Acta path. microbiol. scand. Sect. C, 84: 31-36, 1976
FRAGMENTS OF IMMUNOGLOBULINS IN HUMAN FAECES B. HANEBERC and C. ENDRESEN Department of Pediatrics and Microbiology, and Broegelmann Research Laboratory for Microbiology, University of Bergen, School of Medicine, Bergen, Norway
Haneberg, B. & Endresen, C. Fragments of immunoglobulins in human faeces. Acta path. microbiol. scand. Sect. C, 84: 31-36, 1976. Fab-fragments of IgG were easily demonstrated in extracts of faeces from healthy infants and children. Employing immunoelectrophoresis with antisera to whole human serum and to Fab-fragments of IgG, a marked precipitation line, likely to represent such fragments, was evident in the cathodal or y-region. Usually no precipitate was formed with antisera specific for y-heavy chains or Fc-fragments of IgG. Presumably IgG in the gut is partially destroyed before being excreted with faeces. Results of immunoelectrophoresis, gel filtration as well as polyacrylamide gel electrophoresis, using antisera to a-heavy chains, indicated the presence of fragments of IgA in some faecal extracts. Fragments of IgA could not be demonstrated in IgA-deficient patients, and no fragments of either class, or only traces of Fab-fragments of IgG, were found in agammaglobulinaemic patients. No antibody activity against rabbit erythrocytes was found in gel filtration fractions containing such fragments; any reaction was not observed either in the direct agglutination test or in a modified antiglobulin test. Key words: Immunoglobulin-fragments; copro-antibodies. Bjern Haneberg, Barneklinikken, N-50 16 Haukeland sykehus, Norway.
Received 8.ix.75
Accepted 8.ix.75
IgA is found in extracts of faeces from most infants and children, while IgG and IgM are only demonstrable in a small percentage ( 6 ) . The concentrations of IgA, in relation to dry weight of faeces, exceed those of the other classes. This difference can, to some extent, be explained by a preponderance of IgA-containing plasma cells in the intestinal mucosa ( 2 ) , indicating a high level of production. In addition, the relative resistance of high-molecular-weight secretory IgA to proteolytic enzymes ( 11) may contribute to the dominance of IgA in faeces since the number of intestinal plasma cells containing IgA is not equally marked in relation to the numbers of cells containing IgG and IgM
(2). A large part of the IgG and IgM found in the upper intestinal tract (1, 10, 12) is probably more or less degraded during passage through the gut. If exposed to pepsin at low p H even secretory IgA will be fragmented ( 11) , and large amounts of IgA subjected to this process are found in gastric secretion (8). A proteolytic enzyme, probably derived from enteric microorganisms may also split secretory IgA (9). Normally, however, only IgA in faeces seems to possess antibody activity if tested against rabbit erythrocytes (5) ; this is likely to be the result of a different sensitivity of the immunoglobulins to the intestinal enzymes. Since fragments of immunoglobulins can interfere with the quantitation of immuno31
globulins ( 11) we investigated their presence in the faecal extracts used. In this paper, the identification of the fragments normally present in faeces; using conventional antisera and separation techniques, are described. The fragments were also tested for activity against rabbit erythrocytes. MATERIALS AND METHODS
Individuals. One hundred healthy infants and children, l month to 12 years of age, were randomly selected at well-baby clinics and among children of hospital employees. None of the babies were receiving human milk. Three patients with ulcerative colitis in clinical remission and with high faecal levels of IgG, 2 patients with IgA-deficiency and 3 patients with agammaglobulinaemia ( 6 ) were also included. Faecal extracts. Samples of faeces, free from blood, were collected and extracts were made, using 10 ml phosphate buffered saline (PBS), p H 7.2 per g freeze-dried faecal material, as described previously ( 7 ) . Sera. Antisera to whole human serum, IgA, IgG and IgM were obtained by immunization of rabbits and made specific for heavy chains by absorptions with rabbit erythrocytes agglutinated by different immunoglobulin classes ( 7 ) . The corresponding antisera as well as rabbit antisera to human IgG (Fab)*, IgG (Fc)*, IgG (Fd)*, light chains of kappa and lambda types and to free light chains (Bence-Jones) of both types, were purchased (Behringwerke AG, Marburg-Lahn, Germany). A pool of sera from 10 healthy blood donors served as normal human serum. Gammaglobulin 16.5 per cent (Kabi, Stockholm, Sweden) was used as a source of human IgG. Gel filtration. Sephadex G-200 column, 1.5 x 45 cm, equilibrated a t p H 8 with 0.05 M tris-HCIbuffer with 0.14 M NaCl were used for separation of proteins in 10 different concentrated faecal extracts as outlined earlier ( 7 ) . Normal human serum was run as control for elution regions of the various proteins. Agglutination test. Gel filtration fractions were tested for agglutinins to rabbit erythrocytes as previously described ( 5 ) . Antiglobulin test. Two-fold dilutions of faecal extracts or gel filtration fractions of faecal extracts were used for sensitization of rabbit erythrocytes and tested for agglutination with antiserum to IgG (Fab) and not as previously (5) with antiserum to IgG. Polyacrylamide gel electrophoresis. The precipitates obtained from one faecal extract mixed with Manufacturer's terminology.
