Proline/arginine-rich end leucine-rich repeat protein N-terminus is a novel osteoclast antagonist that counteracts bone loss


Address correspondence to: Anna Teti, PhD, Department of Biotechnological and Applied Clinical Sciences, Via Vetoio–Coppito 2, 67100 L'Aquila, Italy. E-mail:


hbdPRELP is a peptide corresponding to the N-terminal heparin binding domain of the matrix protein proline/arginine-rich end leucine-rich repeat protein (PRELP). hbdPRELP inhibits osteoclastogenesis entering pre-fusion osteoclasts through a chondroitin sulfate– and annexin 2–dependent mechanism and reducing the nuclear factor-κB transcription factor activity. In this work, we hypothesized that hbdPRELP could have a pharmacological relevance, counteracting bone loss in a variety of in vivo models of bone diseases induced by exacerbated osteoclast activity. In healthy mice, we demonstrated that the peptide targeted the bone and increased trabecular bone mass over basal level. In mice treated with retinoic acid to induce an acute increase of osteoclast formation, the peptide consistently antagonized osteoclastogenesis and prevented the increase of the serum levels of the osteoclast-specific marker tartrate-resistant acid phosphatase. In ovariectomized mice, in which osteoclast activity was chronically enhanced by estrogen deficiency, hbdPRELP counteracted exacerbated osteoclast activity and bone loss. In mice carrying osteolytic bone metastases, in which osteoclastogenesis and bone resorption were enhanced by tumor cell–derived factors, hbdPRELP reduced the incidence of osteolytic lesions, both preventively and curatively, with mechanisms involving impaired tumor cell homing to bone and tumor growth in the bone microenvironment. Interestingly, in tumor-bearing mice, hbdPRELP also inhibited breast tumor growth in orthotopic sites and development of metastatic disease in visceral organs, reducing cachexia and improving survival especially when administered preventively. hbdPRELP was retained in the tumor tissue and appeared to affect tumor growth by interacting with the microenvironment rather than by directly affecting the tumor cells. Because safety studies and high-dose treatments revealed no adverse effects, hbdPRELP could be employed as a novel biological agent to combat experimentally induced bone loss and breast cancer metastases, with a potential translational impact.


Proline/arginine-rich end leucine-rich repeat protein (PRELP) is a matrix protein expressed in basement membranes, cartilage, and bone matrices.[1, 2] It has a potent antiresorptive effect inhibiting osteoclast formation through its glycosaminoglycan-binding domain encompassing the N-terminus of the protein.[3] A peptide representing this region was denominated heparin-binding domain of PRELP (hbdPRELP) because it was isolated for its capacity to bind heparin and heparan sulfate.[4] Further studies demonstrated that the peptide interacted with chondroitin sulfates as well and that this interaction induced its internalization in osteoclast vesicles through a molecular association with annexin II. Internalization appeared specific for osteoclasts and osteoclast precursors and was followed by transport of the peptide to the nucleus, where it accumulated and physically interacted with the p65 subunit of the nuclear factor-κB (NF-κB) transcription factor, leading to a >50% inhibition of its transcriptional activity.[3] In osteoclast precursors, NF-κB activity is under the control of the receptor activator of NF-κB (RANK)/RANK ligand (RANKL) axis, the most potent pro-osteoclastogenic pathway known so far. By interfering with this important inducer of osteoclastogenesis, hbdPRELP powerfully impaired osteoclast formation both in vitro and in vivo.[3]

Osteoclasts are implicated in the physiological process of bone remodeling in concert with osteoblasts. They remove the bone matrix that will later be replaced by new bone matrix produced and mineralized by the osteoblasts. The balanced activity of the two cell types replaces the damaged bone, leaving the total bone mass intact.[5] Disruption of such equilibrium because of an excess of bone resorption over bone formation causes the bone mass to be reduced and the bones to become fragile and prone to fracture. There are two major diseases in which excess of bone resorption is known to be the cause of severe osteopenic syndromes, osteoporosis and osteolytic bone metastases, which have serious consequences on the health of affected individuals.

Osteoporosis can be primary or secondary to many pathological conditions, or it can be iatrogenic, especially resulting from chronic treatments.[6] It affects 1 in 3 women and 1 in 12 men over the age of 50 years,[7] but children and adolescents also can suffer from secondary or iatrogenic osteoporosis. The impact during childhood may be immediate, resulting in fragility fractures, or delayed, because of suboptimal peak bone mass.[8]

Bone metastases are typical of certain types of cancers characterized by elevated osteotropism, including breast carcinomas.[9] It has been estimated that at least 70% of patients who die with mammary cancers present with skeletal metastases at autopsy.[9] When bone metastases develop, life quality and expectancy dramatically drop because bone relapse is associated with bone pain, pathological fractures, spinal cord compression, and life-threatening hypercalcemia.[9, 10] Breast cancer bone metastases are osteolytic in nature; however, tumor cells do not resorb the bone per se, but they produce factors that stimulate osteoclast formation and bone resorption, creating a physical space into which the tumor expands. At the same time, osteoclast activity releases growth factors stored in the bone matrix, which, in turn, stimulate tumor cell growth. This reciprocal stimulation between bone and tumor cells fuels the so-called vicious cycle that perpetuates itself, expanding the metastases.[11] Such a vicious cycle also appears prominent in primary tumors of bone, such as the osteosarcoma, a malignancy with poor prognosis that affects children and young adults especially.[12]

Cancer survivors often develop secondary osteoporosis, not otherwise seen until recently because the disease was much more rapidly fatal in the past than at present. Exacerbated osteoclast activity is a common feature of both osteoporosis and osteolytic bone metastases, and blocking osteoclast formation represents an attractive approach to fight both diseases. Antiresorptive drugs, such as bisphosphonates and humanized anti-RANKL antibody (denosumab), are currently available to treat patients affected by severe bone loss, but their long-term use has revealed a number of pitfalls, suggesting caution, and in cancer they unfortunately do not prolong life because of their inability to dramatically affect tumor cells.[13, 14] Therefore, the identification of new treatments is desirable to expand the therapeutic options and improve the medical approach to bone-loss diseases.

