Blockade of TNF-Α Rapidly Inhibits Pain Responses in the Central

Blockade of TNF-α rapidly inhibits pain responses in SEE COMMENTARY the central nervous system

Andreas Hessa,1, Roland Axmannb,1, Juergen Rechb,1, Stefanie Finzelb, Cornelia Heindla, Silke Kreitza, Marina Sergeevaa, Marc Saakec, Meritxell Garciac, George Kolliasd, Rainer H. Straube, Olaf Spornsf, Arnd Doerﬂerc, Kay Brunea, and Georg Schettb,2

aInstitute of Experimental and Clinical Pharmacology and Toxicology, bDepartment of Internal Medicine 3, and cDivision of Neuroradiology, University of Erlangen-Nuremberg, 91054 Erlangen, Germany; dInstitute of Immunology, Alexander Fleming Biomedical Sciences Research Center, 16672 Vari, Greece; eDepartment of Internal Medicine I, University of Regensburg, 93053 Regensburg, Germany; and fDepartment of Psychological and Brain Sciences, Programs in Neuroscience and Cognitive Science, Indiana University, Bloomington, IN 47405

Edited* by Charles A. Dinarello, University of Colorado Denver, Aurora, CO, and approved December 29, 2010 (received for review August 19, 2010) There has been a consistent gap in understanding how TNF-α neu- which is a central inducer of inflammation and structural damage tralization affects the disease state of arthritis patients so rapidly, in arthritis (5), appears to have less-pronounced and less-rapid considering that joint inflammation in rheumatoid arthritis is a effects on the symptoms of RA, although its role in protecting chronic condition with structural changes. We thus hypothesized that structural changes in the joints is substantial (6, 7). neutralization of TNF-α acts through the CNS before directly affecting These observations, and the fact that traditional antirheumatic joint inflammation. Through use of functional MRI (fMRI), we dem- drugs such as methotrexate achieve far slower clinical responses onstrate that within 24 h after neutralization of TNF-α, nociceptive than TNF-α blockers, have led us to suspect that blockade of CNS activity in the thalamus and somatosensoric cortex, but also the TNF-α could has additional effects that go beyond the sole in- activation of the limbic system, is blocked. Brain areas showing hibition of joint inflammation. We hypothesized that TNF-α block- blood-oxygen level-dependent signals, a validated method to assess ade could influence processes in the CNS, which control the neuronal activity elicited by pain, were significantly reduced as early patient’s perception of the disease state. Indeed, earlier observa- as 24 h after an infusion of a monoclonal antibody to TNF-α. In con- tions in mice have suggested that TNF-α can induce a depressive-like trast, clinical and laboratory markers of inflammation, such as joint behavior in mice (8). Pain processing and sensation are key mech- swelling and acute phase reactants, were not affected by anti-TNF-α anisms in the CNS elicited by arthritis, and pain is the dominant at these early time points. Moreover, arthritic mice overexpressing symptom of disease, which by far has the strongest impact on the human TNF-α showed an altered pain behavior and a more intensive, disease burden of arthritis. Importantly, effects on pain also drive the widespread, and prolonged brain activity upon nociceptive stimuli standard read-out parameters for measuring therapeutic response in compared with wild-type mice. Similar to humans, these changes, arthritis. Thus, if TNF-α blockade has unique pain-reducing prop- as well as the rewiring of CNS activity resulting in tight clustering erties, its efficacy in reducing clinical disease activity of RA will in the thalamus, were rapidly reversed after neutralization of TNF-α. be high, as such disease-activity measures are strongly influenced These results suggest that neutralization of TNF-α affects nociceptive by pain. In consequence, other treatment strategies, lacking such brain activity in the context of arthritis, long before it achieves anti- CNS effects, will less strongly affect the patient’s disease state, even if inflammatory effects in the joints. showing similar anti-inflammatory or structure-sparing effects. Interestingly, inflammation is considered to lower the thresh- cytokines | antiinflammatory therapy old for pain perception in the CNS, leading to hyperalgesia (9). In contrast to the well-documented role of TNF-α as a proin- rthritis is one of the most disabling chronic human diseases. flammatory cytokine, its role as a mediator of pain is in- ARheumatoid arthritis (RA) affects up to 1% of the population completely characterized. It is known that both receptors for and is characterized by pain, swelling, and stiffness of joints, TNF-α are expressed in dorsal root ganglion neurons and that leading to a serious decay of life quality. Pain is the initial and intra-articular injection of TNF-α blocking agents reduces noci- fl α prevailing symptom of the disease, leading to immobility, which in ception elicited by joint in ammation (10). TNF- is also in- turn causes complications such as osteoporosis and cardiovascular volved in development of mechanical hyperalgesia following α disease. During the last 10 y the pharmacologic treatment of dis- injury (11, 12). Hence, we hypothesized that TNF- leads to eases such as RA has substantially improved because of the de- a major change in pain perception in the CNS during arthritis. α velopment of cytokine blocking agents (1). To search for the rapid effects of TNF- blockade, we visual- Although inflamed joints express a multitude of mediators, in- ized cerebral nociception elicited by arthritis and its reversibility α MEDICAL SCIENCES cluding cytokines, chemokines, and growth factors, which con- upon TNF- blockade using blood-oxygen level-dependent α (BOLD) functional MRI (fMRI) as early as 24 h after initiation of tribute to the pathogenesis of arthritis, inhibition of TNF- has fl emerged as a particularly successful therapeutic strategy (2, 3). the treatment. The BOLD signal re ects changes in hemody- namics linked to increased or decreased neuronal metabolic ac- The therapeutic success of TNF-α blockade in RA is unique and has been largely considered to result from rapid and efficient neutralization of joint inflammation based on breakdown of the fl Author contributions: K.B. and G.S. designed research; A.H., R.A., J.R., S.F., C.H., S.K., in ammatory cytokine network in the affected joint, which results M. Sergeeva, M.G., and G.S. performed research; A.H., G.K., and G.S. contributed new in an improvement of the signs and symptoms of the disease (4). reagents/analytic tools; A.H., R.A., J.R., M. Saake, R.H.S., O.S., A.D., K.B., and G.S. analyzed It has, however, always been stunning, how fast the blockade of data; and A.H. and G.S. wrote the paper. TNF-α improves the patient condition, in particular because dis- The authors declare no conflict of interest. eases like RA are highly chronic, building up a vast amount of *This Direct Submission article had a prearranged editor. inflammatory tissue, and leading to irreversible damage of the See Commentary on page 3461. cartilage and the bone. Thus, rapid resolution of this highly or- 1A.H., R.A., and J.R. contributed equally to this work. ganized inflammatory tissue or tissue damage is very unlikely to 2To whom correspondence should be addressed. E-mail: [email protected]. α explain the fast effect of TNF- blockade. Even more importantly, This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. neutralization of other inflammatory mediators, in particular IL-1, 1073/pnas.1011774108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1011774108 PNAS | March 1, 2011 | vol. 108 | no. 9 | 3731–3736 Downloaded by guest on September 30, 2021 tivity in response to outer sensory stimulation, thus allowing lo- activated brain area size after TNF-α blockade could be attrib- calization of brain areas activated by stimuli (13, 14). As a proof- uted to the CNS regions contralateral to the affected joint (Fig. 2 of-concept, we first evaluated whether TNF-α blockade rapidly A–C). No significant side difference was found for finger tapping, reverses the hypernociception in patients with RA. We then which was used as a control procedure (Fig. 2 A–C). TNF-α performed an in-depth analysis of this rapid effect of TNF-α blockade did particularly change the activity of CNS structures, blockade in a mouse model of arthritis based on the over- which are involved in pain perception, such as the thalamus expression of human TNF-α. and the secondary somatosensoric cortex (Fig. 2 D and E). In- terestingly, centers of the limbic system were also affected, such Results as the cingulate and insular cortex. The cingulate cortex is par- Reduction of Nociceptive Responses in the CNS of Arthritis Patients ticularly involved in forming and processing of emotions and by TNF-α Blockade Preceding Its Anti-Inflammatory Effects. To test memory, whereas the insular region is important for the sub- whether TNF-α blockade indeed affects the CNS, we first per- jective sense of the inner body, including the experience for pain formed a proof-of-concept study in humans. We measured CNS and emotions (Fig. 2D). These findings suggest that not only pain activity in patients with RA by assessing the BOLD signal in the responses are affected by TNF-α blockade but also CNS struc- fMRI before and after exposure to an infusion of 3 mg/kg of the tures relevant for creating pain perception, emotions, and body TNF-α blocker infliximab (IFX) (15, 16). The BOLD signal was experience. induced by compression of the metacarpophalangeal joints of the dominantly affected hand, which is a standard assessment pro- Nociceptive Responses in Arthritic Mice Overexpressing Human TNF-α cedure for RA. Interestingly, the size of the brain area with and Their Rapid Reversal by TNF-α Blockade. To study the effect of enhanced BOLD activity was significantly reduced as early as TNF-α blockade on CNS pain responses elicited by arthritis in 24 h after administration of the TNF-α blocker IFX. Moreover, it more detail, we used an animal model, which is characterized by remained low up to 6 wk after the first infusion (Fig. 1A). In an overexpression of human TNF-α in the peripheral joints be- contrast, BOLD signals elicited by a standardized control pro- cause of knocking the human TNF-α gene into the mouse genome cedure, such as finger tapping, were not affected by TNF-α (17, 18). These TNF-α transgenic (TNFtg) mice spontaneously blockade (Fig. 1B). Importantly, clinical measures of disease ac- develop inflammatory arthritis, which leads to progressive func- tivity, such as joint swelling and joint tenderness, composite disease tional impairment. The disease itself manifests as progressive activity scores (DAS), such as DAS28, and acute phase reactants, swelling of the peripheral joints associated with decreased grip such as blood sedimentation rate, serum C-reactive protein levels, strength, mirroring progressively impaired joint function (Fig. and serum IL-6 levels, did not change measurably within 24 h S1). We first assessed the effect of TNF-α overexpression on after TNF-α blockade but improved later during the treatment (Fig. behavioral pain responses and their reversibility upon TNF-α 1 C–E and Table S1). In contrast, subjective rating of pain inten- blockade by IFX. Spinal reflex arches, as measured by the tail- sity using the visual analog scale (VAS) was decreased within 24 h flick test, were not affected in early arthritis of TNFtg mice but after TNF-α blockade (Fig. 1F). Thus, these data indeed indicate were increased in latency at later disease stages (Fig. S2A). When a fast effect of TNF-α blockade on the pain responses in the CNS testing mechanical hyperalgesia using von Frey filaments, the even before a measurable anti-inflammatory effect is achieved. force (threshold) needed to elicit responses was significantly lower in early (6 wk) and later stages of the disease (10 wk) (Fig. Reduced Activity in CNS Regions Involved in Pain Perception and in 2B). The Hargraves test, assessing thermal hyperalgesia, showed the Limbic System After TNF-α Blockade. A more detailed analysis a similar pattern, as observed with mechanical hyperalgesia (i.e., of the BOLD signal in arthritis patients showed that reduction of increased sensitivity to nociception in early and late stages of the

