Büttner and Büttner: pH Determination in human erythrocytes 75

J. Clin. Chem. Clin. Biochem.Vol. 27, 1989, pp. 75-79© 1989 Walter de Gruyter & Co.

Berlin · New York

Methods for pH Determination in Human Erythrocytes

By Dietlinde Büttner and Johannes Büttner

Institut för Klinische Chemie I, Medizinische Hochschule Hannover

(Received August 26/November 12, 1988)

Summary: The methods described in the literature for pH determination in human erythrocytes, i.e. the directpotentiometric measurement of haemolysed erythrocytes and the 5,5-dimethyl-2,4-oxazolidinedione (DMO)method, were examined and compared. In spite of careful optimization of the experimental technique, astatistically significant difference of a few hundredths of a pH unit remained between the results of the twomethods. The calculation of the cellular water space of the erythrocytes is suggested äs a possible reason forthis difference; in the DMO method this leads to uncertainty in the determination of the DMO concentrationand therefore in the derived intracellular pH values. For clinical use, the direct pH measurement of haemolysederythrocytes, using the technique described here, is recommended.

Introduction

Three different methods are commonly used for thedetermination of the pH value in erythrocytes (1).Firstly the pH value cän be measured directly bypotentiometry ("potentiometric method"), if theerythrocytes are haemolysed after Separation from theplasma (2). It is also possible to determine the pHvalue on the basis of the pH-dependent partition ofweak electrolytes between cells and plasma, e. g. äs inthe "DMO method", where the 5,5-dimethyl-2,4-ox-azolidinedione (DMO) acts äs a weak acid (3). Finally,the pH-dependent nuclear magnetic resonance spectraof specific compounds within the intact blood cellscan be measured ("NMR method") (4). The DMOmethod is, like all related methods, very time-consum-ing, because other analyses are required in additionto the concentration determinations. The NMRmethod requires calibration by potentiometric pHdetermination, and furthermore involves the use ofNMR apparatus. Therefore, only the direct potentio-metric method is suitable for clinical purposes. Theuse of this method äs a matter of routine makesappropriate control necessary. For checking the pHdetermination of the haemolysate the IFCC referencemethod (5) developed for whole blood can be used.

This does not detect specimen-induced errors, such ästhose that may arise when collecting the erythrocytes.For a comprehensive check, an independent methodshould therefore be employed, and the DMO methodis suitable for this purpose. Both the potentiometricand the DMO method can be carried out simultane-ously on the same specimen material, without theresults being subject to the same sources of error.When the DMO method was first introduced formeäsüring the pH value of erythrocytes, only a fewconiparative studies with the potentiometric methodwere carried out, and these did not show a satisfactoryagreement (6,7). Meanwhile, the analytical techniquesfor both methods have undergone further develop-ment. In the DMO method, the use of 14C-labeledDMO with measurement by liquid scintillation spec-trometry in particular has considerably increased thespecificity and sensitivity. The method permits thedirect DMO determination in cell material, whereaswith the older UV spectrometric method the intra-cellular concentration of DMO had to be derived fromthe difference of whole blood and plasma concentra-tions. In the potentiometric method the reduction ofthe amount of specimen and the availability of largelygas-tight and temperature-resistant collecting tubes(monovette) have improved the reliability.

J. Clin. Chem. Clin. Biochem. / Vol. 27,1989 / No. 2

76 B ttner and B ttner: pH Determination in human erythrocytes

The present study examines whether identical resultscan be achieved with the two methods, using currentlyavailable technology and optimal test conditions, andthus whether the DMO method is suitable for check-ing the potentiometric pH determination in humanerythrocytes.

