Plasma Osteopontin Modulates Chronic Restraint Stress-Induced Thymus Atrophy by Regulating Stress Hormones: Inhibition by an Anti-Osteopontin Monoclonal Antibody¹

Osteopontin (OPN)³ is a pleiotropic phosphoglycoprotein that is broadly expressed and up-regulated during inflammation, autoimmune diseases, cancer development, and various stress conditions (1, 2, 3). It interacts with different cell surface receptors, including integrins and certain CD44 isoforms, and can induce PI3K/Akt-dependent NF-κB activation (4). It is difficult to determine the molecular mechanism of a specific effect of OPN due to the interplay with various factors, including its ability to engage multiple integrins and CD44 variants; its posttranslational modifications; its cleavage state; and its localization both intracellularly and extracellularly. OPN has important cytokine and chemokine functions and is a key stress mediator (4). Its roles in mediating oxidative stress (5), mechanical stress (6), and cellular stress (7) have been well documented. We have demonstrated that OPN is at least partially responsible for hind limb-unloading (HU) stress-induced losses in peripheral lymphocytes and thymocytes (8). This type of stress leads to rapid systemic changes in stress hormone production, immune cell distribution, and cytokine/chemokine production; it also affects peripheral immune organs in the immune system (9). OPN-deficient mice showed significantly milder changes in response to this stress. However, the extent to which each of these different stress paradigms affects the HPA (hypothalamus-pituitary-adrenal) axis and the immune system remains to be determined.

Another murine stress model, chronic restraint stress (CRS) has been widely used in studies of the effect of stress hormones (10) and immune cell functions in mice (11, 12). CRS consists of a scheduled confinement and restriction of food and water during restraint. In addition to physical immobilization, psychological stress plays a significant part in this model (13). We used this model to evaluate OPN-deficient mice in both 129 and BALB/c backgrounds to determine their stress response as assessed by thymic mass, changes in corticosterone (CORT), and adrenocorticotropic hormone (ACTH) levels and leukocyte trafficking as compared with their wild-type (WT) counterparts.

To further verify the specific role of OPN in the lymphocyte stress response and its effect on HPA axis hormones, we used mouse fibroblast-derived OPN and anti-OPN mAbs to evaluate, respectively, the ability of OPN to restore the WT phenotype to OPN^−/− mice or the effect of depletion of OPN with anti-OPN mAbs in inhibiting the stress-induced lymphoid organ atrophy and associated hormonal changes in WT mice. Our results show that exogenously supplied OPN can sensitize OPN^−/− mice to CRS-induced thymus atrophy, demonstrating a critical role of OPN in organ atrophy and CORT production. In contrast, WT mice that received the mAb 2C5 exhibited a significant protection from CRS-induced reduction in thymus and spleen cellularity. These results support the conclusion that plasma OPN regulates stress-induced organ atrophy and demonstrate the involvement of OPN in the bidirectional communication between the CNS and the immune system.

OPN^−/− mice in the 129 background were generated (14, 15) and maintained along with isogenic WT controls in the Rutgers Nelson Animal Facility, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care and is under the care of a board-certified veterinarian. The research with these mice was approved by the Rutgers Institutional Animal Care and Use Committee Protocol 97:031.

OPN^−/− mice on the BALB/c background were provided by Drs. M. Shinohara and H. Cantor (Dana Farber Cancer Institute, Boston, MA) (16). The breeding pairs were homozygous OPN^−/− mice with 15 generations of backcrossing to the BALB/c background therefore considered 99.99% pure BALB/c background. Breeding pairs were bred and maintained in the Rutgers Nelson Animal Facility as above. BALB/c WT control mice were purchased from The Jackson Laboratory. All animals used in the experiments were age and sex matched.

Mouse OPN was purified from serum-free medium conditioned by a ras-transformed murine embryonic fibroblast line (275-3-2) (17). The medium was incubated with 1 ml of protein G beads (Pierce), to which the 2A1 anti-OPN mAb had been cross-linked. The beads were washed and packed into a 2-ml disposable column. OPN was eluted from the 2A1-protein G beads with 100 mM glycine and 500 mM NaCl (pH 2.5) and collected into tubes containing a neutralizing pH 8 Tris-Cl buffer. Fractions were analyzed by SDS-PAGE and proteins were visualized by non-ammoniacal silver staining and Western blotting. Positive fractions were pooled, desalted on PD-10 columns (GE Healthcare Bio-Sciences), quantified by ELISA, and lyophilized.

