Journal of the Neurological Sciences 163 (1999) 47–52
TNFa, IFNg and IL-2 mRNA expression in CIDP sural nerve biopsies a, b a Emily K. Mathey *, John D. Pollard , Patricia J. Armati a
b
School of Biological Sciences, A08, University of Sydney, Sydney, NSW, Australia, 2006 Institute of Clinical Neurosciences, Blackburn Building, D06, University of Sydney, Sydney, NSW, Australia, 2006 Received 20 August 1998; received in revised form 8 December 1998; accepted 8 December 1998
Abstract Proinflammatory cytokines contribute to the regulation of the disease process in inflammatory neuropathies. Cellular localisation of cytokine expression in CIDP nerve biopsies should provide further insight into the pathogenic mechanisms of the disease and the individual cells involved. In this study in situ hybridisation was used to determine the exact localisation and identity of cells that express TNFa, IFNg and IL-2 mRNA within the CIDP nerve. Paraffin embedded and frozen sural nerve biopsies from three acute phase CIDP patients were used for the study. Sections of these samples were probed with digoxigenin labelled oligoprobes for TNFa, IFNg and IL-2. The results demonstrate localisation of cytokine expression to the inner rim of the perineurium, epineurial and endoneurial blood vessels and infiltrating inflammatory cells. In addition strong staining for TNFa mRNA was widespread in the endoneurium in areas consistent with / suggestive of Schwann cells. Expression of cytokines in the perineurium and endoneurial blood vessels may have pertinent implications with respect to the breakdown of the blood nerve barrier associated with CIDP. In the very least the potential for an immunomodulatory role may be ascribed to these cells. 1999 Elsevier Science B.V. All rights reserved. Keywords: CIDP; Cytokines; TNFa; IFNg; IL-2; In situ hybridisation
1. Introduction Chronic inflammatory demyelinating polyneuropathy (CIDP) is widely regarded as an autoimmune disorder of the peripheral nervous system (PNS) in which myelin or Schwann cell antigens are the targets of immune attack [3]. It shares many features with Guillain Barre´ Syndrome (GBS) and is considered to be a chronic form of this uniphasic disorder. The pathogenesis of these inflammatory demyelinating neuropathies (IDN) remains unknown but they are clearly immune mediated, and both humoral and cellular mechanisms can be defined [10]. For both diseases, experimental autoimmune neuritis (EAN) is the prototypic autoimmune animal model. In both GBS and CIDP there is accumulating evidence of an important role for antibody and complement and many different target epitopes have been implicated [10]. Several key elements *Corresponding author. Tel.: 161-2-9351-2437; fax: 161-2-93514119; e-mail:
[email protected]
of the complex cellular immune process have been defined, including the demonstration that macrophages are the final common pathway in myelin removal [7,21] and activated T cells initiate breakdown of the blood nerve barrier (BNB) [19,29]. This latter event is necessary before antibody and complement can gain access to the endoneurium. The activity of immune cells is regulated by a complex network of cytokines, several of which are implicated in the initiation and acute phases of immune mediated demyelination [10], in particular interferon gamma (IFNg), tumour necrosis factor a (TNFa) and interleukin-2 (IL-2). Both TNFa and IFNg enhance vascular permeability [4–6,30] and therefore may be of fundamental importance in allowing the influx of antibody and complement into the nerve. The importance of both cytokines is further demonstrated by the fact that antibodies to these molecules ameliorate EAN and administration of IFNg exacerbates the disease [10]. The role of cytokines in IDN has primarily been studied at the tissue level in animals with EAN. However, studies in patients
0022-510X / 99 / $ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S0022-510X( 99 )00009-X
48
E.K. Mathey et al. / Journal of the Neurological Sciences 163 (1999) 47 – 52
have shown increased serum and cerebrospinal fluid levels of TNFa in the acute phase of GBS and the levels correlate with disease course [2,4,28]. In this study we have examined nerve biopsy tissue from CIDP patients biopsied in the acute phase by in situ hybridisation (ISH) to identify mRNAs for TNFa, IFNg and IL-2. If these molecules do play an important role such as increasing BNB permeability, they should be demonstrable at the relevant sites.
