Histological analysis, quantification and statistics

Histological and immunohistological staining was carried out as described in Section 3.8. For Nissl staining, sections from two to four individual CLN5 affected sheep brains were processed at each timepoint to test for inter-animal variation. Sequential sections from one brain of each genotype at each age were then used for all subsequent immunohistochemical and lectin histochemical studies. Cortical thickness measurements, thresholding image analysis methods and statistical analysis are described in Sections 3.9– 3.10. All histological processing and subsequent analyses were performed with no prior knowledge of genotype.

52 Table 4.1 CNS regions and sagittal section levels in which they were investigated (adapted from Oswald, 2004)

Level 1 2 3 4 5

Symbol Lateral Medial

Hippocampus HC ** *** ** ** *

Entorhinal cortex Ent ** *** *

Parieto-occipital cortex POC ** *** * Primary somatosensory cortex S1 ** *** Lateral geniculate nucleus LGN ** ***

Striatum Str ***

Cerebellar cortex CB * ** *** **

Frontal association cortex FA * *** *

Optic tract Opt * *** *

Olfactory cortex Olf * ***

Superior colliculus SC *** **

Corpus callosum Cc ** ***

Thalamus Th ** ***

Caudate nucleus CN ** ***

Primary visual cortex V1 ***

4.3

Results

4.3.1

General organisation and development of the ovine NCL brain

Normal sheep brains grew rapidly from birth to reach an early peak at 4 months. After a 3% decline in mean normal brain mass at 6 months, brain growth recommenced, approaching a plateau at 12 months and mature weight by 24 months (Table 4.2). At birth, affected sheep brains of both genotypes appeared to be normally developed and although divergent in weight from the normal newborn brain, this difference was not significant. The brain masses of CLN5 and CLN6 affected sheep also peaked at 4 months of age, falling behind normal controls at this stage by 11% and 19% respectively, marking the start of progressive brain atrophy. This decline was more obvious in the CLN6 affected brain which was reduced to 69% and 54% of the normal brain weight by 12 and 24 months of age, respectively (Table 4.2). In contrast, the CLN5 affected brain mass changes were less overt, plateauing on average at 79.3  0.8 g from 6 to 12 months, however this converged to 58% of normal brain weight by 24 months of age (Table 4.2).

Table 4.2 Mean brain weight of normal and affected sheep from birth to 24 months of age

Age (m) Normal (g) CLN5-/- (g) CLN6-/- (g) CLN5 / Normal CLN6 / Normal 0 – 1 58.0  1.7* (n† = 11) 55.5  1.4 (n = 2) 52.4  3.2 (n = 9) 96% 90% 2 74.1  2.5 (n = 5) 70.4  2.3 (n = 7) 95% 4 92.1  3.7 (n = 13) 81.8  0.7 (n = 2) 74.7  3.3 (n = 7) 89% 81% 6 89.2  1.7 (n = 17) 79.5  2.2 (n = 5) 70.4  3.6 (n = 6) 89% 79% 9 92.2  2.2 (n = 9) 78.4  1.6 (n = 4) 68.2  2.1 (n = 5) 85% 74% 12 98.2  1.6 (n = 16) 79.5  1.1 (n = 16) 67.7  5.6 (n = 5) 81% 69% 15 76.1  1.8 (n = 7) 66.2  1.9 (n = 8) 18 98.1  1.1 (n = 16) 71.2  1.5 (n = 13) 66.0  2.4 (n = 18) 73% 67% 24 106.1  1.5 (n = 17) 62.0  1.0 (n = 5) 57.5  2.5 (n = 8) 58% 54% * Standard error of the mean (SEM)

† Number of brains collected. One to four brains of each category was used for subsequent analysis No brains were available to be analysed at those ages where values are missing

By 6 months, the disparate effect of the diseases on the cerebral hemispheres was macroscopically apparent, with obvious volume loss in the parietal and occipital lobes, medial and caudal to the suprasylvian sulcus, of both CLN5 and CLN6 affected brains. By 18 months, gross atrophy of the cerebral hemispheres, with concomitant dorsoventrally flattened, narrowed gyri and widened sulci, was evident (Figure 4.2). Whilst the CLN6 affected cerebral cortex was more severely shrunken overall, the CLN5 affected frontal lobe, medial and rostral to the pseudosylvian sulcus, was discernibly more atrophied than that in the CLN6 affected brain. In contrast to the marked cerebral

54 atrophy, the cerebellum and subcortical structures of both genotypes retained normal appearance, even at 24 months.

4.3.2

Regional atrophy and cortical thinning

The unifying pathological hallmark of NCL is severe neurodegeneration. In order to quantify this, Nissl stained sections from CLN5 and CLN6 diseased sheep brains were analysed for neuronal cytoarchitecture and by measurement of cortical thickness, to compare spatiotemporal changes with disease progression (Figure 4.3). Distinct neuronal laminae were evident across the normal cortical mantle (Figure 4.4), with pronounced thickening of the normal cortical layers over time to reach maturity by 12 months (Figure 4.5). In contrast, the NCL affected ovine cerebral cortices never attained full maturity; instead regionally specific cortical thinning and neurodegeneration commenced. Preceding or coinciding with the neuronal loss was perturbation of the cytoarchitectonic layers and the progressive appearance of clusters of densely packed cellular aggregates at the layer I/II boundary throughout the affected cerebral cortex of both genotypes. Control sheep brains did not contain these cellular aggregates, except in the entorhinal cortex. In all Figure 4.2 Lateral view of the normal and affected sheep brain

Perfused brains from A. a 19.4 month old normal Coopworth sheep, B. a 19.2 month old CLN5 affected Borderdale and C. a 19.3 month old CLN6 affected South Hampshire. Note the marked atrophy of the diseased cerebral hemispheres and the relative sparing of the cerebella. Brain weights were 99, 74.86 and 60.75 g respectively. The suprasylvian sulcus (▲), pseudosylvian sulcus (■), parietal (Par), occipital (Occ) and frontal (Fr) lobes are indicated.

