*Article* **Prognostic Role of CSF** β**-amyloid 1–42/1–40 Ratio in Patients Affected by Amyotrophic Lateral Sclerosis**

**Tiziana Colletti <sup>1</sup> , Luisa Agnello <sup>2</sup> , Rossella Spataro <sup>3</sup> , Lavinia Guccione <sup>1</sup> , Antonietta Notaro <sup>1</sup> , Bruna Lo Sasso <sup>2</sup> , Valeria Blandino <sup>4</sup> , Fabiola Graziano <sup>4</sup> , Caterina Maria Gambino <sup>2</sup> , Rosaria Vincenza Giglio <sup>2</sup> , Giulia Bivona <sup>2</sup> , Vincenzo La Bella <sup>1</sup> , Marcello Ciaccio 2,† and Tommaso Piccoli 4,\* ,†**


**Abstract:** The involvement of β-amyloid (Aβ) in the pathogenesis of amyotrophic lateral sclerosis (ALS) has been widely discussed and its role in the disease is still a matter of debate. Aβ accumulates in the cortex and the anterior horn neurons of ALS patients and seems to affect their survival. To clarify the role of cerebrospinal fluid (CSF) Aβ 1–42 and Aβ 42/40 ratios as a potential prognostic biomarker for ALS, we performed a retrospective observational study on a cohort of ALS patients who underwent a lumbar puncture at the time of the diagnosis. CSF Aβ 1–40 and Aβ 1–42 ratios were detected by chemiluminescence immunoassay and their values were correlated with clinical features. We found a significant correlation of the Aβ 42/40 ratio with age at onset and Mini Mental State Examination (MMSE) scores. No significant correlation of Aβ 1–42 or Aβ 42/40 ratios to the rate of progression of the disease were found. Furthermore, when we stratified patients according to Aβ 1–42 concentration and the Aβ 42/40 ratio, we found that patients with a lower Aβ 42/40 ratio showed a shorter survival. Our results support the hypothesis that Aβ 1–42 could be involved in some pathogenic mechanism of ALS and we suggest the Aβ 42/40 ratio as a potential prognostic biomarker.

**Keywords:** ALS; biomarker; beta amyloid

#### **1. Introduction**

Amyotrophic lateral sclerosis (ALS) is the most common degenerative motor neuron disease, which results in progressive muscle weakness and causes death in a few years. The pathogenesis of ALS is not fully understood and several pathological processes have been proposed such as abnormal protein aggregation, mitochondrial dysfunction and oxidative stress [1]. To date, the diagnosis of ALS is on clinical features and electrophysiological parameters, indicating the degeneration of both upper and lower motor neurons [2]. Heterogeneity in terms of clinical presentation often makes an early and accurate diagnosis a real challenge for clinicians. For this reason, there has been a growing interest in identifying candidate biomarkers for ALS, which can help make an early diagnosis and predict disease progression. Among these, the role of a neurofilament (NF) phosphorylated

**Citation:** Colletti, T.; Agnello, L.; Spataro, R.; Guccione, L.; Notaro, A.; Lo Sasso, B.; Blandino, V.; Graziano, F.; Gambino, C.M.; Giglio, R.V.; et al. Prognostic Role of CSF β-amyloid 1–42/1–40 Ratio in Patients Affected by Amyotrophic Lateral Sclerosis. *Brain Sci.* **2021**, *11*, 302. https:// doi.org/10.3390/brainsci11030302

Academic Editor: Melissa Bowerman

Received: 25 January 2021 Accepted: 24 February 2021 Published: 27 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/).

heavy chain (pNF-H) and light chain (NF-L) as potential biomarkers for ALS is defining. NFs have a non-specific and not fully clarified role in the pathogenesis of ALS. The abnormal accumulation of NF aggregates was observed in perycaria and proximal axons of motoneurons both in ALS murine models and patients that seemed to be related to an impairment of intracellular transport [3]. Recently, a few authors have shown that the aggregation of NFs is related to their hyperphosphorylation state [4]. pNF-H and NF-L are increased in the cerebrospinal fluid (CSF) of ALS patients in comparison with control groups [5,6] and the higher levels are associated with a more rapidly evolving disease and shorter survival [7]. The role of other candidate biomarkers (such as Tau proteins) is still under investigation [8–10].

ALS shares common pathways with other neurodegenerative disorders. For example, C9 or f72 repeat expansions and TAR DNA-binding protein (TARDBP) mutations have been described in ALS and frontotemporal lobar degeneration (FTLD), modifying the idea of ALS as a disease confined to the motor system to the extreme phenotypic expression of a clinical/pathological continuum with FTLD [11–13]. Furthermore, the presence of β-amyloid (Aβ) deposits at the cortical level, hippocampus and spinal cord motor neurons have been described in ALS patients [14–16], suggesting the possibility of some overlapping features between ALS and Alzheimer's disease (AD).

AD is the most common cause of dementia, characterized by Aβ and Tau deposition, respectively, in senile plaques and neurofibrillary tangles as a result of a complex mechanism known as the amyloid cascade [17]. The amyloid precursor protein (APP) is processed by α-secretase into a soluble form α of the APP (sAPPα) and carbossi-terminal fragment α (CTFα) and by β-secretase sAPPβ and CTFβ. Subsequently, CTFβ is cleaved into Aβ 1–40 or Aβ 1–42 by γ-secretase and the imbalance of this process leads to an overexpression of Aβ 1–42 that precipitates, forming the senile plaques. The consequence is the hyperphosphorylation of the Tau protein and the formation of neurofibrillary tangles [18]. CSF Aβ 1–42 levels combined with total Tau (tTau) and phosphorylated Tau (pTau) are currently used as diagnostic biomarkers for AD with a high sensitivity and specificity [19–22], ameliorating the diagnostic accuracy in the very early stages of the disease.

