J. clin. Path., 1976, 29, 271-285

Long-term preservation of cells and tissues: a reviewDAVID E. PEGG

From the Division of Cryobiology, Clinical Research Centre, Watford Road, Harrow, Middlesex HAI 3UJ

There has been a steady growth of interest incryobiology during the past 10 years: this decadehas seen a considerable increase in our understand-ing of the effects of low temperatures and of freezingon living cells and a significant development of newand more refined methods of overcoming freezinginjury in order to achieve viable storage at verylow subzero temperatures. This has led to workersin many fields realizing, for the first time, thatcryobiology may be able to provide them with usefultools for their own purposes. Pathology is noexception. One of the very first cell types to bepreserved at very low temperatures was the erythro-cyte, and much of the subsequent theoretical workin cryobiology 4ias used this simple cell as a modelsystem. As a result, a whole range of red cell preser-vation techniques has been evolved, some applicableto clinical transfusion and others more suited to thelaboratory. The haematologist may also wish tostore lymphocytes, thrombocytes, haemopoieticcells, and possibly granulocytes. Microbiologistsneed to preserve microorganisms for referencepurposes, and also tissue culture cells and fetaltissues for virus isolation. Finally, experimentalpathologists may wish to preserve a whole range oftissues for their studies, skin and endocrine tissues,for example.

In this review I propose first to outline our presentunderstanding of the mechanisms of freezing injuryand cryoprotection and then to deal with practicalmethods of long-term storage that have been devel-oped for use in haematology and microbiology.Finally, I shall deal more generally with the develop-ment of preservation techniques for some morecomplex tissues that might find some relevance inexperimental pathology. I have quite arbitrarilydefined 'long-term' to mean 'longer than onemonth' for the purpose of this review.

Cooling of Cells to Temperatures above 0°C

The use of cooling to prolong the survival ofisolated cells and tissues depends on the effect thata reduction in temperature has on chemical reaction

Received for publication 18 December 1975

rates and hence on metabolism, thereby reducing thedemand for oxygen and substrates. Unfortunately,however, cooling to temperatures above 0°C doesnot provide adequate storage periods for manypractical purposes, and there are no common cellsfor which long-term preservation can be obtained inthis way. There are many reasons why this is so:metabolism does not cease at 0°C, nor are allreactions slowed to the same degree; consequently,inter-related metabolic pathways may be 'dislocated'by cooling. Some effects of cooling are franklyharmful: for example, cooling switches off theNa-pump, which is responsible for the regulation ofcell volume, and as a result cooled cells swell(Leaf, 1959); membrane lipids undergo phasechanges which may in themselves be harmful, andwhich also have dramatic effects on the reactionrates of membrane-bound enzymes (Lyons, 1972);poorly soluble materials may precipitate, anddissociation constants change, resulting in changesin the composition and pH of solutions (van denBerg, 1959; van den Berg and Rose, 1959); somecells are damaged or even killed by a reduction intemperature per se, especially if cooling is rapid, aphenomenon known as thermal shock (Lovelock,1955). Taken together the various changes that areinduced by cooling severely limit the time for whichcells can be stored at temperatures above 0°C, butwhen attempts are made to use lower temperatures,freezing occurs, and freezing itself is normallylethal to cells.

Freezing

Freezing is the separation of pure water as ice,which concentrates any solutes present in theremaining liquid phase. This raises the possibilityof two sources of freezing injury, ice itself and thealtered liquid phase. Let us look first at the changesthat occur in the liquid phase.The principal solute in biological fluids is sodium

chloride. When isotonic (0-15 M) sodium chloridesolution is cooled it may supercool a few degrees,but if seeded it freezes at -0-56°C. As cooling iscontinued, so further ice separates, sufficient ateach temperature to concentrate the salt in the

271

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

David E. Pegg

remaining liquid to produce a solution that has thatfreezing point. Thus, the remaining solution isprogressively diminished in volume and increased instrength until at -21d1°C the saline has reached aconcentration of 5-2 M; at this temperature, theeutectic point, the remaining solution solidifies.Therefore, when cells are suspended in isotonicsaline that is frozen, they are subjected to a 32-foldincrease in sodium chloride concentration. In themixed solute systems that occur in practice, similarchanges in osmolality occur, but in addition thereare changes in composition brought about by dif-fering solubility characteristics of the varioussolutes. This may be very important; for example,van den Berg (1959) and van den Berg and Rose(1959) have shown that when a solution containingNaH2PO4 and Na2HPO4 is cooled, its pH falls if themolecular ratio of the two compounds is less than57, but the actual pH is greatly influenced by theother solutes that are present. Mazur (Mazur et al,1969) has coined the term 'solution effects' toinclude all the changes that occur in the liquidphase as a result of freezing and the effects thatthey have on any cells that are also present in thesystem.

It is important to realize that ice formation isnormally entirely extracellular. There are severalreasons for this. In the first place, when heat isremoved by conduction from the external surface ofthe specimen, the coldest point will always be in theextracellular fluid. Secondly, the extracellular fluidforms one large compartment, whereas the intra-cellular space consists of very many small compart-ments: the probability of ice nucleation occurring inany given compartment is directly related to its size,and this makes it inevitable that nucleation willoccur in the extracellular fluid before a significantnumber of cells have frozen internally. Once icehas started to form, it will propagate throughout thatcompartment until equilibrium is reached. Hence,even if a few cells should freeze internally beforeextracellular freezing starts, once extracellular ice hasformed it will continue to grow in that space, andsince cell membranes are impermeable to the mainsolutes present, water will be withdrawn from thecells by the increased external osmolality. Hence thecells will shrink, and so long as cooling is slowenough to allow water to leave the cells to maintainequilibrium, no further intracellular freezing willoccur. Let us look next, therefore, at the effects ofraised external osmolality, and then return later tothe possibility of intracellular freezing at rapidcooling rates.

Clearly, the most obvious effect of raised externalosmolality is that cells shrink. Farrant and Woolgar(1972a, b) measured the mass ofwater in erythrocytes

exposed to solutions of sodium chloride and sucroseand showed that the cells reached a minimum volumeat about 1800-2000 mOsm/kg. Meryman (1968)suggested that when the osmolality is increasedbeyond this value, an actual hydrostatic pressuredifference is produced across the cell membrane,and that this may then damage the cell. However,this seems distinctly unlikely, because intracellularproteins reach such a high concentration underthese conditions, and their osmotic properties are soextremely non-ideal, that very small water move-ments are sufficient to maintain equal water activityon either side of the cell membrane. Certainly themembranes are damaged however; they becomeleaky to cations, and significant cell lysis is observedbut this effect is not sufficient to explain the damageobserved during freezing. Three observations madeon human erythrocytes by Daw et al (1973) illustratethis. First, the amount of haemolysis produced at agiven osmolality without freezing was less than thatobserved when the same osmolality was producedby freezing. Moreover there was less damage whencells were frozen in sucrose instead of sodiumchloride, whereas the same shrinkage and haemolysiswas produced by exposure to sgcrose withoutfreezing. In addition, freezing was found to renderred cell membranes permeable to sucrose, whereasexposure to equivalent osmolalities without freezingdid not. Clearly, additional factors were at workduring freezing.The obvious differences between the simple

hypertonic model system and actual freezing arethat freezing and thawing involve changes intemperature as well as concentration, and thatduring thawing the cells are resuspended in theoriginal osmolality. Lovelock in 1955 showed thaterythrocytes suspended in hypertonic salt solutions,sufficient to produce only minimal haemolysis if thetemperature was held constant at 37°C, were lysedwhen cooled to 0°C (see fig 1). This is a 'thermalshock' phenomenon, possibly analogous to thatwhich occurs naturally with some species of bacteria(Meynell, 1958) and spermatozoa (Smith, 1961a) butwhich occurs with erythrocytes only when they havebeen appropriately sensitized. Lovelock (1955)showed that hypertonic conditions elute lecithinfrom erythrocyte membranes, and he suggested thatthe low melting point of this lipid makes lecithin-rich membranes more pliable at low temperaturesand therefore less liable to fail under the stressesproduced by differential thermal contraction. Thisphenomenon must also occur during freezingbecause the cells are then exposed to rising saltconcentrations and falling temperature at the sametime.

Lovelock (1953a) also carried out a very elegant

272

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

Long-term preservation of cells and tissues: a review

60r

50 F

_ 40f-

E 33E

239 F

10

00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2

Molar concentration of sodium chloride

Fig 1 Haemolysis produced by cooling human redblood cells from 37°C to 0°C at 20°C min-' inconcentrations ofsodium chloride ranging from 0-2molar to 12 molar.

experiment showing that red cells exposed tohypertonic saline were damaged further when theywere returned to isotonic conditions. He showedthat when erythrocytes were frozen in 0-15 M sodiumchloride they could not be recovered without lysisif they spent more than 30 seconds in the temperaturezone between -4 and -40°C. He pointed out that- 4°C is the freezing point of 0 8 M sodium chloride,and that therefore as the cells are cooled throughthis zone they are exposed to external sodiumchloride concentrations increasing from 08 M at-4°C to 5-2 M at -21-1°C. Lovelock then showedthat when erythrocytes were suspended in sodiumchloride solutions varying from 0 6 to 3-6 m, thehaemolysis produced when they were resuspendedin 0-15 M sodium chloride was the same as that whichoccurred when suspensions in isotonic saline werefrozen to the temperature which results in the sameconcentration and then thawed again (see fig 2).

