Regulatory T cells exhibit enhanced migratory characteristics, a feature impaired in patients with multiple sclerosis


  • Tilman Schneider-Hohendorf,

    1. Department of Neurology, Inflammatory Disorders of the Nervous System and Neurooncology, University of Muenster, Muenster, Germany
    Search for more papers by this author
    • These authors contributed equally to to this work.

  • Max-Philipp Stenner,

    1. Department of Neurology, Inflammatory Disorders of the Nervous System and Neurooncology, University of Muenster, Muenster, Germany
    Current affiliation:
    1. Department of Neurology, University of Magdeburg, Magdeburg, Germany
    Search for more papers by this author
    • These authors contributed equally to to this work.

  • Christian Weidenfeller,

    1. Department of Neurology, Inflammatory Disorders of the Nervous System and Neurooncology, University of Muenster, Muenster, Germany
    Current affiliation:
    1. Merck Serono International S. A.– Geneva, Geneva, Switzerland
    Search for more papers by this author
    • These authors contributed equally to to this work.

  • Alla L. Zozulya,

    1. Department of Neurology, Inflammatory Disorders of the Nervous System and Neurooncology, University of Muenster, Muenster, Germany
    Current affiliation:
    1. Merck Serono International S. A.– Geneva, Geneva, Switzerland
    Search for more papers by this author
  • Ole J. Simon,

    1. Department of Neurology, Inflammatory Disorders of the Nervous System and Neurooncology, University of Muenster, Muenster, Germany
    Search for more papers by this author
  • Nicholas Schwab,

    1. Department of Neurology, Inflammatory Disorders of the Nervous System and Neurooncology, University of Muenster, Muenster, Germany
    Search for more papers by this author
  • Heinz Wiendl

    Corresponding author
    1. Department of Neurology, Inflammatory Disorders of the Nervous System and Neurooncology, University of Muenster, Muenster, Germany
    • Chair of the Department of Neurology, Inflammatory Disorders of the Nervous System and Neurooncology, University Clinic of Muenster, Domagkstraße 13, 48149 Münster, Germany Fax: +49-251-83-46812
    Search for more papers by this author


Migration of immune cells characterizes inflammation and plays a key role in autoimmune diseases such as MS. CD4+Foxp3+ regulatory T cells (Treg) have the potential to dampen immune responses but show functional impairment in patients with MS. We here show that murine Treg exhibit higher constitutive cell motility in horizontal migration on laminin, surpass non-Treg in transwell assays through microporous membranes as well as across primary brain endothelium and are present in the naïve CNS to a significantly higher extent compared to spleen, lymph nodes and blood. Likewise, human Treg from healthy donors significantly exceed non-Treg in migratory rates across primary human brain endothelium. Finally, we investigated whether the propensity to migrate is impaired as a feature of autoimmunity and therefore tested patients with MS. Treg from patients with stable relapsing-remitting MS show significantly impaired migratory capacity under non-inflammatory conditions compared to healthy donors. We hypothesize that the enhanced propensity to migrate is a feature of Treg that allows for an equilibrium in parenchymal immune surveillance, e.g. of the CNS. Impaired Treg migration across the intact blood–brain barrier, as observed for Treg from patients with MS, indicates a broader functional deficiency hypothetically contributing to early CNS lesion development or phases of MS remissions.


Naturally occurring CD4+Foxp3+ regulatory T cells (Treg) are essential mediators of peripheral immune tolerance, regulating inflammation in the context of infection, autoimmunity, neoplasia and transplant rejection 1. In addition to balancing immunity within lymphoid tissues, Treg enter non-lymphoid target sites of inflammation, exerting their anti-inflammatory function there 2–5. First, regulatory as well as effector T-cell subsets have to undergo a non-lymphoid homing receptor switch after entering secondary lymphoid tissue 6. Upon encountering specific antigens provided by dendritic cells in the T-cell area 7, the lymphocytes acquire an activated phenotype, expressing distinct surface markers of non-lymphoid homing (chemoattractant receptors) and migration (adhesion molecules) 6. Recent observations suggest that Treg should be equipped with a higher propensity to migrate 6 in order to efficiently suppress effector T cells at target sites of emerging inflammation, as they are hypoproliferative 8, 9 and only form 6–10% of the whole CD4+ T-cell subset. Reports on the accumulation of Treg within the murine CNS during EAE 3 and on containment of EAE relapses by CNS Treg 10, 11 support the concept of their central role in balancing parenchymal immune responses in the CNS.

