- Email: [email protected]
Acquired intestinal aganglionosis and circulating autoantibodies without neoplasia or other neural involvement
GASTROENTEROLOGY 1997;112:1366–1371
Acquired Intestinal Aganglionosis and Circulating Autoantibodies Without Neoplasia or Other Neural Involvement VIRPI V. SMITH,* NORMAN GREGSON,‡ LUKAS FOGGENSTEINER,§ GRAHAM NEALE,§ PETER J. MILLAx Departments of *Histopathology and xGastroenterology, Great Ormond Street Hospital for Children National Health Service Trust and the Institute of Child Health, London; ‡Department of Neurology, United Medical and Dental School, Guy’s Hospital Campus, London; and § Department of Gastroenterology, Addenbrooke’s National Health Service Trust, Cambridge, England
The clinical course, diagnosis, and treatment of 2 patients with acquired intestinal aganglionosis without other neurological involvement or neoplasia are described. They initially presented with constipation and abdominal pain in late childhood. They were found to have enteric ganglionitis with a loss of neurons together with vacuolated nerve cells surrounded by CD3and CD4-positive T lymphocytes. This process initially affected only the colon but later the entire gastrointestinal tract was involved in 1 patient. Associated with this process there were circulating immunoglobulin G class enteric neuronal antibodies in high titer (1:5000– 8000). The staining of central nervous system neuronal nuclei and Western blotting indicated the presence of antineuronal nuclear protein antibodies of the ANNA-1 (anti-Hu) type usually associated with paraneoplastic sensory neuropathy. However, the reaction pattern in enteric neurons was quite different with strong reaction to perikarya and only weak staining of nuclear antigens.
A
bnormalities of enteric nerves or intestinal smooth muscle result in functional intestinal obstruction. The most common of these abnormalities is congenital aganglionosis (Hirschsprung’s disease), in which there is an absence of neurons in both the enteric plexuses because of a variety of developmental defects. The existence of acquired aganglionosis was alluded to by Valdes-Dapena in 19701 with a description of previously ganglionic segments of gut becoming aganglionic, but the mechanism for this was not clear. Since then, acquired forms of aganglionosis have been described in which enteric neurons seem to have been destroyed by an inflammatory process. This occurs most commonly as part of paraneoplastic neurological syndromes2,3 but also in a central neurological syndrome in the absence of cancer.4 In this report, we describe 2 patients who have no central nervous system involvement and no neoplasia but who have acquired aganglionosis as a result of enteric ganglionitis and circulating immunoglobulin (Ig) G class autoanti/ 5e1b$$0042
03-14-97 17:19:18
gasa
bodies directed against enteric neurons. The diagnosis and clinical management of these patients are discussed.
Case Reports Case F.H. A previously healthy Omani girl presented at the age of 10 years with anorexia, abdominal distention, abdominal pain, vomiting, and severe constipation that failed to respond to large doses of aperients or laxatives. Suction rectal biopsy specimens were aganglionic, and a diagnosis of Hirschsprung’s disease was made. Despite repeated rectal wash-outs and intravenous fluids, she required surgical decompression, and total parenteral nutrition was commenced. At laparotomy, no obstruction was found, an ileostomy was raised, and full-thickness samples of large and small bowel were taken for histological analysis. After the ileostomy, her condition improved, and during a 3-month period, no further episodes of obstruction occurred. Therefore, an ileoanal pouch was constructed. After this procedure, she once again developed further episodes of functional obstruction, and another ileostomy was performed. Despite this procedure, passage of contents through the small intestine failed, and total parenteral nutrition was resumed. Levels of plasma electrolytes, full blood count, erythrocyte sedimentation rate, liver and renal function test results, and trace elements were all normal. Circulating Igs were normal except for an increased serum IgM. An autoantibody profile was negative for thyroid, gastric parietal cells, pancreatic islet cells, enterocytes, smooth muscle, mitochondria, and microsomes. Antroduodenal manometry showed normal phasic activity with a frequency of 3 cycles/min in the gastric antrum. In the duodenum, irregular high-amplitude multiphasic complexes were observed, and these were propagated both aborally and adorally. No cyclical fasting activity occurred during 4 hours of recording after an 18-hour fast. Cisapride increased the frequency of the contractions but did not produce phase III– like complexes. Although a normal gastric antral response was present after a milk feed, no postprandial contractile response q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00
WBS-Gastro
April 1997
ACQUIRED AGANGLIONOSIS 1367
was observed in the duodenum. Surface electrogastrography showed a dominant frequency of 3 cycles/min on running spectral analysis,5 and this correlated with normal phasic activity as shown by gastric antral manometry. Computerized tomography and magnetic resonance imaging scans of the brain, peripheral nerve conduction studies, and peripheral electromyogram were normal, but a pure tone audiogram showed sensory neural deafness compatible with aminoglycoside damage. She was known to have been administered several poorly controlled aminoglycosides for central line infections at other institutions. Treatment with prednisolone was started initially at 2 mg/ kg body wt for 4 weeks, decreasing to 0.5 mg/kg body wt on alternate days. Domperidone suppositories (30 mg three times daily) were also administered. On this regime, we were gradually able to introduce enteral nutrition and discontinued total parenteral nutrition. Five years after the original episode and after discontinuation of her steroid therapy at another institution, her condition deteriorated with vomiting of her enteral feeds, severe abdominal pain, abdominal distention, and intestinal obstruction. Her severe abdominal pain was partially controlled by celiac plexus blocks and then celiac plexus ablation together with increasing doses of opiates (pethidine and methadone). Cessation of pain and vomiting was finally achieved with the 5-hydroxytryptamine type 3 antagonist ondansetron (8 mg twice daily). Unfortunately, she had already become addicted to opiates requiring methadone (5 mg twice daily). Because of precarious venous access, she was considered for intestinal transplantation, which has now been performed successfully. However, in the time leading up to the transplantation, she continued to denervate her esophagus and is now showing symptoms of impaired esophageal motility. However, the transplanted stomach and small intestine are functioning satisfactorily, and total parenteral nutrition has been stopped. She is currently fed via a gastrostomy.
