Mammalian target of rapamycin inhibitors and their potential role in therapy in leukaemia and other haematological malignancies


David T. Teachey, MD, Divisions of Hematology and Oncology, Children’s Hospital of Philadelphia, ARC 910 3615 Civic Centre Boulevard, Philadelphia, PA 19104, USA.


The mammalian target of rapamycin (mTOR) is a serine/threonine kinase that functions as a key regulator of cell growth, protein synthesis, and cell-cycle progression through interactions with a number of signalling pathways, including PI3K/AKT, ras, TCL1, and BCR/ABL. Many haematological malignancies have aberrant activation of the mTOR and related signalling pathways. Accordingly, mTOR inhibitors, a class of signal transduction inhibitors that were originally developed as immunosuppressive agents, are being investigated in preclinical models and clinical trials for a number of haematological malignancies. Sirolimus and second-generation mTOR inhibitors, such as temsirolimus and everolimus, are safe and relatively well-tolerated, making them potentially attractive as single agents or in combination with conventional cytotoxics and other targeted therapies. Promising early clinical data suggests activity of mTOR inhibitors in a number of haematological diseases, including acute lymphoblastic leukaemia, chronic myeloid leukaemia, mantle cell lymphoma, anaplastic large cell lymphoma, and lymphoproliferative disorders. This review describes the rationale for using mTOR inhibitors in a variety of haematological diseases with a focus on their use in leukaemia.

Targeting mTOR signalling

Mammalian target of rapamycin (mTOR) inhibitors (MTIs) are a class of signal transduction inhibitors developed as immunosuppressive agents. Because mTOR signalling is aberrantly activated in a number of malignancies, MTIs are being investigated in a number of tumour types in both pre-clinical models and clinical trials. Sirolimus (rapamycin), a macrocyclic lactone produced by Streptomyces hydroscopicus, was the first MTI to be used in a clinical setting (Schmelzle & Hall, 2000). Sirolimus is US Fedral Drug Administration (FDA)-approved as an immunosuppressive agent in solid organ transplantation, but the drug has clear anti-neoplastic activity and is in phase II–III trials against a variety of cancers (Baldo et al, 2008). Sirolimus has poor aqueous solubility and variable bioavailability, requiring therapeutic drug monitoring. A number of second-generation MTIs, including temsirolimus (CCI-779), everolimus (RAD001), and deferolimus (AP23573) have been developed to circumvent those problems. These agents are also being investigated in a number of malignancies. Temsirolimus was the first MTI to gain FDA approval for any malignancy, having been approved for the treatment of advanced renal cell carcinoma (Baldo et al, 2008).

When used as monotherapy, MTIs are relatively well-tolerated. Unlike the commonly used immunophillins, ciclosporin and tacrolimus, MTIs cause little nephrotoxicity and neurotoxicity. MTIs may cause hyperlipidemia, mild myelosuppression, hypertension, and mucositis (Baldo et al, 2008). The toxicities of combining MTIs with conventional cytotoxic agents have not been fully explored in both preclinical and clinical studies.

Because mTOR signalling has been demonstrated to be important in cell growth and survival in a number of haematological malignancies, MTIs are being investigated in a myriad of diseases. Preclinical studies have demonstrated MTIs have activity either when used as single agents and/or when used in combination with cytotoxic chemotherapeutics and other targeted agents in acute lymphoblastic leukaemia, acute myeloid leukaemia, chronic lymphocytic leukaemia, chronic myeloid leukaemia, multiple myeloma, non-Hodgkin lymphoma, myelodysplastic syndrome, and non-malignant haematological disorders, including lymphoproliferative disorders. Numerous clinical trials are underway for these diseases and early clinical data has shown potential activity in a number of these conditions. This review will focus on the use of MTIs in leukaemia, but will also briefly summarize on-going work in other malignant and non-malignant haematological disorders.

The mTOR signalling network

Mammalian target of rapamycins is a 210 kDa protein that has C-terminal homology to phosphatidylinositol-3 kinase (PI3K) and is therefore a member of the PI3K-related kinase family (Wullschleger et al, 2006). mTOR is a serine/threonine kinase that acts as a central regulator of cell growth, survival, metabolism, and proliferation and functions as a sensor to ensure that the cell is in an appropriate nutritional and bioenergenic state to support these processes prior to committing to growth and division (Schmelzle & Hall, 2000). When mTOR is activated, a number of cellular processes occur, including an increase in ribosomal biogenesis, cap-dependent translation (initiation of translation involving 5′-end of mRNA), TOP-protein translation (translation of specific class of mRNAs containing oligopyrimine tracts in 5′untranslated region), expression of metabolism-related genes, cell growth, nutrient and amino acid uptake, and an increase in cell cycle transit time (Wullschleger et al, 2006). Conversely, the activation of mTOR leads to an inhibition of apoptosis and autophagy (Asnaghi et al, 2004; Zeng & Kinsella, 2008). The import of nutrients and amino acids is critical for the generation of ATP and cell metabolism. mTOR regulates these processes in part by up-regulating the protein translation machinery, which results in the synthesis of nutrient and amino acid transporters (e.g. Glut 1), as well as key molecules that promote cell growth and survival, such as Hif-1a, Cyclin D1, and myc (Gera et al, 2004; Majumder et al, 2004).

