mucosa - gutlymphoid follicles. agrid (leitz code no. 519902) representing 0.032 mm2 was moved...

Gut, 1986, 27, 667-674

Distribution of immunoglobulin producing cells isdifferent in normal human appendix and colonmucosa

K BJERKE, P BRANDTZAEG, AND T 0 ROGNUM

From the Laboratory for Immunohistochemistry and Immunopathology, Institute of Pathology, The NationalHospital, University of Oslo, Rikshospitalet, Oslo, Norway.

SUMMARY The densities of IgG-, IgA-, IgM- and IgD-producing immunocytes were determinedby paired immunofluorescence staining and morphometric analysis in the lamina propria ofnormal appendix specimens. Normal colon specimens were used as reference material, mostlypaired from individual subjects. The density (median of cells/mm2 lamina propria area) of IgAimmunocytes tended to be slightly higher in the appendix than in the colon (1259 vs 962) and thesame held true for IgM cells (71 vs 55). Conversely, the overall density of IgG immunocytes wasmuch higher in the appendix than in the colon (95 vs 38). A striking feature was the fact thatalmost 50% of all immunocytes were of the IgG isotype adjacent to lymphoid follicles. It seemedjustified to conclude, therefore, that the abundance of such follicles explains the overallenrichment of IgG-producing cells in normal appendix mucosa. These immunocytes most likelyrepresent follicle derived B cells that have reached terminal maturation locally, whereasprecursors generated from less mature memory clones probably emigrate and home ubiquitouslyto distant sites of the gut lamina propria where they develop into IgA-producing immunocytes.

Since the pioneering work of Crabbe andHeremans1 in the 1960s, several immunohistoche-mical studies have been published about the im-munoglobulin (Ig)-producing immunocytes in nor-mal adult large bowel mucosa (Table 1). All of thereports agree with the remarkable normal pre-ponderance of IgA immunocytes, but large varia-tions appear with regard to the proportions of cellsproducing the IgM and IgG isotypes (3%-34% and2%-35%, respectively). We have discussed else-where some of the technical problems that mayexplain these excessive discrepancies. 14 15 The varia-tions are apparently independent of whether im-munofluorescence or immunoperoxidase methodshave been used and seem to be unrelated to thebowel segment studied (Table 1). When attemptshave been made to estimate mucosal densities ofimmunocytes (cells/mm2 lamina propria area), thediscrepancies are likewise unacceptably large, evenfor IgA-producing cells (Table 1).

Addrcss for corrcspondencc: Dr K Bjerkc Institutc of Pathology. Rikshospi-talct. MN127 Oslo 1. Norway.

Rcceived for publication 26 Scptembcr 1985

The human appendix has been subjected tosurprisingly few immunohistochemical studies de-spite the fact that its abundant content of organisedlymphoid tissue suggests an active immunologicalstate. Indiveri et all examined appendectomy speci-mens and reported e ual numbers of IgA and IgGimmunocytes. Chen' found that the numbers ofIgG-producing cells depended on the inflammatorystage; the lowest number (13%) appeared in speci-mens defined histologically as 'early stage'. Data fornormal appendix mucosa have apparently not beenpublished previously.Our previous quantitative studies have been

based on paired immunofluorescence staining re-corded on double exposed colour slides.5 Ill Wehave now adopted mounting of tissue sections withmedium containing paraphenylenediamine (PPD), aprocedure that makes it possible to carry outprolonged and repeated observations directly in thefluorescence microscope without significant loss ofstaining intensity. '8 The aim of this study was to seepartly whether it is possible to reproduce ourprevious quantitative reference data on immuno-cytes in normal large bowel mucosa by doing direct

667

on March 31, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.27.6.667 on 1 June 1986. D

ownloaded from

http://gut.bmj.com/

Table 1 Various reports on isotype distribution (%) and density (d)* of Ig-producing cells in normal adult human largebowel mucosat

