OUP user menu

Synaptic Changes in the Thalamocortical System of Cathepsin D-Deficient Mice
A Model of Human Congenital Neuronal Ceroid-Lipofuscinosis

Sanna Partanen PhD, Aleksi Haapanen BM, Catherine Kielar MSc, Charles Pontikis PhD, Noreen Alexander MSc, Teija Inkinen MS, Paul Saftig PhD, Thomas H. Gillingwater PhD, Jonathan D. Cooper PhD, Jaana Tyynelä PhD
DOI: http://dx.doi.org/10.1097/nen.0b013e31815f3899 16-29 First published online: 1 January 2008


Cathepsin D (CTSD; EC is a lysosomal aspartic protease, the deficiency of which causes early-onset and particularly aggressive forms of neuronal ceroid-lipofuscinosis in infants, sheep, and mice. Cathepsin D deficiencies are characterized by severe neurodegeneration, but the molecular mechanisms behind the neuronal death remain poorly understood. In this study, we have systematically mapped the distribution of neuropathologic changes in CTSD-deficient mouse brains by stereologic, immunologic, and electron microscopic methods. We report highly accentuated neuropathologic changes within the ventral posterior nucleus (ventral posteromedial [VPM]/ventral posterolateral [VPL]) of thalamus and in neuronal laminae IV and VI of the somatosensory cortex (S1BF), which receive and send information to the thalamic VPM/VPL. These changes included pronounced astrocytosis and microglial activation that begin in the VPM/VPL thalamic nucleus of CTSD-deficient mice and are associated with reduced neuronal number and redistribution of presynaptic markers. In addition, loss of synapses, axonal pathology, and aggregation of synaptophysin and synaptobrevin were observed in the VPM/VPL. These synaptic alterations are accompanied by changes in the amount of synaptophysin/synaptobrevin heterodimer, which regulates formation of the SNARE complex at the synapse. Taken together, these data reveal the somatosensory thalamocortical circuitry as a particular focus of pathologic changes and provide the first evidence for synaptic alterations at the molecular and ultrastructural levels in CTSD deficiency.

Key Words
  • Cathepsin D
  • Electron microscopy
  • Lysosome
  • Neurodegeneration
  • Stereology
  • Synapse
  • Thalamus


Cathepsin D (CTSD; EC is a lysosomal aspartic protease that, like many other lysosomal enzymes, is synthesized as an inactive precursor and proteolytically processed to yield the mature protein (1,2). In addition to its enzymatic function, CTSD has a complex, but poorly characterized, role in both cell death and proliferation, and it has been suggested to function in a novel lysosome-associated apoptosis-like cell death pathway (3,4).

Mutations leading to complete inactivation of CTSD underlie congenital neuronal ceroid-lipofuscinosis (NCL) in human infants and lambs (5,6). These are fatal neurodegenerative disorders characterized by extreme brain atrophy and death within the first days of life (5,6). In contrast, CTSD-deficient (CTSD−/−) mice are normal at birth but develop severe neurologic signs, including tremor, epileptic seizures, and motor problems, beginning at the age of 2 weeks (7). Subsequently, these mice also develop atrophic changes in their lymphoid system and small intestine; they die prematurely at the age of 26 ± 2 days (8). Cathepsin D-deficient mice exhibit the main neuropathologic characteristics of NCLs, including neuronal deposition of autofluorescent storage material with granular or fingerprint-type ultrastructure (7,9,10), and immunohistologic evidence for the accumulation of subunit c of the mitochondrial adenosine triphosphate synthase (7). In the brains of CTSD−/− mice, neuronal loss is accompanied by pronounced microglial activation (11). Activated microglia are capable of producing nitric oxide, which has been suggested to play an important role in the neuronal apoptosis and intestinal necrosis in CTSD−/− mice (11). However, inhibition of nitric oxide synthesis provided only a modest improvement in the life span of these mice (11). The involvement of autophagy in neuronal degeneration and storage deposition has also been implicated in CTSD−/− mice (7), but despite active research, the molecular mechanisms involved in the neurodegeneration caused by CTSD deficiency remain obscure, and more detailed information about the events in different brain areas is needed to gain insights into disease pathogenesis.

To gain a better understanding of the cellular and molecular mechanisms involved in the pathogenesis of CTSD deficiency, we have systematically mapped neuropathologic alterations in the brains of CTSD−/− mice using a combination of unbiased stereology, immunologic techniques, and electron microscopy. We report a series of localized neurodegenerative and reactive events within the somatosensory thalamocortical system of CTSD−/− mice and provide the first evidence for synaptic abnormalities in the NCLs at both the molecular and ultrastructural levels.

Materials and Methods


Cathepsin D-deficient mice produced by Saftig et al (8) were maintained on a mixed C57BL6J strain background in the animal facility of the Helsinki University, Biomedicum, where food and water were available ad libitum, and light-dark cycle was 12-12 hours. The study protocol was approved by the ethical committee of Helsinki University.


The following mouse monoclonal antibodies were used for immunohistochemical and immunofluorescence staining of paraffin sections (IHP) or cryosections (IHC) and Western blotting (WB): α-Synaptobrevin (Stressgen Biotechnologies, Victoria, Canada, 1:1000 for IHP and WB; Chemicon International, Temecula, CA, 1:500 for IHC), α-syntaxin (Stressgen Biotechnologies, 1:2000 for IHP and 1:1500 for WB; Chemicon International, 1:250 for IHC), α-synaptophysin (Stressgen Biotechnologies, 1:200 for IHP, 1:750 for WB, and, 1:1000 for IHC), and α-synaptosomal-associated protein of 25 kd (synaptosomal-associated protein of 25 kd [SNAP25]; BD Transduction Laboratories, Lexington, KY; 1:1500 for IHP, 1:1000 for WB and IHC) or the microglial markers F4/80 (monoclonal rat anti-mouse F4/80; Serotec, Oxford, UK; 1:100 for IHC) and CD68 (clone PG-M1; DAKO, Cambridge, UK; 1:100 for IHC and IHP). Polyclonal antisera against the calcium-binding protein parvalbumin (PV) (polyclonal rabbit anti-PV; Swant, Bellinzona, Switzerland; 1:5000) and the astrocytic marker glial fibrillary acidic protein (GFAP; DAKO, Cambridge, UK; 1:5000) were used for IHC. Polyclonal antisera against the lysosome-associated membrane protein 1 (Lamp-1; Sigma-Aldrich, Steinheim, Germany; 1:500) and the autophagosome marker light chain 3 of neuronal microtubule-associated protein 1A/B (LC3; a kind gift of professor Takashi Ueno, Juntendo University Medical School, Tokyo, Japan; 1:100) were used for immunofluorescence.

