Abstract
Acute lymphoblastic leukemia (ALL) is a heterogeneous group of hematologic malignancies that arise from clonal proliferation of immature lymphoid cells in the bone marrow, peripheral blood and other organs. There are approximately 3000 new adult cases diagnosed every year in the United States with a 5-year overall survival ranging from 22% to 50%. Most adult patients with ALL who achieve a complete response will ultimately relapse and for this subset of patients the only hope of curative therapy is successful re-induction to achieve a complete response followed by allogeneic transplant. Conventional vincristine has been used in all phases of ALL therapy but its efficacy is limited by cumulative toxicity, typically neuropathic in nature. Historically, the dose of conventional vincristine has been capped at 2 mg to avoid severe neurotoxicity. Liposomal vincristine [as vincristine sulfate liposomal injection (VSLI)] constitutes encapsulating vincristine in a sphingomyelin/cholesterol envelope. This process is thought to enhance drug delivery to the target tissues, decrease neurotoxicity by reducing the percentage of free drug in the plasma and therefore results in increased efficacy with acceptable toxicity. Results from recent trials using VSLI in the setting of relapsed/refractory Ph-negative ALL have been encouraging. VSLI as salvage monotherapy has been successful in inducing complete responses in a minority of adults with relapsed/refractory ALL so that they can be bridged to stem-cell transplantation. Rigorous post-approval testing needs to be conducted to clarify its utility in the clinic.
Introduction and background
Acute lymphoblastic leukemia (ALL) is a heterogeneous group of hematologic malignancies that arise from clonal proliferation of immature lymphoid cells in the bone marrow, peripheral blood and other organs. It represents approximately 80% of the acute leukemias of childhood but only about 20% of acute leukemias in adults [American Cancer Society, 2012]. However, since acute leukemia is much more common in adults than children, the total number of cases is similar in both groups with slightly more than 3000 cases/year in children and slightly less than 3000 cases/year in adults. Advances in understanding the dose and schedule of traditional chemotherapeutic agents have increased cure rates to 80% in pediatric ALL [Dores et al. 2012]. Despite a high frequency of complete responses (CRs) in adults, relapses are common and overall the likelihood of long-term survival is poor ranging from 22% to 50% at 5 years [Pulte et al. 2013]. The difference in outcomes between adult and pediatric ALL is explained partly by the mix of genetic subtypes and possibly also by the specific drug regimens employed. Pediatric chemotherapy regimens tend to be more aggressive in comparison and young adults treated on pediatric regimens may do better than those on conventional adult regimens [Stock et al. 2008]. However, a definitive, prospective phase III trial has not been performed.
Multiple induction regimens have been developed for adult patients in an attempt to improve response since achieving high proportions of CRs and rapidly achieving CRs have been associated with improved long-term outcomes [Larson et al. 1995; Kantarjian et al. 2004; Weiss et al. 1996; Tavernier et al. 2007].
Lamanna and colleagues conducted a multicenter, prospective randomized trial comparing the two-drug regimen consisting of cytarabine and mitoxantrone (ALL-2) to a conventional four-drug regimen (L-20) for induction and found that even though the ALL-2 regimen was superior to L-20 in terms of frequency of response and time to CR, long-term outcomes were not different, raising questions about whether a certain subset of adults with ALL have curable disease and the exact specifics of the regimen are relatively unimportant [Lamanna et al. 2013].
Most adult patients with ALL who achieve a CR will ultimately relapse. Once patients relapse the only hope of curative therapy is successful re-induction followed by allogeneic transplant. This fact was highlighted in a report by Fielding and colleagues who examined outcomes in 609 adults with recurring ALL and found that the type of treatment received after first remission did not have an influence on outcome after relapse except in a subgroup of people who went on to receive allogeneic stem-cell transplant (SCT) after relapse. The latter group showed durable responses in a small subset of patients [Fielding et al. 2007]. Thus, attaining a CR in order to bridge patients to SCT is currently the goal of salvage therapies.
