The function neutralizing anti-TrkA antibody MNAC13 reduces ... - PNAS

1 downloads 0 Views 835KB Size Report
Feb 20, 2007 - McMahon SB (1996) Philos Trans R Soc London B 351:431–440. ... Hefti FF, Rosenthal A, Walicke PA, Wyatt S, Vergara G, Shelton DL, Davies ...
The function neutralizing anti-TrkA antibody MNAC13 reduces inflammatory and neuropathic pain Gabriele Ugolini*, Sara Marinelli*†, Sonia Covaceuszach*, Antonino Cattaneo*‡, and Flaminia Pavone†§ *Lay Line Genomics, Via di Castel Romano 100, 00128 Rome, Italy; †Consiglio Nazionale delle Ricerche, Institute of Neuroscience, Psychobiology, and Psychopharmacology, Via del Fosso di Fiorano 64, 00143 Rome, Italy; and ‡European Brain Research Institute (EBRI), Via del Fosso di Fiorano 64, 00143 Rome, Italy

Nerve growth factor (NGF) is involved in pain transduction mechanisms and plays a key role in many persistent pain states, notably those associated with inflammation. On this basis, both the NGF ligand and its receptor TrkA (tyrosine kinase A) represent an eligible target for pain therapy. Although the direct involvement of NGF in pain modulation is well established, the effect of a direct functional block of the TrkA receptor is still unknown. In this study, we have demonstrated that MNAC13, the only anti-TrkA monoclonal antibody for which function neutralizing properties have been clearly shown both in vitro and in vivo, induces analgesia in both inflammatory and neuropathic pain models, with a surprisingly long-lasting effect in the latter. The formalin-evoked pain licking responses are significantly reduced by the MNAC13 antibody in CD1 mice. Remarkably, treatment with the anti-TrkA antibody also produces a significant antiallodynic effect on neuropathic pain: repeated i.p. injections of MNAC13 induce significant functional recovery in mice subjected to sciatic nerve ligation, with effects persisting after administration. Furthermore, a clear synergistic effect is observed when MNAC13 is administered in combination with opioids, at doses that are not efficacious per se. This study represents a direct demonstration that neutralizing antibodies directed against the TrkA receptor may display potent analgesic effects in inflammatory and chronic pain. analgesia 兩 behavior 兩 mice 兩 nerve growth factor

P

ersistent pain represents a major health problem (1). The area of chronic and inflammatory pain is very fragmented and diversified, because of the fact that pain can be assessed at various levels of severity and is associated with various pathologies, implying different origins and mechanisms of action. Treatment depends on type and severity of pain. For example, mild inflammatory pain is conventionally treated with nonsteroidal antiinflammatory drugs (NSAIDs), whereas more severe pain is treated with opioids. All of the therapeutic approaches in use show several drawbacks (1). For instance, NSAIDs are associated with gastrointestinal and renal morbidity. Opioids appear to be quite effective in fighting cancer pain; nevertheless, because of their central mechanism of action, they are associated with several side effects and with tolerance development, hence limiting their use. An urgent need therefore exists to develop more specifically efficacious drugs directed against new molecular targets, with particular emphasis on the therapeutic area of chronic and inflammatory pain. Nerve growth factor (NGF) is a multifunctional molecule (2, 3) that exerts its biological functions in a variety of neural and nonneural cells, by means of two types of receptors (4): the TrkA (tyrosine kinase A) receptor (5–7) and the p75 neurotrophin receptor (8, 9), a member of the molecular family of tumor necrosis factor receptors. NGF plays a key role in pain-transduction mechanisms in the adult nervous system (10). Peripheral nociceptors strongly express the TrkA and p75 receptors, developmentally and functionally depending on NGF (10, 11). NGF is a peripherally produced mediator of several persistent pain states, notably those associated with inflammation (12), and NGF treatment in neonatal or mature www.pnas.org兾cgi兾doi兾10.1073兾pnas.0611253104