Due to the pathogenetic similarities among neurodegenerative diseases, possible common pathways between AD and ALS have been investigated. Preclinical studies demonstrated the interaction between superoxide dismutase (SOD) and Aβ and evidence of the amyloid cascade has been reported [23] with an increase of sAPP in the CSF from ALS patients [24] and the post-mortem evidence of the over-expression of APP and Aβ in the hippocampi of ALS patients [17]. On the other hand, it is known that APP regulates glial cell-derived neurotrophic factor (GDNF) expression, having a role on neuromuscular junction formation and probably also in neuromuscular degenerative diseases [23].

Whether or not Aβ has a role in the pathogenesis of ALS is far from being clear but it has been recently proposed that the CSF Aβ 1–42 protein concentration is higher in ALS patients and that it is related to disease severity at the time of diagnosis [25].

Our aim is to evaluate the role of the CSF Aβ 1–42 and Aβ 1–40 concentration and the Aβ 42/40 ratio as a potential predictor factor for progression and overall survival in ALS.

#### **2. Patients and Methods**

#### *2.1. Patients*

Ninety-three (93) ALS patients (M/F: 1.11) were enrolled from the ALS Clinical and Research Center, Department of Biomedicine, Neuroscience and advanced Diagnostics (Bi.N.D.), University of Palermo, Italy, from January 2001 to October 2020. All ALS patients were diagnosed according to El-Escorial revised criteria [2] combined with the neurophysiological ones [26]. The revised ALS Functional Rating Scale (ALSFRS-R) [27] was used to score the severity of the symptoms of ALS patients; a higher score indicated normality and a lower score defined a locked-in condition. ∆FS ((ALSFRS-R at onset–ALSFRS-R at time of diagnosis)/diagnostic delay) was used to define the disease progression [28]. According to the ∆FS, patients could be classified in three groups: slow progression (∆FS < 0.5), intermediate progression (∆FS <sup>≥</sup> 0.5 < 1) and rapid progression (∆FS <sup>≥</sup> 1) [28]. We considered co-morbidities for each patient.

All patients underwent a cognitive/behavioral assessment and the administration of neuropsychological tests such as the Frontal Systems Behavioral Scale (FrSBe), Mini Mental State Examination (MMSE) and Edinburgh Cognitive and Behavioral ALS Screen (ECAS) (S-TAB.1). Fewer than 30% showed some degree of behavioral/cognitive impairment according to the Italian Validation of ECAS but none of them were demented. All patients were tested for the most common ALS-related genes and no known mutations associated with ALS were detected.

ALS patients underwent a lumbar puncture (LP) and a CSF analysis as routine procedures of the diagnostic work-up. For the biomarker analysis, ALS patients were subdivided into three subgroups according to the rate of progression based on ∆FS (i.e., slow: ALS-s; intermediate: ALS-i; rapid: ALS-r). All demographic and clinical features of the selected ALS patients are shown in Table 1. None of the patients enrolled assumed any specific drug for ALS treatment at the time of the LP and all of them started riluzole immediately after the diagnosis was made. None of them participated in clinical trials.

**Table 1.** Demographic and clinical characteristics of the total cohort and amyotrophic lateral sclerosis (ALS) patients stratified based on their rate of progression: slow (ALS-s), intermediate (ALS-i) and rapid (ALS-r). Data are expressed as a median with an interquartile range (IQR).


<sup>A</sup> ∆FS at diagnosis = (ALSFRS-R at onset–ALSFRS-R at diagnosis)/diagnostic delay (months). <sup>a</sup> Forced vital capacity. <sup>b</sup> Body mass index. \* Kruskal–Wallis one way analysis of variance on ranks. \*\* chi-squared test. Bold font indicates a statistical significance (*p* < 0.05).

> All patients gave informed written consent. The study was approved by the local Ethics Committee. All of the clinical and biological assessments were carried out in accordance with the World Medical Association Declaration of Helsinki.

#### *2.2. CSF Collection and Analytical Techniques*

All CSF samples were collected in the morning hours and then sent to the Central Hospital Laboratory for a routine analysis. For biomarker detection, the CSF samples were centrifuged in case of blood contamination, aliquoted in polypropylene tubes and stored at –80 ◦C within one hour until further analysis according to international guidelines [29]. The CSF routine chemical parameters are shown in Table 2.


**Table 2.** Cerebrospinal parameters in ALS patients and in patients with slow (ALS-s), intermediate (ALS-i) and rapid (ALS-r) progression. Data are expressed as a median with an interquartile range (IQR).

\* Kruskal–Wallis one way analysis of variance on ranks. \*\* y = yes, n = no

The CSF Aβ 1–42 and Aβ 1–40 were measured by a chemiluminescent immunoassay CLEIA (Lumipulse G b-amyloid 1–40, Lumipulse G b-amyloid 1–42, Fujirebio Inc. Europe, Gent, Belgium) on a fully automatic platform (Lumipulse G1200 analyzer, Fujirebio Inc. Europe, Gent, Belgium). We used as reference cut-off for the Aβ 1–42 value and the Aβ 42/40 ratio < 650 pg/ML and < 0.055, respectively, as suggested by the manufacturer.

#### *2.3. Statistical Analyses*

All statistical analyses were performed using SIGMAPLOT 12.0 software package (Systat Software Inc., San Jose, CA, USA).

A Shapiro–Wilk test was performed to test the normality of the data. We expressed demographic, clinical and biochemical variables as a median with interquartile ranges (IQR). We performed Kruskal–Wallis one way analysis of variance on ranks to compare non-parametric data, a one way ANOVA to compare parametric data and a chi-squared test to assess differences between the groups. We analyzed non-parametric data with Spearman's rank correlation coefficient and parametric data with Pearson's correlation coefficient, considering *p* values < 0.05 as significant.

A survival analysis was performed with the Kaplan–Meier method and survival curves were compared with the log-rank test. Univariate and multivariate Cox regression analyses were performed to predict risk factors for overall survival.

