Allogeneic hematopoietic stem cell transplantation in adult acute lymphoblastic leukemia: potential benefit of medium-dose etoposide conditioning

The outcomes of adult acute lymphoblastic leukemia (ALL) patients with chemotherapy or autologous hematopoietic stem cell transplantation (HSCT) are unsatisfactory. Therefore, allogeneic (allo) HSCT has been applied to those patients in first complete remission (CR1), and has shown a long-term survival rate of approximately 50 %. In terms of myeloablative conditioning (MAC) regimen, higher dose of cyclophosphamide (CY) and total body irradiation (TBI) (the standard CY + TBI) has been generally applied to allo HSCT. Other MAC regimens such as busulfan-based or etoposide-based regimens have also been used. Among those, medium-dose etoposide (ETP) in addition to the standard CY + TBI conditioning regimen appears to be promising for allo HSCT in adult ALL when transplanted in ALL patients aged under 50 years in CR1 and also in CR2, showing an excellent outcome without increasing relapse or transplant-related mortality (TRM) rates. By contrast, reduced-intensity conditioning (RIC) regimens have also been applied to adult ALL patients and favorable outcomes have been obtained; however, relapse and TRM rates remain high. Therefore, an allo HSCT conditioning regimen which deserves further study for adult ALL patients aged under 50 years in CR1 and CR2 appears to be medium-dose ETP + CY + TBI and RIC is suitable for patients aged over 50 years or for younger patients with comorbid conditions. On the contrary, new therapeutic strategies for adult ALL patients are increasingly utilized with better outcomes; namely, various tyrosine kinase inhibitors for Philadelphia chromosome (Ph)-positive ALL, human leukocyte antigen-haploidentical HSCT, and pediatric-inspired regimens for Ph-negative ALL. Therefore, the optimal treatment modality should be selected considering patient’s age, Ph-positivity, donor availability, risk classification, efficacy, and safety.


Introduction
Adult acute lymphoblastic leukemia (ALL) is a type of intractable hematological malignancies, showing a longterm survival rate of approximately 30 % with a high incidence of relapse when treated with intensive chemotherapies [30,53,70,74]. Allogeneic (allo) hematopoietic stem cell transplantation (HSCT) from a human leukocyte antigen (HLA)-matched donor has been applied to ALL patients, and showed a superior response to chemotherapy and autologous (auto) HSCT [23,70], although there is some controversy regarding the efficacy of allo HSCT for ALL patients, especially for standard-risk patients [58,70]. Although it was suggested that some high-risk patients should be transplanted in first complete remission (CR1) [58], standard-risk patients in CR1 had a more beneficial effect by allo HSCT [23]. A similar result was shown in another report that compared the efficacy of allo HSCT versus auto HSCT for standard-risk adult ALL patients in CR1, showing a 5-year (Y) disease free survival (DFS) rate of 69 % versus 45 % and a 5-Y relapse rate of 14 % versus 52 %, respectively [14]. Furthermore, a 5-Y estimated non-relapse mortality (NRM) rate was 16 % versus 2 %. These outcomes of allo HSCT for adult ALL patients are unsatisfactory. Generally, long-term survival rates of 40-60 % are obtained when both child and adult high-risk patients in CR1 or second complete remission (CR2) received allo HSCT by conventional myeloablative conditioning (MAC) regimens [27,34,35,42,70]. The outcome decreased depending on the remission status at transplantation; namely, 40-60 % in CR1, 20-40 % in CR2, 10-20 % in ≧ third complete remission (CR3), and approximately 10 % in non-CR [5,18,26,34,35,42,47]. A standard conventional MAC regimen for allo HSCT in patients with ALL is 120 mg/kg cyclophosphamide (CY) plus 12 to 13.2 gray (Gy) fractionated total body irradiation (TBI). With this regimen, a 3-Y overall survival (OS) or DFS rate of approximately 50 % was obtained when transplanted in CR1 [58,70].
