Mechanisms of Osteoclast Dysfunction in HumanOsteopetrosis: Abnormal Osteoclastogenesis and Lack of
Osteoclast-Specific Adhesion Structures
ANNA TETI,1 SILVIA MIGLIACCIO,2 ANNA TARANTA,2,3 SILVIA BERNARDINI,2,7
GIULIO DE ROSSI,4 MATTEO LUCIANI,4 METELLO IACOBINI,5 LIDIA DE FELICE,6
RENATA BOLDRINI,4 CESARE BOSMAN,7 ALESSANDRO CORSI,1 and PAOLO BIANCO1
ABSTRACT
Osteoclasts from a patient affected by osteopetrosis were examined in vivo and in vitro. Iliac crest biopsy revealedan osteosclerotic pattern, with prominent numbers of osteoclasts noted for hypernuclearity and incomplete ad-herence to the bone surface. A population comprising tartrate-resistant acid phosphatase (TRAP)-positive, mul-tinucleated and mononuclear cells, and alkaline phosphatase-positive stromal fibroblasts was obtained in vitrofrom bone marrow. Mononuclear TRAP-positive precursors spontaneously fused in culture to form giant osteo-clast-like cells. These cells expressed the osteoclast marker MMP-9 and calcitonin receptor, and lacked the mac-rophage marker, Fc receptor. Expression and distribution of c-src, c-fms, and CD68, and response to steroidhormones relevant to osteoclast differentiation and function were apparently normal, whereas cell retraction inresponse to calcitonin was impaired. TRAP-positive multinucleated cells did not form osteoclast-specific adhesionstructures (clear zone, podosomes, or actin rings). Bone resorption rate was severely reduced in vitro. Focaladhesions and stress fibers were observed en lieu of podosomes and actin rings. Adhesion structures contained lowlevels of immunoreactive vitronectin receptor, most of this integrin being retained in cytoplasmic vesicles. Thesedata provide the first characterization of abnormal differentiation and function of human osteopetrotic osteoclast-like cells. (J Bone Miner Res 1999;14:2107–2117)
INTRODUCTION
OSTEOPETROSIS, also called Albers–Schonberg disease,is a dysplastic bone disorder characterized by a gen-
eral increase of bone mass that obstructs osseus foraminaand impairs normal medullary hematopoiesis. Due to a gen-eralized increase of radiodensity, the severe form of thedisorder has also been called marble bone disease.(1–3)
Even though osteopetrosis is genetically, clinically, and bio-chemically heterogeneous, the primary underlying mecha-nism is a failure in osteoclastic bone resorption.(1,2)
Studies conducted since the first description have dem-onstrated the genetic inheritance of this pathological con-dition.(3) Several investigators have lately emphasized asubdivision of the human phenotype of the disease in twodifferent types: a malignant, infantile autosomal recessiveform(4–7); and a benign autosomal dominant inheritedform(8–10) whose candidate gene has been localized in chro-mosome 1p21.(11) To date, several intermediate forms ofthe disease with a mild course have been described topossess autosomal recessive inheritance.(12) One of suchintermediate forms, described in about 50 cases world-
1Department of Experimental Medicine, University of L’Aquila, L’Aquila, Italy.2Department of Histology and Medical Embryology, University “La Sapienza,” Rome, Italy.3Istituto Dermopatico dell’Immacolata, Rome, Italy.4Divisions of Haematology and Pathology, Bambin Gesu Hospital, Rome, Italy.5Institute of Pediatrics, University “La Sapienza,” Rome, Italy.6Department of Cellular Biotechnologies and Haematology, University “La Sapienza,” Rome, Italy.7Department of Experimental Medicine, University “La Sapienza,” Rome, Italy.
wide,(13–15) is also called marble brain disease because itfeatures skeletal abnormality associated with renal tubuleacidosis and cerebral calcifications. Sly et al.(14) demon-strated that deficiency in carbonic anhydrase II (CAII), anenzyme that catalyzes the reversible hydration of CO2, al-lowing correct osteoclast bone resorption,(16) is the primarydefect in this form of the syndrome. Isolation and sequenc-ing of the complementary DNA for the CAII gene hasallowed identification of the molecular basis of the defect ina Belgian family, that is, a missense mutation in exon 3(Hys-107–→Tyr).(17) This defect is combined with a spliceacceptor site mutation in an American family.(18)
This inherited bone disorder is not confined to humans.Several animal forms (mouse and rat) have also been de-scribed.(19,20) For instance, the op/op osteopetrotic mouselacks osteoclasts and fails to express the colony stimulatingfactor, CSF-1, which turned out to be critical for the normaldevelopment of the osteoclast lineage.(21–23) The defect hasbeen recognized to be a point mutation in the coding regionof the CSF-1 gene which produces a truncated, nonfunc-tional protein.(24) Recently, a number of gene knock-outmice have been demonstrated to develop osteopetrosis.Disruption of the c-src proto-oncogene produces an osteo-petrotic phenotype which displays numerous but dysfunc-tional osteoclasts.(25–27) In contrast, knock-out of the c-fosproto-oncogene,(28) of the PU.1 and of the NF-kB tran-scription factor genes(29,30) results in phenotypes character-ized by the absence of osteoclasts.
Although the primary defect in osteopetrosis is recog-nized to be associated with failure in osteoclast differentia-tion and/or function, it has recently been suggested thatosteoblast abnormalities may contribute to the pathogen-esis of the disease. This aspect is likely to be associated withboth alteration of bone matrix protein synthesis(31,32) orwith the aberrant production of cytokines critical fornormal osteoclast development and function, such as theCSF-1.(33)
Despite substantial advances in the study of animal os-teopetrosis, little is known yet on the cellular and moleculardefects in the human syndrome. This is mostly due to boththe rarity of the disease and genetic heterogeneity. In thisstudy we provide evidence that: cultured osteoclast-likecells from one case of human osteopetrosis fail to resorbbone in vitro and lack the resorption-associated, osteoclast-specific adhesion structures (podosomes and actin rings);and spontaneous fusion of mononuclear precursors intodysfunctional osteoclast-like cells occurs in cultures of mar-row cells.
