Implications of basic research in clinical practice: toward a personalized medicine in T-cell Acute Lymphoblastic Leukemia (T-ALL)

Acute Lymphoblastic Leukemia (ALL) is the most common cancer in children and one of the main causes of death among childhood blood disorders. There are two subtypes according to the affected lymphoid progenitor: B-ALL and T-ALL. The T-ALL is the less common and historically was associated with poor prognosis in both adults and children, although at present, treatment outcomes do not differ significantly between the two types of ALL. The T-ALL subtype is the most complex and heterogeneous at the genetic level and currently the one with less new therapeutic alternatives available. This trend is changing thanks to the remarkable progress that is being made in understanding the biology involved. Advances in genomic research during last decade have largely contributed to this progress. Moreover many efforts are being made to identify which of this new basic data is relevant for clinical practice. This will allow us to better define the risk and take decisions on the best treatment to apply to each patient. Therefore we are moving towards a personalized patient management that ultimately will result in an increase in survival and progress to T-ALL cure. This review summarizes the most relevant and applicable biological findings in T-ALL made in recent years and their therapeutic implications that will influence the clinical practice in the future. *Corresponding author: Dra. Eulàlia Genescà ALL Research Group-IJC Carretera de Canyet s/n, Camí de les Escoles. Edifici IMPPC.08916, Badalona (Barcelona), Spain, Tel: +(93) 557 28 08, Fax: + (93) 465 14 72, E-mail : egenesca@carrerasresearch.org Received October 07, 2015; Accepted October 26, 2015; Published November 02, 2015 Citation: Berbis G, Ribera J, Ribera JM, Genescà E (2015) Implications of basic research in clinical practice: toward a personalized medicine in T-cell Acute Lymphoblastic Leukemia (T-ALL). Mol Biol 4: 142. doi:10.4172/2168-9547.1000142 Copyright: © 2015 Berbis G, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
Acute lymphoblastic leukemia (ALL) is characterized by an oncogenic multistep process leading to the blockage of maturation and malignant transformation of lymphoid progenitors [1]. The incidence of ALL is not homogeneous throughout life; it has an early peak at 4 or 5 years (incidence of 4-5 cases per 100,000 people/ year), a decrease in incidence in young adults and a slight increase after 50 years (incidence rate of up to 2 cases per 100,000 person/ year). The cure rate is lower in adults than in children, with a long term disease-free survival (DFS) of approximately 80% in children and only 35-45% in adults (www. seer.cancer.gov/statistics). Specifically, T-ALL corresponds to 15% of pediatric acute leukemias and 25% of the adult cases and survival does not differ from B-precursor ALL. This subtype is characterized by a high level of genetic heterogeneity and complexity making it attractive for basic research, despite being the rare ALL subtype.
The aim of this review is to summarize the latest scientific advances in T-ALL, both in basic and clinic research highlighting the data that will have an impact in clinical practice. We will try to emphasize where the research in this subtype of ALL is moving forward in order to identify which are the key issues to be solved in the medium to long term to improve the treatment and survival of patients with this disease.

High through-put techniques for detailed genetic information on T-ALL patients Copy Number Alterations (CNAs) in T-ALL
In recent years the work based on the use of the large scale and high resolution genomic techniques has been of vital importance for the understanding of ALL and especially revealing in T-ALL. Since Irving et al. were successfully able to identify copy neutral losses of heterozygosity (CN-LOH, also known as LOH) in the leukemic cells in 2005, using single nucleotide polymorphism (SNP) arrays [2], many other studies have been published involving new genes in the development of the T-ALL [3]. Among others, focal copy number alterations involving deregulation of TAL1, LMO2, PTEN, FBXW7 and MYB, have been identified [4]. Tables 1 and 2 summarize findings in CNAs in T-ALL.

