Systemic administration of anti-NGF increases A-type potassium currents and decreases pancreatic nociceptor excitability in a rat model of chronic pancreatitis

Abstract

We have previously shown that pancreatic sensory neurons in rats with chronic pancreatitis (CP) display increased excitability associated with a decrease in transient inactivating potassium currents (I_A), thus accounting in part for the hyperalgesia associated with this condition. Because of its well known role in somatic hyperalgesia, we hypothesized a role for the nerve growth factor (NGF) in driving these changes. CP was induced by intraductal injection of trinitrobenzene sulfonic acid (TNBS) in rats. After 3 wk, anti-NGF antibody or control serum was injected intra-peritoneally daily for 1 wk. This protocol was repeated in another set of experiments in control rats (receiving intraductal PBS instead of TNBS). Pancreatic nociceptors labeled with the dye Dil were identified, and patch-clamp recordings were made from acutely dissociated DRG neurons. Sensory neurons from anti-NGF-treated rats displayed a lower resting membrane potential, increased rheobase, decreased burst discharges in response to stimulatory current, and decreased input resistance compared with those treated with control serum. Under voltage-clamp condition, neuronal I_A density was increased in anti-NGF-treated rats compared with rats treated with control serum. However, anti-NGF treatment had no effect on electrophysiological parameters in neurons from control rats. The expression of Kv-associated channel or ancillary genes Kv1.4, 4.1, 4.2, 4.3, and DPP6, DPP10, and KCHIPs 1–4 in pancreas-specific nociceptors was examined by laser-capture microdissection and real-time PCR quantification of mRNA levels. No significant differences were seen among those. These findings emphasize a key role for NGF in maintaining neuronal excitability in CP specifically via downregulation of I_A by as yet unknown mechanisms.

a cardinal symptom of chronic pancreatitis (CP) is pain, which is also one of the most difficult to treat, a fact that reflects our inadequate understanding of its pathogenesis (1). In recent years, however, some progress has been made in this area, aided by the development of a robust and reproducible rat model of painful CP with construct, face, and predictive validity (19). Using this model, we have demonstrated that CP results in the sensitization of pancreatic nociceptors, with increased spontaneous firings and excitability (22). We therefore postulated that, as with other painful inflammatory conditions, CP caused an increased “afferent barrage” from sensitized primary neurons with amplification and persistence of pain (2). Our studies also suggested an ionic basis for these changes as we found a significant reduction in transient A-type potassium currents (I_A) in pancreas-specific nociceptive neurons (22).

In this study, our aim was to determine the role of nerve growth factor (NGF) in driving the electrophysiological changes in nociceptive neurons in CP. NGF is a logical and attractive candidate in this regard based on its known effects on neuronal sensitization via modulation of several ion channels important for both signal transduction (such as TRPV1) and excitability (Kv and Nav currents) (5, 9, 12, 21). Furthermore, resected specimens from patients with CP also have increased expression of NGF and its high-affinity receptor, TrkA, in a manner that correlates with pain (8). We have previously shown that NGF is significantly upregulated in our model of CP (23), and most recently, we demonstrated that antagonism of NGF suppressed the hyperalgesia associated with CP, along with a reduction in TRPV1 expression and currents (27). In this paper, we report the neutralizing antibody treatment reverses the changes of I_A and gene expression in pancreas specific sensory neurons of CP rats.

MATERIALS AND METHODS

Induction of CP and cell labeling.

CP was induced in adult male Sprague-Dawley rats as previously described (19). Care and handling of animals were approved by the Institutional Animal Care and Use Committee at Stanford University and were in accordance with the guidelines of the International Association for the Study of Pain. In brief, the common bile duct was temporarily occluded at the hepatic end, and 0.5 ml of TNBS solution (TNBS 6 mg/ml in 10% ethanol and PBS, pH 8.0) was slowly infused into the pancreatic duct through a 30-gauge needle connected to the PE-10 tubing via the duodenal papilla. For control rats, PBS was substituted for TNBS. Postoperatively, animals were monitored and treated for pain.

