Clin. exp. Immunol. (1977) 28, 332-340.
Co-operation between T and B lymphocytes from human tonsils in the response to mitogens and antigens S. ROMAGNANI, E. MAGGI, A. AMADORI, M. G. GIUDIZI & N. RICCI Clinical Immunlmi(og Laboratory, Institute of Medical Pathology, University of Florence, Florence, Italy (Received 21 September 1976)
S U M N1 A R Y
Purified B lymphocytes obtained from human tonsil cell populations by removing E rosetteforming cells by density sedimentation did not proliferate at three days in response to PHA and Con A, but showed a significant 3H-labelled thvmidine incorporation when the PHA response was assessed at day 6 of culture. The 6th-day response, which xas completely abolished by the reduction of T-cell contamination to less than 0 1% by re-rosetting and a second separation, was due in part to a direct activation by PHA of contaminating T cells and in part to a T cell-mediated B-cell response. When purified B cells were stimulated for 3 days by PHA in the presence of an equal number of autologous or homologous mitomvcin-treated T lymphocytes a highly significant uptake of 3H-labelled thymidine was demonstrated. The majority of blast cells obtained at day 4 in these cultures were unable to form E rosettes and showed surface immunoglobulin by immunofluorescence stain. This response was markedly decreased by previous treatment of B cells with mitomycin C and it was abolished when B cells were killed by heating at 56 C for 1 hr. Purified B lymphocytes from human tonsils did not respond in vitro when cultured for 6 days in the presence of soluble antigens (PPD and Candida). However, a highly significant response to the same antigens could be demonstrated when B cells were cultured in the presence of autologous mitomycin-treated T cells. These models of T-B co-operation could provide an interesting tool for studying the differentiation and antibody production in vitro of human B lymphocytes.
I NTRODUCTI ON
The studv of immune responses involving distinctive cell types has been greatly facilitated by the development of techniques for separating and purifying different lymphocyte populations (Edelman, Rutishauer & Millette, 1971; Greaves & Brown, 1974; Schlossman & Hudson, 1973; Wigzell, Sundqxvist & Yoshida, 1972). We have approached this problem by the application of a simple and reproducible technique based on the capacity of lymphocytes to form very stable 'spontaneous' rosettes with neuraminidase-treated SRBC (Galili & Schlesinger, 1974). The in vitro functional properties of purified populations of T and B lymphocytes obtained by this separation technique from human tonsils and the cellular collaboration in the proliferative response to mitogens and soluble antigens have been investigated. Correspondence: Dr T\1. Ricci, Cattedra di Immunologia Clinica, Istituto di Patologia Mledica, Universit~i di Firenze, Policlinico di Careggi. 50134 1Firenze, Italy.
332
Co-operation between T and B Iymphocytes
333
MATERIALS AND METHODS Cell separation technique. Human palatin tonsils obtained at tonsillectomy were washed and placed in TC 109 medium (Difco) containing 5% of FCS (Gibco), 100 u/ml of penicillin G, 100 jug/ml of streptomycin, 50 ug/ml of kanamycin and 4 u/ml of heparin. The tonsils were cut into small pieces in an ice bath, the cells dispersed by teasing, filtered through a nylon strainer and layered over a Ficoll-Urovison mixture (sp. gr.tl 077) in glass tubes, followed by centrifugation at 400 g for 30 min at room temperature. The cell layer on the Ficoll-Urovison mixture was collected, washed twice and resuspended at a concentration of 5 x 106/ml in the same medium containing 50% of FCS. FCS was previously absorbed with SRBC and heated at 560C for 30 min. Fresh SRBC were washed six times with medium and incubated for 30 min at 370C with neuraminidase (Behringwerke) (25 u/0 5 x 109 cells/ml). After three washings the SRBC were counted and adjusted at a concentration of 0 5 x 109/ml. One ml of this suspension was mixed with 2 ml of the lymphocyte suspension. The mixture was centrifuged for 5 min at 200 g and incubated in an ice bath for at least 1 hr. It was then gently resuspended and layered over a Ficoll-Urovison mixture in conical glass tubes, followed by centrifugation at 400 g for 30 min at room temperature. The cell layer at interface was collected and the cells washed twice with medium. The cells of the pellet were resuspended with 2 ml of medium, and mixed with 4 vol. of0-87% NH4Cl solution. After 5 min incubation at room temperature the cells were centrifuged and washed twice with medium. In part of the experiments, to remove from the cell layer at interface the contaminating cells forming 'spontaneous' rosettes, a second separation was carried out by re-rosetting the interface cells. Detection of cell-surface markers. For the detection of E rosette-forming cells, lymphocytes (1 x 106/ml) were suspended in 0-2 ml of medium containing 50% of SRBC-absorbed heat-inactivated FCS and mixed with an equal volume of a suspension of neuraminidase-treated SRBC (1 x 108/ml) in 60 x 10 mm plastic tubes. After 10 min incubation at 370C, the cells were centrifuged for 5 min at 200 g and incubated overnight in an ice bath. The cells were then resuspended and counted in a haemocytometer. At least 300 lymphocytes were counted from each replicate tube and only rosettes possessing more than three adhering erythrocytes were considered. Surface immunoglobulin was detected by immunofluorescence, using a polyvalent rabbit antiserum to human immunoglobulin prepared in the author's laboratory. Immunofluorescent staining of lymphocytes by the direct technique was performed by incubating 1 x 106 lymphocytes resuspended in 0-1 ml of medium with 0-1 ml of conjugated antiserum (protein concentration 180 ug/ml) at 4°C for 30 min. Phagocytic cells were detected by neutral red staining, according to the technique described by Arnaiz-Villena, Gyongy6ssy & Playfair (1974). Stimulators. PHA-M (Difco), dissolved with 5 ml of medium, was used at concentrations ranging between 0-1 and 100 ul/ml. PWM (Gibco), dissolved with S ml of medium, was used at concentrations ranging between 0-2 and 200 ,ul/ml. Con A (Pharmacia) dissolved in the same medium was used at concentrations ranging between 0- 1 and 100 ,ug/ml. PPD (Weybridge) was used at concentrations ranging between 0-5 and 50 jug/ml. Candida (Hollister-Stier) was dialysed against medium and used at concentrations ranging between 2 and 200 ul/ml. In vitro cultures. The technique of in vitro cultures and of cell harvesting has been described in detail in previous papers (Romagnani, Biliotti & Ricci, 1975; Romagnani et al., 1976). In brief, the cultures were performed in microtitre plates containing 96 V-bottomed wells; 0-2 ml of the cell suspension (1 x 106 cells/ml) containing 10% of AB pooled serum were added to each well, followed by the addition of either medium alone or medium containing the appropriate concentration of the stimulant. In the experiments in which the collaboration between T and B cells was investigated, 0-1 ml of medium containing the app:opriate concentration of the mitomycin-treated cells were mixed with 0-1 ml of the other cell suspension supplemented by 20% AB serum. The mitomycin treatment consisted of incubating 4 x 106 cells in 2 ml of final medium containing 200 jug mitomycin at 37°C for 30 min. The cells were then washed three times in final medium. All the cultures were performed in triplicate and incubated at 37°C in a humidified atmosphere at 5% carbon dioxide in air for 72-120 hr. Sixteen hours before harvesting 05 ,uCi of methyl-3H-thymidine (Radiochemical Centre, Amersham) in 20 p1 of medium were added to each well. The cultures were then harvested onto glass fibre filters utilizing a multiple automated sample harvester (Skatron, Lierbyen). The filter strips were dried and added to 2 ml of Insta-gel (Packard) in the scintillation vials. The samples were then counted in a Packard 2425 liquid scintillation spectrophotometer.
