Pergamon
PII:
Neuroscience Vol. 78, No. 2, pp. 449–459, 1997 Copyright ? 1997 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/97 $17.00+0.00 S0306-4522(96)00575-1
THE ROLE OF NERVE GROWTH FACTOR IN A MODEL OF VISCERAL INFLAMMATION N. DMITRIEVA,* D. SHELTON,† A. S. C. RICE,‡ and S. B. McMAHON*§ *Department of Physiology, St Thomas’ Hospital Medical School (UMDS), Lambeth Palace Road, London SE1 7EH, U.K. †Department of Neuroscience, Genentech Inc., South San Francisco, CA 94080, U.S.A. ‡Department of Anaesthetics, Imperial College School of Medicine at St Mary’s, St Mary’s Hospital, London W2 1PG, U.K. Abstract––There is growing evidence that nerve growth factor may be an important mediator of the sensory disorders associated with inflammation. This hypothesis was tested in a rat model of cystitis. In this model, an experimental inflammation is created in anaesthetized rats with an irritant chemical. Within 1 h, bladder reflexes, activated by the sensory innervation of this viscus, become exaggerated, mimicking the disorders seen in humans with chronic cystitis. The development of this hyper-reflexia following experimental inflammation was quantified using the technique of repeated cystometrograms. By several measures, bladder reflex excitability increased about three-fold after 5 h. Firstly, the study investigated whether inflammatory changes can be prevented by pharmacological antagonism of nerve growth factor. A synthetic fusion protein was used, consisting of the extracelluar domain of the high-affinity nerve growth factor receptor, trkA, coupled to the Fc portion of an immunoglobulin. Previous work has shown that this molecule can sequester nerve growth factor and reduce its bioavailability both in vitro and in vivo. Treatment of animals with the fusion molecule at 1 mg/kg, immediately before inflammation of the bladder, largely, and very significantly, prevented the expected increases in reflex excitability of this organ. Pretreatment with a related fusion protein, capable of sequestering the neurotrophin brain-derived neurotrophic factor and neurotrophin-4/5, but not nerve growth factor, was without effect. Similarly, a control fusion molecule, without neurotrophin-sequestering capacity, did not reduce the inflammationinduced hyper-reflexia. Systemic treatment with the nerve growth factor-sequestering molecule, but not control molecules, partially and significantly reversed established inflammatory changes, by about 30–60%, depending on outcome measure. The nerve growth factor-sequestering protein had no significant effects on bladder reflex excitability in the uninflamed state. It was also without significant effect on capsaicin-induced contractions of the urinary bladder. Administration of exogenous nerve growth factor into the lumen of the urinary bladders of normal anaesthetized rats produced a rapid and marked bladder hyper-reflexia similar to that seen with experimental inflammation. These findings are consistent with other circumstantial evidence that nerve growth factor may interact with visceral sensory systems. Together, the data strongly suggest that nerve growth factor produced in inflamed tissues is a critical mediator of the sensory disorders associated with inflammation. ? 1997 IBRO. Published by Elsevier Science Ltd. Key words: nerve growth factor, brain-derived neurotrophic factor, visceral inflammation, chronic cystitis, hyper-reflexia, bladder.
Nerve growth factor (NGF) is a secretory protein which is known to play a critical role in the development of the peripheral nervous system. Animals with a null mutation of the NGF gene12 or the gene §To whom correspondence should be addressed. Abbreviations: ANOVA, analysis of variance; BDNF, brain-derived neurotrophic factor; CD4-IgG, synthetic fusion protein (part of CD4 receptor fused to part of an immnuoglobulin); DMSO, dimethylsulphoxide; MT, micturition threshold; NC, number of micturition contractions; NGF, nerve growth factor; NT-4/5, neurotrophin4/5; RT-PCR, reverse transcriptase polymerase chain reaction; TCT, total micturition contraction time; trkAIgG, synthetic fusion protein (part of trkA receptor fused to part of an immnuoglobulin); trkB-IgG, synthetic fusion protein (part of trkB receptor fused to part of an immnuoglobulin). 449
encoding for the high-affinity NGF receptor, trkA,47 show at birth a loss of about 70% of primary sensory neurons, specifically the small-diameter neurons that are known to have predominantly unmyelinated axons and subserve mainly nociceptive functions. It is not surprising therefore that these animals appear profoundly analgesic. The same animals also exhibit an almost complete loss of sympathetic neurons. There is evidence that NGF has other biological effects on pain-signalling systems in developing animals. Thus, Ritter et al.43 reported that the presence of NGF was essential for the normal maturation of one specific class of nociceptors, the Aä high-threshold mechanoreceptor. In animals treated with neutralizing antibodies to NGF in the first and second postnatal week, this population
450
N. Dmitrieva et al.
Fig. 1. Example of normal reflex activity seen in rat urinary bladder and the hyper-reflexia that developed after inflammation with turpentine oil. Each panel shows a cystometrogram. A pump infused saline into the bladder (via a transurethral cannula) at a constant rate of 0.05 ml/min. The pressure within the bladder was constantly monitored. In the control state, increasing bladder volume initially caused a gradual increase in vesical pressure. At a critical point of distension a series of large reflex contractions of the bladder ensued. For each cystometrogram the threshold volume at which contractions commenced (the micturition threshold), the number of contractions in the 14 min of distension and the total time occupied by the contractions were measured. The top left trace shows the cystometrogram recorded in one animal in the control state. The other panels show cystometrograms recorded at various times after exposing this bladder to the inflammatory agent turpentine. Note the progressive increase in reflex excitability of the bladder evidenced by reduced MT, and increases NC and TCT.
