Covid-19, the disease caused by pandemic SARS-CoV-2 infection, had significant impact on patients with hematologic conditions¹; a meta-analysis involving 3377 patients with hematologic malignancies who were affected by Covid-19 reported a mortality rate of 34%.² A similarly dismal outcome was documented among 175 patients with chronic myeloproliferative neoplasms (MPN), collected in a European observational study, where mortality rate was 30% for the entire cohort, reaching 48% in primary overt myelofibrosis (MF).³ Covid-19 was also associated with higher incidence of thrombosis in patients with essential thrombocythemia (ET), compared to MF and polycythemia vera (PV) (20% vs 5% for both, respectively).⁴ Finally, MPN patients surviving the acute phase may suffer from additional long-term sequelae from Covid-19, that furtherly increase mortality and morbidity.⁵

The JAK1 and JAK2 inhibitor (JAKi) ruxolitinib is approved for the treatment of patients with MF and hydroxyurea resistant/refractory PV.⁶ By inhibiting JAK–STAT signaling, ruxolitinib has profound effects on different cell compartments of the immune system, including T cells, natural killer and dendritic cells, in addition to potently dampening inflammatory cytokine production.⁷ These properties have been mechanistically linked to the increased rate of infections in MPN patients receiving ruxolitinib, and, conversely, were explored successfully in the setting of steroid-refractory acute graft vs host disease following allogeneic stem cell transplantation.⁸ In the above cited European study in MPN, rapid discontinuation of the drug was implicated in 75% of deaths occurring in the ruxolitinib-treated cohort; these were ascribed to a previously described “discontinuation syndrome”, a potentially fatal complication due to a cytokine storm that follows the abrupt suspension of ruxolitinib.^{9, 10} In fact, observational studies support the effectiveness of ruxolitinib to quench the hyper inflammatory reaction accompanying Covid-19 in the general population.^{11, 12} Due to the immunomodulatory properties of ruxolitinib, the question arises whether response to vaccination for SARS-CoV-2 in patients under stable ruxolitinib therapy might be impaired.

We prospectively assessed serologic response following the first injection of SARS-CoV-2 mRNA vaccine in 30 consecutive patients with PV, ET and MF who were referred to the Center of Research and Innovation of Myeloproliferative Neoplasms (CRIMM), Florence. Patients signed an informed consent to participate in the study, that was approved by the local Ethical Committee of Azienda Ospedaliera Careggi, Florence. Patients were eligible if they had a diagnosis of MPN according to the 2016 WHO criteria and all the following at the time of study entry: no history of positivity for SARS-CoV-2 by PCR on swab; negativity of serum anti-nucleoprotein antibodies; no clinical suspicion of Covid-19. A cohort of 14 healthy volunteers without prior SARS-CoV-2 infection was used as a reference group. The vaccines used were the Moderna and Pfizer vaccine in 25 (83%) and five (17%) MPN patients, and 10 (71%) and four (29%) healthy controls, respectively. Blood samples were collected before first vaccination (T0) and right before the second dose administration (T1, day 21 for Pifzer, day 28 for Moderna). Serologic tests for SARS-CoV-2 antibodies were performed to demonstrate presence of IgG antibodies against spike (S) protein, receptor binding domain (RBD) and neutralizing antibodies. A cut-off value of test positivity was established for each antibody type according to manufacturer's instructions; patients above the upper cut-off level were considered as “responders”, and those below as “non responders”. Categorical variables were expressed as frequency and percentage. Chi-square test was used to compare categorical variables.

Clinical characteristics of the patients are outlined in Table S1. There were 10, seven and 13 patients with PV, ET and MF (two and three patients were post-PV and post-ET MF, respectively). Of these, 18 were on a stable dose of ruxolitinib since at least 3 months (ruxo-patients), while 12 were not currently treated, nor had received before, ruxolitinib (no-ruxo patients), including five patients under watch-and-wait and five under hydroxyurea therapy since at least 3 years. The current median dose of ruxolitinib was 20 mg daily (range, 10–50 mg), and the median duration of ruxolitinib therapy was 7.3 years (range, 0.8–13.8 years). Figure 1 shows the levels of individual anti-SARS-CoV-2 antibodies at T0 and T1. Anti-S IgG, anti-RBD IgG and neutralizing Ab were not detected before vaccination in any of the three groups, accordingly to predefined cut-off levels. In general, the extent of specific antibody response after first dose vaccination (T1 time point), measured as binding antibody unit (BAU)/mL for anti-S and anti-RBD immunoglobulin, and relative index for neutralizing antibodies (Figure 1(A)–(C)), was significantly lower in ruxo-patients compared to healthy controls and the no-ruxo group. The latter conversely did not differ significantly from controls, suggesting that the potentiality to mount adequate immune response is maintained in most MPN patients who were not receiving ruxolitinib. In detail, mean anti-S BAU levels/ml were 111.4, 513.4 and 510.8 for ruxo, no-ruxo and controls, respectively (Figure 1(A)); corresponding anti-RBD BAU levels/mL were 65.2, 631.7 and 254.2 (Figure 1(B)); and, for neutralizing antibodies, the mean relative index was 1.1, 3.5 and 3.8 (Figure 1(C)). All these values were statistically different when comparing ruxo patients vs no-ruxo and controls, but not between controls and no-ruxo patients (p values are reported in Figure 1(A)–(C)).

