We wish to respond to the opinions stated in the accompanying letter regarding our manuscript, “Deregulated hypoxic response in myeloid cells: A model for high‐altitude pulmonary oedema (HAPE)” by Gojkovic et al.

In essence, Swenson and Bärtsch state that the deletion of VHL in myeloid cells cannot be a model for human HAPE because, in their view, human HAPE is in no way an inflammatory syndrome, whereas the mouse model is the result of an inflammatory change in the pulmonary milieu. Our sense here is that the Swenson and Bärtsch's chief concern is that this model is based on a genetic modification of cells within the immune system.

There are a number of points where we disagree with their stated views; some relate to the nature of animal model systems themselves, and some to the nature of HIF response and inflammation. To take these in turn:

Swenson and Bärtsch point out that a small animal model for HAPE has been anticipated by the field and would be broadly useful. We agree. In order to establish model systems for human disease in animal systems, it is critical to understand the nature of both the human physiology underlying the disease, and both the common and divergent physiologies of the relevant animal.

Mice undergo a rapid and uniform onset of pulmonary hypertension, and a remodelling of both pulmonary and cardiac features when exposed to hypoxia (eg¹). These phenomena of pulmonary hypertension and tissue remodelling are generally reversible when mice are re‐introduced to normoxic environments.

These responses to hypoxia are not the same as those seen in humans, who do not show the same uniform induction of pulmonary hypertension when exposed to environmental hypoxia, and do not undergo a reversible and extensive hypertensive pulmonary and cardiac remodelling following extended hypoxic exposure.

There are thus fundamental differences in the pulmonary responses to environmental hypoxia in mice and humans. Some investigators might argue that this would make any mouse model for human pulmonary responses to hypoxia irrelevant. We feel that this is a mistake, however, since there are many imperfect mouse models for human disease that are nonetheless informative; mouse models for many cancers, for example, are at best highly imperfect mirrors of the human disease, without being irrelevant for understanding the biology of malignancy or for assessing potential treatments.

Our model for HAPE bypasses hypoxia itself, and introduces a deregulated hypoxic response in a normoxic pulmonary setting; this results in a pulmonary oedema that in many respects mirrors that seen in human HAPE. We believe that this alone makes this a relevant model in mice, with the caveat based on what is stated above, that is, that mice and humans respond quite differently to environmental hypoxia itself.

Other concerns of Swenson and Bärtsch include their noting that we see some number of inflammatory cells in broncho‐alveolar lavage in our model, which they point out only occurs late in some cases of HAPE in humans. Human HAPE has a time course extending from the time of first exposure to hypoxia; whereas a genetic model in mice will show effects from the time of the genetic alteration. Thus, adult mice with this deletion necessarily represent a later stage of the disease than would be seen in an initial clinical presentation of HAPE. This does not mean they are not a useful model for HAPE, although it does mean that they do not represent a model for the full extent of clinical progression of the disease. A relatively simple way to address this in the context of our findings would be to introduce an inducible deletion system, for example with a tamoxifen‐ or tetracycline‐inducible cre recombinase, and trigger the loss of VHL at a specific point in an adult mouse, and then, follow progression. We believe this could well allow a model for the overall progression of HAPE, and a model for its’ earliest stages.

Swenson and Bärtsch also appear to draw some form of equivalence between VHL deletion and HIF activation on the one hand, and inflammation on the other. We would point out that loss of VHL does not directly produce inflammation, and loss of HIF does not completely ablate inflammatory capacity. HIF has a much more subtle role in myeloid cells: it regulates metabolism and survival, changes many proliferative and cytokine responses to inflammatory stimuli, and is, in sum, a key aspect of the inflammatory response.² But it is far too simplistic to equate loss of VHL, and thus increased HIF, solely with inflammation,or with the direct induction of an inflammatory state. The loss of VHL in myeloid cells is a molecular surrogate for a deregulated hypoxic response in those cells; a response that would likely occur in a state of environmental hypoxia.

We do not argue here or in our published findings that simply increasing pulmonary inflammation would create a model for HAPE; if Swenson and Bärtsch believe that is what we are stating, and is what occurs in this model, we can understand their objection to our characterizing it as such.

We do believe that myeloid cells can contribute to alterations in physiology that are far more complex than those following the simple triggering of a gross inflammatory response. The physiological alterations that characterize this mouse model are, we believe, quite possibly an underlying factor in human HAPE. Our data indicate that changes in myeloid cell response to hypoxia can affect pulmonary physiology; this is likely true regardless of the presence or absence of myeloid cells in a broncho‐alveolar lavage. We believe a more subtle understanding of myeloid cells will better illuminate the biology of many diseases and dysfunctions; and we feel our published findings indicate that HAPE may well be amongst these.

