Leukemia Research 38 (2014) 756—763
`
`
`
`
`9;.
`
`Contents lists available at ScienceDirect
`
`Leukemia Research
`
`journal homepage: www.elsevier.com/locate/Ieukres
`ELSEVIER
`
`
`
`
`Lenalidomide and metronomic melphalan for CMML and higher risk
`MDS: A phase 2 clinical study with biomarkers of angiogenesis
`
`Rena Buckstein“, Robert Kerbelb, Matthew Cheunga, Yuval Shakedc, Lisa Chodirkera,
`Christina R. Leeb, Martha Lenisd, Cindy Davidsona, Mary-Anne Cussend, Marciano Reise,
`Alden Chesneye, Liying Zhangd, Alexandre Mamedovd, Richard A. Wellsi"f
`a Department ofMedical Oncology/Hematology, Odette Cancer and Sunnybrook Health Sciences Center, Toronto, Canada
`b Department ofMedical Biophysics and Platform Biological Sciences, Sunnybrook Research Institute, Toronto, Canada
`C Department ofMolecular Pharmacology, Rappaport Faculty ofMedicine Technion — Israel Institute of Technology, Haifa. Israel
`‘5 Department of Clinical Trials, Odette Cancer and Sunnybrook Health Sciences Center, Toronto, Canada
`5 Departments of Pathology and Laboratory Medicine, Sunnybrook Health Sciences Center, Toronto, Canada
`‘ Department ofBiological Sciences, Sunnybrook Research Institute, Toronto, Canada
`
`ARTICLE INFO
`
`
`ABSTRACT
`
`
`
`Article history:
`Raw“ 9Janu_3ry 2014
`Rece‘ve‘j ‘“ rev‘sed form 2 Math 2014
`Accepted 28 March 2014
`Available online 5 April 2014
`
`ig’gfigjesis
`MDS
`CMML
`Biomarker
`Circulating endothelial cells
`VEGF
`
`Metronomic. low dose chemotherapy may have anti-angiogenic effects and augment the effects of
`lenalidomide in MDS and CMML We evaluated the clinical efficacy, tolerability and anti—angiogenic
`effects of melphalan 2 mg and lenalidomide 10 mg for 21 days/28 in CMML (n = 12) and higher risk MDS
`(n=8) patients in a prospective phase ll study. The primary endpoint was overall response and sec-
`ondary endpoints included survival. progression-free survival, toxicity and biomarkers of angiogenesis.
`The median age was 73 years, 55% were pretreated and transfusion dependent. The overall response rate
`was 3(15%)of19 evaluable patients but 25% in CMML and 33% in pCMML. Dose reductions and/or delays
`were common due to myelosuppression. Transient spikes in circulating endothelial cells that declined
`below baseline were seen in responders and patients with CMML, suggesting anti-angiogenic activity.
`In conclusion, lenalidomide and metronomic low dose melphalan demonstrate signals of clinical and
`possible anti-angiogenic activity in patients with pCMML that require future validation. This trial was
`registered at clinicaltrialgov under # NCT00744536.
`© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
`license (http://creativecommons.org/licenses/by-nc—nd/3.0/).
`
`1. Background
`
`Myelodysplastic Syndromes (MDS) and chronic myelomono-
`cytic leukemia (CMML) comprise a heterogeneous group of stem
`cell disorders resulting in ineffective hematopoiesis and clonal
`instability. The clinical presentation is typically that of cytopenias
`and functional cell defects and the development of acute myel-
`ogenous leukemia (AML) in approximately a third of patients [1].
`Patients with CMML have monocytosis and may also have fea—
`tures of myeloproliferative neoplasms including splenomegaly and
`organ infiltration with monocytes. Patients with higher risk MDS
`as defined by the International Prognostic Scoring System (IPSS) [2]
`comprising 25—30% of cases, or higher risk CMML defined by the
`MD Anderson Scores [3.4] have shorter overall and leukemia—free
`
`* Corresponding author at: Odette Cancer Center, 2075 Bayview Avenue, Toronto,
`ON M4N3M5, Canada. Tel.: +1 416 480 5847; fax: +1 416 480 6002.
`E-mail address: rena.buckstein®sunnybrook.ca (R. Buckstein).
`
`survivals. Aberrantly increased angiogenesis in the bone marrow
`may be a factor contributing to disease progression. Bone mar-
`row microvessel density increases as the blast percentage increases
`[5] and this increase is associated with inferior survival. Vascu-
`lar endothelial growth factor (VEGF), the major positive regulator
`of angiogenesis,
`is present at increased concentrations in MDS
`blood and marrow [5,6]. Recently angiopoietin 1 (Aug-1) overex—
`pression was found to be associated with higher risk and higher
`blast percentage MDS and to correlate with the development of
`AML. Furthermore, high levels of Ang-1 were predictive of shorter
`overall survival independent of karyotype, IPSS and age [7]. It has
`been hypothesized that these pro-angiogenic cytokines may pro—
`mote leukemia cell propagation and survival via both paracrine and
`autocrine signaling [8]. Angiogenic factors may interact with other
`mechanisms, including differentiation and apoptosis in CMML and
`MDS [9,10]. By contributing to self- renewal of leukemia and MDS
`progenitors and elaboration of inflammatory cytokines such as
`tumor necrosis factor alpha (TNF-a), it is logical to speculate that
`therapeutic targeting of angiogenesis may improve hematopoiesis,
`
`http://dx.doi.org/l0.1016/j.leukres.2014.03.022
`0145—2126/© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http:/lcreativecommons.org/licenses/by-nc-nd/3.0/ ).
