Targeting Effector Memory T Cells with a Selective Peptide Inhibitor of Kv1.3 Channels for Therapy of Autoimmune Diseases

Abstract

The voltage-gated Kv1.3 K⁺ channel is a novel target for immunomodulation of autoreactive effector memory T (T_EM) cells that play a major role in the pathogenesis of autoimmune diseases. We describe the characterization of the novel peptide ShK(L5) that contains l-phosphotyrosine linked via a nine-atom hydrophilic linker to the N terminus of the ShK peptide from the sea anemone Stichodactyla helianthus. ShK(L5) is a highly specific Kv1.3 blocker that exhibits 100-fold selectivity for Kv1.3 (K_d = 69 pM) over Kv1.1 and greater than 250-fold selectivity over all other channels tested. ShK(L5) suppresses the proliferation of human and rat T_EM cells and inhibits interleukin-2 production at picomolar concentrations. Naive and central memory human T cells are initially 60-fold less sensitive than T_EM cells to ShK(L5) and then become resistant to the peptide during activation by up-regulating the calcium-activated K_Ca3.1 channel. ShK(L5) does not exhibit in vitro cytotoxicity on mammalian cell lines and is negative in the Ames test. It is stable in plasma and when administered once daily by subcutaneous injection (10 μg/kg) attains “steady state” blood levels of ∼300 pM. This regimen does not cause cardiac toxicity assessed by continuous EKG monitoring and does not alter clinical chemistry and hematological parameters after 2-week therapy. ShK(L5) prevents and treats experimental autoimmune encephalomyelitis and suppresses delayed type hypersensitivity in rats. ShK(L5) might prove useful for therapy of autoimmune disorders.

Autoimmune diseases afflict millions worldwide and may have a common pathogenic mechanism. Pathogenesis may involve the “awakening” of dormant disease-specific autoreactive T cells—for instance myelin-specific T cells in patients with multiple sclerosis (MS)—by molecular mimicry or other undetermined mechanisms. Once awakened, autoreactive T cells might differentiate from a naive state into continuously activated memory T cells as a consequence of repeated autoantigen stimulation and contribute to inflammatory damage by migrating rapidly into tissues, secreting inflammatory cytokines, and exhibiting immediate effector function (Sallusto et al., 1999). Several lines of evidence support this scheme. First, a majority of myelin-specific T cells from patients with MS are costimulation-independent activated effector memory T (T_EM) cells (Lovett-Racke et al., 1998; Scholz et al., 1998; Markovic-Plese et al., 2001; Wulff et al., 2003). Second, transfer of myelin-specific T_EM cells into naive rat recipients induces experimental autoimmune encephalomyelitis (EAE), a model for MS (Beeton et al., 2001b). Third, T cells from patients with diabetes mellitus type 1 that are specific for the disease-associated autoantigen glutamic acid decarboxylase 65 are continuously activated memory cells (Viglietta et al., 2002). Last, a majority of T cells in the synovium of patients with rheumatoid arthritis and in skin lesions of patients with psoriasis are T_EM cells, as are T cells that cause delayed type hypersensitivity (DTH) (Ezawa et al., 1997; Friedrich et al., 2000; Soler et al., 2003). Memory B cells, especially those belonging to the class-switched CD27⁺IgD^- subset, also probably contribute to the pathogenesis of many autoimmune diseases (Iglesias et al., 2001; O’Connor et al., 2001; Corcione et al., 2004). Therapies that target T_EM and class-switched memory B cells without impairing the activity of other lymphocyte subsets would therefore target the pathogenic cells in patients with autoimmune diseases but without compromising acute immune responses.

An exciting new therapeutic target for immunomodulation of T_EM and class-switched memory B cells is the voltage-gated Kv1.3 K⁺ channel. T_EM cells up-regulate Kv1.3 upon activation and their antigen-driven proliferation is exquisitely sensitive to Kv1.3 blockers (Wulff et al., 2003). Naive and T_CM cells in contrast are significantly less sensitive to Kv1.3 blockers to begin with and rapidly become resistant to Kv1.3 blockade by up-regulating the calcium-activated K⁺ channel K_Ca3.1 (Wulff et al., 2003; Chandy et al., 2004). B cells, like T cells, change their potassium channel dependence from K_Ca3.1 to Kv1.3 as they differentiate from naive into class-switched CD27⁺IgD^- memory B cells (Wulff et al., 2004). Kv1.3 blockers inhibit the proliferation of these cells without affecting naive and CD27⁺IgD⁺ memory B cells. By targeting T_EM cells and class-switched memory B cells with Kv1.3 blockers, it might be possible to ameliorate autoimmune diseases without compromising the bulk of the immune response. The functionally restricted tissue distribution of Kv1.3 and the fact that, in vivo, Kv1.3 blockade ameliorates EAE, bone resorption in periodontal disease, and DTH in animal models without causing obvious side effects has enhanced the attractiveness of Kv1.3 as a therapeutic target (Koo et al., 1997; Beeton et al., 2001b; Valverde et al., 2004). Although Kv1.3 blockers would suppress all activated T_EM (for example, T_EM cells specific for vaccine antigens) and class-switched memory B cells, a Kv1.3-based therapy would be a significant improvement over current therapies that broadly and indiscriminately suppress the entire immune system. An additional advantage of Kv1.3 blockers is that they are reversible. Thus, one could titrate the therapeutic effect of Kv1.3 blockers when needed and stop therapy in the face of infection, unlike chemotherapeutic agents or targeted monoclonal antibody therapies, which take months to subside.

Despite extensive efforts, selective and potent inhibitors of the Kv1.3 channel have not been developed (Chandy et al., 2004). The most potent Kv1.3 inhibitor is the peptide ShK from the Caribbean sea anemone Stichodactyla helianthus (Pennington et al., 1995). ShK blocks Kv1.3 (K_d, ∼10 pM), suppresses proliferation of T_EM cells at picomolar concentrations (Wulff et al., 2003), and ameliorates EAE (Beeton et al., 2001b). A potential drawback of ShK is its affinity (K_d, 28 pM) for the neuronal Kv1.1 channel (Kalman et al., 1998). Although no side effects were observed with ShK in EAE trials (Beeton et al., 2001b), ingress of high concentrations of ShK into the brain could lead to unwanted neurotoxicity. Other inhibitors including correolide, trans-PAC, CP-339818, UK-78282, Psora-4, margatoxin, and luteolin are less selective for Kv1.3 (Chandy et al., 2004). The development of a more selective ShK derivative is therefore necessary.

We have developed ShK(L5), a synthetic analog of ShK that blocked Kv1.3 with picomolar affinity and exhibited greater than 100-fold selectivity for Kv1.3 over Kv1.1 and other channels. Selectivity was achieved by attaching a negatively charged l-phosphotyrosine (l-pTyr) via a hydrophilic linker to ShK-Arg¹. ShK(L5) suppressed T_EM cell proliferation in vitro at picomolar concentrations without compromising the function of naive and T_CM cells. In proof-of-concept in vivo studies, ShK(L5) ameliorated EAE caused by the transfer of myelin-specific T_EM cells into naive rat recipients and suppressed the DTH response also caused by T_EM cells. ShK(L5) may have use as a therapy for multiple sclerosis and other T and B cell-mediated autoimmune diseases.

