The dismal prognosis of patients with metastatic melanoma has dramatically improved since the introduction of novel immunotherapies (Donia et al., 2019). However, although some patients experience dramatic and durable tumor regression, more than half of the patients treated with current immunotherapies based on checkpoint blockade do not obtain a sustained benefit and progress within five years, leaving a tremendous unmet clinical need. Myeloid cells, including tumor-associated macrophages (TAMs), have long been suspected of promoting an immunosuppressive and immunotherapy–refractory tumor microenvironment (TME; Pathria et al., 2019). Manipulating the immunosuppressive tumor microenvironment by targeting myeloid cells such as TAMs may be the key to improve the clinical outcome of metastatic melanoma.

Midkine is a heparin-binding growth factor with pleiotropic functions. Over-expression of Midkine has been shown to promote tumor proliferation, apoptosis evasion, angiogenesis, and inflammation in many tumor types (Filippou et al., 2020). Building on their previous results (Olmeda et al., 2017), Cerezo-Wallis et al. showed that secretion of Midkine by melanoma cells boosted an inflamed but immune refractory tumor microenvironment leading to resistance to immune checkpoint blockade. By over-expressing Midkine in human and mouse melanoma cell lines that were grafted in mice, the authors showed that Midkine modulated autocrine and paracrine signals attracting immune-suppressive macrophages. These Midkine-educated macrophages impaired T-cell activity. In contrast, depletion of Midkine in human melanoma cells resulted in a secretome that failed to induce TAMs. To understand the signaling pathways affected by the over-expression of Midkine in human melanoma cells, RNA sequencing was performed on Midkine-transduced isogenic cell line pairs. Intriguingly, Midkine over-expression in melanoma cells induced NF-κb targets that explain secretome features leading to the immunosuppressive microenvironment. More importantly, combined inhibition of Midkine and immune checkpoints exerted synergistic effects in animal models. This was further corroborated by bioinformatically teasing out Midkine-associated gene sets that could predict clinical response to immune checkpoint inhibition.

These findings are striking, yet it remains unclear how Midkine secretion is regulated in melanoma. Midkine mRNA expression does not directly correlate to its release; hence, other currently unknown pathways determine Midkine translation, trafficking, and release. How are these pathways regulated? The fact that some melanoma cell lines such as SK-Mel-147 express high levels of Midkine (Olmeda et al., 2017) suggests that “stable” molecular features (e.g., genomic alterations) are involved in the regulation of Midkine. Overall, the results presented by Cerezo-Wallis et al. indicate that Midkine shapes an immune-suppressive TME in melanoma; however, a plethora of other soluble factors influence the immune-TME, and their interaction with Midkine-induced changes should be clarified. Importantly, Midkine-rich tumors appear to be “inflamed” despite being immune evasive. Current signatures associated with response to immune checkpoint inhibitors in multiple cancer types, such as the T-cell-inflamed gene expression profile, are defined by T-cell infiltration, IFN-γ activity, and high levels of adaptive immune-resistant genes (Cristescu et al., 2018). Therefore, it would be interesting to understand whether the integration of Midkine-derived signatures, focusing on scoring the TME via the myeloid compartment, to current IFN-γ-related signatures may improve the predictive value of such biomarkers in melanoma. In their study, Cerezo et al. compared the prognosis of patient populations within the top and bottom 15th percentiles of Midkine-gene expression profiles; it is critical to understand whether the immune-resistant phenotype is limited to the top 15% tumors, and how to consider the large fraction of “Midkine-intermediate” tumors that account for 70% of patients. More direct Midkine measures in the TME, such as protein expression analysis, can help establish the optimal cutoff values to identify Midkine-induced immune-suppression.

Collectively, targeting Midkine and Midkine-induced changes has excellent potential to reshape the TME into an immunotherapy-permissive state. Further research to bring these findings closer to patients with checkpoint-resistant melanoma is highly warranted.

