Immunolocalization and Activation of Transcription Factor Nuclear Factor κB in Dysimmune Neuropathies and Familial Amyloidotic Polyneuropathy

Background  Recently, immunoreactivity of transcription factor nuclear factor κB (NF-κB) was found in peripheral nerves from patients with Guillain-Barré syndrome, chronic inflammatory demyelinating polyneuropathy (CIDP), and familial amyloidotic polyneuropathy (FAP), suggesting a role in their pathogenesis.

Objective  To investigate expression and activation of NF-κB in nerve biopsy specimens from patients with peripheral neuropathies of different origins.

Patients  Nerve biopsies from 17 patients (5 with CIDP, 3 with vasculitis, 4 with Charcot-Marie-Tooth disease, and 5 with FAP) and 3 normal sural nerves were studied by immunocytochemistry and Western blot of nuclear extracts for the activated form of NF-κB. Nuclear factor κB DNA-binding activity was studied by electrophoretic mobility shift assay.

Results  Immunobinding for the activated form p65 of NF-κB was found in 2% to 5% of endoneurial vessel walls, in the external myelin of 5% to 10% of fibers, and in a few axons in CIDP specimens. It was also found in 5% to 15% of epineurial and endoneurial vessels in vasculitis specimens and at the level of amyloid deposits in FAP nerves. Nuclear factor κB immunoreactivity was not correlated to type of inflammatory cells, but it often corresponded to the deposition of the terminal complement complex C5b9. Western blot analysis of nuclear extracts showed a single band corresponding to 65 kDa in all affected nerves. Nuclear factor κB DNA-binding activity was revealed by electrophoretic mobility shift assay in specimens from patients with CIDP, vasculitis, and FAP.

Conclusion  Our novel findings suggest a crucial role of NF-κB in inflammatory neuropathies and FAP.

NUCLEAR FACTOR κB (NF-κB) is a ubiquitous transcription factor expressed in different cells, involved in immune and inflammatory responses, and engaged in cell cycle control and differentiation, including apoptosis and oncogenesis.^1-3 The most prevalent form of NF-κB is a heterodimer consisting of the DNA-binding subunits p50 and p65. It is present in the cytoplasm as an inactive complex associated with an inhibitor called IκB. Many external or internal stimuli can activate NF-κB, leading to the dissociation of the NF-κB complex after phosphorylation and degradation of IκB. The freed NF-κB translocates into the nucleus, where it activates inducible genes, such as those encoding for cytokines, chemokines, cell adhesion molecules, growth factors, immunoreceptors, matrix metalloproteinases, acute-phase proteins, oxidative stress–related enzymes, and other transcription factors.⁴

Many new data suggest that NF-κB plays an important role in neurodegenerative disorders. β-Amyloid protein (Aβ) is a potent activator of NF-κB in neuronal cultures.⁵ In Alzheimer disease, NF-κB is increased in close proximity to Aβ-containing plaques,⁵ and neurofibrillary tangles and dystrophic neurites also manifest NF-κB immunoreractivity.⁶ β-Amyloid protein elicits a biphasic response in cultured neurons: a low-dose neurotrophic response and a high-dose neurotoxic response. The activation of NF-κB has been found to be the underlying mechanism of the neuroprotective effect of low-dose Aβ, and its inhibition potentiates Aβ-mediated neuronal apoptosis.⁷ The trophic pathway of NF-κB generally involves expression of protective gene products such as the calcium-binding protein calbindin and antioxidant enzymes.⁸ Neuroprotection by aspirin is mediated through blockade of NF-κB activity.⁹ Such apparently contradictory data could be compatible with a neurotoxic role of NF-κB activation in glial cells and a neuroprotective role in neurons.¹⁰

There are few studies regarding the role of NF-κB in peripheral nervous system disorders. Recently, it has been investigated in acute inflammatory demyelinating polyneuropathy and chronic inflammatory demyelinating polyneuropathy (CIDP).¹¹ Immunoreactivity of the activated form of NF-κB was found in macrophages, suggesting a role of this factor in the genesis of inflammatory demyelination. In familial amyloidotic polyneuropathy (FAP), a transthyretin-related disorder, up-regulation of the p50 NF-κB subunit was found in peripheral nerves by immunocytochemistry.¹² From this and other observations, the authors postulated that the interaction of transthyretin fibrils with the receptor for advanced glycation end products can lead to the pathogenesis of neurodegeneration by contributing to NF-κB activation. In our study, we investigated the immunolocalization and activation of NF-κB in peripheral neuropathies of different origins.

