Scand J Haematol(1977) 19, 201-206

A Radioisotope Dilution Assay

for Unlabelled Vitamin BIZ-Intrinsic Factor Complex

Employing the Binding Intrinsic Factor Antibody:

Probable Evidence for Two Types of Binding Antibody

ELIZABETH JAOOB, H. A. W. O’BRIEN & D. L. MOLLIN

Department of Haematology, St. Bartholomew’s Hospital, London, England

A new radioisotope dilution assay for vitamin Biz-intrinsic factor complex is described. The method is based on the use of the binding type intrinsic factor antibody (the binding reagent), which when combined with the intrinsic factor-vitamin Biz complex (labelled ligand), is quantitatively adsorbed onto zirconium phmphate gel at pH 6.25. The new assay has been shown to provide a measure of intrinsic factor comparable with other intrinsic factor assays, but it has the important advantage of being able to measure the unlabelled vitamin Biz-intrinsic factor complex (unlabelled Iigand), and will, therefore, be valuable in the study of physiological events in the gastrointestinal tract. During the study, it was found that there is some evidence for at least two types of binding intrinsic factor antibody: One which combines preferentially with the intrinsic factor-vitamin BIZ complex and one which combines equally well with this complex or with free intrinsic factor.

Key words: radioisoto’pe dilution assay - vitamin BizIF complex

Accepted folr publication March 7 , 1977

Correspondence to: Dr. Elizabeth Jacob, Hematology and Nutrition Laboratory, Veterans Administration Hospital, 130 West Kingsbridge Road,

Bronx, New York 10468, USA

A previous communication (Jacob & O’Brien tract, it is desirable to be able to study IF 1972) described a simple assay of radio- bound to unlabelled vitamin B12 (BizIF). active labelled vitamin Biz-intrinsic factor The present communication describes a complex (57CoB1~IF) employing the binding radioisotope dilution assay which enables intrinsic factor antibody and zirconium B12IF to be measured. phosphate gel at pH 6.25. In certain situa- tions, such as when studying the fate of intrinsic factor (IF) in the gastrointestinal

Principle of the assay Binding IF antibody obtained from sera of

202 ELIZABETH JACOB, H. A. W. O'BRIEN & D. L. MOLLIN

patients with pernicious anaemia is em- ployed as the reagent to bind the B d F ligand. The bound ligand (antibody-BizIF complex) is separated from the free ligand (B121F) by the use of zirconium phosphate gel at pH 6.25. In the presence of excess ligand, the amount of bound ligand ad- sorbed into the gel is dependent only on the amount of binding reagent present. This system can therefore be used to ob- tain standard aliquots of pools of B d F for a radioisotope dilution assay. When the pool consists only of the radioactive ligand (57CoBizIF, stock solution), the radioactivity of the aliquot, that is, the amount of radio- activity adsorbed onto the gel, gives the control value. When the stock solution is diluted with nonradioactive ligand (BnIF), the radioactivity of the aliquot will be re- duced below the control value, depending on the relative proportions of the radio- active and nonradioactive ligands (Figure 1).

Figure 1. Principle of radioisotope dilution assay. Upper section shows pool of intrinsic factor saturated with radioactive vitamin Biz to which IF binding type antibody is added to sample an aliquot of the pool which is then quantitatively ad- sorbed onto zirconium phosphate gel at pH 6.25. Lower section shows ef- fect of adding unlabelled vitamin BizIF to the pool of labelled BIZIF and the subsequent reduction in radioactivity of the aliquot adsorbed onto the zirconium gel.

MATERIALS AND METHODS

Zirconium phosphate gel at pH 6.25 was made ac- cording to the medhod of Hansen et al (1966).

"Co-labelled vitamin BIZ (70 pCi/pg) was ob- tained from the Radiochemical Center, Amers- ham. This was diluted with nonradioactive vita- min Biz to give a specific activity of approxi- mately 8 ,uCi/pg.

Albumin coated charcoal was prepared accord- ing to the method of Gottlieb et al (1965).

