CA2001358A1 - Specific removal of ldl from blood - Google Patents

Abstract

Abstract Disclosed herein are particularly efficient methods to reduce cholesterol in patients by treating whole blood with agarose beads conjugated to anti-LDL
antibodies or, especially when conjugated through the mediation of an amino functionality, to a polyanionic sulfate. In preferred embodiments, the beads are characterized as comprising 0.5-3% agarose, preferably 2%
agarose, and having diameters of 200-300 micron.
Preferred antibodies are cross-reactive with C3D1 anti-bodies.

Description

2C~ 5~

SPECIFIC REMOVAL OF LDL FROM BLOOD

Technical Field The invention relates to clinical practices for regulating blood components. Specifically, it concerns methods and materials useful in the lowering ~f cholesterol levelc in plasma through the removal of LDL
from whole blood.

Backaround AIt Extracorporeal treatment of blood and plasma for various therapeutic purposes has been known for some time.
It haq been understood that it is, of course, more convenient to conduct such procedures on whole bloed as opposed to plasma since the step of separating blood cells from supernatant is eliminated and the process therefore becomes simpler and more efficient to conduct. However, typically, extracorporeal treatment is conducted by pass-2S ing the fluid over a solid matrix or a particulate sup-port, especially in a column format. Therefore, the pres-ence of red blood cells in whole blood is often detrimental simply from the standpoint of f luid dynamics.
Nevertheles~, the extracorporeal treatment o f whole blood 30 is not unprecedented. ~ ~;
For example, Losgen, H., et al, Biomat, Ned Dev, Art Ora (1978) 6:151-173 describes the use of lar~e agarose beads of 1-10 mm diameters prepared from Sepharose R. The beads were approximately 4% agarose and used to remove various components from plasma by passage of whole ~ .

~6~ ~ 35~

blood over the beads. The beads could be activated with cyanogen bromide to immobilize proteins as specific re-actants as well. The beads themselves did not seem to cause extensive hemolysis at reasonable flow rates.
Lotan, N., et al, Ent J Art Orq (1983) 6:207-213 reported the removal of paraquat from whole blood using extracorporeal adsorbents by passing whole blood through a column packed with sorbent beads containing Fuller's earth entrapped in cross-linked agarose. The beads were 1.2-2 mm in diàmeter and were formed by mixing Sepharose 4B gel with 17% (9~ by weight) Fuller's earth powder and in~ect-ing the mlxture as drops into cold organic solvent. The beads were treated with epichlorohydrin to stabilize them.
Comparison was made with systems using cellulose-coated activated charcoal. The Fuller's earth/agarose beads were said to be more specific for paraquat removal than the cellulose-treated activated charcoal.
Marcus, L., et al, Biomat, Med Dev, Art Ora (1982) 10z157-171 describes a hemoperfusion method of detoxification using polyacrolein microspheres of about 200 micron diameter encapsulated in 6.4~ agarose to obtain beads of approximately 1 mm. A review of alternate sup- -ports useful for hemoperfusion is also given in this reference. Sideman, S., et al, Life suPport SYS (1983) 1:113-125 disclose composite beads made by encap~ulating hydrous zirconium oxide powder in agarose for the removal of inorganic phosphate from whole blood. Margel, S., React PolYm Ion Exch Sorbents (1983) 4:241-250 describes -~
the u~e of various agarose polymeric microspherical beads of 1 mm diameter for use in hemoperfusion.
Thus, there are a number of preparations of solid supports, typically adsorbent materials encapsulated with a high concentration agarose to obtain spheres of 1 mm or more to remove materials from the bloodstream. Such supports have been used both nonspecifically, and in con-' .: ., :;:,.
''"'',: :/.

, 2~ S~

nection with specific agents such as albumin for the removal of bilirubin and UDP-glucuronyl transferase to effect the glucuronidation of toxins by passage of blood over the solid support.
S The invention herein concerns the removal of apolipoprotein B (apoB) from plasma. Removal of apoB is of interest because apoB is the only protein component of low density lipoprotein (LDL) and a ma~or component of very low density lipoprotein (VLDL). LDL is the principal cholesterol carrier of human plasma, and therefore removal of this component is generally capable of lowering the cholesterol level in plasma as well. This is of significance in treatment and prevention of coronary heart disease.
In the past, LDL has been successfully removed from plasma by plasmapheresis, i.e., after separation of the components of whole blood (Thompson, G.R., Lancet (1981) 1:1246-1248; Stoffel, W., et al, Proc Natl Acad Sci USA (1981) 78:611-615). In fact, plasmapheresis at 2 week intervals with partial replacement of the plasma by albumin preparations has become an established therapy for some patients showing hypercholesterolemia. The plasma cholesterol is thus controlled by removal of the LDL car-rier and plasma cholesterol can be reduced by as much as 50~. Alternatives to this replacement include perfusion of pla~ma over columns of heparin agarose or dextran sulfate cellulose, or through molecular filters, or the precipitation of LDL from plasma at acidic pH. A specific method for removal of LDL from plasma is adsorption to anti-LDL antibody agarose beads which can be performed on-line during plasmapheresis as disclosed by Saal, S.D., et al, Am J Med (1986) 80:583-589 and by Stoffel, W., et al, Lancet (1981) 1:1005-1007.
However, to applicants' knowledge, the first description of direct removal of LDL using affinity x~l3sa chromatography from whole blood was described by Ostlund, R.E., et al, Artificial Orqans (198~ 366-374, published in November of 1987. Thi~ article, incorporated herein by reference, describes the use of 2% agarose beads of 212-300 microns derivatized with goat anti-LDL Ig to remove LDL in vitro from whole blood.
Other processes relevant to the reduction of ~' cholesterol in blood include U.S. Patents 4,656,261 and 4,654,420 which describe a water insoluble hard cellulose gel chemically sulfated to obtain a polymer which adsorbs LDL and VLDL from body fluids. The resulting support is di~clo~ed as appropriate only for plasmapheresis. U.S.
Patent 4,637,994 discloses a water insoluble porous hard ;
gel, to which is bound a qulfated moiety such as dextran i~ulfate for use in removing LDL or VLDL from body fluids.
The disclosure distinguishes the hard gels included therein from ~'30ft~' gels, such as agarose. Monoclonal antibodies reactive with specified epitopes on apolipoprotein B have also been prepared (EPO Application 0257778, published 20 July 1987).
Thus, there remains a need for an efficient method to remove LDL from plasma as a means to reduce circulating cholesterol levels by perfusing whole blood through an appropriate solid support derivatized to an LDL-~pecific ligand, the blood, when pa~sed through this solid support, should not show hemolysis or activation of complement or other~deleterious effects. In addition, the removal of LDL, and thereby the cholesterol-carrying ;
capacity of plasma, should be conducted with adsorbents 30 that efficiently and selectively remove only this , component. ;