MATERIALS AND METHODS
Patient
A female patient, aged 9 years, was first seen at the Di-vision of Haematology, Bambin Gesu Hospital, Rome,Italy. She was admitted to the Hospital at 2 months of agefor severe sepsis, pneumonia, and anemia. The infant had aradiographic diagnosis of osteopetrosis based on “bone inbone” (endobone) configuration in spine and limbs, sclero-sis of the base of the skull, and broadening of the long
bones ends. The patient had short stature, hepatospleno-megaly, recurrent episodes of pneumonia and sepsis, diso-donthiasis, and caries. Karyotype was 46XX. Parents and a11-year-old brother were apparently unaffected. Repeatedblood cell counts demonstrated a condition of neutropeniawhich had lasted since infancy.
A bone biopsy of the iliac crest was performed and rou-tinely processed for paraffin embedding. A bone marrowaspirate was obtained for clinical purposes. Part of this ma-terial was used in this study, as detailed below. A bonemarrow aspirate was also obtained from an 8-year-old sub-ject during remission of thrombocytopenic idiopathic pur-pura. This material analyzed for clinical purposes revealednormal parameters and was employed as control. Materialsfrom both patient and control subject were used for thisstudy with the informed consent of parents.
Materials
Cell culture media, serum, and reagents were fromGIBCO (Uxbridge, U.K.). Sterile glass ware was from Fal-con Becton Dickinson (Meylan, France). The anti-avb3 an-tibody, LM609, and the anti-Fc receptor, CD16, antibodywere from Chemicon International Inc. (Tamecula, CA,U.S.A.). The anti-pp60c–src polyclonal antibody, sc-19, andfluorescein isothiocyanate–conjugated secondary antibod-ies were from Santa Cruz Biotechnology Inc. (Heidelberg,Germany). The anti–c-fms polyclonal antibody, 06–175, wasfrom UBI (Lake Placid, NY, U.S.A.). Antibody to MMP-9was from Oncogene (Cambridge, MA, U.S.A.). Antisera toCD68, S100, CD1, and CD78b were from DAKO Corp.(Carpinteria, CA, U.S.A.). The anti-b3 polyclonal antibodywas generously donated by Dr. Guido Tarone, Diparti-mento di Genetica, Biologia e Chimica Medica, Sezione diBiologia, University of Torino, Torino, Italy. 1,25-dihy-droxyvitamin D3 (1,25(OH)2D3) was a gracious gift of Dr.Domenico Criscuolo and Mario Piatti, Hoffman LaRoche(Milano, Italy). Reagents for reverse transcriptase-polymerase chain reaction (RT-PCR) were from Promega(Milan, Italy). gATP was purchased by NENTM Life Sci-ence Products, Inc. (Boston, MA, U.S.A.). HybondTM-Nmembrane was from Amersham Int. plc (Chalfont Buck-inghamshire, U.K.). All other reagents were of the purestgrade from Sigma Chemical Co. (St. Louis, MO, U.S.A.).
Cell cultures
The total bone marrow cell fraction from the control andthe osteopetrotic samples were dispersed in Iscove’s modi-fied minimum essential medium supplemented with 20%fetal calf serum, 100 IU/ml penicillin, and 100 mg/ml strep-tomycin, and cultured in plastic dishes at 37°C, in a humidi-fied atmosphere of 95% air and 5% CO2. After 24 h,nonadherent cells were removed by aspiration and exten-sive washing, the adherent cells were grown to 80% con-fluence, and subcultured by standard trypsin procedure.Aliquots of the cells were grown in the presence of 10−8 M1,25(OH)2D3, and/or 10−7 M estradiol.
TETI ET AL.2108
Phosphatase enzyme activity
Cells were fixed in 3% paraformaldehyde in 0.1 M cac-odylate buffer for 15 minutes, then extensively washed inthe same buffer. Alkaline phosphatase (ALP) and tartrate-resistant acid phosphatase (TRAP) activities were detectedhistochemically using Sigma kits no. 85 and no. 386, respec-tively.
Fluorescence microscopy
Cells were fixed in 3% paraformaldehyde in PBS,washed (three times) in PBS, then, when required, perme-abilized with 0.5% Triton X-100 in PBS for 3 minutes at0°C.
Microfilaments were decorated by incubation with 10 mg/ml rhodamine-conjugated phalloidin. avb3 and b3 integrins,c-src and c-fms proto-oncogenes, and CD68 were detectedby indirect immunofluorescence. Incubation with the pri-mary antibody (diluted 1:100) was performed for 1 h at37°C followed by fluorescein-conjugated anti-mouse oranti-rabbit secondary antibodies (diluted 1:200) for further1 h at 37°C. Cells were then observed by conventional epi-fluorescence in a Zeiss Axioplan microscope (Jena, Ger-many) and photographed with a Kodak Tmax 400 film (Ko-dak, Rochester, NY, U.S.A.).
Bone resorption
Cells were cultured in 24-plate multiwells containing 4 ×4 mm bovine bone slices for 7 days. Cells were then fixed in3% paraformaldehyde in PBS, stained with 1% toluidineblue, and observed by conventional light microscopy. Mul-tinucleated cells per each bone slice were counted, thensections were cleaned free of cells by sonication, stainedwith 1% toluidine blue and observed again by conventionallight microscopy. The resorption pits per each bone slicewere enumerated. Data were then expressed as number ofpits/multinucleated cell (mean ± SEM).
Reverse transcriptase-polymerase chain reaction
RNA was extracted by the guanidinium thiocyanate-phenol-chloroform method. Thirty-two nanograms of RNAwere reverse transcribed and amplified in 50 ml of AMV-Tfl buffer containing 1 ml of 10 mM dNTP, 2 ml of 25mM MgSO4, 5 U AMV reverse transcriptase, 5 U Tfl-polymerase, and 10 pM of each primer. PCR conditionswere 30 cycles, 94°C (30 s), 53°C (30 s), and 68°C (30 s). Thefollowing primer pairs were used to amplify 903 bp of thehuman calcitonin receptor cDNA (GenBank L00587):
forward 58-CCTTTGCTTCTATTGAGCTG-38 (102–121),
reverse: 58-GGTACTACTTCAACCAGGTG-38 (896–1005). PCR amplified products were separated by 1% aga-rose gel electophoresis, blotted to a HybondTM-N mem-brane, and hybridized overnight with gATP-labeledoligonucleotide (58-GGGATGGATGGCTGTGCTGG-38,
474–493) recognizing the internal sequence of the PCR-amplified calcitonin receptor fragment.
