Abstract
This is an exciting time in the treatment of acute lymphoblastic leukemia (ALL) given the advances in the relapsed/refractory setting. The development of antibody treatments (including antibody drug conjugates with toxins) offers a different treatment approach compared with conventional chemotherapy regimens. Moreover, the use of bispecific T-cell-engager antibodies (BiTEs) such as blinatumomab harness the cytotoxic activity of T cells against CD19-positive lymphoblasts. Another strategy involves the use of chimeric antigen receptor (CAR) T cells. CAR T cells have demonstrated promising results in the relapsed/refractory setting. However, the use of BiTEs and CAR T cells is also associated with a distinct set of adverse reactions that must be taken into account by the treating physician. Apart from the above strategies, the use of other targeted therapies has attracted interest. Namely, the discovery of the Philadelphia (Ph)-like signature in children and young adults with ALL has led to the use of tyrosine kinase inhibitors (TKI) in these patients. The different drugs and strategies that are being tested in the relapsed/refractory ALL setting pose a unique challenge in identifying the optimum sequence of treatment and determining which approaches should be considered for frontline treatment.
Keywords
Background: treatment of acute lymphoblastic leukemia in adults
Acute lymphoblastic leukemia (ALL) in adults is predominantly a disease of the elderly with an incidence of 1.6 cases/100,000 people/year [Marks, 2015]. A number of factors such as elevated white blood cell count (WBC), hypodiploidy, presence of the Philadelphia (Ph) chromosome or the mixed lineage leukemia (MLL)/11q23 rearrangement, five or more chromosomal aberrations, prolonged time to remission, as well as persistence of minimal residual disease (MRD) are associated with worse outcome [Stock, 2010; Hochberg et al. 2013; Inaba et al. 2013]. The treatment of ALL in adults remains particularly challenging and there is an unmet need for novel therapeutic approaches. The focus of this review will be new therapies as well as promising preclinical developments in Philadelphia (Ph) chromosome-negative ALL with a focus on precursor-B (pre-B) ALL.
Upfront treatment regimens are typically prolonged and involve multidrug combinations. Several different protocols have shown efficacy [Litzow and Ferrando, 2015]. The treatment scheme is commonly divided into induction, followed by consolidation, and a protracted outpatient maintenance phase (usually 2–3 years). Pediatric-inspired regimens are more intense compared with traditional protocols for adults and have led to improvements in outcomes in adolescents and young adults (AYA) with ALL [Lukenbill and Advani, 2013]. However, application of this approach to older patients (⩾45 years of age) has been associated with increased morbidity/mortality in one series [Huguet et al. 2009]. Moreover, a particularly vulnerable population is elderly patients (typically defined as ⩾60 years of age) who may not be able to tolerate chemotherapy well (discussed in [Marks, 2015]). In a phase II study that enrolled patients above 40 years old, those in the 61–70 year old subgroup had a 79% complete response (CR) rate but 21% died during induction treatment [Daenen et al. 2012; Marks, 2015]. Another recent study revealed that 5 year overall survival has not improved in patients with ALL ⩾ 70 years (comparing 1992–2001 with 2002–2011 treatment periods) and the 5 year overall survival remained below 10% [Guru Murthy et al. 2015].
At the time of relapse, there is no standard treatment regimen and ultimately allogeneic hematopoietic stem cell transplant (AHSCT) is the only chance of cure. The prognosis for relapsed patients is dismal with a 5 year overall survival of less than 10% and is associated with high relapse rates [Fielding et al. 2007]. These outcomes may be explained by the fact that the lymphoblasts of relapsed/refractory cases have already manifested resistance to multidrug-intensive first-line chemotherapy.
The lymphoblast: a closer look
In patients presenting with acute leukemia, an important first step is the delineation of blast lineage. Morphologic, immunophenotypic and cytochemical assays are employed to distinguish ALL from acute myeloid leukemia (AML). In ALL blasts can be of B-cell lineage (B-ALL [~80%]) or T-cell lineage (T-ALL [~20%]) and are negative for myeloperoxidase or nonspecific esterase assays. Immunophenotypic assays are important in distinguishing ALL of B or T lineage (discussed in [McGregor et al. 2012; Chiaretti et al. 2014]). B-ALL is identified by the presence of B-lineage markers such as CD19, CD20, CD22 and CD79a. The immunophenotype of the neoplastic cells may vary; for example, CD20 is associated with more mature stages of B-lineage, while blasts of less mature stages are positive for markers such as CD10 or CD34. The hallmark of T-ALL is the detection of (cytoplasmic) CD3. Other T-cell lineage markers are CD1a, CD5, CD7 and CD2. In a sizeable number of ALL cases aberrant expression of myeloid markers may occur. Cases of acute leukemia of ambiguous origin represent a separate group and diagnosis may be challenging [Porwit and Bene, 2015].
Mature and pre-B lymphoblasts almost uniformly express CD19 while CD22 is expressed in the overwhelming majority of cases [Shah et al. 2015]. CD20 is detected in a subset of B- ALL (approximately 40%) and has been associated with a worse prognosis [Thomas et al. 2009]. Importantly, CD20 expression is upregulated during induction therapy, making it a particularly attractive therapeutic target [Dworzak et al. 2008].
Our understanding of lymphoblast biology and its mutational landscape has significantly expanded compared with a couple of decades ago [Lee et al. 2015b; Perez-Andreu et al. 2015; Roberts and Mullighan, 2015]. The information derived from deep sequencing and other molecular techniques may allow targeting a few specific pathways in the context of ALL. A distinct group of patients with pre-B ALL has recently attracted attention [Den Boer et al. 2009; Roberts et al. 2014a]. This group of patients, although not Ph chromosome-positive, have a very similar gene expression pattern and prognosis to those harboring the latter chromosomal aberration and are called ‘Ph-like’ [Den Boer et al. 2009]. Ph-like ALL is estimated to have lesions in genes implicated in B-cell development in the majority of cases [Den Boer et al. 2009]. An analysis of 1725 patients with pre-B ALL revealed that more than 1 in 4 young adults harbor Ph-like ALL while the prevalence is less in children [Roberts et al. 2014a]. The authors also reported that Ph-like ALL is associated with worse survival and serves as an independent adverse prognostic marker.
