Elevated levels of nerve growth factor have been linked to the onset and persistence of many pain-related disorders and asthma. Described here are the design, expression, refolding, and purification of a monomeric (nonstrand-swapped) form of the binding domain of the nerve growth factor receptor, designated TrkAd5. We have shown that TrkAd5 produced recombinantly binds nerve growth factor with picomolar affinity. TrkAd5 has been characterized using a variety of biophysical and biochemical assays and is shown here to be stable in both plasma and urine. The palliative effects of TrkAd5 are demonstrated in animal models of inflammatory pain and allergic asthma. We conclude that TrkAd5 will prove effective in ameliorating both acute and chronic conditions where nerve growth factor acts as a mediator and suggest a role for its application in vivo as a novel therapeutic.
Many persistent pain states are associated with increased levels of nerve growth factor (NGF) (Woolf et al., 1994; McMahon et al., 1995), and the application of exogenous NGF can produce rapid and prolonged behavioral hyperalgesia in humans (Petty et al., 1994; Dyck et al., 1997) and in animal pain models (Lewin et al., 1993; Andreev et al., 1995). In addition, patients suffering from allergic diseases, including asthma, have increased serum levels of nerve growth factor (Bonini et al., 1996; Frossard et al., 2004).
Increased endogenous nerve growth factor levels may cause hyperalgesia by both neurogenic and non-neurogenic means. Importantly, nerve growth factor causes a rapid sensitization of nociceptive sensory neurons (Shu and Mendell, 1999) via TrkA receptors (Averill et al., 1995), resulting in sensitization of the capsaicin vanilloid receptor TRPV1 (Chuang et al., 2001) and downstream intracellular signaling (Shu and Mendell, 1999; Chuang et al., 2001; Bonnington and McNaughton, 2003). In addition, the tetrodotoxin-resistant voltage-gated sodium channel Nav 1.8 (Kerr et al., 2001) and the purinergic receptor P2X3 (Ramer et al., 2001) are modulated by nerve growth factor; both are expressed on primary sensory neurons and are associated with sensitization to pain.
We initially showed that the nerve growth factor binding site on the TrkA receptor corresponds to the immunoglobulin-like domain 4 and 5 (Holden et al., 1997). In addition, we showed that all of the binding activity resided in the d5 domain (TrkAd5). This domain was expressed with a histidine-tag in Escherichia coli, purified, and the structure was determined (Robertson et al., 2001). At the concentrations required for crystal formation, the protein folds as a strand swapped dimer, where the A strand from each monomer is transposed such that the domain is incapable of binding nerve growth factor. Here, we demonstrate the production and purification of a nonhistidine-tagged TrkAd5 and describe a novel refolding strategy that results in high concentrations of monomeric protein. In addition, we show that the protein is biologically active in vitro, is stable in blood and urine, and is able to be lyophilized and reconstituted with full activity. We further demonstrate an extremely potent effect in vivo in an animal model of inflammatory pain and an in vitro model of allergic asthma. Conventional pharmaceuticals do not provide effective therapy for conditions such as interstitial cystitis and asthma, and we conclude that TrkAd5 represents a novel therapeutic agent for these disease states and many others in which raised nerve growth factor levels play a role in the etiology of the disease. TrkAd5 is currently being developed for use in clinical trials and has been designated ReN1820 (ReNeuron Ltd., Guildford, Surrey, UK).
Materials and Methods
Expression Vector Construction. Thirteen primers were designed to incorporate silent mutations within the first 30 codons of the open reading frame of TrkAd5 and their theoretical mRNA secondary structures determined by analysis using the MFOLD structure prediction program (M. Zuker, Pasteur Institute, Paris, France). From the results generated, six were chosen and used as forward primers in a polymerase chain reaction with the wild-type pET24a(+)TrkAd5 as template. The six mutant polymerase chain reaction products were ligated into the pET24a(+) vector (Novagen, Madison, WI) at the NdeI/XhoI sites and then amplified by transformation into a nonexpression host, XL1 blue. Purified, mutant constructs were confirmed by sequencing (MWG-Biotech, Ebersberg, Germany) and then transformed into BL21(DE3) E. coli for expression studies. Samples of uninduced/induced cultures from single colonies were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis.