`Genentech 2090
`
`Hospira v. Genentech
`|PR2017-00737
`
`Genentech 2090
`Hospira v. Genentech
`IPR2017-00737
`
`

`

`K Buckstein et aI. / Leukemia Research 38 (2014) 756—763
`
`757
`
`delay the transition to leukemia and improve survival in these high-
`risk sub—groups.
`There have been a limited number of largely negative clinical
`trials of anti—angiogenic therapies in MDS/AML or CMML, possi-
`bly because the agents were tested as monotherapies in highly
`advanced disease. In a German. French and Belgian (GFM) clini-
`cal trial, bevacizumab in conventional dosing was tested in MDS
`patients with excess blasts. Although the drug had pharmacody-
`namic activity as reflected by reduced VEGF and bone marrow
`microvessel density. only 1/20 patients had a significant hemato-
`logic response [11]. Minimal clinical activity was seen in another
`phase 2 trial of SU5416, a first generation small molecule VEGF
`receptor inhibitor for patients with refractory myeloproliferative
`diseases although CMML comprised only 18% of the total popula-
`tion [12].
`Low—dose, metronomic chemotherapy (LDMC) is a strategy
`for optimizing the effects of chemotherapeutics by administering
`traditional cytotoxic drugs (e.g. melphalan. cyclophosphamide. vin-
`blastine, etc.) at lower doses without prolonged rest periods. LDMC
`is thought to inhibit tumor angiogenesis through multiple possi-
`ble mechanisms — by direct killing of activated endothelial cells in
`the tumor neovasculature [13,14]. inhibition of tumor cell HIP-10L
`expression and by inhibition of circulating endothelial precursor
`cells [15]. Other possible non-angiogenic mechanisms of activity
`under investigation include immunomodulation (with the deple—
`tion of T regulatory cells and the upregulation of cytotoxic T cells)
`and cancer stem cell targeting [16.17]. Studies in animal models of
`tumors [13.14.18] have demonstrated that maximal inhibition of
`angiogenesis and tumor regression occurs when LDMC is combined
`with anti—angiogeneic drugs targeting the VEGF pathway. Recent
`phase II and Ill clinical trial results provide further support for their
`treatment combination [19,20].
`Lenalidomide has clinical activity in non delSq MDS [21] and one
`mechanism of action may include its demonstrated anti-angiogenic
`activity [22] through the inhibition of VEGF and TNF-a induced
`endothelial cell migration and Akt phosphorylation [23].
`Continuous oral low dose melphalan has demonstrated activ-
`ity and safety in higher risk MDS and AML [24,25] particularly in
`hypocellular variants.
`This phase II study explored the efficacy and safety of combi—
`nation therapy with both lenalidomide and low dose continuous
`melphalan in patients with CMML or higher risk MDS and examined
`the effects on biomarkers of angiogenesis.
`2. Patients and methods
`
`2.1. Study design
`
`This was a single-arm, single center phase 2 prospective trial of combination
`therapy with lenalidomide and melphalan in patients with CMML 1—2 or higher
`risk (int-2 or high) MDS defined by the lPSS [2] or transfusion dependent MDS with
`blast % 25. Patients had to demonstrate adequate hepatic and renal functions and
`performance status (ECOG 52). In patients with proliferative CMML‘l or 2 (pCMML.
`WBC > 12 x 109/L), risk group was assigned using an MD Anderson CMML prognostic
`scoring system [3]. Patients could receive up to 12 cycles of lenalidomide 10 mg po
`combined with melphalan 2 mg po daily for 21 days out of 28. Two dose reductions
`of lenalidomide and melphalan were permitted (see online Appendix A).
`Patients remained on studydrugs until progression. death. unacceptable adverse
`event or 12 months. Patients were followed until progression. death or data lock
`(March 2013). All patients provided informed consent.
`
`22. Assessment ofefl‘icacy
`
`The primary endpoint was overall response rate as defined by modified inter—
`national working group standardized response criteria which included complete
`remission (CR). partial remission (PR). marrow CR (mCR) and hematologic improve-
`ment (HI) [26]. The secondary endpoints were cytogenetic remission rates. safety.
`OS. PFS, and biomarkers of angiogenesis including circulating endothelial cells
`(CECs) and precursors (CEPs). plasma VEGF and VEGFR 1—2 levels.
`Bone marrows for response assessments were performed days 1 of cycles 3. 5,
`8 and 11.