According to the predefined cut-off levels, all the 14 healthy volunteers were considered as responders to vaccination for the anti-S IgG, anti RBD IgG and neutralizing antibodies (100% each), compared to 38.8%, 33.3% and 33.3%, respectively, for the ruxo-patients (p < 0.001 vs controls for each antibody type), and 91.6% (p = not significant), 91.6% (p = not significant) and 58.3% (p = < 0.01) in the no-ruxo group (Figure 1(D)–(F)).

In summary, these findings, with the limitation of the small number of subjects included, make a strong and urgent argument for an impaired early response to SARS-CoV-2 vaccine in patients receiving ruxolitinib. Further and future studies are needed to address whether such unresponsive status persists after the second dose of vaccine, as suggested by a study performed in Israel, where the rate of seropositivity (anti-S1/S2 IgG) after complete vaccination in patients with MPN was 42% for those using JAKi.¹³ It will be important to address whether also responses mediated by T-cell and other myeloid cells are impaired by ruxolitinib treatment, owing to their key role in SARS-CoV-2 infection.^14-16 Although clear-cut relationships between specific anti-SARS-CoV-2 immunoglobulin titers and protection against the virus has not been unequivocally established in the general population, MPN patients receiving ruxolitinib should be urged to continue to adopt the best preventive measures against Covid-19 even after receiving vaccination, in the light of the evidences presented herein. Furthermore, it is also suggested that, since MPN patients not receiving ruxolitinib overall developed antibody titers that were comparable to healthy volunteers, but a proportion of them not did actually produce neutralizing antibodies, initiation of ruxolitinib therapy in a naïve patient might be prudentially delayed after completion of vaccination, unless urgently needed. This notwithstanding, we reinforce that patients with MPN, as for any hematologic malignancies, should be vaccinated, since the possibility of protection at any extent outweighs minor risks,² as supported also by recommendations from the American Society of Hematology (https://www.hematology.org/covid-19).

Covid-19 是由大流行性 SARS-CoV-2 感染引起的疾病，对患有血液病的患者¹产生了重大影响；一项涉及 3377 名受 Covid-19 影响的血液系统恶性肿瘤患者的荟萃分析报告死亡率为 34%。²在一项欧洲观察性研究中收集的 175 名慢性骨髓增生性肿瘤 (MPN) 患者中记录了类似的令人沮丧的结果，其中整个队列的死亡率为 30%，原发性显性骨髓纤维化 (MF) 的死亡率达到 48%。³与 MF 和真性红细胞增多症 (PV) 相比，Covid-19 还与原发性血小板增多症 (ET) 患者的血栓形成率较高相关（两者分别为 20% 和 5%）。⁴最后，在急性期幸存的 MPN 患者可能会遭受来自 Covid-19 的额外长期后遗症，这会进一步增加死亡率和发病率。⁵

JAK1 和 JAK2 抑制剂 (JAKi) 鲁索替尼被批准用于治疗 MF 和羟基脲耐药/难治性 PV 患者。⁶通过抑制 JAK-STAT 信号传导，鲁索替尼对免疫系统的不同细胞区室产生深远影响，包括 T 细胞、自然杀伤细胞和树突细胞，此外还可以有效抑制炎性细胞因子的产生。⁷这些特性与接受鲁索替尼的 MPN 患者的感染率增加存在机械关联，相反，在同种异体干细胞移植后类固醇难治性急性移植物抗宿主病的情况下，这些特性已被成功探索。⁸在上述提到的欧洲 MPN 研究中，在 ruxolitinib 治疗的队列中，75% 的死亡与药物的快速停药有关；这些归因于先前描述的“停药综合征”，这是一种潜在的致命并发症，原因是鲁索替尼突然停药后出现细胞因子风暴。^{9, 10}事实上，观察性研究支持鲁索替尼在普通人群中抑制 Covid-19 伴随的过度炎症反应的有效性。^{11, 12}由于鲁索替尼的免疫调节特性，出现了一个问题，即接受稳定鲁索替尼治疗的患者对 SARS-CoV-2 疫苗接种的反应是否可能受损。

