Cytokines are proteins produced by cells of the immune system. Unusually high levels can trigger systemic pathological inflammation, referred to as a cytokine storm. The inference is that this results from a correct but over‐reactive immune response. Pathological changes consistent with cytokine storms have been observed with SARS‐CoV‐2–induced COVID‐19 disease, particularly in hospitalized patients with the severe forms of the disease. Therefore, it has been proposed that, while the virus makes people sick, what puts them in the hospital—and potentially kills them, is not the virus itself, but rather an over‐reactive immune response to the virus.^{1, 2}

In our view, the problem is not an appropriate, yet overzealous immune response. Instead, we propose that the virus incites a ‘wrong’ immune response while simultaneously suppressing a more appropriate one. Furthermore, these distinct 'appropriate' and 'incorrect' responses tilt immune balance in opposite directions.

An important survival strategy is to subvert immune attack. Specifically, it would be advantageous for the virus to suppress the most relevant T helper pathways. A large body of evidence indicates that the T helper 1 (Th1) and Th2 pathways are especially vital to clear most viral infections. Th2‐mediated robust antibody and Th1‐mediated cytotoxic T cell responses are needed to bind and neutralize the millions of viral particles produced during the acute phase of infection and to destroy infected cells used by the virus to produce myriad copies of itself. SARS‐CoV‐2³ and other related viruses⁴ appear to do so by antagonizing the production and signalling of type 1 interferons (T1IFNs), which are essential initiators and amplifiers of Th1 immune responses.⁵

Recent analyses indicate bats as the primary reservoir for the SARS‐CoV‐2 lineage and pangolins may have been an intermediate host to facilitate transmission to humans. However, it is at least as likely that the virus evolved in bats the ability to replicate in humans, perhaps decades ago.⁶

T1IFNs are not only critical to drive Th1 cytotoxic T cell pathways, but they also play an important role in attenuating Th17 pathways, which are responsible for the recruitment of phagocytic cells. Indeed, pathways leading to Th1 and Th17 responses are mutually antagonistic. So, by targeting the T1IFN pathway, SARS‐CoV‐2 negates patients’ capacity to mount an effective Th1 response, while simultaneously enhancing a Th17 response. This scenario would predict that if a cytokine storm developed in an acutely ill patient, it would favour activation of the Th17 pathway, that which is far less likely to aid in clearing the virus. Not surprisingly that is what is found. SARS‐CoV‐2–associated cytokine storms usually consist of an overabundance of cytokines such as IL‐6,^{7, 8} which is highly associated with the Th17 pathway.

Mechanistically, the bias towards Th17 would be driven by two independent responses. The first is by altering production of interleukin‐12 (IL‐12) but not interleukin‐23 (IL‐23), which, respectively, are key drivers of Th1 and Th17 responses. T1IFNs are critical for the induction of IL‐12,^{5, 9} and they also inhibit conversion of interleukin‐1 (IL‐1) into IL‐1β, another key cytokine in the Th17 pathway.⁵ In this circumstance, since T1IFNs suppress Th17–associated IL‐1 conversion to its active form, the immune system would tilt towards a vigorous Th17 and against a Th1 response.

Interference with IL‐12 may also contribute to the loss of Bcl‐6–expressing T follicular helper cells leading to the absence of germinal centres in some COVID‐19 patients.^{9, 10} This absence impairs antibody production and potentially negates B cell memory responses. Thus, the consequence of targeting T1IFN production is that the virus incites a not‐so‐helpful Th17 response and avoids the highly relevant Th1 and Th2 responses.

So, if the cytokine storm is not solely a matter of immune over‐responsiveness but rather one where the immune response is ‘barking up the wrong tree’ (in this case the Th17 rather than the Th1 tree), therapies that attenuate Th17 responses and possibly enhance Th1 responses might prevent or mitigate the cytokine storm syndrome. This could be accomplished by targeting the IL‐6 pathway and/or by using T1IFNs themselves (specifically INF‐β). Several monoclonal antibodies targeting IL‐6 and its receptors are currently in clinical trials for COVID‐19 patients, as are INF‐βs.

A corollary to this discussion is that focusing the immune response towards the Th1 pathway should lead to better outcomes, whereas amplifying the Th17 pathway would be a negative prognostic indicator of outcome. This could contribute to the greater susceptibility in older people infected with SARS‐CoV‐2, as this population has a greater tendency to polarize their immune response towards the Th17 pathway. A Th17 vs Th1 bias, however, could also explain the heterogeneous susceptibility of younger patients, and even children, to develop severe disease in the face of SARS‐CoV‐2 infection. Conversely, a Th1 bias might identify those patients most amenable to successful outcomes with targeted immunosuppressive approaches. These concepts are already in practice in the field of autoimmunity where there are strong associations between Th17‐biased immune responses and various autoimmune conditions, and where Th17 antagonists have transformed the outcomes for these patients.

In summary, we propose that the cytokine storm observed with COVID‐19 and severe symptomatology is incited by a Th17‐biased immune response, which just so happens to be a less productive pathway to clear the virus. The short‐term survival of SARS‐CoV‐2 would be favoured if the virus could force the immune system towards this incorrect response by inhibiting T1IFN pathways and thwarting the ‘correct’ and relevant Th1 pathway. Therefore, what is being referred to as a cytokine storm may really be more of a cytokine diversion. While this may seem like splitting hairs, it has important implications for therapy. It follows that therapies should not throw out the baby (Th1) with the bathwater (Th17).

