Alterations in the phenotype and function of immune cells in ovariectomy-induced osteopenic mice

Abstract

BACKGROUND: Within the last few years, much evidence has been presented on the involvement of the immune system in certain types of bone loss, such as activated T cells in rheumatoid arthritis and in periodontitis. Estrogen deficiency induces bone loss; however, how this deficiency affects the immune system has not been sufficiently studied. METHODS: To evaluate the effects of estrogen withdrawal on the status and functionality of the immune system, mice were ovariectomized or sham-operated, and 5 weeks after surgery, when osteopenia had developed, several parameters were analysed in spleen and in bone marrow. We analysed bone turnover, cell phenotype by flow cytometry, cell function by cell proliferation assays, and the expression of several genes related to the process. RESULTS: Five weeks after ovariectomy, augmented osteoclastogenesis persisted in the bone marrow. In addition, the ovariectomized mice had more B-cells and CD3+ T-cells expressing the receptor activator of NF-κB ligand (CD3+/RANKL+). The ovariectomized mice had lower serum alkaline phosphatase activity, a normal amount of T cells, lower percentages of CD11b⁺ and CD51⁺ cells in the bone marrow, and a lower serum interferon-γ level compared with sham-operated controls. CONCLUSIONS: The data suggest that, 5 weeks after ovariectomy, bone turnover remains imbalanced, with increased osteoclastogenesis and a decreased rate of bone formation. Moreover, there is an increase in B-cell formation, with normal and decreased percentages of T cells and myelomonocytic cells (CD11b⁺), respectively, in the bone marrow. Decreased serum interferon-γ levels could be involved in the increased osteoclastogenesis found in the present work.

Introduction

Recent research has uncovered molecules crucial to bone metabolism, and in particular to the generation, development and activation of osteoclasts, the cells that resorb bone. These molecules include the receptor activator of the NF-κB ligand (RANKL, also known as TRANCE, ODF and OPGL), its receptor RANK, and osteoprotegerin, all of which are key molecules for osteoclastogenesis. This was demonstrated through the use of mice deficient in these molecules (Theill et al., 2002). RANKL and RANK were first described in immune system cells, where they are expressed in activated T cells, and in dendritic cells, respectively, and are implicated in cell survival and immunomodulation (Anderson et al., 1997). In bone, the binding of RANKL on stromal or osteoblastic cells to RANK on pre-osteoclasts, together with macrophage colony stimulating factor (M-CSF), is necessary and sufficient for the generation, differentiation and activation of osteoclasts (Lacey et al., 1998), although other molecules in the bone marrow environment, such as interleukin (IL)-1, IL-6 and tumour necrosis factor (TNF)-α, intervene in the process (Jilka, 1998). These discoveries have led to the creation of a new field of research, osteoimmunology (Arron and Choi, 2000).

Much evidence has linked the immune system to bone loss (Theill et al., 2002). Some of this evidence comes from the study of animal models of rheumatoid arthritis and periodontitis. It has been demonstrated that activated T cells express RANKL, which, in turn, is able to stimulate osteoclastogenesis and bone loss (Kong et al., 1999; Teng et al., 2000). Moreover, estrogen regulates T-cell functions through an estrogen receptor-mediated pathway, and T-cell subset alterations were found in postmenopausal women with osteoporosis (Olsen and Kovacs, 1996). Furthermore, it has been shown that ovariectomy induces B-cell lymphopoiesis, and that these cells can be a source of osteoclasts (Masuzawa et al., 1994; Sato et al., 2001). Therefore, immune system cells may induce bone loss not only during inflammation but also under conditions of estrogen deficiency.

Estrogen deficiency induces bone loss and osteoporosis (Jilka, 1998; Riggs et al., 1998), and since both bone and immune cells have active receptors for estrogens (Bellido et al., 1993; Komi and Lassila, 2000; Igarashi et al., 2001) it is conceivable that estrogen deficiency induces changes not only in bone but also in the immune system, and that these changes could be associated with increased osteoclastogenesis and bone loss. The status of the immune system after estrogen deficiency is established; however, it needs further study.

