- Email: [email protected]
Specific removal of IgE by therapeutic immunoadsorption system
Journal of Immunological Methods, 118 (1989) 161-168
161
Elsevier JIM05105
Specific removal of IgE by therapeutic immunoadsorption system H i r o s h i Sato, K o i c h i W a t a n a b e , Junichi A z u m a , T e r u h i s a K i d a k a a n d M a s a t a k e H o r i Takeda ChemicalIndustries, Ltd., Medical Supply and Equipment Department, 17-85Jusohonmachi 2-chome, Yodogawa-ku, Osaka 532, Japan
(Received3 October 1988, accepted 8 November 1988)
A therapeutic immunoadsorption system was developed that can remove IgE effectively and specificall) from the plasma of patients with an allergy or other hyper-IgE syndrome. The immunoadsorbent (IA) consists of immunoaffinity purified anti-IgE antibody (a-IgE ab) immobilized on controlled pore glass beads (50 nm pore size). Adsorption isotherms for IgE, which were reduced by the Freundlich adsorption equation, were obtained with IA that immobilized various amounts of a-IgE ab. An optimum amount of a-IgE ab to be immobilized was selected. IA worked sufficiently in a wide range of IgE concentrations. Clinical treatment requires an amount of 41 mg of IgE to be removed from a patient's plasma for 3 h. An IA for clinical use was designed to contain 10 g of the support binding 325 mg or more of the antibody. In fact, our study in vitro simulating a clinical case showed that serum IgE was removed by IA, as expected: the level decreased from 11 000 to 3000 U / m l after a 3 h perfusion (1 U = 2.3 ng). A very small amount of a-IgE ab (goat IgG) was found to be detached from IA by flowing plasma; the average level was 20 n g / m l , which seems to be safe. However, we installed the second column in a circuit that adsorbs a-IgE ab leaked into plasma, because the amounts of a-IgE ab infused into the patient must be minimized. The second column contained IgE immobilized on the same support, since IgE as a ligand adsorbed more a-IgE ab than did anti-goat IgG antibody. This is an effective and safe therapeutic immunoadsorption system and has been subjected to clinical tests. Key words: IgE; Immunoadsorption; Immobilization; Anti-IgE
Introduction
Type I hypersensitivities are inflammatory reactions mediated by IgE antibodies bound to mast
Correspondence to: H. Sato, Takeda Chemical Industries, Ltd., Medical Supply and Equipment Department, 17-85 Jusohonmachi 2-chome, Yodogawa-ku, Osaka 532, Japan. Abbreviations: IA, immunoadsorbent; a-IgE ab, anti-IgE antibody; CPG, controlled pore glass; DEAE, diethylaminoethyl; "Iris, tris(hydroxymethyl)aminomethane;NMWL, nominal molecular weight limit; PBS, phosphate-buffered saline; RT, room temperature; BSA, bovine serum albumin; EOG, ethyleneoxide gas.
cells and basophils; the reactions result from the release of pharmacologically active factors upon combination with allergens. The concentration of IgE in patients with allergic asthma and atopic dermatitis (eczema) is much higher than that in normal individuals. To treat such patients, plasmapheresis, or plasma exchange has been applied with or without drugs (Gartmann et al., 1978; Ishikawa et al., 1982; Bambauer et al., 1984; Dau, 1988). These investigators reported a beneficial clinical course and significant decreases in serum IgE after plasmapheresis. However, it is difficult to analyze the mechanisms of the therapy because plas-
0022-1759/89/$03.50 © 1989 ElsevierSciencePublishers B.V. (BiomedicalDivision)
162 mapheresis removes substances non-selectively and requires substituted plasma such as frozen fresh plasma from healthy donors. Therapeutic on-line plasmapheresis with IA has the advantage that it removes substances specifically in the plasma and excludes the risk of virus infections that might be introduced from the donor plasma. Some IAs for reacting IgE were prepared: burro anti-IgE (Fc)-Sepharose 2B (Solley et al., 1976), protein A-Sepharose CL-4B (Johansson and Ingan~is, 1978; Zikhn et al., 1984), concanavalin A-Sepharose 4B (Ray and Raychaudhuri, 1982), anti-IgE serum-bromoacetyl cellulose (Carrel et al., 1971), and tryptophan-polyvinyl alcohol (Behm et al., 1987). Most of these were analytical tools, and a-IgE ab as a ligand seemed to have high affinity for IgE. Therapeutic IA must meet the following requirements: high affinity, little non-specific reaction, no activation of complement and coagulation systems, little leakage of the ligand from the cartier, adequate strength, tolerable to sterilization, and no toxicity; furthermore, specialized test methods for sterility of IA are required (Sato et al., 1987a). A concrete specification of the IA in affinity to IgE was set: circulating serum IgE must be decreased from 10000 to 3000 U / m l after 3 h immunoadsorption therapy in an adult. We have developed a therapeutic immunoadsorption system for removing IgE; we report characteristics of the IgE adsorbent and second adsorbent column which are used together in the system. The second column, human IgE immobilized on CPG beads, removes a-IgE ab detached from the IgE adsorbent column.
Materials and methods
lgE Human IgE was isolated from the supernatant of the culture fluid of a human myeloma cell line, U266. Purification was achieved by a method reported by Ikeyama et al. (1986). IgE was used as an immunogen in goats, as a ligand for affinity purification of a-IgE ab, as a ligand in the second column, and as a solute in the adsorption experiments in vitro.
Anti-lgE antibody Antibody against human IgE was raised in goats by subcutaneously injecting its emulsion with Freund's complete adjuvant. Antisera were purified by precipitation with saturated ammonium sulfate to obtain immunoglobulins followed by DEAE cellulose chromatography to separate an IgG fraction. The IgG fraction was further purified by affinity chromatography, a matrix which consisted of IgE immobilized on Formyl Cellulofine (Seikagaku Kogyo, Tokyo, Japan), where aIgE ab bound to the IgE matrix was eluted by 0.2 M acetic acid containing 0.15 M NaC1 followed by neutralization with 1 M Ttis. The resulting affinity-purified a-IgE ab preparation was concentrated by ultrafiltration (Pellicon System equipped with membranes of 100 000 NMWL; Millipore, Tokyo, Japan). Immobilization All the following processes were carried out under pyrogen-free and aseptic conditions. Activation of carrier A previous study indicated that a pore size of CPG of 50 nm or larger was necessary for good adsorption (Sato et al., 1986a). 470 g of CPG (CPG-500 or CPG-1400; pore size, 50 nm and 140 nm, respectively; 120/200 mesh; Electro-Nucleonics, Fairfield, N J, U.S.A.) was washed with a six-fold volume of acetone and then with 50 liters of water. Fine particles generated were removed by decantation during both washings and the later processes. Water was removed by filtering through a glass filter. CPG was placed in 1500 ml of 10% (w/v) 3,-aminopropyltriethoxysilane(Tokyo Kasei, Tokyo, Japan), the pH adjusted to pH 5 with conc. HC1 and the mixture heated at 80 o C, followed by stirring for 2 h at 80 ° C. The CPG with the introduced aminopropyl group was washed with 60 liters of water and dried at l l 0 ° C overnight after removal of moisture by filtration. 93 g of aminopropyl CPG was placed in 1000 ml of 2.5% (v/v) glutaraldehyde solution and the mixture was then stirred at RT for 1 h. The glutaraldehyde solution was discarded, the beads washed with water and the water removed by filtration. The weight of the beads was 199 g on a wet basis.