Expression Arrays
The use of expression arrays has also contributed to an increase in understanding of leukemia development, progression and relapse. When comparing the expression pattern of human T-ALL samples and normal T cells at different stages of differentiation, major clusters of T-ALL are obtained that significantly correlate with the immunologic and oncogenetic expression annotation. This has allowed the redefinition of a limited number of T-ALL subgroups: a) Immature Subtype (that includes Early T-cell Precursor, ETP leukemias), characterized by the absence of CD1a, CD4 and CD8 immunomarkers and the presence of stem cells or myeloid markers such as CD117, CD34, HLA-DR, CD13, CD33, CD11b, or CD65 [5][6], and the expression of high levels of LYL1 gene [6]; b) Cortical Subtype, characterized by aberrant expression of the transcription factor family members with homeobox domain such as TLX1, TLX3, and HOXA genes [6], and the expression of CD1, CD4 and CD8 immunomarkers [6] and the presence of Pre-T1 or T2/ Pre-T3 (E) (7); c) Mature Subtype, characterized by the expression of the TAL oncogene [6], CD4, CD8 and CD3 immunomarkers [6] and Tαβ cell receptor [7] Table 1 summarizes the main chromosome abnormalities that characterize these three molecular subtypes. However, some exceptions in these correlations must be noted, highlighting the genetic diversity of T-ALL and the involvement of the affected genes in different steps of T-cell differentiation. This is the case for TLX3 gene expression, which is also expressed in leukemias that cluster within the immature subgroup in addition to the cortical subgroup [7]; and the LIM-only domain gene (LMO2), whose expression can be found either in mature or immature leukemias [6].

PHF6
Inactivating mutation ; deletion 16(c)/38(a) (78) IKZF1 and EP300. They also found mutations in the histone modifier genes (EZH2, EED, SUZ12, SETD2 and EP300). It is of note that the mutational spectrum identified in the ETP leukemia was similar to that in acute myeloid leukemia (AML) with poor prognosis, with affected pluripotent genes that define the character of the myeloid cell [8].
Recently, a study of whole exome sequencing in adult ETP leukemias has identified mutations in the DNMT3A methyltransferase gene in 16% of patients (11/68), with a similar frequency to that detected in AML. Other mutations were found in FAT1 (25%, 17/68) and FAT2 (20%, 14/68) cadherins, involving important genes for cell adhesion and interaction between leukemic cells and microenvironment in T-ALL development and/or progression [9]. Using the same approach, De Keersmaecker et al. identified mutations in the CNOT3 gene (3.8%, 8/211). This mutation causes decreased expression of the gene, suggesting that CNOT3 could act as a tumor suppressor in T-ALL. In addition, mutations in RPL10 (5.2%, 11/211) and RPL5 (1.9%, 4/211) genes were identified. Both genes encode ribosomal proteins that are part of the 60S ribosomal complex. The most frequently observed mutation in RPL10 was an amino acid change (Arg98Ser) that caused increased proliferation and a defect in ribosome biogenesis in lymphoid cells [10] The authors suggest that these mutations could help leukemic cells to decrease their hyperactive translational activity, which could be beneficial for the fitness of the blasts. Identification of mutations affecting the translational ability of leukemic cells is a new and a very exciting field of research to explore.

