882 Gut, 1992, 33, 882-886 Immunosenescence and mucosal immunity: significant effects of on secretory IgA concentrations and intraepithelial Gut: first published as 10.1136/gut.33.7.882 on 1 July 1992. Downloaded from counts

E Arranz, S O'Mahony, J R Barton, A Ferguson

Abstract gation of gastrointestinal symptoms, and in Concentrations of immunoglobulins (Ig) and whom the final diagnosis was of functional levels of isotype specific to three gastrointestinal symptoms or of a minor clinical dietary in serum, pure parotid saliva, problem without evidence of immunological, and in intestinal secretions obtained by whole infectious, neoplastic, or allergic disease. Infor- gut lavage from groups of healthy elderly mation was recorded on factors possibly relevant subjects (aged >70 years) and ofyounger adult to mucosal immunity, including smoking and controls (aged 25-50 years) were measured. In alcohol consumption, dietary habits, dentition, addition, counts of lamina propria and intra- and medication. epithelial lymphoid cells were performed in histologically normal jejunal biopsy specimens Series A from elderly and younger subjects. Elderly For studies on serum and saliva, elderly volun- subjects had significantly higher concentra- teers were recruited from old people attending tions of serum and salivary IgA and of salivary the geriatric day centre at the Royal Victoria IgM (both, p<0.01), and of salivary IgA anti- Hospital and younger subjects from patients of bodies than did the younger subjects, but the the Gastro-Intestinal Unit at the Western amount of immunoglobulin and in General Hospital, Edinburgh. whole gut lavage fluid was similar in the two age groups. Jejunal biopsy specimen cell Series B counts showed higher IgA counts Whole gut lavage fluid was obtained from http://gut.bmj.com/ and lower intraepithelial lymphocyte counts in patients of the Gastro-Intestinal Unit who were the elderly group (p<0.01), with similar counts having gut lavage with a polyethylene-glycol- of IgM and IgG plasma cells, eosinophils, and electrolyte lavage solution, Golytely, as pre- mast cells in the two groups. There is evidence paration for barium enema or colonoscopy. If of significant effects of old age on the mucosal after clinical assessment the final diagnosis was . as described above, the patients were included in the study. on October 2, 2021 by guest. Protected copyright. There have been many research studies ofageing Series C and immunity, but they have concentrated on Jejunal biopsy specimens were selected from a systemic immunity in man or have been con- file of stored slides, if (i) the patient fell within ducted in aged rodents.' 3 There have been no the age ranges defined, (ii) histology had been thorough studies in humans of the influence of reported to be normal by a consultant path- on secretory immune function or on ologist, (iii) assays of brush border disacchari- intestinal cellular immunity. We report here the dases were normal, and (iv) if review of the case results of measurements of immunoglobulins notes showed no evidence, on final clinical and ofisotype specific antibodies to three dietary appraisal, of immunological or other significant antigens in serum, pure parotid saliva, and in disease as specified above. intestinal secretions obtained by whole gut lavage from a group of healthy elderly subjects (aged >70 years) and from younger adult SPECIMEN COLLECTION controls (aged 25-50 years). In addition, counts of lamina propria and intraepithelial lymphoid Pure parotid saliva Gastro-Intestinal Unit, University of Edinburgh cells were performed in histologically normal A timed five minute collection of parotid saliva and Western General jejunal biopsy specimens from elderly and was obtained after a fast of at least three hours, Hospital, Edinburgh younger subjects. via a Carlsson-Crittenden cup placed over the E Arranz duct orifice.4 S O'Mahony parotid J R Barton A Ferguson Methods Intestinalfluid Correspondence to: Whole gut lavage fluid (WGLF) was obtained by Professor A Ferguson, Department of Medicine, PATIENTS AND VOLUNTEERS giving the patients isotonic non-absorbable University of Edinburgh, 'Elderly' subjects were aged 70 years or more; polyethylene-glycol-electrolyte lavage solution Western General Hospital, Edinburgh EH4 2XU 'younger' subjects were aged 20-50 years. Some (Golytely) to a maximum of4 litres orally after an Accepted for publication were healthy volunteers. Others were patients overnight fast.5 Specimens were collected for the 18 November 1991 who had had full clinical assessment and investi- study when they became clear, without faecal Immunosenescence andmucosal immunity: significant effects ofoldage on secretory IgA concentrations andintraepithelial lymphocyte counts 883

