A novel fluorescent toxin to detect and investigate Kv1.3 channel up-regulation in chronically activated T lymphocytes

T lymphocytes with unusually high expression of the voltage-gated Kv1.3 channel (Kv1.3 high cells) have been implicated in the pathogenesis of experimental autoimmune encephalomyelitis, an animal model for multiple sclerosis. We have developed a fluoresceinated analog of ShK (ShK-F6CA), the most potent known inhibitor of Kv1.3, for detection of Kv1.3 high cells by flow cytometry. ShK-F6CA blocked Kv1.3 at picomolar concentrations with a Hill coefficient of 1, and exhibited >80-fold specificity for Kv1.3 over Kv1.1 and other K V channels. In flow cytometry experiments, ShK-F6CA specifically stained Kv1.3-expressing cells with a detection limit of ~600 channels per cell. Rat and human T cells that had been repeatedly stimulated 7-10 times with antigen were readily distinguished on the basis of their high levels of Kv1.3 channels (>600 channels/cell) and ShK-F6CA staining from resting T cells or cells that had undergone 1-3 rounds of activation. Functional Kv1.3 expression levels increased substantially in a myelin-specific rat T cell line following myelin antigen stimulation, peaking at 15-20 hours and then declining to baseline over the next 7 days, in parallel with the acquisition and loss of encephalitogenicity. Both calcium- and PKC-dependent pathways were required for the antigen-induced Kv1.3 up-regulation. ShK-F6CA might be useful for rapid and quantitative detection of Kv1.3 high expressing cells in normal and diseased tissues, and to visualize the distribution of functional channels in intact cells. HPLC, MALDI-TOF


