Lysosomal storage disorders (LSDs) are a heterogeneous group of inherited diseases with a collective frequency of ∼1 in 7000 births, resulting from the deficiency in one or more enzymes or transporters that normally reside within the lysosomes. Pathology results from the progressive accumulation of uncleaved lipids, glycoproteins and/or glycosaminoglycans in the lysosomes and secondary damages that affect the brain, viscera, bones and connective tissues. Most treatment modalities developed for LSD, including gene therapy (GT), are based on the lysosome-specific cross-correction mechanism, by which close proximity of normal cells leads to the correction of the biochemical consequences of enzymatic deficiency within the neighboring cells. Here, GT efforts addressing these disorders are reviewed with an up-to-date discussion of their impact on the LSD disease phenotype in animal models and patients.
Lysosomal storage disorders
Lysosomal storage disorders (LSDs) are a heterogeneous group of inherited diseases resulting from the deficiency in one or more enzymes or transporters that normally reside within the lysosomes (1). Lysosomes are intracellular organelles delimited by a membrane and contain hydrolytic enzymes synthetized in the endoplasmic reticulum. Once produced, these enzymes are recognized and specifically guided to lysosomes thanks to the interaction between their unique marker mannose-6-phosphate (M6P) and its specific receptor. Part of the newly synthetized enzyme, instead of binding to the intracellular M6P receptor (M6PR), is secreted and can be specifically recognized and bound by extracellular M6P receptors present either on the same cell membrane or on neighboring cells, giving rise to a phenomenon known as cross-correction (2,3). Lysosomal enzyme deficiency is responsible for a progressive accumulation of uncleaved lipids, glycoproteins and/or glycosaminoglycans that lead to secondary accumulation of other macromolecules, organelle and cellular dysfunction, alterations of cellular morphology, impaired autophagy, oxidative stress and neuroinflammation and impaired function of organs and tissues (4,5). Storage and secondary damages may affect the brain, viscera, bones and connective tissues.
Individually LSDs are rare and affect between 1:25 000 to <1:250 000 live births, but they have a collective frequency of ∼1 in 7000 births, which makes them an important public health problem worldwide. To date, mutations in more than 50 different proteins were identified as causing LSDs, and this list continues to grow.
Despite sharing similar pathogenic mechanism, LSDs differ for several disease-specific features, such as the different pattern of visceral organ involvement and the presence and severity of nervous system involvement. LSDs are usually classified according to the nature of the primary stored material, and broad categories include sphingolipidoses, mucopolysaccharidoses (MPSs), mucolipidoses, glycoproteinoses, oligosaccharidoses and glycogen storage diseases. Alternatively, LSDs can also be classified according to the type of protein deficiency, i.e. lysosomal hydrolases, transmembrane proteins, co-factors or co-activators required for lysosomal enzyme function, proteins protecting lysosomal enzymes, proteins involved in post-translational processing of lysosomal enzymes, enzymes involved in targeting mechanisms for protein localization to the lysosome and proteins involved in intracellular trafficking. Each LSD has a distinct clinical and pathological picture, which is determined by the nature of the accumulating substrate and the cell types in which it accumulates. Moreover, most LSDs present with different clinical variants and are also classified based on the patients' age at symptom onset—infantile, juvenile and adult forms are typically identified. The most severe infantile forms frequently present with brain pathology: affected individuals generally appear normal at birth, but symptoms appear soon after birth; neuropathology is progressive and ultimately leads to death at an early age. In adult forms, instead, disability results mainly from peripheral symptoms, and the progression is slower. Juvenile forms are intermediate between infantile and adult forms. Importantly, residual enzyme activity, which usually does not exceed 5% and can be associated to mutations that do not completely abolish folding, processing and catalytic activity of the protein, can account for some of the observed phenotypic variability in LSDs. As general rule, the lower the residual activity, the earlier the age at symptom onset and the more severe the disease phenotype appears. However, patients with a similar genetic background, and sometimes with the same mutation, can show different clinical symptoms and progression, likely because of modifying genes and environmental factors that can influence the clinical disease presentation and course.
