Abstract
Background: Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease with complicated pathogenesis that poses challenges with respect to diagnosis and monitoring of disease progression.
Objectives: To identify a biomarker panel that elucidates ALS disease pathogenesis, distinguishes patients with ALS from neurologic disease controls, and correlates with ALS disease characteristics, and to determine the effect of HFE gene variants, a potential risk factor for sporadic ALS, on the biomarker profile.
Methods: We obtained CSF samples by lumbar puncture from 41 patients with ALS and 33 neurologic disease controls. All patients were genotyped for HFE polymorphisms. We performed a multiplex cytokine and growth factor analysis and immunoassays for iron-related analytes. Classification statistics were generated using a support vector machine algorithm.
Results: The groups of patients with ALS and neurologic disease controls were each associated with distinct profiles of biomarkers. Fourteen biomarkers differed between patients with ALS and the control group. The five proteins with the lowest p values differentiated patients with ALS from controls with 89.2% accuracy, 87.5% sensitivity, and 91.2% specificity. Expression of IL-8 was higher in those patients with lower levels of physical function. Expression of β2-microglobulin was higher in subjects carrying an H63D HFE allele, while expression of several markers was higher in subjects carrying a C282Y HFE allele.
Conclusions: A CSF inflammatory profile associated with amyotrophic lateral sclerosis (ALS) pathogenesis may distinguish patients with ALS from neurologic disease controls, and may serve as a biomarker panel to aid in the diagnosis of ALS pending further validation. Some of these biomarkers differ by HFE genotype.
Glossary
- ALS=
- amyotrophic lateral sclerosis;
- ALSFRS-R=
- ALS Functional Rating Scale-revised;
- BSA=
- bovine serum albumin;
- FGF=
- fibroblast growth factor;
- G-CSF=
- granulocyte colony stimulating factor;
- GM-CSF=
- granulocyte-monocyte colony stimulating factor;
- IFN=
- interferon;
- IL=
- interleukin;
- MCP=
- monocyte chemoattractant protein;
- MIP=
- macrophage inflammatory protein;
- VEGF=
- vascular endothelial growth factor.
Amyotrophic lateral sclerosis (ALS) is a complex, progressive neurodegenerative disease that is difficult to diagnose early in its course because initial symptoms and signs often are similar to those of more common conditions, and there is no specific diagnostic test for ALS. The pathogenesis of ALS is largely unknown; thus identification of biomarkers associated with ALS could eventually assist early diagnosis and aid understanding of the disease by providing insights into its pathogenesis.1 As defined by the NIH Biomarkers Definitions Working Group, a biomarker is “a characteristic that is objectively measured and evaluated as an indicator of normal biologic processes, pathogenic processes or pharmacologic responses to a therapeutic intervention.”1 Biomarkers may also be used as indicators of disease progression and as measures of treatment effects. Finally, a comprehensive panel of biomarkers may help guide stratification of patients for different treatments based on etiopathogenesis.
A number of studies have been undertaken to identify biomarkers associated with ALS.2–8 While many of these studies have provided clues about pathogenetic mechanisms involved in the disease, most have examined only a small number of proteins. Several studies have examined proteomic profiles of patients with ALS compared to various control groups.8–10 These proteomic studies have provided valuable examples of proteins that can distinguish patients with ALS from control groups, but most proteomic techniques are better suited for abundant proteins. We provide in this study an analysis of a panel of analytes that are associated with inflammation and trophic support that may more accurately reflect early stages of disease and cell response.
Our first hypothesis was that a panel of biomarkers could be identified in the CSF that will support the clinical diagnosis. Extensive evidence has suggested ALS pathogenesis involves excessive neuroinflammation,11 aberrant growth factor regulation,12 and iron dyshomeostasis,13 among other factors. Despite many attempts to identify causes of sporadic ALS, no genetic polymorphisms have been identified to account for a large number of cases.14–16 The H63D polymorphism in the hemochromatosis gene (HFE) has been examined in multiple published studies,17,18 with an overall OR for patients with ALS possessing at least one H63D allele of 1.26 (95% CI 1.09–1.46) making this genetic variant the most frequently associated with ALS. HFE is involved in mediating iron homeostasis, inflammatory responses, and innate immunity.19 Therefore, we also tested the hypothesis that HFE gene variants will be associated with altered profiles in the panel of CSF biomarkers.