Production and Purification of Recombinant TrkAd5. Five liters of 2YT medium, 50 μg/ml kanamycin, was inoculated with a 25-ml overnight culture of E. coli BL21(DE3) pET24a(+)TrkAd5 7 at 37°C. The culture was induced by the addition of 5 ml of 1 M isopropylthio-β-d-galactoside. Cells were harvested by centrifugation, and pellets were stored frozen at –80°C until required. Inclusion bodies were purified and processed as described previously (Robertson et al., 2001). Inclusion body pellets were stored overnight at –80°C, and the pellet was subsequently resuspended in 50 ml of solubilization buffer (8 M urea, 40 mM Tris, pH 8.5, 100 mM NaCl, 50 mM dithiothreitol, and 1 mM EDTA) and rocked on a platform for 3 h at room temperature (RT). Solubilized inclusion bodies were loaded on a Q-Sepharose Waters AP-5 column, pre-equilibrated with solubilization buffer containing 10 mM dithiothreitol. TrkAd5 was step eluted using solubilization buffer containing 200 mM NaCl. The protein peak was collected and incubated at RT for 14 to 21 days. Urea solubilized TrkAd5 was refolded and purified on a HiLoad Superdex 200 XK26/70 column, pre-equilibrated with elution buffer (20 mM Tris, pH 8.5, and 100 mM NaCl). Peaks corresponding to the monomer fraction were pooled. For animal studies, endotoxins were removed from TrkAd5 using a 20-ml DetoxiGel column (Pierce Chemical, Rockford, IL) and found to be 1.71 endotoxin units/ml by Limulus assay (BioWhittaker Europe, Verviers, Belgium). The TrkAd5 fractions were concentrated to 0.8 to 1 mg of protein/ml, 10% (v/v) glycerol was added, and the protein was filter sterilized, snap frozen in liquid nitrogen, and stored at –80°C.
PC12 Cell Neurite Outgrowth. The ability of TrkAd5 to sequester nerve growth factor was assessed in the rat phaeochromocytoma cell line PC12 by noting the reduction in neurite outgrowth using the method described previously (Allen et al., 2001; Robertson et al., 2001). This method is routinely used to visually assess activity. However, quantification of TrkAd5 activity in PC12 cells was carried out using the apoptosis assay as described below.
PC12 Cell Apoptosis Assay. PC12 cells were plated out on collagen coated 96-well plates in 100 μl of serum-free media at a cell density of 1 × 104 cells/well, after washing three times in serum-free Dulbecco's modified Eagle's medium [containing 1% (v/v) penicillin/streptomycin solution and 2 mM l-glutamine]. Nerve growth factor, TrkAd5/Tris buffer, or both were added at required concentrations with 10× Dulbecco's modified Eagle's medium (D-2554; Sigma Chemical, Poole, Dorset, UK) used to make up the total volume per well to 200 μl and incubated at 37°C with 5% CO2 for 3 days. Metabolic turnover was assessed by addition of 40 μl/well of premixed CellTiter 96AQueous assay reagents (Promega, Madison, WI). Plates were incubated for up to 4 h. Optical density was measured at 492 nm.
In Vitro Nerve Growth Factor Competition ELISA. Binding of TrkAd5 to nerve growth factor was assessed using a competitive ELISA format. The 96-well plates (Nunc MaxiSorp; Nunc A/S, Roskilde, Denmark) were coated with 50 μl/well of a 1:1000 dilution of nerve growth factor capture antibody (Sigma Chemical) in 50 mM sodium carbonate solution, pH 9.6, and incubated overnight at 4°C. Nonspecific binding sites were blocked by addition of 200 μl of blocking buffer [50 mM sodium carbonate containing 1% (w/v) BSA and 5% (w/v) sucrose] per well and incubation at RT for 2 h. Washing was with TBS-T [containing 0.1% (v/v) Tween 20]. Standards (50 μl) of human recombinant nerve growth factor-β (Sigma Chemical) (0–1 ng/ml) were prepared with and without TrkAd5 and preincubated at RT for 10 min before addition to plate for 1 h at RT. Then, 50 μl of 0.1 μg/ml biotinylated anti-human nerve growth factor monoclonal antibody (R&D Systems, Minneapolis, MN) in TBS-T with 1% (w/v) BSA was added to each well and the plate incubated at RT for 1 h. Next, 50 μl of 0.5 μg/ml streptavidin alkaline phosphatase polymer (Zymed Laboratories, South San Francisco, CA) in TBS-T with 1% (w/v) BSA was added per well and incubated at RT for 1 h. Finally, 50 μl of 4-methyl umbelliferyl phosphate (Sigma Chemical) substrate solution was added per well. Fluorescence readings were measured on a Fluoroskan II (LabSystems, Waltham, MA) at excitation/emission wavelengths of 355/460 nm after 10, 20, and 30 min.
