Glucocerebrosidase deficiency in zebrafish affects primary bone ossification through increased...

OR I G INA L ART I C L E

Glucocerebrosidase deficiency in zebrafish affectsprimary bone ossification through increased oxidativestress and reduced Wnt/β-catenin signalingIlaria Zancan1, Stefania Bellesso2,†, Roberto Costa1,†, Marika Salvalaio2,Marina Stroppiano4, Chrissy Hammond5, Francesco Argenton3,Mirella Filocamo4, and Enrico Moro1,*1Department of Molecular Medicine, 2Department Women’s and Child’s Health, 3Department of Biology,University of Padova, I-35131 Padova, Italy, 4Centro di Diagnostica Genetica e Biochimica delle MalattieMetaboliche Istituto Giannina Gaslini, Genova 16147, Italy and, and 5Department of Biochemistry, Physiology& Pharmacology, University of Bristol, BS8 1TD Bristol, UK

*To whom correspondence should be addressed at: Department of Molecular Medicine, University of Padova, Via Ugo Bassi 58/B, I-35121 Padova, Italy.Tel: +39 498276341; Fax: +39 49 8276300; Email: [email protected]

AbstractLoss of lysosomal glucocerebrosidase (GBA1) function is responsible for several organ defects, including skeletal abnormalitiesin type 1 Gaucher disease (GD). Enhanced bone resorption by infiltratingmacrophages has been proposed to lead tomajor bonedefects. However, while more recent evidences support the hypothesis that osteoblastic bone formation is impaired, a clearpathogenetic mechanism has not been depicted yet. Here, by combining different molecular approaches, we show that Gba1loss of function in zebrafish is associated with defective canonical Wnt signaling, impaired osteoblast differentiation andreduced bone mineralization. We also provide evidence that increased reactive oxygen species production precedes the Wntsignaling impairment, which can be reversed upon human GBA1 overexpression. Type 1 GD patient fibroblasts similarly exhibitreduced Wnt signaling activity, as a consequence of increased β-catenin degradation. Our results support a novel model inwhich a primary defect in canonical Wnt signaling antecedes bone defects in type 1 GD.

IntroductionImpaired lysosomal acid β-glucocerebrosidase activity and accu-mulation of undegraded glucosylceramide (GC) are the hallmarksof Gaucher disease (OMIM #230800, #230900, #2301000), the mostcommon lysosomal storage disorder, affecting one every 60 000people worldwide (1). More than 330 different mutations havebeen identified in the glucocerebrosidase (GBA1) gene encodinglysosomal acid β-glucocerebrosidase and most severe forms ofthe disease are often associated with compound heterozigosity

(2) (www.hgmd.org). Patients have been traditionally dividedinto three major groups on the basis of the absence or presenceand severity of primary involvement of central nervous system.In patients with the non-neuronopathic type 1 form (OMIM#230800), the burden of affected organs includes hepatomegaly,splenomegaly, leukocytopenia, thrombocytopenia and skeletalalterations, which often are recognizable as ‘Erlenmeyer flaskdeformities’ and progressive osteopenia (3).

The etiology of bone alterations has beenmatter of debates inthe past few years since two opposing hypotheses have been

†These authors contributed equally to this work.Received: August 29, 2014. Revised: October 13, 2014. Accepted: October 14, 2014.

© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Human Molecular Genetics, 2014, 1–15

doi: 10.1093/hmg/ddu538Advance Access Publication Date: 17 October 2014Original Article

1

HMG Advance Access published November 10, 2014 at U

niversitÃ degli Studi di Padova on N

ovember 24, 2014

http://hmg.oxfordjournals.org/

Dow

nloaded from

www.hgmd.org

www.hgmd.org

www.hgmd.org

http://www.oxfordjournals.org


proposed. The former suggests that glycolipid-laden macro-phages, namely ‘Gaucher cells’, infiltrate the bone marrow andtrigger an increased cytokine-dependent bone resorption byosteoclast activity (4). In the latter hypothesis, supported byboth recent clinical evidences and studies in a conditionalmouse model, an impaired osteoblastic bone formation is pro-posed to lead to unbalanced bone remodeling, which exacerbatesduring late disease stages (5,6). While the inhibition of osteoblastprecursor cell proliferation by GC has been evoked (5), the mech-anistic link between loss of glucocerebrosidase function andbone abnormalities is still missing.

Among cellular pathways affecting bone remodeling, canon-ical Wnt signaling has been shown to play a key role in bonehomeostasis (7).Wnt pathway activation relies on a complex cas-cade of soluble factors, that, upon binding to a dual receptor com-plex (FRZ/LRP5 or FRZ/LRP6), trigger the inactivation of acytoplasmic multiprotein complex and relieve the signaling me-diator, β-catenin, from its proteasomal degradation. In the pres-ence of Wnt ligands, undegraded β-catenin translocates to thenucleus where it associates with TCF/LEF transcription factorsto promote target gene transcription (8). Mutations, affectingthe Wnt pathway at different levels (extracellular and cytoplas-mic), have been described in several bone-related disorders andpolymorphisms in the Wnt pathway modulators, DICKKOPF-1(DKK-1) and SCLEROSTIN 1 (SOST-1), have been associated withbone mineral density (9).

Herewe describe a complementary approach based on theuseof zebrafish transgenesis and geneticmanipulation togetherwithin vitro analysis to dissect type 1 Gaucher disease-related bonepathogenesis.

We first generated a loss of Gba1 function fish model by tar-geting the splicing donor site of Gba1 exon 2 by means of anti-sense morpholino oligos. We show that increased oxidativestress and dysfunctional canonical Wnt pathway transductionoccur as a consequence of Gba1 impairment. We next character-ized a stable genetic mutant, called sa1621, in which a point mu-tation in the splicing donor site of Gba1 exon 4 is present. Wedemonstrate that sa1621 homozygote mutants (gba1sa1621/sa1621)display a significant decrease of canonical Wnt pathway activ-ity. We then provide evidences that both models display osteo-penia arising from reduced osteoblast differentiation anddecreased bone mineralization. We finally show that canonicalWnt pathway impairment occurs in fibroblasts of Gaucher dis-ease affected patients and propose the use of soluble Wnt path-way modulators as alternative reliable bone serum biomarkersfor GD patients.

ResultsGba1 mutants and morphants display boneabnormalities

A strong aminoacidic conservation (53.5% identity, 68.1% similar-ity) between the fish Gba1 (Genbank Accession no XP_687471.3)and human GBA1 (Genbank Accession no NP_000148.2), and anearly expression of zebrafish gba1 transcriptswerefirst confirmed(Supplementary Material, Fig. S1). We, therefore, designed a spli-cing-blocking morpholino oligo (Gba1MO), targeting the splicingdonor of Gba1 exon2. We ruled out the use of a designed transla-tion-blockingmorpholino, targeting the ATG initiation codon, forapparent off-target effects. As control, we designed both a five-basemismatch-controlmorpholino targeting the same sequenceand afive-base unrelatedmismatch-controlmorpholino.We firsttitrated the efficiency of themorpholino and found that at a dose

of 1.75 mg/ml (∼17.5 pg per embryo), a 60% of exon 2 skippingwasproduced (Fig. 1A). Higher doses of Gba1MOdid not increase exon2-skipping efficiency. Indeed, the specificity of the Gba1MO wasconfirmed by RT-PCR, which revealed aberrant splicing eventsin 2 dpf morphants. Sequencing of the misspliced transcript,in fact, confirmed exon 2 skipping and generation of an aberranttranscript (see chromatograms in Fig. 1A). To further demon-strate the efficient targeting of Gba1 at a proteomic level, weperformed western blots using a cross-reacting antibody andfound in morphants almost a 5-fold decrease of total Gba1protein levels (Fig. 1B). Morphant fish did not display gross ab-normalities, except a slight curvature of the trunk. Body lengthwas not statistically different from that of control fish (datanot shown).

We next investigated whether in fish morphants bone devel-opment and mineralization were affected. To perform thesetasks, we carried out in situ hybridizations for the hyperthrophicchondrocyte col10a1 and early osteoblast-specific runx2b mar-kers, which label distinct populations involved in endochondraland intramembranous ossification processes during skeletal de-velopment, respectively. We found that the expression of bothmarkers was significantly reduced in Gba1 morphant (Fig. 1C).We next microinjected the splicing morpholino in Tg(Ola.Sp7:NLS-GFP)zf132 transgenic fish, which label osterix(sp7)-expressingcells. Confocal analysis at 4 dpf demonstrated a significant re-duction of osterix-transgene expression, thus confirming anearly dysfunction of the osteoblast population (Fig. 1D). Never-theless, we asked whether the cartilage tissue was similarlyimpaired, but Alcian blue staining, labeling mucopolysacchar-ides in cartilage (Fig. 1E) and confocal analysis of Tg(Col2a1aBAC:mCherry)hu5900 (data not shown), demonstrated no evidentalterations. Indeed, multiple Alizarins staining on 10 days post-fertilization (dpf) larvae showed amarked decrease in bonemin-eralization occurring in morphants at both the cephalic regionsand vertebrae centra levels (Fig. 1E).

To better characterize the type of vertebral abnormalities,we sought to perform thin and ultrathin sections at earlierstages (6 dpf ). In Gba1 morphants, we detected morphologicalalterations in the notochord, ranging from outstretched noto-chord cells to a collapsed notochord in fishwith themost severephenotype. On TEM analysis, we observed distinct ultrastruc-tural phenotypes, which included alterations of the medialnotochord sheath, deteriorated mitochondria and notochordcell shape, and a very limited accumulation of intracellularvesicles (Supplementary Material, Fig. S2). Reduced osteoblastdifferentiation and bone mineralization in Gba1 morphantswere not secondary to defects in blood supply, as blood vesselarchitecture and microarchitecture were not affected in Tg(krdl:EGFP)s843 microinjected with Gba1 morpholino (Sup-plementary Material, Fig. S3, upper panel). We also ruled out apossible defect in the differentiation of neural crest cells,as Gba1 knockdown did not modify transgene expression inTg(sox10:mRFP)vu234 (Supplementary Material, Fig. S3, bottompanel). No apparent increased levels of bone resorption werealso identified by tartrate-resistant acid phosphatase (TRAP)staining at 14 dpf in Gba1 morphants (Supplementary Material,Fig. S4).

