Studies on the respiratory response to disturbances

of acid-base balance, with deductions concerning

the ionic composition of cerebral interstitial fluid’

V. FENCL,2 T. B. MILLER,3 AND Ja R. PAPPENHEIMER4 Debartment of Physiology, Harvard Medical School, Boston, Massachusetts

1

FENCL, V., T. B. MILLER, AND J. R. PAPPENHEIMER. Studies on the respiratory response to distwbances of acid-base balance, with de- ductions concerning the ionic cornposition of cerebral interstitial @id. Am. J. Physiol. 2 I o(3) : 459-472, I g66.-Respiratory re- sponses to inhaled (202 were studied on five unanesthetized goats during chronic metabolic acidosis or alkalosis. Measure- ments were made of alveolar ventilation, pH, [HCOa-1, and [Cl-] of arterial plasma and CSF during each steady-state period of CO2 inhalation at each level of acidosis or alkalosis, Net transependymal fluxes of HCOa- and Cl- were deter- mined. Results lead to the following principal conclusions: I) Resting ventilation is a single function of [H*] in cerebral interstitial fluid during inhalation of COZ, perfusion of the ventriculocisternal system with varying [HCOa-1, and during all degrees of chronic metabolic acidosis or alkalosis consistent with life at normal Pao,. 2) Respiratory adaptations to chronic acidosis or alkalosis are accounted for quantitatively by ob- served changes in ion transport between blood and CSF. 3) Concentrations of Hf, HCO~-, and Cl- in cerebral interstitial fluid are equal to those in CSF, even when large differences in concentration between plasma and CSF are maintained by the blood-brain barrier.

the over-all response. The unified theory of Winterstein (35, 36), ascribing all changes of ventilation to changes in [H+] within the central nervous system, gave way to the multiple factor theory advanced by Gray (IO) fol- lowing Nielsen’s (22) demonstration that arterial CO2 and [H+] apparently exerted independent actions on respiration during chronic metabolic acidosis. In recent years the interpretation of respiratory responses in terms of the composition of arterial blood has been questioned, owing largely to the demonstration by Leusen (I 7) that changes in the composition of fluid perfusing the cere- bral ventricles could alter ventilation independently of the composition of arterial blood, Extensions of Leusen’s experiments by Loeschcke and others (x8, 20) have led to the hypothesis that ventilation may be controlled by the [H+] in CSF bathing a chemoreceptive area at the surface of the medulla. This hypothesis has recently been rendered untenable, at least in its simplest form, by our own observations on unanesthetized goats (24) showing that no one of the three related variables in CSF perfusate, Pco~, [HCOS-1, or [H+] was uniquely responsi- ble for regulation of breathing. Large changes of ventila- tion could be produced by varying PCQ and [HCOS-] at constant pH or alternatively by varying Pco? and pH at constant [HCOZ-1. Pulmonary ventilation be- came a single function of pH only when the pH was re- ferred to a locus part way along the concentration gradi- ent of HCOS- between CSF-perfusate and interstitial fluid close to cerebral capillaries taking part in the ex- change of HCOS-. The absolute values of pH or [HCOS-] in interstitial fluid could not be determined from our data, which were consistent with several possible con- centration profiles between CSF and cerebral capillaries. Two hypothetical profiles were discussed which led to unique relations between ventilation and pH; in hypothe- sis A, the normal concentration difference of HC03- be- tween blood and CSF was assumed to be maintained. by an ion pump located at the blood-brain barrier and in this case the ventilation was a single function of intersti- tial fluid pH in the range 7.32-7.24 at all values of [HCOa-] and Pcoz in CSF perfusate; in hypathesis B, the ion pump was assumed to be located at the ependymal

cerebrospinal fluid; regulation of respiration; chronic acidosis, alkalosis; respiration in goat; cerebral HCOa- transport; brain tissue fluid pH; blood-brain barrier

M ECHANXSMS UNDERLYING control of breathing during disturbances of acid-base balance have been a subject of controversy ever since the classical studies of Haldane and Priestley ( I I). The central points at issue concern the uniqueness of molecular CO2 or of [H+] as stimu- lating agents, the locus of stimulation in the CNS, and the relative role played by peripheral chemoreceptors in

Received for publication 4 August 1965. 1 This work was supported by continuing grants from the

American Heart Association and by Grant NB-03-6 15-03 from the National Institute of Neurological Diseases and Blindness.

2 Feilow of the American Heart Association. Present address : Institute for Cardiovascular Research, Prague, Czechoslovakia,

3 Established Investigator, American Heart Association. 4 Career Investigator, American Heart Association.

459

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

460

walls or pial surfaces of the brain and in this case ventila- tion became a single function of interstitial fluid pH in

the range 7.42-7.34 (Fig. 13 of ref. 24). We have now obtained information distinguishing be-

tween these two hypotheses. The information is derived from measurements of HCOS- transport and control of respiration during chronic metabolic acidosis and alkalo- sis, under conditions where various differences in [HCOs-J are naturally created and maintained between CSF and blood. The results show that alveolar ventila- tion is a single function of pH in large cavity Auid during all degrees of chronic acidosis and alkalosis and over a wide range of inspired CO2 concentrations. Thus the ventilation breathing CO2 during severe chronic meta- bolic alkalosis is not significantly different from that breathing air during severe acidosis, provided that the pH in cisternal fluid is the same in both instances. More- over, the relationship between ventilation and pH of cisternal fluid during acidosis or alkalosis is not sig- nificantly different from the relationship between ven- tilation and interstitial fluid pH as determined pre- viously from perfusion experiments, the interstitial pH being calculated on the assumption of an ion pump at the blood-brain barrier (hypothesis A of ref. 24). Combination of these results with those obtained previously leads to the three principal conclusions of the present paper, r) that the effects on respiration of disturbances in acid-base balance are mediated by changes in [H+] of cerebral interstitial fluid, whether these changes be caused by in- haled CO*, perfusion of the ventricles with varying [HCOS-1, or by chronic acidosis or alkalosis; 2) that equilibrium concentrations of H+, HCOS-, and Cl- in cerebral interstitial fluid are approximately the same as those in large cavity fluid, even when large differences are created between blood and CSF during chronic dis- turbances of acid-base balance; 3) that the respiratory response to chronic metabolic acidosis or alkalosis can be explained quantitatively by observed changes in intra- cerebral ion transport.

40 ART.

b3fl

20 mM/L GOAT M5,6f kg

NH4 Cl

GOAT H 5,70kg

DAYS

FIG. I, Typical dose schedule for maintenance of acidosis or alkalosis in the goat. NH&l or NaHCO3 given by rumen tube.

‘ENCL, T. B. MILLER, AND J. R, PAPPENHEIMER

CSF PI-I

,mmHg / , / , / / ,

30 L20 O r l

CSF

20

IO 1 I t 1 I I I 1 I I IO 20 30 40

ARTERIAL [HC05], mM/ kg H,O

FIG. 2. [HCOa-1, pH, and Pcoz in CSF and blood during chronic variations of blood [HCOJ induced by administration of NH&l or NaHCO3. Data obtained by simultaneous anaerobic sampling of arterial blood and cisternal or ventricular CSF in 5 unanes- thetized goats at rest, breathing air. The least-squares line in the bottom panel is: [HC03-jCSF = 11.3 + 0.352 [HC03-lsrt,, In the middle panel the line for arterial Pcoz was fitted to the points by eye. The broken line labeled CSF was obtained from average Pc02 differences between arterial blood and cisternal CSF listed in Table I. To obtain the lines drawn through the points in the upper panel the data were grouped according to acid-base con- dition and the average CSF pH in each condition was plotted against the corresponding average blood [HCOa-1, a smooth line was drawn through these points, and deviations of experimen- tal points from the line were treated statistically. The cross- hatched area includes &I SE, the area between the broken lines &I SD.

