*2.2. GCase Activity is Up-Regulated in GBA2-Deficient LCLs Particularly at the PM*

Several lines of evidence suggest the existence of a cross-talk among the enzymes involved in sphingolipid (SL) catabolism, since modification in the activity/expression of one enzyme could affect that of others [31]. Besides NLGase, another important enzyme involved in GlcCer catabolism is β-glucocerebrosidase (GCase) encoded by the *GBA* gene. Unlike NLGase, GCase is mainly a lysosomal enzyme and only partially associated with the external leaflet of the PM [15]. To investigate the possibility of a cross-talk between GCase and NLGase, we measured the activity of GCase both intracellularly and at the cell surface in control and patient derived LCLs. As shown in Figure 3–panel A, GCase activity was increased in total cell lysates derived from patients' cells with respect to controls. The average activity of GCase was 102 ± 14 pmoles/106 cells/h and 71 ± 17 pmoles/106 cells/h in patient and control derived LCLs, respectively. A marked increase, about three-fold, of GCase activity was also observed at the PM level in patients compared to control LCLs (patients: <sup>33</sup> ± 9 pmoles/106 cells/h; control: 11 ± 4 pmoles/10<sup>6</sup> cells/h). Interestingly, the augmented enzymatic activity is associated with an increase in the GCase protein levels but not in *GBA* mRNA expression (Figure 3B,C).

**Figure 3.** Activity and protein levels of β-glucocerebrosidase (GCase) and *GBA* mRNA expression. (**A**) GCase activity associated with the total cell lysate and plasma membrane of control (light grey) and patient (dark grey) derived LCLs. Activities were expressed as pmoles/10<sup>6</sup> cells/h. Data are expressed as mean ± SD (*n* = 4, \*\*\* *p* < 0.0001 vs. controls). (**B**) Immunoblot of GCase and control GAPDH accompanied by the semi-quantitative graph of normalized GCase/GAPDH. Data are expressed as mean ± SD (*n* = 4 \*\*\* *p* < 0.0001 vs. controls). (**C**) The graph represents the *GBA* mRNA levels of controls (light grey) and patients (dark grey) relative to the average value of controls (*n* = 4), after normalization with the endogenous β-actin gene (*ACTB*). Values represent the mean ± SEM of two independent triplicate experiments.

A substantial body of evidence shows the presence and action of mature and active lysosomal enzymes at the PM in addition to the lysosomes [15,32]. We decided to investigate the activity of two main hydrolases involved in SL catabolism, β-galactosidase and β-hexosaminidase, both intracellularly and at the PM level. We observed an increase only of the β-galactosidase activity associated with the cell surface of LCLs obtained from patients with respect to controls (Table 3).



#### **3. Discussion**

Several lines of evidence indicate that the regulation of GlcCer levels is important for cell homeostasis. GlcCer is a minor component of almost all membranes of eukaryotic cells suggesting an evolutionary strategy aimed to limit its presence. Indeed, de novo biosynthesized GlcCer is mainly used as a building block for the biosynthesis of complex GSLs [33].

In mammalian cells, GSL catabolism occurs by the sequential hydrolysis of the saccharidic chain by removing the reducing sugar. Lysosomes are involved in the catabolism of the endocytic portion of the cell PM and could be considered the principal site, together with the endoplasmic reticulum and Golgi complex, responsible for GSLs turnover [34]. On the other hand, the fine tuning of the GSLs composition is triggered directly at the cell PM by the action of specific glycohydrolases. In particular, the same enzymes that are associated with the lysosomes, such as sialidase Neu1, beta-hexosaminidase, beta-galactosiadase, and GCase, are present at the cell surface, even if in very small amounts, along with the sialidase Neu3 and NLGase, which are enzymes primarily residing at the PM [5]. Loss of function mutations in the lysosomal glycohydrolases determine the onset of lysosomal storage disorders, characterized by the accumulation of non-catabolized substrates.

NLGase is currently the most studied enzyme among the PM glycohydrolases. NLGase deficient mice, obtained both, by gene knockout and by pharmacological inhibition of the enzyme, showed an abnormal GlcCer accumulation in multiple tissues, including brain, liver, and testis. These data were quite surprising considering that the large amount of the GSL catabolism occurs in lysosomes. In addition, these mice were characterized by impaired liver regeneration and male infertility but no neurological involvement was observed [17–19]. However, in humans *GBA2* gene mutations are found in spastic ataxia and spastic paraplegia patients (SPastic Gait locus #46, SPG46) [23–25,35]. Among them, a c.1780G>C (p.Asp594His) missense mutation located in exon 11 of the *GBA2* gene was identified in a Cypriot consanguineous spastic ataxia family [22].

