Concentrations of Adrenocorticotropic Hormone (ACTH_1–24) Too Low to Effect Cortisol Enhance Osteogenesis In Vitro and In Vivo

Abstract

After finding that minimal amounts of adrenocorticotropic hormone (ACTH_1–24) prevented osteonecrosis in rabbits, we studied bone formation at nanomolar ACTH_1–24, in vivo in rabbits, and in vitro in human osteoblasts. ACTH_1–24 in rabbits at 0.6 μg/kg/day had no measurable effect on cortisol. Groups of five rabbits given 0.6 μg/kg/day of ACTH_1–24 enhanced trabecular bone in the rabbit femoral head relative to saline-treated controls (controls), increased bone volume/total volume (BV/TV) by micro-computed tomography, p<0.03. Xylenol orange and calcein labeling in vivo showed increased trabecular bone formation with 0.6 μg/kg/day of ACTH_1–24, p = 0.0089 versus controls. In contrast, the cortex of the femoral shaft was unaffected, BV/TV p>0.95 ACTH_1–24 versus controls. Bone marrow mRNA by PCR showed no change in osteoclast markers, and confirmed increased osteoblast markers, p<0.05. In vitro, ACTH_1–24 elevated expression of Collagen 1, alkaline phosphatase (ALP), osteocalcin (BGLAP), and RunX2 in human osteoblasts differentiated on polyethylene terephthalate (PET) membranes. Optimal response was at 10⁻⁹ to 10⁻¹² M. VEGF, VEGF receptors FLT-1 and FLK-1, and ACTH_1–24 receptors MC2R were upregulated at 10⁻¹² M ACTH_1–24. Pathway analysis included increased BMP2, Smad1, Wnt-1, β-Catenin and TGF-β pathways. Because bone-forming osteoblasts are metabolically highly active, we studied mRNA expression of mitochondrial complex 1 (NDUFA5, NDUFS2, NDUFB1, NDUFB6) members with key roles in energy production. This increased at 10⁻¹² M ACTH_1–24. An ELISA for mitochondrial complex 1 activity showed maximum activity at 10⁻⁹ M and high activity at 10⁻¹² M ACTH_1–24. Thus, long-term very low dose ACTH_1–24 increases bone formation in vivo and in vitro.

NEW & NOTEWORTHY

This is the first study to assess directly the effects of very low concentrations of ACTH_1–24, picomolar to micromolar, on regulation of bone growth. It shows effects on bone mass in the femoral heads and by labeling bone formation in rabbits in vitro, and effect on key proteins related to bone growth mechanisms in cell preparations of human bone forming cells in vitro.

Graphical Abstract

INTRODUCTION

Bone formation, remodeling and turnover are regulated by interaction of neurological, endocrine, and autocrine or paracrine mechanisms. Recent data provide evidence that pituitary hormones, including ACTH_1–24, and parathyroid hormone (PTH), are endocrine regulators of bone metabolism [1–3]. In the last decade, hundreds of animal and clinical studies showed an important role for PTH treatment in human bone formation and stated that it can be viewed as a potential treatment to increase the amount of new calcified bone [4], while for the role of ACTH_1–24 in bone formation there are only animal studies and no clinical trials. While ACTH_1–24 has long been used in the treatment of gout and is an alternative therapeutic option, especially in difficult-to-treat patients [5], it is also used in therapy of the nephrotic syndrome, particularly for patients who fail conventional immunosuppressive therapy [6], and as a treatment of children with infantile spasms [7], there are no evidence of using ACTH_1–24 in patients with osteoporosis.

We showed that ACTH_1–24 protects against glucocorticoid-induced osteonecrosis of rabbit bone in vivo [8], and in that study [8] initial work on osteoblast proliferation and differentiation were included using MC3T3 osteoblast-like cells. In another study in vitro, we found that two-hour 10⁻⁹ M ACTH_1–24 treatment enhanced human osteoblast differentiation in 6-well plates [9]. An in vitro study [4] showed that 10⁻⁸ M ACTH_1–24 enhanced human skeletal stem cell differentiation to osteoblasts by alkaline phosphatase (ALP) activity, alizarin red staining, and real time PCR for ALP and osteocalcin (Bglap), consistent with our results. On the other hand, Sato et al. reported that ACTH_1–24 at 10⁻¹⁰ M inhibited ALP activity and expression of Bglap in murine primary osteoblasts isolated from calvaria collected from newborn mice, and in MC3T3-E1 cells [10]. Our experience with mice (not shown here) suggests that mice do not model well effects of ACTH_1–24 in humans or rabbits, although this topic might require additional study.

Here we evaluated how very low dose ACTH_1–24 treatment, 0.6 μg/kg/day injected subcutaneously for 5 weeks, affects bone formation in rabbits in vivo, and effects of picomolar to nanomolar ACTH_1–24 on human osteoblast differentiation and mineralization, using cells differentiated in vitro on polyethylene terephthalate (PET) membranes [11].

