Robin Ali, PhD
University College London
To explore the challenges that will face the development of therapies for NCL disease caused by mutations in intracellular transmembrane proteins (specifically juvenile CLN3 disease and CLN6 disease)
To determine whether any of the following approaches provide benefit
(1) over-expression of CLN3 and CLN6 is toxic to cells
(2) AAV-mediated expression of CLN6 delays retinal or brain neurodegeneration
(3) neural stem cell-mediated intraocular delivery of neuroprotective factors provides benefit.
Production of AAV vectors
AAV constructs carrying the human or murine CLN3 gene under control of the CMV promoter were cloned. The constructs also contained a GFP marker gene (IRES-driven) that allowed the identification of transduced cells. The CLN3 constructs were packaged into AAV9 capsids, as this vector serotype was expected to be effective in transducing neurons after administration into the central nervous system. Similar AAV constructs carrying the human or murine CLN6 gene were cloned and packaged into the AAV8 serotype, a serotype known to transduce photoreceptor cells after subretinal delivery into the eye. Control vectors carrying the GFP marker gene (AAV-GFP) or without a transgene (AAV-empty) were produced for both serotypes to control for effects relating to vector administration. All vectors were assessed for quality (endotoxin levels, non-AAV proteins) and titres. Where possible, the expression of transgene was assessed in vitro prior to use of the vectors in vivo.
Production of genetically engineered neural stem cell cultures with forced expression of one or more neuroprotective factors
To evaluate whether a sustained cell-based intraocular administration of neurotrophic factors (NTF) results in attenuation of photoreceptor degeneration in in a CLN6 mouse model, different neuroprotective NTFs (glial cell line-derived growth factor (GDNF), brain-derived growth factor (BDNF), pigment epithelium-derived growth factor (PEDF), ciliary neurotrophic factor (CNTF)) were expressed in neural stem (NS) cells derived from the cerebral cortex of embryonic mice. When these cells are cultivated under adherent conditions in the presence of mitogens, they give rise to homogeneous cultures of symmetrically dividing tripotent neural stem (NS) cells. To modify these cells, the cDNAs of the different NTFs were cloned into polycistronic lentiviral vectors additionally encoding different combinations of reporter and resistance genes under regulatory control of the cytomegalovirus enhancer/chicken ß-actin (CAG) promoter. NS cells for control experiments were modified with the same vectors but lacking the cDNAs for the NTFs. Because expression levels of the NTFs from the polycistronic vectors are proportional to the expression levels of the fluorescent reporter proteins, single NS cells with the highest expression levels of the reporter genes from transduced bulk cultures were selected by fluorescence activated cell sorting (FACS), and clonally expanded to generate clonal cell lines with elevated expression levels of NTFs. After several rounds of transductions and clonal expansions, several BDNF-, GDNF-, PEDF- and CNTF-expressing NS cell lines were selected with different expression levels of the transgenes for further characterization and for transplantation experiments. Immunocytochemical and Western blot analyses revealed stable expression of the NTFs in these cell lines, and confirmed secretion of the factors into the culture supernatants. When the modified cells were differentiated into astrocytes or neurons in vitro, they continued to express the NTFs. Stable expression of the transgenes was also observed in intravitreally transplanted cells over a time period of at least six weeks, the longest post-transplantation interval analyzed.
Detail onset and progression of retinal pathology and visual defects in CLN6 & CLN3 mouse models
To allow treatment efficacy to be assessed, survival of photoreceptor and retinal neuron subclasses, glial activation, integrity of retinal pigment epithelium, accumulation of autofluorescent storage material, optic nerve pathology, and extent of remaining visual function have been assessed by optomotor testing and electrophysiological measures of visual function (ERGs and VEPs) in different ages of C57Bl6/J congenic CLN6 and CLN3 mouse models. These findings have been correlated with existing knowledge on the integrity of the visual system in these mice and provide landmarks of disease progression that are crucial for assessing the impact of therapy in the visual system.
CLN3 mouse model
Assessment of the retina in the CLN3 mouse revealed that there was no consistent retinal pathology or visual defect in this animal model within the first year of life. As the initial administration of a therapeutic is generally associated with a small acute drop in visual function/retinal thickness (especially after subretinal administration), the very slowly progressing loss of photoreceptors in the CLN3 mouse created an unsurmountable hurdle to the development of a novel therapeutic regimen. For this reason, these mice have not been used in the development of ocular therapies.
