Curcumin Has Beneficial Effects on Lysosomal Alpha-Galactosidase: Potential Implications for the Cure of Fabry Disease

Fabry disease is a lysosomal storage disease caused by mutations in the GLA gene that encodes alpha-galactosidase (AGAL). The disease causes abnormal globotriaosylceramide (Gb3) storage in the lysosomes. Variants responsible for the genotypic spectrum of Fabry disease include mutations that abolish enzymatic activity and those that cause protein instability. The latter can be successfully treated with small molecules that either bind and stabilize AGAL or indirectly improve its cellular activity. This paper describes the first attempt to reposition curcumin, a nutraceutical, to treat Fabry disease. We tested the efficacy of curcumin in a cell model and found an improvement in AGAL activity for 80% of the tested mutant genotypes (four out of five tested). The fold-increase was dependent on the mutant and ranged from 1.4 to 2.2. We produced evidence that supports a co-chaperone role for curcumin when administered with AGAL pharmacological chaperones (1-deoxygalactonojirimycin and galactose). The combined treatment with curcumin and either pharmacological chaperone was beneficial for four out of five tested mutants and showed fold-increases ranging from 1.1 to 2.3 for DGJ and from 1.1 to 2.8 for galactose. Finally, we tested a long-term treatment on one mutant (L300F) and detected an improvement in Gb3 clearance and lysosomal markers (LAMP-1 and GAA). Altogether, our findings confirmed the necessity of personalized therapies for Fabry patients and paved the way to further studies and trials of treatments for Fabry disease.


Introduction
In the last 20 years, the keyword "curcumin" increased its presence from 0 to 0.2% of the total publication volume in biochemistry, medicine, and pharmacology. Roughly 2000 papers on curcumin were published in 2021 (as represented by English articles found on Scopus belonging to the BIOC, MEDI, and PHARM subject areas). Curcumin is a turmeric-derived compound, the interest in which in Western medicine is based on its historical usage in Chinese medicine. In fact, turmeric (Curcuma longa L.) has a long tradition as a treatment for different illnesses that range from oxidative stress-related pathogenesis to anorexia [1][2][3][4].

Results and Discussion
In this study, we tested the effect of curcumin treatment on AGAL activity over a panel of five mutants; specifically: c.109G>A (p.Ala37Thr, A37T), c.730G>C (p.Asp244His, D244H), c.898C>T (p.Leu300Phe, L300F), c.838C>A (p.Gln280Lys, Q280K), and c.805G>A (p.Val269Met, V269M). In particular, p.A37T causes the atypical renal-dominant phenotype [48], while p.D244H, p.Q280K, and p.V269M are related to the classic phenotype [49][50][51]; no information was given on the phenotype associated with p.L300F [52]. Immortalized fibroblasts transfected with individual pCMV6-AC plasmids carrying the five GLA mutants (IF-GLA-MUTs), immortalized fibroblasts transfected with the pCMV6-AC plasmid carrying wt-GLA (IF-GLA), and immortalized fibroblasts transfected with the empty vector (IF-NULL) obtained as described in [28] were treated with 20 µM curcumin or with no drug for 48 h. The time and dose of curcumin were selected based on the data available in literature, particularly those regarding cytotoxicity on fibroblasts [53][54][55]. After treatment, the AGAL-specific activity was tested on cell protein extracts. As shown in Figure 1, the wild type and all mutants except one (IF-GLA-V269M) showed a significant AGAL improvement upon curcumin treatment. An immunoblot reflected the enzyme activity results ( Figure 2). The activity fold-increase varied between 1.4 (IF-GLA-D244H) and 2.2 (IF-GLA-L300F and IF-GLA-Q280K).
We tested the hypothesis that it could be used to enhance AGAL activity in FD cells both alone or in synergy with the two PCs (DGJ or galactose).

