The purpose of this study is to obtain increased knowledge concerning central pain and fatigue processing in rheumatoid arthritis, and how these conditions are influenced by treatment with Tumor Necrosis Factor (TNF) blockade with adalimumab.

Background

Rheumatoid arthritis (RA) RA is characterized by joint inflammation causing peripheral pain, however the relation between peripheral pathology and pain intensity is weak. There is substantial evidence that also pain modulation mechanisms are dysregulated in RA. For example, generalized pain frequency is higher in RA than in normal population, about 15% compared to 3%. Earlier data from our group show a changed pattern of pain-modulation in established RA compared to early RA and controls (Keystone 2004).

There are several treatments in RA directed at relieving symptoms (NSAIDs) as well as modifying disease (DMARDs and biologics). One of the established biologics is the Tumor Necrosis Factor (TNF)-blocking agent adalimumab which has proven efficacious for decreasing disease activity and retardation of joint destruction in RA (Keystone 2004, Weinblatt 2006). In addition, adalimumab has proven efficient in reduction of fatigue in RA (Yount 2007)

Pain mechanisms in rheumatoid arthritis

Pain in RA has traditionally been regarded as nociceptive pain due to peripheral inflammation of joints (arthritis). However, although affected joints are typically inflamed and swollen, peripheral pathology cannot fully explain the amount of pain a patient experiences (Thompson 1997). Central nervous system (CNS) mechanisms have been implicated in the pain experience by RA patients and we have previously been able to demonstrate a generalised increase in pain sensitivity in RA patients (Leffler 2002). The importance of CNS changes in endogenous pain modulation (i.e, facilitation, central sensitisation/disinhibition) is further underscored by the high co-morbidity between RA and generalised pain syndromes such as fibromyalgia (FM)(Neumann and Buskila 2003), where specific dysfunctions of endogenous pain modulation have been documented. The pronounced co-morbidity between FM and rheumatic inflammatory diseases, and the fact that a generalised increase in pain sensitivity has been reported in patients with long (> 5 years) but not short (< 1 year) duration of RA (Leffler 2002) suggest that disturbances in pain modulation can also be important for pain perception in many patients with RA. Endogenous pain inhibitory mechanisms are physiologically closely linked to autonomic nervous system activity. The same type of autonomic nervous system dysfunction (i.e., basal sympathetic hyperactivity and hyporeactivity) has been reported in RA patients (Evrengul 2004, Goldstein 2007) and FM patients (Cohen 2001).

Besides pain regulation, recent studies have focused on the involvement of autonomic tone also in systemic inflammatory responses. A major breakthrough was the characterization of the cholinergic anti-inflammatory pathway. Efferent activity in the vagus nerve leads to systemic release of acetylcholine (ACh). Specific binding to alfa 7 nicotinic ACh receptors on tissue macrophages efficiently inhibits the release of proinflammatory substances such as TNF, IL-1, IL-6 and IL-8 (Tracey 2002). This inflammatory reflex pathway is rapid, localized and integrated in contrast to the well defined humoral regulation of inflammation. Dysfunction of autonomic regulation, with a relative reduction of vagal tone is a feature of RA and proven to correlate with systemic levels of inflammatory cytokines (Goldstein 2007). The importance of interaction between autonomic pain regulation and neural anti-inflammatory pathways is indicated by several findings. Flight-or-fight responses activated by pain can result in increase of epinephrine and norepinephrine, that can inhibit macrophage activation and regulate TNF secretion (Tracey 2002). Moreover, a reduction of antiinflammatory cytokines (IL-4 and IL-10) (Uceyler 2006) and an increase in pro-inflammatory cytokines (TNF, IL-1, IL6, IL-8) have been shown in syndromes characterized by generalized pain (Salemi 2003, Bazzichi 2007). Thus, these data indicate that modulation of systemic and chronic inflammation may have significant effects on pain processing.

In recent years increasing focus has been placed upon the interactions between the immune system and the CNS with the recognition of an immune - to brain pathway activating the sickness response. In addition to blood-borne signalling, neural pathways via the sensory vagus and glossopharyngeal nerves have been identified (Watkins 2005). Recent animal studies have revealed that peripheral inflammation, nociceptive input as well as pronounced mental stress can activate microglia cells throughout the CNS and that these stimulated microglia cells start to produce pro-inflammatory substances (TNF, IL-

1, IL-6). These substances initiate central inflammatory changes with increased pain sensitivity (allodynia, hyperalgesia and most likely spontaneously ongoing pain) as a result (Milligan 2009, Beattie 2002). Moreover, intrathecal administration of substances potentiating release of antiinflammatory cytokines such as IL-10, have been proven efficiently analgesic (Johnston 2004).

These data are in accordance with recent results where collagen induced arthritis resulted in a decrease in the threshold for thermal and mechanical stimuli, beginning on the day of onset. This was accompanied by increased inflammatory response in spinal cord astrocytes, and treatment with anti-TNF was both profoundly analgesic and also restored inflammatory changes in astrocytes (Inglis JJ 2007). These results thus indicate that pain regulation may be influenced by peripheral inflammation and can also be reversed by potent anti-inflammatory therapy.

Altogether, understanding the relationship between peripheral nociceptive input (due to inflamed joints), autonomic NS activity, endogenous pain modulatory systems and CNS inflammation for the symptoms in patients with RA is important and can lead to improved strategies in treatment of pain mechanisms as well as fatigue.