In this study, we demonstrated that hbdPRELP was efficacious at preventing and curing both osteoporosis and breast cancer osteolytic metastases in a variety of preclinical in vivo models. The peptide was well tolerated and also proved effective in inhibiting the tumor itself, both in orthotopic sites and in distant metastases in visceral organs, reducing cachexia and improving survival.

Materials and Methods


In a previous work, we showed that the human PRELP N-terminal end did bind to heparin, whereas a truncated version lacking the N-terminal 22 amino acids did not.[4] To add a short spacer to the very heparin-binding domain, the currently used synthetic peptide (hbdPRELP) was extended with two amino acids to represent the N-terminal 24 amino acids of the human PRELP. This synthetic peptide (NH2-QPTRRPRPGTGPGRRPRPRPRPTPC-COOH) was synthesized by Schafer-N (Copenhagen, Denmark) or Genscript (Piscataway, NJ, USA) with an additional cysteine residue in its C-terminal end, as already described.[3] Aliquots of the peptide were conjugated with Alexa-488 via their cysteine according to standard protocols. Level of labeling was checked by high-performance liquid chromatography and mass spectrometry. Additional peptides were synthesized with an N-terminal biotin tag. The control peptide was 14 amino acids long and corresponded to the heparin-binding domain of another matrix protein, chondroadherin (NH2-CKFPTKRSKKAGRH-COOH).[3] All peptides were verified by mass spectrometry.


Procedures involving animals and their care were conducted in conformity with national and international laws and policies (European Economic Community Council Directive 86/609, OJ L 358, 1, December 12, 1987; Italian Legislative Decree 116/92, Gazzetta Ufficiale della Repubblica Italiana no. 40, February 18, 1992; National Institutes of Health Guide for the Care and Use of Laboratory Animals, National Institutes of Health Publication no. 85–23, 1985) and received institutional approval. Animals were purchased from Charles River (Milan, Italy).

Treatment of mice subjected to intracardiac injection of tumor cells

Human breast cancer MDA-MB-231 cells stably transfected with the firefly lucifearse (MDA-LUC cells) were injected (1 × 105/100 µL PBS) into the left ventricle of 4-week-old female BALB/c-nu/nu athymic mice anesthetized with ip injections of pentobarbital (60 mg/kg bw) and treated as described in Fig. 1A and Fig. 4A (number of animals/group = 20). Animals were daily monitored for cachexia (evaluated by body weight waste), behavior, and survival. At the end of the experiment (35 days from tumor-cell injection), mice were euthanized and subjected to X-ray analysis and to anatomical dissection for evaluation of bone and visceral metastases, respectively. Percents of mice that developed cachexia, survived for the length of the experiment, and developed visceral metastases over the total number of mice employed in each experiment, as well as the percents of hindlimbs that developed osteolytic lesions over the total number of hindlimbs were then calculated.

Figure 1.

Effect of hbdPRELP on breast cancer–induced disease. (A) Female BALB/c-nu/nu mice were intracardially injected with 1 × 105 cells/100 µL PBS of MDA-MB-231 cells stably transfected with luciferase (MDA-LUC). From day 1 post-cell inoculation, animals were treated 5 days/week with vehicle (PBS), control peptide (15 mg/kg bw), hbdPRELP (3.7 and 15 mg/kg bw), or alendronate (1 mg/kg bw) until day 35. Evaluation of (B) cachexia (decrease of body weight), (C) survival, and (D) bone metastasis incidence. (E) ELISA of serum levels of CTX. (F) µCT images of proximal tibias and quantification of (G) trabecular bone volume/total tissue volume (BV/TV). Histomorphometric evaluation of osteoclast (H) number and (I) surface/bone surface. (J) Representative pictures of histological sections of tibias stained with hematoxylin/eosin to quantify (K) the tumor area (T = tumor). Data are (B–E) the mean analyzed by the chi-square test, (G–I, K) the mean± SEM or (F, J) representative of 20 mice/group. In B, ap = 0.011 versus vehicle and control peptide. In C, ap = 0.03 versus vehicle and control peptide. In D, ap = 0.04 and bp = 0.028 versus vehicle and control peptide. In E, ap = 0.005 and bp = 0.003 versus vehicle and control peptide. In G–I, K, ap < 0.045 and bp < 0.008 versus vehicle and control peptide.

Treatment of mice subjected to intratibial implant of tumor cells

Four-week-old female BALB/c-nu/nu mice were anesthetized as described above, then a syringe with a 27½G needle was inserted in the proximal end of the tibia and 5 × 104 MDA-LUC cells suspended in 10 μL PBS were injected into the intramedullary space (number of hindlimbs/group = 8). Mice were then treated as described in Fig. 1A and evaluated for the percent of hindlimbs harboring bone tumors over the total number of hindlimbs injected with the tumor cells, as well as for the extension of osteolytic lesions, as described below.

