IL-7 Up-Regulates TNF-α-Dependent Osteoclastogenesis in Patients Affected by Solid Tumor

Background Interleukin-7 (IL-7) is a potent regulator of lymphocyte development, which has also significant effects on bone; in fact it is a potent osteoclastogenic factor. Some human solid tumors produce high IL-7 levels, suggesting a potential IL-7 role on tumor development and progression. Methodology We studied 50 male patients affected by solid tumors, and their blood samples were collected at tumor diagnosis. PBMCs were isolated and cultured with/without IL-7 to study its influence on osteoclastogenesis. Serum and cell culture supernatant IL-7 levels were measured by ELISA. The quantitative analysis of IL-7 expression on T and B cells was performed by Real-Time PCR. Principal Findings Serum IL-7 levels were highest in osteolytic cancer patients, followed by cancer patients without bone lesions, and then healthy controls. We showed the IL-7 production in PBMC cultures and particularly in monocyte and B cell co-cultures. A quantitative analysis of IL-7 expression in T and B cells confirmed that B cells had a high IL-7 expression. In all cell culture conditions, IL-7 significantly increased osteoclastogenesis and an anti-IL-7 antibody inhibited it. We demonstrated that IL-7 supports OC formation by inducing the TNF-α production and low RANKL levels, which synergize in promoting osteoclastogenesis. Conclusions We demonstrated the presence of high serum IL-7 levels in patients with bone metastasis, suggesting the use of serum IL-7 level as a clinical marker of disease progression and of bone involvement. Moreover, we showed the capability of IL-7 to stimulate spontaneous osteoclastogenesis of bone metastatic patients and to induce osteoclastogenesis in cancer patients without bone involvement. These findings add further details to the disclosure of the mechanisms controlling bone metastasis in solid tumors.


INTRODUCTION
Interleukin-7 (IL-7) is a pleiotropic immune regulatory protein predominantly produced by stromal cells and by cells at the inflammatory sites [1]. IL-7 is fundamental for the early development of lymphocytes and is a regulator of peripheral T cell homeostasis by modulating the expansion of peripheral T-cell populations in states of T cell depletion [2]. The production of IL-7 by some human solid tumors suggests its potential impact on the process of tumorigenesis [3,4], but it is unclear how IL-7 is involved in solid tumor development and progression. IL-7 stimulates the progression of some types of lymphomas and leukaemias [5,6] and in non-small cell lung cancer, it was reported that IL-7 targeted gene therapy may be effective in modifying host anti-tumor responses [7]. Recent studies on breast cancer describe a quantitative association between the IL-7 signalling complex and some clinico-pathological parameters: there is a trend towards a higher expression of IL-7 and molecules of its signalling pathway in breast cancer patients with poor prognosis [4]. Moreover, IL-7/ IL-7R mRNA was detected in different tumors, such as colorectal [8], renal [9], lung and central nervous system cancers [10]. IL-7 is involved in the control of osteoclastogenesis, however depending on the model considered, it displays either an inhibitory or an activator effect on OCs [11], [12]. The OCs derive from cells of the monocytic-macrophagic lineage, which fuse to form bone-resorbing cells in the presence of Macrophage Colony Stimulating Factor (M-CSF) and Receptor Activator of Nuclear factor kB Ligand (RANKL) [13] or Tumor Necrosis Factora (TNF-a) [14,15]. Weitzmann et al. demonstrated that systemic administration of IL-7 stimulates osteoclastogenesis through T cells by RANKL and TNF-a [16]. In addition, they showed that TNF-a produced by T cells synergises with RANKL increasing bone destruction in murine models of estrogen deficiency [17,18].
The addition of exogenous cytokines is necessary to stimulate osteoclastogenesis in patients without osteolysis and in healthy controls. Previously, we demonstrated that peripheral blood mononuclear cells (PBMCs) from cancer patients with osteolysis differentiate spontaneously into OCs. In addition, we showed that in patients with osteolysis the activation of osteoclastic precursors depends on circulating tumor cells or on factors released from the tumor site, such as TNF-a [15].
To better identify the factors involved in the spontaneous osteoclastogenesis, present in bone metastatic patients, we focused our studies on IL-7, since it have been associated with haematological malignancies and inflammatory diseases characterized by a local and/or systemic bone loss [19][20][21]. IL-7 involvement in the bone metastasis formation and in the spontaneous osteoclastogenesis of cancer patients adds a further detail to the mechanisms of bone resorption and might be useful to design novel therapeutic approaches for treatment and prevention of bone metastasis.

