CA1161368A - Controlled protein fractionation - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE:
A controlled protein fractionation method in which a cold protein-containing solution is circulated in contact with one side of a membrane and the temperature is controlled by circu-lating a protein precipitant solution in contact with the oppo-site side of the membrane. The flow rate is increased and back pressure is exerted on the precipitant solution to cause it to penetrate the membrane, mix with the protein solution and pre-cipitate protein in fine particles along an inside membrane wall. The precipitated protein may be concentrated and separated by centrifugation or diafiltration. The membrane is preferably in the form of a plurality of hollow fibers. Control of process variables increases yield and purity, decreases processing time and reduces degree of denaturization. The method is especially applicable to carrying out the Cohn cold alcohol fractionation process.

Description

11613~8 BACKGROUND OF TIIE INVENTION:
F I E LD OF THE I NVF,NT I ON:
This invcntion relates generally to a protein fractionation process. More specifically, the invention is directed to novel combinations of technology that allow for the simultaneolls addi-tion of potentially h.lrmflll dcnatllrin~ solvents to protein solu-tions at controlled rates, the simultaneous efficient exchange of heat generated by the addition of these solvents to liquid, the control of concentration of protein, solvent, pH and ionic strength in complex mixtures of proteins such as blood plasma.
Also this methodolo~y can be utilized to remove organic solvents from protein solutions circumventing the harsh denaturing effect of solvent removed by freeze drying, flash evaporation or lyo-philization and eliminating need for centrifugation in certain phases of the CohQ cold ethanol process, specifically Cohn methods 6 and 9 The metho~ is ideally adapted to a continuous cold ethanol process for ~rotein or plasma fractionation.
Finally, addition of the organic solvent in the manner described in this inYention allows for the most efficient formation of small particles during precipitation. The temperature control, p~l control~ ionic strcngth control, the formation of small particles during precipitation and the agglomerating effect of the use of hollow fibers are synergistic in effect to reduce the total process time to prepare Cohn Fraction V (albumin) from S to 7 days to less than 3 (lays with high yield, high purity, and minimal denaturation.
THE PRIOR ART:
The fractionation of bloo~ proteins i5 generally practiced according to the alcohol precipitant method introduced by E. J, Cohn (U. S. Patents No. 2,390,074 and No, 2,770,616~. While fractionation of protein~ by this process has met with success, ,a~

