i^y and Cell Bioto.s;y {\99\)
69. 387-393
Phenotype and activation of milk-derived and peripheral blood lymphocytes from normal and coeliac subjects C. E. GIBSON,* B. A. EGLINTON,t L A. PENTTILA* AND A. G. CUMMINSt *Depariinenl of Medicine. Royal Adelaide Hospital and University of Adelaide. Adelaide and the ^Ga.stroenteroiogy Unit. The Queen Elizabeth Hospital. Woodville. Adelaide. SA 5011. Aii.slralia (Submilted 18 April 1991. Accepted for publication 29 October 1991.) Summary The phenotype of milk-derived and peripheral blood lymphoeytes from normal and coeliac subjects was assessed for CD3. aP-TcR (T cell receptor). yS-TcR. CD4. CD8. HML-I (human mucosal lymphocyte) determinants, and activation was measured by interleukin-2 receptor {1L-2R) expression. Milk cells from normal and coeliac subjects were analysed by manual immunofluoreseence and milk and blood cells from normal subjects were analysed by flow cytometry. Milk eells from two eoeliac subjects were tested for proliferation to gluten antigen. The CD4:CD8 ratio of milk lymphoeytes from both normal and eoeliac subjects was similar (0.78-1.1). but lower than that present in blood (1.5-2.1). The IL-2R expression of milk lymphocytes from both coeliac and normal subjeets was increased by 3-6 times eompared with peripheral blood eells. For example. IL-2R was present on 27.3% of milk CD3+ lymphocytes and on 8.0% of blood CD3+ lymphocytes from normal subjects. Y5- and HML-I+ T cells were inereased 4.2-fold and 12-fold respeetively compared with blood lymphocytes. Milk cells from coeliac subjects showed specific proliferation to gluten but not to soya bean antigen. We conclude that milk cells have a 'mucosal" phenotype. with increased Y5-TCR. CD8+ and HML-I+ T cells, and have an inereased proportion of aetivated eells. Milk eells from coeliae subjects have no 'toxic' phenotype. but show functional reactivity by speeific proliferation to gluten antigen. INTRODUCTION Human milk contains approximately 10^ cells/ niL. although the concentration of these cells declines slowly during lactation (1). However, the volume of milk increases, so that a relatively constant number of 10^-10^ cells is fed daily to a baby during 2-6 months of breast feeding (1). These cells consist of maerophages, colostral corpuscles (large fat-containing maerophages), 10-20% lymphocytes, and a few neutrophils (23). The neutral pH of the stomach in early infancy in conjunction with the buffering eapacity of milk is likely to lead to the survival of milk lymphocytes. Funetionally, milk lymphoeytes are hyporesponsive to mitogens and most antigens, but sometimes show reciprocal stimulation to certain antigens for whieh the peripheral systemie response is poor (4,5). Milk cells can potentially engraft in the neonatal gut and transfer immune responses (6,7).
For example, breast-fed infants have higher spontaneous, mitogen and antigen-specilie responses in proliferation assays than bottle-fed infants (8). Similarly, tuberculin hypersensitivity is present in breast-fed but not bottle-fed infants of mothers who have tuberculin hypersensitivity (9,10). Coeliac disease is an example of a mucosal hypersensitivity reaction to gluten in cereals in genetically susceptible (DR3 or DR7, DQw2) individuals, resulting in intestinal damage (11,12). The question arises in coeliac disease whether natural transplantation of milk cells is a faetor that initiates the mucosal hypersensitivitv. However, the phenotype and degree of activation of milk cells from coeliae subjeets is unknown. In a companion study, the present authors have already demonstrated that sensitized CD3+ and CD4+ T lymphocytes in the peripheral blood of coeliac subjects, but not normal subjeets, express membrane-bound IL-2R(a T eell activation marker) after incubation with
Correspondenee: Dr A. G. Cummins. Gastroenterology Unit. The Queen Elizabeth Hospital. Woodville. SA 501 1. Australia. .'ibbn'vialions ii.scJ in ihi.s paper: Coneanavalin-A. Con-A; FITC. fluoreseein isothiocyanate; HML-1. human mueosal lymphocyte; IL-2R. interleukin-2 reeeptor; LMIF. leueoeyte migration inhibition factor; PBS. phosphatebuffered saline; PE. phycoerythrin; TcR. T cell receptor.
