INFECTION AND IMMUNITY, Feb. 1979, p. 347-352 0019-9567/79/02-0347/06$02.00/0
Vol. 23, No. 2
Cell-Mediated Immune Response to Bacterial Products in Human Tonsils and Peripheral Blood Lymphocytes MOSHE M. DRUCKER,' Y. AGATSUMA,1 ILANA DRUCKER,' ERWIN NETER,2'3 J. BERNSTEIN,' AND P. L. OGRA2'3* Departments of Pediatrics and Microbiology, State University of New York at Buffalo,2 and Division of Infectious Diseases and Virology,1 and Bacteriology,3 Children's Hospital, Buffalo, New York 14222 Received for publication 9 November 1978
Lymphoproliferative responses of tonsillar tissue lymphocytes and peripheral blood lymphocytes to phytohemagglutinin and specific bacterial product antigens were studied in children undergoing tonsillectomy and adenoidectomy. Tonsillar tissue lymphocytes responded to optimal concentrations of phytohemagglutinin, Varidase, and streptolysin-O in a manner similar to peripheral blood lymphocytes. Higher base-line mitogenic activity in tonsillar lymphocytes was frequently associated with the presence of Staphylococcus aureus in the tonsils. Tonsillar tissue lymphocytes from 23% of the subjects with the highest base-line mitogenic activity manifested a decreased response to in vitro stimulation with mitogens or antigens. In subjects with such preactivated tonsillar lymphocytes, the proliferative responsiveness of blood lymphocytes to mitogen and antigens was markedly increased after tonsillectomy and adenoidectomy. These observations suggest the existence of in vitro correlates of cellular immunity to bacterial products in the mucosal surfaces. In addition, it is proposed that tonsils may possess immunosuppressive activity for peripheral blood lymphocytes, which may be related to local tonsillar infections.
Recently, the in vitro correlates of cell-mediated immunity to viral antigens have been demonstrated in human tonsillar tissue lymphocytes (TTL) and peripheral blood lymphocytes (PBL) (6, 8). It has been suggested that human TTL respond differently from PBL to purified protein derivative and mitogens (17). Human PBL have been stimulated by oral bacterial antigens, and increased responses have been observed in subjects with chronic oral infections such as gingivitis or periodontitis (11, 13). The present study was designed to determine the presence and degree of cell-mediated immune responses in TTL and PBL, using as stimulants phytohemagglutinin (PHA) and some ubiquitous bacterial products, employing the techniques of in vitro lymphoproliferation. In addition, we attempted to correlate the reactivity of TTL before in vitro mitogenic or antigenic stimulation with the constituents of preexisting tonsillar flora. The effect of tonsillectomy and adenoidectomy (T&A) on the lymphoproliferative responses of PBL was investigated and related to the proliferative activity of the unstimulated TTL.
included in this study. Ten other patients who underwent other forms of surgery, such as repairs of inguinal or umbilical hernias or endoscopy under anesthesia, were included as controls. The indications for tonsillectomy were: recurrent suppurative tonsillitis, recurrent cervical lymphadenopathy, or increased incidence of upper respiratory tract infections. The average age was 10.4 years, with a range of 6 to 20 years. The study population consisted of 19 females and 24 males. All patients were clinically well at the time the lymphocyte function was studied. Bacterial cultures. Both tonsils were swabbed and cultured before surgery. After T&A, the excised tonsils were washed vigorously in sterile normal saline, and the washings were cultured. The tonsils were cut and minced, and a sample of the minced tonsils was also cultured. The relative concentration of different groups of bacteria recovered in each specimen was evaluated. Detection of Epstein-Barr virus infection. The detection of Epstein-Barr virus nuclear antigen in lymphocytes, antibodies to this antigen in plasma, and antibodies to viral capsid antigen of Epstein-Barr virus were determined as described previously (4, 14). Lymphocyte cultures. Serum and heparinized blood specimens were collected immediately before and 1 month after tonsillectomy. Lymphocytes were separated from the blood and from minced tonsils by Ficoll-Hypaque fractionation (1), washed twice in MATERIALS AND METHODS Hanks balanced salt solution, and suspended in a final Patient population. Forty-three patients undergo- concentration of 106 cells per ml in RPMI 1640 meing T&A at the Children's Hospital of Buffalo were dium, containing 100 U of penicillin G and 100 ,ug of 347
348
DRUCKER ET AL.