3~8 a significant body of evidence has accumulatcd l~hich indicates that proteins isolated principally through the use of alcohol and other protein precipitants, are not natural products. Both human albumin and particularly immune gamma globulin invariably contain aggregates when isolated with ethanol. These aggregates are present in amounts which indicate that denaturation has occurred. Many commercially available products prepared by the ethanol fractionation and use of other protein precipitants con-tain up to 25 per cent of aggregates (see U. S. Patent No.
4,136,094 to Condie).
The method proposed by Cohn depends upon balancing the pre-cipitating action of the organic solvent with the solvent actions of the electrolytes present. In such a method, five independent variables have bcen controlled only to varying degrees: electrolyte conccntration, alcohol concentration, hydro-gen ion concentration, temperature, and protein concentration.
To the extent that these variables are controlled depends the purity, the yields, and the extent of denaturation of each of the plasma proteins isolated.
The Cohn process is currently practiced as a batch process.
Cooled plasma is treated with reagents in very large vessels.
That is, after pH adjustment of the plasma or subfraction thereof, the protein solution is cooled to a specific temperature at or below 0 C to decrease solvent denaturing of the protein and to achieve conditions for correct precipitation. Precipi-tation is achieved by adding the precipitant (usually ethanol) with stirring, thc quantity being predetcrmined to achieve a final concentration apl)ropriate to thc scparation of the ~esired protein fraction. 'l'he art as practiced today can neither control localized temperature variations, localized protein concentra-tion variations, nor localized variation in alcohol concentra-z 11~;13~i8 tion. It is not unusual to see localized temperature variation as great as ~ 15 C. Stirring o~ the mixture cannot be carried out efficiently without extensive foaming which is undesirable.
The process is essentially a slow one with the result that the concentration of prccipitant fluctuates continuously up to the point at which all the precipitant has been added. Con-sequently, precipitation of fractions takes place progressi~ely and a long period of aging is required in order to approach final equilibrium. It is in practice very rare to achieve the equilibrium condition required and the final product is almost invariably contaminated and denatured.
Such bulk systems also have the disadvantage that large volumes o~ plasma arc continually at risk due to plant failure and staff errors. Incorporation under appropriate conditions to obtain complete fractiona~ion cycle with the Cohn method requires 160 hours or approximately 7 days.
Cohn in U. S. Patent No. 2,469,193 taught that when a readily denatured protein such as gamma globulin is to be precipitated, considerable care should be exercised in the addition of the precipitant to the plasma or subfractionation thereof. Thus, it is recommended that after a suitable pH and temperature adjust-mént of the plasma or subfraction thereof, the precipitant be added by way of a semi-permeable membrane to avoid denaturation.
This has not been possible in large scale systems. However, addition of alcohol by diffusion through semi-permeable membranes does not control wide fluctuations in protein concentration, alcohol concentration or temperature fluctuation.
It is naturally to be expected that the process is essen-tially slow and the concentration of the precipitant varies continuously up to the point which all the precipi~ant has been added. Long periods of aging are required in order to approach 11t;1368 final equilibrium. The inordinate time requirements are docu-mented in U. S. Patent No. 3,826~740 to Jewitt, a method and apparatus for treating multi-phase systems, and U. S. Patent No.
3,769,009 of John G. Watt, directed to an impro~ed system of blood plasma fractionation.
Use of the Cohn method currently requires a large installa-tion, the yields of certain fractions are low, and it requires prolonged exposure of proteins to high alcohol content which has a denaturing effect on some of these proteins. Up to the - 10 present time, control of the five ~ariables has only been par-tially achieved and always under less than optimal conditions.
The fact that the Cohn cold ethanol fractionated gamma globulin contains ~olecular aggregates makes it unsuitable for intra-venous administration. This fact is documented and mentioned in Chanbrom (J. S. Yatent No. 3,763,135, and Falksvenden U. S.
Patent No. 3,8~9,436. Even following closely the recommended Cohn method, it has been shown that proteins isolated with alcohol as a precipitant cause albumin and immune serum globulin aggregates.
Watt (U. S. Patent No. 3,764,009) has attempted to control temperature and alcohol addition by introducing a method for the fractionation of protein solutions, particularly plasma, in which spatially projected convergent jets are combined with the protein precipitant to form a mixture in which plasma protein is instantaneously precipitated and from which the protein is subsequently separated. In practice Watt claims that the spatially projected jets are sufficiently fine, that upon com-bination with the plasma the mixing is substantially instantane-ous. The exothermic reaction of adding the alcohol to plasma and the dissipation of the generated heat is accomplished, Watt claims, if the jets merge smoothly to form a composite 1~61368 stream, impinging therea~ter on a cool surface so that undue rise in temperaturc, ~hich is liable to cause protein denatura-tion, is prevented. 'I`urbulence which is another problem in the addition of the alcohol is minimized or reduced to an insignificant level.
'I'he ~Ic~rcc tllat tllc fivc vuriablcs arc controlle~ ~eter-mines the yield, purity, and extent of denaturation of the resultant protein fractions. ~enaturation of protein can occur from a combination of any or all of the following - turbulence and foaming, high alcohol concentrations, and heat generate~
by the alcohol addition or failure to keep temperature of the solution constant or, in the last step, when alcohol is removed from the protein. With the exception of the concentration of protein which is only initially controlled at the start of the process, these variables are all essentially uncontrolled until near the final stages of each step. During the critical phases which take place where localized differences in alcohol concen-tration occur, there is high heat, high alcohol concentration, and low concentration of proteins. These wide fluctuations resulting from the inability to control critical factors result in denaturation of proteins, low purity, and low yields. In order to minimize two critical variables, namely the alcohol concentration and temperature, the alcohol is added very slowly to the protein solution and with a minimum amount of agitation ~o avoid foam denaturation. 'I'his then requires approximately 7 days for the completion of the total cycle. In addition to the extensive time requirements, the protein concentrations have never been adequately controlled. FinallyJ because of the pro-found denaturating action and exothermic release of heat when added to aqueous solutions at high concentrations, ethyl alcohol must be added slowly and at fairly low concentrations (SS per il61368 cent) resulting in requirements for combined volume four timesthat of the starting liquid or plasma.
SU~ARY OF T~IE INVENTION:
The present invention incorporates a number of principles for the precise control at any time of the six major variables of the Cohn mcthod includillg (1) protein concentration, (2) rate of alcohol addition, (3) temperature control by a unique heat exchange system, (4) control of pll and (5) conductivity, and (6) controlled agglomeration of the precipitating protein.
This allows for the first time control of the significant variables of the Cohn method. Time required for completion of the complete fractionation cycle is reduced from 7 to 3 days, yields and purity are increased, and degree of denaturation is significantly reduced Broadly stated, the invention is a controlled rapid protein fractionation method which comprises the following steps. A cold protein-containing solution, such as plasma, is circulated through a confined space defined by one side of a porous membrane, such as a hollow fiber. The temperature of the protein solution is con-trolled by circulating a liquid protein precipitant solution in ; contact with the opposite side of the membrane. By increasing the flow rate and adjusting the back pressure, the precipitant solution is caused to penetrate the membrane and cause the pre-cipitation of fine particles of plasma protein along the first side of the membrane. Then the precipitated protein particles are separated. The system may be backflushed with buffers to adjust plJ levels, Na+ conccntration and ionic strength.
BRIEF DESCRIPTION Ol: TIIE DRAWINGS:
The invention is illustrated by the accompanying drawings in which corresponding parts are identified by the same numeral5 and in which:

