The diagnosis of neurological diseases will become increasingly important in the future, so that valid biomarkers are also of interest in this field. This moves the focus primarily on diseases that are bound to increase due to demographic changes (neurodegenerative diseases like Alzheimer’s disease or vascular diseases like cerebral infarction). These diseases require biomarkers that are associated with the respective condition in order to secure a diagnosis early on and detect any progression. Particularly needed are biomarkers that capture the effect of new drugs and therapeutic strategies.
Not long ago one model of the pathogenesis of Alzheimer’s disease was generally accepted. According to this model, neuropathological changes typical of the disease and thus dementia existed, or both were absent. But current studies now confirm the view that neuropathological changes manifest decades before the onset of initial clinical symptoms. In a recently published review article, Jack et al.  presented a hypothetical model that correlates the changes over time in the biomarkers (biochemically and/or via imaging) with clinically defined stages of the disease. Under this model, there are, initially, pathological changes to β-amyloid (β-amyloid 1–42 in cerebrospinal fluid and/or β-amyloid PET imaging), followed by those to the neurodegenerative markers (total tau and phospho-tau protein in CSF and/or fluorodeoxyglucose-PET and cerebral atrophy in MRI; Figure 1). The latter correlate with the clinical severity of the disease. Recently published data indicate that familial Alzheimer’s disease is due to an overproduction of Aβ42, while sporadic Alzheimer’s disease is due to a reduced clearance of Aβ42 .
The diagnosis of Alzheimer’s disease is currently based primarily on the verification of dementia and other specific clinical symptoms. To improve the diagnostic precision and an earlier diagnosis, the determination of biomarkers in the CSF has been proposed as a correlate of neuropathological changes that define Alzheimer’s disease, namely, total tau/phospho-tau protein and β-amyloid.
German S3 guideline “Diagnosis and treatment of dementia” issued by the German Association of Neurology (DGN)
First of all, an examination of blood parameters is generally recommended, due to the low risk for people with dementia and low cost, to detect a possibly present reversible cause of dementia. Examples of possible causes of dementia include endocrinopathies, vitamin deficiency diseases, metabolic encephalopathies, intoxication, electrolyte disturbances, hematologic disorders, chronic infectious diseases, and late forms of leukodystrophies, such as ceroid lipofuscinosis.
As part of the basic diagnostic assessment, the following serum or plasma tests are recommended: blood count, electrolytes (Na, K, Ca), fasting blood glucose, TSH, erythrocyte sedimentation rate or CRP, GOT, gamma-GT, creatinine, urea, vitamin B12 (recommendation 11).
In the case of clinically unclear situations or in specific suspected diagnoses, targeted further laboratory tests should be performed. Examples are: differential blood count, blood gas analysis, phosphate, HBA1c, homocysteine, fT3, fT4, thyroid antibodies, cortisol, parathyroid hormone, ceruloplasmin, vitamin B6, Borrelia serology, Pb, Hg, Cu, syphilis serology, HIV serology, drug screening, urine test strips, folic acid (recommendation 12).
CSF analysis, too, can provide information on non-degenerative causes of dementia, in which a patient’s history, physical findings and other technical additional tests exhibit no pathological findings (recommendation 15). For the purposes of an initial diagnosis, it supports the differentiation between primary neurodegenerative dementia conditions and other causes of dementia syndromes. In particular, one should mention here the exclusion of an inflammatory disease of the brain if so indicated by the patient’s medical history, physical findings or additional diagnostic testing (recommendation 14).
If a CSF analysis is carried out in connection with dementia, it is the parameters of the basic CSF profiles that are to be examined (recommendation 15), i.e., cell count, total protein, lactate concentration, glucose, albumin quotient, intrathecal IgG production and oligoclonal banding. If clinically indicated, it can be useful additionally to analyze the intrathecal IgA and IgM production.