To support the findings of a primary defect in the differenti-ation of early bone-forming cells, we took advantage of a mutantretrieved from a forward genetic screening (gba1sa1621/sa1621).These mutant fish are characterized by a single G>A substitutionin the splicing donor site of exon4, which produces an aberranttranscript with a premature stop codon. The translated Gba1mu-tant protein is made of 190 amino acids, instead of the 518 amino

2 | Human Molecular Genetics

at UniversitÃ

degli Studi di Padova on Novem

ber 24, 2014http://hm

g.oxfordjournals.org/D

ownloaded from

http://hmg.oxfordjournals.org/lookup/suppl/doi:10.1093/hmg/ddu538/-/DC1








acids of thewild-typeGba1 protein (Fig. 1F). Mutant fish exhibitedsignificantly shortened body axis beginning from 7 dpf, and amarked trunk curvature, when compared with age-matchedcontrol siblings (Fig. 1F and Data not shown). In situ hybridiza-tions on fixed larvae derived from incrosses between Gbamutant

carriers (gba1sa1621/+) demonstrated an evident decrease ofrunx2b and col10a1 expression in almost 25% of fish at 2 and4 dpf, respectively (Fig. 1G). We next evaluated bone minerali-zation at 10 dpf by Alizarin red staining and, as shown inFigure 1H, we found a strong decrease of calcium deposition in

Figure 1. Bone-related genes expression and bone mineralization are impaired in GBA1 loss-of-function models. (A) (left) RT-PCR on cDNAs from 48 hpf larvae

microinjected with 1.5 mg/ml (1), 1.75 mg/ml (2), 2 mg/ml (5) Gba1 MO and 1.75 mg/ml (3), 2 mg/ml (4) control morpholino. β-Actin was used as internal control.

Arrowhead indicates the shorter fragment produced by antisense morpholino-mediated knockdown. (middle) Representative chromatograms showing the exon 2

skipping produced by morpholino injection. The arrowhead indicate the nucleotide level at which the exon-skipping is produced. (right) Comparison between

representative control larva (mismatch) and a morpholino injected (Gba1 MO), at 3 dpf, showing the lack of apparent phenotypic differences. Only a mild curvature of

the trunk is detectable. (B). Representative western blot, showing the marked decrease of Gba1 protein levels after morpholino-mediated knockdown. Protein extracts

are from 100 2 dpf mismatch-control and morphant microinjected larvae. Protein extract from the human Hek-293 cell line was used as control for protein band

detection. β-Actin was used as loading control. (C) (left) Reduced runx2b levels in 2 dpf Gba1 morphants. A representative larva is shown for each group (control and

morphants). In the bottom panel, a magnification of the runx2b-stained area is shown. (right) Reduced col10a1 mRNA levels in 4 dpf Gba1 morphants. A representative

larva is shown for each group (control and morphants). In the bottom panel, a magnification of the col10a1-stained opercle (op) and brachiostegal (bs) bone is shown.

(D) Reduced transgene expression in Tg(Olasp7:nuGFP) fish after Gba1 knockdown. On the left a representative confocal Z-stack projection is shown: on the right

ImageJ-based analysis of the opercle volume in control (N = 5) and morphants (N = 5). (E) Alcian blue and Alizarin staining in a representative mismatch-control larva

and Gba1 morphant at 10 dpf, showing lack of vertebrae centra ossification in Gba1 morphants. In the small inset a magnification of the cephalic area, showing the

spongy aspect of parasphenoid bone in morphants. On the right, a graphical representation of the statistically significant difference in the percentage of larvae with

more than eight ossified centra between control (N = 70) and morphants (N = 108). (F) (left) Representative chromatograms of wild-type siblings and gba1sa1621/sa1621

fish, showing the G to A substitution detectable in mutants. The produced truncated protein is shown in grey. (right) Representative 3 dpf wild-type sibling and

gba1sa1621/sa1621fish are shown. Note the marked curvature of the trunk in the mutant. (G) Reduced runx2b (left) and col10a1(right) mRNA levels in gba1sa1621/sa1621

detected by in situ hybridizations with antisense riboprobes. (H) Decreased bone mineralization in gba1sa1621/sa1621fish detected by Alizarin red staining. A small inset

highlighting the reduced opercle size in mutant fish is shown. The numbers below each panel indicate the fraction of fish with the observed phenotype. In all panels,

fish are represented with anterior to the left. Data are expressed as mean ± SEM of three independent experiments (*P < 0.05; t-test).

Human Molecular Genetics | 3

at UniversitÃ




ownloaded from


Figure 2. Loss of Gba1 function affects different cell populations. (A) Gba1 functional knockdown affects thrombocyte formation and differentiation. Gba1 MO-injected

larvae show an evident decrease of transgene expression in Tg(-6.0itga2b:EGFP)la2 fish at 5 dpf detected by in situ hybridization. The numbers below each panel represent

the fraction of fish with the observed phenotype. The inset depicts a magnification of the trunk area. Black arrowheads indicate positive stained cells. (B) Bar graphshowing the quantitative decrease of EGFP decrease measured by ImageJ analysis of acquired images from in situ hybridizations. Numbers on the Y-axis are related to

the positively stained areas in arbitrary units, assessed in 20 fish for each group (***P < 0.0005; t-test). (C) Gba1 loss of function reduces erythroid precursors and mature


at UniversitÃ




ownloaded from


mutant larvae at both rostral regions (opercle, inset of Fig. 1H)and vertebrae centra.

Gba1 loss of function affects multiple cell lineages in fish

We next asked whether other cell populations were involvedin the downstream effects of Gba1 impairment. Given thehematological phenotype of Gaucher patients, we tested themorpholino in Tg(-6.0itga2b:EGFP)la2 fish, in which thrombocyteprecursors and mature thrombocytes are labeled. We observeda strong reduction of transgene expression in morphant fish at5 dpf, supporting a marked thrombocytopenia (Fig. 2A and B).We next asked whether also erythrocytes were affected by Gba1loss of function.We, therefore, evaluated qualitatively and quan-titatively by fluorescentmicroscope, confocal scanning and FACSanalysis the number of erythrocytes in Tg(gata1:dsRed)sd2 micro-injected with Gba1 and mismatch-control morpholinos. Asshown in Figure 2C and D, we detected a significantly reducednumber of erythrocytes in morphants at 3 dpf by fluorescentmicroscope acquisitions and FACS analysis, respectively. A simi-lar reduction at 3 dpf was also measured in the offspring gener-ated by the incross of heterozygous fish (gba1sa1621/+) in the Tg(gata1:dsRed)sd2 background (data not shown).

Since erythrocytes were depleted inmorphants and Gaucherpatients usually exhibit increased macrophage activation, weanalyzed c-myb, which is a marker of definitive hematopoiesisand is important for the initial stage of myelopoiesis (10). Wefound that Gba1 morphants display ectopic expression of c-myb, supporting early hematopoietic defects triggered by Gba1dysfunction (Fig. 2E and F). In addition, we microinjected themorpholino in the gz15Tg transgenic line (named also Lipan),which is a double transgenic line [Tg(ela3l:EGFP)/Tg(fabp10a:DsRed)], labeling the exocrine pancreas in green and the liverin red under epifluorescence. As shown in Figure 2G and H, wefound a progressive and significant hepatomegaly in Gba1 mor-phants after 13 dpf, reminiscent of the visceral phenotype inGD. To address whether the same hepatomegaly was presentin fish mutants, we incrossed heterozygous fish (gba1sa1621/+)in the Lipan transgenic background and found a significant in-crease of liver size already at 10 dpf (not shown). To addresswhether the liver size increase was progressive also in adultstages, we sacrificed 3-month-old homozygousmutants and ob-served a marked increase in both liver and spleen size (Fig. 1I).Therefore, both Gba1 morphants and mutants display a de-crease in the eryrthrocyte population and increase in liversize, thus mimicking the anemia and hepatomegaly seen inGD patients. Moreover, morphant fish exhibit a pronounced de-crease of Tg(-6.0itga2b:EGFP)la2 expression and increase in c-mybtranscripts, thus supporting a disruption of the hematopoieticdifferentiation program, with a consequent thrombocytopeniaand increased myelopoiesis.

