How Best to Detect Endogenous FBctor? References - CiteSeerX

151, 5.9, and “not detectable,” respectively. Coulometric measurement of chloride in the chioridometer (Radiometer; Radiometer America, Westlake, OH 44145) gave a value of 115 mmol/L for the same specimen. Blood gases measured in a different specimen gave a calculated HCO of 22.2 mmoh/L. Before admission, values for chloride and CO2 obtained in the referring hospital in the Astra 8 (Beckman Instruments, Brea, CA 92621) were 117 mmoJJL and 23 mmol/L, respectively. Bromide and iodide are known to interfere in the Kodak procedure for measurement of chloride and CO2. In this instance the interference led to the diagnosis (previously unsuspected) of iodide intoxication. The measured iodide concentration was 93 ing/dL (7 mmol/L), obtained for a sample drawn some 15 h after admission. This infant had received much more of a saturated solution of potassium iodide than the dose prescribed by his pediatrician. He was admitted to the hospital for investigation of a febrile illness and, apart from a fleeting rash, was remarkably free of symptoms of iodide intoxication. The infant survived and was subsequently discharged Steven J. Soldin Stacey Nicholson Arnold Einhorn Susan Koppes Depts. of Lab. Med. and Pediatric Med. Children’s Hospital National Medical Center Washington, DC 20010 and Depts. of Child Health & Develop., Pathol. and Pharmacol. George Washington Univ. School of Med. Washington, DC 20037

How Best to Detect Endogenous Digftalis-iike FBctor? To the Editor: There is much interest in the possibility that the cardiac glycoside receptor on the sodium pump may be “designed” to bind an endogenous ligand present in the mammalian body. Such a physiological endogenous counterpart to cardiac glycoside may serve as a specific regulator of the sodium pump (1,2). Despite several attempts to elucidate the precise nature of this endogenous digitalis-like factor, it has eluded purification and definite character2392

ization. Because there is no specific assay method, various different procedures have been used to detect it. However, the limited sensitivity and selectivity of these procedures are the major problems, difficult to overcome (3). Gruber et al. (4) first described the digoxin-like immunoreactivity for the endogenous digitalis-like substance. Their approach was based on the premise that a compound that binds to a specific endogenous receptor may also bind to an antibody raised against the exogenous ligand (drug). However, it has been known that anti-digoxin antiserum cross-reacts to some extent with most steroids (5). Indeed, many steroids give falsely positive results for digoxin (6), as do many lipids and bile acids (6-8). Moreover, recent findings indicate the dissociation of digoxin-like immunoreactivity from digitalis-like biological activity (9). Overall, digoxin antibodies are not necessarily directed at molecular determinants critical for biological activity (10). It is now evident that the cligoxin-like immunoreactivity may not represent the endogenous ligand with digitalis-like properties. In this context, we agree with the view of Clerico et al. (11) that immunological methods should be used only as a screening or preliminary test to detect the possible existence of digitalislike factor. If one assumes that the presumptive endogenous higand has digitalis-like properties, its identification requires that it share at least some of the biological activities of digitalis: competition for specific binding sites in target tissues, inhibition of the sodium pump, and (or) inhibition of what seems to be the molecular basis of the sodium pump, the enzyme Na/K-transporting ATPase (BC 3.6.1.37). Competition for specific binding sites would appear to be the sine qua non for an endogenous digitalis-like factor. We emphasize that more attention should be paid to whether isolated enzyme or intact cells are used to determine the biological activities. It should be recognized that the isolated NafK-transporting ATPase is devoid of the buffering action of the membrane and thus is more susceptible to non-receptor-mediated inhibition. In fact, inhibition of the isolated enzyme may not provide evidence for inhibition of the sodium pump in intact cells (12). The sidedness of the ligands and the glycoside-binding sites is not maintained in studies of the isolated enzyme. Any extrapolation of results from these studies to the actual interaction between a candidate substance and the sodium pump in intact cells requires caution.