1~61368 Figure 1 is a schematic representation of one form of system incorporating the present invention;
Figures 2a and 2b ar~ schematic representations of one form of' fractionation cell showing one mode of operation in Figure 2a and another mode of operation in Figure 2b;
Figure 3 shows graphically the rate and concentration of alcohol in plasma, the temperature of ingoing and outgoing plasma-and the temperature of ingoing and outgoing alcohol, as measured in the practice of Example l;
Figure 4 shows graphically the rates of alcohol addition and concentration in plasma, temperature of ingoing and out-going plasma before and after alcohol addition and the tempera-ture of the ingoing and outgoing heat exchanging alcohol measured in the practice o F,xample 2; and Figure S shows graphically tlle conditions measured in the practice of ~xample~ 3.
DE'rAILED D~SCI~IPTiON OF l'flE PltEFERRED EMBODIMENT:
Referring to the drawings, and particularly to Figure 1, there is shown one exemplary system for carrying out the frac-tionation method o this invention. The system includes aninsulated plasma vessel 10, an insulated alcohol vessel 11, and a buffer vessel 12, each equipped with a stirrer 13, heat ~ exchange means 14 and temperature indicating means lS. In ;:~ addition, tank 12 is provided with conductivity sensor 16 and ~ pH sensor 17, A pair of alcohol and plasma mixing and heat : ~: exchange fractionation vessels 18 are provided, arranged in parallel, As best seen schematically in Figure 2, each vessel or cell 18 includes a plurality of porous membrane fibers 19, all communi-cating at one end witll an entry chamber 20 and communicating atthe other end with a discharge chamber 21. Each vessèl is pro-116~3~8 vided ~iith an inlet port 22 and ~ischarge port 23, each providedwith a temperature sensor 15. All of the fibers l9 extend through and are surrounded by a central chamber 24 having an inlet port 25 and discharge port 26, each also having a tempera-ture sensor 15. Figure 2a shows the cell in one mode of opera-tion with ethanol (indicated by shading) passing through fibers 19 and plasma passing through the space surrounding the fibers.
Iigure 2b shows the opposite mode 2 in which the plasma passes through the fibers and the alcohol, again indicated by shading, passes through the surrounding space.
Plasma vessel 10 is connected to the mixing cells 18 by means of pipe or tube forming line 27 extending from vessel 10 to the entry ports Z2 of the mixing cells. Line 27 includes a controllcd metering pump 28 to circulate the plasma. In addition, a temperature sensor ~S, pll seJIsor 17, alcohol concentration scnsor 29, protein scnsor 30, 10w gauge 31, pressure control 33, and pressure gauge 34 arc all provided to detect and control conditions of liquid flowing through the line. Three-way valves 32 in the line enable thc flow to be directed as desired. The discharge ports 22 of the mixing cells 18 are connected back to plasma vessel 10 through pipes or tubes forming line 35, these lines having temperature sensors l5, pH sensors 17, alcohol sensors 29, flow gauges 31? pressure valves 36, and pressure gauge 34 to permit detection and control of flow conditions.
Alcohol vessel 11 is connected through line 37 and pump 28 to the entry port 25 to the mixing cell 18 and through line 38 to the discharge port 26 of that cell. Line 37 includes flow gauge 31, pressure valve 36, and pressure control 33. Buffer vessel 12 is similarly connected through lines 39 and 40 to the other of the mixin~ cells.
I`he process involves preferably the use of hollow fibers ~161368 e /h~
A of either Teflon~)divinyl acrylonitrile, polysulfone or cellu-lose acetate constitution. Alternate systems can include the utilization of spiral wound tubular membranes, plate and frame systems utilizing mcml)rarlcs o~ similar composition ancl pore size.
The preferred hollow fiber membranes are manufactured as single, coherent structures that can tolerate pressures on either side of the active membrane surface without rupture or damage to a support structure. Ilollow fibers are made by "spinning" a polymeric solution which upon setting or solidify-ing forms the anisotropic membrane in a tubular configuration.These hollow fibers consist of a very thin smooth membrane surface on the inside of lumen of the fiber with a controlled pore density and molecular rejection coefficient, and a rough spongy mcmbrane structure on the outside o~ the fiber with a looser or more opcn por¢ configuration.
' This membrane structure lends itself to pressurizing both the inside and the outside of the hollow fiber, Operating the hollow fiber module in an ultrafiltration mode is accomplished by pressurizing the inside of the hollow fiber. Operating the hollow fiber module in a "backflush" mode is accomplished by pressurizing the outside or shell side of the hollow fiber. The capability of backflushing hollow fibers is important to the use of hollow fibers for the addition oE cold ethanol in the Cohn process.
Although ethancl is a preferred protein precipitant, a number of alternative solvents and solutions may he used. These include methanol, butanol, acctonc, mcmbers of the glycol series, poly-ethylene glycol, dioxanc, etc , néutral salts such as phosphate, sulfates, etc., or a mixture of reactants such as alcohol and salts, and inally block copolymersof cthylene oxide and poly-isopropylene. Although described in terms of fractionation of 3~8 blood plasma, the inventioll is applicable to the fractionation of protein solutions from whatever source, animal (human, bovine, porcine, etc.), bacterial, or plant.
Plasma at between 4 C and -7 C is pumped through the lumen of the hollow fibers. This temperature can be sensitively and accurately controlle~ by ~umping 50 per cent to 80 per cent ethanol at temperatures between -20 C to -10 C on the outside of the hollow fibers. The cold ethanol is added to the recircu-lating plasma inside the fibers simply by increasing the flow rate and adjusting the back pressure on the outside of the hollow fibers sufficient to- cause the alcohol to penetrate the pores of the hollow fiber. By contacting the cold ethanol with the plasma in this mann~r, local heat ~uildup is minimal due to the velocity (2-10 ft/sec) of the plasma through the hollow fiber and the thin ~monomolecular~ layer of alcohol continuously contacting the plasma at the membrane surface. This high surface area contact created by the hollow fiber membranes causes very small particles to precipitate promoting sharp fractions, high yield, and minimal contact time for complete precipitation.
This critical temperature control of plasma at the alcohol plasma interface is achieved by adjustments of the following easily controlled factors. The temperature and flow rate of the alcohol act in concert to effect efficient and instantaneous heat exchange and act synergistically with the flow rate and temperature of the plasma. Therefore, heat is effectively dis-sipated, localized concentrations of alcohol are minimized, and the plasma protein concentration is kept at concentrations favor-ing optimal formation of optimally sized protein precipitates, The addition of buffers can be carried out in similar "backflushing" manner allowing for rapid and smooth equilibration to various pll levels, Na+ concentration, and ionic strength.

