CART peptide was discovered 16 years ago. After its implication as an anorectic peptide it attracted interest from the pharmaceutical industry because the elusive CART receptor could represent an attractive target for the treatment of obesity.1–4 Since then CART peptides have been implicated in many more processes, including reward and addiction, stress response, anxiety and depression, cardiovascular control, neuroprotection, follicular dominance, and bone remodeling; however, as yet, no CART-based therapy has resulted.5–15
CART is transcribed and translated as a prepropeptide and processed by prohormone convertases to either CART 55-102 and CART 62-102 in rat and CART 49-102 or CART 42-102 in humans.16, 17 The short form is identical between mouse, rat, and human and the long form is one amino acid different at the N-terminus. Both isoforms are biologically active but differences in potency and slight differences in functionality have been reported.18, 19 CART is expressed in central and peripheral neurons where CART is proposed to act as a neurotransmitter.20–23 In addition, it is expressed in endocrine cells in the gastrointestinal tract, pituitary, adrenal glands, and the pancreas.21, 24–29 In line with the expression in endocrine organs and its suggested role as a hormone, different isoforms of CART can also be detected in the circulation, where they exhibit a diurnal variation.30
The role of CART in bone is established in several mouse models. First, CART-deficient mice have a low bone mass phenotype and, surprisingly, very few other defects.14, 31 These include impaired insulin secretion upon glucose challenge and a very modest increase of body weight in aged knockout mice. Second, MCR4-deficient mice display a high bone mass phenotype that was shown to be caused by increased serum CART levels and CART expression in the hypothalamus.32 Third, female transgenic α1(I)collagen-Cart mice display a robust increase in bone mass. These mice ectopically express CART in osteoblasts and this results in a twofold increase in serum CART levels.33 In addition to these mouse models, polymorphisms in the CART gene have been shown to be linked to higher bone mineral density (BMD) in humans.34
The mechanism by which CART modulates bone remodeling is not completely understood but involves reduction of receptor activator of NF-κB ligand (RANKL) production by osteoblasts, thereby reducing osteoclast activity.14 CART has not been found to be expressed in bone cells, nor has it been found to be capable of directly affecting the functionality of bone cells in vitro.14, 33 However, CART is expressed in the ovary, in neurons projecting to gonadotropin-releasing hormone (GnRH) immunopositive neurons in the hypothalamus, and in the pituitary.13, 35, 36 This suggests that CART could potentially modulate the hypothalamic-pituitary-gonadal (HPG) axis and thereby can affect bone physiology.
The ovary produces several factors that have been shown to affect bone mass. Granulosa cells synthesize and secrete E2 that signals in bone through bone cell–specific expression of estrogen receptors.37 Also, gonadal inhibin production has been implicated in the regulation of bone mass, and there is evidence that suggests that the follicle-stimulating hormone (FSH) receptor is expressed in osteoclast cells and FSH can directly affect bone mass.38–40 In addition, INSL3 is a circulating factor that is produced by theca cells of developing follicles and has been shown to affect bone mass through RXFP2 receptor signaling in bone cells.41–44
CART and E2 have been shown to mutually affect each other in several studies. In the bovine ovary, CART expression in granulosa cells is responsible for reduced E2 production and subsequent follicular atresia.13 Additional evidence that link CART and E2 together comes from studies with a rat model of ischemia-induced brain damage and a rhesus monkey ovariectomy (OVX) model in which E2 treatment induced differential expression of CART in the brain.12, 45 Furthermore, CART expression has been shown to be dependent on the estrous cycle.36, 46
The objective of the present study was to determine whether peripheral CART peptide treatment can increase bone mass in a premenopausal and postmenopausal mouse model and can be used as a therapy for the treatment of bone disease.
Materials and Methods
Animals and husbandry
CART peptide treatment was performed in 10- to 12-week-old male and female C57/Bl/6J mice unless specified otherwise. Transgenic α1(I)collagen-Cart mice were previously described and generously provided by G. Karsenty.33 Mice were housed in macrolon cages (5 animals/cage) at 19°C to 21°C. The relative humidity was 50% to 60% and an artificial light cycle of 12 hours alternating with 12 hours of darkness was offered. Standard pelleted food (SDS, Witham, Essex, UK) and water was provided ad libitum during the in-life period. The animals were observed for morbidity and mortality daily during the in-life period. All animal experiments were approved by the Animal Ethics Committee.