Disclosure of the Invention The invention satisfies the need for a rapid and efficient way to treat blood extracorporeally to diminish :, 2C~358 its capacity to support the presence of cholesterol. The method is based on the direct perfusion of whole blood through an adsorbent which selectively removes LDL from the plasma. The process is completed by returning the LDL-depleted blood to the patient.
Thus, in one aspect, the invention is directed -to a method to treat whole blood which comprises passing the blood over a solid phase consisting essentially of 0.5-3~ agarose beads, preferably 2~ agarose beads, of 10 relatively small dimensions (60-500 micron diameter, preferably 200-300 micron diameter) which have beén cross-linked and derivatized to a ligand specifically reactive with LDL or VLDL. The ligand is typically an antibody or a polyanion which is specific for apolipoprotein 8-containing moieties, such as LDL or VLDL, the best known example of which i~ dextran sulfate. This method can be used directly in con~unction with the bloodstream of the patient by placing the adsorbent in an extracorporeal shunt through which blood i9 withdrawn and passed and then returned to the patient. The invention also relates to the 0.5-3~ (preferably 2~) agaro~e 60-500 micron (preferably 200-300 micron) beads derivatized to ligands, which constitute the affinity support.
In another aspect, the invention relate~ to a method to selectively remove LDL from plasma which method comprises passing whole blood through a support consisting essentially of 0.5-3%, preferably 2~, agaro~e beads of 60-500 micron, preferably 200-300 micron, diameter to which . is derivatized a monoclonal antibody preparation crosq-reactive with a particular monoclonal preparation secreted by the cell line described below. The monoclonal prepara-tion is reactLve with the N-terminal portion of apolipoprotein B, and is designated C3Dl. The invention also includes monoclonal antibodies cross-reactive with 35 C3D1 and cell lines capable of secreting them. -,' ~ ';' ~' :

2Q~1358 In still another aspect, the invention i9 .
directed to a method to treat whole blood which comprises passing the blood over a solid phase consisting es-sentially of 0.5-3%, preferably 2%, agarose beads of 60-500 micron, preferably 200-300 micron, diameter which have been cross-linked and derivatized to a polyanion specific for LDL and VLDL, preferably through an amino functionality. This support, too, can be used directly with the bloodstream of the patient and is useful to remove LDL or VLDL from blood or plasma. In another aspect, the invention is directed to these derivatized supports.
The invention is also directed to a column sterilization method.
Brief DescriPtion of the Drawinas Figure 1 shows a map of the apoB protein along with the binding sites for various monoclonal antibody preparations.
Figure 2 shows a diagram of a hemoperfusion ap-paratus which includes the immunosupport of the invention.
Figure 3 shows the bindinq of anti-LDL to sup-port beads as a function of their agarose concentration.
Figure ~ shows LDL binding capacity of the various agarose concentration beads prepared as for Figure 3.
,. .
Mode~ of CarrYinq Out the Invention The methods described below employ agarose bead supports con~ugated to antibody preparations wherein the antibodies are immunoreactive with LDL or con~ugated to an LDL or VLDL-specific polyanion, such as dextran sulfate.
With regard to the embodlment using an~ibodies, it is understood that only a portion of the antibodies is required for this immunoreactivity--namely the variable ZS~ 8 :

regions of these antibodies. Therefore, the column~ may -employ antibodies or their derivatives, wherein these derivatives include, for example, F(ab')2 or Fab~ or Fab fragments. Thus, as used herein, ~antibodies or derivatives thereof immunoreactive with LDL " refers both to intact antibodies and to specifically immunoreactive fragmentc.
It is also well known that a particular antibody or immunoreactive derivative reacts with only a portion of an antigen in most cases, i.e., with a particular epitope.
Immunoglobulins or their immunoreactive derivatives which react with the same epitope are cross-reactive . Cross-reactivity among such immunoreactive molecules can be determined by the ability of one such immunoglobulin to compete with the other for binding to the antigen.
Immunoglobulins which are cross-reactive will block each other's binding to the antigen; those not cross-reactive, and which bind to different epitopes on the antigen will not. As seen in Figure 1, the apoB protein, of 4536 amino acids, has a number of identified epitopes which are specific to particular monoclonal preparations de~cribed below. Alternate monoclonal preparations may be obtained which cross-react with the same epitopes as, for example, C3D1 which represents the most useful monoclonal for the purpose of LDL removal when bound to agarose.
For use in the invention, antibodies specific for LDL or its ma~r component, apolipoprotein B, are con~ugated to the agarose beads. As described below, ;
polyclonal antisera raised against purified LDL or apolipoprotein B, and previously purified by affinity chromatography using LDL con~ugated supports are conveniently used. These antibody preparations are raised in suitable mammals, such as goats, sheep, or mice and are purified from the immune serum using a technique suf-35 ficiently specific to obtain antibodies of the required ~ ;

