WO2000018882A1 - Methods and apparatus for lipid membrane disruption - Google Patents

Abstract

Disclosed are methods and apparatus for disrupting lipid membranes in a composition that comprises the lipid membranes, the method including subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions.

Description

Methods and Apparatus for Lipid Membrane Disruption

BACKGROUND OF THE INVENTION

Reference to Related Applications

The invention claims priority under 35 U.S.C. § 119(e) to U.S. Provisional Patent Application No. 60/101 ,849, filed September 25, 1998.

Field of the Invention

The invention relates to improved methods and apparatus for lipid membrane disruption.

Description of Related Art

In bioprocessing or other types of chemical processing, there are many instances when a product is contained within or associated with a lipid membrane. In such cases, it may be desirable to disrupt the lipid membrane so as to recover the product, or aid in product recovery.

For example, if the lipid membrane is a cell, disrupting the lipid membrane of the cell can allow the product to be separated from the remainder of the cell. Such separations are common in recovering products produced in bacterial fermentations.

A number of different techniques have been developed to disrupt lipid membranes. For example, compositions comprising the lipid membranes may be dried, and the resulting powder may be crushed and ground. The methods used for crushing and grinding such powders may be similar to those used widely in industry and elsewhere but with the very important difference that the particle size, especially of bacteria, is very much smaller than that usually dealt with. This imposes several problems not least of which is that the machines must be made to high precision and that this must be maintained in the face of constant wear. One other important difficulty is that the phenomenon of caking associated with fine powders (10 to 1 micron diameter) is met with almost immediately and not in the later stages in the process.

Dry grinders and crushers include ball and vibratory mills, roller mills, plate mills, and powder attrition mills. The comminution process proceeds by the introduction of cracks at points of strain introduced by imposing either a tensile or shear stress in each lipid membrane. This is the disruption phenomenon even in beater systems such as hammer mills and ball mills where the initial stress is compressive.

Energy is expended producing stresses that result in cracks or breakages. This energy appears as heat, which must be taken into account when working with heat denaturable materials such as biological products. Since the degree of fineness is comparatively great, surfaces in ball or hammer mills such as the liners, balls, or abrasives, appear rough in comparison. These too suffer attrition often quite rapidly and it is common to find a high degree of metal contamination in these systems. This is particularly undesirable when the metals contain toxic elements such as nickel or copper, as do the stainless steels.

Many mills used for dry grinding and crushing can also be used for wet milling of lipid membranes. The common feature of such devices is that the breakage of the lipid membranes is probably produced directly, and predominantly, by the hydraulic or liquid shear rather than by an initial compressional stress as in dry milling. Unfortunately, the rate of attrition is slow, especially if no abrasive is added and the danger of subsequent contamination is high. In addition, changes in the rheological properties due to cell rupture may slow down attrition markedly. The release of DNA and of wetting or foaming polymers is particularly troublesome in vibration mills and it is a common practice to add some antifoam agent. Additionally, the grinding media in wet milling apparatus, such as glass beads, may wear, resulting in contamination of the final product by the worn off fine particles.

Shear in liquids may also be used for disruption of lipid membranes. Shear in liquids may be defined as a frictional force set up during flow. Liquid flows may be arranged so that turbulence is created. The onset of turbulence creates vortices near to which there may be significantly higher shear rates than are present in the bulk liquid. These shear rates may degrade the cell walls of various microorganisms. However, such liquid shearing processes suffer from undesirably low disruption efficiencies.

Other ways known for disintegration of microorganisms include the French, Hughes, and X presses. Liquid or hydrodynamic shear may be used to disrupt microorganism cell walls by using the French press. In its simplest form, the French press consists of a stationary cylinder having a small orifice and needle valve at its base, and a piston with some sort of pressure type washer. Pressure is applied to the piston by means of a laboratory hydraulic press capable of delivering 10-20 tons total pressure. The Hughes press consists of devices that force a frozen suspension or paste of cells through a small gap into a receiving chamber, at very high pressures (often 10-80,000 psi). The X press is designed such that the cell suspension may be passed through a circular hole several times to increase the breakage rate in an aseptic manner. However, these presses, while suitable for use in the laboratory, do not produce product at a high enough throughput to make them commercially valuable in large scale biopharmaceutical or chemical manufacturing processes. A semi-continuous press has been constructed for the disintegration of microorganisms and other biological material by freeze-pressing, i.e., pressure extrusion of frozen material through a narrow hole. The material to be freeze-pressed is frozen in the form of cylindrical rods, which fit into the pressure chamber and are extruded by a piston forced back and forth by means of a hydraulic pump. For example, at a sample temperature of - 35 degrees C and a press temperature of -20 degrees C, about 90% disruption may be achieved in a single passage of undiluted Bakers yeast, at a concentration of 270 milligrams per gram, through the orifice of the pressure chamber. With this press about 10 kilograms of material can be fresh freeze-pressed per hour. K. E. Magnusson, et al., "Large-Scale Disintegration of Micro-Organisms by Freeze-Pressing", Biotechnology and Bioengineering, 18:975-986 (1976). This document, and all other documents cited to herein, are incorporated by reference as if reproduced fully below. However, this mechanism may not be advantageously used in large scale bioprocessing because of its relatively low throughput.

High pressure homogenization is an alternative method for large- scale cell disruption. In high pressure homogenizers, disruption is accomplished by passing cell suspension, under high pressure through an adjustable, restricted orifice discharge valve. There are fewer operating parameters to consider using this apparatus, than with high speed ball mills. The major parameters are operating pressure, temperature and number of passes through the valve.

A drawback is that energy requirements, as well as capital investment, for these processes are high. In addition, since most of the energy input is converted to heat, effective cooling is required to prevent heat denaturation or inactivation of products. C.R. Anglar, Disruption of Microbial Cells, in Comprehensive Biotechnology: the Principles, Applications and Regulations of Biotechnology in Industry, Agriculture and Medicine 2 (Murray Moo-Young Ed.) (1985).

Osmosis is another method of cell or microorganism disruption. In this method, cell disruption occurs as a result of hydrostatic pressure exerted against the membrane and, as the membrane bursts, the force disappears. Thus, osmosis affords the gentlest of mechanical methods of cell disruption. Microorganisms are in osmotic equilibrium with their environment and can be disrupted by suddenly diluting the medium in which they are present. Osmosis by itself, however, suffers from undesirably low efficiencies of disruption.

Ultrasound, drawing and extraction, autolysis, enzymic attack on cell walls, use of bacteriophages, and other methods are described in D.E. Hughes et al., The Disintegration of Micro-Organisms in Methods of Microbiology (J. R. Norris et al., Ed.) (1971 Academic Press). These methods do not generally provide satisfactory results in product recovery, especially at industrial scales.

There is therefore a need for improved methods and apparatus for lipid membrane disruption that overcome the problems in the art as noted above.

SUMMARY OF THE INVENTION

In an aspect, the invention relates to a method of disrupting lipid membranes in a composition that comprises the lipid membranes, the method comprising subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions. BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a solid drum freezing apparatus according to the invention.

FIG. 2 shows a thin wall drum freezing apparatus according to the invention.

FIG. 3 shows a thin layer freezing apparatus according to the invention.

FIG. 4 shows a freeze granulator apparatus according to the invention.

FIG. 5 shows a freeze granulator and compression roller apparatus according to the invention.

FIG. 6 shows a compression roller apparatus according to the invention.

FIG. 7 shows a grooved roller apparatus according to the invention.

FIG. 8 shows a freeze granulator apparatus according to the invention.

