Non-Ablative Chemotherapy Followed by HLA-Mismatched Allogeneic CD3+ T-Cells Infusion Causes An Augment of T-Cells With Mild CRS: A Multi-Centers Single-Arm Prospective Study on Elderly Acute Myeloid Leukemia and int-2/High Risk Myelodysplastic Syndrome Patients

Objective To evaluate the efficacy and safety of standard or low-dose chemotherapy followed by HLA-mismatched allogeneic T-cell infusion (allo-TLI) for the treatment of elderly patients with acute myeloid leukemia (AML) and patients with intermediate-2 to high-risk myelodysplastic syndrome (MDS). Methods We carried out a prospective, multicenter, single-arm clinical trial. Totally of 25 patients were enrolled, including 17 AML patients and 8 MDS patients. Each patient received four courses of non-ablative chemotherapy, with HLA-mismatched donor CD3+ allo-TLI 24 h after each course. AML patients received chemotherapy with decitabine, idarubicin, and cytarabine, and MDS patients received decitabine, cytarabine, aclarubicin, and granulocyte colony-stimulating factor. Results A total of 79 procedures were performed. The overall response rates of the AML and MDS patients were 94% and 75% and the 1-year overall survival rates were 88% (61–97%) and 60% (13–88%), respectively. The overall 60-day treatment-related mortality was 8%. Compared with a historical control cohort that received idarubicin plus cytarabine (3 + 7), the study group showed significantly better overall response (94% vs. 50%, P=0.002) and overall survival rates (the 1-year OS rate was 88% vs. 27%, P=0.014). Post-TLI cytokine-release syndrome (CRS) occurred after 79% of allo-TLI operations, and 96% of CRS reactions were grade 1. Conclusion Elderly AML patients and intermediate-2 to high-risk MDS patients are usually insensitive to or cannot tolerate regular chemotherapies, and may not have the opportunity to undergo allogeneic stem cell transplantation. Our study showed that non-ablative chemotherapy followed by HLA-mismatched allo-TLI is safe and effective, and may thus be used as a first-line treatment for these patients. Clinical Trial Registration https://www.chictr.org.cn/showproj.aspx?proj=20112.


INTRODUCTION
Adoptive CD3 + T-cell infusion, such as donor lymphocyte infusion (DLI), is an established treatment strategy for activating donor-derived T-cells to eliminate leukemic cells (1,2). The recent emergence of immune checkpoint inhibitors (3,4) and chimeric antigen receptor T-cell immunotherapy technologies (5,6) has led to a new understanding of the role of T-cell immunity in tumor treatment, including the abilities of genetically modified or unmodified T-cells with tumor antigenrecognition ability to kill tumor cells (7,8). However, T-cell activation also triggers adverse reactions such as cytokine-release syndrome (CRS) (9,10). Most gene-modified T-cell infusion (TLI) technologies are currently based on autologous T-cells, while most non-modified TLIs are based on allogeneic hematopoietic stem cell transplantation (allo-HSCT; i.e. infusion of donor-derived T-cells after allo-HSCT), and most of these infusions involve HLA-matched donor T-cells, also called donor lymphocyte infusion (DLI) (11,12). Although some clinical trials of HLA-mismatched allogeneic TLI (allo-TLI) after non-myeloablative chemotherapy have shown encouraging effects (13)(14)(15), this approach is rarely used in clinical practice and the inherent immune mechanism of this technology remains unclear.
Elderly patients with acute myeloid leukemia (AML) respond poorly to current therapies and have low survival rates, with a remission rate of 43%-65% and a therapy-related mortality (TRM) rate of 11%-23% for patients older than 55 years (16)(17)(18)(19), and a 5-year disease-free survival rate for elderly AML patients of only 15% (20). This poor prognosis for elderly AML patients is primarily related to more cytogenetic abnormalities and elderly individuals do not tolerate chemotherapy well (21). Myelodysplastic syndrome (MDS) is a hematological malignancy characterized by hematopoietic disorder, ineffective hematopoiesis, and an increased risk of progression to AML (22). Although MDS patients classified at intermediate-2 (int-2) to high risk of AML according to the International Prognostic Score System (IPSS) require chemotherapy (23), these patients are usually insensitive to or unable to tolerate chemotherapy.
Long-term myelosuppression after chemotherapy is one of the treatment challenges in patients with MDS, leading to uncontrollable infection, bleeding, and an increased risk of TRM. Although HSCT has been considered as the only curative option for malignant hematologic disease, elderly AML patients or int-2 to high-risk MDS are often unable to receive HSCT due to advanced age, poor Eastern Cooperative Oncology Group (ECOG) scores, comorbidities, and a lack of suitable donors (24). There is thus an urgent need to develop appropriate treatments to maximize the response rate and reduce the incidence of adverse events in elderly patients with AML or MDS.
We designed a prospective, multicenter, single-arm clinical study to investigate a novel treatment for elderly AML patients and int-2/high-risk MDS patients. The protocol included four courses of non-ablative chemotherapy, each followed by one dose of HLA-mismatched allogeneic TLI (allo-TLI). We evaluated the response, survival, and safety, and observed the immune therapy-related reactions to explore the potential antileukemia mechanism of this novel T-cell therapy.