Micro-computed tomography (µCT) analysis

Tibias and femurs fixed in 4% formaldehyde were mounted and scans acquired in a SkyScan 1174 (Bruker microCT, Kontich, Belgium) with a voxel size of 6 µm (X-ray voltage 50 kV). The scans were more than 180 degrees with a 0.3-degree rotation step. Image reconstruction was carried out employing a modified Feldkamp algorithm using the Skyscan Nrecon software. Beam hardening correction and Fourier transform-based ring artifact reduction were applied to the reconstructed images. 3D and 2D morphometric parameters were calculated for the trabecular bone of selected regions of interest, 150 slides 400 µm from the growth plate. Threshold values were applied for segmenting trabecular bone corresponding to bone mineral density values of 0.6/cm3 calcium hydroxyapatite. 3D parameters were based on analysis of a Marching Cubes type model with a rendered surface. Calculation of 2D areas and perimeters was based on the Pratt algorithm.

Histology and histomorphometry

For histomorphometric analysis, tibias were dissected, fixed in 4% formaldehyde in 0.1 M phosphate buffer, pH 7.2, embedded in glycol-methacrylate, and sectioned longitudinally to the frontal plate using a Reichert-Jung 1150/Autocut microtome. Sections (4 µm thick) were then stained for the osteoclast-specific marker TRAcP using the Sigma-Aldrich (St. Louis, MO, USA) kit #386, according to the manufacturer's instructions. Histomorphometric measurements were carried out with an interactive image analysis system (IAS 2000, Delta Sistemi, Rome, Italy), consisting of a color video-equipped computer linked to the microscope by a videocamera.

Evaluation of osteolysis

Mice intracardially or intratibially injected with tumor cells were weekly anesthetized and subjected to X-ray analysis (36 KPV for 10 seconds) using a Cabinet X-ray system (Faxitron model n.43855A; Faxitron X-Ray Corp., Buffalo Grove, IL, USA) to follow the onset and progression of osteolytic lesions. Radiographs were scanned using the Bio-Rad scanning densitometer (Hercules, CA, USA), model GS800, and quantification of the area of interest was done using the Bio-Rad Quantity One image analysis software.

Bioluminescence reporter imaging (BRI)

Animals injected with MDA-LUC cells were weekly anesthetized and ip injected with 250 mmol/L of luciferin, 5 minutes before photon recording by the Aequoria 2D luminescence imaging system (Hamamatsu Photonics, Naka-ku, Japan). The bioluminescent signal was quantified by measuring the amount of highlighted pixel in the regions of interest shaped around each site of photon emission.

Treatment of mice subjected to orthotopic tumor cell injection

Four-week-old female BALB/c-nu/nu mice were anesthetized as described above, and MDA-LUC cells (1.5 × 106/50 µL PBS) were injected into the fat pad of the inguinal left breast using a tuberculin syringe with a 27½G needle and treated as described in Fig. 1A and Fig. 4A (number of animals/group = 16). One week after tumor-cell injection, tumors began to be evident, and we started to calculate tumor volume twice a week by measuring their three axes using a caliper and applying the formula of the volume of an ellipse [4/3 × π × (a × b × c)]. Tumor growth was also weekly evaluated by BRI, as described above.

Tumor histology and immunohistochemistry

Orthotopic tumors were excised, fixed in 4% formaldehyde in 0.1 M phosphate buffer, pH 7.2, and embedded in paraffin. Sections were cut using a Reichert-Jung 1150/Autocut microtome. Slide-mounted tissue sections (4 μm thick) were deparaffinized in xylene and hydrated serially in 100%, 95%, and 80% ethanol and stained with hematoxylin/eosin, or endogenous peroxidases were quenched in 3% H2O2 in PBS for 1 hour, then sections were incubated with the anti-Ki67 or anti-CD31 primary antibody for 1 hour at room temperature. Sections were washed in PBS and antibody binding was revealed using the Ultra-Vision Detection System anti-Polyvalent HRP/DAB kit according to the manufacturer's instructions.

Fluorescence microscopy

Orthotopic tumors from animals treated with vehicle or biotin tagged hbdPRELP were fixed and cut in 4-μm-thick sections, as described above. Sections were then incubated with fluorescent Alexa-488 streptavidin and observed by fluorescence microscopy, using a Zeiss Axioplan (Zeiss, Inc., Thornwood, NY, USA) microscope. Nuclei were counterstained with 4,6-diamidino-2-phenylindole (DAPI).


Statistical analysis of data expressed as incidence (frequency at which a new event occurs during a specific period) was performed by the chi-square test,[15, 16] a method that does not provide error bars. Statistical analysis of two population means was performed by the unpaired Student's t test. Statistical differences comparing multiple means were analyzed by the analysis of variance (ANOVA). Data were expressed as mean (chi-square test), or mean ± SEM (unpaired Student's t test and ANOVA). A p value <0.05 was conventionally considered statistically significant and reported, along with the statistic method and the number of animals or the number of experiments, in the figure and table legends.


In vivo bone targeting by hbdPRELP

We first assessed hbdPRELP tissue targeting in healthy mice by ip injection of hbdPRELP tagged with Alexa Fluor 488 (Alexa-488). Thirty minutes after injection, there was detectable Alexa-488 fluorescence in liver, kidney, spleen, muscle, and lung. In contrast, no fluorescence was retained in these organs after 5 hours (Supplemental Fig. S1A), and in brain fluorescence was not detectable either after 30 minutes or after 5 hours from injection (Supplemental Fig. S1A). In contrast, in tibias and calvarias there was an accumulation of the fluorescent signal even after 5 hours from Alexa-488-tagged hbdPRELP injection (Supplemental Fig. S1A), suggesting a longer retention of the peptide in the skeletal tissue, possibly because of specific binding.

