Gene therapy for global brain diseases: one small step for mice, one giant leap for humans

Abstract

This scientific commentary refers to ‘Global CNS correction in a large brain model of human alpha-mannosidosis by intravascular gene therapy’, by Yoon et al. (doi:10.1093/brain/awaa161).

This scientific commentary refers to ‘Global CNS correction in a large brain model of human alpha-mannosidosis by intravascular gene therapy’, by Yoon et al. (doi:10.1093/brain/awaa161).

The gene therapy field has recently reached another significant milestone in its march towards revolutionizing the treatment of genetic diseases. Onasemnogene abeparvovec-xioi (Zolgensma^®), an adeno-associated viral (AAV) vector-based gene therapy, has been approved by the FDA for the treatment of spinal muscular atrophy (SMA). It joins voretigene neparvovec (Luxturna^TM), another FDA-approved AAV-based therapy—for the treatment of retinal dystrophy—in making gene therapy an effective and potentially commercially viable treatment option. It is noteworthy that both these initial successes were realized in rare conditions, an arena that has become ground zero for gene therapy clinical trials. In this issue of Brain, Young Yoon and co-workers test a gene therapy approach in a gyrencephalic model of another rare disease, alpha-mannosidosis, and show that gene therapy may have potential for the treatment of global brain diseases (Yoon et al., 2020).

Zolgensma^® has shown an impressive ability to improve motor function and survival in patients with SMA, mainly by rescuing motor neurons within the spinal cord (Mendell et al., 2017). Importantly, it has also shown that AAV serotype 9 (AAV9) can cross the blood–brain barrier in humans and deliver the therapeutic SMN1 gene to motor neurons following a single non-invasive intravenous administration. This has provided hope to those studying other neurological diseases, especially the large number of rare intractable conditions that affect multiple regions of the brain. However, these conditions will represent a far more complex challenge than SMA. There is nothing simple about delivering therapeutic genes with the required efficiency to multiple neural populations in multiple regions of the most complex organ of the body. Numerous preclinical studies in mouse models of monogenic brain disease have suggested that intravenous administration of an AAV9 vector is sufficient to cure these monogenic disorders. However, the debate over whether this will work sufficiently well in the larger and more complex human brain rages on. After all, the gyrencephalic human brain is >600 times larger in volume than the lissencephalic rat brain, and the total surface area of the human cerebral cortex is 377 times larger than that of the rat (Hofman, 1985, 1988). While the cell types present in smooth and convoluted brains are largely the same (Ferrer et al., 1986a, b) (Fig. 1), differences are seen in the orientation of neurons as a result of folding, the formation of local circuit subsystems and the development of the ipsilateral cortico-cortical fibrillary system characteristic of the gyrencephalic brain.

Figure 1.

Neuronal architecture in lissencephalic and gyrencephalic mammals. (A) Neuronal types in layer VI of the cerebral cortex of the mouse (Mus musculus). ax = axons; FP = flattened pyramidal neuron; H = horizontal neuron; HP = horizontal pyramidal cell; IP = inverted pyramidal neuron; LP = large pyramidal neuron; M = Martinotti cell; S = multipolar neuron with locally arborizing axon; SP = small pyramidal neuron. *Multi-apical pyramidal cells. Scale bar = 100 µm. (B) Camera lucida drawing of neurons in the sixth layer of the (i) gyral and (ii) fissural regions of the cat cerebral cortex. aP = atypical pyramidal cells; B = bipolar cells; F = fusiform neurons; LC = local circuit neuron; P = pyramidal neuron. *Inverted pyramidal cell in the gyrus and a horizontal pyramidal cell in the fissural region. Small arrows point out the course of the axons and collaterals. (C) Neurons in the sixth layer of the (i) gyrus and (ii) sulcus of the cerebral cortex of the human infant. B = bipolar neuron; LC = local circuit neuron; sMp = spinous multipolar neuron with long descending axon. *Inverted pyramidal cell in C(i), and a tangential (horizontal) pyramidal cell in C(ii). Reproduced with permission from Ferrer et al. (1986a, b).

In this issue of Brain, Yoon and co-workers provide evidence that intravenous AAV-mediated gene therapy may also be effective in gyrencephalic brains (Yoon et al., 2020). They begin by showing that in adult mice, intravenous administration of the vector AAVhu.32 leads to greater levels of brain transduction than use of the closely related AAV9. They then use the larger gyrencephalic feline brain as a ‘stepping stone’ towards modelling what will happen in the human brain, and show that intravenous administration of AAVhu.32 mediates widespread reporter gene delivery. Interestingly, despite using a minimal human GUSB promoter that should express in all cell types, only neuronal expression was observed. The researchers then conduct a preclinical study in a cat model of alpha-mannosidosis; a neurodegenerative lysosomal storage disease caused by mutations in the alpha-mannosidase gene (MANB) that results in global brain pathology. AAVhu.32 carrying a therapeutic copy of the feline MANB gene was administered to the cat model via either intravenous or intra-arterial injection. The resulting therapeutic response was dose-dependent, and intra-arterial administration of the vector provided better therapy. Improvements were seen in ataxia, survival and neuropathological parameters.

