Levamisole receptor phosphorylation: effects of kinase antagonists on membrane potential responses in Ascaris suum suggest that CaM kinase and tyrosine kinase regulate sensitivity to levamisole

Since the era of modern chemotherapy began(Ehrlich and Shiga, 1904), it has been recognized that the development of resistance to therapeutic drugs can limit their efficacy. Although this was first established for trypanosome parasites with resistance to trypan red(Browning, 1907), we now know that the development of resistance occurs to antibacterial, antiviral and anticancer drugs as well. There is a Darwinian struggle between the pathogen causing the disease and the drug used to treat it. If the disease organism is to survive in the presence of the drug, it has to find a way to overcome its therapeutic effects. A pessimistic view is that development of resistance is inevitable if we continue to use drugs; and a pragmatic approach is to explore ways of counteracting or delaying the onset of resistance so that we preserve the efficacy of these drugs.

We are interested in the mechanisms by which parasitic nematodes become resistant to anthelmintic drugs. The mechanisms that underlie resistance may be more complex in helminths than in viruses and bacteria, in part because of the larger genome of these pathogens. We have focused our attention on the drug levamisole, which acts as a selective agonist on a sub-set of nicotinic acetylcholine receptors (nAChRs) in nematodes(Robertson et al., 1999; Richmond and Jorgensen, 1999). At therapeutic concentrations, levamisole produces depolarization and contraction of nematode somatic muscle, which leads to paralysis and elimination of the parasite without affecting the host nicotinic receptors. We have observed levamisole-activated single-channel currents in levamisole-sensitive (SENS) and levamisole-resistant (LEV-R) nematode isolates(Robertson et al., 1999). One of the striking differences between resistant and sensitive nAChR receptors is that, in LEV-R, the average probability of the channel being in the open state, P_open, is some ten times less than in SENS, while the average channel conductance is the same. We have also observed in SENS and LEV-R isolates that the P_open for individual channels varies dramatically between patches. For example, in the presence of 30μmol l^-1 levamisole, P_open values in SENS varied between 0.090 and 0.003 at a holding potential of -50 mV, which is a 30-fold difference between the highest and lowest P_openvalue. These observations give rise to the question of what causes the wide variation in P_open values: could it be that opening of nAChRs in nematodes is regulated?

The levamisole receptors of nematodes, like those in vertebrates, are understood to be composed of five subunits that surround a central non-selective cation pore. In Caenorhabditis elegans, the subunits that have been described include UNC-38 (an α-subunit), UNC-29 and LEV-1(β-subunits; Fleming et al.,1997). It is anticipated that other subunits are also involved in forming nAChRs on nematode muscle. A common feature of all signal-transduction proteins, including ion channels, is regulation of their activity by kinase enzymes. The addition of a bulky phosphate group to a serine, threonine or tyrosine residue in the protein introduces a highly charged group into a region that was only moderately polar before and, in the case of an ion channel, affects its opening. We have examined the amino acid sequence of theα-subunit of the Ascaris levamisole receptor, ASAR-1 (GenBank Accession No. AJ011382), and have identified consensus phosphorylation sites(Kennely and Krebs, 1991) for protein kinase C (PKC), protein kinase A (PKA), protein kinase G (PKG),Ca²⁺/calmodulin-dependent (CaM) kinase and tyrosine kinase. These sites are all on the cytoplasmic domain of the subunit between the transmembrane regions TM3 and TM4 (Fig. 1A). Some of these sites are overlapping, suggesting that differing kinases could produce the same effects. Phosphorylation of the nAChRs in nematodes, as in mammalian receptors, is expected to lead to changes in P_open (Colledge and Froehner, 1997; Hopfield et al., 1988; Reuhl et al.,1992) and may contribute to the variation in values we have observed in single-channel experiments.