With regard to the actual neurochemical dementia diagnostics, all CSF and serum samples should be sent to the lab, unfrozen, as quickly as possible. Ideally, polypropylene tubes should be used, because otherwise β-amyloid 1–42, total tau and phospho-tau may be lost. According to current knowledge, CSF-based neurochemical dementia diagnostics, in the context of an initial diagnosis, support the differentiation between primarily neurodegenerative dementia conditions and other causes of dementia syndromes (recommendation 16). The combined analysis of the parameters β-amyloid 1–42 and total tau and/or β-amyloid 1–42 and phospho-tau is superior to the analysis of only one parameter and is therefore recommended (recommendation 17). By contrast, an isolated analysis of the apoliprotein E genotype as a genetic risk factor is not recommended for diagnostics due to the lack of diagnostic accuracy and predictive significance (recommendation 13).
However, the differential diagnostic selectivity of these markers within the group of neurodegenerative diseases, and in contrast to vascular dementia, is inadequate (recommendation 18), so that the results of CSF-based neurochemical dementia diagnostics should be assessed only on the basis of the findings of a routine analysis of cerebrospinal fluid and all other diagnostic information that is available (recommendation 19). The parameters mentioned are not suitable as progression parameters, according to current knowledge.
Revised NIA-AA criteria for the diagnosis of Alzheimer’s disease
The recently published revised NIA-AA (National Institute on Aging-Alzheimer’s Association guidelines for the neuropathologic assessment of Alzheimer’s disease) criteria [3–5], according to the above-mentioned model of Alzheimer’s disease , makes both a semantic and a conceptual difference between the pathophysiology of Alzheimer’s disease (AD-P) and the various resulting clinical symptoms (AD-C). Only the five most common biomarkers studied in the literature have been included formally in the diagnostic criteria. It should be noted that the term “biomarker” is increasingly being used in connection with Alzheimer’s disease not only in relation to measured parameters in body fluids, but also in terms of imaging. According to the revised criteria, the biomarkers can be divided into two categories: on the one hand, biomarkers of Aβ accumulation, i.e., an abnormal retention of tracers in imaging as well as low levels of Aβ42 in CSF, and on the other hand, biomarkers of neuronal degeneration and neuronal damage, i.e., elevated levels of tau protein in CSF or a reduced intake of fluorodeoxyglucose (FDG) in positron emission tomography (PET) . As already mentioned above, the view is gaining ground that biomarkers of Aβ accumulation become abnormal approx. 10–20 years before the occurrence of clinical symptoms, while biomarkers of neurodegeneration exhibit dynamic changes at a later time. Some studies have suggested that this is only the case immediately before the onset of clinical symptoms, and the progression of these symptoms go hand in hand with a “deterioration” of the relevant biomarkers.
Approximately 30% of cognitively inconspicuous elderly patients show evidence of AD-P, and many of these individuals meet the neuropathological criteria for Alzheimer’s disease, but do not have obvious cognitive symptoms [6–8]. These 30% are a perfect fit for the prevalence, observed in various studies, of amyloid positivity in CSF diagnosed both by means of imaging and neurochemistry in cognitively normal individuals over 65 years [8–13] and roughly correspond to the prevalence of Alzheimer’s disease observed a decade later .
The biomarker of Aβ accumulation, as already mentioned, is a reduced Aβ42 in CSF and a positive PET. Among the biomarkers for neuronal damage, tau and phospho-tau are generally deemed equivalent, but phospho-tau is considered more specific to Alzheimer’s disease compared to other dementia conditions. It is noteworthy that for clinical purposes, the use of biomarkers is currently not recommended. The reasons cited for this are that 1) the clinical criteria for diagnostics have sufficient accuracy, 2) the design of the criteria that include the use of biomarkers need to be scrutinized more closely, 3) there is still room for improvement with respect to the standardization of biomarkers, and 4) biomarkers are not equally available everywhere. At this time, they should essentially be used for studies, and in clinical routine only if available and deemed suitable by the clinician. The test results for the biomarkers can be classified in three categories: clearly positive, clearly negative and indeterminate.
A certain contradiction to the revised criteria of Alzheimer’s disease comes in the form of the clearly formulated skepticism about the use of biomarkers and their downgrading vis-à-vis clinical criteria. However, the proposed diagnostic classification of Alzheimer’s-related dementia and MCI (mild cognitive impairment) is impossible without a determination of biomarkers.