RESULTS Cell purity and recovery The results of purification experiments of T and B lymphocytes from human tonsils are given in Table 1. In these experiments T cells were identified by E rosetting and B cells by fluorescent stain with anti-Ig. As shown, greater than 97%/O of the 'interface' cell population had Ig on the cell surface as detectable by the direct fluorescent antibody technique, whereas the B-cell proportion of 'pellet' population was reduced to less than 4%/o. Analysis of the E-rosetting properties of these cell populations demonstrated that less than 3%/O of the 'interface' cell population and greater than 92% of the 'pellet' population formed E rosettes. Re-rosetting and a second separation of the 'interface' cells reduced the T-cell contamination to less than 0 1/o. Phagocytic cells, as shown by neutral red staining, were usually 1-4% in
S. Romagnani et ail.
334
TABLE 1. Surface characteristics of lymphocyte preparations from human tonsils Cell population
Unseparated (T B) T cells: single cycle B cells: single cycle B cells: double cycle Values giv en experiments.
E reactivity
Staining with fluorescent anti-Ig
42-6+ 3 1 92 7+ 0 5 2 4+ 0.2 99
represent mean + s.e.
of fifteen different
TABLE 2. Response to mitogens of unseparated and purified T and B lymphocytes after 3 days of culture
Cell population
Stimulant None PHA PWM Con A
T+B
T
B
479+ 105 10,122+ 1092 2842+ 496 4044+ 1359
210+30 9705+ 1112 2399+ 287 2962+400
153+26 453+99 346+ 65 283+79
3H-labelled thymidine incorporation is expressed in c.p.m. + s.e. (eleven experiments).
In
x
,
5
3
4
5
6
Time (days) FIG. 1. PHA response of unseparated, T and B lymphocytes from human tonsils. 3H-labelled thymidine incorporation was assessed on days 3-6 of culture. (-) Unseparated cells; (A) T cells from a single separation cycle; (e), B cells from a single cycle; ((), B cells from a double cycle. B cells obtained after a single cycle were contaminated by 1-3% T cells; T-cell contamination in B cells obtained after a double cycle was lower than 0 1%.
Co-operation between T and B lymphocytes
335
x c
4
0.75
3
50
12.5
200
Number of T cells x 103 FIG. 2. PHA response of untreated (e) and mitomycin-treated (o) highly purified B cells to which different concentrations of autologous T cells were added. Corresponding concentrations of T cells alone were also stimulated by PHA (A). 3H-labelled thymidine incorporation was assessed at day 6.
the 'interface' cell population and less than 1% in the 'pellet' population. The mean cell recovery in twenty-eight consecutive experiments was equal to 57%. The ranges of cell recovery and cell purity with respect to immunoglobulin receptors and E-rosetting properties varied minimally from experiment to experiment. Cell viability of the two populations, as assessed with eosin, was always greater than 90% in all the experiments.
Response to mitogens ofunseparated and purified T and B lymphocytes In these experiments mitogenic activation of unseparated (T+ B) cells was compared to that of the two isolated populations (T or B) measuring 3H-labelled thymidine incorporation. Table 2 illustrates the results obtained using optimal doses of PHA, PWM and Con A to stimulate tonsil lymphocytes for three days. T tonsillar lymphocytes, as well as T+ B cells responded to PHA, PWM and Con A. B lymphocytes were not activated by Con A and gave only a very small response with PHA and PWM. TABLE 3. Reconstitution of the B-proliferative PHA by addition of autologous mitomycin-treated T lymphocytes
response to
Stimulant
Cell population B T Tm Tm+B
Bm+B
None
PHA
132+ 12
274+ 64 8727+ 911 1453+ 287 5408+756 306+ 130
206± 24 68+ 6 273+ 137 210+ 30
3H-labelled thymidine incorporation is expressed in c.p.m. + s.e. (twenty-one experiments). T and B = T and B pure populations from human tonsils; Tm and Bm = T and B cells treated with mitomycin; Tm+B and Bm+B = mixture in equal number of the above mentioned populations.
S. Romagnani et al.
336 8_
T
0 Tm
*Tm+B
6-
0
2
12-5
200 100 25 50 Number of Tm cells x 103 FIG. 3. Reconstitution of the B-cell proliferative response to PHA by addition of increasing concentrations of autologous mitomycin-treated T lymphocytes (Tm). 3H-labelled thymidine incorporation was assessed at day 3. In the cultures containing B cells alone no response was observed.
However, when 3H-labelled thymidine incorporation into the B cell-enriched fraction stimulated with PHA was assessed on days 4-6 of culture a more significant response could be demonstrated in 7 consecutive experiments (Fig. 1). The proliferation peak of B cell-rich populations was usually at day 6 or later, while T cells, as well as unfractionated cells, showed the maximum 3H-labelled thymidine incorporation at the day 3 or 4. When the proportion of T lymphocytes contaminating the 'interface' population was reduced to less than 0.1% by re-rosetting, B cells were not activated by PHA even though the 3H-labelled thymidine incorporation was assessed on days 4 through 6 of culture (Fig. 1).
Influence ofcontaminating T cells on the PHA response of B cell-rich populations To establish whether the PHA response of the 'interface' cell population obtained by a single rosetting, which could be observed at days 4-6, was due to an activation of T contaminating cells (usually 1-3%/o) or it represented a T-mediated B-cell response, increasing concentrations of purified T cells were added to a constant number of untreated and mitomycin-treated B cells (Bm) in which the contamination of T cells was reduced to less than 0*1% by re-rosetting. The results of these experiments are reported
~~~~~~~~~~~~T °E 5 -2
3
4
5
6
Time (days)
FIG. 4. PHA response of B cells in the presence of an equal number of autologous mitomycin-treated T lymphocytes. 31-labelled thymidine incorporation was assessed on days 2-6. (0), B cells obtained after a single separation cycle; (A), mitomycin-treated T cells; (-), B cells+autologous mitomycin-treated lymphocytes.