FIGS. 9A-B show a multiple screw apparatus according to the invention.

FIG. 10 shows a auger apparatus according to the invention.

FIGS. 11-20 show details of auger devices according to the invention. DETAILED DESCRIPTION OF THE INVENTION

The inventor has surprisingly and unexpectedly discovered that methods of disrupting a lipid membrane comprising subjecting the lipid membrane to controlled freezing or thawing conditions can result in increased disruption of lipid membranes, at lower cost and with reduced contamination, as compared with prior art lipid membrane disruption methods. This results from the understanding that a number of different effects occur during freezing or thawing of the lipid membrane.

For example, by controlling the growth of ice crystals formed by any water present in or near lipid membranes, the inventor has discovered that it is possible to carefully manipulate how the ice crystals impact the lipid membranes. Controlled ice crystal growth, achieved for example by controlled freezing or thawing that is in turn achieved for example by manipulating heat flux through the composition at points of interest, thus may lead to selective disruption of lipid membranes, or an enhanced overall disruption of lipid membranes. Additionally, lipid membranes are much more fragile at cooler temperatures than at warmer temperatures. This fragility can be utilized to enhance disruption of the lipid membrane as compared to higher temperature processes.

This represents an advance as compared to previous methods of lipid membrane disruption that froze or thawed lipid membranes in an uncontrolled fashion and without regard for controlled ice crystal growth. Prior art methods often result in less efficient disruption than the present invention and consequently, lower yields of any biological products contained within or associated with the lipid membranes.

Additionally, an advantage of the inventive process is that it permits large-scale recovery of biological materials that might be otherwise denatured at higher temperatures. The practice of controlled freezing according to the invention promotes reduced denaturing as compared to higher temperature recovery processes.

These and other advantages of the invention now will be described in more detail below.

Continuing with a detailed description of the invention, controlled freezing or thawing of lipid membranes may take place in the recited compositions with varying amounts of surrounding water. For example, the lipid membranes in the recited compositions may be substantially dehydrated or they may be suspended in an aqueous media. The amount of surrounding water can be controlled to adjust the freezing or thawing conditions and to improve disruption, if desired.

When the recited compositions are substantially dehydrated, the primary effect of controlled freezing will be on the fragility of the lipid membrane, and the impact of ice crystallization on any water that remains associated with the lipid membrane. While lipid membrane fragility is significant for dehydrated compositions, at low/freezing temperatures generally, the fragility of the lipid membranes is increased and the flexibility is decreased. Lower temperatures are thus better than higher temperatures for disruption, especially if any internal stresses are developed in the frozen lipid membranes.

When the recited compositions are dehydrated, the recited composition also may freeze into a pseudo-glassy state. Under such circumstances, the glassy state that may form upon freezing may incorporate both the lipid membrane and materials contained inside or associated with the lipid membrane. When the composition is in the glassy state, lipid membrane shattering may result, particularly when the composition is subjected to disruption augmentation techniques. Lower temperatures tend to promote shattering of the lipid membrane, and avoid localized heating that may denature biological products. Additionally, in general the more dehydrated the composition in these preferable embodiments, the better the shattering performance may be. Note that the presence of intracellular or extracellular ice crystals may not be needed for lipid membrane disruption in such circumstances because the lipid membranes themselves are in a glassy state and are therefore brittle and subject to shattering.

Such dehydrated lipid membranes may be obtained as filter cakes or solid streams from centrifuges. Freezing of the dehydrated lipid membranes may be carried out using any of the freezing methods detailed herein. In a preferable embodiment, liquid nitrogen is used as a cryogen to obtain a glassy state composition at a relatively low temperature. Shattering of the dehydrated lipid membranes in the glassy state may be preferably carried out using any one of the mechanical shear disruption augmenting techniques discussed herein.

When aqueous media is present in the recited compositions, there are at least two basic types of frozen states: complete or substantially complete dendritic, and partial dendritic. In the first state, the composition comprising the lipid membranes is substantially solid. In the second state, the composition is semi-solid. Dendritic ice crystals are described further in R. Wisniewski, Developing Large-Scale Cryopreservation Systems for Biopharmaceutical Systems. BioPharm 11(6):50-56 (1998) and R. Wisniewski, Large Scale Cryopreservation of Cells. Cell Components, and Biological Solutions, BioPharm 11(9):42-61 (1998). Selecting between the states has implications for selectively disrupting the lipid membranes. For example, the completely dendritic state generally has a higher first pass disruption yield than the partial dendritic state. However, the partial dendritic state may be more amenable for further processing using disruption augmenting techniques.

A first step in understanding the difference between the complete and partial dendritic states is to consider dendritic growth characteristics. In dendritic growth of ice crystals, the dendritic ice crystals do not start growing immediately from the cold wall. Rather, the cold wall acts to cool the liquid volume to approximately 0 °C. When the temperature of the liquid approaches 0 °C, the wall may begin to develop a thin layer of ice up to only a few millimeters thick. This thin layer may turn into a solid front, which then develops surface instabilities and changes from a "flat" surface to a "wavy" surface. The peaks of the waves slowly may become dendritic needle-like crystals.

The development of these dendrites is dependent on several variables, including, but not limited to, front velocities, solute concentrations and solute characteristics, temperature gradients along the dendrites, convection currents in the liquid phase, undercooling of dendritic tips, and liquid phase undercooling. Solid-liquid front velocities are chiefly a function of heat flux at the front. Faster removal of heat promotes faster front growth that, in turn, impacts upon dendrite formation. The spacing between the dendrites depend on the front velocity - it is smaller at high velocities and larger at lower velocities.

For example, the initial spacing between the peaks might change, depending on the solid front movement velocity. If the velocity is slow, some of the peaks may disappear and others may grow into dendrites, thus coarsening the spacing between the peaks. At very low velocities (front velocities of less than about 3 mm/hr., i.e. slow freezing) the dendrites may grow in a "cellular crystal growth" fashion - as columns of ice touching each other and expelling most of the solutes into the remaining liquid phase. At higher front velocities (front velocities from about 80 to about 200 mm/hr, i.e. rapid freezing), the spacing between dendrites tends to be wider. At very rapid freezing (front velocities greater than about 200 mm/hr.), the crystals may form a network of very fine crystals with no directional (columnar/cellular or dendritic) pattern. Any solutes present may be partly rejected/concentrated in the unfrozen portion but may be also trapped between the dendrites, thus possibly forming very large ice crystal-solute interface areas.

In a preferable embodiment, the front velocity ranges from about three to about 250 mm/hr, more preferably from about five to about 120 mm/hr and most preferably from about eight to about 80 mm/hr. At velocities greater than or less than these ranges, there is an increased likelihood that the front will grow and behave almost like a flat interface. This may result in exclusion of the lipid membranes from the frozen portion of the composition and concentration of the lipid membranes in the remaining liquid phase. Within the preferable ranges, typical free interdendritic space in developed dendritic growth may be on the order of tens of micrometers or more for coarse dendrites. Generally speaking, lipid membranes of interest have a diameter of roughly 100 Angstroms to tens of micrometers. The inter-dendritic spacing can be correlated to the size of the lipid membranes such that the dendrites "squeeze" the lipid membranes. The inter-dendritic spacing can be regulated by increasing or decreasing the heat flux out of the system (thereby influencing thermal effects and the resulting front velocities), and by selection of solutes (discussed more below).