163
Immobilization of anti-IgE antibody (preparation of IgE adsorbent) To prepare IA with different amounts of antibody immobilized, glutaraldehyde-activated CPG was reacted to affinity-purified a-IgE ab solutions with various concentrations of the antibody in 0.01 M PBS pH 7.2 for 3 h. The IgG fraction was also immobilized on CPG. Immobilization amounts were calculated from the difference in concentrations of the antibody between pre- and post-immobilization. More than 95% of the antibody applied was found to be immobilized. The resulting beads were washed with saline, then water, and the moisture was removed by filtration. CPG-500 immobilizing a-IgE ab reacted to dimethyl amine borane (Tokyo Kasei) in 0.2 M borate buffer, pH 9.0 at RT for 2 h; 12 mg of this reducing agent was used per 1 g of CPG. These beads were washed with saline. Then the beads were washed with 0.1 M acetate buffer, pH 4.0 containing 1 M NaC1 followed by 0.1 M borate buffer, pH 8.0 containing 1 M NaC1; the beads were washed twice more in the same manner; the beads were again washed with water. Finally, the beads were immersed in a 2% mannitol solution, and the excess solution was removed by filtration. The resulting beads were freeze-dried by Unitop 400L (Virtis Co., Gardiner, NY, U.S.A.).
Immobilization of IgE (preparation of second column adsorbent) The procedure for immobilizing IgE was the same as that described above for a-IgE ab, but a different ligand was used.
Immobilization of anti-goat lgG antibody Rabbit anti-goat IgG antibody (Cappel Labs., Cochranville PA, U.S.A.) was immobiliTed on CPG-500 in a manner similar to the above.
Batch IgE adsorption An adequate amount of IgE adsorbent was mixed with bovine plasma containing various amounts of IgE and incubated while being stirred gently at 20 °C for 5 days. The difference in IgE concentrations between pre- and post-incubation was used to calculate the amounts of IgE adsorbed.
Column IgE adsorption Bovine plasma containing various amounts of IgE was fed downward into a column (5-10 mm ID) at several flow rates at RT. IgE concentrations in plasma at the post-column were determined with time.
Adsorption of anti-IgE antibody Batch and column adsorption tests for a-IgE ab were performed in a manner similar to the IgE adsorption tests.
Determination of IgE IgE 'Mitsui' II (Kainos Labs., Tokyo, Japan) was used to determine the concentrations of IgE in plasma. The concentrations of IgE in the buffers were determined by the relation 1.0 m g/ m l = 1.400 optical density units at 280 nm.
Determination of anti-IgE antibody 100 #1 of IgE solution (4/~g/ml of PBS pH 7.2) was placed in a microtiter well (Immulon 600, C.A. Greiner und Sohne, Nurtingen, F.R.G.). The plates were incubated overnight at 4 o C. The solution was discarded and the wells were rinsed three times with PBS. 300 /~l of gelatin (Wako Pure Chem. Ind., Osaka, Japan, 10 m g / m l of PBS pH 7.2) was added to the well. The plates were incubated overnight at 4 ° C. After the solution was discarded and the wells were rinsed three times with PBS pH 7.2 containing 0.05% Tween 20 (PBS-T), 100 /~1 of affinity-purified a-IgE ab standards or samples appropriately diluted were added. The plates were incubated overnight at 4 ° C. The solution was discarded and the wells were rinsed four times with PBS-T. 100 /xl of biotinylated anti-goat IgG (Bethesda Res. Labs., Gaitherburg, MD, U.S.A.) diluted 1/20000 with PBS-T containing 1% BSA (Miles Labs., Elkhart, IN, U.S.A.) was added to the wells and the plates were incubated with agitation at 25°C for 2 h. The solution was discarded and the wells were rinsed four times with PBS-T. 100 /~1 of Streptavidin-horseradish peroxidase conjugate (Bethesda Res. Labs.) diluted to 1/2000 with PBS-T containing 1% BSA was added to the well and the plates were incubated with agitation at 25 °C for 1 h; the solution was then discarded. The wells were rinsed five times with PBS-T, then twice with PBS.
164
100 ~tl of o-phenylenediamine (Zymed Labs., So. San Francisco, CA, U.S.A.) was added to the well and allowed to stand for 30 min at RT. The reaction was stopped by adding 50 /~1 of 2 M sulfuric acid. Optical densities were read at 490 nm in an ELISA Analyzer ETY-96 (Toyo Sokki, Tokyo, Japan). The lowest detection of a-IgE ab was 0.1 ng/ml. The overnight incubation at 4 ° C could sometimes be replaced by incubation with agitation at 25 ° C for 2 h.
EOG sterilization Dry IAs were sterilized with 20% EOG, where the operating temperature, relative humidity, pressure, and time were 30 o C, 50%, 1 k g / c m 2, and 3 h, respectively. These conditions of EOG sterilization provided 12D, where D, or decimal reduction time was defined as the time required to inactivate 90% of the Bacillus subtilis spores present. The D value was estimated using the spores (Spordex Scale, A M S C O / M e d i c a l Products, Erie, PA, U.S.A.) (data not shown).
IgE adsorbent column The IgE adsorbent column for clinical use consisted of 400 mg of affinity-purified a-IgE ab immobilized on 10 g of CPG-500 in an acrylic resin column (29 mm ID, 50 mm H); it was sterilized.
Second column (for adsorbing a-IgE ab) The second column for clinical use consisted of 26 mg of human IgE immobilized on 10 g of CPG-500 in the same column as the IgE adsorbent; it was sterilized.
Results
Adsorption of IgE The effects of purification and the immobilization amounts of a-IgE ab on IgE adsorption were studied. Fig. 1 shows the adsorption isotherms of IgE at 20 o C, where the IgE adsorbents used were not sterilized. Fig. 1 indicates, using a log-log, that these isotherms fit straight lines well; in other words, they are well fitted to the Freundlich equation; but at the same time they are also well fitted to the Langmuir equation (not shown), which
1000 o ........ O...~." r : r ' ~ . . . . . . . .
.~, . - . ~
-
-
n , ~~, , ~ l ~ r
-~
& --~ 200 v
I00
m
....
......
r
- .O- J- .....
~..O-
20 I0
I
I0
I
10000 1000 IGE CONCENTRATION (U/ML) .... - o 2 MG/G-CPG"~ ---~ 20 MG/G-CPGI 100
100000
--~ ~l .G/G-CPG_] A,,.NITY,u,,,iE0
---.o MG/G-CPG --4, 20 MO/G-CPG
IGG TRACTION
Fig. 1. Adsorption isotherms of IgE on tgE adsorbent immobilizing a-IgE ab at 20 o C. The IgE adsorbents (CPG-500-based) were not sterilized.
suffices in a system with completely homogeneous surfaces and negligible interaction between adsorbed IgE molecules. The data were reduced using the Freundlich equation Y = m . X", where Y is the amount of substance adsorbed, X is the equilibrium concentration of the substance, and m and n are constants. The calculated m and n were 200-320 and 0.06-0.10, respectively for the affinity-purified a-IgE ab. In Fig. 1, the affinity-purified antibody immobilized was found to be able to adsorb 13 times more IgE than was the IgG fraction antibody. Among several immobilization amounts of the affinity-purified antibody, the amounts of IgE adsorbed with the antibody immobilized at 2 or 20 m g / g of CPG were greater than those at 38 or 59 m g / g of CPG. However, these slopes were almost the same. Fig. 2 shows the transient movement of the adsorption waves through the columns, with the corresponding changes in the outflow concentration of IgE expressed as % of the initial IgE level, where bovine plasma containing IgE was passed at a speed of 1.2 /~g of IgE per 1.0 mg of antibody/rain. With CPG-500 immobilizing 10 mg of antibody per 1 g of CPG, the IgE concentration reached the initial value at 3 h. On the other hand, although the adsorption wave with CPG-500 immobilizing 35 mg of antibody per 1 g of CPG was similar to that immobilizing 10 mg until 70-80 min of feed, at a later time the wave was wide, and the breakthrough curve did not become saturated.
165 %
100o°
I0 M o l o - C P
80-
x
z
60
c~o
40
i
I000
o.o-
f
•
°
I ? . . . . . . ,.e.....
I
•
°~°*"D"'@"° . . . . " ' " 4 1 " " ' Q " " "'"
IGE • . . . . . . . . . . . . ~00
~
10'
m
20
20"C
0
"~'R~ I
i
I
60
,
I
120
,
IB0
I
,
240
Fig. 3 shows the time course changes in IgE concentrations at the inlet and outlet of the IgE adsorbent column. The experimentation scale was designed at one-thirty-third of a situation in which a column would be used clinically. The IgE level at the inlet decreased rapidly from 11 000 to 4500 and 3000 U / m l at 2 h and 3 h perfusions, respectively. IgE appeared at the outlet after the perfusion began, but was kept at 1000-1500 U / m l throughout the perfusion.