Mass Spectrometric Characterization of TrkAd5 and N-Terminal Sequencing. Molecular weight of proteins was measured accurately using MALDI-TOF mass spectrometry as described previously (Robertson et al., 2001).
A sample of TrkAd5 was run on a 15% Tricine gel (Schagger and von Jagow, 1987) and then transferred onto ProBlott membrane (Applied Biosystems, Foster City, CA) using standard Western blotting techniques. Protein bands were visualized following staining with 0.1% (w/v) Serva Blue G. N-terminal sequencing was carried out using an Applied Biosystems 492 cLC protein sequencer using standard cycles.
Shelf-Life Stability at 4°C. Sterile filtered TrkAd5 [0.5 mg/ml in 20 mM Tris, pH 8.5, 148 mM NaCl, and 10% (v/v) glycerol] was stored at 4°C. At specific time points, 100 μl of sample was prepared for HPLC analysis by the addition of 40 μl of protein standard (1 mg/ml insulin) and 1860 μl of HPLC solvent A (4.4% acetonitrile and 0.1% TFA). The prepared sample was injected onto an LC-304 C5 column (Supelco, Bellfonte, PA), equilibrated at a flow rate of 1 ml/min using 100% solvent A for 10 min, and then eluted over 50 min with a linear gradient ending at 100% solvent B [95.5% acetonitrile and 0.1% trifluoroacetic acid (v/v)]. TrkAd5 had an over all retention time of 39 min against the insulin standard of 32 min.
Sterile filtered TrkAd5 was stored at 4°C for up to 3 months. The endpoint sample was also tested using the PC12 cell apoptosis assay and in vitro nerve growth factor competition ELISA.
Lyophilization. TrkAd5 was dialyzed against three changes of 5% (w/v) sucrose, snap frozen in liquid nitrogen, lyophilized, and stored at –80°C.
Iodination of TrkAd5 and Stability Time Courses. TrkAd5 was iodinated as described previously (Robertson et al., 2001). We used 50,000 cpm per time point for the stability time course. Protein stability was studied over 1920 min, and 10 time points were taken. Radiolabeled TrkAd5 was added to 3 ml of rat serum, human serum, or human urine, and a 1000-fold amount of unlabeled TrkAd5 was added as a carrier protein. At each time point, 100-μl samples were taken, snap frozen on dry ice, and stored at –80°C until the assay endpoint. Samples were then thawed, and 100 μl of 15% TCA was added. Following 20-min incubation on ice, samples were spun at 2500 rpm, the supernatant was removed, and the pellets were counted.
Chemically Induced Cystitis Study. The method used was similar to that described previously (Dmitrieva et al., 1997). All animal studies were carried out in accordance with the declaration of Helsinki. Briefly, 21 female Wistar rats (190–220 g) were anesthetized with urethane (1.25 g/kg i.p.), which produced a stable level of anesthesia lasting the entire experiment. The carotid artery was cannulated. Body temperature was measured and maintained close to 37°C. The bladder was catheterized transurethrally with a 1.1-mm polythene catheter. A ventral midline laparotomy was performed, enabling complete bladder emptying to be confirmed.