.
,. ' ,.' ~ ,:
' ~, Z~180~S8 specificity. Either the purified whole antibodies or theimmunoreactive fragments are used. Alternatively, a monoclonal preparation can be obtained by immunization of a subject mammal with LDL or apolipoprotein B and fusion of the antibody-secreting cells with myelomas or otherwise immortalizing them to obtain cell lines capable of secret-ing anti-LDL antibodies. The immortalized cell lines are ~creened for desired antibody secretion using standard ELISA or other immunoassay techniques as a preliminary screen. However, selection of the ideal candidate will be either by asqaying the supernatants after their con~uga-tion to the agarose support of the invention and/or by a competitive immunoassay for cross-reactivity with the monoclonal antibody disclosed herein, designated C3Dl.
These particular candidates are especially suitable for use on the supports and in the methods of the invention.
Standard methods are used to con~ugate the monoclonal or polyclonal preparations to the support. The derivatization should result in 2-10 mg/ml gel for monoclonal preparations cross-reactive with C3Dl or for polyclonal preparations.
In the alternative, rather than using antibodies a~ a source of affinity for apolipoprotein B -- i.e., LDL
or VLDL, the agarose supports of the invention may be con~ugated, preferably using the functionality of an amino group, to an LDL/VLDL-specific polyanion, such as dextran sulfate. The polyanlon specifically adsorbs LDL and VLDL
from the blood.
Such polyanions are typically polysulfated products of various saccharides or alcohols. A
particuarly preferred polyanion is dextran sulfate.
However, the polysulfated forms of other mono~accharides, oligosaccharide~ or polysaccharides can also be used, as well as those of polyhydroxy compounds such a~ glucuronic or ascorbic acids, or polyhydric alcohols such as glycerol. Polysulfation of the alcohol groups of these compound~ also provide suitable polyanions. Other apolipoprotein B-specific ligands include the sulfation products of starch, chitin, pectin, chondroitin, and the S like The preferred polyanion, dextran sulfate, is available in a range of molecular weights and a range in percentage sulfation. For use in the present invention, the dextran sulfate should have a molecular weight of 2,000-10,000, preferably around 5,000, and should have a sulfur content of 15-20~, preferably around 17~. Dextran sulfate i~ commercially available and can be made by reaction of a polysaccharide produced by Leuconostoc mesentieroides with, for example, chlorosulfonic acid.
The polyanionic sulfate ligand is preferably attached to the agarose supports of the invention by mediation of an amino group linkage supplied to the ligand as described below. Absent the mediation of an amino group, con~ugation of the ligand to the agarose supports Of the invention for example using alternate methods such as dLrect reaction using cyanogen bromide, epichlorohydrins, a polyoxirane compound such as bisepoxide or triazine halide, all as described by Tani et al in V.S. Patent 4,637,994, fails to supply adequate derivatization of agarose with ligand to make the support an effective adsorbent of VLDL, LDL or apoB.
While not~i~ntending to be bound by any theory as to the rationale for this result, it is believed that the hard gel supports used by Tani et al are more easily ;
derivatized and therefore do not require the innovation of an amine mediated linkage.
Thus, while a variety of techniques standard in the art can be used for con~ugation of the specific affin- ;~
ity antibodies to the agarose, the most effective method for attachment of the polyanionic sulfate affinity ligand is mediation by an amino ~roup. Two major approaches for effecting this mediation are herein described; both involve derivatization of the polyanionic sulfates so as to provide amino groups. In one approach, the polyanionic sulfate is activated first with cyanogen bromide, and then derivatized to a diaminoalkane such as diaminohexane. In an alternative approach, the polyanionic sulfate is first activated with a diglycidyl ether and then reacted with a diaminoalkane. The resulting polyanionic sulfates containing amino groups can then be directly bound to the agarose beads using a number of cross-linking agents including, for example, cyanogen bromide and glutaraldehyde, as exemplified below.
The resulting polyanionic sulfate derivatized lS agarose ghould have the ligand immobilized at a concentra-tlon in a range of 0.5-20 mg of the ligand/ml column volume, preferably 3 mg/ml. Below this range of effective concentration, the binding of LDL to the support is inadequate for effective removal of LDL from serum.
The derivatized supports, using either antibody preparations or polyanionic sulfate as affinity ligands are then suitable not only for plasmapheresis techniques, but also for hemoperfusion using whole blood. Of course, the derivatized supports can also be used for standard in vitro chromatographic techniques for adsorption and assay of LD~, VLDL, or other apolipoprotein ~ containing moieties .
By l'hemoperfusionl- is meant the passage of whole blood through a solid support to obtain a product blood which is different in composition from the blood initially passed through the adsorbent. The process can be conducted in several ways. For example, a portion of blood can be removed as a batch, treated with anti-coagulsnt, passed through the immunoadsorbent, and then used for whatever purpose it is intended. If, however, 5~

the hemoperfusion is intended for therapeutic purposes in an individual patient, this is generally conducted as a continuous process where the blood is recycled into the patient after passage through the appropriate support. A
typical, but of course nonlimiting, arrangement is shown in Figure 2. As shown in the figure, blood is removed from the patient and treated with an anticoagulant shown in the figure as heparin. Alternative anticoagulants can al~o be used, and are preferred, including citrates such as citrate phosphate dextrose (CPD) or the commonly used preparation anticoagulant citrate dextrose (ACD). The blood containing the anticoagulant is then pumped through a drip chamber to prevent bubble formation, and then through a column containing the adsorbent and back through lS an additional drip chamber to the patient. A variety of designs for the hemoperfusion system can be used, as well as a variety of configurations with regard to the column containing the adsorbent.
In one aspect, the invention herein is directed to an improvement in hemoperfusion to remove LDL from blood which comprises the use of 60-500 micron, preferably 200-300 micron, 0.5-3%, preferably 2%, agarose beads as a support. Particularly preferred are diameters of 212-300 microns; they have been sieved in water from 50-100 mesh ~garo8e. Chromatographic size (60-140 micron) and ;
hemoperfusion size (300-450 microns) are less desirable. ~ :~
The 212-300 micron ~200-300 micron) beads showed accept-able flow rates.
The success of this method employing anti-B
antibodies as affinity ligand is described in detail in Ostlund, R.E., et al, Artificial Orqans (1987) 11:366-374, ;~
published in November 1987 and incorporated herein by reference. Briefly, beads were prepared at vari~us agarose concentrations of 1~, 2~, 4%, 6~ and 8~ agarose, 50-100 wet mesh (Bio-Rad, Richmond, California) and cross-.. ~.

