Abnormalities of vasoactive intestinal polypeptide-containing nerves in Crohn's disease

GASTROENTEROLOGY 79:853-860,198O

Abnormalities of Vasoactive Intestinal Polypeptide-Containing Nerves in Crohn’s Disease ANNE E. BISHOP, JULIA M. POLAK, M. G. BRYANT, S. R. BLOOM and S. HAMILTON Departments of Histochemistry and Department of Pathology,

and Medicine, Hammersmith Hospital, Du Cane Road; St. Mark’s Hospital, City Road, London, United Kingdom

The possible involvement of nerves containing vasoactive intestinal polypeptide in Crohn’s disease was investigated by immunocytochemistry and radioimmunoassay of specimens from 17 patienfs with well-dejïned clinical and histoiogic features of the disease. The characteristic pattern of slender fibers, evenly distributed across the gut wall, was seen in specimens taken from controls, which consisted of (a) specimens from uninvolved areas of gut from carcinema resection (n = 37) and (b) jejunoileal specimens obtained during bypass operation for obesity (n = 8) as wel1 as in four of the six specimens from patients with ulcerative colitis. In contrast, this characteristic pattern was lost in all 17 patients with Crohn’s disease, the pattern being replaced by thickened and more intensely immunostained fibers. These changes were consistently found in the mucosa and submucosa, and in 13 of the Crohn’s disease cases, the abnormal pattern was totally transmural, involving both the myenteric and submucous plexus as well as the muscle layers. There was a >ZOO% increase in VIP content, as determined by radioimmunoassay, in Crohn’s disease (294 f 29 pmol/ g wet wt, mean f SEM) in comparison with (a) ulcerative colitis (93 f 5 pmol/g [P < O.OOZ]),and (b) Received November 1, 1979, Accepted May 22, 1980. Address requests for reprints to: Dr. J. M. Polak, M.D., M. R. C. Path., Department of Histochemistry, Royal Postgraduate Medical School, Hammersmith Hospital, Du Cane Road, London W12 OHS, United Kingdom. This work was carried out with the generous support of the Medical Research Council (U.K.), and Janssen Pharmaceutical, Ltd. (U.K.). We thank the following colleagues for providing US with specimens: Dr. B. C. Morson, Dr. S. Tabaqchali, Dr. B Lake, Professor R. B. Welbourn, and Mr. J. Spencer. Dr. Hamilton’s current address is: Department of Pathology, The Johns Hopkins University School of Medicine and Hospital, 800 North Wolfe Street, Baltimore, Maryland 21205. 0 1980 by the American Gastroenterological Association 0011%5085/80/110853-08$02.25

controls consisting of carcinoma resection (108 f 39) and bypassed gut from obese patients (86 f 27 [P < 0.001]). At least part of the previously documented autonomie nerve changes in Crohn’s disease are, thus, due to an increase in vasoactive intestinal polypeptide innervation. Since the original observations of a substantial increase in ganglion cells in Crohn’s disease,’ an involvement of autonomie nerves in this inflammatory bowel disease has been accepted and wel1 documented in text books.2,3 Recently, the concept of a bipartite (cholinergic and adrenergic) autonomie nervous system has been further challenged by the discovery of peptide-containing nerves.4 A number of biologically active peptides, first extracted from the brain or gut, are now known to have a very wide peripheral distribution in the autonomie nervous system.5 These brain-gut neuropeptides include vasoactive intestinal polypeptide (VIP), bombesin, neurotensin, substance P, enkephalin, somatostatin, and cholecystokinin/gastrin. Vasoactive intestinal polypeptide is the most abundant and is present in large quantities throughout the gut.’ We report here an investigation, by immunocytochemistry and radioimmunoassay, of the pathology of the VIP-containing nerves in Crohn’s disease. Material and Methods Fresh, surgical specimens of smal1 and large bowel were obtained from 19 patients with suspected Crohn’s disease. For comparison, similar tissue was obtained from 6 patients with ulcerative colitis, and normal tissue from 8 patients undergoing jejunoileal bypass and 17 patients undergoing resection for carcinoma of the bowel (Table 1). Each specimen was processed within 1 hr of removal. A detailed macroscopic analysis was made, and al1 relevant features, such as muscle hypertrophy, edema, hemorrhage, ulceration, and fissuring, were recorded.