RESULTS
IL-7 levels dosed in supernatants and sera are higher in cancer patients than in healthy controls We measured the IL-7 levels in 25 patient and healthy control PBMC supernatants, collected at days 5 and 10 of culture. We observed that IL-7 levels were significantly higher in patients than in healthy controls at day 5, p,0,0006 (Fig. 1A). At day 10, the IL-7 concentration was comparable to the one at day 5, both in patients and in healthy controls (data not shown). To identify the IL-7 source we performed co-colture experiments of monocytes plus B or T cells and we measured its levels in supernatants at day 5 and 10. We showed that IL-7 levels were significantly higher in cancer patients' co-coltures of both monocytes plus T or B cells than in healthy controls (p,0,0006) and remained nearly steady between day 5 and 10. In our model, the purified B cell population is responsible for IL-7 production: co-coltures of monocytes and B cells showed the highest IL-7 levels. Both in PBMC cultures and in monocytes plus T and B cells derived from patients without bone lesions we measured low levels of IL-7, comparable to healthy controls.
Besides, we tested serum levels of IL-7 in 50 patients affected by solid tumors with or without metastasis at the time of diagnosis. Fig. 1B showed that IL-7 serum levels were significantly higher in patients with bone metastasis compared to patients without bone lesions and healthy controls, p,0,0001. The mean IL-7 value was 17,6861,24 pg/ml for osteolytic patients, 10,7360,10 pg/ml for patients without bone lesion and 5,5660,30 pg/ml for healthy controls. We examined the potential differences in serum IL-7 levels according to the various types of tumor and specifically we considered prostate cancer, NSCLC and SCLC. We identified significant differences in the mean value of IL-7 (p#0,02), which were 19.8966.30 pg/ml for prostate cancer, 14.665.95 pg/ml for NSCLC and 7.5664.09 pg/ml for SCLC.

Evaluation of IL-7 expression in T and B cells with (RQ-PCR)
We analysed the expression of IL-7 on freshly purified T and B cells by RQ-PCR. The analyzed transcripts exhibited high linearity amplification plots (r.0.99) and the efficiencies of both PCR reactions were quite similar. As shown in Fig. 2, we did not measure significant differences in the expression of IL-7 in both patients and healthy controls' T cells, while in healthy controls' B cells the IL-7 expression was five-fold lower than in patients. The relative expression of IL-7 was seven-fold higher in B cells than in T cells in cancer patients. We did not appreciate any significant differences in IL-7 expression in cancer patients with or without bone metastasis (Fig. 2).

