It is a widely believed paradigm that intimal angiogenesis in coronary atherosclerosis contributes to plaque instability through both hemorrhagic plaque expansion and cholesterol deposition.¹ However, this article supports an alternative concept in which angiogenesis is actually essential for plaque stabilization and healing.^2,3 The prevailing paradigm has its origins primarily in histopathology work from the 1980s by Barger’s group reported by Kamat et al.⁴ We argue that misinterpretation of their results has skewed the understanding of vulnerable plaque for >35 years. This misconception still endures in recent plaque angiogenesis reviews, affecting understanding of acute coronary syndromes (ACS) pathophysiology.¹

Most coronary plaque ruptures/erosions do not lead to ACS. ACS research focuses primarily on vessel breakdown through mechanisms such as inflammation. We have instead proposed that vessel repair mechanisms that oppose instability fail, leading to ACS.³ ACS is then a double hit of rupture/erosion and failed vascular healing. Here, we assert that intimal angiogenesis, needed to maintain cellular reparative functions (ie, healing), if inadequate, leads to coronary plaque instability. Angiogenesis is the supply line to reparative cells in healing regions. In plaques with long necrotic cores, this supply line to compromised regions may not be reliably established.

In their 1984 report, Barger’s group injected silicone into human coronary arteries in vitro and filmed the flow into mural microvessels.⁴ They described a primarily short centripetal angiogenesis pattern from the vasa vasorum to the intima. Their discussion linked angiogenesis to plaque hemorrhage with rapid plaque expansion as a trigger for ACS. Our data and literature analysis, including re-examination of the results of Barger and colleagues, do not support primarily centripetal angiogenesis. Instead, we find axial angiogenesis extending many millimeters yet perhaps remaining insufficient to reach compromised intima.

We acknowledge that leaky immature microvessels lead to small intimal hemorrhages that contribute red blood cell cholesterol.² However, hemorrhage is not a significant source of coronary plaque rapid expansion.² This is unlike human carotid arteries or most animal arteries, often used as models in intimal angiogenesis research, in which hemorrhagic plaque expansion is common.¹ We instead propose that coronary angiogenesis is a critical stabilizing factor. Our conclusions are derived from an examination of human coronary plaque in the long axis rather than in conventional cross section, as well as re-evaluating published data, including the work by Barger and colleagues. These conclusions are also supported by data that angiogenesis inhibitors (eg, sunitinib and sorafenib) increase, not decrease, the risk of vascular occlusion in humans.

Examining plaques axially rather than in conventional cross section was critical in our questioning of the current paradigms. The most common morphology found after ACS is thin-walled plaques with a large necrotic core by cross section.² However, only a minority of these plaques (<20%) progress to ACS. Most heal. This brings into question whether these plaques should be called vulnerable plaques. Since 2014, our data have supported the novel concept that ACS risk is a function of necrotic core axial extent.^3,5 Three recent in vivo optical coherence tomography studies, which we reviewed elsewhere, have supported this mechanism.³ Studying long axial necrotic cores, we observed that intimal microvessels tracked long distances predominately in the luminal intima above the core on a longitudinal rather than centripetal course. A representative long axial plaque is shown in the Figure. The data supports the model that, in the presence of long cores, these immature microvessels track many millimeters in the intima above the core parallel to the lumen. These findings are consistent with our re-examination of the 1980s data in the next paragraph. The cellularity needed for promoting healing requires maintaining this angiogenesis, which is challenging with long cores.

Figure. Human coronary artery axial histopathology showing a long necrotic core (NC); microvessels become sparse in a thin intimal cap. A, Coronary artery axial section (Masson trichrome stain) with intima (I), media (M), and adventitia (A). The long NC extends beyond the image and approaches the right luminal surface (thin intimal cap). Microvessels course longitudinally over the core with decreasing density as the intima thins, seen in magnifications of boxes B and C. B(1) and C(1), Endothelium fluorescently stained for von Willebrand factor (red/orange). B(2) and C(2), Contrast-enhanced endothelium appears white within vessel lumens (yellow arrows) by converting the red channel to gray scale. Intima thickness diminishes from B(1,2) across C(1,2). Microvessel density is highest in the thicker cap of B(1,2) and tapers across C(1,2) until absent (no yellow arrows) where the intima is thinnest (<200 μm). We assert that angiogenesis occurs axially over long NCs and is essential for intimal cap healing. Inadequate microvessel extension over long axial NCs, as shown here, confers vulnerability. Thin-capped fibroatheroma NCs are defined only in cross-section (D). This may explain why <20% lead to acute coronary syndromes; most may have limited axial extent and therefore maintain adequate angiogenesis to heal. Concept illustrated in drawings of NC in cross-section (E) and long axis (F). Artist was Oran Suta.

