autophagy receptors such as p6, NBR1, NDP52, and optineurin (OPTN), which connect the damaged mitochondria to the forming autophagosomes. Mitophagy may also be executed in a Parkin-independent manner. In this case, different proteins (FUNDC1, AMBRA1, NIX, and BNIP3) intervene to signal the mitochondria that should be degraded.

#### **3. Mitophagy: The Master Regulator of the Mitochondrial Population**

Mitochondria are essential intracellular organelles that supply substrates and energy to execute numerous cell functions, such as metabolism, differentiation, apoptosis, cell movement, and differentiation. In contrast to other intracellular components, mitochondria are constituted by two membranes, the outer mitochondrial membrane (OMM) and the inner mitochondrial membrane (IMM), which fully surround the mitochondrial matrix. Between the OMM and IMM, another mitochondrial subcompartment exists, the intermembrane space (IMS) [44]. Another unique feature of mitochondria is that they have their own genome (mitochondrial DNA, mtDNA), which encodes 13 proteins that are essential components of the oxidative phosphorylation (OXPHOS) system, the process by which ATP is formed [45]. A series of members (complexes I-V, C-I-V) of the mitochondrial electron chain (mETC) found in the IMM permit the transfer of electrons from NADH or FADH2 to O2 [46]. The energy produced during this movement creates a proton gradient that is used by the last component of the mETC (C-V, ATP synthase) to synthesize ATP [46]. The impairment of electron transfer or stress conditions affect the production of reactive oxygen species (ROS), of which C-I and C-III are the main producers [47]. Mitochondria are also central hubs for calcium (Ca2+) signaling [48]. At rest, mitochondria have low Ca2+ concentrations [Ca2+] (~100 nM range or lower). However, upon stimulation, mitochondrial [Ca2+] can increase to the range of hundreds in micromolar concentration [49]. This happens due to the highly specialized contact sites (mitochondria-associated membranes, MAMs) that exist between mitochondria and the main intracellular Ca2+ store of cells, the ER [50]. These interaction sites represent critical hubs for the regulation of diverse cellular processes (such as energy metabolism, inflammation, redox regulation, and lipid and protein transfer), and recently, MAMs have been described to play an important role in the onset and progression of several human diseases by regulating Ca2+ transmission between the ER and mitochondria [51]. Once released from the ER, Ca2+ can enter mitochondria owing to the close proximity of the ER to mitochondria, the electrochemical driving force (mitochondrial membrane potential) that is created by electron transfer, and the activity of the components of the mitochondrial Ca2+ uniporter (MCU) [52,53]. Mitochondria are normally present in cells in the form of a dynamic network, where the mitochondrial mass increases as a consequence of mitochondrial biogenesis. The control and reshaping of the mitochondrial population can occur through different mechanisms [54]. These mechanisms include (i) the control of protein quality through mitochondrial proteases, the mitochondrial unfolded protein response, or proteasome-dependent degradation; (ii) the budding of mitochondrion-derived vesicles; and (iii) the targeting of some or all mitochondria to lysosomes through mitophagy.

Mitophagy regulation is not yet a completely understood process. During shortterm starvation, the mitochondrial pool is not depleted, so as to not further reduce the cellular production of energy, while oxidative metabolism is mainly sustained by general autophagy [28]. This fact necessarily implies a difference in regulation between autophagy and mitophagy that allows the cautious sparing of mitochondria, which are among the principal end-users of the material provided by autophagy. A role in this sense seems to be played by fission restriction. In fact, fragmented mitochondria appear to be a preferred target for mitophagy: when their number is reduced, mitophagy itself is restricted.

When the ultimate goal is to eliminate mitochondria, there are different physiological mechanisms that can be activated. The first example is programmed mitophagy. There are several situations in the cell that can require the activation of programmed mitophagy, independent of the wellness of mitochondria. An example is the mitochondrial depletion that occurs in reticulocytes during differentiation through the activity of NIP3-like protein X (NIX/BNIP3L). Other examples include the elimination of male-derived mitochondria after egg fertilization [55] and the reshaping of the mitochondrial population during cardiomyocyte [56] or muscle cell differentiation, which induces a change from carbohydrateto fatty acid-driven OXPHOS [57]. Stimulations that can normally trigger mitophagy can

be affected by mitochondrial defects, such as a decline in transmembrane potential and excessive ROS production.

Mitophagy involves some fundamental steps. First, as stated above, mitochondria must assume the dimensions necessary to easily enter autophagosome vesicles. Therefore, they are normally resized through fission processes. In addition, they need to be properly displayed on the surface to trigger the formation of vesicles, which will engulf them. Typically, "eat-me signals" can be ubiquitin-dependent or not. The best-known example of a ubiquitin-dependent mechanism is the PTEN-induced kinase 1 (PINK1)/Parkin axis. PINK1 and Parkin belong to a series of genes referred to as PARK genes, which include α-syn (PARK1/4), Parkin (PARK2), PINK1 (PARK6), protein deglycase-1 (DJ-1, PARK7), leucine-rich repeat kinase 2 (LRRK2, PARK8), and ATP13A2 (PARK9). The name of this group of genes (Parkin genes) comes from the finding that mutations in these genes have been linked to familiar forms of PD. In particular, approximately 100 mutations in the Parkin gene have been identified as causing autosomal recessive Parkinsonism [58].

PINK1 is a mitochondrial serine/threonine-protein kinase, and Parkin is an E3 ubiquitin ligase; these proteins induce different functions at the cellular level but act in a common pathway to regulate mitophagy.

PINK1 is a ubiquitous protein characterized by a mitochondrial targeting sequence (MTS), a transmembrane domain, and a highly conserved serine/threonine kinase domain. At present, approximately 30 pathogenic PINK1 mutations that impair its kinase activity and provoke loss of function have been identified [59–62].

Normally, PINK1 is imported into mitochondria via the activity of the translocase of the inner membrane (TIM)–translocase of the outer membrane (TOM) complex. Once PINK1 arrives in the IMM, it is subjected to a series of proteolytic cleavages that reduce the full-length form of PINK1 into fragments, which are then degraded by the proteasome [63–65]. In the presence of alterations in mitochondrial membrane potential, the activity of the TIM/TOM complex is reduced, and PINK1 begins to accumulate on the OMM. Here, after being stabilized by a molecular complex including TOM proteins [66,67], PINK1 phosphorylates Parkin. The phosphorylation converts Parkin from an autoinhibited enzyme to an active ubiquitin (Ub)-dependent enzyme [68,69]. In this state, Parkin actively ubiquitinates several mitochondrial proteins at the OMM. The ubiquitination events promote the recruitment of the Ub-binding autophagy receptors p62/Sequestome, NBR1, NDP52, optineurin (OPTN), and TAX1BP1 (TBK1), which connect damaged mitochondria to phagosomes for clearance in lysosomes [70–72]. In recent years, different studies have identified pathways regulating mitophagy that are PINK1–Parkin-independent. These mechanisms may act in parallel or in addition to PINK1–Parkin-dependent mitophagy and involve a series of OMM mitophagy receptors that bind LC3 and recruit mitochondria to autophagic vesicles. Among them, the most studied are the proapoptotic members of the BCL2 family, NIX and BNIP3 [73,74] and FUNDC1 [75], which regulate the mitophagy process during ischemic/hypoxic conditions, and the BECN1 regulator AMBRA1. Interestingly, it has been proven that AMBRA1 regulates both Parkin-dependent and Parkin-independent mitophagy [76] (Figure 2).

