Approaching the Next Revolution? Evolutionary Integration...

Approaching the Next Revolution? EvolutionaryIntegration of Neural and Immune PathogenSensing and Response

Kevin J. Tracey

Feinstein Institute for Medical Research, Manhasset, New York 11030

Correspondence: [email protected]

Mammalian immunity evolved by the process of natural selection that produced differentialsurvival and reproduction advantages through combinations of hereditary traits underlyingthe response to pathogens. Primitive animals sense the presence of microbial pathogensthrough recognition of pathogen-derived molecules in their rudimentary immune andnervous systems. No molecular biological mechanism assigns primacy of pathogen sensingmechanisms to immune cells over neurons. Rather, in animals as diverse as Caenorhabditiselegans to mammals, neural reflexes are activated by the presence of pathogens and trans-duce neural mechanisms that control the development of immunity. A coming revolution inimmunological thinking will require immunologists to incorporate neural circuits into under-standing pathogen signal transduction, and the molecular mechanisms of learning, thatculminate in immunity.

On considering memory, one finds an ironicperspective, tainted by shadows of two

major scientific fields that, historically at least,did not collaborate. Immunological memory,mediated by lymphocytes, and neurologicalmemory, mediated by neurons, evolved overmillions of years in response to environmentalchanges. Closer inspection within both fieldsreveals key features of common origin betweenneural and immune information collection andretrieval. Major evolutionary advantages aroseat points of intersection of these systems, man-ifested as beneficial physiological responses toenvironmental stimuli during infection, injury,and metabolic stress. In 1989, Charles Janewayproposed that the first revolution in immuno-

logical thinking led to the domination of thehumoral theory of immunity, and that an ap-proaching second revolution would integrateinnate and adaptive immunity by understand-ing the role of pathogen-associated molecularpattern receptors (Janeway 1989). Today, I be-lieve that the groundwork has been laid for athird revolution in immunological thinkingthat will integrate the role of neurological feed-back circuits into innate and adaptive immuni-ty, including the role of molecular mechanismsthrough which neurons sense microbes and reg-ulate the output of hematopoietic-derived im-mune cells. I also suggest that costimulation ofneural reflex circuits by microbial productsplays a major role in the immune response to

Editors: Ruslan M. Medzhitov

Additional Perspectives on Innate Immunity and Inflammation available at www.cshperspectives.org

Copyright # 2014 Cold Spring Harbor Laboratory Press; all rights reserved

Advanced Online Article. Cite this article as Cold Spring Harb Perspect Biol doi: 10.1101/cshperspect.a016360

1

on July 11, 2021 - Published by Cold Spring Harbor Laboratory Press http://cshperspectives.cshlp.org/Downloaded from

http://cshperspectives.cshlp.org/

infection, and to the subsequent developmentof immunity (Fig. 1). If correct, this idea couldrevolutionize our thinking about immunity andtake us beyond innate and adaptive immunityto an integrated view of neurological and im-munological recognition and learning.

EVOLUTIONARILY ANCIENT SENSINGMECHANISMS IN NEUROSCIENCE ANDIMMUNOLOGY

It begins with sensing a change in the environ-ment. Consider the evolutionarily primitiveanimal, Caenorhabditis elegans, which harborsrudimentary immune and nervous systems.Moving and feeding within a bacterial lawn, ituses neural and immune sensory mechanismsto detect the presence of pathogenic microbesand toxins (Pradel et al. 2007). On encounteringspecific molecules, sensory neurons transducethe chemical environmental information intoaction potentials, which initiate neural respons-es that modify the animal’s behavior and mo-tion. The presence of pathogens also activates its

innate immune cells, which incite defensive re-actions, including enhanced MAP kinase andXBP-1 signaling, accumulation of unfoldedproteins in endoplasmic reticulum, and othercellular and metabolic responses (Styer et al.2008; Aballay 2013). Sensory neural signals ini-tiate primitive response circuits linked to corre-sponding motor neural signals that determinebehavioral and physiological responses. Thus,the worm’s neural and innate immune sensorysystems independently detect the presence ofthe pathogens in the environment.

Primitive innate immune-like cells in C. el-egans, which predate the evolution of compe-tent Toll-like receptor and NF-kB signal trans-duction mechanisms, are also activated by thepresence of pathogens or injury (Aballay 2013).Gene expression patterns in the innate immunesystem, and animal behavior mediated by thenervous system, are altered in a pathogen spe-cies-dependent manner. The result of these re-sponses is that the animal can defend itselfagainst certain pathogen encounters, althoughnot all. The innate neural and immunological

Inducers

Sensors

Mediators

Target

Behavioralresult

Immunological

Humoralcellular

Immune systemNervous system

Immune systemNervous system

Physiologicalresponse and

memory

Neuronal(action

potentials)

Neurological

EndogenousExogenous

Figure 1. Pathogens or products of cellular inflammation and injury costimulate immune cells and neuronsduring the earliest stages of infection. Neural input activates reflex circuits, which modulate the nervous systemand the immune system. Hematopoietic-derived cellular responses produce humoral and cellular signals thatinfluence immune cells and neurons. These signal transduction pathways culminate in mediating the behavior ofthe animal and in initiating learning.

K.J. Tracey

2 Advanced Online Article. Cite this article as Cold Spring Harb Perspect Biol doi: 10.1101/cshperspect.a016360



sensing events are both operative at the earlieststage of the encounter, but the speed of neuraltransmission offers a key advantage to coordi-nating the animal’s specific responses.

There are 302 neurons in C. elegans, and theinterconnections and response characteristics ofthese have been elucidated. A characteristic fea-ture of the worm’s response to infection is thatone family of neural circuits, activated whensensory neurons detect the presence of path-ogenic pseudomonas, transmits neurotrans-mitter-dependent signals to mediate pathogenavoidance behaviors, and to suppress the un-folded protein response in innate immune cells(Sun et al. 2011). In light of today’s advancedknowledge, and our detailed molecular mecha-nistic understanding about inhibitory reflexcontrol of innate immune responses in mam-mals, perhaps the perspective becomes clearer.The regulatory intersection between the ner-vous and innate immune systems in C. elegans,which has been retained in today’s mammals,originated from evolutionary pressure to opti-mize physiological homeostasis, and survival.