Bowel segment IgA (%) IgM (%) IgG (%) IgA (d) IgM (d) IgG (d) Methodt References

DIFDIF J

1523 115 47 DIFDIFDIFDIFDIFDIF

1497 401 201 DIF749 120 17 DIF1375 900 365 DIF1186 73 49 DIF790 63 118 IPT|806 70 121 IPT J

PAP707 137 451 PAP962 55 38 DIF

Crabbe and Heremans (1966)'

Gelzayd et al (1968)2Soltoft et al (1973)3Skinner and Whitehead (1974)4Baklien and Brandtzaeg (1975)5

Brown et al (1975)6Green and Fox (1975)7Meuwissen et al (1976)'O'Donoghue and Kumar (1979)9Rognum et al (1979)"'

Rosekrans et al (1980)"

Otto and Gebbers (1981)12Regadera et al (1983)'3Present study

*d=density of immunocytes in lamina propria (cells/mm2 section area).tSome of the data from other laboratories have been recalculated to make comparison possible.tDIF=direct immunofluorescence technique; PAP=peroxidase-antiperoxidase technique; IPT=indirect immunoperoxidase technique.Italicised figures refer to studies from the authors' laboratory.

cell counting combined with histomorphometry.The latter approach was in fact required for themajor purpose of this study, namely to examinewhether the presence of organised lymphoid tissuewould influence the composition of the adjacentmucosal immunocyte population. For this purposewe decided to carry out a comparative investigationof normal large bowel and appendix mucosa.

Methods

MUCOSAL SPECIMENSTissue specimens were obtained from patients oper-ated for cancer of the large bowel (at least 6 cm

away from the tumour) and were considered normalaccording to histological features. In addition, weobtained normal samples from four kidney donors inwhom peripheral circulation was artificially main-tained. Paired samples from colon and appendixwere obtained in nine cases. Altogether the studywas based on specimens from seven women andeight men with a median age of 48 years (range 7-79years).

IMMUNOHISTOCHEMICAL PROCEDURESThin tissue slices (2-3 mm thick) were excised fromgross specimens as soon as possible after theirremoval, oriented on filter paper, extracted in cold0-01 M phosphate buffer (pH 7-6) containing 0 15 MNaCl for 48 hours to remove diffusible extracellular

Ig components, fixed in cold ethanol, and processedfor paraffin embedding.'9

Serial sections were cut at 6 .tm perpendicular tothe mucosal surface and were evaluated by paireddirect immunofluorescence staining with rabbit IgGconjugates monospecific for human IgA, IgG, IgM,or IgD. The reagents were prepared in our labora-tory by labelling with fluorescein isothiocyanate(FITC) or tetramethylrhodamine isothiocyanate(TRITC); their characteristics have been reportedelsewhere.2t) The contrasting pairs of conjugates,applied for 20 hours at room temperature onadjacent sections, appear in Table 2.

FLUORESCENCE MICROSCOPYThe microscope used was a Leitz Orthoplan equip-ped with an Osram HBO 200 W lamp for TRITC(red) and an XBO 150 W lamp for FITC (green)

Table 2 Isotype specificities of paired fluorochromeconjugates used for immunocyte enumeration

Serial Rhodamine ('red') Fluorescein ('green')tissue section conjugates conjugates

1 a chain (IgA) y chain (IgG)2 y chain (IgG) a chain (IgA)3 ,u chain (IgM) a chain (IgA)4 6 chain (IgD) R chain (IgM)

ColonRectumRectumRectumRectumColonRectumRectum/colonColon/ileumColon/rectumRectumColon/rectumColonRectumColonColonColon

91-585-290-465-792-690-889-383-071-384 552 190-781-480-981-154 690-2

4-59.36-817-52-9565.311-819-113-634-15 66 57-013-210-64-6

4-05.52-816-84-53-65-55-29-61-9

13-83-712-212 15-7

34-830

668 Bjerke, Brandtzaeg, and Rognum



j.com/


ownloaded from

http://gut.bmj.com/

Ig production in appendix and colon mucosa

fluorescence. A Ploem-type epi-illuminator pro-vided narrow band excitation light and selectivefiltration of the contrasting emission colours. Thetissue sections were mounted in polyvinyl alcohol(PVA) containing PPD (0.2-2-0 g/l).'8 The sectionswere remounted in PVA alone when storage wasrequired after the first counting (which took placewithin three days after the initial mounting).