~61368 The addition of cold alcohol and buffer solutions can be acc:urately controlled by the pressure differential between the inside (lumen) of the hollow ~ibers and outside (shell) of the hollow fibers. Ihe high velocity of the plasma flowing through the lumen of the fibers promotes the most efficient contact of alcohol with plasma and acts as an agglomcrator to create ~ine particles of precipitating plasma while minimizing any heat of reaction caused by thc mixing o~ plasma and cold alcohol.
The formation of optimal particle size during precipitation, temperature control, p}l control, and agglomerating effect of the hollow fibers are synergistic in effect to reduce the total process time to make Cohn ~raction V (human albumin) from 5 to 7 days to less than 3 days Another advantag~ of the proposed use of hollow fibers to add protein precipitants such as cold ethanol to plasma is the elimination of the need for open batch tanks, thereby reducing the risk of pyrogenic lots of plasma proteins. This process can be carried out in closed tanks since the pl-l, temperature, and alcohol addition are controlled within the hollow fiber module. Special mixing devices and alcohol addition devices in the tanks are also eliminated since mixing and alcohol addi-tion is carried out in the hollow fiber Since the temperature, particle size, and pll can be finely controlled by the use of hollow fibers, the nee~ ~or vcry expensive low shear pumps is eliminated and more mo~erately priced low shear modified centri-fugal pumps can be employed.
It has heen suggested by some experts that shear caused by pumping and flow through narrow channel (hollow fibers) tubes causes considerable denaturation. (Charm, S.F.: Shear Inactiva-tion in Processing Biologic Material, DIIEW P~blication No.NIH 78-1422, pg. 27). Ilowever, Romicon hollow fibers have been ~161368 used to process enzymes on a commercial scale with minimal loss of enzyme activity. These commercial systems also employ centri-~ugal type pumps, '' Other experts (Dunhill, P.: The Action of Shear on Enzymic Proteins, DI~EW Publication No. NIH 78-1422, pg. 40) have proposed that shcar itsclf is llOt ~CtrinlCnta1 t~oes TlOt cause denaturation) to enzymes and labile protelns, but that exposure to oxygen ~air) and/or elevated temperatures are the denaturing conditions causing loss of enzyme activity an~ protein integrity.
This hypothesis appears to be supported by the work done with Romicon hollow fibers processing both enzymes and plasma proteins. 'It is possible that what was interpreted as shear denaturing of enzymes and proteins was actually denaturation cau~ed by elevated temperatures and exposure to air, particu-larly foaming, This ability to control temperature with the cold ethanol on either the outsides or insides of the hollow fibers similar to a tubular heat exchanger over an area o lS ft2 or 26.5 ft2 per module almost eliminates localized heating, while the elimina-tion o~ open tanks and mixcrs in 'the batch tanks prevents foam-ing and excessive ex~osure to air. This enables the use of cen-' trifugal pumps in conjunction with hollow fib,ers for plasmapumping while controlling the cold alcohol addition by ùse of a refrigerated pressurized tank or a refrigerated tank with centri-fugal pump. Denaturation or protein aggregation will be mini-mized under these conditions as well as minimizing the risk of pyrogen contaminatioll.
Alternatively the alcohol may be added to the plasma by reversing the compartments, i.e., alcohol flowing through the ' inside of the hollow ~iber with plasma circulating on the out-side. See Figure 2. In mode 1, Figure 2a, the alcohol is circulated inside the hollow fibcr. In mode 2, Figure 2b, thealcohol is circulated outside the hollow fiber.
Pinally, the capacity to control protein concentration elim-inates the need for largc storage vessels used in current prac-tice to accommodate up to 4 fold increase in volume following alcohol addition.
The invention is further illustrated by the following:
F.xample 1 Processing Data - Cohn Run ~12 - Mode 1, Figure 2a -Cryo-Precipitated Plasma Separation of Fraction I
The pilot plant used for these examples is illustrated in Pigure 1. The alcohol an~ plasma compartments in cells 18 are shown in Figures 2a and 2b. The mixing of alcohol and plasma and the hcat exchangc occurs also in this interace. Starting cryo-precipitated plasma (1.6 liters, 12.2 mS at 24 C, contain-ing 93 grams total protein: 54.4 grams albumin and 13.5 grams IgG) is cooled to +1 C without permitting ice to form. Three milliliters of sodium acetate buffer (4.0 M flAc + 0.2 M sodium acetate, p~l 4.0) is added to adjust the pH of the plasma from 7.86 to 7.1+0.1 (all pH determinations are 1:5 dilutions of sample to 0.9 per cent saline). Alcohol (80 per cent) for addition is stirred in a closed tank at a temperature o -11 C.
The plasma is stirred in a closed tank and temperature adjusted to comply with desired conditions. The protein is recirculated through the outside of the hollow fiber at a rate of 2500 ml/min.
At the same time, thc cold ethanol is recirculated through the inside of the Çibers at a rate of 300 ml/min. To obtain a final ethanol concentration in the plasma of 8 per cent, 178 ml 30 of 80 per cent ethanol werc addcd to the plasma by increasing the rate of the ethanol flow until 20 psi of back pressure is 1~61368 observed. Under these conditions the addition of alcohol wascompleted in 70 minutes. During this time-the plasma ethanol concentration went from 0 per cent to 8 per cent, the plasma tcmperature droppecl from 0 C to -2.5 C and the ethanol tempera-ture climbed from -11 C to -8.5 C. The rate and concentration of alcohol in plasma, the temperature of ingoing plasma (prior to mixing with alcohol), the temperature of outgoing plasma (after alcohol addition), and the temperature of ingoing and outgoing alcohol are illustrate~ in Figure 3.
Precipitate I was removed by centrifugation for 15 minutes, 27000 x G. Precipita'te I was resuspended wlth ice to 420 ml.
This precipitate is mainly fibrinogen. It contains 1.9 grams total protein (0.6 gram albumin and 0.4 gram IgG). Supernatant I contains 87 grams to,tal protein ~51 grams albumin, 12.1 grams IgG~ in a volume of 1550 ml. The conductivity of Supernatant I
is 8.7 mS at 24 C. The protein cooling tank and the hollow fiber outside are cleaned with distilled water and rinsed with 20 per cent ethanol.
Separation of Fraction II ~ III
! 20 Supernatant I is placed into the plasma cooling tank and maintained at -2.5 to -3 C. The pH is adjusted to 6.8+0.05 , by the addition of~3.1 ml of sodium acetate buffer. The super-natant is brought from 8 per cent ethanol to ~0 pér cent ethanol by the addition of 310 ml of 80 per cent ethanol using the same flow rate and pressure conditions as outlined in the "Separation of Fraction I". Figure 3 shows that during this 30 minute addi-tion, the supernatant I tempcraturc is lowered to -6 C as thc ethanol temperature climbs from -13 C to -6 C.
Precipitate II + III is removed by centrifugation for 15 minutes, -6 C, 27000 x G. Precipitate II ~ III is resuspended with ice to a volume of 1.16 liters. It contains 19.7 grams of total protein (1.9 ~rams albumin and 10.5 ~rams IgG). Super-natant II + III contains 60.5 grams total protein (43.7 grams albumin and 1.4 grams IgG) in a volume of 1.7 liters. Its con-ductivity is 4,8 mS at 24 C. The protein cooling tank and hollow fiber outside are rinsed with distilled water and 40 per cent ethanol. The 80 per cent ethanol is allowed to drop to -15 C in preparation for the separation of Fraction IVl.
Separation of Fraction IVl Supernatant II + III is placed back into the plasma cooling tank and maintained at -6 C. The pH is adjusted to 5.4+0.1 with the addition of 13 ml of sodium acetate buffer. The super-natant is diluted to 18 per cent ethanol with the addition of 190 grams of distilled crushed ice which is allowed to mix until the ice is dissolved, 'This step takes 25 minutes.
Precipitate IVl is remaved by centrifugation for lS minutes, -6 C, 270()0 x G. It is rcsuspended with ice to a final volume o 830 ml. Precipitate IVl contains 6,6 grams of total protein ~1.6 grams albumin and 0.9 gram IgG). Supernatant IVl contains 50.4 grams of total protein ~39.4 grams albumin and 0,2 gram IgG) in a volume of 1.8 liters. Its conductivity is 5.3 mS
at 24 C.
The protein cooling tank is cleaned with distilled water.
The 40 per cent ethanol is evacuated from the outside of the hollow fiber.
Separation of l'raction IV4 _ . _ Supernatant IVl is placed into the cooling tank and main-tained at -6 C. The pll is adjustcd to 5.8~0.05 by the addition of 20 ml of 1.0 M NallC03. l'he supernatant is brought from 18 per cent ethanol to 40 per ccnt ethanol by the addition of 998 ml of 80 per cent ethanol. 'The same initial flow rates are used as in separation of ~raction I, but during the addition . ~_, ~ =. . = - .... . . .... .. . .