Thus, we now look upon the damage occurringduring slow cooling, when ice is exclusively extra-cellular, as being due solely to solution effects andconsider that this damage occurs in two stages:in the first stage, latent damage is sustained by themembranes, principally, but not necessarily solely,as a result of the concentrated electrolyte solutionsproduced by freezing; in the second stage,this damageis manifested by the additional stresses of cooling(thermal shock) and resuspension in isotonic salineas a result of thawing (dilution shock). Farrant andMorris (1973) have recently stressed the importanceof thermal shock, arguing that since cooling occursbefore redilution, thermal shock is likely to havedestroyed the sensitized cells before dilution shockcan do so.

This discussion so far has been limited to slowcooling where ice is exclusively extracellular. Sincethe processes that sensitize the cell membrane todamage on cooling or dilution are chemical innature, and since all chemical reaction rates aretemperature-dependent, it would be anticipatedthat accelerating the cooling rate would increasecell survival because the cells would then be exposedto the damaging solution effects for shorter timesat high temperatures. Direct experiment has con-firmed this expectation but has also shown thatsurvival eventually falls off if the cooling rate isincreased still further. Mazur (1963) has produced avery convincing explanation for this phenomenon.He pointed out that red cells, which have a highpermeability for water, have a high optimum coolingrate, whereas yeast, which is relatively impermeableto water, survives better at low cooling rates: itfollows from the permeability data that yeast cellswill tend to freeze internally at lower cooling ratesthan erythrocytes, and he was able to show thatsignificant amounts of supercooled intracellularwater start to occur in yeast at 10°C/min, and at3000°C/min in erythrocytes. These happen to bethe cooling rates giving maximum survival witheach of these cells, and if one postulates that intra-cellular freezing is damaging, by some mechanism asyet unknown, then the actual effect of cooling rateon cell survival can be explained. There is nowdirect experimental confirmation, both from lightmicroscopy (Diller et al, 1972) and from electron

100

60 t

60

E

640-

n A

20

0 -2 4 6 0 -10 2 -4 -15Temp ('C)

Fig 2 Haemolysis produced by exposing red cellssuspended in 015 M NaCI solution to the temperaturesindicated and rewarming them (-). The haemolysisproduced by exposing red cells to solutions having theindicatedfreezing points, and then returning them toisotonic saline are also shown (-). The agreementbetween the data is apparent. (Data from Lovelock(1953a) reproduced by permission of BlackwellScientific Publications)

. . .I db.--

273

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

David E. Pegg

microscopic examination of freeze-substituted ma-terial (Walter et al, 1975), that intracellular freezingdoes occur during rapid cooling and is associatedwith cell destruction. Figure 3 shows the effect ofcooling rate on the survival of four cell types; noteparticularly that mouse bone marrow cells, whichare typical nucleated mammalian cells, show lessthan 2% survival at any cooling rate. This isexplained by postulating that accelerating the coolingrate causes internal freezing to occur before the rateis fast enough to reduce significantly the damage fromsolution effects. If this is correct, a useful survivalrate would be obtained if some method could befound to reduce solution effects at cooling ratesslow enough to avoid intracellular freezing. It isfortunate that the accidental discovery of thecryoprotective action of glycerol did exactly this.

Cryoprotection

The historic discovery of the cryoprotective effect ofglycerol was made in 1948 when Polge et al (1949)found that fowl spermatozoa that had been cooledto -79°C in 1 1 M glycerol recovered with littledamage after thawing. Mammalian erythrocytesbehaved similarly. Lovelock (1953b) explained thecryoprotective action of glycerol on the basis ofcolligative action; he demonstrated that the rise insalt concentration when isotonic sodium chloridesolution was frozen was much greater than the risein glycerol concentration when biologically accept-able concentrations of glycerol were frozen (see fig4). It followed, of course, that when both sodiumchloride and glycerol are present, they will both be

80-

70 -

1ecst r bc--

50-

- 40-!:'30

" hamster tc cells?to-

mouse marrowI10- Icels

0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~10 100 1000 ioCC

Cooling rote (0,mir)

Fig 3 Graph showing the effect of cooling rate on thesurvival ofyeast cells, tissue culture cells (hamster TCcells), human red blood cells, and mouse bone marrowcells. Note the differing optimum cooling rates, and thefact that less than 2% of the bone marrow cells surviveat any cooling rate. (Reproduced by permission ofIPCScience and Technology Press Ltd)

)o t

_oE6 r

c

In

oc 4 -

ac,Z 2 -1o

04

0 6

08

0 -10 -20 -30 -40 -50Temp (°C )

Fig 4 Graph showing the rise in concentration of theindicated initial mole fractions ofglycerol ( )compared with isotonic sodium chloride (0-0027 mf-- -). It is apparent that the increase in mole fractionofNaCl in a solution that also contains significantamounts ofglycerol will be greatly reduced. (Reproducedby permission of Blackwell Scientific Publications)

concentrated to the same proportional extent, sinceit is the removal of solvent water to form ice that isconcentrating both solutes. He showed that thereduction in salt concentration achieved by theaddition of glycerol to a suspension of erythrocytesin physiological saline exactly paralleled the reduc-tion in haemolysis when the suspension was frozen.Whatever the glycerol concentration, haemolysisstarted when the sodium chloride concentrationreached a mole fraction of 0-014, and reached 5%whenever the mole fraction of sodium chlorideincreased to 0-02, irrespective of temperature orglycerol concentration (see fig 5).Nash (1962) studied many other neutral solutes

that are cryoprotective for erythrocytes and foundthat they had the common properties of ability topenetrate cells, lack of toxicity, and a high affinity forwater. Clearly, the latter two characteristics arevital; the absence of intrinsic toxicity is a self-evident requirement, and the ability to produceconcentrated solutions with a low freezing point isimportant because such compounds will be the mosteffective 'salt buffers'. However, the fact that manycommonly used cryoprotective agents penetrate cellmembranes actually produces some problems and iscertainly not a necessary property for cryoprotection.

Penetrating cryoprotective agents like glycerol anddimethylsulphoxide (DMSO) permeate a good dealmore slowly than water and consequently they allproduce osmotic transients, the severity and durationof which vary with the compound and the cell inquestion. In general, osmotic disturbances have

274

Un2

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

Long-term preservation of cells and tissues: a review

-03

z

t 020

=O-OOM2 =0 15M3 =0-30M4 = 0-50M5 =0 75M6 = OOM7 = 150 M8 = 200M

-20Temp(°C )

Fig 5 Graph showing the rise in mole fraction ofNaCIin solutions initially containing the indicated molalitiesofglycerol and isotonic with respect to NaCl. Whenred blood cells were suspended in these solutions 5%haemolysis was observed at a different temperature ineach case, but the corresponding molefraction ofNaCIwas always 0 016-0 021. (Data from Lovelock (1953b)reproduced by permission ofBlackwell ScientificPublications)

more severe effects during thawing and resuspensionof the cells in cryoprotectant-free medium (whenthe solute is leaving the cells) than during initialequilibration and cooling (when it is entering thecells). This is due to the greater sensitivity of cellsto swelling than to shrinkage. It is certainly veryimportant to minimize osmotic disturbances duringthawing and subsequent manipulations since theycan contribute very significantly to the damagesuffered by stored cells and tissues (Thorpe et al,1975). When this is done it is found that many of themost effective cryoprotectants do penetrate cells,and Lovelock actually showed that if erythrocytesare rendered relatively impermeable to glycerol bytreatment with Cu++ ions, they are less completelyprotected by glycerol.

In 1955 Bricka and Bessis showed that erythrocytessuspended in high concentrations of dextran or

polyvinylpyrrolidone (PVP) would survive freezingat very low temperatures, and it was subsequentlyshown that PVP would protect bone marrow andtissue culture cells (Mazur et al, 1969). For manyyears it was supposed that such macromolecularcompounds must have a different mechanism ofaction from glycerol, because the mole fraction ofthe protective compound was very low. However,the solution properties of macromolecules are wellknown to be very non-ideal (Jellinek and Fok, 1967),especially in high concentration, and Farrant andWoolgar (1970) have shown that the presence of15-30% of PVP depresses the rise in salt concentra-

ion of the system PVP-NaCI-H20 sufficiently toenable red cells to survive exposure to -10°C for20 minutes. Even though these compounds do notpenetrate into the cells they will, of course, reducethe build-up of salt concentration inside the cells,since intracellular solute concentration will be inequilibrium with external salt concentration, whichis reduced. An additional factor has been pointedout by Woolgar (1972), who showed that haemolysiswas reduced by the presence of external colloid(PVP) when frozen red cell suspensions were thawed;this was probably due to the balance that the PVPprovided for the intracellular colloid once the cellmembranes had become leaky to cations. Othermore specific actions of certain high molecularweight materials may also play a part in cryo-protection, but it seems likely that the most import-ant action of high molecular weight, non-penetratingcryoprotective agents is the same as that proposedby Lovelock for glycerol.