MethodsGell material , DMO incubation and haemolysatepreparationVenous donor blood is collected in a 25 ml monovette (serummonovette from Sarstedt, N mbrecht, FRG, plastic granulateremoved) with 25 μΐ Liquemin (from Hoffmann-La R che,Basel, Switzerland, 25 · 106 IE/1) and the blood cells are sepa-rated by centrifugation in a centrifuge (heatable Varifuge fromHeraeus-Christ, Osterode, FRG) heated to 37 °C for 5 miftutesat 3500 min"1. The plasma supernatant is transferred into a 20ml disposable syringe from Braun, Melsungen, FRG and theremixed with 50 μΐ [14C]5,5-dimethyl-2,4-oxazolidinedione solu-tion (preparation CFA 575 from Amersham-Buchler, Braun-schweig, FRG, specific activily 1.85 GBq/mmol (50 mCi/mmol),dissolved in 5 ml Ringer9s solution) and 50 μΐ of a solution ofnon-labelled DMO (Sigma, Deisenhofen, FRG), (5 g/l in Rin-ger's solution). The erythrocyte sediment is freed from plasmaresidues and "buffy coat" by means of a water jet from glasscapillary. The erythrocytes are then added again to the autolo-gous DMO-containing plasma and thoroughly mixed (totalconcentration of DMO ca. 0.1 mmol/1, activity ca. 6000 disin-tegrations/min per 100 μΐ).Incubation is carried out for 10 minutes with constant shakingin a tonometry apparatus (8) with two parallel tonometer bails,each filled with approximately 10 ml incubate and gasified with0.06 volumes CO2 in O2. Using a silicone tube, the contents ofthe two tonometer balls are transferred, with continued gasifi-cation, into blood gas monovettes (safety monovette for theblood gas analysis, max. 4.5 ml volume, from Sarstedt N m-brecht, FRG, two per tonometer ball); any air bubbles areexpelled via the Luer opening from the monovettes, and theSuspension is centrifuged at 37 °C (see above). The pH of theplasma supernatant is determined immediately after the cen-trifugation (extracellular pH = pHc). After expelling the resid-ual plasma supernatant and the uppermost erythrocyte layermixed with plasma, a specimen of the cell sediment is taken forthe haematocrit determination (see below) and the remainingcell sediment in the monovette is frozen in liquid N2. After aminimum freezing time of 10 minutes it is thawed and refrozen,since, s phase microscopy has shown, the cells are only com-pletely haemolysed after the second thaw. Directly after prep-aration of the haemolysate the pH is measured, and specimensare subsequently taken from the same material for the deter-mination of the DMO and water contents.

Measuring technique

Measurements of pH with the micro blood System according toAstrup (BMS 2, Radiometer, Willich, FRG) at 37 °C. Calibra-tion with calibrating buffers from Radiometer. Quality controlalternatively with control specimens "Qualicheck" (Radiometer,Willich, FRG) and "Certain" (Corning, Halstead, Essex, UK).In addition the apparatus setting was checked with referencebuffer Solutions of the National Bureau of Standards, Wash-ington (NBS, No. 186-I-c and 186-II-c). The precision in theseries was determined s s = 0.001 (n = 25).

[™C]DMO measwemem with the liquid scintillation counterTricarb 460 CD from Packard, Frankfurt, FRG. Preparationof specimens for scintillation counting: About 100 μΐ plasma

or haemolysate test samples are weighed exactly in glass vials(Packard, Frankfurt, FRG) (the measurement of the volumeof the sample by pipetting with an Eppendorf pipette provedtoo inexact because of the thick consistency of the haemolysate).The haemolysate specimens were digested with Lumasolve(from Lumac, Schaesberg, Netherlands) mixed with isopropa-nol (l + l parts by volume), then decolorized with 0.5 ml perspecimen hydrogen peroxide (perhydrol, 30%, Merck, Darm-stadt, FRG). ScintiUator: Lumagel (Lumac, Schaesberg, Neth-erlands) acidified by the addition of 0.5 m l/1 hydrochloric acid(l volume pari HC1 and 8 volume parts lumagel). Because ofthe great differences in quench between plasma and haemoly-sate Cocktails, an internal Standard was used for quench cor^rection. The counting efficiency was determined individuallyfor each specimen: 10 μΐ [14C]«-hexadecane (1.128 - 106 disinrtegrations/min per g from Amersham-Buchler, Braunschweig,FRG, assigned value = 8720 disintegrations/min per 10 μΐ)were added from a Hamilton syringe. Measurements were madeat least one week after the preparation of the Cocktails. Onlyin this manner was quench stability guaranteed between thefirst and second measiirements of the specimens, allowing fordecay of chemiluminescence and volatilization of the remainingO2. The counting efficiency of the tests lay between 0.7 (hae-molysate) and 0.9 (plasma).Calculation of the intracellular pH value (pH) by using theequation (9) derived from Jacobs (10) (pKa- value for DMO6.13 (3), DMO concentratioiis given s disintegrations/minvalue per gram of haemolysate water (c;) or plasma water (cc)):

