OUP user menu

Increased Expression of Endoplasmic Reticulum Stress-Related Signaling Pathway Molecules in Multiple Sclerosis Lesions

Aoife Ní Mháille BSc, Stephen McQuaid PhD, Anthony Windebank MD, PhD, Paula Cunnea PhD, Jill McMahon PhD, Afshin Samali PhD, Una FitzGerald PhD
DOI: http://dx.doi.org/10.1097/NEN.0b013e318165b239 200-211 First published online: 1 March 2008


Activation of endoplasmic reticulum (ER) stress-related cell signals has been reported in several neurologic disorders and may contribute to neurodegeneration. Endoplasmic reticulum stress is also linked to ischemic injury. However, activation of an ER stress response has not been investigated in multiple sclerosis (MS) lesions. We detected increased expression of ER stress-associated C/EBP homologous protein, immunoglobulin heavy chain-binding protein, and X-box-binding protein 1 in multiple cell types, including oligodendrocytes, astrocytes, T cells, and microglia in active MS lesions. Semiquantitative analysis of expression in active, chronic active, and chronic inactive lesions indicated that levels of immunoglobulin heavy chain-binding protein were significantly higher in acute lesions than in non-MS controls or MS normal-appearing white matter, and that ER stress-associated C/EBP homologous protein was upregulated to the greatest extent at the edges of chronic active lesions. Because demyelination may be triggered by a tissue response to ischemia-like conditions, changes in the hypoxia-related antigen D-110 were also investigated, and it was found that increased ER stress-associated C/EBP homologous protein expression can occur in either the presence or absence of D-110. A possible link between a perturbed ER and lesion development in MS suggests a signaling pathway that may represent a new therapeutic target in MS.

Key Words
  • BiP
  • CHOP
  • D-110
  • Endoplasmic reticulum stress
  • Hypoxia
  • Multiple sclerosis
  • XBP1


Multiple sclerosis (MS) is an inflammatory disease of the central nervous system associated with the development of large plaques of demyelination, oligodendrocyte destruction, and axonal degeneration. The sequence of molecular events leading to the formation of plaques is not fully understood, but a number of signaling pathways linked to particular forms of cellular stress have been implicated. For example, cellular responses triggered by oxidative stress or mitochondrial dysfunction (1-4), nitric oxide (5, 6), and excitotoxicity (7, 8) may contribute to oligodendrocyte loss and myelin or axonal degeneration; raised levels of these have been reported in MS. A possible link between hypoxic-like conditions and "dying back" oligodendrogliopathy during MS has also been proposed by Aboul-Enein and Lassmann (3), Aboul-Enein et al (9), and Lassmann (10). This, together with reports that endoplasmic reticulum (ER) stress contributes to the destruction of a variety of ischemic cells, including neurons (11, 12), led us to hypothesize that dying back oligodendrocytes found in some MS lesions can be the result of an ER stress response. The pattern of expression of markers consistent with the activation of ER stress and/or hypoxia within MS lesions has not previously been investigated.

Perturbation of the normal ER processes and equilibrium, which can be caused by many factors, activates a common signaling pathway termed the unfolded protein response (13-16). The unfolded protein response is generally a prosurvival mechanism that aims to restore ER homeostasis through the induction of 3 ER transmembrane proteins, namely, inositol-requiring enzyme 1, activated transcription factor 6, and pancreatic ER kinase-like ER kinase. Activation of the 3 transmembrane sensors leads to the induction of proteins and chaperones, including X-box binding protein (XBP1) and immunoglobulin heavy chain binding protein (BiP), which functions to restore the ER to its normal physiologic state. If ER stress persists and protective mechanisms fail, signaling becomes proapoptotic (17). C/EBP homologous protein (CHOP) was the first transcription factor to be linked to proapoptotic events induced after ER stress (17). C/EBP homologous protein tips the balance in favor of cell death through its downregulation of the anti-apoptotic protein Bcl-2 and by induction of the pro-apoptotic BH3-only protein Bim (18, 19).

To establish the potential relevance of ER stress to MS pathogenesis, we performed a detailed semiquantitative immunohistochemical and molecular analysis of a large cohort of MS tissue samples. Specifically, we analyzed the variation in levels of the ER stress-related molecules CHOP, BiP, and XBP1 in MS lesions and its association with the hypoxia-related protein D-110.