`
`2.3. Adverse event assessment
`
`Adverse events were identified and graded 1—5 using Common Toxicity Criteria
`v. 3.0. Hematologic toxicities were recorded throughout the study. and are reported
`as median percentage declines in absolute counts and increases in grade from base-
`line occurring during the first 12 weeks of therapy.
`
`2.4. Evaluation ofangiogenesis
`
`Soluble plasma VEGF.VEGFR1 and 2 levels were measured at baseline, day 1 of
`cycles 1—3 and twice thereafter at intervals of 3 months. and at study discontinu—
`ation. CECs and CEPs were measured at the same time points and enumerated as
`previously described [27.28] (see online Appendix A).
`
`2.5. Sample size and statistical analysis
`
`The sample size for this study was based on Simon's optimal two-stage design
`(alpha=0.05 and power=80%) with 9 patients enrolled in the first stage. The trial
`would be terminated if 0 responses were observed during the first stage. Otherwise.
`an additional 8 patients would be accrued in the second stage. The regimen would
`be active if 2 or more responses were observed from 17 patients accrued. A total
`of 20 patients were planned to accommodate for an anticipated 20% dropout rate.
`Data were presented descriptively using median. interquartiles (IQR), and ranges
`for continuous variables: using proportions for categorical findings. Overall Sur-
`vival (OS) and Progression-free Survival (PFS) were presented using Kaplan—Meier
`curves and log-rank test. To compare baseline biomarkers between groups of
`patients. non-nonparametric median test was conducted. To assess for correlations
`between selected biomarkers or blood counts. Spearman correlation coefficients
`were calculated with p-value<0.05 considered statistically significant. All results
`were conducted using Statistical Analysis Software (SAS version 9.3 for Windows)
`package.
`
`3. Results
`
`The study received local Institutional Review Board approval
`and accrued between January 2008 and March 2011. This trial
`was registered at http://www.clinicaltrials.gov/ as NCT00744536.
`Twenty patients with histologically confirmed CMML or higher
`risk MDS were enrolled at the Odette Cancer Center for this study
`(Table 1). A preponderance of patients enrolled in this study had
`CMML (n=12, 60%) and 9/12 had a total leukocyte count greater
`than 12 x 109/L. The median age was 73 years (range 52-87) and the
`median time from MDS diagnosis was 5.9 months (range 04—55).
`The median blast percentage was 6% (1—18) and 55% were transfu—
`sion dependent (TD) at enrollment. Thirty five percent had received
`a prior erythropoietic stimulating agent (ESA). 30% prior hydrox—
`yurea and 19% prior valproic acid. Of those classifiable by the IPSS,
`58% had Int-2 and 42% had high—risk disease. For the 12 CMML
`patients, the MD Anderson CMML prognostic scores were primar-
`ily intermediate—1 (42%) and Int-2 (50%) [3]. Sixty seven percent
`of the 9 pCMML patients (WBC> 12 x 105/1) had Int-2 risk scores.
`Patients received a median of 4 cycles (range 1-12) with a median
`cycle length of 30 days (range 20—63). The median follow up was
`8 months (range (1—46) and median time on study was 5 months
`(range 1—12).
`
`3.1. Response
`
`response assessment
`19/20 patients were evaluable for
`(Table 2). One patient achieved a marrow CR at 68 days, 1 hemato—
`logic improvement (Hl)—PI.T at 63 days and 1 l-lI-erythroid at 147d
`for a total response rate of 15%. All three responses were seen in
`patients with proliferative CMML—1 (median number of cycles =9,
`IQR 4—9). Nine patients (47%) had stable disease (median # cycles 6,
`IQR 4-9). Four patients (21%) were graded as failures due to wors—
`ening cytopenias without progression or death on study (median
`number of cycles = 2.5. IQR 1—4) and 3 patients (16%) had progres—
`sive disease (median number of cycles = 3, IQR 3—3 ). The 9 patients
`with pCMML had a median WBC of 46 x 109/L (range 19—312) at
`baseline that declined by 82% over the first 3 cycles. Similarly. the
`LDH, which was elevated at baseline (313 IU/L). declined by 37% to
`
`

`

`758
`
`R. Buckstein et al. / Leukemia Research 38 (2014) 756—763
`
`\lO
`
`1
`1
`
`".
`
`------
`
`\
`
`\\
`
`\
`
`\
`
`'n
`+ -
`.
`1.
`:
`i
`
`:r
`\i
`Rik
`:
`a
`:
`I
`
`.
`'1
`.
`I
`.
`l
`L
`l
`i
`[171‘
`..
`: \
`:
`l
`
`:I
`
`lI
`:
`i
`
`m2"
`
`Eg
`
`U
`c
`to
`a
`E
`.t
`.l: 50
`.—
`E
`u
`E 40
`.
`:7
`3‘
`
`3 30X
`
`7m
`g
`8 20
`s
`5
`'g
`m 10Um
`B
`
`60
`
`0
`
`.:.