Co-operation between T and B lymphocytes
337
TABLE 4. Impaired reconstitution of the Tmediated B-cell response to PHA by previous treatment of B lymphocytes with mitomycin or heat
Stimulant
Cell population T B Tm
Tm+B Tm+Bm
Tm+B56 T56+B
None
PHA
209+45 150+37 64+ 4 254+ 112 136+ 24 78+ 6 68+ 4
7325+ 1000 344+145 1030+ 252 5011+ 1129 2044+ 482 1403+ 296 72+ 8
3H-labelled thymidine incorporated is expressed in c.p.m. + s.e. (seven experiments). T and B = T and B pure populations from human tonsils. Tm and Bm = mitomycintreated cells; B56 and T56= cells treated at 560C for 1 hr.
in Fig. 2. It can be observed that the PHA response at day 6 was better in the Bm cultures in the presence of high concentration of T cells, but the cultures containing untreated B cells showed greater PHA response than Bm cultures when a small number of T cells was added. PHA response of B cell-rich populations in the presence of mitomycin-treated T cells (Tm) In twenty-one experiments in which the PHA response of T, B and B + Tm lymphocytes were compared at day 3, it was apparent that T cells and B + Tm responded well, while B lymphocytes did not show a significant response. The response of B + Tm cells stimulated with PHA was significantly greater than that of B+Tm unstimulated lymphocytes and of Tm cells stimulated with PHA in absence of B cells (Table 3). When B cells were stimulated with PHA in the presence of increasing concentrations of autologous Tm cells, a positive relationship between the number of Tm cells added and the degree of TABLE 5. Reconstitution of the B-proliferative response to PHA and Con A by addition of autologous or homologous mitomycintreated T lymphocytes
Stimulant
Cell population B Tm*
B+Tm* Tmt B+Tmt
None
PHA
Con A
137+ 17 68+ 8 206+ 28 72+ 9 609+ 131
887+20 200+ 32 2809+ 730 220+ 37 3106+ 759
402+ 147 103+ 6 1072+ 343 103+ 6 1360+ 375
Mitomycin-treated autologous T cells. t Mitomycin-treated homologous T cells. *
The results represent the mean value+ s.e. of c.p.m. obtained in four experiments. 3H-labelled thymidine incorporation was assessed at day 3.
S. Romagnani et al.
338
TABLE 6. Antigen-induced 3H-labelled thymidine incorporation in purified populations of human T and B cells obtained from tonsils of skin test-positive subjects
Lymphocyte population Exp. no.
Stimulant (dosage)
1
No antigen PPD (10 pg) No antigen l g) PPD (10 No antigen PPD (10pg) No antigen PPD (10 ug) No antigen Candida (20 p1) No antigen Candida (20 p1) No antigen
2
3 4 5
6 7 8
Candida (20 p1) No antigen Candida (20 p1)
T+B
T
B
538 14,188 2167
943 24,875 321 8078 130 13,464 106 1189 912 6278 102 1822 106 751 799 2858
95 205 272 243 110 110 255 111 85 153 140 167 255 220 362 428
12,511 131 8135 269 2387 518 5709 246 877 269 1572 855 3796
3H-labelled thymidine incorporation in triplicate wells containing 0-2 x 106 cells was assessed after 6 days in culture. Results are expressed as mean ct/min. TABLE 7. Reconstitution of the B-proliferative response to PPD and Candida by addition of autologous mitomycin-treated T lymphocytes Stimulant Cell
population B Tm B+Tm
None
PPD
None
Candida
314+ 88 94+ 12 712+23
264+ 88 101+ 34 7105+2969
249+ 101 48+ 8 520+ 192
204+ 61 56+ 6
1140+539
3H-labelled thymidine incorporation is expressed in ct/min+ s.e. (four experiments). Purified B cells (B) from human tonsils of sensitized subjects were stimulated for 6 days with the specific antigen in the absence or in the presence of an equal number of autologous mitomycin-treated T lymphocytes (Tm).
thymidine uptake could be demonstrated (Fig. 3). No significant response was observed when B cells were stimulated with PHA in the presence of T cells killed by heating for 1 hr at 56°C (Table 4). The 3H-labelled thymidine incorporation in the B + Tm cultures stimulated with PHA started at day 2, was maximum at day 3, and decreased in the following days (Fig. 4). It was consistently observed at day 3 after addition to B cells of either autologous or homologous Tm cells (Table 5). The response of B+ Tm cells stimulated with PHA was markedly decreased by the treatment of B cells with mitomycin C and showed the same values of 3H-labelled thymidine incorporation found in the cultures of Tm cells alone, when B cells were killed by heating at 56°C for 1 hr (Table 4). Response to soluble antigens When the cells were taken from the tonsils of sensitized individuals both unfractionated and T cells responded at day 6 to PPD or Candida, whereas B lymphocytes did not respond (Table 6). However, a significant incorporation of 3H-labelled thymidine could be demonstrated at day 6 in cultures contain-
Co-operation between T and B Iymphocytes
339
ing B+ Tm cells in comparison to those containing Tm cells and the antigen or B + Tm cells without the antigen (Table 7). DISCUSSION Highly purified B lymphocytes populations were obtained from human tonsils by a simple and reproducible technique based on the capacity of lymphocytes to form very stable E rosettes with neuraminidase-treated SRBC. The efficacy of the cell separation procedure compared favorably with the more sophisticated immunoadsorbent column methods (Chess, McDermott & Schlossman, 1974a; Wigzell et al., 1972). The cell recovery was satisfactory, the cell populations were viable after separation, remained viable in culture and retained their surface characteristics. The results obtained by the direct assessment of the functional properties of human T and B cells stimulated by mitogens were consistent with those previously ieported by Greaves, Janossy & Doenhoff (1974) and Geha & Merler (1974), but at variance with those of Philips & Roitt (1972) and Chess et al. (1974). In fact, with regard to the capacity of purified B cells to proliferate in vitro in the presence of PHA we showed that if the response was assessed at day 3 a slight or no significant 3H-labelled thymidine incorporation could be noted, whereas it was well demonstrable when the culture time was prolonged to 4-6 days. However, no response could be demonstrated when the B cell-rich population obtained by a single resetting was totally depleted from contaminating T cells by a second separation cycle. Thus, it is clear that the PHA response shown on the days 4-6 by the B cell-rich population obtained by a single rosetting was due to the presence of the few contaminating T lymphocytes (1-3%). It could be suggested that this response was related to a direct activation of contaminating T cells potentiated by non-T cells. An alternative explanation would be that it was a T cell-mediated B-cell response. From our experiments it is reasonably clear that the 3H-labelled thymidine incorporation induced by PHA in cultures containing a small number of contaminating T cells was due either to a direct activation of these cells or to a proliferative response of B cells, facilitated in some way by the presence of a few T lymphocytes. That was particularly shown by the study of the PHA response in cultures containing highly purified B cells to which increasing concentrations of T lymphocytes were added. When more than 5-10% T cells were added to these cultures it was impossible to discriminate between the responses of the two cell populations because the T-cell response widely covered that of B cells. However, the addition of so few T cells as 0.75-3% to untreated B cells was sufficient to restore, particularly at days 5 and 6, a 3H-labelled thymidine incorporation significantly greater than that shown by mitomycin-treated B cells in the presence of the same concentration of T lymphocytes, whereas T cells cultured alone in the same amount did not show any response at this time. Thus, it is clear that the greater PHA response of cultures containing untreated B cells was due to a T cell-mediated B-cell proliferation. On the other hand, the 3H-labelled thymidine uptake found in the cultures containing mitomycin-treated B cells and a small number of T cells was probably due to a potentiating effect of mitomycin-treated B cells on the activation of T cells by PHA, as recently shown by Delespesse et al. (1976), or to a more convenient ratio between the concentration of PHA and the number of cells in culture, taking into account that the same dose of PHA appeared to be toxic for the corresponding amount of T cells alone. For this reason, on the light of our data, caution is advised to interpretate as pure B-cell response those in vitro proliferations of B cell-rich populations which are contaminated also by so few T cells as 1-3%. The experiments performed by culturing B cells in the presence of autologous (or homologous) mitomycin-treated T cells were also consistent with the possibility of a permissive and/or potentiating effect of T cells on the B-cell proliferative response. The study of surface markers of blast cells obtained in these culture systems at day 4 showed the presence of a little number of E rosette-forming blast cells in the cultures in which Tm cells were stimulated with PHA. On the other hand, no blast cells with Ig surface markers were found in the Tm cultures, while the great majority of blast cells from cultures containing B + Tm lymphocytes stimulated with PHA showed Ig receptors on their surface. However, in this experimental model the addition to B cells of at least 25% or more mitomycin-treated T lympho-
