Current and experimental therapeutics for Fabry disease

Abstract Fabry (or Anderson‐Fabry) is a rare pan‐ethnic disease affecting males and females. Fabry is an X‐linked lysosomal storage disease, affecting glycosphingolipid metabolism, that is caused by mutations of the GLA gene that codes for α‐galactosidase A. Fabry disease (FD) can be classified into a severe, classical phenotype, most often seen in men with no residual enzyme activity, that usually appear before 18 years and a usually milder, nonclassical (later‐onset) phenotype that usually appear above 18 years. Affected patients show multifactorial complications, including renal failure, cardiovascular problems, and neuropathy. In this review, we briefly report the clinical trials so far performed with the available therapies, and then we focus on the in vitro and the in vivo experimental models of the disease, to highlight the relevance in improving the existing therapeutics and understand the mechanism of this rare disorder. Current available in vivo and in vitro models can assist in better comprehension of the pathogenesis and underlying mechanisms of FD, thus the existing therapeutic approaches can be optimized, and new options can be developed.


| INTRODUCTION
Fabry disease (FD, OMIM #301500) is an X-linked lysosomal storage disorder, involving glycosphingolipid metabolism. FD is due to a wide range of mutations in the GLA gene on the X chromosome (Xq22.1), resultant in a defect of the lysosomal enzyme α-galactosidase A (α-gal A). 1 This induces to gradual deposits of globotriaosylceramide (Gb3) in cells in whole body, producing multi-systemic impacts, including serious and progressive cardiac and renal impairment 2 (Figure 1). Gb3 and its metabolite, lyso-Gb3, are crucial in the pathogenesis of FD. Lyso-Gb3 is produced by the deacylation of Gb3 by acid ceramidase. 3 Lyso-Gb3 accumulation worsens the disease pathology both reducing α-gal A activity and promoting smooth muscle cells proliferation, 4 an element that may promote the increased intima-media thickness observed in Fabry patients. 5 Lyso-Gb3 destroys nociceptive neurons, leading to kidney fibrosis and inhibiting endothelial nitric oxide synthase. 6 Lyso-Gb3 is strongly increased in plasma of classical FD male patients as well as in tissues and plasma of mice, but it does not correlate with the severity of the disease and cannot represent a valid surrogate biomarker. 7,8 α-gal A deficiency induces the activation of numerous pathways and impacts the onset of disease symptoms. Downstream effects, including inflammation, fibrosis, and the reactive oxygen species production also appear to exert a role in the pathogenesis. 9,10 Fabry is a multisystemic disorder that initiates at cellular level (disruption of basic metabolic processes and a cascade of events), followed by a cascade of organ dysfunctions and structural alterations that ultimately progresses over years or decades. The age of onset is widely variable as well as clinical manifestations and progression. The primary disease begins as early as the fetal phase of development. 11 FD can be classified into a severe, classical phenotype, most often seen in men with no residual enzyme activity, that usually appear before 18 years; and a usually milder, nonclassical (later-onset) phenotype that usually appear above 18 years. 12 The first clinical manifestations in FD classical phenotype occur typically between 3 and 10 years of age with features in boys more severely than girls. 13,14 The most frequent sign in classical Fabry hemizygous males, with no residual α-gal A activity, are cornea verticillata, neuropathic pain, cardiomyopathy, strokes (cerebrovascular), angiokeratoma, proteinuria, arrhythmia, cochleo-vestibular, renal and gastrointestinal disorders. 15 Renal, cerebrovascular and cardiac disorders are manifested after the age of 20 years.
In later-onset FD patients, the symptoms are milder or may be limited to one organ since they have residual α-gal A activity (5-25% of normal) and have a later onset (average age: 40-60 years). 12,16 Long-term clinical manifestations comprise progressive renal failure, stroke, and hypertrophic cardiomyopathy. 12 Women often present signs and symptoms of FD, although less severe compared with men. 17 Fortunately, the knowledge about FD natural history and potential therapeutic approaches are unceasingly evolving and consequently the recommended testing, treatment procedures and monitoring need to be revised.
In this review, we report the current and experimental therapeutic approaches for FD. In addition, we focus on the in vivo and in vitro models of the disease, which are valuable for dissecting the pathogenesis and mechanism of FD and for optimizing the existing therapeutics or developing new options.