We obtained synthetic rat CART 55-102 peptides commercially from different providers (003-62, Phoenix Pharmaceuticals, Burlingame, CA, USA; 46-2-55, American Peptide Company, Sunnyvale, CA, USA; H-4446, Bachem, Bubendorf, Switzerland). Human CART 39-89 peptide and human CART 49-89 peptide were produced recombinantly in Escherichia coli. Briefly, the sequence coding for hCART49-89 and hCART 39-89 was codon-optimized for expression in E. coli, synthesized, and subcloned into the pET32a expression vector (Novagen, Madison, WI, USA). A fusion protein with an N-terminal thioredoxin-A, followed by a HIS-tag and a tobacco etch virus (TEV) protease cleavage site under control of a T7/lac promoter was placed upstream of the CART peptide sequence. We transformed this construct into E. coli BL21(DE3) (Invitrogen, Carlsbad, CA, USA). We used a fed-batch fermentation process in a 7-liter bioreactor (Applikon, Foster City, CA, USA) for expression of the protein. We used minimal salt medium with 60 g/L glycerol. We maintained the temperature at 30°C, pH at 7, and oxygen at 20% saturation. When glycerol was almost depleted in the medium, we added 20g/L lactose for induction of expression. Cells were harvested, centrifuged, and lysed by sonication in a Rosett Cell (Branson, Danbury, CT, USA). Polyethyleneimine (0.2% vol/vol) was added to the lysate and the precipitate was centrifuged. We loaded the supernatant on a Ni-NTA superflow (Qiagen, Hilden, Germany) column, eluted with a buffer containing 500 mM imidazole. The eluate was dialyzed against a 50-mM sodium phosphate buffer (pH 7.3) and the N-terminal tag was cleaved by TEV protease, containing an N-terminal HIStag itself. After purification of the cleaved CART peptide, it was loaded on a Ni-NTA superflow column and the flow-through was concentrated using a VivaSpin concentrator (3000 kDa MWCO; Sartorius Stedim Biotech, Aubagne, France). Endotoxins and other contaminants were removed on a Q sepharose FF column (GE Healthcare, Little Chalfont, UK). Buffer exchange and separation of the different folded CART peptide variants was performed by a size exclusion chromatography (Superdex-30), using 50 mM ammonium acetate. The fractions containing the correctly folded hCART peptide were pooled, checked for endotoxins by a limulus amebocyte lysate (LAL) QCL-1000 assay (Lonza, Basel, Switzerland), and freeze-dried.
We analyzed synthetic and recombinant peptides by nondenaturing polyacrylamide gel electrophoresis (PAGE) in the presence or absence of dithiothreitol (DTT), and liquid chromatography–mass spectrometry (LC-MS) (Supplemental Fig. 1) to determine purity and quality. We performed LC-MS on a XBridge BEH300 C18 column (3.5 µm, 1 × 150mm; Waters) using a gradient of (A) 0.1% HCOOH in MilliQ (Supplemental Fig. 1A) and (B) acetonitrile (Supplemental Fig. 1B) (0–30 minutes, 10% to 30% A; 30–30.1 minutes, 30% to 90% A; 30.1–35 minutes, 90% A; 35–35.1 minutes, 90% to 10% A; 35.1–50 minutes, 10% A). The flow rate was 50 µL/min (MIC splitter). Column temperature was 40°C. We used MicrOTOF-Q (Bruker, Billerica, MA, USA) to determine the mass of the peptides.
We subsequently formulated the peptides in slow-release pellets by Innovative Research of America (IRA, Sarasota, FL, USA). We determined the quality of the peptides in the pellets and the release of the peptides from the pellets by nondenaturing PAGE either in the presence or absence of DTT and LC-MS (Supplemental Fig. 2).
We implanted slow-release pellets (IRA, Sarasota, FL, USA) containing CART peptides and placebo pellets in intact male and female mice, or in female mice directly after they were subjected to sham or OVX surgery. Mice were anesthetized with isoflurane (Isofluo; Abbott Park, IL, USA) and an incision was made in the neck. We placed slow-release pellets between the skin and the muscle at the lateral side of the neck using a trocar. We sutured the skin using silk suture (Silkam 3/0). For sham and bilateral OVX surgery, we anesthetized the mice using isoflurane. Briefly, bilateral incisions were made in the flank and the ovaries were externalized. Fallopian tubes were ligated below the ovaries, which were then excised. Muscle and skin wounds were closed using silk suture. Sham surgery consisted of externalizing the ovaries from the abdominal cavity and replacing them without being excised. Perioperational painkilling was applied by subcutaneous injection with Rimadyl (5 mg/kg) 30 minutes before the start of the procedures and by daily injections for up to 3 days postsurgery. Combined CART and E2 treatment was performed by twice-daily subcutaneous (s.c.) injections with E2 (synthesized in house) in 0.5% wt/vol gelatin and 5% wt/vol mannitol in water or with E2 containing slow-release pellets (IRA, Sarasota, FL, USA).
We analyzed the right femur of intact male and sham, OVX, and intact female mice by in vivo micro–computed tomography (µCT) at different time points after start of treatment. At the end of the experiment, we collected blood and performed an autopsy to retrieve the slow-release pellets to determine the remaining amount of CART peptides.