Evidence for the relevance of Treg in the human CNS to date is sparse. While Tzartos et al. found no evidence for the presence of Treg in active MS lesions 12, a recent study by Fritzsching et al. (personal communication, abstract in Multiple Sclerosis, Sep 2009; vol. 15: p. 72) described the detection of low numbers of Treg in the CNS and in accordance with an earlier study elevated cell numbers in the cerebrospinal fluid of patients with MS 13. Since increasing evidence supports an anti-inflammatory role for Treg at parenchymal sites of inflammation 14, one could speculate that the repeatedly reported impairment in antiproliferative capacity of Treg found in patients with MS 15, 16 is just one expression of a more thorough Treg dysfunction. Whether Treg migration to sites of active inflammation in the CNS of patients with MS is impaired has been elusive so far.

We here combined various murine and human models quantifying transmigratory capacity and locomotion to determine how constitutive, innate Treg motility translates into diapedesis across CNS endothelium.


Murine Treg exhibit enhanced migratory capacity in vitro and in vivo

We first characterized lymph node-derived regulatory and non-regulatory T-cell subsets with regard to their expression of surface markers indicative for adhesion, migration and activation. In line with previous results for CCR6 17, murine Treg consistently showed a significantly higher expression for all inspected markers apart from CCR7, where the higher expression was not significant (p=0.126), and a significantly lower expression of CD62L than on non-Treg. However, collagen/laminin receptors VLA-1 and VLA-2 were expressed very weakly on both T-cell subsets (n=5) (Supporting Information Fig. 1A–D).

When applied to a laminin-coated glass slide for 3 h of time-lapse videomicroscopy, Treg revealed an enhanced motility compared to non-Treg (n=3) (Supporting Information Fig. 2A–F). Moving cells were individually tracked to measure laminin-specific, horizontal motility and speed excluding non-moving periods. Treg covered the distance of 248.1 μm (mean)±20.47 (SEM) with a mean speed of 1.53 μm/min±0.13 in 3 h, whereas non-Treg reached a mean distance of 97.47 μm±9.38 with a mean speed of 0.65 μm/min±0.06. The average percentage of locomotion during the video-capturing was comparable between the two T-cell subsets (Treg 85.95±1.78; non-regulatory T cells 83.37±2.84); compared to the difference in total distance, the difference in beeline distance was smaller with Treg covering 88.8 μm±9.51 and non-Treg covering 49.24 μm±5.25, indicating that Treg exhibited a higher rate of direction changes during laminin-specific 2D migration compared to non-Treg.

To analyze T-cell diapedesis, we used freshly isolated, primary CNS endothelium as an in vitro model of the blood–brain barrier (BBB) cultured in a transwell migration assay. Naïve, lymph node-derived CD4+ T cells were applied on the luminal side of the cultured murine brain microvascular endothelial cell (MBMEC) layer and were collected from the three compartments after 18 h as delineated in Fig. 1C (upper chamber, MBMEC layer and lower chamber) to check whether Treg accumulated among CD4+ T cells. Fig. 1B depicts a representative population of CD4+ T cells incubated for 18 h to serve as a reference. Between 4.8 and 6.3% Treg were found in all experiments (n=5, data not shown). When no attracting stimuli was added to the medium, CD4+ T cells showed very low migration (data not shown) so we used FBS, which is known to contain low concentrations of different cytokines as a chemoattractant agent. Eighteen hours after application of the CD4+ T cells to an FBS gradient, Treg accumulated to 20.7% of the entire CD4+ T-cell population within the MBMEC fraction (n=5, 15.1–29.8%). In the basolateral compartment, Treg enriched to 10.8% of total CD4+ T cells (n=5, 8.4–20.2%) (Fig. 1D).

Figure 1.

Murine Treg exhibit higher migratory capacity in vitro and in vivo when compared to their non-regulatory counterparts. (A) Gating for the murine flow cytometry data: left panel shows the lymphocyte live gate; right panel depicts the gating on the CD4+ T-cell subset. Ratio of T-cell subsets identified by flow cytometry. (B) T cells were cultured for 18 h to determine the percentage of Foxp3+/− and the CD44+/− T-cell subsets. One representative experiment is shown (n=5). (C) Scheme of the experimental setup including the transwell insert with the microporous membrane (3 μm pore Ø) and three collecting compartments. (D) CD4+ murine T cells were cultured on a MBMEC layer in presence of 1% FBS in the lower chamber. After 18 h of incubation, cells were harvested from the upper chamber, the MBMEC layer and the lower chamber and analyzed by flow cytometry to determine the percentage of Foxp3+/– and the CD44+/– T-cell subsets. One representative experiment is shown (n=3). (E) CD4+ murine T cells were cultured on a MBMEC monolayer in the presence of 1% FBS and 20 μg/mL CCL20 in the lower chamber, analysis was performed as described in Figure D (n=3). (F) CD4+ T cells were isolated from peripheral blood, spleen, lymph nodes and CNS of untreated C57BL/6 mice after perfusing with PBS and subjected to flow cytometry after Foxp3 staining. The average percentage of Foxp3 among total CD4+ T cells was calculated. ***p<0.001, **p<0.01, unpaired, two-tailed t-test, error bars indicate SE (n=5).

