Abstract
Extracellular signal-regulated protein kinase 5 (ERK5) has been implicated during development and carcinogenesis. Nkx3.1-mediated Cre expression is a useful strategy to genetically manipulate the mouse prostate. While grossly normal at birth, we observed an unexpected phenotype of spinal protrusion in Nkx3.1:Cre;Erk5 fl/fl (Erk5 fl/fl) mice by ~6–8 weeks of age. X-ray, histological and micro CT (µCT) analyses showed that 100% of male and female Erk5 fl/fl mice had a severely deformed curved thoracic spine, with an associated loss of trabecular bone volume. Although sex-specific differences were observed, histomorphometry measurements revealed that both bone resorption and bone formation parameters were increased in male Erk5 fl/fl mice compared to wild type (WT) littermates. Osteopenia occurs where the rate of bone resorption exceeds that of bone formation, so we investigated the role of the osteoclast compartment. We found that treatment of RANKL-stimulated primary bone marrow-derived macrophage (BMDM) cultures with small molecule ERK5 pathway inhibitors increased osteoclast numbers. Furthermore, osteoclast numbers and expression of osteoclast marker genes were increased in parallel with reduced Erk5 expression in cultures generated from Erk5 fl/fl mice compared to WT mice. Collectively, these results reveal a novel role for Erk5 during bone maturation and homeostasis in vivo.
Similar content being viewed by others
Introduction
ERK5 (MAPK7 or BMK) belongs to the family of mitogen-activated protein kinases (MAPKs), members of which typically function as signalling nodes to integrate information from extracellular stimuli and different intracellular signalling cascades. Through this activity, MAPKs control numerous biological processes during development and homeostasis, including cellular proliferation, differentiation and survival. ERK5 has a uniquely large C-terminal domain, which contains nuclear localisation and export signals required for dynamic shuttling of ERK5. On the other hand, the N-terminal catalytic domain of ERK5 shares a 50% homology with ERK1/2 and inhibitors developed against MEK1/2, such as PD98059 and U0126, have been demonstrated to have off-target effects on the MEK5/ERK5 pathway. Unlike other members of the MAPK family, MEK5 is the only MEK (MAPK kinase) reported to directly interact with and activate ERK5 via phosphorylation1,2,3.
Deletion of Mek5 and Erk5 in mice has revealed their essential role during development, where almost identical phenotypes are observed4,5,6,7. Erk5 null mice die around embryonic day 10.5 due to cardiovascular defects with impaired vasculogenesis and angiogenesis. This is consistent with findings that ERK5 plays a critical role in endothelial cell function8,9. Growth in the head region in these mice was severely retarded. In addition, ERK5 has been implicated in the physiology of neurones10, muscle11 and immune cells12,13,14 by controlling proliferation, differentiation and cell survival. ERK1/2 and ERK5 have distinct roles in the regulation of brain-derived neurotrophic factor expression15. In addition to its essential role in development, there is an increasing body of evidence that ERK5 plays a role in tumour development (reviewed in ref.16) and previous work from our laboratory has implicated ERK5 as a key driver for prostate carcinogenesis17,18,19. Hence, we initiated experiments to test the impact of in vivo Nkx3.1:Cre-mediated Erk5 deletion in a transgenic mouse model20,21.
Formation of the axial skeleton begins with the subdivision of paraxial mesoderm into somites. Signals (sonic hedgehog) emanating from the notochord influence the early somites to form the sclerotome. Sclerotomal cells form the beginnings of the vertebrae. Vertebral structures subsequently develop and segment under the influence of various cellular signals, including Pax-9 and homeobox-containing genes22,23,24. A previous study to examine the pattern of Nkx3.1 mediated Cre recombinase expression, via the use of ROSA26 reporter mice, documented expression of Nkx3.1:Cre in a range of tissues, including somites, dorsal aspects of the spinal ridge, ribs and skull25. However, despite expression being noted in somites, no significant skeletal phenotype has been described in Nkx3.1 deleted mice to date (reviewed in ref.26).
In this study, our novel finding is that Nkx3.1:Cre driven Erk5 deletion associates with a phenotype of a severely curved spine and substantially reduced bone mass/osteopenia. We also report increased osteoclast differentiation in primary bone marrow cultures generated from wild type (WT) mice treated with small molecule ERK5 pathway inhibitors and in mice with Nkx3.1:Cre-mediated Erk5 deletion compared to WT, indicating a link between Erk5 and osteoclast activity in vivo.
Results
Nkx3.1-Cre Erk5 null mice developed severe spinal abnormalities
To circumvent Erk5 null induced embryonic lethality, conditional deletion of Erk5 was intended to target the prostate by Cre recombinase inserted into the Nkx3.1 locus. However, we observed marked phenotypic abnormalities in male and female Nkx3.1:Cre;Erk5 fl/fl mice (homozygous for the Erk5 conditional allele), referred to hereafter as Erk5 fl/fl mice. Newborn pups all had the same normal gross morphology, regardless of genotype, but by 10–12 weeks of age, Erk5 fl/fl male and female mice were noticeably and significantly smaller in size compared to WT [Nkx3.1:Cre;Erk5 +/+ or mice which did not express Cre recombinase] and Erk5 fl/+ [heterozygous for the Erk5 conditional allele; Nkx3.1:Cre;Erk5 fl/+] littermate controls (Fig. 1A and S1A and Tables S1 and S2). Erk5 fl/fl male and female mice, from ~6–8 weeks onwards, exhibited a protruding spine when compared to WT and Erk5 fl/+ littermate controls (Fig. 1A), indicative of a potential underlying bone defect causing abnormal curvature of the thoracic spine. Overall, among 33 Erk5 fl/fl mice (17 male and 16 female), all developed a spinal phenotype. In contrast, 17 (7 male and 10 female) Erk5 fl/+ and 42 (18 male and 24 female) WT littermate control mice did not demonstrate any clinical evidence of spinal defect. Moreover, a curled tail phenotype (Fig. 1B) was observed in ~30% Erk5 fl/fl mice (5/17 males and 5/16 females; > 6 weeks old). There was no evidence of abnormalities in other major organs examined, including the prostates of Erk5 fl/fl male mice (Fig. S1B).