INTRODUCTION
Human T lymphocytes express two potassium channels, the voltage-gated K + channel Kv1.3 and the calcium-activated K + channel IKCa1, that are involved in proliferation and cytokine secretion (1)(2)(3)(4)(5)(6). Recently, we reported that myelin-reactive encephalitogenic rat T cells expressed unusually high numbers of Kv1.3 channels following eight or more repeated antigenic stimulations in vitro (7). Adoptive transfer of these Kv1.3 high T cells into rats induced experimental autoimmune encephalomyelitis (EAE) 1 . EAE is an animal model for multiple sclerosis (MS), a chronic inflammatory disease of the central nervous system characterized by immune-mediated focal demyelination and axonal damage resulting in severe neurological deficits (8)(9)(10). Studies with several myelin-specific rat T cell lines revealed a correlation between encephalitogenicity and the number of expressed Kv1.3 channels (7,11). In addition, in vivo Kv1.3 blockade ameliorated adoptive EAE, suggesting a crucial role of Kv1.3 in the pathogenicity of these cells (7,12). A rapid method to detect Kv1.3 high lymphocytes might therefore facilitate studies of Kv1.3's role in the pathogenesis of MS and its potential as a therapeutic target for this disease.
The patch-clamp technique is widely used to determine functional Kv1. 3 and IKCa1 channel levels in lymphocytes, but requires highly specialized equipment and is time-consuming, allowing the study of only a few dozen cells per day. RT-PCR and Western blot analysis provide a measure of channel-transcript or channel-protein expression in lymphocytes, but not the 1 Aeea: aminoethyloxyethyloxy-acetyl; APC: antigen presenting cell; ChTX: charybdotoxin; Con A: concanavalin A; CsA: cyclosporin A; EAE: experimental autoimmune encephalomyelitis; F6CA: Fluorescein-6 carboxyl; IL-2: interleukin-2; MBP: myelin basic protein; MgTX; margatoxin; MS: multiple sclerosis; NFAT: nuclear factor of activated T cells; PBMC: peripheral blood mononuclear cells; PKC: protein kinase C; ShK: Stichodactyla helianthus toxin; TCGF: T cell growth factor; TEA: tetraethylammonium; TMR: tetramethylrhodamin; TT: tetanus toxoid. number of functional channels in the cell membrane. All known Kv1.3-specific antibodies target intracellular epitopes (13) in the channel making it necessary to permeabilize lymphocytes before they can be stained with these reagents. In addition, non-specific staining imposes a further limitation to the use of antibodies for protein detection. Furthermore, no reports have yet described any success in using these antibodies for immunostaining lymphocytes necessitating the development of novel tools to identify Kv1.3 high lymphocytes in tissues.
Despite the clear need for fluorescently labeled toxins as markers of channel expression and distribution in intact cells, there have been only a few reports of channel-binding peptides that have been successfully tagged with fluorophores. Fluorophore-labeled channel-binding peptides were reported previously for sodium channels (14,15), voltage-activated calcium channels (16) and NMDA-receptors (17). Most recently, a fluorophore-tagged hongotoxin analog was used to detect Kv1.1 and Kv1.2 channels in rat brain sections and Kv1.3 in Jurkat cells (18,19). Several Kv1.3-blocking polypeptides have been discovered by us and by others in scorpion venom and sea anemone extracts (5,(20)(21)(22)(23)(24). Since these polypeptides bind with extremely high affinities to Kv1.3, they might be used in much the same way as antibodies. Unlike Kv1.3-specific antibodies, these polypeptides bind to the outer vestibule of Kv1.3 and can therefore reach their binding pocket in live intact lymphocytes. ShK, a 35 amino acid polypeptide from the Caribbean sea anemone Stichodactyla helianthus, is the most potent known inhibitor of Kv1.3 (K d 11 pM), and once bound to the channel does not wash off easily (5). ShK binds via high affinity interactions to residues in the channel's outer vestibule (5,25). If suitably tagged with fluorophores, ShK could be used as a molecular probe to detect Kv1.3 high lymphocytes by flow cytometry.
Generation of the ShK conjugates -Fmoc-amino acids (Bachem A.G., CH-4416 Bubendorf, Switzerland) included: Ala, Arg(Pmc), Asn(Trt), Asp(OtBu), Cys(Trt), Gln(Trt), Glu(OtBu), Gly, His(Trt), Ile, Leu, Lys(Boc), Met, Phe, Pro, Ser(tBu), Thr(tBu) and Tyr(tBu). Stepwise assembly was carried out starting with 10 g of Fmoc-Cys(Trt)-resin (0.65 mmol/g) on a Labortec SP4000 peptide synthesizer. Following final removal of the Fmoc-group from the N-terminal Arg residue, a resin aliquot was removed and the hydrophilic linker Fmoc-Aeea-OH (Fmocamino-ethyloxy-ethyloxyacetic acid) was coupled as an HOBT ester. This resin was subsequently divided into four portions for preparation of the biotinyl (for ShK-biotin), fluorescein-6 carboxyl (for ShK-F6CA), tetramethylrhodamine-6 carboxyl (for ShK-TMR) or the biotinyl-(Aeea) 4 derivatives. Each of these residues was also coupled as an HOBT ester after deblocking the Fmoc group. Following N-terminal derivatization, each of the peptides was cleaved from the resin and simultaneously deprotected with reagent K (29) for 2 h at room temperature. The free peptide was then filtered to remove the spent resin beads and precipitated with ice cold diethyl ether, collected on a fine filter by suction, washed with ice cold ether and finally extracted with 20% AcOH in H 2 O. Oxidative folding of the disulfide bonds and its subsequent purification were as previously described with the addition of 25% MeOH to the solution to maintain solubility (23). Oxidative folding was facilitated by addition of 1.5 mM reduced glutathione and 0.75 mM oxidized glutathione. Each sample was purified by preparative RP-HPLC using a Rainin Dynamax C18 column. HPLC-pure fractions for each sample were pooled and lyophilized. Structures and purity of all analogs were confirmed by HPLC, amino acid and MALDI-TOF analysis.
The PAS T cell line was a kind gift from Dr. Evelyne Béraud (Marseille, France). This long-term cell line, specific for MBP, was generated in Lewis rats and expresses only two types of K + channels: Kv1.3 and IKCa1 (7). Once activated with MBP and injected into naïve Lewis rats, PAS T cells induce EAE (7,12,32). They were maintained in culture by alternating rounds of antigen-induced activation and rounds of expansion in IL-2 containing medium. For antigen stimulation, PAS T cells (3 x 10 5 /ml) were incubated for 2 days with 10 µg/ml of MBP and 15 x 10 6 /ml syngeneic irradiated (2500 rads) thymocytes as antigen-presenting cells (APCs) in RPMI 1640 Dutch modification containing 4 mM glutamine, 1 mM Na pyruvate, 1% non essential amino acids, 1% RPMI vitamins, 100 U/ml penicillin, 100 µg/ml streptomycin, and 50 µM β-mercaptoethanol (basic medium) supplemented with 1% syngeneic rat serum. For the IL-2dependent growth phase, PAS cells were seeded in basic medium supplemented with 10% FCS and 5% T cell growth factor (TCGF). After 5 days of expansion in this medium PAS T cells were restimulated with MBP. TCGF was produced by activating Lewis rat (5-8 weeks old; Charles River Laboratories, Wilmington, MA) splenocytes (2 x 10 6 /ml) with 2 µg/ml Con A in basic medium supplemented with 10% FCS. After 48 hours cells were pelleted and 15 mg/ml αmethyl mannoside (Sigma) added to the supernatant to inactivate Con A. After thorough mixing the supernatant was passed through a 0.2 µm filter and stored at -20ºC.
Rat and human mononuclear cells were isolated from the spleens of Lewis rats or from the blood of healthy volunteers, and enriched for T cells by nylon wool purification for rat T cells or with CD3 + RosetteSep (StemCell Technologies, Vancouver) for human T cells. Rat T cells were activated with 5 µg/ml Con A; human T cells were activated with 50 ng/ml anti-CD3 Ab (Biomeda) in the presence of autologous irradiated (2500 rad) peripheral blood mononuclear cells (PBMC) as APCs for 48 hours. Human TT-or MBP-specific T cells were generated from PBMCs from a healthy volunteer. Cells (2x10 8 ) were stimulated with 10 µg/ml TT or MBP.
After 2 days 5% TCGF was added and the cells were expanded for 5 days. Cells were restimulated in regular 7-day cycles with TT or MBP in the presence of autologous irradiated PBMCs. Experiments were performed after the 9 th stimulation when 95% of cells expressed an effector memory phenotype.

