- Email: [email protected]
Cytotoxic cells are activated in cellular infiltrates of alcoholic chronic pancreatitis
GASTROENTEROLOGY 1997;112:1656–1663
Cytotoxic Cells Are Activated in Cellular Infiltrates of Alcoholic Chronic Pancreatitis ROBERT E. HUNGER,* CHRISTOPH MUELLER,* KASPAR Z’GRAGGEN,‡ HELMUT FRIESS,‡ ¨ CHLER‡ and MARKUS W. BU Departments of *Pathology and ‡Visceral and Transplantation Surgery, University of Bern, Bern, Switzerland
See editorial on page 1762. Background & Aims: Perforin messenger RNA (mRNA) expression has been shown to be a specific in vivo activation marker for cytotoxic cells. In this study, the contribution of cell-mediated cytotoxicity in the pathogenesis of alcoholic chronic pancreatitis is assessed. Methods: Tissue sections of patients with alcoholic chronic pancreatitis were analyzed for perforin mRNA expression by in situ hybridization. In a further step, the phenotype and the relative frequency of perforin mRNA–expressing cells were determined. Results: In the normal pancreas, perforin mRNA–expressing cells are rarely present (mean, 0.3 cells/mm2). In contrast, the frequency of perforin mRNA–expressing cells is increased severalfold in diseased tissue specimens (mean, 6.6 cells/mm2). The frequency of perforin mRNA–expressing cells is high in the CD56/ (18%) and CD8/ cell population (12%) but low in the CD4/ cell population (1%). Conclusions: The significantly elevated frequencies of perforin mRNA–expressing cells in the pancreas of patients with alcoholic chronic pancreatitis suggest an involvement of cell-mediated cytotoxicity in the pathogenesis of this disease. The preferential localization of these activated cells close to areas with parenchyma provides circumstantial evidence that autoreactive cytotoxic cells may contribute to tissue destruction in alcoholic chronic pancreatitis.
C
hronic pancreatitis is a disease characterized by irreversible destruction of the pancreas and replacement of the parenchyma by a dense fibrous tissue containing numerous lymphocytes. This process may lead to endocrine and exocrine insufficiencies. In developed countries, 60%–70% of the patients have a long history of heavy consumption of alcohol.1 Information on the mechanisms operative during tissue destruction in chronic pancreatitis is very limited. However, some observations seem to indicate an involvement of cell-mediated cytotoxicity in the pathogenesis of this disease. In particular, there are numerous T cells of both major T-cell subsets (CD4/ and CD8/ cells) in the cellular infiltrates. Moreover, major / 5E1C$$0038
04-18-97 17:39:23
gasa
histocompatibility complex (MHC) class II expression is up-regulated in tissue specimens of patients with chronic pancreatitis.2 Our understanding of the molecular mechanisms involved in cell-mediated cytotoxicity has increased dramatically during the past few years. Experiments performed with gene-deficient mice lead to the elucidation of at least two major pathways of cell-mediated cytotoxicity. The first pathway requires exocytosis of several proteins, including the pore-forming protein perforin and one or several members of a serine protease family called granzymes. The importance of granule exocytosis in cellmediated cytotoxicity in vivo has been clearly shown in perforin-deficient3 and granzyme B–deficient4 mice. Because genes coding for granzymes and perforin are generally expressed only after specific activation, detection of cells containing transcripts of these genes is often used as a marker to specifically localize activated cytotoxic and natural killer (NK) cells in situ.5 The second major pathway that is reported to be perforin independent and does not require exocytosis can be mainly attributed to Fas-mediated induction of apoptosis in target cells.6 In the present study, we compare the presence and localization of activated cytotoxic cells defined by the expression of perforin messenger RNA (mRNA) in pancreatic tissue sections from patients with chronic pancreatitis and from normal controls. We attempted to define the preferential phenotype of these perforin mRNA–expressing cells to obtain information on the mode of antigen recognition involved in this activation process.
Materials and Methods Patients All patients included in this study had chronic pancreatitis and a long history of excessive alcohol consumption. The clinical data and patient characteristics are shown in Table 1. Abbreviations used in this paper: FITC, fluorescein isothiocyanate; MHC, major histocompatibility complex; NK, natural killer. q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00
WBS-Gastro
May 1997
CYTOTOXIC T CELLS IN CHRONIC PANCREATITIS 1657
Table 1. Clinical and Histopathologic Data of Patients With Alcoholic Chronic Pancreatitis Duration of pain (yr)
Staging (Cambridge)
Patient
Sex
Age (yr)
A B C D
F M M M
47 43 56 60
10 2 2 1
Marked Moderate Marked Marked
E F G H I J K L M N
M M M F F M M M M M
46 54 39 49 42 46 49 51 34 44
15 2 2 2 4 1 4 8 8 1
Moderate Marked Marked Marked Marked Marked Marked Marked Marked Moderate
Indication for surgery Pain, MPDS, RA CBS, IT, RA Pain, CBS, IT Bleeding, MPDS, IT, vessel alterations Pain, IT CBS, pain, IT, MPDS IT, CBS, MPDS, RA IT, pain, MPDS MPDS, pain, IT CBS, MPDS, IT, pain IT, vessel compression Pain, IT, weight loss IT, pain, MPDS Cyst
Degree of fibrosis ( %)
Degree of inflammation
Exocrine insufficiency
Endocrine insufficiency
69 71 79 84
2 2 2 2
Yes Yes Yes Yes
None None None None
68 58 98 90 97 96 64 83 98 81
2 2 2 2 3 3 2 2 3 2
Yes Yes Yes Yes Yes Yes Yes Yes Yes No
None SD SD SD None OD None OD OD None
NOTE. Clinical and histopathologic data of patients with alcoholic chronic pancreatitis were assessed as described in Materials and Methods. Exocrine insufficiency was determined using the fluorescein dilaurate serum test as described previously.26 CBS, common bile duct stenosis; IT, inflammatory tumor (pancreatic head enlargement of ú4 cm); MPDS, main pancreatic duct stenosis; OD, clinical overt diabetes mellitus; RA, recurrent attacks; SD, subclinical diabetes mellitus.
For assessment of perforin mRNA expression on tissue sections, pancreatic tissue specimens from patients A–J (3 women and 7 men; mean age, 46.5 years) were used. For the isolation of pancreas-infiltrating lymphocytes, specimens from patients G– N (2 women and 6 men; mean age, 45 years) were processed. Normal tissue specimens of the pancreata of 7 multiorgan donors (3 women and 4 men; mean age, 37 years) and 2 patients (49-year-old woman and 61-year-old man) with pancreatic tumors were used as controls.
Histopathologic Evaluation of Tissue Sections On all tissue sections, the extent of fibrosis was assessed by an image analysis system as described previously.7 Briefly, a high-resolution color video camera CF20DX with CFDX camera control 0.98 software (KAPPA; Messtechnik GmbH, Gleichen, Germany) on a light microscope (Leitz DMRB; Leica Microskopie und Systeme GmbH, Wetzlar, Germany) was used to acquire the picture. The area occupied by fibrotic tissue was determined on an AST Power Premium Computer GX P/ 90 (AST Research Inc., Irvine, CA) using Image-Pro Plus for Windows software (version 1.3; Media Cybernetics, Silver Spring, MD). The extent of inflammatory infiltrates was graded according to the following score system: 0, no infiltrates; 1, limited; 2 and 3, moderate; and 4, high numbers of infiltrating cells. As shown in Table 1, all patients included in this study showed moderate inflammatory infiltration (scores 2 and 3).
Tissue Preparation Tissue specimens of chronic pancreatitis were obtained at surgical procedures. From each freshly removed specimen, one part was embedded in OCT compound (Miles, Elkhart, IN) and stored at 0707C, until use for immunostainings. An-
/ 5E1C$$0038
04-18-97 17:39:23
gasa
other part was immersed in freshly prepared 4% paraformaldehyde (in phosphate-buffered saline [PBS]), heated in a microwave oven as described previously,8 and embedded with paraffin by routine techniques for further processing for in situ hybridization. The remaining pancreatic tissue was used for the isolation of infiltrating cells as described below.