In this work, ovariectomy was carried out in the mouse as a model of estrogen deficiency. Ovariectomy induces a rapid augmentation in bone turnover, causing an increase in both bone resorption and bone formation (Jilka et al., 1998; Tanizawa et al., 2000). In the mouse, indices of bone remodelling peak at the first or second week after ovariectomy and then fall progressively. Five weeks after ovariectomy, bone turnover has been normalized almost completely, and the mouse has developed osteopenia due to the large amount of trabecular bone lost (Jilka et al., 1998; Tanizawa et al., 2000; Miyazaki et al., 2004).

The purpose of the present study was to analyse several aspects of the immune system in the spleen and bone marrow in mice that had developed osteopenia as a consequence of estrogen deficiency induced by ovariectomy. We studied the phenotype and function of both cellular types, several parameters in serum, such as calcium, alkaline phosphatase (ALP) activity, TNF-α, IL-6 and interferon-γ levels, and the expression of several genes 5 weeks after ovariectomy, when bone turnover normalized after an initial imbalance provoked by the abrupt interruption in the supply of estrogens.

Materials and methods

Media and reagents

Recombinant mouse soluble receptor activator of NF-κB ligand (sRANKL) and recombinant murine M-CSF were obtained from PeproTech (London, UK). Monoclonal CD3 (clone 145–2C11), CD28 (clone 37.51), CD3ε-fluorescein isothiocyanate (FITC) (clone 145–2C11), CD11b-FITC (clone M1/70), CD4-phycoerythrin (PE) (clone GK1.5), CD8-PE (clone 53–6.7), CD51-PE (clone RMV-7), RANKL-PE (clone IK22/5) and rat PE-negative control (clone KLH/G2a-1-1) were from e‐Bioscience (San Diego, CA). Monoclonal CD19-FITC (clone 6D5), CD25-RPE (clone PC61.5.3), rat anti-mouse RANK (clone LOB14-8), rat FITC-negative control (clone LO-DNP-16) and F(ab ¢)2 goat anti-rat immunoglobulin G-FITC were from Serotec (Oxford, UK).

Cells were obtained in calcium- and magnesium-free phosphate-buffered saline (PBS; Invitrogen, Carlsbad, CA, USA), and incubated in phenol red-free RPMI-1640 (spleen cells) or in phenol red-free Minimum Essential Medium α Modified (α-MEM) (bone marrow cells). Media was supplemented with 10% charcoal-stripped, heat-inactivated fetal bovine serum (FBS), 2 mmol/l glutamine, 100 IU/ml penicillin G and 100 µg/ml streptomycin (all from Gibco, Invitrogen). Calcitriol (1α,25-dihydroxy vitamin D₃) was from Alexis (San Diego, CA, USA). PCR primers and enzymes for reverse transcription were from Invitrogen. All other chemical reagents were from Sigma (Sigma-Aldrich, St Louis, MO, USA).

Mice

The ethics committee of our institution approved all animal procedures. Ten-week-old female C57BL/6 mice (Harlan Interfauna Ibérica, Barcelona, Spain) were subjected to either dorsal ovariectomy or sham operation under general anaesthesia by using 150 mg/kg of ketamine (Merial, Lyon, France) and 5 mg/kg acepromazine (Calmo Neosan, Pfizer, NY, USA). Animals were kept at 21°C with a 12-light:12-h dark cycle and were allowed free access to a pelleted standard mouse laboratory diet containing 0.88% calcium and 0.59% phosphorus (Panlab, Barcelona, Spain) and tap water. The mice were killed 5 weeks after surgery under halothane anaesthesia (Fluothane; Zeneca, Macclesfield, UK) by cardiac puncture, and the blood, spleen and femora were removed aseptically. The uterus of each mouse was also removed and weighed to confirm the success of ovariectomy surgery. Blood was allowed to clot and the serum was separated and frozen at –80°C until analysis. The spleen and femora were processed immediately. None of the mice exhibited evidence of infectious disease, impaired growth (see Results), immunosuppression or other side-effects.

