In FD storage of Gb3 occurs in all cardiac cellular types: myocytes, endothelial and smooth muscle cells of intramyocardial vessels, endocardium, valvular fibroblasts and conduction tissue [16, 17]. Intracellular glycosphingolipids organize in concentric lamellar bodies (zebra bodies) causing engulfment of the cytoplasm and cellular enlargement. Storage in myocytes causes a mechanical impairment, initially detectable only through advanced echocardiographic techniques as subclinical diastolic and systolic dysfunction [18]. Intracellular storage leads to cardiac wall thickening resulting in overt LVH that can mimic HCM. Progression of LVH and worsening of diastolic function can lead to further remodeling, including atrial enlargement and atrial fibrillation [19, 20]. Storage in intramyocardial vessel walls causes structural and functional changes leading to ischemia and ultimately replacement fibrosis with possible progression to systolic dysfunction [20]. Fibrosis and involvement of conduction tissue represent the substrate of ventricular arrhythmias and conduction disturbances [21] (Fig. 1). Valvular involvement, rarely severe, may contribute to disease progression and development of symptoms.

Right ventricular involvement is common in FD and mainly represented by right ventricular hypertrophy and speckle-tracking abnormalities, rarely associated with overt right ventricular dysfunction, at variance with other infiltrative disorders like cardiac amyloidosis with comparable degree of right ventricular wall thickening [22, 23].

Although accumulation of Gb3 in lysosomes represents the main pathophysiological mechanism in FD, growing evidence indicates that secondary pathways of damage, activated by tissue-infiltrating and circulating glycosphingolipids, play a central role in the pathophysiology of the disease (Fig. 2) [24].

Indeed, a significant correlation between duration and intensity of lysoGb3 lifetime exposure and overall disease burden has been reported in both male and female patients with classic FD [25] suggesting that beyond storage, lysoGb3 may also represent a pathogenic factor [26, 27]. In fact, similarly to what has been observed in other inherited glycosphingolipidoses in vitro studies showed that intra-lysosomal Gb3 storage through steric hindrance may disrupt several lysosomal functions, including endocytosis and autophagy, thus interfering with mitochondrial energy production, and triggering apoptosis [28, 29, 30] (Fig. 2).

The autophagy–lysosome pathway (ALP) is an essential recycling pathway regulating cell survival and programmed death. Disruption of this pathway is a common feature of lysosomal storage disorders, including FD [31, 32]. Accordingly, pathology studies on cardiac biopsies from patients with different disease severity revealed increasing rates of cell death with subsequent development of replacement fibrosis [33].

In addition, glycolipid accumulation may induce an oxidative damage of DNA, myofibrils, and mitochondria leading to degradation of contractile elements, reduced ATP synthesis, and cell death [34, 35]. These mechanisms are probably responsible for the increased passive and decreased active forces described in isolated Fabry cardiomyocytes [35].

Steric hindrance is also responsible for the altered mitochondrial function and abnormal energy production processes reported in lysosomal storage disorders. Lücke and colleagues [36] observed a significant reduction of energy metabolism and mitochondrial functions, namely lower activities of respiratory chain enzymes I, IV, and V, in cultured fibroblasts from FD patients. Reductions in adenosine diphosphate ADP, adenosine monophosphate AMP and adenosine triphosphate ATP were also observed. Accordingly, 31(P)-CMR-spectroscopy demonstrated energy depletion of myocardial tissue in FD patients with LVH. Interestingly, in the same study energy metabolism improvement following enzyme-replacement therapy (ERT), preceded LVH regression [37]. Energy depletion coupled with the action of trophic factors like sphingosine, are thought to activate pathways of cellular hypertrophy common to sarcomeric HCM and other phenocopies [38].

Dysfunction of the endoplasmic reticulum represents another consequence of steric hindrance determined by Gb3 lysosomal storage. Consequences of endoplasmic reticulum dysfunction are mainly represented by release of oxidative stress products and mostly by induction of the unfolded protein response observed in cells of FD patients [39].

Unfolded protein response regulates many components of the secretory pathway to restore protein homeostasis, including protein folding, maintenance of calcium, but also inflammasome activation thus representing a well-established pro-inflammatory trigger [40]. Besides interference with intracellular organelles functions, Gb3 storage seems to also affect cell membrane structures. Birket and colleagues [41] demonstrated enhanced function of sodium and calcium channels, resulting in higher and shorter spontaneous action potentials, in FD cardiomyocytes obtained from induced pluripotent stem cells. These findings are in line with similar results on neuronal ion channels, thus suggesting that stored glycosphingolipids may alter ion channel expression and/or cell membrane trafficking, interfering with the electrical properties of cardiomyocytes [42]. Indeed, Namdar and colleagues [43] hypothesized that an increased conduction velocity in atrial and ventricular myocardium may underlie the electrocardiographic abnormalities in FD, including the presence of a short PR interval in the absence of an accessory pathway.

