The present study provides the first evidence that targeted disruption of the nNOS gene significantly attenuates nerve injury-induced mechanical hypersensitivity. Moreover, nerve injury-induced mechanical hypersensitivity in WT mice was significantly inhibited by systemic and intrathecal administration of L-NAME and by intrathecal treatment with 7-NI. Peripheral nerve injury increased the expression of nNOS in the DRG, although it did not alter the expression of nNOS in the spinal cord. Our findings suggest that nNOS, especially in the DRG, may play a critical role in nerve injury-induced mechanical hypersensitivity.
Conflicting pharmacological evidence has been reported regarding the effects of systemic or spinal treatment with specific or non-specific NOS inhibitors on neuropathic pain. Meller et al.  first reported that intrathecal administration of L-NAME (20 nmol) on day 3 after loose ligation of the sciatic nerve blocked the thermal hyperalgesia in rats for a period of 2 h. A subsequent study also demonstrated that injecting 3-bromo-7-NI (a specific nNOS inhibitor, 10 mg/kg) intramuscularly into the mid-thigh region daily for 7 days beginning on day 7 post-surgery significantly alleviated nerve injury-induced thermal hyperalgesia from weeks 5 and 6 onward in rats . Yoon et al.  further showed that both mechanical and cold allodynia induced by tight ligature of the left L5 and L6 spinal nerves were dose dependently suppressed during the maintenance phases (1 to 3 weeks post-ligature) by intraperitoneal L-NAME (10, 50, 100, and 200 μM/kg). In support of these findings, our pharmacological study showed that spinal nerve injury-induced mechanical hypersensitivity in mice could be dramatically attenuated by systemically (100 mg/kg) and intrathecally (30 μg) injected L-NAME and by intrathecally injected 7-NI (8.15 μg) on day 7 post-nerve injury (Figs. 2, 3, 4). Interestingly, Yamamoto and Shimoyama  found that the development of thermal hyperesthesia evoked by sciatic nerve constriction injury was significantly delayed by intrathecal pre-treatment (10 min before injury), but not post-treatment (15 min after injury), with NOS inhibitors Nω-nitro-L-arginine (30 μg) and L-NAME (100 μg). However, Pan et al.  reported that allodynia induced by tight ligation of the left L5 and L6 spinal nerves was not affected by intrathecal injection of either NG-monomethyl-arginine (a non-specific NOS inhibitor, 30 μg) or 1-(2-trifluoromethylphenyl) imidazole (a nNOS inhibitor, 30 μg) at the third or fourth week post-nerve injury. Additionally, Luo et al.  observed that neither pre-surgical (once daily for 6 days beginning 30 min before surgery) nor post-surgical (days 11 through 13) 7-NI (50 mg/kg) given subcutaneously or intraperitoneally affected the development or maintenance of tactile allodynia evoked by tight ligature of the left L5 and L6 spinal nerves in rats. The reason for the discrepancies among these pharmacological studies is unclear but may be related to the drug potency, delivery method, dose, and administration times. It is worth noting that only one dose of a nonspecific or specific nNOS inhibitor was administered by Luo et al.  and Pan et al . It is possible that the higher doses used in these two studies might have had a significant inhibitory effect on nerve injury-induced neuropathic pain. Our nNOS knockout mouse study clearly demonstrated that deficiency of nNOS significantly blunted spinal nerve injury-induced mechanical hypersensitivity (Fig. 1). In addition, nNOS knockout mice exhibited impaired nerve injury-induced thermal hypersensitivity (data not shown). Although it has been reported that spinal eNOS is upregulated in nNOS KO mice , it appears that this upregulation does not fully compensate for nNOS function in neuropathic pain. Thus, our genetic knockout and pharmacologic studies support the notion that nNOS, particularly at the spinal cord level, may contribute to the mechanisms that underlie the development and/or maintenance of neuropathic pain.
