Current analgesic therapies typically inhibit cyclooxygenase enzymes or activate opiate receptors. Each of these front line analgesics for treatment of inflammatory and neuropathic pain has some negative clinical consequences or lacks efficacy in a significant proportion of patients. Adjuvant analgesics, including antidepressants (SSRIs), anticonvulsants (gabapentin, pregabalin) and anesthetics (mexiletine, lidocaine) are sometimes effective in relieving pain. However, they do not represent widely effective treatments for relief of chronic inflammatory or neuropathic pain or hyperalgesia.
The NAAG receptor, mGluR3, is widely expressed by neurons and glia in the nervous system including the PAG and RVM . The observation of NAAG immunoreactivity (Figures 1a–c) in the ascending and descending pain pathway supports the potential of NAAG peptidase inhibitors to affect the response to inflammation.
The efficacy of NAAG peptidase inhibition in the PAG is consistent with the report that intra-PAG injection of the group II mGluR agonist L-CCG-I also reduces responses in the formalin model . In previous studies, we found that peripheral , intrathecal  and intracerebroventricular  administration of NAAG peptidase inhibitors had similar analgesia-like properties in the formalin test. The data in this manuscript provide the first insight into the potential of NAAG and the inhibitors of its inactivation to influence the ascending (PAG) and descending (RVM and PAG) pain pathways. Given the expression of mGluR3 receptors and NAAG in the spinal cord and spinal sensory neurons [10, 33], our previous studies and the data presented here suggest that, like opiate peptides, NAAG has the potential to moderate the response to inflammatory pain at several different levels within this pathway. The efficacy in this model of NAAG peptidase inhibition in the PAG and RVM support the conclusion that, at a minimum, the release of NAAG in these regions modulates the spinal pain response via the descending inhibitory pain pathway. This hypothesis needs to be further tested in other models that more clearly assess perception of pain since flinching in the formalin model can be critiqued as a spinal reflex that does not reflect cortically based cognition of pain. Relevant to this question, we previously reported that systemic and intrathecal injections of two NAAG peptidase inhibitors were highly effective in the sciatic nerve ligation and carrageenan pain models [5, 6]. Additionally, we found NAAG peptidase inhibition to be analgesic in a model of bone cancer pain .
A series of studies demonstrate that NAAG peptidase inhibition elevates extracellular NAAG levels with the consequent activation of a group II mGluR (mGluR3), an activity that inhibits the release of small amine transmitters including glutamate, GABA, and aspartate, via presynaptic inhibition [16–20]. While other mechanisms of action are possible, presynaptic inhibition might well be responsible for the efficacy of NAAG peptidase inhibitors in animal models of several nervous system disorders .
Figure 10a presents the first demonstration of inflammation-induced increase in glutamate release in the PAG, a result consistent with increased ascending excitatory signals from the site of inflammation. Given the efficacy of ZJ43 microinjection into the PAG in reducing flinching and increasing extracellular NAAG levels, we examined the effect of NAAG peptidase inhibition on this inflammation-induced glutamate release. Systemically applied ZJ43 (50 mg/kg) reaches a concentration of 2 nM in the brain 30 minutes after injection  and significantly reduces NAAG hydrolysis in the rat brain in vivo (Olszewski et al., submitted). Its efficacy in blocking this inflammation-induced glutamate release in the PAG is consistent with our models of NAAG activation of presynaptic mGluR3 to inhibit transmitter release [3, 33]. The difference between the RVM and PAG with respect to formalin-induced glutamate release could reflect the role of the PAG, but not the RVM, in the ascending pain pathway. While these data demonstrate a role for NAAG in the control of inflammation-induced glutamate release in the PAG, they are not sufficient to prove that the NAAG peptidase inhibition-mediated decrease in glutamate release mediates the observed reduction in the inflammation-induced motor response.
Microinjection of ZJ43 into the RVM also reduced the response to footpad inflammation (Figures 6
7) and systemic treatment with this inhibitor also elevated RVM NAAG levels (Figure 9b). In contrast to the PAG, however, inflammation did not significantly elevate glutamate levels in the RVM. Since microinjection of excitatory amino acids into the RVM is analgesic , it would not be expected that formalin treatment would necessarily produce a substantial increase in glutamate release or that inhibition of glutamate release in the RVM would mediate analgesia. One interpretation of these data is that NAAG activation of mGluR3 receptors inhibited the release of other transmitters in the RVM with the consequent effect on the local circuitry [37, 38]. For example, inhibition of GABA release could indirectly result in an increase in release of other transmitters, whose actions mediate analgesia in the RVM [39, 40]. Alternatively, the effect of formalin injection on glutamate release in the RVM might have been restricted to a volume of tissue that was smaller than that sampled by the microdialysis probe resulting in a failure to detect increases in glutamate levels above the background in the sampling area. However, a small study (n = 3) obtained using a smaller (1 mm) dialysis probe tip in sampling the RVM provided no evidence of an inflammation stimulated increase in glutamate release Figure 10b.
Heterotropic group II mGluR (mGluR2 and mGluR3) agonists reduce inflammatory pain responses and also may represent a novel analgesic strategy . However, these compounds were tested in mGluR2 and mGluR3 knockout mice in animal models of schizophrenia and were found to be effective in mGluR3 but not mGluR2 knock outs [41, 42]. In the same animal models, NAAG peptidase inhibition was effective in the mGluR2 but not the mGluR3 knockout mice . These data support the conclusion that the heterotropic mGluR2/3 agonists and mGluR2 positive allosteric modulators have the potential to be effective mGluR2 based analgesic strategies in contrast to NAAG peptidase inhibition that represents an mGluR3 specific strategy. Also relevant to the differences between these two analgesic approaches, pharmacotherapies, such as antidepressants, sedatives and anxiolytics, that increase the activity of endogenous transmitters tend to enhance the normal ongoing physiology and thus can have less potential for secondary effects than continuous agonist-based receptor activation.
The concept that orally available NAAG peptidase inhibitors [2, 26] might ultimately be used clinically for the treatment of inflammatory and neuropathic pain begs the question as to their potential secondary effects inasmuch as the peptide and mGluR3 are widely distributed in the nervous system. Studies in mice do not suggest that negative secondary effects result from NAAG peptidase inhibition . For example, we found no significant neurological deficits in mice in which the major NAAG peptidase, glutamate carboxypeptidase II, had been knocked out . Similarly, chronic treatment with a NAAG peptidase inhibitor was without detectable side effects in a study where the drug increased the lifespan of mice in a model of amyotrophic lateral sclerosis . Acute treatment with ZJ43 similarly lacks detectable effects in open field behavior , prepulse inhibition of acoustic startle , or the 1.5 hr delay novel object recognition test [Olszewski et al., submitted]. Neither systemic NAAG peptidase inhibition  nor microinjection of 150μg of ZJ43 into the PAG and RVM (Figure 8) alter the reaction time of rats in the hot plate test. This lack of apparent side effects is consistent with the established pattern of peptide co-transmitter release under conditions of high neuronal activity with little released during normal to low levels of activity. Consistent with this concept, the basal extracellular concentrations of NAAG and glutamate in the dialysates, uncorrected for recovery, were 53 ± 06 nM and 1,250 ± 130 nM in PAG, 74 ± 5 nM and 820 ± 70 nM in RVM respectively. This difference in extracellular concentrations also is found in other brain regions [19, 20] despite the fact that NAAG is present in mM concentrations in the mammalian nervous system . Such a pattern of release is consistent with our model of NAAG feeding back on presynaptic mGluR3 to dampen synaptic release of primary amine transmitters under conditions of high activity .