Materials and Methods

Synthesis of ShK Analogs

For the Fmoc-Pmp analogs, the ethyl protecting groups were removed by treating the Fmoc-Pmp-(Ethyl)₂-OH with aqueous 6 N HCl at reflux. After 16 h, a white solid precipitated out that was then isolated by filtration and washed with water until the washings were neutral (Hammerschmidt and Hanbauer, 2000). Partial removal of the ethyl protecting groups from Pmp-phosphonate resulted in either Pmp, Pmp-Et, or Pmp(Ethyl)_2. All Fmoc-amino acid derivatives were obtained from Bachem AG (Bubendorf, Switzerland), except for Fmoc-d-TyrPO₃(benzyl)-OH, which was obtained from Nova Biochem (San Diego, CA), and Fmoc-Pmp(Ethyl)-OH and Fmoc-d-Pmp(Ethyl)₂-OH, which were obtained from Chem Impex (Wood Dale, IL). Solid-phase assembly was initiated with Fmoc-Cys(Trt)-2-chlorotrityl resin to minimize potential racemization of the C-terminal Cys residue (Fujiwara et al., 1994). Automated assembly was carried out on an ABI-431A peptide synthesizer (Applied Biosystems, Foster City, CA). Fmoc-Aeea-OH was coupled to the N terminus after the assembly of ShK. The resin was divided into nine aliquots. Either Fmoc-Tyr(PO₃Bzl)-OH, Fmoc-d-Tyr(PO₃Bzl)-OH, Fmoc-Tyr(PO₃Me₂)-OH, Fmoc-Pmp-OH, Fmoc-d-Pmp-OH, Fmoc-Pmp(Ethyl)-OH, Fmoc-Pmp(Et)₂-OH, Fmoc-Tyr(tert-butyl)-OH, or Fmoc-p-amino-l-phenylalanine(tert-butyloxycarbonyl)-OH was coupled, using diisopropylcarbodiimide and 1-hydroxybenzotriazole, to one of the resin aliquots. The deblocked peptide resin was cleaved and deprotected with reagent K (King et al., 1990) containing 5% triisopropylsilane for 2 h at RT. Met(O) was reduced by addition of solid NH₄I to the cleavage cocktail at t - 15 min. (Nicolas et al., 1995). For the peptides containing Tyr(PO₃Me₂)-OH, a cleavage cocktail containing 1 M trimethylsilyl bromide in TFA (trifluoroacetic acid) containing thioanisole as a scavenger for 18 h at 4°C was used (Tian et al., 1993). Incomplete removal of the methyl protecting groups is common when using this method, and two of the species [p-phosphotyrosine) and Tyr(PO₃HMe)] are easily purified by RP-HPLC. The Tyr(PO₄Me₂) containing analog was cleaved via standard Reagent K cleavage keeping both methyl groups intact. In each case, the cleavage mixture was filtered and the crude peptide was precipitated into ice-cold diethyl ether. The precipitate was collected, yielding approximately 75 mg of peptide from 200 mg of resin. The crude product was dissolved in 20 ml of 50% aqueous AcOH and diluted into 0.75 l of H₂O. The pH of the solution was adjusted with NH₄OH to 8.2, and it was allowed to fold overnight with the addition of glutathione (2 mM:1 mM) (reduced:oxidized). All analogs were purified using RP-HPLC using a linear gradient of water versus acetonitrile buffered with trifluoroacetic acid as described previously (Pennington et al., 1995, 1996a,b). Pure fractions were pooled and lyophilized, resulting in a trifluoroacetate salt of each peptide. Purity of the peptides was greater than 95%. Each sample was confirmed by reversed-phase high-performance liquid chromatography, amino acid analysis, and matrix-assisted laser desorption ionization/time of flight mass spectrometry and adjusted to account for peptide content before bioassay. ShK and margatoxin were obtained from Bachem Biosciences (King of Prussia, PA). Luteolin was purchased from Sigma-Aldrich (St Louis, MO).

Ion Channels

We used the IUPHAR nomenclature for the ion channels described in this article (Gutman et al., 2003). Cells stably expressing mKv1.1, rKv1.2, mKv1.3, hKv1.5, and mKv3.1 have been described previously (Grissmer et al., 1994). Cell lines stably expressing other mammalian ion channels were gifts from several sources: mKv1.7 in CHL cells and hKCa2.3 in COS-7 cells from Aurora Biosciences Corp. (San Diego, CA); hKv1.4 in LTK cells from Michael Tamkun (University of Colorado, Boulder, CO); hKv2.1 in HEK293 cells from Jim Trimmer (University of California, Davis, CA); Kv11.1 (HERG) in HEK293 cells from Craig January (University of Wisconsin, Madison, WI); HEK293 cells expressing hKCa1.1 or rKCa2.1 or hKCa3.1 from Khaled Houamed (University of Chicago, Chicago, IL); hNav1.4 in HEK-293 cells from Frank Lehmann-Horn (University of Ulm, Germany), and Cav1.2 in HEK-293 cells from Franz Hofmann (Munich, Germany). RBL-2H3 (expressing Kir2.1) and N1E-115 neuroblastoma cells (expressing Nav1.2) were obtained from the American Type Culture Collection (Manassas, VA). hKv1.6 and rKv3.2 (both in pcDNA3) were obtained from Protinac GmbH (Hamburg, Germany) and transiently-transfected into COS-7 cells with Fugene-6 (Roche, Mannheim, Germany) according to the manufacturers’ protocol.

Lymphoid Cells and Cell Lines

Histopaque-1077 gradients (Sigma-Aldrich) were used to isolate splenocytes from Lewis rats and human peripheral blood mononuclear cells (PBMCs) from the blood of healthy volunteers. Human myelin oligodendrocyte glycoprotein- or tetanus toxoid-specific T_EM cells were generated as described previously (Wulff et al., 2003). The encephalitogenic CD4⁺ Lewis rat T cell line PAS (Beraud et al., 1993) was a gift from Evelyne Béraud (University of Marseille, Marseille, France), and RPMI 8226 plasmacytoma cells were a gift from Shastri Gollapudi (University of California, Irvine, CA). Jurkat and Burkitt cells were purchased from the American Type Culture Collection.

Electrophysiological Analysis

Experiments were conducted in the whole-cell configuration of the patch-clamp technique. K_V currents were elicited by 200-ms depolarizing pulses from a holding potential of -80 to 40 mV as described previously (Zhou et al., 1998; Wulff et al., 2000; Bardien-Kruger et al., 2002; Kolski-Andreaco et al., 2004; Vennekamp et al., 2004). Each channel blocker was tested at multiple concentrations. The measured reduction in peak current at 40 mV for each concentration was used to generate a dose-response curve, and the K_d and Hill coefficient were determined with Origin software (OriginLab Corp., Northampton, MA) as described previously (Zhou et al., 1998; Wulff et al., 2000; Bardien-Kruger et al., 2002; Kolski-Andreaco et al., 2004; Vennekamp et al., 2004). The K_d was also determined from the on (T_ON) and off (T_OFF) rates for channel block. After stabilization of peak Kv1.3 current amplitude, 70 pM ShK(L5) was perfused onto the cell, and peak current values plotted as a function of time were fitted to a single exponential function to determine T_ON. After reaching equilibrium block, perfusion was switched back to blocker-free bath solution. Peak currents were plotted as described above to determine T_OFF. K_ON, K_OFF, and K_d were calculated assuming a simple bimolecular reaction between ShK(L5) and Kv1.3: K_ON = 1 - T_ON × K_OFF / [T_ON × ShK(L5) concentration]; K_OFF = 1 / T_OFF; K_d = K_OFF / K_ON (Peter et al., 2001). For Kv11.1 channels, current block was measured both at 20 and -50 mV (tail current). For K_Ca channels, K_ir2.1, and swelling-activated chloride currents, we measured the change in slope conductance by the ShK analogs and, for Na⁺ and Ca²⁺ currents, the reduction of minimum current.

Staining for Flow Cytometry and Fluorescence Microscopy

The T cell phenotypes of human PBMCs, rat splenocytes, and PAS T cells were determined by flow cytometry. PBMCs were triplestained with anti-CD3 antibody conjugated to Cy-chrome (BD Pharmingen, San Diego, CA), anti-CD45RA antibody conjugated to phycoerythrin (BD Pharmingen), and anti-CCR7 antibody conjugated to fluorescein isothiocyanate (R&D Systems, Minneapolis, MN). Rat splenocytes and PAS-stained T cells were double-stained with anti-CD3 antibody conjugated to Cy-chrome and anti-CD45RC antibody conjugated to fluorescein isothiocyanate (BD Pharmingen). The stained cells were analyzed with a FACScan (BD Biosciences, San Jose, CA).

Two approaches were used to evaluate Kv1.3 protein expression in PAS T cells. First, PAS cells were stained with ShK-F6CA (10 nM; Bachem Bioscience Inc.), a fluorophore-tagged ShK analog, and analyzed by flow cytometry as described previously (Beeton et al., 2003). For competition experiments, PAS cells were preincubated with excess unlabeled ShK(L5) (100 nM) before addition of 10 nM ShK-F6CA. Second, PAS cells were permeabilized and stained with anti-Kv1.3 antibody (Koch et al., 1997) (gift from Hans-Gunther Knaus, Innsbruck, Austria) followed by a secondary antibody conjugated to Alexa-488 (Molecular Probes, Eugene, OR). Stained cells were visualized with a Zeiss LSM-510 META confocal microscope (Carl Zeiss GmbH, Jena, Germany), fluorescence intensities were measured for individual cells (n = 10-15), and statistical analysis carried out using the Mann-Whitney U test.