388
C. E. GIBSON in
gluten fraction III antigen (13), and that soluble IL-2R concentration in blood is elevated 5-6 times in coeliac subjects (14). As there is a common mucosal immune system, some of these activated cells from the gut could circulate to the breast and be secreted into milk (15). There are few studies of the phenotype of milk T lymphocytes, and these have been confined to assessing total T and B cells or the CD4 and CD8 subsets of T cells (4.16,17). A higher proportion of CD8+ than CD4+ T cells is present in milk than blood. A major difficulty that continues to limit further progress is the difficulty of analysing milk lymphocytes in the presence of contaminating macrophages, eolostral corpuscles and emulsified milk fat (3). These macrophages cannot be removed successfully by adherence, by carbonyl iron ingestion, or by density gradient centrifugation. This has made flow cytometry difficult as these macrophages, eolostral corpuscles and their debris are a heterogeneous population with respect to size and morphology, and may not be gated easily from lymphocytes on forward and side scatter characteristics. The aims of this study were to extend the phenotype of milk-derived T' cells from normal subjects, to assess their activation, and to compare these results with those obtained from analysis of milk from coeliac subjects. Both manual immunofluorescence and flow cytometry were used to characterize the milk lymphocytes. In addition, milk lymphocytes from coeliac subjects were tested for specific proliferation in response to gluten antigen in two subjects, and for production of leucocyte migration inhibition factor (LMIF) in one subject. MATERIALS AND METHODS Subjects Human expressed milk was donated with consent from non-coeliac mothers attending the Maternity Section of the Queen Elizabeth Hospital, and from subjects who were known to have coeliac disease. Milk was collected within 3 weeks of deli very. Coeliac disease was diagnosed by a small bowel biopsy using the criteria of villous atrophy, surface epithelial effacement and crypt hyperplasia. These subjects improved on a gluten-free diet. Blood was taken when the coeliac subjects had been on a gluten-free diet for 3-12 months. Guidelines for human experimentation of the National Health & Medical Researeh Council of Australia were followed, and the study was approved by the Human
Ethics Committees of The Royal Adelaide Hospital and The Queen Elizabeth Hospital. Separation of milk cells Milk was clarified by shaking with mineral oil. The oily layer was removed, and milk cells were sedimented by centrifugation (600 g, 15 min) from the aqueous layer. Milk lymphocytes were washed in cold phosphate-buffered saline (PBS), and were further purified by eentrifuging through 30% Percoll (Pharmacia Fine Chemicals, Uppsala, Sweden)/PBS. Milk lymphocytes were washed again in PBS and incubated with monoclonal antibodies for phenotyping, or resuspended for proliferation or leucocyte migration inhibition assays in complete culture medium (RPMI-1640, Flow Laboratories, Sydney, Australia), containing 10% fetal calf" serum, 25 mmol/L HEPES buffer, 2 mmol/L Lglutamine, 100 ng/mL streptomycin, and 100 units/mL penicillin. Peripheral blood mononuclear cells from normal or coeliac donors were purified on a Ficoll-Paque (Pharmacia) gradient, and resuspended in PBS or complete culture medium. Immunofluorescence analysis Milk or peripheral blood cells were stained with mouse IgG monoclonal antibodies against CD3, CD4 (kindly donated by the Department of Human Immunology, Institute of Medical and Veterinary Science, Adelaide), CD8 (Leu-2a. clone SKI). ap-TcR (T cell receptor, clone WT31. Becton Dickinson, Sydney, Australia), Y6-TCR (TCR, TCR5I, T Cell Sciences, Cambridge. MA), human mucosal lymphocyte determinants (HML-l, Immunotech, Marseilles, France: 18), and IL-2R (clone 2A3, Beeton Dickinson). Antibody labelling was detected using either direct fluorescein isothiocyanate (FITC) or phycoerythrin (PE) conjugates or using a secondary goat FITC F(ab')2 anti-mouse antibody (Cappel, Organon Teknika. Veedijk, Belgium). Immunofluorescence was analysed by microscopic examination using a Zeiss epiluminescence microscope with a narrow band FITC filter. Mononuclear cells with a lymphocytic morphology were examined, while macrophages and eolostral corpuscles were excluded. Further milk samples were examined by flow eytometry with the FACScan (Becton Dickinson). As milk cells displayed wide forward and side scatter characteristics, often without segregation into discernible populations, a gate for collecting
