Human parathyroid hormone (1–28)NH2 [hPTH(1–28)NH2] is the smallest of the PTH fragments that can fully stimulate adenylyl cyclase in ROS 17/2 rat osteoblast-like osteosarcoma cells. This fragment has an IC50 of 110 nM for displacing 125I-[Nle8,18, Tyr34]bovine PTH(1–34)NH2 from HKRK B7 porcine kidney cells, which stably express 950,000 human type 1 PTH/PTH-related protein (PTHrP) receptors (PTH1Rs) per cell. It also has an EC50 of 23.9 nM for stimulating adenylyl cyclase in ROS 17/2 cells. Increasing the amphiphilicity of the α-helix in the residue 17–28 region by replacing Lys27 with Leu and stabilizing the helix by forming a lactam between Glu22 and Lys26 to produce the [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 analog dramatically reduced the IC50 for displacing 125I-[Nle8,18, Tyr34]bPTH(1–34)NH2 from hPTHIRs from 110 to 6 nM and dropped the EC50 for adenylyl cyclase stimulation in ROS 17/2 cells from 23.9 to 9.6 nM. These modifications also increased the osteogenic potency of hPTH(1–28)NH2. Thus, hPTH(1–28)NH2 did not significantly stimulate either femoral or vertebral trabecular bone growth in rats when injected daily at a dose of 5 nmol/100 g body weight for 6 weeks, beginning 2 weeks after ovariectomy (OVX), but it strongly stimulated the growth of trabeculae in the cancellous bone of the distal femurs and L5 vertebrae when injected at 25 nmol/100 g body weight. By contrast [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 significantly stimulated trabecular bone growth when injected at 5 nmol/100 g of body weight. Thus, these modifications have brought the bone anabolic potency of hPTH(1–28)NH2 considerably closer to the potencies of the larger PTH peptides and analogs. (J Bone Miner Res 2000;15:964–970)
Replacingthe polar Lys27 with nonpolar Leu (as in human parathyroid hormone–related protein [hPTHrP]) in the hydrophobic face of the amphiphilic α-helix of PTH in the 17–28 region and forming a cyclic lactam linkage between Glu22 and Lys26 to stabilize the helix together increase the ability of hPTH(1–31)NH2 to stimulate adenylyl cyclase in rat ROS 17/2 osteosarcoma cells about 6-fold and raise its ability to stimulate trabecular bone growth in the femurs of ovariectomized (OVX) rats to that of hPTH(1–34).(1–3) The question then arose as to whether these modifications could enhance the bioactivities of smaller hPTH fragments, which have reduced adenylyl cyclase–stimulating activities and little or no bone anabolic activity. According to Neugebauer et al. hPTH(1–28)NH2 is the smallest of the C-terminally truncated hPTH fragments that can fully stimulate adenylyl cyclase in ROS 17/2 cells albeit with a somewhat higher concentration needed to obtain 50% of the maximum stimulation (EC50).(4) Even so, it could be predicted from the brief report of Miller et al. that hPTH(1–28)NH2 would not stimulate either adenylyl cyclase or bone growth because hPTH(1–28) was not osteogenic in their OVX rat model.(5)
Here we present the results of experiments that show that the same modifications of the C-terminal α-helix dramatically increase the bioactivities of hPTH(1–28)NH2. We will show that [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 has acquired the receptor-binding ability of hPTH(1–31)NH2 and [Leu27]cyclo(Glu22-Lys26)hPTH(1–31)NH2 and an adenylyl cyclase–stimulating ability, which is twice that of hPTH(1–31)NH2 and hPTH(1–28)NH2. We also will show that contrary to expectations, a high dose of hPTH(1–28)NH2 can stimulate the growth of trabeculae in the cancellous bone of distal femurs and L5 vertebrae of OVX rats and that [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 can stimulate bone growth at a much smaller dose.(5)
MATERIALS AND METHODS
The PTH fragments used in these experiments were made at the Institute for Biological Sciences using the same Fmoc procedure and continuous flow peptide synthesizer (PerSeptive Biosystems Model 9050, Framingham, MA, U.S.A.) described previously.(1)
Adenylyl cyclase activity
Adenylyl cyclase activity was measured in the cells of 4- to 6-day-old cultures of ROS 17/2 cells in 24-well plates. The activity was reflected in the rate of formation of [3H]cyclic adenosine monophosphate (AMP) from the cellular adenosine triphosphate (ATP) pool that had been labeled with [3H]adenine before the cells were exposed to hPTH peptides.(1,3,6) The cells were incubated for 10 minutes after adding a peptide. The reaction was stopped with 10% trichlororacetic acid, and the [3H]cyclic AMP was then separated and measured.(6) The EC50's of the resulting dose-response curves were determined by fitting the data to a sigmoidal function (Table Curve, Jandel, San Rafael, CA, U.S.A.).