It will be appreciated that all these cryoprotectiveagents affect only the part of the graph relatingdamage to cooling rate which has been ascribed tosolution effects. This was very elegantly demonstra-ted by the study of Leibo et al (1969), who showedfor mouse bone marrow cells that increasing theglycerol concentration produced progressively greatersurviving fractions at progressively lower coolingrates, but left the 'intracellular freezing' part of thecurve completely unaffected (see fig 6).

Cell Preservation

Long before much of the fundamental informationreviewed above was available, experimental investi-

80-

70-

60-

50-

-40-1-25Mv<

30-Z>30-/\_C-0.8 M-m

10-04M

0.1 10 100 100 1000Cooling rate (0/min)

Fig 6 Graph showing the effect of cooling rate on thesurvival of mouse haemopoietic cells cooled in thepresence of various concentrations ofglycerol. Note thatincreasing the glycerol concentration moved the optimalcooling rate downwards and increased maximum survival.(Data from Leibo et al (1969) reproduced by permissionofIPC Science and Technology Press)

275

W0

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

276

gation had shown that many cells could be preservedby 'slow' (-1°C min-') cooling in the presence ofmoderate (1-2 molar) concentrations of glycerol orDMSO, and this came to be regarded as the 'standardtechnique'. The accumulation of information on themode of action of cryoprotectants and the effect ofcooling rate has shown that this is far too simple aview; each cell should be considered separatelysince variation of the cooling rate, the cryoprotectiveadditive, and the technique of post-thaw handlingwill produce different, sometimes very different,optimum conditions for the storage of each cell.Moreover, it should not be assumed that themaintenance of a single cooling rate throughout theprocess is optimal; in the original work on spermato-zoa, Polge and Lovelock (1952) obtained their bestresults by cooling at 1 °C min-1 to - 15°C and thenat 5°C min-, and the same technique has been usedfor bone marrow (Pegg, 1964a) and other cells(Smith, 1961b). Luyet and Keane (1955) developeda method known as two-step cooling in whichcells, suspended in a medium containing a suitablecryoprotective agent, are cooled rapidly to a criticalsubzero temperature, are held at that temperaturefor a sufficient length of time, and are then cooledrapidly to the storage temperature, usually - 196°C.Studies recently carried out by Walter et al (1975)have shown that during the holding time, when thecells are exposed at a low temperature to a highexternal osmolality, they shrink, and that whenthey are then cooled rapidly, they do not freezeinternally. Thus, the mechanism of cryoprotectionis precisely the same as that in controlled-ratecooling, but because the cells are dehydrated at alower temperature, where chemical reaction ratesare slower, it seems reasonable to predict that overallsurvival may be improved over the controlled-ratetechnique where the cells spend longer at highertemperatures during freezing. So far the two-stepmethod has been shown to be applicable to bullspermatozoa (Luyet and Keane, 1955), humanlymphocytes, and hamster fibroblastic tissue culturecells (Farrant et al, 1974); -25°C seems to be asuitable holding temperature and 10 minutes asufficient time. The maximum reported survival rateshave exceeded those obtained by controlled-ratemethods, but so far the method has proved successfulonly with samples of 0 5 ml or smaller.

It will be realized therefore that there are a varietyof techniques available for cell preservation, and inthe present state of knowledge there is no certainway of predicting the optimum method for an asyet untested cell. In the rest of this paper I proposeto review the methods that have been worked outfor cells and tissues that are of particular interest topathologists.

David E. Pegg

Applications in Haematology

Of all the medical specialities, haematology hasprovided the greatest number and some of the mostimportant applications of cryobiology. The clinicaltransfusion of red blood cells and platelets renderssome form of storage mandatory; currently, thereis a resurgence of interest in bone marrow trans-plantation and granulocyte transfusion, so preserva-tion of these cells is also relevant; lymphocytes arestored for laboratory reference purposes in tissuetyping and cytogenetic analysis. We will examine therequirements for successful preservation of each ofthese cell types in turn.

PRESERVATION OF ERYTHROCYTESFigure 3 shows the relationsip between cooling rateand survival for unprotected human erythrocytes.The high water permeability of these cells makesthem very resistant to intracellular freezing, andconsequently it is possible to obtain usable recoveryrates in the absence of any cryoprotectant. Theaddition of low concentrations of glycerol, forexample, or of non-penetrating agents that dehydratethe cells slightly before freezing makes it possible toobtain virtually 100% survival with rapid cooling.The development of practical preservation methodsbased on these observations is a simple matter ifsmall volumes are sufficient and sterility is notrequired, as is the case, for example, with serologicaltesting. Meryman and Kafig (1955) described onesuch technique; blood, to which sufficient glucosehad been added to yield a final concentration of 7 %,was sprayed from a syringe fitted with a fine needleonto the surface of liquid nitrogen. The dropletsfroze separately and then sank to the bottom of theliquid nitrogen where they were stored; thawing wascarried out by sprinkling into warm saline. Hunts-man et al (1960) found 130% sucrose to give highersurvival rates; they recommended the dilution ofEDTA-anticoagulated blood with one half itsvolume of 400% w/v sucrose solution immediatelybefore dispensing drop by drop into a container ofliquid nitrogen. Alternatively, the cell suspensionmay be held as a film on a piece of stainless steelgauze, which is then dipped into liquid nitrogen(Bronson and McGinniss, 1962), a technique thatavoids the possibility of frozen droplets fromdifferent blood samples getting mixed.

Provided that a well insulated Dewar vessel isused for storage, the samples may be stored in thegas phase above the liquid nitrogen where the tem-perature should be below - 160°C.

Attempts to adapt this sort of technique to thesterile storage of pint volumes of blood for trans-fusion met with great difficulties. Special flat or

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

Long-term preservation oJ cells and tissues: a review

corrugated metal cannisters were developed andtechniques were devised for accelerating heattransfer from them into liquid nitrogen duringimmersion cooling (Strumia et al, 1960) but itproved impossible to devise a method that wouldgive satisfactory haemolysis rates and post-trans-fusion survival figures without the use of relativelyhigh concentrations of cryoprotectants. Once thisrestriction had been accepted, and with it the needto remove the cryoprotectant after thawing andbefore transfusion, then the requirement for veryrapid cooling was removed, because increasing thecryoprotectant concentration made much slowercooling perfectly acceptable. Glycerol has been themost widely used cryoprotectant, and since ery-throcytes will withstand very high concentrations ofglycerol, their storage in the presence of 3-4 Mglycerol presents little difficulty, and the coolingrate does not need to be controlled. With such highconcentrations of glycerol the storage temperaturemay even be as high as - 45°C with reasonablyacceptable levels of haemolysis, although - 80°C isreally preferred. The serious technical difficulty withthis technique is the removal of glycerol afterthawing since this must be performed slowly ifserious osmotic imbalance, with consequent haemoly-sis, is to be avoided. Slow stepwise dilution ordialysis against glycerol-free saline are effective butare time-consuming, and sterility is difficult tomaintain. However, if the glycerol technique isbeing used for small volumes of blood intended forserological testing, then dialysis is both simple andconvenient; one hour's dialysis in a 20 cm piece ofVisking tubing immersed in a 200 ml beaker ofphysiological saline is satisfactory for 5 ml samples(Weiner, 1961).The first really practical method of removing

glycerol from pint volumes of blood for transfusioninvolved the use of the Cohn ADL fractionator forcontinuous washing of the cell suspension. Thethawed glycerolized blood was run into the sterilizedbowl revolving at 4000 rev/min, and followed by 4litres of solution of progressively decreasing glycerolcontent, starting at 11 M and falling to zero overabout 90 minutes. The residual glycerol content wastypically 0O05-0-1 M and the post-transfusionrecovery was 85-90% (Pyle, 1964). There wereconsiderable difficulties in the washing and sterilizingof the bowls but this problem was overcome by theintroduction of disposable plastic bowls. Othermanufacturers have now entered the market withsimple continuous-flow centrifuges using disposablebowls, and the problem is now one of cost-typically£18 per wash at the time of writing.An alternative and somewhat simpler method that

avoids the need to centrifuge was introduced by

Huggins (1963). He observed that when erythrocytesare suspended in slightly acid media containing lessthan 0-02 M salts but rendered isotonic with sugars,the cells aggregate to form large clumps that sedi-ment rapidly. When these cells are returned toisotonic saline or plasma they disperse and can berecovered with very little haemolysis. Huggins usedthis phenomenon to remove dimethylsulphoxide orglycerol from thawed red cell suspensions: first theagglomerating sugar solution was run into the cellsuspension, then the suspension was stirred and timewas allowed for the glycerol to diffuse out, the stirringwas stopped, and the cells settled rapidly undergravity. The supernatant was then removed and threewashes of approximately 1 litre each were carriedout in a similar way. Finally, the cells were resuspen-ded in saline. By using special blood bags this processcan be completed in 20 minutes and one techniciancan handle 5 pints simultaneously. The disadvantagesof this method are the large volume of wash solutionrequired, the rather high cost of the special bloodbags (although they are much cheaper than dispos-able centrifuge bowls), and the fact that recoverytends in most hands to be somewhat less efficientthan with centrifugal washing methods.Both of the methods described so far have been

extensively used in the USA but in Europe, andincreasingly in North America also, blood bankshave preferred to use a lower concentration ofglycerol (19 M) although this requires more rapidcooling and storage in liquid nitrogen. Severalvariations of the technique are in use: the methoddescribed by Rowe (1974) is as follows: packed redcells are mixed with an equal volume of 3-8 Mglycerol and frozen rapidly (90°C min-1) inspecial flattened plastic bags. They are stored inliquid nitrogen refrigerators at -196°C. Whenrequired for transfusion, the bag is thawed by agita-tion in a 45°C water bath and washed with 300-500ml of 15% mannitol or 3 5% sodium chloridesolution followed by 2 litres of physiological salineeither in three batches using a four-tail bag and aconventional centrifuge or an automatic serialwashing centrifuge, or continuously using a con-tinuous-flow, disposable-bowl centrifuge. Thismethod gives an in vitro cell recovery of 97 % anda 24-hour in vivo survival of 95 ± 20%.Thus there is a range of techniques available for

the long-term preservation of erythrocytes forlaboratory and clinical use. When liquid nitrogenstorage temperatures are used, there is no perceptibledeterioration during storage, and ATP, 2-3 DPG,and intracellular K+ levels are all fully maintained.