Functional Studies

Proliferation of human and rat T cells was determined with [³H]thymidine incorporation assays as described previously (Beeton et al., 2001a,b; Wulff et al., 2003). For measurements of IL2 production, PAS T cells were activated with MBP in the presence or the absence of ShK or ShK(L5) for 8 h, and the culture supernatants were collected as described previously (Beeton et al., 2001a). IL2 was detected in supernatants using the rat IL2 Quantikine kit (R&D Systems) according to manufacturer’s instructions. Effect of exogenous IL2 (20 units/ml; Sigma-Aldrich, St Louis, MO) on proliferation of PAS T cells was determined as described previously (Beeton et al., 2001a).

Circulating Half-Life Determination and Plasma Stability

Known amounts of ShK(L5) were added to Lewis rat serum, and the blocking activity on Kv1.3 channels was tested by patch-clamp to establish a standard dose-response curve. Serum samples from Lewis rats obtained at various times after single subcutaneous or intravenous injections of ShK(L5) were tested for Kv1.3 blocking activity by patch-clamp and the levels of ShK(L5) determined from the standard curve as described previously (Beeton et al., 2001b). In other experiments, Lewis rats received single daily injections of 10 μg/kg ShK(L5) and serum levels of ShK(L5) were determined 24 h after each injection on days 1 to 5. To determine plasma stability of ShK(L5), rat plasma spiked with a known amount of ShK(L5) was incubated at 37°C for varying durations and then tested for Kv1.3 blocking activity; the amount of residual ShK(L5) in these samples was determined from the standard curve.

Cytotoxicity Assays and Ames Test

Human PBMCs, PAS, Jurkat, RPMI 8226, and Burkitt cells were grown for 48 h in the presence or the absence of 100 nM ShK(L5). Cells were then stained with the LIVE/DEAD viability/cytotoxicity kit (Molecular Probes) according to the manufacturer’s instructions, and the percentage of live and dead cells was determined using a fluorescence microplate reader (CytoFluor; Applied Biosystems, Foster City, CA). Triton X-100 (0.1%) was used as a positive control for cell death. For the Ames test, the mutagenic activity of ShK(L5) was determined on the Salmonella typhimurium tester strain TA97a by Nelson Laboratories (Salt Lake City, UT).

EKG Studies to Evaluate Cardiac Toxicity

Electrocardiographic studies with implanted EKG transmitters (Data Sciences International, Arden Hills, MN) were used for heart rate variability analysis in animals that received ShK(L5) or vehicle. Experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee of UC Davis. Six Lewis rats (9-11 weeks old; weight = 219 ± 9 g) were anesthetized with a mixture of ketamine (80 mg/kg) and xylazine (7.5 mg/kg) administered by intramuscular injection. EKG transmitters were placed in the peritoneal cavity of each rat and two EKG leads were tunneled subcutaneously to the right shoulder and to the xiphoid space caudal to the ribcage. An analgesic, carprofen (5 mg/kg, subcutaneous), was administered at the end of surgery. Two weeks after surgery, we performed baseline EKG recording on the rats for 2 h (day 1). We then injected the vehicle (PBS + 2% rat serum) subcutaneously and continued recording for another 8 h. The animals were then returned to their rooms. On day 2, we performed baseline EKG recording for 2 h after which ShK(L5) (10 μg/kg dissolved in vehicle) was injected subcutaneously and EKG recording continued for another 8 h. Data recorded from 1.5 to 3.5 h after the injections were used for standard heart rate variability parameters-analysis in both time and frequency domains using Nevokard software (Bio-Impedance Technology, Inc., Chapel Hill, NC).

Subchronic Toxicity Studies

Lewis rats (9-11 weeks old; weight, 199 ± 7 g) received subcutaneous injections of ShK(L5) 10 μg/kg/day (n = 6) or saline (n = 6) for 2 weeks. Rats were weighed daily. The Comparative Biology Laboratory at University of California, Davis performed chemical (COBAS MIRA Plus; Roche Diagnostic Systems, Branchburg, NJ) and hematological (HEMAVET® 850 Multispecies Hematology Analyzer; CDC Technologies, Oxford, CT) analysis on blood samples drawn at the end of 2 weeks. Single-cell suspensions prepared from thymuses and spleens removed from six animals given ShK(L5) and six given saline were stained with antibodies specific for various T and B cell markers (BD Pharmingen) and analyzed by flow cytometry.

Prevention and Treatment of Acute Adoptive EAE and Prevention of DTH in Lewis Rats

Female inbred Lewis rats 9 to 11 weeks old were purchased from Harlan-Sprague-Dawley (Indianapolis, IN) and housed under barrier conditions with irradiated rodent chow and acidified water ad libitum. All experiments were in accordance with National Institutes of Health guidelines and approved by the Institutional Animal Care and Use Committee at the University of California, Irvine. ShK(L5) was dissolved in PBS + 2% Lewis rat serum (saline) for subcutaneous injection. Acute adoptive EAE was induced as described previously (Beeton et al., 2001a,b) with 6 to 8 × 10⁶ myelin basic protein (MBP)-activated PAS cells. MBP was extracted from frozen guinea pig spinal cords (Harlan Bioproducts, Indianapolis, IN) as described previously (Deibler et al., 1972). Rats were weighed daily and observed twice daily for clinical signs of EAE. For prevention trials, rats received 10 μg/kg/day ShK(L5) from days 0 to 5, whereas control rats received saline. For treatment trials, administration of ShK(L5) (10 μg/kg/day) or saline was begun after the onset of disease (rats had a limp tail, were hunched, and had lost ≥6% of their weight over 24 h) and continued for 3 days.

For DTH trials, Lewis rats were immunized with an emulsion of ovalbumin in complete Freund’s adjuvant (Difco, Detroit, MI). Seven days later, they received an injection of ovalbumin dissolved in saline in the pinna of one ear and saline in the other ear. Rats then received subcutaneous injections of ShK(L5) (10 μg/kg/day) or vehicle (PBS + 2% Lewis rat serum). Ear swelling was measured 24 and 48 h later using a spring-loaded micrometer (Mitutoyo, Spokane, WA).

Results

ShK(L5), a Novel ShK Analog that Exhibits 100-fold Selectivity for Kv1.3 Over Kv1.1

ShK blocks the neuronal Kv1.1 channel and the Kv1.3 channel with roughly equivalent potency. Neurotoxicity is therefore a concern under circumstances that compromise the blood-brain barrier and allow the entry of sufficient amounts of ShK to block Kv1.1 channels. Our strategy to design a Kv1.3-specific inhibitor was guided by our finding that ShK-F6CA containing fluorescein-6-carboxylate (F6CA), attached through a 20-Å Aeea linker to the N terminus of ShK exhibited 80-fold selectivity for Kv1.3 over Kv1.1 (Beeton et al., 2003). Because F6CA can exist as a restricted carboxylate or also as a cyclized lactone, it was not clear whether the Kv1.3 specificity of ShK-F6CA was a result of the negative charge of F6CA, the hydrophobicity created by this large bulky fluorescein nucleus, potential planar π-π electronic stacking, or a combination of all of these potential contributions. To distinguish between these possibilities and with the intention of developing a nonfluorescent Kv1.3-selective inhibitor, we generated a series of 12 novel N-terminally-substituted ShK analogs to probe some of these interactions. By attaching tyrosine, phenylalanine, or their derivatives (varying in charge, size, and hydrophobicity) through an Aeea linker to the N terminus of ShK, we could probe the effects of charge and hydrophobicity to gain insight into our selectivity enhancement seen with F6CA substitution.

