| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
INVITED REVIEW
HIGHLIGHTED TOPICS
A Physiological Systems Approach to Human and Mammalian Thermoregulation
Little Garth, High Street, Harston, Cambridge, United Kingdom
 |
ABSTRACT |
|---|
 TOP ABSTRACT REFERENCES |
|---|
homeothermy; thermoregulation; reference signal set point; reciprocal crossing inhibition
To understand why there must be as least some degree of variability, consideration must first be given to the apparent fundamental raison d'etre of a central nervous system (CNS). If its function were to be simply that of connecting particular sensors to particular response effectors, the inherent danger of concentrating the many specific sensor-to-effector pathways through a single structure would surely have forbidden such an evolutionary development. Throughout the evolutionary development of a CNS, any advantage gained by the concentration of the many sensor-to-effector pathways through a central structure had to outweigh the disadvantage of its vulnerability to damage. That advantage was surely that it facilitated the communications between sensor-to-effector pathways. It is now evident that the integration of the responses to an infinitely variable pattern of external and internal occurrences depends on a massive assembly of interpathways connections. These interconnections between pathways through the CNS render it unlikely that any particular response is solely the consequence of one particular stimulus. Hence there is the inevitability of some degree of variability in any stabilized quality or quantity of the internal environment.
Engineers will point out that, even in a singular regulatory process, to activate a corrective response, there must be some degree of divergence of the stabilized quality or quantity from the set level. Thus there is some variation in even the most precisely regulatory processes. This could apply equally to biological processes of homeostasis, but there is clearly more than that to the variability of the parameters of mammalian homeostasis, including homeothermy. Some of this is likely to be the consequence of the effects of activities in neural interconnections between sensor-to-effector pathways, which vary the magnitude of a particular response to a particular species of sensor.
The expectation must be that the stability of body temperature depends on the existence of some definable neuronal properties of thermosensors, of thermoresponse effectors, and of the interconnections between them through the CNS. It can be further surmised that the determination of the relationships between thermal disturbances and the thermocorrective responses will be influenced by other concurrent nonthermal disturbance and response activities via the CNS. It was the patterned change in the level at which body temperature is seemingly being maintained during infection that led Liebermeister (16) to describe the occurrence of fever as the resetting of a biological thermostat.
The essential difference between the so-called cold and warm sensors is that they have differing activity-temperature profiles. Within a limited and seemingly physiological range of temperature variation, the activities of "warm" sensors increase as temperature increases, whereas those of the "cold" sensors decrease as temperature increases. The effector responses to thermal change are separable into those that vary heat production within the body and those that vary the exchange of heat with the external environment. A likely basic sensor-to-effector relationship is expressible as in Fig. 1. The assumption in this neuronal model is that the warm sensor inputs and the cold sensor inputs are each separately summed to provide the two activating drives, the one to heat loss effectors, and the other to heat production effectors. With the universality of interpathway influences within the CNS, there could well be reciprocally crossing influences between these two sensor-to-effector pathways, as well as excitatory and inhibitory influences derived from other nonthermal pathways through the CNS. These are also indicated in Fig. 1A. A noteworthy point is that, whereas generally the diametric changes in the activities of the warm and cold sensors as local temperature is changed are overlapping, those of the heat production effectors (HP) and the evaporative heat loss effectors (HL) occur on either side of a null point or null range. The cause of this distinction between the input from thermosensors and the output to thermoregulatory correction effectors lies, presumably, within the CNS.