Fig. 1. Rapid reversal of pain induced BOLD signals after TNF-α blockade in patients with RA. (A and B) Area of BOLD signal based on fMRI scans elicited by joint compression (A) and ﬁnger tapping (B) in patients (n = 5) with RA before, 1, 14, and 42 d after administration of TNF-α blocking agent IFX. (C) VAS (reaching from 0 to 10) for arthritis-related pain before, 1, 14, and 42 d after administration of IFX; (D) DAS28 before, as well as 1, 14, and 42 d after administration of IFX; (E) swollen and (F) tender joint count based on the assessment of 28 joints. (G) Maps of BOLD activity in a single patient before (Top) as well as 1 (Middle) and 42 d (Bottom) after administration of IFX. Asterisks indicate signiﬁcant difference to baseline (P < 0.05).

3732 | www.pnas.org/cgi/doi/10.1073/pnas.1011774108 Hess et al. Downloaded by guest on September 30, 2021 SEE COMMENTARY

Fig. 2. Mapping of BOLD changes after TNF-α blocking therapy. (A–C) Changes in the area size of the BOLD signal elicited by joint compression (blue bars) or finger tapping (red bars) comparing day 0 with day 1 (W2/W1) and day 0 with day 42 (W3/W1) after IFX treatment: overall (A) and specific for the contralateral (B)and ipsilateral hemisphere (C). (D and E) Changes in the area size of the BOLD signal elicited by joint compression (D)orfinger tapping (E) comparing day 0 with day 1 (W2/W1) at the following brain regions: thalamus (Th), secondary (S2) and primary (S1), somatosensoric cortex, parietal cortex (Par), posterior cingulate cortex (PCC), anterior cingulate cortex (ACC), lateral (LPFC) and medial (MPFC) prefrontal cortex, anterior (AIns) and posterior (PIns) insular cortex, cerebellum (Cb), motor cortex (M1), and periaqueductal gray (PAG).