streptomycin per ml and supplemented with 10% autologous plasma. ERF cells. Sheep erythrocytes were washed three times with Hanks Balanced salt solution, and 1 volume of packed cells was suspended in 40 volumes of fetal calf serum. Sedimented lymphocytes (106) were mixed with 0.025 ml of sheep erythrocyte suspension and 0.05 ml of fetal calf serum. The mixture was incubated for 15 min in a 370C water bath and centrifuged for 5 mi at 275 x g. After overnight refrigeration at 40C, the cells were gently resuspended, and the percentage of lymphocytes bearing >3 sheep erythrocytes was counted as E-rosette-forming (ERF) T-lymphocytes. To determine the nature of the cell type involved in the in vitro lymphoproliferative responses, the ERF and non-ERF cells from PBL and TTL were tested separately after in vitro stimulation. The fractionation of the T-cell-rich (ERF lymphocytes) and non-ERF lymphocyte population was performed as described previously (2). Briefly, the whole blood or tonsillar cell specimens of the subjects were cultured in presence of PHA and labeled with ['H]thymidine. After 24 h, the cultures were harvested and subjected to Ficoll-Hypaque fractionation. The lymphocyte-rich fractions were collected and mixed with sheep erythrocytes to form E-rosettes with T-lymphocytes and non-ERF lymphocytes using another Ficoll-Hypaque fractionation. The erythrocytes in each fraction were lysed by treatment with acetic acid, and the [3H]thymidine uptake in different lymphocyte fractions was determined by scintillation counting. Mitogens and antigens. The lymphocytes were stimulated by the following mitogens and antigens: purified PHA (Wellcome Research Laboratories, Beckenham, England), in concentrations of 0.8 and 8 pg/ml; streptolysin 0 (SLO; Difco Laboratories, Detroit, Mich.), in a concentration of 0.5 minimum hemolytic units per ml; and Varidase (preservative free, kindly supplied by Lederle Laboratories, Pearl River, N.Y.), used in a dilution containing 1,000 U of streptokinase per ml. Purified tetanus toxoid (kindly supplied by the Biologic Laboratories of the Department of Public Health, Boston, Mass.) was used in a dilution that contained 1.1 Lf/ml. Staphylococcal filtrates were prepared from Staphylococcus aureus strains isolated from tonsils and kept on nutrient agar slants. The bacteria were grown overnight in brain heart broth in 5% C02 at 370C. The bacterial suspensions were centrifuged, and the supernatant was filtered through a 0.45-pm Millex filter (Millipore Corp., Bedford, Mass.). Lymphocyte stimulation. The in vitro lymphocyte transformation was assessed using a micromethod, as follows. The antigens were introduced in a 0.02-ml volume into the wells of ultraviolet-sterilized, U-bottomed Cooke microtiter plates (Cooke Laboratory Products, Alexandria, Va.), followed by 0.2 ml of the cell suspension. The plates were sealed and incubated for 4 days in 5% CO2 at 370C. Subsequently, the cells were labeled with [8H]thymidine, 10 ,uCi/ml, in RPMI 1640. The cell cultures were harvested 20 h later using a multiple automated sample harvester (MASH, model U; Microbiological Associates, Bethesda, Md.), and the paper disks containing cells were processed for determination of radioactivity. Each sample was run in triplicate. The stimulation index
INFECT. IMMUN. was defined as the ratio of the mean counts per minute (CPM) of the stimulated lymphocytes divided by the mean CPM of unstimulated lymphocytes. For the sake of the present studies, a stimulation index of 3 or higher with the antigens and 10 or higher with PHA was considered as evidence of in vitro lymphocyte stimulation. The 8 CPM was expressed as the net difference of CPM observed between the stimulated and unstimulated cultures. To explore the role of TTL on the proliferative response of PBL, samples of TTL were preincubated with either mitomycin C (40 tLg/ml for 30 min), concanavalin A (10 jig/ml for 24 h), or culture medium alone. The preincubated TTL were then cocultured with PBL of the same subjects and stimulated with PHA at a concentration of 0.8 jig/ml of the culture (5,
15).
