A Phase I Study of Quizartinib Combined with Chemotherapy in Relapsed Childhood Leukemia: A Therapeutic Advances in Childhood Leukemia & Lymphoma (TACL) Study

This article is featured in Highlights of This Issue, p. 3985

FMS-like tyrosine kinase 3 (FLT3) is a class III receptor tyrosine kinase that dimerizes and autophosphorylates when bound by FLT3 ligand, activating downstream pathways that induce cell growth and inhibit apoptosis. The most common FLT3 mutation is internal tandem duplication (ITD) on exon 14 of the FLT3 gene, resulting in constitutive activation and autophosphorylation of FLT3 (1, 2). In both adults and children with AML, FLT3-ITD mutations with high allelic burden (ratio > 0.4) confer a dismal prognosis (3–6). Therefore, the current standard of care for children with FLT3-ITD mutations is intensive induction chemotherapy followed by hematopoietic stem cell transplantation (HSCT). High levels of FLT3 wild-type receptor (FLT3-WT) also promote constitutive activation of the FLT3 receptor and carry poor prognosis similar to those with FLT3-ITD mutations (7, 8). Among children with AML, about 30 percent have FLT3 disease: FLT3-ITD mutations (15%), kinase domain (KD) mutations (5%), or overexpression of FLT3-WT (10%; refs. 6, 9). In addition, c-KIT is a receptor tyrosine kinase that is commonly overexpressed or mutated in childhood AML (19%) and serves as a potential target for therapy (10). In children with acute lymphoblastic leukemia (ALL), the highest levels of FLT3-WT expression occur in MLL-rearranged (MLL-r) patients, which account for 80% of infant and 5% of childhood ALL, and those with hyperdiploidy, which account for 30% of childhood ALL (11–13).

First-generation FLT3 inhibitors were developed for the treatment of solid tumors, demonstrating activity against a variety of kinases including RAF kinase, PDGFR, VEGFR, c-KIT, and FLT3. These less selective inhibitors demonstrated toxicity consistent with their off-target effects (14–16). In children, the first-generation FLT3 inhibitors lestaurtinib, midostaurin, and sorafenib have been studied most extensively. Quizartinib (AC220) is a second-generation FLT3 inhibitor demonstrating superior potency and selectivity in preclinical models when compared with first-generation inhibitors. In addition to nanomolar potency against FLT3-ITD mutations, quizartinib also demonstrates preclinical activity against FLT3-WT and c-KIT (17, 18). Clinical trials in adults with relapsed or refractory AML demonstrate clinical activity, manageable toxicity, and utility as a bridge to HSCT. The adult phase I single-agent clinical trial demonstrated responses in non-FLT3-ITD AML, leading investigators to hypothesize that quizartinib may have clinical activity against overexpression of FLT3-WT or c-KIT (19).

This is a first-in-child phase I study of quizartinib performed in the Therapeutic Advances in Childhood Leukemia/Lymphoma Consortium (TACL). The primary objective of this study was to determine a safe and biologically active dose of quizartinib given in sequential combination with cytarabine and etoposide (AE) in young patients with relapsed or refractory AML or MLL-r ALL.

The trial was open to accrual between September 2011 and March 2013. Data for all patients were current as of March 2015. Patients were required to be between >1 month and ≤21 years of age. Eligible patients were required to have ALL in ≥ first relapse (>25% bone marrow blasts) with MLL-r or hyperdiploidy or to have AML in ≥ first relapse with ≥ 5% bone marrow blasts. Children with relapsed AML were included regardless of FLT3-ITD status due to quizartinib activity against overexpression of WT-FLT3, c-KIT overexpression and mutations, single-agent clinical activity in adult FLT3-WT patients, and the combination of intensive AML reinduction therapy used in this study. Other requirements included adequate liver [serum bilirubin <1.5 times upper limit of normal (ULN) for age, ALT < 5 times ULN for age], renal (derived from the Schwartz formula), cardiac (echocardiogram with shortening fraction ≥ 27%), and adequate performance status (Karnofsky or Lansky >50%). Exclusion criteria included uncontrolled systemic infection, significant cardiovascular disease, and active central nervous system involvement (CNS3). Children with prolonged QT, using the Fridericia's formula to correct for heart rate (QTcF ≥ 450 ms) at study entry were not eligible. Institutional review boards at participating centers approved the study and participating patients or their parents signed written informed consent. The original clinical trial was registered at www.clinicaltrials.gov as NCT01411267.