Global cellular and molecular defects due toglucocerebrosidase deficiency are triggered duringearly stages of development

Since anemia, thrombocytopenia and a reduced number ofrunx2b-expressing osteoblasts could be due to overinducedapoptosis or decreased cell proliferation, we performed in vivoproliferation and apoptosis assays at 2–6 dpf, in both morphantsand controls. Using cleaved caspase 3 immunohistochemistry(Supplementary Material, Fig. S5A) and in vivo TUNEL labeling(not shown) for apoptosis, EdU-labeling (not shown) and in vivophosphohistone tracking (Supplementary Material, Fig. S5B) forproliferation, we observed no global evident differences in therate of cell proliferation and apoptosis, except for a slight, butnot significant, decrease in the proliferation of osterix-expressingcells. To better understand early molecular defects underlyingthe global cellular alterations, we performed genome-widemicroarray profiling and compared control and morphant geneexpression patterns at 2 dpf. To perform this task, we collectedeight different RNA extracts, each extracted from 150 larvae mi-croinjected with the Gba1 morpholino or control morpholino.In each individual microinjection experiment control and Gba1morpholinowere injected in embryos derived from the same off-spring to rule out potential intra-experiments variability. There-fore, we collected a total of four distinct biological replicates foreach condition. We identified 131 genes that were significantlyupregulated (of which 62 were >1.5-fold) and 63 genes that weredownregulated in Gba1 morphants (Fig. 3A). Noteworthy, boththe most up- and downregulated genes were associated with anoxidative-stress response. We found the genes plod1 (procollagen-lysine, 2-oxoglutarate 5-dioxygenase 1) and pon3 (paraoxonase 3)upregulated >3-folds, while gpx1b (glutathione peroxidase 1b) ex-pression levelswere significantly decreased of∼1.5-fold. Interest-ingly, in-depth microarray analysis allowed us to demonstratethat many differentially expressed genes were involved in intra-cellular vesicle trafficking, mitochondrial activity and transcrip-tional activity [Gene Expression Omnibus, accession no.GSE54754]. To confirm some of the data retrieved by microarrayanalysis, we carried out quantitative real-time-PCR (q-PCRs) ana-lysis on the same pooled RNA extracts used inmicroarray experi-ments from the 2 dpf control andmorphantmicroinjected larvae.As shown in Figure 3B, for plod1 and pon3 we detected the samedegree of differential gene expression, thus confirming the out-put of microarray analysis. To further verify the obtained resultsat a proteomic level, we carried out western blot analysis using across-reacting PLOD1 antibody on 2 dpf pooled morphant andcontrol protein lysates from 150 microinjected larvae. As shownin Figure 2B, we detected an increase of Plod1 protein levels inmorphants when compared with matched control fish lysates.Therefore, for the plod1 target gene, we could conclude thatmicroarray profile matched the results obtained with protein

erythrocytes. Representative Tg(gata1:dsRed)sd2 fish showing reduced transgene expression after Gba1 knockdown. In the small inset, a magnification of the trunk area is

shown. The numbers below each panel represent the fraction of fish with the observed phenotype. (D) Bar graph showing the significant reduction of gata1-positive cells

assessed by FACS analysis in Gba1 morphants fish at 3 dpf. FACS analysis was carried out on cells sorted from 35 dissociated larvae in each experiment. The bar graph

depicts the mean ± SEM of four independent experiments (106 recorded events) (*P < 0.05; t-test). (E) Gba1 impairment increases the number of ectopic cMyb

(+)-hematopoietic precursors. In the small inset, a magnified caudal view of the trunk is depicted. Numbers represent the fraction of fish with the observed

phenotype. (F) Bar graph showing the quantitative increase of c-myb expression measured by ImageJ analysis of acquired images from in situ hybridizations. Numbers

on the Y-axis are related to the positively stained areas in arbitrary units, assessed in 20 fish for each group (**P < 0.005; t-test). (G) Gba1 morphants display enlarged

liver at 13 dpf. In the upper panel, a representative confocal Z-stack projection of the liver from Lipan control and morphant fish at 13 dpf. (H) Quantitative

measurements of the liver area calculated by ImageJ on acquired images of the Lipan transgenic fish microinjected with control and Gba1 morpholino at different life

stages (5 dpf Ncontrol: 14, NGBA1MO: 20; 10 dpf Ncontrol: 12, NGBA1MO: 12; 13 dpf Ncontrol: 18, NGBA1MO: 10). Data are expressed as mean ± SEM of three independent

experiments (*P < 0.05; t-test). (I) Representative liver and spleen from 3-month-old control sibling and gba1sa1621/sa1621fish, showing the enlarged liver and spleen size

in the homozygous mutant (g:gut; s:spleen).


at UniversitÃ




ownloaded from




analysis. Since the upregulation of plod1 and the downregulationof gpx1b could account for an early oxidative stress response,we decided to perform additional in vivo analysis. To addressthis issue, we, therefore, carried out in vivo staining of 24 dpf(Fig. 3C) and 48 hpf (data not shown)morphant and control larvaewith DCFDA (dichlorofluorescein, DCF, diacetate), which is aprobe for reactive oxygen species (ROS). The DCFDA is cleavedintracellularly by non-specific esterase to form DCFH, which isfurther oxidized by ROS to form the fluorescent compound DCF.As shown in Figure 3C, we observed a significant increase offluorescence in morphant fish microinjected with the DCDFAcompound at 24 hpf, when compared with controls. The direct

effect of ROS-mediated increase of fluorescence was specific, asthe co-injection of a human GBA1 mRNA was able to stronglyreduce ROS release.

Canonical Wnt pathway activity is specifically hamperedby glucocerebrosidase impairment

Since early apoptotic events or cell proliferation defects wereruled out, we focused our attention to potential cellular pathwaysinvolved in cell differentiation process. To screen for a candidateaffected cell signaling, we tested the Gba1 morpholino on differ-ent generated signaling pathway reporter lines, in which a

Figure 3. Transcriptomic analysis and in vivo labeling reveal early oxidative stress in Gba1 morphants. (A) Heat map showing the differential expression of transcripts in

Gba1 morphants versus control fish. A partial list of differentially expressed genes (>2-fold increased expression for upregulated genes and <1.5-fold expression for

downregulated genes) is reported. (B) Quantitative polymerase chain reaction (qPCR) analysis for pon3, plod1 and gpx1b, highlighting significantly increased transcripts

levels for pon3 and plod1, while gpx1b mRNA levels decreased in morphants and significantly augmented in rescued larvae. RQ-PCRs were performed on the RNA

extracts from the four biological replicates used in microarray analysis. Each pooled RNA was extracted from 150 microinjected larvae. Data are expressed as

mean ± SEM of five independent experiments (*P < 0.05; t-test). On the bottom right, a representative western blot analysis demonstrates the correspondent increase

between mRNA and protein levels in Gba1MO. Protein lysates were obtained from 150 microinjected larvae as reported in Results. Numbers on the top of each band

represent quantitative measurements by ImageJ analysis. (C) (top) Oxidative stress increase detected by fluorescent DCDFA staining in Gba1 morphants.

Representative whole fluorescent imaging of control, morphant and rescued morphant fish at 24 hpf. (bottom) Quantitative bar graph showing the percentage of fish

with DCFDA-positive staining from three independent experiments.


at UniversitÃ




ownloaded from


reporter gene (EGFP or mCherry) is driven by cell signaling re-sponsive elements (11) (see Fig. 4; Supplementary Material,Fig. S6). By both fluorescent microscopy imaging and in situ hy-bridization analysis, we found a specific drop of reporter gene ex-pression for the responsive line Tg (7xTCFX.lasiam:EGFP)ia4, inwhich domains of active canonical Wnt signaling are labelled(Fig. 4A) (12). To address whether the Wnt signaling defect wasspecific and reproducible in stable Gba1 mutants, we crossed

the transgenic line with the gba1sa1621/+ carriers and analyzedthe offspring from incrossed heterozygote gba1sa1621/+ in theWnt reporter background. As shown in Figure 4B (top), we de-tected in homozygous fishmutants (gba1sa1621/sa1621) a strong de-crease of reporter expression, which consistently remained loweven later than 7 dpf (see bottompanel of Fig. 4B). Since canonicalWnt signaling relies on its major transducer, β-catenin, we per-formed immunoblot analysis of β-catenin in morphants and

Figure 4. Canonical Wnt signaling is disturbed in Gba1 morphants and gba1sa1621/sa1621fish. (A) Representative 24 hpf (top) and 48 hpf (bottom) Wnt reporter fish

microinjected with control morpholino and Gba1 morpholino, showing the marked decrease of reporter activity. (B) Representative 48 hpf (top) and 7 dpf (bottom)

gba1sa1621/sa1621fish in the Wnt reporter background, showing the strong decrease of reporter activity. The small inset shows a magnified view of the cephalic region

of the gba1sa1621/sa1621fish at 7 dpf. (C–E) Representative western blot analysis of total lysates from 100 pooled control and morphant fish at 2 dpf for β-catenin (C), p-

Gsk3β (pSer9) (D) and Gsk3α (E). β-catenin and p-Gsk3β protein level decrease and increase are shown in C and D, respectively. Numbers on the top of each band

represent quantitative measurements by ImageJ analysis. Note that in panel C and E the same β-actin loading control was used. (F) RQ-PCR analysis of GFP showing

the significant decrease of reporter expression in Wnt reporter transgenics after Gba1 morpholino-mediated knockdown. Note that the GFP mRNA levels were

significantly restored when fish were coinjected with the human GBA1 mRNA. RNA extracts were obtained from 100 pooled microinjected control, morphant and

rescued morphant fish in five independent experiments. (G) RQ-PCR analysis of GFP showing the significant decrease of reporter expression in gba1sa1621/sa1621fish in

the Wnt reporter background. RQ-PCRs were carried out on RNAs from pooled 2 dpf homozygous mutants and wild-type siblings (N = 3 for each condition). Data are

expressed as mean ± SEM of four independent experiments (*P < 0.05; t-test). (H) Quantitative polymerase chain reaction (qPCR) analysis for axin1, showing the

significant upregulation of axin1 in Gba1 morphant fish. RNA extracts were obtained from 100 microinjected control, morphant fish in five independent experiments.