Leakage of anti-IgE antibody Anti-IgE antibody was found in flowing plasma at the outlet of the IgE adsorbent column in spite of being bonded covalently; the average con-
TI-GOAT
IGG
I
0. I
I
J
I
10
100
1000
ANTI-IGE CONCENTRATION (NG/ML)
Fig. 4. Adsorption isotherms of a-IgE ab on second column immobilizing IgE or anti-goat IgG antibody. 2.6 mg of IgE or 15 mg of anti-goat IgG antibody was immobilized on 1 g of CPG-500; the resulting adsorbents were then sterilized.
centration of a-IgE ab was 20 ng/ml in the perfusion test simulating clinical situations.
Adsorption of anti-IgE antibody Fig. 4 compares the adsorption isotherms of a-IgE ab with different ligands. Immobilizing human IgE as a ligand gave much greater adsorption than did immobilizing anti-goat IgG antibody particularly at low concentrations of a-IgE ab. Table I shows the % of removal of a-IgE ab by the second column immobilizing anti-goat IgG antibody or IgE. The % removal of a-IgE ab inTABLE I % REMOVAL OF a-IgE ab IN COLUMN ADSORPTION BY THE S E C O N D C O L U M N
IsE(UIML)
12000~-I~ 10000V~
IN VITRO CIRCULATION PLASMA VOLUME: 77 ML
Np
FLOWRATE: 0.66 ML/MIN IGE ADSORBENT: 0.3 G (12 MG OF ANTI-IGE)
X
8 0 0 0 F ' Q , , ~ , 6000
.
o
.
0
N
i
Fig. 2. IgE adsorption breakthrough curve. 0.4 g of the CPG500-based adsorbent, sterilized, was placed in a column (0.5 cm ID) and plasma containing IgE was fed at 1.2/xg of IgE/mg of antibody/min. The initial IgE levels were 2800 U / m l for 10 mg/g CPG and 10000 U / m l for 35 mg/g CPG.
200o L L _
~
300
TIME (MIN)
I
A
l
1
.
.
.
.
.
i 2
.
.
=
=
.
_
OUrLFT
J
3
4
The second column consisting of the immobilized ligand on CPG-500 was sterilized. Plasma containing a-IgE ab at 100 ng/ml was fed to the second column at various flow rates. The amounts of a-lgE ab adsorbed are given as % of the initial amount.
Ligands
Immobilization amount (mg/g CPG)
Anti-goat IgG antibody
1 8 8 15
Flow rate (ml/min)
Removal (%)
20 20 4 10
10 65 80 80
10 5 20 10
60 70 86 94
TIME (HOUR)
Fig. 3. In vitro IgE adsorption.0.3 g of the CPG-500-basedIgE adsorbent, sterilized, contained 12 mg of a-IgE ab; 77 ml of plasma containing IgE was perfused over the column at a flow rate of 0.66 ml/min.
IgE
1 1 2.6 2.6
166
creased with an increase in the immobilization amounts of ligands or with a decrease in the flow rates of plasma. The immobilized IgE removed more a-IgE ab than did the immobilized anti-goat IgG antibody in the column experiment and in the equilibrium adsorption.
Discussion
The IgE adsorbent is goat polyclonal a-IgE ab covalently immobilized on CPG beads, packed in a column, and sterilized with EOG. Pathophysiologically, targeting specific IgE antibodies is important in managing allergic patients. However, our adsorbent can remove all specific IgE antibodies because goat polyclonal a-IgE ab produced by immunization with whole IgE is employed. Subsequently, serum IgE rather than specific IgE antibodies was taken into account as a variable in this study. Therapeutic IA must be provided in sterilized form. EOG was used for the IA described here in spite of causing a 40% loss in adsorption capacity (data not shown). We selected it because it is popular and hence, more likely to be accepted for medical devices; it can be applied to dry IA which facilitates transportation, storage and stability; it is simple to use and good for mass production; and it does not require any special facilities. Sterilization with 7-rays (Sato et al., 1986b) or propylene oxide aqueous solution (Sato et al., 1985) has inherent difficulties. The adsorption isotherms of IgE, using the affinity-purified a-IgE ab produced a 13-fold greater adsorption of IgE than those of the IgG fraction. This difference has advantages in terms of practical use: adopting the affinity-purified antibody as a ligand can reduce the amounts of ligand to be immobilized, and can therefore reduce the column size dimension, which means a lower extracorporeal volume of the patient's blood. The use of the affinity-purified a-IgE ab can also reduce the amounts of a-IgE ab detached into the plasma from the carrier, because the amounts of antibody detached depend on the. immobilization amounts (Sato et al., 1987b). These benefits are in terms of the safety aspect of the therapeutic adsorption system.
The IgE IA using the affinity-purified a-IgE ab worked well over a wide range of IgE concentrations. The constant, n was 0 < n < 1, indicating favorable adsorption throughout (Hiester et al., 1963). However, in comparing the immobilization amounts, more than 38 mg of the antibody per I g of CPG showed less IgE adsorbed. This means that not all of the antibody molecules densely immobilized on the surface are effectively reacted IgE molecules because of steric hindrance. This difference of immobilization amounts lies in efficiency of adsorption, but not in absolute amounts of IgE adsorbed. If more antibody is immobilized, more IgE is adsorbed until saturation is reached under the present conditions. In the column adsorption study (Fig. 2), although both IAs showed a similar profile at the beginning, the IA with the greater a-IgE ab had the wider adsorption wave at a later time. A theory of affinity chromatography has been developed to describe a heterogeneous, nonequilibrium system in which the diffusion of solutes from a bulk fluid to the outer surface of the adsorbents, the diffusion inside the adsorbent particles, and the adsorption reaction are all occurring simultaneously. The two diffusion processes are considered to control the rate of adsorption (Katoh et al., 1978). The data presented in Fig. 2 indicate that IgE molecules were rapidly adsorbed on apparent (geometrical) surfaces of both IAs immobilizing 10 and 35 mg of a-IgE ab per 1 g of CPG at an early stage. However, as the adsorption proceeds, it would take longer for the IgE molecules to penetrate into the pores, the surface of which immobilizes more a-IgE ab. In other words, the intraparticle diffusion of IgE is taken to be a rate-controlling step. Theoretically, the smaller the amount of a-IgE ab immobilized, the better the column adsorption as well as the equilibrium adsorption described above. However, to remove clinically requested large amounts of IgE from the plasma by a single, as small as possible IgE adsorbent column, greater amounts of the antibody per CPG must be immobilized, because a patient's extracorporeal blood volume is strictly limited, and because a long column with a small diameter, which causes a dangerously high pressure drop, cannot be used. 41 mg of IgE must be adsorbed per treatment