2. Materials and methods
2.1. CIDP tissue Biopsied sural nerves were selected from the Neurology archives, Department of Medicine, University of Sydney. Three patients fulfilling the clinical, electrophysiological and histological criteria for diagnosis of CIDP were selected. These nerves were biopsied during an acute phase of the disease. Previous studies of these nerves have shown the classical changes of inflammatory demyelination [3]. Experiments used material from archival paraffin blocks and where available, frozen tissue. Tissue was fixed in picric acid saline (2.5% glacial acetic acid in a saturated solution of picric acid in 0.9% saline) for a minimum of 24 h then paraffin embedded. Paraffin blocks were stored at room temperature until sectioning. For frozen tissue, the fresh tissue was snap frozen in iso-pentane and stored in liquid nitrogen until required. Five-mm paraffin sections and 8-mm cryosections were collected onto slides coated with 2% 3-aminopropyl-triethoxysilane (Sigma) made up in dry acetone.
2.2. Control tissue Hybridisation with the TNFa, IFNg or IL-2 probes was also conducted on sural nerve biopsy material from patients with CMT1A as well as on normal nerves.
2.3. In situ hybridisation ( ISH) 2.3.1. Pretreatment of tissue Paraffin sections were dewaxed by immersion in dipentene followed by rehydration in graded alcohols, 100%, 100%, 95%, 70% and 50%. Sections were permeabilised with 0.3% Triton-X100 in PBS for 15 min, washed and then incubated in Type XIL bacterial protease (1 mg / ml in PBS) (Sigma) at 228C. Protease was washed off by immersion in PBS for 5 min. Sections were postfixed with 4% paraformaldehyde in PBS for 5 min at 228C. Following fixation the tissue was acetylated with 0.25% acetic anhydride (Sigma) in 0.1 M triethanolamine (Sigma) followed by washing with distilled water. The sections were then incubated in hybridisation buffer (63SSC, 13 Denhardt’s solution, 0.1 mM ATP, 2 mM sodium pyro-
phosphate and 200 mg / ml denatured herring sperm DNA) (Sigma) for 30 min at 378C.
2.3.2. Hybridisation Hybridisation was performed in hybridisation buffer containing 30 nmol of digoxigenin-labelled hIFNg (R&D Systems Cat. No. BPR 216) or hTNFa (R&D Systems Cat. No. BPR 49) oligonucleotide probe at 378C for 18 h. After hybridisation, sections were washed in 63SSC at 228C for 10 min, 23SSC and 0.23SSC twice each for 5 min at 378C. Sections were then immersed in 23SSC for 1 min. 2.3.3. Digoxigenin detection Before detection sections were immersed in buffer 1 (100 mM Tris–HCl / 150 mM NaCl at pH 7.5) at 228C prior to an incubation in buffer 1 containing 2% normal sheep serum, 3% BSA and 0.3% Triton X-100 for 1 h. Sections were then incubated in sheep anti-digoxigenin (Boehringer Mannheim) diluted 1:500 with buffer 1 containing 1% normal sheep serum and 0.3% Tritron X-100 for 3 h. Unbound conjugate was removed by washing in buffer 1 for 15 min then in buffer 2 (100 mM Tris–HCl, 100 mM NaCl, 50 mM MgCl 2 at pH 9.5) for 15 min at 228C. NBT / BCIP stock solution (Boehringer Mannheim) (1:50 dilution in buffer 2) was applied to each section and incubated in the dark. The development of the reaction product was monitored by light microscopy over 7–12 h. The reaction was terminated by immersion in buffer 3 (10 mM Tris–HCl, 1 mM EDTA at pH 8.0) after which the sections were mounted using Aquamount. Slides were viewed and photographed with a Zeiss Photomicroscope using Kodak Ektachrome colour reversal film at 100 ASA. 2.4. Immunohistology Replicate sections were incubated in mouse monoclonal antibodies to MHC class II molecules (Coulter Immunology, FL, USA) diluted 1:80 (in PBS containing 5% fetal calf serum) for 30 min. Sections were washed then incubated in peroxidase-conjugated F(ab9) 2 fragment of sheep antimouse IgG (Sigma) diluted 1:50 for 30 min. After further washing, the reaction product was visualised with 0.04% diaminobenzidine.