In the example shown in Fig. 1A, l-phosphotyrosine (l-pTyr) a negatively charged (net charge -2) post-translationally modified aromatic amino acid, was attached via the AEEA linker to ShK-Arg¹ to generate a novel analog called ShK(L5). ShK and ShK(L5) were tested on Kv1.3 and Kv1.1 channels stably expressed in L929 cells. Figure 1B shows the effects of ShK and ShK(L5) on Kv1.3 and Kv1.1 currents elicited by 200-ms depolarizing pulses from a holding potential of -80 to 40 mV. Both peptides reversibly blocked Kv1.3 and Kv1.1 in a dose-dependent manner with Hill coefficients of 1 (Fig. 1, B-D). K_d values were determined from the dose-response curves shown in Fig. 1C using Origin software. ShK blocked Kv1.3 (K_d = 10 ± 1 pM) and Kv1.1 (K_d = 28 ± 6 pM) with roughly equivalent potency, as expected (Fig. 1C). In contrast, ShK(L5) was 100-fold selective for Kv1.3 (K_d = 69 ± 5 pM) over Kv1.1 (K_d = 7.4 ± 0.8 nM) (Fig. 1, B and C). The time course of Kv1.3 current block by ShK(L5) and its washout is shown in Fig. 1D. The time constant (T_ON) of ShK(L5) wash-in was 131 ± 21 s (n = 7), whereas the time constant (T_OFF) for peptide wash-out was 150 ± 28 s (n = 4). The K_d (57 ± 7 pM) calculated from the K_ON (15 × 10⁶± 0.5 × 10⁶ M^-1s^-1) and K_OFF (0.0059 ± 0.0013 s^-1) values is consistent with the K_d (69 ± 5 pM) determined with the use of Origin software.

Fig. 1.

Generation of a selective Kv1.3 blocker. A, molecular model of ShK based on the published NMR structure. The Lys²², critical for channel blockade, is highlighted in orange. l-pTyr was attached to the α-amino group of Arg¹ of ShK (highlighted in cyan) through an Aeea linker (right). The structures of the linker and l-pTyr were modeled with AM1 in Hyperchem. B, effect of ShK (top) and ShK(L5) (bottom) on Kv1.3 and Kv1.1 currents in stably transfected cells. C, dose-dependent inhibition of Kv1.3 (open symbols) and Kv1.1 (closed symbols) by ShK (blue) and ShK(L5) (red). K_d values on Kv1.3 = 10 ± 1 pM (ShK) and 69 ± 5 pM (ShK(L5)); K_d values on Kv1.1 = 28 ± 6 pM (ShK) and 7.4 ± 0.8 nM (ShK(L5)). D, time course of wash-in and wash-out of ShK(L5) on Kv1.3. Cells were held at a holding potential of -80 mV and depolarized for 200 ms to 40 mV every 30 s. E, K_d values shown for inhibition of Kv1.3 and Kv1.1 by ShK analogs. K_d values for ShK-F6CA and ShK-Dap²² are from published sources (Kalman et al., 1998; Beeton et al., 2003; Chandy et al., 2004).

Other ShK analogs were tested on Kv1.3 and Kv1.1 channels (Fig. 1E). ShK(D5) containing d-phosphotyrosine was 35-fold selective for Kv1.3 over Kv1.1 but was an order of magnitude less potent than ShK(L5). ShK(L6) containing l-pTyr-monomethyl showed modest (11-fold) Kv1.3 specificity, whereas ShK analogs containing l-pTyr-dimethyl or l-Tyr were not selective for Kv1.3 over Kv1.1 (Fig. 1E). Analogs that contained phenylalanine or its derivatives (varying in bulk, π electron density, and charge) were modestly specific or not specific for Kv1.3 over Kv1.1 (Fig. 1E). The 100-fold specificity of ShK(L5) for Kv1.3 over Kv1.1 is greater than that of ShK-F6CA (80-fold), ShK(D5) (35-fold), ShK-Dap²² (33-fold), or any other ShK analog tested (Fig. 1C).

ShK(L5) Is a Highly Specific Kv1.3 Inhibitor

We assessed the specificity of ShK(L5) on a panel of 20 ion channels (Table 1). ShK(L5) blocked the Kv1.3 channel in T cells with a K_d (76 pM) equivalent to its K_d on the cloned channel (69 pM). It was 100-fold selective for Kv1.3 over Kv1.1, 260-fold selective over Kv1.6, 280-fold selective over Kv3.2, 680-fold selective over Kv1.2, and >1000-fold selective over all other channels tested. It is noteworthy that it was 1600-fold Kv1.3-selective over KCa3.1, the calcium-activated K⁺ channel that regulates activation of human naive and T_CM cells (Wulff et al., 2003). Native ShK was less selective than ShK(L5). ShK was 2.8-fold selective for Kv1.3 (K_d = 10 ± 1 pM) over Kv1.1 (K_d 28 ± 6 pM), 20-fold selective over Kv1.6 (200 ± 20 pM), 500-fold selective over Kv3.2 (K_d = 5000 ± 1000 pM), and >1000-fold selective-over Kv1.2 (10 ± 1 nM) and KCa3.1 (K_d = 28 ± 3 nM). Margatoxin, a peptide from scorpion venom that has been touted as a specific Kv1.3 inhibitor (Lin et al., 1993; Koo et al., 1997; Middleton et al., 2003) was also not specific. It was 5-fold selective for Kv1.3 (110 ± 12 pM) over Kv1.2 (K_d = 520 ± 1 pM), 9-fold selective over Kv1.1 (10 ± 1 nM), and >1000-fold selective over Kv1.6 and Kv3.2 (K_d > 100 nM). Luteolin, a nutriceutical sold for autoimmune diseases (http://www.lutimax.com) on the basis of its being a Kv1.3 inhibitor (Lahey and Rajadhyaksha, 2004), blocked Kv1.3 weakly (K_d = 65 ± 5 μM) and exhibited no selectivity over Kv1.1 (K_d = 77 ± 5 μM), Kv1.2 (K_d = 63 ± 4 μM), or Kv1.5 (K_d = 41 ± 3 μM). The exquisite specificity of ShK(L5) for Kv1.3, together with its picomolar affinity for the channel, makes it a potentially attractive immunosuppressant.

TABLE 1.

Selectivity of ShK(L5)

Channels	Kd of ShK(L5)
	pM
Kv1.1	7000 ± 1000
Kv1.2	48,000 ± 7000
Kv1.3 (cloned)	69 ± 5
Kv1.3 (native)	76 ± 8
Kv1.4	137,000 ± 3000
Kv1.5	100,000 (N.E.)
Kv1.6	18,000 ± 3000
Kv1.7	100,000 (N.E.)
Kv2.1	100,000 (N.E.)
Kv3.1	100,000 (N.E.)
Kv3.2	20,000 ± 2000
Kir2.1	100,000 (N.E.)
Kv11.1 (HERG)	100,000 (N.E.)
KCa1.1	100,000 (N.E.)
KCa2.1	100,000 (N.E.)
KCa2.3	100,000 (N.E.)
KCa3.1	115,000 ± 5000
Nav1.2	100,000 (N.E.)
Nav1.4	100,000 (N.E.)
Swelling-activated T cell Cl- channel	100,000 (N.E.)
Cav1.2	100,000 (N.E.)

N.E., no effect.

ShK(L5) Preferentially and Persistently Suppresses Human T_EM Cell Proliferation

To assess the in vitro immunosuppressive activity of ShK(L5), we compared its ability to suppress anti-CD3 antibody-stimulated proliferation of human T_EM cell lines versus human PBMCs that contain a mixture of naive and T_CM cells. Flow cytometry confirmed the cell surface phenotypes of the two populations studied. The T_EM lines were >90% CCR7^-CD45RA^- (Fig. 2A), whereas PBMCs contained 65% CCR7⁺CD45RA⁺ (naive) and 18% CCR7⁺CD45RA^- (T_CM) cells (Fig. 2B). Figure 2C shows that ShK(L5) and ShK were 60-fold more effective in suppressing the proliferation of T_EM cells (IC₅₀ = ∼80 pM) compared with PBMCs (IC₅₀ = 5 nM, p < 0.05). The lower sensitivity of PBMCs might be explained by a rapid up-regulation of KCa3.1 channels in naive and T_CM cells upon stimulation as has been reported previously (Ghanshani et al., 2000; Wulff et al., 2003). In keeping with this interpretation, PBMCs activated for 48 h to up-regulate KCa3.1 expression, then rested for 12 h and re-activated with anti-CD3 antibody, were completely resistant to ShK(L5) block (Fig. 2D, top arrow). PBMCs that had been suppressed by ShK(L5) during the first round of stimulation exhibited identical resistance to ShK(L5) when the cells were washed, rested, and re-challenged with anti-CD3 antibody. These results corroborate an earlier report showing that naive and T_CM cells escape Kv1.3 inhibitors by up-regulating KCa3.1 channels (Wulff et al., 2003). Thus, ShK(L5) preferentially and persistently suppresses the proliferation of T_EM cells.