In experiments on the large parasitic nematode of swine, Ascaris suum, Trim et al. (1999)described inhibitory effects of tamoxifen on the membrane potential responses to levamisole and acetylcholine and suggested the involvement of PKC in modulating the activity of nAChRs. However, a more selective PKC antagonist,chelerythine had no effect. In view of the difficulty in interpreting the effects of tamoxifen, we have examined the effects of a number of other protein kinase inhibitors on A. suum muscle, to test the hypothesis that kinase activity is required to support the activity of nAChRs and is involved in regulating the electrical response to the anthelmintic levamisole and acetylcholine. We tested the effects of both levamisole and acetylcholine,because they may select for a different subset of nAChRs in A. suum,and found that responses to these compounds are labile and inhibited by CaM kinase II antagonists and a tyrosine kinase antagonist. These observations suggest that the responses to acetylcholine and levamisole are not fixed but vary, and this plasticity depends on the phosphorylation state of the nAChRs. They also suggest that resistance to the anthelmintic levamisole may involve changes in the phosphorylation state of the receptor and may not require amino acid changes at the drug's binding site. The work has relevance to the understanding of mechanisms of anthelmintic resistance.

Adult Ascaris suum (Chitwood and Chitwood, 1950) worms were collected from the abattoir in Storm Lake City, Iowa and maintained in Locke's solution [NaCl (155 mmol l^-1), KCl (5 mmol l^-1), CaCl₂ (2 mmol l^-1), NaHCO₃ (1.5 mmol l^-1) and glucose (5 mmol l^-1)] at 35°C in the dark. Parasites were used within 5 days of collection.

A body-wall muscle-flap preparation was prepared from a 1 cm section, 3 cm caudal to the head of adult Ascaris and pinned, cuticle-side down,onto a Sylgard™-lined chamber, where the intestine was removed. The preparation was microperfused continuously with Ascaris peri-enteric fluid [APF: NaCl (23 mmol l^-1), Na acetate (110 mmol l^-1), KCl (24 mmol l^-1), CaCl₂ (6 mmol l^-1), MgCl₂ (5 mmol l^-1), glucose (11 mmol l^-1) and Hepes (5 mmol l^-1), pH 7.6 with NaOH]. In experiments to test the effect of low-Ca²⁺, APF was prepared by replacing Ca²⁺ in APF on an equimolar basis with Mg²⁺. Application of the perfusate was via a fine microtube placed with a micromanipulator (approximately 500 μm) over the muscle cell bag. The rate of perfusion was 1.5 ml min^-1 and this allowed rapid change of the solution bathing the cell being recorded from. The temperature in the chamber was maintained at 32-33°C.

A two-microelectrode current-clamp technique was used for measuring the membrane potential and input conductance changes of the Ascarismuscle cell bags. Micropipettes made from borosilicate glass (Clarke Electromedical, Reading, UK) with resistances in the range of 20-40 MΩwhen filled with 2 mol l^-1 potassium acetate were used for recording. Two microelectrodes were carefully inserted into one muscle cell bag with minimum damage. An Axoclamp 2B amplifier, 1320A Digidata Interface,pClamp 8.0 software (all from Axon Instruments, Union City, CA, USA) and Pentium III PC were used to display, record and analyze the membrane potential and injected current. One micropipette was used for recording of membrane potential, while the second was used for injection of current pulses(hyperpolarizing: 40 nA, 500 ms filtered at 0.3 kHz).

In the present study, we measured the first fast peak membrane potential change response to acetylcholine or levamisole as the nicotinic response to these drugs. In approximately one-third of the responses, we also observed a second component or tail to the response (arrows, Fig. 1B). This component has not been described previously, perhaps because it is more labile than the first component. It may have arisen as a secondary effect of entry of Ca²⁺ following stimulation of the nicotinic receptors, or following stimulation of metabotrophic (muscarinic) receptors(Hwang et al., 1999). In the present study, we focused exclusively on the effects of the kinase antagonists on the peak of the first component because it results from activation of extrasynaptic nicotinic receptors on the bag region of the muscle(Martin, 1982), and these receptors are accessible to direct stimulation by perfused drugs and to recording by standard two-electrode current-clamp technique.

H-7, staurosporine, genistein and KN-93 were purchased from Calbiochem (San Diego, CA, USA). Levamisole hydrochloride and acetylcholine chloride were purchased from Sigma (St Louis, MO, USA). Stock solutions for staurosporine,genistein and KN-93 were prepared in dimethyl sulfoxide (DMSO) and frozen for use following dilution in APF. The maximum concentration of DMSO used (0.1%)did not affect membrane potential or responses to levamisole and acetylcholine. Stock solutions of H-7 and levamisole were prepared each week in APF and stored at 4°C. Stock and working solutions of acetylcholine were prepared freshly every day in APF.