of sensory neurons appeared to adopt a different physiological phenotype and became low-threshold mechanoreceptors. Less is known regarding the biological role of NGF in the adult animal. The current data indicate that NGF continues to exert actions on nociceptive systems. Thus, trkA receptors are still found on large numbers of, specifically, small-diameter primary sensory neurons6,37 and NGF protein continues to be expressed in the targets of these neurons.15,20,21,53 Recently, we have shown that constitutively produced NGF regulates the sensitivity of painsignalling systems. We used a recombinantly produced fusion molecule, consisting of part of the trkA receptor fused to part of an immunoglobulin (trkA-IgG),45 to sequester NGF in the skin of adult animals for 14 days. The nerves supplying the treated skin showed a reduced retrograde transport of NGF and the animals became progressively hypoalgesic to noxious thermal and chemical stimuli.37 These behavioural changes are paralleled by reduced sensitivity of cutaneous nociceptors to the same stimuli.7 There is growing evidence that NGF may be an important peripheral mediator of several inflammatory pain states. In particular, the cutaneous hyperalgesia associated with experimental inflammation produced by carrageenin and Freund’s adjuvant is blocked by the sequestration of NGF.27,37,58 In the
visceral domain there is as yet only circumstantial evidence that NGF may similarly function as an inflammatory mediator (see Discussion). In the present studies a well-characterized model of experimental inflammation of the urinary bladder was used.34,35 This model demonstrates many of the features of interstitial cystitis in humans. That is, micturition reflexes are greatly exaggerated and animals show signs of ongoing discomfort and referred hyperalgesia. Primary sensory neurons innervating the urinary bladder are known to be sensitized to peripheral stimuli, and these changes are likely to account for the observed sensory and reflex abnormalities.38,39 In the current experiments we ask whether the consequences of urinary bladder inflammation can be prevented (and reversed once established) by the pharmacological antagonism of NGF using the sequestering molecule trkA-IgG, and whether administration of exogenous NGF is capable of mimicking the effects of inflammation on bladder reflex excitability. EXPERIMENTAL PROCEDURES
Animal maintenance and surgery This study was carried out on 59 female Wistar rats, weighing 225–260 g. The animals were anaesthetized with a single dose of urethane (1.25 g/kg, i.p.), which produced
NGF in visceral inflammation
Fig. 2. Average changes in measures of reflex excitability of the urinary bladder (measured by cystometrograms as described in Fig. 1) in groups of animals treated with (i) 10% DMSO in saline for 1 h; (ii) 25% turpentine oil for 1 h; and (iii) NGF (20 µg/ml in 10% DMSO) for 1 h. Cystometrograms were determined in the control state and then 1, 3 and 5 h after instillation of the appropriate agent into the bladder lumen. Note the progressive increase in reflex excitability, by all outcome measures, after turpentine inflammation. NGF, but not vehicle, produced similar changes in bladder motility to slow filling. a stable level of anaesthesia lasting for the entire experiment. The trachea and one carotid artery were 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. Measurement of bladder motility The primary outcome measure in these experiments was bladder motility, assessed by cystometrograms. When the bladder is slowly filled, intravesical pressure gradually increases. At some volume (the micturition threshold, MT) a series of a large active contractions (micturition
451
Fig. 3. Effects of pretreatment with neurotrophinsequestering molecules on bladder reflex hyperexcitability associated with turpentine inflammation. The average changes in measures of reflex excitability of the urinary bladder are shown in groups of animals treated with turpentine for 1 h. (i) Turpentine alone; (ii) turpentine plus systemic pretreatment with 1 mg/kg CD4-IgG; (iii) turpentine plus pretreatment with 1 mg/kg trkB-IgG; and (iv) turpentine plus pretreatment with 1 mg/kg of trkA-IgG. Reflex excitability was determined in the control state and 1, 3 and 5 h after turpentine. Note that trkA-IgG, but not CD4-IgG and trkB-IgG, largely prevented the reflex hyperexcitability caused by turpentine inflammation, by all outcome measures. contractions) is elicited. These contractions are reflex in nature and are triggered by afferent activity originating in the bladder wall.55 They therefore represent one measure of afferent excitability. It is known that in the inflamed bladder these reflexes become hyperexcitable.1,30,35 Bladder motility was assessed by slow filling of the bladder with normal saline through a transurethral cannula, at 0.05 ml/min for 14 min (i.e. bladder volume was increased from 0 to 0.7 ml in this period). The filling was controlled by means of an infusion pump. The pressure within the bladder associated with this slow filling was monitored by a pressure transducer connected to a side arm of the filling catheter. The rate of filling used was within the physiological range.31 In normal animals, bladder pressure increased gradually for