Patients and Donors
We enrolled 25 patients at three hospitals from October 27, 2017 to September 22, 2020. The patients were followed-up until March 2, 2021, with a median follow-up time of 30.0 months (range: 1.2-35.9 months). The inclusion criteria were as follows: 1) diagnosed with de novo AML and age ≥55 years, or de novo int-2 to high-risk MDS and age 18-80 years; 2) unfit (see Supplementary Materials 1) for or unwilling to undergo allo-HSCT; 3) ECOG performance status ≤2; and 4) available donor willing to donate T-cells. The diagnosis was defined according to the 2016 World Health Organization (WHO) criteria (25). The risk stratification of AML was assessed using the 2019 National Comprehensive Cancer Network (NCCN) guidelines, and the risk stratification of MDS was assessed using the IPSS system (23). Donor criteria were as follows: 1) ≤45 years old; 2) relative or legal relation of the patient (e.g. spouse); 3) consent to donate lymphocytes after fully informed; 4) match in fewer than seven of 10 HLA loci; 5) no blood-borne or bone marrow disease; and 6) tolerated the process of granulocyte colony-stimulating factor (G-CSF) mobilization and lymphocytes collection.
A historical control cohort comprised 30

Treatment Design
AML patients received induction therapy with decitabine (DAC, 15 mg/m 2 , day -10~-6) followed by a reduced dose of idarubicin 8 mg/m 2 , day -5~-3, and cytarabine 100 mg/m 2 , day -5~-1 (IA 3 + 5 regimen). Patients with MDS received DAC (15 mg/m 2 , day -9~-5) plus a 4-day induction chemotherapy regimen consisting of cytarabine (10 mg/m 2 , day -4~-1), aclarubicin (10 mg/day, day -4~-1), and G-CSF(150 mg every 12 h, day -5~-1). G-CSF was discontinued when the neutrophil count reached 2×10 9 /L. Patients who achieved partial remission (PR) received a second course of the same induction chemotherapy. AML patients who failed to achieve PR; or complete remission (CR) after the second course of induction chemotherapy will be withdrawn from the trial and received other treatments. For consolidation treatment, AML patients who achieved CR received three further courses of DAC+ (IA 3 + 7). MDS patients who achieved bone marrow or blood responses received three further DAC+ 6-day of CAG regimens. CD3 + T-cells from the same donor were allocated equally among four doses, with one dose administered per cycle without graft versus host disease (GvHD) prophylaxis 24 h (day 0) after the end of each chemotherapy. The interval between treatment cycles was 60 days. Patients who developed signs of CRS were given acetaminophen instead of steroids. Patients who relapsed or experienced severe adverse events were withdrawn from the trial. Patients who completed four allo-TLI cycles were followed up closely every 3 months. The allo-TLI protocol and patient outcomes are shown in Figure 1.
Patients in the historical control cohort received IA(3 + 7) (idarubicin 8 mg/m 2 for 3 consecutive days and cytarabine 100 mg/m 2 for 7 consecutive days). Hydroxyurea and hemapheresis were permitted as rescue treatments to control white blood cells. No allo-TLI was given after chemotherapy.