animals can lead to behavioral hyperalgesia (13). In inflammatory pain, a central role is also played by mast cells, which are recruited by NGF to the injured or painful site, and are induced by NGF to release inflammatory mediators (14–16). Inflammation-related pain can be significantly reduced by neutralizing NGF bioactivity in animal models (17–19), implying that an enhanced level of this neurotrophin is necessary to generate the full hyperalgesic response. Remarkably, the inhibition of other neurotrophins does not result in antagonizing the induced hyperalgesia, suggesting that this effect is specific for NGF (18). The involvement of NGF in the evolution of chronic neuropathic pain is much less defined, even if some clinical studies support its involvement in chronic pain states, showing that s.c. injection of NGF into the forearm and masseter muscle of healthy volunteers produces allodynia and hypersensitivity in the surrounding skin (20, 21). However, the role of NGF in chronic pain remains controversial. On one hand, NGF exerts neuroprotective and trophic action on peptidergic small-diameter dorsal root ganglia cells, reversing many of the histological changes after nerve injury (4, 22), and regulates the collateral sprouting of intact nociceptive sensory axons after denervation (23). On the other hand, peripheral neuropathic pain is correlated with elevated NGF levels (24, 25), and direct administration of NGF into the sciatic nerve produces hyperalgesia (26) Moreover, NGF function inhibition results in analgesia in different neuropathy-related pain protocols, such as chronic constriction injury and partial sciatic nerve transection (27, 28). Thus, NGF and its receptors emerge as excellent molecular targets for pain treatment, and inhibition of NGF signaling identifies a previously uncharacterized class of pain drugs (29). Although the direct involvement of NGF in the cascade of events occurring during chronic and inflammatory pain is rather well established, no study has been performed so far to directly assess the effect of blocking the TrkA receptor, rather than scavenging the NGF ligand. The functional neutralization of the TrkA receptor is not obviously equivalent to the block of NGF, because of the complexity of the two-receptor system mediating NGF activity (4), implying that the outcome of NGF action depends on the fine balance in the signaling from the two receptors. To address this question, a tool is needed that is specifically able to disrupt TrkA signaling in response to NGF. MNAC13 is a well characterized Author contributions: G.U. and S.M. contributed equally to this work; G.U., S.M., A.C., and F.P. designed research; G.U., S.M., S.C., and F.P. performed research; S.M., S.C., and F.P. analyzed data; and G.U., A.C., and F.P. wrote the paper. Conflict of interest statement: A.C. is president and shareholder of Lay Line Genomics, a biotech company that is involved in the development of therapeutics for neurological diseases based on recombinant antibodies and neurotrophic factors. G.U. and S.C. are employed by and S.M. is a postdoctoral fellow in the same biotech company, as shown in the affiliation. Abbreviations: CCI, chronic constriction injury; NGF, nerve growth factor; TrkA, tyrosine kinase A. §To

whom correspondence should be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/cgi/content/full/ 0611253104/DC1. © 2007 by The National Academy of Sciences of the USA

PNAS 兩 February 20, 2007 兩 vol. 104 兩 no. 8 兩 2985–2990

NEUROSCIENCE

Communicated by Rita Levi-Montalcini, European Brain Research Institute, Rome, Italy, December 18, 2006 (received for review June 21, 2006)

Fig. 1. MNAC13 specifically binds TrkA and leads to inhibition of NGF-dependent signaling. (A) The MNAC13 antibody (here represented as monovalent antigen binding fragment) binds to TrkA extracellular domain (ECD), preventing TrkA from being activated by its natural ligand NGF. (B) MNAC13 does not bind to the p75 receptor: Surface Plasmon Resonance experiments were performed by using a BIACore instrument (BIACore, Uppsala, Sweden). (C) Western blot results showing that MNAC13 leads to reduction of phospho-TrkA levels after NGF stimulation in mouse 3T3 cells overexpressing TrkA. (D) Western blot results showing that MNAC13 leads to inhibition of NGF-dependent phosphorylation of the PLC␥1 adaptor protein in mouse 3T3 cells overexpressing TrkA. In C and D, cells were incubated for 60 min at 37°C in the presence of MNAC13 monoclonal antibody or irrelevant monoclonal antibody SV5 (mock), or no monoclonal antibody (No Mab). Antibodies were used at two different concentrations: 300 ␮g/ml (h) and 200 ␮g/ml (l). NGF (⫹) at 100 ng per milliliter was added for 10 min at 37°C after preincubation with monoclonal antibodies. In the negative control (⫺), NGF treatment was omitted. MW, molecular mass, in kilodaltons.

anti-TrkA monoclonal antibody with remarkable functionneutralizing properties (30, 31) that can be exploited to discriminate the specific contribution of TrkA signaling. To investigate the analgesic potential of blocking NGFmediated activation of the TrkA receptor, the effects of MNAC13 administration were evaluated in a typical model of inflammation-derived persistent pain (formalin-evoked pain) (32) and in the chronic constriction injury (CCI) model of neuropathic pain (33). In addition, because a functional interaction between NGF and opioid system was reported in experimental models of inflammatory pain (34, 35), the possible synergistic effects of the anti-TrkA antibody with opiates, such as morphine and fentanyl, were investigated. Results MNAC13 Specifically Binds TrkA and Leads to Inhibition of NGFDependent Signaling. The anti-TrkA mouse monoclonal antibody

MNAC13 was initially selected because of its ability to prevent NGF binding to TrkA, by directly binding to the receptor and because it has been shown to inhibit TrkA bioactivity both in vitro and in vivo (30, 36). In accordance with previous results, a new BIACore study showed that MNAC13 (Fig. 1A) did not display any binding activity toward the p75 receptor, being absolutely specific for TrkA (Fig. 1B). In a phosphorylation assay performed on mouse 3T3 cells overexpressing TrkA, the binding of MNAC13 to TrkA did indeed result in a significant reduction of the amount of receptor phosphorylated at Tyr-490, the binding site for the SHC adaptor (7) (Fig. 1C), and at Tyr-794 (data not shown), the binding site for PLC␥ (37). Consistently, downstream TrkA signaling was impaired after MNAC13 treatment: in particular, a significant decrease was observed in the level of phosphorylation of PLC␥1 (Fig. 1D). 2986 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0611253104