RESULTS ERF cells. The T-lymphocytes in peripheral blood as evidenced by E-rosette formation were found to constitute 53 ± 10% of PBL just before T&A and 68 ± 12% PBL 1 month after T&A. Similar numbers of ERF cells were observed in the 10 other surgical control (non-T&A) ageand sex-matched subjects tested simultaneously. The T-lymphocytes in tonsillar tissue collected at the T&A comprised approximately 23 ± 20% of TTL. Base-line activity. The CPM in the absence of any mitogenic or antigenic stimulation in all PBL specimens tested was found to be below 3,000 and frequently below 1,000, and the geometric mean CPM in PBL was about 964. The unstimulated CPM in TTL ranged between 300 to 30,000, and the geometric mean CPM was 2,136. Ten (23%) of the 43 subjects studied were found to have exceptionally high base-line proliferative activity in the absence of any induced stimulation. The individual CPM in the TTL of these subjects ranged from 3,000 to 30,000, with a mean value of about 5,000. For the sake of further discussion, the TTL whose base-line proliferative CPM was above the geometric mean were termed preactivated tonsillar lymphocytes (PTL). Sixty-six percent of these subjects were found to harbor S. aureus in the tonsils, whereas only 30% of other subjects whose TTL base-line CPM was below the geometric mean exhibited staphylococcal colonization in the tonsils (Table 1). The association of Staphylococcus with PTL was statistically significant (P < 0.05). It should be pointed out that the staphylococcal strains isolated from tonsils belonged to different phage types. In addition, serial dilutions of culture filtrates prepared from the different isolates of S. aureus were tested for their mitogenic effect on human cord blood lymphocytes. All the filtrates that were tested caused significant lymphoproliferation even in dilutions of 1:1,000. No
CELL-MEDIATED IMMUNE RESPONSE IN HUMAN TONSILS
VOL. 23, 1979
TABLE 1. Association of tonsillar unstimulated lymphoproliferative activity and the presence of S. aureus No. of subjects S. aureus
CPM < 2,136a CPM > 2,136a 14 7 Present 7 15 Absentb a Represents geometric mean value. b Chi-square test, P < 0.05.
Total 21 22
349
PBL or TTL after mitogenic or antigenic stimulation in subjects with base-line CPM of 5,000 or more. The 8 CPM of stimulated PBL and TTL varied widely in different individuals; there was some degree of correlation, however, between the values in PBL and TTL belonging to the same individuals. The degree of this correlation and its nature seemed to depend on the mitogen or antigen employed for stimulation. Differential response by T- and B-lymphocytes. The response of PHA stimulation was deficient in the T-cell-depleted population of PBL and TTL, but the T-cell-rich population exhibited marked proliferative responses (Table 4). Similarly, the SLO response was most conspicuous in the T-cell-rich population, and little or no SLO-specific proliferative activity was observed in T-cell-depleted PBL or TTL (Table 4). Effect of T&A on proliferative activity of PBL. In the subjects whose tonsils did not manifest any PTL activity, the lymphoproliferative responses in peripheral blood were essentially similar before and after T&A (Table 5). On the other hand, the subjects who had high unstimulated (5,000 CPM) proliferative activity in tonsils evidenced a marked increase in the proliferative response to PHA, Varidase, SLO, and tetanus toxoid in the peripheral blood lymphocytes after T&A, as compared to the responses observed before T&A (Table 5). Similar studies carried out in non-T&A sub-
difference could be demonstrated between strains derived from subjects with PTL and strains isolated from other subjects. Attempts to isolate viruses from tonsillar tissue of subjects with PTL, on routine tissue culture infectivity, and to isolate Epstein-Barr virus or its nuclear antigen proved unsuccessful. Antibodies to viral capsid antigen were found in 3 of the 13 subjects tested, but only 1 subject had evidence of PTL. Anti-Epstein-Barr virus nuclear antigen antibodies were detected in one individual who was found to have PTL. Subjects with PTL did not differ from other subjects in regard to sex, race, clinical course, tonsil size, or histological morphology. Lymphoproliferative response after stimulation. PHA in suboptimal concentrations (0.08 ,ug/ml) resulted in stimulation of 66% of PBL but only 10% of TTL (Table 2). However, PHA in optimal concentration (0.8 Aug/ml) stimulated all of PBL and 74% of TTL. Varidase and SLO stimulated 74 to 78% of PBL, but only 33 to 49% of TTL. PBL responded to tetanus toxoid TABLE 2. Lymphoproliferative response to various in 59% of cases, whereas TTL responded to the stimulants in TTL and PBL same antigen in only one case (3%); (Table 2). % Responders No. of subA comparison of the responses of TTL and jects Tonsil Blood PBL after stimulation with PHA indicated that PTL rarely if ever demonstrated significant in PHA (0.08 ug/ml)' 10 66 41 vitro stimulation, and PBL of the same subjects 74 100 PHA (0.8 Ig/ml)a 42 usually manifested significantly reduced in vitro 49 78 Varidaseb 41 SLOb 42 33 74 stimulation (Table 3). Similar results were obTetanus toxoidb 3 59 32 served when the CPM after stimulation with Varidase and streptokinase was related to the aSignificant response expressed as 10-fold or levels of base-line CPM in different tonsillar greater increase in stimulation index. b tissues and PBL studied. As shown in Table 3, Significant response expressed as threefold or little or no increase in CPM was observed in the greater increase in stimulation index. TABLE 3. Proliferative response in TTL and PBL after stimulation with mitogens or antigens, relative to base-line unstimulated activity in TTL
lanlated TTLJ
m CPM in
No. of sub-
jects
Mean 8 CPM in 1L after stimulation with: SLO Varidase PHA
43,000 21,000 100 2,800 with PTL. subjects aRepresents 200 1,500 5,000
12 21
1,680 2,850 -3,500
2,200 3,900 680
Mean 8 CPM in PBL after stimulation with: SLO Varidase PHA
68,570 39,300 8,500
2,153 3,220 640
1,900 3,855 430
350
INFECT. IMMUN.
