Galactosialidosis is an autosomal recessive lysosomal storage disorder characterized by a combined deficiency of β-D-galactosidase (EC and N-acetyl-α-neuraminidase (EC, due to a primary defect of a third gene product, the PPCA(13). The three-enzyme deficiency directly relates to one of the functions of PPCA which is to associate with and protect the two glycosidases, regulating their activity and stability in lysosomes(1, 46). After the first report of a combined deficiency in a patient, originally diagnosed as variant of GM1-gangliosidosis(7), other cases were identified among patients of different ethnic origin and with variable severity of the symptoms(810). Based on clinical manifestations and age of onset of the disease, three clinical phenotypes are now distinguished: a severe early infantile form, fatal at or soon after birth, and associated with fetal hydrops, CNS involvement, visceromegaly, renal insufficiency, and skeletal and eye abnormalities; a milder late infantile type with absent or minor mental deterioration; and a juvenile/adult form, mainly found in Japan, which is characterized by slowly progressive CNS symptoms, skeletal and eye abnormalities, dysmorphism, and angiokeratoma. The number of diagnosed patients worldwide has been recently surveyed and estimated at 20 for the infantile forms and at approximately 50 for the juvenile/adult form(2).

Analysis of the primary structure of human PPCA revealed its identity with lysosomal cathepsin A, a carboxypeptidase belonging to the serine protease family of enzymes(3, 11). Indeed PPCA was shown to function as cathepsin A at acidic pH and deamidase/esterase at neutral pH(1214). All three enzymatic activities appeared to be deficient in galactosialidosis patients, although the number of patients tested has so far been limited(1418). Here, we present a survey on 20 galactosialidosis patients with different clinical phenotype. We have tested cathepsin A activity in cultured fibroblasts derived from these patients as well as their obligate heterozygote parents. Although all the patients have a severe enzyme deficiency, in some cases, there is a clear correlation between residual activity and clinical severity.


Origin and types of galactosialidosis patients. Skin biopsy and/or cultured fibroblasts of 20 galactosialidosis patients from 16 families were received either for diagnostic investigation or for basic molecular and biochemical studies. The patients are listed in Table 1 and designated by sequential numbers, GS1a to 16. They are identified by their initials and by cell strain number, as classified in our cell bank (The European Cell Bank, Rotterdam). Cells of GS13 are stored in the Camden Cell Repository under no. GM0806. The country of origin of the patients and their case reports are given in the last two columns. GS1a to 8b exhibited severe clinical manifestations within the first few months of life, sometimes at or a few days after birth. Fetal hydrops was diagnosed in cases GS3, 4, 6, and 7. Late-infantile patients were diagnosed at least 6 mo after birth and presented with either mild or severe symptoms at the time of diagnosis. GS13 (and affected sibs), as well as GS14 and 15 have been followed up at the age of 18, 15, and 19 y. GS16 had relatively mild symptoms until adult age but deteriorated after his 30th y and died at the age of 48.

Table 1 Type and origin of galactosialidosis patients

Cell culture and tissues. Cells were cultured routinely in Ham's F10 medium supplemented with FBS (10% for skin fibroblasts, 15% for amniocytes) or with 5% FBS and 4% Ultroser G (Life Technologies, Inc.) (for chorionic villus cells). Fetal liver and brain tissues were obtained after termination of the pregnancy because the fetus was diagnosed to be affected with galactosialidosis(8). Tissues were stored at-70°C until use. Whenever possible, cultured cells were maintained at low passage. They were harvested for enzymatic analysis by trypsinization, 7 d after the last subculture. Chorionic villus cultures and amniocytes were used at passage 2-3. Cell pellets were frozen at -70°C until use.

Tissue homogenates were prepared in water using a small Potter tube (100μL), and sonicated. Cultured cells were homogenized by sonication. However, for neuraminidase assay all samples were homogenized on ice in a Potter tube without further sonication and immediately used. Total protein concentration was measured by the bicinchoninic acid method(19), as specified by the manufacturer (Pierce, Rockford, IL). All extracts were diluted in water for neuraminidase assays or in saline to a protein concentration standardized for each enzyme assay and type of cell or tissue.