3. Results
3.1. In situ hybridisation TNFa, IFNg and IL-2 were each expressed consistently in all sections of the CIDP biopsies within cells with the size and morphology consistent with that of lymphocytes and macrophages. These cytokine mRNA expressing cells
E.K. Mathey et al. / Journal of the Neurological Sciences 163 (1999) 47 – 52
were distributed throughout the endoneurium as well as in and around endoneurial and epineurial blood vessels. Representative photomicrographs taken from the frozen sections are shown in Figs. 1 and 2. Similar staining patterns were seen in the paraffin sections, however, the staining was less intense. Perineurial cells and cells infiltrating the perineurium were also positive for TNFa, IFNg and IL-2 mRNA. The message for IFNg and TNFa was more abundant than that for IL-2. mRNA for both TNFa and IFNg was abundantly present throughout the endoneurium but also in association with epineurial and endoneurial blood vessels (Fig. 1A,B and 2A). In addition strong staining for TNFa mRNA was widespread in the endoneurium in areas consistent with / suggestive of Schwann cells. mRNA for TNFa was particularly dense in the region of the perineurium. The negative controls were (i) omission of the cytokine probe (Fig. 1D) and (ii) hybridisations using normal nerve. Hybridisation on sections of normal nerve produced no specific staining with TNFa, IFNg or IL-2 probes (Fig. 1C and 2B). The positive control, digoxigenin-labelled poly-
49
d(T) DNA oligoprobe, demonstrated that the biopsy tissue mRNA was both intact and detectable on infiltrating inflammatory cells, cells in the vasculature and endoneurial cells which were strongly stained (not shown).
3.2. Immunohistology Major histocompatibility complex (MHC) class II molecules were markedly upregulated. MHC class II staining was evident throughout the endoneurium, on endothelial cells and cells surrounding epineurial vessels and on cells within the perineurium (Fig. 2D).
4. Discussion We found consistent, well localised and specific staining of the proinflammatory cytokine mRNAs for TNFa, IFNg and IL-2 in all CIDP nerve sections. This expression was predominantly found in infiltrating inflammatory cells, which were evident within the endoneurium, and surround-
Fig. 1. Photomicrograph of (A) frozen sections of CIDP nerve probed for TNFa mRNA showing staining of the perineurium, cells within the endoneurium, as well as the epineurial and endoneurial blood vessels. (B) High power of same nerve showing TNFa expressing cells around endoneurial blood vessels and on cells within the endoneurium. Section of normal nerve (C) shows no specific staining when probed for TNFa and (D) no staining when the probe was omitted from CIDP nerve sections. Bars510 mm; v, blood vessel; p, perineurium.
50
E.K. Mathey et al. / Journal of the Neurological Sciences 163 (1999) 47 – 52
Fig. 2. Photomicrograph of (A) frozen section of CIDP probed for IFNg mRNA showing staining of infiltrating cells, perineurium and epineurial blood vessels. No specific staining was obtained when normal nerve was probed for IFNg mRNA (B). Hybridisation with the IL-2 probe on CIDP nerve sections produced staining of infiltrating cells within the endoneurium and epineurium (C). Staining for MHC class II on CIDP nerve sections shows upregulated expression within the endoneurium, on perineurial cells as well as on cells around epineurial blood vessels (D). Bars510 mm; v, blood vessel; p, perineurium.
