Main

α-Mannosidosis is an inherited disorder of glycoprotein metabolism resulting from the absence or defective function of lysosomal α-mannosidase (EC 3.2.1.24). This leads to the accumulation of undegraded mannose-rich oligosaccharides in lysosomes and multisystem storage (1). The disease has been described in humans (reviewed in Ref.1), cattle, and cats (reviewed in Ref.2), and is characterized in all three species by progressive neurologic deterioration and premature death. Ingestion of certain toxic plants containing swainsonine, an indolizadine alkaloid that is a potent inhibitor of both lysosomal α-mannosidase and Golgi mannosidase II, also induces a similar condition, commonly in grazing livestock (2).

The disease in humans is rare, with an autosomal recessive mode of inheritance. A continuum of disease phenotypes is evident, ranging from a severe infantile form to a milder juvenile-adult form (1, 3). Clinical features include progressive mental retardation, facial dysmorphia, and skeletal abnormalities. Hearing loss, recurrent bacterial infections, corneal and lenticular opacities, hepatomegaly, and hernias are also commonly observed. Destructive synovitis and polyarthropathy also have been found in several patients surviving beyond childhood (4–6). Significant heterogeneity in disease severity has been reported between siblings (4, 7).

α-Mannosidosis in Angus and Angus-derived cattle breeds was of considerable economic importance in Australia and New Zealand until carrier detection programs were instituted. Affected animals surviving the neonatal period exhibit progressive ataxia, aggression, and failure to thrive, and die before 18 mo of age (8, 9). More severe disease has been observed in Galloway cattle, whose affected calves either aborted, were stillborn, or died soon after birth (9, 10). Specific mutations have recently been reported for each of these breeds (11), and PCR-based assays have been developed for routine carrier detection (12).

Feline α-mannosidosis has been reported in Persian (13–15), domestic shorthair (16, 17), and domestic longhair (18) breeds. Clinical heterogeneity in disease severity is also apparent among these breeds, with the mildest form in the domestic longhair cats. A disease mutation was recently identified in Persian but not in domestic longhair cats, also indicating molecular heterogeneity in feline α-mannosidosis (19).

Both bovine and feline models of α-mannosidosis have enabled detailed studies of pathology, oligosaccharide analysis, cellular pathology of the CNS (20), and also response to BMT (21). Swainsonine-induced disease with subsequent withdrawal has also allowed similar studies, as well as afforded additional information regarding reversibility of lesions, particularly in the CNS (22–27).

The genomic and cDNA sequences for human (28–30) and bovine lysosomal α-mannosidase (11) and feline genomic lysosomal α-mannosidase (19) have been determined, thus enabling the development of future therapeutic strategies using enzyme and gene replacement therapies. Expression systems are currently being developed that will enable large amounts of recombinant human lysosomal α-mannosidase to be produced in cell culture (Thomas Berg, personal communication).

A preliminary report recently described naturally occurring α-mannosidosis in domestic pet guinea pigs (31). In the present report, we describe in detail the clinical, biochemical, and pathologic features of this new animal model for α-mannosidosis.

Materials and Methods

History and presentation.

Several offspring from successive litters of domestic pet guinea pigs showed stunted growth and progressive mental dullness. Premature death by misadventure occurred at 12–16 wk of age. Subsequently, an affected animal was submitted to a veterinary diagnostic laboratory, and histopathology revealed widespread cytoplasmic vacuolation of cells in the CNS and visceral tissues. α-Mannosidosis was diagnosed after enzymology of liver extracts and leukocytes and thin-layer chromatography of the urine (31).

The parents of the original animals were not immediately related, inasmuch as they had been obtained from different sources. Affected animals of both sexes had been bred by the owner, and breeding outcomes suggested an autosomal recessive mode of inheritance, with three affected animals identified by clinical signs from a total of 20 live and two dead guinea pigs born in six successive litters. Two affected male guinea pigs, GP-A and GP-B, were donated by the owner for further detailed evaluation. All studies relating to these two animals were approved by the Women's and Children's Hospital animal ethics committee.

Enzymology.