Preparation of
35
S-Labeled RNA Probes
A 1953–base pair complementary DNA fragment of the human perforin gene9 provided by J. Tschopp, (Institute of Biochemistry, University of Lausanne, Lausanne, Switzerland), cloned into the expression vector pBluescript SK, was used to prepare 35S-labeled antisense and sense RNA probes using T3or T7-RNA polymerase reaction, respectively, as previously described in detail.10
In Situ Hybridization In situ hybridization was performed as previously described in detail.10 Briefly, sections of paraffin-embedded tissue specimens were dewaxed with xylene and fixed in 4% paraformaldehyde in PBS for 20 minutes at room temperature, rinsed in PBS, and subsequently treated with proteinase K (Boehringer, Mannheim, Germany) at 1 mg/mL at 377C for 30 minutes. After refixation and acetylation, hybridization was performed with 2 1 105 cpm of 35S-labeled RNA probe per microliter of hybridization solution for 18 hours at 487C. After digestion of single-stranded, nonhybridized RNA with ribonuclease A and T1 and extensive washing, slides were dipped in NTB2 nuclear track emulsion (Eastman Kodak, New Haven, CT). Slides were exposed for 28 days in a light-tight box at 47C, developed, and subsequently counterstained with nuclear fast red (0.05% in 5% aluminum sulfate) by standard techniques.
WBS-Gastro
1658 HUNGER ET AL.
GASTROENTEROLOGY Vol. 112, No. 5
Immunofluorescence Perforin was detected by immunofluorescence on cryostat sections using fluorescein isothiocyanate (FITC)-labeled anti-human perforin monoclonal antibody clone 2d4.11 FITClabeled isotype matched mouse monoclonal antibodies were used as a negative control.
Isolation of Infiltrating Cells Tissue specimens of patients with chronic pancreatitis were teased into small pieces and incubated in 20 mL PBS containing 5% horse serum and 1 mg/mL collagenase type V (Sigma Chemical Co., St. Louis, MO) on a magnetic stirrer. After 10 minutes, 10 mL of the supernatant was removed and 10 mL fresh PBS containing 5% horse serum and 1 mg/mL collagenase type V was added. This step was repeated four to five times. Supernatants with isolated cells were pooled, passed through a 40-mm cell strainer (Falcon; Becton Dickinson, San Jose, CA), and washed with cold PBS containing 5% horse serum. After centrifugation, cells were resuspended in 38% Percoll (Pharmacia Biotech, Uppsala, Sweden), underlaid with 73% Percoll, overlaid with 19% Percoll, and centrifuged at 900g for 15 minutes at room temperature. Cells in the lower interphase were collected, washed, and resuspended for staining.
Staining and Fluorescence-Activated Cell Sorter Sorting of Isolated Cells Isolated cells were resuspended in PBS containing 5% horse serum and 0.05% NaN3 and stained for 30 minutes on ice with directly conjugated antibodies against CD4, CD8, CD56 (NK cells), CD19 (pan-B cell marker), and CD45 (leukocyte common antigen) (all obtained from Pharmingen, San Diego, CA). After completion of the staining procedure, cells were fixed in freshly prepared 4% buffered paraformaldehyde for 5 minutes and resuspended in PBS. A FACScan (Becton Dickinson) was used to analyze the isolated cells, and a FACSVantage (Becton Dickinson) was used for the separation of CD4-, CD8-, and CD56-positive cell subsets after gating on the lymphoid cell population in the forward and side light scatter of the isolated cell population. Cells were directly sorted onto poly-L-lysine–coated microscopy slides (Menzel Gla¨ser, Braunschweig, Germany) using lysis II software (Becton Dickinson). In situ hybridization of sorted cells was performed as described above for tissue sections.
Statistics Data are expressed as means { SD. The Mann–Whitney rank sum U test was used to evaluate the data, and differences were considered to be significant with P values of õ0.05.
Results Histopathologic Evaluation The extent of fibrosis determined in the pancreatic tissue specimens of all patients included in this study is / 5E1C$$0038
04-18-97 17:39:23
gasa
shown in Table 1. All tissue specimens analyzed showed extended areas with fibrotic tissue and circumscribed areas with remaining pancreatic parenchyma. Cellular infiltrates predominantly consisting of lymphocytes were consistently present in the dense fibrous tissue. The degree of inflammatory infiltrates was scored as described in Materials and Methods. All patients showed comparable amounts of inflammatory infiltrates (Table 1). No histological signs of acute inflammation were found. A representative example of a tissue section is shown in Figure 1A. The control specimens from 7 multiorgan donors and from 2 patients with pancreatic tumors showed normal pancreatic tissue without signs of histopathologic alterations. Phenotypic Analysis of Isolated Infiltrating Cells To determine the preferential phenotype of the infiltrating lymphocytes, cells were isolated from tissue specimens of patients with chronic pancreatitis and stained with the appropriate primary antibodies for CD4, CD8, CD56, a marker for NK cells, as well as CD19, a pan B-cell marker, and CD45, a leukocyte common antigen. The results of the subsequent analysis by FACScan are shown in Table 2. The majority of isolated cells are CD45/ (54%–95.3%). Within the population of CD45/ cells, the large majority of cells are either CD4 or CD8 positive T cells. In patients with chronic pancreatitis, CD4 T cells represent 30.9% (SD, {11.6%) and CD8 T cells represent 26.1% (SD, {12.0%) of the CD45/ infiltrating cells. The frequency of CD45//CD56/ cells, generally referred to as NK cells, was comparatively low with a mean frequency of 6.8% (SD, {7.4%). To determine the phenotype of the remaining CD45/ cells, isolated cells from 2 patients were also stained for CD19 expression. In the lymphoid cell population isolated from tissue specimens from patients H and L, 7.9% and 17.5%, respectively, expressed CD19, indicating that the remaining CD45/ cells, negative for CD4, CD8, and CD56, were almost exclusively B cells. Detection of Perforin-Expressing Cells In Situ To assess the contribution of activated cytotoxic cells in tissue destruction, we performed in situ hybridization with a 35S-labeled probe to detect perforin mRNA. Whereas in normal pancreatic tissues only a few perforin mRNA expressing cells are present, as shown in Figure 2, all tissue specimens of patients with chronic pancreatitis showed high frequencies of perforin mRNA– expressing cells (Figure 2, alcoholic pancreatitis). Perforin mRNA–expressing cells are predominantly localWBS-Gastro
May 1997
CYTOTOXIC T CELLS IN CHRONIC PANCREATITIS 1659
Figure 1. Representative pancreatic tissue sections (A–C and E and F ) from patients with alcoholic pancreatitis and (D ) from a multiorgan donor as a normal control. (A) Stained with H&E. (B and D ) Hybridized with a radioactively labeled antisense probe for the specific detection of perforin mRNA. (C ) Hybridized with a sense probe (negative control). (E ) Immunofluorescence staining with FITC-labeled anti-human perforin antibody. (F ) Staining with an FITC-labeled irrelevant isotype-matched antibody.
ized close to pancreatic ducts, within dense fibrous tissue as patchy infiltrates and in transitional areas between remaining intact pancreatic and fibrous tissue. Figure 1B and D show representative examples of tissue sections hybridized with a 35S-labeled antisense RNA probe for the detection of perforin mRNA. All pancreatic tissue sections of patients with alcoholic pancreatitis (Figure 1B) show specific signals for perforin in a similar pattern, / 5E1C$$0038
04-18-97 17:39:23
gasa
although at variable frequencies, whereas no specific signal is detected in the pancreatic tissue sections of multiorgan donors (Figure 1D). In addition, tissue sections were hybridized with a sense probe. No specific accumulation of silver grains is detected on these negative controls (Figure 1C). As an internal control, perforin mRNA expression has been compared between tissue near the surface of the specimen and deeper areas. No WBS-Gastro
1660 HUNGER ET AL.
GASTROENTEROLOGY Vol. 112, No. 5
Table 2. Surface Phenotype of Isolated Pancreas-Infiltrating Lymphoid Cells Patient
CD4 (%)
CD8 (%)
CD56 ( %)
CD45 ( %)
G H I J K L M N
17.0 33.8 47.3 35.4 12.0 32.9 40.0 28.5
17.6 13.2 19.5 23.6 50.0 19.1 31.5 34.6
ND 2.7 4.1 7.2 21.0 5.5 ND ND
ND 54.0 88.4 95.3 78.2 72.1 89.1 66.2
NOTE. Isolated cells were gated on a FACScan by side and forward light scatter for the lymphoid cell population, and the relative frequencies of the indicated cell populations were determined subsequently. ND, not determined.
difference in the frequency of perforin mRNA–expressing cells can be detected with this analysis. Furthermore, the maximum number of silver grains, reflecting the maximum amount of RNA detected in a given cell, is comparable in all tissue specimens from patients with chronic pancreatitis (data not shown). Detection of perforin-containing cells in infiltrates of tissue specimens of patients with chronic pancreatitis by immunofluorescence (Figure 1E) showed an identical distribution pattern as found by in situ hybridization for perforin mRNA. Stainings with isotype-matched antibodies showed no specific signal (Figure 1F).