LSD treatment modalities
Most treatment modalities known for LSD are based on the lysosome-specific ‘cross-correction’ mechanism described above, by which close proximity of normal cells leads to the correction of the biochemical consequences of enzymatic deficiency within the neighboring cells.
Enzyme replacement therapy
Enzyme replacement therapy (ERT), consisting of the intravenous (i.v.) administration of the recombinant functional enzyme, has been applied to many LSDs like Pompe disease, type I Gaucher disease, Fabry disease and MPS-I and MPS-II (6). Despite the fact that this is one of the most successful treatments available for LSD patients, significant issues and limitations still remain to be addressed. Among them is enzyme targeting to defective cells. Different cell-surface receptors, such as M6PR, are required for successful uptake of lysosomal enzymes; diversity in the membrane density of these receptors in different cells and tissues has been observed. M6PR concentration was found very high in the heart and kidneys, whereas in the muscles and brain its concentration is very low (7). Therefore, since successful ERT requires targeting of multiple cell types, the ideal drug may be one that includes enzymes with various sugar residues and isoforms, in order to take advantage of the many cellular receptors involved in endocytosis. Eukaryotic expression systems were developed to carry out the appropriate post-translational modifications of the enzymes, primarily the generation of M6P residues. Recombinant enzymes are currently obtained from cultures of overexpressing Chinese hamster ovary cells or human fibroblasts.
Although systemic manifestations improve with ERT, this treatment is generally poorly effective on central nervous system (CNS) disease manifestations due to the inability of most lysosomal enzyme to efficiently cross the blood–brain barrier (BBB) (8). Preclinical studies are currently directed at promoting strategies to allow the transport of the administered enzyme across the BBB, such as the conjugation of the enzyme with molecules recognized by a specific BBB carrier or direct enzyme delivery into the cerebroventricular space (9–13). Moreover, immune responses against the injected protein frequently occur in ERT-treated patients. As a consequence, alteration of the pharmacokinetic, hypersensitivity reactions and very occasionally neutralization of the administered enzyme were observed (14,15). Finally, the small number of patients affected by individual LSDs makes the pharmacological development of this therapeutic approach likely to be unprofitable to produce and test the protein for most of these disorders.
Overall, the high cost of the recombinant enzymes, difficulties with patients' compliance because of the requirement for life-long treatment and the frequency of injections, the development of antibodies and the difficulty for the enzyme to efficiently cross the BBB are significant issues for many patients and their families.
Substrate reduction and enzyme enhancement therapies
Substrate reduction therapy (SRT) and enzyme enhancement therapy (EET) act at a cellular level directly in the biological pathway affected by the disease (16,17). Through the use of molecules that restrain biosynthesis of metabolites upstream of the deficient catabolic pathway, SRT partially inhibits the synthesis of the substrate of the defective hydrolase in order to reduce substrate influx into the catabolically compromised lysosome, limiting storage. Instead EET, also known as pharmacological chaperone therapy, can enhance the residual activity of the defective lysosomal enzyme thanks to its stabilization. Both these strategies hold great potential since they should have a better biodistribution than recombinant enzyme. Indeed, they employ low-molecular-weight molecules that can be taken orally and cross the BBB. Regarding SRT, one drug is already approved by Food and Drug Administration for use—N-butyl-deoxynojirimycin (Miglustat®) represents a potential treatment for a variety of LSDs, including Gaucher's and Fabry's diseases. Both preclinical and clinical results showed significant improvement in all major efficacy end points and indicated that treatment was increasingly effective with time (18–21). However, benefit of SRT is less evident in neurological storage disorders (22). Other SRT molecules with greater specificity and improved delivery to the brain are now being developed.