Safety of in vivo treatment with hbdPRELP

We found no adverse effects in mice treated with hbdPRELP at a dose as high as 50 mg/kg administered daily, 5 days/week for 8 weeks, as demonstrated by the normal histological appearance of liver, kidney (Supplemental Fig. S1B), lung, spleen, and heart (not shown), the normal weights of liver and spleen (Supplemental Fig. S1C), and the normal serum biomarkers of kidney, pancreas, and liver functions (Supplemental Table S1). Finally, enzyme-linked immunosorbant assay (ELISA) in serum samples showed no immune response in hbdPRELP-treated mice versus controls (Supplemental Fig. S1D).

Impact of hbdPRELP on bone microarchitecture

We next performed a µCT analysis of tibia proximal secondary spongiosa of mice employed for testing the safety. We observed that, after 8 weeks of treatment, hbdPRELP increased the trabecular bone volume over total tissue volume, with an increment of trabecular number and a decrement of trabecular separation (Supplemental Table S2). In these samples, we noted, by histomorphometry, that hbdPRELP reduced osteoclast number and surface/bone surface, with no effect on the osteoblast parameter (Supplemental Table S3). Taken together, these results demonstrate that hbdPRELP is safe, targets the bone, and increases the basal bone mass through an anti-osteoclastic effect.

Effect of hbdPRELP on exacerbated osteoclast activity

Next, we addressed the effect of hbdPRELP in an acute model of increased osteoclast formation. To this aim, we treated mice with retinoic acid, which rapidly increased the TRAcP 5b isoform in the serum and the osteoclast number and surface over bone surface in histological sections. These increments were markedly inhibited by the treatment with hbdPRELP. In contrast, a control peptide was ineffective compared with vehicle-treated mice (Supplemental Fig. S2).

Effect of hbdPRELP on experimentally induced osteoporosis

We next investigated whether hbdPRELP could counteract overt osteoporosis. To this aim, we induced a generalized bone loss in mice by bilateral OVX and applied a curative protocol starting the treatments 5 weeks post-OVX (Supplemental Fig. S3). After 6 weeks of therapy, µCT analysis of distal femurs showed that, similar to the reference drug alendronate, hbdPRELP significantly counteracted the decrease of bone volume/total volume observed in vehicle-treated OVX mice, with an improvement of trabecular number and separation and no significant effect on trabecular thickness (Supplemental Table S4). In histological sections of tibia proximal spongiosa, we observed a significantly lower osteoclast number and surface over bone surface in OVX mice treated with hbdPRELP compared with the control groups (Supplemental Table S5). Moreover, hbdPRELP reduced the OVX-induced increase of the serum bone resorption marker collagen type 1 cross-linked telopeptide (CTX) (Supplemental Table S5). In contrast, no change in the osteoblast parameter was observed (Supplemental Table S5). These data confirmed our previous results using a preventive treatment protocol[3] and demonstrate that hbdPRELP has a powerful anti-osteoporotic effect, leading to an improvement of the bone phenotype also by a curative treatment.

Effect of hbdPRELP on breast cancer-induced metastases

We next tested the effect of hbdPRELP on the development of experimental bone metastases. Four-week-old female Balb/c nu/nu athymic mice were injected in the left ventricle with bioluminescent human breast cancer MDA-LUC cells and treated with vehicle, hbdPRELP, or our control peptide starting the day after tumor-cell injection, according to an adjuvant therapy (Fig. 1A). Compared with the control groups, hbdPRELP-treated animals showed dose-dependent decrease of cachexia (Fig. 1B), increase of survival (Fig. 1C), decrease of incidence of bone metastases (Fig. 1D, Table 1), together with a reduction of the bone resorption marker CTX (Fig. 1E). Consistently, in hbdPRELP-treated mice, decreased levels of trabecular bone volume/total tissue volume and osteoclast parameters were observed compared with the control groups (Fig. 1F–I). Tumor area in tibias, evaluated in histological sections stained with hematoxylin/eosin, was also reduced (Fig. 1J, K). Our peptide was able to reduce bone metastasis incidence and serum CTX to a similar extent as alendronate (Fig. 1B–E), whereas, at variance with hbdPRELP, alendronate failed to reduce cachexia (Fig. 1B) and to improve survival (Fig. 1C). Weekly bioluminescence and X-ray analyses showed a significant decrease of tumor growth and osteolytic area over time by treatment with hbdPRELP at the highest dose (Fig. 2A–D). The peptide induced a slight inhibition of the incidence of visceral metastases (Fig. 2E), whereas, again in contrast with alendronate, it caused a strong and significant reduction of their size (Fig. 2F). Histological sections of lungs confirmed the antitumoral effect of hbdPRELP in these organs (Fig. 2G). The outcome was dependent on frequency of administrations, with increasing effects from 1 to 5 days/week treatment regimens (Fig. 3A–H), whereas higher doses (30 and 60 mg/kg) did not improve the efficacy of hbdPRELP (data not shown).

Table 1. Incidence of Bone Metastases (%) in Mice Intracardially Injected With MDA-LUC Cells and Treated ip With Vehicle, 3.7 or 15 mg/kg bw hbdPRELP, Administered Daily 5 Days/Week for 4 Weeks (See Fig. 1A)
TreatmentVehiclehbdPRELP 3.7 mg/kghbdPRELP 15 mg/kg
  1. Number of hindlimbs/group = 8.
  2. ap < 0.05 versus vehicle (chi-square test).
2nd week50%43%29%
3rd week57%43%29%a
4th week64%50%29%a
Figure 2.

