Aquaporin-4 antibodies, CNS acidosis and neuromyelitis optica: a potential link.

BACKGROUND: Neuromyelitis optica (NMO, Devic's syndrome) is a severely disabling disorder of the central nervous system characterized by optic neuritis and longitudinally extensive myelitis. In around 80% of cases, NMO is caused by autoantibodies to astrocytic aquaporin-4 (AQP4), the most abundant water channel in the CNS. Acute NMO attacks are frequently accompanied by elevated levels of lactate in the cerebrospinal fluid (CSF). As a strongly dissociated anion (pK'=3.7) directly changing the strong ion difference, lactate causes a reduction in the dependent anion [HCO3-] and a rise in [H+], resulting in "metabolic" acidosis in the CSF. CSF acidosis also develops during respiratory failure due to brainstem or high cervical spinal cord lesions, the most common cause of death in NMO. However, lactic acid and more generally, a decrease in pH, has been shown to increase the membrane expression of AQP4 in astrocytes. An increase in AQP4 membrane expression during acute NMO attacks could potentially enhance the complement-mediated humoral immune reaction against AQP4-expressing astrocytes characteristic for NMO and, thus, result in more severe astrocytic damage. Moreover, lactate and acidosis have been shown to cause astrocytic swelling and to affect astrocytic viability, potentially rendering astrocytes more susceptible to AQP4-Ab-mediated damage. Finally, increased AQP4 expression could be an independent risk factor in NMO and other forms of CNS inflammation, as indicated by the finding of grossly attenuated experimental autoimmune encephalomyelitis in AQP4-null mice. Therefore, we hypothesize that CSF acidosis might play a role in the pathophysiology of AQP4-Ab-positive NMO and that alterations in CSF pH might possibly influence the outcome of acute attacks in this condition. In addition, we discuss potential clinical implications and make proposals on how to test the hypothesis. Finally, other factors that influence astrocytic AQP4 membrane expression and might play a role in NMO are discussed.

Correlation of cerebral spinal fluid pH and HCO3- with disease progression in ALS.

Amyotrophic lateral sclerosis (ALS) is a progressive and fatal disease, which is characterized by progressive degeneration of spinal and bulbar innervating motor neurons. However, the underlying mechanisms of motor neuron death remain poorly understood. Several candidate disease biomarkers have been detected in cerebrospinal fluid of ALS patients. The present study analyzed various cerebral spinal fluid gas parameters in ALS patients and compared these values to controls, as well as patients with cervical spondylosis, Parkinson syndrome, and spinocerebellar degeneration. Cerebral spinal fluid pH positively correlated with the ALS functional rating scale in total and limb-type ALS patients. In addition, cerebral spinal fluid pH positively correlated with shorter disease duration (less than 22 weeks). These results suggested that cerebral spinal fluid pH provides a biomarker for ALS and could reflect mechanisms of disease progression in ALS patients.

Hyperventilation in severe diabetic ketoacidosis.

OBJECTIVE: To explore whether the carbon dioxide-bicarbonate (P(CO(2))-HCO(3)) buffering system in blood and cerebrospinal fluid (CSF) in diabetic ketoacidosis should influence the approach to ventilation in patients at risk of cerebral edema. DATA SOURCE: Medline search, manual search of references in articles found in Medline search, and use of historical literature from 1933 to 1967. DESIGN: A clinical vignette is used--a child with severe diabetic ketoacidosis who presented with profound hypocapnia and then deteriorated--as a basis for discussion of integrative metabolic and vascular physiology. STUDY SELECTION: Studies included reports in diabetic ketoacidosis where arterial and CSF acid-base data have been presented. Studies where simultaneous acid-base, ventilation, respiratory quotient, and cerebral blood flow data are available. DATA EXTRACTION AND SYNTHESIS: We revisit a hypothesis and, by reassessing data, put forward an argument based on the significance of low [HCO(3)](CSF) and rising Pa(CO(2))- hyperventilation in diabetic ketoacidosis and the limit in biology of survival; repair of severe diabetic ketoacidosis and Pa(CO(2))-and mechanical ventilation. CONCLUSION: The review highlights a potential problem with mechanical ventilation in severe diabetic ketoacidosis and suggests that the P(CO(2))--HCO(3) hypothesis is consistent with data on cerebral edema in diabetic ketoacidosis. It also indicates that the recommendation to avoid induced hyperventilation early in the course of intensive care may be counter to the logic of adaptive physiology.

Cerebrospinal fluid acid-base status during normocapnia and acute hypercapnia in equine neonates.