By contrast, reduced-intensity conditioning (RIC) regimens have been applied to older patients and younger patients with comorbid conditions. Fludarabine (Flu)based regimens are often utilized, showing an approximately 50 % OS rate; however, the frequencies of relapse and complications following HSCT are relatively high. Similar outcomes have been observed between RIC and MAC regimens in the Philadelphia chromosome (Ph) + ALL patients who were minimal residual disease (MRD)-negative when transplanted in CR1 [2]. Therefore, RIC regimens are appropriate for patients aged ≧50 years or those <50 years with serious comorbid conditions and who are MRD-negative.
Recently, excellent outcomes have been obtained with pediatric-inspired regimens for adult Ph − ALL patients [15,16] and also with HLA-haploidentical (haplo) HSCT for ALL patients lacking an HLA-matched donor [1,12,40,51,54]. Furthermore, in Ph + ALL patients, the introduction of tyrosine kinase inhibitors (TKIs) has revolutionized the treatment paradigm of Ph + leukemia patients [28,38,39,50,74], although these strategies have some limitations. Therefore, it should be determined which type of treatment is optimal for each patient, considering patient's age, Ph-positivity, donor availability, risk, efficacy, and safety.
MAC regimens other than the CY + TBI regimen for adult ALL patients MAC regimens other than the CY + TBI regimen, such as 60 mg/kg etoposide (ETP) + 12 -13.2 Gy TBI or 16 mg/kg busulfan (BU) + 120 mg/kg CY, were also utilized in allo HSCT for hematological malignancies, including ALL (Table 1) [6, 7, 25, 37, 49, 57, 59-61, 64, 66]. A single dose or a one-day dose of ETP was generally used in ETP-based regimens. The outcome of BU (oral) + CY regimen was not superior to that of CY + TBI regimen. The outcome of ETP + TBI regimen was similar to that of CY + TBI regimen, although ETP as a substitute for CY is less immunosuppressive, resulting in more frequent engraftment failure [57]. An early phase I/II study for patients with advanced hematological malignancies of 33 patients with acute leukemia and 13 patients with ALL who were not in CR1, and 14 patients with other advanced hematological malignancies showed that a single dose of 60 mg/kg ETP was the maximal tolerated dose, as determined by relapse and survival rates (15 % and 54 %, respectively) with the dose ranging from 25 to 70 mg/kg [2].
Comparison of the medium-dose ETP + CY + TBI regimen in the Hokkaido University Hospital with that in French group We tried to examine whether more intensified conditioning regimens of adding 10-25 mg/kg/day for 2 days of ETP to the CY (60 mg/kg × 2) + TBI (2 Gy × 6) regimen could improve the outcome of patients with hematological malignancies since 1990 in the Hokkaido University Hospital. Our preliminary study in the early 1990s showed that higher ETP doses (20 and 25 mg/kg/day for 2 days) were more toxic to the patients; therefore, we fixed the ETP dose to 15 mg/kg/day for 2 days in 1993. Since then, almost all of the patients with hematological malignancies admitted to the Hokkaido University Hospital were treated with this regimen. Ten years later, we recognized that this regimen was more suitable for ALL (n = 11, the 5-Y OS: 100 %) as compared with total cases including other hematological malignancies (n = 27, the 5-Y OS: 74 %) and that remission status was important for obtaining a better outcome. A 5-Y DFS rate in CR1, CR2, and non-CR were 91 %, 83 %, and 41 %, respectively [71].