MATERIALS AND METHODS

Animals, Reagents, and Treatment.

New Zealand White Rabbits, males, ~1 year old, were used. The animal protocol was approved by the Pittsburgh Veteran Center Animal Care and Use Committee (the VA IRB/ IACUC). There were no rabbits excluded, none died or were ill, and there was no attrition or drop out. Subjects were allocated to treated and untreated groups by alteration. There was no selection and animals assigned to treated and untreated were from the same animal groups received. Thus, assignment to groups was completely random. Controls (saline only) or ACTH_1–24 (ACTH; Cosyntropin) treated rabbits had average weight of 4.2 kg with SD 0.4 kg, with variability of ~10% in weight of each group. In preliminary work, two rabbits were evaluated for short term (4 hour) effects of ACTH_1–24, treated with 0.6 μg/kg, 1 μg/kg, and 30 μg/kg of rabbit weight ACTH_1–24 by injection. Subsequently, five rabbits each were treated with saline only or 0.6 μg/kg/day of ACTH_1–24, Monday through Friday. All analysis was blinded to investigators. This included that sections cut were done by a different person than did the analysis and stored by groups without identification. The rabbits were treated by the separate person. Code was not used in the study. Original data was not deposited but will be provided on request. Digital object identifiers were not used and are not relevant.

Materials.

Reagents and media were from Thermo-Fischer or as stated. Normal human osteoblast precursors (NHOst), and proliferation media were from Lonza (Basel, Switzerland). Aliquots were from ages 27–33 years old with two male and two female sex specimens; all lots of cells showed similar results with minor quantitative variation (cf. Figs 7 and 8). Interestingly, direct comparison of males and females (blue and red dots in Fig 8) showed no significant difference. Cells were differentiated into human osteoblasts (HOB) for five weeks in osteogenic medium on polyethylene terephthalate membrane inserts with 0.4 μm perforations, to optimize epithelial-like matrix synthesis and mineral transport [11]. Osteogenic differentiation medium was Dulbecco’s modified essential medium, low glucose (1.0 g/l) with 10% fetal bovine serum, antibiotics-antimycotics, supplemented with 10 mM 2-glycerol phosphate and 30 μg/ml ascorbic acid. Calcium in media was adjusted to 2 mM. Differentiation media were replaced every 3 days.

Figure 7. The effect of low-dose ACTH_1–24 activity of Mitochondrial complex 1 in osteoblasts.

Mitochondrial complex 1 plays a key role in energy production, and is also upregulated by ACTH_1–24, with maximum activity at 10⁻⁹ M (green trace) and high activity at 10⁻¹²M ACTH_1–24 (red trace), compared to Saline (control) treatment (black trace) and background (lilac trace) over 30 minutes by ELISA. Compared with Fig 4C, showing variable results for individual proteins, but in a consistent range overall with maximum from 10⁻⁹ to 10⁻¹² M ACTH_1–24. We did not include higher concentrations here. In Fig 8C we included 1 μM, at which concentration all mitochondrial protein effects dropped relative to 10⁻⁹ M. Effects at 1 μM could be interesting, exceeding concentrations of interest in long-term very low dose ACTH_1–24.

A and B. Two separate osteoblast isolates were grown and differentiated on PET membranes are shown. Mean and SD for dual measurements in each of two isolates are calculated; individual points cannot be plotted due to overlap in this case. Note that activity of ACTH_1–24 in A isolate (left) is higher than in the B isolate (right). This is in keeping with different maturation potential of different SSC isolates (see Discussion). Activity at 10⁻⁶ M was much lower, near the control (not illustrated).

Figure 8. A survey of expression by PCR in ACTH_1–24 treated human osteoblasts of proteins linked to osteoblast (maturation) differentiation pathways.

The classes of proteins are discussed in [27] with illustrations of major classes and roles in biochemistry of bone. Assayed proteins include Bone Morphogenetic Protein-2 (A), SMADs 1,2, and 3 (Suppressors of Mothers Against Decapentaplegic, transcription factors affecting osteoblast differentiation) (B-D), Transforming Growth Factor-beta (E), Wnt1(F), beta-Catenin (G), biglycan-19 (H), and integrins A5 and B1 (I-J). The results using cells from female donors are in red, from male donors in blue. Male to female differences were not significant.

RNA isolation and real-time PCR.