As CCR6 is expressed on both T-cell subsets (Supporting Information Fig. 1D), we tested whether CCL20 (the CCR6 ligand) contributes to the preferential migration of Treg in the MBMEC layer. Although enrichment of Treg within the MBMEC layer was nearly completely abrogated (5.7–6.7%), the accumulation of Treg in the lower chamber was threefold enhanced by addition of CCL20 from 10.8 to 34.1% of migrated cells (Fig. 1E). Activation of the MBMEC layer 24 h before starting the migration assay with murine TNF-α and IFN-γ revealed a similar Treg accumulation as under non-inflammatory conditions while, as expected, the total counts of migrated cells from the lower chamber increased under inflammatory conditions (n=3, data not shown).

To verify our findings in vivo, we examined naïve C57BL/6 mice for ratios of Treg versus non-Treg in the CNS, spleen, lymph nodes and peripheral blood by flow cytometry after animal perfusion with PBS (Fig. 1F). We were able to isolate approximately 2×104–1×105 leukocytes with a Percoll density gradient from the CNS of healthy mice. Strikingly, Treg were present to a significantly higher extent in the CNS compared to the three other examined organs (mean±SE blood: 4.5±0.5, lymph nodes: 10.6±0.9, spleen: 12.1±1, CNS: 19.55±1.4, n=5).

Taken together, murine Treg showed higher expression of surface markers indicative for activation, adhesion and migration, and exhibited higher motility in 2D migration on a laminin substrate. Moreover, they were enriched in the basolateral fraction of a transendothelial migration assay as well as in or closely adhering to the top of endothelial cell monolayer itself. The specific chemoattractant stimulus with CCL20 augmented the basolateral accumulation of Treg and prevented their enrichment in the endothelial cell monolayer. The higher migratory capacity of Treg was reflected by an enrichment of Treg within the CNS of naïve WT mice.

An MBMEC layer significantly enhances the migratory capacity of murine Treg

To quantify the total amount of migrated T cells and to preclude other reasons for an enrichment of Foxp3+ T cells in the lower compartment, such as suppression of non-regulatory T-cell migration by Treg or short-term induction of Foxp3-expressing T cells in the course of diapedesis of de facto non-Treg, we isolated the CD25high Treg and CD25 non-regulatory T-cell fractions to use these subsets in migration assays. We first applied solely the T-cell fractions to microporous membranes without a MBMEC monolayer, using an FBS gradient. Although non-Treg showed a migratory rate of 565±38.5 cells/104 beads, Treg amounted to 1018±53.2 cells/104 beads, a rate that was 30.6% higher (Fig. 2A). As expected, this difference in migratory rates was higher in the presence of CCL20 (by 40%, Treg 1704±125.5 cells/104 beads, non-Treg 814±68.2 cells/104 beads). In the presence of MBMEC monolayer the total amount of migrated cells decreased due to the cellular barrier. Thus, non-Treg showed a migratory rate of 93±36.8 cells/104 beads, whereas Treg reached an elevated rate of 279±53 cells/104 beads, resulting in a difference of 66.7% of migration index (Fig. 2B). An even higher difference in the migratory rate of 78% was reached by addition of CCL20 chemokine (Treg 546±27.6 cells/104 beads, non-Treg 120±6.4 cells/104 beads). Figure 2C summarizes three independent experiments as shown in Fig. 2A and B. The migration indices of Treg, normalized to the migratory rates of non-Treg, significantly increased in the presence of MBMEC (p=0.03). Taken together, these experiments demonstrate that the assumed differences in migratory capabilities are consistent for isolated Treg or non-Treg that are facing a microporous membrane. Enrichment of Treg is hence neither due to any suppression of migration of non-Treg nor due to induction of Foxp3-expressing non-Treg. The difference in migratory rates is augmented in the presence of MBMEC as a cellular barrier as well as by CCL20 as a specific, chemotactic stimulus.