pKa + log

Determination of water content by weighing a 100 mg (approx.)specimen before and after treatment in the drying cabinet at80 °C for about 20 hours, followed by drying in a desiccatorover P?O5 to constant weight (deviation from previous weighing0.3 mg maximum).

Haematrocrit determination: Withdrawal of the cell materialafter the centrifugation: a microhaematocrit capillary is filledvia the Luer opening of the monovette. Centrifiigation: Hen>ofuge from Heraeus-Christ, Osterode, FRG, with 12000 min"1.The packed cells had an average haematocrit of 0.96(s = 0.0063, n =,16).

Results

The data and measurernent results of a total of 16tests on blood specimens of healthy donofs are sum-marized in table 1. The pH values and all furthermeasurement data are average values of 4 measure-ments on 4 subspecimens (in each case 2 tests from lof the 2 tonometer fillings of the saine blood speci-men). Due to the standardized gasification, the pHvalues were all in the reference interval for arterialblood (pH between 7.37 and 7.45). The values for theintracellular pH (pHj (potentiometric methods) andpHi (DMO method)) were lower than the associatedpHc values, in accordance with the stronger acid en-vironment in the erythrocytes. The two methods pro-duced comparable results for the pHj. However, theresults of the potentiometric meas rements were reg-ularly slightly higher than those of the DMO method(see tab. 1). The average value of the difference be-tween pH] (potentiometric method) and pHj (DMOmethod) was 0.027 with a Standard deviation of 0.012.

J. Clin. Chem. Clin. Biochein. / Vol. 27,1989 / No. 2

Büttner and Büttner: pH Determination in human erythrocytes 77

Tab. 1 . Results ofthe determination of intracellular pH in human erythrocytes. Potentiometric direct determinalion in haemolysate,and the DMO method applied to the same specimen

TestNo.

123456789

10111213141516

s

pHe Watcr Content*)

7.4147.3797.4187.4247.4007.4187.4507.3757.3837.4027.3977.4047.4207.4107.3837.374

Haemolysate

0.6620.6620.6680.6700.6640.6630.6600.6600.6530.6660.6640.6660.6820.6790.6690.677

0.6670.0075

Plasma

0.9210.9110.9050.9100.9160.9160.9180.9080.9090.9110.9110.9130.9220.9190.9190.921

0.9140.0053

Concentration pHjquotientCi/Cc

0.5850.5970.5780.5900.5910.5900.5750.5680.5620.5830.5720.5780.6000.5930.5940.625

DMO

7.1657.1387.1637.1797.1557.1737.1947.1107.1137.1507.1367.1497.1837.1677.1407.155

potent.

7.1917.1477.1947.2007.1847.2067.2077.1417.1487.1887.1877.1867.1917.1817.1797.169

DifferencepHj(potenL)-pHj(DMO)

0.0260.0090.0310.0210.0290.0330.0130.0310.0350.0380.0510.0370.0080.0140.0390.014

0.0270.012

*) mass fraction

The difference between the average values of the twomethods is statistically significant (p < 0.001 , pairedt-test). The slope of the standardized principal com-ponent model (see tab. 2) was however not signifi-cantly different from l . Whereas occasionally the pHjvalues of the two different methods are almost iden-tical (differences of 0.01 pH units), the deviations inthe extreme case are 0.05 pH units. An explanationfor these differences cannot be deduced from themeasurement results: The measured quantities corre-late only poorly (concentrations quotient Cj/cc or pHgradient pHc— pH») or not at all (water content ofhaemolysate and plasma) with the pHj differences ofthe two methods. As expected, the pHj values of bothmethods show a dependence on pHc, and this de-pendence is linear (correlation coefficient of the linearregression for pHi (DMO method) r = 0.86 and for

j (potentiometric method) r = 0.83).