In the same study, Ph-like ALL could be further subdivided based on genetic lesions including CRLF2 rearrangements with and without JAK1/JAK2 mutations. The frequency of CRLF2-elevated expression was not uniform across age groups (2.5 times more frequent in adolescents compared with children). Other aberrations included fusions involving ABL1, ABL2 and PDGFRB or lesions (mutations, rearrangements) affecting critical pathways such as Ras and JAK-STAT [Roberts et al. 2014a]. The authors of the same study included preclinical data indicating that dasatinib or imatinib were effective in inhibiting the growth of pre-B cells with ABL1 or ABL2 fusions. Ruxolitinib, which is an approved JAK2 inhibitor for patients with primary myelofibrosis, [Mascarenhas and Hoffman, 2012] was effective in inhibiting the growth of leukemic cells with ATF7IP-JAK2 or IGH-EPOR aberrations [Roberts et al. 2014a]. Hence, a subgroup of Ph-like patients may be candidates for tyrosine kinase inhibitors (TKIs), discussed in more detail in the ‘Targeting Ph-like ALL’ subsection.
Moreover, a different study, presented in abstract form utilizing xenografts, revealed that STAT5 was activated in cases of ABL1 class kinase fusions [Roberts et al. 2014b]. Another study demonstrated aberrations affecting the JAK-STAT signaling pathway, while aberrations affecting the RAS gene were noted in a distinct group of patients (discussed in [Harrison, 2013]). A recent abstract analyzing a small number of patients indicated that the use of fluorescent in situ hybridization (FISH) for IGH-CRLF2 and mutational analysis of JAK2 were able to determine, with modest sensitivity but high specificity, Ph-like ALL [Herold et al. 2014]. This report is important, as advanced mutational screenings/sequencing may not be available in all facilities.
ALL and antibody therapies
Antibody therapies in the context of ALL are based on utilizing epitopes expressed on leukemic cells with a preference to epitopes that will interfere the least with normal hematopoiesis. Leukemic cells can be eliminated by multiple different mechanisms. These mechanisms include internalization of conjugated toxins, complement activation, direct cytotoxic effects and recruitment of patient's T cells to interact and eliminate leukemic cells (Figures 1 and 2).

Outline of the novel therapies targeting lymphoblasts. (A) Antibodies or antibody conjugates can recognize epitopes expressed on the surface of the lymphoblast. (B) Blinatumomab can act as a bridge to bring in close proximity CD3 T cells and the CD19 expressing lymphoblasts mediating the activation of the former and the destruction of the lymphoblast through a granzyme-mediated pathway. (C) Ex vivo-modified lymphocytes are a relatively new concept of targeted therapy and are effective at targeting CD19-expressing lymphoblasts.

Mechanisms of antibody-mediated cytotoxic effect against lymphoblasts. Antibodies can exert their action through a variety of mechanisms. (A) Naked antibodies can elicit recruitment of complement or cell-mediated cytotoxicity. (B) Cytotoxic moieties of antibody conjugates enter the cell and can affect the mitotic apparatus, prevent mRNA translation, or inflict DNA damage. The mechanism of action depends on the type of cytotoxic moiety.
Anti-CD20 antibodies
Rituximab
Rituximab is a monoclonal anti-CD20 antibody. In the context of ALL, it has been used mainly in combination with chemotherapy. The German Multicenter Study Group for Adult ALL (GMALL) explored the combination of rituximab and chemotherapy with encouraging results [Hoelzer et al. 2007, 2010]. Adult CD20+ pre-B ALL patients (15–55 years of age, n = 181) received rituximab (375 mg/m2) before each induction course and before each consolidation course for a total of 8 doses [Hoelzer et al. 2010]. Patients were compared with patients recruited earlier in the same study without rituximab (n = 82). These latter patients received the same chemotherapy regimen. Although there was no improvement in complete remission (CR) rate with the addition of rituximab in the standard risk group, the rate of molecular CR (MRD < 10–4) was higher at various time points in the standard risk group (57% versus 27% at day 24 and 90% versus 59% at week 16) [Hoelzer et al. 2010]. Importantly, the MRD detection was PCR-based (described in [Bruggemann et al. 2006]). This translated into an improved rate of continued CR at 3 years for patients receiving rituximab (64% versus 48%), p = 0.009 and overall survival at 5 years (80% versus 47%). Overall survival at 5 years was also increased in the high-risk group with the addition of rituximab (55% versus 36%).
Another study performed at MD AndersonCancer Center in United States demonstrated almost identical results in pre-B ALL, CD20+ patients < 60 years of age. However, the backbone chemotherapy regimen used in this trial was hyperCVAD (instead of a BFM-based regimen) and patients ⩾ 60 years of age were included [Thomas et al. 2006]. In addition, this trial used a historical control comparison instead of consecutive patients treated within the context of a clinical trial. In patients < 60 years of age, the addition of rituximab was associated with an improved rate of CR duration and overall survival compared with historical controls at 3 years (91% versus 66%, p < 0.02 and 89% versus 53%, p < 0.01) [Thomas et al. 2006]. However, this benefit did not extend to patients ⩾ 60 years of age, suggesting a different biology in these latter patients. Although a phase III study is ongoing in Europe to address the benefit of rituximab in a randomized manner, many institutions, including ours, are incorporating rituximab into the upfront treatment of CD20+ pre-B ALL patients < 60 years of age given the strong data from these two studies and the correlation of MRD response with these results.
