Immune cells after prolonged Natalizumab therapy: implications for effectiveness and safety


S. Marousi, Department of Neurology, ‘G. Gennimatas’ General Hospital of Athens, 154 Mesogeion Ave, 15669, Athens, Greece

Tel.: 0030 6972009400

Fax: 0030 2107470468




Previous studies on Natalizumab (NAT) have shown increased circulation of most white blood cells (WBC) in multiple sclerosis (MS) patients shortly after its introduction.


To describe peripheral immune cell phenotypes after more than 2 years of continuous NAT therapy and test for associations with clinical response to therapy.


Peripheral immune cell subsets were analyzed in 44 NAT-MS patients receiving NAT for over 24 months, and in 22 NAT-free control-MS patients.


NAT-MS patients displayed significantly higher numbers of all WBC when compared with controls. B lymphocytes exhibited a more pronounced increase when compared with CD4+, CD8+ and NK T-cells (P = 0.011). CD4/CD8 ratio was significantly decreased in NAT-MS patients (P = 0.018) and showed no correlation with the number of NAT doses. The reduced CD4/CD8 ratio was attributable to the EDSS improvement’ group only, irrespective of age, sex and disease severity.


The study suggests that there is no desensitization effect after prolonged NAT exposure. A reduced CD4/CD8 ratio was associated with long-term response to therapy; thus, those patients who most benefitted from the drug might be at greater risk for opportunistic infections like progressive multifocal leucoencephalopathy (PML). We provide implications for future research for the CD4/CD8 ratio as a possible contributor to the recently developed risk stratification scheme for PML.


Natalizumab (NAT) is a humanized monoclonal antibody against the α4 subunit (α4-integrin) of the very late antigen-4 (VLA-4) expressed on the surface of most white blood cells (WBC). NAT inhibits the transendothelial migration of autoreactive lymphocytes into sites of inflammation by blocking the adhesion to endothelial cells of the α4-integrin with the vascular cell adhesion molecule 1 (VCAM1) and the mucosal vascular addressin cell adhesion molecule 1 (MADCAM1). Thereby, it is currently used in the treatment of relapsing multiple sclerosis (MS) [1] and Crohn's disease [2].

Previous studies have described increased peripheral circulation of all the VLA-4-expressing blood cells and significant depletion of B and T lymphocytes from the central nervous system (CNS) during the first few months after induction to NAT therapy [1, 3-5]. The increased peripheral cell counts have been attributed to altered lymphocyte activation and trafficking into the secondary lymphoid organs whereas the opposite effect in the CNS has been thought to result from the direct inhibition of cell transmigration through the blood brain barrier [3, 6].

The widescale use of NAT during the last 6 years has been complicated by the occurrence of progressive multifocal leucoencephalopathy (PML), a severe opportunistic CNS infection caused by JC virus, initially described in immunocompromised HIV patients [7]. Recently, efforts have been focused on the identification of risk factors for NAT-related PML, and new biological markers are sought to most accurately predict the risk for PML in the individual patient [8, 9].

To date, it remains unknown what specific immune cell profiles are present after prolonged and uninterrupted delivery of NAT. Moreover, there are currently no available data regarding the possible correlations between NAT-induced peripheral immune cell changes and long-term clinical response to therapy. The present pilot study describes the specific peripheral immune cell phenotypes of MS patients after more than 2 years of continuous NAT therapy, and tests whether there is any significant immunological association with the clinical response to therapy.


Forty four consecutive MS patients treated monthly with NAT (NAT-MS) with mean 30 doses (range 24–40) and 22 NAT-free RMS patients (controls) were included in the study. Control-MS patients were previously exposed to common immunomodulatory drugs (i.e. interferons, glatiramer acetate) and were treatment-free during the last 3 months prior to blood sampling, before switching to NAT. Patients who had received steroid therapy during the last 2 weeks before blood collection were excluded from the study.

