Search for new biomarkers of pediatric multiple sclerosis: application of immunoglobulin free light chain analysis

Patients and samples

CSF and serum samples, collected from 64 pediatric patients (age range 1–17 years), were stored at −20 °C temperature until used. The patient groups were defined as follows:

1. Patients with definite MS diagnosis (n=21);

2. Patients with other demyelinating neurological disorders (n=20), including (ADEM, n=3), neuromyelitis optica (NMO, n=2), clinically isolated syndrome (CIS, n=8), radiologically isolated syndrome (RIS, n=2), transverse myelitis (TMY, n=2), optic neuritis (ON) with the seropositivity to myelin oligodendrocyte glycoprotein (MOG+, n=1), Guillain-Barré syndrome (GBS, n=2);

3. Patients with non-demyelinating neurological diseases (n=15). These included cases with vasculitis (n=1), Susac syndrome (n=1), Rasmussen’s encephalopathy (RE, n=3, one of them electrical status epilepticus in sleep [ESES]), epilepsy (n=2), Hashimoto disease (n=1), migraine (n=3), cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL, n=1), Sydenham’s chorea (n=1), monoplegia of left upper extremity (n=1), non-MS (patient under observation, n=1);

4. Control group (negative control) of patients (n=8) with no evidence for demyelinating, inflammatory, autoimmune, or other well-defined neurological disease.

MS diagnosis was based on dissemination in time and space as demonstrated by clinical and MRI findings [3]. The study was approved by the institutional review board (Rabin Medical Center, 0104-15-RMC). All participants have signed the written informed consent forms.

Oligoclonality test

A standard procedure for detection of IgG bands was used by employing high-resolution agarose gel electrophoresis (Hydragel six CSF) coupled with immunofixation (Sebia, Moulineaux, France) [10], [11].

FLC analysis by Western blotting

CSF and serum samples were obtained from each patient participating in this study. Preparation of CSF and serum samples for Western blot analysis was performed as described in [11] with some modifications. CSF sample was mixed with electrophoresis sample buffer (S.B.x5), 2:1 (vol/vol). Serum sample was diluted with phosphate buffered saline (PBS), 1:40 (vol/vol), and the diluted serum sample was mixed with S.B.x2, 1:3 (vol/vol). CSF and serum samples were run on 10%–20% Nu-Sep Tris-Tricine gels (Gradipore, Frenchs Forest, Australia) under non-reducing conditions and blotted onto nitrocellulose (Schleicher and Schuell, Dassel, Germany). Rabbit antibodies to human Ig κ and λ light chains (DAKO, Carpinteria, CA, USA) were used as primary antibodies. Peroxidase conjugated goat anti-rabbit IgG antibodies were used as secondary antibody (Sigma-Aldrich, Israel). Super-Signal West Pico Chemiluminescent Substrate (Pierce, Rockford, IL, USA) was used to visualize immunoreactive bands on X-ray films (FujiFilm Corp., Tokyo, Japan) at exposure times from 5 to 50 s.

Software [10] was used for quantitative evaluation of the intensity of FLC immunoreactive bands. The obtained intensity value in the tested CSF sample was normalized relative to that in the control sample: intensity of FLC band of positive control divided by intensity of FLC band of tested sample. The positive control sample, representing a mixture of CSF samples obtained from 15 patients with definite MS diagnosis, was included in each electrophoretic run alongside the tested samples. The following FLC indices (I) were calculated:

1. I₀ and I₁ – CSF λ and κ monomer level indices, respectively,

2. I₂ – CSF monomeric κ/λ ratio index,

3. I₃ – index to compare the κ/λ monomer ratio in the CSF vs. that in the serum (κ/λ_SERUM/κ/λ_CSF),

4. I₄ – CSF λ dimer level index,

5. I₅ – CSF λ dimerization index: λ dimer/monomer ratio (λD/M),

6. I₆ – index to compare the λ dimerization in the CSF to that in the serum (λD/M_SERUM/λD/M_CSF).

Of importance, I₀, I ₁, I₂, I₄ and I₅ indices are calculated using normalized intensity values, while I₃ and I₆ are calculated without normalization. (Notably, the lower index values point to the higher FLC levels, higher κ/λ ratios and higher λ dimerization values.)

