Physiological Discontinuity in An Epithelium with An Apparently Uniform Structure

It is often assumed that in an epithelium made up of cells of uniform structural appearance each cell contributes equally to the performance of that epithelium. This paper describes work on the Malpighian tubule of the bloodsucking insect Rhodnius prolixus where such an assumption turns out to be totally incorrect.

Larvae of Rhodnius take very large meals of blood. This blood is hypo-osmotic to the haemolymph of the insect by some 50 m-osmol (Maddrell & Phillips, 1975). To avoid dilution of its haemolymph and, in addition, to reduce its size and to concentrate the nutritious part of the meal, Rhodnius rapidly excretes a large volume of sodium chloride-based fluid which is some 120 m-osmol hypo-osmotic to its haemolymph. This leaves the osmotic concentration of the haemolymph unchanged at about 370 m-osmol with the contents of the gut now iso-osmotic to it (Table 1). The excretion of hypo-osmotic fluid is carried out by the Malpighian tubules of the insect. As the first step in this, the upper or distal regions of the Malpighian tubules secrete fluid iso-osmotic to the haemolymph and containing both sodium and potassium chloride. This fluid then passes through the lower lengths of the Malpighian tubules where reabsorption occurs of a hyper-osmotic potassium chloride solution (1700 m-osmol) leaving a K-poor hypo-osmotic urine to be eliminated (Maddrell & Phillips, 1975). The process is summarized in Fig. 1.

The structure of the lower Malpighian tubule was first examined by Wigglesworth (1931), who showed, with light microscopy, that it was constant in structure throughout its length, apart from a small region just above the terminal ampulla that marks the junction of the tubule with the rectum. This preterminal region with its thicker wall and reduced lumen is easily seen in fresh material; it extends 0·5–1 mm above the ampulla (see fig. 1 c in Wigglesworth’s (1931) paper). The experiments described in this paper did not include this region and were all done on the apparently uniform part of the lower tubule which is about 20 mm long in the fifth stage insect.

The ultrastructure of the lower tubule was investigated by Wigglesworth & Salpeter (1962) and was compared with that of the upper fluid-secreting part of the tubule. Appropriately enough they found that the two parts of the tubule were very distinctly different in ultrastructure. The cells of the upper tubule all appeared similar to each other. No regional differences in physiological performance have been found in this part of the tubule (Maddrell, 1969). The lower Malpighian tubule showed no evident regionalization of structure and appeared to consist of cells of only one structural type. It turns out however that this apparent uniformity hides a dramatic discontinuity in physiological performance.

Fifth-stage larvae of Rhodnius prolixus from a laboratory culture maintained at 28 °C were used in all the experiments. The insects were used 1–2 weeks after the moult from the fourth stage.

The experimental techniques used were similar to those developed by Maddrell & Phillips (1975) for the earlier investigation of the lower Malpighian tubule and their paper should be consulted for full details. Basically, isolated Malpighian tubules were investigated while bathed in drops of appropriate saline solutions placed in a bath of liquid paraffin (mineral oil). The lower tubule secretes no fluid; so to provide particular conditions in the lumen, saline of known composition was perfused through the lumen from a glass cannula pushed through the tubule wall. Perfusion was driven by a motor-operated syringe (100 μl, Hamilton). Osmotic concentrations were measured cryoscopically or were estimated from conductivity measurements. Radioactive chemicals were obtained from the Radiochemical Centre, Amersham. Measurements of radioactivity were made using standard scintillation counting techniques in conjunction with an Intertechnique ABAC SL 40 scintillation counter.

All values are quoted as mean ± s.e. (number of observations).