The length of free dendrites may depend in part on the front velocity and on the temperature gradient along the dendrites. The free dendrite may refer to the length of the dendrite sticking into the liquid phase, or, alternatively, the thickness of the "mushy zone", e.g. a mixture of needles and liquid phase between them. At the tips of the dendrites, the temperature is close to 0 oC, and decreases gradually to match the wall temperature along the dendrite length and the solidified mass away from the front. The temperature of liquid between the dendrites also decreases with nearness to the cold wall.

Dendritic growth also depends on the concentration level of solutes and the character of the solutes. The solutes may affect the spacing and the temperature gradient along the free dendrite. As the front progresses, the water is being frozen out of the solution and the dendrites become thicker. Simultaneously, the concentration of solutes between the dendrites in the liquid phase increases. At this point, the recited composition passes through one of the partial dendritic states. This state may be referred to as the "slushy" state in that the mixture of dendrites, aqueous media, lipid membranes, solutes, etc. make up a slushy mixture.

As cooling continues, with certain solutes such as salts, the solute concentration reaches a eutectic concentration and temperature. The solution between the dendrites then solidifies, reaching the complete or substantially complete, or solid, dendritic state.

For some molecular species, instead of reaching the eutectic point, as the solutes are increased in concentration, the solution in the interdendritic spaces become more and more viscous. As a result, water diffusion out of the interdendritic spaces towards the dendrites' surfaces slows. Eventually, the solution in the interdendritic spaces becomes "rubbery" and later solidifies into a "glassy" amorphous state at certain temperature ranges. At this point, water diffusion towards the dendritic crystals comes to a halt, trapping a certain amount of water molecules inside the glassy interdendritic space. This is another partial dendritic state, referred to as the glassy state.

For faster freezing, more water is trapped in the glassy state and the temperature range at which the "glass" exists is lower - liquid water may act as a plasticizer of sorts. Faster freezing thus puts the frozen material into a "nonequilibrium state" regarding the interdendritic solutes. When warmed, the solutes between the ice crystals become less viscous and the water can slowly migrate towards the dendrites.

Cycling the temperature up and down may remove part of the water from solution, with resulting dendrite growth and reduction in the size of the interdendritic spacing. Other variables that impact dendrite growth at this point include, but are not limited to interdendritic spacing, which impacts the diffusional migration path, and how high the temperature rises during cycling, which impacts the viscosity and the diffusion coefficient.

Partial dendritic freezing or thawing can also be used. In such a state, there still will be semi-liquid solutes between the dendrites. This can be achieved by a combination of temperature and solute composition control - e.g., temperature kept high enough to have the ice crystals present but solutes between them still liquefied, e.g., above the eutectic and glassy state temperatures.

In a partial dendπtic freezing or thawing state, disruption augmenting techniques may be used to enhance disruption of the lipid membrane. For example, mechanical shear generated by apparatus such as a piston, a hydraulically/pneumatically driven plate, etc. can be applied, crushing the lipid membranes between the ice crystals. Bags containing the lipid membrane and aqueous media could also be pressed, so long as the bag material was selected to stand up to the stress at low temperature. The partial dendritic approach is different from the "French press." In a French press, the completely frozen (solid) product is pressed through a small orifice under very high pressures. Such an approach is a low throughput approach. Instead, in the inventive partial dendritic approach, the aqueous media and the lipid membranes may be in the form of a

"slush" that is near 0 oC. The process of recrystallization also may create some disruption. Recrystallization is discussed at more detail in R. Wisniewski, Developing Large-Scale Cryopreservation Systems for Biopharmaceutical Systems, BioPharm 11(6):50-56 (1998); R. Wisniewski, Large Scale Cryopreservation of Cells, Cell Components, and Biological Solutions. BioPharm 11(9):42-61 (1998). Additionally, if the frozen lipid membranes are processed using disruption augmenting techniques, as discussed further below, the presence of the recrystallized ice can serve as a type of grinding media, which enhances the tendency of the frozen, fragile membranes to be disrupted. Uses of recrystallized ice generally as grinding media are additionally discussed in more detail below.

Recrystallization effects are most significant when ice crystals are small and form during rapid cooling under conditions of nonequilibrium. Such small crystals are thermodynamically metastable because of their high surface energies. They tend to have minimal surfaces and to form larger, more stable crystals. The rate of ice crystal growth and recrystallization is temperature dependent with the maximum recrystallization effect for aqueous solutions at about 260 K. On both sides of that maximum, the rate of recrystallization decreases: sharply at 260-273 K and slowly at 260-160 K. At about 140 K, ice crystals do not grow; recrystallization rates are significantly retarded if the temperature is kept below 180 K. The temperature of the composition may be cycled through the recrystallization range, leading to ice crystal growth that would further squeeze and disrupt the lipid membrane. It is possible to cycle back and forth through the recrystallization zone, achieving larger and larger ice crystals, thus enhancing disruption. The cycling may be symmetrical or asymmetrical in terms of the amount of time-temperature change held on both sides of the temperature level of recrystallization peak, depending upon the particular circumstances.

The recrystallization may also cause partial sintering of the remaining larger crystals. These recrystallization effects may develop mechanical forces and rearrangements acting upon the lipid membranes sufficient to disrupt frozen, fragile membranes. This method may be relatively gentle enough, for example, to disrupt an outer cell membrane without damaging the inner cell membrane. Mechanical compression may be used in another preferable embodiment to aid recrystallization. A combination of temperature, holding time and mechanical compression may enhance recrystallization effects resulting in internal stress and rearrangement of the frozen matrix.

The effects of recrystallization can be further enhanced through control of the initial portion of the freezing or thawing. For example, fast freezing of the volume may be accomplished by applying a cryogen in a controlled manner. For example, too slow a cooling rate, and the ice crystals might not be thermodynamically instable enough. Too fast a cooling rate (e.g. by direct plunging into liquid nitrogen) might produce ice crystals that are too small, resulting in rejection of the lipid membranes from the matrix. The effects of a overly fast cooling rates may be ameliorated to an extent by using an aqueous media with enhanced viscosity. In such a case, if the viscous aqueous media reduced rejection of the lipid membranes from the matrix, and the lipid membranes are trapped between the small crystals, disruption can be enhanced. Similar effects may be achieved by controlling the thawing of the aqueous media and lipid membrane.

Having controlled the freezing and thawing of the recited composition may also offer disruption advantages over uncontrolled freezing and/or thawing. Ice recrystallization due to thawing may be ineffective if the freezing of the composition took place at too slow a rate, i.e. was uncontrolled. Alternatively, uncontrolled thawing (for example, in a large sample thawing nonuniformly from its surface to its center or different sample areas thawing differently depending on their location within the composition) makes it very difficult to accomplish careful manipulation of the lipid membranes to achieve selective disruption, such as disrupting an outer cell membrane without disrupting the inner cell membrane.

Temperature cycling can provide controlled freezing and thawing conditions inside the sample. For example, controlled thawing may be conducted very slowly in large samples. This controlled thawing minimizes temperature gradients across the sample, and allows large samples to be held at a roughly uniform temperature, such as the recrystallization temperature, as needed. The heating portion of the cycle may be conducted volumetrically (by microwave radiation, for example), or at a very slow rate. Thawing may be accomplished by precisely controlled temperature profiles - the holding periods (temperature plateaus) may be added at levels where the recrystallization effects are of significance. These temperature plateaus may be combined with application of mechanical compression/decompression. The controlled freezing and thawing steps may be repeated to achieve desired disruption results. An additional advantage of this controlled freezing and thawing is that cell disruption may take place under conditions that are relatively unlikely to lead to biological product denaturation. This may be critical in recovery of certain biological products that are likely to denature under harsher or higher temperature denaturation processes.