Serum containing the binding type of intrinsic factor antibody was obtained from patients with pernicious anaemia. The titer of the antibody in the sera was measured by the zirconium gel assay (Jacob et a1 1966, Hansen et a1 1971). In order to eliminate the problem of transcobalamin I1 in the serum binding radioactive vitamin Biz to the gel (Hansen et a1 1967) the vitamin Biz binding sites of the serum were saturated with excess non- radioactive vitamin Biz (5 ng/ml serum) and the unbound vitamin Bi2 removed with 2.0 ml of al- bumin coated charcoal. The Biz saturated anti- body serum was then diluted with 0.15 M NaCl so that it contained 25-35 ng units of antibody per ml.

D D

D D

IF -57CoB12 SOLUTION D

D.G D D

I F - ~ ~ C ~ B ~ ~ SOLUTION 0 D

UNLABELLED D [9 t

ADD ANTIBODY

D D

ADD ZIRCONIUM GEL

0

4 UNITS RADIOAC'I IV lTY

0

0

2 UNITS RADIOACTIVITY

B 1 2 - I F

t3D D D O D D D D

RADIOACTIVE V!TAYIN 812

INTRINSIC FACTOR

BINDING INTRINSIC FACTOR ANTIBODY

INTRINSIC FACTOR ASSAY 203

Specimens of gastric juice were obtained from 10 healthy adult volunteers after pentagastrin stimulation (6 pg/kg body wt. i.m.). The juice was filtered through glass wool, the pH adjusted to 10 to inactivate pepsin, and 30 min later brought back to pH 7.4 and stored in aliquots at -20° C .

The vitamin BIZ binding sites on the IF were saturated by adding labelled or unlabelled BIZ to gastric juice in excess of its total binding capa- city, the excess vitamin being removed by treating with albumin-coated charcoal.

Intrinsic factor quantities are expressed in ng units, i.e., the amount which will bind 1 ng of vitamin BIZ is equal to 1 ng unit.

RESULTS

Standard curve. For the standard curve, 0.25 ml aliquots of stock solution of satu- rated 57CoB~21F containing approximately 8 ng units of IF were taken in a series of duplicate polystyrene tubes. The first two tubes served as controls. To the other pairs of tubes, increasing amounts of saturated unlabelled BizIF ranging from 4 to 20 ng units were added. The vollume in each tube was made up to 1.8 ml with 0.15 M NaCl. 0.2 ml of the solution containing the bind- ing antibody (5 to 7 ng) was added to each tube. The tubes were incubated folr 30 min at room temp. To each tube 10 ml of zir- conium phosphate gel (pH 6.25) was added and the contents mixed. The tubes were centrifuged at 1500 g for 10 min; the gel was washed 3 times with 10 ml of 0.1 M ammonium acetate solution (pH 6.25) and the radioactivity remaining on the gel counted. The results of a typical standard curve are shown in Figure 2.

The assay. 1.0 ml of the unknown sample of IF to be assayed was first treated with 1.0 ml (200 ng) of nonradioactive vitamin Biz solution prepared in 0.15 M NaC1, to ensure that all available IF was saturated

with the vitamin. Excess vitamin was re- moved with 2.0 d of albumin coated charcoal, giving a final volume of 4.0 ml. Duplicate amounts of 0.1 ml, 0.2 ml, and 0.3 ml of the vitamin Biz saturated un- known sample were then added to 0.25 ml aliquots of the stock solution. The volume in each tube was made up to1 1.8 ml with 0.15 M NaCl. 0.2 ml of the antibody solu- tion was added, the mixture incubated for 30 min at room temp., 10 ml of zirconium phosphate gel (pH 6.25) added, the gel washed three times with 10 ml of 0.1 M ammonium acetate solution (pH 6.25), and the amount of radioactivity adsorbed onto the gel measured. The amount of unlabelled vitamin B12IF in the unknown was then read off the standard curve and the result multiplied by 4 since 1.0 ml of the un- known sample was diluted to a final volume of 4.0 ml. By reference to Figure 2, it will be seen that the assay is most sensitive when the amount of B12IF in the unknown is not moce than 1V2 times the amount of 57CoBi21F in the control tubes. If the result

3 4 8 I2 16 20 ng UNLABELLED BIZ-

I F ADDED

Figure 2. Typical standard curve for radioisotope dilution assay. Each point represents the mean of two observations. The coefficient of variation was 2.6 %.