Differently from SRT, EET appears to be mutation dependent, as only diseases caused by mutations impairing enzyme activity or catalytic site are expected to benefit from EET. Various chemical and pharmacological chaperones such as substrate analogs, active-site inhibitors, cofactors or effector molecules have already been tested in research settings (17,21,23). Potential pharmacological chaperones may be already commercially available, having been licensed for other indications. N-(n-nonyl)deoxynojirimycin for Gaucher's disease and 1-deoxygalactonojirimycin for Fabry's disease are good examples of chemical chaperones that have shown satisfactory response in vitro, increasing enzyme activity and reducing storage products, and good efficacy in vivo (21,24).
Hematopoietic cell transplantation
Hematopoietic cell transplantation (HCT) from healthy compatible donors has been widely used to treat patients with LSDs. HCT provides metabolically competent cells at tissue sites, including the CNS, which may correct enzyme deficiencies by active secretion of the functional enzyme for cross-correction, and contribute to storage and tissue debris removal as well as down-regulating local inflammation. Indeed, macrophages and microglia represent major effectors of the catabolism of the storage material, and their replacement by metabolically competent cells can restore a critical scavenger function (25–27).
Based on this sound rationale and on the favorable outcome of HCT in adrenoleukodystrophy, which shares some similarities with some neuropathic LSDs, many LSD patients underwent allogeneic HCT from healthy compatible donors. A variety of stem cell sources were employed, with cord blood (CB) becoming widely used in the most recent years due to the several advantages in transplanting pediatric patients (28,29). Importantly, LSD showing a marked benefit derived from the transplant could be identified, such as MPS type I-Hurler syndrome (MPS-IH) (30), a lysosomal storage disease characterized by multisystem morbidity and death in early childhood, and Krabbe's disease (globoid cell leukodystrophy, GLD) (31). These early findings favorably impacted clinical practice in the form of treatment indications, i.e. in the case of MPS-IH patients for whom allogeneic HCT represents the treatment of choice if appropriate qualifying conditions are met, and/or in the establishment of newborn screening programs, as for GLD (32), allowing for activation of an immediate transplant procedure for the affected patients. Subsequent efforts contributed to the assessment of the long-term post-HCT outcome of patients affected by these disorders. A recent international study analyzed the long-term outcome of 217 patients with MPS-IH successfully engrafted with a median follow-up age of 9.2 years. Despite the overall clinical conditions of the transplanted patients' cohort were relatively good and a marked transplant-related phenotype amelioration was confirmed, considerable residual disease burden was observed in the majority of the treated children, with high variability between subjects. Importantly, key factors influencing the clinical long-term outcome, particularly considering cognition development, were identified in cognitive function preservation at the time of HCT, young age at transplantation and normal α-l-iduronidase (IDUA) enzyme level obtained post-HCT (the latter for superior long-term outcome in most organ systems) (33). Similarly, transplanted Krabbe babies experienced benefit exclusively upon very early transplantation (<4 months of chronological age) and in most cases declined significantly at long-term outcome. Long-term observation confirmed that also other LSDs are refractory to disease correction by HCT when severe CNS involvement, as well as rapidly progressive disease in symptomatic stage affected the patients at the time of transplant (34–39). The likely reasons for this sub-optimal outcome of HCT in these patients could be (i) the preexisting damage to the nervous system (and skeleton) and the possible establishment of pathogenic cascades not easily interrupted/accessible by functional enzyme delivery and metabolic correction of a fraction of the microglia/myeloid cell pool, and (ii) an insufficient or not timely delivery of the functional lysosomal enzyme to the affected tissues, and in particular to the affected brain. While precocious treatment administration is the only possible approach for addressing the first issue, quantity and timing of enzyme delivery to the affected tissues, including the brain, could be ameliorated by means of gene transfer into the cell product and modulation of the transplant protocol. Genetic engineering of autologous stem cells represents a proficient approach to increase the enzyme dose associated to hematopoietic stem cell (HSC) transplantation [see below—ex vivo gene therapy (GT)], as shown in preclinical LSD models and patients (40–42). Similarly, optimization of the transplant conditions could substantially contribute to increasing HCT-associated benefit by improving the speed and extent of enzyme delivery to affected tissues, and particularly to the brain. Indeed, the following key factors favorably affect the rate of myeloid/microglia reconstitution after HCT: Substantial improvements in the outcome of HCT could be expected upon proper application and manipulation of these variables.
