Background
The differentiation of naïve T cells into an activated memory phenotype is characterized by an extensive change in the expression of trafficking determinants, resulting in the acquisition of homing receptors that enable the cells to migrate from the circulation into peripheral tissues. This change in T cell homing potential is affected by the microenvironment where initial antigen recognition occurred, as memory T cells preferentially return to regions of the body similar to those where the initial antigen was encountered [
1]. For instance, studies in mice have demonstrated that CD4+ T cells activated in cutaneous lymph nodes upregulate trafficking determinants specific for the skin, such as P-selectin ligand, while T cells responding to antigen in intestinal lymph nodes express high levels the gut-associated adhesion molecule integrin α4β7 and acquire responsiveness to the intestinal CC-chemokine ligand CCL25 [
2]. The specific profile of adhesion molecules and chemoattractant receptors expressed by individual T cells allows the cells to interact with the vascular endothelium at anatomical sites where the cognate ligands are selectively expressed, targeting the cells to specific tissues [
3,
4]. The relevance of such interactions has been amply demonstrated in humans, where postcapillary venules in the inflamed skin selectively express E-selectin and CCL17, while lamina propria of the small intestines displays the mucosal addressin MAdCAM-1 and CCL25 [
5‐
7]. Interestingly, T cells isolated from the small intestines and the skin are almost completely dichotomous in their expression of trafficking determinants; T cells from the skin express cutaneous leukocyte antigen (CLA) and CCR4, but lack integrin α4β7, while T cells from the small intestines are positive for α4β7, but not CLA or CCR4 [
6,
8‐
10]. Consequently, circulating T cells can be divided into tissue specific subsets, each of which have the ability to traffic through certain tissue compartments, but which are excluded from others.
While trafficking of T cells to the skin and gut is well characterized, less is known about the mechanisms governing homing of T cells across the choroid plexus into the CSF during immune surveillance of the healthy brain [
11]. Although it has been hypothesized that such CSF-specific trafficking determinants exist, their molecular specificity has been evasive. Some progress has been made in defining trafficking determinants for inflamed CNS microvessels: CD8+ T cells activated in cervical lymph nodes draining intracerebral tumors in the mouse acquire a phenotype characterized by high expression of the two integrins α4 and α1 as well as a modest increase in the expression of P-selectin ligand [
12]. Functional data supports a role for P-selectin/E-selectin and their ligands (including CLA) as well as integrin α4β1 (VLA-4) in extravasation of lymphocytes in the CNS during experimental autoimmune encephalomyelitis (EAE), but the relevance of these findings for trafficking of T cells across non-inflamed vessels during immune surveillance of the healthy brain remains unclear [
13‐
16]. T cells in the CSF of healthy individuals or patients with non-inflammatory neurological diseases (NIND) are characterized by high expression of integrin αLβ2 (LFA-1), α4β1, CCR5 and CXCR3 [
17,
18], a phenotype they share with activated T cells isolated from a wide range of other organs [
19‐
22]. Therapeutic α4-integrin blockade suppressed access of several cell types (CD4>CD8 T cells; CD19+ B cells; CD138+ plasma cells) into CSF [
23], identifying the first functional CSF entry determinant for human lymphocytes.
In this study we addressed the expression of trafficking determinants known to be involved in organ-specific trafficking to the skin (CLA and CCR4) and the small intestines (integrin α4β7 and CCR9) on CSF T cells from patients with NIND. We reasoned that any bias in CD4+ T cell expression of these known organ-specific trafficking determinants in the CSF might aid in identifying CNS-specific adhesion molecules or chemokine receptors. By contrast, equal presence of tissue-committed T cells in the blood and CSF would indicate that memory T cells enter CSF according to their activation or differentiation state but possibly not by specific homing molecules. We observed comparable frequencies of CD4+ memory T cells displaying CLA and CCR4 as well as integrin α4β7 and CCR9 in paired blood and CSF samples. The results demonstrated that, in contrast to most other tissues, skin- and gut-homing T cells are not excluded from the CSF, suggesting that the recruitment of CD4+ T cells to the CNS is more related to previous antigen exposure and activation than to the expression of an organ-specific adhesion molecule or chemokine receptor. Although negative with regard to roles for CCR9/α4β7 or CCR4/CLA, for CSF homing, the results carry implications for the use of novel immunotherapeutic strategies to treat neurological diseases. In particular, several new techniques for treating diverse neurological disorders require the generation of neuroantigen-specific T cells by peripheral administration of antigen either subcutaneously or to the gastrointestinal mucosa. Our current results indicate that skin- and gut-homing memory T cells readily access the CNS compartment, providing a rationale for the feasibility of such strategies.