854

Table

BISHOP

1.

ET AL.

GASTROENTEROLOGY Vol. 79, No. 5, Part 1

Details of Patients Specimen site”

Group Crohn’s disease Ulcerative colitis Bypass surgery Carcinoma resection

Patient’s weight (kg)

No.

M

F

17

32 (14-63)”

7

10

6

36 (29-48)

3

3

6

37 (28-45)

1

7

126 (115-140)

17

68 (43-91)

12

5

59 (39-70)

Patients with rectum diarrhea

ileum 3

3

6

2

2

4

-

-

6

_

0

45 (37-68)b 55(41-63)

colon and rectum

ileum and colon

Sex Mean age (yr)

8 -

colon

_

_

-

-

11

-

3

6

17

14

’ Numbers in parentheses are ranges of patient’s ages. b Numbers in parentheses are ranges of patient’s weights. c Number of specimens taken.

Sampling

Method

The Crohn’s disease and ulcerative colitis specimens were sampled from the areas of maxima1 pathology and, where possible, from areas 1 cm proximal and distal. Samples from larger specimens were taken in addition from more distant areas above and below the lesion. Macroscopically normal samples were taken at resection for carcinoma (as far away from the tumor as possible), and at jejunoileal bypass, performed for obesity. The dimensions of specimens and the distances between samples and lesion areas were recorded. Al1 samples measured at least z x 2 cm and consisted of mucosa together with muscle wall. Each sample was bisected longitudinally. One half was processed for immunocytochemistry, and the other was frozen for subsequent extraction of peptide for radioimmunoassay. Part of the frazen tissue from each specimen was also lyophilized.

Tissue

Fixation

Characterization

Histology Serial sections were taken for conventional histology (hematoxylin and eosin) and immunocytochemistry. An analysis of the histologie features of each section was first made by two independent observers. Only clear-cut cases of Crohn’s disease (n = 17), ulcerative colitis (n = 6), and macroscopically and histologically normal pieces of control bowel were selected for further investigations. Diagnoses were made on the basis of histologie criteria of Morson and Dawson.’

Immunocytochemistry

and Processing

The tissue samples were al1 fixed in benzoquinone, either in solution or vapor forme7 Pieces of gut measuring no more than 2 x 2 cm were fixed by immersion in 0.4% benzoquinone in 0.01 M phosphate-buffered saline (PBS), pH 7.1-7.4, for 2 hr.’ lO-Pm Sections were cut in a cryostat and mounted on formol-gelatin coated slides.

Table 2.

For fixation in benzoquinone vapor, samples no larger than 1 x 1 cm were rapidly frozen in melting Arcton and were freeze-dried overnight at -40°C. The tissue was exposed to benzoquinone vapor for 3 hr at 60°C and infiltrated with paraffin wax.

Immunocytochemistry was carried out by the indirect immunofluorescence’ and the unlabeled antibody enzyme” (peroxidase-antiperoxidase) methods. Al1 dilutions of reagents were made in PBS. For indirect immunofluorescence, cryostat or freezedried paraffin wax sections were incubated with VIP anti-

of Antisera Optimal dilution

Peptide concentration for absorption (nmol/

Peptide antiserum raised against VIP Bombesin Enteroglucagon Gastrin Met-enkephalin Neurotensin Somatostatin Substance P o IF = immunofluorescence

technique.

IF”

PAPb

1:400 1:200 1:lOOO 1:200 1:400 1:400 1:400 1:zOOo

1:lZOO 1:5000 1:5000 1:zOOO 1:1wO 1:4OOO 1:6000 1:6000

b PAP = peroxidase-antiperoxidase

Region specificity C-terminal n ,, I, I, I, I, I, I, ,, I, Midportion c-terminal metbod.

ml diluted antiserum)

0.1 10 10 1 10 0.1 10 1

November

1980

VIP-CONTAINING

serulm (1 :300) for 16-20 hr. The fluorescein-conjugated goat anti1 Tabbit globulin second layer was applied at a diluSections were tion of 1: 150 for 1 hr at room temperature. and exammou mted in PBS-glycerine (1:9 PBSglycerine) microscope. For the perined I und er a Leitz fluorescente

Figu re

1.