The disease setting, and in particular neuroinflammation and microglia activation, which promote recruitment of bone marrow-derived cells to the CNS (43,44). This was observed in models of several neuropathologic conditions, such as nerve axotomy (45,46), CNS autoimmunity (46), ischemia (45), stroke (45), scrapie (47,48), Parkinson's (49) and Alzheimer's diseases (50–52), toxic demyelination (53) and in LSDs (41,44), as in the Twitcher mouse (animal model for GLD).
The administration of a proper preparatory regimen prior to HCT. As shown by Mildner et al., irradiation is required (although not sufficient) to enable donor cells contributing the parenchymal microglia pool. The initial interpretation was that some degree of irradiation-mediated damage could have been critical for entry of circulating cells into the brain (54). More recently, Capotondo et al. instead demonstrated that brain ablation/killing of brain-resident bona fide microglia progenitors mediated by irradiation and, to a greater extent, by busulfan favors microglia turnover by creating space for the engraftment and proliferation of HSC-derived early brain immigrants (55).
The transplantation of HSC fractions retrieved from the bone marrow mechanically (56) or upon mobilization (Capotondo, Gentner, personal communication). Indeed, the presence of bone fide microglia progenitors in the circulation of the recipients is necessary for their migration to the brain and contribution to microgliosis.
Gene Therapy for LSDs
Rationale of gene therapy for LSDs
GT has been proposed as a promising strategy to treat lysosomal disorders and ameliorate the affected patient's prognosis. The therapeutic potential of gene replacement strategies, initially exploited for primary immunodeficiencies where a positive selective advantage of the corrected cells was present, is highly favored also in LSDs due to the cross-correction mechanism that allows spreading of the functional exogenous enzyme in the affected tissues upon secretion by engineered depot cells or organs. Indeed, GT, according to the route of administration and the type of target tissue or organ, may establish a cellular pump releasing the therapeutic enzyme into the local tissue and/or extracellular fluids for the cross-correction of affected cells within the tissue and/or widespread in the body. Examples are represented by liver-directed gene transfer that allows release of enzyme into the bloodstream, or vector administration into the CNS that provides direct metabolic correction of specific cell types representing major disease targets, such as oligodendrocytes and neurons, and generates a local source of enzyme that is transported at a distance, i.e. along axons (57). Also replacement of affected tissue-resident cells by their genetically corrected counterparts or progenitors, such as in the transplantation of ex vivo transduced hematopoietic or neural stem cells (NSCs), could exert a similar effect.
Broadly, there are three conditions where GT, either in vivo or cell based, could benefit LSD patients' management: Critical issues and goals for the implementation of GT for LSDs in these contexts are represented by the development and availability of effective and safe approaches that would guarantee the achievement of the above-defined targets in each disease setting.
Disorders that are refractory to correction after allogeneic HCT and/or ERT and/or other conventional treatment modalities; this class mostly comprises rapidly progressive LSDs with severe CNS involvement [i.e. MPS-IIIA, MPS-IIIB, metachromatic leukodystrophy (MLD), Batten's disease]. In this case, GT is intended for providing a treatment option to the affected patients lacking a valuable treatment modality.
Diseases for which allogeneic HCT and ERT have a role currently, but where residual disease burden is generally observed in the treated patients in the medium to long term (i.e. MPS-IH, globoid leukodystrophy). Here, GT is aimed at substantially ameliorating the limited efficacy of currently available therapeutic options and/or at ameliorating treatment outcome at refractory disease sites.
Conditions for which ERT (and/or HCT) already provides a favorable outcome (i.e. milder MPS-I variants, MPS-VI, Gaucher, Pompe). In this case, GT could constitute a preferable option for safety and feasibility (adherence to chronic treatment, immunization, costs for ERT) concerns, and possibly provide superior extent and/or kinetics of substrate reduction and clinical benefit compared with standard of care.
