Peripheral neuropathy in metachromatic leukodystrophy: current status and future perspective

The accumulation of metachromatic material in peripheral nerves in MLD have first been reported by Jacobi [40]. Metachromatic material consists of Schwann cells and endoneural macrophages that are filled with characteristic lysosomal inclusions of sulfatides, also called inclusion bodies. The sulfatides are normal in structure but cause a lower cerebroside-sulfatide ratio in myelin composition and a disruption in myelin metabolism [41]. Schwann cells and phagocytes die, and demyelination of myelin in the PNS and CNS occurs. Rarely, evidence of actual destruction of the axons can be observed.

Remarkably, no correlation between demyelination and the presence of metachromatic material in peripheral nerves has been found [22, 42‐45]. This raises the question whether peripheral neuropathy in MLD is (partially) due to other causes in addition to sulfatide accumulation. On the other hand, sulfatide levels in the cerebrospinal fluid (CSF) and sural nerve do reflect the severity of peripheral neuropathy (measured by nerve conduction studies), while they are not proportional to central white matter injury (assessed by the Gross Motor Function Measure 88–items score, somatosensory evoked potentials, and MR spectroscopy) [42].

Segmental demyelination and reduction in number of myelinated fibers are most severe in late-infantile MLD and in more advanced stages of the disease. Larger myelinated fibers tend to be more affected, resulting into the loss of normal bimodal distribution of myelin sheath thickness. Remyelination may occur and is mostly seen in adult MLD patients. The observed increased g-ratios (ratio between axonal diameter and myelinated fiber diameter) suggest that the thick myelinated fibers are remodeled into thin myelinated fibers [46].

On the other hand, one recent study reported homogenous enlargement of the peripheral nerves on ultrasound in a patient with advanced late-infantile MLD. The echo-intensities of the nerves were normal to reduced, possibly due to the expression of accumulated inclusion bodies [47]. These findings should be taken with caution since they have not been validated in other MLD patients. However, cranial nerve and cauda equina enhancement on MRI might also suggest nerve enlargement secondary to the accumulation of metachromatic material [48‐51], although contrast enhancement could also result from a disturbed blood–nerve barrier [52]. Hypertrophic changes and onion bulbs, as are seen in hypertrophic neuropathies and chronic inflammatory demyelinating polyneuropathy, have only rarely been noticed.

Inclusion bodies, including zebra, tuffstone, prismatic, lamellar and granular bodies, are the characteristic cell alterations observed in neural and non-neural tissue of MLD patients. They consist of metachromatic material and can already be found in the peripheral nerves of asymptomatic patients, even before birth [53‐56]. Numbers of inclusion bodies are higher in patients with late-infantile MLD, due to higher sulfatide levels and lower ASA activity compared to later onset forms. Besides, some studies found that tuffstone bodies are more frequent in late-infantile MLD, while zebra bodies are more frequent in juvenile and adult MLD. However, whether different inclusion body types have different roles in disease pathogenesis is unclear as different types can blend into each other and most likely reflect different orientations and packing of metachromatic material instead of different disease mechanisms [5, 56].

Notably, Cravioto et al. [57] and Argyrakis et al. [53] also described several abnormalities other than inclusion bodies. These are morphological alterations of the endoplasmatic reticulum and mitochondria in Schwann cells, and accumulation of glycogen in mitochondria, Schwann cells and axons. These abnormalities could reflect a metabolic derangement of these cells, causing premature cell death, and may explain the lack of a correlation between demyelination and presence of metachromatic material. Nevertheless, it is too early to draw any firm conclusions based on two individual cases [58].

Hematopoietic cells from bone marrow, peripheral blood or umbilical cord blood are able to cross the blood-brain and blood-nerve barrier, differentiate into macrophages/microglia, and deliver ASA into the CNS and PNS [76]. Allogeneic HCT has been proven to correct ASA deficiency in MLD patients if stable engraftment following transplantation has been accomplished [79]. Nonetheless, the replacement of ASA deficient host cells by ASA producing donor cells is slow, resulting in a delay estimated at 12–24 months until the disease stabilizes. This makes HCT unsuitable for symptomatic MLD patients or (asymptomatic) patients with the late-infantile MLD. Considering time, unrelated umbilical cord blood is currently preferred over bone marrow and peripheral blood because stored umbilical cord blood can be identified and transplanted faster than other sources [12, 79, 81].