Case J.L. A 23-year-old white man presented with recurrent episodes of intestinal obstruction. He had no family history of autoimmune disease, malignancy or gastrointestinal pathology, was a smoker, and drank moderately. He had been well until the age of 5 years when he developed type 1 diabetes mellitus. At the age of 11 years, he began to experience episodes of abdominal pain associated with apparent constipation that were often followed by bouts of diarrhea. These episodes became progressively more frequent. At the age of 23 years, he was admitted to a hospital with abdominal pain and distention. A laparotomy was performed, and several adhesions were divided. However, over the ensuing year, three further laparotomies were required for intestinal obstruction. No apparent mechanical cause was found, and total parenteral nutrition was commenced. When the patient was stable and established on total parenteral nutrition, urea, creatine, electrolytes, full blood count, erythrocyte sedimentation rate, C-reactive protein, serum electrophoresis, and thyroid function tests were all normal. Circu-
/ 5e1b$$0042
03-14-97 17:19:18
gasa
lating Igs (IgG, IgM, and IgA) and complements (C3 and C4) were normal, whereas the CH50 level was reduced (38 U/ mL). Tests for circulating immune complexes were also within normal limits. Autoantibody screen for antinuclear, antimitochondrial, anti–smooth muscle, anti–gastric parietal cell, antireticulin, and anti–liver-kidney-muscle antibodies was negative. Serum magnesium level was persistently low, and gglutamyltransferase and alkaline phosphatase levels were increased. Finally, an ileostomy was fashioned with good symptomatic relief. Six months later the ileostomy was closed, a subtotal colectomy was performed, and the ileum was anastomosed to the sigmoid colon. Total parenteral nutrition was discontinued, and a course of prednisolone (10 mg/day) was administered for a trial period of 6 weeks. On withdrawal of steroids, he had further episodes of intestinal obstruction. Since restarting prednisolone (10 mg/day), he is well but still requires parenteral nutrition 3 days per week. No peripheral neuropathy was noted at any time.
Results Intestinal Biopsy Specimens Full-thickness intestinal samples from F.H. and J.L. were fixed and processed routinely into paraffin wax. Sections were stained with H&E and immunostained for white blood cells using the avidin-biotin-peroxidase complex method.6 The white cell markers (Dako Ltd., High Wycombe, England) included antibodies for T lymphocytes (CD4 and CD3), monocytes and macrophages (CD68), neutrophils (CD15), activated white cells (CD45 and CD30), and B lymphocytes (CD20). From patient F.H. snap-frozen full-thickness ileum resected 2 years after the previous sample was also available and was used for assessment of acetyl cholinesterase activity and immunostaining for neurofilament triplet protein using neurofilament antibody clones NN18, NE14, BF10, NR4, N52 (Sigma Chemical Co., Poole, Dorset, England), RT97 (gift from Professor Brian Anderton, Institute of Psychiatry, University of London, London, England), SM132 (Sternberger, Baltimore, MD), and neural cell adhesion molecule (Eric 1; gift from Professor Frank Walsh, Guy’s Hospital, London, England). The large bowel resected at the time of the pullthrough operation from patient F.H. was aganglionic, but in the samples of ileum, a few degenerate neurons surrounded by mononuclear cells were observed. Although full-thickness ileum from patient J.L. appeared unremarkable, sections of the resected colon and appendix showed marked myenteric ganglionitis with degeneration and loss of neurons. The inflammatory infiltrate in both patients consisted of mononuclear cells with immunoreactivity mostly associated with CD4- and CD3-positive T lymphocytes (Figure 1) although a few B cells WBS-Gastro
1368 SMITH ET AL.
GASTROENTEROLOGY Vol. 112, No. 4
patient F.H. showed strong mainly perikaryal but also weak nuclear immunostaining (IgG) with enteric neurons at a dilution of 1:8000 (Figure 2). A similar pattern was observed with serum from patient J.L. up to a dilution of 1:5000. No IgM immunoreactivity was detected with sera from any of the patients. Serological Testing for Circulating Autoantibodies Against Central Nervous System Neurons
Figure 1. Section of intestine from patient J.L. immunostained for a pan T-lymphocyte marker (CD3) showing that the inflammatory infiltrate predominantly consists of T cells (original magnification 1601).
were also observed. From patient F.H., the progression of the disease process could be assessed in a further ileal sample 2 years from the first specimen of ileum. The lymphocytic infiltrate and neurons had disappeared from the previously inflamed myenteric plexus that had originally contained degenerate neurons. The whole bowel had become denervated with no neurofilament immunostaining in the myenteric plexus and no acetyl cholinesterase-positive nerve fibers in the circular muscle. However, there was evidence of glial elements that were recognized by immunoreactivity to neural cell adhesion molecule. Serological Testing for Autoantibodies Against Enteric Neurons Sera were obtained from F.H. and J.L. together with sera from control subjects. The control sera consisted of sera from a healthy adult and from 14 control patients aged 3 months to 46 years, and these were studied for the presence of circulating antibodies against enteric neurons. The control patients had inflammatory bowel disease (n Å 2), aganglionosis (n Å 1), ileal stricture (n Å 1), esophageal achalasia (n Å 1), hypoganglionosis (n Å 1), intestinal myopathy (n Å 1), unspecified functional intestinal obstruction (n Å 3), juvenile Batten’s disease (n Å 1), rectal ganglionitis (n Å 1), mucosal ulceration (n Å 1), and unspecified functional intestinal obstruction with severe abdominal pain (n Å 1). Sera were serially diluted and incubated with cryostat sections of neonatal piglet small intestine. The incubation with the primary serum was followed by a polyclonal anti-human IgG and IgM antibodies (Dako Ltd.). The avidin-biotin-peroxidase complex technique6 was used to detect the presence of the antibody. No immunoreactivity with enteric neurons was observed with sera from the control patients. Serum from / 5e1b$$0042
03-14-97 17:19:18
gasa
Sera from F.H. and J.L. were tested together with sera from 6 of the control patients with enteric neuromuscular disease described above in random sequence together with other sera sent for diagnostic screening for paraneoplastic antineuronal antibodies (1500 samples/ year). The sera were serially diluted and applied to cryostat sections of human cerebellum, cerebral cortex, and brain stem. The bound antibody was visualized by the indirect immunoperoxidase technique.7 Positive control sera from 3 patients with paraneoplastic neurological syndrome known to contain antineuronal antibodies (antiHu, anti-Yo, and anti-Ri), and a known negative control serum were run in parallel. Of the eight sera from patients with enteric neuromuscular disease tested, only sera from F.H. and J.L. produced staining of the cytoplasm as well as the nuclei of all neurons in sections of human cerebellum, frontal cortex, and brain stem (Figure 3). The nuclei of glia and vascular endothelium were not stained, and the pattern was similar to that observed with control sera with antinuclear neuronal antibodies type I from patients with small cell lung carcinoma and sensory encephalopathy. The sera from the 6 control patients with enteric neuromuscular disease and the serum from the healthy adult produced no staining of brain sections at a dilution of 1:100. The end point dilution determined by immuno-
Figure 2. Serum (1:1280 dilution) from patient J.L. applied to a section of normal neonatal piglet small intestine showing strongly positive myenteric neurons. Note that the immunoreactivity can be mainly seen in the neuronal perikarya (original magnification 4001).