Conventional vincristine
Vincristine has formed the backbone of most of the ALL regimens used in adults and in the pediatric age group for over half a century. It is a vinca alkaloid derived from the periwinkle plant which acts on the cell during active mitosis and causes depolymerization of the cellular micro-tubules causing metaphase arrest leading to apoptosis [Carter and Livingston, 1976; Owellen et al. 1972]. While, in vitro, the therapeutic efficacy of vincristine increases in proportion to duration of exposure of the neoplastic cells to the drug, the value of this in the clinic has not been demonstrated [Wood et al. 2001; Owellen et al. 1977]. Conventionally vincristine is given as a bolus injection and there is wide interpatient variation in the half-life, volume of distribution and drug clearance. There is rapid uptake of the drug into the cells as evidenced by the t1/2α of approximately 8 minutes. The t1/2β is about 14 hours which reflects lower plasma concentrations. Most studies in adults ‘cap’ the dose at 2 mg because attempts to give doses of more than 2 mg have resulted in significant neurotoxicity [Weiden and Wright, 1972]. This total dose of 2 mg for adults is typically much lower than the pediatric dose of 2 mg/m2 and some believe that it may relate to the poorer outcome seen in adult patients. This has led to formulation of new drug delivery systems intended to enhance adult dosing with an improvement in antitumor activity without an increase in the toxicity.
Liposomal formulations
Liposomes were first described in the 1960s [Bangham et al. 1965] and they soon generated considerable research as drug carriers for antimicrobial and anticancer formulations. Liposomes consist of nanoparticles which are spherical in shape and composed of a phospholipid bilayer which have the ability to encapsulate and deliver hydrophilic and lipophilic molecules. Since they resemble cell membranes, they are not immunogenic [Cattel et al. 2004]. One of the earliest liposomal antimicrobial formulations was amphotericin B. Clinical studies have demonstrated that liposomal amphotericin B had a similar efficacy but a better safety profile as compared with conventional amphotericin B. The incidence of infusion-related adverse effects as well as nephrotoxicity were significantly less in the liposomal amphotericin group. The dosage for the liposomal amphotericin B ranged from 3 to 6 mg/kg as compared with conventional amphotericin B (range 0.6–1 mg/kg) [Walsh et al. 1999; Leenders et al. 1997].
Another successful liposomal formulation that followed was doxorubicin. Again, the liposomal form had similar efficacy but appeared to have reduced cardiac toxicity associated with anthracyclines. When pegylated liposomal doxorubicin was compared with continuous infusion doxorubicin in the treatment of Ph-negative ALL, no differences in survival were observed between the two arms. There was a suggestion of a greater likelihood of induction failure with refractory disease (17% versus 3%, p = 0.10) and a higher cumulative incidence of relapse (52% versus 32% at 2 years, p = 0.20) in the pegylated liposomal doxorubicin arm. The pegylated doxorubicin arm had reduced toxicity compared with the conventional arm. However, the inferior antileukemic activity has limited its use in this disease [O’Brien, 2004; Hunault-Berger et al. 2011]. This failure of liposomal doxorubicin is a reminder to us that liposomal encapsulation does not necessarily ensure improved efficacy.
Liposomal vincristine is a three-part formulation which needs to be prepared by the pharmacist prior to administration. It consists of the cholesterol liposomes, vincristine sulfate and sodium phosphate which is used as a buffer. Encapsulation involves mixing 1 ml of vincristine with sphingomyelin/cholesterol liposomal injection in a sterile vial. Then sodium phosphate is added and this mixture is heated at 63°C (60–65°C) and mixed by gently shaking it for 10 minutes so that the liposomes are properly loaded. The vial then needs to be equilibrated for 30 minutes. Once prepared, the drug is stable for no more than 12 hours at controlled room temperature.