#### **4. Relationship between Autophagy and Mitophagy in MS**

Multiple sclerosis (MS) is a progressive and chronic disease that affects approximately 3 million persons worldwide. MS is an inflammatory condition in which activated immune cells enter the central nervous system (CNS) and cause progressive demyelination, gliosis, and neuronal loss. The symptoms vary from individual to individual [77]. The most common symptoms are walking difficulties, sensory disturbances, vision problems, and cognitive and emotional impairments. Typically, MS starts with an unexpected onset of neurological impairments, and the majority of individuals display a relapsing–remitting (RR) course of the disease in which recurrent periods alternate with relapse phases. This course may be followed by a secondary progressive phase in which inflammatory attacks are more frequent and cause irreversible neurological impairments. A small percentage

of individuals may present with the primary progressive form of the disease, which is characterized by the absence of remission periods and a progressive worsening of symptoms [78]. Currently, the pathogenesis and etiology of MS are unclear. MS is considered a multifactorial disease, and genetic predisposition and environmental factors may play important roles in disease progression. Furthermore, mitochondrial dysfunction as well as the impairment of the QC systems of mitochondria have been identified in different MS samples and represent evidence that the mitochondrial compartment has a major role in MS [9]. In addition, recent investigations have described an important contribution of autophagic processes. The first evidence that autophagy could be involved in MS was reported in 2009, when a strong correlation was found between the expression of the autophagic marker ATG5 and the clinical disability observed in the experimental autoimmune encephalomyelitis (EAE) MS animal model. Moreover, in this work, the authors found increased expression of ATG5 in T cells obtained from RR-MS patients and in postmortem brain tissue from individuals with secondary progressive MS [79]. Unfortunately, the authors did not address the role of autophagy in T cells and MS. They only speculated that autophagy may help to increase the survival of T cells and help to propagate the immune response. Similarly, other work detected ATG5 increases in terms of both mRNA levels and protein amounts in T cells obtained from MS patients who were treatment naïve [80]. Increases in ATG5 also correlated with the presence of proinflammatory cytokines, thus displaying a possible relationship between the inflammatory status and ATG5 expression in MS. However, they did not perform a detailed analysis of the clinical activity state [80]. T cells present different subpopulations. Among them, T regulatory cells (Treg) are particularly relevant in autoimmune disease because they prevent inflammation and preserve the tolerance to self-antigens. Recently, it has been demonstrated that the autophagic mediator AMBRA1 associates with the protein phosphatase PP2A to sustain Treg differentiation by increasing the expression of Forkhead box P 3 (FOXP3), an essential transcription factor for the differentiation of Treg cells [81]. In addition, the AMBRA1–PP2A–FOXP3 molecular axis was found to be essential for regulating the optimal autophagic levels necessary for T-cell stimulation and differentiation. Consistently, AMBRA1 conditional KO mice display reductions in FOXP3 levels with consequent impairments in Treg differentiation and activity. Most importantly, AMBRA1 deficiency worsens the disease pathogenesis in an EAE MS animal model [81]. Finally, work of Akatsuka et al. not only demonstrates the important role of AMBRA1 in the regulation of T cells, but also highlights decreased mitochondrial functioning and metabolism in these cells [82]. All these findings demonstrate that AM-BRA1 is an essential factor that regulates both autophagic and mitochondrial behaviors and, probably, also the mitophagic process in T cells.

In MS, T-cell activities may be modulated by the complement-regulating molecule CD46 [83]. This factor is also described as an autophagic inducer [84], and its levels are documented to be increased in the serum and cerebrospinal fluid (CSF) of MS patients [85]. The increased T-cell autoreactivity in MS may also be promoted by IRGM1, a GTPase that regulates the survival of immune cells through autophagy. Consistent with this finding, IRGM1 deletion increases the apoptosis of T cells, reduces their proliferative capacity, and ameliorates the clinical score of the EAE mouse model [86]. Considering that subsequent studies have demonstrated that IRGM1 is localized to the mitochondrial compartment and regulates the mitochondrial metabolism and mitochondrial fission induced by mitophagy [87,88], the increased T-cell autoreactivity observed in MS may be due to an impairment in the mitophagic process. In addition to its effects on T cells, autophagy plays a role in dendritic cells (DCs), the most potent antigen-presenting cells (APCs) in the immune system. In particular, autophagy starts in response to bacterial and viral infection. By generating transgenic mice with silencing of ATG7 in DCs, Bhattacharya and colleagues demonstrated the importance of DCs and autophagy in MS. Indeed, they showed that the specific loss of autophagy in DCs significantly delayed disease progression and reduced disease severity in EAE mice [89]. As reported above, AMPK is the main positive regulator of autophagy. This kinase works by sensing the AMP/ATP ratio and activates autophagy

to combat energetic imbalance. It has been demonstrated that following exposure to proinflammatory cytokines, AMPK activates and triggers autophagy in oligodendrocyte precursor cells (OPCs) [90]. This change is due to a metabolic switch from OXPHOS to glycolysis and impairment of mitochondrial dynamics, leading to increased oxidative stress and reduced mitochondrial Ca2+ uptake and ATP production. As a consequence, OPCs fail to differentiate into mature and myelinating oligodendrocytes [90]. In support of these in vitro findings, recent work demonstrated that metabolic stress-induced autophagy is a key element in an in vivo MS model. Indeed, MCU-deficient (MCU-def) mice subjected to EAE displayed elevated clinical scores, excessive inflammation, and demyelination [91]. Morphological and functional analyses performed with the spinal cords of MCU-def mice revealed important mitochondrial damage, accompanied by an elevated presence of autophagosomal markers and a decrease in ATP synthesis and mitochondrial gene expression. Overall, these data confirm that the presence of mitochondrial dysfunction provokes the inhibition of Ca2+ buffering, ATP synthesis, and mitochondrial gene expression, causing a metabolic collapse that prompts autophagy and worsens MS-like conditions. Furthermore, since autophagic activation accompanied by the downregulation of PGC1α (a master regulator of mitochondrial biogenesis) has been observed, it is possible to speculate that the mitochondrial QC system is also affected. However, studies have not verified whether autophagy activities lead to autophagic mitochondrial removal.

Markers of autophagic processes may represent reliable potential biomarkers for monitoring the progression of disease. Increased amounts of Parkin, ATG5, and inflammatory cytokines are present in both the serum and CSF obtained from MS patients. Analyses comparing MS patients to healthy individuals and patients affected by other neurodegenerative conditions have been conducted [16]. Moreover, subsequent work demonstrated that increases in both autophagic and mitophagic markers correlated with the active phases of the disease and with circulating lactate levels, demonstrating the presence of an impaired metabolic status in MS patients [92]. Notably, several studies have associated lactate levels with MS progression [93]. Other independent research groups have confirmed that circulating autophagy and mitophagy markers are increased in MS biofluids [94,95]. In addition, the circulating levels of mitochondrial adenine nucleotide translocase 1 (ANT1) and oxidative stress markers have also been investigated. Interestingly, MS patients display increased oxidative stress, accompanied by reduced levels of the mitochondrial marker ANT1, suggesting that the mitochondrial QC systems are activated to promote the removal of nonfunctioning mitochondria. Consistent with this, reduced circulating levels of the OMM protein translocator protein 18 kDa (TSPO) and increased amounts of the mitochondrial disease marker growth/differentiation factor 15 (GDF-15) have been found in MS individuals and correlate with the severity of the disease [96,97] (Table 1).

It is clear that autophagy and mitophagy as well as the mitochondrial quality control system are important contributors in MS. In the last few years, an increasing number of studies have correlated the activities of such molecular mechanisms with the progression of the disease. Furthermore, circulating elements of autophagy and mitophagy may be detected in human samples from MS individuals, thus suggesting the possibility of using them as novel biomarkers. However, MS shows a great heterogeneity with regard to the clinical symptoms as well as therapy response. In addition, MS manifests in different forms (clinically isolated syndrome, RR MS, secondary progressive MS, and primary progressive MS), where the relapse rate and disability progression differentiate one from the other. Only when the dynamics and response of autophagy and mitophagy are well characterized in regard to all these conditions will we be able to claim to have identified the real contributions of them in MS, and we could use autophagic and mitophagic elements as innovative markers for MS disease progression.


**Table 1.** Summary of autophagy- and mitophagy-related markers in biofluids of MS-, AD-, and PD-affected persons.

#### **5. Involvement of Autophagy Mechanisms in AD Progression**

AD was first described in the early 20th century and is characterized by a progressive deterioration of cognitive function. Memory loss and dementia represent the most common symptoms. The cardinal pathological hallmarks of AD are extracellular (amyloid) plaques and intracellular and extracellular neurofibrillary tangles (NFTs). Amyloid plaques are composed of deposits of Aβ, α-syn, Ub, and apolipoprotein E. NFTs are characterized by hyperphosphorylated tau protein and apolipoprotein E. These aggregates induce neuronal toxicity by impeding neural communication and provoking cell death either directly or by preventing the delivery of an optimal nutrient supply to brain cells [3].