2~ 8 linked and desulfated as described by Kristiansen, T., et al, Meth EnzYmol (1974) 34:331-341 and Porath, J., et al, J Chromat (1971) 60:167-177, respectively. The bead~ were then passed through a 300 micron hand sieve, and retained on a 212 micron sieve. The beads were washed 70 times with water on the smaller sieve. The beads were activated using cyanogen bromide as described by March, S.C., et al, Anal Biochem (1974) 60:149-152. CNBr activated beads showed excellent flow rates.
After activation, the activated gel cake was added to 2-4 tlmes its volume of 0.2 M sodium bicarbonate, pH 9 containing the desired antibody or derivative thereof and rotated for two hours at room temperature. Unreacted 9 ites were blocked by rotating the gel overnight at 4C
with 0.2 M glycine, pH 8 containing 0.15 M sodium chloride.
Figure 3 ~hows ~he effect of bead agarose concentration on the capacity to be derivatized with anti-body. The antibody preparations specific for LDL were raised in goats, harvested by plasmapheresis and purified by affinity chromatography on LDL agarose as described by Semenkovich, C.F., et al, J Lab Clin Med (1985) 106:42-47.
Fragments were prepared by digestion with papain to produce Fab (Mage, M.G., Meth EnzYmol (1980) 70:142-150) or with pepsin to produce F(ab')2 as described by Hudson, L., et al, Practical Immunoloav (1976) Oxford: Blackwell Scientific Publications, pp. 186-188. As shown in Figure 3, when the beads of various concentrations were reacted as described above with 15 mg affinity purified anti-LDL
IgG per ml beads, the amount of anti-LDL antibody bound per ml gel increaqes slightly as the percentage of agaroqe in the gel increases. ~The open circles qhow the re~ults when whole antibody is used; the closed circles show the results using Fab fragments of these antibodies.) The ~ ~:

:

~ , : ,, ",~

~: i, , :: - i " ." ~ : . . . .

2~1~13S~3 amounts of antibody bound to gel were in the range of 4-10 mg/ml of gel.
The ability of the resulting derivatized beads to bind LDL was tested using an in vitro test tube assay 5 for adsorption. In the assay, the beads to be tested are suspended at 10% volume/volume in water a~d 100-200 micro-liter of the suspension are pipetted into 1.5 ml conical plastic microfuge tubes. The tubes are centrifuged for 5 seconds, the supernatant fluid removed under suction, and a 40-80 microliter sample of EDTA-bovine serum albumin buffer (0.15 M NaCl, 50 mM tris, 1 mM EDTA, 2 mg/ml BSA, pH 7.4) containing 2.5 mg 125I-labeled LDL protein/ml were added. The tubes are then vortexed and reset in the microfuge and then rotated end over end for two hours at room temperature.
In some cases, using the above protocol, the binding was 64% completed in 15 minutes and completed in 2 hours. The beads are then washed 6 times with EDTA/BSA
buffer and eluted 3 times for two hours with 0.5 ml 1 M
acetic acid. The amount of LDL originally bound is then determined in the eluate; about 5~ of the LDL consistently ~;~
remains in the pellet. Agarose beads without attached ~;
antibody or with unrelated antibody bind negligible ;~
amounts of labeled LDL.
The results of this assay for the various percentage ag~rose bead~ are shown in Figure 4. As shown in the Figure, the capacity of the gel to bind LDL was much greater at lower agarose concentrations of the gel regardless of whsther the anti-LDL antisera or fragments were u~ed. ~The solid circles are the fragments, and the open circles the whole antibody.) As shown in the figure, the capacity of the gel to bind LDL was much greater at lower agarose concentrations for the gel regardless of whether the anti-LDL antisera or fragments were used. At 2~ and 1~ gels, approximately 3.5 mg of LDL were bound per '".:~'' ....

,,, :. : .: ., - .,,, :,,, , . . , ., ,, .. ~ . , , . , , " , . . ......... . ..
,,, .,.. .. ... ~ ., , ~ - .,; .. . . . .

2~ 35~3 ml of gel while at 4% and 6~ gels this had fallen to l.S
mg per ml and at 8~ gels almost no binding occurred.
These results occurred despite the slightly higher capac-ity of the higher percentage agarose gels to bind antibody or antibody fragment. Thus, the superiority of 1-2~
agarose as an LDL immunosorbent is not due to increased amounts of antlbody attached to the beads. The enhanced ability of 1-2% agarose to adsorb LDL was also shown to be independent of whether the antibodies were bound to the gel through CNBr or glutaraldehyde derivatization.
Alternatively, the beads could be activated using glutaraldehyde according to the procedure of Cambiaso, C.L., et al, Immunochem ~1975) 12:273-278. In this case, the activated beads are washed S times in 0.1 M
sodium bicarbonate, pH 8.5 and reacted with 6 volumes of antlbody in 0.1 M sodium bicarbonate, pH 8.5 for two hours at room temperature. Unreacted groups are blocked as described above. All gel beads are washed with 1 M acetic acid with three cycles of alternating acidic and basic washes of 0.5 M NaCl containing 0.1 M sodium acetate, pH

4.5 and 0.5 M NàCl containing 0.1 M sodium bicarbonate, pH
8.3, and with water.
The superiority of 2% agarose as compared to other percentages in bead composition appears to be specific for LDL or other apolipoprotein B adsorption. In a comparable experiment, the beads were derivatized to affinity-purified rabbit anti-goat IgG using the CNBr method, and the resulting adsorbents were tested, in a ~^
manner analogous to that described above for LDL
adsorption, for their ability to adsorb 125I-labeled goat Ig. In this case, the 2% beads were markedly inferior to higher percentage agarose beads -- 2% beads adsorbed about 3 mg goat IgG per ml gel; 4% and higher beads adsorbed about 6 mg/ml or approximately twice as much.
;

