- Email: [email protected]
Efficient separation of human T lymphocytes from venous blood using PVP-coated colloidal silica particles (Percoll)
Journal o f Immunological Methods, 38 (1980) 43--51 © Elsevier/North-Holland Biomedical Press
43
EFFICIENT SEPARATION OF HUMAN T LYMPHOCYTES FROM VENOUS BLOOD USING PVP-COATED COLLOIDAL SILICA PARTICLES (PERCOLL)
H.E. FEUCHT, M.R. HADAM, Fr. FRANK and G. RIETHM~)LLER 1 Institut fiir Immunologic, Schillerstr. 42, 8 Munich 2, F.R.G.
(Received 17 March 1980, accepted 24 June 1980)
In a comparative study, human peripheral T lymphocytes were separated as E rosettes by density centrifugation through various gradient media. Sheep red blood cells (SRBC) were removed by dissociation of the E rosettes at 37°C with subsequent centrifugation on a similar density gradient prewarmed to 37°C. In particular, gradients made of Ficoll ®Urovison ® were compared with Percoll ® gradients with regard to both separation steps. Using Percoll gradients, a maximal T cell recovery of 75% was obtained, whereas Ficoll separation yielded only 46%. T lymphocytes separated with Percoll exhibited equal viability compared to Ficoll isolated cells and consisted of 98% EAET-RFC. No inhibition of cellular function by Percoll treatment was detected, whereas Ficoll treatment led to an impaired mitogenic response. An inherent mitogenicity of Percoll was not observed. The method described results in considerably shortened centrifugation times due to the low viscosity of the Percoll medium and simultaneously seems to be less harmful to the rather fragile rosettes. Reproducibility was found to depend on careful control of density and osmolarity of the Percoll medium.
INTRODUCTION In l y m p h a t i c organs o f t h e m o u s e , distinct s u b p o p u l a t i o n s o f t h y m u s derived {T) l y m p h o c y t e s , c o m p r i s i n g o n l y a few p e r c e n t o f t h e t o t a l T-cell p o p u l a t i o n have been s h o w n to e x e r t specific i m m u n o r e g u l a t o r y f u n c t i o n s ( C a n t o r and Boyse, 1977). Peripheral b l o o d is the m o s t feasible source o f cells to d e f i n e similar T l y m p h o c y t e s u b p o p u l a t i o n s in man. This source, h o w e v e r , places considerable c o n s t r a i n t s o n non-selective m e t h o d s o f isolation. T h e m o s t w i d e l y applied t e c h n i q u e c o m b i n e s t h e binding o f sheep red b l o o d cells t o h u m a n T l y m p h o c y t e s w i t h t h e s u b s e q u e n t c e n t r i f u g a t i o n o f t h e cell aggregates (E rosettes) t h r o u g h d e n s i t y gradients consisting o f F i c o l l - s o d i u m - m e g l u m i n a m i n o t r i z o a t e solutions ( B ¢ y u m , 1 9 6 8 ; J o n d a l et al., 1 9 7 2 ) . This p r o c e d u r e requires p r o l o n g e d c e n t r i f u g a t i o n t h r o u g h a highly viscous m e d i u m and
1 Supported by the Deutsche Forschungsgemeinschaft, Bonn.
44 results in a considerable loss of T lymphocytes. Removal of red blood cells by lytic agents such as distilled water or a m m o n i u m chloride invariably leads to impairment of l y m p h o c y t e functions (Kay et al., 1977). To overcome some of these disadvantages we have tested a new density gradient medium consisting of polyvinyl pyrrolidone coated silica particles (Percoll ®) (Pertoft, 1969). This material requires only 10 min centrifugation time and allows the removal of dissociated erythrocytes from T lymphocytes at 37 ° C, a temperature at which many T cells are lost in Ficoll gradients. The results reported here show that density gradient centrifugation with Percoll allows rapid separation of T lymphocytes from peripheral blood with high purity and with distinctly improved recovery. MATERIALS AND METHODS
Density gradient media Percoll For the determination of density by refractometry a standard curve was obtained by plotting the densities of various Percoll solutions against refractive indices. Densities were determined by gravimetry in 50 ml measuring flasks. Refractive indices were measured at 25°C in a Bausch and Lomb refractometer. The viscosity of the media was determined in an Ostwald viscosimeter. The osmolarity was measured in a Knauer HalbmikroOsmometer. Preparation o f isotonic Percoll solution. One hundred volumes of undiluted Percoll (Pharmacia Fine Chemicals, Uppsala, Sweden) of refractive index Ril.3525, density 1.1283 g/ml, osmolality 20 mosmoles/kg H20, pH 8.8, were mixed with 9 vol of 10X concentrated Hank's BSS (Gibco BioCult, Glasgow, Scotland), thus obtaining a solution with an osmolarity ranging from 280 to 300 mosmoles/1. Osmolarity was controlled for each experiment. The isotonic Percoll solution had a refractive index of 1,3530 corresponding to a density of 1.1240 g/ml. For cell separation the appropriate densities were obtained by diluting the isotonic Percoll with 0.15 M NaC1. To deplete E rosette forming lymphocytes (E-RFC) from l y m p h o c y t e suspensions, an isotonic Percol] solution was adjusted to a refractive index of 1.3450 corresponding to a density of 1.0714 g/ml (100 vol isotonic Percoll + 75 vol 0.15 M NaC1). The viscosity of this separation medium proved to be 1.641 cP. The centrifugation of dissociated rosettes was carried out at 37°C in isotonic Percoll solution with a refractive index of 1.3462 according to a density of 1.082 g/ml (100 vol isotonic Percoll + 55 vol 0.15 M NaC1). The viscosity was 1.637 cP. The exactly adjusted Percoll solutions were sterilized by filtration through 0.45 pm Millipore filters. The solutions were kept at 4°C up to 2 weeks w i t h o u t observing changes in their physical properties.