X-ray analysis of whole body and isolated spines from 7 and 10–12 week old WT, Erk5 fl/+ and Erk5 fl/fl male and female mice showed no major differences between Erk5 fl/+ and WT mice (Fig. 1C and D). In contrast, we observed severe spinal curvature in the thoracic region of male and female Erk5 fl/fl mice at 7 weeks (Fig. 1E) and 10–12 weeks of age (Fig. 1C and D), regardless of curled tail phenotype. Micro-computed tomography (μCT) analysis of newborn pups (2 days after birth) demonstrated that there was no spinal defect present in Erk5 fl/fl mice compared to WT littermates (Fig. 1F), suggesting that the defect manifests during post-embryonic skeletal maturation. μCT analysis of isolated spines from older (10–12 weeks) male (Fig. 1G) and female (Fig. S1D) Erk5 fl/fl mice confirmed the marked curvature in the spine as seen in plain X-ray imaging. The affected vertebrae in male and female Erk5 fl/fl mice showed striking wedge-shaped deformities, indicative of vertebral collapse, as can be seen from the detailed images of WT and Erk5 fl/fl spines in the bottom panels of Fig. 1G and S1D. No spinal phenotype was observed in Nkx3.1 Cre/Cre;Erk5 +/+ mice, which were homozygous for the Nkx3.1:Cre allele (Fig. S1C). Collectively, these data suggest that the loss of Erk5, rather than Nkx3.1-driven Cre expression, is the causative event for the spinal phenotype.
Trabecular bone loss in Erk5 fl/fl mice
Histological analysis of the thoracic spine confirmed the deformation of the thoracic vertebrae in Erk5 fl/fl mice compared to WT mice (compare 4x images in Fig. 2A). Moreover, the vertebrae in the affected region from Erk5 fl/fl mice were variably larger in size and structurally dis-organised when compared to WT control mice. Vertebrae from Erk5 fl/fl mice were characterized by a significant increase in marrow space in association with a profound loss of trabecular bone mass (compare 10x and 20x images in Fig. 2A). The spinal cords showed no gross pathology associated with compression but inter-vertebral discs were enlarged with some structural malformation of nucleus pulposus (NP) tissue between collapsed thoracic vertebrae in Erk5 fl/fl mice compared to WT mice. However, it is noteworthy that the inter-vertebral disc NP tissue located between thoracic vertebrae that had not collapsed in Erk5 fl/fl mice appeared similar to WT (Fig. S1E).
The severe deformation of the thoracic vertebrae would make a quantitative analysis of the bone architecture difficult to interpret. Therefore μCT analysis of bone architecture was performed using the 5th lumbar vertebrae (L5) because generally, these were not badly deformed. High-resolution μCT analysis of L5 isolated from WT, Erk5 fl/+ and Erk5 fl/fl mice (Fig. 2B) demonstrated a dramatic loss of bone volume in Erk5 fl/fl male and female mice when compared to WT. Although the lumbar vertebrae showed no major deformation overall, some minor changes to the end plates and the central canal were observed (highlighted by yellow arrows). Both male and female Erk5 fl/fl mice showed substantial loss of trabecular bone volume (Fig. 2B and C and Table S3). The bone loss was more pronounced in male Erk5 fl/fl mice, where a 54% decrease in bone volume per tissue volume (BV/TV) was seen compared to WT. Female Erk5 fl/fl mice showed a lower, but still substantial 35% decrease in BV/TV compared to WT. The loss of trabecular bone volume was accompanied by a significant reduction in trabecular number (Tb.N) in both male (54.1% compared to WT) and female (39% compared to WT) Erk5 fl/fl mice. There was no significant difference in trabecular thickness (Tb.Th) in male and female Erk5 fl/fl mice compared to WT. However, a deterioration of bone architecture was observed, as indicated by decreased trabecular connectivity and a significant increase in the structure model index (SMI) in both male and female Erk5 fl/fl mice compared to WT. Erk5 fl/+ male mice also displayed reduced bone volume (15%) and trabecular number (15.8%) compared to WT, but this was less severe than Erk5 fl/fl, indicating the possibility of a dosage-dependent effect of Erk5 loss.
To address whether bone abnormalities were present in the appendicular skeleton in addition to the axial skeleton, we analysed distal femurs by μCT (Fig. S2A and B and Table S4). There was a substantial decrease in BV/TV in both Erk5 fl/fl and Erk5 fl/+ males (44% and 24% respectively) compared to WT. Female Erk5 fl/fl mice also demonstrated significantly reduced BV/TV (23%) compared to WT, although this was not as marked as in the Erk5 fl/fl male mice. Erk5 fl/+ females showed no significant bone loss in the distal femur. In keeping with trends seen for the spine, changes in bone architecture were observed in the distal femurs of male and female Erk5 fl/fl mice. Specifically, there was a significant reduction in Tb.Th and a significant increase in trabecular pattern factor (Tb.Pf) in both male and female Erk5 fl/fl mice compared to WT. In both Erk5 fl/+ and Erk5 fl/fl males, there was a significant reduction in Tb.N and increase in trabecular separation (Tb.Sp) compared to WT but these parameters were not changed in female Erk5 fl/+ or Erk5 fl/fl mice. It is noteworthy that the overall shape and morphology of the long bones of Erk5 fl/fl mice appeared similar to WT and the growth plate in distal femurs of male Erk5 fl/fl mice had no significant abnormalities (Fig. S2C), indicating that there was no major defect with endochondral ossification.
