Oncogenic FLT3 internal tandem duplications (ITD) and CD45/PTPRC control osteoclast functions and bone microarchitecture

Abstract

Activating internal tandem duplications (ITD) in the juxtamembrane domain of receptor tyrosine kinase FLT3 occur frequently in patients with acute myeloid leukemia (AML). Constitutive active FLT3-ITD mutations induce aberrant signaling and promote leukemic cell transformation. Inactivation of the attenuating receptor protein tyrosine phosphatase CD45 (PTPRC) in FLT3-ITD mice resulted in the development of a severe hematopoietic phenotype with characteristics of AML. In addition, abnormal bone structures and ectopic bone formation were observed in these mice, suggesting a previously unknown role of FLT3 to control bone development and remodeling. While Ptprc knockout and Flt3-ITD mutant mice showed a largely normal bone microarchitecture, micro-CT analysis of femurs from Flt3-ITD Ptprc knockout mice revealed trabecularization of the cortical bone. This resulted in increased trabecular bone volume at the metaphysis, while the cortical bone at the diaphysis was thinner and less dense. In the metaphysis, severely reduced osteoclast and osteoblast numbers were observed. Reduced capacity of ex vivo differentiation of CD11b-positive bone marrow stem cells to mature osteoclast was accompanied by their abnormal morphology and reduced size. Transcriptome analysis revealed reduced expression of osteoclastogenic genes. Unexpectedly, cumulative resorption activity of osteoclasts was increased. Size and structure of resorption pits of differentiated osteoclasts remained similar to those observed in osteoclast cultures derived from control animals. Enhanced proliferation of cells in osteoclast cultures derived from FLT3-ITD–expressing mice was mediated by increased expression of STAT5 target genes. Transcriptome analysis of differentiated osteoclasts showed dysregulated signaling pathways influencing their differentiation as well as the coupling of bone resorption and formation. Taken together, inactivation of attenuating CD45 in mice expressing oncogenic FLT3-ITD resulted in marked abnormalities of the osteo-hematopoietic niche, which can be explained by aberrant STAT5 activation.

Lay Summary

Acute myeloid leukemia (AML) is frequently linked to mutations in the receptor tyrosine kinase FLT3. These mutations, called FLT3-ITD, cause the gene to become constitutive active, leading to abnormal cell behavior and cancer development. While investigating Flt3-ITD mutant mice corresponding to human AML patients with FLT3 mutations, we previously found that these mice had an unexpected phenotype of bone resorption when CD45/Ptprc, a phosphatase that normally regulates FLT3 activity, was additionally knocked out. In the present study, we investigated the details of this bone resorption phenotype. MicroCT scans revealed unusual bone patterns like thinner cortical areas and increased spongy bone structures in other areas. These changes were linked to fewer bone-building cells (osteoblasts) and bone-resorbing cells (osteoclasts). While remaining osteoclasts appeared to smaller in size, they behaved abnormally, showing increased overall activity. Further analysis revealed that these changes were driven by altered signaling pathways linked to an overactive protein, STAT5, which is a known target of FLT3-ITD. Ptprc knockout without FLT3-ITD background did not lead to any of the observed changes. Thus, this study highlights a previously unknown capacity of AML-associated FLT3 mutations to regulate bone remodeling offering new insights how the bone and blood systems are interconnected.

receptor tyrosine kinase, FLT3 ITD, AML, protein tyrosine phosphatase, CD45, bone density, osteoclast, osteoblast differentiation

Introduction

FMS-like tyrosine kinase 3 (FLT3) is a class III receptor tyrosine kinase (RTK), promoting proliferation, survival, and differentiation of hematopoietic progenitors. It plays a central role in the development of lymphoid and myeloid cells.^1–3 Ligand-independent, constitutive FLT3 activity caused by internal tandem duplications (ITD) in the juxtamembrane domain of FLT3 has been linked to disruption in cellular homeostasis and the development of acute myeloid leukemia (AML).^4,5 Activation of several signaling pathways by constitutive FLT3 ITD kinase activity mediates leukemic cell transformation, including the aberrant activation of the transcription factor signal transducer and activator of transcription 5 (STAT5).^6,7

Protein tyrosine phosphatases (PTP), such as receptor type PTPRC, also known as CD45, are able to antagonize the kinase activity of FLT3 ITD in vitro and in vivo.^8–10 PTPRC is expressed in hematopoietic cells and cells with hematopoietic origin–like osteoclasts (OCs).^11,12 Aberrant PTP functions are associated with human diseases.^13–15 In addition, PTPRC regulates retention and motility of hematopoietic progenitors as well as bone remodeling. Ptprc knockout (KO) mice show an abnormal trabecular distribution pattern and defective OC function, leading to an osteopetrotic bone phenotype.¹² We have recently shown that, in patients with AML carrying FLT3 ITD mutations, reduced PTPRC expression was negatively correlated with patient survival.¹⁰ In contrast, patients with AML with WT FLT3 show no correlation between PTPRC expression and patient survival, suggesting a negative role of PTPRC in modulating FLT3 ITD activity.^10,16,17

Mice with the bi-allelic knock-in of FLT3 ITD (hereafter referred to as Flt3^ITD/ITD mice) develop a fully penetrant myeloproliferative neoplasm, comparable to human chronic myelomonocytic leukemia.¹⁰ PTPRC deficiency in Flt3^ITD/ITD mice resulted in exacerbation of the FLT3 ITD–based hematopoietic phenotype and a severely shortened lifespan. Flt3^ITD/ITD Ptprc^-/- mice develop severe monocytosis, anemia, hepatosplenomegaly, and organ structure aberrancies along with increased transforming signaling of FLT3 ITD and its downstream target STAT5. Furthermore, the inactivation of PTPRC in Flt3^ITD/ITD mice led to cortical porosity, reduced bone density in the femoral diaphysis, and ectopic bone formation at the spleen.¹⁰

Due to increased numbers of monocytic osteoclast progenitor cells resulting from monocytosis, this severe bone phenotype might be caused by changes in osteoclastogenesis and bone resorption. To decipher the mechanism(s) underlying these aberrancies, we comprehensively analyzed the effects of FLT3 ITD and inactivation of PTPRC on bone remodeling and osteoclastogenesis. We report that Flt3^ITD/ITD mice lacking PTPRC developed an abnormal trabecular and cortical bone structure and showed altered bone cell numbers. OC differentiation of Flt3^ITD/ITD Ptprc^-/- cells was impaired with OCs appearing smaller and less mature compared with WT and Ptprc^-/- OCs. On the contrary, the bone-resorption activity of Flt3^ITD/ITD Ptprc^-/- OC cultures was increased. Osteoclastic progenitor cells with constitutive FLT3 ITD activity showed higher proliferation activity due to STAT5 activation, which was further increased by the inactivation of PTPRC. Taken together, altered osteoclastogenesis and increased osteoclast function may be the main driver of the severe bone phenotype of Flt3^ITD/ITD Ptprc^-/- mice.