In all experiments, acetylcholine was added to the bag region of the cell via the microcatheter in the perfusate for 20 s; levamisole was applied for 10 s. Peak membrane potential changes and input conductance were measured. All kinase antagonists were added to the preparation in the perfusate for at least 2 min before the application of acetylcholine or levamisole.

All results are presented as means ± S.E.M. Significance was tested using paired t-test (for antagonist influences) or one-way analysis of variance (ANOVA; for influence of Ca²⁺ on the levamisole concentration—response relationship). The influences of Ca²⁺on the effects of increasing levamisole concentrations on membrane potential were estimated using the Hill equation: % response =1/(1+[EC₅₀/X_A]ⁿ), where EC₅₀ is the concentration of agonist (X_A)producing 50% of the maximum response, and n is the Hill coefficient(slope). Prism Pad software (Version 3.0, San Diego, CA, USA) was used for all statistical determinations.

We made recordings from muscle cell bags with stable membrane potentials more negative than -20 mV and input conductances of <4.0 μS. Recordings were rejected for analysis if there was a non-drug-induced change in resting input conductance of >0.2 μS or a resting membrane potential change of>4 mV during the recording.

Fig. 2A shows a representative example of the effect of a 10 s superfusion of the bag region of the muscle with 10 μmol l^-1 levamisole. The first response recorded was a 7 mV depolarization peak associated with a change in input conductance from 2.26 μS to 2.86 μS. The membrane potential recovery after the levamisole pulse was slow, with a t_1/2 of 94 s. Fig. 2B shows a representative trace from another preparation of the effect of 20 s of 10 μmol l^-1 acetylcholine at a resting membrane potential of -23.8 mV. The peak depolarization of the first response was 3.4 mV associated with a conductance change of 0.31 μS. Note, however, that the recovery from acetylcholine was faster (t_1/2=13 s) than with levamisole. The slower recovery time of levamisole, when compared with acetylcholine, was a consistent feature, making the measurement of the effect of kinase antagonists on levamisole responses more difficult to determine because of the longer stable recording conditions that were required. Several factors may have contributed to the slow recovery with levamisole, including the more lipophilic nature of levamisole (Robertson and Martin, 1993), its resistance to degradation by cholinesterase(Eyre, 1970) or its ability to antagonize phosphatases (Schnieden,1981). However, we did not investigate this phenomenon further.

H-7 is a broad-spectrum serine—threonine kinase inhibitor with activities against the AGC kinase group of PKA, PKC and PKG, and inhibitor constants (K_is) in the range of 0.1-6.0 μmol l^-1 (Kawamoto and Hidaka,1984; Boulis and Davis,1990; Hoffman and Newlands,1991; Li et al.,1992; Reuhl et al.,1992; Hu and Li,1997; Ku et al.,1997; Calbiochem 1998 data sheet CN: 371955). We superfused H-7 at a concentration of 30 μmol l^-1 over the bag region of the muscle cell to determine the effect of H-7 on responses to 10 μmol l^-1levamisole and 10 μmol l^-1 acetylcholine(Figs 2A,B). The mean control peak levamisole response was 7.6±2.0 mV. In the presence of 30 μmol l^-1 H-7, the mean response was 6.0±1.0 mV (N=5) or 79% of the control. The mean control peak acetylcholine response was 4.1±0.3 mV and, in the presence of 30 μmoll^-1 H-7, it was 3.5±0.6 mV (N=5) or 85% of the control. H-7 did not induce a statistically significant effect on the response to either compound(Table 1). These observations suggest that PKA, PKC and PKG do not have a modulatory effect on nematode nAChRs.

Staurosporine is a broad-spectrum kinase inhibitor with activities against PKA, PKG, PKC, myosin light chain kinase (MLCK), CaM kinase and tyrosine kinase and K_is in the range of 0.7-70 nmoll^-1(Bergstrand et al., 1992; Coultrap et al., 1999; Li et al., 1992; Nishimura and Simpson, 1994;Calbiochem 1999 data sheet CN: 569397). It combines with the ATP-binding site of the kinase enzyme to inhibit its function. We perfused 1μmoll^-1 staurosporine over the muscle bag to determine its effect on responses to levamisole and acetylcholine. Fig. 3A shows the inhibitory effect of staurosporine on representative recordings of levamisole responses. Table 1 shows that the mean control peak levamisole depolarization was 6.8±1.9 mV (N=6),and, in the presence of 1 μmoll^-1 staurosporine, the mean peak levamisole response was 3.9±1.1 mV (N=6) or 57% of the control. The difference was statistically significant(P<0.0001).