OBJECTIVE: To determine normal acid-base status of the CSF and to compare it with changes during acute hypercapnia in equine neonates. ANIMALS: 10 clinically normal foals between 1 and 12 days old. PROCEDURE: CSF and arterial and venous blood samples were collected every 15 minutes during 45 minutes of normocapnia and 90 minutes of hypercapnia in isoflurane-anesthetized foals. CSF samples were collected via a subarachnoid catheter placed in the atlanto-occipital space. RESULTS: Comparison of blood and CSF gases during normocapnia indicated that CSF was significantly more acidic than blood. The lower pH was attributable to higher CO2 and lower bicarbonate concentrations than those in blood. During hypercapnia, CSF CO2 increased and pH decreased parallel to changes in blood, but changes were not a great as similar changes in venous blood, indicating that some degree of buffering occurs in the CSF of foals. CONCLUSIONS: Normal CSF acid-base status in equine neonates is similar to that in other domestic species. The blood-brain and blood-CSF interfaces in neonates allow rapid diffusion of CO2, but allow only slow diffusion of bicarbonate. Equine neonates are capable of buffering respiratory-induced acid-base changes in the CSF, but the buffering capacity is less than that of the vascular compartment. CLINICAL RELEVANCE: Neonatal foals may develop severe respiratory compromise, resulting in hypoxemia and hypercapnia. Because the ability of the CSF to buffer acid-base changes in neonates is reduced, hypercapnia may contribute to the CNS abnormalities that often develop in sick neonates. Thus, normal blood gas values should be maintained in diseased equine neonates.

Acid-base state of cerebrospinal fluid during pregnancy and its effect on spread of spinal anaesthesia.

To assess the possible relationship between changes in acid-base state of cerebrospinal fluid (CSF) and enhanced spread of spinal anaesthesia during pregnancy, we have measured CSF pH, carbon dioxide tension (PCO2) and HCO3- values in 73 women undergoing spinal anaesthesia with hyperbaric amethocaine 8 mg. Patients were allocated to one of four groups according to gestational period: non-pregnant group (n = 13), first trimester group (8-13 weeks, n = 19), second trimester group (14-26 weeks, n = 11) and third trimester group (27-39 weeks, n = 30). The pH of the CSF was greater in the second and third trimester groups than in the non-pregnant group. CSF PCO2 decreased by 0.53-0.8 kPa throughout pregnancy. CSF HCO3- was decreased throughout pregnancy. Overall, no clinically significant correlation was found between maximum cephalad spread of analgesia and CSF pH, PCO2 or HCO3-. We conclude that pregnancy-induced changes in acid-base state of CSF have little effect on the spread of spinal anaesthesia, although there is a clinically different spread of spinal anaesthesia between non-pregnant and pregnant states.

Efecto del bicarbonato de sodio sobre la actividad colinesterasa en cerebro, eritrocitos, plasma e higado de ratas intoxicadas con paratión

El presente trabajo se realizó con la finalidad de evaluar el efecto que tiene el bicarbonato sobre la actividad de la acetil colinesterasa de diversos tejidos, en animales de experimentación intoxicados con paratión. Para ello se formaron cuatro grupos de diez ratas cada uno. El grupo A sirvió de control . Los grupos B y C recibieronm por vía intraperitoneal paratión y además este último recibió bicarbonato de sodio luego de quine minutos de ser administrado el organofosforado. El grupo D recibió únicamente bicarbonato de sodio. Se dosaron las actividades colinesterásicas de los cuatro grupos en plasma , eritrocito, cerebro e hígado, por el método de Ellman modificado, y se compararon las actividades en los diferentes tejidos entre los grup[os de estudio con el fin de ver como variaba la actividad de la enzima inhibida por el paratión, por efecto del bicarbonato. Se encontró que la inhibición de la actividad colinesterásica, por el efecto del paratión, era significativa a nivel plasmático, cerebral y hepático, mas no a nivel eritrocítico; además se vió que el bicarbonato por si solo inhibia la actividad de la enzima en los mismos tejidos, de manera significativa. Por otro lado se encontró que la actividad colinesterásica inhibida por el órgano fosforado, se recuperó luego de la administración del bicarbonato de sodio. Con los resultados obtenidos se propone que la acción del bicarbonato sería desactivando al insecticida mismo y no sobre la colinesterasa.

Exchange of CO2 and bicarbonate between the circulation and cerebrospinal fluid in piglets.