204 ELIZABETH JACOB, H. A. W. O’BRIEN .& D. L. MOLLIN

does not fall in this sensitive range, it is desirable to repeat the assay diluting the unknown with 0.15 M NaC1.

Reproducibility. One gastric juice was as- sayed in this way 12 times in 1 d. The mean was 21.6 ng unitdm1 with a standard error of 0.7. Another gastric juice, which was stored in small aliquots, was assayed in duplicate 6 times on different days over a two-week period. The mean was 37.3 ng units/ml with a standard error of 1.0.

Comparison with direct zirconium gel as- say. Since the direct zirconium gel assay has been shown to give results comparable with other methods (Jacob & O’Brien 1972), this radioisotope dilution assay was compared with it.

21 gastric juices were divided into two aliquots. To one aliquot S7Co-vitamin BIZ was added in excess of the total vitamin BIZ

each case, the vitamin BIZ was removed with albumin coated charcoal. The IF in the aliquot with the 5 7 C 0 B ~ ~ added was assayed by the previously described tech- nique (Jacob & O’Brien 1972) using the binding antibody and zirconium gel at pH 6.25. The amount of IF vitamin BIZ com- plex in the aliquot with the unlabelled vita- min BIZ was measured by the above tech- nique. The results of the two assays are shown in Figure 3. There was excellent correlation between the two assays (r = 0.99).

Two types of sera with binding type IF antibody. During the course of this study, 4 different binding antibody sera were used. It was noted that there were two types of binding IF antiboldy. When the radioisotope dilution was done with IF saturated with vitamin Biz, all 4 sera produced the same results, but when it was done with free IF,

binding capacity and to the other, unlnbel- 3 of the sera produced one result and the led vitamin BIZ was added in excess. In fourth prolduced a different result. These

2201 E

5 180-

2 140- v)

I Figure 3. Comparison between radio- isotope dilution assay and direct zir- conium gel assay (r = 0.99) (y = 1.035

220 I F CONTENT BY DIRECT ZIRCONIUM GEL ASSAY

(ng UNITS/rnl) x -1.52).

INTRINSIC FACTOR ASSAY

Apparent IF content (ng unitshl) by radiodilution assay

205

IF content by direct zirconium gel assay

TABLE 1 Apparent IF content of 6 gastric juices by the radioisotope dilution assay with 2 sera

compared with the IF content by the direct zirconium gel assay With the antibody from patient G.P. (type A serum), the true IF content is only

given when the IF is saturated with B12 With the antibody from patient G.C. (type B serum), the IF content is measured

equally well with either B d F or free IF

Serum (ng unitdml) I Free IF 1 IF saturated with

vitamin Biz

G.P. (type A)

G.C. (typeB)

23 85 16

57 206

92

14 7 0

56 220

93

22 88 15

59 210

89

two types of sera have been tentatively designated A and B. Typical results using one of the type A sera and the type B se- rum are shown in Table 1. It is clear that in type A sera, the binding antibody ap- pears to have a greater affinity for BizIF than for free IF, whereas in the type B serum the antibody has equal affinity for both IF and BnIF.

DISCUSSION

The differences observed in this study be- tween A and B type sera could theoretically be due to the effects of blocking type IF antibody. This is present in all sera con- taining binding type IF antibody and it is possible that large amounts of blocking type antibody could interfere with the formation of the binding type antibody-IF complex. However, there was no1 relation between the blocking antibody content of the sera and the behavior of the binding type antibody. This, therefolre, does no't seem to1 be a likely explanation. More probably, the two types of sera are due to the existence of two dif-

ferent types of binding IF antibody. If this is so, the IF molecule must have at kast 3 different antigenic determinants, one which produces the blocking type antibody and 2 which produce the two binding type anti- bodies. The one which produces the bind- ing antibody in the type A sera is pre- sumably closely related to the Biz binding site of IF and the one which produces the antibody in type B sera is independent of that site. It is obvioasly important that more studies of a larger series of binding type IF antibodies be undertaken to further elu- cidate this point.

