Specification of Sprouty2 functions in osteogenesis in in vivo context

ABSTRACT

Sprouty proteins are modulators of the MAPK/ERK pathway. Amongst these, Sprouty2 (SPRY2) has been investigated as a possible factor that takes part in the initial phases of osteogenesis. However, the in vivo context has not yet been investigated and the underlying mechanisms taking place in vitro remain unknown. Therefore, in this study, the impact of Spry2 deficiency was examined in the developing tibias of Spry2 deficient (-/-) mouse. The investigation was performed when the osteogenic zone became clearly visible and when all three basic bone cells types were present. The main markers of osteoblasts, osteocytes and osteoclasts were evaluated by immunohistochemistry and RT-PCR. RT-PCR showed that the expression of Sost was 3.5 times higher in Spry2-/- than in the wild-type bone, which pointed to a still unknown mechanism of action of SPRY2 on the differentiation of osteocytes. The up-regulation of Sost was independent of Hif-1α expression and could not be related to its positive regulator, Runx2, since none of these factors showed an increased expression in the bone of Spry2-/- mice. Regarding the RANK/RANKL/OPG pathway, the Spry2-/- showed an increased expression of Rank, but no significant change in the expression of Rankl and Opg. Thanks to these results, the impact of Spry2 deletion is shown for the first time in the developing bone as a complex organ including, particularly, an effect on osteoblasts (Runx2) and osteocytes (Sost). This might explain the previously reported decrease in bone formation in postnatal Spry2-/- mice.

INTRODUCTION

Sprouty (Spry) genes encode cytoplasmic membrane-associated proteins. Initially described in Drosophila, the SPRY family of proteins is highly conserved. Four mammalian orthologs (SPRY1-4) exist, which modulate receptor tyrosine kinase (RTK) signaling pathways, although variations of their expression were observed, depending on cell types.^1,2

Transgenic mice lacking SPRY and SPRED (Sprouty-related, EVH1 domain-containing proteins) family proteins show abnormalities in endochondral (long) bones including a decreased bone growth (achondroplasia-like phenotype) and decreased trabecular bone mass.^3–5 Endochondral bones, unlike the intramembranous ones, involve a chondrogenic developmental step. The developing growth plate may be divided into three compartments: an external fibrous tissue surrounding the periphery, a cartilaginous, and a bony tissue. The cartilaginous part consists of several zones, where chondrocytes differentiate towards the diaphysis: the resting zone, where a pool of less differentiated chondroblasts is localized, proliferating zone, where chondroblasts multiply and further differentiate, and zone of hypertrophy undergoing calcification, degradation and replacement by osseous tissue. The calcifying matrix of cartilage starts to be removed by osteoclasts/chondroclasts in the line of erosion, and the formation of the primary spongiosa involves a cooperation between osteoblasts and osteoclasts in the zone of ossification.⁶

Chondrogenesis in long bones was reported to be impaired in Spry2 deficient (-/-) mice: chondrocytes undergo increased cell proliferation, but their terminal differentiation is impaired in the femur at embryonic (E) day 18.5.⁵ Further, using an in vitro approach, Spry2 overexpression in osteoblastic lineage (MC3T3-E1) resulted in abnormal cell proliferation and differentiation.⁷ However, interpretations of the in vitro studies were hindered by the fact that proper bone development in vivo requires a tight molecular cross-talk networks among osteoblasts, osteocytes and osteoclasts.

To follow the consequences of Spry2 deficiency in vivo on the osteogenic cells (osteoblasts, osteocytes and osteoclasts), the prenatal growth plate within the tibia was examined in the Spry2-/- and compared with that from wild-type mice. Attention was focused on initial stages of osteogenesis when the osteogenic zone became apparent in the forming growth plate at prenatal day E18. Since several bone cells interact and play complementary integrated roles in this area, major markers of osteoblasts (early: RUNX2, and late: osteocalcin – BGLAP), osteocytes (sclerostin – SOST) and osteoclasts (receptor activator of nuclear factor κB - RANK) were investigated. To get an overview on the RANK–RANKL–OPG system, RANK and OPG were analyzed simultaneously. All factors were investigated at the level of transcripts (RT-PCR) and proteins (immunohistochemistry).