For experiments involving electrophysiology, the pancreas was injected with the lipid soluble fluorescence dye 1,1′-dioleyl-3,3,3′,3-tetramethylindocarbocyanine methanesulfonate (DiI; Molecular Probes, Eugene, OR), 25 mg in 0.5 ml of methanol at 8–10 sites on the exposed pancreas in 2-μl volumes, before TNBS and PBS infusion. For experiments involving laser-capture micro-dissection (LCM) to isolate mRNA for gene expression analysis, pancreas-specific DRG neurons were retrogradely labeled with cholera toxin B (CTB) (Molecular Probes, Eugene, OR), 2 mg/ml in PBS at 8–10 sites in 2-μl volumes per site in a second surgical procedure 2 wk after TNBS injection.

Anti-NGF treatment.

Three weeks post-TNBS injection, rats were injected either with 16 μg/kg body wt of neutralizing polyclonal NGF antibody in 0.5 ml of PBS (R & D Systems, Minneapolis, MN) or an equal volume of nonimmune normal goat serum (MP Biomedicals, Solon, OH) daily by the intraperitoneal route for 7 days. This concentration of NGF antibody has been successfully used for relief of pain in another model of visceral hyperalgesia (20).

Histological assessment.

The severity of CP was assessed by a previously described histological grading system that uses a semi-quantitative score on a scale of 0–3 for glandular atrophy, fibrosis, and inflammatory infiltrate (7, 16).

Electrophysiological study of pancreatic nociceptors.

At day 7 following injection of anti-NGF or vehicle [normal goat serum (NGS)], DRG neurons were harvested as described previously (22, 23). Briefly, after decapitation and spinal cord exposure, DRGs were bilaterally dissected out along T9–T13 and trimmed in ice-cold MEM supplemented with penicillin-streptomycin (GIBCO, Grand Island, NY), then transferred into HBS solution containing type 2 collagenase (Worthington, Lakewood, NJ) and incubated for 1.5 h at 36.5°C. A single-cell suspension was obtained by trituration, and dissociated cells were centrifuged twice (270 g for 5min). After resuspending the pellet in neurobasal media (GIBCO), cells were plated in poly-l-ornithine coated dishes (P35G-1.5-14-C, MatTek) and incubated for 3 h, and DiI-labeled neurons were identified on the patch-clamping stage by brief exposure to Lambda XL (Sutter, San Diego, CA) light using an inbuilt fluorescence microscope and a rhodamine band-pass filter (excitation wave length 530–560 nm; emission 573–648 nm; Nikon TE200). Cells were continuously superfused (1.5 ml/min) at room temperature with normal external solution containing (in mM) 135 NaCl, 5.4 KCl, 0.1 NaH₂PO₃, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 10 glucose, pH adjusted to 7.4 with NaOH (295–310 mosM). Recording pipettes typically had a resistance of 2.5–4 MΩ for whole cell when filled with normal external solution. The pipette was back filled with solution containing (in mM) 115 K-gluconate, 25 KCl, 5 NaCl, 10 HEPES, 1 CaCl₂, 2 ATP-Mg₂, and 1.12 EGTA, pH 7.25, mosM 290. In experiments that required eliminating Na⁺ current and K_Ca current, [Na⁺]_o was substituted by equimolar choline and [Ca²⁺]_o was reduced to 20 μM, and intracellular EGTA was increased to 11.2 mM to achieve intracellular free Ca²⁺ below 1 × 10⁻⁸. Signals were acquired using an Axopatch 200B amplifier and digitized with a Digidata 1200 (Axon Instruments, San Jose, CA) by setting with low-pass, four-point Bessel filtered at 2 or 5 kHz and then digitized at 5 or 20 kHz. Data were stored and analyzed offline. Voltage or current commands were delivered to the amplifier under computer control using pClamp 8 (Axon Instruments). Access resistance and cell membrane resistance, capacitance, and time constants were monitored by software programmed switching to the pClamp membrane test protocol. Analysis and interpretation of results were done while blinded to the nature of the treatment that the rats received.

Laser capture of pancreatic sensory neurons and measurement of mRNA.