METHODS

r) General. Repeated studies of the respiratory re- sponse to inhaled CO2 were made on each of five un- anesthetized goats which were made chronically acidotic or alkalotic by intraruminal administration of NH&l or NaHCOs. Each animal was provided with carotid loops and with permanently implanted guide tubes over the lateral ventricles and cisterna magna. Measurements were made of alveolar ventilation, gas exchange, pH, [HCOZ-1, and [Cl-] of arterial plasma and cisternal CSF

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

CEREBRAL FLUIDS AND RESPIRATION DURING CHRONIC ACID-BASE DISTURBANCES 461

during each steady-state period of COP inhalation at each (chronic) level of metabolic acidosis or alkalosis. Usually four to six periods of CO2 inhalation were stud- ied on each experimental day, the inspired CO2 ranging from o to IO % for normal and alkalotic animals and o to 5 ?& in metabolic acidosis. Net transependymal fluxes of HCOS- and Cl- were measured in two goats during ventriculocisternal perfusion with various concentrations of HCOS- while the animals were in chronic metabolic acidosis or alkalosis.

Operative procedures and techniques for perfusion of the ventriculocisternal system have been described in ref. 25; respiratory measurements and analytical tech- niques in ref. 24; theory and procedures for measurement of CSF formation and transependymal flux rates in refs. 12 and 24. New procedures and modifications of pre- vious techniques are described in detail below.

2) ~roduchn and maintenance of metabolic acidosis or alkalo- sis. Priming doses of NH&l (0.5-0.7 g/kg) or NaHC03 ( 1-2 g/kg) were dissolved in 1-2 liters of warm water, mixed with IOO g of baby cereal, and introduced into the rumen through a stomach tube. The cereal was effective in preventing diarrhea. [HCOS-] in arterial plasma was usually determined twice or more each day and supplementary doses of NH&l or NaHC03 were given as needed to maintain plasma [HCOS-] in the range 8-15 mEq/liter (acidosis) or 35-45 mEq/liter (alkalosis). Ruminants can excrete large amounts of alkali but are poorly adapted for excretion of acid. Usu- ally 0.2 g/kg NH&l, given in a single daily dose, was sufficient to maintain plasma [HCOS-] in the range IO & 2 mEq/liter, whereas daily administration of 2-4 g/kg NaHC03, given in multiple doses, was necessary to

maintain plasma [HCO,] in the range 40 & 5 mEq/ liter. Treatment was continued on each animal for 4-5 days, the experimental measurements being made on the last 2 days of each period of treatment. Typical dose schedules and variation of plasma [HCOs+] are shown in Fig. I.

3) Respiratory measurements. In our previous studies (24) the respiratory inflow was supplied from an aviation demand valve which created undesirably large inspira-

TABLE I. Cisternal-arterial ho2 di$erences during inhalation of narious CO2 mixtures --

Ci sternal-Arterial PCO 2 IX ff erence, mm Hg (Mean ZIZ SE)

Pwo2, mm Hg

Normal* (II goats, I I 7 samples)

Acidosis (5 goats, 26 samples)

Alkalosis (3 goats, I 7 samples)

25-30 II .x&2.4 ro.r*r.2

30-35 10.2~to.6 7 -8zto.g

35-40 8.gko.5 7.8+1.1

40-45 8.gho.6 8.x+0.8

45-50 7.7ho.8 7.6h2.0

50-55 7.gz1.6 7 .g+o.8

* Includes 87 measurements obtained during perfusion of the ventricular systems of 6 goats, published previously as Fig. 5 of ref. 24.

l

Zl t ARTERIAL [ HC03-1, mM/ kg t$O

1 I I 1 I 1 I 1 I

IO 20 30 40

FIG. 3. Variation of arterial pH with arterial [HCO 3-1 in chronic metabolic acidosis and alkalosis. Data from 5 unanes- thetized goats at rest, breathing air. The solid line through the points was fitted by eye. For comparison the plot of average CSF pH taken from Fig. 2 is shown as a broken line. The changes in CSF [H+] are about >i those in blood.

tory pressures when respiratory minute volume ex- ceeded 25 liters/mine For the present work the range of respiratory responses was extended to 50 liters/min and respiratory inflow was supplied from a loo-liter bal- anced spirometer, refilled either continuously or inter- mittently from a high-pressure cylinder. Maximum in- spiratory and expiratory pressures were less than ~5 cm Hz0 and inspiratory volumes measured from the spirom- eter usually agreed with the continuously recorded ex- piratory minute volumes within 2 %.

4) Sampling procedures, measurement of PH. CSF is poorly buffered, and a change of only 0.0 I pH units is asso- ciated with a 20 % change in alveolar ventilation (Figs, g, IO). Traces of acid or alkali in the sampling system or loss of CO2 into minute air bubbles adhering to the barrel of the sampling syringe may cause spurious shifts of pH which are large compared to physiologically sig- nificant variations. The relatively large variations of pH in samples of CSF drawn from the same animal under the same physiological conditions have not been empha- sized by previous investigators. Many published values for CSF pH are improbably high, leading to values of Pcoz corresponding to improbably small or even nega- tive cisternal-arterial differences in Pcoz, particularly when buffering power of CSF is reduced in acidosis. We used the following procedure for sampling and an- alyzing CSF anaerobically. The cistern (or lateral ven- tricle) was punctured through the guide tube with a sharp probe which was left in place throughout the day of experiment; the animals showed no sign of discom- fort. The probe was connected to a three-way stopcock via a 4-cm length of flexible polyvinyl tubing. A g-ml glass, capillary-tipped syringe was rinsed thoroughly with boiled saline and small air bubbles adhering to the sides were removed by tapping+ The syringe was placed in the three-way stopcock with its tip pointed upward and the barrel at the level of the auditory meatus. One milliliter of CSF was allowed to enter the syringe under

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

462 V. FENCL, T. B. MILLER, AND J. R. PAPPENHE1MER

the natural pressure of CSF and immediately discarded through the side arm; 1.5 ml of CSF was then obtained for analysis. A Radiometer capillary glass electrode was fitted with an entrance tube which could be inserted to the center of the sample sy- ringe and whichalso made an airtight seal over the nozzle of the syringe. The sample was discharged into the pH capillary by gentle positive pressure. Usually less than I 5 ‘set elapsed between with- drawal of the sample and final measurement of its an- aerobic pH. Successive pH readings on samples taken from the same syringe often increased by .OI--.02 units in the course of a minute or two, Sampling of both CSF and arterial blood (the latter from an indwelling needle in a carotid loop) was accom- plished without disturbing the animal. The total volume of CSF in a goat is 20-25 ml (25) and the rate of pro- duction of CSF is about o. I 5 ml/min (I 2). Consequently, s-ml sampl es could be with- drawn every hour without fear of causing significant al- terations in the volume or composition of CSF.

Even with these precau- tions, the standard deviation of multiple measurements of CSF pH, made under the same conditions, was &.o I 7 pH units as shown in Fig. 2. On the basis of experience with more than IOO samples of CSF, collected under the most favorable conditions, we conclude that direct esti- mates of CSF pH may be less accurate than indirect estimates based on the meas- ured PCQ of arterial plasma, the gasometric analysis of CSF [HCOL-1, and average values for cisternal-arterial

CO, pressure difference. Table I shows mean values for

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

CERlXBRAL FLUIDS AND RESPIRATION DURING CHRONIC ACID-BASE DISTURBANCES 463

CSF El-],mM /kg+&0

140 r

I IO 1 I I 1 ARTERIAL [Cl-], mM/kg Ii20

I I 1 90 100 I IO 120 130 140

FIG. 4. Variation of CSF [Cl-] associated with variation in ar- terial [Cl-] in chronic metabolic acidosis and alkalosis. Data from 5 goats. Least-squares line is : [Cl-]CSF = 83.9 + 0.398 [Cl-Iart. Concentrations were corrected to a standard 300 milliosmols/kg &O to allow for effects of dehydration during administration of salts. The corrections ranged from o to ;I%.

cisternal-arterial Pcop differences based on analysis of 160 pairs of simultaneously drawn samples of cisternal fluid and arterial blood. Comparison of column I with columns 2 and 3 shows that cisternal-arterial Pcoa differ- ences during acidosis or alkalosis are not significantly different from normal at any given Pace,,.