In vivo µCT analyses
We anesthetized female and male mice with a combination of fentanyl (Hypnorm; VetaPharma, Leeds, UK) and midazolam (Dormicum; Roche, Mijdrecht, The Netherlands) via subcutaneous injection. We scanned the femurs in a Skyscan 1076 µCT scanner (Skyscan, Kontich, Belgium). We performed a 180-degree scan for each femur using the following settings: 9-µm resolution, 1.0-mm aluminum filter, 2100-ms exposure time, 0.9-degree rotation step, Frame Averaging 1, X-ray source settings 50 kV and 200 µA. We performed three-dimensional reconstruction with the NRecon software and analysis with the manufacturer's CT-analyzer software (Skyscan, Kontich, Belgium). We performed trabecular bone analysis on a 2-mm slice 150 µm proximal to the distal growth plate of the femur. We analyzed cortical bone on a 2-mm slice 4.150 mm proximal from the growth plate. All three-dimensional and two-dimensional trabecular and cortical parameters were calculated by the CT-analyzer program as described in “Structural parameters measured by the Skyscan CT-analyser software” (http://www.skyscan.be/next/ctan_ctvol_02.pdf).
We performed histopathological analysis on female mice after 60 days of treatment with CART peptide. All internal organs were examined for visual signs of abnormalities and their weights were recorded. Samples of organs and tissues from the animals were preserved in 4% buffered formaldehyde, dehydrated, and embedded in paraffin wax. Sections (3- to 4-µm-thick) from vagina, uterus, ovary, and mammary gland were stained with hematoxylin and eosin (HE) staining method and examined histopathologically. For bone histomorphometric analysis, one tibia was fixated in 4% buffered formaldehyde, dehydrated, and embedded in paraffin. Sections (3- to 4-µm-thick) made from these blocks were stained with the HE staining method and osteoblasts were counted on three sections, each 100 µm apart. The second tibia was fixated in Burckhardt's fixative (pH 7.4) for 24 hours and then kept in absolute ethanol. The decalcified tibia was embedded in methyl methacrylate (MMA), and 5-µm-thick sections were made. The trabecular bone volume (BV) was measured on three von Kossa–stained coronal plane sections, each 100 µm apart. For the osteoclast counts, sections were stained with a commercial kit for acid phosphatase leukocyte (Sigma, St. Louis, MO, USA) and osteoclasts were counted on three sections, each 100 µm apart. Total area (mm2) was measurement by using a specific software program developed in house.
We measured epithelial thickness using an in-house-developed software program. Average thickness was determined in three different areas with a total length of approximately 1 mm.
Alkaline phosphatase and tartrate-resistant acid phosphatase activity in bone lysates
We removed the proximal end of the femur and flushed the bones with PBS containing protease inhibitor (Complete protease inhibitor cocktail; Roche, Mannheim, Germany). The femur was minced into small pieces (<1 mm) using a power homogenizer (GLH; OMNI International) in 100 mM potassium phosphate buffer (pH 7.8) containing protease inhibitor (Complete protease inhibitor cocktail; Roche). Triton X-100 was added (0.2%) and the lysates were incubated for 18 hours at 4°C on a shaker. The next day we centrifuged the samples for 15 minutes at 20800g at 4°C. The supernatant was collected and the protein concentration was measured by using a commercial kit from Pierce (BCA protein assay reagent, 23225). Alkaline phosphatase (ALP) activity was measured in an optiplate (white, 96 wells; Perkin Elmer, Waltham, MA, USA) using CDP-star-ready to use solution (Roche). After 30 minutes incubation in the dark, we measured the chemiluminescence using the Wallac Victor2 1420 multilabel counter (Perkin Elmer) according to the manufacturer's instruction.
We measured tartrate-resistant acid phosphatase (TRAP) activity using p-nitrophenyl phosphate (pNPP) substrate buffer (pNPP tablets; Sigma-Aldrich, St. Louis, MO, USA; 0.1 M NaCl, 0.01% triton X, pH5.5) with 50 mM tartrate solution (pH 5.2). After 1.5 hours of incubation at 37°C, we measured the absorbance at 405 nm using the Wallac Victor2 1420 multilabel counter (Perkin Elmer).
Blood and urine analyses
For serum samples, we collected blood via the vena cava under isoflurane anesthesia. Blood was allowed to clot and serum was separated by centrifugation at 20800g at room temperature for 15 minutes. For plasma samples, we collected blood in centrifuge tubes containing lithium-heparine (9 µL/mL blood) and separated it by centrifugation at 20800g at room temperature for 15 minutes.