Ofatumumab
Another antibody engineered to recognize and bind the large and small loop of CD20 is ofatumumab (reviewed in [Reagan and Castillo, 2014]). Binding occurs in an area distinct to that of rituximab. Ofatumumab has been US Food and Drug Federation (FDA) approved for patients with relapsed/refractory chronic lymphocytic leukemia (CLL) and studies have revealed robust complement-dependent cytotoxicity (CDC) and antibody-dependent T-cell-mediated cytotoxicity (ADCC) mechanisms of action, making it an attractive therapeutic agent [O’Brien and Osterborg, 2010; Osterborg, 2010]. In a phase II trial, newly diagnosed CD20+ pre-B ALL patients received hyperCVAD in combination with ofatumumab. All but one patient achieved CR and MRD negativity (measured by flow), with 64% of patients achieving MRD negativity after induction. The 1-year continuous rate of remission and overall survival were both 91% [Jabbour et al. 2014]. More than two-thirds of patients developed febrile neutropenia. Other grade 3 or above toxicities included elevated transaminases and bilirubin (30% and 26% respectively). However, these toxicities were attributed to chemotherapy and not ofatumumab. These results are encouraging and further follow up and larger studies will be needed to determine the position of ofatumumab in the treatment of ALL.
Obinutuzumab
Obinutuzumab has potent activity in CLL and has been FDA approved for the upfront treatment of CLL. No clinical studies in ALL have been performed yet with the glycoengineered type 2 anti-CD20 antibody. However, encouraging preclinical results have been reported [Awasthi et al. 2014] with obinutuzumab in ALL cell lines and xenografts. Future studies are awaited as it may offer another option for the treatment of patients with ALL.
Anti-CD22 antibodies
Epratuzumab
CD22 has been another focus of targeting leukemic cells as it is expressed in the overwhelming majority of B-cell leukemias as previously discussed. Epratuzumab is a humanized antibody which attracted attention in the treatment of adult pre-B ALL based on encouraging preclinical data as well as trials in patients with non-Hodgkin lymphoma (NHL) and pediatric ALL [Furman et al. 2004; Leonard et al. 2004; Raetz et al. 2008].
A trial through the adult cooperative group, SWOG (S0910) evaluated epratuzumab in combination with a backbone chemotherapy regimen of clofarabine and cytarabine (as described in [Advani et al. 2010]) in CD22+ relapsed/refractory pre-B ALL [Advani et al. 2014]. CD22+ was defined as ⩾20%. Epratuzumab was administered on a weekly schedule at a dose of 360 mg/m2 for four doses. No allergic reactions to epratuzumab were noted. The median age was 41 years (range, 21–69) and 61% of patients were in first relapse. A total of 10% of patients had received a prior AHSCT. The CR/CRi (CRi: CR with incomplete count recovery) rate was 52% and 1 patient achieved MRD negativity which compared favorably with a prior trial of clofarabine/cytarabine (S0530) [Advani et al. 2010]. However, a randomized comparison is needed to address the true additional benefit of epratuzumab and a phase III study in children with relapsed/refractory ALL (IntReALL) is ongoing and should help answer this question.
A French phase II study investigated the benefit of epratuzumab combined with vincristine and dexamethasone in older adult CD22+ patients (92% of patients were above 55 years old) with relapsed/refractory ALL [Chevallier et al. 2015]. CD22+ in this study was defined as ⩾30%. Ph chromosome-positive ALL cases were included and 12% of the 25 evaluable patients had received prior AHSCT. The combination of dexamethasone followed by epratuzumab was given on Days 1,8,15, and 22. CR/CRi was noted in 20% of the patients with a median leukemia free survival of 3.8 months and a median overall survival of 4 months. Overall, treatment was well tolerated with pancytopenia being the major adverse event. Given the heterogeneity of cases included and the small number of patients, it is difficult to fully assess the benefit of this combination. Based on the above studies, epratuzumab seems to have limited single agent activity and needs to be combined with other cytotoxic therapy to have significant benefit. The phase III randomized study in children with relapsed/refractory ALL, noted above, should help further address this question.
Anti-CD22 immunotoxin, HA22 (CAT-8015)
The quest for more effective antibody constructs against lymphoblasts has led to the development of antibodies conjugated to potent toxin moieties. HA22 combines the lethality of the exotoxin A of Pseudomonas with the selectivity of targeting the CD22 epitope [Mussai et al. 2010]. HA22 development was based on the results of the first generation antibody (BL22) which had less potent affinity for CD22 [Mansfield et al. 1996]. The HA22 immunoconjugate is rapidly internalized upon binding to CD22 and the toxin inactivates elongation factor 2, which is a crucial component in the translation of mRNA sequences. Preclinical studies utilizing blasts from pediatric cases of ALL, established the potent cytotoxic activity of HA22 through apoptotic pathways [Mussai et al. 2010]. The study of an HA22-resistant ALL cell line uncovered a mechanism involving hypermethylation of the DPH4 promoter [Wei et al. 2012]. Incubation of this cell line with the hypomethylating agent 5-azacytidine prevented the development of resistance.
In the clinical trial setting, a phase I study evaluated HA22 in pediatric patients with relapsed/refractory CD22 positive B-ALL [Wayne et al. 2009]. Overall, one of the seven patients treated attained morphological and flow cytometric CR. However, one patient developed high titer neutralizing antibodies and a case of grade 4 vascular leak syndrome (VLS) was documented. The hallmark of VLS is the increased vascular permeability with subsequent development of fluid retention and hemodynamic compromise that may lead to organ failure [Baluna and Vitetta, 1997]. In another study of 21 patients (aged 4–21 years), 14 patients were treated with dexamethasone to reduce the capillary leak syndrome (CLS) [Wayne et al. 2011]. An overall response rate of 29% was reported [four CRs and one partial response (PR)] but a significant percentage of patients (14%) developed neutralizing antibodies to the immunoconjugate. A phase I–II trial is underway in adult patients with relapsed/refractory ALL [ClinicalTrials.gov identifier: NCT01891981].
Combotox
Another approach has combined anti-CD19 and anti-CD22 antibodies linked with a deglycosylated ricin-A chain through a thiol-containing cross-linker in a 1:1 ratio (Combotox) [Herrera et al. 2000, 2003, 2006]. The rationale was that ricin, which is a lethal toxin, could be delivered only in cells expressing CD19 or CD22 and preclinical (in vitro) studies revealed that the combination of two antibodies was more potent than either one [Herrera et al. 2000]. Moreover, the combination of two antibodies could suppress the evolution of lymphoblasts lacking either epitope, which is a potential mechanism of resistance to antibody-based therapy. The mechanisms that lead to augmented efficacy are not well understood.