MNAC13 Induces Analgesic Effects on Formalin-Induced Pain. Formalin injection resulted in the typical biphasic response, showing the highest peak after 5 min and a second phase of licking that started 15 min after the treatment. Fig. 2A shows the two phases of formalin-evoked licking response in mice after s.c. administration into the right hindpaw (s.c.injected) of either saline or mouse-irrelevant IgGs 18 h before the beginning of the test. No significant differences were observed among treatments (t15 ⫽ ⫺0.415, P ⫽ 0.6841 for the early phase and t15 ⫽ 0.667, P ⫽ 0.5152 for the late phase), demonstrating that IgG-injected mice could be used as control groups. Preliminary experiments in which both the whole monoclonal antibody format (Mab) and the fragment antigen-binding format (Fab) of MNAC13 were administered either 1 or 18 h before formalin test, showed a higher efficacy of MNAC13 Mab when injected 18 h before testing [see supporting information (SI) Fig. 5 and SI Text]. Therefore, we chose the latter time interval for our experiments. Our hypothesis is that when a short time elapses between MNAC13 administration and formalin test, the analgesic effects could be at least partially masked by a potentially proinflammatory action exerted on immune cell receptors by the Fc portion of the whole monoclonal antibody. Alternatively, this result could suggest a predominantly central (spinal cord) site of action for the antibodies, which have time to diffuse systemically in the 18-hour period. MNAC13 s.c.-injected induced analgesic effects in the late phase of formalin test. Several doses of the antibody significantly decreased the licking response (F7,66 ⫽ 5.806; P ⬍ 0.0001), starting from the very low dose of 1.8 ␮g per mouse and reaching the maximal effect at 15 ␮g per mouse (halving of the licking response). As far as the early phase is concerned, only the dose of 15 ␮g per mouse resulted in a significant analgesic effect (F7,66 ⫽ 3.439; P ⬍ 0.01) (Fig. 2B). Because the early phase of the formalin response is evoked by Ugolini et al.

direct formalin stimulation of peripheral nerve endings, whereas the late one is due to subsequent inflammation, MNAC13 appears to be more specifically effective on the inflammatory component of formalin-induced pain (32). When the same dose, i.e., 15 ␮g per mouse of MNAC13, was i.p. injected, it produced similar effects on licking behavior, showing that the antibody resulted analgesic (F1,32 ⫽ 13.66; P ⬍ 0.001 and F1,32 ⫽ 21.657; P ⬍ 0.0001, for the early and late phases, respectively) regardless of the route of administration (F1,32 ⫽ 1.094; P ⫽ 0.3035 and F1,32 ⫽ 0.309; P ⫽ 0.5824, for the early and late phases, respectively) (Fig. 2C). MNAC13 Blocks Mechanical Allodynia in a Mouse Model of Neuropathic Pain. Besides inducing analgesia in a relevant model of

inflammatory pain, repeated i.p. injections of MNAC13 resulted in a significant reduction of mechanical allodynia induced by CCI of the sciatic nerve. Three days after ligature and before the beginning of the anti-TrkA treatment, animals were tested for their mechanical threshold of the hindpaw ipsilateral to the lesion (D3). In a first set of experiments (Fig. 3A), we demonstrated that four i.p. injections of MNAC13 (50 ␮g per mouse, from day 3 to day 6) were able to significantly reduce mechanical allodynia (treatment: F1,10 ⫽ 12.203, P ⬍ 0.01; time: F6,60 ⫽ 3.211, P ⬍ 0.01; treatment ⫻ time F6,60 ⫽ 5.090, P ⬍ 0.001), starting from the fifth day after surgery. On this basis, a second set of experiments was designed, in which mice were injected with two different doses of MNAC13 (either 30 or 70 ␮g per mouse repeated i.p. injections from day 3 to day 10) and observed for a longer period (up to day 31). Starting from day 4, a significant dose-dependent enhancement of the mechanical threshold was observed after MNAC13 treatment (treatment: F2,28 ⫽ 31.452, P ⬍ 0.0001; time: F15,420 ⫽ 13.324, P ⬍ 0.0001; treatment ⫻ time: F30,420 ⫽ 4.417, P ⬍ 0.01). In particular, the effect was maintained throughout the observation period, with a peak after the last antibody administration (day 11). Surprisingly, after a gradual decline of the antiallodynic effect, reaching a minimum at day 17, i.e., 1 week after the end of the treatment, MNAC13 again reduced neuropathic pain to a higher extent, from day 21 up to day 31 (as shown by Fig. 3B). MNAC13 Enhances the Analgesic Effects of Opiates. Because of the

significant side effects and tolerance development observed with opioid-based pain therapies greatly limiting their use, it was of interest to assess whether MNAC13 was able to reduce the effective doses of opioids in animal models. It is noteworthy that a significant analgesic effect, during both early (F4,39 ⫽ 4.417; P ⬍ 0.01) and late (F4,39 ⫽ 6.41; P ⬍ 0.001) phases of the formalin test, was also observed when MNAC13 and morphine were coadministered, at Ugolini et al.