| THERAPEUTIC APPROACHES
FD therapies are mainly based on intravenous replacement of the altered enzyme using agalsidase α or β. However, recently, therapeutic approaches for some patients affected by FD have been extended ( Figure 2).

| Enzyme-based treatments
One of the first therapeutic approach for FD was enzyme replacement therapy (ERT). The treatment successfully improved patient quality of life and alleviating kidney failure, but unmet clinical needs persist. 18 Ex vivo and in vivo approaches are necessary to increase the enzyme duration in plasma and ameliorate the delivery. 19 ERT with recombinant α-galactosidase was approved in Europe in  19,20 In USA, only Fabrazyme is licensed.
Although ERT has been in clinical use since 2001, the optimal dosage, treatment target and beginning of the treatment need to be elucidate. 1,21 This is relevant because ERT is costly and is a lifelong obligation for FD patients. 21 Since it is a rare disorder, it results difficult to analyze the data with meta-analyses and metasyntheses.
It has been demonstrated that short-term treatment with Replagal reduced plasma Gb3 levels by 50%. 22 Also, patients on dialysis or who had received a transplant showed comparable level clearance, in particular, plasma Gb3 levels decreased by 43% after 27 weeks of treatment. 23 Variations in Gb3 levels have been investigated in several pediatric trials. [24][25][26] In 2001, Schiffmann and collaborators published the first randomized, double-blind, placebo-controlled Phase II/III trial with Replagal. 27 Notably, this trial showed from primary end points significant statistically and clinically amelioration of neuropathic pain. In the secondary end points, the patients showed great improvement in creatinine clearance and cardiac conduction, implying that the therapeutic effects of the drug are prevalent. 27 Numerous other randomized, F I G U R E 1 Schematic representation of the pathogenic mechanism that occurs in Fabry disease. GB3, globotriaosylceramide; GLA, Alpha-Galactosidase A controlled trials and open-label studies validated the safety and efficacy of Replagal. [28][29][30][31] Clinical trials reported that Fabrazyme induces the clearance of Gb3 in glomerular and mesangial endothelial cells. 19,32 Additionally, the treatment can reduce pain and critical clinical events also in aged patients. 33 Furthermore, Fabrazyme treatment was able to ameliorate vascular functions in FD patient from the first infusion and then maintained for the whole period of observation (1 year). 34 The aspects of ERT are substantially reviewed in. [35][36][37][38] Overall, Replagal and Fabrazyme share comparable biochemical and structural characteristics, even though some physicians indicated the recommended dosage of Fabrazyme five times greater. The switch to Replagal resulted well tolerated. 39,40 In contrast, a cohort study reported that Fabrazyme treatment led to improved heart and biochemical amelioration (even in the presence of antibodies) and a greater decrease in plasma lyso-Gb3 levels in men with classical FD compared to Replagal. 20 However, the two groups showed no differences in GFR and clinical event rate. 20 ERT is usually well tolerated, however may appear reactions due to the infusion, which lead to dyspnea, rash or hyperpyrexia and rarely anaphylaxis. 19,41 Premedication of steroid or diphenhydramine are used to diminish these symptoms. 19 Moreover, the infusion rate is a key issue (faster rates are linked to higher intolerance) but also the frequency of the infusions. Classical FD male patients should start ERT promptly, with adjunct therapeutic approaches if necessary. Adult women with classic mutation should initiate ERT if the symptoms involve major organs or, if asymptomatic, when the histological, imaging or laboratory investigations reveal damage in the primary organs. 42