We determined serum alkaline phosphatase using a Hitachi-911 analyzer (Roche). We used commercial assays to measure plasma 17β-estradiol (DRG Instruments, Marburg, Germany; EIA-2693), plasma progesterone (Cayman Chemical Company, Ann Arbor, MI, USA; cat no: 582601), serum osteoclast-derived TRAP form 5b (TRACP5b) (IDS, Scottsdale, AZ, USA; MouseTRAPTM Assay, SB-TR103), serum prolactin (GenWay Biotech Inc, San Diego, CA, USA; cat no 40-101-325032), plasma osteocalcin (Millipore, Billerica, MA, USA; single plex assay cat no MBN-41k-10c), plasma RANKL (Abcam, Cambridge, UK; RANKL Mouse ELISA kit, ab100749), and plasma collagen type 1 propeptide (P1NP) (IDS; cat no. AC-33F1). We measured luteinizing hormone (LH) and FSH in serum using an immunofluorescence assay (IFMA) previously described by van Casteren and colleagues.47 We determined urinary calcium, phosphate, and creatinine using a Hitachi-911 analyzer (Roche, Germany).
Three days before the start of the experiment, mice received drinking water containing 0.4 mg/mL of potassium iodide. Rat CART 55-102 peptide was iodinated using IODO-GEN (Pierce, Rockford, IL, USA) precoated iodination tubes according to the procedure the manufacturer prescribed. We administered 125I-CART with a specific activity 0.5 µCi/µg peptide intravenously (i.v.) or s.c. to sham or OVX-operated C57/BL6 mice (n = 3 per group) at a dose volume of 1.0 mL/100g of body weight. We collected blood by orbital puncture under gas anesthesia at different time intervals after administration of the compound. Nine blood samples were obtained for each individual administration route in sham and OVX mice. We collected a maximum of three blood samples of ±80 µL per mouse, and after the third collection the mouse was euthanized. Then the most important organs were collected to determine the radioactive content. In order to separate 125I-labelled peptide from free 125I label, 50 µL of each blood sample was precipitated by addition to 1.95 mL 10% (vol/vol) trichloric acetic acid solution. After mixing, the blood was centrifuged and the radioactivity of the fractions was counted in a scintillation counter.
Real-time PCR analysis
We isolated total RNA using Trizol (GibcoBRL, Chagrin Falls, OH, USA). We performed first-strand synthesis using Oligo(dT)15 (Promega, Madison, WI, USA) and pd(N) (Amersham Pharmacia Biotech, Inc., Little Chalfont, UK) primers and Superscript II RNase H-Reverse Transcriptase (Invitrogen), according to the manufacturers' instructions. cDNA from the synthesis reaction was subjected to quantitative, real-time PCR with SYBR Green mastermix (Applied Biosystems, Foster City, CA, USA) in the ABI PRISM® 7900HT Sequence Detection System (Applied Biosystems). For quantification, standard curves were prepared for both the target and an endogenous reference HP1. We calculated normalized expressions values for each sample by dividing the reference and target value for each sample. Primer sequences used: HP1 forward 5′-gcccaagatggacgcaatc-3′, reverse 5′-ccgaggcgccagtcttc-3′; InhA forward 5′-aagatgtctcccaggctatccttt-3′, reverse 5′-tggccggaatacataagtgaaga-3′; and Insl3 forward 5′-gagctgctgcagtggctaga-3′, reverse 5′-agaagcctggagaggaagctg-3′.
All parameters are expressed as mean ± SEM of the individual data. Statistical analysis was performed using a nonpaired Student's t test or analysis of variation (ANOVA); p < 0.05 was considered as significant.
Treatment of intact female mice with CART peptides increases bone mass
To establish that peripheral CART treatment can affect bone mass, we treated intact mature female mice with synthetic rat CART 55-102 peptides from different commercial providers. Slow-release pellets that contained either 0, 20, 50, or 100 µg rat CART 55-102 peptide were implanted subcutaneously in the neck to provide sustained release of the peptides for 60 days. Quality of the peptides was confirmed by LC-MS and PAGE before and after formulation into slow-release pellets (Supplemental Figs. 1 and 2). At the end of the experiment the pellets were retrieved from the mice and the presence of CART peptides was determined by MS. Results indicated that CART peptides were undetectable in the pellets after 60 days of slow release in the body (Supplemental Fig. 2).
After 60 days of treatment, no significant differences in body weight (placebo versus 100 µg CART peptide, 25.8 ± 0.5 g versus 25.5 ± 0.6 g) or organ weights were observed between placebo and treatment groups.