Because of the intrinsically low receptor number in ROS 17/2 cells, competitive displacement of radioligand binding to hPTH1Rs (type 1 PTH/PTHrP receptors) was conducted using the cell line HKRK B7, a subclone of LLC-PK1 porcine renal epithelial cells that is stably transfected with complementary DNA (cDNA) encoding the human PTHR1 and that expresses 950,000 hPTH1Rs per cell. Cells were seeded at 105/cm2 in 24-well plates and cultured as previously described for 2 days before use.(7) The radioligand 125I-[Nle,8,18Tyr34]bovine (b)PTH(1–34)NH2, iodinated by the chloramine-T method and purified by high-performance liquid chromatography (HPLC), was added to confluent cell monolayers in cold binding buffer to which competing nonradioactive peptides were added immediately.(7) The plates were incubated for 6 h at 15°C and then washed extensively with cold buffer and the cells solubilized for counting of bound radioactivity.(7) Specific binding was obtained by subtracting from total bound radioactivity per well the counts per minute of nonspecific binding observed in the presence of an excess (1000 nM) of nonradioactive hPTH(1–34)NH2. In two large experiments reported here, total binding ranged from 25% of 150,000 cpm added per well to 33% of 110,000 cpm added per well, whereas nonspecific binding was 3–4% of total bound radioactivity. Results are presented as percentages of the maximal specific binding (%B/Bmax) observed in the presence of the tracer alone. The data subsequently were analyzed by nonlinear regression using a four-parameter logistic model of the form y = min + (max – min)/[1 + (IC50/x)n], where y is %B/Bmax, x is the concentration of added competing peptide, n is the slope, and IC50 is the concentration of peptide causing 50% displacement of specifically bound tracer. Curve fitting and derivation of the IC50 values was performed using the “Solver” function in Microsoft Excel to accomplish iterative least squares minimization of the variance between the observed and predicted data (“max” was constrained to 100%).
Bone growth assay
In this study we have used OVX, 3-month-old, sexually mature rats. This model combines the easy manageability of rats with the principal features of human osteoporosis. Moreover, contrary to the current belief, it now appears that rats also remodel their vertebral trabecular bone as do humans.(8)
Normal, sham-OVX and OVX Sprague-Dawley rats were bought from Charles River Breeding Laboratories (St. Constant, QC, Canada). On arrival, they were randomly separated into groups of seven to eight animals each [vehicle-injected sham-operated, vehicle-injected OVX, OVX injected with 5 nmol of hPTH(1–28)NH2/100 g of body weight, OVX injected with 25 nmol of hPTH(1–28)NH2/100 g of body weight, and OVX injected with 5 nmol of [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2/100 g of body weight]. The animals were fed Purina rat chow (1.0% calcium and 0.6% phosphorus) and had free access to water. These experiments were approved by the National Research Council's (NRC's) Institute for Biological Sciences Animal Care Committee.