In a retrospective analysis in which 37 ALL patients in 28 CR1, 7 CR2, and 2 non-CR were treated with allo HSCT by a medium-dose ETP (15 mg/kg × 2) + CY (60 mg/kg × 2) + TBI (2 Gy × 6) regimen, a 3-Y OS, relapse, and transplant-related mortality (TRM) rate were 89 %, 8 %, and 5 %, respectively [59]. Among 10 Ph + ALL patients transplanted in CR1, all patients were alive, indicating that this regimen is markedly effective even in high-risk Ph + ALL patients. Six out of the ten patients who were polymerase chain reaction-positive for the bcr-gene rearrangement at transplantation became MRD-negative after allo HSCT. Four other patients were MRD-negative at transplantation. No patients received TKIs before and after allo HSCT.
Another ETP (40 mg/kg in 8 h: 4 consecutive infusions of 10 mg/kg lasting 2 h each) + CY + TBI regimen was reported by a French group [25]. A 6-Y OS rate in 39 ALL patients aged < 50 years or 15-60 years, who received allo BMT in CR1 was 75 % and 41 %, respectively, whereas a 6-Y OS rate was 40 % in those who received late high-dose chemotherapy followed by auto HSCT.
Major difference between ours and theirs are as follows: we used an ETP dose of 15 mg/kg/day × 2 days and they used 40 mg/kg/day × 1 day. Furthermore, they used bone marrow (BM) stem cells alone; by contrast, among 37 patients, we used BM in 86 % of the transplants, peripheral blood (PB) stem cells in 11 %, and cord blood (CB) stem cells in 3 %. The outcome seems to be no difference in each stem cell used. Our study had 69 % high-risk ALL patients, including 27 % Ph + patients [59], whereas their study had 33 % high-risk ALL patients, including 21 % Ph + patients. Although they used only an HLA-matched sibling donor, we used an HLA-matched related or unrelated donor and an HLA-mismatched related or unrelated donor.
A prospective multi-center phase II clinical trial with the medium-dose ETP + CY + TBI regimen for adult ALL patients in Japan A prospective multi-center phase II clinical trial has been conducted in Japan to validate whether the medium-dose ETP + CY + TBI regimen is an excellent conditioning regimen for ALL patients in CR1 and CR2. The sequential order of each agent was not always fixed depending on the physician's discretion. Of the 50 patients (47 CR1 and 3 CR2), 84 % were high-risk patients, including 38 % Ph + patients. A 1-Y OS, EFS, relapse, and NRM rate were 80 %, 76 %, 10 %, and 14 %, respectively [60]. Furthermore, an estimated 2-Y OS, EFS, relapse, and NRM rate were 67 %, 65 %, 15 %, and 20 %. The relapse rate in the recent study was slightly higher than the previous rate (10 % at 1 year and 15 % at 2 years versus 8 % at 3 years); it was 10 % at 6 years in the French group study [25]. Although the 1-and 2-Y NRM rates (14 % and 20 %) in the recent study were higher than the 3-Y TRM rate (6 %) in the previous one, these rate was comparable with that of the French group study (15 %) [25]. Therefore, these results strongly confirm that this regimen is remarkably effective for adult ALL patients in CR1, showing a higher OS rate without increasing TRM/NRM and relapse rates. The outcome was not influenced by the difference of donor type (related versus unrelated), age difference (≧40 years old versus < 40 years old), stem cell source (BM versus PB), Ph-positivity (positive versus negative), risk group (high versus standard), or disease status (CR1 versus CR2). TKIs were used in Ph + ALL patients only before allo HSCT. All patients were engrafted with median neutrophil recovery (≧500/μl) at day 16, similar to the result of the previous report [59]. No patients died within 100 days.