Materials and Methods

Patient Details

Informed consent for research on all brain tissues and local ethical approval was obtained for the conduct of this study. Archival formalin-fixed paraffin-embedded (FFPE) MS tissue derived from 27 neurologically diagnosed and neuropathologically confirmed autopsy cases and 9 normal control non-MS patients without other neurologic disease, collected at the UK MS tissue bank in London and the Belfast MS Donor Tissue Bank in the Department of Neuropathology, Royal Group of Hospitals Trust, UK, was used. All FFPE sections (6 μm) underwent preliminary screening after standard Luxol fast blue (LFB; VWR International, Poole, UK)/hematoxylin and eosin (H&E; VWR International) and human leukocyte antigen (HLA)-DR (Dako, Ely, UK) staining to determine their pathologic state as lesion or normal-appearing white matter (NAWM) based on the Medical Research Council MS pathologic classification and International Classification of Diseases of the Central Nervous system (20). For molecular studies, snap-frozen postmortem material collected at the London MS Tissue Bank, UK, was used. Cryostat sections (12 μm) were screened using Oil-red O (ORO; Sigma-Aldrich, Dublin, Ireland) and anti-myelin oligodendrocyte glycoprotein (MOG; clone Z12, 1:100, kindly supplied by Professor Richard Reynolds, Department of Cellular and Molecular Neuroscience, Imperial College London, UK). Clinical and pathologic features of all patients, including the total number of lesions examined, are summarized in Table 1.

View this table:


Immunohistochemistry was performed on FFPE sections using the following antibodies and dilutions: monoclonal antibody to MOG (1:100; source as indicated above) macrophages/microglia (HLA-DR; CR3/43; 1:500; Dako); hypoxia-related protein (D-110; 1:2000 in 10% fetal calf serum/Tris-buffered saline (TBS); kindly supplied by Professor Hans Lassmann, Medical University of Vienna, Vienna, Austria); polyclonal antibodies to the ER stress proteins (CHOP; sc793; 1:400; and XBP1; sc-7160; 1:100; both Santa Cruz Biotechnology, Inc., Heidelberg, Germany; BiP; AB32618; 1:200; Abcam, Cambridge, UK).

All FFPE sections were deparaffinized, and antigen retrieval was performed in either 0.01 mol/L citrate (Sigma-Aldrich; pH 6.0) at 700 W in a microwave for 20 minutes (HLA-DR; D-110, CHOP, and BiP) or in 0.01 μm Tris-EDTA (Sigma-Aldrich; pH 9.0) in a pressure cooker at full steam pressure for 2 minutes (XBP1; 31-32). D-110-immunostained sections were pretreated with 10% fetal calf serum (Invitrogen, Dublin, Ireland)/TBS (Sigma-Aldrich) for 20 minutes at room temperature. All antibodies were then incubated on sections overnight at 4°C and detected using peroxidase-labeled Envision anti-mouse or anti-rabbit secondary antibodies (Dako) with diaminobenzidine (Dako) as chromogen. Sections from all MS lesions were also dual-labeled for HLA-DR and LFB. Briefly, sections were immunolabeled for HLA-DR and after diaminobenzidine incubation were washed in tap water and stained with LFB. After differentiation, sections were counterstained with hematoxylin, dehydrated, cleared, and mounted. The specificity and use of all these antibodies for immunocytochemistry on tissue sections have been previously determined (10, 21, 22).

Negative controls were generated by omission of primary antibodies or by section pretreatment with proteinase K (Sigma-Aldrich). Cryostat frozen sections (12 μm) were fixed with either 4% paraformaldehyde when staining for MOG or 10% neutral-buffered formalin (Sigma-Aldrich) for ORO staining (23). After paraformaldehyde fixation, MOG immunocytochemistry was performed as previously described (24).

Lesion Characterization

Formalin-fixed paraffin-embedded sections stained with H&E/LFB and by immunohistochemistry for HLA-DR antigen were used to classify lesions as active, chronic active (CAL), or chronic inactive. Ongoing demyelination was also confirmed by the presence of LFB-positive myelin fragments within phagocytosing macrophages, as determined by HLA-DR/LFB dual labeling. Active lesions (ALs) showed pronounced pallor in LFB staining (Fig. 1A, i), and macrophages containing LFB-positive particles of myelin were observed within the lesion center (LC) (Fig. 1A, ii-iii). Chronic active LCs had less myelin staining (Fig. 1A, iv) and were essentially devoid of macrophages, although large numbers ramified microglia, and occasional LFB-positive macrophages were detected at the lesion edges (LEs) (Fig. 1A, v-vi). A few persisting microglial cells were noted in some inactive centers (Fig. 1A, vii). Microglial cells were sparse in chronic inactive lesions (CILs) and were almost completely devoid of myelin fragments (Fig. 1A, viii-ix).