Beyond either field, there is little need toaccount for separateness in functional mecha-nisms that operate in a parallel time frame, andalso intersect to control the magnitude of hostdefense and response. But immunology andneuroscience have only recently embraced thistheory, so there is value in reviewing the conceptfrom the perspectives originating within each.From an historic perspective in immunology,the principal focus has been to understand thedevelopment and maintenance of immunity,and to provide a complete molecular mechanis-tic basis for how lymphocytes acquire and retainthe information necessary to remember the na-ture of infectious threats. This has been highlysuccessful for establishing the principles to ex-plain the diversity and repertoire of long-termhumoral and cellular immune memory that re-sults from antigen processing, presentation, andclonal development of B and T cells. A detailedmolecular understanding for the basis of innateimmunity was also produced. When likened toreflex or instinctual behavior, pathogen recog-nition by innate immune cells is encoded in thegenome, enables a relatively rapid and consis-

tent response to pathogen-associated and dam-aged-associated molecular patterns, and ishighly conserved. These mechanisms of immu-nity have been exceedingly well described with-in a knowledge sphere circumscribed by cells ofhematopoietic origin, operating in response topathogens, absent control from other systems.It is axiomatic within immunological think-ing that the earliest stages of sensing pathogensoccurs through activation of innate immunepattern-recognition receptor signals, and thatlater sensing occurs through adaptive immuneresponses.

In an historic time line parallel with theseunfolding events, the principle focus of neuro-science has been to understand the functions ofthe brain and nervous system. This establishedneurons as the principal cells controlling infor-mation, learning, and memory; the chemicalbasis of neurotransmission; and the principlesof synaptic neural communication underlyingneuronal signaling. In the mid-20th century,Sherrington famously recognized that the sim-ple reflex, comprised of a sensory arc, interneu-rons, and a motor arc, is the fundamental unitfor information processing on which the entirenervous system evolved (Sherrington 1906).From C. elegans and to modern mammals, sen-sory neurons respond to chemical, mechanical,or temperature changes in the environment.This information is transduced via ion channelsand gradient potentials that generate action po-tentials that are relayed to other neurons. Dur-ing pathogen threat, or in response to other en-vironmental changes, information processingin the nervous system begins with sensory inputtransmitted in the afferent arc of a reflex circuit.

Sensory neural signals in mammals arerouted through nuclei residing within the cen-tral nervous system, primarily in the brain stem.These nuclei produce the output that regulatesorgan function by transmission through themotor arc to complete the reflex circuit. Thesimple reflex unit is the basic building blockthat evolutionary pressure used to integrateand assemble the mammalian nervous systemas the regulatory and coordinating system ofhost responses to infection, injury, and otherthreats. Nuclei that provide the relay stations

Integration of Neural and Immune Pathogen Sensing

Advanced Online Article. Cite this article as Cold Spring Harb Perspect Biol doi: 10.1101/cshperspect.a016360 3



for regulatory reflexes are housed in the evolu-tionarily ancient regions of the mammalianbrain stem. These brain stem nuclei control thehomeostatic set points of all basic physiologicalfunctions associated with health, and the re-sponse to environmental change.

Individual neural reflex circuits are stimu-lated in response to specific sensory inputs aris-ing within the body organs. The reflexively con-trolled release of neurotransmitters from themotor arc controls the output of each organ.Ultimately, the physiological, or involuntary,behavior of the host is the result of the sum totalaction of all reflexes connecting its organs to itsbrain stem. These involuntary reflexes functionwith limited or no input from the cerebral cor-tex. They operate largely beneath conscious con-trol. Primitive animals, like C. elegans, which(presumably) lack the cerebral faculties of con-sciousness and self-awareness, also show reflexbehavior in response to changes in pressure,temperature, the chemical composition of theenvironment, and the presence of pathogens.

In considering the role of regulatory feed-back on the molecular mechanisms underlyingimmunity, it bears special note here that there isa modern tendency to center discussions ofmammalian neural information processing onthe cerebral cortex. Termed “cortical conceit,”one may view the sophisticated human cerebralencephalon as an undoubted source of inesti-mable pride, a grand evolutionary accomplish-ment (Dubos 1998). But the advanced cerebralcortex found in mammals is a relatively newaddition to biology, one that is largely dispen-sable for understanding basic mechanisms ofphysiological and immunological homeostasis.Surgically or genetically decorticated mammalscan well maintain physiological and immuno-logical set points within a healthy range, as longas the brain stem nuclei and corresponding re-flex circuits are functional. Basic reflex circuits,having developed in an evolutionary time,maintain physiological responses during infec-tion and injury without requiring input fromthe cortex to operate competently. Signals de-scending from the higher brain can influencethe functional output of the reflex nuclei, butthese psychological implications need not be

considered here to understand the basic physi-ological principles governing homeostasis. So,in defining rules that govern how the nervoussystem modulates the immune system to estab-lish immunological homeostasis, we can dis-pense with the need to consider individualacts of conscious will, or choices made. Thesystem operates in the background of consciousexistence, driven by sensory inputs that elicitreflex responses.

These reflex circuits originating in the evo-lutionary precursors of the modern mammali-an brain stem nuclei conferred the ability tomodulate cardiac and pulmonary function inresponse to infection, and countless other pro-tective and beneficial organ responses. Animalsexpressing such competent beneficial reflexesgained a preservation and reproductive advan-tage over breeding populations lacking it. Intime, the accumulation of individual simple re-flexes endowed evolving species with optimalflexibility necessary to use effective defensiveresponses and maintain homeostasis in re-sponse to a range of changes in the internaland external milieu. The identity and functionof these regulatory reflexes in mammals was re-vealed in classic physiological studies of the an-atomically accessible body systems. Considerthe cardiovascular system, which is beautifullypresented in the thorax when the sternum isdivided, to behold the beating heart.