All cell enumerations were carried out by thesame observer throughout the study and based onthe use of a x40 Leitz immersion objective. Defini-tive cell profiles showing bright cytoplasmic stainingwere counted and also more faintly fluorescent cellswith purely green or red colour and a distinctnucleus.

IMMUNOCYTE ENUMERATION ANDHISTOMORPHOMETRYCell enumerations were based on a lamina propriasection area at full height (from the muscularismucosae to the surface epithelium), limited by adistinct crypt at each side and including the domes oflymphoid follicles. A grid (Leitz code No. 519902)representing 0.032 mm2 was moved systematicallythroughout this area which included on an average26 grids. A sufficiently large area was alwaysincluded to ascertain a stabilised mean number ofIgA cells/grid (Fig. 1). The mean was usuallystabilised at about 20 grids (95% confidence inter-val, 16-25 grids).

After completion of the cell enumerations, thecover slip was removed and the section was stainedwith haematoxylin and eosin (H & E). A penciltracing of the actual lamina propria area was thenmade with a Leitz Orthoplan microscope equippedwith a Leitz drawing device and a x40 objective.The included area was determined by means of aKontron MOP-AMO 3 instrument (Munich, GFR).

DATA TREATMENT AND STATISTICAL METHODSThe density of IgA and IgG immunocytes wasdetermined in two sections with reversed fluoresc-ence colours (Table 2) and the mean values wereused for these isotypes (When the counting wasrepeated, each mean was based on four determina-tions). The density of IgM immunocytes, however,was calculated by extrapolation from the IgA/IgM-cell ratio determined in the third section (Table 2).The statistical analyses were performed as

two-tailed tests.21 Differences were consideredsignificant when p<0.05. For estimation of locationparameters, the median with 95% confidenceinterval (95% CI) was used. Confidence interval wascalculated using the Bernoulli-Wilcoxon procedure.22For evaluation of differences between the first andthe second series of cell enumerations, the Wilcoxon'ssigned midrank test was used.22 Comparisonsbetween groups were based on the Wilcoxon'smidrank sum test.22 For testing differences indispersion between the first and second enumeration,

5 10 15 20 25Total number of grids

Fig. 1 Influence ofthe total number ofgrids evaluated on the estimated mean number ofIgA-producing cells in threespecimens from appendix and threefrom colon mucosa. The means are stabilised at about 20 grids.

669



j.com/


ownloaded from

http://gut.bmj.com/

670

the Sandvik's test with correction for ties wasused.23

REPRODUCIBILITYEnumerations were repeated blindly in the samesections after four to six weeks for the first fivesamples from both appendix and colon. The medianvalues for IgA cells were virtually identical whereasthose for IgG cells tended to be slightly althoughstatistically significantly (p<002) lower and lessdispersed in the second enumeration (Table 3).

Results

OVERALL LAMINA PROPRIAThe median density of IgA-producing immunocytesin the lamina propria between the follicles of theappendix (1259 cells/mm2; 95% CI, 833-1545)tended to be higher than in the colon (962 cells/mm2; 95% CI, 695-1435), but there was no signifi-cant difference (Fig. 2). The same held true for IgMimmunoc2ytes (71 cells/mm2; 95% CI, 24-128 vs 55cells/mm; 95% CI, 33-88). IgG immunocytes, onthe other hand, showed in the appendix a density(95 cells/mm2; 95% CI, 61-251) that was more thantwice as high as (p<002) in the colon (38 cells/mm2;95% CI, 11-61).The numbers of IgM immunocytes in the nine