Liposomes are nanoparticles which are composed of a phospholipid bilayer with an aqueous core which can be used to encapsulate drugs. They are preferentially taken up by the process of passive diffusion by organs that have fenestrated microvessels and the reticuloendothelial system. This may lead to higher concentrations as well as prolonged exposure of the drug in the tumor tissue [Bangham et al. 1965; Cattel et al. 2004]. In addition, the toxicity profile may be improved as the amount of ‘free’ drug in the plasma is decreased and the liposomes have poor penetration into tissues with tight endothelial junctions [Allen and Cullis, 2013; Webb et al. 1998; Krishna et al. 2001]. Pharmacokinetic and murine model studies have shown an increased half-life of vincristine, decreased free drug levels in the circulation which may result in less drug toxicity and increased efficacy [Webb et al. 1995]. Vincristine sulfate liposomal injection (VSLI) has been shown to be more efficacious than the free drug resulting in improved survival in at least two different animal models [Mayer et al. 1990, 1993; Boman et al. 1994].
Phase I trial
In a phase I trial conducted by Gelmon and colleagues 25 patients with previously treated solid tumors were enrolled with dose escalations of VSLI ranging from 0.5 to 2.8 mg/m2 every 3 weeks with cohorts of at least three patients for each dose levels [Gelmon et al. 1999]. For all patients who received VSLI doses at or higher than 1.5 mg/m2, the area under the curve (AUC) was significantly increased as compared with the standard 2 mg dose of vincristine used in clinical practice. The dose-limiting toxicities were myalgias, constipation and peripheral neuropathy. Other grade 3–4 toxicities included fatigue (14%), alopecia (14%) and anemia (14%). Nausea, fever without demonstrable infection, granulocytopenia, thrombocytopenia, neuropathy and myalgias all occurred with an incidence of 7%. A dose of 2 mg/m2 over 1 hour for 21 days was recommended for phase II trials as there was increasing frequency of toxicities above levels of 2 mg/m2.
Pharmacokinetic data in the study revealed that there was a significant increase in Cmax and AUC with dose escalation. There was significant interpatient variability in vincristine concentration versus time curves, the latter being seen in other studies involving vincristine. Still, the systemic exposure was estimated to be up to 150 times higher as compared with free drug.
From a clinical point of view, there was one partial response in a patient with pancreatic cancer, two patients had transient improvements lasting less than the 4 weeks required for the classification of response and three patients had stable disease. Following completion of this phase I trial, the use of VSLI in solid tumors was abandoned.
Phase II trials in non-Hodgkin’s lymphoma
Vincristine has played an important role in the treatment of lymphoproliferative disorders and, therefore, lymphoma was chosen for phase II studies. Rodriguez and colleagues conducted a pivotal trial consisting of 77 patients with refractory, aggressive non-Hodgkin’s lymphoma (NHL) who had few therapeutic options [Rodriguez et al. 2009]. The dose of VSLI used was 2 mg/m2 over 1 hour administered every 2 weeks with a median total dose of 7.9 mg/m2 (4 doses). The median number of prior treatments was 3, and approximately one-third of the population had undergone prior SCT. As expected patients with disease that was sensitive to prior treatments did better than those that were refractory with 1-year overall survival of 61% and 33%, respectively. Median survival was also higher among those with disease sensitive to VSLI. The overall response rate was 25% with CRs in 5% of patients. Despite all patients being previously exposed to neurotoxic agents in their prior regimens, only 29% and 3% had grade 3 and 4 neurotoxicities, respectively. Other grade 3 and 4 toxicities reported were hematologic (69%), fatigue (7%) and pyrexia (2%)
Sarris and colleagues conducted a phase II trial of VSLI in patients with relapsed NHL using a VSLI dose of 2 mg/m2 every 14 days for up to 12 injections [Sarris et al. 2000]. A total of 35 patients were evaluable for response: most were heavily pretreated and all had received vincristine in prior regimens. Overall responses were seen in 40% with a greater degree of response seen in those patients with transformed and aggressive NHL. Among the responders, the median time to progression was 5.5 months. Grade 3 neurotoxicity was seen in 11 out of 35 patients and was the cause of termination of treatment in 5 patients. The trial was extended to patients with relapsed or refractory ALL once tolerability was established [Thomas et al. 2006]. A total of 16 heavily pretreated patients were included with 50% of them having Ph-negative chromosome ALL. Overall response frequency was 14% (one complete and one partial response) and an additional 36% (five patients) had reductions in marrow leukemia infiltrate (MLI) but subsequently progressed. Two patients went on to matched sibling allogeneic SCT including one who had a partial response to VSLI. One of those patients was still alive at the time of last follow up. No significant, unexpected toxicity was observed in the population but the dose intensity of VSLI was limited due toxicity. The median number of doses administered was only two [Thomas et al. 2006].