With each cardiac contraction it is possibleto observe, measure, and record numerous ex-perimental end points reflecting the intricatephysiological working of the reflexively con-trolled heart. Cutting the vagus nerve to theheart produces an increase in heart rate, indi-cating that heart rate is tonically suppressed bydescending neural signals from the brain stem.This descending inhibitory signal is transmittedin the motor arc of a reflex initiated when anincrease in heart rate is sensed by afferent bar-oreceptor neurons. Action potentials arrivewithin brain stem nuclei, and from there, neuralsignals exit the brain stem to be transmittedback to the heart to decrease heart rate relativeto a predetermined set point. Other reflexes,activated when heart rate decelerates below itsset point, culminate in signaling via other mo-

K.J. Tracey




tor arcs that increase heart rate. From observa-tions of this sort, physiologists and neuroscien-tists built detailed models and developed rulesof feedback control based on individual unitsof reflex action to explain how cardiovascularhomeostasis is maintained during a range ofchanges in the internal and external milieu.Similar rules have been developed and appliedto other organ systems. The regulatory neuralsystems that reflexively control physiologicalhomeostasis operate in real time, are highly sen-sitive to changes in external and internal envi-ronment, and ultimately enable organ systemsto function within a narrow range of metabolic,thermal, nutrient, and behavioral parametersthat is compatible with cellular and animalhealth.

The vagus nerve is a principal nerve in mam-mals that transmits afferent, sensory informa-tion from the organs to the brain stem, and itis a mechanism for informing the nervous sys-tem about the status of the internal milieu. Fromthe visceral organs, it travels a meanderingcourse across the abdomen, thorax, and neck,entering the brain stem, where its sensory nerveendings synapse on, and relay information to,neurons in the nucleus tractus solatarius. Inter-neurons arising there relay signals to the dorsalmotor nucleus of the vagus nerve, representing amajor cholinergic outflow tract that returns tothe peripheral organs. The motor arcs stimulat-ed by vagus sensory afferent neurons return tothe organs via efferent neurons that travel alongtwo major routes: the sympathetic chain, andthe vagus nerve itself. At the nerve terminus,motor neurons from these two major outflowpaths release acetylcholine, norepinephrine, va-soactive peptide, and other neuropeptides thatbind to cognate receptors in the target organ’sepithelial and parenchymal cells, to modulatecellular metabolism, function, and secretory ca-pacity. Studies of the cardiopulmonary, gastro-intestinal, renal, hepatic, and endocrine systemsestablished the anatomic and neurophysiologi-cal principles underlying reflex control of phys-iological homeostasis. Until quite recently, how-ever, the existence of regulatory reflexes exertingcontrol over the immune system had evadeddescription. Immunological thinking had not

viewed immunity as a reflexively regulated re-sponse.

REFLEX REGULATION OF INNATEIMMUNITY

This began to change following the discoverythat cytokines produced in response to infec-tion and injury are necessary and sufficient me-diators in disease pathogenesis. This providedexperimental tools and strategies that enabledthe illumination of neural circuits that regulateinnate immunity. Termed “the cytokine theoryof disease,” cytokines produced in response toinfection and injury mediate cellular, metabol-ic, and pathological effects underlying the ma-jor physiological and clinical manifestations ofillness (Tracey 2007). Restated, disease occurswhen there is a loss of immunological homeo-stasis producing an unbalanced or excessivecytokine response. In such cases, inhibiting ormodulating the cytokine response to restore im-munological homeostasis can prevent or reversedisease, and it should be stressed that the cyto-kine theory does not state that all cytokines are“bad.” By analogy, the germ theory does notstate that all microbes are pathogenic. Rather,the cytokine theory indicates that in some in-stances of cytokine imbalance or excess, treat-ment of diseases can be specifically accom-plished with targeted therapeutics.

The cytokine theory of disease was validatedin mammals, including humans, within a fewyears after early direct evidence became avail-able that tumor necrosis factor (TNF) is a nec-essary and sufficient mediator of acute septicshock during infection (Tracey et al. 1986,1987). The first principles for administeringanti-TNF monoclonal antibodies as therapywere established in baboons infected with Es-cherichia coli (Tracey et al. 1987). Treated ani-mals survived despite the presence of replicatingbacteria in their bloodstream, proving that thedisease (septic shock) required the activity ofTNF. Disease was attributable to the activity ofthe cytokine, regardless of the status of the path-ogen (Tracey et al. 1987). Today it is common towitness this principle being applied to patientsreceiving monoclonal anti-TNF for the treat-





ment of rheumatoid arthritis, inflammatorybowel disease, and other conditions. From apractical point of view, the earliest underlyingor inciting etiological agent can be dissociatedfrom the disease. The physician standing at thepatient’s bedside is witness to signs and symp-toms of inflammation, organ damage, and tox-icity mediated by the direct action of cytokines.In many, if not most, diseases, excessive cyto-kine activity can be considered as a failure ofimmunological homeostasis.

It was from this perspective, in the 1990s,that my colleagues and I began to directly ad-dress the possibility that reflex neural circuitshad a critical role in regulating cytokine releaseand maintaining immunological homeostasis.We focused on the TNF response in animalmodels of infection and injury. The choice ofTNF as the target of these studies was important,because it represents an early-stage response,rapidly induced in response to pathogenic stim-ulation. The measurable TNF response fallswithin a time window peaking 90 min after ex-posure to endotoxin, quite compatible withmaking observations that could be interpretedin the context of experimental manipulation ofneural circuits. To determine whether neural sig-nals suppress TNF release from tissue macro-phages, we applied stimulating electrodes tothe vagus nerve where it traversed the neck onits descending course to the reticuloendothelialsystem (Borovikova et al. 2000). Electrical stim-ulation of the vagus nerve significantly inhibitedTNF release in red pulp and marginal zone mac-rophages in spleen, suggesting that descendingsignals in the motor arc of an “inflammatoryreflex” participate in regulating the innate im-mune response and maintaining immunologi-cal, or “cytokine homeostasis” (Tracey 2002; Ro-sas-Ballina et al. 2008).

