Biochimica et Biophysica Acta 1832 (2013) 636–649

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbad is

Specific functioning of Cav3.2 T-type calcium and TRPV1 channels underdifferent types of STZ-diabetic neuropathy☆

Eugen V. Khomula a,b, Viacheslav Y. Viatchenko-Karpinski b, Anya L. Borisyuk b, Dmytro E. Duzhyy b,Pavel V. Belan a,b, Nana V. Voitenko a,b,⁎a International Center of Molecular Physiology of Natl. Acad. of Sci. of Ukraine, 4 Bogomoletz str., Kyiv 01024, Ukraineb State Key Laboratory of Molecular and Cellular Biology, Bogomoletz Inst. of Physiol. of Natl. Acad. of Sci. of Ukraine, 4 Bogomoletz str., Kyiv 01024, Ukraine

Abbreviations: [Ca2+]i, cytosolic Ca2+ concentration;holding membrane potential; IB4, isolectin B4; LVA/HVNTCN, nonpeptidergic thermal C-type nociceptive (neuroPDN, peripheral diabetic neuropathy; PWL, paw withdrawactivation; STZ, streptozotocin; T-PCD, T-type PCD; TPS, th☆ Disclosure summary: the authors have nothing to declof interest in regard to this study.⁎ Corresponding author at: Bogomoletz Inst. of Physiol. o

Bogomoletz str., Kyiv 01024, Ukraine. Tel.: +380 442562E-mail addresses: eugen_kh@biph.kiev.ua (E.V. Khom

(V.Y. Viatchenko-Karpinski), borisyuk@biph.kiev.ua (A.Ldduzhyy@biph.kiev.ua (D.E. Duzhyy), pasha@biph.kiev.unana@biph.kiev.ua (N.V. Voitenko).

0925-4439/$ – see front matter © 2013 Elsevier B.V. Alhttp://dx.doi.org/10.1016/j.bbadis.2013.01.017

a b s t r a c t

a r t i c l e i n f o

Article history:Received 16 November 2012Received in revised form 31 December 2012Accepted 22 January 2013Available online 30 January 2013

Keywords:DiabetesPDNDRGPatch clampVOCCCalcium imaging

Streptozotocin (STZ)-induced type 1 diabetes in rats leads to the development of peripheral diabetic neuropathy(PDN) manifested as thermal hyperalgesia at early stages (4th week) followed by hypoalgesia after 8 weeks ofdiabetes development. Here we found that 6–7 week STZ-diabetic rats developed either thermal hyper- (18%),hypo- (25%) or normalgesic (57%) types of PDN. These developmentally similar diabetic rats were studied inorder to analyze mechanisms potentially underlying different thermal nociception. The proportion of IB4-positive capsaicin-sensitive small DRG neurons, strongly involved in thermal nociception, was not alteredunder different types of PDN implying differential changes at cellular and molecular level. We further focusedon properties of T-type calcium and TRPV1 channels, which are known to be involved in Ca2+ signaling and path-ological nociception. Indeed, TRPV1-mediated signaling in these neurons was downregulated under hypo- andnormalgesia and upregulated under hyperalgesia. A complex interplay between diabetes-induced changes infunctional expression of Cav3.2 T-type calcium channels and depolarizing shift of their steady-state inactivationresulted in upregulation of these channels under hyper- and normalgesia and their downregulation underhypoalgesia. As a result, T-type window current was increased by several times under hyperalgesia partially un-derlying the increased resting [Ca2+]i observed in the hyperalgesic rats. At the same time Cav3.2-dependent Ca2+

signaling was upregulated in all types of PDN. These findings indicate that alterations in functioning of Cav3.2T-type and TRPV1 channels, specific for each type of PDN, may underlie the variety of pain syndromes inducedby type 1 diabetes.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Peripheral diabetic neuropathy (PDN) is one of the earliest, the mostfrequent and troublesome complications of diabetes mellitus occurringin about 66% of patients with type 1 diabetes [1]. As in humans, develop-ment of PDN in rats is accompanied with various alterations in painsensation (hyper-, hypoalgesia, allodynia) or leaves pain sensation

DRG, dorsal root ganglia; HMP,A, low/high voltage activated;ns); PCD, peak current density;al latency; SSI, steady-state in-ermal pain sensitivityare. There are no other conflicts

f Natl. Acad. of Sci. of Ukraine, 4049; fax: +380 442562053.ula), viatchen@biph.kiev.ua. Borisyuk),a (P.V. Belan),

l rights reserved.

unchanged (normalgesia) [2–5] and these alterationsmay be consideredas manifestation of different PDN types. Rats with streptozotocin (STZ)-induced diabetes, a model of diabetes mellitus type 1, exhibit strongthermal hyperalgesia at 4th week, progressing to hypoalgesia after8 weeks of diabetes development [2]. At the same time, 6–7 week STZ-diabetes represents a transition period of PDN, when the rats havingthe same age and terms of diabetes are expected to reveal eitherhyper-, or norm- or hypoalgesia [5] providing a model for PDN-type-specific remodeling of cellular and molecular systems.

Impairment of Ca2+ homeostasis [6–9] and remodeling of voltage-and ligand-gated ion channels [10–12] in nociceptive neurons havebeen implicated in development of PDN. Among ion channels modulat-ed in STZ-diabetes T-type calcium channels [13] and TRPV1 [14,15] areboth of exclusive importance for pathological nociception and directlyinvolved in intracellular calcium signaling. T-type calcium channels arecrucially involved in both acute [16–21] and neuropathic pain [22–25].Moreover, in vivo selective block of T-type calcium channels effectivelyreverses thermal hyperalgesia in STZ-diabetes [13]. On the other hand,TRPV1 channels, “molecular integrators” of painful stimuli [26–30],also play a key role in variations of thermal pain sensitivity (TPS) duringPDN development, that was directly demonstrated by unchanged TPS in

637E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

STZ-diabetic vs non-diabetic TRPV1 deficient mice [15]. The changes inIB4-positive small-sized DRG neurons [31], expressing both TRPV1 andT-type channels [32,33], are of particular interest. These neurons,being nonpeptidergic thermal C-type nociceptors (NTCN) [34], arestrongly involved in TPS [35] and play a key role in neuropathic pain[36,37].

Here rats with thermal hyper-, hypo- and normalgesic PDN devel-oped at the same age and duration of STZ-diabetes were used to de-termine PDN type-specific remodeling of TRPV1 and T-type calciumchannels and their Ca2+ signaling in NTCN neurons.

2. Materials and methods

2.1. Experimental animals

MaleWistar ratswere housed in the animal facility of the BogomoletzInstitute of Physiology (Kyiv, Ukraine), which was maintained at 22 °C,55% relative humidity, with natural light/dark cycle. The animalsreceived a standard laboratory diet and tap water ad libitum. All experi-mental protocols were approved by the Animal Care and Use Committeeof the Bogomoletz Institute of Physiology (Kyiv, Ukraine) andwere in ac-cordance with the National Institutes of Health Guide for the Care andUse of Laboratory Animals. Every effort was made to minimize animalsuffering and the number of animals used.

2.2. Induction of experimental diabetes

We used a well establishedmodel of streptozotocin (STZ) injectionsto induce diabetic neuropathy in young male Wistar rats (30–50 g,21–23 days old) [9,38]. Experimental diabetes was induced in rats bya single intraperitoneal injection of STZ solution (80 mg/kg, i.p.). STZwas prepared in saline (0.9% NaCl adjusted to pH 4.3 with citric acid)on ice; the solutionwas discarded if bubblingwas noted.We used intact(naive) age-matched rats for most control in vitro experiments. Bloodglucose levels were checked on the 3rd day after the injection (to verifydiabetes onset) and just before electrophysiological experiments(6–7 weeks after injections), using a blood glucometer (Accu-check Ac-tive; Roche Diagnostics, Indianapolis, IN, USA). Rats with values of>270 mg/dL (15 mM) were considered hyperglycemic and were in-cluded in the respective experimental group.

2.3. Assessment of thermal nociception (behavioral experiments)

Nociceptive responses to thermal stimulation were measured usingHargreaves' method [16,39]. A paw thermal stimulation system (Plan-tar Test (Hargreaves's Apparatus), Ugo Basile, Italy) was used in theseexperiments. Each animalwas placed in a plastic chamber of the systemto accommodate for 15 min. A radiant heat source mounted on a mov-able holder beneath the glass floor was positioned to deliver a thermalstimulus to the plantar side of the hind paw. When the animal with-draws the paw, a photocell detects interruption of a light beam reflec-tion and the automatic timer shuts off. This method has a precision of±0.05 s for the measurement of paw withdrawal latency (PWL). Toprevent thermal injury, the light beam automatically discontinued at20 s if the rat failed to withdraw its paw. PWL was obtained for eachtested rat as a mean of 10measurements with 5 min interval interleav-ing left and right hind paws.

2.4. Acutely dissociated dorsal root ganglia neurons

We prepared dissociated dorsal root ganglia (DRG) cells and usedthem within 6–8 h for whole-cell recordings as described previously[40]. In brief, we dissected two L4 and two L5 DRGs from each animal,cut each one into three to four pieces, and incubated the pieces inTyrode's solution containing 140 mM NaCl, 4 mM KCl, 2 mM MgCl2,2 mM CaCl2, 10 mM glucose, and 10 mM HEPES, adjusted to pH 7.4

with NaOH and supplemented with 1 mg/ml protease Type XIV(Sigma) and 0.5 mg/ml collagenase Type I (Worthington BiochemicalCorporation) water bathed at 35 °C for 18–20 min. L4 and L5 gangliawere chosen since they contain cell bodies of sensory neurons innervat-ing peripheral receptive fields of the hind paws. Following incubation,gangliawerewashed three times (Tyrode's solution, room temperature)and dissociated into single cells in three cycles of trituration with glasspipettes having progressively narrower tip diameters. For recordings,the isolated neurons were plated onto an uncoated glass coverslip,placed in a culture dish, and perfused with Tyrode's solution. All record-ingswere obtained from isolatedDRG cells without visible processes. Allexperiments were done at room temperature. We routinely observedsmall, medium, and large sized neurons in the same preparation but,in this study, focused only on cells with a soma diameter less than30 μm (small DRG neurons) [41].

2.5. Fluorescent imaging

TILL Photonicswide-field imaging system (TILL Photonics, Gräfelfing,Germany) containing a monochromator Polychrome V and Imago CCDcamera was installed on an inverted microscope (IX71, Olympus) andcontrolled by TILLvision software (TILL Photonics). An oil immersion ob-jective (40× UV, NA 1.35; Olympus) was used for fluorescent imaging.Transmitted light images were captured using DIC of the same system.