We assessed the bone mass of the mice treated with CART peptide by in vivo µCT analysis of the femur. After 28 days of treatment, trabecular bone volume fraction (BV/TV) in the distal femur was significantly increased by 45% and trabecular bone mineral density (BMD) by 55% in the 100 µg CART peptide dose group compared to placebo, but not in the 20 µg and 50 µg groups (Supplemental Table 1). After 60 days of treatment, trabecular BV/TV was significantly increased by 36% and 79%, and trabecular BMD by 47% and 73%, in the 50 µg and 100 µg dose group, respectively, compared to placebo (Table 1; Fig. 1A). The observed increase in trabecular bone mass was caused by increased trabecular thickness and trabecular number. The trabecular pattern factor was reduced in most experiments, indicating increased connectivity and strength after CART peptide treatment (Tables 1 and 2; Supplemental Tables 1, 2, and 3). Similar increases in trabecular bone mass were observed with rat CART 55-102 peptide from the different providers.
Table 1. In Vivo Micro-CT Analysis of Female C57Bl/6 Mice After 60 Days Treatment With 60-Day Slow Release Pellets Containing Synthetic Rat CART 55-102, Recombinant Human CART 39-89, or Recombinant Human 49-89 Peptide
Data are presented as mean ± SEM of the individual data.
BMD = bone mineral density; BV = bone volume; BV/TV = bone volume/tissue volume; Tb.Pf = trabecular pattern factor; Tb.Th = trabecular thickness; Tb.N = trabecular number; C.Th = cortical thickness.
Table 2. In Vivo Micro-CT Analysis of Male C57Bl/6 Mice After 60 Days Treatment With 60-Day Slow Release Pellets Containing Synthetic Rat CART 55-102 Peptide, and Male and Female Transgenic α1(I)collagen-Cart Mice
Data are presented as mean ± SEM of the individual data.
BMD = bone mineral density; BV = bone volume; BV/TV = bone volume/tissue volume; Tb.Pf = trabecular pattern factor; Tb.Th = trabecular thickness; Tb.N = trabecular number; C.Th = cortical thickness; WT = wild-type; TG = transgenic α1(I)collagen-Cart mice.
p < 0.05,
p < 0.01, and
p < 0.001, treated versus placebo or WT versus TG.
Cortical bone parameters were determined in the femoral diaphysis. Sustained release of rat CART 55-102 peptide did not affect cortical bone BMD, cortical bone volume, or cortical thickness after 60 days of treatment in a dose-dependent manner, compared to placebo (Table 1; Supplemental Tables 1, 2, and 3).
In order to determine if other CART peptides isoforms were capable of increasing bone mass in mice, we produced two recombinant CART peptides in E. coli, human CART 39-89, and human CART 49-89. Intact female mice were treated with 60-day slow-release pellets containing 50 and 100 µg purified human CART 39-89 peptide or 20, 100, and 500 µg human CART 49-89 peptide. The bone mass in the distal femur of these mice was determined by µCT after 60 days of treatment (Table 1). Human CART 39-89 peptide at a dose of 100 µg increased trabecular BV/TV by 8% and trabecular BMD by 30% compared to placebo, whereas the 50-µg dose did not have a significant effect on BV/TV or BMD. Human CART 49-89 peptide increased BV/TV by 26%, 21%, and 53% in the 20-, 100-, and 500-µg dose groups, respectively, compared to placebo. BMD was increased significantly by 22%, 22%, and 42% in the 20-, 100-, and 500-µg dose groups, respectively. Human CART 49-89 peptide also increased cortical BMD in all three dose groups.
We determined if the increase in bone mass was maintained after the end of the treatment period. To this purpose the slow-release pellets were left in the body after the 60-day treatment period and bone mass was measured by in vivo µCT at days 90 and 120. Results showed that bone mass declined rapidly after CART peptides were depleted from the slow-release pellets and bone mass returned to the level of placebo-treated mice within 30 days after treatment with CART peptides stopped (data not shown).
Analysis of markers of bone formation and bone resorption revealed a significant decrease of serum TRAcP5b in mice that received 60 days of sustained-release treatment with CART peptide (Fig. 1B). Markers of bone formation were not different between placebo and CART peptide–treated groups as exemplified by serum ALP, osteocalcin, and procollagen type I N-terminal propeptide (PINP) levels (Fig. 1B). These findings were confirmed by determination of ALP and TRAP activity in lysates of bones from placebo and CART peptide–treated mice at day 60 of treatment. There was a slight, but not significant, increase in bone ALP activity and a significant reduction in bone TRAP activity compared to placebo (Fig. 1B). The urinary calcium/creatinine ratio was reduced in the CART peptide–treated mice whereas the phosphate/creatinine ratio remained unchanged. Histological examination of bone sections revealed a nonsignificant reduction in osteoclast number of 31%, a 15% decrease in osteoblast number, and a 45% increase in BV/TV in the tibia (Fig. 1B) of CART-treated mice. The morphology of osteoblast, osteocyte, and osteoclast cells was normal in both placebo and CART-treated mice.