Acute and chronic graft-versus-host disease (GVHD) generally developed at rates of 20-40 % (grade ≧II: approximately 20 %) and 30-50 % (extensive type: approximately 30 %) in MAC regimens, respectively [2,6,7,57,[59][60][61]64]. The occurrence of acute GVHD in the medium-dose ETP + CY + TBI regimen appears to be more frequent (66 %). Although grade ≧II acute GVHD were observed in 58 % of transplants, uncontrollable severe acute GVHD was not observed. Chronic GVHD was observed in 56 % of patients (extensive type: 41 %), showing a slightly more frequent incidence. In Japan, we use both related and unrelated HLA-matched donors (approximately 35 % and 50 %, respectively), whereas HLA-matched sibling donors were generally selected in other ETP-based regimens in Western countries. The outcome of allo HSCT from an HLA-matched unrelated donor is almost the same as that from an HLA-matched related donor in Japan [26]. This favorable phenomenon may be due to Japanese genetic homogeneity [43,44,48,55]. Furthermore, although we use HLA-mismatched related donors in approximately 5 % of transplants and HLA-mismatched unrelated donors in approximately 10 % of transplants, no unfavorable outcome was observed.
The upper age limit for allo HSCT is thought to be 50-55 years, although we performed allo HSCT in an ALL patient aged 58 years using the medium-dose ETP + CY + TBI regimen. When the patient was older, severe complications developed more often. Therefore, the medium-dose ETP + CY + TBI regimen appears to be appropriate for patients aged < 50 years because of its toxicity. In fact, the outcomes of RIC and MAC in allo HSCT for adult ALL patients aged ≧45 years were analyzed in the adult ALL Working Group of the JSHCT, suggesting that patients aged ≧50 years should be transplanted by RIC regimens [69].
A TRM rate of only 5 % by the medium-dose ETP + CY + TBI regimen was observed in ALL patients aged 15-58 years (54 % were from < 40-years-old patients) in our previous study [59], and a NRM rate of 14 % was observed in ALL patients aged 17-49 years in our recent study [60]. Such low TRM/NRM rates in adult ALL patients are remarkable when compared with other ETP-based regimens, in which higher doses of ETP were used. Possible explanations for this difference appears to be due to the divided administration of a lower dose of ETP and relatively younger patients enrolled (54 % were < 45-years-old). Even if lower doses of ETP were used, the relapse rate remained < 10 %. Furthermore, no patients developed second malignancies after a median follow-up period of 90 months as of June, 2014. This finding is supported by the fact that the cumulative risk factor after 15 years was 1.26 % in 366 out of 1376 ALL patients who received allo HSCT with fractionated TBI + ETP 40-60 mg/kg [8].
Although CB was not used in the present phase II trial, a similar outcome between CB and other stem cells is expected with the medium-dose ETP + CY + TBI regimen as shown in several studies [20,31,52,68].
ETP-based conditioning regimens were also used in child and adolescent ALL patients [19,72]. The ETP doses were comparable with our regimen; however, most of the patients were transplanted at advanced stages and ETP was administered in a single dose, resulting in higher TRM rates.

Outcome of Ph + adult ALL patients transplanted with allogeneic HSCT
Although some Ph + ALL patients were included in each trial shown in Table 1, the outcome of allogeneic HSCT for only Ph + ALL patients is shown in Table 2 [4,17,21,22,29,62,65,75]. Ph + ALL patients who received allo HSCT after conditioning regimens excluding ETP had a poor prognosis. ETP (higher doses: 50-60 mg/kg)-based conditioning regimen improved the outcome to some extent. However, the relapse and TRM/NRM rates remained high.
On the contrary, the introduction of TKIs in combination with chemotherapies has drastically changed the outcomes of Ph + ALL patients, resulting in increased CR rate compared with historical controls. However, since the probability of relapse remains high, allo HSCT in CR1 is required to obtain a better outcome. There are several reports regarding the efficacy of imatinib (IM) use before allo HSCT with 2-Y -5-Y OS of 40-78 % [28,33,38,50,74].
In terms of IM use in ALL patients after allo HSCT, a 5-Y DFS in prophylactic arm and in MRDtriggered arm was 84 % and 60 % (p = 0.89), respecticvely and a 5-Y OS was 80 % and 75 % (p = 0.84), respectively [50]. Although prophylactic use of IM significantly reduced the incidence of molecular recurrence after allo HSCT compared with MRDtriggered use of IM (40 % versus 69 %, p = 0.046), the outcome was shown to be no difference between both arms. Relapse probability was significantly higher in patients who became MRD-positive (p = 0.017).