individual doses of each compound that were per se ineffective (Fig. 4A). As a matter of fact, the TrkA antibody treatment was able to interact with a subthreshold dose of 1 mg/kg of the opiate, producing an analgesic effect comparable with that induced by the effective dose of 2.5 mg/kg of morphine. Analogous pharmacological synergy was also observed when an ineffective dose of MNAC13 was combined with an ineffective dose (0.005 mg/kg) of fentanyl (F4,37 ⫽ 9.219, P ⬍ 0.0001; and F4,37 ⫽ 3.543, P ⬍ 0.01 for the early and late phases, respectively); this combination resulted in an analgesic effect comparable with that observed after the administration of the opioid alone at the dose of 0.01 mg/kg (Fig. 4B). Administration of the peripheral opioid antagonist naloxone methiodide (10 mg/kg), which does not cross the blood–brain barrier, failed to counteract the combined analgesic response induced by MNAC13 plus morphine (t14 ⫽ 1.887, P ⫽ 0.080 and t14 ⫽ ⫺0.337, P ⫽ 0.7413 for the early and late phases, respectively), suggesting that the opioid component of the combined antinociceptive effect exerts its action in the CNS. Discussion Our data provide the first direct demonstration that neutralizing antibodies directed against the TrkA receptor induce potent analgesic effects on both inflammatory and neuropathic pain in mice. Inhibition of TrkA signaling therefore represents an efficient way of reducing both acute and chronic pain. In fact, the licking behavior recorded in the formalin pain model was significantly reduced by the anti-TrkA antibody MNAC13, as was the neuropathic pain induced by chronic constriction injury. These results strongly support the idea that NGF signaling through TrkA is essential to account for the hyperalgesic effects of NGF, although a role for the p75 receptor cannot be ruled out (29). Remarkably, this study adds independent evidence to the emerging idea that the NGF signaling system plays key pronociceptive and proinflammatory actions (29), highlighting the important role of the TrkA receptor. The interaction of NGF with the TrkA receptor represents a key event in the activation and sensitization of primary afferents nociceptors and in following central changes. The NGF/ TrkA system is a sort of master control system in the spreading of inflammation around the nociceptive terminal, because it is functionally placed upstream of many different molecular partners involved in inflammation (including the NGF/TrkA system itself), and it regulates the expression and function of ion channels conveying nociceptive signals (C-fiber nociceptors), including the transient receptor potential vanilloid receptor 1 (TRPV1) (38, 39), the P2⫻3 ATP receptors (40), and the tetrodotoxin (TTX)resistant SNS sodium channel (41). TRPV1 undergoes sensitization PNAS 兩 February 20, 2007 兩 vol. 104 兩 no. 8 兩 2987

NEUROSCIENCE

Fig. 2. Analgesic effects of MNAC13 on inflammatory pain in CD1 mice. (A) Comparison between licking behavior responses of saline (SAL: n ⫽ 7) and irrelevant IgG (60 ␮g per mouse, n ⫽ 10) s.c.-injected mice, during the early and late phases of formalin test. (B) Dose/response effect of MNAC13, s.c.-injected into the right hindpaw 18 h before testing, on the early (0 –15 min) and late (15– 40 min) phase of formalin test (IgG: 60 ␮g, n ⫽ 9; MNAC13 mAb: 0.9 ␮g, n ⫽ 8; 1.875 ␮g, n ⫽ 10; 3.75 ␮g, n ⫽ 9; 7.5 ␮g, n ⫽ 9; 15 ␮g, n ⫽ 10; 30 ␮g, n ⫽ 10; 60 ␮g per mouse, n ⫽ 9). *, P ⬍ 0.05; **, P ⬍ 0.01 vs. IgG-injected (n ⫽ 9) mice (Tukey/Kramer). (C) Analgesic effects of s.c .and i.p. administration of MNAC13 on the early (0 –15 min) and late (15– 40 min) phase of formalin test (s.c.: 15 ␮g per mouse; i.p.: 15 ␮g per mouse; 18 h before the test. s.c.: IgG, n ⫽ 9; MNAC13, n ⫽ 8; i.p.: IgG, n ⫽ 10 ; MNAC13, n ⫽ 9). Data are presented as means ⫾ SEM.

