Mechanisms underlying clinical efficacy of Angiotensin II type 2 receptor (AT2R) antagonist EMA401 in neuropathic pain: clinical tissue and in vitro studies

Background The clinical efficacy of the Angiotensin II (AngII) receptor AT2R antagonist EMA401, a novel peripherally-restricted analgesic, was reported recently in post-herpetic neuralgia. While previous studies have shown that AT2R is expressed by nociceptors in human DRG (hDRG), and that EMA401 inhibits capsaicin responses in cultured hDRG neurons, the expression and levels of its endogenous ligands AngII and AngIII in clinical neuropathic pain tissues, and their signalling pathways, require investigation. We have immunostained AngII, AT2R and the capsaicin receptor TRPV1 in control post-mortem and avulsion injured hDRG, control and injured human nerves, and in cultured hDRG neurons. AngII, AngIII, and Ang-(1-7) levels were quantified by ELISA. The in vitro effects of AngII, AT2R agonist C21, and Nerve growth factor (NGF) were measured on neurite lengths; AngII, NGF and EMA401 effects on expression of p38 and p42/44 MAPK were measured using quantitative immunofluorescence, and on capsaicin responses using calcium imaging. Results AngII immunostaining was observed in approximately 75% of small/medium diameter neurons in control (n = 5) and avulsion injured (n = 8) hDRG, but not large neurons i.e. similar to TRPV1. AngII was co-localised with AT2R and TRPV1 in hDRG and in vitro. AngII staining by image analysis showed no significant difference between control (n = 12) and injured (n = 13) human nerves. AngII levels by ELISA were also similar in control human nerves (4.09 ± 0.36 pmol/g, n = 31), injured nerves (3.99 ± 0.79 pmol/g, n = 7), and painful neuromas (3.43 ± 0.73 pmol/g, n = 12); AngIII and Ang-(1-7) levels were undetectable (<0.03 and 0.05 pmol/g respectively). Neurite lengths were significantly increased in the presence of NGF, AngII and C21 in cultured DRG neurons. AngII and, as expected, NGF significantly increased signal intensity of p38 and p42/44 MAPK, which was reversed by EMA401. AngII mediated sensitization of capsaicin responses was not observed in the presence of MAP kinase inhibitor PD98059, and the kinase inhibitor staurosporine. Conclusion The major AT2R ligand in human peripheral nerves is AngII, and its levels are maintained in injured nerves. EMA401 may act on paracrine/autocrine mechanisms at peripheral nerve terminals, or intracrine mechanisms, to reduce neuropathic pain signalling in AngII/NGF/TRPV1-convergent pathways.