Fig. 2.

ShK(L5) preferentially suppresses the proliferation of human T_EM cells. Human PBMCs (A) and a human T_EM line (B) were stained with antibodies against CD3, CD45RA, and CCR7. Staining intensities of CD45RA and CCR7 were determined by flow cytometry in the CD3⁺-gated population. C, dose-dependent inhibition by ShK (blue) and ShK(L5) (red) of [³H] thymidine incorporation by PBMCs (open symbols, a mixture of naive/T_CM cells) and T_EM cells (closed symbols) stimulated for 48 h with anti-CD3 antibody. D, preactivated human PBMCs (naive/T_CM cells) that up-regulate KCa3.1 expression (Ghanshani et al., 2000) become resistant to ShK(L5) inhibition when reactivated with anti-CD3 antibody. These cells have previously been shown to become sensitive to the K_Ca3.1-specific inhibitor TRAM-34 (Ghanshani et al., 2000).

ShK(L5) Inhibits Proliferation of and IL2 Production by Rat T_EM Cells; Exogenous IL2 Partially Overrides Suppression

As a preamble to evaluating therapeutic effectiveness of ShK(L5), we examined its ability to suppress proliferation of a memory T cell line, PAS, that causes an MS-like disease in rats (Beraud et al., 1993). As a control, we used rat splenic T cells. To confirm the differentiation status of the two cell populations, we assessed the expression of CD45RC, a marker of naive T cells (Bunce and Bell, 1997). Rat splenic T cells were 76% CD45RC⁺ (i.e., mainly naive cells), whereas PAS cells were CD45RC^-, suggesting that they are memory cells (Fig. 3A). To determine whether PAS cells are in the T_EM or T_CM state, we examined Kv1.3 expression before and 48 h after activation. T_EM but not T_CM cells are expected to significantly up-regulate Kv1.3 levels upon stimulation (Beeton et al., 2001b, 2003). Patch-clamp experiments revealed a striking increase in Kv1.3 current amplitude after MBP-stimulation of PAS cells consistent with their being T_EM cells (Fig. 3B). As an independent measure of the number of Kv1.3 channels on PAS cells, we used ShK-F6CA, a fluorescently labeled ShK analog that binds specifically to Kv1.3 (Beeton et al., 2003). The intensity of ShK-F6CA staining determined by flow cytometry reflects the number of Kv1.3 tetramers expressed on the cell surface (Beeton et al., 2003). ShK-F6CA (10 nM) staining intensity increased with MBP-activation of PAS cells, and an excess of unlabeled ShK(L5) (100 nM) competitively inhibited ShK-F6CA staining (Fig. 3C). As a final test, we performed confocal microscopy on quiescent and MBP-stimulated PAS cells that had been fixed and stained with a Kv1.3-specific antibody. In keeping with data in Fig. 3, B and C, resting PAS T cells had a Kv1.3 staining intensity of 4.4 ± 0.6, and this value increased to 10.6 ± 2.3 (p < 0.005) after antigen-induced activation (Fig. 3D), showing augmentation in Kv1.3 protein expression after activation. Thus, MBP-activated PAS cells are CD45RC^- Kv1.3^high T_EM cells, whereas rat splenic T cells used in our experiments are predominantly in the naive state.

Fig. 3.

ShK(L5) preferentially suppresses the proliferation of rat T_EM cells. A, CD45RC staining of rat splenic T cells (left) and PAS T cells (right) detected by flow cytometry. B, Kv1.3 currents exhibited by quiescent (top) and myelin antigen-activated (bottom) PAS T cells. C, flow cytometry profiles of ShK-F6CA-staining in quiescent (top) and myelin antigen-activated (bottom) PAS T cells. Unstained cells (black lines) and cells stained with ShK-F6CA (green filled). Competition of ShK-F6CA staining by unlabeled ShK(L5) is red-filled. D, confocal images of Kv1.3 immunostaining in quiescent (top) and myelin antigen-activated (bottom) PAS T cells. Statistical analysis was carried out using the Mann-Whitney U test. E, dose-dependent inhibition by ShK (blue) and ShK(L5) (red) of [³H]thymidine incorporation by rat naive/T_CM (open symbols) and T_EM (closed symbols) cells activated with Con A (1 μg/ml). F, dose-dependent inhibition by ShK (blue) and ShK(L5) (red) of IL2 secretion by PAS T cells 7 h after stimulation with MBP. G, ShK(L5)-induced inhibition of myelin-antigen triggered [³H]thymidine incorporation by PAS T cells (open symbols) is reversed by the addition of 20 units/ml IL2 (closed symbols).

MBP-triggered proliferation of PAS cells was suppressed ∼1000-fold more effectively by ShK(L5) and ShK (IC₅₀ = ∼80 pM) than mitogen-induced proliferation of rat splenic T cells (Fig. 3E, IC₅₀ ≈100 nM; p < 0.05). These results corroborate the findings with human T cells (Fig. 2). ShK(L5) inhibited MBP-induced IL2 production by PAS cells (Fig. 3F), and exogenous IL2 partially overrode ShK(L5) suppression of PAS cell proliferation (Fig. 3G). Earlier studies reported similar findings with less specific Kv1.3 inhibitors on human, rat and mini-pig T cells (Chandy et al., 1984; Koo et al., 1997; Beeton et al., 2001a). In summary, ShK(L5) is a powerful and selective inhibitor of human and rat T_EM cells and may therefore have therapeutic use in autoimmune diseases by preferentially targeting T_EM cells that contribute to the pathogenesis of these disorders (Chandy et al., 2004).

ShK(L5) Plasma Values after Subcutaneous Administration

Before embarking on in vivo studies in a rat EAE model, we used a patch-clamp bioassay to ascertain whether circulating levels of ShK(L5) after subcutaneous injection were sufficient to inhibit T_EM cells. Serum samples from ShK(L5)-treated and control rats were tested for blocking activity on Kv1.3 channels. Control serum did not exhibit detectable blocking activity, indicating an absence of endogenous channel blockers. To standardize the assay, known amounts of ShK(L5) were added to rat serum, and these samples were tested on Kv1.3 channels. The spiked serum samples blocked Kv1.3 currents in a dose-dependent fashion (K_d, 77 ± 9 pM) that was indistinguishable from the effect of ShK(L5) effect in the absence of serum (Fig. 4A). Levels of ShK(L5) in treated animals were determined by comparison with the standard curve. ShK(L5) was detectable in serum 5 min after a single subcutaneous injection of 200 μg/kg (Fig. 4B). Peak levels (12 nM) were reached in 30 min and the level then fell to a baseline of approximately 300 pM over 420 min (Fig. 4B). The disappearance of ShK(L5) from the blood could be fit by a single exponential (Fig. 4C). The circulating half-life was estimated to be ∼50 min.

Fig. 4.

Circulating half-life and stability of ShK(L5). A, known amounts of ShK(L5) were added to rat serum (Δ) or to PBS (■) and blocking activity was determined on Kv1.3 channels stably expressed in L929 cells. B, a single dose of 200 μg/kg of ShK(L5) was injected subcutaneously into four rats. Blood was drawn at the indicated times and serum was tested by patchclamp to determine the amount of ShK(L5). C, data fit to a single exponential decay. Half-life ≈ 50 min. D, five Lewis rats received single daily subcutaneous injections of 10 μg/kg/day ShK(L5) for 5 days. Blood was drawn each morning (24 h after the previous injection) and tested for blocking activity on Kv1.3 channels by patch-clamp. E, rats received a single dose of 10 μg/kg ShK(L5) either subcutaneously (open bars; n = 4) or intravenously (closed bars; n = 4). Blood was drawn at the indicated times. Serum was tested by patch-clamp to determine the amount of ShK(L5) in blood. F, a half-blocking dose of ShK(L5) was added to rat plasma (□) or to PBS containing 2% rat plasma (■) and incubated at 37°C for varying duration. Aliquots were taken at the indicated times and blocking activity determined on Kv1.3 channels.