CLN6 mouse model
Detailed characterisation of the ocular phenotype in the mouse model of vLINCL (Cln6nclf mouse) allowed to create a morphological and functional description of the disease phenotype. Analyses of retinas from Cln6nclf mice at different developmental ages revealed a slowly progressing photoreceptor degeneration. Apoptotic photoreceptors were present at 2 weeks of age, with loss of photoreceptor cells evident at the end of the first postnatal month, and nearly complete in 9 months old mice (Fig. 5 A). Retinal degeneration (Fig. 5 B) was paralleled by elevated expression levels of glial fibrillary acidic protein in astrocytes and Müller cells (Fig. 5 A), and a dysregulation of lysosomal proteins such as cathepsin D and Lamp1 as revealed by immunohistochemistry, Western blot analyses and enzyme activity assays. Furthermore, electron microscopic analyses revealed the presence of lysosomal storage material with curvilinear and fingerprint-like inclusions in various retinal cell types. Characterisation of retinal function at various ages, identified early changes to the a- and b-wave form in electroretinography. A decrease in scotopic (dark-adapted) a-wave amplitude (a function of the rod photoreceptors) was present as early as 3 weeks of age, the youngest age at which reliable ERG measurements can be performed. The loss of function progressed rapidly before slowing around 3 months (Fig. 5 C). The scotopic b-wave (a measure of rod bipolar cell function) was significantly lower than in WT mice, but this difference was not progressive. No notable differences were present in the photopic (light-adapted) ERG. The morphological and functional characterisation of the animal model suggests that the rod photoreceptor cells are the primary cell type affected in this disease.
To conclude, the defined steps of photoreceptor degeneration suggest that nclf mice might serve as an ideal animal model for experimental therapeutic approaches aimed at attenuating vision loss in neuronal ceroid lipofuscinosis. These results have been published by Bartsch U et al. Invest Ophthalmol Vis Sci (2013).
Fig. 5 Disease phenotype in retinas of Cln6nclf mice. (A) The photoreceptor layer (onl) was similar in thickness in 0.5 months old wild-type (a) and Cln6nclf mice (b), and expression of GFAP was restricted to retinal astrocytes in both genotypes (a, b). In 1 month old mutants (d), the photoreceptor layer was slightly reduced in thickness when compared to age-matched wild-type mice (c). Photoreceptor loss was more pronounced in 4 months old mutants (f), and almost complete in 9 months old Cln6nclf mice (h). Retinal degeneration in CLN6nclf mice was paralleled by elevated expression levels of GFAP in retinal astrocytes and Müller cells (d, f, h) when compared to age-matched wild-type mice (c, e, g). (B) The number of rows of photoreceptor nuclei was similar in 0.5 month old CLN6nclf and wild-type mice, but then decreased in Cln6nclf mutants with increasing age of the animals. In 9 months old mutants, the photoreceptor layer was composed of about 3 rows of photoreceptor nuclei as opposed to about 11 rows of nuclei in age-matched wild-type retinas. (C) Scotopic a-wave amplitude, a measure of rod photoreceptor activity, was affected in 1 month old Cln6nclf mice, decreasing rapidly over the subsequent months (data shown at 2 months of age)
AAV-mediated gene transfer in CLN6 & CLN3 mouse mutants
Retinal gene therapy in Cln6nclfmice
The characterization of the visual phenotype revealed that Cln6nclfmice present with photoreceptor cell death and reduced photoreceptor activity from 2 weeks and 3 weeks of age, respectively. These findings indicate firstly that the mutation in Cln6 predominantly affects photoreceptor survival and secondly that therapeutic interventions most likely need to be administered before the age of 2 weeks to prevent the onset of the retinal degeneration.
To assess toxicity associated with transgene expression, the first step in the development of gene therapy, AAV2/8.cmv.hCLN6 vectors were produced as described above and injected subretinally into eyes of adult wild type animals. A range of titers of AAV2/8 vectors carrying Cln6 has been used.
To conclude, preliminary data showed that high expression levels of Cln6 are detrimental to retinal function and morphology. Currently, lower expression levels of Cln6 in photoreceptor cells are being assessed for their therapeutic efficacy.
Kleine-Holthaus et al., presentation at the 14th International Congress on Neuronal Ceroid Lipofsucinoses, Oct. 2014, Cordoba, Argentina and will be published soon.
Delivering gene therapy to the brain in Cln3-deficient mice
Because the cross-correction of adjacent cells is not possible in disorders caused by a transmembrane protein deficiency, the requirements of a gene therapy vector may differ to those needed in an enzyme deficient form of NCL. Since overexpression of Cln3 is associated with cell toxicity in Drosophila, we also needed to characterise whether vector driven Cln3 expression has similar effects, or is associated with localised inflammation. We have been assessing the capacity of an AAV2/9-CMV vector to drive either the reporter gene GFP alone (AAV2/9-CMV-GFP), or a bicistronic vector expressing both GFP and either mouse Cln3 (AAV2/9-CMV-GFP-mCln3) or human CLN3 (AAV2/9-CMV-GFP-hCLN3). This was addressed first in vitro in primary cultures and subsequently in vivo, to determine the cell-type specificity of transduction and whether adverse events are associated with gene delivery. These studies revealed that a high multiplicity of infection was needed to produce even low-level transduction of cells in vitro, that neurons were transduced with higher efficiency than glia, and that bicistronic vectors appeared to express GFP at lower levels. Most importantly, at these titres none of these vectors resulted in any overt cytotoxicity.