VIP-immunostained disease (X 350).

nerves in the myenteric

NERVES IN CROHN’S DISEASE

855

oxidase-antiperoxidase (PAP) technique, sections w‘ere pretreated for 30 min each with 0.3% hydrogen peroxidc 9 to remove endogenous peroxidase and normal goat ser ‘urn (dilution 1: 30) to black possible background staining. The VIP antiserum was used at a dilution of 1: 1200 for

plexus of: A, normal colon

(x,

350); B, colon removed from a patient with Cro hn’s

856

BISHOP ET AL.

goat 16-20 hr at 4% The second layer of unconjugated antirabbit globulin (1: 50) and third layer of PAP complex (1: 300) were both applied for 30 min at room temperature. The peroxidase of the PAP complex was made visible by the diaminobenzidine method of Graham and Karnovsky.ll The sections were dehydrated, mounted in synthetic medium, and examined under a Leitz transmitted-light microscope. In addition to the VIP antiserum, antisera to the following peptides were also used: substance P, somatostatin, met-enkephalin, cholecystokinin, bombesin, enteroglucagon, and neurotensin’z-18 (Table 2). A blind assessment of the immunocytochemical results was carried out by two independent observers.

GASTROENTEROLOGY

the assay was sensitive to changes of 3 fmol/assay tube (95% confidence limit). Addition of up to 200 ~1 of 0.1 M formic acid to the assay did not interfere with the binding of antibody to label. Extracts of liver with similar protein concentration to the gut extracts (determined by 280 nm absorption) also showed no nonspecific interference.

Statistics The results of the radioimmunoassay of VIP content were examined statistically cation of a one-way analysis of variante.

Vasoactive

Controls Controls were those proposed by Sternberger,” including the use of VIP antiserum previously incubated with excess VIP (0.1-10 nmol/ml diluted antiserum). Nonimmune rabbit serum was used as first layer and either the fluorescein-conjugated globulin or the peroxidase-antiperoxidase complex was applied singly.

Tissue

Extracts

For radioimmunoassay, tissue samples were kept deep-frozen at -20°C until the time of extraction. Individual tissues were weighed while stil1 frozen and plunged into vigorously boiling water (5 ml/g wet wt of tissue) for 5-10 min depending on size, in order to destroy proteolytic enzymes. After cooling to room temperature, the tissues were finely minced and then homogenized in the same water made up to 5 ml/g. Homogenization was carried out using an Ultraturrax homogenizer (ranke and Kunkel, West Germany) at 20,000 rpm. The resulting extracts were allowed to mix gently for 1 hr at room temperature by “over and over” rotation in order to solubilize peptides and were then centrifuged at 2500 g for 20 min. The supernatants were decanted and saved, and the tissue pellet was rehomogenized in 4 ml of water /g wet wt of tissue. The solubilization and centrifugation processes were repeated, and the fïrst and second extract supernatants were pooled. The total volume of the water extract was measured, and the sample was stored at -2O’C until assay. The tissue debris was then extracted twice more first with 5 ml then with 4 ml of 0.1 M formic acid per gram of tissue

(original wet weight), as for the water extraction. The formic acid extracts were pooled, the volume measured and the sample stored at -20°C. Samples were assayed at three separate dilutions 1/8, 1/80, and 1/800.

Radioimmunoassay The development of a radioimmunoassay for the measurement of plasma concentrations of VIP has been reported elsewhere. *’In the present study, a similar assay system was adopted, differing only in that no plasma was added to the assay tubes. The antiserum was used at a dilution of 1: 320,000, and

Vol. 79, No. 5, Part I

Intestinal

Polypeptide

measurement by the appli-

Antiserum

The VIP antiserum for immunocytochemistry and radioimmunoassay was raised in New Zealand white rabbits against natura1 porcine VIP coupled to bovine serum albumin by the carbodiimide methodXg and was found to be c-terminally directed. On radioimmunoassay, the VIP antiserum showed no cross-reactivity with glucagon, secretin, or gastric inhibitory polypeptide-which are structurally related to VIP-neither did it cross-react with any of the known neuropeptides. The VIP-immunostaining was not quenched by prior incubation of the antiserum with these peptides (concentrations 0.1-40 nmol/ml diluted antiserum).

Table

3.

