Meet the First Author, see p 706

The perinatal mouse retina is the current model of choice to define the mechanisms of sprouting angiogenesis.¹ After birth, blood vessels grow from the optic nerve head towards the retinal periphery while branching laterally to establish a planar network that is exquisitely suited for quantifying vascular extension and branching.¹ Cre recombinase–mediated ablation of loxP-flanked (floxed) genes is widely used to uncover angiogenesis genes in the mouse retina, with temporal control available through the Cre recombinase estrogen fusion protein with ligand binding mutation (CreER), whose activity is tamoxifen-dependent.^2,3 The correct interpretation of experiments employing CreER-mediated gene ablation requires controls to exclude off-target effects, such as transgene insertion into essential genomic loci or toxic effects of activating CreER. Despite reports of impaired cell proliferation and DNA damage in several cell types after CreER activation,⁴ published work has not addressed whether CreER toxicity affects the endothelial cells that line all blood vessels and drive angiogenesis.

We have compared retinal angiogenesis in littermate perinatal day 7 pups expressing or lacking the endothelial cell–selective Cdh5-CreER^T2 transgene with the Mouse Genome Informatics identifier 5705396.⁵ Mice were injected at perinatal day 2 and 4 with peanut oil or peanut oil containing tamoxifen doses representative of those commonly reported.³ Neither CreER^T2 nor tamoxifen alone impaired vascular network extension or branch density (Figure [A] and [B]). In contrast, and despite the absence of floxed target genes, both vascular parameters were significantly reduced in the retinas of tamoxifen-injected pups expressing CreER^T2 compared with tamoxifen-injected pups lacking CreER^T2 (Figure [A] and [B]; similar defects were seen with the tamoxifen metabolite 4-hydroxytamoxifen; data not shown). Linear regression analysis revealed a significant, negative relationship of both vascular parameters with tamoxifen dose in Cdh5-CreER^T2 mice (Figure [A] and [B]; extension: regression coefficient, −0.00095, R²⁼0.46, P=1.69×10⁻⁵; branch density: regression coefficient, −0.0026, R²⁼0.54, P=1.69×10⁻⁶). Vascular network extension and branch density were also significantly reduced in the retinas of tamoxifen-injected pups carrying the tamoxifen-inducible, endothelial cell–selective Pdgfb-iCreER^T2 transgene (Mouse Genome Informatics identifier 3793852⁵; Figure [C] and [D]). The bodyweight of CreER^T2 mice was not reduced compared to CreER^T2-deficient littermates, excluding that a general developmental delay had impeded angiogenesis (data not shown). We conclude that tamoxifen-activated CreER^T2 compromises retinal angiogenesis. In contrast, Tie2-Cre, which is constitutively expressed in endothelial cells (Mouse Genome Informatics identifier 2450311),⁵ did not impair retinal angiogenesis (Figure [C] and [D]), although this does not exclude toxicity of constitutive Cre activity in other circumstances.

Figure. Endothelial Cre recombinase estrogen fusion protein with ligand binding mutation (CreER)^T2 activity impairs retinal angiogenesis. Flat mounted perinatal day (P) 7 retinas were stained with the vascular endothelial marker IB4 (isolectin B4) and fluorescent streptavidin.³ Scale bars: 500 µm. A and B, Cdh5-CreER^T2–expressing and wildtype littermates were injected at P2 and P4 with 25 µL peanut oil containing 0, 50, 100, or 150 µg tamoxifen. A, Micrographs and (B) quantification of vascular extension and branch density. Dotted boxes indicate areas shown at higher magnification. Red and yellow lines indicate vascular extension and retinal radius. The green box indicates a representative area analyzed for vascular branch density. C and D, Pdgfb-iCreER^T2–expressing and wildtype littermates were injected at P2 and P4 with 25 µL peanut oil containing 100 µg tamoxifen. Tie2-Cre litters were not injected. C, Micrographs and (D) quantification of vascular extension and branch density. Data are presented as mean±SD fold change relative to littermate controls; each data point represents the average of several retinal leaflets. Cdh5-CreER^T2 experiments: controls n=5 (0 µg), n=5 (50 µg), n=10 (100 µg), n=7 (150 µg); CreER^T2 n=5 (0 µg), n=4 (50 µg), n=13 (100 µg), n=9 (150 µg); Pdgfb-iCreER^T2 experiments: controls n=5, CreER^T2 n=7; Tie2-Cre experiments: controls n=5, Tie2-Cre n=3. Two-way ANOVA with Holm-Sidak multiple comparison test, nonsignificant (ns), P>0.05; *P<0.05; **P<0.01, ***P<0.001.