RESULTS
Bone biopsy
A clear-cut osteosclerotic pattern was obvious in sectionsof iliac crest biopsies. Bone trabeculae were massive andirregular in shape. Inner cores of unresorbed mineralizedcartilage (endochondral trabeculae) were present in all tra-beculae, which often also featured complex systems of ce-ment (arrest, reversal) lines (Fig. 1a). A very high value ofBV/TV, which estimates the amount of bone tissue (BV)per tissue volume (TV), was obtained by histomorphometry(44.92 ± 2.8; normal value 21.56 ± 4.52).
Prominent numbers of osteoclast-like cells (Fig. 1b; os-teoclast number/bone surface 4 3.85 ± 0.7; normal value0.13 ± 0.10; osteoclast surface/bone surface 4 16.47 ± 3.6;normal value 0.97 ± 0.83), noted for their large size andmultinuclearity (number of nuclei/osteoclast 4 15.00 ± 5.0;Figs. 1c and 1d), were associated with the trabecular pro-files and occasionally impinged upon the adjacent marrowspaces. These osteoclasts were also noted for their pooradherence to the bone surface, as suggested by the smallproportion of their profile that was seen in contact withbone. Upon histochemical/immunohistochemical charac-terization, these cells exhibited strong staining for TRAPand for the osteoclast/macrophage marker, CD68, whereasthey were negative for S100, CD1, and CD78a (data notshown). By transmission electron microscopy, nuclear in-clusion bodies were observed in osteoclasts, whereas cyto-plasmic features were unremarkable (Fig. 2). Of interest,typical ruffled borders and clear zones were never observedat sites of bone resorption.
Isolation of osteoclast-like cells in culture: Generalcharacteristics and response tocalciotropic hormones
Growth of the total population of marrow adherent cellsfrom the osteopetrotic patient resulted in a 80% confluentmonolayer by 2 weeks. Large and irregularly shaped mul-tinuclear giant cells were prominent against a backgroundof mononuclear cells in such primary cultures (Fig. 3). Allgiant cells were intensely reactive for the osteoclast markersTRAP (Fig. 4a) and MMP-9 (Fig. 5a), but negative to themacrophage marker Fc receptor (Fig. 5b). In addition, RT-PCR revealed detectable levels of calcitonin receptormRNA expressed in these cultures (Fig. 5c). Nuclei andorganelles were located in the central area of the giant cells,surrounded by a flat cytoplasm firmly attached to the sub-strate (Figs. 3 and 4a). The paramarginal area did not showmembrane ruffling, lamellipodia or filopodia, nor clear-zone-like organization (Figs. 3 and 4a). Giant cells con-tained 23 ± 8 nuclei (range 7–45), a higher number com-pared with normal osteoclasts.(34) Upon histochemicalcharacterization, mononuclear cells appeared to comprise adual population of fibroblast-like, ALP-positive cells (mar-
OSTEOCLAST DYSFUNCTION IN HUMAN OSTEOPETROSIS 2109
row stromal fibroblasts; Fig. 4b), and ALP-negative andTRAP-positive cells (putative mononuclear osteoclast pre-cursors, Fig. 4a). Cultures from the age-matched controlsubject revealed the presence of TRAP-positive mono-nuclear cells against a background of ALP-positive marrowstromal fibroblasts (Figs. 4c and 4d). Most of the TRAP-positive mononuclear cells had morphological features ofthe monocyte-macrophage lineage. No multinucleated os-teoclast-like cells where observed in this culture, indepen-dent of time and treatment with 1,25(OH)2D3.
Repeated counts performed on the same culture dishes atdifferent time points revealed an increase in giant cell num-bers with time in the bone marrow of the osteopetroticpatient only (one representative experiment: 2 day culture,14/cm2; 7 day culture, 51/cm2, fold increase 3.64), indicatingthat new osteoclast-like cells were being spontaneously gen-erated. Of interest, passaging of the ostepetrotic primarycell population resulted in subcultures with essentially iden-tical cellular composition, morphology, and expression ofhistochemical markers, except a slight reduction in TRAPactivity. TRAP-positive monocyte/macrophage-like cellswere, instead, completely lost in the subcultures of the con-trol subject.
Treatment of the primary cultures with 10−8 M1,25(OH)2D3, which is known to induce osteoclast forma-tion from the bone marrow in a variety of species, includinghuman,(23,34) resulted in a 1.8-fold increase in osteoclast-
like giant cell numbers (one representative experiment:control 38/cm2; 1,25(OH)2D3-treated 67/cm2). Estradiol,alone or in combination with 1,25(OH)2D3, reduced by30% the number of multinucleated cells relative to un-treated cultures (estradiol-treated 27/cm2; 1,25(OH)2D3 +estradiol-treated: 44/cm2). Treatment with 10−7 M salmoncalcitonin failed to produce detectable changes in cell shapein giant osteoclast-like cells. Of note, excess osteoclast-likegiant cell formation could still be induced by 1,25(OH)2D3
in subcultures.
Bone resorption assay and organization ofadhesion structures
To probe the efficiency of osteoclast-like giant cells inresorbing bone in vitro, a pit assay system was used.(35)
Cells firmly adhered to bone matrix and fully spread on thesubstrate, mostly retaining the overall morphology they ex-hibited on plastic (Fig. 6). Cells were cultured on bone for7 days, then pit formation was enumerated. Very little boneresorption was observed, with a rate of 0.17 ± 0.7 pits/cell(mean normal value 1–4 pits/cell, depending on species andtime of culture).