In this study we used a preventative protocol in which the subcutaneous injections of peptide dissolved in an acidic saline vehicle (0.15 M NaCl in water containing 0.001N HCl) were started at the end of the second week after OVX before a significant trabecular loss would have started and then were given once each day, 6 days/week for the next 6 weeks (i.e., until the end of the eighth week after OVX).(9)
At the end of the experiment the femurs and L5 vertebrae were removed. The vertebrae and the distal halves of the femurs minus their epiphyses were fixed in acetate-buffered 10% formalin for 1 week and then demineralized by stirring in 5% trichloroacetic acid for 9–12 days, dehydrated, cleared, and embedded in paraffin wax. Serial 8-μm-thick transverse sections of the dehydrated, cleared, paraffin-embedded L5 vertebral bodies and 10-μm-thick longitudinal sections of the paraffin-embedded distal femur halves were cut with a Leica RM-2035 microtome and stained for 2 minutes at 55°C with Sanderson's rapid bone stain (Surgipath Medical Industries, Inc., Winnipeg, MB, Canada).
The abilities of the hPTH fragments to stimulate trabecular growth in the cancellous bone of vertebrae and distal femurs were assessed by measuring total trabecular volumes (mm3) and the mean trabecular thicknesses or widths (area [μm2]/perimeter [μm] of about 80 measurements per bone). The trabecular volume (expressed in mm3) was obtained by summing the trabecular areas in five, 8-μm serial sections in the case of the vertebrae and in five, 10-μm serial sections in the case of the distal femurs. The trabecular areas and perimeters were measured with an M4 imaging system from Imaging Research, Inc. (St. Catherine's, ON, Canada). The software used for these analyses was Imaging Research's morphometry version 3.01.7. All of the histomorphometric measurements were made by the same ‘blinded’ observer for either the distal femurs or the L5 vertebrae.
The bone data were expressed as means ± SEMs. Statistical comparisons were made by one-way analysis of variance (ANOVA). Scheffe's test was used for multiple comparisons and p < 0.05 was considered to be significant.
Human PTH(1–34)NH2 displaced 125I-[Nle8,18, Tyr34]bPTH (1–34)NH2 from the human PTH1Rs on HKRK B7 cells with an IC50 of 2.8 nM (Fig. 1). The IC50 of hPTH(1–31)NH2 was 6.6 nM, which was marginally higher (p = 0.052) than that of [Leu27]cyclo(Glu22-Lys26)hPTH(1–31)NH2 (4.8 nM; Fig. 1). The relative affinity of hPTH(1–28)NH2 for the human receptors was much less than those of the larger fragments, as indicated by its IC50 of 110 nM (Fig. 1). Introduction of the Leu27 substitution plus the Glu22-Lys26 lactam greatly increased the binding affinity of hPTH(1–28)NH2 as indicated by the reduction in IC50 from 110 to 6.3 nM for [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2, which also was similar to the IC50 of [Leu27]cyclo(Glu22-Lys26)hPTH(1–31)NH2 (Fig. 1).