Cell culture

The spleen cells were isolated by careful disintegration of spleen with a scalpel in PBS or by passing the tissue through a sterile 90-µm nylon mesh in PBS. The bone marrow cells were isolated from the femur using centrifugation (Cenci et al., 2000). Briefly, after removal of bone-adherent soft tissues, femur ends were cut off with scissors and inserted into a 0.6-ml microcentrifuge tube with a hole in the bottom and placed in a 1.5-ml carrier microcentrifuge tube. Cells were extracted from the bones by centrifugation at 4°C and resuspended immediately in PBS.

Spleen cells were carefully layered over Histopaque 1077 (Sigma) and centrifuged for 30 min at 400 g at room temperature. Mononuclear cells were collected, washed twice in PBS, resuspended in RPMI containing FBS and then incubated. The bone marrow suspension cells were treated similarly, except that they were allowed to adhere overnight in six-well plastic plates to deplete mature stromal cells (Dao et al., 1997; Cenci et al., 2000), with the exception of flow cytometry analysis and RNA extraction, in which this step of adhesion was omitted. After incubation, non-adherent cells were harvested by pipetting, and mononuclear cells were obtained as described for spleen cells. Mononuclear non-adherent bone marrow cells were resuspended in α-MEM containing FBS. Spleen and bone marrow cells were counted using a haemocytometer, and aliquots of these cells were used for RNA isolation.

Serum biochemistry

Serum calcium, measured by reaction with o-cresolphthalein, and ALP were assayed using commercial kits (Sigma). Serum levels of TNF-α, IL-6, interferon-γ and IL-2 were measured by an enzyme-linked immunosorbent assay according to the manufacturer’s instructions (Diaclone, Besançon, France). Total tartrate-resistant acid phosphatase (TRAP) was determined by a rapid microplate colorimetric assay with modifications (Lau et al., 1987). p-Nitrophenyl phosphate (10 mmol/l) was used as substrate in a buffer containing 100 mmol/l sodium acetate, pH 5.0, 20 mmol/l sodium tartrate and 0.1% (v/v) Triton X-100. Serum samples (10 µl) were added to 200 µl of substrate and incubated at 37°C for 60 min. The reaction was stopped by the addition of 100 µl 0.5 mol/l NaOH, and absorbance was read at 405 nm with a microplate reader (Model 550, Biop-Rad, Richmond, CA, USA).

Resorption pit assay

Bone marrow preosteoclasts were characterized by assessing their ability to form resorption pits on calcium phosphate-coated culture wells. For this, 5 × 10⁴ non-adherent mononuclear bone marrow cells in α-MEM containing 10% FBS were seeded onto calcium phosphate-coated osteological discs (Millenium Biologix, Kingston, Ontario, Canada) and incubated for 12 days in a humidified atmosphere of 5% CO₂ in air in the presence of vehicle, M-CSF (20 ng/ml) or M-CSF (20 ng/ml) plus murine sRANKL (20 ng/ml). The medium was changed three times per week, eliminating half of the medium and replacing it with fresh medium plus stimuli. After 12 days, adherent cells were removed with bleach solution (6% NaOCl and 5.2% NaCl). The discs were examined for the presence of resorption lacunae by light microscopy after von Kossa staining, according to the manufacturer’s instructions.

Cell proliferation and viability analysis

Cell proliferation and viability were evaluated with an XTT (sodium 3¢-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis (4-methoxy-6-nitro) benzene sulfonic acid hydrate)-based colorimetric assay (Roche Diagnostics, Mannheim, Germany) according to the manufacturer’s instructions. Briefly, 1.5 × 10⁵ spleen or 4 × 10⁴ non-adherent bone marrow mononuclear cells in 100 µl were placed in 96-well microtitre plates (Orange Scientific, Braine-l’Alleud, Belgium) supplemented with glutamine, antibiotics and 10% FBS at 37°C in a humidified atmosphere of 5% CO₂ in air for 4 days. The stimuli and their combinations used are indicated in Table II, and were added in triplicate. After incubation, 50 µl of XTT substrate was added to each well and incubated in the conditions described above. Absorbance of each well was determined at 450 nm using a microplate reader. Proliferation data in Table II represent the quotient of proliferation for each stimulus (in triplicate) and proliferation of the same cells in the same culture plate incubated with vehicle.