Results and Discussion
In this study, we tested the effect of curcumin treatment on AGAL activity over a panel of five mutants; specifically: c.109G>A (p.Ala37Thr, A37T), c.730G>C (p.Asp244His D244H), c.898C>T (p.Leu300Phe, L300F), c.838C>A (p.Gln280Lys, Q280K), and c.805G>A (p.Val269Met, V269M). In particular, p.A37T causes the atypical renal-dominant phenotype [48], while p.D244H, p.Q280K, and p.V269M are related to the classic phenotype [49][50][51]; no information was given on the phenotype associated with p.L300F [52]. Immortalized fibroblasts transfected with individual pCMV6-AC plasmids carrying the five GLA mutants (IF-GLA-MUTs), immortalized fibroblasts transfected with the pCMV6-AC plasmid carrying wt-GLA (IF-GLA), and immortalized fibroblasts transfected with the empty vector (IF-NULL) obtained as described in [28] were treated with 20 μM curcumin or with no drug for 48 h. The time and dose of curcumin were selected based on the data available in literature, particularly those regarding cytotoxicity on fibroblasts [53][54][55]. After treatment, the AGAL-specific activity was tested on cell protein extracts. As shown in Figure  1, the wild type and all mutants except one (IF-GLA-V269M) showed a significant AGAL improvement upon curcumin treatment. An immunoblot reflected the enzyme activity results ( Figure 2). The activity fold-increase varied between 1.4 (IF-GLA-D244H) and 2.2 (IF-GLA-L300F and IF-GLA-Q280K).   We also explored the potential of curcumin as a PC enhancer; i.e., a molecule able to strengthen the chaperoning effect of PCs. Cells were treated for 48 h with 10 μM DGJ either alone or in combination with 20 μM curcumin, and the protein extracts were analyzed. Figure 3 shows that the presence of curcumin improved AGAL stabilization induced by DGJ in four out of the five tested mutants (L300F, D244H, Q280K, and V269M) in a fold-change range of 1.1 (IF-GLA-V269M) to 2.3 (IF-GLA-L300F). As shown in Figure  4, immunoblots reflected the results.  We also explored the potential of curcumin as a PC enhancer; i.e., a molecule able to strengthen the chaperoning effect of PCs. Cells were treated for 48 h with 10 µM DGJ either alone or in combination with 20 µM curcumin, and the protein extracts were analyzed.  We also explored the potential of curcumin as a PC enhancer; i.e., a molecule able to strengthen the chaperoning effect of PCs. Cells were treated for 48 h with 10 μM DG either alone or in combination with 20 μM curcumin, and the protein extracts were ana lyzed. Figure 3 shows that the presence of curcumin improved AGAL stabilization in duced by DGJ in four out of the five tested mutants (L300F, D244H, Q280K, and V269M in a fold-change range of 1.1 (IF-GLA-V269M) to 2.3 (IF-GLA-L300F). As shown in Figure  4, immunoblots reflected the results.   In addition to DGJ, which is approved for therapeutical use, galactose is a low-affinity PC for AGAL, thus it requires a high dosage for patients to receive a beneficial effect. Investigation of the potential use of galactose supplementation showed very promising results for other rare diseases (mainly congenital disorders of glycosylation) [56][57][58][59][60]. Therefore, we tested whether the presence of curcumin could improve its stabilizing effect. We analyzed protein extracts derived from cells treated for 48 h with 100 mM galactose either alone or in combination with 20 μM curcumin. As Figure 5 shows, galactose potentiation benefited four out of the five tested mutants in a fold-change range of 1.1 (IF-GLA-V269M) to 2.8 (IF-GLA-L300F). Immunoblots reflecting the enzyme activity results are shown in Figure 6. Interestingly, different mutants showed different behaviors when moving to the combined therapies with PCs. For example, mutant IF-GLA-A37T did not show DGJ potentiation upon curcumin treatment ( Figure 3). On the contrary, its activity improved upon combined treatment with galactose and curcumin with respect to galactose monotherapy ( Figure 5). The opposite behavior was detected for IF-GLA-D244H, which was responsive to the combination of DGJ and curcumin but not to galactose and curcumin. These results were particularly interesting in a disease that requires personalized therapies.  