Cellular and behavioural models to predict responses to immunoglobulin G treatment in complex regional pain syndrome

Complex regional pain syndrome (CRPS) was first described during the American Civil War. Silas Weir Mitchell began to recognize unusual symptoms in soldiers with partial nerve injuries, such as the development of extreme pain in a distal limb, even when the acute injury had subsided. Today, cases of CRPS following partial nerve injury are rare, with the syndrome more often developing following non-nerve-injury trauma to a distal limb. Clinical presentation is extremely varied; the acute presentation can resemble septic inflammation. However, upon investigation there would be no neutrophils present and inflammatory markers are always normal. It is thought that this clinical picture is caused in part by neurogenic inflammation with anti-dromic substance P and calcitonin gene-related peptide (CGRP) secretion. A rare complication of this can be malignant oedema, which can lead to repeated skin infection and eventual amputation 1.

Treatment options for CRPS are limited and have low efficacy, especially in patients with long-standing CRPS (>1 year duration) who are much less likely to recover spontaneously. In recent years, an important role for immune mechanisms in sustaining chronic pain has been recognized, and evidence for immune involvement in CRPS suggests that immune modulation may be an effective treatment for the syndrome. A randomized clinical trial in 12 patients with long-standing CRPS set out to investigate the effect of intravenous immunoglobulin (IVIg), if any, on the symptoms of CRPS 2 and found that a subset of patients experienced important benefit. Twenty-five per cent (n = 3) of the subjects experienced an alleviation of their symptoms by more than 50%, while a further 17% (n = 2) experienced pain relief of between 30 and 50% (P < 0·001) 2. Based on earlier results 3, it was postulated that patients who responded well to the immunoglobulin (Ig) treatment may have been suffering from an autoimmune condition, with secretion of antibodies directed against peripheral sensory nerves. These pre-existing serum autoantibodies may synergize with the consequences of trauma to cause or sustain chronic pain.

In a first attempt to identify specific autoantibodies in long-standing CRPS, and as abnormal autonomic receptor activation may play a role in causing clinical CRPS signs, such as sweating and nail-growth changes, it was considered that CRPS-specific autoantibodies might cause autonomic receptor activation; of note, it is possible that this particular subset of antibodies would not necessarily be involved in the actual pain generation. Anti-autonomic autoantibodies had already been detected previously in serum samples from patients with short-term CRPS 4. To ascertain anti-autonomic autoantibodies in long-standing CRPS patients, a laboratory study was carried out using a novel adult cardiomyocyte model 5. Although cardiomyocytes are not involved in the CRPS pathophysiology, these cells are useful for detecting autoantibodies directed against autonomic receptors, as any functional receptor effect will be indicated by changes in the pattern of the cardiomyocytes' beatings. Cardiomyocytes treated with serum-IgG preparations from CRPS patients and controls (29 healthy patients, seven with neuropathic pain, nine with myasthenia and 12 with fibromyalgia) were placed into a pulsating electric field to induce calcium influx and contraction. In the CRPS cells, both the baseline calcium levels and the calcium transient were reduced; however, the level of cell contraction was the same as that of the control cells, suggesting calcium-independent myofibril sensitization. The calcium effect was confirmed in patch-clamping experiments where calcium influx was reduced in the CRPS group compared to the control preparations. Eleven of 18 CRPS serum-IgG preparations induced functional or calcium abnormalities, while only one in 57 control preparations induced abnormalities (P < 0·0001). These results suggest that long-term CRPS is associated with specific anti-autonomic autoantibodies. Discussions in the field have traditionally assumed that although there might be an immune involvement in the initial CRPS stages the patients' pain would later be maintained by brain factors but, conversely, our results argue that there is an ongoing, potentially treatable immune abnormality.

Additionally, of the 11 serum-IgG preparations available from CRPS patients who participated in the previous IVIg treatment trial 2, all preparations from subjects who responded to IVIg treatment (n = 4) were active in the cardiomyocyte assay, but the majority of preparations from non-responders to IgG (n = 4/7) were also active. This therefore indicates that CRPS-specific autoantibodies are not restricted to IVIg responders.

The study group also investigated the effect of CRPS serum-IgG in a novel animal model via passive transfer 6. Serum-IgG preparations from 12 CRPS patients and 12 controls from the previous trial were administered to mice. Behaviour in the open field, stimulus-evoked pain and motor co-ordination were observed in order to ascertain whether the transfer of IgG antibodies produced signs of CRPS. Rearing behaviour was reduced significantly in the CRPS-IgG-treated group, and motor impairment was also observed; however, these mice were not suffering from CRPS, as assays for hyperalgesia revealed no results. Because injury or trauma normally precedes CRPS, in a modified approach mice were again injected with either saline, healthy-IgG or CRPS-IgG, and the study group then performed an incision of the skin and muscle of animals' right hind paws 7. This caused significant changes; the CRPS animals developed mechanical hyperalgesia and increased oedema compared to the controls in the traumatized limb only.

The CRPS mice additionally developed markedly raised levels of substance P (which has been implicated in CRPS development, with abnormally high substance P activity observed previously in the skin of CRPS-patients' affected areas) in their operated paws (mean difference to not-operated limb 7·5 fmol/mg, P < 0·001) 7. This shows that passive transfer of CRPS to rodents using serum-IgG from patients with long-standing CRPS elicits important signs reflecting the clinical disease.

In this behavioural passive transfer assay, similar to the cardiomyocyte model, it was shown that preparations from CRPS subjects, but not controls, are active regardless of IVIg response. Of the six serum-IgG preparations taken from patients with long-standing CRPS, one was from an IVIg responder, one was from a responder who later became a non-responder, one was from a non-responder and three were from patients who had never had IVIg. All these sera were active, in that in all groups the CRPS-injected mice developed abnormalities compared to the control mice. It is therefore possible that some non-responders to IVIg therapy can be treated with other anti-autoimmune interventions.

Disclosure

A. G. has received grant support, travel support, speaker fees and consultancy fees from CSL Behring, Biotest, BPL, Baxter, Grifols, Axsome and Pfizer.