Substance P activates responses correlated with tumour growth in human glioma cell lines bearing tachykinin NK1 receptors

The neuropeptide substance P (SP), by stimulating tachykinin NK1receptors (NK1R), triggers a number of biological responses in human glioma cells which are potentially relevant for tumour growth. First, radioligand binding studies demonstrated the presence of tachykinin NK1R on SNB-19, DBTRG-05 MG and U373 MG, but not on U138 MG and MOG-G-GCM human glioma cell lines. Second, application of SP or neurokinin A (NKA) to NK1R+glioma cell lines increased the secretion of interleukin 6 (IL-6) and potentiated IL-6 secretion induced by IL-1β. SP also up-regulated the release of transforming growth factor β1 (TGF-β1) by the U373 MG glioma cell line. Third, SP induced new DNA synthesis and enhanced the proliferation rate of NK1R+, but not of NK1R−glioma cell lines. Also, NKA stimulated the proliferation and cytokine secretion in NK1R+glioma cell lines. All the stimulant effects of SP/NKA on NK1R+glioma cell lines were completely blocked by a specific tachykinin NK1R antagonist, MEN 11467. These data support the potential use of tachykinin NK1R antagonist for controlling the proliferative rate of human gliomas. © 1999 Cancer Research Campaign

Substance P (SP), an undecapeptide of the tachykinin family of peptides, is the preferential endogenous ligand for the tachykinin NK 1 receptor (NK 1 R) (Maggi et al, 1993;Otsuka and Toshioka, 1993).
SP activates phospholipase C and stimulates the release of interleukin 6 (IL-6) and prostaglandin E 2 from human fetal astrocytes in culture (Palma et al, 1997), indicating an involvement of this neuropeptide in modulating astrocyte functions. Although there is little evidence for the expression of tachykinin NK 1 R by astrocytes in the normal adult brain (Maggi, 1997 for review), there is evidence for up-regulation of this receptor by reactive astrocytes: proliferating glial cells express high concentrations of NK 1 Rs after transection of the optic nerve (Mantyh et al, 1989). Moreover, SPimmunoreactive astrocytes were observed in multiple sclerosis plaques (Kostyk et al, 1989) and in the forebrains of human infants (Michel et al, 1986). Interestingly, high expression of SP receptors (Henning et al, 1995), as well as the presence of the SP itself, (Allen et al, 1985) has been described in human malignant gliomas such as astrocytomas and glioblastomas. These observations suggest a possible role of tachykinin NK 1 Rs in supporting the development and growth of human astrocytomas.
The human astrocytoma cell line U373 MG expresses high levels of tachykinin NK 1 R (Heuillet et al, 1993), and responds to applied SP by releasing taurine (Lee et al, 1992) and a panel of cytokines including IL-6, IL-8, leukaemia inhibitor factor (LIF) and granulocyte-macrophage colony-stimulating factor (GM-CSF) (Gitter et al, 1994;Palma et al, 1994Palma et al, , 1995Palma and Manzini, 1998). The release of cytokines by malignant glioma cells has been associated with glioma progression (Gillespie, 1996); for example, IL-6 has been detected in tumour cysts and cerebrospinal fluids in patients harbouring malignant gliomas (Weller et al, 1991;Frei et al, 1992).
In addition to modulating cytokine production, SP induces DNA synthesis and activates the mitogen-activated protein (MAP) kinase pathway in U373 MG cells . However, the actual ability of this neuropeptide to produce proliferation of human glioma cells has not been determined until now.
In this study, we have further investigated the possible role of the tachykinin NK 1 R in modulating growth/development of human astrocytomas by addressing: (a) the expression of highaffinity tachykinin NK 1 R in different human glioma cell lines; (b) the presence/absence of functional responses (IL-6 secretion, DNA synthesis) to applied tachykinins in glioma cells lines positive or negative for expression of the tachykinin NK 1 R (NK 1 R + and NK 1 Rrespectively); (c) the ability of SP to induce not only DNA synthesis but also actual proliferation of human glioma cells; (d) whether applied tachykinins induce the secretion of immunosuppressive cytokines (IL-10 and TGF-β1) from human gliomas; (e) the possibility that neurokinin A (NKA), which is often coexpressed with SP in the central nervous system, may affect glioma cells in the same way as SP does; and (f) the effect of a potent and selective antagonist for tachykinin NK 1 R on the responses of human glioma cells to SP.
In binding studies, MEN 11467 showed high affinity for NK 1 Rs expressed in human cell lines (K i = 0.4 ± 0.1 and K i = 1.34 ± 0.32 for IM-9 and U373 MG respectively) and virtually no affinity (K i ≥ 10 000 nM) for the NK 1 R present in rat urinary bladder membranes, a result more than 1000-fold selective for the human compared with the murine receptors. In addition, MEN 11467 has negligible effects on the binding of [ 125 I]NKA to hamster urinary bladder membranes (NK 2 receptor) or of [ 3 H]senktide to guinea pig cerebral cortex membranes (NK 3 receptor) (Cirillo et al, 1998).