340
S. Romnagnani et at.
cytes was necessary to register a significant proliferative response. Similar results were recently reported by Brochier et al. (1976), who were able to restore a proliferative response of B cells to PHA by the addition of 25-50/, irradiated T lymphocytes. Any way it should be noted that in this system, in contrast to that represented by the addition to B cells of a small number of untreated T lymphocytes, the proliferation peak was at day 3 instead of day 6. With regard to the response of purified T- and B-cell populations to soluble antigens our results are in agreement with those previously reported by Chess et al. (1974b). T, as well as unseparated, but not B cells responded in vitro to PPD and Candida at day 6 when the cells were taken from tonsils of individuals sensitized to these antigens. Thus, it is evident that a few contaminating T cells, as those present in a B cell-rich population obtained by a single separation cycle and which were able to allow a proliferative response to PHA at 4-6 days, are not sufficient to induce the response to soluble antigens. However, when unresponding B cells were stimulated by PPD or Candida in the presence of an equal number of mitomycin-treated T cells a very significant 3H-labelled thymidine incorporation could be demonstrated. Thus, it may be suggested that also soluble antigens, as well as mitogens, are capable to induce a T cell-mediated B-cell proliferation. The mechanisms by which T cells allow B cells to respond in vitro to mitogens and antigens is not clarified. The results of Brochier et a!. (1976) argue against the possibility that a soluble factor is responsible for the B-cell response to PHA. On the other hand we showed that living T cells are necessary to allow the activation of B cells, whereas either the activating effect of mitomycin-treated T cells showed by us or that of irradiated T cells demonstrated by Brochier et al. (1976) was significantly lower in comparison to that of untreated T cells. Any way these in vitro models of T-B interaction in the response to mitogens and antigens might provide a useful tool for studying the differentiation and the in vitro antibody production of human B cells. This study was supported in part by funds of the Consiglio di Amministrazione dell'Universita di Firenze, cap. XIb and in part by the CNR (contratto no 7500684/04). The authors are greatful to the colleagues of the Clinica Otorinolaringoiatrica dell'Universita di Firenze for the generous supply of tonsils. REFERENCES BROCHIER, J., SAMARUT, C., GUEHO, J.P. & REVILLARD, J.P. (1976) T-dependence of human B lymphocyte proliferative response to mitogens. It. Arch. Allergy, 51, 101. CHESS, L., MACDERMOTT, R.P. & SCHLOSSMAN, S.F. (1974a) Immunologic functions of isolated human lymphocyte subpopulations. I. Quantitative isolation of human T and B cells and response to mitogens. 5. Immunol. 113, 1113. CHESS, L., MACDERMOTT, R.P. & SCHLOSSMAN, S.F. (1974b) Immunologic functions of isolated human lymphocyte subpopulations. II. Antigen triggering of T and B cells in vitro. 5. Immunol. 113, 1122. DELESPESSE, G., DUCHATEAU, J., GAUSSET, PH. & GOVAERTS, A. (1976) In vitro response of subpopulations of human tonsils lymphocytes. I. Cellular collaboration in the proliferative response to PHA and Con A. Y. Immuiiol. 116, 437. EDELMAN, G.M., RUTISHAUER, U. & MILLETTE, F.C. (1971) Cell fractionation and arrangement on fibers, beads and surfaces. Proc. nat. Acad. Sci. (Wash.), 68, 2153. GALILI, U. & SCHLESINGHER, M. (1974) The formation of stable E rosettes after neuraminidase treatment of either human peripheral blood lymphocytes or of sheep red blood cells. J. Immunol. 112, 1628. GEHA, R.S. & MERLER, E. (1974) Response of human thymus-derived (T) and non-thymus-derived (B) lympho-
cytes to mitogenic stimulation in vitro. Europ. J. Immunol. 4, 193. GREAVES, M.F. & BROWN, G. (1974) Purification of human T and B lymphocytes. 7. Immunol. 112, 420. GREAVES, M.F., JANOSSY, G. & DOENHOFF, M. (1974) Activation of human T and B lymphocytes by polyclonal mitogens. Nature (Lond.), 248, 698. PHILLIPS, B. & RoITT, I.M. (1972) Evidence for transformation of human B lymphocytes by PHA. Nature New Biology, 241, 254. ROMAGNANI, S., BILIOTTI, G. & Ricci, M. (1975) Depression of grass pollen induced lymphocyte transformation by serum from hyposensitized patients. C/in. exp. Immunol. 19, 83. ROMAGNANI, S., AMADORI, A., BITI, G., BELLESI, G. & RIccI, TIN. (1976) Inl vitro lymphocyte response to phytomitogens in untreated and treated patients with Hodgkin's disease. itt. Arch. Allergy, 51, 378. SCHLOSSMAN, S.F. & HUDSON, L. (1973) Specific purification of lymphocyte populations on a digestible immunoabsorbent. J. Immunol. 110, 313. WIGZELL, H., SUNDQVIST, K.G. & YOSHIDA, T.O. (1972) Separation of cells according to surface antigens by the use of antibody-coated columns. Fractionation of cells carrying immunoglobulins and blood group antigens. Scand. 5. Immunol. 1, 75.