Histologic Processing

For light microscopy analysis of cryopreserved samples, brains were dissected from CTSD−/− mice and age-matched control+/+ littermates at postnatal day (P) 16, P20, and P23 ± 1 (n = 5-6 for P23 ± 1, n = 2 for P16, and n = 3 for P20). The brains were fixed for 24 hours in a solution containing 4% paraformaldehyde in PBS (pH 7.4), cryoprotected at 4°C in Tris-buffered saline (TBS; 50 mmol/L Tris, pH 7.6) containing 30% sucrose and 0.05% NaN3. Brains were subsequently processed and sectioned as described previously (12) with serial 40-μm frozen coronal and sagittal sections stored at −40°C in cryoprotectant solution (TBS/30% ethylene glycol/15% sucrose/0.05% sodium azide) in a 96-well plate well before histologic and immunohistologic processing of the free-floating sections. For paraffin-embedded samples, brains from 23 ± 1-day-old CTSD−/− and CTSD+/+ mice (n = 4) were fixed in 4% neutral-buffered formaldehyde and subsequently processed, embedded in paraffin, and cut into 4-μm-thick sections.

Nissl Staining

Every fifth section was mounted onto gelatin-chrome alum-coated Superfrost microscope slides (VWR, Dorset, UK), air-dried overnight, and incubated for 45 minutes at 60°C in a solution of 0.05% Cresyl fast violet and 0.05% acetic acid (VWR). Stained sections were then rinsed in distilled water and differentiated through alcohol series before clearing in xylene (VWR) and coverslipping with [p-xylene-bis(pyridinium bromine)] (VWR).

Immunohistochemical and Immunofluorescence Stainings

Every sixth 40-μm free-floating cryosection was stained for markers of astrocytosis (GFAP and F4/80; n = 5), synapses (synaptobrevin [Syb], synaptophysin [Syp], SNAP25, and syntaxin [Syx]; n = 6), or PV expressed in GABAergic neurons in the reticular thalamic nucleus (n = 5). Sections were first incubated in 1% H2O2 in TBS for 30 (glial markers) and 15 minutes (synaptic markers), rinsed in TBS, and blocked with 15% normal serum in TBS for 40 minutes. Sections were incubated with primary antibody diluted in TBS containing 0.3% Triton X-100 and 10% normal serum overnight at 4°C. After washing, sections were incubated in biotinylated secondary antisera (Vector Laboratories, Peterborough, UK; 1:2000) in TBS containing 0.3% Triton X-100 containing 10% normal serum for 2 hours at 4°C. After washing, sections were incubated with an avidin-biotin-peroxidase complex in TBS (Vectastain Elite ABC kit; Vector Laboratories) for 2 hours. Sections were then rinsed in TBS, and immunoreactivity was visualized by incubation in 0.05% 3,3′-diaminobenzidine HCl (Sigma, Dorset, UK) and 0.001% H2O2 in TBS for 6 to 20 minutes (a range of time that represented saturation for this reaction). Sections were then transferred to excess ice-cold TBS and rinsed several times, mounted, air-dried, cleared in xylene, and coverslipped with [p-xylene-bis(pyridinium bromine)] (VWR). To obtain reproducible and quantitative results, particular care was taken to fully saturate the color reaction and to treat all sections stained for the same antigen as 1 batch.

Paraffin-embedded CTSD−/− and control mouse brain sections were immunohistochemically stained for synaptic markers (Syb, Syp, SNAP25, and Syx). Sections were dewaxed in xylene, antigen retrieval was performed by microwaving in citric acid buffer, and endogenous peroxidase activity was blocked by incubating the sections in methanol containing 1.6% H2O2 at room temperature for 30 minutes. Before mounting, the sections were counterstained with hematoxylin.

For double labeling immunofluorescence studies, the tissue sections were treated with the primary antibodies after blocking as above, and the primary antibodies were detected using species-specific fluorescent secondary antibody conjugates (Alexa Fluor 488 and Alexa Fluor 568; Molecular Probes Inc., Eugene, OR). The double stainings were visualized using Leica TCS SP2 AOBS laser scanning confocal microscope with 488- and 561-nm excitation lines (Leica Microsystems, Wetzlar, Germany). The images were acquired using sequential scanning to avoid channel cross talk. The double stainings for Syb (with 3-amino-9-ethylcarbazole peroxidase substrate) and CD68 (with Alexa Fluor 488 conjugates) were visualized using an Olympus AX70 microscope (Olympus Optical, Co., Hamburg, Germany).

Volume Estimations

Cavalieri estimates of the volumes of selected brain regions were made in Nissl-stained sections using StereoInvestigator software (MicroBrightField, Inc., Williston, VT). Briefly, an appropriately spaced sampling grid (150 μm for cortex, thalamus, corpus callosum, cerebellar gray matter, and white matter; 100 μm for hippocampus and striatum; 50 μm for the deep cerebellar nuclei), was randomly superimposed over sections, and the number of points covering each central nervous system (CNS) region was counted from 1 hemisphere, and the corresponding regional volume was calculated as described previously (12). All cerebellar regions were measured from 12 consecutive sections, thus, not representing the total volume of these regions, but giving a good estimate of differences between CTSD−/− and control mice.

Thickness Measurements

Cortical thickness measurements were made in Nissl-stained sections through different areas of cerebral cortex, including prefrontal cortex, primary motor cortex (M1), barrel-field area of the primary somatosensory cortex (S1BF), and primary visual cortex, as defined by Paxinos and Franklin (13). This analysis was performed by drawing 10 lines through the cortex to 3 adjacent sections in which the examined cortical area was present, with the length of each perpendicular line extending from the white matter to the pial surface measured using ImageJ software. Similarly, the thicknesses of individual cortical laminae were measured in M1 and S1BF from 3 adjacent Nissl-stained sections, with individual measures obtained for laminae I, IV (S1BF only), V and VI, whereas a combined measure was taken for laminae II and III.

Threshold Analysis of Syb, Syp, SNAP25, and Syx Thalamus

The optical density of Syb, Syp, SNAP25, and Syx immunoreactivity was assessed using a semiautomated thresholding image analysis system (Image Pro Plus; Media Cybernetics, Silver Springs, MD). Analysis was performed blind to genotype, as previously described (12). Forty nonoverlapping images were captured through the thalamus on 3 consecutive sections via a live video camera (JVC, 3CCD, KY-F55B), mounted onto a Zeiss Axioplan microscope using a ×40 objective, and saved as JPEGs. All parameters, including lamp intensity, video camera setup, and calibration, were maintained constantly throughout image capturing.