Lesion classification. (A) Lesions within formalin-fixed paraffin-embedded material were subclassified as active, chronic active, or chronic inactive. (i) Active lesions (ALs) showed pale staining in conventional Luxol fast blue (LFB) myelin stain and were sharply demarcated from the surrounding perilesional white matter. (ii) Immunohistochemistry with the macrophage activation marker human leukocyte antigen (HLA)-DR shows pronounced macrophage staining in the lesion. (iii) Luxol fast blue-positive material is detected within many phagocytosing macrophages (arrows) using HLA-DR/LFB staining. (iv) There was greater loss of myelin in the inactive centers of chronic ALs (CALs). (v) Human leukocyte antigen-DR staining in CAL was abundant at the edge and reduced within the plaque. (vi) Luxol fast blue-myelin debris is detected within macrophages at the edge of CALs (arrows). (vii) Chronic inactive lesions (CILs) show a complete loss of myelin. (viii) Lack of macrophage infiltration in CILs. (ix) Absence of LFB-positive macrophages in CILs. Scoring regions are shown as follows: lesion center (LC), lesion edge (LE), and perilesion (PL). Inset images represent the entire section, and the approximate location (red box) within each section corresponding to the higher magnification images is shown. Scale bars = (i, ii, iv, v, vii, viii) 500 μm; (iii, vi, ix) 12.5 μm. (B) Characterization of frozen autopsy material before molecular analysis. Lesions with active demyelination shared the features of Oil-red O (ORO) staining of foamy macrophages (red) and reduced staining of the myelin (MOG; brown). Oil-red O-negative lesions devoid of MOG activity were classified as CIL (scale bars = 500 μm). Lower panels show magnifications of regions marked by an asterisk in the corresponding upper panel. Scale bars = 12.5 μm.

In frozen sections, ORO activity in conjunction with MOG loss was used to define lesion activity. Lesions that contained ORO-positive macrophages at their borders in conjunction with reduced MOG activity were defined as chronic active. In contrast, CILs were largely devoid of ORO-positive macrophages with complete MOG loss (Fig. 1B).

In both MS FFPE and frozen tissues, white matter with no myelin abnormalities was recorded as perilesional (PL) if a lesion was present in the tissue block or as NAWM if no lesion was present in the block. Tissue blocks from control cases contained normal white matter assessed as pathologically normal by H&E/LFB and HLA-DR (paraffin) or ORO (frozen) staining.

Semiquantitative Analysis of Immunocytochemistry

Microscopic assessments were primarily done to describe overall regional variations in the expression of ER stress proteins and D-110. A semiquantitative analysis of specific cell type expression levels was adopted. To describe inherent variations in expression of these molecules within lesion-positive sections, different pathoanatomic areas were assessed: LC, the central zone of myelin destruction; LE, representing the rim of white matter separating the lesion from the surrounding white matter, and PL, consisting of the white matter surrounding a lesion (Fig. 1A). Expression levels were also recorded in lesion-free MS NAWM and the white matter of control non-MS patients. Twenty microscopic fields were assessed at 40× objective from all regions scored.

In all regions scored, low, moderate, or abundant immunoreactivity was detected. To quantify this variation, the following scoring system was applied. A score of 0 described no antigen expression; 1, low expression (0%-30% of cells); 2, moderate expression (30%-60%); and 3, abundant expression (60%-100%).

Selected histologically or immunocytochemically stained sections were digitally scanned using an Aperio Scanscope T3 (Aperio Technologies, Inc., Bristol, UK) with a 40× objective. From these scans, selected images can be viewed or displayed at a range of magnifications.

Statistical Analysis

After semiquantitative evaluation, median scores for CHOP, BiP, and D-110 in control non-MS, NAWM, and in each region (i.e. PL, LE, and LC) from acute, primary progressive (PP), and secondary progressive (SP) cases were subjected to statistical analysis. X-box binding protein-positive acute material was not subjected to statistical evaluation due to the small number of sections available for staining. Analysis of variance (nonparametric) was completed with a Kruskal-Wallis test followed by a Dunn posttest using GraphPad Prism software (GraphPad Software, Inc., San Diego, CA).

