Recovery From Acute Haemolymph Acidosis in Unfed Locusts: I. Acid Transfer to the Alimentary Lumen is the Dominant Mechanism

Regulation of acid-base status is one of the most critical aspects of organismal homeostasis. The ability of insects to preserve haemolymph acid-base status and the mechanisms by which such regulation may be achieved are poorly understood (Phillips et al. 1986). In this study we examine the capacity of the desert locust Schistocerca gregaria to recover from acute haemolymph acidosis (HC1 injections) and quantify the contribution of some of the possible mechanisms responsible for this recovery.

In the terrestrial animals that have been studied, recovery from extracellular acid-base disturbances may occur by a number of mechanisms, including ventilatory adjustment of ⁠, mobilization of blood buffer compounds, excretion of acid equivalents from the animal by epithelial transport and transfer of acid equivalents to another body compartment (Truchot, 1987). Increases in ventilatory carbon dioxide excretion aid in restoring haemolymph acid-base status after exposure to elevations in environmental or endogenous in locusts (Harrison, 1989a; Harrison et al. 1991). However, haemolymph acid-base and ventilatory responses to metabolic acid loads have not been studied in insects. Metabolic acid loads may be produced in animals during a variety of circumstances, including fasting, ketosis or in association with anaerobic energy metabolism (Masoro and Siegel, 1977). In cockroaches, ventilation increases when the pH of saline irrigating the ventral nerve cord is decreased at constant (Snyder et al. 1980). The results of Snyder et al. suggest that insects have the capacity to reduce haemolymph in response to depressed haemolymph pH.

Transfer of acid-base equivalents out of the haemolymph allows organisms to regulate pH and [HCO₃⁻] at constant ⁠. In terrestrial insects, acid-base equivalents may be transferred out of the animal by renal excretion. Alternatively, acid-base equivalents may be transferred from the haemolymph into the intracellular space or the alimentary lumen. The pH of the alimentary lumen of locusts differs from that of haemolymph throughout most of its length (Bodine, 1925; Swingle, 1931; Thomson et al. 1988a). Because the alimentary lumen in locusts contains a large fraction of body water and bicarbonate (Harrison, 1989 b), this compartment may have considerable physiological capacity for temporary storage of acid-base equivalents. Transport of acid-base equivalents across renal or alimentary epithelia has been demonstrated in a variety of insects (Ramsay, 1956; Irvine et al. 1988; Thomson et al. 1988a,b; Lechleitner et al. 1989; Chamberlin, 1990; Dow and O’Donnell, 1990; Stagg et al. 1991). However, it is not known whether regulatory adjustments of acid-base transport by the excretory or alimentary epithelia occur during haemolymph pH regulation. In this study, we test for transfer of acid equivalents to the alimentary lumen in response to haemolymph acidosis. In a companion paper (Harrison and Phillips, 1992), we examine the role of renal acid and nitrogen excretion in the recovery from HC1 injections.

Locusts (Schistocerca gregaria) were obtained from a colony maintained as described previously (Thomson et al. 1988a). Adult female locusts, 14–25 days past their final moult, were used in all experiments. Animals were fed on lettuce ad libitum for 2h at 35 °C, and then kept solitarily in cardboard cartons with access only to distilled water-soaked cotton for 20–24h at 21 °C (±0.5°C). Locusts were then injected with 25μl of 400 mmol l⁻¹ HC1 or NaCl using a 100μl Unimetrics syringe via the fourth or fifth abdominal intersegmental membrane. Bleeding from the injection site was generally minimal. Animals were then returned to their individual cartons.

Owing to limitations in sample sizes, not all variables were measured for each animal, and all samples were terminal. Therefore, effects of HC1 injections were evaluated by comparing values for HC1-and NaCl-injected animals under similar conditions. Variables measured concurrently from single animals are described below. For each experimental series, half of the locusts were injected with HC1 and half with NaCl.

