Protein determination in cerebrospinal fluid by protein dye-binding assay

Abstract: In this study, Coomassie brilliant blue (CBB) and pyrogallol red/molybdate (PRM) protein dye-binding assays for total protein determination in cerebrospinal fluid (CSF) are compared. Using human albumin (HA) as a protein calibrator, protein concentration in CSF samples (n = 73) ranged from 55-1960 mg/L (median: 315 mg/L) with the CBB assay, and from 95-2450 mg/L (median: 395 mg/L) with the PRM assay. Linear regression analysis indicated yCB = 0.824xp"' - 8 (r = 0.99). The discrepancy between the values was investigated by comparing the response of the two assays to different proteins. Compared with HA, the PRM assay showed a more uniform response to human albumin/globulin (A/G) and bovine gamma globulin (G) than did the CBB assay, but it gave high colour yields with bovine myelin basic protein. When CSF was assayed using A/G as a protein calibrator, agreement between the methods improved (yCB = 0.960xP' + 0 [r = 0.99]), indicating that comparability is dictated by the choice of protein calibrator. Of the two assays studied, the PRM assay is recommended for CSF protein determination because it gives a more uniform and linear response to human albumin and globulin over a wider working range.

Key words: Cerebrospinal fluid. Coomassie brilliant blue. Protein assay. Pyrogallol red.

Introduction

Determination of the total protein content in cerebrospinal fluid (CSF) is of clinical importance because raised protein levels are associated with neoplasm, infectious disease, and inflammatory or traumatic disorders that result in impairment of the blood-brain barrier or intrathecal synthesis of immunoglobulins.'

Total protein content can be determined by biuret assay,2M Lowry assay,', turbidimetry (using trichloroacetic acid [TCA],'8 sulfosalicylic acid [SSA]7,8 or benzethonium chloride."') or precipitation with trichloroacetic acid/Ponceau S (TCA/PS).2 Alternative methods include nephelometry (using polyethylene glycol" or benzalkonium chloride") and assays based upon bicinchoninic acid," copper/bathocuproine 16 and silver-binding. 17

However, the Coomassie brilliant blue (CBB) protein dye-binding assay, which measures the shift in absorbance maxima when CBB binds protein, currently is the most widely recommended method for protein determination in CSF.**zb The pyrogallol red (PR)/molybdate (PRM) protein assay is an alternative dye-binding assay that monitors the shift in absorbance, from 460 to 600 nm, when PR binds protein.27 This assay was developed for total protein determination in urine,27-l9 and is recommended in preference to the CBB assay."

Here, we report the results of a comparative evaluation of the CBB and PRM assays to detect protein in CSF, a comparison not published previously.

Materials and methods Materials

CBB protein dye reagent (catalogue no. 610-A), Microprotein PRTM dye reagent (catalogue no. 61 I-A), human albumin (HA) protein calibrator (catalogue no. 610-50), human albumin/globulin (A/G) protein calibrator (catalogue no. 540-10) and bovine myelin basic protein (MBP, catalogue no. M1891) were purchased from Sigma Diagnostics Inc. (SigmaAldrich Co. Ltd., Poole, Dorset, UK). Bovine gamma globulin (G) (protein calibrator I; catalogue no. 500-0005) was purchased from Bio-Rad Laboratories Ltd. (Hemel Hempstead, Herts, UK). HA protein calibrator was used as supplied (500 mg/L). A/G, MBP and G were diluted to 500 mg/L in 0.15 mol/L sodium chloride (NaCI), containing 1 g/L sodium azide.

Samples

Cerebrospinal fluid: Residual samples of clinical specimens (n = 73), collected without preservative and centrifuged at 4500 rpm for 10 min, were stored at -70*C. The samples were thawed at room temperature, gently vortex-mixed and dispensed for protein assay using a 5-40 liL Finnpipette (Labsystems, Helsinki, Finland), with performance precision of 20.06 0.05 gL (coefficient of variation, %CV: 0.25). A small number of samples (n = 8) were tested both fresh and following overnight freezing at -70*C, using both protein assays.

Protein assay

CBB protein dye-binding assay: CSF (20 ILL) or protein calibrator (5-20 lL; HA, A/G, G or MBP at 500 mg/L), in a sample volume of 20 gL (adjusted with saline - 0.15 mol/L NaCI, containing 1 g/L sodium azide), were mixed gently with protein assay solution (1 ml) (protein dye reagent diluted with four volumes of ultrapure water in a plastic container), according to the manufacturer's instructions. After 10 minutes, absorbance values were measured (A 595 nm) in a plastic microcuvette (1 cm) using a Jenway 6100 spectrophotometer (Dunmow, Essex, UK), zeroed with a reagent blank (20 pIL saline, plus 1 mL dye reagent). Protein concentrations were calculated by extrapolating the absorbance values from the linear portion of the calibration curve. Protein-rich CSF samples were reassayed at lower volume (2.5-10 EL). The colour yields of the individual proteins were compared over a wider range of protein amount by assaying 10-40 ptL of protein calibrator (HA, A/G, G or MBP at 500 mg/L) in a sample volume of 40 pL.