Results are expressed as mean ± SEM of five independent experiments (*P < 0.05; ***P < 0.0005; t-test).


at UniversitÃ




ownloaded from




control protein lysates, obtained by protein extraction from 150larvae for each condition. We found a 2-fold decrease of totalβ-catenin protein levels in morphants (Fig. 4C). Total β-cateninlevels are tightly regulated by the action of themultiprotein com-plex, comprised of Glycogen synthase kinase 3 beta (Gsk3β) andthe scaffolding protein Axin1. By immunoblot, we tested thelevels of Gsk3β and found a 3-fold increase of Gsk3β proteinlevels in the previously obtained morphant lysates (Fig. 4D). In-deed, levels of the Glycogen synthase kinase 3 α (Gsk3α), anotherβ-catenin regulator (13) were not affected by Gba1 loss of func-tion, as assessed bywestern blot on the samemorphant and con-trol protein extracts. We also quantitatively measured thedecrease of Wnt reporter activity by RQ-PCRs in transgenic fishmicroinjected with the Gba1 morpholino and control morpholi-no. To perform this task, we collected 100 larvae at 2 dpf foreach condition (control and morphants) for a total amount offive biological independent replicates. We consistently foundabout a 2-fold decrease of total GFP mRNA levels (Fig. 4F and G)in morphant fish when compared with age-matched mis-match-control larvae, thus supporting the direct effect of GBA1loss of function on the transcriptional activity of theWnt report-er. Noteworthy, the Wnt pathway deregulation was significantlyreverted upon human GBA1mRNA overexpression inmorphants(Fig. 4F), thus supporting a direct link betweenWnt signaling de-fects and Gba1 loss of function. To further confirm the canonicalWnt signaling involvement in Gba1 functional impairment, wecollected single genotyped homozygous mutant larvae and per-formed several RQ-PCRs on RNAs from pooled 2 dpf homozygousmutants and wild-type siblings (N = 3 for each condition). Asshown in Figure 4G, we detected a significant decrease ofWnt re-porter transgene expression inmutant fish when comparedwithage-matched control siblings. Notably, the decrease of reporterexpression was similar to that assessed in morphant RNA ex-tracts when compared with mismatch controls. Microinjectionof hGBA mRNAs was partially able to rescue the decrease of re-porter expression in mutant larvae (data not shown). We nexttried to assess the levels of axin1 mRNAs by RQ-PCR analysis onmorphants and mismatch-control RNA extracts from the previ-ously isolated pools. As shown in Figure 4H, wemeasured signifi-cantly higher axin1 mRNA levels in 2 dpf morphants when

compared with mismatch-control larvae. The same degree ofaxin1 upregulation was not, however, detected in gba1 mutantsextracts (data not shown).

Bone defects are reversibly corrected when a humanGBA1 is overexpressed in morphants

Osteopenia has been shown to develop during early life stages inhumans, thus justifying bone improvements when ERT is startedin young patients (14). We, therefore, explored whether overex-pression of a human GBA1mRNAwas also able to reverse the os-teopenic phenotype in fish. As shown in Figure 5A and B, weobserved a significant recovery in the expression of the col10a1marker, which we previously showed to be affected by Gba1loss of function. We next assessed whether bone mineralizationcould also be restored to normal levels after hGBA overexpressionin Gba1 morphants and we found that both cephalic bones andvertebrae centra mineralization were consistently recovered(Fig. 5C and D).

CanonicalWnt signaling activity is significantly impairedin type 1 Gaucher disease patients

Given an important role of the canonical Wnt signaling in boneremodeling (15), we next addressed whether Wnt signaling de-fects could also be detected in type 1 GD patients. To investigatethis possibility, we performed different assays. We first carriedout immunofluorescence analysis for β-catenin and glucocereb-rosidase in fibroblasts from a healthy volunteer and a type 1 GDpatient. The type 1 GD patient was a compound heterozygousfor [R170P] + [c.1225(−11delC) − (14T>A)] (16). As shown in Fig-ure 6A, we detected a strong reduction in β-catenin protein levelsand almost no immunoreactivity for glucocerebrosidase in thepatient. Western blot analysis on total protein lysates from thesame patient’s fibroblasts confirmed a quantitative decrease oftotal β-catenin (>1.5-fold) and glucocerebrosidase (nearly 3-fold)levels (Fig. 6B). We next evaluated by RQ-PCRs the mRNA levelsof the intracellular, Axin1 and the extracellular, DICKKOPF-1(DKK1), negative regulators of the Wnt pathway in RNA extractsfrom cultured fibroblasts of (i) the same type 1 GD patient,

Figure 5. Col10a1 decreased expression and bonemineralization are rescued upon human GBA1 overexpression (A) Representative in situ hybridization for col10a1 in 4 dpf

control, Gba1MO and rescued fish. Number below each panel represents the fraction of fishwith the observed phenotype. (B) Graphical diagram showing an ImageJ based

volume analysis of the col10a1-stained opercles in the different conditions. (C) Lateral (top) and ventral (bottom) views of 10 dpf Alizarin-stained larvae, showing the

recovery of bone mineralization after human GBA1 forced expression. In the bottom panel mineralization of the ceratohyal cartilage is highlighted by arrowheads. (D)

Graphic representation of the quantitative measurements of vertebrae centra in different conditions. Results are expressed as mean ± SEM of five independent

experiments (*P < 0.05, **P < 0.005, ***P < 0.000; t-test).


at UniversitÃ




ownloaded from


compound heterozygous for [R170P] + [c.1225(−11delC)− (14T>A)](16), (ii) an N370S homozygous type 1 patient and (iii) a L444Phomozygous type 3 GD patient. As shown in Figure 6C and D,we detected a significant increase of both Axin1 and DKK-1mRNAs in the compound heterozygote for the [R170P] + [c.1225(−11delC) − (14T>A)] alleles, while DKK-1 mRNA levels were al-most undetectable in the N370S homozygote. Notably, we alsodetected significantly reduced levels for the human GPX1mRNA in the N370S homozygote. Since the identification of

bone turnover markers in relation to GD manifestations hasbeenmatter of debate (6,17,18), we investigated whetherWnt-re-lated modulators could be used as traceable blood serum mar-kers, in a cohort of 30 patients affected by type 1 and 3 GD witha known genotype (Table 1). As shown in Figure 6F, we observedthat only type 1 GD patients homozygous for [N370S] + [N370S]and compound heterozygous for [N370S] + [L444P] displayed sig-nificantly decreased osteocalcin levels. When measuring bloodserum levels for DKK-1 and sclerostin (SOST-1), another known

Figure 6. CanonicalWnt signaling is impaired in type 1 GD patients. (A) Representative immunostaining for β-catenin/VPS28 (top) andGBA1 (bottom) in control fibroblasts

and in fibroblasts from a type 1 Gaucher patient [R170P] + [c.1225(−11delC) − (14T>A)] (abbreviated to GDp1). Red arrowheads indicate spikes and the nuclear β-catenin

staining in control cells. Cells from the Gaucher patient display a marked decrease of β-catenin and GBA staining. (B) Western blot analysis of total cell lysates with

GBA1 and β-catenin antibodies, showing decreased levels of β-catenin in GDp1 patient fibroblasts. Numbers on the top of each band represent quantitative

measurements by ImageJ analysis. (C and D) Increased levels of Axin1 and DKK1 mRNAs in GDp1 fibroblasts, as detected by qPCR analysis. Results are expressed as

mean ± SEM of three independent experiments, with *P < 0.05, **P < 0.005; t-test). (E) Differential expression of GPX-1 in type 1 GD fibroblasts from patients with

different genotypes. Fibroblasts from a homozygous [N370S] + [N370S] patient display a significant reduction of GPX-1 mRNA levels, while those from the patient with

the rare genotype [R170P] + [c.1225(−11delC)−(14T>A)] were associated with significantly increased GPX-1 mRNA levels. (F) Blood serum levels of OSTEOCALCIN,

DICKKOPF-1 (DKK-1) and SCLEROSTIN (SOST-1) measured by ELISA show significant decrease of OSTEOCALCIN in homozygous [N370S] + [N370S] and compound

heterozygous [N370S] + [L444P] (left bar graph), significantly decreased DKK-1 levels in compound heterozygous [N370S] + [L444P] (middle bar graph) and significantly

decreased SOST-1 levels in homozygous [N370S] + [N370S] patients (right bar graph). (G) (left) Correlation of DICKKOPF-1 (DKK-1) and OSTEOCALCIN serum levels.

X-axis indicates DKK-1 serum level; Y-axis indicates OSTEOCALCIN level in Gaucher patients. r = 0.46 (P < 0.001). (right) Correlation of SCLEROSTIN-1 (SOST-1) and

OSTEOCALCIN serum levels. X-axis indicates SCLEROSTIN serum levels; Y-axis indicates OSTEOCALCIN level in Gaucher patients. r = 0.56 (P < 0.001).


at UniversitÃ




ownloaded from


extracellular Wnt pathway modulator, we found that DKK-1serum levels were significantly reduced in compound heterozy-gotes [N370S] + [L444P], while SOST-1 levels were significantly di-minished in homozygotes [N370S] + [N370S]. We next assessed apotential correlation between osteocalcin and the Wnt modula-tors DKK-1 and SOST-1. We found significant evidence that oste-calcin levels positively correlate with both DKK-1 and SOST-1levels in type 1 GD patients, regardless of a specific genotype(Fig. 6G).