When the vagus nerve was divided beforeexposing animals to endotoxin, we observedthat TNF production by splenic macrophageswas enhanced (Borovikova et al. 2000). By anal-ogy to the tonic inhibitory influence of vagusnerve signaling, which inhibits heart rate, thisfact indicated that constitutively active or tonicvagus nerve signals slow or inhibit the innateimmune response to endotoxin. By following

the track of the vagus nerve from the brainstem, down the neck, across the thorax, andinto the abdomen, and either dividing the nervein some cases, or electrically stimulating it alongits course in others, we discovered that the pointof convergence of the vagus nerve signals to reg-ulate serum TNF during endotoxemia is in thespleen (Huston et al. 2006; Rosas-Ballina et al.2008). Electrical signals applied to the cervicalvagus nerve descend to the celiac ganglion,which is the origin of motor neurons that inner-vate the spleen. Vagus nerve and splenic nerveaction potentials inhibit the release of TNF inred pulp and marginal zone macrophages,which together account for .90% of the peakserum TNF levels following the acute onset ofendotoxemia.

Three implications from these findings areimmediately obvious. First, the TNF producinginnate immune response to endotoxin is regu-lated by neural signals. The constitutive inhibi-tory activity of neural circuits establishes a setpoint for the innate immune response to endo-toxin. Second, the immune system is anatomi-cally and functionally innervated. Although ex-tensive prior anatomic evidence indicated thatall of the major organs of the reticuloendothelialsystem receive neural connections from thebrain stem, the prior anatomic findings hadnot provided a scientific framework for experi-mental studyof the functional operation of theseneurons. Approaching this as a functional, neu-rophysiological property of neuronal activity toregulate the innate immune response to endo-toxin opened a door to delineating molecularmechanisms. Third, it is overwhelmingly prob-able that the basic principles of reflex actionactivated by sensory input also apply to otheraspects of innate immunity and adaptive immu-nity. The evolutionary emergence of lympho-cytes, and adaptive immunity, occurred roughlyin parallel with the emergence of more sophisti-cated neural networks arising from evolution-arily expanding neural circuits and connectivity.It is nearly impossible to imagine an evolvingsituation in which the environmental changescaused by infection and injury did not intersectwith signals that simultaneously influence boththe immune and nervous systems.

K.J. Tracey




As broadly defined, immunity is the state ofacquiring (learning) and retaining (remember-ing) systems to confer protection against infec-tion. On accepting the theory that neural sig-nals, which regulate immune responses, arecostimulated during the activation of innate im-munity, and provide regulatory signals in theearliest stages of acquiring immune memory,then it can be strongly argued that a completeunderstanding of immunity requires incorpo-rating neural mechanisms. To determine themechanism for reflex inhibition of innate im-munity, we experimentally followed the path ofthe neural signals arising in the vagus nerve andterminating in the spleen (Rosas-Ballina et al.2011). Within the spleen, we observed splenicnerve endings terminating in synapse-like struc-tures adjacent to lymphocytes. In transgenicmice, which coexpressed green fluorescent pro-tein in cells that express choline acetyltransfer-ase, the rate-limiting enzyme that catalyzes thebiosynthesis of acetylcholine, we observed thatT lymphocytes capable of synthesizing acetyl-choline resided in close proximity to splenicnerve endings. Electrical stimulation of the va-gus nerve stimulated the adrenergic splenicnerve to release norepinephrine, which in turnstimulated Chat-expressing T cells to releaseacetylcholine. Lymphocyte-derived acetylcho-line is the neurally activated molecular mediatorthat binds to a7 nicotinic acetylcholine recep-tors expressed by red pulp and marginal zonemacrophages (Wang et al. 2003). Ligand recep-tor interaction inhibits the release of TNFthrough a molecular mechanism that has beenattributed to formation of a heteroprotein com-plex comprised of a7 binding to JAK-STAT thatsuppresses activation of the nuclear transloca-tion of nuclear factor kB and to stabilizing mi-tochrondrial membrane permeability (de Jongeet al. 2005; Lu et al. 2014). The functional im-portance of these T cells in this neural circuit isstriking, because T-cell-deficient nude mice lacka working inflammatory reflex. Electrical stim-ulation of the vagus nerve in nude mice failsto inhibit TNF. This phenotype is reversed bypassive transfer of Chat-expressing T cells intonaıve recipient nude mice, which restores thefunctional inflammatory reflex in these animals

(Rosas-Ballina et al. 2011). Thus, the inflamma-tory reflex is defined by vagus nerve action po-tentials that activate the release of an inhibitoryneurotransmitter, acetylcholine, from special-ized lymphocytes to suppress the magnitudeof the TNF response in spleen. A prototypicalneural circuit prevents overexpression of TNF toestablish healthy immunological homeostasis.

Our initial observations on the neural reg-ulation of TNF release enabled us to formulatethe fundamental principles for reflex control ofimmunity. Discrete neurons terminating in spe-cific molecular mechanisms modulate the im-mune response to infection (Tracey 2009; An-dersson and Tracey 2012a). Inhibitory neuralcontrol of innate immunity by acetylcholine isnot restricted to controlling TNF, because ace-tylcholine interaction with a7 nicotinic acetyl-choline receptors in macrophages significantlyinhibits activation of the inflammasome, theprotein complex required for the release of lead-erless cytokines, including IL-1, HMGB1, andIL-18 (Borovikova et al. 2000; Lu et al. 2014).Expression of a7 nAChR is essential to the in-tegrity of the inflammatory reflex, because a7nAChR knockout mice have an impaired in-flammatory reflex. Vagus nerve stimulation failsto suppress inflammasome activity in theseanimals; and, moreover, activation of innate im-munity in the a7 nAChR knockout mice pro-duces significantly higher levels of TNF andthe leaderless cytokines (Wang et al. 2003; Luet al. 2014).