Enzyme activities in leukocytes, plasma, and skin fibroblasts were determined fluorometrically, using the appropriate 4-methylumbelliferone derivatives, according to standard methods (32, 33). Lysosomal α-mannosidase activity was assayed at pH 4.0. Whole blood was collected into EDTA tubes, and leukocytes were isolated by dextran sedimentation (34). Plasma and leukocytes were frozen at −80°C until assayed.

Liver homogenates (∼20% wt/vol) from both α-mannosidosis-affected guinea pigs and one unrelated normal control were also prepared for enzyme assay. Liver was minced finely in 0.15 M NaCl and 0.1% (vol/vol) Triton X-100 and then were homogenized in a Potter-Elvehjem homogenizer using three strokes. Homogenates were centrifuged for 10 min at 3000 ×g at 4°C, and the supernatant extracts were collected. Protein concentration in the supernatant was assayed by the method of Lowry et al. (35). Supernatants (∼100 μg protein) were assayed for α-mannosidase activity by using 4-methylumbelliferyl-α-mannopyranoside (Melford Laboratories Ltd., Suffolk, UK) with a final substrate concentration of 5 mM, in 0.1 M citrate–0.2 M phosphate buffer (36) at variable pH values ranging from 3.0 to 8.0. Samples were incubated for 60 min at 37°C, and the reaction was terminated by the addition of 1.5 mL 0.2 M glycine buffer, pH 10.7. Fluorescence was measured at excitation 366 nm and emission 446 nm.

Liver samples for subcellular fractionation were also collected postmortem from GP-B and a normal control guinea pig and were immediately placed on ice. Homogenates were prepared and fractionated on an 18% (vol/vol) Percoll gradient as described previously (37). β-Hexosaminidase (38) and lysosomal α-mannosidase activities (pH 4.0, method as above) were determined in fractions across the gradient.

Urinary and tissue oligosaccharide TLC.

Urine was stored at −20°C until use, and then thawed and centrifuged before analysis. Urine creatinine concentration assayed by an autoanalyzer method (Sychron CX Systems, Beckman Instruments Inc., Fullerton CA) was used to determine the volume of urine to prepare for analysis [volume (μL) = 176/creatinine (mmol/L)]. Aliquots of urine were desalted over ∼600-μL mixed bed resin columns (AG-501-X8-formate, Bio-Rad Laboratories, Hercules, CA) and were defatted with chloroform:methanol in a ratio of 0.4:1:1. The upper phase was collected and lyophilized. Samples were applied to Silica Gel 60-coated glass plates (Merck, Darmstadt, Germany) and were developed twice in an n-butanol-acetic acid-water system (2:1:1 vol/vol). To detect oligosaccharides, resorcinol and orcinol were used, as described previously (39). Glucose, lactose (both obtained from Ajax Chemicals, Auburn, NSW, Australia), maltoheptose (Aldrich Chemical Co., Milwaukee, WI), and NANA, NANA-lactose, mannose, and glucuronic acid (all obtained from Sigma Chemical Co., St. Louis, MI) standards were used, loading 3 μg of each per lane. Urine from a human patient with α-mannosidosis was also prepared as above and analyzed for comparison.

Tissues collected at postmortem were frozen at −80°C until further analysis. Samples (0.2 g) of cerebrum, kidney, and liver were minced finely in deionized water (60% wt/vol) and were sonicated on ice for 3 × 5 sec bursts. Homogenates were freeze-thawed five times and centrifuged for 5 min at 13,800 ×g at 4°C. The pellet was washed with 100 μL H2O and recentrifuged, and both supernatants were pooled (final volume ∼500 μL). Ten microliters of supernatant were lyophilized and analyzed by TLC, as described above.

Clinical evaluation.

Corneal clouding was evaluated with a portable slit-lamp and the lens, using a direct ophthalmoscope. Radiographs were taken under general anesthesia [atropine 0.025 mg/kg (Atrosine Mitis, Parnell Laboratories Pty Ltd, Australia), xylazine 5 mg/kg (Rompun, Bayer Aust. Ltd, Australia), and ketamine 5–15 mg/kg (Ketamine Injection, Parnell Laboratories Pty Ltd), subcutaneously], at a constant focal film distance with detail-intensifying screens. Body weights were recorded regularly.

