Respiratory Gas Exchange at Lungs, Gills and Tissues: Mechanisms and Adjustments

The aim of this report is to outline the mechanisms and the adjustments of gas exchange and transport systems in vertebrates, using models suitable for quantitative analysis.

A generalized and highly simplified scheme of the gas exchange and transport system is depicted in Fig. 1. The elements of the gas transport chain are ventilation, medium/blood diffusion, perfusion (circulation), blood/tissue diffusion, and oxidative tissue metabolism. There is a gradient from inspired gas to tissue cells and an oppositely directed gradient, both consisting of and steps, which reflect the resistances to O₂ and CO₂ transfer of the individual links of the gas transport chain.

The factors determining the individual partial pressure steps are analysed using the simplest possible models. First, the external gas exchange occurring in various types of gas exchange organs, with air or water as external respiratory medium, is considered. Then the internal (tissue) gas exchange is discussed, using simple models for interaction of blood flow and diffusion. Finally some physiologically important phenomena which occur during transition from one steady state to another (from rest to exercise, from normoxia to hypoxia) are analysed.

A complete coverage of the pertinent literature is impossible in this brief account of a wide research area. Therefore, only a very restricted, personally biassed, reference list is appended. More detailed references to the literature can be found in recent reviews by White (1978), Wood and Lenfant (1979a) and Dejours (1981).

In this section, emphasis will be placed on the comparative aspects of the function of gas exchange organs in vertebrates. More detailed accounts have been published elsewhere (Piiper & Scheid, 1977; Piiper & Scheid, 1981).

In gas exchange organs of vertebrates the external respiratory medium (air or water) is brought into intimate contact with the internal gas transport medium (blood).

In comparing air and water breathing the capacitance coefficients of the medium, β_in for CO₂ and O₂, are the decisive factors. For air (gas phase), β_m is equal for all (ideal) gases. For water, β for O₂ and CO₂ are markedly different, the ratio being about 30 (the exact figure is dependent on temperature, salinity and buffering). The ratio β (water)/ β (gas) is close to unity for CO₂, but only about 0·033 for O₂. These relationships have the following consequences for external gas exchange (Rahn, 1966; Dejours et al. 1970; Dejours, 1972).

1. To achieve the same O₂ uptake (more precisely, the same G_vent for O₂), water breathers must ventilate much more than air breathers.

2. Since is about equal for water and air, the increased ventilation with water breathing means an equally increased G_vent for CO₂, whereby is markedly diminished in expired water and in arterial blood. This is the reason for the large discrepancy in arterial between mammals (about 40 torr) and fish (about 1–4 torr).

3. According to the Henderson-Hasselbalch equation

For the ideal models, β is the only significant property of the medium with respect to gas transfer. In real gas exchange organs, however, a number of other properties are important. These are :

1. Diffusion properties, characterized by the diffusion coefficient, d, or Krogh’s diffusion constant, K ( = d. α), determine the development of partial pressure gradients within the medium (interlamellar water in fish gills ; surrounding air or water in skin breathing; ‘stratification’ in mammalian lungs).

2. Viscosity, η is a major determinant of the mechanical resistance to respiratory medium flow, both with air and water breathing.

3. Density, ρ, determines the inertia of the medium and is, therefore, of importance in respiratory flow varying with time within the respiratory cycle.

Since K is much smaller, and η and ρ are much higher in water than in air, water breathing is generally more costly, (i.e. requiress more energy per volume of medium respired, than does air breathing).

In both skin and lungs gas exchange takes place between a homogeneous medium and blood flowing through a dense capillary network. There are, however, two important differences :

1. In lungs the medium is alveolar gas, the composition of which differs from atmospheric air according to transfer rates and G_vent. The alveolar-capillary barrier is very thin and the surface area is large. Therefore G_diff is high, in first approximation not limiting O₂ uptake and CO₂ output.