Bladder motility was assessed by slow filling of the bladder with normal saline through a transurethral catheter, at 0.05 ml/min for 14 min (bladder volume was increased from 0 to 0.7 ml in this period); rate of filling used was within the physiological range. In normal animals, during filling, bladder pressure increased gradually for approximately 5 min, beyond which a series of regular micturition contractions were elicited, typically two to five within the 14-min period of measurement. Only animals with no visible signs of bladder inflammation had clear urine and that showed normal baseline cystometrograms were chosen for further experimentation.
After control determinations, animals were subjected to one of two treatments, and cystometrograms were subsequently undertaken at 1, 3, and 5 h. Briefly, 0.5 ml of a turpentine/olive oil mixture (50:50) was instilled into the bladder for 1 h after which the turpentine was drained. This treatment produced a sterile inflammatory immune response with invasion of immune cells and development of hyper-reflexia. The inflammation started within 1 h of turpentine installation and progressively increased over the next few hours. To facilitate the analysis of treatment for each animal the slope of the regression line was calculated for each outcome measure and compared between the saline and TrkAd5 by unpaired t tests.
In Vitro Organ Bath Studies. Organ bath studies were carried out as described previously (De Vries et al., 2001). Male Hartley guinea pigs (44–66 g; Harlan CPB, Zeist, The Netherlands) were used in all experiments. The Animal Care Committee of the Utrecht University approved the animal studies. The animals were sacrificed by cervical dislocation, and isolated tracheal rings (three cartilage segments per ring) were placed in an isometric organ bath setup, containing warmed (37°C) Krebs' solution (pH 7.4, with 8.3 mM glucose) gassed with 95% O2, 5% CO2. The experiment started with four washout periods lasting 15 min each. During these washouts, a tension was applied of 2000, 2000, 4000, and 2000 mg, respectively. A histamine concentration-response curve (10–8–10–3 M) was subsequently conducted to measure contractility of the tracheal rings. Nerve growth factor was applied at a concentration of 20 ng/ml 30 min before the start of the response curve. TrkAd5 at 7.64 μM was applied 5 min before addition of nerve growth factor.
In a second set of experiments, guinea pigs were sensitized to ovalbumin (grade V; Sigma Chemical) by injections of a gel containing 20 μg of ovalbumin/ml and 200 mg of Al(OH)3/ml as adjuvant. Six injections were placed on the same day: one injection of 0.5 ml intraperitoneally, one injection of 0.1 ml nuchally, two injections of 0.1 ml axillarly, and two injections of 0.1 ml inguinally. Control animals were not sensitized and received instead a vehicle gel, not containing ovalbumin. Fourteen days after the sensitization procedure, tracheal rings were isolated. The experiment started with four washout periods lasting 15 min each. During these washouts, a tension was applied of 2000, 2000, 4000, and 2000 mg, respectively. TrkAd5 (7.4 μM) was added, and the response to ovalbumin challenge (1 mg/ml) was then measured. Sixty minutes later, following two 15-min washouts, the histamine-response curve was measured.
Design of Construct. TrkAd5 without extraneous amino acids was required for potential clinical application. The TrkAd5 construct was therefore recloned to eliminate the histidine-tag and any vector-derived amino acid sequence. This resulted in a lack of expression, probably because of the presence of positioning of the AUG start site at the base of a long stem-loop structure (Fig. 1a), which has the potential to interfere with the AUG translation initiation codon (Tessier et al., 1984). Predicted secondary structure of the expression construct was redesigned using MFOLD (M. Zuker, Pasteur Institute), such that the mRNA structure was altered while maintaining the amino acid sequence of TrkAd5 (Fig. 1b). In total, six constructs were designed, incorporating silent mutations within the first 20 to 30 nucleotides of the open reading frame, predicted to alter the mRNA structure to make both the ribosome binding site and the translational start more accessible. One construct was designed for optimal prokaryotic codon use. Recombinant proteins were then expressed (Fig. 1c). Of the constructs created, five produced high yields of protein of the correct molecular weight. One of these constructs d5#7, whose MFOLD structure can be seen in Fig. 1b, was chosen for further analysis. The construct designed for optimal codon use did not express protein (Fig. 1c, lane 7)
Protein Expression, Purification, and Refolding. Inclusion body preparations were isolated and processed from 5 liters of bioreactor cultures of d5#7. Final purification on an S200 Superdex column allowed rapid refolding and separation of the monomer away from any aggregate and/or dimeric species. The effects of incubating the protein for different times in urea showed that extended incubation effectively converts aggregate and dimer into monomeric protein (Fig. 2). The monomeric state of the protein was also assessed by native gel silver stain analysis with more than 95% homogeneity of the sample (data not shown).