The effect of the amount of antibody on the columns was also studied, and it was found that after 5 mg of antibody was bound per ml beads, there was no further improvement in the capacity of the immunoadsorbent to bind LDL. This result was obtained both for 2~ and 4~ beads.
As the amount of Fab, for example, reacted with the activated beads was increased in the range of 0-32 mg/ml, the amount of Fab bound increased monotonically to about 15 mg/ml. However, the number of mg of LDL capable of binding per ml of beads leveled off at about 2 mg/ml for 4~ and at about 4 mg/ml for 2 percent agarose beads after, in each case, about 5 mg/ml had been bound.
It has been shown that satisfactory binding can be achieved at bead sizes as low as 60 microns and using a variety of derivatization techniques but that optimum results are obtained for CNBr or glutaraldehyde derivatization and 23 gel.
As ~hown in Table 1, antibody in the amount of about 3-15 mg per ml gel could be bound using various methods including the CNBr and glutaraldehyde methods described above, as well as methods involving trichloro-triazine and adsorption methods. Considerable variation resulted in the resulting ability of the beads to bind LDL; clearly the best results were achieved for CNBr or glutaraldehyde.

, ~ ., .,.,': ;,:'' '' "',''" ', ,. .' . ., :, .
', ,, ": '''', ,~'' ~
~
: . .

, . , .. . ,. " . . ~ . .

2~358 Table I
Effectivenes~ of Various Methods Eor Attachment of AntibodY to Aaarosë Beads Mg LDB
M~ A~a M~ Ab Mg LDL bound/
~ aC~ Bead Agarose Added/ l~ound/ bound/ mg AB
Melh~ si~e (~) m1 gel m1 gel ml ~e1 bound ~JI3L 21Z-300 2 IS.0 9.8 5.70~0.37 0.582 ~;lu~ ldelly(1e 212-300 ~ 30.0 L5.0 5.01+0.60 0.334 ~ l, adsorbed ~o pl'~.-?i il-A-dgaro~eC
) C,oa~ anti-LDL 60-140 4 25.0 2.6 1.87~0.09 0.71?
~2~ l~al~bit anti-L~L 60-140 4 35.7 8 8 0.94 0.10 ~-OII-s~cciltimide è~er 80-300 6 20.010.1 2.52~0.43 0.250 llich1Orotriazinee 212-300 2 30.0 12.0 0.65~0.04 O.OS
rllclllorotriazinef 212-300 2 30.0 4.2 0.08~0.01 0.02 _. , Comparison of various methods for attaching antibodies to agarose beads.
The amount of LDL bound to the immunosorbent i5 expressed as mg LDL
protein/ml gel volume, as determined by the micro test tube assay.
aAb is afinity-purified anti-human LDL raised in goats.
bAgarose beads were derivatized with diaminohexane using the CNBr technique and then exposed to 2.5~ glutaraldehyde, followed by anti-LDL
IgG, as described ln Methods.
CProtein A-Sepharose 4B-Cl beads (2 mg protein A/ml gel) were rotated overnight with affinity-purified anti-LDL antibody. After washing twice witll EDTA-BSA buffer, the gels were assayed for 1~5I-LDL binding in the ;;
usual manner, using 1 M acetic acid as eluting agent. ~;
aAffigel-10 (Bio-Rad, Richmond, CA, U.S.A.). A 10-atom spacer arm separates the actlve ester from the agarose bead. Fab antibody fragments ~:
were bound to gel beads in 0.1 ~ sodium bicarbonate, pH 8, for 4 h at 4C.
ePrepared as in Finlay, T.H., et al, Anal Biochem (19 _ ) 87:77-90.
fPrepared as in Ultroael, maanoael, and trisacrYl. Practl~al auide for u~e ln afflnity chromatoqraDhY anc related techniaue~. V~lleneuve-La-Garenne, France: Reactifs IBF-Societe Chimique Pointet-Girard, 1983:131. ~ .
: - . .:, ., ' ' ".' ".~
, ~ i, , ,", ,,, ,",i:
"" ''~ ;~,'.

~

:,: -,'~'~, ' .
,.' ~

'~`. .,' , , ~,, . ;: .: . , . ': .

2~ 35~3 Neither carbohy~rate nor antibody leakagegreater than 100 pg/ml column effluent could be detected from cross-linked desulfated 2~ agarose immunoadsorbent beads derivatized with CNBr; and the immunoadsorbent did 5 not activate complement in vitro. Nonspecific binding of other proteins was also shown to be satisfactorily low.
The invention also includes agarose bead 5Up- .:
ports derivatized to dextran sulfate through the mediation of amino group linkages as described herein, and to methods of plasmapheresis and hemoperfusion using these supports. Detailed examples of the preparation of dextran sulfate derivatized supports is exemplified below.
Note should be taken that the immunosorbent can be reused if properly sterilized. Since autoclaving destroys the activity, alternative sterilization methods must be used, and a variety of such methods are suggested ln the above-referenced paper. A particularly preferred method employs a mixture of phosphoric acid in alcohol at a slightly elevated temperature. An illustrative mixture containing 0.34% phosphoric acid and 80% ethanol at 37C
is capable of sterilizing the immunoadsorbents for reuse.
An improved sterilant, 60-80~, preferably 70~
ethanol containing 60-120 mM, preferably 90 mM phosphate ad~usted to pH 2.5-4.0, preferably 3.0 and employed at elevated temperatures, e.g., 37C for 4-24 hours, prefer-ably 8-16 hours, has been developed. A solution of 70%
ethanol and 90 mN phQsphate, pH 3 was tested for 8 hours at 37C on 15 carrier samples contaminated with Clostridium sporoaenes resistant to 2 minute exposure to 2-5 N HCl. The carriers were withdrawn and rinsed in a tube of thioglycolate broth for 15 minutes to remove excess sterilant, and then cultured. No surviving cultures were found. Doubling the 8 hour exposure time for a safety factor, a sterilization time of 16 hours is quite similar to that of commercially available chemical 2ai~135t~3 cterilants (10-20 hours) and is a fourfold improvement in k lling time over the similar illustrative solution of 80 ethanol containing 0.34% phosphoric acid. Sixteen hours exposure of goat anti-LDL agarose beads to the tested sterilant resulted in a minimum recovery of 76.4 + 3.7 (SEM) of LDL binding capacity.