Efficacy and safety of both in vivo and ex vivo GTs (Fig. 1) have been extensively demonstrated in several disease animal models of human LSDs and early, but promising clinical results obtained in humans are emerging, as detailed in the following paragraphs.
State of the art of gene therapy for LSDs
In vivo gene therapy
Systemic or intraparenchymal administration of GT vectors had been successfully applied to many LSD animal models with an emerging target represented by the CNS. Early proof of concept studies demonstrated that administration of vectors, either adenoassociated vectors (AAVs) or lentiviral vectors (LVs), into the systemic circulation or into the liver vascularity could provide successful delivery of therapeutic lysosomal proteins to peripheral organs, given a proper management of immune issues (58–62). Single vector administration allows converting the liver into a factory organ for systemic secretion of therapeutic proteins that was proved to be as effective as ERT in LSD murine models, potentially eliminating problems with compliance and costs (63,64). Following the favorable clinical experience in hemophilia, liver-directed in vivo GT is now approaching the clinical arena also in LSDs, with expected future candidate diseases including MPS-VI (64).
The most interesting recent advance in this area is likely represented by the use of site-specific genome editing to provide long-term, stable therapeutic expression of lysosomal enzymes (65). To overcome the limited targeting efficiency or insufficient expression from the endogenous promoter, the serum albumin locus was chosen as genomic safe harbor, characterized by very high expression level and easy tractability of the liver by gene delivery and in vivo editing, for zinc finger nuclease (ZFN)-mediated site-specific integration of the donor sequences encoding for human α-galactosidase A, acid β-glucosidase, iduronate-2 sulfatase or IDUA lysosomal enzymes, deficient in Fabry's and Gaucher's diseases and in Hurler's and Hunter's syndromes. Four weeks after administration, the lysosomal enzymes were detectable in liver lysates of treated mice and in the serum of wild-type mice, providing promising preliminary evidence of the potential applicability of this approach for LSD treatment.
Importantly, along with vector development, systemic administration is also becoming amenable to targeting the CNS. Rafi et al. have explored the outcome of the single i.v. injection of viral vector at Post-natal Day 10 in the severe animal model of globoid leukodystrophy (66). This has resulted in increased enzyme activity in the CNS and peripheral nervous system (PNS), clinical phenotype improvement with lack of tremors and continued weight gain and extended life span up to 20–25 days longer than untreated mice. Similarly, Weismann et al. observed restoration of βGal activity in the liver and serum; reduction in GM1-ganglioside content in the brain and spinal cord; clearance of lysosomal storage throughout the cortex, hippocampus, brainstem and spinal cord; decrease in astrogliosis; improved performance in multiple tests of motor function and behavior; and improved survival of GM-1 mice treated by systemic AAV9 GT (67). Also, Ruzo et al. obtained similar promising brain findings in MPS-IIIA mice (correction of pathological accumulation of glycosaminoglycans in CNS) treated by systemic AAV9 administration (68). In these settings, moderate widespread enzyme expression in the CNS of LSD mice was sufficient to achieve significant biochemical impact with phenotypic amelioration and extension of life span. However, whether this approach will indeed be sufficient to treat or prevent neurodegeneration and brain damage in patients is difficult to predict at present.
Systemic GT was also employed to induce tolerance to CNS-directed GT. Neonatal systemic AAV induced tolerance to CNS GT in MPS-I dogs and nonhuman primates (69). MPS-I dogs treated systemically in the first week of life with a vector expressing canine IDUA did not develop antibodies against the enzyme and exhibited robust expression in the CNS upon intrathecal AAV delivery at 1 month of age, resulting in complete correction of brain storage lesions. Newborn rhesus monkeys treated systemically with AAV expressing human IDUA developed tolerance to the transgene, resulting in high cerebrospinal fluid (CSF) IDUA expression and no antibody induction after subsequent CNS GT.