116~3~8 the ethanol flow is increased until 25 psi of bacX pressure is observed. Figure 3 shows that during this 80 minute addition, the plasma temperature is maintained between -6 C and -9 C
as the ethanol temperature is kept below -7 C.
Precipitate IV4 is removed by centrifugation for 15 minutes, -6 C, 27000 x G and is resuspended in ice to obtain a final volume of 730 ml. It contains 6.6 grams of total protein (1.6 grams albumin and 0.9 gram IgG). Supernatant IV4 contains 40,2 grams of total protein ~35.8 grams albumin and 0.0 gram IgG) in a volume of 2.68 liters. Its conductivity is 2.5 mS
at 24 C.
The protein cooling tank is cleaned with distilled water and is separated from the hollow fiber. The hollow fiber and ethanol cooling tank are cleaned at this time.
Separation of l~'ration V
Supernatant IV4 is placed into the cooling tank and maintained between -7 C and -11 C. The pll is adjusted to 4.8+0.05 by the addition of 38 ml of sodium acetate buffer. The plasma becomes milky white. This step takes 60-70 minutes and the plasma temperature is below -~ C throughout.
Precipitate V is rcmoved by centrifugation for 15 minutes, at -6 C, 27000 x G and is resuspended to a volume of 3.0 liters - with ice. This precipitate is mainly albumin. It contains 37.5 grams of total protein by Biuret determination (37.7 grams albumin by immunochemical assay and 0.0 gram IgG). Supernatant V contains 2.6 grams total protein (0.1 gram albumin and 0,0 gram IgG) in a volume of 2.18 liters. Its conductivity is 3.0 mS at 24 C.
The essential features such as total protein yields of IgG
and albumin are summarized in 'I'able I.