Co-operation between T and B lymphocytes from human tonsils in the response to mitogens and antigens S. ROMAGNANI, E. MAGGI, A. AMADORI, M. G. GIUDIZI & N. RICCI Clinical Immunlmi(og Laboratory, Institute of Medical Pathology, University of Florence, Florence, Italy (Received 21 September 1976)
S U M N1 A R Y
Purified B lymphocytes obtained from human tonsil cell populations by removing E rosetteforming cells by density sedimentation did not proliferate at three days in response to PHA and Con A, but showed a significant 3H-labelled thvmidine incorporation when the PHA response was assessed at day 6 of culture. The 6th-day response, which xas completely abolished by the reduction of T-cell contamination to less than 0 1% by re-rosetting and a second separation, was due in part to a direct activation by PHA of contaminating T cells and in part to a T cell-mediated B-cell response. When purified B cells were stimulated for 3 days by PHA in the presence of an equal number of autologous or homologous mitomvcin-treated T lymphocytes a highly significant uptake of 3H-labelled thymidine was demonstrated. The majority of blast cells obtained at day 4 in these cultures were unable to form E rosettes and showed surface immunoglobulin by immunofluorescence stain. This response was markedly decreased by previous treatment of B cells with mitomycin C and it was abolished when B cells were killed by heating at 56 C for 1 hr. Purified B lymphocytes from human tonsils did not respond in vitro when cultured for 6 days in the presence of soluble antigens (PPD and Candida). However, a highly significant response to the same antigens could be demonstrated when B cells were cultured in the presence of autologous mitomycin-treated T cells. These models of T-B co-operation could provide an interesting tool for studying the differentiation and antibody production in vitro of human B lymphocytes.
I NTRODUCTI ON
The studv of immune responses involving distinctive cell types has been greatly facilitated by the development of techniques for separating and purifying different lymphocyte populations (Edelman, Rutishauer & Millette, 1971; Greaves & Brown, 1974; Schlossman & Hudson, 1973; Wigzell, Sundqxvist & Yoshida, 1972). We have approached this problem by the application of a simple and reproducible technique based on the capacity of lymphocytes to form very stable 'spontaneous' rosettes with neuraminidase-treated SRBC (Galili & Schlesinger, 1974). The in vitro functional properties of purified populations of T and B lymphocytes obtained by this separation technique from human tonsils and the cellular collaboration in the proliferative response to mitogens and soluble antigens have been investigated. Correspondence: Dr T\1. Ricci, Cattedra di Immunologia Clinica, Istituto di Patologia Mledica, Universit~i di Firenze, Policlinico di Careggi. 50134 1Firenze, Italy.
332
Co-operation between T and B Iymphocytes
333
MATERIALS AND METHODS Cell separation technique. Human palatin tonsils obtained at tonsillectomy were washed and placed in TC 109 medium (Difco) containing 5% of FCS (Gibco), 100 u/ml of penicillin G, 100 jug/ml of streptomycin, 50 ug/ml of kanamycin and 4 u/ml of heparin. The tonsils were cut into small pieces in an ice bath, the cells dispersed by teasing, filtered through a nylon strainer and layered over a Ficoll-Urovison mixture (sp. gr.tl 077) in glass tubes, followed by centrifugation at 400 g for 30 min at room temperature. The cell layer on the Ficoll-Urovison mixture was collected, washed twice and resuspended at a concentration of 5 x 106/ml in the same medium containing 50% of FCS. FCS was previously absorbed with SRBC and heated at 560C for 30 min. Fresh SRBC were washed six times with medium and incubated for 30 min at 370C with neuraminidase (Behringwerke) (25 u/0 5 x 109 cells/ml). After three washings the SRBC were counted and adjusted at a concentration of 0 5 x 109/ml. One ml of this suspension was mixed with 2 ml of the lymphocyte suspension. The mixture was centrifuged for 5 min at 200 g and incubated in an ice bath for at least 1 hr. It was then gently resuspended and layered over a Ficoll-Urovison mixture in conical glass tubes, followed by centrifugation at 400 g for 30 min at room temperature. The cell layer at interface was collected and the cells washed twice with medium. The cells of the pellet were resuspended with 2 ml of medium, and mixed with 4 vol. of0-87% NH4Cl solution. After 5 min incubation at room temperature the cells were centrifuged and washed twice with medium. In part of the experiments, to remove from the cell layer at interface the contaminating cells forming 'spontaneous' rosettes, a second separation was carried out by re-rosetting the interface cells. Detection of cell-surface markers. For the detection of E rosette-forming cells, lymphocytes (1 x 106/ml) were suspended in 0-2 ml of medium containing 50% of SRBC-absorbed heat-inactivated FCS and mixed with an equal volume of a suspension of neuraminidase-treated SRBC (1 x 108/ml) in 60 x 10 mm plastic tubes. After 10 min incubation at 370C, the cells were centrifuged for 5 min at 200 g and incubated overnight in an ice bath. The cells were then resuspended and counted in a haemocytometer. At least 300 lymphocytes were counted from each replicate tube and only rosettes possessing more than three adhering erythrocytes were considered. Surface immunoglobulin was detected by immunofluorescence, using a polyvalent rabbit antiserum to human immunoglobulin prepared in the author's laboratory. Immunofluorescent staining of lymphocytes by the direct technique was performed by incubating 1 x 106 lymphocytes resuspended in 0-1 ml of medium with 0-1 ml of conjugated antiserum (protein concentration 180 ug/ml) at 4°C for 30 min. Phagocytic cells were detected by neutral red staining, according to the technique described by Arnaiz-Villena, Gyongy6ssy & Playfair (1974). Stimulators. PHA-M (Difco), dissolved with 5 ml of medium, was used at concentrations ranging between 0-1 and 100 ul/ml. PWM (Gibco), dissolved with S ml of medium, was used at concentrations ranging between 0-2 and 200 ,ul/ml. Con A (Pharmacia) dissolved in the same medium was used at concentrations ranging between 0- 1 and 100 ,ug/ml. PPD (Weybridge) was used at concentrations ranging between 0-5 and 50 jug/ml. Candida (Hollister-Stier) was dialysed against medium and used at concentrations ranging between 2 and 200 ul/ml. In vitro cultures. The technique of in vitro cultures and of cell harvesting has been described in detail in previous papers (Romagnani, Biliotti & Ricci, 1975; Romagnani et al., 1976). In brief, the cultures were performed in microtitre plates containing 96 V-bottomed wells; 0-2 ml of the cell suspension (1 x 106 cells/ml) containing 10% of AB pooled serum were added to each well, followed by the addition of either medium alone or medium containing the appropriate concentration of the stimulant. In the experiments in which the collaboration between T and B cells was investigated, 0-1 ml of medium containing the app:opriate concentration of the mitomycin-treated cells were mixed with 0-1 ml of the other cell suspension supplemented by 20% AB serum. The mitomycin treatment consisted of incubating 4 x 106 cells in 2 ml of final medium containing 200 jug mitomycin at 37°C for 30 min. The cells were then washed three times in final medium. All the cultures were performed in triplicate and incubated at 37°C in a humidified atmosphere at 5% carbon dioxide in air for 72-120 hr. Sixteen hours before harvesting 05 ,uCi of methyl-3H-thymidine (Radiochemical Centre, Amersham) in 20 p1 of medium were added to each well. The cultures were then harvested onto glass fibre filters utilizing a multiple automated sample harvester (Skatron, Lierbyen). The filter strips were dried and added to 2 ml of Insta-gel (Packard) in the scintillation vials. The samples were then counted in a Packard 2425 liquid scintillation spectrophotometer.