32
antiserum to IgA were repeatedly washed in large volumes of PBS and dissolved in buffer containing either 8 M urea or 1 per cent sodium dodecyl suIphate before electrophoresis in 7.5 per cent polyacrylamide gel with urea ( 3 ) and in 7.5 and 15 per cent gel with dodecyl sulphate (15). Precipitates prepared from normal human serum with antiserum to IgA were used as control. O t h e r methods. Double diffusion and immunoelectrophoresis in agar were carried out as formerly described ( 7 ) . Fragmentation of IgG by trypsin (Sigma Chemical Corp., St. Louis, Missouri, U.S.A.) 1 per cent w/v, was performed as outlined by Wa!ler et al. (14). EXPERIMENTS AND RESULTS
Immunoelectrophoresis of extracts of faeces from 12 healthy individuals, using antisera to whole human serum, invariably showed a strong precipitation line in the y-region in addition to the IgA-line and 1 or 2 lines in the a-region (Fig. 1-1). The mobility of the marked y-line was intermediate between those of serum IgA and IgG. A reaction of partial identity with this precipitation line and those obtained with antisera to IgG (Fab), IgG ( F d ) and to light chains of both kappa and lambda types was observed (Fig. 1-111, IV, V ) . However, no precipitate was formed with antisera to either IgG (Fig. 1-VI) or IgM, nor with antisera to IgG (Fc) or free light chains (Bence-Jones). The antisera to IgG (Fab) or IgG ( F d ) gave no precipitate in double diffusion against purified IgA, and the strong precipitation reaction with faecal extracts was not inhibited by absorption of the antisera with rabbit erythrocytes agglutinated by IgA or IgM. Thus, the precipitate did not solely represent light chains. I n a control system with trypsin-treated IgG, the antisera to IgG reacted with both Fab- and Fc-components, while the antiserum to whole serum reacted mainly with the Fabcomponent as identified by immunoelectrophoresis. Here too, absorption of the antisera with other classes of immunoglobulins, to remove anti-light-chain activity, did not influence the reactions. Gel filtration through a Sephadex G-200 column indicated that the components reacting with the antiserum to IgG (Fab) were
Fig. 1. Immunoelectrophoresis of normal faecal extract against antiserum to I ) whole human serum, 11) IgA, 111) IgG (Fab), IV) IgG ( F d ) , V ) kappa light chains and V I ) IgG. An additional weak precipitation line produced by the antiserum to IgA is indicated by the arrow.
+
eluted in the same region as a1 antitrypsin and albumin (Fig. 2 ) . Hence, the components were likely to represent monovalent Fab-fragments and not F(ab’)*. The antiserum to IgG (Fab) gave a strong precipitate in double diffusion with extracts of faeces from all of 100 healthy infants and children. Only a faint precipitate between antisera to IgG or IgG (Fc) and 8 of these individual extracts was seen. Thus, Fab-fragments of IgG seemed to be present in faeces from normal individuals. The faecal IgG (Fab) precipitate was also seen in patients with extremely low serum IgA and no demonstrable IgA in faeces. Faecal extracts from 3 patients with agamma3 Acta path. microbiol. scand. Sect. C, 84, 1
I
globulinaemia had no y-precipitates on immunoelectrophoresis. In double diffusion, however, a weak reaction with the extract from one of them against antiserum to IgG (Fab) was observed. The gel filtration fractions likely to contain Fab-fragments of IgG did not agglutinate rabbit erythrocytes. Antiglobulin tests of extracts of faeces and gel filtration fractions, using antiserum to IgG (Fab), did not increase the titre or give visible agglutination of rabbit erythrocytes. It therefore seems as if the Fab-fragments of IgG in faeces fail to react with these erythrocytes. Immunoelectrophoresis of the extracts of faeces from 2 of the healthy individuals re-
33
100
11,
antiirypsin
,,,,,,, ,,,, ,,,,,,, , ,,, ,,,, ,,, ,, ,,,, ,,
10
20
PO
40
,
50 Fracllon no.
Fig. 2. Gel filtration through Sephadex G-200 (1.5 x 45 cm) of a ten-fold concentrate of normal faecal extract. The flow rate was 11 ml/cm2/ hour and fractions, about 2 ml each, were collected during continuous registration of UV-light transmittance. T h e presence of precipitates in double diffusion of the fractions and various antisera is indicated. The dashed line indicates a faint precipitate.