我们对转诊至佛罗伦萨骨髓增生性肿瘤研究与创新中心 (CRIMM) 的连续 30 名 PV、ET 和 MF 患者首次注射 SARS-CoV-2 mRNA 疫苗后的血清学反应进行了前瞻性评估。患者签署了参与该研究的知情同意书，该同意书得到了佛罗伦萨 Azienda Ospedaliera Careggi 当地伦理委员会的批准。如果患者根据 2016 年 WHO 标准诊断为 MPN，并且在进入研究时满足以下所有条件，则符合条件： 没有通过拭子 PCR 检测出 SARS-CoV-2 阳性史；血清抗核蛋白抗体阴性；没有临床怀疑 Covid-19。一组 14 名未感染过 SARS-CoV-2 的健康志愿者被用作参考组。分别在 25 名 (83%) 和 5 名 (17%) 的 MPN 患者以及 10 名 (71%) 和 4 名 (29%) 的健康对照组中使用了 Moderna 和 Pfizer 疫苗。在第一次接种疫苗之前（T0）和第二次给药之前（T1，Pifzer 第 21 天，Moderna 第 28 天）收集血样。进行了 SARS-CoV-2 抗体的血清学测试，以证明存在针对刺突 (S) 蛋白、受体结合域 (RBD) 和中和抗体的 IgG 抗体。根据制造商的说明，为每种抗体类型建立了测试阳性的临界值；高于上限的患者被视为“有反应者”，低于上限的患者被视为“无反应者”。分类变量表示为频率和百分比。卡方检验用于比较分类变量。

表S1概述了患者的临床特征。PV、ET 和 MF 分别有 10、7 和 13 名患者（2 名和 3 名患者分别为 PV 后和 ET 后 MF）。其中，18 人从至少 3 个月开始服用稳定剂量的鲁索替尼（ruxo 患者），而 12 人目前没有接受过鲁索替尼治疗，也没有接受过鲁索替尼（无 ruxo 患者），其中包括 5 名接受观察和治疗的患者等待和五个羟基脲治疗至少 3 年。目前鲁索替尼的中位剂量为每天 20 毫克（范围，10-50 毫克），鲁索替尼治疗的中位持续时间为 7.3 年（范围，0.8-13.8 年）。图 1 显示了 T0 和 T1 时个体抗 SARS-CoV-2 抗体的水平。三组接种前均未检测到抗 S IgG、抗 RBD IgG 和中和抗体，根据预定义的截止水平。一般来说，第一剂疫苗接种后特异性抗体反应的程度（T1 时间点），以抗 S 和抗 RBD 免疫球蛋白的结合抗体单位 (BAU)/mL 和中和抗体的相对指数来衡量（图 1(A) )–(C))，与健康对照组和无 ruxo 组相比，ruxo 患者显着降低。相反，后者与对照组没有显着差异，这表明在大多数未接受鲁索替尼的 MPN 患者中保持了产生足够免疫反应的潜力。详细地说，ruxo、no-ruxo 和对照的平均抗 S BAU 水平/ml 分别为 111.4、513.4 和 510.8（图 1（A））；相应的抗 RBD BAU 水平/mL 分别为 65.2、631.7 和 254.2（图 1（B））；并且，对于中和抗体，平均相对指数为 1.1，3.5 和 3.8（图 1（C））。在比较 ruxo 患者与无 ruxo 和对照时，所有这些值都有统计学差异，但在对照和无 ruxo 患者之间没有差异。p值在图 1(A)-(C)) 中报告。

根据预先设定的临界值，所有 14 名健康志愿者都被认为是抗 S IgG、抗 RBD IgG 和中和抗体疫苗接种的应答者（各 100%），相比之下，分别为 38.8%、33.3% 和 33.3%，分别对于 ruxo 患者（p  < 0.001 与每种抗体类型的对照），以及 91.6%（p  = 不显着）、91.6%（p  = 不显着）和 58.3%（p  = < 0.01） ruxo 组（图 1（D）-（F））。

总而言之，这些发现，由于包括少数受试者的限制，为接受鲁索替尼的患者对 SARS-CoV-2 疫苗的早期反应受损提出了强有力和紧迫的论据。正如在以色列进行的一项研究所表明的那样，需要进一步和未来的研究来解决在第二剂疫苗后这种无反应状态是否持续存在，其中 MPN 患者在完全接种疫苗后的血清阳性率（抗 S1/S2 IgG）为使用 JAKi 的用户为 42%。¹³重要的是要解决由 T 细胞和其他骨髓细胞介导的反应是否也因鲁索替尼治疗而受损，因为它们在 SARS-CoV-2 感染中起关键作用。^14-16尽管尚未在普通人群中明确确定特定的抗 SARS-CoV-2 免疫球蛋白滴度与对病毒的保护之间的明确关系，但应敦促接受鲁索替尼治疗的 MPN 患者继续对 Covid-19 采取最佳预防措施即使在接受疫苗接种后，根据本文提供的证据。此外，还有人建议，由于未接受鲁索替尼治疗的 MPN 患者总体上产生的抗体滴度与健康志愿者相当，但其中一部分实际上并未产生中和抗体，因此可能会谨慎推迟在未接受过鲁索替尼治疗的患者中开始接受鲁索替尼治疗。完成疫苗接种，除非紧急需要。尽管如此，我们强调 MPN 患者，²也得到美国血液学会 (https://www.hematology.org/covid-19) 建议的支持。