A study in pediatric patients [Herrera et al. 2009] was followed by a phase I study in adults (aged 19–72 years) with relapsed or refractory ALL [Schindler et al. 2011]. Inclusion criteria included expression of CD19 or CD22 in at least 50% of lymphoblasts. The treatment was administered every other day (total 3 doses per cycle) with dose levels ranging between 3 mg/m2 (minimum dose) to 8 mg/m2 (maximum dose). Overall, 1 patient attained PR, 1 had stable disease, while 15 patients had disease progression. Peripheral blast count suppression was often not long lasting after administration of the third dose of Combotox. In terms of toxicity, the patient receiving the highest dose developed severe VLS; while grade 3 elevations of liver function tests was observed in 2 out of 7 patients receiving 7 mg/m2. Notably, in a mouse model the concomitant administration of Combotox and cytarabine demonstrated a strong synergistic effect [Barta et al. 2012]. A phase I clinical trial of this combination is in progress [ClinicalTrials.gov identifier: NCT01408160].
Inotuzumab ozogamicin
Since CD22 is expressed from the early stages of B-cell lineage, it is an attractive target for the treatment of B-ALL. Moreover, a recent study of 163 patients (age range 0.6–25 years) with relapsed/refractory pre B-ALL revealed that cases with less than 90% CD22 expression were rare [Shah et al. 2015]. Inotuzumab ozogamicin (IO) (reviewed in [Jain et al. 2014]) is a fusion of a monoclonal anti-CD22 antibody attached to calicheamicin, a potent cytotoxic agent derived from Micromonospora echinospora [Thorson et al. 2000; Dijoseph et al. 2004]. Similar to other immunoconjugates, IO attaches to CD22 and then exerts its cytotoxic effect once internalized by binding to the minor groove of the DNA, inducing double-strand DNA breaks [Zein et al. 1988]. This treatment is easier than blinatumomab (discussed below) in that no continuous infusion is needed and no hospitalization is required unless a patient requires it due to adverse events. However, there is a risk of veno-occlusive disease, particularly in patients receiving AHSCT. The details of this are discussed below. In addition, unlike the immunotoxins discussed above, there do not appear to be issues with respect to the development of neutralizing antibodies or CLS.
A phase II study of IO enrolled patients with relapsed or refractory pre-B ALL and tested a weekly versus a once every 3–4 weeks schedule [Kantarjian et al. 2013]. Overall, 90 patients were enrolled with 65% of patients being more than 18 but less than 60 years of age. A total of 90% of patients had an Eastern Cooperative Oncology Group (ECOG) performance status of 0–1. Salvage status (S) included S1 (32%) and ⩾S3 (30%). Overall, 23% of patients had a diploid karyotype, 17% were Ph chromosome-positive, and 9% had chromosomal translocation 4;11 [t(4;11)]. A total of 11% of patients had undergone prior AHSCT.
The overall response rate was 58%: 19% CR, 30% CR with incomplete platelet recovery (CRp), and 9% marrow CR with incomplete recovery of platelets or neutrophils. Patients with Ph + ALL or t(4;11) had a statistically significant inferior response rate compared with the rest of the enrolled patients. Patients in S1 had a more favorable response rate compared with those of S2 or higher (p = 0.056). The rate of MRD negativity by flow cytometry on the 50 patients assessed was 72%. The median overall survival was 6.2 months. In addition, patients with CR had a notably more favorable overall survival compared with those with resistant disease (approximately 13 months versus 3 months). Moreover, responses were short lived in the absence of AHSCT but a significant proportion of patients (36 out of 90 patients) were able to proceed to AHSCT. In terms of adverse events, liver enzyme and bilirubin elevations were less prevalent in the weekly schedule. However, grade 3–4 adverse effects were not different between the two schedules. In the weekly schedule, grade 3–4 elevated aminotransferase was noted in two patients compared with one in the other group of patients. Overall, two patients had grade 3–4 elevation of bilirubin in the once every 3–4 weeks schedule while in the weekly schedule, none of the patients had similar grade elevation. A recent study identified that prognostic markers predicting a lower response to IO were adverse baseline cytogenetics, such as complex karyotype or chromosome 17 abnormalities; Ph chromosome-positive chromosomal rearrangement, t(4;11), and a high peripheral blast count [Jabbour et al. 2015]. Not surprisingly, patients beyond salvage 1 had worse survival outcomes.
A recent abstract submission to the European Hematology Association 20th congress 2015 [Deangelo et al. 2015] reported results from a phase III trial of patients in salvage 1 or 2 treated with IO or standard of care chemotherapy. The results of the first 218 patients in an ‘intent to treat analysis’ revealed that CR/CRi, median duration of response and MRD negativity was statistically significant in favor of IO (CR/CRi 80.7 versus 33.3, p < 0.0001). Data regarding safety are available for 259 patients. The most common grade 3 or above adverse events were cytopenias and some cases of veno-occlusive liver disease (VOD). Grade 3 or above VOD was noted in 13 patients treated with IO versus 1 patient in standard of care chemotherapy. The incidence of VOD has varied in the different studies but this may reflect differences of the preparative regimen and the period of time from the last dose of IO to the time of AHSCT. There are reports with respect to gemtuzumab ozogamicin and VOD (same calicheamicin backbone) [Wadleigh et al. 2003] but not IO. Currently, the analysis of overall survival between the two arms is awaiting longer follow up.
Just recently, based on the results of this trial, IO received FDA breakthrough therapy designation and will be another good option for patients with relapsed/refractory disease. Given the impressive activity of IO, a current study [ClinicalTrials.gov identifier: NCT01925131] is looking at combining lower dose IO with cyclophosphamide, vincristine and prednisone (S1312) which may potentially decrease VOD and may potentially increase CR, given the synergy of this chemotherapy with IO.