Pre-analytical and analytical issues in the determination of biomarkers in cerebrospinal fluid
How justified this reservation about the use of biomarkers is with respect to the lack of standardization is borne out by the fact that, even when identical assays were used, substantial variability was observed in connection with the absolute concentration levels of biomarkers between different centers, resulting in significantly different cut-off values . Very impressive are the results of a study published last year that dealt with the influence of the material of the sample tubes on the measurement results [16, 17]. The concentrations of tau protein, phospho-tau protein and Aβ1–42 were investigated in dependence on the sample tubes used, for a total of 11 different types of tubes. There were significant differences between the different types of tubes with maximum deviations from the medians of –48% to +31% for Aβ1–42, –8% to +8% for tau protein and –4% to +6% for phospho-tau protein (Table 1). While the results for the two latter biomarkers are in a range of still acceptable analytical imprecision, the results for Aβ1–42 clearly fall outside this range. The complementary physical analysis of the composition of the sample tubes yielded the surprising result that only one of eleven sample vessels declared to be propylene tubes was in fact made from pure propylene. The other ten were copolymers with at least an admixture of polyethylene. What was also surprising was that the type of tube made from pure propylene did not yield the best results at all. This suggests that an additional surface treatment influences the adsorption properties of the sample tube significantly.
The high pre-analytical sensitivity of Aβ1–42 for different polymer/plastic surfaces can therefore lead to reductions in the measured concentrations by 20%–60%, making it necessary to account for this type in future consensus protocols. Ultimately, this means that different types of tubes must be considered with different pre-analytical protocols and cut-off values. One must also consider in this context that in previous published studies not some of the test subjects designated as “controls” may in fact have been persons without cognitive abnormalities, but with already existing Alzheimer’s pathologies .
A corresponding approach to solving the above problems is found in the recently published recommendations of the “Alzheimer’s Biomarkers Standardization Initiative (ABSI)” , which for the first time consider all possible pre-analytical and analytical aspects of determining the biomarkers of Alzheimer’s disease. They are summarized in Table 2.
Concerning the problem of surface adsorption to the wall of the sample tube, also identified therein as a “key issue”, it is recommended that sample tubes with the smallest possible volume be used, which should be filled to at least 50%. In addition, each laboratory should always use identical polypropylene tubes, and under no circumstances use glass or polystyrene tubes. Some findings in a study mentioned in the recommendations, but not yet published, could be significant for the future. According to this, pre-treatment of the sample material with non-ionic detergents, such as Tween-20, prevented the adsorption of β-amyloid to the walls of the sample tube .
Even though the ABSI does not (yet) solve the problem of definitive cut-off values, it represents an important approach with practical and helpful implications: for the first time, comprehensive recommendations are given for the pre-analysis and analysis in connection with the determination of the biomarkers of Alzheimer’s disease. It is to be hoped that this is a first step towards a much needed standardization and that it will soon yield a certified quality control sample, which is not yet available.
Over the past two decades, several new biomarkers of Parkinson’s disease have been proposed. To date, however, there is not one sufficiently validated biomarker. The Parkinson Progression Marker Initiative (PPMI) is an international multicenter study, supported by regulatory authorities, foundations, industry and academic institutions, to identify biomarkers of the progression of Parkinson’s disease. On the one hand, they are to create a better understanding of the etiology and, in particular, progression, while on the other hand, they are to provide important tools to improve the probability of therapeutic success . The PPMI was launched in 2010 and will ultimately include a total of 400 newly diagnosed cases of Parkinson’s disease and 200 healthy test subjects: over a period of 5 years, biomarkers obtained from clinical, imaging and biological materials (CSF, serum, urine) will be captured by means of standardized protocols. The tapping of cerebrospinal fluid is justified on the grounds of promising results of individual studies, particularly with respect to α-synuclein. Inclusion of a Parkinson’s patient in the study depends on dopamine transporter scintigraphy. At present, 80% of patients and/or control subjects are already included in the study. Also participating are 21 study centers in the USA (16) and Europe (5).
All data are recorded in the PPMI database and quickly posted to the PPMI website (www.ppmi-info.org). In addition, biological materials will also be shared with external scientists if their proposed projects are approved by an independent panel of experts.