Rats injected with CTB into the pancreas were perfused transcardially with 150 ml of ice-cold Tyrodes's solution containing 5 U/ml heparin. T9 and T10 DRGs from both sides were snap frozen in Tissue-Tek O.C.T. on dry ice. Ten-micrometer frozen sections were fixed for 1 min in 75% ethanol and washed in 50% ethanol and DEPC, followed by consecutive rehydration in 75, 95, and 100% and xylene. Sections were air dried under fume hood. Pancreas-specific afferent neurons were identified under fluorescent light using a FITC filter. Laser capture was performed on a Arcturus Pixel IIe system (Applied Biosystems, Foster City, CA).

Total RNA was obtained from laser-captured DRG sections from CP rats treated with either anti-NGF or control serum using a Qiagen RNA kit per manufacturer's instructions (Qiagen, Valencia, CA). RNA from T9 and T10 DRG segments was then combined before cDNA synthesis and preamplified for 5, 10, and 14 cycles. Pre-amplified cDNA was diluted 1:20 in ×1 TE before quantitative real-time measurements for the genes Kv1.4, 4.1, 4.2, and 4.3, and DPP6, DPP10, KCHIPs 1–4, GAPDH, β-actin, and β-tubulin per manufacturer's instructions (Invitrogen). Changes in mRNA levels of test genes were calculated by the DDCt method using GAPDH as a normalizer and expressed relative to the average DDCt of the control serum-treated group for each gene of interest.

Data analysis.

Data from patch-clamp experiments were analyzed by pClamp 9 and Origin7. Results were expressed as means ± SE, with n being the number of cells. The paired Student's t-test was used to evaluate differences between mean values. P values of ≤0.05 were considered to indicate a statistically significant difference. Gene expression data and immunohistochemistry were analyzed by statistical tests (t-tests for comparisons of means and χ² tests for comparisons of proportions) using GraphPad Prism 5 (GraphPad Software, La Jolla, CA).

RESULTS

Anti-NGF treatment does not affect inflammation in rats with TNBS-induced CP.

The effects of anti-NGF treatment on pancreatic morphology in rats with TNBS-induced CP were assessed by examination of H&E stained tissue sections. As shown in Fig. 1, intraductal TNBS induced typical morphological changes of CP that was unaffected by anti-NGF treatment. The mean histological scores were 7.3 ± 0.99 in rats treated with anti-NGF compared with 8.5 ± 0.34 in rats treated with normal goat serum (P = 0.47; n = 6 in each group).

Fig. 1.

Histological assessment of the pancreas following various treatments. H & E stained pancreatic tissue at 3 wk postinfusion followed by 1 wk of additional treatment as indicated. A: pancreas from a rat receiving intraductal PBS. B: pancreas from a rat with trinitrobenzene sulfonic acid (TNBS)-induced chronic pancreatitis (CP), which was treated with normal goat serum (NGS) C: pancreas from a rat with TNBS-induced CP, which was treated with anti-nerve growth factor (NGF) treatment. D: mean histological scores in the CP groups treated with NGS or anti-NGF.

Anti-NGF treatment results in suppression of pancreas-specific neuronal excitability in rats with CP.

Using patch-clamp techniques, we next examined the excitability of DiI-labeled DRG neurons, ∼15–30 μM in size, from rats with TNBS-induced CP. After positive current injection under current-clamp mode, burst discharges (action potentials) were frequently recorded in neurons from rats treated with normal goat serum but only occasionally noted in neurons from rats treated with anti-NGF (Fig. 2). Neuronal membrane properties were also measured (Fig. 3), with significant changes in several key parameters. Resting membrane potential (RMP) was more negative in pancreatic sensory neurons from the anti-NGF group compared with the NGS group (−55.25 ± 2.03 mV, n = 12 vs. −48.9 ± 2.21 mV, n = 11; P < 0.05); the mean rheobase was 0.44 ± 0.09 nA in the anti-NGF group (n = 13) compared with 0.14 ± 0.04 nA in the NGS group (n = 16; P < 0.001); the mean number of spikes elicited by ×2 rheobase current was also lower (1.36 ± 0.4 in the anti-NGF group vs. 3.0 ± 1.79 in the NGS group, n = 11 each; P = 0.02), as was mean input resistance (MΩ) (124.8 ± 15.35, n = 23 in the anti-NGF group vs. 291.6 ± 24.09, n = 27 in the NGS group; P < 0.0001); voltage threshold, however, was unchanged (−26.55 ± 1.575 mV, n = 11 in the anti-NGF group vs. −30.00 ± 1.22 mV, n = 12 in the anti-NGS group; P = 0.09) as was action potential amplitude (82.24 ± 7.14 mV, n = 14 vs. 83.28 ± 4.75 mV, n = 26; P = 0.9). Overall, anti-NGF treatment clearly resulted in markedly less membrane excitability.