The monoclonal anti-OPN Abs used (2A1, 3D9, 1G4, and 2C5) were generated and characterized in our laboratory (18). Abs were purified from ascites fluid obtained from mice injected with the hybridomas encoding them.

Eight- to 10-wk-old mice were subjected to an established CRS protocol with some modifications (11). Briefly, OPN^+/+ and OPN^−/− mice were each divided into control and stress groups. Mice used in a study were randomized by evenly distributing mice from the same litter to different treatment groups so as to minimize the influence of litter and age variations. Mice were immobilized individually in well-ventilated cylindrical wire mesh restrainers sized 12 cm (length) × 3 cm (diameter) that were closed off at both ends. The restrained mice were maintained horizontally in their home cages during the restraint sessions and released into the same cages during the free sessions. They were restrained for 12 h daily followed by a 12-h recovery period. Food and water were provided during the recovery period ad libitum; therefore, food and water deprivation during the restraint sessions should be considered as part of the stress. Control animals were housed in their home cages with food and water provided ad libitum at all times; when appropriate, they were injected with PBS to control for Ab or OPN injections. Six mice were used in each treatment group whenever possible or the data from parallel experiments were combined for the statistical analysis. BALB/c and 129 mice were restrained for two and three 12-h periods, respectively. Due to strain genetic differences, BALB/c mice are more sensitive to restraint treatment and significant organ atrophy appeared after two cycles of CRS; however, three cycles were required for 129 mice to exhibit significant organ atrophy (K.X.W., unpublished data).

At the end of the final restraint stress session, animals were euthanized by CO₂ inhalation and the blood, spleen, and thymus were harvested. Blood was drawn immediately after euthanasia by cardiac puncture. Plasma samples were collected in 50 mm EDTA in PBS and stored at −70°C. The spleen and thymus were excised and put into 1 ml of cell culture medium. The weight of each organ was recorded.

Purified mouse fibroblast OPN was rehydrated from lyophilized stock and diluted in sterile PBS before use. This OPN is phosphorylated randomly at a few serine/threonine sites (out of some 30 potential sites) and presumably behaves as unphosphorylated OPN (19). OPN^−/− mice were divided into three groups: A, control group; B, CRS group injected with PBS; and C, CRS group injected with OPN in PBS. Mice in group C were injected i.p. daily with 5 μg of OPN in 100 μl of PBS starting 3 days before the first restraint cycle and continued through the restraint sessions. Mice received 25–30 μg of OPN by the end of treatment depending on the number of restraint cycles. The mice in group B received 100 μl of PBS, using the same schedule as group C, and were similarly restrained. Mice in control group A were kept in their home cages undisturbed.

The mAbs were diluted in sterile PBS to 0.66 μg/μl. WT OPN^+/+ mice were divided into the control group (A), CRS group injected with PBS (B), and CRS group injected with anti-OPN mAb in PBS (C). Mice in group C were injected with 100 μg of anti-OPN mAb in 150 μl of PBS i.p. starting 24 h before the first restraint cycle and then immediately before each restraint session. The total amount of anti-OPN mAb received at the end of treatment was 300–400 μg depending on the number of restraint cycles. CRS group B received 150 μl of PBS and was subjected to restraint as group C. Control group A mice were kept in their home cages undisturbed.

The levels of CORT in plasma samples were assessed with a CORT ELISA kit from IBL America (catalog no. RE52211) according to the manufacturer’s instructions. The levels of ACTH were measured using an ACTH ELISA kit from MD Biosciences (catalog no. ACTH.96) according to the manufacturer’s instructions.