Results of Blind Assessment of Vasoactive Intestinal Polypeptide-Containing Nerves”,b

Crohn’s disease AA A A AA A A AABA AAB BB B B B B

Ulcerative colitis AA ACAC cc C CCCE CDE

Carcinoma resection cc cc CE DC DD DD DD D D D D DE

BBBCB BBBC BCB

DE DE EE E

CCBB

E

Bypass Al1 E

” Key: A = Total transmural increase in number of immunostained VIP-containing nerves and intensity of immunostain; nerves thickened and distended. B = Total transmural increase in VIP-containing nerves; only few nerves thickened and distended. C = VIP-containing nerves increased, but only in mucosa and submucosa. D = Slight increase in VIP-containing nerves in mucosa and submucosa. E = NO change, normal. b Each line in table refers to grading of single or multiple samples from 1 patient.

November

VIP-CONTAINING

1980

Figure 2. An aggregation

of VIP-immunoreactive

857

fibers in colonic mucosa of a patient with Crohn’s disease (X 350).

Results Conventional

NERVES IN CROHN’S DISEASE

Histology

In accordance with the selection procedure, the classica1 histologie characteristics of each disease were seen in the abnormal specimens.* The carcinema resection and bypass tissues were found to have no histologie abnormalities. Immunocytochemistry Controls. Al1 25 control specimens showed the characteristic distribution pattern of VIP-containing nerve fibers consisting of fine, beaded fibers distributed across the whole width of the gut wall, weaving in and out of the muscle layers, and forming meshes in the submucosa. The fibers ran along the length of the villi and finished in close contact with the mucosa. In the myenteric and submucous plexi they surrounded nonimmunoreactive ganglion cells (Figure 1A). In addition, very occasional immunoreactive ganglion cells were observed in the submucous plexus (Table 3). Ulcerative colitis. Four out of the six ulcerative colitis specimens showed a pattern of distribution of VIP fibers identical to that observed in the control specimens, with no alteration in the intensity of the immunostain or in the number of immunoreactive ganglion cells in the submucous plexus. The

remaining two cases showed, however, an increase in VIP-immunostained nerve fibers (grades AA and ACAC) (Table 3). Crohn’s disease. In contrast with the controls and most of the ulcerative colitis specimens, al1 the Crohn’s disease tissues exhibited changes in VIP innervation. The mucosa and submucosa of each Crohn’s disease specimen showed these changes, but 13 of the 17 cases had a total transmural increase in VIP-immunostained nerves. The changes took the form of thicker and more brightly immunostained fibers in each layer of the gut wall, with a greater population of fibers in the muscle and more densely packed and disorganized meshes in the submucosa. In some specimens, immunoreactive fibers could be found in the densely packed nerves observed in the granulomatous areas after conventional histologie staining, an observation not made in either the control or the ulcerative colitis specimens (Figure 2). Intensely immunostained fibers filled both the myenteric and submucous plexi, which were often enlarged and infiltrated the muscle layers (Figure 1B). A clear increase was seen in the number of immunostained ganglion cells in the submucous plexus (Figure 3). Semiquantitative assessment showed a range of 6-8 immunoreactive cel1 bodies in ganglia in the Crohn’s disease samples (0-2 in controls). These features were detected not only in the areas of maximal pathology, but also in adjacent, apparently normal bowel where conventional histologie staining

858

BISHOP

GASTROENTEROLOGY

ET AL.

plexus of colon of a patient FigfJ1.e 3. The submucous nonimmunoreactive ganglion cells (X 400).

with

Crohn’s

only reveal slight neural changes. Similar changes in VIP-immunostained nerves were observed in specimens from the ileum, colon, and rectum (Table 3). Using the described antiserum to VIP (C-terminally directed) immunostaining in the smal1 and large bowel was almost exclusively in nerves. The application of antisera to peptides other than VIP showed no alteration in either peptide-containing nerves (substance P and met-enkephalin) nor in endocrine cells (neurotensin, somatostatin, and enteroglucagon).

can

disease.

VIP immunoreactive

fibers

surround

Vol. 79, No. 5, Part

immunoreactive

Al1 immunostaining obtained with the antiserum to VIP could be removed by the prior absorption of the antiserum with 0.1 nmol VIP/ml diluted antiserum (Table 2). Neither the normal rabbit serum nor the conjugated final layers gave any immunostaining.

