Selective Demonstration of Mural Nerves in Ganglionic and Aganglionic Colon by Immunohistochemistry for Glucose Transporter-1

Transcription

1 Selective Demonstration of Mural Nerves in Ganglionic and Aganglionic Colon by Immunohistochemistry for Glucose Transporter-1 Prominent Extrinsic Nerve Pattern Staining in Hirschsprung Disease Yutaka Kakita, MD; Kiyohiko Oshiro, MD; D. Sean O Briain, MB, MRCPath; Prem Puri, MS, FRCS, FRCS(Ed) Objective. Hypertrophic nerves have long been considered a histopathologic feature of the aganglionic segment in Hirschsprung disease, but they remain incompletely explained. The purpose of this study was to define the nature and diagnostic importance of hypertrophic nerves in Hirschsprung disease and to clarify their relation to nearby smaller nerve fibers. Methods. We used an immunoperoxidase staining technique to compare the distribution of nerve markers erythrocyte-type glucose transporter (GLUT-1), a marker of perineurium, and nerve growth factor receptor, a marker of both nerve fibers and perineurium in aganglionic tissue (1 resected specimens and rectal biopsies) and control tissue (6 autopsy specimens and 17 rectal biopsies) of children. Results. In control ganglionic tissue, the myenteric and submucosal areas contained only occasional GLUT-1 positive nerves (usually less than 50 m in diameter), but extramural extrinsic (serosal) nerves were invariably positive for GLUT-1. In aganglionic tissue, GLUT-1 positive nerves in the myenteric and submucosal areas were frequent and included both large ( m) and small ( 50 m) diameter nerves. Nerve growth factor receptor positive fibers were frequent in all layers of all tissue studied. In aganglionic bowel, a distinct perineurium could be identified in the largest nerves, but nerve growth factor receptor had poor discrimination for small perineuriumsheathed nerves. Conclusion. Most nerves, of both large and small diameter, in the myenteric and submucosal plexus of aganglionic bowel are GLUT-1 positive. Serosal extrinsic nerves stain identically, supporting the interpretation that the mural nerves are of extrinsic origin. Mural GLUT-1 positive nerves, when they are multiple and especially when they are greater than 50 m in diameter (a figure which may be used as a threshold for hypertrophic nerves), are suggestive of Hirschsprung disease. (Arch Pathol Lab Med. 000;1: ) The erythrocyte-type glucose transporter (glucose transporter-1; GLUT-1) is a member of a family of glucose transporter molecules responsible for the facilitated diffusion of glucose into cells. 1 3 It is expressed predominantly in red blood cell membranes, as well as in endothelial cells involved in maintaining the blood-brain barrier. Malignant tissue may also contain GLUT-1.,5 Peripheral nerves, which like the brain depend heavily on glucose for metabolism, have a blood-nerve barrier in which GLUT-1 also plays a major role, and the perineurium of these fibers has been demonstrated to stain intensely with antibodies to GLUT-1 by immunohistochemical techniques. 6 9 Perineurial staining has been used to assess the presence of tumor invasion of nerves. 10 Hirschsprung disease (HD) is a disorder of intestinal innervation in which the principal histologic features are the absence of ganglion cells and the presence of increased Accepted for publication March 31, 000. From the Children s Research Centre, Our Lady s Hospital for Sick Children (Drs Kakita, Oshiro, and Puri) and St James Hospital (Dr O Briain), Dublin, Ireland. Reprints: Prem Puri, MS, FRCS, FRCS(Ed), Children s Research Centre, Our Lady s Hospital for Sick Children, Crumlin, Dublin 1, Ireland. numbers of hypertrophic nerve bundles in the aganglionic segment of bowel. The origin of these hypertrophic nerve bundles (also termed trunks or fascicles), which predominantly involve areas of the myenteric and submucosal plexuses, has long been a subject of controversy Whole-mount studies of thick sections stained by immunohistochemistry have indicated that the nerve fibers are continuous with extrinsic fibers 11 ; they enter the muscle wall in association with blood vessels, but subsequently become tortuous and irregularly expanded and finally branch to form coarse plexuses. Several immunohistochemical studies have failed to identify specific features to distinguish these fibers from other intramural nerves, 1 1 but recently staining with nerve growth factor receptor (NGFR), a neural marker that stains a large proportion of mural nerves, has shown a distinct pattern of perineurial staining in both the extrinsic nerves of the mesentery and the hypertrophic nerves of the aganglionic area of HD, but not in other mural nerves. 