Effects of IL-7 on osteoclastogenesis
The OC differentiation was verified by the presence of multinucleated/TRAP positive cells from cancer patient and healthy control PBMCs (Fig. 3A-B lines, respectively). PBMCs were cultured at different doses of IL-7, that resulted in significantly increased osteoclastogenesis with 2,5 and 10 ng/ml of IL-7 in osteolytic patients and in osteoclastogenesis stimulation for patients  without bone lesions and healthy controls, p,0,0001. At 15 ng/ ml, IL-7 had not a stimulatory effect on osteoclastogenesis (Fig. 3C). By adding a neutralizing anti-IL-7 antibody on osteolytic patients' PBMC, we observed a dose-dependent inhibition of spontaneous osteoclastogenesis, p,0,01 (Fig. 3D). We characterized OCs also for the expression of a V b 3 , a typical OC marker. Both OCs derived from culture with and without IL-7 expressed a V b 3 , as shown in Fig. 3E.
The ability of IL-7 to induce OC formation was tested in cultures of purified monocytes alone and in co-cultures with purified T or B cells treated or untreated with IL-7 at 2,5 and 15 ng/ml. Our results showed that monocytes alone were not able to support significant OC formation in either the presence or absence of IL-7, in both cancer patients and healthy controls (data not shown). Moreover, we demonstrated that co-cultures of monocytes and T cells both from bone metastatic patients and healthy controls resulted in a significant induction of OC formation by adding 2,5 ng/ml compared to basal condition, p , 0,001, while the dose of 15 ng/ml did not elicit a significant stimulatory effect on osteoclastogenesis (Fig. 4A). Co-cultures of monocytes and B cells in presence of IL-7 showed an increase in OC number, but it was not statistically significant (Fig. 4B). In all culture conditions, we observed a comparable osteoclastogenesis in patients without bone metastases and in healthy controls (data not shown).
Since we showed that spontaneous osteoclastogenesis depends on contact between monocytes and lymphocytes [15], we investigated whether T and B cells were necessary for IL-7dependent OC differentiation. We cultured purified monocytes with IL-7-treated conditioned medium from T and B cells and we did not observe OC differentiation (data not shown). Thus, cell-tocell interaction between lymphocytes and monocytes are required for solid tumor dependent osteoclastogenesis.   . IL-7 increases OC bone resorbing activity. PBMC resorption activity with/without IL-7 were showed for bone metastatic patients (line A) and healthy controls (line B). The basal resorbing activity of OCs derived from patients was significantly higher compared to healthy controls, # p#0,0001 (C). The percentage of resorption area was significantly increased with 2,5 ng/ml of IL-7, both in patients and in healthy controls, *p#0,002 (C). At 15 ng/ml IL-7 did not elicit a significant variation of resorption area. The data represent the means6SE from ten independent experiments. doi:10.1371/journal.pone.0000124.g005

IL-7 stimulates OC resorbing activity
Since we showed that IL-7 had effects in stimulating osteoclastogenesis, we verified whether there was a similar stimulatory action on bone resorbing activity. By analysing the resorption area on biocoated slices, we demonstrated that in presence of IL-7 at 2,5 ng/ml the OCs increased the bone resorbing activity of osteolytic patients and stimulated the healthy controls' resorption ( Fig. 5A-B lines, respectively). The basal resorption activity of OCs derived from patients was significantly higher compared to healthy controls, # p#0,0001 (Fig. 5C). After stimulation with IL-7 at 2,5 ng/ml, the bone resorbing activity of OCs was significantly augmented both in patients and in healthy controls, *p#0,002 (Fig. 5C). At 15 ng/ml of IL-7, we did not observe any statistically significant variation on bone resorbing areas.

IL-7 affects TNF-a and RANKL release in culture
We assessed the IL-7 ability to induce TNF-a and RANKL secretion in PBMC cultures and monocyte plus T or B cell cocoltures at day 5 and 10. The basal levels of TNF-a were similar in PBMC media at day 5 and 10 and they significantly increased by adding 2,5 ng/ml of IL-7, p,0,002 (Fig. 6A). Co-cultures of monocytes and T or B cells showed basal TNF-a levels higher at day 10 than at day 5. Monocytes plus T cells displayed an IL-7 dose-dependent increase of TNF-a level both on days 5 and 10, p,0,01 (Fig. 6B). At 15 ng/ml of IL-7, TNF-a values reached a plateau level both in PBMC cultures and in monoytes plus T cell co-coltures (Fig. 6A-B). At day 5, co-coltures of monocytes and B cells showed a significant enhancement of TNF-a amount at 15 ng/ml, whereas at day 10 we measured a significant increase in TNF-a levels at 2.5 ng/ml, that was unchanged at 15 ng/ml of IL-7, p#0,01 (Fig. 6C).
We found low RANKL levels in culture media collected from PBMC and monocyte plus T or B co-coltures. In particular, in the IL-7 treated cells there were not significant differences compared to unstimulated controls (Table 1). To further study the involvement of TNF-a and RANKL we cultured cancer patients' PBMC in presence of IL-7 and a neutralizing anti-TNF-a antibody or osteoprotegerin (OPG) in different concentrations. Both the inhibitors determined a dose-dependent inhibition of bone resorbing activity (data not shown) and of osteoclastogenesis, that was stronger in cultures treated with anti-TNF-a (Fig. 7A) than with OPG (Fig. 7B). Thus, our data demonstrated that IL-7 supports osteoclastogenesis by inducing mainly the TNF-a production and low levels of RANKL.