We examined microfilament distribution by decorationof F-actin with fluorescent phalloidin (Figs. 7a–7c). Thetypical punctate distribution of F-actin associated with po-dosomes in normal osteoclasts and their precursors from a
FIG. 1. Iliac crest biopsy. Light microscopic photomicrographs of sections of the iliac crest of the osteopetrotic patient.(a) Massive and irregular shaped bone trabeculae (B) with inner cores of unresorbed cartilage (large arrows), surroundingthe bone marrow tissue (bm) are visible. Small arrows show cement lines. Magnification ×100. (b) Section showing severalosteoclasts (arrows) lining the bone surface. Magnification ×450. (c and d) Irregularly shaped, hypernucleated osteoclasts(arrows) are visible in this field. Magnification ×1000.
TETI ET AL.2110
variety of species, including human,(36–43) was never ob-served. In contrast, stress fibers, which are never found innormal osteoclasts (or even in inflammatory polykarya),
were readily apparent in most of our osteoclast-like cells.The bundles of microfilaments crossed irregularly the cyto-plasm and terminated in areas resembling focal adhesions(Fig. 7c). Staining with fluorescent phalloidin also con-firmed the absence of both lamellipodia and filopodia.
In the normal donor bone marrow culture, the mono-nuclear monocyte/macrophage-like cells had prominentthin microfilaments distributed in the paramarginal area,especially at the level of membrane lamellipodia and filo-podia (Fig. 7d). Approximately one third of these cells alsoshowed typical podosomes distributed in the ventral mem-brane (Figs. 7e and 7f). Stress fibers and focal adhesionswere observed in the stromal fibroblasts, but not in themonocytic cells.
Distribution of the vitronectin receptor
We next examined the distribution of the vitronectin re-ceptor, the avb3 integrin, a relevant marker of the osteo-clast phenotype as well as a critical molecule for bone re-sorption.(44–49) Immunoreactivity for the LM609 antibody,which recognizes the intact avb3 integrin, was observed inthe multinucleated cells of the osteopetrotic patient (Fig.8a). However, an unusual granular pattern of cytoplasmiclabeling was obtained. When an antibody recognizing theb3 subunit of the vitronectin receptor was used in immuno-fluorescence studies, most of the signal was again associated
FIG. 2. Transmission electron microscopy photomicro-graphs of osteoclasts. Part of the iliac crest biopsy of theosteopetrotic patient was processed for transmission elec-tron microscopic analysis. (a), (b), and (c) show osteoclasts(OC) in contact with the bone surface (B). Arrows indicatenuclear inclusion bodies. Of note, no ruffled borders norclear zones are apparent. An irregular profile of the osteo-clast–bone interface is obvious, consistent with an unusualpattern of adhesion and resorption. (a and c) Magnification×2600. (b) Magnification ×3250.
FIG. 3. Phase contrast micrographs of the cultured cellpopulation of the osteopetrotic patient. (a) A multinucle-ated cell (arrow) is visible together with numerous mono-nuclear, irregularly shaped mononuclear cells. Magnifica-tion ×600. (b) Hypernucleated (arrows) osteoclast-like cell.Magnification ×800.
OSTEOCLAST DYSFUNCTION IN HUMAN OSTEOPETROSIS 2111
with cytoplasmic granules of variable size, whereas labelingof the focal adhesions was negligible (Figs. 8b–8d). No po-dosomal pattern, commonly detected in functional humanosteoclasts,(40) was observed. The b3 integrin was insteadregularly distributed in the podosomes of the normal donormonocyte/macrophage-like cells (data not shown).
c-src, c-fms, and CD68 expression and distribution
We next examined expression and distribution of the c-src proto-oncogene, whose disruption is known to causeosteopetrosis in a mouse model.(25) Figure 9a shows thatthe multinucleated cells cross-reacted with a specific poly-clonal antibody recognizing the c-src gene product,pp60c–src. This protein was mostly distributed in small andnumerous cytoplasmic vacuoles which were located in theperinuclear area and extended to the cell periphery. Nucleiwere negative.
Since CSF-1 is critical for osteoclastogenesis(21–24) andmature osteoclast function,(50–52) we evaluated the expres-sion of the c-fms tyrosine kinase receptor, which recognizesCSF-1 and is expressed by osteoclasts.(53) Figure 9b dem-onstrates that the c-fms receptor can be immunodetected inthe multinucleated cells, with a diffuse distribution which islikely to be membrane associated. In addition, localizationin cytoplasmic vacuoles was also observed.
Finally, we tested the cultured cells for expression of theCD68 antigen, which has been shown both in macrophagesand osteoclasts.(54) As for the osteoclasts in the bone bi-opsy, cultured multinucleated cells (Fig. 9c), as well asTRAP-positive mononuclear cells (data not shown), ex-
pressed CD68. c-fms, c-src, and CD68 were normally dis-tributed in the cells from the normal donor (data notshown).
DISCUSSION
To our knowledge, this study provides the first in vitrocharacterization of osteoclast-like cell dysfunction in hu-man osteopetrosis. Using cultures of total adherent bonemarrow fraction, as per commonly used methods,(34,55) weobtained a unique heterogeneous cell population enrichedin, and spontaneously generating, dysfunctional osteoclast-like giant cells. Strong expression of TRAP activity, ofMMP-9, of CD68, and of the vitronectin receptor, lack ofexpression of the Fc receptor and detectable mRNA levelsof calcitonin receptor concurred in defining an osteoclast-like phenotype of these cells.(56) Likewise, the expression ofTRAP activity and CD68, and the unusually large size andnumbers of nuclei also established a remarkable phenotypicsimilarity between the cells we isolated in culture and theosteoclasts observed in the bone biopsy sections.
In the bone marrow cultures of the osteopetrotic patient,osteoclast-like giant cells were readily apparent from thestart and further increased with time in the absence of anydifferentiating factor, indicating that they were constitu-tively generated. Generation of osteoclast-like cells was re-tained in such cultures even after repeated passages. Spon-taneous generation of osteoclast-like cells does not usuallyoccur in cultures of total marrow adherent cells. Osteoclastgeneration in vitro can be induced by added exogenous
FIG. 4. Histochemically staining for TRAP and ALP enzymes. Conventional light micrographs showing (a) severalTRAP-positive multinucleated (arrowheads) and mononuclear cells (arrows), and (b) ALP-positive mononuclear cells(arrows) in the bone marrow culture of the osteopetrotic patient. (c) TRAP-positive mononuclear cells only (arrows), and(d) ALP-positive mononuclear cells (arrows) were observed in the bone marrow culture of an age-matched normal donor.(a and c) Magnification ×350. (b and d) Magnification ×1000.