A. C. B. ▲▲▲▲ Par Occ Fr

affected cortical regions, the neuronal atrophy was more pronounced in the upper layers (II-III) with the lower cortical layers (V-VI) better preserved, especially the lamina V pyramidal cells.

It was previously reported that the cytoarchitecture of the cerebellum, hippocampus and all cortical regions of the CLN6 affected brains appear normal at birth (Oswald et al., 2005; Kay et al., 2006). However, even by the earliest age in the current study (2 months), loss of layer definition was observed in the CLN6 affected primary visual and parieto-occipital regions with aggregation of cells at the I/II laminar boundary (Figure 4.4). These degenerative changes spread from the visual and parieto-occipital cortices to the somatosensory cortex at 6 months, reaching the motor cortex by 12 months, and the entire cortical mantle by 18 months, corroborating the findings of Oswald et al (2005). Cortical thickness measurements demonstrated that the rate of active thinning differed between the regions but followed a similar pattern (Figure 4.5), to the cytoarchitectural changes albeit delayed. The visual and parieto-occipital regions were affected most and earliest (from 2 months). The thickness of the CLN6 affected somatosensory cortex increased comparatively normally up to 9 months, delayed atrophy commenced in the entorhinal and frontal cortices from 12 months, and the motor cortex was relatively spared until 18 months of age. By 24 months, significant atrophy was seen across the CLN6 affected cerebral cortex with the primary visual, entorhinal, and primary motor areas reduced to 37%, 56%, and 61% of the respective normal thicknesses.

The CLN5 affected brain was also normal in appearance at birth, except for a mild disturbance in the laminar architecture and small cellular clusters at the layer I/II interface of the primary visual cortex. The same clusters and loss of layer definition reached the parieto-occipital and somatosensory CLN5 cortices by 4 months of age. Degenerative changes in the CLN5 entorhinal, frontal association and motor cortices were detected from 6 - 9 months, some 3 - 6 months earlier than in the CLN6 affected brain. However, despite the earlier cortical laminar reorganisation in the CLN5 brain, the affected brains of both genotypes reached the same pathological endpoint at 24 months of age with obvious layer I/II cellular aggregates and few cortical neurons remaining (Figure 4.4).

As in the CLN6 model, active thinning in the CLN5 affected brain became apparent at different ages in different cortical regions. There was a near linear decline in the visual, parieto-occipital and somatosensory cortices from birth. Atrophy became apparent in the frontal cortex from 9 months and was not apparent in the entorhinal and motor cortex until 15 months (Figure 4.5). The neurodegenerative cascade began earlier in most regions of the CLN5 affected brain than in the CLN6 affected brain, however, despite this earlier onset of atrophy, cortical thickness measurements from CLN5 and CLN6 affected sheep brains converged, and were very similar by 24 months (Figure 4.5). The major point of difference was seen in the cortices of the frontal lobe. At 24 months there was an 11% greater reduction in the thickness of the CLN5 frontal association cortex compared to that of the

56 CLN6 brain (P  0.0001). Additionally the motor cortex of the CLN5 affected sheep brain shrank dramatically from 15 months of age to be 59% that of normal thickness by 18 months, a 20% greater reduction than in the same region in an age-matched CLN6 affected animals. This discrepancy was still statistically significant (P = 0.0006) at 24 months when the CLN5 and CLN6 motor cortex thicknesses were reduced to 54% and 61% of normal respectively (Figure 4.4).

Progressive atrophic changes in the cortical grey matter were accompanied by white matter tract changes. Occipital white matter loss with disease progression was macroscopically overt (Figure 4.3). The thickness of the corpus callosum in normal brains increased to plateau at 1323  30 μm (n = 7) by 18 - 24 months, whereas it remained relatively unchanged in NCL affected sheep throughout postnatal development (CLN6 affected 724  30 μm (n = 5); CLN5 affected 756  18 μm (n = 8)). The subcortical structures were remarkably preserved in sharp contrast to the gross atrophy of the cortex in both ovine disease models. Nissl staining revealed no overt depletion of cells in the affected thalamic nuclei, colliculi or striatum and neuronal populations in these resembled those seen in control sections. Cerebellar cortical thickness was also relatively constant from birth to 24 months in normal (513  5 μm; n = 10), CLN5 affected (511  5 μm; n = 18) and CLN6 affected (495  7 μm; n = 5) sheep brains, demonstrating the relative sparing of the cerebellum in ovine CLN5 and CLN6 NCLs (Figure 4.5).

Figure 4.3 Progressive cortical atrophy in CLN5 and CLN6 affected sheep

Nissl stained sagittal sections (Level 5) show gross atrophy of the cerebral cortex in CLN5 and CLN6 affected sheep at 12 months, which is more pronounced at 24 months, especially in the visual cortex. The vulnerability of the motor cortex is also highlighted in CLN5 affected sheep. In contrast, the cerebellum is relatively spared in both disease models.

CLN5 Affected