At present, the origin of AD and the mechanisms occurring in the pathogenesis of AD are not well defined. Inflammation seems to play an important role: mediators of inflammation, such as cytokines, adhesion molecules, and prostaglandins, drive degeneration in different neural AD models [98]. Consistent with this finding, aggregated peptides increase proinflammatory agent production, and inflammatory molecules are detected in the CSF, serum, and plaques obtained from AD patients. Oxidative and nitrosylative damage provoked by ROS and reactive nitrogen species (RNS) are determinants of the initiation and progression of AD [99]. Oxidatively damaged membrane phospholipids and increased oxidative stress in neurons are frequently present in neurons exposed to Aβ [100]. Furthermore, AD brains extracted at autopsy have decreased amounts of vitamins A and E and β-carotene [101] and display a higher production of free radicals and increased expression of neuronal nitric oxide synthase (nNOS) [102]. This increased nNOS correlates with an increased apoptosis of hippocampal neurons. In the last 10 years, an increasing number of studies have demonstrated the critical contributions of autophagy and mitophagy to AD pathogenesis [103]. Several studies have reported an increased presence of Aβ in autophagosomes [104]. Interestingly, autophagosomes also contain amyloid precursor protein (APP) and its processing enzymes, in particular, a component of the γ-secretase complex, suggesting an additional source of Aβ. Consistent with this finding, the induction of autophagy correlates with Aβ production, and autophagy-deficient

animals (with ATG7 knockdown) display reduced Aβ secretion [105]. Additionally, the hyperphosphorylation of tau correlates with increased autophagic levels. Indeed, postmortem AD brain samples are characterized by LC3- and p62-positive autophagosomes, and the hyperphosphorylation of tau has been recognized in autophagy-deficient mice [106,107]. Although these observations highlight a dangerous correlation between autophagy and AD, other studies suggest that autophagy and mitophagy may exert beneficial effects against AD [103]. The abnormal accumulation of autophagosome vesicles is present in AD neurons [104]. This accumulation is related to compromised lysosomal function, which results in lysosomes that are no longer able to degrade autophagosomes. The overexpression of Parkin and PINK activates mitophagy, restores mitochondrial function, and reduces Aβ production [108,109]. Similar results have been obtained from another independent experiments that demonstrated that mitophagy is essential for reducing Aβ levels, abolishing tau hyperphosphorylation, preventing cognitive impairments in an AD mouse model, and suppressing neuroinflammation [110].

To confirm the crucial role of autophagic and mitophagic dynamics in AD, different studies have evaluated the presence of elements belonging to these processes in biofluids from persons with AD. The first investigation was performed in 1995, in which ventricular CSF from postmortem AD patients was analyzed. In this study, the authors detected increased levels of the lysosomal protein cathepsin D [111]. However, a subsequent report performed with lumbar CSF samples from living AD patients found no change in the levels of cathepsin B [112]. This finding was confirmed in other work that investigated a broad range of lysosomal proteins in CSF samples from living AD patients and found no variations in diverse cathepsin forms (A, B, D, and L); however, the study did find altered expression for five other lysosomal proteins in the AD samples: early endosomal antigen 1 (EEA1), LAMP1, LAMP2, RAB3, and RAB7 [113]. By contrast, a recent study analyzed the levels of proteins associated with lysosomal function in the CSF of AD persons by conducting solid-phase extraction and parallel reaction monitoring mass spectrometry and found only minor or absent changes in their levels [114]. Unfortunately, the levels of proteins directly related to autophagy and mitophagy processes were not investigated in that study. A follow-up study at 12 and 24 months identified autophagic elements (BECN1, p62, and LC3) in peripheral blood mononuclear cells (PBMCs) obtained from the blood of AD patients and demonstrated that their levels varied during the course of the disease and correlated with the inflammatory environment [115]. Recently, autophagic elements have also been assessed directly in AD blood samples. Indeed, increased levels of the autophagic marker ATG5 are present in the plasma of patients with dementia who meet the criteria for probable AD. Unfortunately, the authors did not identify the subtype of dementia or confirm the AD status. These limitations were overcome in a recent investigation assessing the circulation of autophagic and mitophagic markers in the serum of patients affected by mild–moderate late-onset AD, mild cognitive impairment (MCI), vascular dementia (VAD), and mixed dementia (MD). In this work, the authors found decreased levels of ATG5 and Parkin in patients affected by AD, MCI, and MD. By contrast, they detected increased levels of these markers in VAD patients [17]. This investigation suggests that autophagy and mitophagy markers are possible biomarkers for AD and that they are differentially affected in different dementia types, which may help to discriminate AD-type dementias from VAD. Additionally, the fact that AD samples have decreased levels of autophagy and mitophagy markers confirms the presence of an impaired degradative system in AD persons (Table 1).

Summing up, autophagy and mitophagy represent well-established mechanisms in AD and may exert a protective role. Accordingly, most research highlights the reduced recruitment of both autophagy and mitophagic factors in cell cultures, in vivo AD models, and human samples obtained from AD-affected patients, including in the body fluids of the CSF and blood. Here, autophagic and mitophagic partners also correlate with the inflammatory status and change during the course of the disease, thus opening up the possibility of using autophagic and mitophagic elements as markers for the progression of AD. However, before ascribing merit to these molecules as potential screening, prognostic, diagnostic, or disease-monitoring markers for AD, it is important to consider different aspects. The diagnosis of AD cannot be achieved until the patient displays dementia symptoms. In addition, different dementia types exist and vary between individuals. Very few studies have monitored the variation of circulating markers of autophagy and mitophagy during the different dementia types. Furthermore, these studies lack validation of the investigated markers with accepted methods for diagnosing AD, such as amyloid PET imaging. Again, all the investigations performed did not provide follow-up studies and did not analyze the effects of the disease-modifying drugs commonly used for AD therapy on autophagy and mitophagy circulating markers. Undoubtedly, more detailed analyses and larger cohort studies are necessary to verify whether autophagic and mitophagic circulating elements may represent promising biomarkers for AD.

#### **6. Current Knowledge of the Relationship between PD and Autophagy Dynamics**

Resting tremor, bradykinesia, rigidity, and postural instability represent the four cardinal signs of PD, the most common neurological movement disease, and PD is characterized by a progressive loss of dopaminergic neurons in the substantia nigra pars compacta. PD is considered a multifactorial disease since both genetic and environmental factors play important roles. Approximately 90% of the cases are sporadic, while the remaining 10% are caused by monogenic mutations in at least 23 genes. Similarly, a number of cellular mechanisms are involved in PD pathogenesis. Among them, the uncontrolled intracellular aggregation of α-syn, in the form of LBs and Lewy neurites, represents the main hallmark of the disease. It is not surprising that the first evidence of the genetic mechanisms of PD was a mutation in the α-syn gene, and different forms of the α-syn protein (oligomers, protofibrils, and unfolded monomers) have been found in human PD brain samples. α-syn is a presynaptic neuron protein abundantly expressed in the nervous system; it is present in proximity to synaptic vesicles and folds into α-helical structures. The primary role of this protein is to attenuate neurotransmitter release and synaptic vesicle recycling. In PD, α-syn generates β-sheet structures that are prone to aggregation, which leads to pathologic conditions with toxic gain-of-function effects. Mitochondrial dysfunction is another crucial element during PD pathogenesis in both sporadic PD and familial Parkinsonism. Different postmortem studies have highlighted the existence of deficiencies in components of the mETC, and compounds (toxins and pesticides) were found to promote the Parkinsonian phenotype and neuron loss by impairing complex I of the mETC. Furthermore, α-syn alone induces effects by interacting with the mitochondrial membrane, accumulating inside the organelle, and leading to mitochondrial dysfunction and oxidative stress by damaging C-I [116]; alternatively, α-syn can interact with the mitochondrial transporter TOM20 [117]. Another gene causing autosomal dominant PD is LRRK2, a protein involved in diverse signaling pathways, including vesicular trafficking, protein translation, and the control of mitochondrial dynamics. Mutations in this member of the leucine-rich kinase family have been found in approximately 1–2% of sporadic and 5% of familial PD cases. The most frequent mutation of LRRK2 (G2019S) induces an increase in LRRK2 activity. G2019S-LRRK2 PD postmortem human tissues, animal models, and cellular models are characterized by important mitochondrial dysfunction, with impaired ATP production, mitochondrial fragmentation, mtDNA damage, and oxidative stress representing the main features. Recently, it has been demonstrated that this mutation induces impairment in mitophagic clearance [118]. In addition, the loss of function, mutation, and overexpression of the mitophagic regulatory members PINK1 and Parkin provoke impaired mitochondrial turnover and cause autosomal recessive PD. To mediate mitophagy, PINK1 and Parkin cooperate to recognize and label damaged mitochondria with polyubiquitin (p-Ub) chains [119]. Postmortem brains from LB disease patients are characterized by p-Ub chain structures that colocalize with markers of mitochondria and autophagy [120]. Mutant forms of PINK1 are unable to move to mitochondria upon stress signaling, thereby avoiding mitophagic induction. Similarly, Parkin mutations (such as S65N, G12R, and R33Q) decrease the capacity of PINK1 to phosphorylate and activate Parkin itself. Correct mitochondrial turnover is