vealed a double precipitation line with antisera to IgA (Fig. 1-11). One of the lines gave a reaction of identity with the IgA-line against antisera to whole human serum. I t is possible that the other line (indicated by arrow) represents fragments of IgA, as demonstrated for IgG. The results of Sephadex G-200 gel filtration support this; fractions from the region in which the fragments of IgG were found, gave a precipitate in double diffusion and in single radial immunodiffusion with antisera to IgA (Fig. 2). This precipitate was distinctly weaker than the usual IgA-precipitate, being visible only after a t least 48 hours’ incubation. Moreover, absorptions of the antisera to IgA with other immunoglobulin classes did not remove this precipitate which was not found in the gel filtration fractions of faecal extracts from patients with IgA-deficiency, and thus, it was not cross-reacting with Fab-fragments of IgG. Polyacrylamide gel electrophoresis of precipitates from faecal extract obtained by antiserum to IgA and dissolved in urea, gave at least 4 bands distinct from the intact immunoglobulin (Fig. 3 ) . The results of parallel controls using normal human serum precipitated by the same antiserum to IgA gave no bands below the usual immunoglobulin level. After electrophoresis in urea, slices of
34
a n unstained gel, corresponding in position to the stained bands, were put into a small amount of PBS, dialysed and freeze-dried before they were dissolved in 100 pl PBS. These extracts of the gel all gave a precipitate in agar against antiserum to IgA. Electrophoresis using sodium dodecyl sulphate gave a t least 3 bands, indicating the presence of fragments with molecular weights 50,000100,000 (Fig. 3 ) . DISCUSSION
The present results of immunoelectrophoresis, double diffusion and gel filtration provide evidence for the occurrence of Fab-fragments of IgG in extracts of faeces from healthy infants and children. The absence of reactivity with antisera to y-heavy chains may be the result of proteolytic action on intestinal IgG, which is probably more pronounced in the gut than in the control system in which trypsin is acting on IgG in vitro. Results of this in uitro degradation indicated that the antisera to y-heavy chains are capable of reacting with the Fab- as well as the Fc-frag-
Fig. 3. Polyacrylamide gel electrophoresis in 8 M urea (A, B) and in 1 per cent sodium dodecyl sulphate (C, D) of precipitates of faecal extract (A, C, D) and of normal human serum ( B ) obtained using an antiserum to IgA. T h e urea gel was 7.5 per cent and the dodecyl sulphate gel 7.5 per cent ( C ) and 15 per cent (D). Stained bands below the band corresponding to intact immunoglobulins are indicated by arrows.
ment. The possibility that some fragments of IgG may also be secreted as “unfinished” molecules, however, cannot be ruled out. T h e electrophoretic mobility of the heavy precipitate in the y-region, probably representing Fab-fragments of IgG, was less cathodal than that of native serum IgG. This precipitate might be misinterpreted as representing IgA. However, only a small part of this precipitate reacted with antisera to IgA. Nevertheless, in some faecal extracts, the double lines obtained on immunoelectrophoresis suggest that fragments of IgA were also present. Such double lines have previously been observed following peptic digestion of intestinal juice ( 1 1) and following exposure of serum as well as colostral IgA by an enteric microbial enzyme ( 9 ) . T h e results of gel filtration experiments indicate that the molecular size of any fragments of IgA present was close to that of Fab-fragments of IgG, a, antitrypsin and albumin. The multiple bands obtained by polyacrylamide gel electrophoresis in urea also suggests that it is not only intact IgA in faeces which is precipitated by the antiserum to IgA. Dodecyl sulphate was used in order to obtain information about the size of the molecules reacting with the antiserum to IgA. The results, like those obtained in urea-electrophoresis, were not unequivocal since the IgAz subclass lacks the disulphide bridges between the heavy and light chains, and therefore will probably be fragmented during analysis (4). However, only one or two “small-molecular-bands”, representing light chains are then reported to be formed ( 4 ) . In our experiments, a-heavy chain determinants were demonstrated in all the regions of the urea gel corresponding to stained bands, The serum control appeared to confirm that the smaller molecules to be observed when this technique is used were not solely caused by in vitro fragmentation of IgA, although this subclass makes up a relatively small part of the IgA in serum compared to secretions (4, 13). Fragments of IgA may interfere with the quantitative determination of IgA in single 3*
