Sex-specific hormonal rescue of bone growth in PAPPA2-deficient mice

Abstract

Aims

Recombinant human (rh) insulin-like growth factor-1 (IGF1) treatment has been shown to restore growth velocity and height in children harbouring pregnancy-associated plasma protein-A2 (PAPP-A2) mutations. The present study aimed to evaluate whether pharmacological modulation of the growth hormone (GH)/IGF1 system elicits sex-specific improvements in bone growth in a pappalysin 2 (Pappa2)-deficient mouse model of growth impairment, and to identify the associated signalling pathways.

Methods

Pappa2 knockout (Pappa2^ko/ko) mice of both sexes received daily treatment with rhGH, rhIGF1, or rhPAPP-A2 from postnatal day (PND) 5 to PND 35. The IGF1 system, bone remodelling factors, cannabinoid receptors, and intracellular signalling were analyzed in bone.

Results

At the onset of puberty, hormone-treated knockout Pappa2 (Pappa2^ko/ko) mice exhibited increased body length and weight, along with a higher femur weight-to-length ratio following rhGH or rhPAPP-A2 administration. In Pappa2^ko/ko females, hormone treatments reduced the bone resorption marker osteopontin, particularly following rhPAPP-A2, which also led to decreased expression of cannabinoid receptors (CNR1, CNR2, and GPR55). Compared to rhGH, both rhPAPP-A2 and rhIGF1 specifically decreased the IGF binding protein Igfbp5 and increased expression of the bone formation markers osteocalcin and collagen I in females. These effects were accompanied by elevated levels of phosphorylated signal transducer and activator of transcription 3 (STAT3). Pappa2 deficiency and hormone therapy were associated with numerous sex-specific differences in bone IGF1 binding, bone remodelling, cannabinoid receptors, and signal transducer and activator of transcription 3 (STAT3) signalling.

Conclusion

These findings support rhPAPP-A2 as a promising therapeutic candidate for mitigating postnatal bone growth retardation by increasing bone formation and reducing resorption through cannabinoid receptor modulation in a female-specific manner.

Cite this article: Bone Joint Res 2026;15(5):519–536.

Article focus

• The pregnancy-associated plasma protein-A2 (PAPP-A2) protease increases insulin-like growth factor-1 (IGF1) bioavailability and plays a key role in regulating postnatal growth.

• Recombinant human (rh) IGF1 treatment improves growth, bone mass, and bone microarchitecture in children with homozygous PAPP-A2 mutations.

• Animal models suggest that rhPAPP-A2 alleviates postnatal growth retardation caused by low IGF1 bioavailability in a female-specific manner.

Key messages

• A 30-day postnatal treatment with rh growth hormone(GH) or rhPAPP-A2 increased bone weight and length in pappalysin 2 (Pappa2)-deficient mice.

• In Pappa2-deficient females, rhPAPP-A2 reduced bone resorption in association with decreased expression of cannabinoid receptors.

• rhPAPP-A2-induced signal transducer and activator of transcription 3 (STAT3) phosphorylation was associated with bone formation and reduced IGF1-binding protein expression specifically in females.

Strengths and limitations

• Hormone therapy in conditions of low IGF1 availability was associated with bone-specific changes affecting IGF1 binding, intracellular signalling, bone remodelling, and cannabinoid receptor expression in females.

• rhPAPP-A2 potentials as a targeted therapy for postnatal growth failure and bone remodelling disturbances demand further studies, particularly in prepubertal girls.

Introduction

The growth hormone (GH)/insulin-like growth factor (IGF) axis plays a central role in skeletal growth and remodelling by coordinating bone formation and resorption.¹ Pituitary-derived GH promotes the production of IGFs, IGF-binding proteins (IGFBPs), and IGF-acid labile subunit (IGFALS) mainly in the liver but also in bone and other peripheral tissues. While free IGF1 has a serum half-life of approximately ten minutes,² its association with IGFBPs and IGFALS stabilizes the molecule and facilitates endocrine transport by forming binary or ternary complexes.^3,4 Since free IGF1 is the active form that binds to IGF receptors, its release from IGFBPs is essential for local bioavailability and function.⁵ This process is regulated by metalloproteases such as pregnancy-associated plasma protein-A2 (pappalysin 2 or PAPP-A2), which cleaves IGFBP3 and IGFBP5, thereby increasing IGF1 bioavailability at the tissue level.^6-9

Members of the GH/IGF axis exert both overlapping and independent effects on osteogenic cells, orchestrating postnatal growth.¹⁰ Osteoblasts and osteoclasts, the main cells responsible for bone formation and resorption, respectively, are integral for skeletal development and remodelling.^11,12 During development, chondrocytes secrete cartilage matrix, forming perichondrium, which will define the developing skeleton. Chondrocytes, in addition to osteoblasts and osteoclasts, respond directly to GH that stimulates proliferation and supports bone growth and maintenance.^13-16 Recent evidence also points to the involvement of G-protein-coupled receptors in regulating chondrocyte proliferation within the epiphyseal growth plate.¹⁷ The balance in the activity and development of chondrocytes, osteoblasts, and osteoclasts is essential for bone metabolism, and recent evidence shows that bone remodelling promoted by osteoblasts and osteoclasts is subject to fine post-transcriptional regulation processes,¹⁸ as well as the paracrine action of osteocytes.¹⁹

IGF1, via its canonical receptor in bone cells, activates several downstream pathways, including phosphoinositide 3-kinase (PI3K) and protein kinase B (AKT), which have significant implications for bone growth.²⁰ Emerging research in neuroskeletal biology has identified the endocannabinoid system (ECS) as a novel regulator of bone physiology. Through interactions with sympathetic nerve terminals, the ECS regulates both bone formation and resorption, suggesting that cannabinoid receptor ligands could represent therapeutic options for skeletal disorders.^21,22 Moreover, the identification of metabolic markers predictive of bone quality remains a key area of research,²³ and components of the ECS (cannabinoid receptors, ligands and enzymes involved in the synthesis and breakdown of cannabinoids) emerge as potential biomarkers of bone remodelling and metabolism.

In this study, we conducted a comparative investigation of the effects of three recombinant human (rh) proteins, including GH (rhGH), IGF1 (rhIGF1), and pappalysin 2 (rhPAPP-A2), administered daily from postnatal day (PND) 5 to PND35 in mice. We used a mouse model of PAPP-A2 gene deletion (Pappa2^ko/ko), which displays reduced longitudinal bone growth and impaired bone matrix composition compared to wild-type (Pappa2^wt/wt) mice.^24-26 We studied the effect of each treatment on the bone of Pappa2^ko/ko mice of both sexes, examining: 1) auxological parameters and skeletal proportions; 2) gene expression of IGFBPs; 3) key genes involved in bone development and remodelling; 4) cannabinoid receptor gene expression; and 5) phosphorylation downstream signalling proteins. Our aims were to evaluate therapeutic interventions that could mitigate the skeletal consequences of PAPP-A2 deficiency, and to further elucidate the molecular mechanisms by which GH/IGF1 signalling influences bone development.

Methods

Ethics statement

We have adhered to the ARRIVE guidelines and have included the ARRIVE checklist in the Supplementary Material. All procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council, 2011) and the European Directive 2010/63/EU on the protection of animals used for scientific purposes.^27,28 Efforts were made to minimize animal suffering and the number of animals used.