Dual Immunofluorescence

In selected MS tissue blocks, dual-immunofluorescence labeling was used to map the cellular distribution of CHOP. After microwave antigen retrieval and blocking of autofluorescence (25), sections were incubated overnight at 4°C in a mix of anti-CHOP and 1 of the cell-specific monoclonal antibodies: astrocyte-specific protein (glial fibrillary acidic protein [GFAP] clone 6F2; 1:200; Dako); macrophage-specific protein (CD68; clone PGM1; 1:100; Dako); T-lymphocyte-specific protein (CD3; clone F7.2.38; 1:50; Dako), MOG, or oligodendrocyte-specific protein (Nogo-A; clone 11C7; 1:5000; kindly provided by Professor M. Schwab, Brain Research Institute, Dept. of Neurology, University of Zurich, Zurich, Switzerland). After washes in TBS, bound antibodies were detected by incubation in a mixture of goat anti-rabbit Alexa-568 and goat anti-mouse Alexa-488 (both 1:500; Invitrogen) for 2 hours at 37°C. After final rinses in TBS, sections were mounted in 4′,6-diamidino-2-phenylindole Vectashield (Vector Laboratories, Inc., Dublin, Ireland). Mounted sections were analyzed, and images were acquired using a Leica DF350 FX digital camera and processed using Leica FW4000 software (Leica, Milton Keynes, UK).

RNA Extraction and Polymerase Chain Reaction

Total RNA was extracted and pooled from four 20-μm cryostat tissue sections derived from each of 7 MS and 4 control non-MS patients using Trizol reagent (Invitrogen) and based on the method of Haddock et al (26). Genomic DNA was removed using an RNeasy MinElute Clean up kit (QIAGEN, Crawley, UK). RNA quantity and quality were determined using a Nanodrop ND-1000 Spectrophotometer (Labtech International, Ringmer, UK) and an Agilent 2100 Bioanalyser (Agilent Technologies, Dublin, Ireland). An RNA integrity number calculated by a built-in Agilent Technologies software algorithm (27) with a value ranging from 1 (poor quality) to 10 (high quality) was generated for each sample. Polymerase chain reaction-based studies used samples with an RNA integrity number greater than or equal to 5 because this has been shown to be necessary to generate reproducible data using polymerase chain reaction methodology (28).

Copy DNA was synthesized from 1 μg of RNA according to the manufacturer's instructions (Invitrogen) using oligo dT primer. Amplification of CHOP sequence was modified from Mertani et al (29). Briefly, 2 μm of forward (5′-GCA CCT CCC AGA GCC CTC ACT CTC C-3′) and reverse (5′-GTC TAC TCC AAG CCT TCC CCC TGC G-3′) primers were combined with 200 μmol/L deoxynucleotide triphosphate mix (Invitrogen), GoTaq reaction buffer, and 0.25 U of GoTaq DNA polymerase (Promega, Hampstead, UK). Solutions were denatured at 95°C for 3 minutes, followed by 28 cycles of 94°C for 30 seconds, 70°C for 30 seconds, and 72°C for 30 seconds, and a final extension at 72°C for 10 minutes.

The ribosomal 18S subunit was used as the housekeeping gene with the following primers: forward primer, 5′-CTT AGA GGG ACA AGT GGC G-3′ reverse primer, 5′-GGA CGT CTA AAG GGC ATC ACA-3′. Solutions were denatured at 95°C for 3 minutes, followed by 29 cycles of 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds, and a final extension at 72°C for 10 minutes. No template controls or absence of reverse transcriptase during cDNA synthesis were used as negative controls.


ER Stress Response Proteins in Actively Demyelinating MS Lesions

The profiling of ER stress markers was first investigated in 2 acute MS cases. In total, 14 areas of lesional activity, defined by the infiltration of macrophages containing LFB-positive fragments, were analyzed. Luxol fast blue/H&E and HLA-DR-stained tissue from 1 acute MS case containing multiple lesions is depicted in Figure 2(i, ii). D-110 was expressed at high levels in all actively demyelinating lesions (n = 14) in these samples (Fig. 2, iii). C/EBP homologous protein and BiP expression mirrored that seen for D-110 in these lesions, with strong staining of a variety of cell types within all LCs and edges (Fig. 2, iv, v). Unlike D-110 and CHOP, BiP expression was not confined to areas of ongoing demyelination but was also detectable in completely demyelinated areas. D-110 and BiP stainings were significantly increased in ALs compared with control non-MS tissue at the edges (p < 0.01) and in the centers (p < 0.001 and p < 0.01, respectively). X-box binding protein was examined in a smaller number of lesions that could not be subjected to statistical analysis. In the cases examined, there was variability in XBP1 expression in that lesions showed low, moderate, or abundant levels of immunoreactivity (Fig. 2, vi) regardless of the stage of demyelination.