Six normal male guinea pigs from a separate nonbarrier colony were weighed regularly and radiographed at 14 and 18 wk of age to provide limited control data. These animals were housed in large floor pens containing 30–40 animals, were fed pelleted rations ad libitum, and were supplied with fresh greens on alternate days. Tissue and blood samples were also collected from adult male animals being culled from this colony to provide control data.

Hematology and clinical chemistry.

Air-dried blood films were stained with May-Grünwald Giemsa and were examined at ×400. Serum samples from both α-mannosidosis-affected guinea pigs and six adult male normal controls were analyzed for total thyroxine (T4) and total tri-iodothyronine (T3) concentrations. For the T4 assay, a Magic Lite kit (Chiron Diagnostics Corporation, East Walpole, MA) radioimmunoassay was used, and T3 was determined with a microparticle enzyme-linked immunoassay on an Abbott IMX system (Abbott Laboratories, Abbott Park, IL).

To determine kidney, liver, and pancreatic function, we performed biochemistry on heparinized plasma samples from one α-mannosidosis affected guinea pig (GP-B at 23 wk of age) and seven adult male normal controls, using standard autoanalyzer methods (Beckman Sychron CX Systems).

Light and electron microscopy.

GP-A and GP-B were killed by anesthetic overdose at 17 and 23 wk of age, respectively. Tissues were also collected from normal adult male control guinea pigs. Tissue samples for routine light microscopy were fixed in 10% buffered formalin and were paraffin embedded. Sections were stained with H&E and Luxol fast blue/PAS. Brain sections were also incubated with a biotinylated mouse anti-human MAb to amyloid precursor protein (clone 22C11), a sensitive marker of axonal injury, according to previous methods (40). Selected cryostat sections of snap-frozen brain from GP-B were protected by celloidin before staining by the PAS method. Tissues for electron microscopy were rapidly placed in 2% glutaraldehyde/2% paraformaldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, fixed overnight at 4°C, and were processed routinely and embedded in Spurrs resin. One-micrometer-thick survey sections stained with toluidine blue were evaluated at ×100–400 to assess overall distribution of vacuolation due to lysosomal storage. Selected samples were evaluated further under electron microscopy.

RESULTS

Clinical examination.

At 6–8 wk of age, the growth of both affected guinea pigs was stunted compared with that of littermates, and they had vague behavioral abnormalities; both conditions became more severe with increasing age (Fig. 1). Clinical features were characterized by excessive time spent eating, subdued activity, abnormal posture with an inability or reluctance to raise the head and fully extend forepaw toes, and abnormal hindlimb positioning and gait. Gait deficits were more apparent on smooth surfaces. Hopping responses were delayed in all four limbs, particularly in the hindlimbs, and hindlimb proprioception was also abnormal. Both affected guinea pigs vocalized normally, but GP-B at 23 wk of age showed facial asymmetry with unilateral exophthalmos and mild strabismus in the same eye. GP-B also had constantly moist skin under his chin from at least 12 wk of age, suggesting difficulty eating and diffuse brainstem involvement.

Figure 1
figure 1

Body weights (g) in normal male (▵, n= 6) and α-mannosidosis male guinea pigs GP-A (▪) and GP-B (•) with increasing age. Error bars = 1 SD.

Full size image

The affected guinea pigs had dull coats and eyes; however, ophthalmological examination was normal. The face was slightly broader, with hypertelorism and mild shortening of the nose. Both animals also had a history of dermatitis that was much more severe than in normal animals within the same colony, suggesting impaired immune function. Cutaneous lesions resolved with the administration of Ivermectin (MSD Agvet, Australia Pty Ltd).