15 This perineurial staining pattern, however, is only identifiable in large nerves, so that the distribution of the hypertrophic nerves cannot be fully traced. As GLUT-1 antibodies enable the perineurium to be identified in both large and small nerve fibers, we studied material from healthy and HD tissues to assess the 131 Arch Pathol Lab Med Vol 1, September 000 Hypertrophic Nerves in Hirschsprung Disease Kakita et al

2 distribution of perineurial-bound nerve fibers and to determine if their demonstration has diagnostic usefulness. MATERIALS AND METHODS Resected specimens of colon were obtained at the time of pullthrough operation from 1 patients with HD, aged 1 days to 7 years; 10 patients had classic rectosigmoid HD, and patients had total aganglionosis of the colon. The extent of the aganglionic and ganglionic zones was mapped by conventional morphology, and in each case tissue was selected from the aganglionic zone as well as the ganglionic zone (at the proximal margin). Tissue was also taken from the most distal ganglionic zone (the transitional zone). Eight sections each, taken from sites extending from the distal rectum to cecum, were selected from the patients with total colonic aganglionosis. Control colon showing no more than mild autolysis was taken at autopsy from 6 patients (aged months to years) in whom there was no evidence of colonic disease, and for each case sections from rectum, sigmoid, and ascending and transverse colon were selected. Suction rectal biopsies taken at and cm above the pectinate line were obtained from 0 children who were being investigated because of constipation or intestinal obstruction. In all cases, fresh tissue was embedded in OCT compound (Miles Inc, Elkhart, Ind) and frozen at 70 C. Immunohistochemical studies were performed using the avidin-biotin complex (ABC) method. Eight-micrometer-thick cryostat sections were mounted on poly-l-lysine coated glass slides and fixed in % paraformaldehyde for 10 minutes at C. The sections were immersed in methanol with 0.3% hydrogen peroxide to block endogenous peroxidase and then incubated with 5% normal rabbit serum (for NGFR staining) or goat serum (for GLUT-1 staining) to minimize nonspecific reaction, followed by rinsing in phosphate-buffered saline. These sections were incubated with anti-ngfr (1:10, Boehringer, Mannheim, Germany) and anti GLUT-1 (1 g/ml, Chemicon, Temecula, Calif, USA) overnight at C. After thorough rinsing in phosphate-buffered saline, biotinylated rabbit anti-mouse immunoglobulin (1:00, Dakopatts, Glostrup, Denmark) for NGFR staining and goat anti-rabbit immunoglobulin (1:00, Dakopatts) for GLUT-1 staining were applied for 60 minutes at room temperature. After rinsing in phosphatebuffered saline, sections were incubated with avidin-biotin-horseradish peroxidase complex preparation (Dakopatts) for 30 minutes at room temperature. Visualization of peroxidase was achieved by 3-3 -diaminobenzidine tetrahydrochloride (Sigma Chemical Co, St Louis, Mo), 5 mg in 100 ml of phosphate-buffered saline containing 0.015% hydrogen peroxide. In addition, formalin-fixed, paraffin-embedded tissue was selected in one case and stained by the ABC immunoperoxidase technique. Staining was unsuccessful in undigested tissue and in slides predigested with trypsin, but was successful with hyaluronidase (5 mg/ml, Sigma) predigestion. Maximum and minimum widths in the short axis of nerves stained with NGFR and GLUT-1 were measured on a Samba image analysis workstation (Alcatel, Grenoble, France). RESULTS GLUT-1 Staining Glucose transporter-1 strongly stained the perineurium but did not stain the nerve fibers within the perineurium, thus giving a ringlike staining pattern (Figures 1 through 5). There was also strong staining of erythrocytes, but these could easily be distinguished from perineurium except in small capillaries, which were sometimes difficult to distinguish from the smallest nerve fibers. Occasional ganglion cells stained