Eligible patients received cytarabine (1 g/m²/dose i.v. every 12 hours) and etoposide (150 mg/m²/dose i.v. once daily) on days 1–5. Quizartinib was administered orally once daily on days 7–28. Patients were to receive up to 2 cycles of protocol therapy. Patients who achieved a complete response (CR) or complete response without platelet recovery (CRp) after the second course of therapy could exit the study to pursue alternative treatment or were eligible to continue single-agent quizartinib as continuation therapy.

Adult phase I studies determined the MTD to be 200 mg fixed daily dosing, with grade 3 QTcF prolongation proving dose limiting (19). However, PIA testing demonstrated biologic activity via complete inhibition of phospho-FLT3 in a limited number of FLT3-ITD patients at 18 mg daily and in a larger cohort of FLT3-ITD patients at 60 mg daily. Data from subsequent phase II studies have indicated that the optimal dose in adults balancing activity and reduced QT prolongation is 60 mg; however, adult phase I studies available during the development of this study demonstrated responses at 40 mg (20). In addition to dosing data described above, this clinical trial was the first to combine quizartinib with intensive chemotherapy despite not having single-agent pediatric toxicity data. Therefore, we selected 25 mg/m²/daily as dose level one (DL1). This is roughly equivalent to 40 mg fixed dosing using the average adult body surface area of 1.73 m².

A 3+3 design (21) escalated doses of quizartinib in approximately 50% dose increments beginning with 25 mg/m²/once daily. Blood from each patient was assayed for biologic activity using PIA testing. If DL1, DL2 (40 mg/m²/day), and DL3 (60 mg/m²/day) were tolerated, escalation would stop for assessment of quizartinib PIA. Additional patients would be accrued at DL3 to ensure that PIA was assessable in at least 9 patients; if 7 of 9 patients achieved PIA of >90% at 3 of 4 trough time points, DL3 would be considered biologically active and escalation would stop. If not, dose escalation would proceed to DL4 (90 mg/m²/day) and DL5 (130 mg/m²/day). Secondary endpoints included the achievement of CR, determination of FLT3 PIA levels, and in vitro sensitivity testing. Two-way ANOVA was utilized to investigate differences in marrow blasts inhibition among the patient groups as a result of increasing doses of quizartinib. The Spearman rank correlation was used to assess the relationship between the quizartinib in vitro sensitivity and change in bone marrow blasts on study. Change in bone marrow blast count was evaluated using a paired t test. Unless otherwise stated, P values refer to two-sided tests.

Dose-limiting toxicity (DLT) was assessed in the first course only. DLT was defined as a serious adverse event at least possibly attributable to quizartinib. The DLT definition was consistent with published principles adopted for Children's Oncology Group (COG) relapsed acute leukemia clinical trials that combine molecularly targeted drugs with intensive chemotherapy. In general, the guidelines incorporate tolerance of resolving grade 3 and some resolving grade 4 toxicities with stringent safety monitoring (22). On the basis of prior COG studies utilizing cytarabine/etoposide in AML, nonhematologic toxicities occurring in >10% of patients were either excluded (anorexia/alopecia) or required resolution with supportive care within a specified period of time (nausea/vomiting/diarrhea/metabolic abnormalities). Attributions were made by individual investigators and reviewed by the study committee. Hematologic DLT was defined as failure to recover a peripheral ANC > 500/μL and nontransfusion–dependent platelet count > 20,000/μL due to documented bone marrow aplasia/hypoplasia (as opposed to malignant infiltration or other cause) for greater than or equal to 56 days, which is approximately 2 SDs greater than the average duration of neutropenia in the aforementioned cytarabine/etoposide regimens.

Any patient who received at least one dose of study drug was included in the toxicity tabulations. Any patient that experienced a DLT as defined in the protocol only needed to have received one dose to be evaluable for toxicity. Otherwise, patients must have received at least 75% of the prescribed doses of quizartinib (i.e., at least 17 of the prescribed 22 doses of quizartinib) on study to be evaluable for DLT. Patients who were not evaluable for DLT were replaced.