The findings of Yoon et al. are a step forward in developing a gene therapy strategy for alpha-mannosidosis. They may also help enhance gene therapy for other conditions, especially those in which there is a possibility of soluble enzymes leaking out of transduced cells and cross-correcting the defect in untransduced neighbouring cells, thereby amplifying the therapeutic effect. The study also provides some important reminders. The first is the importance of the delivery vector in dictating whether a gene therapy approach will work. Another is that we cannot simply assume that a new vector will behave as predicted, any new vector will require thorough evaluation in its own right. AAV9 and AAVhu32 are both members of phylogenetic clade F, and Yoon et al. describe how their capsid proteins differ by 12 amino acids. Nevertheless, the higher levels of transduction, strong neuronal tropism and differing transduction patterns seen with AAVhu32 relative to AAV9 are not easy to predict, especially given that AAVhu32 transduced other types of neural cells when injected directly into the brain parenchyma of mice (Cearley et al., 2008).

The route of administration for a gene delivery vector is an important consideration. Yoon et al. demonstrate that intra-arterial administration of AAVhu32 has greater therapeutic efficacy than intravenous administration. The authors point out that this is in spite of a limited number of studies in non-human primates showing little difference in transduction efficiency between the two routes of administration using the AAV9 vector. Systemic routes of administration have the added advantage of delivering gene therapy to any visceral organs that may be affected in addition to the brain. However, what we cannot gauge from this study is whether the intra-arterial route of administration is superior to administration of the vector via the CSF. This could be achieved via intracerebroventricular, intra-cisterna magna or intrathecal administration, in effect using the CSF to maximize delivery throughout the CNS. There are ongoing clinical trials using intravenously administered AAV9 vectors for other neurodegenerative lysosomal storage disorders (e.g. mucopolysaccharidosis IIIA, ClinicalTrials.gov: NCT02716246). But more researchers have opted for intra-CSF administrations (e.g. in mucopolysaccharidosis IIIA via intracerebroventricular administration, EudraCT number 2015-000359-26; in CLN6 Batten disease via intrathecal administration, ClinicalTrials.gov: NCT02725580). This also raises the question of whether bypassing the blood–brain barrier and injecting directly into the CSF would alter the neuronal tropism of the AAVhu32 vector. Indeed, would broadening the tropism of the vector to include glial cells be advantageous, for example by providing more cells to act as enzyme-producing ‘factories’? It is interesting to observe that there is a significant decline over time in levels of alpha-mannosidase in the plasma and in the CSF of treated cats. A decline in plasma levels could be attributed to turnover of cells in visceral organs and dilution of the non-integrating AAV genome. However, the decline in enzyme levels within the CNS is surprising given that expression is coming exclusively from a stable population of non-dividing postmitotic neurons.

The challenge now lies in evaluating whether the AAVhu32 vector has the potential to be effective in the even larger and more complex human brain. A logical step forward would be testing it in non-human primates. Information is also required on the prevalence of pre-existing neutralizing antibodies against AAVh32 in paediatric and adult human populations. The normal serum chemistry and lack of any measurable adverse effects in the cats treated in the current study is encouraging, particularly as high doses of intravenously administered AAV vectors in SMA (Mendell et al., 2017) and haemophilia B (Nathwani et al., 2011) clinical trials have resulted in elevated serum liver enzymes, which required transient corticosteroid treatment to normalize.

The promise of gene therapy lies in the potential to cure genetic diseases from a single administration. To realize this for global brain disorders, the hunt for more effective and efficient vectors must go on. The study by Yoon et al. will therefore be welcomed, particularly by scientists and clinicians working in the rare paediatric neurodegenerative arena where the prognosis for patients is bleak. No effective treatments are available but novel approaches such as gene therapy could prove to be a light at the end of a dark tunnel.

Funding

A.A.R. is supported by the UK Medical Research Council (MR/N026101/1, MR/S036784/1, MR/R025134/1, MR/R015325/1, MR/S009434/1), the Wellcome Trust Institutional Strategic Support Fund/UCL Therapeutic Acceleration Support (TAS) Fund (204841/Z/16/Z), Action Medical Research (GN2485) and Asociación Niemann Pick de Fuenlabrada. P.G's research is supported by the UK Medical Research Council (MRS0191111, MC_U12266B), Great Ormond Street Hospital Charity and SPARK Children’s Medical Research (V4420), UCL Therapeutic Acceleration Support (TAS) Fund (556551).

Competing interests

The authors report no competing interests.