CNS-directed GT was also demonstrated to exert significant therapeutic benefit in several LSD animal models; both LV and AAV were employed with similarly promising results (70–77). Initial data obtained in LSD mice were replicated also in large disease animal models, providing stringent demonstration of the feasibility and transferability of the approach (78,79). Moreover, the mechanism of vector/enzyme distribution to the CNS tissue was analyzed, confirming enzyme transport along axons and identifying the CSF as a valuable enzyme delivery vehicle within the brain and spinal cord (57,73) (in press). Importantly, CNS-directed GT has already entered clinical testing in several LSDs. AAVs were employed for direct gene delivery into the brain parenchyma in patients affected by Canavan's disease (80,81), late-infantile neuronal cereoid lipofuscinosis (82,83), MPS type IIIA and B (84) (Tardieu—ESGCT 2015) and MLD (85). Due to the inherent risk of vector- and transgene-directed immune responses, most of these clinical trials employed immunosuppressive agents transiently to guarantee a durable effect of the treatment. Results of these clinical trials have proved the feasibility and overall safety of gene transfer into the brain via multiple sites of vector injection; however, limited information can be drawn in terms of treatment-related clinical benefit because of the symptomatic stage of most of the treated children at the time of GT administration. A more selective strategy in patients' enrollments as well as biochemical end points will likely contribute to a greater estimation of the clinical impact of these approaches.
The intra-cerebrospinal administration of AAV was also recently proved as efficacious in correcting LSD (neuro)pathology in diverse LSD animal models. Most recent examples report on the favorable phenotypic outcome of AAV8 administration in MPS-IIIA and B mice (86,87) and AAV2 in a dog model of late-infantile ceroid lipofuscinosis (LINCL) (in press). Plans for clinical translation of this approach are on the way.
Ex vivo gene therapy
Different cell types at diverse stage of maturation could be sampled from the patients, corrected in their lysosomal enzyme deficiency using GT vectors and reintroduced into LSD patients as therapeutic agents providing enzymes for uptake by deficient cells in the affected tissues, where the corrected cells could replace their enzymatically defective equivalent generating a resident and metabolically competent cell population. Despite committed cells, such as oligodendroglial progenitors (88) or neuronal progenitors (89), the unique properties of stem cells, in particular their capacity to regenerate, persist long term and the ability to migrate and differentiate into various cell types, render them a much more attractive tool for this purpose. Importantly, stem cell-based approaches could also provide means for replacing or repairing damaged tissues. Among stem cell types employed for LSD ex vivo GT, HSCs provided most promising results. A similar broad therapeutic potential could also be envisaged for NSCs that have been so far mostly been employed in cell therapy applications in LSD animal models. In both cases, key advantages associated to the use of engineered cells of autologous origin include (i) a substantial reduction of transplant side effects, risks, morbidity and mortality, by avoidance of the risk of graft-versus-host disease (HSCs) and rejection (NSCs) and (ii) an improvement of the therapeutic potential of the allogeneic transplant procedure by means of gene transfer. Indeed, autologous HSCs (and NSCs) can be genetically modified to constitutively express higher-than-normal levels of the therapeutic enzyme and become a quantitatively more effective source of enzyme than normal donors' cells also at the level of the affected tissues of LSD patients and mice (90,91). To this goal efficient gene transfer into the autologous cells to be transplanted is critical and strictly dependent on the use of appropriate vectors.
Therapeutic benefit was observed in several LSD animal models and patients upon transplantation of engineered HSCs and normal and/or engineered NSCs.
