European Journal of Clinical Investigution ( 1 990) 20, 5 16-524
Antibodies t o cytoskeletal proteins in patients with Crohn’s disease W. J. MAYET, A. G. PRESS, E. HERMANN, R. MOLL*, M. MANNS, K. EWE & K. H. MEYER ZUM BUSCHENFELDE, I Medical Clinic, Institute of Pathology, University of Mainz, Mainz, FRG Received 1 1 October 1989 and in revised form 26 March 1990
Abstract. The immunologic basis of inflammatory bowel disease has been the focus of interest of a series of studies on Crohn’s disease and the process of immune sensitization at the gastrointestinal mucosal level is functionally poorly understood. To date only few contradictory reports concerning the incidence of autoantibodies in patients with this disease exist. The aim of this study was to investigate the sera drawn from 60 patients suffering from biopsy-proven Crohn’s disease to evaluate the prevalence of autoantibodies against nuclear antigens and cytoskeletal proteins. Using standard methods, no anti-nuclear antibodies or antibodies to extractable nuclear antigens could be detected. All sera were also negative for antibodies to double-stranded DNA, anti-mitochondria1 antibodies, and antibodies to gastric parietal cells. Using sensitive enzyme-linked immunosorbent assays with purified antigens and Western blotting with cytoskeletal proteins of human intestinal cells, the following antibodies could be demonstrated: cytokeratin 18 autoantibodies (IgG 20.0%; IgM 6.7%; IgA 13.3%), actin antibodies (IgG 36.7%; IgM 48.3‘%,IgA 26.7‘%), desmin antibodies (IgG 6.7%; IgM 15.08%; IgA 5.00/,), vimentin antibodies (IgG 3.3%; IgM 16.7%; IgA 10.0%) and tropomyosin antibodies (IgG 3.3%; IgM 3.3‘%,IgA 5.0‘%).Statistically significant correlations could be found for levels of cytokeratin 18 antibodies (IgM-type) and the BEST index of activity, and for levels of desmin antibodies (IgM-type) and the van HEES index of activity. Highest levels could be measured for actin antibodies (IgG-type) in patients with isolated disease manifestation in the colon. The mechanism of induction of autoantibodies against cytoskeletal components in Crohn’s disease still remains obscure. Unmasking of hidden antigens after cell injury during the inflammatory process of disease might lead to sensitization and antibody production. The pattern of antibodies in patients with Crohn’s disease seems to be different compared with that of connective tissue diseases. Correspondence: Professor Dr K.-H. Meyer zum Biischenfelde. I Medizinische Klinik und Poliklinik der Johannes GutenbergUniversitlt, Langenbeckstrasse 1. D-6500Mainz, FRG.
516
Keywords. Antibodies, Crohn’s disease, cytoskeletal proteins, ELISA. Introduction Crohn’s disease represents a chronic inflammatory bowel disease with possible involvement of the whole gastrointestinal tract. Aetiology of Crohn’s disease remains unknown; however, certain features of this disease have suggested immunologic factors of possible aetiologic importance. Apart from antibodies to bacterial antigens, anti-nuclear antibodies (ANA), smooth muscle antibodies (SMA) and antibodies against intermediate filaments could be detected [ I ,2]. Other authors did not obtain evidence for humoral and cellular immunological abnormalities [3]. ANA have gained importance as diagnostic markers of several connective tissue diseases, whereas antibodies to cytoplasmic components are often associated with organspecific autoimmune diseases. These contradictory reports were the impetus to investigate the sera drawn from 60 patients with Crohn’s disease using standard methods but also polyacrylamide gel electrophoresis followed by Western blotting and recently established enzyme-linked immunosorbent assays (ELISA) which could be easily performed in routine diagnostics to evaluate the prevalence of autoantibodies against nuclear antigens and cytoskeletal components. Materials and methods Serum samples
Serum samples were obtained from 142 donors, 60 from patients suffering from Crohn’s disease proven by endoscopy, X-ray and/or histology. No patient suffered from an infectious disease like malaria, brucellosis or an apparent viral infection. The mean age of the Crohn group was 32 (range 12-71 years), 34 were females and 26 males. The BEST [4] and van HEES [5] index were used to evaluate activity and severity of Crohn’s disease. The former contains eight parameters such as abdominal pain, number of liquid stools, score of general well being and others. This index estimates
ANTIBODIES TO CYTOSKELETAL PROTEINS IN CROHN'S DISEASE Table I . Localization of Crohn's disease in different segments of the gastrointestinal tract in 60 patients. BEST index > I50= active disease vs. HEES index > 100=mild disease; > 150=moderate disease Upper Gi Terminal ileum Colon+ rectum lleocolitis
2 II
2 6
0 5
3
l30.5+ 102.5 X4.4k69.7
207.7k7 142.6k33.8
17
13
4
3
96.4k86.1
144.9k24.1
25
12
13
8
89.2k77.5
157,4+33
I
m. male; f, female: upper GI, upper gastrointestinal tract; SD, standard deviation.