我们希望回应所附信件中关于Gojkovic等人的论文“骨髓细胞中低氧反应失调：高海拔肺水肿（HAPE）的模型”的意见。

本质上，Swenson和Bärtsch指出，髓样细胞中VHL的缺失不能成为人类HAPE的模型，因为在他们看来，人类HAPE绝不是炎症综合症，而小鼠模型则是人类HPE炎症改变的结果。肺环境。我们的理解是，Swenson和Bärtsch的主要关注点在于该模型基于免疫系统中细胞的遗传修饰。

我们在很多方面不同意他们所说的观点；有些涉及动物模型系统本身的性质，有些涉及HIF反应和炎症的性质。依次进行以下操作：

Swenson和Bärtsch指出，HAPE的小型动物模型是该领域所预期的，将广泛使用。我们同意。为了在动物系统中建立人类疾病的模型系统，至关重要的是要了解该疾病所基于的人类生理学以及相关动物的共同和不同生理学的本质。

小鼠在暴露于缺氧状态时会迅速而均匀地发生肺动脉高压，并同时发生肺和心脏功能的重塑（例如¹）。将小鼠重新引入常氧环境后，这些肺动脉高压和组织重塑现象通常是可逆的。

这些对低氧的反应与在人类中所见的反应不同，在人体暴露于环境低氧时，它们并没有表现出相同的统一诱导性肺动脉高压，并且在长期低氧暴露后也不会经历可逆的，广泛的高血压性肺和心脏重构。

因此，小鼠和人类对环境缺氧的肺反应存在根本差异。一些研究人员可能会争辩说，这将使用于人类肺部对缺氧反应的任何小鼠模型都不相关。但是，我们认为这是一个错误，因为对于人类疾病，有许多不完善的小鼠模型仍然可以提供信息。例如，许多癌症的小鼠模型充其量不过是人类疾病的高度不完善的镜像，与理解恶性生物学或评估潜在治疗方法无关。

我们的HAPE模型绕过了缺氧本身，并在常氧性肺环境中引入了失调的低氧反应。这会导致肺水肿，在许多方面都可与人类HAPE所见相似。我们认为，仅此一项就使它成为小鼠中的一个相关模型，但需要注意的是，基于以上所述，即小鼠和人类对环境缺氧本身的反应截然不同。

Swenson和Bärtsch的其他担忧包括他们注意到我们的模型中支气管肺泡灌洗液中发现了一些炎症细胞，他们指出这种现象仅在人类HAPE的某些情况下才发生。人类HAPE的时程从首次接触到缺氧的时间有所延长。而小鼠的遗传模型将显示出遗传改变后的影响。因此，与HAPE最初的临床表现相比，具有这种缺失的成年小鼠必然代表疾病的晚期。这并不意味着它们不是HAPE的有用模型，尽管这确实意味着它们不代表疾病临床进展的全部范围。根据我们的发现，解决此问题的相对简单的方法是引入可诱导的删除系统，例如，使用他莫昔芬或四环素诱导的cre重组酶，并在成年小鼠的特定点触发VHL的丧失，然后随病情进展。我们认为，这很可能为HAPE的整体发展提供模型，并为其最早阶段提供模型。

Swenson和Bärtsch似乎在一方面使VHL缺失和HIF活化，另一方面与炎症之间产生了某种形式的等价关系。我们将指出，VHL的丧失不会直接产生炎症，而HIF的丧失并不能完全消除炎症能力。HIF在髓样细胞中起着更为微妙的作用：它调节新陈代谢和存活，改变许多对炎症刺激的增殖和细胞因子反应，并且总的来说，它是炎症反应的关键方面。²但是，将VHL的损失等同于仅与炎症或直接诱发炎症状态相比，使VHL丧失，从而使HIF升高太简单了。髓样细胞中VHL的丧失是这些细胞中低氧反应失控的分子替代因子。在环境低氧状态下可能发生的反应。

我们在这里或在我们已发表的研究结果中没有争论仅仅增加肺部炎症会创建HAPE模型。如果Swenson和Bärtsch相信这就是我们所说的，并且是该模型中发生的事情，那么我们可以理解他们对我们如此表征的异议。

我们确实相信，髓细胞可以促进生理变化，而这种变化远比简单触发总的炎症反应复杂得多。我们认为，表征这种小鼠模型的生理变化很可能是人类HAPE的潜在因素。我们的数据表明，髓细胞对缺氧反应的变化会影响肺部生理。无论支气管肺泡灌洗液中是否存在髓样细胞，这都是可能的。我们认为，对髓样细胞的更细微的了解将更好地阐明许多疾病和功能障碍的生物学特性。并且我们认为我们发表的发现表明HAPE很可能在其中。