PHENOTYPE AND ACTIVATION OF MILK LYMPHOCYTES
data from lymphocytes was set corresponding to the scatter characteristics of blood lymphocytes. Dual colour analyses were performed for IL-2R and CD3, CD4 and CD8, and for HML-1 in combinationwithCD4orCD8. CD14(Leu-M3, clone MO-P9) and CD 19 (Leu 12, clone 4G7) markers (Becton Dickinson) were used to identify macrophages/colostral corpuscles and B lymphocytes, respectively. An irrelevant mouse IgG monoclonal antibody of the same isotype was used as a control. For dual colour analyses, cursors were set at the same value, which was determined using the relevant single label. All results had control readings of the irrelevant antibody subtracted from the single histogram analysis, or from the relevant quadrants in the case of dual colour analyses. Proliferation assay Cells were incubated in complete culture medium for 3 days with gluten fraction III (100 )ig/mL), soya bean hydrolysate (100 ng/mL), or Concanavalin-A (Con-A; 10 ng/mL, Sigma, St Louis, MO) as described previously (13). [^H]Thymidine (1 nCi) was added to each aliquot of cells during the final 24 h of incubation. Cells were harvested and radioactivity measured using a Beckman liquid scintillation counter. Proliferation was expressed as the stimulation index, which is the ratio of stimulated cells minus background to unstimulated cells minus background. Indirect LMIF assay Peripheral blood mononuclear or enriched milk cells from either coeliac or normal subjects were stimulated with 100 ng/mL of each test antigen (13). Control wells contained cells without antigen. Supernatants were collected after centrifugation and assayed for leucocyte migration inhi-
389
bition activity as described previously (18). The area of migration was measured and a migration index was calculated using the formula: Migration index = Mean migration area of four experimental wells Mean migration area of four control wells RESULTS Phenotype and activation markers of milk and blood cells by manual immunofluorescence The phenotypes, including activation markers, of milk and blood lymphocytes from normal and coeliac subjects are given in Table 1. The CD4:CD8 ratios of lymphocytes in milk from normal and coeliac subjects were 0.78 and 0.96, respectively, while the ratios in peripheral blood were 1.5 and 2.1 respectively. IL-2R was expressed by 46% of milk-derived CD3+ lymphocytes from coeliac patients and by 64% of CD3+ lymphocytes in the milk of normal subjects. In contrast, the expression of 1L-2R by CD3+ cells from peripheral blood of either coeliac or normal subjects was less than 10%. Flow cytometry of milk cells The phenotypes of normal milk and peripheral blood cells, as determined by flow cytometry, are given in Table 2. A considerable proportion of maerophages and colostral corpuscles (64% CD 14+ cells) contaminated the milk lymphocytes, even when the gate had been selected to define blood lymphocytes by forward and side scatter characteristics. Nevertheless, two colour analyses were applied to further examine the phenotype of the lymphocytes. The proportion of Y5-T cells in milk was 4.2-fold higher than in blood (15.6% i'.s' 3.7%; Table 2). The ratio of CD4:CD8 was 1.1, which is similar to the value
Table 1. Phenotype of milk-derived and peripheral blood lymphocytes from normal and eoeliac subjects. Blood
Milk Determinant CD3 CD4 CD8
IL-2R
Normal (" = 6)
Coeliac (/? = 4)
Normal ("=12)
Coeliac (" = 6)
28 ± 6 I4±3 I8± 5 18± 7
41 ± 16 24 ±6* 25 ± 4 19±5
63 ± 6
62 ± 9 41 ±6 20 ± 3
38 ±8 25±4 5± 1
4±3
Milk cells were assessed by manual immunofluoreseenee and blood cells by flow eytometry. Results expressed as mean percentage of positive mononuelear cells ± s.d. *T\vo samples available for counting.