METHODS
Patients and samples.
We obtained blood samples by venipuncture and CSF samples by lumbar puncture at the time of evaluation from patients who were undergoing CSF examination as part of their diagnostic evaluation in the outpatient Neurology clinic. Patient demographics are described in table 1. These patients were grouped into those with ALS (clinically definite, probable, probable laboratory-supported, or possible ALS),20 and those who presented with neurologic symptoms but ultimately were found not to have ALS (neurologic disease controls). Demographic and selected medical data were obtained from the patients’ medical records. The ALS Functional Rating Scale Score-Revised (ALSFRS-R)21 was completed for each patient with ALS at an outpatient visit within 2 months of the lumbar puncture. All patients provided informed consent. This study was approved by the Institutional Review Board of the Penn State Milton S. Hershey Medical Center and Penn State College of Medicine. CSF samples were obtained between 8 am and 12 pm, to limit changes related to a circadian rhythm, from a mostly Caucasian population. Samples were frozen immediately after collection and were later thawed on ice and centrifuged to remove any particulate matter. Protease inhibitor cocktail (Sigma-Aldrich; St. Louis, MO) was then added 1:100, and samples were refrozen at –80°C in 200 μL aliquots.
Table 1 Comparison of CSF samples from patients with amyotrophic lateral sclerosis (ALS) and neurologic disease controls
HFE genotyping.
DNA was purified from white blood cells using the QIAamp DNA Mini kit (Qiagen; Valencia, CA). All patients were genotyped for the H63D and C282Y HFE polymorphisms by restriction fragment length analysis as previously reported.17
Multiplex cytokine bead assay.
We performed multiplex analysis on undiluted CSF supernatants using the Bio-Plex Human 27-plex panel of cytokines and growth factors (Bio-Rad; Hercules, CA). The proteins in this panel as well as additional analytes are listed in table 2. Briefly, 1% bovine serum albumin (BSA) (Sigma-Aldrich; St. Louis, MO) was added to 200 μL of each CSF sample and standards were reconstituted in PBS with 1% BSA. Fifty μL of each sample or standard was added in duplicate to a 96-well filter plate and mixed with 50 μL of antibody-conjugated beads for 1 hour at room temperature. After 1 hour, wells were washed and 25 μL of detection antibody was added to each well. After a 30-minute incubation, wells were washed and 50 μL of streptavidin-PE was added to each well and incubated for 10 minutes. A final wash cycle was then completed and 125 μL of assay buffer was added to each well. The plate was then analyzed using a Bio-Plex 200 workstation (Bio-Rad). Analyte concentration was calculated based on the respective standard curve for each cytokine.
Table 2 All markers screened in our panel of potential biomarkers
Immunoassays.
CSF levels of β2-microglobulin (US Biologic; Swampscott, MA) and transferrin (Bethyl Laboratories; Montgomery, TX) were assayed by ELISA according to the manufacturers’ protocols.
Atomic absorption spectroscopy.
The amount of iron in the CSF was determined by digesting the CSF in ultrapure Nitric Acid (JT Baker, 9598-00; Phillipsburg, NJ), 1:4 v/v, and samples were heated to 60°C for 24 hours. The digested samples were diluted 1:100 in ddH2O, and then analyzed on a Perkin Elmer Atomic Absorption Spectrometer 600 series (Waltham, MA).
Statistical analysis.