TETI ET AL.2112
factors, such as 1,25(OH)2D3 or parathyroid hormone, inbone marrow cultures from small mammals.(23) In contrast,in vitro generation of osteoclasts from human marrow ismuch less consistently observed under similar experimentalconditions.(56) Indeed, our cultures of total marrow from anage-matched human control subject were consistently freefrom osteoclast-like giant cells, even after treatment with1,25(OH)2D3, which only resulted in the appearance of in-creased numbers of mononuclear TRAP-positive cells. Wetherefore believe that the observed pattern of spontaneousgeneration of osteoclast-like giant cells in bone marrow cul-tures from our osteopetrotic patient reflects an abnormalityof the diseased bone marrow system being reproduced inculture, and mirroring the enhanced generation of osteo-clasts in vivo reflected in the high counts of osteoclastsobtained in the bioptic material. Indeed, excess osteoclast
formation is a common feature in animal models of osteo-petrosis, and has been reported previously in other in vivoobservation in humans.(1–3) It has been suggested that acompensatory enhanced proliferation of osteoclast precur-sors may occur in these systems in response to the impairedbone resorption.(25) It is interesting to note here that theretention of a highly enhanced osteoclast generation in cul-ture over several passages might as well reflect an intrinsiccharacteristic of the osteopetrotic bone marrow system.
Our cultured osteoclast-like cells showed apparent nor-mal response to 1,25(OH)2D3, which increased the numberof giant cells generated in vitro, as expected for normal andosteopetrotic osteoclasts.(34,55,57) Furthermore, decrease ofosteoclast-like cell formation was observed in estradiol-treated cultures. In contrast, the response to calcitonin, asassessed by cell retraction,(58,59) was impaired. This sug-gests either that expression levels of calcitonin receptors isnot sufficient to elicit a response, or that altered behaviorprevents cell retraction in response to the hormone. Thisissue requires further investigation. Most important, ourosteoclast-like cells were severely dysfunctional with re-spect to bone resorption, as assessed by the in vitro pitassay.
Strikingly, the multinucleated osteoclast-like cells andthe mononuclear TRAP-positive cells of the osteopetroticpatient revealed an altered cytoskeletal organization andadhesion pattern as compared to normal osteoclasts andtheir putative precursors of virtually all species, includinghumans.(36–43) A substantial body of evidence demonstratesthat the development of correct adhesion is mandatory forbone resorption,(36,37) and that the peripheral area of func-tional osteoclasts, which forms the so-called clear zone ei-ther in vivo and in vitro, is endowed with podosomes andactin rings,(36–43) which tightly and dynamically seal themembrane to the substrate. Our cultured osteoclast-likecells showed no evidence of podosomes or actin rings, norof lamellipodia or filopodia arrays, strongly suggesting thata key defect of these hypofunctional osteoclasts could re-side in the adhesion properties, as well as in their motility.In fact, stress fiber organization is a typical feature of staticadhesion,(60) and is found in stromal cells and fibroblasts,which are nonmotile, matrix-forming cells, but not in the
FIG. 5. Expression MMP-9, Fc receptor, and calcitoninreceptor in the bone marrow culture of the osteopetroticpatient. (a) Positive immunofluorescence reaction forMMP-9 in a multinucleated cell. (b) Negative immunoflu-orescence reaction for Fc receptor in a multinucleated cell.The inset shows Fc receptor expression in HL-60 cells usedas positive control for this reaction. Magnification ×1200.(c) RT-PCR using the human calcitonin receptor primersdescribed in materials and methods: lane 1, COS-7 cellstransiently transfected with the human calcitonin receptorcDNA (positive control); lane 2, bone marrow culture ofthe patient.
FIG. 6. Conventional light micrographs of a bone slice. (a)A multinucleated cell from the osteopetrotic patient is vis-ible (arrow). (b) The same field of (a) (arrows) after themultinucleated cell was removed by sonication. Magnifica-tion ×500.
OSTEOCLAST DYSFUNCTION IN HUMAN OSTEOPETROSIS 2113
monocyte/macrophage family (including the osteoclast pre-cursors) and in normal osteoclasts,(34,38,39) which are highlymotile populations. Although the cultures from our controlsubject did not form multinucleated osteoclast-like cells, atvariance with the osteopetrotic patient, the putative osteo-
clast precursors showed normal actin distribution and po-dosome assembly. Due to the lack of osteoclasts in thecontrol subject, a final conclusion on a potential link be-tween actin organization and the osteopetrotic defect can-not be drawn. However, based on our observations and on
FIG. 7. Fluorescent phalloidin decoration of microfilaments. (a–c) Culture of bone marrow cells from the osteopetroticpatient showing osteoclasts with stress fibers (arrows). In (c) cells were partially lifted by gentle pipetting to show patchesof microfilaments reminiscent of focal adhesions (arrows). Magnification ×1000. (d–g) Cultures of bone marrow cells froman age-matched normal donor. (d) Mononuclear, monocyte/macrophage-like cells (arrows) with thin microfilamentbundles are apparent against a background of stress fiber-presenting stromal fibroblasts. (e–g) Some of the monocytic cells(arrows) show typical podosomes noted for their punctate pattern. Magnification ×1200.
FIG. 8. Immunofluorescence staining for vitronectin receptor in the bone marrow culture of the osteopetrotic patient. (a)Total avb3 receptor. (b–d) b3 subunit. Note the distribution of the receptor in large perinuclear vesicles (arrows). (a–c)Magnification ×800; (d) magnification ×1200.
TETI ET AL.2114
reports in the literature,(36–43) it seems reasonable to hy-pothesize that the cytoskeletal/adhesion array observed inthe patient is defective.