2. In amphibian skin the medium is atmospheric air or water. The cutaneous capillary plexus is located beneath the epithelium which has a considerable thickness (to provide protection against mechanical injury and desiccation). Therefore, G_diff is low whereas G_vent formally approaches infinity. According to Fick’s law of diffusion,

In applying the model to the real situation of blood capillaries in gas exchange organs, a number of complicating features must be taken into consideration.

1. Part of the resistance to diffusion is located within the blood (i.e. in the plasma, red cell membrane and within the red cells).

2. Analogously, the medium, particularly when it is water, may offer a resistance to diffusion.

3. The physico-chemical processes associated with gas exchange in the blood (e.g, combination of O₂ with haemoglobin, dehydration of carbonic acid (bicarbonate) to CO₂, exchange of bicarbonate and chloride ions between red cells and plasma) may be rate-limiting.

Values of D derived from physiological measurements contain all these resistances to O₂ or CO₂ transfer. Because of reaction limitation, the less specific term ‘transfer factor’ may be preferable to the conventional term ‘diffusing capacity’.

Of decisive importance, for the transport of both O₂ and CO₂ by blood, is the increase of the ‘effective solubility’ (measured by the capacitance coefficient (β_b)) by reversible chemical combination as O₂-haemoglobin and as bicarbonate, respectively.

The capacitance coefficient β_b for O₂ is largely proportional to the concentration of haemoglobin (O₂ capacity) but varies with according to the shape of the O₂ dissociation curve (= plot of O₂ saturation of blood against ). The shape of the O₂ dissociation curve in turn is determined by the chemical structure of haemoglobin, temperature, pH and ( = Bohr effect), and by the intraerythrocyte concentration of organic phosphates (adenosine triphosphate, guanosine triphosphate, 2,3-diphosphoglycerate, inositolpentaphosphate) and of other substances (Cl⁻, ⁠) functioning as specific regulators of O₂ affinity. The effects and mechanisms are analysed in several recent reviews (e.g. Bauer, 1974; Bartels & Baumann, 1977; Wood & Lenfant, 1979b).

The β_b value for CO₂ results mainly from reversible formation of bicarbonate with increasing by the buffering action of haemoglobin, plasma proteins and phosphates. Effects are exerted by temperature, the acid-base status and the O₂ saturation of haemoglobin (= Haldane effect).

Both and depend upon the respective partial pressures, according to the slope of the blood dissociation curves. For perfusive transport it is sufficient to use the slope of the straight line crossing the dissociation curve at the arterial and venous values. For calculation of medium-blood transfer, however, particular step-by-step techniques may be required to account for the curvature (Bohr integration). In most instances is considerably higher than ⁠, and the range of variation of in blood (and in tissue) is much less than that of ⁠.

With the simultaneous circulatory transport of O₂ and CO₂ in opposite directions, both and are increased by the Bohr and Haldane effects, respectively. In hypoxia and in exercise is increased due to lowering of mean blood ⁠. This property provides an automatic adjustment of G_pert to the challenged O₂ transport system.

The functional properties of gas exchange organs of vertebrates – gills, skin and lungs – can be described in terms of four models illustrated in Fig. 3 (Piiper & Scheid, 1972, 1975).

The rows of secondary lamellae carried by the gill filaments form a fine sieve for respiratory water. Gas exchange takes place in the blood lacunae of the secondary lamellae which receives venous blood from the ventral aorta and whose arterialized outflow is into the arterial system. The anatomical arrangement is such that water and blood flows are in opposite directions (counter-current model).

Skin breathing is important in all extant amphibians being the only means of gas exchange in those slamanders (terrestrial and aquatic) which possess neither lungs nor gills. Gas exchange takes place in the dense subepithelial capillary network, the inflow to which is in part from the arterial system, in part from a branch of the pulmonary arch carrying venous blood. The oxygenated cutaneous blood flows into the venous system. This is in contrast to the arrangement of pulmonary outflow in tetrapods and lungfish which allows (complete or partial) separation of oxygenated from venous blood.