A new actor in the T-ALL scenario: the MicroRNAs
Advances in technology have not only contributed to identifying tumor suppressor genes and oncogenes but have also revealed that noncoding genes such as MicroRNAs (miRNA) are deregulated in almost all types of human cancer [11]. MiRNAs are one of the most abundant regulatory molecules in multicellular organisms and influence the expression of many genes. Therefore it is not surprising that they also play a key role in T-cell differentiation [12] and T-cell leukemia development [13]. The miR-17-92 cluster and specifically the miR-19b, was the first miRNA described to be overexpressed in human T-ALL samples [14]. Later, the existence of a translocation in the TCR/D locus affecting NOTCH1 on one side (through the t(9;14)(q34;q11)) and the miR-17-92 cluster (through t(13;14)(q32;q11)) was described by Mavrakis et al. [15]. They demonstrated that the coexpression of the active form of NOTCH1 and the miR-19 in mouse hematopoietic progenitors accelerated NOTCH1 induced leukemia in mice. It was also observed that BIM, PTEN, PRKAA1, DOCK5 and PPP2R5E genes were regulated by the miR-19 [15]. The first screening of miRNA in human T-ALL samples was published in 2011. Five miRNA (miR-19b, -20a, -26a, -92 and -223) that accelerate leukemia in a mouse model of T-ALL were identified. These miRNAs act cooperatively inhibiting tumor suppressor genes such as PTEN, BIM, and FBXW7 [16] After this, the Baldus group observed overexpression of the miR-221 and the miR-222 and underexpression of the 151-3p, -19a, -20b, -342-3p, -363 and 576-3p miRNAs in adult ETP ALL samples when analyzed by microarrays [17] They showed that overexpression of miR-222 decreases the messenger RNA levels of the ETS1 gene, a regulator of macrophage differentiation. This group also noted that high expression of miRNAs were positively associated with the expression of genes with prognostic impact on the ETP leukemia [17]. Recently, it has been reported that the expression of the miR-150, one of the miRNAs widely studied in B cells, is increased in CD4 + or CD8 + T-cells and decreased in CD4 + -CD8 + T cells (DP T-cells). The miR-150 also regulates NOTCH3 expression in both DP and CD4+ or CD8+ T cells. The authors have suggested a possible role of this miRNA in T-ALL, since it regulates proliferation and apoptosis in T-ALL cell lines [18]. In the case of the mature T-ALL subtype, it has been observed that the expression of the TAL1 oncogene leads to overexpression of miR-223 and the consequent deregulation of this miRNA's targets such as FBXW7, a key tumor suppressor in T-cell leukemias [19]. Finally miR-142-3p, which is overexpressed in T-ALL (16), has also been involved in a possible mechanism of glucocorticoid resistance, since this miRNA is able to regulate the translation of the α glucocorticoid receptor isoform α (GRα) [20]. In table 3 are summarized the miRNAs involved in T-ALL.
More recently, long non coding RNAs (lncRNAs), RNAs defined as transcripts with a length of at least 200 nucleotides that lack proteincoding potential and have evolutionary conservation [21], have also been shown to be involved in T-ALL development [22]. Again, high throughput techniques have introduced a new player in the T-ALL field and add an additional level of complexity to the disease, which needs to be further explored.

Adult T-ALL and pediatric T-ALL
As the number of genes known to be involved in T-ALL increases, and the number of samples tested and analyzed also increases, adult and pediatric T-ALL are emerging as two distinct molecular entities. In the extensive sequencing study done by the group of Cools, 39 matched diagnosis and remission samples (19 pediatric cases and 20 adult cases) were analyzed, it was observed that the number of mutations in adults was 2.5 times higher than in children (21.0 versus 8.2) with a clear correlation between the age and the number of mutations. However, this fact did not imply any prognostic significance. Likewise, it was observed that adults mostly had mutations in regulatory genes affecting the methylation pattern such as KDM6A (4.5%, 3/67) and MAGEC3 (3.0%, 2/67), transcription regulators like PHF6 (17.9%, 12/67) and tumor suppressor genes as CNOT3 and FBXW7 (11.9%, 8/67) [10]. Furthermore, mutations in proteins that are part of the 60S ribosomal complex were identified to be child specific [10]. All these data suggest that the pathogenesis of T-ALL in adults and children could arise from a different origin with different key signaling pathways affected. While in the adult alterations in epigenetic factors would have a decisive role in the acquisition of a multi-complex cancer genotype, in children a unique genetic alteration such as the MLL fusion alone may be sufficient to trigger an aggressive leukemia in infants [23]. This is a new area of research to explore in T-ALL, which will be developed during the coming years with the associated therapeutic consequences.