A patient was considered evaluable for response if the patient received all or part of protocol therapy, and either (i) met the definition of progressive disease at any time, or (ii) had bone marrow collected and was under follow-up for a sufficient period to evaluate their disease at the end of course 1. Deaths due to toxicity after receiving all or part of protocol therapy were considered as nonresponders.

Adverse events were graded using the NCI Common Toxicity Criteria (CTCAE), version 4.0. Events were collected beginning with the first dose of study therapy until 30 days following removal from protocol therapy, and investigators evaluated each for severity and relationship to quizartinib and cytarabine/etoposide. Electrocardiograms (ECG) were performed at screening, and on days 7, 14, 21, 28 prior to quizartinib dosing and 2 and 4 hours postdose. Each ECG was performed in triplicate, reviewed locally, and transmitted for central review.

Bone marrow aspirates and/or biopsies were performed at baseline and on day 29 of each course. If the marrow was hypoplastic and peripheral blood counts had not recovered to ANC ≥ 1,000/μL and platelets ≥100,000/μL in the absence of leukemia, marrow testing was repeated not less than once every 14 days. Responses were defined per standard criteria and included CR, CRp, complete response with incomplete hematologic recovery (CRi), stable disease (SD), and progressive disease (PD; ref. 23). Of note, PD was defined as increase of at least 25% of the absolute number of bone marrow or circulating leukemic blasts, development of extramedullary disease, or other laboratory or clinical evidence of progression.

Detailed procedures have been published previously (24). The TF/ITD cell line was originally generated in the Small laboratory as described (23). Low passage aliquots have been maintained in the Brown laboratory. The cells are authenticated every 6 months (last March 2015) using RT-PCR for FLT3 juxtamembrane domain and sequencing of the PCR products to confirm expression of the original FLT3 ITD construct. Peripheral blood was drawn from patients before chemotherapy, before and 2 hours after the first dose of quizartinib on day 7 and at trough time points on days 14, 21, and 28. Plasma was isolated from peripheral blood by centrifugation. For each time point, 5 × 10⁶ TF/ITD cells were incubated with 500 μL of plasma for 1 hour. Clarified lysate was immunoprecipitated with anti-FLT3 antibody, electrophoresed with SDS-PAGE, and blotted with anti-phospho-tyrosine antibody, then stripped and reprobed with anti-FLT3 antibody. Proteins were visualized by chemiluminescence and densitometric analysis was performed.

Detailed procedures are published (25). RNA was isolated from baseline bone marrow or peripheral blood blasts enriched by Ficoll centrifugation. Contaminating germline DNA (gDNA) was removed from all samples by a DNase1 digestion, and 1 μg cDNA was synthesized. The juxtamembrane domain was amplified and gel electrophoresed, with the presence of an additional, higher molecular weight band indicating a FLT3/ITD. The kinase domain was amplified and digested with EcoRV. Pre- and postdigestion samples were gel electrophoresed, with a double banded profile indicating the presence of a point mutation. PCR products were sent to sequencing for confirmation.

RNA was isolated using the RNeasy kit (Qiagen). All samples were treated with DNase1 to remove contaminating gDNA. One microgram of RNA was used to synthesize cDNA using the Verso cDNA synthesis kit (Thermo Scientific). Quantitative real-time PCR was performed using SYBR green and primers JMD-F (5′-TGTCGAGCAGTACTCTAAACA-3′) and JMD-R (5′-ATCCTAGTACCTTCCCAAACTC-3′) with GAPDH as a control. Samples were run in duplicate on a CFX96 Real-Time instrument (Bio-Rad) and the results were analyzed on the Bio-Rad CFX96 Manager.

Bone marrow or peripheral blood blasts were enriched using density centrifugation (Ficoll-Paque Plus; GE Healthcare) from baseline bone marrow aspirates or whole blood. Remaining red blood cells were lysed using a red blood cell lysis buffer (0.155 mol/L NH⁴Cl, 0.01 mol/L KHCO³, 0.1 mmol/L EDTA) and washed with PBS. Cells were plated with increasing doses of quizartinib (0, 5, 10, 20, and 50 nmol/L) with DMSO control and maintained for 48 hours at 37°C + 5% CO₂ in AIM-V media (Life Technologies) supplemented with 20% FBS (Gemini Biosciences), 1% penicillin–streptomycin and 1% l-glutamine (Life Technologies). WST-1 (Roche) was added to each well and incubated for an additional 30 minutes and read within 4 hours on a Bio-Rad 680 Microplate Reader. Inhibition was plotted as a percentage of the untreated control.