Images were subsequently analyzed using Image Pro Plus software and an appropriate threshold that selected the foreground immunoreactivity above background. This threshold was then applied as a constant to all subsequent images analyzed per batch of animals and reagent used to determine the specific area of immunoreactivity for each antigen in thalamus. Each field measured 120 μm wide, with a height of 90 μm. Therefore, the total area compiled from 40 fields in thalamus corresponded to 432,000 μm2. Macros were recorded to transfer the data to an Microsoft Excel spreadsheet for statistical analysis. Data were plotted graphically as the mean percentage area of immunoreactivity per field ± SEM.

Line Profile Analysis of Immunohistochemical Stainings

The distribution of immunoreactivity for GFAP, F4/80, and SYP was analyzed in S1BF using a line profile method. Images were captured using a DAGE-MTI CCD-100 camera (DAGE-MTI Inc., Michigan City, IN) from 3 adjacent sections of S1BF. Each image was 1,300 × 1,030 pixels at a resolution of 150 dots per inch. These images were filtered in ImageJ software (Wayne Rasband, National Institutes of Health) using Gaussian blur filter with radius of 8 pixels. Ten lines were drawn from lamina I to lamina VI upon every image, and the relative pixel brightness was measured along the length of each line using a scale of 1to 255, where white = 1 and black = 255. Data from the seline profiles from individual animals were exported to an Excel spreadsheet for subsequent analysis. The length of each individual line was normalized to correct for corticalatrophy. Finally, the brightness value in each pixel was inverted to calculate the intensity of the staining because the staining intensity (SI) is inversely correlated to the brightness of the image. A histogram composed of 100bins was then calculated, with each bin representing 1%of the length of the original line, and the SI in each bin representing the corresponding mean SI value. The results are expressed as percentage difference between the CTSD−/− and control mice as follows: 100% × (SICTSD−/−− SICTSD+/+) / SICTSD+/+.

Measurement of Total Neuronal Number

Stereo Investigator software was used to obtain optical fractionator estimates of the number of Nissl-stained somatosensory relay neurons in the ventral posterior thalamic nucleus (VPM/VPL), PV-positive inhibitory neurons in the reticular thalamic (Rt) nucleus, and Nissl-stained cortical neurons in laminae IV and VI of S1BF. These optical fractionator estimates were obtained as previously described (14), with a random starting section chosen, followed by every sixth Nissl-stained or PV-stained section thereafter. Only neurons with a clearly identifiable nucleus were sampled, and all counts were performed using a ×100 oil objective (NA 1.4). The following sampling scheme was applied to each region of interest: VPM/VPL dissector frame, 74 × 45 μm; grid area, 175 × 175 μm; Rt dissector frame, 74 × 24 μm; grid area, 100 × 100 μm; S1BF lamina IV and lamina VI dissector frame, 41 × 26 μm; grid area, 175 × 175 μm.

Data Analysis

Statistical significance (p) of the altered mean values of the structural, threshold, line profile, and neuron number data was analyzed with SPSS v12.0 (SPSS, Inc., Chicago, IL) and Microsoft Excel programs using Student t-test. The volume data were expressed in cubic millimeters and thickness values in micrometers. The SIs of the line profile analysis were expressed in arbitrary SI units (see above). Values of all quantitative results are represented as mean ± SEM. The mean coefficient of error for all individual optical fractionator and Cavalieri estimates was calculated according to the method of Gundersen and Jensen (15) and was less than 0.08 in all these analyses.

Electron Microscopy

Mice were anesthetized using subcutaneous injections of ketamine (Orion Pharma, Espoo, Finland; 75-100 mg/kg) and killed by perfusion fixation with 0.1 mol/L phosphate buffer containing 4% paraformaldehyde and 2.5% glutaraldehyde. Brains were removed and immersed in fresh fixative for 24 hours, washed, and stored in 0.1 mol/L phosphate buffer before cutting free-floating 70-μm coronal sections. Sections containing regions of interest (VPM/VPL and S1BF) were postfixed in 1% osmium tetroxide in 0.1 mol/L phosphate buffer for 45 minutes. After dehydration through an ascending series of ethanol and propylene oxide, all sections were embedded on glass slides in Durcupan resin. Regions of interest (~1 × 1 mm) were then cut from a randomly selected section using a scalpel and glued onto a resin block. Ultrathin sections (~60 nm) were cut and collected on formvar-coated grids (Agar Scientific, Stansted, UK), stained with uranyl acetate and lead citrate in an LKB "Ultrostainer," and then quantitatively assessed in a Philips CM12 transmission electron microscope. A timed-counting protocol was used to estimate synaptic density in regions of interest, as described previously (16). Briefly, ultrathin sections containing the region of interest were placed on coded grids, and the number of synaptic profiles observed during a 15-minute analysis period were recorded for 2 separate counting sessions per grid. The total distance traveled over each section during each counting period was measured to ensure that a similar total area had been included in each analysis. Analysis was performed blind to genotype. Raw data were collated using Excel spreadsheets, and statistical tests were performed using Graphpad Prism software.

Western Blotting

Synaptic protein complexes were analyzed from the cerebral cortices and thalami of control and CTSD−/− mice by homogenizing the tissue in 40 mmol/L Tris-HCl, pH 8.8, containing protease inhibitors. Homogenates were then centrifuged at 1,700 × g for 10 minutes at 4°C, and the protein concentration was determined using a bicinchoninic acid protein kit (Interchim, Montluçon, France). Under nonreducing conditions, 20 μg from each homogenate was subjected to 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel. The gels were blotted onto PVDF membranes, blocked with 4% (wt/vol) bovine serum albumin in TBS containing 0.1% Tween 20 and incubated with the primary antibodies overnight at 4°C. Immunoreactive bands were visualized by enhanced chemiluminescence after incubation with goat anti-mouse immunoglobulin G coupled to horseradish peroxidase (Bio-Rad, Hercules, CA; 1:4000).


CTSD−/− Brains Show Widespread Atrophy

Macroscopically, the brains of CTSD−/− mice seemed relatively normal but proportionately smaller than those of controls at the age of 24 days. Consistent with this observation, Nissl-stained sections revealed widespread atrophy of the CNS in the affected mice (Fig. 1A). Cavalieri estimates of regional volume showed significant effects upon both white and gray matter of CTSD−/− mice: Of all structures, the corpus callosum was the most significantly affected in mutant mice, being reduced by 32% compared with controls (Fig. 1B). Each cerebral region examined, including the striatum (−22%), thalamus (−22%), hippocampus (−19%), and cortex (−16%), displayed a markedly decreased volume in CTSD−/− mice, as did the deep cerebellar nuclei (−24%) and cerebellar white matter (−22%) (Fig. 1B). Of all structures in mutant mice, the cerebellar gray matter was least affected (−12%) (Fig. 1B).