Flow cytometry and DNA staining

Cells were labelled with antibodies at the manufacturer’s recommended dilutions and conditions. Mononuclear spleen and bone marrow cells (1 × 10⁶) were incubated for 30 min on ice with the indicated antibodies and then washed and resuspended in PBS containing 1% bovine serum albumin and 0.1% sodium azide. Non-specific signal was estimated by incubation with rat FITC- and PE-conjugated IgG isotype controls. For analysis of the cell cycle, cells were labelled with Coulter DNA-Prep (Beckman Coulter, Fullerton, CA, USA) according to the manufacturer’s instructions. Labelled cells were analysed with an EPICS-XL flow cytometry system (Beckman Coulter). Data were expressed as the percentage of positive cells since, although ovariectomy induced approximately 20% increment in spleen and bone marrow total cell content, there were no significant differences between ovariectomized and sham-operated mice.

RNA isolation and semiquantitative RT-PCR

Total RNA was extracted from mononuclear bone or spleen cells with Trizol reagent (Invitrogen) according to the manufacturer’s protocol, and the integrity of the RNA preparations was examined by agarose gel electrophoresis. One microgram of total RNA was reverse-transcribed in a 20-µl reaction volume into single-stranded cDNA with a first-strand cDNA synthesis kit using an oligo-dT primer (Invitrogen). The subsequent PCR was performed with specific primers for each gene with Taq polymerase from Sigma. To ensure equal starting quantities of cDNA for the experiment and to allow the semiquantification of the PCR products, reverse-transcribed RNA samples were amplified with primers specific for glyceraldehyde phosphate dehydrogenase (GAPDH). Amplifications were done using a GeneAmp 9600 thermal cycler (Perkin-Elmer), with the temperature cycling set according to the primer length and Tm value. The number of cycles for each primer pair was determined according to a linear amplification curve established from primary experiments. For each PCR reaction, several samples underwent 12–15 extra cycles of amplification to guarantee the semiquantification of the experiment. Amplified PCR products were separated on 2% agarose gel and stained with ethidium bromide for visualization. The intensity of ethidium bromide-stained bands was quantified using the histogram function in Adobe Photoshop (version 7.0) and was normalized with the GAPDH housekeeping gene.

The sequence of primers used were the following: GAPDH, sense 5′-ACC ACA GTC CAT GCC ATC AC-3′ and antisense 5′-TCC ACC ACC CTG TTG CTG TA-3′; CD25, sense 5′-CTC TCC TAC AAG AAC GGC AC-3′ and antisense 5′-TCA CTA GCC AGA AAT CGG TGG-3′; core binding factor a1 (Cbfa1), sense 5′-CCG CAC GAC AAC CGC ACC AT-3′ and antisense 5′-CGC TCC GGC CCA CAA ATC TC-3′; RANKL, sense 5′-CAT TTG CAC ACC TCA CCA TC-3′ and antisense 5′-AAG GGT TGG ACA CCT GAA TG-3′; IL-6, sense 5′-ATG AAG TTC CTC TCT GCA AGA GAC T-3′ and antisense 5′-CAC TAG GTT TGC CGA GTA GAT CTC-3′; TRAP, sense 5′-ACT TCC CCA GCC CTT ACT ACC-3′ and antisense 5′-TCA GCA CAT AGC CCA CAC CG-3′; TNF-α, sense 5′-TCT TCT GTC TAC TGA ACT TCG G-3′ and antisense 5′-GTA GAG AAT GGA TGA ACA CCC-3′; interferon-γ, sense 5′-TCT TGG CTT TGC AGC TCT TCC-3′ and antisense 5′-CGA ATC AGC AGC GAC TCC TTT TC-3′.

Statistical analysis

Analysis of variance test was used for statistical comparisons between means of different groups. Results are presented as mean ± SD. The numbers of mice used in each experiment appear in the tables or in the text of the figures; they correspond to three surgical interventions in different weeks. A P value < 0.05 of was considered statistically significant. The entire statistical analysis was carried out using the Statistical Package for Social Sciences (SPSS, Chicago, IL, USA), v. 11.0 for Windows.

