BiologyandPhilosophy(2006)21:189–211 (cid:1)Springer2006 DOI10.1007/s10539-005-7908-y Contribution to the whole (H). Can squids show us anything that we did not know already? ANDREW PACKARD StazioneZoologica‘‘AntonDohrn’’,VillaComunale,Napoli80121,Italy;(e-mail:andrew@pack- ards.de) Received30November2004;acceptedinrevisedform25May2005 Key words: Community, Homeotaxy, Horizontal control of living things, Myogenic, Peer conformity,Physiology Abstract. For a multicellular organism to proceed from egg to adult it must: (i) undergo cell division,(ii)differentiate,(iii)remainaunifiedwhole(H ).Theserequirementsareatrightanglesto o each other. The first two are achieved through hierarchical processes (vertical control) that are relatively well understood, the third through non-hierarchical processes (horizontal control) physiologicalevidenceforwhichisabundant,thoughnotwidelyrecognizedasaformofcontrol. The essay gives an example of a tissue – the skin of a living squid – whose horizontal network propertiescometolightwhennervous(vertical)controlisremoved.Itoffersthenamehomeotaxy or‘peerconformity’forthegeneralprinciple(alliedtothecommunityeffect,Gurdon1988)that constrainsthepartsofthewholetobeinthesamestatewithinanygivenlayerofthenetwork– wherelayerscorrespondtoontogeneticstagesinthedevelopmentofthetissue–anddiscussesthe questionofaneedandanameforthisprincipleinBiology. Savouring of an intangible ‘holism’ and of Driesch’s dubious ‘entelechy’, the oldquestionofhowitisthattheorganismbehavesasawhole,andnotjustasa collectionofparts,cannotbesaidtobeofmajorconcerntotoday’sbiologists. Oneobstacleisthatholisticconceptsarenotoriouslydifficulttoformulate;the words and concepts available for dealing with wholes are perceived as inade- quate; we seem to have advanced little beyond the broad statements of the organismal and Gestalt schools. Another is the sheer success of reductionist methodology allied to molecular technologies. While this essay is essentially aboutthefirstoftheseproblems,itisheavilyinfluencedbytheexistenceofthe second. Asaworking‘wholeanimal’biologist,Iseeitasaquestionofcontrols,and seek a pragmatic way round the problem by asserting, on general biological grounds and as far as possible in non-technical language, some simple prin- ciples amply demonstrable in living things. To pose the question about wholes and parts in a dynamic form with rea- sonablechanceofbeingunderstoodbyseminaraudiences,Ihavebeenaskingit inacontextwithwhichallstudentsandpractitionersinthelifesciencesshould be more or less familiar. What is it that holds the blastomeres together so that 190 they continue to act as one (egg or organism) after the first division(s) of the oocyte?and–forreasonswhichwillbecomeevident–havebeenlinkingitwith a second question what is the opposite of differentiation? Theanswergiventothefirstquestion–forsimplicityabouttheblastomeres of a frog’s egg and their implicit loss of individuality subsequent to the first division – tends to be in terms of some specific and more or less well-known mechanism holding them together: desmosomes, adherins, gap junctions (see later) or the follicular envelope surrounding the egg, etc. Thereisofcoursenothingnewinthisappealtoamechanism,orprocess,in order to explain the behaviour of the blastomeres, rather than invoking a generalrule,orcharacteristic,ofcellularorganisationcommontoallMetazoa andMetaphytaandtoallstagesofdevelopment.ButasWoodgerpointsoutin histreatmentofthe‘MechanicalExplanation’inBiology(Woodger1929:260) ‘an explanation of any phenomenon always involves two factors: general laws andaspecifiedsetofentitiessubjecttothoselaws’.Heisquotingthephysicist C.D.BroadBroad1919).Biology,atthelevelconsidered,hasratherfew‘laws’ by comparison with physics. But some it needs, if only to enable one to dis- criminateempiricallybetweenthetwo–betweenwhatisgeneralandwhatisa specific instance. Origins of multicellular organisms From the point of view of phylogeny, the importance of questions about the relationship between the ‘unit’ of life (the cell) and the whole (H), is the light theycanthrowontheevolutionarystepbetweenone-celledcreatures(Protista) and