disease of TNFtg mice) (Fig. S2C). In contrast, visceral noci- TNFtg mice to the level of WT mice within 24 h. When quantifying ception, measured by intraperitoneal injection of MgSO4, was the BOLD activity by measuring the area size and peak amplitude significantly reduced rather than enhanced in TNFtg mice, both in of the BOLD signal, we observed a consistent and significant in- early and late disease stages, suggesting a desensitization to vis- crease in the somatosensoric cortex of TNFtg mice, which was ceral nociception in TNFtg mice (Fig. S2D). The Rotarod test completely abolished after TNF-α blockade (Fig. 3C). Smaller, measuring motor activity was not impaired at an early disease but still significant, increases (Fig. 3C) were also observed in the stage but significantly decreased at later stages (Fig. S3A). Similar thalamus, as well as in the association cortex in TNFtg compared data were obtained when the suspended-tail test was applied, with WT mice, which were again completely reversed upon which showed a significant decrease in overall activity in TNFtg blockade of TNF-α (Fig. 3D). mice but only in later stages of disease (Fig. S3B). On the basis of these observations, we aimed at precisely To test whether this nociceptive sensitization is reversible, we mapping the interaction of the temporal structure of the BOLD treated TNFtg mice with the TNF-α blocker IFX. Whereas the signals in arthritis (Fig. 4A). Again, amplitudes of BOLD signals latency of spinal pain reflex arches, as measured by the tail-flick were generally higher in TNFtg mice than in WT mice. More- test, were normalized 72 h after TNF-α blockade (Fig. S2E, in- over, even low-impact subthreshold stimulation led to increased creased sensitivity to mechanical nociceptive stimuli in TNFtg brain activity in TNFtg mice, which was not found in WT mice mice was completely reversed as early as 24 h after injection (Fig. (red color in Fig. 4A at S1/S2 and peaks in Fig. 4B). This pattern S2F). This profound and rapid antinociceptive effect of TNF-α was completely reversed upon TNF-α blockade as early as 24 h blockade was further substantiated by the Hargreaves test, which after injection. Aside from higher amplitudes, brain activation normalized within 24 h after TNF-α blockade and remained at was more widespread in TNFtg mice, affecting many more brain the level of WT mice over 72 h (Fig. S2G). Within this short time regions than in WT mice. In the latter, the nociceptive response period after TNF-α blockade, no apparent change of clinic-analog was localized to several well-defined brain areas (vertical pattern parameters (such as paw swelling or grip strength) or histopath- in Fig. 4A, white circles). This spreading of brain activation in

ologic signs of arthritis (such as synovitis) could be observed. TNFtg mice was partially reversed upon blockade of TNF-α.In MEDICAL SCIENCES Moderate effects on visceral nociception were also observed, as addition, the increased BOLD activity was maintained even the reduced sensitivity of TNFtg mice to visceral nociception was during the intervals between stimulations and did not completely reversed to WT levels within 72 h after TNF-α blockade (Fig. regress to baseline levels (horizontal red lines in Fig. 4A; red line S2H). TNF-α blockade also completely restored motor activity in Fig. 4B is permanently high above baseline). TNF-α blockade within the first 24 h (Rotarod test) (Fig. S3C). Similarly, but rapidly reversed this prolonged activation, and the BOLD signal slightly less pronounced, decreased overall mobility and activity in the time periods between the stimulations dropped dramati- of TNFtg mice as measured by the suspended-tail test also im- cally, at times below the baseline level. This effect was particu- proved upon TNF-α blockade(Fig. S3D). larly pronounced in somatosensoric cortical and limbic regions of the brain (asterisks at dark blue intervals in Fig. 4A and green Mapping of Nociceptive Brain Responses Elicited by Arthritis and line Fig. 4B). Their Modulation by TNF-α Blockade. Comparative fMRI analysis of the brain-activity patterns elicited by heat stimulation of the hind Rewiring of the Pain Matrix in the CNS by Chronic Arthritis and Its paw revealed a significant increase of the BOLD activity in Reversibility by TNF-α Blockade. Higher-order changes in brain TNFtg mice in the so-called pain matrix (19, 20) of the brain function may not only be reflected in changes of response compared with the WT mice (Fig. 3 A–C). TNF-α blockaded to properties of single brain structures. Instead, dynamic changes of a fast and complete reversal of the increased brain activity in the complex network connectivity (i.e., within the pain matrix)

Hess et al. PNAS | March 1, 2011 | vol. 108 | no. 9 | 3733 Downloaded by guest on September 30, 2021 Fig. 3. Reversible enhancement of central pain responses by TNF-α–mediated arthritis. Functional MRI of the brain of 10-wk-old WT and human TNFtg mice without and with human antiTNF-α antibody IFX (each n = 10). (A and B) 2D axial scans with heatmap (A) or superimposed on the corresponding anatomical image (B) show enhanced neuronal activity in the primary (S1) and secondary (S2) somatosensoric cortex, as well as in association cortex (RS) and the thalamus (Th) of TNFtg mice, which is reversible 24 h after IFX administration. (C) 3D reconstruction of functional brain activity. (D) Bar graphs showing area size (Left) and peak height (Right) of the BOLD signal in WT mice (blue) and TNFtg mice treated with either vehicle (red) or the human anti–TNF-α antibody IFX (green). Four key groups of brain regions (brainstem, thalamus, sensory cortex, and association cortex) involved in central nociception are shown. Asterisks indicate signiﬁcant differences of WT to TNFtg plus those of TNFtg/IFX to TNFtg, with * and + indicating a P value of < 0.05 and ** and ++ a P value of < 0.025) (n = 10).