weakly. Quantification of fibers showing perineurium staining with GLUT-1 is shown in the Table and the maximum fiber diameter for each group is represented in Figure 6. Figure 1. A, Extrinsic nerves in the mesentery of the ganglionic portion of bowel resected for Hirschsprung disease when stained for glucose transporter-1 (GLUT-1) show a ringlike staining pattern representing perineurial staining; the nerve fibers are negative. Intravascular (arrow) and extravasated erythrocytes also stained positively (original magnification 50). B, An adjacent section stained for nerve growth factor receptor shows staining of both perineurium and nerve fibers (original magnification 50). The perineurium of serosal (or extramural) nerves stained in all cases of resected HD (Figure 1, A) and healthy control (autopsy) colon. A similar pattern of staining was seen in hypertrophic nerves in the myenteric plexus of aganglionic colon (Figure, A). Serosal nerves extending into the myenteric and submucosal plexuses were seen, particularly in aganglionic areas of HD (Figure 3). In one case of rectosigmoid HD in which formalin-fixed, paraffin-embedded tissue was stained following hyaluronidase predigestion, the staining pattern was similar to the corresponding frozen tissue section. Occasional positive ringlike fibers were present in the submucosal and myenteric plexuses in of the 6 control autopsy colon cases but were not identified in the remaining cases. The fibers ranged from 13 to 5 m, but were predominantly small. They were more frequent in the rectum than in the proximal sections (transverse, descending, and sigmoid colon) and more frequent in the submucosal than in the myenteric plexus. In 16 control rectal biopsies that contained ganglion cells and showed a normal acetylcholinesterase staining pattern, no ringlike fibers were present in 6, scanty fibers in 8 ranged in diameter from 13 to 3 m, and in there were moderate numbers of fibers that measured up to 6 and 77 m. In one of these biopsies, Arch Pathol Lab Med Vol 1, September 000 Hypertrophic Nerves in Hirschsprung Disease Kakita et al 1315

3 Figure 3. Extrinsic nerves penetrating to the myenteric layer of the aganglionic area of bowel with Hirschsprung disease are outlined by glucose transporter-1 immunostaining (original magnification 50). Figure. A, The perineurium of a hypertrophic nerve stained by glucose transporter-1 is outlined in this section of the myenteric plexus from the aganglionic portion of a case of Hirschsprung disease (original magnification 00). B, Nerve growth factor receptor stained both nerve fibers and perineurium in an adjacent section. A distinct perineurium can be seen in some areas; in other areas, the staining of the components merges, making distinction difficult (original magnification 00). Figure. ganglion cells lying within the perineurial ring in a rectal biopsy of healthy bowel. This feature raises the possibility of intestinal neuronal dysplasia (glucose transporter-1, original magnification 500). ganglion cells were identified within the perineurial ring (Figure ), a feature that has been reported in intestinal neuronal dysplasia. 16 In rectal biopsies showing HD, positive fibers were prominent, and many fibers (at least 3 in each case) greater than 50 m were readily found in each (Figure 6). In 10 cases of typical rectosigmoid HD, abundant GLUT-1 positive fibers were present in the submucosal and myenteric plexuses of the aganglionic segments, and numerous fibers greater than 50 m were readily found in each (Figure 5, A). Reduced numbers of fibers with a smaller maximum diameter were found in the transitional zones. In the proximal ganglionic zone, the pattern of 8 of the 10 cases (Figure 5, B) was similar to healthy control colon, but in cases (Figure 5, C) there were moderate numbers of fibers up to 75 and 88 m in diameter; these large fibers involved both submucosal and myenteric plexuses. In both of these cases we noted an increase of intramucosal acetylcholinesterase staining in the mucosa, and occasional giant ganglia were found in the submucosa in one. These features are inconclusive but raise the possibility of intestinal neuronal dysplasia. In the cases of total colonic aganglionosis, GLUT-1 fibers were not found in the submucosa, although multiple sections from different regions were examined. However, small numbers of fibers were present in the myenteric plexus; these fibers were small (up to 3 m) in one case, but included larger fibers (up to 7 m) in the second. Fibers greater than 100 m in diameter were specific for aganglionic tissue; they were not seen in the transitional zone, but they occurred in only 6 of 10 resection specimens and in none of the HD rectal biopsies. NGFR Staining In the control colon specimens, we noted strong NGFR staining of nerve fibers and ganglion cells of the submu Arch Pathol Lab Med Vol 1, September 000 Hypertrophic Nerves in Hirschsprung Disease Kakita et al

