The Hsp90 inhibitor SNX-7081 synergizes with and ...

Leukemia & Lymphoma, July 2012; 53(7): 1367–1375 © 2012 Informa UK, Ltd. ISSN: 1042-8194 print / 1029-2403 online DOI: 10.3109/10428194.2011.647310

ORIGINAL ARTICLE: RESEARCH

The Hsp90 inhibitor SNX-7081 synergizes with and restores sensitivity to fludarabine in chronic lymphocytic leukemia cells with lesions in the TP53 pathway: a potential treatment strategy for fludarabine refractory disease O. Giles Best1,5, Yiping Che4, Nisha Singh2, Cecily Forsyth3,5, Richard I. Christopherson4,5 & Stephen P. Mulligan1,4,5 Leuk Lymphoma Downloaded from informahealthcare.com by University of Sydney on 03/09/14 For personal use only.

1Northern Blood Research Centre, Kolling Institute and 2Cytogenetics, PALMS, Royal North Shore Hospital, Sydney, Australia, 3Jarrett Street Specialist Centre, North Gosford, Australia, 4School of Molecular and Microbial Biosciences, University of Sydney,

Sydney, Australia, and 5CLL Australian Research Consortium (CLLARC) regimens rely on a functional TP53 signaling pathway and, hence, the association between TP53 deletions and mutations and drug resistance in many cancers, including CLL [4]. The ineffectiveness of the FCR regimen against CLL cells with TP53 lesions highlights the substantial need for novel therapeutic strategies for this subset of patients. Recent evidence suggests that the heat shock protein-90 (Hsp90) may represent a novel therapeutic target in many cancers, including CLL [5,6]. Many of the proteins reliant on Hsp90 for correct functioning have critical roles in the survival and proliferation of tumor cells [7], including the Akt and nuclear factor-κB (NF-κB) proteins in CLL [8]. Compelling in vitro evidence has led to more than 40 clinical trials of Hsp90 inhibitors against a range of cancers [9]. However, trials of the geldanamycin-derivative 17-AAG in advanced cancers have shown limited clinical activity, poor solubility and dosedependent toxicities [10,11]. Consequently novel inhibitors have been manufactured, including a family of compounds developed by Serenex (now Pfizer, Inc.). Several groups [12–14] have demonstrated that the Serenex compounds are effective against various solid and hematological malignancies, with activity superior to the geldanamycin-derived drugs. It is also likely that Hsp90 inhibitors may be more efficacious and their toxicity minimized by combination with more conventional agents. Given the array of Hsp90 clients, it has been suggested that Hsp90 inhibitors may synergize with any drug that induces stress in cancer cells. Several studies support this, demonstrating synergy between geldanamycinderived compounds and established anti-cancer drugs or radiation against a range of malignancies [15–19], including CLL. To date, however, there is no evidence of a similar synergy between a synthetic Hsp90 inhibitor and a more conventional agent against CLL cells. Here we demonstrate significant synergy between one of the compounds developed by Serenex, SNX-7081, and

Abstract Drug resistance in chronic lymphocytic leukemia (CLL) associated with lesions in the ATM/TP53 pathway represents a major challenge in clinical management. Evidence suggests that heat shock protein-90 (Hsp90) inhibitors may represent a therapeutic option in combination with more conventional therapies. We explored the effects of combining the Hsp90 inhibitor, SNX-7081, with fludarabine in vitro against CLL cells and hematological cell lines. In seven cell lines and 23 patient samples synergy between SNX-7081 and fludarabine (2-FaraA) was apparent in the three TP53 mutated cell lines and at significantly lower concentrations in TP53 or ATM dysfunctional patient cells. In 11/13 2-FaraAresistant patient samples, SNX-7081 reduced the 50% inhibitory concentration to within a clinically achievable range. Synergy between SNX-7081 and 2-FaraA was evident in both the cell lines and patient samples as a significant decrease in cell viability. Our data suggest that combining SNX-7081 and fludarabine may be effective in the treatment of fludarabine-refractory CLL. Keywords: CLL, TP53, fludarabine, Hsp90 inhibitor, SNX-7081

Introduction The clinical course of chronic lymphocytic leukemia (CLL) is heterogeneous. Several biological factors identify patients at risk of disease progression and poor treatment response rates. In particular, mutations or deletions of the TP53 and ATM genes identify a patient group with extremely poor prognosis, highlighted by recent trials of the fludarabine, cyclophosphamide, rituximab (FCR) regimen [1,2]. The TP53 and ataxia telangiectasia mutated (ATM) proteins are critical in preventing tumor development and maintaining genomic integrity through inhibition of cell replication following DNA damage [3]. Many chemotherapeutic