Effect of hbdPRELP on breast cancer–induced disease. Female BALB/c-nu/nu mice were intracardially injected with 1 × 105 cells/100 µL PBS of MDA-MB-231 cells stably transfected with luciferase (MDA-LUC). From day 1 post-tumor cell inoculation, animals were treated ip 5 days/week with vehicle (PBS), control peptide (15 mg/kg bw), hbdPRELP (3.7 and 15 mg/kg bw), or alendronate (1 mg/kg bw) for 35 days. (A) BRI analysis and (B) quantification of the bioluminescence signal recorded at the 2nd, 3rd, and 4th week from cell injection. (C) X-ray and (D) quantification of osteolytic area, evaluated by densitometric analysis, performed at the 2nd, 3rd, and 4th week from cell injection. (E) Incidence of visceral metastases and (F) visceral metastasis size evaluated by BRI. (G) Histological sections of lung metastases stained with hematoxylin/eosin (asterisk = tumor). Scale bar = 400 µm. Data are representative (A, C, G), the mean analyzed by the chi-square test (E) or the mean± SEM (B, D–F) of 20 mice/group. In B, D, ap < 0.05 versus vehicle (ANOVA). In F, ap = 0.045 versus vehicle (unpaired Student's t test). In E, data are not statistically significant (p = 0.2).

Figure 3.

Effect of various protocols of administration of hbdPRELP on breast cancer–induced disease. Four-week-old female BALB/c-nu/nu mice were intracardially injected with a suspension (1 × 105 cells/100 µL PBS) of MDA-LUC cells. From the day after cell inoculation, animals were treated ip with vehicle (PBS) or hbdPRELP (15 mg/kg bw) 1, 2, and 5 days/week until the end of the experiment (35 days). Evaluation of (A) cachexia, (B) survival, (C) bone metastasis incidence, and (D) serum levels of the bone resorption marker CTX. (E) X-ray and (F) quantification of the osteolytic area, evaluated by densitometric analysis, performed at the 3rd and 4th week from cell injection. (G) Incidence of visceral metastases and (H) visceral metastasis volume. (I) Combination treatment of hbdPRELP with doxorubicin (DXR). Four-week-old female BALB/c-nu/nu mice were intracardially injected with a suspension (1 × 105 cells/100 µL PBS) of MDA-LUC cells. From the day after cell inoculation, animals were treated ip with vehicle (PBS) or hbdPRELP (15 mg/kg bw), doxorubicin (DXR, 0.2 mg/kg bw), or the latter two compounds in combination (hbdPRELP + DXR), administered daily 5 days/week until the end of the experiment. In A, ap = 0.048 versus vehicle; in B, C, ap = 0.035 versus vehicle (chi-square test). In D, ap = 0.01 versus vehicle (unpaired Student's t test). In F, ap = 0.048 versus vehicle (ANOVA). In G, data are statistically nonsignificant (p > 0.24) (chi-square test). In H, ap = 0.047 versus vehicle (unpaired Student's t test). In I, ap = 0.03, bp = 0.045, and cp = 0.04 versus vehicle (chi-square test). Number of mice/group = 10.

In this context, we also addressed if the treatment with hbdPRELP could improve the outcome of conventional chemotherapy on the incidence of bone metastases. Animals treated with hbdPRELP and doxorubicin (DXR, 0.2 mg/kg) alone or in combination showed a trend to a lesser incidence of bone metastases in the combined treatment versus each of the single therapies (Fig. 3I).

Effect of hbdPRELP on overt metastases

To test if the peptide was also active on overt metastatic disease, mice were subjected to intracardiac inoculation of MDA-LUC cells and after 2 weeks they were analyzed by BRI to identify those with initial metastases (Fig. 4A, B), which were divided in two groups receiving the vehicle or 15 mg/kg of hbdPRELP, daily 5 days/week for 3 weeks (Fig. 4A, B). We observed a trend of a decrease of cachexia (Fig. 4C) and improvement of survival (Fig. 4D) in hbdPRELP-treated animals versus controls. We also observed a significant decline of the incidence of bone metastases (Fig. 4E) associated with a decrease of serum CTX levels (Fig. 4F), whereas osteolytic area was slightly, but not significantly, reduced (arbitrary units, 3rd week: vehicle 3.9 ± 1.5, hbdPRELP 3.1 ± 0.8; 4th week: vehicle 6.5 ± 2.2, hbdPRELP 4.8 ± 3.1, p > 0.08). In contrast, visceral metastasis incidence and size were not affected (Fig. 4G, H). These results suggest that hbdPRELP is effective on overt bone metastases but fails to cure visceral metastases when they are already developed.

Figure 4.

Curative protocol of treatment with hbdPRELP. (A, B) Four-week-old female BALB/c-nu/nu mice were intracardially injected with a suspension (1 × 105 cells/100 µL PBS) of MDA-LUC cells and monitored for the onset of metastases by BRI. (B) Those showing a bioluminescent signal (white areas) after 2 weeks from tumor-cell injection were treated ip with vehicle or 15 mg/kg bw hbdPRELP administered daily, 5 days/week until the end of the experiment. Evaluation of incidence of (C) cachexia, (D) survival, and (E) bone metastases. (F) Measurement of serum levels of CTX. (G) Incidence of visceral metastases and (H) visceral metastasis volume. In E, ap = 0.040 versus vehicle (chi-square test). In F, ap = 0.02 versus vehicle (unpaired Student's t test). In C, D, G, H, data are statistically not significant; C, D, G, p > 0.61 (chi-square test) and H, p > 0.9 (unpaired Student's t test), respectively. Number of mice/group = 10.