DRUCKER ET AL.
TABLE 6. In vitro proliferative responses of PBL to PHA in coculture with drug-treated TTLa 8 CPM reMean CPM sponge to Cell population PHA Control PHA 54 40,526 40,580 PBL
jects before and 1 month after other surgical procedures revealed no significant difference in 8 CPM in response to PHA, Varidase, SLO, and tetanus toxoid and other mitogens. The 8 CPM responses before and after surgery were approximately similar. Cocultivation of PBL and TTL exhibited significant reduction in the 8 CPM. The inhibition of 8 CPM in PBL was more clearly demonstrated by cocultivation of PBL with concanavalin Aactivated TTL. The proliferative activity in such cultures was three- to fourfold lower than in cultures of PBL alone (Table 6). The suppressor effect of TTL was masked to a large extent by preincubation with mitomycin C before coculturing with PBL (Table 6).
TTL TTLMl
TTLCo
phoproliferative activity were frequently associated with S. aureus infection in the tonsils. S. aureus has been shown to induce lymphoproliferative activity in in vitro settings (18). It produces strong soluble mitogens (18), carries cellbound antigens such as protein A, which seem to be specific B-cell stimulators in humans (3), and contains antigens such as teichoic acid which significantly stimulate lymphoproliferation in presensitized lymphocytes (9). In view of the frequent occurrence of staphylococcal colonization in the nasopharynx, these data suggest that frequent infection with Staphylococcus may be a potent stimulus for mucosal cell-mediated immune response in in vivo settings. Stimulation of human lymphocytes by bacteria is supported by the observation that patients with septic bacterial infections have higher than normal spontaneous blast transformation in PBL (19). Epstein-Barr virus infection is also well known to induce lymphoproliferation in vivo and in vitro (20). However, in the present study we could not demonstrate any obvious
TABLE 4. Lymphoproliferative response to PHA and SLO in unseparated, T-cell-rich and T-celldepleted populations of PBL and TTL Mean 8 CPM' Stimulant Tonsil
Blood
Unseparated lympho- PHA SLO cytes
48,550 2,700
62,000 2,800
T-cell-rich population PHA SLO (ERF cells)
46,930 3,600
59,512 4,800
5,650 893 4,310
8,750 1,043 4,500
17,352 21,432 PBL + TTL 4,080 163 36,789 36,952 PBL + TTLMl 210 PBL + TTLco 16,220 16,430 a TTL were preincubated with mitomycin C (TTLMl) or concanavalin A (TTLco). Data represent four patients.
DISCUSSION It has been previously suggested that the spontaneous lymphoproliferative activity of TTL is much higher than observed in corresponding PBL (10). However, the mechanism underlying this phenomenon was not explored in further detail. In the present study it was shown that TTL with high unstimulated lym-
Population
3,100 150 190
1,759 1,580 T-cell-depleted popula- PHA 120 200 tion (non-ERF cells) SLO a Represents the mean value of six experiments.
TABLE 5. Changes in the proliferative response of PBL after T&A in relation to the prestimulation activity in TTL Proliferative activity in PBL
Unstimulated
No. of sub-
jects ~~ Stimulant ~ ~ ~~et
mean CPM in TTL
5,000
a NS, Not significant.