ing blood vessels. TNFa and IFNg were also prominent within the perineurium. Previous studies of CIDP biopsies [13,26,27], on these same nerves [18] and on others, have found that these infiltrating cells were macrophages with fewer CD4 and CD8 lymphocytes. In addition to infiltrating mononuclear cells and Schwann cells, mRNA for TNFa and IFNg was found within perineurial cells. Recent studies from our laboratory and others [9,20] have emphasised the important role of the BNB in IDN. We have shown that demyelinating antibody such as antigalactocerebroside, given systemically to rats can only access the endoneurium if the BNB is made more permeable [19,29,30], and that such changes are most effectively induced by activated T cells and cytokines such as TNFa [30]. The role of the BNB is further emphasised by recent studies of the pathology of acute motor axonal neuropathy (AMAN) formerly called the Chinese paralytic syndrome. Immunopathological studies by Hafer-Macko et al. [8] have shown that changes associated with GM-1 antibody and complement deposition are confined to the motor terminals since, in the absence of inflammatory infiltrates,
it is only in this region, or in the proximal nerve root zone that the BNB is deficient. In IDN, however, the production of TNFa, IFNg and IL-2 by infiltrating mononuclear cells and resident perivascular cells can promote increased permeability of the BNB [15] allowing the entry of antibodies to the endoneurium which may result in demyelination [29]. As Schwann cells are able to synthesise TNFa mRNA in vitro after stimulation with IFNg [14] the dense staining for TNFa mRNA within the endoneurium may indicate that Schwann cells are a source of some intrafasicular TNFa. This would further indicate an active immunological rather than a passive role for Schwann cells in IDNs. Further, both TNFa and IFNg can induce inflammatory mediators or effectors of tissue destruction such as matrix metalloproteinases (MMPs). MMPs are proteolytic enzymes strongly implicated in BBB and BNB breakdown in MS and inflammatory demyelination neuropathy respectively, via their proteolytic activity on the basal lamina components, essential elements of the barrier [11,23–25]. As the BNB is comprised of both endothelial and
E.K. Mathey et al. / Journal of the Neurological Sciences 163 (1999) 47 – 52
perineurial components it is therefore of considerable interest that the same suite of cytokines expressed by endoneurial and epineurial perivascular cells were also expressed by perineurial cells. Hence the finding of this study that cytokines associated with BNB permeability changes are actually produced in endoneurial and perineurial sites in patients with CIDP is an important one and consistent with current views on disease pathogenesis. There is accumulating evidence for a role of antibody in IDN (reviewed by Hartung, 1996). Both the response of patients to plasma exchange therapy and the recent burgeoning literature on glycolipid antibodies suggest an important role for humoral factors in these disorders. Except for those regions with a BNB deficiency, leakiness of the BNB and hence antibody and complement access will depend upon local inflammatory cell accumulation and activation and cytokine production. Hence the findings presented in this paper are highly relevant not only to GBS and CIDP which show a multifocal pathology characterised by inflammation and demyelination, but also to conditions such as multifocal motor neuropathy with conduction block, in which focal block may be associated with the effect of anti-GM-1 antibodies on Na 1 and K 1 channels of the motor axons [31]. It is also of interest that Prineas and McLeod (1976) noted an increase in the subperineurial space in CIDP nerve, suggestive of an accumulation of serum proteins possibly due to perineurial malfunction [22]. A similar increase in subperineurial space was a consistent feature of our study and a strong signal for TNFa production by perineurial cells was evident in each patient. As well, in EAN induced by autoreactive T cell lines, the perineurium also appears similarly damaged [20]. Thus expression of proinflammatory cytokines by perineurial cells appears to be related to increased permeability of this region also. As well as TNFa and IFNg mRNA, we found strong expression of IL-2 mRNA. Such IL-2 mRNA is consistent with acute disease exacerbation and supports the concept of ongoing or recurring local reactivation of T cells within nerve fascicles [12]. It is well documented for example that T cell activation and clonal proliferation are mediated by the autocrine and paracrine effects of IL-2 [16]. Furthermore, IFNg has been shown to augment antigen presentation and reactivation of CD4 T cells by Schwann cells in culture (Lilje and Armati, submitted). Thus localised TNFa, IFNg and IL-2 production by infiltrating inflammatory and resident perivascular cells could well promote T cell reactivation in inflamed nerve. The combined effects of T cell restimulation and heightened antigen presentation could therefore exacerbate and amplify the disease process. TNFa and IFNg expressing cells within the endoneurium can also upregulate the expression of MHC class II or ICAM-1 molecules on Schwann cells in vitro [1] and on Schwann cells in CIDP and GBS nerve [17,18]. The marked upregulation of MHC class I and II molecules within the endoneurium was clearly demonstrated in this
51
study and was presumably due to the abundant IFNg and TNFa present [17]. Such upregulation on resident or infiltrating macrophages as well as non-professional antigen presenting cells such as Schwann cells and perineurial and endoneurial fibroblasts could in turn restimulate infiltrating T cells and hence contribute to the chronic disease process.