contamination. The specimens were immedi- IgM at 1:500, 1:250, and 1:200 dilutions, respec- ately centrifuged and treated with protease tively); secondary antibody (affinity purified inhibitors. donkey anti-sheep/goat IgG at 1:50 dilution); Serum, saliva, and gut lavage specimens were and tertiary antibody (sheep peroxidase/anti-

stored in aliquots at -70C. peroxidase at 1:80 dilution). All incubations Gut: first published as 10.1136/gut.33.7.882 on 1 July 1992. Downloaded from were carried out in a humid chamber at room temperature for 60 minutes. Antisera were ELISA TECHNIQUES obtained from the Scottish Antibody Production All samples were assayed for total IgA, IgG, and Unit, Carluke, Scotland; positive and negative IgM and for specific antibodies of these isotypes specificity controls were included in each batch; to three representative food protein antigens (iii) Toluidine blue and carbol chromatrope gliadin, ovalbumin, and lactoglobulin, by for mast cells and eosinophils respectively. enzyme linked immunosorbent assays (ELISAs) Blinded cell counts, on stained sections which as previously described.6 Reference standards had been coded and mixed, were performed with for Ig concentrations were human colostral IgA a Leitz microscope, using appropriate eyepieces (Sigma Chemical Co, Poole, Dorset) and a and graticules, calibrated with a stage micro- human reference serum for IgG and IgM meter. Villus and crypt length measurements (Protein Reference Unit, Royal Hallamshire were performed with a 1 cm linear eyepiece Hospital, Sheffield). Serum from an untreated graticule and a x 10 and x40 objective lens coeliac disease patient was used as a reference respectively. standard for antibodies. Plates were read in Intraepithelial lymphocyte counts were per- an MR580 microELISA reader (Dynatech, formed by a method previously described7 with a Billingshurst, Sussex, UK), set at a wave length 1 cm square graticule and a x 100 oil immersion of405 nM (OD405), when the standard reached an lens, expressing the results in numbers of intra- arbitrarily selected OD405 of 1-0, and antibody epithelial per 100 villus enterocytes. levels in test specimens were expressed as a Counts of plasma cells, mast cells, and percentage ofthe optical density ofthe standard. eosinophils were performed with the same The following method was used for measure- graticule and eyepiece as for intraepithelial ment of secretory IgA in WGLF: ELISA plates lymphocyte counts. Only well orientated were coated as for total IgA (anti-a chain), and sections were used; fields were examined standard colostrum IgA and samples were added systematically, starting with the base of the in duplicate for overnight incubation; two graticule at the muscularis mucosae, counting different conjugates were used including anti- sequential fields of lamina propria vertically to human IgA (for total IgA, dilution 1:5000) and the luminal surface, then realigning the graticule

anti-human secretory component/alkaline phos- on the immediately adjacent part of the http://gut.bmj.com/ phatase conjugate (dilution 1:2000) (The muscularis mucosae and continuing the process. Binding Site, Birmingham B15 2SQ, UK). The Epithelium was excluded, and where only part of optical density values obtained for total IgA and a field comprised lamina propria, the proportion secretory component were compared for each was estimated by eye. At least 40 lamina propria gut lavage sample. graticule fields per section were counted. The grid area with a x 100 objective was 0-0132 mm2.

The results were expressed as numbers of on October 2, 2021 by guest. Protected copyright. JEJUNAL BIOPSY SPECIMENS: positive cells/mm2 oflamina propria tissue. STAINS AND COUNTING TECHNIQUES Biopsy specimens from both groups were formalin fixed, embedded in paraffin wax, and STATISTICAL METHODS stained with: Differences in immunoglobulin concentrations, (i) Haematoxylin and eosin, in order to study antibody levels, and cell counts were assessed the villus/crypt morphology and intraepithelial using the Mann-Whitney U test (two-tailed) and lymphocyte counts; correlations with Spearman's rank correlation (ii) Immunoperoxidase technique to study coefficient. A p value <0 05 was considered immunoglobulin containing cells. Briefly, significant. sections were pretreated with trypsin 0-1% in Tris buffered saline pH 7-6 at 37°C for 20 minutes, and then serially incubated with Results primary antibody (sheep anti-human IgG, IgA, DETAILS OF SUBJECTS Demographic data for the patients in whom was assessed are summarised TABLE I Demographic data ofelderly andyounger subjects - series A, B, and C in Table I. Sex distribution was similar in the Series A Series B Series C elderly and young groups. Denture use was Serum/saliva Gut Lavage Jejunal specimens significantly more common in the elderly, and, Elderly Young Elderly Young Elderly Young in the subjects whose serum and saliva were No 43 37 16 14 27 25 studied, alcohol use was less frequent than in Male/female 14:29 14:23 6:10 4:10 6:21 7:18 young subjects. Mean age 79.3 36-2 79-2 32-2 75 5 36-3 (range) (70-94) (26-50) (71-92) (23-47) (70-87) (27-50) Smokers 14 10 2 6 7 9 Alcohol 15 23* 4 8 17 20 AND ANTIBODIES 1** 15/15 4/25 SERUM IMMUNOGLOBULINS Dentures 41 26** 16 In elderly subjects, the serum IgA concentration *p

