The Multiple Sclerosis Cooperative Etiological Study in Italy: preliminary analysis of CSF findings.

295 newly diagnosed Multiple Sclerosis (MS) patients were investigated for interdependence of CSF abnormalities: leukocyte count, CSF/serum albumin ratio, CSF IgG index and intrathecal synthesis of oligoclonal IgG. Only 7% of patients had no CSF abnormality. The most frequent abnormal finding was the presence of intrathecal synthesis of oligoclonal IgG. Polyacrylamide isoelectric focusing appeared more sensitive than agarose electrophoresis for the detection of abnormal CSF fractions. CSF IgG index had low negative and high positive predictive value when compared with electrophoretic methods. Suspected MS had the lowest frequency of abnormal CSF parameters. Patients with oligoclonal CSF bands (OB+) did not differ in age, sex, disease duration or annual bout rate from patients without oligoclonal bands (OB-). CSF leukocyte count was higher in OB+ than in OB- patients. Low CSF cell count was a hallmark of progressive disease. In OB+ patients CSF cell count correlated with CSF IgG index during the first five years of disease and declined with increasing disease duration. In later disease phases CSF IgG index correlated negatively with CSF/serum albumin ratio. 5 years follow-up of this series of patients is in progress to assess the diagnostic and prognostic values of CSF abnormalities in Multiple Sclerosis.

Cerebrospinal fluid (CSF) parameters and clinical course of multiple sclerosis.

The separate correlations among CSF and clinical parameters (mononuclear cell count, assay of intrathecal IgG synthesis, assessment of blood-brain barrier permselectivity to albumin and alpha 2-macroglobulin, detection of oligoclonal intrathecal IgG isoelectric spectro-type; age, disease duration, disability, number of bouts) were investigated by multiple linear regression in 100 multiple sclerosis patients subdivided according to the course and the phase of the disease. Results indicate that the complexity of the oligoclonal CSF IgG isoelectric spectrotype increases, but the overall amount of intrathecal IgG synthesis decreases with increasing number of bouts or disease duration in relapsing remitting and in progressive courses respectively. The CSF mononuclear pleocytosis was low in progressive forms, and it appeared to be linked to the intrathecal IgG synthesis changes. The IgG synthesis rate was proportional to CSF pleocytosis, but a negative correlation between pleocytosis and complexity of CSF IgG spectrotype was found. About 40% of patients had a blood-brain barrier damage. The increased permeability to albumin was related to the number of previous bouts, whereas the CSF pleocytosis seemed to be accompanied by a loss of barrier selectivity to large serum molecules. High CSF IgG index, high mononuclear cell count and low number of prominent oligoclonal CSF IgG fractions characterize the less active or short-lasting disease. Low CSF IgG index with numerous, faint, abnormal IgG CSF bands, low mononuclear cell number and increased barrier permeability to albumin are associated with long-lasting or steady progressive disease.

Blood-cerebrospinal fluid barrier permeability to serum IgG subfractions and measurement of intrathecal IgG synthesis.

CSF/serum gradients of IgG subfractions separated by isoelectric focusing (IF) have been measured by high resolving laser densitometry. In patients with normal blood-CSF barrier permeability (N.25) and with barrier damage due to acute idiopathic polyneuropathy (N.15) and to medullary compression (N.17), the CSF/serum gradients of IgG subfractions were negatively correlated with their pI. This electrostatic selectivity appeared to be reverted in barrier damage due to acute meningoencephalitis (N.15). In a series of multiple sclerosis patients (N.31), the CSF/serum gradients of IgG subfractions lacking CSF oligoclonal bands have been used to assess the overall barrier permeability to serum IgG. All intra-BCB synthesized IgG subfractions could be measured by densitometry, whereas with other quantitative formulae, 23-26% of the results were false negatives; the total intrathecal IgG amount ranged from 0.01 to 11 mg/dl. The most frequent and prominent fractions appeared to be cathodic. Electrostatic and steric barrier selectivity must be taken into account when the amount of intrathecal IgG synthesis has to be measured.

Acute changes in blood-CSF barrier permselectivity to serum proteins after intrathecal methotrexate and CNS irradiation.

Ten children affected by acute lymphoblastic leukaemia without CNS involvement were treated with a CNS prophylaxis protocol. Intrathecal methotrexate and CNS irradiation (60Co) administered at different times both induced an increase in blood-CSF barrier permeability to serum proteins (albumin, IgG, alpha 2 macroglobulin). The relationship between permeability coefficients of proteins was analysed by theoretical porous or vesicular blood-CSF barrier models. The analysis indicated that both therapeutic procedures affect endothelial pinocytosis. An increase in radius of pinocytotic vesicles from 400 to 1500 A seemed the most relevant change. The damage of endothelial barrier permselectivity could be involved in acute and late delayed toxic effects of intrathecal methotrexate and of CNS irradiation.

Heterogeneous models for blood-cerebrospinal fluid barrier permeability to serum proteins in normal and abnormal cerebrospinal fluid/serum protein concentration gradients.

Cerebrospinal fluid (CSF)/serum concentration gradients (Q) of individual proteins (albumin, IgG, alpha 2-macroglobulin) have been studied in controls and in patients in whom the lumbar CSF flow is altered (medullary compression) or the blood-CSF barrier (BCB) function impaired (acute idiopathic polyneuropathy and acute meningoencephalitis). The analysis of relationships among protein Q has been performed by total and multiple regressions and the actual BCB permeability to individual proteins has been interpreted according to the accepted theoretical porous or vesicular BCB models. The exponential Q-IgG vs. Q-albumin total regression, and the poor Q-alpha 2-macroglobulin vs. Q-albumin regression found in controls, together with the different multiple regressions among proteins and the high Q-IgG vs. Q-albumin partial regression coefficients found in medullary compression, acute idiopathic polyneuropathy and acute meningoencephalitis, indicated that different permeability mechanisms can be postulated. Heterogeneous, fairly independent permeability BCB mechanisms maintain the normal CSF/serum protein concentration gradient. Pinocytotic vesicles or pores of radius exceeding 1000-1500 A, probably located at the capillary endothelium, account for the main serum-derived CSF protein fraction(s) with large hydrodynamic radius (R). A more selective endothelial vesicular transport with a radius of 250 A transfers a negligible amount of protein from serum into CSF. Proteins with small R also enter the CSF through a set of selective pores of radius 120 A, probably at the level of the choroidal epithelium. Pinocytotic vesicles with a radius of 250 A and increased rate of formation induce the accumulation of proteins below an obstruction of lumbar CSF flow. An increased formation rate of vesicles with a radius of 450 A can explain the increased capillary permeability in nerve roots in acute idiopathic polyneuropathy. Loss of selectivity was the main feature of BCB in acute meningoencephalitis, and it seemed to be due to pores or vesicles with a radius larger than 1000-1500 A. The heterogeneity of BCB mechanisms must be taken into account when the intrathecal synthesis of a protein, also derived from serum (for example IgG), has to be measured.