Regional atrophy of the cathepsin D (CTSD)−/− CNS. (A) Low-power micrographs of Nissl-stained parasagittal sections reveal the overall atrophy of the CNS in CTSD−/− mice compared with controls+/+. (B) Cavalieri estimates of regional volume indicate the extent of atrophy in grey and white matter structures of CTSD−/− mice compared with controls. Data are expressed as mean regional volume in μm3 ± SEM. *, p ≤ 0.05; †, p ≤ 0.001; ‡, p ≤ 0.000. Scale bar = (A) 2,000 μm.

Cortical Thinning in CTSD−/− Mice is the Result of Lamina-Specific Changes

To determine whether CTSD deficiency affected cortical subfields that serve different functions to a similar extent, we surveyed the thickness of prefrontal cortex, M1, S1BF, and visual cortex, and also measured the individual laminar thicknesses within these regions (Fig. 2A). Statistically significant thinning was evident in most cortical subfields but occurred to different extents in S1BF (−12%), visual cortex (−11%), M1 (−6%), and PF (not significant) (Fig. 2B).

These changes in overall cortical thickness were accompanied by distinct effects on the thickness of individual laminae, which occurred to different extents in each cortical subfield. In M1 of CTSD−/− mice, laminae I, V, and VI were of normal thickness, whereas the combined thickness of laminae II and III was significantly reduced (by 15%) compared with controls (Fig. 2B). In S1BF, the thickness of superficial laminae I to III and lamina V were indistinguishable between genotypes. In contrast, laminae IV (−20%) and VI (−23%) were markedly and significantly thinner in CTSD−/− mice compared with controls (Fig. 2B).


Cortical lamina-specific thinning in cathepsin D (CTSD)−/− mice. (A) Nissl-stained sections from 2 representative cortical regions, primary barrel field of primary somatosensory cortex (S1BF), and primary motor cortex (M1) reveal the significant atrophy of these regions in CTSD−/− compared with controls+/+. (B, upper panel) Measurements of total cortical thickness show significant atrophy of S1BF, M1, and primary visual cortex (V1), but not prefrontal cortex in CTSD−/− mice. (B,lower panels) Measurements of individual laminar thickness, using the boundaries depicted in (A), indicate the different lamina-specific effects in S1BF and M1. Thickness data are expressed as mean cortical thickness (in micrometers ± SEM). *, p ≤ 0.05; †, p ≤ 0.000. Scale bar = (A) 500 μm. M1, primary motor cortex; PF, prefrontal cortex; S1BF, somatosensory cortex; V1, primary visual cortex.

Localized Microglial Activation is Accompanied by Widespread Astrocytosis in CTSD−/− Mice

Staining for F4/80 revealed that microglial activation showed extreme regional accentuations in the brain of CTSD−/− mice, being most pronounced within the thalamus and olivary nuclei. F4/80-immunoreactive microglia were present in the ventrolateral part of the anteroventral thalamic nucleus, and in the VPM/VPL and posterior thalamic nuclei and the lateral superior olive, rostral, and dorsal periolivary region, but were virtually absent from adjacent nuclei (Fig. 3A). In addition, F4/80-positive microglia were present in the substantia nigra and also formed 2 densely staining bands immediately dorsal and ventral to the corpus callosum of CTSD−/− mice. Only faint staining for F4/80 was observed in the brains of control mice (Fig. 3A).


Glial activation and enhanced synaptic staining in cathepsin D (CTSD)−/− mice. (A) Immunohistochemical staining of parasagittal brain sections shows the widespread activation of astrocytes (glial fibrillary acidic protein [GFAP]), restricted activation of microglia (F4/80), and enhanced staining for a synaptic marker (synaptophysin [Syn]) in CTSD−/− mice. The glial activation is particularly marked within the thalamus (anteroventral ventrolateral thalamic nucleus, white arrow; ventral posteromedial thalamic nucleus [VPM]/ventral posterolateral thalamic nucleus [VPL], black arrow) and olivary region (arrowhead) of CTSD−/− mice. In age-matched controls+/+, the GFAP-positive fibrous astrocytes are confined to the white matter, and the controls displayed minimal F4/80 immunoreactivity. Immunohistochemical staining for Syp reveal increased immunoreactivity in CTSD−/− mouse brains compared with controls. Scale bar = 2,000 μm. (B) Representative images of GFAP-, F4/80-, and Syp-stained sections through the somatosensory cortex [S1BF] of CTSD−/− (left panels) and control (right panels) mice, with the corresponding line profile analysis. The line profile shows the pixel intensity of the respective immunohistochemical stainings along the cortex, expressed in percentage difference between CTSD−/− and control mice (see Materials and Methods section for details). Note the increase in GFAP and F4/80 stainings in laminae IV and VI of CTSD−/− mice, where there was a coinciding relative decrease in SYP staining. The cortical samples are set to equal thickness to allow better comparison of the staining pattern in CTSD−/− and control animals. Scale bar = 500 μm. (C) Optical fractionator estimates of neuronal number in laminae IV and VI of S1BF and in the VPM/VPL and reticular (Rt) nuclei of CTSD−/− and CTSD+/+ mice. Optical fractionator data are expressed as mean estimated neuronal number ± SEM. *, p ≤ 0.001, †, p ≤ 0.01.

In contrast to microglial activation, immunohistochemical staining for GFAP showed widespread astrocytosis in CTSD−/− mouse brains (Fig. 3A). However, there were also marked differences in the extent of astrocytosis in different CNS regions. Glial fibrillary acidic protein immunoreactivity was particularly prominent in the individual thalamic and olivary nuclei that displayed intense F4/80 immunoreactivity (Fig. 3A). Within the cortex, GFAP-immunoreactive astrocytes were not distributed evenly but were present more abundantly in laminae IV and VI. In age-matched controls, GFAP-positive astrocytes were observed in limited brain regions only, including the white matter tracts corpus callosum and fornix, hippocampus, and the cerebellar white matter (Fig. 3A).

Colocalization of Glial Activation, Synaptic Changes, and Neuron Loss in the Thalamocortical System of CTSD−/− Mice

Astrocytosis and microglial activation are often considered markers of ongoing neurodegeneration (17). Therefore, we investigated neuron survival in VPM/VPL, the thalamic nucleus that displayed the most intense GFAP and F4/80 immunoreactivity (Fig. 3A), and its target cortical region, the primary S1BF. Optical fractionator estimates of neuronal number revealed a marked and significant loss of neurons in VPM/VPL (−41%; p = 0.000) of CTSD−/− mice compared with controls, but no loss of PV-positive neurons in the neighboring reticular nucleus, which provides inhibitory input to the other thalamic nuclei (Fig. 3C). In S1BF, there was a significant reduction (−27%; p = 0.000) in the number of granule neurons in lamina IV, which receive afferent input from VPM/VPL. Lamina VI of S1BF, which provides feedback projections to VPM/VPL, showed a tendency toward reduced neuronal number (−10%), but the decrease was statistically not significant (p = 0.167; Fig. 3C).