Fig. 3B shows the effect of staurosporine on representative recordings of acetylcholine responses. Table 1 shows that, as with levamisole, there was a depressant effect on the acetylcholine responses. The mean control acetylcholine response was 5.4±0.6 mV (N=7) and this was reduced in the presence of 1 μmoll^-1 staurosporine to 2.8±0.3 mV (N=7) or to 51% of the control(P=0.0016).

As we observed a significant effect with the broad-spectrum kinase antagonist staurosporine, but not with H-7, we can suggest that the difference is due to the effect of staurosporine on a kinase that is not affected by H-7. One of the protein kinase enzymes not affected by H-7, but affected by staurosporine, is CaM kinase II. A selective CaM kinase II antagonist was therefore used to investigate the role of this enzyme.

To follow up the above studies, we tested KN-93, a selective CaM kinase II inhibitor with a K_i of 0.4 μmoll^-1(Mamiya et al., 1993; Ishida et al., 1995; Anderson et al., 1998; Ledoux et al., 1999; Nakayama et al., 2001;Calbiochem 2001 data sheet CN: 422708) on levamisole and acetylcholine responses (Fig. 4). 10μmoll^-1 KN-93 reduced the control mean peak levamisole response from 6.2±1.0 mV (N=8) to 2.7±0.3 mV (N=8) or 44% of the control (Table 1; P=0.035). An inhibitory effect on the acetylcholine responses was also observed; the control response of 4.7±0.6 mV (N=9) was reduced to 2.0±0.4 mV (N=9) by 10 μmoll^-1 KN-93(Table 1, P=0.0004). The effect of KN-93 on levamisole and acetylcholine responses suggests that CaM kinase II is involved in phosphorylating nAChRs on the muscle bag and promoting the opening of the ion-channel receptor.

We tried routinely to wash off the effects of the kinase inhibitors, but reversal of the effects of the antagonists was not always apparent. This was probably due to the lipophilic nature of the antagonists and the large size of the bag muscle region. However, we did occasionally observe reversal with staurosporine and KN-93 when we used acetylcholine as an agonist. Fig. 3B illustrates the reversal on washing with acetylcholine as the agonist and staurosporine as the kinase antagonist. Fig. 4Billustrates reversal of the effect of KN-93 on washing: here, 20 min was required to observe definitive recovery towards the control response values.

An effect of CaM kinase II in modulating levamisole receptors implies that the response will also be sensitive to Ca²⁺. We tested the effect of levamisole at different concentrations in preparations bathed for longer than 30 min in low-Ca²⁺ APF and in normal Ca²⁺ APF. Elimination of Ca²⁺ from the bathing solution is known to reduce cytoplasmic Ca²⁺ concentrations because low-Ca²⁺extracellular solutions reduce the opening of Ca-activated Cl-channels(Thorn and Martin, 1987). In normal Ca²⁺ APF, the resting membrane potential was-22.9±3.5 mV and input conductance was 2.63±0.38 μS(N=4). These values are similar to those recorded in low-Ca²⁺ APF, where membrane potential was -21.7±0.8 mV and input conductance was 2.16±0.12 μS (N=4, P=0.11, t-test).

Fig. 5 shows the effect of different concentrations of levamisole on a preparation bathed in normal Ca²⁺ APF Fig. 5A)and one bathed in low-Ca²⁺ APF solution(Fig. 5B). Notice that the responses to levamisole at lower concentrations are bigger in the presence of Ca²⁺. Fig. 5C shows a plot of the mean (N=4) peak depolarizations in the presence and absence of Ca²⁺. The best fit to the observed responses in the presence of Ca²⁺ had an EC₅₀ of 1.2μmoll^-1, with a Hill slope of 1.21 and a maximum response of 7.1 mV. In low-Ca²⁺ APF, the best-fit EC₅₀ was 3.1μmoll^-1, the Hill slope was 3.9 and the maximum response was 6.8 mV. Note again that the presence of Ca²⁺ is associated with an increase in the size of the response (P=0.045, ANOVA).