Luminal pH was measured prior to injections and 24 h after HC1 or NaCl injections in a third group of animals. The pH of the various sections of the alimentary lumen was measured using a glass microelectrode inserted through a small tear (<1 mm) in the epithelium after rapid (<5 min) dissection to expose the gut. The reference electrode used was a PE-10 KCl-agar bridge placed in the lumen, in series with a calomel electrode. The potential difference between the reference and pH electrodes was measured with the Keithley electrometer.

We measured the total ammonia concentration (Amm_tot) and of the contents of alimentary lumen 24 h after HC1 or NaCl injections in a fourth group of animals. Luminal contents were removed with forceps after rapid (<5min) dissection. was measured immediately by gas chromatography. Samples of gut contents used for measurement of Amm_tot were immediately placed in 19 times their volume of 5 % TCA to prevent loss of NH₃ and frozen. This solution was later diluted 1500 times with distilled water, mixed and centrifuged (5 min, 1500 g). The supernatant was analyzed for Amm_tot using an assay modified from Kun and Kearney (1974). The masses of the luminal contents of the crop and midgut were determined 24 h after injections on a fifth group of animals from the difference between the mass of intact segments and the mass of segments with luminal contents removed by washing with isosmotic saline.

Non-bicarbonate buffer values of the crop and midgut contents were measured 24 h after NaCl or HC1 injections using the fifth group of animals. Samples were diluted 1000 times with 100 mmol l⁻¹ KC1, to approximate the KCl-rich fluids found in these compartments (Dow, 1981). Carbon dioxide was removed by adding HC1 to lower solution pH below 4, followed by stirring for 2h. Samples were then alkalized by the addition of KOH (Radiometer TTT80 titrator and model ABU80 Autoburette), while pH was measured with a Radiometer GK2321C pH electrode. We calculated nonbicarbonate buffer value (mequiv kg⁻¹ pH unit⁻¹) as the quantity of base equivalents required to titrate 1kg of bicarbonate-free sample by 1 pH unit. Buffer values were measured over the range of pH corresponding to the difference in the pH of HC1-and NaCl-injected animals (crops 5.3–6.0; midguts 6.2–7.0). Buffer values were corrected for the buffering due to the dilution saline by subtracting the buffer values for 100 mmol l⁻¹ KC1 solutions.

The phrase ‘addition of 1 μequiv of acid to a compartment’ indicates that the compartment has been acidified in a manner equivalent to the addition of 1 μequiv of HC1. The conditions of this study do not permit discrimination between acid equivalent transfer out of a compartment and base equivalent addition to the compartment.

Data were tested for normality (Kolmogorov’s test) and heteroscedasticity (Fmax test). Data satisfying assumptions of analysis of variance (ANOVA) were tested for significance with ANOVA or t-tests. Post-hoc contrasts for multiple comparisons were performed using the Bonferroni procedure to maintain the overall level for statistical significance of P<0.05 (Wilkinson, 1987).

Fifteen minutes after acid injections, haemolymph [H⁺] was increased by 10.3×10⁻⁸ mmol l⁻¹, while haemolymph was approximately halved (Fig. 1). Recovery to values statistically equivalent to resting occurred in 8 h ([H⁺]) or 24 h ⁠. Injection of 10μmol of NaCl into the haemolymph had no effect on haemolymph [H⁺] or (post-hoc contrasts, P > 0.05), indicating that the injection procedure did not affect haemolymph acid-base status (Fig. 1).

Acid injection had no effect on haemolymph protein or inorganic phosphate concentrations (Table 1). As these compounds account for a large fraction of the haemolymph non-bicarbonate buffer value in S. gregaria (Harrison et al. 1990), we constructed a Davenport diagram of haemolymph acid-base status after acid injection assuming that the haemolymph non-bicarbonate buffer value also remained constant throughout the experimental period. The Davenport diagram (Fig. 2) illustrates that recovery of acid-base status occurs primarily by movement along the CO₂ isobar, indicating that recovery occurs predominantly by non-respiratory mechanisms.

