Altered Trafficking and Processing of GALC Mutants Correlates with Globoid Cell Leukodystrophy Severity

Abstract

Globoid cell leukodystrophy (GLD, Krabbe disease) is due to autosomal recessive mutations in the lysosomal enzyme galactosylceramidase (GALC). Many GLD patients develop infantile-onset of progressive neurologic deterioration and death by 2 years of age, whereas others have a later-onset, milder disease. Cord blood transplant slows disease progression much more effectively when performed presymptomatically, highlighting the importance of early diagnosis. Current diagnosis is based on reduced GALC activity, DNA sequence, and clinical examination. However, presymptomatic diagnosis is hampered by imperfect genotype-GALC activity-phenotype correlations. In addition, three polymorphisms in the GALC gene are variably associated with disease mutations and have unknown effects on GALC activity and disease outcome. Here, we study mutations that cause infantile or later-onset GLD, and show that GALC activity is significantly lower in infantile versus later-onset mutants when measured in the lysosomal fraction, but not in whole-cell lysates. In parallel, infantile-onset mutant GALCs showed reduced trafficking to lysosomes and processing than later-onset mutant GALCs. Finally, the cis-polymorphisms also affected trafficking to the lysosome and processing of GALC. These differences potentially explain why the activity of different mutations appears similar in whole-cell extracts from lymphocytes, and suggest that measure of GALC activity in lysosomes may better predict the onset and severity of disease for a given GLD genotype.

Significance statement: Globoid cell leukodystrophy (GLD, Krabbe disease) is diagnosed by measuring galactosylceramidase (GALC) activity and DNA analysis. However, genotype and phenotype often do not correlate due to considerable clinical variability, even for the same mutation, for unknown reasons. We find that altered trafficking to the lysosome and processing of GALC correlates with GLD severity and is modulated by cis-polymorphisms. Current diagnosis of GLD is based on GALC activity of total cell lysates from blood, which does not discriminate whether the activity comes from the lysosome or other subcellular organelles. Measurement of GALC activity in lysosomes may predict which infants are at high risk for the infantile phenotype while distinguishing other children who will develop later-onset phenotypes without onset of symptoms for years.

Keywords: Krabbe disease; galactosylceramidase; globoid cell leukodystrophy; lysosomal processing; lysosomal storage disorders; protein trafficking.

Figure 1.

GALC trafficking and processing in transfected cells is impaired in infantile-onset mutants but not in a later-onset mutant. A, A scheme is shown of GALC structure and common missense mutations found in GLD of European ancestry (Wenger, 2011a). The two most common infantile-onset missense mutations of GLD are T529M (exon 14) and Y567S (exon 15). The most frequent later-onset mutation is G286D (exon 8). B, Forty-eight hours after transfection of HEK-293T cells, total cell lysates were analyzed with a polyclonal anti-GALC antibody (Lee et al., 2010) that detects the 80 kDa precursor, and both the 50 and 30 kDa processed forms of GALC protein. There were nonspecific bands ∼90 and 40 kDa (asterisks). β-Tubulin was used for loading control. FLAG and HA fused to the C terminus of GALC were comparable with the untagged protein, in contrast fusion with eGFP markedly decreased GALC expression. GALC-FLAG and GALC-HA had similar enzyme activities to the nontagged GALC, whereas GALC-GFP had lower enzymatic activity. Oligodendrocyte GALC enzymatic activities at 4 DIV from WT and Galc-deficient twitcher mice, an authentic mouse model of GLD, were used for comparison. C, Compared with wild-type, the later-onset mutation (G286D) is significantly processed, whereas the infantile forms (T529M and Y567S) are much less processed. D, Lysosomal fractions, which were purified from the lysate of GALC-transfected HEK-293T cells and analyzed by Western blotting with anti-GALC, β-tubulin, and LAMP2 antibodies, contain mostly the processed forms of GALC, but not the precursor form. In the lysosomal fractions, there are almost no processed forms of GALC for the infantile forms (T529M and Y567S) in contrast to the later-onset form (G286D). E, F, In total cell lysates, the GALC activity of the later-onset mutant (G286D) is indistinguishable from that of the infantile forms (T529M and Y567S), whereas in lysosomal fractions, there is residual GALC activity for the later-onset mutant, but almost no activity for the infantile-forms, revealing a correlation between lysosomal GALC activity and severity of disease. Error bars indicate SD of triplicates. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test). n.s., Not significant. G, Immunofluorescence for WT or the three mutant GALCs in HEK-293T cells with anti-GALC (red) and anti-KDEL, anti-58k, anti-EEA1, or anti-LAMP2, antibodies (all green) showed that WT and later-onset mutant GALC colocalized with all four subcellular markers, whereas the infantile-onset GALC mutants were not (or minimally) detectable in the lysosome, and mainly found in the ER and Golgi. The same analyses in transfected oli-neu cells at 4 DIV (H) and primary rat Schwann cells (I) with anti-GALC (red) and anti-LAMP2 (green) confirmed that the WT GALC and later-onset mutant are present in the lysosome, whereas the infantile-forms are not. Arrows indicate overlapping lysosomal puncta and GALC (yellow). Orthogonal views (XY, XZ, and YZ) of a z-stack are shown. White lines indicate section positions. Nuclei were labeled with DAPI. DIV, Days of in vitro differentiation. Scale bars: G–I, 5 μm.