#### **3. Results**

A retrospective observational study was performed on 93 ALS patients to analyze the role of Aβ 1–42, Aβ 1–40 and the Aβ 42/40 ratio as candidate biomarkers for ALS. As a few studies have shown that the CSF Aβ levels are correlated with the rate of progression [21], we stratified ALS patients into three subgroups: ALS-s (*n* = 19; M/F: 2.16), ALS-i (*n* = 31; M/F: 1) and ALS-r (*n* = 35; M/F: 1.18).

In our study, the total cohort of ALS patients had a median age at onset of 67 years. 96.6% of ALS patients were sporadic with a spinal onset in 70.3% of the whole cohort. At the time of diagnosis, ALS patients showed median values of a forced vital capacity (FVC)% of 80.5 (IQR = 54.75–93.25), of a body mass index (BMI) of 24.8 kg/m2 (IQR = 21.5–27.12) and of a ∆FS of 0.81 (IQR = 0.5–1.33). The Kruskal–Wallis one way ANOVA with the rate of progression (∆FS) as a factor showed statistically significant differences in the age of onset (lower in the ALS-s group, *p* < 0.001), diagnostic delay (longer in the ALS-s group, *p* < 0.001) and survival (longer in the ALS-s group, *p* < 0.001); no statistically significant differences were found for the M/F ratio, education, FVC% and BMI (Table 1). The CSF biochemical profile was similar in the three subgroups (Table 2). The neuropsychological assessments with FrSBe, MMSE and ECAS showed no cognitive/behavioral impairments (Table S1).

Analyzing data with the Shapiro–Wilk test, we found that the CSF Aβ 42/40 values were normally distributed while the Aβ 1–42 and Aβ 1–40 ones were not. As shown in Figure 1, the median values of the CSF Aβ 1–42 concentration and the mean values of the Aβ 42/40 ratio resulted above the reference cut-off (< 650 pg/mL and < 0.05, respectively).

β β β

Δ

Δ

β

β β

β β

β β

β

β

β

β

**Figure 1.** Cerebrospinal fluid (CSF) Aβ levels in ALS patients and in patients with slow (ALS-s), intermediate (ALS-i) and rapid (ALS-r) progression. (**A**) Aβ 1–42; (**B**) Aβ 1–40, (**C**) Aβ 42/40 ratio. Solid dots in A–C represent known outliers.

> Spearman's correlation analyses for the CSF Aβ 1–42 and Aβ 1–40 levels showed no significant correlations (Table 3) while the Pearson's correlation analysis showed a significant correlation of Aβ 42/40 ratio values with the age at onset (*r* <sup>2</sup> <sup>=</sup> <sup>−</sup>0.274, *<sup>p</sup>* = 0.008) and MMSE scores (*r* <sup>2</sup> = 0.396, *p* = 0.019) (Table 4).

> **Table 3.** Spearman's correlation of the CSF Aβ 1–42 and Aβ 1–40 with demographic, clinical and neuropsychological features of ALS patients.


Frontal Systems Behavioral Scale (FrSBe). Mini Mental State Examination (MMSE). Edinburgh Cognitive and Behavioral ALS Screen (ECAS). Bold font indicates a statistical significance (*p* < 0.05).

**Table 4.** Pearson's correlation of the CSF Aβ 42/40 ratio with demographic, clinical and neuropsychological features of ALS patients.


Bold font indicates a statistical significance (*p* < 0.05).

To verify if the CSF Aβ proteins could affect the survival of ALS patients, we stratified ALS patients according to the median values of Aβ 1–42, Aβ 1–40 and the Aβ 42/40 ratio, obtaining three subgroups for each analyzed protein: L-1-Q (i.e., patients with values lower than the first quartile), IQR (i.e., patients with values between the first and the third quartiles) and U-3-Q (i.e., patients with values upper than the third quartile). Only for the Aβ 42/40 ratio did the Kaplan–Meier analysis with a Holm–Sidak post-hoc test show that L-1-Q patients had a significantly shorter survival (27 (IQR: 17–41) months) in comparison with U-3-Q (39 (IQR: 26–60) months) (log-rank = 6.617; *p* = 0.037) (Figure 2).

β **Figure 2.** Kaplan–Meier survival curves of ALS patients stratified according to the median CSF values of the Aβ 42/40 ratio: lower than the first quartile (L-1-Q), interquartile range (IQR) and upper than the third quartile (U-3-Q).

Interestingly, patients in the L-1-Q showed a higher median age in comparison with other subgroups (L-1-Q: 71 (66.5–75.25); IQR: 66.5 (63–71.75); U-3-Q: 65.5 (61–70.5); *p* = 0.019).

β β β β Δ β Subsequently, we performed univariate and multivariate Cox regression analyses to test the predictor role of different demographic and clinical features of ALS patients and the CSF levels of Aβ 1–42, Aβ 1–40 and the Aβ 42/40 ratio. As shown in Table 5, at the univariate regression analysis, the age at onset (*p* = 0.001), diagnostic delay (*p* = 0.001), ∆FS at diagnosis (*p* < 0.001) and Aβ 42/40 ratio (*p* = 0.026) were significantly associated with overall survival. We then considered variables that were positively related to survival at the univariate analysis for the multivariate Cox regression analysis. As shown in Table 6, the diagnostic delay (*p* =0.025), ∆FS at diagnosis (*p* = 0.032) and Aβ 42/40 ratio (*p* = 0.015) were independent predictors of overall survival. Furthermore, the multivariate Cox regression analysis was performed to investigate the role of co-morbidities in overall survival but no significant data were obtained (Table S2).

Δ

−

−

Δ

β


**Table 5.** Univariate Cox regression analysis for the overall survival for ALS patients.

b = regression coefficient; SE = standard error; HR = hazard ratio; CI = confidence interval. Bold font indicates a statistical significance (*p* < 0.05).