in the young (Table II). Serum concentrations of TABLE IV Concentration oftotal IgA and secretory IgA IgM class antibodies to three dietary proteins (median (range)) in whole gut lavagefluidfrom elderly and were also significantly lower in the elderly than in young subjects (Series B) the young (Table III). Elderly Young Gut: first published as 10.1136/gut.33.7.882 on 1 July 1992. Downloaded from No 16 14 Total IgA (jsg/ml) 189-7 186-5 IMMUNOGLOBULINS AND ANTIBODIES IN PAROTID (470-7-11-9) (531-2-13) Secretory IgA (Isg/ml) 167-6 111-4 SALIVA (439-6-14-9) (508-23-4) In pure parotid saliva, IgA concentrations were SIgA, % of total 92 80 appreciably higher in the elderly than in the (100-74-1) (100-51) young (Table II). There was a significant positive correlation between the salivary IgA concen- tration and age in the elderly group (r=0-309, when elderly and young subjects were com- p<005). Concentrations of IgM and IgG were pared. The assay for SIgA concentration in also significantly higher in the elderly. Concen- WGLF was developed only recently, and trations of IgA class antibody to gliadin, sufficient material for analysis remained from 12 ovalbumin, and ,B lactoglobulin in saliva were elderly and 12 younger subjects. Values were high in elderly subjects, and trace amounts of similar for total IgA, SIgA, and the percentage IgG class antibody to two of the three antigens of IgA as secretory (mean 92%, range 100-74 1 were detected. in the elderly group; and mean 80%, range 100-51 in younger controls) (Table IV).