Finally, when considering the complexity of myocardial pathophysiology in FD, it should be emphasized that patients with the same gene variant, even in the same family, may present different severity of clinical manifestations, suggesting that also other genetic factors, like the presence of additional GLA variants, concomitant variants in other genes and environmental factors, may influence the phenotypic expression [44]. With this regard, it’s worth to mention that in patients with an established diagnosis of FD, in the presence of significant LVH in children and adolescents, severe LVH in young adults or evidence of male-to-male transmission of LVH phenotype, the presence of concomitant sarcomeric gene mutations should be ruled out, particularly when response to FD-therapy is scarce or null [45]. Similar secondary mechanisms of damage mediated by Gb3 and lysoGb3 beyond the lysosomal storage, have been demonstrated also in other affected tissues. In arterial vessels, lysoGb3 promotes smooth muscle cell proliferation, causing wall remodeling with increase of intima media thickness and arterial stiffness of small and large arteries, including myocardial and renal vessels [46]. In addition, lysoGb3 at concentrations consistent with levels detectable in FD patients, can impair endothelial nitric oxide synthase, further contributing to the widespread vasculopathy observed in Fabry patients, particularly in classic FD [47, 48].

In kidneys, lysoGb3 is thought to contribute to podocyte loss and glomerulus fibrosis through different mechanisms, triggering apoptosis and inflammatory activation similarly to what described for myocardial tissue [49, 50]. With this regard it should be reminded that both cardiomyocytes and podocytes are terminally differentiated cells with very low, if any, rates of turn-over. A recent proteome analysis study showed that $\alpha$-Gal A-deficient podocytes present dysregulation of proteins involved in lysosomal trafficking and function, metabolism and thermogenesis regulation, cell-cell interactions and cell cycle. Of note treatment with induction of $\alpha$-Gal A expression normalized protein expression only in part. A similar abnormal protein expression was also observed in endothelial and epithelial cells, suggesting a FD-specific rather than a cell-specific intracellular pathways disruption [51].

At neuronal level it has been demonstrated that lysoGb3 at concentrations detectable in plasma of FD males, damages nociceptive neurons, likely accounting for the reported small-fiber neuropathic pain affecting classic FD patients [52]. Similarly, duration of exposure to lysoGb3 significantly correlated with thermal sensory limen and the cold detection threshold of hand and arms [53].

Increasing evidence supports a role of inflammation in early pathogenesis and progression of cardiac damage [50, 54]. Accumulation of Gb3 and lyso-Gb3 can trigger a chronic inflammatory response either promoting the release of—or behaving themselves as—damage mediators [55] (Fig. 1). Altered peptides derived from oxidative stress or from abnormal enzymes digestion may act as neoantigens activating the immune system. Unfolded protein response, as previously mentioned is also a potent trigger of inflammatory response. Glycolipids can act themselves as antigens when presented to natural killer T cells. The Gb3-mediated effects can be abolished by antibodies blocking the toll-like receptor 4, demonstrating a pivotal role of this inflammatory pathway [50, 54]. Activation of toll-like receptor 4 pathway may also enhance TGF-$\beta$ response leading to remodeling of the extracellular matrix and myocardial fibrosis [49].

Endomyocardial biopsy studies showed the presence of myocarditis in up to 56% of patients with FD cardiomyopathy [56]. In the presence of positive antiheart and antimyosin antibodies and negative polymerase chain reaction for viral genomes, an immune mediated process was supposed and a significant correlation with disease severity was observed suggesting that myocardial inflammation may contribute to progression of FD cardiac damage and resistance to ERT.

In addition, signs of systemic and myocardial chronic inflammatory activation are consistently detected in patients with FD [57].

The widespread use of new cardiac magnetic resonance (CMR) techniques, namely T1 and T2 mapping, further supported the view of FD as a storage/inflammatory disease providing a qualitative assessment of myocardial tissue in terms of myocardial lipid content and inflammation.

Systematic evaluation of patients with different gender, age and type of disease (classic or late-onset) led to identify sequential phases of cardiac damage evolution [58, 59, 60] suggesting a three-stage progression of Fabry cardiomyopathy: an initial phase, starting since childhood, characterized by myocardial storage without signs of inflammation nor overt LVH. In the second phase, secondary pathways particularly inflammation and hypertrophy take place leading to clinical manifestations and ECG and imaging abnormalities. In this phase T1 lowering, indicative of myocardial glycolipid storage and T2 elevation indicative of myocardial edema/inflammation, may precede LVH, mostly in female patients. Markers of this phase, characterized by myocardial inflammation and injury, are the release of troponin and an initial increase of NT-proBNP. In the absence of specific therapy, cardiac damage progresses towards severe LVH and fibrosis, precipitated by concomitant mechanisms, in particular myocardial ischemia. Higher NT-proBNP levels are typical of this phase, characterized by overt symptoms and poor response to FD-specific therapies. Myocardial perfusion impairment is another early feature of cardiac involvement, likely reflecting microcirculation dysfunction induced by storage and inflammation. Whether myocardial inflammation is the main mechanism leading to myocardial fibrosis remains to be fully elucidated. It remains also unclear whether secondary pathways triggered by Gb3 storage, including systemic and tissue inflammation, may become storage-independent and thus resistant to FD-specific therapies aimed at halting or preventing further Gb3 storage.