In al1 cases, VIP immunoreactivity from both control and diseased tissue extracts diluted in parallel fashion to the standard curve. A one-way analysis of the variante of the results revealed that the VIP content of the Crohn’s disease tissues was significantly higher (P c 0.001) than that of any of the controls. The differente in VIP content was also found to be significant after lyophilization of the tissues. NO significant differente was found between the VIP levels in the ulcerative colitis and control specimens.

We have demonstrated a substantial increase in the content of VIP and of VIP-immunostained

Table 4.

Vasoactive Intestinal Polypeptide Content of Crohn’s Disease, Ulcerative Colitis, Carcinoma Resection, and Bypass Procedure Specimens VIP”

Tissue

Radioimmunoassay

and

Discussion

Con trols

Results

A large, approximately 200%, increase in VIP content was detected in the Crohn’s disease specimens in comparison with al1 the other types of specimens examined (Table 4 and Figure 4).

1

294 93 108 86

Crohn’s disease Ulcerative colitis Carcinoma resection Ileum from bypass procedure 0 Values

expressed

as pmol/g

wet weight

+ SEM.

f f + f

29 5 38 27

November

VIP-CONTAINING

1980

Crohn’s Disease Figure

4. Histogram

displaying

Ulcerative Carcinoma By-Pass Colitis Resection results

of radioimmunoassay

termination of VIP content of Crohn’s disease, ative colitis, carcinoma resection, and bypass dure specimens.

deulcerproce-

nerves in the gut wal1 of patients with one type of inflammatory bowel disease, Crohn’s disease, though not in the other main type, ulcerative colitis, as compared with controls (carcinoma resection and bypassed gut from obese patients). Although the increase in number, size, and intensity of the immunostain of VIP-containing fibers was always seen in the mucosa and submucosa, in most cases (13 out of 17) the changes were transmural. These observations represent, on average, a greater than 200% increase in VIP content of the gut wall, as measured by radioimmunoassay of tissue extracts. This increase in VIP was also observed in tissues, from each of the four groups, which had been lyophilized, indicating a similar degree of water loss in al1 specimens, Thus, edematous changes in the abnormal tissues did not appear to significantly alter the VIP concentration in the tissues. Abnormalities of the autonomie nerves in Crohn’s disease, including a rise in the number of ganglion cells and a genera1 nerve proliferation (neuromatous hyperplasia) were first noted more than 20 yr ago.’ However, their significante remains unknown. Specific VIP immunostained indicates that these autonomic nerve changes are greater than those revealed by conventional staining and can be accounted for, at least in part, by the significantly increased content of VIP-immunostained nerves reported here. The absente of VIP-immunostained nerve changes in the majority of ulcerative colitis specimens examined (four out of six), together with the

NERVES IN CROHN’S

DISEASE

859

finding of consistent VIP increases in the mucosa and submucosa of Crohn’s disease patients, may provide the basis for a useful diagnostic tool. However, itappears that there may be a spectrum of VIPimmunostaining nerve changes in both these diseases. Thus, the double-blind analysis of a large number of endoscopic biopsy specimens may be necessary to fully validate this proposal. One possible explanation for this phenomenon may reside in the severe transmural inflammatory process, as seen in Crohn’s disease, which may act as an irritant stimulus for the proliferation of intrinsic’” VIP ganglion cells, a stimulus possibly absent in al1 but the most severe cases of ulcerative colitis. It is interesting to note that the changes in VIP innervation in Crohn’s disease can be detected in histologically normal bowel adjacent to the areas of maxima1 pathology. This finding argues against the changes being a mere secondary response to the irritant stimulus of the inflammatory process, although the possibility of previous inflammation, subsequently healed, cannot be ruled out. Comparison between the VIP-immunostained nerve changes and the histologie and clinical data revealed no absolute correlation. Vasoactive intestinal polypeptide is powerfully active in the gut. It causes diarrhea and affects not only motility but also, as its name suggests, blood flow.‘l There is firm evidente that it can function as a neurotransmitter.” It is possible that the increase in VIP innervation described here could have a significant effect on gut function in Crohn’s disease, and its discovery may further help in the understanding of the pathophysiology of this condition.