Correspondence: Dr. Giles Best, Northern Blood Research Centre, Level 11, Kolling Institute, Royal North Shore Hospital, St Leonards, Sydney, NSW 2065, Australia. Tel: 61(02)99264860. Fax: 61(02)99265716. E-mail: [email protected] Received 4 August 2011; revised 4 October 2011; accepted 4 December 2011

1367

1368

O. G. Best et al.

fludarabine (2-FaraA) against primary CLL samples and hematological cell lines harboring mutated ATM or TP53. Our data suggest that the combination of SNX-7081 with fludarabine may have significant therapeutic benefit for patients who are resistant to fludarabine-based regimens by rendering the cells fludarabine-sensitive. Clinical trials of this combination may be warranted, particularly for the treatment of aggressive CLL.

Methods

FISH analysis

Cell lines and patient samples

Leuk Lymphoma Downloaded from informahealthcare.com by University of Sydney on 03/09/14 For personal use only.

subdivided by exposure to etoposide and nutlin-3a. A previous study demonstrated a strong correlation between samples harboring mutations of TP53 or ATM and categorization as CAT1 or CAT2 dysfunctional, respectively [22]. ZAP-70 and CD38 expression was assessed by flow cytometry as described previously [23]. Positive cases were defined as containing more than 10% ZAP-70 or 7% CD38 positive cells. The ZAP-70 antibody was produced by Millipore (clone 2F3.2). All other antibodies were obtained from Becton Dickinson.

Details of the patient samples and cell lines used are shown in Tables I and II, respectively. Patients were diagnosed according to criteria defined by the International Working Group on CLL [20]. Peripheral blood samples were obtained with ethical approval and informed consent. Mononuclear cells (PBMCs) were harvested by centrifugation of blood through a Ficoll density gradient, and stored in liquid nitrogen. All patient samples contained 85% CD5 /CD19 cells as determined by two-color flow cytometry, and were 80% viable by trypan blue exclusion after thawing. The MEC1 and MEC2 cell lines were derived from a patient with CLL in prolymphocytoid transformation [21]. While there is debate regarding how closely these cell lines reflect the biology of CLL, the cell lines are B-cell derived. Importantly, in the context of the present study, they are 2-FaraA resistant [5], likely resulting from mutation of TP53. All cell lines and patient cells were maintained in RPMI-1640 with 10% (v/v) fetal calf serum (FCS), 2 mM l-glutamine and 1% penicillin/streptomycin.

Analysis of ATM/TP53 function, ZAP-70 and CD38 All CLL patient samples were screened for functional defects in the ATM/TP53 pathway as described elsewhere [22]. Briefly, samples were exposed to 50 μM etoposide or 50 μM etoposide plus 5 μM nutlin-3a for 24 h. Expression of TP53 and p21 was assessed by flow cytometry. Cells were defined as functionally normal if TP53 and p21 expression increased in response to etoposide. Cells that displayed little or no up-regulation were deemed dysfunctional and were further

Interphase fluorescence in situ hybridization (FISH) analysis was performed using standard techniques as previously described [5] using the Cytocell CLL panel (Cambridge, UK).

Assessment of cytotoxicity Drug-induced cell death was determined by MTT [3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay and flow cytometry. Viable cells reduce MTT to form blue formazan crystals. Cells from each cell line (1 105) or patient with CLL (5 105) were treated with drug(s) for the times and at the doses indicated in 96-well plates and a final volume of 200 μL. Four hours prior to the end of treatment, 20 μL of 5 mg/mL MTT was added to each well and incubated for a further 3 h at 37°C. The reaction was stopped and the crystals solubilized by the addition of 100 μL dimethylsulfoxide (DMSO). Absorbance readings were taken from each well at 540 nm. Fifty percent inhibitory concentration (IC50) values were calculated from dose–response analyses following 48 h treatment using BioDataFit software (http://www.changbioscience.com/stat/ec50.html). As loss of mitochondrial membrane potential is an early indication of apoptotic cell death [24], we used the mitochondrial membrane potential dye DilC1(5) (Invitrogen) and propidium iodide (PI; Sigma) to measure the percentage of apoptotic cells present. Cells were treated for the times and at the doses of SNX-7081 and 2-FaraA indicated, harvested, washed and stained with DilC1(5) for 30 min. For the last 5 min of incubation, 10 μL of 0.01 mg/mL PI was added, with immediate analysis by flow cytometry. Cells negative for DilC1(5) and PI were considered apoptotic.

Table I. Details of the seven cell lines and their sensitivity to 2-FaraA in the presence and absence of SNX-7081*.