390
C. E. GIBSON/r/\.|/,.
Table 2. Flow cytometry of milk and blood cells from normal subjects. Milk 7 = 8-15)
Determinant CD3 (T lymphocytes) aP-TcR
Blood {n= 18-38)
25.6± 14.9 (15) 24.5+ 15.6(8) 4.0 ± 3 . 6 ( I I ) 13.6±8.7(12) 12.2 ± 7 . 0 (12) IO.5±4.6(15) 7.0±3.5(13) 4.6±3.3(12) l.O±O.7(12) 1.2 ±0.8 (8) 3.7± 1.6(12) 4 . 7 ± 1.6 (12) l . 2 ± 0 . 8 (11) 3.3± 1.9(12) 10.2 ± 5 . 3 (12) 64.0 ± 18.2 (15)
Y6-TCR
CD4 CD8 IL-2R
CD3 CD4 CD8 CD19 CDI4 HML-1 CD4 CD8 CDI9 (B lymphocytes) CD 14 (macrophages)
68.6 ± 7.6 67.4 ± 5 . 8 2.6 ± 1.8 42.9 ± 7 . 9 25.9 ± 6 . 7 6.2 ± 2 . 7 5.5 ± 2 . 4 4.2 ± 2 . 4 1.3 ± 0 . 6 1.7± 1.1 0.1 ± 0 . 1 1.1 ± 0 . 9 ND ND 12.6 ± 4 . 8 2.1 ± 1.6
(38) (20) (36) (38) (38) (38) (35) (38) (37) (18) (18) (38) (38) (18)
Lymphocytes were selected on forward and side scatter eharacteristics before single and dual colour analyses were used to determine their phenotype and activation. Results are expressed as mean percentage of positive cells ± s.d. («). ND: not done.
obtained by manual immunofluorescence. The propoftion of CD3+ cells that expressed IL-2R was 27.3%, which was lower than the value obtained by manual immunofluorescence (64%) but still 3.4-fold higher than that present on blood cells (Table 2). The ratio of CD4 to CD8 cells that expressed IL-2R was 4.6:1. Approximately 6% of CD 14+ macrophages also expressed IL-2R. The proportion of HML-I+ cells within the CD3+ population was 18.4% which is a higher value than the 1.1% present in adult peripheral blood. The majority of milk cells that expressed HML-1 were CD8+.
Proliferation of milk cells in response to antigen and mitogen stimulation
The stimulation index of milk cells ftom control and coeliac subjects in response to Con-A and two dietary antigens is given in Table 3. Milk lymphocytes from normal subjects were hyporesponsive to the mitogen, Con-A, as well as to gluten and soya bean food antigens, whereas milk cells from coeliac subjects proliferated in response to Con-A and showed specific stimulation by gluten antigen but not by soya bean antigen (Table 3).