Figure 2.

Isolated murine Treg efficiently migrate through microporous membranes, but show improved migration across MBMEC. Quantitative comparison of murine Treg and non-regulatory T-cell migration (for gating, see Supporting Information Fig. 4). (A) T cells were applied onto a microporous membrane adding 20 μg/mL CCL20 (left bars) or 1% FBS (right bars) to the lower chamber. (B) T cells were applied to a MBMEC layer adding 20 μg/mL CCL20 (left bars) or 1% FBS (right bars) to the lower chamber. For analysis, reference beads were added to the migration chambers before collecting and staining the migrated cells with anti-CD4 and anti-Foxp3. Quantification was performed analyzing 1×104 reference beads by flow cytometry. In (A) and (B), one representative experiment out of three performed is shown; error bars indicate the standard error generated by three individually analyzed transwells for each condition. (C) Summary of three independent experiments as shown in (A) and (B). The migration index is calculated normalizing the mean cell counts/1×104 reference beads of CD4+CD25high to that of CD4+CD25 T cells, which is set to 1 (dashed line, p=0.03, unpaired, two-tailed t-test).

Human Foxp3+ Treg migrate across in vitro human brain endothelium at higher rates than non-Treg

To determine whether human Treg feature similar characteristics in transendothelial migration as their murine counterparts, we used a well-established in vitro model of the human BBB 18. Primary human brain microvascular endothelial cells (HBMEC) cultured on transwell membranes were used for these experiments. As evidence of their preserved physiological behavior, HBMEC displayed a phenotype more supportive of diapedesis upon activation with IFN-γ and TNF-α, expressing the adhesion ligands such as ICAM-1 and VCAM-1 that were otherwise not detectable (Supporting Information Fig. 3). Similar to the murine experiments, 5% of human PBMC added to the upper transwell compartment crossed the HBMEC layer in 12 h migration experiments as compared to an average of 15% when the barrier only consisted of the coated porous membrane (n=12, not shown).

In line with the murine experiments, the proportion of Treg among CD4+ T cells was significantly higher within the fraction of PBMC that had crossed HBMEC than among the initial PBMC sample added to the upper compartment, the latter approximating the Treg blood frequencies of healthy donors (HD) (n=10, Fig. 3: %Foxp3+ among CD4+ T cells, mean±SD: 3.32±1.36%, range 1.83–6.03% (blood) versus 11.31±5.07%, range 2.81–19.39% (migrated)).

Figure 3.

Human Treg from HD are significantly enriched after transendothelial migration. Quantitative comparison of human HD Treg in relation to non-regulatory T-cell migration: PBMC from HD (n=10) were added to the apical chamber for migration either through coated porous membranes (3 μm pore Ø) or across HBMEC under baseline and inflammatory conditions as indicated. After 12 h, PBMC in the basolateral compartment were collected, stained with anti-CD4 and anti-Foxp3 and analyzed by flow cytometry (gating is depicted in Supporting Information Fig. 4). The y-axis represents the proportion of Foxp3+ Treg among CD4+ T cells before (blood) and after migration. **p<0.01; Friedman Test, Dunn's Multiple Comparison test.

Similarly, in vitro simulation with IFN-γ and TNF-α did not significantly alter the migratory superiority of Treg (14.14±5.29%, range 5.48–22.56% migrated Foxp3+ among CD4+ T cells). Again, as seen in the murine experiments, when migrating across porous membranes in the absence of HBMEC, Treg consistently accumulated within the migrated CD4+ compartment as well, but to a lower and non-significant extent (6.16±2.3%, range 3.16–10.51% migrated Foxp3+ among CD4+ T cells).

Taken together, under basal, non-inflammatory conditions, human Foxp3 Treg migrate through porous membranes and brain endothelium at higher rates than their non-regulatory counterparts.

Treg from patients with relapsing-remitting MS (RR-MS) are impaired in their migratory capability

We further speculated that the enhanced migratory propensity of Treg might contribute to the equilibrium in tissue immune surveillance under physiological conditions. To further investigate this concept, we tested the migratory potential of Treg derived from RR-MS patients, which have been reported to be dysfunctional by several groups. To date, Treg dysfunctionality has been attributed to their suppressive, antiproliferative capacity in vitro, which has been shown to be reduced in MS 19. Whether migratory abilities are affected and could therefore contribute to the disturbed immune cell homeostasis in the CNS as well has been elusive so far. Of note, the antiproliferative function of Treg from HD has been shown to decline with age 19. To exclude potential differences due to an alleged general deterioration of Treg function with age, we matched age and sex of patients and controls.