2.6. Immunohistochemistry

Cells were incubated in Tyrode's solution supplemented with10 μg/ml isolectin B4 (IB4) [42] conjugated to Alexa Fluor® 568 dye(Invitrogen) in the dark for 10–12 min. Then a coverslip with thecells was mounted in a 500-μl recording chamber perfused with aroom temperature (22 °C) Tyrode's solution at a gravity driven flowrate of 2 ml/min. Cells were visualized using a standard Rodamine Fil-ter Set (Chroma Technology, USA). Fluorescent images were capturedunder standardized settings from 15 to 20 randomly selected smallDRG cells on each dish before any electrophysiological recordingsduring the first 15 min of each experiment. Attribution of each cell toone of three classes by ability of IB4-binding (strongly IB4-positive,weakly IB4-positive and IB4-negative) and analysis of obtained distri-bution were performed offline following Fang and collaborators [43].The mean intensity of halo of IB4 staining around the neuronal plasmamembrane was determined for each neuron using image analysistools of TILLvision software (TILL Photonics). Then the relative intensitywas calculated separately for neurons of each coverslip. The 0 and 100%intensity values for a particular coverslip were calculated by averagingthe halo intensity of the two least intensely (0%, a) and two most in-tensely stained cell profiles (100%, b). For each tested neuron, its halointensity of (c) was used to determine its relative intensity (percentageof maximum) as (100×(c−a)/(b−a))%. Neurons were judged as IB4positive (IB4+) if their relative intensities exceeded 20%, otherwisethey were scored as negative. Then IB4+ neurons were classed asstrongly IB4+ if their intensity was≥40% and weakly IB4+ if their in-tensity was in a range of 20–40%. Fractions of neurons in the respectiveclasses obtained for a single coverslip were then averaged for each ani-mal group.

2.7. Fluorescent calcium imaging

To measure cytosolic Ca2+ concentration ([Ca2+]i), the intrapipettesolution was supplemented with 200 μM of the high-affinity Ca2+ indi-cator Fura-2. After 10 min of intracellular dialysis, Fura-2 fluorescencereached a steady-state level. Autofluorescence of unloaded cells had sig-nal strength of less than 5% of the fluorescence of loaded cells. Fura-2fluorescence was recorded at 510±20 nmwavelength during alternat-ing 340- and 380-nmexcitation, and the ratio of fluorescence at 340 and380 nm (R) was calculated on a pixel-by-pixel basis. The frame capture

638 E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

period was 2 ms at an interval of 50 ms. Each neuron was specified as aregion of interest in the digital image (TILLvision software) and an addi-tional backgroundareawas recorded in eachfield for on-line subtractionof background fluorescence [44]. Fluorescent recordings of [Ca2+]i tran-sients were performed simultaneously with electrophysiological re-cordings of corresponding currents by means of synchronization ofPatchMaster and TILLvision software. [Ca2+]i was calculated for eachneuron from the fluorescent ratio, R, averaged over the region of inter-est corresponding to the neuron according to the equation [45]:

Ca2þh i

i¼ K � R−Rmin

R−Rmax; ð1Þ

The constants Rmin, Rmax and K determined by calibration were 0.29,3.8 and 690 nM, correspondently.

The amplitude of [Ca2+]i transient was measured from aprestimulated level to the peak. Neurons with a resting level of[Ca2+]i>300 nM were excluded from statistics.

2.8. Electrophysiology

Electrophysiological recordings were performed using a standardwhole-cell technique. Electrodes were pulled from borosilicate glassmicrocapillaries with a filament (Sutter Instrument, Novato, CA) andhad a resistance from 3 to 4 MΩ when filled with an internal solution.The recordings were performed using an EPC-10⁄2 amplifier controlledby PatchMaster software (HEKA, Freiburg, Germany). FitMaster soft-ware (HEKA, Freiburg, Germany) was used for an offline data analysis.Currents were low-pass filtered at 2–5 kHz. Series resistance (Rs) andcapacitance (Cm) valueswere taken directly from readings of the ampli-fier after electronic subtraction of the capacitive transients. A series re-sistance was compensated to the maximum extent possible (usually~60–80%).

Multiple independently controlled glass syringes served as reser-voirs for a gravity-driven local perfusion system. All drugs wereprepared as stock solutions: capsaicin, 10 mM; L-cysteine and Ni2+,100 mM. Drugs were freshly diluted to the appropriate concentrationsat the time of experiments. L-Cysteine and Ni2+ were prepared in H2O,and capsaicin was prepared in DMSO. The maximum concentration ofDMSO in any one experiment was 0.02%; at that concentration, DMSOhad no effect on the calcium currents or membrane potential [17].

For voltage-clamp experiments, the external solution used to isolatecalcium currents contained the following (in mM): 2 CaCl2, 2 MgCl2,158 tetraethylammonium (TEA)-Cl, 10 Glucose and 10 HEPES adjustedto pH 7.4 with TEA-OH. The internal solution used for all voltage-clamprecordings contained the following (in mM): 146 CsCl, 2 MgATP, 2MgCl2, 0.5 GTP-Na, 5 2Na-Phosphocreatin and 10 HEPES, adjusted topH 7.3 with CsOH. The internal solution was always supplementedwith 200 μM Fura-2 serving also as a Ca2+ buffer.

All chemicals were obtained from Sigma (St. Louis, MO) unlessotherwise noted.

Following previous works [38,46,47] T-type calcium current (LVAcomponent) was isolated using a protocol (inset of Fig. 4A) that includ-ed two double-pulse stimulations consisting of a preconditioning pulse(to either−95 mV or−60 mV) for 3 s followed by a “command” pulseto −45 mV for 250 ms, during which currents and corresponding[Ca2+]i transients were simultaneously recorded. The stimulation withthe preconditioning pulse to −95 mV was used to elicit a total (a sumof LVA and HVA components) current (Fig. 4A1, a black trace). In 10 s(after [Ca2+]i recovery) the stimulation with the preconditioningpulse to −60 mV was applied to inactivate the LVA component and toelicit the HVA current only (Fig. 4A1, a gray trace). The T-type calciumcurrent (Fig. 4A2) was obtained by subtraction of the HVA currentfrom the total current. The same subtraction was also performed for[Ca2+]i transients (Fig. 4A3). The command pulse was chosen to be to

−45 mV to minimize the HVA component while keeping the LVA com-ponent near its maximum.

Responses to capsaicin were recorded at the end of experimentalprocedure by bath application of Tyrode's solution supplementedwith capsaicin (2 μM).

2.9. Analysis

Statistical comparisons were performed using unpaired Student's ttest, one-way ANOVA, and χ2 or Fisher's exact test. All quantitativedata are expressed as means of multiple experiments±SEM.

2.10. Analysis of electrophysiological data

The amplitude of T-type current was measured as a difference be-tween the current peak value and a current value at the end of adepolarizing command pulse in order to avoid a residual HVA current.T-type current inactivation kinetics was estimated as a decay constantof single exponential fit of each recorded current. Activation kineticswas estimated as a time-to-peak of each recorded current.

Voltage dependencies of activation and steady-state inactivationwere described with single Boltzmann distributions of the followingforms:

for activation:

G Vð ÞGmax

¼ 1

1þ exp − V�V1=2

k

� � ð2Þ

for inactivation:

I Vð ÞImax

¼ 1

1þ exp V−V1=2

k

� � ð3Þ

where conductance (G(V)) was defined as PCD /(V−Er) (PCD is a peakcurrent density defined as Ipeak/Cm and (V−Er) is an electrodrivingforce for a membrane potential (V) and reversal potential (Er)obtained from interpolation of I(V) dependence); Gmax is the maximalconductance and Imax is the maximal peak current amplitude; V1/2 is avoltage at which half of the current is activated or inactivated, and krepresents the voltage dependence (slope) of the distribution. Thefitted values for V1/2 and k are reported with 95% linear confidencelimits.

The predicted peak current density of the T-type current evokedby a depolarization step to −45 mV (PCD−45, Fig. 7) was calculatedfor different holding membrane potentials (V) as a point-by-pointproduct of the steady-state inactivation (I(V)/ Imax) and the experi-mental PCDs recorded for steps from −95 to −45 mV (that isPCD−45(−95)) (Fig. 4C); the products were further divided by a de-gree of channel availability at −95 mV (I(−95)/ Imax) taken fromFig. 5Bb (0.77 in control, 0.92 in hypo-, 0.91 in hyper-, and 0.89 innormalgesia):

PCD−45 Vð Þ ¼ PCD−45 −95ð Þ �I Vð Þ.

Imax

I −95ð Þ�Imax

ð4Þ

The prediction for the window current densities (iwindow, definedas Iwindow/C) was calculated for different membrane potentials (V)by point-by-point product of PCD−45 (Fig. 6A) and activation curves(G(V)/Gmax)(Fig. 5Ab) divided by a degree of channel activation at−45 mV (G(−45)/Gmax) (taken from Fig. 5Ab: 0.64 in control, 0.72

639E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

in hypo-, 0.68 in hyper-, 0.71 in normalgesia) and multiplied by aratio of driving forces at V and at −45 mV ((V−Er)/(−45−Er)):

iwindow Vð Þ¼ PCD−45 Vð Þ �G Vð Þ.

Gmax

G −45ð Þ�Gmax

� V−Er−45−Er

ð5Þ

We paid close attention to the fidelity of voltage-clamp recordings.Cellswere discarded from further quantitative analysis of current kinet-ics if we noted signs of poor voltage control such as slow tail currents,sudden increase in the amplitude of inward current with a depolarizingstep of 10 mV increments, and/or irregular shape of the inward current.

2.11. Estimation of window T-type current contribution to the resting[Ca2+]i

The estimation is based on a fact, that in a resting state a Ca2+ influxdue to the window current (Iwindow) should be compensated by Ca2+

extrusion current (Iextr), i.e. Iwindow= Iextr. The value of Iextr was estimat-ed based on an amplitude (A) and decay time constant (τtr) of [Ca2+]itransient (Fig. 4D) induced by the T-type current (Fig. 4B) having PCD,iT, and a decay time constant, τi. To simplify calculations of charge (Q)introduced into a neuron the T-type current was considered asmono-exponential, and consequentlyQ= iT×C×τI, where C is a neuroncapacitance. We considered that for small changes of [Ca2+]i, from theresting level induced by T-type current the Ca2+ extrusion is propor-tional to changes in [Ca2+]i: Iextr=k1×Δ[Ca2+]i, where k1 is a constantand Δ[Ca2+]i=[Ca2+]i−[Ca2+]i, resting. For small [Ca2+]i transients wealso suggested a proportionality between the amplitude of [Ca2+]i tran-sient and the charge introduced by the T-type current: A=k2×Q, wherek2 is the other constant. From the previous equations k2 can beexpressed as k2=A/Q=A/(iT×C×τi). Next we have assumed thatrate of [Ca2+]i recovery is proportional to a rate of Ca2+ extrusion rep-resented by Iextr with the same constant k2: dΔ[Ca2+]i/dt=−k2× Iextr.After substitution of Iextr for k1×Δ[Ca2+]i we obtain: dΔ[Ca2+]i /dt=−k2×k1×Δ[Ca2+]i, that is an equation for exponentially decaying[Ca2+]i transient with a decay time constant of 1/(k2×k1) definedabove as τtr. Thus, k1 can now be calculated as: k1=k2/τtr= iT×C×τi /(A×τtr). Substituting k1 into the equation Iextr=k1×Δ[Ca2+]i and hav-ing in mind that Iwindow= Iextr the contribution of the window currentto the resting [Ca2+]i may be estimated as Δ[Ca2+]i= Iwindow/k1:

Δ Ca2þh i

i¼ A� τtr

τi� iwindow

iTð6Þ

where iwindow= Iwindow/C is the T-type window current density.