Binding assays
Binding assays were conducted on confluent intact cells in 24-well plastic culture dishes, as described previously  with minor modifications. Briefly, cells were rinsed and 500 µl of RPMI-1640 supplemented with 0.2% glucose and 1% bovine serum albumin was added to each well for 30 min. The buffer was then aspirated and fresh buffer containing [ 3 H]SP (specific activity 40 Ci mmol -1 ; Amersham, Buckinghamshire, UK) or [ 3 H][Sar 9 ,Met(O 2 ) 11 ]SP (specific activity 40.6 Ci mmol -1 ; New England Nuclear, Du Pont de Nemours, NEN Division, Dreiech, Germany) was added in a final volume of 500 µl. The non-specific binding was defined as that displaceable by unlabelled 10 µM SP or [Sar 9 ,Met(O 2 ) 11 ]SP. The plates were incubated for 2 h at 4°C. The reaction was stopped by aspirating the medium and then rinsing each well three times with 1 ml of 0.9% sodium chloride in 10 mM HEPES (pH 7.2) wash buffer. The cells were solubilized in 5% sodium dodecyl sulphate in 0.01 M hydrochloride acid (0.5 ml) at 37°C for 1 h and radioactivity was quantified in a liquid scintillation spectrometer.
For saturation binding experiments, cells were incubated with increasing concentrations of [ 3 H]SP (0.1-5 nM) or [ 3 H][Sar 9 ,Met(O 2 ) 11 ]SP (0.1-5 nM). For competition binding experiments, cells were incubated with 2 nM of both radioligands, in the presence of varying amounts of competitor. Binding data were analysed using the iterative curve fitting program, Ligand (Munson and Rodbard, 1980).

DNA synthesis
Human glioma cells were plated on 24-well tissue culture plates (2×10 4 cells per well in 1 ml of medium, unless otherwise noted) in their appropriate media with low concentration or without FBS as indicated throughout the text. The stimulants were then added and left in the culture media for the entire length of the experiment. [ 3 H]methyl-thymidine [specific activity 82.5 Ci mmol -1 , NEN-Du Pont De Nemours Italiana, Cologno Monzese (MI), Italy] at a concentration of 1 µCi ml -1 was added to each cell culture for the last 24 h of incubation. At the end of each experimental point, the cells were treated with 1 ml of 5% cold trichloroacetic acid (TCA) for 10 min at 4°C. The acid was removed, and the cells were washed with double-distilled water (three times, 2 ml each time). The cells were removed from each well with 0.5 ml of 0.1 N sodium hydroxide and incubated for 15 min. Each sample was then neutralized with 0.1 N hydrochloric acid and the radioactivity was quantified in a liquid scintillation spectrometer.

Proliferation studies
Human glioma cells were plated on 24-well tissue culture plates (2×10 4 cells per well in 1 ml of medium, unless otherwise noted) in their appropriate media in serum-free conditions or with lowconcentration FBS as indicated. The stimulants were then added and left in the culture medium for the entire experiment length. At the end of each experimental point, the medium was aspirated and the living cells detached with 200 µl of trypsin. The cells were then diluted in the vital colorant trypan blue (1:1) and counted in haemocytometer counting chambers (Newbauer and Nageotte).