The dynamic exchange of CO2 and HCO-3 between the central circulation and peripheral sites such as the cerebrospinal fluid (CSF) is not completely understood. To examine this we administered a tracer dose of NaHCO3 labeled with the stable isotope 13C into the central circulation of nine 3- to 4-wk-old anesthetized, instrumented piglets. Serial samples of arterial and venous blood, breath, and CSF were obtained to determine the 13C/12C ratio by isotope-ratio mass spectrometry. Gas-exchange and hemodynamic parameters were obtained through standard techniques. The patterns of tracer washout in venous blood, arterial blood, and breath were nearly indistinguishable from each other, and the mean residence time was 113 +/- 24 min (average +/- SD). In contrast, the CSF 13C/12C ratio was not equivalent to that in venous blood, arterial blood, or breath until 20 min after tracer administration. Tracer washout data were used to determine the best-fit three-exponential-term equation and to calculate compartmental parameters of a three-compartment mammillary system in which CO2 and HCO3- residing in the central compartment (compartment 1) exchanges relatively rapidly with one peripheral compartment (compartment 2) and slowly with the other (compartment 3). The turnover time for the CSF was 6.8 +/- 2.5 min and the estimate for compartment 2 was 7.1 +/- 3.7 min. Accordingly, CSF appears to be one of the rapidly interchanging peripheral compartments.

Quantitative cerebrospinal fluid acid-base balance in acute respiratory alkalosis.

Data on canine cisternal cerebrospinal fluid (CSF) ions in acute respiratory alkalosis are limited and fragmentary. We hypothesized that with the fall in arterial PCO2 (PaCO2) and in the face of normal osmoregulation, CSF [Na+] remains relatively constant and CSF [Na+-Cl-] narrows to account in part for the fall in CSF [HCO3-]. We therefore measured blood and CSF acid-base variables and ions of two groups of pentobarbital-anesthetized, mechanically ventilated dogs (n = 10 in each group). In the control group, PaCO2 was kept constant and changes in serum and CSF ions were minimal. In Group II (acute respiratory alkalosis), both PaCO2 and cisternal CSF PCO2 decreased by 10 mm Hg. Five hours after induction of respiratory alkalosis, mean CSF [HCO3-] decreased significantly by 4.4 +/- 1.2 mEq/L (mean +/- SD). The fall in CSF [HCO3-] was similar to changes in CSF strong ion difference (SID = Na(+)+K(+)+Ca(2+)+Mg(2+)-CL(-)-lactate), which decreased 4.4 +/- 1.9 mEq/L. Concentrations of the four major CSF cations did not change significantly. Cisternal CSF lactate rose significantly by 1.2 +/- 0.9 mEq/L, accounting for 25% of the change in CSF [HCO3-]. The remaining (75%) change in CSF [HCO3-] was accounted for by changes in CSF [Cl-].

Cl- replacement alters the ventilatory response to central chemoreceptor stimulation.

To determine whether cerebrospinal fluid (CSF) Cl- has a role in determining the stimulus to the central respiratory chemoreceptors under conditions of constant CSF pH, CO2, and HCO3- concentrations, the ventral medullary surface of the anesthetized rat was perfused with mock CSF of various ion composition and pH. Four mock CSF perfusates were used: two normal pH control perfusions and two acidic solutions. One acidic perfusate was formulated in the traditional manner by substituting Cl- for HCO3-. The second acidic perfusate, and one of the normal pH control perfusates, had approximately 15% of the Cl- replaced with isethionate, an impermeant strong anion. When the two acidic solutions were perfused over the ventral medulla, consistently larger increases in both tidal volume and minute ventilation were observed with the isethionate-containing acidic solution, despite conditions of identical pH and PCO2. The unequal ventilatory effects of the two acidic perfusions suggest that Cl- transport may be a factor determining the stimulus to the central respiratory chemoreceptors.

Long-term monitoring of CSF lactate levels and lactate/pyruvate ratios following subarachnoid haemorrhage.

Ventricular cerebrospinal fluid (CSF) lactate concentrations and lactate/pyruvate (L/P) ratios were measured daily in 20 patients from day 1 to day 12 after subarachnoid haemorrhage due to ruptured aneurysms. Patients without symptomatic vasospasm were classified in Group 1, patients with symptomatic vasospasm were classified in Group 2, and patients who were Hunt and Kosnik grade 4 on admission clinically were classified in Group 3. Patients in all three groups had high CSF lactate concentrations on day 1, and, especially in Group 3, the high lactate was accompanied by an increased L/P ratio and a decreased CSF bicarbonate. Lactate concentrations in Group 1 decreased throughout the observation period. Lactate concentrations in Group 2 also decreased but then began to increase again on days 5 to 7, correlating well with the onset of cerebral vasospasm. The delayed increase of CSF lactate in Group 2 was also accompanied by increases in the CSF pyruvate level and the CSF L/P ratio. Daily monitoring of CSF lactate may thus serve as a chemical marker for cerebral vasospasm.