Mean
vaCPM P value
Pre-T&A
Post-T&A
PHA (0.8 ,ug/ml) Varidase SLO Tetanus toxoid
14 13 14 12
53,000 2,356 1,860 2,630
59,504 2,010 1,630 2,800
NSa NS
PHA (0.8 pg/ml) Varidase SLO Tetanus toxoid
5 5 5 5
8,500
62,040 1,838 1,540 1,660
Vol. 23, No. 2
Cell-Mediated Immune Response to Bacterial Products in Human Tonsils and Peripheral Blood Lymphocytes MOSHE M. DRUCKER,' Y. AGATSUMA,1 ILANA DRUCKER,' ERWIN NETER,2'3 J. BERNSTEIN,' AND P. L. OGRA2'3* Departments of Pediatrics and Microbiology, State University of New York at Buffalo,2 and Division of Infectious Diseases and Virology,1 and Bacteriology,3 Children's Hospital, Buffalo, New York 14222 Received for publication 9 November 1978
Lymphoproliferative responses of tonsillar tissue lymphocytes and peripheral blood lymphocytes to phytohemagglutinin and specific bacterial product antigens were studied in children undergoing tonsillectomy and adenoidectomy. Tonsillar tissue lymphocytes responded to optimal concentrations of phytohemagglutinin, Varidase, and streptolysin-O in a manner similar to peripheral blood lymphocytes. Higher base-line mitogenic activity in tonsillar lymphocytes was frequently associated with the presence of Staphylococcus aureus in the tonsils. Tonsillar tissue lymphocytes from 23% of the subjects with the highest base-line mitogenic activity manifested a decreased response to in vitro stimulation with mitogens or antigens. In subjects with such preactivated tonsillar lymphocytes, the proliferative responsiveness of blood lymphocytes to mitogen and antigens was markedly increased after tonsillectomy and adenoidectomy. These observations suggest the existence of in vitro correlates of cellular immunity to bacterial products in the mucosal surfaces. In addition, it is proposed that tonsils may possess immunosuppressive activity for peripheral blood lymphocytes, which may be related to local tonsillar infections.
Recently, the in vitro correlates of cell-mediated immunity to viral antigens have been demonstrated in human tonsillar tissue lymphocytes (TTL) and peripheral blood lymphocytes (PBL) (6, 8). It has been suggested that human TTL respond differently from PBL to purified protein derivative and mitogens (17). Human PBL have been stimulated by oral bacterial antigens, and increased responses have been observed in subjects with chronic oral infections such as gingivitis or periodontitis (11, 13). The present study was designed to determine the presence and degree of cell-mediated immune responses in TTL and PBL, using as stimulants phytohemagglutinin (PHA) and some ubiquitous bacterial products, employing the techniques of in vitro lymphoproliferation. In addition, we attempted to correlate the reactivity of TTL before in vitro mitogenic or antigenic stimulation with the constituents of preexisting tonsillar flora. The effect of tonsillectomy and adenoidectomy (T&A) on the lymphoproliferative responses of PBL was investigated and related to the proliferative activity of the unstimulated TTL.
included in this study. Ten other patients who underwent other forms of surgery, such as repairs of inguinal or umbilical hernias or endoscopy under anesthesia, were included as controls. The indications for tonsillectomy were: recurrent suppurative tonsillitis, recurrent cervical lymphadenopathy, or increased incidence of upper respiratory tract infections. The average age was 10.4 years, with a range of 6 to 20 years. The study population consisted of 19 females and 24 males. All patients were clinically well at the time the lymphocyte function was studied. Bacterial cultures. Both tonsils were swabbed and cultured before surgery. After T&A, the excised tonsils were washed vigorously in sterile normal saline, and the washings were cultured. The tonsils were cut and minced, and a sample of the minced tonsils was also cultured. The relative concentration of different groups of bacteria recovered in each specimen was evaluated. Detection of Epstein-Barr virus infection. The detection of Epstein-Barr virus nuclear antigen in lymphocytes, antibodies to this antigen in plasma, and antibodies to viral capsid antigen of Epstein-Barr virus were determined as described previously (4, 14). Lymphocyte cultures. Serum and heparinized blood specimens were collected immediately before and 1 month after tonsillectomy. Lymphocytes were separated from the blood and from minced tonsils by Ficoll-Hypaque fractionation (1), washed twice in MATERIALS AND METHODS Hanks balanced salt solution, and suspended in a final Patient population. Forty-three patients undergo- concentration of 106 cells per ml in RPMI 1640 meing T&A at the Children's Hospital of Buffalo were dium, containing 100 U of penicillin G and 100 ,ug of 347
348
DRUCKER ET AL.