| Investigational Exogenous Enzyme Replacement Therapies
Among the novel ERT, α-pegunigalsidase is a pegylated form of α-galactosidase generated in a PlantCell Ex system. The use of plantbased compounds is rapidly increasing in the area of medicine, since it is more efficient at lower costs. 43 In addition, preclinical data showed that the vascular half-life of plant-based biologics is higher than existing ERT. It has been reported reduced uptake by the liver and improved uptake by kidney and heart when compared to presently available ERT preparations that are characterized by terminal mannose-6-phosphate glycosylation residues. 44 Existing ERT therapies showed 2 hours of half-life, while the plant-based approach showed 53-121 hours of half-life. Moreover, the drug is well tolerated with mild adverse events. 44 Notably, during the 9 months of treatment, most of the patients showed a substantial decrease in Gb3 accumulation. 45 Clinical trials tried to compare this approach to αand β-galactosidase to understand the potential of pegunigalsidase. An open-label switchover study is ongoing (NCT03180840), to assess the pharmacokinetics of α-pegunigalsidase. 46

| Gene-based therapy
Gene-based therapy represents a potential approach for different rare genetic disorders, among which FD. For the gene therapy in vivo, a vector is infused into the FD patient and then the cells, including liver F I G U R E 2 Graph on the current and investigational therapeutic approaches for Fabry disease with the relative mechanism of action (MoA) and the current clinical phase. FD, Fabry disease; ERT, enzyme replacement therapy; RNA, ribonucleic acid cells, promptly undertake gene editing to express the missing protein. 47 A recent approach is based on the adeno-associated virus (AAV)mediated gene transfer infusion in order to enhance the enzyme level.
Studies on α-galactosidase A knockout (GLAko) mouse model demonstrated that this approach was able to induce the production of α-galactosidase by the liver in a dose-dependent manner. 48 A comparable method utilizing another vector was able to strongly increase α-galactosidase levels after single-stranded AAV8 vector administration and lyso-Gb3 and Gb-3 levels decreased comparable to wild-type animals. 49 In the ex vivo method, hematopoietic stem cells from the patient are harvested, underwent gene editing and subsequently are engrafted back to the patient after myeloablative therapy. To date, there are different ex vivo gene therapy approaches for FD treatment.
Huang and collaborators reported that CD34-positive hematopoietic stem cells, modified using recombinant lentivirus-mediated gene infused into autologous recipients, showed good engraftment and continuous α-galactosidase production at 1 and 2 years. 19 In a phase II trial using ex vivo gene therapy was able to increase α-gal A activity in 2 patients. 50 Recently, to increase the delivery of human α-gal protein in vivo nanoparticle-formulated mRNA in mouse and nonhuman primate were used, with no necessity for myeloablative therapy or viral vectors administration. In particular, the mRNA for hGLA was encapsulated with lipid nanoparticles and was able to increase α-galactosidase expression in cardiac, liver, and kidney tissues, causing increased Gb-3 clearance. 49 Gene therapy approaches have the common aim to increase α-galactosidase enzyme activity. Before adopting gene therapy for FD, sufficient α-gal A activity and stable viral copy number should be considered compared to current ERT.