To determine whether these effects upon neuron survival were accompanied by synapse loss within the thalamocortical system of CTSD−/− mice, we stained sections for the presynaptic marker, Syp. Unexpectedly, we observed an overall increase in the intensity of Syp staining in the cerebral cortex of CTSD−/− mice compared with controls (Fig. 3A). Indeed, the laminar distribution of Syp immunoreactivity, apparent in control mice, was lost in the mutant mice (Fig. 3A). To further investigate this redistribution of Syp expression and its potential relationship to the lamina-specific activation of glial cells, we used line profile analysis to compare the intensity and distribution of Syp, GFAP, and F4/80 immunoreactivity in S1BF, which receives afferent input from VPM/VPL. In S1BF of control mice, GFAP and F4/80 staining were virtually absent, and Syp staining increased gradually toward the deeper parts of the cortex (Fig. 3B). In contrast, within the S1BF of CTSD−/− mice, GFAP staining increased in a dorsoventral manner and exhibited 2 distinct intensity peaks, similar to F4/80 staining (Fig. 3B). Synaptophysin staining, although generally elevated in CTSD−/− mice, also showed 2 bands of relatively reduced SI, which coincided perfectly with the bands of increased GFAP SI in laminae IV and VI (Fig. 3B). These data indicate that in the S1BF of CTSD−/− mice, glial activation is accompanied by redistribution of the synaptic marker, Syp, in a lamina-specific manner.

Developmental Course of the Neuropathologic Changes

To clarify the development of the neuropathologic alterations, we examined CTSD−/− and control mouse brains at different ages, P16, P20, and P23 ± 1. Glial activation was first evident in the VPM/VPL of CTSD−/− mice: GFAP staining was already visible as GFAP-positive spots within the thalamus at P16 and became pronounced within the VPM/VPL of CTSD−/− mice by P20 (Fig. 4A). At P16 and P20, abnormal GFAP staining was restricted to the thalamus in CTSD−/− mice. By P24, GFAP staining had spread throughout the thalamus and the whole brains of CTSD−/− mice, whereas it was absent from the thalamus of control mice (Fig. 4A). Microglial activation, as indicated by F4/80 staining, was minimal in CTSD−/− mice at P16 but became more evident in VPM/VPL at P20, and by P24, it had spread also to other thalamic nuclei (Fig. 4B). In control mice, F4/80 staining was practically absent in all of these time points (Fig. 4B). Based on these data, it seems that activation of astrocytes slightly precedes that of microglia in the thalamus of CTSD−/− mice. The number of neurons within the VPM/VPL was significantly reduced in CTSD−/− mice at P20 and P24 but not at P16 (Fig. 4C), indicating that loss of neurons coincided with microglial activation in this thalamic nucleus.


Progressive astrocytosis, microglial activation, and neuronal loss in cathepsin D (CTSD)−/− mice. (A) Immunohistochemical staining for glial fibrillary acidic protein (GFAP) reveals the pronounced upregulation of this marker of astrocytosis with increased age in the ventral posteromedial thalamic nucleus (VPM)/ventral posterolateral thalamic nucleus (VPL) of CTSD−/− mice compared with age-matched controls+/+. Glial fibrillary acidic protein-immunoreactive astrocytes formed positive spots within the thalamus at P16. The staining became pronounced but was still confined within the VPM/VPL at P20, spreading through the thalamus with increasing intensity at P24. In control mice, there was minimal GFAP staining within the thalamus. (B) Localized microglial activation within the thalamus of CTSD−/− mice is revealed by F4/80 immunoreactivity. In these mutant mice, F4/80 staining was minimal at P16. More intensely stained F4/80-positive microglia were present in the VPM/VPL at P20, and the staining then spread to other regions of the thalamus at P24. F4/80 staining was virtually not present in age-matched controls+/+. (C) Optical fractionator estimates of neuronal number in the VPM/VPL thalamic nucleus of developing CTSD−/− and CTSD+/+ mice. Optical fractionator data are expressed as mean estimated neuronal number ± SEM. *, p ≤ 0.001.

The Amount of Presynaptic Proteins Is Reduced Within the Thalamus of CTSD−/− Mice

Using the somatosensory thalamocortical system as an example of interconnected neuronal pathways that display both neuronal loss and glial activation in CTSD−/− mice, we next extended our analysis to include the expression of the presynaptic proteins Syp, Syb, SNAP25, and Syx in VPM/VPL. Immunohistochemical stainings revealed that SNAP25 and Syx immunoreactivity were markedly reduced in VPM/VPL of CTSD−/− mice (Fig. 5A), as confirmed by thresholding analysis of low-power images through the whole thalamus (Fig. 5B). Although Syp immunoreactivity was also locally reduced within VPM/VPL, it was apparently increased in the adjacent thalamic nuclei (Fig. 5A). Because of these opposing effects, the alterations in Syp SI did not reach statistical significance (Fig. 5B). Synaptobrevin staining was not reduced in the thalamus of CTSD−/− mice.


Reduced immunohistochemical staining for presynaptic proteins and loss of neurons in ventral posteromedial thalamic nucleus (VPM)/ventral posterolateral thalamic nucleus (VPL) of cathepsin D (CTSD)−/− mice. (A) Representative images of coronal cryosections showing the ventral posterior (VPM/VPL) thalamic nucleus of CTSD−/− and littermate control+/+ mice stained with antibodies against synaptobrevin (Syb), synaptophysin (Syp), synaptosomal-associated protein of 25 kd (SNAP25), and syntaxin (Syx). Scale bar = 1000 μm. (B) Corresponding quantitative image analysis within the whole thalamus. *, p ≤ 0.05.

Aggregation of Presynaptic Proteins in CTSD−/− Mice

In addition to the regional changes in the expression of presynaptic markers, large globular aggregates of Syp, Syb, and SNAP25 were evident in the thalamus of CTSD−/− mice, being more readily apparent in paraffin-embedded thin sections of the brain (Figs. 6, 7). Both Syb- and Syp-immunopositive deposits were abundant within VPM/VPL of CTSD−/− mice (Fig. 6), whereas SNAP25 formed aggregates in the zona incerta thalamic nucleus of these mice (Fig. 7). Synaptosomal-associated protein of 25 kd-immunoreactive aggregates were also scattered across all laminae of the S1BF (Fig. 7) and, to a lesser extent, of the visual and auditory cortices, but never in the entorhinal cortex of CTSD−/− mice. In contrast, the distribution of all these presynaptic markers seemed unchanged in the inhibitory Rt nucleus, which did not display any significant loss of PV-positive inhibitory neurons (Fig. 3C). Taken together, these data provide evidence that redistribution of Syp staining and aggregation of presynaptic proteins occurs in the same nuclei of the thalamocortical system that display neuronal loss and pronounced glial activation.