RESULTS

I. Distribution of ions between plasma and CSF during chronic metabolic acidosis and alkalosis. The steady-state dis- tribution of HCOZ-, H+, and Cl- between plasma and CSF during various degrees of metabolic acidosis or alkalosis are summarized in Figs. 2, 3, and 4, respectively. In Figs. 2 and 3 the degree of metabolic acidosis is indi- cated on the abscissa by the arterial [HCOS-] which, in contrast to the acidity, is affected only secondarily by respiratory compensation. The relation between arterial pH and [HCOX-] is shown in Fig. 3. Detailed values for relevant variables are given for one goat in Table 2, which contains control data and data from repeated periods of acidosis and alkalosis. Similarly detailed data for each of four additional goats are available from a table deposited with the American Documentation Institute. 5

Reference to the bottom panel of Fig. 2 shows that CSF [HCO,] varies linearly with arterial [HCOS-1. There is, however, a considerable degree of regulation of [HCO,] in CSF so that each 3.5 mmolal change in arterial plasma is associated with only a I mmolal change in CSF. The changes in [Cl-] are equal and opposite to those of [HCOS-] as shown in Fig. 4. The CSF [HCOZ-] is about 5 mmolal greater than plasma [HCOS-] during severe acidosis and is 18 mmolal less than that of plasma during severe alkalosis. Most of the data summarized in

TABLE 3. &m;barisan between measured and culculated CSF pH values at various arterial [HC&] z’n chronic metabolic acidosis and alkalosis

Arterial [HCO3-J, rim/kg H20 -

-

Avg measured * Calculatedt

7 a267 7.272 7 -290 7.288

7,303 7.300 7.3’5 7.320

* Taken from Fig. 2. tpJ%m = 6.122 + log [HCQ-ICSF

where [HCQ-] 0.0306 Pcs~co~

CSF is defined by bottom panel of Fig. 2 and PCSFCO~ is defined by Table I and middle panel of Fig. 2.

Fig, 2 refer to cisternal CSF. However, I I of the points refer to ventricular samples taken under normal, acidotic, or alkalotic conditions. The [HCOS-] in ventricular

t from that in cistern al fluid was not significantly differen fluid.

Changes in CSF pH are shown in the upper panel of Fig. 2. The scatter of individual points about the mean is large compared to the total change in pH but is small in terms of absolute value. Means and SE of CSF DH at aver- age arterial bicarbonates of 13, 28, and 40 mLmolal were 7.279 h ,005, 7.303 =t ,005, and 7.325 + .003, respec- tively. The differences between these means are highly significant (P < ,005) and the continuous change of mean CSF pH indicated by the solid central line is ex- actly sufhcient to explain the observed changes of repira- tion as will be shown below. An independent check of the mean values for pH shown in Fig. 2 may be obtained in- directly from the mean [HCOS-] in CSF (lower panel), the mean PacoZ mean values for

shown in the middl l-arterial PC02

e d

panel, and the Xerences based cisterna

on 160 paired measurements summarized in Table I. As shown in Table 3, the mean values for CSF pH calcu- Iated indirectly do not differ by more than 0.005 pH units al. (I

from the mean 9) have recently

values shown in emphasized the

Fig. 2. Mitchell et constancy of pH in

samples of lumbar spinal fluid during moderate acidosis or alkalosis. However, they were not concerned with the significance of small changes such as those shown in Fig. 2 and it is samples in

possible that alkalosis and

their results, based o in acidosis, are

n only three not statisti- five

detailed corn investigators -

.parison will be

cally d ifferent from our own; a of our results with those of pre

more VlOUS

considered in the DISCUSSION. The conclusion, based on Fig. 2 and Table 3, that CSF pH varies slightly but con- tinuously with plasma [HCOJ, is crucial for the sub- sequent development of the present paper,

The arterial Pcoz shown in the middle panel of Fig. 2

is a reciprocal measure of the steady-state alveolar venti- lation, which is increased twofold as arterial [HCo,-] changes from 40 mmolal in alkalosis to IO mmolal in acidosis. The significance of this change in relation to the small change in pH of CSF will become apparent be- low.

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

464 V. FENCL, T. I3. MILLER, AND J. R. PAPPENHEIMER

0 A ;I , NET FLUX, pM/min

tiCOf

ALKALOSIS

I -2

-4

+10

+8

+6

C 0.

CI- ACtDOSlS

NORMAL /

2) NQ%XX of [HCOS-] between CSF and blood. In normal goats the average [HCOa-] is 26 mmolal in arterial plasma, 28 mmolal in cerebral capillary plasma, and only 22 mmolal in cisternal fluid. In our previous paper (24) we showed that the concentration difference of 6 mmolal between capillary blood and CSF is maintained without net flux. Net Aux of HCOS- occurred only when [HCOZ-] of CSF was changed experimentally by per- fusion of the ventriculocisternal system with abnormal concentrations of HCO,. In the present investigation the [HCOB-] in pl asma has been varied over a wide range with associated chronic concentration differences be- tween CSF and plasma ranging from +5 mmolal in acidosis to - I 8 mmolal in alkalosis (Fig. 2). In order to establish the relation between [HCOS-] in CSF and that in interstitial fluid, it is essential to determine whether or not net flux of HCOZ- occurs under these altered condi- tions. If net flux occurs, there will be a concentration gradient in interstitial fluid; if net flux is zero, there will be no concentration gradient in interstitial fluid, as dis- cussed previously (24). Net fluxes of HCO, and Cl- were therefore measured by the technique of ventriculocis- ternal perfusion in each of two goats which were sub- jected to chronic acidosis and alkalosis. Net fluxes were calculated as described previously (24)

li = iri(Ci - Cf) - 6% -I- W(% - Cf>

where fi = net flux (pM/min), V = rate of perfusion (ml/min), c = concentration (mM/liter), CI, = clear- ance of inulin (ml/min), and subscripts i, o, and f refer, respectively, to ventricular inflow, cisternal outflow, and freshly formed CSF. Concentrations of HCO, and Cl- in freshly formed CSF (cf) were assumed to be equal to the values found in CSF (c& under the given condi- tions of acid-base balance. The above equation excludes

Ah-1 mM/kg l-l20

E - c,)

FIG. 5. Transependymal net fluxes of HCO3+ and Cl- meas- ured during ventriculocisternal perfusion with artificial CSF containing various [HCO 3-1 and [Cl-], in each of 2 goats during metabolic acidosis and alkalosis. Net fluxes were calculated ac- cording to the equation given in the text. Detailed data from one goat are presented in Table 4. Fluxes are plotted against differences in concentration be- tween CSF perfusate (C) and capillary plasma (c,) . Lines labeled “alkalosis” and “acido- sis” were drawn by eye through the experimental points. Broken lines, labeled “normal,” were adapted from Fig. IO of ref. 24. Zero net flux, indicated by the intercept with the abscissa, corresponds in each case to the chronic concentration difference of I-KQ- or Cl- between CSF and plasma for each acid- base condition,

net addition of ions to ventricular fluid from choroid plexus secretion and net loss of ions from the system via absorption in bulk through the arachnoid valves; it re- fers specifically to transependymal flux through intersti- tial fluid containing neural elements. Results are sum- marized in Fig. 5, A and B, and detailed data from one animal are given in Table 4.