Phase I trial in ALL
To improve the efficacy and further assess tolerability in combination therapy, a phase I multicenter trial with VSLI and concurrent dexamethasone in relapsed or refractory ALL was conducted [Thomas et al. 2009]. The purpose was to assess the maximum tolerated dose (MTD) of a 4-week cycle. Dose escalations for VSLI ranged from 1.5 to 2.4 mg/m2. Pulse dose dexamethasone was given concurrently. The median number of doses that were given was 4 and the median cumulative dose of VSLI was 19.2 mg. Common toxicities included constipation, fatigue, peripheral neuropathy, anemia and fever. Grade 3–4 toxicities consisted of hematologic (28%), neurologic (17%) and endocrine toxicities (11%). The MTD was established at 2.25 mg/m2.
The overall response frequency was 22% with 19% (n = 7) achieving a CR and 3% (n = 1) a partial response. A total of 11% of the patients did not technically achieve a response but had hematologic improvement. Four out of the seven patients who achieved CR and one patient who had hematologic improvement went on to potentially curative SCT.
Phase II trial in Ph-negative ALL
This lead to the pivotal phase II single-arm, open-label trial of weekly 2.25 mg/m2 (no dose capping) of VSLI in adults with advanced, previously responsive, relapsed and refractory B- or T-cell lineage Ph-negative ALL [O’Brien et al. 2013]. Traditionally, patients with Ph-positive ALL have had a poor prognosis but with the advent of tyrosine kinase inhibitors (TKIs), the prognosis has improved and therefore such patients were excluded from the VSLI studies. All patients had previously been treated with standard vincristine and 77% had ongoing neuropathy. Of the 65 patients enrolled in the study, 48% had a history of prior allogeneic SCT and more than half had received 3 or more lines of therapy. Overall response rate in this population with VSLI monotherapy was 35% with 20% of the patients having a CR or a complete response with incomplete hematologic recovery (CRi). A total of 22% with CR/CRi had more than 5 prior lines of therapy and 26% had previously undergone allogeneic SCT. Responses were seen in both B- and T-cell disease. A total of 19% of the complete responders were successfully bridged to SCT with a median overall survival of 8.9 months. Two (3%) patients were still alive at the time of last follow up. In addition, the toxicity profile was predictable and was comparable with that of standard vincristine even though patients received a median of four doses. The most common serious adverse events were peripheral neuropathy, febrile neutropenia, constipation and tumor lysis syndrome. Grade 3 neuropathy was seen in 23%. Based on these results, the drug received approval from the US Food and Drug Administration on 9 August 2012 for the treatment of adult patients with Ph-negative ALL in second or greater relapse or whose disease has progressed following two or more antileukemia therapies. It is not commercially available outside the United States. There are several ongoing trials that continue to test liposomal vincristine in ALL and uveal melanoma (Table 1).
Ongoing clinical trials with VSLI.
HyperCVAD, hyperfractionated cyclophosphamide, vincristine doxorubicin, and dexamethasone alternating with cycles of methotrexate and cytarabine; ALL, acute lymphoblastic leukemia.