Effects of sodium lactate infusion on cisternal lactate and carbon dioxide levels in nonhuman primates.

OBJECTIVE: To further the understanding of lactate-induced panic in patients with panic disorder, the authors examined cisternal lactate and carbon dioxide levels in nonhuman primates after infusions of sodium lactate comparable to those used in studies of human beings. METHOD: CSF and venous blood lactate, pH, PCO2, PO2, and bicarbonate were measured in five ketamine-anesthetized nonhuman primates, without mechanical ventilation, before and after they underwent infusions of sodium lactate. In addition, the same measurements were made for three of the five subjects who were given saline infusions. RESULTS: Despite the development of the characteristic peripheral biochemical effects of infused sodium lactate--increased lactate and bicarbonate levels and metabolic alkalosis--no increases in central lactate or carbon dioxide levels were observed. Saline infusions produced no biochemical effects on venous and cisternal measures. CONCLUSIONS: The results of this study are in keeping with previous findings of nonpermeability of the blood-brain barrier to anionic compounds such as lactate. They therefore support theories of lactate panic based on cognitive and/or brainstem misevaluation of peripheral somatic sensations.

Maturational differences in acetazolamide-altered pH and HCO3 of choroid plexus, cerebrospinal fluid, and brain.

The carbonic anhydrase inhibitor acetazolamide is useful for analyzing ion transport, pH regulation, and fluid formation in developing central nervous system. We used the 14C-labeled dimethadione technique to measure alterations in steady-state pH, and to estimate the HCO3 concentration [HCO3], in choroid plexus (CP), cerebrospinal fluid (CSF), and cerebral cortex of 1- and 3-wk-old Sprague-Dawley rats treated with acetazolamide or probenecid. These drugs can suppress transport of HCO3 and other anions in some cells, consequently altering intracellular pH. In 1-wk-old infant rats whose CSF secretory process is incompletely developed, 1 h of acetazolamide treatment did not significantly change CP intracellular pH or [HCO3]. However, in 3-wk-old rats, in which the ability of CP to secrete ions and fluids is almost fully developed, acetazolamide caused marked increases in CP cell intracellular pH and [HCO3]. In contrast, acetazolamide-induced alkalinization was not observed in CSF or cerebral cortex of the 1- and 3-wk-old animals. The other test agent, probenecid (an inhibitor of anion transport but not of carbonic anhydrase), did not alter the pH of any region at any age investigated. Overall, the results are interpreted in light of developmental changes in carbonic anhydrase and previous findings from kinetic analyses of ion-translocating systems in CP. Acetazolamide may interfere with a CP apical membrane HCO3 extrusion mechanism not fully operational in infant rats.

Differential respiratory effects of HCO3- and CO2 applied on ventral medullary surface of rats.

To estimate whether H+ is the unique stimulus of the medullary chemosensor, ventilatory effects of HCO3- and/or CO2 applied on the ventral medullary surface using an improved superfusion technique and of CO2 inhalation were compared in halothane-anesthetized spontaneously breathing rats. Superfusion with low [HCO3-]-acid mock cerebrospinal fluid (CSF) (normal Pco2) induced a significant increase in ventilation, with an accompanying reduction in endtidal Pco2 (PETco2). High [HCO3-]-alkaline CSF depressed ventilation. Changes in Pco2 of superfusing CSF, on the other hand, had no significant effect despite the similar changes in pH. Simultaneous decrease in [HCO3-] and Pco2 of mock CSF with normal pH also maintained stimulated respiration. CO2 inhalation during superfusion with various [HCO3-] solutions caused further increase in ventilation as PETco2 increased. The results suggest that the surface area of the rat ventral medulla contains HCO3- (or H+)-sensitive respiratory neural substrates which are, however, little affected by CO2 in the subarachnoid fluid. A CO2 (or CO2-induced H+)-sensitive chemosensor responsible for the increase in ventilation during CO2 inhalation may exist elsewhere functionally apart from the HCO3- (or H+)-sensitive sensor in the examined surface area.

CSF acidosis augments ventilation through cholinergic mechanisms.