3. Results

3.1. TPS variation in the transition period of PDN

Within three days after the injection of STZ, most (70%) rats devel-oped hyperglycemia (mean glucose concentration 29±2 mM) andwere considered as diabetic. At the same time, saline-injected animalsmaintained a normal blood glucose level (4.1±0.2 mM) and did notalter a thermal nociception threshold (n=10, p>0.9). In order to testexistence and direction of changes in thermal nociception for ratswith 6–7 weeks of STZ-induced diabetes (n=60) their pawwithdraw-al latency (PWL), as a measure of TPS [16,39], was compared to agematched control (n=90; Fig. 1A). PWL distributions for both animalgroups were Gaussian (Lilliefors normality test, p>0.05). There wasno significant difference between means of the distributions (Fig. 1B):corresponding values were 12.2±0.2 s in control and 13.0±0.5 sunder the diabetes. On the other hand a standard deviation of PWL dis-tribution was significantly greater under diabetes (4.2 s) as comparedto control (1.6 s) (p=8·10−15, F-test, F=0.16, Fcrit=0.68, Fig. 1C).

Thus, PWL distribution under diabetic conditions had no shift but itwas significantly wider than in control.

95% confidence interval of PWL distribution for control rats was8.9÷15.5 s. Based on this result, the diabetic rats were divided intothree groups (Fig. 1A): hyperalgesic (PWLb8.9 s, significantly exag-gerated response), hypoalgesic (PWL>15.5 s, significantly attenuat-ed response) and normalgesic (unchanged response, PWL within8.9÷15.5 s). Hyper-, hypo- and normalgesic rats consisted 18% (11of 60), 25% (15 of 60) and 57% (34 of 60) of diabetic animals, respec-tively (Fig. 1D). The averaged PWLs for hyper-, hypo- and normalgesicgroups were 7.7±0.2 s, 18.7±0.5 s, and 13.2±0.5 s, correspondingly(Fig. 1E). A blood glucose level of diabetic animals was significantly dif-ferent from that of control rats; at the same time no significant differ-ences in the blood glucose level and body weight were observedbetween experimental rats of different diabetic groups (Table 1).

Thus, hyper-, norm- and hypoalgesic animals are simultaneouslypresent within the population of rats of the same age and terms ofSTZ-induced diabetes. This was considered asmanifestation of differenttypes of PDN. 6–7 week STZ diabetic rats with hyper-, hypo- andnormalgesic types of PDNwere further used to analyzemechanisms un-derlying alterations of thermal nociception induced by type 1 diabetes.

3.2. Development of PDN does not affect the proportion of nonpeptidergicC-type nociceptors

Diabetes development may potentially result in cell death [48].Moreover, a differential loss of peptidergic and nonpeptidergic primarynociceptors correlates with changes in TPS [49]. Therefore, we checkedwhether different types of PDN were associated with changes in a pro-portion of nonpeptidergic versus peptidergic C-type nociceptors.

Using isolectin B4 (IB4) [50] for in vitro labeling of nonpeptidergic neu-rons [34], small DRGneurons (n=608), considered as C-type nociceptors[51], were divided into three classes based on the intensity of IB4 binding[43]: strongly IB4-positive,weakly IB4-positive and IB4-negative. Fractionsof these neuronal classes were unchanged in all PDN groups compared tocontrol (p>0.05, χ2-test; Fig. 2). The IB4-positive small DRG neurons,considered as nonpeptidergic polymodal C-type nociceptors [34,43],were the most abundant and constituted 73±6%, 76±7%, 78±8%, and76±7%of control, hyper-, hypo- andnormalgesic groups, corresponding-ly. Known as being involved in thermal nociception [35] and neuropathicpain [36], this neuronal class became the focus of further experiments.

Thus, there was no significant redistribution between C-typenociceptors of IB4-positive, weakly IB4-positive and IB4-negative neuro-nal classes under different types of PDN. Therefore, differences in TPSare unlikely to have developed due to specific changes in the proportionof nonpeptidergic versus peptidergic C-type primary nociceptors, rath-er implying impairments of molecular signaling systems in theseneurons.

3.3. Signaling by TRPV1 channels is oppositely modified under hyperalgesicversus hypoalgesic PDN

TRPV1 channels are involved in TPS [26,52] and, therefore,changes in their functioning in the primary nociceptors may poten-tially result in TPS abnormalities observed in rats with PDN. Thus,we studied TRPV1 signaling under different types of PDN. For thatIB4-positive small-size DRG neurons were held in a voltage-clampmode at −60 mV and challenged with TRPV1 channels agonist, cap-saicin. Most tested neurons (>85%) responded to capsaicin applica-tion (2 μM, 15 s) with an inward current (Fig. 3A) and rise in[Ca2+]i (Fig. 3C). These capsaicin-sensitive IB4-possitive small-sizeDRG neurons were considered as NTCN neurons and were the focusof all further experiments in this work.

No statistically significant difference between the animal groupswas observed (p>0.05, χ2-test) in a fraction of NTCN neurons amongIB4-positive small-size DRG neurons. However, significant PDN type-

Fig. 1. Different types of TPS are expressed in ratswith 6–7 weeks of STZ-induceddiabetes. (A)Normalizeddistribution of PWLs for control (n=90) and agematcheddiabetic (n=60) rats. Thebin width is 1.6 s. The diabetic rats were divided into three groups: normalgesic (Dn) having PWL in a range of 8.9÷15.5 s (corresponding to 95% confidential interval for control rats),hyperalgesic (D+)with PWL less than 8.9 s, andhypoalgesic (D−)with PWLexceeding 15.5 s. (B,C)Means (B) of PWLdistributions for control anddiabetic ratswerenot significantly different,(Student's t-test p>0.05) while their standard deviations (C) differed significantly (***— Fisher's F-test pb0.001 vs control). (D) Fractions of hyper-, hypo- and normalgesic rats within the di-abetic group. (E) PWLs of control, hyper-, hypo- and normalgesic rats.

640 E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

specific changes in capsaicin-induced currents and [Ca2+]i transientswere found in these neurons. A peak current density (PCD) ofcapsaicin-induced responses under hyperalgesic PDN was increasedby 45±11% as compared to control (201±16 pА/pF, n=15 underhyperalgesia vs. 139±12 pА/pF, n=41 in control; pb0.05) (Fig. 3B).On the contrary, in neurons of hypo- and normalgesic rats PCD was

Table 1Blood glucose level and body weight of experimental rats.

Control Diabetica

Hyperalgesic Normalgesic Hypoalgesic

Blood glucose level, mM 4.2±0.2b 25±4 31±2 28±1Body weightc, g 158±6 148±17 119±20 130±24n 20 7 6 6

a There was no significant differences between hyper-, hypo- and normalgesic diabeticrats (p>0.1, ANOVA one-way test).

b t-test: pb0.001 for control vs each diabetic group.c There was no significant differences between hyper-, hypo- and normalgesic diabetic

and control rats (p>0.1, ANOVA one-way test).

significantly decreased as compared to control by 43±15% and 48±12%, correspondingly (79±21 pА/pF, n=14 under hypoalgesia and73±16 pА/pF, n=15 under normalgesia; pb0.05). Similarly, ampli-tudes of capsaicin-induced [Ca2+]i transients were significantly in-creased (by 112±11%, n=15, pb0.001) under hyperalgesic PDN anddecreased under hypoalgesic (by 67±8%; n=14, pb0.001) andnormalgesic (by 67±7%, n=15, pb0.001) PDN as compared to control(Fig. 3D).

A ratio of amplitudes of capsaicin-induced [Ca2+]i transient toPCD reflects a degree of involvement of intracellular Ca2+ regulatingsystems in the neuronal Ca2+ response to the same amount of Ca2+

entry. As compared to control this ratio was increased by 43±3%(n=15, pb0.01) under hyperalgesic PDN and decreased underhypoalgesic (by 55±8%, n=14, pb0.01) and normalgesic (by 38±10%, n=15, pb0.05) PDN (Fig. 3F) implicating different impairmentsof intracellular Ca2+-regulating systems under hyperalgesic versusnormalgesic and hypoalgesic PDN.

These results indicate that in NTCNneurons TRPV1-mediated signal-ing is significantly altered under PDN conditions. Moreover, these

0%

20%

40%

60%

80%

100%

strongly IB4+ weakly IB4+ IB4-N

euro

nal f

ract

ion,

% D+D-Dn

C

Fig. 2. Proportion between IB4-positive and IB4-negative small DRG neurons was not altered in rats with 6–7 weeks of STZ-induced diabetes. Bars represent fractions of stronglyIB4-positive (IB4+), weakly IB4+ and IB4-negative (IB4-) small DRG neurons in a control (C) (n=189 stained neurons), hyper- (D+) (n=147), hypo- (D−) (n=137) andnormalgesic (Dn) (n=140) groups. Fractions of strongly IB4+, weakly IB4+ and IB4− neurons were not significantly different between 4 studied groups (p>0.05, χ2-test).An inset shows a transmitted light (DIC optics) (top) and fluorescent (bottom) images of the same strongly IB4+ small DRG neuron. Note an intensive fluorescent ring associatedwith the neuronal plasma membrane. Scale bar, 20 μM.

641E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

alterations are oppositely directed under hyperalgesic and hypoalgesicPDN implying TRPV1 channels as a molecular system responsible fordifferences in TPS. A significant correlation between capsaicin re-sponses and TPS (Fig. 3E) also corroborates this idea. However, thesame responses to capsaicin observed in normalgesic and hypoalgesicPDN suggest an involvement of other molecular systems in the TPSalterations.

3.4. PDN-type-specific alterations of T-type calcium channel signaling inNTCN neurons

Recent studies demonstrate that T-type calciumchannels [47,53] arekey mediators of nociceptive signaling by primary sensory neurons inacute and neuropathic pain [16–21], including TPS in the hyperalgesicPDN [13,38,54]. Therefore, we further examined T-channel-mediatedsignaling in NTCN neurons under different types of PDN.