Analysis of bone turnover in Erk5 fl/fl mice
Bone homeostasis is achieved via the controlled activities of two major cell types: osteoblasts synthesize and deposit new bone matrix while osteoclasts break down and resorb bone. The bone loss observed in the Erk5 fl/fl mice could therefore be due to a decrease in bone formation, an increase in bone resorption or a combination of both. In order to address this question, we performed dynamic histomorphometry on L5 vertebrae of 10–12 week old mice. In Erk5 fl/fl males, the histomorphometry showed a significant 69% increase in osteoclast surface (Oc.S) and a 68% increase in osteoclast number (N.Oc) per bone surface (BS) (N.Oc/BS) (Fig. 3B) as assessed by tartrate-resistant acid phosphatase (TRAP) staining (Fig. 3A). There was a non-significant trend for increased osteoclast surface and a significant increase in osteoclast number in Erk5 fl/+ males, again indicating the possibility of a dosage-dependent effect of Erk5 loss (Fig. 3B and Table S5). There was no statistically significant increase in bone resorption parameters in Erk5 fl/+ or Erk5 fl/fl female mice. Calcein double labels were used to assess bone formation parameters. While there was no difference in mineral apposition rate, we observed a significant increase in mineralizing surface per bone surface (MS/BS) and bone formation rate per bone surface (BFR/BS) in Erk5 fl/fl males compared to WT. There was no significant difference in any bone formation parameters in Erk5 fl/+ males or Erk5 fl/+ or Erk5 fl/fl female mice compared to WT controls (Table S5).
Collectively, data from histomorphometry analysis are consistent with the observed bone loss in the vertebrae of Erk5 fl/fl mice being associated with increased osteoclast activity. Key mediators of osteoclastogenesis include the receptor activator of NF-κB ligand (RANKL) and its negative regulator, osteoprotegerin (OPG). We measured the serum levels of these secreted factors in WT and Erk5 fl/fl male mice (7–11 weeks of age). There was a significant decrease in serum RANKL levels and a significant increase in serum OPG levels, resulting in a significant decrease in the ratio of RANKL to OPG from 0.164 in WT to 0.047 in Erk5 fl/fl male mice (p = 0.007) (Fig. 3C). It is possible that the loss of bone mass in Erk5 fl/fl mice may have resulted in the activation of compensatory mechanisms, such as the increased production of OPG and decreased production of RANKL.
Reduction of Erk5 expression increased osteoclast numbers and expression of Rank, Cathepsin K (Ctsk) and Nuclear factor of activated T-cells (Nfatc1) in osteoclasts derived from Erk5 fl/fl mice
To further investigate the involvement of Erk5 in the osteoclast compartment, macrophage colony-stimulating factor (M-CSF) dependent bone marrow-derived macrophages (BMDMs, osteoclast lineage pre-cursors) and osteoclasts (RANKL-stimulated BMDMs) were generated from the long bones of WT and Erk5 fl/fl mice. Quantitative PCR (QPCR) was performed using RNA extracted from these BMDM and osteoclast cultures. BMDM and osteoclast cultures derived from Erk5 fl/fl mice exhibited significantly lower expression of Erk5 Exon 4 (the floxed region) when compared to WT controls (Fig. 4A and B), in keeping with an Nkx3.1:Cre mediated event. Moreover, BMDM cultures generated from Erk5 fl/fl mice had significantly reduced levels of ERK5 protein compared to WT controls as assessed by Western blot (Fig. 4B and S3). To confirm reduced expression of Erk5 in vertebral tissue sections, we tried immunohistochemical and in situ hybridization, by RNA Scope, staining of vertebral tissue sections for ERK5 protein and Erk5 mRNA transcripts respectively but this was unsuccessful. To demonstrate successful areas of Cre recombination, a proportion of mice were bred to carry the Z/EGFP reporter transgene27. The Z/EGFP reporter mouse expresses lacZ throughout all embryonic and adult developmental stages. The expression of Nkx3.1:Cre during development leads to the excision of the lacZ gene, allowing expression of the second reporter, EGFP and so this essentially acts as a surrogate marker of Cre expression. GFP was found to be expressed in the macrophage and osteoclast cultures (Figure S2D) generated from Nkx3.1:Cre-expressing, Z/EGFP +ve mice, indicating the expression of Cre in the bone. RANKL dose-response experiments showed significantly increased numbers of osteoclasts in Erk5 fl/fl cultures versus WT at concentrations of 10 ng/mL RANKL and above (Fig. 4C). Furthermore, the MEK5 inhibitors BIX02188 and BIX02189 stimulated RANKL-induced osteoclast formation in cultures from WT mice (Fig. 4D), suggesting that Erk5 negatively regulates osteoclast differentiation. Next, we characterized the expression of a panel of osteoclast markers. Consistent with enhanced osteoclast differentiation, we found a significant increase in the expression of Rank (the receptor for RANKL), Ctsk (the gene encoding cathepsin K, the protease released by osteoclasts during bone resorption) and Nfatc1 (a transcription factor that is crucial for osteoclast differentiation) in Erk5 fl/fl cultures versus WT (Fig. 4E). Collectively, these data demonstrate that reduction of Erk5 expression in the osteoclast compartment correlates with increased osteoclastogenesis and function.
Discussion
While investigating the role of Erk5 in the prostate using Nkx3.1-Cre recombinase expression, we observed a novel, dramatic spinal deformity in Erk5 fl/fl mice. Erk5 fl/fl mice were phenotypically normal at birth but developed the spinal defect by 2 months of age, suggesting that the loss of Erk5 during the process of post-natal skeletal maturation is associated with the observed spinal phenotype. The spinal deformity was found to be linked with a profound reduction of trabecular bone mass in the vertebral column of both male and female Erk5 fl/fl mice. From histomorphometry studies, we found that the low bone mass in male Erk5 fl/fl mice was associated with increased bone-remodeling, with both significantly enhanced bone resorption and bone formation parameters being observed. It was striking that almost all (~90%) of the entire bone surface was undergoing re-modelling, either resorption or bone formation, in male Erk5 fl/fl mice.