MALDI-TOF Analysis and N-Terminal Sequencing. Monomeric fractions were shown by MALDI-TOF to have a molecular mass of 13,593 Da (Fig. 3). This corresponds to the calculated molecular mass of TrkAd5 with the initiating methionine removed and with the addition of one carbamylation adduct, which occurs during extended incubation in urea. N-terminal sequencing of TrkAd5 showed the first nine residues to be PASVQLHTA, corresponding to the predicted amino acid sequence.
The Nerve Growth Factor-Sequestering Ability of TrkAd5. PC12 cells differentiate and extend processes in a dose-dependent manner upon addition of nerve growth factor (Fig. 4a). TrkAd5 was shown to inhibit nerve growth factor-induced neurite outgrowth in PC12 cells (Fig. 4b) and to exacerbate the apoptotic response in a nerve growth factor-dependent cell survival assay, in which serum is withdrawn from PC12 cells (Fig. 4c). At a concentration of 5 ng/ml nerve growth factor, addition of TrkAd5 resulted in a dose-related reduction in cell survival to background levels, with an approximate IC50 of 1 μM. Both assays confirm the in vitro sequestration of nerve growth factor by TrkAd5.
A competition sandwich ELISA assay was designed for rapid assessment of binding capability of the active monomer. Preincubation of standard solutions of nerve growth factor with various concentrations of TrkAd5 (0–4.5 μM) gave a dose-dependent reduction in fluorescence signal (Fig. 4d) with an IC50 of approximately 0.05 μM.
Stability and Lyophilization. To determine any effect lyophilization may have on the bioactivity of the TrkAd5, the protein was lyophilized, stored overnight at –80°C, and then reconstituted to its original concentration. This procedure showed no adverse affects on the bioactivity of the protein as determined using the apoptosis assay (Fig. 4c).
The shelf-life stability of TrkAd5, stored at 4°C, was determined over a period of 75 days. Samples were assayed by HPLC with no appreciable degradation observed (data not shown). As a prerequisite to animal pharmacokinetic studies, the in vitro stability of radiolabeled TrkAd5 was measured, using TCA precipitation, over a 32-h time course in human and rat serum and human urine (Fig. 5). After 32 h in human and rat serum, 90% of the protein was still intact. The protein was less stable in urine with a half-life of approximately 32 h.
Effect of TrkAd5 in a Model of Chemically Induced Acute Cystitis. The ability of TrkAd5 to reduce inflammatory changes was investigated in an established rat chemically induced cystitis paradigm (McMahon and Abel, 1987; Dmitrieva et al., 1997). Experimental inflammation was induced by turpentine infusion and followed by injection of either saline (i.v.) or TrkAd5 (200 μg i.v., equivalent to 14.8 nmol) 1 h later at the onset of bladder irritation. Cystometrograms were taken for up to 5 h after the instillation of the irritant, and bladder responsiveness was assessed by micturition threshold, number of contractions, and the total duration of contractions. At 1 h after irritation of the bladder, both groups of animals treated with either saline (n = 13) or TrkAd5 (n = 8) developed similar degrees of hyper-reflexia; bladder reflexes were exaggerated and showed decreased micturition threshold, increased number of contractions, and increased total duration of contractions. Saline-treated animals thereafter continued to show progressive hyper-reflexia with bladder reflexes increasing from 216% at 1 h to 329% of control at 5 h. In the TrkAd5-treated animals, there was a reversal of progression of hyper-reflexia from 194 to 138% of control (unpaired t test; p = 0.0006) (Fig. 6, top). In saline-treated animals total time of contraction increased from 278 to 410% of control; however, in TrkAd5-treated animals, contraction times were reduced from 214 to 158% of control (unpaired t test; p = 0.0102) (Fig. 6, middle). Similarly, micturition threshold was reduced from 68 to 51% of control in saline-treated animals but increased from 67 to 85% of control in TrkAd5-treated animals (unpaired t test; p = 0.0275) (Fig. 6, bottom).