paired samples of appendix and colon were signifi-cantly correlated (r=071; p<002) whereas no suchrelationship was revealed for IgA (r=0.28) or IgGimmunocytes (r=-018).IgD immunocytes were virtually absent in both

mucosal tissues. In one patient there was a densityof eight IgD-producing cells/mm2 in the appendixand 24 such cells/mm2 in the colon, whereas in theremaining samples we found only one cell.When ignoring IgD immunocytes, the percen-

tages of IgA-, IgM- and IgG-producing cells in theappendix were 87 3%, 5.6% and 7.0%, respectively(Fig. 3). In the colon the same figures were 902%,

Table 3 Reproducibility of immunocvte enumeration(cellslmm2) in tissue sections from the appendix (n=5)and colon (n=5)

Immunocyte First Secondisotype enumeration enumeration

IgA Median: 1258 124995% confidenceinterval: 743-1636 724-1644Observed range: 658-1734 670-1731

IgG Median: 44 3795% confidenceinterval: 9-247 8-240Observed range: 0-390 0-365

2200O

U

1800-

I1~O

0@ 60

0-

-I

oC 3000

100

Bjerke, Brandtzaeg, and Rognum

IgA

* 00

..... l...0.*....oo....

-; 00*0

.... oO

. "88"'

AppendixGoon o

IgM IgG

: 00~~~~

00* 0.... s @o..

1W,§@@@~~~~~~~~~~~~....-Fig. 2. Density ofIg-producing cells in appendix andcolon mucosa. Medians (horizontal solid lines) and95% confidence interval (horizontal broke lines) areindicated.

IgAAppendix

90 Colon

170

F Ig01u Igm~~~~~~~gC o i

Fig. 3. Median isotype percentages ofIg-producing cellsin appendix and colon mucosa. Vertical bars indicate the95% confidence intervals.



j.com/


ownloaded from

http://gut.bmj.com/

Ig production in appendix and colon mulcosa

466%, and 30)%, the proportion of IgG cells beingsignificantly lower than in the appendix (p<0 (5).

FOL LICLE ASSOCIATED l.AMINA PROPRIAIn the 0( 1 mm zone adjacent to lymphoid follicles inthe appendix mucosa including the domes, there wasan increased proportion of IgG-producing immuno-cytes compared with the more distant lamina propria(Fig. 4); almost 50(% of the immunocytes in thiszone were IgG-positive (Table 4). IgG immunocyteslikewise tended to be enriched adjacent to lymphoidfollicles in the colon mucosa but it contained too few

'Table 4 Isotypv ratio (IgAlIgG) of Ig-producingimmunocyles in afppendix muncosa

Zon,e (()1/ ,tit,)Overall l(aoital( ulfjacent toproprii l!inphlloild oflcles0i /_1 ) (0I = I /)1

Mclediin ratio: 10)3 1.195', confidcncc intcrval: 53 -17 4 0) -1l6Ohscivcd raingc: 28-31 3 (). S--

:In onc sanmplc it was tcchnicailly imilpossiblc to evaluate this zone.

Fi,- 4. ITo adaceth .sec(Ttiolls *f p"re-'ashed tanldCt/i(lOI.Ol-fledoor1oilOlf)/)L'7bidiA Ii(RSa. (ai) 1 (k 1 stailiedseCti' li for orientation. 1, = Imn/phoid /folicle. 1)-dome.(b. c) (c omparabl eful boot sect/oh sii/tjected to pairedstilins.jfor IlA (/h. grecli fluorescentce) tandlgG(c, reel fioore.c('ltce). Neo preferential accuonidationi of'IG(;-/prOdlhh1n cells inI tiw (doowt01(1 (dolnos ile t/i lo'llicle,v/here('as IgA -producing ce('s are abltindati also betwveenthe crv)t.s on i/ic rig/it. Botli crspt andi sorft 'epithellion s/ovi sec1tcive uptakeof(' IgA in columnnarcel/s, *signifVi1 (1-t1er11al transport of this is0otApe.(Magmiification.' x 90).

such structures for performance of a systematicstudy.