SENSORY STIMULATION OFNEUROLOGICAL REFLEXES IN IMMUNITY

As is the case for all reflexes, the motor arc of theinflammatory reflex is activated in response tosignals arising in the sensory arc. Beyond theevidence cited above that the interaction ofC. elegans with pathogens begins by sensing sig-nals transmitted in afferent neurons, a series ofpivotal studies have provided insight into theorigin and function of the sensory arc of theinflammatory reflex in mammals. Originally,Linda Watkins and her colleagues (1995)made the seminal discovery that the fever re-sponse induced by the intra-abdominal admin-





istration of IL-1 requires an intact and func-tional vagus nerve, because when they surgicallydivided the vagus nerve, animals maintainednormothermia despite exposure to intra-ab-dominal IL-1. Thus, the presence of IL-1 inthe abdomen, or more specifically in the hepaticcirculation, is sensed by the vagus nerve, whichpropagates action potentials ascending to syn-apses in the nucleus tractus solatarius in thebrain stem (Fairchild et al. 2011). Sensory vagusnerve action potentials mediated by IL-1 havebeen experimentally shown using recordingelectrodes applied to the vagus nerve at variouspoints along its course from the abdomen to thebrain stem (Niijima 1996). Compound actionpotentials arising in response to intrahepatic IL-1 have also been recorded in the splenic nerve,which is a pure motor nerve lacking sensoryfibers (Niijima et al. 1991). This gives direct evi-dence that sensory vagus nerve signals activatecompound action potential signals descendingto the spleen via the motor arc of the inflamma-tory reflex.

Other inflammatory mediators have beenimplicated in stimulating sensory action poten-tials in the vagus nerve, including endotoxin,TNF, prostaglandins, substance P, and other en-dogenous and exogenous products of infectionand injury (Chiu et al. 2012). The sensory andmotor arcs of the inflammatory reflex withinthe vagus nerve are positioned to anatomicallyand functionally respond to and modulate thehost response to infection and threat in the vis-ceral organs of the body compartment. The im-portance of this inhibitory neural circuit tomaintaining immunological homeostasis is in-dicated by the findings that animals deficient inthis reflex are rendered increasingly sensitive toexcessive TNF release. This raises the next ques-tion: Can reflexes stimulate, rather than inhibit,innate immune responses to maintain immu-nological homeostasis?

As far back as 1874, it had been proposedthat inflammation is mediated by action poten-tials transmitted in peripheral neurons (Chiuet al. 2012). This stemmed from the observa-tions that electrical stimulation of dorsal rootneurons produced vasodilation in the inner-vated skin. Today we understand that the in-

flammatory effect of these motor neurons is at-tributable to neurotransmitters released intothe tissue including neuropeptides, catechol-amines, and substance P. These molecules inter-act with cognate receptors expressed on endo-thelial cells, smooth muscle cells, lymphocytes,monocytes, and macrophages that in turn me-diate vasodilation, increased capillary perme-ability, neutrophil recruitment, and swelling.Signaling through these stimulating neuralpathways produces an enhancement of inflam-mation attributable to the biological mecha-nisms of specific proinflammatory neurotrans-mitters activated by that specific neural circuit.As these motor circuits are stimulatory, definedby their activity to increase the output of innateimmune responses, there is evidence supportinga functional approach to studying the regulationof innate immunity by combined inhibitory andexcitatory neural signaling mechanisms.

A basic principle in neuroscience is that theinteraction of inhibitory and excitatory neuralcircuits mediates the output of a neural net-work, and determines the final state of organfunction. Major forms of information process-ing in the nervous system are routinely definedby the identity of the excitatory or inhibitoryneurotransmitter released in response to specif-ic action potentials. Accordingly, depending onthe neurotransmitters released, neural circuitscan either inhibit inflammation, as in the caseof cholinergic signals, or enhance inflamma-tion, as in the cases of proinflammatory neuro-peptides, adrenergic signals, and substance P.The functional output depends on the neuro-transmitter receptors, rather than the neuro-transmitters per se, as in the case of adrenergicsignaling by norepinephrine secreting neurons,which can enhance or inhibit inflammation,depending on whether the principle receptorsengaged are a or b adrenergic GPRs. Indeed,most innervated tissues receive input from cir-cuits that release immune inhibitory and im-mune excitatory signals, so the net effect of neu-ral control on the immune response will dependon the sum total interaction of molecular mech-anisms resulting from ligand receptor signaltransduction in a cell-specific and neurotrans-mitter-specific path.

K.J. Tracey




To this point, I have argued that the outputof neural reflex circuits regulates the innateimmune response to pathogens, and that thesecircuits can be activated by IL-1 and other en-dogenous products of inflammation and tissuedamage. This is in keeping with pervasive im-munological thinking that the innate immunesystem is the primary sensor of the presence ofpathogens. Next, consider that the regulatoryintersection between the nervous and immunesystems can also originate with pathogen sens-ing. It is clear that innate immune systemreceptors respond to pathogen-associated mo-lecular patterns during early phases of infec-tion, and that clonal expansion of lymphocytesexpressing specific pathogen sensing receptorsconfers long-term protection against sub-sequent infections. But how early is “early,”and what role might the nervous system playif it detects the presence of microbes or tissuedamage first, before innate immune responses?Can neurons respond to bacteria directly, ormust there be an intermediate sensing step ofmicrobial recognition by the innate immunesystem?