might occur. Graph theory has proven to have a high impact on rapid effect of TNF-α blockade on their subjective disease state. the quantitative analysis of complex networks (i.e., brain network However, the basis of this rapid effect has not been addressed organization) (19, 20). We therefore wondered if graph theo- convincingly so far, although CNS effects of TNF-α blockade retical analysis explaining the functional connectivity of CNS have always been suspected to play a role in this process. We regions provides insight into the dynamic of rewiring of the pain were stunned by the fact that TNF-α blockade triggered a pro- matrix in arthritis and after neutralization of TNF-α. After re- found and significant change of CNS activity in RA patients as moving the global (mainly task-driven) signals, cross-correlation early as 24 h after its administration, compared with baseline of the fMRI time profiles of all structures, which were signifi- CNS activity. In contrast, joint swelling, as well as composite cantly activated, revealed distinct patterns of interactions within DAS (DAS28) did not change significantly within the first 24 h, the pain matrix (Fig. S4A, lower triangular half). There was an although they improved later. These observations suggest that increased degree of connectivity, clustering, and modularity in functional CNS changes clearly precede the improvement of arthritic TNFtg mice compared with WT mice (Fig. S4A and inflammatory signs of arthritis in human patients. Nociceptive Table S2). These changes largely resulted from the formation of signals elicited by joint compression did induce the activity in a tight cluster of the thalamus, the periaqueductal gray, and the CNS structures, which are typically involved in pain perception, amygdala (Kamada-Kawai plots, Fig. S4B). TNF-α blockade led such as the thalamus and the somatosensoric cortex, but also in to a rapid partial dissolution of this cluster, resulting in a more parts of the limbic system, such as the cingulum and the insular diffuse network, indexed by an overall decreased connectivity, region, which are relevant for the inner body sensation as well as clustering, and modularity similar to that observed in WT mice. negative emotional experience, including pain experience. Sup- pression of neuronal activity in these centers may likely explain Discussion the rapid improvement of the symptom state of patients with RA In this study we show that neutralization of the proinflammatory after TNF-α blockade, because the perception of nociceptive cytokine TNF-α rapidly affects CNS pain responses elicited by stimuli in the CNS is impaired and activation of CNS centers arthritis. Administration of TNF-α blockers rapidly abrogates involved in shaping the emotional state of the individual—as well CNS activity linked to nociceptive stimuli in patients with ar- as body perception—is blocked. thritis, as well as in mice with TNF-α–driven arthritis. This effect In murine arthritis triggered by TNF-α overexpression, we occurs immediately after antibody administration and well be- observed that behavioral measurements of nociceptive responses fore anti-inflammatory effects of TNF-α blockade are observed. showed a sensitization to somatic nociception (hyperalgesia) Our data suggest that TNF-α appears to trigger hypernociception during arthritis. Moreover, investigation of arthritic mice by because of changed pain processing in the CNS and a strong fMRI unequivocally demonstrated an altered pattern of brain linkage to activation of limbic areas involved in pain experience, activation, as measured by an enhanced BOLD activity. Even emotions, and body sensation. low-impact subthreshold stimulation led to increased BOLD Cytokine blockade has dramatically changed the treatment of signals, which could not be depicted in nonarthritic controls. In inflammatory diseases, particularly of inflammatory arthritides, support of this notion, TNF-α blockade caused a significant and such as RA. Despite RA being the embodiment of a chronic rapid reduction of these nociceptive BOLD signals in the disease, patients with RA experience a very rapid improvement somatosensoric and the association cortex, as well as the thala- of disease-related symptoms upon blockade of TNF-α. Although, mus, within 24 h. Upon neutralization of TNF-α, both thermal it takes several weeks to observe a consistent and objective re- and mechanical hyperalgesia induced by arthritis were com- duction of joint inflammation (16), many patients report on a pletely reversed within 24 h, an observation that supports the

3734 | www.pnas.org/cgi/doi/10.1073/pnas.1011774108 Hess et al. Downloaded by guest on September 30, 2021 Fig. 4. Temporal proﬁles and reversibility of TNF-

α–induced changes of central pain response. (A) SEE COMMENTARY Temporal proﬁles showing the intensity of BOLD signals obtained by fMRI of 10-wk-old WT mice and human TNFtg mice treated either with vehicle or with human anti–TNF-α antibody IFX. Colors indicate peak height of the BOLD signal from very low (dark blue) to very high (red). The x axis indicates time with at total of three sequences, each of which comprises four pain stimuli (S1 to S4) with increasing intensity (40, 45, 50, 55 °C). The y axis reﬂects 32 different brain regions as follows: motor cortex (M1), cerebellum (Cb), ventral pallidum (VP), globus pallidus (GP), nucleus accumbens (Acb), striatum (CPu), periaqueductal gray (PAG), zona incerta (ZI), hypothalamus (HT), bed nucleus of stria terminalis (BST), amygdala (Amd), hippocampus (Hip), septal area (Sep), piriform cortex (Pir), perirhinal/ectorhinal cortex (Prh/Ect), ento- rhinal cortex (Ent), insular cortex (Ins), frontal association cortex (FrA), cingulate cortex (Cg), retrosplenial cortex (RS), secondary (S2) and primary somatosensory cortex (S1), ventral postero- lateral/posteromedial thalamic nucleus (VPL/VPM), medial thalamus (MT), lateral posterior thalamic nucleus (LP), lateral (LG) and medial geniculate nucleus (MG), pretectal area (PTA), superior (SC) and inferior colliculus (IC), substantia nigra (SN), ventral tegmental area (VTA). (B) Curves showing average time-dependent changes of the amplitudes of BOLD signals in WT (blue), hTNFtg mice (red), and the latter treated with the human anti–TNF-α antibody IFX (green).

results obtained in RA patients. Within this short period no explanation for the rapid improvement of the symptom state in- apparent effect on clinical analog parameters (such as paw duced by TNF-α blocking therapy. These data also suggest that swelling) or histopathologic signs of arthritis (such as synovitis) anticytokine therapy targeting TNF-α may achieve its (fast or rapid) could be observed. This result indicates that the CNS effect of clinical benefit remote from the inflamed joint through influencing TNF-α blockade precedes its anti-inflammatory effect. CNS processes. Our findings also extend previous observations on Interestingly, we even observed a decrease in the BOLD signal the effects of TNF-α in the CNS (i.e., the regulation of expression below the normal levels observed in nonarthritic WT mice. This of clock genes), which may explain the high prevalence of fatigue in effect was again particularly pronounced in somatosensoric cor- inflammatory diseases (22). Finally, our findings also raise the tical and limbic regions of the brain. One explanation for this possibility that the assessment of responsiveness of BOLD activity decrease of the BOLD signal below the baseline level could be an in the CNS of RA patients and patients with other forms of in- increased activity of the cerebral inhibitory network developed flammatory arthritis may be used as a surrogate marker for pre- during this chronic pain state (21). In this context, it is worth dicting clinical responses to therapeutic intervention in a much mentioning that brain network analysis detected a tight cluster faster way than it is currently realized. consisting of thalamic structures, the periaquaeductal gray, and limbic areas, such as the amygdala in TNFtg mice. This finding Materials and Methods supports the concept of an increased efficacy of the cerebral in- Mice. The human TNF-α and TNFtg mice (strain Tg197) were previously de- hibitory network in arthritis. This rewiring may have developed scribed (17). Treatment with IFX (Centocor), a neutralizing chimeric human- murine monoclonal antibody directed against human TNF-α, was done by a as an adaptation to chronic pain preventing the abnormally in- fl single intravenous injection of a dose of 10 mg/kg antibody 24 h before testing creased ow of nociceptive signals to higher CNS structures, es- (15). Clinical evaluation was performed weekly, starting at 4 wk after birth. pecially to the cortex. Importantly, TNF-α blockade rapidly Arthritis was evaluated in a blinded manner, as described previously (18). All induced partial dissolution of this cluster; this happened within animal experiments were approved by the local ethics committee of the 24 h after TNF-α blockade, in parallel with the reduction of the University of Erlangen-Nuremberg. nociceptive BOLD activity and the restoration of the normal MEDICAL SCIENCES nociceptive behavior. Tests for Nociceptive Behavior. The tail-flick test was conducted using a tail- In summary, these data show that TNF-α leads to more in- flick analgesia meter (Columbus Instruments), Hargreaves test by IITC Plantar tensive, widespread, and prolonged brain activity upon nocicep- Analgesia Meter, and measurement of paw-withdrawal latency upon heat α stimulation. The mechanical nociception test was performed by applying an tive stimulation, which is reversible upon neutralization of TNF- . ascending series of von Frey hairs (23). In visceral nociception tests, mice were This reversal of pain sensitization occurs rapidly and is not pri- intraperitoneally injected with MgSO4 (120 mg/kg) (23). Depression was marily linked to the anti-inflammatory effects of TNF-α blockade. tested by the tail-suspension test and locomotor activitiy was tested by Ugo In patients with RA, a very similar rapid and profound down- Basile 7750 accelerating Rotarod (Ugo Basile) (24). regulation of nociceptive brain activity can be observed, which did precede the alleviation of joint inflammation. Although sub- Functional MRI in Mice. WT and TNFtg mice (n = 10 per group, 10 wk) were clinical anti-inflammatory effects of TNF-α blockade within the anesthetized with isoflurane and placed on a cradle inside the magnetic first 24 h after exposure to TNF-α blockade cannot be ruled out resonance tomograph (Bruker BioSpec 47/40, quadrature head coil) (25). The completely, the lack of a drop in acute phase reactant and cyto- contact heat stimuli sequences (40, 45, 50, and 55 °C, plateau for 5 s, ramp 15 s) were presented at the right hind paw with 3-min and 25-s intervals, three kine levels within this short time interval does at least not support fi times, using a custom-made computer-controlled Peltier heating device. A such a concept. Our ndings emphasize that changes of nocicep- series of 750 sets of functional images (matrix 64 × 64, field of view 15 × 15 tive responses in the brain precede the anti-inflammatory effect mm, slice thickness 0.5 mm, axial, 22 slices) were sampled using gradient of TNF-α blockers. Importantly, these findings may be a good echo-based Echo Planar Imaging Technique (single shot: TR = 4,000 ms,