T-type currents (Fig. 4B,C) and corresponding [Ca2+]i transients(Fig. 4D,E) isolated at −45 mV after a preconditioning at −95 mV(Fig. 4A) were substantially altered in NTCN neurons under PDN con-ditions as compared to control. T-type PCD (T-PCD) (Fig. 4C) was in-creased by 38±8% under hyperalgesic PDN compared (5.0±0.4 pА/pF, n=15) to control (3.6±0.3 pА/pF, n=41; pb0.05). On the con-trary, T-PCD was decreased by 26±8% under hypoalgesic PDN(2.7±0.3 pА/pF, n=14; pb0.05) and by 26±9% under normalgesicPDN (2.6±0.2 pА/pF, n=15; pb0.01) compared to control. At thesame time inactivation kinetics of T-type current was not significantlydifferent between control (26±2 ms), hyper- (27±4 ms), hypo-(27±3 ms) and normalgesic (23±2 ms) groups (ANOVA, p>0.05).

Amplitudes of corresponding [Ca2+]i transients were significantlyincreased in all PDN groups compared to control (n=6) (Fig. 4D,E):by 66±24% under hyperalgesia (n=6, pb0.05), by 55±18% underhypoalgesia (n=5, pb0.05) and by 88±18% under normalgesia(n=5, pb0.05). At the same time, a ratio of amplitude of [Ca2+]itransient to T-PCD was significantly increased only under hypoalgesic(by 261±15%, n=5, pb0.01) and normalgesic (by 290±62%, n=5,pb0.01) PDN, whereas under hyperalgesic PDN changes were not sig-nificant (52±32% n=6, p>0.5, Fig. 4F) compared to control (n=6).The increase of the ratio implies that intracellular Ca2+ response tothe same amount of Ca2+ entry was specifically enhanced underhypo- and normalgesia corroborating specific disruptions of intracel-lular Ca2+-regulating systems of NTCN neurons under different typesof PDN.

Thus, we have found that T-channel-mediated signaling is sub-stantially modified in NTCN neurons under PDN condition. Moreover,observed changes are oppositely directed under hyperalgesic versusnormalgesic and hypoalgesic PDN.

3.5. Cav3.2 (α1H) isoform of T-type channels is mainly expressed inNTCN neurons under control and diabetic conditions

Cav3.2 isoform of T-type channels is most abundantly expressed inDRG neurons [55]. To study functional contribution of Cav3.2 T-typechannels to T-type current in NTCN neurons under control and PDNconditions we examined pharmacological properties of T-type currentusing such Cav3.2-specific pharmacological tools as T-type channelblockers, Ni2+ and amiloride, at low micromolar concentrations [56]as well as an endogenous T-type channel agonist, L-cysteine [57].

T-type currentwas significantly and reversibly blocked by bothNi2+

(50 μM) and amiloride (500 μM) in NTCN neurons of all experimentalgroups. Ni2+ blocked 77±6% of T-PCD in the control (n=6), 79±5%in the hyperalgesic (n=5), 75±7% in the hypoalgesic (n=5) and82±5% in the normalgesic (n=5) animal groups. Amiloride blocked84±4% of T-PCD in the control (n=7), 86±3% in the hyperalgesic(n=5), 82±3% in the hypoalgesic (n=5), and 83±4% in thenormalgesic (n=5) groups. L-Cysteine (100 μM) significantly and re-versibly enhanced T-type currents in neurons of control rats as well asin neurons of PDN animals. L-Cysteine increased the T-PCD by 63±11% in the control (n=7), by 53±9% in the hypoalgesic (n=6), by58±9% (n=6) in the hyperalgesic and 56±8% in the normalgesic(n=5) groups. The effects of Ni2+, amiloride and L-cysteine were sta-tistically significant in all groups (pb0.001), but therewasno significantdifference in the effect between the groups for each pharmacologicaltool (ANOVA, p>0.05).

Taken together these findings demonstrate the same high contribu-tion (>90%) of Cav3.2 T-type channels to the total T-type current inNTCN neurons of control and PDN rats with no detectable alteration inthe proportion of Cav3.2 isoform in the total pool of T-type channelsduring the transition period of PDN.

3.6. PDN alters biophysical properties of T-type channels in NTCN neurons

Changes in voltage-dependent activation and steady-state inacti-vation (SSI) of T-type channels substantially influence the neuronalexcitability [58]. Therefore, we studied these biophysical propertiesof T-type channels in NTCN neurons under different types of PDN.

The voltage-dependence of T-type channel activation, assessed viaa conventional double-pulse protocol (Fig. 5Aa), revealed no signifi-cant difference between midpoints (V1/2) as well as slope factors (k)of the activation curves in control (V1/2=−48±1 mV, k=5.9±0.8 mV, n=19), hyperalgesic (V1/2=−49±2 mV, k=5.2±0.3 mV,n=9), hypoalgesic (V1/2=−50±1 mV, k=5±1 mV, n=8) andnormalgesic (V1/2=−50±1 mV, k=5.8±0.7 mV, n=13) types ofPDN (Fig. 5Ab; ANOVA, p>0.05). No significant differences betweenthe control and PDN groups (ANOVA, p>0.05) were also observed

-250

-150

-100

-50

0

PC

D, p

A/p

F

* **

*

DnD-D+C

### ###

0%

50%

100%

150%

200%

[Ca2+

]i, %

of c

ontr

ol

***

*** ***

DnD-D+C

### ###

-200

0%

50%

100%

150%

200%

[Ca2+

]i / P

CD

, % o

f con

trol

**

***

DnD-D+C

######

-300

-250

-200

-150

-100

-50

0

6 11 16 21 26

PWL, s

PC

D, p

A/p

F C

D+

D-

Dn

R 2 =0 .42

(dF = 83, p~10 -16 )

A B

C D

E F

10 s

1 nA

Capsaicin (2 mM)

C

D+

D-

Dn

10 s

200nM

Capsaicin (2 mM)

C

D+

D-

Dn

Fig. 3. TRPV1 channels were downregulated under hypo- and normalgesia, while hyperalgesia was associated with their upregulation. (A) Representative traces of currents induced by acti-vation of TRPV1 channelswith capsaicin (2 μM) inNTCNneurons of control and diabetic ratswith hyper-, hypo- and normalgesic types of PDN. (B) PCD of TRPV1 channels in theNTCNneuronsof PDN rats were significantly different from control (*— pb0.05, **— pb0.01) and oppositely changed in hyper- versus hypo- and normalgesic rats. (C) Representative traces of [Ca2+]i tran-sients corresponding to the currents shown in a. Baselines of traces were aligned to emphasize differences in amplitudes. (D) Normalized amplitudes of [Ca2+]i transients were significantlydifferent from control (*— pb0.05, **— pb0.01) and also oppositely changed in hyper- versus hypo- and normalgesic rats. (E) PCD of TRPV1 channels plotted against PWL of correspondinganimals. A significant positive correlationwas found between the PCD and PWL. (F)Normalized ratios of amplitudes of capsaicin-induced [Ca2+]i transients to the respective PCD. *—pb0.05, **— pb0.01, values significantly different from control; ###— pb0.001 for hypo- and normalgesic groups versus hyperalgesia. n neurons: 41 in C, 15 inD+, 14 inD− and 15 inDn groups.N rats:15 in C, 5 in D+, 5 in D− and 5 in Dn groups.

642 E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

for a time-dependence of activation measured as a time-to-peak overa range of potentials from −65 mV to −30 mV (Fig. 5Ac). Thus, acti-vation properties of T-type channels seem to be not affected underPDN conditions.

The voltage-dependence of SSI (Fig. 5Bb) was obtained from T-PCDsrecorded at−40 mV in another conventional double-pulse protocol in-volving the application of a series of preconditioning pulses (Fig. 5Ba).We found about 7 mV depolarizing shift in the midpoint (V1/2) of SSIcurve in all PDN groups compared to control (Fig. 5Bc): correspondingvalues of V1/2 were −89±2 mV in control cells (n=19), −81±2 mVunder hyperalgesia (n=9; pb0.01), −82±2 mV under hypoalgesia(n=8; pb0.01) and −83±1 mV under normalgesia (n=13;pb0.01). The slope factor of the SSI curve was not significantly affected

under PDN as compared to the control (5.4±0.2 mV): correspondingvalues were 6.0±0.2 mV under hyperalgesia, 5.4±0.4 mV underhypoalgesia and 6.0±0.3 mV under normalgesia (ANOVA, p>0.05).Thus, in contrast to unchanged activation properties, inactivation ofT-type channels is significantly affected in a similar way for differenttypes of PDN.

3.7. Specific upregulation of T-type channels under hyperalgesia andnormalgesia

Diabetes-induced changes in T-PCD described in Section 4 (Fig. 4C)were found using a holding membrane potential (HMP) of −95 mV.Due to a diabetes-induced shift in SSI (Fig. 5Bb) a degree of the T-PCD

0%

100%

200%

300%

400%

500%

**## ** ##

DnD-D+C

-6

-4

-2

0

PC

D, p

A/p

F

##*

*

##*

DnD-D+C1s

10n

M

C

D+

D-

Dn

0T-type

20p

A

0

total

HVA

50 ms

2s

10n

M

HVAtotal

1 3

2

-45m V,250 ms

3s

h = -60 mV

-95m V

50 ms

50p

A

C

D+

D-

Dn0

E F

A

C D

B

0%

50%

100%

150%

200%

250%

** **

DnD-D+C

Δ[C

a2+]i,

% o

f co

ntr

ol

Δ[C

a2+]i

/ PC

D, %

of

con

tro

l

Fig. 4. Each PDN type is associatedwith specific changes in T-type channel signaling. (A) Electrophysiological isolation of T-type currents (1, 2) and corresponding [Ca2+]i transients (3).The T-type current (A2)was obtained by a digital subtraction of high-voltage activated current (HVA, evoked from holding potential of−60 mV) from the total voltage activated calciumcurrent (A1) evoked after a 3 s preconditioning at a potential of−95 mV. Both currents were evoked by a depolarizing step to−45 mV (see an inset). The same subtraction procedurewas performed for the respective [Ca2+]i transients (A3). (B) Representative traces of T-type currents (after subtraction procedure) in the NTCN neurons of control, hyper-, hypo- andnormalgesic rats. (C) Pooled results of peak current densities of T-type channels, T-PCD, in the NTCN neurons of four studied groups. (D) Representative traces of [Ca2+]i transients(after subtraction procedure) corresponding to currents shown in a panel (B). (E) Pooled results of normalized amplitudes of [Ca2+]i transients induced by activation of T-type channels.(F) Pooled normalized ratios of amplitudes of [Ca2+]i transients to the respective PCD. *— pb0.05 and **— pb0.01 versus control; ##— pb0.01 for hypo- and normalgesic groups versushyperalgesia. n currents: 41 in C, 15 in D+, 14 in D− and 15 in Dn groups. n transients: 6 in C, 6 in D+, 5 in D− and 5 in Dn groups.