Anti-CD19 antibodies
SAR3419
SAR3419 is an immunoconjugate construct combining a powerful microtubule disruptor with an anti-CD19 antibody. Namely, maytansine and its derivatives (DM1 and DM4) have a mechanism of action reminiscent of that of the vinca alkaloids but much more potent [Remillard et al. 1975]. Preclinical studies utilizing ALL xenografts revealed that SAR3419 had potent and specific activity against lymphoblasts without significant toxicity [Carol et al. 2013]. In addition, in the same study, SAR3419 demonstrated activity in preventing/delaying leukemic relapses in a model of chemoresistant ALL xenografts. Unfortunately, the phase II clinical trial MYRALL [ClinicalTrials.gov identifier: NCT01440179] that enrolled patients with relapsed or refractory ALL was terminated due to the modest activity of SAR3419.
SGN-CD19A
SGN-19A is another immunoconjugate targeting CD19 and links a humanized anti-CD19 antibody to monomethyl auristatin F. Auristatin is a potent inhibitor of tubulin assembly and can perturb the mitotic process [Doronina et al. 2003]. The proposed mechanism of action involves internalization of the immunoconjugate followed by the release of the auristatin through lysosome processing [Law et al. 2011]. The auristatin then interacts with tubulin and inhibits its polymerization. Interim results of a phase I study including patients with relapsed/refractory ALL and highly aggressive lymphoma were reported at a American Society of Hematology (ASH) 2014 meeting [Fathi et al. 2014]. In this study both adult (median age 37; n = 38) and pediatric patients (median age 11; n = 11) were enrolled. Overall, 40 out of the 49 enrolled patients had relapsed/refractory ALL. SGN-CD19A was administered in two schedules: once a week (0.3–2.3 mg/kg) or once every 3 weeks (4–5 mg/kg). In the weekly schedule, 6 out of the 25 patients with B-ALL or B-lymphoblastic lymphoma responded and 5 patients with CR/CRp, while on the once every 3 weeks schedule 4 out of 8 patients attained CR/CRp. Notably, 6 out of 8 patients with CR/CRp attained MRD negativity. Drug-related toxicities appeared to be similar in cohorts and included nausea, dry eyes, and blurry vision. Grade 3–4 corneal events were observed in 34% of adult patients and 9% of pediatric patients, likely related to the drug. Other notable adverse events and drug-limiting toxicities in adult patients included a cytokine release syndrome (CRS) case and one grade 4 transaminase elevation.
Blinatumomab
The quest for more powerful and targeted engagement of leukemic cells has led to efforts of cell surface engineering [Swartz et al. 2012] using host cytotoxic T cells to eliminate lymphoblasts. Blinatumomab is the FDA-approved single chain bispecific (anti-CD19 and anti-CD3) T-cell-enganger antibody (BiTE) for the treatment of relapsed/refractory Ph-negative ALL (recently reviewed in Bumma et al. 2015). Preclinical studies demonstrated that the ingenious engineering to link two potent antibody moieties that recognize CD3 and CD19 was associated with high specificity but also potent cytotoxic activity [Loffler et al. 2000; Dreier et al. 2002]. The mechanism of action is based on bringing in close proximity T cells of the patient (through the anti-CD3 antibody moiety) with B-lymphoblasts (through anti-CD19). The next step involves the release of cytotoxic granules by the T cell and activation of the perforin–granzyme pathway in the B-lymphoblast. Importantly, the activated T cells are able to engage serially multiple leukemic cells [Hoffmann et al. 2005]. Although, blinatumomab does not require a second costimulatory molecule to exert its action it is not plagued by aberrant or uncontrolled activation of T cells. Initial studies in NHL patients revealed that optimal administration requires continuous infusion due to the short half-life and initial studies with short infusion schedules were discontinued secondary to diminished benefit/risk outcomes [Nagorsen et al. 2012].
The initial studies evaluating blinatumomab in B-ALL were in MRD-positive disease. This is a unique setting to test new drugs and likely will be a platform to evaluate other drugs in the future. The results with blinatumomab in this initial trial were impressive [Topp et al. 2011]. Patients must have achieved a CR but could have persistent MRD or have relapsed MRD following front line treatment/post consolidation. Blinatumomab was administered as a continuous intravenous infusion (15 µg/m2/day) for 4 weeks with brief inpatient hospitalizations at the beginning of each cycle to monitor for CRS. Overall, 80% of patients achieved a complete molecular remission. Typically, patients with MRD-positive disease have a poor prognosis but the reported relapse-free survival in a follow-up study at a median of 33 months remained at an impressive 61% [Topp et al. 2012]. Some patients have had durable remissions without proceeding to AHSCT. Importantly, two of the four patients that relapsed had CD19 negative lymphoblasts. The most common adverse events included symptoms of asthenia, fevers, chills, as well as metabolic derangements and decreased immunoglobulin levels. Grade 3–4 adverse events included leukopenia, lymphopenia, catheter related infections, and neurological complications (headache, syncope, seizure, somnolence; n = 1 for each preceding neurological complication). All grade 3–4 adverse events were transient apart from a patient with leukopenia. Blinatumomab was permanently discontinued in the case of a severe seizure.
Another key publication is the phase II single-arm open-label study of blinatumomab in patients with Ph chromosome negative relapsed/refractory ALL [Topp et al. 2014]. A total of 36 patients (aged 18–77 years) were treated in three cohorts of dose escalation. The first cohort was treated with 15 µg/m2/day while the second had a lower starting dose prior to attaining the same dose as cohort 1. The third cohort dose was escalated to 30 µg/m2/day towards the end of the first cycle and thereafter. Overall, 69% of patients attained CR or CR with partial hematologic recovery with 22 patients (88%) achieving MRD negativity (MRD level below 10–4 by PCR). The median overall survival was 13.2 months for those achieving CR. Approximately half of patients who achieved CR or CR with partial hematologic recovery underwent AHSCT. Overall, 66% of patients who did not undergo AHSCT relapsed. The side effect profile was similar to the previous study. A total of 17% of patients (n = 6) developed nervous system or psychiatric complications necessitating intervention, and in 2 patients blinatumomab was permanently discontinued. Overall, two patients with a high disease burden also developed grade 4 CRS.
