This article has a correction

Please see: Ann Rheum Dis 2004;63:328

Ann Rheum Dis 63:67-70 doi:10.1136/ard.2002.005256
  • Extended report

Metabolic activation stimulates acid secretion and expression of matrix degrading proteases in human osteoblasts

  1. M George1,
  2. B Stein1,
  3. O Müller2,
  4. M Weis-Klemm2,
  5. T Pap3,
  6. W J Parak1,
  7. W K Aicher2
  1. 1Institute for Applied Physics and Centre for Nanoscience, Ludwig-Maximilians, University Munich, Munich, Germany
  2. 2Centre for Orthopaedic Surgery, University of Tübingen Medical School, Tübingen, Germany
  3. 3Department of Internal Medicine, University of Magdeburg, Magdeburg, Germany
  1. Correspondence to:
    Dr W K Aicher
    Research Laboratories, Centre for Orthopaedic Surgery, The University of Tübingen Medical School, Pulvermühlstrasse 5, D 72070 Tübingen, Germany;
  • Accepted 12 June 2003


Background: Both cellular and matrix components of healthy bone are permanently renewed in a balanced homoeostasis. Osteoclastic bone resorption involves the expression of vacuolar-type ATPase proton pumps (vATPase) on the outer cell membrane and the secretion of matrix degrading proteases. Osteoblasts modulate the deposition of bone mineral components and secrete extracellular matrix proteins.

Objectives: To investigate the ability of osteoblasts and osteosarcoma to secrete acid and express matrix degrading proteases upon metabolic activation. To examine also the potential contribution of vATPases to proton secretion expressed on osteoblasts.

Methods: Osteoblasts were isolated from trabecular bone and characterised by reverse transcriptase-polymerase chain reaction and immunohistochemistry. Proton secretion was analysed by a cytosensor microphysiometer.

Results: Osteoblasts not only express matrix degrading proteases upon stimulation with tumour necrosis factor or with phorbol ester but they also secrete protons upon activation. Proton secretion by osteoblasts is associated partially with proton pump ATPases.

Conclusion: These data suggest that, in addition to monocyte derived osteoclasts, cytokine activated mesenchymal osteoblasts and osteosarcoma cells may contribute to the acidic milieu required for bone degradation.

Bone tissue is constantly in a process of remodelling. For bone degradation, osteoclasts express matrix degrading proteases, including different cathepsins1 and matrix metalloproteinases (MMPs).2–4 Further, osteoclasts secrete protons by a vacuolar (H+) ATPase (vATPase).5–7 Mutations of a vATPase subunit cause osteopetrosis8 and deletion of a vATPase subunit results in osteosclerosis.9

Previously the vATPase activity was associated with osteoclasts only. But recently, vATPases have been described in a variety of other cell lineages as well.10–13 Such vATPases act in intracellular pH regulation as well as in proton secretion.13,14 This prompted us to study proton secretion by osteoblasts.

Parathyroid hormone has been shown to induce proton secretion by osteoblasts or osteosarcoma cells.15 We have shown that synovial fibroblasts secrete protons upon metabolic activation,16 and fibroblasts isolated from the synovial-like interface membrane can resorb bone without the help of osteoclasts.10,17 These results suggested that activated mesenchymal cells secrete considerable amounts of protons and thereby contribute to bone degradation under certain conditions. As osteoblasts are closely related to fibroblasts,18 we investigated osteoblasts as a second proton source in bone besides osteoclasts.

Protons have an important role in collagen degradation of demineralised bone areas or in cartilage as an acidic milieu is a prerequisite for optimal enzymatic activity of collagenolytic enzymes such as cathepsins, and also for solubilisation of collagen fibres before enzymatic degradation by collagenases.19 Consequently, enhanced proton secretion contributes to degeneration of articular cartilage in rheumatoid arthritis16 and in other pathological conditions of the musculoskeletal apparatus as well.8,9 Therefore, we investigated mechanisms of proton secretion in human osteoblasts.


Preparation of osteoblasts

Osteoblasts were isolated from trabecular bone and expanded in Dulbecco’s modified Eagle’s medium enriched with 20% fetal calf serum and antibiotics. Osteosarcoma lines SAOS-2 and MG-63 (ATCC) served as osteoblast controls. The cells were characterised by reverse transcriptase-polymerase chain reaction (RT-PCR; table W1 (available at and immunohistochemistry. Secretion of MMPs was evaluated by enzyme linked immunosorbent assay (ELISA).

