The main findings of this study are: (1) The spared nerve injury model caused a rapid and long-lasting neuropathy with mechanical and cold allodynia; (2) ARA 290 produced dose-dependent relief of both mechanical and cold allodynia; (3) A spreading microglia response (i.e. iba-1-IR and phenotype) was apparent from L5 at week 2 following nerve damage to L2-6 at week 20; (4) No effect of nerve injury on the astrocyte response was observed at weeks 2 and 20 following nerve damage; and (5) ARA 290 suppressed iba-1R in a dose dependent manner.
Neuropathic pain in animals (due to experimental nerve damage) and humans (due to sarcoidosis or diabetes mellitus type 2) responds well to treatment with ARA 290, in that it produces relief of spontaneous pain (humans) and allodynia (humans and animals) [25, 26, 31, 35]. Studies in mice that lack the β-common-receptor show further that ARA 290 is without behavioral effect (ie. allodynia is not relieved by ARA 290), implicating this receptor as site of action of ARA 290 [25, 26]. The β-common-receptor forms a heterocomplex together with EPO receptor and it is believed that this receptor complex, which we designate the innate repair receptor (IRR), is the molecular site of action of both EPO and ARA 290 [27, 29, 36]. Exogenous EPO, similar to ARA 290, reverses allodynia and reduces neuronal apoptosis and proinflammatory cytokine production, neuronal regeneration and the release of anti-inflammatory cytokines . We do not use EPO in our studies as, in contrast to ARA 290, it comes with severe side effects including enhanced hematopoiesis and cardiovascular complications (eg. hypertension, thrombosis, myocardial infarction). In common with previous studies [25, 26], we show here that ARA 290 has effective and prolonged (up to 20 weeks) anti-allodynic effects.
There is ample evidence that peripheral nerve injury results in a strong spinal inflammatory response . For example, we previously showed in mice that surgical damage to the sciatic nerve causes the increase of expression of pro-inflammatory markers including iba-1 mRNA, GFAP mRNA and CCL2 mRNA, within 7 days following nerve damage . CCL2 plays an important role in the invasion of monocytes from peripheral blood as well as resident macrophages towards the spinal cord lesion site following peripheral nerve damage. In our current study the inflammatory response following SNI was apparent from the increase in iba-1-IR. The iba-1-IR response showed a marked expansion from level L5 in week 2 following SNI, to 5 adjoining segments, L2 to L6, at week 20. In addition to the spreading of iba-1-IR to multiple segments, the intensity of the response also increased over time as shown by a higher degree of iba-1-IR and phenotypic signs of activation. We are the first to show this spreading inflammatory response in the spared nerve injury model of neuropathic pain. Similar observations were made earlier in experimental models of spinal cord injury and nerve root avulsion [37, 38]. Previous reports of glial response following peripheral nerve injury showed that the response area is confined to the spinal cord segments innervated by the damaged nerve [39, 40]. However, these responses were measured within a 2-week time frame. This is in agreement with our observation of lack of spreading at week 2. Caudal and cranial expanding inflammation, as observed here, may explain the increase in severity of NP symptoms over time and development of symptoms in areas of the body not innervated by the damaged nerve . Similarly, various experimental reports indicate that the inflammatory responses may spread to contralateral spinal cord areas [8, 42]. In this study we did not quantify contralateral inflammation. Our data do suggest that time is an important factor in the spreading of the microglia response.
Iba-1-IR reflects microglia activation in addition to localization and morphology. Our data show increased iba-1-IR after SNI, which is dose-dependently and long-term reduced by ARA 290 treatment coupled to a dose-dependent and long-term reduction of mechanical and cold allodynia. This long-term effect suggests a disease modulatory effect of ARA 290. We argue that ARA 290 initiates a cascade of events involving several transduction factors of which activation of the IRR is the first step that eventually silences or reduces the inflammatory process [29, 36]. Since the activation or recruitment of microglia is largely mediated through the local production of CCL2 [43, 44], a possible scenario is that ARA 290 reduces the release of CCL2 via activation of the IRR on neuronal and immune cells . However, both at 2 and 20 weeks after SNI and ARA 290 treatment, relief of allodynia was not complete, indicating that the central response to peripheral nerve damage involves multiple systems including neuroinflammation and probably also up-regulation of excitatory pathways and synaptic plastic changes. Of interest is that ARA 290 treatment causes a reduction in NMDA mRNA (subunits NR1, NR2A and NR2B) in SNI animals, suggestive of an additional role, apart from immune-modulation, for ARA 290 in the treatment of neuropathic pain by suppression of excitatory glutamatergic activity .
In contrast to a markedly increased iba-1-IR after SNI, no change in GFAP-IR or astrocytic phenotype (ie. activation) was observed in animals with an SNI treated with vehicle after 2 and 20 weeks of lesion. This observation stands in contrast with reports describing involvement of astrocytes adjacent to microglia in NP [5, 7, 10, 45, 46]. The absence of astrocytosis in our SNI model may be explained by a time-limited astrocyte response (ie. < 2 weeks or between 2–20 weeks). The involvement of astrocytes following peripheral nerve injury reported in the literature varies with some studies showing a relatively short-lived increase in GFAP-IR , while others show an increase in GFAP-IR after 14 days that was still present after 150 days . Further studies using more dense observations over time are required to get a reliable indication of the kinetics of the astrocyte response to peripheral nerve damage.
We have argued that the spinal cord is the predominant site of action of ARA 290 following peripheral nerve damage. Indeed, there is ample evidence that peripheral nerve injury activates an innate immune response activated in the spinal cord [5–11]. Furthermore, a complete block of the peripheral nerve with local anesthetics will not prevent central inflammation following peripheral nerve damage but only delays the development of pain, suggestive of a predominant central effect [47, 48]. Still, at this point we cannot exclude an additional peripheral effect of ARA 290. Indeed, EPO specifically reduces axonal TNF-α in Schwann cells after peripheral nerve injury, resulting in attenuation of NP symptoms . Hence, in addition to a central nervous system effect, modulation of the peripheral nerve immune response could also be part of the mechanism of action of ARA 290, but these specific effects remain to be investigated.