643E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

changes depends onHMP. Therefore, we compared T-PCD evoked at thesame potential of −45 mV (T-PCD−45) under conditions of differentHMP in NTCN neurons of hyper-, hypo- and normalgesic animals versuscontrol ones (see Eq. (4) in Materials and methods). The obtained de-pendencies of T-PCD−45 onHMP clearly show specific and complex pat-tern of changes observed under each type of PDN (Fig. 6A).

To visualize the relative changes in T-PCD−45 their dependencies foreach PDN typewere normalized point by point to corresponding controlvalues (Fig. 6B,C). Since activation properties of T-type channels werealmost the same for all experimental groups (Fig. 5Ab) such normalizedT-PCD were considered as dependent only upon HMP.

Firstwe looked atHMP≤−100 mV,where T-type channelswere al-most fully recovered from inactivation. We found that T-PCD underhyperalgesia was not significantly different from control, while it was

almost twice lower compared to control under hypo- and normalgesia(Fig. 6C).

For HMP≥−90 mV T-PCD was significantly increased underhyperalgesia compared to control, suggesting functional upregulationof T-type channels. At the same time, under hypoalgesia T-PCD wasnot significantly different from control (Fig. 6B), indicating the absenceof T-type channel upregulation.

In case of normalgesia T-PCD values were very similar to thoseunder hypoalgesia for HMP≤−75 mV. However, under normalgesiain contrast to hypoalgesia, T-PCD values were significantly increasedcompared to control for HMP≥−70 mV (Fig. 6B), suggesting function-al upregulation of T-type channels but less pronounced than inhyperalgesia. It is also interesting to note that there was a large (almosttwofold) and significant difference of T-PCD in a case of hyperalgesia

A B

-100

-90

-80

-70

-60

V1/

2, m

V

** ** **

DnD-D+C

0%

50%

100%

-120 -100 -80 -60

Vholding, mV

I/Im

ax

CD+D-Dn

0

20

40

60

-70 -60 -50 -40 -30

Vcommand, mV

TT

P, m

s

CD+D-Dn

0%

50%

100%

-80 -60 -40 -2 0

Vcommand, mV

CD+D-Dn

G/G

max

50 ms

20p

A

-55 mV

-115 mV

-65 mV

-60m V -40m V,250 ms

-115…-55

mV,3s

50 ms

20p

A

-70 mV

-30 mV

-70…-30m V250 ms

-95m V,3s

-60m V

a

b

c

Fig. 5. PDN increase T-type channel availability inNTCNneuronsof ratswith6–7 weeks of STZ-induceddiabetes. (A)Voltage-dependent activationof theT-type currentwasnot affected in PDNrats. (Aa) Representative series of current traces demonstrates voltage-dependent activation of the T-type current. The currents were evoked by test pulses to−70 through−30 mV in 5 mVincrement after a 3 s preconditioning at−95 mV. An inset illustrates the stimulation protocol. The T-type currentwas isolated from the total current by theprocedure described in Fig. 4A1. (Ab)Normalized peak conductance defined as Ipeak/(V−Er)was plotted against test potentials for theNTCNneurons of control and PDN rats. Solid lines arefits obtained using an equationG/Gmax=1/(1+exp(−(V−V1/2)/k)) (see Materials and methods for details). Note that the average fits for control and all PDN groups almost overlap. (Ac) Pooled results of time to peak, TTP, plottedagainst test potentials for the NTCN neurons of control and PDN rats. (B) T-type channel availabilitywas increased in PDN rats due to a shift of steady-state inactivation, SSI. (Ba) Representativecurrent traces illustrate SSI of the T-type current (note a presence of stable non-inactivating HVA current). The currents were evoked by test steps to −40 mV after a 3 s preconditioning atpotentials from −115 to −55 mV in 10 mV increment. An inset illustrates a stimulation protocol. (Bb) Normalized peak T-type currents plotted against the preconditioning potentials forthe NTCN neurons of control and PDN rats. Solid lines are fits obtained using an equation I=Imax/(1+exp((V−V1/2)/k)), giving potential of half-maximal availability (V1/2) and slope factor(k) characterizing SSI. Note a parallel shift to the right for all PDN groups compared control. (Bc) The potential of half-maximal availability for different animal groups (values obtained fromfitted curves in (Bb)) demonstrating a similar shift of SSI curves for all PDN groups. **— pb0.01 vs. control; p>0.05 between all PDN groups. n: 19 in C, 9 in D+, 8 in D− and 13 in Dn groups.

644 E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

versus hypoalgesia within the whole tested range of HMPs (Fig. 6D). Atthe same time the difference in T-PCD between hyper- and normalgesiabecame insignificant for HMP≥−70 mV (Fig. 6D).

Thus, T-type channels seem to be significantly upregulated nearthe resting membrane potential specifically in case of hyper- andnormalgesic PDN.

3.8. PDN affects T-type window current and resting [Ca2+]i in NTCNneurons

An important functional consequence of the depolarizing shift of SSIunder PDNwas a larger overlap area between activation and inactivationcurves that may result in enhancement of a window Ca2+ current viaT-type channels [59]. To test this hypothesis, the window current was

calculated at different membrane potentials for control, hyper-, hypo-and normalgesic groups (see Eq. (5) in Materials andmethods). The cal-culations revealed a tendency to increase of T-type window under PDNconditions compared to control, although significance was reachedonly under hyperalgesic (at HMP≥−60 mV) and normalgesic (atHMP≥−55 mV). A sixfold increase was observed under hyperalgesianear the resting membrane potential (at −60 mV) (Fig. 7A). As theT-type window current may contribute to the resting level of [Ca2+]i[60–62] we compared a resting [Ca2+]i in NTCN neurons clamped at−60 mV between all experimental groups of animals (Fig. 7B). Wefound that the resting [Ca2+]i was significantly increased underhyperalgesia (230±20 nM, n=9, pb0.001) and unchanged undernormalgesia (73±9 nM, n=7, p>0.95) compared to control (72±11 nM, n=41). At the same time the resting [Ca2+]i was slightly

0

2

4

6

8

10

-110 -90 -70 -50

HMP, mV

T-P

CD

/ T

-PC

DC

* **

**

*

****

* **

*

* * * **

-6

-5

-4

-3

-2

-1

0

-110 -90 -70 -50

HMP, mV

T-P

CD

-45,

pA

/pF

D-

D+

Dn

C

**

*

*

*********

*****

** * **

0

1

2

3

-110 -100 -90 -80

HMP, mV

T-P

CD

/ T

-PC

DC

*** *

**

**

*** *

0

0.2

0.4

0.6

0.8

1

1.2

-110 -90 -70 -50

HMP, mV

T-P

CD

/ T-P

CD

Hyp

er

#

#

#

#

#

#

#

#

#

#

#

#

#

##

## # # ##

A B

C D

Fig. 6. PDN-type-specific remodeling of T-type Ca2+ channels. (A) Predicted peak current densities of the T-type currents evoked by a depolarizing step to−45 mV, T-PCD−45, for the NTCNneurons of control and PDN groups of animals plotted against the holding membrane potentials, HMP. (B) T-PCD−45 of hyper-, hypo- and normalgesic groups point-by-point normalized tocontrol (T-PCD/T-PCDC) and plotted against the HMP. Note a robust amplification of the T-type current in the physiological range of HMP under hyperalgesia. (C) A part of (B) enlarged forthe HMP from −110 to −80 mV. Note a decreased functional expression of T-type channels under hypo- and normalgesia rather than under hyperalgesia. (D) T-PCD−45 of hypo- andnormalgesic groups point-by-point normalized to hyperalgesic ones (T-PCD/T-PCDHyper) and plotted against the HMP emphasizing a difference of T-type current in the NTCN neurons betweenthe hyperalgesic versus hypo- and normalgesic rats in thewhole range of HMP. *— pb0.05 vs control; #— pb0.01 vs hyperalgesia; black * and# are for hyper- and hypoalgesia; gray * and# arefor normalgesia.

645E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

increased in the case of hypoalgesia (120±20 nM, n=13, pb0.05) inspite of unchanged T-typewindow current, possibly due to impairmentsof Ca2+ clearance from the cytosol [63,64]. Thus, the hyperalgesic PDN isassociated with the drastic increase of both the resting [Ca2+]i andT-type window current in NTCN neurons.

-35

-30

-25

-20

-15

-10

-5

0

-110 -90 -70 -50 -30

HMP, mV

CD

win

dow, f

A/p

F

D-

D+

Dn

C

* **

*

**

*

**

*

*

*

2+

A B

Fig. 7. PDN-type-specific alterations in T-typewindowcurrent and resting [Ca2+]i. (A)Apredictedof animals plotted against theHMPs. (B) Resting [Ca2+]i at theHMPof−60 mV in theNTCNneuro* is for normalgesia. n neurons in (B): 10 in C, 9 in D+, 13 in D− and 7 in Dn groups.

4. Discussion

In this study we have investigated functioning of Ca2+ regulatingTRPV1 and T-type calcium channels under hyperalgesic versushypoalgesic and normalgesic PDN in NTCN neurons that terminate in

0

100

200

300

***

*

[Ca

]i, n

M

C D+ D- Dn

density of T-typewindowcurrent, CDwindow, for theNTCNneurons of control andPDNgroupsns of different animal groups. *—pb0.05 vs control;black * is for hyper- andhypoalgesia; gray

646 E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

lamina II and play an important role in neuropathic pain [34]. We havedemonstrated that differences in TPS between hyper-, hypo- andnormalgesic diabetic rats are likely due to differential changes in func-tioning of TRPV1 and T-type calcium channels within the existing poolof NTCNneurons rather than due to elimination of this specific neuronalclass or redistribution between nonpeptidergic and peptidergic primarynociceptors. Our results provide a better understanding of potentialmolecularmechanisms involved in expression of different types of PDN.

4.1. Specific remodeling of TRPV1 channels was associated with differenttypes of PDN

TRPV1 channels were reciprocally dysregulated under hyper- andhypoalgesia in the transition period of PDN development in STZ-diabeticrats of the same age and duration of diabetes (Fig. 3A,B). Moreover, a sig-nificant correlation was found between PWL and TRPV1-mediated PCD(Fig. 3E). In addition, TRPV1-mediated currents and [Ca2+] transientswere correspondently threefold and sixfold higher in hyperalgesia versusboth hypoalgesia and normalgesia. Together these findings corroboratean important role of TRPV1 channels in expression of hyperalgesia andhypoalgesia in a transition period of PDN development. Our results arein agreement with previous reports about association between changesin TRPV1 expression and direction of abnormalities in TPS shown for an-imals of different ages and durations of STZ-diabetes (2–4th and 8thweeks) [14,15].