PRESERVATION OF PLATELETSPlatelets are considerably more difficult to preserve

277

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

278

than erythrocytes. Baldini et al (1960) obtained only20% survival of platelets after storage for 24 hoursat - 75°C in the presence of 1 2 M glycerol. Higherconcentrations of glycerol produced lower survivalfigures in their hands, but Cohen and Gardner(1966) found that glycerol toxicity could be greatlyreduced by using extremely gentle methods ofdeglycerolization following thawing; the maximumtolerated concentration then rose to 1-7 M, but theproportional survival was only 23 %. Since plateletsseem to be particularly susceptible to osmoticshock, it is not surprising that DMSO, whichdiffuses much more rapidly than glycerol, shouldhave given somewhat better results. lossifides et al(1963) reported an estimated 30% recovery afterstorage at - 196°C with 2-1 M DMSO in plasma forfive weeks. Djerassi and Roy (1963) found that alower concentration of DMSO (0 7 M) combinedwith 0-28 M dextrose yielded 70% platelet recoveryimmediately after infusion into x-irradiated animals,and 30% survival 24 hours later. Fresh platelets,in their hands, showed a 65% survival at 24 hours,so the preserved platelets were about one half aseffective as the fresh material. The value of the addedglucose is doubtful, however; Murphy et al (1974)found it to be without significant effect, but theyconfirmed the superiority of 0 7 M DMSO overprevious experience with glycerol. Valeri's laboratoryhas now developed the DMSO technique to thepoint where it can be used clinically (Handin andValeri, 1972; Valeri et al, 1974). They have stressedthe desirability of removing DMSO before trans-fusion and, by adding the DMSO slowly and remov-ing it by a stepwise dilution technique, they haveobtained an in vivo survival figure of 47% comparedwith 65% for fresh platelets. The lifespan of theplatelets was normal and there were no adversetransfusion reactions. In this study the cooling ratewas controlled at 1 °C min-' and the storagetemperature was - 150°C, but in a later reportequaly good results were obtained when coolingand storage were achieved simply by placing the 30ml bags in a - 80°C refrigerator; the cooling ratewas then approximately 2-3°C min-'. A full clinicalassessment of this technique is now urgently required.

PRESERVATION OF BONE MARROWBarnes and Loutit (1955) were the first workers toshow that haemopoietic cells from the spleens ofinfant mice would survive storage at - 79°C if theywere cooled at approximately 1 C min-' to - 15°Cand then more rapidly, and if the suspending mediumcontained 2-0 M glycerol. These results have beenconfirmed by many workers using actual bonemarrow cells ofmany species, viability being assessedby the ability to recover lethally irradiated animals.

David E. Pegg

DMSO has also been shown to be effective but therelative efficiency of the two compounds is in somedoubt: Ashwood-Smith (1961) found DMSO to besuperior, whereas Kurnick (1968) found that glycerolwas more effective and van Putten (1965) obtainedidentical results with each compound. For clinicaluse the balance probably tips in favour of glycerolfor two additional reasons: glycerol is a physiologicalcompound that has been widely used for red cellpreservation without any evidence of harmful sideeffects, whereas there are some doubts about DMSO(Smith et al, 1967), and this compound has theundesirable effect of causing the recipients of signifi-cant quantities to acquire a most unpleasant sulphideodour. Other cryoprotectants have been studied inexperimental animals, and PVP has been shown tobe effective in rats (Persidsky et al, 1965). vanPutten (1965) made the interesting observation thata combination of 1-4 M glycerol and 10% PVP ismore effective than either compound separately inthe preservation of monkey bone marrow cells, andthis is probably the technique of choice for clinicaluse. In spite of early worries about the toxicity ofPVP, there seems to be no reason why a low molecu-lar weight fraction such as K-15 should not be usedclinically. Unfortunately the relatively precisespleen-colony assay for stem cell survival cannot beused with human marrow, and since direct testing inlethally irradiated individuals is clearly inapplicable,one is forced to rely on in vitro techniques forquantitative measurements of survival rates forpreserved human bone marrow cells. These are notentirely satisfactory (Pegg, 1964b) although improve-ments in the semisolid culture method of Bradleyand Metcalf (1966) should soon provide morereliable data. In the meantime one is forced to relyon the evidence from studies in experimentalanimals and the results of limited clinical studies(Pegg, 1964a). On this evidence the followingprocedure is suggested:

1 Dilute the bone marrow suspension with anequal volume of balanced salt solution contain-ing 3 M glycerol, possibly with 20% K 15 PVP,and allow 30 minutes at 4°C for equilibration.

2 Cool the suspensions at 1-2°C min-' to - 100°Cand store at - 196°C.

3 Thaw rapidly by agitation in a 40°C water bath.4 Administer the suspension intravenously with-

out prior dilution or manipulation.Attempts have been made to remove the cryoprotec-tant before administration, but this results inunacceptably high cell losses (Pegg, 1964c) and isnotrecommended with presently available techniques.

PRESERVATION OF GRANULOCYTESIt has been known for many years that lethally

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

Long-term preservation of cells and tissues: a review

irradiated animals may be recovered by the infusionof large quantities of peripheral blood leucocytes(Cavins et al, 1964) although it is uncertain whetherthis is due to the presence of adequate numbers ofstem cells in the peripheral blood, or to temporarysupport for a sufficient time to permit naturalrecovery of the irradiated marrow. The clinicalapplication of these observations was difficult untilthe development of cell separators made it relativelyeasy to collect the very large numbers of cellsrequired, and there is now a resurgence of interestin the possibility of using granulocyte transfusion inleucopenic patients with severe bacterial or fungalinfection (Lowenthal and his colleagues, 1975;Vallejos and his colleagues, 1975). Clearly, such aclinical programme would be simplified if it werepossible to store granulocytes for long periods oftime. Unfortunately, however, granulocytes areextremely susceptible to freezing injury, andattempts to preserve them in the presence of DMSOby slow cooling to - 196°C have yielded very lowsurvival rates (Cavins et al, 1965; Cavins et al, 1968).Knight et al (1975) have shown that human granu-locytes are severely damaged by rapid cooling to- 15°C even if they are supercooled, and that thisthermal shock can be prevented by cooling at03°C min-'. However, as soon as freezing occursand the suspension cools again after the evolutionof latent heat, the majority of the cells are lost. Themechanism of this damage is not yet known butmay be due to the sensitivity of lysosomes to highconcentrations of electrolytes (Lee and Allen, 1972).

PRESERVATION OF LYMPHOCYTESLymphocytes resemble bone marrow cells in theirrequirements for satisfactory preservation, andsimilar high recovery rates can be obtained with theoptimum techniques. Both glycerol and DMSOhave been used as cryoprotectants, and it is generallyagreed that DMSO is more effective. Ashwood-Smith (1964) used 2-1 M DMSO and preserved mouselymphocytes by cooling at 4°C min-1 followed bystorage at - 196°C. Pegg (1965) used somewhatslower cooling to preserve human lymphocytes in1-4 M DMSO. The interaction between DMSOconcentration and cooling rate was studied byFarrant et al (1972), who found that the maximumrecovery of CON-A responsive human lymphocyteswas obtained at 0-3°C/min in 1-4 M DMSO and at1°C min-' in 0-7 M DMSO. Thorpe et al (1975)have studied the influence of post-thaw manipula-tions on the recovery of frozen mouse lymphocytes:they found that dilution of the suspension to removeDMSO should be carried out at 25°C rather than at0°C, and that the presence of 10% serum, and theuse of low g forces for centrifugation, were beneficial.

