Department of Optometry and Vision Science, University of Auckland, Auckland, New Zealand.
Exp Neurol. 2013 Dec;250:74-93. doi: 10.1016/j.expneurol.2013.09.007. Epub 2013 Sep 13.
Pattern recognition has been used for the complete and statistically rigid classification of retinal neurons in vertebrates such as the adult cat, primate, rat and goldfish. Here, we label the mouse retina with antibodies against seven amino acids and use pattern recognition to characterize distinct retinal neurochemical cell classes based on their unique amino acid signatures. We followed the development of the cell classes in the X-inactivation transgenic mouse expressing the lacZ reporter gene on one X-chromosome. This mouse allows clonally related cells to be identified through differential β-galactosidase activity due to random X-chromosome inactivation. Pattern recognition analysis partitioned the retina into nine neuronal classes at birth, increasing to 19 classes at eye opening and 26 classes by adulthood. Emergence of new cell classes was partly attributed to new neuron types and partly to the splitting of classes from early ages from refinement of their amino acid profiles. All six GABAergic amacrine cell classes and most ganglion cell classes appeared by P7 whilst all the glycinergic amacrine cell classes did not appear till adulthood. Separable bipolar cell classes were not detected till eye opening. Photoreceptor cell classes were detected at P3 but inner and outer segments did not form separable classes until adulthood. More importantly, we show that cells which share common amino acid profiles also shared cell dispersion patterns. GABAergic amacrine cell classes with conventional and displaced counterparts transgressed clonal boundaries whereas GABAergic amacrine cell classes found exclusively in the inner nuclear layer and all glycinergic amacrine cell classes did not transgress. Ganglion cells displayed both dispersion patterns. This study provides a comprehensive neurochemical atlas of the developing mouse retina, tracking the amino acid levels within distinct neuronal populations and highlighting unique migratory patterns within subpopulations of inner retinal neurons.
模式识别已被用于对脊椎动物（如成年猫、灵长类动物、大鼠和金鱼）的视网膜神经元进行完整且严格的分类。在这里，我们使用针对七种氨基酸的抗体对小鼠视网膜进行标记，并使用模式识别根据其独特的氨基酸特征来描述不同的视网膜神经化学细胞类群。我们跟踪了在表达 lacZ 报告基因的 X 失活转基因小鼠中细胞类群的发育，该基因在一条 X 染色体上表达。这种小鼠允许通过差异β-半乳糖苷酶活性识别克隆相关细胞，因为 X 染色体随机失活。模式识别分析将出生时的视网膜分为 9 个神经元类群，在睁眼时增加到 19 个类群，成年时增加到 26 个类群。新细胞类群的出现部分归因于新的神经元类型，部分归因于由于其氨基酸谱的细化而从早期开始将类群分裂。所有 6 种 GABA 能无长突细胞类群和大多数神经节细胞类群在 P7 时出现，而所有甘氨酸能无长突细胞类群直到成年时才出现。直到睁眼时才检测到可分离的双极细胞类群。在 P3 时检测到光感受器细胞类群，但内节和外节直到成年时才形成可分离的类群。更重要的是，我们表明具有共同氨基酸特征的细胞也具有细胞分散模式。具有传统和置换对应物的 GABA 能无长突细胞类群跨越了克隆边界，而仅存在于内核层中的 GABA 能无长突细胞类群和所有甘氨酸能无长突细胞类群则没有跨越。神经节细胞显示出两种分散模式。这项研究提供了发育中老鼠视网膜的全面神经化学图谱，跟踪了不同神经元群体中的氨基酸水平，并突出了内视网膜神经元亚群中的独特迁移模式。