Enzyme assays. For cathepsin A assay, homogenates were diluted in 0.9% sodium chloride with 0.2% BSA to a protein concentration of 0.5 mg/mL for leukocytes and cultured cells, 1 mg/mL for chorionic villi, and 5 mg/mL for liver and brain. Cathepsin A activity was measured according to the methods of Taylor and Tappel(20) and Roth(21) as modified by Galjart et al.(14). Briefly, 10 μL of diluted homogenate were incubated with 90 μL of substrate solution consisting of 1.5 mMN-carbobenzoxy-L-phenylalanyl-L-alanine (Z-Phe-Ala) in 50 mM 4-morpholineethanesulfonic acid buffer, pH 5.5. After incubation for 30 min at 37°C, the sealed 96-well plates were heated on a boiling waterbath for 15 min; 10 μL of the mixed contents of each well were transferred to another 96-well plate, and 300 μL of alanine reagent were added to each well. Alanine reagent was freshly prepared by adding 500 μL ofo-phthaldialdehyde (10 mg/mL in 96% ethanol) and 500 μL ofβ-mercaptoethanol (5 μL/mL in 96% ethanol) to 30 mL of 50 mM borate buffer, pH 9.5. Fluorescence was read after 10 min in a Titertek fluorimeter(excitation 355 nm; emission 460 nm). Readings were corrected for homogenate and substrate blanks. β-Galactosidase and neuraminidase activities were measured using 4-methylumbelliferyl substrates as described elsewhere(8, 22).


Cathepsin A activity in fibroblasts from galactosialidosis patients with different clinical phenotypes. Table 2 summarizes the activities of cathepsin A, β-galactosidase, and neuraminidase measured in cultured fibroblasts from 20 patients with galactosialidosis. Values are the mean of two to three independent assays. Within the early infantile group, all patients had negligible or nondetectable cathepsin A activity; measurements were lower than 4 mU/mg of protein, which is less than 1.5% of mean control level. In the groups of late-infantile and juvenile/adult patients, activities ranged from 4 to 12 mU/mg of protein, corresponding to 1.5-5% of the mean control level. In two cell strains, GS13 and GS16, we consistently detected a residual activity higher than in other samples. Repeated measurements of GS13 (late-infantile) resulted in an average activity of 12.3 mU/mg (n = 5; SD 3.7), whereas three independent assays of GS16 (juvenile/adult) gave a mean of 10.0 mU/mg (SD 4.2). All cell strains had strongly reduced β-galactosidase and neuraminidase activities, which has so far been the hallmark of galactosialidosis.β-Galactosidase ranged from 39 to 105 nmol/h/mg of protein(i.e. 5.6-15.2% of controls) with the highest level present again in cells of patient GS13, who also had the highest cathepsin A activity. For all other patients, the amount of residual cathepsin A activity had no correlation with the levels of residual β-galactosidase. Also for neuraminidase, the highest residual activity (5.5 nmol/h/mg) was found in cells of patient GS13.

Table 2 Cathepsin A, β-galactosidase, and neuraminidase activities in fibroblasts of the 20 patients with galactosialidosis and patients with related lysosomal storage diseases

The results for the galactosialidosis patients in Table 2 are compared with the activities of the three enzymes in patients with other lysosomal storage disorders associated with deficiencies ofβ-galactosidase and/or neuraminidase. Nearly complete absence ofβ-galactosidase in GM1-gangliosidosis or neuraminidase in sialidosis had no apparent effect on cathepsin A activity. As anticipated, fibroblasts of patients with mucolipidosis II (I-cell disease), which is associated with multiple hydrolase deficiency, had indeed very lowβ-galactosidase, neuraminidase, and cathepsin A activities. Fibroblasts of two mucolipidosis III patients with a milder form of I-cell disease showed only partial deficiency of cathepsin A and β-galactosidase (20-25%) but a remarkably low neuraminidase (5%).

Considerable cathepsin A activity was detected in normal leukocytes (not shown). This should allow the use of blood samples for the (postnatal) diagnosis of galactosialidosis (range: 71-146 mU/mg; mean 106, SD 20,n = 17). Samples from patients were, however, not available for investigation.