Animals

Adult male and female mice with constitutive Pappa2 gene deletion (Pappa2^ko/ko) and littermate controls (Pappa2^wt/wt) were generated as previously described.^24-26,29Pappa2 lines (on a C57BL/6 genetic background) were generated by breeding heterozygous transgenic mice. The mice were housed at the Animal centre for Experimentation of the University of Málaga, where they were subjected to a 12-hour reverse light/dark cycle (lights off at 8:00 am) with simulated dawn/sunset effects. Mice had access to standard rodent chow and water ad libitum, and environmental conditions were maintained at 23°C and 60% humidity. Genotyping of mice was performed via polymerase chain reaction (PCR) using ear-clip tissue, as previously described.³⁰ Body weight and length (including tail) were monitored every three days periodically from PND5 to PND35 (both included).

Treatment

rhGH protein (cat no. ABIN803848; antibodies-online, Germany), rhIGF1 protein (cat. no. 100-11; PeproTech; ThermoFisher Scientific USA), and rhPAPP-A2 protein (cat. no. ABIN7578676; antibodies-online GmbH) were dissolved in sterile saline (0.9% NaCl). The rhIGF1 and rhGH solutions were administered at a dose of 0.3 mg/kg/day, while rhPAPP-A2 was dispensed at a dose of 0.1 mg/kg/day. Saline, rhGH, rhIGF1, and rhPAPP-A2 were administered by daily interscapular subcutaneous injections, in a volume of 10 ml/kg body weight, for 30 days from PND5 to PND35. The dose selections were based on prior studies^29,31,32 and previous data. To prevent hypoglycemic episodes, the mice received hormone treatments during a non-fasting period, two hours after the onset of the dark cycle. Subsequently, all mice were killed 24 hours after the final administration, following a 12-hour fasting period.

Samples collection

At 35 days of age, the mice were anaesthetized by the administration of equithesin (3 ml/kg i.p.; chloral hydrate 2.1 g, sodium pentobarbital 0.46 g, MgSO₄ 1.06 g, propylene glycol 21.4 ml, ethanol (90%) 5.7 ml, H₂O 3 ml) and killed by decapitation. Femora and tibiae were collected, weighed, and length measured with a digital caliper. The bones were frozen on dry ice and stored at -80°C until used for protein and RNA isolation or kept at 4°C for histological analysis.

Histology and immunohistochemistry

For postfixtation of the tissue, the tibiae were immersed in paraformaldehyde solution (4% in phosphate buffered saline 0.1 M) for 48 hours with gentle agitation and then washed for 1 hour with distilled water. To properly perform bone dissection, decalcification, the complete removal of calcium salt and minerals, was first carried out. The bone was slowly decalcified using the calcium chelator ethylenediaminetetraacetic acid (EDTA). To do this, each femur was immersed in a 20% EDTA solution (pH adjusted to 7.4) and gently agitated with a rotary shaker at 4° and 100 rpm for two weeks, renewing the solution two to three times/week until the bone was soft and pliable. The samples were embedded in paraffin wax, and 10-µm-thick sections were obtained and mounted on slides for subsequent haematoxylin-eosin staining and immunohistochemistry. After dewaxing, the sections were washed several times with 0.1 M phosphate-buffered saline (PBS, pH 7.3) and then incubated in sodium citrate (50 mM, pH 9) for 15 minutes at 80°C for antigen retrieval. After several washes in PBS, the sections were incubated in 3% hydrogen peroxide (H₂O₂) for 20 minutes to inactivate endogenous peroxidase and then blocked with 10% donkey serum in PBS containing 0.1% NaN₃ and 0.2% Triton X-100 (Merck KGaA, Germany) for one hour. They were then incubated with the primary antibodies anti-cannabinoid receptor (CB) 1 (diluted 1:100; ABR, cat. no. PA1-745, lot. no. 424-121) and anti-CB2 (diluted 1:100; ABR, cat. no. PA1-746A, lot. no. 452-114) for 48 hours at 4°C, renewing the primary antibody solution 24 hours after incubation began. Tissue sections incubated without primary antibodies served as negative controls. The sections were then washed in PBS and incubated in a biotinylated donkey, anti-rabbit secondary antibody dilution (1:500) for one hour (GE HealthCare, USA). The sections were washed again in PBS and incubated with a 1:2,000 dilution of ExtrAvidin peroxidase (Sigma-Aldrich, USA) for one hour. After several washes in PBS, immunolabelling was revealed by exposure to 0.05% diaminobenzidine (DAB; Sigma-Aldrich), 0.05% nickel ammonium sulphate and 0.03% H₂O₂ in PBS. After several washes in phosphate buffer (PB), the sections were mounted on slides, dehydrated in ethanol, rinsed with xylene, and coverslipped with Eukitt mounting medium (PanReac AppliChem, Spain). High-resolution digital photomicrographs of the bone slices were taken under the same light and brightness/contrast conditions by an Olympus BX41 microscope equipped with an Olympus DP70 digital camera (Olympus Europa, Germany).

RNA isolation and RT-qPCR analysis

Frozen femora were pulverized using a Qiagen TissueLyser II sample disruptor (Qiagen, Germany). Total RNA was extracted following the TRIzol method according to the manufacture’s instruction (Gibco; Thermo Fisher Scientific, USA). Total RNA was quantified using a spectrophotometer to ensure A260/280 ratios of 1.8 to 2.0. Following the reverse transcript (RT) reaction from 1 μg of RNA, real-time RT polymerase chain reaction (PCR) was performed in a CFX96TM Real-Time PCR Detection System (Bio-Rad, USA) using specific primer probe sets (Supplementary Table i) of TaqMan Gene Expression Assays (Thermo Fisher Scientific) and the fluorescein amidite (FAM) dye label format for TaqMan Gene Expression (Thermo Fisher Scientific). To analyze osteopontin (Spp1) and Wnt1-inducible signaling pathway protein 1 (Wisp1) gene expression, gene-specific PCR primers (PrimePCR PreAmp for Probe Assay, Bio-Rad; Supplementary Table i) were used for unbiased preamplification of small amounts of complementary DNA (cDNA). Gene expression analysis was then performed using a second gene-specific PCR primer with the FAM dye label format (PrimePCR Probe Assay, Bio-Rad). Melting curve analysis was performed to ensure that only a single product was amplified. Cycle threshold (Ct) values were normalized relative to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) messenger RNA (mRNA) levels.