Table 3. Proliferation of purified milk lymphocytes from normal and coeliac subjects after incubation tor 3 days with Con-A. gluten fraction III and soya bean hydrolysate.*
Milk Normal Coneanavalin A Gluten fraction III Soya bean hydrolysate
1.3(0.9-2.9)* 1.4(O.8-1.5)§ 1.0(0.7-1.2)
Coeliac 21 (7.0-36)* 4.9 (4.7-5.1 )S 0.8 (0.2-1.4)
Bloodt Coeliac 109 4.1 2.9
*[-'H]-thymidine incorporation was expressed as the stimulation index (median and range), which is the ratio of counts per minute of stimulated to unstimulated eells. 1"These blood data are taken from a concurrent study and serve as a comparison to the milk data (13). These coeliae subjects were not pregnant or post-partum. Group means of stimulated milk eells were compared between normal and coeliae donors using Student's /-test after logio(.v+ I) transformation to normalize the data: t/^ = 0.02; §/'
69. 387-393
Phenotype and activation of milk-derived and peripheral blood lymphocytes from normal and coeliac subjects C. E. GIBSON,* B. A. EGLINTON,t L A. PENTTILA* AND A. G. CUMMINSt *Depariinenl of Medicine. Royal Adelaide Hospital and University of Adelaide. Adelaide and the ^Ga.stroenteroiogy Unit. The Queen Elizabeth Hospital. Woodville. Adelaide. SA 5011. Aii.slralia (Submilted 18 April 1991. Accepted for publication 29 October 1991.) Summary The phenotype of milk-derived and peripheral blood lymphoeytes from normal and coeliac subjects was assessed for CD3. aP-TcR (T cell receptor). yS-TcR. CD4. CD8. HML-I (human mucosal lymphocyte) determinants, and activation was measured by interleukin-2 receptor {1L-2R) expression. Milk cells from normal and coeliac subjects were analysed by manual immunofluoreseence and milk and blood cells from normal subjects were analysed by flow cytometry. Milk eells from two eoeliac subjects were tested for proliferation to gluten antigen. The CD4:CD8 ratio of milk lymphoeytes from both normal and eoeliac subjects was similar (0.78-1.1). but lower than that present in blood (1.5-2.1). The IL-2R expression of milk lymphocytes from both coeliac and normal subjeets was increased by 3-6 times eompared with peripheral blood eells. For example. IL-2R was present on 27.3% of milk CD3+ lymphocytes and on 8.0% of blood CD3+ lymphocytes from normal subjects. Y5- and HML-I+ T cells were inereased 4.2-fold and 12-fold respeetively compared with blood lymphocytes. Milk cells from coeliac subjects showed specific proliferation to gluten but not to soya bean antigen. We conclude that milk cells have a 'mucosal" phenotype. with increased Y5-TCR. CD8+ and HML-I+ T cells, and have an inereased proportion of aetivated eells. Milk eells from coeliae subjects have no 'toxic' phenotype. but show functional reactivity by speeific proliferation to gluten antigen. INTRODUCTION Human milk contains approximately 10^ cells/ niL. although the concentration of these cells declines slowly during lactation (1). However, the volume of milk increases, so that a relatively constant number of 10^-10^ cells is fed daily to a baby during 2-6 months of breast feeding (1). These cells consist of maerophages, colostral corpuscles (large fat-containing maerophages), 10-20% lymphocytes, and a few neutrophils (23). The neutral pH of the stomach in early infancy in conjunction with the buffering eapacity of milk is likely to lead to the survival of milk lymphocytes. Funetionally, milk lymphoeytes are hyporesponsive to mitogens and most antigens, but sometimes show reciprocal stimulation to certain antigens for whieh the peripheral systemie response is poor (4,5). Milk cells can potentially engraft in the neonatal gut and transfer immune responses (6,7).
For example, breast-fed infants have higher spontaneous, mitogen and antigen-specilie responses in proliferation assays than bottle-fed infants (8). Similarly, tuberculin hypersensitivity is present in breast-fed but not bottle-fed infants of mothers who have tuberculin hypersensitivity (9,10). Coeliac disease is an example of a mucosal hypersensitivity reaction to gluten in cereals in genetically susceptible (DR3 or DR7, DQw2) individuals, resulting in intestinal damage (11,12). The question arises in coeliac disease whether natural transplantation of milk cells is a faetor that initiates the mucosal hypersensitivitv. However, the phenotype and degree of activation of milk cells from coeliae subjeets is unknown. In a companion study, the present authors have already demonstrated that sensitized CD3+ and CD4+ T lymphocytes in the peripheral blood of coeliac subjects, but not normal subjeets, express membrane-bound IL-2R(a T eell activation marker) after incubation with
Correspondenee: Dr A. G. Cummins. Gastroenterology Unit. The Queen Elizabeth Hospital. Woodville. SA 501 1. Australia. .'ibbn'vialions ii.scJ in ihi.s paper: Coneanavalin-A. Con-A; FITC. fluoreseein isothiocyanate; HML-1. human mueosal lymphocyte; IL-2R. interleukin-2 reeeptor; LMIF. leueoeyte migration inhibition factor; PBS. phosphatebuffered saline; PE. phycoerythrin; TcR. T cell receptor.