Cytosensor microphysiometer analysis

Proton secretion was analysed by a cytosensor microphysiometer, and pericellular acidification was measured as described previously.16 In brief, proton release of the cells after stimulation increases the acidification rate (fig W1 ( For induction experiments cells were stimulated with tumour necrosis factor α (TNF; 1–100 ng/ml) or phorbol myristate acetate (PMA; 0.1 ng/ml to 10 μg/ml). To block Na+/H+ transmembrane proton pumps or vacuolar-type H+ ATPases (vATPases), Amiloride (0.5 and 1 mmol/l) or Bafilomycin A1 (1 and 2 µmol/l; Calbiochem, Bad Soden, Germany) were added to the osteoblasts, and proton secretion was recorded.


Functional characterisation of the osteoblasts

To characterise the cells under study, the expression of osteocalcin was detected by immunohistochemistry (fig W2A (, and osteoblast associated genes, including alkaline phosphatase, bone morphogenic protein (BMP)-2, BMP-4, osteocalcin, and osteopontin, were detected by RT-PCR analysis, confirming that the cells under investigation were osteoblast-like cells (fig W2B (

Proton secretion by osteoblasts

Phorbol ester (PMA) activates protein kinase C and parathyroid hormone induced pericellular acidification in SAOS-2 osteosarcoma by a protein kinase C dependent pathway.15 Consequently, we tested osteoblasts and osteosarcoma cells for proton secretion upon activation with PMA. In early passage osteoblasts (passage 4–8), PMA induced a dose dependent acidification, and 35% proton secretion above equilibrium was reached (fig W3A ( After long term culture (10–12 passages) osteoblasts responded with lower sensitivity to PMA induced proton secretion (fig W3A ( Further, time course experiments showed that proton secretion was activated immediately after addition of PMA, reaching the plateau as early as 10 minutes after induction (fig 1A). PMA-induced proton secretion in SAOS-2 osteosarcoma cells confirmed that mesenchymal cells secrete protons upon metabolic activation, and a proton secretion above 40% was induced with PMA (fig W3A (

Figure 1

Activation of proton secretion (rmax/req in per cent) in osteoblasts upon stimulation by PMA or TNF. (A) Osteoblasts were preincubated for about 20 minutes in running medium to reach metabolic quiescence characterised by a low spontaneous acidification (r(t) = req, white zone). After 15 minutes of preincubation the fluctuations of the acidification value drop below 5%. Then cells are activated with PMA (grey zone), and the metabolic responses are recorded as a function of time. Each data point represents the mean value of n individual measurements of xi (n = 5–10) and the error bars represent the normalised standard deviations. Addition of 1 μg/ml PMA induced an immediate increase in acidification, reaching a plateau at about 115% after 20 minutes. Addition of 5 μg/ml PMA induced a higher response reaching more than 120% acidification. Addition of 10 μg/ml PMA induced an immediate acidification response and a plateau was not observed after 40 minutes of induction. (B) Osteoblasts were activated with TNF (grey zone) as described above (see (A)). Addition of 20 ng/ml TNF induced a slow increase in acidification reaching about 115% after 20 minutes. Addition of 50 ng/ml TNF induced a response reaching more than 120% acidification. Addition of 100 ng/ml TNFα at first reduced the acidification rate below the equilibrium level, then an acidification response was recorded reaching about 125% after 20 minutes

TNFα stimulates proton secretion in osteoblasts

We next tested TNFα for its induction of proton secretion by osteoblasts. Early passage osteoblasts responded to TNFα at higher concentrations (10–100 ng/ml) with a proton secretion of 20–35% above the equilibrium level. Low concentrations of TNF (1–2 ng/ml) induced an acidification rate of about 20% above the equilibrium level. The TNF-induced proton secretion was reduced in late passage osteoblasts (fig W3B (( Further, time course experiments showed that the TNF response was delayed to some extent (fig 1B) in comparison with the PMA induced acidification (fig 1A), and the maximal proton secretion was reached 20 minutes after induction (fig 1B).

As different mechanisms may contribute to pericellular acidification, we used Bafilomycin A1, a specific vATPase blocker (1 µmol/l, 2 µmol/l), and Amiloride, a blocker of Na+/H+ exchange ATPases at a higher concentration (500 µmol/l, 1 mmol/l), to reduce the proton secretion. Both compounds reduced proton secretion by 5 and 12% or by 10 and 12%, respectively (fig 2). This indicates that the involvement of proton pump ATPases of either the vacuolar type (vATPase) or Na+/H+ exchange ATPases works across the extracellular membrane of osteoblasts, thus contributing to the acidification reported here.

Figure 2

Blocking proton secretion by osteoblasts upon addition of Bafilomycin A1 or Amiloride. Osteoblast-like cells were incubated in medium containing 1 mM and 2 mM Bafilomycin A 1 or 500 mM and 10000 mM Amiloride. Pericellular acidification was recorded by cytosensor microphysiometer. Each data point represents the mean of two individual measurements and the error bars represent the normalised standard deviations. Both the vATPase proton pump blocker, Bafilomycin A1, and the blocker of Na+/H+ transmembrane exchange pathways, Amiloride, reduced proton secretion by 5–12%.