The cycling could be also done during freezing (changing the cold wall temperature to remove more water from the interdendritic space due to waves of temperature change) and also to affect the growth of ice crystals. The periods and temperature amplitude need to be optimized for maximum desired effects. For example, changing the period an/or the temperature amplitude may impact the water content in the glassy state and the dendrite shape.

Initial lipid membrane concentrations may also impact freezing/thawing and ultimately disruption. If higher concentrations of lipid membranes are used, the lipid membranes may fill in more interdendritic space. This may mean that, under external forces, the ice crystals may dislocate easier because the presence of the lipid membranes may weaken the overall ice matrix structure. Under such conditions, there may be more "grinding" motion (longer processing time) of ice crystals needed, but the force to start the dislocation of ice crystals can be reduced. The matrix produced by rapid freezing may also be influenced by lipid membrane initial concentration. Low initial concentrations give thin layers between the ice crystals, while very high concentrations may give ice crystals frozen within a glassy mass of frozen cells.

As noted above, the selection of solutes plays a role in controlling the freezing or thawing according to the invention. Of interest are selection of the type and concentration of solutes, their eutectic or glassy state solidification temperature and water content, their molecular hydration, their molecular size (which impacts the diffusion coefficient). Also of interest are the molecular interactions of solutes with solutes and with the ice surface and with other surfaces in the composition.

This points to another impact of the initial concentration of solutes. The amount of water trapped in the interdendritic spaces may also depend on initial concentration of solutes. As more solutes are initially present, the thickness of the glassy states increases and the diffusional path for water through the interdendritic spaces towards the dendrites also increases. The result may be more water trapped after solidification. Additionally, the solutes may form a rubbery embedding mass around cells, particularly if the solutes are those that can bind water. Consequently, more temperature cycles may be needed to move the water towards dendrites with, quite likely, higher final water content. To get "harder glass" one may need to cool even more.

If the lipid membranes are in a solution that contains glass-forming solutes, sucrose, for example, then the cells will be embedded into solidified solutes. The initial concentrations of lipid membranes and solutes impact how much interdendritic volume the lipid membranes and solutes might occupy. Higher lipid membrane concentrations and lower solute concentration might tend to form mostly squeezed lipid membrane networks with lipid membrane to lipid membrane, and lipid membrane to dendrite ice contacts. R. Wisniewski, Large Scale Cryopreservation of Cells. Cell Components, and Biological Solutions. BioPharm 11(9):42-61 (1998).

In the partial dendritic approach, the composition of solutes may be selected to achieve the desired ice crystal formation at the desired temperature. This is based on solute ejection from ice crystals and formation of the interdendritic matrix from the solutes. These solutes may solidify as eutectics or in the glassy state with a certain amount of water trapped (Wg') at the freezing glassy temperature (Tg'). R. Wisniewski, Developing Large-Scale Cryopreservation Systems for Biopharmaceutical Systems, BioPharm 11(6):50-56 (1998). Depending on the solute selected, the Tg' may vary widely.

If a very low Tg' solute is selected, the solution between the dendrites may remain a liquid even at the final freezing temperature, thus insuring that the composition would remain in a partial dendritic state. For example, some solutes may be selected such that unfrozen material may be present inter-dendritically even at temperatures below -25 oC . However, the solute concentration should not be too high - the ice crystals should be in a dominant phase with the lipid membranes between them. Otherwise, the liquid phase may provide a dampening effect that might partially or completely protect the lipid membranes from disruption.

The rate of freezing also plays a role in the formation of ice crystals within or near the lipid membranes, particular when the lipid membrane is a cell membrane. Slow cooling may induce cell osmotic dehydration between the dendrites due to solute concentration increase around the cells. As the cells attempt to achieve osmotic equilibration, they may shrink and lose enough free water that mostly bound water remains. Further temperature decrease may not lead to significant intracellular ice formation, because bound water freezes at very low temperatures, and may form glassy states when associated with molecular structures. When freezing occurs quickly, the cells may not osmotically equilibrate, with the result that there may be substantial amounts of free intracellular water left. This water will form more ice crystals inside the cell than in the slow freezing case.

Another way that controlled freezing/thawing may be used to achieve increased disruption is the use of crystallizing solutes in the recited compositions. The crystallized form of the solutes may help to disrupting the lipid membranes. After processing, when the temperature of the composition is increased, the solutes may dissolve again, permitting easy removal. Of course, any such solutes are preferably product- compatible and easy to remove from solution, unless they are a part of the buffer.

The invention may be practiced in such as way as to recover a variety of products that are partially or completely associated with or contained within lipid membranes. For the purposes of this invention, lipid membranes include, but are not limited to cells, cellular components, cell fragments, cell walls with embedded structures, liposomes and vesicle structures. Cells may include single cell organisms, and cells from multiple cell organisms (such as cells from tissues). For the purposes of this invention, products are defined to include, but not limited to, viral particles, inclusion bodies, intermembrane proteins, enzymes, other proteins, cell organelles, DNA, RNA and other nucleic acids, other types of cellular products, and other materials synthetically produced in or near the lipid membrane, etc. These products all may be recovered using the invention. Suitable compositions for use in practicing the invention include, but are not limited to, aqueous or non-aqueous suspensions of lipid membranes, frozen granules containing lipid membranes, or preparations of dehydrated lipid membranes.

The controlled freezing and thawing of the present invention may be accomplished using a variety of apparatus. For example, in one preferable embodiment, the apparatus used is a bulk vessel. Disclosure regarding preferable bulk vessels may be found in L. Leonard and R. Wisniewski, U. S. Patent Applications No. 08/895,777, 08/895,782, and 08/895,936, the disclosure of which is incorporated herein by reference. Alternatively, in another preferable embodiment, the controlled freezing and thawing of the present invention may be accomplished using a freeze granulator. Disclosure regarding freeze granulators may be found in R. Wisniewski, U. S. Patent Applications No. 09/003,283 and 09/003,288, the disclosure of which is incorporated herein by reference.

Freeze granulators can be used to cause disruption of lipid membranes for a variety of reasons, including product release. During freezing, the freeze granulator produces large number of ice crystal granules, preferably in an agitated suspension comprising the lipid membranes.

Granules may also be obtained by directing a spray of the recited composition into a stream or bath of cryogen, such as liquid nitrogen. Examples of apparatus that may be useful for performing this task are disclosed in Temple et al., U.S. Patent No. 4,655,047, and Milankov et al., U.S. Patent No. 4,982,577.

The granules obtained from the freezing by submersion in liquid nitrogen are different than those from freeze granulation. While the granules from submersion in liquid nitrogen are generally hard, high density solids, granules produced using a freeze granulation process may have internal pores created by the cryogens entrapped during granule formation. The solid, hard granules have many fine crystals and very thin layers between the crystals where the cells reside. Subsequent processing by a disruption augmenting technique, such as mechanical shearing, of the solid granules may produce inter-crystalline stress quickly upon applying force. Application of stress moves these crystals against each other and causes cell damage almost instantly. In contrast, for the porous granules produced by freeze granulation, there is generally first a collapse state while the stress is still moderate, and after further application of force the stress in material would build up rapidly. Additionally, the internal structure of freeze granulation granules may be different - there tend to be more relatively larger crystals separated by "sheared dendrites" which are like compacted "snow." This may be one of the reasons that the freeze granulation granules are less dense that the granules produced by submersion in cryogen baths. The porosity of the freeze granulation granules may be increased by using certain types of cryogens, such as solid carbon dioxide, that sublime after the granule is formed and leave a void. Granules obtained by submersion in liquid cryogen (for example, nitrogen) often sinter during storage and become difficult to handle. In contrast, porous granules generally do not significantly sinter under typical storage conditions.