167 provided that the serum IgE level is decreased from 10000 to 3000 U / m l in an adult. The equilibrium adsorption of IgE indicated that the amount of IgE adsorbed by the sterilized adsorbent was 180 # g / m g of the affinity-purified a-IgE ab at a concentration of 3000 U / m l of IgE (data not shown). As the efficiency of adsorption in a column is assumed to be 70% of that in the batch equilibrium manner, 1 mg of the antibody immobilized could adsorb 126 #g of IgE; 325 mg of the antibody would be required to adsorb 41 mg of IgE per single treatment. In fact, 325 mg or more of the antibody was designed to be immobilized on 10 g of CPG-500 for clinical use (see materials and methods section). The perfusion test in vitro simulating clinical situations was carded out assuming a one-pool model. Since the most effective dimension of a column for the IgE adsorption was not employed because of the high pressure drop down the column, IgE appeared at the outlet of the column from the beginning of the perfusion (Fig. 3). Nonetheless, it did not exceed 1500 U/ml. The appearance of IgE at the outlet seems to be due to the short column size. On the other hand, IgE at the inlet decreased rapidly in spite of the appearance of IgE at the outlet. The reduction profile of the inlet IgE levels met the original requirement. The amounts of antibody (IgG) detached from our CPG-based IA were less than those from a Sepharose-based IA (Sato et al., 1987b). Nevertheless, the amounts were not negligible. The IgE adsorbent column was found to be safe in both repeated immunoadsorptions with a normal monkey and a single immunoadsorption with a monkey producing specific IgE antibodies, where 9 to 15 ng of a-IgE ab/ml of flowing plasma was infused into the animal during the perfusion (Azuma et al., 1988). However, because such small amounts of a-IgE ab, which can trigger a Type I hypersensitivity reaction (Ishizaka et al., 1979), appear in the plasma stream of the patient, and because the difference in sensitivity to a-IgE ab between monkeys and humans is unknown, we incorporated the second column which can adsorb a-IgE ab to allow a margin of safety. To overcome the problem of a Type I hypersensitivity reaction caused by a-IgE ab, Yamamoto et al. (1988a)
recently began using the Fab portion of a-IgE ab, which allows the sensitized mast cells to release little histamine in vitro. In terms of the adsorption capacity it was better to adopt IgE than anti-goat IgG antibody as a ligand; the antigen immobilized rather than the antibody immobilized was preferable in the present range of comparatively low concentrations of a-IgE ab. Of course, situations seem to vary depending on the combination of antigen and antibody. The results of the column adsorption also indicate that IgE was preferable as a ligand; its effectiveness was observed to be higher at the greater amount of immobilization or at the slower flow rate of plasma. Very small amounts of IgE would be detached from the second column (not determined); however, as it is of human origin and the amounts may be much smaller than present in the serum IgE of patients, the amount of IgE detached is considered to be acceptable. Immobilization of IgE at 2.6 m g / g of CPG was selected for further studies. In studies with animals, the largest amount of a-IgE ab that did not give any toxic reactions was estimated to be 20 #g per human adult (Sato and Kidaka, 1988). If 86% of 20 ng of a-IgE ab/ml are removed by the second column according to the results in Table I, a total of 7.3/~g is infused into an adult (2.6 liters of plasma assumed). This amount is less than the 20 #g described above; therefore, the system can be considered safe from the viewpoint of detached a-IgE ab. In fact, a plasma perfusion over the IgE adsorbent column followed by the second column was carried out with a healthy volunteer and a patient suffering bronchial asthma and atopic dermatitis in a clinical setting and no toxicity was observed (Yamamoto et al., 1988b). This immunoadsorption system removes IgE effectively and safely.
Acknowledgements This work was supported by a Grant-in-Aid from National Research and Development Program for Medical and Welfare Apparatus of the Japanese Ministry of International Trade and Industry.
168
The authors thank Mrs. M. Chikano for technical assistance.
References Azuma, J., Sato, H., Watanabe, K., Kidaka, T. and Hori, M. (1988) Development of IgE adsorption system with immunoadsorbents. In press (Japanese). Bambauer, R., Jutzler, G.A., Micka, K., Austgen, M., Schlimmer, P. and Tredelenburg, F. (1984) Drng-resistant bronchial asthma successfully treated with plasma exchange. J. Clin. Apheresis 2, 200. Behm, E., Kuroda, T., Yamawaki, N., Tsuda, N., Loth, F., Schwachula, G., Sabrowski, E., Falkenhagen, D. and Klinkmann, H. (1987) In vitro investigations with selective adsorbents for IgE and IgM. Biomater. Artif. Cells Artif. Organs 15, 101. Carrel, S., Theilk~is, L., Morell, A., Skvaril, F. and Barandun, S. (1971) Isolation of normal immunoglobulin E by means of an immunoadsorbent. Biochem. J. 122, 405. Dau, P.C. (1988) Remission of hyper-IgE syndrome treated with plasmapheresis and cytotoxic immunosuppression. J. Clin. Apheresis 4, 8. Gartmann,. J., Grob, P. and Frey, M. (1978) Plasmapheresis in severe asthma. Lancet ii, 40. Hiester, N.K., Vermeulen, T. and Klein, G. (1963) In: R.H. Perry, C.H. Chilton and S.D. Kirkpatrick (Eds.), Chemical Engineers' Handbook, 4th edn. McGraw-Hill, New York, 16-18. Ikeyama, S., Nakagawa, S., Arakawa, M., Sugino, H. and Kakinuma, A. (1986) Purification and characterization of IgE produced by human myeloma cell line, U266. Mol. Immunol. 23, 159. Ishikawa, I., Fukuda, Y., Kitada, H., Yuri, T., Shinoda, A., Hayakawa, Y., Sawano, K. and Usui, T. (1982) Plasma exchange in a patient with hyper-IgE syndrome. Ann. Allergy 49, 295. Ishizaka, T., Foreman, J.C., Sterk, A.R. and Ishizaka, K. (1979) Induction of calcium flux across the rat mast cell membrane by bridging IgE receptors. Proc. Natl. Acad. Sci. U.S.A. 76, 5858. Johansson, S.G.O. and Ingan~is, M. (1978) Interaction of polyclonal human IgE with protein-A from Staphylococcus aureus. Immunol. Rev. 41, 248. Katoh, S., Kambayashi, T£ Deguchi, R. and Yoshida, F. (1978) Performance of affinity chromatography columns. Biotechnol. Bioeng. 20, 267.
Ray, P.K. and Raychaudhuri, S. (1982) Differential binding affinity of immobilized concanavalin A-Sepharose 4B for normal and myelomatous immunoglobulins. Biomedicine 36, 206. Sato, H. and Kidaka, T. (1988) Some aspects of immunological safety assessments of antibody which might be detached from therapeutic immunoadsorbents. Submitted to J. Biomed. Mater. Res. Sato, H., Kidaka, T. and Hori, M. (1985) Sterilization of therapeutic immunoadsorbents with aqueous propylene oxide solution. Int. J. Artif. Organs 8, 109. Sato, H., Kidaka, T. and Hori, M. (1986a) Effect of pore size of porous bead carriers immobilizing antibody on IgE adsorption. J. Biomed. Mater. Res. 20, 853. Sato, H., Kidaka, T. and Hori, M. (1986b) Sterilization of therapeutic immunoadsorbents by ionizing radiation. Int. J. Artif. Organs 9, 131. Sato, H., Kidaka, T. and Hori, M. (1987a) Sterilization of therapeutic immunoadsorbents with chemical sterilants or ionizing radiation. In: T. Oda, Y. Shiokawa and N. Inoue (Eds.), Proceedings of the First International Congress of the World Apheresis Association: Therapeutic Plasmapheresis (VI). ISAO Press, Cleveland, OH, p. 509. Sato, H., Kldaka, T. and Hori, M. (1987b) Leakage of immobilized IgG from therapeutic immunoadsorbents. Appl. Biochem. Biotechnol. 15, 145. Sato, H., Watanabe, K., Azuma, J., Kidaka, T. and Hori, M. (1988) Preclinical evaluation of therapeutic immunoadsorbent (IA) for removing IgE. Efficacy and Safety. Abstracts of the 2nd International Congress of the World Apheresis Association. p. 13. Solley, G.O., Gleich, G.J., Jordon, R.E. and Schroeter, A.L. (1976) The late phase of the immediate wheal and flare skin reaction. Its dependence upon IgE antibodies. J. Clin. Invest. 58, 408. Yamamoto, Y., Fukukawa, T., Toba, H. and Tsuyuguchi, I. (1988a) Use of monoclonal antibody for the preparation of IgE specific immunoadsorbent. Abstracts of the 2nd International Congress of the World Apheresis Association. p. 14. Yamamoto, Y., Toba, H., Hanamoto, S., Tsuyuguchi, I., Fukukawa, T., Sato, H., Watanabe, K., Azuma, J. and Hori, M. (1988b) Efficacy and safety of IgE-specific adsorption system in human trials. In press (Japanese). Zik~m, J., Zav~tzal, V. and Krauz, V. (1984) Heterogeneity of human polyclonal IgE reacting with staphylococcal protein A. Folia Microbiol. 29, 264.