Twenty-two eligible patients were enrolled (Table 1). The median age was significantly lower in ALL patients, reflecting that 3 of 4 patients had relapsed MLL-r infant leukemia. Of 18 AML patients, 9 were FLT3-ITD and 9 were FLT3-WT.

Eighteen patients were evaluable for toxicity. Four patients were not evaluable for toxicity according to protocol definition and were replaced (Supplementary Table S1). Toxicities were consistent with intensive relapsed leukemia regimens. None of 3 patients at DL1 experienced a DLT. One of 6 patients at DL2 experienced a DLT. This was a 19-year-old female with relapsed MLL-r ALL and a history of prior asparaginase-related pancreatitis who experienced grade 3 lipase elevation on study. One of 9 patients at DL3 experienced a DLT. This 19-year-old female with FLT3-ITD AML experienced grade 4 hyperbilirubinemia on day 12, and quizartinib was stopped on day 15 due to investigator decision. Attribution by the treating physician was documented as possibly related to quizartinib. The event did not resolve prior to the patient's death, which was due to overwhelming aspergillosis and multi-organ failure occurring 12 days after the last dose of quizartinib and 8 days after study removal. Toxicities attributable to quizartinib and chemotherapy for evaluable and inevaluable patients are listed in Tables 2 and 3, respectively. A dose-dependent increase in QTcF was observed (Supplementary Table S2). One patient experienced grade 3 QTcF prolongation, which occurred on the last day of quizartinib administration. This was not considered a DLT because it quickly resolved and did not limit further doses of quizartinib.

Seventeen patients were evaluable for response: 3 had ALL, 7 had FLT3-WT AML, and 7 had FLT3-ITD AML (Table 4). Five patients were not evaluable for response because they were removed from protocol therapy prior to disease assessment without meeting PD criteria (Supplementary Table S3). Of those evaluable for response, 2 achieved CR, 1 CRi, 1 CRp, 10 SD, and 3 had PD. Responses in the 7 evaluable FLT3-ITD AML patients included 1 CR, 1 CRp, 1 CRi, and 4 SD. Patients with FLT3-ITD AML demonstrated marked reduction in bone marrow blast counts after protocol therapy. This was not true of FLT3-WT or MLL-r ALL patients (Fig. 1). Of the 10 patients with SD, 4 were FLT3-ITD AML, 5 were FLT3-WT AML and 1 was MLL-r ALL. The relative percent change in bone marrow blast percentage was strikingly different in the 4 FLT3-ITD AML patients with SD (mean reduction 86%) compared with the 5 others with SD (mean increase of 5%, P = 0.002). Three FLT3-ITD patients (1 CR, 1 CRp, 1 SD) and 1 FLT3-WT (CR) received HSCT after protocol therapy. Because of small sample size, and in the context of the phase I study, we are not able to draw statistically significant conclusions regarding overall survival. Supplementary Table S4 illustrates a comparison of 1-year survival between evaluable FLT3-WT and FLT3-ITD AML patients.

PIA FLT3 inhibition was nearly complete at all patients at all dose levels (Fig. 2). In 19 patients with PIA assessment, 9 had 100% inhibition, 9 had 97%–99% inhibition, and 1 had 94% inhibition (Supplementary Table S5). As DL3 (60 mg/m²) had 1 of 9 patients with DLT and 9 of 9 patients with >90% inhibition by PIA, this dose level was defined as the recommended phase II dose.

FLT3 genotyping was confirmed in 20 of 22 patients for whom marrow at study entry was submitted (Supplementary Fig. S1). The FLT3 genotype of 2 patients was based on reports from outside clinical laboratories. The size of the FLT3-ITD insertions ranged from 18 to 105 base pairs, and the mutant:wild-type allelic ratio ranged from 0.62 to 2.82. One patient had no detectable wild-type FLT3, indicating loss of heterozygosity (LOH) for FLT3-ITD (Patient 10, Supplementary Table S5). No patients had detectable D835 point mutations. Quantitative expression of FLT3 transcript was measured using qPCR (Supplementary Table S5; Supplementary Fig. S2), demonstrating a high degree of interpatient variability in the AML patients, particularly in the FLT3 WT cohort.