Table 3. Frequency of Perforin mRNA–Expressing Cells in the Respective Cell Subsets Patient
CD4 ( %)
CD8 ( %)
CD56 ( %)
G H I J K L M
0.9 1.0 1.0 1.7 1.2 0.0 1.0
23.9 10.6 7.2 9.4 12.5 ND 8.0
ND 23.1 9.3 35.0 6.1 16.6 ND
NOTE. The indicated cell populations were sorted on a FACSvantage, and in situ hybridization was performed. The percentage of cells expressing perforin mRNA was determined by counting positive cells in the corresponding cell populations. ND, not determined.
Perforin mRNA Expression in T-Cell Subsets To determine the preferential phenotype of the perforin mRNA–positive cells, isolated cells from tissue samples from patients with chronic pancreatitis were stained with the appropriate primary antibodies for CD4, CD8, and CD56 for subsequent sorting of the respective cell subpopulations. Cells were directly sorted onto polyL-lysine–coated glass slides and hybridized in situ with perforin probes. This procedure is required because of the different fixation requirements of in situ hybridization and immunohistochemistry, thus leading to a significant loss of perforin mRNA on tissue sections immunohistochemically stained before in situ hybridization (data not shown). The frequency of perforin mRNA– expressing cells is shown in Table 3. Although perforin mRNA is expressed only in a minor fraction of the CD4 T cells (1.0%; SD, {0.5%), high frequencies of CD8 (11.9%; SD, {6.2%) and CD56/ (18.0%; SD, {11.6%) cells express this mRNA at detectable levels. The comparison of the frequencies of perforin mRNA– expressing cells obtained by enumerating in situ hybridizations of tissue sections and by analyzing isolated and sorted cells shows a good correlation (correlation coefficient, 0.92; numerical data not shown).
Discussion
Figure 2. Perforin mRNA–expressing cells per square millimeter of tissue section in patients with alcoholic pancreatitis and normal controls. The mean ({SD) of all values in each group is indicated on the right. The differences between the two groups are statistically significant (P õ 0.0005).
/ 5E1C$$0038
04-18-97 17:39:23
gasa
Despite numerous hypotheses on the etiopathogenesis of chronic pancreatitis, the pathogenetic steps of this disease still remain enigmatic. In developed countries, most patients with chronic pancreatitis (60%– 70%) have a long history of alcohol abuse before clinical onset of disease.12 With the exception of studies reporting an increased MHC class II expression in affected tissues from patients with chronic pancreatitis and the observation of T cell–enriched cellular infiltrates, studies WBS-Gastro
May 1997
CYTOTOXIC T CELLS IN CHRONIC PANCREATITIS 1661
on the contribution of the immune system, either humoral or cellular in nature, to the pathogenesis of chronic pancreatitis in general and to alcoholic pancreatitis in particular are limited. Phenotypical analysis on immunohistochemically stained pancreas sections of patients with chronic alcoholic pancreatitis showed numerous T cells. Both major T-cell subsets, CD4 and CD8 T cells, were, in general, present at comparable frequencies (data not shown). These data are further substantiated by the cytofluorometric analysis of isolated pancreas-infiltrating cells (Table 2) in which the mean of the CD4/CD8 ratio is 1.4:1.0. However, the variation of the CD4/CD8 ratio determined in isolated pancreas-infiltrating cells is considerable. This variation was also observed on pancreatic tissue sections from the same patients, stained immunohistochemically with the respective reagents. Thus, selective loss or enrichment of a given T-cell subset during the isolation procedure is very unlikely. CD56/ cells were consistently found in small numbers (Table 2). They have been described to represent mostly NK cells and large granular lymphocytes that kill target cells in an MHC-independent manner.13,14 Thus, CD56 is often used as a phenotypic marker for NK cells. The CD450 cells found in the lymphoid cell population, as defined by gating in the forward and side light scatter, are likely to represent contaminating cells, either fibrocytes, fibroblasts, or acinar cells. Indirect evidence for this assumption is provided by the finding that the relative frequency of CD450 cells was considerably higher before lymphocyte enrichment of the isolated cell populations. In situ hybridization of sections of pancreatic tissue from patients with chronic pancreatitis consistently showed a high frequency of perforin mRNA–positive cells in cellular infiltrates (Figure 2). All patients analyzed showed clearly higher frequencies of activated cytotoxic cells than controls. The presence of perforin was also confirmed on the protein level by immunohistochemistry and consistently reflected the results obtained by in situ hybridization (Figure 1B and E). Together with the finding that these activated cells are often found in contact to remaining parenchyma, this might indicate an involvement of these cells in tissue destruction. Perforin mRNA expression in vivo is reported only in activated CD4- and CD8-positive T cells and NK cells. The relative frequencies of perforin mRNA in the CD4-, CD8-, and CD56-positive cell subsets (Table 3) showed that the majority of perforin mRNA–expressing cells were found within the CD8/ T-cell subset, whereas in the CD56/ and CD4/ subsets only 27% and 13%, respectively, of all perforin mRNA–positive cells were found. Thus, the majority of perforin-mediated killing / 5E1C$$0038
04-18-97 17:39:23
gasa
can be attributed to MHC class I–dependent killing, a minor but significant part (CD56/ NK cells) to an MHC unrestricted killing, and only a small part to MHC class II–dependent cytotoxicity. Besides this perforin-dependent mechanism of cell-mediated cytotoxicity, perforinindependent cytotoxic effector mechanisms, in particular Fas/FasL interaction, also have been described recently.6,15 However, for technical reasons, the contribution of this latter mechanism, which appears to be mainly operative in CD4/ cytotoxic T cells, cannot be assessed in this study. Furthermore, this perforin-independent mechanism seems to be mainly involved in regulation of the immune response rather than in effector mechanisms in host defense or autoimmune tissue destruction.16 As shown in Figure 2, there are large variations between the different tissue samples analyzed. To assess whether these variations are caused by technical artifacts or truly represent actual differences in the extent of cellular activation, additional hybridizations and evaluations of the tissue sections hybridized with perforin antisense probes have been performed. The observed comparable frequencies of cells expressing perforin mRNA at detectable levels in the superficial and deeper areas of pancreatic tissues indicate that the methods used for tissue processing allowed for a homogenous tissue fixation and preservation of mRNA within the entire tissue specimens. The observation that comparable maximum expression levels of perforin mRNA in positive cells are found in all tissue specimens analyzed from patients with chronic pancreatitis further indicates that, because of the standardized procedure of tissue processing, only minimal variations in the preservation of mRNA occurred between different tissue specimens. This latter aspect is further substantiated by the result of control hybridization with an insulin RNA antisense probe, which showed comparable signals for the presence of insulin mRNA on tissue sections from patients with chronic pancreatitis with intact islets of Langerhans (data not shown). Thus, taken together, the observed variations in the frequency of perforin mRNA–expressing cells in tissue specimens from patients with chronic pancreatitis very likely reflect different activation states of cytotoxic effector cells. The observed variations in the expression of the perforin gene at consistently high levels in all tissue samples examined are reminiscent of acute episodes of activation of the cellular immune response, as has been shown in tissue destruction during viral infections17 and transplant rejection.18 It will be of particular interest to define the nature of the triggering antigens that leads to these recurrent episodes of activation of mainly CD8-positive, MHC class I–restricted cytolytic T cells. Because antigens presented by MHC class I molecules are derived from endogWBS-Gastro