Table 2. Extraintestinal manifestations in 60 patients with Crohn's disease None Anal lesion
46 7
E.nor/o.~t>~ 5
Arthritis
2
20 4 I I
26 3 4
X
I
I
2 5
77.0k68.9 96k77.7 212.2i-614 156k60.8
147.2k30.5 155.6k25.7 193.Xk21.1 174.5k54.5
m. male; f. female: E. nodosum, erythema nodosum; SD, standard deviation.
predominantly patients' complaints (BEST > 150=active disease). The van HEES index nearly exclusively contains data which can be raised objectively and is laboratory orientated (van HEES < 150 =mild disease; > 150 and < 2 10 =moderate disease; > 2 10 = severe disease). Mean BEST index for males was 64.5 (range - 17-216) and for females 66 (range -26-317). Mean van HEES index for males was 157 (range 96-229) and for females 142 (range 103-217). Forty-four out of 60 patients with Crohn's disease were in remission (BEST index < 150; 18 males and 26 females). 16 out of 60 patients suffered from clinical active disease (BEST index > 150; 8 males and 8 females). Further data concerning clinical features are given in Tables 1 and 2. Eighty-two sera were obtained from healthy blood donors (HBD). All serum samples were stored at - 20°C until used. Serologic l e s t s
All serum samples were tested for 15 different autoantibody specificities. Tests were done in triplicates. The methods used for each antibody are listed below: ANA, anti-mitochondria1 antibodies (AMA) and antibodies to parietal cells (PCA) were detected using indirect immunofluorescence microscopy (IFT) on rat stomach, kidney and liver tissue, HEp2-cells and CVIcells. HEp2-cells were commercially prepared (Antibodies Inc., Davis, CA, USA). CVI-cells were grown on cover slips air-dried and fixed with methanol/ EGTA for 4 min at - 20°C. Positive sera were serially diluted in PBS to obtain antibody titres. Stained tissues were analysed with a Zeiss Axiophot microscope equipped with epifluorescence optics (Zeiss, Oberkochen, FRG). A super-pressure mercury lamp (HBO
517
50) served as light source. An excitation filter BP 485/ 20 and a cut-off filter LP 520 were used. The results were documented by photography with Ilford XP-I film. Antibodies to extractable nuclear antigens (ENA) were detected by Ouchterlony immunodiffusion (ID) and counterimmunoelectrophoresis (CIE). Screening of sera was done by ID [6] of undiluted sera with an antigen concentration of 50 mg ml-' in 0.650/m agarose gel [7]: rabbit thymus extract (RTNE) (Paesel, Frankfurt, FRG) and human spleen extract (HSE) prepared as described by Venables et al. [8] served as antigens. CIE was performed as described by Tan [9] using standardized reference antisera (anti-Ro, La, Sm, RNP; Center for Disease Control, Atlanta, GA, USA). Additionally sera were tested by performing polyacrylamide gel electrophoresis (PAGE) and Western blotting. Segments of normal human small intestine obtained during the surgical removal of malignant tumours were used for the isolation of intestinal epithelial cells, followed by lysis and extraction using low- and high-salt buffers and Triton X- 100, essentially as previously described by Franke et al. [lo], which yielded cytoskeletal residues. Cytoskeletal preparation of human smooth muscle rich in actin and desmin was obtained by microdissection of the tunica muscularis propria of stomach followed by similar extraction steps. PAGE (10%) of RTNE, HSE. commercially available cytoskeletal antigens, cytoskeletal material of human intestinal cells and smooth muscle tissues was performed according to the system of Laemmli [ I I ] in the presence of O.l"/u NaDodS04 (SDS). The gels were stained with silver [ 121. For twodimensional gel electrophoresis, non-equilibrium pH gradient electrophoresis [ 131 was applied in the first dimension; in the second dimension SDS-PAGE was performed using the elevated salt-buffer system described by Achtstatter et al. [14]. Western blotting was performed as described previously [15,16] or as described in [ 171. Antibodies to double-stranded DNA (dsDNA) were measured by FARR- and PEG-assay [I81 and IFT on Crithidia luciliae (Antibodies Inc, Davis, CA, USA). Rheumatoid factors (RF) were measured by latex fixation and Waaler-Rose test (Behringwerke, Marburg, FRG). A CA-ELISA
Wells of a microtitre plate (M 129 A; Dynatech, Plochingen, FRG) were coated for 16 h at room temperature (24°C) with 100 pi of the respective purified antigen (cytokeratin no. 18 bovine, 981059 Boehringer, FRG; vimentin from bovine lens, 981 192 Boehringer; desmin from chicken stomach, 98 I 1 17 Boehringer; actin from rabbit muscle, 98 1052 Boehringer; tropomyosin from bovine muscle, T-4895 Sigma, FRG) diluted in 0.05%] sodiumcarbonate buffer pH 9.6. Concentrations ranging from 100 pg to 3 pg ml- I for coating antigen were tested and 5 pg ml-' was found to give optimum results for actin, desmin, vimentin and tropomyosin and 10 pg ml-' for cyto-