388
C. E. GIBSON in
gluten fraction III antigen (13), and that soluble IL-2R concentration in blood is elevated 5-6 times in coeliac subjects (14). As there is a common mucosal immune system, some of these activated cells from the gut could circulate to the breast and be secreted into milk (15). There are few studies of the phenotype of milk T lymphocytes, and these have been confined to assessing total T and B cells or the CD4 and CD8 subsets of T cells (4.16,17). A higher proportion of CD8+ than CD4+ T cells is present in milk than blood. A major difficulty that continues to limit further progress is the difficulty of analysing milk lymphocytes in the presence of contaminating macrophages, eolostral corpuscles and emulsified milk fat (3). These macrophages cannot be removed successfully by adherence, by carbonyl iron ingestion, or by density gradient centrifugation. This has made flow cytometry difficult as these macrophages, eolostral corpuscles and their debris are a heterogeneous population with respect to size and morphology, and may not be gated easily from lymphocytes on forward and side scatter characteristics. The aims of this study were to extend the phenotype of milk-derived T' cells from normal subjects, to assess their activation, and to compare these results with those obtained from analysis of milk from coeliac subjects. Both manual immunofluorescence and flow cytometry were used to characterize the milk lymphocytes. In addition, milk lymphocytes from coeliac subjects were tested for specific proliferation in response to gluten antigen in two subjects, and for production of leucocyte migration inhibition factor (LMIF) in one subject. MATERIALS AND METHODS Subjects Human expressed milk was donated with consent from non-coeliac mothers attending the Maternity Section of the Queen Elizabeth Hospital, and from subjects who were known to have coeliac disease. Milk was collected within 3 weeks of deli very. Coeliac disease was diagnosed by a small bowel biopsy using the criteria of villous atrophy, surface epithelial effacement and crypt hyperplasia. These subjects improved on a gluten-free diet. Blood was taken when the coeliac subjects had been on a gluten-free diet for 3-12 months. Guidelines for human experimentation of the National Health & Medical Researeh Council of Australia were followed, and the study was approved by the Human
Ethics Committees of The Royal Adelaide Hospital and The Queen Elizabeth Hospital. Separation of milk cells Milk was clarified by shaking with mineral oil. The oily layer was removed, and milk cells were sedimented by centrifugation (600 g, 15 min) from the aqueous layer. Milk lymphocytes were washed in cold phosphate-buffered saline (PBS), and were further purified by eentrifuging through 30% Percoll (Pharmacia Fine Chemicals, Uppsala, Sweden)/PBS. Milk lymphocytes were washed again in PBS and incubated with monoclonal antibodies for phenotyping, or resuspended for proliferation or leucocyte migration inhibition assays in complete culture medium (RPMI-1640, Flow Laboratories, Sydney, Australia), containing 10% fetal calf" serum, 25 mmol/L HEPES buffer, 2 mmol/L Lglutamine, 100 ng/mL streptomycin, and 100 units/mL penicillin. Peripheral blood mononuclear cells from normal or coeliac donors were purified on a Ficoll-Paque (Pharmacia) gradient, and resuspended in PBS or complete culture medium. Immunofluorescence analysis Milk or peripheral blood cells were stained with mouse IgG monoclonal antibodies against CD3, CD4 (kindly donated by the Department of Human Immunology, Institute of Medical and Veterinary Science, Adelaide), CD8 (Leu-2a. clone SKI). ap-TcR (T cell receptor, clone WT31. Becton Dickinson, Sydney, Australia), Y6-TCR (TCR, TCR5I, T Cell Sciences, Cambridge. MA), human mucosal lymphocyte determinants (HML-l, Immunotech, Marseilles, France: 18), and IL-2R (clone 2A3, Beeton Dickinson). Antibody labelling was detected using either direct fluorescein isothiocyanate (FITC) or phycoerythrin (PE) conjugates or using a secondary goat FITC F(ab')2 anti-mouse antibody (Cappel, Organon Teknika. Veedijk, Belgium). Immunofluorescence was analysed by microscopic examination using a Zeiss epiluminescence microscope with a narrow band FITC filter. Mononuclear cells with a lymphocytic morphology were examined, while macrophages and eolostral corpuscles were excluded. Further milk samples were examined by flow eytometry with the FACScan (Becton Dickinson). As milk cells displayed wide forward and side scatter characteristics, often without segregation into discernible populations, a gate for collecting
