Zebrafish mafbb Mutants Display Osteoclast Over-Activation and Bone Deformity Resembling Osteolysis in MCTO Patients

Multicentric carpotarsal osteolysis (MCTO) is a rare skeletal dysplasia with osteolysis at the carpal and tarsal bones. Heterozygous missense mutations in the transcription factor MAFB are found in patients with MCTO. MAFB is reported to negatively regulate osteoclastogenesis in vitro. However, the in vivo function of MAFB and its relation to MCTO remains unknown. In this study, we generated zebrafish MAFB homolog mafbb mutant utilizing CRISPR/Cas9 technology. Mafbb deficient zebrafish demonstrated enhanced osteoclast cell differentiation and abnormal cartilage and bone development resembling MCTO patients. It is known that osteoclasts are hematopoietic cells derived from macrophages. Loss of mafbb caused selective expansion of definitive macrophages and myeloid cells, supporting that mafbb restricts myeloid differentiation in vivo. We also demonstrate that MAFB MCTO mutations failed to rescue the defective osteoclastogenesis in mafbb−/− embryos, but did not affect osteoclast cells in wild type embryos. The mechanism of MCTO mutations is likely haploinsufficiency. Zebrafish mafbb mutant provides a useful model to study the function of MAFB in osteoclastogenesis and the related MCTO disease.


Introduction
The bZip factor MafB (v-maf musculoaponeurotic fibrosarcoma oncogene ortholog B) is a member of the large Maf transcription factor family [1]. It is expressed in multiple tissues including the spinal cord, retina and hematopoietic cells, and plays diverse functions in tissue development and cellular differentiation [1][2][3][4][5]. MafB is highly expressed in monocytes and macrophages, and is important for macrophage differentiation [1,4,6]. It has also been demonstrated that MafB regulates osteoclastogenesis [7].
Osteoclasts are bone-resorbing cells that are derived from the hematopoietic monocytemacrophage lineage [8,9]. They are located at or near the bone surface, and degrade the bone in a specialized extracellular compartment called Howship's lacunae by secreting acid and lytic enzymes, like tartrate-resistant acid phosphatase (TRAcP) and Cathepsin K(CTSK) [10]. Osteoclast cells are indispensable for normal bone development and remodeling [11].
The formation of osteoclast cells is regulated by a series of signaling molecules. Two main factors, macrophage colony stimulating factor (M-CSF) and receptor activator of NF-κB ligand (RANKL) positively regulate the osteoclast differentiation [12]. After binding to their respective receptors (M-CSFR or RANK) on the cell surface, they are able to induce the expression of nuclear factor of activated T-cells cytoplasmic 1 (NFATc1)-the master regulator of osteoclastogenesis [7]. Osteoprotegerin (OPG) is an important inhibiting factor of osteoclastogenesis by blocking RANKL binding to RANK [13]. It is known that MafB inhibits the expression of NFATc1 and the osteoclast-associated receptor (OSCAR) during RANKL-mediated osteoclastogenesis, and thereby negatively regulates osteoclastogenesis in vitro [7]. However, the in vivo function of MafB in osteoclastogenesis has not been demonstrated.
Mutations in MAFB have been reported to be responsible for the multicentric carpotarsal osteolysis syndrome (MCTO) [14][15][16]. MCTO is a rare skeletal and nephropathic disorder, and the responsible mutations are all missense mutations in the amino-terminal transactivation domain of MAFB. The patients show aggressive osteolysis, predominantly in the carpal and tarsal bones [17,18]. It is hypothesized that the patients might have overproduced osteoclasts, which eventually leads to underdeveloped and deformed bone formation. However, how MCTO mutation leads to the disease pathogenesis remains unknown. Understanding the function and mechanism of MAFB in normal and pathological conditions will help identify novel therapeutic targets.
Animal models are powerful tools to study the function of genetic mutations involved in disease development. MafB null mutant (MafB −/− ) mice and mice containing an MCTO mutation (MafB MCTO/MCTO ) have been generated [6,19]. These mice show nephropathic symptoms of glomerular sclerosis which is similar to MCTO patients. However, the osteoclast formation and bone development defects in these mice have not been reported. As an alternative animal model, zebrafish has recently been used to study bone development and bone disease. The osteoclast cells and skeletal physiology are similar between zebrafish and mammals [20][21][22]. The biochemical networks and metabolic pathways in zebrafish and mammals are also largely conserved. Moreover, zebrafish embryos are relatively transparent, and the osteoclast cell and bone development can be easily observed by real-time live imaging [23].
There are two paralogs of human MAFB in zebrafish, mafba and mafbb, and mafbb is more preferentially expressed in myeloid lineages during embryogenesis [24][25][26]. We generated mafbb mutants by CRISPR-Cas9 technology, and characterized osteoclast and bone development in the mutants. We then tested the function of MAFB MCTO mutations in mafbb −/− embryos. Our results demonstrated that MAFB negatively regulates osteoclast differentiation in vivo and MCTO mutation is a loss-of-function of MAFB. Zebrafish mafbb mutant provides a useful model to study the in vivo function of MafB and the MCTO disease.