Statistical analysis

Fisher’s exact two-tailed test was applied to compare the three distinct MS patient subgroups (characterized by the differences in their FLC monomer-dimer patterns; see Results) with respect to the clinical and MRI data.

FLC patterns in the MS patients

Analysis of the CSF-serum samples from control individuals and MS patients revealed striking differences between these two groups in relation to their FLC monomer-dimer (M-D) patterns in the CSF. The control individuals (n=8) demonstrated a weak or no immunoreactivity in the area corresponding to FLC monomers (25 kDa) and dimers (50 kDa) in the CSF. In contrast, 19 out of 21 MS patients showed significantly increased levels of κ-FLC monomers and dimers, and/or increased levels of λ dimers (Figure 1).

Figure 1:

Western blot analysis of the FLC monomer-dimer patterns in MS patients (B, C, D) and control individual/negative control (A).

FLC bands were detected using anti-κ (a.κ) and anti-λ (a.λ) antibodies. 1-CSF positive control (MS), 2 and 3 – the tested CSF and serum samples, respectively. (B–D) – MS patients with “κ-type”, “λ-type” and “mixed type” FLC patterns, respectively.

To establish the clear-cut FLC criteria allowing the differentiation of MS from other non-MS neurological diseases, we determined quantitatively the intensities of immunoreactive FLC bands which were then used to calculate the FLC indices (I) as already described. The calculated FLC indices of pediatric patients with the clinically definite MS diagnosis are displayed in Table 1. The 10 MS cases (patients #1–10) showed high intensity of immunoreactive κ-FLC bands (I₁ mean=1.12±0.96) together with the relatively low immunoreactivity of λ monomers (I₀ mean=9.21±8.71) and dimers (I₄ mean=8.29±7.87). Also, κ/λ ratio of monomeric FLC in the CSF was significantly higher than that in the serum of the same patient (I₃ mean=0.19±0.13) (“κ type” FLC pattern, Figure 1B). In six MS cases (patients #11–16), the major abnormality was manifested by high intensity of dimeric λ-FLC bands (I₄ mean=0.85±0.67); in these cases the value of λ dimerization was much higher in the CSF than that in the serum (I₆ mean=0.09±0.007). Of note, the κ monomer levels varied widely (I₁ from 0.67 to 18) among these five patients (“λ type” FLC pattern, Figure 1C). Three MS cases (patients #17–19) showed high levels of κ FLC (monomers and dimers) and λ dimers (“mixed type” FLC pattern, Figure 1D). In all these MS cases the immunoreactivity of λ monomer bands was relatively weak (I₀ mean=9.21±8.71).

Based on these calculations, the FLC criteria were established to support the diagnosis of pediatric MS (conditions A or B):

1. In case of MS (conditions A or B), I₀≥0.2 and I₁<20.

2. A: I₁≤3 and I₂≤1 and I₃≤0.7;

3. B: I₄≤0.8 or I₄≤3 and I₅≤2 and I₆≤0.05.

The obtained results were also helpful to determine three different pathological patterns of FLC observed in MS:

1. “κ type” FLC pattern: only conditions A are satisfied,

2. “λ type” FLC pattern: only conditions B are satisfied,

3. “Mixed type” FLC pattern: both conditions (A and B) are satisfied.

FLC patterns in the non-MS patients

In patients with non-MS neurological diseases, the FLC patterns of 31 out of 34 cases were different from those observed in MS (Table 2). In two out of seven CIS cases the immunoreactivity of FLC monomer and dimer bands in the CSF was very weak as compared to MS (patients #24 and #26). In other two CIS patients (#23 and #25) the κ-FLCs were not detectable, while the λ-FLC bands were very weak (Figure 2A, case 23, track 4). The CIS patient #27 showed only slight increase in FLC levels, which was incompatible with MS (I₁=10.5 and I₄=6.5). In one of seven CIS cases (patient #22), κ monomer level was significantly increased (I₁=1.96), but in contrast to MS, no abnormalities were found by comparing the κ/λ monomer ratio in the CSF versus that in the serum (I₃=1.63). It is remarkable, that this patient showed seropositivity to myelin basic protein (MBP). The CIS patient #28 showed FLC pattern typical of MS; of note, this patient converted to MS during the 1st year of the patient follow-up.