Profile of expression in acute multiple sclerosis lesions. (i) Many lesions (marked with asterisks) scattered throughout the white matter of tissue from a 51-year-old female patient with a clinical disease duration of 18 weeks. Abundant expression of the immune activation marker human leukocyte antigen-DR (ii) is duplicated by intense D-110 (iii), C/EBP homologous protein (CHOP) (iv), and immunoglobulin heavy chain binding protein (BiP) staining (v) within the same plaque region. X-box binding protein (XBP1) was also found to be expressed in some acute lesions (vi). Red asterisks within inset images indicate the approximate location within each section shown in higher magnification images. At increased magnifications, D-110 staining was observed in different cell types, including those morphologically resembling foamy macrophages (vii), astrocytes (viii), and oligodendrocytes (ix). C/EBP homologous protein (x), BiP (xi), and XBP1 (xii) were also expressed by a variety of morphologically distinct cell types. Letter codes: O, oligodendrocyte; A, astrocyte; LC, lesion center; LE, lesion edge; M, Macrophage. Scale bars = (i) 1 mm; (ii-vi) 50 μm; (vii-xii) 12.5 μm.

Expression of D-110, CHOP, XBP1, and BiP was not restricted to 1 cell type. D-110 was noted in a variety of cell types, including macrophages (Fig. 2, vii), astrocytes (Fig. 2, viii), and cells that most closely resembled small, rounded oligodendrocytes (Fig. 2, ix). Expression of all proteins was not only present in the cytoplasm, but intense staining was also observed in the nucleus, except for BiP, which was only observed in the cytoplasm (Fig. 2, x-xii).

CHOP Expression Levels are Highest at the Edges and Centers of CALs

Studies were next expanded to include tissues from the patients with the more prevalent forms of MS, that is, PP and SP MS. Strong expression of BiP was observed in all active lesional areas. However, CHOP was found to be more variably expressed. Table 2 provides an overview of the variability in CHOP expression (absent, low, moderate, or high) according to lesion type (active, chronic active, chronic inactive). Abundant levels of CHOP were found in 55% of active/acute, 74% in chronic ALs, and 32% in CILs. Although fewer lesions were analyzed for XBP1 (34 in total), overall expression levels were lower than those of CHOP. For example, in 11 of 29 active and chronic ALs examined (39%), a moderate level of XBP1 immunoreactivity was recorded, whereas only 5 of 29 (17%) displayed abundant levels.

View this table:

A plot of median scores (Fig. 3) demonstrated that CHOP expression was most significantly upregulated at the edges and centers of CALs. A maximal increase of 3-fold was recorded when the CAL edge was compared with control white matter (p < 0.05) and NAWM (p < 0.001). Central regions of CALs also showed significantly raised levels of CHOP (2-fold) compared with NAWM (p < 0.01). Similarly, the edges of CALs had levels of CHOP that were significantly greater than those in the PL areas of ALs (p < 0.01) (Fig. 3). However, there was a highly significant difference in CHOP expression when all chronic inactive regions were compared with CAL edges (p < 0.001).


Semiquantitative analysis of C/EBP homologous protein (CHOP) and immunoglobulin heavy chain binding protein (BiP) expression. (A) Median scores for CHOP and BiP in control non-multiple sclerosis tissue (C); normal-appearing white matter (NAWM); active lesions (ALs); chronic active lesions (CALs); chronic inactive lesions (CILs) isolated from primary progressive and secondary progressive patients. Anatomic regions scored were perilesion (PL), lesion edge (LE), and lesion center (LC). Sample number (n) consisted of the total number of each type of tissue block assessed from each anatomic region. C/EBP homologous protein expression was significantly upregulated at the edge of CALs when compared with NAWM and all regions within CILs (p < 0.001). (B) Summary of significant differences when all regions scored were cross-compared for CHOP and BiP expression. A-LC, active LC; A-LE, active LE; A-PL, active perilesion; CA-LC, chronic active LC; CA-LE, chronic active LE; CI-LC, chronic inactive LC; CI-LE, chronic inactive LE; CI-PL, chronic inactive perilesion.

Levels of BiP in all normal and diseased tissue regions were high, with scores more than 2 in most regions. With the exception of CALs, in which there was a significant decrease in expression when both LEs and centers were compared with AL centers (p < 0.05), the expression profile of BiP staining across all lesion types was generally high and essentially consistent within the different regions scored (Fig. 3).