Adenylyl cyclase stimulation
The EC50 of hPTH(1–28)NH2 for stimulating adenylyl cyclase was 23.9 nM ± 5.3 nM (Fig. 2), which was higher than the reported values for hPTH(1–34)NH2 and hPTH(1–31)NH2 (i.e., 16 and 19, respectively).(1,4) Besides having a greatly increased receptor-binding affinity, [Leu27]cyclo (Glu22-Lys26)hPTH(1–28)NH2 was a more active adenylyl cyclase stimulator; its EC50 was only 9.6 nm ± 1.7 nM (Fig. 2). Although this EC50 is lower than those reported for hPTH(1–34)NH2 and hPTH(1–31)NH2, it is still much higher than the reported value for [Leu27]cyclo(Glu22-Lys26)hPTH(1–31)NH2 (i.e., 3.3 nM).(1)
Bone growth: distal femurs
By 8 weeks after OVX, the mean total trabecular volume in five, 10-μm-thick serial sections of the distal femurs of the vehicle-treated OVX rats (0.21 mm3 ± 0.02 mm3) was 50% of the volume (0.43 mm3 ± 0.04 mm3; p < 0.01) in the distal femurs of the accompanying vehicle-treated sham-operated rats (Fig. 3). Injecting 5 nmol of hPTH(1–28)NH2/100 g of body weight once daily, 6 days/week from the end of the second week to the end of the eighth week after OVX reduced the trabecular loss and the mean volume (0.31 mm3 ± 0.03 mm3) was not significantly less than in the sham-operated rats (Fig. 3). In contrast, injecting 5.0 nmol of [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2/100 g of body weight raised the volume significantly to 0.57 mm3 ± 0.03 mm3, which was 2.6 times higher than in the vehicle-treated OVX rats (p < 0.01), higher, though not significantly so (p > 0.05), than in the sham-operated rats, and significantly higher (p < 0.01) than in the rats treated with 5.0 nmol of hPTH(1–28)NH2/100 g of body weight (Fig. 3).
hPTH(1–28)NH2 was able to stimulate bone growth, but only at a much larger dose. Thus, injecting 25 nmol of the fragment increased the mean trabecular volume to 0.75 mm3 ± 0.05 mm3, which was 3.5 times the mean volume in the OVX rats (p < 0.01) and 1.8 times the mean volume (p < 0.01) in the sham-operated animals (Fig. 3).
These changes in the mean trabecular volume in the distal femurs were paralleled by changes in the mean trabecular thicknesses. The mean trabecular thickness in the distal femurs of the OVX rats (34.8 μm ± 1.5 μm) was not significantly less (p > 0.05) than in the sham-operated rats (39.4 μm ± 2.3 μm) by the end of the eighth week after OVX (Fig. 4). In the rats injected with 5.0 nmol of hPTH(1–28)NH2/100 g of body weight the mean trabecular thickness by the end of the eighth week after OVX was 40.4 μm ± 1.2 μm, which was not significantly different (p > 0.05) from the thicknesses in the OVX and sham-operated rats (Fig. 4). By contrast, injecting 5.0 nmol of [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 increased the mean femoral trabecular thickness to 67.6 μm ± 1.9 μm by the end of the eighth week after OVX, which was significantly thicker than in the OVX (p < 0.01) and sham-operated (p < 0.01) rats (Fig. 4). Injecting 25 nmol of hPTH(1–28)NH2/100 g of body weight increased the mean thickness to 88.0 μm ± 2.6 μm, which also was significantly higher than in the OVX (p < 0.01) and sham-operated (p < 0.01) rats (Fig. 4).
Bone growth: L5 vertebrae
The mean trabecular volume (0.16 mm3 ± 0.01 mm3) in 40, 8-μm slices of the L5 vertebral bodies at the end of the eighth week after OVX was not significantly different from the volume (0.15 mm3 ± 0.01 mm3) in the sham-operated animals (Fig. 5). In animals injected with 5.0 nmol of hPTH(1–28)NH2 the volume at the end of 8 weeks was 0.18 mm3 ± 0.01 mm3, which was not significantly different (p > 0.05) from the volumes in the sham-operated and OVX animals (Fig. 5). However, injecting either 5.0 nmol of [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 or 25.0 nmol of hPTH(1–28)NH2, respectively and equally, increased the trabecular volume to 0.24 mm3 ± 0.01 mm3 and 0.24 mm3 ± 0.01 mm3, which were significantly (p < 0.01) higher than the volumes in the sham-operated and OVX animals (Fig. 5).