The effect of CART on bone mass is gender-specific
Mature male mice were treated with rat CART 55-102 for 60 days. Slow-release pellets providing sustained release for 60 days containing either 0, 50, 100, or 200 µg rat CART 55-102 peptide were implanted subcutaneously in the neck, and changes in bone mass were determined by in vivo µCT at day 28 and day 60. Measurement of the trabecular bone in the distal femur and the cortical bone of the femoral diaphysis indicated that at both 28 and 60 days there was no increase in BV/TV, BMD, or any other trabecular or cortical bone parameter in any of the dose groups (Table 2; Supplemental Table 1).
Next we examined male transgenic α1(I)collagen-Cart mice at the age of 6 and 18 months. Measurement of the trabecular bone in the distal femur by µCT in male transgenic mice did not reveal a significant change in BV/TV, BMD, or any other structural parameter compared to wild-type (Table 2). The female α1(I)collagen-Cart transgenic mice, however, did show a significant 45% increase of BV/TV in the trabecular bone of the distal femur at an age of 10 months and a 236% increase at 18 months (Table 2).
Ovariectomy abolishes the effect of CART on bone
In order to determine if the ovary played a role in the gender-specific activity of CART peptide, we performed sham or ovariectomy surgery in mature female mice and subsequently treated the mice with rat CART 55-102 peptide. The effect on trabecular bone mass in the distal femur was determined after 28 and 60 days of treatment with a slow-release pellet containing 100 µg rat CART 55-102 peptide. Again, CART peptide could induce a significant 61% increase of trabecular BV/TV in sham-operated mice after 28 days and 53% at the end of the 60-day treatment period (Fig. 2A; Supplemental Table 2). Successful OVX surgery was confirmed by the observed 72% reduction of uterus wet weight compared to sham mice at autopsy (sham versus OVX, 854 ± 9 versus 236 ± 5 mg, p < 0.01). In addition, OVX surgery resulted in a 56% decrease in trabecular BV/TV in the distal femur 28 days post-surgery and a 47% decrease after 60 days. Treatment of OVX mice for 60 days with pellets containing 100 µg rat CART 55-102 did not result in increased trabecular bone mass compared to placebo-treated OVX mice (Fig. 2A).
Next we performed ovariectomy in 20-week-old wild-type and α1(I)collagen-Cart transgenic mice. At 4 weeks post-surgery the trabecular bone in the distal femur of the mice was examined by in vivo µCT (Fig. 2B). The results indicate that α1(I)collagen-Cart mice were not protected from OVX-induced bone loss. α1(I)Collagen-Cart OVX mice displayed a significant 30% decrease in trabecular BV/TV compared to sham-operated α1(I)collagen-Cart mice. Unexpectedly, the wild-type OVX mice did not display a reduction in trabecular BV/TV at the 4 weeks post-OVX time point in this experiment. The sham α1(I)collagen-Cart mice did have a significantly increased BV/TV of 26% compared to wild-type sham mice. Because the effect of CART on bone mass increased with age in the transgenic mice and also increased with duration of CART peptide treatment, we examined the response of α1(I)collagen-Cart mice not only after 4 weeks but also 4 months post-OVX. Four months after OVX the α1(I)collagen-Cart mice did not show any recovery from the OVX-induced bone loss (Fig. 2B). α1(I)collagen-Cart OVX mice displayed a significant decrease in BV/TV of 48% compared to sham-operated α1(I)collagen-Cart mice, which was not significantly different from the OVX-induced bone loss of 41% in the wild-type mice. At this time point the α1(I)collagen-Cart sham mice showed a nonsignificant increase in BV/TV of 20% (Fig. 2B) but a significant increase in trabecular BMD of 13% (Supplemental Table 4).
To obtain insight in the pharmacokinetic properties of CART peptides, 10 µg/kg 125I-labeled rat CART 55-102 peptide was either injected i.v. or s.c. in sham or OVX mice. Pharmacokinetic curves and parameters are presented in Figure 3. CART peptide displayed an unexpected long half-life (t½) of 4.5 hours in sham-operated and 4.9 hours in OVX mice. The peptide showed a low clearance and a moderate distribution volume after i.v. administration. The bioavailability of the peptide after s.c. injection was ≥88% compared to i.v. administration in sham-operated mice. In the OVX mice s.c. bioavailability was approximately 100%. Furthermore, in OVX mice 125I-CART peptide displayed equal distribution and elimination characteristics compared to intact mice (Fig. 3). Therefore, difference in exposure could not have accounted for the observed lack of activity of CART peptides in bone in OVX mice.
The addition of 90 µg/kg unlabeled CART did not affect the pharmacokinetic parameters of 125I-CART peptide, indicating that distribution and elimination were not affected at higher CART peptide concentrations (data not shown). Measurement of the accumulation of 125I-CART peptide in the different organs revealed that only the kidney showed accumulation of detectable amounts of 125I-labeled CART during the first 2 hours after i.v. administration. No significant accumulation was found in adrenal, bone, brain, heart, liver, lungs, ovaries, pancreas, spleen, and uterus. These findings are in line with the expected clearance of a 5-kD peptide by the kidneys. However, as CART peptide clearance was low and the terminal elimination half-life was relative long, CART peptide is most likely bound to a serum protein, which prevents quick elimination from the circulation by the kidney. A similar long t½ was also found upon pharmacokinetic examination of CART peptides in dog and cynomolgus monkey (data not shown).