Clonal heterogeneity and relapse
Genome-wide approaches have also helped to define the clonal heterogeneity of T-ALL at diagnosis. The importance of this subclonal heterogeneity was recently revealed through the analysis of CNAs in paired diagnosis-relapse pediatric T-ALL samples [24]. Clappier et al. noted that while leukemic cells at diagnosis and at relapse share a large number of genetic lesions, additional genetic alterations that are underrepresented in the initial diagnosis sample are also detected at relapse but in a higher frequency. Thus, at the time of diagnosis leukemia could be composed of several subclones that share the same common pre leukemic ancestor, but only one or a few of them would be able to acquire a selective advantage in a particular microenvironment (niche) and proliferate to develop leukemia. Relapse would be mainly composed of clones underrepresented at diagnosis and selected during treatment. The mutational screening of paired samples (diagnosis-relapse) of pediatric ALL by NGS identified the NT5C2 gene to be specifically mutated at relapse [25]. Sequencing of the gene in a large number of patients confirmed that mutations in NT5C2 were found in 19% of relapsed T-ALL patients. Mutations in this gene create resistance to 6-mercaptopurine and 6-thioguanine, two nucleoside analogues used in T-ALL treatment. That was the first demonstration that leukemia developed at relapse arises from a clone refractory/resistant to the treatment [25]. A similar study published by Meyer et al., showed mutations in NT5C2 in 2/7 of the samples tested at diagnosis, at a frequency of mutated alleles of 0.01% and 0.02% respectively, indicating that, in some cases, resistant mutations already exist before treatment at a very low frequency [26]. Collectively, these data clearly suggest that treatment at diagnosis and at relapse should be different, as is currently being carried out in clinics, but it is not informative regarding how to predict or anticipate the existence of a resistant clones before starting treatment. In addition, it is difficult to explain the origin of clone resistance in cases where no genetic alterations are detected at diagnosis. Probably the generation of resistant subclones is much more dynamic and clearly influenced by the selective pressures of the microenvironment, such as the treatment process itself. Therefore new approaches to better identify and define the resistant clones should be developed.

Clinical implication: prognostic value of biomarkers
To assess whether the information obtained in basic research is relevant to the clinics, basic data need to be correlated with clinical information of patients included in treatment protocols, in order to establish the prognostic value of the different molecular markers.

Minimal residual disease (MRD)
Prognostic factor independent of treatment response (35), (91) Prognostic factor independent of treatment response (36), (92) ETP : Early T-cell Precursor; TCR: T-cell cellular receptor; NA: not analyzed these markers increases, the number of studies with different results, and sometimes with discordant results, also increases. The clearest example is the predictive value of mutations in NOTCH1/FBWX7 genes. In children, two different groups with two different treatment protocols (the English MRC UKALL 2003 [27], and the German ALL-BFM 2000 [28]) confirmed NOTCH1 mutations as markers of favorable prognosis in this subtype of ALL. However, the American Pediatric Oncology Group (POG) found no predictive value in a study on 47 patients analyzing 5 different treatment protocols [29]. Likewise, in the multicenter French study, which included children treated with two different protocols (EORTC 5881/58951), NOTCH1 mutations only had favorable prognostic value on early response to treatment [30]. In the case of adults, the prognosis of NOTCH1/FBWX7 value was initially even more difficult to analyze because there were fewer and discordant results, probably, due to the use of different treatment protocols (GRAALL03 LALA-94/05, by the French group [31] and GMALL 05/93 and 06/99 24 by the German group [32] (Table 4)).
Recently, some light has been shed on the role of NOTCH1/FBWX7 and its prognostic value. A study of the Eastern Cooperative Oncology Group (E2993ECOG) has confirmed the favorable prognostic value of mutations in NOTCH1/FBWX7 in a series of 53 adult patients [33]; and last year the GRAALL group identified a subgroup within the high risk treatment protocol with a clear favorable outcome (5-year OS approximately of 90% ) defined by negative MRD (<10 -4 ), the presence of NOTCH1/FBWX7 mutations and the absence of N/K RAS and PTEN mutations [34]. In summary, the different impact depending on the use of different treatment protocols by the cooperative ALL groups makes necessary to perform additional studies with large series of patients, with detailed biological and clinical data. This is only possible if cooperative international initiatives are developed.

