Dear Editor:

This letter describes our work identifying the neuronal targets of a clinical-stage stroke therapeutics CN-105, and proposing a novel neuroprotective strategy involving nAChR antagonism. Stroke is a devastating disease with high morbidity and mortality. CN-105 was originally designed to mimic the anti-inflammatory activities of endogenous apolipoprotein E (apoE). Despite its proven efficacy in various animal models of brain injury and well-established safety profile in clinical trials,^{1, 2} our understanding of CN-105’s mechanism of action remains incomplete. Early reports suggested that apoE-derived peptides and a number of oligoarginine species may interact directly with various neuronal targets including the nicotinic acetylcholine receptors (nAChR).^{3, 4} The long-held view regarding nAChR was that its activation was neuroprotective, best exemplified by the cognitive-enhancing effects of nAChR agonists or positive allosteric modulators.⁵ One common role of nAChR, which received limited attention in stroke, is its potentiation of glutamate release at the presynaptic terminal.⁶ Glutamatergic neurotransmission is arguably the primary reason for the propagation of excitotoxicity in stroke.^{7, 8} Although postsynaptic nAChR activated by ACh or nicotine could desensitize proximal NMDA receptors,⁹ we hypothesized that acute action of nAChR antagonist on presynaptic neurons may serve to downregulate the detrimental cascade associated with glutamate excitotoxicity.

After given supratherapeutic dosing of CN-105 several orders of magnitude higher than clinical practice, Cynologous monkey exhibited symptoms of mydriasis and ptosis. A gradual recovery to normal activity took about two serum clearance half-lives of CN-105 (4 h, Table S1; Figure S1). In C57BL6 mice, CN-105 had an LD₈₀ of 25 mg/kg (Figure 1A) with symptoms of spasm and respiratory suppression. Following intubation and mechanical ventilation, dosing of 50 mg/kg was not associated with mortality (Figure 1A), but led to extended sedation after the removal of isoflurane anaesthesia (Movie S1).

Given apoE-peptides’ antagonistic activity upon nAChR,^{3, 10} we hypothesized that the respiratory suppression of CN-105 was due to its inhibition of nAChRs present in various respiratory control pathways. An initial electrophysiological study using an HEK293 cell line overexpressing α7-nAChR showed that 10 μM CN-105 strongly suppressed the ACh induced current, similar to the inhibitory effect of apoE (Figure S2). In the presence of 10 nM to 10 μM CN-105, we observed a typical antagonistic suppression of the ACh-induced currents (Figure 1B). Such effect was sub-type specific, as the IC₅₀ of CN-105 differed by two orders of magnitude for α7-, α4β2- and α1β1δε-nAChR (Figure 1C), the latter of which might be responsible for CN-105’s respiratory toxicity at supratherapeutic dosage. The voltage dependency of CN-105 was measured in parallel with apoE130-149 (Figure 1D), in which we observed maximal inhibition at –80 mV. In the primary culture of rat neurons, 1 mM ACh triggered strong calcium influx, which was significantly diminished by a priori incubation with CN-105 prior to ACh infusion (Figure 1E). These results indicated that CN-105 may block the ACh-induced influx of Ca²⁺ and the propagation of action potential (AP).

Based on the reported binding site of the apoE140-148 peptide on nAChR,¹⁰ we constructed a model of the α7 nAChR-CN-105 complex. The putative binding site of CN-105 was proximal to the orthosteric binding site for ACh (Figure S3A) and would most likely block the binding of ACh. Unsurprisingly, the arginine side chains of CN-105 were involved in extensive hydrogen bonds (Figure S3B). Poisson–Boltzmann analyses of the complex structure (Figure 2A) showed that positive charges of CN-105 significantly weakened the negative electric field that is critical for the passage of cations, diminishing the propensity of cation inflow. The two preceding neutral residues were vital for the nAChR activity (Figure S3B). The peptide in which the first valine was replaced by an arginine lost most of its nAChR activity (Table S2) in spite of possessing an extra positive charge.