Materials and methods

Mouse lines

C57BL/6 J (WT), B6.129-Ptprc^tm1Holm/H (subsequently named Ptprc^-/-), Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- female and male mice were used in this study (see also Table S1).¹⁰ The actual number of mice is indicated in the legend to each of the figures. Primers used to validate Flt3 ITD and Ptprc KO mutations are given in Table S2. The committee of the Thuringian State Government on Animal Research (TVA reg. no. UKJ-20-009)–approved animal procedures and experiments were performed according to the German Animal Welfare Act. Mice were treated in accordance the guiding principles in the care and use of animals. Mice were subjected to a 12-hour light/dark cycle and were fed a standard diet with water ad libitum. Enrichment was provided in the form of cardboard houses and bedding material. Animals of particular genotypes were randomly assigned to groups and the subsequent analyses were performed in a blinded fashion.

Bone μCT and histomorphometry

Twelve- to 15-week-old mice received oral analgesics, were anesthetized, and subsequently perfused with PBS and 4 % PBS-buffered paraformaldehyde (PFA) for 10 minutes. The lumbar vertebrae, tibiae, and femora were removed and kept in PBS-buffered PFA for 48 hours.

For μCT measurements, bones were rinsed twice with distilled water (ddH₂O) and stored in 50% ethanol at 4°C. Bone microarchitecture was analyzed ex vivo using the viva CT40 (Scanco Medical, Switzerland). Imaging was performed on the femur and the fourth lumbar vertebra, with a resolution of 10.5 μm using X-ray energy set at 70 kVp, 114 mA, and an integration time of 200 ms. Evaluation of trabecular bone in the distal femur focused on the metaphysis, specifically 20 slices below the last slide, where primary spongiosa was observed, with a total of 150 slices. In the vertebral bone, measurements were taken over 150 slices between both growth plates. Cortical bone analysis was conducted at the femoral midshaft using 150 slices, and the assessment utilized predefined scripts provided by Scanco.

For histomorphometry, femora were decalcified with Osteosoft (Merck, Germany) for 7 days. Decalcifying solution was replaced daily. The bones were dehydrated in an ascending ethanol series prior to embedding in molten paraffin. To determine numbers of bone cells, bones were sectioned in 4-μm slices and stained for tartrate-resistant acid phosphatase (TRAP). Numbers of osteoclasts (N.Oc/BS), stained in pink, as well as osteoblasts (N.Ob/BS), osteocytes (N.Ocy), and adipocytes (N.Adipo) were identified per bone surface by their morphology and their location and were counted using Osteomeasure software (Osteometrics, USA) following international standards.

Quantification of bone turnover markers

Fourteen-week-old mice were anesthetized with 2% isoflurane and killed by cervical dislocation. The abdomen was opened and the heart was punctured with an 18G cannula to draw blood. Separation of serum and other blood components was achieved by incubation at room temperature for 30 minutes, followed by centrifugation with 14 000 × g for 10 minutes. Levels of CTX I, TRAP5b, and procollagen type I N-terminal propeptide (P1NP) were determined in the blood serum according to the manufacturer’s protocol using Serum CrossLaps (CTX-I) ELISA, Mouse TRACP-5b ELISA, or rat/mouse PINP enzyme immunoassay (EIA) (kits from Immunodiagnostic Systems [IDS]), respectively.

Primary OC culture

CD11b-positive BM cells purified via magnetic cell sorting were seeded with a cell density of 3 × 10⁵ cells/cm² and cultured in alpha-MEM (Sigma Aldrich, Germany) with 10% heat-inactivated FCS (Sigma Aldrich, Germany), 1% penicillin/streptomycin (Sigma Aldrich, Germany), and 50 ng/mL murine M-CSF (Immunotools, Germany) for 3 days. Osteoclasts were then generated using the media described previously with addition of 20 ng/mL murine RANKL (R&D Systems, Germany) for the remainder of the culture (3–10 days). Medium was changed on day 3 and day 5. Where indicated, medium was supplemented with 20 nM AC220 in FLT3 ITD inhibition experiments.

Osteoclasts were stained for TRAP activity at maximum differentiation (between 3 and 5 days) and imaged with the Evos FL Auto Microscope (Invitrogen ThermoFisher Scientific). The total number of multinucleated OCs per well was counted. Using the multipoint tool in the ImageJ software, the area occupied by mature OCs was quantified for 20 OCs in each scan, randomly selected from different regions of the scan. Cells with extended size and fully formed actin rings were defined as mature OCs, while small multinucleated cell were counted as immature OCs. TRAP-positive multinucleated cells with spread cytoplasm were manually annotated in 1 scan per mouse. TRAP levels in cell supernatants were determined using para-nitro phenylphosphate dephosphorylation assay on day 5 of differentiation after 48 hours of accumulation period. Osteoclast proliferation was quantified at the indicated stages of the differentiation process using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega, Germany).