DISCUSSION
Over the last years, many studies investigated the cross-talk between immune system and bone, and IL-7 appeared as an important factor in these interactions [12,22,23]. IL-7 is known to be a powerful inducer of T and B lymphopoiesis, moreover many studies link IL-7 to inflammatory diseases [21,24,25] and bone loss in estrogen deficiency conditions [22,26]. Our study focused on IL-7 involvement in the spontaneous osteoclastogenesis occurring in cancer patients with osteolytic metastasis. We showed significantly higher IL-7 levels in supernatants derived from PBMC cultures and co-coltures of monocytes and T or B cells of cancer patients compared to healthy controls. Our data describe for the first time the presence of elevated IL-7 levels in sera of male patients affected by solid tumors, in fact the only data available in literature, concerning IL-7 serum levels in patients affected by solid tumors are related to gynecologic cancers [27][28][29], while elevated serum IL-7 levels were previously detected in other pathologies with bony involvement, such as multiple myeloma Figure 6. IL-7 up-regulates TNF-a production in cell culture media. PBMC and monocytes plus T or B cells were cultured in absence or presence of IL-7. Supernatants were collected at days 5 and 10 to dose TNF-a. PBMCs showed the same TNF-a basal level, which increased at 2,5 ng/ml of IL-7 and reached the maximum value at 15 ng/ml, (A).

Monocytes plus T cells (M+T) (B) and monocytes plus B cells (M+B) (C)
had a TNF-a basal level higher at day 10 than at day 5, but both supernatants showed the peak value at 2,5 ng/ml of IL-7, and remained roughly costant at 15 ng/ml. The data were means6SE of 25 supernatants collected at day 5 and 10 for both patients and healthy controls. doi:10.1371/journal.pone.0000124.g006 [19], rheumatoid arthritis [30] and systemic juvenile rheumatoid arthritis [31]. It is interesting to note that patients affected by bone metastatic cancer had an IL-7 value higher than non-bone metastatic ones, and these latter showed an intermediate IL-7 serum value between bone metastatic patients and healthy controls. The high IL-7 serum levels might be interpreted as a signal of disease progression towards the metastatic phenotype in some non-bone metastatic tumor patients. The analysis of the available data showed that NSCLC and prostate cancer had a statistically significant higher IL-7 serum levels compared to SCLC cases. However, in all the specific cancer groups (NSCLC, prostate cancer and SCLC) we found that patients with bone metastasis had higher IL-7 levels than patients without osteolysis. These findings might infer that in different tumor types the increase of IL-7 serum levels might be useful for early diagnosis of metastatic bone lesions. Literature data support the IL-7 production by different tumor types [4,8,10] apart from stromal cells in bone microenvironment. Previously reported data showing the role of B cells in IL-7 production [6,19,23,32], were confirmed by our finding that B cells produced IL-7 and this production was higher than T cell one.
In this study we demonstrated that IL-7 at 2,5 ng/ml increased osteoclastogenesis in patients' cultures of PBMCs and co-coltures of monocytes plus T or B cells, while at the dose of 15 ng/ml, IL-7 did not affect osteoclastogenesis. To explain this result we hypotesize that, at high doses, IL-7 stimulates a differentiation switch from OCs towards dendritic cells since in cultures with 15 ng/ml of IL-7 we observed only few OCs and mainly undifferentiated macrophages or cells with dendritic cells morphology. This hypothesis is supported by some studies showing that IL-7 regulates dendritic cells differentiation from PBMC cultures [33,34]. Since literature data link IL-7 to TNF-a and RANKL-dependent osteoclastogenesis [16,19,35], we measured their levels in culture media. In detail, we showed that IL-7 enhances in a dose-dependent manner TNF-a levels in PBMC culture media and in co-coltures of monocytes and T or B cells. We hypothesize that TNF-a levels reached maximal values in the PBMC cultures probably due to a cooperative release of TNF-a by monocytes [21], T and B cells stimulated by IL-7. We observed a little and not statistically significant dose-dependent increase of RANKL release from day 5 to day 10 of culture compared to unstimulated condition. According to literature data [36,37] we think that these low levels of RANKL are synergistic with TNF-a in promoting osteoclastogenesis. In fact we observed that both anti-TNF-a antibody and OPG caused a dose-dependent inhibition of IL-7-induced osteoclastogenesis. The anti-TNF-a antibody reached an OC inhibition plateau at 90% while OPG treatment of cell cultures reached a plateau of OC differentiation inhibition at 50%. Further increase in OPG concentrations, in an attempt to reach an higher level of inhibition, was found to be toxic for the cultured cells. This observation suggests that in our model TNF-a might have a stronger role in OC formation than RANKL; it is conceivable that both factors may synergize in OC differentiation.
These last results demonstrate that TNF-a plays a fundamental role in IL-7-induced osteoclastogenesis in patients affected by solid tumor. Moreover, the pivotal role exerted by TNF-a in our systems was also supported by the fact that in freshly isolated T and B cells from patients we did not find RANKL expression, while TNF-a was strongly present. TNF-a enhancement by IL-7 is also present in other pathological conditions, characterized by a local and/or systemic bone loss, such as rheumatoid arthritis [30] or postmenopausal osteoporosis [38].
In conclusion, we propose a new physiopathologic mechanism with a key role of IL-7 in the formation of solid tumor bone metastasis. We demonstrated that IL-7, produced mainly by B cells in cell culture, directly sensitizes T cells to produce proosteoclastogenic factors, such as TNF-a and RANKL, and enhances spontaneous osteoclastogenesis. It could be useful to analyse a large number of patients for each different tumor type, in order to set an IL-7 cut off value, representative of a warning threshold for clinicians. This could allow to follow the bone metastatic disease both at diagnosis and during treatments.