DiscussionUnbalanced bone remodeling due to reduced bone formation andincreased bone resorption represents themost prevailing currentskeletal aspect of non-neuronopathic forms of Gaucher disease.As towhether the former or the latter is more determinant in theclinical outcome has been longly debated (6,17,18).While a previ-ously supported macrophage-based theory claims the centralrole of Gaucher cells in promoting an inflammatory response atthe expense of bone mineralization (4,18), recent evidencesdepict a novel scenario, in which the role of inflammatoryagents-producing cells becomes less crucial in the onset ofbone pathogenesis. In support to this hypothesis, a recentlygenerated conditionalmousemodel has demonstrated that a pri-mary osteoblast differentiation and proliferation defect underliesthe decreased bone formation and leads to a progressive osteope-nic phenotype (5). In the conditional knockoutmodel, a defectiveprotein kinase C was proposed to be responsible for decreasedosteoblast proliferation, whereas increased osteoclastic activitywas ruled out. Though this challenging paradigm has opened anovel attractive overview of the disease, several key questions re-mained so far unanswered. One of these questions has beenwhether key signaling pathways may be dysfunctional in type 1GD patients. Here, we describe a novel complementary animalmodel, by which this issue has been partly addressed. We exam-ined two different fish models. In the first Gba1 loss of functionwas produced by in vivo morpholino-based gene knockdown;while in the second a stable inheritable mutation was generatedin the Gba1 gene by a forward genetic approach. Notably, in bothmodels a splicing defect in the gba1 gene was produced, at theexon 2 level for morpholino-injected fish and in exon 4 for thegba1sa1621/sa1621 mutant fish, leading in both cases to a truncatedGba1 protein. Although we were unable to measure the loss oflysosomal acid β-glucocerebrosidase activity in both models byan enzymatic assay, we demonstrated that some key pheno-types, involving type 1 GD-related hematological and skeletal as-pects were recapitulated. Fish morphants displayed reducedGATA-1:DsRed-expressing erythrocytes reminiscent of the GD-re-lated anemia and 6.0itga2b:EGFP-expressing thrombocytes re-sembling thrombocytopenia detected in patients. However, wealso preliminarily found an increased ectopic c-myb expressionin morphants. This observation provides support to a previouslyraised hypothesis suggesting that defects of intrinsic hematopoi-etic stem/progenitor cells may be linked to GBA1 deficiency (19).Moreover, an increase in the number of c-myb-expressing hem-atopoietic precursors may suggest a pro-myeloid differentiationprogram, associated with reduced 6.0itga2b:EGFP-expressingthrombocytes (20) and GATA-1-expressing erythrocytes (21). Inagreement with an early developmental expression and pattern-ing of gba1 gene during fish bone formation, we next showed thatloss of its function impaired osteoblast differentiation through areduced expression of key bone-related markers, such as runx2band col10a1. The zebrafish Runx2b protein is an ortholog ofRUNX2, a master transcription factor involved in osteoblast andchrondrocyte differentiation (22,23). RUNX2 has been shown tobe regulated by FGF and Wnt signaling in the osteoblast lineage(24) and positively regulates Col10a1 expression in mice (25). Wedemonstrated that co-injection of a human GBA1 mRNA withthe antisense morpholino oligo in early developmental stageswas able to recover bone alterations, by increasing col10a1 ex-pression in fish, thus supporting the observation that majorbone improvements are detectable in younger children undergo-ing ERT treatment (14). We also provided evidences that twomajor cellular pathways were strictly affected by Gba1 functional

Table 1. Characteristics of affected and non-affected individualsrecruited for blood serum analysis

Subjects Gender Status (type ofGaucher disease)

Genotype§

1 F 1 [L444P] + [W312S]2 M 3 [L444P] + [L444P]3 M 1 [N370S] + [L444P]4 F 1 [N370S] + [L444P]5 M 1 [N370S] + [N370S]6 M 1 Na7 F 1 Na8 F 1 [N370S] + [RecNciI]9 F 3 [L444P] + [L444P]10 F 1 [N370S] + [N370S]11 F 1 [N370S] + [N370S]12 F 1 [N370S] + [E388K]13 F 1 [N370S] + [E388K]14 M 3 [T231R] + ?]15 F 1 [L444P;E326K] + [L444P;

E326K]16 M 3 [N370S] + [N370S]17 M 1 [N370S] + [RecNciI]18 M 1 [N370S] + [L444P]19 M 1 [N370] + [?]20 F 1 [L444P] + [L444P]21 F 3 [N370S] + [L444P]22 F 1 [N370S] + [V214X]23 F 1 [D409H] + [?]24 F 3 [N370S] + [L444P]25 M 1 [N370S] + [total gene del]26 F 1 [N370S] + [total gene del]27 F 1 [N370S] + [IVS2G>A]28 F 1 N370S] + [IVS2G>A]29 M 1 [N370S] + [L444P]30 M 1 [N370S] + [RecNciI]

Control31 M Non-affected32 F Non-affected33 F Non-affected34 F Non-affected35 F Non-affected36 M Non-affected37 M Non-affected38 M Non-affected39 M Non-affected40 M Non-affected

na, not available, genotyped elsewhere; ?, still unknown allele; §Reference cDNA

sequence: GenBank accession no. M16328.1. Mutations at the protein level are

described following the traditional nomenclature within the Gaucher field,

which considers amino acid 1 the first amino acid after the signal peptide.

According to current mutation nomenclature guidelines (http://www.hgvs.org/

mutnomen), ascribing the A of the first ATG translational initiation codon as

nucleotide +1, 39 amino acids should be added.


at UniversitÃ




ownloaded from

http://www.hgvs.org/mutnomen







impairment before the onset of bone abnormalities. Molecularanalysis and in vivo labeling demonstrated the occurrence of anearly increased oxidative stress, in line with recent observations(26). A tight link between oxidative stress and reduced osteoblas-togenesis has been largely described elsewhere (27,28). Note-worthy, Gba1 forced expression in morphants rescued theincrease of ROS, thus enforcing the hypothesis of a tight link be-tween correct Gba1 function and antioxidant response.

A second major affected cellular pathway involved in Gba1impaired activity was found to be the canonical Wnt pathway.We showed that a significant decrease in theWnt pathway activ-ity is detectable in both zebrafish Gba1 loss-of-function models.Moreover, by showing reduced accumulation of theWnt pathwaytransducer, β-catenin and increased levels of the negative regula-tors Gsk3β, Axin1 and Dkk1, we provided a potential mechanisticexplanation of the diminished Wnt pathway activity. Since ourfirst observation of Wnt pathway deregulation was already at24 hpf, we hypothesized that runx2b and col10a1 downregulationcould have been determined by reduced TCF/LEF transcriptionalactivity. To support this hypothetical mechanism, the forced ex-pression of human GBA1mRNA inmorphants was able to rescueWnt reporter activity deregulation already at 2 dpf and recoverthe expression of the downstream col10a1 target at 4 dpf. A ca-nonical Wnt pathway alteration due to reduced β-catenin levelsand increased Axin1 and DKK1 transcripts was similarly identi-fied in a Type 1 GD patient (16). Notably, in a recent study ofbone marrow mesenchymal and hematopoietic stem cells ofType 1 GD patients significantly elevated DKK-1 levels were de-tected (19).

Nonetheless, we unexpectedly observed significantly reducedserum DKK-1 and SOST-1 levels in type 1 GD patients. This latterobservation seems in apparent contrastwith the notion that bothproteins are Wnt pathway negative regulators. However, we alsomeasured significantly decreased levels of osteocalcin, but incontrast to a previous work (18), we analyzed a bigger cohort ofpatients, selected according to the genotype. Our findings indi-cated that only [N370S] + [N370S] homozygous or [N370S] +[L444P] compound heterozygous patients showed a significantdecrease of osteocalcin levels, thus highlighting a possible strictcorrelation between the type 1 GD N370S-associated genotypeand the bone phenotype. Lower DKK-1 and SOST-1 blood levelscould be explained by the fact that they represent an indirectmeasurement of osteoblast and osteocyte population, thus con-firming a depauperation of bone-forming cells in affected pa-tients in agreement with previous studies (29,30).

Moreover, since upregulation of DKK-1 is a prerequisite forlate stage differentiation of osteoblasts (31), it is possible that re-duced DKK-1 levels in patients sera may explain the progressivebone loss seen in type 1 GD.

In conclusion, we provide a potential novel mechanistic linkbetween bone defects in type 1 Gaucher disease and GBA1 lossof function. According to our observations, impairment in theglucocerebrosidase activity triggers an early deregulation of thecanonical Wnt pathway activity which has been demonstratedto play a pivotal role in bone homeostasis (7,15). It is thus tempt-ing to speculate that manipulation of this pathway may comple-ment the already established ERT therapy, the newly developedsubstrate reduction therapy, and the promising described ap-proach with molecular chaperones (32,33).

Material and MethodsAll procedures involving fish husbandry and manipulation wereevaluated and accepted by the Local Ethical Committee at the

University of Padova. Fibroblasts from type 1 Gaucher patients,supplied by the Biobank (G. Gaslini) have been obtained for ana-lysis and storage with the patients’ (and/or a family member’s)written informed consent. The consent was sought using aform approved by the local Ethics Committee. The mutantsa1621 fish line was obtained by the Zebrafish InternationalResource Center at Eugene (OR, USA). The following fish lineswhere used for the analysis of gba1 knockdown: Tg(dusp6:d2GFP)pt6(34), Tg(7xTCFXla.siam;EGFP)ia4(12); Tg(fli1a:EGFP)y1(35),Tg(sox10:mRFP)vu234 (36), Tg(kdrl:EGFP)s843 (37), Tg(Col2a1aBAC:mCherry)hu5900 (38); Tg(Ola.Sp7:NLS-GFP)zf132(39), TgBAC(col10a1:Ci-trine)hu7050(40); Tg(12xGli-HSV.Ul23:GFP)ia11(11); Tg(EPV.TP1-Mmu.Hbb:EGFP)ia12(11); (Tg(12xSBE:EGFP)ia16(11); Tg(BMPRE:EGFP)ia18

(11), Lipan (41); Tg(gata1:dsRed)sd2 (42), Tg(-6.0itga2b:EGFP)la2 (43).