WHOLE GUT LAVAGE FLUID Studies of intestinal fluid, however, showed no CELL COUNTS differences in total immunoglobulin concen- The biopsy specimens from elderly and young trations (Table II) or specific antibody values patients all looked histologically normal; villus and crypt lengths were similar in the two These and counts of TABLE II Immunoglobulin concentrations (median (range)) in serum and secretions ofelderly groups. measurements, andyoung subjects (SeriesA, B) intraepithelial lymphocytes and lamina propria lymphoid cells, are summarised in Table V. Elderly Young In the elderly patients, counts of intraepithelial No Median (range) No Median (range) p Value lymphocytes (expressed per 100 villus entero- cytes) were significantly lower (p<004) and Serum (series A): 43 2737-8 37 2062 of lamina propria IgA containing plasma cells IgA ([tg/ml) p<0-001 http://gut.bmj.com/ (9413-1031) (3682-845) significantly higher (p<001) than in the younger IgM([g/ml) 43 922-2 37 1357 p<005 (5498-270) (3445-322) controls. Counts of other lamina propria cells IgG ([tg/ml) 43 11223 37 10527 NS (IgG plasma cells, IgM plasma cells, mast cells, (25 490-6090) (15 921-7743) Parotid saliva (series A): eosinophils) were similar in the two groups. IgA ([tg/ml) 43 205.4 37 113-4 p<0-001 (3229-30.7) (553-216) IgM([tg/ml) 43 1-6 37 1.1 p<005 (42.6-0) (5.5-0) Discussion IgG ([tg/ml) 43 1-7 37 0.5 p<0-001 on October 2, 2021 by guest. Protected copyright. (45.8-0) (34.2-0.1) Traditional methods of immunological investi- Whole gut lavage fluid (series B): gation applied to systemic immunity have clearly IgA(Ig/ml) 16 69-9 14 69-1 NS shown a phenomenon of immunosenescence.'-3 (226-8) (274-10-6) IgM([kg/ml) 16 3-6 14 5-8 NS The is atrophic by the age of 60 years (20-1-1-4) (35.8-0) and peripheral T lymphocytes have impairment IgG(,ug/ml) 16 0-1 14 0-7 NS (9.8-0) (2-1-0) of proliferative capacity, a reduced secretion of in response to various signals, and a TABLE III Levels of antibodies to three dietary proteins in serum and secretions (median reduced rate of receptor expression when (range)) ofelderly andyoung subjects (Series A, B) appropriately stimulated.3 8-11 B cells are intrinsically normal in old age but Elderly Young function in an aberrant way. This is reflected by No IgA IgM IgG No IgA IgM IgG high concentrations of immunoglobulins in the blood, high titres of various autoantibodies,'2 '3 Serum antibodies (series A): Gliadin 43 8-2 39.5 17 37 4.5 88.3* 15-5 and yet an impaired generation of specific anti- (0-150) (15-155) (0-111) (0-60-6) (14-222) (26-63) body in response to encounter.'2 14 Ovalbumin 43 5 14-6 23-4 37 3-2 21.6** 30-5 (1*3-81-6) (2-3-76) (0-98.5) (1-150) (5-49) (2-1-37) Several factors probably contribute to this ,BLactoglobulin 43 9-3 20-3 36-3 37 9.2 29.2** 29-8 abnormality of B cells: deficient help and (2-150) (4-50) (3-108-4) (0-9-37) (8-90-8) (3-7-101-6) Parotid saliva antibodies (series A): also deficient T suppression of inappropriate Gliadin 43 11.5* 4-5 0 37 3-8 4-6 0 function3; the fact that after a lifetime of (1-6-81) (0-66) (0-5-3) (0-304) (0-24-5) (0-2-3) exposure, Ovalbumin 43 15.3** 0-1 0 37 9-1 1 0 idiotype/anti-idiotype immuno- (3-4-116-7) (0-21-3) (0-6-1) (0-73.5) (0-15-5) (0-7-9) suppressive loops have formed'5; or that most 6 Lactoglobulin 43 31-8** 1-5 0.3* 37 20.5 1 0 B (5-150) (0-53) (0-11-3) (3-101-6) (0-255) (0-0 6) cells are precommitted leaving little capacity Whole gut lavage fluid antibodies (series B): for a brisk response to further antigen encounter. Gliadin 16 6-2 6-3 0 14 3-2 5 3 0-3 There is general agreement that serum IgA (0-85-9) (0.5-62 8) (0-6.5) (0-57-8) (0-36-2) (0-3.3) Ovalbumin 16 3.5 1-2 0 14 3-3 1-6 0.5 concentrations are increased in old age in a range (0-71-5) (0-5.9) (2-2.5) (0.6-32-9) (0-4 7) (0-44) of species including man.'6 7 Previous work on ,BLactoglobulin 16 5.3 1-2 0 14 5 2 0-2 (1-42) (0-7) (0-3-3) (0-76) (0-20-6) (0-3.6) secretory immunity in man in old age has mainly concerned studies with mixed saliva. '7 These *p

TABLE V Measurement ofvilli and crypts, and counts ofintraepithelial lymphocytes and secretions in old age is similar to that of lamina propria lymphoid cells, injejunal biopsy specimensfrom elderly andyoung subjects the young. (Series C) The high concentrations of IgA in serum and Elderly (n=27) Young (n=22) saliva are parallelled by high levels of specific