Patients
Samples from peripheral blood (PB) were obtained from 50 male patients affected by solid tumors and 50 healthy controls, matched for age and sex. The main clinical characteristics of patients, aged from 50 to 84 years (median 65611,01 SD), are listed in Table 2. The patients' samples were collected at diagnosis to exclude biases, including recent chemo/hormono-therapy and pre-existing or non-related bone pathologies. We decided not to include breast cancer patients to avoid bias deriving from women treated with steroids or hormonal therapies.
Healthy controls showed a normal bone metabolism (evaluated by bone densitometry) and were not under medications, such as steroids that could increase osteoclastogenesis. Informed consent from patients and healthy controls was obtained to comply with institutional policies.

Cell cultures
PBMCs, collected from cancer patients and healthy controls, were isolated after centrifugation over a density gradient, Lymphoprep (Nycomed Pharma, Norway) and cultured in a-MEM, supplemented with 10% FBS, penicillin 100 U/ml and streptomycin 100 mg/m (Cambrex, Walkersville MD). To obtain fully differentiated human OCs, PBMCs were cultured with/without rhIL-7 at different doses (0,5-1-2,5-10 and 15 ng/ml), R&D Systems (Abingdon, UK). Culture supernatants were collected on days 5 and 10, when medium was either refreshed and supplemented or not with IL-7. For some experiments, patient PBMCs were cultured with different doses of anti-IL7 (50-100-300 ng/ml), R&D Systems (Abingdon, UK). Cultures were stopped after 15 days, mature OCs were identified as multinucleated cells containing three or more nuclei and positive for the expression of TRAP (Tartrate-resistant acid phosphatase, Sigma Aldrich, St. Louis MO) and a V b 3 (vitronectin receptor) kindly provided by Prof. G. Tarone. To evaluate bone resorbing activity, for 21 days PBMCs were cultured on Biocoat Osteologic bone cell culture system (BD Biosciences Bedford, MA) with/without IL-7, then cells were removed and resorption lacunae were identified by light microscopy.