1662 HUNGER ET AL.
GASTROENTEROLOGY Vol. 112, No. 5
enously produced proteins, such as cellular proteins, or virally encoded antigens, one might speculate that recurrent virus infections or an altered pattern of endogenously produced proteins, recognized by virus-specific or autoreactive T cells, induce this cellular activation. In this context it will be of foremost interest to analyze whether the cellular activation events operative during episodes of acute pancreatitis resemble the recurrent activation episodes in chronic pancreatitis described in the present study. Such an analysis might clarify the controversy of whether recurrent episodes of alcoholic acute pancreatitis will eventually lead to alcoholic chronic pancreatitis, as has been hypothesized previously.19 – 21 However, because of the lack of appropriate tissue specimens from patients with acute pancreatitis, such an analysis could not be included in this study. It will be of further interest to assess whether cell-mediated cytotoxicity is also operative in tissue destruction of other alcohol-associated disorders. The finding that numerous CD4 and CD8 T cells infiltrate the lesions in alcoholic liver disease and that these cells persist during the progression of disease may suggest that cell-mediated cytotoxicity is also operative in destruction of the liver parenchyma.22 However, further studies of perforin expression and T-cell function are required to be able to compare the contribution of cellmediated cytotoxicity in tissue destruction in these disorders. Furthermore, experiments comparing the vb usage of the T-cell receptors in the activated infiltrating T cells with the vb usage of blood T lymphocytes would give more insight in the clonality of the T-cell response. Such studies could clarify if only a few antigens are responsible for the T-cell activation or if a broad range of antigens provoke a polyclonal T-cell response. Activated cytotoxic cells of both the CD8 and the CD56 subsets may influence the course of disease not only through their potent cell-mediated cytotoxic capacities but also through the secretion of lymphokines, in particular of interferon gamma.23 Together with the reported up-regulation of transforming growth factor b in the inflammatory process,24 this may significantly affect the functional differentiation of CD4/ T cells. Both cytokines are known to promote the expansion of Th1-type T cells, leading to a predominant cellular immune response.25 Further evidence that an imbalance in the immune system is of etiopathological importance in chronic pancreatitis has been recently provided in an animal model in which the absence of the immunoregulatory molecule CTLA-4 leads to a chronic pancreatitis-like disease.26 In conclusion, the observed increased frequency of perforin mRNA–expressing cells in the pancreas of patients with alcoholic chronic pancreatitis indicates a preferential / 5E1C$$0038
04-18-97 17:39:23
gasa
activation of these effector cells of cell-mediated cytotoxicity. The consistent observation of a close association between activated cytotoxic cells with remaining intact pancreatic parenchyma may indicate that these cells are directly involved in the destruction of parenchymal cells. It may also indicate in situ activation of these cells, thus providing circumstantial evidence for an involvement of autoreactive cytotoxic T cells and NK cells in the pathogenesis of this disease.
References 1. Steer ML, Waxman I, Freedman S. Chronic pancreatitis. N Engl J Med 1995;332:1482–1490. 2. Bedossa P, Bacci J, Lemaigre G, Martin E. Lymphocyte subsets and HLA-DR expression in normal pancreas and chronic pancreatitis. Pancreas 1989;5:415–420. 3. Ka¨gi D, Ledermann B, Bu¨rki K, Seiler P, Odermatt B, Olsen KJ, Podack ER, Zinkernagel RM, Hengartner H. Cytotoxicity mediated by T cells and natural killer cells is greatly impaired in perforindeficient mice. Nature 1994;369:31–37. 4. Heusel JW, Wesselschmidt RL, Shresta S, Russel JH, Ley TJ. Cytotoxic lymphocytes require granzyme B for rapid induction of DNA fragmentation and apoptosis in allogeneic target cells. Cell 1994;76:977–987. 5. Griffiths G, Mueller C. Expression of perforin and granzymes in vivo: potential diagnostic markers for activated cytotoxic cells. Immunol Today 1991;12:415–419. 6. Henkart PA. Lymphocyte-mediated cytotoxicity: two pathways and multiple effector molecules. Immunity 1994;1:343–346. 7. Friess H, Malfertheiner P, Isenmann R, Ku¨hne H, Beger HG, Bu¨chler MW. The risk of pancreaticointestinal anastomosis can be predicted preoperatively. Pancreas 1996;13:202–208. 8. Held W, MacDonald HR, Weissman IL, Hess MW, Mueller C. Genes encoding tumor necrosis factor a and granzyme A are expressed during development of autoimmune diabetes. Proc Natl Acad Sci USA 1990;87:2239–2243. 9. Lichtenheld MG, Olsen KJ, Lu P, Lowrey DM, Hameed A, Hengartner H, Podack ER. Structure and function of human perforin. Nature 1988;335:448–451. 10. Mueller C, Gershenfeld HK, Lobe CG, Okada CY, Bleackley RC, Weissman IL. A high proportion of T lymphocytes that infiltrate H-2 incompatible heart allografts in vivo express genes encoding cytotoxic cell–specific serine proteases, but do not express the MEL-14-defined lymph node homing receptor. J Exp Med 1988; 167:1124–1136. 11. Baetz K, Isaaz S, Griffiths GM. Loss of cytotoxic T lymphocyte function in Chediak–Higashi syndrome arises from a secretory defect that prevents lytic granule exocytosis. J Immunol 1995; 154:6122–6131. 12. Sarles H, Sahel J, Staub JL, Bourry J, Laugier R. Chronic pancreatitis. In: Howat HT, Sarles H, eds. The exocrine pancreas. Philadelphia: Saunders, 1979:402–439. 13. Lu PH, Negrin RS. A novel population of expanded human CD3/CD56/ cells derived from T cells with potent in vivo antitumor activity in mice with severe combined immunodeficiency. J Immunol 1994;153:1687–1696. 14. Metha BA, Schmidt-Wolf IGH, Weissman IL, Negrin RS. Two pathways of exocytosis of cytoplasmic granule contents and target cell killing by cytokine-induced CD3/CD56/ killer cells. Blood 1995;68:3493–3499. 15. Stalder T, Hahn S, Erb P. Fas antigen is the major target molecule for CD4/ T cell–mediated cytotoxicity. J Immunol 1994;152: 1127–1133. 16. Ka¨gi D, Ledermann B, Bu¨rki K, Zinkernagel RM, Hengartner H.
WBS-Gastro
May 1997
17.
18.
19.
20. 21.
22.
CYTOTOXIC T CELLS IN CHRONIC PANCREATITIS 1663
Molecular mechanisms of lymphocyte-mediated cytotoxicity and their role in immunological protection and pathogenesis in vivo. Annu Rev Immunol 1996;14:207–232. Mueller C, Ka¨gi D, Aebischer T, Odermatt B, Held W, Podack ER, Zinkernagel RM, Hengartner H. Detection of perforin and granzyme A mRNA in infiltrating cells during infection of mice with lymphocytic choriomeningitis virus. Eur J Immunol 1989;19: 1253–1259. Mueller C, Shelby J, Weissman IL, Perinat Frey T, Eichwald EJ. Expression of the protease gene HF as a marker in rejecting allogeneic murine heart transplants. Transplantation 1991;51: 514–517. Ammann RW, Heitz PU, Klo¨ppel G. Course of alcoholic chronic pancreatitis: a prospective clinicomorphological long-term study. Gastroenterology 1996;111:224–231. Amman RW, Muellhaupt B. Progression of alcoholic acute to chronic pancreatitis. Gut 1994;35:552–556. Klo¨ppel G, Maillet B. The morphological basis for the evolution of acute pancreatitis into chronic pancreatitis. Virchows Arch A Pathol Anat Histopathol 1992;420:1–4. Chedid A, Mendenhall CL, Moritz TE, French SW, Chen TS, Morgan TR, Roselle GA, Nemchausky BA, Tamburro CH, Schiff ER, McClain CJ, Marsano LS, Allen JI, Samanta A, Weesner RE, Hen-
/ 5E1C$$0038
04-18-97 17:39:23
gasa
23. 24.
25. 26.
27.
derson WG, VAC Study Group 275. Cell-mediated hepatic injury in alcoholic liver disease. Gastroenterology 1993;105:154–266. Farrar MA, Schreiber RD. The molecular cell biology of interferong and its receptor. Annu Rev Immunol 1993;11:571–611. Van Laethem JL, Deviere J, Resibois A, Rickaert F, Vertongen P, Ohtani H, Cremer M, Miyazono K, Robberecht P. Localization of transforming growth factor b1 and its latent binding protein in human chronic pncreatitis. Gastroenterology 1995;108:1873– 1881. Romagnani S. Lymphokine production by human T cells in disease states. Annu Rev Immunol 1994;12:227–257. Waterhouse P, Penninger JM, Timms E, Wakeham A, Shahinian A, Lee KP, Thompson CB, Griesser H, Mak TW. Lymphoproliferative disorders with early lethality in mice deficient in Ctla-4. Science 1995;270:985–988. Malfertheiner P, Bu¨chler M, Mu¨ller A, Ditschuneit H. Fluorescein dilaurate serum test: a rapid tubeless pancreatic function test. Pancreas 1987;2:53–60.
Received November 5, 1996. Accepted February 11, 1997. Address requests for reprints to: Markus W. Bu¨chler, M.D., Department of Visceral and Transplantation Surgery, University of Bern, Inselspital, CH-3010 Bern, Switzerland. Fax: (41) 31-382-4772.