not guaranteed by only the mitophagic process; it is also regulated by fission and fusion events. Among the mitochondrial fusion proteins, the best characterized are mitofusins (MFNs). It has been demonstrated that MFN-1 and MFN-2 are substrates for PINK1 and Parkin and that they can be ubiquitinated by both PINK1 and Parkin [121,122]. Consistent with this, mutations in and the loss of PINK1 and Parkin impair MFN-1/2 ubiquitination in PD patient cells [123]. MFNs also work as bridges between mitochondria and the ER to preserve the appropriate functioning of MAMs. These contact sites between the ER and mitochondria act as primary signaling hubs for cells and regulate lipid homeostasis, calcium dynamics, apoptosis, the stress response, and autophagosome vesicle formation. Diverse recent studies have highlighted a contribution of these contact subdomains in the progression of PD [124]. Several proteins encoded by genes involved in PD (α-syn, Parkin, and PINK1) are located in MAMs and have been found to regulate correct ER–mitochondrion tethering. For example, Parkin deletion increases MFN2 amounts and increases Ca2+ transfer from the ER to mitochondria [125]. Notably, this event is a crucial mediator of the regulation of autophagy [126]. In addition, DJ-1, which provokes a rare form of autosomal recessive PD [127], increases ER–mitochondrion communication and preserves the optimal Ca2+ transfer between the ER and mitochondria, thereby having a cytoprotective role that is essential for maintaining mitochondrial functioning [128] and, probably, autophagy levels. As reported above, alterations in tau protein are highly related to AD disease. In truth, pathogenic mutations in tau protein are present in different neurodegenerative disorders, defined as tauopathies. Interestingly, different studies suggest that tau mutations are also present in PD. By comparing characterized tau mutations related to tau toxicity and aggregation in PD (P301L and A152T) [129–131], a recent investigation explored, for the first time, the concomitant activity of the three different forms of autophagy (autophagy, CMA, and mA) [132]. Here, the authors found different activity levels for the three autophagic forms and demonstrated that the pathogenic tau mutation A152T resulted in a blockage of both CMA and mA, but caused a compensatory activation of autophagy. Oppositely, the P301L mutation provoked an inhibition of the degradation of tau aggregates by any of the three catabolic pathways [132]. A deeper understanding of the different recruitment of the diverse autophagic forms may help to increase our knowledge of the molecular mechanisms existing in PD as well as in the other tauopathies.

Detecting autophagy- and mitophagy-related proteins in peripheral human biospecimens may represent a promising method for identifying PD statuses and controlling the progression of the disease. Significant decreases in LC3B, BECN1, ATG5, and lysosomal associated membrane protein (LAMP) 2 are present in CSF samples obtained from early-stage PD patients. Interestingly, among these autophagic partners, only LC3B shows a significant correlation with α-syn, total tau levels, and the clinical severity of patients [133]. Notably, recent studies have demonstrated that tau protein also participates in the pathology of PD [134]. A reduction in LAMP2 levels in the CSF of PD patients was found in another study. In this study, the authors found a decrease in LAMP1 levels, but they did not identify variations in LC3 levels [135]. An important difference exists between these two studies. In the first, circulating proteins were detected by using ELISA technology, while the second employed an immunoblotting technique. LAMP2 levels were also analyzed in a study comparing PD patients, PD patients harboring LRRK2 mutations, and healthy control subjects with or without LRRK2 mutations. The main finding of that investigation was that LAMP2 protein levels were reduced in the PD patients harboring LRRK2 mutations. Similar to the study mentioned above, LAMP2 levels were not related to the clinical states of the patients. However, a positive correlation between LAMP2 and oxidative stress has been shown [136]. Autophagic markers and related proteins were analyzed in circulating PBMCs obtained from PD patients. The results demonstrated that the steady-state autophagy in PD patients was profoundly different from that observed in healthy individuals and correlated with augmented expression of α-syn [137]. Unfortunately, investigations aimed at detecting mitophagic elements in human biofluids are lacking. The existing literature can inform us about the "mitochondrial signature" that exists in PD patients. Significant decreases

in mtDNA copy number have been observed in patient blood cells and in CSF samples from early-stage PD [138,139]. Increased mtDNA levels were present in the CSF of PD patients carrying LRRK2 mutations [140] and in the sera of PD persons with mutations in Parkin or PINK1. This study demonstrated an association between mitophagy impairment (represented by a Parkin- or PINK1-mutant genotype) and mtDNA in circulating biofluids from PD patients for the first time. Finally, an interesting study found reduced methylated mtDNA in PD patients, suggesting that affected patients may have disrupted mtDNA gene expression and replication [141] (Table 1).

Undoubtedly, the discovery that many genes involved in both autophagy and (in particular) mitophagy are mutated in familial Parkinsonism and in sporadic PD makes these molecular processes fundamental for this neurodegenerative disease. The fact that markers of these catabolic systems may be detected in human samples also opens up the possibility of creating new real-time monitoring approaches for the progression of the disease. However, today, these studies' results are still incomplete, and some of them report controversial results, probably due to the limited sizes of the cohorts analyzed, the different types of samples analyzed, and, most importantly, not always accounting for the clinical history of each single patient. It is thus clear that more detailed and larger longitudinal, stratified, and standardized analyses are needed.

#### **7. Principles and Current Strategies for Targeting Autophagy in Neurodegeneration**

Several therapeutic approaches ameliorate the consequences and symptoms of neurodegenerative disorders. Unfortunately, current treatments become less effective as the neurodegenerative status advances, and most importantly, none of them prevents the onset or progression of the disease. The evidence reported in the previous sections demonstrates that autophagy processes have an important role during the development of the most common neurodegenerative diseases. Hence, these findings suggest that the modulation of autophagic and mitophagic processes may be a possible innovative therapeutic approach for combating neurodegeneration (Figure 3).

**Figure 3.** Strategies for targeting autophagy in neurodegeneration.

Autophagy upregulation has been demonstrated to be an effective strategy for increasing the clearance of neurodegenerative disease-causing proteins in different cellular and mouse models, thereby reducing toxicity. Autophagic activation may be induced by blocking mTOR activities. Rapamycin is the best known mTOR inhibitor, and it has been demonstrated that this compound increases the clearance of tau protein and decreases tau toxicity [142,143]. In addition, rapamycin activates autophagy to remove other protein aggregates, such as long polyglutamines and polyalanine-expanded proteins [143]. Unfortunately, one limit of rapamycin is its limited absorption. Different analogs of rapamycin (rapalogs) have been developed in recent years. Among them, temsirolimus has been found to increase the autophagy clearance of hyperphosphorylated tau and ameliorate learning and memory impairments [144]. Another compound mediating mTOR inhibition is the proneurogenic and antihistaminic compound latrepirdine. It has been demonstrated that latrepirdine also improves learning behaviors and reduces Aβ and α-syn aggregates in an AD mouse model [145]. Autophagy can also be activated by mTOR-independent pharmacological agents. For example, resveratrol exerts neuroprotective effects in PD models by increasing autophagy through an AMPK-dependent mechanism [146]. The same effects on AMPK activity are induced by the recently identified small molecules A769662, GSK621, RSVA314, and RSVA405. A769662 and GSK621 promote the autophagic clearance of α-syn aggregates by inducing the phosphorylation of AMPK and ULK1 [147]. RSVA314 and RSVA405 activate AMPK by a CaMKKβ-dependent mechanism to activate autophagy and promote the degradation of Aβ 20852062. The molecule AUTEN-67, which antagonizes the autophagic inhibitor phosphatase MTMR14, increases autophagic flux, promotes neuron longevity, and prevents neurodegenerative symptoms in AD models [148]. Interestingly, the same research group also demonstrated that another molecule, AUTEN-99, is capable of improving autophagy and exerting neuroprotective effects in PD models [149]. Lithium is widely used to treat bipolar disorders and depression. In addition, recent investigations have demonstrated that lithium administration activates autophagy by inhibiting inositol monophosphatase in an mTOR-independent manner [150] and exerts neuroprotection in AD [151]. Consistent with this, a clinical trial evaluating long-term treatment with lithium in AD patients revealed the amelioration of multiple cognitive parameters in the lithium group. Furthermore, analyses conducted with CSF samples revealed a significant reduction in phosphorylated tau [152]. It is clear that an improvement in autophagic machinery should permit an increase in the clearance of protein aggregates. Despite this understanding, studies have also demonstrated that pharmacological interventions aimed at blocking the autophagic process may be useful for counteracting a neurodegenerative status. In an α-syn transgenic mouse model, the overexpression of α-syn reduced dendritic and synaptic markers, which were reduced after exposure to the anti-autophagic compounds bafilomycin-A1 (Baf-A1) and chloroquine (CQ). Interestingly, the authors also found a reduction in α-syn inclusions after treatment with rapamycin. This finding suggests that the aggregation of α-syn is not exclusively mediated by the mTOR-dependent regulation of autophagy. Considering that mTOR is involved in multiple cellular processes, it is possible that other mechanisms are involved in α-syn metabolism [153]. Of note, it is important to specify that both CQ and Baf-A1 have a broad spectrum of biological activities. Similar to what was observed for α-syn aggregates, the inhibition of autophagy promoted by CQ induced a reduction in total tau levels in rat hippocampal extracts. Increase and decreases in tau levels have been observed after rapamycin administration [154]. The inhibition of autophagy also seems to exert beneficial effects in MS. Indeed, by suppressing the inflammatory process in EAE, CQ administration ameliorates the clinical signs of the disease [155]. A subsequent study unveiled that the effect of CQ in reducing the severity of the clinical course of EAE is mediated by a direct effect on DCs. In this work, the authors demonstrated that CQ-treated cells displayed reduced expression of molecules involved in antigen presentation, which resulted in reduced T-cell activation and proliferation [156]. The deleterious role of autophagy in MS has also been demonstrated in the cuprizone (CPZ) demyelination model, which permits us to determine the contribution of other