518
W. J . MAYET et al.
keratin 18. After washing once with PBS, each well was filled with PBS containing 1 % BSA and left at room temperature for 2 h in order to block unsaturated sites. After washing the wells with PBS containing 0.05% Tween 20 (three times), 100 pI of the respective serum diluted I :500 in PBS was added and incubated for 1 h at room temperature. Rabbit-anti-human peroxidase (IgG, gamma-chain specific; Dako P 214. IgM, p-chain specific; Dako P 322. IgA, a-chain specific; Dako P 216) diluted 1 : 1000 in PBS containing 0.1% BSA was added after washing the wells again three times with PBS containing 0.05‘!41 Tween 20. After 1 h incubation time at room temperature the wells were washed again three times with PBS containing 0.05% Tween 20 and 100 p1 of freshly prepared substrate ( 100 mmol 1 I acetate buffer, pH 5.4 containing 0.002‘!41 H202 and 0.43 g ml-’ ophenylene diamine; P-3888, Sigma, Deisenhofen, FRG) were added to each well. After I0-min incubation at room temperature the reaction was stopped by the addition of 100 pl 1 mmol I-’ H2S04. Optical densities (D)were measured at 492 nm in a doublebeam spectrophotometer (Beckmann, Model 25, Fullerton, CA, USA). ~
Statistical analysis
The upper limit of normality in the five ELISA systems was taken as two standard deviations above mean normal control levels. By performing Mann-Whitney test the average rank of patients with Crohn’s disease and HBD was calculated for each of the five autoantibodies including their respective isotypes (IgG, IgM, IgA). Correlations between these antibodies were saught using Spearman’s correlation coefficient. Linear regression analysis was performed to evaluate relationships between clinical features (i.e. anatomic location of involvement, extra-intestinal manifestations and surgical treatment), the indices of BEST and van HEES (i.e. clinical activity of disease) and antibody levels measured in the ELISA. Study for the reproducibility of the assay
Sera from five patients with Crohn’s disease that reacted positively in the desmin ELISA were assayed in different experiments on different days and another five sera were tested in five different experiments within one assay procedure to study the reproducibility of the assay (i.e. inter- and intra-assay coefficients). Results
Titration curves of three sera positive for desmin antibodies and of a pool of HBD are shown in Fig. 1. The ELISA systems provided a good discrimination between ACA-positive and ACA-negative sera from 60 patients with Crohn’s disease (Fig. 2a-e). As the respective cut-off point values above the mean optical density of the 82 HBD plus two standard deviations
2.5
2.0
I
I+
1.5
1.0
0.5
0
50
I00
500
1000
5000
Reciprocal o f serumdilution Figure 1. Titration curves for three anti-desmin positive sera ( A ) and a serum pool of healthy blood donors ( 0 )
were chosen: 0.470 (IgG), 0.398 (IgM) and 0.405 (IgA) forcytokeratin 18 antibodies,O.l71 (IgG), 0.121 (IgM) and 0.371 (IgA) for actin antibodies, 0.493 (IgG), 0.27 1 (IgM) and 0.550 (IgA) for desmin antibodies, 0.179 (IgG), 0.123 (IgM) and 0.548 (IgA) for vimentin antibodies, 0.330 (IgG), 0.41 I (IgM) and 0.543 (IgA) for tropomyosin antibodies. Antibodies detected in patients’ sera