Background
Neuropathic pain has a prevalence of 1-8% in the general population, and is a major unmet medical need due to limited efficacy and tolerability of currently available drug treatments. Hyperexcitability and abnormal sprouting of primary afferent sensory nerve fibres underpin peripheral neuropathic pain [1].
The angiotensin II (AngII) type 2 receptor (AT 2 R) was identified as a novel target for neuropathic pain: several small molecule AT 2 R antagonists with >1000-fold selectivity over the AT 1 R receptor produced dose-dependent analgesia in multiple rodent neuropathic pain models, with the analgesia abolished in mice null for the AT 2 R [2]. Our studies in human and rat DRG neurons demonstrated that EMA401, a selective AT 2 R antagonist, inhibited neurite outgrowth and capsaicin responses in cultured human and rat DRG neurons, in an in vitro model of sensitization [3]. In a recent randomized, double-blind, placebo-controlled clinical trial in patients with post-herpetic neuralgia, administration of the orally active AT 2 R antagonist, EMA401, for 4 weeks produced significant pain relief above placebo and was well-tolerated [4].
The underlying mechanisms of AT 2 R signalling in clinical neuropathic pain states, and mode of action of the peripherally-restricted AT 2 R antagonist EMA401, require elucidation. We have previously shown that AngII, the endogenous ligand for the AT 2 R, acted in vitro on AT 2 R resulting in sensitization of TRPV1 and increased neurite outgrowth in DRG neurons via increased cAMP, which were both inhibited in the presence of EMA401 [3]. TRPV1, the heat and capsaicin receptor, is known to be sensitized when phosphorylated by cAMP, and desensitized when dephosphorylated [5]. Neurite outgrowth is also sensitive to cAMP levels, and our previous study showed the neurite promoting effects of AngII in the presence of added neurotrophic factors, which have similar effects [3]. In animal models of neuropathic pain the analgesic action of AT 2 R antagonists involves inhibition of enhanced AngII/AT 2 R induced p38 and p42/ p44 MAPK activation [2].
Systemic AngII is known to be derived from the action of kidney derived renin on AngI in the circulation. However, local renin angiotensin system (RAS) components have been described in brain [6], heart, blood vessels [7], and DRG [8,9], constituting potential local sources of AngII, separate from the circulating AngII. In addition to AngII, other components of the RAS, AngIII and Ang- (1)(2)(3)(4)(5)(6)(7), are also ligands at the AT 2 R; however the levels of these endogenous ligands in clinical peripheral neuropathic tissues and signalling pathways are unknown.
AngII to AngIII conversion in rodent CNS has been reported to mediate descending inhibitory pathways, while Ang-(1-7) is a biologically active heptapeptide formed endogenously from either AngI or AngII, with vasodilatory and antiproliferative activities that oppose the constrictive and proliferative effects of AngII [10,11]. Circulating levels of Ang-(1-7) in humans are reported to increase following long term administration of ACE and AT 1 receptor blockers [12][13][14].
We have studied the expression of AngII, AT 2 R and TRPV1 in clinical tissues, and quantified the levels of AngII, AngIII and Ang-(1-7) by ELISA. The effects of AngII, the AT 2 R antagonist EMA401 and the synthetic AngII agonist EMA1087 (Compound 21) [15] on calcium influx and neurite outgrowth were studied in cultured rat DRG neurons. The in vitro effects of AngII, NGF and EMA401 on expression of pp38 and pp42/44 MAPK in rat sensory neurons were measured using immunofluorescence. AngII, AT 2 R and TRPV1 expression were also studied in cultured human DRG neurons using immunofluorescence, and signalling pathways using calcium imaging.

Results of in vitro studies
These studies showed co-expression of AT 2 R, AngII and TRPV1 in small diameter cultured hDRG neurons using immunofluorescence ( Figure 3). AngII was expressed in 75.6 ± 6.3% small diameter neurons (≤50 μm diameter, 509 neurons), with a mean diameter of 37.5 ± 1.8 μm, and co-localised with virtually all AT 2 R positive hDRG neurons. AngII was co-localised in cultured DRG neurons with AT2R and TRPV1, as illustrated in Figure 3h-k. AngII treated rDRG neurons showed a significant increase in pp42/44 signal intensity compared to vehicle treated controls (*P < 0.05), similar to the positive controls treated with NGF (*P < 0.05, Figure 4). Signal intensity was reduced to an extent in cultures treated with AngII combined with EMA401 (P > 0.05). Similar increases in pp38 signal intensity were observed in NGF (*P < 0.05) and AngII treated neurons (*P < 0.05), which were reduced to an extent after co-incubation with EMA401 (P > 0.5, Figure 4).
A schematic diagram showing the signalling pathway involving AT2R and TRPV1 is shown in Figure 6.