Mammalian target of rapamycin can form two distinct complexes, mTORC1 and mTORC2 (Fig 1) (Bhaskar & Hay, 2007). mTORC1 is sensitive to MTIs, including sirolimus, and is thought to regulate cell growth, proliferation, autophagy and translation in response to nutrients and energy availability. Data suggest that mTORC2 is insensitive to MTIs in some cell types whereas mTORC2 remains sensitive to MTIs in other cancer cell types (Rosner & Hengstschlager, 2008). mTORC2’s function, regulation and response to MTIs remains unclear, and seems to vary by cell type. mTOR is activated by a number of upstream signalling pathways, including PI3K/AKT, RAS/MAPK/RSK, cytokine signalling (IKK), TCL1, BCR-ABL, and nutrient (amino acid) sensing via the Rag GTP-binding proteins (Shaw & Cantley, 2006; Kharas et al, 2008). The main downstream targets of activated mTORC1 are S6K1 and the inhibitor of cap-dependent translation, 4E-BP-1 (Wullschleger et al, 2006). Figure 1 summarizes the mTOR signalling pathway.

Figure 1.

 mTOR signalling cascade. mTOR regulates a number of key cellular processes in mammalian cells, including protein translation. mTOR can bind to GβL, Mlst8, PRAS40, and RAPTOR, forming the MTI sensitive complex, mTORC1 (Bhaskar & Hay, 2007). In comparison, the components of mTORC2 include mTOR, GβL, mSIN1, RICTOR, and PROTOR/PRR5 (Bhaskar & Hay, 2007). mTORC2, in concert with PDK1, activates AKT by phosphorylation. Activation of growth factor receptors, including IL-7R, IGF-1R, c-kit, and flt-3, via insulin, hormones, and growth factors, leads to activation of IRS-1. The activation of IRS-1 in turn leads to PI3K upregulation. PI3K can also be activated by directly associating with the growth factor receptor at the cell membrane. Activated PI3K generates PIP3, which can recruit AKT to the cell membrane so that PDK1 and TORC2 can activate it (Wullschleger et al, 2006). The tumour suppressor PTEN negatively regulates PI3K by dephosphorylating its second messengers, i.e. PIP3 (Mills et al, 2001). Inactivating mutations of PTEN, which are found in many tumour types, leads to excess activation of AKT, mTOR, p70S6 kinase 1 (S6K1) and can increase sensitivity to mTOR inhibition (Neshat et al, 2001). Activated AKT then can phosphorylate TSC2, resulting in the inactivation of the TSC1:TSC2 complex, allowing for RHEB to activate mTORC1 (Wullschleger et al, 2006). The main downstream targets of activated mTORC1 are S6K1 and the inhibitor of cap-dependent translation, 4E-BP-1 (Wullschleger et al, 2006). mTORC1 kinase phosphorylates S6K1. Phosphorylated-S6K1 induces TOP-translation and ribosomal biosynthesis and blocks apoptosis by phosphorylating the pro-apoptotic molecule BAD. In addition, P-S6K1 acts as a negative feedback mechanism for the mTOR pathway by down-regulating IRS-1 (Harrington et al, 2005). mTORC1 regulates cap-dependent protein translation via phosphorylation of 4E-BP-1 (Huang et al, 2003). When hypophosphorylated, 4E-BP1 binds tightly to eIF-4E, blocking the association of eIF-4E with eIF-4G. This blocks the formation of the eIF-4F translation initiation complex, which is necessary for cap-dependent translation. When 4E-BP1 is phosphorylated by mTORC1, it is released from eIF4E, thereby facilitating translational initiation of mRNAs for a number of key intracellular proteins, including c-MYC, cyclin D1, and ornithine decarboxylase (Faivre et al, 2006). Cyclin D1 forms a complex with CDK4 (cyclin dependent kinase 4) which is required for activation via phosphorylation of Rb (retinoblastoma protein) (Ewen et al, 1993). mTOR also facilitates the elimination of the cyclin dependent kinase inhibitor p27kip1 through interactions with p34cdc2, allowing cell cycle progression under the regulation of cyclin-dependent kinases, including cyclin-A (Faivre et al, 2006). Arrows represent activation; Lines with circles represent inhibition. mTOR, mammalian target of rapamycin; PI3K, phosphoinositide 3-kinase; IRS, insulin receptor substrates; PTEN, phosphatase and tensin homologue deleted on chromosome ten; TSC, tuberous sclerosis; Rheb, ras homologue enriched in brain; p34cdc2, cyclin-dependent controlling kinase p34; p27kip1, cyclin-dependent kinase inhibitor kip1; cdks, cyclin-dependent kinases; pRb, retinoblastoma protein; S6, ribosomal protein S6; 4E-BP1, eIF-4E binding protein; eIF, eukaryotic initiation factors; GβL, G protein beta subunit-like; mTORC, mTOR complex; PIP2, phosphatidylinositol bisphosphate; PIP3, phosphatidylinositol triphosphate; PDK1, pyruvate dehydrogenase kinase, isozyme 1; Mlst8, mammalian lethal with sec-13; PRAS40, proline-Rich Akt substrate of 40kDa; mSIN1, mammalian stress-activated protein kinase-interaction protein 1; PROTOR/PRR5, Protein observed with Rictor-1/Proline-rich protein 5; GF, growth factor. Colour schematic: Yellow, mTOR; grey, Other proteins that bind to mTOR to form mTORC1/2; orange, targets phosphorylated by mTOR; pink, other down-stream-effectors; green, up-stream signalling molecules; purple, growth factor and receptor; red, drug.