45
Ficoll For comparative studies two Ficoll solutions of slightly different densities and different concentrations of Ficoll 400 ° (Pharmacia Fine Chemicals, Uppsala) were prepared. Ficoll I (density 1.081 g/ml) corresponds to a solution with a final concentration of 0.08 g Ficoll 400/ml in 10% (w/v) Urovison ® (Schering AG, Berlin). The density of 1.081 g/ml was controlled by gravimetry and the osmolarity was found to be 295 mosmoles/1. The viscosity of this suspension was 4.12 cP. Ficoll I solution was used for separation of mononuclear cells from whole blood, for E-RFC depletion and for centrifugati0n at 37 ° C. Ficoll II (density 1.080 g/ml) was prepared similarly using only 0.06 g Ficoll 400/ml in 10% (w/v) Urovison. The osmolarity was 290 mosmoles/1, viscosity proved to be 3.79 cP. Ficoll II solution was used for E-RFC depletion and for centrifugation at 37 ° C.
Cell separation Separation of non-adherent lymphocytes (NAL) Heparinized blood from normal human volunteers was diluted 1 : 3 with phosphate-buffered saline (PBS) and layered on Ficoll I. After centrifugation for 30 min at 400 Xg (20°C) mononuclear cells were harvested from the interphase and washed twice at 200 X g for 10 min in PBS. Fifty million cells were resuspended in 10 ml medium RPMI 1640 (Gibco Bio-Cult) supplemented with 50% non-inactivated fetal calf serum (FCS; Gibco) and incubated in 100 mm X 20 mm Petri dishes (Falcon Div. Becton Dickinson, Oxnard, CA) for 60 min at 37°C in 5% CO2 atmosphere. Non-adherent l y m p h o c y t e s (NAL) were harvested by gentle aspiration and washed twice in RPMI medium at 200 X g for 10 min.
E-rosette formation Ten million NAL and 4 X 108 SRBC pretreated with AET (2-aminoethylisothiouronium bromide; Serva Feinbiochemica, Heidelberg) were suspended in 2.4 ml medium supplemented with 60% fetal calf serum absorbed with SRBC. AET treatment of SRBC was performed according to Kaplan and Clark (1974). This lymphocyte-SRBC suspension was incubated for 15 min at 37 °C and centrifuged for 5 min at 100 X g. After further incubation on ice for 1 h rosettes were evaluated in a h e m o c y t o m e t e r . In order to examine the purity of separated T cells, diagnostic E rosette tests were performed in a similar way. Small aliquots consisting of 5 X 10 s T cells were rerosetted either with 2 X 106 EAE w o r with 5 X 106 Euntreate d. Rosettes were kept on ice overnight before counting.
Separation of E-RFC The 100 g cell pellets were gently resuspended in a total volume of 3.5 ml RPMI-FCS. The cell suspension was layered on 4 ml density gradient medium in a 100 mm X 17 mm polystyrol tube (Greiner Labortechnik, Nfir-
46 tingen, G.F.R.). When Percoll was used, centrifugation was carried out at 1000 X g for 10 min, whereas Ficoll gradients had to be centrifuged for 40 min at 400 X g. The interphases depleted of E-RFC (termed 'B' enriched) were collected. Dissociation o f pelleted rosettes at 3 7° C After decanting the supernatant density gradient medium, the E-RFC enriched pellet was washed twice in RPMI 1640 and finally resuspended in 2 ml RPMI 1640 supplemented with 10% FCS. Dissociation was carried out at 37°C with frequent mixing for about 15 min. The dissociated cell suspension was layered on 4 ml of prewarmed density gradients in 100 mm X 17 mm tubes. Centrifugation was as for the separation of E-RFC. T cells were recovered from the gradient interphase. The cells were washed twice in RPMI medium and analysed for purity by E rosette formation. Viability was assessed by trypan blue exclusion. Dissociation of SRBC from T lymphocytes was facilitated by aspirating the cell mixture through a 22-gauge needle at the beginning of the 37°C incubation. When the dissociated rosettes were layered on top of the prewarmed gradient through this fine cannula, clump formation could also be avoided and the rosette-forming cells were distributed as a single cell suspension. Phytohemagglutinin (PHA ) stimulation PHA stimulation was performed in round b o t t o m e d microtiter plates (NUNC Inter Med, Roskilde, Denmark) containing 2 X l 0 s lymphocytes in 150 pl RPMI 1640 supplemented with 10% AB serum, 100 U/ml penicillin and 100 #g/ml streptomycin (Gibco Bio-Cult). Fifty microliters of various phytohemagglutinin P dilutions (Difco Laboratories, Detroit, MI, U.S.A.) were added to give final concentrations of 250 pg/ml, 25 t~g/ml, 2.5 pg/ml and 0.25 pg/ml. After 40 h of incubation at 37°C in 5% CO2 the cells were pulsed with [3H]thymidine at a concentration of 1 ttCi per well for another 12 h period. The cells were harvested with a multiple harvester on glass fiber filters (Sartorius, GSttingen, G.F.R.) Results were expressed as incorporated cpm and as stimulation indices. To simulate separation conditions, 8 X 106 NAL suspended in 200 pl RPMI medium were pre-incubated either in 4 ml Percoll (Ril.3450) or in 4 ml Ficoll I (1.081 g/ml) for 60 min at room temperature and another 15 min at 37 ° C. After 3 washings cells were subjected to PHA stimulation as described. In a series of experiments lymphocytes were stimulated with PHA (25 pg/ml) in the presence of Percoll (Ril.3450) or Ficoll (1.081 g/ml) solutions (final dilutions of Percoll or Ficoll: 25%, 5%, 2.5%, v/v). Intrinsic mitogenic activity of Percoll and Ficoll was tested in controls w i t h o u t PHA. RESULTS To compare their effectiveness in T cell separation the various gradient