Our findings imply that CreER toxicity has likely confounded the interpretation of some prior retinal angiogenesis studies, including our own, because they have compared tamoxifen-injected CreER-expressing pups to treatment-naive or tamoxifen-injected CreER-negative littermates. As CreER-mediated gene targeting remains a key method for angiogenesis research, we propose that future retinal studies should include tamoxifen-injected CreER control mice lacking floxed genes. Moreover, potential CreER toxicity should be considered when studying angiogenesis in other organs.

Nonstandard Abbreviations and Acronyms

CreER

Cre recombinase estrogen fusion protein with ligand binding mutation

Cre recombinase estrogen fusion protein with ligand binding mutation

This work was supported by the British Heart Foundation (FS/16/61/32740, FS/18/65/34186, PG/19/37/3439) and Wellcome (205099/Z/16/Z). This study used mice on a C57/Bl6J background and was approved by the local Animal Welfare Ethical Review Body and the UK Home Office. The authors declare that data will be made available upon reasonable request.

None.

*J.T.B. and R.L.B. contributed equally to this article.

For Sources of Funding and Disclosures, see page 850.

中文翻译：

---

他莫昔芬激活的 CreERT 独立于基因缺失损害视网膜血管生成。

认识第一作者，见第 706 页

围产期小鼠视网膜是当前定义发芽血管生成机制的首选模型。¹出生后，血管从视神经乳头向视网膜周边生长，同时横向分支形成一个平面网络，非常适合量化血管延伸和分支。¹ Cre 重组酶介导的 loxP 侧翼（floxed）基因消融被广泛用于揭示小鼠视网膜中的血管生成基因，可通过具有配体结合突变的 Cre 重组酶雌激素融合蛋白 (CreER) 进行时间控制，其活性为他莫昔芬-依赖。^2,3使用 CreER 介导的基因消融实验的正确解释需要控制以排除脱靶效应，例如转基因插入基本基因组位点或激活 CreER 的毒性作用。尽管有报道称 CreER 激活后几种细胞类型的细胞增殖和 DNA 损伤受损，但⁴篇已发表的工作并未说明 CreER 毒性是否会影响排列所有血管并驱动血管生成的内皮细胞。

我们比较了同窝围产期第 7 天幼崽的视网膜血管生成，该幼崽表达或缺乏内皮细胞选择性Cdh5-CreER ^T2转基因与小鼠基因组信息学标识符 5705396。5^小鼠在围产期第 2 天和第 4 天注射花生油或含有他莫昔芬的花生油代表通常报告的剂量。³ CreER ^T2和他莫昔芬单独都不会损害血管网络延伸或分支密度（图 [A] 和 [B]）。相比之下，尽管缺乏 floxed 靶基因，与缺乏 CreER T2 的注射他莫昔芬的幼崽相比，注射他莫昔芬的幼崽表达 CreER T2 的视网膜中的两种血管参数均^显着^降低（图 [A] 和 [B]；三苯氧胺代谢物 4-羟基三苯氧胺存在类似缺陷；数据未显示）。线性回归分析显示，在Cdh5-CreER ^T2小鼠中，两种血管参数与他莫昔芬剂量呈显着负相关（图 [A] 和 [B]；扩展：回归系数，-0.00095，R ²⁼ 0.46，P =1.69×10 ^-5；分支密度：回归系数，-0.0026，R ²⁼ 0.54，P =1.69×10 ^-6 )。携带他莫昔芬诱导的内皮细胞选择性Pdgfb-iCreER的注射他莫昔芬幼崽的视网膜中的血管网络扩展和分支密度也显着降低^T2转基因（小鼠基因组信息学标识符 3793852 ⁵；图 [C] 和 [D]）。^{与缺乏 CreER T2}的同窝小鼠相比， ^{CreER T2}小鼠的体重，但一般发育迟缓阻碍了血管生成（数据未显示）。我们得出结论，他莫昔芬激活的 CreER ^T2会损害视网膜血管生成。相比之下， Tie2-Cre在内皮细胞中组成型表达（小鼠基因组信息学标识符 2450311）， ⁵不损害视网膜血管生成（图 [C] 和 [D]），尽管这不排除组成型 Cre 活性在其他情况。