A number of mechanisms that lead to mTOR deregulation have been identified. These include overexpression of growth factors (such as insulin-like growth factor, IGF), overexpression or mutations of growth factor receptors [e.g. IGFR, human epidermal growth factor EGF receptor (HER/EGFR)], point mutations in the PIK3CA (p110alpha PI3K) gene, loss of tumour suppressor genes (e.g. PTEN or TSC1:TSC2 complex), and gain-of-function mutations in mTOR or mTOR-linked pathways (e.g. formation of the aberrant protein BCR-ABL in Ph + leukaemia cells or stimulation of PI3K by aberrant ras/raf/MAPK pathway intermediates) (Inoki et al, 2005). Of these mechanisms, the most common ones are related to the overexpression or constitutive activation of PI3K or AKT and/or the loss of PTEN (Inoki et al, 2005). Through mTOR-mediated deregulated signalling or pathway cross-talk, increased mTOR activity supports cancer cells by stimulating the synthesis of proteins necessary for cell growth, proliferation, survival, angiogenesis, nutrient uptake and metabolism.


Acute lymphoblastic leukaemia (ALL)

Acute lymphoblastic leukaemia is a malignancy of lymphoid origin, arising from transforming events that occur in early B cell progenitors. It is the most common cancer in children, accounting for 35% of new paediatric cancer diagnoses (Plasschaert et al, 2004). Unfortunately, 20–25% of children and 80% of adults with ALL relapse and the majority of these patients succumb to their cancer despite aggressive therapy (Plasschaert et al, 2004). Current ALL treatment protocols use combinations of multiple cytotoxic chemotherapeutics with overlapping toxicity and the potential for long term sequelae, especially in the most intensively treated patients. New agents with activity against ALL are needed, and targeted biological agents have the potential to add efficacy without additional toxicity in patients with refractory ALL.

As MTIs have activity against lymphocytes and abnormal signalling can lead to neoplastic transformation, our group hypothesized that ALL cells may be dependent on mTOR signalling and studied the effects of MTIs on ALL blasts (Brown et al, 2003). We demonstrated that sirolimus inhibited proliferation and induced apoptosis in ALL cell lines and improved survival in a Eu-RET transgenic mouse model of leukaemia/lymphoma (Brown et al, 2003). Since that initial report, the mTOR signalling pathway has been extensively studied by our group and others in preclinical models of ALL (Brown et al, 2003, 2007; Avellino et al, 2005; Teachey et al, 2006a, 2008; Hirase et al, 2008; Houghton et al, 2008). MTIs (sirolimus, temsirolimus, and everolimus) have been shown to be effective not only against cell lines and transgenic mouse models, but also against primary human ALL cells using in vitro (bone marrow stromal cell culture) and in vivo [non-obese diabetic severe comined immunodeficient (NOD/SCID) xenografts] models (Avellino et al, 2005; Teachey et al, 2006a; Crazzolara et al, 2007).

The use of primary human ALL cells xenografted into immunodeficient mouse strains, such as NOD/SCID mice, is a powerful tool to study ALL biology and response to therapy. ALL cell lines can be useful tools, especially for signal transduction experiments, but they are difficult to establish from primary blasts, do not represent the heterogeneity of primary disease, and are thus suboptimal models for many preclinical studies. NOD/SCID xenografted ALL maintains its phenotypic characteristics even after serial passage and response of leukaemic blasts to chemotherapeutic agents in the NOD/SCID xenograft model has been shown to correlate directly with human response to therapy (Liem et al, 2004). Since treatment response in the mice correlates with human disease, these models can be used to compare chemotherapeutic responders to non-responders to delineate mechanisms of resistance. One problem with laboratory investigation of primary leukaemia cells has been the limited quantity of blasts for analysis. The NOD/SCID xenograft model allows for significant expansion of ALL in the mouse in order to generate sufficient quantities of cells for study.

Our group has studied the activity of MTIs using xenografts generated from 13 different ALL patients and found that MTIs were effective in 62% of samples (Brown et al, 2008). MTIs appear to be active against both pre-B and pre-T ALL, however, they may be most active in pre-T cell disease (Houghton et al, 2008). While MTIs have been clearly shown to kill ALL cells, debate exists in the literature as to whether it is through apoptosis or autophagy (Avellino et al, 2005; Crazzolara et al, 2007). As more data is accumulating that suggests ALL cells are dependent on the PI3K/AKT/mTOR signalling pathway, the activity of PI3K inhibitors, AKT inhibitors, and multi-kinase inhibitors (mTOR plus PI3K) is being investigated in ALL (Brown et al, 2008; Kharas et al, 2008; Levy et al, 2008). Also, as cancer cells may become resistant to MTIs through upregulation of other intermediaries in the PI3K/AKT/mTOR signalling pathway, combinations of MTIs with PI3K inhibitors and AKT inhibitors are being actively explored to overcome mTOR resistance with promising results (Breslin et al, 2005; Brown et al, 2008; Kharas et al, 2008).