Zebrafish Maintenance and Embryo Handling
The wild type (WT) AB and transgenic zebrafish were maintained, handled, and bred according to standard protocols from the Institutional Animal Care Committee of Shanghai Jiao Tong University. Adult zebrafish were raised in a circulating water system under a 14 h/10 h light/dark cycle at 26-28 • C and fed two times per day. Adult male and female zebrafish were kept separately in the same mating box in the evening and mated the following morning. The embryos were collected and kept at 28.5 • C in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl 2 , and 0.33 mM MgSO 4 ) with a density of 100 embryos per 10-cm-diameter Petri dish. Embryos were staged by hours post-fertilization (hpf) and days post-fertilization (dpf).

Generation and Analysis of Tg(mafbb:GFP) and Tg(ctsk:mGFP)
To generate Tg(mafbb:GFP), the plasmid (pT2-cryR; mafbbCE1-P1Egfp, from Addgene, 20 ng/µL) and 30 ng/µL Tol2 mRNA were co-injected into WT embryos at the one-cell stage. The embryos showing a positive expression of GFP were raised to adults (F0 founder). F0 fish were outcrossed to WT fish, and GFP positive embryos were raised to adults (F1). F1 fish were outcrossed WT again to obtain stable transgenic lines. Transgenic lines were established from two different F0 founders.
For analysis, the embryos were mounted in 1% low-melt agarose and imaged under a confocal microscope Leica SP8 microsystems (Leica, Heidelberg, DE, Germany).

Generation of mafbb Knockout Mutants
mafbb knockout mutants were generated through the CRISPR-Cas9 system and performed following the protocol as described in [27]. Two mafbb specific guided RNAs (as in Figure 1B) were designed to target the beginning site of the exon. One-cell stage WT embryos were injected with 1 nL of the solution containing 100 ng/µL Cas9 mRNA and 20 ng/µL gRNA. Injected F0 fish were grown to adulthood and outcrossed to WT fish. F1 mutant offsprings were identified with T7 endonuclease I (T7E1) assay (M0302S, NEB, Ipswitch, MA, USA) using primers around the target loci. Each target loci was amplified by PCR from the genomic DNA and the mutation was revealed by DNA sequencing. F1 fish were outcrossed to WT to obtain stable F2 mutant lines. Primers used in T7E1 assay and PCR amplification are listed in Table A1 (Appendix A).

Synthesis of Antisense Probes and Whole-Mount In Situ Hybridization (WISH)
We searched the NCBI genebank for sequences of ctsk (Cathepsin K) (Source: ZFIN; Acc: ZDB-GENE-001205-4) and rank/tnfrsf1b (tumor necrosis factor receptor superfamily, member 1B (Source: ZFIN; Acc: ZDB-GENE-070410-133). Briefly, the total RNA extracted from 6 dpf AB larvae was reverse transcribed to cDNA. The cDNA was then used to get a 1 kb fragment of ctsk/rank through PCR. The PCR products were cloned into the Ecor1/Xbal site of the pCS2 (+) backbone vector. SP6 RNA-polymerase (P1088, Promega, Madison, WI, USA) was used to synthesize the digoxigenin-labeled RNA probes. PCR primer sequences are from the literature [28] and listed in Table A2. The PCR products were confirmed by sequencing.