Subtle differences in skeletal appearance were observed on radiographs, including irregular long bone cortical thickening in the proximal humerus and distal femur, possibly due to altered cortical bone remodeling. Long bones and vertebrae were also shorter compared with age- and sex-matched controls, and vertebral dorsal and lateral spinous processes and the acromion processes of the scapulae were also reduced in length. No obvious differences in vertebral bone density or vertebral shape in the thoracolumbar spine was observed, except for wider pedicles of the dorsal vertebral arches in the lumbar spine. Development and fusion of secondary centers of ossification appeared normal. Skull shape was altered, with broader and shorter nasal bones (dorsoventral view), and the zygomatic arches in GP-B were thicker. Remaining skull bones were not obviously thickened.

From 16 wk of age, the older affected guinea pig was housed for 7 wk with a sexually mature normal female guinea pig. The female was monitored for pregnancy for an additional 65 d after euthanasia of the male, and no pregnancy was detected. Normal gestation length in the guinea pig is 68–70 d.

Hematology and clinical chemistry.

Blood film examination revealed several clear vacuoles in many lymphocytes. Some lymphocytes appeared normal, and a small proportion contained large numbers of vacuoles. Large, slightly acidophilic inclusions (Kurloff bodies) (41) were also observed in some lymphocytes; however, these were also present in normal controls. Mild elevations in plasma aspartate aminotransferase and γ-glutamyltranspeptidase were observed in GP-B, suggesting a mild cholestatic hepatopathy. However, all remaining parameters were within normal limits (urea, creatinine, glucose, amylase, cholesterol, total bilirubin, alanine aminotransferase, alkaline phosphatase, and total protein). Thyroid function tests (total T3 and total T4) in the α-mannosidosis affected animals were normal.

Enzymology.

There was a marked deficiency of lysosomal α-mannosidase activity in leukocytes, plasma, and skin fibroblasts at pH 4.0 from both α-mannosidosis affected animals compared with normal control animals (Table 1). α-Mannosidase activity measured over a pH range of 3.0–8.0 in liver extracts confirmed a deficiency of lysosomal α-mannosidase (pH 4.0) activity in both affected animals (Fig. 2). In contrast, a peak of activity was observed at pH 4.0 in the normal control. Two additional peaks of enzyme activity were observed at pH 5.3–5.4 and pH 6.2–6.4 in both normal and α-mannosidosis affected animals, which was attributed to normal activity of nonlysosomal α-mannosidases (42). After fractionation of liver samples, a deficiency of α-mannosidase activity (pH 4.0) was localized specifically to lysosomal-rich fractions in the α-mannosidosis affected guinea pig (Fig. 3), confirming a deficiency of lysosomal α-mannosidase.

Table 1 Activity of lysosomal enzymes in leukocytes, plasma, and skin fibroblasts in guinea pigs with α-mannosidosis and normal controls* Mean value; range given in brackets.† nmol/min/mg protein.‡ % normal given in brackets.§ nmol/min/ml.∥ Not done.¶n= 3.**n= 2.
Full size table
Figure 2
figure 2

pH-dependence of α-mannosidase activity in liver extracts from normal and α-mannosidosis guinea pigs (▵, normal, n= 1; •, GP-A α-mannosidosis; ▪, GP-B α-mannosidosis).

Full size image
Figure 3
figure 3

Distribution of β-hexosaminidase and α-mannosidase activity, pH 4.0, in Percoll fractions from (A) normal guinea pig liver (□, β- hexosaminidase; ▪, α- mannosidase) and (B) α-mannosidosis guinea pig liver (○, β- hexosaminidase; •, α- mannosidase). Five microliters of each fraction were assayed for β-hexosaminidase activity, and 30 μL of each fraction were assayed for α-mannosidase activity. Elevated β-hexosaminidase activity at the bottom of the gradient indicates lysosomal-enriched fractions, and corresponds with a peak in pH 4.0 α-mannosidase activity in the normal liver which is absent in the α-mannosidosis liver.

Full size image

Urinary and tissue oligosaccharides.