From Fig. 5A it is seen that zero net flux occurred when CSF [HCO,-] was 6 mmolal greater than in plasma during acidosis and I 8 mmolal less than in plasma during alkalosis. Similarly, zero net flux of Cl- occurred when [Cl-] was 3 mmolal greater than plasma in acidosis and 25 mmolal greater than plasma in alkalosis. Fluxes in normal animals are shown as intermediate lines, drawn without experimental points, but adapted from Fig, IO of our previous publication (24). Reference to Figs. z and 5 of the present work shows that the point of zero net flux corresponds in each case to the resting, steady-state concentration difference of HCOS- or Cl- between CSF and plasma at the stated degree of acidosis or alkalosis. In Fig. 6 the flux rate of HC03- is plotted as a function of the differences between [HCOS-] in perfusion fluid and [HCOS-] which occurred chronically in CSF under each acid-base condition. All of the points fit a common line passing through the origin. We conclude that in chronic conditions there is no net flux and that no gradi- ent of concentration exists in pathways allowing free diffusion of the ions within cerebral interstitial fluid. Net flux occurs only if the composition of CSF is altered to nonequilibrium values by means of artificial perfusion of the ventricular system. The common slope of lines re- lating flux rate to imposed concentration differences (Figs. 5 and 6) indicates that the passive permeability characteristic of barriers interposed between CSF and blood are unaltered during acidosis or alkalosis.

The significant changes in composition of cerebral

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

TABL

E 4.

Net

jhx

of H

CO

,- an

d C

l- in

C

SF

syst

em d

urin

g ch

roni

c m

etab

olic

acid

osis

an

d al

kalo

sis:

G

oat

Ns

A.

MEA

SURE

D Q

UANT

ITIE

S (c

once

ntra

tions

de

term

ined

du

ring

last

15

m

in.

of

indi

cate

d pe

riod)

Con

ditio

n Ti

me,

m

in

Acid

osis

O

-62

1.30

8 1.

441

25.5

23

.5

14.4

14

9.0

144.

7 13

1.0

85.2

69-1

60

1.30

7 1.

428

41.0

36

.5

16.8

13

5.0

134

129.

0 88

.9

165-

208

1,30

4 1.

376

7.4

10.0

11

.9

163.

0 x5

4 12

5.0

93.4

Alk

alos

is

o-55

1.

722

1.70

0 15

.5

17.8

45

.5

153.

5 14

8.0

93.3

94

.3

61-1

09

1.82

3 1.

783

' 34

.8

33.6

43

.7

136.

5 13

4.0

93.3

95

.4

122-

180

1.67

9 1.

679

50.6

46

.6

44.7

12

1.0

120.

5 93

-5

90.3

Con

ditio

n T

ime,

m

in

Aci

dosi

s

Alk

alos

is

O-6

2

69-1

60

165-

208

o-55

1 61

-109

122-

180

Flow

R

ate

ml/m

in

I [H

C03

-l t

m/k

g H

20

Vent

. C

ist.

inflo

w

outfl

ow

iri

it0

Vent

. C

ist.

Art.

Vent

. C

ist,

Art.

Cis

t.

inflo

w

outfl

ow

plas

ma

inflo

w

outfl

ow

plas

ma

outfl

ow

Cm

(%

of

1

co

=a

=i

CO

=a

inflo

w)

B.

CALC

ULAT

ED

QUA

NTIT

IES

Ste

ady-

M

ean

Cer

ebra

l st

ate

vent

. ca

pilla

ry

cone

. C

ont.

cont

. co

ne

. in

CS

F di

ffere

nce

E =D

=C

SF

C-C,

“C

CSF

24.3

16

.6

16*8

+7

.9

+7.4

38.2

18

.9

17.6

+1

9*3

+20.

6

9.0

14.0

15

.6

-5.0

-6

.6

17.0

47

.7

29.1

-3

0.8

-13.

2

34.0

45

.9

28.2

-1

1.9

+5.8

48.1

46

.9

28.8

+1

*2

+19*

3

I IC

I-1

1. m

wkg

H

p fn

ulfn

[cl-]

I'

mM

/kg

H20

Ste

ady-

M

ean

Cer

ebra

l st

ate

vent

. ca

pilla

ry

cone

l C

onc.

co

nt.

cont

. in

C

SF

diffe

renc

e E

cP

%

SF

E-C

P E-

C C

SF

146.

8 12

9 13

6.4

+17.

8 +1

0,4

134.

0 12

7 13

4.0

+7*0

0

157.

3 12

3 13

4.0

+34.

3 +2

3.3

150.

2 91

.9

121.

0 +5

8,3

+29.

0

134.

9 91

.9

121.

0 +4

3.0

+14

*0

121.

4 92

.0

121.

0 +2

9.0

0

Inul

in

clea

ranc

e %

&I

ml/m

in

0.09

2

0.04

6

0.02

0

0.12

8

0.12

9

0.10

5

Net

flu

x /M

/min

-

HC

o3

Cl

+I.0

+3

.8

+2.9

+1

*3

-2.9

+8

.9

-3.1

+6

.7

+1.7

+3

*4

+4.8

+0

.9

by 10.220.33.2 on April 3, 2017http://ajplegacy.physiology.org/Downloaded from

466

fluids brought about by chronic metabolic acidosis and alkalosis involve redistribution of ions to new equilibrium concentrations. A change in acidity of blood causes a change in active ion transport between blood and CSF; net flux due to this altered rate of active transport con- tinues until the entire extracellular compartment (inter- stitial fluid plus CSF) attains a concentration such that the electrochemical gradient causing passive flux exactly balances flux occurring by active transport in the oppo-

HCOZ NET FLUX, HM/min

l / /

9 x” /

/

Ik /

a/

A’ A A [HCO;] mM/kg H,O

x*’ Perf usate- CSFchronie

1 I 1 I ‘x I -I

-10 ,/ 0 +I0 +20 +30

- /‘x X NORMAL 26& I

X A ACIDOSIS 15* I

42 & 5

FIG. 6. Net transependymal fluxes of HCO3- in chronic meta- bolic acidosis, alkalosis, and normal conditions. Fluxes are plotted against the concentration difference between perfusate and

[HCOS-] occurring naturally in CSF at each acid-base condition (CSF chronic). lD t f a a rom normal goats taken from ref. 24. All points fit a common line passing through the origin. Therefore, during normal conditions, chronic acidosis or alkalosis, [HCO 3-1 in large cavity CSF is maintained without net transependymal flUX.

\;rA, L/min, BTPS c 5C

4c

30

20

IO

0

W&r+, mM/kg H,O

A 9-11, X28-33, +35-44

V. FENCL, T. 13. MILLER, AND J. R. PAPPENHEIMER

site direction. The ion pump could be located at the blood-brain barrier or at the ependymal linings of the ventricular system. The respiratory changes considered in sections 3 and 4 below suggest that the location of this pump is at the blood-brain barrier.

3) Ras@atory changes. The experiments of Nielsen (2 2)

showed that during metabolic acidosis in man the COP response curve is shifted to the left of normal, and the experiments of Alexander et al. (e) showed that during metabolic alkalosis the COP response curve is shifted to the right. Figure 7 shows a repetition of these classical experiments on an unanesthetized goat. The data were obtained over a period of 3 months during which the animal was twice made acidotic and alkalotic. Detailed data are given in Table 2; similar data, obtained from four other goats, are included in the table deposited with ADI mentioned previously. As shown in Fig. 7 the COZ response curve during acidosis was shifted 30 mm Hg to the left of that obtained during alkalosis, a shift which is considerably greater than reported by previous investigators who employed milder degrees of acidosis or alkalosis in man.