Fifteen minutes after acid injection, calculated haemolymph was significantly elevated (Fig. 2; post-hoc contrast, F=6.21, P<0.05); however, from 1 to 24h after injection, haemolymph values did not differ significantly from control values (Fig. 2; post-hoc contrasts, P>0.05). Calculated haemolymph [HCO₃⁻] returned to values equivalent to controls in 24h (Fig. 2; post-hoc contrasts).

Twenty-four hours after HCI injections, gut luminal pH values were significantly lower than pre-injection values in the crop, midgut and at the point of Malpighian tubule entry (Fig. 3, t-tests, P<0.05). NaCl injection had no effect on the luminal pH of the alimentary canal (Fig. 3, t-tests, P >0.05). Midgut and crop luminal pH values were significantly lower after HC1 injections than after NaCl injections (Table 2, t-tests, midgut: t =4.01, P<0.001; crop: t=2.29, P<0.05).

In the midgut, values were significantly lower after HC1 injection than after NaCl injection (Table 2, t-test, t=2.41, P <0.05). In the crop, values were low and similar after HC1 or NaCl injection (Table 2, t-test, P >0.05). NaCl-and HCl-injected animals did not differ in their luminal Amm_tot values (Fig. 4), indicating that effects of NH₄⁺ transfer to the alimentary lumen on acid elimination from the haemolymph need not be considered. Using compartment buffer values and volumes (Table 2), we estimated that 50% of the injected acid load was transferred to the lumen of the midgut and 24% into the lumen of the crop.

This study provides the first unequivocal evidence that insects regulate haemolymph acid-base status. After an injection of HCl that reduced haemolymph pH by 0.5 pH units and haemolymph [HCO₃⁻] by 50 %, locusts completely recovered haemolymph pH, and [HCO₃⁻] within 24h. The dominant mechanism by which this recovery occurred was transfer of net acid equivalents out of the haemolymph into the alimentary lumen.

Initially, approximately 50 % of the injected HC1 was buffered by bicarbonate and approximately 50% by non-bicarbonate buffers (calculated as described in Harrison et al. 1990). The elevated haemolymph 15 min after the HC1 injection suggests that initially the tracheal system did not release all of the carbon dioxide derived from bicarbonate buffering. However, from 1 to 24 h after the HC1 injections, haemolymph did not differ from that in control animals. The tracheal system effectively eliminated all of the carbon dioxide produced from bicarbonate buffering, but we found no evidence for respiratory compensation (reduced ⁠) for reduced haemolymph pH during long-term recovery (1–24 h).

Haemolymph levels of protein and inorganic phosphate did not vary after HC1 injections. It is possible that concentrations of unmeasured buffer compounds (organic phosphates or citrate; Harrison et al. 1990) increased after HC1 injections. However, since protein and inorganic phosphate account for 85 % of the non-bicarbonate buffer value in the haemolymph of quiescent S. gregaria (Harrison et al. 1990), it seems likely that an increased haemolymph non-bicarbonate buffer value was not a mechanism of blood pH regulation.

If metabolic consumption of protons within the haemolymph is small, recovery of haemolymph acid-base status at constant and haemolymph buffer value must occur by transfer of acid equivalents out of the haemolymph. Supporting this assumption, the cell fraction (and presumably the metabolic capacity) of locust haemolymph is small (<2 %, J. F. Harrison, unpublished data). We estimate that 75 % of the acid equivalents removed from the haemolymph during the recovery period are transferred to the lumens of the midgut and crop (Table 2). A further 15 % of the acid equivalents are excreted as ammonium (Harrison and Phillips, 1992). The midgut is a quantitatively more important sink for acid equivalents than is the crop, as indicated by its larger volume and its greater change in pH and calculated [HCO₃⁻]. There was no evidence for a decreased pH in the lumen of the hindgut after HC1 injections (Fig. 3). The unchanging luminal pH within the hindgut after acid injection is consistent with the finding that the pH of faecal pellets does not change after acid injection (Harrison and Phillips, 1992).