Peripheral whole blood (3 ml) from all patients was collected in the morning. Blood samples from NAT-MS patients were obtained just before the next drug infusion. Flow cytometry was performed on a FACS Calibur (Beckman Coulter Cytomics FC 500) using directly conjugated monoclonal antibodies (mAbs) from Beckman Coulter and DakoCytomation, against the surface markers CD3, CD4, CD8, CD19, CD16, CD56, CD14 and TCRαβ/γδ. mAbs were titrated according to manufacturer instructions. Flow cytometry measurements were held by a single investigator (G.P.) blinded to clinical data, with CXP software (Beckman Coulter), in the routine diagnostic laboratory of our hospital. As a measure of ‘response to treatment’ we considered ‘EDSS stability or improvement’ through the NAT treatment period.

Statistical analysis was performed using the PASW 18.0 software (IBM Corp., Armonk, NY, USA). All variables were found to be normally distributed after testing for normality with the Kolmogorov–Smirnov's test, and independent samples t-tests were applied to compare means between groups. Bonferroni's adjusted correction was applied to test for possible Type I statistical error. A binary logistic regression was used for adjusted comparisons, expressed by Odds Ratios (OR) with 95% Confidence Intervals (CI). Pearson's correlation coefficient expressed the relation between CD4/CD8 ratio and the number of NAT infusions.

The study was approved by the scientific and ethical committee of our hospital. All participants produced their informed consent at inclusion.


NAT-MS and control-MS patients did not significantly differ in age and sex distributions, or in disease severity measured by the EDSS. Baseline clinical characteristics of all studied groups are seen on Table 1.

Table 1. Baseline characteristics of study participants
 Control-MS patients (n = 22)All NAT-MS patients (n = 44)P1-valueResponding NAT-MS patients (n = 28)Non-responding NAT-MS patients (n = 16)P2-value
Stable EDSS NAT-MS patients (n = 15)Improving-EDSS NAT-MS patients (n = 13)
  1. SD, standard deviation; EDSS, Expanded Disability Severity Scale.

  2. P1-value compares between controls and NAT-MS patients (t-test for the comparison of means, Chi-square with Fisher's exact for the comparison of proportions).

  3. P2-value compares between stable-EDSS, improving-EDSS and non-responding NAT-MS patients (One-way ANOVA for the comparison of means, Pearson's Chi-square for the comparison of proportions).

Age, mean (SD)42.9 (9.5)45.5 (9.9)0.3343.5 (8.5)44.6 (10.8)49.8 (9.8)0.17
Sex, females (%)16 (73)26 (59)0.4210 (67)8 (62)8 (50)0.63
Baseline EDSS, mean (SD)4.7 (2.2)4.2 (2.0)0.363.97 (2.13)4.69 (1.89)4.17 (1.9)0.51
NAT doses, mean (range)30 (24–40)28.5 (3.5)30.4 (3.6)30.8 (4.0)0.22

The studied cell counts across the different categories of MS patients as well as the P-values for the comparisons between them are described in Table 2. All NAT-MS patients (either responding or non-responding) displayed significantly increased WBC, total lymphocytes, monocytes, eosinophils, basophils, CD4+T, CD8+T, NK-T and B lymphocytes when compared with controls. The two groups had similar counts in neutrophils, TCRγδ, dendritic and CD4-/CD8- T-cells. B lymphocytes exhibited a more pronounced increase (2.6-fold) when compared with CD4+ (1.4-fold), CD8+ (1.8-fold) and NK (1.9-fold) T cells (P = 0.011).

Table 2. Cell counts across the different categories of studied MS patients
Cell counts (normal range)NAT-MS patients (n = 44)Control-MS patients (n = 22)P1-valuesResponding NAT-MS patients (n = 28)P2-valuesNon-Responding NAT-MS patients (n = 16)P3-values
  1. NAT, natalizumab; MS, multiple sclerosis; WBC, white blood cells; SD, standard deviation; NK, natural killer cells; TCRγδ, T-cell receptor γδ cells; IQR, interquartile range.