Neutral Red Staining and Benzidine Staining
Optimal staining of macrophages in live embryos was obtained by incubating 3 dpf or 5 dpf embryos in 2.5 µg/mL neutral red/E3 medium (A600652, Sangon Biotech, Shanghai, China) at 28.5 • C in the dark for 6-8 h according to the protocol in [30]. Benzidine staining was done according to [31] with some modification. Larvae were incubated in benzidine (B108444, Aladdin, Shanghai, China) staining solution (2 mL 5 mg/mL benzidine/methanol, 16.7 µL 3 M NaOAc solution, 100 µL H 2 O 2 and 2.483 mL H 2 O) for 30 min in dark, then washed by PBT and fixed in 4%PFA overnight. The stained embryos were imaged under SZX16 stereomicroscope or BX53 microscope (Olympus, Tokyo, Japan). Results are expressed as mean ± SEM, (* p < 0.05, ** p < 0.01, t test). The statistical significance was displayed as (F) Images of embryos at 3 dpf. The black arrow points to the curved tail. (G) Images of adult zebrafish at 9 mpf. The projecting lower jaw (green arrows), the curved spine (black arrows) and the asymmetric caudal fin (red arrows) are shown in mafbb −/− mutants. WT, wild type; HET, mafbb d11/+ ; HOM, mafbb d11/d11 ; mpf, month post fertilization. Results are expressed as mean ± SEM, (* p < 0.05, ** p < 0.01, t test). The statistical significance was displayed as (F) Images of embryos at 3 dpf. The black arrow points to the curved tail. (G) Images of adult zebrafish at 9 mpf. The projecting lower jaw (green arrows), the curved spine (black arrows) and the asymmetric caudal fin (red arrows) are shown in mafbb −/− mutants. WT, wild type; HET, mafbb d11/+ ; HOM, mafbb d11/d11 ; mpf, month post fertilization.

Tartrate-Resistant Acid Phosphatase (TRAcP) Staining
TRAcP staining on zebrafish and zebrafish scales was performed as described previously by using ACID PHOSPHATASE, LEUKOCYTE (Procedure No. 387, Sigma-Aldrich, St Louis, MO, USA) with modification [32]. Briefly, adult zebrafish were fixed in 4% PFA at 4 • C overnight. The fish were then eviscerated to collect the scales for TRAcP staining. For some fish, the rest parts were used to do alizarin red staining. For TRAcP staining on 2 mpf zebrafish, the muscular tissue near the spine was removed first with blades and forceps after fixation.
Scales or whole zebrafish were preprocessed for 2 h at room temperature in tartrate buffer [0.2 M acetate buffer (pH 5.2) with 50 mM L(+)-Tartrate buffer (pH 4.9)] and then incubated in TRAcP staining solution (70 µg/mL Fast Garnet GBC Base Solution; 1 mM Sodium Nitrite Solution; 125 µg/mL Naphthol AS-BI Phosphoric Acid Solution; 0.1 M Acetate Solution; 6.7 Mm L(+)-Tartrate buffer) in the dark for 2 h. The samples were washed three times by PBT before taking images. For quantification of scale TRAcP staining, weak staining means little or no staining at the scale border as in Figure 2I1, and strong staining means a large area of staining as in I2-3 or more at the scale border.

MicroCT Scans
10-mpf (months post-fertilization) were fixed in 4% PFA O/N, then stored in ethanol and scanned with a Skyscan 1176 microCT system (Bruker, Billerica, MA, USA) at 45 kV and 450 µA. We used a voxel size of 18um as resolution. The 3D evaluation was conducted using CTAn (Bruker, Billerica, MA, USA). Quantification of bone morphology was performed on the hypural bone next to the caudal fin, similar to the studies in [33,34]. Determined parameters were bone volume density (BV/TV) and the mean value of grey-level intensity which is corresponded to the relative bone density.

Alcian Blue Staining and Alizarin Red Staining
Cartilage was stained by alcian blue (AB) solution (015-13805, Wako, Osaka, Japan) according to the protocol described in [35]. 5 dpf embryos were fixed in 4% PFA at 4 • C overnight and bleached in 1% KOH/3% H 2 O 2 solution until the pigments were cleared. After that, embryos were incubated with acid alcohol (1% HCl in 70% ethanol) for 20 min before transferred into a 0.1% alcian blue in acid alcohol for 2 h. The specimens were mounted in 70% glycerol and photographed after washed three times, 30 min per time by acid alcohol.
For larvae teeth staining, 8 dpf larvae were fixed and bleached as for alizarin red staining, following the procedure in [35]. The larvae were incubated with 1 mg/mL Alizarin red S (71001954, China National Medicines Corporation Ltd., Beijing, China) in 1% KOH for 1 h and digested for several hours in 1 mg/mL trypsin (A100260-0250, Sangon Biotech, Shanghai, China) in 2% borax (10020808, China National Medicines Corporation Ltd., Beijing, China).
Adult zebrafish bone staining was similar to larval bone staining. Briefly, 9-month old zebrafish were anesthetized with 0.02% tricaine and sacrificed on ice before fixed in 4% PFA at 4 • C overnight. The specimens were bleached for 1 day and transferred to 30 % saturated sodium tetraborate overnight. Afterwards, the specimens were placed in 1% KOH with 1 mg/mL Alizarin Red overnight and 1% trypsin and 2% borax were added until 85% of the soft tissue was dissolved. The specimens were transferred through a series of 1% KOH/glycerol solutions until they settled at the bottom. The specimens were transferred to 70% glycerol for long-term storage and photographing.