Hess et al. PNAS | March 1, 2011 | vol. 108 | no. 9 | 3735 Downloaded by guest on September 30, 2021 TEef = 24.38 ms, NEX = 2) within 50 min. Finally, 22 corresponding ana- the human brain (30). The voxels, which were significantly activated by the tomical T2 reference images (RARE, slice thickness 0.5 mm, field of view 15 × above criterion, were counted as the activated volume per given brain region. 15 mm, matrix 256 × 128, TR = 2,000 ms, TEef =56 ms) were taken as pre- The mean corresponding peak activity was determined for each stimulation viously described in detail (26). temperature for mice and for tapping and compression in humans, averaged over all activated brain structures and finally over all subjects of one experi- Patients. Five female patients with RA, failing on standard treatment with mental group, respectively, to provide a global group comparison. disease-modifying antirheumatic drugs and having active disease with joint tenderness and swelling, received the TNF-α blocker IFX at a dose of 3 mg/kg Graph Theoretical Analysis. Functional connectivity patterns were computed as an intravenous infusion. Mean (± SD) age was 56.3 ± 8.2 y and mean (± as cross-correlations of the residuals after the global signal mean was re- SD) disease duration was 8.5 ± 3.3 y. The number of tender and swollen moved by linear regression and represented as correlation matrices. Similarity joints, joint pain (VAS ranging from 0 to 10), overall disease activity calculated by the DAS based on 28 joints (DAS28) (27), C-reactive protein, and IL-6 across these correlation matrices was calculated as Spearman rank correla- fi levels were assessed at baseline and 1, 14, and 42 d after the infusion. tion. Modularity and clustering coef cients, as well as several centrality measures, were computed from the weighted pattern of positive correlations Functional MRI in Humans. All anatomical and fMRI data were acquired on a 3T present between brain regions (31). The network modularity index was scanner (Magnetom Trio; Siemens) using a standard eight-channel phased- obtained by optimally subdividing the network into modules, such that most array head coil. The ethics committee of the University Clinic of Erlangen connections are made within modules and only few connections exist be- approved all procedures and written informed consent was obtained from all tween modules (32, 33). Clustering was computed from the weighted func- patients. For anatomic datasets, we used a T1-weighted MPRAGE-sequence tional-connectivity matrix (34) and scaled relative to a population of 100 fi × × × ( eld-of-view = 256 mm, matrix size = 256 256, voxel size = 1.0 1.0 1.0 random networks with identical degree distribution. Node-strength cen- 3 mm , slices = 176, slice thickness = 1 mm, TR = 1,900 ms, TE = 1.13 ms). For trality was measured as the sum of each node’s positive cross-correlations. each subject, two experiments with different stimulation conditions (first, The networks are visualized using a force-based algorithm after Kamada- finger-tapping and second, compression of the metacarpophalangeal joints) Kawai for achieving that all edges are of more or less equal length, and there were performed. In each of them, 93 whole-brain images were obtained exist as few crossing edges as possible (35). with a gradient-echo, echo-planar scanning sequence (TR = 3,000 ms, TE = 30 ms, flip angle 90°; field-of-view, 220 mm2, acquisition matrix 64 × 64, 36 axial ± slices, slice thickness 3 mm, gap 0.75 mm). Statistical Analysis. Data are presented as mean SEM. Group mean values were compared by the Student’s t tests in a task-specific single-test frame- fi Functional MRI Analysis. Functional analysis was performed for mice and work to assess signi cant differences between the different mouse strains. humans using Brain Voyager QX (Version 10.3) and our own software MagnAn (25), as previously described (28). In summary, after preprocessing [motion- ACKNOWLEDGMENTS. This study was supported by the Deutsche For- corrected using sinc interpolation Gaussian spatial (human: FWHM = 4 mm, schungsgemeinschaft FG 661/TP4 and SPP1468-Immunobone (to A.H. and mouse: 0.469 mm) and temporal (FWHM = 3 volumes) smoothing], general G.S.); the Bundesministerium für Bildung und Forschung projects 01EM0514, 01GQ0731, 0314102, Ankyloss and Immunopain (to G.S. and A.H.); the Mas- linear modeling analysis with separate predictors for each stimulus was per- terswitch, Kinacept, and Adipoa projects of the European Union (to G.S.); formed. The statistical parametric mapping obtained were corrected for the Interdisciplinary Centre for Clinical Research and the Erlanger Leistungs- multiple comparisons, false-discovery rate threshold at z-score level of 3.3, and bezogene Anschubfinanzierung und Nachwuchsförderung (ELAN) fund of different groups of activated voxels were labeled as belonging to certain brain the University of Erlangen-Nuremberg (to G.S. and A.H.); K.B. is Doerenkamp structures based on (i) the mouse atlas from Paxinos (29) or (ii) the Mai atlas of Professor for Innovations in Animal and Consumer Protection.