Thus, a convenient technique would be to mix thelymphocyte suspension with an equal volume of1-4 M DMSO in BSS containing 10% serum, cool at0 3°C min-1, and store at - 196°C. When required,the suspensions should be thawed rapidly in a 40°Cwater bath, then placed in a 25°C bath and diluted10-fold using 10% serum in BSS taking about 2minutes to complete the dilution, and finallycentrifuged at 60 g to deposit the cells.The two-step cooling procedure can also be used

to preserve small volumes of lymphocyte suspension(Farrant et al, 1974); 0-2 ml aliquots in 0-7 MDMSO can be cooled rapidly in small glass vials byplacing them in a bath at - 25°C, and after 10minutes the vials are transferred to liquid nitrogenat - 196°C. The cells are thawed and diluted asbefore, and 100% recovery has been reported. Thistechnique is remarkably effective but may turn outto be somewhat limited in its practical applicationsince preliminary results indicate lower recoveryrates when the sample volume is increased to 1 ml.

Applications in Microbiology

Viruses and bacteria differ from the mammaliancells that we have considered so far in that many ofthem, although not all, are remarkably resistant tofreezing injury so that recovery can be obtained evenwithout the help of cryoprotectants. Indeed, manyspecies can be fieeze-dried. Space does not permit ofa comprehensive review of freezing and drying ofmicroorganisms here; for this, reference should bemade to Fry (1966) and to Grieff and Rightsel(1966), but a brief survey, with emphasis on basicprinciples, will be given here.

PRESERVATION OF BACTERIASome bacteria, and most spores, will survivepractically any freezing and thawing procedure;spores, of course, contain relatively little freezablewater so their resistance is hardly surprising. Mostmicrococci, streptococci, and staphylococci arehighly resistant, while the bacillus, lactobacillus, andpseudomonas species are rather more easily damaged.In general, even if the organisms will withstandfreezing and thawing per se, the survival rate fallsoff with storage time, unless the temperature isbelow - 60°C. This characteristic is highly dependenton the composition of the suspending medium(Woodburn and Strong, 1960), sugars in particularoften being protective. It seems most likely that inbacteria, as in mammalian cells, death is caused bythe high salt concentrations produced both internallyand externally by freezing, and that sugars actcolligatively to reduce electrolyte concentrations.When the temperature is much lower, for example

279

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

280

- 196°C, progressive loss during storage is notgenerally observed.Some bacterla exhibit thermal shock, that is,

they can be destroyed by cooling rapidly even

without freezing; this is true of log-phase E. coli(Meynell, 1958) for example. Whether intracellularfreezing is often a significant factor in freezingdamage to bacteria is disputed, but Mazur hasproduced convincing evidence that it is in the caseof rapidly cooled Pasteurella tularensis (Mazur et al,1957), and it seems likely to be of general importance.Both thermal shock and intracellular freezing can,

however, be avoided by slow (-1°C min-1) cooling.The protective effect of sugars has already beenreferred to: 10% glucose or sucrose may be used,but 1-3 M glycerol is even more effective (Postgateand Hunter, 1961): storage should be at -80°C or- 196°C for maximum stability, although the moreresistant organisms will survive well at much highertemperatures, even at - 200 or - 30°C.

It has been mentioned that many bacterial cellswill withstand freeze-drying. In this process water issublimed at a low temperature under reducedpressure. During the freezing phase that precedesdrying, the cells are subjected to the same influencesof rising solute concentration and temperaturechange that have been discussed at length already;steps should be taken to avoid thermal shock andintracellular freezing by cooling slowly, and electro-lyte damage may be minimized by using cryoprotec-tants. Cryoprotective compounds have the additionalproperty of holding a certain amount of water in thespecimen at the end of drying, which is importantbecause it has been shown that optimum storage isobtained when the residual moisture level is around1% (Scott, 1960). The temperature at which dryingis carried out is also of great importance, and if thisis allowed to rise above - 30°C the recovery ratefalls dramatically with many species (Greaves,1956). Presumably this effect is due to exposure ofthe cells to strong electrolyte solutions for longperiods of time, and it is very disturbing to notethat many laboratory freeze-driers in common usedo not permit any control of the drying temperature.It is also important to include in the preservationmedium some inert material that will give body to thefreeze-dried product to facilitate handling andresuspension; one suitable compound is dextran,which has the additional advantage of being cryo-protective. A convenient medium for freeze-dryingbacteria is 7-5 % sucrose, to provide cryoprotectionand maintain a satisfactory residual moisture, 8 '

dextran to produce added bulk, and 2% sodiumglutamate which mops up carbonyl groups in themedium and improves stability during storage(Muggleton, 1960). After cooling slowly (> 1 °C/

David E. Pegg

min) drying is carried out at - 30°C or below to aresidual moisture content of about 1 %, and thesamples are then sealed in the absence of oxygen andstored at room temperature.

PRESERVATION OF VIRUSESViruses vary in their sensitivity to freezing, but manyare highly resistant. For example, members of thepox virus, adenovirus, and papovavirus groups arecommonly and very successfully stored at - 20°C.However, this is probably an unwise choice oftemperature, because a significant liquid phase witha very high solute concentration will exist in mostmedia, and in fact herpes simplex virus has beenshown to survive better at + 40C than at - 20°C;temperatures below - 60°C are much more effective.Similarly, it has been found that members of themyxovirus and arbovirus groups can be stored at- 60'C, particularly if additives such as 50%sucrose or 35% sorbitol are included. If solid carbondioxide refrigerators are used it is important toexclude CO2 from the ampoules. In the case ofstrongly cell-associated viruses, such as herpeszoster, the virus can be preserved by storing infectedcells in medium containing dimethyl sulphoxide. Aconcise review of virus freezing, which gives manyreferences to techniques for individual species, hasbeen provided by Melnick (1965). Lyophilizationcan also be used with many but not all viruses, 20%serum or skimmed milk usually being included inthe suspending medium. Further details of methodsfor individual species can be obtained from thereview by Grieff and Rightsel (1966).

PRESERVATION OF CELLS AND TISSUES FOR

VIRUS ISOLATIONTechniques similar to those used to preserve lympho-cytes or bone marrow cells have been found to beeffective for many tissue culture cell lines, for instanceHeLa and L cells have been stored for five years bycooling at approximately 1°C min-1 in serum with0-65 M glycerol (for L cells) or 2 75 % glycerol (forHeLa cells) and holding at - 79°C (Scherer, 1960).Porterfield and Ashwood-Smith (1962) found 1-4 MDMSO to be superior to 1-4 M glycerol for chickfibroblasts and human embryo lung cells. Thesefindings were confirmed by Dougherty (1962), whofound 1°C min-1 to be the optimum cooling rateand also demonstrated the value of including serumin the medium. Thawing was carried out rapidly byimmersion in a 37°C water bath, and the cryo-protectant was removed by dilution with growthmedium over a period of about 5 minutes. Thismethod is of wide but not universal applicability.Chinese hamster lung fibroblasts gave very lowsurvival rates by this method, but survival rose to

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

Long-term preservation of cells and tissues: a review

60% when they were cooled in 1-25 M glycerol at100°C min-1 (Mazur et al, 1970), and rapid thawingwas vital. Good survival at 10°C min-1 was obtainedwhen 15 % w/v PVP (MW 40 000) was used as thecryoprotectant. This study underlines the importanceof determining the optimum cooling rate and study-ing more than one cryoprotectant when faced witha new cell line to store. More recently, it has beenshown that hamster lung fibroblasts suspended in065 M DMSO can be cooled rapidly to - 26°C,held at that temperature for 10 minutes, and thencooled rapidly to - 196°C. After rapid thawing 75 %survival was obtained (Farrant et al, 1974). Two-step cooling will probably be applicable to other celllines also.Organ cultures are sometimes used for viral

growth and the investigation of antiviral drugs, butthese tissues will not survive storage by the con-ventional '1 °C min-1 + 1 to 2 M DMSO' method.Morris and his colleagues (1973) studied the inter-action of DMSO concentration and cooling ratewith embryonic human trachea and obtainedexcellent recovery, as tested both by ciliary activityand viral growth potential, when the tracheal ringswere cooled at 0 3°C min-' in Eagles mediumcontaining 2% serum and 3-3 M DMSO. It isprobable that other tissues will yield satisfactorypreservation methods if similar systemic studies arecarried out.

PRESERVATION OF PROTOZOAL PARASITESMalarial parasites (Collins and Jeffery, 1963),trypanosomes (Polge and Soltys, 1957), Leishmania(Mieth, 1966), and Trichomonas (McEntegart, 1954)have all been successfully preserved by freezing withvarious cryoprotectants. In many cases, however,the cooling rate has not been controlled, nor theeffect of different cryoprotectants studied. Callowand Farrant (1973) subjected Leishmania tropica tosuch an investigation and found DMSO to besuperior to glycerol, sucrose, and PVP, and showedthat cooling at 1-9°C min-' in 1-5 M DMSO gavemaximal survival.

Applications in Experimental Pathology

I would like to conclude this review with a briefsurvey of cryopreservation techniques that have beendeveloped for a variety of cells and tissues that mightbe of interest to experimental pathologists, andfinally to mention some quite recent work on thepreservation of more complex tissues than those wehave dealt with so far.

