Changing the firing threshold for normal optic nerve axons by the application of infra-red laser light

Application of IR laser light at the recording site increases the DAP

The depolarizing afterpotential (DAP) in myelinated axons is generated by internodal charging from local active nodes⁷ and the amplitude and duration of the DAP depends on the internodal membrane conductance. This conductance is increased by depolarization, as internodal voltage-gated K⁺ channels activate⁸, and this makes the amplitude and the rate of decay of the DAP an indirect, although sensitive, index of membrane potential. Furthermore, our experiments with 4-aminopyridine (4-AP), see below, have confirmed that the DAP amplitude is controlled by 4-AP sensitive K⁺ channels, likely to be of the K_V1 sub family see⁹, K_V1.2. The amplitude of the DAP following a supramaximal compound action potential increased when laser light was applied with a 400 mA driving current (equivalent to 600 mJ per minute) (Fig. 1). Comparing the effects of applying laser light and the effects of altering whole-bath perfusion temperature, indicates that the laser diode when operating continuously, brings about local changes in temperature of a few single °C. As is also apparent in Fig. 1, the application of laser light differs subtly from raised perfusion temperature, in that it reliably caused a small drop in the amplitude of the compound action potential, that we understand is as a result of a rapid effect of laser light (discussed more fully below).

Application of IR laser light at the site of nerve electrical stimulation makes optic nerve axons less excitable

Switching on the IR laser irradiation at the stimulation site caused an immediate effect on the height of the recorded compound action potential. It was always reduced. Where the application of light was prolonged over minutes, the reduction in amplitude gradually lessened over this period, although it stayed reduced. Where a less than supramaximal action potential was evoked in the nerve, and the technique of threshold-tracking was employed to keep the response constant eg³, an increase in threshold was observed usually lasting throughout the duration of light application (Fig. 2A–C). Where the laser diode was powered at 400 mA, and with an ambient bath temperature of 32 °C the average increase in threshold measured at the end of 3 min applications was 5.26 ± 1.32% (mean ± sem, P = 0.017, one-sample t-test, n = 5), (Fig. 3A), consistent with an average local temperature change affecting the nerve close to 3 °C. This finding was anticipated from previously reported findings for changing bath ambient temperature³, and also supports the idea of a temperature dependent membrane potential. The small size of the temperature change revealed is the reason that the manoeuvre can be repeated without any noticeable acute damage to the axons.

The recordings in Fig. 2A and B provide clear evidence for an increase in superexcitability during the IR light application. The superexcitability trace shown in Fig. 2B was obtained by comparing the threshold measured in response to a control stimulus, with one preceded by a conditioning stimulus, 10 ms before. Superexcitability is a normal integral part of the recovery cycle following an action potential, and can be defined as a period during which axons become easier to excite following an action potential eg¹⁰. It can be contrasted with the refractory period that occurs closer to the response to the conditioning stimulus, during which threshold is increased. The recordings in Fig. 2E (and Supplementary Figure S1) show a fall in refractoriness during the application. The recorded increase in superexcitability at 10 ms, and fall in refractoriness at 4 ms, following a conditioning stimulus, occur gradually over the minute following the onset of light application in Fig. 2B, E, and their new level outlasts the period of less-than-adequate threshold-tracking occurring immediately after illumination (indicated by the fall in the peak response). This increase in superexcitability is to be expected from the effect of laser light on the DAP, and suggests that the nerve fibres hyperpolarize over this time. The fall in refractoriness implies Na⁺ channels are able to escape inactivation (reprime) more quickly during irradiation, and again suggests a hyperpolarization may be occurring. It therefore seems likely that a part of the reduction in excitability is caused by a warming induced membrane potential hyperpolarization, that we have already explored^1,2,3. The most direct measurements of membrane potential changes with IR light application revealed that the warming effect is complete with a time course of a minute to minutes (eg Fig. 3B).

It is to be expected that near-IR light will be absorbed by water and lipid, both within and surrounding the isolated nerve, and because the nerve is constantly perfused in our nerve bath, heat energy will be continuously lost. We propose that the rate of axonal warming will depend critically on the properties of the interface between the fibre optic and the nerve, and this explains why the effect of exactly similar manoeuvres in different experiments can give variable results. We can also say that unless the end of the light guide is in physical contact with the nerve, no functional changes in the nerve are found, even if the light guide is very near the nerve when viewed under a stereo microscope, and so this contact is essential.