Ventilation is influenced by the acid-base status of the brain extracellular fluids (ECF). CO2 may affect ventilation independent of changes in H+. Whether the acidic condition directly alters neuronal firing or indirectly alters neuronal firing through changes in endogenous neurotransmitters remains unclear. In this work, ventriculocisternal perfusion (VCP) was used in anesthetized (pentobarbital sodium, 30 mg/kg) spontaneously breathing dogs to study the ventilatory effects of acetylcholine (ACh), eucapnic acidic (pH approximately 7.0) cerebrospinal fluid (CSF), and hypercapnic acidic (pH approximately 7.1) CSF in the absence and presence of atropine (ATR). Each animal served as its own control. Base line was defined during VCP with control mock CSF (pH approximately 7.4). With ATR (4.8 mM) there was an insignificant downward trend in minute ventilation (VE). ACh (6.6 mM) increased VE 53% (n = 12, P less than 0.01), eucapnic acidic CSF increased VE 41% (n = 12, P less than 0.01), and hypercapnic acidic CSF increased VE 47% (n = 6, P less than 0.01). These positive effects on ventilation were not seen in the presence of ATR. This suggests that acidic brain ECF activates ventilatory neurons through muscarinic cholinergic mechanisms. Higher concentrations of ACh increased ventilation in a concentration-dependent manner. Higher concentrations of ATR decreased ventilation progressively, resulting in apnea. The results suggest that ACh plays a significant role in the central augmentation of ventilation when the brain ECF is made acidic by either increasing CSF PCO2 or decreasing CSF bicarbonate.

Pial arteriolar vessel diameter and CO2 reactivity during prolonged hyperventilation in the rabbit.

Hyperventilation reduces intracranial pressure (ICP) acutely through vasoconstriction, but its long-term effect on vessel diameter is unknown. In seven rabbits with a cranial window implanted 3 weeks earlier, the effect of prolonged hyperventilation on vessel diameter was studied. Anesthesia was maintained for 54 hours with a pentobarbital drip (1 mg/kg/hr). The pH, CO2, and HCO3- levels were measured in arterial blood and cisterna magna cerebrospinal fluid (CSF). The diameter of 31 pial arterioles was measured with an image splitter. After baseline measurements, pCO2 was reduced from 38 to 25 mm Hg and allowed to return to 38 mm Hg for 10 minutes every 4 hours. There was an initial vasoconstriction of 13%, which progressively diminished by 3% every 4 hours. Thus, by the 20th hour, vessel diameters at a pCO2 of 25 mm Hg had returned to slightly above baseline values obtained at a pCO2 of 38 mm Hg. The temporary return of pCO2 to 38 mm Hg every 4 hours caused vasodilation: 12% at 4 hours, gradually increasing to 16% at 52 hours. Thus, at 52 hours, the vessel diameters were 105% of baseline at a pCO2 of 25 mm Hg and increased to 122% at a pCO2 of 38 mm Hg. Arterial pH had returned to baseline at 20 hours, and CSF pH had returned at 24 hours. Bicarbonate in blood and CSF remained decreased throughout the experiments. In three control experiments during which normocapnia was maintained, vessel diameter and pH and bicarbonate levels remained unaltered over the same period. The CO2 reactivity, tested by brief periods of hyperventilation every 4 hours, also did not change. These results indicate that hyperventilation is effective in reducing cerebral blood volume for less than 24 hours and that it should be used only during actual ICP elevations. If used preventively, its effect may have worn off by the time ICP starts to rise for other reasons, and further decreases in pCO2 cannot be obtained. Moreover, the reduction in buffer capacity with lower bicarbonate renders the vessels more sensitive to changes in PaCO2. This could lead to more pronounced elevations in ICP during transient rises in PaCO2, such as during endotracheal suctioning in head-injured patients.

Blood-brain barrier to hydrogen ion during acute metabolic acidosis in piglets.