`
`I
`E
`E
`5
`;
`fl
`:-- .............‘
`E
`"
`
`..
`
`400
`
`350
`
`m CO
`
`N U! CsauuenbmwHumumpew-(1/m)H01
`
`200
`
`150
`
`100
`
`U1 0
`
`
`
`Cycle 1
`n=12
`
`Cycle 2
`,"=10
`
`Cycle 3
`n=10
`
`Cycle 4
`n=9
`
`Cycle 5
`n=7
`
`Cycle 6
`n=6
`
`Cycle 7
`"=6
`
`Cycle 8
`n=6
`
`—0— Monocytes -£+- WBC
`
`Ill‘ll LDH
`
`Fig. 1. WBC: White blood cells x 109/L. LDH: Lactate dehydrogenase (lU/L). lQR: lnterqurtile range.
`
`a normal level (lQGIU/L) after 3 cycles (Fig. 1). One notable patient
`with pCMML—l who had initially presented with rapidly increasing
`leukocytosis, red cell and platelet transfusion dependence, symp-
`tomatic pleural effusions and peripheral edema had resolution of
`her effusions and edema by cycle 4 and became red cell transfu-
`sion independent by 147 days. She remained alive, effusion/edema
`free with a normal WBC and LDH at a follow up of 3.9 years from
`enrollment. She remained offall therapy, red cell and platelet trans-
`fusion independent despite only receiving 9 months of treatment.
`In contrast, another patient with pre—treated pCMML—l had a dra-
`matic decrease in his WBC (from 312 to 78 x 109/L) and LDH (from
`900 to 4501U/L) after 1 cycle of therapy but remained persistently
`symptomatic with unremitting ascites and peripheral edema that
`worsened markedly on the 1—week drug holiday. As a result. he was
`taken off study after 3 cycles, despite achieving a Hl-P.
`Only 1 patient with stable disease completed all 12 cycles and
`1 patient became red cell transfusion independent on study. The
`reasons for premature discontinuation included death (n =3), pro-
`gressive disease (n = 5) and toxicity (n =11).
`
`3.2. Survival
`
`In total, 5 progressed and three died on study. At follow—up,
`19 patients have died: 10 due to progressive disease and 9 from
`other causes (myocardial infarction (n=1), congestive heart fail—
`ure (n = 3), pneumonia (11 = 2). post hip surgery complication (n =1),
`liver failure (n=1) and unknown (n=1). The median overall and
`progression—free survivals were 8.5 months (95% Cl 5.9—12.7) and
`7.7 months (95% Cl 4.2—123) respectively and are depicted in Fig. 2.
`
`3.3. Assessment ofsafety
`
`Despite a high rate of baseline cytopenias (Table 2B) worsening
`hematologic toxicity was prevalent. Within the first 12 weeks of
`therapy, 20% grade 3 and 45% grade 4 neutropenia, 10% grade 3
`and 75% grade 4 thrombocytopenia were documented (Table 2C).
`While day 1 cell counts for subsequent cycles often improved com-
`pared with nadirs. 45% and 60% had grades 3-4 neutropenia and
`
`thrombocytopenia respectively at these time points. This repre-
`sented a worsening in grade neutropenia and thrombocytopenia
`(compared with baseline) for 40% and 55% of patients. As a result.
`dose delays and/or reductions were common. Of the 16 patients
`that had 2 or more cycles, 9/ 16 had a lenalidomide dose reduction
`by or during cycle 3:6/9 to 5mg daily and 3/9 to 5mg alternate
`days. in two additional patients dosage was reduced to 5 mg daily
`after cycle 4. Similarly, 8/ 16 patients had a melphalan dose reduc-
`tion to 2 mg alternate days by or during cycle 3. and by cycle 5, 4/8
`were close reduced, 2 to 2mg alternate days and 2 to 2 mg twice
`weekly. Four patients discontinued therapy within the first cycle
`for toxicity or disease progression and 7 patients by cycle 3. Three
`patients died on study, 1 from transfusion associated circulatory
`overload (TACO) post red cell transfusion in cycle 1, 1 from pneu-
`monia in cycle 1 and 1 from progressive disease and blast crisis
`post cycle 9. No death was felt to be directly treatment related.
`The grade 3—4 non—hematologic toxicities (and attributions)
`are summarized in Table 2A. The most common non-hematologic
`toxicities were infection (pneumonia (n=2). febrile neutropenia
`(n = 3). cellulitis (n = 1)), neurologic (syncope n = 2, ischemic event
`n = l ). and cardiac (atrial fibrillation (n =1), congestive heart failure
`(n =1) and TACO (n= 1, grade 5). There was 1 episode of grade 3
`epistaxis.