PHENOTYPE AND ACTIVATION OF MILK LYMPHOCYTES
data from lymphocytes was set corresponding to the scatter characteristics of blood lymphocytes. Dual colour analyses were performed for IL-2R and CD3, CD4 and CD8, and for HML-1 in combinationwithCD4orCD8. CD14(Leu-M3, clone MO-P9) and CD 19 (Leu 12, clone 4G7) markers (Becton Dickinson) were used to identify macrophages/colostral corpuscles and B lymphocytes, respectively. An irrelevant mouse IgG monoclonal antibody of the same isotype was used as a control. For dual colour analyses, cursors were set at the same value, which was determined using the relevant single label. All results had control readings of the irrelevant antibody subtracted from the single histogram analysis, or from the relevant quadrants in the case of dual colour analyses. Proliferation assay Cells were incubated in complete culture medium for 3 days with gluten fraction III (100 )ig/mL), soya bean hydrolysate (100 ng/mL), or Concanavalin-A (Con-A; 10 ng/mL, Sigma, St Louis, MO) as described previously (13). [^H]Thymidine (1 nCi) was added to each aliquot of cells during the final 24 h of incubation. Cells were harvested and radioactivity measured using a Beckman liquid scintillation counter. Proliferation was expressed as the stimulation index, which is the ratio of stimulated cells minus background to unstimulated cells minus background. Indirect LMIF assay Peripheral blood mononuclear or enriched milk cells from either coeliac or normal subjects were stimulated with 100 ng/mL of each test antigen (13). Control wells contained cells without antigen. Supernatants were collected after centrifugation and assayed for leucocyte migration inhi-
389
bition activity as described previously (18). The area of migration was measured and a migration index was calculated using the formula: Migration index = Mean migration area of four experimental wells Mean migration area of four control wells RESULTS Phenotype and activation markers of milk and blood cells by manual immunofluorescence The phenotypes, including activation markers, of milk and blood lymphocytes from normal and coeliac subjects are given in Table 1. The CD4:CD8 ratios of lymphocytes in milk from normal and coeliac subjects were 0.78 and 0.96, respectively, while the ratios in peripheral blood were 1.5 and 2.1 respectively. IL-2R was expressed by 46% of milk-derived CD3+ lymphocytes from coeliac patients and by 64% of CD3+ lymphocytes in the milk of normal subjects. In contrast, the expression of 1L-2R by CD3+ cells from peripheral blood of either coeliac or normal subjects was less than 10%. Flow cytometry of milk cells The phenotypes of normal milk and peripheral blood cells, as determined by flow cytometry, are given in Table 2. A considerable proportion of maerophages and colostral corpuscles (64% CD 14+ cells) contaminated the milk lymphocytes, even when the gate had been selected to define blood lymphocytes by forward and side scatter characteristics. Nevertheless, two colour analyses were applied to further examine the phenotype of the lymphocytes. The proportion of Y5-T cells in milk was 4.2-fold higher than in blood (15.6% i'.s' 3.7%; Table 2). The ratio of CD4:CD8 was 1.1, which is similar to the value
Table 1. Phenotype of milk-derived and peripheral blood lymphocytes from normal and eoeliac subjects. Blood
Milk Determinant CD3 CD4 CD8
IL-2R
Normal (" = 6)
Coeliac (/? = 4)
Normal ("=12)
Coeliac (" = 6)
28 ± 6 I4±3 I8± 5 18± 7
41 ± 16 24 ±6* 25 ± 4 19±5
63 ± 6
62 ± 9 41 ±6 20 ± 3
38 ±8 25±4 5± 1
4±3
Milk cells were assessed by manual immunofluoreseenee and blood cells by flow eytometry. Results expressed as mean percentage of positive mononuelear cells ± s.d. *T\vo samples available for counting.