452
N. Dmitrieva et al.
the first 5 min or so of filling, beyond which a series of regular micturition contractions were elicited, typically 2–5 within the 14-min cystometrogram. Only those animals which did not have any visible sign of bladder inflammation, had clear urine and showed normal baseline cystometrograms to two or three control fillings were chosen for further treatment. After control determinations, animals were subjected to one of several treatments (see below) and cystometrograms were subsequently undertaken at times 1, 3 and 5 h after the start of these treatments. These were recorded for later analysis. Three parameters of bladder responsiveness to distension were measured for each cystometrogram: (i) the number of active contractions during the 14-min cystometrogram (NC); (ii) the volume at which the first active contraction occurred (MT); and (iii) the total time occupied by active contractions during the 14-min filling period, i.e. the total contraction time (TCT). Inflammation and chemical treatments of the urinary bladder The bladders of some animals were subject to an experimental inflammation. In these animals, 0.5 ml of 50% turpentine oil was instilled into the bladder for 1 h, after which the turpentine was drained. This treatment produces a sterile inflammatory response with the invasion on immune cells and development of hyper-reflexia. The inflammation starts within 1 h of turpentine instillation and progressively increases over the next few hours.35 Since the principal hypothesis driving these experiments was that NGF produced by experimental inflammation was an important cause of the sensory changes, some animals, without inflammation, were subjected to exogenous NGF treatment to see whether this could simulate the changes seen with turpentine inflammation. Previous work13 showed that NGF dissolved in 10% dimethylsulphoxide (DMSO) was capable of sensitizing bladder afferents. Although NGF in saline was not systematically studied, its effect was less pronounced then the effect of NGF in DMSO. Therefore in this study some animals were treated with either 0.5 ml of 10% DMSO alone or 0.5 ml of 20 µg/ml NGF in 10% DMSO for 1 h, and then cystometrograms were performed in the subsequent hours. The NGF was human recombinant protein. Effects of neurotrophin-sequestering agents To test further the hypothesis that inflammatory changes were mediated by NGF, recombinantly produced fusion proteins were used, which consisted of the extracellular domain of high-affinity trk receptors fused with the fc portion of an IgG.45 We have previously shown that these ligand-trap molecules are capable of selectively blocking the survival-promoting effects of neurotrophins in an in vitro culture assay.37 We have also previously shown that the trkA-IgG, a selective blocker of NGF effects, has biological effects when given in vivo.37 This study used trkA-IgG to sequester NGF, trkB-IgG to sequester the neurotrophin brain-derived neurotrophic factor (BDNF) and neurotrophin-4/5 (NT-4/5) and a control recombinantly produced synthetic protein, consisting of part of CD4 receptor fused to part of an immunoglobulin (CD4-IgG), not known to have any antineurotrophin actions. The fusion proteins were given i.v. at a dose of 1 mg/kg either immediately before the administration of turpentine to the bladder to determine whether inflammatory changes could be prevented, or 3 h after turpentine treatment, when inflammatory changes were established to determine whether they might be reversed. The trkA-IgG-sequestering molecule was also used in some further control animals. In some of these the effects of the molecule on normal micturition reflexes (i.e. without chemical inflammation) were tested. In others, the effects of trkA-IgG on capsaicininduced bladder reflexes were tested. In this latter case, the
trkA-IgG was again given i.v. at 1 mg/kg, and 3 h later cystometrograms were undertaken using a 10 µM solution of capsaicin, dissolved in 10% ethanol in saline. Assessment of cystometrograms The primary purpose of these experiments was to assess the bladder hyper-reflexia that develops following turpentine treatment and ask whether this was attenuated by pharmacological antagonists of NGF and mimicked by exogenous NGF. Three outcome measures of excitability (MT, NC and TCT) were determined at each of four times (control, and 1, 3 and 5 h after treatment). To facilitate the analysis of treatments, for each animal the slope of the regression line derived from the measurements made at these four time-points was calculated for each outcome measure. Thus, in the case of turpentine inflammation alone, the NC in each cystometrogram typically increased from 4 in the control state, to 7 at 1 h, 10 at 3 h and 12 at 5 h. The average slope of the regression line through these data was 1.54 contractions/h. By comparing the slopes calculated in this way in different groups of animals, it was possible to assess whether turpentine-induced hyper-reflexia was modified by particular agents. The different treatment groups were compared by analysis of variance (ANOVA) and, in a limited number of cases, by unpaired t-tests. For all but one treatment, groups of five to 10 animals were used. The exception was a group of 20 animals used for turpentine treatment alone. All data are expressed as means&S.E. RESULTS
The control cystometrograms in all treatment groups were similar. The average NC was low (3.1&0.3, n=59), as was the average TCT (157.3&13.8 s, n=59), while baseline MTs were moderately high (0.45&0.03 ml, n=59). Examples of representative cystometrograms are shown in Figs 1 and 5. Cystometrograms after inflammation Turpentine oil instilled into the bladder produced hyper-reflexia. An example is shown in Fig. 1, and average data are presented in Fig. 2. NC increased at 1.73&0.17/h on average, TCT increased at 46.0&7.6 s/h and MT decreased at 0.049&0.008 ml/ h. In all cases, these changes were significant (i.e. the rate of change was significantly different from 0). In contrast, in uninflamed bladders (Fig. 2, left) the excitability of reflexes did not change significantly. Cystometrograms after intravesical nerve growth factor treatment Intravesical NGF produced a progressive hyperreflexia similar to that seen with intravesical turpentine oil. That is, NC increased at 2.02&0.65/h, TCT increased at 51.55&12.7 s/h and MT decreased at 0.085&0.024 ml/h. All of these increases were significantly different from 0 (P<0.05 in each case). The changes seen with NGF were not significantly different from those seen with turpentine (P>0.05 in each case).