The angiogenesis patterns we observed are in conflict with currently held paradigms and initially seemed to be in conflict with the results of the Barger group, leading us to re-examine their data.⁴ In their second article, the authors quantified regional microvessel densities from the coronary cinematography described in their first article. They published a table of microvessel densities for the adventitia, inner media, and intima from cross sections of diseased human coronaries. The average microvessel densities (not paired data) were 9.8±1.3 and 10.3±4.6 in the adventitia and intima, respectively. However, the inner media vessel density was only 2.2±0.7 and was not explained in the article. This is inconsistent with predominately centripetal penetration. Microvessel densities should correlate across artery wall layers if penetration is centripetal with microvessels growing directly from adventitia through media to intima. We found no significant paired correlation between the microvessel densities of the intima and inner media (paired t test, P<0.0009), nor was there a significant correlation between the adventitia and intima microvessel densities for each vessel (Pearson paired correlation r=0.25, P=0.24).

In arteries with significant plaque, all data are consistent with intimal microvessel propagation in a predominately longitudinal course. Our efforts to obtain the original silicone injection films were unsuccessful. On close examination of their still images (see Figure 1 in Reference 4), silicone is rarely seen penetrating directly from the adventitia into the luminal intima.⁴ These uncommon centripetal penetrations connect to microvessels coursing above the core, parallel to and near the lumen. These were not the conclusions drawn from their article. Re-examination of their data is instead consistent with the longitudinal intimal angiogenesis shown in our analysis of plaques axially.

Intact healing mechanisms, dependent on the microvasculature, are critical for preventing intimal thinning, plaque rupture/erosion, and ACS.³ This includes maintaining the presence and function of noncontractile, synthetic (-ACTA2) smooth muscle cells.^2,3 We have shown that longitudinal angiogenesis occurs over long necrotic cores and fails to reach some areas in the intimal cap (Figure). We propose that these areas are at risk for thinning, rupture/erosion, and ACS as a result of impaired healing mechanisms. In this model, ACS requires a double hit of both rupture/erosion and impaired vascular healing from inadequate angiogenesis. Clinically, this model supports moving away from angiogenesis inhibition to a strategy of augmenting healing.

The authors thank Peter Caradonna for his technical support in histopathology generation.

Histopathology was funded by National Institutes of Health R01 HL55686 (Dr Brezinski), National Institutes of Health R01 EB02638/HL63953 (Dr Brezinski), and a University of New England Office of Scholarship and Research grant (Drs Brezinski and Willard).

None.

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.

https://www.ahajournals.org/journal/circ

人们普遍认为，冠状动脉粥样硬化中的内膜血管生成通过出血性斑块扩张和胆固醇沉积导致斑块不稳定。¹然而，本文支持另一种概念，即血管生成实际上对于斑块稳定和愈合至关重要。^2,3流行的范式主要起源于 Kamat 等人报道的 Barger 小组 20 世纪 80 年代的组织病理学工作。⁴我们认为，超过 35 年来，对其结果的误解已经扭曲了对易损斑块的理解。这种误解在最近的斑块血管生成综述中仍然存在，影响了对急性冠状动脉综合征（ACS）病理生理学的理解。¹

大多数冠状动脉斑块破裂/侵蚀不会导致 ACS。ACS 研究主要关注炎症等机制引起的血管破裂。相反，我们提出对抗不稳定的血管修复机制失败，导致 ACS。³ ACS 是破裂/侵蚀和血管愈合失败的双重打击。在这里，我们断言维持细胞修复功能（即愈合）所需的内膜血管生成如果不充分，就会导致冠状动脉斑块不稳定。血管生成是愈合区域修复细胞的供应线。在具有长坏死核心的斑块中，可能无法可靠地建立通往受损区域的供应线。

在 1984 年的报告中，Barger 的研究小组将硅胶注射到体外人体冠状动脉中，并拍摄了流入壁微血管的情况。⁴他们描述了从血管滋养管到内膜的主要短向心血管生成模式。他们的讨论将血管生成与斑块出血联系起来，斑块快速扩张是 ACS 的触发因素。我们的数据和文献分析，包括对 Barger 及其同事的结果的重新审查，主要不支持向心性血管生成。相反，我们发现轴向血管生成延伸了许多毫米，但可能仍不足以到达受损的内膜。

我们承认，未成熟的微血管渗漏会导致小内膜出血，从而导致红细胞胆固醇升高。²然而，出血并不是冠状动脉斑块快速扩张的重要原因。²这与人类颈动脉或大多数动物动脉不同，后者通常用作内膜血管生成研究的模型，其中出血斑块扩张很常见。¹相反，我们认为冠状血管生成是一个关键的稳定因素。我们的结论源自对长轴而非传统横截面的人类冠状动脉斑块的检查，以及重新评估已发表的数据，包括 Barger 及其同事的工作。这些结论也得到了血管生成抑制剂（例如舒尼替尼和索拉非尼）增加而不是减少人类血管闭塞风险的数据的支持。