As examples of data obtained through PPMI and shared via the website so far, two studies will be discussed briefly: In the first one, fluctuations throughout the day and inter-assay variabilities were analyzed through repeated measurements of α-synuclein, DJ-1 (PARK-7 gene product), Aβ1-42 and Aβ1–40 in CSF (11 time points over 26 h after setting up a lumbar catheter early in the morning in 13 control subjects and 12 patients with Parkinson’s disease). The same was done again 10–14 days later for the control subjects, but not the patients. It was shown that all biomarkers exhibited constant values over the 2-week period examined. The sampling at multiple times over 26 h, however, demonstrated for several biomarkers both in control subjects and Parkinson’s patients a positive correlation with the duration and the total protein content in CSF. In the process, α-synuclein (Figure 2), Aβ1–42 and Aβ1–40 followed a similar pattern, while DJ-1 showed a largely constant progression (http://www.ppmi-info.org/2012/05/mds-presentation-on-ppmi-diurnal-and-intersubject-variability-of-cerebrospinal-fluid-biomarkers-in-parkinsons-disease-and-healthy-volunteers/).
Another study examined the association of baseline levels of biomarkers (Aβ1–42, t-tau and p-tau 181, α-synuclein) with clinical criteria in patients thus far untreated and who had recently been included. Overall, all biomarkers exhibited significantly lower values compared to the healthy control subjects. In addition to various associations of individual biomarkers with clinical scores, lower levels of α-synuclein in cerebrospinal fluid were associated with a significantly higher odds ratio of the diagnosis of Parkinson’s disease .
Although the previously obtained, still rather preliminary, results of PPMI have not yet led to the establishment of a practical biomarker in cerebrospinal fluid, this overdue approach represents a decisive step on the path to a successful (being comprehensive and well-standardized) strategy for implementing biomarkers of Parkinson’s disease, and should ultimately lead to improved therapeutic options.
Cerebral infarction (stroke)
Ischemic cerebral infarction
A useful biomarker for clinical use should meet these requirements: it should be detectable in the course of cerebral ischemia as early as possible, and it should be as specific to the brain as possible. However, it must be emphasized that due to the fact alone that even a clinically valid biomarker will hardly ever provide information on anatomical features, biomarkers can only supply supplementary information for other diagnostic procedures, including imaging. In a recently published meta-analysis of studies from 1996 to 2007, which examined various biomarkers in connection with cerebral infarction, however, not a single marker provided additional predictive power compared to existing and validated clinical models .
The task of a future biomarker for routine clinical use must be to identify high-risk patients who are in need of a rapid reperfusion strategy and who must be directed promptly to the appropriate centers. It must also contribute to the individualization of therapy through risk stratification of reperfusion hemorrhage, allow for the relative volume of the penumbra to be estimated, and should be able to provide additional prognostic information . An overview of previously examined biomarkers of cerebral infarction during the ischemic cascade is shown in Figure 3.
Instead of using only a single biomarker, it is conceivable, of course, to combine several such markers. In the study by Vanni et al., a panel of four biomarkers (BNP, D-dimer, MMP-9, S100B) was used in an emergency room for triage in connection with a suspected brain infarction. Even when these four biomarkers were combined by way of a multi-marker index (MMX), the diagnosis of an actual cerebral infarction was missed in about 25% of patients. Only once the MMX was combined with the Cincinnati Prehospital Stroke Scale (CPSS), which focuses on the key symptoms of a stroke that are present, was the diagnostic accuracy improved significantly . The same panel was used in another multicenter prospective study. This one showed for patients with acute symptoms a sensitivity of 86% for the diagnosis of an ischemic cerebral infarction and/or 94% for the diagnosis of a hemorrhagic stroke. This study, too, achieved improvement by adding the case history and/or clinical parameters of age, sex and atrial fibrillation. This result was validated in an independent cohort of patients .
Thus, at this point, one must unfortunately state (still) that there is no biomarker robust enough due to its discriminative properties for routine clinical use regarding the diagnosis and management of patients suffering a brain attack. Unlike, say, the heart, the brain poses particular challenges for laboratory diagnostics given its heterogeneity of different cell populations and varying tolerance for ischemia, the complexity of the ischemic cascade and the existence of the blood-brain barrier. Nevertheless, the future diagnostic use of biomarkers is of high clinical interest for three reasons: Accelerated triage and early decisions regarding the therapeutic approach, individualized treatment and identification of etiology and prognosis of the brain attack. In all likelihood, this will have much less to do with the use of a single biomarker than, rather, a combination of several biomarkers.