By contrast, neither TKI use pre-nor post-allo HSCT was found to significantly impact outcomes [28]. However, this is a retrospective study and there are some limitations regarding the heterogeneity in conditioning regimens, allotype, and stem cell source, as well as the small sample size.
Therefore, the outcome of chemotherapy combined with TKIs followed by allo HSCT in CR1 with MAC regimens other than ETP-based regimens appears to be not always superior to that of allo HSCT performed by using the medium-dose ETP + CY + TBI regimen. It is meaningful to evaluate whether administration of other TKIs pre-or post-allo HSCT is beneficial for the outcome of Ph + ALL patients by the prospective clinical trial in the future.

Importance of administration schedule of ETP
The administration schedule of ETP is an important factor for reducing adverse effects, although it depends on combined chemotherapy and/or TBI. We used a 3-h infusion on 2 consecutive days with 30 mg/kg of ETP, whereas 60 mg/kg of ETP was usually administered in a single dose of a 4-h infusion [2,6,7,64]. The maximally tolerated dose (MTD) is thought to be 2.4 g/m 2 in 1 to 1.5-h infusions on 3 consecutive days, with 4.2 g/m 2 as continuous infusion (29 to 69-h) [9,24]. When ETP, CY, and TBI were combined, the MTDs were 1.8 g/m 2 for ETP, and 180 mg/kg for CY.

Possible mechanisms of efficacy of the medium-dose ETP + CY + TBI
It is important to clarify how the medium-dose ETP + CY + TBI regimen can eradicate residual leukemic cells more efficiently than the CY + TBI regimen or others. The graft-versus-leukemia (GVL) effect is hardly observed against ALL; however, allo HSCT is superior to auto HSCT in terms of the outcome [18,23,70]. Furthermore, chronic GVHD appears to induce a GVL effect in ALL patients [32], although others could not confirm a decreased relapse risk in the higher incidence of chronic GVHD among the recipients of allo PB [10,56].
ETP upregulated the expression levels of interleukin (IL)-8 and macrophage inflammatory protein-1α in promonocytic acute myelogenous leukemia cells, most likely inducing an enhancement of the anti-leukemic effect [45]. A low dose of ETP can enhance leukemia rejection in mice by rendering them immunogenic and susceptible to CD8 + T cell deaths [63]. Several immunomodulations by ETP are also observed in other tumor cells [73]. A preliminary in vitro study showed that CY synergized with ETP in terms of cytotoxic effect against leukemic cells (unpublished observation). Indeed, it is known that ETP exhibits a synergy with 4-hydroperxycyclophosphamide against a promyelocytic leukemic cell line determined by in vitro     cytotoxicity assay [11]. Therefore, some of these mechanisms may be responsible for the superiority of the medium-dose ETP + CY + TBI regimen in ALL patients.
Reduced-intensity conditioning regimens for adult ALL patients As shown in Table 3, various RIC regimens for allo HSCT were applied for adult ALL patients [3,13,35,36,41,67]. The stem cell source was mainly either BM or PB from an HLA-matched or mismatched related and an HLA-matched or mismatched unrelated donor. The 2-Y or 3-Y OS were 30-60 %. A better outcome was observed in patients transplanted in CR1 compared with those transplanted in more advanced stages. The relapse rates were 20-60 %. The TRM/NRM rates were 20-40 %. The occurrence of acute GVHD (≧grade II) and chronic GVHD was 40-60 % and 30-70 % (extensive type: 30-50 %), respectively. GVL effect against ALL could be induced by RIC regimens and allo HSCT with RIC regimens was feasible for patients with high-risk ALL patients in remission [36]. It is notable that the relapse rate in the RIC regimen for adult Ph − ALL patients in Japan was higher (26 % vs 10 %) than that in the medium-dose ETP + CY + TBI regimen [46,60].