Effect of TrkAd5 on the Responsiveness of Guinea Pig Tracheal Smooth Muscle. The effect of TrkAd5 on the nerve growth factor-induced hyper-responsiveness of isolated guinea pig tracheal rings was investigated in an organ bath setup. As previously demonstrated (De Vries et al., 2001), incubation with 20 ng/ml nerve growth factor induced a hyper-responsiveness following application of histamine. We show here that the addition of 7.64 μM TrkAd5 before nerve growth factor exposure was able to block the concentration-related hyper-responsiveness to histamine, presumably by sequestration of nerve growth factor (Fig. 7a). Addition of TrkAd5 alone had no effect compared with controls (Fig. 7a). In a second series of experiments, the effect of TrkAd5 on ovalbumin challenge was investigated in tracheal rings isolated from guinea pigs that had been ovalbumin sensitized with ovalbumin 2 weeks previously (Fig. 7b). TrkAd5 was shown to inhibit this direct allergen-induced tracheal contraction (Fig. 7b). In addition, TrkAd5 was also shown to inhibit the hyper-responsiveness to histamine 60 min after ovalbumin challenge (Fig. 7c).
Asthma and inflammatory pain states are known to be associated with increased levels of endogenous nerve growth factor. For example, nerve growth factor is up-regulated in the joints of human arthritic patients (Aloe et al., 1992a); mRNA levels of nerve growth factor and TrkA are increased in samples from patients with chronic pancreatitis (Zhu et al., 1999); and increased bladder nerve growth factor levels are associated with painful conditions, including interstitial cystitis (Lowe et al., 1997). A related animal model exists for each of these conditions, with studies suggesting nerve growth factor has a pivotal role in nociceptor sensitization in each case (McMahon and Abel, 1987; Aloe et al., 1992b; Dmitrieva et al., 1997; Braun et al., 1998; De Vries et al., 1999; Toma et al., 2000). Neurotrophins levels have been shown to be raised in the urine (Okragly et al., 1999) and bladder (Lowe et al., 1997) of women suffering with interstitial cystitis. Studies with a cyclophosphamide-induced bladder inflammation model have demonstrated alterations in micturition reflex (Yoshimura and de Groat, 1999), probably mediated by neurotrophins produced in the bladder during cystitis (Vizzard, 2000). Thus inflammation-induced changes in neurotrophins may sensitize afferent and postganglionic nerves, resulting in bladder over activity, and may be involved in neuroplasticity of the lower urinary tract pathways (Lowe et al., 1997; Okragly et al., 1999; Vizzard, 2000; Qiao and Vizzard, 2002).
Studies in animal pain models in which TrkA immunoadhesins are administered have shown that sequestration of endogenous nerve growth factor is able to block the hyperalgesia associated with inflammation (McMahon et al., 1995; Zahn et al., 2004). However, these are large molecules and are unlikely to be clinically useful. Previously, we have shown that a histidine-tagged TrkAd5 domain reversed the electrophysiological correlates of complete Freund's adjuvant inflammation in the guinea pig (Djouhri et al., 2001). We describe here the design, expression, refolding, and purification of a small nontagged monomeric 13.5-kDa domain of the TrkA receptor, capable of sequestering nerve growth factor in an in vivo model of cystitis and an in vitro model of allergic asthma and of reversing symptoms associated with these conditions.