RESULTS Cell purity and recovery The results of purification experiments of T and B lymphocytes from human tonsils are given in Table 1. In these experiments T cells were identified by E rosetting and B cells by fluorescent stain with anti-Ig. As shown, greater than 97%/O of the 'interface' cell population had Ig on the cell surface as detectable by the direct fluorescent antibody technique, whereas the B-cell proportion of 'pellet' population was reduced to less than 4%/o. Analysis of the E-rosetting properties of these cell populations demonstrated that less than 3%/O of the 'interface' cell population and greater than 92% of the 'pellet' population formed E rosettes. Re-rosetting and a second separation of the 'interface' cells reduced the T-cell contamination to less than 0 1/o. Phagocytic cells, as shown by neutral red staining, were usually 1-4% in
S. Romagnani et ail.
334
TABLE 1. Surface characteristics of lymphocyte preparations from human tonsils Cell population
Unseparated (T B) T cells: single cycle B cells: single cycle B cells: double cycle Values giv en experiments.
E reactivity
Staining with fluorescent anti-Ig
42-6+ 3 1 92 7+ 0 5 2 4+ 0.2 99
represent mean + s.e.
of fifteen different
TABLE 2. Response to mitogens of unseparated and purified T and B lymphocytes after 3 days of culture
Cell population
Stimulant None PHA PWM Con A
T+B
T
B
479+ 105 10,122+ 1092 2842+ 496 4044+ 1359
210+30 9705+ 1112 2399+ 287 2962+400
153+26 453+99 346+ 65 283+79
3H-labelled thymidine incorporation is expressed in c.p.m. + s.e. (eleven experiments).
In
x
,
5
3
4
5
6
Time (days) FIG. 1. PHA response of unseparated, T and B lymphocytes from human tonsils. 3H-labelled thymidine incorporation was assessed on days 3-6 of culture. (-) Unseparated cells; (A) T cells from a single separation cycle; (e), B cells from a single cycle; ((), B cells from a double cycle. B cells obtained after a single cycle were contaminated by 1-3% T cells; T-cell contamination in B cells obtained after a double cycle was lower than 0 1%.
Co-operation between T and B lymphocytes
335
x c
4
0.75
3
50
12.5
200
Number of T cells x 103 FIG. 2. PHA response of untreated (e) and mitomycin-treated (o) highly purified B cells to which different concentrations of autologous T cells were added. Corresponding concentrations of T cells alone were also stimulated by PHA (A). 3H-labelled thymidine incorporation was assessed at day 6.
the 'interface' cell population and less than 1% in the 'pellet' population. The mean cell recovery in twenty-eight consecutive experiments was equal to 57%. The ranges of cell recovery and cell purity with respect to immunoglobulin receptors and E-rosetting properties varied minimally from experiment to experiment. Cell viability of the two populations, as assessed with eosin, was always greater than 90% in all the experiments.
Response to mitogens ofunseparated and purified T and B lymphocytes In these experiments mitogenic activation of unseparated (T+ B) cells was compared to that of the two isolated populations (T or B) measuring 3H-labelled thymidine incorporation. Table 2 illustrates the results obtained using optimal doses of PHA, PWM and Con A to stimulate tonsil lymphocytes for three days. T tonsillar lymphocytes, as well as T+ B cells responded to PHA, PWM and Con A. B lymphocytes were not activated by Con A and gave only a very small response with PHA and PWM. TABLE 3. Reconstitution of the B-proliferative PHA by addition of autologous mitomycin-treated T lymphocytes
response to
Stimulant
Cell population B T Tm Tm+B
Bm+B
None
PHA
132+ 12
274+ 64 8727+ 911 1453+ 287 5408+756 306+ 130
206± 24 68+ 6 273+ 137 210+ 30
3H-labelled thymidine incorporation is expressed in c.p.m. + s.e. (twenty-one experiments). T and B = T and B pure populations from human tonsils; Tm and Bm = T and B cells treated with mitomycin; Tm+B and Bm+B = mixture in equal number of the above mentioned populations.
S. Romagnani et al.
336 8_
T
0 Tm
*Tm+B
6-
0
2
12-5
200 100 25 50 Number of Tm cells x 103 FIG. 3. Reconstitution of the B-cell proliferative response to PHA by addition of increasing concentrations of autologous mitomycin-treated T lymphocytes (Tm). 3H-labelled thymidine incorporation was assessed at day 3. In the cultures containing B cells alone no response was observed.
However, when 3H-labelled thymidine incorporation into the B cell-enriched fraction stimulated with PHA was assessed on days 4-6 of culture a more significant response could be demonstrated in 7 consecutive experiments (Fig. 1). The proliferation peak of B cell-rich populations was usually at day 6 or later, while T cells, as well as unfractionated cells, showed the maximum 3H-labelled thymidine incorporation at the day 3 or 4. When the proportion of T lymphocytes contaminating the 'interface' population was reduced to less than 0.1% by re-rosetting, B cells were not activated by PHA even though the 3H-labelled thymidine incorporation was assessed on days 4 through 6 of culture (Fig. 1).
Influence ofcontaminating T cells on the PHA response of B cell-rich populations To establish whether the PHA response of the 'interface' cell population obtained by a single rosetting, which could be observed at days 4-6, was due to an activation of T contaminating cells (usually 1-3%/o) or it represented a T-mediated B-cell response, increasing concentrations of purified T cells were added to a constant number of untreated and mitomycin-treated B cells (Bm) in which the contamination of T cells was reduced to less than 0*1% by re-rosetting. The results of these experiments are reported
~~~~~~~~~~~~T °E 5 -2
3
4
5
6
Time (days)
FIG. 4. PHA response of B cells in the presence of an equal number of autologous mitomycin-treated T lymphocytes. 31-labelled thymidine incorporation was assessed on days 2-6. (0), B cells obtained after a single separation cycle; (A), mitomycin-treated T cells; (-), B cells+autologous mitomycin-treated lymphocytes.