many-celled animals and plants (Metazoa and Metaphyta). Plants and animals have common mechanisms of cell division (mitosis and meiosis) and these are supposed to have evolved only once. Gerhart and Kirschner (1997) write: ‘One development of great importance for future metazoan multicellularity was the loss of the cell wall in some unicellular eukaryoteancestor.Thelackofacellwall…permittedtheancestorsofanimal cells to interact directly with each other through apposed plasma membranes, to adhere to each other, to crawl on surfaces, to differentiate into complex shapes, to engulf other cells by phagocytosis, and to engage in junctional communication with other cells. Cell adhesion and junctional communication are characteristics of the formation of epithelia and the segregation of an internal milieu, which are found in all metazoa’. (See Mueller 2003 and other contributions to the same number of Integrative and Comparative Biology). The list of attainments mentioned does not include the extracellular matrix (ECM). The matrix is of special interest for the theme of this essay (Oschman 1984).Asitisacontinuumofpolymerfibresandfibrils,itcanactasintegrator. The matrix’s complex of large molecules – principally in metazoa the long- chain triple helix collagen fibres – is secreted by the cell and remains in com- munication with it through its other components: adhesive glycoproteins 191 attachedtointegrinsspanningthecellmembrane(Ingber1998).Intheformof the basal lamina of epithelia, it mediates movements and transformations of cell type during the embryonic development of animals – from the sponges onwards (Morris 1993) – and provides a 3-D microenvironment that helps define tissue specificity during organogenesis by signalling directly to the nucleus (see Bissell et al. 2003). Billionsofyearsbeforetheappearanceoftheextracellularmatrixofanimals, however, a functionally analogous role in constructing and integrating the whole was performed by the matrix of extracellular polymeric substances (exopolymer biofilm) still found in the famous ‘living fossil’ stromatolite communities (Reid et al. 2000) of Southern Australian beaches and the Bahamas. Thebiofilmofstromatolitesisproducedbybacteriathatinhabitthetopmost layerofthemultilayeredstructure1.Hereitentrapsfineparticleswhencovered by the tide, resulting in a sediment which is further colonised by the bacteria, and thus to further layers and growth of the whole – incidentally rendering distinctionsbetweenorganismandenvironmentsingularlydifficulttosustain2! Wholes and parts Bywhole(H)Ihaveinmindanindividualclearlyboundedinspaceandtime– such as an arrow-worm, a squid, a frog or a human being. By parts, I have in mind the differentiated and differentiating cells of such organism. For the purposesofthisessay,andfordidacticreasons,the‘whole(H)’isalsointended as the isolated ‘laboratory’ organism H that is still the object of much re- o search.Such‘H’isofcourseanartefact,notonlybecausemany,perhapsmost, organisms making up the biosphere are not single individuals clearly bounded inspaceandtime.Moreimportantly,asjustillustrated,noindividualorganism is really separable from the community, the environment or other network of relationships constituting the wider whole (H ) – whether or not it be treated w as such in the laboratory. The way in which this multicellular organism proceeds in time from egg to adult is highly predictable. It is alive, and whatever else we may mean by this dynamic state it presupposes intrinsic controls which non-living things do not have. The controls contained in the genome unfold as epigenetic instructions expressed in the processes of cell multiplication and differentiation. Thus the 1 Seebelow,‘quorumsensing’bythemicrobialcommunitiesofeverydayinfections. 2 ‘Cyanobacteriaaretheprimaryproducersinthissystemprovidingenergy,directlyorindirectly, fortheentirestromatolitemicrobialcommunity.