SIL-TAL1 Rearrangement is Related with Poor Outcome: A Study from a Chinese Institution

SIL-TAL1 rearrangement is common in T-cell acute lymphoblastic leukemia (T-ALL), however its prognostic implication remains controversial. To investigate the clinical characteristics and outcome of this subtype in Chinese population, we systemically reviewed 62 patients with newly diagnosed T-ALL, including 15 patients with SIL-TAL1 rearrangement. We found that SIL-TAL1+ T-ALL was characterized by higher white blood cell count (P = 0.029) at diagnosis, predominant cortical T-ALL immunophenotype (P = 0.028) of the leukemic blasts, and a higher prevalence of tumor lysis syndrome (TLS, P<0.001) and disseminated intravascular coagulation (DIC, P<0.001), which led to a higher early mortality (P = 0.011). Compared with SIL-TAL1− patients, SIL-TAL1+ patients had shorter relapse free survival (P = 0.007) and overall survival (P = 0.002). Our NOD/SCID xenotransplantation model also demonstrated that SIL-TAL1+ mice models had earlier disease onset, higher leukemia cell load in peripheral blood and shorter overall survival (P<0.001). Moreover, the SIL-TAL1+ mice models exerted a tendency of TLS/DIC and seemed vulnerable towards chemotherapy, which further simulated our clinical settings. These data demonstrate that SIL-TAL1 rearrangement identifies a distinct subtype with inferior outcome which could allow for individual therapeutic stratification for T-ALL patients.

Introduction T-cell acute lymphoblastic leukemia (T-ALL) accounts for approximately 15% and 25% of the newly diagnosed ALL in children and adults, respectively, and is often linked with a poor prognosis. T-cell acute lymphoblastic leukemia 1 (TAL1, also known as stem cell leukemia, SCL) is a frequent target for chromosomal translocation, interstitial deletion, or mutation in T-ALL [1,2]. It encodes a class II basic-helix-loop-helix (bHLH) transcription factor that is a master regulator for hematopoietic lineage commitment. TAL1 has been proved essential for the development of mouse HSCs [3]. SIL-TAL1 rearrangement is a common TAL1 related alteration and occurs in 16,26% of T-ALL cases. It results from a 90 kb-interstitial-deletion in the TAL1 gene locus that fuses with the 5 non-coding portion of SIL, thus leading to an aberrant overexpression of TAL1 protein [4,5].
Patients bearing SIL-TAL1 rearrangement (SIL-TAL1 + ) are defined by distinct clinical and biological characteristics such as a high white-blood-cell count and hemoglobin, T-lineage immunophenotype with CD2 expression, and low incidence in adult patients [6][7][8]. Some of these features are generally associated with more unfavorable clinical features. However, the prognosis of this rearrangement remains controversial [6,7,[9][10][11][12]. TAL1 rearrangements had been historically linked with a better outcome in some reports [7,10,12]. With respect to SIL-TAL1, the results from Mansur MB indicated a negative impact on overall survival (OS) in Brazilian pediatric patients [6], whereas Ballerini P suggested that SIL-TAL1 expression did not significantly affect either leukemia-free survival (LFS) or OS [11]. Therefore, the clinical features and prognostic significance of SIL-TAL1 rearrangement deserves further study.
In the present study, we retrospectively reviewed a serial of 62 patients diagnosed with T-ALL in our department. Clinical characteristics and outcome was compared for patient subgroup with and without SIL-TAL1 rearrangement. Moreover, we established several reliable T-ALL xenograft murine models, which would further demonstrate the disease phenotype and responses to chemotherapy of this distinct T-ALL subtype.