Figure 2:

Western blot analysis of the FLC monomer-dimer patterns in the non-MS patients.

FLC bands were detected using anti-κ (a.κ) and anti-λ (a.λ) antibodies. (A) 1 – positive control (MS, CSF); 2 and 3 – TMY (patient #35) CSF and serum, respectively; 4 and 5 – CIS (#23), CSF and serum, respectively; 6 and 7 – Sydenham’s chorea (#50) CSF and serum, respectively. (B) 1 – positive control (MS, CSF), 2 and 3 – RE (#42) CSF and serum, respectively, 4 and 5 – NMO (#32) CSF and serum, respectively.

In ADEM, each of the three tested cases showed different FLC patterns, but none of the patterns resembled MS. The abnormally high levels of κ monomers were observed in two NMO cases: I₁=0.71 (patient #32) (Figure 2B, track 4), and I₁=1.09 (patient #33). However, the other FLC indices in these NMO patients showed no abnormalities typical of MS (Table 2). One patient with vasculitis (# 44, Susac syndrome) displayed the FLC pattern typical of MS. In the second vasculitis patient (#45), κ-FLCs were not measurable and only slight increase of λ dimer level was observed. Also, in one of the two RIS cases (patient #36) FLC pattern was close to normal, while the second RIS patient (#37) showed the MS-like FLC pattern (this patient had a family history of MS).

Taken together, the FLC patterns in seven out of 35 non-MS cases were close to normal (i.e. similar to the negative control): CIS (n=2), RE (n=1), epilepsy (n=1), monoplegia of left upper extremity (n=1), migraine (n=1), non-MS case under observation (n=1). Mild/moderate abnormalities of FLC patterns were observed in other non-MS patients (n=16), which included patients with CIS (n=3), TMY (n=2), GBS (n=2), RIS (n=1), migraine (n=1), Sydenham’s chorea (n=1), Hashimoto disease (n=1), ADEM (n=1), ON & MOG+ (n=1), RIS (n=1) and CADASYL (n=1). Nine out of 35 non-MS cases demonstrated highly abnormal FLC patterns which, however, differed from the FLC patterns in MS: ADEM (n=2), CIS (1), RE (n=2), epilepsy (n=2), NMO (n=2). In three non-MS cases (Susac syndrome [n=1], CIS [n=1] and RIS [n=1]) the FLC patterns resembled those of MS.

FLC pattern analysis and oligoclonality test versus clinical diagnosis

The FLC diagnostic criteria were applied to support or disfavor the MS diagnosis (Tables 1 and 2). We found that in 19 out of 21 patients with clinically definite MS, the results of our FLC analysis supported this diagnosis (sensitivity 90.5%). The results of the oligoclonality test showed that 17 cases were oligopositive, three cases remained inconclusive, and one case was oligonegative (sensitivity 81%).

Our FLC criteria were also applied for the non-MS patient group: 32 out of 35 non-MS cases tested disfavored the MS diagnosis (specificity 91.4%). The oligoclonality test rejected the MS diagnosis in 23 out of 35 cases, in four cases the results were inconclusive, and eight cases were oligopositive (specificity 65.7%).

FLC pattern analysis in MS versus MRI and clinical symptoms

Comparison of the FLC data with those of MRI and clinical assessment at the time of MS diagnosis (Table 3) showed that “mixed type” and “λ type” patients demonstrated a more active onset of the disease as compared to the “κ type” MS patients. Fisher’s exact two-tailed test showed significant differences when comparing the MS patient group with ”mixed plus λ types” versus that with “κ type” with respect to:

1. amount of brain lesions with a T2 number load >20, p=0.015;

2. amount of active brain lesions with a number of gadolinium enhanced lesions (Gd+)>3, p=0.003;

3. presence of lesions in spinal cord, p=0.009.