Co-operation between T and B lymphocytes
337
TABLE 4. Impaired reconstitution of the Tmediated B-cell response to PHA by previous treatment of B lymphocytes with mitomycin or heat
Stimulant
Cell population T B Tm
Tm+B Tm+Bm
Tm+B56 T56+B
None
PHA
209+45 150+37 64+ 4 254+ 112 136+ 24 78+ 6 68+ 4
7325+ 1000 344+145 1030+ 252 5011+ 1129 2044+ 482 1403+ 296 72+ 8
3H-labelled thymidine incorporated is expressed in c.p.m. + s.e. (seven experiments). T and B = T and B pure populations from human tonsils. Tm and Bm = mitomycintreated cells; B56 and T56= cells treated at 560C for 1 hr.
in Fig. 2. It can be observed that the PHA response at day 6 was better in the Bm cultures in the presence of high concentration of T cells, but the cultures containing untreated B cells showed greater PHA response than Bm cultures when a small number of T cells was added. PHA response of B cell-rich populations in the presence of mitomycin-treated T cells (Tm) In twenty-one experiments in which the PHA response of T, B and B + Tm lymphocytes were compared at day 3, it was apparent that T cells and B + Tm responded well, while B lymphocytes did not show a significant response. The response of B + Tm cells stimulated with PHA was significantly greater than that of B+Tm unstimulated lymphocytes and of Tm cells stimulated with PHA in absence of B cells (Table 3). When B cells were stimulated with PHA in the presence of increasing concentrations of autologous Tm cells, a positive relationship between the number of Tm cells added and the degree of TABLE 5. Reconstitution of the B-proliferative response to PHA and Con A by addition of autologous or homologous mitomycintreated T lymphocytes
Stimulant
Cell population B Tm*
B+Tm* Tmt B+Tmt
None
PHA
Con A
137+ 17 68+ 8 206+ 28 72+ 9 609+ 131
887+20 200+ 32 2809+ 730 220+ 37 3106+ 759
402+ 147 103+ 6 1072+ 343 103+ 6 1360+ 375
Mitomycin-treated autologous T cells. t Mitomycin-treated homologous T cells. *
The results represent the mean value+ s.e. of c.p.m. obtained in four experiments. 3H-labelled thymidine incorporation was assessed at day 3.
S. Romagnani et al.
338
TABLE 6. Antigen-induced 3H-labelled thymidine incorporation in purified populations of human T and B cells obtained from tonsils of skin test-positive subjects
Lymphocyte population Exp. no.
Stimulant (dosage)
1
No antigen PPD (10 pg) No antigen l g) PPD (10 No antigen PPD (10pg) No antigen PPD (10 ug) No antigen Candida (20 p1) No antigen Candida (20 p1) No antigen
2
3 4 5
6 7 8
Candida (20 p1) No antigen Candida (20 p1)
T+B
T
B
538 14,188 2167
943 24,875 321 8078 130 13,464 106 1189 912 6278 102 1822 106 751 799 2858
95 205 272 243 110 110 255 111 85 153 140 167 255 220 362 428
12,511 131 8135 269 2387 518 5709 246 877 269 1572 855 3796
3H-labelled thymidine incorporation in triplicate wells containing 0-2 x 106 cells was assessed after 6 days in culture. Results are expressed as mean ct/min. TABLE 7. Reconstitution of the B-proliferative response to PPD and Candida by addition of autologous mitomycin-treated T lymphocytes Stimulant Cell
population B Tm B+Tm
None
PPD
None
Candida
314+ 88 94+ 12 712+23
264+ 88 101+ 34 7105+2969
249+ 101 48+ 8 520+ 192
204+ 61 56+ 6
1140+539
3H-labelled thymidine incorporation is expressed in ct/min+ s.e. (four experiments). Purified B cells (B) from human tonsils of sensitized subjects were stimulated for 6 days with the specific antigen in the absence or in the presence of an equal number of autologous mitomycin-treated T lymphocytes (Tm).
thymidine uptake could be demonstrated (Fig. 3). No significant response was observed when B cells were stimulated with PHA in the presence of T cells killed by heating for 1 hr at 56°C (Table 4). The 3H-labelled thymidine incorporation in the B + Tm cultures stimulated with PHA started at day 2, was maximum at day 3, and decreased in the following days (Fig. 4). It was consistently observed at day 3 after addition to B cells of either autologous or homologous Tm cells (Table 5). The response of B+ Tm cells stimulated with PHA was markedly decreased by the treatment of B cells with mitomycin C and showed the same values of 3H-labelled thymidine incorporation found in the cultures of Tm cells alone, when B cells were killed by heating at 56°C for 1 hr (Table 4). Response to soluble antigens When the cells were taken from the tonsils of sensitized individuals both unfractionated and T cells responded at day 6 to PPD or Candida, whereas B lymphocytes did not respond (Table 6). However, a significant incorporation of 3H-labelled thymidine could be demonstrated at day 6 in cultures contain-
Co-operation between T and B Iymphocytes
339
ing B+ Tm cells in comparison to those containing Tm cells and the antigen or B + Tm cells without the antigen (Table 7). DISCUSSION Highly purified B lymphocytes populations were obtained from human tonsils by a simple and reproducible technique based on the capacity of lymphocytes to form very stable E rosettes with neuraminidase-treated SRBC. The efficacy of the cell separation procedure compared favorably with the more sophisticated immunoadsorbent column methods (Chess, McDermott & Schlossman, 1974a; Wigzell et al., 1972). The cell recovery was satisfactory, the cell populations were viable after separation, remained viable in culture and retained their surface characteristics. The results obtained by the direct assessment of the functional properties of human T and B cells stimulated by mitogens were consistent with those previously ieported by Greaves, Janossy & Doenhoff (1974) and Geha & Merler (1974), but at variance with those of Philips & Roitt (1972) and Chess et al. (1974). In fact, with regard to the capacity of purified B cells to proliferate in vitro in the presence of PHA we showed that if the response was assessed at day 3 a slight or no significant 3H-labelled thymidine incorporation could be noted, whereas it was well demonstrable when the culture time was prolonged to 4-6 days. However, no response could be demonstrated when the B cell-rich population obtained by a single resetting was totally depleted from contaminating T cells by a second separation cycle. Thus, it is clear that the PHA response shown on the days 4-6 by the B cell-rich population obtained by a single rosetting was due to the presence of the few contaminating T lymphocytes (1-3%). It could be suggested that this response was related to a direct activation of contaminating T cells potentiated by non-T cells. An alternative explanation would be that it was a T cell-mediated B-cell response. From our experiments it is reasonably clear that the 3H-labelled thymidine incorporation induced by PHA in cultures containing a small number of contaminating T cells was due either to a direct activation of these cells or to a proliferative response of B cells, facilitated in some way by the presence of a few T lymphocytes. That was particularly shown by the study of the PHA response in cultures containing highly purified B cells to which increasing concentrations of T lymphocytes were added. When more than 5-10% T cells were added to these cultures it was impossible to discriminate between the responses of the two cell populations because the T-cell response widely covered that of B cells. However, the addition of so few T cells as 0.75-3% to untreated B cells was sufficient to restore, particularly at days 5 and 6, a 3H-labelled thymidine incorporation significantly greater than that shown by mitomycin-treated B cells in the presence of the same concentration of T lymphocytes, whereas T cells cultured alone in the same amount did not show any response at this time. Thus, it is clear that the greater PHA response of cultures containing untreated B cells was due to a T cell-mediated B-cell proliferation. On the other hand, the 3H-labelled thymidine uptake found in the cultures containing mitomycin-treated B cells and a small number of T cells was probably due to a potentiating effect of mitomycin-treated B cells on the activation of T cells by PHA, as recently shown by Delespesse et al. (1976), or to a more convenient ratio between the concentration of PHA and the number of cells in culture, taking into account that the same dose of PHA appeared to be toxic for the corresponding amount of T cells alone. For this reason, on the light of our data, caution is advised to interpretate as pure B-cell response those in vitro proliferations of B cell-rich populations which are contaminated also by so few T cells as 1-3%. The experiments performed by culturing B cells in the presence of autologous (or homologous) mitomycin-treated T cells were also consistent with the possibility of a permissive and/or potentiating effect of T cells on the B-cell proliferative response. The study of surface markers of blast cells obtained in these culture systems at day 4 showed the presence of a little number of E rosette-forming blast cells in the cultures in which Tm cells were stimulated with PHA. On the other hand, no blast cells with Ig surface markers were found in the Tm cultures, while the great majority of blast cells from cultures containing B + Tm lymphocytes stimulated with PHA showed Ig receptors on their surface. However, in this experimental model the addition to B cells of at least 25% or more mitomycin-treated T lympho-