`
`3.4. Biomarkers
`
`Angiogenesis related biomarkers were available in 19 patients
`at baseline, 16 patients at cycle 2 day l, 15 patients at cycle 3
`day 1 with reducing numbers thereafter due to patient attrition
`(Fig. 3). CEPs were too infrequent to measure for most patients and
`are not reported. End of study (EOS) readings were available in 5
`patients with either response or stable disease at a median time of
`37 days (range 28—50 days) post last dose. When plotted over time
`for all patients. there were no discernable declines in serum VEGF
`or circulating VEGFR] or VEGFR2 although an apparent transient
`spike in CECs was noted during the first cycle that then reverted
`to baseline or declined by cycle 6 (data not shown). Sample sizes
`are generally too small to detect statistically significant differences
`
`

`

`R. Bucksteiri et al. / Leukemia Research 38 (2014) 756—763
`
`759
`
`
`
`
`
`Survivaldistributionfunction
`
`Duration of Overall Survival (months)
`
`
`
`
`
`5 4
`
`no. at risk
`8
`14
`0.0 — 20
`0
`1
`1
`2
`l—|—__T~*~'—l_l—l_*‘_’fif'—l_—I
`0
`6
`12
`18
`24
`30
`36
`42
`48
`
`
`
`
`
`Survivaldistributionfunction
`
`Duration of Progression-free Survival (months)
`
`Fig. 2. (A) Overall survival. (B) Progression-free survival.
`
`in the changes between responders (n=3), stable disease (n=9)
`and patients with progressive disease and failure (n=6) but it is
`instructive to observe any trends according to response (Fig. 3a—c).
`Responding patients had higher levels of baseline CECs compared
`with non responders (10.35 cells/u]. IQR 635-187 versus 4.48
`cells/u], IQR 2.24-13.8, p=.06). Responders also appeared to have
`a 19 fold spike in total (primarily viable) CECs after the first cycle
`of therapy (from 10.35 cells/u]. IQR 63-1872 to 193.7 cells/u], IQR
`189—2684) before reverting back to baseline and declining in sub-
`sequent cycles. This compares with a 6 fold spike in total CECs after
`the first cycle in patients with progressive disease or failure (from
`3.8 cells/u], IQR 3.4—4.4 to 23.9 cells/u], [QR 1—46). Similarly, in
`CMML we also documented transient spikes in total CECs after cycle
`1 (from 6.67 cells/u], IQR 3.7—16.2 to 24 cells/pl, IQR 5.2 to 50) that
`gradually declined by 86% at cycle 6, only to increase again toward
`the end of study drugs (Fig. 3d-e).
`The median baseline VEGF levels were higher for CMML than
`for non-CMML patients (43.45 versus 26.5 pg/ml) but they were
`not statistically different (p = 0.22). Baseline CECs were significantly
`higher in the CMML patients compared with the other WHO MDS
`subtypes (6.67 versus 3.45 cells/u]. p= .03), (Table 3). In the CMML
`patients, VEGF, sVEGFR1. sVEGFRZ and CEC were not significantly
`correlated with the monocyte count, total WBC or LDH.
`CECs were weakly correlated with sVEGFR1 at baseline (r= 0.45,
`p = .05) and with serum VEGF on day 1 cycle 2 (r=O.68, p = .02).
`
`4. Discussion
`
`Metronomic chemotherapy is under active investigation in the
`clinic, especially for the treatment of solid tumors. As recently
`reviewed, there are more than 80 published phase 2—3 studies
`encompassing 3688 patients with a variety of cancers that have
`demonstrated the clinical benefits of LDMC. The most commonly
`used drugs are cyclophosphamide. and oral 5—FU (5 fluorouracil)
`pro-drugs capecitabine or UFT (tegafur—uracil). LDMC is frequently
`combined with other therapies (64%) [29]. Anti—angiogenic ther—
`apy in general and LDMC specifically have been infrequently
`evaluated in blood cancers but show promise in the lymphoid
`neoplasms. particularly with the drug cyclophosphamide [30—32].
`The LDMC/anti-angiogenic field was recently bolstered by the pos—
`itive results of the CAIRO3 — a clinical trial of the Dutch Colorectal
`Cancer Group. CAIRO3 demonstrated a significant PFS benefit to
`maintenance treatment with bevacizumab+oral daily low dose
`capecitabine in patients with metastatic colorectal carcinoma [20].
`Before azacitidine was available there were limited therapeutic
`options for patients with CMML and higher risk MDS. Building
`on our local experience with metronomic cyclophosphamide and
`celecoxib in relapsed or refractory aggressive lymphomas [31]
`we chose to evaluate the efficacy, safety and anti-angiogenic
`activity of metronomic melphalan and lenalidomide in CMML and
`MDS patients. During the course of the study, azacitidine became
`
`

`

`760
`
`R. Buclcstein et aI. / Leukemia Research 38 (2014) 756—763
`
`525
`
`125
`
`i 25
`
`.E5Eae
`E
`'i
`E:5
`u
`S
`s
`g
`g
`E 0.2
`
`5
`
`,
`
`.
`.
`
`
`
`.
`.
`.
`I
`I
`.
`.