47 media were tested in simultaneous experiments on blood samples from 5 donors. As a first parameter, the recoveries of cells were evaluated. Table 1 denotes the mean values from 5 experiments. The total cell recovery, defined as the sum of fraction 'B' enriched after E-RFC depletion plus fraction 'T' after 37°C centrifugation, is given in percent of nonadherent lymphocytes submitted to E rosette formation. As Table 1 shows, cell separation with Percoll provides the highest recovery of total cells 75.9% + 2.9. After gradient centrifugation through Ficoll solutions I and II cell recoveries are markedly reduced: 61.9%-+ 3.9 and 46.1% ± 7.4 respectively (see also Fig. 1). The higher yields after Percoll gradients seem to be due to more efficient T cell recoveries, since the 'B' enriched fractions after Percoll and Ficoll I are nearly identical. Considerable loss of cells occurs during centrifugation through the Ficoll II solution, and with this solution fewer 'B' enriched cells are obtained. By microscopical examination of various samples from this gradient, unrosetted lymphocytes were detected floating at several positions. No lymphocytes were found either in Percoll or in Ficoll I gradients after E rosette centrifugation. After careful resuspension of pelleted rosettes only 3--5% of unrosetted cells could be detected in Percoll or in Ficoll I pellets (rosette values: 96.8% + 0.8 in Percoll versus 94.6% ± 1.2 in Ficoll I). Pellets of E-RFC centrifuged through Ficoll II gradients always contained unrosetted lymphocytes (pellet value: 86.3% ± 4.6 rosettes). As demonstrated in Table 1 and Fig. 2, dissociated rosettes are best separated with Percoll at a higher density of 1.082 g/ml (Ril.3462). Calculating the possible yield of T cells from a sample (= % E-RFC × number of NAL), 65% of T lymphocytes could be recovered from the Percoll gradient, whereas only 46--40% were obtained after Ficoll gradient centrifugation. In recent experiments the mean recovery of 65% of T cells could be increased to 75% by layering the dissociated rosettes on top of the Percoll gradient with a 22-gauge needle. The rosette forming cells are thus distributed as a single cell suspension on the prewarmed Percoll medium. It should be mentioned that after centrifugation of dissociated E rosettes at 37°C on Ficoll gradients discrete cell layers at the interphases were not consistently observed. Moreover, Ficoll-separated T cells were occasionally contaminated by SRBC. In contrast, the discrete Percoll T cell interphases proved to be devoid of contaminating red cells. Microscopical examination of both gradient medium and pellet fraction showed that very few lymphocytes from the interphase entered the Percoll gradient at 37°C and virtually no lymphocytes reached the red cell pellet. During centrifugation at 37°C numerous lymphocytes penetrated the Ficoll gradients, so that unrosetted lymphocytes and undissociated rosettes were found in the red cell pellet.
Purity o f T cells (Table 1) The purity of the isolated T lymphocytes was determined immediately
75.9±2.9 61.9±3.9 46.1±7.4
Percoll FicollI FicollII
a Non-adherent lymphocytes.
Total cell number ( " B " enr. + " T " ) (% of NAL a)
Cell separation via
23.0±3.7 23.3±4.0 12.5±1.3
" B " enriched (% of NAL)
52.9±2.4 37.3±1.8 33.5±7.6
% of NAL
" T " cells
65.6±3.7 46.2±2.3 40.0±8.7
% of expected "T" cells
Ceil recovery after separation on different density gradient media (mean values from 5 experiments).
TABLE 1
97.6±0.4 97.6 ± 0.4 96.6±0.3
% E-RFC (AET)
Purity
94.2±1.3 91.6±1.3 88.3±3.7
% E-RFC (untreated)
Oo
49
°/o ,
100
100-
+ 50
50-
Percoll Ficolll
Ficoll2
t
Percoll Ficolll
Ficoll2
~t
Fig. 1. T o t a l cell r e c o v e r y ( ' B ' e n r i c h e d + ' T ' ) e x p r e s s e d as p e r c e n t a g e of n o n - a d h e r e n t lymphocytes. Fig. 2. R e c o v e r y o f T cells e x p r e s s e d as p e r c e n t a g e o f e x p e c t e d T cells. *T cell yield r o u t i n e l y o b t a i n e d b y l a y e r i n g dissociated E r o s e t t e s o n p r e w a r m e d Percoll g r a d i e n t s t h r o u g h a 22-gauge needle.
after the separation procedure by the EAE w and E u n t r e a t e d rosette test. No significant differences were noted in EAET-RFC between Percoll- and Ficollseparated T cells (97.6% and 96.6% EAET-RFC). The purity in terms of Euntreated-RFC was slightly increased: 94% in Percoll T cells compared to Ficoll T cells, 91% and 88%.
Viability of T cells The viability of cells was assessed by the trypan blue exclusion test. T cells prepared b y Percoll gave viability counts of 98--99%. Ficoll T cells yielded 95--97%.
TABLE 2 [3H]thymidine incorporation by PHA stimulated non-adherent lymphocytes pretreated w i t h e i t h e r Ficoll I or Percoll. Pretreatment a
Thymidine uptake in c p m ( P H A conc. 25 ~ g / m l )
Thymidine uptake in c p m without PHA
Stimulation i n d e x (SI)
None Ficoll I Percoll
231 8 3 3 + 4 6 6 1 165 787 + 6 0 6 7 218 4 6 2 _+ 4 2 9 8
1 1 5 0 + 132 1 1 2 2 + 243 1 1 3 0 _+ 186
201 147 193
a 8 × 106 n o n - a d h e r e n t l y m p h o c y t e s s u s p e n d e d in 200 btl RPMI m e d i u m were i n c u b a t e d e i t h e r in 4 ml Percoll (R i 1 . 3 4 5 0 ) or in 4 m l Ficoll I ( 1 . 0 8 1 g / m l ) for 60 m i n a t r o o m t e m p e r a t u r e a n d a n o t h e r 15 rain at 37°C. A f t e r 3 washings cells were s u b j e c t e d t o P H A stimulation.