Clinical Comparison of T-ALL Patients
We systemically collected clinical data from 62 de novo T-ALL patients hospitalized in our department from 2006 to 2012. The bone marrow specimen of each patient was obtained before treatment and the diagnosis was made based on a multiparametric approach, including examination of clinical characteristics, morphologic features, immunophenotype, and cytogenetic and molecular findings [13]. SIL-TAL1 fusion transcript was detected using reverse transcription polymerase chain reaction (RT-PCR) [14]. The  PCR  primer  pairs  were:  F  59-TCCCGCTCCTACCCTGCAA-39 and R 59-CGTCGCGGCCCTTTAAGTC-39 [14]. The evaluation criteria for tumor lysis syndrome (TLS) and disseminated intravascular coagulation (DIC) was as described by Tosi et al [15] and Taylor et al [16], respectively. Clinical characteristics and outcome was compared for patients with and without SIL-TAL1 rearrangement. Complete remission (CR) was defined as bone marrow morphology with less than 5% blasts, a neutrophil count of 1610 9 /L or more, a platelet count of 100610 9 /L or more, and no evidence of extramedullary leukemia. Overall survival (OS) was calculated from the first day of induction therapy to their death or the last day of observation. Relapse free survival (RFS) was calculated from the initiation of induction therapy.

Comparison of SIL-TAL1 + and SIL-TAL1 2 T-ALL in Xenograft Models
We had inoculated leukemia cells from four T-ALL patients, including one SIL-TAL1 + patient, to NOD/SCID mice [17]. Clinical data of the four patients and detailed protocols for xenotransplantation and evaluation had been described previously [17]. In this study, data obtained from the secondary passage (P2) of these murine models were reanalyzed (Table S1). Comparison of disease phenotype and survival was made between the mice with and without SIL-TAL1 rearrangement. The disease onset was defined as the first time human CD45 + cells detected in peripheral blood. Overall survival of the mice was calculated from disease onset till the death.

In vivo Drug Treatment of SIL-TAL1 + Murine Models
In this study, we serial transplanted leukemia cells from the secondary passage of SIL-TAL1 + model to NOD/SCID mice for in vivo drug treatment. The detailed protocol was described as before [17,18]. Briefly, 21 mice aged 4 to 6 weeks were pretreated intraperitoneally with anti-mouse CD122 monoclonal antibody (TM-b1, Rat IgG2b, Bio X cell, USA). After conditioning, each mouse was inoculated via tail vein with 1.0610 7 thawed leukemia cells that were suspended in 300 ml PBS. After inoculation, engraftment was monitored by weekly blood collections and FACS analysis. When human CD45 + (hCD45 + ) cell proportion in peripheral blood reached 1%, 21 mice were randomly grouped into treatment (vincristine or dexamethasone) and control groups (seven mice for each group). The treatment groups received vincristine (0.5 mg/kg every 7 days for 4 weeks), or dexamethasone (20 mg/kg Monday-Friday for 4 weeks) through intraperitoneal administration [17,19]. The control group received equal volume of PBS. Since treatment, the percentage of human CD45 + cells in the peripheral blood was monitored weekly. The end point was defined as the first indication of morbidity (.20% weight loss, lethargy and ruffled fur). The survival was calculated from the initiation of inoculation until the end point. To confirm the engraftment of leukemia cells in NOD/SCID mice at the end point, immunophenotypical and genetic characteristics were inspected by flow cytometry, immunohistochemistry and RT-PCR. The procedures were described as before [17,18].

Assessment of Serum Biochemical and Coagulation Parameters in Murine Models
The leukemia cells from the secondary passage of one SIL-TAL1 + and two SIL-TAL1 2 T-ALL models (SIL-TAL1 2 a and b) were inoculated to NOD/SCID mice as described above. When hCD45 + cell proportion in peripheral blood reached 1%, the mice were randomly grouped into treatment (vincristine or dexamethasone) and PBS control groups (six mice for each group), and then received drug treatment as described above. At the end point, blood serum of the mice was collected for routine serum biochemical and coagulation test. Serum samples from six healthy NOD/SCID mice were also tested to provide normal baseline.