4 Semiquantitative Assessment of the Number of Glucose Transporter-1 Positive Nerve Fibers in 16 Patients With Hirschsprung Disease and Control Patients Tissue Hirschsprung disease Rectal suction biopsies 10 Rectosigmoid resections Aganglionic Transitional Ganglionic Total colonic aganglionosis Control tissue 16 Rectal suction biopsies 6 Autopsies Rectal tissue Descending colon Fiber Quantity* * Quantification of fibers:, abundant;, moderate numbers;, occasional fibers; and, no fibers. 6 3 cosal and myenteric plexuses and of nerve fibers within the circular and longitudinal muscle layers. Large extrinsic nerves in the serosa of both healthy and HD tissue (Figure 1, B) stained, and in these a distinct perineurium, which was separated from the nerve fibers by a narrow space, also stained strongly. In aganglionic tissue, hypertrophic nerve fibers in the myenteric and submucosal zones stained strongly, and in the larger fibers a distinct perineurium, similar to that seen in the serosal nerves in healthy colon, also stained (Figure, B). Frequent smaller nerve fibers also stained in these zones, and in some of these fibers a distinct perineurium could be detected focally when a space existed between nerve and perineurium. However, in most of the smaller fibers the space between nerve and perineurium was reduced or absent, so that it could not be established whether the peripheral staining represented nerve fibers or perineurium. Comparison of GLUT-1 and NGFR Staining In healthy control bowel and in ganglionic HD (Figure 1), the large extrinsic nerves of the serosa stained both for Figure 5. A, In aganglionic bowel of Hirschsprung disease stained for glucose transporter-1 (GLUT-1), the submucosal area shows a marked increase in the number of fibers staining in a ringlike pattern. Many fibers appear hypertrophic, but there are also frequent small-diameter fibers. B, The submucosal portion of the ganglionic area of bowel with Hirschsprung disease shows very occasional small fibers with a ringlike pattern of perineurial staining for GLUT-1. A similar pattern is seen in healthy controls. C, An unusual pattern in the ganglionic bowel of Hirschsprung disease, seen in cases, was the presence of frequent small-diameter fibers with a ringlike staining pattern in the submucosa (GLUT-1 immunostaining, original magnification 50 [A through C]). Arch Pathol Lab Med Vol 1, September 000 Hypertrophic Nerves in Hirschsprung Disease Kakita et al 1317

5 Figure 6. Maximum diameter of submucosal glucose transporter-1 positive fibers from 10 patients with rectosigmoid resections of Hirschsprung disease (resection) and 1 patients who underwent suction rectal biopsy because of a question of Hirschsprung disease (biopsy). Origins of the tissue specimens are indicated as aganglionic (open circle), transitional zone (open diamond), and ganglionic (filled circle). GLUT-1 (ringlike pattern) and NGFR (both perineurium and nerve stained). In the muscle and submucosal layers, the rare intramural GLUT-1 positive fibers were also NGFR positive (adjacent sections were compared), but large numbers of NGFR-positive fibers failed to stain with GLUT-1. In the aganglionic bowel (Figure ) all the large hypertrophic fibers were NGFR positive and showed a distinct perineurium, but in smaller nerve fibers a distinct perineurium could not be identified. With GLUT-1, the perineurium of the large hypertrophic nerve fibers stained strongly in a ringlike pattern. However, in contrast to NGFR staining, staining was also identified around many of the smaller nerve fibers. The smallest fibers with a distinct perineurium measured 1 m in diameter, in contrast to NGFR staining, for which the smallest fibers in which perineurial