Because MTIs are less likely to be effective in a clinical setting when used as single agents against leukaemia, combination treatment is the next logical step in the therapeutic development of MTIs in ALL. It is important to choose rationally-designed combinations, building on an understanding of the mechanism(s) of action of MTIs in ALL blasts. MTIs have been shown to be effective and potentially synergistic in combination with a number of chemotherapeutics in vitro, including methotrexate, dexamethasone, etoposide, asparaginase, and doxorubicin (Saydam et al, 2005; Teachey et al, 2008). MTIs have also been studied in combination with methotrexate and vincristine in vivo using NOD/SCID models with a marked response to both combination regimens (Crazzolara et al, 2007; Teachey et al, 2008). The combination of temsirolimus and methotrexate resulted in cure in some xenografted animals. ALL cells treated with temsirolimus had marked reduction of cyclin D1 and dihydrofolate reductase, potentially increasing the sensitivity of ALL cells to methotrexate and explaining the combined effect (Teachey et al, 2008). While this combination appears extremely promising in preclinical studies, both methotrexate and MTIs can cause mucositis and clinical trials are needed to determine tolerability. Nevertheless, everolimus and methotrexate have been successfully used in combination in patients with rheumatoid arthritis without significant toxicity, including mucositis (Bruyn et al, 2008).

Another promising combination studied by our group and others in ALL is corticosteroids with MTIs (Wei et al, 2006; Brown et al, 2008; Teachey et al, 2008). Wei et al (2006) screened a database of drug-associated gene expression profiles in ALL cells to evaluate gene expression signatures of glucocorticoid sensitivity as compared to resistance. They found the profile generated by sirolimus matched the signature of glucocorticoid sensitivity and demonstrated that sirolimus could restore steroid sensitivity to steroid-resistant ALL. Similar work performed by other groups suggests that MTIs may reverse glucocorticoid resistance in ALL cells, an important finding especially since ALL patients frequently develop steroid resistance at relapse (Gu et al, 2008; Haarman et al, 2008).

Patients with ALL that express the Philadelphia chromosome (Ph + ALL) have a particularly poor prognosis; however, the development of tyrosine kinase inhibitors against Bcr-Abl appears promising and will hopefully improve outcome. Resistance to these tyrosine kinase inhibitors is a real clinical concern. Because BCR-ABL is upstream of the PI3K/AKT/mTOR signalling pathway, MTIs maybe effective in Ph + ALL, including Bcr-Abl TKI resistant disease (Kharas et al, 2004). Promising data suggests that Ph + ALL cells may be especially sensitive to mTOR inhibition (see section on chronic myeloid leukaemia below) (Hirase et al, 2008; Kharas et al, 2008).

Based on the preclinical work that investigated MTIs in ALL, there are a number of on-going clinical trials evaluating MTIs in patients with ALL as single agents and in combination with other agents (Table I). Two Phase I/II trials of MTIs in patients with relapsed or refractory malignancies, including patients with ALL, have been completed (Yee et al, 2006; Rizzieri et al, 2008). Both of these trials had one patient each with ALL and both patients tolerated therapy, but neither had an objective response. Rheingold et al (2007) recently reported interim results of an on-going phase 1 trial of sirolimus in children with relapsed/refractory ALL. All patients tolerated sirolimus and three of seven patients had stable disease.

Table I.   Ongoing clinical trials.
DiseasePhaseLocation(s)Clinical numberAdditional information
  1. *A number of clinical trials are on-going using sirolimus post-stem cell transplant as part of GVHD prophylaxis in patients with haematological maligancies. ASCT0431 is the only one that randomizes patients to sirolimus versus no sirolimus with the hypothesis that sirolimus will improve survival via a direct action of sirolimus on ALL blasts.

  2. **CML in late accelerated phase or blast crisis.

  3. AML, acute myeloid leukaemia; ALL, acute lymphoblastic leukaemia; CLL, chronic lymphocytic leukaemia; CML, chronic myeloid leukaemia; NHL, Non-Hodgkin lymphoma; HL, Hodgkin lymphoma; MM, multiple myeloma; GVHD, graft versus host disease; ppx, prophylaxis; ATG, antithymocyte globulin; MEC, mitoxantrone, cytarabine, etoposide.