  2. P1₁-values for the comparisons between NAT-MS and control-MS patients.

  3. P2- and P3 values compare between responding and non-responding NAT-MS vs control-MS patients, respectively.

Total WBC [3900–10200/mm3], mean (SD)8531 (2285)7005 (1485)0.0068537 (2554)0.0168522 (1798)0.007
Total lymphocytes [1000–3500/mm3], mean (SD)3211 (876)1904 (654)<0.0013115 (818)<0.0013379 (973)<0.001
Monocytes [150–600/mm3], mean (SD)656 (273)510 (136)0.021638 (251)0.036687 (312)0.023
Neutrophils [1800–7000/mm3], mean (SD)3983 (1348)4294 (983)0.3414083 (1558)0.5823809 (889)0.127
Eosinophils [10–600/mm3], mean (SD)419 (404)144 (92)0.002452 (413)0.004363 (246)<0.001
Basophils [20–85/mm3], mean (SD)65 (26)42 (18)<0.00167 (27)<0.00161 (24)0.009
CD4+ T3 lymphocytes [500–1500/mm3], mean (SD)1335 (429)979 (332)0.0011297 (417)0.0051402 (454)0.002
CD8+ T3 lymphocytes [250–1000/mm3], mean (SD)704 (287)398 (179)<0.001724 (268)<0.001670 (291)0.001
T4/T8 ratio, [2.2–2.6], mean (SD)2.1 (0.9)2.7 (1.1)0.0181.99 (0.77)0.0082.34 (1.06)0.282
NK-T lymphocytes (CD3+ CD16+ CD56+) [80–350/mm3], mean (SD)354 (178)186 (140)<0.001326 (140)0.001403 (227)0.001
B lymphocytes (CD19+) [100–500/mm3], mean (SD)698 (278)264 (145)<0.001655 (248)<0.001774 (317)<0.001
TCRγδ (αβ−γδ+) [20–250/mm3], median (IQR)28 (7–41)12 (7–25)0.09731 (7–40)0.09422 (7–51)0.284
Dendritic (CD14−CD16+) [0–70/mm3], mean (SD)38 (37)27 (26)0.18932 (24)0.42749 (53)0.092
CD4−CD8− T3 lymphocytes [<10% of total WBC], mean (SD)115 (94)75 (43)0.063107 (68)0.085129 (128)0.087

CD4/CD8 ratio was found to be significantly decreased in the NAT-MS group when compared with controls (P = 0.018) and showed no correlation with the number of NAT infusions (Pearson's = 0.09, P = 0.55). However, when separately analyzed responding or non-responding NAT-MS patients, this significant CD4/CD8 ratio reduction was present only in the responding NAT-MS group (P = 0.008) and persisted after adjustments for age, sex and baseline EDSS [OR (95% CI) = 0.44(0.20–0.99), P = 0.05] when compared with control-MS patients.

A further separate analysis of NAT-MS responders in those with ‘EDSS stability’ (n = 15) or ‘EDSS improvement’ (n = 13) revealed that the reduced CD4/CD8 ratio was attributable to the ‘EDSS improvement’ group only [mean (standard deviation) = 1.61(048)] and was significantly lower than the control value (P = 0.005). Application of the Bonferonni's correction (adjusted alpha value = 0.004) confirmed that this result truly showed at least a significant trend. The association was also present in the adjusted analysis for age, sex and baseline EDSS [OR(95% CI) = 0.09(0.01–0.59), P = 0.012].


The present study showed that prolonged (>24 doses) and uninterrupted treatment with NAT resulted in a significant increase in circulating lymphocytes, monocytes, eosinophils, basophils, CD4+T, CD8+T, NK-T and B-lymphocytes, but not in neutrophils, dendritic, TCRγδ and CD4/CD8-double negative T cells. This is the first study to describe specific peripheral immune cell profiles at a later stage of NAT treatment. Our results coincide with past reports of immune changes very early after NAT introduction [1, 3-5], add new data on specific cell subsets (dendritic, TCRγδ and CD4-/CD8-T-cells), and clearly suggest that there is no cell desensitization effect after repeated NAT exposure. The increased cell counts probably resulted from both altered VLA-4-expressing lymphocyte activation and restricted transendothelial migration of cells into secondary lymphoid organs due to the α4-integrin-VCAM1/MADCAM1 inhibition.