To determine bone resorption activity, CD11b-positive BM cells were seeded on bovine bone slices (IDS, Germany). Medium was changed on day 3 of differentiation and cells were differentiated for 10 days. Medium was supplemented once again on day 7 with 10% FCS, 50 ng/mL murine M-CSF, and 20 ng/mL murine RANKL. Bone resorption was assessed by scanning electron microscopy of lacunae on bone slices. Then, the bone samples were cleaned and sputter-coated with gold (thickness ~2 nm) using a sputter-coater CCU-010 (Safematic GmbH, Switzerland) to prevent surface charging. Specimens were investigated with a field emission SEM LEO-1530 Gemini (Carl Zeiss NTS GmbH, Germany) at 12-kV acceleration voltages. To emphasize the shallow lesions on the surface, images were taken using the side-mounted Everhart–Thornley detector in back-scattered mode. Images of resorption pits were taken at 200× and 2000× magnification. Cumulative resorption activity was analyzed using cell supernatants on day 10 of differentiation. CTX-I levels were quantified using CrossLaps for Culture (CTX-I) ELISA (IDS, Germany).

Primary osteoblast culture and differentiation

The diaphyses of the long bones were cut into small pieces and incubated in DMEM supplemented with 10% FCS, 1% penicillin/streptomycin, and 125 U/mL collagenase II rotating with 1000 rpm for 60 minutes. Subsequently, bone pieces and cells were centrifuged at 300 × g for 10 minutes, washed with osteoblast (OB) medium (DMEM with 10% FCS, 1% penicillin/streptomycin, and 2.5 μg/mL amphotericin B), and plated in 6-cm petri dishes in OB medium. For physiological and functional analysis of OBs, bone pieces and cells were cultured in OB medium for 7 days. Afterwards, cells were trypsinized and counted with a Countess cell counter (Invitrogen). In 48-well plates, 2 × 10⁴ cells/well were seeded in OB differentiation medium containing 10 mM β-glycerophosphate and 200 μM ascorbic acid-2-phosphate. Osteoblast differentiation was continued for 12 to 21 days with regular medium changes (3×/wk). Osteoblast differentiation was performed in technical triplicate. After 12 days of differentiation, OB alkaline phosphatase (ALP) activity was determined using p-Nitrophenyl phosphate (pNPP). ALP enzyme activity was analyzed in relation to the protein concentration of each sample.

To visualize the area occupied by OBs in the well, OBs were stained for ALP on day 14 of differentiation. Here, washed cells were fixed with cold PBS-buffered 10% formaldehyde solution and subsequently incubated with ALP staining solution for 15 minutes at room temperature. ALP-positive areas were recorded using an HP Scanjet G4050 (HP, Inc, Wilmington, USA). Subsequently ALP-positive areas were calculated as a percentage of the total well.

Transcriptomics

Total RNA of OCs was obtained when first mature OCs appeared in the culture (day 3–5). Osteoclast cultures derived from 4 mice for each genotype were processed. Due to abnormal principal components analysis (PCA) values, 1 FLT3 ITD and 1 FLT3 ITD PTPRC KO sample was excluded from the data processing. For library preparation, an input of 500 ng total RNA per sample were used. For this purpose, NEBNext Poly(A) mRNA Magnetic Isolation Module, NEBNext Ultra II Directional RNA Library Prep Kit for Illumina, and NEBNext Multiplex Oligos for Illumina (Unique Dual Index UMI Adaptors) were used. Sequencing was performed with a NovaSeq 6000 SP Reagent Kit v1.5 (100 cycles) on a NovaSeq 6000 sequencer with single end sequencing (101 bp). In short, read data were preprocessed to remove adapters and low-quality bases (cutadapt 4.2) and aligned (STAR 2.7.6a) to the mouse reference genome (GRCm38.p6). Afterward, the expression of protein coding genes was quantified (feature counts v2.0.3; Ensembl 100). Read counts are available at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE98667. Downstream analysis was performed in R (4.1.3) for statistical computing and graphics.DESeq2 (1.34) was used to normalized quantified genes by variance stabilizing transformation (VST) for principal component analysis (PCA). The PCA between samples was based on commonly expressed genes with a read count >0 in all samples. Differentially expressed genes (DEGs) between WT and Ptprc^-/-, WT and Flt3^ITD/ITD, and WT and Flt3^ITD/ITD Ptprc^-/- samples were determined using DESeq2¹⁸ and controlled for multiple testing using a false discovery rate (FDR) <0.05 (Table S8). Relative change in expression of a gene in comparison to WT expression was given as log₂ fold-change. Gene Ontology (GO) terms were assessed with clusterProfiler¹⁹ (4.2.2) using 2 different methods to detect overrepresentation of DEGs (Fisher’s exact test). Results were corrected for multiple testing using FDR. In further analysis, genes of Flt3^ITD/ITD Ptprc^-/- samples vs WT were filtered based on specific GO terms using a significance threshold of p < .05. To visualize the expression patterns of these genes across all groups, heat maps were generated. Underlying tables including log₂ fold-changes of DEGs and p values can be found in Table S9. The expression of genes of interest was analyzed by comparing normalized gene expression transcripts per million (TPM) and visualized in box plot graphs.

Statistical analysis

Comparisons between groups were made with ordinary 1-way ANOVA followed by Tukey's multiple-comparisons test (post hoc test) or 2-way ANOVA, as indicated in the Figures 1–6 in GraphPad Prism 9. Differences were considered significant if p <.1. Transcriptomic data were plotted in GraphPad Prism 9 using the statistical evaluation of the DESeq2 (p values and adjusted p-values as indicated in the figure legends).