Podosomes and actin rings have been demonstrated to becritical to the resorbing osteoclast function.(36–43) In par-ticular, podosomes are highly dynamic adhesion structureswhich form and disassemble within minutes,(61) as opposedto focal adhesions which are more stable matrix-contactareas, disrupted within hours.(60) The multinucleated cellsderived from our patient form focal adhesion structuresonly. In other cell types it has been demonstrated that theseare sites where actin bundles (stress fibers) end in contactwith the adhesion talin/vinculin/a-actinin complex, which inturn binds the integrin and, as a consequence of this, theextracellular matrix.(60) In the giant cells of our patient, theavb3 integrin was only marginally found at the adhesionsites, most of it being retained in cytoplasmic vesicles. Thisobservation indicates a potential defect in the membranedelivery pathway of this receptor. At this time it is not clear,however, whether this defect is related to a specific impair-ment of the vitronectin receptor transfer to membrane, orto a more general failure of the secretory activity of the cell.Alteration of osteoclast interaction with the substrate has
been demonstrated to lead to inhibition of bone resorp-tion.(46–49) Therefore, the defect observed in our culturemay critically contribute to the impairment of resorptionactivity shown both in the bone biopsy and in the in vitro pitassay test.
The cultured TRAP-positive multinucleated cellsexamined in this study showed apparently normal pp60c-src
and c-fms receptor expression and distribution. Theseare two of the most important genes influencing os-teoclast function (c-src and c-fms) and differentiation(c-fms).(21,22,25–27,50,51,62) c-src and the c-fms ligand, CSF-1,are implicated in the pathogenesis of osteopetrosis in ani-mal models.(24,25) This apparent normal expression and dis-tribution, however, do not rule out the occurrence of de-fects of these genes in our patient, due to the fact thatimmunofluorescence analysis may also detect mutant mol-ecules, which may be normally expressed and distributed,but nonfunctioning. A more detailed molecular analysis isnecessary to address this point.
In conclusion, this study provides a novel insight into thecellular abnormality in a case of human osteopetrosis. Theresults indicate that changes in both function and osteoclas-togenesis process may occur in this patient. While boneresorption is reduced, osteoclast-like cell generation seemsto be increased and dysregulated, leading to excess of non-functioning osteoclasts both in vivo and in vitro. The studyalso documents for the first time that aberration of cyto-skeletal/adhesion array may represent a key underlyingmechanism of osteoclast defect in human osteopetrosis. Fu-ture work on this cell population will allow to clarify thealtered molecular mechanism and to attempt the genotypiccharacterization of this human bone dysplastic pathology.
ACKNOWLEDGMENTS
We thank Ms. Susanne Voit and Mr Luigi Pellegrino fortheir excellent technical help. The Telethon grant E.456 isgratefully acknowledged. This work was also supported bya grant from the “Ministero dell”Universita e della RicercaScientifica (cofinanziamento 1997). Ms. Susanne Voit andDr Anna Taranta are recipients of fellowships from theTelethon grant E.456.
REFERENCES
1. Mundy GR 1987 Osteopetrosis. In: Stein JH (ed.) InternalMedicine. Little, Brown, Boston, MA, U.S.A., pp. 2115–2117.
2. Teitelbaum SL 1993 Molecular defects of bone development.In: Favus MJ (ed.) Primer on the Metabolic Bone Disease andDisorders of Mineral Metabolism. Lippincott Raven Publisher,Philadelphia, PA, U.S.A., pp. 10–15.
3. Whyte MP 1996 Sclerosis bone dysplasias. In: Favus MJ (ed.)Primer on the Metabolic Bone Disease and Disorders of Min-eral Metabolism. Lippincott Raven Publisher, Philadelphia,PA, U.S.A., pp. 363–379.
4. Loria-Cortes R, Quesada-Calvo E, Cordero-Chaverri E 1977Osteopetrosis in children: A report of 26 cases. J Pediatr 91:43–47.
5. Kaplan FS, August CS, Fallon MD, Dalinka M, Axel L, Had-
FIG. 9. Immunofluorescence staining for (a) c-src (b) c-fms, and (c) CD68 in the bone marrow culture of the os-teopetrotic patient. Magnification ×1000.
OSTEOCLAST DYSFUNCTION IN HUMAN OSTEOPETROSIS 2115
dad JG 1988 Successful treatment of infantile malignant osteo-petrosis by bone marrow transplantation: A case report. JBone Joint Surg 70:617–623.
6. Beard CJ, Key L, Newburger PE, Ezekowitz RA, Arceri R,Miller B, Proto P, Ryan T, Anast C, Simons ER 1986 Neutro-phil defect associated with malignant infantile osteopetrosis. JLab Clin Med 108:498–505.
8. Johnston CC Jr, Lavy N, Lord T, Vellios F, Merrit AD, DeissWP 1968 Osteopetrosis: A clinical, genetic, metabolic and mor-phologic study of the dominant inherited benign form. Medi-cine 47:149–167.
10. Grodum E, Gram J, Brixen K, Bollerslev J 1995 Autosomaldominant osteopetrosis: Bone mineral measurements of theentire skeleton of adults in two different subtypes. Bone16:431–434.
11. Van Hul W, Bollerslev J, Gram J, Van Hul E, Wuyts W, Be-richou O, Vanhoeanacker F, Willems PJ 1997 Localization of agene for autosomal dominant osteopetrosis (Albers–Schonberg disease) to chromosome 1p21. Am J Human Genet61:363–369.
12. Kahler SG, Burns JA, Aylsworth AS 1984 A. mild autosomalrecessive form of osteopetrosis. Am J Med Genet 17:451–464.
13. Whyte MP 1993 Carbonic anhydrase II deficiency. Clin Orthop294:52–63.
14. Sly WS, Hewett-Emmet D, Whyte MP, Yu Y-SL, Tashian RE1983 Carbonic anhydrase II deficiency identified as the pri-mary defect in the autosomal recessive syndrome of osteope-trosis with renal tubular acidosis and cerebral calcification.Proc Natl Acad Sci USA 80:2752–2756.