Whole Kidney Marrow Cell Collection and Cytology
The adult fish kidney marrow was dissected and placed into tubes containing 400 µL FBS solution. The single hematopoietic cells from kidney marrow were generated by pipetting and filtration through 40-µm filters. Single-cell suspensions were diluted to 15,000-30,000 cells/mL and cytocentrifuged at 400 rpm for 4 min with cytospin 4 (Sigma-Aldrich, St Louis, MO, USA). Blood smears were processed by May-Grünwald and Giemsa double stain (63590/48900, Sigma-Aldrich, St Louis, MO, USA) for morphological analysis and differential cell counts.

Flow Cytometry and Cell Sorting and Counting
We sorted ctsk + cells from Tg (ctsk:mGFP) in WT and mafbb +/− embryos at 3 dpf and 5dpf. Embryos were anesthetized with 0.02% tricaine and washed with sterile E3 solution. Single cells were collected by shredding larvae with a blade and incubated for 20 min (37 • C) with 38 µg/mL Liberase (05401119001, Roche, Basel, Switzerland). 10% FBS was added to stop the reaction, followed by filtration (40 µm filter) and centrifugation (5000 rpm, 4 • C, 15min). The supernatant was removed, and cells were resuspended with 800 µL 2% FBS. GFP negative and positive cells were sorted with FACS AriaII (Becton, Dickinson and Company, NJ, USA).

Gene Expression by Real-Time qPCR
Gene expression was evaluated using real-time qPCR. Briefly, total RNA was extracted from embryos with TRIzol reagent (10296028, Thermo Fisher Scientific, Waltham, MA, USA). cDNAs were synthesized from total RNA using the PrimeScript TM RT reagent Kit with gDNA Eraser (RR047A, Takara, Shiga, Japan). TB Green Premix Ex Taq TM ii (RR820A, Takara, Shiga, Japan) was used for qPCR analysis. Each target gene was calculated using the 2 −∆∆CT method [36]. The primers for different target genes and β-actin (the reference gene) are listed in Table A3.

Generation of Zebrafish mafbb Mutants Using the CRISPR/Cas9 System
Zebrafish contains two mafb genes, mafba and mafbb. The two genes are expressed in distinct and overlapping areas during development. Previous studies have suggested that MafB is highly expressed in myelomonocytic lineages of hematopoietic cells in mice [3]. In zebrafish, mafbb is highly expressed in myeloid lineages and macrophages, we therefore focused on mafbb. To further explore the expression of mafbb in zebrafish, we generated a transgenic line using the construct (mafbbCE1-P1E:GFP) in which an enhancer and a cognate promoter element of mafbb gene are used to drive the expression of the enhanced green fluorescent protein [38]. At 24 hpf, transgenic embryos showed good GFP expression in primitive myeloid (pm) cells, the caudal hematopoietic tissue (CHT), posterior cardinal vein (pcv), hindbrain (hb) and otic vesicle (ot) during the early stages of development ( Figure 1A,A1-4), similar to the analysis from whole-mount in situ hybridization (WISH) in zebrafish embryos [24], and to the GFP expression in mouse transgene containing MafB 5 -upstream fragment fused to GFP [3]. Tg(mafbb:GFP) was crossed to Tg(mpeg1:mCherry) which directs mCherry expression in macrophages [39]. We observed that a large number of mCherry + macrophages co-expressed GFP in the head and the CHT region at 3 and 5 dpf ( Figure 1A,A5-10). These data support the robust expression of mafbb in macrophages, and indicate a potential function of mafbb in macrophage and osteoclast cells.
To study the role of mafbb in zebrafish osteoclast differentiation and bone development, we generated mafbb mutant utilizing CRISPR/Cas9 technology. The 318 aa protein Mafbb is encoded by the 1612 bp gene mafbb, which contains one exon ( Figure 1B). A guide RNA that targets the N-terminal region of mafbb exon ( Figure 1B) was designed and co-injected with Cas9 mRNA into one-cell stage WT embryos. After screening several founders that transmitted to F1 progeny, we established a stable line named mafbb d11/d11 that has a frameshift mutation caused by a deletion of 11 bp. We also used a second gRNA with a different target site and obtained a second mutant line named mafbb d4/d4 with a deletion of 4 bp ( Figure 1C,D). The two mafbb mutant lines demonstrate the same phenotypes. Here, we show the results from mafbb d11 allele for the rest of the paper.