Low bone mass (osteopenia/osteoporosis) occurs when bone turnover is imbalanced in favour of bone resorption i.e. where the rate of bone resorption by osteoclasts exceeds the rate of bone formation by osteoblasts. Our dynamic histomorphometry data in male Erk5 fl/fl mice point towards dis-regulated osteoclast function upon loss of Erk5 as being the primary cause for the observed bone loss. However, we did not observe increased osteoclast surface or number in Erk5 fl/fl female mice and the loss of bone mass was greater in male compared to female Erk5 fl/fl mice, indicating that there are sex-specific differences. Gonadal sex steroid hormones (androgens and estrogens) are known to play an important role in bone homeostasis28,29. In particular, estrogen is known to inhibit osteoclastogenesis and osteoblast apoptosis30. Osteoporosis is a common disease in post-menopausal women, when estrogen levels have declined. It is established that while there is a net loss of bone mass, both bone resorption and bone formation rates are elevated in post-menopausal osteoporosis31, which is in keeping with our observations in male Erk5 fl/fl mice. The spinal deformity develops in Erk5 fl/fl female mice during the first two months of life and we suggest that the presence of estrogen in the female Erk5 fl/fl mice at the time when bone histomorphometry parameters were assessed (10 weeks) mediates a protective effect in terms of inhibiting osteoclastogenesis/bone resorption and reducing the elevated bone formation rate when compared to male Erk5 fl/fl mice. Testosterone levels in male Erk5 fl/fl mice were found to be normal (data not shown) but future studies, beyond the scope of this work, to determine whether estrogen treatment could have a protective effect in the male Erk5 fl/fl mice would be interesting and informative.
Development and maintenance of bone tissue involves precise spatial and temporal interplay between osteoclast and osteoblast function. RANKL is a key regulator of osteoclastogenesis and is negatively regulated by OPG. It is increasingly apparent that the major sources of both of these factors are osteoblasts and osteocytes, highlighting an important interplay between bone formation and bone resorption32,33. Our gross observations of the phenotype and extent of spinal deformity in Erk5 fl/fl mice are similar to those from a classical study of osteoprotegrin (OPG)-deficient mice34, where it was reported that while embryonic bone formation appeared normal, adolescent and adult OPG-deficient mice exhibited a marked decrease in total bone density and trabecular bone porosity. Furthermore, the OPG-deficient mice also presented with curvature of the spine, had an increased incidence of vertebral bone fractures in the first two months of age and exhibited very active bone re-modeling, with increases in both osteoclast and osteoblast surface per bone surface. We examined the serum levels of RANKL and OPG in 7–11 week old WT and Erk5 fl/fl male mice by ELISA and found that at this age, there was a significant reduction in the ratio of RANKL to OPG in Erk5 fl/fl mice compared to WT, which would normally result in lower bone resorption levels. Although the RANKL/OPG ratio is very important for the maintenance of bone homeostasis, our results indicate that changes in this ratio are not what drives the low bone mass phenotype in these mice. Rather, this phenotype is most likely driven by increased responsiveness of the osteoclast precursors to RANKL, as suggested from our primary BMDM culture studies using a RANKL dose-response curve. The resulting increase in osteoclast formation could lead to bone loss in vivo, and this bone loss may result in the activation of compensatory mechanisms, such as the increased production of OPG and decreased RANKL production. Increased serum OPG levels have been observed in cases of post-menopausal osteoporosis35.
It is of interest that while Erk5 has not been implicated in genome-wide association studies (GWAS) in the osteoporosis field to date36, a network-based meta analysis of gene expression profiles in women with bone mineral variations identified Mekk3 (Mek kinase 3), which encodes an upstream kinase of MEK5 in the ERK5 pathway, as one of five candidate genes to be associated with bone mineral density37. Furthermore, Mekk3 was identified as being induced by M-CSF in an oligonucleotide microarray study of gene expression patterns in mouse bone marrow mononuclear cells during the process of osteoclast differentiation38. Another upstream kinase of MEK5 in the ERK5 pathway, Mekk2, has previously been shown to be important in the regulation of osteoblast activity39. Collectively, our data are in keeping with a role for ERK5 pathway in normal bone homeostasis and clinically relevant osteoporosis.
Besides the spine, μCT of the long bones in the hind limbs also revealed significant loss of bone mass and trabecular structure in the Erk5 fl/fl mice, indicating that there are effects in the appendicular skeleton as a result of Nkx3.1-mediated Cre recombinase expression. Our finding that Erk5 expression is reduced in BMDM and osteoclast cultures support the notion of an impact of loss of Erk5 in the appendicular skeleton. Examination of the growth plate from µCT images of distal femurs revealed that there was no major defect in growth plate cartilage/endochondral ossification, in male Erk5 fl/fl mice compared to WT. However, the mineralised cartilage area appeared to be thinner, most probably due to the increased osteoclast activity in these mice. The additional phenotype of curly tail in a proportion of Erk5 fl/fl mice is intriguing, and may result from the spine defect, similar to those seen in mouse models for spina bifida, neural tube defects and other transgenic mouse models40,41.
Importantly, we demonstrated that GFP, a surrogate marker for Nkx3.1:Cre activity, was expressed in BMDM and osteoclast cultures generated from Nkx3.1:Cre-expressing, Z/EGFP + ve mice. In parallel, Erk5 mRNA expression was significantly reduced in BMDM and osteoclast cultures and total ERK5 protein expression was significantly reduced in BMDM cultures generated from Erk5 fl/fl mice compared to WT, which is in keeping with Cre-mediated recombination having occurred. The efficiency of Cre-mediated recombination is never 100% and so our observed partial reduction in Erk5 expression is indicative that deletion is occurring in only a subset of the cells. Despite this, osteoclast numbers and expression of osteoclast differentiation markers including Rank, Ctsk and Nfatc1 were significantly increased in BMDM and osteoclast cultures generated from Erk5 fl/fl mice compared to WT. As further validation of our findings in Erk5 fl/fl cultures, we also demonstrated that inhibition of ERK5 pathway via treatment with the MEK5 inhibitors, BIX02188 and BIX02189, significantly increased osteoclast numbers in cultures generated from WT mice. Collectively, these data indicate that reduction of Erk5 expression or activity in osteoclast pre-cursors is associated with increased osteoclastogenesis.