Statistical Analysis
The analysis of categorical variables was performed using Fisher's exact test for 262 tables. Student's t test was applied to continuous variables. All the survival data were compared using log-rank test, and graphically represent by Kaplan and Meier analysis. All calculations were performed using the SPSS software version 16.0 (SPSS, Chicago, IL).

TLS and DIC were Associated with Early Death in SIL-TAL1 + Patients
As SIL-TAL1 + patients had a higher incidence of TLS (9/15, 60%) and DIC (11/15, 73.3%) ( Table 1), we evaluated each of the 15 SIL-TAL1 + patients for DIC and TLS grading in Table S2 according to the diagnostic criteria mentioned before [15,16]. For the 9 patients developed TLS, 4 patients were classified as grade I and the rest 5 were in grade II. The onset of TLS and DIC was shown in Figure 1. Three patients had TLS at diagnosis; the other 6 patients developed TLS when pre-treatment or induction therapy began. Though we had more DIC cases, the occurrence of DIC was almost parallel with TLS except for the earlier onset in some cases, with 5 occurred at diagnosis, 2 at pre-treatment stage and 4 after induction therapy. Moreover, we noticed a high mortality (4/15, 26.7%) in SIL-TAL1 + patients before induction therapy. All of the 4 patients had TLS, while 3 of them had DIC and died of severe hemorrhage. On the contrary, only 2 patients died before induction in SIL-TAL1 2 group (2/47, 4.3%). The difference was statistically significant (P = 0.011).

Treatment and Outcome of Patients
After pre-treatment and induction therapy, most patients had well responses. The complete remission (CR) rate was 90.9% (10/ 11) for the SIL-TAL1 + patients and was 91.1% (41/45) for SIL-TAL1 2 patients (P = 0.983). However, all the SIL-TAL1 + patients relapsed in a period no longer than 4 months. The median RFS was only 2 months ( Figure 2A). The overall survival was also poor. Four patients died within 1 month after relapse. The median OS of SIL-TAL1 + patients was 4 months ( Figure 2B). In comparison, the SIL-TAL1 2 patients had significantly longer RFS (Figure 2A, median 12 month, P = 0.007) and OS ( Figure 2B, median 25 month, P = 0.002) than SIL-TAL1 + patients.

Disease Phenotype and Outcome of the SIL-TAL1 + Murine Models
The mice transplanted with SIL-TAL1 + T-ALL cells (SIL-TAL1 + mice) had earlier disease onset (1460 days) than those mice transplanted with SIL-TAL1 2 T-ALL cells (SIL-TAL1 2 mice) (28.9613.4 days, Figure 3A, P,0.001). At the end point, engraftment and infiltration of leukemia cells was detected in multiple organs, which mimicked the clinical characteristics of SIL-TAL1 + T-ALL ( Figure S1). The percentage of hCD45 + cells in peripheral blood were significantly higher in SIL-TAL1 + mice than those in SIL-TAL 2 mice (96.361.5% vs. 53.0621.8%, Figure 3B, P,0.001). As the OS of each model was compared, we found that the 6 SIL-TAL1 + mice have significant shorter survival than the other 18 SIL-TAL1 2 mice ( Figure 3C, P,0.001).

Drug Treatment in SIL-TAL1 + Murine Model
We used tertiary passage of mice for drug treatment. As shown in Figure 4A, for vincristine group, the percentage of hCD45 + cells could be controlled at no more than 80%, while dexamethasone hardly had effect on suppressing tumor growth. We then compared the survival of mice receiving different treatment strategies. Interestingly, the survival of dexamethasone group (median 35 days) was longer than that of vincristine treated group (median 26 day, P = 0.002, Figure 4B), even though the dexamethasone group had higher hCD45 + cell level. In the vincristine group, instead, when the hCD45 + cell reached near 80%, all mice died soon after vincristine administration. As a result, the vincristine group had slightly longer survival than saline control group (24 days, P = 0.033, Figure 4B).