Cell line Mutated MEC1 MEC2 U266 Wild-type Raji MOLT4 IM9 Null HL60

Origin

2-FaraA IC50 (μM) (single agent)

2-FaraA IC50 (μM) ( SNX-7081)

DRI (2-FaraA) at Fe 0.5

DRI (SNX-7081) at Fe 0.5

2-FaraA SNX-7081 CI at Fe 0.5

PLL PLL ALL-B

161 30.9 188 56.67 98.2 9.76

5.22 0.42† 5.17 0.33† 4.79 0.29†

30.84 36.36 20.50

4.57 3.12 2.13

0.25 0.03 0.36 0.05 0.32 0.25

Burkitt-B ALL-T MM

2.67 0.20 3.38 0.18 11.8 1.76

2.09 0.12 2.80 0.36 2.91 1.54‡

1.28 1.21 3.95

12.64 1.71 2.00

0.86 0.02 1.35 0.12 1.10 0.15

APML

2.76 0.24

2.48 0.24

1.11

2.70

1.30 0.10

*Sensitivity to 2-FaraA as a single agent and in combination with SNX-7081 was assessed in the seven cell lines using the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide] assay. Fifty percent inhibitory (IC50) values were calculated using BioDataFit software and are shown as the mean ( SD) of three independent experiments. Combination indices (CIs) shown were calculated using the median effect principle of Chou and Talalay [25] at a fractional effect (Fe) of 0.5. Combination indices 1 are indicative of synergy. A dose reduction index (DRI) of 1 indicates a reduction in the IC50 for the respective drug. †Indicates a 95% and ‡indicates a 75% reduction in the IC of 2-FaraA. 50 PLL, prolymphocytic leukemia; ALL-B, acute lymphoblastic leukemia-B lineage; Burkitt-B, Burkitt lymphoma-B lineage; ALL-T, acute lymphoblastic leukemia-T lineage; MM, multiple myeloma; APML, acute promyelocytic leukemia.

Hsp90 inhibition and fludarabine sensitivity 1369 Table II. Details of CLL patient samples.*


Patient Normal 1 2 3 4 5 6 7 8 9 10 CAT1 11 12 13 14 15 16 17 CAT2 18 19 20 21 22 23

ZAP-70 (%) CD38 (%) Treatment

2-FaraA IC50 (μM) (single agent)

2-FaraA IC50 (μM) ( SNX7081)

CI at Fe 0.5

DRI (2-FaraA) at Fe 0.5

DRI (SNX7081) at Fe 0.5

11q (%)

17p (%)

/ / / ND (80)

/ / ND (12)

4.34 8.7 3.51 6.88 14.0 2.87 14.7 26.31 4.62 19.1

1.35 3.35 0.19 68.7 16.0 1.22 7.69 84.2 85.81 58.1

FCR FCR NT NT NT NT FCR NT Multiple FCR

5.43 3.62 3.37 0.97 1.12 7.11 3.62 3.35 2.14 3.45

2.08‡ 4.30 2.42 2.50 2.42 2.63‡ 0.98‡ 1.47‡ 1.39 3.10

0.53 1.29 0.87 2.64 0.22 0.37 4.64 0.48 0.67 0.93

2.61 0.84 1.39 0.39 0.46 2.70 3.69 2.28 1.54 1.11

6.90 10.00 8.00 14.80 265.00 248.85 561.22 26.00 43.57 33.23

ND

(89) (39) ND (9) (27)

ND 38.3 2.98 ND 1.75 0.68 77.8

ND 23.6 3.78 ND 2.38 15.4 13.8

CLB FCR NT Multiple FCR Multiple FCR

19.10 103.00 229.00 7.50 215.83 12.99 7.11

1.82† 4.48† 1.73† 3.5‡ 9.49† 1.9† 3.17‡

0.12 0.24 0.59 0.53 0.18 0.22 0.63

10.49 22.99 132.37 2.14 22.74 6.84 2.24

34.44 5.18 44.71 15.03 6.82 12.74 5.31

(69) (97) (ND)

60.2 6.46 53.7 19.3

98.5 99.7 31.3 92.3

13.45 8.12 62.17 3.65

1.9† 2.22‡ 2.01† 1.76‡

0.17 0.89 0.07 0.61

7.08 3.66 30.93 2.07

32.11 2.18 24.26 88.01

(58)