Discussion
In this study we observed that AngII was expressed in a large proportion of small diameter neurons in human DRG, co-localised with AT 2 R and TRPV1, and that AngII levels were preserved in injured human DRG and nerve tissues. Co-localisation of AngII expression with AT 2 R in hDRG, and its levels in injured nerves, supports the concept of an intrinsic neuronal angiotensinergic system, and suggests that EMA401 may act via autocrine/intracrine in addition to paracrine mechanisms in DRG neurons, to reduce augmented neuropathic pain signalling.
Quantitative ELISA studies in human tissues showed that the major AT 2 R ligand in human peripheral nerves was AngII, and that its levels were similar in injured human nerves and painful neuromas compared to control nerves. Evidence supports the concept of an intrinsic neuronal angiotensinergic system, with intraneuronal AngII formation in sensory neurons, which appears to be maintained in injured neurons-hence its potential autocrine/intracrine role in pathological nociceptive mechanisms. AngIII and Ang-(1-7) levels were below the detection limits, in accord with other studies [8]. Previous studies have also described dense AngII positive nerve fibres and high levels in peripheral organs [16], and localisation in small and medium sized neurons in trigeminal ganglia [8,9]. In situ hybridization in rat trigeminal ganglia have revealed expression of AngII precursor Angiotensinogen (Ang-N) mRNA in the cytoplasm of numerous neurons, and post in situ hybridization immunocytochemistry marked some of these for Ang II; substance P was found colocalized with Ang II [8]. These studies support intraneuronal AngII formation in sensory neurons, separate from circulating AngII, and its potential autocrine/intracrine role in nociceptive mechanisms.
It is important to note that we have measured AngII levels in human nerves and DRG by ELISA in frozen tissue extracts; as AngII may degrade when extracted in frozen tissues, freshly extracted tissues need to be studied in future for AngII, other putative endogenous RAS ligands, and their metabolites. The levels obtained in our study are higher than those reported in other studies of rat and human trigeminal ganglia [8,9], and may reflect differences in peptide extraction prior to assay.
Injury and associated inflammation may increase local AngII, contributing to sensitization of nociceptors, particularly in sprouting nerve fibres. The molecular regulation of AT 2 R and AngII expression in injured human or rodent sensory neurons is unknown, and may be influenced by target-organ derived or intrinsic factors. In our previous study [3], AT 2 R levels were reduced in human nerve segments proximal to injury (lesion distal to the DRG, or 'peripheral axotomy'), but they were preserved or high in chronic painful neuromas, which comprise regenerating nerve fibres. These AT 2 R receptor level changes parallel those of key pain receptors and ion channels regulated by target-derived growth factors e.g. TRPV1, Nav 1.7 and Nav1.8 by NGF after nerve injury. The mode of analgesic action of EMA401 may involve inhibition of augmented AngII/AT 2 R signalling in NGF convergent pathways, a postulate which is supported by our in vitro results, described below. The lack of CNS penetration of EMA401, and the time-course of its clinical effects in patients with post-herpetic neuralgia (i.e. gradually progressive over 4 weeks, unlike anticonvulsants) argue against AngII acting as a classical neurotransmitter in dorsal spinal cord.
Co-expression of AngII, AT2R and TRPV1 was observed in cultured small diameter hDRG neurons using immunofluorescence, in accord with the hDRG tissue findings. AngII was co-localised in cultured DRG neurons with AT2R and TRPV1, as illustrated in  In AngII treated rat neuronal cultures we observed significantly increased levels of pp38 and pp42, similar to the effect of NGF, which were reduced in the presence of EMA401. These results indicate that AngII mediated sensitization involves p38 and p44/42 phosphorylation, and that EMA401 is effective in blocking this in vitro, in agreement with animal models of pain [2,17]. It appears that the mechanism of AngII mediated p38 and p44/42 phosphorylation involves interaction of the AT 2 R and TrkA, which underlies neuronal differentiation of NG108 cells [18], sensitization of mature neurons as observed in this study, and increased neurite outgrowth.
In our previous study [3], neurons treated with combined AngII and NTFs (NGF, GDNF, NT3) had significantly longer neurites compared with NTF treatment alone, suggesting a synergistic effect of AngII with the neurotrophic factors, which was reduced by co-treatment with EMA401. In this study neurite lengths were also significantly increased in neurons treated with AngII and EMA1087 (C21) alone, but less than with added NTFs. The effect of EMA401 on neurites was diminished outgrowth, rather than degeneration of established neurites. Thus AngII treatment of DRG neurons leads to increased neurite length and TRPV1 sensitization by increased cAMP, both of which are blocked by EMA401 [3]. While AT 2 R activation appears to have a synergistic effect with TrkA, other pathways such as PI3K [19] may also be involved, and need investigation. Other studies have reported a role for the AT 2 R in mediating neurite outgrowth in vitro via estrogen activation [20,21], and in a rodent model of inflammatory hypersensitivity [22]. These mechanisms may contribute to the analgesic effects of AT 2 R antagonists in such animal models, and in patients with chronic pain, particularly those more prevalent in females.