116~368 ~, o ~_ # I o ~o o ~ C~ o ~_ o o :~ ~1 o a~ ~ o~ ~ ~ ~D O O O O
.~
~.
o ~1 o o . ~
o ~ o o o o o o o ~ ,, ~

~U ~ ~ ~ ~ ~r a~ co d~ 0 ~1 ~ O ~ t`~ t`J Iti ~ O a~
~ o a~
~ æ ~
.r o ~ o coo !D ,L ~ ~ ~ O ~ ~ a~ ~ o H ~ ~ . ~Y u ~ In ~ , O
~ ,~
,~ *
~1 d~ O ~ 1 o ,1 1~ ~
~ ,C
`~3''~ O
~_ ~ ~ co o cou~ r ~ ~ o u~
. , O ~ ~ r~ ~ o ~n o ~ o a~ ~ ~ ,~
~ a~

,::: ' _ ' O ln 1 0 ~ g .~
~ ~o Q ~ H b ~4 ~ H H
.~ ~ J ~ J
v ~ ~ ~ 3 ~ ~ 3 ~ 3 ~ ~ ~ o ~
O C: rJ rJ r~ -1 rl .,~ r; .,~ rf ,~ b _I
" ~ ~: R. S~
O ~ 'U .~J 'U C~
~ r~ 3 C~ #

li613~8 l.xamp1e 2 Processing Data - Cohn ~un ~14 - Mode l and Mode 2 -Stabilizcd ~luman Plasma Stabilized human ~lasma was prepared by the method outlined in U. S. Patents No. 3,998,946 and No. 4,136,094 (Condie).
Stabilization o plasma results in the removal of fibrinogen, the clotting fuctors, tllc colllpl~ cllt syst~l~, the kinillogclls, the lipoproteins, and the proteolytic enzyme plasmin plasminogen.
Since it is free of fibrinogen, se~aration of Fraction I is not required. For the example given, the protein is recirculated through the outside of the hollow fiber and is referred to as Mode #1 (~ig. 2a). When thc addition o alcohol is accomplished with the protein on the inside of the hollow fiber, it is referred to as Mode #2 (Fig. 2b). Rates of alcohol addition and concentra-tion in plasma, temperature of ingoing and outgoing plasma before and after alcohol a~dition, and the temperature of the heat exchanging alcohol (ingoing and outgoing~ are presented graphically in Figure 4.
Separation of Fraction II + III
Starting stabilized plasma (1.8 liters, 11.4 mS at 24 C, containing 120.1 grams total protein: 89.1 grams albumin and 11.4 grams IgG) was cooled to ~1 C without permitting ice to form. Four milliliters of sodium acetate buffer (4.0 M HAc +
0.2 M sodium acetate, pH 4.0) was added to adjust the pH from 7.53 to 7.1+0.1 (all pH determinations are 1:5 dilutions of sample to 0.9 per cent saline). Alcohol (80 per cent) for addi-tion is stirred in a closed tank at a temperature of -11.5 C, The plasma is stirred in a closed tank and temperature adjusted to desired conditions. Ihe protein is recirculated through the outside of the hollow fiber (Mode ~1) at a rate of 2500 ml/min.
At the same ~ime, thc cold cthanol is rccirculated through tho inside of the fibers at a rate of 300 ml/min.
To obtain a final ethanol concentration in the plasma of 8 per cent, 178 ml of 80 per cent ethanol was added to the plasma by increasing the rate of the ethanol flow until 25 psi of back pressure is observed. Under these conditions the addition o~
alcohol was completed in 25 minutes. During this time the plasma ethanol concentration went from 0 per cent to 20 per cent, the plasma temperature dropped from 0 C to -S.0 C and the ethanol temperature climbed from -11.5 C to -5.5 C ~see Figure 4-II + III).
When Mode #2 is used, the protein is recirculated through the inside of the hollow fiber at a rate of 400 ml/min. As the ethanol is recirculated through the outside of the hollow fiber, the back pressure due only to the plasma is 12 psi. A screw clamp is applied to the ethanol out line until the total back pressure reaches 20 psi. This addition takes 70 minutes. The ethanol temperature is kept b¢low -5 C and the protein tempera-ture is kept below -3 C throughout the addition.
Precipitate II + III was removed by centrifugation for 15 minutes, at -5 C, 27000 x C Precipitate II + III was resus-pended with ice to 1270 ml. This precipitate is mainly IgG.
Tt contains 13.2 grams total protein (3.4 grams albumin and 8.4 grams IgG). Supernatant II + III contains 98.1 grams total protein (75.9 grams albumin, 2.5 grams IgG) in a volume of 2160 ml. The conductivity o Supernatant II ~ III is 5.0 mS at 24 C.
The protein cooling tank an~ the hollow fiber outside are cleaned with distilled water and rinsed with 40 per cent ethanol.
ration of ~raction IVl Supernatant II ~ III is placed back into the cooling tank and maintained at -3.5 C. Ihe pH is adjusted to 5.4~0,1 with the addition of 12 ml of sodium acetate buffer. The supernatant - `