Thin layer chromatography demonstrated seven distinct orcinol-positive bands, larger than the lactose standard, in high concentrations in urine, brain, and kidney extracts from the α-mannosidosis guinea pig and not in the normal control (Fig. 4). The pattern of oligosaccharide excretion in urine and tissues from the α-mannosidosis guinea pig was almost identical to the pattern observed in urine from a human α-mannosidosis patient. Orcinol-positive bands were present in liver extracts from both α-mannosidosis and normal guinea pigs, with an additional band present in the normal guinea pig. The presence of multiple positive bands in normal liver has been observed previously in other normal animals (Margaret Jones, unpublished observation).

Figure 4
figure 4

TLC of urine and tissue extracts from α-mannosidosis and normal guinea pigs, and urine from a human α-mannosidosis patient. In this legend, α-m denotes α-mannosidosis; Norm denotes normal. Lane 1, glucose (G), lactose (L), and maltoheptaose (M) standards. Lane 2, human α-m urine. Lanes 3–10 from guinea pig extracts:Lane 3, α-m urine;Lane 4, Norm urine;Lane 5, α-m brain;Lane 6, Norm brain;Lane 7, α-m kidney;Lane 8, Norm kidney;Lane 9, α-m liver;Lane 10, Norm liver;Lane 11, mannose (Ma), glucuronic acid (GA), NANA (N) and NANA-lactose (NL) standards. Asterisk and bracket indicate differences in oligosaccharide pattern between α-m and Norm livers (Lanes 9 and 10).

Full size image

Gross pathology.

At necropsy, multiple cysts were present in both kidneys from GP-A, with fewer cysts present in kidneys from GP-B. Liver and spleen sizes were normal. However, the mesenteric lymph node in GP-B appeared enlarged. Thyroid dimensions were measured in GP-B and were no different from those of normal control guinea pigs. In GP-B, forelimbs were unable to extend as fully as those of normal control animals. All major joint surfaces were normal, except for erosive lesions observed on the distal trochlea surface and intercondylar fossa of both distal femurs in both α-mannosidosis affected animals. Deficient dietary vitamin C is known to cause skeletal and joint abnormalities in guinea pigs (43). However, vitamin C supplementation in the water and fresh vegetables supplied daily was considered adequate for both affected animals. No other gross pathology was observed.

Light and electron microscopy.

Although generally visible on H&E-stained paraffin sections, lysosomal vacuolation was most clearly demonstrated on toluidine blue-stained 1-μm sections. Table 2 details the grading system used to describe distribution and severity of lysosomal vacuolation. Light microscopy of 1-μm sections revealed marked cytoplasmic vacuolation in neurons in most regions of the CNS, including the cerebral cortex, hippocampus, thalamus, cerebellum, midbrain, pons, medulla, and the dorsal and ventral horns of the spinal cord (Table 2;Fig. 5, A and B). Large neurons were generally more severely affected than small neurons. The dendritic processes in cerebellar Purkinje cells were frequently swollen with numerous fine vacuoles, and smaller numbers of vacuoles were present in the Purkinje cell perikarya (Fig. 5B). There was also an impression of a mild decrease in Purkinje cell numbers in the older α-mannosidosis affected animal; however, this needs to be quantified in greater numbers of animals in future studies. Prominent vacuolation was also present in astrocytes, pericytes (Fig. 5C), dorsal root and trigeminal ganglion neurons (Fig. 5D), renal tubule epithelial cells, thyroid follicular cells (Fig. 5E), pancreatic acinar cells (Fig. 5F), epididymal epithelial cells (Fig. 5G), and macrophages in lymphoid tissues (Table 2). Less prominent vacuolation was found in some regions of the brain, including the caudate nucleus, and in other cell types such as endothelial cells, oligodendrocytes, Schwann cells, fibroblasts, smooth muscle cells, chondrocytes, osteocytes, and osteoblasts in various tissues. No vacuoles were observed in hepatocytes. Vacuole contents did not stain when routine or special stains were used, except for a few PAS-positive neurons in layers III and VI of the cerebral cortex, in celloidin-protected frozen sections. Amyloid precursor protein-positive axonal spheroids of varying size and staining intensity were scattered throughout corpus callosum, subcortical white matter, cerebellar Purkinje cell layer, and deep white matter (Fig. 5C), caudate nucleus, thalamus, and brainstem. Gliosis and demyelination were not observed.