Tab. 2. Statistical data for comparison of the methods.

Mean valueStand, principal componentmodel

iüterceptslope

Standard error of residuesCorrelation coefficientlinear regression

pH.DMO-method

7.154

Sy.x

r

pHipotentio-metricmethod

7.181

1.0120.86

0.0080.85

Discussion

After careful optimization of the »specimen prepara-tion and of the analytical procedures, the intracellularpH values in human erythrocytes determined with thetwo methods under examination showed only a low,albeit statistically verified average difference of 0.027pH units. Larger differences are reported in the lit-erature (6, 7, 11). The differences found in the presentinvestigations vary with the individual tests, but showsystematically higher pH; values with the potentio-metric measurement than with the DMO method.This raises the question ofthe cause of this systematicdeviation.

In the potentiometric method, the haemolysate is con-taminated by the trapped plasma in the intercellularspace. For a mean pH difference (pHc—pHj) of 0.22and a haematocrit value of 0.96, the increase of thepH value due to the plasma contamination is calcu-lated to be less than 0.001 of a pH unit. This value isin agreement with figures in the literature (12) andlies at the limit öf the sensitivity of the pH measure-ment.In own results, plasma contamination is not a possiblesource of error, because the DMO method was appliedto the same specimen material, likewise without mak-ing allowance for any plasma contamination. A math-ematical correction for the plasma contamination inthe DMO method would result in a somewhat greatereffect because ofthe higher pH difference (pHc-pH,)(mean value 0.25 compared with 0.22 in the potenti-ometric method).

J. Clin. Chem. Clin. Biochem. / Vol. 27,1989 / No. 2

78 Bütlner and Büttner: pH Determination in human erythrocytes

Other interference factors in the potentiometricmethod are:

1) the so-called 'liquid junction potential" (13),2) ghicose breakdown by glycolysis (14) or via thehexose monophosphate shunt (15),3) a CO2 shift during the centrifugation (if the tem-perature of 37 °C is not maintained) (13).

These interference factors cannot always be ruled outwith certainty. However, they do not provide an ex-planation for the observed discrepancies between thetwo techniques, since they would reduce the pH, val-ues, whereas the potentiometrically determined pHjvahies are larger than in the DMO method. Interfer-ence factors possibly increasing the pHj are äs follows:

1) Losses of O2 or CO2 during the specimen prepa-ration (16). In order to avoid a gas exchange of thissort, the tests were carried out under anaerobic cori-ditions in gas-tight collecting tubes.

2) The apparent pHi could also be increased by con-tamination of the liquid junction blood/KCl (17). Toavoid such contamination, the KC1 bridge was rinsedand replaced after each measurement series.

Consequently there was no apparent reason for er-roneous results in the pH determination with thepotentiometric method.

An analysis of the errors of the DMO method wasmade easier by being able to check for and eliminateerrors of a fundamental nature, e. g. permeability ofthe erythrocyte membrane with respect to the disso-ciated DMO component (1), physicochemical inhom-ogeneities through intracellular compartmentalization(18), and a binding of the DMO to plasma (19) anderythrocyte proteins, in particular haemoglobin (12).It can also be assumed that the pKa values for DMOin plasma and cells agree. The ion strengths of serumand erythrocytes differ so little (20) that they give riseto only insignificant differences in the pKa values(Waddell & Butler (3)). Thus the theoretical precon-ditions for correct measurement are fulfilled. With thegreat number of separate determinations which arerequired in the DMO method, the possibility of ex-perimentally induced systemic errors is, however, par-ticularly large, äs Robson et al. (21) have alreadypointed out. The pHi values determined with theDMO method are largely governed by the quotientof the DMO concentratioüs in cells (q) and plasma(ce). It is therefore obvious to look for error sourcesin its formulation. The concentration figures for theDMO must be related to the water space in which

the DMO is distributed. In our calculations, accordingnormal practice, the analysed DMO amount (meas-ured äs disintegrations/min per g specimen) was re-lated to the total water in plasma or haemolysate(here, for reasons of coniparison without deductionof the plasma water, which makes up about 0.04 ofthe haemolysate). In order to detect test-specific fluc-tuations of the watef content (in tue erythrocyte adependence of water content on pHe (22, 23) exists,known äs Gibbs-Donnan effect)* the latter was deter-mined for each test.