High-binding ELISA plates were coated with 0.8 μg/ml anti-OPN Ab (AF808; R&D Systems) in PBS overnight at 4°C. Coated wells were blocked with 1% BSA and 5% sucrose in PBS and incubated for 1 h before samples were applied to wells. Plasma samples were diluted 1/100 in assay diluent (PBS plus 1% BSA) and 100 μl of diluted samples was added to the wells. After a 2-h incubation at room temperature, the plate was washed and detection was performed by incubating the plate with 100 μl of biotinylated anti-OPN mAb at 0.1 μg/ml (BAF808; R&D Systems) at room temperature for 2 h. After the plate was washed, a secondary detection reagent, 100 μl of streptavidin-HRP (1/200 dilution, DY998; R&D Systems) was added to the plate and incubated for 20 min. For color development, 100 μl of 3,3′,5,5′-tetramethylbenzidine liquid substrate (T8665; Sigma-Aldrich) was added to the washed plate and incubated for 15–20 min. Color development was terminated with 50 μl of the stop solution and absorbance was determined by a spectroMax microplate reader (Molecular Devices) at 450 nm. Recombinant mouse OPN (441-OP; R&D Systems) was used as a protein standard in the OPN ELISA. The assays were conducted in triplicate.

To verify the involvement of OPN in the stress response revealed in our previous research using the HU model (8), we used the CRS model (11). To reduce the possibility that differences in genetic background could affect the response to stress, OPN^−/− mice in a BALB/c background were also tested in parallel with 129 mice. When subjected to CRS, a statistically significant difference in body weight reduction between WT and OPN-deficient mice was observed (p = 0.046); the WT BALB/c mice exhibited 8.0 ± 0.6% reduction in body weight, whereas the OPN-deficient BALB/c mice showed a 4.4 ± 1.1% body weight reduction (20). As shown in Fig. 1,A, CRS caused a 60% reduction in thymus weight in 129 OPN^+/+ mice but only a 30% reduction in the OPN-deficient mice. A 30% reduction of spleen weight was observed in WT mice compared with a 10% reduction in OPN^−/− mice; similar responses were observed using WT and OPN-deficient BALB/c mice (Fig. 1 B). These results indicated that CRS caused lymphoid organ atrophy in both BALB/c and 129 WT mice to a significantly greater extent than in OPN^−/− mice. This confirms that OPN indeed plays a role in mediating stress-induced responses in immune organs.

Because OPN has been found to regulate CORT production in response to HU stress (8), we were interested in determining whether OPN directly affects CORT production or acts by regulating upstream hormones in the HPA axis. Corticosteroid provides negative feedback regulation to other stress hormones in the HPA axis by inhibiting the secretion of ACTH and corticotropin-releasing hormone (21). Thus, in addition to evaluating the CORT levels, we evaluated plasma levels of ACTH to more closely localize where OPN interacts with the HPA axis.

The levels of CORT and ACTH in blood samples harvested immediately after the termination of CRS were tested with commercial ELISA kits. Results showed that the level of CORT was highly up-regulated in stressed WT mice. However, in OPN^−/− mice, there was no significant difference in CORT levels between control (unstressed) and stressed mice (Fig. 2,A). Interestingly, the basal level of CORT in unstressed mice was significantly higher in OPN^−/− mice as also shown in the HU model, implying that OPN plays a role in controlling the production of CORT; in the absence of OPN, CORT production is apparently elevated, leading to a persistent high level of CORT in the circulation in the absence of applied stress. On the other hand, the classic negative feedback mechanism of CORT toward ACTH in response to chronic stress suggests that up-regulation of CORT could lead to a reduction of ACTH. The results from the ACTH ELISA reflect this reciprocal relationship by showing that the levels of ACTH in WT mice were higher in control mice but largely suppressed in the stressed mice (Fig. 2 B). However, in the OPN^−/− mice, the ACTH levels were reproducibly significantly lower than that in the WT mice and did not respond to CRS, suggesting that ACTH secretion was inhibited by the persistent high level of CORT in the OPN^−/− mice. These results suggest a critical role for OPN in the regulation of HPA axis functions.