Figure 2.

Benign cis-polymorphisms affect total enzymatic activity and trafficking of GALC. A, Diagram of the genomic structure of GALC and common benign polymorphisms (blue) found in GLD patients. A new numbering classification, starting from the upstream translational start site, was used to define the residue numbers. The G286D later-onset mutation is always found with the c.550T or c.1685C polymorphism. The T529M infantile-onset mutation is always found with the c.550T, c.742A, or c.1685C polymorphism. The Y567S infantile-onset mutation is always found with the c.1685C polymorphism. These three common polymorphisms were introduced by site-directed mutagenesis. B, GALC activities were measured in total lysates prepared from cells transfected with wild-type or mutant GALC in combination with indicated cis-polymorphisms. Both c.550T and c.742A reduced WT GALC activity, whereas only c.550T or c.1685C reduced G286D or Y567S activity, respectively. Error bars indicate SD of triplicates. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test). C, HEK-293T cells transfected with WT or mutant GALCs in combination with indicated cis-polymorphisms were costained for GALC (red) and LAMP2 (green). Nuclei were labeled with DAPI. Orthogonal views (XY, XZ, and YZ) of a z-stack. White lines indicate section positions. D, Quantification of colocalized GALC and LAMP2 signals was obtained by counting the number of yellow puncta. C, D, All polymorphisms reduced lysosomal targeting of WT GALC. The c.550T and c.1685C polymorphisms reduced lysosomal targeting of G286D and Y567S, respectively. More than three independent experiments were performed with qualitatively similar results. Quantification was performed on 8–16 cells. Error bars indicate SD. **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test). Scale bar: C, 5 μm.

Figure 3.

cis-polymorphisms affect GALC processing. A, Lysates from HEK-293T expressing GALC mutations in combination with various cis-polymorphisms were analyzed by Western blot using anti-GALC antibody. Both the GALC mutations and some of the polymorphisms affect the level and processing of GALC. B, Quantification of levels in four independent experiments, normalized by cotransfection of luciferase plasmid, showed that both the c.550T and c.742A polymorphisms reduced total GALC (defined as the sum of the GALC levels for the 80, 50, and 30 kDa forms) in the WT context, but c.1685C did not have any effect. In the case of later-onset G286D mutant, both c.1685C and c.550T reduced total GALC levels. The c.1685C polymorphism reduced total GALCs of both T529M and Y567S. Interestingly, none of the mutations with (or without) polymorphisms completely abolished GALC expression. Error bars indicate SD of triplicates. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test). C, The ratio of 50 versus 80 kDa levels and (D) 30 versus 80 kDa levels showed that both c.550T and c.742A reduce the processing of WT. Both c.1685C and c.550T reduce processing of the later-onset G286D GALC. None of the polymorphisms had any effect on the infantile-onset T529M GALC, whereas c.1685C reduced the infantile-onset Y567S GALC processing, indicating a critical pathogenic role of this polymorphism in patients harboring the Y567S mutation. Error bars indicate SD of triplicates. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test).

Figure 4.

cis-polymorphisms affect the level of precursor 80 kDa GALC. A, Lysates of cells transfected with the GALC mutants and cis-polymorphisms were analyzed by Western blot with anti-FLAG antibody, which detected only the 80 kDa precursor form. The FLAG antibody did not detect the 30 kDa fragment, suggesting that the FLAG tag at the C terminus might slightly modify the 30 kDa structure and mask the epitope. B, Quantification of the 80 kDa GALC densities showed that the c.550T polymorphism increased the unprocessed WT GALC, whereas none of the polymorphisms significantly affected the unprocessed G286D GALC. Both c.1685C and c.742A polymorphisms reduced the unprocessed level of T529M GALC. The polymorphism c.1685C did not have much effect on the unprocessed Y567S GALC. Error bars indicate SD of triplicates. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test).

Figure 5.