Increase in bone mass by CART peptides is not mediated by estradiol
Analysis of the plasma of placebo and rat CART 55-102–treated mice did not reveal any significant changes in E2, progesterone, LH, FSH, or prolactin after 30 days (data not shown) or 60 days of treatment (Fig. 4A). However, on average serum E2, prolactin, and LH were increased in the 100 µg CART peptide treated group by 25%, 22%, and 75%, respectively, and serum progesterone and FSH were decreased by 28% and 27%, respectively, after 60 days of treatment. No morphological changes were observed in the reproductive tissues of CART peptide–treated mice. Histology of the ovary did not reveal any differences in follicular maturation, follicular atresia, or the presence of corpus lutea in CART peptide–treated mice (Fig. 4C). In addition, the mammary glands and the uteri of the CART peptide–treated mice displayed normal morphology. Endometrial surface epithelium thickness was comparable between placebo and CART treated mice (31.7 ± 0.66 µm versus 32.4 ± 1.26 µm).
To determine if E2 is a mediator of the effect of CART peptide on bone, we established whether E2 supplementation in OVX mice could rescue the bone-building effect of CART peptide. First, we established the amount of E2 needed to restore OVX-induced bone loss. Different concentrations were tested and twice-daily injections with 0.25 µg E2 could restore OVX-induced bone loss back to the level of sham-operated mice. Subsequently, mature female mice were subjected to sham or OVX surgery and treated with either placebo pellets, 60-day slow-release pellets containing 100 µg rat CART 55-102 peptide, twice-daily injection with 0.25 µg of E2, or combined treatment with a 100-µg rat CART 55-102 peptide pellet and twice-daily injection with 0.25 µg of E2. After 60 days of treatment the trabecular bone in the distal femur was measured by µCT and uteri wet weight was recorded. OVX surgery reduced uterus wet weight by 74% and twice-daily injections with 0.25 µg of E2 in OVX mice fully restored uterus wet weight back to the level observed in sham mice (Fig. 5B). As depicted in Figure 5A, treatment of sham animals with CART peptide significantly increased trabecular BV/TV by 51% and OVX surgery reduced BV/TV by 45%. Treatment of OVX mice with twice-daily injections with 0.25 µg of E2 over a period of 60 days completely restored the OVX-induced bone loss (Fig. 5A; Supplemental Table 3). However, the combined treatment of CART peptide and E2 did not result in an additional increase in bone mass, because there was no significant difference in trabecular BV/TV between the OVX CART peptide group and the OVX CART/E2 treatment group.
We next performed expression analysis of two other ovary-derived factors that can affect bone mass, InhA and Insl3. Expression of Insl3 was reduced by 27% and InhA expression was increased by 8% in the ovary of CART peptide–treated mice compared to placebo; however, the changes were not significant (Fig. 4B).
In the present study we showed that peripheral subcutaneous sustained release of CART peptides increased trabecular bone mass in mature intact female mice. Both synthetic rat CART 55-102 peptide and recombinant human CART 49-89 peptide or human CART 39-89 peptide were able to increase bone mass in female mice. This indicates that multiple isoforms of the peptide can evoke the same biological effect on bone. However, the endogenous mature forms of CART peptide seemed to be somewhat more potent than CART 39-89 peptide, and CART 49-89 peptide also increased cortical BMD. With the exception of CART 49-89 peptide, cortical bone was unaffected by CART peptide treatment in all experiments. This could reflect the lower turnover rate of cortical bone compared to trabecular bone. Increased treatment duration or more potent CART peptides could in theory then affect cortical bone. Alternatively, different bone-cell populations exist48–50 at trabecular and cortical sites, responding differently to CART peptide.
CART peptide treatment did not increase the bone mass in male mice, neither did we observe increased bone mass in male α1(I)collagen-Cart mice at the age of 6 or 18 months. Singh and colleagues33 only presented data on increased bone mass in the female α1(I)collagen-Cart mice, which was confirmed in our study, but not on male transgenic mice. Bartell and colleagues51 published a study in which they examined male CART-deficient mice. In this study they did not observe any effect on bone mass or strength whereas female CART knockout mice displayed a low trabecular bone mass phenotype.14 These combined data lead us to conclude that the increase in bone mass by CART peptides is gender-specific and absent in male mice.