CHOP Is Expressed in a Variety of Cell Types in Demyelinating Lesions

Figure 4(i-iii) shows representative changes in expression of CHOP observed at the LE (ii) and centers (iii) of chronic active MS lesions compared with NAWM (i). Dual immunofluorescence for MOG (green) and CHOP (red) also demonstrates large numbers of CHOP-expressing cells present at the edges of chronic ALs (iv)-(v).


Detection of C/EBP homologous protein (CHOP) in a variety of cell types. Representative images show upregulation of CHOP at the edge (ii) and in the center (iii) of chronic active lesions (CALs) compared with multiple sclerosis normal-appearing white matter (NAWM) (i). Red asterisks in insets indicate the approximate location within each section from which each magnified image was taken. Dual immunofluorescence for myelin oligodendrocyte glycoprotein (MOG)/CHOP (iv, v), CD3/CHOP (vi), CD68/CHOP (vii), glial fibrillary acidic protein (GFAP)/CHOP (viii), and Nogo-A/CHOP (ix) are shown. (iv) Pronounced depletion of MOG immunoreactivity in the presence of abundant CHOP expression is observed through the lesion edge. (v) Higher magnification of the highlighted area (iv) demonstrates more clearly the abundance of CHOP expression. (vi) C/EBP homologous protein immunoreactivity was also present in T cells within perivascular cuffs as indicated by CD3/CHOP dual immunofluorescence (arrows). (vii) C/EBP homologous protein + ve/CD68 + ve (*, inset image at higher magnification) and CHOP + ve/CD68 − ve (arrows) cells were present in lesions. (viii) C/EBP homologous protein + ve/GFAP + ve (*, inset image at higher magnification) and CHOP + ve/GFAP − ve (arrows) cells were also present in lesions. (ix) Nogo-A was used for the identification of oligodendrocytes in demyelinated lesions. C/EBP homologous protein immunoreactivity was present in Nogo-A-positive cells (*). The cellular antigens were immunodetected with Alexa-488 (green), and CHOP is immunodetected with Alexa-568 (red). All sections were counterstained with 4′,6-diamidino-2-phenylindole. LC, lesion center; LE, lesion edge. Scale bars = (iv) 50 μm; (i-iii) 25 μm; (v-viii) 12.5 μm; and (ix) 8 μm and inset images (vii-ix).

Because CHOP was found to be most consistently upregulated in CALs, this region was selected for dual immunofluorescence with a panel of cell-specific markers to determine which cell types showed increased expression of ER stress proteins. Many small CD3-positive T cells within perivascular infiltrates expressed CHOP (Fig. 4, vi; arrows). CD68/CHOP dual immunofluorescence revealed that most CD68-positive macrophages at the LC were CHOP negative (Fig. 4, vii; arrows). GFAP/CHOP dual immunofluorescence demonstrated that a number of CHOP-expressing cells in lesional material were GFAP-positive astrocytes (Fig. 4, viii; asterisk). It was also evident that other GFAP-negative cell types expressed high levels of CHOP (Fig. 4, viii; arrows). In addition, large numbers of Nogo-A-positive cells also labeled positively for CHOP (Fig. 4, ix; asterisk). Omission of either primary antibody from the dual-labeling procedure resulted in no emission of specific fluorescent signals from the channel used to detect that antigen.

CHOP Expression Can Occur in the Presence or Absence of D-110 in Lesions from Primary or Secondary Progressive MS Patients

Although 65% of lesions examined from cases of PP or SP MS showed abundant levels of CHOP, the staining did not always coincide with expression of the D-110 antigen. For example, tissue samples from 17 of the 25 MS patients examined showed absent or very low levels ofD-110-immunoreactive cells regardless of the stage of demyelination. The other 8 cases demonstrated considerable intraindividual and/or interindividual variation in their D-110 staining pattern. One of these, an SP patient with no neuropathologic findings, suggesting a coincident hypoxic condition, consistently displayed abundant D-110- and CHOP-immunoreactive cells in all ALs examined (Fig. 5, top panel). By contrast, another patient with SP MS showed abundant expression of CHOP within ALs that were D-110 negative (Fig. 5; lower panel).