In Table 3 the incidence of ACA (IgG-, IgM- and IgAtype) as determined by five ELISA systems in 60 patients with Crohn’s disease is shown. All sera classified to be positive in the ELISAs recognized the respective antigens in Western blot (data not shown). In this group of patients no ANA, antibodies to dsDNA, PCA, antibodies to ENA and AMA could be detected. All sera were RF negative. No antibodies could be found in the sera of 82 HBD. Performing the ELISAs, the optical densities of the Crohn’s group were significantly different from the optical densities of the HBD group as determined by statistical analysis. Low but significant correlations could be found for levels of cytokeratin 18 antibodies (IgM-type) and the BEST index of activity (0.30; P < 0.05) and for levels of desmin antibodies (IgM-type) and the van HEES index of activity (0.282; PcO.05). Table 4 shows positive correlations of different antibodies. Highest optical density ranges could be measured for actin antibodies (IgG-type) in patients with isolated disease manifestation in the colon. Vimentin and desmin antibodies (IgG-type) could be found only in sera from patients with colon involvement. Most of the patients had ACA against several isotypes and also against various cytoskeletal components (Table 3). Figure 3a shows that sera of all 12 patients classified as positive in the ELISA recognized cytokeratin 18 using an extract of
ANTIBODIES TO CYTOSKELETAL PROTEINS IN CROHN’S DISEASE
( a )
\
1.2
(b)
Keratin
)
Actin
0.6
-1
0
f0
. 0
.. 3
0
t
*:
8
3
._.. t----
!
0
(d 1 1.0 I
1.0 0
E N 6,
e +
0.8 -
0
0
0.8 0
0.6 -
t
_._....___
.
0
0.2 Oa4
n
0.8
% .
Ii (el
1
0.6
Viment i n 0-
11
.L....
c
0
om
I
0.2
..........
n
.
Tropomyosin
0
J
0 ‘ IgG
I gM
IOA
Figure 2. Results of ACA ELISAs in patients with Crohn’s disease. Cut off (-
519
- --
-); mean of healthy blood donors (.)+SD.
Table 3. Incidence of ACA in patients with Crohn's disease
Keratin
IgG IgM IgA
12 4 8
(20.0%) Actin IgG 22 (36.7%) Desmin IgG 4 (6.7%) Vimentin IgG 2 (3.3%) Tropomyosin IgG 2 (3.3%) IgM 9 ( I 5.0%) IgM 2 (3.3%) IgM 29 (48.3'Xi) IgM 10 (16'70/;1) (6.7%)) IgA 3 (5.0%)) (13.3"%) IgA 3 (5.3%) IgA 6 (IO'O'X,) IgA I 6 (26.7"/;,)
(b) Patient No.
Keratin
Actin
Desmin
Vimentin
Tropomyosin
I 2 3 4 5 6
GM
G
M M
G G
G
M
I 8 9 10
GMA
GM G
G
II
GM M
12 13 14 15 16
17 18 19 20 21 22 23 24 25 26 21 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 41 48 49 50 51 52 53 54 55 56 57 58 59 60
A G
M G
M M
G GM M GMA GM GM GMA A A GMA GMA MA M GM G A M
G
G
A MA A
A
MA M M M M
MA
M
MA A M
GM G A
G
A A
M GM G A A M
G
M MA M M
G
M M M G G
M A
M
M
M A
GM GM G
G
M
GM A A
G
A
M
A
G
G GM
A A
A
A A
ANTIBODIES TO CYTOSKELETAL PROTEINS IN CROHN’S DISEASE
total cytoskeletal proteins of human intestinal cells as source of antigen (Western blot after SDS-PAGE). These results were further confirmed in two-dimensional gel electrophoresis (Fig. 3b). The immunoblots also show the presence of autoantibodies against other cytokeratin polypeptides (Fig. 3a, b). Actin-positive sera (ELISA) also recognized actin by performing Western blot with a cytoskeletal preparation of human smooth muscle (data not shown).