Protein extraction and western blot analysis

Total protein was extracted using 1 ml of ice-cold cell lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.25% deoxycholic acid, 1 mM EDTA, 1% Triton) supplemented with a cocktail of phosphatase (1 mM NaVO₄, 1 mM NaF) and protease inhibitors (cOmplete tablets, Merck, Germany). Western blot analyses were performed using various primary antibodies (Supplementary Table ii); 30 μg of protein extracted from the femora was electrophoresed on 4% to 12% Criterion XT Precast Bis-Tris gels (Bio-Rad, USA) for 30 minutes at 80 V and two hours at 150 V. Subsequently, proteins on the gels were transferred to nitrocellulose membranes for one hour at 80 V using a wet transfer kit (Bio-Rad). Total protein transferred to the membrane was visualized by Ponceau S red staining. The membrane was washed twice in tris-buffered saline with tween-20 (TBS-T) buffer (10 mM Tris-HCl, 150 mM NaCl, 0.1% Tween 20, pH 7.6) for five minutes and blocked with 2% bovine serum albumin (BSA; Roche, Germany) diluted in TBS-T buffer (BSA-TBST) for one hour at room temperature on a shaking platform. The membrane was then incubated with the respective primary antibody, diluted in 2% BSA-TBST, overnight at 4°C. Mouse β-actin was used as the reference protein. After several washes in TBS-T, the membrane was incubated with the corresponding horseradish peroxidase (HRP)-conjugated anti-rabbit or anti-mouse secondary antibody (Promega, USA) diluted (1:10,000) for one hour at room temperature. After washing three times with TBS-T for ten minutes, the membranes were incubated with Western Blotting Luminol Reagent (Santa Cruz Biotechnology, USA) for five minutes. The membrane-bound protein was then visualized by chemiluminescence (ChemiDoc Imaging System, Bio-Rad). To detect the phosphorylated and total forms of the proteins, stripping steps were performed by incubation with stripping buffer (2% SDS, 62.5 mM Tris HCL pH 6.8, 0.8% β-mercaptoethanol) for 30 minutes at 50°C. The bands were quantified by densitometric analysis using ImageJ software (National Institutes of Health, USA). Normalization was performed using the reference protein β-actin from the same membrane. Results were presented as the ratio of total protein expression to β-actin, and the ratio of phosphorylated protein expression to its total protein expression.

Statistical analysis

Analysis of variance (ANOVA) and multiple Pearson correlation analyses were performed with SPSS Statistics 26 software (IBM, USA). Values were expressed as mean (SEM) for each experimental group (n = 7 to 15 for auxological and mRNA analyses; n = 6 for protein analysis). The significance of difference within and between groups was analyzed by factorial between-subjects ANOVA and three-way ANOVA, whose factors were treatment (rhGH, rhIGF1 or rhPAPP-A2 vs saline), genotype (Pappa2^wt/wt vs Pappa2^ko/ko), sex (males vs females), and/or age (from PND5 to PND35), followed by univariate analysis (e.g. genotype × treatment in each timepoint), as well as Tukey’s post-hoc test for multiple comparisons. Changes over time were analyzed in each sex, genotype and treatment separately. Given the sex differences found in this experiment, a two-way ANOVA (genotype and treatment as factors) was also performed. No data points were excluded in the analyses. A p-value of less than 0.05 was considered statistically significant.

Results

Hormone treatment differentially reversed the auxological reduction of Pappa2^ko/ko mice

Body weight and length of hormone-treated Pappa2^ko/ko mice of both sexes showed a gradual increase over time compared to the lower values that characterize this growth-deficient mouse model, and this was treatment dependent (Figure 1). Overall, regardless of sex or genotype, effects on body weight and length were more significant in response to rhGH compared to rhIGF1 or rhPAPP-A2, while effects in rhIGF1-treated mice were more notable than in those treated with rhPAPP-A2 (Figure 1). Treatment effects on body weight and length were also more significant in males than in females (for additional statistical analysis see Supplementary Tables iii to v). By the end of the study, measurements of the hormone-treated Pappa2^ko/ko mice were not statistically different from the control saline-treated Pappa2^wt/wt mice. At the end of all experiments (PND35), significant differences were observed in the body weight (Figures 1a to 1f) and length (Figures 1g to 1l) of hormone-treated Pappa2^ko/ko mice of both sexes.

Fig. 1.

Effects of rhGH, rhIGF1 and rhPAPPA2 treatments on growth curves of pappalysin 2 (Pappa2^wt/wt) and Pappa2^ko/ko mice (males and females). Increasing body weight in a) to c) males and d) to f) females after treatment from postnatal day (PND)5 to PND35. Increase in body length in g) to i) males and j) to l) females after treatment from PND5 to PND35. Data are represented as mean (standard error of the mean) (n = 7 to 15 per experimental group). Univariate analysis (genotype × treatment in each timepoint) after factorial between-subjects analysis of variance: *p < 0.05, **p < 0.01, ***p < 0.001. See Supplementary Tables iii to v for additional statistical results. ko, knockout; rh, recombinant human; rhGH, rh growth hormone; rhIGF1, rh insulin-like growth factor-1; rhPAPPA2, rh plasma protein-A2; wt, wild-type.

An overall effect of treatment resulted in increased body weight in male Pappa2^wt/wt mice treated with rhGH from PND5, but no effects on body weight were found on the control mice when treated with rhIGF1 or rhPAPP-A2. An overall effect of treatment also resulted in increased body weight in male Pappa2^ko/ko mice treated with rhIGF1 or rhPAPP-A2 from PND5 onwards, and in the male Pappa2^ko/ko mice treated with rhGH from PND17 onwards, compared with the control mice. Univariate analysis (genotype × treatment in each timepoint) of body weight revealed a rhGH effect from PND17 to PND35, a rhIGF1 effect from PND23 to PND35, and a rhPAPP-A2 effect from PND32 to 35 (Figures 1a and 1b).

In female Pappa2^wt/wt mice, there was also an increase in body weight in response to rhGH, rhIGF1 and rhPAPP-A2 starting from PND5. In addition, there was an overall effect of treatment in female Pappa2^ko/ko mice resulting in increased body weight in those treated with rhGH or rhPAPP-A2 from PND5 onwards, and in female Pappa2^ko/ko mice treated with rhIGF1 from PND23 onwards. Univariate analysis (genotype × treatment in each timepoint) of body weight revealed a rhGH effect from PND8 to PND35, and rhIGF1 and rhPAPP-A2 effects at PND35 (Figures 1d to 1f).

An overall treatment effect resulted in increased body length in male Pappa2^wt/wt mice treated with rhGH or rhPAPP-A2 from PND5, and in the male Pappa2^wt/wt mice treated with rhIGF1 from PND20. Interestingly, an overall treatment effect also resulted in increased body length in the rhPAPP-A2-treated male Pappa2^ko/ko mice from PND5, but no treatment effects on body length were found when male Pappa2^ko/ko mice were treated with rhGH or rhIGF1. Treatments with rhGH or rhIGF1 only affected body length in male Pappa2^ko/ko mice after PND29. Univariate analysis (genotype × treatment in each timepoint) of length weight revealed a rhGH effect from PND8 to PND35, a rhIGF1 effect from PND23 to PND35, and a rhPAPP-A2 effect from PND32 to 35 (Figures 1g and 1h).

There was a treatment effect on body length in female Pappa2^wt/wt mice, with rhGH, rhIGF1 and rhPAPP-A2 from PND5. Treatment with rhIGF1 or rhPAPP-A2 affected body length in female Pappa2^ko/ko mice starting at PND14 and PND8. Univariate analysis (genotype × treatment in each timepoint) of length weight revealed effect of rhGH, rhIGF1 and rhPAPP-A2 at PND 35 (Figures 1j to 1l). See Supplementary Tables iii to v for additional statistical results.

Overall effects of age, genotype, and sex were found on body weight and length independently of treatment, except for the absence of a sex effect on body length in rhPAPPA2-treated mice (Supplementary Table v). Many interactions between factors (age, treatment, genotype, and sex) were also found, highlighting the significant interaction between treatment and sex and the lack of interaction between treatment and age in the body weight and length of rhPAPPA2-treated mice (Supplementary Table v). These results indicated that the differential effects of rhPAPP-A2 on body weight and length are sex-dependent, and that there is a sustained effect of rhPAPP-A2 on body weight and length throughout the prepubertal period. In addition, interactions between treatment, sex, genotype, and/or age were only found on body length in the rhPAPPA2-treated mice (Supplementary Table v).