Total RNA was isolated by oligo dT affinity (RNeasy, Qiagen, Hilden, Germany); cDNA was synthesized with random hexamers and Moloney murine leukemia virus reverse transcriptase (Superscript III; Invitrogen). Real-time PCR used an MX3000P instrument (Stratagene, San Diego, CA) with SYBR green to monitor synthesis. Reactions were performed in duplicate in 25 μl reaction volumes with 12.5 μl of pre-mixed dye, NTPs, buffer, and polymerase (SYBR Green Master Mix, Stratagene), to which 250 nM primers and 1 μl of first strand cDNA were added. After 10 min at 95°C, the mixture was amplified in cycles of 30 sec at 95°C, 30 sec at 59°C, and 1 min at 72°C. Product specificity was verified by agarose gel electrophoresis. Product abundance relative to controls was calculated assuming linearity to log(initial copies) [12]. Rabbit and human primers are listed in Table 1.

Table 1.

Rabbit and Human PCR Primers

Rabbit Primers
ALP - alkaline phosphatase
Forward	5’-ACTGTGGACTACCTCTTG-3’
Reverse	5’-GGTCAGTGATGTTGTTCC-3’	Product size 76 bp
Col1A1 - collagen 1 alpha 1 subunit
Forward	5’-CACATGCGTGCAGAACGGCG-3’
Reverse	3’-CGCGTCTTCGGGGCAGACAG-5’	Product size 184 bp
BMP2 bone morphogenetic protein 2
Forward	5’-GCTTCGACGTCACCCCTGCC-3’
Reverse	3’-AGGGGGTGCCCCTTCCCATC-5’	Product size 211 bp
TRAP (ACPA5) tartrate resistant acid phosphatase
Forward	5’- GAAACCACGACCACATTGGC -3’
Reverse	3’- GTCACCGTGTCCAGCATGTA -5’	Product size 151 bp
BGLAP - bone gamma carboxyglutamate protein (osteocalcin)
Forward	5’- GCTCAGCCTTCGTGTCCA -3’
Reverse	3’- CCTGCCCGTCAATCAGTT -5’	Product size 75 bp
ATPa3 (V-ATPase subunit a3)
Forward	5’- CTGTTCGGCTACCTCGTGTT -3’
Reverse	3’- TAGGACGGACTGGACCACTT -5’	Product size 171 bp
ATPd2 (V-ATPase subunit d2)
Forward	5’- CGCTCAGCACCTTTTTCACC -3’
Reverse	3’- CTAGCGGTGTTTCCACCAGT -5’	Product size 207 bp
CTSK (cathepsin K)
Forward	5’- TCAAAGTACCCCCGTCTCGT -3’
Reverse	3’- TGAAAGCCCAACAGGAACCA -5’	Product size 150 bp
Human Primer
Col1A1 - collagen type I alpha 1 chain
Forward	5’-AGGGACACAGAGGTTTCAGTGGTT-3’
Reverse	5’- GCAGCACCAGTAGCACCATCATTT-3’	Product size 198 bp
RunX2 – RUNX family transcription factor 2
Forward	5’-TGGTTACTGTCATGGCGGGTA-3’
Reverse	5’-TCTCAGATCGTTGAACCTTGCTA-3’	Product size 101 bp
ALPL – alkaline phosphatase, biomineralization associated
Forward	5’-ACCACCACGAGAGTGAACCA -3’
Reverse	5’-CGTTGTCTGAGTACCAGTCCC -3’	Product size 79 bp
BGLAP - bone gamma-carboxyglutamate protein (Osteocalcin)
Forward	5’- CACTCCTCGCCCTATTGGC -3’
Reverse	5’- CCCTCCTGCTTGGACACAAAG -3’	Product size 112 bp
VEGF - vascular endothelial growth factor A
Forward	5’- AAGGAGGAGGGCAGAATCAT -3’
Reverse	5’- ATCTGCATGGTGATGTTGGA -3’	Product size 225 bp
FLT1 - fms related receptor tyrosine kinase 1 (VEGFR1)
Forward	5’- TTTGCCTGAAATGGTGAGTAAGG -3’
Reverse	5’- TGGTTTGCTTGAGCTGTGTTC -3’	Product size 117 bp
FLK1/KDR - kinase insert domain receptor (VEGFR2)
Forward	5’- GGCCCAATAATCAGAGTGGCA -3’
Reverse	5’- CCAGTGTCATTTCCGATCACTTT -3’	Product size 109 bp
MC2R – melanocortin 2 (adrenocorticotropic hormone) receptor
Forward	5’- GACTGTCCTCGTGTGGTTTTG -3’
Reverse	5’- GGCTGCCCAGCATATCAGAT -3’	Product size 166 bp
NDUFA5 – NADH: ubiquinone oxidoreductase subunit A5
Forward	5’- GAGAAGCTGGCTATGGTTAAAGCG -3’
Reverse	5’- CCACTAATGGCTCCCATAGTTTCC -3’	Product size 154 bp
NDUFB1 – NADH: ubiquinone oxidoreductase subunit B1
Forward	5’- GTCCCTATGGGATTTGTCATTGG -3’
Reverse	5’- CAGTTAGCCGTTCATCACTCTT -3’	Product size 60 bp
NDUFB6 – NADH: ubiquinone oxidoreductase subunit B6
Forward	5’- CCACAGAAGATGGGGCCTATG -3’
Reverse	5’- TCCAGACAGGTACAAGTACATGA -3’	Product size 133 bp
NDUFS2 – NADH: ubiquinone oxidoreductase subunit S2
Forward	5’- ACCCAAGCAAAGAAACAGCC -3’
Reverse	5’- AATGAGCTTCTCAGTGCCTC -3’	Product size 214 bp
BMP2 – bone morphogenetic protein 2
Forward	5’- ACTACCAGAAACGAGTGGGAA -3’
Reverse	5’- GCATCTGTTCTCGGAAAACCT -3’	Product size 113 bp
TGFβ – transforming growth factor, beta 1
Forward	5’- CTAATGGTGGAAACCCACAACG -3’
Reverse	5’- TATCGCCAGGAATTGTTGCTG -3’	Product size 209 bp
SMAD1 - SMAD family member 1
Forward	5’- TTGGCACAGTCTGTGAACCATGG -3’
Reverse	5’- GTAACATCCTGGCGGTGGTATTC -3’	Product size 116 bp
SMAD2 - SMAD family member 2
Forward	5’- GGGTTTTGAAGCCGTCTATCAGC -3’
Reverse	5’- CCAACCACTGTAGAGGTCCATTC -3’	Product size 149 bp
SMAD3 - SMAD family member 3
Forward	5’- TGAGGCTGTCTACCAGTTGACC -3’
Reverse	5’- GTGAGGACCTTGTCAAGCCACT -3’	Product size 156 bp
WNT1 – wingless-type MMTV integration site family, member 1
Forward	5’- CTCTTCGGCAAGATCGTCAACC -3’
Reverse	5’- CGATGGAACCTTCTGAGCAGGA -3’	Product size 112 bp
β-Catenin – catenin beta 1 (CTNNB1)
Forward	5’- CGTGGACAATGGCTACTCAAGC -3’
Reverse	5’- TCTGAGCTCGAGTCATTGCATAC -3’	Product size 285 bp
BGN- Biglycan
Forward	5’- GAGACCCTGAATGAACTCCACC -3’
Reverse	5’- CTCCCGTTCTCGATCATCCTG -3’	Product size 131 bp
ITGA5 - Integrin α5
Forward	5’- GCCGATTCACATCGCTCTCAAC -3’
Reverse	5’- GTCTTCTCCACAGTCCAGCAAG -3’	Product size 139 bp
ITGB1 - Integrin β1
Forward	5’- GGATTCTCCAGAAGGTGGTTTCG -3’
Reverse	5’- TGCCACCAAGTTTCCCATCTCC -3’	Product size 143 bp