在我们对当前范例的质疑中，轴向检查斑块而不是传统的横截面至关重要。ACS 后最常见的形态是薄壁斑块，横截面有大的坏死核心。²然而，这些斑块中只有少数 (<20%) 进展为 ACS。大多数痊愈。这就提出了这些斑块是否应该被称为易损斑块的问题。自 2014 年以来，我们的数据支持了这样一个新概念：ACS 风险是坏死核心轴向范围的函数。^3,5最近的三项体内光学相干断层扫描研究（我们在其他地方回顾过）支持了这一机制。³通过研究长轴坏死核心，我们观察到内膜微血管主要在核心上方的管腔内膜中沿纵向而不是向心轨迹长距离移动。图中显示了具有代表性的长轴向斑块。数据支持这样的模型：在存在长核心的情况下，这些未成熟的微血管在平行于管腔的核心上方的内膜中追踪许多毫米。这些发现与我们在下一段中重新审视 20 世纪 80 年代的数据是一致的。促进愈合所需的细胞结构需要维持这种血管生成，这对于长核心来说是一个挑战。

数字。 人冠状动脉轴向组织病理学显示长坏死核心（NC）；薄薄的内膜帽中的微血管变得稀疏。A，冠状动脉轴切面（马森三色染色），内膜 (I)、中膜 (M) 和外膜 (A)。长 NC 延伸超出图像并接近右侧管腔表面（薄内膜帽）。随着内膜变薄，微血管在核心上纵向延伸，密度逐渐降低，如框B和C的放大倍数所示。B(1)和C(1)，内皮细胞对冯维勒布兰德因子进行荧光染色（红色/橙色）。B(2)和C(2)，通过将红色通道转换为灰度，对比度增强的内皮在血管腔内呈现白色（黄色箭头）。内膜厚度从B(1,2)到C(1,2) 逐渐减小。微血管密度在B(1,2)较厚的帽中最高，并在C(1,2)上逐渐减小，直到内膜最薄 (<200 μm) 处消失（无黄色箭头）。我们断言，血管生成在长 NC 上轴向发生，对于内膜帽愈合至关重要。如此处所示，长轴 NC 上的微血管延伸不足会导致脆弱性。薄顶纤维粥样斑块 NC 仅在横截面 ( D )中定义。这或许可以解释为什么<20%会导致急性冠状动脉综合征；大多数可能具有有限的轴向范围，因此维持足够的血管生成以愈合。NC 图纸中横截面 ( E ) 和长轴 ( F ) 所示的概念。艺术家是奥兰·苏塔。

我们观察到的血管生成模式与当前持有的范式相冲突，并且最初似乎与 Barger 小组的结果相冲突，导致我们重新检查他们的数据。⁴在第二篇文章中，作者根据第一篇文章中描述的冠状动脉摄影术量化了区域微血管密度。他们发布了患病人类冠状动脉横截面的外膜、内中膜和内膜的微血管密度表。外膜和内膜的平均微血管密度（未配对数据）分别为 9.8±1.3 和 10.3±4.6。但其内部介质容器密度仅为2.2±0.7，文中未作解释。这与主要向心渗透不一致。如果渗透是向心的，微血管直接从外膜穿过中膜生长到内膜，则微血管密度应该与动脉壁层相关。我们发现内膜和内中膜的微血管密度之间不存在显着的配对相关性（配对t检验，P <0.0009），每个血管的外膜和内膜微血管密度之间也不存在显着的相关性（Pearson配对相关性r = 0.25，P =0.24）。

在具有明显斑块的动脉中，所有数据都与内膜微血管在主要纵向过程中的传播一致。我们试图获得原始的有机硅注射薄膜，但没有成功。仔细检查其静止图像（参见参考文献 4 中的图 1），很少看到有机硅直接从外膜渗透到管腔内膜。⁴这些不常见的向心穿透连接到在核心上方流动的微血管，平行于内腔并靠近内腔。这些不是从他们的文章中得出的结论。相反，重新检查他们的数据与我们对斑块的轴向分析中显示的纵向内膜血管生成一致。

依赖于微血管系统的完整愈合机制对于预防内膜变薄、斑块破裂/侵蚀和 ACS 至关重要。³这包括维持非收缩性合成 (-ACTA2) 平滑肌细胞的存在和功能。^2,3我们已经表明，纵向血管生成发生在长坏死核心上，并且无法到达内膜帽的某些区域（图）。我们认为，由于愈合机制受损，这些区域面临变薄、破裂/侵蚀和 ACS 的风险。在该模型中，ACS 需要破裂/侵蚀和血管生成不足导致血管愈合受损的双重打击。在临床上，该模型支持从抑制血管生成转向增强愈合的策略。

作者感谢 Peter Caradonna 在组织病理学生成方面的技术支持。

组织病理学由国立卫生研究院 R01 HL55686（布热津斯基博士）、国立卫生研究院 R01 EB02638/HL63953（布热津斯基博士）以及新英格兰大学奖学金和研究资助办公室（布热津斯基博士和威拉德博士）资助。

没有任何。

本文表达的观点不一定代表编辑或美国心脏协会的观点。

https://www.ahajournals.org/journal/circ