Currently available FD-specific treatments include ERT (agalsidase alfa and agalsidase beta) and an oral pharmacological chaperone (migalastat). Both therapies are indicated in patients with an established diagnosis of FD, with the pharmacological chaperone indicated only in those patients with an amenable variant.

Long-term follow-up studies and registry data demonstrated that ERT may halt or slow progression of cardiac disease reducing the rate of cardiovascular events, particularly when started early [2, 15]. Regression of mild LVH has been reported in patients with the classic and cardiac phenotype, with some evidence that LVH may be prevented by early treatment [61]. In a recent study ERT-naïve patients presented attenuated T1 lowering, with small reductions in maximum wall thickness and stabilized LVMI after 1 year of ERT [63].

When cardiac involvement is advanced, response to ERT is limited [61, 63] and there is evidence that despite 1-year-ERT an increase of T2 signal and LGE area with worsening global longitudinal strain can be observed at CMR.

Endomyocardial biopsy studies have shown that ERT reduces endothelial Gb3 inclusions in myocardial vessels, while clearance of Gb3 from cardiomyocytes is less significant [64]. The development of neutralizing antibodies binding exogenous enzyme molecules and preventing them to reach target cells, is among the possible causes for a limited efficacy of ERT in some patients [65].

Chaperone molecules are iminosugars that in amenable variants of FD bind to the catalytic domain of $\alpha$-Gal A promoting its proper folding and trafficking to the lysosome, thus increasing enzymatic activity and Gb3 degradation. The same molecules at higher doses exert an inhibitory action on $\alpha$-Gal A. Amenable GLA variants are defined by an increase by at least 20% of enzymatic activity in patients’ lymphocytes cultured with 20 mM Migalastat.

A list of GLA variants resulted amenable in vitro is available online in a public online database. Clinical trials and open-label extension studies have shown that treatment with Migalastat is associated with a small but significant decrease in indexed left ventricular mass assessed by echocardiography [66]. However, no CMR data on Migalastat effects on myocardial damage are currently available. In addition, recent real-world studies demonstrated that in some genetic variants a significant discrepancy between predicted in vitro amenability and the effective in vivo increase in $\alpha$-Gal A activity and clinical response can be observed [67]. These pitfalls of Migalastat treatment may recognize several causes including the possible dosage-dependent inhibitory effects of Migalastat and the intrinsic limitations of the in vitro amenability test. Accordingly, a careful monitoring of the biochemical and clinical response to chaperone therapy is required when treatment is initiated to confirm clinical efficacy.

New therapies under development are represented by second-generation ERTs, substrate reduction therapies and gene and mRNA therapies [5, 68].

New generation ERTs include plant-derived ERTs developed to reduce ADAs formation and improve enzyme biodistribution. Pegunigalsidase alpha is a novel pegylated form of $\alpha$-Gal A produced in a PlantCell Ex system (Protalix Biotherapeutics, Carmiel, Israel) characterized by a much longer circulatory half-life and increased heart and kidney uptake compared to currently available ERTs.

In the phase III BRIDGE trial (NCT03018730) patients switched from agalsidase alfa to Pegunigalsidase alfa showed stabilization, or at least slower progression of kidney failure (eGFR slope improved from –5.1 to 0.23 mL/min/1.73 m${}^2$/year in both male and female) [69]. Other clinical trials (NCT02795676; NCT03018730, NCT03180840) evaluating Pegunigalsidase alfa treatment are still ongoing,

Substrate reduction therapy (SRT) is represented by oral iminosugars blocking the glucosylceramide synthase enzyme thus inhibiting glycosphingolipid synthesis to lower the cellular amount of Gb3. This therapeutic approach, already validated in Gaucher disease, can be administered irrespective of genotype. There are two SRT formulations, venglustat and lucerastat, currently under investigation in phase II and III clinical trials respectively [70].

Both novel SRT agents are promising potential oral therapeutics for the nearby future with no limitation regarding specific mutations as seen in chaperone therapy.

In phase I/II clinical trials treatment with lucerastat led to a significant reduction of glycosphingolipids, glucosylceramide, lactosylceramide, and globotriaosylceramide plasma levels compared to baseline, with a good safety and tolerance profile [71]. A randomized multi-center double-blind clinical phase III study with lucerastat is currently ongoing.