Minimal Residual Diseases (MRD)
Analysis of minimal residual disease (MRD) by flow cytometry (for aberrant immunophenotypes), or by detecting fusion genes by qPCR (for TCR translocated leukemias or others), or TCR rearrangements, is currently the most powerful prognostic factor for treatment response and survival in T-ALL, in both children and adults [35]. MRD allows us to track the response to treatment in each patient and in each phase of treatment, and the early identification of those patients with a high risk of relapse. This allows us to tailor treatment in these patients and even anticipate the rescue therapy if subclinical relapse appears. Consequently, all therapeutic ALL protocols currently incorporate the analysis of MRD to monitor patients at key moments to decide on their treatment.
Current techniques of massive high resolution genomic analysis are being also applied to monitor MRD in ALL. Three different research groups have shown that it is possible to apply NGS in ALL for MRD detection [36]. With the use of NGS, MRD moves from tracking a single clone to clone screening (including the annotation of normal lymphocyte rearrangements). This us to increase the sensitivity of the test and detect minor clones that are important for later relapse, avoiding false negative results.

Treatment
Most of the drugs used in the treatment of ALL were developed before 1970. Since then, the treatment of ALL has continued to improve and reached a survival rate in pediatric ALL of about 90% and between 30-40% in adults. This is thanks to the development of numerous treatment protocols by different cooperative ALL working groups, who optimized them by testing different combinations of drugs, dosage and administration sequences. An important improvement has been the inclusion of MRD detection, measured in centralized specialized platforms to give robustness to the MRD data. Along with this, the identification of specific risk groups according to the molecular characteristics of the diseases, such as the Philadelphia chromosome subgroup (Ph + ) has allowed the administration of specific and effective treatment with ABL tyrosine kinase inhibitors. However, using similar treatment strategies in children and adults, the cure rate in adults continues to be very low. This could be explained by a higher presence of genetic variants of poor prognosis in adults and, in many cases, the adult does not tolerate the intensity and density of the treatment that is administered in a child.

Standard treatment
In Spain, patients diagnosed with an ALL follow treatment protocols from the PETHEMA (Programa Español de Tratamientos en Hematología) cooperative group. T-ALL patients are included in high-risk protocols except those with cortical subtype and less than 100x10 9 leucocytes/L at the time of diagnosis. The key elements of these protocols are the use of high-dose methotrexate and asparaginase and the assessment of MRD to take the decision of post-consolidation therapy (delayed intensification and maintenance or allogeneic stem cell transplantation). It is striking that they are the same as the ones used for B-precursor high risk Ph-negative ALL. There are, however, differences in treatment protocols between adults and children. Pediatric treatment protocols use more intensive regimens of essential cytotoxic drugs. In young adults physicians are using treatment protocols inspired by pediatric treatment regimens, which has led to an improvement in overall survival [37]. However, the application of these regimens is only feasible in patients up to 45-55 years as over that age the increase in mortality counterbalances the positive effect of the treatment [37]. In general, adult patients with T-ALL and high-risk criteria are treated with chemotherapy followed by allogeneic stem cell transplantation.

New treatments
Targeting the BCR-ABL1rearrangement with ABL1 tyrosine kinas inhibitors (imatinib, dasatinib, nilotinib, and ponatinib, has represented a major advance in the treatment of children and adults with Ph + ALL [38]. A minority of patients with T-ALL show genetic alterations involving the ABL1 gene, as NUP214-ABL1, EML1-ABL1, ETV6-ABL1, for which the efficacy of these inhibitors is under investigation [39]. One of the drugs approved for the treatment of patients with T-ALL who are resistant or in second or subsequent relapse is nelarabine. This deoxyguanosine analog, which accumulates specifically in T cells, has shown promising results in Phase I and II, in either adults or children with T-ALL and T-lymphoblastic lymphoma at relapse [40]. The most important toxicity is neurologic [40]. The Children's Oncology Group conducted a pilot trial with nelarabine combined with intensive chemotherapy treatment in newly diagnosed children with high-risk T-ALL, with encouraging results [41], and has included nelarabine in the current treatment protocol (COGAALL0434) for children with T-ALL.