…Mostofthesespeciesarehighlymotileandcan adjusttheirpositionandorientationwithinthesedimentmatrixinordertooptimizetheiraccessto irradianceandnutrients.Asindividualspecieshavedifferentphysicalandmetabolicproperties,this motility generally results in segregated distributions of species, which in turn contributes to the laminatedtexturesobserved’(Beboutetal.2001). 192 question posed to seminar audiences could be answered by stating that it is epigenesis, and the continuity of the genome from cell to cell through mitosis, that holds the organism together over time as one whole (H ). The answer, in o other words, is implicit in the description of the processes of embryogenesis and of differentiation. Forvariousreasonsthiswillnotsuffice,andnotjustbecausethedescription of developmental processes is still far from complete or because it leaves out explicit reference to reciprocal relationships between organism and environment. Thequestionhasariseninaverypractical way during myownworkon the remarkablecoloursandcolourchangesofsquidsandoctopuses–whicharesuch an important part of their behavioural repertoire and widely known to be underthecontrolofeyesandbrain.Ifapieceofskinistakenfromasquidora cuttlefish(habituallyfromanindividualthatisalreadyfullydifferentiated,and without much regard to its embryology or the ontogenetic dimension) and is placedunderthemicroscope,coloursareseentocomeandgoastensionwaxes and wanes in the muscle fibres surrounding its many tiny pigment spots. The optical signal generated by expansion and retraction of a pigment spot were harnessed by A.V. Hill and his collaborator to record the shape of the mechanical response – or myogram – (Hill and Solandt 1935) and led the comparativephysiologistErnstFlorey(1966) towrite ofsuchthings asbrown twitches and red tetani. Efforts to understand the colour patterns of these cephalopod animals by studies at the cellular (rather than at the behavioural) level immediately encounter two additional problems, however. First, the elements to which the colours are reducible are not single cells but chromatophore organs: each constructedfromseveraldifferentcelltypes.Themanymusclefibresbelonging to any one of these organs exhibit what is called myogenicity; they mostly act synchronously because they are intimately coupled3, and many chromato- phores act simultaneously with other chromatophore organs (whether or not they are under the command of nerves, see below) as if they too were coupled in some way. Second, in isolation or in preparations removed from their ori- ginallocation,thechromatophoreensembledoesnotbehavephysiologicallyas it does in the whole squid or octopus. Separation of the piece of skin (or of a pieceofgut,seefootnote2)fromitswholesquidcontext,damagesthenetwork linking the organs and interferes with its rhythms. On the other hand, attempts to study the intrinsic activity of these coupled ensembles, and the contribution that the coupling makes to normal activity in the intact animal (H ), meet with other kinds of difficulty. In the squid, the o 3 One early account of electrical coupling between cells was Florey’s and Kriebel’s (1969) description of junctions between neighbouring muscle fibres on the squid chromatophore. The squidembryowasanother(Potteretal.1966).Familiarexamplesofcoupledensemblesthatare ‘continuousfromwithin’(Loewenstein,1981),becausethecytoplasmoftheircellsislinkedthrough gapandotherkindsofjunctions,aretheheart,thegut,theuterusandthebladder(seebelow). 193 switching‘off’and‘on’ofcoloursbynervesandbrainduringthegenerationof natural patterns can mask evidence of coupling between cells, evoke it differ- entiallybyuncouplingtheparts,orcantightlyentrainactivity,interferingwith the underlying rhythms and spontaneity of the ensemble. A classical experi- mental way round the difficulty is toalter the system by removing only one of its components – in this case removing nervous control from the whole, or a partofit–effectivelyreducingtheindividualto(H -1 )(seebelow).Anotheris o n to study individuals in which this has occurred naturally or accidentally4. Two forms of control In order to proceed from egg to adult as a continuous self-regulating whole, our multicellular organism (H ) must: o 1. undergo cell division, 2. differentiate, 3. remain one (organism or individual). These requirements are of equal status and may be asserted as basic prin- ciples. The principles are of cell division and of differentiation (here bracketed together)andtheprincipleofonenessorunity.Thelatterdoesnotderivefrom the former. In analytical terms they are at the same level but scientifically – logically, qualitatively and heuristically – they are very different. They are at right angles to each other. The key to the difference is in the two kinds of control. Control being understood as the set of algorithms fulfilling the two requirements. Thefirstrequirementisfulfilledbyprocessesthatareessentiallyhierarchical and can be called ‘vertical’ control (Figure 1). During development it is the series of discontinuous steps in a sequence of causes and effects rich in infor- mation and with feedback both between levels, and between organism and environment, that carry it from simple to complex, from undifferentiated to differentiated5. Vertical (or hierarchical) control extends beyond embryo- and organo-genesis to physiological control exerted between levels by organ sys- tems that are themselves the product of differentiation – for instance the classical control exerted by the nervous, hormonal and immune systems, etc. The sense of a vertical dimension is reflected in some of the language used to describe their interactions (higher and lower motor control, etc). In the motor system on which this essay is grounded, nerves control muscles both during developmental time (for instance during differentiation of muscle fibre types andduringsynapseformation,wherethecontrolisalsoreciprocal)andduring 4 Forinstanceinstudiesofthebladderinhemiplegia,orofthepre-termuterus(whichiseffectively ina‘denervated’condition:seelater). 5 See Davidson et al. 2002 for a recent synthesis of the hierarchical genetic cascades in animal development. 194 Figure1. physiological time as the exercise of nervous commands over the activity of muscle. Thesecondkindofcontrol–thatmaintainsthephysiologicalonenessofthe organism (and fulfils requirement no. 2) – is non-hierarchical, is continuous in space and time, is not obviously either cause-and-effect related or information rich; andin ordinarysenses oftheword it isnotgenerally perceivedasaform of control (see Discussion). It is called ‘horizontal’ because, being exerted withinlevels,itcanbethoughtofasorthogonaltoverticalcontrol.Asweshall see, within any horizontal level of the conceptual hierarchy – indicated by the ellipsesinFigure 2–thecontinuityofHmaintainedinspaceandtimeisalsoa physical continuum. The‘organism’inFigure 2hasacertainpolarity(asymmetry),withthehead end or animal pole (+) dominating over the tail end or vegetal pole ((cid:1)). The shaded ellipses are horizontal continuities of cell type and of stage of differ- entiationattwolevelsalongthisaxis.Withinonelevel,cellsareconsidered,for Figure2. 195 thepurposesofthisessay,tobeinthesamephysiological,ordynamic,state.I call this condition homeotaxy or peer conformity. Homeotaxy and horizontal control To return to the squid and its array of coloured spots. The word homeotaxy (meaning ‘same arrangement’) has been coined to describe the unified state of thearraywhenunder‘horizontalcontrol’.Itmakesnoreferencetotheunifying mechanisms involved. Horizontal control of the skin in a squid that is fully differentiated asserts itselfparipassuwithremovaloftheverticalcontrolexercisedbythebrainand its motor neurones. The condition (H -1 ) resembles that of a denervated gut o n or bladder and isachieved experimentallyby severing thenervesupplying one sideoftheanimal6,whereupontheautonomousactivityofdifferentensembles, lyingatlowerlevelsthanH ,begintosurface7.Twodaysaftertheoperation,a o standing wave of myogenic (muscle-generated) darkening shows on the oper- atedside–apparentlytriggeredbysomeresidualinfluenceofnerves(Figure 3). Days later, when the cut nerve has completely degenerated, the picture has changed. Chromatophores remain either quiescent (muscles relaxed, hyper- polarized and therefore minimum colours) or they flip from this baseline