Aggregation of immunoreactive clusters of synaptobrevin (Syb) and synaptophysin (Syp) in the thalamus of cathepsin D (CTSD)−/− mouse. Representative coronal paraffin sections of the thalamus of CTSD−/− and littermate control+/+ mice immunohistochemically stained for (A, C, E) (Syb) and (B, D) (Syp). Immunohistochemical staining reveals the accumulation of immunoreactive clusters of Syp and Syb in ventral posterolateral thalamic nucleus (VPL) and in ventral posteromedial thalamic nucleus (VPM). (E) In the control+/+ mice, such aggregates were not detected. Scale bars = (A, B) 50 and (C-E) 10 μm.


Synaptosomal-associated protein of 25 kd (SNAP25) staining is clustered in the somatosensory cortex and thalamus of cathepsin D (CTSD)−/− mice. Representative coronal paraffin sections of the barrel field of primary somatosensory cortex (S1BF) and thalamus (Thal) of CTSD−/− mice and littermate controls+/+, immunohistochemically stained for SNAP25 reveals accumulation of SNAP25-positive aggregates in S1BF and dorsal sector of zona incerta (ZID)/ventral sector of zona incerta (ZIV) of thalamus in CTSD-deficient mice (C-F) but not in controls (A, B). Medial eminiscus (mL) between ZID/ZIV and ventral posterolateral thalamic nucleus (VPL)/ventral posteromedial thalamic nucleus (VPM) are indicated in pictures (B, D). Note the absence of SNAP25-immunoreactive aggregates in VPM/VPL. Scale bars = (A-D) 50 and (E, F) 10 μm.

To clarify the subcellular compartment where the presynaptic protein aggregates appeared, we performed double labeling immunofluorescence studies. These stainings revealed good colocalization of the Syb-positive aggregates with the autophagosomal marker LC3 (Fig. 8A) and the lysosomal marker Lamp-1 (Fig. 8B) in the VPM/VPL of CTSD−/− mice, suggesting that the presynaptic protein aggregates reside in autophagosomes/autophagolysosomes in CTSD−/− mice. Moreover, there was also a pool of small vesicular Lamp-1-positive structures devoid of Syb staining, indicating that the normal lysosomes did not contain presynaptic protein aggregates in CTSD−/− mice (Fig. 8B).


Subcellular localization of the presynaptic protein aggregates. Double-immunofluorescence staining for the presynaptic protein aggregates (synaptobrevin [Syb]) and (A) the autophagosomal marker (LC3) or (B) lysosomal marker (lysome-associated membrane protein [Lamp-1]) in the ventral posteromedial/ventral posterolateral thalamic nucleus of cathepsin D (CTSD)−/− and control+/+ mice. Confocal laser scanning microscopy of coronal brain sections in which the presynaptic aggregates (marked by arrows) are visualized by Syb staining in the green channel and the markers of subcellular organelles in the red channel. Arrowheads indicate unstained oval structures likely to be neuronal nuclei, and asterisks indicate granular Lamp-1-positive structures likely to represent normal lysosomes. Colocalization is visualized as yellow color in the overlay image. Scale bars = 20 μm.

As an attempt to establish whether the presynaptic protein aggregates resided in microglial cells, we double-stained paraffin sections of the CTSD−/− mouse brains for Syb and the microglia-specific marker CD68 (Fig. 9A, B). There was no overlap between these stainings, indicating that Syb-positive protein aggregates did not reside in the activated microglial cells (Fig. 9C).


Immunochemical staining of the presynaptic protein aggregates and activated microglia in the thalamus of cathepsin D (CTSD)−/− mouse. Coronal paraffin sections of the thalamus of CTSD−/− mouse were first stained for synaptobrevin (Syb), which was visualized as red color using the peroxidase substrate 3-amino-9-ethylcarbazole (A, arrows). After this, the sections were stained for the microglial marker CD68 using fluorescent Alexa Fluor 488 antibody conjugates (B, arrowheads). An overlay of the images with pseudocolors (C) shows no overlap of Syb-positive protein aggregates (arrows) and activated microglial cells (arrowheads). Scale bar= 20 μm.

CTSD Deficiency Leads to Changes in the Amount of Syp/Syp Complex

On the basis of immunohistochemical evidence for abnormal aggregates of presynaptic proteins in CTSD−/− mice, we next studied the presynaptic protein complexes by immunoblotting. Of the examined proteins, Syb, Syx, and SNAP25 together form the SNARE complex needed for docking and release of synaptic vesicles from the presynaptic terminals, and Syp is likely to have a regulatory role in this process via formation of a Syp/Syb-heterodimer (18). The amount of the SNARE complex (73 kd) seemed unchanged in the cortex and thalamus of CTSD−/− mice (Fig. 10). However, Syp/Syb-complex (56 kd), which regulates the availability of Syb for formation of SNARE complexes, was far more abundant in the brains of CTSD−/− mice than in control mice (Fig. 10). Accordingly, the amount of the Syp-homodimer (76 kd) was reduced in the brains of the affected mice compared with controls, particularly in the thalamus (Fig. 10). No changes in the abundance of the monomeric forms of Syb, Syp, SNAP25, or Syx were observed between the affected and control animals (Fig. 10 and data not shown). Taken together, these data show, at the molecular level, that the presynaptic protein machinery associated with synaptic transmission is affected in CTSD−/− mice and emphasize the importance of presynaptic events in the pathogenesis of CTSD deficiency.


Western blot analysis of presynaptic proteins in cathepsin D (CTSD)−/− and control mouse brains. Representative results of immunoblotting analyses indicated no apparent changes in the steady-state levels of the SNARE complex (73 kd) in the cortex and Thal of CTSD−/− mice (n = 4) compared with controls+/+, probed here with antibody against synaptobrevin (Syb) (upper panel). Immunoblotting with Syb antibody revealed an abnormally high amount of synaptophysin (Syp)/Syb complex (56 kd) in the cortex and thalamus (Thal) of mutant mice (middle panel). Immunoblotting with antibody against Syp confirmed the finding and also indicated a markedly decreased level of Syp homodimer (76 kd) in CTSD−/− mice compared with controls (lower panel).