Immunotherapy
A breakthrough in the treatment of ALL has been the incorporation of specific therapies using monoclonal antibodies, such as anti-CD20 (rituximab), anti-CD19 (blinatumumab, which is also anti-CD3), anti-CD52 (alemtuzumab), anti-CD22 (epratuzumab andinotuzomab ozogamycin) and anti-CD33 (gemtuzumab ozogamycin), among others [42]. Recently, genetically modified T-lymphocytes (CAR T-cells) designed to recognize surface antigens of B-cells such as CD19 and effectively kill leukemic cells have been included in the therapeutic armamentarium [43]. Most of the monoclonal antibodies and all the CAR T-cells are directed against malignant B-cells, although normal B-cell lymphocytes are also target. All of the immunotherapeutic approaches mentioned before are directed to B-cell precursors ALL. For T-ALL there only are few studies with alemtuzumab. This monoclonal antibody, that recognizes both T-and B-cells, has been investigated in small trials, but its development has been slow because of its modest activity and significant side effects [44].

Future Therapies
New therapies on T-ALL should be based on the improving understanding of the basic molecular data and the specific subtype to be applied. In other words, they must be as personalized as possible. In this regard, the molecular characterization of the poor prognosis ETP-ALL subtype opens the possibility of using drugs tested for the AML treatment such as FLT3 [45] and the JAK/STAT inhibitors [46]. The overwhelming evidence for an inappropriate regulation of epigenetic factors during malignant transformation has created an enormous interest in pharmacological intervention of epigenetic processes. In this sense, new targeted therapies for T-ALL could be based on the use of methyltransferase inhibitors or histone deacetylase inhibitors, such as 5-azacytidine and vorinostat, respectively [47]. Another new possible targeted therapy strategy could be the use of cyclin-CDK complex: D3 CDK4/6inhibitors, which have been shown to be effective in inhibiting proliferation of blasts in samples from patients at diagnosis and at relapse, in vivo [48].
One form of targeted treatment is to use a downstream signaling molecule that has regulatory functions in the signaling pathway affected in the leukemic. This could be the case for protein phosphatases. Inhibition of phosphatase activity of calcineurin by using FK506 and cyclosporine drugs blocks proliferation and increases apoptosis in a mouse model of T-ALL [49]. In addition, blasts that do not express calcineurin are not capable of developing leukemia when these cells are re-injected in secondary recipient mice [50]. Another good target could be the PI3K cellular intermediate, since there is a lot of evidence that the PI3K/Akt/mTOR signal pathway contributes to T-ALL development [51].

Conclusions
The improvement in survival and the progress towards healing patients with T-ALL will ultimately come from the relevant molecular information that will allow us to define the risk and the treatment to apply to each patient and increase the therapeutic armamentarium with new alternative therapies. Progress in the molecular knowledge of T-ALL is increasing quickly, since the technological support in which this knowledge is based is continually evolving. The sequencing platforms are continuously improving to detect very rare alterations and to process a big number of samples at lower cost. Also, new techniques such as single cell sequencing will help us to understand and identify the resistant clones. However, the biggest challenge lies in being able to transfer this information to clinical practice. This requires studies with human samples and the development of humanized animal models for T-ALL, such as xenotransplantation. This is not possible without preserving good quality of samples from patients within collections or biobanks, in which samples are collected and stored according to a strict protocol for handling and covering the legal aspects of the collection. In addition, the evaluation of the prognostic significance of new putative biomarkers in patients treated homogeneously should continue, and the establishment of partnerships with other national and international groups to increase the series and compare the data is mandatory. The identification of resistant clones and the development of specific therapies directed against these resistant cells localized in specific niches is highly desirable. These therapies should be able to remove the resistant clone before the patients experience overt relapse (preventive therapy).