1. Smolen JS, Steiner G (2003) Therapeutic strategies for rheumatoid arthritis. Nat Rev 18. Diarra D, et al. (2007) Dickkopf-1 is a master regulator of joint remodeling. Nat Med Drug Discov 2:473–488. 13:156–163. 2. Firestein GS (2003) Evolving concepts of rheumatoid arthritis. Nature 423:356–361. 19. Apkarian AV, Bushnell MC, Treede RD, Zubieta JK (2005) Human brain mechanisms of 3. McInnes IB, Schett G (2007) Cytokines in the pathogenesis of rheumatoid arthritis. Nat pain perception and regulation in health and disease. Eur J Pain 9:463–484. Rev Immunol 7:429–442. 20. Borsook D, Becerra L (2007) Phenotyping central nervous system circuitry in chronic 4. Brennan FM, Jackson A, Chantry D, Maini R, Feldmann M (1989) Inhibitory effect of pain using functional MRI: considerations and potential implications in the clinic. Curr TNFa antibodies on synovial cell interleukin-1 production in rheumatoid arthritis. Pain Headache Rep 11:201–207. Lancet 334:244–247. 21. Shmuel A, Augath M, Oeltermann A, Logothetis NK (2006) Negative functional MRI 5. Dinarello CA (2010) IL-1: Discoveries, controversies and future directions. Eur J response correlates with decreases in neuronal activity in monkey visual area V1. Nat Immunol 40:599–606. Neurosci 9:569–577. 6. Singh JA, et al. (2009) Biologics for rheumatoid arthritis: An overview of Cochrane 22. Cavadini G, et al. (2007) TNF-alpha suppresses the expression of clock genes by reviews. Cochrane Database Syst Rev 4:CD007848. interfering with E-box-mediated transcription. Proc Natl Acad USA 104:12843–12848. 7. van den Berg WB (2001) Uncoupling of inflammatory and destructive mechanisms in 23. Whishaw IQ (1999) The Behavior of the Laboratory Rat: A Handbook with Tests, eds arthritis. Semin Arthritis Rheum 30(5, Suppl 2):7–16. Whishaw IQ, Kolb B (Oxford University Press, Oxford), pp 478–486. 8. O’Connor JC, et al. (2009) Interferon-gamma and tumor necrosis factor-alpha mediate 24. Jones BJ, Roberts DJ (1968) The quantiative measurement of motor inco-ordination in the upregulation of indoleamine 2,3-dioxygenase and the induction of depressive-like naive mice using an acelerating rotarod. J Pharm Pharmacol 20:302–304. behavior in mice in response to bacillus Calmette-Guerin. J Neurosci 29:4200–4209. 25. Hess A, Sergejeva M, Budinsky L, Zeilhofer HU, Brune K (2007) Imaging of hyperalgesia 9. Reinold H, et al. (2005) Spinal inflammatory hyperalgesia is mediated by pros- in rats by functional MRI. Eur J Pain 11:109–119. taglandin E receptors of the EP2 subtype. J Clin Invest 115:673–679. 26. Hennig J, Nauerth A, Friedburg H (1986) RARE imaging: A fast imaging method for 10. Boettger MK, et al. (2008) Antinociceptive effects of tumor necrosis factor alpha clinical MR. Magn Reson Med 3:823–833. neutralization in a rat model of antigen-induced arthritis: evidence of a neuronal 27. Prevoo ML, et al. (1995) Modified disease activity scores that include twenty-eight- target. Arthritis Rheum 58:2368–2378. joint counts. Development and validation in a prospective longitudinal study of 11. Marchand F, et al. (2009) Effects of Etanercept and Minocycline in a rat model of patients with rheumatoid arthritis. Arthritis Rheum 38:44–48. spinal cord injury. Eur J Pain 13:673–681. 28. Knabl J, et al. (2008) Reversal of pathological pain through specific spinal GABAA 12. Schäfers M, et al. (2008) Selective stimulation of either tumor necrosis factor receptor receptor subtypes. Nature 451:330–334. differentially induces pain behavior in vivo and ectopic activity in sensory neurons in 29. Paxinos G, Franklin KBJ (2001) The Mouse Brain in Stereotactic Coordinates (Elsevier vitro. Neuroscience 157:414–423. Academic Press, Burlington, MA). 13. Ogawa S, Lee TM, Kay AR, Tank DW (1990) Brain magnetic resonance imaging with 30. Mai JK, Paxinos G, Voss T (2008) Atlas of the Human Brain (Elsevier Academic Press, contrast dependent on blood oxygenation. Proc Natl Acad USA 87:9868–9872. Burlington, MA). 14. Thulborn KR, Waterton JC, Matthews PM, Radda GK (1982) Oxygenation dependence 31. Rubinov M, Sporns O (2010) Complex network measures of brain connectivity: Uses of the transverse relaxation time of water protons in whole blood at high field. and interpretations. Neuroimage 52:1059–1069. Biochim Biophys Acta 714:265–270. 32. Girvan M, Newman ME (2002) Community structure in social and biological networks. 15. Knight DM, et al. (1993) Construction and initial characterization of a mouse-human Proc Natl Acad Sci USA 99:7821–7826. chimeric anti-TNF antibody. Mol Immunol 30:1443–1453. 33. Newman ME (2004) Analysis of weighted networks. Phys Rev E Stat Nonlin Soft 16. Maini R, et al.; ATTRACT Study Group (1999) Infliximab (chimeric anti-tumour necrosis Matter Phys 70:056131. factor alpha monoclonal antibody) versus placebo in rheumatoid arthritis patients 34. Onnela JP, Saramäki J, Kertész J, Kaski K (2005) Intensity and coherence of motifs in receiving concomitant methotrexate: A randomised phase III trial. Lancet 354:1932–1939. weighted complex networks. Phys Rev E Stat Nonlin Soft Matter Phys 71:065103. 17. Keffer J, et al. (1991) Transgenic mice expressing human tumour necrosis factor: A 35. Kamada T, Kawai S (1989) An algorithm for drawing general undirected graphs. Inf predictive genetic model of arthritis. EMBO J 10:4025–4031. Process Lett 31:7–15.