In vivo at 22 °C, fluid flows through the lower tubule at 150 nl min⁻¹ and is subject to reabsorption for only 30 s and yet in this time its osmotic concentration falls from 370 to 250 m-osmol and its potassium concentration from about 80 to less than 5 mm. In vitro the osmotic concentration and potassium level of fluid perfused through the tubule can change at rates as high as those observed in vivo (Maddrell & Phillips, 1975). In view of the speed of this process it seemed likely that reabsorption must begin at the upper end of the lower tubule and continue along its whole length. In an attempt to test this, some initial experiments were done with the lower half of the lower tubule, i.e. that part of the lower tubule nearest to the hindgut and farthest from the upper Malpighian tubule (Fig. 1). Fluid iso-osmotic with the bathing fluid (standard saline, 342 m-osmol) but containing 90 mm K (and 61·6 mm Na) was perfused through the lower halves of lower Malpighian tubules stimulated with 10⁻⁴M 5-HT. The emergent fluid contained 22· 7 ± 1· 7 mm K (n = 10) and had an osmotic concentration of 240 ± 4 m-osmol (n = 5). This result shows that the lower half of the lower tubule can decrease both the potassium and osmotic concentration of K-rich fluid flowing through it. It seemed then that reabsorption of KC1 was likely to occur along the whole length of the lower tubule. In fact, as we shall see, this is not the case.

That the picture was not so simple was shown in an experiment in which K-rich fluid (Na = 14·5 mm, K = 137·1 mm; 342 m-osmol) was perfused through whole lower tubules from their upper ends at a rate of 75 nl min⁻¹ while they were bathed in a hypo-osmotic fluid containing 5-HT (Na = 101 mm, K = 5 mm; 210 m-osmol). Fluid emerged from the lower ends of these tubules with lowered osmotic concentration as expected (113·9 ± 8·2 m-osmol, n = 5), but the rate of emergence of fluid was 112·0 ± 3·6 nl min^-1(n = 5). That is, there had been a net movement of water into the tubule in spite of the fact that the fluid emerging from its lower end was considerably hypo-osmotic to the bathing fluid, an osmotic gradient which ought, if anything, to have led to a movement of water out of the tubules.

The experiment which led to the explanation of these curious results is depicted in Fig. 2. K-rich fluid was perfused through the entire length of the lower tubule. A 50 μl bathing drop containing 5-HT, 5 mm K and having an osmotic concentration 60 m-osmol lower than the perfused fluid was positioned either near the upper end or near the lower end of the lower tubule. With the bathing drop near the upper end the fluid emerging from the tubule was close to iso-osmotic to the bathing fluid, and from the elevated rate of its emergence, water had evidently entered the lumen of the tubule from the surrounding drop. By contrast, when the bathing drop was set close to the lower end of the tubule, the rate of fluid emergence was scarcely changed – it was slightly lower than the rate of perfusion – but the osmotic concentration of the fluid had been greatly reduced to less than 100 m-osmol. Four further such experiments gave very similar results. These experiments show a clear-cut difference in the ability of different parts of the lower tubule to reduce the osmotic concentration of the K-rich fluid passing along its lumen. The upper part is not able, by reabsorbing solute, to reduce the osmotic concentration of K-rich fluid perfused through it but it does allow passive water movements in response to an osmotic gradient. That it does not absorb solute to any significant extent follows from comparison of the rates of perfusion and emergence. In the case shown in Fig. 2, 75 nl min⁻¹ of fluid at 340 m-osmol entered the tubule and 93 nl min⁻¹ of fluid at 280 m-osmol left it, i.e. 75 × 340 = 25500 p-osmol min⁻¹ entered the tubule and 93 × 280 = 26040 p-osmol min⁻¹ left it. Plainly the solute content of the perfused fluid had changed very little during its passage through the tubule. The lower region by contrast is able to absorb solute and produce an osmotic gradient, which of course requires that its wall has a relatively low osmotic permeability. The extent to which solute was absorbed can again be calculated from the rates of perfusion and emergence of fluid and the osmotic concentrations. In the case shown in Fig. 2, solute entered the tubule at 25 500 p-osmol min^-1 but only 73 × 95 = 6935 p-osmol min^-1 left its lower end. Evidently the difference, 18565 p-osmol min⁻¹, was absorbed by the tubule into the bathing drop.