Table 2 Distribution of lysosomal storage in tissues from two guinea pigs affected with α-mannosidosis by light microscopy* 0, no vacuoles; 1+, few vacuoles; 2+, many vacuoles not distending cytoplasm; 3+, many vacuoles distending cytoplasm.
Full size table
Figure 5
figure 5

Resin-embedded 1-μm sections from an affected guinea pig (toluidine blue). (A) Cerebral cortical neurons with numerous clear vacuoles distending the cytoplasm (arrows) (×246). (B) Cerebellum. Purkinje cell neurons with cell body vacuolation (arrows) and swollen dendritic process (arrowhead) containing numerous vacuoles (×197). (C) A large axonal spheroid (arrow) visible within the cerebellar white matter, and an adjacent blood vessel with heavily vacuolated pericytes (arrowheads) and mildly vacuolated endothelial cells (×339). (D) Trigeminal ganglion cells containing multiple fine vacuoles in the cytoplasm (arrows) (×246). (E) Marked vacuolation in thyroid follicular cells, with some follicles containing free-floating follicular cells within the colloid (×212). (F) Exocrine pancreatic epithelium with severe foamy vacuolation (×197). (G) Numerous clear vacuoles of variable size in epididymis epithelium (×189).

Full size image

Cerebral cortex, cerebellum, medulla, caudate nucleus, thyroid, and pancreas were examined ultrastructurally. Most lysosomal storage vacuoles in both the CNS and visceral tissues were lucent, or contained fine fibrillar or flocculent material typical of a glycoprotein storage disease (Fig. 6A). Rare large neurons in the cerebral cortex contained fine lamellated membranous structures and fine fibrils in stacks (Fig. 6B). The contents of axonal spheroids were variable and included mixtures of dense bodies, empty vesicles, mitochondria, membranous whorls, and flocculent material. Several aggregates of tightly packed tubular lattice structures identical to those described by Jellinger (44) were observed in axonal spheroids in cerebellar white matter. Some axonal spheroids were clearly surrounded by a myelin sheath, but not obviously in others with an otherwise similar appearance (Fig. 6C).

Figure 6
figure 6

(A) Electron microscopy of neuronal perikaryon from cerebral cortex of an affected guinea pig. Cytoplasm distended by largely empty, membrane-bound vacuoles (arrows). Dendritic process (d) (×2232). (B) Electron microscopy of cerebral cortex. Storage vacuoles in some neurons contained fine fibrils in stacks suggestive of lipid (arrows) (×5386). (C) Electron microscopy of caudate nucleus white matter. Axonal spheroids (thin arrows) appeared variable and contained many degenerate organelles. Myelin sheath attenuated (thick arrow) (×4728).

Full size image

DISCUSSION

Morphologic and biochemical examination of two guinea pigs described in this paper confirmed the diagnosis of α-mannosidosis. Reduced lysosomal α-mannosidase activity leads to progressive accumulation of mannose-rich oligosaccharides in many tissues, as clearly demonstrated by TLC of tissue extracts in these guinea pigs. Long-lived postmitotic cells such as neurons are particularly vulnerable to lysosomal storage. Neuronal involvement in the α-mannosidosis guinea pigs, as with many other lysosomal storage diseases, was severe and widely distributed throughout the brain, corresponding with clinical neurologic disease. As in human cases of α-mannosidosis (45, 46), there were many vacuolated neurons in the cerebral cortex, brainstem, cerebellum, spinal cord, and trigeminal and dorsal root ganglia, with lesser involvement in the basal ganglia. Loss of cerebellar Purkinje cells has also been observed in some but not all cases of human (4, 45) and feline α-mannosidosis. Among cats, however, this feature appears to be observed only in the longer-lived domestic longhair breed (18). Demyelination was not observed in the guinea pigs and is an inconsistent feature in human (46), feline (15, 18), and bovine α-mannosidosis (47). Axonal damage appears to be more common in animal models than in humans (3), and there were many axonal spheroids in affected guinea pig brains, with a morphology similar to that in bovine α-mannosidosis (47, 48). Moreover, the correlation between location and severity of axonal spheroids and clinical neurologic changes in a number of animal models of lysosomal storage diseases suggests that axonal spheroids may cause more neurologic derangement than any other morphologic change, including intraneuronal storage, and spheroids also tend to persist even when neuronal storage is reversed (20).