This investigation, for the first time in an in vivo context, aimed to specify the possible impact of Spry2 deletion on bone-forming cells during endochondral bone development.

MATERIAL AND METHODS

Animals

Mice (Mus musculus) of the ICR strain were studied at prenatal stage 18. The initial breeding pairs of the mouse line carrying mutant alleles of Spry2-/- were established by Shim et al.⁸ and were kindly provided by Dr. Ophir Klein.⁹ Mice were maintained in the animal facilities of the Institute of Experimental Medicine, The Czech Academy of Sciences, Prague, Czech Republic. Housing and sacrifice of animals were carried out in strict accordance with the national and international guidelines (ID 39/2009). This study was performed under the supervision of a professional committee for the guarantee of good life-conditions of experimental animals at the Institute of Experimental Medicine, Czech Academy of Sciences, Czech Republic (permit number 81/2017).

The pregnant mice were sacrificed by cervical dislocation and the fetuses were removed at E18. Immediately after removing the embryos, their wet body weights were measured in order to precisely determine their developmental stage.¹⁰

Histology, immunohistochemistry (IHC) and immunofluorescence (IF)

Mouse hind limbs were fixed in 4% PFA (24 h), decalcified in EDTA, dehydrated in alcohol, treated with xylen and embedded in paraffin. Histological sections (5 µm) were then prepared. Morphological/histological criteria were analyzed after trichrome staining (hematoxylin, alcian blue, sirius red) and TRAP staining, as previously described.¹¹

Immunohistochemical detection was performed to show the expression of markers and better specify cell heterogeneity within the developing bone. Antigen retrieval was carried out in citrate buffer (pH = 6.0) for 5 min at 98°C for all markers. The following primary antibodies were applied: SPRY2 (ab50317, abcam, UK), BGLAP (ab93876, abcam, UK), SOST (AF1589, R&D Systems, USA), RANK (ab13918, abcam, UK), RANKL (sc-7628, Santa Cruz Biotechnology, USA), OPG (sc-8468, Santa Cruz Biotechnology, USA), RUNX2 (sc-10758, Santa Cruz Biotechnology, USA). The primary antibodies were diluted 1:50 (RANKL, OPG, RUNX2), 1:100 (SOST, SPRY2), or 1:150 (BGLAP, RANK) and applied overnight at 4°C.

Peroxidase-conjugated streptavidin-biotin system (Vectastain, USA) and chromogen substrate diaminobenzidine (DAB, K3466; Dako, USA) reactions were used to detect the positive cells as brown. Slides were counterstained with hematoxylin. SPRY2 antibody was tested in Spry2-/- and wild-type sections (Suppl. 3).

Immunofluorescent analysis was also performed for all the markers listed above. Antigens retrieval and dilutions of antibodies were performed as mentioned for IHC. Then Alexa Fluor® 488 (Thermo Fischer, USA) or Alexa Fluor® 594 (Abcam, UK) was used (1:200) for 40 min at RT. Nuclei were detected by ProLong® Gold Antifade reagent with DAPI (Thermo Fischer, USA).

Sost-positive osteocytes/TRAP-positive osteoclasts were counted in Spry2-/- (3/5 animals) and wild-type group (3/5 animals), respectively. Cells were counted in one medial section of tibia for each sample at the same magnification. The number of cells and the total area involved in the analysis were counted/measured using Image J. Results were expressed as total number of cells/total area in each analyzed sample. For the statistical analysis, Mann–Whitney (p ≤ 0.05) test was applied.

Tissue separation

The region of interest (shown in Suppl. 1) was processed as follows: the tibia (E18) in Spry2-/- mice and their wild-type littermates were separated from surrounding soft tissue using a stereoscope. The cartilaginous part was removed and the resting bone part, including the osteogenic part (bony component) of the growth plate and diaphysis, was used for analysis. The separation of fresh tibia was carried out in sterile PBS. One sample for RNA isolation contained two tibias.