WBS-Gastro
April 1997
ACQUIRED AGANGLIONOSIS 1369
Figure 3. Indirect immunoperoxidase of human brain. Sections of the brain stem treated with (A) normal control human serum (1:200 dilution) treated with (B) serum from patient F.H. (1:400 dilution). Sections of frontal cortex treated with (C) serum from patient F.H. (1:400 dilution), (D) serum from patient J.L. (1:400 dilution), and (E) with a positive control serum from a patient with anti-Hu antibodies and small cell lung carcinoma (1:500 dilution) and counterstained with Mayer’s hematoxylin. Notice nuclei of cell in the blood vessels (small arrow) and glia stained with hematoxylin in E and their absence in B, C, and D (bar Å 25 mm in A and B and 50 mm in C, D, and E).
peroxidase was 1:1600 and 1:4000 with cerebral cortex and cerebellum, respectively (F.H.), and 1:8000 and 1:2000 (J.L.). Western Blotting Antigen specificity was determined by Western blotting7 against human cerebral cortex and cerebellum, rat forebrain, and small cell lung carcinoma cell line NCl-H69. Control sera containing anti-Hu and anti-Ri as well as normal negative sera were run as primaries in parallel. Only sera from patients F.H. and J.L. showed any binding to protein on Western blotting of brain tissue. The sera reacted with protein bands of apparent molecular weights of 38 and 45 kilodaltons in the brain and with similar proteins of approximately 38–39 kilodaltons in / 5e1b$$0042
03-14-97 17:19:18
gasa
the small cell lung carcinoma cell line NCl-H69, as did the control anti-Hu serum (Figure 4).
Discussion Aganglionosis occurs in 1 of 4500 live births and is diagnosed most commonly in the neonatal period. However, in about 7%–8% of patients, aganglionosis is not recognized until later in life.8 We describe 2 patients with episodes of pseudoobstruction caused by an acquired progressive aganglionosis as a result of a severe T-cell– mediated inflammatory disorder involving the enteric nerve plexuses and associated with a circulating IgGclass autoantibody directed against enteric neurons. At the aganglionosis stage, the appearances in H&E-stained sections could not be distinguished from classical Hirschsprung’s disease. These patients together with 3 subjects WBS-Gastro
1370 SMITH ET AL.
GASTROENTEROLOGY Vol. 112, No. 4
Figure 4. Western blot against proteins of human cerebral cortex and small cell lung carcinoma. Strips contain proteins from small cell lung carcinoma and human cerebral cortex in lanes 1 and 2. Lane 1, serum from patient F.H.; lane 2, serum from positive control patient with anti-Hu antibodies and small cell lung carcinoma; lane 3, serum from patient J.L.; and lane 4, serum from normal control (first lane standards). All sera were at 1:1000 dilution. kD, Kilodalton.
described in a preliminary communication by DeGiorgio et al.9 show that at least some patients who present late with aganglionosis could have acquired it rather than it being a congenital disease. As in patients with paraneoplastic syndrome,2,10,11 sera from our patients contained high titers of an IgG-class circulating antibody against nerve cells. In the paraneoplastic syndrome, the antineuronal antibodies have been shown to be of variable antigenic specificity and encompass proteins such as anti–neuronal nuclear antibodies, which include anti-Hu (HuC, HuD, and HelN1) and anti-Ri proteins12 as well as an anti-Purkinje cell cytoplasmic antibody (anti-Yo).11 The immunostaining with the sera from our patients was mostly cytoplasmic with only a weak nuclear component in the enteric nerve cells, and this was different from the anti-Hu–associated pattern in which the nuclear staining predominates.11 However, the results of the Western blots and the staining pattern in the brain sections do not allow distinction from the anti-Hu proteins found in paraneoplastic neurological syndrome. It is generally accepted that the presence of autoantibodies directed against neurons is an indication of a paraneoplastic syndrome often with only a small neoplastic lesion, which aberrantly expresses neuron-specific antigens.12 However, in patients with sensory encephalopathy and anti-Hu antibodies, often no tumor is discovered even at autopsy.13 We found no evidence of neoplasia in our patients, despite extensive investigations during a period of up to 8 years. It has been suggested that the immune response is associated with a complete tumor remission.12 However, it is also possible that the antibodies could arise through mechanisms not associated with malignancy. Parasitic infections, such as trypanosomiasis / 5e1b$$0042
03-14-97 17:19:18
gasa
(Chagas’ disease), have been postulated to elicit an autoimmune response resulting in neuronal degeneration producing esophageal achalasia and megacolon.14 Interestingly, no circulating enteric neuronal antibodies were found in our achalasia control. The autoimmune injury in our patients appeared clinically to be limited to the enteric nervous system, although the antibody is capable of binding in vitro to central neurons. We postulate that the apparent absence of central nervous system symptoms is caused by the fact that the symptoms are the result of local T-cell–mediated responses rather than the production of an antibody, with antibody production being another consequence of the T-cell response. An alternative explanation could be that quantitative differences in epitopes in the central nervous system, dorsal root ganglia, and enteric nervous system are important because it is thought that there is a critical density of epitopes for the antibody to cause disease (D. Vergani, personal communication, June 17, 1996). Aganglionosis could be caused by other mechanisms such as apoptosis.15 However, we did not observe any features to suggest programmed cell death or other explanations for acquired aganglionosis. Indeed, on treatment with prednisolone, the progression of the disease seemed to be arrested; a relapse was observed at withdrawal; and at reintroduction of prednisolone, there was clear-cut clinical improvement providing further support in addition to the immunologic and histological features that the disease is caused by a T-cell–mediated immune response. As well as the loss of purposeful intestinal contractile activity, F.H. experienced severe abdominal pain. A variety of mechanisms could have induced the pain, including hyperactivity in primary afferent or central nervous system nociceptive neurons, loss of central inhibitory connections, and increased activity in sympathetic efferents.16,17 The lack of sustained effect of celiac plexus ablation suggests that the latter was not important in F.H.’s pain. However, the denervation of the gut could have resulted in hypersensitivity of the intact parts of the visceral nociceptive pathways probably at a spinal or central level. This notion is supported by the response to the 5-hydroxytryptamine antagonist ondansetron because it is known that 5-hydroxytryptamine is important in visceral nociception.18,19 We have discussed the clinical course, diagnosis, and treatment of 2 patients with acquired intestinal aganglionosis. In these patients, there appeared to be a T-cell– mediated inflammatory response directed against enteric neurons that also resulted in the production of circulating neuronal antibodies. In addition, there was an absence of other neurological or neoplastic disease. We suggest WBS-Gastro
April 1997
ACQUIRED AGANGLIONOSIS 1371
that in patients with late-onset persistent constipation associated with myenteric ganglionitis, the presence of circulating enteric neuronal antibodies should be sought.