streptomycin per ml and supplemented with 10% autologous plasma. ERF cells. Sheep erythrocytes were washed three times with Hanks Balanced salt solution, and 1 volume of packed cells was suspended in 40 volumes of fetal calf serum. Sedimented lymphocytes (106) were mixed with 0.025 ml of sheep erythrocyte suspension and 0.05 ml of fetal calf serum. The mixture was incubated for 15 min in a 370C water bath and centrifuged for 5 mi at 275 x g. After overnight refrigeration at 40C, the cells were gently resuspended, and the percentage of lymphocytes bearing >3 sheep erythrocytes was counted as E-rosette-forming (ERF) T-lymphocytes. To determine the nature of the cell type involved in the in vitro lymphoproliferative responses, the ERF and non-ERF cells from PBL and TTL were tested separately after in vitro stimulation. The fractionation of the T-cell-rich (ERF lymphocytes) and non-ERF lymphocyte population was performed as described previously (2). Briefly, the whole blood or tonsillar cell specimens of the subjects were cultured in presence of PHA and labeled with ['H]thymidine. After 24 h, the cultures were harvested and subjected to Ficoll-Hypaque fractionation. The lymphocyte-rich fractions were collected and mixed with sheep erythrocytes to form E-rosettes with T-lymphocytes and non-ERF lymphocytes using another Ficoll-Hypaque fractionation. The erythrocytes in each fraction were lysed by treatment with acetic acid, and the [3H]thymidine uptake in different lymphocyte fractions was determined by scintillation counting. Mitogens and antigens. The lymphocytes were stimulated by the following mitogens and antigens: purified PHA (Wellcome Research Laboratories, Beckenham, England), in concentrations of 0.8 and 8 pg/ml; streptolysin 0 (SLO; Difco Laboratories, Detroit, Mich.), in a concentration of 0.5 minimum hemolytic units per ml; and Varidase (preservative free, kindly supplied by Lederle Laboratories, Pearl River, N.Y.), used in a dilution containing 1,000 U of streptokinase per ml. Purified tetanus toxoid (kindly supplied by the Biologic Laboratories of the Department of Public Health, Boston, Mass.) was used in a dilution that contained 1.1 Lf/ml. Staphylococcal filtrates were prepared from Staphylococcus aureus strains isolated from tonsils and kept on nutrient agar slants. The bacteria were grown overnight in brain heart broth in 5% C02 at 370C. The bacterial suspensions were centrifuged, and the supernatant was filtered through a 0.45-pm Millex filter (Millipore Corp., Bedford, Mass.). Lymphocyte stimulation. The in vitro lymphocyte transformation was assessed using a micromethod, as follows. The antigens were introduced in a 0.02-ml volume into the wells of ultraviolet-sterilized, U-bottomed Cooke microtiter plates (Cooke Laboratory Products, Alexandria, Va.), followed by 0.2 ml of the cell suspension. The plates were sealed and incubated for 4 days in 5% CO2 at 370C. Subsequently, the cells were labeled with [8H]thymidine, 10 ,uCi/ml, in RPMI 1640. The cell cultures were harvested 20 h later using a multiple automated sample harvester (MASH, model U; Microbiological Associates, Bethesda, Md.), and the paper disks containing cells were processed for determination of radioactivity. Each sample was run in triplicate. The stimulation index
INFECT. IMMUN. was defined as the ratio of the mean counts per minute (CPM) of the stimulated lymphocytes divided by the mean CPM of unstimulated lymphocytes. For the sake of the present studies, a stimulation index of 3 or higher with the antigens and 10 or higher with PHA was considered as evidence of in vitro lymphocyte stimulation. The 8 CPM was expressed as the net difference of CPM observed between the stimulated and unstimulated cultures. To explore the role of TTL on the proliferative response of PBL, samples of TTL were preincubated with either mitomycin C (40 tLg/ml for 30 min), concanavalin A (10 jig/ml for 24 h), or culture medium alone. The preincubated TTL were then cocultured with PBL of the same subjects and stimulated with PHA at a concentration of 0.8 jig/ml of the culture (5,
15).