WBS-Gastro
Cytotoxic Cells Are Activated in Cellular Infiltrates of Alcoholic Chronic Pancreatitis ROBERT E. HUNGER,* CHRISTOPH MUELLER,* KASPAR Z’GRAGGEN,‡ HELMUT FRIESS,‡ ¨ CHLER‡ and MARKUS W. BU Departments of *Pathology and ‡Visceral and Transplantation Surgery, University of Bern, Bern, Switzerland
See editorial on page 1762. Background & Aims: Perforin messenger RNA (mRNA) expression has been shown to be a specific in vivo activation marker for cytotoxic cells. In this study, the contribution of cell-mediated cytotoxicity in the pathogenesis of alcoholic chronic pancreatitis is assessed. Methods: Tissue sections of patients with alcoholic chronic pancreatitis were analyzed for perforin mRNA expression by in situ hybridization. In a further step, the phenotype and the relative frequency of perforin mRNA–expressing cells were determined. Results: In the normal pancreas, perforin mRNA–expressing cells are rarely present (mean, 0.3 cells/mm2). In contrast, the frequency of perforin mRNA–expressing cells is increased severalfold in diseased tissue specimens (mean, 6.6 cells/mm2). The frequency of perforin mRNA–expressing cells is high in the CD56/ (18%) and CD8/ cell population (12%) but low in the CD4/ cell population (1%). Conclusions: The significantly elevated frequencies of perforin mRNA–expressing cells in the pancreas of patients with alcoholic chronic pancreatitis suggest an involvement of cell-mediated cytotoxicity in the pathogenesis of this disease. The preferential localization of these activated cells close to areas with parenchyma provides circumstantial evidence that autoreactive cytotoxic cells may contribute to tissue destruction in alcoholic chronic pancreatitis.
C
hronic pancreatitis is a disease characterized by irreversible destruction of the pancreas and replacement of the parenchyma by a dense fibrous tissue containing numerous lymphocytes. This process may lead to endocrine and exocrine insufficiencies. In developed countries, 60%–70% of the patients have a long history of heavy consumption of alcohol.1 Information on the mechanisms operative during tissue destruction in chronic pancreatitis is very limited. However, some observations seem to indicate an involvement of cell-mediated cytotoxicity in the pathogenesis of this disease. In particular, there are numerous T cells of both major T-cell subsets (CD4/ and CD8/ cells) in the cellular infiltrates. Moreover, major / 5E1C$$0038
04-18-97 17:39:23
gasa
histocompatibility complex (MHC) class II expression is up-regulated in tissue specimens of patients with chronic pancreatitis.2 Our understanding of the molecular mechanisms involved in cell-mediated cytotoxicity has increased dramatically during the past few years. Experiments performed with gene-deficient mice lead to the elucidation of at least two major pathways of cell-mediated cytotoxicity. The first pathway requires exocytosis of several proteins, including the pore-forming protein perforin and one or several members of a serine protease family called granzymes. The importance of granule exocytosis in cellmediated cytotoxicity in vivo has been clearly shown in perforin-deficient3 and granzyme B–deficient4 mice. Because genes coding for granzymes and perforin are generally expressed only after specific activation, detection of cells containing transcripts of these genes is often used as a marker to specifically localize activated cytotoxic and natural killer (NK) cells in situ.5 The second major pathway that is reported to be perforin independent and does not require exocytosis can be mainly attributed to Fas-mediated induction of apoptosis in target cells.6 In the present study, we compare the presence and localization of activated cytotoxic cells defined by the expression of perforin messenger RNA (mRNA) in pancreatic tissue sections from patients with chronic pancreatitis and from normal controls. We attempted to define the preferential phenotype of these perforin mRNA–expressing cells to obtain information on the mode of antigen recognition involved in this activation process.
Materials and Methods Patients All patients included in this study had chronic pancreatitis and a long history of excessive alcohol consumption. The clinical data and patient characteristics are shown in Table 1. Abbreviations used in this paper: FITC, fluorescein isothiocyanate; MHC, major histocompatibility complex; NK, natural killer. q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00
WBS-Gastro
May 1997
CYTOTOXIC T CELLS IN CHRONIC PANCREATITIS 1657
Table 1. Clinical and Histopathologic Data of Patients With Alcoholic Chronic Pancreatitis Duration of pain (yr)
Staging (Cambridge)
Patient
Sex
Age (yr)
A B C D
F M M M
47 43 56 60
10 2 2 1
Marked Moderate Marked Marked
E F G H I J K L M N
M M M F F M M M M M
46 54 39 49 42 46 49 51 34 44
15 2 2 2 4 1 4 8 8 1
Moderate Marked Marked Marked Marked Marked Marked Marked Marked Moderate
Indication for surgery Pain, MPDS, RA CBS, IT, RA Pain, CBS, IT Bleeding, MPDS, IT, vessel alterations Pain, IT CBS, pain, IT, MPDS IT, CBS, MPDS, RA IT, pain, MPDS MPDS, pain, IT CBS, MPDS, IT, pain IT, vessel compression Pain, IT, weight loss IT, pain, MPDS Cyst
Degree of fibrosis ( %)
Degree of inflammation
Exocrine insufficiency
Endocrine insufficiency
69 71 79 84
2 2 2 2
Yes Yes Yes Yes
None None None None
68 58 98 90 97 96 64 83 98 81
2 2 2 2 3 3 2 2 3 2
Yes Yes Yes Yes Yes Yes Yes Yes Yes No
None SD SD SD None OD None OD OD None
NOTE. Clinical and histopathologic data of patients with alcoholic chronic pancreatitis were assessed as described in Materials and Methods. Exocrine insufficiency was determined using the fluorescein dilaurate serum test as described previously.26 CBS, common bile duct stenosis; IT, inflammatory tumor (pancreatic head enlargement of ú4 cm); MPDS, main pancreatic duct stenosis; OD, clinical overt diabetes mellitus; RA, recurrent attacks; SD, subclinical diabetes mellitus.
For assessment of perforin mRNA expression on tissue sections, pancreatic tissue specimens from patients A–J (3 women and 7 men; mean age, 46.5 years) were used. For the isolation of pancreas-infiltrating lymphocytes, specimens from patients G– N (2 women and 6 men; mean age, 45 years) were processed. Normal tissue specimens of the pancreata of 7 multiorgan donors (3 women and 4 men; mean age, 37 years) and 2 patients (49-year-old woman and 61-year-old man) with pancreatic tumors were used as controls.
Histopathologic Evaluation of Tissue Sections On all tissue sections, the extent of fibrosis was assessed by an image analysis system as described previously.7 Briefly, a high-resolution color video camera CF20DX with CFDX camera control 0.98 software (KAPPA; Messtechnik GmbH, Gleichen, Germany) on a light microscope (Leitz DMRB; Leica Microskopie und Systeme GmbH, Wetzlar, Germany) was used to acquire the picture. The area occupied by fibrotic tissue was determined on an AST Power Premium Computer GX P/ 90 (AST Research Inc., Irvine, CA) using Image-Pro Plus for Windows software (version 1.3; Media Cybernetics, Silver Spring, MD). The extent of inflammatory infiltrates was graded according to the following score system: 0, no infiltrates; 1, limited; 2 and 3, moderate; and 4, high numbers of infiltrating cells. As shown in Table 1, all patients included in this study showed moderate inflammatory infiltration (scores 2 and 3).
Tissue Preparation Tissue specimens of chronic pancreatitis were obtained at surgical procedures. From each freshly removed specimen, one part was embedded in OCT compound (Miles, Elkhart, IN) and stored at 0707C, until use for immunostainings. An-
/ 5E1C$$0038
04-18-97 17:39:23
gasa
other part was immersed in freshly prepared 4% paraformaldehyde (in phosphate-buffered saline [PBS]), heated in a microwave oven as described previously,8 and embedded with paraffin by routine techniques for further processing for in situ hybridization. The remaining pancreatic tissue was used for the isolation of infiltrating cells as described below.