In addition to DGJ, which is approved for therapeutical use, galactose is a low-affinity PC for AGAL, thus it requires a high dosage for patients to receive a beneficial effect. Investigation of the potential use of galactose supplementation showed very promising results for other rare diseases (mainly congenital disorders of glycosylation) [56][57][58][59][60]. Therefore, we tested whether the presence of curcumin could improve its stabilizing effect. We analyzed protein extracts derived from cells treated for 48 h with 100 mM galactose either alone or in combination with 20 µM curcumin. As Figure 5 shows, galactose potentiation benefited four out of the five tested mutants in a fold-change range of 1.1 (IF-GLA-V269M) to 2.8 (IF-GLA-L300F). Immunoblots reflecting the enzyme activity results are shown in Figure 6. Interestingly, different mutants showed different behaviors when moving to the combined therapies with PCs. For example, mutant IF-GLA-A37T did not show DGJ potentiation upon curcumin treatment ( Figure 3). On the contrary, its activity improved upon combined treatment with galactose and curcumin with respect to galactose monotherapy ( Figure 5). The opposite behavior was detected for IF-GLA-D244H, which was responsive to the combination of DGJ and curcumin but not to galactose and curcumin. These results were particularly interesting in a disease that requires personalized therapies.
The effectiveness of curcumin was then evaluated concerning the phenotypic response to the monotherapy or combined therapy on the mutant with the highest fold-increase for all the treatments. The IF-GLA-L300F mutant was treated for up to 50 days with drug administration every seven days. At the end of the long-term treatments, cells were lysed, lipid extraction was accomplished following a described protocol, and Gb3 content was evaluated via LC-MS/MS [61,62]. As shown in Figure 7, Gb3 clearance was significantly improved upon curcumin treatment both in monotherapy (panel A) or combined therapy with DGJ (panel B). The Gb3 quantity in treated cells ( Figure 7A,B) was comparable to that of IF-GLA (Supplementary Figure S1A).
Jehn et al. [63] recently described alterations in the lysosomal pathway in the context of FD. In particular, they reported an increased expression in lysosomal hydrolases that was potentially due to Gb3 accumulation in lysosomes. Pereira et al. [64] described higher levels of lysosome-associated membrane protein 1 (LAMP-1) in FD lymphocytes compared to healthy controls, and we were able to confirm this upregulation in our FD cell model (Supplementary Figure S1B). The IF-NULL cells also showed higher levels of acidic αglucosidase (GAA) than IF-GLA (Supplementary Figure S1C).
To evaluate the effect of curcumin treatment on potential lysosomal biomarkers, the IF-L300F cell line was treated with curcumin with or without DGJ. As shown in Figure 7, curcumin treatment resulted in a reduction in LAMP-1 levels (panel C), which had been previously described upon ERT treatment [64,65]. In addition, GAA activity is reduced upon curcumin treatment, in curcumin monotherapy (panel D), or in combined therapy (panel E). These results highlighted the beneficial effect of curcumin treatment on the FD cell model. improved upon combined treatment with galactose and curcumin with respect to galactose monotherapy ( Figure 5). The opposite behavior was detected for IF-GLA-D244H, which was responsive to the combination of DGJ and curcumin but not to galactose and curcumin. These results were particularly interesting in a disease that requires personalized therapies.   The effectiveness of curcumin was then evaluated concerning the phenotypic response to the monotherapy or combined therapy on the mutant with the highest foldincrease for all the treatments. The IF-GLA-L300F mutant was treated for up to 50 days with drug administration every seven days. At the end of the long-term treatments, cells were lysed, lipid extraction was accomplished following a described protocol, and Gb3 content was evaluated via LC-MS/MS [61,62]. As shown in Figure 7, Gb3 clearance was significantly improved upon curcumin treatment both in monotherapy (panel A) or combined therapy with DGJ (panel B). The Gb3 quantity in treated cells ( Figure 7A,B) was comparable to that of IF-GLA (Supplementary Figure S1A). Curcumin has been described to disrupt Hsp90's molecular function [66][67][68]. The inhibition mechanism is not fully understood; one hypothesis is the disruption of p210 bcr/abl with the Hsp90/p23 complex [66]. Sang et al. demonstrated the major role of Hsp90 in the curcumin-mediated protective effect in an Alzheimer's disease cell model [69]. In particular, silencing Hsp90 significantly attenuated the rescuing effect of curcumin, while Hsp90 overexpression facilitated its effect. Jehn et al. previously demonstrated that AGAL rescue in an FD model had a prominent effect on Hsp90 expression, thus suggesting a role of the molecular chaperon in AGAL folding [63]. We hypothesized that Hsp90 has a prominent role in mediating curcumin effects in FD, thereby allowing AGAL