PRESERVATION OF TUMOUR CELLSSome experimental tumours can be stored at - 79°Cor - 196°C and retain their transplantability without

the use of any cryoprotectant (Ising, 1960) althoughit is likely that, in some cases at least, this is due topreservation of tumour viruses rather than intactcells (Gye et al, 1949). The addition of glycerol(Hauschka et al, 1959) or DMSO (Greenberg, 1966)has generally yielded more consistent results, butprecise information on optimum conditions andquantitative estimations of survival are available forvery few malignant cells. Farrant et al (1973) havestudied leucocytes from the peripheral blood ofpatients suffering from acute myeloid leukaemia orchronic lymphatic leukaemia; the cells were cooledat various rates in medium containing 5 % DMSO andrecovery was assessed by thymidine incorporation.They found populations of leukaemic cells thatsurvived at rates both faster and slower than thatgiving maximal survival of normal lymphocytes,and each population could be recovered after re-freezing at its own optimum rate. This interestingstudy emphasized once more the importance ofinvestigating the interdependence of the cryo-protectant, its concentration, and the cooling ratefor each cell type, and introduced the possibility ofusing cryobiological techniques to separate sub-populations of cells. More experimental work usingqualitative assays of cell recovery is required in thisfield.

PRESERVATION OF EMBRYOSA recent innovation of considerable potentialsignificance is the development of a technique for thepreservation of early mammalian embryos. Whitting-ham et al (1972) described experiments in which theeffects of suspending medium, cooling rate, finaltemperature, and warming rate were studied onmouse embryos up to the blastocyst stage. It wasfound that maximum recovery was obtained whenthe embryos were cooled at 0-3°C min-1 in 1 MDMSO and rewarmed at 4°C min-.1 Survivalreached 70% with two cell embryos and 20% withearly blastocysts. When they were transferred tofoster mothers 65% of the embryos gave rise to apregnancy and 40% of these produced normal full-term fetuses. The insensitivity to very slow coolingcombined with a high sensitivity to rapid thawing isquite remarkable, and so far is unexplained. It ishowever beyond doubt since the same finding wasrepeated independently in a similar study carried outat the same time by Wilmut (1972). Wilmut andRowson (personal communication) have sub-sequently obtained successful preservation of cowembryos using 2 M DMSO and a cooling rate of0 2°C min-.

PRESERVATION OF SIMPLE TISSUESBillingham and Medawar (1952) showed that rabbit

281

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

282

skin could be preserved if it was first soaked in 1 6 Mglycerol for 1 hour and then cooled slowly to- 150°C. Thawing was carried out rapidly byplunging the frozen skin into Ringer's solution at37°C, and viability was proved by grafting. Lehr etal (1964) found that DMSO could be used withhuman skin, and that a range of cooling rates from0 4 to 8 0°C min-1 were equally effective whereasthawing had to exceed 50°C min-1 for maximumsurvival. Cornea is structurally somewhat similar toskin and can be preserved by similar methods;using 1P8 M DMSO and slow cooling it was import-ant to add the DMSO at 4°C rather than at roomtemperature in order to minimize toxic effects(O'Neill et al, 1967).

In the 1950s Smith and Parkes showed that a widevariety of embryonic tissues could be preserved by'slow' cooling in media containing serum and 2 Mglycerol: this included ovarian granulosa (Smith,1952), anterior pituitary (Smith, 1961c), adrenalcortex (Parkes, 1955), and thyroid (Parkes, 1959). Itwas important to allow sufficient time for the glycerolto permeate the tissue and to handle the tissuefragments carefully after thawing.

PRESERVATION OF MORE COMPLEX TISSUESNone of the techniques described so far has beensuccessful for the preservation of more complextissues, much less of whole organs like heart, liver,and kidneys. The reasons for this failure are obscure,and various possibilities have been discussed in detailelsewhere (Pegg, 1973). However, an interestingtechnique has been developed by Farrant (1965)and Elford and Walter (1972) which may eventuallymake cryopreservation of quite complex tissues areality. Farrant was studying possible mechanisms offreezing injury in smooth muscle tissue when heobserved that greatly improved functional recoverycould be obtained if the concentration of DMSOwas increased during cooling and decreased duringrewarming so that freezing was never allowed totake place. The changes in DMSO concentrationwere made in steps, sufficient time being allowed fordiffusion through the tissue at each step. Elford andWalter (1972) found that function could be improvedstill further if the major anion in the medium had amolecular weight greater than 200 and if the initialpH of the medium, measured in the presence of 7 7 MDMSO, exceeded 8-4. Reversing the relative concen-trations of sodium and potassium in the bathingmedium improved the speedofrecoveryafter rewarm-ing. Maximum recovery was obtained when mediumcontaining 60 mm PIPES (piperazine-NN'-bis-2-ethanesulphonate) and 165 mm K+ was used; theDMSO concentration was increased in five stages at37°C, -7°C, -14°C, -22°C, and -39°C, and

David E. Pegg

cooling was then continued to - 79°C in a finalDMSO concentration of 7-7 M. The steps werereversed during warming and there was no freezingat any time. Contractile response to histamine wasmeasured in an organ bath at 370C, and the transientresponse was as powerful as that of unfrozen muscle,although this was followed by a gradual relaxationwhich was not seen in the controls. Electron micro-scopic appearances were indistinguishable from thoseof fresh muscle. In this technique there is no increasein electrolyte concentration although there arechanges in pH during cooling and, of course, inDMSO concentration. This technique is much morecomplicated to use than any of the more conventionalmethods described previously, but this may beacceptable in special circumstances; it is hoped, forexample, that some success in subzero organpreservation may eventually be obtained in this way.

Equipment for Cryopreservation

Throughout this review emphasis has frequentlybeen laid on the importance of cooling rate, ofstorage temperature, and, during two-step techni-ques, of the intermediate holding temperature.Several sophisticated controlled-rate coolingmachines have been described (Pegg et al, 1973;Hayes et al, 1974) and some are commerciallyavailable; they are convenient to use and provide arecord of the cooling curve actually obtained, whichis a valuable safeguard. All the machines currentlyon the market in the United Kingdom use liquidnitrogen as the refrigerant. Any potential purchasershould however ensure that a cam follower, orsome similar programming device, is used: dif-ferential thermocouple controllers have seriousdrawbacks which are discussed elsewhere (Pegg et al,1973) and their use is not recommended. It is quitepossible however to build simple cooling apparatusin the laboratory, and devices that rely on theimmersion of an insulated cooling vessel in a bathof liquid nitrogen or methylated spirit cooled withsolid carbon dioxide are perfectly practicable.Rather less satisfactory are the insulated hollowplugs designed to produce slower cooling wheninserted into the neck of a liquid nitrogen tank; thecooling rates obtained are quite variable and verydependent on the load of samples being cooled.The constant-temperature bath, typically at

- 25°C, required for two-step cooling is bestprovided by placing a bath of silicone oil or glycerolsolution in an electrical refrigerator set to thattemperature.

Storage of cells is most conveniently and reliablyachieved by the use of liquid nitrogen refrigerators.There is a wide choice of refrigerators on the market

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

Long-term preservation of cells and tissues: a review

offering liquid or gas-phase storage or both, in sizesvarying from some 7 litres to 640 litres capacity, andwith or without elaborate inventory control systems.Very small and unspillable refrigerators are alsoavailable for the transport of frozen samples. Wherehigher temperatures are satisfactory, reliable electri-cal refrigerators operating at - 80°C are available.A list of manufacturers of specialist equipment is

appended.

ReferencesAshwood-Smith, M. J. (1961). Preservation of mouse bonemarrow at -79'C with dimethyl sulphoxide. Nature(Lond.), 190, 1204-1205.

Ashwood-Smith, M. J. (1964). Low temperature preservationof mouse lymphocytes with dimethyl sulphoxide. Blood,23, 494-501.

Baldini, M., Costea, N., and Dameshek, W. (1960). Theviability of stored human platelets. Blood, 16, 1669-1692.

Barnes, D. W. H. and Loutit, J. F. (1955). The radiationrecovery factor: preservation by the Polge-Smith-Parkestechnique. J. nat. Cancer Inst., 15, 901-905.

Berg, van den L. (1959). The effect of addition of sodium andpotassium chloride to the reciprocal system: KH2, P04-Na2, HPO4-H20 on pH and composition during freezing.Arch. Biochem., 84, 305-315.

Berg, van den, L. and Rose D. (1959). Effect of freezing onthe pH and composition of sodium and potassium phos-phate solutions: the reciprocal system KH2PO4-Na2,HPO4-H20. Arch Biochem., 81, 319-329.

Billingham, R. E. and Medawar, P. B. (1952). The freezing,drying and storage of mammalian skin. J. exp. Biol., 29,454-468.

Bradley, T. R. and Metcalf, D. (1966). The growth of mousebone marrow cells in vitro. Aust. J, exp. Biol. med. Sci.,44, 287-299.

Bricka, M. and Bessis, M. (1955). Sur la conservation deserythrocytes par congelation a basse temperature en pr6-sence de polyvinyl-pyrrolidone et de dextran. C.R. Soc.Biol. (Paris), 149, 875-877.

Bronson, W. R. and McGinniss, M. H. (1962). The preserva-tion of human red blood cell agglutinogens in liquidnitrogen: study of a technic suitable for routine bloodbanking. Blood, 20, 478-484.

Callow, L. L. and Farrant, J. (1973). Cryopreservation of thepromastigote form of Leishmania tropica var. Major atdifferent cooling rates. Int. J. Parasit., 3, 77-88.