Because the peak serum level after 200 μg/kg (12 nM) significantly exceeds the requirement for selective blockade of Kv1.3 channels and T_EM cell function, we tested lower doses. After a single injection of 10 μg/kg, the peak serum concentration of ShK(L5) reached ≈500 pM within 30 min (data not shown), a concentration sufficient to block >90% Kv1.3 but not affect Kv1.1. Repeated daily administration of this dose (10 μg/kg/day) resulted in steady-state levels of ∼300 pM (measured 24 h after injection; Fig. 4D), which is sufficient to cause 60 to 70% suppression of T_EM cells with little effect on naive/T_CM cells. The “steady-state” level is unexpected given the estimated circulating half-life of ∼50 min and indicates that ShK(L5) “accumulates” on repeated administration. To determine whether the “depot” was in the skin or elsewhere in the body, we measured blood levels of ShK(L5) 10 h after rats received single intravenous or subcutaneous injections of 10 μg/kg ShK(L5). The peptide disappeared with the same time course after administration by either route (Fig. 4E), indicating that the skin is not responsible for the steady-state level of 300 pM ShK(L5) reached after a single 10 μg/kg daily injection (Fig. 4D), and the depot(s) resides elsewhere.

Our success in achieving a steady-state level of 300 pM ShK(L5) after daily single injections of 10 μg/kg/day suggests that the peptide may be stable in vivo. To examine its stability, we incubated ShK(L5) in rat plasma or in PBS containing 2% rat plasma at 37°C for varying durations and then measured Kv1.3 blocking activity. In both sets of spiked samples (plasma and PBS) we observed a 50% reduction in Kv1.3-blocking activity in approximately 5 h, presumably due to peptide binding to the plastic surface of the tube, and the level then remained steady for the next 2-days (Fig. 4F). As an added test of stability, we compared the Kv1.3- versus Kv1.1-blocking activities of sera from ShK(L5)-treated rats. If ShK(L5) is modified in vivo, either by dephosphorylation of pTyr or cleavage of the Aeea-pTyr side chain, it would yield ShK(L4) and ShK, respectively, neither of which is selective for Kv1.3 over Kv1.1 (Fig. 1E). Serum samples from ShK(L5)-treated animals exhibited the same selectivity for Kv1.3 over Kv1.1 as ShK(L5), indicating that the peptide does not undergo the modifications stated above. Taken together, these results indicate that ShK(L5) is remarkably stable in plasma and attains pharmacologically relevant serum concentrations after single daily subcutaneous injections of 10 μg/kg.

Toxicity Studies

We conducted several in vitro and in vivo tests to determine whether ShK(L5) exhibits any toxicity (Table 2). Human and rat lymphoid cells incubated for 48 h with a concentration (100 nM) of ShK(L5) >1200 times greater than the Kv1.3 half-blocking dose or the IC₅₀ for T_EM suppression (70-80 pM) exhibited minimal cytotoxicity. The same high concentration of ShK(L5) was negative in the Ames test on tester strain TA97A, suggesting that it is not a mutagen. Both in vitro tests failed to detect any significant toxicity.

TABLE 2.

Toxicity study of ShK(L5)

In vitro tests	100 nM ShK(L5)
Cytotoxicity (% dead cells)
Human PBMCs	7.5 ± 4.3
PAS T cells	8.1 ± 0.8
Jurkat cells	5.5 ± 3.3
Burkitt lymphoma	3.1 ± 0.9
RPMI 8226 myeloma	6.5 ± 2.1
Ames test	Negative

Acute in vivo tests	Saline	ShK(L5) 10 μg/kg
Electrocardiograma
Heart rate	302 ± 13	311 ± 20
SDNN	13.3 ± 3.0	17.8 ± 4.4
CV%	6.7 ± 1.4	9.2 ± 2.2
SDANN5 min	5.0 ± 2.0	6.9 ± 2.3
rMSSD	6.8 ± 2.2	9.8 ± 3.5
HF (n.u.)	71 ± 21	79 ± 37
HF (%)	50 ± 8	53 ± 10
LF (n.u.)	68 ± 4	64 ± 10
LF (%)	50 ± 8	47 ± 10
LF/HF	1.1 ± 0.4	1.3 ± 0.7

Subchronic in vivo tests	Saline	ShK(L5) 10 μg/kg/day for 2 weeks
Weight gain (%)	7.2 ± 1.8	6.2 ± 1.7
Complete blood count
Hematocrit (%)	40.3 ± 1.4	39.0 ± 4.9
Hemoglobin (g/dl)	15.3 ± 0.5	15.0 ± 1.5
MCV (fl)	48.5 ± 0.2	48.3 ± 0.3
MCH (pg)	18.5 ± 0.8	18.5 ± 0.6
MCHC (g/dl)	38.0 ± 1.8	38.4 ± 1.3
Total white cells (×103/mm3)	7.1 ± 2.1	7.1 ± 2.5
Total red cells (×106/mm3)	8.3 ± 0.3	8.1 ± 1.0
Total platelets (×103/mm3)	656 ± 214	606 ± 106
Blood chemistry
Alkaline phosphatase (U/L)	170 ± 26	150 ± 18
Glucose (mg/dl)	139 ± 21	150 ± 18
Blood urea nitrogen (mg/dl)	17.1 ± 2.6	15.0 ± 1.7
Creatinine (mg/dl)	0.6 ± 0	0.6 ± 0.1
Albumin (g/dl)	5.0 ± 0.3	4.5 ± 0.4
Thymic cell populations (%)
CD4-CD8-	3.6 ± 1.1	4.3 ± 0.7
CD4+CD8+	77.8 ± 6.1	76.8 ± 4.1
CD4+CD8-	8.5 ± 1.7	11.2 ± 2.0
CD4-CD8+	10.0 ± 3.3	7.6 ± 1.3
CD3+	89.5 ± 1.6	93.2 ± 3.5
Splenic populations (%)
CD3+	72.4 ± 4.4	65.4 ± 0.1
CD3+CD45RC+	35.6 ± 2.6	39.8 ± 1.1
CD3+CD45RC-	23.6 ± 2.3	26.5 ± 1.3
CD3+CD4+	62.7 ± 0.1	66.6 ± 1.2
CD3+CD8+	26.9 ± 0.1	25.0 ± 0.2
IgM+	38.8 ± 1.5	33.3 ± 0.3

Data are expressed as mean ± S.D.

^aTested with t tests, P < 0.05 on all parameters.

SDNN, standard deviation of all normal-to-normal RR intervals; CV%, 100 × SDNN/average RR interval; SDANN_{5 min}, standard deviation of the mean of normal RR intervals for each 5-min period; rMSSD, root-mean-square of successive difference; HF (n.u.), high frequency (0.75-2.5 Hz) power in normalized unit; LF (n.u.), low frequency (0.2-0.75 Hz) power in normalized unit; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration.

Drug-induced blockade of Kv11.1 (HERG) channels has contributed to major cardiac toxicity and the withdrawal of several medications from the market. ShK(L5) has no effect on Kv11.1 channels at 100 nM (>1430-fold the K_d for Kv1.3), and our chosen therapeutic regimen (10 μg/kg/day, 300 pM steady-state circulating level) should therefore not cause cardiotoxicity. As a further test, we performed heart rate variability analysis in conscious rats administered vehicle (PBS + 2% rat serum) on day 1, followed by 10 μg/kg/day ShK(L5) on day 2. ShK(L5) had no effect on heart rate or the standard HRV (heart rate variability) parameters in either the time or the frequency domain (Task Force of the European Society of Cardiology and the North American Society of Pacing Electrophysiology, 1996).

Encouraged by the acute toxicity experiments, we performed a subchronic toxicity study in which rats were administered daily subcutaneous injections of 10 μg/kg ShK(L5) or vehicle for 2 weeks (n = 6 in each group). ShK(L5)-treated animals gained weight to the same degree as rats receiving vehicle (Table 2). Hematological and blood chemistry analysis showed no difference between ShK(L5)- and vehicle-treated rats, and flow cytometric analysis revealed no differences in the proportions of thymocyte or lymphocyte subsets (Table 2). Together, these studies suggest that ShK(L5) is safe.