Effect of hbdPRELP on breast cancer cell growth in the bone

To distinguish between the effect of hbdPRELP on homing of the tumor cells to the bone microenvironment and tumor growth in the same microenvironment, we eliminated the former by directly inoculating the tumor cells into the tibia medullary cavity. The treatment with hbdPRELP by an adjuvant protocol (Fig. 1A) did not significantly alter the incidence of osteolytic lesions, although there was a tendency (Table 2), suggesting that part of the inhibition observed when the tumor cells were injected in the heart was because of the impairment of their homing to the bone microenvironment. In contrast, hbdPRELP decreased the growth of tumor cells (Fig. 5A, B) and the extension of tumor-induced osteolysis (Fig. 5C, D). µCT analysis confirmed a less intense disruption of trabecular bone in tibias of mice treated with hbdPRELP relative to vehicle-treated mice, with a consequent higher percentage of trabecular bone volume over total tissue volume and a better preserved trabecular microarchitecture (Fig. 5E). Moreover, the serum bone resorption marker CTX (Fig. 5F), the intensity of TRAcP staining, and the number and surface of osteoclasts per bone surface in histological sections of tibias (Fig. 5G) were reduced in mice treated with hbdPRELP compared with control mice.

Table 2. Incidence of Osteolytic Lesions (%) in Mice Intratibially Injected With MDA-LUC Cells and Treated ip With Vehicle or 15 mg/kg bw hbdPRELP, Administered Daily 5 Days/Week for 4 Weeks (See Fig. 1A)
  1. Number of hindlimbs/group = 8. Data are statistically not significant (chi-square test).
2nd week25%25%
3rd week50%37.5%
4th week62.5%50%
Figure 5.

Effect of hbdPRELP on tumor growth in bone. BALB/c-nu/nu mice were intratibially injected with MDA-LUC cells (1 × 104 cells/10 µL PBS) and with 15 mg/kg bw hbdPRELP 5 days/week from day 1 post-cell inoculation for 4 weeks. (A) BRI analysis and (B) quantification of the bioluminescence signal recorded at the 2nd, 3rd, and 4th week from cell injection. (C) X-ray, performed the 4th week from cell inoculation, of hindlimbs of mice treated with vehicle or hbdPRELP. White contours = osteolytic lesions. (D) Osteolytic area evaluated by densitometric analysis of X-ray. (E) Tibias from vehicle or hbdPRELP-treated mice analyzed by µCT and quantification of trabecular BV/TV%. (F) ELISA for serum CTX. (G) Histochemical detection of TRAcP in proximal tibias of vehicle- and hbdPRELP-treated animals (Nomarsky microscopy). Arrows = osteoclasts, B = bone, T = tumor. Scale bar = 50 µm. Measurements of OcS/BS and OcN/BS are shown below the images. In B, D, ap = 0.045 versus vehicle (ANOVA). In E, F, ap = 0.01 versus vehicle (unpaired Student's t test). In G, ap = 0.048 versus vehicle (unpaired Student's t test). Number of hindlimbs/group = 8.

Effect of hbdPRELP on orthotopic tumors

To address whether hbdPRELP had a direct inhibiting effect on tumor growth, we obtained orthotopic tumors by inoculation of MDA-LUC cells in the mammary fat pad of BALB/c nu/nu mice. Using the adjuvant protocol of treatment (Fig. 1A), we found a reduction of tumor expansion in hbdPRELP-treated mice, which was significant up to the 24th day of treatment (Fig. 6A, B). However, this effect was not maintained over time because at the end of the experiment (28th day), differences between control and treated groups were no longer significant (Fig. 6A–C). Gross evaluation of tumor samples at autopsy revealed a larger heterogeneity of the tumor size in the group of mice treated with hbdPRELP compared with controls (Fig. 6D). This transient inhibition of orthotopic tumor growth was also observed, although less pronounced, in mice treated by the curative protocol (Fig. 6E), in which they received the hbdPRELP starting from the gross appearance of the tumor (volume = 1 cm3, approximately 2 weeks after the implant, Fig. 4A).

Figure 6.

Effect of hbdPRELP on orthotopic tumor growth. BALB/c-nu/nu mice were subjected to injection of MDA-LUC cells (1.5 × 106/50 µL PBS) into the mammary fat pad and (A–D) treated from day 1 post-cell inoculation (preventive treatment) with vehicle or 15 mg/kg bw hbdPRELP 5 days/week. (A) Tumor volume. (B) BRI recording. Groups of mice were sacrificed after 28th day post-treatment, then tumors were excised and (C) weighted. (D) Macroscopic evaluation of explanted tumors. (E) Mice subjected to orthotopic injection of MDA-LUC cells were treated as described above starting when tumors had reached a volume of 1 cm3 (curative treatment), then tumor volume was measured twice a week until the end of the experiment (fold increase versus tumor volume at the 1st day of treatment, t0). (F–M) Groups of mice were sacrificed at day 21 post-treatment. (F) BRI analysis (arrow = bioluminescent signal). (G) Quantification of the bioluminescence shown in (F). (H) Weights of explanted tumors. (I, J) Histological tumor sections were stained with hematoxylin/eosin and the numbers of (I) cells in mitosis and (J) apoptotic cells were evaluated. (K) Histological sections immunohistochemically stained for the cell proliferation marker Ki67 (inset pictures) and quantification of Ki67-positive nuclei per field (graph). Scale bar = 50 µm. (L) Quantification of percent of necrotic area per tumor area. Inset: representative pictures of the whole tumors stained with hematoxylin/eosin. Black line = necrotic area. (M) Tumor sections were immunohistochemically stained for detection of the endothelial marker CD31 (arrows = CD31 positive endothelium). Results are (A–C, E, G–L) the mean ± SEM or (D, F, M) representative of 16 mice/group (ap < 0.05, bp = 0.002, and cp = 0.01 versus vehicle; unpaired Student's t test).