PRM protein dye-binding assay: CSF (40 gL) or protein calibrator (10-40gL; HA, A/G, G or MBP at 500 mg/L), in a sample volume of 40 jL (adjusted with saline), were mixed gently with Microprotein-PRTM reagent (1 mL) according to the manufacturer's instructions. After 10 minutes, absorbance values were measured (A6 .) in a plastic microcuvette (1 cm) using a Jenway 6100 spectrophotometer, zeroed with a reagent blank (40 lL saline, plus 1 mL dye reagent). Protein concentrations were calculated by extrapolating the absorbance values from the linear portion of the calibration curve. Protein-rich CSF were re-assayed at lower volume (5-20 pl). The colour yields for the individual proteins were compared over a wider range of protein amount by assaying 20-80 lIL of protein calibrator (HA, A/G, G or MBP at 500 mg/L), in a sample volume of 80 gL.

Results

Protein concentration values were higher with the PRM assay than the CBB assay. Values for the CSF samples tested (n = 73) ranged from 55-1960 mg/L (median: 315 mg/L) with the CBB assay, and from 95-2450 mg/L (median: 395 mg/L) with the PRM assay. Linear regression analysis indicated yCB = 0.824xPM - 8, in units of mg/L (r = 0.99) (Figure 1). Both assays gave comparable reproducibility as indicated by precision studies on three CSF samples (Table 1).

Comparison of the response of the two assays to different proteins (Figure 2) indicated that the PRM assay showed a more uniform response to A/G and G, relative to HA. Thus, for 20 pg protein, the absorbance ratio for HA: A/G: G was 1:0.94:0.84 with the PRM assay, and 1:0.79:0.51 with the CBB assay. The extent of linearity was greater with the PRM assay (40 Rg HA, 20 Rg A/G) than the CBB assay (10 jig HA, Ig A/G). MBP gave higher colour yields than HA with the PRM assay, but lower colour yields than HA with the CBB assay (Figure 2). Thus, for 20 pg protein, the HA : MBP ratio was 1:1.53 with the PRM assay, and 1:0.87 with the CBB assay.

The use of A/G as a protein calibrator improved agreement between the PRM and CBB assays. Thus, CSF values ranged from 70-2600 mg/L (median: 425 mg/L) with the CBB assay, and from 1132740 mg/L (median: 438 mg/L) with the PRM assay. Linear regression analysis indicated yBB = 0.960xPR' + 0, in units of mg/L (r = 0.99) (Figure 3). Precision studies indicated that the assays retained comparable reproducibility when A/G was used as the protein calibrator (Table 2).

The use of G or MBP as the protein calibrator highlighted the importance of this choice when using the CBB and PRM assays. Calibrating with G, the CSF protein concentrations ranged from 125-400 mg/L (median: 720 mg/L) with the CBB assay, and from 1203200 mg/L (median: 475 mg/L) with the PRM assay. Linear regression analysis indicated yC = 1.338x' + 57, in units of mg/L (r = 0.99) (Figure 4). Calibrating with MBP, the CSF protein concentrations ranged from 55-2100mg/L (median: 325 mg/L) with the CBB assay, and from 38-1000 mg/L (median: 165 mg/L) with the PRM assay. Linear regression analysis indicated yBB = 2.008xP" - 8, in units of mg/L (r = 0.99) (Figure 5).

Comparison of a small number of specimens (n = 8), both before and after freezing, indicated 97% ( 4%) recovery of protein (with both assays), without effect upon their relative response. Protein concentrations in fresh samples were 232-430 mg/L (CBB) and 280530 mg/L (PRM), and in frozen samples were 231443 mg/L (CBB) and 280-515 mg/L (PRM).

Discussion

Biochemical analysis of CSF is of clinical importance for the diagnosis and management of infectious disease, trauma, infarction, neoplasia and degeneration of the central nervous system? Total protein levels are determined routinely in the hospital laboratory, and reference intervals of 120-640 mg/L have been established using a variety of methods, including the Du Pont aca and Kodak Ektachem dry chemistry systems.4 CSF total protein levels are requested commonly to support a diagnosis, but sometimes can be used for differential diagnosis (e.g. a cut off value >1 g/L reliably distinguishes septic from aseptic meningitis').