The decrease in pH of the alimentary lumen at the point of Malpighian tubule entry suggests that a portion of the acid transfer into the midgut may be mediated by the Malpighian tubules. Some of the fluid secreted by the Malpighian tubules is known to flow forward into the midgut in locusts, particularly in unfed animals (Dow, 1981). Recently, we have found that the pH of secreted Malpighian tubule fluid decreases during haemolymph acidosis when tubules are cannulated in vivo (Stagg et al. 1991). However, acid secretion by the Malpighian tubules can only account for at most 20 % of the acid equivalents transferred to the lumen of the midgut (Stagg et al. 1991). Therefore, the majority of the acid equivalents transferred from haemolymph to midgut lumen must be transported by the midgut epithelia and/or transferred from the crop lumen. In larval lepidopteran midgut, base equivalent transfer to the lumen has been well-demonstrated (Chamberlin, 1990; Dow and O’Donnell, 1990). However, acid-base transport by the locust midgut has not been studied.

Presumably, the acid transfer to the lumen of the crop is mediated by the salivary glands, as the crop epithelium is relatively impermeable to ions and insect salivary glands are known to secrete ions, enzymes and various metabolites (Miles, 1967; Oschman and Berridge, 1970; Bay, 1978). In the only insect in which the acid-base status of salivary fluid has been examined, the saturniid moth Antheraea peryi, the secreted fluid is alkaline and high in bicarbonate (Kafatos, 1968). The low pH of the locust crop indicates that the situation must be quite different in S. gregaria.

Why is acid equivalent transfer into the fore-and midgut the dominant mechanism of haemolymph acid-base regulation in unfed locusts? Unfed locusts have been shown to retain food within the alimentary canal, improving fractional digestion and absorption and reducing the excretion rate of urinary pellets (Baines et al. 1973). In addition, in unfed locusts, the pH of the hindgut lumen is already near the minimum luminal pH sustainable by the hindgut epithelia in vitro (Thomson et al. 1988b). Together, these effects limit the capacity of unfed locusts to increase acid excretion in urinary pellets (Harrison and Phillips, 1992). Therefore, non-respiratory recovery from haemolymph acid loads requires transfer of acid equivalents either to the intracellular space or to some other storage compartment within the animal. Transfer of acid equivalents to the alimentary lumen provides unfed locusts with a temporary mechanism of haemolymph pH regulation which does not compromise intracellular acid-base status or require increases in ventilation that would increase water loss. The acid equivalents stored within the alimentary lumen are presumably excreted later, when locusts feed and produce frequent urinary pellets. This hypothesis has not yet been tested.

While transfers of acid equivalents to the fore-and midgut are important mechanisms of haemolymph pH regulation after HCI injection, it is not yet known whether locusts actually utilize such mechanisms under natural conditions. Unfed locusts excrete acid equivalents, perhaps derived from oxidation of sulphur-and phosphate-containing proteins (Harrison and Phillips, 1992). Potentially, acid equivalent transfer to the fore-and midgut may function to remove non-volatile metabolic acid loads during long-term fasting (Masoro and Siegel, 1977) when faecal production is very low. While changes in midgut pH may reduce the activity of digestive enzymes (reviewed by Applebaum, 1985), digestion and absorption is virtually complete within 6h after a meal in locusts (Baines et al. 1973). After digestion and absorption have been completed, suspending regulation of the pH in the alimentary lumen may be inconsequential.

This work was partially supported by NATO and Killam PDF’s to J.F.H., an NSERC undergraduate research award to C.J.H.W. and an NSERC operating grant to J.E.P. J. Fewell made helpful comments on the manuscript.