The same alveolar ventilations shown in Fig. 7 are plotted in Fig. 8 as a function of [H+] in cisternal CSF, and it is clear that the CO2 response curves were a single function of [H+] in large cavity fluid whether the animal was acidotic or alkalotic. The relationship between ven- tilation and CSF pH is nonlinear, especially at low ven- tilations; semilogarithmic coordinates provide a linear description of the results as shown in Figs. g and IO,

Figure IO summarizes respiratory data from the five goats used in the present experiments. Since the animals varied in weight from 32.5 to 70 kg, all ventilations were corrected to a standard resting, steady-state CO2 pro- duction of zoo ml/min STPD, which is the average metabolic CO2 production of animals weighing 35-40 kg (24). For example, if the resting CO2 production of an animal was 300 ml/min STPD, then all ventilations of this animal were multiplied by 3s for inclusion in Fig. g, The data comprise 81 measurements of ventilation with

FIG. 7. Ventilatory response to CO2 inhalation as affected by chronic metabolic acidosis and alkalosis. Alveolar ventilations are plotted against Pace,. Goat Hs, 61 kg. From data in Table 2.

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

CEREBRAL FLUIDS AND RESPIRATION DURING CHRONIC ACID-BASE DISTURBANCES

4c

35

30

25

20

f5

IO

- VA, Lhnin

BTPS

Pc%+ mM/kg Hz0 A A 9-11

x 28-33

l 35-44

CSF pti 7.35 7.30

0 Ll, 7,25 7.?0 7.15 I

4s I I I I

50 55 60 65 70 [t-l+]CSF, nM/kg t-l,0

FIG. 8. Ventilatory response to CO2 inhalation in goat H5 in normal condition and during chronic metabolic acidosis and al- kalosis. The same alveolar ventilations shown in Fig. 7 are plotted against [Hfl in cisternal CSF. From data in Table 2.

simultaneous sampling of arterial blood and CSF (31 controls, 3 I in acidosis, and I g in alkalosis). The range of corrected alveolar ventilations extends from 2.5 liters/

min while breathing air in severe alkalosis to 40 & IO liters/min while breathing CO2 mixtures in all states of alkalosis or acidosis. At the lowest ventilations, obtaining during severe alkalosis, inspired gas i n order to

it was necessarv to maintain normal

add ox ygen to oxygen satura-

tions, The entire 2o-fold range of alveolar ventilation shown in Fig. IO was associated with a pH range of only 0.2 units. Two animals died from hypoventilation during chronic alkalosis; in goats, the maximum chronic CSF pH compatible with life appears to be less than 7.40.

A statistical analysis of results summarized in Fig. IO

is shown in Table 5A which gives least-square equations and correlation coefficients for the best straight lines re- lating log VA to CSF [Hf] for each condition separately. Table 5B summarizes an analysis of variance and shows that the regression coefficients a and b of Table 5A are not significantly different. We therefore conclude that alveolar ventilation of resting goats is a single function of CSF [Hi-] under all conditions of our experiments, namely, at all degrees of chronic acidosis or alkalosis con- sistent with life and while breathing COZ mixtures rang- ing from 0 to I 0 % at normal arterial oxygen pressures.

50 VA, Urnin

BTPS

30-

JO-

8-

6-

4L

CSF pH 7,25 7.20

I 1 I 7.15 l I 55 60 65 70

[H+] CSF, nM/kg Hz0

7.40

4 40

7.35

I 45

7.30

f 50

da FZc. g. Vent2 atory response to CO:! inhalation in goat Hs. Same

.ta as in Fig. 8, but plotted on semilogarithmic coordinates.

Respiratory data for air-breathing periods (I 3 periods of control, 13 during acidosis, and IO during alkalosis) are shown separately in Fig. I I against a background of the over-all response in order to emphasize that the slight changes of CSF pH during chronic acidosis or alkalosis shown in Fig. 2 are quantitatively sufficient to account for the observed respiratory adaptations.

Since there is no net flux of HCO, between large cavity fluid and blood, we may assume that the composi- tion of CSF is identical with that of interstitial fluid with respect to [H+] or [HCO,-] if the ion pump is located at the blood-brain barrier (Iyfiothesis A, ref. 24). In con- trast, if the ion pump were located at the ventricular ependyma (/~yp&esis B of ref. 24) then interstitial [H+] would be close to that of arterial plasma and ventilation would be a separate function of CSF pH at each degree of metabolic acidosis or alkalosis. The single relationship shown in Fig. IO suggests that the ion pump is located at the blood-brain barrier and that the concentrations of HCOS- and H+ in CSF are identical with those in inter- stitial fluid under chronic conditions.

4) hnparkun of respiratory response to changas in blood [HC&-] with responses to alterations in [HCOZ-] of ventriculu- cisternaL perfusion j~id. The relationship between alveolar ventilation and CSF pH under conditions of zero net flux for HCOS- (Figs. 8-1 I) implies that the ionic com- position of interstitial fluid is similar to that of CSF, the concentration gradient to blood being maintained by an ion pump at the blood-brain barrier. This conclusion is consistent with hypothesis A of our previous paper (24). A direct comparison of this hypothesis with our present

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

468 V. FENCL, T. B. MILLER, AND J. R. PAPPENHEIMER

30.

2Ob

3-

- thin, BTPS

ALVEOLAR A - VENTtLATION

(per 200 mt/min CO, production)

A

A’

x 24.6 - 32.2

CSF pH / 7.35 7.30 7.25 7.20 7.15

4 I t I I

I I I I I 1 1 40 45 50 55 60 65 70

[I-I+] CSF, nM/kg Ii,0

FXG. IO. Alveolar ventilation vs. CSF pH in various chronic acid-base conditions. Eighty-one steady-state measurements in 5 goats (3 I controls, 3 I acidosis, I g alkalosis) while breathing O--I 0% COP. Alveolar ventilations corrected to standard CO2 production of 200 ml/min, STPD. The best fit line for the pooled data is shown, See Table 5A for statistical analysis.

Intercept a Ype

Correlation Coefficient

Control -2.14 0.0571 0.886 Acidosis -1.88 o .0526 0.802 Alkalosis -2.23 0.0612 0.810

Pooled data - I .8g 0.0531 O*99O

B: analysis of variance fur testing equality of regression coe$icients (30)

Degrees of Sum of the Mean Residual Due to Freedom Squares Squares

Deviation from null hypothesis Separate Common

regressions regression

F= o.oq28/o.oqg = I.‘791

.25 > P > - IO

4 ‘75 79

0.1712

I .7g28 I .g64o

o. 0428 0.0239

result is shown in Fig. I z. It is clear from this comparison that the relationship between ventilation and CSF pH during chronic acidosis and alkalosis is not significantIy different from the relationship between pulmonary ven- tilation and the pH of interstitial Auid during perfusions of the ventriculocisternal system with varying [HCOz+], it being assumed that the ion pump is located at the blood-brain barrier and that neuronal elements respon-

sive to pH are located three-fourths of the distance along the functional concentration gradient for HCOS- be- tween CSF perfusate and interstitial fluid at the site of ion exchange with cerebral blood. A further quantitative test of this hypothesis was obtained by perfusion of the ventricular system with low [HCOS-] during chronic metabolic alkalosis and by perfusion with high [HCOS-] during acidosis (Table 6). Under these conditions the pH of the perfusate -varied from 7* I 3 to 7.53, extremes which would be incompatible with life of the animal if they referred to fluid surrounding respiratory neurons. In contrast, the pH of interstitial fluid, calculated from hypothesis A, remained within the range 7.28-7.36 and the ventilation in each case corresponded to that pre- dicted from Fig. 12. These results emphasize the con- clusion that neural elements responsive to [H+] are lo- cated at some distance from the large cavities.