161
Elsevier JIM05105
Specific removal of IgE by therapeutic immunoadsorption system H i r o s h i Sato, K o i c h i W a t a n a b e , Junichi A z u m a , T e r u h i s a K i d a k a a n d M a s a t a k e H o r i Takeda ChemicalIndustries, Ltd., Medical Supply and Equipment Department, 17-85Jusohonmachi 2-chome, Yodogawa-ku, Osaka 532, Japan
(Received3 October 1988, accepted 8 November 1988)
A therapeutic immunoadsorption system was developed that can remove IgE effectively and specificall) from the plasma of patients with an allergy or other hyper-IgE syndrome. The immunoadsorbent (IA) consists of immunoaffinity purified anti-IgE antibody (a-IgE ab) immobilized on controlled pore glass beads (50 nm pore size). Adsorption isotherms for IgE, which were reduced by the Freundlich adsorption equation, were obtained with IA that immobilized various amounts of a-IgE ab. An optimum amount of a-IgE ab to be immobilized was selected. IA worked sufficiently in a wide range of IgE concentrations. Clinical treatment requires an amount of 41 mg of IgE to be removed from a patient's plasma for 3 h. An IA for clinical use was designed to contain 10 g of the support binding 325 mg or more of the antibody. In fact, our study in vitro simulating a clinical case showed that serum IgE was removed by IA, as expected: the level decreased from 11 000 to 3000 U / m l after a 3 h perfusion (1 U = 2.3 ng). A very small amount of a-IgE ab (goat IgG) was found to be detached from IA by flowing plasma; the average level was 20 n g / m l , which seems to be safe. However, we installed the second column in a circuit that adsorbs a-IgE ab leaked into plasma, because the amounts of a-IgE ab infused into the patient must be minimized. The second column contained IgE immobilized on the same support, since IgE as a ligand adsorbed more a-IgE ab than did anti-goat IgG antibody. This is an effective and safe therapeutic immunoadsorption system and has been subjected to clinical tests. Key words: IgE; Immunoadsorption; Immobilization; Anti-IgE
Introduction
Type I hypersensitivities are inflammatory reactions mediated by IgE antibodies bound to mast
Correspondence to: H. Sato, Takeda Chemical Industries, Ltd., Medical Supply and Equipment Department, 17-85 Jusohonmachi 2-chome, Yodogawa-ku, Osaka 532, Japan. Abbreviations: IA, immunoadsorbent; a-IgE ab, anti-IgE antibody; CPG, controlled pore glass; DEAE, diethylaminoethyl; "Iris, tris(hydroxymethyl)aminomethane;NMWL, nominal molecular weight limit; PBS, phosphate-buffered saline; RT, room temperature; BSA, bovine serum albumin; EOG, ethyleneoxide gas.
cells and basophils; the reactions result from the release of pharmacologically active factors upon combination with allergens. The concentration of IgE in patients with allergic asthma and atopic dermatitis (eczema) is much higher than that in normal individuals. To treat such patients, plasmapheresis, or plasma exchange has been applied with or without drugs (Gartmann et al., 1978; Ishikawa et al., 1982; Bambauer et al., 1984; Dau, 1988). These investigators reported a beneficial clinical course and significant decreases in serum IgE after plasmapheresis. However, it is difficult to analyze the mechanisms of the therapy because plas-
0022-1759/89/$03.50 © 1989 ElsevierSciencePublishers B.V. (BiomedicalDivision)
162 mapheresis removes substances non-selectively and requires substituted plasma such as frozen fresh plasma from healthy donors. Therapeutic on-line plasmapheresis with IA has the advantage that it removes substances specifically in the plasma and excludes the risk of virus infections that might be introduced from the donor plasma. Some IAs for reacting IgE were prepared: burro anti-IgE (Fc)-Sepharose 2B (Solley et al., 1976), protein A-Sepharose CL-4B (Johansson and Ingan~is, 1978; Zikhn et al., 1984), concanavalin A-Sepharose 4B (Ray and Raychaudhuri, 1982), anti-IgE serum-bromoacetyl cellulose (Carrel et al., 1971), and tryptophan-polyvinyl alcohol (Behm et al., 1987). Most of these were analytical tools, and a-IgE ab as a ligand seemed to have high affinity for IgE. Therapeutic IA must meet the following requirements: high affinity, little non-specific reaction, no activation of complement and coagulation systems, little leakage of the ligand from the cartier, adequate strength, tolerable to sterilization, and no toxicity; furthermore, specialized test methods for sterility of IA are required (Sato et al., 1987a). A concrete specification of the IA in affinity to IgE was set: circulating serum IgE must be decreased from 10000 to 3000 U / m l after 3 h immunoadsorption therapy in an adult. We have developed a therapeutic immunoadsorption system for removing IgE; we report characteristics of the IgE adsorbent and second adsorbent column which are used together in the system. The second column, human IgE immobilized on CPG beads, removes a-IgE ab detached from the IgE adsorbent column.
Materials and methods
lgE Human IgE was isolated from the supernatant of the culture fluid of a human myeloma cell line, U266. Purification was achieved by a method reported by Ikeyama et al. (1986). IgE was used as an immunogen in goats, as a ligand for affinity purification of a-IgE ab, as a ligand in the second column, and as a solute in the adsorption experiments in vitro.
Anti-lgE antibody Antibody against human IgE was raised in goats by subcutaneously injecting its emulsion with Freund's complete adjuvant. Antisera were purified by precipitation with saturated ammonium sulfate to obtain immunoglobulins followed by DEAE cellulose chromatography to separate an IgG fraction. The IgG fraction was further purified by affinity chromatography, a matrix which consisted of IgE immobilized on Formyl Cellulofine (Seikagaku Kogyo, Tokyo, Japan), where aIgE ab bound to the IgE matrix was eluted by 0.2 M acetic acid containing 0.15 M NaC1 followed by neutralization with 1 M Ttis. The resulting affinity-purified a-IgE ab preparation was concentrated by ultrafiltration (Pellicon System equipped with membranes of 100 000 NMWL; Millipore, Tokyo, Japan). Immobilization All the following processes were carried out under pyrogen-free and aseptic conditions. Activation of carrier A previous study indicated that a pore size of CPG of 50 nm or larger was necessary for good adsorption (Sato et al., 1986a). 470 g of CPG (CPG-500 or CPG-1400; pore size, 50 nm and 140 nm, respectively; 120/200 mesh; Electro-Nucleonics, Fairfield, N J, U.S.A.) was washed with a six-fold volume of acetone and then with 50 liters of water. Fine particles generated were removed by decantation during both washings and the later processes. Water was removed by filtering through a glass filter. CPG was placed in 1500 ml of 10% (w/v) 3,-aminopropyltriethoxysilane(Tokyo Kasei, Tokyo, Japan), the pH adjusted to pH 5 with conc. HC1 and the mixture heated at 80 o C, followed by stirring for 2 h at 80 ° C. The CPG with the introduced aminopropyl group was washed with 60 liters of water and dried at l l 0 ° C overnight after removal of moisture by filtration. 93 g of aminopropyl CPG was placed in 1000 ml of 2.5% (v/v) glutaraldehyde solution and the mixture was then stirred at RT for 1 h. The glutaraldehyde solution was discarded, the beads washed with water and the water removed by filtration. The weight of the beads was 199 g on a wet basis.
163
Immobilization of anti-IgE antibody (preparation of IgE adsorbent) To prepare IA with different amounts of antibody immobilized, glutaraldehyde-activated CPG was reacted to affinity-purified a-IgE ab solutions with various concentrations of the antibody in 0.01 M PBS pH 7.2 for 3 h. The IgG fraction was also immobilized on CPG. Immobilization amounts were calculated from the difference in concentrations of the antibody between pre- and post-immobilization. More than 95% of the antibody applied was found to be immobilized. The resulting beads were washed with saline, then water, and the moisture was removed by filtration. CPG-500 immobilizing a-IgE ab reacted to dimethyl amine borane (Tokyo Kasei) in 0.2 M borate buffer, pH 9.0 at RT for 2 h; 12 mg of this reducing agent was used per 1 g of CPG. These beads were washed with saline. Then the beads were washed with 0.1 M acetate buffer, pH 4.0 containing 1 M NaC1 followed by 0.1 M borate buffer, pH 8.0 containing 1 M NaC1; the beads were washed twice more in the same manner; the beads were again washed with water. Finally, the beads were immersed in a 2% mannitol solution, and the excess solution was removed by filtration. The resulting beads were freeze-dried by Unitop 400L (Virtis Co., Gardiner, NY, U.S.A.).