conditionintoanactivestatecarryingtheirpeerswiththem:notablyintheform of fast (1 cm/s) waves of coloured twitches that run across the skin in various directions and with broken frequencies, propagating into all denervated regions, but not into neighbouring, intact, skin nor into parts where nerves persist. These phase waves are associated with transient reversing depolarisa- tions. Evidently members of the chromatophore community (ensemble) are now tightly entrained physiologically. Examined closely, there are seen to be several ensembles. Large brown spots, likewise small yellow spots, may all be relaxed, and intermediate-sized red spots all be tonically expanded (muscles contracted) as is the case in Figure 3; fast phase waves may be purely red/ brown. From time to time the standing wave of darkening is abolished by a slow wave of relaxation (wave of hyperpolarization?), only to return many minutes later. Thehomeotaxyobservedinthisexamplehasadevelopmentalsignature.Each ofthedifferentresting-sizeandcolourclassesofchromatophoreorganisalsoan age class. Each behaves co-operatively as a separate matrix, or network, and sometimes entrains other classes or networks. Put more formally, connectivity betweenmembersofagivencolour/size/ageclassofchromatophoreiscloseto 6 Thiscutsnervefibresofffromtheircentralcellbodies.Unlikesomeoftheotherexamplesquoted (uterus,bladder,gut),theskinofsquidshasnoindependentperipheralnetworkofnerves. 7 In species less richly innervated than the one illustrated here (Loligo vulgaris) autonomous activityoftheensemblesoccurswithoutsurgicalintervention. 196 Figure3. Myogeniccontrol.Thenervesupplyingthenearsideofthebodyofthissquid(Loligo vulgaris)wascuttwodaysbeforethephotographwastaken.Contractionsofthousandsofmuscle fibres on the operated side (centre of photograph) are producing a standing wave of darkening amongstcertaincategoriesof red spots(chromatophores) justbelow the surface. Elsewhere, the skinisundernervouscontrol;chromatophoresarerelaxedandthereforenotvisible. 100%: i.e. obligatory – whence the name peer conformity. Between networks (corresponding to slightly different shades of blue in Figure 2) it is less than 100%.Cooperativeactivityceasesonlywhenthetissuedies8. TheonlyexcusefortakingalibertywithreadersofBiologyandPhilosophyto expoundthissomewhatremoteexampleofageneralphenomenon(myogenicity) isitsstrongillustrative(i.e.heuristic)appeal.Manymorefamiliarexamplesof muscular structures that behave myogenically temporarily freed from ‘higher’ controlaretobefoundintextbooksofhumanandofcomparativephysiology. Thebeatingoftheheart,thebirthofababy,thechangingsizeofthepupils,the erection of the penis (Christ 1997), the movements of the gut and bladder, all proceedbecausethecellsmakinguptheorganortissueconcernedarefunctional syncytia(orcoupledensembles)thatgeneratetheirownrhythms.Butinnoneof these examples of peripheralautomaticity9 can the experimenter see both the small-scaleandthelarge-scalepatternsofconnectivitywithintheensemblewith thedetailavailableinFigure 3.Noneofthemcombinethespatialandtemporal rangeandresolutionsuppliedbyvideo-recordingsoflivingsquids,whereevery chromatophoreisnaturallycolour-codedandsymbolizesadevelopmentalstage 8 For further details see Packard (2001) and http://www.gfai.de/www_open/perspg/g_heinz/bio- model/squids/squids.htm. Note that we are not here concerned with the temporal and other characteristicsofwaves,norwiththepacemakerandtriggeringconditions.Pacemakeractivityis nonethelessafundamentalpropertyofcoupledensembles. 9 SeeBozler(1948)(alsoHess1954)fortheusageofthetermautomaticityinthisessay. 197 in the history of the tissue. In other wholes (H ), or modified wholes (H -1 ), o o n suchinformationissimplymissing10. The following paragraphs will help the reader place the above in a wider context. 