The lungs are formed by a number of parabronchi (or tertiary bronchi), in parallel arrangement, most of which connect the mediodorsal secondary bronchi with the medioventral secondary bronchi. Air passes through the parabronchi, both during inspiration and expiration, in the major part of the lungs unidirectionally, in a smaller part (neopulmo) bidirectionally. Gas exchange takes place in the peri-parabronchial tissue consisting of an interwoven air capillary and blood capillary network. The simplest adequate model for gas transfer in avian lungs is the serial multi-capillary or cross-current model (Scheid & Piiper, 1972; Scheid, 1979).

The airways of mammalian lungs constitute a highly branching system of several orders of bronchi, leading to bronchi carrying alveoli and lastly to alveolar ducts the walls of which are entirely made up by alveoli surrounded by a blood capillary network. Since the renewal fraction of alveolar gas per breath is small, the variations in the composition of alveolar gas are relatively small, and for a simplified analysis a constant composition of alveolar gas may be assumed (ventilated pool model).

The same functional model may be used for the lungs of amphibians and some reptiles. However, in lungs of other reptiles there is a marked tendency to develop non-alveolated regions resembling avian air sacs (Duncker, 1978), which requires a modified cross-current model.

The decisive parameter for the overall gas exchange performance of a gas exchange organ, or its model, is the total conductance, ⁠. A comparison of G_tot for the various ipodels yields the picture shown in Fig. 4. The following decreasing order of gas exchange efficiency is obtained for the models (the * infinite pool’ model is a limiting case resulting from all models when G_vent approaches infinity) :

The gas exchange efficacy in real gas exchange organs is considerably less than in idealized models due to functional inhomogeneities, dead space, vascular shunts and other factors (Piiper & Scheid, 1977).

The reason for the adoption of a certain type of gas exchange organ by the different vertebrate groups cannot be sought in the gas exchange requirements alone. Nevertheless, the following may be stated.

1. As water-breathing is energetically costly (see above), it is important for fishes to use the scarce dissolved O₂ as effectively as possible. This is achieved by the counter-current strategy.

2. Birds, many of which are capable of sustained flight at high altitudes, require particularly efficient gas exchange organs. However, the higher tolerance of hypoxia by birds as compared to mammals probably results from other, unknown, factors besides the efficient cross-current type gas/blood arrangement in lungs.

Physiologically important adjustments are made (1) to increased metabolism, (2) to changes in the respiratory medium, and (3) to disturbances by disease.

1. In exercise the conductances are increased. In man and mammals, G_vent (i.e. ⁠) increases proportionally to and ⁠, by increasing both tidal volume and breathing frequency. G_perf also rises, due to increased cardiac output, ⁠, brought about by increased cardiac frequency and stroke volume, but also due to increase of of blood produced by lowering of venous and by increase of blood haematocrit. The extent of increase of G_diff is unclear, and its extent is probably rather limited. Therefore, the role of diffusion limitation is expected to increase in exercise. The adjustments seem to be similar in birds (cf. Fedde, 1976; Bouverot, 1978) and in fish (cf. Randall, 1970; Johansen, 1971).

2. Similar adaptive changes occur during environmental hypoxia. However, since only the O₂ availability is reduced, the hyperventilation must lead to hypocapnia, which may be compensated by adjustment of the bicarbonate concentration in blood. In environmental hypercapnia, increased G_vent alleviates the acidosis. In water-breathing animals, however, even a large increase in G_vent would have little effect, and the main adjustment observed is increase of blood bicarbonate leading to compensation of the respiratory acidosis (Heisler, 1980).

3. The compensations for anatomical and functional derangements in the respiratory gas transport system in various diseases are not only of interest for clinical physiology, but also contribute to the understanding of the basic mechanisms involved. Examples of such compensatory mechanisms include increased ventilation of lungs with impaired gas exchange function, renal compensation of respiratory acidosis due to disturbed lung function, increased cardiac output in anaemia, and hypoxic vasoconstriction in lung regions with airway obstruction.