Multifactorial analysis of analyte expression was performed using GeneSpring GX version 7.3.1 (Agilent Technologies; Santa Clara, CA). Normal distribution of analyte expression was assessed with the Kolmogorov-Smirnov test using SigmaStat 2.03 (SPSS, Inc.; Chicago, IL). Biomarkers were compared between groups via t test or Mann–Whitney U test, as appropriate, and differences were considered significant if p < 0.05. Correlations between markers and ALSFRS-R, duration of symptoms, and age were assessed by the Spearman or Pearson correlation coefficient, as appropriate, with p < 0.05 considered significant. Classification statistics were generated using a support vector machine algorithm with radial basis and polynomial dot product kernel functions and a diagonal scaling factor ranging from zero to five. Classification of each sample by disease status was determined by crossvalidation in a leave one out strategy in which each sample was sequentially blinded. Each sample is individually removed from the sample set and the algorithm seeks to classify the sample compared to the others.
RESULTS
CSF was obtained from 39 patients with sporadic ALS and two patients with familial ALS for a total of 41 patients with ALS (9 with bulbar onset, 32 with limb onset). CSF was also obtained from 33 neurologic disease controls. The diagnoses for these neurologic disease control patients are listed in table e-1 on the Neurology® Web site at www.neurology.org. The characteristics of the two groups are given in table 1. All biomarkers on the panel were detectable in the CSF of the majority of the samples, with the exception of IL-1β, which was not detected in any sample. Expression levels of analytes that increased or decreased with increasing age of subjects were age-adjusted by linear regression (table e-2). The expression of two biomarkers differed by sex (table e-2).
Between patients with ALS and neurologic controls, the expression of 11 biomarkers was significantly higher in the CSF of patients with ALS, while the expression of two biomarkers was significantly higher in controls, as shown in table 3. A support vector machines algorithm was used to classify each sample as ALS or control in a leave one out crossvalidation strategy with an accuracy of 86.5% using all assayed markers. The best accuracy was obtained using only the markers with the five most significant p values (IL-10, IL-6, GM-CSF, IL-2, and IL-15) with a first order polynomial dot product kernel function and a diagonal scaling factor of one. Limiting the analysis to the five most significant biomarkers enabled us to distinguish patients with ALS from controls with 89.2% accuracy, 87.5% sensitivity, and 91.2% specificity.
Table 3 Biomarkers significantly different between patients with amyotrophic lateral sclerosis (ALS) and neurologic disease controls by t test or Mann-Whitney U test, as appropriate
No markers differed significantly between patients with ALS with limb onset vs bulbar onset or significantly correlated with duration of symptoms. The expression of IL-8, however, was higher in those patients with lower ALSFRS-R scores, an indicator of disease progression (table e-3).
Fourteen patients with ALS carried an H63D HFE variant (3 homozygotes, 11 heterozygotes), and 7 control patients (1 homozygote, 6 heterozygotes) carried an H63D HFE variant. Both the ALS patient group and the neurologic disease control patient group had five subjects carrying C282Y HFE variants (one control homozygote, all others heterozygous). Heterozygotes and homozygotes were grouped together because of the relatively small number of homozygotes. Additionally, the two disease groups were considered together to determine the association between biomarker expression and HFE genotypes. Biomarkers differing by HFE genotype are given in table 4. There were no significant differences of iron concentration in the CSF for any of the comparison groups (data not shown).
Table 4 Biomarkers were stratified by HFE genotype at the 63rd amino acid position and the 282nd amino acid position
DISCUSSION
The results of this study support our hypotheses that the profile of inflammatory and anti-inflammatory cytokines, growth factors, and analytes reflective of iron homeostasis differs between patients with ALS and a group of neurologic disease control patients. It has been proposed that ALS results from triggering events causing a cascade leading toward selective motor neuron death in genetically susceptible individuals.22 This study takes a novel approach toward the development of a panel of CSF biomarkers at one point during the ALS disease process. Our approach does not clarify whether inflammatory processes precede disease onset or result from it, but it does reveal inflammatory activity early in the disease process. No markers differed between patients with ALS with limb vs bulbar onset suggesting similar pathologic processes irrespective of the site of disease onset.