radial immunodiffusion ; smaller molecules diffuse more readily through agar, resulting in erroneously high concentrations (11) . I n our system, however, the late appearance and the weak precipitation of the presumed fragments of IgA seemed to exclude the possibility of any such influence on the results of quantitation. The fragments found may explain the multiple precipitin rings observed by others (10). The findings in extracts of faeces from patients with immunoglobulin deficiencies seemed to confirm the specificity of the reactions observed. T h e presence of fragments of IgG appeared to be independent of the presence of fragments of IgA. Fragments of IgA were not seen in faecal extracts from IgAdeficient patients. In IgA-deficient individuals, the relatively large amounts of IgM excreted with faeces retain some antibody activity against rabbit erythrocytes ( 6 ) . However, no antibody activity of faecal IgG was demonstrated, even when present in faeces in amounts corresponding to those of IgM. Neither did the antiserum to IgG (Fab) , reacting strongly with the presumptively large amounts of fragmented IgG, lead to agglutination of the rabbit erythrocytes in the antiglobulin test. I t therefore seems probable that the activity of IgG in the gut is shortlived.
REFERENCES 1. Bull, D . M., Bienenstock, J . & TomaJi, T . B. jr.: Studies on human intestinal immunoglobulin A. Gastroenterology 60: 370-380, 1971. 2. Crabbt, P . A . & Heremans, J. F.: The distribution of immunoglobulin-containing cells along the human gastrointestinal tract. Gastroenterology 51: 305-3 16, 1966. 3. Davis, B. 1.: Disc electrophoresis. 11. Method and application to human serum proteins. Ann. N.Y. Acad. Sci. 121: 404-427, 1964. 4. Grey, H . M., Abel, C . A., Yount, W . J. & Kunkel, H . G.: A subclass of human yA-globulins (yA,) which lacks the disulphide bonds linking heavy and light chains. J. exptl. Med. 128: 1223-1236, 1968. 5 . Haneberg, B.: Human fecal agglutinins to rabbit erythrocytes. Scand. J. Immunol. 3 : 71-76, 1974.
35
6. Haneberg, B. & Aarskog, D.: Human fecal immunoglobulins in healthy infants and children, and in some with diseases affecting the intestinal tract or the immune system. Clin. exp. Immunol. I n press 1975. 7. Haneberg, B. & Tsnder, 0.: Immunoglobulins and other serum proteins in feces from infants and children. Scand. J. Immunol. 2: 375-383, 1973. 8 . McClelland, D . B. L., Finlayson, N . D . C., Samson, R . R., Nairn, I . M . & Shearman, D . I . C . : Quantitation of immunoglobulins in gastric juice by electroimmunodiffusion. Gastroenterology 60: 509-514, 1971. 9. Mehta, S. K., Plaut, A . G., Calvanico, N . J . & Tomasi, T . B. jr.: Human immunoglobulin A: Production of an Fc fragment by an enteric microbial proteolytic enzyme. J. Immunol. 1 1 1 : 1274-1276, 1973. 10. Plaut, A . G . & Keonil, P.: Immunoglobulins in human small intestinal fluid. Gastroenterology 56: 522-530, 1969.
36
1 1 . Samson, R . R . , McClelland, D . B. L. & Shearman, D . J . C.: Studies on the quantitation of immunoglobulins in human intestinal secretions. Gut 14: 616-626, 1973. 12. Savilahti, E.: Immunoglobulin-containing cells in the intestinal mucosa and immunoglobulins in the intestinal juice in children. Clin. exp. Immunol. ZI: 415-425, 1972. 13. Vaerman, 1.-P.: Subdivision of IgA in two subclasses. I n : Studies on IgA immunoglobulins in man and animals, Louvain, 1970, p. 69-86. 14. Waller, M . , Richard, A . J . & Mallory, 1.: Immunochemical and serological studies of enzymatically fragmented human IgG globulins. 11. Hydrolysis with subtilisin, elastase, trypsin and chymotrypsin. Immunochemistry 6: 207214, 1969. 15. Weber, K . & Osborn, M . : T h e reliability of molecular weight determinations by dodecylsulphate-polyacrylamide gel electrophoresis. J. Biol. Chem. 244: 4406-4412, 1969.