NGF in visceral inflammation
453
Pretreatment with neurotrophin-sequestering molecules The effects of pretreatment with trkA-IgG, trkBIgG and the control fusion molecule CD4-IgG on the excitability changes produced by turpentine inflammation are summarized in Fig. 3. ANOVA revealed there were no differences between turpentine, CD4IgG and trkB-IgG groups (P=0.92 for NC, P=0.72 for TCT and P=0.18 for MT). However, pretreatment with trkA-IgG largely attenuated the hyperreflexia to turpentine by all measures (Fig. 3). The NC increased in the trkA-IgG group by only 0.62&0.23/h. This was only 36% of the increase seen with turpentine alone (P<0.01). The TCT increased in the trkA-IgG group by only 19.8&13.7 s/h. This was 43% of the turpentine alone value (P<0.05). Finally, the MT decreased at 0.0277&0.009 ml/h, only 56% of the value for turpentine alone. This last difference narrowly failed to reach statistical significance (P=0.062).
Posttreatment with neurotrophin-sequestering molecules In some animals, CD4-IgG or trkA-IgG was given systemically 3 h after a turpentine inflammation had been established, and bladder motility monitored for a further 2 h. With turpentine alone, the hyperexcitable state is well maintained during this phase. As shown in Fig. 4, CD4-IgG posttreatment did not lead to any change in excitability by any of the measures, i.e. the hyper-reflexia continued unabated (P>0.65 for all outcome measures). When trkA-IgG was given 3 h after turpentine, some of the hyper-reflexia that had developed was reversed. An example is shown in Fig. 5. By all three measures, trkA-IgG produced a significant reduction in reflex excitability within 30 min (Fig. 4). This antihyperalgesic effect was maintained for up to 2 h (Fig. 4). The reversal of hyper-reflexia was about 30–50% complete (depending on outcome measure; Fig. 4).
Effects of trkA-IgG on the excitability of the uninflamed bladder Control experiments were undertaken to determine whether the effects of trkA-IgG, described above could also be seen in the uninflamed state. In five animals, control cystometrograms were undertaken and trkA-IgG then given systemically at a dose of 1 mg/kg, as before. Bladder excitability was then redetermined 2 h later. In contrast to the results obtained for inflamed bladders, trkA-IgG did not produce any reduction in reflex excitability (Table 1). In fact, there were no significant changes in any of the outcome measures (P>0.05 in each case).
Fig. 4. Average changes in bladder reflex excitability produced by neurotrophin-sequestering molecules. For each of the outcome measures (NC, MT and TCT), the first two data points show the levels of bladder reflex excitability established in the control state and 3 h after inflammation with turpentine. After 3 h of inflammation, when bladder reflexes were enhanced, either trkA-IgG or CD4-IgG was given (1 mg/kg, i.v.). Cystometrograms were then determined at 0.5, 1 and 2 h subsequently. The degree of reflex excitability, as a percentage of the levels seen after 3 h of inflammation, was determined. Note that trkA-IgG, but not CD4-IgG, partially reversed the established hyper-reflexia, by all outcome measures.
Effects of trkA-IgG on capsaicin-induced contractions of the urinary bladder Capsaicin produces contractions when instilled into the normal bladder, and this is believed to occur because of a direct excitatory action of capsaicin on
454
N. Dmitrieva et al.
Fig. 5. Examples of cystometrograms recorded from one animal subjected to turpentine inflammation and, when a reflex hyperexcitable state was established, then treated with trkA-IgG (1 mg/kg, i.v.). The turpentine produced a progressive increase in bladder reflex excitability, which was partially reversed after trkA-IgG treatment.
Table 1. Effects of systemic trkA-IgG treatment on the motility of the normal bladder and on the hyper-reflexia induced by intravesical capsaicin (% control)
1. 2. 3. 4. 5. 6.
Control After trkA-IgG Control With capsaicin Control With capsaicin and trkA-IgG
Number of contractions
Micturition threshold
Total contraction time
100&33.3 140&38.6 100&50.4 176.9&15.4* 100&27.6 238.1&13.7*†
100&12.5 82.5&15 100&34.4 10.6&5.7* 100&18.1 32.1&18.6*†
100&38.3 158.5&55.4 100&51.2 212.8&20.9* 100&34.9 214.9&32.3*†
For each outcome measure (NC, MT and TCT), the level of bladder excitability was determined in the control state, and trkA-IgG was then given (1 mg/kg, i.v.). Two hours later, reflex excitability was redetermined. Note that in the uninflamed state, NGF sequestration did not reduce excitability (rows 1 and 2). Capsaicin instilled into the bladder produced a transient increase in excitability (see Table 1). Levels of excitability were determined in a control cystometrogram and then in one undertaken with 10 µM capsaicin in normal animals and animals pretreated with trkA-IgG (1 mg/kg, i.v., 2 h before first cystometrogram). Note that capsaicin produced similar changes in excitability with or without trkA-IgG (rows 3–6). *P<0.05 compared with control. †P>0.05 compared with capsaicin without trkA-IgG.
sensory neurons. This excitatory action was studied by undertaking cystometrograms, as above, with a 10 µM solution of capsaicin (Fig. 6). The first cystometrogram undertaken with capsaicin shows higher levels of reflex excitability than seen with saline. The outcome measures studied (NC, TCT, MT) were all increased two- to three-fold during the first cystometrogram with capsaicin (Table 1). These increases were significant (P<0.05 in all cases). Within 15–30 min of capsaicin treatment, the urinary
bladder become hypo-reflexic (Fig. 6), presumably because of the desensitization of sensory neurons that follows the initial excitation. The effects of capsaicin were evaluated in an identical fashion in other animals that had received trkA-IgG systemically (1 mg/ kg) 2 h earlier. In these animals, capsaicin still induced an immediate hyper-reflexia, followed by desensitization. The magnitude of the hyper-reflexia (measured as NC, TCT and MT during the first cystometrogram with capsaicin) was not significantly
NGF in visceral inflammation
455
Fig. 6. Examples of cystometrograms from two animals, one without (A) and one pretreated with trkA-IgG (B). For each animal, a control cystometrogram, a cystometrogram performed filling the bladder with 10 µM capsaicin, and a cystometrogram 25 min. after exposing the bladder to capsaicin are shown. Note that capsaicin initially enhanced reflex excitability of the bladder, and then produced some desensitization. TrkA-IgG treatment did not affect the responses to capsaicin.