In the freeze granulation granules, the lipid membranes may be thus mixed with the snow crystals, and other ice crystals. The snow crystals also may include the "late arriving" nuclei of the crystals built by falling carbon dioxide particles or fine liquid nitrogen droplets prior to the granule formation (at the end of the heat of fusion temperature plateau period). These crystals form rapidly and thus may be in a nonequilibrium state suitable for later recrystallization.

While in the case of the granules obtained from droplets submerged into liquid nitrogen there is no practical control regarding internal ice crystal structure and the distribution of cells between the ice crystals, the granules formed in freeze-granulator can be "manipulated", e.g. by delivery of cryogens (rate, continuous/intermittent), agitation (speed and pattern) and application of machine jacket (warming the walls at the temperature plateau - melting a thin layer of freezing ice crystals and product at the wall), the cell treatment prior and during granulation may be determined as well as the internal structure of granules (fewer/more larger ice crystals versus the snow fraction, fraction of pores from cryogen sublimation or evaporation, "compactness" of the snow, overall granule density, etc.). For example, extending the time duration of the temperature plateau at increased agitation rate would produce more snow fraction due to breaking the dendrites from larger crystals, while at slower agitation would produce larger fraction of bigger crystals. The change (reduction) in cryogen delivery at the end of temperature plateau (prior to the granule formation) would reduce internal granule porosity.

As a result the interior of the granules is quite complex and this might reflect upon the subsequent mechanical process of compression. The mechanical processes may be different for these two types of granules:

-for hard granules - compression and shear (simultaneously)

-for porous granules - compression (volume reduction first) followed by "kneading" under moderate compression and shearing under high compression (subsequent steps).

As noted above, granules with minimum internal porosity are most desirable for compression. For recrystallization only, without mechanical compression, the fraction of the snow is preferably as large as possible and the snow portion should preferably contain many ice crystal nuclei. These tend to be formed at the end of the temperature plateau - e.g. prior to the granule formation/mass braking. To accomplish this, the delivery rate of cryogen(s) should preferably increase at the end of the temperature plateau. Disclosure regarding the delivery rate of cryogen(s) may be found in L. Leonard and R. Wisniewski, U.S. Patent Application Nos. 08/895,777, 08/895,782, and 08/895,936. Delivery of more carbon dioxide particles or liquid nitrogen droplets tends to mean more nuclei formed. The size of the ice nuclei may be controlled by the size of liquid nitrogen droplets (coarse versus fine spray, spraying nozzle design, liquid nitrogen pressure, etc.) and size of solid carbon dioxide particles (fine or coarse spraying orifices, higher or lower liquid carbon dioxide pressure prior to the nozzle, the shape of the carbon dioxide horn, etc.). Formation of fine ice nuclei is promoted by formation of fine liquid nitrogen droplets and very small solid carbon dioxide particles. Size of the ice nuclei may range from about 0.1 micrometer to about 10 millimeters, more preferably from about 5 micrometers to about 4 millimeters.

The porous, freeze-granulation granules may be treated differently during storage and thawing than solid granules, which may be treated similarly to the complete dendritic state described above.

Recrystallization of the "snow" may cause rearrangements of the frozen matrix because the lipid membranes are mixed and frozen with the "snow" fraction. This may be beneficial at low temperatures when the granule is still relatively hard. If the temperature increases the granule may "soften." Additionally, the internal porosity may also contribute to decrease in overall granule hardness. Further warming may cause a partial granule "collapse" due to the granules' softness and internal pores. These processes may not be beneficial for lipid membrane disruption. Subsequent lowering of temperature may produce more compact granules, e.g., more suitable for mechanical compression treatment.

Extended holding periods at lower temperatures, but still within the range of recrystallization, may be preferable. Such periods may be considered storage or "curing" periods. The length of these periods is preferably such that there is relatively small amounts of granule surface fusion (i.e. minimal forming of large blocks). Some temperature cycling around these holding temperature levels also may be used, if needed for recrystallization, etc.

In another preferable embodiment, having frozen and stored the recited composition, e.g. in the form of granules, at approximately -50 to - 86 °C, one may attempt to treat it with a temperature shock to create increased internal stress (i.e. tribological stress). For example, the recited composition, e.g. granules, may be plunged into liquid nitrogen, thus inducing a rapid temperature change to about -196 °C. This is designed to generate tribological stress between the ice crystals that are present. This temperature shock technique is preferably undertaken when minimal amounts of solutes are present in the recited composition. High solute concentration or the significant presence of glassy state material may interfere with the effectiveness of this technique.

As noted above, since the freeze granulation process occurs near 0

°C, any biological products contained within or associated with lipid membranes are well protected -i.e. there is much reduced thermal denaturation danger.

If the lipid membrane disruption process is accomplished using freeze granulation, there are at least two ways to end the disruption process. In the first case, the temperature is elevated to a few °C above zero; for example by applying the heating jacket. The ice crystals then melt and a suspension/slurry of disrupted membranes and biological product can be discharged for further processing/extraction, etc. The temperature can be maintained low enough (approximately -10 to +37 °C ) to permit protection of the biological products. An advantage of this process is that the ice crystals do the job and then melt, with the result that the original water content is restored in unfrozen composition suspension. Additionally, the ice crystals may be used as a grinding media replacement in grinding mills, with improved results versus conventional glass bead grinding mills in that the problems of wearing and cracking beads, need for cooling via jackets, need for cell mass-bead separation, etc. are much reduced or eliminated.

A second way to end the disruption process is to make granules from the disrupted membranes and aqueous media using a typical freeze- granulation procedure as described in R. Wisniewski, U. S. Patent Applications No. 09/003,283 and 09/003,288. For example, "typical" granulation may be used. In this process, the agitation may be slowed, and the spraying of cryogenic fluids may be continued until a granule forming point is reached. At this point, the frozen aqueous media and disrupted membranes break into granules, the temperature plateau ends and further cooling produces final freezing of the water remaining in the granules while the bed of granules may be agitated. Frozen granules may be stored and taken in portions for further processing according to biopharmaceutical production needs.

To release biological products contained between the inner and outer membranes (like intermembrane proteins from E. coli bacteria), process optimization may need to be done, depending upon the specific case. Factors that may be used in this optimization process include the speed of agitation in a freeze-granulator, time of processing/number of cycles, control of processing temperature, cell type, cell concentration, presence of solutes, and concentration of solutes. Additionally, disruption augmentation techniques may be tacked on to improve recovery yields of a biological product.

While bulk freezing techniques offer a number of advantages, for example ease of material handling, certain problems may be associated with bulk freezing. For example, freezing times or lipid membrane settling during freezing may become a problem. In such cases, it is possible to practice the invention using freezing in thin layers as an alternative to freeze granulation (which also may be used to address such problems). Thin layer freezing or thawing offers advantages of reduction of the freezing time as well as reduction in lipid membrane settling and more uniform final lipid membrane distribution, as compared to bulk freezing. Additionally, temperature control of the freezing front is more straightforward than for bulk freezing or thawing, as it may approximate two-dimensional freezing models.

Thin layer freezing may be accomplished using any method for providing a relatively thin layer of aqueous media and lipid membranes. Preferable thicknesses for thin layers in the context of this invention range from about 0.5 mm to about 50 mm, more preferably from about 1 mm to about 25 mm.