Results

Serum parameters in ovariectomized mice

The success of ovariectomy was determined from the uterine hypoplasia caused by estrogen deficiency (Table I and Figure 3A). Ovariectomy induced a dramatic decrease in uterine weight but not body weight. Ovariectomy increased bone resorption, as shown by increased values of serum calcium (P < 0.01; Table I). Serum levels of ALP, a formation marker of bone remodelling, were lower in ovariectomized mice (P < 0.01). Furthermore, ovariectomy caused a significant decrease in serum levels of interferon-γ (P < 0.01), a T-cell produced cytokine that inhibits osteoclastogenesis in vitro. Other markers (TNF-α, IL-6, IL-2 and TRAP) remained unchanged. The number of resorption pits generated by mononuclear non-adherent ovariectomy bone marrow cells after incubation with RANKL and M-CSF was more than twice that found in sham-operated operated mice (P < 0.05; Table I and Figure 1). The number of pits formed when cells were incubated with vehicle or M-CSF alone was very small (<10 pits/well), with no difference between ovariectomized and sham-operated mice.

Cell proliferation analysis

To evaluate the functionality of mononuclear spleen and non-adherent bone marrow cells, in vitro cell proliferation assays were carried out in response to different stimuli (Table II). Spleen cells of ovariectomized mice showed a significantly higher rate of cell proliferation only in response to 4β-phorbol-12-myristate-13- acetate (PMA). The proliferation of bone marrow cells from ovariectomized mice showed a significantly smaller proliferation rate in response to CD3 + CD28 (P < 0.05). Proliferation in response to other stimuli, such as PMA, concanavalin A (Con A), phytohaemagglutinin (PHA) and M-CSF, remained unaffected after ovariectomy.

Flow cytometry analysis of spleen and bone marrow cells

To examine the effects of estrogen deficiency on cell subpopulations, mononuclear cells from spleen and bone marrow were analysed by flow cytometry. The lymphoid region, which contains the greatest number of lymphocyte/blast cells, was gated for analysis both in spleen as well as in bone marrow (Figure 2).

In spleen, ovariectomy caused an increase in CD25⁺ cells (IL-2 receptor α-chain positive cells; P < 0.05) (Table III), with no change in the percentages of CD3⁺, CD3⁺/CD4⁺, CD3⁺/CD8⁺, CD19⁺, CD11b⁺ (type 3 complement receptor positive cells) and CD51⁺ (αV-chain of vitronectin receptor positive cells) cells. Ovariectomy caused a greater change in bone marrow cell phenotype (Table IV). In the lymphoid region of the bone marrow, ovariectomized mice showed an accumulation of CD19⁺ (P < 0.05) and CD3⁺/RANKL⁺ (P < 0.01) cells and decreases in the percentages of CD11b⁺ (P < 0.05) and CD51⁺ cells (P < 0.01) (Table IV).

We also analysed the percentage of spleen and bone marrow cells in the G1 (diploid cells), G2 (tetraploid cells) and S (DNA synthesis) phases of the cell cycle by propidium iodide staining (results not shown). In spleen, there were no significant differences among groups. In contrast, the bone marrow of ovariectomized mice showed a lower percentage of cells in the G1 phase (75.97 and 73.61% respectively; P < 0.05) and an accumulation of cells in the G2 phase (9.73% and 13.53% respectively; P < 0.01), showing increased haematopoiesis 5 weeks after ovariectomy.

Gene expression

The analysis of gene expression of spleen and bone marrow mononuclear cells did not exhibit major differences between ovariectomized and sham-operated mice (Figure 3B and C). In bone marrow, a significant increase in CD25 gene expression was found after ovariectomy (P < 0.05). The expression of interferon-γ, Cbfa-1, RANKL, IL-6, TRAP and CD61 did not change with ovariectomy in these mononuclear cells. Analysis of the gene expression of spleen mononuclear cells did not produce any major differences between ovariectomized and sham-operated mice.