While our data indicate the osteoclast as being the primary compartment responsible for the observed bone loss, it is possible that other cell types could be affected by loss of Erk5 expression driven by Nkx3.1-Cre. Interestingly, a recent study demonstrated that ERK5 is implicated in degenerated human intervertebral disc NP tissues42. ERK5 expression was found to be reduced in degenerated compared to normal tissues and treatment with TNF-α, which is implicated in NP cell degeneration, reduced Erk5 mRNA expression and the expression of NP cell marker genes. In our study, the NP tissue between collapsed thoracic vertebrae in Erk5 fl/fl mice was found to have an abnormal, distorted appearance whereas NP tissue located between thoracic vertebrae that had not collapsed in Erk5 fl/fl mice appeared similar to WT. Furthermore, our histomorphometry studies were performed in lumbar vertebrae that showed no major deformation in intervertebral NP tissue, yet these vertebrae still show increased bone resorption. We suggest, therefore, that the NP deformation most likely results from vertebral collapse. Future studies to target Erk5 ablation using an osteoclast-specific Cre, such as Ctsk-Cre, or other cell lineage-specific-Cre would be very informative to further address the question of the causal relationship between loss of Erk5 in the osteoclast compartment with the spinal deformity.
To date, in vitro investigations of ERK5 in osteoblast and osteoclast biology have produced conflicting findings with both differentiation promoting and suppressive effects being documented43,44,45,46. Cyclic fluid shear stress has been shown to stimulate the proliferation of osteoblasts via an ERK5 signalling pathway43,44. However, in contrast to those studies, it has also been reported that the expression of key phenotypic markers for osteoblast differentiation, osteocalcin, alkaline phosphatase and osterix, are reduced after treatment with the MEK5 inhibitor, BIX02189, suggesting that MEK5 can suppress osteoblast differentiation45. Our data are in keeping with the former studies as we observe increased bone formation parameters in male Erk5 fl/fl mice although we suggest that this could be an indirect effect of enhanced osteoclast activity. A very recent study has shown that ERK5 activation, through the induction of c-Fos, is essential for osteoclast-like differentiation of the monocytic RAW264.7D clone and 4B12 cells in response to RANKL and RANKL + M-CSF respectively46. The findings of Amano et al. are in contrast to our report as we demonstrate that reduction of Erk5 expression, in the case of Erk5 fl/fl mice, or inhibition of ERK5 pathway, in the case of treatment with small molecule MEK5 inhibitors, associated with increased osteoclast numbers and expression of osteoclast differentiation markers. Besides using primary bone marrow cells from our in vivo mouse model as opposed to cell lines, the ERK5 pathway inhibitor concentrations we used were considerably lower than those used by Amano et al. (100 nM versus 1–8 µM) because at high concentration of these inhibitors, we observed a high degree of toxicity and it is likely that selectivity for MEK5 is lost given that the reported IC50 for MEK5 kinase inhibition is 1.5 nM for both BIX0218 and BIX02189. Furthermore, Amano et al. used the inhibitor, XMD8-92, which is now known to have off-target effects against bromo domain-containing proteins47. For the first time, our study sheds important insights into the functional role of ERK5 in bone biology in an in vivo model.
In conclusion, we demonstrate an important role of Erk5 in bone development and homeostasis in vivo. Loss of Erk5 driven by Nkx3.1-Cre associated with a severe spinal deformity and an osteopenic phenotype. Furthermore, our evidence indicates that reduction of ERK5 pathway activity and Erk5 gene expression promotes osteoclastogenesis. Our model could be of benefit to further our knowledge and understanding of the molecular events regulating bone homeostasis, spinal deformities and/or osteopenia.
Materials and Methods
Mouse Strains and Breeding
All small animal (murine) experiments were (i) approved by the Animal Welfare and Ethical Review Board (AWERB) at the University of Glasgow and (ii) performed in accordance with relevant guidelines and regulations. Nkx3.1:Cre mice48 were intercrossed with mice harbouring the conditional inactivatable Erk5 allele (where Exon 4 is flanked by LoxP sites)49 to generate Nkx3.1:Cre;Erk5 fl/+ and Nkx3.1:Cre;Erk5 fl/fl mice. A proportion of these mice were interbred with mice containing the Z/EGFP reporter transgene27. This transgene results in the expression of β-galactosidase in most tissues. The β-geo insert of the transgene is flanked by lox p sites and the expression of Cre recombinase results in the excision of the β-geo insert, activating the constitutive expression of GFP where Cre is expressed. Mice were genotyped by PCR by TransnetyxTM. Mice were of a mixed background (C57BL/6J) with strain-matched littermates (either Cre+ve WT or mice not expressing Cre) used as control mice.
Radiography
Whole body radiographs and radiographs of isolated spines of transgenic animals after sacrifice were performed using the Siemens Angiographie (AX) system. Exposure factors were: 40 kV; 50 mAs; 115 cm film to focus distance (FFD). Agfa CR 35-X developing cassettes were used.
Histological Analysis
For histological analysis, mice were sacrificed using CO2 narcosis in order to maintain an intact skeleton, then bone tissues and a spectrum of other tissues, including liver, lung, kidney, heart, pancreas, prostate, leg muscle and small intestine, were collected and fixed intact in 10% formalin for 24 hr and subsequently processed for paraffin embedding. Sections from these paraffin blocks were stained with haematoxylin and eosin (H + E).
μCT analysis
Whole mouse spines and 2-day-old pups were imaged using a Skyscan1076 in vivo μCT scanner (Skyscan, Kontich, Belgium) and were scanned at a resolution of 18 μm, with a source voltage of 50 kV, a 0.6° rotation step and a 0.5 mm aluminium filter. The scans were reconstructed using NRecon software, and visualised using CTVox and Dataviewer software (all Skyscan).
High resolution scans of mouse lumbar vertebrae and distal femurs were obtained using a Skyscan 1272 system at 4.5 μm resolution, 50 kV source voltage, 0.5° rotation step and a 0.5 mm aluminium filter. Scans were visualised as described above, and analysed using CtAn software (Skyscan). For the 5th lumbar vertebrae, the bone parameters were measured in the vertebral body, excluding the cortex and the end plates. For the distal femur, the analysis was performed in 200 slices 0.1 mm proximal of the primary spongiosa.