Ultrastructural Evidence for Axonal Degeneration and Synaptic Loss in VPM/VPL, but not S1BF, of CTSD−/− Mice

Finally, we examined synaptic and axonal morphology in the thalamocortical system at an ultrastructural level. As expected, accumulations of electron-dense storage product were present in many cell soma throughout the VPM/VPL thalamic nucleus in CTSD−/− mice but were absent in preparations from control mice (Ref. (7); Fig. 11A, C, D). Although we found clear evidence for accumulation of electron-dense storage product in neuronal cell bodies, it is possible that similar material was also present in nonneuronal cell types. Most myelinated axon profiles in VPM/VPL of CTSD−/− mice showed signs of degeneration, as indicated by the breakdown of axon membranes, disruption of axonal organelles, and unraveling of myelin sheaths (compare Fig. 11B with C, E, and F). Evidence for widespread synaptic degeneration, indicated by electron-dense synaptic profiles (16), was not present in VPM/VPL of CTSD−/− mice, although synaptic profiles seemed relatively scarce in these mice. Nevertheless, there was evidence of "pinching" of presynaptic nerve terminals from postsynaptic spines by nonneuronal cell types (possibly an astrocyte; Fig. 11G). Qualitative analyses suggesting a paucity of synapses in VPM/VPL were supported by quantitative assessment of synapse number, showing a markedly reduced number of synapses in the affected animals compared with controls (p < 0.005; 2-tailed t-test; Fig. 11H). Thus, the lack of degenerating synaptic profiles in VPM/VPL was most likely because synaptic degeneration and subsequent removal of all synaptic debris in this nucleus was essentially complete at the time point examined.


Ultrastructural analysis revealed widespread axonal degeneration and synapse loss in the ventral posteromedial (VPM)/ventral posterolateral (VPL) nucleus of the thalamus in cathepsin D (CTSD)−/− mice. Representative electron micrographs from VPM/VPL of control+/+ (A, B) and CTSD−/− (C, E, G) mice, and the barrel field of primary somatosensory cortex (S1BF) of CTSD−/− (D, F) mice. In control mice, neuronal cytoplasm was free from accumulations of electron-dense storage material (A), and most axons (B, black arrow) and synapses (B, white arrow) seemed normal. Accumulations of electron-dense storage material were present throughout VPM/VPL in CTSD−/− mice (C), occurring most often around cell soma (C, white arrow). Deposits of storage material were not as common in S1BF (D; note a few storage deposits scattered among healthy mitochondria). Most axon profiles seemed disrupted in VPM/VPL of CTSD−/− mice, providing evidence for a breakdown and loss of axonal membranes and organelles as well as an unraveling of the myelin sheath (C, black arrow; E). The presence of well-preserved organelles such as mitochondria (E; black arrow) in neighboring cells suggests that axonal pathology was not the result of poor tissue processing. Synaptic profiles were commonly observed in the S1BF of CTSD−/− mice and seemed normal (F). Synaptic profiles were not commonly observed in VPM/VPL of CTSD−/− mice, but occasional examples of a cell process (possibly belonging to an astrocyte; G, white arrowhead) "pinching off" a presynaptic nerve terminal (G; white arrow) from its postsynaptic cell (G; black arrow) were found. Nonstereologic quantification of synaptic density (see Materials and methods section) suggested a significant reduction of synaptic density in VPM/VPL of CTSD−/− compared with control mice, but no change in S1BF (H; mean and SEM; 2-tailed t-test). *, p ≤ 0.05; NS, not significant.

In contrast to VPM/VPL, ultrastructural examination of S1BF from the same CTSD−/− mice showed very few changes compared with control preparations. Accumulations of electron-dense storage product in cell soma were rare and much less prominent than in VPM/VPL. There was no evidence of axonal degeneration, and most synapses seemed normal (Fig. 11B). Quantitative assessment of synapse number in lamina IV of S1BF revealed no difference between CTSD−/− and control mice (p > 0.05; 2-tailed t-test). However, a couple of examples of degenerating synaptic profiles were identified (data not shown), suggesting that pathologic changes at the level of axons and synapses were only beginning to occur in S1BF during the period when axonal degeneration and synapse loss were prevalent in the VPM/VPL nucleus of the thalamus.


Neuropathologic Findings in CTSD−/− Mice Resemble Those in Human Congenital NCL Patients

In the present study, we have used CTSD−/− mice to reveal novel molecular and cellular aspects underlying the pathogenesis of CTSD deficiencies. We observed structural alterations and atrophy of various brain regions, including both white and gray matter in CTSD−/− mouse brains. These findings were in good agreement with the extreme and generalized brain atrophy previously observed in human and ovine congenital NCL caused by CTSD deficiency (5,6,19). In accordance with the remarkable loss of myelin and axons in the subcortical white matter of patients with congenital NCL (19), the corpus callosum showed the highest degree of atrophy in CTSD−/− mice. Astrocytosis in brains of the affected mice was severe and generalized, similar to that in CTSD−/− human patients and lambs (5,6,19). Despite these similarities, the degree of brain atrophy and neuronal loss is markedly milder in CTSD−/− mice than in CTSD−/− humans or sheep, possibly reflecting functional and structural differences in the brain between these species. In comparison with mouse models of other NCL types, however, this mouse model of CTSD deficiency shows the most severe phenotype and brain atrophy, thus being parallel to the relative phenotypes of the corresponding human NCL diseases.

The Somatosensory Thalamocortical System Shows Severe Degenerative Changes in CTSD−/− Mice

A striking feature of CTSD−/− mice was the remarkably pronounced microglial activation within VPM/VPL nucleus of the thalamus. In further analyses, this thalamic nucleus was revealed as a particular focus of degenerative changes, exhibiting not only intense glial activation but also neuronal loss, redistribution of presynaptic proteins, axonal degeneration, and loss of synapses. The regional specificity of these changes was evident because they did not occur in the neighboring thalamic nuclei. Although thalamic pathology in CTSD−/− mice has already been reported (11), the severe involvement of discrete nuclei within this region was unexpected because the cerebral cortex has been reported to be the most severely affected brain region in both human and ovine CTSD deficiency (5,6,19). Experiments during the course of the disease indicated that, indeed, it was the thalamic VPM/VPL in which the glial activation was first seen in CTSD−/− mice, with the earliest signs of astroglial activation being visible at P16, followed by marked activation of both astrocytes and microglia, as well as significantly reduced neuronal number at P20. Previously, the variable involvement of individual thalamic nuclei has been noted at autopsy of infantile NCL patients (20), magnetic resonance images of patients with different forms of NCL (21), and in mouse models of infantile and juvenile NCL (22,23), suggesting that the thalamus is an important center in which the pathologic processes may originate in multiple forms of NCL.

Pathologic alterations in VPM/VPL of CTSD−/− mice were accompanied by degenerative changes in the functionally associated laminae IV and VI of S1BF cortex. Line profile analysis revealed novel evidence for lamina-specific glial activation in CTSD−/− mice. The most intense glial activation occurred in laminae IV and VI of S1BF cortex, the same laminae showing alterations in Syp expression. Lamina IV of S1BF also showed significant neuronal loss and thinning in CTSD−/− mice, whereas lamina VI was significantly reduced in thickness but showed only a mild tendency toward reduced neuronal number, suggesting that the packaging of neurons may be abnormal in lamina VI. Although this spatial correlation between the glial activation and neuronal loss may simply reflect activation of glial cells during neurodegeneration, it may alternatively reflect the influence of astrocytes upon synaptic efficacy (24,25,26). Taken together, these data indicate that the somatosensory thalamocortical system is a particular focus of neurodegenerative events in CTSD−/− mice.