ALLIICOG Transplant CentresNCT00795886Randomized trial comparing sirolimus plus standard GVHD ppx versus standard GVHD ppx alone after stem cell transplant for ALL*
ALL/NHLIPhiladelphia, PANCT00068302Sirolimus for relapsed/refractory ALL or NHL
ALL/NHL CML**I/IIPhiladelphia, PANCT00776373Sirolimus plus etoposide and cytarabine for relapsed/refractory lymphoid malignancies
AMLI/II Philadelphia, PANCT00780104Sirolimus plus MEC chemotherapy for high risk AML
AMLIMelbourne, AustraliaNCT00636922Everolimus plus cytarabine in elderly with AML
AMLIParis, FranceNCT00544999Everolimus plus cytarabine and daunorubicin in relapsed AML
AMLIIRome, ItalyNCT00775593Temsirolimus and clofarabine for relapsed or refractory AML
AMLI/IIBavaria, GermanyNCT00762632Everolimus plus nilotinib for c-kit+ CML
CLL/B-NHLIIHouston, TXNCT00290472Temsirolimus for relapsed/refractory CLL or B cell NHL
CML**IMultiple centres in U.S., China, and SingaporeNCT00101088Temsirolimus and imatinib for CML accelerated phase
NHLIOntario, CanadaNCT00659568Temsirolimus for advanced lymphoma
NHLICleveland, OHNCT00671112Everolimus plus bortezomib for relapsed refractory MCL and other NHL
NHLIIMultiple centres U.S.NCT00436618Everolimus for refractory or advanced NHL
NHLITokyo, JapanNCT00622258Everolimus for refractory or relapsed NHL
NHL/HLI/IIMultiple centres U.S.NCT00704054Deforolimus for relapsed/refractory NHL HD
MCLIIMunich, GermanyNCT00727207Everolimus for relapsed/refractory MCL
NHL/HL MMI/IIMultiple centres U.S.NCT00474929Everolimus and Sorafenib for relapsed or refractory NHL, HD, or MM
MMINew York, NYNCT00317798Sirolimus and ATG for relapsed MM
MMIMultiple centres U.S.NCT00729638Everolimus and lenalidomide for relapsed MM
Advanced malignanciesIHouston, TXNCT00678233Temsirolimus plus IMC-A12 (anti-IGF-1R ab) for locally advanced or metastatic malignancy, including haematological
Advanced malignanciesISan Antonio, TXNCT00060645Deforolimis for relapsed/refractory malignancies, including NHL, HD, and MM

Given the potential activity of MTIs against ALL, and considering that haematopoietic stem cell transplantation (HSCT) is used as a major salvage strategy for patients with relapsed or refractory ALL and sirolimus has been used as graft-versus-host disease (GVHD) prophylaxis in a number of transplant trials, the use of sirolimus in the the post-HSCT setting has been proposed. This was tested in a phase II study, with promising results (Pulsipher et al, 2008). As a result, the Children’s Oncology Group has initiated a nationwide phase III randomized trial, ASCT0431, evaluating the addition of sirolimus to GVHD prophylaxis during HSCT for relapsed ALL. The primary hypothesis of this trial is that the addition of sirolimus to GVHD prophylaxis will increase leukaemia-free survival compared to a regimen of standard agents, through the novel benefit of using a drug that has the potential to both control GVHD and directly suppress leukaemic blasts. A strategy such as this would be a major advance in antileukaemia therapy and transplantation.

Acute myeloid leukaemia (AML) and myelodysplastic syndromes (MDS)

Acute myeloid leukaemia is a clonal disorder of myeloid haematopoietic stem/progenitor cells. While the prognosis for patients with AML has been poor, in general the outcome has improved with aggressive chemotherapeutic regimens and HSCT, at the cost of increased toxicities and long-term sequelae. A number of newer and more targeted agents with promise are in use and under development, including gemtuzumab ozogamicin, FLT-3 inhibitors, and farnesyl transferase inhibitors. Nevertheless, the only targeted therapies that have made significant improvements to date in the prognosis of AML have been used in patients with acute promyelocytic leukaemia with the additions of arsenic and all-trans retinoic acid (Sanz et al, 2009).

Recent interest has focused on targeting the PI3K/AKT/mTOR pathway in AML, as a majority of patient’s blasts have constitutive activation of AKT with subsequent phosphorylation of down-stream targets of mTOR, including 4E-BP1 and S6K1 (Xu et al, 2003). As direct inhibitors of AKT and PI3K inhibitors remain in early development, the primary focus thus far has been to evaluate MTIs in AML. Promising results have been demonstrated using monotherapy with MTIs in preclinical models of AML; however, these have not translated into substantial clinical benefit in early phase trials (Recher et al, 2005; Yee et al, 2006; Perl & Carroll, 2007; Rizzieri et al, 2008). This discrepancy may be due to the fact that early phase clinical trials are performed in patients with more aggressive disease (relapsed and/or refractory disease), and/or that the majority of preclinical work testing MTIs in AML has occurred in vitro. Until very recently, there were no murine models of AML that would allow testing of agents in vivo after the development of measurable disease. Prior preclinical work involved exposing AML cells to drugs in vitro and then testing the ability of cells to engraft in xenografted animals. This is in contrast to xenograft models of other haematological diseases, including ALL, where agents can be tested in vivo after the development of measurable disease. Despite these findings, interest remains as targeting the mTOR pathway may enhance the cytotoxicity of existing chemotherapeutic agents and other targeted agents. Xu et al (2005a) demonstrated that sirolimus enhanced the sensitivity of AML blasts to etoposide in vitro and the combination could prevent engraftment of AML cells in NOD/SCID mice better than either single agent alone if cells were treated in vitro prior to injection. Other groups have demonstrated that blocking mTOR increases the sensitivity of AML cells to HDAC (histone deacetylase) inhibitors and inhibitors of glycolysis (Xu et al, 2005b; Nishioka et al, 2008). As other inhibitors of the PI3K/AKT/mTOR pathway are developed and tested in clinical trials, these agents may prove superior to MTIs, because targeting PI3K with LY294002, AKT with perifosine, and both mTOR and PI3K with the dual inhibitor PI103 have shown promise in preclinical studies (Xu et al, 2003; Papa et al, 2008; Park et al, 2008). Because combination therapy with MTIs and either cytotoxics or biologicals may be beneficial in patients with AML, clinical trials testing MTIs in AML are actively enrolling patients (Table I). At the University of Pennsylvania, a phase I/II trial of sirolimus plus etoposide, mitoxantrone, and cytarabine has compelling preliminary data and is being broadened into a multi-centre randomized trial (Luger et al, 2006). In addition, trials using MTIs in elderly patients with AML who cannot tolerate more aggressive cytotoxic therapy and in combination with cytotoxics or biological agents in relapsed or refractory AML are ongoing (Table I).