Bone Histomorphometry
The mice received two injections of calcein (200 µl IP injection of 2 mg/ml) at 5 and 2 days before culling. Specimens were then fixed in 10% buffered formalin for 24 hr, and stored in 70% ethanol. After uCT analysis, the samples were dehydrated through an alcohol series, cleared in xylene, embedded in methyl methacrylate, and 5 μm sections were cut on a Leica RM2265 microtome.
For analysis of bone resorption parameters, the specimens were stained for TRAP to identify osteoclasts and the bone counterstained using Aniline Blue as described by Chappard et al.50. The sections were imaged on a Zeiss Axioimager microscope with a 10x lens and a QImaging Retiga 4000 camera resulting in a pixel size of 1.487 μm.
For analysis of bone formation parameters, the sections were stained without deplastification for 3 min in 0.1% Calcein Blue (Sigma), pH 8. The sections were washed twice in water, dehydrated through an alcohol series, cleared in xylene and coverslipped using Eukit (Sigma). The mineralised bone was imaged using a DAPI filter set, and the calcein labels using an FITC filter set on a Zeiss Axioimager microscope with a 20x lens and a QImaging Retiga 4000 camera resulting in a pixel size of 0.372 μm.
Bone histomorphometric analysis was performed using a custom in-house developed image analysis program based on ImageJ, available at https://www.liverpool.ac.uk/ageing-and-chronic-disease/bone-hist/ 51.
RNA Isolation and Quantitation
RNA was isolated using RNeasy Mini Kit (Qiagen) as per manufacturer’s instructions, including a DNAse digestion step. RNA samples were quantified either by Qbit Assay (Invitrogen) according to manufacturer’s instructions or using a NanoDrop™ spectrophotometer (Thermo Scientific).
Quantitative Real Time Polymerase Chain Reaction (QPCR) – Taqman
cDNA was prepared from RNA samples using High Capacity cDNA Transcription Kit (Applied Biosystems). QPCR was performed to evaluate relative transcript expression levels in cDNA samples using the Taqman technique. Specific primer/probe combinations (Table S6) and Taqman Universal PCR Mastermix (Applied Biosystems) were added to cDNA samples (according to manufacturer’s instructions) before being run on an Applied Biosystems 7500 Fast Real-Time PCR System machine. The QPCR conditions were: 20 sec at 50 °C, 10 min at 95 °C, followed by 40 cycles of 15 sec at 95 °C and 1 min at 60 °C. Technical replicates for each sample and a minimum of three independent samples for each genotype/cell line were included. Hmbs, the gene encoding hydroxymethylbilane synthase, was used as the ‘house-keeping’ or reference gene and 7500 Software v2.0.5 (Applied Biosystems) used 2−ΔΔCT method to determine the relative gene expression. Primer/Probe sets were designed using the Universal Probe Library (UPL) Assay design centre (Roche).
Enzyme-linked immunosorbent assay (ELISA)
Serum samples were collected from mouse whole blood by centrifugation for 15 min at approximately 800 × g. We used the RANKL (TNFSF11) Mouse ELISA Kit (ab100749) and Osteoprotegerin Mouse ELISA Kit (ab100733) (ELISA) from Abcam® (Cambridge, UK) to measure the serum levels of RANKL and OPG in 7–11 week old male WT and Erk5 fl/fl mice, according to manufacturer’s instructions. Serum samples were diluted 1:15 for RANKL ELISA and 1:50 for OPG ELISA using Diluent A which was provided with the kits. Standard curves and experimental samples were performed in duplicates. Data are expressed as nanograms per milliliter.
Generation and culture of BMDMs and osteoclasts
Bone marrow cells were flushed out of the long bones of the mice and the cell suspension cultured in αMEM supplemented with 10% FCS, Pen/Strep and 100 ng/ml M-CSF (Prospec Bio) for three days. Next the non-adherent cells were removed by washing the cultures with PBS, and the adherent BMDMs harvested using Acutase (Sigma) or cell dissociation buffer (Enzyme-Free; PBS-based) (Gibco, Life Technologies). For osteoclast formation experiments, cells were plated in 96-well plates at 15 × 103 cells in 100 µl medium supplemented with 25 ng/ml M-CSF and up to 100 ng/ml RANKL (a kind gift of Dr. J. Dunford, University of Oxford) per well. The culture medium was refreshed at day 2 and day 4 and the cells fixed on day 5 with 4% buffered formalin and stained for TRAP. TRAP positive cells containing three or more nuclei were counted as osteoclasts. For RNA isolation, BMDM cells were seeded in 6-well plates at 5 × 105 cells per well in medium supplemented with 25 ng/ml M-CSF only (for macrophage cultures) or 25 ng/ml M-CSF plus 100 ng/ml RANKL (for osteoclast cultures) and cultured for 5 days as described above. For Western Blot analysis, BMDM cells were seeded in 6-well plates at 5 × 105 cells per well in medium supplemented with 25 ng/ml M-CSF and cultured for 3 days.
Western Blotting
Whole cell lysates (WCL) were prepared by lysing cells in lysis buffer [50 mM Tris pH 7.6, 150 mM NaCl, 1% Triton X-100, 0.5% deoxycholate, 0.1% SDS, 1 mM sodium ortho-vanadate, 5 mM sodium fluoride, 50 µg/mL phenylmethylsulfonyl fluoride (PMSF), protease inhibitor cocktail mix 1 (Calbiochem) and PhosSTOP (Roche)]. WCL were resolved by SDS-PAGE on 4–12% gradient polyacrylamide gels (Invitrogen) at 180 V for 1 hr and transferred electrophoretically by wet transfer system onto PVDF membrane (Milipore) at 100 V for 1 hr. Blots were blocked for 1 hr with 5% skimmed milk, rinsed and probed with the required primary antibody (diluted in 5% BSA, 0.1% Tween-20 containing TBS and 0.05% sodium azide) overnight at 4 °C. Following incubation with appropriate horseradish peroxidase-conjugated secondary antibody (Cell Signalling, 7076 or 7074, 1:5000) (in 5% skimmed milk), bands were visualised using Bio-RAD ChemiDoc™ Imaging System and software. Antibodies used were: ERK5 (Cell Signalling, 3372, 1:1000) and HSC70 (B-6) (Santa Cruz Biotechnology, sc-7298, 1:1000). Densitometry was performed using Image J software.