Cavins, J. A., Djerassi, 1., Aghai, E., and Roy, A. J. (1968).Current methods for the cryopreservation of humanleukocytes (granulocytes). Cryobiology, 5, 60-69.

Cavins, J. A., Djerassi, I., Roy, A. J., and Klein, E. (1965).Preservation of viable human granulocytes at low tempera-tures in dimethyl sulphoxide. Cryobiology, 2, 129-133.

Cavins, J. A., Scheer, S. C., Thomas, E. D., and Ferrebee,J. W. (1964). The recovery of lethally irradiated dogs giveninfusions of autologous leukocytes preserved at -80°C.Blood, 23, 38-43.

Cohen, P. and Gardner, F. H. (1966). Platelet preservation.IV. Preservation of human platelet concentrates bycontrolled slow freezing in a glycerol medium. New Engl.J. Med., 274, 1400-1407.

Collins, W. E. and Jeffery, G. M. (1963). The use of dimethylsulphoxide in the low-temperature frozen preservation ofexperimental malarias. J. Parasit., 49, 524-525.

Daw, A., Farrant, J., and Morris, G. J. (1973). Membraneleakage of solutes after thermal shock or freezing. Cryo-biology, 10, 126-133.

Diller, K. R., Cravalho, E. G., and Huggins, C. E. (1972).Intracellular freezing in biomaterials. Cryobiology, 9,429-440.

Djerassi, I. and Roy, A. (1963). A method for preservationof viable platelets: combined effects ofsugars and dimethyl-sulphoxide. Blood, 22, 703-717.

Dougherty, R. M. (1962). Use of dimethyl sulphoxide forpreservation of tissue culture cells by freezing. Nature(Lond.), 193, 550-552.

Elford, B. C. and Walter, C. A. (1972). Effects of electrolytecomposition and pH on the structure and function ofsmooth muscle cooled to -79'C in unfrozen media.Cryobiology, 9, 82-100.

Farrant, J. (1965). Mechanism of cell damage during freezingand thawing and its prevention. Nature (Lond.), 205,1284-1287.

Farrant, J., Knight, S. C., McGann, L. E., and O'Brien, J.(1974). Optimal recovery of lymphocytes and tissue culturecells following rapid cooling. Nature (Lond.), 249, 452-453.

Farrant, J., Knight, S. C., and Morris, G. J. (1972). Use ofdifferent cooling rates during freezing to separate popula-tions of human peripheral blood lymphocytes. Cryo-biology, 9, 516-525.

Farrant, J., Knight, S. C., O'Brien, J. A., and Morris, G. J.(1973). Selection of leukaemic cell populations by freezingand thawing. Nature (Lond.), 245, 322-323.

Farrant, J. and Morris, G. J. (1973). Thermal shock anddilution shock as the causes of freezing injury. Cryo-biology, 10, 134-140.

Farrant, J. and Woolgar, A. E. (1970). Possible relationshipsbetween the physical properties of solutions and celldamage during freezing. In The Frozen Cell: A CibaFoundation Symposium, edited by G. E. W. Wolstenholmeand M. O'Connor, pp 97-119. Churchill, London.

Farrant, J. and Woolgar, A. E. (1972a). Human red cellsunder hypertonic conditions: a model system for investi-gating freezing damage. 1. Sodium chloride. Cryobiology,9, 9-15.

Farrant, J. and Woolgar, A. E. (1972b). Human red cellsunder hypertonic conditions: a model system for investi-gating freezing damage. 2. Sucrose. Cryobiology, 9,16-21.

Fry, R. M. (1966). Freezing and drying of bacteria. InCryobiology, edited by H. T. Meryman, pp. 665-696.Academic Press, London and New York.

Greaves, R. I. N. (1956). The preservation of bacteria.Canad. J. Microbiol., 2, 365-371.

Greenberg, S. (1966). Use of dimethyl sulfoxide to preventdamage to leukemic cells in freezing. N. Y. St. J. Med., 66,652-654.

Grieff, D. and Rightsel, W. (1966). Freezing and freeze-drying ofviruses. In Cryobiology, edited by H. T. Meryman,pp. 697-728. Academic Press, London and New York.

Gye, W. E., Begg, A. M., Mann, I., and Craigie, J. (1949).The survival of activity of mouse sarcoma tissue afterfreezing and drying. Brit. J. Cancer, 3, 259-267.

Handin, R. I. and Valeri, C. R. (1972). Improved viabilityof previously frozen platelets. Blood, 40, 509-513.

Hauschka, T. S., Mitchell, J. T., and Niederpruem, D. J.(1959). A reliable frozen tissue bank: viability and stabilityof 82 neoplastic and normal cell types after prolongedstorage at - 78°C. Cancer Res., 19, 643-653.

Hayes, A. R., Pegg, D. E., and Kingston, R. E. (1974). Amultirate small-volume cooling machine. Cryobiology, 11,371-377.

Huggins, C. E. (1963). Preservation of blood for transfusionsby freezing with dimethylsulfoxide and a novel washingtechnique. Surgery, 54, 191-194.

283

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

David E. Pegg

Huntsman, R. G., Hurn, B. A. L., and Lehmann, H. (1960).Storage of red cells for blood-grouping after freezing inliquid nitrogen. Brit. med. J., 2, 118.

lossifides, I., Geisler, P., Eichman, M. F., and Tocantins,L. M. (1963). Preservation of the clot-retracting activity ofplatelets by freezing in dimethylsulfoxide and plasma.Transfusion, 3, 167-172.

Ising, U. (1960). Chromosomal pattern in some mouse

ascites tumours after deep freezing. Exp. Cell Res., 19,475-488.

Jellinek, H. H. G. and Fok, S. Y. (1967). Freezing of aqueouspolyvinylpyrrolidone solutions. Kolloidzeitschrift, 220,122-133.

Knight, S. C., Farrant, J., and O'Brien, J. (1975). In defenceof granulocyte preservers. Lancet, 1, 929. (Letter.)

Kurnick, N. B. (1968). The present status of low-temperaturepreservation of bone marrow. Proceedings of the 11thCongress of the International Society of Blood Transfusion,Sydney, 1966 (Bibliotheca haematologica, No. 29), editedby L. Hollander, Part 3, pp. 788-796. Karger, Basle andNew York.

Leaf, A. (1959). Maintenance of concentration gradients andregulation of cell volume. Ann. N. Y. Acad. Sci., 72, 396-404.

Lee, D. and Allen, C. (1972). The effect of increasing con-

centrations of potassium chloride on rat liver lysosomes.Biochem. J., 128, 142P.

Lehr, H. B., Berggren, R. B., Lotke, P. A., and Coriell, L. L.(1964). Permanent survival of preserved skin autografts.Surgery, 56, 742-746.

Leibo, S. P., Farrant, J., Mazur, P., Hanna, M. G., andSmith, L. H. (1969). Effects of freezing on marrow stemcell suspensions: interactions of cooling and warmingrates in the presence of PVP, sucrose or glycerol. Cryo-biology, 6, 315-332.

Lovelock, J. E. (1953a). The haemolysis of human red blood-cells by freezing and thawing. Biochim. biophys. Acta, 10,414-426.

Lovelock, J. E. (1953b). The mechanism of the protectiveaction of glycerol against haemolysis by freezing andthawing. Biochim. biophys. Acta, 11, 28-36.

Lovelock, J. E. (1955). Haemolysis by thermal shock. Brit.J. Haemat., 1, 117-129.

Lowenthal, R. M., Grossman, L., Goldman, J. M., Storring,R. A., Buskard, N. A., Park, D. S., Murphy, B. C.,Spiers, A. S. D., and Galton, D. A. G. (1975). Granu-locyte transfusions in treatment of infections in patientswith acute leukaemia and aplastic anaemia. Lancet, 1,353-358.

Luyet, B. J. and Keane, J., Jr. (1955). A critical temperaturerange apparently characterised by sensitivity of bull semen

to high freezing velocity. Biodynamica, 7, 281-292.Lyons, J. M. (1972). Phase transitions and control of

cellular metabolism at low temperatures. Crvobiology, 9,341-350.

McEntegart, M. G. (1954). Prolonged survival ofTrichomonasvaginalis at - 79°C. Nature (Lond.), 183, 270-271.

Mazur, P. (1963). Kinetics of water loss from cells at sub-zero temperatures and the likelihood of intracellularfreezing. J. gen. Physiol., 47, 347-369.

Mazur, P., Farrant, J., Leibo, S. P., and Chu, E. H. Y. (1969).Survival of hamster tissue culture cells after freezing andthawing: interactions between protective solutes andcooling and warming rates. Cryobiology, 6, 1-9.

Mazur, P., Leibo, S. P., Farrant, J., Chu, E. H. Y., Hanna,M. G., Jr., and Smith, L. H. (1970). Interactions ofcooling rate, warming rate, and protective additive on thesurvival of frozen mammalian cells. In The Frozen Cell:A Ciba Foundation Symposium, edited by G. E. W.Wolstenholme and M. O'Connor, pp. 69-88. Churchill,

London.Mazur, P., Rhian, M. A., and Mahlandt, B. G. (1957).