References 1. Valdes-Dapena M. Megacolon in the adult. In: Valdes-Dapena AM, Stein GN, eds. Morphological pathology of the alimentary canal. Philadelphia: Saunders, 1970:659–660. 2. Lennon VA, Sas DF, Busk MF, Scheithauer B, Malagelada J-R, Camilleri M, Miller LJ. Enteric neuronal autoantibodies in pseudoobstruction with small-cell lung carcinoma. Gastroenterology 1991;100:137–142. 3. Gerl A, Storck M, Schalhorn A, Muller-Hocker J, Jauch KW, Schildberg FW, Wilmanns W. Paraneoplastic chronic intestinal pseudoobstruction as a rare complication of bronchial carcinoid. Gut 1992;33:1000–1003. 4. Horoupian DS, Kim Y. Encephalomyeloneuropathy with ganglionitis of the myenteric plexus in the absence of cancer. Ann Neurol 1982;6:628–632. 5. Devane SP, Ravelli AM, Bisset WM, Smith VV, Lake BD, Milla PJ. Gastric antral dysrhythmias in children with chronic idiopathic intestinal pseudo-obstruction. Gut 1992;33:1477–1481. 6. Hsu SM, Raine L, Fanger H. Use of avidin biotin peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabled antibody (PAP) procedures. J Histochem Cytochem 1981;29:577–580. 7. Ball JA, Warner T, Reid P, Howard RS, Gregson NA, Rossor MN. Central alveolar hypoventilation associated with paraneoplastic brain-stem encephalitis and anti-Hu antibodies. J Neurol 1994; 241:561–566. 8. Smith VV, Milla PJ. Aganglionosis may not only be due to Hirschsprung’s disease (abstr). Gut 1993;34:S12. 9. DeGeorgio R, Bassotti G, Stanghellini V, Barbara G, Tazzari P-L, Santini D, Suozzi A, Germani U, Velio P, Minni F, Morelli A, Corinaldesi R. Clinical morpho-functional and immunological features of idiopathic myenteric ganglionitis (abstr). Gastroenterology 1996;110:A655. 10. Jean WC, Dalmau J, Ho A, Posner JB. Analysis of the IgG subclass distribution and inflammatory infiltrates in patients with anti-hu– associated paraneoplastic encephalomyelitis. Neurology 1994; 44:140–147.
/ 5e1b$$0042
03-14-97 17:19:18
gasa
11. Altermatt HJ, Rodriguez M, Scheithauer BW, Lennon VA. Paraneoplastic anti-Purkinje and type 1 antineuronal nuclear autoantibodies bind selectively to central, peripheral and autonomic nervous system cells. Lab Invest 1991;65:412–420. 12. Rosenblum MK. Paraneoplasia and autoimmunologic injury of the nervous system: the anti-hu syndrome. Brain Pathol 1993; 3:199–212. 13. Dalman J, Graus F, Rosenblum MK, Posner JB. Anti-Hu associated paraneoplastic encephalomyelitis/sensory neuropathy. A clinical study of 71 patients. Medicine 1992;71:59–72. 14. Schuffler MD. Neuromuscular abnormalities of the small and large intestine. In: Whitehead R, ed. Gastrointestinal and oesophageal pathology. New York: Churchill Livingstone, 1989:329– 353. 15. Singaram C, Koch J, Gaumnitz EA, Goyal RK, SenGupta A. Nature of neuronal loss in human achalasia (abstr). Gastroenterology 1996;110:A259. 16. Cervero F, Janig W. Visceral nociceptors: a new world order? TINS 1992;15:374–378. 17. Fields HL. Painful dysfunction of the nervous system. In: Fields HL, ed. Pain. New York: McGraw–Hill, 1987:133–145. 18. Randich A, Gebhart GF. Vagal afferent modulation of nociception. Brain Res Rev 1992;17:77–99. 19. Thurston CL, Randich A. Quantitative characterization and spinal substrates of antinociception produced by electrical stimulation of the subdiaphragmatic vagus in rats. Pain 1991;44:201–210. Received July 17, 1996. Accepted January 6, 1997. Address requests for reprints: Virpi V. Smith, Ph.D., Department of Histopathology, Camelia Botnar Laboratories, Great Ormond Street Hospital for Children, Great Ormond Street, London WC1N 3JH, England. Fax: (44) 1-71-813-1170. The authors thank the following who made important contributions to the diagnosis and management of these 2 patients: Dr. J. Hopkinson for making the original histological diagnosis of myenteric ganglionitis in case J.H. and for allowing us to characterize the lymphocytic infiltrate in the intestine; Dr. I. Filipe for the original paraffin blocks from case F.H., and B. D. Lake and Dr. D. G. D. Wight for noting the ganglionitis; R. Medcalf, Dr. A Tozzi, and H. Kaur for their excellent technical assistance in the screening of the sera; Dyanne Rampling for immunostaining the lymphocytes; and Dr. Jia Newcome for normal human brain tissue from the Multiple Sclerosis Brain Bank, Institute of Neurology.