The thalamus forms a gateway from the peripheral sensory organs to the cortex. Somatosensory information is normally relayed from VPM/VPL to lamina IV neurons in S1BF. This input is regulated by cortical neurons in lamina VI, which project back to the thalamus. The presence of neurodegenerative changes in these interconnected pathways raises the possibility of a dysfunction within the thalamocortical circuitry of CTSD−/− mice. Despite no loss of inhibitory neurons in the Rt nucleus of affected mice, the loss of feedback from neurons in lamina VI may decrease the modulatory influence of S1BF upon VPM/VPL. This may result in a greater excitatory input from VPM/VPL to its target neurons in the cortex and cause excitotoxic effects in lamina IV. Although the axonal pathology observed in VPM/VPL is consistent with this suggestion, the functional consequences of these alterations remain unclear. As such, it will be important to test somatosensory thalamocortical circuitry functions in CTSD−/− mice. Interestingly, elevated somatosensory evoked potentials have been reported in patients with the Finnish variant of late-infantile NCL (CLN5) (27), reflecting ongoing dysfunction within the somatosensory thalamocortical system.

Synaptic Alterations in CTSD−/− Mice

Unexpectedly, immunoreactivity for Syp, a synaptic vesicle membrane protein, was increased in the cortical and subcortical brain regions of CTSD−/− mice as compared with controls. However, this did not apply to VPM/VPL, where the staining for Syp and 2 other synaptic markers, SNAP25 and Syx, was markedly decreased, as might be expected with the observed loss of neurons and synapses in this brain region. Interestingly, synaptic loss in CTSD−/− mice was evident only in VPM/VPL, but not in S1BF, suggesting that the degenerative changes in the thalamus precede those in the cortex.

Another distinctive feature of CTSD−/− mice was the abnormal accumulation of presynaptic proteins: Syp- and Syb-immunoreactive aggregates were confined to VPM/VPL, whereas SNAP25-positive aggregates were more widely distributed in the affected mice. As shown by double labeling experiments, the Syb-positive protein aggregates localized within the autophagosomal/autophagolysosomal compartment. It is unclear whether these aggregates may have a pathogenic role in CTSD deficiency, but vesicular stalling and generation of cytosolic or intravesicular aggregates in axons may be associated with blocked axonal transport in a variety of neurodegenerative conditions (28). Indeed, our ultrastructural evidence of gross axonal pathology with disrupted organelles and unraveled myelin sheaths within VPM/VPL of CTSD−/− mice would be compatible with the suggestion of axonal transport deficits. Previously, Koike et al (29) reported axonal pathology and accumulation of autophagosomes in cortical neurons as well as corpus callosum of CTSD−/− mice and suggested that axonal transport may be impeded in these mice. Although we were unable to confirm the type of cells in which the aggregates resided, the distribution of these aggregates was clearly distinct from that of microglial cells, thus indirectly implying that the presynaptic protein aggregates may reside within neurons. Thus, the involvement of presynaptic protein accumulation within autophagosomes and its potential consequences to axonal transport and signaling in CTSD−/− mice warrants further examination.

Synaptic transmission requires the release of neurotransmitters from synaptic vesicles by regulated exocytosis. This is facilitated by formation of the SNARE complex composed of Syb, SNAP25, and Syx (18). Synaptophysin participates the regulation of this process by forming a Syp/Syb-heterodimer, which then controls the ability of Syb to enter the SNARE complex (18,30,31). Thus, the increased amount of Syb/Syp-heterodimer in the cortex and thalamus of CTSD−/− mice may be associated with excessive neuronal activity, which could occur during seizures in these mice (7). Indeed, elevated level of Syp/Syb-complex has been reported in the rat kindling model of epilepsy (32). Alternatively, the abnormally high levels of the Syp/Syb-heterodimer can result from a failure of this complex to dissociate normally. Dissociation of the heterodimer precedes the formation of the SNARE complex and is controlled by neuronal activity via a poorly understood calcium-dependent mechanism (18,33). Although the Syp/Syb-dimer can be experimentally dissociated by proteolytic cleavage using tetanus toxin (18), the proteases responsible for the dissociation under physiologic conditions are unknown, and it remains to be seen whether CTSD or other endosomal/lysosomal proteases may have a role in this process. The presynaptic protein complexes may also be targeted to the autophagosomal compartment for degradation, and the Syb-containing aggregates arise from a failure of this autophagic process in CTSD−/− mice. Together, these data provide the first direct in vivo evidence for synaptic abnormalities in CSTD deficiency at the tissue, molecular, and ultrastructural levels. Our findings are also in line with previous in vitro studies, suggesting that synaptic malfunction in infantile NCL is reflected by a reduced synaptic vesicle pool size in primary neuronal cultures of the CLN1 knockout mouse (34).

In conclusion, we report a series of localized pathologic changes in brains of CTSD−/− mice, particularly pronounced within the thalamocortical pathways between VPM/VPL and laminae IV and VI of the S1BF. These changes included not only loss of neurons and synapses but also axonal degeneration, redistribution of presynaptic markers, and abnormal accumulation of the synaptic Syp/Syb complex. The present data suggest that presynaptic modulation may be disrupted in CTSD−/− mice, leading to aggregation of presynaptic proteins within autophagosomal compartment and, further, to axonal and neuronal degeneration in CTSD deficiencies. These data also emphasize the role of CTSD, a lysosomal protease, in influencing the presynaptic organization.


The authors thank Sirkka Kanerva and Steve Mitchell for expert assistance with immunoblots and electron microscopy, as well as Drs. Mikko Liljeström and Mika Hukkanen at the Institute of Biomedicine/Cell Imaging Unit, University of Helsinki, for assistance with fluorescent microscopy. Prof. Takashi Ueno is gratefully acknowledged for the LC3 antiserum and Dr. Ulla Lahtinen for critical reading of the article.


  • The authors Partanen and Haapanen contributed equally to this work.

  • This study was financially supported by grants from the Academy of Finland (214343; JT), National Institutes of Health (NS41930; JDC), European Commission (LSHM-CT-2003-503051; JT, JDC), The Batten Disease Support and Research Association (JDC, CK, THG), The Natalie Fund (JDC), The Batten Disease Family Association (JDC), Biotechnology and Biological Sciences Research Council (THG), and Scottish Hospital Endowments Research Trust (THG).

  • Sanna Partanen is a member of the Finnish Graduate School of Neuroscience.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
View Abstract