`'I
`
`Cycle 1
`“=3
`"=2
`
`Cycle 6
`Cycle 4
`Cycle 3
`n=5
`"=3
`n=8
`n=2
`n=5
`—o—cMML --A-- Non-CMML
`
`Cycle 9
`"=4
`
`E05
`n=3
`
`a.“
`Cycl: 2
`CMML
`"=12
`Nan-CMML n=7
`
`615
`
`Nu-
`
`
`
`
`
`
`
` 25 vzcr(pg/lull-Medianwllhlnlerquanlles
`
`
`
`Cy le 1
`n=s
`n=6
`"=3
`
`Cycle 6
`Cycle 3
`Cycle 2
`n=5
`n-7
`n=6
`"=3
`n=2
`n=z
`n-a
`n:2
`--- PDand Failure +51: "-0 Response
`
`Cycle 4
`n=2
`
`Cycle 9
`n=2
`n=2
`
`EOS
`n=3
`
`T
`I
`...l
`I
`;
`: ,
`*-~e
`; -
`l
`
`
`
`g 1251:..=
`
`E..
`Ez 25
`gi
`5iu
`s
`
`l.
`
`EEE"
`
`3‘
`
`I
`
`0.2
`
`5D
`Pmrallure
`Response
`
`625
`
`EE
`
`EE3upu:...>
`
`er
`= 125
`5
`‘E
`_
`
`é;:
`-_°
`-e
`"
`E,
`
`25
`
`5
`SD
`PD+Failure
`Response
`
`Cycle 1
`n=9
`"=5
`n=3
`
`Cycle 6
`Cycle 3
`Cycle 2
`n:5
`n=a
`n=9
`n=z
`"=4
`n=Z
`n=3
`"=3
`-e§- PDand Failure +50 we Response
`
`Cycle 4
`n=2
`
`Cycle 9
`n=3
`
`E05
`n=3
`
`10000
`
`nacl0
`
`s:aao maoa
`sVEGFFlZ[pg/ml)vMedlallwilhlnlemvanlles3ea
`
`Cycle 2
`n=10
`n=6
`
`Cycle 6
`Cycle 4
`Cycle 3
`"=5
`n=2
`n=l
`"=2
`"=5
`+CMML ‘f-Non-CMML
`
`Cycle 9
`n=4
`
`E05
`n=1
`n=1
`
`5
`Cycle 1
`CMML
`n=12
`Nan-CMML "=7
`
`menu
`
`
`
`:3» sawE'=‘E
`é’ soon
`§3e
`=3 7am:
`
`u5
`
`sona
`
`ES
`a scan
`
`.iE
`
`«on
`Cycle 1
`CMML
`mi:
`Non-CMML “-7
`
`cycle 2
`"-10
`«=6
`
`Cycle 6
`Cycle a
`Cycle 3
`n-5
`n-2
`net
`“-2
`n:5
`+CMML —-n-- Non-CMML
`
`Cycle 9
`n-4
`
`-
`
`-
`:05
`n-2
`"-2
`
`Ie 1
`n 9
`n s
`n 3
`
`lo 6
`Ie 3
`le 2
`n 5
`n a
`n 9
`n 2
`n a
`n 2
`n 3
`n 3
`--n--Poand Fallun: +SD "-0-- Response
`
`I14
`n 2
`
`le
`n 3
`
`E
`n a
`
`
`
`
`
`
`
`4000
`
`Falluve
`Resp use
`
`Fig. 3. (A) CEC according to response. (B) VEGF according to response. (C) VEGFRZ according to response. (D) CEC according to CMML or not. (E) VEGF according to CMML
`or not. (F) VEGFRZ according to CMML or not. CEC: Circulating endothelial cells. VEGF: Vascular endothelial cell growth factor. VEGFRZ: Vascular endothelial growth factor
`receptor-2. SD: Stable disease. PD: Progressive disease.
`
`available for int—2 and high risk MDS and explains the predomi-
`nance of CMML patients on this study.
`To our knowledge, this is the first clinical trial to evaluate LDMC
`as possible anti—angiogenic therapy in CMML and MDS. We chose
`the immunomodulatory drug lenalidomide because it is modestly
`active as monotherapy in non-delSq MDS [21] and in higher risk
`del5q MDS [33] even after azacitidine failure [34]. One mechanism
`of action in non-del5q MDS may include the inhibition of angiogen-
`esis [22,23]. We chose to partner lenalidomide with low continuous
`dose melphalan because of its demonstrated activity and safety in
`higher risk MDS and AML [24.25]. We hypothesized that when used
`in such a metronomic schedule this drug may be functioning in part
`via anti—angiogenic mechanisms.
`In this study. we found a modest overall response rate of
`15%, and a response rate of 25% in CMML overall (3/12) and 33%
`in proliferative CMML (3/9). Our patients experienced significant
`myelosuppression at our starting doses of melphalan 2 mg 0d and
`
`lenalidomide 10 mg ed for 21d/28 schedule. This necessitated drug
`reductions and dose delays in the majority of the patients within
`2 cycles and 5 patients came off study early for thrombocytopenia.