Statistical analysis
Statistical analyses (t-test; ANOVA, using Dunnett’s post-hoc test) were performed using GraphPad Prism v7.02 or SPSS v21 software. For all graphs, mean ± SEM (error bars) are presented.
References
Drew, B. A., Burow, M. E. & Beckman, B. S. MEK5/ERK5 pathway: the first fifteen years. Biochim. Biophys. Acta 1825, 37–48 (2012).
Nishimoto, S. & Nishida, E. MAPK signalling: ERK5 versus ERK1/2. EMBO Rep. 7, 782–786 (2006).
Wang, X. & Tournier, C. Regulation of cellular functions by the ERK5 signalling pathway. Cell. Signal. 18, 753–760 (2006).
Hayashi, M. & Lee, J. D. Role of the BMK1/ERK5 signaling pathway: lessons from knockout mice. J. Mol. Med. (Berl.) 82, 800–808 (2004).
Wang, X. et al. Targeted deletion of mek5 causes early embryonic death and defects in the extracellular signal-regulated kinase 5/myocyte enhancer factor 2 cell survival pathway. Mol. Cell. Biol. 25, 336–345 (2005).
Regan, C. P. et al. Erk5 null mice display multiple extraembryonic vascular and embryonic cardiovascular defects. Proc. Natl. Acad. Sci. USA 99 (2002).
Yan, L. et al. Knockout of ERK5 causes multiple defects in placental and embryonic development. BMC Dev. Biol. 3, 11 (2003).
Hayashi, M. et al. Targeted deletion of BMK1/ERK5 in adult mice perturbs vascular integrity and leads to endothelial failure. J. Clin. Invest. 113, 1138–1148 (2004).
Roberts, O. L., Holmes, K., Müller, J., Cross, D. A. & Cross, M. J. ERK5 and the regulation of endothelial cell function. Biochem. Soc. Trans. 37, 1254–1259 (2009).
Cavanaugh, J. E. Role of extracellular signal regulated kinase 5 in neuronal survival. Eur. J. Biochem. 271, 2056–2059 (2004).
Sunadome, K. et al. ERK5 regulates muscle cell fusion through Klf transcription factors. Dev. Cell 20, 192–205 (2011).
Sohn, S. J., Lewis, G. M. & Winoto, A. Non-redundant function of the MEK5-ERK5 pathway in thymocyte apoptosis. EMBO J. 27, 1896–1906 (2008).
Rovida, E. et al. ERK5/BMK1 is indispensable for optimal colony-stimulating factor 1 (CSF-1)-induced proliferation in macrophages in a Src-dependent fashion. J. Immunol. 180, 4166–4172 (2008).
Ananieva, O. et al. ERK5 regulation in naïve T-cell activation and survival. Eur. J. Immunol. 38, 2534–2547 (2008).
Su, C., Underwood, W., Rybalchenko, N. & Singh, M. ERK1/2 and ERK5 have distinct roles in the regulation of brain-derived neurotrophic factor expression. J. Neurosci. Res. 89, 1542–1550 (2011).
Simões, A. E., Rodrigues, C. M. & Borralho, P. M. The MEK5/ERK5 signalling pathway in cancer: a promising novel therapeutic target. Drug Discov. Today (2016).
Mehta, P. B. et al. MEK5 overexpression is associated with metastatic prostate cancer, and stimulates proliferation, MMP-9 expression and invasion. Oncogene 22, 1381–1389 (2003).
McCracken, S. R. et al. Aberrant expression of extracellular signal-regulated kinase 5 in human prostate cancer. Oncogene 27, 2978–2988 (2008).
Ramsay, A. K. et al. ERK5 signalling in prostate cancer promotes an invasive phenotype. Br. J. Cancer 104, 664–672 (2011).
Ahmad, I. et al. HER2 overcomes PTEN (loss)-induced senescence to cause aggressive prostate cancer. Proc. Natl. Acad. Sci. USA 108, 16392–16397 (2011).
Patel, R. et al. Sprouty2, PTEN, and PP2A interact to regulate prostate cancer progression. J. Clin. Invest. 123, 1157–1175 (2013).
Carlson, B. M. Human embryology and developmental biology. 5th edn, Elsevier Saunders (2014).
Gilbert, S. F. Developmental biology. 9th edn, Sinauer Associates, Inc. (2010).
Kaplan, K. M., Spivak, J. M. & Bendo, J. A. Embryology of the spine and associated congenital abnormalities. Spine J. 5, 564–576 (2005).
Stanfel, M. N. et al. Expression of an Nkx3.1-CRE gene using ROSA26 reporter mice. Genesis 44, 550–555 (2006).
Abate-Shen, C., Shen, M. M. & Gelmann, E. Integrating differentiation and cancer: the Nkx3.1 homeobox gene in prostate organogenesis and carcinogenesis. Differentiation 76, 717–727 (2008).
Novak, A., Guo, C., Yang, W., Nagy, A. & Lobe, C. G. Z/EG, a double reporter mouse line that expresses enhanced green fluorescent protein upon Cre-mediated excision. Genesis 28, 147–155 (2000).
Nelson, E. R., Wardell, S. E. & McDonnell, D. P. The molecular mechanisms underlying the pharmacological actions of estrogens, SERMs and oxysterols: implications for the treatment and prevention of osteoporosis. Bone 53, 42–50 (2013).
Martin, A. C. Osteoporosis in men: a review of endogenous sex hormones and testosterone replacement therapy. J. Pharm. Pract. 24, 307–315 (2011).