precursor stabilization. Its action on the exosome/microvesicle secretion pathway has also been observed, which suggests its potential role in the treatments of LSDs [70]. Further studies will be needed both to explore the mechanism of action of curcumin in FD and to focus on a wider panel of mutations and precisely analyze their responsiveness to the mono-or combined therapies. We are aware that curcumin preparations might contain curcuminoid contaminants and would be better described as curcuminoid extracts. The effectiveness of curcumin was then evaluated concerning the phenotypic response to the monotherapy or combined therapy on the mutant with the highest foldincrease for all the treatments. The IF-GLA-L300F mutant was treated for up to 50 days with drug administration every seven days. At the end of the long-term treatments, cells were lysed, lipid extraction was accomplished following a described protocol, and Gb3 content was evaluated via LC-MS/MS [61,62]. As shown in Figure 7, Gb3 clearance was significantly improved upon curcumin treatment both in monotherapy (panel A) or combined therapy with DGJ (panel B). The Gb3 quantity in treated cells ( Figure 7A,B) was comparable to that of IF-GLA (Supplementary Figure S1A).  At the end of the treatment, the immunoblot showed a reduction in the levels of lysosome-associated membrane glycoprotein 1 (LAMP-1) upon curcumin treatment with or without DGJ (C). In addition, a reduction in GAA activity was determined via an enzyme activity assay both in monotherapy (D) (two-tailed unpaired t-test, **** = p < 1 × 10 −4 , n = 6, p. 3.95 × 10 −6 ) or in combined therapy with DGJ (E) (two-tailed unpaired t-test * = p < 5 × 10 −2 , n = 6, p. 3.3 × 10 −2 ). FD patients show multiple clinical phenotypes. The classical form of FD usually presents symptoms such as neuropathic pain, cornea verticillata, and angiokeratoma, as well as long-term manifestations such as hypertrophic cardiomyopathy, cardiac rhythm disturbances, progressive renal failure, and strokes. In nonclassical FD, also known as late-onset or atypical FD, patients have residual enzyme activity and lower levels of the deacetylated substrate. This disease form is characterized by milder manifestations that affect just a single organ [71]. However, nonclassical FD patients may experience the same long-term effects as in the classical form [72].
The central role of dysregulated autophagy in LSDs was proposed many years ago [73], and its predominance in FD was recently highlighted. In fact, autophagy is among the mechanisms that underlie the FD phenotype together with overall lysosomal dysfunction, lipid dysmetabolism, and inflammation [74]. Tens of different biological effects of curcumin have been described over the years, and more applications are still being recorded yearly [22]. These mainly concern human health and mostly refer to the improvement of common pathological conditions that include both mild or severe conditions. Molecular mechanisms have already been studied [5,21]. It is worth noting that the beneficial effects of curcumin on cardiac and kidney function have been associated with the direct action of the molecule on autophagy and inflammatory pathways [75][76][77][78][79][80][81][82].
The beneficial effects of curcumin also have been described in cases of rare diseases such as Niemann-Pick type C disease (NPC) [17], neuronal ceroid lipofuscinosis (NCL) [83], and Tay-Sachs disease [18]. It is of utmost importance that for NPC, a triple combination therapy using miglustat, curcumin, and ibuprofen was investigated that resulted in a greater neuroprotective benefit compared with single and dual therapies [84].
Herein, we did demonstrate in an FD cell model that both curcumin or curcumin combined with DGJ produced an improvement in the phenotype of FD. Heart failure and renal involvement are significant issues for FD patients. Thus, the action of curcumin on different molecular mechanisms that leads to a general improvement might underlie its possible benefits in a large cohort of patients with different phenotypes.
These results represent a preliminary study, and further research will be needed to fully address the feasibility of moving from in vitro models to patients. Nevertheless, our results are promising because they open the field to using curcumin for both classic and non-classic FD phenotypes. Of course, the number of genotypes analyzed will need to be widened, and the effects of curcumin will have to be tested in different models (mainly patients' fibroblasts or lymphoblasts carrying different mutations). In addition, the timing and dosage of curcumin will have to be defined.