Fig. 3. Antiallodynic effects measured with a dynamic plantar aesthesiometer after four i.p. injections of 50 ␮g per mouse MNAC13 (from day 3 to day 6 after ligation of the sciatic nerve) in CCI mice tested for 14 days (IgG, n ⫽ 7; MNAC13, 50 ␮g per mouse, n ⫽ 5) (A) and after eight i.p. injections of 30 or 70 ␮g per mouse MNAC13 (from day 3 to day 10 after ligation of the sciatic nerve) in CCI mice tested from day 3 up to day 31 (B). (IgG, n ⫽ 11; MNAC13, 30 ␮g per mouse, n ⫽ 8; MNAC13, 70 ␮g per mouse, n ⫽ 12.) Black arrows represent the antibody i.p. administrations. BL (baseline) refers to mechanical threshold before operation. Black lines represent mechanical thresholds of sham-operated controls (n ⫽ 6). Data are presented as means ⫾ SEM. *, P ⬍ 0.05; **, P ⬍ 0.01 vs. IgG-injected mice (Tukey/Kramer).

after NGF-driven activation of phosphatidyl-inositol 3-kinase (PI3K), protein kinase C (PKC), and calmodulin-dependent kinase II (CaMKII) pathways (42, 43), and TRPV1 levels are increased after NGF-induced activation of p38 mitogen-activated protein kinase (MAPK) (44).

We have demonstrated that treatment with the MNAC13 antiTrkA antibody impairs TrkA signaling, resulting in the inhibition of SHC-mediated as well as PLC␥1 pathways. Consistently, a MNAC13-dependent inhibition of TrkA phosphorylation could also be revealed (data not shown) by using an antiphopshoTrk

Fig. 4. The combined effect of ineffective doses of MNAC13 (1 ␮g per mouse) and morphine (Mo 1 mg/kg) (A) or fentanyl (FEN 5 ␮g/kg) (B) results in analgesia on inflammatory pain similar to analgesia induced by effective doses of the opiates (morphine 2.5 mg/kg and fentanyl 10 ␮g/kg). The anti-TrkA antibody shows synergistic action with the opiates both in the early (0 –15 min) and late (15– 40 min) phases of formalin test. Irrelevant IgG (1 ␮g per mouse) and saline-injected (Sal) mice are used as control groups for MNAC13 and opioids, respectively. Naloxone methiodide (Nx-met 10 mg/kg) does not modify the response to MNAC13 plus morphine combination. (A) IgG⫹Sal, n ⫽ 8; MNAC13⫹Sal, n ⫽ 9; IgG⫹Mo 1 mg/kg, n ⫽ 9; MNAC13⫹Mo 1 mg/kg, n ⫽ 11; MNAC13⫹Nx-met⫹Mo 1 mg/kg, n ⫽ 9; IgG⫹Mo 2.5 mg/kg, n ⫽ 9; (B) IgG⫹Sal, n ⫽ 10; MNAC13⫹Sal, n ⫽ 11; IgG⫹FEN 5 ␮g/kg, n ⫽ 6; MNAC13⫹ FEN 5 ␮g/kg, n ⫽ 6; IgG⫹ FEN 10 ␮g/kg, n ⫽ 7. Data are presented as means ⫾ SEM. *, P ⬍ 0.05; **, P ⬍ 0.01 vs. IgG ⫹ Sal-injected mice (Tukey/Kramer). 2988 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0611253104

Ugolini et al.

Ugolini et al.

peripheral sites, failed to antagonize the synergistic effect between morphine and MNAC13, indicating that the interaction between opiates and the anti-TrkA occurs centrally. This hypothesis is also supported by the observation that if the antibody was administered only 5 min before formalin, a time interval that is not sufficient for MNAC13 to diffuse to a central site, no antinociception was observed (data not shown). These data highlight the strategic positioning of the TrkA receptor in the molecular pathways underlying both inflammatory and neuropathic pain and provide the background to further validate TrkA as a clinically considerable target (29) for forms of chronic and inflammatory pain that are presently badly or insufficiently treated. In this context, it is important to observe that the targeting of the TrkA receptor may not be therapeutically equivalent to that of the NGF ligand for a number of reasons, including the following: (i) the ligand and the receptor may be available in very different quantities at the injured site, (ii) an anti TrkA receptor antibody might be shown or engineered to harbor additional effector properties besides a pure ligand-competition mechanism of action. In conclusion, the evidence provided supports the therapeutic concept of antibodies targeting the TrkA receptor as a viable option for the development of new analgesic drugs antagonizing NGF function. Methods MNAC13 Antibody Production and Purification. MNAC13 mouse

hybridoma cells were cultured as described (30). Monoclonal antibody MNAC13 was purified from MNAC13 hybridoma supernatant, according to a protocol described in ref. 51. Surface Plasmon Resonance Using BIACore. SPR experiments were

performed by using a BIACore instrument. TrkA and p75 immunoadhesins (30, 52) were immobilized on CM5 sensor chip by cross-linking the amine groups according to the manufacturer’s instructions. The surface plasmon resonance signal for immobilized TrkA and p75 immunoadhesins was found to be 7,882 and 3,943 resonance units respectively after completion of the chip regeneration cycle. MNAC13 monoclonal antibody was injected in Hepes-buffered saline buffer (running buffer) at a concentration of 50 nM at a flow rate of 30 ␮l/min. TrkA and PLC␥1 Phosphorylation-Inhibition Assay. BALB/C 3T3-