IR laser light affects optic nerve axons rapidly

A consistent finding (during our recordings from F-fibres), is a rapid increase in threshold occurring without apparent delay on light application (Fig. 2C), and in other experiments a fall in the amplitude of a supra-maximal compound action potential. When recording supra-maximal compound action potentials, the whole of the action potential was affected, including the M and S components (data not shown), suggesting that this effect is not specific for any one nerve fibre category. Whether this fast increase in threshold can be seen when threshold-tracking depends critically on how quickly the tracking feedback-loop can compensate for changes in response amplitude. In the F-fibres, this rapid effect on threshold may not be sustained and may decay with continuous irradiation, whereas what we believe to be a hyperpolarization mediated increase in threshold probably remains, accounting for the overall trajectory of threshold in Fig. 2E (bold trace). This fast increase in threshold can also be seen as a brief upward shift in the threshold when the axons are in the relative refractory period measured at 4 ms after a conditioning stimulus, before changes in membrane potential reverse the direction of travel for the threshold-measurement (see Supplementary Figure S1). Also, the maximal change in threshold caused by the rapid effect is too large to be accommodated by the threshold-temperature relation we have already defined^2,3. How might this rapid effect be explained? We can put forward two possibilities. The IR light may have direct effects on nodal ion channels, for example activating delayed-rectifier K⁺ channels¹¹, or increasing Na⁺ channel inactivation. Na⁺ channel inactivation substantially impacts excitability^3,12. A second possibility is that there is a rapid change in the properties of nodal membrane and/or trans-myelin (or Barrett and Barrett⁷) resistance acting to increase the effective nodal capacitance, and hence reduce excitability. However, our evidence suggests that an ion channel mechanism may not be tenable, because blocking K⁺ channels (Figs. 2D and S1) does not prevent the rapid change in excitability, and refractoriness is reduced when the laser light is applied, potentially also ruling-out increased Na⁺ channel inactivation as the mechanism (Fig. 2E, F). This is because the primary mechanism involved in generating refractoriness, once nodal K⁺ channels are blocked, is the delayed escape of Na⁺ channels from inactivation (also known as repriming), and this mechanism is undoubtedly accelerated by the laser.

Recording extracellular membrane potential indicates the change in membrane potential occurs over a minute

While it is not possible to record membrane potential in individual axons in the optic nerve, it has proved possible to record a fraction of the membrane potential of all the axons extracellularly eg¹, across the grease barrier in the nerve bath. There are advantages associated with this approach. First, it is not necessary to damage the axons at the site of recording, and secondly the technique will pick up changes in potential even when some of the axons cannot be electrically stimulated further down the nerve, and so cannot contribute to a threshold measurement. However, the technique cannot provide a definitive value of membrane potential, and the size of the membrane potential changes observed depends on the quality of the seal at the barrier, that can change even within a single experiment. Nevertheless, where the seal resistance is high, mV changes in DC potential can be recorded across the barrier, that appear to be around half the change in potential across the membrane of an individual axon. Finally, it is possible to compare the effects of applying laser light and fluctuations in ambient bath-chamber temperature at closely similar times and in the same recordings (Fig. 3B).

Superfusion of ZD7288 increases the change in threshold, consistent with IR laser light hyperpolarizing axons

When the IR light is applied to the site of stimulation, threshold increases, and this is usually sustained over a 3 min application period (Fig. 3A). In order to help prove that IR light hyperpolarizes the axons using a pharmacological approach, we recorded the threshold-increases before and after superfusion of 50 µM ZD7288, a drug that blocks I_h¹³. I_h is generated by HCN channels that are activated by hyperpolarization^8,14. I_h should be able to supress a temperature rise-dependent hyperpolarization, so blocking the channels would be predicted to increase the hyperpolarization, and therefore increase the change in threshold brought about by laser light. We have found that ZD7288 increased the threshold-rise by 85.11 ± 28.19%, (P = 0.034, paired t-test, n = 6), although the responses were variable, Fig. 4. We present a mathematical treatment of this effect and explain how this is manifest. Further, the addition of ZD7288 to the perfusion solution gives rise to an increase in the resting threshold, an effect already reported in Austerschmidt et al.³, and we think this means the axon resting membrane potential hyperpolarizes with the loss of I_h. This is expected to increase the resting potential by − 5 mV, or perhaps more, explained in the section immediately below.

Explaining the effect of blocking I
_h

Equation (1) allows the estimation of the change in resting membrane potential when the temperature increases, by the expedient of allowing electroneutral Na⁺ entry to have a temperature dependent rate³. The equation incorporates resting Na⁺ and K⁺ conductances, in the simplest case neither of which exhibit voltage-dependence, a Na⁺-pump current returning both the electroneutral influx of Na⁺. This approach assumes that the intra and extra-axonal concentrations of Na⁺ and K⁺ are unchanging throughout. It is also assumed that the total resting membrane conductance is the sum of the Na⁺ and K⁺ conductances, independent of the contribution of individual channel subtypes such as voltage-gated Na⁺ channels and I_h. Modifying the g_Na and g_K variables to allow an increase in membrane conductance on hyperpolarization, according to the voltage-dependence of I_h detailed in Mayer and Westbrook¹⁴, shows how the intrinsic temperature dependence of the resting membrane potential is predicted to be severely limited by the presence of I_h (Fig. 5, below).

At the resting potential there is no net current across the membrane. Therefore with a resting Na⁺ and K⁺ conductance and a parallel expulsion of Na⁺ entering electroneutrally:

where g_K is the membrane conductance for K⁺ ions, E_K the K⁺ ion equilibrium potential, g_Na is the membrane conductance for Na⁺ ions, E_Na is the Na⁺ ion equilibrium potential. I_EN is the pump current resulting from the electroneutral movement of Na⁺ across the membrane, and E_m is the resting membrane potential.