Many methods have been described for total protein determination in CSF,216 and comparative studies have recommended different methods; ,4,12-14,17,25 each of which has its own advantages and limitations. For example, the biuret method is accurate but requires high sample volumes; the Lowry method is sensitive but prone to interference; turbidimetric methods (TCA, SSA) have a wide working range but are temperature sensitive; the TCA-PS method is not sensitive to temperature but is prone to manipulative loss of protein; and the CBB assay is simple but shows a variable response to different proteins.2 26

Recently, the CBB assay has been recommended widely for total protein determination in CSF,'*zb having proved to be rapid, sensitive, economic and resistant to interference. In its automated form,21,24,26 the CBB assay compares well with the Du Pont aca dry chemistry system.26

The PRM protein dye-binding assay was developed for total protein determination in urine, and correlates well with the biuret reference method."29 It has the advantages of the CBB method, a more uniform response to different proteins,27 and less tendency to dye precipitation in cuvettes and automated systems.21

The results presented here indicate that, when applied to CSF, higher protein concentration values are obtained with the PRM assay than the CBB assay. This is consistent with previous studies reporting differences in the protein concentration values in CSF, when determined by two or more methods.'1,19,23,25 The difference can be explained by underestimation of globulin in the CBB assay, and can be minimised by the use of AIG (rather than HA) as a protein calibrator.

The A/G calibrator used in the present study is supplied commercially (63% human albumin, 37% human globulin) and used routinely (in undiluted form) for discrete analyser total protein determination in serum/plasma - it matches the CSF albumin : globulin ratio closely, which has been estimated at 70:30.?11 Although modification of the CBB method by adding sodium dodecyl sulphate (SDS) to the dye reagent has been recommended to overcome the variable response of the assay to different proteins,11,15 it was concluded that the choice of calibrator is more important than the method used."

Both PRM and CBB protein dye-binding assays are simple, rapid and economic (current reagent costs are kO.067 and 0.147 per sample, respectively), and demonstrate similar performance characteristics when applied to CSF. Both show rapid colour development (2-3 min) and comparable stability of colour yield (15-30 min); however, the PRM assay gives a more uniform response to albumin and globulin, and is therefore preferable for total protein determination in CSF because the A:G ratio varies in pathological samples.' Furthermore, in contrast to the CBB assay, the linearity of the PRM assay spans the reference interval for CSF protein (120-640 mg/L),4 and extends towards the cut-off value (1 g/L) used for diagnostic purposes.

The CBB assay gives higher colour yields (absorbance per microgram of protein) than does the PRM assay, and, thus, is more sensitive to low protein content in CSF; however, such samples are of little clinical significance. MBP gives an unusually high colour yield with the PRM assay but this is not a problem as its levels normally do not exceed 2.5 pg/L in CSF,31 and elevations in excess of 5000-fold would be necessary to increase total protein levels erroneously.

In conclusion, the PRM assay is an acceptable alternative to the CBB assay for total protein determination in CSF.

[Reference]

References

[Reference]

1 Silverman LM, Christenson RH. Clinical significance of cerebrospinal fluid protein analysis. In: Burns CA, Ashwood ER, eds. Tietz textbook of clinical chemistry, 2nd edn. Philadelphia, PA: WB Saunders Company, 1994: 724-6.

2 Rice EW, Loftis JW. Critique of the determination of proteins in cerebrospinal fluid. Evaluation of the biuret method of Goa and the TCA-turbidimetric method of Meulemans. Clin Chem 1962; 8: 56.

3 Blijenberg BG, Roetring HA, Zwang L, Leijnse B. Spinal fluid protein revisited: a reappraisal of the biuret procedure. Clin Chem Clin Biochem 1985; 23: 225-30.

4 Lott JA, Warren P. Estimation of reference intervals for total protein in cerebrospinal fluid. Clin Chem 1989; 35: 1766-70.

[Reference]

5 Daughaday WH, Lowry OH, Rosebrough NJ, Fields WS. Determination of cerebrospinal fluid protein with the Folin phenol reagent. J Lab Clin Med 1952; 39: 663-5.

6 Zondag HA, van Boetzelaer GL. Determination of protein in cerebrospinal fluid. Sources of error in the Lowry method. Clin Chim Acta 1960; 5: 155-6.

7 Bossak HN, Rosenberg AA, Harris A. A quantitative turbidimetric method for the determination of spinal fluid protein. _ Vener Dis Infect 1949; 30: 100-3.

8 Meulemans 0. Determination of total protein in spinal fluid with sulphosalicylic acid and trichloroacetic acid. Clin Chim Acta 1960; 5: 757-61.

9 Iwata J, Nishikaze 0. New micro-turbidimetric method for determination of protein in cerebrospinal fluid and urine. Clin Chem 1979; 25: 1317-9.