We next studied whether CN-105’s inhibition of nAChR and nAChR-mediated calcium influx would affect downstream release of glutamate. In rat hippocampal brain slices, evoked excitatory postsynaptic current (eEPSC) mediated by NMDA receptor and AMPA receptor were significantly lower in the presence of CN-105 than the control levels (Figure 2B, Figure S4). CN-105 did not directly inhibit glutamate-induced currents in NMDA or AMPA receptors (Figure S5). Rather, CN-105’s attenuation of the spontaneous EPSC (sEPSC) frequency and the reversion by the cholinesterase inhibitor donepezil (Figure 3A), suggested an intercellular and most likely presynaptic action of CN-105. Confining the agonistic effect of donepezil to nAChR by adding the muscarinic AChR inhibitor atropine exhibited a similar effect to adding donepezil alone, indicative of an nAChR-dependent mechanism. When tetrodotoxin was added to block all AP firings, CN-105 was no longer able to affect the NMDA receptor-mediated EPSC (Figure 3B). These results suggested that CN-105 could suppress the presynaptic release of glutamate in an AP- and AChR-dependent manner. That is, nAChR⁺ glutamatergic neurons may be the primary cellular target of CN-105 in suppressing excitatory neurotransmission in the brain.

In a rat model of transient ischemic stroke, CN-105 (0.1 to 0.4 mg/kg) significantly reduced the infarct volume, similar to the group given the free-radical scavenger edaravone (Figure 4A). At the infarct margin of the primary somatosensory cortex, we enumerated the α7-nAChR⁺ cells which had higher density in the CN-105 treated groups than in the vehicle-treated groups (Figure 4B). The number of β3-GABA_AR⁺ cells in this area was unaffected by CN-105 (Figure 4C). Similar to the trend of α7-nAChR⁺ cells, the overall density of viable neurons (as NeuN⁺ cells) was higher in CN-105 treated groups (Figure 4D), which suggested that specific interactions between CN-105 and neurons carrying α7-nAChRs in the somatosensory cortex may help protect neuronal tissues from the ischemic-reperfusion injury. Our current understanding of CN-105’s neuronal mechanism of action is depicted in Figure 4E. Although chronic inhibition of the cholinergic pathway may be detrimental, our results demonstrated that acute and selective antagonism of nAChR may actually protect the brain from excitotoxicity.

In conclusion, we demonstrated that one of the neuroprotective mechanisms of CN-105 was the dampening of presynaptic glutamate release via nAChR inhibition, arising from a unique electrostatic gating effect on the cation channel. Our current observations emphasize that, in addition to its effects on glia, the direct interaction between apoE (and the peptide derivatives) and neuronal ion channels may play an important role in mediating its neuroprotective effects. Our results also suggest that it may be worthwhile to reconsider the role of nAChR as a potential therapeutic target for stroke neuroprotection.

亲爱的编辑：

这封信描述了我们确定临床阶段中风治疗 CN-105 的神经元靶点的工作，并提出了一种涉及 nAChR 拮抗剂的新型神经保护策略。中风是一种具有高发病率和死亡率的毁灭性疾病。CN-105 最初旨在模拟内源性载脂蛋白 E (apoE) 的抗炎活性。尽管它在各种脑损伤动物模型中被证明有效，并且在临床试验中得到了公认的安全性，^{1, 2}我们对 CN-105的作用机制的理解仍然不完整。早期报告表明，apoE 衍生肽和许多寡精氨酸物种可能直接与各种神经元靶标相互作用，包括烟碱型乙酰胆碱受体 (nAChR)。^3、4关于 nAChR 的长期观点是其激活具有神经保护作用，最好的例子是 nAChR 激动剂或正变构调节剂的认知增强作用。⁵在中风中受到有限关注的 nAChR 的一个常见作用是增强突触前末端的谷氨酸释放。⁶谷氨酸能神经传递可以说是中风兴奋性毒性传播的主要原因。^{7, 8}尽管由 Ach 或尼古丁激活的突触后 nAChR 可以使近端 NMDA 受体脱敏⁹ ，但我们假设 nAChR 拮抗剂对突触前神经元的急性作用可能有助于下调与谷氨酸兴奋性毒性相关的有害级联反应。

在给予比临床实践高几个数量级的 CN-105 超治疗剂量后，食蟹猴表现出瞳孔散大和上睑下垂的症状。逐渐恢复到正常活动需要大约两个 CN-105 的血清清除半衰期（4 小时，表 S1；图 S1）。在 C57BL6 小鼠中，CN-105 的 LD ₈₀为 25 mg/kg（图 1A），具有痉挛和呼吸抑制症状。插管和机械通气后，50 mg/kg 的剂量与死亡率无关（图 1A），但在去除异氟醚麻醉后会延长镇静时间（电影 S1）。