That the upper part of the lower tubule has a high osmotic permeability was shown in a series of experiments in which fluids of varying osmotic concentrations were perfused through this part of the tubule and the rate of emergence and osmotic concentration (estimated from conductivity measurements) of the emergent fluid were determined. The results show that fluid perfused through the upper part of the lower tubule rapidly comes close to osmotic equilibrium with the fluid surrounding it (Table 2). The results also show clearly that there is virtually no change in the solute content of the perfused fluid and that 5-HT has no obvious effects.

From the results described earlier (p. 136), it is evident that the lower part of the lower tubule, when stimulated by 5-HT, can set up and maintain a large osmotic gradient across its wall. Can it maintain such a gradient in the absence of 5-HT? In other words, is its osmotic permeability dependent on its being stimulated? To test this, fluid of low osmotic concentration, 140 m-osmol, containing 2·5 mm K, was perfused at 75 nl min^-1 through the lower tubule arranged so that a 3 mm length very close to its lower end was bathed in a fluid of 280 m-osmol and containing 5 mm K, but with no 5-HT present. During the first 30 min the emerging fluid had an average osmotic concentration of 155 m-osmol and collected at a rate of 68 nl min⁻¹. After the addition of 10⁻⁴ M 5-HT to the bathing fluid, the osmotic concentration of the emerging fluid slowly rose during the next 20 min to 175 m-osmol and the rate of collection of fluid declined to 61 nl min⁻¹. Two further experiments gave similar results. Evidently the lower region of the lower tubule has a low osmotic permeability which is not dependent on stimulation.

So far then it is clear that regions close to the upper end of the lower tubule are osmotically permeable and do not absorb solute (potassium chloride) from the lumen while the region close to the lower end of the tubule has a low osmotic permeability and when stimulated will absorb KC1 from the lumen. Fig. 3 shows the results of an experiment which displays these effects very clearly. In this experiment, K-rich fluid hyper-osmotic to the bathing fluid was perfused through the entire lower Malpighian tubule, first from its lower end and then, with exactly the same long length of tubule in the bathing drop, from its upper end.

When perfused antidromically, i.e. from its lower end, fluid emerged slowly at an osmotic concentration close to that of the bathing medium. Presumably what happened first was an extensive absorption of KC1 in the lower part leaving a hypo-osmotic fluid to run through the osmotically permeable upper region further along the tubule so that water rapidly moved out by osmosis. When perfused orthodromically i.e. from the upper end, fluid rapidly emerged and had an osmotic concentration considerably lower than the bathing medium. Presumably an osmotic uptake of water from the bathing medium in the osmotically permeable region was followed by an absorption of KC1 leaving a rapid flow of hypo-osmotic fluid to run out of the lower end of the tubule.

We have seen that regions close to the two ends of the lower tubule differ very much in their properties. At what point along its length does a change occur or is there a gradual shift in properties along the whole length? Two types of experiments were used to investigate this, one following the osmotic permeability and the other the absorption of KCl. As a preliminary, 25 dissected lower Malpighian tubules were measured with an ocular micrometer and found to have a length of 19·1 ± 0·5 mm.

To investigate the longitudual distribution of osmotic permeability, experiments were done in which the whole lower tubule was bathed in a stimulant-free fluid markedly higher in osmotic concentration (330 m-osmol; 160 mm Na, o mm K) than the K-rich fluid perfused through it (200 m-osmol; 90 mm K, 10 mm Na). For each experiment the lower tubule was cannulated at its upper end and fluid run through it at about 110 nl min⁻¹. Differing amounts of the tubule were then pulled out from the bathing drop by moving the bathing drop further away from the fixed cannula. In this way progressively less of the osmotically permeable upper regions remained in the bathing drop. The rate at which fluid collected at the lowermost cut end of the tubule was measured. The results (Fig. 4) show that until more than 50% of the tubule had been pulled out of the bathing drop, the rate of fluid emergence was very little different from that expected if complete osmotic equilibration had occurred. However, as the length of tubule pulled out changed from 60 to 70%, the rate of emergence of fluid rose rapidly to close to the level expected if no osmotic equilibration had occurred. This behaviour is consistent with an abrupt change in osmotic permeability at a point close to 70% of the way along the lower tubule from its upper end. Or, in other words, relative osmotic impermeability is confined to the lowermost 30% of the length of the lower Malpighian tubule. As before, the low osmotic permeability of this part of the tubule was not dependent on stimulation.