Mobilizing and Collecting Donor CD3 + T-Cells
Donors received G-CSF 8 mg/kg/day for 5-6 consecutive days from day -4. Peripheral blood lymphocytes were harvested on day −1 and 0 by using a white blood cell apheresis procedure (COBE Spectra, Lakewood, Colorado, USA). After apheresis, the donor cells were divided equally into four aliquots; the first aliquot was used immediately and the remaining three were cryopreserved in liquid nitrogen. The median number of mononuclear cells was 2.58×10 8 /kg (IQR: 2.30-3.49×10 8 /kg) and the median number of CD3 + T-cells infused per course was 0.44×10 8 /kg (IQR: 0.36-0.60×10 8 /kg). The dose per Kg for the TLI refers to the weight of the patient.

Micorchimerism Detection
Microchimerism (donor cells <1%) was initially examined using bone marrow sample of patients on day 7 of the first course of chemotherapy and before each of the next three courses of chemotherapy. Microchimerism was detected using real-time quantitative polymerase chain reaction (RQ-PCR) of the insertion-deletion polymorphism site (InDel) with a sensitivity    (32) was used to evaluate the severity of aGvHD. The 60-day TRM rate was categorized as mortality without disease progression or death within 60 days after treatment. Non-recurrence mortality was defined as mortality not due to disease recurrence or progression. The hematopoietic recovery time was defined as the duration from the end of chemotherapy to an absolute neutrophil count ≥0.5×10 9 /L or platelet count was ≥20×10 9 /L for 3 consecutive days. CRS was graded according to the CRS revised grading system (33).

Statistical Analysis
The web tool "StatBox" (https://www.cnstat.org/statbox) is used to estimate sample size (34). Statistical analysis and presentation were carried out using SPSS 24.0 (Statistical Product and Service Solutions, Armonk, New York, USA) and GraphPad Prism Version 7.00 (San Diego, California, USA). Survival curves were prepared using the Kaplan-Meier method. Longitudinal trends in temperature, T-cell numbers, and cytokine levels in peripheral blood from pre-to post-allo-TLI day 7 were assessed using mixed model repeated measures (Generalized Linear Mixed Model) (35) in patients who had completed three or four cycles of allo-TLI before November 1, 2019. Correlations between two groups of quantitative data that did not conform to a Gaussian distribution were analyzed by Spearman's test. We compared the 30 patients in the historical cohort with 17 AML patients who received HLA-mismatched allo-TLI, and prepared survival curves using the Kaplan-Meier method. A two-tailed P value <0.05 was considered statistically significant.

RESULT Clinical Characteristics of Patients and Donors
A total of 25 patients were included in this study, including 17 elderly AML patients and eight patients with IPSS int-2 or highrisk MDS. A total of 79 HLA-unmatched allo-TLI procedures were completed. Of the 25 patient/donor pairs, 16 (64%) were matched at fewer than five of 10 HLA alleles. Nine of the 17 AML patients (53%) were considered high-risk, 62% of the MDS patients were int-2 and 38% were high-risk. The detailed clinical characteristics are presented in Table 1. The historical cohort comprised 30 elderly AML patients (median age 66 years, IQR: 63-70 years; 16 males and 14 females). There was no significant difference in risk stratification between the current AML patients and controls. The detailed patient characteristics are listed in Table 2.
This was a single-arm study, and we therefore used a historical cohort as a control group. The CR rates for the AML and historical groups were 77% vs. 40%, the CR+CRi rates were 82% vs. 47%, and the OR rates were 94% vs. 50%, respectively (P=0.016, 0.017 and 0.002, respectively). The 1-year OS rate was also significantly better compared with the historical controlgroup (88% vs. 27%; P=0.014) ( Table 2 and Figure 2D). The median OS was 7.6 months for the historical group and 24.4 months for the AML group.