Examples The following examples are intended to il-lustrate but not to limit the invention.

Example 1 Reduction of LDL Levels in Doqs Agarose beads con~ugated to goat anti-LDL Ig were prepared as described in Ostlund, R. (supra). The beads were washed with 1% Liquinox to reduce leakage of nonspeclfically adsorbed antibody and the sorbent was shown to contain 3.24 mg antibody/ml. The beads were packed to within 1 cm of the top of a cylindrical 300 ml polycarbonate hemoperfusion canister (7.1 cm diameter x 8.2 cm long, Johnson & Johnson Cardiovascular Corp., King of Prussia, PA) under 100 cm water pressure. Blank columns were prepared in the same manner and treated with CNBr and glycine without addition of antibody. The ;~;
25 columns were disinfected by perfusing with 1.5 liters 1 M ;
acetic acid and stored until use. ~;
Two mongre~ls weighing 20 kg were perfused once per week for four weeks with blank columns and once per week for four weeks with antibody columns. The anlmals were placed in a sling and sedated with acepromazine. The hemoperfusion used an arrangement similar to that shown in Figure 2, containing a Gambro AK-10 blood monitor and standard adult hemodialysis tubing. Pres~ure monitors were connected before and after the column. The tubing was primed and acetic acid washed out of the column in a 2Qt:~358 ., 19 downward direction with 1.5 l of a solution of anticoagulant citrate dexcrose solution A (ACD-A) mixed with 13 volumes saline to 1 volume ACD-A (ACD-saline).
Sixteen gauge catheters were placed in each external jugular vein and flushed with ACD-saline. The left catheter was used to draw blood and was connected to a three-way stopcock into which was infused ACD-A at a rate 1/13 that of blood flow. Venous return to the animal was not connected until blood cells appeared in the venous drip chamber, so as to avoid providing saline to the animal. Blood was pumped at 15 ml/min for one hour and flowed through the column in an upward direction, then washed from the column in a downward direction for an ad-ditional 20-30 min. Samples for assay were taken from the animal before infusions of saline and after the procedure wa~ terminated. Samples were taken from the blood lines before and after the column after 40 minutes of blood -withdrawal. Biochemical and blood cell values were cor-rected for dilution based on the red cell concentration of the sample. The columns could be disinfected for storage after washing with saline by perfusion with 1.5 l 1 M
acetic acid at room temperatures.
The blood was pumped easily through the columns;
at a flow rate of 15 ml/min, the trans-column pressure was 15 mm Hg and at a flow rate of 30 ml/min, it was 25 mm Hg.
The arterial and venous blood pressures were 35 and 50 mm Hg, respectlvely, at~a 15 ml/min blood flow, showing that resi~tance to blood flow was due to catheters and tubing rather than the column. No clotting was observed in the columns or tubing for the most part, although small clots, not of troublesome dimensions, are sometimes observed after exten~ive perfusion.
The results in Table 2 show the effectLvene~s of the columns in removing Apo~:

2~ 5~3 -Table 2 Blood Lines Animal Pre-Col. Post-Col. Pre-Perf. Post-Perf.
5ApoB, mcg/ml ApoB, mcg/ml Dog 1 59.8+5.2 4.3+1.5 77.2+3.3 51.1+2.8 Dog 2 38.7+1.3 13.4+5.2 50.6+1.9 33.9+3.5 Percent of Pre-Procedure Sample from Animal:
, CombLned Data 77.1+3.7 14.5+5.8 100 66.5+3.3 Data for Dogs 1 and 2 represent 4 and 3 ;~
procedures, respectively. Data are corrected for red blood cell count and presented as mean + SEM.
The average ApoB reduction wa~ 81% across the column and 33.5~ in the animal before and after the procedure. ~he dog is not a precise model for the human, as the ma~or llpoprotein of the dog is high-density lipoprotein, and thus this procedure would not be expected to decrease the level of cholesterol in the plasma of dogs, per se. Measurements of blood cell counts, white blood cells and platelets, showed that blood cell 10~8 over the columns we~e clinically negligible. It was also shown that the procedure did not activate or consume com-plement, nor did the animals appear to raise significant levels of antibodies to the columns.

Example 2 Selection of Monoclonal Antibodies A panel of monoclonal antibodies was prepared by in~ection of mice with human LDL and preparation of im-2~ 58 mortalized spleen cells using the Kohler and Milstein technique. A number of cell lines which secreted antibod-ies capable of binding LDL were obtained. These antibod-ies were purified by affinity to LDL agarose and eluted with 1 M acetic acid.
A number of the antibodies were mapped against the LDL protein by competition for LDL binding and re-activity with synthetic peptides. The results of this assay are shown in Figure 1, indicating that the monoclonals C3Dl and D3D5 bind approximately at the amino terminus, D7.2 in the middle, and BlB3 in the carboxy terminal regions. BlB3 antibody was localized at the receptor binding domain of LDL.
When tested in standard immunoassays, C3Dl was Of average binding capacity to LDL, while BlB3 showed the highest affinity. These results were not the same as those obtained in the alternative assay described above, wherein LDL was provided to microfuge tubes containing the antibody derivatized to agarose beads. The immunosorbents were prepared as described above using CNBr-derivatized cross-linked 212-300 micron beads of varying agarose concentration. All of the antibodies bound significantly to the beads. In most cases, the level of binding was similar across all agarose percentages and comparable to that of goat polyclonal antibody.
However, the ability of the immobilized antibod-ies to bind the LDL in this assay was markedly different.
For 2~ agarose derivatized beads, the C3Dl monoclonal antibody-bound beads were roughly twice as effective a~
similar beads prepared with goat polyclonal antibodies.
Three alternative monoclonal lines were relatively ineffective, despite the superiority of these lines, in particular, BlB3, in standard immunoassays.
The above results were obtained using artificial serum solutions of LDL-cholesterol and freshly drawn serum ''''''''''' ~,~
- ' ' ~' .