Thin layer freezing/thawing methods according to the invention include, but are not limited to, bags, flat containers or trays, or spraying thin layers of material on cold rotating drums or moving heat conducting belts, bands, screens, etc..

In the case of bag freezing or thawing, there are a number of structural details to keep in mind: any cold plates contacting the bags are preferably relatively smooth, the bags are preferably placed vertically to reduce air pockets along the cooled side walls, the side walls of the bags are preferably parallel, and the bag side wall external surfaces are preferably smooth. This last structure is preferable because micro- corrugated bag surfaces present in typical blood bags collect myriads of tiny air pockets that hinder conductive heat transfer to or from the bags. In a preferable embodiment, the bags may also be frozen continuously between two cooled belts or bands, as a thin bag can be frozen in about 20 min. by contact cooling

In the case of trays, a significant structural detail is the contact area of the tray with a conductive heat transfer surface. This may be maximized to maximize the conductive heat transfer. In a tray embodiment, the trays should preferably be relatively flat and "match" the flatness of the conductive heat transfer surface they are placed on.

Another useful way of cooling trays or bags is to put them in blast freezers that blow high velocity cold air around them. Of course, in such embodiments, the air flow distribution should be controlled carefully to insure proper cooling of the trays or bags. Freezing of bags or trays may be done in food blast freezers, for example, with a conveyor moving the trays or bags through the "air blasted" zones. In such a way more uniform distribution of blasting cold air is possible. However, the effects of conveyor support versus an open cooled upper surface may cause undesirable freezing asymmetry.

In the preferable embodiment where the aqueous media and liquid membranes are sprayed onto a cold rotating drum, it is possible to have drums that are hollow or solid. A solid drum embodiment is shown in FIG. 1. Solid drum apparatus 100 includes solid drum 102, product spray 104, cryogen sprays 106A-B, scraper 108, product chute 110, and chute for collection of cracked products chips 112. In operation solid drum 102 may be sufficiently cooled by first spraying the outer surface of the drum with the cryogen spray 106A, wherein the cryogen may be liquid nitrogen or liquid carbon dioxide, etc. Following this first spraying, the aqueous media containing the liquid membrane may be sprayed or coated onto the drum, through product spray 104. Additional cryogen may be sprayed through cryogen sprays 106B, to increase the rapidity of the freezing of the aqueous media and liquid membranes. Scrapers 108 or equivalent structures may be used to clean the drum surface and recover the frozen aqueous material in liquid membranes. These materials drop into product chute 110. Chute for collection of cracked product chips 112 serves to collect product chips that may drop off of the drum before the product reaches the scraper. Scraping the frozen product from the coated drum may produce partly "fluffy" product which would require preliminary compression similar to the case of the porous granules above, e.g., preliminary compression, then "kneading" and high compression and shear.

With respect to hollow drums, it is possible to use both thick walled and thin walled drums in the practice of this invention. Thick walled drums offer the advantage that they provide more thermal capacity than thin walled drums -- the drum cannot rapidly warm up locally where the aqueous media and lipid membrane is sprayed (this holds true as well for the solid drum embodiments).

Alternatively, a thin walled drum, as shown in FIG. 2, may be used. Thin walled drum apparatus 200 includes thin wall drum 208, product spray 104, cryogen sprays 106A-B, internal cryogen sprays 202, compression rollers 204, compression rollers scrapers 206, scraper 108, and product chute 110. In operation thin-walled drum 208 is cooled on its outer surface by cryogen sprays 106A, and on its internal surface by cryogen sprays 202. Suitable cryogens include, but are not limited to, liquid nitrogen or liquid carbon dioxide. In a preferable embodiment, one or more of interior cryogen sprays 202 are located on an opposite side of the thin walled drum 208's surface from where product 104 is sprayed. In another preferable embodiment, not shown, thin walled drum 206 may have its inner surface thinned/ribbed to provide better liquid cryogen evaporation and higher heat transfer. Product spray 104 sprays product onto the outer surface of thin walled drum 208. Cryogen sprays 106B serve to further cool the product. In addition, interior cryogen spray 202 cool the thin walled drum 208, which then cools the sprayed on product through conduction. Compression rollers 204 serve to compress the frozen product, producing disruption augmenting shear, as is discussed more below. Compression rollers scrapers 206 serve to remove any product that adheres to the outer surface of compression rollers 204. Scraper 108 removes any product that remains adhered to the surface of thin walled drum 208 after it has passed through compression rollers 204. The dislodged product is collected in product collection chute 110. Of course, other drum configurations may occur to one of skill; so long as they can be used to create controlled freezing or thawing conditions of the aqueous media and lipid membrane, they are within the scope of this invention.

In another preferable embodiment, the processing may also be performed on the inner drum surface, but then proper frozen aqueous media and lipid membrane removal from the interior should occur. This can be accomplished, for example, by use of an auger.

In an alternative preferable embodiment, the thin layer freezing occurs on a band or belt stretched between rotating drums or rollers as shown in FIG. 3. These bands or belts may be of stainless steel or other appropriate materials, and may consist of links or a continuous belt or band. Examples of suitable apparatus may be found in Schermutzki, U.S. Patent No. 4,154,379, or Ulrich et al., U.S. Patent No. 5,326,541 , hereby incorporated by reference. Advantages of such an embodiment over the drum embodiment include a larger capacity process because the belt surface may move faster than the drum surface, and longer product residence times, because of a larger potential surface area. FIG. 3 shows thin layer freezing apparatus 300 which includes first freeze roller 302, second freeze roller 326, product spray 304, product 306, top cryogen sprays 308, bottom cryogen sprays 310, compression roll cryogen spray 312, compression roll 314, compression roller scraper 316, product chute 318, scraper 320, cryogen spray 322, and band 324. In operation band 324 is stretched between freeze roller 302 and second freeze roller 326. Cryogen sprays 322 are used to cool first freeze roller 302. Product is sprayed through product spray 304 onto band 324. Product 306 travels across the gap between the first freeze roller 302 and the second freeze roller 326. Product 306 is cooled by top cryogen sprays 308, and bottom cryogen sprays 310. Compression rollers 314 serve to augment the disruption of product 306. Compression roller cryogen sprays 312 serve to cool compression rollers 314. Compression roller scrapers 316 serve to remove any adhering product from the compression rollers. Scraper 320 serves to remove frozen product from second freeze roller 326, whereupon the product drops into product chute 318. In addition to controlled freezing or thawing, it is possible to use disruption augmenting techniques in the practice of this invention. These techniques may increase the disruption of lipid membranes contained in the aqueous media, and thus potentially improve biological product recovery.

Generally speaking, a wide variety of disruption augmenting techniques may be used in the practice of this invention. These include, but are not limited to, addition of lytic materials, use of mechanical shear, and osmotic shock, homogenizing, and other equivalent methods. Such disruption augmenting techniques are generally used to augment disruption caused by controlled freezing or thawing according to the invention. In a preferable embodiment, osmotic shock is used to augment disruption caused by controlled freezing or thawing. Generally, the mechanics of osmotic shock are well understood by one of skill. In practicing the invention, the amount of solutes in the aqueous media preferably is not at a level that would provide cell protection between the ice crystals in the embedded state. The initial concentration of solutes should be preferably zero or a de minimis amount. In such case, the lipid membrane may swell osmotically with water and be prone to rupture during freezing. This is a valuable technique particularly when the lipid membranes are the membranes of cells. Intracellular ice crystal formation is thus promoted, and disruption may be augmented.