We pointed out earlier that staurosporine, but not H-7, has significant effects on the amplitude of the response to levamisole and acetylcholine. As staurosporine at higher concentrations may also antagonize the effects of tyrosine kinases, and there are consensus tyrosine kinase sites present on the ASAR-1 α-subunit of A. suum(Fig. 1), we tested the effects of genistein on the levamisole and acetylcholine response. Genistein competes with ATP non-competitively, as a substrate for tyrosine kinases, to inhibit phosphorylation, with K_is of ≤1 μmoll^-1(Huang et al., 1992; Young et al., 1993; Wang et al., 1997; Saxena and Henderson, 1997; Neye and Verspohl, 1998;Calbiochem 2001 data sheet, CN: 345834). We tested the effect of 90μmoll^-1 genistein on the levamisole and acetylcholine responses(Fig. 6).

Genistein reduced the amplitude of the levamisole response(Fig. 6). Table 1 shows that the mean control peak levamisole response was 6.4±0.7 mV (N=8), and the mean response in the presence of 90 μmoll^-1 genistein was 3.3±0.3 mV (N=8) or 51% of the control (P=0.001).

Fig. 6B shows a representative trace of the effect of 90 μmol l^-1 genistein on the response to acetylcholine. The effect was small(Table 1), with the mean control acetylcholine peak response reaching 5.6±0.8 mV (N=6)and the acetylcholine response in the presence of 90 μmol l^-1genistein reaching 5.0±0.6 mV (N=6) or 89% of the control. Although the change in response was small, it did reach statistical significance (P=0.03). The effect of genistein on the levamisole and acetylcholine responses suggests that tyrosine kinase is involved in maintaining the nematode nicotinic receptor.

Caenorhabditis elegans is a soil nematode that has been subjected to detailed genetic analysis. Genes encoding nAChR subunits in C. elegans have been described (Fleming et al., 1997), including unc-38, unc-63, unc-29 and lev-1. Both unc-38 and unc-63 encode α-like subunits, while unc-29 and lev-1 encode non-α-subunits. Homologues in Ascaris suum (ASAR-1, GenBank Accession No. AJ011382), Trichostrongylus colubriformis(Wiley et al., 1996) and Haemonchus contortus (Hoekstra et al., 1997) have 78%, 84% and 83% identity, respectively, to UNC-38. All of these nAChR subunits possess consensus sites for phosphorylation by PKA, PKC, PKG, CaM kinase II or tyrosine kinase. Phosphorylation of the levamisole receptor is expected to alter the activity of receptors and may be one means by which sensitivity to levamisole is affected in resistant isolates.

Development of resistance to levamisole therapy in parasitic nematodes has not been linked to changes in amino acid sequences of nAChR subunits(Wiley et al., 1997; Hoekstra et al., 1997). In T. colubriformis, levamisole resistance has a recessive, sex-linked single gene component and a small polygenic component on autosomes. In H. contortus, resistance is polygenic and associated with a 4-25-fold increase in the concentration of levamisole required to produce paralysis(Sangster et al., 1991, 1998).

In the present study, we examined the effects of selected kinase antagonists to test the hypothesis that phosphorylation of nAChRs is involved in regulating the response to levamisole. We observed no significant effects on levamisole and acetylcholine responses with H-7, an antagonist of PKA, PKC and PKG (Kawamoto and Hidaka,1984; Boulis and Davis,1990; Hoffman and Newlands,1991; Li et al.,1992; Reuhl et al.,1992; Hu and Li,1997; Ku et al.,1997). The broad-spectrum antagonist staurosporine, which inhibits PKA, PKC, PKG, CaM kinase and tyrosine kinase(Bergstrand et al., 1992; Coultrap et al., 1999; Li et al., 1992; Nishimura and Simpson, 1994),reduced the response to both acetylcholine and levamisole in A. suummuscle. Reductions in response were also recorded in the presence of KN-93, a selective CaM kinase II antagonist (Mamiya et al., 1993; Ishida et al.,1995; Anderson et al.,1998; Ledoux et al.,1999; Nakayama et al.,2001), and genistein, a selective tyrosine kinase antagonist(Huang et al., 1992; Young et al., 1993; Wang et al., 1997; Saxena and Henderson, 1997; Neye and Verspohl, 1998). These effects suggest the involvement of CaM kinase II and tyrosine kinase in supporting the opening of the ion channels. The lack of effect of H-7, a protein kinase antagonist that is not active on CaM kinase II or tyrosine kinase, and the antagonistic effects of staurosporine, a broad-spectrum protein kinase antagonist, support this interpretation. The potentiating effects of Ca²⁺ on the levamisole responses may be explained by the involvement of CaM kinase II. Observations with H-7 and staurosporine, which depress muscle contraction to acetylcholine in the filarid Acanthocheilonema viteae (Minardi et al., 1995), are consistent with the view that CaM kinase II regulates the response to acetylcholine in parasitic nematodes.