Strikingly, Treg from untreated patients with RR-MS in stable phases of the disease did not accumulate among migrated CD4+ T cells under non-inflammatory conditions, exhibiting transmigratory rates comparable to their non-regulatory counterparts (n=12, Fig. 4A: %Foxp3+ among CD4+ T cells, mean±SD: 3.27±1.54%, range 1.4 to 7.4% (blood) versus 5.11±2.62%, range 2.48–10.96% (migrated)). No significant differences in blood frequencies of CD4+Foxp3+ T cells were observed between HD and patients with RR-MS, which is in accordance to previous reports 14.

Figure 4.

Treg from patients with RR-MS are impaired in their migratory rate across non-inflamed brain endothelium. Quantitative comparison of Treg versus non-regulatory T-cell migration derived from human HD and patients with MS (gating is depicted in Supporting Information Fig. 4). (A) Percentage of Treg among CD4+ T cells from untreated patients with RR-MS in a stable phase of the disease (n=12) in the blood, migrated through untreated or inflamed HBMEC. (B) N-fold change in the Treg proportion of HD and patients with RR-MS. CD4+ T cells comparing the migrated fraction to the initial sample added to the apical compartment ([%Foxp3+ among CD4+]migrated/[%Foxp3+ among CD4+]before migration). (C) Absolute numbers of CD4+ T cells from HD and patients with RR-MS after transendothelial migration under non-inflammatory conditions. *p<0.05; **p<0.01; Friedman Test, Dunn's Multiple Comparison Test; n. s. not significant.

As expected, administration of inflammatory cytokines to the endothelium significantly increased the proportion of migrated Treg (12.52±4.84%, range 6.87–21.09% migrated Foxp3+ among CD4+). When directly comparing the changes in Treg frequencies due to transmigration between patients with RR-MS and HD, we found that transendothelial Treg migration in our cohort of patients with MS was significantly impaired under basal conditions, but could be restored to levels comparable to those observed for HD-derived Treg with TNF-α and IFN-γ pre-treatment (Fig. 4B: n-fold change of [%Foxp3+ among migrated CD4+] and [%Foxp3+ among CD4+ in the initial sample]: 3.81±2.04, range 1.15–6.69 (HD, non-inflamed endothelium) versus 4.81±2.71, range 1.85–10.84 (HD, inflamed endothelium) versus 1.85±1.4, range 0.82–5.12 (RR-MS, non-inflamed endothelium) versus 4.19±1.69, range 2.21–7.3 (RR-MS, inflamed endothelium)). Absolute numbers of migrated CD4+ T cells did not differ between HD and patients with RR-MS, neither under inflammatory nor non-inflammatory conditions (Fig. 4C: total number of migrated CD4+ T cells, mean±SD: 453±505 for HD, n=10; 342±177 for patients with RR-MS, n=5). Hence, it can be excluded that the diminished Treg proportions observed among migrated RR-MS T cells under non-inflammatory conditions are due to increased Foxp3 T-cell migration.


We here report enhanced migratory abilities of murine, unprimed Treg in vitro and in vivo when compared to unprimed non-Treg, a feature shared by human HD Treg. In contrast, Treg of patients with RR-MS exhibit significantly impaired migratory capabilities under non-inflammatory conditions. Hence, we conclude that the observed enhanced propensity to migrate is a basic, innate feature of Treg and that this feature crucially contributes to the maintenance of tissue immune homeostasis, specifically in the CNS. This mechanism is impaired in patients with MS and could thus possibly facilitate the initiation of CNS inflammation.