FRAGMENTS OF IMMUNOGLOBULINS IN HUMAN FAECES B. HANEBERC and C. ENDRESEN Department of Pediatrics and Microbiology, and Broegelmann Research Laboratory for Microbiology, University of Bergen, School of Medicine, Bergen, Norway
Haneberg, B. & Endresen, C. Fragments of immunoglobulins in human faeces. Acta path. microbiol. scand. Sect. C, 84: 31-36, 1976. Fab-fragments of IgG were easily demonstrated in extracts of faeces from healthy infants and children. Employing immunoelectrophoresis with antisera to whole human serum and to Fab-fragments of IgG, a marked precipitation line, likely to represent such fragments, was evident in the cathodal or y-region. Usually no precipitate was formed with antisera specific for y-heavy chains or Fc-fragments of IgG. Presumably IgG in the gut is partially destroyed before being excreted with faeces. Results of immunoelectrophoresis, gel filtration as well as polyacrylamide gel electrophoresis, using antisera to a-heavy chains, indicated the presence of fragments of IgA in some faecal extracts. Fragments of IgA could not be demonstrated in IgA-deficient patients, and no fragments of either class, or only traces of Fab-fragments of IgG, were found in agammaglobulinaemic patients. No antibody activity against rabbit erythrocytes was found in gel filtration fractions containing such fragments; any reaction was not observed either in the direct agglutination test or in a modified antiglobulin test. Key words: Immunoglobulin-fragments; copro-antibodies. Bjern Haneberg, Barneklinikken, N-50 16 Haukeland sykehus, Norway.
Received 8.ix.75
Accepted 8.ix.75
IgA is found in extracts of faeces from most infants and children, while IgG and IgM are only demonstrable in a small percentage ( 6 ) . The concentrations of IgA, in relation to dry weight of faeces, exceed those of the other classes. This difference can, to some extent, be explained by a preponderance of IgA-containing plasma cells in the intestinal mucosa ( 2 ) , indicating a high level of production. In addition, the relative resistance of high-molecular-weight secretory IgA to proteolytic enzymes ( 11) may contribute to the dominance of IgA in faeces since the number of intestinal plasma cells containing IgA is not equally marked in relation to the numbers of cells containing IgG and IgM
(2). A large part of the IgG and IgM found in the upper intestinal tract (1, 10, 12) is probably more or less degraded during passage through the gut. If exposed to pepsin at low p H even secretory IgA will be fragmented ( 11) , and large amounts of IgA subjected to this process are found in gastric secretion (8). A proteolytic enzyme, probably derived from enteric microorganisms may also split secretory IgA (9). Normally, however, only IgA in faeces seems to possess antibody activity if tested against rabbit erythrocytes (5) ; this is likely to be the result of a different sensitivity of the immunoglobulins to the intestinal enzymes. Since fragments of immunoglobulins can interfere with the quantitation of immuno31
globulins ( 11) we investigated their presence in the faecal extracts used. In this paper, the identification of the fragments normally present in faeces; using conventional antisera and separation techniques, are described. The fragments were also tested for activity against rabbit erythrocytes. MATERIALS AND METHODS
Individuals. One hundred healthy infants and children, l month to 12 years of age, were randomly selected at well-baby clinics and among children of hospital employees. None of the babies were receiving human milk. Three patients with ulcerative colitis in clinical remission and with high faecal levels of IgG, 2 patients with IgA-deficiency and 3 patients with agammaglobulinaemia ( 6 ) were also included. Faecal extracts. Samples of faeces, free from blood, were collected and extracts were made, using 10 ml phosphate buffered saline (PBS), p H 7.2 per g freeze-dried faecal material, as described previously ( 7 ) . Sera. Antisera to whole human serum, IgA, IgG and IgM were obtained by immunization of rabbits and made specific for heavy chains by absorptions with rabbit erythrocytes agglutinated by different immunoglobulin classes ( 7 ) . The corresponding antisera as well as rabbit antisera to human IgG (Fab)*, IgG (Fc)*, IgG (Fd)*, light chains of kappa and lambda types and to free light chains (Bence-Jones) of both types, were purchased (Behringwerke AG, Marburg-Lahn, Germany). A pool of sera from 10 healthy blood donors served as normal human serum. Gammaglobulin 16.5 per cent (Kabi, Stockholm, Sweden) was used as a source of human IgG. Gel filtration. Sephadex G-200 column, 1.5 x 45 cm, equilibrated a t p H 8 with 0.05 M tris-HCIbuffer with 0.14 M NaCl were used for separation of proteins in 10 different concentrated faecal extracts as outlined earlier ( 7 ) . Normal human serum was run as control for elution regions of the various proteins. Agglutination test. Gel filtration fractions were tested for agglutinins to rabbit erythrocytes as previously described ( 5 ) . Antiglobulin test. Two-fold dilutions of faecal extracts or gel filtration fractions of faecal extracts were used for sensitization of rabbit erythrocytes and tested for agglutination with antiserum to IgG (Fab) and not as previously (5) with antiserum to IgG. Polyacrylamide gel electrophoresis. The precipitates obtained from one faecal extract mixed with Manufacturer's terminology.