To investigate the longitudinal distribution of solute (KC1) uptake, 10–15 μl drops of 5-HT-containing, K-free saline (280 m-osmol; 140 mm Na) were positioned at various points along the length of the lower Malpighian tubule while K-rich fluid (280 m-osmol; 14 mm Na, 129 mm K) was perfused through the lumen. The osmotic concentration (determined by conductivity) of the fluid emerging from the tubule was measured and in Fig. 5 is plotted against the position of the drop. The results are consistent with an ability to take up KC1 confined to the lowermost one third of the tubule, i.e. approximately the same part of the tubule which has a low osmotic permeability.

In the earlier work on absorption in the lower tubule (Maddrell & Phillips, 1975) it was shown that KC1 uptake only occurred in the presence of a stimulant such as the diuretic hormone or 5-HT. To check that this applies equally to the lower regions of the lower tubule, K-rich fluid (10 mm Na, 137 mm K) containing ⁴²K was perfused at 60 nl min^-1 through the entire length of the lower tubule which had two 40 μl bathing drops positioned one close to the upper end and one close to the lower end.

The bathing fluid contained 5 mm K and had an osmotic concentration of 340 m-osmol. The concentration of ⁴²K in the drops emerging was followed and 5-HT at 10⁻⁴ M was added first to the drop bathing the upper region and then to the drop bathing the lower region. As Fig. 6 shows, addition of stimulant to the upper part had no effect but when added to the drop bathing the lower region the ⁴²K content of the perfused fluid abruptly fell to less than 20% of the previous level. Three other experiments gave comparable results so that one can conclude that absorption of potassium in the lower portion of the lower tubule is dependent on stimulation.

So far then it has been shown that the upper two-thirds of the lower tubule has a relatively high osmotic permeability and does not participate in KC1 uptake, whereas the lowermost one-third of the lower tubule has a much lower osmotic permeability, and, when stimulated, rapidly absorbs KC1 from the lumen. Are these differences reflected in other features of the lower tubule?

The permeability of the two different parts of the lower tubule to urea, fucose, Cl⁻ and SO₄²⁻ was determined. This was done by bathing either of the two regions in a drop of standard saline containing a radioactively labelled test substance which was radioactively labelled. From the rate of appearance of label in fluid perfused through the lumen, the length of tubule in the bathing fluid and the rate of perfusion, the apparent permeability of the tubule wall was calculated (Table 3). Rather surprisingly, the permeability of the two different regions to sulphate ions is the same. To chloride ions and urea, both small molecules, the upper part of the tubule is rather more permeable than is the wall of the lower part of the tubule, though the difference is not large bearing in mind the very large difference in osmotic permeability. The figures for fucose are not easily interpreted. Not only are they higher than for the much more permeable wall of the fluid-secreting part of the Malpighian tubule (whose permeability to fucose is 0·60 nl mm⁻² min⁻¹, Maddrell & Gardiner, 1974) but fucose crosses the wall of the lower part of the lower tubule some five times more readily than the wall of the upper length of the lower tubule. Further research will be required to understand these peculiar results. Perhaps they indicate a lumen-directed transport process; if so, its significance is not at all clear.

Apart from the findings with fucose, the results show rather little regional difference in permeability to solutes along the lower Malpighian tubule.

The lower tubule rapidly transports alkaloids such as nicotine, morphine and atropine into the lumen (Maddrell & Gardiner, 1976). To test whether this ability might be localized in one region of the tubule, drops of saline containing ¹⁴C-labelled nicotine were positioned close to one end or the other of the tubule and the rate of appearance of label in fluid perfused through the tubule was followed. With the bathing drops containing 1 mm nicotine, nicotine appeared in the perfused fluid at more than 25 pmol min⁻¹(n = 5) when 4 mm lengths of either the upper of lower regions were tested. These rates are, pro rata, similar to those determined in the earlier work (Maddrell & Gardiner, 1976) and show that both upper and lower regions of the lower tubule can transport nicotine into the lumen.