Histomorphometry.

Static histomorphometry was done by micro-computed tomography (microCT) of rabbit femoral heads using a 1172 SkyScan (Bruker, Allentown MA) at 6 μm resolution with a 0.5 mm aluminum filter. Bone samples fixed overnight in 3.7% formalin, transferred to 70% ethanol, and kept at −20°C until used. Images of femoral heads were reconstructed and analyzed using the NRecon, CT analyzer and CTvox software; cross sectional images used Data Viewer (Bruker).

For dynamic histomorphometry, animals were labeled with xylenol orange at 20 mg/kg 2 weeks before sacrifice and with calcein at 10 mg/kg 1 week before sacrifice. Bone samples for fluorescent labels were frozen in blocks at the time of animal sacrifice and cut as frozen sections using carbide blades at 6 μm on a CryoStar NX70 cryostat using the Kawamoto method of tape transfer [13]. Sections cut were coded and stored, with the examination of labels by a blinded observer not involved in cutting or labeling.

ELISAs and Glucose measurements.

Rabbit blood serum ACTH_1–24, Cortisol, and VEGF were measured by ELISA (Biomatik, Kitchener, Ontario). Glucose levels were measured by CareSens-N Blood Glucose Meter (i-SENS, Inc., Seoul, Korea). Mitochondrial complex I enzyme activity was measured using the colorimetric complex I enzyme activity microplate assay from Abcam (Waltham, MA).

Statistics.

Statistical analysis used GraphPad Prism, (Insight Partners, NY, NY) (RRID:SCR_002798), software that combines scientific graphing, comprehensive curve fitting and data organization, designed for biological research applications for data analysis, hypothesis testing, and modeling. Results are presented as mean ± SD, for three or more measurements, unless stated. Comparisons of differences used the nonparametric unpaired t-test. Results with p <0.05 were considered significant.

RESULTS

The optimum concentration of ACTH_1–24 for rabbit treatment.