Discussion
Outcomes in the salvage setting for adult Ph-negative ALL remain dismal despite intensive clinical research over the last four decades. Although CR rates of 90% with initial multi-agent chemotherapy can be seen, most patients relapse and this event heralds a grim prognosis. Salvage therapy can induce CR in 20–40% of the patients in first relapse but, again, the disease-free survival is 2–6 months [Petersdorf et al. 2001]. The only potential curative treatment for this population is an allogeneic SCT but many patients cannot proceed to transplant due to suboptimal disease control. In contrast to myeloid leukemias, there is little graft-versus-leukemia (GVL) effect for ALL and therefore in order for patients to have the possibility of cure, they need to be in complete remission at the time of the transplant. VSLI, either as monotherapy or in combination, may be a potential treatment which can help bridge this high-risk group of patients. Conventional vincristine has been used in all phases of ALL therapy but its efficacy has been limited by dose-limiting toxicity. Because the dose is typically capped at 2 mg for adults, and this dose may be suboptimal, it may be one of the contributing factors for the superior results seen in children compared with adults. These pharmacokinetic limitations can be overcome by encapsulating vincristine in a sphingomyelin/cholesterol envelope. Liposomal vincristine is which may allow for increased efficacy and decreased toxicity. VSLI as salvage monotherapy has been successful in inducing CRs in a minority of adults with relapsed/refractory ALL so that they can be bridged to SCT.
In an attempt to understand the true impact of this agent in adult ALL we produced the following model: ALL is a rare disease with 1.7 newly diagnosed cases in 100,000 men and women per year. It is estimated that 6070 patients will be diagnosed with ALL in the United States in 2013 as per the SEER database [National Cancer Institute, 2013]. If we extrapolate from current statistics, approximately 40% (2458 cases) of the cases will occur in people over the age of 20. If we consider only people eligible for transplant (patients age less than 65 years), the total falls to 1802 cases. Out of the 1802 cases, 75% will have Ph-negative ALL (1351 cases). A total of 30–45% of 1351 cases (approximately 500 cases) will be cured with initial induction and consolidation therapy and about 5–11% (108 cases) will die during induction. Thus 745 cases annually will represent relapsed/refractory adult Ph-negative ALL suitable to be bridged to transplant. These are the patients that may be able to benefit from VSLI as a bridge to allogeneic SCT. Extrapolating from the phase II trial conducted by O’Brien and colleagues, 20% (149 cases) of those with relapsed/refractory adult Ph-negative ALL will have a CR or CRi which is necessary for allogeneic SCT [O’Brien et al. 2013]. Out of the 149 cases, 19% (28 cases) will go on to have a successful transplant. Thus, based on the aforementioned calculations, it appears that perhaps 28 patients/year may ultimately benefit from therapy with VSLI. Therefore, for VSLI to be a commercially viable product, it will need to have an expanded FDA indication to be used in other chemotherapy regimens. It has the potential to be used in frontline treatments for Ph-negative adult and pediatric ALL and aggressive NHL. It may also have a role in refractory lymphomas and some solid tumors such as Wilm’s tumor and sarcomas in the pediatric population. However, the challenges of increased cost and the difficulty in preparing the drug in the absence of convincing clinical data is an important caveat to its expanded use.
VSLI has been primarily studied in a younger population which may be bridged to transplant. Older patients may not benefit as much from such therapy and therefore the role of VSLI in this population warrants further investigation. Also, although it is possible to use higher doses with neuropathy comparable with conventional vincristine, neuropathy is still a major concern and will still be a cause of dose reductions and termination of treatments in some patients. In addition, loading of vincristine into liposomes may be fraught with potential technical difficulties in the community setting as it requires a degree of expertise and needs to be done immediately prior to administration. Finally, as with other liposomal preparations, even though the overall dose of VSLI is higher than that of conventional vincristine, we have yet to determine a VSLI dose which has the same therapeutic efficacy and toxicity as that of the conventional vincristine. In addition, VSLI has a wholesale price of US$9750 for a 5 mg vial, compared with conventional vincristine whose average wholesale price ranges from US$12 to US$40 for a 2 mg dose. Finally, in order to definitively establish whether or not VSLI is superior to conventional vincristine in Ph-negative ALL, a prospective randomized trial will need to be conducted.
Footnotes
Conflict of interest statement
The authors have no conflicts of interest to declare.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.