To demonstrate directly that stress-induced organ atrophy is promoted by the presence of OPN, purified mouse OPN was injected into OPN^−/− mice before and during CRS. OPN^−/− mice were injected i.p. with 5 μg of OPN daily for 3 days before subjecting to restraint. Similar injections were made at the beginning of each 24-h restraint cycle to maintain levels of exogenous OPN in the circulation during the restraint. As shown in Fig. 3, CRS led to a larger reduction of thymus weight (38%, p = 0.016) in BALB/c OPN^+/+ mice compared with OPN^−/− mice (27%, p = 0.078). Elevation of plasma OPN levels by repeated i.p. injections of OPN into BALB/c OPN^−/− mice resulted in a larger (51%, p = 0.021) reduction of thymus weight. These results have been closely reproduced in several other experiments using OPN^−/− mice in the 129 background (20). To confirm the presence of OPN circulating in the blood during this experiment, plasma samples harvested at the end of the experiments were assayed for OPN levels. Fig. 4 shows that OPN levels were up-regulated by CRS by ∼25% in WT mice. In naive OPN knockout mice, as expected, OPN was undetectable. OPN was only detectable in the plasma samples from the experiments in which exogenous OPN was supplied to OPN^−/− mice. The concentration of OPN correlated to the number of injections each animal received. Animals receiving a total of five injections (25 μg) had twice as much OPN in the plasma compared with animals receiving three injections. However, even with up to six injections, the OPN level in the plasma reached only 30% of the WT level. Nevertheless, these results clearly demonstrated that OPN in the plasma is essential for promoting stress-induced lymphoid organ atrophy.

It was encouraging to find that OPN can restore partially the WT phenotype of stress-induced organ atrophy in OPN^−/− mice. The other side of the coin is whether sequestering of OPN in WT mice would protect the lymphoid organs from stress-induced atrophy. To address this question, four different mAbs (2A1, 3D9, 2C5, 1G4) (18, 22) were evaluated for their effectiveness in preventing stress-induced thymus atrophy in WT mice. Each mAb recognizes a distinct epitope (18, 22) on the mouse OPN molecule as indicated in Fig. 5. WT BALB/c or 129 OPN mice were injected with 100 μg of mAb 24 h before starting the CRS cycle and again at the beginning of each restraint cycle. Of the four mAbs tested, only 2C5 supported a statistically significant increase in thymus weight compared with the CRS-only group (Fig. 6). Injection of 2C5 blocked stress-induced reduction in thymus weight to the level of that in the OPN^−/− mice after CRS (Fig. 3). The other three anti-OPN mAbs were without effect. Two additional experiments yielded similar results. We know (18, 22) that 2C5 recognizes an epitope in mouse OPN that is near the amino-terminal side of the RGD integrin binding site and thus may block the interaction of OPN with integrins, although this last remains to be confirmed. Since mAbs recognizing other regions of the OPN molecule (1G4, 2A1, 3D9) were not effective, this result suggests that a RGD-integrin interaction may be important for OPN mediation of the stress response.

When OPN levels in mice injected with an anti-OPN mAb were measured, an average 20% increase of OPN was detected (20). It is known that although Ab binding to the target protein may inactivate the function of that protein by blocking a functional site or by reducing the free state of the target protein, it can also inhibit the turnover of the protein, thereby causing the accumulation in the plasma of the target protein. Nevertheless, administration of 2C5 moderately reversed the severe thymus atrophy caused by stress.

Measurement of CORT levels revealed that injection of OPN into OPN^−/− mice elevated CORT levels in the plasma of the mice after CRS (Fig. 7,A). This indicates that exogenous OPN supplied to OPN^−/− mice partly restores the WT phenotype; the presence of OPN in the circulation may modulate stress hormones and other unknown factors to cause the increased organ atrophy in response to stress. To determine whether an anti-OPN mAb injection could affect the CORT level in response to stress, we examined CORT levels in the plasma of mice receiving the 2C5 mAb. As expected, the injection of 2C5 reduced CORT production in the plasma of WT mice subjected to CRS (Fig. 7 B). This result, along with the observation that OPN promotes CORT production in OPN^−/− mice subjected to CRS, further substantiates the role of OPN in controlling CORT production via the HPA axis.

OPN is up-regulated in various pathological and stress situations (4, 23). Our previous research has revealed that OPN is critical for HU-induced lymphoid organ atrophy through modulating corticosteroid production. However, the role of OPN in other physical and psychological stress responses has remained unexplored. Both HU and CRS are physical stress models containing a significant psychological component that activates neural transmitters and stress hormones (24, 25). We have demonstrated here that OPN^−/− mice are resistant to CRS-induced lymphoid (largely thymus) organ atrophy. The stress-induced up-regulation of CORT production was significantly reduced in OPN^−/− mice. Thymus atrophy was easily detectable in CRS-treated mice, whereas spleen weight loss was sometimes insignificant, possibly because the stress level had not reached the threshold required to cause spleen atrophy. In the more stressful HU model, both spleen and thymus atrophy were consistently observed.