Some experiments using ⁴²K were done to measure the rate at which potassium ions would cross the wall of the lower tubule from a bathing drop to the lumen. Rather as with SO₄²⁻ ions, no differences were found in the potassium permeability of the two regions of the lower tubule. The permeability of the upper region was 0·32 ± 0·15 nl mm⁻² min⁻¹ (n = 6) and of the lower region 0·25+10·17 nl mm^-2 min^-1 (n = 5). It might seem, then, that the lower region of the tubule has the low permeability to potassium which would be needed if it is to reabsorb this ion effectively. However, these first experiments were done with bathing solutions that contained no stimulant. When they were repeated with 5-HT at 10⁻⁴ M added to the bathing drop there was a most dramatic increase in the apparent potassium permeability of the lower region of the lower tubule although the upper region remained unaffected (Fig. 7). In some extreme cases the apparent potassium permeability rose from less than 0·1 nl mm⁻² min⁻¹ to more than 100 nl mm⁻² min⁻¹. The mean value of the apparent permeability after 5-HT treatment was 54 nl mm^-2 min⁻¹, n = 7, with a range from 18 to 144 nl mm⁻² min⁻¹.

At the moment it is not clear what the significance is of this stimulation-dependent increase in lumen-directed potassium flux but its occurrence is another marked difference between the upper and lower regions of the lower Malpighian tubule.

The results presented in this paper have an important consequence. The usual assumption that all the cells of an epithelium of uniform structural appearance are functionally equivalent ought not to be relied upon. In turn, it follows that attempts to interpret the function of epithelial cells on the basis of their easily observable ultrastructural appearance may not be very rewarding. It is most striking that the different regions of the lower tubule of Rhodnius can have such pronounced differences in properties and yet apparently show no evident ultrastructural differences. It will plainly be worth while to look much more carefully at the ultrastructure of the tubule both in sections and in preparations made by freeze-fracture techniques to see if differences appear at a more detailed level of structural analysis. Such a study is now in progress.

The finding that the important process of reabsorption of KC1 occurs only in the lowermost 30 % of the lower tubule raises the question of the function of the remainder of this part of the tubule. It does not secrete fluid (Maddrell & Phillips, 1975) but we have seen in the present paper that it can carry out active secretion of alkaloids into the lumen. The history of the study of Malpighian tubules shows a dramatic increase in the number of processes known to occur in them (Maddrell, 1977). Since there is no necessity to suppose that all their functions or even a majority of them have been uncovered it is presumably the case that the upper regions of the lower tubule fulfil some unknown auxiliary functions. The paradoxical behaviour of the lower tubule towards fucose (p. 142) may be an indication of some such process.

The speed of solute reabsorption in the lower tubule is worth emphasis. In vivo at 22 °C, fluid flows through the lower tubule at 0·6 mm s⁻¹ and is in contact with the reabsorptive region for less than 10 s. Yet in this time its potassium concentration falls from above 80 mm to less than 5 mm, that is a fall of 1 mm takes only 130 ms. The osmotic concentration declines from 370 to 250 m-osmol in the same time, a rate of change of 12 m-osmol s⁻¹. This corresponds to a net flux of potassium chloride across the tubule wall of about 0·80 μmol cm⁻² min⁻¹. These figures are significantly above the rates of solute absorption achieved by vertebrate epithelia. For example, sodium chloride is absorbed in the rabbit gall bladder at a rate of 0·13 μmol cm⁻² min⁻¹ (Diamond, 1964) and in the proximal kidney tubule of the rat at a rate of 0·46 μlmol cm⁻² min⁻¹ (Giebisch et al. 1964). These determinations on vertebrate epithelia were made at 37 °C ; it can be calculated from measurements of the osmotic concentration of the fluid excreted at this temperature (Maddrell, 1964) that the rate of net solute reabsorption in the lower tubule of Rhodnius at 37 °C must be at least as high as 0·95 μmol cm⁻² min⁻¹.