1. Inamachine,homeotaxyisensuredbysuchthingsastheconductanceand complianceofthematerialsused,aswellasbynutsandboltsandlinksof various kinds. In living systems it is associated with electrical conduction within layers, with conformational spread (see next bullet), with intercel- lular communication through gap junctions (GJIC) and with intercellular ligands and their receptors. Gap junctions, of several molecular forms, exist in all tissues and, when open, provide low resistance channels for transfer of charge, clamping members of the ensemble to a common po- tential (see Loewenstein 1999) and for passage of molecules between cells permitting their ‘metabolic cooperation’ (Sheridan and Atkinson 1985). 2. Atthemacromolecular(andsubcellular)level,allofthoselivingprocesses which depend on a dynamic change in the conformation of allosteric proteins (DNA replication, muscle contraction, energy production, etc.) are now believed to do so through conformational spread (CS) (Bray and Duke 2004). CS is the, domino-like, free-energy based propagation of the conformationalchangeintheseproteins.It‘coordinatestheactionoflarge numbersofproteinsinextendedcomplexes’, withincells,ontheirsurfaces and presumably also in their surroundings. Cited examples (with assem- bled evidence) of this fundamental form of homeotaxy, are the coupled gating of ryanodine calcium channels in the endoplasmic reticulum of heart muscle (where the change in allosteric state spreads between protein units or sub-units of the receptor lattice), and the switch from one qua- ternaryconformationtoanotherthatpropagatesalonglinearpolymersto bring about the characteristic bending of a bacterial flagellum11. 3. The classical example of automaticity and unified action has always been the vertebrate heart, which can beat independently of the vertical controls that modulate heartrate. It isusually regarded as a specialcase, however, becausethemembranesofthemusclecellsformingitswallsareincomplete andconstituteastructural(aswellasfunctional)syncytiumthroughwhich a wave of excitation can spread in all directions. 4. In sponges Leys et al. 1999) and other lower animals (Mackie 1965; Mackie and Passano 1968) the coordination is through electrically conducting epithelia. An early metazoan with dramatic defence behaviour 10 Notexcepting,fromthisstatement,thesophisticatedandhighlysuccessfulimagingofnormal activityinthebrainwiththeclinicaltoolknownasfunctionalmagneticresonanceimaging(fMRI). ItstemporalandspatialresolutionlimitsarediscussedbyLogothetis(2002). 11 ‘Toatruebeliever,therecanbenoplainerdemonstrationofconformationalspreadthanthe beatingofciliaandflagella’(BrayandDuke2004:60).CSisa‘theoreticalconstruct’,combining manydifferentlinesofevidence,forwhichtherecanbe‘noproof……thatwillsatisfyallskeptics’ sinceactualvisualisationwouldrequireamicrosecondtimescaleandspatialresolutionlessthan 1nmBrayandDuke2004:55). 198 Figure4. Non-neural coordination of the whole (H ). The hydromedusan Sarsia tubulosa (a) in o fishingand(b)defensiveposeaftercollisionwithanobject.Enlargementof(b)(right)showsroutes of electrical waves (arrows) from point of collision through ectoderm (1) and endoderm (2) to epithelio-muscularlayersthatservetoretracttheumbrellamargin(3),tentacles(4)andmanubrium (5).Responsetime200ms.(AdaptedfromMackieetal.1967). coordinatedbysuchnon-neuralmechanismisthejellyfish,Sarsiatubulosa. On colliding with an object this essentially diploblastic animal completely involutes (Figure 4). The action is the result of electrical waves spreading alongtheectodermallayers(withoccasionaljumpsintonervouslayers)to the endodermal and executive epithelio-muscular layers through the mesogloea (Mackie et al. 1967). In many other coelenterates similar behaviour is coordinated by the diffuse neural network12. In early chor- dates,epithelialconductioncoordinatesthemembersofachainofsalps.It is found both in early (Bone and Mackie 1975; Mackie and Bone 1976, 1977) and in later chordates (Roberts 1969) running over the surface as skin impulses which then interact with the nervous system. Epithelial conduction (electrical and/or calcium waves) survives in advanced verte- brates, for instance in the lung alveolar epithelium (Boitano et al. 1992). 12 Evidentlytherewere originallytwo systemsfor horizontal control,the second(neural) oper- ating with synapses, became concentrated and centralised to serve vertical control (cf. Mackie 1970).