transfected cells (3T3-TrkA), expressing 106 human TrkA molecules per cell were kindly provided by Stefano Alema` (Consiglio Nazionale delle Ricerche, Institute of Cell Biology, Rome, Italy) and were maintained in DMEM (Life Technologies, Milan, Italy), supplemented with 10% FCS. The 3T3-TrkA cells were plated in 35-mm Petri dishes at a density of 106 cells per dish. The next day, 3T3-TrkA cells were washed and incubated for 1 h at 37°C in the presence or absence of MNAC13 monoclonal antibody and of the irrelevant monoclonal antibody SV5 (two different concentrations: 200 and 300 ␮g/ml) in serum-free medium supplemented with 0.05% BSA. After antibody pretreatment, 100 ng/ml NGF was added for 10 min at 37°C. Cells were washed with PBS and scraped in 250 ␮l of cold RIPA buffer supplemented with phosphatases and proteases inhibitors (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 1% Na deoxycholate, 10 mM EDTA, protease inhibitor mixture purchased from Roche Diagnostics, Monza, Italy, 1 mM sodium orthovanadate, 50 mM NaF, and 1 mM okadaic acid); insoluble material was removed with a 5-min, 9,000 ⫻ g centrifugation. Extracts were separated on SDS polyacrylamide 10% gels and transferred to nitrocellulose by standard protocols (53). After blocking in PBS (5% nonfat dry milk) for 1 h at room temperature with gentle agitation, the filters were incubated overnight at 4°C with the following primary antibodies: anti-phospho-TrkA antibody PNAS 兩 February 20, 2007 兩 vol. 104 兩 no. 8 兩 2989

NEUROSCIENCE

antibody (kindly provided by M. V. Chao, New York University School of Medicine, New York, NY) specifically recognizing a phosphotyrosine residue known to interact with the PLC␥ (37). However, on the basis of our results, we cannot rule out the possibility that the PI3K intracellular pathway is also involved, because the activation of this specific pathway is believed to play a crucial role in the sensitization of nociceptive neurons by NGF (42, 43). Further experimental efforts will be necessary to fully characterize MNAC13 mode of action, because a multiplicity of mechanisms might be in play. These mechanisms include the possibility of an influence of the monoclonal antibody on spontaneous dimerization, considering that TrkA activity, in the absence of NGF, may be affected by mutagenesis of extracellular domain sequences (45). The role of NGF in the development and maintenance of neuropathic pain is a complex and not well defined issue. However, in recent years, growing evidence supports the hypothesis that a block of NGF action could provide effective pain relief in this chronic pain condition. In the CCI model, an increase of NGF levels is observed in the cells of the injured area and in dorsal root ganglia at the level of the lesion, suggesting the involvement of NGF in the development of hyperalgesia in this model of neuropathic pain (24). Direct administration of NGF into the sciatic nerve induces hyperalgesia in rats (26), whereas NGF inhibition reduces thermal and mechanical hyperalgesia in CCI model (24, 27, 46), partial nerve transection (28) and partial transection of the spinal cord (47). Our data reinforce the hypothesis that NGF and TrkA also play important roles in modulating chronic neuropathic pain. In the present research, the antiallodynic properties of MNAC13 in the CCI model reveal a quite unexpected and remarkable temporal profile, identifying a long-lasting analgesic effect of the anti-TrkA treatment, which could depend on the modulation of gene expression patterns. It is known that retrograde NGF signaling enhances the expression of several proteins involved in the sensitization processes, such as substance P, ion channels, and other neurotrophic factors (29). It could be hypothesized that the anti-TrkA treatment affects the expression levels of these proteins. There is evidence that TrkA expression in adult nociceptive neurons may be up-regulated by NGF itself (48). On this basis, it may be hypothesized that the long-term effect of MNAC13 at least partially relies on the suppression of such feed-forward mechanism, after anti-TrkA-mediated prolonged inhibition of NGF signaling. The mechanistic dissection of this long-term effect remains the object of further studies, that might prospect MNAC13 as a disease-modifying drug for chronic pain states, inducing a therapeutically beneficial feedback loop. Another important finding of the present research is that the anti-TrkA treatment was able to potentiate the antinociceptive effect of morphine and fentanyl in the formalin test. Previous evidence showed a negative interaction between opiates and NGF signaling pathways in primary sensory neurons (34). Conversely, the inflammation-related increase of NGF induced by carrageenan is attenuated by morphine (35). Morphine and other opiates are currently used as analgesics for several types of severe persistent pain, although their use is limited by several drawbacks, including development of tolerance and physical dependence (49). The analgesic effect induced by combining subthreshold doses of opiates and MNAC13 opens the possibility of significantly lowering the effective dose of opiates, which would have a tremendous impact in terms of clinical application. Most of the effects of opiates have been reported to be central, even if a growing number of recent data reported a peripheral effect of these drugs on pain modulation (50). A synergy between central and peripheral effects has also been reported (50). It would be important, in this context, to investigate the site of the interaction observed between the antibody and morphine. Naloxone methiodide, an opioid antagonist that acts only at the