Consider that peripheral sensory neuronsform a dense meshlike network that envelopsand covers all tissues exposed to the externalenvironment, including the epithelial lining ofthe skin and soft tissues, as well as the pulmo-nary, genitourinary, and gastrointestinal sur-faces. Wherever microbes might gain entry, asensory neural net is in place to respond. Spe-cialized “nociceptor” neurons can be activatedto transmit afferent action potentials by expo-sure to specific molecules. In C. elegans, with itslimited repertoire of neurons, glycolipids con-taining ascarylose (3,6-dideoxy-L-arabino-hex-ose), termed “ascarosides,” stimulate specificnocioreceptor neurons that lead to discretebehavioral responses (Ludewig and Schroeder2013). For example, exposure of worms toascr#3 mediates either attraction or repulsionof hermaphrodites and males, whereas chemicalmodification of ascr#3 by addition of a trypto-phan adduct (producing “Icas#3”) producesentirely different aggregation and attraction be-havior. In this example, discrete chemical mod-ification of awell-characterized molecular entity

provides a differential sensory input into reflexcircuits that produce visibly perceptible behav-ioral responses.

Stimulation of nocioreceptors producesgraded potentials that propagate short distanc-es, and can lead to stimulation of action poten-tials in a frequency that reflects the intensity ofthe original stimulus. These action potentialstravel longer distances along axons from the pe-ripheral tissue into the central nervous system,where the incoming information can be relayedto the brain stem, and can activate an interneu-ron to stimulate a motor arc of a simple reflexthat returns to the periphery. Another featureof some nocioceptive neurons is that signalspropagating toward the central nervous systemcan be diverted at branch points to travel backtoward the periphery in a phenomenon termedan “axon reflex” (Yaprak 2008). By this mecha-nism, TNF, IL-1, prostaglandins, substance P,and other molecular products of inflammationstimulate nocioceptive neurons to stream sig-nals to the brain stem and back into the region-al tissues. Functional expression of receptorsfor cytokines (e.g., IL-1R, TNFR), pathogen-as-sociated molecules (e.g., TLR3, TLR4, TLR7,TLR9), and damage-associated molecules (e.g.,RAGE, P2X3) have all been implicated in medi-ating neuronal hyperpolarization and signaling(Chiu et al. 2012).

From the neuroscience perspective, there isnothing particularly curious about the nervoussystem sensing molecular input from the milieuinterior and exterior. Changes in the molecularcomposition of the environment directly medi-ate the generation of action potentials, and ini-tiate the neural platform for responding tothreat, which stimulates learning. If one ap-proaches this from within a narrow immunol-ogy perspective, there will be significant bias infavor of theories that cells within the innate im-mune system, not the nervous system, are theprimary sensing apparatus to detect pathogens.From an evolutionary perspective, however,there is no reason to ascribe innate sensing pri-macy to the immune system. Indeed, the factsargue to the contrary. Atomic structural mech-anistic studies of molecular signaling throughneurons in C. elegans, the evolutionary precur-





sors to mammalian neuronal reflexes, react dif-ferentially to subtle changes in molecular struc-ture. It can be strongly argued that neurons, aswell as innate immune cells, evolved to “sense”molecules derived from pathogens in a mannerthat initiates highly specific and coordinatedresponses that benefit survival, and that evolu-tion would preserve these advantages.

The recent discovery by Clifford Woolf andcolleagues that mammalian nocioceptors sensepathogens independently from the innate im-mune system indicates that some adjustmentwill have to be made to prior immune-centricsensing theories (Chiu et al. 2013). Inoculationof bacteria into mice produced pain and inflam-matory responses, even when animals were ren-dered genetically deficient in TLR and MyD88-dependent signal transduction mechanisms.N-formylated peptides derived from Gram-negative and Gram-positive bacteria are sensedby G-protein-coupled formyl peptide receptors(FPRs) on neurons. Administration of E. coli–derived peptide (fMLF), and Staphylococcusaureus–derived peptide (fMIFL) to mice-acti-vated calcium flux in nocioceptive neuronsthat mediate pain and mechanical hypersensi-tivity. Exposure of the S. aureus–derived pore-forming toxin, a-haemolysin, induced a con-centration-dependent calcium flux in sensoryneurons, which express A disintegrin and met-alloprotease 10 (ADAM10), implicated in mem-brane pore assembly. The assembly of molecularpores in response to a-haemolysin is sufficientto mediate ion flux and generate action poten-tials in sensory neurons. Mice rendered defi-cient in nocioceptive neurons (Nav1.8-Cre/diphtheria toxin A mice) and exposed to S. au-reus–developed enhanced inflammatory re-sponses, including increased neutrophil andmonocyte infiltration, increased lymph nodeswelling, and increased production of TNF.Thus, the interaction of pathogen-associatedmolecules on neurons directly activated sensoryinformation, transmitted as action potentials,to mediate a locally inhibitory neural circuitthat suppresses inflammation (Chiu et al.2013). This direct mechanistic evidence estab-lishes that pathogen-derived molecules can besensed by neurons, that specific molecular

products activate specific neuronal mechanismsthat culminate in the generation of action po-tentials, that localized neural reflexes contributeto regional regulation of innate immunity, andthat all of this occurs at the earliest, immediatestages of host responses that are associated withlearning and, eventually, memory.

CONCLUSIONS

Viewing the onset of immunity as commencingwith neural sensing independent from, and co-stimulatory with, immune cell sensing has pro-found implications to immunological thinking.For students interested in the functional activityof hematopoietic cells, it will remain an option,of course, to continue investigation into the in-teraction between specific microbes and patho-gen-associated molecules with specific cognatereceptors expressed in cells from the myeloidlineage. This can be furthered in exquisite mo-lecular detail by pursuing signal transductionmechanisms that converge on gene expressionpatterns. Students interested in understandinghow immunological memory develops in mam-mals during infection, and those interested inunderstanding how the innervated host oper-ates as a physiologically regulated animal withinconstraints of a stable internal milieu, will take adifferent approach. The revolutionary new ap-proach to immunological thinking will embracethe mechanisms of inhibitory and excitatoryneural feedback loops that function from theonset of infection, and are coordinated by brainstem reflexes and regional axonal reflexes, tomodulate the output of the innate immune cellsthat independently sense the presence of micro-bial agents.