In the embodiment where the lipid membranes are cells, it may be preferable to wash the cells using deionized, purified water prior to freezing. Alternatively, as shown in FIG. 4, in embodiments that make use of freeze-granulators, and to increase the effect of osmotic shock, a batch of ice water (kept as close to 0 °C as possible) may be dumped rapidly into the freeze granulator after the power requirement reaches its high value. At that point, there are a large number of ice crystals, and the material inside the freeze granulator is approaching typical granule formation conditions. The water is preferably of deionized, distilled or ultra pure quality (18 Mohm) to provide a strong osmotic shock and swelling, particularly in the case where the lipid membranes are cell. In this embodiment, when the cells have been previously dehydrated during previous cooling periods, the swelling therefore is accentuated, with concomitant increases in disruption. The volume of the water should be such to make the solution much more diluted. The water temperature is preferably close to 0 oC, in order not to melt too many ice crystals. At the same time that the water is added, it is possible to add additional cryogen. This permits the water heat capacity to be matched by the addition of the cryogen, and reduced the melting effects that may be caused by the addition of water. This may be desirable because there tends to be more shear effect if the number and size of ice crystals are maintained after water addition.

Mechanical shear is another useful disruption augmenting technique. The type and amount of mechanical shear applied may be varied depending upon the controlled freezing or thawing history of the composition.

For example, take two of the partial dendritic states: the rubbery state and the slushy state. In the rubbery state, most of the water is frozen, ice crystals are present, intracellular free water may be frozen, but the bound water in cells may not be. Disruption in this state might work mostly on the principle of reaching the cell membrane fragility level. This might be accomplished by "kneading" - e.g. the movement of ice crystals may be significant and last some time. Batch or continuous, single or multiple shaft kneading devices can be used. Cooling during kneading may be required due to the mechanical energy input.

In the slushy state, disruption may require vigorous mechanical

"beating," by an agitator for example, because impacts between the ice crystals, with lipid membranes in-between, form an important source of disruption. The ice crystals may act almost like the glass beads in glass bead grinding mills. Cooling during beating, due to high energy input, may also be useful. Beating may be also conducted at temperatures near/below the membrane fragility level - it may be still possible to maintain the slushy state if low freezing temperature solutes are used, e.g., at temperatures significantly below 0 deg C. The temperature level is preferably low enough to prevent or reduce action of proteolytic enzymes (time- temperature factors).

In either state, fast or slow freezing may be used. In slower freezing, the large ice crystals formed thereby are acceptable because multiple relative ice crystal motion may be applied. Using fast freezing, with resulting small crystals, also may be acceptable, provided that good cooling is applied to reduce crystal melting due to mechanical work. However, if the ice crystals are too large, the number of ice crystal to ice crystal contacts may be too small. Conversely, if the ice crystals are too small, the ice crystals may melt too quickly. This suggests that there is an optimum ice crystal size, depending upon the particular application. Preferably, the diameter of the ice crystals range from about 200 to 1000 micrometers.

When using freeze granulators in the practice of the invention, it is possible to increase the shear exerted upon the lipid membranes by increasing the rotational speed of the agitator. This is preferably done after producing a sufficient number of ice crystals by spraying cryogens such as liquid carbon dioxide or/and liquid nitrogen into the aqueous media and lipid membranes contained in the freeze granulator to produce controlled freezing and or thawing. As noted above, the ice crystals formed in the freeze granulator may work as beads (analogous to a bead mill) for damaging the biological membranes. Disruption is also augmented by the fragility of lipid membranes at freezing conditions -- for example, lipid membranes are far less fluid at freezing conditions than they are room temperature.

When mechanical shearing is used as an disruption augmenting technique, the mechanical energy input, from intensive mixing for example, may produce enough heat to cause the ice crystals to begin melting. It is possible to counteract this heat addition by adding cryogens, such as liquid carbon dioxide or liquid nitrogen, at rates calculated to offset the heat addition. In certain circumstances, even though addition of cryogens may be sufficient to counteract the heat addition, the average ice crystal size may decrease. If this becomes a problem, a "cyclic" process can be applied. In such a process, the mixing energy applied to the aqueous media and lipid membranes can be temporarily decreased to let the ice crystals' size increase. This is then followed by an increase in mixing energy to continue the disruption augmentation by shearing.

In embodiments where the freeze granulator has a jacket and optionally cooled or heated shaft or paddies, it can also be used for granule thawing and the final cell breaking after warming the granules into "slush" - or partial dendritic state. Once the partial dendritic state has been reached, very rapid agitation may be applied for shearing the lipid membranes.

Another disruption augmentation technique for applying shear to lipid membranes is compression. Versions of compressing machinery may include two or more rollers with frozen aqueous media and lipid membranes fed into the gap between the rollers. In a preferable embodiment, the rollers are gravity fed. The rollers can be cooled to compensate for mechanical energy input occurring as a result of the work done in crushing. More than one set of rollers may be used in the compression step, to achieve an enhanced shear history for the frozen aqueous media and lipid membranes. Additionally, the rollers may have assisting auger or oscillatory-motion feeders at the nip of the rollers to push material between the rollers (no bridging). The rollers can work with both completely frozen and partially frozen (partially dendritic) materials. In a preferable embodiment, the aqueous media and lipid membranes are sprayed or coated onto one or more cooled drums, as have been described above, to achieve thin layer freezing. An additional roller or rollers are pressed up against the freeze rolls to provide compressive shear. The additional rollers may rotate at the same angular velocity as the freeze rolls, or may rotate at a different angular velocity to add an additional degree of shear. The additional rollers may be also cooled by spraying the cryogens if a very low crushing temperature is desirable, or to counteract heat generated by the mechanical act of crushing. This is shown, in a preferable embodiment, in FIG. 2.

Compression can be also applied during "holding'Vstorage periods within the ice recrystallization temperature range to increase recrystallization level.

In an alternative preferable embodiment, the aqueous media and lipid membranes are formed into granules, using the freeze granulation methods disclosed above, or equivalent methods. These granules are then fed into compression rollers to augment disruption. Various aspects of this embodiment are illustrated in FIGS. 5-6.

It is possible to spray solid carbon dioxide particles or droplets of liquid nitrogen or liquid carbon dioxide into the granules entering the space between the compression rollers. This may provide internal volumetric cooling of the frozen aqueous media and lipid membranes and offset the mechanical energy generated by the compression rollers.

Additives, including but not limited to polyethylene glycol (PEG), glycerol, or sucrose solution, may also be sprayed between the compression or freeze rollers to protect released biological product and possibly improve disruption of the lipid membranes. The surfaces of the rollers may be smooth or grooved. The grooved embodiment is illustrated in FIG. 7. Either roller may be grooved, or both may be grooved. Deeper grooves ("hills and valleys") may require synchronized roller drives. Also one roller may be driven and the second "following" the drive roller by being pressed against the first one.

In a preferable embodiment, as illustrated FIG. 8, suspensions or concentrated suspensions (such as a paste) may be passed through the disruptive machinery after being mixed with ice particles (small cubes, frozen droplets, etc.). The ice may be made using purified water or from water with additives in a separate process (non-biological). The ice may then be cooled down by spraying with cryogens or by bulk mixing with cryogens to achieve a low temperature, preferably -50 °C, or lower. The ice particles can also be produced using liquid nitrogen or solid carbon dioxide (spraying into, or mixing/spraying similarly to the freeze granulator). This type of production of ice particles, because it is non-biological, may not require the stringent equipment surface qualities required of biological equipment, thus reducing cost. Additionally, it may be possible to vent the gases from this part of the process to atmosphere, because no biological containment may be needed (e.g., no low temperature HEPA filters, etc.), also resulting in a lower system cost.