50
Separation time Because of the shorter centrifugation time required for Percoll gradients, 1 h can be saved when T cells are separated via this medium. Cells in Ficoll media must be centrifuged for 30 min for E-RFC depletion and at 37°C up to 50 min centrifugation is required to obtain discrete cell bands. Because of the distinctly lower viscosity of Percoll gradients, cells need to be centrifuged for 10 min only.
PHA stimulation PHA stimulation of lymphocytes was performed to assess cell function after treatment with Percoll or Ficoll. At the same time mitogenic effects of these gradient media on lymphocytes were studied. Representative data from one experiment are presented. Non-adherent lymphocytes showed optimal stimulation at a PHA concentration of 25 pg/ml. Preincubation of lymphocytes in Percoll neither reduced nor stimulated thymidine incorporation, whereas preincubation in Ficoll led to a 20% reduction in incorporation of thymidine (Table 2). When non-adherent lymphocytes were stimulated with PHA in the presence of Percoll or Ficoll, it appeared that a medium containing 25% Ficoll led to an impaired mit0genic response. At lower concentrations no differences between Percoll and Ficoll were found. Non-adherent lymphocytes were not stimulated when incubated in the presence of Percoll or Ficoll w i t h o u t PHA (data not shown). DISCUSSION Percoll is a new, modified density gradient medium derived from colloidal silica sols which has been used in recent years for separation of biological materials (Wolff, 1975). It has low osmolality, low viscosity and high density (Pertoft et al., 1978). The PvP coated particles do not seem to penetrate through biological membranes (Pertoft et al., 1977). These advantages may be exploited for cell separation if certain conditions are observed. Since the b u o y a n t density of cells appears to be a function of both the osmolality of the surrounding medium and the real density of the cells (Cercek and Cercek, 1978), Percoll has to be rendered isotonic before use in cell separation. Percoll is supplied as a solution of high density and therefore appropriate lower densities need to be prepared prior to use. Dilution with 0.15 M NaC1 achieves the adjustment to any desired density w i t h o u t changing the osmolarity or increasing the viscosity which would prolong centrifugation times. Since it was not possible to separate human T and B lymphocytes to high purity by mere centrifugation of mononuclear cell suspensions through discontinuous density gradients (Kurnick et al., 1979; Ulmer and Flad, 1979; Feucht, unpublished observation), we employed Percoll as a one-step gra-
51
dient for separation of E rosettes. When the effects of various densities of Percoll solutions on this separation were tested, good correlations were found between the densities of the gradient media and the purities of E-RFC pellets (i.e. at higher densities only compact rosettes reach the pellet). An inverse relation exists between the gradient density and E-RFC recovery (i.e. in gradients of higher densities increasing numbers of rosettes are retained on the interphase). Thus a density of 1.071 g/ml, corresponding to R i l . 3 4 5 0 appeared to be a good compromise with regard to the purity and recovery of E-RFC. Similarly a density of 1.082 g/ml, corresponding to R i l . 3 4 6 2 , was established for centrifugation at 37°C. L y m p h o c y t e s are retained at the interphase without contaminating SRBC. At lower densities lymphocytes enter the gradient medium, whereas at higher densities SRBC will contaminate the interphase. There is some evidence that uncontrollable loss of cells occurs more readily in media of high viscosity than of low viscosity. With respect to E-RFC separation, it seems possible, that in low viscosity medium more rosettes are pelleted without disruption during centrifugation. In addition, it has been shown that fewer cells adhere to centrifuge tube walls after centrifugation in colloidal silica-aluminium modified-PvP density gradients than in Ficoll gradients (Dettman and Wilbur, 1979). This may explain the differences in T cell recoveries between Percoll and Ficoll. In agreement with other investigators we could n o t detect any inhibition of cell functions b y Percoll treatment (Nathanson et al., 1977), in contrast with Ficoll treatment (Kurnick et al., 1979). Inherent mitogenicity of Percoll could be excluded. In conclusion, separation of T cells on Percoll gradients gives higher cell recoveries more rapidly than separation with different Ficoll solutions. REFERENCES B~yum, A., 1968, Scand. J. Clin. Lab. Invest. 21 (Suppl. 97), 97. Cantor, H. and E.A. Boyse, 1977, Immunol. Rev. 3 3 , 1 0 5 . Cercek, L. and B. Cercek, 1978, Br. J. Cancer 38, 163. Dettman, G.L. and S.M. Wilbur, 1979, J. Immunol. Methods 27,205. Jondal, M., G. Holm and H. Wigzell, 1972, J. Exp. Med. 136, 207. Kaplan, M.E. and C. Clark, 1974, J. Immunol. Methods 5 , 1 3 1 . Kay, H.D., G.D. Bonnard, W.H. West and R.B. Herbermann, 1977, J. Immunol. 118, 2058. Kurnick, J.T., L. (}stberg, M. Stegagno, A.K. Kimura, A. C}rn and O. SjSberg, 1979, Scand. J. Immunol. 10, 563. Nathanson, S.D., P.L. Zamfirescu, S.I. Drew and S. Wilbur, 1977, J° Immunol. Methods 18, 225. Pertoft, H., 1969, Exp. Cell Res. 57,338. Pertoft, H., K. Rubin, L. Kjellen, T.C. Laurent and B. Klingeborn, 1977, Exp. Cell Res. 110,449. Pertoft, H., T.C. Laurent, T. Laas and L. Kagedal, 1978, Anal. Biochem. 88, 271. Ulmer, A.J. and H.D. Flad, 1979, J. Immunol. Methods 30, 1. Wolff, D.A., 1975, in: Methods in Cell Biology, Vol. 10, ed. D.M. Prescott (Academic Press, New York) p. 85.