Table 4. Correlation of ACA established using Spearman’s correlation coefficient Desmin-lgM + Keratin-lgM: Vimentin-lgM Keratin-lgM: Tropomyosin-IgM + Keratin-IgM: Tropomyosin-IgM + Desmin-IgM: Tropomyosin-lgG + Actin-lgG: Keratin-IgA Vimentin-IgA: Tropomyosin-lgA + Vimentin-IgA: Keratin-lgA + Desmin-lgA: Actin-igA + Keratin-IgA: Tropomyosin-lgA + Keratin-lgA: Tropomyosin-lgA + Desmin-IgA:
+
+
(a
1
521
r = 0.54; P < 0,005 r = 0.59; P < 0.005 r = 0 . 6 5 ; P
Antibodies t o cytoskeletal proteins in patients with Crohn’s disease W. J. MAYET, A. G. PRESS, E. HERMANN, R. MOLL*, M. MANNS, K. EWE & K. H. MEYER ZUM BUSCHENFELDE, I Medical Clinic, Institute of Pathology, University of Mainz, Mainz, FRG Received 1 1 October 1989 and in revised form 26 March 1990
Abstract. The immunologic basis of inflammatory bowel disease has been the focus of interest of a series of studies on Crohn’s disease and the process of immune sensitization at the gastrointestinal mucosal level is functionally poorly understood. To date only few contradictory reports concerning the incidence of autoantibodies in patients with this disease exist. The aim of this study was to investigate the sera drawn from 60 patients suffering from biopsy-proven Crohn’s disease to evaluate the prevalence of autoantibodies against nuclear antigens and cytoskeletal proteins. Using standard methods, no anti-nuclear antibodies or antibodies to extractable nuclear antigens could be detected. All sera were also negative for antibodies to double-stranded DNA, anti-mitochondria1 antibodies, and antibodies to gastric parietal cells. Using sensitive enzyme-linked immunosorbent assays with purified antigens and Western blotting with cytoskeletal proteins of human intestinal cells, the following antibodies could be demonstrated: cytokeratin 18 autoantibodies (IgG 20.0%; IgM 6.7%; IgA 13.3%), actin antibodies (IgG 36.7%; IgM 48.3‘%,IgA 26.7‘%), desmin antibodies (IgG 6.7%; IgM 15.08%; IgA 5.00/,), vimentin antibodies (IgG 3.3%; IgM 16.7%; IgA 10.0%) and tropomyosin antibodies (IgG 3.3%; IgM 3.3‘%,IgA 5.0‘%).Statistically significant correlations could be found for levels of cytokeratin 18 antibodies (IgM-type) and the BEST index of activity, and for levels of desmin antibodies (IgM-type) and the van HEES index of activity. Highest levels could be measured for actin antibodies (IgG-type) in patients with isolated disease manifestation in the colon. The mechanism of induction of autoantibodies against cytoskeletal components in Crohn’s disease still remains obscure. Unmasking of hidden antigens after cell injury during the inflammatory process of disease might lead to sensitization and antibody production. The pattern of antibodies in patients with Crohn’s disease seems to be different compared with that of connective tissue diseases. Correspondence: Professor Dr K.-H. Meyer zum Biischenfelde. I Medizinische Klinik und Poliklinik der Johannes GutenbergUniversitlt, Langenbeckstrasse 1. D-6500Mainz, FRG.
516
Keywords. Antibodies, Crohn’s disease, cytoskeletal proteins, ELISA. Introduction Crohn’s disease represents a chronic inflammatory bowel disease with possible involvement of the whole gastrointestinal tract. Aetiology of Crohn’s disease remains unknown; however, certain features of this disease have suggested immunologic factors of possible aetiologic importance. Apart from antibodies to bacterial antigens, anti-nuclear antibodies (ANA), smooth muscle antibodies (SMA) and antibodies against intermediate filaments could be detected [ I ,2]. Other authors did not obtain evidence for humoral and cellular immunological abnormalities [3]. ANA have gained importance as diagnostic markers of several connective tissue diseases, whereas antibodies to cytoplasmic components are often associated with organspecific autoimmune diseases. These contradictory reports were the impetus to investigate the sera drawn from 60 patients with Crohn’s disease using standard methods but also polyacrylamide gel electrophoresis followed by Western blotting and recently established enzyme-linked immunosorbent assays (ELISA) which could be easily performed in routine diagnostics to evaluate the prevalence of autoantibodies against nuclear antigens and cytoskeletal components. Materials and methods Serum samples
Serum samples were obtained from 142 donors, 60 from patients suffering from Crohn’s disease proven by endoscopy, X-ray and/or histology. No patient suffered from an infectious disease like malaria, brucellosis or an apparent viral infection. The mean age of the Crohn group was 32 (range 12-71 years), 34 were females and 26 males. The BEST [4] and van HEES [5] index were used to evaluate activity and severity of Crohn’s disease. The former contains eight parameters such as abdominal pain, number of liquid stools, score of general well being and others. This index estimates
ANTIBODIES TO CYTOSKELETAL PROTEINS IN CROHN'S DISEASE Table I . Localization of Crohn's disease in different segments of the gastrointestinal tract in 60 patients. BEST index > I50= active disease vs. HEES index > 100=mild disease; > 150=moderate disease Upper Gi Terminal ileum Colon+ rectum lleocolitis
2 II
2 6
0 5
3
l30.5+ 102.5 X4.4k69.7
207.7k7 142.6k33.8
17
13
4
3
96.4k86.1
144.9k24.1
25
12
13
8
89.2k77.5
157,4+33
I
m. male; f, female: upper GI, upper gastrointestinal tract; SD, standard deviation.