|
In the absence of any clear physiological evidence of the nature of the set-level determinant, James Hardy (11), who was probably influenced by the interest at the John B. Pierce Foundation Laboratory at New Haven, Connecticut, in the engineering of air-conditioned buildings, considered mammalian thermoregulation in terms of the regulatory principles employed in engineered physical systems. The most widely understood principle of the ways the temperature of a physical mass can be regulated is for a representation of the actual temperature to be compared with a fixed value representative of the required level of thermal stability. Any difference between the two values, referred to as the error, is used to determine the quality and/or the quantity of the corrective response. Hardy intended this as an analogy, and not that this is how the basic operational principle of mammalian homeothermy has to be. Being readily comprehendible, and in the absence of any alternative proposal, this analogy came to be considered as the most likely explanation rather than just an analogous one. Furthermore, the generality of the synaptic excitatory and inhibitory influences of one neuron on another made it easy to see how a neuronal equivalent to an engineered physical thermostat could have been created. An excitatory signal from a sensor neuron and a constant and constantly generated inhibitory influence of a central "reference" neuron could be acting on both sensor-to-effector pathways through the CNS. Either of the arrangements depicted in Fig. 1, B and C, would prevent the activity of signals from either the cold sensors or the warm sensors from influencing the appropriate response effector until the excitatory influence is greater than the interjected inhibitory influence. If, then, the intensity of the inhibitory influence is unaffected by temperature variation, or by other nonthermal influences, it could be the provider of a sustained thermoregulatory reference.
If it could be shown that in that part of the CNS known to be essential for normal thermal stability, there are continuously discharging neurons with an impulse frequency that is unaffected by local temperature change, the stable reference signal generator hypothesis would be greatly strengthened. Thus a considerable effort was put into the electrophysiological search for such neurons in the hypothalamic region of the brain. Many interesting and enlightening details emerged, especially of the occurrence, the locations, and the differential densities, of warm- and cold-sensitive neurons both centrally and peripherally and of the convergence of the afferent pathways from cutaneous, spinal cord, and hypothalamic thermosensors (for review of literature, see Ref. 15). The activities of many hypothalamic neurons are unaffected, or little affected, by local changes in temperature, but this could be because they are unrelated to the maintenance of homeothermy. Clear and unequivocal neuronal evidence supportive of the reference signal hypothesis has remained elusive, yet the notion of a stable reference signal with which the variable is being constantly compared remains a popular way of considering the physiology of homeothermy.
There was an intriguing intervention into central events in homeothermy when Feldberg and Myers (8) observed shifts in body temperature when what were then putative central neurotransmitter substances norepinephrine (NE) and 5-hydroxytryptamine (5-HT) were introduced into the cerebral ventricles of cats. NE caused a fall in body temperature, and 5-HT caused a rise. It was postulated, thereupon, that the body temperature set level in mammals could be invested in the relative extents of the release of two different endogenous transmitter substances within the hypothalamic region of the brain. Although both interesting and intriguing, this postulate was more brave than rational. Interneuronal communication generally involves the release of a chemical synaptic messenger in response to some triggering event. These observations could be evidence of endogenous neurotransmitters, the release of which influences the strength of the processes of heat production and heat loss. It is unlikely, however, that they could constitute the set-level determinant. That must surely lie not in transmitters per se but in some pattern of neuronal events that culminate in their release. A vast amount of subsequent research has shown that many naturally occurring substances, which could be synaptic transmitters or modulators, also cause changes in body temperature when similarly applied to the brain. Many of these have been pain-takingly catalogued in a series of publications in Neuroscience and Behavioral Reviews by Clark and coworkers between 1979 and 1986 (see Ref. 6 as a representative paper). Although the effects on body temperature may be consistent in a particular set of circumstances, the effects of these substances can vary between species, between sites of application, between the applied dosages, and between the prevailing external environments. Thus much of the synaptic involvement in homeostasis has still to be clarified. Sadly, so much diligent research has added little to the understanding of the homeothermic set-point determinant. As is mentioned below, however, some possible indication did emerge from these neurochemical studies.
In the course of this search for the "Holy Grail" of mammalian homeothermy, there was an emerging awareness of the fact that the set-level does not necessarily require a stable reference signal. Hammel, as I remember it, was clearly concerned about the growing acceptance of the stable reference signal generator as an all but proven requisite of homeostasis; and Hensel (13) commented that what is essential for the maintenance of body temperature at a set level is the provision of at least two input signals with different temperature coefficients. The comparison of a signal that varies with temperature with one that does not fits with this criterion, but is no more than a particular version of the principle. The differing activity-temperature profiles of signals derived from cold and warm sensors are also in compliance. Werner (20) concluded that there is no substantial evidence supportive of the notion that mammalian homeothermy depends on a stable signal generator.