Weitzmann, M. N. & Pacifici, R. Estrogen deficiency and bone loss: an inflammatory tale. J. Clin. Invest. 116, 1186–1194 (2006).
Feng, X. & McDonald, J. M. Disorders of bone remodeling. Annu. Rev. Pathol. 6, 121–145 (2011).
Nakashima, T., Hayashi, M. & Takayanagi, H. New insights into osteoclastogenic signaling mechanisms. Trends Endocrinol. Metab. 23, 582–590 (2012).
Weitzmann, M. N. The Role of Inflammatory Cytokines, the RANKL/OPG Axis, and the Immunoskeletal Interface in Physiological Bone Turnover and Osteoporosis. Scientifica (Cairo) 2013, 125705 (2013).
Bucay, N. et al. osteoprotegerin-deficient mice develop early onset osteoporosis and arterial calcification. Genes Dev. 12, 1260–1268 (1998).
Yano, K. et al. Immunological characterization of circulating osteoprotegerin/osteoclastogenesis inhibitory factor: increased serum concentrations in postmenopausal women with osteoporosis. J Bone Miner. Res. 14, 518–527 (1999).
Liu, Y. J., Zhang, L., Papasian, C. J. & Deng, H. W. Genome-wide Association Studies for Osteoporosis: A 2013 Update. J Bone Metab. 21, 99–116 (2014).
He, H. et al. Network-Based Meta-Analyses of Associations of Multiple Gene Expression Profiles with Bone Mineral Density Variations in Women. PLoS One 11, e0147475 (2016).
Cappellen, D. et al. Transcriptional program of mouse osteoclast differentiation governed by the macrophage colony-stimulating factor and the ligand for the receptor activator of NFkappa B. J. Biol. Chem. 277, 21971–21982 (2002).
Yamashita, M. et al. Ubiquitin ligase Smurf1 controls osteoblast activity and bone homeostasis by targeting MEKK2 for degradation. Cell 121, 101–113 (2005).
De Castro, S. C. et al. Lamin b1 polymorphism influences morphology of the nuclear envelope, cell cycle progression, and risk of neural tube defects in mice. PLoS Genet. 8, e1003059 (2012).
van Straaten, H. W. & Copp, A. J. Curly tail: a 50-year history of the mouse spina bifida model. Anat. Embryol. (Berl.) 203, 225–237 (2001).
Liang, W. et al. Differential expression of extracellular-signal-regulated kinase 5 (ERK5) in normal and degenerated human nucleus pulposus tissues and cells. Biochem. Biophys. Res. Commun. 449, 466–470 (2014).
Li, P. et al. Cyclic fluid shear stress promotes osteoblastic cells proliferation through ERK5 signaling pathway. Mol. Cell. Biochem. 364, 321–327 (2012).
Zhao, L. G. et al. The MEK5/ERK5 pathway mediates fluid shear stress promoted osteoblast differentiation. Connect. Tissue Res. 55, 96–102 (2014).
Kaneshiro, S., Otsuki, D., Yoshida, K., Yoshikawa, H. & Higuchi, C. MEK5 suppresses osteoblastic differentiation. Biochem. Biophys. Res. Commun. 463, 241–247 (2015).
Amano, S., Chang, Y. T. & Fukui, Y. ERK5 Activation Is Essential for Osteoclast Differentiation. PLoS One 10, e0125054 (2015).
Lin, E. C. K. et al. ERK5 kinase activity is dispensable for cellular immune response and proliferation. Proc. Natl. Acad. Sci. USA 113, 11865–11870 (2016).
Thomsen, M. K., Butler, C. M., Shen, M. M. & Swain, A. Sox9 is required for prostate development. Dev. Biol. 316, 302–311 (2008).
Wang, X. et al. Activation of extracellular signal-regulated protein kinase 5 downregulates FasL upon osmotic stress. Cell Death Differ. 13, 2099–2108 (2006).
Chappard, D., Alexandre, C. & Riffat, G. Histochemical identification of osteoclasts. Review of current methods and reappraisal of a simple procedure for routine diagnosis on undecalcified human iliac bone biopsies. Basic Appl. Histochem. 27, 75–85 (1983).
van ‘t Hof, R. J., Rose, L., Bassonga, E. & Daroszewska, A. Open source software for semi-automated histomorphometry of bone resorption and formation parameters. Bone 99, 69–79 (2017).
Acknowledgements
We thank the Cancer Research UK Beatson Institute core research services, including the biological services unit and the histology department. In particular, we acknowledge Colin Nixon in the histology department for help with staining of tissues. We acknowledge Elaine MacDuff, consultant pathologist NHS Greater Glasgow and Clyde, for histopathological examination of tissues. We acknowledge Gillian Cameron and Nicola Brannan at the University of Glasgow Small Animal Veterinary Hospital for help with X-ray imaging. Erk5 fl/fl mice were provided by Dr Kathy Tournier, University of Manchester. We acknowledge Prof. Ingunn Holen, Sheffield University, an independent expert in the bone field, for advice regarding staining of bone tissue. This work was supported by Cancer Research UK (grant number A15151, A10419 and A17196) and Prostate Cancer UK (PG10-10).
Author information
Authors and Affiliations
Contributions
C.J.L., R.J.V.H., I.A., M.W., O.S., K.B. and H.Y.L. conceived experiments. C.J.L., R.J.V.H., G.C., A.K., E.T., L.R., A.D., A.P., I.A., M.W., E.J.M., C.F. and M.S. carried out experiments and analysed data. C.J.L., R.V.H. and H.Y.L. wrote the paper.
Corresponding authors
Ethics declarations
Competing Interests
The authors declare that they have no competing interests.
Additional information
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Loveridge, C.J., van ’t Hof, R.J., Charlesworth, G. et al. Analysis of Nkx3.1:Cre-driven Erk5 deletion reveals a profound spinal deformity which is linked to increased osteoclast activity. Sci Rep 7, 13241 (2017). https://doi.org/10.1038/s41598-017-13346-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-017-13346-8
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.