3736 | www.pnas.org/cgi/doi/10.1073/pnas.1011774108 Hess et al. Downloaded by guest on September 30, 2021 COMMENTARY

Mapping the immunological homunculus

Betty Diamond and Kevin J. Tracey1 The Feinstein Institute for Medical Research, Manhasset, NY 11030

uring the first century, the Ro- riod after administration of anti-TNF, be- tems, including cardiovascular, gastroin- man physician Cornelius Celsus fore any significant improvement in joint testinal, and respiratory, which enabled the D defined four cardinal signs of swelling or grip strength, and before his- mapping of reflex units. For example, by inflammation: redness, swell- topathological improvement of synovitis. recording signals to the heart, it was possi- ing, heat, and pain. These signs and symp- As in the human subjects, there was sig- ble to define a reflex that controls heart toms occur during infection by invasive nificant abolishment of BOLD signals in the rate: increased heart rate activates afferent pathogens or as a consequence of trauma. somatosensory cortex. Brain activation was arcs that in turn elicit firing of efferent arcs Today, we understand the molecular basis more extensive in the transgenic mice and in the vagus nerve, which slows heart rate. of these physiological responses as medi- affected more brain regions compared with More recently, these principles have ated by cytokines and other factors pro- WT mice, and this spreading of brain acti- been applied to mapping the somewhat duced by cells of the innate immune sys- vation was down-modulated by administra- more elusive immune system. The com- tem. Cytokines are both necessary and suf- tion of the TNF inhibitors. Administration bined use of neurophysiological stimulating ficient to cause pathophysiological alter- of anti-TNF reversed the prolonged acti- techniques and current molecular strate- ations manifested as the four cardinal signs. vation of BOLD activity that was gies to measure immune response estab- Importantly, this knowledge has enabled lished the “inflammatory reflex” (3). the development of highly selective ther- The nervous system is Cytokines activate afferent signals in the apeutical agents that target individual cyto- vagus nerve, which culminate on an efferent kines to prevent or reverse inflammation. hardwired to monitor arc that inhibits further cytokine produc- For example, selective inhibitors of TNF, tion. The molecular basis of this cytokine- a major inflammatory cytokine, have revo- the presence of cytokines inhibiting action is attributable to acetyl- lutionized the therapy of rheumatoid ar- choline, the principal neurotransmitter of thritis, inflammatory bowel disease, and and molecular products the vagus nerve, interacting with α7 nico- other autoimmune and autoinflammatory tinic acetylcholine receptor subunits ex- diseases affecting millions worldwide. Now, of invaders. pressed by cytokine-producing cells. Signal in PNAS, Hess et al. (1) use functional MRI transduction through this receptor–ligand to monitor brain activity and report that interaction down-regulates the activity of patients with rheumatoid arthritis who re- observed during the intervals between NF-κB, a primary pathway for regulating ceive anti-TNF develop significant changes stimulations, and was particularly pro- cytokine transcription and synthesis. Ac- in brain activity before resolution of inflam- nounced in the somatosensory cortex and cordingly, α7 KO mice rendered deficient in limbic system. this pathway are exquisitely sensitive to in- mation in the affected joints. fi To accomplish this, the authors measured These intriguing ndings provide a flammation caused by pathogen-associated blood oxygen level-dependent (BOLD) glimpse into the future, when it is likely molecules and experimental arthritis. The fl fl signals in the brain after compressing the that sensory and motor brain regions as- efferent arm of the in ammatory re ex, fl metacarpal phalangeal joints of the arthritic sociated with immune responses will be termed the cholinergic antiin ammatory hand. They observe enhanced activity in the mapped as an immunological homunculus pathway, has enabled the development of (2). Of particular significance here is that experimental nerve stimulators and highly brain regions associated with pain percep- fi tion, including the thalamus, somatosensory this approach bridges two major elds of selective pharmacological agents that sup- cortex, and limbic system, regions known to medicine, neurology and immunology. press cytokine-mediated disease in stan- process body sensations and emotions as- dardized preclinical models. sociated with the pain experience (1). Brain Reflexes Regulate Immune Responses activity was significantly reduced within 24 h Recent advances at the intersection of Sensing Inflammation and Infection after treatment with TNF inhibitors, a time these fields have provided direct evidence Less is known about the mapping of the frame that preceded any observable evi- that neural circuits reflexively control the afferent or sensory arc of the inflammatory dence of reduced signs of inflammation in onset and resolution of inflammation, and reflex, and there is significant interest in affected joints. Clinical composite scores, previously undescribed molecular mecha- the question of how the nervous system comprising measurements of C-reactive pro- nisms have been defined (3). The funda- monitors the state of inflammation. The tein, a circulating marker of inflammation mental principle of neurological reflexes, as seminal work of Watkins et al. (4) dem- severity, were not improved until after 24 h. originally established by Sir Charles Scott onstrated that the fever-causing activities This suggests that selective inhibition of Sherrington in the early 20th century, is that of the pyrogenic cytokine IL-1 required an TNF has a primary early effect on the ner- neural circuits cooperate to maintain the intact vagus nerve, because administration vous system pain centers. organism’s homeostasis. He elucidated the of IL-1 into the abdomen failed to pro- The authors further explore this result in role of the nervous system in integrating duce fever if the vagus nerve had been cut. genetically engineered transgenic mice that reflexes to provide stable physiological These researchers proposed a critical role overexpress TNF and develop spontane- systems. He famously remarked that “Ex- of the glomus cell, specialized to detect ous arthritis. Administration of anti-TNF istence of an excited state is not a pre- significantly improved the development of requisite for the production of inhibition; mechanical hyperalgesia in the standard- inhibition can exist apart from excitation no Author contributions: B.D. and K.J.T. wrote the paper. ized von Frey filament test system and also less than, when called forth against an ex- The authors declare no conflict of interest. improved the response to thermal hyper- citation already in progress, it can suppress See companion article on page 3731. ” fi algesia in the Hargreaves test. This clinical or moderate it. These principles were rst 1To whom correspondence should be addressed. E-mail: improvement occurred within a short pe- defined in relatively accessible organ sys- [email protected]. www.pnas.org/cgi/doi/10.1073/pnas.1100329108 PNAS | March 1, 2011 | vol. 108 | no. 9 | 3461–3462 pH and other chemical changes in the in- (10). Moreover, intrathecal administration under conditions of prolonged alterations ternal milieu, in sensing IL-1 (5). Glomus of antibody was significantly more effec- of electrical activity (13). A testable hy- cells express IL-1 receptors, and it is a tive compared with systemic anti-TNF, pothesis emerges from this reasoning: The plausible hypothesis that IL-1 binding to suggesting that the principal effects of systemic inflammation associated with glomus cells stimulates the release of do- anti-TNF are mediated through direct in- rheumatoid arthritis opens the barrier to pamine, which, in turn, triggers afferent teraction with neurons. Considered in light anti-TNF, which inhibits glial cell-derived signaling in the vagus nerve. Ascending of the current paper, it is plausible that TNF and alters the strength of synapses in action potentials culminate on the nucleus local accumulation of TNF in the region of the brain’s pain regions. This might explain tractus solitarius and are then relayed to sensory pain circuit neurons drives activa- the observed reversal of prolonged activa- higher brain regions. Sensory neural cir- tion of the brain’s pain centers. Thus, it is tion of BOLD activity during intervals be- cuits can also be activated by LPS, and possible that the rapid early improvement tween stimulations in the arthritic mice. Toll-like receptor 4 (TLR4), the principal from anti-TNF results from acutely neu- endotoxin receptor, is expressed by neutralizing TNF and blocking its ability to Immunological Homunculus rons in the nodose ganglia (6). These stimulate pain fibers. Predictably, this neurons have been implicated in trans- circuit would not require any significant Advances in brain imaging modalities and mitting afferent signals in the vagus nerve. improvement in the severity of inflam- molecular biology are mapping brain net- Another mechanism by which the nervous mation in the arthritic joints in order for works and circuits with high resolution and system can sense the presence of injury is benefit to be perceived. precision. These maps reveal sensory and dependent on the expression of TLR4 in The current study also raises another motor circuits that are somatotopically microglial cells, which is required for the possibility, that anti-TNF antibodies tra- and functionally organized, and form the development of neuropathic pain in a mu- versed the blood–brain barrier and directly basis for understanding the neural circuits rine model of nerve transection (7). This altered the activity of brain neurons in the that coordinate physiological responses TLR4 pathway, which may be activated pain centers. Until quite recently, it has to the external and internal environment. even in the absence of infection by been dogmatic that antibodies do not en- Early advances were achieved by mapping HMGB1 released from injured cells (8), ter the brain, but this has been overturned the relatively accessible domains of body stimulates TNF release in the develop- by studies in a murine model of the auto- sensations and skeletal muscle motor ment of chronic pain states. Together, immune disease lupus (11). This work function. Now, knowledge of the in- these results indicate that the nervous identified antibodies that bind to dsDNA flammatory reflex as a prototypical circuit system is hardwired to monitor the pres- (a common antibody phenotype in this that regulates inflammation, combined ence of cytokines and molecular products syndrome) and to the NMDA receptor with the ability to visualize brain activity in of invaders. expressed on brain neurons. These anti- patients and mice with inflammation, The presence of TNF can also be sensed bodies readily access the brain, because the makes it possible to make additional by sensory neurons, which express type I blood–brain barrier can be opened during advances. It is interesting to consider and type II TNF receptors (9). Local or relatively mild stress caused by adminis- that in addition to defending the host regional administration of TNF elicits tration of endotoxin, cytokines, or even from infection, the immune system func- pain, partially by inciting tissue damage epinephrine. On entering the brain, the tions as a sensory organ that transmits and activating release of other molecules interaction of these antibodies with neu- information in real time to the central capable of activating sensory neurons. rons in the hippocampus and other regions nervous system about the tissue response TNF receptor–ligand interaction on sen- mediates cell damage and cognitive or be- to injury and infection. This presents im- sory neurons may activate pain fibers di- havioral dysfunction (12). portant possibilities for understanding rectly, however, serving as a pathway that In light of the present study, it is inter- fundamental mechanisms that maintain may activate the brain’s pain centers as esting to consider whether inflammatory physiological homeostasis. Progress in described here. Others have shown that mediators produced during arthritis might this field is all but certain to reveal the administration of anti-TNF antibodies into open the blood–brain barrier to enable identity of other circuits that reflexively the cerebrospinal fluid of rats subjected to anti-TNF direct access to brain neurons. regulate the immune system and to pro- experimental arthritis significantly attenu- TNF expressed by glial cells in brain regu- duce maps that will guide understanding ated pain-related behavior and the de- lates synaptic scaling, a mechanism that of the neurological basis of immunity velopment of inflammation in the joints adjusts the strength of neuronal synapses and physiology.