30.4 6.74

94.4 3

Multiple Multiple FC/CHOP F, C, R, FC, FCM FCR ND

8.12 24.10

1.91‡ 9.83‡

0.34 0.43

4.25 2.45

9.74 39.49

∗Chronic lymphocytic leukemia (CLL) patient samples were selected on the basis of ATM/TP53 function. Normal: no evidence of a functional defect; CAT1: TP53 dysfunction; CAT2: ATM dysfunction. Sensitivity to 2-FaraA as a single agent and in combination with SNX-7081 at a ratio of 100:1 was evaluated using the MTT [3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay and expressed relative to an untreated control. Combination indices (CIs) were calculated using the median effect principle of Chou and Talalay [25], where values 1 are indicative of synergy. The combination index for each patient at a fractional effect (Fe) of 0.5, which corresponds to 50% cell survival, is shown. A dose reduction index (DRI) of 1 suggests a reduction in the 50% inhibitory concentration (IC50) for the respective drug. Integrity of the 11q and 17p genetic regions was assessed by fluorescence in situ hybridization (FISH); / and / indicate no loss and a heterozygous deletion, respectively. The clone size of each deletion is shown in parentheses. Samples were assessed for ZAP-70 and CD38 expression by flow cytometry, with cut-offs of 10 and 7%, respectively, being considered as positive. Treatment up to the date of sampling is indicated. †Indicates a 95% and ‡indicates a 75% reduction in the IC50 of 2-FaraA. NT, not treated; F, fludarabine; C, cyclophosphamide; R, rituximab; M, mitoxantrone; CLB, chlorambucil; CHOP, cyclophosphamide, daunorubicin, vincristine, prednisolone; Multiple, three or more treatments with the agents listed; ND, not determined.

Analysis of synergy The synergistic effects of 2-FaraA and SNX-7081 were determined by performing dose–response analyses by MTT assay with combinations of the two drugs at a fixed ratio based on the IC50 for each drug as a single agent. Treatments were carried out for 48 h. The median effect principle described by Chou and Talalay [25] was used to calculate a combination index (CI) at a range of fractional effect levels for the drug combination, where, for example, a fractional effect of 0.5 represents a 50% reduction in the number of viable cells relative to a vehicle-treated control. For each of the patient samples or cell lines, median effect dose–response relationships were drawn for 2-FaraA and SNX-7081 as single agents and in combination [example shown in Figure 1(A)], where Fa and Fu are the fraction of cells affected and unaffected, respectively, and D represents the drug dose. The concentrations of 2-FaraA and SNX-7081 necessary to induce a range of fractional effects from 0.1 to 0.9 were ascertained from the plots and used to calculate CI values according to the following equation: CI A1/A2 B1/B2 where A1 and B1 indicate the concentrations of the drugs alone and A2 and B2 indicate the doses required in combination to elicit a given fractional effect. In this model, CI values of 1, equal to 1 and 1 are indicative of synergy, additivity

and antagonism, respectively. The effect of the combination of SNX-7081 and 2-FaraA was further assessed by calculating a dose reduction index (DRI) for each drug against cell line and patient sample. DRI values were calculated according to the following equation: DRI [A1]/[A2] where A1 and A2 represent IC50 values for SNX7081 or 2-FaraA as single agents or in combination with the other drug, respectively. A DRI of 1 is indicative of a reduction in drug dosage.

Cell cycle analysis and colony formation Cell cycle analysis was performed using PI staining and flow cytometry. Following treatment, cells were fixed and permeabilized in ice-cold 70% ethanol for at least 2 h at 20°C. Cells were then pelleted by centrifugation and stained in a solution containing 50 μg/mL PI, 0.1 mg/mL RNAse-A and 0.1% Triton-X100. After 30 min at room temperature the DNA content of the cells was analyzed by flow cytometry and the proportion of cells in each cycle phase assessed using ModFit software. During normal proliferation, MEC1 and MEC2 cells spontaneously aggregate and form colonies in culture [21]. The effects of 2-FaraA and SNX-7081 alone and in combination were

O. G. Best et al.

Log (fa/fu)

A

-2

B

R2= 0.9751

2

R2= 0.9824

1

R2= 0.9511

0

-1

Surviving Fraction (%)

1370

0

1

2

3

-1

4

Log (D)

-2

NORMAL

100

CAT 1

80

CAT 2

60 40 20 0

2-FaraA SNX7081

-3

120

0

1

2-FaraA+SNX7081

NORMAL Combination Index

WILD-TYPE Combination Index

D

3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.4

0.6

2.5 2.0 1.5 1.0 0.5

0.8

0.4

0.5

0.6

0.8

0.7

0.9

2.5

3.0 2.5

CAT 1 Combination Index

MUTATED Combination Index

100

0.0 0.2

2.0 1.5 1.0 0.5

2.0 1.5 1.0 0.5 0.0

0.0 0.2

0.4

0.6

0.8

0.2

3.0

0.3

0.4

0.5

0.6

0.7

0.8

0.9

2.5

2.5

CAT 2 Combination Index

NULL Combination Index


C

5 10 50 2-FaraA (µM)