Morpholino-mediated knockdown

For knockdown experiments, we designed one translation-block-ing morpholino (5′-ATAAAAAGAGCCGTTTCTCTCATCC) and asplicing-blocking morpholino (5′-TAAGAGCACTCACCTGCACCTGTGC). For control experiments, we used both a five-base mis-match-control morpholino (TAAcAcCACTgACCTcCACCTcTGC)and a five-base unrelated morpholino (GTtAATACcAGgATAgATTgATTG). All morpholinos were purchased from Genetools (Philo-math, OR, USA). Morpholino stocks were resuspended inDNAase/RNAse freewater. For microinjection experiments, mor-pholinos were resuspended in Danieau buffer (8 m NaCl,0.7 m KCl, 0.4 m MgSO4, 0.6 m Ca(NO3), 2, 5 m HEPES, pH7.6) and Red Phenol (Sigma, Milan, Italy) at the working concen-tration. Microinjections were performed in collected one-cellstage embryos under a light microscope. Pigmentation was pre-vented using a 0.003% 1-phenyl-2-thiourea solution in the first24 h.

Genotyping of mutant

Tail clip genotyping was performed on tail clips after 10 mg/mlproteinase K overnight digestion at 55°C. Genomic DNA waspurified by phenol-chloroform extraction and ethanol preci-pitation. Purified DNA was dissolved in DNAse-free water.PCR was performed using the set of oligos, GBAforward(for)(5′-GGACCAGCTGCTCAGGACAGT) and GBAreverse(rev) (5′CTGACCCGAAAGGTAGCAAA) at the following conditions: predena-turation 5 min at 94°C, and 35 cycles of 94°C for 1 min, 56°C for30″ and 72°C for 1 min. Amplicons were then treated with EXO-SAP (Roche Diagnostics, Monza, Italy) and sequenced on bothstrands.

Alizarin and Alcian stainings

Skeletal staining was performed as previously described (44).In addition Alizarin staining was performed using a modifiedprocedure. Briefly, larvae were fixed 1 h at room temperature in4% buffered paraformaldehyde, washed in 50% ethanol and de-hydrated overnight in 95% ethanol. Staining was carried out inAlizarin solution (0.25%, w/v, in 2% KOH) for 3 h and larvaewere then briefly cleared in 2% KOH (w/v) and stored in KOH/glycerol (20 : 80).

Transmission electron microscopy

Fish were fixed in 3% glutaraldehyde in 0.1 cacodylate sodiumbuffer and processed as described previously (45). Ultrathin sec-tions were viewed on a Zeiss 902 electron microscope.


at UniversitÃ




ownloaded from


Whole-mount in situ hybridization andimmunohistochemistry

Embryos were fixed in 4%-buffered p-formaldehyde (PFA) in PBS.RNA in situ hybridizations were performed essentially as previ-ously described (46). The following probes were used: col10a1,cyp26b1 and osteopontin (6). The GBA1 antisense probe wasderived from an 821 bp cDNA cloned in pCRII-TOPO plasmid(Lifetechnologies, Milan, Italy) using the following set of oligos,GBAfor (5′-TGTCTCTGTCTTCCGGAGCT) and GBArev (5′-ATGTCATGGGCGTAGTCCTC). The plasmid was linearized with HindIIIand the probe was transcribed by T7 polymerase using DIG-dUTPs. Sense control probewas preparedwith the same plasmid,linearized with XbaI and transcribed with Sp6. The runx2b probewas generated by amplifying a 523 bp fragment and cloning it inpCRII-TOPO. Antisense- and sense-labeled riboprobes wereobtained by T7 and SP6 transcription, respectively. GFP andmCherry antisense riboprobes were generated by KpnI digestionof a Tol2 middle entry vector containing the GFP and mCherrycassette, respectively, and followed by T7 transcription.

Quantitative real-time PCR

RNA extractionZebrafish morphants embryos and mutant sa1621 at 48 hpf werehomogenized in Trizol reagent (Lifetechnologies) and total RNAwas isolated using the standard trizol-chloroform-ethanol ex-traction procedure. RNAs were resuspended in 20 µl of waterRNAse free. RNA samples were checked for integrity by capillaryelectrophoresis (RNA 6000Nano LabChip, Agilent Technologies,Santa Clara, CA, USA). A total of 2 µg of RNA was reverse tran-scribed into cDNA using a SuperScript III Reverse Transcriptase(Lifetechnologies), according to standard procedures. The cDNAwas subsequently subjected to SYBR Green-based real-time PCRusing a RotorGene 3000 (Corbett, Concorde, NSW). Primers arelisted in Table 2.

RT-PCR data analysisRT-PCR data were analyzed using a manually set threshold andthe baseline was set automatically to obtain the threshold cycle(Ct) value for each target. GAPDH was used as an endogenoushousekeeping control gene for normalization. Relative gene ex-pression among samples was determined using the comparativeCt method (2 – ΔΔCt). Results are expressed as the mean ± SEM inrelative expression.

Immunofluorescence on type 1 Gaucher fibroblasts

Patient history has been described elsewhere (16). Fibroblasts weremaintained in Dulbecco’smodified Eagle’smedium supplementedwith 4.5 g/l glucose, 10% fetal calf serum, 50 mg/ml gentamicin and4m glutamine. Thirty thousandsfibroblasts perwell were seededon polylysine-coated coverslips in a 24-well plate. After 24 h cellswere fixed with 4% buffered paraformaldehyde. Blocking wasachieved with a 10% (v/v) sheep serum, 1% BSA (w/v), 0.1% TritonX-100 and 0.3 M glycine. After few washes in PBS-Tween (0.1%,v/v), cells were incubated overnight with primary antibody at 4°Cand after three washes in PBS-Tween (0.1%, v/v), a 2 h secondaryantibody incubation was performed. Coverslips were finallymountedwithDAPI (Lifetechnologies) on glass slides and observedunder C2 confocal microscope (Nikon, Milan, Italy). The followingantibodieswereused:antiGSK3β (Sigma,1 : 100) rabit anti β-catenin(Abcam,Milan, Italy, 1 : 100), goat anti Vps28 (Abcam, 1 : 100), rabbitanti GBA (Novus Biologicals, Milan, Italy, 1 : 100).

Apoptosis and proliferation assays

For apoptosis analysis with a Cleaved caspase 3 antibody (Cellsignaling, Milan, Italy, 1 : 100) fish at 2–6 dpf were fixed in 4%PFA in PBS, permeabilized in methanol washes and acetone for20 min and finally incubated in primary antibody in PBS/DMSO1% overnight at 4°C. After three washes for 30 min incubation

Table 2. List of q-PCR oligonucleotides (oligos) used with the Accession number of the related target gene

Oligo name Sequence 5′→3′ Genbank Accession number

GFP for ACGTAAACGGCCACAAGTTC AEVGFPBGFP rev AAGTCGTGCTGCTTCATGTG AEVGFPBzGAPH for GTGGAGTCTACTGGTGTCTTC ENSDARG00000043457zGAPDH rev GTGCAGGAGGCATTGCTTACA ENSDARG00000043457zPlod1 for CTGATGGGTTCAGACGGTTT ENSDARG00000059746zPlod1 rev TTGGCCTGCTGAAACTTCTT ENSDARG00000059746zPON1 for AAAGGCTCGGCACACTTAGA ENSDARG00000032496zPON1 rev TTCAAGCCAGTGCTCAGAAA ENSDARG00000032496zGpx1b for CTGCGAGACAAGGAGGAAAC ENSDARG00000006207zGpx1b rev TGCAGCTCGTTCATCTGAGT ENSDARG00000006207zAxin1 for GAGAGACAGCCATGGAGAGG ENSDARG00000026534zAxin1 rev TGCTCATAGTGTCCCTGCAC ENSDARG00000026534zSmad1 for CTGTGAAGGATCACGTCGAG ENSDART00000033566zSmad1 rev GCCCAGTCAACACAGTCTCA ENSDART00000033566hDkk1 for 5′-CAGGCGTGCAAATCTGTCT-3′ ENSG00000107984hDkk1 rev 5′-CCCATCCAAGGTGCTATGAT ENSG00000107984hAxin1 for 5′-ACAGGATCCGTAAGCAGCAC-3′ ENSG00000103126hAxin1 rev 5′-GCTCCTCCAGCTTCTCCTC-3′ ENSG00000103126hSmad1 for 5′-GCTTACCTGCCTCCTGAAGA-3′ ENSG00000170365hSmad1 rev 5′-ACCATCCACCAACACACTTG-3′ ENSG00000170365hGPX-1 for 5-CGGGACTACACCCAGATGAA ENSG00000233276hGPX-1 rev 5′-CCGGACGTACTTGAGGGAAT ENSG00000233276hGAPDH for 5′-CACAATATCACTTTACCAAGAGTTAAAAGC ENSG00000111640hGAPDH rev 5′-CGAGCCACATCGCTCAGAC ENSG00000111640

rev, reverse; for, forward.


at UniversitÃ




ownloaded from


with an alkaline phosphatase-conjugated secondary antibodywas performed (Sigma, 1 : 500). Staining was carried out with anNBT/BCIP solution (Roche Diagnostics), according to manufac-turer’s instructions. Alternatively, fish at the same developmen-tal stages were fixed in 4% PFA overnight, dehydrated in severalmethanol washes and kept in 100% methanol for 30 min. Afterrehydratation, fish were digested with collagenase (Sigma,1 mg/ml), and incubated in TdT reaction (Apotag, Chemicon,DBA, Milan, Italy) for 2 h at 37°C. Fluorescein-conjugated anti-digoxigenin antibody incubation was performed at 4°C overnightand images were taken with C2 Nikon confocal microscopy.

For proliferation assay, the Click-it Edu proliferation assay kit(Lifetechnologies) was considered. Briefly, fish larvae at 2–4 dpfwere incubate for 30 min in 10 m EdU on ice. After severalwashes with cold embryo medium, a 30 min to 4 h incubationat 28.5°C for EdU incorporation was carried out. After fixation in4% PFA larvae were treated with 10 μ Proteinase K (Sigma) andtreated with the reaction cocktail as suggested by the manufac-turer’s instructions. Images were taken on fixed larvae with aC2 Nikon confocal system (Nikon).