Fig. 2.

The effects of anti-NGF treatment on burst patterns in DRG neurons from rats with chronic pancreatitis. A: electrophysiological properties were recorded from Dil-labeled neurons after dissection of DRG (arrow indicates cells labeled with Dil). Bar scale = 20 μm. B–E: representative traces of burst (action potentials) induced by 300-ms depolarizing current pulses injected through the patch pipette at rheobase (B and D) and two times rheobase (C and E) in DRG neurons from rats treated with either normal goat serum (NGS; B and C) or anti-NGF antibody (D and E).

Fig. 3.

The effects of anti-NGF treatment on electrophysiological parameters in DRG neurons from rats with chronic pancreatitis. A: average resting membrane potential in the two groups measured in current clamp (I = 0). B: average rheobase (minimum current required to evoke an action potential) in the two groups measured in I-clamp. C: average number of spikes or action potentials after injection of ×2 rheobase current in the two groups. D: average of input resistance calculated by whole-cell voltage measurements in response to injections of hyperpolarizing current. See text for details. *Significant difference (P < 0.05).

To understand whether this effect was specific to the sensitized state, we repeated the experiment on rats that received intraductal PBS only; these rats do not develop CP. Dil-labeled DRG neurons were examined for changes in resting membrane potential, action potential evocation, rheobase, and input resistance in response to treatment with anti-NGF or NGS. As seen in Fig. 4, there was no significant change in any of the neuronal parameters examined.

Fig. 4.

The effects of anti-NGF treatment on electrophysiological parameters in DRG neurons from rats without chronic pancreatitis. The same protocol and measurements were applied as in Fig. 3 except that rats were given intraductal PBS instead of TNBS. A: average resting membrane potential in the two groups measured in current clamp (I = 0). B: average rheobase (minimum current required to evoke an action potential) in the two groups measured in I-clamp. C: average number of spikes or action potentials after injection of ×2 rheobase current in the two groups. D: average of input resistance calculated by whole-cell voltage measurements in response to injections of hyperpolarizing current.

Anti-NGF treatment results in upregulation of I_A potassium currents in pancreas-specific primary sensory neurons in rats with CP.

We have previously shown that the A-type K⁺ current is suppressed in pancreas-specific afferent neurons in CP (22). We, therefore, investigated the role of this current in the attenuated excitability seen with anti-NGF treatment. Our results demonstrate that the total outward K⁺ current density was significantly increased in the DRG neurons of rats treated with anti-NGF (116.5 ± 13.41 nA; n = 13) compared with those treated with NGS (73.19 ± 13.12, n = 9; P = 0.04) when using a voltage protocol with holding potential at −100 mV and step voltage pulses from −60 up to +30 mV in 5-mV intervals (Fig. 5, A and B). Switching the holding potential to −50 mV, we achieved a blockade of most transient currents, revealing a sustained component that was sensitive to TEA (5 mM), indicating outward rectifying K⁺ currents (I_K), which was subtracted from the total K⁺ currents to yield the transient K⁺ current (I_A). DRG neurons from rats treated with anti-NGF treatment exhibited an increase of I_A density (pA/pF) of 58.19 ± 5.975 (n = 13) compared with the NGS group (37.52 ± 5.989; n = 9; P = 0.03). These values are consistent with a near normalization as seen in our previous report where TNBS-induced CP resulted in a 25% decrease in I_A peak current compared with normal controls (22).

Fig. 5.