DISCUSSION

I> Comjmifion of CSF during ch.ronic acidosis and alkalosis. There is now a large literature reviewed by Mitchell et al. (19) describing the regulation of [HCOS-] in CSF during disturbances of acid-base balance in humans. A summary of the human data is shown in Fig. 13 for comparison with the present data on goats shown in Fig. 2. CSF [HCO,l in goats is 3-5 mmolal less than in humans at any given plasma [HCOB-1, and regula- tion of CSF [HCO,] is slightly more efficient in the goat, as evidenced by the smaller slope. The changes of

ho9 are also about the same in man and goat, thus indicating that respiratory adaptations to chronic acido- sis or alkalosis are similar. The important difference between our conclusions and those of previous investiga- tors concerns the regulation of CSF pH. Our results show that CSF pH changes continuously as a function of plasma [HCO,]; the change is very small in absolute terms but it is statistically significant and of precisely the right magnitude to explain the observed respiratory adaptations. In contrast, previous investigators have re- ported that CSF pH remains unchanged during chronic metabolic disturbances of acid-base balance (cf. Mitchell et al. (I g) for review). We believe that this discrepancy is only an apparent one which derives from the technical difficulty of obtaining meaningful direct pH values in samples of CSF, as discussed above in the METHODS sec-

tion. Direct measurements of CSF pH were reported by each of the nine groups of investigators listed on Fig. I 3, and it is true that a mass plot of their data (not shown in Fig. 13) reveals no statistically significant alteration of CSF pH as plasma [HCOS-] varies from IO to 40 mmolal. During acidosis (plasma [HCOS-] = 5-15 mmolal) the mean CSF pH was 7.326, SD & 0.065; during alkalosis (plasma [HCOS-] = 29-40 mmolal) the mean CSF pH was 7.31 I, SD III 0.015. The standard deviation of 0.065 pH units during metabolic acidosis is larger than the entire change of mean pH found in our experiments on goats (Fig. 2) and is perhaps an expression of the diffi- culty of obtaining reproducible pH measurements in the

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

CEREBRAL FLUIDS AND RESPIRATION DURING CHRONIC ACID-BASE DISTURBANCES 469

50

40

30

20

15

IO

8

6

4

2

i)A L/Min

BTPS

AIR BREATHING

c PERIODS

& S.E.M.

CSF pH 7.20 7AS

I I I I I I I I IL

45 50 55 60 65 70

CSF [H+], nM/kg H,O

FIG. I I. Resting steady-state alveolar ventilation (corrected to standard CO2 production of 200 ml/min, STPD) in relation to CSF (H+) in alkalosis, acidosis, and normal conditions and compared with the over-all respiratory response. The air-breathing periods are presented as the mean values I&EM of VA and CSF [H+] for each acid-base condition. Mean values are significantly different from each other and the slope of the plot of VA vs. CSF [H+] breathing air is not different from that of the over-all response to inhaled COY. The over-all respiratory response is represented statistically as the best fit line for the pooled data shown in Fig. IO. The g$& confidence interval (C.I.) indicates that there is but I chance in 20 that the true relation between log v* and CSF [H+] lies outside of this interval.

poorly buffered CSF. A further complication in experi- ments on humans involves the fact that CSF samples are drawn by lumbar puncture. The Pcoz in these samples will be set by the ratio of metabolism to blood flow in the spinal cord and this may not be the same as in ven- tricular or cisternal fluid, especially during changes of cerebral blood flow induced by the altered PcoB of acido- sis or alkalosis. A discrepancy of only 2-3 mm Hg in CO2 pressure between lumbar and ventricular fluid would cause a relatively large discrepancy in the pH.

An alternative estimate of CSF pH in humans is shown in the top panel of Fig. I 3 which depicts the mean CSF pH calculated indirectly from the measured CSF [HC03-] of the lower panel and the mean jugular-bulb blood PCO~ of the middle panel. The data for jugular COz were obtained from references (8, I 5, 16, 23, 26, 28, and 34) and it is assumed that the jugular Pcog is equal to that in cisternal fluid as shown by Bradley and Semple (4). Th e errors in determination of CSF [HCOS-]

50

40

30

20

t5

IO

8

6

4

3

- 0~ , L/min

BTPS

-

7.40

acidosis or al kolosis

riculo- cisternal

PH ‘7.35 7.30 7.25 7.20 7.15

1 I

45

I I

50

I I I I I ’

55 60 65 70

tH’3, nM/kg HP

I

40

FIG. 12. Respiratory response to inhaled CO2 during chronic disturbances of acid-base balance and during perfusion of ven- triculocisternal system with various bicarbonate concentrations. Alveolar ventilations corrected to a standard COP production of 200 ml/min, STPD. The shaded area includes +z SEM of 81 meas- urements of ventilation on 5 goats during inhalation of O-IO% CO2 with chronic arterial [HCOa-] ranging from g to 44 mmolal. The abscissa for the shaded area refers to direct measurements of pH on samples of cisternal or ventricular CSF. Crosses (X) are derived from measurements on 8 goats breathing o-5% CO2 dur- ing ventriculocisternal perfusion with [HCOa-] ranging from 7 to

40 mmolal and with chronic arterial [HCOZ-] of x0-45 mmolal. The abscissa for the crosses refers to pH of interstitial fluid in equi- librium with Pco2 of CSF and having a [HCOa-] corresponding to a point x of the distance along the concentration gradient for HCOS- in interstitial fluid and defined by

[HCO 3-1 interstitial = x [HCO 3-]ckronio, CSF + >~[HCO3-I,c~f~sat~

as required by hypothesis A of our previous paper (24). The data include the 7 measurements on 2 goats in acidosis and alkalosis shown in Table 6 as well as values derived from 6 normal goats and published previously as Table 4 of ref. 24.

and jugular PCO~ are small and the mean values so ob- tained should provide a relatively reliable estimate of mean CSF pH. It is clear from Fig. r3 that CSF pH, calculated indirectly, varies as a continuous function of plasma [HCOZ-] in humans as it does in goats.

The chronic equilibrium conditions described above are not comparable with the effects of acute administra- tion of HCl or NaHC03. The small, transient net flux of HCO, through the blood-brain barrier which follows a change in blood [HCOS-] produces a rapid change in the composition of fluid surrounding neurons in the vicinity of capillaries, owing to the small volume of in- terstitial Auid relative to the flux. In contrast, the same

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

470

7.3 5

7.30

7.25 t

CALCULATED

CSF pH

50-

40-

30-

Pco* 1 mm Hg

CALCULATED JUGULAR (= CtSTERNAL), 0 0 0

0

D

ARTER t AL

CSF WI mM /Kg Hz0

0 0 -b 30 1 116 ,*' fl 'x

20

t

+A+

+i* , , ,

=g, pLD # 0 , x

/

Do * PAULI ET AL, 1962 *

*o” A* A AGREST ET AL, 1962

0 0 POSNER ET AL, 1965 A MITCHELL ET AL. 1965

I ARTERIAL [HCO,], mM/Kg H20 h 30

20

IO

FIG. 13. Chronic nonrespiratory disturbance of acid-base bal- ance in human subjects. Data are presented for comparison with Fig. 2. Lines relating CSF [HCOS-] and arterial Pcoz to arterial [HCOS-] were fitted to the points by eye. The line for “calcu- rated jugular ( = cisternal)” Pco2 in the middle panel was con- structed from the equation

APCO~ (jugular-arterial) = 18.0 - 0.22 Pacoz

This equation is the best straight-line fit to data published in refs. 8, 15, 16, 23, 26, 28, and 34. CSF pH in the top panel is de- fined by

CSF pH = 6.126 + log [HCQ-ICSF

0.03 I4 PC02 (jugular)

The identity of jugular and cisternal Pco established by Bradley and Semple (4).