Alternatively, fish from the Tg(Ola.Sp7:NLS-GFP)zf132 linewereoutcrossed with a Tg(H2b:RFP) line and heat shocked at 37°C for1 h at 1 dpf. Images from the opercle were taken at 3–4 dpf withthe C2 confocal system and processed by ImageJ analysis. In par-ticular, the number of RFP-labeled opercle cells was assessed byadjusting the thresholded acquired image of each sample abovethe background of faintly stained cells. Manders coefficient wasused to measure the double RFP/GFP co-labeled opercle cells incontrol and morphant larvae.

Western blot

Total proteinswere extracted from zebrafishmorphants and con-trols at 48 hpf with Tissue extraction lysis buffer II (Invitrogen).Samples of denaturated proteins (10 μg) were separated on Nu-PAGE® Novex® 4–12% Bis-Tris Gels (Invitrogen) and transferredto PVDF membranes. The membranes were subsequently incu-bated with antibodies: GSK3β (1 : 50 000) (Sigma), β-catenin (1 :1000) (Sigma) and β-actin (1 : 5000) (Santa Cruz, Dallas, TX, USA)at 4°C overnight after incubation in Western blocker solution(Sigma) for 2 h. Meanwhile, 10 µg of type 1 Gaucher fibroblastand control fibroblast protein lysates were examined using west-ern blotting according to procedures described above. The mem-branes were incubated with: GBA (1 : 1000; Novus Biologicals);β-catenin (1 : 1000; Sigma) and β-actin (1 : 5000; Santa Cruz) anti-bodies. After the incubation with horseradish peroxidase conju-gated secondary antibody (1 : 2000; Sigma) for 1 h at roomtemperature, visualization was performed with SuperSignalWest Pico Chemiluminescent Substrate detection kit (ThermoScientific, Milan, Italy) followed by exposure to X-ray film(Thermo Scientific). β-Actin served as an endogenous control.

Microarray

Four pools of GBA1 MO and control MO-injected fish wereindependently collected after 48 h and total RNA was isolatedaccording to standard procedures. RNA quality was assessed byNanochip Agilent Bioanalyzer and microarray analysis wasperformed using the 44 × 4 Agilent platform, according to themanufacture’s protocol. Raw datawere intra- and interarray nor-malized. Normalized data were then tested with the SAM soft-ware (Stanford University, USA) and processed by the Davidplatform (http://david.abcc.ncifcrf.gov). t-Test statistical analysiswas carried out with Welch correction of α = 0.01.

Rescue with human GBA1 overexpression andCerezyme treatment

The human glucocerebrosidase cDNA cloned into the commer-cial vector pCMV6-XL5 (Origene, USA) was used as template forT7-mediated RNA transcription. Isolated RNA was DNAase trea-ted for 2 h and recovered by MEGAclear Spin column purification(Euroclone). Qualitative and quantitative assessment was carriedout by agarose gel and Nanodrop Instrument, respectively. Up to2.5 ng of hGBA1 mRNA per embryo were microinjected with17.5 pg gba1 morpholino without any evident toxic side effect.Alternatively, coinjection with the GBA MO was carried outwith 1 × 10−5 U of Cerezyme per embryo in Danieau buffer(58 m NaCl, 0.7 m KCl, 0.4 m MgSO4, 0.6 m Ca(NO3)2,5.0 m HEPES, pH 7.6.)

In vivo reactive oxygen species detection(DCDFA staining)

For ROS analysis, fish were either microinjected or incubatedwith a 10 m DCF diacetate (DCDFA, Lifetechnologies) at one-cell stage embryo and visualized by a fluorescent microscope at520 nm at 24–48 hpf.

Tartrate-resistant acid phosphatase assay

Fish at 10–14 dpf were processed using an adapted colorimetricTRAP assay (Sigma), according to manufacturer’s instruction.Bleaching for larval clearing was achieved by 10% peroxidehydrogen in water for 4 h.

Zebrafish dissociation and flow cytometry analysis

The protocol for dissociation of zebrafish cells was previously de-scribed (47); however, several modifications were introduced.Protease solution was replaced with 1× PBS, 0.25% trypsin phenolred free (Gibco, Life Technologies, Milan, Italy), 1 m EDTA, pH8.0, 2.2 mg/ml Collagenase P (Sigma) and resuspension mediumwas replaced with Opti-MEM (Gibco), 1% filtered heat inactivatedFBS (Gibco), 1× Penicillin–Streptomycin solution (Sigma). Disso-ciated cells were filtered by a 70 µm nylon membrane and sub-jected to FACS (BD FACSCanto II system). FSC, SSC andfluorescence (564–606 nm) have been analyzed. Unfluorescentdissociated zebrafish cells at 3 dpf were used as negative control,P1 gate contained all autofluorescent and unfluorescent cells. 106

events were recorded. P2 gate was setup outside the P1 gate, col-lecting highly fluorescent events (564–606 nm). FITCfluorescence(515–545 nm)was also used as internal negative control. Sampleswere collected by 35 dissociated fish larvae. Fluorescent control,Gba1 morphants and unfluorescent samples were generated bymicroinjecting the offspring from the same mating process, be-tween a Tg(gata1:dsRed)sd2(+/−) fish and a wild-type larva in eachreplica.

Supplementary MaterialSupplementary Material is available at HMG online.

AcknowledgementsWe thank Luigi Pivotti and Martina Milanetto for excellent tech-nical assistance in fish husbandry and maintenance, WalterGiuriati for the TEM analysis, Giuliana Cangemi, Cinzia Gattiand Massimo Aureli for providing substantial support in the en-zymatic assays. We also thank Stephan Shulte-Merker (Utrecht,


at UniversitÃ




ownloaded from

http://david.abcc.ncifcrf.gov








Nederland) and Michel Bagnat (NC, USA) for providing with theTg(Ola.Sp7:NLS-GFP)zf132 line and the Tg(hsp:Lamp1-RFP) con-struct, respectively. Patient samples were obtained from the‘Cell Line and DNA Biobank from Patients Affected by GeneticDiseases’ (G. Gaslini Institute) – Telethon Network of GeneticBiobanks (Project No. GTB07001).

Conflict of Interest statement. None declared.

FundingThis work has been financially supported by Genzyme Gener-ation Program 2010 to M.F. and the Italian Ministry of Health(Ricerca Finalizzata GR-2008-1139743) to E.M.

References1. Grabowski, G.A. (2012) Gaucher disease and other storage dis-

orders. Hematol. Am. Soc. Hematol. Educ. Program, 2012, 13–18.2. Hruska, K.S., LaMarca,M.E., Scott, C.R. and Sidransky, E. (2008)

Gaucher disease: mutation and polymorphism spectrum inthe glucocerebrosidase gene (GBA). Hum. Mutat., 29, 567–583.

3. Mikosch, P. (2011) Miscellaneous non-inflammatory muscu-loskeletal conditions. Gaucher disease and bone. Best Pract.Res. Clin. Rheumatol., 25, 665–681.

4. Reed, M., Baker, R.J., Mehta, A.B. and Hughes, D.A. (2013) En-hanced differentiation of osteoclasts from mononuclear pre-cursors in patients with Gaucher disease. Blood Cells Mol. Dis.,51, 185–194.

5. Mistry, P.K., Liu, J., Yang, M., Nottoli, T., McGrath, J., Jain, D.,Zhang, K., Keutzer, J., Chuang, W.L., Mehal, W.Z. et al. (2010)Glucocerebrosidase gene-deficient mouse recapitulates Gau-cher disease displaying cellular and molecular dysregulationbeyond themacrophage. Proc. Natl. Acad. Sci. USA, 107, 19473–19478.

6. van Dussen, L., Lips, P., Everts, V.E., Bravenboer, N., Jansen, I.D., Groener, J.E., Maas, M., Blokland, J.A., Aerts, J.M. and Hol-lak, C.E. (2011) Markers of bone turnover in Gaucher disease:modeling the evolution of bone disease. J. Clin. Endocrinol.Metab., 96, 2194–2205.

7. Baron, R. and Kneissel, M. (2013) WNT signaling in bonehomeostasis and disease: from human mutations to treat-ments. Nat. Med., 19, 179–192.

8. MacDonald, B.T., Tamai, K. and He, X. (2009) Wnt/beta-cate-nin signaling: components, mechanisms, and diseases. Dev.Cell, 17, 9–26.

9. Estrada, K., Styrkarsdottir, U., Evangelou, E., Hsu, Y.H., Dun-can, E.L., Ntzani, E.E., Oei, L., Albagha, O.M., Amin, N.,Kemp, J.P. et al. (2012) Genome-wide meta-analysis identifies56 bone mineral density loci and reveals 14 loci associatedwith risk of fracture. Nat. Genet., 44, 491–501.

10. Valledor, A.F., Borras, F.E., Cullell-Young, M. and Celada, A.(1998) Transcription factors that regulate monocyte/macro-phage differentiation. J. Leukoc. Biol., 63, 405–417.

11. Moro, E., Vettori, A., Porazzi, P., Schiavone, M., Rampazzo, E.,Casari, A., Ek, O., Facchinello, N., Astone, M., Zancan, I. et al.(2013) Generation and application of signaling pathway re-porter lines in zebrafish. Mol. Genet. Genomics, 288, 231–242.

12. Moro, E., Ozhan-Kizil, G., Mongera, A., Beis, D., Wierzbicki, C.,Young, R.M., Bournele, D., Domenichini, A., Valdivia, L.E.,Lum, L. et al. (2012) In vivo Wnt signaling tracing through atransgenic biosensor fish reveals novel activity domains.Dev. Biol., 366, 327–340.