We investigated sural nerve biopsies obtained with informed consent from 17 patients, 5 with CIDP, 3 with vasculitis, 4 with Charcot-Marie-Tooth disease (CMT) (2 with CMT types 1 and 2 with CMT type 2), and 5 with FAP. Diagnosis was based on results of clinical and laboratory studies in patients with CIDP and vasculitis; genetic analysis was performed in all patients with CMT and FAP. Table 1 summarizes the clinical characteristics of the patients. The patients with CIDP or vasculitis had received no immunosuppressive or immunomodulatory treatment before biopsy, except for 1 patient (patient 5) with CIDP who had been taking prednisone (50 mg/d) for 3 months, with moderate improvement. Three normal sural nerves served as control specimens. All the specimens had been frozen in isopentane, cooled in liquid nitrogen, and stored at –70°C.

Seven-micrometer transverse sections were incubated for 60 minutes at room temperature in mouse monoclonal antibody against the NF-κB p65 subunit (1:500; CHEMICON International, Inc, Temecula, Calif), which recognizes an epitope overlapping the nuclear location signal of the p65 subunit of the NF-κB heterodimer. Therefore, it selectively binds to the activated form of NF-κB. To block nonspecific binding of the antibody, sections were preincubated with 1:10 diluted normal goat serum. Control for staining specificity included omission of the primary antibody and replacement with irrelevant monoclonal antibody or with species-specific nonimmune serum. Immunodetection was performed using a biotin-avidin system (Dako, Milan, Italy), followed by horseradish peroxidase staining with 3,3-diaminobenzidine tetrahydrochloride. Serial sections were also examined for the macrophage-differentiating antigen CD68 (1:100), CD22 (1:100), CD4 (1:50), and CD8 (1:100) cell subsets and the terminal complement complex C5b-9 (membrane attack complex [MAC]) (1:50), all from Dako. Two or more entire nerve sections for each patient were semiquantitatively analyzed by 2 of us (A.M. and G.V.). Reproducibility was assessed by comparing 2 sets of data collected independently, indicating good reliability. Some analogue pictures from serial sections were displayed simultaneously on 2 monitors for direct comparison. The study was authorized by the Medical School Ethical Committee, University of Messina.

Because at least 30 mg of nerve tissue is necessary for isolation of an appropriate amount of nuclear proteins, we were able to perform protein analysis in 7 of 17 patients (patients 2 and 5 with CIDP, patient 7 with vasculitis, patient 11 with CMT, and patients 13, 14, and 17 with FAP) and in 2 normal nerves. All the methods, including isolation of nuclear protein, immunoblotting analysis of NF-κB, and electrophoretic mobility shift assay (EMSA), have recently been detailed elsewhere.¹³

Table 2 summarizes the immunocytochemical results. A few CD68 cells, indicative of macrophages, and some CD4 and CD8 T lymphocytes, but no CD22 B cells, were found in the endoneurium and perineurium in all patients with CIDP. Inflammatory infiltrates with abundant CD4, CD8 cells, CD68 macrophages, and some CD22 B cells were detected in epineurial and endoneurial vessel walls in specimens from patients with vasculitis. Some endoneurial and epineurial vessels were positive for MAC in CIDP and vasculitis specimens (Figure 1A and E). Immunoreactivity for MAC was also found in a thin rim surrounding the outermost layer of myelin sheath of 5% to 10% of nerve fibers in CIDP specimens and in 2% to 5% of axons in vasculitis specimens (Figure 1C and G). Strong MAC positivity was observed at the level of all amyloid deposits in patients with FAP (Figure 2A). Charcot-Marie-Tooth disease nerves and control nerves showed no inflammatory cells and no MAC immunostaining. Nonspecific perineurial immunoreactivity for MAC was seen in all nerves studied.

Moderate NF-κB immunoreactivity was found in 2% to 5% of endoneurial vessel walls, in the outermost layer of myelin sheath of 5% to 10% of nerve fibers, and in a few axons in CIDP specimens (Figure 1B and D). It was also seen in 5% to 15% of epineurial and endoneurial vessels in vasculitis specimens (Figure 1F). In FAP nerves, amyloid deposits exhibited mild NF-κB immunoreactivity (Figure 2B). Endoneurial vessels expressed NF-κB in CMT nerves sporadically, with mild intensity. No correlation was found between the presence and type of inflammatory cells and NF-κB immunoreactivity. Nuclear factor κB immunoreactivity was sometimes found at the same sites of MAC deposition in CIDP specimens and in vessel walls in vasculitis specimens. A correspondence between NF-κB and MAC was almost always seen at amyloid deposits in FAP specimens. No immunosignal was detected in control specimens.

Western blot analysis of nuclear extracts, using antibody against the NF-κB p65 subunit, showed a single band of high intensity, corresponding to 65 kDa, in the specimens from all diseased nerves and a milder band in control nerves (Figure 3A).

Nuclear factor κB DNA-binding activity was increased in patient 2 with CIDP, patient 7 with vasculitis, and patients 13, 14, and 17 with FAP. A faint signal was seen in patient 5 with CIDP taking corticosteroids, in patient 11 with CMT, and in 2 normal nerves (Figure 3B and Table 2).