We used LCLs obtained from three patients of this family, who are homozygous for the *GBA2* c.1780G>C missense mutation, to evaluate the effect of the mutation on NLGase activity. By an in vitro enzymatic assay based on the use of CBE to block GCase activity, we found that the mutation strongly affects NLGase activity. Indeed, in pathological cells we were not able to detect any NLGase activity in the total cell lysate, and only a very low residual activity at the cell surface when compared with LCLs derived from four healthy controls. Similar to the results observed in mice, we found that the NLGase loss of function in patient derived LCLs is responsible for an increase of GlcCer content that reaches two-fold to that found in control cells. In addition, patient LCLs show an increased activity of GCase with respect to that measured in controls. In particular, GCase activity associated with the PM of patient derived LCLs is three-fold higher than that of control LCLs. This presumably compensatory effect has already been described in fibroblasts derived from patients affected by Gaucher disease, where the GCase loss of function induced an increase in NLGase activity [31]. Indeed, by the evaluation of the total cell β-glucocerebrosidase activity we did not find statistically significant differences between control and patient derived LCLs (Figure S1).

Taken together, the data, herein reported, further support an important role of NLGase in GlcCer metabolism and the existence of a cross-talk among the enzymes involved in GSL catabolism. Despite this new evidence, the challenges for future studies remain: (i) to explain why NLGase loss of function with concomitant increase of GCase could result in GlcCer accumulation and (ii) which are the molecular mechanisms linking the NLGase-dependent GlcCer accumulation with the onset of spastic ataxia.

#### **4. Materials and Methods**

#### *4.1. Cell Culture*

Seven LCLs (3 patients and 4 controls) were available, which were expanded and sub-cultured for the purposes of this study. Cells were grown in culture medium RPMI (Roswell Park Memorial Institute medium) supplemented with 10% FBS (fetal bovine serum), 1% penicillin/streptomycin and 1% glutamine, and expanded for a period of 6 weeks at 37 ◦C in a 5% CO2 incubator. Genotypes of all patient derived LCLs sub-cultures were reconfirmed with Sanger sequencing and healthy controls were confirmed as homozygous wild-type.

All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and approved by the National Bioethics Committee of Cyprus (EEBK/EΠ/2013/28, date of approval 14 May 2015).

#### *4.2. Evaluation of Enzymatic Activities in Cell Lysates*

The enzymatic activities associated with total cell lysates were determined by an assay based on the use of fluorogenic substrates as previously described [36]. To evaluate NLGase activity, cell lysates were pre-incubated for 30 min at room temperature in McIlvaine buffer (pH 6) with 1 mM CBE (Conduritol-B-epoxide, Merck, Darmstadt, Germany), a specific inhibitor of GCase [37]. For the measurement of GCase activity, cell lysates were pre-incubated for 30 min at room temperature in McIlvaine buffer (pH 5.2) containing 0.1% Triton X100(Merck, Darmstadt, Germany) with 5 nM AMP-dNM (adamantane-pentyl-dNM; *N*-(5-adamantane-1-yl-methoxy-pentyl) deoxynojirimycin, (A generous gift from Prof. Aerts JM form Leiden University) a specific inhibitor of NLGase. The total β-glucocerebrosidase assay was performed using the same procedure without inhibitors and detergents. At the end of the pre-incubation, the reactions were started by the addition of 25 μL of 4-Methylumbelliferyl β-D-glucopyranoside (MUB-β-Glc, Glycosynth, Warrington, UK) at a final concentration of 6 mM. To measure β-galactosidase and β-hexosaminidase activities, the fluorogenic substrates used were 4-Methylumbelliferyl β-D-galactopyranoside (MUB-β-Gal) and 4-Methylumbelliferyl *N*-acetyl-β-D-glucuronide (MUG) (all from Glycosynth, Warrington, UK), respectively. Aliquots of cell lysates were incubated with 25 μL of McIlvaine buffer 4× (0.4 M citric acid /0.8 M Na2HPO4) pH 5.2 and the specific fluorogenic substrates at a final concentration of 500 μM. Water was added to reach the final volumes of 100 μL. At different time points the reaction was stopped by adding 9 volumes of 0.25 M glycine pH 10.7 (Sigma-Aldrich, St. Louis, MO, USA). The fluorescence was detected by a Victor microplate reader (Perkin Elmer, Waltham, MA, USA). Standards of free 4-methylumbelliferone (MUB) were used to construct calibration curves. The enzymatic activities were expressed as pmoles of product/10<sup>6</sup> cells /h.