Parameter Median (range) Median (range) p Value antibody of IgA class to three dietary protein Gut: first published as 10.1136/gut.33.7.882 on 1 July 1992. Downloaded from Villous height (,um) 286-7 313-6 NS antigens, markers of a 'steady state' antibody (347-5-215-3) (490 2-208 3) status of the individuals concerned. Techniques Crypt depth (gm) 94-2 96 5 NS (132-2-63-6) (170-2-63-1) are now available to allow measurement of the Intraepithelial lymphocytes 17-5 23 p<005 (% enterocytes) (316-12) (37-15) kinetics of immune response to a vaccine or IgA-plasma cells 744-8 310-5 p<001 other enterically administered antigen and these (/mm lamina propria) (1133-294) (680-119) IgM-plasma cells 244-8 221-7 NS should now be applied in old people and could be (5426-172) (317-6-47) used to test the hypothesis that intestinal stasis IgG-plasma cells 28-4 22-7 NS (141-7-7-6) (43 5-9 4) and/or changes in colonic flora induce polyclonal Mast cells 255-8 233-6 NS synthesis of IgA via immunomodulatory Gram (335-89) (337-129-5) Decreased Eosinophils 190-3 127-3 NS negative bacterial products.20 (317-22-7) (3024-38 4) efficiency of Kupffer cell function could be relevant.20 Serum IgM antibody levels to the three food antigens studied were lower in the elderly group, taken ofdentition or dental hygiene so the mixed as was the level of IgA antigliadin antibody in saliva is likely to be heavily contaminated with serum. These findings are in agreement with a crevicular fluid. previous report,2' although we found serum IgG In this report we present an essentially antibody levels to be similar in the two age descriptive study, a first attempt to identify any groups. Low levels ofserum IgM antibody could gross aberrations of secretory immune function be clinically relevant, for these have been shown in elderly humans. We have made use of volun- to correlate with a higher frequency of infection teer healthy old people to study serum and pure with capsulate .2223 Old age does not, parotid saliva, and material which has become however, have a global effect on IgM, as there available from old people, in retrospect consid- were similar levels of IgM and IgM antibodies in ered to have essentially normal gastrointestinal the saliva and gut lavage fluid from old and tracts, but who have had either whole gut lavage young subjects. or small bowel biopsy in the course of investi- Our studies of jejunal biopsy specimen path- gation in a busy gastrointestinal unit. ology showed no evidence of mucosal atrophy in Absolute concentrations of immunoglobulins the jejunum of old people; this supports a recent and antibodies in a secreted fluid will be influ- report by Corazza et al.24 We have had similar http://gut.bmj.com/ enced by flow rate'8 as well as by the immuno- counts for most of the lamina propria cell logical function of the tissue concerned. This is types studied in biopsy specimens from old and particularly relevant to the interpretation of our younger people, which adds weight to the findings of generally higher concentrations of significance of the findings that differences immunoglobulins in the saliva of elderly sub- emerge for two cell types. A high count of IgA jects. This could be explained entirely ifsalivary plasma cells has already been discussed. The flow rate is significantly slower in old people than finding ofa low intraepithelial lymphocyte count on October 2, 2021 by guest. Protected copyright. in younger individuals. There are no data on flow (expressed as lymphocytes per 100 villus rates of pure parotid saliva in old age, but in a enterocytes) requires confirmation and further study of mixed saliva, mean flow rates and follow up in prospectively collected biopsy protein concentrations were similar in groups of specimens. It is necessary to use frozen sections individuals aged 26-44 and 65-83 years. 19 for detailed examination of the phenotype of We have found that a general increase in IgA intraepithelial lymphocyte. Further studies production occurs in the gastrointestinal tract as should include counts of CD4 and CD8 positive well as in the serum in old age. The pure parotid cells; identification of the T cell receptor types salivary IgA concentration was appreciably (alpha beta or gamma delta) and the presence of higher in the elderly group, showing a positive activation markers. correlation between IgA and age. IgM and IgG It will be particularly interesting if deficiency concentrations were also increased. Although of a subset of intraepithelial lymphocyte is concentrations of IgA in whole gut lavage fluid found, and if this could be linked to the expan- were similar in old and younger people, IgA sion of IgA locally. This experiment of nature plasma cell counts were significantly higher; this might contribute very useful data to the investi- method is probably more sensitive as an index gation of whether or not intraepithelial lympho- of mucosal IgA status than analysis of fluid cyte function as locally active suppressor cells.25 obtained in the course of standard ward or outpatient lavage bowel preparation. We are now This work has been supported by grants from the Sandoz developing an experimental protocol of steady Foundation for Gerontological Research and from the Scottish state whole gut perfusion which will allow us to Hospitals' Endowment Research Trust. We thank Mrs J Johnston and Mr J Bode for technical assistance, Dr S Walsh for access to measure the hourly rate of secretion of immuno- her patients, and the staff of the GI Investigation suite for their globulins and other substances. The present invaluable support in the collection of specimens. study of whole gut lavage fluid has, however, 1 Makinodan T, Kay MMB. Age influence on the immune enabled us to establish that most of the IgA system. Adv Immunol 1980; 29: 287-330. in intestinal secretions is bound to secretory 2 Wade AW, Green-Johnson J, Szewzuck MR. Functional changes in systemic and mucosal lymphocyte repertoires component - that is, polymeric - and that in with age: an update review. Aging. Immunol Infect Dis this respect the molecular form of IgA in the 1988; 1: 65-97. 886 Arranz, O'Mahony, Barton, Ferguson

3 Thoman ML, Weigle WO. The cellular and subcellular basis during the normal immune response. VIII. Analysis of the ofimmunosenescence. AdvImmunol 1989; 46: 221-61. cellular basis for the increased auto-antiidiotype antibody 4 Barton JR, Riad M, Gaze MN, Maran AGD, Ferguson A. production in aged mice. J Exp Med 1983; 157: 1635-45. Mucosal in smokers, and in patients 16 Buckley CE, Buckley EG, Dorsey FC. Longitudinal changes with epithelial head and neck tumours. Gut 1990; 31: in serum immunoglobulin levels in older humans. Proc Soc 378-82. Exp Biol Med 1974; 33: 2036-9.