Immobilization of IgE (preparation of second column adsorbent) The procedure for immobilizing IgE was the same as that described above for a-IgE ab, but a different ligand was used.
Immobilization of anti-goat lgG antibody Rabbit anti-goat IgG antibody (Cappel Labs., Cochranville PA, U.S.A.) was immobiliTed on CPG-500 in a manner similar to the above.
Batch IgE adsorption An adequate amount of IgE adsorbent was mixed with bovine plasma containing various amounts of IgE and incubated while being stirred gently at 20 °C for 5 days. The difference in IgE concentrations between pre- and post-incubation was used to calculate the amounts of IgE adsorbed.
Column IgE adsorption Bovine plasma containing various amounts of IgE was fed downward into a column (5-10 mm ID) at several flow rates at RT. IgE concentrations in plasma at the post-column were determined with time.
Adsorption of anti-IgE antibody Batch and column adsorption tests for a-IgE ab were performed in a manner similar to the IgE adsorption tests.
Determination of IgE IgE 'Mitsui' II (Kainos Labs., Tokyo, Japan) was used to determine the concentrations of IgE in plasma. The concentrations of IgE in the buffers were determined by the relation 1.0 m g/ m l = 1.400 optical density units at 280 nm.
Determination of anti-IgE antibody 100 #1 of IgE solution (4/~g/ml of PBS pH 7.2) was placed in a microtiter well (Immulon 600, C.A. Greiner und Sohne, Nurtingen, F.R.G.). The plates were incubated overnight at 4 o C. The solution was discarded and the wells were rinsed three times with PBS. 300 /~l of gelatin (Wako Pure Chem. Ind., Osaka, Japan, 10 m g / m l of PBS pH 7.2) was added to the well. The plates were incubated overnight at 4 ° C. After the solution was discarded and the wells were rinsed three times with PBS pH 7.2 containing 0.05% Tween 20 (PBS-T), 100 /~1 of affinity-purified a-IgE ab standards or samples appropriately diluted were added. The plates were incubated overnight at 4 ° C. The solution was discarded and the wells were rinsed four times with PBS-T. 100 /xl of biotinylated anti-goat IgG (Bethesda Res. Labs., Gaitherburg, MD, U.S.A.) diluted 1/20000 with PBS-T containing 1% BSA (Miles Labs., Elkhart, IN, U.S.A.) was added to the wells and the plates were incubated with agitation at 25°C for 2 h. The solution was discarded and the wells were rinsed four times with PBS-T. 100 /~1 of Streptavidin-horseradish peroxidase conjugate (Bethesda Res. Labs.) diluted to 1/2000 with PBS-T containing 1% BSA was added to the well and the plates were incubated with agitation at 25 °C for 1 h; the solution was then discarded. The wells were rinsed five times with PBS-T, then twice with PBS.
164
100 ~tl of o-phenylenediamine (Zymed Labs., So. San Francisco, CA, U.S.A.) was added to the well and allowed to stand for 30 min at RT. The reaction was stopped by adding 50 /~1 of 2 M sulfuric acid. Optical densities were read at 490 nm in an ELISA Analyzer ETY-96 (Toyo Sokki, Tokyo, Japan). The lowest detection of a-IgE ab was 0.1 ng/ml. The overnight incubation at 4 ° C could sometimes be replaced by incubation with agitation at 25 ° C for 2 h.
EOG sterilization Dry IAs were sterilized with 20% EOG, where the operating temperature, relative humidity, pressure, and time were 30 o C, 50%, 1 k g / c m 2, and 3 h, respectively. These conditions of EOG sterilization provided 12D, where D, or decimal reduction time was defined as the time required to inactivate 90% of the Bacillus subtilis spores present. The D value was estimated using the spores (Spordex Scale, A M S C O / M e d i c a l Products, Erie, PA, U.S.A.) (data not shown).
IgE adsorbent column The IgE adsorbent column for clinical use consisted of 400 mg of affinity-purified a-IgE ab immobilized on 10 g of CPG-500 in an acrylic resin column (29 mm ID, 50 mm H); it was sterilized.
Second column (for adsorbing a-IgE ab) The second column for clinical use consisted of 26 mg of human IgE immobilized on 10 g of CPG-500 in the same column as the IgE adsorbent; it was sterilized.
Results
Adsorption of IgE The effects of purification and the immobilization amounts of a-IgE ab on IgE adsorption were studied. Fig. 1 shows the adsorption isotherms of IgE at 20 o C, where the IgE adsorbents used were not sterilized. Fig. 1 indicates, using a log-log, that these isotherms fit straight lines well; in other words, they are well fitted to the Freundlich equation; but at the same time they are also well fitted to the Langmuir equation (not shown), which
1000 o ........ O...~." r : r ' ~ . . . . . . . .
.~, . - . ~
-
-
n , ~~, , ~ l ~ r
-~
& --~ 200 v
I00
m
....
......
r
- .O- J- .....
~..O-
20 I0
I
I0
I
10000 1000 IGE CONCENTRATION (U/ML) .... - o 2 MG/G-CPG"~ ---~ 20 MG/G-CPGI 100
100000
--~ ~l .G/G-CPG_] A,,.NITY,u,,,iE0
---.o MG/G-CPG --4, 20 MO/G-CPG
IGG TRACTION
Fig. 1. Adsorption isotherms of IgE on tgE adsorbent immobilizing a-IgE ab at 20 o C. The IgE adsorbents (CPG-500-based) were not sterilized.
suffices in a system with completely homogeneous surfaces and negligible interaction between adsorbed IgE molecules. The data were reduced using the Freundlich equation Y = m . X", where Y is the amount of substance adsorbed, X is the equilibrium concentration of the substance, and m and n are constants. The calculated m and n were 200-320 and 0.06-0.10, respectively for the affinity-purified a-IgE ab. In Fig. 1, the affinity-purified antibody immobilized was found to be able to adsorb 13 times more IgE than was the IgG fraction antibody. Among several immobilization amounts of the affinity-purified antibody, the amounts of IgE adsorbed with the antibody immobilized at 2 or 20 m g / g of CPG were greater than those at 38 or 59 m g / g of CPG. However, these slopes were almost the same. Fig. 2 shows the transient movement of the adsorption waves through the columns, with the corresponding changes in the outflow concentration of IgE expressed as % of the initial IgE level, where bovine plasma containing IgE was passed at a speed of 1.2 /~g of IgE per 1.0 mg of antibody/rain. With CPG-500 immobilizing 10 mg of antibody per 1 g of CPG, the IgE concentration reached the initial value at 3 h. On the other hand, although the adsorption wave with CPG-500 immobilizing 35 mg of antibody per 1 g of CPG was similar to that immobilizing 10 mg until 70-80 min of feed, at a later time the wave was wide, and the breakthrough curve did not become saturated.
165 %
100o°
I0 M o l o - C P
80-
x
z
60
c~o
40
i
I000
o.o-
f
•
°
I ? . . . . . . ,.e.....
I
•
°~°*"D"'@"° . . . . " ' " 4 1 " " ' Q " " "'"
IGE • . . . . . . . . . . . . ~00
~
10'
m
20
20"C
0
"~'R~ I
i
I
60
,
I
120
,
IB0
I
,
240
Fig. 3 shows the time course changes in IgE concentrations at the inlet and outlet of the IgE adsorbent column. The experimentation scale was designed at one-thirty-third of a situation in which a column would be used clinically. The IgE level at the inlet decreased rapidly from 11 000 to 4500 and 3000 U / m l at 2 h and 3 h perfusions, respectively. IgE appeared at the outlet after the perfusion began, but was kept at 1000-1500 U / m l throughout the perfusion.