数字。 具有配体结合突变 (CreER) ^T2活性的内皮 Cre 重组酶雌激素融合蛋白会损害视网膜血管生成。平装围产期 (P) 7 视网膜用血管内皮标记 IB4 (isolectin B4) 和荧光链霉亲和素染色。³比例尺：500 µm。A和B，在 P2 和 P4 向表达Cdh5-CreER ^T2的野生型同窝仔猪注射 25 µL 含有 0、50、100 或 150 µg 他莫昔芬的花生油。A、显微照片和（B) 量化血管延伸和分支密度。虚线框表示以更高放大倍率显示的区域。红线和黄线表示血管延伸和视网膜半径。绿色框表示对血管分支密度进行分析的代表性区域。C和D，在 P2 和 P4 向表达Pdgfb-iCreER ^T2的野生型同窝仔猪注射 25 µL 含有 100 µg 他莫昔芬的花生油。没有注射Tie2-Cre窝。C，显微照片和（D）血管延伸和分支密度的量化。数据表示为相对于同窝对照的平均值±SD倍数变化；每个数据点代表几个视网膜小叶的平均值。Cdh5-CreER ^T2实验：对照 n=5 (0 µg)、n=5 (50 µg)、n=10 (100 µg)、n=7 (150 µg)；CreER ^T2 n=5 (0 µg), n=4 (50 µg), n=13 (100 µg), n=9 (150 µg)；Pdgfb-iCreER ^T2实验：对照 n=5，CreER ^T2 n=7；Tie2-Cre实验：对照 n=5，Tie2-Cre n=3。采用 Holm-Sidak 多重比较检验的双向 ANOVA，无显着性 (ns)，P >0.05；* P <0.05；** P <0.01，*** P <0.001。

我们的研究结果表明，CreER 毒性可能混淆了对包括我们自己在内的一些先前视网膜血管生成研究的解释，因为他们将注射他莫昔芬的表达 CreER 的幼崽与未接受治疗或注射他莫昔芬的 CreER 阴性同窝仔猪进行了比较。由于 CreER 介导的基因靶向仍然是血管生成研究的关键方法，我们建议未来的视网膜研究应包括注射他莫昔芬的缺乏 floxed 基因的 CreER 对照小鼠。此外，在研究其他器官的血管生成时，应考虑潜在的 CreER 毒性。

非标准缩写和首字母缩略词

克里尔	Cre重组酶雌激素融合蛋白与配体结合突变

Cre重组酶雌激素融合蛋白与配体结合突变

这项工作得到了英国心脏基金会 (FS/16/61/32740、FS/18/65/34186、PG/19/37/3439) 和 Wellcome (205099/Z/16/Z) 的支持。这项研究使用了 C57/Bl6J 背景的小鼠，并得到了当地动物福利伦理审查机构和英国内政部的批准。作者声明将根据合理要求提供数据。

没有任何。

*JTB 和 RLB 对本文的贡献相同。

有关资金来源和披露信息，请参见第 850 页。