References 1. Davis DR, Dockerty MB, Mayo CW: The myenteric plexus in regional enteritis: a study of ganglion cells in the ileum in 24 cases. Surg Gynecol Obstet 101208, 1955 2. Morson BC, Dawson IMP: Inflammatory disorders. In: Gastrointestinal Pathology. Edited by BC Morson, IMP Dawson. Oxford, London, Edinburgh, Blackwell Scientific Press, 1972, p 243-298 3. Whitehead R: Pathology of Crohn’s disease of the colon. In: Inflammatory Bowel Disease. Edited by JB Kirsner, RG Shorter. Philadelphia, Lea & Febiger, 1975, p 182-196 4. Polak JM, Bloom SR: Peptidergic nerves of the gastrointestina1 tract. Invest Cel1 Pathol 1:361, 1978 5. Polak JM, Bloom SR: Peptidergic innervation of the gastrointestinal tract. In: Gastrointestinal Hormones and Pathology of the Digestive System. Edited by M Grossman, V Sieranza, N Basso, E Lezoche. New York, London, Plenum Press, 1977, p 27-49 6. Polak JM, Bloom SR: The neuroendocrine design of the gut. J Clin Endocrinol Metab 8:313, 1979

860

BISHOP ET AL.

7. Pearse AGE, Polak JM: Bifunctional reagents as vapour and liquid phase fixatives for immunocytochemistry. Histochemistry J 7:179, 1975 8. Bishop AE, Polak JM, Bloom SR, Pearse AGE: A new universal technique for the immunocytochemical localisation of peptidergic innervation. J Endocrinol 7725, 1978 9. Coons AH, Leduc EH: Studies on antibody production. 1. A method for the histochemical demonstration of specific antibody and its application to a study of the hyperimmune rabbit. J Exp Med 10249, 1955 10. Sternberger LA: The unlabelled antibody enzyme method. In: Immunocytochemistry. Edited by LA Sternberger. Englewood Cliffs, N.J. Prentice Hall, Inc., 1974, p 129-171 11. Graham RC, Karnovsky MJ: The early stages of absorption of injected horse radish peroxidase in the proximal tubules of mouse kidney: ultra-structural cytochemistry by a new technique. J Histochem Cytochem 14291, 1966 12. Wharton J, Polak JM, Bloom SR, et al: Substance P-like immunoreactive nerves in mammalian hing. Invest Cel1 Pathol 2:3, 1979 13. Polak JM, Pearse AGE, Grimelius L, et al: Growth-hormone releasing inhibiting hormone (GH-RIH) in gastrointestinal and pancreatic D cells. Lancet i:1220, 1975 14. Polak JM, Sullivan SN, Bloom SR, et al: Enkephalin-like immunoreactivity in the human gastrointestinal tract. Lancet i:972, 1977

GASTROENTEROLOGY

Vol. 79, No. 5. Part

1

15. Polak JM, Pearse AGE, Szelke M, et al: Specific immunostaining of CCK cells by use of synthetic fragment antisera. Experientia (Basel) 33:762, 1977 16. Wharton J, Polak JM, Ghatei MA, et al: Bombesin-like immunoreactivity in the lung. Nature (Lond) 273:769, 1978 17. Grimelius L, Polak JM, Solcia E, Pearse AGE: The enteroglucagon cell. In: Gut Hormones. Edited by SR Bloom. Edinburgh, Churchill Livingstone, 1978, p 365-368 18. Polak JM, Sullivan SN, Bloom SR, et al: Specific localisation of neurotensin to the N cel1 in human intestine by radioimmunoassay and immunocytochemistry. Nature (Land) 270:183, 1977b 19. Mitchell SJ, Bloem SR: Measurement of fasting and postprandial VIP in man. Gut 19:1043, 1978 20. Jessen KR, Polak JM, Van Noorden S, et al: A new approach to the demonstration of the enteric plexus origin of peptidecontaining nerves by immunocytochemistry and radioimmunoassay (abstr). Gastroenterology 76:1161, 1979 21. Makhlouf GM, Said SI: The effect of vasoactive intestinal polypeptide (VIP) on digestive and hormonal function. In: Gastrointestinal Hormones. Edited by JC Thompson. Austin, University of Texas Press, 1975, p 599-610 22. Fahrenkrug JF, Haglund U, Jodal M, et al: Possible physiological implications of nervously released vasoactive intestinal polypeptide (VIP) in the gastrointestinal tract. Stand J Gastroenterol 13:59, 1978