#### *4.3. Evaluation of Enzymatic Activities at the Cell Surface of Living Cells*

PM-associated activities of total β-glucocerebrosidase, GCase, NLGase, β-galactosidase and β-hexosaminidase were assessed in living cells, plated in a 96-well microplate at a density of 200,000 cells/well, by a high throughput live cell-based assay as previously described [15,27,31,36]. To distinguish between GCase and NLGase activities, cells were pre-incubated for 30 min at room temperature in DMEM-F12 without phenol red (Thermo Fisher Scientific, Waltham, MA, USA) containing 5 nM AMP-DNM or 1 mM CBE, respectively [38]. Total β-glucocerebrosidase assay was performed using the same procedure without any inhibitors. Activities were assayed using the artificial substrate MUB-β-Gal for β-galactosidase, MUG for β-hexosaminidase, and MUB-β-Glc for β-glucocerebrosidases GCase and NLGase. The fluorogenic substrates were solubilized in DMEM-F12 without phenol red at pH 6, with final concentrations of 250 μM, 1 mM, and 6 mM, respectively. Aliquots of medium (10 μL) were analyzed at different time points by a Victor microplate reader (Perkin Elmer, Waltham, MA, USA), after adding 190 μL of 0.25 M glycine, with a pH of 10.7. Standards of free MUB were used to construct calibration curves. The enzymatic activities were expressed as pmoles of product/10<sup>6</sup> cells/h. The experimental design included internal controls. In particular, this method is based on the observation that the fluorogenic substrates commonly used for the in vitro assay of glycohydrolytic activities are not taken up by living cells. To assess that the substrate hydrolysis occurs only upon the activity of PM enzymes, a series of controls was performed. In the used experimental conditions, we did not observe any intracellular fluorescence, evaluated by both fluorescent microscopy and fluorimetric analysis of the cells lysed in 0.25 M glycine (pH 10.7), indicating that the substrates were not able to cross the cell membrane. Moreover, we verified that the artificial substrates did not undergo either spontaneous or secreted enzyme-driven hydrolysis by the establishment of an appropriate control without cells or by the solubilization of MUB-substrates directly in cell culture medium in the presence or not of cells.

#### *4.4. Real-Time PCR*

Total RNA was isolated from the LCLs using the RNeasy® Midi kit (Qiagen, Hilden, Germany) as described in the manufacturer's instructions. cDNA synthesis was performed using 1μg of total RNA according to the instructions of the Protoscript® M-MuLV II First Strand cDNA Synthesis Kit (New England Biolabs, Ipswich, Massachusetts, USA). Real-time PCR was carried out using the CFX96 Real-Time system (Bio-Rad, Hercules, CA, USA) and the amplification was done using the SsoFast EvaGreen Supermix according to the manufacturer's instructions (Bio-Rad, Hercules, CA, USA). The sequences of the primers are available upon request. *GBA1* and *GBA2* mRNA expression levels were normalized to the actin house-keeping gene (*ACTB*) and relative mRNA expression was calculated according to the ΔΔCT method.

#### *4.5. Immunoblotting*

Equivalent amounts of proteins associated with total cell lysates, determined by DC Protein Assay (Bio-Rad, Hercules, CA, USA), were separated on polyacrylamide gels and then transferred to PVDF (Polyvinylidene fluoride) membranes by electroblotting [29]. Blots were incubated with monoclonal rabbit anti-GCase (ab128879, Abcam, Cambridge, UK) or polyclonal rabbit anti-GAPDH (G9545, Sigma-Aldrich) primary antibodies at 4 ◦C overnight, followed by incubation with goat anti-rabbit HRP-conjugated (7074, Cell Signaling) secondary antibody and detection with a chemiluminescent kit (WESTAR ηC, Cyanagen, Bologna, Italy). Digital images were obtained by the chemiluminescence system Alliance Mini HD9 (UVItec, Cambridge, UK).