staining could be distinguished from nerve staining was 3 m. This finding allowed a population of nerves (ranging in diameter from large to small) with prominent perineurial staining to be readily identified and quantified. These nerve fibers show a staining pattern similar to that of extrinsic nerves and may be interpreted as intramural extensions of the extrinsic nerves. In addition, some NGFR-positive small fibers were found to be GLUT-1 negative and were interpreted as representing intrinsic fibers. COMMENT Hypertrophic nerve fibers have long been recognized as a feature of HD. Evidence that they are of extrinsic origin has been provided by whole-mount sections in which continuity with serosal or mesenteric nerves has been demonstrated and which also demonstrate the extreme tortuosity of the fibers. 11 Immunohistochemical demonstration of NGFR staining of the perineurium of both hypertrophic and extrinsic fibers provides additional evidence. 15 The present study indicates morphologic and functional identity by the finding of GLUT-1 positivity in both extrinsic and hypertrophic fibers. Nerve growth factor receptor stains nerve fibers, including those with no perineurium, in addition to staining perineurium. In the larger fibers a distinct subperineurial space allows the stained perineurium to be distinguished from the stained nerve, but in the small fibers a distinct perineurium may not be identified. In contrast, GLUT-1 stains only perineurium (and not nerve fibers), so that the very smallest perineurium sheathed nerves can be identified easily. This allows the extent of perineurium-lined fibers to be determined in both healthy and abnormally innervated colon. Small numbers of GLUT-1 positive fibers are found in the healthy colon; they are slightly more prominent in the rectum compared to the sigmoid and are more prominent in the submucosal plexus than in the myenteric plexus. These fibers are usually 15 to 50 m in diameter. Glucose transporter-1 positive fibers show a staining pattern identical to that of the extrinsic fibers and probably represent the normal extrinsic fibers extending through the muscle layers to form synapses with submucosal and myenteric ganglion cells. In contrast, in the aganglionic region of HD, GLUT-1 positive fibers are markedly increased in number. This increase includes both small-diameter (15 50 m) fibers and larger fibers; several fibers larger than 50 m were found in all the HD cases, and in many cases frequent fibers of 100 to 150 m were found. A moderate increase in GLUT-1 positive fibers is usually found in the transitional zone, and in occasional cases fibers greater than 50 m are present. In most cases, the proximal ganglionic segment has a distribution of fibers similar in frequency and size to the healthy colon. These findings indicate that the intramural GLUT-1 positive fibers in HD show abnormalities: (1) the largest fibers are markedly increased in size compared to intramural GLUT-1 positive fibers of the healthy ganglionic bowel, and () fibers of all diameters appear increased in number, probably mainly due to increased tortuosity causing increased numbers of transections of individual fibers on the microscopic section; in addition, branching of the large fibers to form the smaller fibers and an absolute increase in fiber numbers may occur. A recent morphometric study 16 indicated that most aganglionic segments contain nerve trunks greater than 0 m in diameter, but that no fibers of that diameter were found in healthy segments of bowel or in normal rectal biopsies. The majority of our observations correspond to this threshold, but we identified small numbers of fibers larger than this in apparently healthy bowel, leading us to select 50 m as our threshold for abnormal fibers. This difference may be due to patient selection or to the use of different antibodies (GLUT-1 or S100). 16 A minority of cases show variations on this pattern. Total colonic aganglionosis, which represents % to 5% of all HD cases, is distinct from other forms of HD; it usually has a normal acetylcholinesterase pattern and lacks hypertrophic nerves. 17 This distinct pattern is supported by the present study, in which submucosal GLUT-1 fibers (whether small or large) were not found, but myenteric fibers were identified in small numbers; in one case fibers were of normal size, but in the second case they were slightly enlarged (up to 7 m). Increased numbers of fibers, including enlarged fibers, were found in the proximal ganglionic bowel in cases of HD. This finding may represent transitional zone extension to the proximal surgical margin. Similarly, increased numbers of fibers, in Arch Pathol Lab Med Vol 1, September 000 Hypertrophic Nerves in Hirschsprung Disease Kakita et al