More recently, results from 189 patients with Ph chromosome-negative relapsed/refractory ALL treated with blinatumomab in the context of an international phase II study were published [Topp et al. 2015]. The median age of enrolled patients was 39 years (range: 18–79). Prior lines of salvage therapy included: 1 (41%), 2 (22%), and >2 (17%). A CR/CR with partial recovery of peripheral counts was noted in 43% of patients, lower than in the smaller study discussed above. Response to blinatumomab did not correlate with age, salvage status, or AHSCT but did correlate with blast percentage. Fewer than 50% of bone marrow blasts were associated with better outcome. The two most common adverse events were neurological events (52%) and febrile neutropenia (28%). Neurological events were mainly grade 1–2 (39%). Almost half of all neurological events involved tremors, dizziness, confusion, ataxia, encephalopathy and somnolence. Moreover, grade 4 adverse events were noted in 30% of patients, with neutropenia being the most common of these. A pre-phase treatment with steroids for patients with high leukemic burden and lower starting dose of blinatumomab ameliorated CRS.
Currently, blinatumomab is FDA-approved for the treatment of relapsed/refractory Ph chromosome-negative ALL. We have been increasingly using this treatment (unless a clinical trial is available) for our patients with relapsed/refractory disease. This treatment has generally been well tolerated unless a patient develops neurologic symptoms or CRS. We do treat our patients with high tumor burden with dexamethasone prior to administration. Although this drug has been approved, a phase III study (TOWER) was conducted evaluating blinatumomab versus standard chemotherapy in relapsed ALL and this study has recently completed accrual. The major disadvantages of this treatment are the need for continuous infusion, bag changes, and the cost. However, home care can also be utilized. Once patients are through the period of the CRS, if they do not have neutropenic fever, they can be treated as outpatients. The data in Ph chromosome-positive patients is not clear since this group was not included in the initial trial, but is likely to be beneficial and trials are planned. The future use of blinatumomab is likely to expand and there are current trials testing its efficacy upfront in combination with chemotherapy for adults with Ph chromosome-negative ALL (E1910), upfront as a single agent in older patients with Ph chromosome-negative ALL (S1318), and a larger study in the setting of MRD positive disease (BLAST).
Anti-CD52 antibodies
Alemtuzumab
CD52 has attracted attention in the context of immunotherapy as it is highly expressed in most lymphoblasts [Ginaldi et al. 1998] and can be targeted in both B and T cell ALL, whereas most of the current antibodies target B-ALL. There is no consensus regarding expression of CD52 in hematopoietic stem cells [Gilleece and Dexter, 1993; Morisot et al. 2006]. Experiments employing transgenic mice expressing human CD52 treated with the monoclonal anti-CD52 antibody, alemtuzumab, demonstrated depletion of circulating CD52 cells [Hu et al. 2009]. However, lymphocyte depletion from lymphoid organs was not as robust. This study indicated that alemtuzumab action was mainly through ADCC but in vitro experiments have indicated that complement may also have a role [Zent et al. 2008]. Hence, the mechanism of action of alemtuzumab is not completely elucidated [Alinari et al. 2007] and its use is accompanied by transient cytopenias. Early case reports reported that alemtuzumab had limited single agent activity in ALL patients relapsing after AHSCT or after standard chemotherapy [Piccaluga et al. 2005].
More recently, a phase II study from France reported the use of alemtuzumab with mandatory coadministration of granulocyte colony stimulating factor (G-CSF) in patients with ALL [Gorin et al. 2013]. The study was designated in a two-step scheme. The first part encompassed escalating dosing of alemtuzumab to 30 mg/dose and then administration thrice weekly for 12–18 infusions. In maintenance, (second step) only patients who tolerated the 30 mg/infusion and were in CR were allowed. Overall, 4 of the 12 patients entered CR but responses were not long lasting (all patients but one that received donor lymphocyte infusions progressed). Adverse events included infections (aspergillosis, cytomegalovirus (CMV) pulmonary infection and CMV reactivation). Moreover, one patient developed tumor lysis while 25% of patients developed fever, chills and rash. These toxicities are similar to what has been observed using alemtuzumab in the treatment of relapsed/refractory CLL, where it has been used the most to date.
Given the toxicity and relatively low response rate, alemtuzumab is unlikely to be an integral part of treatment for relapsed disease. Data using alemtuzumab in newly diagnosed ALL [Stock et al. 2009] to decrease MRD seem more encouraging. Stock and colleagues treated patients with newly diagnosed ALL with a ‘module’ of alemtuzumab after standard induction and post-remission therapy. The relapse-free survival was encouraging given the population and historical data. However, toxicity was still experienced and hematological toxicities included grade 3–4 myelosuppression and in post-remission therapy cases of CMV viremia, herpes zoster reactivation and herpes simplex infection were reported. To this end, and given other new agents (blinatumomab), this is unlikely to be further developed in B-ALL, although consideration should be given in T-ALL.
Chimeric antigen receptor T cells
Introduction
The paradigm of blinatumomab indicated that targeting lymphoblasts for destruction by the patient immune system is feasible. However, it required the use of an exogenous antibody that acted as a bridge between the patient’s T cells and the lymphoblast. The possibility of ex vivo engineering of patient’s T cells to recognize lymphoblasts is a different, compelling approach that falls under the spectrum of adoptive cellular immunotherapies (discussed in [Dotti et al. 2014; Gill and Porter, 2014]). Patient T cells must be harvested through leukapheresis. The cells are then manipulated by inserting the genetic information necessary to express the targeting protein complex recognizing the lymphoblasts; resulting cells are called chimeric antigen receptor (CAR) T cells. The gene vectors used are mainly lentiviral or retroviral [Kochenderfer and Rosenberg, 2013]. In addition, incorporation of signaling/costimulatory domains into the construct was shown to prolong persistence of engineered cells [Kowolik et al. 2006; Milone et al. 2009]. CD28 or CD137 (4-1BB) were used successfully as costimulatory domains with more advanced constructs (third generation) composed by multiple costimulatory domains (discussed in [Grupp, 2014]).