Discussion
This study demonstrated that CD4+ memory T cells in the CSF exhibit a distinctly different expression pattern of adhesion molecules and chemokine receptors compared to most tissue-infiltrating T cell populations. It is well established that T cells isolated from the small intestines and the skin are nearly homogenously positive for mutually exclusive pairs of trafficking determinants; almost all T cells from the skin express CLA and CCR4, but lack integrin α4β7 or CCR9, while T cells from the small intestines are positive for integrin α4β7 and CCR9, but not CLA or CCR4 [
5,
6,
8,
10]. This highly polarized pattern is believed to result from selective recruitment of T cells expressing the appropriate trafficking determinants from a pool of tissue-committed and non-committed T cells in the systemic circulation. It is plausible that additional tissue-committed T cell populations exist, since T cells from most other tissue compartments, including bronchoalveolar lavage (BAL) fluid and liver, do not express CLA, integrin α4β7, CCR4 or CCR9, suggesting that hitherto unknown combinations of adhesion molecules and chemokine receptors mediate organ-specific homing to these organs [
9,
10,
29]. CSF from NIND patients contained, on the contrary, comparable numbers of CD4+ memory T cells expressing trafficking determinants specific for the skin (CLA and CCR4) and the small intestines (integrin α4β7 and CCR9) as paired samples from peripheral blood. Even though percentages of CD4+ memory T cells expressing CCR4 and integrin α4β7 were slightly lower in CSF compared to blood, the results demonstrate that skin- and gut-homing T cells access the CSF. These findings argue that T cell homing to the CSF is characterized by recruitment of a wide spectrum of tissue-specific T cell subpopulations. Aside from the established role of α4-integrin [
23], the trafficking determinants involved in CSF trafficking remain to be defined.
These results address a key question regarding the feasibility of several experimental therapies for neurological conditions utilizing antigenic stimulation in the periphery (such as the skin or the gut) to produce a population of T cells, which are intended to act in the CNS. Such therapies include peripheral immunization against A-beta for Alzheimer's disease [
30,
31], sensitization with self antigens to generate myelin-reactive T cells for the amelioration of spinal contusion and neurodegeneration [
32,
33], as well as injection of altered peptide ligands [
34,
35] and oral administration of myelin to induce tolerance for the treatment of MS [
34]. In recent years, such innovative approaches have been assessed in experimental animal models and in clinical trials, with provocative results [
35].
To our knowledge, the only other tissue compartment besides CSF, which contains tissue-committed T cells specific for a different compartment is synovial fluid. Analogous to the present observations, synovial fluid from patients with arthritis contained comparable numbers of CD4+ memory T cells expressing CLA, integrin α4β7 and CCR4 as peripheral blood [
10]. Both CSF and synovial fluid are sterile tissue fluids and the pathways for immune surveillance of such fluids may be different from the parenchyma itself. It has generally been assumed that cells in the CSF are derived from the brain parenchyma and reflect processes ongoing in the tissue compartment. Recent intravital microscopy studies have, however, demonstrated that leukocyte migration through the blood-brain barrier surrounding deep parenchymal vessels of the brain is an uncommon event in the non-inflamed murine brain [
16], while fluorescently labeled splenocytes readily were detected in the meninges and choroid plexus within two hours after adoptive transfer [
14]. Vasculature in the choroid plexus and meninges, but not in the non-inflamed human brain, express adhesion molecules supporting interactions with circulating leukocytes, suggesting that activated memory T cells may enter the CSF directly from the systemic circulation as part of immune surveillance of the CNS [
36]. It is conceivable that in addition to the well characterized pathways of organ-specific homing to tissues in direct contact with the outside environment such as skin and gut, there may be distinct mechanisms controlling homing of antigen-experienced lymphocytes to sterile tissue fluids. The majority of CNS infections are derived from pathogens, which initially entered the body through mucosal sites or the skin, rendering it likely that memory T cells primed in these locations may mount the appropriate intrathecal immune response. In contrast, pathogens causing infections in the skin or the intestines are predominantly locally derived and a compartmentalized immune response may under these circumstances be phylogenetically beneficial.
Conclusion
This study demonstrated that CLA+/CCR4+ and α4β7+/CCR9+ memory T cells are present in the CSF during immune surveillance of the CNS. Homing of T cells to the CSF during immune surveillance may be more tuned to distinguish between naïve and antigen-experienced T cells than among various tissue-specialized memory T cell populations. The ready access of skin- and gut-homing CD4+ T cells to the CNS compartment via CSF provide a rationale for the use of immunotherapeutic strategies, such as oral tolerance or therapeutic immunization, where immunogens are administered using an oral or subcutaneous route.
Methods
CSF samples
Blood and CSF were obtained from a total of 29 patients (20 women). The majority of the patients were referred for diagnostic lumbar puncture. In addition, three patients with known MS being evaluated for intrathecal baclofen treatment for intractable spasticity were included. The collection of blood and CSF samples was approved by the Institutional Review Board of the Cleveland Clinic Foundation (CCF) and written consent was obtained from all subjects. Because of limited cell numbers, the analysis of cell phenotypes in each sample was restricted to one or two trafficking determinants, along with cell lineage markers. The age of the patients ranged between 18–80 years (mean 43 years). When reviewing the charts, it turned out that 22 patients had non-inflammatory neurological diseases {NIND; headaches (6 patients), paresthesias (5), CSF circulation disturbances (3), chronic pain (3), vertigo (2), polyneuropathy (2), and mitochondrial cytopathy (1)}, six patients had clinically definite MS in remission [
37], and one patient presented with pleiocytosis during follow-up after postinfectious myelitis. Three of the MS patients were treated with immunomodulatory drugs at the time of sampling (glatiramer-acetate or interferon-β1b). The CSF leukocyte count of the NIND patients ranged between 0–2 cells/μl (mean 0.8 cell/μl), while the CSF leukocyte count of the MS patients ranged between 2–41 cells/μl (mean 9.3 cells/μl). RNA was isolated from peripheral blood mononuclear cells (PBMCs) and CSF cells from the patient with postinfectious myelitis (CSF leukocyte count 13 cells/μl).