WBS-Gastro
Acquired Intestinal Aganglionosis and Circulating Autoantibodies Without Neoplasia or Other Neural Involvement VIRPI V. SMITH,* NORMAN GREGSON,‡ LUKAS FOGGENSTEINER,§ GRAHAM NEALE,§ PETER J. MILLAx Departments of *Histopathology and xGastroenterology, Great Ormond Street Hospital for Children National Health Service Trust and the Institute of Child Health, London; ‡Department of Neurology, United Medical and Dental School, Guy’s Hospital Campus, London; and § Department of Gastroenterology, Addenbrooke’s National Health Service Trust, Cambridge, England
The clinical course, diagnosis, and treatment of 2 patients with acquired intestinal aganglionosis without other neurological involvement or neoplasia are described. They initially presented with constipation and abdominal pain in late childhood. They were found to have enteric ganglionitis with a loss of neurons together with vacuolated nerve cells surrounded by CD3and CD4-positive T lymphocytes. This process initially affected only the colon but later the entire gastrointestinal tract was involved in 1 patient. Associated with this process there were circulating immunoglobulin G class enteric neuronal antibodies in high titer (1:5000– 8000). The staining of central nervous system neuronal nuclei and Western blotting indicated the presence of antineuronal nuclear protein antibodies of the ANNA-1 (anti-Hu) type usually associated with paraneoplastic sensory neuropathy. However, the reaction pattern in enteric neurons was quite different with strong reaction to perikarya and only weak staining of nuclear antigens.
A
bnormalities of enteric nerves or intestinal smooth muscle result in functional intestinal obstruction. The most common of these abnormalities is congenital aganglionosis (Hirschsprung’s disease), in which there is an absence of neurons in both the enteric plexuses because of a variety of developmental defects. The existence of acquired aganglionosis was alluded to by Valdes-Dapena in 19701 with a description of previously ganglionic segments of gut becoming aganglionic, but the mechanism for this was not clear. Since then, acquired forms of aganglionosis have been described in which enteric neurons seem to have been destroyed by an inflammatory process. This occurs most commonly as part of paraneoplastic neurological syndromes2,3 but also in a central neurological syndrome in the absence of cancer.4 In this report, we describe 2 patients who have no central nervous system involvement and no neoplasia but who have acquired aganglionosis as a result of enteric ganglionitis and circulating immunoglobulin (Ig) G class autoanti/ 5e1b$$0042
03-14-97 17:19:18
gasa
bodies directed against enteric neurons. The diagnosis and clinical management of these patients are discussed.
Case Reports Case F.H. A previously healthy Omani girl presented at the age of 10 years with anorexia, abdominal distention, abdominal pain, vomiting, and severe constipation that failed to respond to large doses of aperients or laxatives. Suction rectal biopsy specimens were aganglionic, and a diagnosis of Hirschsprung’s disease was made. Despite repeated rectal wash-outs and intravenous fluids, she required surgical decompression, and total parenteral nutrition was commenced. At laparotomy, no obstruction was found, an ileostomy was raised, and full-thickness samples of large and small bowel were taken for histological analysis. After the ileostomy, her condition improved, and during a 3-month period, no further episodes of obstruction occurred. Therefore, an ileoanal pouch was constructed. After this procedure, she once again developed further episodes of functional obstruction, and another ileostomy was performed. Despite this procedure, passage of contents through the small intestine failed, and total parenteral nutrition was resumed. Levels of plasma electrolytes, full blood count, erythrocyte sedimentation rate, liver and renal function test results, and trace elements were all normal. Circulating Igs were normal except for an increased serum IgM. An autoantibody profile was negative for thyroid, gastric parietal cells, pancreatic islet cells, enterocytes, smooth muscle, mitochondria, and microsomes. Antroduodenal manometry showed normal phasic activity with a frequency of 3 cycles/min in the gastric antrum. In the duodenum, irregular high-amplitude multiphasic complexes were observed, and these were propagated both aborally and adorally. No cyclical fasting activity occurred during 4 hours of recording after an 18-hour fast. Cisapride increased the frequency of the contractions but did not produce phase III– like complexes. Although a normal gastric antral response was present after a milk feed, no postprandial contractile response q 1997 by the American Gastroenterological Association 0016-5085/97/$3.00
WBS-Gastro
April 1997
ACQUIRED AGANGLIONOSIS 1367
was observed in the duodenum. Surface electrogastrography showed a dominant frequency of 3 cycles/min on running spectral analysis,5 and this correlated with normal phasic activity as shown by gastric antral manometry. Computerized tomography and magnetic resonance imaging scans of the brain, peripheral nerve conduction studies, and peripheral electromyogram were normal, but a pure tone audiogram showed sensory neural deafness compatible with aminoglycoside damage. She was known to have been administered several poorly controlled aminoglycosides for central line infections at other institutions. Treatment with prednisolone was started initially at 2 mg/ kg body wt for 4 weeks, decreasing to 0.5 mg/kg body wt on alternate days. Domperidone suppositories (30 mg three times daily) were also administered. On this regime, we were gradually able to introduce enteral nutrition and discontinued total parenteral nutrition. Five years after the original episode and after discontinuation of her steroid therapy at another institution, her condition deteriorated with vomiting of her enteral feeds, severe abdominal pain, abdominal distention, and intestinal obstruction. Her severe abdominal pain was partially controlled by celiac plexus blocks and then celiac plexus ablation together with increasing doses of opiates (pethidine and methadone). Cessation of pain and vomiting was finally achieved with the 5-hydroxytryptamine type 3 antagonist ondansetron (8 mg twice daily). Unfortunately, she had already become addicted to opiates requiring methadone (5 mg twice daily). Because of precarious venous access, she was considered for intestinal transplantation, which has now been performed successfully. However, in the time leading up to the transplantation, she continued to denervate her esophagus and is now showing symptoms of impaired esophageal motility. However, the transplanted stomach and small intestine are functioning satisfactorily, and total parenteral nutrition has been stopped. She is currently fed via a gastrostomy.