In spite of this, errors can arise in relating the DMOamount to the water content of the specimen. It is byno means sure that the water from plasma and cellsis completely available äs a solution space. On thecontrary, results are available which throw doübt onthis, at least for erythrocytes, where part of the cellwater is bound by hydration, in particular by hae-moglobin (24, 25). It is precisely with erythrocytes,which have a relatively low water content, that evenslightjy incorrect estimations of the water space wouldhave considerable influence on calculated concentra-tions. If, äs a result of the presence of bonded water,the osmotically active water space is less than thewater content, then the calculated DMO concentra-tions in the erythrocytes, the resulting concentrationquotient and the pHj values derived from them willbe too low.

In our investigations the DMO method does in factfurnish values which are lower than those from thepotentiometric method. An overestimation of thewater phase with the DMO method may thus be thereason for differences observed between the two meth-ods. Theoretically, such a bias shoüld regularly leadto the same deviation. With the complexity of otherfactors which influence the results of the two methodsto a greater or lesser degree, and can superimposethemselves on this effect, a constant deviation canhardly be expected.

On the basis of these considerations, the DMOmethod for checking the direct potentiometric deter-mination of the pH value in erythrocytes seems to beadvisable only if it is possible to define precisely thewater phase of the erythrocytes äs solution space. Onthe other hand, the deviation obtained after the op-timization of the methods is so small that both meth-ods appeär to be sufficiently reliable for practicalpurposes. Because of its simple Implementation, thedirect potentiometric pH determination of haemoly-sate is the more suitable method for clinical purposes.

J. Clin. Chem. Clin. Biochem; / Vol. 27,1989 / No. 2

Büttner and Büttner: pH Determination in human erythrocytes 79

References1. Roos, A. & Boron, W. F. (1981) Intracellular pH. Physiol.

Rev. 6l, 296-434.2. Purcell, M. K., Still, G. M., Rodman, T. & Close, H. P.

(1961) Determination of the pH of hemolyzed packed redcells from artcrial blood. Clin. Chem. 7, 536-541.

3. Waddell, W. J. & Butler, T. C. (1959) Calculation of intra-cellular pH from the distribution of 5,5-dimethyl-2,4-oxa-zolidine-dionc (DMO). Application to skeletal muscle ofthe dog. J. Clin. Invest. 38* 720-729.

4. Moon, R. B. & Richards, J. H. (1973) Determination ofintracellular pH by 31P magnetic resonance. J. Biol. Chem.248, 7276-7278.

5. Maas, A. H. J., Weisberg, H. F., Burnett, R. W., Müller-Plathe, O., Wimberley, P. D., Zijlstra, W. G., Durst, R. A.& Siggaard-Andersen, O. (1987) Approved IFCC Methods.Reference method (1986) for pH measurement in blood. J.Clin. Chem. Clin. Biochem. 25, 281 -289.

6. Thomason, R. (1963) The use of 5,5-Dimethyl-2,4-oxazo-lidinedione for determination of intracellular pH. Scand. J.Clin. Lab. Invest. 75,45-51.

7. Paymaster, N. J. & Englesson, S. (1966) Calculation of pHof human erythrocyte from the distribution of 5,5-Di-methyl-2,4-oxazolidinedione (DMO). Acta Anaesth. Scand.4, 219-224.

8. Laue, D. (1951) Ein neues Tonometer zur raschen Äquili-brierung von Blut mit verschiedenen Gasdrucken. PflügersAren. Ges. Physiol. 254, 142-143.