鉴于 apoE 肽对 nAChR 的拮抗活性，^{3, 10}我们假设 CN-105 的呼吸抑制是由于其抑制了各种呼吸控制途径中存在的 nAChR。使用过表达 α7-nAChR 的 HEK293 细胞系进行的初步电生理研究表明，10 μM C​​N-105 强烈抑制 ACh 感应电流，类似于 apoE 的抑制作用（图 S2）。在 10 nM 至 10 μM C​​N-105 存在下，我们观察到 ACh 诱导电流的典型拮抗抑制（图 1B）。这种效应是亚型特异性的，因为 CN-105 的 IC ₅₀与 α7-、α4β2- 和 α1β1δε-nAChR 相差两个数量级（图 1C），后者可能是 CN-105的原因在超治疗剂量下的呼吸毒性。CN-105 的电压依赖性与 apoE130-149 并行测量（图 1D），其中我们观察到最大抑制在 –80 mV。在大鼠神经元的原代培养中，1 mM ACh 引发强烈的钙流入，在 ACh 输注之前与 CN-105 的先验孵育显着减少了钙流入（图 1E）。这些结果表明CN-105可以阻断ACh诱导的Ca ²⁺流入和动作电位(AP)的传播。

根据报道的 nAChR 上 apoE140-148 肽的结合位点，¹⁰我们构建了 α7 nAChR-CN-105 复合物的模型。CN-105 的推定结合位点靠近 ACh 的正构结合位点（图 S3A），并且很可能会阻断 ACh 的结合。不出所料，CN-105 的精氨酸侧链参与了广泛的氢键（图 S3B）。复杂结构的泊松-玻尔兹曼分析（图 2A）表明，CN-105 的正电荷显着削弱了对阳离子通过至关重要的负电场，从而降低了阳离子流入的倾向。前两个中性残基对 nAChR 活性至关重要（图 S3B）。尽管具有额外的正电荷，其中第一个缬氨酸被精氨酸取代的肽失去了大部分的 nAChR 活性（表 S2）。

我们接下来研究了 CN-105 对 nAChR 和 nAChR 介导的钙内流的抑制是否会影响谷氨酸的下游释放。在大鼠海马脑切片中，由 NMDA 受体和 AMPA 受体介导的诱发兴奋性突触后电流 (eEPSC) 在 CN-105 存在下显着低于对照水平（图 2B，图 S4）。CN-105 不直接抑制 NMDA 或 AMPA 受体中谷氨酸诱导的电流（图 S5）。相反，CN-105 '自发 EPSC (sEPSC) 频率的衰减和胆碱酯酶抑制剂多奈哌齐的逆转（图 3A）表明 CN-105 的细胞间和最可能的突触前作用。通过添加毒蕈碱型 AChR 抑制剂阿托品来限制多奈哌齐对 nAChR 的激动作用表现出与单独添加多奈哌齐相似的作用，表明存在 nAChR 依赖性机制。当添加河豚毒素以阻止所有 AP 发射时，CN-105 不再能够影响 NMDA 受体介导的 EPSC（图 3B）。这些结果表明，CN-105 可以以 AP 和 AChR 依赖性方式抑制谷氨酸的突触前释放。也就是说，nAChR ⁺谷氨酸能神经元可能是 CN-105 抑制大脑中兴奋性神经传递的主要细胞靶点。

在大鼠短暂性缺血性中风模型中，CN-105（0.1 至 0.4 mg/kg）显着减少了梗死体积，类似于给予自由基清除剂依达拉奉的组（图 4A）。在初级体感皮层的梗塞边缘，我们列举了CN-105 治疗组比载体治疗组密度更高的 α7-nAChR ^{+细胞（图 4B）。}该区域的 β3-GABA _A R ⁺细胞数量不受 CN-105 的影响（图 4C）。与 α7-nAChR ⁺细胞的趋势相似，活神经元的总密度（如 NeuN ⁺细胞）在 CN-105 治疗组中更高（图 4D），这表明 CN-105 和体感皮层中携带 α7-nAChR 的神经元之间的特定相互作用可能有助于保护神经元组织免受缺血再灌注损伤。我们目前对 CN-105的神经元作用机制的理解如图 4E 所示。虽然胆碱能通路的慢性抑制可能是有害的，但我们的研究结果表明，nAChR 的急性和选择性拮抗实际上可以保护大脑免受兴奋性毒性的影响。

总之，我们证明 CN-105 的神经保护机制之一是通过 nAChR 抑制抑制突触前谷氨酸的释放，这是由于对阳离子通道的独特静电门控效应引起的。我们目前的观察强调，除了对神经胶质细胞的影响外，apoE（和肽衍生物）与神经元离子通道之间的直接相互作用可能在介导其神经保护作用中起重要作用。我们的研究结果还表明，重新考虑 nAChR 作为中风神经保护的潜在治疗靶点的作用可能是值得的。