Detection of IL-6, IL-10 and TGF-β1 in supernatants of human glioma cell lines
Glioma cells were detached by trypsinization, washed in serumfree RPMI-1640, seeded in their appropriate FBS medium at 3.5×10 5 cell per well in 24-well culture plates, and allowed to adhere for 24 h at 37°C. Then the cultured medium was removed, the cells were washed three times in serum-free RPMI-1640, and fresh media containing low (1%) or high (10%) FBS were added. Cells were cultured for 18 h, unless otherwise indicated, in the presence or absence of SP, NKA, IL-1β. When necessary, the NK 1 receptor antagonists were co-administered with the stimulants. In all experimental conditions, cell viability was always greater than 99%, as determined by the Trypan Blue exclusion method. After the incubation time, the supernatants were collected and spun free of cells and debris. IL-6, IL-10 and TGF-β1 levels were assayed by specific ELISA kits (Quantikine, Research and Diagnostic System, Minneapolis, MN, USA or Biotrak, Amersham). To activate latent TGF-β1 to the immunoreactive form, 0.5 ml of cell culture supernatants were incubated for 10 min at room temperature with 0.1 ml 1 N hydrochloric acid. The acidified samples were then neutralized by adding 0.1 ml 1.2 N sodium hydroxide/0.5 M HEPES before being assayed. Significant levels of latent TGF-β1 were found in FBS, whereas no IL-6 or IL-10 was detected in FBS media even when FBS was present at 10%. The data were analysed using linear regression of the Microplate Manager Program (Bio-Rad Laboratories, Richmond, CA, USA), and simultaneous fitting of families of sigmoidal doseresponse curves using the four-parameter logistic equation with statistical analysis and elaboration of IC 50 and EC 50 values was performed using the program Allfit version 1.05 (De Lean et al, 1978).

Statistical analysis
All data in the text and figures are means ± s.e.m. Statistical analysis of data was performed with Anova one-way and Tukey tests.

Identification and characterization of SP binding sites (NK 1 R) on human glioma cell lines
By studying the specific binding of [ 3 H]SP, evidence for expression of tachykinin NK 1 R was obtained for SNB-19 and DBTRG-05 MG cell lines in addition to U373 MG cells. In these cell lines, the specific binding of [ 3 H]SP was usually >80% of total binding, as defined in the presence of 10 µM unlabelled SP. The binding of [ 3 H]SP was saturable and Scatchard analysis indicated a single population of high-affinity binding sites: the K d and B max values obtained in the three cell lines are reported in Figure 1 Palma et al, 1994Palma et al, , 1995Palma and Manzini, 1998) but also when cultured in low-concentration FBS (1%) (Figure 2). Even in the more stringent culture conditions, SP and NKA induced a concentration-dependent release of IL-6. The EC 50 values were 2 ± 0.4 and 3.8 ± 0.5 nM for SP in SNB-19 and U373 MG glioma cells, respectively; the corresponding EC 50 values for NKA were about tenfold higher than for SP (37 ± 9 and 26 ± 7 nM for U373 MG and SNB-19 cells, respectively). DBTRG-05 MG cells, which do not spontaneously secrete IL-6, did not respond with augmented IL-6 secretion in response to added SP (n = 3). However, if challenged with IL-1β (100 U ml -1 ), a secretion of IL-6 was established (686 ± 20 pg ml -1 ) from DBTRG-05 MG cells; in these conditions, SP produced a concentration-dependent potentiation of the response to IL-1β (739 ± 2, 893 ± 22, 1105 ± 26 and 1186 ± 15 pg ml -1 at 1, 10, 100 and 1000 nM of SP, respectively).
The involvement of the tachykinin NK 1 R in the stimulant action of SP on basal or stimulated (in DBTRG-05 MG cells) secretion of IL-6 was confirmed by studying the inhibitory action of the potent and selective tachykinin NK 1 R antagonist MEN 11467 (Figure 3). MEN 11467 produced a concentration-dependent and comparable inhibition of the responses to SP in the three NK 1 R + glioma cell lines On its own, MEN 11467 neither significantly affected the basal secretion of IL-6 from SNB-19 and U373 MG cell lines (486 ± 26 and 420 ± 49 pg ml -1 ; 900 ± 16 and 843 ± 5 pg ml -1 in the absence or presence of 100 nM MEN 11467 for SNB-19 and U373 MG cells respectively), nor the IL-6 secretion induced by IL-1β from DBTRG-05 MG cells (690 ± 17 pg ml -1 and 604 ± 11 pg ml -1 in the absence or presence of 100 nM MEN 11467, respectively).
To confirm the specificity of the stimulatory effects of SP, the effect of this peptide was also investigated in NK 1 Rglioma cell lines; up to 100 nM, SP neither induced IL-6 secretion nor affected IL-1β-stimulated IL-6 release from U138 MG or MOG-G-CCM cells (data not shown, n = 2 for each cell line).
In contrast, a spontaneous release of the inactive form of TGF-β1 was detected (1033 ± 92 pg ml -1 ) after 24 h of culture in the  presence of low-concentration FBS (1%). SP potentiated, in a concentration-dependent manner, the secretion of the inactive form of TGF-β1 ( Figure 4A). This response involves the activation of tachykinin NK 1 R because MEN 11467 (100 nM) completely blocked the SP (100 nM)-stimulated release of TGF-β1 ( Figure 4B). However, the presence of active TGF-β1 molecules was never observed in the supernatants of SP-triggered U373 MG cells. TGF-β1 release was not modulated by stimulation with LPS (20 ng ml -1 ) or IL-1β (100 U ml -1 ).