. ~:.,.

Z Cl~ 5~3 -to provide the LDL in the microfuge tubes. Using freshly drawn serum, 2~ agarose beads containing C3Dl were capable of removing 11.8 mg cholesterol/ml of gel, whereas immunoadsorbents made from the other three antibody preparations bound less than 3.2 mg cholesterol/ml.
Literature values for immunoadsorbents are approximately 3-5 mg cholesterol/ml. Similar results were obtained when derivatization was effected using glutaraldehyde.
Eleven monoclonal preparations obtained as above were assessed for cross-reactivity with C3D1 in binding as3ays to LDL attached to microtiter plates, but no cross-reactivity was shown. None of these antibodies showed ;
binding capacity comparable to C3D1 when bound to 2 agarose.
Example 3 Production of C3D1 Cell culture containing C3Dl was deposited atthe American Type Culture Collection on 1988, with accession no. The C3D1 cell line was expanded in mice by in~ecting with the cell culture and recovery of the ascites fluid.
The C3D1 cell line can also be proliferated in vitro. For in vitro culture, C3D1 cells from a single frozen vial are recovered into a T flask which is subsequently expanded to multiple new T flasks. Cells from these flasks are used to inoculate a 500 ml ~pinner vessel, and these cells progressively expanded to 3-l and 14-1 spinner vessels, respectively. A 100-1 perfusion reactor is then inoculated with cells from 2X14-1 spinner vessels and grown in standard media supplemented with 5%
fetal bovine serum. Approximately 250 1 of conditioned medium are produced and the IgG concentrated from the med~um using LDL affinity columns.

Z~ 58 Example 4 Derivatization of Aqarose with Dextran Sulfate A. Derivatization of dextran sulfate with amino S functionality A.l In one method, 2 g CNBr was added to 12 ml water at 4C and stirred. Dextran sulfate, 29 in 2 ml ~;
water, was added and the pH wa~ raised to 11.0 with sodium ~
hydroxide and maintained for 20 minutes. After lowering ;
10 the pH to 9.0, the cyanogen bromide-activated dextran ;;
sulfate was added to 18 ml of 2.8 M diaminohexane, pH 9, and rotated for 24 hours at room temperature. The result-ing aminohexyl-dextran sulfate was separated from unreacted CNBr and diaminohexane by dialysis against water in 1 kd MW cutoff dialysis tubing. The product contained 0.7 moles amino group per mole sulfate group.
A.2 In a second method, 0.5 g dextran sulfate was dissolved in 0.5 ml water and added to 0.6 ml 0.6 N
sodium hydroxide and 0.6 ml butanediol diglycidyl ether.
The mixture was rotated at 25C for 10 hr and then neutralized with HCl. The remaining diglycidyl ether was removed by extraction 11 times with diethyl ether. 0. 5 ml of the resulting ac~ivated dextran sulfate preparation was added to 0. 5 ml 2. 8 M diaminohexane and the pH ad~usted to 12. The sOlUtion was rotated at 45C for 24 hr and then dialyzed again~t water to remove unreacted diaminohexane.
The product -dextran sulfate - OCH2CHOHCH2O(CH2)4OCH2CHOHCH2NH(CH2) 6 NH2 contained 0.7 moles amino group per mole sulfate.

B. Con~uaation to Aqarose suPport~ ,:
B.l In a method using CNBr, 2~, 4% or 6~
crosslinked agarose beads of 212-300 microns diameter, .. . ~...
' ' ; ~' ' ~ ' ' .' ~ , ' "';

Z~ 553 prepared as described in Ostlund, R.E., et al, ArtifiCial Orqans (1987) 11:366-374, cited above, were activated with cyanogen bromide as there described, and reacted with aminohexyl-derivatized dextran sulfate prepared as in paragraph A.l of this example using a concentration of 38 mM amine. The reacted support was washed, and residual amino groups were blocked by reaction with 2.5~
glutaraldehyde in 0.1 M sodium bicarbonate, pH 8.5 for 30 min at 25C, washing and then rotating for 24 hr with 0.2 M glycine pH 8Ø
The prepared support was then tested for capacity to adsorb LDL in the microfuge tube assay described above. The results, as a function of agarose bead concentration, are shown in Table 3, for beads lS derivatized using the aminohexyl dextran sulfate prepared as in A.1.

Table 3 20 Agarose LDL Binding Dextran SO4 Conc mq LDL/ml ~elConcentration ma/ml qel 2% 7.19 + 0.05 2.4 4% 5.26 + 0.12 8.8 6~ 5.76 + 0.01 8.3 A~ these results show, the 2~ agarose beads derivatized to dextrsn sulfate had a higher efficiency than higher percentage beads, especially in view of the low concentration of the affinity dextran sulfate ligand required.
B.2 In a second method, 2% crosslinked amlnohexyl agarose bead~ of 212-300 microns diameter were ~;
treated with 2.S% glutaraldehyde in 0.1 M sodium~carbonate ~ ;
buffer for 30 min at room temperature, as described in Ostlund et al (supra). Residual glutaraldehyde was ' ~ . ' ,,:: ,, Zq~ S8 removed, and either the aminohexyl-dextran sulfate as described in paragraph A.L of this example, at a concen-tration of 453 mM, or the diglycidyl ether linked aminohexyl-dextran sulfate prepared in paragraph A.2 of this example at 68 mM amino concentration was added. The mixtures were incubated at 2SC for 24 hr. The residual amino groups were blocked as described in paragraph B.l and the resulting derivatized gels tested for ability to adsorb LDL using the microfuge tube assay described above.
The results are shown in Table 4.