1. Hess A, et al. (2011) Blockade of TNF-α rapidly inhibits 6. Hosoi T, Okuma Y, Matsuda T, Nomura Y (2005) Novel 10. Boettger MK, et al. (2010) Spinal tumor necrosis factor pain responses in the central nervous system. Proc Natl pathway for LPS-induced afferent vagus nerve activa- alpha neutralization reduces peripheral inflammation Acad Sci USA 108:3731–3736. tion: Possible role of nodose ganglion. Auton Neurosci and hyperalgesia and suppresses autonomic responses 2. Tracey KJ (2007) Physiology and immunology of the 120:104–107. in experimental arthritis: A role for spinal tumor ne- cholinergic antiinflammatory pathway. J Clin Invest 7. Tanga FY, Nutile-McMenemy N, DeLeo JA (2005) The crosis factor alpha during induction and maintenance 117:289–296. CNS role of Toll-like receptor 4 in innate neuroimmun- of peripheral inflammation. Arthritis Rheum 62:1308– 3. Tracey KJ (2009) Reflex control of immunity. Nat Rev ity and painful neuropathy. Proc Natl Acad Sci USA 1318. Immunol 9:418–428. 102:5856–5861. 11. Kowal C, et al. (2006) Human lupus autoantibodies 4. Watkins LR, et al. (1995) Blockade of interleukin-1 in- 8. Yang H, et al. (2010) A critical cysteine is required for against NMDA receptors mediate cognitive impair- duced hyperthermia by subdiaphragmatic vagotomy: HMGB1 binding to Toll-like receptor 4 and activation ment. Proc Natl Acad Sci USA 103:19854–19859. Evidence for vagal mediation of immune-brain com- of macrophage cytokine release. Proc Natl Acad Sci 12. Faust TW, et al. (2010) Neurotoxic lupus autoanti- munication. Neurosci Lett 183:27–31. USA 107:11942–11947. bodies alter brain function through two distinct 5. Goehler LE, et al. (1997) Vagal paraganglia bind bio- 9. Boettger MK, et al. (2008) Antinociceptive effects of mechanisms. Proc Natl Acad Sci USA 107:18569– tinylated interleukin-1 receptor antagonist: A possible tumor necrosis factor alpha neutralization in a rat 18574. mechanism for immune-to-brain communication. Brain model of antigen-induced arthritis: Evidence of a neu- 13. Stellwagen D, Malenka RC (2006) Synaptic scaling me- Res Bull 43:357–364. ronal target. Arthritis Rheum 58:2368–2378. diated by glial TNF-alpha. Nature 440:1054–1059.

3462 | www.pnas.org/cgi/doi/10.1073/pnas.1100329108 Diamond and Tracey