Our finding that AngII and EMA1087 (C21), a small molecule AT 2 R agonist, caused neurite outgrowth in our assay is in agreement with the neurite promoting effects of C21 previously described in the neuroblastoma-glioma hybrid NG108 cells, that was blocked by the AT 2 R receptor antagonist (PD-123,319) [15,23]. PD123,319 was effective in reducing neurite growth in a CFA inflammatory pain rodent model [22], and EMA401 inhibited neurite outgrowth in vitro [3], confirming their antagonist effects at the AT 2 receptor, and potential efficacy in pain states associated with abnormal nerve sprouting. Conceptually, while promoting AT 2 receptor mediated neuronal outgrowth could be beneficial after certain pathologies (e.g. stroke), aberrant or collateral neuronal regeneration with hypersensitivity may underpin some painful conditions in the periphery mediated via AT 2 R [22], in accord with our clinical efficacy data with EMA401 in postherpetic neuralgia patients, and human DRG nociceptor models in vitro [3,4]. An important consideration is that the neurite promoting effects of C21 described previously in NG108 cells are derived from undifferentiated neuroblastoma-glioma hybrid cells, presented as % cells with neurites [18], with neurite extension representing neuronal differentiation of the NG108 cells, and AngII effects measured by the number of cells expressing neurites. This reflects neuronal differentiation, with C21 promoting the neuronal phenotype.
The findings of our study are, however, derived from measurements of neurite lengths from mature, well differentiated neurons, and describe the effects of AngII on neurite lengths in individual neurons, relevant to hyperinnervation.
As AT 2 R is a GPCR, the AngII mediated increase in cAMP [3] is in keeping with adenylyl cyclase activation, likely to involve a G αs mechanism. The consequent TRPV1 sensitization was not observed in the presence of the MAP kinase inhibitor, PD98059 in hDRG and rDRG neurons in this study. AngII effects have been associated with increased neuronal excitability and promotion of neurite outgrowth via multiple mechanisms of AT 2 R activation, depending on the cell type [19]. cAMP is also known to activate MAPK to stimulate cell growth [24], which in post-mitotic non-dividing neurons may manifest as neurite outgrowth. The increased pp38 and pp44/42 expression in AngII treated neurons indicate that AngII mediated neurite outgrowth and neuronal sensitization both involve MAPK/ERK activation, that is attenuated by EMA401. AngII thus appears to have similar effects as NGF in causing neurite outgrowth and TRPV1 sensitization [25], but not in the presence of staurosporine, PD98059 (MAPK inhibitor) and GW441756 (TrkA inhibitor).
Our results showed that both AngII and its analogue EMA1087 (C21) caused TRPV1 sensitization, suggesting that like AngII, C21 has a pro-nociceptive effect. While the effects of EMA1087 are more similar to AngII for neurite outgrowth but less so in sensitization effects, EMA1087 did show significantly enhanced responses as did AngII. EMA1087 is a synthetic molecule and AngII a peptide, which could explain differential effects in assays. AngII appeared to result in greater sensitization of human neurons compared with rat neurons, possibly reflecting a species specific difference.
Previous studies have shown pro-inflammatory effects of AngII in a variety of tissues including kidney, heart and blood vessel wall, by up-regulating the expression of both AT 1 R and AT 2 R, and activation of a number of signalling pathways including p44/42 MAPK, p38 MAPK, c-JUN and NF-κB [26,27]. AngII and the inflammatory mediator lipopolysaccharide (LPS) are reported to upregulate the expression of AT 1 R, AT 2 R and LOX-1, and IL-1β in cardiomyocytes, with activation of MAPKs, c-Jun and NF-kB, suggesting a positive feedback between AngII and inflammation [28]. CD68 positive tissue macrophages are also reported to increase expression of angiotensinogen and renin in a model of inflammatory hypersensitivity [22]. There is thus potential for EMA401 to block a number of convergent pathways activated by a variety of inflammatory mediators. The sensitizing effect of NGF is well known [25,[29][30][31] and is the basis of our in vitro hDRG model of hypersensitivity, modelling the elevated NGF levels observed in tissues from chronic pain conditions [32][33][34]; AngII treatment caused further sensitization, suggesting functional synergy between AT 2 R and TrkA. Other pathways need investigation, such as GDNF signalling, which may have similar additive effects.
AngII has been reported to increase K + channel activity in hippocampal neurons, that was reversed by the AT 2 R antagonist PD123,319 [35]. In a recent study, this mechanism of hyperpolarisation was proposed to underlie the lack of pain in Buruli ulcers caused by mycolactone, the polyketide product of m. ulcerans [36]. While this may reflect differences in signalling between hippocampal and DRG neurons, a number of clinical and neuropathological aspects of Buruli ulcer have drawn caution against such an interpretation [37].