Adverse Events and Hematopoietic Recovery
All patients had grade 3-4 hematologic toxicity, including neutropenia and thrombocytopenia. Among non-hematologic toxicities, 32% of patients developed grade 3-4 infections. Severe adverse events involving major organs, such as respiratory failure, severe cardiac dysfunction, and renal dysfunction affected ≤2%. Severe liver dysfunction affected 4% of patients and was mainly attributed to antifungal drugs. The 60-day TRM rate of all patients was 8% (2/25), of AML patients was 6% (1/17). The median times to neutrophil and platelet recovery in the AML group were 12 days (95%CI: 11.4-14.3 days) and 10 days (95% CI: 9.6-13.5 days), and the equivalent times in the MDS group were 7 days (95%CI: 6.1-10.8 days) and 8 days (95%CI: 6.9-10.6 days), respectively. Details of the adverse events and hematopoietic recovery are shown in Table 4. There were no significant differences in the median times to neutrophil and platelet recovery between the current AML patients and the historical control group (all P>0.05) ( Table 2).

Microchimerism
Nine patients were not willing to pay for microchimerism detection. A total of 16 patients were examined for microchimerism. Donor microchimerism was detected 37 out of 41 times (90%). Except for one MDS patient who had an average microchimerism of 0.163% (0.106-0.216%) of copies for all tests but no abnormal clinical symptoms or signs, all other patients displayed microchimerisms of 0.001%-0.022% of copies ( Figure 3A). The average number of microchimerism copies decreased gradually with the increasing of allo-TLI courses ( Figure 3B).

Increase of Recipient-Derived Peripheral T-Cells After HLA-Mismatched Allo-TLI
Based on the 56 HLA-mismatched allo-TLI procedures, numbers of CD3 + , CD4 + , and CD8 + T-cells increased gradually to a peak on day 7 (P<0.001 compared with D0; Figure 6A and Supplementary Figure 1). Comparing the increasing curve of CD3 + T-cells with that for neutrophils showed that the increase   Patients who did not reach OR were classified as no response. c Relapse was defined as bone marrow blasts ≥5%, reappearance of blasts in peripheral blood, and/or extramedullary disease after CR or CRi. in T-cells was not caused by hematologic recovery ( Figure 6B).
To determine the origin of the increased T-cells, we sorted peripheral blood T-cells from one patient. We analyzed the sorted T-cells by the chimerism detection, which confirmed that the increased peripheral T-cells were all derived from the recipient in this patient (see Supplementary Table 3). In addition, patients with shorter platelet recovery times (shorter than the median recovery time) had a higher number of CD3 + T-cells (P=0.028; Figure 6C and Supplementary Table 4). However, there was no significant difference in CD3 + T-cell counts between patients with shorter and longer neutrophil recovery times (P=0.967; Supplementary Figure 2 and Supplementary