**Table 6.** Multivariate Cox regression analysis for overall survival for ALS patients. Significative variables believed to be significant at the univariate analysis were considered for multivariate analysis.


b = regression coefficient; SE = standard error; HR = hazard ratio; CI = confidence interval. Bold font indicates a statistical significance (*p* < 0.05).

#### **4. Discussion**

Our study was aimed at exploring the potential role of Aβ as a prognostic biomarker in ALS. For this purpose, we designed a retrospective observational study that included 93 patients. The CSF Aβ 1–42 and Aβ 1–40 levels and the Aβ 42/40 ratio were determined and correlated with demographic, clinical and neuropsychological features of ALS patients.

In recent years, CSF Aβ levels have been investigated to define their role as potential diagnostic and prognostic biomarkers for ALS and many studies in this field have been reported. However, the two largest studies about this topic found contrasting results. On one hand, higher CSF Aβ 1–42 levels were associated with a poorer prognosis [25] while, on the other hand, an interesting correlation of a higher concentration in patients with better performance was found, reporting increased CSF levels compared with a control group [26].

Even though the CSF Aβ and especially the Aβ 42/40 ratio represent a specific diagnostic biomarker for AD, the idea of shared mechanisms among different neurodegenerative disorders has led many authors to investigate the role of Aβ as a potential modulator of their rate of progression and overall survival. The CSF Aβ 1–42 levels were correlated to conversion from mild cognitive impairment to dementia and the progression of cognitive deficits in AD [27] as well as with the progression of cognitive impairments in Parkinson's disease (PD) [28]. Indeed, lower CSF Aβ 1–42 levels are related to a progressive deposition of Aβ in senile plaques at the cortical level [29]. An intracellular deposition of Aβ 1–42 was also detected in the anterior horn of motor neurons of patients affected by motor neuron disease (MND) [13] while extracellular aggregates of Aβ 1–42 were detected in the hippocampus of ALS and ALS-FTD patients [17]. Studies on murine models of ALS (i.e., SOD1 G93A mice) correlated the overexpression of Aβ with an earlier onset of motor symptoms [30]. Furthermore, few cases of co-morbidity between ALS and AD in a patient showing an overlapping clinical picture have been reported [15,25].

In our study, ALS patients were subdivided into three subgroups (i.e., ALS-s, ALS-i and ALS-r) to analyze the contribution of the CSF Aβ levels on the rate of progression. No statistically significant differences were detected among three subgroups for the CSF Aβ 1–42, Aβ 1–40 and the Aβ 1–42/40 ratio. Indeed, no statistically significant correlation between the Aβ and clinical features including the rate of progression was found. We found a significant correlation between the Aβ 42/40 ratio with the age at onset and MMSE scores.

When analyzing the contribution of the CSF Aβ 1–42, Aβ 1–40 and the Aβ 1–42/40 ratio on overall survival of ALS patients we found that patients with lower Aβ 42/40 ratio values showed a shorter survival in comparison with those with higher values. This finding was confirmed by univariate and multivariate Cox regression analyses, which showed that the Aβ 42/40 ratio could act as an independent predictor for overall survival for ALS patients. A decrease in the CSF Aβ 42/40 ratio values could be indicative of a decrease of the CSF Aβ 1–42 levels because it might deposit in different districts of the central nervous system (CNS) as previously described [29]. In ALS, the presence of intracellular or extracellular aggregates of Aβ 1–42 is probably related to an accumulation of APP following neuronal injury. This accumulation could be due to an impairment of axonplasmatic transport or enhanced biosynthesis of APP, representing an early neuroprotective phase to contrast extracellular and intracellular stresses. As the neuronal injuries continue, a shift toward a neurotoxic phase can occur. APP could be subjected to cleavage in Aβ by alternative mechanisms: caspase 3 giving rise to intracellular aggregates, an accumulation of which gives an increase in oxidative stress, while β-secretase contributes to extracellular deposition [31]. All of these mechanisms may contribute to a decrease of the CSF Aβ 1–42. Studies on murine models of ALS correlated the production of Aβ by β-secretase and consequently deposition as a key event that could improve motor functions and survival [32]. For this purpose, those authors treated asymptomatic and symptomatic SOD-1 G93A mice with a monoclonal antibody able to interfere with β-secretase activity, avoiding the formation of intracellular or extracellular Aβ aggregates: treated asymptomatic ALS mice showed a delay of the onset of symptoms, motor failure and death; however, the same effects were not obtained in treated symptomatic ALS mice.

Another interesting result that we obtained was related to the evidence that ALS patients with lower Aβ 42/40 ratio values presented a higher age at onset than those with higher values. These data were enforced with the finding that there was a statistically significant correlation of Aβ 42/40 ratio values with the age at onset. Considering that the age at onset was considered a strong prognostic factor for ALS [33] and that in cognitively normal subjects the concentration of the CSF AD biomarkers, including the Aβ 42/40 ratio, is associated with age [34], the decrease of the Aβ 42/40 ratio in the CSF of ALS patients might indicate that the triggering of the Aβ cascade could represent an early event that leads to an asymptomatic form of dementia that fails to fully become symptomatic as death occurs. Thus, the coexistence of an elevated age at onset and a low CSF Aβ 42/40 could represent a more severe prognostic condition that could influence survival time. The evidence that ALS patients with a low CSF Aβ 42/40 ratio had a high median age at onset make us speculate that the intracellular or extracellular deposition of Aβ would accelerate the course of the disease, worsening their survival. Indeed, we cannot exclude that a few patients in our population could have a preclinical condition of AD or the presence of age-related amyloid deposition.

