High motility group overexpression accelerates T-cell leukemogenesis

Acute lymphoblastic leukemia (ALL) is the most common malignancy seen in childhood. Improvements in therapy for ALL have resulted in marked increases in overall survival. Specifically, cure rates of 80–90% for B-lineage acute leukemias have been achieved, in part due to a better understanding of the biological basis of the disease. Studies identifying key genetic alterations and linking them with clinical outcomes data have allowed patients with B-lineage ALL to be risk-stratified at the time treatment decisions are being made. Patients with high risk mutations can be given intensified therapy, and those with low risk genetic changes can receive less aggressive therapy, minimizing toxicity.

Unfortunately, the prognosis for T-cell ALL is poorer, with 5-year overall survival rates of 70% in the pediatric population and 10–40% in the adult population [1]. While many patients with T-ALL respond to initial therapy, a significant proportion ultimately experience treatment failure and subsequent relapse. In contrast to B-lineage ALL, no risk stratification currently exists for T-ALL, and data are just emerging regarding possible prognostic factors for this disease. A more thorough understanding of the cellular and molecular biology of T-ALL is required, with the goal of identifying novel therapeutic targets.

Insight into the molecular pathogenesis of T-ALL has blossomed over the past two decades. Many recurrent genetic alterations have been identified in T-ALL, and these can be grouped into several broad categories [2]. NOTCH mediated signaling is critical for early T-cell development, and constitutively activated NOTCH seems to play an important role in T-cell leukemogenesis. In fact, activating NOTCH1 mutations have been found in approximately 50% of human T-ALL samples tested [3]. Mutations in a variety of different tumor suppressor genes have been identified in 10–20% of T-ALL. These include the tumor suppressors WT1, RUNX1 and ETV6. Signal transduction proteins – including PTEN, NRAS, JAK1 and IL7R – have been found to be mutated in smaller percentages of T-ALL.

Defects in cell cycle transit have also been found in T-cell malignancies, and are thought to contribute significantly to T-ALL leukemogenesis. Importantly, deletion within the CDKN2A/2B tumor suppressor locus has been found in more than 70% of tested T-ALL samples [4]. The CDKN2A tumor suppressor locus includes both INK4a and ARF, which are generated via alternative reading frames. The most common deletion in T-ALL occurs within exon 2, which affects the transcripts of both INK4a and ARF [5]. Since these two proteins act in concert with p53 and Rb to regulate the cell cycle, the result of their inactivation is enhanced transit through the cell cycle and increased proliferation.

Finally, alterations in several genes involved in chromatin remodeling and DNA repair have been identified in many cases of T-ALL; an example is HMGA1, which is very frequently mutated in a variety of leukemias [6]. The high mobility group A (HMGA) family consists of four chromatin remodeling proteins that modulate gene expression through several different mechanisms. HMGA proteins may interact directly or indirectly with DNA and transcription factors, but ultimately modulate the interaction between the two through conformational changes in either chromatin or transcription factors themselves. Ultimately, HMGA proteins interfere with cell cycle regulation through a variety of mechanisms, including inhibition of Rb phosphorylation, binding and inhibition of p53, and increasing the activity of AP1. HMGA proteins, including HMGA1, have been shown to be up-regulated in a wide variety of malignancies including solid tumors such as pancreatic, colon and gastric carcinomas, breast cancer, lung cancer, and head and neck tumors [7]. HMGA1 has also been found to be up-regulated in several human leukemia lines.

In this issue of Leukemia and Lymphoma, Di Cello and colleagues [8] use Cdkn2a knockout and HMGA1a transgenic mouse models to explore a possible cooperative role between mutations in HMGA1 and the Cdkn2a tumor suppressor locus in T-cell leukemogenesis. The Resar group had previously demonstrated that HMGA1a transgenic mice, which overexpress HMGA1, develop aggressive T-ALL [5]. Here, Di Cello et al. report that complete loss of Cdkn2a accelerated leukemogenesis in HMGA1a transgenic mice. The HMGA1a transgenic/Cdkn2a null mice displayed marked splenomegaly and significantly decreased overall survival compared with HMGA1a transgenic/Cdkn2a wild-type (WT) mice, with a mean survival of 5 months in the HMGA1a transgenic/Cdkn2a null cohort versus 10.7 months in the HMGA1a transgenic/Cdkn2a WT cohort. The HMGA1a transgenic/Cdkn2a null leukemias were transplantable, with a reduced latency as might be expected. Importantly, the immunophenotype of these leukemias was consistent with a T-cell origin, with expression of Thy 1.2, CD3, CD8 and αβTCR.

Because their murine genetic studies indicated a cooperation between HMGA1a overexpression and Cdkn2a loss, the authors then explored the frequency of HMGA1 mutation in human T-ALL using the publicly available Oncomine database, and found that HMGA1 is one of the most frequently overexpressed genes in pediatric T-ALL compared with control bone marrow samples [8]; indeed, HMGA1 is among the top 10% of genes overexpressed by > 3-fold. Of note, analysis of levels of HMGA1 expression in two other T-ALL datasets (from the Children's Oncology Group and the Dutch Oncology Registry) showed higher HMGA1 gene expression, although statistically significant differences relative to controls were not observed. Nevertheless, these observations in patient samples, combined with the in vivo murine results, support a role for HMGA1 overexpression in T-ALL pathogenesis.

There are two significant implications of the results of Di Cello et al. First, the development of a mouse model that recapitulates salient features of T-ALL in the presence of these two common mutations will be a useful tool for future study. It will be important to further delineate the molecular biological underpinnings of HMGA1 mediated leukemogenesis, and potential cooperating lesions. The availability of this mouse model to other investigators will allow new insights into the pathogenesis of T-ALL, with the potential for testing drugs in a clinically applicable model. Second, Di Cello et al. show that two of the most frequent mutations in T-ALL cooperate in a mouse model to produce a very aggressive leukemia. It will be interesting to further investigate whether this holds true in human T-ALL leukemias, and whether these mutations could be possible prognostic markers for adult and/or pediatric T-ALL. Identification of true prognostic markers in T-ALL is desperately needed, and risk stratification of this disease has the potential to substantially improve outcomes. Moreover, the recognition that HMGA1 over- expression contributes to T-ALL should prompt the search for additional inhibitors of the HMGA1 protein. Despite toxicity issues that will need to be addressed, promising preliminary results have been seen with flavopiridol and FR900482 [9,10], agents that block HMGA1 function. This could represent a novel therapeutic avenue in T-ALL.

Potential conflict of interest:

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.