DISCUSSION
DLI has been used as a type of allo-TLI for decades; however, non-ablative chemotherapy followed by HLA-mismatched stem cell transplantation, also referred to as "micro-transplantation", has also been used to treat myeloid malignancies, with a better response, faster blood cell recovery, and lower TRM rate than conventional regimens, especially in elderly AML patients and patients with MDS (13-17, 19, 36-39). In these previous studies, although CD34 + hematopoietic stem cells were mobilized and isolated from HLA-mismatched donors and infused into the recipient, most researchers believed that the main functional anti-leukemia cells were T-cells, and donor CD3 + T-cells were found to interfere with survival after microtransplantation (38,40). We therefore used HLA-mismatched allogeneic CD3 + TLI as immune therapy for elderly AML patients and int-2/high-risk MDS patients. Donor T-cell isolation was carried out according to the protocol for DLI in our institute (41,42), resulting in a median infusion of 0.44×10 8 /kg CD3 + T-cells per dose, which was similar to the DLI dose regularly used in our institute for allogeneic stem cell transplantation (41). However, the optimal TLI dose for HLA-mismatched allo-TLI still needs to be determined. We used DAC plus IA(3 + 5) followed with HLAmismatched allo-TLI for AML cohort and achieved a CR+CRi of 82%, a 1-year OS rate of 88%, and a median OS of 24.4 months. This result is significantly better than that of a historical control in our center, the OR rates were 94% vs. 50% for this study and the historical control, and the 1-year OS rate were 88% vs. 27%, respectively. We also used the DAC+CAG regimen for int-2/ high-risk MDS patients, and showed an OR rate of 75% and a median OS of 21.5 months.
Most previous HLA-mismatched stem cell transplantation studies used standard-dose chemotherapy as induction therapy, followed by consolidation with high-dose cytarabine. In this study, we designed HMA plus reduced-dose IA for the AML cohort, which is based on the following considerations: Most of the elderly AML patients were secondary AML preceded by MDS or with myelodysplasia-related changes (43); poor-risk cytogenetics increased with advancing age, with 51% of     and achieved a 50% CR+CRi rate and median OS of 11 months (46). Our protocol also showed a significantly better response and survival compared to our historical controls. These results suggest that DAC+IA followed by HLA-mismatched allo-TLI improved response and survival rates in elderly AML patients, and can be an alternative therapy for "fit" elderly AML patients other than intensive chemotherapy or Bcl-2 inhibitor based therapies. The CAG regimen, initially used by researchers in Japan, consists of low-dose cytarabine, low-dose aclarubicin, and G-CSF (47,48), and has shown similar or better response/survival with lower treatment-related toxicity in AML and MDS patients (49)(50)(51)(52). The DAC+CAG regimen has shown similar response and survival, but with reduced infections than standard regimens in elderly AML patients. A meta-analysis by Zhang et al. based on 10 randomized controlled trials including 590 int-2/high-risk MDS patients, found an OR rate of 70% (95% confidence interval: 58%-8l%) for DAC+CAG, which was significantly better than that for DAC or CAG alone (53). In the current study, the DAC+CAG regimen was used for int-2/high-risk MDS patients. Some new treatment strategies have recently demonstrated better responses than regular treatments in int-   (55). The current treatment protocol was thus similar to the above treatments, but with more excellent safety. The current response and survival rates were similar to those reported in previous HLA-mismatched stem cell transplantation studies. Guo et al. used mitoxantrone plus cytarabine (MA 3 + 7) as induction chemotherapy and high-dose cytarabine as consolidation chemotherapy, resulting in CR induction of 80% and a 2-year OS rate of 39.3% (14). A multicenter study (13) of elderly patients with de novo AML used IA, MA, and DAC+CAG as induction and consolidation with high-dose cytarabine. The CR rates for 60-64-and 65-69-year-olds were 75.4% and 70.2%, and the equivalent 1-year OS rates were 87.7% and 85.8%, respectively. Hu et al. (15) used DAC plus mitoxantrone and cytarabine (MA)(3 + 7) as induction chemotherapy for patients with int-2 to high-risk MDS, followed with DAC plus high-dose cytarabine as consolidation chemotherapy, with an OR rate of 81% and 1-year OS rate of 94.1%. However, the median times to neutrophil and platelet recovery in their study were 14 and 17 days, respectively, which were longer than the 7 and 8 days in the current study. This suggests that standard or even low-dose, rather than high-dose, chemotherapies can be used as conditioning regimens for HLA-mismatched allo-TLI. However, a large cohort study with a randomized control group is needed to confirm our results.