CLINICAL CHEMISTRY, Vol. 34, No. 11, 1988

1

S

-1ev

(Co.,cntratLo

of

cc.po.nd.

mauL)

Fig. 1. Effectof progesterone (0) on [3H]ouabain binding to intact human erythrocytes compared with that of ouabain (#{149}) according to the method described previously (17) Several substances have been proposed as endogenous digitalis-like factors, including unsaturated fatty acids (13), lysophosphatidylcholine (14), dehydroepiandrosterone sulfate (15), and ascorbic acid (16). However, none of these compounds appear to be the nat.. ui-al ligands of the digitalis receptor of Na/K-transporting ATPase because of their limited affinity and specificity. Purifications of these substances were mainly based on the detection methods involving the isolated enzyme. We found that these compounds purported to be “endogenous digitalis” have no significant effect on the binding of [3lfiouabain to intact human erythrocytes (17). Recently, progesterone and cortisol have been added to the list of candidate substances (18). Indeed, these steroids do show the dose-dependent inhibition of [3Hlouabain binding to NafK4-transporting ATPase isolated from hog brain (18). We investigated whether progesterone (Figure 1) and cortisol possess the [3Hlouabain-displacing activity from intact cells. At the concentrations tested (2 X iO to 2 x iO mol/L), these steroids had no effect on [3lflouabain binding to intact human erythrocytes. This observation strongly indicates that progesterone and cortisol are unlikely to be the endogenous ligands of the cardiac glycoside receptor of the enzyme. We were actually able topuria polar, novel digitalis-like factor with potent [3H}ouabain-displacing activity from intact human erythrocytes (17, 19). We believe that the biological activities in intact cells should be determined for the detection of endogenous digitalis-like factor.

References 1. Poeton L. Endogenous sodium pump inhibitors: a role in essential hypertension. Chin Sci 1987;72:647-55.

2. Graves SW, Williams GH. Endogenous digitalis-like natriuretic factors. Annu Rev Med 1987;38:433-44. 3. Sagnella GA, MacGregor GA. Problems in the identification and measurement of a sodium transport inhibitor in human plasma. Khin Wochenschr 1985;63(Suppl 110:150-3.

4 Gruber KA, Whitaker JA, Buckalew VM. Endogenous digitalis-like substance in plasma of volume-expanded dogs. Nature (London) 1980287:743-5. 5. Butler VP, Tse-Eng D. Immunoassay of digoxin and other cardiac glycosides. Methods Enzymol 1982;84:558-77. 6. Soldin SI, Papanastausiou-Dianiandi A, Heyes J, et al. Are immunoassays for digoxin reliable? Clin Biochem 1984;17:317-20. 7. Young A, Giesbrecht E, Soldin 5,1. A study of lipid effects on the digoxin immunoassay and on the binding to and activity of Na/K-ATPase. Chin Biochem 1986; 19:195-200.

8. Toaehand PA, Oedfleld PR, Murphy GM, et al. Tentative identification of a digoxinlike immunoreactive substance. Ther Drug Monit 1988;10:168-71. 9. Kelly RA, O’Hara 1)8, Canessa ML, et al. Characterization of digitalis-like factors in hwnan plasma. Interactions with NaKATPase and cross-reactivity with cardiac glycoside-specific antibodies. J Biol Chem 1985;260:11396-405.

10. Hnatowich M, Labella F. Endogenous digitalis-like factors: in vitro comparison of biological and immunological activities of peptide and steroid candidate. Eur J Pharmacoh 1985;106:567-75. it Clerico A, Ghione S, Del Chicca MG, et al. Problems in standardization of digitalislike substance assays by means of competitive immunological methods. Clin Chem 1987;33:340-1.

12. Sagnella GA, MacGregor GA. Characteristics of a (Na-KiATPase inhibitor in extracts of tea. AmJ Clin Nutr 1984;40:3641. 13. Tamura M, Kuwano H, Kinoshita T, et al. Identification

of hinoleic and oleic acids

as endogenous Na’,K-ATPase inhibitors from acute vohume-expanded hog plasma. J Biol Chem 1985;260:9672-7. 14. Kelly RA, (YHara 1)8, Mitch WE, et al. Identification of NaK-ATPase inhibitors in human plasma as nonesterifled fatty acids and hysophosphohipids. J Biol Chem 1986;261:11704-11. 15. Vasdev 5, Longerich L, Johnson E, et al. Dehydroepiandrosterone sulfate as a digitalis-like factor in plasma of healthy human adults. Rea Commun Chem Pathol Pharmacol 1985;49:387-99. 16 Kuske H, Moreth K, Renner 1), et al. Sodium pump inhibitor in the serum of patients with essential hypertension and its partial purification from hemofiltrate. Kiln Wochenschr 1987;65(Suppl Vffl):53-9. 17. Goto A, Yamada K, Ishii M, et al. The effects of urinary digitalis-like factor on cultured vascular smooth muscle cells. Hypertension 1988;11:645-50. 18. Longei-ich L, Brent DA, Johnson RL, et