[Ca2+] transients evoked by TRPV1 activationwere also reciprocallyaltered in hyper- and hypoalgesia (Fig. 3C,D). However, the significantchanges in a transient-to-current ratio (Fig. 3F) indicate that the differ-ences in [Ca2+] transients are not proportional to the correspondingchanges in PCD under hyper- and hypoalgesic PDN. This likely reflectsimpairments of intracellular calcium-regulating and calcium-bufferingsystems observed previously [8,9,63,64].

Noteworthily, the restoration of thermal pain sensitivity innormalgesic animals does not indicate recovery from PDN, sinceTRPV1 and T-type channels and their Ca2+ signaling are still signifi-cantly impaired. Hence, post-STZ normalgesia may be truly consid-ered as a separate type of PDN.

It is interesting to note that TRPV1 signaling under normalgesia isdownregulated to the same extent as under hypoalgesia. We suggestthat different remodeling of Cav3.2 T-type channels under normalgesicand hyperalgesic PDN (Fig. 7B) underlies the differences in TPS betweenthese PDN types.

4.2. PDN-type-specific remodeling of Cav3.2 T-type channels

Kinetics of inactivation (Fig. 4), time- and voltage-dependence ofactivation (Fig. 5) and pharmacological properties of T-type currents(Section 3.5) in the NTCN neurons did not differ between control,hyper-, hypo- and normalgesic animals. These findings indicate thesame high contribution (>90%) of Cav3.2 isoform to the total T-typecurrent in control and during the transition period of PDN. This con-clusion is consistent with previous findings reporting that Cav3.2 isthe most abundant isoform of T-type channels in DRG neurons [55]and underlie the T-type current in small [13] and medium [38] DRGneurons in control and under early hyperalgesic PDN conditions.

We have found that SSI of T-type current is right shifted by about7 mV independently on the type of PDN, resulting in increased chan-nel availability at the same HMP (Fig. 5Bb). This seems to be a very im-portant finding, because even tiny modifications in SSI voltagedependence of the T-type channelsmay dramatically modify neuronalfiring [58]. Noteworthily, no SSI shift was reported in the small DRGneurons in 3–8 days after i.v. STZ injection [13]. However, the shiftwas detected in the 2nd week of STZ-diabetes [54]. Thus, the SSIshift has an onset between the 1st and 2nd weeks of STZ-diabetesand, according to our results, persists at least up to the 7th week, im-plicating effects of prolonged hyperglycemia in its development [65].

In spite of the similarly increased availability of T-type channels, re-modeling of these channels is specific for each PDN type (Fig. 6A). In theNTCN neurons of hyperalgesic rats functional expression of T-typechannels was not altered compared to control (Fig. 6B,C). In spite ofthis, increased availability only was enough to produce a significantupregulation of Cav3.2 T-type channels, as indicated by a significant in-crease in the T-PCD found at HMP≥−95 mV (Figs. 4 and 6). It is nota-ble that the upregulation became dramatic at resting (−70…−50 mV)membrane potentials (Fig. 6B). This upregulation of T-type calciumchannels should certainly contribute to the increased excitability of sen-sory neurons and hyperalgesia [13,19,21,38,66–71]. Thus simulta-neously upregulated Cav3.2 T-type and TRPV1 channels may play inconcert in expression andmaintenance of thermal hypersensitivity dur-ing the transition period of PDN.

In the hypoalgesic animalswe have found that functional expressionof T-type channelswas almost twice lower compared to control (Fig. 6B,C). Comparable downregulation of Cav3.2 alone in healthy rats alsoresulted in the development of hypoalgesia [13,19]. Possibly, a certainlevel of T-type channel functional expression is necessary for the ther-mal nociceptive neurons to be amplifiers of peripheral pain signalsand a decrease in Cav3.2 expression below this level may result in de-creased neuronal excitability that may contribute to hypoalgesia[13,19]. This conclusion is also supported by a greater degree of diabetichypoalgesia in Cav3.2−/− compared to control mice [68]. It is interest-ing to note that in the hypoalgesic animals T-PCD was almost twicelower compared to hyperalgesic ones in the whole tested range ofHMP (Fig. 6D) indicating that the value of T-PCD is associated with aparticular type of TPS abnormality. Thus, downregulation of Cav3.2 oc-curring in concert with TRPV1 downregulation may account for the de-velopment of hypoalgesia during the transition period of PDN.

Wehave not found significant differences in functional expression ofT-type channels (Fig. 6C) and TRPV1-channel remodeling (Fig. 3)between normalgesia and hypoalgesia. However, for HMP≥−70 mVa significant functional upregulation of T-type channels was observedunder conditions of normalgesia, rather than hypoalgesia (Fig. 6B).This may increase excitability of NTCN neurons of normalgesic ratsand produce a compensatory effect on TRPV1 channel downregulationleading to normalgesia.

Thus, altered functional expression of T-channels together with theSSI shift results in either upregulation of T-channels (observed inhyper- and normalgesia) or downregulation (observed in hypoalgesia).

Several knownmolecularmechanismsmayunderlie the depolarizingshift in SSI of T-type currents. Themost appealing candidates are alterna-tive splicing of Cav3.2 isoform [72], modulation by G protein β–γ sub-units [73] and Rho kinase dependent phosphorylation of Cav3.2 [74].Many different factors may also contribute to the decreased functionalexpression of T-type channels [21] found in the NTCN neurons underhypo- and normalgesic PDN. However, the physiological relevance ofthese mechanisms to the Cav3.2 function in the DRG neurons has notbeen studied providing several potential avenues of research aimed atdeciphering the molecular basis of diabetes-induced modulation ofT-type channels.

4.3. Diabetes-induced remodeling of Ca2+ signaling by T-type channels

Remodeling of T-type channels shown in the current study could po-tentiallymodulate Ca2+ signaling of NTCNneurons. Indeed, [Ca2+]i in re-sponse to an equal Ca2+ influx via T-type channels is several times higherin hypo- andnormalgesic rats compared to control andhyperalgesic onesas indicated by changes in the transient-to-current ratio (Fig. 4F). Since ithas been earlier shown that Ca2+ mobilization from internal stores israther decreased under STZ-induced diabetes [63,64], this result indi-cates that a fast Ca2+ buffering is substantially downregulated underhypo- and normalgesia.

T-type channels can be a considerable source of Ca2+ entry into theNTCN neurons of hyperalgesic rats due to a sixfold increase of T-type

647E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

window current (up to about 0.5 pA compared to less than 0.1 pA incontrol) found in this study (Fig. 7A). This constant Ca2+ influx shouldintroduce about 3∗10−18 mol/s of Ca2+ corresponding to 0.1 Hz con-stant firing (as far as single AP introduces about 2∗10−17 mol of Ca2+

[75]). Since nociceptive neurons are almost silent in terms of AP gener-ation [76], the T-type window current may substantially contribute tothe overall Ca2+ influx in the NTCN neurons. To pump this additionalCa2+ influx out, a neuron should hydrolyze ~20 μM/min of ATP thatconsiderably increases metabolic load on neurons during STZ-diabetes(up to ~10% of total ATP production rate [77]).

Ca2+ influx introduced by the T-type window current may also in-fluence a resting [Ca2+]i [47,60–62]. We have estimated this influencefor NTCN neurons of hyperalgesic rats based on an approach describedin Materials and methods. Following parameters of T-type current(Fig. 4B,C), related [Ca2+]i transient (Fig. 4D,E) and T-typewindow cur-rent (Fig. 7A) were used for calculations using Eq. (6) of Materials andmethods: A=50 nM, τtr=5 s, iT=5 pA/pF, τi=20 ms and iwindow=20 fA/pF. Our estimations demonstrate that T-type window currentmay contribute about 50 nM to the resting [Ca2+]i, at least partially un-derlying its substantial increase observed in the hyperalgesic diabeticanimals (Fig. 7B).

4.4. The functional consequences of T-type channel remodeling

It has been conclusively demonstrated that upregulation of Cav3.2T-type channels results in a decreased threshold for action potentialgeneration and burst firing in DRG neurons of control [42] andhyperalgesic diabetic rats [38]. Thus, a reciprocal remodeling ofT-channels shown in the current study could potentially modulateperipheral nociceptive input of NTCN neurons, contributing to TPS al-terations. Our finding about a manifold increase of T-type windowcurrent in the NTSN neurons of hyperalgesic and normalgesic ratsalso seems to be very important due to a previously reported positivecorrelation between the T-type window current [47] and cellular ex-citability [42,78].

Ca2+ signaling by T-type channels may possibly regulate a widespectrum of Ca2+-dependent channels expressed in NTCN neurons, inparticular, SK and BK calcium-activated potassium channels. SK andBK channels are certainly expressed in the IB4 positive small-sizedDRG neurons [79,80] and can substantially modulate their excitability[81]. Since the coupling of T-type Ca2+ channels to BK and SK channelshas been reported in a number of bursting neurons [82], altered Ca2+

signaling by T-type channels may influence activity of BK and SK chan-nels and impact excitability of NTCN neurons under PDN conditions.

Recently presynaptic T-type channels were implicated in regula-tion of synaptic transmission specifically between the nociceptiveDRG neurons and dorsal horn neurons of nociceptive lamina I,II[83,84]. Therefore, T-type channel remodeling observed in this workmay alter synaptic transmission between NTCN and nociceptive dor-sal horn neurons of rats with PDN in that way contributing to a differ-ent pain processing at the level of central terminals of NTCN neuronsunder different types of PDN.

4.5. Remodeling of TRPV1 and T-type channels during PDN development

All in all, in rats of the same age and duration of diabetes in the tran-sition period of PDN development we have observed the following: (1)thermal hyperalgesia is associated with cooperative upregulation ofCav3.2 channels and TRPV1; (2) thermal normalgesia is associated withupregulation of Cav3.2 channels and downregulation of TRPV1; (3) ther-mal hypoalgesia was associated with cooperative downregulation ofCav3.2 channels and TRPV1.

We suggest the following scenario of TRPV1 and T-type channelmodifications during diabetes development. At the early stage of diabe-tes there is a cooperative upregulation of Cav3.2 channels (initially dueto the increased functional expression [13] and later on due to the right

shift in a steady-state channel inactivation [54]) and TRPV1 in the NTCNneurons resulting in increased peripheral sensitization and develop-ment of thermal hypersensitivity. Later on a progressive decrease infunctional expression of TRPV1 and Cav3.2 channels leads to the re-duced functional expression of these channels. However persistentright shift in a steady-state inactivation of Cav3.2 channel (and conse-quently increased T-type channel availability) can still maintainupregulation of T-type channel for a certain period of time. This coun-teraction of the upregulation of T-type channel and the downregulationof TRPV1 leads to the normalgesia during the transition period of PDNdevelopment. A further decrease in Cav3.2 functional expression after8 weeks of diabetes can't be compensated by increased T-type channelavailability. At that stage cooperative downregulation of TRPV1 [15] andCav3.2 channel leads to the development of hypoalgesia.