SP induces DNA synthesis and proliferation in NK 1 R + glioma cell lines
The aim of these experiments was to verify the ability of SP to stimulate growth of human glioma cell lines by measuring both DNA synthesis (assessed by evaluating thymidine uptake) and cell proliferation (by counting the number of cells at different times from exposure to SP). In a first series of experiments ( Figure 5), U373 MG cells were plated at low concentration (2×10 4 cells per well in 24-well tissue culture plates) in serum-free medium, and time course experiments were performed in the presence of SP. No significant increase in the number of cells was observed in control conditions ( Figure 5); SP (10-100 nM) increased DNA synthesis already after 24 h of stimulation, whereas the number of cells present in the wells at this time was similar for treated and untreated samples. However, the addition of SP determined a doubling of cell number after 48 h of culture ( Figure 5).
In another series of experiments (Figure 6), the ability of SP to stimulate the growth of U373 MG cells was determined in the presence of low-concentration serum (0.8%), which stimulates cell growth to some extent; as shown in Figure 6, a stimulant action of SP on cell number above the spontaneous growth was observed since 96 h from addition of the neuropeptide and a stimulant effect was evident up to 192 h ( Figure 6). In these conditions, the mitogenic effect of SP was not concentration-dependent (1-1000 nM); the proliferative response observed at 1 nM was similar to that obtained at the highest concentration (55 333 ± 8671, 94 000 ± 11 222 and 102 000 ± 6060 cells per well after 96 h of incubation in medium alone, SP 1 nM and SP 1000 nM, respectively).
A third series of experiments showed (Figure 7) that application of SP determines a mitogenic effect on other NK 1 R + human glioma cell lines, SNB-19 and DBTRG-05 MG cell lines. In contrast, the NK 1 Rcell lines U138 MG and MOG-G-CCM did not show any increase in DNA synthesis (after 48 h of stimulation, data not shown, n = 2 for each cell line) or proliferation after challenge with SP (55 000 ± 4582 and 61 666 ± 7023 cells per well; 81 000 ± 11 357 and 77 000 ± 4582 cells per well after 96 h of incubation in absence or in the presence of 100 nM SP for U138 MG and MOG-G-CCM, respectively).