Discussion

METHODOLOGICAI CONSIDERATIONSThis study was partly undertaken to establish amethod for direct determination of mucosal im-munocyte densities. Like most previous studies(Table 1), our work was based on fluorescencemicroscopy. Valnes et at4 have shown that there areonly minor discrepancies between immunofluoresc-ence and immunoperoxidase staining as to theenumeration of Ig-producing cells. The unaccept-ably large variations appearing in Table 1 cannot,therefore, be ascribed to differences in stainingmethods per se. We have found that the mostreliable tissue preparation method to reveal allisotypes of Ig-producing cells, is the prewashing-and ethanol fixation procedure> used in the presentand all of our previous studies. In addition, we havebased our immunocyte counts on paired immuno-fluorescence stainin that affords internal control offalse-positive cells' "20 and makes it possible toexamine simultaneously in the same section the

671



j.com/


ownloaded from

http://gut.bmj.com/

672

inter-relationship between different classes of im-munocytes.

In our laboratory previous studies of intestinalimmunocyte populations have been related to a'tissue unit' that constitutes a 6 tim thick and 500 gmwide mucosal block at full height from the mus-

cularis mucosae to the lumen. 14 The included area oflamina propria will therefore vary among differentspecimens; the total immunocyte number per 'tissueunit' depends both on this variable and on the celldensity. An average measure of the total localimmunocyte response is thereby obtained, and inour opinion this is necessary for well foundedevaluation of disease-associated changes. Compari-sons based solely on cell density may provide falseinformation to this end, a fact also pointed out bySkinner and Whitehead.26

'Tissue unit' recording was difficult to apply onthe appendix mucosa, however, because of itscontent of numerous lymphoid follicles. We there-fore developed a method to determine the immuno-cyte density of the lamina propria in a representativemanner, including all layers of the mucosa. The dataobtained were remarkably similar to those reportedpreviously from our laboratory for normal largebowel mucosa, both with regard to cell densities andisotype proportions (Table 1).5 1'The direct counting method is less time consum-

ing than photographic recording of fluorescent cells.Nevertheless, it is a subjective choice at whatbrightness a cell should be considered an Ig-producing immunocyte. This difficulty, in additionto other obvious variables, must be taken intoaccount when comparing the results of variousauthors. The largest variations appearing in Table 1,however, remain unexplained. It is obvious thatworkers in this field should pay more attentionto the methodological problems inherent in im-munohistochemical enumeration of Ig-producingcells. 5 2(0 25

BIOLOGICAL CONSIDERATIONSStudies of the rabbit appendix have suggested that itplays an immunological role in the mucosal defencesystem. Lymphoid follicles are absent at birth27 28and fail to develop after early ligation of theappendix.29 Elimination of the flora by appendicos-tomy and sterilisation of the lumen causes rapiddegeneration of the follicles.30

It seems justified to assume that the lymphoidfollicles of the appendix are analogous to the Peyer'spatches in having the capacity to generate IgA-cellprecursors that migrate via lymph and blood to thedistant gastrointestinal lamina propria.3' 32 There isevidence to suggest that gastrointestinal immuno-cytes of other isotypes than IgA likewise originate in

Bjerke, Brandtzaeg, and Rognum

lymphoepithelial structures.32 The preferential accu-mulation of IgG-producing cells adjacent to thelymphoid follicles, as shown in this study, indicatesthat most of these immunocytes represent clonesthat tend to undergo terminal maturation locallyrather than after emigration. Parrot3l has suggestedthat some migration of blasts takes place directlyfrom Peyer's patches to nearby crypt regions; theseblasts may mainly end up as IgG immunocytes in thefollicle associated mucosal zone.