Results

Abnormal trabecular and cortical bone structure in mice carrying FLT3 ITD and with inactivation of PTPRC

To visualize and investigate bone structural changes in Flt3^ITD/ITD Ptprc^-/- mice, femurs of 12- to 15-week-old WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice were scanned using μCT. Severe structural changes were observed in the longitudinal reconstruction of the femurs of Flt3^ITD/ITD Ptprc^-/- mice (Figure 1A). Cross-sections of the femoral midshaft also showed an abnormal structure with changes, reflecting trabecularization of cortical bone in the diaphysis of these mice (Figure 1B). Due to the disorganized epiphyseal growth plate, it was challenging to precisely differentiate between the metaphyseal and diaphyseal part of the bone. The comparison of the bone parameters allows the quantification of the bone aberrancies of the mutant mice. Trabecular structures traversed not only the metaphysis but also the entire diaphysis. The abnormal trabecular structure in the diaphysis was also observed in Flt3^ITD/ITD mice (Figure 1A). In addition, the number of trabecular structures (Tb.N) in the femoral metaphysis was increased in Flt3^ITD/ITD Ptprc^-/- mice compared with all other 3 genotypes (Figure 1C). The trabeculae were thinner and less separated compared with the control mice, resulting in an overall unaltered ratio of bone volume to total volume (BV/TV) and BMD (Figure 1D, E). The same trends were observed in the lumbar spine (Figure S1E–H). BMD and BV/TV were both increased in the medullary cavity of the diaphysis Flt3^ITD/ITD Ptprc^-/- mice due to the appearance of trabeculae in this area (Figure S1F, G). A nonsignificant increase in BMD was also detected for Flt3^ITD/ITD mice (Figure 1F). In contrast, BMD and BV/TV were reduced in the cortex of the diaphysis as well as the cortical thickness in Flt3^ITD/ITD Ptprc^-/- mice compared with the controls (Figure 1H–J). The reduced cortical bone structure and the more fragile trabecular structure resulted in a reduced Young’s modulus of femurs of Flt3^ITD/ITD Ptprc^-/- mice, measured with a 3-point bending test (Figure S1J). To address the question if bone aberrancies were genus specific, bone parameters were processed for female and male animals separately. As demonstrated in Figure S2, diaphyseal structures did not show significant dissimilarities (Figure S2A–D). In the metaphysis of the femurs, genus-specific alterations observed in WT mice were also observed in Flt3^ITD/ITD mice. In mice inactivated for PTPRC, these diversities were mainly absent (Figure S2E–H). In order to obtain insight into the time course of bone aberrancies, μCT measurements of 2-week-old mice were carried out. At this age, the bone density of femora and sacrum of Flt3^ITD/ITD Ptprc^-/- mice was already substantially reduced compared with control mice (data not shown).

Figure 1

Aberrant bone structures of Flt3^ITD/ITD Ptprc^-/- mice. Bone parameters for 12- to 15-week-old C57BL/6 J (WT), Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice were determined by analysis of μCT scans of the femur. (A) Images of μCT scans of femoral longitudinal sections. (B) Images of μCT scans of diaphysis cross-sections. (C) Trabecular number (Tb.N), (D) trabecular separation (Tb.Sp), and (E) trabecular thickness (Tb.Th) in the femoral metaphysis. (F) Quantification of the BMD and (G) quantification of the ratio of bone volume to total volume (BV/TV) in the diaphysis inside the cortex (BM area). (H) Cortical thickness (C.Th), (I) quantification of the BMD, and (K) quantification of the ratio of bone volume to total volume (BV/TV) of the cortex in the femur diaphysis. (C–J) statistical analysis by 1-way ANOVA followed by post hoc test: ^#p <.1; ^*p <.05; ^**p <.01; ^***p <.001; ^****p <.0001. Mouse numbers used are as follows—(C, D, E) WT: 16; Ptprc^-/-: 18; Flt3^ITD/ITD: 16; and Flt3^ITD/ITD Ptprc^-/-: 15. (F) WT: 6; Ptprc^-/-: 4; Flt3^ITD/ITD: 5; and Flt3^ITD/ITD Ptprc^-/-: 6. (G) WT: 10; Ptprc^-/-: 8; Flt3^ITD/ITD: 9; and Flt3^ITD/ITD Ptprc^-/-: 8. (H, I, J) WT: 16; Ptprc^-/-: 18; Flt3^ITD/ITD: 16; and Flt3^ITD/ITD Ptprc^-/-: 20.

In addition, the growth plate appearance was altered. In addition to the shorter femur length of Flt3^ITD/ITD Ptprc^-/- mice (Figure S1K), the epiphyseal plate of the femur was disrupted compared with the femurs of WT, Ptprc^-/-, and Flt3^ITD/ITD mice (Figure S1L, M). In H&E-stained femoral sections, the palisades of the epiphyseal plate were less organized in Flt3^ITD/ITD Ptprc^-/- mice. Together with the structural alterations and the significantly changed bone parameters, these results demonstrate the severity of the bone phenotype in mice carrying the knock-in of Flt3 ITD and the knockout of Ptprc.

FLT3 ITD and lack of PTPRC lead to altered bone cell numbers and distribution in trabecular and cortical bone

To understand the cellular changes underlying the observed bone phenotype in Flt3^ITD/ITD Ptprc^-/- mice, the abundance of bone cells per surface—namely, OCs, OBs, osteocytes, and adipocytes—was analyzed in decalcified, TRAP-stained femur sections from 12- to 15-week-old WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice. Trabeculae of the metaphysis showed the presence of TRAP-positive OCs, OBs, osteocytes, and adipocytes (Figure 2A). TRAP-positive OCs in Flt3^ITD/ITD Ptprc^-/- mice were profoundly reduced compared with all other genotypes (Figure 2B). Osteoclast numbers were also reduced in Ptprc^-/- mice to a lesser extent compared with WT mice. In the metaphysis of Flt3^ITD/ITD Ptprc^-/- femurs, OBs were diminished compared with controls (Figure 2C). Parameters of active bone formation evaluated using dual calcein labeling showed no abnormalities (Figure S3). Adipocytes were almost completely absent in the Flt3^ITD/ITD Ptprc^-/- metaphysis and strongly reduced in Flt3^ITD/ITD and Ptprc^-/- mice (Figure 2D).

Figure 2

Aberrant bone cell numbers in Flt3^ITD/ITD Ptprc^-/- mice. (A) Tartrate-resistant acid phosphatase (TRAP) staining of femur sections of the cancellous bone in the metaphysis at 40× magnification of the indicated mouse genotypes of 12- to 15-week-old mice. Pink arrows indicate TRAP-positive osteoclasts (OCs), black arrows indicate osteoblasts (OBs), and white arrows indicate adipocytes. Black scale bar: 40 μm. (B–D) Histomorphometric quantification of (B) OC number (N.Oc/BS), (C) OB number (N.Ob/BS), and (D) adipocyte number (N.Adipo) in the femoral metaphysis per bone surface. (E–F) Histomorphometric quantification of (E) OC number (N.Oc/BS) and (F) OB number (N.Ob/BS) in the femoral diaphysis per bone surface. (B–F) statistical analysis by 1-way ANOVA followed by post hoc test: ^*p <.05; ^**p <.01; ^***p <.001; ^****p <.0001.