It is well documented that mice subjected to chronic physical stress exhibit an increased rate of lymphocyte apoptosis, redistribution of immune cells to the periphery, and atrophy in the thymus and the spleen (11, 26). Based on these observations, lymphoid organ atrophy is a convenient marker for monitoring stress-induced changes in the immune system. The thymus is a primary lymphoid organ that consists of immature T cells and provides the environment for T cell development. It manifests dynamic physiological changes and is exquisitely sensitive to stress and toxic insult. It quickly responds to chemical and physical challenges, consequently leading to loss of cortical lymphocytes by apoptosis followed by organ atrophy (27). Our data demonstrate both that injection of OPN into OPN-deficient mice enhances the CRS-induced thymus atrophy and that injection of a specific anti-OPN mAb (2C5) into WT mice ameliorates stress-induced thymus atrophy; changes in CORT levels were also partially reversed. This study reveals that OPN is one of the factors contributing to stress-induced organ atrophy and that it plays a significant role in immune cell survival/redistribution following chronic physical stress.

Our research has demonstrated that OPN is necessary for stress-induced corticosteroid up-regulation, possibly by affecting the production or metabolism of corticosteroid. We hypothesize this on the basis of the elevated basal level of CORT in OPN^−/− mice. The increased level of CORT in OPN-deficient mice could result from either augmented anabolism or diminished catabolism of CORT. The reduced ACTH levels in OPN^−/− mice and the absence of a response during CRS implies either that a high basal level of CORT represses ACTH release through negative feedback regulation of CORT or that OPN has a direct positive effect on ACTH production. A proposal for how OPN might regulate HPA axis hormones is depicted in Fig. 8. We hypothesize that OPN mediates an ACTH-dependent stimulation of CORT synthesis. Organ atrophy is caused by stress induced ACTH-CORT up-regulation, which, in an OPN-dependent manner, leads to cell apoptosis in the spleen and thymus (Fig. 8,A). In the absence of OPN (Fig. 8 B), the basal level of CORT is increased, which sends a feedback signal to reduce ACTH production. The increased CORT:ACTH ratio desensitizes the animal to CRS, resulting in reduced lymphoid organ atrophy in the OPN-deficient mouse. OPN deficiency resembles a stress state that leads to a persistent CORT up-regulation. When physical stress is applied, the HPA axis fails to respond to the new stress. Therefore, OPN is critical in regulating HPA axis hormones to balance their response to stress. These findings suggest that circulating OPN mediates stress responses possibly through regulating HPA hormone levels and, furthermore, reveals that OPN is an important link between the immune system and the endocrine system.

Stress sends signals to the brain to induce the cascading release of the stress hormones ACTH and CORT. The HPA axis is a major part of the neuroendocrine system that controls reactions to stress and regulates various physiological processes, including immune responses. The level of these hormones in the plasma changes in response to stress and each of them has been shown to interact closely with the immune system in a bidirectional manner (28). Production of cytokines can stimulate the release of glucocorticoids; in turn, HPA activation by cytokines has been found to play a critical role in restraining and shaping immune responses. Thus, cytokine-HPA interactions represent a fundamental mechanism of the maintenance of homeostasis and the development of disease during stress or infection (29). For example, HPA axis hormones play an important role in autoimmune diseases such as multiple sclerosis, adjuvant-induced arthritis, eosinophilia myalgia syndrome, and systemic lupus erythematosus (30, 31). Additionally, since OPN expression is highly perturbed in cancer, autoimmune, and inflammatory diseases, manipulating OPN levels may have a wide impact on overall health (16, 32, 33). By determining the role of lymphocyte death in a stressed immune system and the factors contributing to this process, we may be able to develop therapies to maintain a healthy immune system.

We thank Dana Cifelli, Lucy Lee, Megha Rajpal, Getta Denhardt, and David Shen for important contributions to various aspects of this research.

The authors have no financial conflict of interest.