43
EFFICIENT SEPARATION OF HUMAN T LYMPHOCYTES FROM VENOUS BLOOD USING PVP-COATED COLLOIDAL SILICA PARTICLES (PERCOLL)
H.E. FEUCHT, M.R. HADAM, Fr. FRANK and G. RIETHM~)LLER 1 Institut fiir Immunologic, Schillerstr. 42, 8 Munich 2, F.R.G.
(Received 17 March 1980, accepted 24 June 1980)
In a comparative study, human peripheral T lymphocytes were separated as E rosettes by density centrifugation through various gradient media. Sheep red blood cells (SRBC) were removed by dissociation of the E rosettes at 37°C with subsequent centrifugation on a similar density gradient prewarmed to 37°C. In particular, gradients made of Ficoll ®Urovison ® were compared with Percoll ® gradients with regard to both separation steps. Using Percoll gradients, a maximal T cell recovery of 75% was obtained, whereas Ficoll separation yielded only 46%. T lymphocytes separated with Percoll exhibited equal viability compared to Ficoll isolated cells and consisted of 98% EAET-RFC. No inhibition of cellular function by Percoll treatment was detected, whereas Ficoll treatment led to an impaired mitogenic response. An inherent mitogenicity of Percoll was not observed. The method described results in considerably shortened centrifugation times due to the low viscosity of the Percoll medium and simultaneously seems to be less harmful to the rather fragile rosettes. Reproducibility was found to depend on careful control of density and osmolarity of the Percoll medium.
INTRODUCTION In l y m p h a t i c organs o f t h e m o u s e , distinct s u b p o p u l a t i o n s o f t h y m u s derived {T) l y m p h o c y t e s , c o m p r i s i n g o n l y a few p e r c e n t o f t h e t o t a l T-cell p o p u l a t i o n have been s h o w n to e x e r t specific i m m u n o r e g u l a t o r y f u n c t i o n s ( C a n t o r and Boyse, 1977). Peripheral b l o o d is the m o s t feasible source o f cells to d e f i n e similar T l y m p h o c y t e s u b p o p u l a t i o n s in man. This source, h o w e v e r , places considerable c o n s t r a i n t s o n non-selective m e t h o d s o f isolation. T h e m o s t w i d e l y applied t e c h n i q u e c o m b i n e s t h e binding o f sheep red b l o o d cells t o h u m a n T l y m p h o c y t e s w i t h t h e s u b s e q u e n t c e n t r i f u g a t i o n o f t h e cell aggregates (E rosettes) t h r o u g h d e n s i t y gradients consisting o f F i c o l l - s o d i u m - m e g l u m i n a m i n o t r i z o a t e solutions ( B ¢ y u m , 1 9 6 8 ; J o n d a l et al., 1 9 7 2 ) . This p r o c e d u r e requires p r o l o n g e d c e n t r i f u g a t i o n t h r o u g h a highly viscous m e d i u m and
1 Supported by the Deutsche Forschungsgemeinschaft, Bonn.
44 results in a considerable loss of T lymphocytes. Removal of red blood cells by lytic agents such as distilled water or a m m o n i u m chloride invariably leads to impairment of l y m p h o c y t e functions (Kay et al., 1977). To overcome some of these disadvantages we have tested a new density gradient medium consisting of polyvinyl pyrrolidone coated silica particles (Percoll ®) (Pertoft, 1969). This material requires only 10 min centrifugation time and allows the removal of dissociated erythrocytes from T lymphocytes at 37 ° C, a temperature at which many T cells are lost in Ficoll gradients. The results reported here show that density gradient centrifugation with Percoll allows rapid separation of T lymphocytes from peripheral blood with high purity and with distinctly improved recovery. MATERIALS AND METHODS
Density gradient media Percoll For the determination of density by refractometry a standard curve was obtained by plotting the densities of various Percoll solutions against refractive indices. Densities were determined by gravimetry in 50 ml measuring flasks. Refractive indices were measured at 25°C in a Bausch and Lomb refractometer. The viscosity of the media was determined in an Ostwald viscosimeter. The osmolarity was measured in a Knauer HalbmikroOsmometer. Preparation o f isotonic Percoll solution. One hundred volumes of undiluted Percoll (Pharmacia Fine Chemicals, Uppsala, Sweden) of refractive index Ril.3525, density 1.1283 g/ml, osmolality 20 mosmoles/kg H20, pH 8.8, were mixed with 9 vol of 10X concentrated Hank's BSS (Gibco BioCult, Glasgow, Scotland), thus obtaining a solution with an osmolarity ranging from 280 to 300 mosmoles/1. Osmolarity was controlled for each experiment. The isotonic Percoll solution had a refractive index of 1,3530 corresponding to a density of 1.1240 g/ml. For cell separation the appropriate densities were obtained by diluting the isotonic Percoll with 0.15 M NaC1. To deplete E rosette forming lymphocytes (E-RFC) from l y m p h o c y t e suspensions, an isotonic Percol] solution was adjusted to a refractive index of 1.3450 corresponding to a density of 1.0714 g/ml (100 vol isotonic Percoll + 75 vol 0.15 M NaC1). The viscosity of this separation medium proved to be 1.641 cP. The centrifugation of dissociated rosettes was carried out at 37°C in isotonic Percoll solution with a refractive index of 1.3462 according to a density of 1.082 g/ml (100 vol isotonic Percoll + 55 vol 0.15 M NaC1). The viscosity was 1.637 cP. The exactly adjusted Percoll solutions were sterilized by filtration through 0.45 pm Millipore filters. The solutions were kept at 4°C up to 2 weeks w i t h o u t observing changes in their physical properties.