elements independent of the immune system during demyelination/remyelination. The administration of CPZ with rapamycin resulted in increased demyelination compared with treatment with CPZ alone [157]. Furthermore, rapamycin increases axonal damage and leukocyte infiltration when administered together with CPZ [158], suggesting that by administrating agents blocking the autophagic process, it may be possible to reduce demyelination. Despite this finding, other studies have demonstrated that rapamycin ameliorates histological and clinical signs in MS models, particularly EAE models [159]. In addition, a clinical trial in which RR-MS patients received rapamycin for 6 months highlighted some degree of reduction in the volumes of sclerotic plaques, accompanied by a significant decrease in T-responder cells [160]. To the best of our knowledge, there is no specific agent that can modulate the selective autophagic removal of mitochondria during neurodegeneration. Several efforts are ongoing to try to overcome this shortcoming. For example, mitophagy may be improved by small molecules activating the PINK1–Parkin pathway. The ATP analog kinetin triphosphate has been identified as a potent PINK1 activator. Indeed, this neosubstrate accelerates PINK1-dependent Parkin recruitment to damaged mitochondria and prevents the apoptosis induced by oxidative stress in neuronal cells [161]. Unfortunately, long-term oral kinetin administration does not prevent the neurodegeneration induced by α-syn in a PD model [162]. A recent high-throughput screening identified two other small molecules (T0466 and T0467) that affect the PINK1–Parkin axis. These compounds successfully promote Parkin translocation to mitochondria, suppress mitochondrial aggregation in dopaminergic neurons, and improve locomotor defects in the *Drosophila* PINK1 model [163]. Finally, it has been suggested that Rho-associated protein kinase (ROCK) inhibitors may exert a neuroprotective effect by increasing the activity of the Parkin-mediated mitophagy pathway. In one investigation, the authors performed a screen of ~3000 compounds with the aim of identifying compounds that promote Parkin translocation to mitochondria. As a result, they found that several ROCK inhibitors increased the recruitment of Parkin to damaged mitochondria and found compound SR3677 to be the most efficacious. In addition, SR3677 also exerted neuroprotective effects and restored locomotor abilities in a *Drosophila* PD model [164].

The activation of autophagy seems to be efficacious in increasing the clearance of protein aggregates. Autophagic activation may be obtained by using rapamycinRAPAMYCIN and its analogs, such as TEMSIROLIMUS. These compounds permit inhibiting the mammalian target of rapamycin (mTOR). A similar effect was obtained by using the proneurogenic and antihistaminic compound LATREPIRIDINE. Autophagy may also be activated by potentiating the activity of 5 adenosine monophosphate-activated protein kinase (AMPK) with RESVERATROL and a series of small molecules (A769662, GSK621, RSVA314, and RSVA405) recently identified. Autophagy activation may also be obtained in an mTOR/AMPK-independent manner. LITHIUM was found to block the pro-autophagic function of inositol monophosphatase (IMPase), and two molecules (AUTEN-69 and -99) antagonize the autophagic inhibitor phosphatase MTMR14. Oppositely, autophagy inhibition represents a possible therapeutic approach against multiple sclerosis (MS). In particular, diverse studies suggest that the anti-autophagic compound chloroquine (CQ) suppresses inflammation and reduces the clinical score of an MS mouse model.

#### **8. Conclusions and Future Perspectives**

In this review, we summarize studies showing how autophagy and mitophagy have crucial roles in neurodegenerative disorders. It is clear that compromised autophagy and mitophagy dynamics mediate the pathogenesis and progression of diseases characterized by the uncontrolled accumulation of protein aggregates, such as AD and PD. These mechanisms are in contrast with those in MS, in which the neurodegenerative condition is not due to aberrant protein inclusions and autophagy and mitophagy appear to be excessively activated and to provoke deleterious conditions causing cell death or the impairment of normal cellular functions.

Interestingly, these scenarios are present when autophagy markers or any biochemical or molecular markers are analyzed in body fluids from persons affected by specific neurodegenerative conditions. Indeed, most of the investigations performed highlight a reduced activity of the autophagy processes in AD and PD; however, these processes are increased in human MS samples. Furthermore, autophagy levels correlate with clinical outcomes. Overall, these findings clearly show that changes in autophagy and mitophagy elements may be reliable markers for predicting/controlling disease progression and helping to monitor clinical status. It is, thus, necessary to point out that most of the investigations performed have utilized serum. This biofluid is more accessible and safer and has fewer contraindications than CSF, and it would certainly facilitate the continuous tracking of disease progression. Another point should be highlighted: the findings presented herein also suggest that the modulation of autophagy dynamics with pharmacological strategies may be an effective method for slowing down the progression of some neurodegenerative diseases. Different studies performed in cellular and animal models support this possibility. For instance, the inhibition of autophagy appears to reduce inflammation and clinical signs in MS models, while its activation reduces protein aggregation and consequent motor and learning behavior defects in AD and PD models. Consistent with these findings, long-term treatment with the autophagy inducer lithium in a clinical trial restored multiple cognitive parameters in AD patients.

However, it should be noted that several caveats exist in the studies mentioned above. Currently, there are no good methods for accurately quantifying autophagy and mitophagy levels in samples obtained from human patients. For this reason, it is difficult to understand whether increases or decreases in autophagy levels are not due to the impairment of the correct autophagic flux. Experiments with inhibitors of the autophagosomal fusion with the lysosomal membrane may help to overcome this limitation and may be performed in human tissues, such as skin and muscle biopsies and postmortem brain tissues. However, their application in more accessible human samples, such as sera and CSF, seems very hard to achieve. Similarly, today, it remains difficult to perform real-time monitoring of all the autophagic dynamics, from autophagosomal vesicle formation to degradation, in a patient.

Additionally, it is difficult to ensure that pharmacological agents modulating autophagy in cultured cells and animal models also do so in vivo in patients affected by a neurodegenerative condition. Furthermore, drugs activating/inhibiting autophagy may exert several other effects and modulate gene expression and protein, lipid, and nucleotide synthesis. In addition, to date, no effective treatment aimed at selectively modulating only the mitophagic pathway has been tested. Finally, at first glance, serum may represent a valid alternative to CSF for monitoring clinical status. However, outside of the CSF, the concentrations of markers related to the CNS are often low, and circulating antibodies and proteases may alter the effective concentrations of proteins in peripheral tissues [165].

In summary, more work is required to fully clarify all the connections that exist between autophagy processes and neurodegenerative status. Despite this limitation, the current literature demonstrates that the recent decades have been characterized by a great improvement in the understanding of the connection between autophagic dynamics and neurodegeneration. Currently, autophagy is accepted worldwide to be a fundamental aspect of the onset and progression of almost all neurodegenerative diseases. A better understanding of the molecular partners participating in this cellular process may help to identify susceptible patients, control disease progression, and perform active monitoring of treatment responses. Finally, the pharmacological modulation of autophagy may represent an attractive tool for identifying new disease-modifying therapies to combat different types of neurodegenerative diseases.