Table 2. Extraintestinal manifestations in 60 patients with Crohn's disease None Anal lesion
46 7
E.nor/o.~t>~ 5
Arthritis
2
20 4 I I
26 3 4
X
I
I
2 5
77.0k68.9 96k77.7 212.2i-614 156k60.8
147.2k30.5 155.6k25.7 193.Xk21.1 174.5k54.5
m. male; f. female: E. nodosum, erythema nodosum; SD, standard deviation.
predominantly patients' complaints (BEST > 150=active disease). The van HEES index nearly exclusively contains data which can be raised objectively and is laboratory orientated (van HEES < 150 =mild disease; > 150 and < 2 10 =moderate disease; > 2 10 = severe disease). Mean BEST index for males was 64.5 (range - 17-216) and for females 66 (range -26-317). Mean van HEES index for males was 157 (range 96-229) and for females 142 (range 103-217). Forty-four out of 60 patients with Crohn's disease were in remission (BEST index < 150; 18 males and 26 females). 16 out of 60 patients suffered from clinical active disease (BEST index > 150; 8 males and 8 females). Further data concerning clinical features are given in Tables 1 and 2. Eighty-two sera were obtained from healthy blood donors (HBD). All serum samples were stored at - 20°C until used. Serologic l e s t s
All serum samples were tested for 15 different autoantibody specificities. Tests were done in triplicates. The methods used for each antibody are listed below: ANA, anti-mitochondria1 antibodies (AMA) and antibodies to parietal cells (PCA) were detected using indirect immunofluorescence microscopy (IFT) on rat stomach, kidney and liver tissue, HEp2-cells and CVIcells. HEp2-cells were commercially prepared (Antibodies Inc., Davis, CA, USA). CVI-cells were grown on cover slips air-dried and fixed with methanol/ EGTA for 4 min at - 20°C. Positive sera were serially diluted in PBS to obtain antibody titres. Stained tissues were analysed with a Zeiss Axiophot microscope equipped with epifluorescence optics (Zeiss, Oberkochen, FRG). A super-pressure mercury lamp (HBO
517
50) served as light source. An excitation filter BP 485/ 20 and a cut-off filter LP 520 were used. The results were documented by photography with Ilford XP-I film. Antibodies to extractable nuclear antigens (ENA) were detected by Ouchterlony immunodiffusion (ID) and counterimmunoelectrophoresis (CIE). Screening of sera was done by ID [6] of undiluted sera with an antigen concentration of 50 mg ml-' in 0.650/m agarose gel [7]: rabbit thymus extract (RTNE) (Paesel, Frankfurt, FRG) and human spleen extract (HSE) prepared as described by Venables et al. [8] served as antigens. CIE was performed as described by Tan [9] using standardized reference antisera (anti-Ro, La, Sm, RNP; Center for Disease Control, Atlanta, GA, USA). Additionally sera were tested by performing polyacrylamide gel electrophoresis (PAGE) and Western blotting. Segments of normal human small intestine obtained during the surgical removal of malignant tumours were used for the isolation of intestinal epithelial cells, followed by lysis and extraction using low- and high-salt buffers and Triton X- 100, essentially as previously described by Franke et al. [lo], which yielded cytoskeletal residues. Cytoskeletal preparation of human smooth muscle rich in actin and desmin was obtained by microdissection of the tunica muscularis propria of stomach followed by similar extraction steps. PAGE (10%) of RTNE, HSE. commercially available cytoskeletal antigens, cytoskeletal material of human intestinal cells and smooth muscle tissues was performed according to the system of Laemmli [ I I ] in the presence of O.l"/u NaDodS04 (SDS). The gels were stained with silver [ 121. For twodimensional gel electrophoresis, non-equilibrium pH gradient electrophoresis [ 131 was applied in the first dimension; in the second dimension SDS-PAGE was performed using the elevated salt-buffer system described by Achtstatter et al. [14]. Western blotting was performed as described previously [15,16] or as described in [ 171. Antibodies to double-stranded DNA (dsDNA) were measured by FARR- and PEG-assay [I81 and IFT on Crithidia luciliae (Antibodies Inc, Davis, CA, USA). Rheumatoid factors (RF) were measured by latex fixation and Waaler-Rose test (Behringwerke, Marburg, FRG). A CA-ELISA
Wells of a microtitre plate (M 129 A; Dynatech, Plochingen, FRG) were coated for 16 h at room temperature (24°C) with 100 pi of the respective purified antigen (cytokeratin no. 18 bovine, 981059 Boehringer, FRG; vimentin from bovine lens, 981 192 Boehringer; desmin from chicken stomach, 98 I 1 17 Boehringer; actin from rabbit muscle, 98 1052 Boehringer; tropomyosin from bovine muscle, T-4895 Sigma, FRG) diluted in 0.05%] sodiumcarbonate buffer pH 9.6. Concentrations ranging from 100 pg to 3 pg ml- I for coating antigen were tested and 5 pg ml-' was found to give optimum results for actin, desmin, vimentin and tropomyosin and 10 pg ml-' for cyto-