Preparation of
35
S-Labeled RNA Probes
A 1953–base pair complementary DNA fragment of the human perforin gene9 provided by J. Tschopp, (Institute of Biochemistry, University of Lausanne, Lausanne, Switzerland), cloned into the expression vector pBluescript SK, was used to prepare 35S-labeled antisense and sense RNA probes using T3or T7-RNA polymerase reaction, respectively, as previously described in detail.10
In Situ Hybridization In situ hybridization was performed as previously described in detail.10 Briefly, sections of paraffin-embedded tissue specimens were dewaxed with xylene and fixed in 4% paraformaldehyde in PBS for 20 minutes at room temperature, rinsed in PBS, and subsequently treated with proteinase K (Boehringer, Mannheim, Germany) at 1 mg/mL at 377C for 30 minutes. After refixation and acetylation, hybridization was performed with 2 1 105 cpm of 35S-labeled RNA probe per microliter of hybridization solution for 18 hours at 487C. After digestion of single-stranded, nonhybridized RNA with ribonuclease A and T1 and extensive washing, slides were dipped in NTB2 nuclear track emulsion (Eastman Kodak, New Haven, CT). Slides were exposed for 28 days in a light-tight box at 47C, developed, and subsequently counterstained with nuclear fast red (0.05% in 5% aluminum sulfate) by standard techniques.
WBS-Gastro
1658 HUNGER ET AL.
GASTROENTEROLOGY Vol. 112, No. 5
Immunofluorescence Perforin was detected by immunofluorescence on cryostat sections using fluorescein isothiocyanate (FITC)-labeled anti-human perforin monoclonal antibody clone 2d4.11 FITClabeled isotype matched mouse monoclonal antibodies were used as a negative control.
Isolation of Infiltrating Cells Tissue specimens of patients with chronic pancreatitis were teased into small pieces and incubated in 20 mL PBS containing 5% horse serum and 1 mg/mL collagenase type V (Sigma Chemical Co., St. Louis, MO) on a magnetic stirrer. After 10 minutes, 10 mL of the supernatant was removed and 10 mL fresh PBS containing 5% horse serum and 1 mg/mL collagenase type V was added. This step was repeated four to five times. Supernatants with isolated cells were pooled, passed through a 40-mm cell strainer (Falcon; Becton Dickinson, San Jose, CA), and washed with cold PBS containing 5% horse serum. After centrifugation, cells were resuspended in 38% Percoll (Pharmacia Biotech, Uppsala, Sweden), underlaid with 73% Percoll, overlaid with 19% Percoll, and centrifuged at 900g for 15 minutes at room temperature. Cells in the lower interphase were collected, washed, and resuspended for staining.
Staining and Fluorescence-Activated Cell Sorter Sorting of Isolated Cells Isolated cells were resuspended in PBS containing 5% horse serum and 0.05% NaN3 and stained for 30 minutes on ice with directly conjugated antibodies against CD4, CD8, CD56 (NK cells), CD19 (pan-B cell marker), and CD45 (leukocyte common antigen) (all obtained from Pharmingen, San Diego, CA). After completion of the staining procedure, cells were fixed in freshly prepared 4% buffered paraformaldehyde for 5 minutes and resuspended in PBS. A FACScan (Becton Dickinson) was used to analyze the isolated cells, and a FACSVantage (Becton Dickinson) was used for the separation of CD4-, CD8-, and CD56-positive cell subsets after gating on the lymphoid cell population in the forward and side light scatter of the isolated cell population. Cells were directly sorted onto poly-L-lysine–coated microscopy slides (Menzel Gla¨ser, Braunschweig, Germany) using lysis II software (Becton Dickinson). In situ hybridization of sorted cells was performed as described above for tissue sections.
Statistics Data are expressed as means { SD. The Mann–Whitney rank sum U test was used to evaluate the data, and differences were considered to be significant with P values of õ0.05.
Results Histopathologic Evaluation The extent of fibrosis determined in the pancreatic tissue specimens of all patients included in this study is / 5E1C$$0038
04-18-97 17:39:23
gasa
shown in Table 1. All tissue specimens analyzed showed extended areas with fibrotic tissue and circumscribed areas with remaining pancreatic parenchyma. Cellular infiltrates predominantly consisting of lymphocytes were consistently present in the dense fibrous tissue. The degree of inflammatory infiltrates was scored as described in Materials and Methods. All patients showed comparable amounts of inflammatory infiltrates (Table 1). No histological signs of acute inflammation were found. A representative example of a tissue section is shown in Figure 1A. The control specimens from 7 multiorgan donors and from 2 patients with pancreatic tumors showed normal pancreatic tissue without signs of histopathologic alterations. Phenotypic Analysis of Isolated Infiltrating Cells To determine the preferential phenotype of the infiltrating lymphocytes, cells were isolated from tissue specimens of patients with chronic pancreatitis and stained with the appropriate primary antibodies for CD4, CD8, CD56, a marker for NK cells, as well as CD19, a pan B-cell marker, and CD45, a leukocyte common antigen. The results of the subsequent analysis by FACScan are shown in Table 2. The majority of isolated cells are CD45/ (54%–95.3%). Within the population of CD45/ cells, the large majority of cells are either CD4 or CD8 positive T cells. In patients with chronic pancreatitis, CD4 T cells represent 30.9% (SD, {11.6%) and CD8 T cells represent 26.1% (SD, {12.0%) of the CD45/ infiltrating cells. The frequency of CD45//CD56/ cells, generally referred to as NK cells, was comparatively low with a mean frequency of 6.8% (SD, {7.4%). To determine the phenotype of the remaining CD45/ cells, isolated cells from 2 patients were also stained for CD19 expression. In the lymphoid cell population isolated from tissue specimens from patients H and L, 7.9% and 17.5%, respectively, expressed CD19, indicating that the remaining CD45/ cells, negative for CD4, CD8, and CD56, were almost exclusively B cells. Detection of Perforin-Expressing Cells In Situ To assess the contribution of activated cytotoxic cells in tissue destruction, we performed in situ hybridization with a 35S-labeled probe to detect perforin mRNA. Whereas in normal pancreatic tissues only a few perforin mRNA expressing cells are present, as shown in Figure 2, all tissue specimens of patients with chronic pancreatitis showed high frequencies of perforin mRNA– expressing cells (Figure 2, alcoholic pancreatitis). Perforin mRNA–expressing cells are predominantly localWBS-Gastro
May 1997
CYTOTOXIC T CELLS IN CHRONIC PANCREATITIS 1659
Figure 1. Representative pancreatic tissue sections (A–C and E and F ) from patients with alcoholic pancreatitis and (D ) from a multiorgan donor as a normal control. (A) Stained with H&E. (B and D ) Hybridized with a radioactively labeled antisense probe for the specific detection of perforin mRNA. (C ) Hybridized with a sense probe (negative control). (E ) Immunofluorescence staining with FITC-labeled anti-human perforin antibody. (F ) Staining with an FITC-labeled irrelevant isotype-matched antibody.
ized close to pancreatic ducts, within dense fibrous tissue as patchy infiltrates and in transitional areas between remaining intact pancreatic and fibrous tissue. Figure 1B and D show representative examples of tissue sections hybridized with a 35S-labeled antisense RNA probe for the detection of perforin mRNA. All pancreatic tissue sections of patients with alcoholic pancreatitis (Figure 1B) show specific signals for perforin in a similar pattern, / 5E1C$$0038
04-18-97 17:39:23
gasa
although at variable frequencies, whereas no specific signal is detected in the pancreatic tissue sections of multiorgan donors (Figure 1D). In addition, tissue sections were hybridized with a sense probe. No specific accumulation of silver grains is detected on these negative controls (Figure 1C). As an internal control, perforin mRNA expression has been compared between tissue near the surface of the specimen and deeper areas. No WBS-Gastro
1660 HUNGER ET AL.
GASTROENTEROLOGY Vol. 112, No. 5
Table 2. Surface Phenotype of Isolated Pancreas-Infiltrating Lymphoid Cells Patient
CD4 (%)
CD8 (%)
CD56 ( %)
CD45 ( %)
G H I J K L M N
17.0 33.8 47.3 35.4 12.0 32.9 40.0 28.5
17.6 13.2 19.5 23.6 50.0 19.1 31.5 34.6
ND 2.7 4.1 7.2 21.0 5.5 ND ND
ND 54.0 88.4 95.3 78.2 72.1 89.1 66.2
NOTE. Isolated cells were gated on a FACScan by side and forward light scatter for the lymphoid cell population, and the relative frequencies of the indicated cell populations were determined subsequently. ND, not determined.
difference in the frequency of perforin mRNA–expressing cells can be detected with this analysis. Furthermore, the maximum number of silver grains, reflecting the maximum amount of RNA detected in a given cell, is comparable in all tissue specimens from patients with chronic pancreatitis (data not shown). Detection of perforin-containing cells in infiltrates of tissue specimens of patients with chronic pancreatitis by immunofluorescence (Figure 1E) showed an identical distribution pattern as found by in situ hybridization for perforin mRNA. Stainings with isotype-matched antibodies showed no specific signal (Figure 1F).