Our study had a few limitations. First, the sample size. We studied 96 patients but, for our analyses, we stratified these into three groups resulting in quite small numbers. Related to that is the difference of the age of onset among groups that could partially reduce the significance of our results. Another limitation was the lack of a follow-up for the cognitive formal evaluation. This could have been useful for a correlation with the clinical progression but we found no correlation at the baseline and none of our patients developed clinically significant dementia. Finally, the lack of a control population represented a further limitation. Our goal was to assess the role of the CSF Aβ 1–42, Aβ 1–40 and the Aβ 42/40 ratio in the clinical progression of patients affected by ALS and a comparison with a control group could have enriched our work but we did not consider this to be mandatory.

#### **5. Conclusion**

In our study, we aimed to evaluate the potential role of Aβ in predicting the prognosis in ALS patients. We found that Aβ 42/40 is an independent predictor for survival and could be proposed as a potential prognostic biomarker as suggested by previous reports. Further studies are needed to confirm our findings in a larger population but we consider that we have added a piece to the understanding and management of the disease.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2076-342 5/11/3/302/s1, Table S1: Neuropsychologic assessment of ALS patients by FrSBe, MMSE and ECAS at the time of diagnosis (baseline). Table S2: Multivariate Cox regression analysis for overall survival for ALS patients considering their co-morbidities.

**Author Contributions:** T.C. collected samples, performed statistical analyses, reviewed the data and drafted the manuscript; L.A. made the assays, reviewed the data and the manuscript; R.S. performed clinical evaluations, reviewed the data and the manuscript; L.G. performed neuropsychologic evaluations, reviewed the data and the manuscript; A.N. collected samples, reviewed the data; V.B. performed neuropsychologic evaluations, reviewed the data and the manuscript; F.G. performed clinical evaluations, reviewed the data and the manuscript; B.L.S. made the assays, reviewed the data and the manuscript; C.M.G. made the assays, reviewed the data and the manuscript; R.V.G. made the assays, reviewed the data and the manuscript; G.B. made the assays, reviewed the data and the manuscript; V.L.B. designed the study, performed clinical evaluations and reviewed the data and the manuscript; M.C. designed the study, wrote the original draft, reviewed the data and the manuscript; T.P. designed the study, performed statistical analyses, wrote the original draft, reviewed the data and the manuscript. All authors have read and agreed to the published version of the manuscript.

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

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of University Hospital "Paolo Giaccone" (Institutional Ethic Committee Palermo 1 n◦ 07/2017).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study. The informed consent contained a statement that the biological material may also be used for research purposes.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to current privacy laws.

**Acknowledgments:** This study performed at University Hospital of Palermo is the result of transdisciplinary collaboration of ALS Research and Clinical Center of Palermo, Institute of Clinical Biochemistry, Clinical Molecular Medicine and Laboratory Medicine and IRCSS Bonino Pulejo. The Authors feel grateful for such and lively collaboration.

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

#### **References**


### *Review* **COVID-19 and Alzheimer's Disease**

**Marcello Ciaccio 1,2,\*, Bruna Lo Sasso 1,2, Concetta Scazzone <sup>1</sup> , Caterina Maria Gambino <sup>1</sup> , Anna Maria Ciaccio <sup>3</sup> , Giulia Bivona <sup>1</sup> , Tommaso Piccoli <sup>4</sup> , Rosaria Vincenza Giglio 1,† and Luisa Agnello 1,†**


**Abstract:** The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) is a neurotropic virus with a high neuroinvasive potential. Indeed, more than one-third of patients develop neurological symptoms, including confusion, headache, and hypogeusia/ageusia. However, long-term neurological consequences have received little interest compared to respiratory, cardiovascular, and renal manifestations. Several mechanisms have been proposed to explain the potential SARS-CoV-2 neurological injury that could lead to the development of neurodegenerative diseases, including Alzheimer's Disease (AD). A mutualistic relationship between AD and COVID-19 seems to exist. On the one hand, COVID-19 patients seem to be more prone to developing AD. On the other hand, AD patients could be more susceptible to severe COVID-19. In this review, we sought to provide an overview on the relationship between AD and COVID-19, focusing on the potential role of biomarkers, which could represent precious tool for early identification of COVID-19 patients at high risk of developing AD.

**Keywords:** AD; biomarkers; SARS-CoV-2; neuroinflammation; neurodegenerative nisease; nervous system

#### **1. Introduction**

The Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) is the pathogen responsible for COVID-19 disease, which is characterized by a wide spectrum of symptoms, from fever and cough to multiple organ dysfunctions [1]. Additionally, SARS-CoV-2 can induce, directly or indirectly, several complications involving different organs [2,3]. Nowadays, the clinical course of the infection is unpredictable and characterized by high inter-individual variability. However, more than 80% of COVID-19 patients present ageusia or anosmia, which occurs early during the infection and represent pathognomonic features of the disease [4].

SARS-CoV-2, as well as all members of the human coronaviruses (CoVs) family, is an opportunistic pathogen of the central nervous system (CNS) [5]. The neurological signs and symptoms associated with SARS-CoV-2 infection, such as confusion, headache, hypogeusia/ageusia, hyposmia/anosmia, dizziness, epilepsy, acute cerebrovascular disease [4], are caused by the direct invasion of the virus into the CNS, and the subsequent interaction between SARS-CoV-2 spike protein and the angiotensin-converting enzyme 2 (ACE2) [6–8]. Post-mortem studies revealed the presence of both SARS-CoV-2 antigen and RNA in the brain tissue of COVID-19 patients [9].

**Citation:** Ciaccio, M.; Lo Sasso, B.; Scazzone, C.; Gambino, C.M.; Ciaccio, A.M.; Bivona, G.; Piccoli, T.; Giglio, R.V.; Agnello, L. COVID-19 and Alzheimer's Disease. *Brain Sci.* **2021**, *11*, 305. https://doi.org/10.3390/ brainsci11030305

Academic Editor: Fernando Aguado

Received: 5 February 2021 Accepted: 24 February 2021 Published: 27 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/).