32
antiserum to IgA were repeatedly washed in large volumes of PBS and dissolved in buffer containing either 8 M urea or 1 per cent sodium dodecyl suIphate before electrophoresis in 7.5 per cent polyacrylamide gel with urea ( 3 ) and in 7.5 and 15 per cent gel with dodecyl sulphate (15). Precipitates prepared from normal human serum with antiserum to IgA were used as control. O t h e r methods. Double diffusion and immunoelectrophoresis in agar were carried out as formerly described ( 7 ) . Fragmentation of IgG by trypsin (Sigma Chemical Corp., St. Louis, Missouri, U.S.A.) 1 per cent w/v, was performed as outlined by Wa!ler et al. (14). EXPERIMENTS AND RESULTS
Immunoelectrophoresis of extracts of faeces from 12 healthy individuals, using antisera to whole human serum, invariably showed a strong precipitation line in the y-region in addition to the IgA-line and 1 or 2 lines in the a-region (Fig. 1-1). The mobility of the marked y-line was intermediate between those of serum IgA and IgG. A reaction of partial identity with this precipitation line and those obtained with antisera to IgG (Fab), IgG ( F d ) and to light chains of both kappa and lambda types was observed (Fig. 1-111, IV, V ) . However, no precipitate was formed with antisera to either IgG (Fig. 1-VI) or IgM, nor with antisera to IgG (Fc) or free light chains (Bence-Jones). The antisera to IgG (Fab) or IgG ( F d ) gave no precipitate in double diffusion against purified IgA, and the strong precipitation reaction with faecal extracts was not inhibited by absorption of the antisera with rabbit erythrocytes agglutinated by IgA or IgM. Thus, the precipitate did not solely represent light chains. I n a control system with trypsin-treated IgG, the antisera to IgG reacted with both Fab- and Fc-components, while the antiserum to whole serum reacted mainly with the Fabcomponent as identified by immunoelectrophoresis. Here too, absorption of the antisera with other classes of immunoglobulins, to remove anti-light-chain activity, did not influence the reactions. Gel filtration through a Sephadex G-200 column indicated that the components reacting with the antiserum to IgG (Fab) were
Fig. 1. Immunoelectrophoresis of normal faecal extract against antiserum to I ) whole human serum, 11) IgA, 111) IgG (Fab), IV) IgG ( F d ) , V ) kappa light chains and V I ) IgG. An additional weak precipitation line produced by the antiserum to IgA is indicated by the arrow.
+
eluted in the same region as a1 antitrypsin and albumin (Fig. 2 ) . Hence, the components were likely to represent monovalent Fab-fragments and not F(ab’)*. The antiserum to IgG (Fab) gave a strong precipitate in double diffusion with extracts of faeces from all of 100 healthy infants and children. Only a faint precipitate between antisera to IgG or IgG (Fc) and 8 of these individual extracts was seen. Thus, Fab-fragments of IgG seemed to be present in faeces from normal individuals. The faecal IgG (Fab) precipitate was also seen in patients with extremely low serum IgA and no demonstrable IgA in faeces. Faecal extracts from 3 patients with agamma3 Acta path. microbiol. scand. Sect. C, 84, 1
I
globulinaemia had no y-precipitates on immunoelectrophoresis. In double diffusion, however, a weak reaction with the extract from one of them against antiserum to IgG (Fab) was observed. The gel filtration fractions likely to contain Fab-fragments of IgG did not agglutinate rabbit erythrocytes. Antiglobulin tests of extracts of faeces and gel filtration fractions, using antiserum to IgG (Fab), did not increase the titre or give visible agglutination of rabbit erythrocytes. It therefore seems as if the Fab-fragments of IgG in faeces fail to react with these erythrocytes. Immunoelectrophoresis of the extracts of faeces from 2 of the healthy individuals re-
33
100
11,
antiirypsin
,,,,,,, ,,,, ,,,,,,, , ,,, ,,,, ,,, ,, ,,,, ,,
10
20
PO
40
,
50 Fracllon no.
Fig. 2. Gel filtration through Sephadex G-200 (1.5 x 45 cm) of a ten-fold concentrate of normal faecal extract. The flow rate was 11 ml/cm2/ hour and fractions, about 2 ml each, were collected during continuous registration of UV-light transmittance. T h e presence of precipitates in double diffusion of the fractions and various antisera is indicated. The dashed line indicates a faint precipitate.