5 O'Mahony S, Barton JR, Crichton S, Ferguson A. Gut lavage: 17 Finkelstein MS, Tanner M, Freedman ML. Salivary and Gut: first published as 10.1136/gut.33.7.882 on 1 July 1992. Downloaded from a new approach to the study ofintestinal humoral immunity. serum IgA levels in a geriatric outpatient population. J Clin Gut 1990; 31: 1341-4. Immunol 1984; 4: 85-91. 6 O'Mahony S, Arranz E, Barton JR, Ferguson A. Dissociation 18 Brandtzaeg P. Human secretory immunoglobulins-VII. Con- between systemic and mucosal humoral immune responses centrations of parotid IgA and other secretory proteins in in coeliac disease. Gut 1991; 32: 29-35. relation to the rate of flow and duration of secretory 7 Ferguson A, Murray D. Quantitation of intraepithelial lym- stimulus. Arch OralBiol 1971; 16: 1295-310. phocytes in human jejunum. Gut 1971; 12: 988-94. 19 Ganguly R. Oropharyngeal tract host defenses in aging. In: 8 Peterson WJ. Immunity, age and loss of immunohomeostasis. Part B. Mestecky J, McGhee JR, Bienenstock J, Ogra PL, Gerontology 1984; 3: 259-69. eds. Recent advances in mucosal immunology. New York: 9 Antonaci S, Jirillo E, Bonomo L. Immunoregulation in aging. Plenum Press. 1987: 1409-16. Diagn Clin Immunol 1987; 5: 55-61. 20 Horan MA, Fox RA. Ageing and the immune response - a 10 Brill S, Kukulansky T, Tal E, Abel L, Polgin Y, Dassa C, etal. unifying hypothesis? MechAgeingDev 1984; 26: 165-81. Individual changes in T lymphocyte parameters of old 21 Scott H, Rognum TO, Midvedt T, Brandtzaeg P. Age-related human subjects. MechAgeingDev 1987; 40: 71-9. changes of human serum antibodies to dietary and colonic 11 Nagel JE, Chopra RK, Powers DC, Adler WH. Effect of age bacterial antigens measured by an enzyme-linked immuno- on the human high affinity IL-2 receptor ofPHA-stimulated sorbent assay. Acta Path Microbiol Immunol Scand. [C]. peripheral blood lymphocytes. Clin Exp Immunol 1989; 75: 1985; 93:65-70. 286-91. 22 Phair JP, Hsu CS, Hsu YL. Ageing and infection. In: CIBA 12 Felser JM, Raff MJ. Infectious diseases and aging: immuno- Foundation Symposium, 1988; 134: 143-59. logic perspectives. JAm GeriatrSoc 1983; 31: 802-7. 23 Peterson WJ. Immunity, age and loss of immunohomeostasis. 13 Hiiimans W, RadI J, Bottazzo GF, Doniach D. Autoanti- Gerontology 1984; 3: 259-69. bodies in highly aged humans. Mech Ageing Dev 1984; 26: 24 Corazza GR, Frazzoni M, Gatto MRA, Gasbarrini G. Ageing 83-9. and small-bowel mucosa: a morphometric study. 14 Saltzman RL, Peterson PK. Immunodeficiency of the elderly. Gerontology 1986; 32: 60-5. Rev Infect Dis 1987; 9: 1127-39. 25 Brandtzaeg P, Sollid LM, Thrane PS, Kvale D, Bjerke K, 15 Goidl EA, Choy JW, Gibbons JJ, Weksler ME, Thorbecke GJ, Scott H, et al. Lymphoepithelial interactions in the mucosal Siskind GW. Production of auto-antiidiotypic antibody immune system. Gut 1988; 29: i 116-30. http://gut.bmj.com/ on October 2, 2021 by guest. Protected copyright.