| Nonenzyme replacement approaches
To reduce Gb3 deposits, nonenzyme replacement approaches are also used and comprise the inhibition of glucosylceramide synthase. Chaperone therapy is able to stabilize and increase the endogenous enzyme activity and is now approved for a specific group of FD mutations. These mutations that produce a form of α -Gal A which responds to chaperone binding with a relevant increase in function are known as amenable mutations.
Chemical chaperones are able to attach to the defective enzyme, leading to its proper folding, maturation and trafficking to the functional site. 51 A recent approved pharmacologic chaperone is the low molecular weight iminosaccharide Migalastat. Migalastat (Galafold™) stabilizes and improves trafficking of amenable mutations of α-gal A enzyme from the endoplasmic reticulum to lysosomes and increases its lysosomal activity. 52,53 It is administered orally showing a significant distribution and it is able to cross the blood brain barrier. The randomized trial ATTRACT intended to study in FD patients the switch from ERT to Migalastat.
Migalastat was able to maintain at a low-level plasma globotriaosylsphingosine, and, interestingly, induced a significant decrement in left ventricular mass index compared to ERT patients.
Regarding side effects, headache, nausea, urinary infection and pyrexia were the most frequent. 54 In 2018, the Food and Drug Administration approved Migalastat basing on FACETS, a randomized, double-blind, placebo-controlled phase III study. 52 35 to 50% FD amenable mutations can be treated with Migalastat, however, the in vitro assays reported discordant results in the effectiveness of this treatment for some mutations and additional investigations are needed. 54 Migalastat is not indicated for use in FD patients with nonamenable GLA mutations or with severe renal failure. 55 Another iminosaccharide used as FD treatment is Lucerastat (Idorsia Pharmaceuticals), which functions as inhibitor, since it is able to avoid Gb3 deposits, reducing the amount of ceramide converted in glycosphingolipid. 56 Lucerastat is currently under investigation in a phase 3 trial (NCT03425539) in order to evaluate the safety and efficacy in FD patients as oral monotherapy.

Another substrate reduction inhibitor is Venglustat by Sanofi
Genzyme; it is under phase 2 clinical trial (NCT02489344) to determine its effectiveness in male FD patients. 19

| Adjunctive Therapies
For FD patients is necessary to also follow the standards of care for patients affected by cardiomyopathy and chronic renal disorder. To contain the renal failure, angiotensin receptor blockers and angiotensin-converting enzyme inhibitors resulted valid treatment for FD patients, concomitant with a low-sodium diet. 57 FD is a risk factor for stroke and statins therapy should be take into account. 58 A study 59 reported the effectiveness of ERT and antiproteinuric therapy with ACE inhibitors and/or angiotensin II receptor blockers in patients with severe Fabry nephropathy. Patients showed preservation of kidney function if Fabrazyme treatment was initiated at a younger age, and urine protein to creatinine ratio was maintained at or below 0.5 g/g with both antiproteinuric therapies. 59