Local production of IgG has been demonstratedin Peyer's patches of rats and mice,3136 and we haverecently shown that there is a remarkably highproportion (38%) of IgG immunocytes among theIg-producing cells found in the dome area of humanPeyer's patches.37 It is worth noting that these IgGimmunocytes show much poorer J-chain expressionthan their counterparts in distant lamina propria(37% vs 82%), and a similar trend is seen for theIgA-producing cells present at these two sites (85%vs 97% J-chain positivity). Relatively decreasedexpression of J chain is likewise a characteristic ofthe IgG-producing cells that accumulate preferen-tially adjacent to solitary follicles elsewhere in thegut and particularly in the appendix.37 Accumula-tion of J chain-negative IgG- and IgA-producing cellis still more striking in the tonsils, however38

Altogether, it seems that stimulated B cells withdown-regulated J chain-expressing potential tend toremain locally. They apparently belong to maturememory clones and therefore express mainly IgG,whereas most members of relatively early memoryclones probably emigrate from the lymphoepithelialstructures after stimulation and differentiate to Jchain-positive immunocytes at distant mucosalsites.3PWe found that the proportion of IgM immuno-

cytes was similar in colon and appendix mucosa.Soltoft et a13 reported 6*5% IgD immunocytes in therectum, but like Crabbe and Heremans' and Brownet a16 we found almost no IgD-producing cells in theintestinal lamina propria. By contrast, salivary,nasal, and lacrimal glands contain a significantproportion of such cells, both normally and es-pecially in selective IgA deficiency.39 We haveproposed that the tonsils and other lymphoepithelialstructures of the upper aeroalimentary tract maysupply glandular sites of this region with immuno-cyte precursors.38 39 There is no evidence to suggestwhether the appendix preferentially supply a certaingut segment with such precursors, but experimentsin rats have indicated that lodging of IgA immuno-cytes takes place mainly within the segment wherethey have been generated-that is, in the duodenumor in the colon.4"We have found that normal human appendix



j.com/


ownloaded from

http://gut.bmj.com/

Ig production in appendix and colon mucosa 673

mucosa contains relatively more IgG-producing cellsthan the colonic counterpart. This difference can beascribed to preferential accumulation of IgG im-munocytes adjacent to the numerous lymphoidfollicles, in the appendix. We propose that thisaccumulation reflects local terminal maturation of Bcells belonging to relatively mature memory clones.Generation of such clones in lymphoepithelial struc-tures may depend on the magnitude of topicalimmunological stimulation and apparently increasesin the order Peyer's patches, appendix, and tonsils.The potency of a lymphoid tissue to function asprecursor source of relatively immature B-cellclones that can emigrate and seed secretory sitesmay hence be reversely related to the local IgGresponse.

Supported by the Norwegian Cancer Society andAnders Jahre's Fund. We thank Dr Sci Stig Larsen,Medstat, Center for Administration, Design andStatistical Analysis in Medical Research, Str0mmen,Norway, for carrying out the statistical analyses.

References1 Crabbe PA, Heremans JP. The distribution ofimmunoglobulin-containing cells along the human gas-trointestinal tract. Gastroenterology 1966; 51: 305-16.

2 Gelzayd EA, Kraft SC, Kirsner JB. Distribution ofimmunoglobulins in human rectal mucosa. I. Normalcontrol subjects. Gastroenterology 1968; 54: 334-40.

3 Soltoft J, Binder V, Gudmand-Hoyer E. Intestinalimmunoglobulins in ulcerative colitis. Scand J Gas-troenterol 1973; 8: 293-300.

4 Skinner JM, Whitehead R. The plasma cells in in-flammatory disease of the colon: a quantitative study. JClin Pathol 1974; 27: 643-6.

5 Baklien K, Brandtzaeg P. Comparative mapping of thelocal distribution of immunoglobulin-containing cells inulcerative colitis and Crohn's disease of the colon. ClinExp Immunol 1975; 22: 197-209.