al. Identification of progesterone and cortisol as immunoreactive plasma digitalis-like factors in pregnancy. Res Commun Chem

ples calculated from the sum of overnight, morning, and remaining daytime samples, all predict an overnight AER >12 g/min better than does the corresponding uncorrected albumin concentration; these unpublished results were obtained for diabetic subjects. We would therefore suggest that the clinical strategy suggested by Howey et al. is unnecessarily restrictive and may be lacking in sensitivity and specificity, particularly when applied to diabetic subjects. Present evidence indicates that better sensitivity and specificity in the detection of early increases in albumin excretion in diabetics will be achieved by correcting urine albumin concentration for either creatinine concentration or time and volume, and there is little to choose between different times to carry out urine collections for this purpose.

Pathol Pharmacol 1988;59:383-93. 19. Goto A, Yamada K, Ishii M, et al. Presence of digitalis-hike factor in mammalian plasma. Biochem Biophys Res Cornmun 1988;152:322-7. L Goto K. Yamada

M. Ishii T. Sugimoto Second Dept. of Intern. Med. University of Tokyo 7-3-1 Hongo, Bunkyo-ku Tokyo 113, Japan

Which Specimen to Screen for Microalbuminurla To the Editor: Recently Howey et al. (Chin Chem 1987;33:2034-8) advocated measurement of the albumin concentration in the first morning urine specimen, with a cutoff value of 30 mg/L, to screen for pathologically increased albumin excretion by diabetic subjects. They stated that correction of this figure for creatinine concentration, or for time and volume, did not improve the intraindividual variability of albumin excretion or the predictive value of the albumin test. We would challenge these suggestions on the following grounds: #{149} Theirdatawereobtainedinasmall sample (n = 11) of non-diabetic individuals. #{149} It may not be correct to extrapolate from a non-diabetic population to provide a screening protocol for diabetic patients. For instance, changes in glycemic control in diabetics during the night may affect urine flow rate, which will alter albumin excretion to a different extent than in normals. #{149} Several workers have demonstrated that the validity of urine albumin concentration alone is less than that of the albumin/creatinine ratio (ACR) in predicting microalbuminuria in samples from recumbent or ambulatory subjects (1-3). This would be expected under conditions in which urine flow rate underwent changes throughout the day. #{149} As others (3, 4) have shown, the ACR in an untimed specimen collected in the clinic shows a greater than 90% correlation with timed overnight albumin excretion rate (AER). Moreover, ACR, or AER in morning samples, random daytime samples, or 24-h sam-

References

1. Gatling W, Knight C, Hill RD. Screening for early diabetic nephropathy. Diabetic Med 1985;2:451-5. 2. Marshall SM, Alberti KG. Screening for early diabetic nephropathy. Ann Chin Biochem 1986;23:195-7. 3. Watts GF, Shaw KM, Poiak A. The use of random urine samples to screen for mlcroalbuininuria in the diabetic clinic. Practical Diabetes 1986;3(2):86-8. 4. Nathan DM, Rosenbaum C, Protasowicki D. Single void urine samples can be used to estimate quantitative microalbuminuria. Diabetes Care 1987;1O:414-8. David J. F. Rowe Rysiek M. Szydlo Dept. of Chem. Pat/wi. The General Hospital Southampton S09 4XY,

U.K.

Gerald

F. Watts

Dept. of Chem. Pathol. & Metab. Med. St Thomas’s Hospital London SE1 7EH, U.K. The authors respond:

of the paper

in question

To the Editor: We thank Rowe and colleagues for their comments and wish to make the following points in reply. Contrary to their claim, we made no statement on the relative predictive value of differing reporting formats for the albumin test. Ideally, such a screening test for important treatable disease should have 100% nosological sensitivity, with subsequent identification of false positives by repeat testing or by performing an alternative test with different nosological characteris-

CLINICAL

CHEMISTRY,

Vol.34,

No. 11, 1988

2393