Several groups have published their experience with CAR T cells and a comprehensive review was recently published [Tasian and Gardner, 2015].
Preclinical data and clinical trials
Studies using a single costimulatory domain targeting B lymphoblasts were pursued at the University of Pennsylvania, USA using the CD19/4-1BB/CD3 zeta construct. T cells engineered ex vivo to carry this construct are called CTL019.
CTL019 cells demonstrated antileukemic activity in CLL [Porter et al. 2011]. CTL019 cells were then used in two pediatric patients with ALL; both had relapsed ALL and remission could not be achieved despite using multiple lines of treatments [Grupp et al. 2013]. Overall, one of the two patients had a sustained CR with CTL019 while the second relapsed with CD19 negative lymphoblasts.
A follow up landmark report included 25 children and 5 adults with relapsed or refractory CD19+ ALL using (lentivirally transduced) CTL019-engineered T cells [Maude et al. 2014]. The majority of patients had B-ALL with multiple relapses, and 72% of the pediatric cases had undergone AHSCT. Notably, three patients who were refractory to blinatumomab were also included as well as a patient with CD19+ T-ALL. The overwhelming majority of the patients received chemotherapy in an attempt to deplete T cells prior to infusion of the ex vivo manipulated T cells. An impressive 90% of patients achieved a morphological CR. Overall, 22 out of 25 patients tested were also MRD negative 1 month after infusion. In the same study, CTL019 also exhibited activity in patients with documented blasts in the CNS and in 66% of patients who were refractory to blinatumomab. However, 30% of patients with CR experienced relapse including one patient who developed myelodysplastic syndrome and evolution to acute myeloid leukemia [Maude et al. 2014]. The overall survival was 78% at 6 months but it should be noted that relapses occurred rapidly (approximate range 1.5–8.5 months). To this end, long term data from a larger series of patients are needed to estimate progression-free survival and overall survival. In patients who had received prior AHSCT, CTL019 was not associated with graft versus host disease (GVHD) or with an inferior overall survival.
Reminiscent of the side effects noted in patients treated with blinatumomab, patients in this trial also experienced symptoms related to CRS as well as CNS toxicity. However, the incidence and severity of the side effects were higher compared with those reported with blinatumomab. The side effect profile is not unexpected, as T-cell activation is a key step in this treatment modality, although the mechanism of CNS toxicity is not well understood. All patients in this study developed CRS but in a significant percentage of patients (27%) CRS was severe, necessitating intensive medical care [Maude et al. 2014]. Hypotension and respiratory distress as well as coagulation derangements (notably low fibrinogen) were the main manifestations in this subgroup of patients. Tocilizumab proved to be an effective treatment with rapid resolution of CRS manifestations but readministration was necessary in a few cases. A sizeable number of patients (n = 13) experienced neurological adverse events. The neurological adverse events ranged from delirium to generalized encephalopathy. No persistent neurological adverse events or their sequelae are reported. Of note, fewer than 3% CD19 positive lymphocytes were noted on essentially all patients with response and infusions of intravenous immunoglobulin (IVIG) were administrated to prevent infections.
In a phase I study from the US National Cancer Institute (NCI) [Lee et al. 2015a] which included patients with relapsed or refractory CD19+ ALL or NHL, a different construct (retroviral-expressing CD28) was used. A total of 21 patients were enrolled (age range 1–30 years) with 7 patients having refractory disease and number of relapses ranging from 1–8. Impressively, 70% of the 20 patients with ALL had a CR while 60% attained MRD negativity. Notably, two patients with MRD negativity who were not transplanted relapsed within 5 months. CRS occurred in 16 patients but it was grade 4 only in 3 patients and required tocilizumab in three patients (in one case combined with steroids). Overall, six patients experienced neurological toxicity which was reversible. Frequent grade 3 toxicities included anemia (68%), fever (43%) and hypokalemia (43%). Grade 4 hematological toxicities included decrease in neutrophil count (89%) and leukopenia (68%).
Encouraging results were also reported from a different group [Brentjens et al. 2013]. Overall, five patients with relapsed B-ALL (age range: 23–66 years old) were included with two patients being MRD positive and two having refractory disease. The duration of CR ranged between 5 weeks to 34 months and 3 out of 5 patients had a normal karyotype. All patients received conditioning with cyclophosphamide prior to CAR T-cell infusion and subsequently attained MRD negativity. A total of four out of the five patients subsequently went to AHSCT and three remained in CR (the fourth patient died from pulmonary embolism while in CR). The fifth patient although attained MRD negativity, relapsed in 3 months (the patient was not eligible for AHSCT). Grade 3 febrile neutropenia was noted in three patients while one patient had grade 4 neutropenia. Hypotension was noted in the majority of patients.
Although the previous studies focused on targeting CD19, CAR T cells are being also engineered to recognize CD22 and preclinical data revealed potent cytotoxicity [Haso et al. 2013]. More specifically, NALM6-GL (B-ALL cell line) tumor cells were infused in a xenotransplant mouse model. CAR T cells, engineered to recognize CD22 or mock-transduced CAR T cells, were infused three days later in these mice. There was a statistically significant survival difference in favor of CD22 recognizing CAR T cells.
Several other studies have been reported in abstract form (reviewed in [Tasian and Gardner, 2015]) signifying the great potential of CAR T cells.
Outlook of chimeric antigen receptor T-cell treatments
Clinicians should be familiar with CRS and the other common adverse events associated with the use of CAR T cells. It is important to note that the remarkable CR achieved with the CAR T-cell approach was not sustained in all patients and long=term data are not yet available. Therefore, it is not clear if this approach will be able to replace AHSCT. However, CAR T cells are a particularly attractive approach in patients who have failed AHSCT. The long-term persistence of CAR T cells and durable remissions in some patients who have been refractory to multiple therapies is extremely encouraging.