An NK 1 R antagonist inhibits SP-induced growth in glioma cell lines
The tachykinin NK 1 R antagonist MEN 11467 (100 nM) completely reverted the SP-induced thymidine uptake by U373 MG cells, whereas that induced by IL-1β (10 and 40 U ml -1 ) was not significantly reduced ( Figure 8). Moreover, the enhanced proliferation of NK 1 R + glioma cell lines induced by SP was inhibited by MEN 11467 (Table 1).

DISCUSSION
The present findings expand previous observations Sharif et al, 1996) on the role played by tachykinins (via NK 1 R) in producing biological responses which are potentially relevant for the development and growth of human gliomas. In particular, the present study provides the following novel observations: (a) in addition to U373 MG cells, two glioma cell lines have been characterized as being NK 1 R + (SNB-19 and DBTRG-05 MG), whereas two other cell lines were found to be NK 1 R -(MOG-G-CCM and U138 MG); (b) the demonstration of the mitogenic properties of SP in terms not only of DNA synthesis but also as an actual increase in glioma cell number; (c) a positive correlation has been established between functional responses (modulation of IL-6 secretion and proliferation) of human glioma cell lines and the expression of NK 1 R; (d) in keeping with the above, the activation of human glioma cell lines by tachykinins was blocked by the highly potent and selective tachykinin NK 1 R antagonist MEN 11467; (e) in addition to SP, NKA also stimulates cytokine secretion and enhances proliferation in human glioma cell lines via the NK 1 R; (f) stimulation of the NK 1 R also determines the release of the immunosuppressant cytokine TGF-β1 from the U373 MG cell line.
Application of SP or of tachykinin NK 1 R selective agonists determines a number of functional responses in U373 MG glioma cells: the stimulation of phosphatidyl inositol (Pi) turnover, mobilization of intracellular calcium, release of taurine, activation of MAP kinase and secretion of cytokines (Eistetter et al, 1992;Lee et al, 1992;Gitter et al, 1994;Palma et al, 1994Palma et al, , 1995Palma and Manzini, 1998). Therefore, the activation of NK 1 R initiates a set of cellular responses which eventually includes stimulation of DNA synthesis, cell division and proliferation. The actual proliferation of human glioma cells detected in this study occurs both in serum-free conditions and in the presence of serum; the latter observation ( Figure 6) indicates that SP can cooperate with other growth factors in regulating the proliferation of NK 1 R + cells. In contrast, the former observation indicates that SP is in itself a sufficient stimulus for promoting glioma cell proliferation. Even in the most stringent serum-free conditions, an actual proliferative response to SP was evident only at > 24 h from its applications; therefore, the question arises as to whether the continuous occupancy of NK 1 R during this time lag is required for the triggering of the proliferative response, or, as an alternative, whether the initial stimulus provided by SP determines glioma cell proliferations indirectly through the release of other factors (e.g. IL-6, see below). As a matter of fact, SP application induces a lasting expression of κB-dependent genes (Lieb et al, 1998), such as expression of mRNA coding for the mitogenic cytokine IL-6 ). It appears, therefore, conceivable that the actual proliferation of glioma cells, as detected in this study, had involved some intermediate step in the sequence of events initiated by the occupancy of tachykinin NK 1 R. The secretion of cytokines such as IL-6 or TGF-β1 could provide an autocrine /paracrine loop for glioma progression. In fact, both IL-6 and TGF-β1 roles facilitate glioma cell proliferation (Jennings et al, 1991;Munoz-Fernandes and Fresno, 1993).
At the present time, we cannot totally exclude that, in addition to its role as growth promoter, SP may also have enhanced glioma cell survival and plating efficiency. However, the initial number of cells present in the wells was similar in the presence or absence of SP, suggesting that a putative effect of this neuropeptide on plating efficiency, if any, was a minor one. In contrast, further studies should be performed to clarify whether SP can increase the number of cells entering the cell cycle and/or induce an improved survival by preventing apoptosis, which is a common event in tumour cells lacking in growth factors.