Of the 22 eligible patients enrolled, sufficient material to perform 48-hour in vitro sensitivity (IVS) testing was available for 15 patients at study entry (7 FLT3-ITD AML, 5 FLT3-WT AML, and 3 ALL). To be considered sufficient, samples were required to contain at least 1 × 10⁶ viable leukemic blasts by Trypan blue exclusion and morphologic examination. As shown in Fig. 3A, a striking difference in IVS was demonstrable, with FLT3-ITD samples demonstrating a clear dose-dependent cytotoxic effect at the 0—20 nmol/L concentration levels at which quizartinib is primarily active against FLT3, and below the concentrations at which more nonspecific kinase inhibition or other off-target effects would be expected. The FLT3-WT AML samples showed a more modest cytotoxic effect and the MLL-r ALL samples showed no cytotoxic effect (Fig. 3; Supplementary Table S5; P < 0.005 comparing the 3 groups). Of the 15 patients with IVS data, 10 also had bone marrow blast counts available before and after therapy. As shown in Fig. 3 B, there was a statistically significant positive correlation (Spearman rank correlation coefficient r = 0.7, P = 0.04) between IVS and change in bone marrow blast percentage, suggesting that the IVS assay has the potential to serve as a biomarker of clinical response to quizartinib.

This clinical trial demonstrated that quizartinib in combination with intensive AML chemotherapy is safe and biologically active at a dose of 60 mg/m² given once daily. The unique design selected a RP2 dose (RP2D)-based on both safety and pharmacodynamic endpoints. There was no dose escalation past the 60 mg/m² dose based upon adult phase I/II data demonstrating complete responses at doses as low as 40 mg, potent inhibition of FLT3 phosphorylation at doses as low as 12 mg, and dose-dependent prolongation of QTcF (19). The current adult phase III study of quizartinib for FLT3/ITD AML starts all patients at 30 mg fixed dosing, escalating to 60 mg fixed dosing based on QTcF evaluation (www.clinicaltrials.gov identifier NCT02039726). Despite intensive ECG monitoring, our study only demonstrated only one transient grade 3 elevation of QTcF that was not dose limiting. Although there was a dose-dependent increase in QTcF from baseline (Supplementary Table S2), 60 mg/m² daily dosing did not restrict study drug administration in any patients. Our results support further study of quizartinib in children with FLT3/ITD mutations at 60 mg/m² once daily dosing. However, our PD testing demonstrates FLT3 inhibition at lower doses and it is possible that future pediatric clinical trials may mimic adult quizartinib development and study its efficacy at lower doses.

This trial was the first in adults or children to give quizartinib in combination with intensive chemotherapy. This strategy addressed the inherent difficulty in enrolling children with relapsed leukemia onto single-agent phase I studies while also providing the vital combination data required to incorporate quizartinib into de novo AML therapy. Importantly, we demonstrated that adding quizartinib to intensive chemotherapy does not result in detectable additive toxicity. Infection, myelosuppression, and febrile neutropenia were common and consistent with intensive AML regimens, but there were no hematologic DLTs and no indication that the combination increased the severity of these toxicities.