RESULTS ERF cells. The T-lymphocytes in peripheral blood as evidenced by E-rosette formation were found to constitute 53 ± 10% of PBL just before T&A and 68 ± 12% PBL 1 month after T&A. Similar numbers of ERF cells were observed in the 10 other surgical control (non-T&A) ageand sex-matched subjects tested simultaneously. The T-lymphocytes in tonsillar tissue collected at the T&A comprised approximately 23 ± 20% of TTL. Base-line activity. The CPM in the absence of any mitogenic or antigenic stimulation in all PBL specimens tested was found to be below 3,000 and frequently below 1,000, and the geometric mean CPM in PBL was about 964. The unstimulated CPM in TTL ranged between 300 to 30,000, and the geometric mean CPM was 2,136. Ten (23%) of the 43 subjects studied were found to have exceptionally high base-line proliferative activity in the absence of any induced stimulation. The individual CPM in the TTL of these subjects ranged from 3,000 to 30,000, with a mean value of about 5,000. For the sake of further discussion, the TTL whose base-line proliferative CPM was above the geometric mean were termed preactivated tonsillar lymphocytes (PTL). Sixty-six percent of these subjects were found to harbor S. aureus in the tonsils, whereas only 30% of other subjects whose TTL base-line CPM was below the geometric mean exhibited staphylococcal colonization in the tonsils (Table 1). The association of Staphylococcus with PTL was statistically significant (P < 0.05). It should be pointed out that the staphylococcal strains isolated from tonsils belonged to different phage types. In addition, serial dilutions of culture filtrates prepared from the different isolates of S. aureus were tested for their mitogenic effect on human cord blood lymphocytes. All the filtrates that were tested caused significant lymphoproliferation even in dilutions of 1:1,000. No
CELL-MEDIATED IMMUNE RESPONSE IN HUMAN TONSILS
VOL. 23, 1979
TABLE 1. Association of tonsillar unstimulated lymphoproliferative activity and the presence of S. aureus No. of subjects S. aureus
CPM < 2,136a CPM > 2,136a 14 7 Present 7 15 Absentb a Represents geometric mean value. b Chi-square test, P < 0.05.
Total 21 22
349
PBL or TTL after mitogenic or antigenic stimulation in subjects with base-line CPM of 5,000 or more. The 8 CPM of stimulated PBL and TTL varied widely in different individuals; there was some degree of correlation, however, between the values in PBL and TTL belonging to the same individuals. The degree of this correlation and its nature seemed to depend on the mitogen or antigen employed for stimulation. Differential response by T- and B-lymphocytes. The response of PHA stimulation was deficient in the T-cell-depleted population of PBL and TTL, but the T-cell-rich population exhibited marked proliferative responses (Table 4). Similarly, the SLO response was most conspicuous in the T-cell-rich population, and little or no SLO-specific proliferative activity was observed in T-cell-depleted PBL or TTL (Table 4). Effect of T&A on proliferative activity of PBL. In the subjects whose tonsils did not manifest any PTL activity, the lymphoproliferative responses in peripheral blood were essentially similar before and after T&A (Table 5). On the other hand, the subjects who had high unstimulated (5,000 CPM) proliferative activity in tonsils evidenced a marked increase in the proliferative response to PHA, Varidase, SLO, and tetanus toxoid in the peripheral blood lymphocytes after T&A, as compared to the responses observed before T&A (Table 5). Similar studies carried out in non-T&A sub-
difference could be demonstrated between strains derived from subjects with PTL and strains isolated from other subjects. Attempts to isolate viruses from tonsillar tissue of subjects with PTL, on routine tissue culture infectivity, and to isolate Epstein-Barr virus or its nuclear antigen proved unsuccessful. Antibodies to viral capsid antigen were found in 3 of the 13 subjects tested, but only 1 subject had evidence of PTL. Anti-Epstein-Barr virus nuclear antigen antibodies were detected in one individual who was found to have PTL. Subjects with PTL did not differ from other subjects in regard to sex, race, clinical course, tonsil size, or histological morphology. Lymphoproliferative response after stimulation. PHA in suboptimal concentrations (0.08 ,ug/ml) resulted in stimulation of 66% of PBL but only 10% of TTL (Table 2). However, PHA in optimal concentration (0.8 Aug/ml) stimulated all of PBL and 74% of TTL. Varidase and SLO stimulated 74 to 78% of PBL, but only 33 to 49% of TTL. PBL responded to tetanus toxoid TABLE 2. Lymphoproliferative response to various in 59% of cases, whereas TTL responded to the stimulants in TTL and PBL same antigen in only one case (3%); (Table 2). % Responders No. of subA comparison of the responses of TTL and jects Tonsil Blood PBL after stimulation with PHA indicated that PTL rarely if ever demonstrated significant in PHA (0.08 ug/ml)' 10 66 41 vitro stimulation, and PBL of the same subjects 74 100 PHA (0.8 Ig/ml)a 42 usually manifested significantly reduced in vitro 49 78 Varidaseb 41 SLOb 42 33 74 stimulation (Table 3). Similar results were obTetanus toxoidb 3 59 32 served when the CPM after stimulation with Varidase and streptokinase was related to the aSignificant response expressed as 10-fold or levels of base-line CPM in different tonsillar greater increase in stimulation index. b tissues and PBL studied. As shown in Table 3, Significant response expressed as threefold or little or no increase in CPM was observed in the greater increase in stimulation index. TABLE 3. Proliferative response in TTL and PBL after stimulation with mitogens or antigens, relative to base-line unstimulated activity in TTL
lanlated TTLJ
m CPM in
No. of sub-
jects
Mean 8 CPM in 1L after stimulation with: SLO Varidase PHA
43,000 21,000 100 2,800 with PTL. subjects aRepresents 200 1,500 5,000
12 21
1,680 2,850 -3,500
2,200 3,900 680
Mean 8 CPM in PBL after stimulation with: SLO Varidase PHA
68,570 39,300 8,500
2,153 3,220 640
1,900 3,855 430
350
INFECT. IMMUN.