Enzyme activities in obligate heterozygotes. Fibroblast strains of 11 parents of galactosialidosis patients were tested for the three enzymes under study. All cell strains were tested at least twice in a series of three experiments. The mean activities are presented in Table 3 and compared with average control values. Two of the heterozygotes (F12 and M12) had cathepsin A activities well below the total control range. The activity in the other nine cell strains fell in the lower half of the control range. The average cathepsin A activity for the 11 carriers is at 60% of the mean control level. β-Galactosidase activity assayed in the same series of experiments was not reduced; the mean level for the carriers was 110% of the average control value. A similar comparison for neuraminidase showed an overall slight reduction from the mean value of controls for the 11 carriers tested (84%). However, the heterozygotes, F12 and M12, with clearly reduced levels of cathepsin A activity, also exhibited neuraminidase activities below the control range.

Table 3 Enzyme activities in cultured skin fibroblasts of obligate heterozygotes for galactosialidosis

Cathepsin A activity in fetal cells and tissues. Cathepsin A activity was assessed in normal cultured amniotic fluid cells and chorionic villi. Although clearly lower than in cultured skin fibroblasts, the activity in amniotic fluid cells (n = 11, 64-129 mU/mg; mean 92 ± 23) and chorionic villi (n = 11, 18-61 mU/mg; mean 40 ± 14) can be accurately measured in small samples. Since 1979 we have investigated eight pregnancies at risk for galactosialidosis (families 4, 5, and 12). The prenatal diagnosis in these cases was based on β-galactosidase and neuraminidase activities in cultured amniotic fluid cells, or chorionic villi, as the involvement of cathepsin A was not known at that time. Normal enzyme activities were shown in seven cases, and the children born were not affected. In one previously reported case(8), clear deficiency ofβ-galactosidase and neuraminidase indicated an affected fetus, which was confirmed by the demonstration in skin fibroblasts of the deficiency of the same two enzymes and later of cathepsin A (case GS4b; Table 1).

Brain and liver tissues of patient GS4a, who died at birth with severe hydrops fetalis, and of the aborted fetus, GS4b, were previously shown to be deficient in β-galactosidase and neuraminidase(8). Cathepsin A in these tissues appeared completely absent, whereas considerable activity was present in control fetal samples which had been stored for a similar length of time at -70°C: brain, controls (n = 3) 4.6-5.0 mU/mg; patient and fetus 0.0 mU/mg; liver, controls (n = 2) 44-55 mU/mg; patient 1.4 mU/mg.


The carboxypeptidase activity of PPCA was discovered after the identification of the protein's primary structure(3). This original finding made galactosialidosis the first lysosomal storage disorder associated with a primary defect of a lysosomal protease, the cathepsin A. Reduced or deficient cathepsin A activity was indeed demonstrated in fibroblasts of galactosialidosis patients with the early infantile, late infantile, and juvenile/adult type of the disease(1418).

We have now surveyed 20 galactosialidosis patients from 16 unrelated families: 12 patients with the early infantile type and eight with the late infantile or juvenile/adult type. Accurate assessment of residual activity of cathepsin A in galactosialidosis has become possible by using a fluorescence-based method which is more sensitive than standard colorimetric assays. Cathepsin A was completely deficient (0-2%) in all patients with the early infantile type of galactosialidosis. In all patients with the late infantile or juvenile/adult form, some residual activity was measured (2-5%), the highest activity being consistently found in two previously described patients (GS13 and GS16) with mild phenotype and slow progression of the disease(27, 30, 31). Thus, these results suggest a correlation between residual cathepsin A activity and clinical phenotype. However, because the clinical manifestations in galactosialidosis resemble, to some extent, those of sialidosis patients with an isolated neuraminidase deficiency, the severity of the symptoms may be primarily related to the absence of neuraminidase activity rather than the deficiency of cathepsin A as such. Indeed neuraminidase is nearly absent (<2%) in all severe early infantile patients, whereas in the late infantile patients higher activities were measured (2-4%); the highest activity (10%) was found in the mild phenotypic patient GS13. Apparently the small amount of PPCA in this patient is able to rescue some of the neuraminidase molecules. It should be noted, however, that cultured fibroblasts may not reflect adequately the effects of protective protein deficiency on neuraminidase andβ-galactosidase in important tissues such as brain and liver.