Survival of Pasteurella tularensis in sugar solutions aftercooling and warming at sub-zero temperatures. J. Bact.,73, 394-397.

Melnick, J. L. (1965). Preservation of viruses by freezing.Fed. Proc., 24, Suppl. 15, S280-S283.

Meryman, H. T. (1968). Modified model for the mechanismof freezing injury in erythrocytes. Nature (Lond.), 218,333-336.

Meryman, H. T. and Kafig, E. (1955). Rapid freezing andthawing of whole blood. Proc. Soc. exp. Biol. N. Y., 90,587-589.

Meynell, G. G. (1958). The effect of sudden chilling onEscherichia coli. J. gen. Microbiol., 19, 380-389.

Mieth, H. (1966). Tiefgefrierkonservierung verschiedenerBlut- und Gewebeprotozoen in flussigem Stickstoff. Z.Tropenmed. und Parasit., 17, 103-108.

Morris, G. J., Buckland, R., Rubenstein, D., and Farrant, J.(1973). Preservation of human tracheal organ cultures at- 196°C. J. med. Microbiol., 6, 123-126.

Muggleton, P. W. (1960). Freeze-drying of bacteria withspecial reference to BCG. In Recent Research in Freezingand Drying, edited by A. S. Parkes and A. U. Smith, pp.229-237. Blackwell, Oxford.

Murphy, S., Sayar, S. N., Abdou, N. L., and Gardner, F. H.(1974). Platelet preservation by freezing. Use of dimethylsulphoxide as cryoprotective agent. Transfusion, 14, 139-144.

Nash, T. (1962). The chemical constitution of compoundswhich protect erythrocytes against freezing damage. J.gen. Physiol., 46, 167-175.

O'Neill, P., Mueller, F. O., and Trevor-Roper, P. D. (1967).On the preservation of cornea at - 196°C for full-thicknesshomografts in man and dog. Brit. J. Ophthal., 51, 13-30.

Parkes, A. S. (1955). Viability of adrenocortical tissuetransplanted after freezing and thawing. Proc. roy. Soc. B.,144, 314-328.

Parkes, A. S. (1959). Functional grafts of fresh or storedthyroid tissue. J. Endocr., 18, xxxv-xxxvi.

Pegg, D. E. (1964a). Freezing of bone marrow for clinical use.Cryobiology, 1, 64-71.

Pegg, D. E. (1964b). In vitro assessment of cell viability inhuman bone marrow preserved at - 79°C. J. appl. Physiol.,19, 123-126.

Pegg, D. E. (I 964c). Cytology of human bone marrowsubjected to prolonged storage at - 79°C. J. appl. Physiol.,19, 301-309.

Pegg, D. E. (1973). Theory and experiments towards sub-zero organ preservation. In Organ Preservation, edited byD. E. Pegg, pp. 108-122. Churchill-Livingstone, Edinburghand London.

Pegg, D. E., Hayes, A. R., and Kingston, R. E. (1973).Cooling equipment for use in cryopreservation. Cpryo-biology, 10, 271-281.

Pegg, P. J. (1965). The preservation of leucocytes for cyto-genetic and cytochemical studies. Brit. J. Haemat., 11,586-591.

Persidsky, M. D., Richards, V., and Leef, J. (1965). Cryo-preservation of bone marrow with low molecular weightpolyvinylpyrrolidone. Cryobiology, 2, 74-78.

Polge, C. and Lovelock, J. E. (1952). Preservation of bullsemen at -79°C. Vet. Rec., 64, 396-397.

Polge, C., Smith, A. U., and Parkes, A. S. (1949). Revivalof spermatozoa after vitrification and dehydration at lowtemperatures. Nature (Lord.), 164, 666.

Polge, C. and Soltys, M. A. (1957). Preservation of trypano-somes in the frozen state. Trans. R. Soc. trop. Med. Hyg.,51, 519-526.

Porterfield, J. S. and Ashwood-Smith, M. J. (1962). Preser-

284

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

Long-term preservation of cells and tissues: a review

vation of cells in tissue culture by glycerol and dimethylsulphoxide. Nature (Lond.), 193, 548-550.

Postgate, J. R. and Hunter, J. R. (1961). On the survival offrozen bacteria. J. gen. Microbiol., 26, 367-378.

Putten, L. M. van (1965). Quantitative aspects of the storageof bone marrow cells for transplantation. Europ. J.Cancer, 1, 15-22.

Pyle, H. M. (1964). Glycerol preservation of red blood cells.Cryobiology, 1, 57-60.

Rowe, A. W. (1974). A 'cookbook' for low glycerol rapidfreezing. Lab. Manage., 20-30.

Scherer, W. F. (1960). Effects of freezing speed and glyceroldiluent on 4-5 year survival of the HeLa and L cells. Exp.Cell Res., 19, 175-176.

Scott, W. J. (1960). In Recent Rescarch in Freezing andDrying, edited by A. S Parkes and A. U. Smith, p. 188.Blackwell, Oxford.

Smith, A. U. (1952). Cultivation of ovarian granulosa cellsafter cooling to very low temperatures. Exp. Cell Res., 3,574-583.

Smith, A. U. (1961a). Biological Effects of Freezing andSupercooling,p. 440. Edward Arnold, London. Williamsand Wilkins, Baltimore.

Smith, A. U. (1961b). Biological Effects of Freezing andSupercooling. Edward Arnold, London. Williams andWilkins, Baltimore.

Smith, A. U. (1961c). Biological Effects of Freezing andSupercooling, p. 157. Edward Arnold, London. Williamsand Wilkins, Baltimore.

Smith, E. R., Hadidian, Z., and Mason, M. M. (1967). Thesingle- and repeated-dose toxicity of dimethyl sulphoxide.Ann. N. Y. Acad. Sci., 141, 96-109.

Strumia, M. M., Colwell, L. S., and Strumia, P. V. (1960).Preservation of blood for transfusion. IV. In vitro recoveryafter freezing and thawing of red cells modified with sugars.V. Post-transfusion of red cells modified with sugars,frozen, and stored in the frozen state. J. Lab. clin. Med.,56, 576-593.

Thorp., P. E., Knight, S. C., and Farrant, J. (1975).Optimal conditions for the preservation of mouse lymphnode cells in liquid nitrogen using cooling rate techniques.Cryobiology, 12, (in press).

Valeri, C. R., Feingold, H., and Marchionni, L. D. (1974)A simple method for freezing human platelets using 6%dimethylsulphoxide and storage at -80°C. Blood, 43,131-136.

Vallejos, C., McCredie, K. B., Bodey, G. P., Hester, J. P.,and Freireich, E. J. (1975). White blood cell transfusionsfor control of infections in neutropenic patients. Trans-fusion, 15, 28-33.

Walter, C. A., Knight, S. C., and Farrant, J. (1975). Ultra-structural appearance of freeze-substituted lymphocytesfrozen by interrupting rapid cooling with a period at-26°C. Cryobiology, 12, 103-109.

Weiner, W. (1961). Reconstitution of frozen red cells. Lancet,1, 1264-1265.

285

Whittingham, D. G., Leibo, S. P., and Mazur, P. (1972).Survival ofmouse embryos frozen to - 196°C and - 269°C.Science, 178, 411-414.

Wilmut, I. (1972). The effect of cooling rate, warming rate,cryoprotective agent, and stage of development onsurvival of mouse embryos during freezing and thawing.Life Sci., 11, 1071-1079.

Woodburn, M. J. and Strong, D. H. (1960). Survival ofSalmonella typhimurium, Staphylococcus aureus, andStreptococcus faecalis frozen in simplified food substrates.Appl. Microbiol., 8, 109-113.

Woolgar, A. E. (1972). A study of the effects of freezing onhuman red blood cells. PhD Thesis. Council for NationalAcademic Awards.

Appendix

SUPPLIERS OF CRYOPRESERVATION EQUIPMENT

Cooling apparatus

Dewar flasks

Electrical refrigerators

Liquid nitrogen refrigeratorsand liquefied gas containers

Freeze driers

Planar Products Ltd,Windmill Rd,Sunbury on Thames,MiddlesexCryoson,UK Agent-Cryotech,PO Box 16,Wallingford, Oxon.Union Carbide (UK) Ltd,8 Grafton St,London WIA 2LRDay-Impex Ltd,Station Works,Earls Colne,Colchester, EssexBoro Labs Ltd,Paices Hill,Aldermaston, Berks.Lec Refrigeration Ltd,Bognor Regis, SussexUnion Carbide (UK) Ltd,8 Grafton St,London WIA 2LRBritish OxygenCryoproducts,Manor Royal,Crawley, SussexEdwards High Vacuum,Manor Royal,Crawley, SussexVertis, UK Agent,Techmation Ltd,58 Edgware Way,Edgware,Middlesex HA8 8JP

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from

and tissues: a review.Long-term preservation of cells

D E Pegg

doi: 10.1136/jcp.29.4.2711976 29: 271-285 J Clin Pathol

http://jcp.bmj.com/content/29/4/271.citationUpdated information and services can be found at:

These include:

serviceEmail alerting

the online article. article. Sign up in the box at the top right corner of Receive free email alerts when new articles cite this

Notes

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on April 9, 2018 - Published by http://jcp.bmj.com/Downloaded from