SIL-TAL1 + Murine Model Had more Evidence of TLS and DIC
Since we found that SIL-TAL1 + patients were associated with TLS and DIC, as well as the paradoxical survival of drug treated SIL-TAL1 + murine models, we carried out the serum biochemical and coagulation test on the drug treated murine models for further confirmation. As shown in table 2 (calculated from Table S3), compared to the SIL-TAL1 2 murine models, the SIL-TAL1 + murine models had significantly higher LDH (p = 0.009), potassium (p,0.001) and uric acid levels (p,0.001), which revealed that the SIL-TAL1 + models tended to develop TLS. Similarly for the coagulation tests, the SIL-TAL1 + murine models had significantly prolonged prothrombin time (PT, p,0.001) and activated partial thromboplastin time (APTT, p,0.001), lower fibrinogen level (p,0.001) and marginal elevated D-dimer level (p = 0.076), indicating a disregulated coagulation status in these mice. Besides, in the SIL-TAL1 + murine models, the vincristine treatment group had longer PT and APTT, and higher D-dimer and fibrinogen degradation product (FDP) level than the dexamethasone treatment group (all p,0.001), demonstrating that vincristine treatment was more likely to induce DIC.

Discussion
SIL-TAL1 rearrangement is a common form among heterogeneous TAL1 related alterations. Only a few clinical researches have been hitherto performed to describe the features of SIL-TAL1 + T-ALL. In the present study, we analyzed the characteristics of SIL-TAL1 + T-ALL retrospectively in 15 Chinese patients presented to our department, and made a comparison with the other 47 SIL-TAL1 2 T-ALL patients administered during the same period. The  The percentage of human CD45 + cells in mice peripheral blood was 96.361.5% for SIL-TAL1 + models and 53.0621.8% for SIL-TAL1 2 models (P,0.001). (C) The median survival for SIL-TAL1 + models was 5 days, which was shorter than any of the three SIL-TAL1 2 models: a (median 13 days, P = 0.001), b (median 43 days, P = 0.001), and c (median 16 days, P = 0.001). There were significant differences between the four groups (P,0.001). doi:10.1371/journal.pone.0073865.g003 SIL-TAL1 + T-ALL Has Poor Outcome PLOS ONE | www.plosone.org incidence of SIL-TAL1 + T-ALL in our study was comparable with previous literatures [4][5][6][7][8][10][11][12]. We showed that the SIL-TAL1 + patients were linked with relatively younger age, higher WBC count, more cortical immunophenotype, and a poor outcome. These clinical features were similar to the Brazilian report [6]. However, we noticed that a large proportion of de novo patients with SIL-TAL1 developed TLS and DIC during, or even before chemotherapy, and resulted in a higher mortality in the early stage, which has also been simulated by our SIL-TAL1 + murine model. These features have not yet been observed in any form of TAL1 related alterations. Higher tumor burden might account for these features, which was indicated by extremely higher WBC count and elevated LDH level than Brazilian reported in the peripheral blood.
Hyperleukocytic leukemia is conventionally defined as leukemia with an initial WBC count or blast count greater than 100610 9 /L and associated with a very high early mortality rate (20,40%) [20]. Symptoms of hyperleukocytosis are primarily due to leukostasis and subsequent endothelial damage [20]. The clinical implication of hyperleukocytosis of ALL is different from those in  Table 2. Comparison of serum biochemical and coagulation parameters between SIL-TAL1 + and SIL-TAL1 2 murine models.  acute myeloid leukemia (AML) [21,22]. Eguiguren et al. reported that in pediatric ALL, hyperleukocytosis is significantly associated with age of less than 1 year, T-cell lineage, leukemic cell ploidy ,50 chromosomes, presence of a mediastinal mass, and central nervous system leukemia at diagnosis [23], some of which were also presented in our study. Despite a higher WBC count in ALL than AML, leukostasis is rarely seen in ALL (2% to 6%), and management focuses primarily on the treatment of TLS and DIC [22]. TLS is an emergency of malignant tumors caused by rapid degradation of tumor cells. WBC count over 100610 9 /L and LDH over 2 times of upper limit of normal level are considered of high risk to develop TLS [24]. Based on the risk stratification criteria, most of our patients with SIL-TAL1 belonged to the high risk group, thus we were not astonished to find that two thirds of these patients developed TLS. Usually TLS is treatment related, while in some cases with high tumor load, TLS may occur spontaneously [24]. We did observe three patients with SIL-TAL1 developed TLS before chemotherapy, but most TLS were controllable, since all the TLS cases were within grade I or II, which were not life-threatening.