Table 3. Frequency of Perforin mRNA–Expressing Cells in the Respective Cell Subsets Patient
CD4 ( %)
CD8 ( %)
CD56 ( %)
G H I J K L M
0.9 1.0 1.0 1.7 1.2 0.0 1.0
23.9 10.6 7.2 9.4 12.5 ND 8.0
ND 23.1 9.3 35.0 6.1 16.6 ND
NOTE. The indicated cell populations were sorted on a FACSvantage, and in situ hybridization was performed. The percentage of cells expressing perforin mRNA was determined by counting positive cells in the corresponding cell populations. ND, not determined.
Perforin mRNA Expression in T-Cell Subsets To determine the preferential phenotype of the perforin mRNA–positive cells, isolated cells from tissue samples from patients with chronic pancreatitis were stained with the appropriate primary antibodies for CD4, CD8, and CD56 for subsequent sorting of the respective cell subpopulations. Cells were directly sorted onto polyL-lysine–coated glass slides and hybridized in situ with perforin probes. This procedure is required because of the different fixation requirements of in situ hybridization and immunohistochemistry, thus leading to a significant loss of perforin mRNA on tissue sections immunohistochemically stained before in situ hybridization (data not shown). The frequency of perforin mRNA– expressing cells is shown in Table 3. Although perforin mRNA is expressed only in a minor fraction of the CD4 T cells (1.0%; SD, {0.5%), high frequencies of CD8 (11.9%; SD, {6.2%) and CD56/ (18.0%; SD, {11.6%) cells express this mRNA at detectable levels. The comparison of the frequencies of perforin mRNA– expressing cells obtained by enumerating in situ hybridizations of tissue sections and by analyzing isolated and sorted cells shows a good correlation (correlation coefficient, 0.92; numerical data not shown).
Discussion
Figure 2. Perforin mRNA–expressing cells per square millimeter of tissue section in patients with alcoholic pancreatitis and normal controls. The mean ({SD) of all values in each group is indicated on the right. The differences between the two groups are statistically significant (P õ 0.0005).
/ 5E1C$$0038
04-18-97 17:39:23
gasa
Despite numerous hypotheses on the etiopathogenesis of chronic pancreatitis, the pathogenetic steps of this disease still remain enigmatic. In developed countries, most patients with chronic pancreatitis (60%– 70%) have a long history of alcohol abuse before clinical onset of disease.12 With the exception of studies reporting an increased MHC class II expression in affected tissues from patients with chronic pancreatitis and the observation of T cell–enriched cellular infiltrates, studies WBS-Gastro
May 1997
CYTOTOXIC T CELLS IN CHRONIC PANCREATITIS 1661
on the contribution of the immune system, either humoral or cellular in nature, to the pathogenesis of chronic pancreatitis in general and to alcoholic pancreatitis in particular are limited. Phenotypical analysis on immunohistochemically stained pancreas sections of patients with chronic alcoholic pancreatitis showed numerous T cells. Both major T-cell subsets, CD4 and CD8 T cells, were, in general, present at comparable frequencies (data not shown). These data are further substantiated by the cytofluorometric analysis of isolated pancreas-infiltrating cells (Table 2) in which the mean of the CD4/CD8 ratio is 1.4:1.0. However, the variation of the CD4/CD8 ratio determined in isolated pancreas-infiltrating cells is considerable. This variation was also observed on pancreatic tissue sections from the same patients, stained immunohistochemically with the respective reagents. Thus, selective loss or enrichment of a given T-cell subset during the isolation procedure is very unlikely. CD56/ cells were consistently found in small numbers (Table 2). They have been described to represent mostly NK cells and large granular lymphocytes that kill target cells in an MHC-independent manner.13,14 Thus, CD56 is often used as a phenotypic marker for NK cells. The CD450 cells found in the lymphoid cell population, as defined by gating in the forward and side light scatter, are likely to represent contaminating cells, either fibrocytes, fibroblasts, or acinar cells. Indirect evidence for this assumption is provided by the finding that the relative frequency of CD450 cells was considerably higher before lymphocyte enrichment of the isolated cell populations. In situ hybridization of sections of pancreatic tissue from patients with chronic pancreatitis consistently showed a high frequency of perforin mRNA–positive cells in cellular infiltrates (Figure 2). All patients analyzed showed clearly higher frequencies of activated cytotoxic cells than controls. The presence of perforin was also confirmed on the protein level by immunohistochemistry and consistently reflected the results obtained by in situ hybridization (Figure 1B and E). Together with the finding that these activated cells are often found in contact to remaining parenchyma, this might indicate an involvement of these cells in tissue destruction. Perforin mRNA expression in vivo is reported only in activated CD4- and CD8-positive T cells and NK cells. The relative frequencies of perforin mRNA in the CD4-, CD8-, and CD56-positive cell subsets (Table 3) showed that the majority of perforin mRNA–expressing cells were found within the CD8/ T-cell subset, whereas in the CD56/ and CD4/ subsets only 27% and 13%, respectively, of all perforin mRNA–positive cells were found. Thus, the majority of perforin-mediated killing / 5E1C$$0038
04-18-97 17:39:23
gasa
can be attributed to MHC class I–dependent killing, a minor but significant part (CD56/ NK cells) to an MHC unrestricted killing, and only a small part to MHC class II–dependent cytotoxicity. Besides this perforin-dependent mechanism of cell-mediated cytotoxicity, perforinindependent cytotoxic effector mechanisms, in particular Fas/FasL interaction, also have been described recently.6,15 However, for technical reasons, the contribution of this latter mechanism, which appears to be mainly operative in CD4/ cytotoxic T cells, cannot be assessed in this study. Furthermore, this perforin-independent mechanism seems to be mainly involved in regulation of the immune response rather than in effector mechanisms in host defense or autoimmune tissue destruction.16 As shown in Figure 2, there are large variations between the different tissue samples analyzed. To assess whether these variations are caused by technical artifacts or truly represent actual differences in the extent of cellular activation, additional hybridizations and evaluations of the tissue sections hybridized with perforin antisense probes have been performed. The observed comparable frequencies of cells expressing perforin mRNA at detectable levels in the superficial and deeper areas of pancreatic tissues indicate that the methods used for tissue processing allowed for a homogenous tissue fixation and preservation of mRNA within the entire tissue specimens. The observation that comparable maximum expression levels of perforin mRNA in positive cells are found in all tissue specimens analyzed from patients with chronic pancreatitis further indicates that, because of the standardized procedure of tissue processing, only minimal variations in the preservation of mRNA occurred between different tissue specimens. This latter aspect is further substantiated by the result of control hybridization with an insulin RNA antisense probe, which showed comparable signals for the presence of insulin mRNA on tissue sections from patients with chronic pancreatitis with intact islets of Langerhans (data not shown). Thus, taken together, the observed variations in the frequency of perforin mRNA–expressing cells in tissue specimens from patients with chronic pancreatitis very likely reflect different activation states of cytotoxic effector cells. The observed variations in the expression of the perforin gene at consistently high levels in all tissue samples examined are reminiscent of acute episodes of activation of the cellular immune response, as has been shown in tissue destruction during viral infections17 and transplant rejection.18 It will be of particular interest to define the nature of the triggering antigens that leads to these recurrent episodes of activation of mainly CD8-positive, MHC class I–restricted cytolytic T cells. Because antigens presented by MHC class I molecules are derived from endogWBS-Gastro