Hormone treatment modified skeletal proportions in Pappa2^ko/ko mice

When skeletal proportions were analyzed (see Supplementary Table vi for additional data), interactions between genotype, sex, and treatment were found in the ratio of femur weight to body weight (FBWR) and the ratio of tibia weight to body weight (TBWR) of mice treated with rhGH, rhIGF1 and rhPAPP-A2, the ratio of femur length to body length (FBLR) of mice treated with rhGH and rhPAPP-A2, and the ratio of femur weight to femur length (FWLR) and the ratio of tibia length to body length (TBLR) of those treated with rhPAPP-A2. See Supplementary Tables viiA-C for complete statistical data. These results indicated that the effects of rhPAPP-A2 treatment on bone parameters were sex- and genotype-dependent. When males and females were analyzed separately, interactions between treatment and genotype were found in FBLR and TBLR after hormone treatments in males (Supplementary Tables viiD-F). However, in females, interaction between treatment and genotype were only found in FBWR after hormone treatments, as well as FWLR after rhIGF1 and rhPAPP-A2 treatments (Supplementary Tables viiG-I). Interestingly, treatment effects on FWLR were significant in male and female mice treated with rhGH and rhPAPP-A2 (Supplementary Tables viiA-I). Effects of rhPAPP-A2 treatment on bone parameters were more evident in males than females (Supplementary Tables viiF and I).

When analyzing the bone parameters by pairwise comparison, higher values of FBWR (Figure 2a), TBWR (Figure 2b), FBLR (Figure 2c), and TBLR (Figure 2d) were observed in the saline-treated male Pappa2^ko/ko mice compared to respective male Pappa2^wt/wt mice. In contrast, lower values in FWLR were found in saline-treated Pappa2^ko/ko mice of both sexes (Figure 2e).

Fig. 2.

Effects of recombinant human growth hormone (rhGH), rh insulin-like growth factor-1 (rhIGF1), and rh plasma protein-A2 (rhPAPPA2) treatments on auxological parameters and skeletal proportions of pappalysin 2 (Pappa2^wt/wt) and Pappa2^ko/ko mice (males and females). a) Femur-to-body and b) tibia-to-body weight ratios. c) Femur-to-body and d) tibia-to-body length ratios. Ratios of e) weight-to-length of femur and f) tibia. Data are represented as mean (standard error of the mean) (n = 7 to 15 per experimental group). Tukey´s post hoc test: treatment vs saline: *p < 0.05, **p < 0.01, ***p < 0.001; Pappa2^wt/wt vs Pappa2^ko/ko: ^#p < 0.05, ^##p < 0.01, ^###p < 0.001; male vs female: ^$p < 0.05, ^$$p < 0.01, ^$$$p < 0.001. See also Supplementary Tables vi and vii. ko, knockout; wt, wild-type

Regarding pairwise comparison, it should be highlighted that higher values of most bone parameters (FBWR, TBWR, FBLR, TBLR, and FWLR) were observed in male Pappa2^wt/wt mice treated with rhPAPP-A2 compared to controls (Figures 2a to 2f). However, this was not observed in female Pappa2^wt/wt mice treated with rhPAPP-A2. In contrast, rhPAPP-A2 reduced the values of TBWR, FBLR, and TBLR in male Pappa2^ko/ko mice (Figures 2b to 2d). Interestingly, higher values of FWLR were specifically observed in Pappa2^ko/ko mice of both sexes treated with rhPAPP-A2 (Figure 2e), suggesting greater bone density induced by rhPAPP-A2 treatment. Overall, similar results to rhPAPP-A2 were obtained with rhGH treatment. No remarkable changes were found in the ratio of tibia weight to tibia length (TWLR) (Figure 2f).

Hormone treatment modified the expression of bone IGF1-binding proteins

An interaction between genotype, sex, and treatment was found on Igfals gene expression after rhGH treatment only (Supplementary Table viiiA). Interactions between treatment and sex were found on Igfbp5 and Igfals expression after rhGH treatment, and Igfals expression after rhIGF1 treatment (Supplementary Tables viiiA and B), suggesting that the effects of hormone treatment on Igfals and/or Igfbp5 were sex-dependent. In addition, interaction between treatment and genotype was found on Igfbp5 expression after rhGH, rhIGF1 and rhPAPP-A2 treatment (Supplementary Tables viiiA-C), suggesting that the effects of hormone treatment on Igfbp5 were Pappa2-dependent. Interestingly, treatment effects on Igfbp5 and Igfals expression were observed in the femur of mice treated with rhIGF1 and rhPAPP-A2, but not rhGH (Supplementary Tables viiiA-C). Indeed, Igfbp5 expression was higher in the femur of Pappa2^wt/wt mice of both sexes after rhIGF1 and rhPAPP-A2 treatment (Figure 3a). In contrast, Igfbp5 expression was lower in the femur of female Pappa2^ko/ko mice after rhIGF1 and rhPAPP-A2 treatment. In addition, rhIGF1 and rhPAPP-A2 increased Igfals expression in the femur of male mice of both genotypes (Figure 3b).

Fig. 3.

Effects of recombinant human growth hormone (rhGH), rh insulin-like growth factor-1 (rhIGF1), and rh plasma protein-A2 (rhPAPPA2) treatments on relative messenger RNA (mRNA) levels of a) Igfbp5, b) Igfals,c) osteocalcin, d) osteopontin, e) type I collagen, and f) Wisp1 in the femur of pappalysin 2 Pappa2^wt/wt and of Pappa2^ko/ko mice (male and female). Data are represented as mean (standard error of the mean) (n = 6 to 9 per experimental group). Tukey’s post hoc test: treatment vs saline: *p < 0.05, **p < 0.01, ***p < 0.001; Pappa2^wt/wt vs Pappa2^ko/ko: ^#p < 0.05, ^##p < 0.01, ^###p < 0.001; male vs female: ^$p < 0.05, ^$$p < 0.01, ^$$$p < 0.001. See also Supplementary Tables i and viii. ko, knockout; wt, wild-type.

Hormone treatment modified the expression of bone remodelling regulators

Interactions between genotype, sex, and treatment were found on osteopontin expression after rhIGF1 and rhPAPP-A2 treatment (Supplementary Tables viiiA-C), and on osteocalcin and Wisp1 expression only after rhIGF1 treatment (Supplementary Table viiiB). Interactions between treatment and genotype were also found on osteopontin expression in male mice treated with rhGH (Supplementary Table viiiD) and osteocalcin expression in female mice treated with rhGH (Supplementary Table viiiG). When the sexes were analyzed separately, interactions between treatment and genotype were observed in osteocalcin, osteopontin, collagen I, and Wisp1 expression in male mice treated with rhIGF1, but not in rhIGF1-treated female mice, suggesting that the sex-specific effect of rhIGF1 on bone remodelling regulators is Pappa2-dependent (Supplementary Tables viiiE and H). Similarly, an interaction between treatment and genotype was also found on osteopontin expression after rhPAPP-A2 treatment in males, but not females (Supplementary Tables viiiF and I). Treatment effects on osteocalcin, osteopontin, and collagen I expressions were prevalent in mice treated with rhIGF1 and rhPAPP-A2, but not rhGH, in a sex- and genotype-independent manner (Supplementary Tables viiiA-C). Thus, increased osteocalcin and decreased osteopontin expression was observed in the femur of male Pappa2^wt/wt mice and female Pappa2^ko/ko mice after rhIGF1 and rhPAPP-A2 treatment (Figures 3c and 3d). The effects of rhGH on osteopontin expression were similar to rhIGF1 and rhPAPP-A2 treatment. However, increased collagen I expression was observed in Pappa2^ko/ko mice of both sexes treated with rhIGF1 and rhPAPP-A2, but not after rhGH treatment (Figure 3e). Treatment with rhIGF1 induced a specific decrease in Wisp1 expression in the femur of male Pappa2^ko/ko mice (Figure 3f).

