Dear Ina -

Thanks for your kind note. I wish I could suggest that faster progress was being made in this area. However, I just came across a new study "Changes in immune and glial markers in the CSF of patients with Complex Regional Pain Syndrome," Guillermo M. Alexander, Marielle J. Perreault, Erin R. Reichenberger and Robert J. Schwartzman Brain Behav Immun. 2006 Nov 25; [Epub ahead of print], which suggests a somewhat less robust role for the role of IL6 (60% of all tested CRPS patients had elevated levels) than had been suggested by the authors' prior study, G.M. Alexander, M.A. van Rijn, J.J. van Hilten, M.J. Perreault and RJ. Schwartzman, "Changes in cerebrospinal fluid levels of pro-inflammatory cytokines in CRPS," Pain 116 (2005), pp. 213–219.

The "Discussion" portion of the most recent article - except for a couple of charts that wouldn't copy - follows:

The pathophysiology of CRPS is not well understood, but the evidence indicates that it includes different biological pathways involved in inflammation and central processing of afferent input. Inflammation, tissue damage or nerve lesions can lead to hypersensitivity and allodynia at the site of injury. In some individuals, the pain persists long after the initiating event has healed. There are a number of mechanisms that attempt to explain the pathophysiology of these chronic pain states. Some of these mechanisms are centered on neuronal sensitization (Ikeda et al., 2006 and Woolf and Salter, 2000) and others on neuroimmune interactions and the activation of glial cells (DeLeo and Yezierski, 2001, Marchand et al., 2005, Tsuda et al., 2005 and Watkins and Maier, 2005).

Studies in animal models of exaggerated pain demonstrate that following tissue injury or inflammation, glial cells (astrocytes and microglia) become activated (Tsuda et al., 2005). Some of the signaling molecules implicated in glial activation include; MCP1, fractalkine, ATP, pro-inflammatory cytokines, substance P and glutamate (Abbadie, 2005, Abbadie et al., 2003, Inoue, 2006, Klein et al., 1997, Nakajima and Kohsaka, 2001, Nishiyori et al., 1998 and Svensson et al., 2003). Once activated, microglia and astrocytes secrete a number of substances known to excite dorsal horn neurons and influence the establishment and maintenance of neuropathic pain (Watkins and Maier, 2000). These substances include pro-inflammatory cytokines, nitric oxide, excitatory amino acids, prostaglandins and ATP (Abbadie, 2005, Marchand et al., 2005 and Wieseler-Frank et al., 2004).

This study set out to investigate changes in levels of multiple biological markers in the CSF of individuals afflicted with CRPS and in patients suffering with other non-painful or painful conditions. The use of markers combined with the clinical examination is essential in determining the presence or absence of disease and monitoring its response to therapy.

CRPS patients in this study demonstrated elevated CSF levels of both glutamate and calcium when compared to normal individuals. The release of glutamate and substance P in the spinal cord dorsal horn following inflammation or tissue injury (Hunt and Mantyh, 2001 and Willis, 2001) can potentially lead to calcium dependent long term potentiation resulting in hyperalgesia (Ikeda et al., 2006). Increased extra cellular calcium has been shown to activate nitric oxide synthase (NOS), which is required for the maintenance of hyperalgesia in animal models of persistent pain (Meller and Gebhart, 1993). In addition, variations in extracellular calcium have also been shown to affect neurotransmitter quantal size as well as the probability of transmitter release at central synapses. This effect is mediated by group I metabotropic glutamate receptors (Hardingham et al., 2006).

We evaluated CSF levels of the pro-inflammatory cytokine IL-6, in order to extend our previous work showing an elevation in CSF IL-6 in many CRPS patients. The control patients for CSF IL-6 levels were individuals with radiculopathies, peripheral neuropathies, and NPH. These patients demonstrated CSF IL-6 levels (n = 18, 1.32 pg/ml) similar to previously reported values in normal human volunteers (n = 24, 1.27 pg/ml) (Steensberg et al., 2006). In agreement with our previous work, the CRPS patients in this study showed significantly (F(1, 20) = 8.59, p < 0.01) elevated levels of CSF IL-6 as compared to the control group. The increase in CSF IL-6 was not universal, and was only seen in approximately 60% of the CRPS patients in this study. It is also not specific, as patients with other conditions such as Alzheimer’s and Parkinson’s disease (Blum-Degen et al., 1995), seizures (Lehtimaki et al., 2004), sepsis (Verboon-Maciolek et al., 2006) and spondylolisthesis (this study) show significant increases in CSF IL-6 levels. However, our data do show that the elevation of CSF IL-6 in CRPS is greater than that seen in Alzheimer’s and de novo Parkinson’s disease where neuroinflammation has been proposed as part of the neurodegerative process (Blum-Degen et al., 1995).