vealed a double precipitation line with antisera to IgA (Fig. 1-11). One of the lines gave a reaction of identity with the IgA-line against antisera to whole human serum. I t is possible that the other line (indicated by arrow) represents fragments of IgA, as demonstrated for IgG. The results of Sephadex G-200 gel filtration support this; fractions from the region in which the fragments of IgG were found, gave a precipitate in double diffusion and in single radial immunodiffusion with antisera to IgA (Fig. 2). This precipitate was distinctly weaker than the usual IgA-precipitate, being visible only after a t least 48 hours’ incubation. Moreover, absorptions of the antisera to IgA with other immunoglobulin classes did not remove this precipitate which was not found in the gel filtration fractions of faecal extracts from patients with IgA-deficiency, and thus, it was not cross-reacting with Fab-fragments of IgG. Polyacrylamide gel electrophoresis of precipitates from faecal extract obtained by antiserum to IgA and dissolved in urea, gave at least 4 bands distinct from the intact immunoglobulin (Fig. 3 ) . The results of parallel controls using normal human serum precipitated by the same antiserum to IgA gave no bands below the usual immunoglobulin level. After electrophoresis in urea, slices of
34
a n unstained gel, corresponding in position to the stained bands, were put into a small amount of PBS, dialysed and freeze-dried before they were dissolved in 100 pl PBS. These extracts of the gel all gave a precipitate in agar against antiserum to IgA. Electrophoresis using sodium dodecyl sulphate gave a t least 3 bands, indicating the presence of fragments with molecular weights 50,000100,000 (Fig. 3 ) . DISCUSSION
The present results of immunoelectrophoresis, double diffusion and gel filtration provide evidence for the occurrence of Fab-fragments of IgG in extracts of faeces from healthy infants and children. The absence of reactivity with antisera to y-heavy chains may be the result of proteolytic action on intestinal IgG, which is probably more pronounced in the gut than in the control system in which trypsin is acting on IgG in vitro. Results of this in uitro degradation indicated that the antisera to y-heavy chains are capable of reacting with the Fab- as well as the Fc-frag-
Fig. 3. Polyacrylamide gel electrophoresis in 8 M urea (A, B) and in 1 per cent sodium dodecyl sulphate (C, D) of precipitates of faecal extract (A, C, D) and of normal human serum ( B ) obtained using an antiserum to IgA. T h e urea gel was 7.5 per cent and the dodecyl sulphate gel 7.5 per cent ( C ) and 15 per cent (D). Stained bands below the band corresponding to intact immunoglobulins are indicated by arrows.
ment. The possibility that some fragments of IgG may also be secreted as “unfinished” molecules, however, cannot be ruled out. T h e electrophoretic mobility of the heavy precipitate in the y-region, probably representing Fab-fragments of IgG, was less cathodal than that of native serum IgG. This precipitate might be misinterpreted as representing IgA. However, only a small part of this precipitate reacted with antisera to IgA. Nevertheless, in some faecal extracts, the double lines obtained on immunoelectrophoresis suggest that fragments of IgA were also present. Such double lines have previously been observed following peptic digestion of intestinal juice ( 1 1) and following exposure of serum as well as colostral IgA by an enteric microbial enzyme ( 9 ) . T h e results of gel filtration experiments indicate that the molecular size of any fragments of IgA present was close to that of Fab-fragments of IgG, a, antitrypsin and albumin. The multiple bands obtained by polyacrylamide gel electrophoresis in urea also suggests that it is not only intact IgA in faeces which is precipitated by the antiserum to IgA. Dodecyl sulphate was used in order to obtain information about the size of the molecules reacting with the antiserum to IgA. The results, like those obtained in urea-electrophoresis, were not unequivocal since the IgAz subclass lacks the disulphide bridges between the heavy and light chains, and therefore will probably be fragmented during analysis (4). However, only one or two “small-molecular-bands”, representing light chains are then reported to be formed ( 4 ) . In our experiments, a-heavy chain determinants were demonstrated in all the regions of the urea gel corresponding to stained bands, The serum control appeared to confirm that the smaller molecules to be observed when this technique is used were not solely caused by in vitro fragmentation of IgA, although this subclass makes up a relatively small part of the IgA in serum compared to secretions (4, 13). Fragments of IgA may interfere with the quantitative determination of IgA in single 3*