Imaging Pain can be conceptualised as a primarily motivational state to induce a behavioural drive with the purpose to restore homeostasis. The multidimensionality of pain perception is supported by studies using functional magnetic resonance imaging (fMRI) to assess cerebral activation during stimulus-evoked pain in human subjects. These studies have documented activation of brain areas traditionally associated with the perception of sensory features (e.g. somatosensory cortices) and regions associated with emotional and motivational aspects of pain (prefrontal cortex (PFC), anterior cingulate cortex (ACC) and insular cortex (IC)). In addition, brain areas involved in the regulation of autonomic nervous system and endogenous pain modulation are also activated. In a recent fMRI study of patients with chronic low back pain (LBP) activation of the somatosensory cortex was only seen during brief periods of spontaneously increasing pain intensity. During periods of ongoing LBP only brain regions of importance for emotional and cognitive aspects of pain (i.e. prefrontal cortex and cingulate) were activated (Baliki 2006). These findings indicate that the perception of chronic, ongoing pain requires very limited involvement of the pure somatosensory areas. These results are in accordance with two recent positron emission tomography (PET) studies of joint pain. In patients with pain due to knee osteoarthritis, increased activity was reported in the prefrontal cortex, cingulate and insula during arthritic pain compared to experimental heat pain (Kulkarni 2007). Accordingly, in another PET study, Schweinhardt et al.

(Schweinhardt 2008) reported a stronger activation in the dorsolateral prefrontal cortex in response to provoked joint pain in RA patients compared to experimental heat pain. The results were interpreted as increased activation of areas implicated in emotional and cognitive processing of noxious stimuli, reflecting the greater emotional salience of clinical pain compared to experimental pain and the importance of coping strategies potentially influencing the perception of chronic pain. In healthy volunteers, activation of the prefrontal cortex is related to the degree of anxiety during the experience of experimental pain (Ochsner 2006) and the prefrontal cortex has been reported to be more consistently activated in studies of clinical pain compared to studies of experimental pain (Apkarian 2005). Little is known about the relation between cerebral pain related activity, clinical symptoms, degree of systemic inflammation and various coping strategies.

The aim of the study We hypothesise that RA patients have a generalised increase in pain sensitivity (i.e., not restricted to inflamed joints) and that this increase in pain sensitivity is mediated through increased central nervous system inflammation affecting the ability to recruit endogenous pain modulatory mechanisms in a negative way. Furthermore, we hypothesise that treatment with TNF antagonists, in addition to reductions in peripheral inflammation, also affect central nervous system inflammation and improve the ability to activate endogenous pain modulatory mechanisms thus reducing pain and other symptoms (i.e., fatigue).

The study adresses the following questions:

1. Do patients with RA exhibit an increased sensitivity to randomized painful stimulation at an inflamed joint and at a neutral painfree area, resp., compared to healthy controls?
2. Does the cerebral activation pattern assessed by fMRI differ between patients and controls during experimental painful stimulation at a painful (inflamed joint) and a non-painful (thumbnail) area?
3. Can disturbances in autonomic nervous system activity be linked to the pain sensitivity, cerebral processing of noxious stimuli and/or the degree of inflammation in RA patients?
4. Can the subjective perception of fatigue be related to any of the following; pain sensitivity, abnormalities in cerebral pain processing, inflammatory parameters and autonomic activity?
5. How are the above parameters affected by treatment with a TNF antagonist (adalimumab) compared to placebo?

Patients with active RA and incomplete response to methotrexate (MTX) will be recruited to the study as well as healthy clinical subjects. RA patients and controls will be compared concerning pain pressure thresholds, pain processing using brain fMRI, activity in the autonomous nerve system and inflammatory markers in blood samples as well as assessment of pain, fatigue using standardized assessments. Following inclusion, patients with RA will be randomized into active treatment (A) and placebo control group (B) using standardized procedures for randomized, double blind, placebo controlled studies. The treatment group will receive adalimumab directly following the baseline assessment and subsequently eow during the study. The placebo control group will receive placebo injections at baseline, two weeks and at four weeks and subsequently adalimumab. The overall aim with the study is to obtain increased knowledge concerning pain processing and fatigue in RA, and how these conditions are influenced by treatment with TNF blockade with adalimumab.

Inclusion Criteria:

• Age ≥18
• Fulfilling American College of Rheumatology (ACR) criteria for RA.
• Disease duration ≤ 5 years.
• Either under treatment with methotrexate (in a maximum tolerable up to 20 mg/week orally or subcutaneously), or previous treatment with methotrexate withdrawn due to documented side effects.
• Patients should be bio-naïve.
• Disease activity: Disease Activity Score (DAS28)>3.2 and Swollen joint count (SJC)>1 and Tender Joint Count (TJC)>1.

Exclusion Criteria:

For fMRI - left handedness and all forms of metallic implants.

• Fulfilling ACR criteria for fibromyalgia.
• Severe ischemic heart disease.
• Concurrent treatment for depression/anxiety with antidepressant drugs.
• Contraindication to adalimumab.
• Active or latent tuberculosis.
• Chronic infections including hepatitis B or C.
• Malignancy, multiple sclerosis, Systemic lupus erythematosus.
• Other reason as evaluated by the PI.