In summary, our results demonstrate that specific and simultaneousalterations in functioning of Cav3.2 T-type and TRPV1 channels in theNTCN neurons may underlie the variety of pain syndromes induced bytype 1 diabetes.

Acknowledgments

The authors declare no competing financial interests. This work wassupported by NASU Biotechnology, STCU #5510, DFFD F46.2/001 andF47/066 grants.

References

[1] C. Gooch, D. Podwall, The diabetic neuropathies, Neurologist 10 (2004) 311–322.[2] N.A. Calcutt, J.D. Freshwater, A.P. Mizisin, Prevention of sensory disorders in diabetic

Sprague–Dawley rats by aldose reductase inhibition or treatment with ciliaryneurotrophic factor, Diabetologia 47 (2004) 718–724.

[3] K.K. Beiswenger, N.A. Calcutt, A.P. Mizisin, Dissociation of thermal hypoalgesia andepidermal denervation in streptozotocin-diabetic mice, Neurosci. Lett. 442 (2008)267–272.

[4] I.G. Obrosova, Diabetic painful and insensate neuropathy: pathogenesis and poten-tial treatments, Neurotherapeutics 6 (2009) 638–647.

[5] M. Dobretsov,M.M. Backonja, D. Romanovsky, J.R. Stimers, Animalmodels of diabeticneuropathic pain, Neuromethods 49 (2011) 147–169.

[6] T.J. Huang, N.M. Sayers, P. Fernyhough, A. Verkhratsky, Diabetes-induced alterationsin calcium homeostasis in sensory neurones of streptozotocin-diabetic rats are re-stricted to lumbar ganglia and are prevented by neurotrophin-3, Diabetologia 45(2002) 560–570.

[7] E. Kostyuk, N. Voitenko, I. Kruglikov, A. Shmigol, V. Shishkin, A. Efimov, P.Kostyuk, Diabetes-induced changes in calcium homeostasis and the effects ofcalcium channel blockers in rat and mice nociceptive neurons, Diabetologia 44(2001) 1302–1309.

[8] I. Kruglikov, O. Gryshchenko, L. Shutov, E. Kostyuk, P. Kostyuk, N. Voitenko,Diabetes-induced abnormalities in ER calcium mobilization in primary and sec-ondary nociceptive neurons, Pflugers Arch. 448 (2004) 395–401.

[9] L. Shutov, I. Kruglikov, O. Gryshchenko, E. Khomula, V. Viatchenko-Karpinski, P.Belan, N. Voitenko, The effect of nimodipine on calcium homeostasis and pain sensi-tivity in diabetic rats, Cell. Mol. Neurobiol. 26 (2006) 1541–1557.

[10] J.N. Campbell, R.A. Meyer, Mechanisms of neuropathic pain, Neuron 52 (2006)77–92.

[11] A. Veves,M. Backonja, R.A.Malik, Painful diabetic neuropathy: epidemiology, naturalhistory, early diagnosis, and treatment options, Pain Med. 9 (2008) 660–674.

[12] A.I. Basbaum, D.M. Bautista, G. Scherrer, D. Julius, Cellular and molecular mecha-nisms of pain, Cell 139 (2009) 267–284.

[13] R.B. Messinger, A.K. Naik, M.M. Jagodic, M.T. Nelson, W.Y. Lee, W.J. Choe, P. Orestes,J.R. Latham, S.M. Todorovic, V. Jevtovic-Todorovic, In vivo silencing of the Ca(V)3.2T-type calcium channels in sensory neurons alleviates hyperalgesia in rats withstreptozocin-induced diabetic neuropathy, Pain 145 (2009) 184–195.

[14] S. Hong, J.W. Wiley, Early painful diabetic neuropathy is associated with differentialchanges in the expression and function of vanilloid receptor 1, J. Biol. Chem. 280(2005) 618–627.

[15] R.M. Pabbidi, S.Q. Yu, S. Peng, R. Khardori, M.E. Pauza, L.S. Premkumar, Influenceof TRPV1 on diabetes-induced alterations in thermal pain sensitivity, Mol. Pain4 (2008) 9.

[16] S. Pathirathna, B.C. Brimelow, M.M. Jagodic, K. Krishnan, X. Jiang, C.F. Zorumski,S. Mennerick, D.F. Covey, S.M. Todorovic, V. Jevtovic-Todorovic, New evidencethat both T-type calcium channels and GABAA channels are responsible forthe potent peripheral analgesic effects of 5alpha-reduced neuroactive steroids,Pain 114 (2005) 429–443.

[17] S.M. Todorovic, V. Jevtovic-Todorovic, A.Meyenburg, S.Mennerick, E. Perez-Reyes, C.Romano, J.W. Olney, C.F. Zorumski, Redox modulation of T-type calcium channels inrat peripheral nociceptors, Neuron 31 (2001) 75–85.

648 E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

[18] S.M. Todorovic, A. Meyenburg, V. Jevtovic-Todorovic, Mechanical and thermalantinociception in rats following systemic administration of mibefradil, a T-typecalcium channel blocker, Brain Res. 951 (2002) 336–340.

[19] E. Bourinet, A. Alloui, A. Monteil, C. Barrere, B. Couette, O. Poirot, A. Pages, J. McRory,T.P. Snutch, A. Eschalier, J. Nargeot, Silencing of the Cav3.2 T-type calcium channelgene in sensory neurons demonstrates its major role in nociception, EMBO J. 24(2005) 315–324.

[20] S. Choi, H.S. Na, J. Kim, J. Lee, S. Lee, D. Kim, J. Park, C.C. Chen, K.P. Campbell, H.S. Shin,Attenuated pain responses in mice lacking Cav3.2 T-type channels, Genes BrainBehav. 6 (2007) 425–431.

[21] S.M. Todorovic, V. Jevtovic-Todorovic, T-type voltage-gated calciumchannels as targetsfor the development of novel pain therapies, Br. J. Pharmacol. 163 (2011) 484–495.

[22] A. Dogrul, L.R. Gardell, M.H. Ossipov, F.C. Tulunay, J. Lai, F. Porreca, Reversal of exper-imental neuropathic pain by T-type calcium channel blockers, Pain 105 (2003)159–168.

[23] M.E. Hildebrand, T.P. Snutch, Contributions of T-type calcium channels to the path-ophysiology of pain signaling, Drug Discov. Today Dis. Mech. 3 (2006) 335–341.

[24] T.P. Snutch, L.S. David, T-type calcium channels: an emerging therapeutic targetfor the treatment of pain, Drug Dev. Res. 67 (2006) 404–415.

[25] S.M. Todorovic, V. Jevtovic-Todorovic, Regulation of T-type calcium channels in theperipheral pain pathway, Channels (Austin) 1 (2007) 238–245.

[26] M.J. Caterina, M.A. Schumacher, M. Tominaga, T.A. Rosen, J.D. Levine, D. Julius, Thecapsaicin receptor: a heat-activated ion channel in the pain pathway, Nature 389(1997) 816–824.

[27] S.E. Jordt, M. Tominaga, D. Julius, Acid potentiation of the capsaicin receptor deter-mined by a key extracellular site, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 8134–8139.

[28] M.J. Caterina, D. Julius, The vanilloid receptor: a molecular gateway to the painpathway, Annu. Rev. Neurosci. 24 (2001) 487–517.

[29] D. Julius, A.I. Basbaum, Molecular mechanisms of nociception, Nature 413 (2001)203–210.

[30] M. Tominaga, M.J. Caterina, A.B. Malmberg, T.A. Rosen, H. Gilbert, K. Skinner, B.E.Raumann, A.I. Basbaum, D. Julius, The cloned capsaicin receptor integrates multiplepain-producing stimuli, Neuron 21 (1998) 531–543.

[31] C.L. Stucky, G.R. Lewin, Isolectin B(4)-positive and -negative nociceptors are func-tionally distinct, J. Neurosci. 19 (1999) 6497–6505.

[32] C.G. Cardenas, L.P. Del Mar, R.S. Scroggs, Variation in serotonergic inhibition ofcalcium channel currents in four types of rat sensory neurons differentiated bymembrane properties, J. Neurophysiol. 74 (1995) 1870–1879.

[33] J.C. Petruska, J. Napaporn, R.D. Johnson, J.G. Gu, B.Y. Cooper, Subclassified acutelydissociated cells of rat DRG: histochemistry and patterns of capsaicin-, proton-,and ATP-activated currents, J. Neurophysiol. 84 (2000) 2365–2379.

[34] C.J.Woolf, Q.Ma, Nociceptors—noxious stimulus detectors, Neuron55 (2007) 353–364.[35] L. Vulchanova, T.H. Olson, L.S. Stone,M.S. Riedl, R. Elde, C.N. Honda, Cytotoxic targeting

of isolectin IB4-binding sensory neurons, Neuroscience 108 (2001) 143–155.[36] W.D. Snider, S.B.McMahon, Tackling pain at the source: new ideas about nociceptors,

Neuron 20 (1998) 629–632.[37] C.L. Chen, D.C. Broom, Y. Liu, J.C. de Nooij, Z. Li, C. Cen, O.A. Samad, T.M. Jessell, C.J.

Woolf, Q. Ma, Runx1 determines nociceptive sensory neuron phenotype and is re-quired for thermal and neuropathic pain, Neuron 49 (2006) 365–377.

[38] M.M. Jagodic, S. Pathirathna, M.T. Nelson, S. Mancuso, P.M. Joksovic, E.R. Rosenberg,D.A. Bayliss, V. Jevtovic-Todorovic, S.M. Todorovic, Cell-specific alterations of T-typecalcium current in painful diabetic neuropathy enhance excitability of sensory neu-rons, J. Neurosci. 27 (2007) 3305–3316.

[39] K. Hargreaves, R. Dubner, F. Brown, C. Flores, J. Joris, A new and sensitive methodfor measuring thermal nociception in cutaneous hyperalgesia, Pain 32 (1988)77–88.

[40] V. Shishkin, E. Potapenko, E. Kostyuk, O. Girnyk, N. Voitenko, P. Kostyuk, Role ofmitochondria in intracellular calcium signaling in primary and secondary sensoryneurones of rats, Cell Calcium 32 (2002) 121–130.

[41] R.S. Scroggs, A.P. Fox, Calcium current variation between acutely isolated adultrat dorsal root ganglion neurons of different size, J. Physiol. 445 (1992)639–658.

[42] M.T. Nelson, P.M. Joksovic, E. Perez-Reyes, S.M. Todorovic, The endogenousredox agent L-cysteine induces T-type Ca2+ channel-dependent sensitizationof a novel subpopulation of rat peripheral nociceptors, J. Neurosci. 25 (2005)8766–8775.

[43] X. Fang, L. Djouhri, S. McMullan, C. Berry, S.G. Waxman, K. Okuse, S.N. Lawson, In-tense isolectin-B4 binding in rat dorsal root ganglion neurons distinguishes C-fibernociceptors with broad action potentials and high Nav1.9 expression, J. Neurosci.26 (2006) 7281–7292.