Because glycoproteins are major constituents of cellular membranes and extracellular matrix, lysosomal storage is also present in multiple visceral tissues in α-mannosidosis. Vacuolation has been observed in lymphocytes, hepatocytes, Kupffer cells, fibroblasts, endothelial cells, histiocytes, and some epithelial cell types in human α-mannosidosis (4, 49–52). In our guinea pig cases, storage vacuoles were most clearly demonstrated on semithin sections of resin-embedded tissues and were widely distributed in visceral tissues, especially in the secretory epithelia of thyroid, pancreas, renal tubules and epididymis, and in reticuloendothelial cells. Hepatocyte storage generally appears to occur in more severe cases of α-mannosidosis in humans (51) and cats (13, 18), thereby suggesting a milder disease spectrum in the guinea pigs. Pancreas, kidney, and tissue macrophages were also prominent sites of extraneural storage in bovine and feline cases (17, 47). Thyroid vacuolation has been reported in some cases of feline α-mannosidosis (13) and more particularly in caprine and bovine β-mannosidosis (53, 54). Impaired thyroid function was also demonstrated in the β-mannosidosis models but was not observed in the α-mannosidosis guinea pigs. Vacuolated epididymis epithelium in fucosidosis-affected dogs appears almost identical to that seen in the α-mannosidosis guinea pigs, and rendered the dogs infertile (55). Further test breedings are required to determine the fertility of the α-mannosidosis guinea pigs. The skeletal abnormalities in the guinea pigs appeared much less severe than in human α-mannosidosis patients, although this condition is very variable in humans (1). However, the degenerative joint disease present in both guinea pigs may have parallels with the destructive polyarthropathy observed in some adult human α-mannosidosis cases (4–6).

Ultrastructural features of the lysosomal inclusions in guinea pigs were consistent with those previously described in human (46, 49, 51, 52), feline (13, 17, 18), and bovine α-mannosidosis (47, 48), with the majority of inclusions in neuronal and visceral cell types lacking any contents or containing fine reticulogranular material typical of oligosaccharide storage. Occasional lamellar or membranous inclusions were similar to those in a human α-mannosidosis patient (46) and in the inherited and swainsonine-induced α-mannosidosis cat models (17, 18, 26). In the cat models, these types of inclusions were found in limited numbers of PAS-positive pyramidal neurons which were also immunoreactive for GM2 ganglioside (reviewed in Ref.56). In addition, these pyramidal neurones exhibited ectopic dendrites, suggesting that abnormal dendrite growth is induced by accumulation of GM2 ganglioside. The presence of occasional PAS-positive neurons as well as membrane leaflets in inclusions in small numbers of cortical neurons in the α-mannosidosis guinea pigs suggests that ganglioside storage may also occur in guinea pigs. Golgi studies and immunostaining of these regions will be important in determining whether cellular pathology is similar in these two different models of α-mannosidosis.

Swainsonine administered to normal guinea pigs produced histologic lesions that were almost identical to those observed in the inherited α-mannosidosis guinea pig model (57, 58). Notable differences were that vacuoles were observed in hepatocytes but not in ependymal cells in the swainsonine-induced disease. Kupffer cells, involved in both the induced and inherited guinea pig diseases, demonstrated vacuolation at an earlier stage of the disease than hepatocytes in swainsonine-induced α-mannosidosis. The absence of hepatocyte vacuolation in inherited α-mannosidosis may therefore indicate a milder or earlier stage of the disease.