In the periphery, the CD4/CD8 ratio has been previously shown to gradually decrease during the first six NAT infusions, but to remain within normal limits (=1.2–2.6) [10]. The current results indicated that after 2 years of NAT exposure, the reduced CD4/CD8 was not associated with the number of drug infusions. The reduced CD4/CD8 ratio was the only parameter found to be significantly related with the long-term clinical response to therapy, defined as ‘EDDS stability or improvement’. A further analysis revealed that the reduction in CD4/CD8 ratio was associated only with the ‘EDSS improving’ and not with the ‘EDSS stability’ group, irrespective of age, sex and disease severity.

NAT treatment is de facto expected to block the transmigration of VLA-4 expressing autoreactive T cells in the CNS and is known to reduce the CD4/CD8 ratio in the CSF of MS patients to HIV(+) patient levels (<1.0) [10], who were the first patients in whom PML was identified as a detrimental CNS opportunistic infection [11]. In NAT-treated MS patients, the reduced serum CD4/CD8 ratio is the epiphenomenon of the disproportionate migratory inhibition of CD4+ vs CD8+ T cells by NAT, and reflects impairment in the T-helper immune surveillance. Thus, one might hypothesize that as low CD4/CD8 ratio NAT-MS patients present a sustained response to therapy like PML they might also be the ones in greater risk for CNS opportunistic infections. According to the recently reported risk stratification algorithm for PML, their risk would be critically defined by prior use of immunosuppressants, seropositivity for JCV antibodies and duration of NAT deliverance [9].

The finding that prolonged NAT use results in a disproportionate increase in B- lymphocytes when compared with CD4 + , CD8 +  and NK-T lymphocytes confirms with previous reports on shorter treatment periods [4, 5]. B cells are thought to be the natural reservoir for JC virus in the periphery [12] and probably carry it into the CNS, leading to PML [13]. This is possible because the blocking of VLA-4 expressing cells from NAT is not absolute, and there are still some cells getting into the CNS [3]. Consequently, it may be argued that a combination of impaired T cell immune surveillance (low CD4/CD8 ratio) and increased JC carrying B cells could critically augment the risk for PML, especially in responding patients.

Limitations of the study include the relatively small number of included patients, which may not allow for powerful assumptions. Nevertheless, all the past studies on immune profiles early after NAT introduction had been conducted in considerably fewer patients, about half our sample size [3-5, 10]. Probably, the CSF would be a more relevant biological compartment with regard to CNS immunocompetence; however, our intention was to trace a peripheral immune-related biomarker, indicative of the long-term clinical response to NAT treatment, which would be easy to access in the every day clinical practice. Moreover, one could argue that patients not responding to NAT might have developed neutralizing antibodies (Nabs) which were not considered in our study. Interestingly enough, persistence of Nabs against NAT has been described in only 6% of treated patients, especially those with documented early infusion-related reactions [14], and in clinical practice the utility of checking for Nabs remains controversial [15]. Notably, none of our patients had any infusion-related reaction, and the evidence from their altered immune cell profiles when compared with MS controls clearly pointed to an effective drug-integrin bonding.

Last but not least, one should stress the purely suggestive nature of our hypotheses regarding the association of low CD4/CD8 ratio with increased risk for PML as well as the potential value of this ratio as a biomarker for treatment surveillance.

The usefulness of NAT as a therapeutic tool in the near future will be based upon individualized criteria, which will permit us at the same time to stratify patients according to their specific risk for complications like PML, and identify those who benefit the most from the drug. Our study shows that low CD4/CD8 ratio coincides with a sustained and prolonged response to therapy; furthermore, it provides possible implications for future research, by implying that peripheral blood CD4/CD8 ratio might serve as a new and easy-to-measure variable to the currently proposed algorithm for the risk stratification for PML, after wider validation and careful clinical evaluation.



Conflicts of interest