15. Sly WS, Hu PY 1995 The carbonic anhydrase deficiency syn-drome: Osteopetrosis with renal tubular acidosis and cerebralcalcification. In: Scriver CR, Beaudet AL, Sly WS, Valle D(eds.) The Metabolic and Molecular Bases of Inherited Dis-ease. McGraw-Hill, New York, NY, U.S.A., pp. 4113–4124.
16. Laitala T, Vaananen HK 1996 Inhibition of bone resorption invitro by antisense RNA and DNA molecules targeted againstcarbonic anhydrase II or two subunits of vacuolar H+ ATPase.J Clin Invest 93:2311–2318.
17. Venta PJ, Welty RJ, Johnson TM, Sly WS, Tashian RE 1991Carbonic anhydrase deficiency syndrome in a Belgian family iscaused by a point mutation at an invariant histidine residue(107 His→Tyr): Complete structure of the normal human CAII gene. Am J Hum Genet 49:1082–1090.
18. Roth DE, Venta PJ, Tashian RE, Sly WS 1992 Molecular basisof human carbonic anhydrase II deficiency. Proc Natl Acad SciUSA 89:804–1808.
19. Walker DG 1975 Control of bone resorption by hematopoietictissue: The induction and reversal of congenital osteopetrosisin mice through the use of bone marrow mononuclear phago-cytes. J Exp Med 156:1604–1614.
20. Marks SC Jr 1989 Osteoclast biology: lessons from mammalianmutations. Am J Med Genet 34:43–54.
21. Wiktor-Jedrzeiezak W, Bartocci A, Ferrante AW, Ahmed-Ansari A Jr, Sell KW, Pollard JW, Stanley ER 1990 Totalabsence of colony-stimulating factor 1 in the macrophage-deficient osteopetrotic (op/op) mouse. Proc Natl Acad SciUSA 87:4828–4832.
22. Felix R, Cecchini MG, Hofstetter W, Elfors PR, Stutzer A,Fleisch H 1990 Impairment of macrophage colony-stimulatingfactor production and lack of resident bone marrow macro-phages in the osteopetrotic op/op mouse. J Bone Miner Res5:781–789.
23. Suda T, Nakamura I, Jimi E, Takahashi N 1997 Regulation ofosteoclast function. J Bone Miner Res 12:869–879.
24. Yoshida H, Hayashi S-I, Kunisada T, Ogawa M, Nishikawa S,Okamura M, Sudo T, Shultz LD, Nishikawa S-I 1990 The mu-
rine mutation osteopetrosis is in the coding region of the mac-rophage colony-stimulating factor gene. Nature 345:442–444.
25. Soriano P, Montgomery C, Geske R, Bradley A 1991 Targeteddisruption of the c-src proto-oncogene leads to osteopetrosis inmice. Cell 64:693–702.
26. Boyce BF, Yoneda T, Lowe C, Soriano P, Mundy GR 1992Requirement of pp. 60c–src expression for osteoclasts to formruffled borders and resorb bone in mice. J Clin Invest 90:1622–1627.
27. Lowe C, Yoneda T, Boyce BF, Chen H, Mundy GR, Soriano P1993 Osteopetrosis in Src-deficient mice is due to an autono-mous defect of osteoclasts. Proc Natl Acad Sci USA 90:4485–4489.
28. Grigoriadis AE, Wang ZQ, Cecchini MG, Hofstetter W, FelixR, Fleisch HA, Wagner EF 1994 c-fos: A key regulator ofosteoclast-macrophage lineage determination and bone re-modeling. Science 266:443–448.
29. Tondravi MM, McKercher SR, Anderson K, Erdmann JM,Quiroz M, Maki R, Teitelbaum SL 1997 Osteopetrosis in micelacking haemopoietic transcription factor PU.1. Nature 386:81–84.
30. Iotsova V, Caamano J, Loy J, Yang Y, Lewin A, Bravo R 1997Osteopetrosis in mice lacking NF-kB1 and NF-kB2. Nat Med3:1285–1289.
31. Shalboud V, Jackson ME, Lian JB, Stein GS, Marks SC Jr 1991Gene expression during skeletal development in three osteo-petrotic rat mutations: Evidence for osteoblast abnormalities. JBiol Chem 266:9847–9856.
32. Chackalaparampil I, Peri A, Nemir M, Mckee MD, Lin P-H,Mukherjee BB, Mukherjee AB 1996 Cells in vivo and in vivofrom osteopetrotic mice homozygous for c-src disruption showsuppression of synthesis of osteopontin, a multifunctional ex-tracellular matrix protein. Oncogene 12:1457–1467.
33. Shalboud V, Jackson ME, Paradise C, Stein GS, Lian JB,Marks SC Jr 1902 Heterogeneity of colony stimulating factor-1gene expression in the skeleton of four osteopetrotic mutationsin rats and mice. J Cell Physiol 166:340–350.
34. MacDonald BR, Takahashi N, McManus LM, Holahan J,Mundy GR, Roodman GD 1987 Formation of multinucleatedcells that respond to osteotropic hormones in long-term humanbone marrow cultures. Endocrinology 120:2326–2333.
35. Caselli GF, Mantovanini M, Gandolfi CA, Allegretti M,Fiorentino S, Pellegrini L, Melillo G, Bertini R, Sabbatini W,Anacardio R, Clavenna G, Sciortino G, Teti A 1997 Tartro-nates: A new generation of drugs affecting bone metabolism. JBone Miner Res 12:972–981.
36. Teti A, Marchisio PC, Zambonin Zallone A 1991 Clear zone inosteoclast function: Role of podosomes in regulation of bone-resorbing activity. Am J Physiol 261:C1–C7.
37. Vaananen HK, Horton MA 1995 The osteoclast clear zone is aspecialized cell-extracellular matrix adhesion structure. J CellSci 108:2729–2732.
38. Marchisio PC, Naldini L, Cirillo D, Primavera MV, Teti A,Zambonin Zallone A 1984 Cell–substratum interaction of cul-tured avian osteoclasts is mediated by specific adhesion struc-tures. J Cell Biol 99:1696–1705.