Hemorrhagic cerebral infarction
In contrast to the ischemic cerebral infarction, the literature on biomarkers of intracerebral hemorrhage is much sparser. A recent study on 60 patients with primary hemispheric intracerebral hemorrhage showed an increase in the so-called high-mobility group protein box-1 (HMGB1, a nuclear DNA-binding protein) in the serum within 12 h after onset of symptoms and a correlation with the severity of the hemorrhage . In another recent study, an ischemic attack was differentiated from a hemorrhagic attack by means of the biomarkers S100B and sRAGE (soluble receptor of advanced glycation end products) measured in the plasma by way of the appropriate ELISAs within 6 h of the onset of symptoms: here, S100B was significantly increased in the hemorrhagic stroke, while sRAGE showed a significant decrease .
For two reasons, however, the determination of biomarkers in the context of the hemorrhagic brain attack will likely not attain the same significance as it did for the ischemic cerebral infarction: first, imaging methods will certainly not be abandoned because they are needed to rule out intracerebral hemorrhage; second, the significance of biomarkers lies, after all, in the early diagnosis of the ischemic cerebral infarction so as to decide on a reperfusion strategy as quickly as possible, which would be prohibited in the case of intracerebral hemorrhage. At most, biomarkers could play a role in intracerebral hemorrhage in the future with respect to assessing the prognosis.
Conflict of interest statement
Authors’ conflict of interest disclosure: The authors stated that there are no conflicts of interest regarding the publication of this article.
Research funding: None declared.
Employment or leadership: None declared.
Honorarium: None declared.
Jack CR Jr., Knopman DS, Jagust WJ, Shaw LM, Aisen PS, Weiner MW, et al. Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol 2010;9:119–28.CrossrefWeb of ScienceGoogle Scholar
Mawuenyega KG, Sigurdson W, Ovod V, Munsell L, Kasten T, Morris JC, et al. Decreased clearance of CNS beta-amyloid in Alzheimer’s disease. Science 2010;330:1774.Google Scholar
Albert MS, DeKosky ST, Dickson D, Dubois B, Feldman HH, Fox NC, et al. The diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 2011;7:270–9.Web of ScienceCrossrefGoogle Scholar
Jack CR Jr., Albert MS, Knopman DS, McKhann GM, Sperling RA, Carrillo MC, et al. Introduction to the recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 2011;7:257–62.CrossrefWeb of ScienceGoogle Scholar
McKhann GM, Knopman DS, Chertkow H, Hyman BT, Jack CR Jr., Kawas CH, et al. The diagnosis of dementia due to Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 2011;7:263–9.CrossrefWeb of ScienceGoogle Scholar
Price JL, Morris JC. Tangles and plaques in nondemented aging and “preclinical” Alzheimer’s disease. Ann Neurol 1999;45:358–68.Google Scholar
Savva GM, Wharton SB, Ince PG, Forster G, Matthews FE, Brayne C. Age, neuropathology, and dementia. N Engl J Med 2009;360:2302–9.Google Scholar
Aizenstein HJ, Nebes RD, Saxton JA, Price JC, Mathis CA, Tsopelas ND, et al. Frequent amyloid deposition without significant cognitive impairment among the elderly. Arch Neurol 2008;65:1509–17.Google Scholar
Fagan AM, Mintun MA, Mach RH, Lee SY, Dence CS, Shah AR, et al. Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid Abeta42 in humans. Ann Neurol 2006;59:512–9.Google Scholar
Jack CR Jr., Lowe VJ, Senjem ML, Weigand SD, Kemp BJ, Shiung MM, et al. 11C PiB and structural MRI provide complementary information in imaging of Alzheimer’s disease and amnestic mild cognitive impairment. Brain 2008;131:665–80.Web of ScienceGoogle Scholar