2 in humans has been

small flux rate entering the large cavities produces very slow changes of concentration. We may expect the pH of interstitial fluid to change almost immediately, thus altering respiration and causing inverse changes of Pcop, which in turn produce the well-known transient, para- doxical shifts in the pH of large cavity fluid.

The chronic distribution of concentrations of H+, HC03-, and Cl- (Figs. 2, 13) does not represent the electrochemical gradients. The electrical potential differ- ence between CSF and plasma is normally about 5 mv but is increased to 15 mv in severe acidosis and is de- creased to near zero in alkalosis (13). These changes in

V. FENCL, T. B. MILLER, AND J* R. PAPPENHEIMER

potential, first demonstrated in acute experiments (33), have recently been shown to persist in chronic acidosis or alkalosis (9) The potential changes are sensitive to pH of blood but not to changes in pH of fluid perfusing the ventricular system (x 3) ; it seems reasonable to relate them to the altered distribution of ions across the blood- brain barrier described in the present paper, but their precise significance in terms of specific ion transport re- mains to be determined.

2) Dependence of ventilation on interstitial fluid PH. The data summarized in Fig. I 2 indicate that the relationship between ventilation and [H+] in interstitial fluid during chronic acidosis and alkalosis is identical with that which obtains in perfusion experiments when the [H+] is referred to a point three-fourths of the distance along the functional concentration gradient for HCOZ- from the perfusate to the site of ion exchange with blood. This unique relation between ventilation and interstitial fluid pH appears to be valid over a 2o-fold variation of alveolar ventilation induced by inhalation of O-IO % COP, per- fusion of the ventriculocisternal system with [HCOS-] varying from 5 to 45 mmolal, and during all degrees of chronic acidosis or alkalosis compatible with continued life of the animal. The results provide quantitative sup- port for Winterstein’s Yeaction theory” (36). The mul- tiple-factor theory proposed by Gray (I o) appears to be unnecessary, at least with regard to disturbances of acid-base balance. It follows, also, that the peripheral chemoreceptors play no significant role in the respiratory response to acid-base disturbances, a conclusion which is in accord with recent studies by Katsaros (x4) on respiratory response to acid-base disturbances following denervation of the carotid and aortic bodies. Mitchell et al. (19) were forced to the conclusion that peripheral chemoreceptors contribute to the respiratory response because they could detect no significant change in CSF pH during chronic metabolic acidosis or alkalosis. In the interpretation of previous work it is also important to emphasize that the small but significant changes in CSF pH reported in the present paper would be incon- sequential with respect to the respiratory responses of anesthetized animals but are quantitatively sumcient to account for the observed responses in unanesthetized animals.

3) Tunic composition of cerebral interstitial&id. The func- tional characteristics of excitable tissues depend, in part, on the ionic composition of extracellular fluid. It is therefore of great interest to determine whether cells of the central nervous system are exposed to a simple ultra- filtrate of plasma, as in tissues generally, or whether their normal fluid environment is similar to the special composition of fluid found in the ventricles and sub- arachnoid spaces. The brain parenchyma has no lym- phatic drainage and it is impossible to collect tissue exudates or capillary ultrafiltrates for direct analysis. Indirect methods, commonly employed for analysis of volume and composition of peripheral interstitial fluids, are generally inapplicable to brain because of the special permeability characteristics of the blood-brain barrier.

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

CEREBRAL FLUIDS AND RESPIRATION DURING CHRONIC ACID-BASE DISTURBANCES

TABLE 6. Resting alveolar ventilation and calculated interstitial [H+] : ventricu/ocisternul her fusions during chronic metabolic acidosis and alkulosis

1 J

47=

Goat Condition

N5 Acidosis

Alkalosis

R5 Acidosis

<:A, Ii ters/min per 200 mI/min CO2 Production

4.3 8.1

4-J 3.4 3.4

9.1 4.6

[HC03-] m?n/kg Hz0

Arterial

16.8 II .8

46.8 43-3 44.3

10.4 IX .o

Inter- stitial*

17.8 33.6 46.6

24.9 28.4 31.8

13.6 19.2

PWOP mm Hg

CSF Pco2J mm Hg

47 32

51 59 60

pH (IH+], nM/kg HnO)

Interstitial$

7.306 (49.5) 7.284 (52 .o>

7,325 (47.4) 7.179 (66.4) 7-3’3 (43-d 7,393 (40*5) 7,353 (43J) 7-523 bg.6)

7.273 (5” -3) 7.317 143.2)

outflow

7.530 (29.7) 7.131 (74.0)

7, I’7 (74.3) 7,531 (29.4)

* [HCO3+]interqtitial = x [HCOI%SF, cllronic + X CHCO3-Ioutflow; [HC03-1CSF, cllronic = 0.352 CHCo~-lart. + 11*276*

from Table I. j: PHintcrstitisl = 6.122 + log [EJCO3-] interstitial

l

0.0306 CSF-Pco2

t Taken

Nicholls and Kuffler (21) have devised an elegant method for analyzing the ionic composition of interstitial fluid in the avascular central nervous system of the leech. In essence, they used the resting membrane po- tential and the overshoot of the action potential for the biological assay of [K+] and [Na+] in fluid surrounding neurons within leech ganglia. Their results indicated that ions penetrate rapidly from fluid surrounding a ganglion to neurons located 30-50 p within the ganglion; it may be inferred that the interstitial fluid normally has the same ionic composition as the surrounding hemo- lymph.

Results described in the present paper suggest that the respiratory response of the whole animal may be used as a sensitive biological assay for [H+] in interstitial fluid. The [H+] in turn defines the [HCOS-1, since Pcoz in interstitial fluid is identical with that which can be measured in cisternal CSF (4, 24). Finally, the [HCOS-] defines the [Cl-] since the latter is the only significant anion other than HCOS-. If this reasoning is correct, we can conclude from our results that the concentrations of

REFERENCES

I. ,~GREST, A., E. ROEHR, AND A. RUE-GUI~AZU. Sgure-Basen- Gleichgewicht des Liquor cerebrospinalis und des arteriellen Blutes bei Niereninsufizienz. Wirkung der extrakorporalen Hgmodialyse. Klin. MWzscizr. 40 : I 04.51 I 048, I 962.

2. ,~LEXANDER, J. K., j. 31. IvEST, J. A. IvOoD, AND D. w. RICHARDS. Analysis of the respiratory response to carbon dioxide inhalation in varying clinical states of hypercapnia, anoxia, and acid-base derangement. J. Clin. hued. 34 : 5 I I -

532, ‘955. 3. BEKAERT, J., AND G. DEMEESTER. Influence of the potassium

concentration of the blood on the potassium level of the cere- brospinal fluid. Ex@. Med. Surg. I a : 480-501, 1954.

4. BRADLEY, R. D., AND S. J. G. SEMPLE. A comparison of cer- tain acid-base characteristics of arterial blood, jugular venous blood and cerebrospinal fluid in man, and the effect on them of some acute and chronic acid-base disturbances. J. Physiol., London 160: 381-391, 1962.

5, COWIE, J., A. T. LAMBIE, AND J. S. ROBSON. The influence of extracorporeal dialysis on the acid-base composition of blood and cerebrospinal fluid. Clin. Sci. 23 : 397-404, x962.

H+, HC03-, and Cl- in interstitial fluid surrounding respiratory neurons are the same as those in CSF, even when large concentration differences between plasma and CSF are created by alkalosis or acidosis.