Acknowledgements This study was supported by a grant from the National Health and Medical Research Council of Australia [950133.
References [1] Argall KG, Armati PJ, Pollard JD, Watson E, Bonner J. Interactions between CD41 T-cells and rat Schwann cells in vitro. 1. Antigen presentation by Lewis rat Schwann cells to P2-specific CD41 T-cell lines. J Neuroimmunol 1992;40:1–18. [2] Creange A, Belec L, Clair B, Raphael JC, Gherardi RK. Circulating tumor necrosis factor (TNF)-alpha and soluble TNF-alpha receptors in patients with Guillain-Barre syndrome. J Neuroimmunol 1996;68:95–9. [3] Peripheral neuropathy. in: Dyck PJ, Prineas JW, Pollard JD, Dyck PJ, Thomas PK, Griffin J, Low PA, Poduslo J, (Eds.) Chronic inflammatory demyelinating polyradiculoneuropathy, 3, W.B. Saunders, Philadelphia, 1993, pp. 1498–517. [4] Exley AR, Smith N, Winer JB. Tumour necrosis factor-alpha and other cytokines in Guillain-Barre syndrome. J Neurol Neurosurg Psychiatry 1994;57:1118–20. [5] Gold R, Toyka KV, Hartung HP. Synergistic effect of IFN-gamma and TNF-alpha on expression of immune molecules and antigen presentation by Schwann cells. Cell Immunol 1995;165:65–70. [6] Gold R, Zielasek J, Kiefer R, Toyka KV, Hartung HP. Secretion of nitrite by Schwann cells and its effect on T-cell activation in vitro. Cell Immunol 1996;168:69–77. [7] Griffin JW, George R, Lobato C, Tyor WR, Li CY, Glass JD. Macrophage responses and myelin clearance during Wallerian degeneration: relevance to immune-mediated demyelination. J Neuroimmunol 1992;40:153–66. [8] Hafer-Macko CE, Sheikh KA, Li CY, Ho TW, Cornblath DR, McKhann GM, Asbury AK, Griffin JW. Immune attack on the Schwann cell surface in acute inflammatory demyelinating polyneuropathy. Ann Neurol 1996;39:625–35. [9] Hahn AF, Feasby TE, Wilkie L, Lovgren D. Antigalactocerebroside antibody increases demyelination in adoptive transfer experimental allergic neuritis. Muscle Nerve 1993;16:1174–80. [10] Hartung HP, Kiefer R, Gold R, Toyka KV. Autoimmunity in the peripheral nervous system. Baillieres Clin Neurol 1996;5:1–45. [11] Leppert D, Waubant E, Galardy R, Bunnett NW, Hauser SL. T cell gelatinases mediate basement membrane transmigration in vitro. J Immunol 1995;154:4379–89. [12] Lisak RP, Bealmear B. Upregulation of intercellular adhesion molecule-1 (ICAM-1) on rat Schwann cells in vitro: comparison of interferon-g, tumor necrosis factor-a and interleukin-1. J Peripher Nerv Syst 1997;2:233–43. [13] Matsumuro K, Izumo S, Umehara F, Osame M. Chronic inflammatory demyelinating polyneuropathy: histological and immunopathological studies on biopsied sural nerves. J Neurol Sci 1994;127:170–8.