Revolutions can be messy business, and pre-dicting the outcome even messier. But some-times, as perhaps illustrated by Janeway’s pre-science in defining the second revolution inimmunological thinking, the outcome can bepredetermined by facts. At present, nearly25 years after Janeway’s prediction, innate im-munity has been elevated to heightened pre-eminence in immunological thinking. There isan intense and appropriate focus on under-standing the role of innate immunity in the ac-

K.J. Tracey




quisition of adaptive immunity. Looking to thecoming revolution, there is clear evidence todaythat innate immunity is regulated by neural re-flexes activated by pathogens and host-derivedmediators of infection, inflammation, and inju-ry; and there is strongly suggestive evidence thatneural signals modulate adaptive immunity aswell (Wong et al. 2011; Arima et al. 2012; Mina-Osorio et al. 2012). From my vantage point, thishas all the makings of a revolution in immuno-logical thinking that is under way owing to aconvergence, not a clashing, of two major fields.A complete understanding of immunity re-quires fully understanding the neural responseto infection and injury (Andersson and Tracey2012b). Tools provided by molecular biology,neurophysiology, and cell biology render this aparticularly exciting time to embrace this inter-section of neuroscience and immunology, andto consider that advances from worms to mam-mals hold promise for advances in the clinic.

REFERENCES

Aballay A. 2013. Role of the nervous system in the control ofproteostasis during innate immune activation: Insightsfrom C. elegans (ed. Goldman WE) PLoS Pathog 9:e1003433.

Andersson U, Tracey KJ. 2012a. Neural reflexes in inflam-mation and immunity. J Exp Med 209: 1057–1068.

Andersson U, Tracey KJ. 2012b. Reflex principles of immu-nological homeostasis. Annu Rev Immunol 30: 313–335.

Arima Y, Harada M, Kamimura D, Park J-H, Kawano F,Yull FE, Kawamoto T, Iwakura Y, Betz U, Marquez G, etal. 2012. Regional neural activation defines a gatewayfor autoreactive T cells to cross the blood–brain barrier.Cell 148: 447–457.

Borovikova LV, Ivanova S, Zhang M, Yang H, Botchkina GI,Watkins LR, Wang H, Abumrad N, Eaton JW, Tracey KJ.2000. Vagus nerve stimulation attenuates the systemicinflammatory response to endotoxin. Nature 405: 458–462.

Chiu IM, Von Hehn CA, Woolf CJ. 2012. Neurogenic in-flammation and the peripheral nervous system in hostdefense and immunopathology. Nat Neurosci 15: 1063–1067.

Chiu IM, Heesters B, Ghasemlou N, Von Hehn C, Zhao F,Tran J, Wainger B, Strominger A, Muralidharan S,Horswill AR, et al. 2013. Bacteria activate sensory neu-rons that modulate pain and inflammation. Nature 501:52–57.

De Jonge WJ, van der Zanden EP, The FO, Bijlsma MF, vanWesterloo DJ, Bennink RJ, Berthoud H-R, Uematsu S,Akira S, van den Wijngaard RM, et al. 2005. Stimulationof the vagus nerve attenuates macrophage activation by

activating the Jak2-STAT3 signaling pathway. Nat Immu-nol 6: 844–851.

Dubos R. 1998. So human an animal: How we are shaped bysurroundings and events. Transaction, Piscataway, NJ.

Fairchild KD, Srinivasan V, Moorman JR, Gaykema RP,Goehler LE. 2011. Pathogen-induced heart rate changesassociated with cholinergic nervous system activation.Am J Physiol Regul Integr Comp Physiol 300: R330–R339.

Huston JM, Ochani M, Rosas-Ballina M, Liao H, Ochani K,Pavlov VA, Gallowitsch-Puerta M, Ashok M, Czura CJ,Foxwell B, et al. 2006. Splenectomy inactivates thecholinergic antiinflammatory pathway during lethal en-dotoxemia and polymicrobial sepsis. J Exp Med 203:1623–1628.

Janeway C. 1989. Approaching the asymptote? Evolutionand revolution in immunology. Cold Spring Harb SympQuant Biol 54: 1–13.

Lu B, Kwan K, Levine YA, Olofsson PS, Yang H, Li J, Joshi S,Wang H, Andersson U, Chavan SS, Tracey KJ. 2014. a7nicotinic acetylcholine receptor signaling inhibitsinflammasome activation by preventing mitochondrialDNA release. Mol Med doi: 10.2119/molmed.2013.00117.

Ludewig AH and Schroeder FC. 2013. Ascaroside signalingin C. elegans. In WormBook (ed. The C. elegans ResearchCommunity), doi/10.1895/wormbook.1.155.1.

Mina-Osorio P, Rosas-Ballina M, Valdes-Ferrer SI, Al-AbedY, Tracey KJ, Diamond B. 2012. Neural signaling in thespleen controls B-cell responses to blood-borne antigen.Mol Med 18: 618–627.

Niijima A. 1996. The afferent discharges from sensors forinterleukin 1b in the hepatoportal system in the anesthe-tized rat. J Auton Nerv Syst 61: 287–291.

Niijima A, Hori T, Aou S, Oomura Y. 1991. The effects ofinterleukin-1b on the activity of adrenal, splenic andrenal sympathetic nerves in the rat. J Auton Nerv Syst36: 183–192.