`While four patients discontinued treatment within 1 treatment
`cycle due to death, disease progression or severe pancytopenia, it is
`often difficult to ascribe worsening blood counts to drug toxicity or
`natural history in advanced pre—treated MDS and CMML. Our expe—
`rience is consistent with that observed in higher risk del5q MDS
`where lenalidomide monotherapy was also associated with sig-
`nificant myelosuppression and adverse events necessitating dose
`reductions to 5 mg/day in 72% and discontinuation after only 1 cycle
`in 31% [33].
`We chose to include CMML in our patient population because
`there were limited therapeutic options at the time. CMML is now
`a distinct entity defined in the WHO as an MDS/MPN, not strictly
`MDS. It is classified as CMML-1 or 2 according to blast percentage
`in the bone marrow and peripheral blood and has its own clinical
`
`

`

`R. Buckstein et al. / Leukemia Research 38 (2014) 756—763
`
`761
`
`Table 1
`Baseline characteristics n = 20.
`
`Baseline characteristics
`Age at consent (year)
`Median (range)
`Gender
`Male/female
`WHO subtype
`RCMD-RS
`RAEB-1
`RAEB—Z
`CMML-1
`CMML-2
`IPSS risk group
`(n = 1 1
`evaluable,
`includes 3
`CMML with
`low WBC)
`Int-2 MDS
`High risk MDS
`IPSS — MD Anderson CMML (n =12)
`Low
`Int-1
`Int-2
`Time from diagnosis (months)
`Median (range)
`Hgb
`Median (range)
`
`ANC (absolute neutrophils count)
`Median (range)
`Platelets
`Median (range)
`Transfusion dependence at baseline
`TI
`TD
`Previous therapy (:1 = 12)
`ESA
`Lenalidomide
`Immunosuppressive therapy
`Hydroxyurea
`Valproic acid
`Bone marrow blast (96)
`Median (range)
`Karyotype
`Good risk
`Intermediate risk
`Poor risk
`
`N (96)
`
`13/7
`
`l
`2
`5
`10
`2
`
`9
`2
`
`1
`5
`.5
`
`9
`1 l
`
`7
`0
`1
`6
`2
`
`12
`2
`4
`
`73 (52—87)
`
`(6596/3576)
`
`(5.00%)
`(10.00%)
`(25.00%)
`(50.00%)
`(10.00%)
`
`(82.00%)
`(18.00%)
`
`(8.33%)
`(41.67%)
`(50.00%)
`
`5.9 (GA—55.1)
`
`93.5
`(68.0—122.0)
`
`5.2 (03—1247)
`
`56 (7—621)
`
`(45.00%)
`(55.00%)
`
`(58.33%)
`(0%)
`(8.33%)
`(50.00%)
`(16.67%)
`
`6 (1—18)
`
`(66.6%)
`(11.11%)
`(22.2%)
`
`WHO: World health organization. IPSS: International prognostic scoring system,
`Hgb: Hemoglobin,le Transfusion independent, TD: Transfusion dependent.
`
`prognostic scoring systems, with most systems reporting infe—
`rior survival for proliferative forms defined by a WBC> 12 x 109/L
`[3,35,36,37]. There are a limited number of published case reports
`or clinical trials discussing the activity of lenalidomide in CMML
`as monotherapy [38] or in combination with bortezomib [39].
`While the hypomethylating agents azacitidine and decitabine are
`approved for CMML based on activity observed in phase 2 and 3
`trials, CMML (particularly pCMML) is often underrepresented in
`such trials [40—42]. Our study was conceived before the availabil—
`ity of HMA therapy for CMML in Ontario. which has subsequently
`become available but restricts reimbursement to patients with
`WBC<13 x 109/L, based on AZA-001 inclusion criteria [40]. Sim—
`ilarily, the EMA (European Medical Agency) has restricted the
`approval of azacitidine for the treatment of CMML-2 in the absence
`of myeloproliferative aspects, which restricts access to only 10—20%
`of patients. This leaves hydroxyurea as the primary option for
`patients with proliferative CMML which is not always long-lasting
`nor effective. In addition, the OS after azacitidine failure is short [43]
`therefore alternative therapies are needed in these circumstances.
`It is interesting to note that the 3 responding patients all had
`pCMML (out of 9 total, ORR 33%) and remained on drug for a median
`
`Table 2
`Adverse events.