To determine the therapeutic safety index, we administered a 60-fold higher dose (600 μg/kg/day) of ShK(L5) to healthy rats for 5 days and observed no clinical signs of toxicity, and no toxicity was seen when healthy rats received a single injection of 1000 μg/kg ShK(L5). The situation is less sanguine when the blood-brain barrier is compromised, as happens in EAE and MS. Rats with EAE that received ShK(L5) 10 μg/kg/day for 10 days showed no signs of toxicity. In contrast, 40% of rats (5 of 12) administered 600 μg/kg/day for 5 days died on the fifth day when they developed clinical signs of EAE (extrapolated LD₅₀ = 750 μg/kg/day). Because the peak concentration of ShK(L5) in the serum (12 nM) after administration of a single injection of 200 μg/kg is sufficient to block >50% of Kv1.1 channels, toxicity observed in EAE rats administered 600 μg/kg/day ShK(L5) is probably caused by the ingress into the brain of sufficient amounts of ShK(L5) to block Kv1.1. Thus, the effective therapeutic safety index of ShK(L5) is well in excess of 100 in situations in which the blood-brain barrier is not compromised (as seen in autoimmune diseases that do NOT affect the central nervous system), whereas the therapeutic safety index is 75 when the blood-brain barrier is breached.

ShK(L5) Prevents and Treats Acute Adoptive EAE and Prevents DTH in Lewis Rats

ShK(L5) was evaluated for immunosuppressive activity in vivo in two animal models. We tested its ability to prevent and treat acute EAE induced by the transfer of MBP-activated PAS T_EM cells into Lewis rats (Beeton et al., 2001a,b; Beraud et al., 1993), as well as to suppress the DTH reaction mediated by T_EM cells (Soler et al., 2003). PAS cells were activated with MBP for 48 h in vitro and then adoptively transferred (6-8 × 10⁶ viable cells) into Lewis rats. For the prevention trial, rats then received subcutaneous injections of saline (control rats) or ShK(L5) (10 μg/kg/day) for 5 days. In the first prevention trial, control rats developed mild EAE (mean maximum clinical score 2.0 ± 1.2) with an average onset of 5.6 ± 0.6 days (not shown). ShK(L5) reduced disease severity (mean maximum clinical score, 0.7 ± 0.6, p < 0.05). In the second prevention trial, control rats developed more severe EAE (mean maximum clinical score 3.2 ± 0.4) with a mean onset of 4.8 ± 0.4 days (Fig. 5A). ShK(L5) significantly reduced disease severity (mean maximum clinical score 0.6 ± 0.4, p < 0.007) but did not significantly delay disease onset (5.5 ± 0.7 days; p = 0.07). No signs of toxicity were noted in these studies.

Fig. 5.

ShK-L5 prevents DTH and acute adoptive EAE in Lewis rats. A, prevention of EAE. PAS T cells were activated in vitro, washed, and injected intraperitoneally on day 0. Clinical scoring of EAE: 0 = no clinical signs, 0.5 = distal limps tail, 1 = limp tail, 2 = mild paraparesis or ataxia, 3 = moderate paraparesis, 4 = complete hind limb paralysis, 5 = 4 + incontinence, 6 = death. Rats (n = 6/group) were injected subcutaneously with vehicle (□; n = 6) or ShK(L5) (■; n = 6; 10 μg/kg/day) from day 0 to day 5. B, treatment of EAE. PAS T cells were activated in vitro, washed, and injected intraperitoneally on day 0. Treatment with ShK(L5) at 10 μg/kg/day was started when rats developed clinical signs of EAE and was continued for 3 days. C, DTH reaction was elicited against ovalbumin and rats (n = 6/group) were treated with ShK(L5) 10 μg/kg/day for 2 days, after which ear swelling was measured. Statistical analysis was carried out using the Mann-Whitney U test.

In the treatment trial (Fig. 5B) rats were injected with MBP-activated PAS cells, administered saline or 10 μg/kg/day ShK(L5) when they initially developed signs of EAE (limp tail, hunched posture, and loss of 6% or more of their weight over 24 h), and therapy was continued for 3 days. Clinical signs of EAE peaked on day 6 in the control group (score = 3.9 ± 0.7) and on day 7 in the treated group (score = 1.9 ± 0.9; p < 0.05).

As an independent assessment of the immunosuppressive activity of ShK(L5) in vivo, we also examined its effectiveness in inhibiting the DTH reaction that is mediated predominantly by skin-homing T_EM cells (Soler et al., 2003). Lewis rats immunized with ovalbumin and adjuvant were challenged 7 days later with ovalbumin in one ear and saline in the other ear. Rats then received injections of saline (control rats) or ShK(L5) (10 μg/kg/day), and ear thickness was measured as an indication of DTH. All control rats developed ear swelling 24 and 48 h after ovalbumin challenge, whereas the DTH reaction was substantially milder in ShK(L5)-treated animals (Fig. 5C). Thus, ShK(L5) inhibits the T_EM-mediated DTH response and prevents and ameliorates severe adoptive EAE induced by myelinactivated T_EM cells.

Discussion

We have developed a highly specific Kv1.3 inhibitor by attaching the negatively charged amino acid l-pTyr to the N terminus of ShK via a 20-Å hydrophilic linker. ShK(L5) blocks Kv1.3 with a K_d of 69 pM and exhibits selectivity for Kv1.3 of 100-fold over Kv1.1, 260-fold over Kv1.6, 280-fold over Kv3.2, 680-fold over Kv1.2, and >1000-fold over all other channels tested. Other known blockers of Kv1.3 are significantly less selective than ShK(L5). Margatoxin, a peptide from Centruroides margaritatus scorpion venom, suppresses DTH in mini-pigs (Koo et al., 1997) but exhibits only 5-fold selectivity for Kv1.3 (K_d, 110 pM) over Kv1.2 (K_d, 520 pM) and 9-fold selectivity over Kv1.1 (K_d, 10 nM). Kaliotoxin from the scorpion Androctonus mauritanicus suppresses DTH in rats and ameliorates EAE (Beeton et al., 2001a) and inflammatory bone resorption in experimental periodontal disease (Valverde et al., 2004), but it is less potent (Kv1.3 K_d, 650 pM) and less selective (Grissmer et al., 1994) than ShK(L5). The first small-molecule Kv1.3 blockers with nanomolar affinity that were discovered—iminodihydroquinolines WIN-17317 and CP-339818 and the benzhydryl piperidine UK-78282—also block sodium channels (Wanner et al., 1999) and the neuronal Kv1.4 channel (Hanson et al., 1999). The small-molecule Kv1.3 inhibitors developed by Merck—correolide (Felix et al., 1999; Hanner et al., 1999; Koo et al., 1999; Bao et al., 2005), cyclohexyl-subsituted benzamides (Schmalhofer et al., 2002) and candelalides A-C (Singh et al., 2001)—are poorly selective for Kv1.3. Psora-4, the most potent small-molecule Kv1.3 blocker (K_d, 3 nM) is only 16- to 20-fold selective for Kv1.3 over Kv1.1 and Kv1.2 and 2.5-fold selective over the cardiac Kv1.5 channel (Vennekamp et al., 2004). Luteolin, a flavonoid that ameliorates EAE in rats (Hendriks et al., 2004), is sold as a nutriceutical (http://www.lutimax.com; http://www.synorx.com), ostensibly because of its ability to block Kv1.3 channels (Lahey and Rajadhyaksha, 2004). However, luteolin is a weak Kv1.3 inhibitor (K_d, 65 μM) and it is not selective for Kv1.3 over Kv1.1, Kv1.2, or Kv1.5. Other known Kv1.3 small-molecule blockers—sulfamidebenz-amidoindanes (Castle et al., 2000), dichlorophenylpyrazolopyrimidines (Atwal et al., 2001), furoquinoline Ibu-8 (Butenschon et al., 2001), tetraphenylporphyrins (Gradl et al., 2003), 3-alkyl- and 3-aryl-7H-furo[3,2-g]chromen-7-ones (Wernekenschnieder et al., 2004), and khellinone and chalcone derivatives (Baell et al., 2004), charybdotoxin, noxiustoxin, (Grissmer et al., 1994), agitoxin-2 (Garcia et al., 1994), BgK (Cotton et al., 1997), Pandinius imperator toxin1 (Peter et al., 2001), HsTx1—are neither as potent nor as selective as ShK(L5). Because of their lack of Kv1.3 selectivity, these blockers could potentially be neurotoxic if they entered the central nervous system at concentrations sufficient to block neuronal channels. The only Kv1.3 inhibitor with potency and Kv1.3-specificity comparable with ShK(L5) is the recently described synthetic analog (OSK1-Lys¹⁶Asp²⁰) of the OSK1 toxin from scorpion Orthochirus scrobiculosus (Mouhat et al., 2005), but its activity as an immunomodulator remains undetermined.