With the exception of the NPH group, the CSF levels of the chemokine IL-8 were elevated in all of the patient groups in this study as compared to published values for normal control volunteers (15.5 pg/ml) (Natelson et al., 2005). The values for IL-8 in the CSF of the CRPS, radiculopathies, peripheral neuropathies, spondylolisthesis and ALS patients were comparable to levels reported in patients with postherpetic neuralgia (PHN) (35 pg/ml) (Kotani et al., 2000). In patients afflicted with PHN, the increase in CSF IL-8 correlates with both the degree of pain and the duration of disease (Kotani et al., 2000). In this study, IL-8 levels did not correlate (r(1, 49) = 0.06, p = 0.68) with pain levels (VAS scores) and significant differences were not noted (F(1, 49) = 0.29, p = 0.59) in CSF IL-8 levels between patients reporting chronic pain (n = 39, 39.7 pg/ml) and patients reporting no pain (n = 12, 33.7 pg/ml).

In this study, the patients in the ALS group demonstrated significantly greater (F(1, 48) = 7.4, p < 0.01) CSF levels of MCP1 as compared to all other patient groups. The CSF levels of MCP1 in the ALS patients in our study (518 pg/ml) are comparable to previously reported values for MCP1 in ALS patients (570 pg/ml) (Wilms et al., 2003). In their study, the CSF level of MCP1 in their control group (tension headaches) was 285 pg/ml, which is much less that the CSF level of MCP1 in all of the groups in our study, suggesting that all of the groups in this study, including the CRPS patients, demonstrate elevated CSF levels of MCP1.

Following injury or inflammation, MCP1 is expressed by both neurons (Zhang and De Koninck, 2006) and glial cells (Babcock et al., 2003). The major source of MCP1 expression comes from microglia and GFAP positive astrocytes (Babcock et al., 2003). GFAP, a protein member of the intermediate filament family, is strongly expressed in activated astrocytes and its level in CSF was also increased in all patient groups in this study when compared to published values for age-matched neurologically healthy individuals (Rosengren et al., 1994 and Anderson et al., 2003). Given that activated astrocytes are the source for GFAP and one of the major sources of MCP1, it is not surprising that they were positively correlated in the CSF of all patient groups (r(1, 47) = 0.45, p < 0.01), and especially in the CRPS group (r(1, 20) = 0.55, p < 0.01) (Fig. 1).

[Not shown] Fig. 1. Relationship between levels of the chemokine MCP1 (pg/ml) and the protein GFAP (ng/ml) in the cerebrospinal fluid of patients afflicted with CRPS. MCP1 and GFAP were positively correlated (r(1, 20) = 0.55, p < 0.01).

Most of the patients in this study showed elevated levels of nitrate plus nitrite with the patients in the CRPS group demonstrating the greatest elevation. In the CNS, IL-4 and IL-10 are expressed in reactive astrocytes and activated microglia (Hulshof et al., 2002 and Park et al., 2005). IL-4 and IL-10 have been shown to inhibit inducible NOS (iNOS) expression resulting in decreased NO synthesis by glial cells (Koeberele et al., 2004). A reduction of iNOS by IL-4 or IL-10 should result in an inverse correlation between their CSF levels and NO metabolites. There was no correlation between CSF NO metabolites and IL-10 in any of the patient groups (r(1, 42) = 0.06, p = 0.71). However, with the exception of the CRPS patients the CSF levels of NO metabolites (nitrate plus nitrite) in the study patients were inversely correlated with their CSF IL-4 levels (r(1, 14) = 0.60, p < 0.01) in contrast to the CRPS patients which did not show such a correlation (r(1, 11) = 0.32, p = 0.29) (Fig. 2). The lack of correlation between IL-4 and NO metabolites in CRPS patients may be due to the fact that IL-4 levels were not high enough to inhibit iNOS or that NO production in these patients resulted from the induction of other isoforms of NOS. It has been proposed that IL-4 regulates brain inflammation by inducing the death of activated microglia, (Park et al., 2005). The level of IL-4 expression in the CRPS patients may be insufficient to reduce brain inflammation and may be a contributing factor to the mechanisms responsible for the pathophysiology of CRPS.