自引入新型免疫疗法以来，转移性黑色素瘤患者的不良预后已显着改善（Donia 等人，  2019 年）。然而，尽管一些患者经历了显着且持久的肿瘤消退，但在接受当前基于检查点阻断的免疫疗法治疗的患者中，超过一半的患者在五年内并未获得持续的益处和进展，留下了巨大的未满足的临床需求。骨髓细胞，包括肿瘤相关巨噬细胞 (TAM)，长期以来一直被怀疑促进免疫抑制和免疫治疗难治性肿瘤微环境 (TME; Pathria et al.,  2019）。通过靶向 TAM 等骨髓细胞来操纵免疫抑制性肿瘤微环境可能是改善转移性黑色素瘤临床结果的关键。

Midkine 是一种具有多效性功能的肝素结合生长因子。Midkine 的过表达已被证明可促进许多肿瘤类型中的肿瘤增殖、细胞凋亡逃避、血管生成和炎症（Filippou 等人，  2020）。基于他们之前的结果（Olmeda et al.,  2017），Cerezo-Wallis 等人。表明黑色素瘤细胞分泌的 Midkine 促进了发炎但免疫难治性肿瘤微环境，导致对免疫检查点阻断的抵抗。通过在移植到小鼠体内的人和小鼠黑色素瘤细胞系中过度表达 Midkine，作者表明，Midkine 调节了吸引免疫抑制性巨噬细胞的自分泌和旁分泌信号。这些受 Midkine 教育的巨噬细胞损害了 T 细胞活性。相比之下，人类黑色素瘤细胞中 Midkine 的消耗导致分泌组不能诱导 TAM。为了了解受 Midkine 在人黑色素瘤细胞中过表达影响的信号通路，对 Midkine 转导的同基因细胞系对进行了 RNA 测序。耐人寻味的是，黑素瘤细胞中中期因子过表达诱导 NF-κb 靶标，这些靶标解释了导致免疫抑制微环境的分泌组特征。更重要的是，Midkine 和免疫检查点的联合抑制在动物模型中发挥了协同作用。通过生物信息学梳理出可以预测对免疫检查点抑制的临床反应的 Midkine 相关基因组进一步证实了这一点。

这些发现令人震惊，但仍不清楚黑色素瘤中中期因子的分泌是如何受到调节的。Midkine mRNA 表达与其释放不直接相关。因此，其他目前未知的途径决定了 Midkine 的翻译、贩运和释放。这些途径是如何调节的？事实上，一些黑色素瘤细胞系如 SK-Mel-147 表达高水平的 Midkine (Olmeda et al.,  2017) 表明“稳定的”分子特征（例如，基因组改变）参与了 Midkine 的调节。总体而言，Cerezo-Wallis 等人提出的结果。表明 Midkine 在黑色素瘤中形成免疫抑制 TME；然而，过多的其他可溶性因素会影响免疫 TME，应该澄清它们与 Midkine 诱导的变化的相互作用。重要的是，尽管免疫回避，富含 Midkine 的肿瘤似乎“发炎”了。当前与多种癌症类型中免疫检查点抑制剂反应相关的特征，例如 T 细胞炎症基因表达谱，由 T 细胞浸润、IFN-γ 活性和高水平的适应性免疫抗性基因定义（Cristescu等人，  2018）。因此，了解 Midkine 衍生特征（侧重于通过骨髓隔室对 TME 评分）与当前 IFN-γ 相关特征的整合是否可以提高此类生物标志物在黑色素瘤中的预测价值将是一件有趣的事情。在他们的研究中，Cerezo 等人。比较了中期因子基因表达谱顶部和底部 15% 内患者群体的预后；至关重要的是要了解免疫抗性表型是否仅限于前 15% 的肿瘤，以及如何考虑占患者 70% 的大部分“中间因子”肿瘤。TME 中更直接的 Midkine 测量，例如蛋白质表达分析，可以帮助建立最佳临界值，以识别 Midkine 诱导的免疫抑制。

总的来说，靶向 Midkine 和 Midkine 诱导的变化具有将 TME 重塑为免疫治疗许可状态的巨大潜力。进一步研究以使这些发现更接近具有检查点耐药性的黑色素瘤患者是非常有必要的。