390
C. E. GIBSON/r/\.|/,.
Table 2. Flow cytometry of milk and blood cells from normal subjects. Milk 7 = 8-15)
Determinant CD3 (T lymphocytes) aP-TcR
Blood {n= 18-38)
25.6± 14.9 (15) 24.5+ 15.6(8) 4.0 ± 3 . 6 ( I I ) 13.6±8.7(12) 12.2 ± 7 . 0 (12) IO.5±4.6(15) 7.0±3.5(13) 4.6±3.3(12) l.O±O.7(12) 1.2 ±0.8 (8) 3.7± 1.6(12) 4 . 7 ± 1.6 (12) l . 2 ± 0 . 8 (11) 3.3± 1.9(12) 10.2 ± 5 . 3 (12) 64.0 ± 18.2 (15)
Y6-TCR
CD4 CD8 IL-2R
CD3 CD4 CD8 CD19 CDI4 HML-1 CD4 CD8 CDI9 (B lymphocytes) CD 14 (macrophages)
68.6 ± 7.6 67.4 ± 5 . 8 2.6 ± 1.8 42.9 ± 7 . 9 25.9 ± 6 . 7 6.2 ± 2 . 7 5.5 ± 2 . 4 4.2 ± 2 . 4 1.3 ± 0 . 6 1.7± 1.1 0.1 ± 0 . 1 1.1 ± 0 . 9 ND ND 12.6 ± 4 . 8 2.1 ± 1.6
(38) (20) (36) (38) (38) (38) (35) (38) (37) (18) (18) (38) (38) (18)
Lymphocytes were selected on forward and side scatter eharacteristics before single and dual colour analyses were used to determine their phenotype and activation. Results are expressed as mean percentage of positive cells ± s.d. («). ND: not done.
obtained by manual immunofluorescence. The propoftion of CD3+ cells that expressed IL-2R was 27.3%, which was lower than the value obtained by manual immunofluorescence (64%) but still 3.4-fold higher than that present on blood cells (Table 2). The ratio of CD4 to CD8 cells that expressed IL-2R was 4.6:1. Approximately 6% of CD 14+ macrophages also expressed IL-2R. The proportion of HML-I+ cells within the CD3+ population was 18.4% which is a higher value than the 1.1% present in adult peripheral blood. The majority of milk cells that expressed HML-1 were CD8+.
Proliferation of milk cells in response to antigen and mitogen stimulation
The stimulation index of milk cells ftom control and coeliac subjects in response to Con-A and two dietary antigens is given in Table 3. Milk lymphocytes from normal subjects were hyporesponsive to the mitogen, Con-A, as well as to gluten and soya bean food antigens, whereas milk cells from coeliac subjects proliferated in response to Con-A and showed specific stimulation by gluten antigen but not by soya bean antigen (Table 3).
Table 3. Proliferation of purified milk lymphocytes from normal and coeliac subjects after incubation tor 3 days with Con-A. gluten fraction III and soya bean hydrolysate.*
Milk Normal Coneanavalin A Gluten fraction III Soya bean hydrolysate
1.3(0.9-2.9)* 1.4(O.8-1.5)§ 1.0(0.7-1.2)
Coeliac 21 (7.0-36)* 4.9 (4.7-5.1 )S 0.8 (0.2-1.4)
Bloodt Coeliac 109 4.1 2.9
*[-'H]-thymidine incorporation was expressed as the stimulation index (median and range), which is the ratio of counts per minute of stimulated to unstimulated eells. 1"These blood data are taken from a concurrent study and serve as a comparison to the milk data (13). These coeliae subjects were not pregnant or post-partum. Group means of stimulated milk eells were compared between normal and coeliae donors using Student's /-test after logio(.v+ I) transformation to normalize the data: t/^ = 0.02; §/'