Leakage of anti-IgE antibody Anti-IgE antibody was found in flowing plasma at the outlet of the IgE adsorbent column in spite of being bonded covalently; the average con-
TI-GOAT
IGG
I
0. I
I
J
I
10
100
1000
ANTI-IGE CONCENTRATION (NG/ML)
Fig. 4. Adsorption isotherms of a-IgE ab on second column immobilizing IgE or anti-goat IgG antibody. 2.6 mg of IgE or 15 mg of anti-goat IgG antibody was immobilized on 1 g of CPG-500; the resulting adsorbents were then sterilized.
centration of a-IgE ab was 20 ng/ml in the perfusion test simulating clinical situations.
Adsorption of anti-IgE antibody Fig. 4 compares the adsorption isotherms of a-IgE ab with different ligands. Immobilizing human IgE as a ligand gave much greater adsorption than did immobilizing anti-goat IgG antibody particularly at low concentrations of a-IgE ab. Table I shows the % of removal of a-IgE ab by the second column immobilizing anti-goat IgG antibody or IgE. The % removal of a-IgE ab inTABLE I % REMOVAL OF a-IgE ab IN COLUMN ADSORPTION BY THE S E C O N D C O L U M N
IsE(UIML)
12000~-I~ 10000V~
IN VITRO CIRCULATION PLASMA VOLUME: 77 ML
Np
FLOWRATE: 0.66 ML/MIN IGE ADSORBENT: 0.3 G (12 MG OF ANTI-IGE)
X
8 0 0 0 F ' Q , , ~ , 6000
.
o
.
0
N
i
Fig. 2. IgE adsorption breakthrough curve. 0.4 g of the CPG500-based adsorbent, sterilized, was placed in a column (0.5 cm ID) and plasma containing IgE was fed at 1.2/xg of IgE/mg of antibody/min. The initial IgE levels were 2800 U / m l for 10 mg/g CPG and 10000 U / m l for 35 mg/g CPG.
200o L L _
~
300
TIME (MIN)
I
A
l
1
.
.
.
.
.
i 2
.
.
=
=
.
_
OUrLFT
J
3
4
The second column consisting of the immobilized ligand on CPG-500 was sterilized. Plasma containing a-IgE ab at 100 ng/ml was fed to the second column at various flow rates. The amounts of a-lgE ab adsorbed are given as % of the initial amount.
Ligands
Immobilization amount (mg/g CPG)
Anti-goat IgG antibody
1 8 8 15
Flow rate (ml/min)
Removal (%)
20 20 4 10
10 65 80 80
10 5 20 10
60 70 86 94
TIME (HOUR)
Fig. 3. In vitro IgE adsorption.0.3 g of the CPG-500-basedIgE adsorbent, sterilized, contained 12 mg of a-IgE ab; 77 ml of plasma containing IgE was perfused over the column at a flow rate of 0.66 ml/min.
IgE
1 1 2.6 2.6
166
creased with an increase in the immobilization amounts of ligands or with a decrease in the flow rates of plasma. The immobilized IgE removed more a-IgE ab than did the immobilized anti-goat IgG antibody in the column experiment and in the equilibrium adsorption.
Discussion
The IgE adsorbent is goat polyclonal a-IgE ab covalently immobilized on CPG beads, packed in a column, and sterilized with EOG. Pathophysiologically, targeting specific IgE antibodies is important in managing allergic patients. However, our adsorbent can remove all specific IgE antibodies because goat polyclonal a-IgE ab produced by immunization with whole IgE is employed. Subsequently, serum IgE rather than specific IgE antibodies was taken into account as a variable in this study. Therapeutic IA must be provided in sterilized form. EOG was used for the IA described here in spite of causing a 40% loss in adsorption capacity (data not shown). We selected it because it is popular and hence, more likely to be accepted for medical devices; it can be applied to dry IA which facilitates transportation, storage and stability; it is simple to use and good for mass production; and it does not require any special facilities. Sterilization with 7-rays (Sato et al., 1986b) or propylene oxide aqueous solution (Sato et al., 1985) has inherent difficulties. The adsorption isotherms of IgE, using the affinity-purified a-IgE ab produced a 13-fold greater adsorption of IgE than those of the IgG fraction. This difference has advantages in terms of practical use: adopting the affinity-purified antibody as a ligand can reduce the amounts of ligand to be immobilized, and can therefore reduce the column size dimension, which means a lower extracorporeal volume of the patient's blood. The use of the affinity-purified a-IgE ab can also reduce the amounts of a-IgE ab detached into the plasma from the carrier, because the amounts of antibody detached depend on the. immobilization amounts (Sato et al., 1987b). These benefits are in terms of the safety aspect of the therapeutic adsorption system.
The IgE IA using the affinity-purified a-IgE ab worked well over a wide range of IgE concentrations. The constant, n was 0 < n < 1, indicating favorable adsorption throughout (Hiester et al., 1963). However, in comparing the immobilization amounts, more than 38 mg of the antibody per I g of CPG showed less IgE adsorbed. This means that not all of the antibody molecules densely immobilized on the surface are effectively reacted IgE molecules because of steric hindrance. This difference of immobilization amounts lies in efficiency of adsorption, but not in absolute amounts of IgE adsorbed. If more antibody is immobilized, more IgE is adsorbed until saturation is reached under the present conditions. In the column adsorption study (Fig. 2), although both IAs showed a similar profile at the beginning, the IA with the greater a-IgE ab had the wider adsorption wave at a later time. A theory of affinity chromatography has been developed to describe a heterogeneous, nonequilibrium system in which the diffusion of solutes from a bulk fluid to the outer surface of the adsorbents, the diffusion inside the adsorbent particles, and the adsorption reaction are all occurring simultaneously. The two diffusion processes are considered to control the rate of adsorption (Katoh et al., 1978). The data presented in Fig. 2 indicate that IgE molecules were rapidly adsorbed on apparent (geometrical) surfaces of both IAs immobilizing 10 and 35 mg of a-IgE ab per 1 g of CPG at an early stage. However, as the adsorption proceeds, it would take longer for the IgE molecules to penetrate into the pores, the surface of which immobilizes more a-IgE ab. In other words, the intraparticle diffusion of IgE is taken to be a rate-controlling step. Theoretically, the smaller the amount of a-IgE ab immobilized, the better the column adsorption as well as the equilibrium adsorption described above. However, to remove clinically requested large amounts of IgE from the plasma by a single, as small as possible IgE adsorbent column, greater amounts of the antibody per CPG must be immobilized, because a patient's extracorporeal blood volume is strictly limited, and because a long column with a small diameter, which causes a dangerously high pressure drop, cannot be used. 41 mg of IgE must be adsorbed per treatment