Isolation of monocytes, T and B cells
Monocytes, T and B cells were isolated from PBMCs by an immunomagnetic selection using MACs microbeads (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany), according to the manufacturer's instructions. The purity of monocytes, T and B cells, obtained after separation, was 95-98%, as assessed by flow cytometry. Isolated cells were plated according to the human physiological concentrations: 10% monocytes and 30% lymphocytes (70% for T and 30% for B) with/without IL-7.

ELISA (Enzyme-Linked Immunosorbent Assay)
The amount of IL-7 in cell culture supernatant and serum was determined by ELISA kit (sensitivity was 0,156 to 20 pmol/L and 0 to 16 pmol/L, respectively. R&D Systems, Abingdon UK). TNF-a (Biosource, Nivelles Belgium) and RANKL (Biomedica Gruppe, Wien Austria) ELISA kit were performed on culture supernatants, the sensitivity range was 0 to 1500 pmol/L and 0 to 8 pmol/L, respectively.

Quantitative analysis of IL-7 gene expression
Total RNA was extracted from T and B cells and the first-strand cDNA synthesis was performed as previously described [15]. Quantitative analysis of IL-7 was performed with Real-Time Quantitative PCR (RQ-PCR) using b-actin as housekeeping control. RT-PCR was carried out using the iCycler iQ TM system (Bio Rad, CA, USA). PCR primers and TaqMan probes were designed using Primer Express v1.0 software and synthesized by Applied Biosystems (Warrington, UK). TaqMan probe specific for  bone  61  T2N2M1  IV   22  NSCLC  bone, liver  50  T2N3M1  IV   23  NSCLC  adrenal gland  71  T2N3M1  IV   24  NSCLC  bone  57  T2N3M1  IV   25  NSCLC  none  53  T3NXM0  IIIA-B   26  NSCLC  none  56  T3N2M0  IIIA   27  NSCLC  none  81  pT3N2M0  IIIA   28  NSCLC  lung  76  T3N2M1  IV   29  NSCLC  bone  71  T3N2M1  IV   30  NSCLC  bone  66  T3N3M1  IV   31  NSCLC  adrenal gland, brain 54  T3N3M1  IV   32  NSCLC  none  75  T4N0M0  IIIB   33  NSCLC  brain, bone  59  T4N3M1  IV   34  NSCLC  surrenal  45  T4N3M1  IV   35  NSCLC  bone  77  T4N3M1  IV   36  NSCLC  liver  77  T4N3M1  IV   37  NSCLC  brain  63  T4N3M1  IV   38  prostate bone  62  T3N0M1  IV   39  prostate bone  84  pT3bN1M1  IV   40  prostate bladder  71  T3N2M1  IV   41  prostate bone  64  T3N2M1  IV   42  prostate none  68  T4N0M0  IIIB   43  prostate bone  66  T4N0M1  IV   44  prostate bone  64  T4N0M1  IV   45  colon  bone  59  pT3N1M1  IV   46  colon  bone  66  pT4N0M1  IV   47  bladder bone  86  pT3NXM1  IV   48  bladder bone  75  pT3bN2M1  IV   49  pancreas multiple  36  T4N2M1  IV   50  kidney  bone  74  We tested a series of cell lines for the IL-7 expression and we chose , as positive control, MG63, an osteosarcoma cell line that expressed at high level IL-7. As negative control we used H522, a non-small cell lung cancer cell line. Relative IL-7 quantization in T and B lymphocytes, expressed as -fold variation over control (H522), was calculated after determination of the difference between C T (threshold cycle) of the given gene A (IL-7) and that of the calibrator gene B (b-actin) using the 2 2 DD CT method [39]. C T values are means of duplicate measurements. To validate the use of the 2 2DDCT method, serial dilutions (ranging from 10 0 to 10 25 ) of MG63 cDNA were used to obtained a standard curve.

Statistical analyses
Statistical analyses were performed with the Statistical Package for the Social Sciences (spssx/pc) software (SPSS, Chicago, IL). We compared the results by means of student's paired t test. The results were considered statistically significant for p,0,05.