The effects of anti-NGF treatment on Kv currents in DRG neurons from rats with chronic pancreatitis. Total outward K⁺ and I_A were evoked and separated by biophysical protocols switching holding potential between −100 and −50 mV, as illustrated in inset in which currents were generated using a 5-mV voltage step from −60 mV to +30 mV. A and B: representative traces respectively recorded from a single DRG neuron from a rat with CP treated with normal goat serum (NGS) (A) and another one treated with anti-NGF antibody (B). Left: total potassium current (I_TOTAL) at holding potential of −100 mV. Middle: sustained non-inactivating I_K-type current at holding potential of −50 mV. Right: transient inactivating A-type current (I_A) obtained from the difference between I_TOTAL and I_K. C, I/V curves from the two groups corresponding to the depictions of left, middle, and right. D: statistical summary of the differences in I_TOTAL (left), I_K (middle), and I_A (right) between the two treatment groups.

We also tested the effects of anti-NGF treatment in rats without pancreatitis (i.e., with intraductal infusion of PBS alone) on potassium currents and found no difference in I_total or I_A, suggesting once again that the effects of NGF suppression in the dose and duration used were specific for the sensitized state (results not shown).

Anti-NGF treatment does not affect the expression of Kv-associated channel or ancillary genes Kv1.4, 4.1, 4.2, and 4.3, and DPP6, DPP10, KCHIPs 1.

Real-time PCR performed on total RNA was obtained from laser-captured DRG (T9 and T10) sections in both groups of rats. We first used 14 cycles of pre-amplification of RNA since a similar protocol had previously allowed us to detect a significant difference in TRPV1 expression from the same RNA pool (27). The results, presented in Fig. 6, showed that mRNA levels (normalized to GAPDH) for Kv1.4, 4.1, 4.2, and 4.3, and DPP6, DPP10, and KCHIPS 2–4 were similar in both groups (n = 6 each). KCHIP1 was not detectable. We then used 10, and subsequently 5, cycles of pre-amplification and repeated the PCR but again could not detect differences by either protocol in the expression of the targeted genes (results not shown). Furthermore, normalization to housekeeping genes other than GAPDH such as β-actin and β-tubulin or to an average of all three by any pre-amplification protocol also did not detect any differences in the expression of these genes in either group (results not shown).

Fig. 6.

The effects of anti-NGF treatment on mRNA expression of Kv-associated and ancillary genes. Total RNA was obtained from laser-captured DRG sections from CP rats treated with either anti-NGF or control serum. A: section of DRG before laser capture showing fluorescently labeled neurons. B: same section after capture. C: RNA from T9 and T10 DRG segments was combined before cDNA synthesis and pre-amplified for 14 cycles in the presence of proprietary primers for the genes, Kv1.4, 4.1, 4.2, and 4.3, and DPP6, DPP10, and KCHIPs 1–4 (KCHIP1 was not detectable). Changes in mRNA levels of test genes were calculated by the DDCt method using GAPDH as a normalizer and expressed relative to the average DDCt of the serum-treated group for each gene of interest.

DISCUSSION

Within the potassium channel superfamily, the Kv channel group is the largest with 12 families, each of which may consist of more than one channel; these channels give rise to broad types of delayed rectifier currents. A-type currents are both activated and inactivated rapidly and inhibit neuronal excitability (repetitive firing) by regulating the interspike frequency of action potentials. Sustained delayed rectifier type (I_K) currents are activated later but are also involved in neuronal excitability, particularly by modulating the repolarization and after-hyperpolarization phase. Our previous study strongly implicated a role for decreased I_A in the excitability of pancreatic nociceptive neurons in TNBS-induced model of CP (22). This was in keeping with other studies in various models of visceral inflammation including cystitis (24), TNBS-induced ileitis (18), gastric ulcers (4), and neonatal colonic sensitization (17).

In a previously published study, our laboratory examined the effects of the same anti-NGF protocol on pain-related behavior and changes in TRPV1 expression and function (27). In that study, there was a marked reduction in pain responses as measured by referred somatic allodynia as well as pancreatic hyperalgesia. Furthermore, the activity of TRPV1, a key nociceptive receptor that is modulated by NGF, was considerably suppressed by this treatment. Finally, using similar protocols as we have described in this paper, we were able to show a suppression of the mRNA level of TRPV1 in pancreatic neurons from rats treated with anti-NGF, and this was accompanied by evidence of changes in protein expression as well. It is therefore helpful to interpret the present study in this context. The main results of this study are that systemic anti-NGF administration to rats with CP caused a reduction in the excitability of pancreatic nociceptors. This is associated with an increase in I_A but not in the expression of genes that are generally believed to subserve this current. In vitro studies have shown that NGF can suppress potassium (I_K type) currents (25, 26), but there is little knowledge about its effect on I_A currents. Paradoxically, NGF can prevent the reduction in transient A-currents in cutaneous afferent cell bodies after nerve ligation (6). At the same time, NGF has no effect on the expression of other Kv subunits associated with I_A currents such as Kv1.4 and 4.2 (14). Thus this study is the first definitive demonstration of a role for NGF in the suppression of I_A currents in an in vivo model of chronic pain.