Nevertheless, HCT treatment effects on the PNS in most clinical studies (NCT00383448, NCT00176904, NCT01043640, NCT01626092) are considered disappointing when compared to the CNS, although two case studies describe stabilization or improvement of symptoms in the PNS only [82, 83]. For example, Boucher et al. [11] found that 76% of the patients demonstrated worsening peripheral neuropathy after HCT, compared to 31% of the patients with worsened demyelination in the CNS (n = 40, follow-up = 0–30 years). De Hosson et al. [13] found that NCV studies for all patients deteriorated while the white matter lesions on brain MRI were stable for most patients (n = 5, follow-up = 18–29 years). Martin et al. [3] evaluated long-term outcomes after unrelated umbilical cord blood transplantation (UCBT) in late-infantile and juvenile MLD patients. They found that the brain lesions improved in 84% of asymptomatic patients, but that the NCV results continued to decrease, resulting in a decline in gross motor function for all except one patient (n = 19, follow-up = 2–14 years). Finally, Chen et al. [12] compared asymptomatic juvenile MLD patients who underwent unrelated UCBT. Brain MRI abnormalities were stable, but their peripheral neuropathy progressed. Nonetheless, the speed of progression in UCBT patients was slower when compared with their untreated siblings (n = 3, follow-up = 7–17 years).

The use of autologous hematopoietic stem cells transduced with a lentiviral vector containing a healthy copy of the ARSA gene allows supra-normal production (500–1000%) of ASA by donor cells, due to overexpression of the gene by a stronger promoter. This ex vivo gene therapeutic approach could therefore be faster and more effective in cross correction of ASA deficient graft cells when compared to HCT alone [76, 77]. After favorable treatment effects on both the CNS and PNS in MLD mouse models [84‐86], multiple clinical trials on hematopoietic stem cell-directed gene therapy (HSC-GT) for the treatment of MLD have started (NCT02559830, NCT01560182, NCT03392987). The preliminary results and ad-hoc analysis of one of these trials (NCT01560182) have already been published. In this clinical trial, HSC-GT in nine patients with early onset MLD (< 6 years old) in an asymptomatic or early-symptomatic phase, resulted in stable engraftment and correction of ASA deficiency in all hematopoietic cell lines and CSF. At follow-up (18–54 months after HSC-GT), NCV improved in three patients, remained relatively stable in four, and substantially decreased in two, particularly in the first 6–12 months of follow-up. Brain MRI abnormalities were stable or improved in eight patients. Signs of remyelination in the PNS were also found in a few patients, with better remyelination in patients with a higher transduced cell engraftment [78, 87]. Although long-term treatment effects have yet to be determined, the stable or improving NCVs in combination with signs of PNS remyelination indicate that the majority of HSC-GT treated patients do indeed benefit from higher ASA levels, and thereby probably improved enzyme delivery to the PNS when compared to HCT.