We detected immunoreactivity for NF-κB in dysimmune neuropathies, but with a different pattern in CIDP specimens compared with a previous study.¹¹ Immunocytochemistry is the sole technique that can localize a given antigen in its cell. However, immunocytochemical assays have several limitations, such as availability of the antigen sites, specificity and sensitivity of the antibody, and technical artifacts. Therefore, immunocytochemistry should be combined with other types of analysis. Western blot of nuclear extracts and, to a greater extent, demonstration of NF-κB DNA-binding activity by EMSA provide the best evidence of NF-κB activation. To our knowledge, NF-κB has not been investigated using these 2 techniques in human nerve biopsy specimens. Indeed, we found positive NF-κB immunoblotting within nuclear proteins in all diseased nerves examined. A high level of DNA-binding activity on EMSA was observed in 1 of 2 patients with CIDP, in a patient with vasculitis, and in 3 of 3 patients with FAP. This confirms the lower specificity of Western blot compared with the EMSA technique, which has functional implications in quantifying the ability of NF-κB to bind consensus sequence sites of gene promoters. To our knowledge, our study is the first demonstration of NF-κB activation in CIDP and vasculitis specimens, suggesting a crucial role of this transcription factor in inflammatory peripheral nervous system disorders. A positive NF-κB immunoblotting within nuclear proteins in patients who showed absent DNA-binding activity on EMSA could be explained by the presence of non–DNA-binding p65 subunit, which plays a role in controlling inducibility and specificity of NF-κB DNA-binding.^14,15 Nuclear factor κB activation is inhibited by several drugs, including glucocorticoids, cyclosporine, tacrolimus, aspirin, and other nonsteroidal anti-inflammatory medications. This may account for the minimal DNA-binding activity found in patient 5 with CIDP who had been taking prednisone for 3 months, compared with the untreated patients with CIDP, and further explains our results.

There has been recent research on possible mechanisms of neurodegeneration associated with FAP, but how extracellular amyloid deposits lead to cell death is unknown. Maximal NF-κB DNA-binding activity was demonstrated in vitro by EMSA, using PC-12 cells transfected with receptor for advanced glycation end product complementary DNA and exposed to soluble or fibrillar transthyretin.¹² Our novel demonstration of NF-κB activation in FAP nerves underlines the possible role of this factor in inflammatory or apoptotic mechanisms.

In summary, our findings support activation of a NF-κB pathway in CIDP, vasculitis-related polyneuropathy, and FAP. It is not surprising that such a role is not disease-specific, because NF-κB may be involved with similar or dissimilar mechanisms in different diseases. Nuclear factor κB is a ubiquitous essential survival factor and is implicated in various cellular processes and responses. It also has been found to play contradictory roles in different cell types in the same tissue or organ, such as apoptotic and antiapoptotic roles in the brain.¹⁰ Further studies are needed to better elucidate the protective or toxic function of NF-κB in dysimmune neuropathies and FAP and to delineate possible therapeutic implications.

Finally, some interesting findings were obtained in the study of MAC. Deposition of complement factors, including MAC, has been described at the level of Schwann cell membranes in some patients with inflammatory demyelinating polyneuropathy, supporting the role of humoral immunity in this disease.^16-18 Analogously, our findings suggest that in CIDP, antibody reaction against epitopes on the outer surface of the Schwann cell can activate the complement, resulting in myelin damage. Evidence of participation of a humoral mechanism has also been considered in axonal neuropathy, such as vasculitis, in which deposition of immunoglobulins and complement has been detected in vessels with intense cellular infiltrates.¹⁹ We found MAC deposition in a few axons in vasculitis specimens, where with other molecules it may be involved in nerve fiber degeneration. Moreover, we detected in FAP nerves a strong immunoreactivity for MAC, with a precise correlation to amyloid plaques, confirming a recent report.²⁰ At the same levels, we demonstrated mild NF-κB immunoreactivity, which is in accord with the known regulation of complement complex activity by NF-κB.⁴ Complement activation on amyloid deposits may contribute to injury of axons in their vicinity, in a manner similar to the mechanisms proposed in Alzheimer disease.²¹

Correspondence: Anna Mazzeo, MD, Dip di Neuroscienze, Scienze Psichiatriche ed Anestesiologiche, Clinica Neurologica 2, Policlinico Universitario, 98125 Messina, Italy (annamazzeo@yahoo.it).

Accepted for publication August 4, 2003.

Author contributions: Acquisition of data (Drs Mazzeo and Aguennouz); analysis and interpretation of data (Drs Mazzeo, Messina, and Vita); drafting of the manuscript (Drs Mazzeo and Aguennouz); critical revision of the manuscript for important intellectual content (Drs Messina and Vita); obtained funding (Drs Messina and Vita); administrative, technical, and material support (Drs Mazzeo and Aguennouz).

This study was supported by grants from the National Research Council and the Italian Ministry of Education, University, and Research, Rome, Italy.