Case J.L. A 23-year-old white man presented with recurrent episodes of intestinal obstruction. He had no family history of autoimmune disease, malignancy or gastrointestinal pathology, was a smoker, and drank moderately. He had been well until the age of 5 years when he developed type 1 diabetes mellitus. At the age of 11 years, he began to experience episodes of abdominal pain associated with apparent constipation that were often followed by bouts of diarrhea. These episodes became progressively more frequent. At the age of 23 years, he was admitted to a hospital with abdominal pain and distention. A laparotomy was performed, and several adhesions were divided. However, over the ensuing year, three further laparotomies were required for intestinal obstruction. No apparent mechanical cause was found, and total parenteral nutrition was commenced. When the patient was stable and established on total parenteral nutrition, urea, creatine, electrolytes, full blood count, erythrocyte sedimentation rate, C-reactive protein, serum electrophoresis, and thyroid function tests were all normal. Circu-
/ 5e1b$$0042
03-14-97 17:19:18
gasa
lating Igs (IgG, IgM, and IgA) and complements (C3 and C4) were normal, whereas the CH50 level was reduced (38 U/ mL). Tests for circulating immune complexes were also within normal limits. Autoantibody screen for antinuclear, antimitochondrial, anti–smooth muscle, anti–gastric parietal cell, antireticulin, and anti–liver-kidney-muscle antibodies was negative. Serum magnesium level was persistently low, and gglutamyltransferase and alkaline phosphatase levels were increased. Finally, an ileostomy was fashioned with good symptomatic relief. Six months later the ileostomy was closed, a subtotal colectomy was performed, and the ileum was anastomosed to the sigmoid colon. Total parenteral nutrition was discontinued, and a course of prednisolone (10 mg/day) was administered for a trial period of 6 weeks. On withdrawal of steroids, he had further episodes of intestinal obstruction. Since restarting prednisolone (10 mg/day), he is well but still requires parenteral nutrition 3 days per week. No peripheral neuropathy was noted at any time.
Results Intestinal Biopsy Specimens Full-thickness intestinal samples from F.H. and J.L. were fixed and processed routinely into paraffin wax. Sections were stained with H&E and immunostained for white blood cells using the avidin-biotin-peroxidase complex method.6 The white cell markers (Dako Ltd., High Wycombe, England) included antibodies for T lymphocytes (CD4 and CD3), monocytes and macrophages (CD68), neutrophils (CD15), activated white cells (CD45 and CD30), and B lymphocytes (CD20). From patient F.H. snap-frozen full-thickness ileum resected 2 years after the previous sample was also available and was used for assessment of acetyl cholinesterase activity and immunostaining for neurofilament triplet protein using neurofilament antibody clones NN18, NE14, BF10, NR4, N52 (Sigma Chemical Co., Poole, Dorset, England), RT97 (gift from Professor Brian Anderton, Institute of Psychiatry, University of London, London, England), SM132 (Sternberger, Baltimore, MD), and neural cell adhesion molecule (Eric 1; gift from Professor Frank Walsh, Guy’s Hospital, London, England). The large bowel resected at the time of the pullthrough operation from patient F.H. was aganglionic, but in the samples of ileum, a few degenerate neurons surrounded by mononuclear cells were observed. Although full-thickness ileum from patient J.L. appeared unremarkable, sections of the resected colon and appendix showed marked myenteric ganglionitis with degeneration and loss of neurons. The inflammatory infiltrate in both patients consisted of mononuclear cells with immunoreactivity mostly associated with CD4- and CD3-positive T lymphocytes (Figure 1) although a few B cells WBS-Gastro
1368 SMITH ET AL.
GASTROENTEROLOGY Vol. 112, No. 4
patient F.H. showed strong mainly perikaryal but also weak nuclear immunostaining (IgG) with enteric neurons at a dilution of 1:8000 (Figure 2). A similar pattern was observed with serum from patient J.L. up to a dilution of 1:5000. No IgM immunoreactivity was detected with sera from any of the patients. Serological Testing for Circulating Autoantibodies Against Central Nervous System Neurons
Figure 1. Section of intestine from patient J.L. immunostained for a pan T-lymphocyte marker (CD3) showing that the inflammatory infiltrate predominantly consists of T cells (original magnification 1601).
were also observed. From patient F.H., the progression of the disease process could be assessed in a further ileal sample 2 years from the first specimen of ileum. The lymphocytic infiltrate and neurons had disappeared from the previously inflamed myenteric plexus that had originally contained degenerate neurons. The whole bowel had become denervated with no neurofilament immunostaining in the myenteric plexus and no acetyl cholinesterase-positive nerve fibers in the circular muscle. However, there was evidence of glial elements that were recognized by immunoreactivity to neural cell adhesion molecule. Serological Testing for Autoantibodies Against Enteric Neurons Sera were obtained from F.H. and J.L. together with sera from control subjects. The control sera consisted of sera from a healthy adult and from 14 control patients aged 3 months to 46 years, and these were studied for the presence of circulating antibodies against enteric neurons. The control patients had inflammatory bowel disease (n Å 2), aganglionosis (n Å 1), ileal stricture (n Å 1), esophageal achalasia (n Å 1), hypoganglionosis (n Å 1), intestinal myopathy (n Å 1), unspecified functional intestinal obstruction (n Å 3), juvenile Batten’s disease (n Å 1), rectal ganglionitis (n Å 1), mucosal ulceration (n Å 1), and unspecified functional intestinal obstruction with severe abdominal pain (n Å 1). Sera were serially diluted and incubated with cryostat sections of neonatal piglet small intestine. The incubation with the primary serum was followed by a polyclonal anti-human IgG and IgM antibodies (Dako Ltd.). The avidin-biotin-peroxidase complex technique6 was used to detect the presence of the antibody. No immunoreactivity with enteric neurons was observed with sera from the control patients. Serum from / 5e1b$$0042
03-14-97 17:19:18
gasa
Sera from F.H. and J.L. were tested together with sera from 6 of the control patients with enteric neuromuscular disease described above in random sequence together with other sera sent for diagnostic screening for paraneoplastic antineuronal antibodies (1500 samples/ year). The sera were serially diluted and applied to cryostat sections of human cerebellum, cerebral cortex, and brain stem. The bound antibody was visualized by the indirect immunoperoxidase technique.7 Positive control sera from 3 patients with paraneoplastic neurological syndrome known to contain antineuronal antibodies (antiHu, anti-Yo, and anti-Ri), and a known negative control serum were run in parallel. Of the eight sera from patients with enteric neuromuscular disease tested, only sera from F.H. and J.L. produced staining of the cytoplasm as well as the nuclei of all neurons in sections of human cerebellum, frontal cortex, and brain stem (Figure 3). The nuclei of glia and vascular endothelium were not stained, and the pattern was similar to that observed with control sera with antinuclear neuronal antibodies type I from patients with small cell lung carcinoma and sensory encephalopathy. The sera from the 6 control patients with enteric neuromuscular disease and the serum from the healthy adult produced no staining of brain sections at a dilution of 1:100. The end point dilution determined by immuno-
Figure 2. Serum (1:1280 dilution) from patient J.L. applied to a section of normal neonatal piglet small intestine showing strongly positive myenteric neurons. Note that the immunoreactivity can be mainly seen in the neuronal perikarya (original magnification 4001).