The activity-temperature coefficient of the warm sensors and cold sensors can only be described loosely as being reciprocal. The technical difficulties of obtaining sustained activity recordings from single fibers while the local temperature is being varied means that results are relatively sparse. The recorded slopes of the activity responses to temperature changes differ between individual recordings. Furthermore, the immediate activity responses to temperature changes differ markedly from the sustained responses to sustained temperature (15). Thus there is more than an element of license in the suggestion that the mean activity-temperature profiles of cold sensors and of warm sensors are reciprocal over a physiologically meaningful range of temperature variation. There are, however, clear indications that some such relationship exists. This generalization allows the theoretical possibility that body temperature stabilization within a narrow range of variability is not dependent on the central involvement of temperature-insensitive "reference neurons," but on the different response coefficients of the cold and warm sensors to temperature changes, and on the influences these have on the heat production and heat loss effectors.
If that were to be so, there would be no need to hypothesize any additional central set determinant of the set level. Yet if there is no central modulation of signals passing from the thermosensor to the thermoregulatory effector pathways, it could be expected that heat production and heat loss effector activities would overlap much as do the sensor activities. In that case, thermal stability could still be achieved, but uneconomically, with thermoregulatory HP and HL activities in continuous concurrent competition. It is well established, however, that in humans and other mammals, HP by shivering, and HL by panting or sweating, are not concurrent events, but occur one side and the other of a null point or null zone. This separation of the thermoregulatory HP and HL activities could be provided by the convergence on the cold sensor to HP, and the warm sensor to HL, pathways through the CNS of a reference-signallike inhibitory influence (Fig. 1, B and C). Another possibility is that of reciprocal crossing inhibitory (RCI) influences of both sensor-to-effector pathways on each other. The weaker of the two activities would thus be nullified. This notion was discernible in the diagrammatic expression of the relationships between variations in environmental and deep body temperature on the thermoregulatory responses of humans by Wyndham and Atkins (21). The underlying principle of signal separation by reciprocal crossing inhibition was not stated explicitly, but this became evident when the diagram was redrawn in a neuronal format by Bligh (2). This is reproduced in Fig. 2A.
|
In none of the representations was there any clear statement of the possibility that the set-level determination could be vested in the activity-temperature profiles of the thermosensors. Nor was the postulated RCI between the sensor-to-effector pathways explicitly suggested as the means by which the HP and HL activities are caused to occur on the one side and the other of a null point or null range. Only later (3) was an attempt made to marry evidence and theory to provide an integrative consideration of how mammalian homeothermy could be effected. It was proposed that the response coefficients of the thermosensors, together with central RCI, is all that is strictly necessary to provide a basis for the stabilization of body temperature within a small range of variation. The convergence onto the thermosensor-to-thermocorrection pathways, of excitatory and inhibitory influences derived from other central nervous events, could account for variations in the set-level, or in the extent of a null range.
A difference between the models of Hammel (10) and Bligh (2) lies in the depiction of the central thermosensitivity. Whereas in Hammel's model the central thermosensors are interneurons on the pathways through the CNS from extracentral thermosensors, in Bligh's model they are depicted as primary thermosensors uninfluenced by any synaptic signal input. There is no a priori reason to favor the one possibility or the other, and no significance need be attached to this distinction. In both representations, there is RCI between the two principal thermosensor to thermoregulatory pathways, and there is no representation of a stable reference signal generator.
To validate the theoretical soundness of this assertion, Smullin (19) constructed a physical model in which two temperature sensors with overlapping reciprocal responses to temperature connected directly to HP and HL correction effectors with reciprocally crossing negative electrical influences of each sensor to effector pathway on the other (Fig. 3A). This construction effected the thermoregulation of a physical mass with a relationship between thermosensors and HP and HL response effectors that was essentially the same as that now being suggested as the means of achieving mammalian homeothermy. There was no set-point determinant other than the differing properties of the two sensors, and RCI between the two sensor-to-effector connections caused the HP and HL responses to occur on one side and the other of a temperature null point or null range. Additional convergent positive and negative inputs onto the sensor to effector pathways (Fig. 3B) were analogous to convergent influences on pathways through the CNS. Variations in these resulted in thermoregulatory variations that closely mimicked those of different species of mammals, or of one species in different circumstances, including those of fever and hibernation. Figure 3C is a neuronal translation of Fig. 3B. Although this physical model constituted a crucial verification of the theoretical validity of the alternative theory of mammalian homeothermy, it was apparently considered by referees to be outside the legitimate scope of thermal physiology, and it has remained unpublished except briefly.