In obligate heterozygotes cathepsin A activity appears to be reduced, although there is large overlap between the ranges for the nine carriers and the controls. In seven carriers with cathepsin A activity in the lower half of the normal range, β-galactosidase and neuraminidase activities were normal. This suggests that PPCA is normally present in excess with respect to complex formation and “protective capacity” towardβ-galactosidase and neuraminidase. However, in two heterozygotes (parents of patient GS12) with rather low cathepsin A (25% of controls) neuraminidase activity was also reduced to a level below the normal range. We have previously shown that patient GS12 is homozygous for a mutation which leads to a substitution of the Phe412 with Val in PPCA and that the mutant protein fails to dimerize in the ER(32). In the parents, heterozygosity for this mutation may result in the production of mutant, as well as normal, PPCA. It may be speculated that mutant (F412V) monomers in the ER trap some of the normal monomers into a dimeric form, which is, however, not fully competent to leave the ER (i.e. F412V may be regarded as a “dominant negative mutation”). The amount of normal protective protein dimers leaving the ER would thereby be reduced from the expected 50% in a heterozygote to perhaps 25%, which would explain the remarkably low cathepsin A activities in these parents. At this low level the functional amount of PPCA may become limiting for the protection of neuraminidase.

The protective role of cathepsin A toward β-galactosidase and neuraminidase is not reciprocal: complete deficiency of β-galactosidase in GM1-gangliosidosis patients and complete deficiency of neuraminidase in patients with mucolipidosis I does not exert any influence on cathepsin A activity. In fibroblasts of patients with I-cell disease (mucolipidosis II) cathepsin A is severely reduced, as are β-galactosidase, neuraminidase, and most other lysosomal enzymes.

Our studies have so far revealed one general type of defect: the simultaneous deficiencies of catalytic and protective function of PPCA, resulting in secondary impairment of β-galactosidase and neuraminidase leading to galactosialidosis. This could provisionally be designated as a type 1 PPCA defect. We have previously shown that the two functions of PPCA are fully separable, because an in vitro mutagenized protein that lacks cathepsin A activity is able to correct β-galactosidase and neuraminidase, when endocytosed by galactosialidosis fibroblasts(14). Hypothetically, two other types of defect may be envisaged: type 2 with normal cathepsin A activity but defective protective function and, therefore, β-galactosidase and neuraminidase deficiencies, and type 3 with defective cathepsin A activity but a normal protective function and normal β-galactosidase and neuraminidase activities. A type 2 disorder may largely resemble the known clinical type(s) of galactosialidosis; therefore, if this disorder exists, it should probably be found among galactosialidosis patients by screening routinely for cathepsin A activity, as we have done in the present study. It is not known whether, or to what extent, cathepsin A activity per se contributes to the clinical manifestations of galactosialidosis because the clinical consequences of a specific deficiency of cathepsin A activity are not known. In the latter case, a block in the degradation of selected bioactive peptides, which may be the target of cathepsin A in vivo, may be overcome by the action of other exo- and/or endopeptidases.

Prenatal diagnosis of galactosialidosis is reliably possible by the investigation of the secondary deficiencies of β-galactosidase and neuraminidase in amniocytes, as was first shown by us(8) and confirmed more recently by Sewell et al.(24). Several prenatal analyses have been made in pregnancies at risk for galactosialidosis by chorionic villus sampling in our center and elsewhere, but affected cases have not been reported. Although the prenatal diagnosis may be made by β-galactosidase and additionally neuraminidase assay, there are reasons to include cathepsin A in the analysis. First, β-galactosidase activity in normal chorionic villi is relatively low (compared with cultured fibroblasts), whereas a considerable residual but still unknown level of activity must be expected in the villi of an affected pregnancy, such as in fibroblasts of a patient. Second, neuraminidase is generally low in villi, so that a reliable distinction between an affected and a normal or heterozygous fetus is not always possible. It is therefore suggested to measure cathepsin A activity as an essential parameter in the prenatal diagnosis of galactosialidosis.