Another potential gene therapy approach is to restore the ARSA gene in vivo by using an adeno-associated virus (AAV) as vector. This AAV-based gene therapy can be administered directly to the CNS, either via an intraparenchymal or intrathecal route, correcting the ARSA gene in local cells and resulting in an even faster ASA expression, secretion and cross-correction in CNS cells, such as astroglial cells and oligodendrocytes for some AAV serotypes serotypes [88‐92]. This can be of particular importance since astroglial cells and oligodendrocytes might not take up the non-phosphorylated form of ASA, secreted by bone marrow-derived macrophages/microglia, via the mannose 6-phosphate receptor pathway [93]. Besides, in vivo gene therapy is thought to work at distance, e.g. in the peripheral nerves, by spreading of the AAV vector and/or ASA by either diffusion along the myelinated tracks or by retrograde/anterograde axonal transport [94, 95]. Nevertheless, the potential effects of in vivo gene therapy on the PNS have yet to be demonstrated in MLD. So far, intraparenchymal administration of serotype 5 AAV prevented motor coordination impairment in 18-month-old treated ARSA knockout mice, but effects on PNS function could not be judged as also the untreated mice lacked PNS abnormalities [88, 96]. In addition, intraparenchymal administration of serotype 2–5 recombinant AAV did not result in presence of the vector in the sciatic and radial nerves in macaques, whereas a clear diffusion of the vector and a significant increase of ASA activity was observed in the injected brain hemisphere [97]. Finally, a clinical trial on CNS-administered AAV-based gene therapy with serotype rh.10 in human patients with early onset MLD (< 6 years old) (NCT01801709) has been halted due to lack of efficacy [98], and effects on the PNS in these patients have not been reported yet. However, combining CNS-administered and intravenously-administered AAV-based gene therapy might be more promising, as this combination showed synergistic effects on presence of the viral vector, enzyme activity and functional outcomes in both the CNS and PNS in mouse and canine models of Krabbe disease [99, 100].

ERT is used with variable success in treating some lysosomal disorders, including Gaucher disease, Fabry disease, mucopolysaccharidoses type I, II, and VI, and Pompe disease [101]. However, its applicability to MLD is challenged as ASA has a high molecular weight, and is therefore unable to penetrate the blood-brain and blood-nerve barrier. Nevertheless, Matthes et al. [66] found that intravenous ERT reduced sulfatide storage in the brain and peripheral nerves, and led to increased NCVs in early treated MLD mouse models. Since then, the results of multiple clinical trials on intravenous administration of Metazym (HGT-1111, recombinant human ASA) have been reported (NCT01303146, NCT00681811, NCT00633139, NCT00418561). Unfortunately, none of them show any beneficial treatment effect of ERT on the CNS and PNS in human patients so far [102]. Recently, Simonis et al. [103] were able to increase the catalytic rate constant of intravenous administered ASA by protein engineering, resulting in a threefold more reduction of sulfatide storage in the PNS and CNS in humanized ARSA knockout MLD mouse models. This might be promising for all enzyme-based therapies including ERT and gene therapy. In order to avoid the blood-brain barrier, clinical trials consisting of intrathecal administration of recombinant human ASA (HGT-1110) as a for symptomatic late-infantile and juvenile patients (age up to 13 years) have also been started (NCT01510028, NCT01887938); however these results have not been published yet.

There are several other forms of therapy that have been studied in small clinical trials. One of these is the administration of warfarin. Since vitamin K availability could be a rate-limiting step in the production of sphingolipids and the conversion of cerebrosides into sulfatides [104], it was hypothesized that warfarin, a vitamin K antagonist, could mitigate the MLD phenotype by reducing the amount of sulfatide formation. This hypothesis was supported by the studies of Sundaram and Lev that found that administration of warfarin lowers brain sulfatides in mice [105, 106]. Assadi et al. [104] therefore examined the treatment effects of warfarin in four advanced juvenile MLD patients (of whom two patients had a PSAP variant; NCT00683189); however, they did not demonstrate any beneficial treatment effects.

The effect of allogeneic mesenchymal stem cell (MSC) infusion was studied in six MLD patients that previously had an allogeneic bone marrow transplantation (no ClinicalTrials.​gov identifier). In four of them there was clear evidence of improvement of NCV at follow-up between 1 and 2.5 years, with an increase in NCV between six to 12 m/s. They speculated that this improvement is due to Schwann cell differentiation of MSC in vivo or to passive enzyme transfer into peripheral nerves provided by the MSC. However, MSCs are incapable of differentiating into Schwann cells, and also the transient nature of the improvement in one patient suggests that passive enzyme transfer is more likely [107].

Finally, it was thought that administration of supplemental umbilical cord blood cells would increase the speed at which normal levels of circulating blood cells are re-established after UCBT. This was tested in one clinical trial with ALD-101 in patients with late-infantile and juvenile MLD (NCT00654433), and in one clinical trial with ALD-601 in pregnant women with affected fetuses (NCT01003912). Both studies were terminated early due to disappointing results and non-enrollment, respectively.