An increase in these currents by NGF can therefore be expected to decrease the excitability of neurons, but the molecular basis of this remains largely unknown, with the most commonly implicated subunits being Kv1.4 and 4.3. There are four closely related subfamilies in the Kv group: Kv1 (Shaker), Kv2 (Shab), Kv3 (Shaw), and Kv4 (Shal). Kv1, Kv2, and Kv3 channels mediate both fast and slow inactivating outward K⁺ currents (A-type and delayed rectifier-type, respectively). In contrast, Kv4 channels predominantly mediate fast A-type K⁺ currents (3, 15). Although Kv4 channels localized to the soma and dendrites in central neurons are known to be responsible for the somatodendritic A-type K⁺ current (I_SA), the molecular basis of the I_A in dorsal root ganglia neurons remains controversial. Phuket et al. have shown that nociceptive DRG neurons express significant levels of Kv4.1 and Kv4.3, but the expressions of Kv1.4 and Kv4.2 are relatively low and infrequent (15) . Furthermore, dominant-negative suppression and overexpression strategies confirmed the contribution of Kv4 channels to I_A in DRG neurons. Thus Kv 4.3 is the most likely channel responsible for I_A currents in nociceptive neurons. Its contribution was supported by a study of functional colonic pain in rats, showing downregulation of Kv4.3, but not 1.4 or 3.4, in lumbar DRGs (17). However, despite using a dose that resulted in clear changes in excitability of neurons and I_A currents, we were unable to demonstrate a significant increase in the expression of Kv4.3 at the mRNA or protein level.

As with other Kv channels, changes in expression are only one of a myriad of ways in which Kv4.3 function can be affected. Indeed, functional expression of the I_A phenotype requires the interaction of a variety of proteins with the Kv α subunits themselves, which includes K-channel interacting proteins (KCHIPs; which facilitate trafficking of Kv4 channels to the plasma membrane, facilitate slow inactivation, and accelerate the recovery from inactivation) and the dipeptidyl peptidases DPP6 and DPP10 (by facilitating the voltage-dependent gating transitions that activate the channels, resulting in large hyperpolarizing shifts in voltage dependence) (10). However, we could not detect any significant change in the expression of these genes after anti-NGF treatment. We are left with the possibility that antagonism of NGF in this model alters I_A activity by either as yet unidentified potassium channel genes or by translational or posttranslational events of known gene products, including association with specific subunits, phosphorylation, and oxidation (11, 13). We also do not know whether the effects of NGF on Kv currents are mediated by a direct effect on nociceptor neurons (which express both high- and low-affinity receptors for the neurotrophin) or an indirect effect through the induction of other cellular or biological factors.

It can be argued, since NGF is critical for the development of sensory neurons and may also play an important trophic effect in adult neurons, that these effects are nonspecific. To test this, we examined the effects of anti-NGF treatment on excitability in rats without CP and could find no changes in electrophysiological attributes including potassium currents. Thus, at least at the dose and duration of NGF neutralization used in this study, the observed effects appear to be specific to the sensitized state targeting the changes in the expression of various genes and their products, associated with neuronal plasticity in this state. It is conceivable that at higher doses of anti-NGF antibody, nonspecific effects may emerge.

In conclusion, we have shown that neutralization of NGF significantly attenuates the changes in pancreatic nociceptor excitability and increases I_A by mechanisms that are not related to the expressions of Kv subunits or ancillary proteins. These findings emphasize a key role for NGF in pancreatic pain, presenting an important new target for the treatment of pain in this condition.

GRANTS

This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R01 DK-073558 (to P. J. Pasricha) and P30 DK-56339 (Stanford Digestive Disease Center).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).