Pradel E, Zhang Y, Pujol N, Matsuyama T, Bargmann CI,Ewbank JJ. 2007. Detection and avoidance of a naturalproduct from the pathogenic bacterium Serratia marces-cens by Caenorhabditis elegans. Proc Natl Acad Sci 104:2295–300.

Rosas-Ballina M, Ochani M, Parrish WR, Ochani K, HarrisYT, Huston JM, Chavan S, Tracey KJ. 2008. Splenic nerveis required for cholinergic antiinflammatory pathwaycontrol of TNF in endotoxemia. Proc Natl Acad Sci 105:11008–11013.

Rosas-Ballina M, Olofsson PS, Ochani M, Valdes-Ferrer SI,Levine Y, Reardon C, Tusche MW, Pavlov VA, AnderssonU, Chavan S, et al. 2011. Acetylcholine-synthesizing Tcells relay neural signals in a vagus nerve circuit. Science334: 98–101.

Sherrington SCS. 1906. The integrative action of the nervoussystem. Yale University Press, New Haven.

Styer KL, Singh V, Macosko E, Steele SE, Bargmann CI,Aballay A. 2008. Innate immunity in Caenorhabditis ele-gans is regulated by neurons expressing NPR-1/GPCR.Science 322: 460–464.

Sun J, Singh V, Kajino-Sakamoto R, Aballay A. 2011. Neu-ronal GPCR controls innate immunity by regulating non-





canonical unfolded protein response genes. Science 332:729–732.

Tracey KJ. 2002. The inflammatory reflex. Nature 420: 853–859.

Tracey KJ. 2007. Physiology and immunology of the cholin-ergic antiinflammatory pathway. J Clin Invest 117: 289–296.

Tracey KJ. 2009. Reflex control of immunity. Nat Rev Im-munol 9: 418–428.

Tracey KJ, Beutler B, Lowry SF, Merryweather J, Wolpe S,Milsark IW, Hariri RJ, Fahey TJ, Zentella A, Albert JD.1986. Shock and tissue injury induced by recombinanthuman cachectin. Science 234: 470–474.

Tracey KJ, Fong Y, Hesse DG, Manogue KR, Lee AT, Kuo GC,Lowry SF, Cerami A. 1987. Anti-cachectin/TNF mono-clonal antibodies prevent septic shock during lethal bac-teraemia. Nature 330: 662–664.

Wang HH, Yu M, Ochani M, Amella CA, Tanovic M, SusarlaS, Li JH, Yang H, Ulloa L, Al-Abed Y, et al. 2003. Nicotinicacetylcholine receptora7 subunit is an essential regulatorof inflammation. Nature 421: 384–388.

Wang H, Liao H, Ochani M, Justiniani M, Lin X, Hong L, Al-abed Y, Wang H, Metz C, Miller EJ, et al. 2004. Cholin-ergic agonists inhibit HMGB1 release and improve sur-vival in experimental sepsis. Nat Med 10: 1216–1221.

Watkins L, Goehler L, Relton J. 1995. Blockade of interleu-kin-1 induced hyperthermia by subdiaphragmatic vagot-omy: Evidence for vagal mediation of immune-braincommunication. Neuroscience 183: 27–31.

Wong CHY, Jenne CN, Lee W-Y, Leger C, Kubes P. 2011.Functional innervation of hepatic iNKT cells is immu-nosuppressive following stroke. Science 334: 101–105.

Yaprak M. 2008. The axon reflex. Neuroanatomy 7: 17–19.

K.J. Tracey




published online November 6, 2014Cold Spring Harb Perspect Biol Kevin J. Tracey and Immune Pathogen Sensing and ResponseApproaching the Next Revolution? Evolutionary Integration of Neural

Subject Collection Innate Immunity and Inflammation

DiseaseGroup 2 Innate Lymphoid Cells in Health and

Brian S. Kim and David Artis

The Chemokine System in Innate ImmunityCaroline L. Sokol and Andrew D. Luster

SystemInflammation and the Blood Microvascular

Jordan S. Pober and William C. SessaIntracellular Nucleic Acid SensorsMicrobial Sensing by Toll-Like Receptors and

Surya Pandey, Taro Kawai and Shizuo Akira

Lymphohematopoietic InterfaceSinusoidal Immunity: Macrophages at the

Subhankar MukhopadhyaySiamon Gordon, Annette Plüddemann and

Immunity and InflammationTumor Necrosis Factor Superfamily in Innate

John Sedý, Vasileios Bekiaris and Carl F. Ware

Innately Homeostatic−−Allergic InflammationLaurence E. Cheng and Richard M. Locksley Transduction by Pattern-Recognition Receptors

Emerging Principles Governing Signal

Jonathan C. Kagan and Gregory M. Barton

Sensing and ResponseIntegration of Neural and Immune Pathogen Approaching the Next Revolution? Evolutionary

Kevin J. Tracey

Lipid Mediators in the Resolution of Inflammation

al.Charles N. Serhan, Nan Chiang, Jesmond Dalli, et

InflammasomesMarcel R. de Zoete, Noah W. Palm, Shu Zhu, et al. Responses

Transcriptional Control of Inflammatory

Stephen T. Smale and Gioacchino NatoliIL-6 in Inflammation, Immunity, and Disease

KishimotoToshio Tanaka, Masashi Narazaki and Tadamitsu

DNA Degradation and Its DefectsKohki Kawane, Kou Motani and Shigekazu Nagata

http://cshperspectives.cshlp.org/cgi/collection/ For additional articles in this collection, see

Copyright © 2014 Cold Spring Harbor Laboratory Press; all rights reserved


http://cshperspectives.cshlp.org/cgi/collection/

http://cshperspectives.cshlp.org/cgi/collection/

http://cshperspectives.cshlp.org/cgi/adclick/?ad=55917&adclick=true&url=http%3A%2F%2Fwww.genelink.com


Approaching the Next Revolution? Evolutionary Integration...

Documents

Transcript of Approaching the Next Revolution? Evolutionary Integration...