Mintun MA, Larossa GN, Sheline YI, Dence CS, Lee SY, Mach RH, et al. [11C]PIB in a nondemented population: potential antecedent marker of Alzheimer disease. Neurology 2006;67:446–52.Google Scholar
Shaw LM, Vanderstichele H, Knapik-Czajka M, Clark CM, Aisen PS, Petersen RC, et al. Cerebrospinal fluid biomarker signature in Alzheimer’s disease neuroimaging initiative subjects. Ann Neurol 2009;65:403–13.Web of ScienceGoogle Scholar
Rowe CC, Ellis KA, Rimajova M, Bourgeat P, Pike KE, Jones G, et al. Amyloid imaging results from the Australian Imaging, Biomarkers and Lifestyle (AIBL) study of aging. Neurobiol Aging 2010;31:1275–83.CrossrefWeb of ScienceGoogle Scholar
Perret-Liaudet A, Pelpel M, Tholance Y, Dumont B, Vanderstichele H, Zorzi W, et al. Cerebrospinal fluid collection tubes: a critical issue for Alzheimer disease diagnosis. Clin Chem 2012;58:787–9.Google Scholar
Perret-Liaudet A, Pelpel M, Tholance Y, Dumont B, Vanderstichele H, Zorzi W, et al. Risk of Alzheimer’s disease biological misdiagnosis linked to cerebrospinal collection tubes. J Alzheimers Dis 2012;31:13–20.Web of ScienceGoogle Scholar
Snowdon DA, Kemper SJ, Mortimer JA, Greiner LH, Wekstein DR, Markesbery WR. Linguistic ability in early life and cognitive function and Alzheimer’s disease in late life. Findings from the Nun Study. J Am Med Assoc 1996;275:528–32.Google Scholar
Vanderstichele H, Bibl M, Engelborghs S, Le Bastard N, Lewczuk P, Molinuevo JL, et al. Standardization of preanalytical aspects of cerebrospinal fluid biomarker testing for Alzheimer’s disease diagnosis: a consensus paper from the Alzheimer’s Biomarkers Standardization Initiative. Alzheimers Dement 2012;8:65–73.CrossrefGoogle Scholar
Parkinson’s Progression Markers Initiative. MDS Presentation on PPMI-Diurnal and intersubject variability of cerebrospinal fluid biomarkers in Parkinson’s disease and healthy volunteers. Available at: http://www.ppmi-info.org/2012/05/mds-presentation-on-ppmi-diurnal-and-intersubject-variability-of-cerebrospinal-fluid-biomarkers-in-parkinsons-disease-and-healthy-volunteers/. Accessed: 23 Sept 2013.
Parkinson’s Progression Markers Initiative. MDS Presentation on PPMI-Association between CSF biomarkers and clinical phenotype of early Parkinson’s disease. Available at: http://www.ppmi-info.org/2012/05/mds-presentation-on-ppmi-association-between-csf-biomarkers-and-clinical-phenotype-of-early-parkinsons-disease/. Accessed: 24 Sept 2013.
Vanni S, Polidori G, Pepe G, Chiarlone M, Albani A, Pagnanelli A, et al. Use of biomarkers in triage of patients with suspected stroke. J Emerg Med 2011;40:499–505.Google Scholar
Laskowitz DT, Kasner SE, Saver J, Remmel KS, Jauch EC. Clinical usefulness of a biomarker-based diagnostic test for acute stroke: the Biomarker Rapid Assessment in Ischemic Injury (BRAIN) study. Stroke 2009;40:77–85.Google Scholar
Zhou Y, Xiong KL, Lin S, Zhong Q, Lu FL, Liang H, et al. Elevation of high-mobility group protein box-1 in serum correlates with severity of acute intracerebral hemorrhage. Mediators Inflamm 2010;2010:29.Web of ScienceGoogle Scholar
Montaner J, Mendioroz M, Delgado P, Garcia-Berrocoso T, Giralt D, Merino C, et al. Differentiating ischemic from hemorrhagic stroke using plasma biomarkers: the S100B/RAGE pathway. J Proteomics 2012;75:4758–65.Web of ScienceGoogle Scholar
Original German online version at: http://www.degruyter.com/view/j/labm.2014.38.issue-1/labmed-2013-0059/labmed-2013-0059.xml?format=INT. The German article was translated by Compuscript Ltd. and authorized by the authors.