We next assessed whether these AAV2/9 vectors were capable of similar transduction at these titres in vivo. This analysis revealed widespread transduction, and an acute astrocytosis and microglial activation that was mostly localised to near the injection tract, but declined with increased survival time. Based on these data we designed a larger study to determine the extent of vector transduction, inflammatory response and cytotoxicity in wildtype and Cln3 deficient mice at two different ages. These mice were injected unilaterally (with 1μl at 1 x 1012 titre) into the striatum at either 3 months (presymptomatic) and 7.5 months of age (when pathological changes are already evident in these mutant mice), with the contralateral uninjected hemisphere serving as a within-animal control. In order to control for the effects of inserting a needle into the brain, and for injecting a volume of fluid into the striatum, additional sham injected (needle inserted, but no injection made) and vehicle injected mice were also generated. All mice were examined histologically at either 2 and 6 months post injection, with a smaller number of mice analysed by quantitative PCR to determine transgene expression. These cohorts of mice have all been sacrificed and the histological analysis is nearing completion.
To conclude, preliminary data suggest that AAV2/9 vectors appear to be capable of widespread and long lasting transduction of the CNS.
Intraocular transplantation of neural stem cells engineered to (co-)express neuroprotective factors
To evaluate the clonal cell lines for potential neuroprotective effects on photoreceptor cells, they were intravitreally grafted into CLN6nclf mice at the onset of retinal degeneration at postnatal day 14. The contralateral eye of each animal received injections of control NS cells and served as a reference. Photoreceptor numbers in NTF-treated and control eyes were determined 2, 4 and 6 weeks after transplantation in six defined regions of the nasal and temporal retina (Fig. 6). Several different NTF-secreting cell lines (i.e. BDNF-, GDNF-, PEDF- and CNTF-secreting NS cell clones) and combinations of BDNF-, GDNF- and PEDF- expressing NS cell lines have been analysed for the effect on retinal degeneration. Only grafted CNTF-expressing NS cells were capable to attenuate the loss of photoreceptor cells in CLN6nclf mice. These results encourage further research into neuroprotective approaches aimed at attenuating retinal degeneration in NCL disease caused by mutations in intracellular transmembrane protein.
Results of this project task have been submitted for publication: Jankowiak W et al. Sustained neural stem cell-based intraocular delivery of CNTF attenuates photoreceptor loss in the nclf mouse model of neuronal ceroid lipofuscinosis.
Fig. 6: Intravitreally grafted CNTF expressing neural stem cells attenuate photoreceptor degeneration in nclf mice.(A) A CNTF expressing NS cell line was intravitreally grafted into 14 days old nclf mice, and retinas were analyzed two (a), four (c) and six (e) weeks after transplantation. The contralateral eye of each animal received injections of control NS cells and served as a reference (b, d, f). A comparison of central retinal sections revealed a significantly thicker photoreceptor layer (onl) in CNTF-treated retinas when compared to the contralateral control retinas. (B) Determination of photoreceptor numbers in six defined areas of the central nasal and temporal retina revealed the presence of significantly more photoreceptor cells in CNTF-treated retinas (filled bars) than in control retinas (open bars) at all post-transplantation intervals analyzed. Each bar represents the mean value (±SEM) from six retinas. ***, p<0.001 (Newman-Keuls post hoc test after the mixed two-way ANOVA)
The retinal CLN6 gene therapy study has shown that gene supplementation of the CLN6 gene to the mouse photoreceptor cells is insufficient to attenuate photoreceptor cell degeneration. Future experiments will aim to target the bipolar cells specifically.
The CLN3 gene therapy study analysed the delivery of CLN3 to the CNS showing that AAV2/9 vectors are capable of widespread and long lasting transduction of the CNS.
The study using genetically modified neural stem cell grafts to deliver neurotrophic factors to the retina has shown that intravitreally grafted CNTF-secreting stem cell lines are capable of attenuating the loss of photoreceptor cells in CLN6nclf mice. These results encourage further research into neuroprotective approaches aimed at attenuating retinal degeneration in NCL disease caused by mutations in intracellular transmembrane protein..