FLT3-ITD patients clearly demonstrated superior bone marrow responses compared with those with FLT3-WT AML and MLL-r ALL (Fig. 1). Of the 5 patients not evaluable for response, 3 were removed from protocol therapy due to persistence of leukemia detected in the peripheral blood. The other 2 patients were removed because of infection without having received 75% of prescribed quizartinib therapy. We recognize that those with detectable peripheral blasts clearly did not respond to quizartinib, but were not evaluable because they did not meet our protocol definition of progressive disease. Rather than retrospectively change our protocol definitions for evaluability, we detail the circumstances surrounding all patients not evaluable for toxicity and response in Supplementary Tables S1 and S3. Of interest, of those that were removed from protocol therapy for persistence of peripheral blasts on therapy, none harbored FLT3-ITD mutations. Not only did FLT3-ITD patients have a higher overall response rate (ORR; ORR = CR+CRp+CRi) compared with non-ITD patients (43% vs. 14%) but FLT3-ITD AML patients with SD clearly demonstrated improvement in blast count compared with FLT3-WT AML and ALL patients with SD. There was one CR among the FLT3 WT patients (patient 9 in Supplementary Table S5). Interestingly, this patient had the highest FLT3 expression by qPCR of the evaluable patients, and demonstrated a high degree of in vitro sensitivity to quizartinib. Unfortunately, there did not appear to be a response signal in the small number (n = 3) of evaluable MLL-r ALL patients. While this subset of patients was included in the trial due to expected overexpression of WT FLT3 and sensitivity to FLT3 TKI, these 3 individual patients actually expressed relatively low levels of FLT3 by qPCR (Supplementary Fig. S2), a finding associated with low level dependence on FLT3 signaling and modest FLT3 TKI-induced cytotoxicity in prior preclinical studies (12, 13). Importantly, testing of these 3 samples confirmed a lack of in vitro sensitivity to quizartinib (Fig. 3). Full assessment of efficacy of FLT3 TKI in MLL-r ALL and the potential of FLT3 expression level to serve as a biomarker of response will require larger studies.

Comprehensive evaluation of FLT3 inhibition as a promising therapeutic modality requires interrogation of an agent that is sufficiently potent and selective. The PIA assay has been validated as a surrogate measure of pharmacodynamic activity in response to targeted therapies (24). More specifically, the ability to achieve sustained phospho-FLT3 inhibition using the PIA assay has definitively been linked to response (26). To date, the clinical efficacy of FLT3 inhibitors has been limited by their off-target toxicities and their inability to consistently inhibit phospho-FLT3 (26). The PIA assay in this trial demonstrated near total phospho-FLT3 inhibition in every patient across all dose levels, showing that quizartinib is pharmacodynamically superior to any FLT3 inhibitor that has been published in pediatric trials to date. Given the near complete inhibition in all patients, correlation of PIA with response was not possible. In addition, the potency of quizartinib makes selection of a RP2D based on the PIA assay not feasible. Interestingly, our IVS assays of quizartinib-induced in vitro cytotoxicity strongly correlate with response in this clinical trial. Further study in a larger trial will be necessary to determine whether this biologic correlate may provide an additional tool by which to predict likelihood of clinical response to FLT3 inhibition.

This clinical trial demonstrates that quizartinib added to intensive chemotherapy is a safe and active therapy that provides continuous suppression of FLT3 phosphorylation in vivo. Our results provide strong support for further clinical testing of quizartinib in children with FLT3-ITD–mutated AML.

P. Gaynon reports receiving speakers bureau honoraria from JAZZ and Sigma Tau, and is a consultant/advisory board member for JAZZ. G. Gammon is an employee of Daiichi Sankyo. No potential conflicts of interest were disclosed by the other authors.

Conception and design: T.M. Cooper, R. Sposto, P. Gaynon, L. Gore, G. Gammon, P.A. Brown

Development of methodology: T.M. Cooper, P. Gaynon, P.A. Brown

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T.M. Cooper, J. Cassar, E. Eckroth, P. Gaynon, L. Gore, K. August, S.G. Dubois, L.B. Silverman, J. Oesterheld, D. Magoon, P.A. Brown

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T.M. Cooper, J. Cassar, J. Malvar, R. Sposto, P. Gaynon, B.H. Chang, J.A. Pollard, S.G. Dubois, P.A. Brown

Writing, review, and/or revision of the manuscript: T.M. Cooper, J. Malvar, R. Sposto, P. Gaynon, B.H. Chang, L. Gore, K. August, J.A. Pollard, S.G. Dubois, L.B. Silverman, J. Oesterheld, G. Gammon, P.A. Brown

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Cassar, E. Eckroth, R. Sposto, P.A. Brown

Study supervision: T.M. Cooper, P.A. Brown

Other (performed correlative laboratory studies): C. Annesley

This work was financially supported by grants from the Phase One Foundation to Children's Hospital of Los Angeles, TACL Consortium. Quizartinib was provided by Ambit Biosciences Corporation. The correlative laboratory assays were funded by grants (to P. Brown) from the Leukemia and Lymphoma Society (LLS Scholar in Clinical Research Grant #2365-12) and the American Cancer Society (ACS Research Scholar Grant #120237).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.