518
W. J . MAYET et al.
keratin 18. After washing once with PBS, each well was filled with PBS containing 1 % BSA and left at room temperature for 2 h in order to block unsaturated sites. After washing the wells with PBS containing 0.05% Tween 20 (three times), 100 pI of the respective serum diluted I :500 in PBS was added and incubated for 1 h at room temperature. Rabbit-anti-human peroxidase (IgG, gamma-chain specific; Dako P 214. IgM, p-chain specific; Dako P 322. IgA, a-chain specific; Dako P 216) diluted 1 : 1000 in PBS containing 0.1% BSA was added after washing the wells again three times with PBS containing 0.05‘!41 Tween 20. After 1 h incubation time at room temperature the wells were washed again three times with PBS containing 0.05% Tween 20 and 100 p1 of freshly prepared substrate ( 100 mmol 1 I acetate buffer, pH 5.4 containing 0.002‘!41 H202 and 0.43 g ml-’ ophenylene diamine; P-3888, Sigma, Deisenhofen, FRG) were added to each well. After I0-min incubation at room temperature the reaction was stopped by the addition of 100 pl 1 mmol I-’ H2S04. Optical densities (D)were measured at 492 nm in a doublebeam spectrophotometer (Beckmann, Model 25, Fullerton, CA, USA). ~
Statistical analysis
The upper limit of normality in the five ELISA systems was taken as two standard deviations above mean normal control levels. By performing Mann-Whitney test the average rank of patients with Crohn’s disease and HBD was calculated for each of the five autoantibodies including their respective isotypes (IgG, IgM, IgA). Correlations between these antibodies were saught using Spearman’s correlation coefficient. Linear regression analysis was performed to evaluate relationships between clinical features (i.e. anatomic location of involvement, extra-intestinal manifestations and surgical treatment), the indices of BEST and van HEES (i.e. clinical activity of disease) and antibody levels measured in the ELISA. Study for the reproducibility of the assay
Sera from five patients with Crohn’s disease that reacted positively in the desmin ELISA were assayed in different experiments on different days and another five sera were tested in five different experiments within one assay procedure to study the reproducibility of the assay (i.e. inter- and intra-assay coefficients). Results
Titration curves of three sera positive for desmin antibodies and of a pool of HBD are shown in Fig. 1. The ELISA systems provided a good discrimination between ACA-positive and ACA-negative sera from 60 patients with Crohn’s disease (Fig. 2a-e). As the respective cut-off point values above the mean optical density of the 82 HBD plus two standard deviations
2.5
2.0
I
I+
1.5
1.0
0.5
0
50
I00
500
1000
5000
Reciprocal o f serumdilution Figure 1. Titration curves for three anti-desmin positive sera ( A ) and a serum pool of healthy blood donors ( 0 )
were chosen: 0.470 (IgG), 0.398 (IgM) and 0.405 (IgA) forcytokeratin 18 antibodies,O.l71 (IgG), 0.121 (IgM) and 0.371 (IgA) for actin antibodies, 0.493 (IgG), 0.27 1 (IgM) and 0.550 (IgA) for desmin antibodies, 0.179 (IgG), 0.123 (IgM) and 0.548 (IgA) for vimentin antibodies, 0.330 (IgG), 0.41 I (IgM) and 0.543 (IgA) for tropomyosin antibodies. Antibodies detected in patients’ sera
In Table 3 the incidence of ACA (IgG-, IgM- and IgAtype) as determined by five ELISA systems in 60 patients with Crohn’s disease is shown. All sera classified to be positive in the ELISAs recognized the respective antigens in Western blot (data not shown). In this group of patients no ANA, antibodies to dsDNA, PCA, antibodies to ENA and AMA could be detected. All sera were RF negative. No antibodies could be found in the sera of 82 HBD. Performing the ELISAs, the optical densities of the Crohn’s group were significantly different from the optical densities of the HBD group as determined by statistical analysis. Low but significant correlations could be found for levels of cytokeratin 18 antibodies (IgM-type) and the BEST index of activity (0.30; P < 0.05) and for levels of desmin antibodies (IgM-type) and the van HEES index of activity (0.282; PcO.05). Table 4 shows positive correlations of different antibodies. Highest optical density ranges could be measured for actin antibodies (IgG-type) in patients with isolated disease manifestation in the colon. Vimentin and desmin antibodies (IgG-type) could be found only in sera from patients with colon involvement. Most of the patients had ACA against several isotypes and also against various cytoskeletal components (Table 3). Figure 3a shows that sera of all 12 patients classified as positive in the ELISA recognized cytokeratin 18 using an extract of
ANTIBODIES TO CYTOSKELETAL PROTEINS IN CROHN’S DISEASE
( a )
\
1.2
(b)
Keratin
)
Actin
0.6
-1
0
f0
. 0
.. 3
0
t
*:
8
3
._.. t----
!