The 2D migration paradigm is supposed to represent T-cell migratory behavior on extracellular matrix components such as laminin, also dominant in the basement membrane surrounding the endothelium. To closer mimic the in vivo situation, we used primary MBMEC to generate a transversal barrier for CD4+ T-cell migration. Treg maintained their feature of enhanced motility compared to non-Treg: importantly, they also accumulated within or on top of the endothelial layer indicating an advantage of Treg in performing the first steps of transendothelial migration. Specific chemotactic stimuli then seem to draw Treg from the endothelial layer into the surrounding tissue as Treg accumulation within the MBMEC layer is abolished when a CCL20 gradient is added. The presence of elevated numbers of Treg in murine CNS confirmed their enhanced migratory capacity in vivo, further emphasizing the important role of Treg in immune surveillance of the CNS under non-inflammatory conditions. Quantitative migration assays with purified Treg versus non-Treg through microporous membranes proved that the lower migratory capacity of non-Treg was not due to a suppressive influence of Treg. Moreover, the transendothelial migratory superiority of Treg cannot solely be due to interactions with the endothelium. However, the increased difference in migratory rates of Treg and non-Treg in the presence of a MBMEC layer hints to Treg-specific interactions with the endothelial cell layer, either due to direct cell–cell contact or due to a constitutive secretion of soluble factors by the endothelial cells. CCL20 as a soluble stimulus secreted by the MBMEC layer can be excluded since its expression is only found in epithelial cells of the choroid plexus and astrocytes during EAE relapse 20, 21 but not in brain endothelium. More likely, Treg seem to have an advantage in forming stable cell–cell contacts with the brain endothelium, consistent with their higher expression of LFA-1 and CD49d, as they intensively accumulated in or on top of the endothelial cell monolayer compared to their non-regulatory counterparts.

The preferential migration of Treg through a porous membrane in the presence of the chemoattractant CCL20 was expected by their CCR6 cell surface expression and was maintained when T cells migrated across an in vitro model of the BBB. In the non-regulatory fraction, particularly the Th17 cells should be attracted by the CCL20 gradient as they are known to express high amounts of CCR6 compared to other effector cell types 22. This finding further supports the current notion that CCR6 expressing, autoreactive effector Th17 cells may be able to gain entry to the yet non-inflamed CNS, facilitated through CCL20 secretion by epithelial cells of the choroid plexus or brain resident glia cells 21, 23, 24, and induce the subsequent immune responses by producing CCL20 among other inflammatory stimuli 22. In consequence, this might lead to inflammation of the BBB endothelium allowing further, CCL20 independent lymphocyte infiltration into the CNS parenchyma. Treg, exhibiting a stronger migratory response to CCL20 than conventional CD4+ T cells, should therefore have a higher prevalence in the brain tissue compared to their effector counterparts under healthy conditions, consistent with our in vivo finding.

Human Treg have been reported to be present in the CNS in certain neurological disorders, such as gliomas 25, 26. Under conditions of experimental autoimmune neuroinflammation as in EAE, Treg accumulate in the murine CNS 4, 10, most notably in the remission phases 11, counterbalancing encephalitogenic CNS responses. As mentioned above, data on the presence and function of Treg in the human CNS are sparse 12–14, 18. To translate our findings into human pathophysiology, we used an in vitro model of the human BBB to mimic lymphocyte diapedesis in vivo. In contrast to HD, MS patient-derived Treg failed to outmatch their non-regulatory counterparts in crossing the BBB under basal, non-inflammatory conditions. The absence of any inflammatory milieu under these experimental conditions might represent the situation in very early lesion development or in remitting stages of CNS inflammation. Since total numbers of migrated CD4+ T cells did not differ between HD and RR-MS samples, lower Treg percentages under non-inflammatory conditions can be excluded to be due to increased migration of non-Treg. In line with our data on murine Treg transmigration, human HD Treg displayed consistent basolateral accumulation in the absence of endothelial cells. Higher Treg motility compared to non-Treg has previously been suggested as a mechanism of suppression of T effector cell function as Treg were shown to be superior to TH cells in establishing close contact to dendritic cells, subsequently inhibiting their full maturation 27. Our finding of an augmented Treg motility in HD therefore is very well in line with this previous data. Furthermore, our observation of a migratory dysfunction of MS patient derived Treg introduces the idea that the presumed “regulatory deficiency” of CD4+ Treg in MS could at least be partially due to impairment in Treg motility.

Our study provides first evidence of augmented overall cell motility as a constitutive feature of both murine and human naturally occurring regulatory T cells. Adhesion ligand and chemokine receptor patterns expressed by Treg and their non-regulatory counterparts presumably determine site-specific homing and have recently been a matter of substantial interest. Their innate cell motility, however, forms the basis of transendothelial diapedesis to and locomotion within any tissue and has been completely neglected in the past. Our data demonstrate an innate migratory superiority of murine and human Treg over naïve non-Treg. This migratory advantage should contribute to the role of Treg in maintaining tissue immune homeostasis and CNS immune surveillance. However, this can be disturbed under conditions of autoimmunity, as demonstrated for MS patient-derived Treg. Albeit speculative, our findings could have relevance for the understanding of early lesion development and remitting phases during MS course.