Kinase targeting therapies and immunomodulatory agents
Proteasome inhibitors: an outlook in the context of acute lymphoblastic leukemia
Proteasome inhibitors such as bortezomib and carfilzomib have changed the treatment paradigm in multiple myeloma [Cvek, 2012]. Their mechanism of action involves inhibition of the proteasome which affects critical pathways for cellular survival through nuclear factor Kβ signaling, Jun N-terminal kinase (JNK) pathway, histone deacetylase action and suppression of anti-apoptotic proteins [Mujtaba and Dou, 2011; Dou and Zonder, 2014]. Other studies using cell lines derived from T-ALL indicated that expression of Apaf-1 has an important role in bortezomib-induced apoptosis [Ottosson-Wadlund et al. 2013]. The Nalm-6 cell line [Adachi et al. 2006] recapitulates aspects of B-ALL biology and in vitro experiments using bortezomib highlighted an array of antileukemic mechanisms. Namely, bortezomib induced a cell cycle arrest at the G2/M phase and promoted apoptosis and formation of autophagic vacuoles [Wang et al. 2015]. Other studies focusing on B-CLL cell lines have again highlighted a caspase apoptotic pathway as an important cytotoxic mechanism of proteasome inhibitors in the context of leukemia [Almond et al. 2001]. However, the precise mechanisms of action have not been completely elucidated. There have been few reports of their use in the treatment of patients with ALL.
Proteasome inhibitors: focus on bortezomib
Bortezomib has been employed in pediatric trials of relapsed/refractory ALL (discussed in [Du and Chen, 2013]) but in adults, data are more limited [Du and Chen, 2013]. A phase I study of bortezomib included three adult patients with ALL (the majority of patients in this study had AML) [Cortes et al. 2004]. The maximum tolerated dose was 1.25 mg/m2 in a biweekly schedule of 4 weeks (out of 6 weeks). The effect of bortezomib appeared to be very limited and of short duration. Dose limiting toxicities included diarrhea, orthostatic hypotension (prolonged and unresponsive to fluids) and fluid retention.
A recent study was reported where adult patients (n = 9) with relapsed/refractory ALL received bortezomib in combination in most cases with HyperCVAD [Zhao et al. 2015b]. All but one patient achieved CR and (following consolidation) underwent AHSCT. Three patients who attained CR ultimately relapsed. The Therapeutic Advances in Childhood Leukemia and Lymphoma group (TACL) demonstrated similar encouraging results in a pediatric study. Bortezomib was combined with vincristine, dexamethasone, pegylated asparaginase, and doxorubicin in a phase II study [Messinger et al. 2012]. The overall response rate was encouraging at 73%, meeting predefined criteria for early closure. Future studies will be needed to further delineate any potential role of bortezomib in ALL treatment. Carfilzomib is a more potent proteasome inhibitor, and may be another strategy to investigate in further studies.
Spleen tyrosine kinase inhibitors, C61
Lymphoblasts must circumvent pathways that would otherwise prevent uncontrollable proliferation and resistance to programmed cell death. The role of spleen tyrosine kinase (SYK) in malignant hematology has recently attracted attention as a potent anti-apoptotic regulator [Geahlen, 2014; Krisenko and Geahlen, 2015]. Recent preclinical studies have identified C61 as a potent competitive inhibitor of SYK (discussed in [Uckun and Qazi, 2014]). Attempts to enhance stability of C61 and delay clearance have been reported and discussed [Zhao et al. 2015a]. Clinical trials using SYK inhibitors would be a welcome addition to the growing armamentarium of agents targeting ALL and current trials are ongoing.
Targeting Philadelphia-like acute lymphoblastic leukemia
TKIs can target aberrations in Ph-like ALL. Experiments employing xenografts demonstrated efficacy of ruxolitinib and rapamycin [Maude et al. 2012].
Moreover, a seminal case of a pediatric case with EBF1-PDGFRB-positive ALL refractory to conventional induction treatment responded to imatinib [Weston et al. 2013]. In another report, dasatinib was used successfully in the case of an ATF7IP/PDGFRB translocation [Kobayashi et al. 2015]. In the ASH meeting of 2014 a study of 154 patients with Ph-like ALL revealed rearrangements of the erythropoietin receptor (EPOR) in a small subset of patients. These patients may benefit from inhibition of the JAK-STAT pathway [Iacobucci et al. 2016]. Importantly, Roberts and coworkers reported on a small number of patients (11 patients with follow-up data; only one adult) with Ph-like ALL who were treated with imatinib, dasatinib or ruxolitinib with encouraging results [Roberts et al. 2014a]. A clinical trial sponsored by the MD Anderson Cancer Center in United States is currently underway exploring the use of ruxolitinib or dasatinib (based on the molecular profile) in patients with Ph-like ALL [ClinicalTrials.gov identifier: NCT02420717].
Conclusion
The development of new and highly-active targeted therapies for ALL is an exciting and rapidly expanding field. Clinical trials that will bring new therapies in the frontline treatment are starting to occur and their results are eagerly anticipated. Moreover, ALL relapses, despite highly targeted therapies, are indicative of the resilience of the leukemic clones and the need to further understand the molecular signature of leukemic clones. Perhaps the more challenging questions in the short term are the optimal sequence and timing of novel therapies for relapsed/refractory leukemia and if such treatments should be moved into frontline regimens. Moreover, the discovery of predictive markers for response to novel therapies and the role of AHSCT in the era of new treatments are questions that will tantalize the field of ALL research in the future.
Footnotes
Acknowledgements
We would like to thank Dr. Cotta from the department of pathology in the Cleveland Clinic, USA for reviewing part of the manuscript and his suggestions. We apologize to our colleagues that their work is not cited due to space limitations.
Funding
Nikolaos Papadantonakis has no financial disclosures.
Anjali S. Advani has the following financial disclosures: research funding and honoraria from Pfizer, research funding from Amgen and honoraria from Seattle Genetics.
Conflict of interest statement
The authors declare that there is no conflict of interest.