Similarly large concentration differences or ratios can be induced chronically in the case of Kf (3). Recent studies by Cserr (6) indicate that the permeability of ependymal walls to K+ is very large compared to the permeability of the blood-brain barrier, thus implying that the concentration of Kf in cerebral interstitial fluid is close to that in CSF, even when large concentration differences are created between plasma and CSF.

These considerations lead us to the view that the steady-state ionic composition of fluid bathing neurons in the central nervous system is close to that in the large cavities and subarachnoid spaces.

We thank Dr. Jane Worcester of the Harvard School of Pub-

lic Health for help with the statistical analysis of Table 5. It is a pleasure also to acknowledge the technical assistance of James Nicholl, Jr.

6.

8.

9.

10.

Il.

12.

CSERR, H. Potassium exchange between cerebrospinal fluid, plasma, and brain. Am. J. Physiol. 209: 12x9-1226, 1965. FISHER, V. J., AND L. C. CHRISTIANSON, Cerebrospinal fluid acid-base balance during a changing ventilatory state in man. J. Ap@. Phy:,iol. 18: 712-716, 1963. GIBBS, E. I;., W. G. LENNOX, L. F. Nrm, AND F. A. GIBBS. Arterial and cerebral venous blood. Arterial-venous differ- ences in man. J. Riol. Chem. 144: 325-332, Ig42+ GOODRICH, C. Effect of chronic acidosis and alkalosis on rat CSF-blood potential. Physiologist 8 : I 78, I 965. GRAY, J. S. The multiple factor theory of the control of respir- ation. Science x03 : 739-744, I 946. HALDANE, J. S., AND J. G. PRIESTLEY. The regulation of the lung-ventilation. J. Physiol., London 32 : 225-266, I 905. HEISEY, S. R., D. HELD, AND J. R. PAPPENHEIMER. Bulk flow and diffusion in the cerebrospinal fluid system of the goat. Am. J. Physiol. 203 : 775-781, 1962. HELD, D., V. FENCL, AND J* R. PAPPENHEIMER. Electrical potential of cerebrospinal fluid. J. Neurophysiol. 27 : 942-959, 1964.

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from

472 V. FENCL, T. B. MILLER, AND J. R. PAPPENHEIMER

14.

I5*

16.

=7*

l9*

20.

21.

22.

23.

KATSAROS, B. Die Rolle der Chemoreceptoren des Carotis- gebiets der narkotisierten Katze ftir die Antwort der Rtmung auf isolierte Aenderung der Wasserstoffionen-Konzentration und des COs-Drucks des Blutes. Arch. Ges. Physiol. 282 : 157-

178, 1965. KENNEDY, C., AND L. SOKOLOFF. An adaptation of the nitrous oxide method to the study of the cerebral circulation in children; normal values and cerebral metabolic rate in child- hood. J. C/in. havest. 36 : I 130-1137, 1957. KETY, S. S., AND C. F. SCHMIDT. The effects of altered arterial tensions of carbon dioxide and oxygen on cerebral blood flow and cerebral oxygen consumption of normal young men. J. Clin. Invest. 27 : 484-492, 1948. LEUSEN; I. R. Chemosensitivity of the respiratory center. In- fluence of changes in the H+ and total buffer concentration in the cerebral ventricles on respiration. Am. J. Physiol.

176: 45-51, 1954. LOESCHCKE, H. H., H. P. KOEPCHEN, AND K. H. GERTZ. uber den Einfluss von Wasserstoffionenkonzentration und COZ- Druck im Liquor cerebrospinalis auf die Atmung. Arch. Ges. Physiol. 266 : 569-585, 1958. MITCHELL, R. A., C. T. CARMAN, J* W. SEVERXNGHAUS, B. W. RICHARDSON, M. M. SINGER, AND S. SHNIDER. Stability of cerebrospinal fluid pH in chronic acid-base disturbances in blood. J. A#. physiol. go: 443-452, 1965. MITCHELL, R. A., H. H. LOESCHCKE, W. MASSION, AND J. W. SEVERXNGHAUS. Respiratory responses mediated through superficial chemosensitive area on the medulla. J. Appl. Physiol. 18: 523-533, 1963. NICHOLLS, J. G., AND S. W. KUFFLER. Extracellular space as a pathway for exchange between blood and neurons in the central nervous system of the leech: ionic composition of glial cells and neurons. J. Neuro@zysioZ. 27 : 645-67 I, I 964. NIELSEN, M. Untersuchungen tiber die Atemregulation beim

Menschen, besonders mit Hinblick auf die Art des chemischen Reizes. Skand. Arch. Physiol. 87 : Suppl. x 0, x 936. NOVACK, P., H. A. SHENKIN, 3;. BORTIN, B. GOLUBOFF, AND A. M. SOFFE. The effect of carbon dioxide inhalation upon the cerebral blood flow and cerebral oxygen consumption in vascular disease. J. C/in. Invest. 32 : 696-702, 1953.

24.

25.

26.

28.

29.

30.

31.

32.

33,

34.

35.

36.

PAPPENHEIMER, J. R., V. FENCL, S. R. HEISEY, AND D. HELD. Role of cerebral fluids in control of respiration as studied in unanesthetized goats. Am. J. Physiol. 208 : 436-450, 1965. PAPPENHEIMER, J. R., S. R. HEISEY, E. F. JORDAN, AND J. DEC. DOWNER. Perfusion of the cerebral ventricular system in unanesthetized goats. Am. J. Physiol. 203 : 763-774, 1962. PATTERSON, J. L., A. HEYMAN, L. L. BAT.I‘EY, AND R. W. FER-

GUSON. Threshold of response of the cerebral vessels of man to increase in blood carbon dioxide. J. Clin. Invest. 34: x857-1863,

I955* PAULI, H. G., C. VORBURGER, AND F. REUBI. Chronic derange-

ments of cerebrospinal fluid acid-base components in man.

J. A@. Physiol. I 7 : 993-998, 1962. PIERCE, E, C., JR., C. J. LAMBERTSEN, S. DEUTSCH, P. E.

CHASE, H. W. LINDE, R. D. DRIPPS, AND H. L. PRICE. Cerebral

circulation and metabolism during thiopenthal anesthesia and

hyperventilation. J. C&n. Invest. 41 : 1664-167 I, rg6z. POSNER, J. B., A. G. SWANSON, AND F. PLUM. Acid-base

balance in cerebrospinal fluid. Arch. hhrol. I 2 : 479-496, 1965. RAO, C. R. Advanced Statistical Methods in Biometric Research.

New York: Wiley, 1952, p. I 12-117. SCHWAB, M. Das SZure-Basen-Gleichgewicht im arteriellen

Blut und Liquor cerebrospinalis bei chronischer Nierenin-

sufizienz. Hin. Wochschr. 40 : 765-772, I 962. SEVERXNGHAUS, J. MT., R. A. MJTCHELL, B. W. RICHARDSON, AND M. M. SINGER. .Respiratory control at high altitude sug-

gesting active transport regulation of CSF PH. J. Appl.

Physiol. 18: I 155-1166, 1 g63- ~SCHIRGX, R. D., AND J. L. TAYLOR. Slowly changing bio-

electric potentials associated with the blood-brain barrier.

Am. J. Physiol. I g5 : 7-22, 1958.

WASSERMAN, A. J., AND J. L. PATTERSON, JR. The cerebral

vascular response to reduction in arterial carbon dioxide ten- sion. J. Clin. Invest. 40 : 1297-1303, 1961.

WINTERSTEIN, .H. Die Regulierung der Atmung durch das

Blut. Arch. Ges. Physiol. 138: 167-184, 191 I, WINTERSTETN, H. Die Reaktionstheorie der Atmungsregula-

tion. Arch. Ges. Physiol. 187 : 2~2-2~8,

by 10.220.33.2 on April 3, 2017

http://ajplegacy.physiology.org/D

ownloaded from