Genetic therapy is considered the definitive treatment for many genetic disorders including lysosomal storage diseases. In the last decades, both in vivo and ex vivo gene therapy approaches have been explored. In a recent phase II clinical trial, hematopoietic stem cells were retrieved from patients, transfected with lentiviruses (AVR-RD-01, Avrobio) and re-administered to the patient showing a persistent elevation in $\alpha$-Gal A activity [72]. However, despite encouraging safety and efficacy results this research line has been recently deprioritized by the pharmaceutical company.

In vivo approaches with liver targeted adenoviral-mediated gene transfer showed in preclinical studies with $\alpha$-Gal A knockout mouse model, a significant increase of $\alpha$-Gal A activity associated with a marked reduction of lyso-Gb3 [73]. Regarding this approach, it remains unclear whether the uptake of enzyme released by transfected cells by affected tissues will be sufficient to compensate the enzymatic deficit. Indeed, when considering the case of heterozygous females, cross-correction doesn’t seem able to restore adequate enzymatic activity and prevent Gb3 storage. In addition, it is also possible that male patients with classic FD and absent a-Gal A activity could develop ADAs against the expressed enzyme, as observed for ERT, although given the continuous exposure and endogenous synthesis and glycosylation, tolerance towards the new antigen could be expected in most of cases. To specifically target myocardial tissue, novel cardiac-tropic vectors have been developed and are currently tested in non-human primates. Initial studies suggest an increased gene delivery and a reduced immunogenicity compared with conventional adenoviral vectors. Very recent data in adult male patients with classic FD showed safety and efficacy of the vectors with evidence of biochemical efficacy (sustained increase of $\alpha$-Gal A activity and significant decrease of lyso-Gb3 levels), together with promising cardiac improvement in terms of T1 mapping increase at CMR.

The administration of human $\alpha$-GAL A mRNA encapsulated with lipid nanoparticles, has been also tested in mice and non-human primates showing a significant increase of $\alpha$-GAL A levels in liver, heart, and kidney [74].

Cell transplantation represents another potential strategy to treat lysosomal diseases. In a recent study a new ex vivo gene therapy platform was developed using a transplant pack, consisting of a porous membrane and spheroids with scaffolds. These membranes have countless pores of less than 0.1 mm${}^2$ allowing secretion of proteins keeping them separated from the host immune system. The packs with cells overexpressing $\alpha$-Gal A were subcutaneously transplanted into the backs of mice, leading to high levels of enzyme in plasma and livers. Compared to classic ERT this technique has several advantages including continuous enzyme secretion and potentially a one-time treatment to cure diseases [75].

In general, while initial experience and short-term results, are promising, the long-term efficacy and potential adverse effects of gene therapy and cellular therapies remain unclear and require larger studies with longer follow-up.

Cardiovascular complications are the first cause of morbidity and mortality in patients with FD. Therefore, conventional cardiovascular therapies including pharmacological and interventional strategies are essential to improve survival and quality of life in these patients. In a recent consensus document, expert recommendations have been provided regarding treatment and follow-up [76]. It is important to emphasize that FD does not represent a contraindication to conventional invasive therapies including percutaneous and surgical myocardial revascularization procedures or pacemaker and defibrillator implant. Similarly, interventional structural cardiology and radiofrequency ablation can be considered in FD patients with valvular disease or life-threatening arrhythmias, while patients with advanced heart failure can be proposed for cardiac transplant.

With this regard, sudden cardiac death prevention remains challenging in FD, particularly in patients with advanced cardiac involvement. Current recommendations for HCM, including the risk calculator developed by European Society of Cardiology, cannot be applied. Therefore, prognostic stratification relies on the identification of risk factors including advanced LV hypertrophy, extensive fibrosis, and unexplained syncope [77].

Similarly, considering the increased risk of stroke and also of cerebral microbleeds associated with FD, when considering stroke prevention strategies in patients with atrial fibrillation, classical scoring systems (i.e., CHA2DS2VASC and HAS-BLED) should not be applied. As patients with HCM, all patients with FD and atrial fibrillation should be considered candidates to oral anticoagulation. Although there are no studies assessing safety and efficacy of direct oral anticoagulants in FD, these drugs should be preferred (when not contraindicated for severe impairment of renal function) considering the augmented risk of intracranial hemorrhage and nephropathy associated with vitamin-K antagonists administration. It is also important to remind that amiodarone is a cationic amphiphilic molecule that may further worsen lysosomal pH and functions, potentially reducing the effect of ERT. Therefore, long-term treatment with this drug, either for ventricular arrhythmias or rhythm control in atrial fibrillation should be avoided or limited to selected cases with close monitoring of ERT effects [76].