(1:1,000; Cell Signaling Technology, Beverly, MA), the antiphosphotyrosine (Y785) TrkA antibody (1:200; kindly provided by M. V. Chao) (37), antitubulin YOL (1:3 dilution of supernatant, used to normalize results) (54), antiphosphoPLC␥1 (1:1,000; Cell Signaling Technology) and anti-PLC ␥1 (1:1,000, Cell Signaling Technology, used to normalize results). The addition of the secondary antibody (anti-rabbit for antiTrkA and anti-PLC ␥1 antibodies; anti-rat for YOL) coupled to HRP diluted 1:1,000 (1 h at room temperature) was followed by washes in PBST (0.1% Tween 20) and PBS at room temperature with gentle agitation, before HRP-conjugate detection by the Amersham (Piscataway, NJ) electrochemiluminescence protocol [replaced by Pierce (Rockland, IL) only for anti-phosphoPLC␥1]. Behavioral Experiments. For details on behavioral experiments,

see SI Text. CD1 male mice (35–40 g) (Charles River Laboratories, Como, Italy) were used according to the guidelines of the International Association for the Study of Pain. Formalin Test. Mice were s.c. injected into the right hind paw with

20 ␮l of formalin solution (5% in saline) immediately before the test, as described (56). Licking activity, recorded continuously for 40 min, was calculated in blocks of consecutive 5-min periods and analyzed as the early (0–15 min) and the late (15–40 min) phases of the formalin-evoked response. General activity and selfgrooming were also continuously recorded for 40 min during the formalin test. No significant differences were observed for these parameters after MNAC13 treatment. Different groups of mice, 18 h before formalin injection, were i.p. or s.c. injected into the dorsal surface of the right hind paw with MNAC13 antibody or irrelevant mouse IgGs (Sigma–Aldrich, Milan, Italy), as control group (20 ␮l). In additional experimental groups, mice treated with MNAC13 were either i.p. injected with morphine hydrochloride (Laboratori Guieu, Milan, Italy) or s.c. (systemically) injected with fentanyl (Hameln Pharmaceuticals, 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

Renfrey S, Downton C, Featherstone J (2003) Nat Rev Drug Discov 2:175–176. Levi-Montalcini R (1952) Ann NY Acad Sci 55:330–344. Levi-Montalcini R (1987) Science 237:1154–1162. Frade JM, Barde YA (1998) BioEssays 20:137–145. Kaplan DR, Hempstead BL, Martin-Zanca D, Chao MV, Parada LF (1991) Science 252:554–558. Klein R, Jing SQ, Nanduri V, O’Rourke E, Barbacid M (1991) Cell 65:189–197. Patapoutian A, Reichardt LF (2001) Curr Opin Neurobiol 11:272–280. Radeke MJ, Misko TP, Hsu C, Herzenberg LA, Shooter EM (1987) Nature 325:593–597. Chao MV (1991) Curr Top Microbiol Immunol 165:39–53. Julius D, Basbaum AI (2001) Nature 413:203–210. Lewin GR, Barde YA (1996) Annu Rev Neurosci 19:289–317. McMahon SB (1996) Philos Trans R Soc London B 351:431–440. Lewin GR, Ritter AM, Mendell LM (1993) J Neurosci 13:2136–2148. Skaper SD (2001) Mol Neurobiol 24:183–199. Sawada J, Itakura A, Tanaka A, Furusaka T, Matsuda H (2000) Blood 95:2052–2058. Aloe L, Skaper SD, Leon A, Levi-Montalcini R (1994) Autoimmunity 19:141–150. Woolf CJ, Safieh-Garabedian B, Ma QP, Crilly P, Winter J (1994) Neuroscience 62:327–331. McMahon SB, Bennett DL, Priestley JV, Shelton DL (1995) Nat Med 1:774–780. Koltzenburg M, Bennett DL, Shelton DL, McMahon SB (1999) Eur J Neurosci 11:1698–1704. Dyck PJ, Peroutka S, Rask C, Burton E, Baker MK, Lehman KA, Gillen DA, Hokanson JL, O’Brien PC (1997) Neurology 48:501–505. Svensson P, Cairns BE, Wang K, Arendt-Nielsen L (2003) Pain 104:241–247. Gold BG, Mobley WC, Matheson SF (1991) J Neurosci 11:943–955. Diamond J, Holmes M, Coughlin M (1992) J Neurosci 12:1454–1466. Herzberg U, Eliav E, Dorsey JM, Gracely RH, Kopin IJ (1997) NeuroReport 8:1613–1618. Li L, Xian CJ, Zhong JH, Zhou XF (2003) Mol Cell Neurosci 23:232–250. Ruiz G, Ceballos D, Banos JE (2004) Brain Res 1011:1–6. Ro LS, Chen ST, Tang LM, Jacobs JM (1999) Pain 79:265–274. Theodosiou M, Rush RA, Zhou XF, Hu D, Walker JS, Tracey DJ (1999) Pain 81:245–255. Hefti FF, Rosenthal A, Walicke PA, Wyatt S, Vergara G, Shelton DL, Davies AM (2006) Trends Pharmacol Sci 27:85–91.