2.0 1.5 1.0 0.5

2.0 1.5 1.0 0.5 0.0

0.0 0.2

0.4 0.6 Fractional effect

0.8

0.2

0.3

0.4 0.5 0.6 0.7 Fractional effect

0.8

0.9

Figure 1. 2-FaraA and SNX-7081 synergy in cell lines and patient samples. (A) Combination index/fractional effect plots for the cell lines and (B) patient cells grouped according to the mutational or functional status of TP53 as indicated. A combination index of 1 (dotted line) indicates synergy. (C, D) Relationship between synergy and sensitivity to fludarabine alone in the cell lines and patient samples. Results shown represent the combination index and 2-FaraA dose at a fractional effect of 0.5 and the IC50 for 2-FaraA.

investigated on the formation of these clusters. Cells were treated with 10 μM 2-FaraA and various concentrations of SNX7081 for 48 h. After treatment the cells were washed, counted and seeded at 1 105 cells/mL in fresh media for a further 72 h. The number of viable colonies present was determined by counting colonies capable of converting MTT to blue formazan crystals, as described above for the MTT cytotoxicity assay. All

experiments were repeated in triplicate. Images of the colonies were acquired using an Olympus® (Hamburg, Germany) inverted microscope with digital imaging hardware.

Statistical analyses All statistical analyses were carried out using a two-tailed Student’s t-test. p-Values 0.05 were considered significant.

Hsp90 inhibition and fludarabine sensitivity 1371

Results


2-FaraA sensitivity in vitro correlates strongly with ATM/TP53 mutational status of cell lines and CLL patient samples Initially, we demonstrated a strong correlation between TP53/ATM dysfunction (CAT1 and CAT2, respectively) or mutational status and resistance to 2-FaraA in the patient samples [Figure 1(B)] and cell lines (Table I). The mean IC50 for 2-FaraA in the 10 functionally normal CLL patient samples was 3.4 μM (range 0.97–7.1 μM) (Table II). In contrast, the mean IC50 values for the seven TP53/CAT1 and six ATM/ CAT2 samples were markedly higher at 98 μM (range 7.5–230 μM) and 18 μM (range 3.6–62 μM), respectively. The mean IC50 for 2-FaraA differed significantly between the TP53 functional and dysfunctional samples but also between samples in the TP53/CAT1 and ATM/CAT2 groups (p 0.05 in both cases). Patient samples were deemed resistant to 2-FaraA if the IC50 was 7 μM, which has been reported as being the upper limit of the clinically achievable plasma level [26]. Twelve of the 13 patient samples categorized as TP53/CAT1 or ATM/CAT2 dysfunctional were considered to be 2-FaraA resistant. In contrast, only one of the 10 functionally normal cases had an IC50 for 2-FaraA of 7 μM. Among both the TP53/CAT1 and ATM/CAT2 cases, four of six were identified by FISH as having loss of genetic material encompassing the TP53 (17p13) and ATM (11q23) loci, respectively. Importantly, three of the four cases (patients 13, 15 and 23) without loss of either TP53 or ATM, but dysfunctional by flow cytometric assay, were resistant to 2-FaraA. With the exception of the HL60 line, which is TP53-null and sensitive to fludarabine, TP53 mutation was associated with resistance to 2-FaraA in the cell lines (Table I). TP53independent apoptosis in response to DNA-damage has been described previously [27], but only appears to occur in certain cell types when the cells are in cycle. The specific reasons for the fludarabine-sensitivity of the HL-60 line remain unclear.

Synergy between 2-FaraA and SNX-7081 We determined whether the combination of 2-FaraA and SNX-7081 was synergistic against the patient cells and cell lines by using the method of Chou and Talalay to calculate CIs based on the cytotoxic effects of the drugs alone or in combination. A CI of 1 [dotted line, Figures 1(C) and 1(D)] indicates synergy between the agents. For the TP53 mutated MEC1, MEC2 and U266 cell lines, synergy was apparent at fractional effect levels 0.2, indicating that the drugs synergize in these lines when more than 20% of the cells are killed by the combination. However, at no effect level was synergy apparent in the TP53 wild-type or null lines [Figure 1(C)]. A similar analysis of the CLL patient samples demonstrated that the combination of 2-FaraA and SNX-7081 was synergistic in 20 of the 23 samples. In three of the functionally normal samples a CI value of greater than 1 was indicative of antagonism between the drugs. Synergy was apparent at significantly lower fractional effects (p 0.05) in the CAT1 and CAT2 cases ( 0.3 and 0.4, respectively) compared with the functionally normal cases in which synergy was apparent