`
`2A
`Non-hematologic toxicities
`
`Grades 3—4 on
`study
`
`n (96)
`
`Infection
`Neurological
`Cardiac (CHF, arrhythmia)
`Endocrine
`Hemorrhage/bleeding
`Pain
`Constitutional symptoms
`Gastrointestinal
`Musculoskeletal/soft/tissue
`Pulmonary
`
`6 (30%)
`3 (15%)
`3 (15%)
`1 (5%)
`1 (5%)
`1 (5%)
`1 (5%)
`1 (5%)
`1 (5%)
`1 (5%)
`
`Possible/probany/definitely
`related to protocol therapy 11
`(76)
`6 (30%)
`3 (15%)
`0 (0%)
`1 (5%)
`1 (5%)
`1 (5%)
`0 (0%)
`0 (0%)
`0 (0%)
`O (0%)
`
`23
`
`Anemia
`Leukocytes
`Neutrophils
`Platelets
`
`2C
`
`Hematologic toxicity grades® baseline (n = 20)
`1
`2
`3
`
`4
`
`7 (35%)
`1 (5%)
`2 (10%)
`6 (30%)
`
`9 (45%)
`2 (10%)
`1 (5%)
`3 (15%)
`
`3 (15%)
`3 (15%)
`2 (10%)
`6 (30%)
`
`0 (0%)
`0 (0%)
`3 (15%)
`2 (10%)
`
`Hematologic toxicity grades on study (n = 20) (first 12 weeks)
`I
`2
`3
`4
`
`3 (15%)
`8 (40%)
`5 (25%)
`4 (20%)
`Anemia
`4 (20%)
`5 (25%)
`3 (15%)
`3 (15%)
`Leukocytes
`9 (45%)
`4 (20%)
`3 (15%)
`0 (0%)
`Neutrophils
`15 (75%)
`2 (10%)
`2 (10%)
`1 (5%)
`Platelets
`Grade 3-4 non hematologic toxicities on study and relatedness to therapy. 23: All
`grade hematologic toxicities at baseline. 2C: All grade hematologic toxicities on
`study during first 12 weeks.
`
`of 9 cycles suggesting a possible signal of activity and tolerability
`unique to this histology. The anti—proliferative effects of lenalido—
`mide and melphalan in CMML are also evident in the declines seen
`in LDH, WBC and monocyte count (Fig. 1 ). The higher baseline levels
`ofVEGF and CEC's in CMML patients compared with other MDS sub—
`types may point to higher levels of angiogenesis driving this disease
`as previously observed [6.44.45] and a potential 'druggable‘ target
`for appropriately selected patients.
`Mature circulating endothelial cells (CECs) are in most cases
`apoptotic in healthy subjects and more viable in cancer patients,
`and represent an indirect marker of vessel damage and/0r turnover
`and remodeling associated with angiogenesis [46]. CECs are signif—
`icantly increased in different types of cancer [47] but because they
`are rare events, they need to be enumerated in dedicated labo—
`ratory with experienced personnel that use rigorous. robust and
`validated multiparamettic procedures [48]. Changes in CECs and or
`CEP levels during therapy have been found to correlate with clinical
`outcomes of response and/or survival in 5 studies, but baseline lev—
`els have been inconsistently predictive of response in clinical trials
`of anti-angiogenic agents [49]. In one study, a CEC count of > 11uL
`after 2 months was predictive of improved overall and disease-free
`survival and the increase was attributed to an increased fraction of
`apoptotic CEC’s [27]. In that study, a significant decline in VEGF—A
`levels after 2 months of therapy wasassociated with a significantly
`improved time to progression. In the current study, the increase
`in CEC values observed prior to cycle 2 in responding patients and
`those with CMML was primarily of viable cells. This is consistent
`with results from a preclinical model [50], in which a maximally
`tolerated dose (MTD) regimen caused short-term suppression of
`viable CECs and CEPs immediately after drug administration, fol-
`lowed by a robust rebound effect leading to an increased number
`of viable cells prior to the next cycle. The declines we observed
`
`

`

`762
`
`R. Buckstein et aI. / Leukemia Research 38 (2014) 756—763
`
`Table 3
`Biomarkers at baseline (cycle 1 day 1) according to response or histology.
`
`Response
`PD + Failure
`SD
`CMML
`MDS (non-CMML)
`
`N
`3
`6
`9
`12
`7
`
`CEC (IQR)
`1035 (63—187)
`3.7 (3.4—4.5)
`5.3 (2.2—13.8)
`6.7 (3.7—16)
`3.4 (02—455)
`
`VEGFUQR)
`47.7 (30.1—51.4)
`17.96 (15.1—26.5)
`39.9 (21.2—80.7)
`43.4 (25.9—71.5)
`25.5 (15.4—37.8)
`
`VECFRl (IQR)
`1250 (1067—2575)
`100.2 (78—1057)
`104.6 (93.6—119.2)
`106.9 (1002—127)
`95.1 (78—1057)
`
`VEGFRZ (IQR)
`7534(6418.4—7640.9)
`7557.9 (7040—78958)
`7534(6418.4—7640.9)
`7418.9 (57351—79163)
`7375.2 (70407—82447)
`
`PD: Progressive disease, SD: Stable disease, CMML: Chronic myelomonocytic leukemia, MDS: Myelodysplastic syndrome, CEC: Circulating endothelial cells/u]. IQR: Interquar-
`tile range. VEGF: Vascular endothelial growth factor (pg/ml). VEGFR-l: Vascular endothelial growth factor receptor 1(pg/ml), VEGFR-Z: Vascular endothelial growth factor
`receptor 2(pg/ml).
`
`in subsequent cycles in responding patients and those with CMML
`(who remained