10 Flachaire E, Damour 0, Bienvenu J, Aouiti T, Later R. Assessment of the benzethonium chloride method for routine determination of protein in cerebrospinal fluid and urine. Clin Chem 1983; 29: 343-5.

11 Luxton RW, Patel P, Keir G, Thompson EJ. A micro-method for measuring total protein in cerebrospinal fluid by using benzethonium chloride in microtiter plate wells. Clin Chem 1989; 35: 1731-4.

12 Pesce MA, Strande CS. A new micromethod for determination of protein in cerebrospinal fluid and urine. Clin Chem 1973; 19: 1265-7.

[Reference]

13 Karlsson B, Alling C. A comparative study of three approaches to the routine quantitative determination of spinal fluid total proteins. Clin Chim Acta 1980; 105: 65-73.

14 Shephard MDS, Whiting MJ. Nephelometric determination of total protein in cerebrospinal fluid and urine using benzalkonium chloride as precipitation reagent. Ann Clin Biochem 1992; 29: 411-7.

[Reference]

15 Shihabi ZK, Dyer RD. Protein analysis with bicinchoninic acid. Ann Clin Lab Sci 1988; 18: 235-9.

16 Matsushita M, Irino T, Komoda T, Sakagishi Y. Determination of proteins by a reverse biuret method combined with copperbathocuproine chelate reaction. Clin Chim Acta 1993; 216: 103-11.

[Reference]

17 Krystal G, Lam V, Schreiber WE. Application of a silverbinding assay to the determination of protein in cerebrospinal fluid. Clin Chem 1989; 35: 860-4.

18 McIntosh JC. Application of dye-binding method to the determination of protein in urine and cerebrospinal fluid. Clin Chem 1977; 23: 1939-40.

19 Johnson JA, Lott JA. Standardization of the Coomassie blue method for cerebrospinal fluid proteins. Clin Chem 1978; 24: 1931-3.

[Reference]

20 Schleicher E, Wieland OH. Evaluation of the Bradford method for protein determination in body fluids. j Clin Chem Clin Biochem 1978; 16: 533-4.

21 Heick HMC, Begin-Heick N, Acharya C, Mohammed A. Automated determination of urine and cerebrospinal fluid proteins with Coomassie brilliant blue and the Abbott ABA-100. Clin Biochem 1980; 13: 81-3.

22 Hische EAH, van der Helm HJ, van Meegen MT, Blanken HIG. Protein estimation in cerebrospinal fluid with Coomassie brilliant blue. Clin Chem 1982; 28: 1236-7.

23 Macart M, Gerbaut L. An improvement of the Coomassie blue dye binding method allowing an equal sensitivity to various pro

[Reference]

teins: application to cerebrospinal fluid. Clin Chim Acta 1982; 122: 93-101.

24 Gillery P, Locre F, Malgras A, Borel JP. A continuous-flow technique with Coomassie brilliant blue for cerebrospinal fluid protein assay. Clin Chem 1985; 31: 1092-3.

25 Gerbaut L, Macart M. Is standardization more important than methodology for assay of total protein in cerebrospinal fluid? Clin Chem 1986; 32: 353-5.

26 Huang CM. Development and evaluation of an automated dyebinding assay for protein in cerebrospinal fluid. Clin Chem 1988; 34: 980-3.

27 Watanabe N, Kamel S, Ohkubo A et al. Urinary protein as measured with a pyrogallol-red-molybdate complex manually and in a Hitachi 726 automated analyzer. Clin Chem 1986; 32: 1551-4.

[Reference]

28 Phillipou G, James SK, Seaborn CJ, Phillips PJ. Screening for microalbuminuria by use of a rapid low-cost colorimetric assay. Clin Chem 1989; 35: 456-8.

29 Orsonneau JL, Douet P, Massoubre C, Lustenberger P, Bernard S. An improved pyrogallol red-molybdate method for determining total urinary protein. Clin Chem 1989; 35: 2233.

30 Watson MA, Scott MG. Clinical utility of biochemical analysis of cerebrospinal fluid. Clin Chem 1995; 41: 343-60.

31 Painter PC, Cope JY, Smith JL. Reference intervals. In: Burtis CA, Ashwood ER, eds. Tietz textbook of clinical chemistry. 2nd edn. Philadelphia, PA: WB Saunders Company, 1994: 2199.

[Author Affiliation]

Am-teniwyAW'',*************************** INIf 1 I--1el NE,,M, Wl 110OWMARSHAUs ,logo, I * , MI,* Z7jtww,2w0

[Author Affiliation]

Correspondence to: Dr. T. Marshall. E-mail: tom.marshall@sunderland.ac.uk