radial immunodiffusion ; smaller molecules diffuse more readily through agar, resulting in erroneously high concentrations (11) . I n our system, however, the late appearance and the weak precipitation of the presumed fragments of IgA seemed to exclude the possibility of any such influence on the results of quantitation. The fragments found may explain the multiple precipitin rings observed by others (10). The findings in extracts of faeces from patients with immunoglobulin deficiencies seemed to confirm the specificity of the reactions observed. T h e presence of fragments of IgG appeared to be independent of the presence of fragments of IgA. Fragments of IgA were not seen in faecal extracts from IgAdeficient patients. In IgA-deficient individuals, the relatively large amounts of IgM excreted with faeces retain some antibody activity against rabbit erythrocytes ( 6 ) . However, no antibody activity of faecal IgG was demonstrated, even when present in faeces in amounts corresponding to those of IgM. Neither did the antiserum to IgG (Fab) , reacting strongly with the presumptively large amounts of fragmented IgG, lead to agglutination of the rabbit erythrocytes in the antiglobulin test. I t therefore seems probable that the activity of IgG in the gut is shortlived.
REFERENCES 1. Bull, D . M., Bienenstock, J . & TomaJi, T . B. jr.: Studies on human intestinal immunoglobulin A. Gastroenterology 60: 370-380, 1971. 2. Crabbt, P . A . & Heremans, J. F.: The distribution of immunoglobulin-containing cells along the human gastrointestinal tract. Gastroenterology 51: 305-3 16, 1966. 3. Davis, B. 1.: Disc electrophoresis. 11. Method and application to human serum proteins. Ann. N.Y. Acad. Sci. 121: 404-427, 1964. 4. Grey, H . M., Abel, C . A., Yount, W . J. & Kunkel, H . G.: A subclass of human yA-globulins (yA,) which lacks the disulphide bonds linking heavy and light chains. J. exptl. Med. 128: 1223-1236, 1968. 5 . Haneberg, B.: Human fecal agglutinins to rabbit erythrocytes. Scand. J. Immunol. 3 : 71-76, 1974.
35
6. Haneberg, B. & Aarskog, D.: Human fecal immunoglobulins in healthy infants and children, and in some with diseases affecting the intestinal tract or the immune system. Clin. exp. Immunol. I n press 1975. 7. Haneberg, B. & Tsnder, 0.: Immunoglobulins and other serum proteins in feces from infants and children. Scand. J. Immunol. 2: 375-383, 1973. 8 . McClelland, D . B. L., Finlayson, N . D . C., Samson, R . R., Nairn, I . M . & Shearman, D . I . C . : Quantitation of immunoglobulins in gastric juice by electroimmunodiffusion. Gastroenterology 60: 509-514, 1971. 9. Mehta, S. K., Plaut, A . G., Calvanico, N . J . & Tomasi, T . B. jr.: Human immunoglobulin A: Production of an Fc fragment by an enteric microbial proteolytic enzyme. J. Immunol. 1 1 1 : 1274-1276, 1973. 10. Plaut, A . G . & Keonil, P.: Immunoglobulins in human small intestinal fluid. Gastroenterology 56: 522-530, 1969.
36
1 1 . Samson, R . R . , McClelland, D . B. L. & Shearman, D . J . C.: Studies on the quantitation of immunoglobulins in human intestinal secretions. Gut 14: 616-626, 1973. 12. Savilahti, E.: Immunoglobulin-containing cells in the intestinal mucosa and immunoglobulins in the intestinal juice in children. Clin. exp. Immunol. ZI: 415-425, 1972. 13. Vaerman, 1.-P.: Subdivision of IgA in two subclasses. I n : Studies on IgA immunoglobulins in man and animals, Louvain, 1970, p. 69-86. 14. Waller, M . , Richard, A . J . & Mallory, 1.: Immunochemical and serological studies of enzymatically fragmented human IgG globulins. 11. Hydrolysis with subtilisin, elastase, trypsin and chymotrypsin. Immunochemistry 6: 207214, 1969. 15. Weber, K . & Osborn, M . : T h e reliability of molecular weight determinations by dodecylsulphate-polyacrylamide gel electrophoresis. J. Biol. Chem. 244: 4406-4412, 1969.