The exquisite Kv1.3-specificity of ShK(L5) makes it an attractive drug prospect. ShK(L5) is remarkably stable in plasma and it reached steady-state blood levels of ∼300 pM after repeated single daily subcutaneous injections of 10 μg/kg. This blood concentration of ShK(L5) is sufficient to block >90% of Kv1.3 channels without affecting other ion channels and to cause 60 to 70% suppression of T_EM cells while sparing naive/T_CM cells. A concentration of ShK(L5) greater than 1200 times its pharmacological dose was not cytotoxic or mutagenic in vitro. ShK(L5) did not block the cardiac Kv11.1 (HERG) K⁺ channel responsible for drug-induced long QT syndrome (Recanatini et al., 2005), and in vivo administration of ShK(L5) at pharmacological concentrations (10 μg/kg/day) did not alter cardiac function in healthy rats based on continuous EKG monitoring. Repeated in vivo administration of 10 μg/kg/day ShK(L5) for 2 weeks did not change clinical chemistry or hematological profiles. Healthy animals administered a single 1000 μg/kg/day ShK(L5) injection (100 times the pharmacological dose) or repeated injections of 600 μg/kg/day for 5 days (60 times the pharmacological dose) did not exhibit any overt signs of toxicity. These results indicate that pharmacologically relevant blood levels can be attained after single daily subcutaneous injections of ShK(L5), and the effective therapeutic safety index in healthy rats exceeds 100.

ShK(L5) may have use as a therapeutic in autoimmune diseases by preferentially targeting continuously activated autoreactive T_EM cells that have been implicated in MS, type-1 diabetes mellitus, rheumatoid arthritis, and psoriasis (Ezawa et al., 1997; Lovett-Racke et al., 1998; Scholz et al., 1998; Friedrich et al., 2000; Markovic-Plese et al., 2001; Viglietta et al., 2002; Soler et al., 2003; Wulff et al., 2003). ShK(L5) suppressed the proliferation of human and rat T_EM cells (IC₅₀ = ∼80 pM in both cases) and inhibited IL2 production at picomolar concentrations (Figs. 2 and 3). Exogenous IL2 partially overrode this block. Human naive/T_CM cells were initially 60-fold less sensitive to ShK(L5) than human T_EM cells and became completely resistant to the blocker during activation (Fig. 2), presumably by up-regulating K_Ca3.1 (Ghanshani et al., 2000; Wulff et al., 2003). Rat naive/T_CM cells were 1000-fold less sensitive to ShK(L5) than T_EM cells. This species variation in ShK(L5)-sensitivity of naive/T_CM cells compared with T_EM cells—60-fold lower in humans and 1000-fold lower in rats—can be explained by differences in K⁺ channel expression between human and rat naive/T_CM cells. Quiescent human naive/T_CM cells express more Kv1.3 channels per cell (250-400) than K_Ca3.1 channels (10-20) and are therefore more potently inhibited by Kv1.3 blockers than K_Ca3.1 blockers (Ghanshani et al., 2000; Wulff et al., 2003). Quiescent rat naive/T_CM cells, in contrast, express more K_Ca3.1 channels per cell (10-20) than Kv1.3 channels (1-10) and are more sensitive to K_Ca3.1 than Kv1.3 blockers (Beeton et al., 2001b). The rat/human difference in channel expression may underlie the differential sensitivity of rat naive/T_CM cells (IC₅₀ = 100 nM) and human naive/T_CM cells (IC₅₀ = 5 nM) to ShK(L5). In summary, ShK(L5) preferentially suppresses T_EM cells, whereas naive/T_CM cells are less sensitive to the blocker to begin with and then rapidly escape suppression by up-regulating KCa3.1 channels (Ghanshani et al., 2000; Beeton et al., 2001b; Wulff et al., 2003).

Human class-switched memory B cells (e.g., CD27⁺IgG⁺IgD^-) are implicated in the pathogenicity of autoimmune diseases (Iglesias et al., 2001; O’Connor et al., 2001; Corcione et al., 2004). Kv1.3 blockers preferentially suppress the proliferation of late memory B cells whereas naive and early memory B cells (CD27⁺IgD⁺) are significantly less sensitive (Wulff et al., 2004). ShK(L5) could therefore shut down the function of T_EM and class-switched memory B cells that contribute to the development of autoimmune disorders. Of concern is the important role of class-switched memory B cells in humoral immunity (production of IgG antibodies) and the diminished capacity to mount viable immune responses to bacterial challenges that might ensue as a result of channel-based suppression of these cells. It is fortunate that human class-switched memory B cells are less sensitive to block by ShK (IC₅₀ = 1-4 nM) than human T_EM cells (IC₅₀ = 80-400 pM), because they express higher numbers of Kv1.3 channels at rest (∼2000/cell) than T_EM cells (250-400/cell) (Wulff et al., 2004). It might therefore be possible to titrate the dose of Kv1.3 blockers to preferentially suppress one or both groups of memory cells during therapy of autoimmune disease.

We evaluated ShK(L5) in two rat models of T_EM cell-induced disease, EAE induced by adoptive transfer of myelin-specific T_EM cells (Beeton et al., 2001b) and DTH caused by skin-homing T cells (Soler et al., 2003). ShK(L5) prevented EAE if administered as a single daily injection (10 μg/kg/day) from the time of adoptive cell transfer, and it significantly reduced disease severity when therapy was initiated at the onset of symptoms. No toxicity was observed in treated EAE rats, suggesting that the blood and tissue concentrations of ShK(L5) achieved with this treatment regimen are not sufficient to block neuronal channels, including heteromultimeric K_V channels containing Kv1.3 subunits (Koch et al., 1997). ShK(L5) was also effective in suppressing DTH. These proof-of-concept studies demonstrate the therapeutic effectiveness of ShK(L5) in ameliorating T_EM-mediated diseases in rat models. We determined the therapeutic safety index of ShK(L5) in EAE rats (when the blood-brain barrier is likely to be compromised) by administering daily injections of a dose (600 μg/kg/day) 60-fold higher than the therapeutically effective dose. Forty percent of rats with EAE that received this dose died on the fifth day (extrapolated LD₅₀ = 750 μg/kg/day for 5 days), which corresponds to a therapeutic safety index of approximately 75.

In conclusion, ShK(L5) is a more selective Kv1.3 blocker than any other known inhibitor, and it might prove beneficial in autoimmune diseases by targeting both T_EM cells and class-switched memory B cells. Its picomolar affinity for Kv1.3 and remarkable plasma stability coupled with its high therapeutic safety index in healthy rats (>100) as well as in EAE rats (∼75) bodes well for its potential use as a therapeutic immunomodulator. Single daily subcutaneous injections of ShK(L5) are effective in ameliorating EAE and preventing DTH, indicating that this route of peptide delivery would be feasible for therapy.

ABBREVIATIONS

MS

multiple sclerosis

T_EM

effector memory T cell subset

T_CM

central memory T cell subset

EAE

experimental autoimmune encephalomyelitis

DTH

delayed type hypersensitivity

K_Ca3.1

intermediate-conductance Ca²⁺-activated K⁺ channel

K_v

voltage-gated K⁺ channel

ShK

Stichodactyla helianthus toxin

Fmoc

9-fluorenylmethoxycarbonyl

Pmp

p-phosphonomethyl-phenylalanine

Aeea

amino-ethyloxy-ethyloxy-acetic acid

HEK

human embryonic kidney

HERG

human ether-a-go-go-related gene

PBMC

peripheral blood mononuclear cell

IL2

interleukin 2

PBS

phosphate-buffered saline

MBP

myelin basic protein

F6CA

fluorescein-6-carboxyl

l-pTyr

l-phosphotyrosine

CP-339818

1-benzyl-4-pentylimino-1,4-dihydroquinoline

UK-78282

4-[diphenylmethoxy)methyl]-1-[3-(4-methoxyphenyl)propyl]-piperidine

WIN-17317

1-benzyl-7-chloro-4-n-propylimino-1,4-dihydroquinoline hydrochloride.