Preliminary analysis evaluated ACTH_1–24, cortisol, glucose, and vascular endothelial growth factor (VEGF) in rabbit serum for four hours as a function of ACTH_1–24 as a single subcutaneous injection by ELISA (Fig 1). Injection of 30 μg per kg rabbit weight (30 μg per kg) ACTH_1–24 elevated serum ACTH_1–24 (p<0.05) and Cortisol (p<0.0001) starting 30 minutes after the treatment. On the other hand, 1 μg per kg ACTH_1–24 increased serum ACTH_1–24 (p<0.001) and Cortisol (p<0.0001) only at 4 hours and was not significant at earlier times (Fig 1 A–B). Injection of 0.6 μg per kg ACTH_1–24 had no measurable effect on cortisol (p > 0.5, Fig 1B) and had minimal effects on serum VEGF at 4 hours (Fig 1C). Blood glucose was elevated by 30 μg/kg (p<0.005) and 1 μg/kg (p<0.05) of ACTH_1–24 at four hours, but not at 0.6 μg/kg (Fig 1D). We chose 0.6 μg/kg of ACTH_1–24, with no significant effects on cortisol, for chronic treatment without causing hypercortisolism.

Figure 1.

ELISA assays of rabbit blood serum following injections of 0.6, 1, and 30 μg per kg rabbit mass. High doses of ACTH_1–24 increased serum ACTH_1–24, Cortisol, VEGF and glucose at 4 hours, green and purple traces. At 0.6 μg per kg, the concentration used in long-term studies, ACTH_1–24 (red symbols) did not measurably increase serum ACTH_1–24 (A), cortisol (B), or glucose (D), but did increase serum VEGF (C) at four hours (p <0.0001); VEGF was also increased by 1 μg per kg rabbit mass. 30 μg per kg did not affect VEGF, interestingly. Effects on glucose (D) were inconsistent at short times, but increased at the two higher doses, (p <0.003), at four hours.

ACTH_1–24 treatment (0.6 μg/kg/day for 5 weeks) increased bone formation measured by xylenol orange and calcein labeling.

Calcein and xylenol orange labels were visualized in the trabecular bone of femoral heads in treated and control animals (Fig 2A–B). To show the trabecular structure, right panels show the same fields superimposed on phase images. In obtaining these data, femoral heads were sectioned through at 6 μm increments, with one evaluated at every eight sections (~50 μm intervals). Surprisingly large and consistent increases in bone formation occurred with the very low dose ACTH_1–24. In Fig 2B, left and right bars, labeled surfaces are compared (double labels plus 1/2 of single labeled area)/ total area, in control and treated bone daily for five weeks. This gave p = 0.0089 for five-week treatment. Overall, this suggests that effects were due to increased formation of trabecular bone. This result was confirmed by micro-computed tomography (Fig 3, below).

Figure 2.

Five-week low dose ACTH_1–24 treatment produces a highly significant effect on bone formation by evaluation of calcein and xylenol orange labeling. The variability may reflect in part the ~10% variability in animal weights (see Materials and Methods), 4.2 ± 0.4 kg for each group.

A. Examples of calcein and xylenol orange labels at the center of femoral head in treated (top) and control (bottom) animals. To show the trabecular structure, right panels show the same fluorescent labels superimposed on phase images. Each image is 580 across x 750 microns high.

B. Comparison of labeled surfaces (double labels plus 1/2 of single labeled area)/ total area in control and treated has p = 0.0089. The results are in keeping with differences in the same animals by micro computed tomography (Fig 3, below).

Figure 3. Micro computed tomography results on treated and control femoral heads.

A-B. Representative cross sections of the femoral heads of rabbits treated with 0.6 μg/kg/day of ACTH_1–24 for five weeks (A) or given saline only (B).

C-F. Computed tomography results of separate measurements (upper and low part of the femoral head) from each of five ACTH and five saline treated animals are shown.

C. ACTH_1–24 increases trabecular bone volume, p = 0.026.

D. ACTH_1–24 increases intersection surface, p = 0.0302.

E. ACTH_1–24 increases trabecular number (trabeculae per distance), p = 0.0457.

F. ACTH_1–24 decreases porosity relative to saline control, p = 0.029.

ACTH_1–24 treatment increased trabecular bone parameters in the femoral head measured by micro-computed tomography (micro-CT).

Micro-CT images of femoral head from ACTH_1–24 and control (saline) treated rabbits showed that the trabecular bone of ACTH_1–24 treated rabbits has more trabecula and less space in between them (Fig 3A–B). The percentage bone volume was significantly increased after 5 weeks of ACTH_1–24 treatment (Fig 3C, p < 0.03). Intersection surface and trabecular number increased 17–18% (p = 0.030 and p = 0.045 respectively) (Fig 3 D, F). Total porosity decreased 8%, p < 0.03 (Fig 3 E).

ACTH_1–24 treatment did not affect cortical bone from the femoral shaft.