Conclusion
AngII and AT 2 R are co-expressed in nociceptive human sensory neurons, and the levels of AngII, the major endogenous ligand in human peripheral nerves, are preserved after injury. AngII induces p38, p42/p44 mitogen activated protein kinase (MAPK) activation, neurite outgrowth in adult rat DRG neurons, and sensitization of adult rat and human DRG neurons that is blocked by EMA401. Hence increased AngII/AT 2 R signalling in DRG neurons secondary to peripheral nerve injury may have a key role in chronic pain mechanisms, including neuropathic pain. The mode of EMA401 analgesic action appears to involve inhibition of augmented AngII/AT 2 R induced p38 and p42/p44 MAPK activation, and hence inhibition of DRG neuron hyperexcitability and sprouting of DRG neurons. EMA401 is likely to be most effective in conditions of hypersensitivity associated with abnormal nerve sprouting, where AngII may synergize or augment NGF mechanisms. Selective AT 2 R antagonists represent a new class of analgesics for improved relief of neuropathic pain.

Immunostaining for Angiotensin II (AngII) in human tissues Tissues
A range of tissues was used in this study obtained with consents and approvals as described previously [3], including Local Research Ethics Committee, Royal National Orthopaedic Hospital, Stanmore, UK, Material Transfer Agreement, and Netherlands Brain Bank. Specimens were snap frozen in liquid nitrogen and stored at −70°C until use or immersed in Zamboni's fixative (2% w/v formalin, 0.1 M phosphate, and 15% v/v saturated picric acid) for 2 h and stored in phosphate buffered saline (PBS) containing 15% sucrose, 0.01% azide.

Immunohistology
Tissues were supported in optimum cutting tissue (OCT) medium (RA Lamb Ltd, Eastbourne, UK). Tissue sections (15 µm thick) were collected onto coated glass slides and post-fixed in 4% w/v paraformaldehyde in 0.15 M phosphate buffered saline (PBS) for 30 min (for frozen section only). Endogenous peroxidase was blocked by incubation in methanol containing 0.3% w/v hydrogen peroxide for 30 min. After rehydration with PBS buffer, sections were incubated overnight with primary antibody using a range of dilutions ( Table 1).

Analysis of data DRG
Antibody-immunoreactive, nucleated neurons in sensory ganglia (DRG) were counted and their diameter assessed using a calibrated microscope eyepiece graticule and expressed as % total; image analysis (% area) of nerve sections has been described previously [3].