Localization and modulation of cannabinoid receptors in bone tissue

Haematoxylin and eosin staining confirmed the integrity of tibia tissue and delineated the regions of proliferation, growth, and calcification in the epiphysis, metaphysis, and diaphysis, respectively (Figure 4a). Immunohistochemical expression of the cannabinoid receptors CB1 and CB2 was found in specific tibia regions in mice of both sexes and genotypes (Figures 4b and 4c); this was not observed in the negative controls (data not shown). The results showed the presence of CB1 and CB2 receptors in the haematopoietic stem cells of the bone marrow throughout metaphysis and diaphysis. Resting chondrocytes in the epiphysis and proliferating/hypertrophic chondrocytes in the growth plate also expressed both cannabinoid receptors, as did cells of the calcified bone in the tibia diaphysis (Figures 4b and 4c).

Fig. 4.

Histology of tibia tissue of pappalysin 2 (Pappa2^wt/wt) and Pappa2^ko/ko mice. a) Haematoxylin-eosin staining of tibia tissue with a schematic diagram of ossification showing epiphysis (resting cartilage with chondrocytes), growth plate (proliferating cartilage with chondrocytes), metaphysis (ossification zone with mixture of bone matrix and bone marrow), and diaphysis (calcified bone with osteocytes and osteoclasts which surrounds bone marrow). b) and c) Immunohistochemistry for cannabinoid receptor (CB)1 and CB2 revealed positive labelling of both cannabinoid receptors in resting chondrocytes (green arrows) in the epiphysis, proliferating/hypertrophic chondrocytes (purple arrows) in the growth plate, bone matrix/marrow cells in the metaphysis (blue arrows), and occasionally, calcified bone cells in the diaphysis (red arrows), in untreated mice of both sexes and genotypes. ko, knockout; wt, wild-type.

Interactions between genotype, sex, and treatment were found on the gene expression of the cannabinoid receptors CB1 (Cnr1) and CB2 (Cnr2), as well as G protein-coupled receptor 55 (Gpr55) (Figure 5), in the femur of rhGH-, rhIGF1-, and rhPAPPA2-treated mice (Supplementary Tables ixA-C). In addition, a treatment effect on CNR2 expression was found in the femur in response to rhGH, rhIGF1 and rhPAPPA2 (Supplementary Tables ixA-C), while treatment effects on CNR1 and GPR55 expression were only found in the femur of rhGH-treated mice (Supplementary Table ixA).

Fig. 5.

Effects of recombinant human growth hormone (rhGH), rh insulin-like growth factor-1 (rhIGF1), and rh plasma protein-A2 (rhPAPPA2) treatments on relative messenger RNA (mRNA) levels of cannabinoid receptor (Cnr)1, Cnr2 and Gpr55 in the femur of pappalysin 2 (Pappa2^wt/wt) and Pappa2^ko/ko mice (male and female). Data are represented as mean (standard error of the mean) (n = 6 to 9 per experimental group). Tukey’s post hoc test: treatment vs saline: *p < 0.05, **p < 0.01, ***p < 0.001; Pappa2^wt/wt vs Pappa2^ko/ko: ^#p < 0.05, ^##p < 0.01, ^###p < 0.001; male vs female: ^$p < 0.05, ^$$p < 0.01, ^$$$p < 0.001. See also Supplementary Tables i and ix. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; Gpr55, G protein-coupled receptor 55; ko, knockout; wt, wild type.

When the sexes were analyzed separately, interactions between treatment and genotype were found in Cnr2 and Gpr55 expression in males after rhGH and rhPAPP-A2 treatment (Supplementary Tables ixD-F). Treatment effects on Gpr55 expression were only observed in rhIGF1- and rhPAPPA2-treated males. In females, interactions between treatment and genotype, as well as treatment effects on Cnr1, Cnr2, and Gpr55 expression, were observed in females treated with rhGH, rhIGF1 and rhPAPP-A2 (Supplementary Tables ixG-I). Increased Cnr1, Cnr2, and Gpr55 expression was observed in the femur of saline-treated female Pappa2^ko/ko mice compared to respective female Pappa2^wt/wt mice (Figures 5a to 5c). In addition, increased Cnr2 and Gpr55, but not Cnr1, expression was found in the femur of saline-treated male Pappa2^ko/ko mice compared to respectively treated male Pappa2^wt/wt mice (Figure 5a to 5c). Interestingly, Cnr1, Cnr2, and Gpr55 expression was lower in the femur of female Pappa2^ko/ko mice treated with rhGH, rhIGF1, and rhPAPP-A2 compared to saline-treated female Pappa2^ko/ko mice (Figures 5a to 5c). No such changes were observed in males after hormone treatment. On the contrary, increased Cnr2 expression was observed in male Pappa2^wt/wt mice treated with rhPAPP-A2 (Figure 5b, p < 0.05), and increased Gpr55 expression was found in Pappa2^wt/wt mice of both sexes after hormone treatment (Figure 5c).

Hormone treatment modified bone intracellular signalling

Interactions between genotype, sex, and treatment were found on the protein levels of protein kinase A (PKA) and extracellular related kinase (ERK) when they were phosphorylated (p) on threonine 198 (pThr198-PKA) and tyrosine 204 (pTyr204-ERK), respectively, after rhGH treatment (Supplementary Table xA), in pTyr607-PI3K, pThr198-PKA, p-serine (Ser) 108-adenosine monophosphate-activated protein kinase β subunit (AMPKβ; pSer108-AMPKβ), and after rhIGF1 treatment (Supplementary Table xB), and in pSer473-AKT, pSer108-AMPKβ, and pTyr204-ERK2 after rhPAPP-A2 treatment (Supplementary Table xC). In addition, treatment effects on pSer473-AKT, pTyr705-signal transducer and activator of transcription 3 (STAT3), and pThr198-PKA were observed in the femur of mice treated with rhIGF1, as well as a treatment effect on pTyr204-ERK2 in the femur of mice treated with rhPAPP-A2. No treatment effects on protein levels were found in mice treated with rhGH. See Supplementary Table x and Supplementary Figures a and b for additional statistical results, and Supplementary Figure c for unedited blots.

When sexes were analyzed separately, interactions between treatment and genotype were found in pSer473-AKT, pThr198-PKA, and pTyr204-ERK2 in male mice treated with rhGH (Supplementary Table xiA), pTyr607-PI3K, pSer108-AMPKβ, pTyr705-STAT3, and pThr198-PKA in male mice treated with rhIGF1 (Supplementary Table xiC), and pSer108-AMPKβ, pThr198-PKA, and pTyr204-ERK2 in male mice treated with rhPAPP-A2 (Supplementary Table xiC). In females, interactions between treatment and genotype were observed in pSer473-AKT, pTyr705-STAT3, and pThr198-PKA in rhIGF1-treated mice (Supplementary Table xiD), as well as pSer473-AKT in rhPAPPA2-treated mice (Supplementary Table xiF). No significant interactions were found in phosphorylated protein levels in rhGH-treated female mice (Supplementary Table xiB). Among the multiple treatment effects observed, the treatment effects on pTyr204-ERK2 in males mice treated with rhGH, rhIGF1, and rhPAPP-A2 (Supplementary Tables xiA, C, E), as well as the treatment effects on pTyr705-STAT3 in female mice treated with rhIGF1 and rhPAPP-A2 (Supplementary Tables xiD, F) must be highlighted. See Supplementary Table xi for additional statistical results.