Myelodysplastic syndromes are a heterogenous group of disorders characterized by cytopenias caused by defects in haematopoietic stem cell differentiation which frequently transform into AML. Patients are classified by the International Prognostic Scoring System (IPSS) into low, intermediate (groups 1 and 2), and high prognostic risk groups based on patient characteristics, pathology, and tumour biology (Greenberg et al, 1997). As PI3K/AKT/mTOR signalling has been shown to be important in cellular proliferation and malignant transformation, Follo et al (2007) hypothesized that aberrant activation of these survival signals may lead to transformation of MDS into AML. This group found that mTOR and its downstream intermediates, S6K1 and 4E-BP1, were activated in high risk MDS patients (IPSS: intermediate risk group-2 and high risk) and were not activated in low risk patients (intermediate risk group-1 and low risk). They also found that sirolimus was effective in vitro against the CD34+ cells from high risk patients but not those from low risk or normal controls. MTIs were noted to be active in some patients with MDS, resulting in either stable disease or improvement in cytopenias in early phase clinic trials and additional trials are under development (Yee et al, 2006; Rizzieri et al, 2008).

Chronic lymphocytic leukaemia (CLL)

CLL is the most common form of leukaemia and arises from transforming events in CD5+ B cells. With aggressive chemotherapy, stem cell transplant, and novel therapeutics, including monoclonal antibodies, the prognosis for CLL has improved; however, it remains largely an incurable disease (Lin, 2008). Two cell populations are thought to exist in CLL: a very large non-proliferating population of peripheral blood B lymphocytes and a smaller pool of cycling malignant B cells found in pseudofollicles in lymph nodes, spleen, and bone marrow (Dighiero & Hamblin, 2008). The crux of targeted therapy is currently directed at the smaller cycling compartment.

As MTIs have been shown to inhibit proliferation of malignant and non-malignant B lymphocytes and because PI3K was found to be constitutively active in CLL cells from patients, it has been hypothesized that targeting the PI3K/AKT/mTOR signalling pathway may be effective in patients with CLL (Ringshausen et al, 2002, 2005). Preclinical studies have shown that MTIs do not induce apoptosis in CLL cells but can cause cell cycle arrest through targeting the expression of cyclins D3, E, A and survivin (Decker et al, 2003; Ringshausen et al, 2005). In addition, sirolimus was shown to improve survival in a CLL transgenic mouse generated by overexpressing TCL1A under the control of the μ immunoglobulin gene enhancer (Zanesi et al, 2006). These preclinical studies led to a phase I clinical trial of everolimus in patients with CLL (Decker et al, 2008). Three patients on this trial had stable disease and one patient had a partial response. Unfortunately, this trial was stopped after only seven patients were enrolled because four patients developed opportunistic infections. Patients on this trial did not receive infectious prophylaxis, and all had received aggressive and immunosuppressive regimens prior to everolimus. Interestingly, a high infectious risk has not been seen in other trials using everolimus even when used in combination with other immunosuppressives, including corticosteroids and ciclosporin (Dunn & Croom, 2006; Decker et al, 2008). Smith et al (2008) recently presented similar preliminary results from a phase II non-randomized single institution trial at M.D. Anderson, demonstrating stable disease and partial responses in a number of patients with CLL who were treated with temsirolimus. Infections were noted but not to a degree that required the trial to be stopped. Other clinical trials using MTIs in CLL are ongoing (Table I).

Chronic myeloid leukaemia (CML)

Chronic myeloid leukaemia is a myeloproliferative disorder characterized by malignant cells with a BCR-ABL1 (9;22) translocation. The BCR-ABL1 oncogene in CML encodes a 210 kDa oncoprotein, whereas in Ph + ALL the translocation produces a 190 kDa oncoprotein. Both fusion proteins have aberrant tyrosine kinase activity (Piccaluga et al, 2007). Prior to the development of tyrosine kinase inhibitors (TKIs) with activity against Bcr-Abl, CML was only curable with HSCT. Over the past few years, complete responses have been documented with a number of targeted agents, including imatinib, nilotinib, and dasatinib, and front-line use of a TKI is now the standard of care for the disease (Gora-Tybor & Robak, 2008). Imatinib was the first TKI to be used in CML; unfortunately, approximately 25% of patients will either have innate resistance or more commonly acquire resistance to imatinib, because of Bcr-Abl mutations (Gora-Tybor & Robak, 2008). The majority of these patients will respond to second line TKIs, but a subset have a particular mutation (T3151) that is not treatable with current Bcr-Abl targeting TKIs (Gora-Tybor & Robak, 2008). Three reasons to develop alternative agents for patients with CML are: (i) to treat those with resistant disease; (ii) to identify relatively non-toxic agents that target pathways down-stream of Bcr-Abl and could be used in combination with Bcr-Abl targeting TKIs front-line in certain high-risk patients to potentially prevent the development of resistant clones; (iii) to treat patients in accelerated phase or blast crisis as they have a particularly poor prognosis and are less likely to respond to Bcr-Abl targeting TKIs as monotherapy.