167 provided that the serum IgE level is decreased from 10000 to 3000 U / m l in an adult. The equilibrium adsorption of IgE indicated that the amount of IgE adsorbed by the sterilized adsorbent was 180 # g / m g of the affinity-purified a-IgE ab at a concentration of 3000 U / m l of IgE (data not shown). As the efficiency of adsorption in a column is assumed to be 70% of that in the batch equilibrium manner, 1 mg of the antibody immobilized could adsorb 126 #g of IgE; 325 mg of the antibody would be required to adsorb 41 mg of IgE per single treatment. In fact, 325 mg or more of the antibody was designed to be immobilized on 10 g of CPG-500 for clinical use (see materials and methods section). The perfusion test in vitro simulating clinical situations was carded out assuming a one-pool model. Since the most effective dimension of a column for the IgE adsorption was not employed because of the high pressure drop down the column, IgE appeared at the outlet of the column from the beginning of the perfusion (Fig. 3). Nonetheless, it did not exceed 1500 U/ml. The appearance of IgE at the outlet seems to be due to the short column size. On the other hand, IgE at the inlet decreased rapidly in spite of the appearance of IgE at the outlet. The reduction profile of the inlet IgE levels met the original requirement. The amounts of antibody (IgG) detached from our CPG-based IA were less than those from a Sepharose-based IA (Sato et al., 1987b). Nevertheless, the amounts were not negligible. The IgE adsorbent column was found to be safe in both repeated immunoadsorptions with a normal monkey and a single immunoadsorption with a monkey producing specific IgE antibodies, where 9 to 15 ng of a-IgE ab/ml of flowing plasma was infused into the animal during the perfusion (Azuma et al., 1988). However, because such small amounts of a-IgE ab, which can trigger a Type I hypersensitivity reaction (Ishizaka et al., 1979), appear in the plasma stream of the patient, and because the difference in sensitivity to a-IgE ab between monkeys and humans is unknown, we incorporated the second column which can adsorb a-IgE ab to allow a margin of safety. To overcome the problem of a Type I hypersensitivity reaction caused by a-IgE ab, Yamamoto et al. (1988a)
recently began using the Fab portion of a-IgE ab, which allows the sensitized mast cells to release little histamine in vitro. In terms of the adsorption capacity it was better to adopt IgE than anti-goat IgG antibody as a ligand; the antigen immobilized rather than the antibody immobilized was preferable in the present range of comparatively low concentrations of a-IgE ab. Of course, situations seem to vary depending on the combination of antigen and antibody. The results of the column adsorption also indicate that IgE was preferable as a ligand; its effectiveness was observed to be higher at the greater amount of immobilization or at the slower flow rate of plasma. Very small amounts of IgE would be detached from the second column (not determined); however, as it is of human origin and the amounts may be much smaller than present in the serum IgE of patients, the amount of IgE detached is considered to be acceptable. Immobilization of IgE at 2.6 m g / g of CPG was selected for further studies. In studies with animals, the largest amount of a-IgE ab that did not give any toxic reactions was estimated to be 20 #g per human adult (Sato and Kidaka, 1988). If 86% of 20 ng of a-IgE ab/ml are removed by the second column according to the results in Table I, a total of 7.3/~g is infused into an adult (2.6 liters of plasma assumed). This amount is less than the 20 #g described above; therefore, the system can be considered safe from the viewpoint of detached a-IgE ab. In fact, a plasma perfusion over the IgE adsorbent column followed by the second column was carried out with a healthy volunteer and a patient suffering bronchial asthma and atopic dermatitis in a clinical setting and no toxicity was observed (Yamamoto et al., 1988b). This immunoadsorption system removes IgE effectively and safely.
Acknowledgements This work was supported by a Grant-in-Aid from National Research and Development Program for Medical and Welfare Apparatus of the Japanese Ministry of International Trade and Industry.
168
The authors thank Mrs. M. Chikano for technical assistance.
References Azuma, J., Sato, H., Watanabe, K., Kidaka, T. and Hori, M. (1988) Development of IgE adsorption system with immunoadsorbents. In press (Japanese). Bambauer, R., Jutzler, G.A., Micka, K., Austgen, M., Schlimmer, P. and Tredelenburg, F. (1984) Drng-resistant bronchial asthma successfully treated with plasma exchange. J. Clin. Apheresis 2, 200. Behm, E., Kuroda, T., Yamawaki, N., Tsuda, N., Loth, F., Schwachula, G., Sabrowski, E., Falkenhagen, D. and Klinkmann, H. (1987) In vitro investigations with selective adsorbents for IgE and IgM. Biomater. Artif. Cells Artif. Organs 15, 101. Carrel, S., Theilk~is, L., Morell, A., Skvaril, F. and Barandun, S. (1971) Isolation of normal immunoglobulin E by means of an immunoadsorbent. Biochem. J. 122, 405. Dau, P.C. (1988) Remission of hyper-IgE syndrome treated with plasmapheresis and cytotoxic immunosuppression. J. Clin. Apheresis 4, 8. Gartmann,. J., Grob, P. and Frey, M. (1978) Plasmapheresis in severe asthma. Lancet ii, 40. Hiester, N.K., Vermeulen, T. and Klein, G. (1963) In: R.H. Perry, C.H. Chilton and S.D. Kirkpatrick (Eds.), Chemical Engineers' Handbook, 4th edn. McGraw-Hill, New York, 16-18. Ikeyama, S., Nakagawa, S., Arakawa, M., Sugino, H. and Kakinuma, A. (1986) Purification and characterization of IgE produced by human myeloma cell line, U266. Mol. Immunol. 23, 159. Ishikawa, I., Fukuda, Y., Kitada, H., Yuri, T., Shinoda, A., Hayakawa, Y., Sawano, K. and Usui, T. (1982) Plasma exchange in a patient with hyper-IgE syndrome. Ann. Allergy 49, 295. Ishizaka, T., Foreman, J.C., Sterk, A.R. and Ishizaka, K. (1979) Induction of calcium flux across the rat mast cell membrane by bridging IgE receptors. Proc. Natl. Acad. Sci. U.S.A. 76, 5858. Johansson, S.G.O. and Ingan~is, M. (1978) Interaction of polyclonal human IgE with protein-A from Staphylococcus aureus. Immunol. Rev. 41, 248. Katoh, S., Kambayashi, T£ Deguchi, R. and Yoshida, F. (1978) Performance of affinity chromatography columns. Biotechnol. Bioeng. 20, 267.
Ray, P.K. and Raychaudhuri, S. (1982) Differential binding affinity of immobilized concanavalin A-Sepharose 4B for normal and myelomatous immunoglobulins. Biomedicine 36, 206. Sato, H. and Kidaka, T. (1988) Some aspects of immunological safety assessments of antibody which might be detached from therapeutic immunoadsorbents. Submitted to J. Biomed. Mater. Res. Sato, H., Kidaka, T. and Hori, M. (1985) Sterilization of therapeutic immunoadsorbents with aqueous propylene oxide solution. Int. J. Artif. Organs 8, 109. Sato, H., Kidaka, T. and Hori, M. (1986a) Effect of pore size of porous bead carriers immobilizing antibody on IgE adsorption. J. Biomed. Mater. Res. 20, 853. Sato, H., Kidaka, T. and Hori, M. (1986b) Sterilization of therapeutic immunoadsorbents by ionizing radiation. Int. J. Artif. Organs 9, 131. Sato, H., Kidaka, T. and Hori, M. (1987a) Sterilization of therapeutic immunoadsorbents with chemical sterilants or ionizing radiation. In: T. Oda, Y. Shiokawa and N. Inoue (Eds.), Proceedings of the First International Congress of the World Apheresis Association: Therapeutic Plasmapheresis (VI). ISAO Press, Cleveland, OH, p. 509. Sato, H., Kldaka, T. and Hori, M. (1987b) Leakage of immobilized IgG from therapeutic immunoadsorbents. Appl. Biochem. Biotechnol. 15, 145. Sato, H., Watanabe, K., Azuma, J., Kidaka, T. and Hori, M. (1988) Preclinical evaluation of therapeutic immunoadsorbent (IA) for removing IgE. Efficacy and Safety. Abstracts of the 2nd International Congress of the World Apheresis Association. p. 13. Solley, G.O., Gleich, G.J., Jordon, R.E. and Schroeter, A.L. (1976) The late phase of the immediate wheal and flare skin reaction. Its dependence upon IgE antibodies. J. Clin. Invest. 58, 408. Yamamoto, Y., Fukukawa, T., Toba, H. and Tsuyuguchi, I. (1988a) Use of monoclonal antibody for the preparation of IgE specific immunoadsorbent. Abstracts of the 2nd International Congress of the World Apheresis Association. p. 14. Yamamoto, Y., Toba, H., Hanamoto, S., Tsuyuguchi, I., Fukukawa, T., Sato, H., Watanabe, K., Azuma, J. and Hori, M. (1988b) Efficacy and safety of IgE-specific adsorption system in human trials. In press (Japanese). Zik~m, J., Zav~tzal, V. and Krauz, V. (1984) Heterogeneity of human polyclonal IgE reacting with staphylococcal protein A. Folia Microbiol. 29, 264.