different from that seen in the control (non-trkAIgG-treated) animals (Table 1; P>0.1 in all cases). DISCUSSION
In these experiments we set out to test the hypothesis that NGF functions as a mediator of sensory disorders associated with inflammation. A welldescribed model of cystitis in the rat was used.17,18,19,33,35 Bladder reflex activity was studied, rather than, for instance, single afferent fibre activity, because it offers a robust measure of the integrated neuronal response to inflammation. Repeated measures can be made, allowing the time-course of effects to be conveniently studied. It is well recognized that micturition reflexes are triggered by afferent activity from sensory neurons innervating the urinary bladder and running in the pelvic nerve.55 Reflex excitability was therefore used as an indirect measure of the excitability of sensory systems. Previous work in this model showed that bladder reflex hyperexcitability is well correlated with signs of ongoing abdominal discomfort and referred cutaneous hyperalgesia in unanaesthetized animals.35 Moreover, for this viscus in particular, the electrophysiological evidence suggests that one rather homogeneous population of primary sensory neurons encodes not only low levels of bladder distension associated with motility regulation, but also high levels of distension, beyond the normal physiological range that is likely to signal pain.39
With these techniques, local administration of exogenous NGF into the lumen of the bladder has been shown to be capable of inducing bladder reflex hyperexcitability which is very similar to that seen with turpentine inflammation, in terms of both its magnitude and latency of onset. This observation is consistent with our hypothesis on the role of NGF. The NGF was mixed with DMSO since in pilot experiments it was found that NGF applied without DMSO was less effective. With this mode of administration, the NGF is likely to be acting locally on structures within the bladder, since much larger doses (in the order of 1 mg/kg) appear necessary for acute systemic effects on sensory neurons.26,27,28 These effects of exogenous NGF on bladder motility are also consistent with other work. Using an identical means of delivery, NGF can induce a modest plasma extravasation into the bladder with a similarly short latency.13 The extravasation is caused by an effect of NGF on primary sensory neurons rather than on vascular smooth muscle or efferent innervation, since it is absent in capsaicin-treated animals (unpublished observations). Electrophysiological recordings from bladder afferents have also shown that local exogenous NGF treatment at the same dose leads to rapid activation and sensitization of nearly all bladder afferents projecting through the pelvic nerve.13 Such sensitization would readily explain the changes in motility seen in the current work. The other major finding of this study was that the changes in bladder reflex excitability produced by
456
N. Dmitrieva et al.
turpentine inflammation can be blocked by the trkAIgG fusion protein. The reflex excitability of the normal bladder, however, was unaffected by acute trkA-IgG treatment. Acute treatment with trkAIgG also failed to block the effects of the excitant capsaicin, further arguing against a non-specific action of trkA-IgG. Together, these results provide direct support for our hypothesis on the importance of NGF as an inflammatory mediator. Since trkAIgG given immediately before turpentine was able to block nearly all of the expected inflammatory changes in bladder motility, it seems that NGF is of critical importance for these sensory disorders. The IgG fusion molecules are large and do not cross the blood–brain barrier, indicating a peripheral site of action. However, preliminary pharmacokinetic studies indicate that i.v. administration of 1 mg/kg trkA-IgG leads to serum levels between 5 and 20 µg/ ml (data not shown) during the duration of these experiments, a level above that required for effectively blocking NGF biological activity in vitro.37 The very size of the molecule is also likely to present an access problem to interstitial space when it is given i.v. However, the developing inflammatory process will itself aid delivery of the IgGs to the relevant site, since increased vascular permeability is a cardinal sign of inflammation. We did not try in the current work to block the in vivo effects of NGF on bladder motility with trkA-IgG because of the uncertainty of penetration of the IgG molecule into uninflamed tissue. However, in vitro studies have shown that trkA-IgG is a very potent and highly selective antagonist of NGF.38,46 There are other reasons for believing NGF that may be an important inflammatory mediator. Firstly, NGF levels are elevated in inflammation. In the model used here the levels of NGF mRNA, determined by both in situ hybridization and reverse transcriptase polymerase chain reaction (RT-PCR), were increased in the inflamed bladder.5,41 There was also an increase in NGF protein levels in this model. The time-course of NGF up-regulation parallels the sensory abnormalities occurring in this model. Unfortunately, our in situ studies have not provided clear answers as to which cell type(s) are responsible for increased NGF production, although in general the greatest changes were seen immediately below the uro-epithelium in the mucosal layer of the bladder. The literature suggests that there are many potential cell types capable of producing NGF, including fibroblasts, leukocytes, mast cells, Schwann cells of peripheral nerve trunks and even smooth muscle cells of the urinary bladder.9,20–22,25,29,51 Visceral afferents in particular may be sensitive to NGF, since nearly all pelvic afferents are known to express the highaffinity NGF receptor trkA (whereas only some 45% of somatic afferents have this receptor).6,36 Interestingly, sacral visceral afferents also express the trkB receptor36 and BDNF levels were recently found to be elevated in our model of inflammation.41 The