The quantitative analysis of O₂ and CO₂ exchange in tissues is less advanced than that in external gas exchange organs, due mainly to the experimental difficulties involved in determining and in tissues and also due to problems of adequate modelling (cf. Tenney, 1974; Grünewald & Sowa, 1977).

In tissue respiration usually only O₂ is considered. The main reason for this derives from the existence of an absolute limit for tissue ⁠, whereas no such limit exists for ⁠. Moreover, all gradients are small, because of high and high K for CO₂ (equation 8).

Although Krogh’s cylinder is homogeneous with respect to diffusivity and solubility of O₂, most resistance to diffusion is located near the capillary, because here the O₂ flux density is highest. Thus no great inaccuracy is introduced when the model is simplified by completely separating the resistance to O₂ diffusion from the O₂ consuming tissue compartment, which in the Krogh model is predominantly represented by the more peripheral regions of the cylinder. Furthermore, longitudinal diffusion, which is not permitted in Krogh’s model, would reduce the longitudinal O₂ gradient. Moreover, when adjacent parallel capillaries are not perfectly aligned, but overlap, and when their blood flow is in part counter-current, the mean tissue is expected to show less pronounced longitudinal O₂ gradients.

Both models may be used to investigate the adaptive physiological changes in hypoxia (reduction of arterial ⁠) and in activity (increased tissue O₂ consumption) which maintain tissue at an adequate level for oxidative metabolic demands. Clearly the adaptive changes must affect either the circulatory O₂ supply by the blood (specific tissue blood flow, ⁠, and the capacitance coefficient of blood for O₂, β_b) or the blood-tissue diffusion characteristics, as quantified in terms of the specific diffusing capacity (d or d′).

Increase in tissue blood flow, ⁠, is an effective means of increasing tissue O₂ supply. At high blood flows or, more precisely, at high or values a further increase in q becomes ineffective, because the O₂ supply is then mainly limited by diffusion ; this behaviour is evident from eqs. (13) and (15).

Increasing the capacitance coefficient, β_b, has formally the same effect as increase of ⁠. β_b may be increased by increase of haematocrit or by change of the slope of the O₂ saturation − relationship, which is increased in hypoxia. Thus the shape of the blood O₂ dissociation curve provides an automatic adjustment of perfusive O₂ conductance in arterial hypoxia, as well as in venous hypoxia occurring in exercise with increased utilization of blood oxygen.

Physiologically there are two ways to improve diffusion conditions for O₂ in tissues.

(a) An increase of the capillary diameter or radius (r_c) increases d′ in equation (10) and d in equation (14) by increasing the effective surface area available for diffusion.

(b) Opening of closed, unperfused, capillaries increases the capillary density and thereby reduces the effective radius of the O₂ supply cylinder (r₀ in equation (10)) and the effective diffusion distance (x in equation (14)).

The effectiveness of these measures to increase O₂ supply is high when the ratio or is small, meaning predominant diffusion limitation of blood-tissue O₂ transfer. With high values of these ratios increased perfusion would be more effective since in these conditions O₂ supply is preponderantly perfusion-limited.

In real tissues the simple models may become inadequate for many reasons, two of which will be briefly addressed.

In real tissues, even with essentially parallel arrangement of capillaries, like in muscle, complications arise when in adjacent capillaries the arterial and the venous ends are at different levels and the directions of flow are counter-current (cf. Grünewald & Sowa, 1977). The counter-current arrangement leads to a truncated cone model of O₂ supply, and appears to provide more efficient O₂ supply than a co-current arrangement. However, with high diffusive conductance (dense capillary network) shunting of O₂ from the arterial end of one capillary into the venous end of another capillary (or of the same loop-shaped capillary) will occur, whereby the O₂ transport efficiency is decreased.

There is experimental evidence for a rather inhomogeneous distribution of blood flow to volume in apparently homogeneous muscles (e.g. Sparks & Mohrman, 1977). The efficiency of O₂ supply in a system of parallel capillary units with unequal blood flow is reduced because it reaches critical O₂ supply conditions at lower O₂ requirement or at higher total blood flow than in a homogeneously perfused system.