Cross sections of ACTH_1–24 treated (top) and saline control (bottom) bone appeared identical (Fig 4A, top versus bottom images). This was followed by quantitative μ-computed tomography (Fig B). Bone volume/total volume was statistically identical (> 0.95, n = 5 treated and untreated). In this case, there is no trabecular bone, so trabecular parameters (see Fig 3) are not meaningful.

Figure 4. Shaft of femur with and without ACTH_1–24 treatment shows no differences.

A. Cross sections of ACTH_1–24 treated (top) and saline control (bottom) bone. No difference is apparent; this was further studied by quantitative μ-computed tomography (B).

B. Percent bone volume by μ-computed tomography showed no difference (p > 0.95, n = 5 treated and 5 untreated). Since there is no trabecular bone (see A), trabecular parameters (as shown in trabecular bone of the femoral head, in Fig 3) are not meaningful and are not shown.

Effects on osteoclast and osteoblast differentiation studied using marrow cells.

We followed up these studies using PCR for osteoblast and osteoclast targets in bone marrow. Here, we harvested cells from the femur after removal of the femoral head; stromal stem cells and macrophages are plentiful in the red marrow. Serum assays might be used, but aside from TRAP, osteoclast-produced proteins are not highly expressed in the serum. Osteoblast protein controls showed that alkaline phosphatase, bone GLA protein (osteocalcin), Type 1 Collagen, and bone morphogenetic protein 2 all increased significantly (Fig 5 A–D), in keeping with labeling of bone formation (Fig 2) and bone density measurements (Fig 3). On the other hand, there were no significant differences in osteoclast products in keeping with lack of effects on bone degradation (TRAP, Cathepsin K, ATPa3 subunit, and the ATPd2 subunit (Fig 5, E–H).

Figure 5. PCR for osteoblast and osteoclast targets in bone marrow cells.

Cells were used to produce mRNA as described in methods. Significant increases in osteoblast products are seen (A-D) in keeping with effects of bone formation by labeling (Fig 2). In contrast, there are no significant differences in osteoclast products in keeping with lack of effects on bone degradation (E-H).

A. Alkaline Phosphatase expression (p = 0.033, n = 5).

B. Bone gla protein (osteocalcin) (p = 0.012, n = 5).

C. Type 1 collagen (p = 0.046, n = 5).

D. Bone morphogenetic protein 2 (p = 0.011, n = 5)

E. Tartrate-resistant acid phosphatase secreted by osteoclasts (not significant, n = 5).

F. Cathepsin K expressed highly in osteoclasts (not significant, n = 5).

G. ATPa3 subunit of the V-ATPase expressed in osteoclasts (not significant, n = 5).

H. ATPd2 subunit of the V-ATPase expressed in osteoclasts (not significant, n = 5).

ACTH_1–24 treatment up-regulates the expression of osteoblast-related genes in normal human osteoblast precursors differentiating into human osteoblasts (hOB) in vitro.

Human osteoblasts were differentiated from normal human osteoblast precursors, four aliquots, two male and two female, differentiating on PET membranes for 5 weeks [11] with 0, 10⁻¹², 10⁻⁹, and 10⁻⁶ M ACTH_1–24 (Fig 6). ACTH_1–24 in concentrations as low as 10⁻¹² M, and consistently at 10⁻⁹ M, increased expression of the structural bone proteins Collagen type 1, ALP, Osteocalcin (BGLAP), and RunX2 (Fig 6A).

Figure 6. Effects of low dose ACTH_1–24 on structural bone proteins, the ACTH_1–24 and VEGF receptors, and mitochondrial complex proteins of human osteoblasts.

Human osteoblasts were differentiated for 5 weeks on polyethylene terephthalate membranes as described in Methods. Effects of no ACTH_1–24 to 10⁻¹² M, 10⁻⁹ M, and 10⁻⁶ M ACTH_1–24 treatment by PCR. For clarity, p-values are shown in the figure only.

A. Effect on structural proteins and key bone differentiation factors. Individual PCR measurements of type I collagen (Col1A1), bone GLA protein (osteocalcin), alkaline phosphatase (ALPL), and the master differentiation factor RunX2 are shown. Significant effects are seen for all, with maximal effects for collagen and alkaline phosphatase at 10⁻¹² M.

B. Effect on the ACTH_1–24 receptor MC2R and on the key differentiation cytokine VEGF and its receptors FLT-1 and FLK-1. Effects are variable but seen at 10⁻¹² M for the MC2R, VEGF, and FLK-1. Effects on FLT-1 are smaller and less distinct.

C. Variable effects on several mitochondrial complex proteins. This was studied because of the energy-intensive process of bone formation. All were upregulated at 10⁻¹²M ACTH_1–24, but with variable effects and variable specificity for the lowest dose of ACTH_1–24.