functional significance of these findings is unclear, however, since in the present study trkB-IgG did not alter motility changes seen with inflammation. It is possible that BDNF (and indeed NGF) could exert actions at times longer than those studied here, following retrograde transport of the neurotrophins to the sensory neuron cell bodies. A recent study50 reports that increased NGF production in urinary bladder in a model of outlet obstruction is responsible for long-term changes in sensory neuron soma size. The same report also provides evidence that NGF can produce bladder reflex excitability in that model. TrkA-IgG produced a partial reversal of established hyper-reflexia. This suggests a continuing role of NGF in the maintenance of sensory disorders. It is known that NGF up-regulation is maintained in this model (see above). However, the incomplete reversal with posttreatment may indicate that some other process now mediates the inflammatory disorders, or that they become independent of their precipitating cause. Alternatively, it may simply be that NGF availability would need to be blocked for a longer period than was possible in these experiments to reverse the changes completely. Clinical relevance In this study, a model of experimental inflammation induced by turpentine oil was investigated. It is reasonable to ask how relevant these findings might be to human pathology. Certainly the disorders seen in this rat model parallel in many ways those found in conditions of chronic cystitis is humans, the principal symptoms of which are increased bladder frequency, reduced capacity, ongoing pain and pain on micturition. In some cases referred pain and hyperalgesia are present. The clinical condition has an unknown aetiology. However, there are counterparts of all the symptoms in our rat model, and these manifest themselves within a few hours. It is not implausible that NGF up-regulation is a common mechanism, since it is clear that NGF levels are increased in a wide variety of inflammatory states.3,4,14,57,58 Our model may also have relevance to another clinical disorder of unknown aetiology, interstitial cystitis. This chronically painful disease is estimated to affect some 270,000–450,000 women in the U.S.A. alone.44 Histological analysis of bladder biopsies shows pathology of mucosa and detrusor layers of bladder wall. Oedema and vasodilatation are present, associated with chronic inflammatory cell infiltrate.24,42,44 The number of mast cells present is elevated and many are degranulated.2,11,52 Mastocytosis has been proposed as a diagnostic tool of interstitial cystitis.44 Our experimental model also shows prominent signs of oedema and inflammatory cell infiltrate.35 There is also clear evidence for interactions between NGF and mast cells in other
NGF in visceral inflammation
contexts. Thus, mast cells may express receptors for NGF and are strongly degranulated by this neurotrophin.10,23,25,54 In addition, NGF is able to induce mast cell proliferation.8,32,48 Moreover, mast cell products appear to be important intermediary agents in the cutaneous hyperalgesia that develops to exogenous NGF administration.26–28 However, mast cells are also a source of NGF in peripheral tissue.25 This NGF could be immediately delivered in a paracrine manner to the sensory nerve endings which are found in close proximity to mast cells.8,40,49 Physiological interactions between mast cells and sensory neurons are also supported by the fact that mast cells degranulate in the presence of substance P, a neuropeptide responsible for the efferent functions of primary afferent nociceptors.16,46,47
457 CONCLUSIONS
Taken together, these findings suggest that our experimental observations may be relevant to some painful disorders of the urinary bladder in humans. The implication is that anti-NGF strategies may be of some therapeutic benefit in treating these clinical disorders.
Acknowledgements—This work was supported by a grant from the MRC of Great Britain. ND is supported by a studentship from CONICIT. We would also like to thank Caroline Abel, Tabitha Springhall and Sridhar Viswanathan for excellent technical help, and Paul Seed for statistical advice.
REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