By the end of the eighth week after OVX the mean trabecular thickness in the L5 vertebrae was 37.4 μm ± 1.6 μm, which was not significantly different from the 35.9 μm ± 2.2 μm in the sham-operated animals (Figs. 6 and 7). When the injections of 5.0 nmol of hPTH(1–28)NH2/100 g had ended, the mean trabecular thickness was 44.6 μm ± 2.5 μm, which was not significantly different from the OVX and sham-operated values (Figs. 6 and 7). In contrast to the greater effectiveness of 25 nmol of hPTH(1–28)NH2 in increasing the femoral trabecular thickness, 5.0 nmol of [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 and 25 nmol of hPTH(1–28)NH2 nearly equally (p > 0.05) increased the mean vertebral trabecular thickness. Thus, the injections of only 5.0 nmol of [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 increased the mean trabecular thickness to 70.1 μm ± 3.1 μm in contrast to the 25 nmol injections of hPTH(1–28)NH2 needed for the same (73.1 μm ± 2.1 μm) increase (Figs. 6 and 7). Both of these thicknesses were significantly higher (p < 0.01) than those in the L5 vertebrae of the OVX and sham-operated rats.
Barbier et al. and Whitfield et al. reported that they had enhanced adenylyl cyclase–activating and osteogenic potencies of hPTH(1–31)NH2 by introducing two modifications that they expected would increase the peptide's ability to bind to PTH/PTHrP receptors.(1,3) First, they had increased the hydrophobicity of the receptor-binding hydrophobic face of the amphiphilic α-helix in residues 17–28 by replacing Lys27 with Leu, as in the native hPTHrP.(2) This alone reduced the EC50 for adenylyl cyclase stimulation in ROS 17/2 cells to about one-third of that in hPTH(1–31)NH2. They further doubled the fragment's adenylyl cyclase–stimulating effectiveness by lactam formation between Glu22 and Lys26 to stabilize locally the α-helix. The resulting 6-fold enhancement of adenylyl cyclase potency for the [Leu27]cyclo(Glu22-Lys26)hPTH(1–31)NH2 analog was greater than the more modest, nearly significant (p = 0.052) 1.4-fold improvement in binding IC50 for the hPTHR1 observed here (i.e., 4.8 nM vs. 6.6 nM). On the other hand, the effect on receptor-binding affinity of increasing the hydrophobicity and stabilizing the receptor-binding helix was a more dramatic 15-fold increase for the weakly bound hPTH(1–28)NH2 fragment than for hPTH(1–31)NH2, but there was only a 2.5-fold enhancement of adenylyl cyclase potency. But different though these relative enhancements and the HKRK B7 and ROS 17/2 models might be, these modifications can strikingly increase the bioactivity of hPTH(1–28)NH2.
The superior receptor-binding and adenylyl cyclase-stimulating properties of [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 have brought the potency of the peptide for stimulating trabecular growth in the distal femurs and L5 vertebrae of OVX rats much closer to those of the larger fragments [Leu27]cyclo(Glu22-Lys26)hPTH(1–31)NH2 and hPTH(1–34). The most likely explanation for the analog's enhanced osteogenic potency would be a greater apparent affinity for the PTH1R, which would be consistent with the results obtained with the HKRK B7 kidney cells with their high density of human receptors. These modifications also may have increased the analog's stability in vivo and with this its ability to enter the circulation from the injection sites and survive long enough to stimulate its target bone cells.(1,10) Regardless of the reason(s) for its effectiveness, [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 has now replaced hPTH(1–30)NH2 as the smallest osteogenic PTH fragment.(11) Because the native PTH and its N-terminally intact, adenylyl cyclase–stimulating fragments stimulate bone growth in osteoporotic humans as well as mice, rats, and monkeys, [Leu27]cyclo(Glu22-Lys26)hPTH(1–28)NH2 could have a role in the development of an effective anabolic drug for treating osteoporosis, particularly because such a small, yet still significantly osteogenic, fragment might be more readily adaptable for noninjectable delivery than the larger fragments.(12) It also might be more selectively osteogenic than the larger fragments in view of recent reports that hPTH(1–31)NH2 is as anabolic as hPTH(1–34) but is significantly less able than hPTH(1–34) to stimulate bone resorption in humans and mice.(13,14)