Intracellular zinc inhibits KCC2 transporter activity

Abstract

We report that KCC2 activity, monitored with wide–field fluorescence, is inhibited by intracellular Zn²⁺, a key component of neuronal injury. Zn²⁺–mediated KCC2 inhibition produced a depolarizing shift of GABA_A reversal potentials in rat neurons. Moreover, oxygen–glucose deprivation attenuated KCC2 activity in a Zn²⁺–dependent manner. The link between Zn²⁺ and KCC2 activity provides a novel target for neuroprotection and may be important in activity–dependent regulation of inhibitory synaptic transmission.

The K⁺/Cl⁻ co-transporter-2 (KCC2) is the major Cl⁻ outward transporter in neurons, creating a Cl⁻ equilibrium potential negative to the resting membrane voltage and rendering GABA inhibitory. However, following acute and chronic neuronal injury, KCC2 activity is decreased and GABA becomes excitatory^1,2,3, a process that has been tightly linked to neurodegeneration⁴. Another important contributor to neuronal injury and death is an increase of cytosolic Zn²⁺ concentrations ([Zn²⁺]_i)⁵. Although both rise in [Zn²⁺]_i and decrease in KCC2 have been associated with neuronal injury, whether Zn²⁺ itself can influence KCC2 activity is unknown.

To test whether [Zn²⁺]_i modulates KCC2 activity we first monitored KCC2-mediated ion transport in HEK 293T cells expressing the co-transporter⁶. NH₄⁺ was used as a surrogate ion for K⁺ and changes in intracellular pH were monitored using the fluorescent dye BCECF (see Supplemental methods online). KCC2-expressing cells showed a 3–fold faster NH₄⁺-induced acidification rate than cells transfected with empty vector (Fig. 1a,b). Increasing [Zn²⁺]_i by a 2 min application of Zn²⁺ with the Zn²⁺ ionophore pyrithione (ZnPyr) was immediately followed by a decrease of KCC2–mediated acidification rate, with an IC₅₀ of ~50 µM Zn²⁺ (Fig. 1a–c). Application of Zn²⁺ without pyrithione did not change KCC2 activity (not shown), indicating that KCC2 is inhibited by intracellular Zn²⁺. Chelating [Zn²⁺]_i with N,N,N’,N’–tetrakis–(2–pyridylmethyl)– ethylenediamine (TPEN) reversed the effects of ZnPyr (Fig. 1b). Moreover, TPEN alone increased KCC2 activity (Fig. 1a,b), indicating that endogenous levels of [Zn²⁺]_i tonically inhibit KCC2.

Figure 1. Increase in [Zn²⁺]_i inhibits KCC2 activity.

(a) KCC2 activity was monitored with the pH-sensitive dye BCECF in HEK 293T cells tranfected with a KCC2 or empty vector. Cells were incubated for 2 min with vehicle, Zn²⁺ (100 µM) with pyrithione (5 µM) (ZnPyr), or TPEN (10 µM) and the rate of intracellular pH change following application of NH₄Cl (10 mM) was monitored (see Supplemental Methods online). (b) NH₄⁺-mediated acidification rates (mean ± SEM) in control and KCC2-expressing HEK 293T cells treated with vehicle (n=17), ZnPyr (n=14), TPEN (n=14), or ZnPyr followed by TPEN (n=10); *p<0.05, **p<0.005 ANOVA/Tukey, compared to KCC2-expressing NH₄Cl only group. (c) Concentration-inhibition curve of KCC2 activity by Zn²⁺ (n=11 for each concentration, mean ± SEM). (d) KCC2 activity monitored with the Cl⁻ sensitive dye MQAE. KCl (10 mM) was applied to KCC2-expressing HEK 293T cells and the rate of Cl⁻ dependent quenching of the signal was determined. (e) Summary of Cl⁻ transport rates; ZnPyr, or TPEN, compared to KCl alone (n=6 for each treatment; mean ± SEM); **p<0.005 ANOVA/Tukey. (f) Endogenous KCC2 activity monitored with BCECF in immature and mature cortical neurons.

Next, we determined changes in intracellular chloride [Cl⁻]_i directly with the Cl⁻ sensitive dye N-(ethoxycarbonylmethyl)6–methoxyquinolinium bromide (MQAE; see Supplemental Methods online). Application of 5 mM KCl reverses KCC2 transport leading to accumulation of intracellular Cl⁻ (Fig. 1d). Increasing [Zn²⁺]_i with ZnPyr blocked the Cl⁻ influx (Fig 1d,e), while TPEN enhanced it (Fig. 1d,e), similar to the results observed with NH₄⁺-induced acidification.

We then tested whether KCC2 inhibition by Zn²⁺ is also observed in cortical neurons, which endogenously express the co-transporter. In BCECF-loaded immature neurons (7 days in vitro, DIV), expressing low levels of KCC2, NH₄⁺ had little effect on intracellular acidification, similar to empty vector-transfected HEK 293T cells (Fig. 1f). In contrast, in mature neurons (>25 DIV), which express high levels of KCC2, NH₄⁺–induced acidification was pronounced. Importantly, in mature neurons ZnPyr also effectively attenuated KCC2 activity (Fig. 1f).