The effect on the ACTH_1–24 receptor MC2R and on the key differentiation cytokine VEGF and its receptors FLT-1 and FLK-1 were tested. Increases occurred at 1 pM for the MC2R, VEGF, and FLK-1. Effects on FLT-1 were smaller and less distinct (Fig 6B).

Following the hypothesis that effects on osteoblast reflected activity of mitochondrial complex proteins [14], we studied this for four mitochondrial complex 1 proteins NDUFA5, NDUFS2, NDUFB1, NDUFB6 (Fig 6C). All proteins had synthesis increased at 10⁻¹² M ACTH_1–24, with some proteins increased to 10⁻⁹ M dose of ACTH_1–24 and variation at higher concentrations. For exact p values please see Fig 6. Effects of additional proteins known to regulate bone formation largely at the cell differentiation level were further studied, see Fig 8.

ACTH_1–24 treatment may increase human osteoblast energy production by up-regulating mitochondrial complex 1 subunit expression.

We followed measurements of individual mitochondrial complex 1 protein mRNAs (Fig. 6) with a colorimetric Complex I enzyme activity microplate assay (Fig 7). The activity of Mitochondrial complex 1 was upregulated by ACTH_1–24, with maximum activity at 10⁻⁹ M and increased activity at 10⁻¹² M ACTH_1–24 in two different preparations of osteoblasts differentiated on PET membranes [12]. Interestingly, the activity in individual isolated osteoblast preparations was quantitatively different but qualitatively consistent (Fig 7A versus 7B), suggesting that the preparations may vary in differentiation, as we noted with age in studies of mouse transport proteins [15]. We did not include higher concentrations in the ELISA (Fig 7) In Fig 6C we included 1 μM, assays of individual mitochondrial Complex 1 proteins. At 1 μM, all mitochondrial protein effects dropped relative to 10–9 M. While further studies at 1 μM might be of interest, this exceeds concentrations of interest in long-term very low dose ACTH_1–24 treatment.

Initial Pathway Analysis: Expression by PCR in ACTH_1–24 treated human osteoblasts of proteins linked to osteoblast maturation and activity in other studies.

Human osteoblasts differentiated from normal human osteoblast precursors on PET membrane inserts for 5 weeks with 0, 10⁻¹² M, 10⁻⁹ M, and 10⁻⁶ M of ACTH_1–24 were studied. This showed the activation of specific BMP2, Smad 1, Wnt-1, and β-Catenin together with the nonspecific TGF- β pathway (Fig 8). Assayed proteins were Bone Morphogenetic Protein-2, SMADs 1,2, and 3 (Suppressors of Mothers Against Decapentaplegic (Fig 8A), transcription factors affecting osteoblast differentiation) (Fig 8 B–D) and Transforming Growth Factor- β (Fig 8E), Additional regulatory proteins studied were Wnt1(Fig 8F), β-Catenin (Fig 8G), biglycan-19 (Fig 8H), and the surface proteins integrin A5 and B1 (Fig 8 I–J).

DISCUSSION

We found a surprisingly strong effect on trabecular bone formation of amounts of ACTH_1–24 too small to increase cortisol (Fig 1), as shown initially by in vivo labeling with fluors (Fig 2). This is interesting from several points of view. With very high ACTH_1–24 and high glucocorticoids, there is rapid loss of trabecular bone, which recovers rapidly when hypercortisolism is resolved [16]. It is possible that this reflects very small amounts of residual ACTH_1–24, but this is hypothetical.

At best, the situation is complex, and it is likely that commonly used murine models, including in our earlier work, respond differently than humans or rabbits, e.g. [17]. Thus, species specific effects must be considered in any related work.

We questioned the effect of extremely low doses of ACTH_1–24 on trabecular bone formation (Fig 2) as possibly being, at least in part, artifactual. However, very sensitive micro-computed tomography, assayed blindly to the operator (see Methods) (Fig 3) supported the bone formation labeling work. Percent bone volume, intersection surface and trabecular number were increased after 5 weeks ACTH_1–24 treatment, and porosity was decreased (Fig 3C–F, with p values from 0.026 to 0.045). This indicates that, particularly in the rabbit, the skeleton increases bone synthesis at concentrations of ACTH_1–24 that do not affect glucocorticoids measurably (Fig 1).

We harvested marrow from the femoral shaft in the proximal femur and assayed mRNA by PCR for osteoclast and osteoblast-specific products (Fig 5). In keeping with the findings of increased bone formation in Figs 2 and 3 alkaline phosphatase, bone GLA protein (osteocalcin) Type 1 Collagen and bone morphogenetic protein 2 all increased significantly (Fig 5 A–D). In contrast, osteoclast products (TRAP, Cathepsin K, ATPa3 and ATPd2) (Fig 5 E–H) were present, as expected in marrow, but were not significantly changed by low-dose ACTH_1–24, in keeping with lack of effect on bone degradation at the level of sensitivity of this assay.