ELISA for AngII, AngIII and Ang-(1-7) AngII ELISA
Frozen tissue samples of control normal nerve (n = 31), injured nerves (n = 7), and painful neuromas (n = 12), were weighed and extracted in boiling 0.5 M acetic acid for 10 min. Extracts were then concentrated and desalted by applying 0.2 ml extract to an activated Sep-Pak column, washed and desalted with water containing 0.1% TFA. Bound peptides were eluted from the columns using 1 ml of 80% acetonitrile containing 0.1% TFA. The eluate was evaporated to dryness using a speed-vac overnight. Each vial was reconstituted in 1× assay buffer supplied by the manufacturer (see below).
A human specific AngII/AngIII immunoassay kit (EK-002-12, reacting 100% to AngII and AngIII, Phoenix Pharmaceuticals, Burlingame, California, USA) was used according to the manufacturer's instructions (but note AngIII specific immunoassay kit described below showed undetectable levels of AngIII in these nerve extracts). AngII standard (0.04-25 ng/ml) and specimens (50 µl) were added in duplicate and the mean used for subsequent analysis. The absorbance at 450 nm in each well was measured using a plate reader. Standard curve was plotted using a log scale and the concentration of human AngII in each specimen determined using Excel software.
In order to determine the recovery of AngII after extracting in boiling in acetic acid and concentrating using a Sep PaK, 200 µl of standard (at 1 nmol/ml) was added to 800 µl of boiling acetic acid for a further 10 min, and after cooling 200 µl was applied to Sep Pak and dried overnight as above. This was reconstituted with 200 µl of assay buffer and 50 µl used for assay in duplicate.

AngIII
A human specific AngIII immunoassay kit (USCN Life Sciences ELISA kit, E92312Hu, Diagenics Limited, Milton Keynes, England) that reacts 100% to AngIII and with no significant cross-reactivity with analogues, was used according to the manufacturer's instructions. AngIII standard (6.17-500 pg/ml) and specimens (50 µl) were added in quadruplicate and the mean used for subsequent analysis. The absorbance at 450 nm in each well was measured using a plate reader. Standard curve was plotted and the concentration of human AngIII in each specimen determined using Excel.

Ang-1-7
A human specific Ang-1-7 immunoassay kit (USCN Life Sciences ELISA kit, E86085Hu, Diagenics Limited, Milton Keynes, England), reacting 100% to Ang-(1-7) and no significant cross-reactivity with analogues, was used according to the manufacturer's instructions. Ang-(1-7) standard (12.35-1,000 pg/ml) and specimens (50 µl) were added in quadruplicate and the mean used for subsequent analysis. The absorbance at 450 nm in each well was measured using a plate reader. Standard curve was plotted and the concentration of human Ang-(1-7) in each specimen determined using Excel.

In vitro studies Preparation of hDRG neurons
hDRG were obtained from five patients with brachial plexus avulsion undergoing nerve repair surgery, excised as a necessary part of surgical repair i.e. redundant tissue; ganglia were enzyme digested, mechanically dissociated and plated on collagen and laminin coated MatTek dishes (MatTek Corp USA), in Ham's F12 medium containing 10% HIFCS (heat inactivated fetal calf serum), penicillin/ streptomycin (100 μg/ml each), NTFs (NGF 100 ng/ml, GDNF and NT3 50 ng/ml each), as previously described [3], for 48 h at 37°C before further studies.

Calcium imaging in human DRG neurons
48 h after plating, hDRG neurons were loaded with 2 μM Fura2 AM, and responses to capsaicin, AngII, and EMA1087 (C21), were imaged as before [3]. In each experiment, capsaicin sensitive neurons were identified with a brief 200 nM capsaicin stimulus (30 s), and washout of medium. After a rest period of 40 min, a second capsaicin stimulus, of 1 μM was applied with or without drugs. The effects of the kinase inhibitor staurosporine, MAPK inhibitor PD98059 and TrkA inhibitor GW441657 on AngII mediated sensitization were determined by adding either one of the inhibitors 10 min prior to adding AngII, followed 10 min later, by the second capsaicin stimulus. The second capsaicin response was normalised to the first, for calculating the percent response, and the average calculated for each group.

Preparation of rDRG neurons
Bilateral DRG from all levels of 6 adult female Wistar rats were isolated and plated as before [3], in collagen/laminin coated glass bottom plastic petri dishes (MatTek, USA), at 1,000 neurons/dish, in BSF2 medium without NTFs for pp38 and pp42 immunofluorescence and neurite length assay, and with NTFs for calcium imaging, at 37°C. Calcium imaging studies in rat DRG neurons were carried out as for human neurons above.
Following fixation, and 3 min permeabilisation with methanol at −20°C, the neurons were incubated in rabbit polyclonal antibody to phospho-p38MAPK pThr180