45
Ficoll For comparative studies two Ficoll solutions of slightly different densities and different concentrations of Ficoll 400 ° (Pharmacia Fine Chemicals, Uppsala) were prepared. Ficoll I (density 1.081 g/ml) corresponds to a solution with a final concentration of 0.08 g Ficoll 400/ml in 10% (w/v) Urovison ® (Schering AG, Berlin). The density of 1.081 g/ml was controlled by gravimetry and the osmolarity was found to be 295 mosmoles/1. The viscosity of this suspension was 4.12 cP. Ficoll I solution was used for separation of mononuclear cells from whole blood, for E-RFC depletion and for centrifugati0n at 37 ° C. Ficoll II (density 1.080 g/ml) was prepared similarly using only 0.06 g Ficoll 400/ml in 10% (w/v) Urovison. The osmolarity was 290 mosmoles/1, viscosity proved to be 3.79 cP. Ficoll II solution was used for E-RFC depletion and for centrifugation at 37 ° C.
Cell separation Separation of non-adherent lymphocytes (NAL) Heparinized blood from normal human volunteers was diluted 1 : 3 with phosphate-buffered saline (PBS) and layered on Ficoll I. After centrifugation for 30 min at 400 Xg (20°C) mononuclear cells were harvested from the interphase and washed twice at 200 X g for 10 min in PBS. Fifty million cells were resuspended in 10 ml medium RPMI 1640 (Gibco Bio-Cult) supplemented with 50% non-inactivated fetal calf serum (FCS; Gibco) and incubated in 100 mm X 20 mm Petri dishes (Falcon Div. Becton Dickinson, Oxnard, CA) for 60 min at 37°C in 5% CO2 atmosphere. Non-adherent l y m p h o c y t e s (NAL) were harvested by gentle aspiration and washed twice in RPMI medium at 200 X g for 10 min.
E-rosette formation Ten million NAL and 4 X 108 SRBC pretreated with AET (2-aminoethylisothiouronium bromide; Serva Feinbiochemica, Heidelberg) were suspended in 2.4 ml medium supplemented with 60% fetal calf serum absorbed with SRBC. AET treatment of SRBC was performed according to Kaplan and Clark (1974). This lymphocyte-SRBC suspension was incubated for 15 min at 37 °C and centrifuged for 5 min at 100 X g. After further incubation on ice for 1 h rosettes were evaluated in a h e m o c y t o m e t e r . In order to examine the purity of separated T cells, diagnostic E rosette tests were performed in a similar way. Small aliquots consisting of 5 X 10 s T cells were rerosetted either with 2 X 106 EAE w o r with 5 X 106 Euntreate d. Rosettes were kept on ice overnight before counting.
Separation of E-RFC The 100 g cell pellets were gently resuspended in a total volume of 3.5 ml RPMI-FCS. The cell suspension was layered on 4 ml density gradient medium in a 100 mm X 17 mm polystyrol tube (Greiner Labortechnik, Nfir-
46 tingen, G.F.R.). When Percoll was used, centrifugation was carried out at 1000 X g for 10 min, whereas Ficoll gradients had to be centrifuged for 40 min at 400 X g. The interphases depleted of E-RFC (termed 'B' enriched) were collected. Dissociation o f pelleted rosettes at 3 7° C After decanting the supernatant density gradient medium, the E-RFC enriched pellet was washed twice in RPMI 1640 and finally resuspended in 2 ml RPMI 1640 supplemented with 10% FCS. Dissociation was carried out at 37°C with frequent mixing for about 15 min. The dissociated cell suspension was layered on 4 ml of prewarmed density gradients in 100 mm X 17 mm tubes. Centrifugation was as for the separation of E-RFC. T cells were recovered from the gradient interphase. The cells were washed twice in RPMI medium and analysed for purity by E rosette formation. Viability was assessed by trypan blue exclusion. Dissociation of SRBC from T lymphocytes was facilitated by aspirating the cell mixture through a 22-gauge needle at the beginning of the 37°C incubation. When the dissociated rosettes were layered on top of the prewarmed gradient through this fine cannula, clump formation could also be avoided and the rosette-forming cells were distributed as a single cell suspension. Phytohemagglutinin (PHA ) stimulation PHA stimulation was performed in round b o t t o m e d microtiter plates (NUNC Inter Med, Roskilde, Denmark) containing 2 X l 0 s lymphocytes in 150 pl RPMI 1640 supplemented with 10% AB serum, 100 U/ml penicillin and 100 #g/ml streptomycin (Gibco Bio-Cult). Fifty microliters of various phytohemagglutinin P dilutions (Difco Laboratories, Detroit, MI, U.S.A.) were added to give final concentrations of 250 pg/ml, 25 t~g/ml, 2.5 pg/ml and 0.25 pg/ml. After 40 h of incubation at 37°C in 5% CO2 the cells were pulsed with [3H]thymidine at a concentration of 1 ttCi per well for another 12 h period. The cells were harvested with a multiple harvester on glass fiber filters (Sartorius, GSttingen, G.F.R.) Results were expressed as incorporated cpm and as stimulation indices. To simulate separation conditions, 8 X 106 NAL suspended in 200 pl RPMI medium were pre-incubated either in 4 ml Percoll (Ril.3450) or in 4 ml Ficoll I (1.081 g/ml) for 60 min at room temperature and another 15 min at 37 ° C. After 3 washings cells were subjected to PHA stimulation as described. In a series of experiments lymphocytes were stimulated with PHA (25 pg/ml) in the presence of Percoll (Ril.3450) or Ficoll (1.081 g/ml) solutions (final dilutions of Percoll or Ficoll: 25%, 5%, 2.5%, v/v). Intrinsic mitogenic activity of Percoll and Ficoll was tested in controls w i t h o u t PHA. RESULTS To compare their effectiveness in T cell separation the various gradient