2990 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0700697104

Hameln, Germany) 15 min before the beginning of the test. Naloxone methiodide (10 mg/kg, Sigma–Aldrich), known to counteract morphine effects on peripheral receptors (55), at that dose, was i.p. administered 5 min before morphine in a group of mice previously treated with MNAC13 by using the same experimental procedure. Neuropathic Pain Test. After the procedure originally proposed by

Bennett and Xie (33) adapted for mice, the CCI model was used to induce neuropathic pain as described (56). The sham surgical procedure was the same, but no ligatures were placed around the nerve. At postoperative day 3, the production of neuropathic pain was assessed by measuring the sensitivity of both the ipsi- and contralateral hindpaws to normally nonnoxious punctuate mechanical stimuli. The nerve injury-induced mechanical allodynia was tested by using a dynamic plantar aesthesiometer (Model 37400; Ugo Basile, Varese, Italy). Each testing day, the latency time to withdraw ipsilateral and contralateral hindpaws was taken in three measurements with at least 10 seconds between each measurement. The mean of three measurements per paw represented the mechanical threshold for each testing day. After the measurement of postoperative day 3, different groups of mice were i.p. injected with MNAC13 from day 3 to day 6 (four injections) or from day 3 to day 10 (eight injections) and tested for 14 or for 31 days, respectively. Statistical Analysis. The two phases characterizing the formalin

test were separately analyzed by unpaired Student’s t test (for two groups) or by ANOVA (one- or two-way ANOVAs). To analyze data deriving from neuropathic experiments, a two-way ANOVA for repeated measures was used. Post hoc comparisons were carried out by using a Tukey/Kramer test. We thank Dr. Tiziana Grassi and Anna Di Luzio for technical help with hybridoma and cell culture and Mrs. Pina Moliterno for reviewing the English text. 30. Cattaneo A, Capsoni S, Margotti E, Righi M, Kontsekova E, Pavlik P, Filipcik P, Novak M (1999) J Neurosci 19:9687–9697. 31. Covaceuszach S, Cattaneo A, Lamba D (2005) Proteins 58:717–727. 32. Porro CA, Cavazzuti M (1993) Prog Neurobiol 41:565–607. 33. Bennett GJ, Xie YK (1988) Pain 33:87–107. 34. McDowell TS (2004) Neuroscience 125:1029–1037. 35. Amann R, Lanz I, Schuligoi R (2002) Pharmacology 66:169–172. 36. Pesavento E, Margotti E, Righi M, Cattaneo A, Domenici L (2000) Neuron 25:165–175. 37. Rajagopal R, Chen ZY, Lee FS, Chao MV (2004) J Neurosci 24:6650–6658. 38. Chuang HH, Prescott ED, Kong H, Shields S, Jordt SE, Basbaum AI, Chao MV, Julius D (2001) Nature 411:957–962. 39. Winston J, Toma H, Shenoy M, Pasricha PJ (2001) Pain 89:181–186. 40. Ramer MS, Bradbury EJ, McMahon SB (2001) J Neurochem 77:864–875. 41. Fjell J, Cummins TR, Dib-Hajj SD, Fried K, Black JA, Waxman SG (1999) Brain Res Mol Brain Res 67:267–282. 42. Bonnington JK, McNaughton PA (2003) J Physiol (London) 551:433–446. 43. Zhang X, Huang J, McNaughton PA (2005) EMBO J 24:4211–4223. 44. Ji RR, Samad TA, Jin SX, Schmoll R, Woolf CJ (2002) Neuron 36:57–68. 45. Arevalo JC, Conde B, Hempstead BL, Chao MV, Martin-Zanca D, Perez P (2000) Mol Cell Biol 20:5908–5916. 46. Ramer MS, Bisby MA (1999) Eur J Neurosci 11:837–846. 47. Gwak YS, Nam TS, Paik KS, Hulsebosch CE, Leem JW (2003) Neurosci Lett 336:117–120. 48. Mearow KM (1998) Exp Neurol 151:14–25. 49. Przewlocki R, Przewlocka B (2001) Eur J Pharmacol 429:79–91. 50. Stein C, Schafer M, Machelska H (2003) Nat Med 9:1003–1008. 51. Covaceuszach S, Cattaneo A, Lamba D (2001) Acta Crystallogr D 57:1307–1309. 52. Chamow SM, Ashkenazi A (1996) Trends Biotechnol 14:52–60. 53. Harlow E, Lane D (1988) in Antibodies: A Laboratory Manual (Cold Spring Harbor Lab Press, Cold Spring Harbor, NY), pp 471–510. 54. Kilmartin JV, Wright B, Milstein C (1982) J Cell Biol 93:576–582. 55. McGaraughty S, Honore P, Wismer CT, Mikusa J, Zhu CZ, McDonald HA, Bianchi B, Faltynek CR, Jarvis MF (2005) Br J Pharmacol 146:180–188. 56. Luvisetto, S, Marinelli S, Panasiti MS, D’Amato F, R, Fletcher CF, Pavone F, Pietrobon D (2006) Neuroscience 142:823–832.

Ugolini et al.