Osteoclast Cell Development Is Enhanced in mafbb Deficient Embryos
To investigate whether mafbb is involved in osteoclastogenesis, we first examined the osteoclast-like cells in zebrafish embryos through WISH of two osteoclast makers, ctsk and rank [29]. The WISH results showed that ctsk (Figure 2A) and rank ( Figure 2B) were expressed in the pharyngeal arches and the pectoral fin, as in the previous report [28]. At 2 dpf, the expression of ctsk and rank in mafbb +/− and mafbb −/− showed no difference compared to WT. However, at 3 and 4 dpf, the expression of ctsk and rank increased significantly in mafbb +/− and mafbb −/− (Figure 2C,D).
To analyze the osteoclast formation through the development course of mafbb mutants, we generated zebrafish Tg (ctsk:mGFP) using the plasmid expressing membrane bound EGFP (mEGFP) under the control of a medaka cathepsin K (ctsk) promoter [40]. Stable transgenic fish showed GFP expression comparable to the medaka transgenic line from the same plasmid [40], and to the zebrafish bacterial artificial chromosome (BAC) recombineering-based YFP labeled ctsk transgenic line [28]. At 5 dpf, GFP is expressed in the head (including the pharyngeal arches) and the tail region (likely in the caudal fin) ( Figure 2E,E1-3). At this stage, GFP expression was increased in mafbb −/− embryos compared to WT siblings. The percentage of GFP-expressing osteoclast-like cells in mafbb −/− was significantly expanded in comparison to WT (7.8 ± 0.4% vs. 3.5± 0.01%, Figure 2F). At later stages, GFP expression was found in the vertebrate column, around the neural and haemal arches ( Figure 2E4), similar to the expression shown in [40]. In 1-month-old mafbb −/− zebrafish, GFP expression along the neural and haemal arches was strongly increased compared to WT siblings ( Figure 2E8,G).
To further analyze the effects of mafbb on osteoclast differentiation, we sorted ctsk + cells from Tg (ctsk:mGFP) in WT and mafbb +/− embryos, and examined the relevant gene expression related to osteoclastogenesis by RT-PCR ( Figure 2H). The results showed that osteoclast differentiation genes fosab and nfatc1, osteoclast maturation genes ctsk, acp5a and ocstamp [41] were all upregulated in ctsk + cells sorted from mafbb +/− compared to WT at 3 and 5 dpf. However, the inhibitor factor opg (also known as tnfrsf11b) [13] was downregulated in mafbb +/− at 3 and 5 dpf ( Figure 2H). These data support that osteoclast differentiation is enhanced when mafbb is down-regulated.
To evaluate osteoclast cells in adults, we used TRAcP enzyme staining on adult zebrafish scales, which are part of the dermal skeleton and contain both osteoblasts and osteoclasts. The results showed significant enhancement of TRAcP activity in mafbb +/− and mafbb −/− scales, particularly at the scale border compared to the WT ( Figure 2I,J). We also performed TRAcP staining on adult zebrafish, and found enhanced osteoclast activity in the spine and caudal fin of mafbb −/− fish compared to WT siblings ( Figure 2K). Thus, the development and maturation of osteoclast cells are persistently increased in mafbb-deficient zebrafish.

mafbb Deficiency Results in Abnormal Cartilage and Bone Development
To explore whether mafbb mutation causes any bone development defects, we performed alcian blue (AB) staining on 5 dpf embryos to visualize the cartilages of the pharyngeal skeleton [42]. The results showed that Meckel's cartilage (mk) in 15% of mafbb +/− embryos is bent ventrally compared to WT ( Figure 3A). In 75% of mafbb −/− embryos, mk, basihyal (bh), and ceratohyal (ch) are all bent ventrally ( Figure 3A). In a small portion (5%) of mafbb −/− embryos, only three ceratobranchial pairs (cb) were clearly stained instead of 5 as seen in WT ( Figure 3A3). We measured the ceratohyal angel (CHA) to quantify the cartiliage abnormality [43]. The CHA widens from 60 • to around 100 • ( Figure 3B). Hence, mafbb mutation leads to an abnormal arrangement of pharyngeal skeleton. We used Alizarin red (AR) to stain for the bone mineralization in the larvae at 8 dpf. Although mafbb +/− embryos do not show changes in AR staining of bone and teeth compared to WT (data not shown), the mafbb −/− embryos demonstrate a strong reduction in the mineralization, and 10% of the embryos have a reduced number of mineralized teeth ( Figure 3C). To evaluate the bone formation at later stages, we stained 9 mpf adults with AR. The overall bone structure in the majority of mafbb +/− fish was comparable to WT, except for the lack of small holes frequently seen in hypural 1 and 2 next to the caudal fin in WT ( Figure 3F). About half mafbb −/− zebrafish displayed the protruding lower jaw, curved spine and the corresponding abnormal bone structure. The shape of haemal and neural arches are more irregular and the hypurals are frequently fused in mafbb −/− ( Figure 3F). Thus, zebrafish mafbb deficiency leads to defective cartilage and skeleton development. We further used micro-computed tomography(microCT) to examine the bone quality in adult fish [33,34]. As bone volume fraction (BV/TV, bone volume/total tissue volume) is a useful parameter for osteoporotic changes [33], we evaluated BV/TV for the hypural bones ( Figure 3D). BV/TV is decreased in mafbb −/− compared to WT ( Figure 3E). Moreover, the mean grey-level intensity in the hypural bones, as assessed by microCT [44], is also decreased in mafbb −/− compared to WT. These results are consistent with the phenotypes involved in osteoporosis, supporting the enhanced activation of osteoclast cells in the mafbb −/− mutants.