52
E.K. Mathey et al. / Journal of the Neurological Sciences 163 (1999) 47 – 52
[14] Murwani R, Hodgkinson S, Armati P. Tumor necrosis factor alpha and interleukin-6 mRNA expression in neonatal lewis rat Schwann cells and a neonatal rat Schwann cell line following interferon gamma stimulation. J Neuroimmunol 1996;71:65–71. [15] Ohmori Y, Wyner L, Narumi S, Armstrong D, Stoler M, Hamilton TA. Tumor necrosis factor-alpha induces cell type and tissuespecific expression of chemoattractant cytokines in vivo. Am J Pathol 1993;142:861–70. [16] Paillard X, de Waal Malefijt R, Yssel H, Blanchard D, Chretian I, Abrams J, de Vries J. Simultaneous production of IL-2, IL-4 and IFN-G by activated human CD41 and CD81 T cell clones. J Immunol 1988;141:849–55. [17] Pollard JD, Baverstock J, McLeod JG. Class II antigen expression and inflammatory cells in the Guillain Barre´ syndrome. Ann Neurol 1987;21:337–41. [18] Pollard JD, McCombe PA, Baverstock J, Gatenby PA, McLeod JG. Class II antigen expression and T lymphocyte subsets in chronic inflammatory demyelinating polyneuropathy. J Neuroimmunol 1986;13:123–34. [19] Pollard JD, Westland KW, Harvey GK, Jung S, Bonner J, Spies JM, Toyka KV, Hartung HP. Activated T cells of nonneural specificity open the blood–nerve barrier to circulating antibody. Ann Neurol 1995;37:467–75. [20] Powell HC, Myers RR, Mizisin AP, Olee T, Brostoff SW. Response of the axon and barrier endothelium to experimental allergic neuritis induced by autoreactive T cell lines. Acta Neuropathol (Berl) 1991;82:364–77. [21] Prineas JW. Pathology of inflammatory demyelinating neuropathies. Baillieres Clin Neurol 1994;3:1–24. [22] Prineas JW, McLeod JG. Chronic relapsing polyneuritis. J Neurol Sci 1976;27:427–58.
[23] Redford EJ, Smith KJ, Gregson NA, Davies M, Hughes P, Gearing AJ, Miller K, Hughes RA. A combined inhibitor of matrix metalloproteinase activity and tumour necrosis factor-alpha processing attenuates experimental autoimmune neuritis. Brain 1997;120:1895– 905. [24] Rosenberg GA, Dencoff JE, Correa Jr. N, Reiners M, Ford CC. Effect of steroids on CSF matrix metalloproteinases in multiple sclerosis: relation to blood–brain barrier injury. Neurology 1996;46:1626–32. [25] Rosenberg GA, Estrada EY, Dencoff JE, Stetler-Stevenson WG. Tumor necrosis factor-alpha-induced gelatinase B causes delayed opening of the blood–brain barrier — an expanded therapeutic window. Brain Res 1995;703:151–5. [26] Schmidt B, Toyka KV, Kiefer R, Full J, Hartung HP, Pollard J. Inflammatory infiltrates in sural nerve biopsies in Guillain-Barre syndrome and chronic inflammatory demyelinating neuropathy. Muscle Nerve 1996;19:474–87. [27] Schroder HD, Olsson T, Solders G, Kristensson K, Link H. HLADR-expressing cells and T-lymphocytes in sural nerve biopsies. Muscle Nerve 1988;11:864–70. [28] Sharief MK, McLean B, Thompson EJ. Elevated serum levels of tumor necrosis factor-alpha in Guillain-Barre syndrome. Ann Neurol 1993;33:591–6. [29] Spies J, Bonner JG, Westland KW, Pollard JD. Blood nerve barrier breakdown by activated P2 specific T cells. Brain 1995;118:857–68. [30] Spies JM, Pollard JD, Bonner JG, Westland KW, McLeod JG. Synergy between antibody and P2-reactive T cells in experimental allergic neuritis. J Neuroimmunol 1995;57:77–84. [31] Takigawa T, Yasuda H, Kikkawa R, Shigeta Y, Saida T, Kitasato H. Antibodies against GM1 ganglioside affect K 1 and Na 1 currents in isolated rat myelinated nerve fibers. Ann Neurol 1995;37:436–42.