Discussion

In the present work, the status of the immune system was assessed 5 weeks after ovariectomy, when mice had developed osteopenia and bone turnover had tended to normalize (Jilka et al., 1998Tanizawa et al., 2000; Miyazaki et al., 2004).

Five weeks after surgery, increased bone resorption persisted, as shown by augmented levels of serum calcium and by increased resorption pits generated by bone marrow cells (Table I and Figure 1). The number of pits should be proportional to the number of preosteoclasts in bone marrow since bone marrow cells were mononuclear cells depleted of mature stromal cells and incubated with RANKL and M-CSF. However, we cannot discard the possibility that this increase reflects an augmentation in the resorptive activity without a change in the number of preosteoclasts. The results confirm the hypothesis that increased osteoclastogenesis persists 5 weeks after ovariectomy. This agrees with previous work in which an increased number of osteoclastic cells 5 weeks after ovariectomy was described, although osteoclastic cell numbers declined over time (Jilka et al., 1998). Results of the present work show no increases in serum levels of IL-6 and TFN-α, although a trend can be observed (Table I). Nevertheless, inflammatory cytokines are involved in the process since it has been reported that blocking IL-1 and TNF-α by antibodies prevents bone loss after ovariectomy (Kimble et al., 1995).

Two unexpected results can be seen in Table I: the small amount of serum ALP-activity and the low level of interferon-γ after ovariectomy. ALP activity is considered a marker of bone remodelling (Weaver et al., 1997; Minisola et al., 1998), and a high correlation exists between bone-specific and total ALP (Takahashi et al., 1997). After ovariectomy, levels of total and bone-specific ALP increase, peak between the first and second weeks, and then decline progressively (Tanizawa et al., 2000; Miyazaki et al., 2004). Although total ALP activity is not bone-specific, it can be assumed that, if there is no hepatic disease (Minisola et al., 1998), it should reflect changes caused by bone marrow cells (Weaver et al., 1997). This decreased level of ALP activity 5 weeks after ovariectomy implies that bone formation is impaired, and together with the increase in resorption pit-generating cells in bone marrow, could explain the increased calcaemia detected, thus linking estrogen deficiency with osteoblastic function. We cannot discard the possibility that the decreased levels of ALP after ovariectomy, however, are related to the low bone density reached by our strain of mouse (C57BL/6) compared with others, as a consequence of a lower rate of bone formation (Richman et al., 2001). In accordance with our data, after ovariectomy, decreases have been described in the percentage of preosteoblastic marrow stromal cells in early postmenopausal women (Eghbali-Fatourechi et al., 2003), in serum ALP activity (Sakakura et al., 2001), in the number of ALP-positive colony-forming units (Pei et al., 2003) and in ALP activity in individual osteoblasts (Gevers et al., 2002).

Interferon-γ is an in vitro negative regulator of osteoclastogenesis (Gowen and Mundy, 1986; Takayanagi et al., 2000). Nonetheless, in vivo interferon-γ cures osteopetrosis by stimulating osteoclast formation and bone resorption (Key et al., 1995), and interferon-γ receptor-deficient mice have been shown to be protected against ovariectomy-induced bone loss (Cenci et al., 2003). The issue of interferon-γ is complex because this cytokine is likely to have tissue specific effects. Data in the literature about interferon-γ production after gonadectomy are sparse, however, particularly in the bone marrow microenvironment. Cenci and colleagues have described an increase in the percentage of CD4⁺ T cells expressing interferon-γ and in the interferon-γ concentration in culture supernatants of bone marrow CD90⁺ T cells after ovariectomy, concluding that this increase is critical in explaining the bone effects in estrogen deficiency (Cenci et al., 2003). However, most data indicate that estrogens up-regulate and gonadectomy down-regulates interferon-γ production, although these studies were not carried out in bone marrow cells. For instance, it has been reported that interferon-γ production is decreased after mice gonadectomy (Aloisi et al., 2001; Sun et al., 2003), but up-regulated by estrogen in spleen and lymph node lymphocytes (Karpuzoglu-Sahin et al., 2001; Maret et al., 2003). Moreover, a decrease in the interferon-γ production has also been described in postmenopausal women (Yang et al., 2000; Kumru et al., 2004). These data suggest a difference in the regulation of interferon-γ production for T cells from peripheral tissues (spleen and lymph node) and bone marrow T cells after gonadectomy. Although the rationale for such impairment remains unclear, it might be due to differences between systemic and local interferon-γ production. Indeed, interferon-γ implicated in osteoclastogenesis should be produced mainly in the bone marrow, and it remains to be demonstrated whether the decrease in interferon-γ production detected in this and other studies has an effect on the increased osteoclastogenesis that is established after gonadectomy (Teitelbaum, 2004).