39. Marchisio PC, Cirillo D, Teti A, Zambonin Zallone A, TaroneG 1987 Rous sarcoma virus transformed fibroblasts and cells ofmonocytic origin display a peculiar dot-like organization ofcytoskeletal proteins involved in microfilament–membrane in-teractions. Exp Cell Res 169:202–214.
40. Zambonin Zallone A, Teti A, Grano M, Rubinacci A, Ab-badini M, Gaboli M, Marchisio PC 1989 Immunocytochemicaldistribution of extracellular matrix receptors in human osteo-clasts: A b3 integrin is colocalized with vinculin and talin in thepodosomes of osteoclastoma giant cells. Exp Cell Res 182:645–652.
41. Kanehisa J, Yamanaka T, Doi S, Heersche JNM 1990 A. bandof F-actin containing podosomes is involved in bone resorptionby osteoclasts. J Bone Miner Res 5:287–293.
42. Lakkakorpi P, Tukkanen J, Huntunen T, Jarvelin K, VaananenHK 1989 Organization of osteoclast microfilaments during the
TETI ET AL.2116
attachment to bone surface in vitro. J Bone Miner Res 4:817–825.
43. Lakkakorpi P, Vaananen HK 1991 Kinetics of the osteoclastcytoskeleton during the resorption cycle in vitro. J Bone MinerRes 6:817–826.
44. Davies J, Warwick J, Totty N, Philp R, Helfrich M, HortonMA 1989 The osteoclast functional antigen implicated in theregulation of bone resorption is biochemically related to thevitronectin receptor. J Cell Biol 109:1817–1826.
45. Shinar DM, Schmidt A, Halperin D, Rodan GA, Weinreb M1993 Expression of av and b3 integrin subunits in rat osteo-clasts in situ. J Bone Miner Res 4:403–414.
46. Horton MA, Taylor ML, Arnett TR, Helfrich MH 1991 Arg-Gly.Asp (RGD) peptides and the anti-vitronectin receptor an-tibody 23C6 inhibit dentine resorption and cell spreading byosteoclasts. Exp Cell Res 195:368–375.
47. Fisher JE, Caulfield MP, Sato M, Quartuccio HA, Gould RJ,Garsky VM, Rodan GA, Rosenblatt M 1993 Inhibition of os-teoclastic bone resorption in vivo by echistatin, an ‘arginyl-glycyl-aspartyl’ (RGD)-containing protein. Endocrinology132:1411–1413.
48. Rodan SB, Rodan GA 1997 Integrin function in osteoclasts. JEndocrinol 154:S47–S56.
49. Horton MA 1997 Molecules in focus The avb3 integrin ‘vitro-nectin receptor’. Int J Biochem Cell Biol 29:721–725.
50. Insogna K, Tanak S, Neff L, Horne W, Leby J, Baron R 1997Role of c-src in cellular events associated with colony-stimulating factor 1-induced spreading in osteoclasts. Mol Re-prod Dev 46:104–108.
51. Fuller K, Owens JM, Jagger CJ, Wilson A, Moss R, ChambersTJ 1993 Macrophage colony-stimulating factor stimulates sur-vival and chemotactic behavior in isolated osteoclasts. J ExpMed 178:1733–1744.
52. Teti A, Taranta A, Migliaccio S, DeGiorgi A, Santandrea E,Villanova I, Faraggiana T, Chellaiah M, Hruska KA 1998Colony stimulating factor-1-induced osteoclast spreading de-pends on substrate and requires the vitronectin receptor andthe c-src proto-oncogene. J Bone Miner Res 13:50–58.
53. Hofstetter W, Wetterwald A, Cecchini MG, Felix R, Fleisch H,Mueller C 1992 Detection of transcript for the receptor formacrophage colony stimulating factor, c-fms, in murine osteo-clasts. Proc Natl Acad Sci USA 89:9637–9641.
54. Athanasou NA, Alvarez JI, Ross FP, Quinn JM, TeitelbaumSL 1992 Species differences in the immunophenotype of osteo-clasts and mononuclear phagocytes. Calcif Tissue Int 50:427–432.
55. Thavarajah M, Evans DB, Kanis JA 1991 1,25(OH)2D3 in-duces differentiation of osteoclast-like cells from human bonemarrow cultures. Biochem Biophys Res Commun 176:1189–1195.
56. Roodman GD 1996 Advances in bone biology: The osteoclast.Endocr Rev 17:308–332.
57. Cournot C, Petrovic M, Trubert C-L, Cormier C, Girault D,Fischer A, Garabedian M 1993 Cultured circulating mono-nuclear cells from osteopetrotic infants express the osteoclast-assopciated vitronectin receptor and form multinucleated cellsin response to 1,25-dihydroxy vitamin D3. J Bone Miner Res8:61–70.
58. Chambers TJ, McSheehy PMJ, Thompson BM, Fuller K 1985The effect of calcium-regulating hormones and prostaglandinson bone resorption by osteoclast disaggregated from rabbitlong bones. Endocrinology 116:234–239.
59. Gravel MR, Zheng ZG, Sims SM, Dixon SJ 1994 Platelet-acti-vating factor induces pseudopod formation in calcitonin-treated rabbit osteoclasts. J Bone Miner Res 9:1769–1776.
60. Gilmore AP, Burridge K 1996 Molecular mechanisms for focaladhesion assembly through regulation of protein–protein in-teractions. Structure 4:647–651.
61. Stickel SK, Wang YL 1987 Alpha actinin-containing aggre-gates in transformed cells are highly dynamic structures. J CellBiol 104:1521–1526.
62. Tanaka S, Amling M, Neff L, Peyman A, Uhlmann E, Levy JB,Baron R 1996 c-cbl is downstream of c-src in a signalling path-way necessary for bone resorption. Nature 383:528–531.
Address reprint requests to:Anna Teti, Ph.D.
Department of Experimental MedicineVia Vetoio – Coppito 2
67100 L’Aquila, Italy
Received in original form May 26, 1998; in revised form February5, 1999; accepted June 1, 1999.
OSTEOCLAST DYSFUNCTION IN HUMAN OSTEOPETROSIS 2117