DRUCKER ET AL.
TABLE 6. In vitro proliferative responses of PBL to PHA in coculture with drug-treated TTLa 8 CPM reMean CPM sponge to Cell population PHA Control PHA 54 40,526 40,580 PBL
jects before and 1 month after other surgical procedures revealed no significant difference in 8 CPM in response to PHA, Varidase, SLO, and tetanus toxoid and other mitogens. The 8 CPM responses before and after surgery were approximately similar. Cocultivation of PBL and TTL exhibited significant reduction in the 8 CPM. The inhibition of 8 CPM in PBL was more clearly demonstrated by cocultivation of PBL with concanavalin Aactivated TTL. The proliferative activity in such cultures was three- to fourfold lower than in cultures of PBL alone (Table 6). The suppressor effect of TTL was masked to a large extent by preincubation with mitomycin C before coculturing with PBL (Table 6).
TTL TTLMl
TTLCo
phoproliferative activity were frequently associated with S. aureus infection in the tonsils. S. aureus has been shown to induce lymphoproliferative activity in in vitro settings (18). It produces strong soluble mitogens (18), carries cellbound antigens such as protein A, which seem to be specific B-cell stimulators in humans (3), and contains antigens such as teichoic acid which significantly stimulate lymphoproliferation in presensitized lymphocytes (9). In view of the frequent occurrence of staphylococcal colonization in the nasopharynx, these data suggest that frequent infection with Staphylococcus may be a potent stimulus for mucosal cell-mediated immune response in in vivo settings. Stimulation of human lymphocytes by bacteria is supported by the observation that patients with septic bacterial infections have higher than normal spontaneous blast transformation in PBL (19). Epstein-Barr virus infection is also well known to induce lymphoproliferation in vivo and in vitro (20). However, in the present study we could not demonstrate any obvious
TABLE 4. Lymphoproliferative response to PHA and SLO in unseparated, T-cell-rich and T-celldepleted populations of PBL and TTL Mean 8 CPM' Stimulant Tonsil
Blood
Unseparated lympho- PHA SLO cytes
48,550 2,700
62,000 2,800
T-cell-rich population PHA SLO (ERF cells)
46,930 3,600
59,512 4,800
5,650 893 4,310
8,750 1,043 4,500
17,352 21,432 PBL + TTL 4,080 163 36,789 36,952 PBL + TTLMl 210 PBL + TTLco 16,220 16,430 a TTL were preincubated with mitomycin C (TTLMl) or concanavalin A (TTLco). Data represent four patients.
DISCUSSION It has been previously suggested that the spontaneous lymphoproliferative activity of TTL is much higher than observed in corresponding PBL (10). However, the mechanism underlying this phenomenon was not explored in further detail. In the present study it was shown that TTL with high unstimulated lym-
Population
3,100 150 190
1,759 1,580 T-cell-depleted popula- PHA 120 200 tion (non-ERF cells) SLO a Represents the mean value of six experiments.
TABLE 5. Changes in the proliferative response of PBL after T&A in relation to the prestimulation activity in TTL Proliferative activity in PBL
Unstimulated
No. of sub-
jects ~~ Stimulant ~ ~ ~~et
mean CPM in TTL
5,000
a NS, Not significant.
Mean
vaCPM P value
Pre-T&A
Post-T&A
PHA (0.8 ,ug/ml) Varidase SLO Tetanus toxoid
14 13 14 12
53,000 2,356 1,860 2,630
59,504 2,010 1,630 2,800
NSa NS
PHA (0.8 pg/ml) Varidase SLO Tetanus toxoid
5 5 5 5
8,500
62,040 1,838 1,540 1,660