is diluted to 18 per cent ethanol with the addition of 240 grams of distilled crushed ice which is allowed to mix until the ice is dissolved. This step takes 25 minutes. The addition of ice causes the plasma temperature to drop to -8 C (see Figure 4-IVl).
Precipitate IVl is removed by centrifugation for lS minutes, -6 C, 27000 x G. It is resuspended with ice to a final ~olume of 960 ml. Precipitate IVl contains 5.1 grams of total protein (2.0 grams albumin and 1.4 grams IgG). Supernatant IVl contains 86.3 grams of total protein ~77.1 grams albumin and l.0 gram IgG) in a volume of 2.34 liters. Its conductivity is 5.4 mS at 24 C.
The protein cooling tank is cleaned with distilled water.
The 40 per cent ethanol is evacuate~ from the outside of the hollow fiber.
Separation of Fraction IV4 Supernatant IVl is placed into the cooling tank and main-tained at -6 C. The pH is adjusted to S,8~0.05 by the addi-tion of 17 ml of 1.0 M NaHCO3. The supernatant is brought from 18 per cent ethanol to 40 per cent ethanol by the addition of 1294 ml of 80 per cent ethanol. The same flow rates and back pressure are used as in separation of Fraction II + III Dur-- ing this 55 minute addition, the plasma temperature is maintained -between -6 C-and -8.5 C as~ the ethanol temperature is kept below -9 C (see Figure 4-IV4).
When Mode #2-is used, the conditions are the same as in Fraction II + III Mode #l except that the back pressure is allowed to reach 25 psi. The ethanol temperature is kept below -8 C
and the protein temperature is kept below -5 C throughout the addition. The addition was adjusted to completion in 250 minutes.
Precipitate IV4 is removed by centrifugation for 15 minutes, -6 C, 27000 x G and is resuspended in ice to obtain a final volume of 1100 ml. It contains 19.9 grams of total protein -~0 -11~1368 (13.0 gTams albumin and 0.9 gram IgG). Supernatant IV4 contains 64.6 grams of total protein (56.4 grams albumin and 0.0 gram IgG) in a volume of 3.40 liters. Its conductivity is 2.4 mS
at 24 C.
The protein cooling tank is clcane~l with distilled water and is separated from the hollow fiber. The hollow fiber and ethanol cooling tank can be cleaned at this time.
Separation of Fraction V
Supernatant IV4 is placed into the cooling tank and main-tained between -6.5 C and -7.5 C. The pH is adjusted to 4.8+0.05 by the addition of 33 ml of sodium acetate buffer. The plasma becomés milky white. 'lhis step takes 55 minutes and the plasma temperature is below -6 C throughout (see Figure 4-V).
Precipitate V is removed by contrifugation for 15 minutes, at -6 C, 27000 x G and is resuspended to a volume of 2.8 liters with ice. This precipitate is mainly albumin. It contains 57.4 grams of total protein according to Biuret assay ~58.7 grams albumin by immunochemical assay and 0,0 gram IgG). Super-natant V contains 3.8 ~rams total protein (0.8 gram albumin and 0.0 gram IgG) in a volume of 2.56 liters. Its conductivity is

2.8 mS at 24 C.
The essential features concerning IgG and albumin yield for this run are presented in Table II.
Table III summarizes the IgG and albumin yields of ~ive runs of cryopoor plasma and four runs of stabilized human plasma.
The albumin yields averaged 69.5+3,1 per cent for the stabilized human plasma runs and 69.8+1.8 per cent for cryo-precipitated human plasma runs. The albumin isolated during these runs contained less than 2 per cent contamination by other plasma proteins.
The IgG Precipitate II + III contained appreciable concen-~6~368 ., N r ~n ~ N i (~

oo~ co ~ o a~ o o

3~ ~ r~t ~ a~ t~ O r r t~ ~1 ~ In ~r o ~;r o a~ o o I o ~ c~
O I O ~ ~D N ~ er N r~i O

t~ o ~ o a~
tJ~-3 ~P a~ j t~ ~r ~ ~ , u ,~ -t~
,~

_. h ~ ~ ~ e ~ .

~ ~ , 2 ~- 1161368 ~ ~1 ~' .~ ~il ,~' ~ ~ , ~ ~ o r~ ~

~ o ~

IB~

H ~1 J L

trations of albumin with the yields of IgG being 76.6+7.5 per cent for stabilized human plasma and 73.9+4.3 per cent for cryo-precipitate~ human plasma. Reprocessing Precipitate II + III
in this system following Cohn method 6 gave an essentially pure IgG with an overall yield based on the IgG present in starting plasma of 64.7 per cent.
* * *
Plasma Temperature Control During Ethanol Addition There are major points to be made regarding the precise nature of temperature control that is routinely achieved in this system when adding and mixing concentrated ethanol to plasma, First, alcohol may be added through hollow fibers at controlled rates which can be very slow or quite rapid. Figures 3 and 4 illustrate that this addition can be controlled so that ethanol addition is at a constant, steady rate. Second, that regardless of the rate of addition the temperature of the plasma during ethanol addition can be rigorously controlléd so that there is on the average 0.4 C differential between the temperature of plasma immediately before ethanol addition and immediately after mixing in the hollow fiber. This fine control of temperature is achieved and controlled by: ~a) the alcohol temperature and alcohol flow rate through the hollow fiber, and (b) plasma temperature and plasma flow rate through the hollow fiber. As is clearly demonstrated in Figures 3 and 4, the alcohol serves as the major source of heat exchange and this control is enhallced by the design and construction of the hollow fiber system when utilized according to this invention.
Time Required to ~ractionate Plasma According to Cohn Method 9 Another significant feature of the process is the reduction in time required to fractionate human plasma completely to albumin. As can be seen in ligure 3, the time required for 3~8 each step is short with a total elapsed time being 10-1/2 hours, albumin yield averages 70 per cent with less than 2 per cent non-albumin contaminants.
Whcn stabilizcd plasma i~ use~l (Examp]e 2 and Figure 4), the total time is reduce~ to 8 hours an~ results in an essen-tially pure albumin witll 70 per cent yields.
Example 3 ~iafiltration Diafiltration may be used to replace centrifugation and solvent removal. The following is an example of its use to process albumin eliminating centrifugation and solvent removal by other denaturing methods. After the addition of ethanol to 40 per cent, diafiltration may be used to concentrate the protein and to remove the ethanol. The process is identical to Examples 1 or 2 until ater the ad~lition o ethanol to a plasma concentration of 40 per ccnt.
The essential features of plasma temperature, plasma volume, protein concentration, and plasma ethanol concentration are presented in Figure S. Prior to centrifugation, the plasma containing 40 per cent ethanol is concentrated from 3.45 liters to 1.88 liters. This process was adjusted for completion in 250 minutes. The protein temperature is kept below -3 C. The protein concentration of thc plasma is increased from 26 mg/ml to 48 mg/ml. The plasma is pumped through the inside of the hollow fiber at a rate of 3000 ml/min which gives a back pressure reading of 25 psi. Filtrate removal proceeds at an average rate of 390 ml/hr at the tempcrature described.
The concentrate-l plasma is now centrifuged for 15 minutes, at -6 C, 27000 x G. The composition of Precipitate IV4 and Supernatant IV4 is the samc as described for Examples 1 and 2.
The volume of Supernatant IV4 however has been reduced to 1.70 3~8 liters at a protein concentration of 42.5 mg/ml.
Supernatant IV4 is placed into the cooling tank and kept at -6 C. Instcad of re~ucing the p~l to 4.8 followed by centri-fugation, the plasma is simply brought to 0 per cent ethanol by diafiltration with ice and cold water.
Begin recirculating the plasma through the inside of the hollow fiber until 25 psi of back pressure is observed. The flow rate of plasma will be 3000 ml/min. Add 1200 grams of distilled ice to the plasma to reducc the cthanal concentration from 40 per cent to 23 per cent. This addition increases the plasma volume to 2.95 liters. As the ice dissolves, concentra-tion proceeds until the plasma volume is reduced to 1.95 liters.
An additional 120 grams of distilled ice are added to the plasma to reduce the ethanol concentration from 23 per cent to 10 per cent, This a~dition increases the plasma volume to 2,95 liters.
As the ice dissolves, concentration proceeds until the plasma volume is reduced to about 2,50 liters. The filtrate flow rate averages 900 ml/hr during the ice addition which is accomplished in 120 minutes.
Since the ethanol concentration in the p~asma is reduced to 10 per cent, it is a~vantageous to allow the plasma tempera-ture to increase above 0 C. The concentration rate of the plasma is highly temperature dependent and the rate of filtrate out increases from 900 ml/hr to 5800 ml/hr as the plasma temperature increases to +14 C. The reduction of the ethanol concentra-tion from 10 per cent to 0 per cent is accomplished by adding 3.5 liters of cold distilled water to the plasma over a two hour period. Since the rate of concentration is greater than thc rate of water addition, the plaslna volume decreases to a final volume of 1.6 liters and a final protein concentration of 45,5 mg/ml. 'rhe resulting albumin is free of alcohol and IgG, is 3~8 in a conccntratcd statc, and was ncver allowed to precipitate.
Although the separation of albumin by diafiltration is illus-trated, the same method may be used to remove and separate alcohol and othcr solvcllts from l~rotein ~ractions from any step of the overall systcm.
It is apparent that many modifications and variations of this invention as hereinbefore set forth may be made without departing from the spirit and scope thereof. The specific embodiments described are given by way of example only and the invention is limited only by the terms of the appended claims,