C/EBP homologous protein (CHOP) expression in actively demyelinating lesions in the presence or absence of D-110 immunoreactivity. Luxol fast blue (LFB) showing a large brain lesion with an ill-defined border from a 69-year-old female patient with secondary progressive (SP) multiple sclerosis (MS; clinical disease duration of 46 years). Immunohistochemistry for D-110 and CHOP shows abundant expression of both antigens within this actively demyelinating lesion (top panel). In contrast, a 42-year-old female patient with SP MS (clinical duration of 18 years) demonstrates abundant expression of CHOP in the core of a D-110-negative demyelinating lesion (lower panel). Scale bars = 1 mm in LFB images, 25 μm in D-110 images, and 12.5 μm in CHOP images. AL, active lesion.

Molecular Analysis of CHOP Expression

To determine whether or not alterations in CHOP expression can be detected at the transcriptional level, RNA was isolated from snap-frozen material. In total, 2 PP patients containing 7 ALs and 5 SP patients with 13 ALs were used in this study. For comparison, 4 control non-MS cases were included. On average, active lesions showed upregulation of CHOP in comparison to corresponding NAWM and chronic inactive lesional material (Fig. 6).


Molecular analysis of C/EBP homologous protein (CHOP) expression. Reverse-transcriptase polymerase chain reaction showing CHOP and 18S polymerase chain reaction products generated using RNA isolated from between 3 and 10 tissue blocks derived from 7 different multiple sclerosis (MS) patients (P1-P7), as well as from control non-MS subjects (lanes marked control). Levels of CHOP in blocks containing active lesions (boxed lanes) were generally higher than that seen in chronic inactive (♦) lesions or in normal-appearing white matter (*).


The major findings of this study are as follows: 1) markers consistent with the occurrence of ER stress are increased over controls in MS lesions; 2) some CHOP-positive lesions also display high levels of D-110, an epitope associated with hypoxia; and 3) there are variations in the expression patterns of the antigens studied both in lesions at different stages and in different pathoanatomic areas of the lesions.

After the examination of a large number of acute, active, and CILs, raised levels of BiP, CHOP, and XBP1 were detected in diseased samples when compared with NAWM or non-MS controls. Although endogenous levels of CHOP and XBP1 were detectable in control non-MS and NAWM, this increased 2- to 3-fold in the perilesion, LE, and centers of actively demyelinating MS lesions. This was then followed by reversion to control levels across CILs. When significance values for regional comparisons were analyzed, the edge of CALs emerged as the primary site of CHOP activation. Immunoglobulin heavy chain binding protein was also found to be significantly upregulated in the center and at the edge of acute lesions when compared with control non-MS tissue. A significant drop in BiP expression was detected when the centers of active and CALs were compared, leading us to speculate that BiP loss within the LC could have enabled CHOP induction. Furthermore, concentration of CHOP activity at the edge of CALs suggests that CHOP expression may contribute to lesion expansion. In broad agreement with our immunohistochemical analyses, most active lesions displayed a clear upregulation of CHOP mRNA expression in comparison to NAWM and CILs isolated from the same subjects. Interestingly, CILs isolated from 2 patients displayed comparable levels of CHOP mRNA to those of ALs. To investigate alterations of CHOP mRNA expression in specific cell types, enrichment of tissue sampling (e.g. by application of laser capture methodology) will be required (30).

Immunohistochemical and dual-immunofluorescent analysis of ER stress protein expression in acute, active, and chronic MS lesions indicated that oligodendrocytes, astrocytes, and, to a lesser extent, macrophages expressed ER stress molecules. Dual labeling using anti-CD3 also demonstrated the presence of perivascular CHOP-positive T cells. The detection of markers of ER stress in a variety of cell types suggests a pathogenic trigger capable of affecting different cell types within an MS lesion. However, the presence of CHOP-negative cells in all dual-labeled fields examined suggests that other factors within the immediate environment of a cell or within an individual cell may determine whether or not the ER stress signaling pathway is activated. These factors can include the length of exposure to the pathogenic stimulus or the distance from its source.

In support of our findings are data from a microarray study of NAWM isolated from 10 MS patients that showed on average 2.5-fold upregulation in CHOP in diseased tissue (31). A separate smaller study of chronic MS tissue lesions demonstrated altered levels of activated transcription factor 4, the chief regulator of CHOP, in MS lesions (32). Furthermore, in the widely used mouse model of MS, experimental autoimmune encephalomyelitis, Lin et al (33) propose involvement of ER stress via interferon-mediated activation of pancreatic ER kinase-like ER kinase in a protective response that prevents demyelination, axonal, and oligodendrocyte damage.