Materials
The RPMI, fetal bovine serum, and reagents for cell cultures were purchased from Euroclone (Milan, Italy).
The curcumin was from BDH Chemicals Ltd. The Alpha-Galactosidase Polyclonal Antibody (product number: PA5-27349) and GAPDH Loading Control Monoclonal Antibody (product number: MA5-15738) were purchased from ThermoFisher Scientific (Milan, Italy). The lysosome-associated membrane glycoprotein 1 antibody (product number: H4A3, DHSB) was a kind gift from the Telethon Institute of Genetics and Medicine (TIGEM). All of the materials were used without further purification.

Cell Cultures
Immortalized patient-derived fibroblasts carrying a large deletion in GLA exons 3 and 4 were obtained from the Telethon Biobank and stably transfected as described in Monticelli et al. [62]. Cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, 2 mM glutamine, 0.5 mg/mL penicillin, 0.5 mg/mL streptomycin, and non-essential amino acids at 37 • C in 5% humidified CO 2 . Geneticin 0.1 mg/mL was used to maintain the selection. Treatments with drugs were performed in the absence of geneticin. Drugs were dissolved in dimethyl sulfoxide, which represented the control for the untreated samples.

Enzymatic Activity Assays
Fibroblasts from a 90% confluent 20 cm 2 plate were collected in Roche M cOmplete lysis buffer then centrifuged at 14,000× g for 10 min. An AGAL enzymatic activity assay was performed as described in [85] with the modifications described in Monticelli et al. [62].

Gb3 Extraction
The extraction was accomplished according to the protocol outlined by Bligh and Dyer [61] with a few modifications as described in [62]. Briefly, cell pellets were lysed via resuspension in water and freezing-thawing, the soluble proteins were measured, and then lactosylsphingosine was added as an internal standard (2.5 ng of standard/µg of protein). Lipid extraction was performed with chloroform, methanol, water, and hydrochloric acid up to a final condition of (1:1:1:0.05) added in the following order: (i) chloroform/methanol (1:2); (ii) HCl; (iii) chloroform; (iv) water. Centrifuging (1500× g, 45 min at 20 • C) allowed us to obtain upper and lower phases. The samples were eventually dried under nitrogen and then analyzed via liquid chromatography-tandem mass spectrometry. For the UPLC-MS/MS analysis of Gb3, see Monticelli et al. [62].

Miscellaneous
The protein concentration was determined using the Bradford method with BSA as the standard [86]. Immunoblotting was performed under standard conditions as given in [40]. The data handling, analysis, and visualization were performed using the R environment for statistical computing (v4.2.1) with the tidyverse collection of packages (v1.3.1) [87,88] for the data handling and the rstatix (v0.7.0) [89] and ggpubr (v0.4.0) [90] packages for the statistical analysis and visualization, respectively. We performed unpaired two-tailed t-tests using the rstatix::t-test() function. All of the experiments were performed at least in biological duplicate; each biological duplicate was analyzed at least in technical duplicate. Biological replicates were considered in the statistical analysis.

Conclusions
This paper presented the results of a pilot study aimed at improving therapeutic approaches to FD. An AGAL increase in cells is highly beneficial for FD patients regardless of the mechanism leading to this effect. This is demonstrated in the approved therapies (ERT and PCT) [91] based on very different approaches but with the same outcome. We explored the possibility of obtaining this effect in an FD cell model upon curcumin treatment by analyzing the improvement in AGAL activity and the AGAL increase via immunoblotting.
The wide spectrum of biological activities makes curcumin a very unique and interesting molecule in the biomedical field [5]. For this reason, there is a large amount of literature in the field that will certainly benefit the Fabry community. In particular, data on the absorption, distribution, metabolism, and excretion (ADME) of curcumin have been collected over several decades that shows that curcumin bioavailability poses a severe limitation to its application for therapeutic purposes. To overcome this issue, different formulations to enhance curcumin bioavailability have already been widely investigated and clinical outcomes are also available. These research outputs include the combination of curcumin with adjuvants, the discovery of its structural analogues, its nanoformulations, and its inclusion in liposomes or phospholipid complexes [20][21][22]. A recent review on the topic that was published in 2021 highlighted the urgency of clinical trials to analyze the efficiency of nanoforms of curcumin as well as its derivative analogues [21]. Remarkably, one bio-enhanced derivative of curcumin was been successfully tested in a cell model of a rare variant form of Gaucher disease (GD) caused by mutations in the prosaposin gene (PSAP) [92].
Drug development and approval are multi-step processes that require significant investments and have poor chances of successful results [93]. These limitations imply a significant difficulty in improving new drugs for rare diseases. In this context, drug repositioning-using previously approved drugs for new therapeutic purposes-is a strategy for overcoming some problems related to de novo drug discovery. Repositioning reduces the time "between bench and bedside" while keeping research-related costs low. Most importantly, it reduces the risk of failure from more than 95% to around 45% [94][95][96][97][98][99]. Therefore, drug repurposing is a highly recommended practice in the scientific community, particularly for rare diseases [93].
The results described herein represent a case of nutraceutical repositioning for FD that pave the way to improving personalized therapies. We believe that this approach would be highly beneficial for patients, increase the advantages of PCT, and lead to the broadening of the target audience treatable with oral therapy.
Our proposal for FD is even more promising when considering that combined therapeutic approaches for rare diseases have been proposed [100][101][102][103][104] and that most importantly, one of them was recently approved by the FDA for the treatment of cystic fibrosis [105].
In conclusion, we believe that our results pave a new way toward the improvement of FD therapies, notwithstanding the limitations of our study and the further research needed for the translation to a medical approach.