6 cluding enlarged fibers, were found in rectal biopsies that showed normal ganglion cells and acetylcholine staining. One of these biopsies demonstrated intraperineurial ganglion cells, a feature that had been reported as suggestive of intestinal neuronal dysplasia. 18,19 The finding of increased numbers of fibers of all sizes and of enlarged fibers is unexplained, but raises the possibility of neuronal dysplasia in ganglionic areas of bowel. Glucose transporter-1 is a marker of perineuriumbound nerves, corresponding to extrinsic nerves. The increase in GLUT-1 positive nerves in HD is consistent with intramural extension of extrinsic nerves and includes both abnormally large (hypertrophic) nerves greater than 50 m and smaller nerves. Glucose transporter-1 as a marker of these fibers may be a useful positive indicator in the diagnosis of HD. References 1. Bell GI, Kayano T, Buse JB, et al. Molecular biology of mammalian glucose transporters. Diabetes Care. 1990;13: Thorens B, Charron MJ, Lodish HF. Molecular physiology of glucose transporters. Diabetes Care. 1990;13: Gould WG, Holman GD. The glucose transporter family: structure, function and tissue specific expression. Biochem J. 1993;95: Younes M, Lechago LV, Somoano JR, Mosharaf M, Lechago J. Wide expression of human erythroid glucose transporter Glut-1 in human cancers. Cancer Res. 1996;56: Haber RS, Rathan A, Weiser KR, et al. GLUT-1 glucose transporter expression in colorectal carcinoma: a marker for poor prognosis. Cancer. 1998;83: Gerhart DZ, Drewes LR. Glucose transporters at the blood-nerve barrier are associated with perineurial cells and endoneurial microvessels. Brain Res. 1990; 508: Harik SI, Kalaria RN, Andersson L, et al. Immunocytochemical localisation of erythroid glucose transporter: abundance in tissues with barrier functions. J Neurosci. 1990;10: Magnani P, Cherian PV, Gould GW, et al. Glucose transporters in rat peripheral nerve: paranodal expression of GLUT-1 and GLUT-3. Metabolism. 1996; 5: Muona P, Jaakkola S, Salonen V, Peltonen J. Expression of glucose transporter 1 in adult and developing human peripheral nerve. Diabetologica. 1993; 36: Fogt F, Capodieci P, Loda M. Assessment of perineural invasion by GLUT- 1 immunohistochemistry. Appl Immunohistochem. 1995;3: Tam PKH, Boyd GP. Origin, course, and endings of abnormal enteric nerve fibres in Hirschsprung s disease defined by whole-mount immunohistochemistry. J Pediatr Surg. 1990;5: Mackenzie JM, Dixon MF. An immunocytochemistry study of the enteric neural plexi in Hirschsprung s disease. Histopathology. 1987;11: Robey SS, Kuhajda FP, Yardley JH. Immunoperoxidase stains of ganglion cells and abnormal mucosal nerve proliferations in Hirschsprung s disease. Hum Pathol. 1988;19: Deguchi E, Iwai N, Goto Y, et al. An immunocytochemical study of neurofilament and microtubule associated Tau protein in the enteric innervation of Hirschsprung s disease. J Pediatr Surg. 1993;8: Kabayashi H, O Briain DS, Puri P. Nerve growth factor receptor immunostaining suggests an extrinsic origin for hypertrophic nerves in Hirschsprung s disease. Gut. 199;35: Monforte-Munoz H, Gonzalez-Gomez I, Rowland JM, Landing BH. Increased submucosal nerve trunk caliber in aganglionosis; a positive and objective finding in suction biopsies and segmental resections in Hirschsprung s disease. Arch Pathol Lab Med. 1998;1: Caniano DA, Ormsbee HS, Polito W, et al. Total intestinal aganglionosis. J Pediatr Surg. 1985;0: Munakata D, Morita K, Okabe I, Sueoka H. Clinical and histological studies of neuronal intestinal dysplasia. J Pediatr Surg. 1985;0: Kobayashi H, Hirakawa H, Surana R, O Briain DS, Puri P. Intestinal neuronal dysplasia is a possible cause of persistent bowel symptoms after pullthrough operation for Hirschsprung s disease. J Pediatr Surg. 1995;30: Arch Pathol Lab Med Vol 1, September 000 Hypertrophic Nerves in Hirschsprung Disease Kakita et al 1319