mAbs
CD4 PerCP (clone SK3) and CD45RO APC (UCHL1) from BD Biosciences, San Jose, CA; β7 PE (FIB-504), CLA FITC (HECA-452), CXCR3 PE (1C6) and CD4 FITC/PE/APC (RPA-T4) from BD PharMingen, San Diego, CA; CD4 PE-Cy7 (SFCI12T4D11), CD45RA PE-TexasRed (2H4LDH11LDB9) and αE FITC (2G5) from Beckman Coulter, Fullerton, CA; CCR4 (328B) from ICOS, Bothell, WA; CCR9 (96-1) and α4β7 (ACT-1) from Millennium Pharmaceuticals Inc, Cambridge, MA; anti-mouse IgG1 PE/FITC from Southern Biotechnology Associates, Birmingham, AL; anti-mouse IgG Biotin and Streptavidin Cy5 from Jackson ImmunoResearch Laboratories Inc, West Grove, PA.
Flow cytometry
Immunostainings for flow cytometry were performed as previously described [
18]. Paired blood (5 ml) and CSF (10 ml) samples were collected and stained within 20 min of sampling. Staining was performed at room temperature for 15 min (CCR9, α
E, β7) or on ice for 45 min (CCR4, CLA, α4β7). Titrations were performed for each mAb in blood samples to define the concentration that resulted in saturating conditions and an optimal signal to noise ratio. Identical concentrations of mAbs were used for blood and CSF samples, as numbers of CSF cells per staining never exceeded the number of cells in peripheral blood. Erythrocytes were lysed after staining whole blood using FACS lysing solution (BD Biosciences).
Cells were acquired on an LSR (BD Immunocytometry Systems, San Jose, CA) or a MoFlo flow cytometer (Cytomation Inc., Fort Collins, CO) and analyzed using WinList software (Verity Software House Inc., Topsham, ME). Cells were gated according to forward- and side light-scattering properties, and were positively selected for the presence of CD4 in combination with high expression of CD45RO (for all stainings containing CCR4) or the absence of CD45RA (for all other stainings) to identify CD4+ memory T cells. Isotype matched control mAbs were used to define background fluorescence.
RNA isolation and real-time RT-PCR
RNA was generated from PBMCs isolated through density centrifugation on Ficoll (Lymphocyte separation medium; Meidatech Inc., Herndon, VA) and CSF mononuclear cells using Trizol (Invitrogen, Carlsbad, CA). Approximately 1 μg of DNase (Invitrogen) treated total RNA was reverse transcribed using Superscript II (Invitrogen) according to the manufacturer's instructions. PCR reactions were performed in 20-μl capillaries containing 2 mM Mg
2+, 0.25 μl each of forward (5'-AAATGAACCCCACGGATATAGCAG-3') and reverse (5'-GAAAACACGAAGAGCAGATCCGAGA-3') primer [
38], 1×DNA Master SYBR Green (LightCycler-DNA Master SYBR Green I kit; Roche, Indianapolis, IN) and 2 μl of cDNA using a LightCycler (Roche). Reaction conditions for PCR were as follow: denaturation at 95°C for 1 min, followed by 40 cycles of amplification by denaturation at 95°C for 15 s, annealing at 60°C for 5 s, and extension at 72°C for 30 s. The accumulation of products was monitored by SYBR Green fluorescence at the completion of each cycle. Construction of standard curves and analysis was performed with the LightCycler 3 software (Roche) as previously described [
39].
Statistical methods
Paired t-test was used for comparing expression of trafficking determinants on blood and CSF T cells. Reported p-values are two-tailed and considered statistically significant at p < 0.05.
Acknowledgements
We thank Dr J Stine, ICOS corporation, Bothell, WA for generously providing us with the antibody against human CCR4 (clone 328B). This work was supported by National Institutes of Health Grants PO1 NS38667 (to RMR) and RO1 AI046784 (to JJC). PK is the recipient of an advanced postdoctoral fellowship from the National Multiple Sclerosis Society.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PK participated in the conception of the study, designed the experiments, participated in the performance of the experiments, analyzed the data, and drafted the manuscript. BT enrolled the patients in the study and carried out the flow cytometry assays. TW designed the molecular biological studies. JJC designed the 6-color immunofluorescent flow stainings and participated in the analysis and interpretation of the data. RMR participated in the conception and coordination of the study, contributed to the interpretation of the data and helped to draft the manuscript. All authors read and approved the final manuscript.