WBS-Gastro
April 1997
ACQUIRED AGANGLIONOSIS 1369
Figure 3. Indirect immunoperoxidase of human brain. Sections of the brain stem treated with (A) normal control human serum (1:200 dilution) treated with (B) serum from patient F.H. (1:400 dilution). Sections of frontal cortex treated with (C) serum from patient F.H. (1:400 dilution), (D) serum from patient J.L. (1:400 dilution), and (E) with a positive control serum from a patient with anti-Hu antibodies and small cell lung carcinoma (1:500 dilution) and counterstained with Mayer’s hematoxylin. Notice nuclei of cell in the blood vessels (small arrow) and glia stained with hematoxylin in E and their absence in B, C, and D (bar Å 25 mm in A and B and 50 mm in C, D, and E).
peroxidase was 1:1600 and 1:4000 with cerebral cortex and cerebellum, respectively (F.H.), and 1:8000 and 1:2000 (J.L.). Western Blotting Antigen specificity was determined by Western blotting7 against human cerebral cortex and cerebellum, rat forebrain, and small cell lung carcinoma cell line NCl-H69. Control sera containing anti-Hu and anti-Ri as well as normal negative sera were run as primaries in parallel. Only sera from patients F.H. and J.L. showed any binding to protein on Western blotting of brain tissue. The sera reacted with protein bands of apparent molecular weights of 38 and 45 kilodaltons in the brain and with similar proteins of approximately 38–39 kilodaltons in / 5e1b$$0042
03-14-97 17:19:18
gasa
the small cell lung carcinoma cell line NCl-H69, as did the control anti-Hu serum (Figure 4).
Discussion Aganglionosis occurs in 1 of 4500 live births and is diagnosed most commonly in the neonatal period. However, in about 7%–8% of patients, aganglionosis is not recognized until later in life.8 We describe 2 patients with episodes of pseudoobstruction caused by an acquired progressive aganglionosis as a result of a severe T-cell– mediated inflammatory disorder involving the enteric nerve plexuses and associated with a circulating IgGclass autoantibody directed against enteric neurons. At the aganglionosis stage, the appearances in H&E-stained sections could not be distinguished from classical Hirschsprung’s disease. These patients together with 3 subjects WBS-Gastro
1370 SMITH ET AL.
GASTROENTEROLOGY Vol. 112, No. 4
Figure 4. Western blot against proteins of human cerebral cortex and small cell lung carcinoma. Strips contain proteins from small cell lung carcinoma and human cerebral cortex in lanes 1 and 2. Lane 1, serum from patient F.H.; lane 2, serum from positive control patient with anti-Hu antibodies and small cell lung carcinoma; lane 3, serum from patient J.L.; and lane 4, serum from normal control (first lane standards). All sera were at 1:1000 dilution. kD, Kilodalton.
described in a preliminary communication by DeGiorgio et al.9 show that at least some patients who present late with aganglionosis could have acquired it rather than it being a congenital disease. As in patients with paraneoplastic syndrome,2,10,11 sera from our patients contained high titers of an IgG-class circulating antibody against nerve cells. In the paraneoplastic syndrome, the antineuronal antibodies have been shown to be of variable antigenic specificity and encompass proteins such as anti–neuronal nuclear antibodies, which include anti-Hu (HuC, HuD, and HelN1) and anti-Ri proteins12 as well as an anti-Purkinje cell cytoplasmic antibody (anti-Yo).11 The immunostaining with the sera from our patients was mostly cytoplasmic with only a weak nuclear component in the enteric nerve cells, and this was different from the anti-Hu–associated pattern in which the nuclear staining predominates.11 However, the results of the Western blots and the staining pattern in the brain sections do not allow distinction from the anti-Hu proteins found in paraneoplastic neurological syndrome. It is generally accepted that the presence of autoantibodies directed against neurons is an indication of a paraneoplastic syndrome often with only a small neoplastic lesion, which aberrantly expresses neuron-specific antigens.12 However, in patients with sensory encephalopathy and anti-Hu antibodies, often no tumor is discovered even at autopsy.13 We found no evidence of neoplasia in our patients, despite extensive investigations during a period of up to 8 years. It has been suggested that the immune response is associated with a complete tumor remission.12 However, it is also possible that the antibodies could arise through mechanisms not associated with malignancy. Parasitic infections, such as trypanosomiasis / 5e1b$$0042
03-14-97 17:19:18
gasa
(Chagas’ disease), have been postulated to elicit an autoimmune response resulting in neuronal degeneration producing esophageal achalasia and megacolon.14 Interestingly, no circulating enteric neuronal antibodies were found in our achalasia control. The autoimmune injury in our patients appeared clinically to be limited to the enteric nervous system, although the antibody is capable of binding in vitro to central neurons. We postulate that the apparent absence of central nervous system symptoms is caused by the fact that the symptoms are the result of local T-cell–mediated responses rather than the production of an antibody, with antibody production being another consequence of the T-cell response. An alternative explanation could be that quantitative differences in epitopes in the central nervous system, dorsal root ganglia, and enteric nervous system are important because it is thought that there is a critical density of epitopes for the antibody to cause disease (D. Vergani, personal communication, June 17, 1996). Aganglionosis could be caused by other mechanisms such as apoptosis.15 However, we did not observe any features to suggest programmed cell death or other explanations for acquired aganglionosis. Indeed, on treatment with prednisolone, the progression of the disease seemed to be arrested; a relapse was observed at withdrawal; and at reintroduction of prednisolone, there was clear-cut clinical improvement providing further support in addition to the immunologic and histological features that the disease is caused by a T-cell–mediated immune response. As well as the loss of purposeful intestinal contractile activity, F.H. experienced severe abdominal pain. A variety of mechanisms could have induced the pain, including hyperactivity in primary afferent or central nervous system nociceptive neurons, loss of central inhibitory connections, and increased activity in sympathetic efferents.16,17 The lack of sustained effect of celiac plexus ablation suggests that the latter was not important in F.H.’s pain. However, the denervation of the gut could have resulted in hypersensitivity of the intact parts of the visceral nociceptive pathways probably at a spinal or central level. This notion is supported by the response to the 5-hydroxytryptamine antagonist ondansetron because it is known that 5-hydroxytryptamine is important in visceral nociception.18,19 We have discussed the clinical course, diagnosis, and treatment of 2 patients with acquired intestinal aganglionosis. In these patients, there appeared to be a T-cell– mediated inflammatory response directed against enteric neurons that also resulted in the production of circulating neuronal antibodies. In addition, there was an absence of other neurological or neoplastic disease. We suggest WBS-Gastro
April 1997
ACQUIRED AGANGLIONOSIS 1371
that in patients with late-onset persistent constipation associated with myenteric ganglionitis, the presence of circulating enteric neuronal antibodies should be sought.