Macrophage Differentiation Is Altered in mafbb Mutants
Osteoclast cells are developed from the monocyte/macrophage precursors through a multistep process. While repressing osteoclastogenesis, MafB might enhance macrophage differentiation [6]. Similar to mammalian systems, zebrafish macrophage generation occurs in distinct waves during primitive and definitive hematopoiesis. The Ets transcription factor PU.1/Spi-1 is a master regulator of myeloid cell development. It is critical for macrophage differentiation, and also critical for the induction of NFATc1, Ctsk, and TRAcP during osteoclast differentiation [45]. We first examined the primitive macrophage development in mafbb mutants. The expression of pu.1 and the macrophage specific marker mfap4 in the rostral blood island and the ventral tail region are both significantly decreased in mafbb −/− mutant embryos compared to WT from 22-24 hpf ( Figure 4A,B).
We then stained the embryos with pu.1 and mfap4 at later stages for definitive macrophage development. Surprisingly, pu.1 and mfap4 expression in the ventral trunk region (the circulating macrophages) in mafbb −/− is increased compared to WT at 3 and 5 dpf ( Figure 4A,C). Similarly, neutral red (NR)-stained macrophages were increased in the tail of mafbb mutants. However, NR-stained macrophages in the brain (microglia) were significantly decreased in the mutants. A microglia specific marker apoe is also strongly reduced in mafbb −/− (Figure 4B,C). It is known that the embryonic microglia uniquely derive from primitive macrophages [46]. The reduced microglia development supports the reduction of primitive macrophages. To further characterize the macrophage development, we used Tg(mpeg1:mCherry) and performed flow cytometry to analyze the total number of mCherry + -macrophages in different embryos. At 5 dpf, the percentage of mCherry + macrophages in mafbb −/− was significantly increased in comparison to WT (0.75 ± 0.02%, vs. 0.18± 0.01%, Figure 4D). Thus, mafbb mutants lead to expanded definitive macrophage differentiation and inhibit the primitive macrophage development.

Expansion of Definitive Myelopoiesis in mafbb Mutants
Previous studies suggest that MafB might be involved in the development of multiple hematopoietic lineages in addition to macrophages [47]. To check whether lack of mafbb leads to preferential hematopoietic lineage differentiation, we examined the hematopoiesis in mafbb −/− by analyzing the development of various blood cells at different stages ( Figure 5A,B). The results showed that the hematopoietic stem and progenitor cells (HSPCs), as stained by runx1 and cmyb, were decreased in mafbb −/− at 24 hpf and 3 dpf, and were slightly decreased by 5 dpf. rag1 + -T lymphocytes were slightly decreased at 5 dpf. The hbbe1 + and benzidine stained erythrocytes are also decreased in mafbb −/− from 3-5 dpf. In contrast, mpx stained neutrophils were significantly increased in mafbb −/− compared to WT ( Figure 5A, B), similar to the increased staining of myeloid progenitors and macrophages in mafbb −/− mutant during the definitive hematopoiesis. We then examined mpx expression in the primitive neutrophils, but did not see visible changes in mafbb −/− compared to WT (Figure 5A, B).
To study the myeloid lineages in adult zebrafish, we analyzed the hematopoietic cells from the whole kidney marrow (WKM) via cytological assay. The overall kidney morphology between WT and mafbb −/− are similar ( Figure 5C,D). For the hematopoietic cells, the population of lymphoid cells, and of the precursors were all slightly decreased in mafbb −/− , but the population of myelomonocytes was expanded in mafbb −/− , consistent with the phenotypes during embryogenesis ( Figure 5C,D). Thus, mafbb loss of function leads to selective expansion of definitive myelopoiesis including macrophages and neutrophils, and a small reduction of lymphoid and erythroid differentiation.