Trim et al. (1999) have observed inhibitory effects of tamoxifen on the depolarizing responses to levamisole and acetylcholine and suggested that a PKC-dependent phosphorylation of A. suum nAChRs is necessary to maintain the sensitivity of the receptor to acetylcholine. In this model, blockade of PKC results in inhibition and reduction of the response. However, Trim et al.(1999) also tested the selective PKC antagonist chelerythrine and did not observe an effect on the membrane-potential responses to acetylcholine. In our experiments, we did not observe a significant effect with H-7, another kinase antagonist that has inhibitory effects on PKC. As tamoxifen also has an inhibitory effect on calmodulin (Edwards, 2002), we can explain the inhibitory effects of tamoxifen by inhibition of calmodulin and depression of the activity of CaM kinase II. A role for CaM kinase II in regulating nematode body wall muscle activity has been demonstrated in C. elegans, where a single gene, unc-43, codes for CaM kinase II(Reiner et al., 1999). The unc-43 gene was also shown to interact with unc-103, a gene encoding a voltage-gated K⁺ channel, so a clear connection between phosphorylation of ion channels, muscle excitability and CaM kinase II in nematodes has been demonstrated. CaM kinase II may mediate cell excitability through modulation of channel activity or modeling of the synapse, as demonstrated in the vertebrate central nervous system(Rongo and Kaplan, 1999). We do not rule out the possibility of PKC involvement, but our evidence with H-7 suggests that, over the time-span of the experiment (approximately 3 h), it is less important than CaM kinase II. Over a longer time span (12-24 h), where receptor synthesis can play a role, evidence from C. elegans does indicate a role for PKC (tpa-1) in long-term adaptation of nAChRs(Waggoner et al., 2000). In mammalian muscle, PKC and CaM kinase II can regulate nAChR expression, but it is CaM kinase II rather than PKC that is required for Ca²⁺ or activity-dependent control of nAChR gene expression(Macpherson et al., 2002).

In summary, our studies indicate that CaM kinase II and tyrosine kinases are involved in supporting the opening of acetylcholine/levamisole receptors on A. suum somatic muscle. In previous publications, we have observed differences in opening of nAChRs in SENS and LEV-R isolates of the parasitic nematode Oesophagostomum dentatum and hypothesized that phosphorylation may contribute to the drug-resistant phenotype by modulating the opening of levamisole receptors. In the present study, we have used selective kinase antagonists to test the importance of the kinases and phosphorylation process in modulating the receptors. These observations demonstrate that the responses to acetylcholine and the anthelmintic levamisole are not fixed but vary, and this plasticity depends on the phosphorylation state of the nAChRs. The observations emphasize the possibility that resistance to the anthelmintic levamisole does not necessarily involve amino acid changes at the levamisole-binding site but may involve regulation of the activity to the levamisole receptor viachanges in its phosphorylation state.

The mechanisms used by nematodes to modulate their responses to acetylcholine and nicotinic anthelmintics is worthy of further study; it seems increasingly likely that modulation of the receptors may contribute to at least some of the observed drug resistance. Further study necessitates the use of Ascaris because it is the only readily available nematode parasite suitable for electrophysiological study. To date, there are no anthelmintic-resistant isolates of Ascaris, but information derived from Ascaris will inform work on parasitic nematodes where anthelmintic isolates are available.

We are pleased to acknowledge the support of the NIH to R.J.M.: RO1 A147194-02.