The purposes of this study are to determine (1) if the T394 site in MORs is involved in opioid tolerance and (2) if the application of DNA plasmids to upregulate receptors is a feasible way to study behavioral consequences of the expression of mutant receptors in sensory neurons. Viral vectors and mutant mice are two approaches commonly used to alter receptor expression. Aside from the concerns of the toxicity of viral vectors and possible compensatory adaptations in mutant mice, those approaches are complex. We sought a simpler method to accomplish the goal. Using hMOR as an example, we introduced mutant hMOR-T and hMOR into L4-L5 DRGs by intrathecal injection of DNA plasmid. A neuron-specific promoter was used to limit exogenous hMOR expression in neurons (Figure 2). Receptor upregulation could be seen within 3 days following the plasmid injection. The expression of hMOR-T and hMOR were found in small and medium DRG neurons which are known to mediate nociception. The time course of morphine effect was not altered by the mutation (Figure 1). Compared with GFP-injected rats, the upregulation resulted in 3.0- and 2.3-fold increase in morphine potency in hMOR-T and hMOR rat groups respectively (Figure 3), a magnitude analogous to that observed in DRGs transduced with adeno-associated viral vector (AAV) expressing MORs . We showed here that the development of acute and chronic morphine behavioral tolerance is independent of the phosphorylation of T394 in MOR (Figure 4), a result similar to the findings obtained in S375A MOR mutant mice . Thus, phosphorylation of the single site, i.e., T394 in hMORs, is not sufficient to alter the tolerance effect of morphine in vivo.
Even though our plasmid injection resulted in a substantial increase in the expression of hMORs in DRG neurons (Figure 2), one needs to ascertain that similar reduction in the morphine ED50 and identical morphine-induced tolerance observed in hMOR and hMOR-T DRG neurons were not a result of endogenous rat MORs masking morphine effects in the T394 mutant. The studies were therefore repeated using the full agonist, DAMGO. We found that DAMGO ED50 (Figure 5) and DAMGO-evoked IK desensitization (Figure 7) in hMOR and hMOR-T DRGs are significantly different. Thus, T394 is important in determining DAMGO affinity and desensitization. The observations also suggest that the expression of hMOR-T in DRGs was sufficiently large to exhibit its opioid agonist-dependent responses. Since the development and extent of DAMGO-induced tolerance were similar in hMOR and hMOR-T neurons (Figure 6), as observed in morphine-induced tolerance (Figure 4), T394 in the hMOR appears to play a limited role in opioid-dependent tolerance.
Birdsong et al.  reported that prolonged treatment of cells with high-efficacy agonists, which causes MOR desensitization, increases the agonist affinity for MOR. In our study of opioid dose–response relation in hMOR and hMOR-T, DRGs were exposed to agonists longer than 15–30 min. Thus, the dose–response curves were obtained under desensitization conditions. We found that DAMGO had a lower affinity for hMOR-T than that for hMOR (Figure 5), a result consistent with the observation that DAMGO induced less desensitization in hMOR-T cells (Figure 7).
The advantage of our approach includes that DNA plasmids are easy to produce. With proper promoter design, the mutant receptors expression can be targeted to specific cell types (Figure 2). In contrast to the long waiting period (> 3 weeks) for viral vector-induced expression of mutant MORs, the expression reaches its peak level within 3 days of plasmid injection. Using a slow rate of intrathecal plasmid injection procedure (see Methods), the expression of mutant MORs can be largely confined within L4-6 DRGs. It has been well documented that DNA plasmids are seldom incorporated into the genome, have low immunogenicity and rarely produce ill effects in host animals . There are a couple of noticeable limitations to the plasmid injection approach. One is that the expression level of mutant receptors is maintained for a short period. Following a single injection of i.t. hMOR-T or hMOR, the enhancing effect was reduced by 40% in ten days and dissipated in fourteen days. It has been shown that two injections of anti-inflammatory cytokine IL-10 DNA would prime immune responses and result in long term (up to 90 days) upregulation of IL-10 [13, 14]. Unlike the immune mediators, double injections of mutant MOR DNA did not lengthen the duration of the expression (unpublished observation). Nevertheless, the expression of exogenous receptors could be maintained at a steady level following repeated injections of the DNA plasmid every seven days (data not shown). Another possible limitation is that changes induced by the mutant MOR have to be sufficiently large to overcome the influence of endogenous MORs. Since the differences in DAMGO ED50 and IK desensitization in hMOR-T and hMOR DRGs were readily observed (Figures 5 and 7), changes in opioid responses induced by mutant hMORs are likely to be detected under our experimental conditions.
We and others have found that the mutation of a single phosphorylation site, i.e., T394 or S375, was not sufficient to alter chronic behavioral tolerance (Figures 4 and 6) . It is of interest to determine if phosphorylation of other or multiple sites in MOR would alter morphine tolerance. In addition to T394 and S375, a number of serine and threonine sites, e.g., T357, S363 and T370, have been localized in the C-terminus of MOR [15–17]. In model cells, these sites were found to be phosphorylated by different kinases. In a mass spectrometric analysis phosphorylation of the T394 site was not detected . This result is inconsistent with the observation that DAMGO-induced phosphorylation of T394A mutant receptors is much reduced [4, 18]. Additional studies are needed to resolve the inconsistency. The consequences of the phosphorylation of serine and threonine sites in receptor desensitization have been extensively studied in vitro. The opioid-induced desensitization was found to diminish in T394A and S375A mutant receptors [4, 6, 7, 18]. Furthermore, the mechanism underlying receptor desensitization depends on the agonist used [19–21]. In locus ceruleus neurons, morphine-induced desensitization was shown to depend on PKC whereas DAMGO-induced desensitization is mediated by GRK2 and independent of PKC . It is suggested that DAMGO- and morphine-bound MORs assume differential conformations that are targeted by different GRKs to give rise to various efficiencies in receptor internalization [16, 21, 22] and changes in the responses to PKC . We also found that IK desensitization in hMOR expressing DRG neurons depends on opioid agonists. T394 mediates DAMGO-induced desensitization, but does not affect morphine action on IK (Figure 7). The mechanism underlying the agonist-dependent desensitization and its effect on opioid-induced analgesic behaviors have yet to be explored.
The relationship between MOR desensitization and cellular tolerance has been studied in cells or slices obtained from chronic morphine treated animals . Cellular tolerance was defined as a sustained decrease in the morphine efficacy for its coupling effector, e.g., GIRK channels  or as right-shift in the morphine dose–response curve . Cellular tolerance and desensitization were found to be separate processes that occur simultaneously  and β-arrestin2 was shown to impair the MOR resensitization and recycling [25, 26]. The relationship between cellular and behavioral tolerance is not entirely clear. The information will be essential for our understanding of the mechanisms underlying behavioral tolerance. Using our approach, the behavioral tolerance of various mutants can be determined and its relationship with cellular tolerance and desensitization assessed. This knowledge will be useful in designing ways to improve morphine usage in clinics.