GALC in the ER has substantial enzyme activity. A, Purified ER fractions from total cell lysates of transfected HEK-293T cells showed that the 80 kDa GALC precursor form is present in the ER. GRP78 was used as an ER loading control. B, GALC activity assay performed with ER fractions reveals that the precursor protein possesses some enzymatic activity. Error bars indicate SD of triplicates. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test). C, Purified lysosomal fractions contain mainly the processed GALCs. D, The same lysosomal fractions were also used for determination of GALC activities. Error bars indicate SD of duplicates. E, The GALC activities from ER (red bar) and lysosome (blue bar) were combined to show the relative contributions to actual cellular activity. F, Percentage of the activities from each subcellular organelle shows that the GALC in ER contributes significantly to the total activities in the mutants (G286D, T529M, and Y567S), but not in WT.

Figure 6.

Additional later-onset GALC mutations also retain partial trafficking and processing of GALC in lysosomes. Two more later-onset (G57S and L634S) and one more infantile (D544N) GALC mutations were generated with or without the c.1685C polymorphism. Measurements of GALC activities in both total lysates (A) and lysosomal fractions (B) from HEK-293T transfected with the mutant constructs show that both later-onset mutants G57S and L634S have significantly higher residual GALC activities compared with the infantile D544N form, revealing a strong correlation between residual GALC activity and severity of disease. Error bars indicate SD of triplicates. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test). n.s., Not significant. C, In total cell lysates, both later- and infantile-onset mutations are processed less than the WT. However, in lysosomal fractions, the later-onset forms (G57S and L634S) are more processed than the infantile form (D544N). D, Immunofluorescence for the three mutant GALCs in HEK-293T cells with anti-GALC (red) and anti-LAMP2 antibodies (green) shows that later-onset mutant GALCs colocalized (yellow) more with the lysosomal marker, whereas the infantile-onset GALC mutant was minimally detectable in the lysosome. Orthogonal sections (XY, XZ, and YZ) of a z-stack are shown. White lines indicate the position of sections. Nuclei were labeled with DAPI. Scale bar, 5 μm.

Figure 7.

Lysosomal fractions of dermal fibroblasts from later-onset patients have significantly higher GALC activities than those from infantile-onset patients. A, The table shows clinical and molecular information of GLD dermal fibroblasts used in this study, which were from published (Tappino et al., 2010) and unpublished samples (Mirella Filocamo, Telethon Biobank, Italy). K359AfsX3 (lysine 359 substituted by alanine followed by a frame-shift translational termination) is located close to the frequent 30 kb deletion GLD mutation and therefore will generate a truncation similar to the 30 kb deletion. B, GALC activities in total cell lysates or lysosomal fractions of the dermal fibroblasts from GLD patients and age-matched normal controls (<4 years old). GALC activities from total cell lysates of the later-onset mutants are many times indistinguishable from those of the infantile forms. However, GALC activities from the lysosomal fractions showed that the later-onset mutants have more enzyme activity in the lysosome and are more reliably distinguished from those of the infantile forms. WT-1 is male and WT-2 is female. Error bars indicate SD of triplicates. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test). n.s., Not significant.

Figure 8.

4-PBA enhances GALC levels and activity. A, GALC levels were measured from total lysates of transfected cells expressing WT or mutant GALC in combination with various cis-polymorphisms after treatment with 2 mm 4-PBA for 24 h. The chaperone increased the 80 kDa GALC precursor in WT and in all mutants, but not relative levels of processed GALC fragments. B, GALC activities were measured from transfected cells expressing the WT and G286D GALCs with or without treatment with 2 mm of 4-PBA. 4-PBA significantly enhanced the GALC activities in cells expressing WT, G286D, and T529M GALCs. *p < 0.05 (Student's t test). **p < 0.01 (Student's t test). ***p < 0.001 (Student's t test). n.s., Not significant. C, The addition of 4-PBA for 24 h increased the trafficking to the lysosome (LAMP2; green) of WT and all mutant GALCs (red) in transfected HEK-293T cells, possibly due to the overall increase in GALC precursor by 4-PBA. Orthogonal views (XY, XZ, and YZ) of a z-stack. White lines indicate section positions. Scale bar: C, 5 μm.

Figure 9.

Model for how measured GALC activities may not correlate with disease severity in GLD; 5% of normal GALC activity is regarded as the general cutoff for diagnosis of GLD. The diagram represents two different possibilities for the localization of GALC that can produce the same whole-cell lysate GALC activity when measured in patient blood. A, If the majority of remaining GALC is <5% but is found concentrated in the lysosome, then the patient will have milder phenotype and present with a later-onset form of the disease. B, If, however, most of the remaining GALC is localized to the ER or Golgi, due to defective trafficking, then the patient will have a more severe phenotype and may manifest with the infantile-onset form of the disease.