at a mean fractional effect of 0.6 [Figure 1(D)]. In contrast to the ATM/TP53 functional status, we observed no correlation between ZAP-70 and CD38 status and degree of synergy between 2-FaraA and SNX-7081 (Table II). Concomitant with synergy between 2-FaraA and SNX7081 we observed a significant reduction in 2-FaraA IC50 in all the cell lines and patient samples. In all three TP53 mutated cell lines SNX-7081 reduced the IC50 for 2-FaraA by 95%. In the patient samples a similar pattern of increased 2-FaraA sensitivity was observed; all 13 of the CAT1 or CAT2 cases showed a 75% reduction in 2-FaraA IC50, with 95% reduction in 5/7 CAT1 and 2/6 CAT2 cases. The more significant decrease in IC50 for 2-FaraA in the TP53/ATM mutated cell lines and patient cases is reflected by the significantly higher (p 0.05) DRI in these cases compared to the TP53 wild-type or functionally normal cases (Tables I and II). The mean DRI for 2-FaraA in the TP53 mutated cell lines was 29.14 compared to 1.89 in the TP53 wild-type/null lines. In the patient samples the mean DRIs for 2-FaraA in CAT1 and CAT2 were 28.55 and 8.41, respectively, compared with 1.09 in the functionally normal cases. No statistical difference was observed in the DRI for 2-FaraA between cases of CLL with TP53- or ATM-related dysfunction. Further evidence of the synergy between 2-FaraA and SNX-7081 is evident in the positive DRI for SNX-7081 in all the cell lines and patient samples. There was no significant correlation between DRI and TP53 mutational status in the cell lines or patient samples.

Induction of early stage apoptosis in response to 2-FaraA and SNX-7081 To further investigate the effect of combining 2-FaraA and SNX-7081 on the TP53 mutated cell lines and TP53/CAT1 or ATM/CAT2 patient samples we determined the effect on the proportion of cells undergoing apoptosis using the mitochondrial membrane potential dye DilC1(5) and propidium iodide. We observed a significantly greater proportion of apoptotic cells in cultures of the three cell lines and in representative patient samples exposed to the combination of SNX-7081 and 2-FaraA compared with those treated with 2-FaraA or SNX-7081 alone (p 0.05). A significantly greater-than-additive effect (synergy) of the combination was observed at both 50 nM and 100 nM SNX-7081 [Figure 2(A)] against the cell lines and at a range of drug ratios against the representative patient samples [Figure 2(B)].

Effect of SNX-7081 and 2-FaraA on cell cycle and colony formation in cultures of CLL cell lines MEC1 and MEC2 Colony forming assays are commonly used to evaluate the efficacy of anti-cancer agents [28]. Although semi-solid medium is often required for cells to form colonies, certain cell lines and cells in primary culture form spontaneous clusters under normal culture conditions, including the MEC cell lines [21]. We observed this phenomenon in cultures of the MEC1 and MEC2 cell lines and sought to determine the effects of 2-FaraA and SNX-7081, alone and in combination, on the ability to form colonies. Cultures of both lines were treated for 48 h with the indicated concentrations of 2-FaraA and SNX-7081, washed into fresh media, counted

1372

O. G. Best et al.

60

MEC1 MEC2 U266

50 40

*

20 10

00 +1

0

10µ 0µM FdA + SNX7081 (nM)

SNX7081 (nM)

* *

3µM FdA + 3µ SNX7081 (nM)

0 +5 +1 0 0 +2 0 0 +3 0 00

0

*

+3

0

SNX7081 (nM)

+1

0

30

0

20

10

50

*

Fd A 10 30

Fd

50 10 0 20 0 30 0 +1 0 +3 0 +5 +1 0 0 +2 0 0 +3 0 00

*

35 30 25 20 15 10 5 0

3µ M

*

*

3µ

M

2 (ATM) ( CAT2 *

DilC1(5) negative / PI negative cells (%)

CAT 1 (TP53) ( 45 40 35 30 25 20 15 10 5 0

A 10 30



+5

10

50

0

0

SNX7081 (nM)

B

* *

*

30

1 2- 0µM Fa ra A


A

3µM FdA + 3µ SNX7081 (nM)

Figure 2. Apoptosis in response to 2-FaraA and SNX-7081 alone and in combination. Representative histograms showing the proportion of cells with loss of mitochondrial membrane potential. (A) Proportion of apoptotic MEC1, MEC2 and U266 cells following exposure to fludarabine, SNX-7081 or their combination. (B) Mean ( SD) proportion of apoptotic CLL cells from a representative CAT1 and CAT2 dysfunctional sample following treatment with fludarabine, SNX-7081 or their combination. Experiments were performed in triplicate.