Abelli L., Conte B., Somma V., Maggi C. A., Giuliani S. and Meli A. (1989) A method for studying pain arising from the urinary bladder in conscious, freely-moving rats. J. Urol. 141, 148–151. Aldenborg F., Fall M. and Enerback L. (1986) Proliferation and transepithelial migration of mucosal mast cells in interstitial cystitis. Immunology 58, 411–416. Aloe L., Probert L., Kollias G., BracciLaudiro L., Micera A., Mollinari C. and Levi-Montalcini R. (1993) Level of nerve growth factor and distribution of mast cells in the synovium of tumour necrosis factor transgenic arthritic mice. Int. J. Tissue React. 15, 139–143. Aloe L., Tuveri M. A., Carcassi U. and Levi-Montalcini R. (1992) Nerve growth factor in the synovial fluid of patients with chronic arthritis. Arthrit. Rheum. 35, 351–355. Andreev N. Y., Bennett D., Priestley J., Rattray M. and McMahon S. B. (1994) NGF mRNA is increased by experimental inflammation of adult rat urinary bladder. Soc. Neurosci. Abstr. 19, 108. Averill S., McMahon S. B., Clary D. O., Reichardt L. F. and Priestley J. V. (1995) Immunocytochemical localization of trkA receptors in chemically identified subgroups of adult rat sensory neurons. Eur. J. Neurosci. 7, 1484–1494. Bennett D. L., McMahon S. B., Shelton D. and Koltzenberg M. (1996) NGF specifically modulates the thermal and chemical sensitivity of primary afferent nociceptors. VIIIth World Congress on Pain Abstr., p. 120. Bienenstock J., Tomioka M., Matsuda H., Stead R. H., Quinonez G., Simon G. T. and Denburg J. A. (1987) The role of mast cells in inflammatory processes: evidence for nerve/mast cell interactions (Review). Int. Arch. Allergy appl. Immun. 82, 238–243. Brown M. C., Perry V. H., Lunn E. R., Gordon S. and Heumann R. (1991) Macrophage dependence of peripheral sensory nerve regeneration: possible involvement of nerve growth factor. Neuron 6, 359–370. Bruni A., Bigon E., Boarato E., Mietto L., Leon A. and Toffano G. (1982) Interaction between nerve growth factor and lysophosphatidylserine on rat peritoneal mast cells. Fedn Eur. biochem. Socs Lett. 138, 190–192. Christmas T. J. and Rode J. (1991) Characteristics of mast cells in normal bladder, bacterial cystitis and interstitial cystitis. Br. J. Urol. 68, 473–478. Crowley C., Spencer S. D., Nishimura M. C., Chen K. S., Pitts-Meek S., Armanini M. P., Ling L. H., McMahon S. B., Shelton D. L. and Levinson A. D. (1994) Mice lacking nerve growth factor display perinatal loss of sensory and sympathetic neurons yet develop basal forebrain cholinergic neurons. Cell 76, 1001–1011. Dmitrieva N. and McMahon S. B. (1996) Sensitization of visceral afferents by nerve growth factor in the adult rat. Pain, in press. Donnerer J., Schuligoi R. and Stein C. (1992) Increased content and transport of substance P and calcitonin gene-related peptide in sensory nerves innervating inflamed tissue: evidence for a regulatory function of nerve growth factor in vivo. Neuroscience 49, 693–698. Ebendal T., Olson L. and Seiger A. (1983) The level of nerve growth factor (NGF) as a function of innervation. A correlation radio-immunoassay and bioassay study of the rat iris. Expl Cell Res. 148, 311–317. Erjavec F., Lembeck F., Florjanc-Irman T., Skofitsch G., Donnerer J. and Holzer P. (1981) Release of histamine by substance P. Naunyn-Schmiedebergs Arch. Pharmac. 317, 67–70. Habler H. J., Janig W. and Koltzenburg M. (1988) A novel type of unmyelinated chemosensitive nociceptor in the acutely inflamed urinary bladder. Agents Actions 25, 219–221. Habler H. J., Janig W. and Koltzenburg M. (1990) Activation of unmyelinated afferent fibres by mechanical stimuli and inflammation of the urinary bladder in the cat. J. Physiol., Lond. 425, 545–562. Habler H. J., Janig W. and Koltzenburg M. (1993) Receptive properties of myelinated primary afferents innervating the inflamed urinary bladder of the cat. J. Neurophysiol. 69, 395–405. Heumann R. (1987) Regulation of the synthesis of nerve growth factor. J. exp. Biol. 132, 133–150. Heumann R., Korsching S., Bandtlow C. and Thoenen H. (1987) Changes of nerve growth factor synthesis in non-neuronal cells in response to sciatic nerve transection. J. Cell Biol. 104, 1623–1631. Heumann R., Lindholm D., Bandtlow C., Meyer M., Radeke M. J., Misko T. P. and Thoenen H. (1987) Differential regulation of mRNA encoding nerve growth factor and its receptor in rat sciatic nerve during development, degeneration, and regeneration: role of macrophages. Proc. natn Acad. Sci. U.S.A. 84, 8735–8739.
458 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57.
N. Dmitrieva et al. Horigome K., Pryor J. C., Bullock E. D. and Johnson E. M. Jr. (1993) Mediator release from mast cells by nerve growth factor. Neurotrophin specificity and receptor mediation. J. biol. Chem. 268, 14881–14887. Koziol J. A. (1994) Epidemiology of interstitial cystitis. In The Urologic Clinics of North America: Interstitial Cystitis (ed. Hanno P. M.), p. 7. Leon A., Buriani A., Dal Toso R., Fabris M., Romanello S. and Aloe L. (1994) Mast cells synthesize, store, and release nerve growth factor. Proc. natn Acad. Sci. U.S.A. 91, 3739–3743. Lewin G. R. and Mendell L. M. (1993) Nerve growth factor and nociception. Trends Neurosci. 16, 353–359. Lewin G. R., Ritter A. M. and Mendell L. M. (1993) Nerve growth factor-induced hyperalgesia in the neonatal and adult rat. J. Neurosci. 13, 2136–2148. Lewin G. R., Rueff A. and Mendell L. M. (1994) Peripheral and central mechanisms of NGF-induced hyperalgesia. Eur. J. Neurosci. 