Regarding genotype effects on phosphorylated proteins in bone, increased pTyr607-PI3K levels and decreased pSer108-AMPKβ and pThr198-PKA levels were found in the femur of saline-treated male Pappa2^ko/ko mice compared to respective male Pappa2^wt/wt mice (Figures 6a, 6c and 6e), whereas increased pSer473-AKT and pThr198-PKA levels were found in the femur of saline-treated female Pappa2^ko/ko mice compared to respective female Pappa2^wt/wt mice (Figures 6b and 6e).

Fig. 6.

Effects of recombinant human growth hormone (rhGH), rh insulin‑like growth factor‑1 (rhIGF1), and rh plasma protein‑A2 (rhPAPPA2) treatments on the relative protein levels of key intracellular regulators of growth hormone (GH)/insulin‑like growth factor‑1 (IGF1) signalling pathways in the femur of pappalysin 2 (Pappa2^wt/wt) and Pappa2^ko/ko mice (male and female). Phosphorylated forms of a) Tyr607-PI3K, b) Ser473-AKT, c) Ser108-AMPKβ, d) Tyr705-STAT3, e) Thr198-PKA, and f) Tyr204-ERK2. Data are represented as mean (standard error of the mean) (n = 6 per experimental group). Tukey’s post hoc test: treatment vs saline: *p < 0.05, **p < 0.01, ***p < 0.001; Pappa2^wt/wt vs Pappa2^ko/ko: ^#p < 0.05, ^##p < 0.01, ^###p < 0.001; male vs female: ^$p < 0.05, ^$$p < 0.01, ^$$$p < 0.001. See also Supplementary Figures a to c and Supplementary Tables ii, x, and xi. AKT, protein kinase B; AMPKβ, adenosine monophosphate‑activated protein kinase β subunit; ERK2, extracellular related kinase 2; ko, knockout; p, phosphorylated; PI3K, phosphoinositide 3‑kinase; PKA, protein kinase A; Sal, saline; STAT3, signal transducer and activator of transcription 3; Thr, threonine; Tyr, tyrosine; wt, wild‑type.

Regarding treatment effects on phosphorylated proteins in bone, rhIGF1 treatment was specifically associated with increased pTyr607-PI3K levels and decreased pSer108-AMPKβ and pThr198-PKA levels in male Pappa2^wt/wt mice compared to saline-treated Pappa2^wt/wt males (Figures 6a, 6c and 6e). rhPAPP-A2 also increased pTyr204-ERK2 levels in male Pappa2^wt/wt mice (Figure 6f). rhGH, rhIGF1, and rhPAPP-A2 decreased pSer473-AKT levels in male Pappa2^ko/ko mice (Figure 6b) and rhIGF1 and rhPAPP-A2 also increased pSer108-AMPKβ and pTyr204-ERK2 levels (Figures 6c and 6f). In females, rhIGF1 and rhPAPP-A2 also increased pSer473-AKT and pThr198-PKA levels in Pappa2^wt/wt mice (Figures 6b and 6e), while rhIGF1 and rhPAPP-A2 also increased pTyr705-STAT3 levels (Figure 6d). Representative blots are shown in Figure 7. Unedited blots are shown in Supplemetary Figure c.

Fig. 7.

Representative immunoblots of the proteins listed in (Figure 6). AKT, protein kinase B; AMPKβ, adenosine monophosphate‑activated protein kinase β subunit; ERK2, extracellular related kinase 2; ko, knockout; MW, molecular weight; p, phosphorylated; Pappa2, pappalysin 2; PI3K, phosphoinositide 3‑kinase; PKA, protein kinase A; rhGH recombinant human growth hormone; rhIGF-1, rh insulin‑like growth factor‑1; Sal, saline; STAT3, signal transducer and activator of transcription 3; Thr, threonine; Tyr, tyrosine; wt, wild‑type.

Specific correlations between skeletal proportions and bone gene/protein expression are genotype- and treatment-dependent

We analyzed how Pappa2 deletion and hormone treatment affect the relationships between bone-to-body growth and bone gene/protein expression of IGF1-binding factors, bone remodelling regulators, cannabinoid receptors, and intracellular signalling regulators (Figure 8). To this end, multiple correlations were carried out, and treatments and sexes were analyzed separately. Increased body weight correlated with higher levels of: 1) pSer108-AMPKβ in the femur of male mice treated with rhGH, rhIGF1 and rhPAPP-A2; 2) pTyr204-ERK2 in the femur of male mice treated with rhIGF1 and rhPAPP-A2; and 3) osteopontin expression in the femur of male mice treated with rhGH (Figure 8c). In addition, increased body weight correlated with lower expression levels of: 1) CNR2 in the femur of male mice treated with rhGH and rhPAPP-A2; 2) Igfbp5 in the femur of male mice treated with rhGH and rhIGF1; and 3) pSer473-AKT in the femur of male mice treated with rhIGF1 and rhPAPP-A2 (Figure 8c).

Fig. 8.

Pearson correlations of a) to f) body weight and g) to l) femur weight-to-length ratio with messenger RNA (mRNA) levels and protein phosphorylation in the femur of male and female pappalysin 2 (Pappa2^ko/ko) mice after recombinant human growth hormone (rhGH), rh insulin-like growth factor-1 (rhIGF1), and rh plasma protein-A2 (rhPAPPA2) treatments. Significance: *p < 0.05, **p < 0.01, ***p < 0.001. AKT, protein kinase B; AMPKβ, adenosine monophosphate-activated protein kinase β subunit; CB, cannabinoid receptors; p, phosphorlyated; ERK2, extracellular signal-regulated kinase 2; GRP55, G protein-coupled receptor 55; Igfals, insulin-like growth factor binding protein; Igfbp5, insulin-like growth factor-binding protein 5; ko, knockout; PKA, protein kinase A; PI3K, phosphoinositide 3-kinase; STAT3, signal transducer and activator of transcription 3; WISP1, Wnt1-inducible signaling pathway protein 1.

Treatment with rhGH also contributed to negative correlations between body weight and the gene expression of Gpr55, collagen I, and osteopontin (Figure 8a). It should be highlighted that increased body weight correlated with lower gene expression of Cnr1, Cnr2, and Gpr55 in the femur of female mice treated with rhGH, rhIGF1, and rhPAPP-A2 (Figure 8f). In addition, increased body weight correlated with lower Igfbp5 expression and higher pTyr705-STAT3 levels in the femur of female mice treated with rhIGF1 and rhPAPP-A2 (Figure 8f). Treatment of female mice with rhPAPP-A2 also resulted in a positive correlation between body weight and osteocalcin expression, and a negative correlation between body weight and osteopontin expression (Figure 8f).