9. Büttner, D. & Büttner, H. (1980) pH dependency in uptakeof sulfonamides by bacteria. Chemotherapy 26, 153 — 163.

10. Jacobs, M. H. (1940) Some aspects of cell permeability toweak elektrolytes. Cold Spring Harbor Symp. Quant. Biol.8, 30-39.

11. Minakami, S., Tomoda, A. & Tsuda, S. (1975) Accelerationof red cell glycolysis by citrate due to the elevation ofintracellular pH. In: Internationales Symposium überStruktur und Funktion der Erythrozyten (Rapoport, S. &Jung, F., eds.) Berlin, Akademie-Verlag (pag. 53 — 55, Dis-cussion remark: Minakami pag. 55).

12. Bromberg, P. A., Theodore, J., Robin, E. D. & Jensen, W.N. (1965) Anion and hydrogen ion distribution in humanblood. J. Lab. Clin. Med. 66, 464-475.

13. Siggaard-Andersen, O. (1974) The acid-base Status of ihcblood. 4. Auflage, Munksgaard, Copenhagen, pag. 155 —157.

14. Gleichmann, U., v. Stuckrad, H. & Zindler, M. (1965)Methode zur Bestimmung des intracellulären Säurebasen-haushaltes (pH, pCO2, Standardbikarbonat, Basenüber-schuß) in Erythrocyten. Pflügers Arch. Ges. Physiol. 283,43-55.

15. Warth, J. A., Desforges, J. F. & Stolberg, S. (1977) Intraer-ythrocyte pH, pCO2 and the hexose monophosphate shunt.Brit. J. Haematol. 37, 373-377.

16. Labotka, R. J. (1984) Measurement of intracellular pH anddeoxyhemoglobin concentration in deoxygenatcd erythro-cytes by phosphorus-31 nuclear magnetic resonance. Bio-chemistry2J, 5549-5555.

17. Müller-Plathe, O. (1981) Maßnahmen zur Qualitätsverbes-serung von Blutgasanalysen. Mcd. Lab. 34, 177 — 183.

18. Warth, J. & Desforges, J. F. (1978) Intracrythrocyte pHand physiochemical homogeneity. Proc. Soc. Exp. Biol.Med. 759, 136-138.

19. Waddeü, W. J. & Butler, T. C. (1957) Renal excretion of5,5-dimethyl-2,4-oxazolidinedione (product of demcthyla-tion of trimethadione). Proc. Soc. Exp. Biol. Med. 96, 563 —565.

20. Van Slyke, D. D., Hastings, A. B., Murray, C. D. & Sen-droy, J. jr. (1925) Studies of gas and elcctrolyte equilibriain blood. VIII. The distribution of hydrogen, chloride andbicarbonate ions in oxygenated and reduced blood. J. Biol.Chem. 65, 701-728.

21. Robson, J. S., Bone, J. M. & Lambie, A. T. (1968) Intra-cellular pH. Adv. Clin. Chem. 11, 213-275.

22. Funder, J. & Wieth, J. O. (1966) Determination of sodium,potassium, and water in human red blood cells. Scand. J.Clin. Lab. Invest. 18, 151-166.

23. Gunn, R. B., Dalmark, M., Tosteson, D. C. & Wieth, J.O. (1973) Characteristics of chloride transport in humanred blood cells. J. Gen. Physiol. 61, 185-206.

24. Drabkin, D. L. (1950) Spectrophotometric studies. XV.Hydration of macro sized crystals of human hemoglobin,and osmotic concentrations in red cells. J. Biol. Chem. 1.85,231-245.

25. Savitz, D., Sidel, V. W. & Solomon, A. K. (1964) Osmoticproperties of human red cells. J. Gen. Physiol. 48, 79—94.

Dr. rer. nat. D. BüttnerInst. f. Klinische Chemie lMed. Hochschule HannoverKonstanty-Gutschow-Str. 83000 Hannover 61

J. Clin. Chem. Clin. Biochem. / Vol. 27,1989 / No. 2