The present study investigates the integrity of the blood-brain barrier to H+ or HCO3- during acute plasma acidosis in 35 newborn piglets anesthetized with pentobarbital sodium. Cerebrospinal fluid acid-base balance, cerebral blood flow (CBF), and cerebral oxygenation were measured after infusion of HCl (0.6 N, 0.191-0.388 ml/min) for a period of 1 h at a constant arterial PCO2 of 35-40 Torr. HCl infusion resulted in decreased arterial pH from 7.38 +/- 0.01 to 7.00 +/- 0.02 (P less than 0.01). CBF measured by the tracer microsphere technique was decreased by 12% from 69 +/- 6 to 61 +/- 4 ml.min-1.100 g-1 (P less than 0.05). Infusion of 0.6 N NaCl as a hypertonic control had no effect on CBF. Cerebral metabolic rate for O2 and O2 extraction was not significantly changed from control (3.83 +/- 0.20 ml.min-1.100 g-1 and 5.7 +/- 0.6 ml/100 ml, respectively) during acid infusion. Cerebral venous PO2 was increased from 41.6 +/- 2.1 to 53.8 +/- 4.0 Torr by HCl infusion (P less than 0.02) associated with a shift in O2-hemoglobin affinity of blood in vivo from 38 +/- 2 to 50 +/- 1 Torr. Cisternal cerebrospinal fluid pH decreased from 7.336 +/- 0.014 to 7.226 +/- 0.027 (P less than 0.005), but cerebrospinal fluid HCO3- concentration was not changed from control (25.4 +/- 1.0 meq/l). These data suggest that there is a functional blood-brain barrier in newborn piglets, that is relatively impermeable to HCO3- or H+ and maintains cerebral perivascular pH constant in the face of acute severe arterial acidosis. (ABSTRACT TRUNCATED AT 250 WORDS)

Effects of altered CSF [H+] on ventilatory responses to exercise in the awake goat.

We investigated the effects of selective large changes in the acid-base environment of medullary chemoreceptors on the control of exercise hyperpnea in unanesthetized goats. Four intact and two carotid body-denervated goats underwent cisternal perfusion with mock cerebrospinal fluid (CSF) of markedly varying [HCO-3] (CSF [H+] = 21-95 neq/l; pH 7.68-7.02) until a new steady state of alveolar hypo- or hyperventilation was reached [arterial PCO2 (PaCO2) = 31-54 Torr]. Perfusion continued as the goats completed two levels of steady-state treadmill walking [2 to 4-fold increase in CO2 production (VCO2)]. With normal acid-base status in CSF, goats usually hyperventilated slightly from rest through exercise (-3 Torr PaCO2, rest to VCO2 = 1.1 l/min). Changing CSF perfusate [H+] changed the level of resting PaCO2 (+6 and -4 Torr), but with few exceptions, the regulation of PaCO2 during exercise (delta PaCO2/delta VCO2) remained similar regardless of the new ventilatory steady state imposed by changing CSF [H+]. Thus the gain (slope) of the ventilatory response to exercise (ratio of change in alveolar ventilation to change in VCO2) must have increased approximately 15% with decreased resting PaCO2 (acidic CSF) and decreased approximately 9% with increased resting PaCO2 (alkaline CSF). A similar effect of CSF [H+] on resting PaCO2 and on delta PaCO2/VCO2 during exercise also occurred in two carotid body-denervated goats. Our results show that alteration of the gain of the ventilatory response to exercise occurs on acute alterations in resting PaCO2 set point (via changing CSF [H+]) and that the primary stimuli to exercise hyperpnea can operate independently of central or peripheral chemoreception.

Effects of amiloride and diethyl pyrocarbonate on CSF HCO-3 and ventilation in hypercapnia.

Amiloride (10(-3) M), a Na+-H+ countertransport inhibitor, infused into the cisterna magna (10 microliter/min for 40 min) of ketamine-xylazine-anesthetized rabbits decreased the cerebrospinal fluid (CSF) HCO3- response to 3 h of hypercapnia [arterial PCO2 (PaCO2) = 60 Torr] by 21.6% (mean delta CSF [HCO3-]/delta PaCO2 0.232 vs. 0.296 mmol.l-1.Torr-1, P less than 0.05). Diethyl pyrocarbonate (DEPC, 10(-3) M), a histidine-blocking agent, infused into the cisterna magna decreased the CSF HCO3- response to hypercapnia by 25.3% (mean delta CSF [HCO3-]/delta PaCO2, 0.230 vs. 0.308 mmol.l-1.Torr-1, P less than 0.02). DEPC is known to inhibit the ventilatory response to hypercapnia (E. Nattie. Respir. Physiol. 64: 161-176, 1986) by a direct effect at the ventrolateral medulla (E. Nattie. J. Appl. Physiol. 61: 843-850, 1986). In this study amiloride had no significant effect on the ventilatory response to hypercapnia. The interpretation is that a Na+-H+ countertransport protein, perhaps with a histidine at a key location, is involved in CSF acid-base regulation and that amiloride appears to have no effects on the chemoreception process. DEPC appears to have effects on chemoreception and on CSF acid-base regulation.