Previous HLA-mismatched stem cell transplantation studies revealed that GvHD was rare, and the proportion of severe aGvHD was only 1.1% (13). GvHD prophylaxis is therefore not considered necessary. In the present study, one patient with MDS developed grade 3 aGvHD and eventually died. The incidence of aGVHD in this trial is similar as previously reported. There were no apparent differences between this patient and the other patients in terms of age (65 years), donor age (42 years), sex of donor/recipient (male/female), matched HLA loci (≤5/10), risk stratification of disease (MDS at IPSS int-2), and dose of infused CD3 + T-cells (0.57×10 8 /kg). GvHD used to be considered unavoidable in patients undergoing HLAunmatched allo-HSCT, and GvHD prophylaxis was usually given even for HLA-matched allo-HSCT; however, some pioneer practices found that non-ablative conditioning followed by HLA-mismatched allo-HSCT produced donorrecipient chimerism, with no increase in GvHD, and long-term survival (56)(57)(58). The successful practice of microtransplantation has further expanded our understanding of HLA-mismatched allogeneic transplantation. The current study, demonstrated that, although aGvHD was rare, it can still occur after HLA-  mismatched allo-TLI. Therefore close observation and early intervention are encouraged, and further studies are needed to improve our understanding of the mechanism of aGvHD. This study showed that HLA-mismatched allo-TLI was generally safe, except for a very low risk of GvHD. The most common hematologic toxicities were neutropenia and thrombocytopenia. In previous studies of HLA-mismatched allo-HSCT, the median times to neutrophil and platelet recovery were significantly shorter than in the control group. The hematologic recovery time in the current study was similar to previous reports (14). However, there were no significant differences compared our current study with the historical control group. This might be because the historical control group did not include patients with DAC. Severe nonhematologic toxicities (> grade 3) were rare, and the 60-day TRM rates were 8% for all patients and 6% for AML patients. Mono-HMA has also been recommended for elderly AML patients unfit for intensive chemotherapy. A meta-analysis of nine published studies including 718 elderly patients with de novo AML showed that DAC monotherapy resulted in 30-day and 60-day TRM rates of 7% and 17%, respectively (59). Venetoclax plus HMA, as another popular treatment, was reported to have 30-day and 60-day TRM rates of 1% and 7%, respectively, compared with a 30-day TRM rate of 24% following intensive chemotherapy (60). These results indicated that the current treatment protocol was generally equivalent to nonintensive therapies in terms of safety.
The anti-leukemia mechanism of HLA-mismatched allo-TLI is not well known. Previous studies observed that the dose of infused donor CD3 + T-cells interfered with survival after HLAmismatched allo-TLI. A CD3 + T-cell dose ≥1.1×10 8 /kg was associated with better OS and leukemia-free survival than a dose <1.1×10 8 /kg (38,40). In this study, although we used G-CSF to mobilize donor stem cells, the highest CD3 + T-cell dose collected from a donor was 0.87×10 8 /kg (median 0.44×10 8 /kg). However, patients' T-cell counts were significantly increased after HLA-mismatched allo-TLI, and this effect was not due to the recovery of bone marrow after chemotherapy. Based on the chimerism detection in one case, we suspected that the elevation in T-cells originated from the recipient. However, more cases should be enrolled to furtherly explore the origin of the increased T cells after allo-TLI. We also observed that the patient's peripheral CD3 + T-cell counts increased in line with increasing cycles of HLA-mismatched allo-TLI, as reported by Zhu et al., suggesting that more cycles of allo-TLI promote graft versus leukemia effects (61).
Cytokine secretion is a byproduct of T-cell activation. Some cytokines are involved in the anti-tumor immunity of T-cells, while others contribute to the proliferation of hematopoietic cells. Previous studies showed changes in serum concentrations of several cytokines, such as IL-6, IL-8, and TNF-b, after HLAmismatched allo-TLI (62). In the present study, IL-2, IL-10, IL-6, and IFN-g levels increased significantly after HLA-mismatched allo-TLI, but IL-4, IL-17A, and TNF-a were unchanged. We also found that increased CD3 + T-cells levels were associated with platelet recovery time, and patients with shorter than median platelet recovery times had higher peripheral CD3 + T-cell counts. This might be due to an elevation of cytokines, which can promote the generation of platelets. IL-6 is considered to be involved in promoting megakaryocytic maturation and augmenting platelet counts (63), while other cytokines, such as IL-1 (64), IL-3, granulocyte-macrophage colony-stimulating factor (65), and IL-11 (66) stimulate platelet counts. However, whether these cytokines were released during post-allo-TLI CRS was unclear, and complete and detailed cytokine expression profiles need to be analyzed in future research.