ACE-2 expression is a key determinant of viral tropism and COVID-19 pathogenesis. In the brain, ACE-2 is expressed both on neurons and glial cells as well as on endothelial and arterial smooth muscle cells. ACE-2 is also expressed on the temporal lobe and hippocampus, which represent cerebral regions involved in the pathogenesis of Alzheimer's Disease (AD) [6].

It has been hypothesized that SARS-CoV-2 could cause damage in the CNS by direct neurotoxicity or indirectly through the activation of the host immune response, which could lead to demyelination, neurodegeneration and cellular senescence. Thus, it could accelerate brain aging favoring the development of neurodegenerative diseases, including dementia [10]. However, after the acute recovery phase, the long-term consequences of SARS-CoV-2 infection on accelerated aging and age-related neurodegenerative disorders are actually unknown. Noteworthily, SARS-CoV-2 could potentially induce a worsening cognitive decline in AD patients.

On the other hand, dementia could represent an important risk factor for COVID-19 severity and mortality, as shown by preliminary reports [11,12]. Thus, a mutualistic relationship between SARS-CoV-2 infection and AD can be hypothesized. Figure 1 shows the possible association between AD and SARS-CoV-2 infection by summarizing the possible underlying mechanisms, which are described in the next paragraphs.

**Figure 1.** The complex relationship between Alzheimer's Disease and SARS-CoV-2 infection.

In this review, we sought to provide an overview on the relationship between AD and COVID-19, focusing on the potential role of biomarkers. This should represent a starting point for further investigations.

#### **2. COVID-19 in Alzheimer's Disease**

AD represents the most common form of dementia worldwide [13]. The term dementia refers to a wide spectrum of disorders characterized by global, chronic and generally irreversible cognitive deterioration, leading to the progressive alteration of several functions such as memory, the ability to orient oneself, and alterations of the personality and behavior, which compromise the autonomy of the subject in the daily life [13,14]. The incidence of dementia is increasing in the general population. Indeed, the World Health Organization and Alzheimer Disease International Report of 2016 defined it as a global public health priority [15]. Patients with dementia are frail, dependent on caregivers for daily living activities and need the support of several services resources, such as physical exercise and physiotherapy [16]. Thus, the measures introduced by government authorities during the current COVID-19 pandemic, including confinement and isolation, may exacerbate the cognitive decline. Additionally, patients with AD and mild dementia may either be unwilling or unable to follow recommendations from public health authorities such as sanitize their hands, cover their mouth and nose when coughing, maintain physical distance from others, in part due to the severity of their short-term memory loss and overall cognitive impairment [17].

The brain of AD patients is characterized by amyloid plaque deposition and the presence of neurofibrillary tangles, which induce neuronal damage and synapse loss as well as oligodendroglia degeneration and myelin impairment [18].

Post-mortem studies showed that ACE-2 expression is increased in the brain of AD patients in comparison to controls [19]. Additionally, genome-wide association studies (GWAS) showed that the expression of ACE-2 gene is elevated in the brain tissue of AD patients with increased levels in severe forms [20]. Thus, enhanced ACE-2 expression could represent a risk factor for COVID-19 transmission in AD patients. It has been postulated a direct link between AD and ACE-2 expression mediated by oxidative stress. Specifically, aging leads to the imbalance in the redox state, characterized by the generation of excess reactive oxygen species (ROS) or the dysfunction of the antioxidant system, leading to oxidative stress [21]. AD patients show a significant extent of intracerebral oxidative damage associated with the abnormal marked accumulation of Aβ and the deposition of neurofibrillary tangles [21]. Interestingly, ACE2 inhibitors have recently been suggested as potential treatment for neurodegenerative diseases, including AD [22].

Noteworthily, AD and COVID-19 share several risk factors and comorbidities, such as age, gender, hypertension, diabetes and APOE ε4 expression. Such evidence could in part explain the increased prevalence of SARS-CoV-2 infection in AD patients. However, further studies are mandatory in order to clarify the pathophysiological mechanisms linking AD and COVID-19.

#### **3. Patients with COVID-19 Could Develop AD?**

Overall, CoVs can enter the CNS via different routes, including retrograde axonal transport via the olfactory and enteric neurons or infected lymphocytes, which cross the disrupted blood-brain barrier (BBB) [23].

Aging is characterized by a gradual loss of the BBB integrity [24]. Thus, the elderly could be more susceptible to neuroinvasion during SARS-CoV-2 infection.

SARS CoV-2 infects the olfactory neurons and, through the neuro-epithelium of the olfactory mucosa, reaches the olfactory bulb in the hypothalamus [5,25]. The presence of SARS CoV-2 in the olfactory bulb leads to the activation of non-neuronal cells, such as mast cells, microglia, astrocytes, as well as to the tissue release of pro-inflammatory cytokines. SARS-CoV-2 uses the phospholipids of the infected cells to build its own envelope. The consequence is that the cells, in particular the innate immune cells, lose precursors for

the synthesis of the autacoid local injury antagonist amides (ALIamides), which have a pivotal role for controlling the excessive reactivity [26]. Consequently, the resulting neuroinflammation could become uncontrollable, especially in the elderly, which have a less efficient immune system response [27,28]. Neuroinflammation, associated with intense oxidative stress, could induce neurodegeneration, potentially favoring the development of neurodegenerative diseases, such as AD [25,29]. COVID-19 patients with advanced age and comorbidities with an inflammatory basis, such as diabetes, atherosclerosis and sub-clinical dementia, could be at increased risk of developing AD.

Several pathological mechanisms seem to be involved in the potential increased risk of developing AD in COVID-19 patients.

A growing body of evidence suggested a role for neuroinflammation. Systemic inflammation induces the activation of microglia and astrocytes, which in turn secrete pro-inflammatory cytokines, including IL-1β, IL-6, IL-12, TNF-α. Such biomarkers could be involved in the synaptic dysfunction, inducing neurodegeneration, which could potentially lead to AD [30].