Each of the five most significantly different cytokines between patients with ALS and neurologic disease controls can be linked to disease pathogenesis. One marker increased in the neurologic disease control group compared to patients with ALS was IL-10, which is an anti-inflammatory cytokine known to inhibit microglial activation.23 IL-6 is considered to be an inflammatory cytokine, but is also known to have anti-inflammatory and neurotrophic properties.24 GM-CSF is regarded as a proliferator and differentiator of neutrophilic, eosinophilic, and monocytic cellular lineages.25,26 IL-2 is produced mainly in T lymphocytes and acts to stimulate growth and cytotoxicity of activated T lymphocytes.27 IL-15 is a T lymphocyte chemoattractant also shown to activate microglia28 and upregulate the production of other chemoattractants including IL-8 and MCP-1.29 Indeed, expression of MCP-1 as well as MIP-1α and MIP-1β were all significantly increased in patients with ALS compared to controls.
This study is consistent with theories that suggest ALS pathogenesis involves inflammatory activation.30 The biomarker panel also is consistent with the overall concept that microglia recruitment and activation are key components of ALS pathogenesis.5,31-33 Activated microglia are a source of the leukocyte chemoattractant IL-834 and this was the only protein whose expression was related to ALSFRS-R. This relationship showed higher levels of IL-8 in the CSF of those patients with lower ALSFRS-R scores, consistent with evidence suggesting that T lymphocytes accumulate in regions of motor neuron loss.32 Accumulation of T lymphocytes can in turn lead to further recruitment and activation of microglia, and cytotoxic mediators produced by these cells may be detrimental to motor neurons.
Several studies have reported an increased risk of ALS in carriers of the H63D HFE polymorphism.17,18 The C282Y variant of the HFE gene has not been reported to increase risk of ALS although this rarer allelic variant may simply be underrepresented in the studies to date.18 Both of these common HFE gene variants increase cellular iron accumulation,35,36 but the phenotype of cells carrying the allelic variants is different.36 In vivo studies in humans and animals reveal that HFE variants are associated with loss of iron homeostasis, exacerbated inflammatory responses, and alterations in innate immunity,19 all of which may be considered part of the pathogenic process in ALS. The increase in β2M in the CSF associated with the H63D HFE allele may reflect increased turnover of the HFE-β2M complex or may be a marker of more general membrane protein turnover, particularly from immune cells.37 A study in a population of hemochromatosis patients showed that patients carrying the H63D HFE allele had higher plasma levels of the chemokine MCP-1 than patients homozygous for the C282Y variant, or wildtype controls.38 Consistent with previous studies,7 MCP-1 was elevated in patients with ALS compared to controls in our biomarker panel. MCP-1 expression, however, did not differ by HFE genotype. Only one of the markers, IL-8, trended toward an increase in those individuals carrying the H63D allele. In the current study, the expression of several markers was higher in subjects carrying a C282Y HFE allele (table 4), but an association of this allele with ALS or other neurodegenerative disease has not been established. Future studies incorporating larger numbers of subjects with HFE variants as well as subjects homozygous for the H63D and C282Y alleles will be necessary to further define the biomarker profiles associated with these variants.
The biomarker panel distinguished ALS cases from neurologic disease controls, consistent with previous proteomic studies.3,8 Our data expand the previous studies and show promise that a biomarker panel could be developed for aiding in the diagnosis of ALS. Future validation studies should incorporate samples from multiple clinics and from patients with syndromes that mimic ALS. To establish a predictive value for this biomarker panel, serial analysis of individual patients may provide a better assessment of biomarkers reflecting disease progression.
ACKNOWLEDGMENT
The authors thank all the patients who participated in this study. The authors thank Scot Kimball, PhD, and Lydia Kutzler for their technical expertise and Dave Mauger, PhD, for statistical assistance.
Footnotes
-
Supplemental data at www.neurology.org
Editorial, page 11
e-Pub ahead of print on November 5, 2008, on www.neurology.org.
Supported by funds from The Judith and Jean Pape Adams Charitable Foundation, The Paul and Harriett Campbell Fund for ALS Research, The Zimmerman Family Love Fund, and the ALS Association Greater Philadelphia Chapter.
Disclosure: The authors report no disclosures.
Received January 7, 2008. Accepted in final form May 21, 2008.
REFERENCES
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.