13. Doble, B.W., Patel, S., Wood, G.A., Kockeritz, L.K. and Wood-gett, J.R. (2007) Functional redundancy of GSK-3alpha andGSK-3beta in Wnt/beta-catenin signaling shown by usingan allelic series of embryonic stem cell lines. Dev. Cell, 12,957–971.

14. Mistry, P.K.,Weinreb, N.J., Kaplan, P., Cole, J.A., Gwosdow, A.R.and Hangartner, T. (2011) Osteopenia in Gaucher disease de-velops early in life: response to imiglucerase enzyme therapyin children, adolescents and adults. Blood Cells Mol. Dis., 46,66–72.

15. Regard, J.B., Zhong, Z., Williams, B.O. and Yang, Y. (2012) Wntsignaling in bone development and disease: making strongerbone with Wnts. Cold Spring Harb. Perspect. Biol., 4, 4–17.

16. Romano,M., Danek, G.M., Baralle, F.E., Mazzotti, R. and Filoca-mo, M. (2000) Functional characterization of the novel muta-tion IVS 8 (-11delC) (-14T>A) in the intron 8 of theglucocerebrosidase gene of two Italian siblings with Gaucherdisease type 1. Blood Cells Mol. Dis., 26, 171–176.

17. Drugan, C., Jebeleanu, G., Grigorescu-Sido, P., Caillaud, C.and Craciun, A.M. (2002) Biochemical markers of bone turn-over as tools in the evaluation of skeletal involvement inpatients with type 1 Gaucher disease. Blood Cells Mol. Dis.,28, 13–20.

18. Ciana, G., Martini, C., Leopaldi, A., Tamaro, G., Katouzian, F.,Ronfani, L. and Bembi, B. (2003) Bone marker alterations inpatients with type 1 Gaucher disease. Calcif. Tissue Int., 72,185–189.

19. Lecourt, S., Mouly, E., Freida, D., Cras, A., Ceccaldi, R., Heraoui,D., Chomienne, C., Marolleau, J.P., Arnulf, B., Porcher, R. et al.(2013) A prospective study of bone marrow hematopoieticand mesenchymal stem cells in type 1 Gaucher disease pa-tients. PLoS ONE, 8, e69293.

20. Grabher, C., Payne, E.M., Johnston, A.B., Bolli, N., Lechman, E.,Dick, J.E., Kanki, J.P. and Look, A.T. (2011) Zebrafish micro-RNA-126 determines hematopoietic cell fate through c-Myb.Leukemia, 25, 506–514.

21. Galloway, J.L.,Wingert, R.A., Thisse, C., Thisse, B. and Zon, L.I.(2005) Loss of gata1 but not gata2 converts erythropoiesis tomyelopoiesis in zebrafish embryos. Dev. Cell, 8, 109–116.

22. Komori, T. (2005) Regulation of skeletal development by theRunx family of transcription factors. J. Cell Biochem., 95, 445–453.

23. Liu, T.M. and Lee, E.H. (2013) Transcriptional regulatory cas-cades in Runx2-dependent bone development. Tissue Eng.Part B Rev., 19, 254–263.

24. Komori, T. (2011) Signaling networks in RUNX2-dependentbone development. J. Cell Biochem., 112, 750–755.

25. Komori, T. (2010) Regulation of bone development and extra-cellular matrix protein genes by RUNX2. Cell Tissue Res., 339,189–195.

26. Cleeter, M.W., Chau, K.Y., Gluck, C., Mehta, A., Hughes, D.A.,Duchen, M., Wood, N.W., Hardy, J., Mark Cooper, J. and Scha-pira, A.H. (2013) Glucocerebrosidase inhibition causes mito-chondrial dysfunction and free radical damage. Neurochem.Int., 62, 1–7.

27. Wauquier, F., Leotoing, L., Coxam, V., Guicheux, J. andWittrant, Y. (2009) Oxidative stress in bone remodelling anddisease. Trends Mol. Med., 15, 468–477.

28. Almeida, M. (2012) Aging mechanisms in bone. BonekeyRep., 1, 1–7.

29. Winkler, D.G., Sutherland, M.K., Geoghegan, J.C., Yu, C.,Hayes, T., Skonier, J.E., Shpektor, D., Jonas, M., Kovacevich,


at UniversitÃ




ownloaded from


B.R., Staehling-Hampton, K. et al. (2003) Osteocyte control ofbone formation via sclerostin, a novel BMP antagonist.EMBO J., 22, 6267–6276.

30. Li, J., Sarosi, I., Cattley, R.C., Pretorius, J., Asuncion, F., Grisan-ti, M., Morony, S., Adamu, S., Geng, Z., Qiu, W. et al. (2006)Dkk1-mediated inhibition of Wnt signaling in bone resultsin osteopenia. Bone, 39, 754–766.

31. van der Horst, G., van der Werf, S.M., Farih-Sips, H., vanBezooijen, R.L., Lowik, C.W. and Karperien, M. (2005) Downre-gulation of Wnt signaling by increased expression of Dick-kopf-1 and -2 is a prerequisite for late-stage osteoblastdifferentiation of KS483 cells. J. Bone Miner. Res., 20, 1867–1877.

32. Yang, C., Rahimpour, S., Lu, J., Pacak, K., Ikejiri, B., Brady, R.O.and Zhuang, Z. (2013) Histone deacetylase inhibitors increaseglucocerebrosidase activity in Gaucher disease by modula-tion of molecular chaperones. Proc. Natl. Acad. Sci. USA, 110,966–971.

33. Yang, C., Swallows, C.L., Zhang, C., Lu, J., Xiao, H., Brady, R.O.and Zhuang, Z. (2014) Celastrol increases glucocerebrosidaseactivity in Gaucher disease bymodulating molecular chaper-ones. Proc. Natl. Acad. Sci. USA, 111, 249–254.

34. Molina, G.A.,Watkins, S.C. and Tsang, M. (2007) Generation ofFGF reporter transgenic zebrafish and their utility in chemicalscreens. BMC Dev. Biol., 7, 62.

35. Lawson, N.D. and Weinstein, B.M. (2002) In vivo imaging ofembryonic vascular development using transgenic zebrafish.Dev. Biol., 248, 307–318.

36. Kirby, B.B., Takada, N., Latimer, A.J., Shin, J., Carney, T.J.,Kelsh, R.N. and Appel, B. (2006) In vivo time-lapse imagingshows dynamic oligodendrocyte progenitor behavior duringzebrafish development. Nat. Neurosci., 9, 1506–1511.

37. Beis, D., Bartman, T., Jin, S.W., Scott, I.C., D’Amico, L.A., Ober,E.A., Verkade, H., Frantsve, J., Field, H.A., Wehman, A. et al.(2005) Genetic and cellular analyses of zebrafish atrioven-tricular cushion and valve development. Development, 132,4193–4204.

38. Hammond, C.L. and Schulte-Merker, S. (2009) Two popula-tions of endochondral osteoblasts with differential sensitiv-ity to Hedgehog signalling. Development, 136, 3991–4000.

39. Spoorendonk, K.M., Peterson-Maduro, J., Renn, J., Trowe, T.,Kranenbarg, S., Winkler, C. and Schulte-Merker, S. (2008) Ret-inoic acid and Cyp26b1 are critical regulators of osteogenesisin the axial skeleton. Development, 135, 3765–3774.

40. Mitchell, R.E., Huitema, L.F., Skinner, R.E., Brunt, L.H., Severn,C., Schulte-Merker, S. and Hammond, C.L. (2013) New toolsfor studying osteoarthritis genetics in zebrafish. OsteoarthritisCartilage, 21, 269–278.

41. Korzh, S., Pan, X., Garcia-Lecea, M., Winata, C.L., Wohland, T.,Korzh, V. and Gong, Z. (2008) Requirement of vasculogenesisand blood circulation in late stages of liver growth in zebra-fish. BMC Dev. Biol., 8, 84.

42. Traver, D., Paw, B.H., Poss, K.D., Penberthy, W.T., Lin, S. andZon, L.I. (2003) Transplantation and in vivo imaging of multi-lineage engraftment in zebrafish bloodless mutants. Nat. Im-munol., 4, 1238–1246.

43. Lin, H.F., Traver, D., Zhu, H., Dooley, K., Paw, B.H., Zon, L.I. andHandin, R.I. (2005) Analysis of thrombocyte development inCD41-GFP transgenic zebrafish. Blood, 106, 3803–3810.

44. Walker, M.B. and Kimmel, C.B. (2007) A two-color acid-freecartilage and bone stain for zebrafish larvae. Biotech. Histo-chem., 82, 23–28.

45. Parsons, M.J., Pollard, S.M., Saude, L., Feldman, B., Coutinho,P., Hirst, E.M. and Stemple, D.L. (2002) Zebrafish mutantsidentify an essential role for laminins in notochord forma-tion. Development, 129, 3137–3146.

46. Thisse, C., Thisse, B., Schilling, T.F. and Postlethwait, J.H.(1993) Structure of the zebrafish snail1 gene and its expres-sion in wild-type, spadetail and no tail mutant embryos. De-velopment, 119, 1203–1215.

47. Amigo, J.D., Suzuki, K., Teplyuk, V., Straubhaar, J. and Lawson,N. (2006) Global analysis of hematopoietic and vascular endo-thelial gene expression by tissue specificmicroarray profilingin zebrafish. Dev. Biol., 299, 551–562.


at UniversitÃ




ownloaded from


Glucocerebrosidase deficiency in zebrafish affects primary bone ossification through increased...

Documents

Transcript of Glucocerebrosidase deficiency in zebrafish affects primary bone ossification through increased...