Hypoxic alterations and demyelinating lesions have been described in COVID-19 patients [31–33]. Neuroradiological studies showed alterations of functional brain integrity, especially in the hippocampus, in recovered COVID-19 patients at 3-month follow-up. The hippocampus is an area particularly vulnerable to respiratory viral infections, as shown in experimental studies [34]. Hippocampal atrophy is associated with cognitive decline and represents a common characteristic of AD patients [35,36]. Additionally, the altered BBB could allow the infiltration of immune cells, which may contribute to cognitive decline and dementia in COVID-19 patients. Moreover, endothelial dysfunction, which is a pathognomonic characteristic of COVID-19, and loss of pericytes could impair the clearance of cerebral metabolites, including Aβ peptides. The excess and accumulation of Aβ protein in senile plaques, especially in the hippocampus, represent the main pathophysiological mechanism underlying the AD. Some authors showed that severe COVID-19 presents ischemic white matter damage due to the reduced perfusion secondary to hypercoagulability and disseminated intravascular coagulation (DIC), which are common features of severe COVID-19. Neuroimaging and experimental studies showed that ischemic white matter damage occurs at a very early stage of AD, accelerates the progression of the disease and contributes to cognitive decline [37,38]. Moreover, cerebral hypoperfusion can increase the phosphorylation rate of tau [39].

In severe COVID-19, the systemic inflammation characterized by the so-called "cytokine storm" leads to the disruption of the blood–brain barrier and neural and glial cell damage that could be involved in long-term sequelae. Systemic inflammation is recognized as a pathophysiological mechanism underlying AD [40]. Also, pro-inflammatory cytokines alter the capacity of the microglial cells to phagocyte b-amyloid, promoting the accumulation of amyloid plaques [41]. The virus-induced systemic inflammatory storm, associated with a massive release of mediators able to access the CNS due to the increased permeability of the blood-brain barrier, could amplify neuroinflammation and contribute to the neurodegeneration process [42].

Another interesting piece of evidence suggests that the potential increase of AD risk in COVID-19 patients could be related to Aβ, which can act as an antimicrobial peptide. Thus, it could be postulated that the SARS-CoV-2 neuroinvasion could promote Aβ generation, as part of the immune response, and the b-amyloid cascade leading to b-amyloid deposition [43]. However, this is only a hypothesis that must be proved.

McLoughlin et al. showed that COVID-19 hospitalized patients who developed delirium during their hospitalization, after 1-month discharge had lower cognitive scores [44]. However, the difficulty in performing neuropsychological assessment leads to a poor understanding of the neurological impact of SARS-CoV-2 infection. Overall, ARDS is associated with a high prevalence of long-term cognitive impairment in critically ill patients [45]. Specifically, mechanical ventilation, which is a standard therapy to maintain adequate gas exchange during ARDS, also in severe COVID-19 patients, could contribute

to long-term cognitive impairment [46–49]. Experimental studies showed that short-term mechanical ventilation triggers the neuropathology of AD by promoting cerebral accumulation of the Aβ peptide, systemic and neurologic inflammation, and blood–brain barrier dysfunction [50].

The long-term complications of COVID-19 would be expected in the next 10–15 years. Nowadays, it is not possible to assess them because the pandemic started last year. However, in the future, it will be pivotal to evaluate the risk of long-term COVID-19 neurological sequelae, especially in the elderly and patients who developed severe forms.

The potential mechanisms involved in cognitive impairment in COVID-19 patients can be summarized as follows: (i) direct SARS-CoV-2 infection in the CNS; (ii) systemic hyper inflammatory response to SARS-CoV-2; (iii) cerebrovascular ischemia due to endothelial dysfunction; (iv) severe coagulopathy; v) mechanical ventilation due to ARDS or severe disease; (vi) peripheral organ dysfunction.

Table 1 summarizes the potential mechanisms linking SARS-CoV-2 infection and the development of AD.


**Table 1.** Potential mechanisms involved in Alzheimer's Disease (AD) risk in COVID-19 patients.

ACE-2: Angiotensin Converting Enzyme 2; AD: Alzheimer's Disease; APOE: Apolipoprotein E; BBB: Blood Brain Barrier; COVID-19: Coronavirus Disease 2019; CSN: Central Nervous System; NO: Nitric Oxide; SARS-CoV-2: Severe Acute Respiratory Syndrome Coronavirus 2.

#### **4. Biomarkers of Cognitive Decline in COVID-19 Patients**

#### *4.1. Neuronal Injury*

Biomarkers of neurodegeneration in the cerebrospinal fluid (CSF), such as tau proteins, neurofilament light chain protein (NfL), and glial fibrillary acidic protein (GFAp), are increased in COVID-19 patients and associated both with neurological symptoms and disease severity [53–57].

T-tau is a biomarker of neuronal death. Its levels are increased in several neurodegenerative diseases, including AD. Specifically, the biochemical diagnosis of AD relies on the detection of a CSF biomarker profile characterized by the decrease of amyloid beta 1-42

(Aβ 1-42), the ratio Aβ 1-42/1-40 and the increase of t-Tau and p-Tau levels [13]. Some Authors found that COVID-19 patients have an increase of CSF t-Tau levels suggesting the presence of neuronal damage. However, to date, levels of amyloid beta have never been investigated in such patients.

Among intermediate filaments expressed in cerebral cells, GFAp and Neurofilmanents have been evaluated in COVID-19 patients.

GFAP is highly expressed in astrocytes and represents a biomarker of astrocytic activation/injury [58]. AD is characterized by amyloid plaques surrounded by reactive astrocytes, which show an increased expression of intermediate filaments, including GFAP [58]. To date, only two studies evaluated the role of GFAp in COVID-19 patients [53,54]. The authors showed that severe COVID-19 patients had higher plasma concentrations of GFAp than controls.