Claims (10)

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1, A controlled rapid protein fractionation method which comprises:
A) circulating a cold protein solution through a confined space defined by one side of a porous membrane, B) controlling temperature by circulating a liquid protein precipitant solution in contact with the opposite side of said membrane, C) increasing the flow rate and pressure of the pre-cipitant solution on the opposite side of said membrane to cause the protein precipitant solution to penetrate the membrane, mix with the protein solution, and precipitate fine particles of protein from the solution along the first side of the membrane wall, and D) separating the precipitated protein particles.

2. A method according to Claim 1 wherein the temperature of the protein solution is controlled by circulating a cold protein precipitant solution.

3. A method according to Claim 2 wherein:
A) the resulting supernatant is recooled, B) the pH, Na+ concentration and ionic strength of the supernatant are adjusted and the supernatant is recirculated through a confined space defined by one side of a porous membrane, C) the temperature is controlled by pumping a cold liquid protein precipitant solution in contact with the opposite side of said membrane, D) the flow rate and pressure of the precipitant solution on the opposite side of the membrane is increased to cause the protein precipitant solution to penetrate the membrane, mix with the supernatant and precipitate further fine particles of protein along the first side of the membrane wall, and E) the further precipitated protein particles are separated.

4. A method according to Claim 3 wherein:
A) the pH, Na+ concentration and ionic strength of successive supernatants are further adjusted and the supernatant is circulated in contact with one side of the membrane, B) further cold protein precipitant solution is circulated in contact with the opposite side of the membrane to control the temperature of the supernatant, C) the flow rate and pressure of the precipitant solution is increased to cause that solution to penetrate the membrane wall and precipitate further fine protein particles along the first side of the membrane wall, and D) the precipitated protein particles are separated successively until all desired fractions have been recovered.

5. A method according to Claim 3 wherein the protein containing solution is acidified by:
A) circulating an acidic buffer in contact with the opposite side of the membrane, and B) increasing the flow rate and pressure of the buffer to cause the buffer to penetrate the membrane.

6. A method according to Claim 5 wherein:
A) the membrane is in the form of a plurality of small diameter hollow fibers, B) one of said solutions is circulated through the fibers, and C) the other of said solutions is circulated around the outside of the fibers.

7. A method according to Claim 1 wherein the precipitated protein particles are separated by centrifugation.

8, A method according to Claim 1 wherein the precipitated protein particles are concentrated and solvent is removed by diafiltration.

9. A method according to Claim 8 wherein:
A) a solution containing precipitated protein particles is circulated through a confined space defined by one side of a porous membrane, B) another liquid is circulated on the opposite side of the membrane wall, and C) the flow rate and pressure are increased on said solution to cause solvent therein to penetrate the membrane wall and concentrate the protein particles.

10. A method according to Claim 1 wherein:
A) said protein solution is blood plasma at an initial temperature between about 4° C and -7° C, and B) said protein precipitant is ethanol in initial concentration between about 50 per cent to 80 per cent at an initial temperature between about -20°C to -10°C.