Mechanism of action of hbdPRELP on orthotopic tumors

To address the mechanism by which tumor expansion was inhibited, we euthanized the animals at the 21st day after orthotopic cell injection, when the differences between the treated and the control groups were obvious. At this time, hbdPRELP-treated mice presented with a less intense bioluminescent signal in the site of tumor-cell injection (Fig. 6F, G). Consistently, tumor weight was significantly reduced (Fig. 6H), and tumor sections stained with hematoxylin/eosin showed reduced mitosis (Fig. 6I) and increased apoptosis (Fig. 6J) in hbdPRELP-treated mice versus controls. In keeping with these observations, the expression of the cell proliferation marker Ki67 was reduced by the treatment with hbdPRELP (Fig. 6K). A trend of decrease of tumor necrotic area was also observed in tumors from mice treated with hbdPRELP (Fig. 6L), whereas blood vessel formation, evaluated by immunohistochemical staining of the endothelial marker CD31, was not influenced (Fig. 6M). Taken together, these results show that hbdPRELP induces apoptosis and inhibits tumor-cell proliferation in vivo, thus explaining the reduced tumor size in orthotopic sites and, perhaps, in visceral metastases.

Mechanism of action of hbdPRELP on tumor cells

At variance with osteoclasts, in our previous work we had found no internalization of hbdPRELP in MDA-MB-231 cells in vitro3 (Fig. 7A, B). To address the discrepancies of the previous in vitro and the current in vivo results, we first investigated the localization of the peptide in the orthotopic tumors by injection of biotin-tagged hbdPRELP. After 24 hours, we indeed detected the presence of the peptide in the tumor tissue (Fig. 7C). To dissect in more detail the interaction between the peptide and the tumor cells, we investigated whether there was any binding of hbdPRELP to MDA-MD-231 cell surface in vitro. We observed that biotin-tagged hbdPRELP decorated the cell surface of cultured MDA-MB-231 when the binding assay was performed a 4°C to prevent membrane turnover (Fig. 7D), a result also confirmed by flow cytometry analysis (Fig. 7E). However, because at 37°C this binding was not followed by biotin-tagged hbdPRELP internalization (Fig. 7B), we propose a mechanism of action different from that of osteoclasts. In confirmation, again at variance with osteoclasts,3 pretreatment of MDA-MB-231 cells with hbdPRELP did not affect their adhesion to substrate (Supplemental Fig. S4A), nor did it affect their proliferation, migration, and invasion ability3 (Supplemental Fig. S4B–D). Moreover, hbdPRELP failed to influence MDA-MB-231 survival (Supplemental Fig. S4E) as well as MAPKs activation pathways (Supplemental Fig. S4F). Taken together, these results indicate that hbdPRELP does not directly alter tumor cell activity in vitro and suggest that the in vivo effect on cancer cells is likely to be mediated by the interaction of hbdPRELP with the tumor microenvironment.

Figure 7.

Binding of hbdPRELP in vivo and in vitro. Confocal microscopy of (A) mouse osteoclasts and (B) MDA-MB-231 cells vitally incubated in vitro for 20 minutes at 37°C with 15 µM Alexa 488-tagged hbdPRELP to monitor localization of the tagged peptide. Results are representative of three independent experiments and confirm previous findings.[3] In B, cells were counterstained with DAPI (right panel) to detect nuclei. Scale bar = 20 µm. (C) Tumor sections from mice treated ip with vehicle or with 15 mg/kg bw biotin-tagged hbdPRELP (BhbdPRELP) evaluated by fluorescence microscopy upon incubation with Alexa-488 streptavidin (lower panels). Sections were also stained with DAPI to detect the nuclei (upper panels). Scale bar = 10 µm. (D) (a, b) MDA-MB-231 cells were vitally incubated in vitro with a biotin-488 streptavidin-tagged hbdPRELP (BhbdPRELP) for 20 minutes at 4°C. (c) Biotin-488 streptavidin (streptav) was used as negative control. (a', b', c') Cells were counterstained with DAPI to detect nuclei. Scale bars = 50 µm (a, a' and c, c') and 10 µm (b, b'). (E) Flow cytometry analysis of in vitro binding of Alexa-488 streptavidin-tagged hbdPRELP (A-488hbdPRELP) to MDA-MB-231 cell surface. Results are representative of (C) 5 animals/group and (A, B, D,–E) three independent experiments.