**Author Contributions:** S.P., C.G. and P.P. conceived the article; S.P., A.D., M.P. and E.B. wrote the first version of the manuscript with constructive input from C.G. and P.P.; E.B. prepared the display items ("Created with BioRender.com") under the supervision of S.P. The figures are original and have not been published before. S.P., P.P. and C.G. reviewed and edited the manuscript before submission. All authors have read and agreed to the published version of the manuscript.

**Funding:** P.P. is grateful to Camilla degli Scrovegni for continuous support. The Signal Transduction Laboratory is supported by the Italian Association for Cancer Research: Grant IG-23670 (to P.P.) and Grant IG-19803 (to C.G.); A-ROSE; the Telethon Grant GGP11139B (to P.P.); Progetti di Rilevante Interesse Nazionale Grants: PRIN2017E5L5P3 (to P.P.) and PRIN20177E9EPY (to C.G.); an Italian Ministry of Health Grant GR-2013-02356747 (to C.G.); a European Research Council Grant 853057- InflaPML (to C.G.); local funds from the University of Ferrara (to P.P. and C.G.); and Fondazione Umberto Veronesi (to S.P.).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Review* **News about the Role of Fluid and Imaging Biomarkers in Neurodegenerative Diseases**

**Jacopo Meldolesi**

Division of Neuroscience, San Raffaele Institute and Vita-Salute San Raffaele University, via Olgettina 58, 20132 Milan, Italy; meldolesi.jacopo@hsr.it

**Abstract:** Biomarkers are molecules that are variable in their origin, nature, and mechanism of action; they are of great relevance in biology and also in medicine because of their specific connection with a single or several diseases. Biomarkers are of two types, which in some cases are operative with each other. Fluid biomarkers, started around 2000, are generated in fluid from specific proteins/peptides and miRNAs accumulated within two extracellular fluids, either the central spinal fluid or blood plasma. The switch of these proteins/peptides and miRNAs, from free to segregated within extracellular vesicles, has induced certain advantages including higher levels within fluids and lower operative expenses. Imaging biomarkers, started around 2004, are identified in vivo upon their binding by radiolabeled molecules subsequently revealed in the brain by positron emission tomography and/or other imaging techniques. A positive point for the latter approach is the quantitation of results, but expenses are much higher. At present, both types of biomarker are being extensively employed to study Alzheimer's and other neurodegenerative diseases, investigated from the presymptomatic to mature stages. In conclusion, biomarkers have revolutionized scientific and medical research and practice. Diagnosis, which is often inadequate when based on medical criteria only, has been recently improved by the multiplicity and specificity of biomarkers. Analogous results have been obtained for prognosis. In contrast, improvement of therapy has been limited or fully absent, especially for Alzheimer's in which progress has been inadequate. An urgent need at hand is therefore the progress of a new drug trial design together with patient management in clinical practice.

**Keywords:** neurons; astrocytes; Alzheimer's and Parkinson's diseases; fluid and imaging biomarkers; amyloid-β and tau; miRNA; extracellular vesicles; exosomes and ectosomes; PET; radiotracers; radiolabeled molecules

#### **1. Introduction**

Biomarkers are molecules that are highly variable in their origin, nature, and mechanism of action, and they are connected to or are directly involved to single or various peculiar diseases. At present, biomarkers, which are addressed to cells, organs, or structures, exist for almost all diseases, including cancers. Among them, neurodegenerative diseases are receiving the greatest attention. During the last 20 years, articles about their biomarkers published in known journals have totaled over 20,000, including about 4000 reviews. Investigations of their properties, going from specificity to the mechanisms of their action, are often used to clarify various aspects of pathogenesis. Such results often play roles in processes of medical relevance, such as diagnosis, prognosis, and also therapy, and they are useful for patients and also for clinical practice. However, the relevance of biomarkers in clinical practice is variable. Some of them are well known and widely used; for others, however, knowledge is still questioned due to, for example, their limited specificity. In this review such limitations are not further illustrated. The properties of biomarkers presented here are those of general significance, identified in the last few years. Information about additional aspects can be found in other publications [1,2].

**Citation:** Meldolesi, J. News about the Role of Fluid and Imaging Biomarkers in Neurodegenerative Diseases. *Biomedicines* **2021**, *9*, 252. https://doi.org/10.3390/ biomedicines9030252

Academic Editor: Arnab Ghosh

Received: 28 January 2021 Accepted: 26 February 2021 Published: 4 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Compared to biomarkers of other organs, biomarkers of the brain exhibit distinct properties [3,4]. Initial studies about two types, the fluid and the imaging biomarkers, were developed separately. Biomarkers of the first type appeared around 2000. They are collected not in vivo but in fluid, within either of two fluids taken from patients, the central spinal fluid (CSF) and the blood plasma [5,6]. Biomarkers of the second type were developed by the use of radiolabeled molecules. Upon penetration into the living brain, these molecules are bound with high specificity as revealed by positron emission tomography (PET) imaging. A biomarker study by the latter approach, addressed to amyloid-β (Aβ) plaques of Alzheimer's disease (AD), was published in 2004 [7]. Since then, the in vivo studies of the imaging type have continued with growing success (see Section 5). Since the beginning of the two types of biomarker studies, the state of neurodegenerative diseases changed profoundly. In particular, biomarkers did recently revolutionize scientific and medical research by transforming drug trial design and also improving patient management in clinical practice [8].

So far, I have introduced the two types, i.e., fluid and imaging biomarkers. The in vivo imaging of the latter takes place upon their high affinity binding by radiolabeled specific molecules introduced in the brain. In contrast, fluid biomarkers are collected within (CSF) or away (blood) from the central nervous system. Yet, all fluid biomarkers are adequate to identify central molecules/processes, critical for patients suffering neurodegenerative diseases, not only AD but also Parkinson's disease (PD), amyotrophic lateral sclerosis (ALS), and other diseases. Initially, the disease identification was searched by the recovery in the fluids of free specific molecules. These molecules, however, are largely digested during their traffic, from the cells affected by the disease to the accumulation in the fluids. Thus, their identification was often difficult. The problem has been solved recently by changing the study from free molecules to cargo molecules segregated within extracellular vesicles (EVs). This change will be presented in detail in Section 3.

Initially, each neurodegenerative disease was considered dependent on a single specific biomarker, such as Aβ for AD and α-synuclein for PD. Now it is clear that these two, as well as many other biomarkers, are not fully specific but expressed also by patients of other neurodegenerative diseases [9]. Multispecificity of biomarkers requires caution in their operation. Caution is necessary also with another type of problem. Initially considered specific and efficient, some biomarkers have been found to be hardly reproducible, thus inappropriate for research and clinical practice [8]. A final important consideration refers to the age dependence of disease investigation/treatment. Specific biomarkers have been applied not only to mature patients, but also to patients at early stages of disease. By novel and innovated methods, it could be established whether patients at risk of long-term diseases, such as AD or dementia, can be treated only upon full development of their symptoms, or also when symptoms are absent or still at an early stage [10,11].

Summing up, I anticipate the two types of biomarker generation, and the anticipation of their properties confirms their relevance, which is demonstrated for many of them. The others, in contrast, are weak, characterized by poor specificity and limited employment. The review here includes the most important properties of the two types of biomarkers that have been identified during the last years. Following this introduction, Section 2 deals with the fluids, the environment of type one biomarker generation, followed by Section 3 about the role of EVs, and Section 4 about fluid biomarkers. Finally, Section 5 deals with the second type of biomarkers, i.e., the imaging biomarkers.

#### **2. Fluids**

In the Introduction, I already mentioned the two ways leading to the generation of neural biomarkers, i.e., the first based on the analysis of peripheral fluids containing appropriate molecules, the second in vivo, based on the imaging by PET labeling of specific brain molecules. How is it that fluids are essential for the first, very important approach? Two of their properties need to be emphasized. First, molecules of interest for their recognition as biomarkers need to be present in the biological fluids of human body;

second, the analysis of fluids, with ensuing isolation of molecules, is less expensive, much easier, and does not disrupt the body compared to the in vivo analysis of tissues. In other words, the fluids are advantageous in the development of biomarkers.