References 1. Valdes-Dapena M. Megacolon in the adult. In: Valdes-Dapena AM, Stein GN, eds. Morphological pathology of the alimentary canal. Philadelphia: Saunders, 1970:659–660. 2. Lennon VA, Sas DF, Busk MF, Scheithauer B, Malagelada J-R, Camilleri M, Miller LJ. Enteric neuronal autoantibodies in pseudoobstruction with small-cell lung carcinoma. Gastroenterology 1991;100:137–142. 3. Gerl A, Storck M, Schalhorn A, Muller-Hocker J, Jauch KW, Schildberg FW, Wilmanns W. Paraneoplastic chronic intestinal pseudoobstruction as a rare complication of bronchial carcinoid. Gut 1992;33:1000–1003. 4. Horoupian DS, Kim Y. Encephalomyeloneuropathy with ganglionitis of the myenteric plexus in the absence of cancer. Ann Neurol 1982;6:628–632. 5. Devane SP, Ravelli AM, Bisset WM, Smith VV, Lake BD, Milla PJ. Gastric antral dysrhythmias in children with chronic idiopathic intestinal pseudo-obstruction. Gut 1992;33:1477–1481. 6. Hsu SM, Raine L, Fanger H. Use of avidin biotin peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabled antibody (PAP) procedures. J Histochem Cytochem 1981;29:577–580. 7. Ball JA, Warner T, Reid P, Howard RS, Gregson NA, Rossor MN. Central alveolar hypoventilation associated with paraneoplastic brain-stem encephalitis and anti-Hu antibodies. J Neurol 1994; 241:561–566. 8. Smith VV, Milla PJ. Aganglionosis may not only be due to Hirschsprung’s disease (abstr). Gut 1993;34:S12. 9. DeGeorgio R, Bassotti G, Stanghellini V, Barbara G, Tazzari P-L, Santini D, Suozzi A, Germani U, Velio P, Minni F, Morelli A, Corinaldesi R. Clinical morpho-functional and immunological features of idiopathic myenteric ganglionitis (abstr). Gastroenterology 1996;110:A655. 10. Jean WC, Dalmau J, Ho A, Posner JB. Analysis of the IgG subclass distribution and inflammatory infiltrates in patients with anti-hu– associated paraneoplastic encephalomyelitis. Neurology 1994; 44:140–147.
/ 5e1b$$0042
03-14-97 17:19:18
gasa
11. Altermatt HJ, Rodriguez M, Scheithauer BW, Lennon VA. Paraneoplastic anti-Purkinje and type 1 antineuronal nuclear autoantibodies bind selectively to central, peripheral and autonomic nervous system cells. Lab Invest 1991;65:412–420. 12. Rosenblum MK. Paraneoplasia and autoimmunologic injury of the nervous system: the anti-hu syndrome. Brain Pathol 1993; 3:199–212. 13. Dalman J, Graus F, Rosenblum MK, Posner JB. Anti-Hu associated paraneoplastic encephalomyelitis/sensory neuropathy. A clinical study of 71 patients. Medicine 1992;71:59–72. 14. Schuffler MD. Neuromuscular abnormalities of the small and large intestine. In: Whitehead R, ed. Gastrointestinal and oesophageal pathology. New York: Churchill Livingstone, 1989:329– 353. 15. Singaram C, Koch J, Gaumnitz EA, Goyal RK, SenGupta A. Nature of neuronal loss in human achalasia (abstr). Gastroenterology 1996;110:A259. 16. Cervero F, Janig W. Visceral nociceptors: a new world order? TINS 1992;15:374–378. 17. Fields HL. Painful dysfunction of the nervous system. In: Fields HL, ed. Pain. New York: McGraw–Hill, 1987:133–145. 18. Randich A, Gebhart GF. Vagal afferent modulation of nociception. Brain Res Rev 1992;17:77–99. 19. Thurston CL, Randich A. Quantitative characterization and spinal substrates of antinociception produced by electrical stimulation of the subdiaphragmatic vagus in rats. Pain 1991;44:201–210. Received July 17, 1996. Accepted January 6, 1997. Address requests for reprints: Virpi V. Smith, Ph.D., Department of Histopathology, Camelia Botnar Laboratories, Great Ormond Street Hospital for Children, Great Ormond Street, London WC1N 3JH, England. Fax: (44) 1-71-813-1170. The authors thank the following who made important contributions to the diagnosis and management of these 2 patients: Dr. J. Hopkinson for making the original histological diagnosis of myenteric ganglionitis in case J.H. and for allowing us to characterize the lymphocytic infiltrate in the intestine; Dr. I. Filipe for the original paraffin blocks from case F.H., and B. D. Lake and Dr. D. G. D. Wight for noting the ganglionitis; R. Medcalf, Dr. A Tozzi, and H. Kaur for their excellent technical assistance in the screening of the sera; Dyanne Rampling for immunostaining the lymphocytes; and Dr. Jia Newcome for normal human brain tissue from the Multiple Sclerosis Brain Bank, Institute of Neurology.
WBS-Gastro