细胞因子是免疫系统细胞产生的蛋白质。异常高的水平会引发全身性病理性炎症，称为细胞因子风暴。推论是，这是由正确但过度反应的免疫反应引起的。在 SARS-CoV-2 诱导的 COVID-19 疾病中观察到与细胞因子风暴一致的病理变化，特别是在患有严重疾病的住院患者中。因此，有人提出，虽然病毒使人生病，但将他们送进医院并可能杀死他们的不是病毒本身，而是对病毒的过度反应性免疫反应。^{1, 2}

在我们看来，问题不是适当的，但过度热情的免疫反应。相反，我们建议该病毒会引发一种“错误的”免疫反应，同时抑制一种更合适的免疫反应。此外，这些不同的“适当”和“不正确”反应使免疫平衡向相反方向倾斜。

一个重要的生存策略是颠覆免疫攻击。具体来说，病毒抑制最相关的 T 辅助途径将是有利的。大量证据表明，T 辅助 1 (Th1) 和 Th2 通路对于清除大多数病毒感染尤其重要。需要 Th2 介导的强效抗体和 Th1 介导的细胞毒性 T 细胞反应来结合和中和在感染急性期产生的数百万个病毒颗粒，并破坏病毒用来产生无数自身拷贝的受感染细胞。SARS-CoV-2 ³和其他相关病毒⁴似乎是通过拮抗 1 型干扰素 (T1IFNs) 的产生和信号传导来实现的，T1IFNs 是 Th1 免疫反应的重要发起者和放大器。⁵

最近的分析表明，蝙蝠是 SARS-CoV-2 谱系的主要宿主，穿山甲可能是促进传播给人类的中间宿主。然而，这种病毒在蝙蝠身上进化出在人类身上复制能力的可能性至少与几十年前一样。⁶

T1IFN 不仅对驱动 Th1 细胞毒性 T 细胞通路至关重要，而且在减弱负责吞噬细胞募集的 Th17 通路中也发挥着重要作用。事实上，导致 Th1 和 Th17 反应的途径是相互拮抗的。因此，通过靶向 T1IFN 通路，SARS-CoV-2 否定了患者产生有效 Th1 反应的能力，同时增强了 Th17 反应。这种情况可以预测，如果在重症患者中出现细胞因子风暴，它将有利于激活 Th17 通路，而这不太可能有助于清除病毒。毫不奇怪，这是发现的。SARS-CoV-2 相关的细胞因子风暴通常由过多的细胞因子组成，例如 IL- ^6、7、8与 Th17 通路高度相关。

从机制上讲，对 Th17 的偏见将由两个独立的反应驱动。第一个是通过改变白细胞介素-12 (IL-12) 而不是白细胞介素-23 (IL-23) 的产生，白介素-23 (IL-23) 分别是 Th1 和 Th17 反应的关键驱动因素。T1IFNs 对于诱导 IL-12、5、9 至关重要，^它们还抑制白细胞介素-1 (IL-1) 转化为 IL-1β，IL-1β 是 Th17 通路中的另一种关键细胞因子。⁵在这种情况下，由于 T1IFN 抑制 Th17 相关的 IL-1 转化为其活性形式，免疫系统将倾向于强大的 Th17 并对抗 Th1 反应。

干扰 IL-12 也可能导致表达 Bcl-6 的滤泡辅助性 T 细胞丢失，导致一些 COVID-19 患者缺乏生发中心。^{9, 10}这种缺失会损害抗体的产生并可能否定 B 细胞的记忆反应。因此，靶向 T1IFN 产生的结果是病毒引发了不太有用的 Th17 反应，并避免了高度相关的 Th1 和 Th2 反应。

因此，如果细胞因子风暴不仅仅是免疫过度反应的问题，而是免疫反应“错误的树”（在这种情况下是 Th17 而不是 Th1 树），那么减弱 Th17 反应的疗法和可能增强 Th1 反应可能预防或减轻细胞因子风暴综合征。这可以通过靶向 IL-6 通路和/或使用 T1IFN 本身（特别是 INF-β）来实现。几种针对 IL-6 及其受体的单克隆抗体目前正处于 COVID-19 患者的临床试验中，INF-β 也是如此。

该讨论的一个推论是，将免疫反应集中在 Th1 途径应该会导致更好的结果，而放大 Th17 途径将是结果的负面预后指标。这可能会增加感染 SARS-CoV-2 的老年人的易感性，因为该人群更倾向于将免疫反应极化到 Th17 途径。然而，Th17 与 Th1 的偏差也可以解释年轻患者甚至儿童在面对 SARS-CoV-2 感染时发展为严重疾病的异质性。相反，Th1 偏倚可能会识别出那些最容易通过靶向免疫抑制方法获得成功结果的患者。

总之，我们提出在 COVID-19 和严重症状中观察到的细胞因子风暴是由偏向 Th17 的免疫反应引发的，而这恰好是清除病毒的效率较低的途径。如果病毒能够通过抑制 T1IFN 通路并阻止“正确”和相关的 Th1 通路来迫使免疫系统做出这种不正确的反应，那么 SARS-CoV-2 的短期存活将是有利的。因此，所谓的细胞因子风暴实际上可能更像是一种细胞因子转移。虽然这可能看起来像分裂头发，但它对治疗具有重要意义。因此，治疗不应将婴儿 (Th1) 与洗澡水 (Th17) 一起丢弃。