Although the use of antiresorptive drugs is a well-established strategy to treat osteoporosis and bone metastases, current treatments present important limitations. As an example, bisphophonates severely affect the bone quality and increase the risk of long-term adverse effects, including osteonecrosis of the jaw and atypical femoral fractures, especially in cancer patients who receive higher doses relative to osteoporotic patients.[17, 18] Bisphophonates are also used in children affected by osteogenesis imperfecta and diffuse connective disorders or by osteosarcoma, in which they decrease fracture incidence and increase muscle strength, height, and weight, or counteract osteolytic lesions, respectively. However, the impact on bone modeling in children is not well characterized, and paradox effects inducing deformities and fractures have been described.[19, 20] A promising antiresorptive compound is denosumab, a humanized monoclonal antibody blocking the osteoclastogenic cytokine RANKL,[21, 22] which has just been approved by the Food and Drug Administration in the United States. However, the RANK-RANKL axis also regulates immune cells, and a recent report showed that the most common adverse effects of denosumab in patients with cancer include infections, bone pain, and pain at the extremities and at the joints.[23] Therefore, the long-term adverse effects of anti-RANKL therapy, as well as the impact on the immune response, remain to be established. This body of evidence demonstrates the need for improvements of treatments for bone loss diseases. This is further accentuated by an increasing awareness and diagnoses of osteoporosis in long-term cancer survivors, owing to the previously unrecognized deleterious effects of chemotherapeutics on bone and to the poor calcium intake as a consequence of malabsorption.[24]

In this work, we tested the effects on bone of part of an endogenous matrix protein, PRELP, whose 24mer N-terminal derived peptide (hbdPRELP) previously demonstrated a clear and specific anti-osteoclastogenic effect, with no influence on osteoblast differentiation.[3] We showed that this peptide has a powerful effect in vivo, given both preventively and curatively, in various animal models of healthy and diseased mice. In fact, in healthy mice, in retinoic acid–treated mice (in which an acute increase of osteoclast formation was induced), in OVX mice (in which osteoclast activity was chronically enhanced by estrogen deficiency), and in mice carrying bone metastases (in which osteoclastogenesis and bone resorption were exacerbated by tumor-derived factors), hbdPRELP was equally efficient in rescuing a nearly normal osteoclast activity and improving the bone phenotype. These results are in line with those obtained with the reference drug alendronate.

hbdPRELP could represent an interesting treatment because this is a peptide corresponding to the N-terminal domain of a physiological protein, which is unlikely to cause deleterious effects as those observed for the other antiresorptive agents. Consistent with this assumption, at least in our experimental conditions, high doses of hbdPRELP given for a prolonged time frame did not cause adverse effects on a variety of organs, nor did it induce any immunological response. At variance with alendronate, hbdPRELP also reduced the size of visceral metastases and, to some extent, tumor growth in orthotopic sites.

The mechanism of action of hbdPRELP in cancer cells is different from that described in osteoclasts[3] because in vitro we did not observe any peptide internalization or changes in functional parameters. However, the tagged peptide distributed in vivo not only in bone but also within the tumor and has been observed at the tumor cell surface in vitro. hbdPRELP is part of a protein with cell-to-matrix anchoring function. This protein binds collagens I and II via its leucin-rich repeats domain, whereas the N-terminal heparin-binding domain of PRELP can bind cell surface heparan sulfate[25] and chondroitin sulfate proteoglycans,[3] thus serving as a linker between these proteoglycans and the extracellular matrix. PRELP has also been found to inhibit the formation of the complement membrane attack complex, limiting pathological complement activation in inflammatory diseases such as rheumatoid arthritis.[26] It is thus possible that in vitro hbdPRELP does not perturb cancer cells directly, whereas in vivo it could exert its antitumoral effect by affecting microenvironmental components and their interaction with the tumorcells, which in turn could impair tumor cell proliferation and induce apoptosis. However, it has to be noted that the effect of the peptide in orthotopic tumors was not persistent, an event that could rely on some kind of “resistance” that could allow the tumor cells to regain in a manner similar to that observed with the use of chemotherapeutics.

The inhibitory effect of hbdPRELP on visceral metastases could contribute to the reduced cachexia and increased survival of mice. These events were not observed in alendronate-treated mice, suggesting that, at least in our experimental conditions, inhibition of bone metastases was unlikely to be involved in the improvement of the two parameters. These results, along with the greater effects on bone metastases observed when the treatment with hbdPRELP was combined with a conventional chemotherapeutic, could have a translational importance for future applications in human pathologies.

In conclusion, we have demonstrated that hbdPRELP is a new agent to experimentally treat bone diseases induced by exacerbated osteoclast activity, with a substantial beneficial effect, especially in adjuvant treatments, also on tumor growth both in orthotopic and metastatic sites. These results could open up an avenue for new treatment opportunities to combat bone diseases.


All authors state that they have no conflicts of interest.


We are grateful to Dr Gabri van der Pluijm (Leiden University Medical Center, Leiden, The Netherlands) for his valuable support in the assessment of the bioluminescence technique; to Dr Anna Rufo for histomorphometric evaluation of the bone phenotype; and to Dr Rita Di Massimo for her contribution in the editing of the manuscript. This work was supported by a Swiss Bridge Award grant to AT and DH; by grants from the “Associazione Italiana per la Ricerca sul Cancro” (AIRC) to AT and NR; by a European Calcified Tissue Society postdoctoral grant to MC; and by a grant from the Sweden Research Council to DH. MC is recipient of the “1st Mariuccia Borrini fellowship” from the “Fondazione Italiana per la Ricerca sul Cancro” (FIRC).

Authors' roles: Study design: NR, DH, and AT. Study conduct: NR and MC. Data collection: NR, MC, VT, and BP. Data analysis: NR and MC. Data interpretation: NR, MC, MM, DH, and AT. Drafting manuscript: NR, DH, and AT. Revising manuscript content: All. Approving final version of manuscript: All. AT takes responsibility for the integrity of the data analysis.