6 Brown WR, Borthistle BK, Chen ST. ImmunoglobulinE (IgE) and IgE-containing cells in human gastrointes-tinal fluids and tissues. Clin Exp Immunol 1975; 20:227-37.

7 Green FHY, Fox H. The distribution of mucosalantibodies in the bowel of patients with Crohn'sdisease. Gut 1975; 16: 125-31.

8 Meuwissen SGM, Feltkamp-Vroom TM, Brutel de laRiviere A, von dem Borne AEGK, Tytgat GN-Analysis of the lympho-plasmacytic infiltrate inCrohn's disease with special reference to identificationof lymphocyte-subpopulations. Gut 1976; 17: 770-80.

9 O'Donoghue DP, Kumar P. Rectal IgE cells ininflammatory bowel disease. Gut 1979; 20: 149-53.

10 Rognum TO, Brandtzaeg B, Baklien B, Hognestad I.Immunoglobulin-producing cells in the "transitional"mucosa adjacent to adenocarcinomas of the humanlarge bowel. Int J Cancer 1979; 23: 165-73.

11 Rosekrans PCM, Meijer CJLM, van der Wal AM,Cornelisse CJ, Lindeman J. Immunoglobulin contain-ing cells in inflammatory bowel disease of the colon: amorphometric and immunohistochemical study. Gut1980; 21: 941-7.

12 Otto HF, Gebbers JO. Electron microscopic, ultracy-tochemical and immunohistochemical observations inCrohn's disease of the ileum and colon. Virchows Arch[Pathol Anat] 1981; 391: 189-205.

13 Regadera J, Paniagua R, Nistal M, Santamaria L.Quantitative distribution of Ig-containing cells in themucosa of the human large intestine. Cell Mol Biol1983; 29: 387-95.

14 Brandtzaeg P, Baklien K. Immunohistochemical stu-dies of the formation and epithelial transport ofimmunoglobulins in normal and diseased human intes-tinal mucosa. Scand J Gastroenterol 1976; 11: suppl 36:1-45.

15 Brandtzaeg P. Tissue preparation methods, In: BullockGR, Petrusz P, eds. Techniques in immunocytochemis-try. London: Academic Press, 1982; 1: 1-75.

16 Indiveri F, Saba V, Barabino A, Bianchi G. Studioimmunoisto-fluoroscopico del sistema immunocom-petente dell'appendice cecale umana. Arch E Marag-liano Patol Clin 1971; 27: 205-12.

17 Chen ST. Cellular sites of immunoglobulins. II. Therelative proportions of mucosal cells containing IgG,IgA, and IgM, and light polypeptide chains of kappaand lambda immunoglobulin in human appendices.Acta Pathol Jpn 1971; 21: 67-83.

18 Valnes K, Brandtzaeg P. Retardation of immuno-fluorescence fading during microscopy. J HistochemCytochem 1985; 33: 755-61.

19 Brandtzaeg P. Mucosal and glandular distribution ofimmunoglobulin components. Immunohistochemistrywith a cold-ethanol fixation technique. Immunology1974; 26: 1101-14.

20 Brandtzaeg P. Prolonged incubation time in immuno-histochemistry: effects on fluorescence staining ofimmunoglobulins and epithelial components inethanol- and formaldehyde-fixed paraffin-embeddedtissues. J Histochem Cytochem 1981; 29: 1302-15.

21 Sverdrup E. Significance testing in multiple statisticalinference. Scand J Statist 1976; 3: 73-6.

22 Lehmann EB, D'Abrera HJM. Nonparametric: Statis-tical methods based on ranks. San Francisco: HeolenDay Inc, 1975.

23 Sandvik L, Olson B. A nearly distribution-free test forcomparing dispersion in paired samples. Biometrica1982; 69: 482-5.