RT-PCR

RNA was isolated by RNeasy Mini Kit (Qiagen, USA) and transcribed into cDNA using SuperScript VILO (Invitrogen, USA). RT-PCR was performed in 10 μl reaction volume containing the one-step master mix gb Ideal PCR Master Mix (Generi Biotech, Czech Republic) using LightCycler 96 (Roche, Switzerland) with preheating at 95°C for 10 min, followed by 40 cycles of 95°C/15 s and 62°C/1 min. The following markers were examined: Bglap, Sost, Rankl, Rank, Opg, Runx2, Spry2, Hif-1α (TaqMan Gene Expression Assay, Thermo Fischer Scientific, USA). Expression levels were calculated using the ∆∆CT method, with normalization against actin levels (Mouse Actb, Mm02619580_g1, TaqMan Gene Expression Assay, Thermo Fischer Scientific, USA). Each group was analyzed in three biological replicates. Reactions were conducted in three replicates per sample for statistical analysis (Student's t-test, p ≤ 0.05). RT-PCR confirmed that the Spry2 mRNA, detected in the wild-type tibia, was indeed absent in the Spry2-/- samples.

RESULTS

The osteogenic part of the tibial growth plate at prenatal day 18 comprises osteoblasts, osteocytes and osteoclasts

The presence of distinct bone cells was demonstrated based on morphological criteria, immunohistochemical detection of specific markers and TRAP-analysis. Osteoblasts showed a cuboidal shape with single round nuclei and were positive for osteocalcin (Fig. 1A). Osteocytes were defined as entrapped in osteocytic lacunae and expressing sclerostin (Fig. 1B). Finally, osteoclasts were multinucleated cells, positive for TRAP staining (Fig. 1C).

FIGURE 1.

Immunohistochemical/immunofluorescent identification of osteogenic markers and SPRY2 in wild-type tibia at E18.

Localization of osteoblastic cells identified as cuboidal cells with a single nucleus and expressing BGLAP (A), osteocytes entrapped in bone matrix and expressing SOST (B), and osteoclasts identified as large multinuclear cells positive for TRAP staining (C). Detection of SPRY2 in growth plate of tibia (D, E), with focus on resting, proliferating (F) and hypertrophic cartilage together with osteogenic zone (G). BGLAP (H) was localized in osteoblastic (H₁), and osteocytic cells (H₂), SOST (I) in osteocytes (I₁). In growth plate, RANK (J) was detected in osteoclastic cells (J₁), RANKL (K) in osteoblastic cells (K₁), OPG (L) in osteoblastic cells (L₁), and RUNX2 (M) in osteoblastic (M₁) and osteocytic cells (M₂). HZ (hypertrophic zone of cartilage), OZ (osteogenic zone), PB (periosteal bone), PS (primary spongiosa), PZ (proliferating zone of cartilage), RZ (resting zone of cartilage). For immunofluorescent labeling, positive signal is green, autofluorescent erythrocytes are orange in figures E-I, K-M. Positive signal is red, autofluorescent erythrocytes are green in figure J. Yellow arrowheads point to osteoblasts, green ones to osteocytes, and black ones to osteoclasts. Scale bar = 30 µm (A-C), 100 µm (E-M), 50 µm (H₁-M₁, H₂, M₂).

SPRY2 is abundantly expressed in all cell types from the tibia at prenatal day 18

In wild-type mice, intense staining for SPRY2 protein was detected in the osteogenic part and other zones of the growth plate at E18 (Fig. 1D–G): in osteoblasts, osteocytes and even in some osteoclasts (Fig. 1D). SPRY2 was also detected in the periosteum (Fig. 1E) and in the cytoplasm of chondrocytes within the resting, proliferating (Fig. 1F), and hypertrophic zones of the cartilage (Fig. 1G).

Osteogenic markers are expressed in the growth plate at prenatal day 18

In the osteogenic zone within the wild-type growth plate at E18, osteoblasts abundantly expressed BGLAP (Fig. 1H, H₁), RUNX2 (Fig. 1M, M₁), and some were even positive for RANKL (Fig. 1K, K₁) and OPG (Fig. 1L, L₁). Osteocytes were strongly positive for SOST (Fig. 1I, I₁), some were also positive for RUNX2 (Fig. 1M, M₂) and BGLAP (Fig. 1H, H₂). Osteoclasts, and mostly their precursors, were positive for RANK (Fig. 1J, J₁).