Materials and methods

Clinical samples

Twelve patients (9 female, 3 male) suffering from clinically definite RR-MS according to the revised McDonald diagnostic criteria 28 were enrolled in this study. All patients were in a stable phase of the disease, with relatively low scores on Kurzke's expanded disability status scale (EDSS<3.5) and neither currently nor previously receiving any immunomodulatory treatment (age: 41.7±12.6 years, disease duration: 4.9±6.6 years, EDSS: 1.4±0.8). Ten HD (7 female, 3 male) with no previous history of neurologic disease served as controls (age: 34.1±12.2 years). There was no significant difference in age and gender distribution between patients with MS and healthy individuals. The study was approved by the local ethics committee and informed written consent was obtained from all participants.

Isolation of murine T cells

Six-wk-old female C57BL/6 mice were obtained from Harlan Laboratories. For T-cell isolation, cervical, axillary, mesenterial and inguinal lymph nodes or spleens were collected and mechanically homogenized. Obtained cell clusters were isolated with a 40-μm mesh filter (Becton Dickinson) and magnetically separated into a CD4+ or into CD4+CD25high/– fractions using a Miltenyi MACS® kit according to the suppliers manual. A proportion of the CD25high T-cell population was checked for Foxp3 expression with the purity≥85% in all experiments. Peripheral blood was drawn directly from the heart of sacrificed mice. For CNS-derived lymphocyte flow cytometry, a Percoll density gradient was used as described previously 29. In brief, mice were sacrificed with CO2 and immediately perfused with 10 mL of PBS before harvesting the brain and spinal cord. The tissue was, similar to the lymph nodes, mechanically homogenized in PBS, layered on a 30%/50% Percoll gradient and centrifuged without brake at 600×g for 30 min. After removing the top layer of myelin, lymphocytes were harvested at the Percoll interphase.

Isolation of MBMEC

MBMEC were isolated according to Weidenfeller et al.30. The obtained capillary fragments were seeded onto CollagenIV/fibronectin-coated membranes of transwell inserts (6.5 mm Transwell® Pore Polyester Membrane Insert, pore size 3.0 μm, Corning, 2 inserts/mouse brain). Cells were incubated in DMEM high glucose with 2 mM L-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin (PAA), 20% plasma derived bovine serum (First Link), 10 ng/mL basic fibroblast growth factor (Peprotech), 100 ng/mL heparin and 4 μg/mL pyromycin (Sigma-Aldrich) for 3 days followed by an additional 2 days of incubation without pyromycin. At this time, the monolayer reached confluence, which was randomly monitored by TEER measurements (confluence at TEER plateau).

Murine transmigration assays

Freshly isolated and magnetically separated fractions of CD4+, CD4+CD25high or CD4+CD25 T cells (6×105/insert) were applied on 3.0-μm pore polyester membrane transwell inserts (Corning) with or without a MBMEC layer grown onto the microporous membrane in RPMI1640 with 100 U/mL penicillin, 100 μg/mL streptomycin (PAA) and 2% B-27 serum free supplement (Gibco). T cells from three compartments were harvested after an incubation period of 18 h. Each transwell insert was removed from the well plate; cells from the upper chamber were collected by transfer of the cell suspension into a new conical and rinsing with PBS two times to ensure removal of all remaining T cells. T cells from within the MBMEC layer were harvested by incubating the cell layer with Accutase (PAA) for 10 min at 37°C and 4% CO2. The cells were then detached by rinsing with PBS and transferred into a new conical. Cells in the lower chamber were collected and wells were subsequently rinsed with PBS twice to ensure complete removal of cells. For quantification, Calibrite beads (Becton Dickinson) were added prior to harvesting the cells. Cell number was determined by counting 1×104 reference beads with a four-color FACSCalibur flow cytometer (Becton Dickinson).

Primary HBMEC and human transmigration assays

Primary HBMEC (ScienCell) were cultured in HBMEC growth medium supplemented with FBS (ProVitro, Berlin) on 4 μg/mL fibronectin (Sigma-Aldrich). Transendothelial migration experiments were performed as described previously 18. In brief, 3.0-μm pore polyester membrane transwell inserts (Corning) were coated with 100 μg/mL fibronectin and 400 μg/mL collagen type IV (Sigma-Aldrich) for 30–60 min before 1.5×105 HBMEC were added. 500 IU/mL TNF-α and 500 IU/mL IFN-γ (R&D, Minneapolis, MN, USA) were added to the lower compartment 4 h after the addition of HBMEC for some experiments. Incubation time for the endothelial monolayer was carefully titrated according to confluence and firm intraendothelial adhesion, determined by immunohistochemical stainings of the tight junction protein occludin, and the electrical resistance of the endothelial monolayer (TEER). PBMC or CD4+ T cells were seeded onto the confluent BMEC monolayer 16 h after activation of the endothelium and the T-cell phenotypes in the lower compartment were analyzed after a 12-h incubation time.