Allowing for the Na⁺ pump to return a current one third of any inward Na⁺ ionic current

Given the above, the following rearrangement applies:

Two eventualities are apparent. The first is that if the electroneutral transmembrane Na⁺ flux and Na⁺ conductance are reduced to zero, the membrane potential (E_m) becomes equal to E_K, and the value of E_K is shown to be the most negative value possible for E_m. The second is that with the introduction of electroneutral Na⁺ flux, it becomes possible for E_m to be more negative than E_K.

Austerschmidt et al.³ used a membrane resistance of 0.55 GΩ for a length of optic nerve axon, the same value previously chosen by Harris and Attwell¹⁵. This corresponds to a transmembrane conductance of 1.82 nS, and where this conductance comprises the sum of g_Na = 0.29 nS and g_K = 1.53 nS, the resting membrane potential, without the effects of electroneutral Na⁺ entry, is − 82 mV, according to Eq. (3), where E_K = − 100 mV and E_Na is + 60 mV. With these same variable values, the effect of the Na⁺ pump on E_m, in the absence of any electroneutral entry, is − 7.5 mV, and this component is not dependent on temperature, unless the value of g_Na alters. However, adding an electroneutral flux of Na⁺ of 2.0 × 10⁻¹⁶ mol/s increases E_m to − 85.8 mV, and where this is assumed to be occurring at 20 °C, an increase in temperature from 22 to 37 °C with a Q₁₀ of 3 gives a hyperpolarization of − 19.8 mV.

Our experiments predict I_h must act to reduce the magnitude of hyperpolarization with warming (Fig. 5).

Adding a voltage-dependent I_h, activating according to the relation shown in B, necessitated solving Eq. 3 numerically using a recursive program in MatLab. The maximally activated g_h was taken to increase the nerve membrane conductance by 100% or by 50% (orange and blue curves, respectively). I_h was taken to be the sum of ionic currents provided by a parallel voltage-dependent Na⁺ conductance and K⁺ conductance, with an overall reversal potential of − 36 mV¹⁴, and with a steady-state activation curve of this form:

where g_h is the conductance at any membrane potential expressed as a fraction of the maximal conductance, E_m is the membrane potential, V₀ is the midpoint (− 110 mV) and k is the slope of the relation (taken as − 12 mV), cf^3,16. A summary of the arrangement of ion channels and the Na⁺-pump in myelinated axons is available in the literature for peripheral nerve axons (¹⁷ ‘the Bostock model’), that must differ from the F-fibres in optic nerve only by degree (explored in³, and where the functionality of K_V7 in optic nerve axons appears minimal). The major ionic leakage, electroneutral transport, pump, and I_h that concerns the present analysis are suggested to be found in internodal membrane, where voltage-dependent gating affects only I_h, and is based only on steady-state conditions.

The presence of I_h depolarizes the nerve membrane, and more I_h is associated with a reduced slope of the membrane potential temperature-dependence. Exploring a scenario with variables similar to those thought to be applying in our experiments, where local temperature is raised 3 °C by applying laser light from an ambient bath temperature of 32 °C (and where a g_hmax is incorporated that increases the resting membrane conductance by 100% when fully activated), blocking I_h will increase the laser-induced change in membrane potential by 63.2%. This explains why the exposure to ZD7288 makes the threshold increases bigger. While this result is within a single standard error of the proportionally larger threshold increases, these increases, before and after exposure to ZD7288, may also be affected by differing degrees of resting Na⁺ channel inactivation with the pre-exposure effect being moderated by a greater removal of Na⁺ channel inactivation throughout the 3 min long exposure to the light.

Ambient bath temperature affects the threshold increase caused by IR laser light

Without altering either the position of the laser light guide, or changing the current driving the diode, our data suggest that threshold changes brought about by applying laser light depend on the bath temperature. This would make sense if the bath temperature had a role to play in setting the resting potential of the axons, which we have previously reported, and if the laser brought about changes in threshold by warming the nerve. Applying laser light when the nerve is at 26 °C and at 33 °C gives rise to different responses, as shown in Fig. 6, and these differences can be explained easily as the relationship between threshold and temperature is ‘U’ shaped². In the case where the bath temperature is near 26 °C, laser light induced warming of a few single degrees centigrade takes the nerve across the bottom of the ‘U’, indicated as 2 on the inset cartoon, whereas at a warmer bath temperature, the threshold increases more because the nerve is made to ascend the right hand side of the ‘U’, indicated as 1. One possibility is that the small threshold-changes seen at 26 °C equate to the fast changes only. The nerve was exposed to 5 µM bumetanide throughout this recording, revealing that although a part of the electroneutral Na⁺ flux was blocked, it is not sufficiently reduced to prevent temperature dependent changes in threshold, nor all the effects of IR laser light. Our previous findings indicate that bumetanide will reduce electroneutral Na⁺ entry, for example slowing the decline in the action potential amplitude in the presence of the Na⁺-pump blocker ouabain², and blocking the transporter NKCC1 will cause a small (probably single mV) reduction in resting membrane potential as a result³. However, most electroneutral Na⁺ movement appears to be not sensitive to bumetanide.