TLC of urinary and tissue oligosaccharides in the guinea pigs revealed a pattern that was almost identical to that of human α-mannosidosis urine, both in proportion of each oligosaccharide band and in the number of bands present. This contrasts with the patterns observed on TLC in feline and bovine α-mannosidosis (16). The major oligosaccharide species detected in these models were Man3GlcNAc2 and Man2GlcNAc2, respectively, compared with the major component Man2GlcNAc in human α-mannosidosis (reviewed in Ref.59). Analysis of stored oligosaccharides in human, bovine, and feline α-mannosidosis has identified structural differences in the major oligosaccharides in tissues and body fluids (reviewed in Ref.60). In particular, stored oligosaccharides in both cattle and cats contain terminal di-N-acetylchitobiose instead of the single terminal N-acetylglucosamine seen in humans. This is due to a lack of chitobiase activity in carnivores and ungulates; however, it has been detected in guinea pig liver (61). Therefore, stored oligosaccharides in α-mannosidosis guinea pigs may be expected to contain a single terminal N-acetylglucosamine. Further oligosaccharide analysis of guinea pig samples will clarify species differences and will indicate whether oligosaccharide metabolism in the α-mannosidosis guinea pig will provide a better model of metabolism for human α-mannosidosis.

Both the pH dependence of α-mannosidase activity and subcellular fractionation of liver extracts clearly illustrated a deficiency of acidic (lysosomal) α-mannosidase activity in affected guinea pigs. The pH profile of mannosidase activity in liver extracts from the α-mannosidosis guinea pigs also demonstrated two additional peaks in activity corresponding to Golgi (pH 5.3–5.4) and cytosolic (pH 6.2–6.4) α-mannosidases. Similar profiles have been observed in feline (16) and bovine (62) tissue homogenates. Results also demonstrate the presence of significant α-mannosidase activity at pH 4.0 in α-mannosidosis whole tissue extracts, probably attributable to the Golgi α-mannosidase activity, which was removed when lysosomes were isolated from the rest of the subcellular compartments. The relatively high levels of enzyme activity observed in skin fibroblasts in the affected guinea pigs was probably also the result of Golgi α-mannosidase activity.

Dramatic improvements in CNS and visceral storage after BMT in feline α-mannosidosis has encouraged the use of BMT for lysosomal storage diseases with CNS pathology (21). BMT has been performed in two α-mannosidosis patients, because no other effective therapies are currently available. The first patient was 7 y old and died 18 wk after transplantation. Normalization of storage and enzyme activities in liver and spleen were observed, but CNS lesions were not significantly modified by treatment (63). However, a second patient recently showed significant clinical improvement after a successful BMT, with no loss of neurocognitive function 15 mo later (64).

Enzyme replacement therapy is being developed as an alternative to BMT for a number of lysosomal storage diseases, inasmuch as the morbidity associated with BMT procedures can be avoided and earlier intervention is also possible. However, enzyme replacement therapy in lysosomal storage disease animal models such as mucopolysaccharidosis type VII has been shown to reverse pathology in visceral tissues, but provides only minimal benefit to CNS storage because of difficulties with enzyme penetration through the blood-brain barrier (65). Similarly, native human α-mannosidase, with an approximate precursor molecular mass of 110 kD that is processed into three glycopeptides of 70, 42, and 15 kD (29), would be excluded from passage across the blood-brain barrier because of size. Understanding the transport mechanisms by which other large molecules can cross this barrier may generate novel means for the passage of lysosomal proteins across this same barrier. Future studies in the α-mannosidosis guinea pig will be directed toward overcoming the blood-brain barrier. Early diagnosis and onset of therapy before development of irreversible pathology will also be critical to the overall efficacy of therapy, because axonal spheroids and ectopic dendrites that develop in response to neuronal disease were shown to persist after withdrawal of swainsonine administration in induced feline α-mannosidosis (20). Therefore it will also be important to characterize the nature of the molecular defect causing the disease, both to assist in early intervention and to increase our understanding of the disease phenotype in the guinea pigs.

Because the pathology closely resembles the same lysosomal storage disease in humans, the α-mannosidosis guinea pig model may facilitate improved understanding of disease pathogenesis and evaluation of therapeutic strategies, particularly for storage disorders with progressive CNS pathology. The guinea pig also has the advantage of small size and good fecundity, which make it a convenient small laboratory animal model.