Early work on HSC GT for LSDs was conducted employing γ-retroviral vectors for HSC gene transfer; results had been quite disappointing, as in the case of Matzner et al., i.e. who observed a limited benefit associated to the treatment, with modest amelioration of the neuropathology and of the performance of treated animals at behavioral tests, in the murine model of MLD (92,93). These findings suggested that unexpectedly high levels of lysosomal enzymes might be required for the correction of the metabolic defect in the nervous system of LSD animals. A key advance in the field came with the use of LVs that enabled significantly improved HSC transduction efficiency and therapeutic gene expression in progeny. Indeed, by transplanting HSCs transduced with LV carrying the normal cDNA of the disease-causing gene, lysosomal enzyme activity was restored at higher-than-normal values in the hematopoietic system and affected tissues, including the brain, in different LSD animal models, with marked phenotypic amelioration. CNS and PNS disease manifestations were prevented and corrected upon pre-symptomatic and symptomatic treatments, respectively, in MLD mice (41,43). Similarly, LV-based HSC GT, employing an advanced miRNA-regulated construct (91) to tightly regulate lysosomal enzyme expression in hematopoietic cells (94), resulted in efficient delivery of the galactocerebrosidase (GALC) enzyme to the affected nervous system and in a dramatic amelioration of the overall disease phenotype of mice affected by GLD (42). Similar results were also obtained in the murine models of MPS-I (40), MPS-II (95), MPS-IIIA (96,97), cystinosis (98), Pompe's disease (99), Fabry's disease (100) and others. Efficient enzyme delivery was typically associated with metabolic correction of the affected tissues, as shown by the clearance of accumulated storage material within hematopoietic and non-hematopoietic cells, confirming the active secretion of the functional enzyme by the gene-corrected progeny of the transplanted cells and its re-uptake by the resident populations. Importantly, in most of the cases, GT allowed amelioration of the disease phenotype, including the neurological and skeletal defects. Importantly, when appropriate comparison with wild-type HSC transplantation was performed, it confirmed an equal if not superior proficiency of the transduced cells over their wild-type unmanipulated counterpart. The degree of efficacy of GT was proved to be dependent on the levels of enzyme activity in transduced HSCs and in their progeny.
Based on these promising data as well as dedicated safety assessments, early clinical testing of this approach has begun. In particular, a first successful experience has been recently reported for MLD, for the treatment of which a Phase I/II clinical trial is currently going on based on the use of autologous HSCs transduced with LV encoding for the functional arylsulfatase A (ARSA) cDNA and of a conditioning regimen based on the alkylating agent busulfan. Feasibility, medium-term safety and benefit of the procedure were described in the first three treated patients. The data showed sustained polyclonal engraftment of the transduced HSCs, extensive and stable ARSA gene replacement with up to above-normal enzyme expression observed in the reconstituted hematopoiesis and in the CSF of the patients. Additionally, the disease had not progressed in the treated patients even after the projected time of onset from sibling cases. These findings are supported by a recent follow-up on a large patient population (Sessa, Lorioli et al., under peer revision). Following this favorable clinical experience, new candidate diseases have been identified for which clinical testing is expected in the near future.
In the case of NSCs, initial attempts mostly focused on the use of wild-type cells without genetic correction (98,101–103), but the most exciting perspective for clinical application is the one based on NSC generation from patient-specific induced pluripotent stem cells (iPSCs) for which gene transfer constitutes a critical requirement. This approach was successfully tested in a preclinical LSD model by Doerr et al., who generated long-term self-renewing neuroepithelial stem cells and astroglial progenitors from iPSCs derived from patients affected by MLD for cell-based ARSA replacement. Following transplantation of ARSA-overexpressing precursors into ARSA-deficient mice, significant reduction of storage up to a distance of 300 µm from grafted cells was observed, providing a proof of concept that neural precursors generated via reprogramming from MLD patients may be employed as an autologous cell-based vehicle for continuous ARSA supply in MLD-affected brain tissue (104). Importantly, these findings reproduce and extend the results obtained in the same disease model employing allogeneic neurospheres by Givogri et al. (105), thus confirming the actual value of iPSCs as a source of autologous genetically corrected precursor cells for therapeutic transplantation. Griffin et al. in a well-characterized mouse model of Sly's disease explored a similar iPSC-based approach. Human Sly disease fibroblasts were reprogrammed into iPSCs, differentiated into NSCs, genetically corrected with a transposon vector and assessed for their engraftment potential in a humanized mouse model and for their contribution to disease correction into the Sly animal model. Interestingly, corrected iPSC-NSCs not only showed good engraftment and survival capacity in the nonobese diabetic severe combined immunodeficient brain, but also reversed neuropathology in the proximity of the grafts when transplanted post-symptomatically into the striatum of adult Sly disease mice (106). These findings are of great relevance since they demonstrate the potential for ex vivo GT in the brain using human NSCs from autologous, non-neural tissues, thus opening the way for clinical application of this approach.