The protein purification protocol involved prolonged incubation in 8 M urea. Aggregation is a major problem of refolding protein from E. coli. This extended incubation in urea (14–28 days) may be beneficial for the refolding of other recombinant proteins in E. coli where aggregation is an issue. Endotoxins were removed before in vivo studies, and the protein was tested in vitro for the ability to neutralize the effect of nerve growth factor in two PC12 cell bioassays: prevention of neurite outgrowth and nerve growth factor-induced rescue from apoptosis. In a previous study (Robertson et al., 2001), we have shown that this nerve growth factor binding ability is unique to TrkAd5 since the related domain TrkAd4 does not bind or sequester nerve growth factor. In addition, we describe here a competition ELISA in which TrkAd5 was able to sequestrate nerve growth factor with an IC50 of 40 nM. Using surface plasmon resonance, TrkAd5 binds nerve growth factor with an affinity of 94 pM, which is with a 10-fold greater affinity than to neurotrophin-3, and with more than 1000-fold greater affinity than to brain-derived neurotrophic factor or neurotrophin-4 (D. Dawbarn, J. J. Watson, M. S. Fahey, S. J. Tyler, and S. J. Allen, unpublished data).
The effectiveness of TrkAd5 in modulating bladder excitability was assessed in an established in vivo model (McMahon and Abel, 1987; Dmitrieva et al., 1997) of acute inflammatory cystitis. In our study, animals infused with turpentine into the bladder showed features typical of cystitis in humans, with clear changes in bladder activity recorded. Systemic injection of TrkAd5 proved effective in reversing all symptoms. Thus TrkAd5 is, to date, unique in being the only molecule to reverse the symptoms in this chemically induced model of cystitis, suggesting a clear clinical role.
The importance of nerve growth factor as a factor in the development of human allergic disease and asthma was initially suggested by the detection of increased serum levels of nerve growth factor in patients with these conditions (Bonini et al., 1996). A number of different cell types in the lung are known to be involved in production of nerve growth factor, including macrophages, T cells, mast cells, fibroblasts, and epithelial cells (Ehrhard et al., 1993; Burgi et al., 1996; Nilsson et al., 1997). In asthmatics, nerve growth factor has been detected in bronchoalveolar lavage fluid (Undem et al., 1999) with increase in levels with allergen challenge and in fluid and bronchial biopsies from ovalbumin-sensitized and challenged mice (Virchow et al., 1998). Nerve growth factor pretreatment of airways induces bronchial hyper-responsiveness in vivo in guinea pigs and in vitro on tracheal segments (De Vries et al., 1999, 2001).
Using isolated guinea pig tracheal rings, as demonstrated previously (De Vries et al., 2001), incubation with nerve growth factor induced an increased dose-related histamine hyper-responsiveness. We show here that addition of TrkAd5 before nerve growth factor exposure was able to block this hyper-responsiveness. Interestingly, in guinea pigs previously ovalbumin sensitized, TrkAd5 was able to moderate the direct allergen-induced tracheal contraction upon ovalbumin challenge. Moreover, 60 min following this ovalbumin challenge, the histamine-induced hyper-responsivity was inhibited in the rings that had received one dose of TrkAd5, preceding the ovalbumin challenge. This significant finding indicates that TrkAd5 is able to modify the airway late hyper-responsiveness to histamine. Presumably effects are because of the ability of TrkAd5 to sequester endogenous nerve growth factor, and it is of note that administration of TrkAd5 causes the contractile response to histamine to be lower than control after ovalbumin challenge.
Overall, these studies indicate that the TrkAd5 protein is stable in blood and urine and shows potent effects in animal models of cystitis and allergic asthma. Further studies will address pharmacokinetic issues and the clinical use of this protein for these and other disease states.
We thank Dr. Annick de Vries (University of Edinburgh, Edinburgh, Scotland) for advice on nerve growth factor in relation to asthma.
- Received September 20, 2005.
- Accepted November 10, 2005.
This work was supported by The Wellcome Trust. E.v.d.W. was financially supported by the Dutch Asthma Foundation Grant AF 3.2.01.50.
J.J.W. and M.S.F. contributed equally to this work.
ABBREVIATIONS: NGF, nerve growth factor; Trk, tyrosine receptor kinase; RT, room temperature; ELISA, enzyme-linked immunosorbent assay; BSA, bovine serum albumin; TBS-T, Tris-buffered saline/Tween 20; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine; MALDI-TOF, matrix-assisted laser desorption ionization/time of flight; HPLC, high-performance liquid chromatography; TCA, trichloroacetic acid.
- The American Society for Pharmacology and Experimental Therapeutics