and re-plated at a density of 1 105 cells/mL in triplicate wells of a 96-well plate. Following culture for a further 72 h in fresh media, viable colonies were identified by the addition of MTT reagent [representative image in Figure 3(A)]. The number of colonies containing 50 viable cells in each well was quantified. Results shown are means ( SD) from three independent experiments [Figure 3(B)]. Neither 10 μM 2-FaraA nor 50 or 100 nM SNX-7081 had any significant effect on the number of dense blue cell colonies observed, compared with a vehicle-treated control culture. SNX-7081 at the higher concentrations of 200 and 300 nM caused a significant reduction in colony number in cultures of MEC2, but not MEC1 cells. Combinations of 2-FaraA and SNX-7081 resulted in greater-than-additive inhibition of colony formation in both cell lines. However, the two cell lines differed in the concentrations of SNX-7081 necessary to inhibit cluster formation when in combination with 10 μM 2-FaraA. In the MEC2 line the effects were evident at 50 and 100 nM SNX-7081, compared with 200 and 300 nM in the MEC1 line [Figure 3(B)]. The effect of SNX-7081 and 2-FaraA on cell cycle progression and colony forming ability was investigated in cultures of MEC1 and 2 cells. Cell cycle analysis demonstrated that the combination of SNX-7081 with 2-FaraA induced a significant increase (p 0.05) in the proportion of cells in G2/M and a decrease in cells in the G0/G1 phase compared with the drugs as single agents [Figure 3(C)]. A significant increase in cells

with a sub-G1 DNA content, indicative of cells undergoing apoptosis, was also observed in cells exposed to both agents (data not shown). No significant change in the proportion of cells in S-phase was observed.

Discussion Resistance to fludarabine, often resulting from the clonal expansion of cells with lesions in the TP53 pathway, represents a major challenge in the clinical management of CLL, and there is a substantial need for novel therapeutic approaches in these patients. The geldanamycin-derived Hsp90 inhibitors have provided compelling evidence in support of targeting Hsp90 for therapy. However, toxicity and low response rates against advanced cancers have been issues with the clinical utility of geldanamycin-derived compounds, including 17-AAG [29,30], prompting the development of synthetic inhibitors such as SNX-7081 [31]. Our previous study on SNX-7081 demonstrates that SNX-7081 is significantly more potent in vitro than the geldanamycin derivative, 17-AAG, against cell lines and patient samples irrespective of mutations of ATM or TP53 [5]. Although these data support the use of SNX-7081 as a single agent, unprecedented response rates of fludarabine-based regimens suggest that combinations that include the purine analog are currently the most efficacious in the treatment of CLL. Given that combining inhibitors of Hsp90 with other agents may also lower their toxicity,

Hsp90 inhibition and fludarabine sensitivity 1373

A CONTROL

2-FaraA

2-FaraA + SNX-7081

SNX-7081

MEC2

Colonies formed (% of control)

140 120

MEC1

100

MEC2

80 60 40 *

20 0 50

aA

2-

100

200

300

50

SNX7081 (nM)

*

100

200 300

2-FaraA + SNX7081 (nM) 100

100 Percent of cells

80 60 40 20

60 40 20

S

G2/M

08 1 2SN Fa X7 raA 08 + 1

X7

A ra Fa 2-

G0/G1

SN

l tro on

08 X7 SN

Fa

ra

l 2-

tro on C

G0/G1

1 2SN Fa X7 raA 08 + 1

0 A

0

80

C

C

* *

r Fa

Percent of cells


B

S

G2/M

Figure 3. Colony formation in MEC1 and MEC2 cell lines following treatment with 2-FaraA and SNX-7081. (A) Representative images of viable (blue) colonies of MEC1 and MEC2 cells following 48 h treatment and 72 h culture in fresh medium. (B) Quantification of the number of colonies of 50 cells per well following treatment as indicated. All values are relative to vehicle-treated control cultures. *Indicates significant difference (p 0.05) between observed number of colonies and expected from the effects of 2-FaraA and SNX-7081 as single agents. Error bars are mean SD of three independent experiments. (C) Cell cycle analysis of MEC1 (left histogram) and MEC2 (right histogram) cell lines in response to SNX-7081 and 2-FaraA alone and in combination.