Multiple groups have demonstrated that mTOR-dependent pathways are activated in Bcr-Abl transformed cells both in CML and in Ph + ALL (Ly et al, 2003; Kim et al, 2005; Mayerhofer et al, 2005). Bcr-Abl has been shown to regulate translation of critical targets in CML, including S6 and 4EBP-1 via mTOR (Ly et al, 2003). In preclinical studies, sirolimus has been shown to not only be effective when used as a single agent against Bcr-Abl transformed cells, but also to be potentially synergistic when combined with imatinib (Mohi et al, 2004; Mayerhofer et al, 2005). Sirolimus has also been shown to be effective in vitro against resistant CML, including cells with T3151 mutations (Sillaber et al, 2008). Similar results have been described in Ph + ALL (Hirase et al, 2008; Kharas et al, 2008). Accordingly, clinical trials are underway using MTIs in patients with relapsed/refractory CML and Ph + ALL. Sillaber et al (2008) treated six patients with imatinib-resistant CML with sirolimus and two patients had a major response and two others had a minor response. Wetzler et al recently completed a phase I–II study of everolimus in combination with imatinib in patients with imatinib-resistant CML (NCT00093639). Results of this trial are not currently available.

Other haematological malignancies

Targeting mTOR signalling has also been investigated in Hodgkin (HL) and non-Hodgkin lymphoma (NHL), post-transplant lymphoproliferative disorder, and multiple myeloma. Arguably, the subtype of lymphoma that has been the most studied and has the most potential for clinical benefit from targeting mTOR is mantle cell lymphoma (MCL) (Younes, 2008). MCL is an extremely aggressive and incurable form of B cell lymphoma with a median overall survival of 3–5 years (Schmidt & Dreyling, 2008). MCL is characterized by a t(11;14) translocation juxtaposing cyclin D1 with the immunoglobulin heavy chain, resulting in increased production of cyclin D1 (Hartmann et al, 2008). Cyclin D1 is a down-stream target of mTOR and MTIs can inhibit the cap-dependent translation of this protein in many cell types (Hipp et al, 2005). Accordingly, Hipp et al (2005) hypothesized that targeting mTOR would be an effective treatment in MCL by down-regulating cyclin D1 expression. They found that sirolimus was effective against MCL cell lines in vitro; however, while other cyclins (D3, E, and A) were reduced, cyclin D1 expression did not change (Hipp et al, 2005).

Subsequent work by Rudelius et al (2006), demonstrated that AKT, mTOR, and a number of mTOR down-stream signalling intermediates are constitutively activated in MCL, giving an alternative explanation for the potential effectiveness of targeting AKT/PI3K/mTOR signalling; these results were confirmed by Peponi et al (2006). Other preclinical work has demonstrated that MTIs synergize in vitro with a number of agents used to treat MCL, including rituximab, vincristine, doxorubicin, and bortezomib (Haritunians et al, 2007). These findings led to a phase II trial of temsirolimus in patients with relapsed lymphoma. Results from this study are impressive, with a 41% overall response rate and a median time to progression of 6 months (Ansell et al, 2008). Based on these results, additional trials are underway (Table I).

AKT and the down-stream intermediates, 4E-BP1 and S6K1, were shown to be activated in HL cell lines; however, targeting either PI3K with LY294002 or mTOR with sirolimus only showed a modest effect in vitro. Nevertheless, a combination of doxorubicin and sirolimus was found to be synergistic and profoundly inhibited the same cell lines (Dutton et al, 2005). A more pronounced single agent effect was demonstrated in a NOD/SCID xenograft model of HL treated with everolimus (Jundt et al, 2005). mTOR has also been shown to be activated in follicular lymphoma through a Syk-dependent mechanism, and sirolimus has some activity in follicular lymphoma cell lines (Calastretti et al, 2001; Leseux et al, 2008). Similarly, mTOR and its intermediates have been shown to be activated in ALK-positive anaplastic large cell lymphoma and mTOR inhibition is effective in preclinical models of the disease (Jundt et al, 2005; Vega et al, 2006). Chumsri et al (2008) reported that treating a patient with refractory cutaneous anaplastic large cell lymphoma with sirolimus resulted in a durable complete response. Targeting mTOR has also been shown to be effective in preclinical models of diffuse large B cell lymphoma (Wanner et al, 2006). All of these studies have led to a number of phase I-III clinical trials of MTIs as single agents or in combination in aggressive and/or refractory lymphomas (Table I).

Post-transplant lymphoproliferative disorder (PTLD) is a rare but serious complication of transplantation (solid organ or HSCT), resulting from a defective cytotoxic T cell response to viral infection, primarily Epstein–Barr virus (EBV), in the setting of chronic immunosuppression (Lewin, 1997). The goals of treatment are to reduce and/or alter immunosuppression to allow partial T cell recovery and/or to target the EBV-infected B cells with rituximab or anti-viral agents (Lewin, 1997). Some patients with PTLD have very aggressive disease with transformation to lymphoma and need definitive chemotherapy. Recently, El-Salem et al (2007) demonstrated constitutive activation of mTOR signalling in patients with PTLD, regardless of the EBV genome expression status, and preclinical studies have demonstrated the efficacy of MTIs in PTLD (Majewski et al, 2003; El-Salem et al, 2007). Accordingly, a number of investigators have changed immunosuppression to MTIs in PTLD patients (sirolimus and everolimus) with improvement in PTLD and, in some cases, documented complete responses (Pascual, 2007; Boratynska & Smolska, 2008).