Pediatric-inspired regimens for adult Ph − ALL patients
Pediatric-inspired regimens resulted in a better outcome for adult Ph − ALL patients aged 15-55 years regardless of undergoing allo HSCT in CR1 [16]. Although CY + TBI conditioning regimen was used for allo HSCT, a 3-Y relapse rate, NRM, and relapse-free survival (RFS) were 20 %, 16 %, and 65 %, respectively. Of the allo HSCT patients, a 3 Y-NRM in the patients aged 45-55 years and those aged 15-44 years was 26 % and 13 % (p < 0.048), respectively. A 3-Y RFS in the younger and the older was 67 % and 56 % (p < 0.12), respectively and a 3-Y OS was 72 % and 61 % (p < 0.082), respectively. Almost comparable outcomes were observed in patients with allo HSCT and those without allo HSCT, although the survival in patients with MRD (≧1 −3 ) at 6 weeks after induction initiation was longer in the allo HSCT cohort than in the no HSCT cohort. These results indicate standard risk ALL patients aged < 40-45 years with MRD (<1 −3 ) at 6 weeks after induction initiation may not require allo HSCT in CR1. A similar result has been also shown in other study [15]. In the future, it should be determined whether the medium-dose ETP + CY + TBI regimen is suitable for adult standard risk ALL patients aged < 40-45 years based on MRD analysis.

HLA-haploidentical hematopoietic stem cell transplantation for adult ALL patients
Recently, haplo HSCT, which is varied in the settings, has been increasingly performed [51]. Basically, this procedure is still alternative for the patients lacking an HLA-matched related donor, since the outcomes are approximately 20-50 % and limited numbers of ALL cases are reported. In patients with ALL, AML, and other hematological malignancies transplanted with mostly T cell-depleted PB, DFS, relapse rate, and TRM rate were 20-50 %, 20-60 %, and 30-40 %, respectively. The outcomes in ALL patients were generally inferior to those in AML patients [1].
The same group reported that the outcome of haplo HSCT for Ph + ALL patients (n = 101) is comparable to that of HSCT (n = 38) from an HLA-matched related donor. A 5-Y DFS, OS, relapse, and NRM rates in the haplo HSCT group were 66 %, 74 %, 20 %, and 16 %, respectively. Acute (grade ≧II) GVHD was higher (32 % versus 16 %, =0.045) in the halpo HSCT than in the control group, although there was no difference between both groups in chronic GVHD frequency (extensive type: 20 % versus 13 %) [12].
Ruggeri et al. [54] reported that no statistically significant differences were observed between haplo HSCT and CB transplantation for a 5-Y relapse (approximately 40 %), NRM (approximately 30 %), and LFS (approximately 30 %). The outcomes were not so excellent probably due to the inclusion of many patients not in CR1. However, this strategy is valid for ALL patients lacking an HLA-matched donor.

Conclusions
In allo HSCT, the medium-dose ETP + CY + TBI regimen is promising for high-risk ALL patients aged < 50 years in CR1 and also for standard-risk patients in CR2, resulting in an excellent outcome without higher TRM and relapse rates. However, the enrolled case for the prospective multi-center phase II clinical trial with the medium-dose ETP + CY + TBI regimen for adult ALL patients in Japan is still small and a prospective randomized phase III clinical trial is required for determining its genuine efficacy in the future. On the contrary, RIC for adult ALL patients is preferable to older patients aged ≧50 years or younger patients with serious comorbid conditions and who are MRD-negative. Furthermore, recently developed various therapeutic strategies such as haplo-HSCT, pediatric-inspired regimens, and use of TKIs at pre-or post-allo HSCT for adult ALL patients should be taken into consideration to obtain a better clinical outcome (Table 4).