Table 4 LDL Binding Dextran SO4 mq LDL/ml qel Concentration mq/ml ael 15 Dextran SO4 Derivatized by ~A.1 4.72 + 0.24 0.84 ~A.2 3.88 + 0.? 0.52 These results show that linkage of dextran sulfate to agarose beads using glutaraldehyde rather than CNBr as in ~B.1 above results in a lower concentration of dextran sulfate per ml of gel and a correspondingly less efficient removal of LDL from the blood for the aminohexyl derivatized dextran sulfate of 0.7 mole amino/mole sulfate exemplified above. However, at higher levels of ~ ~-amlnohexyl derivatization of the dextran sulfate, linkage -to the beads through glutaraldehyde shows superior results to CNBr linkage.

.; ~ , .

Claims (50)

1. A method to remove LDL selectively from plasma, which method comprises:
passing whole blood containing said plasma through a matrix of 0.5-3% agarose beads of 60-500 microns in diameter derivatized to a ligand capable of adsorbing LDL.

2. The method of claim 1 wherein the agarose beads are 200-300 microns in diameter.

3. The method of claim 1 wherein the beads are 2% agarose beads.

4. The method of claim 2 wherein the beads are 2% agarose beads.

5. The method of claim 1 wherein the ligand is an antibody or derivative thereof immunoreactive with apolipoprotein B (apoB).

6. The method of claim 5 wherein the anti-apoB
antibodies are prepared by inoculation of an animal with purified LDL and recovery of the antiserum.

7. The method of claim 5 wherein the antibody or derivative is bound to agarose beads through derivatization with CNBr or glutaraldehyde.

8. The method of claim 5 wherein the antibody is a monoclonal antibody specifically immunoreactive with apolipoprotein B (apoB).

9. The method of claim 8 wherein the monoclonal antibody or derivative is cross-reactive with C3D1.

10. The method of claim 1 wherein the ligand is an apolipoprotein B-specific polyanionic sulfate.

11. The method of claim 10 wherein the ligand is dextran sulfate.

12. The method of claim 1 which comprises treating whole blood.

13. A method to reduce the level of LDL in plasma, which method comprises treating plasma or the whole blood in which the plasma is contained with 0.5-3%
agarose beads of 60-500 microns in diameter wherein said beads are derivatized to a preparation of a monoclonal antibody or derivative thereof cross-reactive with C3D1.

14. The method of claim 13 wherein the beads are 2% agarose.

15. The method of claim 13 wherein the beads are 200-300 microns in diameter.

16. The method of claim 14 wherein the beads are 200-300 microns in diameter.

17. The method of claim 13 wherein the method comprises treating whole blood.

18. The method of claim 17 which further includes the steps of withdrawing blood from a subject and returning the treated blood to the same subject.

19. The method of claim 18 wherein the process is carried out in a continuous system.

20. A composition of matter which comprises 0.5-3% agarose beads of 60-500 micron diameter derivatized with an antibody or derivative preparation specifically immunoreactive with LDL.

21. The composition of claim 20 wherein the antibody or derivative preparation is specifically immunoreactive with apoB.

22. The composition of claim 21 wherein the preparation is cross-reactive with C3D1.

23. The composition of claim 20 wherein the beads are 2% agarose beads.

24. The composition of claim 20 wherein the beads are 200-300 microns in diameter.

25. The composition of claim 23 wherein the beads are 200-300 microns in diameter.

26. A method to produce antibody useful in treatment of whole blood to reduce LDL content, which method comprises culturing immortalized cells capable of secreting antibodies cross-reactive with C3D1 and recover-ing the antibodies from the cell culture.

27. Monoclonal antibodies useful in treatment of whole blood to reduce LDL content which are cross-reactive with C3D1.

28. An inmortalized cell line which secretes the antibodies of claim 27.

29. A method to remove or reduce LDL
selectively from plasma, which method comprises:
passing whole blood containing said plasma through a matrix of 0.5-3% agarose beads of 60-500 microns in diameter derivatized to an apolipoprotein B-specific polyanionic sulfate.

30. The method of claim 29 wherein the agarose beads are 200-300 microns in diameter.

31. The method of claim 29 wherein the beads are 2% agarose beads.

32. The method of claim 30 wherein the beads are 2% agarose beads.

33. The method of claim 29 wherein the polyanionic sulfate is conjugated to the agarose beads through an amino functionality.

34. The method of claim 33 wherein the polyanionic sulfate provided with amino functionality is bound to agarose beads through derivatization with CNBr or glutaraldehyde.

35. A composition of matter which comprises 0.5-3% agarose beads of 60-500 micron diameter derivatized with an apolipoprotein B-specific polyanionic sulfate.

36. The composition of claim 35 wherein the polyanionic sulfate is conjugated to the agarose beads through an amino functionality.

37. The composition of claim 35 wherein the polyanionic sulfate is dextran sulfate.

38. The composition of claim 35 wherein the agarose beads are 200-300 microns in diameter.

39. The composition of claim 35 wherein the beads are 2% agarose beads.

40. The composition of claim 37 wherein the derivatized concentration of dextran sulfate is at least 0.5 mg per ml gel.

41. A process for derivatizing dextran sulfate to agarose beads which process comprises covalently bonding a dextran sulfate preparation which has been provided with amino functionality to a preparation of agarose beads.

42. The process of claim 41 wherein the conjugation is mediated by glutaraldehyde or cyanogen bromide.

43. The process of claim 41 wherein the agarose beads are 0.5-3% agarose.

44. The process of claim 41 wherein the dextran sulfate is provided an amino functionality by derivatization to a diaminoalkane.

45. A composition of matter prepared by the process of claim 41.

46. A method to sterilize an immunosorbent column which method comprises treating said column with a sterilant consisting essentially of 60-80% ethanol containing 60-120 mM phosphate pH 2.5-4.0 for 8-24 hours at room temperature-40°C.

47. The method of claim 46 wherein the sterliant is 90 mM phosphate in 70% ethanol, pH 3.

48. A method to remove apolipoprotein B-containing moieties from a biological fluid, which method comprises contacting said fluid with the composition of claim 20.

49. A method to remove apolipoprotein B-containing moieties from a biological fluid, which method comprises contacting said fluid with the composition of claim 35.

50. A method to remove apolipoprotein B-containing moieties from a biological fluid, which method comprises contacting said fluid with the composition of claim 45.