Isolation of human immune cells and purification of T-cell subsets

Human PBMC were isolated by centrifugation of donor blood on a Lymphoprep (Fresenius Kabi Norge AS) density gradient. To allow comparative analysis of cells from patients with RR-MS and healthy controls, PBMC were immediately cryopreserved and stored in liquid nitrogen. Human CD4+CD25high Treg were isolated using MACS technology (Miltenyi) according to the supplier's manual.

Flow cytometry

Cells were washed twice in PBS containing 0.1% sodium azide and 1% bovine serum albumin and incubated for 30 min with monoclonal antibodies for different T-cell surface antigens. The following anti-human monoclonal antibodies were used (all fluorochrome-conjugated): anti-CD4 (SK3), (BD Biosciences), anti-CD4 (M-T466) (Ebioscience) and anti-VCAM-1 (1G11B1) (Abcam). The respective isotype controls (mouse IgG1, rat IgG2a, mouse IgG1) were purchased from BD Biosciences. Intracellular staining using anti-human and anti-murine-Foxp3 (clones PCH101 and FJK-16s, respectively) antibodies were performed using Foxp3 staining kits (Ebiosciences) according to the manufacturer's protocol. AntiCD4 (RM4-5), anti-CD44 (IM7), anti-CD73 (TY-11-8), anti-CD62L (MEL-14), anti-CD69 (H1.2F3), anti LFA-1 (2D7), anti-CCR5 (C34-3448), anti-CCR7 (150503), anti-CD49d (9C10) (BD Biosciences), anti-CCR6 (140706) (R&D), anti CD49a (804) (Serotec) and anti-CD49f (GoH3) (Biolegend) monoclonal antibodies were used for flow cytometry of murine T cells. Data were acquired on a FACSCalibur flow cytometer (BD) and analyzed using FlowJo software 7.5 (Tree Star).


HBMEC cultures were fixed at different incubation time points with 4% paraformaldehyde, blocked with 30% donkey serum (PAA) for 60 min, incubated with goat-anti-human ICAM-1 (British Biotechnology) for 1 h and subsequently stained with donkey-anti-goat Cy2 (Dianova) for another 60 min.

Time-lapse videomicroscopy

Cover slips for migration analysis were coated with 20 μg/mL laminin (Sigma-Aldrich (after precoating with 10 μg/mL poly-D-lysine (Sigma-Aldrich)) and were transferred to migration chambers. Each chamber was filled with 200 μL RPMI1640 with 100 U/mL penicillin, 100 μg/mL streptomycin (PAA) and 2% B-27 serum-free supplement (Gibco) containing 6×104 freshly isolated CD4+ T cells, pooled from the lymph nodes of three female C57BL/6 mice, and were subsequently sealed with paraffin/vaseline. Migration chambers were incubated at 37°C for 1 h prior to time-lapse imaging to allow for sedimentation and were then transferred to the microscope (DM IL, Leica) connected to a digital camera (TP-505D, Topica). Images were taken every 20 s at a magnification of 20× for 3 h using an automated software (Time controlled Recorder Tetra V., SVS-Vistek). To provide adequate culturing conditions (37°C), a thermal measurement feedback regulator (STATOP-4849, Chauvin Arnoux) was connected to an infrared heat lamp (Beurer). Time-lapse movie sequences were analyzed for speed (excluding non-moving periods) and covered distance of migrated cells with a custom build software (Autocell, Department of Dermatology, University of Wuerzburg).

Statistical analysis

The murine experiments were statistically analyzed with an unpaired, two-tailed Student's t-test. The human experiments were analyzed with a repeated measures, non-parametric Friedman Test and a Dunn's Multiple Comparison Test as post test. Significance is indicated as *=p<0.05 and **=p<0.01.


The authors would like to thank Professor P. Friedl for providing materials, Julia Schlingmann and Heike Menzel for the collection of clinical samples and Michaela Karches-Böhm for excellent technical help. The authors are grateful to all patients and HD for enabling this study. This study is supported by the BMBF Competence Network of MS (UNDERSTANDMS, Alliance “Immunoregulatory networks in MS,” to H. W.).

Conflict of interest: The authors declare no financial or commercial conflict of interest.