Peripheral nerve injury insult has distinct effects on nNOS expression in the DRG and dorsal horn. Under normal conditions, the DRG contains only a few neurons that exhibit nNOS labeling, the majority of which are very small, whereas spinal dorsal horn exhibits intense nNOS labeling in several neuronal types, especially in the superficial dorsal horn . It has been reported that constriction of the common sciatic nerve in rats increased nNOS expression in the ipsilateral L4-L6 DRG, but not in spinal segments L4-L6 . The increased level of nNOS protein in the DRG was seen in the small, medium, and large cell bodies . Likewise, tight, unilateral ligature of the rat L5 and L6 spinal nerves increased nNOS mRNA in ipsilateral L5/6 DRG neurons, but not in dorsal horn, beginning 1 day after nerve ligation and lasting for at least 13 weeks . A corresponding increase in DRG (but not spinal cord) nNOS protein was also observed and localized mainly to small and occasionally medium-sized sensory neurons [19, 21]. Consistent with these findings in rats, our study showed that unilateral injury of the fifth spinal nerve in mice also increased expression of nNOS protein in the ipsilateral fifth DRG, but not in dorsal horn. An increase in the DRG NOS catalytic activity and a decrease in spinal NOS catalytic activity have been reported following peripheral nerve injury [20, 22]. The evidence described above indicates that peripheral nerve injury up-regulates expression and activity of nNOS in small DRG neurons and down-regulates its expression and activity in the spinal cord. Such adaptations suggest that the increased nNOS in the DRG may be involved in the mechanism underlying the development and/or maintenance of neuropathic pain. It has been demonstrated that intrathecally administered agents can directly affect activities of DRG proteins in pain processing in addition to affecting activities of these proteins in spinal cord neurons . It is very likely that the pharmacologic effects of specific and non-specific NOS inhibitors in the present study were mediated through their actions on the DRG nNOS, although our data could not exclude their effects on spinal cord nNOS.
Consistent with previous reports [5, 31, 32], our study detected basal expression of eNOS and iNOS proteins in the spinal cord and DRG of mice under normal conditions. However, the expression of neither was altered in the spinal cord or DRG on day 7 post-nerve injury. It should be noted that we did not examine iNOS expression at the ligature site or at other time points post-nerve injury. Previous studies showed that local expression of iNOS was induced at 3 days and persisted for at least 26 days at the constriction and distal sites following chronic constriction partial nerve injury [33–36]. Systemic and topical administration of specific or non-specific iNOS inhibitors reversed nerve hyperaemia proximal and distal to the constriction . Targeted disruption of the iNOS gene slowed Wallerian degeneration of myelinated fibers and delayed regeneration of myelinated and smaller caliber fibers in a chronic constriction partial nerve injury model . Slowed nerve degeneration is associated with normal initiation but delayed expression of neuropathic pain , suggesting that local induction of iNOS expression may be important in the pathogenesis of nerve injury-induced neuropathic pain. Thus, our data do not exclude the possibility that iNOS at the ligature site might play a role in the development of neuropathic pain.
The mechanism by which DRG NOS-synthesized NO contributes to the development and/or maintenance of neuropathic pain is unclear. A gaseous molecular, NO diffuses out from the DRG neurons or primary afferent terminals and stimulates guanylyl cyclase in neighboring DRG or dorsal horn neurons to form cyclic guanosine monophosphate (cGMP) . cGMP activates several intracellular processes, including cGMP-dependent protein kinase (PKG), ion channels, and phosphodiesterases (PDEs) . PKG is present in the small DRG neurons and the superficial dorsal horn [38, 39]. Sung et al.  showed that PKG was activated in peripheral terminals of the DRG neurons and retrogradely transported to the DRG after peripheral nerve injury and inflammation; however, intrathecal injection of a PKG inhibitor did not affect nerve injury-induced mechanical hypersensitivity . Interestingly, intraplantar injection of a PDE-5 inhibitor significantly reduced mechanical hypersensitivity in diabetic neuropathy. This effect could be reversed by inhibition of NOS and guanylyl cyclase . In addition, NO modulates the DRG Ca2+ and Na+ channels via cGMP-dependent pathways [43–45]. Therefore, it is very likely that DRG NOS-synthesized NO contributes to the mechanism that underlies the development and/or maintenance of neuropathic pain by activating PDEs and/or ion channels, but not by activating PKG, in the DRG and spinal cord.
In summary, we have provided genetic evidence that the deficiency of nNOS protein impairs mechanical hypersensitivity during the development and maintenance of neuropathic pain. Pharmacologic studies further demonstrated that systemic or spinal administration of specific and non-specific nNOS inhibitors attenuates nerve injury-induced mechanical hypersensitivity. Moreover, peripheral nerve injury up-regulates nNOS expression in the DRG but not in the spinal cord. These findings suggest that the DRG nNOS may play a role in the mechanism underlying neuropathic pain.