For O₂ supply, it is the distribution of and d (or d′) relative to which is the important variable, thus one has to consider the inhomogeneity’. It would be interesting to know if in exercising muscle the inhomogeneity is reduced by local micro-circulatory control mechanisms (adjustment of blood flow and capillary density to the local metabolic level).

Steady state is an ideal condition, appreciated by physiologists, but never fully achieved in reality. Gas transport clearly varies within a muscle fibre twitch, a cardiac cycle, a respiratory cycle, activity–rest cycle, cyclic changes in environment etc. There is particular interest in the last-mentioned changes which have a longer period and, therefore, can be analysed in terms of transition from one steady state to another.

The delay (finite kinetics) may result from two categories of factors.

(2) Furthermore, the delay may be due to the slowness of adaptive changes in the gas transport system after an abrupt change of a variable. An important example is the delayed increase of O₂ uptake after onset of exercise of constant power. The cause is the time requirement of increase in ventilation, cardiac output, muscle blood flow and diffusing conditions in the muscle (Cerretelli et al. 1980; di Prampero, 1981).

An important consequence of delayed increase of after onset of exercise is the O_a debt (or O₂ deficit) (Fig. 6). Assuming constant efficiency of oxidative metabolism, the following amount of O₂, M, is ‘missing’ from the balance :

The O₂ debt and its energy equivalent are attributed to several mechanisms (Fig. 7) :

1. O₂ stores (mainly tissue and venous blood),

2. energy gained from hydrolysis of high-energy phosphates (ATP and creatine phosphate), and

3. energy gained from anaerobic glycolysis, leading to accumulation of lactate.

The common denominator for these changes is energy release, oxidative for (1), anoxidative for (2) and (3). The involved energy equivalences have been determined in vivo (cf. Piiper et al. 1980a). It has been shown that the 0₂ debt incurred after onset of light or medium exercise is energetically explained by hydrolysis of high energy phosphates, mainly phosphocreatine (Piiper et al. 1968). At least part of the 0₂ debt repayment is required for resynthesis of phosphocreatine to the resting level (Piiper & Spiller, 1970).

The relationship between the kinetics of O₂ uptake after onset of exercise with the changes in high-energy phosphates can be considered from two points of view:

(a) A certain metabolic level is associated with a certain degree of hydrolysis of high energy phosphates ; the energy released therefrom is utilized for mechanical work and therefore the adjustment of O₂ supply need not be instantaneous.

(b) The adjustments of O₂ supply are intrinsically slow, giving rise to an O₂ debt which has to be covered by splitting of high-energy phosphates.

In any case, the speed of the adjustments and the functional energy stores must be interrelated in a manner to render possible rapid, but economical, energy release.

The O₂ debt associated with exercise of vertebrate muscles is usually repaid during the recovery. After onset of hypoxia, however, in many lower vertebrates the O₂ uptake is reduced, and after return to normoxia there is little overshoot in O₂ uptake : this behaviour is called O₂ conformity, in contrast to O₂ regulation meaning O₂ consumption independent of O₂ supply (cf. Prosser, 1973).

In many cases the O₂ conformity appears not to be only a passive consequence of shortage of O₂ supply, but it should rather be interpreted as an adjustment to reduced O₂ supply. This certainly was the case in lungless salamanders subjected to deep hypoxia, since they showed recovery of initially increased lactate and decreased high energy phosphates during persisting hypoxia and reduced O₂ uptake (Gatz & Piiper, 1979). Similarly, the reduced oxidative metabolism during diving in habitually diving mammals is the result of specific circulatory and metabolic control mechanisms (Andersen, 1966).

Probably there are transitions from O₂-debt repaid fully (or even in excess), through O₂ debt repaid partially to ‘true’ reduced metabolic state. Their systematic and comparative study in lower vertebrates is expected to be rewarding.