Regarding bone growth, increased femur weight-to-length ratio correlated with: 1) higher levels of pSer108-AMPKβ and pThr198-PKA, as well as lower levels of pSer473-AKT in male mice treated with rhGH and rhPAPP-A2; 2) higher Igfals expression in male mice treated with rhIGF1 and rhPAPP-A2, and female mice treated with rhPAPP-A2. Interestingly, the correlation between femur weight-to-length ratio and pTyr705-STAT3 levels was negative in male mice and positive in female mice when they were treated with rhPAPP-A2. Similar to body weight in females, an increased femur weight-to-length ratio also correlated with lower gene expression of Cnr1, Cnr2, and Gpr55 in female mice treated with rhGH and rhPAPP-A2, as well as lower expressions of Igfals and Igfbp5, respectively (Figure 8l). Treatment of female mice with rhGH and rhIGF1 also resulted in negative correlations between femur weight-to-length ratio and pThr198-PKA (Figure 8k). In addition, similar to female body weight, rhPAPP-A2 treatment in female mice also resulted in a positive correlation between femur weight-to-length ratio and osteocalcin expression (Figure 8l).

A summary of the main effects of a) pappalysin 2 (Pappa2) deletion and b) administration of rhGH, rhIGF1, and rhPAPP-A2 for 30 days (PND5 to 35) on 1) auxology, 2) gene expression of IGF1-binding proteins, key regulators of bone remodelling, and cannabinoid receptors, and 3) protein phosphorylation of key downstream signalling pathways can be found in Figure 9 and Figure 10.

Fig. 9.

Schematic diagrams summarizing the effects of a) pappalysin 2 (Pappa2) deletion and b) administration of recombinant human growth hormone (rhGH), rh insulin-like growth factor-1 (rhIGF1), and rh plasma protein-A2 (rhPAPP-A2) for 30 days (postnatal day (PND) 5 to 35) on 1) auxology, 2) gene expression of insulin growth factor 1 (IGF1)-binding proteins, key regulators of bone remodelling, and cannabinoid receptors, and 3) protein phosphorylation of key downstream signalling pathways in the bone of male (blue arrows) and female (red arrows) Pappa2^ko/ko mice. The black lines indicate the process they activate or inhibit. The effect of genotype and treatment is indicated by an asterisk (*). AC, adenylate cyclase; AKT, protein kinease B; AMPKβ, adenosine monophosphate-activated protein kinase β subunit; ATP, adenosine triphosphate; cAMP, cyclic adenosine monophosphate; CB, cannabinoid receptor; ERK2, extracellular signal-regulated kinase 2; GHR, growth hormone receptor; Gpr55, G protein-coupled receptor 55; GSK3β, glycogen synthase kinase-3 β; Igfals, insulin-like growth factor binding protein; IGFBPS, insulin-like growth factor-binding protein 5; IRS1, insulin receptor substrate-1; Jak, janus kinase; PI3K, phosphatidylinositol 3-kinases; PKA, protein kinase A; STAT3, signal transducer and activator of transcription 3; Wisp1, Wnt1-inducible signaling pathway protein 1.

Fig. 10.

Schematic diagrams summarizing the effects of administration of recombinant human growth hormone (rhGH), rh insulin-like growth factor-1 (rhIGF1), and rh plasma protein-A2 (rhPAPP-A2) for 30 days (postnatal day (PND) 5 to 35) on 1) auxology, 2) gene expression of insulin growth factor 1 (IGF1)-binding proteins, key regulators of bone remodelling, and cannabinoid receptors, and 3) protein phosphorylation of key downstream signalling pathways in the bone of male (blue arrows) and female (red arrows) Pappa2^ko/ko mice. The black lines indicate the process they activate or inhibit. The effect of genotype and treatment is indicated by an asterisk (*). AC, adenylate cyclase; AKT, protein kinease B; AMPKβ, adenosine monophosphate-activated protein kinase β subunit; ATP, adenosine triphosphate; cAMP, cyclic adenosine monophosphate; CB, cannabinoid receptor; ERK2, extracellular signal-regulated kinase 2; GHR, growth hormone receptor; Gpr55, G protein-coupled receptor 55; GSK3β, glycogen synthase kinase-3 β; Igfals, insulin-like growth factor binding protein; IGFBPS, insulin-like growth factor-binding protein 5; IRS1, insulin receptor substrate-1; Jak, janus kinase; PI3K, phosphatidylinositol 3-kinases; PKA, protein kinase A; STAT3, signal transducer and activator of transcription 3; Wisp1, Wnt1-inducible signaling pathway protein 1.

Discussion

The use of animal models of low IGF1 bioavailability, such as Pappa2-deficient mice, has facilitated research into the diverse members of the hypothalamic-pituitary GH/IGF axis and their roles in postnatal growth retardation, as well as how treatment induces growth by restoring longitudinal bone growth and bone mineral density.^33-35 The present study corroborates the reduced length and body weight of Pappa2-deficient mice relative to their wild-type counterparts, consistent with prior studies conducted in mice^8,24,25,36 and humans.^33-35,37,38 Our findings demonstrate the differential effectiveness of one-month treatment with either rhGH, rhIGF1, or rhPAPP-A2 in rectifying auxological parameters, similar to improvements observed in other studies with rhGH and rhIGF1 therapies in humans, with some evidence of sex-dependent responses.^38-40 Notably, while rhGH treatment of wild-type mice resulted in significant increases in body weight and skeletal length, rhPAPP-A2 treatment had no such effect in wild-type animals. This indicates that rhPAPP-A2 appears to act as a context-dependent homeostatic regulator, selectively restoring skeletal integrity under low IGF1 bioavailability.

This study also provides evidence for the presence and differential expression of the cannabinoid receptors CB1, CB2, and GPR55 in the long bones of both Pappa2^wt/wt and Pappa2^ko/ko mice, irrespective of sex. Consistent with previous reports identifying these receptors in various bone cell types,^39-43 we observed immunohistochemical expression in bone marrow, osteocytes, and osteoclasts in the diaphysis, as well as in resting, proliferating, and hypertrophic chondrocytes of the growth plate. The cannabinoid system is known to play a crucial role in bone cell activity and remodelling, affecting bone mass regulation.^40,44 Activation of cannabinoid receptors promotes osteoclast-mediated bone resorption,^22,45 in a sex-dependent manner.^42,46 In this regard, we found higher Cnr2 and Gpr55 gene expressions in the femur of Pappa2^ko/ko mice of both sexes, but higher Cnr1 gene expression only in the femur of Pappa2^ko/ko females (Figure 9). However, rhPAPP-A2 treatment selectively reduced Cnr2 expression in the femur of Pappa2^ko/ko males compared to saline-treated males. While the mechanism remains unclear, this finding suggests that PAPP-A2 may modulate components of the endocannabinoid system, either directly or indirectly, to help support bone homeostasis, particularly under conditions of PAPP-A2 deficiency. The decreased Cnr2 expression observed in male Pappa2^ko/ko mice after rhPAPPA2 treatment was consistent with the observed improvement in auxological parameters, emphasizing its role in fine-tuning bone physiology. Moreover, in female Pappa2^ko/ko mice, treatment with rhGH, rhIGF1 and rhPAPPA2 led to consistent reductions in the expression of Cnr1, Cnr2, and Gpr55, reversing the elevated cannabinoid expression observed in saline-treated Pappa2^ko/ko females (Figure 10). These findings underscore the relevance of the cannabinoid system in regulating bone growth and its modulation by the three recombinant hormones used here, particularly in females.