Spry2 deficiency impacts the expression of osteogenic markers

In order to find out whether Spry2 deficiency could impact osteogenic pathways in the developing tibia, an RT-PCR-based gene expression was analysis was performed. RT-PCR, performed on the osteogenic part of tibias of the Spry2-/- and wild-type mice at E18, revealed statistically significant changes in the expressions of Sost, Runx2 and Rank (Fig. 2B, C, D). The expression of Sost mRNA increased by almost 250% (p = .004674) (Fig. 2B) in Spry2-/- samples, while the expression of Rank increased by 80% (p = .042324) (Fig. 2D). Conversely, a slight decrease in the expression of Runx2 was observed in Spry2-/- (30%, p = .011147) (Fig. 2C). Spry2-/- did not show statistically significant changes in the expression of osteocalcin (Bglap), Rankl, Opg (Fig. 2A, E, F) and Hif-1α (data not shown). These data are available expressed also in –ddCT values for individual animals (Suppl. 2).

FIGURE 2.

Comparison of tibia in Spry2-/- and wt at E18

Expression of osteocalcin (A), sclerostin (B), Runx2 (C), Rank (D), Rankl (E), Opg (F), a tentative scheme of the molecular background of bone phenotype in Spry2-/- bones (G). Asterisk, p ≤ .05. Immunofluorescent detection of SOST (H, I) and RANK (J, K) in wild-type and Spry2-/- tibias, histology of the growth plate was shown after trichrome staining (L, N) and osteoclastic cells were detected by TRAP assay (M, O). PB (periosteal bone). For immunofluorescent labeling, positive signal is green, autofluorescent erythrocytes are orange in figures H, I. Positive signal is red, autofluorescent erythrocytes are green in figure J, K. Scale bar = 100 µm.

The significantly altered expression of mRNA levels was examined also at protein level in the histological sections. While there was a visibly higher number of cells positive for RANK, mostly in peritrabecular regions (Fig. 2J, K), the distribution of SOST-positive cells did not seem to differ when comparing samples from Spry2-/- and wild types (Fig. 2H, I). Quantification of SOST-positive cells in immunolabelled samples revealed slightly increased number without statistical significance in Spry2-/- mice (Suppl. 3D). The slight decrease in Runx2 expression, as detected at mRNA level, was not observed at the protein level (data not shown).

Although the gene expression in Spry2-/- was different from that observed in the wild type, there was no obvious consequence at the bone histological appearance in the mutant, when compared to wild type (Fig. 2L, N). The decrease in TRAP-positive osteoclastic cells in Spry2-/- mice (Fig. 2M, O) appeared as statistically non-significant (Suppl. 3C).

DISCUSSION

Bone development involves its formation including extracellular matrix secretion and mineralization (mediated by osteoblasts), and resorption (mediated by osteoclasts). The differentiation of both cell types is linked and their functions are tuned by osteocytes.¹² In endochondral (long) bones, osteogenesis includes a cartilaginous step. Joo et al.⁵ have reported alterations in long bone development in Spry2-/- mice, which were related to growth plate chondrocytes next to the zone of ossification. Since Spry2 expression precedes the functional differentiation of osteoblasts, an effect on early osteogenesis was suspected.^13,14 Therefore, we searched for early consequences of SPRY2 deficiency in main bone cell types, osteoblasts, osteocytes and osteoclasts, during the prenatal formation of the growth plate in Spry2-/- mice in vivo.

The first step was thus to examine SPRY2 expression in cells of interest in wild-type mice. Immunolabelling of the prenatal tibia showed that all these three cell types were positive for SPRY2, within the zone of ossification, as well as some other cells outside of it (e.g. chondrocytes).