47 media were tested in simultaneous experiments on blood samples from 5 donors. As a first parameter, the recoveries of cells were evaluated. Table 1 denotes the mean values from 5 experiments. The total cell recovery, defined as the sum of fraction 'B' enriched after E-RFC depletion plus fraction 'T' after 37°C centrifugation, is given in percent of nonadherent lymphocytes submitted to E rosette formation. As Table 1 shows, cell separation with Percoll provides the highest recovery of total cells 75.9% + 2.9. After gradient centrifugation through Ficoll solutions I and II cell recoveries are markedly reduced: 61.9%-+ 3.9 and 46.1% ± 7.4 respectively (see also Fig. 1). The higher yields after Percoll gradients seem to be due to more efficient T cell recoveries, since the 'B' enriched fractions after Percoll and Ficoll I are nearly identical. Considerable loss of cells occurs during centrifugation through the Ficoll II solution, and with this solution fewer 'B' enriched cells are obtained. By microscopical examination of various samples from this gradient, unrosetted lymphocytes were detected floating at several positions. No lymphocytes were found either in Percoll or in Ficoll I gradients after E rosette centrifugation. After careful resuspension of pelleted rosettes only 3--5% of unrosetted cells could be detected in Percoll or in Ficoll I pellets (rosette values: 96.8% + 0.8 in Percoll versus 94.6% ± 1.2 in Ficoll I). Pellets of E-RFC centrifuged through Ficoll II gradients always contained unrosetted lymphocytes (pellet value: 86.3% ± 4.6 rosettes). As demonstrated in Table 1 and Fig. 2, dissociated rosettes are best separated with Percoll at a higher density of 1.082 g/ml (Ril.3462). Calculating the possible yield of T cells from a sample (= % E-RFC × number of NAL), 65% of T lymphocytes could be recovered from the Percoll gradient, whereas only 46--40% were obtained after Ficoll gradient centrifugation. In recent experiments the mean recovery of 65% of T cells could be increased to 75% by layering the dissociated rosettes on top of the Percoll gradient with a 22-gauge needle. The rosette forming cells are thus distributed as a single cell suspension on the prewarmed Percoll medium. It should be mentioned that after centrifugation of dissociated E rosettes at 37°C on Ficoll gradients discrete cell layers at the interphases were not consistently observed. Moreover, Ficoll-separated T cells were occasionally contaminated by SRBC. In contrast, the discrete Percoll T cell interphases proved to be devoid of contaminating red cells. Microscopical examination of both gradient medium and pellet fraction showed that very few lymphocytes from the interphase entered the Percoll gradient at 37°C and virtually no lymphocytes reached the red cell pellet. During centrifugation at 37°C numerous lymphocytes penetrated the Ficoll gradients, so that unrosetted lymphocytes and undissociated rosettes were found in the red cell pellet.
Purity o f T cells (Table 1) The purity of the isolated T lymphocytes was determined immediately
75.9±2.9 61.9±3.9 46.1±7.4
Percoll FicollI FicollII
a Non-adherent lymphocytes.
Total cell number ( " B " enr. + " T " ) (% of NAL a)
Cell separation via
23.0±3.7 23.3±4.0 12.5±1.3
" B " enriched (% of NAL)
52.9±2.4 37.3±1.8 33.5±7.6
% of NAL
" T " cells
65.6±3.7 46.2±2.3 40.0±8.7
% of expected "T" cells
Ceil recovery after separation on different density gradient media (mean values from 5 experiments).
TABLE 1
97.6±0.4 97.6 ± 0.4 96.6±0.3
% E-RFC (AET)
Purity
94.2±1.3 91.6±1.3 88.3±3.7
% E-RFC (untreated)
Oo
49
°/o ,
100
100-
+ 50
50-
Percoll Ficolll
Ficoll2
t
Percoll Ficolll
Ficoll2
~t
Fig. 1. T o t a l cell r e c o v e r y ( ' B ' e n r i c h e d + ' T ' ) e x p r e s s e d as p e r c e n t a g e of n o n - a d h e r e n t lymphocytes. Fig. 2. R e c o v e r y o f T cells e x p r e s s e d as p e r c e n t a g e o f e x p e c t e d T cells. *T cell yield r o u t i n e l y o b t a i n e d b y l a y e r i n g dissociated E r o s e t t e s o n p r e w a r m e d Percoll g r a d i e n t s t h r o u g h a 22-gauge needle.
after the separation procedure by the EAE w and E u n t r e a t e d rosette test. No significant differences were noted in EAET-RFC between Percoll- and Ficollseparated T cells (97.6% and 96.6% EAET-RFC). The purity in terms of Euntreated-RFC was slightly increased: 94% in Percoll T cells compared to Ficoll T cells, 91% and 88%.
Viability of T cells The viability of cells was assessed by the trypan blue exclusion test. T cells prepared b y Percoll gave viability counts of 98--99%. Ficoll T cells yielded 95--97%.
TABLE 2 [3H]thymidine incorporation by PHA stimulated non-adherent lymphocytes pretreated w i t h e i t h e r Ficoll I or Percoll. Pretreatment a
Thymidine uptake in c p m ( P H A conc. 25 ~ g / m l )
Thymidine uptake in c p m without PHA
Stimulation i n d e x (SI)
None Ficoll I Percoll
231 8 3 3 + 4 6 6 1 165 787 + 6 0 6 7 218 4 6 2 _+ 4 2 9 8
1 1 5 0 + 132 1 1 2 2 + 243 1 1 3 0 _+ 186
201 147 193
a 8 × 106 n o n - a d h e r e n t l y m p h o c y t e s s u s p e n d e d in 200 btl RPMI m e d i u m were i n c u b a t e d e i t h e r in 4 ml Percoll (R i 1 . 3 4 5 0 ) or in 4 m l Ficoll I ( 1 . 0 8 1 g / m l ) for 60 m i n a t r o o m t e m p e r a t u r e a n d a n o t h e r 15 rain at 37°C. A f t e r 3 washings cells were s u b j e c t e d t o P H A stimulation.