Expansion of Definitive Myelopoiesis in mafbb Mutants
Previous studies suggest that MafB might be involved in the development of multiple hematopoietic lineages in addition to macrophages [47]. To check whether lack of mafbb leads to preferential hematopoietic lineage differentiation, we examined the hematopoiesis in mafbb −/− by analyzing the development of various blood cells at different stages ( Figure 5A,B). The results showed that the hematopoietic stem and progenitor cells (HSPCs), as stained by runx1 and cmyb, were decreased in mafbb −/− at 24 hpf and 3 dpf, and were slightly decreased by 5 dpf. rag1 + -T lymphocytes were slightly decreased at 5 dpf. The hbbe1 + and benzidine stained erythrocytes are also decreased in mafbb −/− from 3-5 dpf. In contrast, mpx stained neutrophils were significantly increased in mafbb −/− compared to WT ( Figure 5A,B), similar to the increased staining of myeloid progenitors and macrophages in mafbb −/− mutant during the definitive hematopoiesis. We then examined mpx expression in the primitive neutrophils, but did not see visible changes in mafbb −/− compared to WT ( Figure 5A,B). To study the myeloid lineages in adult zebrafish, we analyzed the hematopoietic cells from the whole kidney marrow (WKM) via cytological assay. The overall kidney morphology between WT and mafbb −/− are similar ( Figure 5C,D). For the hematopoietic cells, the population of lymphoid cells, and of the precursors were all slightly decreased in mafbb −/− , but the population of myelomonocytes was expanded in mafbb −/− , consistent with the phenotypes during embryogenesis ( Figure 5C,D). Thus, mafbb loss of function leads to selective expansion of definitive myelopoiesis including macrophages and neutrophils, and a small reduction of lymphoid and erythroid differentiation. [14][15][16]. Whether the mutation in MCTO is haploinsufficiency or a dominant negative effect remains unclear. mafbb −/− zebrafish is a null mutant and has increased osteoclastogenesis, and Mafbb protein has 52% amino acid sequence identity with human MafB. We thus used this as a model to determine whether MCTO mutations could rescue the defects. The mafbb −/− embryos were injected with either WT or P59L mutated human MAFB mRNA. The WT MAFB mRNA reduced the overproduced ctsk +cells, while the P59L MAFB mRNA did not ( Figure 6A,B). Similarly, the P63R and S70L MAFB mRNA were not able to reduce ctsk expression in mafbb −/− (Figure 6C,D). In addition, the P59L MAFB mRNA injection in WT embryos did not interfere with the ctsk expression ( Figure 6A,B). These results support that the MAFB MCTO mutations are loss-of-function mutations, and the heterozygosity of MCTO patients might be haploinsufficiency.

MCTO individuals carry heterozygous missense mutations in MAFB with a dominantinheritance pattern
Currently, there are no effective treatments for MCTO patients [48]. Bisphosphonates may be effective by interrupting the osteoclast activity and formation [49]. A previous report suggested that bisphosphonates may slow down the progression of bone destruction in MCTO [16]. We exposed WT, mafbb +/− , mafbb −/− embryos to alendronate (AL), a type of bisphosphonate, and then examined the osteoclastogenesis and the cartilage development by ctsk expression and AB staining after the treatment. AL treatment reduced the increased ctsk + expression in mafbb −/− embryos at 3 and 5 dpf ( Figure 6E-H). More importantly, the cartilage abnormalities such as the jaw protrude, in mafbb −/− mutants were almost completely recovered after 4-day treatment of AL ( Figure 6I,J). Thus bisphosphonates are capable of inhibiting the excessive osteoclastogenesis in mafbb mutants. The mafbb −/− mutant is a valid model to understand the pathophysiology of MCTO, and aid in the identification of novel therapies.