Coagulopathy was another characteristic manifestation in our study. We discovered 11 patients with SIL-TAL1 developing overt DIC in different degrees according to the scoring system given by International Society on Thrombosis and Hemostasis (ISTH). In previous literatures, the DIC was found in about 10% of adult ALL patients, and less than 5% in childhood ALL [25,26]. In our study, the incidence of DIC was much higher in patients carrying SIL-TAL1, and most DIC accompanying with TLS occurred in the early stage of induction therapy, which indicated that DIC may also be related to the breaking down of leukemia cells during chemotherapy. Meanwhile, three SIL-TAL1 + patients died from severe hemorrhage, prompting DIC as a more lethal event than TLS. Among the three patients, one had DIC at diagnosis, the other two patients developed DIC during or soon after leukapheresis and died within one day, in whom coagulopathy might be accelerated by the mechanical destruction from leukapheresis.
SIL-TAL1 + ALL had a poor outcome, which was indicated by the high relapse rate and the short RFS and OS of our patients. The treatment outcome of relapsed ALL is highly heterogeneous. From several large cohort studies in children or adults, the prognosis of recurrent ALL was poor [27,28]. In adults, relapsed patients with young age and long duration of the first remission were predicted to have a good outcome after salvage therapy; otherwise, whatever the prior treatment, relapsed patients cannot be rescued using currently available regimens [28]. Our data showed that most of the SIL-TAL1 + patients cannot survive for 1 month after relapse. These observations suggested that although treatment outcome in patients with T-ALL has improved in recent years, patients with relapsed disease continue to have a dismal outcome and the prognosis of SIL-TAL1 + ALL was probably poor.
With respect to the murine models, our SIL-TAL1 + xenograft model, derived from a seven years old boy, also exhibited rapid increase of blast cells in peripheral blood and overt multiple organ infiltration, as well as a poor survival, suggesting that SIL-TAL1 + T-ALL was highly proliferative and aggressive. Glucocorticoids have long been used in the treatment of ALL and glucocorticoid resistance is an adverse prognostic factor in childhood ALL [29,30]. In our in vivo drug treatment, dexamethasone could hardly retard the progression of engraftment. Given the data from ALL-BFM 95 trial, which indicated that the prognosis was poor in childhood T-ALL with prednisone resistance [31], this boy may be linked to a dismal outcome. Another finding was that the survival of our murine model was not improved by vincristine, although it could suppress tumor growth. Death of mice soon after vincristine administration was similar to the clinical occurrence of DIC, and this presumption was further confirmed by the coagulation test on the mice. The paradox that dexamethasone group had longer survival may be resulted from the protective effect of dexamethasone on endothelial cells against leukostasis.
On the basis of our clinical observations and murine models, our study revealed that SIL-TAL1 + ALL represented a distinct ALL subtype, which was characterized by a higher WBC count and a predominant cortical T-ALL immunophenotype. Moreover, SIL-TAL1 + ALL tended to develop TLS and DIC and related with inferior outcome. These findings could aid the individual therapeutic stratification for T-ALL patients and attract more attention for the treatment of original disease. Besides, our SIL-TAL1 + xenotransplantation model could be used as model of TLS and DIC for drug assessment. However, the present study was retrospective and based on a relatively small number of patients from single clinical center and a few murine models. Therefore, expanded multicenter clinical trial and experimental.