Based on immunohistochemistry, osteoblasts, expressing their major marker, BGLAP (osteocalcin), were also positive for RUNX2 in the wild-type bone. RUNX2 is an inducer and an early marker of osteoblast differentiation,¹⁵ which is in agreement with its presence in the early growth plate. Modulation of Runx2 expression by SPRY2 was documented also in other system.¹⁴

Regarding osteocytes identified by their typical marker, sclerostin,¹⁶ Sost mRNA was strongly up-regulated in Spry2-/- tibias. Immunolabelling showed only a slight non-significant increase of SOST-positive osteocytes in Spry2-/- tibias. Therefore, the Sost mRNA changes were not due to an altered number of cells. The increased expression of Sost was neither mediated by the positive regulator RUNX2.¹⁷ Although RUNX2 was immunodetected in osteocytes, it was shown as decreased in Spry2-/- tibias. The increased Sost expression was caused neither by an increased HIF-1α signaling, a positive regulator of SOST expression,^18–20 since HIF-1α, a downstream molecule of SPRY2,²¹ had a similar expression in wild type and Spry2-/-. Thus, the regulation of Sost remains elusive, although neither connected with Runx2 nor Hif-1α.

The third type of bone-related cells (osteoclasts) was impacted via an increased expression of RANK in Spry2-/- mice. RANK is involved in osteoclastic differentiation from their hematopoietic precursors, through osteoblastic RANK ligand^22,23 showing a decreasing non-significant trend in expression as reported here. This imbalance resulted in only slightly non-significantly decreased number of TRAP-positive osteoclasts in Spry2-/- tibias at E18 as observed by Joo et al.⁵ Immunolabelling showed that RANK was mostly detected in mononuclear precursors located in intertrabecular regions.

Joo et al.⁵ have described Spry2−/− pups as indistinguishable from their wild-type littermates in their body weight with a slightly abnormal ratio between proliferating and hypertrophic zones of chondrocytes at E18. Similarly, when compared to the wild type, the histological appearance of Spry2-/- tibias did not show differences at E18 in our analysis. However, 6 weeks old Spry2-/mice exhibited a decrease in bone volume/tissue volume, trabecular number and thickness, an increased trabecular spacing and a reduced mineralizing function of osteoblasts.⁵ This postnatal phenotype can be supported by our prenatal investigation at molecular level. Although SPRY2 may apparently be involved in osteoclastic differentiation, the decreased bone formation in Spry2−/− mice seems to be caused by abnormal functions of osteoblasts or osteocytes. Since Runx2 deficient mice show an impaired ossification,²⁴ decreased expression of Runx2 may result in decreased osteoblastic differentiation and thus in decreased ossification during development. The same trend can relate to an increased expression of Sost. Indeed, SOST acts as a negative regulator of bone metabolism and transgenic mice overexpressing Sost show a decrease in bone formation rate, bone volume and bone mineral density.^25,26

CONCLUSION

Using the developing long bone as a model organ, the effects of Spry2 deletion on osteogenic cells appear complex in part because of its ubiquitous expression and also due to the multiple cell-specific effects of the encoded protein. This study showed differential tuning of distinct markers for a given cell type, as well as differences in the pattern of expression from one bone cell type to another. The present in vivo study brings complementary data to information gained from studying cell lines in vitro, where the associated loss of specific cell-cell interactions (through heterotypic cell contacts or diffusible molecules) may also impair regulatory mechanisms, as they exist in physiological conditions. Based on our observation, the decreased bone formation and mineralization of postnatal Spry2-/- mice might be related to the abnormal expression of osteoblastic Runx2 and osteocytic Sost.

Funding Statement

This work was supported by the Czech Science Foundation (Grantová Agentura České Republiky) GACR 14-37368G in 2018 and GACR 19-12023S in 2019.

Acknowledgments

Initial data were obtained in the frame of a research supported by the Czech Science Foundation project 14-37368G (2014-2018). This investigation was completed with support by Czech Science Foundation project 19-12023S (2019-2021) and Research Program Progress Q29 by Charles University. In 2018, Alice Ramesova participated in the frame of her student project 110/2018/FVL provided by The University of Veterinary and Pharmaceutical Sciences, Brno.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.