Due to the severe alterations of the bone structure in the diaphysis of Flt3^ITD/ITD Ptprc^-/- mice (Figure 1), diaphyseal bone cell numbers were additionally quantified in cross-sections. While Flt3^ITD/ITD or Ptprc^-/- mice showed no effect on OC numbers in the diaphysis (Figure 2E), OB numbers were elevated in Flt3^ITD/ITD Ptprc^-/- mice (Figure 2F). Osteocyte numbers were not significantly affected by FLT3 ITD or lack of PTPRC in both metaphysis and diaphysis (Figure S1C).

Taken together, while OB and OC numbers were extremely low in the metaphysis of Flt3^ITD/ITD Ptprc^-/- mice, the number of OCs tended to be increased in the diaphyseal bone, potentially correlating the low cortical bone volume. Flt3^ITD/ITD and Ptprc^-/- mice also showed slight alterations in bone cell numbers and bone turnover.

Metabolic bone markers of FLT3 ITD–expressing mice indicate elevated bone resorption

To understand the cellular changes underlying the observed bone phenotype in Flt3^ITD/ITD Ptprc^-/- mice, bone turnover markers reflecting bone resorption and formation were evaluated from serum samples of 12- to 15-week-old WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice. Enhanced levels of bone resorption marker TRAP5b were observed both in Flt3^ITD/ITD as well as Flt3^ITD/ITD Ptprc^-/- mice (Figure 3A). In addition, in FLT3-ITD–expressing mice, the bone resorption marker CTX-I was slightly, but statistically nonsignificantly, elevated (Figure 3B). In addition, P1NP, reflecting bone formation, remained largely comparable in the serum of all 4 genotypes (Figure 3C).

Figure 3

Metabolic bone parameters in Flt3^ITD/ITD Ptprc^-/- mice. Metabolic bone parameter CTX-I concentration (A), TRAP5b activity (B), and P1NP concentration (C) of blood serum derived from 14-week-old WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice were determined. (A–C) Statistical analysis by 1-way ANOVA followed by post hoc test: ^*p <.05; ^**p <.01. Mouse numbers used are as follows—(A) WT: 11; Ptprc^-/-: 14; Flt3^ITD/ITD: 11; and Flt3^ITD/ITD Ptprc^-/-: 9. (B) WT: 7; Ptprc^-/-: 7; Flt3^ITD/ITD: 6; and Flt3^ITD/ITD Ptprc^-/-: 8. (C) WT: 10; Ptprc^-/-: 14; Flt3^ITD/ITD: 12; and Flt3^ITD/ITD Ptprc^-/-: 10.

Impaired OC differentiation and reduced OC size in ex vivo cultures of Flt3^ITD/ITD Ptprc^-/- cells

To understand the impact of FLT3 ITD and PTPRC inactivation on OC function, CD11b-positive BM cells were isolated from 12- to 15-week-old WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice and similar cell numbers were differentiated into OCs ex vivo. The differentiation capacity was assessed by imaging TRAP-stained OC cultures, examining the TRAP activity in the cell supernatants of OC cultures, and measuring OC size. To assess the effects of FLT3 ITD and inactivation of PTPRC on osteoclastogenic gene expression, mRNA was analyzed by next-generation sequencing (NGS).

Differentiated WT and Ptprc^-/- OCs exhibited similar mature morphologies, whereas OCs derived from Flt3^ITD/ITD and Flt3^ITD/ITD Ptprc^-/- cultures appeared smaller and less mature (Figure 4A). Wild-type and Ptprc^-/- OC cultures showed mostly mature, multinucleated, TRAP-positive cells with spread cytoplasm and big rounded shapes (indicated by black arrows). Immature, multinucleated, TRAP-positive OCs without extended size or with visible cell processes were also observed in cultures of all 4 genotypes (indicated by white arrows). A trend towards a reduction in mature and immature OC numbers was detected for Flt3^ITD/ITD Ptprc^-/- cultures (Figure S4A). To quantify the morphological differences, the size of mature OCs was measured in ImageJ. Osteoclasts of Flt3^ITD/ITD and Flt3^ITD/ITD Ptprc^-/- genotypes were smaller compared with WT and Ptprc^-/- OCs (Figure 4B). Moreover, TRAP activity in cell supernatants of OC cultures was measured showing lower TRAP activity in supernatants of Flt3^ITD/ITD Ptprc^-/- OC cultures (Figure 4C).

Figure 4

Reduced osteoclast (OC) differentiation capacity of Flt3^ITD/ITD Ptprc^-/- derived cells. CD11b-positive BM cells were isolated from 12- to 14-week-old WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice, as indicated. After a 3-day expansion phase of myeloid progenitors in the presence of M-CSF (50 ng/mL), differentiation to OCs was induced for 3 to 5 days in the presence of M-CSF (50 ng/mL) and RANKL (20 ng/mL). (A) Images of tartrate-resistant acid phosphatase (TRAP)–stained OCs on the day of maximum differentiation. Black arrows indicate mature OCs with spread cytoplasm and white arrows indicate TRAP-positive multinucleated immature OCs with small, rounded shape or visible cell processes. Black scale bars: 400 μm. (B) OC size was calculated in pixelssquared. OC size was accessed with multipoint tool and area measurement in ImageJ. Data are based on means of 20 mature OCs per biological replicate. (C) TRAP activity of cell supernatants is shown as absorbance at 405 nm. Cell supernatants of OC cultures were harvested on day 5 of differentiation, after an incubation period of 48 hours. (D) Heat map of relative gene expression of differentiated OCs. OC mRNA was isolated as the first mature OCs appeared and the transcriptome was sequenced. Differentially expressed genes (DEGs; identified by DESeq2; p < .05) of Flt3^ITD/ITD Ptprc^-/- OC mRNA were filtered for Gene Ontology (GO) term “bone remodeling”. Log₂ fold-changes to WT of these genes were plotted for all genotypes (Ptprc^-/-, Flt3^ITD/ITD, Flt3^ITD/ITD Ptprc^-/-). (B, C) Statistical analysis by 1-way ANOVA followed by post hoc test: ^*p <.05; ^**p <.01; ^***p <.001. (D) WT, Ptprc^-/-: n = 4; Flt3^ITD/ITD, Flt3^ITD/ITD Ptprc^-/-: n = 3.