Neurofilaments are cytoskeletal proteins of neurons, particularly abundant in axons. Neurofilaments comprise three subunits: neurofilament light chain (NF-L), neurofilament medium (NF-M) and neurofilament heavy (NF-H). Among these, NF-Ls are the most abundant.

Following axonal damage, NFs are released into CSF. Thus, they represent a biomarker of axonal damage and neuronal death. CSF NFs levels are increased in several neurological disorders, including AD [59]. Increased levels of serum and CSF NF-L have been found in severe COVID-19 patients [53–57,60].

Only one study evaluated t-Tau in COVID-19 patients and reported increased levels of t-Tau in severe cases.

To date, a few authors evaluated the CSF biochemical profile of COVID-19 patients due to the difficulty of obtaining such biological fluid. However, preliminary literature evidence raises awareness for potential long-term neurologic sequelae following COVID-19. Although severe COVID-19 patients have CSF biochemical alterations indicative of neuronal and axonal damage, it is not possible to draw definitive conclusions on the cognitive impairment. Longitudinal studies are required to evaluate the potential neurological sequelae and the risk of developing AD.

#### *4.2. Genetic Variants*

The most important known predisposing risk factor for AD is the polymorphism APOE ε4, with the ε4ε4 (homozygous) genotype being associated with a 14-fold increase in AD risk. Specifically, APOE ε4 is correlated with low cerebral blood flow and subcortical ischaemic white matter damage, as well as neuroinflammation in AD patients [51]. Kuo et al. showed that individuals carrying APOE ε4 in homozygous had a higher prevalence of SARS-CoV-2 infection. Additionally, APOE ε4ε4 allele was associated with an increased risk of developing severe COVID-19, independently of dementia, and other comorbidities, including cardiovascular disease, and type-2 diabetes [52]. Thus, APOE ε4 represents a common risk factor for AD and SARS-CoV-2 infection. APOE ε4 could promote vulnerability to viral infection and neurodegeneration. Thus, it can be postulated that the SARS-CoV-2 infection could be a promoting factor for neurodegeneration in individuals with susceptible genetic variants [52].

However, the relationships between APOE ε4, COVID-19, and AD must be elucidated.

#### *4.3. Inflammatory Biomarkers*

Some inflammatory biomarkers, including IL-6, IL-1, and galectin-3 (Gal-3), have been proposed as a link between COVID-19 and AD.

IL-6 represents one of the most studied cytokines in COVID-19. Circulating increased levels of IL-6 are associated with a high risk of developing severe COVID-19 and mortality. Accordingly, it represents a reliable prognostic biomarker in SARS-CoV-2 infection [61]. IL-6 is also a prognostic biomarker of AD. Indeed, its increased levels are associated with the progression of the disease and worse cognitive performance [62]. Thus, IL-6 represents a common biomarker for COVID-19 and AD.

IL-6 exerts its biological effects by the interaction with IL-6R, which can be expressed on the membrane of immune, epithelial and liver cells or it can be present in soluble form. The latter represents an agonist of IL-6. The complex IL-6/IL-6R can activate intracellular pathways involved in the immunoinflammatory response [62,63].

Alterations in IL-6 and IL-6R genes could be involved in the onset and progression of several diseases, including infectious diseases, such as COVID-19, and neurodegenerative diseases, such as AD [64–66]. The "Disease and Function analysis" performed by Strafella et al. showed that IL-6 and IL-6R could be involved in neuroinflammation, synaptic damage, microglia activation and cognitive impairment in AD pathogenesis [63].

Similar to IL-6, IL-1 represents a prognostic biomarker of SARS-CoV-2 infection, with increased levels associated with worse prognosis [67,68]. IL-1 is a pro-inflammatory cytokine produced by several cell types, including glia and neurons. IL-1 levels have been found to increase in the brain of AD [69]. In vitro studies reported that IL-1 could induce neuronal death by the direct effect on neurons or indirectly by glial production of neurotoxic substances. Additionally, IL-1 is involved in the physiological regulation of hippocampal plasticity and memory processes. Literature evidence showed that alterations of IL-1 levels, both positively (increase) and negatively (decrease), are associated with impaired memory functioning. Thus, the increased levels of IL-1 found in COVID-19 patients could enhance cognitive decline, leading to the development of AD [70].

Gal-3 is a carbohydrate-binding protein belonging to the family of lectins. It has pleiotropic functions, with a key role in several physiological and pathological processes, including inflammation and fibrosis [71–73]. Increased levels of Gal-3 have been found in severe COVID-19 patients. It has been postulated that Gal-3 promotes COVID-19 progression by supporting the hyper-inflammation reaction and lung fibrosis, which is associated with the acute phase of diffuse alveolar damage, edema, and hypoxia [74]. Increased levels of Gal-3 have also been described in the serum of AD patients [75]. Studies on AD animal models showed that Gal-3 could be involved in the Aβ aggregation and amyloid plaque formation [76]. Thus, it can be hypothesised that increased levels of Gal-3 in COVID-19 patients could also be involved in the damage leading to the development of AD. However, further studies are mandatory to confirm such a hypothesis.

#### **5. Conclusions**

Nowadays, the question "Can SARS-CoV-2 infection increase the risk for development of Alzheimer's Disease?" actually remains unanswered. There is an urgent need for prospective studies to address such question.

Neurological sequelae, including the cognitive impairment leading to AD, could represent an important complication of COVID-19. Further detailed clinical, laboratory, and neuropathological studies will help to elucidate the underlying pathophysiological mechanisms of the COVID-19 neurological complications. A longitudinal follow-up of COVID-19 patients, especially older adults and severe cases, is required to detect the potential long-term neurological consequences of SARS-CoV-2 infection. In such scenario, biomarkers represent reliable tools for early monitoring of COVID-19 patients and early detection of those at high risk of developing neurological sequelae, such as AD. Currently, there is still little literature evidence to draw definitive conclusions. However, an important relationship between AD and COVID-19 seems to exist.

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

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

#### **References**