1662 HUNGER ET AL.
GASTROENTEROLOGY Vol. 112, No. 5
enously produced proteins, such as cellular proteins, or virally encoded antigens, one might speculate that recurrent virus infections or an altered pattern of endogenously produced proteins, recognized by virus-specific or autoreactive T cells, induce this cellular activation. In this context it will be of foremost interest to analyze whether the cellular activation events operative during episodes of acute pancreatitis resemble the recurrent activation episodes in chronic pancreatitis described in the present study. Such an analysis might clarify the controversy of whether recurrent episodes of alcoholic acute pancreatitis will eventually lead to alcoholic chronic pancreatitis, as has been hypothesized previously.19 – 21 However, because of the lack of appropriate tissue specimens from patients with acute pancreatitis, such an analysis could not be included in this study. It will be of further interest to assess whether cell-mediated cytotoxicity is also operative in tissue destruction of other alcohol-associated disorders. The finding that numerous CD4 and CD8 T cells infiltrate the lesions in alcoholic liver disease and that these cells persist during the progression of disease may suggest that cell-mediated cytotoxicity is also operative in destruction of the liver parenchyma.22 However, further studies of perforin expression and T-cell function are required to be able to compare the contribution of cellmediated cytotoxicity in tissue destruction in these disorders. Furthermore, experiments comparing the vb usage of the T-cell receptors in the activated infiltrating T cells with the vb usage of blood T lymphocytes would give more insight in the clonality of the T-cell response. Such studies could clarify if only a few antigens are responsible for the T-cell activation or if a broad range of antigens provoke a polyclonal T-cell response. Activated cytotoxic cells of both the CD8 and the CD56 subsets may influence the course of disease not only through their potent cell-mediated cytotoxic capacities but also through the secretion of lymphokines, in particular of interferon gamma.23 Together with the reported up-regulation of transforming growth factor b in the inflammatory process,24 this may significantly affect the functional differentiation of CD4/ T cells. Both cytokines are known to promote the expansion of Th1-type T cells, leading to a predominant cellular immune response.25 Further evidence that an imbalance in the immune system is of etiopathological importance in chronic pancreatitis has been recently provided in an animal model in which the absence of the immunoregulatory molecule CTLA-4 leads to a chronic pancreatitis-like disease.26 In conclusion, the observed increased frequency of perforin mRNA–expressing cells in the pancreas of patients with alcoholic chronic pancreatitis indicates a preferential / 5E1C$$0038
04-18-97 17:39:23
gasa
activation of these effector cells of cell-mediated cytotoxicity. The consistent observation of a close association between activated cytotoxic cells with remaining intact pancreatic parenchyma may indicate that these cells are directly involved in the destruction of parenchymal cells. It may also indicate in situ activation of these cells, thus providing circumstantial evidence for an involvement of autoreactive cytotoxic T cells and NK cells in the pathogenesis of this disease.
References 1. Steer ML, Waxman I, Freedman S. Chronic pancreatitis. N Engl J Med 1995;332:1482–1490. 2. Bedossa P, Bacci J, Lemaigre G, Martin E. Lymphocyte subsets and HLA-DR expression in normal pancreas and chronic pancreatitis. Pancreas 1989;5:415–420. 3. Ka¨gi D, Ledermann B, Bu¨rki K, Seiler P, Odermatt B, Olsen KJ, Podack ER, Zinkernagel RM, Hengartner H. Cytotoxicity mediated by T cells and natural killer cells is greatly impaired in perforindeficient mice. Nature 1994;369:31–37. 4. Heusel JW, Wesselschmidt RL, Shresta S, Russel JH, Ley TJ. Cytotoxic lymphocytes require granzyme B for rapid induction of DNA fragmentation and apoptosis in allogeneic target cells. Cell 1994;76:977–987. 5. Griffiths G, Mueller C. Expression of perforin and granzymes in vivo: potential diagnostic markers for activated cytotoxic cells. Immunol Today 1991;12:415–419. 6. Henkart PA. Lymphocyte-mediated cytotoxicity: two pathways and multiple effector molecules. Immunity 1994;1:343–346. 7. Friess H, Malfertheiner P, Isenmann R, Ku¨hne H, Beger HG, Bu¨chler MW. The risk of pancreaticointestinal anastomosis can be predicted preoperatively. Pancreas 1996;13:202–208. 8. Held W, MacDonald HR, Weissman IL, Hess MW, Mueller C. Genes encoding tumor necrosis factor a and granzyme A are expressed during development of autoimmune diabetes. Proc Natl Acad Sci USA 1990;87:2239–2243. 9. Lichtenheld MG, Olsen KJ, Lu P, Lowrey DM, Hameed A, Hengartner H, Podack ER. Structure and function of human perforin. Nature 1988;335:448–451. 10. Mueller C, Gershenfeld HK, Lobe CG, Okada CY, Bleackley RC, Weissman IL. A high proportion of T lymphocytes that infiltrate H-2 incompatible heart allografts in vivo express genes encoding cytotoxic cell–specific serine proteases, but do not express the MEL-14-defined lymph node homing receptor. J Exp Med 1988; 167:1124–1136. 11. Baetz K, Isaaz S, Griffiths GM. Loss of cytotoxic T lymphocyte function in Chediak–Higashi syndrome arises from a secretory defect that prevents lytic granule exocytosis. J Immunol 1995; 154:6122–6131. 12. Sarles H, Sahel J, Staub JL, Bourry J, Laugier R. Chronic pancreatitis. In: Howat HT, Sarles H, eds. The exocrine pancreas. Philadelphia: Saunders, 1979:402–439. 13. Lu PH, Negrin RS. A novel population of expanded human CD3/CD56/ cells derived from T cells with potent in vivo antitumor activity in mice with severe combined immunodeficiency. J Immunol 1994;153:1687–1696. 14. Metha BA, Schmidt-Wolf IGH, Weissman IL, Negrin RS. Two pathways of exocytosis of cytoplasmic granule contents and target cell killing by cytokine-induced CD3/CD56/ killer cells. Blood 1995;68:3493–3499. 15. Stalder T, Hahn S, Erb P. Fas antigen is the major target molecule for CD4/ T cell–mediated cytotoxicity. J Immunol 1994;152: 1127–1133. 16. Ka¨gi D, Ledermann B, Bu¨rki K, Zinkernagel RM, Hengartner H.
WBS-Gastro
May 1997
17.
18.
19.
20. 21.
22.
CYTOTOXIC T CELLS IN CHRONIC PANCREATITIS 1663
Molecular mechanisms of lymphocyte-mediated cytotoxicity and their role in immunological protection and pathogenesis in vivo. Annu Rev Immunol 1996;14:207–232. Mueller C, Ka¨gi D, Aebischer T, Odermatt B, Held W, Podack ER, Zinkernagel RM, Hengartner H. Detection of perforin and granzyme A mRNA in infiltrating cells during infection of mice with lymphocytic choriomeningitis virus. Eur J Immunol 1989;19: 1253–1259. Mueller C, Shelby J, Weissman IL, Perinat Frey T, Eichwald EJ. Expression of the protease gene HF as a marker in rejecting allogeneic murine heart transplants. Transplantation 1991;51: 514–517. Ammann RW, Heitz PU, Klo¨ppel G. Course of alcoholic chronic pancreatitis: a prospective clinicomorphological long-term study. Gastroenterology 1996;111:224–231. Amman RW, Muellhaupt B. Progression of alcoholic acute to chronic pancreatitis. Gut 1994;35:552–556. Klo¨ppel G, Maillet B. The morphological basis for the evolution of acute pancreatitis into chronic pancreatitis. Virchows Arch A Pathol Anat Histopathol 1992;420:1–4. Chedid A, Mendenhall CL, Moritz TE, French SW, Chen TS, Morgan TR, Roselle GA, Nemchausky BA, Tamburro CH, Schiff ER, McClain CJ, Marsano LS, Allen JI, Samanta A, Weesner RE, Hen-
/ 5E1C$$0038
04-18-97 17:39:23
gasa
23. 24.
25. 26.
27.
derson WG, VAC Study Group 275. Cell-mediated hepatic injury in alcoholic liver disease. Gastroenterology 1993;105:154–266. Farrar MA, Schreiber RD. The molecular cell biology of interferong and its receptor. Annu Rev Immunol 1993;11:571–611. Van Laethem JL, Deviere J, Resibois A, Rickaert F, Vertongen P, Ohtani H, Cremer M, Miyazono K, Robberecht P. Localization of transforming growth factor b1 and its latent binding protein in human chronic pncreatitis. Gastroenterology 1995;108:1873– 1881. Romagnani S. Lymphokine production by human T cells in disease states. Annu Rev Immunol 1994;12:227–257. Waterhouse P, Penninger JM, Timms E, Wakeham A, Shahinian A, Lee KP, Thompson CB, Griesser H, Mak TW. Lymphoproliferative disorders with early lethality in mice deficient in Ctla-4. Science 1995;270:985–988. Malfertheiner P, Bu¨chler M, Mu¨ller A, Ditschuneit H. Fluorescein dilaurate serum test: a rapid tubeless pancreatic function test. Pancreas 1987;2:53–60.
Received November 5, 1996. Accepted February 11, 1997. Address requests for reprints to: Markus W. Bu¨chler, M.D., Department of Visceral and Transplantation Surgery, University of Bern, Inselspital, CH-3010 Bern, Switzerland. Fax: (41) 31-382-4772.
WBS-Gastro