Whether this applies to humans requires additional work. At this point, however, we did consider the effect of very low dose ACTH_1–24 on human osteoblasts. Specifically, very low concentrations of ACTH_1–24 increased key bone proteins, Col1A1, osteocalcin, and alkaline phosphatase, as well as up-regulated the master differentiation factor RunX2 (Fig 6A). It is interesting that the expression level of all osteoblast differentiation markers that we studied (Col1a1, RunX2, ALP, and BGLAP) increased at both 1pM and 1nM of ACTH_1–24. Peak of the expression was at 1pM for Col1a1 and ALP compared to 1nM for RunX2. The high RNA expression of Col1a1 and ALP is logical as structural proteins relative to RunX2 as a developmental signal. In other words, an increase of RunX2 at an earlier phase of differentiation would activate the expression of Col1a1 and ALP. Why BGLAP production is not as high is unclear, since it is also a structural protein, although the time-dependency of its production is unknown. This work included assays using female and male sex osteoblasts from Lonza (see Materials); possible sex-specific effects were not seen, although it should be noted that with this in vitro assay no sex hormones were present. Although there are extensive reports of humans with elevated ACTH_1–24, there are no clear reports of normal glucocorticoids with very slightly increased ACTH_1–24. On the other hand, limited data links increased ACTH_1–24 to bone formation markers [18]. The data on effects of picomolar ACTH_1–24 on its receptor MC2R, the cytokine VEGF and the VEGF receptors FLT-1 and FLK-1 (Fig 6B) support the hypothesis of very low dose ACTH_1–24 increasing bone formation by biochemical mechanisms related to these receptors and cytokines, in general. The relationship of VEGF to bone synthesis is established, in vivo in rabbits and in vitro in human stem cells [19, 20].

We hypothesized that the effects of very low concentrations of ACTH_1–24 might reflect in part changes in cellular metabolism including mitochondrial complex 1 activity [14] (Fig 6C and Fig 7). These results were consistent with changes in cellular metabolism. There are a variety of reports consistent with the relationship of mitochondrial activity and bone formation, mainly using mouse models [21, 22, 23, 24], but there are some consistent reports using human osteoblast precursors [25, 26]. Whether this is central to the effects of low dose ACTH_1–24 on bone formation in humans will require further study.

We surveyed by PCR for key proteins in ACTH_1–24 treated and control human osteoblasts proteins linked to osteoblast maturation and activity [9] (Fig 8). The classes of proteins are discussed in [27–31] with illustrations of major classes and roles in biochemistry of bone. Assayed proteins were Bone Morphogenetic Protein-2 (Fig 8A), SMADs 1,2, and 3 (Suppressors of Mothers Against Decapentaplegic), all of which but SMAD2 increased with ACTH_1–24 low dose treatment. Transcription factors affecting osteoblast differentiation all increased with ACTH_1–24 (Fig 8B–D) as did Transforming Growth Factor-β (Fig 8E), Wnt1 (Fig 8F), β-Catenin (Fig 8G), and structural proteins including biglycan (Fig 8H), and integrins A5 and B1 (Fig 8I–J). These initial statistical tests shown used nonparametric unpaired t-tests, and in later work additional analysis including analysis of variance may be useful.

Differentiation factors and transcription factors involved are modeled in Fig 9. The green area represents intracellular signals, the white area nuclear signals, although these factors overlap with intracellular signals. Structural proteins are shown at the bottom in violet, and extracellular signals at the top in brown. Vertical arrows indicate increase with ACTH_1–24, although these are variable with concentration in the picomolar range as shown in Figs 6 and 8.

Figure 9. Model of major growth factors modified by low concentrations of ACTH_1–24 in osteoblast maturation.

The differentiation and transcription factors are defined in Fig 7 and not repeated here, for simplicity. The green area represents intracellular signals, the white area nuclear signals, although these overlap. Structural proteins are shown at the bottom in violet, and extracellular signals at the top in brown. Vertical arrows indicate increase with ACTH_1–24, although variable with concentration in the picomolar to nanomolar range, as illustrated in Figs 5 and 6.

Collectively, our results indicate that very low dose ACTH_1–24 treatment increases bone formation in vivo and in vitro and up-regulates energy production in human osteoblasts. This occurs at concentrations that do not affect cortisol level measurably (Fig 1).

Overall, our results point to the importance of ACTH_1–24 in bone formation and suggest that ACTH_1–24 may be a therapeutic target for treating osteopenia and osteoporosis in the femur, as well as preventing osteonecrosis as shown in our earlier work [8]. Whether the low-dose ACTH_1–24 might be of more general use in promoting formation of trabecular bone is possible, for example in the spine. In addition, more in-depth exploration of each signaling pathway will be useful, in further work.

Data Availability:

All data are available on request. No shared data are used in the work.