Estrogen deficiency induced important changes in cellular subtypes of spleen and bone marrow. Ovariectomy induced an increase in levels of B-cells (CD19⁺) in bone marrow and CD25⁺ cells (IL-2 receptor α chain) in spleen as well as in bone marrow (total cells; results not shown). Both estrogen deficiency (this work and Masuzawa et al., 1994) and IL-7 administration (Miyaura et al., 1997) induce B-lymphopoiesis, a process that may be involved in the mechanism of stimulated bone loss. In this respect, it has been reported that B-cells are a source of osteoclasts (Sato et al., 2001). Moreover, an increase in CD25⁺ T cells in bone marrow of ovariectomized mice (Cenci et al., 2003) and of CD25⁺ and HLA-DR⁺ T cells in postmenopausal women (Yang et al., 2000) has been described, although this effect has not been well explained, particularly if IL-2 levels do not change with ovariectomy (Table I).

Estrogen receptors are detectable in the reticular stromal cells of the thymus (Barr et al., 1982), suggesting that estrogen can modulate T-cell lymphopoiesis. Nevertheless, it is not clear whether ovariectomy modulates T-cell levels. In the present study, ovariectomy did not affect T-cell levels (total cells or percentage), using a CD3 monoclonal antibody. This result agrees with other work, in which it has been reported that T cells (Thy 1.2⁺) do not change (total cells or percentage) after mouse ovariectomy (Masuzawa et al., 1994). An increase in bone marrow T-cell content has been described after ovariectomy (Cenci et al., 2000; Roggia et al., 2001), but it is possible that the authors gated immature B cells as T cells, since the CD90 monoclonal antibody used is specific for CD90.1 (Thy 1.1⁺) and CD90.2 (Thy 1.2⁺), and an increase in Thy 1.1⁺ cells of B-cell lineage after ovariectomy has been described (Erben et al., 1998b). Results of the present study support previous data from other authors showing that ovariectomy either does not affect or diminishes mouse and rat bone marrow T-cell levels (Masuzawa et al., 1994; Erben et al., 1998a; Safadi et al., 2000), as well as the results of a study in postmenopausal women (Yang et al., 2000). However, T-cell-deficient mice do not lose bone after ovariectomy, thus indicating a determinant role of T-cells in bone loss in estrogen deficiency (Cenci et al., 2000). Although it is not possible to discard possible roles of monocytes and NK cells, T cells are responsible for increased production of pro-resorptive cytokines after ovariectomy (Jilka, 1998; Theill et al., 2002), and the decreased interferon-γ level found in the present study after ovariectomy.

Ovariectomy did not change the percentage of RANK-expressing cells. RANKL expression in peripheral tissues is very low (Anderson et al., 1997; Josien et al., 1999). We have found, however, that levels of bone marrow CD3⁺/RANKL⁺ cells are higher after ovariectomy. The increase in RANKL⁺ T cells could participate in the increased osteoclastogenesis after ovariectomy, as has been described (Kong et al., 1999).