[Not shown] Fig. 2. Relationship between CSF levels of the NO metabolites (nitrate plus nitrite) (uM) and IL-4 (pg/ml). All of the included samples had CSF IL-4 levels greater than the sensitivity of the ELISA (0.1 pg/ml). In the NPH, radiculopathies, peripheral neuropathies and spondylolisthesis patients (Controls ), NO metabolites were inversely correlated with CSF IL-4 levels (r(1, 14) = 0.60, p < 0.01). In the CRPS (•) patients, there was no correlation between CSF levels of NO metabolites and CSF IL-4 levels (r(1, 11) = 0.32, p = 0.29).

As a group, the CRPS patients in this study demonstrated elevated CSF levels of IL-6, IL-8, MCP1, GFAP, NO metabolites, glutamate and calcium. It was difficult to establish whether the CRPS patients demonstrated elevated or reduced levels of IL-10 and IL-4 as compared to individuals without neurological diseases, since normative values for these cytokines in the CSF range widely in the literature (Bartosik-Psujek and Stelmasiak, 2005, Natelson et al., 2005, Rota et al., 2006 and Stoeck et al., 2005). However, except for the IL-4 levels in the ALS group, the CSF levels of IL-10 and IL-4 in the CRPS patients were the lowest of all of the other disease groups.

There was no elevation or reduction of a CSF marker that was specific to the CRPS patients. However there were several patterns of markers that could be helpful in both elucidating the mechanisms involved in the disease process and supporting the diagnosis of CRPS. The most common pattern was found in 50% (11 out of 22) of the CRPS patients and consisted of; elevated IL-6, low levels of either IL-4 or IL-10, increased GFAP or MCP1 and increases in at least two of the following markers NO metabolites, calcium or glutamate. The second most common pattern was found in 18% (4 out of 22) of the CRPS patients and it consisted of; normal IL-6, low levels of either IL-4 or IL-10, increased GFAP or MCP1 and increases in at least two of the following markers NO metabolites, calcium or glutamate. A third pattern was found in 14% (3 out of 22) of the CRPS patients and it consisted of; increased GFAP or MCP1 and low levels of either IL-4 or IL-10. The remaining CRPS patients showed normal levels of GFAP and MCP1 and demonstrated at least one of the following; elevated IL-6, low levels of the anti-inflammatory cytokines or increases in either NO metabolites, calcium or glutamate.

There were several limitations of this study: (1) We studied a relatively small number of CRPS patients, and a larger sample is needed in order to determine which patterns are relevant to the pathophysiology of the disease; (2) We did not have CSF samples from normal control volunteers and in many cases had to compare CSF levels to previously reported values in healthy individuals. Using values from other studies is limited by the variability seen in the literature for levels of cytokines and chemokines in CSF. However, much of the variability is due to the lack of sensitivity of the methods employed. In this study, we made comparisons of our data to normative values from other studies that used methods with the highest sensitivity available and when possible from the same manufacturer as the ELISA kits we employed; (3) None of the CSF samples were from patients with early CRPS (less than 6 months); (4) We did not have CSF samples from the same individual at different time points in order to match CSF marker patterns with the severity of their symptoms and; (5) More sensitive assays are needed. There were a number of markers (IL-1β, TNF-α and fractalkine) that may have provided additional information, but in most samples their CSF levels were below the level of detection of the available assays.

Our hope is that the data obtained from this and other similar studies may aid in elucidating the mechanisms involved in the pathophysiology of CRPS. A better understanding of these mechanisms may lead to novel treatments for this very severe, life-altering illness. [Emphasis added.]

I would be happy to email a copy of the full article to anyone who wants it. Just send me a pm with your email address: the file is just slightly too large to attach here. (Personal, non-commercial use only, please.)

So now maybe have a better understanding as to why the allergist/clinical immulogist who saw me at Hopkins a couple of months ago wasn't as bowled over by "Changes in cerebrospinal fluid levels of pro-inflammatory cytokines in CRPS," Pain 116 (2005), pp. 213–219, as I had been walking in there. Call it a work in progress.

Mike