6, 1903–1912. Lindholm D., Heumann R., Hengerer B. and Thoenen H. (1988) Interleukin 1 increases stability and transcription of mRNA encoding nerve growth factor in cultured rat fibroblasts. J. biol. Chem. 263, 16348–16351. Maggi C. A., Abelli L., Giuliani S., Santicioli P., Geppetti P., Somma V. and Meli A. (1988) The contribution of sensory nerves to xylene-induced cystitis in rats. Neuroscience 26, 709–723. Maggi C. A. and Meli A. (1983) Reserpine-induced detrusor hyper-reflexia: an in vivo model for studying smooth muscle relaxants at urinary bladder level. J. pharmac. Meth. 10, 79–91. Marshall J. S., Stead R. H., McSharry C., Nielsen L. and Bienenstock J. (1990) The role of mast cell degranulation products in mast cell hyperplasia. I. Mechanism of action of nerve growth factor. J. Immun. 144, 1886–1892. Mazurek N., Weskamp G., Erne P. and Otten U. (1986) Nerve growth factor induces mast cell degranulation without changing intracellular calcium levels. Fedn Eur. biochem. Socs Lett. 198, 315–320. McMahon S. B. (1988) Neuronal and behavioural consequences of chemical inflammation of rat urinary bladder. Agents Actions 25, 231–233. McMahon S. B. and Abel C. (1987) A model for the study of visceral pain states: chronic inflammation of the chronic decerebrate rat urinary bladder by irritant chemicals. Pain 28, 109–127. McMahon S. B., Armanini M. P., Ling L. H. and Phillips H. S. (1994) Expression and coexpression of Trk receptors in subpopulations of adult primary sensory neurons projecting to identified peripheral targets. Neuron 12, 1161–1171. McMahon S. B., Bennett D. L. H., Priestley J. V. and Shelton D. L. (1995) The biological effects of endogenous nerve growth factor on adult sensory neurons revealed by a trkA-IgG fusion molecule. Nat. Med. 1, 774–780. McMahon S. B., Dmitrieva N. and Koltzenburg M. (1995) Visceral pain. Br. J. Anaesth. 75, 132–144. McMahon S. B. and Koltzenburg M. (1993) Changes in the afferent innervation of the inflamed urinary bladder. In Basic and Clinical Aspects of Chronic Abdominal Pain (eds Mayer E. A. and Raybould H.), p. 155. Elsevier, Amsterdam. Newson B., Dahlstrom A., Enerback L. and Ahlman H. (1983) Suggestive evidence for a direct innervation of mucosal mast cells. Neuroscience 10, 565–570. Oddiah D., McMahon S. B. and Rattray M. (1995) Inflammation produced upregulation of neurotrophin messenger RNA levels in bladder. Soc. Neurosci. Abstr. 21, 1534. Ratliff T. L., Klutke C. G. and McDougall E. M. (1994) The etiology of interstitial cystitis. In The Urologic Clinics of North America: Interstitial Cystitis (ed. Hanno P. M.), p. 21. Ritter A. M., Lewin G. R., Kremer N. E. and Mendell L. M. (1991) Requirement for nerve growth factor in the development of myelinated nociceptors in vivo. Nature 350, 500–502. Sant G. R. and Theoharides T. C. (1994) The role of the mast cell in interstitial cystitis. In The Urologic Clinics of North America: Interstitial Cystitis (ed. Hanno P. M.), p. 41. Shelton D. L., Sutherland J., Gripp J., Camerato T., Armanini M. P., Phillips H. S., Carroll K., Spencer S. D. and Levinson A. D. (1995) Human trks: molecular cloning, tissue distribution, and expression of extracellular domain immunoadhesins. J. Neurosci. 15, 477–491. Skofitsch G., Savitt J. M. and Jacobowitz D. M. (1985) Suggestive evidence for a functional unit between mast cells and substance P fibers in the rat diaphragm and mesentery. Histochemistry 82, 5–8. Smeyne R. J., Klein R., Schnapp A., Long L. K., Bryant S., Lewin A. and Lira S. A. (1994) Severe sensory and sympathetic neuropathies in mice carrying a disrupted Trk/NGF receptor gene. Nature 368, 246–249. Stead R. H., Bienenstock J. and Stanisz A. M. (1987) Neuropeptide regulation of mucosal immunity. Immun. Rev. 100, 333–359. Stead R. H., Dixon M. F., Bramwell N. H., Riddell R. H. and Bienenstock J. (1989) Mast cells are closely apposed to nerves in the human gastrointestinal mucosa. Gastroenterology 97, 575–585. Steers W. D., Creedon D. J. and Tuttle J. B. (1996) Immunity to nerve growth factor prevents afferent plasticity following urinary bladder hypertrophy. J. Urol. 155, 379–385. Steers W. D., Kolbeck S., Creedon D. and Tuttle J. B. (1991) Nerve growth factor in the urinary bladder of the adult regulates neuronal form and function. J. clin. Invest. 88, 1709–1715. Theoharides T. C., Bondy P. K., Tsakalos N. D. and Askenase P. W. (1982) Differential release of serotonin and histamine from mast cells. Nature 297, 229–231. Thoenen H., Bandtlow C., Heumann R., Lindholm D., Meyer M. and Rohrer H. (1988) Nerve growth factor: cellular localization and regulation of synthesis. Cell. molec. Neurobiol. 8, 35–40. Tomioka M., Stead R. H., Nielsen L., Coughlin M. D. and Bienenstock J. (1988) Nerve growth factor enhances antigen and other secretagogue-induced histamine release from rat peritoneal mast cells in the absence of phosphatidylserine. J. Allergy clin. Immun. 82, 599–607. Torrens M. and Morrison J. F. B. (1987) The Physiology of the Lower Urinary Tract. Springer, London. Urfer R., Tsoulfas P., O’Connell L., Shelton D. L., Parada L. F. and Presta L. G. (1995) An immunoglobulin-like domain determines the specificity of neurotrophin receptors. Eur. molec. Biol. Org. J. 14, 2795–2805. Weskamp G. and Otten U. (1987) An enzyme-linked immunoassay for nerve growth factor (NGF): a tool for studying regulatory mechanisms involved in NGF production in brain and in peripheral tissues. J. Neurochem. 48, 1779–1786.
NGF in visceral inflammation 58.
459
Woolf C. J., Safieh-Garabedian B., Ma Q. P., Crilly P. and Winter J. (1994) Nerve growth factor contributes to the generation of inflammatory sensory hypersensitivity. Neuroscience 62, 327–331. (Accepted 16 October 1996)