|
One cannot now examine the simple structures of the ancestral nervous systems from which the mammalian CNS evolved, except by inference. Comparative studies of the simple ganglionic structures of contemporary invertebrates may, however, afford a clue. A study by Kandel (14) of the contraction and relaxation of muscular structures in an invertebrate organism, the marine snail Aplysia, indicated the likelihood of RCI within the ganglion that intervenes between the input from disturbance sensors and the output to the musculature of the siphon that can be extended and contracted. In a somewhat more developed invertebrate, the horseshoe crab, Hartline and Ratcliff (12) found evidence of RCI between the centripetal pathways from the ommatidia of the compound eye. RCI seemed to nullify whichever neural drive was the weaker. This, it seems, enhances contrast in the visual fields. In his discussion of RCI, Dowling (7) records how, in the mammalian eye, retinal RCI between the centripetal pathways of signal flow from photoreceptors could facilitate the enhanced interface contrast of areas of differing light intensity in the visual field. It is now well established that in the mammalian retina both the horizontal cells, which lie nearer to the photoreceptors, and the amacrine cells, which lie nearer to the ganglion cells and the optic nerves, exert these crossing inhibitory influences. These moderate the patterns of activity in the centripetal pathways through the bipolar cells to the ganglion cells. The horizontal cells, in particular, are now believed to play a crucial role in the perceived enhancement of the contrast at the interfaces of areas of different light intensity in the visual field. Dowling adds the comment that not only can RCI be recognized in all visuals systems of both invertebrate and vertebrate organisms but also it can be seen to exist and operate in other sensory systems as well.
RCI between two nerve pathways, by which opposing musculatures are caused to contract, was first recognized more than a century ago by Sherrington (18). This, he averred, occurs in the control of the extensor and flexor musculature of limbs, such that they do not contract concurrently, and likewise in the control of movements of the eyeball. Although RCI has long been accepted as a generality in the operation of opposing motor muscles, it seems not to have been given much consideration in the control of other paired effectors with opposing actions. In the text From Neuron to Brain (17), there is no mention of RCI beyond a reference to Sherrington (18). There is, however, at least some reason to suspect RCI influences between the neural pathways that activate the inspiratory and expiratory musculatures and that these facilitate the absence of an overlap of inspiration and expiration (3).
Although the evidence of RCI is too fragmentary to indicate the full extent of its role in the central modulation of relationships between disturbances and responses, there is just sufficient evidence to allow the suggestion that this particular usage of the massive interactions between pathways through the CNS is probably far more general than is currently supposed. It is worthy of recall that with the embryological kinship between retinal and central nervous tissues, the prominence of RCI in the mammalian retina could have come with the evolution of the retina. If so, that would suggest most strongly that RCI has been a functionally crucial feature of neuronal organization through eons of evolutionary progression.
Considerations of the nature of the mechanism that underlies mammalian homeothermy might well have been somewhat different had it been recognized 1) that the abiding feature of systems of regulation is not the comparison of a variable with a constant but is the interplay of two variables with different response coefficients; 2) that RCI is an almost inevitable feature of the connections between pathways through the central nervous systems by which functional integration is effected; and 3) that those two, together, could provide the basis of separating opposing effector functions on one side and the other of a null point. This trinity of circumstances on which mammalian homeothermy could depend, remains more hypothetical than proven. Its possibility, however, raises the further possibility that similar neuronal features could underlie the management of other stabilized conditions of the internal environment.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. A. Romanovsky Thermoregulation: some concepts have changed. Functional architecture of the thermoregulatory system Am J Physiol Regulatory Integrative Comp Physiol, January 1, 2007; 292(1): R37 - R46. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