Both streams, the ice granules and a suspension of lipid membranes, may be fed to a disruption augmenting machine, such as the compression rollers or the freeze granulator described above. The ice granules are mixed and crushed with the suspension of lipid membranes. The mixture may then be cooled to about - 30 °C or lower, with the result that the cell suspension freezes between the ice particles. Due to the mixing and crushing of granules, the material is prevented from from one solid block of frozen mass. Work done during the compression and shearing of this material may lead to temperature elevation. This may be handled as discussed above, by introduction of cryogens, or by jacketing the various pieces of process equipment and cooling them, preferably to temperatures of about - 50 °C or lower. The external zones/walls of the process equipment may also be sprayed with cryogens. The crushing and mixing continues until the material reaches a stage where maximum shear and compression occur, and the material is discharged to a bin. The process may be repeated by recycling the material or passing it to a second stage.

In an alternative preferable embodiment, the mechanical shearing may be created by use of a screw- or auger-based apparatus. Such apparatus may be configured with a single or multiple screws. A multiple screw apparatus is shown in FIGS. 9A-B. The screws may be contained within a tube, such as in an Archimedes pump-like design, or such as an extruder-like design, although other shear producing configurations are contemplated within the scope of the invention. An advantage of the screw/auger embodiment is that it is naturally contained. In the roller embodiment, a containment housing may be needed around the rollers, whereas the screw/auger based apparatus naturally contains the frozen material. Additionally, the shearing of material in the screw/auger machines can be extended in time and better controlled, as compared to the rollers.

Under extrusion conditions, as under many mechanical shearing conditions, there is also the possibility that changes in ice crystal structure, from hexagonal to other geometries, may occur with pressure increases. Additionally, there is the possibility that the ice crystals may melt at temperatures below 0 °C under pressure, depending on the pressures developed under mechanical shearing. If high pressure ice melting is achieved, then after the shear is ended, the pressure will drop and the mixture may quickly "refreeze". While melting the ice crystals under pressure may provide some improvements in processability (e.g. internal lubrication, etc.), it may lead to decreases in disruption. However, rapid re- freezing after pressure decrease may improve cell disruption and increase yield significantly. Optimization studies may be required to resolve such problems, depending upon particular circumstances.

In one preferable embodiment, the apparatus is configured with a single, smooth, auger contained within a tube as shown in FIG. 10. The auger moves the frozen aqueous media and lipid membranes along the inner surface where shear occurs between the auger edges and the inner surfaces of the tube. Internal shear also occurs between the auger flights, due to mixing. In such embodiment, the tube may be externally cooled, using conventional cooling jackets or by spraying cryogens on the tube. Furthermore, the auger may be hollow and cooled internally. Cooling can be provided by adding or spraying cryogen (such as liquid nitrogen or solid carbon dioxide) into the tube inlet, along with the material to be sheared. However, any gas produced by using the cryogen may have to be vented. Any cooling provided may be designed to reduce melting of any aqueous media present. If too much of such media is allowed to melt, the melted material may serve a lubricating function, thus reducing the shearing effect.

The auger can have cuts and solid bars attached to the inner surface of the tube may fill those gaps, with the result that additional material shear may occur there. Aspects of such augers are illustrated in FIGS. 11-14. The end of the auger preferably has additional shearing devices, such as profiled blades, that rotate against the end extrusion nozzle. These arrangements may take a variety of forms, such as a plate with two semi-circular orifices, or a plate with multiple orifices. The orifices can be internally profiled, including but not limited to converging, converging-diverging, diverging, round edge, knife edge, triangular - or star-shaped cross sections, thick plates with a "screw orifice" (like a static mixer), etc. Aspects of these arrangements are illustrated in FIGS. 15A-20. For example, nozzle or die internal geometry may increase and decrease pressure in material passing through, thus causing melting and re-freezing cycles. Such cycles may occur once or multiple times in passing through the nozzle or die.

In certain preferable embodiments, the apparatus may have two or more screws/augers. In such apparatus, more shearing of the material may be accomplished between the augers even before the material reaches the final extrusion orifices, and any additional shearing structures present there. The augers flights can be configured to maximize material shear during the transport and compression of the material.

The material fed into the auger/screw apparatus may be already frozen, for example frozen granules produced as discussed above, or as separate streams of ice particles, cryogens, and lipid membranes. This second approach offers the advantage, discussed above, that the ice particles may be produced separately, without the need of biological containment equipment. The ice granules may be cooled to very low temperatures (e.g., below about -140 oC ) and then mixed with the lipid membranes, such that freezing and shear may begin in the early stages of the compression/shearing (beginning of the screw/auger). The size of the initial ice particles may be controlled during their production and later optimized for maximum rate of lipid membrane disruption after passing through the machine. The ice particles and/or frozen granules may be "cured" prior to this process to obtain the best size distribution for disruption.

Claims

WHAT IS CLAIMED IS:

1. A method of disrupting lipid membranes in a composition that comprises the lipid membranes, the method comprising:

subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions,

wherein the lipid membranes comprise cells, cellular components, cell fragments, cell walls with embedded structures, liposomes, or vesicle structures.

2. The method of claim 1 , further comprising

recovering products that are partially or completely associated with, or contained within, the lipid membrane,

wherein the products comprise viral particles, inclusion bodies, intermembrane proteins, enzymes, other proteins, cell organelles, DNA, RNA and other nucleic acids, other types of cellular products, and other materials synthetically produced in or near the lipid membrane.

3. The method of claim 1 , wherein a spacing between dendrites present in the composition is adjusted to a size of the lipid membranes such that the lipid membranes are squeezed by the dendrites.

4. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises causing the composition to assume a complete or substantially complete dendritic state.

5. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises causing the composition to assume a partial dendritic state.

6. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises placing dehydrated cells in a glassy state.

7. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises thawing a frozen composition such that the temperature difference between an outer surface of the composition and an interior of the composition is minimized.

8. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises subjecting the composition to non-equilibrium freezing.

9. The method of claim 8, wherein the non-equilibrium freezing is caused by rapid freezing of the composition.

10. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises freezing the composition, followed by subjecting the frozen composition to temperature shock to create tribological stress in the frozen composition.

11. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises rapidly freezing the composition, and subsequently allowing the rapidly frozen composition to undergo recrystallization.

12. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises selecting a solute so as to modify final properties of a frozen state of the composition.

13. The method of claim 1 , wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises growing a ice crystal dendritic front at a velocity ranging from about three to about 250 mm/hr.

14. The method according to claim 1 , further comprising:

augmenting disruption of the lipid membranes using at least one disruption augmenting technique.

15. The method of claim 14, wherein the at least one disruption augmenting technique comprises inducing osmotic shock or providing mechanical shear.

16. The method of claim 15, wherein the mechanical shear is provided, at least in part, by ice crystals formed during the controlled freezing or thawing.

17. The method of claim 15, wherein the mechanical shear is provided by a freeze granulator apparatus.

18. The method of claim 15, wherein the mechanical shear is provided by compressing the composition between two or more compression rollers.

19. The method of claim 15, wherein the mechanical shear is provided by processing the composition through an auger shearing apparatus or a screw shearing apparatus.

20. The method of claim 15, wherein the osmotic shock is induced by the addition of water to the composition during freezing of the composition.

21. The method of claim 15, wherein subjecting the composition that comprises the lipid membranes to controlled freezing or thawing conditions comprises forming frozen granules comprising the composition, and the mechanical shear is provided by compression.

22. The method of claim 15, wherein the mechanical shear is further provided by kneading.