24 Valnes K, Brandtzaeg P, Hanssen LE, Stave R, LarsenS, Londong W. Quantitation of immunoglobulin- andpeptide hormone-producing cell in gastrointestinalmucosa. Comparison of direct immunofluorescenceand the unlabelled antibody peroxidase-antiperoxidasemethod. Histochem J 1983; 15: 1011-20.

25 Brandtzaeg P, Rognum TO. Evaluation of nine diffe-rent fixatives. 1. Preservation of immunoglobulinisotypes, J chain, and secretory in human tissues.Pathol Res Pract 1984; 179: 250-66.

26 Skinner JM, Whitehead R. Immunological aspects ofgastrointestinal pathology. In: Morson BC, ed. Pathol-



j.com/


ownloaded from

http://gut.bmj.com/

674 Bjerke, Brandtzaeg, and Rognum

ogy of the gastro-intestinal tract. Current Topics inPathology. Vol. 63. Springer: Berlin, 1976.

27 Fridenstein A, Goncharenko I. Morphological evi-dence of immunological relationships in the lymphoidtissue of rabbit appendix. Nature 1965; 206: 1113-15.

28 Stramignoni A, Mollo F. Development of the lymphoidtissue in the rabbit's appendix. A light and electronmicroscopic study. Acta Anat 1968; 70: 202-18.

29 Stramignoni A, Mollo F, Rua S, Palestro G. Develop-ment of the lymphoid tissue in the rabbit appendixisolated from the intestinal tract. J Pathol 1969; 9:265-9.

30 Blythman HE, Waksman BH. Effect of irradiation andappendicostomy on appendix structure and responsesof appendix cells to mitogens. J Immunol 1973; 111:171-82.

31 Parrott DMV. The gut-associated lymphoid tissues andgastrointestinal immunity. In: Ferguson A, Mac SweenRNM, eds. Immunological aspects of the liver andgastrointestinal tract. Lancaster: MTP, 1976: 1-32.

32 Brandtzaeg P. Research in gastrointestinal immu-nology: state of the art. Scand J Gastroenterol 1985;20: suppl 114: 137-56.

33 Crabbe PA, Nash DR, Bazin H, Eyssen H, HeremansJF. Immunohistochemical observations on lymphoidtissues from conventional and germ-free mice. LabInvest 1970; 22: 448-57.

34 Sminia T, Janse EM, Plesch BEC. Ontogeny of Peyer'spatches of the rat. Anat Rec 1983; 207: 309-16.

35 Sminia T, Plesch BEC. An immunohistochemical studyof cells with surface and cytoplasmic immunoglobulinsin situ in Peyer's patches and lamina propria of ratsmall intestine. Virchows Arch [Cell Pathol] 1982; 40:181-9.

36 McClelland DBL. Peyer's patch-associated synthesis ofimmunoglobulin in germ-free, specific-pathogen-free,and conventional mice. Scand J Immunol 1976; 5:909-15.

37 Bjerke K, Brandtzaeg P. Immunoglobulin- and Jchain-producing cells associated with the lymphoidfollicles in human appendix, colon and ileum, includingPeyer's patches. Clin Exp Immunol 1986 (In press).

38 Korsrud FR, Brandtzaeg P. Immunohistochemicalevaluation of J-chain expressions by intra- and extra-follicular immunoglobulin-producing human tonsillarcells. Scand J Immunol 1981; 13: 271-80.

39 Brandtzaeg P, Gjeruldsen ST, Korsrud F, Baklien K,Berdal P, Ek J. The human secretory immune systemshows striking heterogeneity with regard to involve-ment of J chain-positive IgD immunocytes. J Immunol1979; 122: 503-10.

40 Pierce NF, Cray WC. Determinants of the localization,magnitude, and duration of a specific mucosal IgAplasma cell response in enterically immunized rats. JImmunol 1982; 128: 1311-15.



j.com/


ownloaded from

http://gut.bmj.com/

mucosa - gutlymphoid follicles. agrid (leitz code no. 519902) representing 0.032 mm2 was moved...

Documents