In addition to activated T-cells and rising cytokine levels, we also observed a significant increase in body temperature after allo-TLI. Temperatures remained high during the first week after allo-TLI, and both hs-CRP and PCT increased during this period. Non-infectious fever, and increased hs-CRP, PCT, cytokines, and T-cells, are part of the immune response related to HLA-mismatched allo-TLI, similar to CRS following chimeric antigen receptor T-cell immunotherapy. Indeed, HLAmismatched HSCT-related CRS was proposed as early as 2009 in a study of non-engraftment haploidentical cellular immunotherapy, termed "haplo immunostorm", which presents with a series of clinical symptoms accompanied by cytokine flux, and has been shown to be sensitive to steroids (67). Sung et al. also reported an incidence of HLA-mismatched stem cell transplantation-related CRS of 40%, of which most cases were mild (62). There is currently no recognized term for this CRS, and we therefore referred to it as "post-TLI CRS". In our study, 79% of cases developed post-TLI CRS, of which 96% were mild. Patients with CRS in this study received indomethacin rather than glucocorticoids in the first week after transfusion. Together, these findings further our understanding of the immune reactions following HLA-mismatched allo-TLI. However, a more formal post-TLI CRS assessment and grading system needs to be established.
This study still has many limitations. The cases number is small, especially for MDS patients. And this was not a randomized controlled study. There are many treatment options for elderly leukemia patients now. The controlled group could be regular intensive chemotherapy such as IA (3 + 7); intensive chemotherapy followed reduced intensity allogenic HSCT; BCL-2 based chemotherapies or HMA and low-dose cytarabine based chemotherapies, none of them was commonly accepted as "the best" first-line induction therapy for elderly AML patients, because of the high heterozygous property that the elderly AMLs have, the same do MDS patients. Although it has been reported that the incidence is very low, we had one case who developed aGvHD and died. The safety of this protocol still needs to be closely observed in future studies. This study has a pleased ORR, but the 1-year PFS rates for the AML and MDS patients were still low. A prospective clinical trial on a maintenance therapy followed four courses of allo-TLI has been registered, trying to overcome the relapse problem. More cases, well-designed randomized control studies should be working on in the following years.
In conclusion, the current study which enrolled 25 patients (17 AML and 8 MDS patients) showed standard or low-dose chemotherapy plus HLA-mismatched allo-TLI were associated with a higher treatment response rate, longer survival and favorable safety in elderly AML patient and int-2/high-risk MDS patients, compared with a historical control group in our center with IA (3 + 7), and with reported data on conventional treatments. Although this pilot study only enrolled 25 patients, we analyzed the results of 79 HLA-mismatched allo-TLI procedures. Based on these data, this study provides new insights into the mechanism of HLAmismatched allo-TLI. Post-TLI CRS is common and usually mild, and a higher T-cell level after allo-TLI may be related to better blood cell recovery, while more cycles of HLA-mismatched allo-TLI may increase recipients' T-cell levels, contributing to improved survival.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Materials. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Research Ethics Committee of the First Affiliated Hospital, College of Medicine, Zhejiang University. The patients/ participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

AUTHOR CONTRIBUTIONS
JS, HH, and ZC designed the study and revised the article. YH and MH performed the trials, processed the data analysis and interpretation and drafted the manuscript. ZQ, WZ, HXH, LJL, TL, YX, SY, YZ, LZL, WH, SF, JC, KW, ML, QL, YW, FH, JCZ, JYZ, YC, and MZ performed the trials and processed the data analysis. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by the National Natural Science Foundation of China (grant no. 82070200).