Pharmacological analysis of the pathways involved in the up-regulation of Kv1.3 channels -
Rested PAS T cells were incubated with 100 nM CsA, 10 nM staurosporine, or 100 nM MgTX for 1 hour. MBP and APCs were added for a further 20-30 hours to activate the cells before patch-clamp and flow cytometry analysis (see above). Statistical analysis was carried out using the non-parametric Mann-Whitney U-test. were four-fold less potent than ShK (Figs. 2B, 2E, 2F). Although the affinity of ShK-biotin for Kv1.3 (K d = 11 ± 2 pM) was comparable to that of ShK, it was completely ineffective when preassembled with phycoerythrin-conjugated streptavidin (Figs. 2C, 2D, 2F), probably because the complex was too large to reach the binding site in the channel pore. Attachment of biotin via a longer linker (33 Å in length) to ShK dramatically reduced its affinity for Kv1.3 (K d >100 nM; not shown in graph). All these polypeptides blocked Kv1.3 with a Hill coefficient of 1. These results demonstrate that ShK retains its ability to block the Kv1.3 channel at picomolar concentrations after the attachment of a fluorophore to Arg 1 .

ShK-F6CA staining and flow cytometry detects Kv1.3 channels in mammalian cells-ShK-F6CA
stained L929 cells stably expressing roughly 2000 Kv1.3 channels/cell, the fluorescence signal in these cells being clearly distinguishable from that in unstained cells (Fig. 4A, left). The addition of an anti-fluorescein Ab conjugated to Alexa-488 (Molecular Probes) did not increase the intensity of the stain, probably because the size of the Ab prevented it from reaching ShK-F6CA docked in the channel vestibule (data not shown). ShK-TMR also stained these cells although less brightly (data not shown), possibly because TMR is a less bright fluorophore with a lower quantum yield than F6CA. An excess of unlabeled Kv1.3 inhibitors (ShK, MgTX and ChTX) competitively inhibited ShK-F6CA staining, whereas an inhibitor of small-conductance calciumactivated K + channels (apamin) had no effect (Fig. 4B). The specificity of ShK-F6CA for cells with Kv1.3 channels was further confirmed by the lack of staining of L929 cells stably expressing an equivalent number of Kv1.1 (Fig. 4A, right) or Kv3.1 (not shown) channels. We extended these findings to human T lymphocytes that correspond to the three groups of cells examined in rats. Freshly isolated unstimulated peripheral blood T lymphocytes expressed about 300-400 channels per cell and did not stain with ShK-F6CA (Fig. 5, (Fig. 6A). The channel levels remained elevated for the next 48 hours during which time PAS cells are at their peak of encephalitogenicity (11,32).

Identification of Kv1.3 high T cells by ShK
Following the addition of TCGF at the 48 th hour Kv1.3 levels progressively declined to a baseline of <500 channels/cell on day-8 paralleling the decrease in encephalitogenicity (11,32). and we had ascertained in other experiments that this procedure was sufficient to wash out MgTX (data not shown). ShK or ShK-F6CA could not be used for this experiment since it was not possible to completely wash them out (5). Suppression of Kv1.3-up-regulation by MgTX is probably due to attenuation of the calcium-signaling cascade upstream to the point of interruption by CsA (4,37). Taken together, these results indicate that Kv1.3 up-regulation requires the activation of both the calcium and PKC-dependent signaling pathways.

T cells expressing unusually high levels of the Kv1.3 channel have recently been
implicated in an animal model for MS (7) and we now describe a tool for the detection of Myelin-specific T cells in MS patients are reported to exhibit properties of memory T cells (38)(39)(40), and such cells could potentially contribute to the pathogenesis in MS because they traffic directly to inflamed tissues and release copious amounts of inflammatory cytokines such as interferon-γ and tumor necrosis factor-α. Kv1.3 high expression may therefore be a functional marker for such pathogenic myelin-reactive memory cells. In keeping with this idea, we have previously reported that MBP-specific rat memory T cells express higher levels of Kv1.3 channels than naïve rat T cells and induce severe EAE following adoptive transfer into rats (7). Generalization of this approach of tagging polypeptide inhibitors of ion channels with fluorophores could lead to the development of novel reagents for flow cytometry. For example, primary acute myeloid leukemia and hematopoietic cell lines aberrantly express increased levels of HERG K + channels, which have been suggested to regulate proliferation in these cells (41,42). integrins (46). The channel may also participate in signaling at the immunological synapse through possible interactions with PKC and p56 lck (47). Through the use of tools such as ShK-F6CA it may be feasible to visualize fluorescently tagged Kv1.3 channels at the immunological synapse during antigen presentation. Fluorescent probes that directly mark functional channels in the membrane may therefore find utility as diagnostic tools that can distinguish between subsets of cells with varying channel phenotypes and as experimental tools to locate channels within the cell during locomotion or immunological synapse formation.