| In vivo models
An animal model of this disorder would be useful to investigate beneficial approaches for patients with FD as well as for the study of the underlying molecular pathophysiology ( Table 1).
The first model reported was in 1997; Oshima and collaborators developed a mouse model for FD by disrupting α-gal A gene by homologous recombination, then the mouse α-gal A gene was isolated and characterized. 56 In particular, α-gal A targeting vector was designed to replace exon 3 and intron 3 of α-Gal gene with a neomycin resistance cassette. The construct was electroporated into 129S4/ SvJae-derived J1 embryonic stem cells. Accurately targeted embryonic stem cells were then inoculated into blastocysts and the resultant chimeric males were bred to C57BL/6 females. These mice showed undetectable α-Gal enzyme activity in cultured fibroblasts as well as the tissues, thus indicating an effective disruption of the α-gal A gene. 64 These α-gal A-deficient mice created by gene targeting are clinically normal at 10 weeks of age but showed globotriaosylceramide deposits and inclusions in the kidneys, closer to the situation that usually occur in clinic. The correction in α-gal A-deficient fibroblasts, using multi drug resistance retroviruses containing the lacking enzyme, is relevant for translating in vivo the correction of the α-gal A deficiency. 56,64 Experimental strategies to produce the clinical phenotype in these mice result useful for the advancement of efficient therapeutic approaches for FD patients.
Successively, the same research group used aged Fabry mice (80 weeks old) and they noticed that the animals did not show clinical sign or organ malfunction. 65 Another research group generated α-gal AKO mouse and no organ malfunction in aged animals was reported. 60 Other investigations described neuronal glycosphingolipid accumulation and somatosensory phenotypes modified in Fabry mice, thus revealing neuropathic pain. However, these findings are conflicting, [61][62][63] probably due to the unavailability of wild type control mice that may concur to the divergences. A theory supporting the asymptomatic Fabry animal is that glycosphingolipid profiles differ between humans and mice. Indeed, the level of Gb3, Gb4, lactosylceramide, red blood cell globosides are lower in mice than humans. 65 Another group established a transgenic mouse (TgG3S) expressing the human α1,4-galactosyltransferase (Gb3 synthase). The purpose of the study was to elevate Gb3 levels in mouse organs (in particular in the heart), which allow to evaluate the active-sitespecific chaperone therapy for FD. 66 Further, it has been generated a symptomatic mouse model by crossbreeding GLAko mice with transgenic mice TgG3S. TgG3S/ GLAko mice showed elevated Gb3 levels in the primary organs and progressive renal failure accompanied by the typical clinical conditions. Upon recombinant α-gal A administration, the urine volume and albumin concentration were strongly decreased in TgG3S/GLAko mice. These data reported that Gb3 storage is a leading pathogenic aspect in the symptomatic phenotype of TgG3S/GLAko mice, and that this model may be appropriate for preclinical studies. 67 Unfortunately, the developed FD mouse models did not shown the typical clinical symptoms, thus another lab established a Fabry rat model using CRISPR/Cas9 technology to delete the GLA gene, to see if larger animals showed a phenotype that better recapitulates the FD clinical symptoms, in particular the neuropathic pain. 68 These rats accumulated glycosphingolipids in all tissues and were completely deficient of α-gal A activity. Notably, they developed neuropathic pain as demonstrated from the altered cation channel and altered N-glycan processing within the Golgi, likely due to substrate deposits within the membranes of this organelle. 68 Recently, the same research group demonstrated that Fabry rats showed also ocular manifestations 69 and cardiorenal phenotypes, 70 including renal tubule impairment and mitral valve thickening.
Further, the same research team reported hearing problems in Fabry aged rats, but additional studies are ongoing. Overall, this symptomatic rat model is a valuable addition to the current Fabry mouse models. 71

| In vitro models
Complementary in vitro models for testing potential therapeutic approaches and to understand the underlying mechanisms are really helpful ( Table 2). An in vitro model of FD was developed from α-gal A KO mouse, in particular, primary cultures of aortic endothelial cells were generated, characterized and maintained in culture. Further, these primary cultures were used to compare the effects of recombinant α-gal A and a strong glucosylceramide synthase inhibitor on Gb3 metabolism (ethylenedioxyphenyl-P4). 72 The main limit in using primary endothelial cells is the limited lifespan, thus Shen and collaborators produced an endothelial cell line from a Fabry hemizygote patient, introducing human telomerase reverse transcriptase genes. The generated cell line showed prolonged lifespan and express different key markers of endothelial cells, while the activity of α-gal A was strongly decreased compared with primary endothelial cells from normal individuals. 74 To study the renal damage in FD, it was necessary to develop a human podocyte model, combining ribonucleic acid (RNA)  tions. Notably, they reported potential new cellular and secreted protein biomarkers. In example, the increase of LIMP-2 (large inhibitor of metalloproteinases 2), a lysosomal protein involved in heart disorders, due to α-gal A deficiency, could induce the protein secretion, suggesting its essential role in FD pathology. 78 Another recent paper reported the use of Jurkat cells, a T-lymphoblastic leukemia cell-line, as FD model. 79  It would be interesting for future investigations to study the efficacy of the so far developed therapies in combination (not only as solo agent). Another important point to reexamine is the optimal time to begin a specific therapy considering the family history, the onset, the FD variant and the sex. The current available in vivo and in vitro models are valuable for a better understanding of the pathogenesis and underlying mechanism of FD. These models should be made available for a larger audience in order to study this rare disease, therefore, the available therapeutics can be improved, and new approaches can be established.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

FUNDING
No funding source.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/cge.13999.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.