We also examined the signalling pathways activated by the binding of GH and IGF1 to their respective receptors, given their pivotal role in bone development. Treatment with rhGH, rhIGF1, or rhPAPP-A2 reduced AKT phosphorylation in the bone of male Pappa2^ko/ko mice. Conversely, rhIGF1 and rhPAPP-A2 administration increased phosphorylation of AMPKβ and ERK2 (Figure 10). It is well established that IGF1 signalling typically activates a cascade involving IRS1, PI3K, ERK, and AKT, thereby regulating cell survival, protein synthesis, and energy use in bone.²⁰ Specifically, the PI3K/AKT pathway plays a critical role in regulating chondrocyte survival and apoptosis,⁴⁷ while AKT/ERK signalling is associated with rapid cellular proliferation during cartilage repair.⁴⁸ Notably, GH also triggers activation of ERK1/2 signalling.⁴⁹ In addition, AMPK activation has been shown to promote bone formation,⁵⁰ while PKA signalling positively regulates chondrogenesis initiation.⁵¹ In this study, we observed increased ERK2 and AMPKβ phosphorylation following rhIGF1 and rhPAPP-A2 treatment in male Pappa2^ko/ko mice, along with enhanced PKA phosphorylation after rhGH treatment. Although AKT phosphorylation was decreased, the activation of these alternative pathways may contribute to improved bone development. Given the cellular complexity of bone, these phosphorylation events likely reflect distinct responses across cell populations.

Female Pappa2^ko/ko mice treated with rhIGF1 and rhPAPP-A2 exhibited increased STAT3 phosphorylation, a noteworthy observation given the association of this protein to chondrogenesis and bone formation.⁴⁹ Similar to males, rhIGF1 treatment in females also led to a concurrent decrease in PKA phosphorylation, suggesting a shift in STAT3 activity. Interestingly, no discernible changes were noted in the phosphorylation of any proteins in female Pappa2^ko/ko treated with rhGH, even though GH is recognized to stimulate chondrocyte¹⁴ and osteoblast proliferation,¹³ through activation of Janus kinase (JAK)–signal transducer and the STAT signalling pathway,^49,52 and has been shown to have direct effects on primary chondrocytes, osteoblasts, and osteoclasts.^15,16,53 However, IGF1 receptor activation also stimulates STAT3 signalling through the JAK and AKT-driven signalling pathways,⁵⁴ and our findings in female Pappa2^ko/ko mice suggest that the increase in STAT3 phosphorylation is likely a result of increased IGF1 bioavailability, as the response was positive to rhIGF1 and rhPAPPA2 but not to rhGH.

The expression of IGFBPs at the bone level is regulated by GH/IGF1 signalling. Sex differences were found in Igfbp5 expression in the bone of Pappa2-deficient mice, which is consistent with previous findings following short-term rhIGF1 administration,²⁶ among others.^55,56 IGF-binding protein deficiencies are associated with short stature and reduced bone mineral density due to alterations in the regulation of circulating IGF1 levels,⁵⁶ potentially affecting mesenchymal stromal cell fate,⁵⁷ and osteoblastic proliferation.⁵⁸ Interestingly, increased Igfals expression was found in the bone of male Pappa2^ko/ko mice of both sexes following rhIGF1 and rhPAPP-A2 administration (Figure 9b). Conversely, female Pappa2^ko/ko mice had reduced Igfals after rhGH treatment and decreased Igfbp5 and Igfals that were maintained and even enhanced expression after rhIGF1 and rhPAPP-A2 administration. Moreover, opposite effects were found on Igfbp5 and Igfals expressions after rhGH administration in a sex-dependent manner. The response of Igfbp5 gene expression in female bone to repeated rhIGF1 administration for 30 days contrasts with previous findings with a single rhIGF1 dose,²⁶ suggesting a potential long-term feedback mechanism in response to rhIGF1 treatment.

The expression of key genes involved in bone development, osteocalcin, osteopontin, and type I collagen, was not modulated by rhGH treatment in the bone of male Pappa2^ko/ko mice (Figure 9b). Conversely, decreased expression of these three genes was observed in female Pappa2^ko/ko animals. Osteocalcin, type I collagen, and osteopontin secreted by osteoblasts are crucial for bone remodelling and development. Osteocalcin is essential for calcium deposition and crystal maturation, contributing to the structural integrity of the skeleton.^12,59-61 On the other hand, osteopontin, mainly expressed by osteoclasts, facilitates bone resorption during remodelling.^61,62 The lack of change or decrease in osteopontin expression in males, and the decrease in osteocalcin and type I collagen expression in females following rhGH treatment suggests a positive outcome in males, but potential negative effects on bone development in females. Similarly, decreased Wisp1 expression in rhIGF1-treated male Pappa2^ko/ko mice could hinder bone formation, although Wisp1 expression is also associated with bone protection, proliferation and regeneration against senescence and apoptosis of the chondrocyte.^63,64 There was a significant increase in the expression of osteocalcin, and a decrease in osteopontin, in female Pappa2^ko/ko mice treated with rhIGF1 and rhPAPP-A2, in addition to an increase in type I collagen gene expression in both male and female Pappa2^ko/ko mice, which aligns with the observed improvements in growth following treatment with rhIGF1 and rhPAPPA2, particularly in females.

In male Pappa2^ko/ko mice there were correlations between skeletal proportions and bone growth regulators after hormone treatments. Specifically, rhPAPP-A2 treatment resulted in positive correlations of the femur weight/length ratio with type I collagen and PKA and AMPKβ phosphorylation, but a negative correlation with AKT and STAT3 phosphorylation. On the contrary, in female Pappa2^ko/ko mice, in response to rhPAPPA2 treatment there was a positive correlation of the femur weight/length ratio with STAT3 phosphorylation and osteocalcin, but a negative correlation with CNR1, CNR2, and GPR55 expression. These correlations point to different mechanisms of action between the responses to rhPAPP-A2 and those of rhGH and rhIGF1, as well as significant sex-dependent variations.

It should be noted that, while PAPP-A2 is known to preferentially cleave IGFBP-3 and -5, IGFBP-4 still remains a potent inhibitor of IGF1 bioavailability.⁵ Although Igfbp4 expression was not assessed as it is not directly targeted by PAPP-A2, IGFBP-4 is known to regulate pubertal and adult skeletal growth in a sex-specific manner.^65,66 Thus, possible compensations in Igfbp4 expression lead by PAPP-A2 deficiency or recombinant treatment could further influence local IGF1 signalling. Future studies should address this potential mechanism to provide a more complete picture of the IGFBP landscape in hormone treatment for bone growth.

In summary, this study indicates that pharmacological modulation of the GH/IGF1 system promotes distinct, sex-specific effects on bone in Pappa2-deficient mice. While rhGH was effective in driving systemic recovery of body size, rhIGF1 and rhPAPP-A2 treatments specifically enhanced local bone formation markers (osteocalcin, collagen I) and STAT3 signalling, an effect that was largely observed in females. Moreover, we identified a novel interaction between the GH/IGF1 axis elements and the ECS, where recombinant treatment reversed the overexpression of cannabinoid receptors specifically in the bone of female mice. Relevant correlations between hormone treatments and molecular markers implicated in bone growth and signalling were observed, positioning rhPAPPA-2 as a targeted therapeutic candidate that mimics physiological restoration of local IGF1 bioavailability and structural support in a sex-dependent manner, thereby offering new insights into the complex mechanisms underlying growth regulation, with implications for growth disorders in humans.

Data Availability

All data generated or analyzed during this study are included in the published article and/or in the supplementary material.