Expression of proteolytic enzymes in activated osteoblasts

To test whether activated osteoblasts may contribute to degradation of extracellular matrix proteins, the expression of proteolytic enzymes was investigated. In osteoblasts mRNA encoding MMP-1 was enhanced by addition of TNF (p⩽0.17) and PMA (p⩽0.17) about 25-fold (fig W4A ( MMP-3 encoding mRNA was stimulated by TNF (p⩽0.136) and PMA (p⩽0.05) on average fivefold as determined by real time quantitative RT-PCR (fig W4A). Further, TNF augmented MMP-1 (p⩽0.13) and MMP-3 (p⩽0.02) concentrations in osteoblast supernatants. Comparably, PMA stimulated MMP-1 (p⩽0.096) as well as MMP-3 secretion (p⩽0.27) (fig W4B ( These results indicate that cells of the osteoblast lineage may contribute to degradation of the extracellular matrix upon activation by both proton and protease production.


Previously, we showed that fibroblasts may secrete protons upon metabolic activation. To study osteoblast associated proton secretion we activated the cells by addition of PMA, as this compound has previously been associated with proton secretion in bone20 and with pericellular acidification through osteoblasts, T cells, or peritoneal macrophages.11,15,21 Interestingly, an overall trend towards a lower acidification response was noted in late passage osteoblasts in comparison with early passage cells or SAOS-2 in osteosarcoma cells. However, detection of osteoblast marker genes supports our suggestion that besides osteoclasts the osteoblasts may secrete protons upon activation. Addition of TNF was used as a physiological activator of the osteoblasts. TNF is a prominent product of monocytic cells and, in addition, TNF activates osteoclast resorptive activity.22 In our experiments, TNF activated proton secretion in human normal osteoblasts.

Activated osteoblasts express different matrix MMPs.23,24 The activation of proton secretion in the presence of enhanced MMP activity may result in a catabolic situation. Our data support these findings as expression of MMP-1 and MMP-3 were up regulated in osteoblasts. However, owing to variations in gene induction in the individual samples the statistical significance was not high. Nevertheless, induction of MMP-1 and MMP-3 were recorded in all individual experiments.

The resorptive capacity of cells in bone is principally associated with osteoclasts or tumour associated macrophages. Proton secretion by individual osteoblast cells is probably rather small as they lack ruffled membrane areas equipped with proton pumps. Of note, osteoblasts outnumber osteoclasts in bone. The lifespan of osteoclasts and their differentiation capabilities are limited at least in vitro in comparison with osteoblasts. In addition, osteoclasts are rather sensitive to apoptosis inducing signals in comparison with osteoblasts.25 Although osteolysis by tumours was shown to require osteoclasts in vivo,26,27 osteoblasts or osteosarcomas may represent a considerable proton source in bone under specific conditions.28 Consequently, for physiological bone turnover processes such as growth of bone, wound healing, or load adaptation, osteoclasts represent the primary cells for mineral and protein matrix degradation. In chronic inflammatory processes, osteoporosis of the elderly, cancer, or under other pathological conditions the slow resorbing but apoptosis resistant osteoblast may contribute to osteolysis.

The specific contribution of different cellular proton sources for pericellular acidification by osteoblasts remains to be investigated. Enhanced glycolysis may account for pericellular acidification. This may explain in part the observation that early passage osteoblasts show higher proton secretion than the same cells at later passages or SAOS-2 cells. However, as cells are serum starved before cytosensor microphysiometry, glycolysis is probably considerably reduced under these conditions. This indicates that additional proton sources are active on osteoblasts. Such proton sources include Na+/H+ ion exchange ATPases or vATPases.28 As we found mRNA encoding the 116 kDa subunit of the vATPase in osteoblasts (not shown) and a reduced pericellular acidification upon addition of the specific proton pump blocker Bafilomycin A1 and Amiloride to the cells, our data suggest that specific cell membrane associated ATPases participate in proton extrusion and in pericellular acidification by osteoblasts or osteosarcoma cells. Therefore, mesenchymal cells may contribute directly to tissue degradation using different proton sources, including vATPase and H+ ion exchange ATPase activities and matrix degrading proteases. To the best of our knowledge, this is the first report of mechanisms of pericellular acidification by proton pumps on osteoblasts or osteosarcoma cells.


We thank Professor HE Gaub for experimental advice and providing the equipment, Angelika Kardinal and Andrea Mast for excellent technical support, and Professors Giehl, Martini, and Sell for suitable tissue samples. This project was supported by DFG grants Ai 16/10–1, Ai 16/14–1, and Pa 698/3–1, by the Schuler Foundation (University of Tübingen) and, in part, by institutional funding.