Furthermore, gene expression analysis revealed differential regulation of OC-specific GO terms. Genes of Flt3^ITD/ITD Ptprc^-/- OCs vs WT were filtered based on these GO terms using a significance threshold of p <.05. To visualize the expression patterns of these genes across all groups, heat maps were generated (Figure 4D) for GO terms “bone remodeling” (0046849; Figure 4D, Table S3) and “bone resorption” (0045453; Figure S5A, Table S4). The GO term “osteoclast differentiation” showed differential expression close to significance (GO: 0030316; Figure S5B, Table S5). Remarkably, most of the genes associated with these OC-related GO terms were downregulated in differentiated OCs derived from Flt3^ITD/ITD Ptprc^-/- mice. Specific genes involved in the OC formation, such as Oscar, Ocstamp, and Dcstamp, showed reduced expression (Figure S5C–E). In addition, gene expression of OC markers like Mmp9, Acp5 (TRAP), and Ctsk (cathepsin K) was strongly decreased (Figure S5F–L). Thus, the gene expression of Flt3^ITD/ITD Ptprc^-/- OCs showed a strong reduction for genes involved in OC differentiation.

Increased bone resorption activity of Flt3^ITD/ITD Ptprc^-/- OC cultures

By analyzing pit formation and quantifying CTX-I in the cell supernatants, effects of FLT3 ITD and inactivation of PTPRC on the resorptive activity of OCs were studied. CD11b-positive BM cells were differentiated into OCs on bovine bone slices and resorption pits were examined by scanning electron microscopy. Resorption was detected in the OC cultures of WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice showing comparable lacunae with varying depths and detail structures, yet no distinctions were observed between OCs derived from all genotypes (Figure 5A, B).

Figure 5

Bone-resorption capacity of differentiated osteoclasts (OCs). CD11b-positive BM cells were isolated from 12- to 14-week-old WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice and seeded on bovine bone slices. After a 3-day expansion phase of myeloid progenitors in the presence of M-CSF (50 ng/mL), differentiation into OCs was initiated for 10 days in the presence of M-CSF (50 ng/mL) and RANKL (20 ng/mL). (A) Scanning electron microscopy of cleaned bone slices after 10 days of OC differentiation; 200× magnification. Black scale bars: 100 μm. (B) Details of resorption lacunae by scanning electron microscopy of cleaned bone slices after 10 days of OC differentiation; 2000× magnification. White scale bars: 10 μm. (C) Cumulative CTX-I concentration of OC cell supernatants. Cell supernatants were collected on day 10 of differentiation, after an accumulation period of 7 days, and CTX-I concentration was measured by ELISA. Statistical analysis by 1-way ANOVA followed by post hoc test: ^*p <.05; ^**p <.01.

To quantify the resorption activity of the OC cultures, the CTX-I concentration of the cell supernatant was measured. The analysis included the cumulative CTX-I release from day 3 to day 10 of differentiation. CTX-I levels were significantly elevated in Flt3^ITD/ITD Ptprc^-/- OC cultures (Figure 5C). In addition, a tendency to increased CTX-I levels was found in Ptprc^-/- OC cultures. The formation of lacunae and the increased CTX-I concentrations suggest a similar or even higher resorption activity of Flt3^ITD/ITD Ptprc^-/- OCs.

FLT3 ITD activity increased proliferation in OC cultures

To understand how impaired OC differentiation and elevated cumulative resorption of Flt3^ITD/ITD Ptprc^-/- cells are in tune with each other, differentially regulated genes were viewed again. The gene expression profile of Flt3^ITD/ITD Ptprc^-/- OCs showed enrichment for several GO terms related to cell proliferation (data not shown). Genes of Flt3^ITD/ITD Ptprc^-/- OCs vs WT were filtered for the GO terms “DNA replication” or “mitosis” with a significance threshold of p < .05. Notably, most of the genes associated with proliferation-related GO terms were upregulated in Flt3^ITD/ITD Ptprc^-/- OCs (Figure 6A). Among these DEGs, the expressions of the antiapoptotic B-cell lymphoma 2 (Bcl2) and enkurin domain containing 1 (Enkd1), which is part of the mitotic spindle apparatus,²⁰ were greatly increased (Figure S5I, J). Furthermore, significant increases in gene expression were found in Flt3^ITD/ITD Ptprc^-/- OCs for replication fork stabilizing Donson and origin recognition complex subunit 1 (Orc1) (Figure S5K, L).