CART affects bone mass by reducing osteoclast number, while maintaining normal osteoblast activity.14, 33 This was confirmed in our experiments as CART peptide treatment resulted in decreased osteoclast TRAP, urinary calcium, and the number of osteoclasts, but not in changed osteoblast activity. How CART peptide mechanistically uncouples bone resorption from bone formation is not clear. CART peptide is not expressed in bone cells and was shown to have no effect on bone cells in vitro.14 Because of the lack of a direct effect on bone and bone cells and the observed gender-specific effect, we examined the role of the ovary. Removal of the ovaries completely abolished the effect of CART on bone. This was observed in both OVX mature wild-type mice treated with CART peptide for 60 days, and in OVX α1(I)collagen-Cart transgenic mice. The presence of the ovary is therefore needed for CART peptide to exert its effect on bone and this could be (1) a direct effect on the ovary, in which CART increases the level of expression of a bone-building ovarian factor; (2) an indirect effect, in which CART only needs the presence of one or more ovarian factors to be able to affect bone; or (3) a factor from another organ or tissue that is modulated by the absence of the ovary, which prevents CART from having its effect on bone.
The most prominent ovarian factor that has bone anabolic properties in mice is E2 and, importantly, CART peptide and E2 have been implicated to be able to mutually regulate each other in several studies.12, 13, 36, 45, 46 In our experiments we did not find clear evidence of increased estrogenic activity in the CART peptide–treated mice. Only a nonsignificant increase in serum E2 of 25% was observed, which is well within the normal range of the estrus cycle. The epithelium of the endometrium, which is very sensitive to changes in estrogenic activity, displayed no differences in epithelium cell height, nor was there an increase in the number of mice in estrus between placebo and CART peptide–treated groups. In addition, normal physiological levels of E2 were also not a prerequisite for the effect of CART peptide on bone. Combined CART peptide and E2 treatment of OVX mice could not restore the bone mass to the level of CART-treated sham mice. Together these findings indicate that CART peptide does not exert its action on bone via E2 alone. However, the described experiments can not exclude the possibility that estrogen prepares cells (eg, for subsequent progesterone activity) and the combination of two ovarian factors are needed for the effect of CART peptides on bone.
We were unable to show the involvement of other ovarian factors. Ovarian InhA and Insl3 expression, and plasma progesterone were unchanged in CART peptide–treated mice. In addition to the absence of differential expression, inhibin A is unlikely to mediate the effect of CART on bone because it is effective in male mice and mainly affects osteoblast activity.38 Alternatively, CART peptide could affect the HPG axis centrally. CART peptide has been shown to be capable of at least partly entering the brain from the general circulation52 and to affect gonadotrophs, lactotrophs, and LH secretion.36, 53–55 Serum LH, FSH, and prolactin were not significantly changed in CART-treated mice compared to placebo. The measured values, although increased and decreased for LH and FSH, respectively, were within the normal range of the estrus cycle and were not accompanied by morphological changes in the ovary. Interestingly, Yarram and colleagues56 showed that transgenic mice overexpressing hCG do have a bone phenotype that is strikingly similar to the effects on bone induced by CART peptides. The hCGαβ+ transgenic mice showed an increased bone mass that was lost upon OVX and no increase in bone mass was observed in the male transgenic mice. Serum hCG was exceptionally high in these mice, up to 2000-fold increased compared to normal circulating LH activity. This was accompanied by only a transient increase in E2 and markedly increased serum progesterone (60-fold), testosterone (6-fold), and prolactin (80-fold). It seems unlikely that the minor changes in LH, FSH, and E2 observed in our study are responsible for the robust increase in bone mass by CART peptide treatment.
In conclusion, the current study shows that CART peptide treatment can be used as a therapeutic agent for the treatment of premenopausal but not postmenopausal bone diseases. The long t½ of CART peptide facilitates prolonged increase of serum CART peptide through sustained release by an implant. In addition, the experiments described here warrant future studies to identify the ovarian factor needed for CART activity on bone and to further understand the underlying mechanism of action.
All the authors state that they have no conflicts of interest.
This research was funded by Merck Sharp & Dohme. We thank J. van der Honing-Barends, R. Willems, S. Vonsovic for LC-MS analysis of CART peptides and pellets, R.J.B. te Poele for CART PAGE and initial purification of CART peptide, W.M. Koot for production of recombinant human CART 39-89 peptide, T.A.E. Achterberg for testing of activity of CART peptides in vitro, and H.A.M.D. van der Heijden for CART PK experiments.
Authors' roles: NECB, CJMV, FGMB, CP, and RHGML performed the experiments. NECB, CJMV, and RHGML contributed to detailed methods and tables and figures. HJK organized the production of recombinant CART peptides. GMTV organized the CART pharmacokinetic experiments. MAMK organized the histopathological experiments. JAG and HG designed and organized the experiments and reviewed the data. HG wrote the manuscript. JAG reviewed and revised the paper.