[44] A. Fuchs, M. Rigaud, Q.H. Hogan, Painful nerve injury shortens the intracellularCa2+ signal in axotomized sensory neurons of rats, Anesthesiology 107 (2007)106–116.

[45] G. Grynkiewicz, M. Poenie, R.Y. Tsien, A new generation of Ca2+ indicatorswith greatly improved fluorescence properties, J. Biol. Chem. 260 (1985)3440–3450.

[46] B.P. Bean, Two kinds of calcium channels in canine atrial cells. Differences in kinetics,selectivity, and pharmacology, J. Gen. Physiol. 86 (1985) 1–30.

[47] E. Perez-Reyes, Molecular physiology of low-voltage-activated t-type calcium chan-nels, Physiol. Rev. 83 (2003) 117–161.

[48] S.T. Britland, R.J. Young, A.K. Sharma, B.F. Clarke, Acute and remitting painful diabeticpolyneuropathy: a comparison of peripheral nerve fibre pathology, Pain 48 (1992)361–370.

[49] M.S. Johnson, J.M. Ryals, D.E. Wright, Early loss of peptidergic intraepidermal nervefibers in an STZ-induced mouse model of insensate diabetic neuropathy, Pain 140(2008) 35–47.

[50] M.B. Gerke,M.B. Plenderleith, Binding sites for theplant lectin Bandeiraea simplicifoliaI-isolectin B(4) are expressed by nociceptive primary sensory neurones, Brain Res.911 (2001) 101–104.

[51] A.A. Harper, S.N. Lawson, Conduction velocity is related tomorphological cell type inrat dorsal root ganglion neurones, J. Physiol. 359 (1985) 31–46.

[52] D.J. Cavanaugh, H. Lee, L. Lo, S.D. Shields, M.J. Zylka, A.I. Basbaum, D.J. Anderson, Dis-tinct subsets of unmyelinated primary sensory fibers mediate behavioral responsesto noxious thermal and mechanical stimuli, Proc. Natl. Acad. Sci. U.S.A. 106 (2009)9075–9080.

[53] M.C. Nowycky, A.P. Fox, R.W. Tsien, Three types of neuronal calcium channel withdifferent calcium agonist sensitivity, Nature 316 (1985) 440–443.

[54] S.A. Shankarappa, E.S. Piedras-Renteria, E.B. Stubbs Jr., Forced-exercise delaysneuropathic pain in experimental diabetes: effects on voltage-activated calciumchannels, J. Neurochem. 118 (2011) 224–236.

[55] E.M. Talley, L.L. Cribbs, J.H. Lee, A. Daud, E. Perez-Reyes, D.A. Bayliss, Differentialdistribution of three members of a gene family encoding low voltage-activated(T-type) calcium channels, J. Neurosci. 19 (1999) 1895–1911.

[56] J.H. Lee, J.C. Gomora, L.L. Cribbs, E. Perez-Reyes, Nickel block of three clonedT-type calcium channels: low concentrations selectively block alpha1H, Biophys.J. 77 (1999) 3034–3042.

[57] P.M. Joksovic, M.T. Nelson, V. Jevtovic-Todorovic, M.K. Patel, E. Perez-Reyes, K.P.Campbell, C.C. Chen, S.M. Todorovic, CaV3.2 is the major molecular substrate forredox regulation of T-type Ca2+ channels in the rat andmouse thalamus, J. Physiol.574 (2006) 415–430.

[58] A. Tscherter, F. David, T. Ivanova, C. Deleuze, J.J. Renger, V.N. Uebele, H.S. Shin,T. Bal, N. Leresche, R.C. Lambert, Minimal alterations in T-type calcium channelgating markedly modify physiological firing dynamics, J. Physiol. 589 (2011)1707–1724.

[59] J.T. Wolfe, H. Wang, E. Perez-Reyes, P.Q. Barrett, Stimulation of recombinant Ca(v)3.2, T-type, Ca(2+) channel currents by CaMKIIgamma(C), J. Physiol. 538 (2002)343–355.

[60] R. Assandri,M. Egger,M. Gassmann, E. Niggli, C. Bauer, I. Forster, A. Gorlach, Erythro-poietin modulates intracellular calcium in a human neuroblastoma cell line, J. Phys-iol. 516 (1999) 343–352.

[61] J. Chemin, A. Monteil, C. Briquaire, S. Richard, E. Perez-Reyes, J. Nargeot, P. Lory,Overexpression of T-type calcium channels in HEK-293 cells increases intracellularcalcium without affecting cellular proliferation, FEBS Lett. 478 (2000) 166–172.

[62] P.Mariot, K. Vanoverberghe, N. Lalevee,M.F. Rossier, N. Prevarskaya, Overexpressionof an alpha 1H (Cav3.2) T-type calcium channel during neuroendocrine differentia-tion of human prostate cancer cells, J. Biol. Chem. 277 (2002) 10824–10833.

[63] A. Verkhratsky, P. Fernyhough,Mitochondrialmalfunction andCa2+dyshomeostasisdrive neuronal pathology in diabetes, Cell Calcium 44 (2008) 112–122.

[64] P. Fernyhough, N.A. Calcutt, Abnormal calcium homeostasis in peripheral neurop-athies, Cell Calcium 47 (2010) 130–139.

[65] D.R. Tomlinson, N.J. Gardiner, Glucose neurotoxicity, Nat. Rev. Neurosci. 9 (2008)36–45.

[66] K.E. Hall, A.A. Sima, J.W.Wiley, Voltage-dependent calcium currents are enhanced indorsal root ganglion neurones from the Bio Bred/Worchester diabetic rat, J. Physiol.486 (1995) 313–322.

[67] K.E. Hall, J. Liu, A.A. Sima, J.W.Wiley, Impaired inhibitory G-protein function contrib-utes to increased calcium currents in rats with diabetic neuropathy, J. Neurophysiol.86 (2001) 760–770.

[68] J.R. Latham, K. Krishnan, D.F. Covey, V. Jevtovic-Todorovic, S.M. Todorovic, Cav3.2 isnecessary for thermal hyperalgesia but attenuates hypoalgesia in experimental dia-betic neuropathy: A possible dual role for T-channels in disease, NeuroscienceMeet-ing Planner, Society for Neuroscience, Chicago, 2009 (Personal communication,Program No. 521.24).

[69] W. Choe, R.B. Messinger, E. Leach, V.S. Eckle, A. Obradovic, R. Salajegheh, V.Jevtovic-Todorovic, S.M. Todorovic, TTA-P2 is a potent and selective blockerof T-type calcium channels in rat sensory neurons and a novel antinociceptiveagent, Mol. Pharmacol. 80 (2011) 900–910.

[70] M.M. Jagodic, S. Pathirathna, P.M. Joksovic, W. Lee, M.T. Nelson, A.K. Naik, P. Su,V. Jevtovic-Todorovic, S.M. Todorovic, Upregulation of the T-type calcium cur-rent in small rat sensory neurons after chronic constrictive injury of the sciaticnerve, J. Neurophysiol. 99 (2008) 3151–3156.

[71] L.A. Swayne, E. Bourinet, Voltage-gated calcium channels in chronic pain: emergingrole of alternative splicing, Pflugers Arch. 456 (2008) 459–466.

[72] X. Zhong, J.R. Liu, J.W. Kyle, D.A. Hanck, W.S. Agnew, A profile of alternativeRNA splicing and transcript variation of CACNA1H, a human T-channel genecandidate for idiopathic generalized epilepsies, Hum. Mol. Genet. 15 (2006)1497–1512.

[73] J. Tao, M.E. Hildebrand, P. Liao, M.C. Liang, G. Tan, S. Li, T.P. Snutch, T.W. Soong,Activation of corticotropin-releasing factor receptor 1 selectively inhibits CaV3.2T-type calcium channels, Mol. Pharmacol. 73 (2008) 1596–1609.

[74] M. Iftinca, J. Hamid, L. Chen, D. Varela, R. Tadayonnejad, C. Altier, R.W. Turner,G.W. Zamponi, Regulation of T-type calcium channels by Rho-associated kinase,Nat. Neurosci. 10 (2007) 854–860.

[75] D.P. McCobb, K.G. Beam, Action potential waveform voltage-clamp commands re-veal striking differences in calcium entry via low and high voltage-activated calciumchannels, Neuron 7 (1991) 119–127.

[76] W. Xie, J.A. Strong, D. Kim, S. Shahrestani, J.M. Zhang, Bursting activity inmyelinatedsensory neurons plays a key role in pain behavior induced by localized inflammationof the rat sensory ganglion, Neuroscience 206 (2012) 212–223.

[77] S. Genc, I.A. Kurnaz, M. Ozilgen, Astrocyte-neuron lactate shuttle may boost moreATP supply to the neuron under hypoxic conditions—in silico study supported byin vitro expression data, BMC Syst. Biol. 5 (2011) 162.

649E.V. Khomula et al. / Biochimica et Biophysica Acta 1832 (2013) 636–649

[78] F.M. Dreyfus, A. Tscherter, A.C. Errington, J.J. Renger, H.S. Shin, V.N. Uebele, V.Crunelli, R.C. Lambert, N. Leresche, Selective T-type calcium channel block in tha-lamic neurons reveals channel redundancy and physiological impact of I(T)win-dow, J. Neurosci. 30 (2010) 99–109.

[79] L.C. Mongan, M.J. Hill, M.X. Chen, S.N. Tate, S.D. Collins, L. Buckby, B.D. Grubb,The distribution of small and intermediate conductance calcium-activated po-tassium channels in the rat sensory nervous system, Neuroscience 131 (2005)161–175.

[80] X.L. Zhang, L.P. Mok, E.J. Katz, M.S. Gold, BKCa currents are enriched in a subpop-ulation of adult rat cutaneous nociceptive dorsal root ganglion neurons, Eur. J.Neurosci. 31 (2010) 450–462.

[81] X.H. Cao, S.R. Chen, L. Li, H.L. Pan, Nerve injury increases brain-derived neurotrophicfactor levels to suppress BK channel activity in primary sensory neurons, J. Neurochem.121 (2012) 944–953.

[82] L. Cueni, M. Canepari, J.P. Adelman, A. Luthi, Ca(2+) signaling by T-type Ca(2+)channels in neurons, Pflugers Arch. 457 (2009) 1161–1172.

[83] J. Bao, J.J. Li, E.R. Perl, Differences in Ca2+ channels governing generation of miniatureand evoked excitatory synaptic currents in spinal laminae I and II, J. Neurosci. 18 (1998)8740–8750.

[84] M.O. Jacus, V.N. Uebele, J.J. Renger, S.M. Todorovic, Presynaptic Cav3.2 channels reg-ulate excitatory neurotransmission in nociceptive dorsal horn neurons, J. Neurosci.32 (2012) 9374–9382.