The Zn²⁺-dependent changes in KCC2 activity should affect the Cl⁻ gradient and thereby the reversal potential of GABA_A receptor–mediated currents (E_GABA). We measured E_GABA using the gramicidin perforated patch technique, which leaves intracellular Cl⁻ undisturbed. A 2 min ZnPyr treatment produced a gradual positive shift in E_GABA as, presumably, the intracellular Cl⁻ concentration decreased with KCC2 inhibition (Fig. 2a). E_GABA stabilized within 6–8 minutes after the ZnPyr application to a level ~20 mV positive to control (Fig. 2b,c,d). ZnPyr did not affect E_GABA under whole-cell recording conditions with equal extracellular and intracellular chloride (Fig. 2d). The Zn²⁺-induced shift in E_GABA reversed within 30 minutes after removing ZnPyr (Fig. 2d), indicating that neurons can tightly regulate [Zn²⁺]_i under non-injurious conditions.

Figure 2. Effects of [Zn²⁺]_i on E_GABA, and KCC2 inhibition by OGD in neurons.

(a) Representative gramicidin-perforated whole-cell currents (V_hold −50 mV) in response to 10 µM GABA before and after a 2 min treatment with ZnPyr (100 µM/5 µM). Scale bars: 300 pA, 1 s. (b) Gramicidin-perforated whole-cell currents before and after inhibition of KCC2 by ZnPyr. Scale bars: 300 pA, 1 s. (c) Current-voltage relationships of peak currents shown in (b). ZnPyr shifted E_GABA from −54 to −33 mV. (d) E_GABA measured with gramicidin-patch in control neurons (n=6), at <1, and 6–10 min after ZnPyr (n=10), and after 30 min rinse (n=3). E_GABA measured in whole-cell (equal extracellular and intracellular chloride conditions) under control conditions (n=10), and after ZnPyr (n=5; mean ± SEM ***p<0.001; ANOVA/Tukey) (e) OGD-induced increase in neuronal [Zn²⁺]_i inhibits KCC2. Neurons were exposed to OGD for 90 min and loaded with the Zn²⁺-sensitive dye FluoZin3. A significant Zn²⁺ rise, sensitive to TPEN (20 µM), was monitored at 1 hour following OGD. This increase in cytoplasmic Zn²⁺, was insensitive to the extracellular Zn²⁺ chelator tricine (1 mM), suggesting the source of the metal was intracellular (n=7–9; mean ± SEM ; *p<0.05; ANOVA/Bonferroni). (f) One hour following OGD treatment, neuronal KCC2 activity was monitored with NH₄⁺/BCECF. TPEN treatment during OGD abolished KCC2 inhibition (n=5–14; mean ± SEM ; **p<0.005; ANOVA/Tukey).

Cerebral ischemia induces acute Zn²⁺ dysregulation in neurons⁷. Therefore, we investigated if KCC2 is inhibited by the sustained increases in neuronal [Zn²⁺]_i following oxygen-glucose deprivation in mature neurons (OGD; Fig. 2e). Application of an extracellular Zn²⁺ chelator did not attenuate the [Zn²⁺]_i rise following OGD (Fig. 2e), strongly suggesting that it is liberated from intracellular binding sites⁸. Importantly, in ischemic neurons NH₄⁺-induced acidification rates were markedly decreased (Fig. 2f). Moreover, neurons exposed to TPEN during OGD had normal rates of acidification (Fig. 2f), indicating that the rise in [Zn²⁺]_i following ischemic insult led to the inhibition of KCC2.

Our results reveal a novel link between [Zn²⁺]_i and KCC2 activity. KCC2 inhibition by Zn²⁺ was rapid, suggesting that Zn²⁺ may directly interact with intracellular cysteine or histidine residues of KCC2. Alternatively, Zn²⁺ could contribute to the known regulation of KCC2 activity by phosphorylation⁹ leading to changes in its expression^9,10, or by oligomerization¹¹. Zn²⁺, however, did not induce changes in KCC2 transcription, surface expression, or oligomeric organization (see Supplemental Figs. 1 and 2 online), some of which have been observed in other neuronal injury models^2,3,4,9,10. Changes in KCC2 activity without changes in expression have also been reported in a deafness model¹².

The downregulation of KCC2 activity following OGD-induced Zn²⁺ rise may account for acute seizure activity following ischemia¹³. A major factor in neuronal injury following OGD is intracellular release of Zn²⁺ and it was recently shown that it is also critical for spreading depression triggered by OGD¹⁴. Interestingly, a decrease in GABA inhibition has been proposed to be a mechanism contributing to neuronal death following cerebral ischemia⁴. Our results indicate that Zn²⁺-dependent inhibition of KCC2 may be an important contributor to neuronal injury. We also suggest that influx of synaptically-released Zn²⁺ or its intracellular release, particularly during periods of intense synaptic activity¹⁵, may dynamically regulate inhibition by modulating KCC2 activity.