Percentages of CD11b⁺ and of CD51⁺ cells were smaller in the bone marrow of ovariectomized mice, although, confirming previous results (Masuzawa et al., 1994), the total cell content remained unchanged (results not shown). CD11b (Mac-1) is a marker of the myeloid lineage, while CD51 is directed against the αV-chain of the vitronectin receptor (CD51/CD61, αVβ3) and is considered an osteoclast marker, although it is expressed in platelets, T cells and granulocytes. Since osteoclasts derive from myelomonocytic lineage cells, one could expect that ovariectomy would increase the monocyte–macrophage precursors in bone marrow. Confirming our results, a time-course study showed that percentages of CD11b⁺ cells and Gr-1⁺ cells (myeloid cells and granulocytes, respectively) decreased but B220⁺ cells (B cells) were selectively increased 2–4 weeks after ovariectomy (Masuzawa et al., 1994). It is possible, though, that the transitory increase in myeloid precursors is lost or that the osteoclast precursor represents only a small fraction of CD11b⁺ and CD51⁺ cells. In addition, it is possible that we need a better specific marker to detect increases in osteoclast lineage cells. In this respect, a transitory increase in ED1⁺ cells (myeloid marker of monocyte and osteoclast lineage) has been described 2 weeks after ovariectomy coinciding with the up-regulation of osteoclast numbers (Erben et al., 1998b; Jilka et al., 1998), the expression returning to normal after 4 weeks (Erben et al., 1998b).

In the present study, cell functionality was similar in ovariectomized and sham-operated mice. Spleen cells from ovariectomized mice showed increased proliferation in response to PMA, a specific activator of protein kinase C (PKC), suggesting that there is an increase in the amount and/or activity of PKC. In fact, it has been described that the PKC-β gene is upregulated during osteoclastogenesis (Lee et al., 2003). Mononuclear bone marrow cells from ovariectomized mice exhibited lessened proliferation in response to specific stimulation of T-cells (CD3 + CD28 monoclonal antibodies), which could be due to impairment of T-cell functionality and could explain the low level of interferon-γ observed in this work.

Gene expression analysis by RT-PCR did not offer substantial data. The only significant difference found was higher CD25 expression in bone marrow. Other authors have found changes in rat osteoprotegerin and RANKL gene expression in the first days after ovariectomy, but not in the fourth week (Bonnelye et al., 2002), suggesting that many of the changes in gene expression occur in the first days after ovariectomy (Tanizawa et al., 2000; Miyazaki et al., 2004).

In summary, the main effects of estrogen deficiency on the immune system seem to involve haematopoiesis. It seems that estrogen receptors are present only in certain immature immune cells or in a small population of CD8⁺ T cells and macrophages (Igarashi et al., 2001). Therefore, the effects of estrogen could be due to direct action on mature lymphoid cells or, more probably, through interaction with other estrogen receptor-expressing cells that can regulate lymphoid cells, such as dendritic or bone marrow stromal cells (Bellido et al., 1993; Komi and Lassila, 2000). In bone marrow, ovariectomy up-regulated CD3⁺/RANKL⁺ cells and B cells, did not affect T-cell levels, and decreased the percentage of myeloid cells, although only B cells increased in absolute numbers after ovariectomy. The decrease in serum interferon-γ level, which can be in the base of the incremented osteoclastogenesis found, could reflect impaired T-cell functionality. Furthermore, ovariectomy significantly diminished serum ALP activity, indicating a lower rate of bone formation 5 weeks after ovariectomy.

Acknowledgements

The authors are indebted to Drs Enrique O’Connor and Guadalupe Herrera for their expert assistance with flow cytometry, and to Mrs Rosa Aliaga and Mrs Elvira Calap for their excellent technical assistance. This work was supported by grant 01/3051 from the Fondo de Investigaciones Sanitarias (FIS).

References

Author notes

1Research Unit, Hospital Clínico Universitario of Valencia, Av. Blasco Ibáñez 17, 46010 Valencia, Spain, 2Research Unit, Faculty of Medicine, University of Valencia, 46010 Valencia, Spain, 3Department of Physiology, Faculty of Medicine, University of Valencia, 46010 Valencia, Spain, 4Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Valencia, 46010 Valencia, Spain, 5Department of Functional Biology and Physical Anthropology, Faculty of Biological Sciences, University of Valencia, 46100 Burjassot, Spain and 6Department of Pediatrics, Obstetrics and Gynecology, Faculty of Medicine, University of Valencia, 46010 Valencia, Spain