There are several potential explanations for activation of ER stress during MS. First, a link between excitotoxicity and ER stress was made when kainic acid-induced neuronal death was found to be abrogated by exposure to salubrinol, an inhibitor of ER stress (34). Raised levels of glutaminase have been detected in MS lesions (7); therefore, it is plausible that BiP, XBP1, and CHOP induction is a downstream effect of excitotoxicity. A second possibility is that ER stress is triggered by cytokines and/or other signaling molecules secreted by invading immune cells. For example, nitric oxide has been shown to induce CHOP in β-cells, macrophages, and neurons (35, 36), and increased amounts of nitric oxide metabolites have been detected in MS lesions (6). Furthermore, tumor necrosis factor, which has been shown to be upregulated in chronic ALs (32), induces unfolded protein response in a fibroblast model of ER stress (37). An intriguing third potential trigger is the development of hypoxic conditions, that is, conditions that would activate an intracellular response similar to that seen after a stroke. There is evidence that a hypoxic-like insult may be damaging or may lead to oligodendrocyte cell death during MS and other neuroinflammatory disorders (3, 9, 10, 38). In our study, we found D-110 to be statistically upregulated at the edge and in the centers of acute lesions when compared with control subjects. Detection of strong D-110 staining in tissue derived from acute MS patients lends further support to the notion that a hypoxic-type response can be part of the pathogenic trigger in the early stages of the disease (10). Moreover, because abundant CHOP and BiP were detected in the same ALs, they may be part of a tissue response to a hypoxic environment.

Together with D-110, heightened levels of CHOP were not restricted to ALs in acute MS patients but were also detected in chronically lesioned material. For example, 1 SP patient showed the same degree of D-110 and CHOP activity observed in the acute MS cases. Lesions within this patient had ill-defined borders and alternating rims of myelinated and demyelinated tissue, consistent with Balo's concentric lesions (Fig. 5). This kind of pathologic finding has been reported by Lucchinetti et al (38) and Stadelmann et al (39) in some MS patients who displayed features consistent with hypoxic-like injury. In another 2 SP patients, clusters of D-110-positive cells were located in NAWM and in proximity to already-established lesions. It is possible that such areas of increased D-110 expression in the NAWM may correspond to the very early prereactive lesions previously described (31, 40).

Whether CHOP induction is causally linked to cell death in MS lesions is still unclear. Because CHOP is mostly considered a cell death mediator, exerting its effect through mechanisms that lead to activation of the apoptosis-related protease caspase 3 (18), we examined a subset of lesions in which high levels of CHOP had been detected for the presence of cleaved poly-adenosine diphosphate-ribose polymerase. Poly-adenosine diphosphate-ribose polymerase cleavage is usually detected after activation of caspase 3. However, using a high-quality immunohistochemical-grade polyclonal antibody to cleaved poly-adenosine diphosphate-ribose polymerase, no expression was detected in any of the lesions examined despite finding high levels in positive control human tonsil tissue (unpublished observations).

Although ER stress-based therapies have not yet reached human clinical trials, there are encouraging indications that the ER signaling pathway is a valid new therapeutic target. For example, recent animal studies demonstrated that specific inhibition of this pathway can protect hippocampal neurons from excitoxicity in vivo (34).

Impairment of the ER has previously been reported in Alzheimer (41, 42), Parkinson (43), and prion diseases (44-46). With our data, we can now add MS as a disease that may benefit from future ER stress-targeted therapies. Given that ER stress-associated signaling is unlikely to occur in isolation, but occurs in association with an inflammatory response in MS lesions, deciphering the molecular events that predominate at a particular point in time and discerning the true relevance of ER stress pose a formidable challenge. Clarity will depend on the application of technologies capable of simultaneously recording cell-specific changes in all pathways activated and the ability to monitor these changes over time.


The authors thank the technical assistance of Gordon McGregor, Queen's University, Belfast, UK; and Dr. Gail Haddock (Sheffield Hallam University, Sheffield, UK) for provision of the tissue RNA isolation protocol. Tissue samples were supplied by the UK Multiple Sclerosis Tissue Bank, funded by the Multiple Sclerosis Society of Great Britain and Northern Ireland, registered charity 207495. Human tissue was obtained from a retrospective store of MS tissues in the Department of Neuropathology, Royal Group of Hospitals Trust, Belfast, UK.


  • Supported by the Multiple Sclerosis Society of Great Britain and Northern Ireland (P.C.), the National University of Ireland, Galway Millennium Research Fund, the National University of Ireland, Galway Foundation Office, and Multiple Sclerosis Ireland (A.N.M.). The authors' work was also supported by funding from Science Foundation Ireland.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
View Abstract