50
Separation time Because of the shorter centrifugation time required for Percoll gradients, 1 h can be saved when T cells are separated via this medium. Cells in Ficoll media must be centrifuged for 30 min for E-RFC depletion and at 37°C up to 50 min centrifugation is required to obtain discrete cell bands. Because of the distinctly lower viscosity of Percoll gradients, cells need to be centrifuged for 10 min only.
PHA stimulation PHA stimulation of lymphocytes was performed to assess cell function after treatment with Percoll or Ficoll. At the same time mitogenic effects of these gradient media on lymphocytes were studied. Representative data from one experiment are presented. Non-adherent lymphocytes showed optimal stimulation at a PHA concentration of 25 pg/ml. Preincubation of lymphocytes in Percoll neither reduced nor stimulated thymidine incorporation, whereas preincubation in Ficoll led to a 20% reduction in incorporation of thymidine (Table 2). When non-adherent lymphocytes were stimulated with PHA in the presence of Percoll or Ficoll, it appeared that a medium containing 25% Ficoll led to an impaired mit0genic response. At lower concentrations no differences between Percoll and Ficoll were found. Non-adherent lymphocytes were not stimulated when incubated in the presence of Percoll or Ficoll w i t h o u t PHA (data not shown). DISCUSSION Percoll is a new, modified density gradient medium derived from colloidal silica sols which has been used in recent years for separation of biological materials (Wolff, 1975). It has low osmolality, low viscosity and high density (Pertoft et al., 1978). The PvP coated particles do not seem to penetrate through biological membranes (Pertoft et al., 1977). These advantages may be exploited for cell separation if certain conditions are observed. Since the b u o y a n t density of cells appears to be a function of both the osmolality of the surrounding medium and the real density of the cells (Cercek and Cercek, 1978), Percoll has to be rendered isotonic before use in cell separation. Percoll is supplied as a solution of high density and therefore appropriate lower densities need to be prepared prior to use. Dilution with 0.15 M NaC1 achieves the adjustment to any desired density w i t h o u t changing the osmolarity or increasing the viscosity which would prolong centrifugation times. Since it was not possible to separate human T and B lymphocytes to high purity by mere centrifugation of mononuclear cell suspensions through discontinuous density gradients (Kurnick et al., 1979; Ulmer and Flad, 1979; Feucht, unpublished observation), we employed Percoll as a one-step gra-
51
dient for separation of E rosettes. When the effects of various densities of Percoll solutions on this separation were tested, good correlations were found between the densities of the gradient media and the purities of E-RFC pellets (i.e. at higher densities only compact rosettes reach the pellet). An inverse relation exists between the gradient density and E-RFC recovery (i.e. in gradients of higher densities increasing numbers of rosettes are retained on the interphase). Thus a density of 1.071 g/ml, corresponding to R i l . 3 4 5 0 appeared to be a good compromise with regard to the purity and recovery of E-RFC. Similarly a density of 1.082 g/ml, corresponding to R i l . 3 4 6 2 , was established for centrifugation at 37°C. L y m p h o c y t e s are retained at the interphase without contaminating SRBC. At lower densities lymphocytes enter the gradient medium, whereas at higher densities SRBC will contaminate the interphase. There is some evidence that uncontrollable loss of cells occurs more readily in media of high viscosity than of low viscosity. With respect to E-RFC separation, it seems possible, that in low viscosity medium more rosettes are pelleted without disruption during centrifugation. In addition, it has been shown that fewer cells adhere to centrifuge tube walls after centrifugation in colloidal silica-aluminium modified-PvP density gradients than in Ficoll gradients (Dettman and Wilbur, 1979). This may explain the differences in T cell recoveries between Percoll and Ficoll. In agreement with other investigators we could n o t detect any inhibition of cell functions b y Percoll treatment (Nathanson et al., 1977), in contrast with Ficoll treatment (Kurnick et al., 1979). Inherent mitogenicity of Percoll could be excluded. In conclusion, separation of T cells on Percoll gradients gives higher cell recoveries more rapidly than separation with different Ficoll solutions. REFERENCES B~yum, A., 1968, Scand. J. Clin. Lab. Invest. 21 (Suppl. 97), 97. Cantor, H. and E.A. Boyse, 1977, Immunol. Rev. 3 3 , 1 0 5 . Cercek, L. and B. Cercek, 1978, Br. J. Cancer 38, 163. Dettman, G.L. and S.M. Wilbur, 1979, J. Immunol. Methods 27,205. Jondal, M., G. Holm and H. Wigzell, 1972, J. Exp. Med. 136, 207. Kaplan, M.E. and C. Clark, 1974, J. Immunol. Methods 5 , 1 3 1 . Kay, H.D., G.D. Bonnard, W.H. West and R.B. Herbermann, 1977, J. Immunol. 118, 2058. Kurnick, J.T., L. (}stberg, M. Stegagno, A.K. Kimura, A. C}rn and O. SjSberg, 1979, Scand. J. Immunol. 10, 563. Nathanson, S.D., P.L. Zamfirescu, S.I. Drew and S. Wilbur, 1977, J° Immunol. Methods 18, 225. Pertoft, H., 1969, Exp. Cell Res. 57,338. Pertoft, H., K. Rubin, L. Kjellen, T.C. Laurent and B. Klingeborn, 1977, Exp. Cell Res. 110,449. Pertoft, H., T.C. Laurent, T. Laas and L. Kagedal, 1978, Anal. Biochem. 88, 271. Ulmer, A.J. and H.D. Flad, 1979, J. Immunol. Methods 30, 1. Wolff, D.A., 1975, in: Methods in Cell Biology, Vol. 10, ed. D.M. Prescott (Academic Press, New York) p. 85.