In the body of mammal animals, including humans, there are 8 types of external fluids. For our purposes, however, only two are relevant: the CSF and the blood plasma. Molecules and the small organelles EVs, released from brain cells such as neurons and astrocytes, navigate in the extracellular fluid space from which they are easily transferred to the CSF [3] (step 2 in Figure 1). At the arachnoid villi, the EVs of the CSF can move to venous blood within large vacuoles (step 4 in Figure 1). In addition, the molecules and EVs have been shown to traffic through the blood–brain barrier [12,13] (step 3 in Figure 1). The latter is the structure known to reduce/exclude the traffic of many other molecules and organelles to and from the brain.

**Figure 1.** Traffic of extracellular vesicles (EVs) released by a neural cell. The neural cell at the top (cytoplasm of grain color, marked 1) releases the two types of EVs, the small exosomes by exocytosis of multivesicular bodies (MVB), and the larger ectosomes by shedding of plasma membrane rafts. Upon release, the vesicles navigate in the extracellular fluid (light blue). Their targeting relevant for fluid biomarker generation can be accumulate either to the central spinal fluid (CSF) of the ventricular system (left, peach color, marked 2) or as shown by the arrows to the blood (right, red-pink color, marked 3). Additional CSF-to-blood EV transfer occurs by large vacuoles operative at the arachnoid villi (bottom, marked 4). Thus, molecules and EVs can move from the extracellular space and the CSF to the blood plasma. In the case of neurodegenerative diseases, the molecules and EVs of such origin account for significant fractions of the total transferred to the fluids.

> Recent studies have shown that biomarkers of the two fluids, CSF and blood plasma, are employed approximately with the same frequency. The choice of one fluid, however, is not due to a negative evaluation of the other. In recent reviews, both are reported as valid [12,14]. Rather, the choice appears to depend on advantages existing in either fluid. High levels of molecules and EVs are present in CSF (Figure 1), which is more invasive

and requires more expenses than the blood for biomarker generation. Yet, the CSF fluid has made it possible to have processes of low concentrated proteins, for example, those of the COVID-19 disease reported this year in patients exhibiting neurological symptoms, evident of cerebral infection [15]. In the blood, plasma withdrawal can be large, frequently repeated, and not expensive. However the levels of molecules/EVs in this fluid are lower than in the CSF. The transfer of various EVs across the blood–brain barrier (BBB) has been shown to be different, with ensuing variability of the plasma levels [16]. In many cases, however, they have emerged in the study of neurodegenerative diseases, with enormous potential as a diagnostic, evaluation of therapeutics, and treatment tool.

Summing up, both fluids are employed for biomarker generation. However, the choices depend on their distinct properties. The processes of biomarker generation, previously defined of liquid biopsy [13], have been shown to identify not only the fluid molecules but also the brain pathologic EVs inaccessible in vivo. In addition, fluids have offered unique opportunities, as seen from recent clinical trials, to improve the quality and applicability of results. The processes involved will be developed by approaches presented in the subsequent Section 4 of this review.

#### **3. From Molecules to Extracellular Vesicles**

Neurodegenerative diseases are heterogeneous disorders characterized by a progressive and severe cognitive and functional decline leading to the progressive loss of function and death of neuronal cells. Due to the complexity of their diagnosis, early detection and treatments are difficult to recognize, and this is critical for the development of successful therapies. In addition, current diagnostic approaches are often poorly effective. Thus, their therapeutic effects are limited [16]. For several years, progress in diagnosis and therapy has been searched by molecules, such as proteins and nucleotides, via the generation of biomarkers, specific in the diagnosis and possibly in the therapy of the diseases investigated. However, during their traffic and maintenance within the fluids, the molecules were extensively digested by proteases and ribonucleases. Thus, the attempts based on free proteins and nucleotides remained with limited success. The attempt changed considerably when the search for biomarkers started to be made by the use of EVs containing the molecules of interest within their lumen (Figure 1). Now it is clear that when segregated within EVs, the molecules are protected, making them suitable candidates for noninvasive biomarkers [16–18].

EVs are small vesicles of two types released by all cells. The first, widely called exosomes, are produced and accumulated within a vacuole of endosomal nature, the multivesicular body (MVB); the second, the ectosomes, are released by shedding of rafts, directly from the plasma membrane (Figure 1). Interestingly, the two EV types share similar properties. Their membranes are different from the two membranes of origin; their cargoes, segregated within their lumen, contain peculiar molecules, namely many proteins, many nucleotides, most often the noncoding miRNAs, lipids, and others. EVs of neuronal and astrocytic origin, when harvested from blood, can be used to interrogate brain pathologic processes. Moreover, in the case of proteins involved in the development of neurodegeneration, their segregation within EVs and even more their transfer to the fluids decrease the levels within the cells and can thus alter the disease progression [17].

Isolation from CSF or blood plasma of specific brain-cell-derived EVs, confirmed by neural markers [12], is theoretically simple. Thus, the approach has been successful recently and offers great promises for the near future [17,18]. The EVs, released by neurons and astrocytes of patients affected by neurodegenerative disease, can be identified by specific biomarkers found in their cargoes, such as Aβ and phosphorylated tau in AD, α-synuclein in PD, and the transactive response DNA/RNA binding protein of 43 kDa (TDP-43) in ALS [13,19–22] (Table 1). Other biomarkers investigate the subclinical declines of the diseases, for example, the decline of middle age in AD. This can be done with tau and insulin signaling biomarkers [23,24]. A problem considered is that of multiple dementias, only some are connected to neurodegenerative diseases such as AD and PD. Among the proteins of EVs from specific neurodegenerative diseases, some have been recently shown to concern synaptic and axon injury, inflammations, stress responses and other defects [14,25–28]. Other dementias that are independent of neurodegenerative diseases still need identification of their specific biomarkers [29].


**Table 1.** Gene expression and disease specificity of proteins/biomarkers.

Well-known fluid biomarkers related to the mentioned proteins are marked by an asterisk (\*). The dependence of several diseases may explain the low specificity of some fluid biomarkers. Abbreviations not used elsewhere in this Review: C9orf72, chromosome9 open reading frame 72 gene; DLB, dementia with Lewy's bodies; FTD, frontotemporal dementia; FUS, fused in sarcoma gene; HD, Huntington disease; MAPT, gene of the microtubuleassociated protein tau; SNCA, synuclein alpha gene; TARDPB, TAR DNA binding protein gene; TREM2, triggering receptor expressed on myeloid cells 2.

In addition to many proteins, EVs include in their cargo various types of nucleotides, with predominance of miRNAs. The latter are members of a noncoding family involved in various functions, the best known being translational gene expression. Within fluids, miRNAs investigated have been numerous. Whether or not they play the role of biomarkers has been discussed. Positive evidence, analogous to that of proteins, has been recently reported but without precise identification of many miRNAs involved [30–33]. Interestingly, biomarkers of miRNA origin were shown active in the ALS disease [31]. Another positive conclusion has been found, but it concerns children, healthy and patients of their diseases [33]. Additional studies have focused on a new form of noncoding RNA, namely circular RNA (circRNA), known to traffic within EVs. At present, however, the function of these RNAs is unknown. CircRNAs might be involved in age-related diseases [34,35]. RNAs of this type may also operate in neuropsychiatric disorders [35]. The next section, focused exclusively on AD, intends to reconsider in more detail the origin, the multiplicity, and the role of its biomarkers, with special attention on their heterogeneity with perspectives of innovative therapies.

#### **4. AD and Its Multiple Fluid Biomarkers**

AD is the neurodegenerative disease of greatest importance for two reasons, namely its much larger number of patients (at least 50 million worldwide) and the many scientists committed to its investigation. The choice of this disease has been made to illustrate its general properties. The task is to provide a current landscape, largely common to PD [21,29,36] but not always to other neurodegenerative diseases (see, for example, [9,17,28,31,32]).

A property common to many neurodegenerative diseases is heterogeneity. Various factors are considered as possible con-causes of its starting. Among these are oxidative stress, neural network dysfunctions, and defects in protein regulation and degradation. In the case of AD, two additional factors need to be considered, i.e., inflammation and immune dysregulation [37]. Such defects are expected to induce alterations in neurons, synapses, axons, and possibly also on glial cells. As already assumed in Section 3, some of these factors have been recognized by the identification of the corresponding biomarkers. For other defects, however, biomarkers are not available yet. Although incomplete, the present knowledge appears of interest to establish the properties of some AD heterogeneous forms, such as responses to treatments. The study of multiple biomarkers, identified during the last months, might be sufficient to characterize at clinical level various aspects of AD, including diagnosis and prognosis [38–40]. Heterogeneity exists also for tau. The fraction inducing tau biomarkers in the CSF was found to stage Alzheimer's disease, and this is potentially useful for tau-directed therapeutics [41]. Other proteins, such as mesenchymal stem cells and exosomes, appear of considerable potential for therapy. The role of some other properties, including heterogeneity, remains to be established [42,43].