0
(d 1 1.0 I
1.0 0
E N 6,
e +
0.8 -
0
0
0.8 0
0.6 -
t
_._....___
.
0
0.2 Oa4
n
0.8
% .
Ii (el
1
0.6
Viment i n 0-
11
.L....
c
0
om
I
0.2
..........
n
.
Tropomyosin
0
J
0 ‘ IgG
I gM
IOA
Figure 2. Results of ACA ELISAs in patients with Crohn’s disease. Cut off (-
519
- --
-); mean of healthy blood donors (.)+SD.
Table 3. Incidence of ACA in patients with Crohn's disease
Keratin
IgG IgM IgA
12 4 8
(20.0%) Actin IgG 22 (36.7%) Desmin IgG 4 (6.7%) Vimentin IgG 2 (3.3%) Tropomyosin IgG 2 (3.3%) IgM 9 ( I 5.0%) IgM 2 (3.3%) IgM 29 (48.3'Xi) IgM 10 (16'70/;1) (6.7%)) IgA 3 (5.0%)) (13.3"%) IgA 3 (5.3%) IgA 6 (IO'O'X,) IgA I 6 (26.7"/;,)
(b) Patient No.
Keratin
Actin
Desmin
Vimentin
Tropomyosin
I 2 3 4 5 6
GM
G
M M
G G
G
M
I 8 9 10
GMA
GM G
G
II
GM M
12 13 14 15 16
17 18 19 20 21 22 23 24 25 26 21 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 41 48 49 50 51 52 53 54 55 56 57 58 59 60
A G
M G
M M
G GM M GMA GM GM GMA A A GMA GMA MA M GM G A M
G
G
A MA A
A
MA M M M M
MA
M
MA A M
GM G A
G
A A
M GM G A A M
G
M MA M M
G
M M M G G
M A
M
M
M A
GM GM G
G
M
GM A A
G
A
M
A
G
G GM
A A
A
A A
ANTIBODIES TO CYTOSKELETAL PROTEINS IN CROHN’S DISEASE
total cytoskeletal proteins of human intestinal cells as source of antigen (Western blot after SDS-PAGE). These results were further confirmed in two-dimensional gel electrophoresis (Fig. 3b). The immunoblots also show the presence of autoantibodies against other cytokeratin polypeptides (Fig. 3a, b). Actin-positive sera (ELISA) also recognized actin by performing Western blot with a cytoskeletal preparation of human smooth muscle (data not shown).
Table 4. Correlation of ACA established using Spearman’s correlation coefficient Desmin-lgM + Keratin-lgM: Vimentin-lgM Keratin-lgM: Tropomyosin-IgM + Keratin-IgM: Tropomyosin-IgM + Desmin-IgM: Tropomyosin-lgG + Actin-lgG: Keratin-IgA Vimentin-IgA: Tropomyosin-lgA + Vimentin-IgA: Keratin-lgA + Desmin-lgA: Actin-igA + Keratin-IgA: Tropomyosin-lgA + Keratin-lgA: Tropomyosin-lgA + Desmin-IgA:
+
+
(a
1
521
r = 0.54; P < 0,005 r = 0.59; P < 0.005 r = 0 . 6 5 ; P