Figure 6

Increased proliferation of FLT3 ITD–positive osteoclast (OC) cultures. CD11b-positive BM cells were isolated from 12- to 14-week-old WT, Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- mice. After a 3-day expansion phase of myeloid progenitors in the presence of M-CSF (50 ng/mL), differentiation into OCs was initiated in the presence of M-CSF (50 ng/mL) and RANKL (20 ng/mL) for 3 to 5 days. (A, F) Heat maps of relative gene expression of differentiated OCs. OC mRNA was isolated as the first mature OCs appeared and the transcriptome was sequenced. Differentially expressed genes (DEGs; identified by DESeq2; p < .05) of Flt3^ITD/ITD Ptprc^-/- OC mRNA were filtered for Gene Ontology (GO) terms (A) “DNA replication” and (B) “mitosis”. Log₂ fold-changes to WT of these genes were plotted for all genotypes (Ptprc^-/-, Flt3^ITD/ITD, Flt3^ITD/ITD Ptprc^-/-). Gene symbols for GO DNA replication are listed in Table S6. (B) Absorption values of the phenazine methosulfate (MTS) assay reflecting proliferative capacity during the myeloid expansion phase after 3 days of M-CSF treatment. (C) Absorption values of the MTS assay reflecting proliferative capacity during OC differentiation after M-CSF and RANKL treatment at indicated time points. (D) MTS assay reflecting proliferative capacity during myeloid expansion phase after 3 days of M-CSF treatment in the presence of FLT3 inhibitor AC220. Data are shown as fold-changes of AC220-treated to DMSO-treated samples (AC220 normalized). (E) MTS assay reflecting proliferative capacity during OC differentiation after M-CSF and RANKL treatment at indicated time points in the presence of FLT3 inhibitor AC220. Data are shown as fold-changes of AC220-treated to DMSO-treated samples (AC220 norm.). (F) Heat map of relative gene expression of the indicated STAT5 targets of Ptprc^-/-, Flt3^ITD/ITD, and Flt3^ITD/ITD Ptprc^-/- OCs as log₂ fold-change to WT OC expression. OC mRNA was isolated as the first mature OCs appeared and the transcriptome was sequenced. Numbers in boxes represent the adjusted p-values (p < .4) of the DESeq2 analysis. (A, B, F) WT, Ptprc^-/-: n = 4; Flt3^ITD/ITD, Flt3^ITD/ITD Ptprc^-/-: n = 3. (B–E) Statistical analysis by 1-way ANOVA followed by post hoc test: ^*p <.05; ^**p <.01; ^***p <.001; ^****p <.0001.

Thus, to investigate whether the presence of FLT3 ITD and the absence of PTPRC affect the proliferation of CD11b-positive BM cells and OC precursors, we analyzed cell proliferation in OC cultures at different stages of the differentiation process using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (MTS). Increased absorbance at 492 nm, and thus cell growth, was observed after 3 days of myeloid expansion in cultures of FLT3 ITD–expressing cells compared with WT and Ptprc^-/- cells (Figure 6B). After 3 days of differentiation, proliferation was still significantly increased in Flt3^ITD/ITD Ptprc^-/- OC cultures (Figure 6C). After 5 and 7 days of OC differentiation, no differences in absorbance and thus proliferation were observed. To test whether the observed increase in proliferative capacity in FLT3 ITD–expressing OC cultures was due to FLT3 ITD activity, cells were treated with AC220 (20 nM) during expansion of myeloid progenitors and OC differentiation. Absorbance values of AC220-treated cells were compared with DMSO-treated cells. A trend toward a decrease in relative absorbance was observed during OC differentiation in Flt3^ITD/ITD Ptprc^-/- cultures (Figure 6D, E). Being only a trend during the expansion phase (Figure 6D), the decrease becomes more pronounced at the beginning of the differentiation phase in Flt3^ITD/ITD Ptprc^-/- cultures, indicating reduced proliferation (Figure 6E). At day 5 after induction of OC differentiation, the cultures of Flt3^ITD/ITD Ptprc^-/- OCs still showed decreased absorbance values and thus reduced proliferation compared with WT, Ptprc^-/-, and Flt3^ITD/ITD cultures. Thus, inhibition of FLT3 ITD significantly affected proliferation in Flt3^ITD/ITD Ptprc^-/- OC cultures. Consequently, FLT3 ITD drives proliferation during the myeloid expansion phase and early OC differentiation.

Considering the enrichment of proliferation-related genes and higher proliferation activity, especially in cultures of FLT3 ITD–expressing OCs, specific activation of STAT5 target genes was addressed in the gene expression analysis. Predominantly, increased gene expression of STAT5 targets was observed in Flt3^ITD/ITD Ptprc^-/- OCs (Figure 6F). Socs3, Ccnd1 (cyclin D1), and Cdc25a showed significant upregulation, and other STAT5 targets, such as Il2ra, Pim1, Ccnd2 (cyclin D2), and Myc (c-Myc), also showed increased expression, although this was not statistically significant. Thus, data indicate that elevated growth rates of Flt3^ITD/ITD Ptprc^-/- OC cultures is mediated by FLT3 ITD–driven STAT5 activity.

FLT3 ITD and lack of PTPRC do not alter osteoblast activities

To characterize activities of OBs, mesenchymal stromal cells were purified from femurs, tibiae, and humeri and subsequently differentiated in the presence of ascorbic acid and β-glycerophosphate. This ex vivo differentiation showed similar OB formation of cells derived from all genotypes (Figure S6A). In addition, ALP activity remained similar (Figure S6B, C).

Discussion

In our previous studies, we demonstrated that inactivation of PTPRC in Flt3^ITD/ITD mice resulted in hematologic abnormalities, including severe monocytosis. Thus, the enhanced concentration of myeloid precursor cells might also translate to increased numbers of monocytic OC precursors, which could affect bone metabolism. Thus, we already previously showed structural alterations of long bones with increased numbers of thin trabeculae in the metaphysis as well as trabeculae traversing the entire diaphysis.²¹ Extension of the metaphysis into the diaphysis was also observed in Flt3^ITD/ITD mice and Ptprc^-/- mice, but to a lesser extent than in Flt3^ITD/ITD Ptprc^-/- mice. The presence of trabecular structures in the diaphysis was already described in Ptprc KO mice and interpreted as osteopetrosis.¹² Trabecularization of cortical bone can be attributed to increased abundance of OBs and OCs at the femoral diaphysis.^22,23 In contrast to the general understanding of osteopetrosis, cortical BMD and bone volume and thickness were reduced in Flt3^ITD/ITD Ptprc^-/- mice. Furthermore, no changes in the bone formation rate as well as serum concentration of classical bone turnover markers CTX-I and P1NP was observed in Flt3^ITD/ITD Ptprc^-/- mice compared with the controls, indicating no systemic osteopetrotic phenotype.