Multiple myeloma is a plasma cell malignancy characterized by monoclonal proliferation of B cells producing a single immunoglobulin and affecting bone marrow and osseous bone (Kyle & Rajkumar, 2008). Treatment has improved over the past decade with the introduction of thalidomide (and thalidomide analogues) and bortezomib; however, even with these agents, aggressive chemotherapy, and stem cell transplant, median overall survival is 2 years in patients with high-risk disease (Kyle & Rajkumar, 2008). The PI3K/AKT/mTOR pathway is altered frequently in multiple myeloma with constitutive activation of AKT and loss of PTEN function (Pene et al, 2002; Shi et al, 2002). AKT and PI3K inhibitors have profound effects against multiple myeloma in preclinical models; however, MTIs were found to have lesser effects, demonstrating efficacy in cells with loss of PTEN function but showing significantly less activity in cells with normal PTEN function (Pene et al, 2002; Shi et al, 2002; Hideshima et al, 2006). Temsirolimus demonstrated efficacy in a xenograft model of multiple myeloma. In other studies, sirolimus was found to sensitize multiple myeloma cells to dexamethasone, and MTIs were found to synergize with other targeted agents, including sunitinib, a HSP90 inhibitor, and lenalidomide (Frost et al, 2004; Raje et al, 2004; Francis et al, 2006; Ikezoe et al, 2006; Yan et al, 2006). Clinical trials in multiple myeloma are ongoing (Table I).

Non-malignant haematological disorders

In addition to the recent interest of MTIs in haematological malignancies, targeting mTOR signalling has been studied in non-malignant haematological disorders, particularly autoimmune disorders, for two reasons: (i) MTIs can cause apoptosis in abnormal lymphocytes, whereas many immunosuppressive agents only inhibit growth; and (ii) MTIs increase peripheral blood regulatory T cells (Tregs). Tregs are a T cell population that suppresses the immune system. Tregs are frequently decreased in autoimmune diseases and increasing Tregs may improve autoimmune disorders (Brusko et al, 2008). We have studied the activity of sirolimus in a rare paediatric haematological disorder, autoimmune lymphoproliferative syndrome (ALPS). ALPS is a disorder of disrupted lymphocyte homeostasis caused by defective fas-mediated apoptosis, leading to chronic lymphoproliferation, systemic autoimmunity, and a propensity to develop secondary cancers (Bleesing et al, 2000). The most common autoimmune manifestation is autoimmune cytopenias and many patients have severe disease. We hypothesized that targeting mTOR would be effective in ALPS through inducing apoptosis in the abnormal lymphocytes found in the disease and/or by increasing Tregs. We found marked improvement in lymphoproliferation and autoimmunity following the treatment of murine models of ALPS (Teachey et al, 2006b). Based on those results, we have treated six refractory ALPS patients with sirolimus and found profound improvement in all patients with complete responses in the majority of patients (Teachey et al, 2009). We have an open clinical trial for ALPS patients and plan to expand the trial to include patients with chronic severe and/or refractory autoimmune cytopenias either as an idiopathic condition (immune thrombocytopenic purpura, autoimmune haemolytic anaemia, autoimmune neutropenia, or Evans syndrome) or as a consequence of a syndrome (systemic lupus erythematosis, rheumatoid arthritis, or inflammatory bowel disease).

Sirolimus has also been used as a immunosuppressive salvage regimen in patients who develop transplant-associated microangiopathy or autoimmune cytopenias (Teachey D.T., Jubelirer T., Baluarte H.J., Wade A. and Manno C.S., in preparation; Yango et al, 2002). Finally, MTIs are under investigation in preclinical models and/or clinical trials in aplastic anaemia and for haematological manifestations of systemic lupus erythematosis (Tisdale et al, 2000; Ramos-Barron et al, 2007).


Interest in MTIs has grown considerably over the recent years as the important role of mTOR signal transduction in cell growth and proliferation has become better elucidated. As a class, MTIs are safe and well-tolerated. Because aberrant activation of the AKT/PI3K/mTOR signalling network is a common finding in many haematological and non-malignant diseases, MTIs have the potential to be efficacious in a variety of disorders. To date, MTIs have demonstrated substantial activity against abnormal haematopoietic cells of multiple lineages. Nevertheless, as MTIs are unlikely to be curative as single agents in many malignancies, elucidating the most effective way of combining MTIs with conventional cytotoxic agents and new targeted therapies is imperative in order to improve cure rates in these difficult-to-treat diseases.


Supported by a Larry and Helen Hoag Foundation Clinical Translational Research Career Development Award, ASCO Young Investigator and Career Development Awards, and the Leukaemia and Lymphoma Society (DTT); NIH 1 K08 CA104882-01A1, grant no. IRG-78-002-30 from the American Cancer Society, the Children’s Cancer Fund, the Florence R.C. Murray Program at the Children’s Hospital of Philadelphia and WW Smith Charitable Trust (VIB); and NIH CA102646, CA1116660, ACS RSG0507101, and the Weinberg Fund of the Children’s Hospital of Philadelphia (SAG).