The present findings demonstrate that, in addition to SP, NKA also stimulates the growth of human glioma cells via the NK 1 R. This observation is of particular interest because SP and NKA can be produced through the expression of the same gene in the CNS and are frequently co-localized and co-released by the same neurons (Maggi et al, 1993;Otsuka and Yoshioka, 1993 for reviews). NKA is the natural ligand with preferential affinity for tachykinin NK 2 receptors, but it can also act as a full agonist at NK 1 Rs (Maggi and Schwartz, 1997 for review). Indeed, because NK 2 receptors are poorly expressed, if not absent, in the mammalian CNS (Otsuka and Yoshioka, 1993), it is currently believed that SP and NKA act in concert by stimulating NK 1 R at most mammalian CNS synapses. Sharif et al (1996) reported that activation of the MAP kinase pathway and [ 3 H]thymidine incorporation into DNA were induced by NKA as efficiently as SP when fixed high equimolar concentrations of tachykinin were used (100 nM and 1 µM respectively). In our hands, NKA induced responses (secretion of IL-6, stimulation of DNA synthesis and cell proliferation) which were quantitatively comparable to those produced by SP. However, our data document that NKA has about a ten fold lower potency in stimulating IL-6 secretion from NK 1 R + human glioma cell lines than SP, in accordance with the lower affinity of NKA for NK 1 Rs, as shown in competition binding vs [ 3 H]SP in U373 MG cells (Heuillet et al, 1993) and in SNB-19 cells (our unpublished observation). Our findings provide unequivocal evidence that the actions of NKA on human gliomas are mediated via NK 1 R stimulation. In fact NKA was active in NK 1 R + but not in NK 1 Rcell lines, and its effects in NK 1 R + glioma cells were blocked by MEN 11467.
We also demonstrated that SP, via NK 1 R occupancy, stimulates the secretion of the immunodepressant cytokine TGF-β1. This finding is of particular interest because it provides an additional mechanism through which tachykinins may facilitate the growth and development of human gliomas. TGF-β1 and IL-6 contribute to the general immunodepression observed in glioma patients (Roszman et al, 1991;Ausiello et al, 1991). IL-6 inhibits the secretion of IL-1β and TNF-α and can counteract the activation of the immune system in this way (Schindler et al, 1990). TGF-β1 induces apoptosis of tumour-infiltrating lymphocytes, significantly reducing their cytotoxic properties, and depresses the activity of natural killer cells (Bodmer et al, 1989;Weller et al, 1995). Therefore, a stimulation by tachykinins of the local release of immunosuppressant cytokines could play a facilitatory role in the growth and development of gliomas by helping tumour cells to evade attack by the immune system.
High-grade human malignant glioma are inevitably lethal neoplasms, and the median survival of patients treated with standard cytoreductive surgery and post-operative radiotherapy is in the range of 1 year. The task of developing additional or novel treatments crucially requires an increase in knowledge of the physiology and molecular biology of these brain tumours. We have gathered evidence indicating that tachykinins (SP and NKA), through a specific stimulation of NK 1 R, have functional effects on several human glioma cell lines, resulting in proliferation, mitogenesis and release of soluble factors which can influence the tumour cell-host interactions including modulation of the immune system.
Altogether, these observations suggest a possible role of tachykinins, via NK 1 Rs in glioma development and growth. It should be noted that SP was found in human glioma tissues (Allen et al, 1985), and human astrocytes themselves are a possible source of tachykinins (Michel et al, 1986). In addition, an overexpression of NK 1 Rs seems to occur in gliomas in relation to their degree of malignancy (Henning et al, 1995). Therefore, an autocrine/paracrine loop providing a facilitatory input on tumour growth could be devised to exist in the glioma itself. In addition, the expression of NK 1 Rs in peritumoral and tumoral blood vessels (Henning et al, 1995) suggests a role of SP in facilitating tumour blood supply by virtue of its NK 1 R-mediated angiogenic and vasodilator properties (Ziche et al, 1990). Tachykinins are multifaceted factors for amplification of the malignant growth of human glioma cells. In view of the obligatory role of NK 1 Rs in all these effects of tachykinins, the pharmacological blockade of NK 1 Rs could represent a new strategy to slow down glioma tumour growth in humans.