Along with this, severely reduced OC numbers in the metaphysis of Flt3^ITD/ITD Ptprc^-/- mice and impaired ex vivo OC differentiation of Flt3^ITD/ITD Ptprc^-/- cells were observed. As the size of mature OCs from Flt3^ITD/ITD and Flt3^ITD/ITD Ptprc^-/- mice was reduced compared with WT and Ptprc^-/- OCs, this suggests that the phenotype is more pronounced when FLT3 ITD activity is further increased by knocking out Ptprc. Interestingly, in the femoral diaphysis, the number of OCs in the Flt3^ITD/ITD Ptprc^-/- mice remained unchanged. This indicates that the origin of bone alterations could start from the metaphysis.

Furthermore, the greatest differences in the expression of OC-specific genes were detected in Flt3^ITD/ITD Ptprc^-/- OCs vs WT OCs compared with the comparisons of Ptprc^-/- and Flt3^ITD/ITD OCs with WT OCs. In addition to the reduced OC formation, the smaller OC size and lower gene expression of Ocstamp and Dcstamp indicate an OC fusion defect in Flt3^ITD/ITD Ptprc^-/- mice. Expression of both genes is essential for fusion of mononucleated TRAP-positive cells into fully functional multinucleated OCs. Ocstamp as well as Dcstamp deficiency completely abolished OC fusion.^24–26 Due to the reduced expression of OC fusion genes like Ocstamp and Dcstamp, it is likely that these cells, while functional in their bone-resorbing activity, are unable to fuse into fully mature giant OCs in Flt3^ITD/ITD Ptprc^-/- mice. In combination with the increased proliferation of Flt3^ITD/ITD Ptprc^-/- OC cultures and thus higher cell numbers, this might explain the slightly increased bone resorption activity of Flt3^ITD/ITD Ptprc^-/- OC cultures. This hypothesis could also explain the increased TRAP5b activity in the blood serum of Flt3^ITD/ITD Ptprc^-/- mice. To study the potential of OC precursor cells to differentiate into mature OCs in the OC culture experiments, similar numbers of CD11b-positive OC precursors were seeded. Flt3^ITD/ITD and Flt3^ITD/ITD Ptprc^-/- mice showed approximately 2-fold elevated numbers of CD11b-positive BM cells compared with the WT and Ptprc^-/-mice (data not shown). Due to the higher number of hematopoietic stem and progenitor cells in Flt3^ITD/ITD Ptprc^-/- mice,¹⁰ it is likely that more OC progenitors were produced that express and secrete TRAP5b. Furthermore, it has been reported that mononuclear TRAP-positive cells are capable of resorbing bone.^27,28 Thus, the increased bone resorption activity of Flt3^ITD/ITD Ptprc^-/- OCs compared with WT, Ptprc^-/-, and Flt3^ITD/ITD OCs might be due to increased proliferation of hematopoietic stem and progenitor cells that start to differentiate but failed to mature into giant multinucleated OCs. However, the smaller OCs could resorb bone, resulting in an increased CTX-I release from Flt3^ITD/ITD Ptprc^-/- OCs. In addition, the reduced size of OCs derived from Flt3^ITD/ITD Ptprc^-/- mice could be also partially due to a spatial competition during OC formation.

FLT3 ITD drives leukemic cell transformation of hematopoietic cells and proliferation via activation of STAT5.^7,22,29 In this study, the presence of FLT3 ITD led to an increase in proliferation and gene expression of proliferation-related GO terms. To test the hypothesis that the increased proliferation is caused by FLT3 ITD activity, OC cultures were treated with the selective FLT3 inhibitor AC220. Here, proliferation was reduced in Flt3^ITD/ITD Ptprc^-/- OC cultures, indicating that FLT3 ITD is the driver of the increased proliferation. This was further validated by increased gene expression of proliferation-promoting STAT5 targets like Ccnd1/2 (cyclin D1/2), Pim1, Cdc25a, and Myc. Thus, it is likely that the oncogenic FLT3 ITD signaling inhibited OC maturation and increased proliferation of myeloid progenitors and OC precursors via its downstream effector STAT5, resulting in the observed impaired OC differentiation.

Studies focusing on OB activity and differentiation revealed that expression of oncogenic FLT3 ITD and inactivation of PTPRC resulted in minor alterations compared with control mice. Since expression of FLT3 ITD as well as PTPRC could be predominantly detected in HSC-derived cell populations and was virtually absent in mesenchymal cells, it can be concluded that the above-described aberrancies in the formation of the osteohematopoietic niche are mainly due alterations of OC abundance, activity, and differentiation as well as the previously described hematopoietic phenotype.

Acknowledgments

The authors thank Ivonne Görlich and Cornelia Luge for excellent technical assistance realizing next-generation sequencing as well as Claudia Waskow for scientific advice and discussions. They also thank Sandra Hippauf, Ina Gloe, and Tina Dybek for excellent technical assistance with μCT analyses and histology.

Author contributions

Carolin Lossius-Cott (Data curation, Investigation, Methodology, Writing—original draft), Akua Annoh (Investigation, Methodology), Martin Bens (Investigation, Validation, Visualization), Sandor Nietzsche (Conceptualization, Investigation, Methodology), Bianca Hoffmann (Data curation, Investigation, Methodology, Software), Marc Thilo Figge (Conceptualization, Formal analysis, Methodology, Validation), Martina Rauner (Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Writing—original draft), Lorenz C. Hofbauer (Conceptualization, Formal analysis, Funding acquisition, Methodology), and Jörg P. Müller (Conceptualization, Formal analysis, Project administration, Resources, Validation, Writing—original draft).

Funding

This work was supported by the Deutsche Forschungsgemeinschaft, grant Mu955/14-1 and 15-1 (to J.P.M.), Ho 1875/28-1 (to L.C.H.), and DFG-FOR 5146 (to L.C.H. and M.R.) and by the Leibniz ScienceCampus InfectoOptics, Jena (to M. T. F.), which is financed by the funding line Strategic Networking of the Leibniz Association (to M.T.F.).

Conflicts of interest

None declared.

Data availability

The data underlying this article are available in the article and in its online supplementary material. RNA-seq data have been deposited in Gene Expression Omnibus (GEO) database and are accessible through accession number GSE273708 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE273708).

Author notes

Martina Rauner, Lorenz C. Hofbauer and Jörg P. Müller contributed equally to the manucript.