An additional approach relevant for AD diagnosis and therapy has been recently reported based on two different miRNAs that are active as EV biomarkers in the blood plasma. Although not identical, the evaluation of the two miRNAs could be considered in parallel. Effects in AD patients, induced by increased miRNA levels in the blood, were revealed by neuronal viability and neuroinflammation followed by a mini-mental state examination. Additional effects were investigated also in fluid, using well-known neural cells such as SH-SY5Y cells, affected by Aβ treatment and evaluated in terms of proliferation, apoptosis, and neuroinflammation. Interestingly, significant upregulation of the first miRNA, miR-485-3p, was shown in patients and cell models, accompanied by severity of DA in vivo and in fluid [44]. The response induced by the second miRNA, miR-331-3p, appears the opposite. Both in vivo and in fluid, the increased miRNA induced significant and persistent attenuation of AD [45]. The two types of results induced by these miRNA opened the possibility of new, promising therapies based on the two biomarkers investigated [44,45].

In conclusion, the actions by EV biomarkers based on multiple properties of proteins and miRNAs have recently offered ample chances of both nature and specificity, showing enormous potential as diagnostic, evaluation and treatment tools. Their new results allow researchers to test hypotheses by proof of concept studies at the preclinical phase, with further opportunities to develop therapeutic discoveries in neurodegenerative diseases [46].

#### **5. Imaging Biomarkers from AD and PD**

As already reported in Section 1, the identification of Aβ as the key factor of AD, which was already demonstrated in fluid, was confirmed in vivo by the development and investigation of an appropriate radiolabeled molecule. Upon its penetration into the living brain, such a molecule made it possible to have a clear Aβ imaging by PET [7]. By now, such an approach of investigation plays roles much more important than its historical role. Advanced technologies of the near future are expected to further improve the present success of such studies. The biomarkers obtained by this approach are usually named imaging biomarkers.

The advantages offered by such biomarkers include properties that are different and integrative with respect to the fluid biomarkers illustrated in the previous sections. Among the results of imaging biomarkers are their demonstrations occurring in the living brain, distinct from the peripheral demonstrations of fluid biomarkers; their data obtained from larger numbers (hundreds) of patients; and the quantitation of their measurements, necessary for many assays. The relevance of these properties emerges from a recent group of papers based on the comparison between the results obtained with fluid and imaging biomarkers. In these reports, the relevance of the results obtained by the imaging techniques is emphasized [36,47,48]. Another difference between the two approaches has

emerged with respect to the amyloid cascade, a concept activated by elevated levels of Aβ and tau, which is completed by severe cognition and functional impairments [49]. In this case, the data by CSF biomarkers were found to emerge before those by PET imaging, demonstrating a high sensitivity that is relevant for the study of early AD stages [50]. Further studies with CSF biomarkers, compared with those obtained by PET and resonance (MRI) imaging, confirmed the relevance of the latter approach in many preclinical AD investigations [51,52].

Among additional problems adequately investigated by brain imaging is the heterogeneity of neurodegenerative diseases. Detailed recent studies have focused not only on the various forms of AD but also on PD and its atypical syndromes. In an initial study, several PDs that were hindered by substantial clinical and pathological heterogeneities were investigated by numerous imaging techniques including PET and MRI [53]. More recently, combinations with new techniques have been established to investigate, in distinct areas and pathways of the brain, the various molecules and processes, i.e., the molecular imaging of Aβ, tau and α-synuclein, as well as neuroinflammation. This integrative "multimodal approach" was found to be superior to single modality-based methods, with expected future advancements in the field [54]. Other innovative combinations of imaging biomarkers, revealed by MRI and PET, have been employed to investigate subtypes of AD. At the level of previously established groups of patients, the average and clinical characteristics appeared similar. In subtype assignments, however, disagreements were considerable. The subtypes therefore need to be further investigated, with an establishment of their harmonization by appropriate methods [55].

Additional new studies have expanded the investigation of in vivo biomarker-driven profiles by the reinforcement of the techniques employed. In a first example, a versatile form of PET imaging was found able to quantify the molecular targets of interest. By such approach, age-related neurodegenerative diseases, inducing dysregulation of synapses, neuroinflammation, protein misfolding, and other dysfunctions, were appropriately identified. Discussion of these processes has led to the identification of novel biomarkers [56]. A second approach was based on the development of optical imaging (OPI) probes and devices, which are affordable by imaging studies but are limited by their low depth of penetration. The combination of the OPI technique with PET, which is characterized by high depth penetration, resulted in the elimination of each limitation. The affording of new radiolabeled fluorophores made the activation of the dual PET/OPI mode possible, with excellent preclinical imaging results in various pathological conditions [57].

Summing up, the imaging biomarkers presented in this section are profoundly different from the biomarkers of the fluid type. Being recognized from the analysis of brain tissue, they do not fit strictly the definition of biomarkers as distinct molecules related to pathology, given in the first lines of the Introduction. Yet, a growing class of radiotracers, addressed to specific proteins such as Aβ and tau, have recently contributed to the growing knowledge about AD. In other words, the targets of the imaging processes can now be identified as molecules based on which the identification of imaging biomarkers is established [58,59].

#### **6. Conclusions**

This review focused on the new developments of two types of biomarkers, i.e., fluid and imagining biomarkers, generated via two distinct operative pathways originated in the brain. As emphasized in Section 5, even if the structure and the mechanisms of action of the two types are largely different, at least some of their effects are similar, to the point that their results, when considered together, give rise to positive integrations [36,47,48].

How relevant is the medical work of biomarkers? The experience accumulated during the last decade appears positive and encouraging. Many of their properties, such as specificity, multiplicity, and affinity, have strengthened their use in the course of AD, PD, and other neurodegenerative diseases. The present investigation about the pathogenesis of AD and the other neurodegenerative diseases is largely due to biomarker-based data. To

understand the issues underlying complex symptoms such as dementia, biomarkers often operate combined with other disciplines [60].

In basic and applied research, the impact of biomarkers has increased, sustained by the improvement of their technologies. In the fluid field, the recent experience with EVs as tools for biomarker generation has improved their success considerably. An additional development is the recent recognition of increased number of proteins and peptides in the CSF of AD patients [61]. The imaging technology has been strengthened by the development of new radiolabeled molecules, by the use of improved imaging techniques, and by their combination with different procedures. In these studies, the best successes has been obtained when working on early stages of AD, including the identification of pathological alterations and a selective cognitive decline [62].

The results obtained by the use of biomarkers are often of relevance in medical practice. This in particular is the case of diagnosis: O often inadequate when based only on medical practice, it has been recently improved by the multiplicity and specificity of biomarkers. By the use of the latter, it has been established that a disease can be diagnosed even before the appearance of symptoms [10,11,51,52]. Moreover, the intensity of the responses can contribute to the prognosis of patients [28,44]. Thus, diagnostic and prognostic results obtained by biomarkers are often analogous. In contrast, therapy is still problematic. Although open to all neurodegenerative diseases, such problem is particularly aggressive for AD and ALS, for which appropriate therapy is not available. An initial attempt to improve the present situation involves the investigation of diseases at early or even presymptomatic conditions. Drugs slowing down the development of the disease should delay its progress, ultimately resulting in a progressive decrease of the AD mature state. Up to now, however, the results of clinical trials to evaluate the state of the AD therapy have been disappointing. In the future, development should be based on multiple criteria, including the state of the disease, its progression, and the activity of biomarkers focused on critical processes [40].

In conclusion, the recent identification and employment of biomarkers have resulted in increased knowledge and improvement of both basic and applied research. However, as emphasized in the first and subsequent sections of this review, various aspects of their function are still not completely or even inappropriately known. In the future, it appears desirable that intense research about biomarkers will include their critical evaluation and their further improvement, especially in terms of new therapy practice.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not Applicable.

**Data Availability Statement:** All the data reported in this review are available at the American Program PubMed.

**Acknowledgments:** I thank Palma Gallana for the support in the final version of this Review and Gabriella Racchetti for the preparation of Figure 1.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**



#### **References**


#### MDPI

St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Biomedicines* Editorial Office E-mail: biomedicines@mdpi.com www.mdpi.com/journal/biomedicines

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel: +41 61 683 77 34

www.mdpi.com

ISBN 978-3-0365-5730-4