340
S. Romnagnani et at.
cytes was necessary to register a significant proliferative response. Similar results were recently reported by Brochier et al. (1976), who were able to restore a proliferative response of B cells to PHA by the addition of 25-50/, irradiated T lymphocytes. Any way it should be noted that in this system, in contrast to that represented by the addition to B cells of a small number of untreated T lymphocytes, the proliferation peak was at day 3 instead of day 6. With regard to the response of purified T- and B-cell populations to soluble antigens our results are in agreement with those previously reported by Chess et al. (1974b). T, as well as unseparated, but not B cells responded in vitro to PPD and Candida at day 6 when the cells were taken from tonsils of individuals sensitized to these antigens. Thus, it is evident that a few contaminating T cells, as those present in a B cell-rich population obtained by a single separation cycle and which were able to allow a proliferative response to PHA at 4-6 days, are not sufficient to induce the response to soluble antigens. However, when unresponding B cells were stimulated by PPD or Candida in the presence of an equal number of mitomycin-treated T cells a very significant 3H-labelled thymidine incorporation could be demonstrated. Thus, it may be suggested that also soluble antigens, as well as mitogens, are capable to induce a T cell-mediated B-cell proliferation. The mechanisms by which T cells allow B cells to respond in vitro to mitogens and antigens is not clarified. The results of Brochier et a!. (1976) argue against the possibility that a soluble factor is responsible for the B-cell response to PHA. On the other hand we showed that living T cells are necessary to allow the activation of B cells, whereas either the activating effect of mitomycin-treated T cells showed by us or that of irradiated T cells demonstrated by Brochier et al. (1976) was significantly lower in comparison to that of untreated T cells. Any way these in vitro models of T-B interaction in the response to mitogens and antigens might provide a useful tool for studying the differentiation and the in vitro antibody production of human B cells. This study was supported in part by funds of the Consiglio di Amministrazione dell'Universita di Firenze, cap. XIb and in part by the CNR (contratto no 7500684/04). The authors are greatful to the colleagues of the Clinica Otorinolaringoiatrica dell'Universita di Firenze for the generous supply of tonsils. REFERENCES BROCHIER, J., SAMARUT, C., GUEHO, J.P. & REVILLARD, J.P. (1976) T-dependence of human B lymphocyte proliferative response to mitogens. It. Arch. Allergy, 51, 101. CHESS, L., MACDERMOTT, R.P. & SCHLOSSMAN, S.F. (1974a) Immunologic functions of isolated human lymphocyte subpopulations. I. Quantitative isolation of human T and B cells and response to mitogens. 5. Immunol. 113, 1113. CHESS, L., MACDERMOTT, R.P. & SCHLOSSMAN, S.F. (1974b) Immunologic functions of isolated human lymphocyte subpopulations. II. Antigen triggering of T and B cells in vitro. 5. Immunol. 113, 1122. DELESPESSE, G., DUCHATEAU, J., GAUSSET, PH. & GOVAERTS, A. (1976) In vitro response of subpopulations of human tonsils lymphocytes. I. Cellular collaboration in the proliferative response to PHA and Con A. Y. Immuiiol. 116, 437. EDELMAN, G.M., RUTISHAUER, U. & MILLETTE, F.C. (1971) Cell fractionation and arrangement on fibers, beads and surfaces. Proc. nat. Acad. Sci. (Wash.), 68, 2153. GALILI, U. & SCHLESINGHER, M. (1974) The formation of stable E rosettes after neuraminidase treatment of either human peripheral blood lymphocytes or of sheep red blood cells. J. Immunol. 112, 1628. GEHA, R.S. & MERLER, E. (1974) Response of human thymus-derived (T) and non-thymus-derived (B) lympho-
cytes to mitogenic stimulation in vitro. Europ. J. Immunol. 4, 193. GREAVES, M.F. & BROWN, G. (1974) Purification of human T and B lymphocytes. 7. Immunol. 112, 420. GREAVES, M.F., JANOSSY, G. & DOENHOFF, M. (1974) Activation of human T and B lymphocytes by polyclonal mitogens. Nature (Lond.), 248, 698. PHILLIPS, B. & RoITT, I.M. (1972) Evidence for transformation of human B lymphocytes by PHA. Nature New Biology, 241, 254. ROMAGNANI, S., BILIOTTI, G. & Ricci, M. (1975) Depression of grass pollen induced lymphocyte transformation by serum from hyposensitized patients. C/in. exp. Immunol. 19, 83. ROMAGNANI, S., AMADORI, A., BITI, G., BELLESI, G. & RIccI, TIN. (1976) Inl vitro lymphocyte response to phytomitogens in untreated and treated patients with Hodgkin's disease. itt. Arch. Allergy, 51, 378. SCHLOSSMAN, S.F. & HUDSON, L. (1973) Specific purification of lymphocyte populations on a digestible immunoabsorbent. J. Immunol. 110, 313. WIGZELL, H., SUNDQVIST, K.G. & YOSHIDA, T.O. (1972) Separation of cells according to surface antigens by the use of antibody-coated columns. Fractionation of cells carrying immunoglobulins and blood group antigens. Scand. 5. Immunol. 1, 75.