So far, what has entered clinical testing in LSDs is transplantation of allogeneic primary NSCs. After demonstration that human NSCs could produce both PPT-1 and TPP1 and result in donor cell engraftment and reduced accumulation of storage material in the brain when tested in the mouse model of neuronal ceroid lipofuscinosis (NCL), human NSC transplantation was tested in an open-label dose-escalation Phase I clinical trial as a potential treatment for infantile and late-infantile NCL (107). Allogeneic unmanipulated cells were transplanted into the cerebral hemispheres and lateral ventricles of 6 patients, accompanied by 12 months of immunosuppression. The treatment was feasible and well tolerated by the patients, supporting further exploration of its therapeutic potential for NCL, and possibly for other neurodegenerative disorders.
Several efforts have been put into developing effective combined approaches to treat the global phenotype of LSDs in animal models. One of the LSDs in which combinatorial approaches were tested most is GLD, a severe demyelinating LSD caused by mutations in GALC. GALC deficiency results in the inability to degrade several substrates, such as galactosylceramide and its toxic lysolipid derivative galactosylsphingosine (psychosine) in myelinating cells, and to a minor extent, in neurons of the CNS and PNS; this, in turn, causes global deterioration of white matter tracts and neurodegeneration. In the classic early infantile form, children present with symptoms by the first 6 months of life, then rapidly lose their motor and cognitive skills and die within a few years of symptom onset (108). This severe phenotype is well recapitulated by the historical natural GALC mouse mutant, the Twitcher mouse. The severity, multisite involvement of the nervous system and great speed of progression render GLD an optimal target for combinatorial approaches. Indeed, individual treatment attempts in GLD murine models provided a variable extent of metabolic correction and pathological amelioration but were overall modestly effective in delaying the onset of symptoms and enhancing lifespan, and failed to address the global disease (42,73,89,109–113). Combinatorial approaches tested so far in animal models include HCT, SRT, systemic/intrathecal delivery of the recombinant GALC protein, intracerebral/intrathecal injection of AAVs, CNS-directed or systemic injection of LV or AAV and NSC transplantation (114–118). These studies showed a variable extent of additivity or synergy of the treatments, often resulting in enzymatic reconstitution, amelioration of some pathological hallmarks, improved motor performance and remarkable extension of lifespan of the treated animals. The potential transferability of these proof-of-concept studies would require validation in large animals models of the disease and, most importantly, safety testing addressing the different components of the combinatorial approach.
GT for LSDs has evolved substantially in recent years, and positive evidences in favor of its therapeutic potential in this disease setting are accumulating. Major efforts are currently directed toward targeting the LSD brain disease and improving its management. The different approaches so far developed in preclinical models are progressively reaching the clinical testing phase, which is and will be instrumental for definition of the future steps that may render gene-based innovative treatments available for the affected patients. In this direction, it is likely expected that combinatorial strategies, allowing to address the complexity of the LSD pathology as well as key critical issues such as the timing of intervention needed for therapeutic benefit at different disease sites, would become of progressively increasing interest, once the safety and efficacy of individual strategies are known through early single-therapy clinical experiences. Certainly, the promising preclinical and clinical data currently available already suggest that GT will be a valuable therapeutic option for LSD patients.
Conflict of Interest statement. The San Raffaele Telethon Institute for Gene Therapy (TIGET) is a joint venture between Telethon and OSR, with no legal personality. A.B. is the Principal Investigator of the TIGET-MLD clinical trial of GT for MLD. The GT of MLD was licensed to GlaxoSmithKline (GSK) in 2014, and GSK became the financial sponsor of the trial. Telethon and OSR are entitled to receive milestone payments and royalties upon commercialization of such therapies.