When the type A serum antibody is used in the dilution assay, it is necessary to satu- rate the IF with vitamin Biz and to remove any excess free BIZ with coated charcoal before performing the assay. This procedure is therefore recommended as a routine un- less it is proven that the antibody being employed has the same a n i t y for IF and BizIF.

Rothenberg (1966) described a radioiso- tope dilution assay using the binding IF antibody and sodium sulphate to precipitate and quantitate the antibody-BiJF. This

206 ELIZABETH JACOB, H. A. W. O’BRIEN .& D. L. MOLLIN

technique was only employed to measure free IF although with certain modifications, it could also be used to measure BizIF. However, Rothenberg was using an anti- body which had a greater affinity for B12IF than for free IF and in his assay system, IF and BizIF would not behave the same and if present in different proportions in different samples, the results would be fal- lacious.

The assay described here will quantitate IF with a good degree of reproducibility and with results similar to those obtained by other methods. The assay further ap- pears to be reasonable specific for IF. Previous studies have shown that zirconium phosphate gel at pH 6.25 selectively ad- sorbes IFBiz binding antibody complex and is not influenced by the presence of other binders in the gastric juice (Jacob & O’Brien 1972). Transcobalamin I1 will bind to the gel at pH 5 but this does not bind to the binding IF antibody and therefore will not interfere with the reaction between the binding reagents and the ligand. Since a large excess of the zirconium phosphate gel is employed, the TC I1 present does not interfere with the assay. With the use of this assay, it is therefore possible to measure the total IF content of any material. By adding radioactive vitamin B12 to another aliquot of the material and performing the assay described by Jacob & O’Brien (1972), the amount of IF not already bound to vita- min BIZ can be determined. The difference between the two assays will then be a mea- sure of the amount of IF already bound to vitamin Biz in the original material. The radioisotope dilution assay would appear to provide a useful method for studying problems such as what happens ta IF and B12IF in the gastrointestinal tract, the en- terohepatic circulation of unlabelled vitamin

BIZ, and the fate of vitamin Biz liberated from food.

ACKNOWLEDGEMENT

We are grateful to the Wellcome Trust, London, for their support.

REFERENCES

Gottlieb C , Lau K-S, Wasserman L R & Herbert V (1965) Rapid charcoal assay for intrinsic fac- tor (IF), gastric juice unsaturated BIZ binding capacity, antibody to IF, and serum unsaturated Biz binding capacity. Blood 25, 875-84.

Hansen H J, Jacob E & Frentz G D (1971) Assay of antiintrinsic factor antibodies. Meth Enzymol 18, 130-33.

Hansen H J, Miller 0 N & Tan C H (1966) Assay of the autohumoral antibody that neutralizes the vitamin Biz combining site of intrinsic fac- tor in serum from patients with pernicious ane- mia. Am J Clin Nutr 19, 10-16.

Hansen H J, Tan C H & Jacob E (1967) Studies concerning transcobalamin-I1 of humans. Clin Res 15, 422.

Jacob E, Hansen H J & Miller 0 N (1966) Rapid assay of antibody (CAB) in serum of pernicious anemia patienis which reacts with intrinsic fac- tor vitamin BIZ complex. Fed Proc 25, 430.

Jacob E & O’Brien H A W (1972) A simple as- say of intrinsic factor-vitamin BIZ complex em- ploying the binding intrinsic factor antibody. J Clin Pathol 25, 320-25.

Rothenberg S P (1966) A radioimmunoassay for human intrinsic factor. J Lab Clin Med 67, 879- 90.