Discussion
Osteoclasts are produced from monocyte/macrophage lineage through a differentiation process induced by M-CSF and RANKL [12]. In bone marrow-derived monocyte/macrophage lineage cells, knockdown of MafB enhances RANKL-mediated osteoclastogenesis, while overexpression of MafB decreased the process [7]. Thus, MafB negatively regulates RANKL-induced osteoclastogenesis in vitro. However, the MafB function in vivo remains unclear. Mice with a homozygous deletion in the MafB gene or transgenic mice containing MCTO mutation (MafB MCTO/MCTO ) have been generated, but their osteoclast development and bone formation defects have not been demonstrated.
In our studies, zebrafish MafB homolog mafbb deletion leads to consistently increased osteoclastogenesis and the subsequent bone growth defects, supporting that MafB also inhibits osteoclast development in vivo. The inhibition of osteoclastogenesis by mafbb might be dosage-dependent, as the mafbb heterozygotes also manifested considerable osteoclast cell overgrowth and bone development problems. Given the presence of an additional MafB homolog, mafba in zebrafish, mafbb homozygous mutants might not represent a MafB knockout, but are probably relevant for the autosomal dominant MCTO. mafba could play a compensatory role in osteoclastogenesis, it will be important to evaluate the function of mafba and the combined function of mafba and mafbb in osteoclastogenesis and bone growth to better reveal the conserved roles of MafB across different species.
In MCTO patients, only heterozygous missense mutations in MAFB are observed, and the mutations have a dominant inheritance pattern. Whether the pathogenesis of MCTO involves haploinsufficiency or a dominant-negative effect remains unknown. MCTO mutations all lie within a short region of the transcriptional activation domain of MAFB (amino acids 54-71), and map to the phosphorylation sites that determine MAFB protein stability [50]. It was hypothesized that MCTO mutations increase the MAFB protein stability and the individual might have enhanced MAFB expression [50]. Using zebrafish mafbb null mutants, we demonstrated MAFB MCTO mutations failed to reduce the excessive osteoclastogenesis in the mutant embryos. This provides evidence that MCTO mutations resemble loss of functions. Our findings are consistent with the finding showing the reduced transactivation activity of the mutated MAFB proteins [50]. We also injected mutant MAFB mRNA in WT embryos, and did not see noticeable effects on ctsk staining. Thus, MCTO mutations are likely haploinsufficiency. In the future, analysis of osteoclast differentiation in zebrafish or mice containing the MCTO mutations will help elucidate the precise pathogenesis of MCTO osteolysis.
While repressing osteoclastogenesis, MafB might enhance macrophage differentiation. Indeed, overexpression of MafB induced macrophage differentiation in chick myeloblasts [4]. However, in MafB-deficient mice, the number of macrophages was not impaired [6]. It was also found that combined knockout of MafB and C-Maf increases self-renewal of macrophages [51]. These discordant results highlight that the complex functions MafB in macrophages. Our studies indicate that MafB likely affects macrophage development differently at different stages. During the early developmental stages, the primitive myeloid progenitors and macrophages are strongly decreased in mafbb mutants. In support of this, the embryonic microglia, which exclusively derive from primitive macrophages, remain decreased in mafbb mutants. To our surprise, starting from 3 dpf, pu.1 expression in myeloid cells, mfap4 + -, NR stained-and Tg(mpeg1:mCherry) labeledmacrophages (circulating macrophages) are all expanded in mafbb mutants. Thus, mafbb deficiency leads to increased differentiation of definitive macrophages. Further studies using conditional and/or inducible deletion of MafB will help to dissect MafB function in macrophages during different stages and in different populations.
In mafbb mutants, HSPCs, the lymphoid cells and erythroid cells are all decreased or slightly decreased. However, the myeloid cells, including neutrophils in addition to macrophages, are significantly expanded, and the myeloid expansion remains till the adult stage. Our findings are consistent with the results by Sarrazin et al. [52]. They demonstrate that MafB deficiency leads to activation of myeloid master regulator Pu.1 and enhanced myeloid commitment in hematopoietic stem cell (HSC) reconstitution. Considering that osteoclast cells also originate from myeloid cell lineage, our result support that mafbb restricts definitive myeloid differentiation in vivo.

Conclusions
Here, we report that zebrafish mafbb deficiency leads to enhanced osteoclastogenesis and bone growth defects resembling MCTO. In addition, loss of mafbb in zebrafish reduces the primitive macrophage development, but expands definitive macrophage and myeloid lineage differentiation. Using mafbb mutants, we further demonstrate that MCTO missense mutations are loss of function mutations, and the pathogenesis of MCTO is likely haploinsufficiency.

Acknowledgments:
We would like to thank Christoph Winkler at National University of Singapore (Singapore) for providing the plasmid (ctsk:mGFP); Li Li at Southwest University (China) for providing us the plasmids of WISH probes.

Conflicts of Interest:
The authors declare no conflict of interest. Table A1. Nucleotide sequence of primers used for mafbb mutants genotyping.