we conducted the first investigation into the effects of combining SNX-7081 with fludarabine, specifically focusing on the efficacy of this combination against TP53/ATM mutated, fludarabine-resistant cells. The data presented demonstrate that the combination of SNX-7081 with fludarabine exerts a strong synergistic action against fludarabine-resistant cells and that this synergy is concomitant with a significant increase in the sensitivity of these cells to fludarabine. This synergy was evident as a significant reduction in the IC50 for fludarabine to within a clinically achievable range (3–7 μM) and as an increase in the proportion of apoptotic cells, most notably in TP53 mutated lines and patient cells. Initially we confirmed the association between ATM/TP53 aberrations and fludarabine resistance using the MTT assay in seven hematological cell lines and 23 CLL patient samples. All three of the cell lines and 12 of the 13 patient samples with TP53 mutation or ATM/TP53 dysfunction exhibited resistance to fludarabine in vitro relative to TP53 normal control cells. These data support those in a previous study in which the authors suggest the MTT assay may have utility in identifying patients likely to exhibit a poor response to fludarabine

treatment [32], and reinforce the association of fludarabinerefractory disease with lesions in the ATM/TP53 pathway. An important aspect of the present study was the use of a unique, easily performed and rapid-throughput TP53 functional assay that enabled us to further classify the fludarabineresistant cases and identify patients unlikely to benefit from the FCR regimen due to mutations of TP53. The majority of such cases could be identified from loss of the ATM and TP53 loci by FISH analysis, but inactivating mutations of TP53 occur in the absence of genetic loss [33]. Importantly, the two CAT1 and one CAT2 dysfunctional cases without loss of TP53 or ATM identified in the present study (cases 13, 15 and 23; Table II) were found to be highly resistant to 2-FaraA (IC50 20 μM). In theory, there may be aberrations other than mutations of TP53 or ATM that confer resistance to 2-FaraA and may elicit a response similar to the pattern of dysfunction detected by our assay, such as underexpression of miR34a [34]. However, our experience to date with this assay [22] suggests that CAT1 or CAT2 dysfunction is strongly associated with mutations of TP53 or ATM, respectively. Given that the data presented here demonstrate a good correlation between CAT1 or CAT2


1374

O. G. Best et al.

dysfunction and 2-FaraA sensitivity in vitro and that patients with non-functional TP53, whether as a consequence of mutation or other mechanisms, respond poorly to FCR [33], functional screening as employed in the present study could be considered for all in vitro and clinical trials of novel therapies for 2-FaraA-resistant CLL. The synergy of SNX-7081 with 2-FaraA we describe is not surprising given the wide range of cellular functions dependent on Hsp90. Synergy of Hsp90 inhibitors derived from geldanamycin with radiation [15] and anti-angiogenic and chemotherapeutic agents, including fludarabine [6], has been described previously. Jones et al. [6] observed a significant reduction in the IC50 for fludarabine against primary CLL cells when combined with geldanamycin, and observed the greatest effect in samples resistant to fludarabine as a single agent. We observed a similar pattern of response to the combination of SNX-7081 and 2-FaraA, but by focusing on samples with fludarabine resistance due to mutation of TP53 or ATM, our data suggest that the combination may be most beneficial in the treatment of those patients least likely to respond to FCR therapy due to fludarabine resistance following primary treatment failure or early relapse. In fact, our data suggesting that the combination may be antagonistic in three of the four 2-FaraA-sensitive cell lines and three of the 10 ATM/TP53 functionally normal cases (Table II) indicate that for translation into the clinical setting, this combination should only be considered following failure of fludarabinebased therapy or when there is evidence of a TP53 pathway lesion. Furthermore, the high proportion of heavily pretreated patients in the TP53/CAT1 and ATM/CAT2 groups suggests that this combination may also be effective as a salvage therapy to re-establish fludarabine sensitivity in patients whose CLL has become fludarabine-refractory. Although the study by Jones et al. [6] illustrates the potential of combining an Hsp90 inhibitor with fludarabine, toxicity associated with geldanamycin would preclude its clinical utilization. Despite being distinct from the geldanamycin derivatives, toxicity may also be an issue with SNX-7081. Our demonstration of a dose reduction index for SNX-7081 of 1 in all the cell lines and patient samples suggests that combination of SNX-7081 with 2-FaraA may avoid or reduce levels of toxicity. The mechanisms of the synergy documented here are currently unclear. However, evidence in the literature has led us to hypothesize that the synergy may depend on the effects of Hsp90 inhibition on proteins involved in DNA repair [16,35,36]. Restricting DNA repair in the presence of DNA damage induced by 2-FaraA may potentiate the initiation of apoptotic signals. This mechanism may be particularly relevant in cases with TP53/ATM dysfunction, since deficiencies in this pathway would allow DNA repair to proceed unchecked following DNA damage. This hypothesis requires confirmation but is in keeping with the stronger synergy we observed in those cell lines and patient cells with defects of the TP53/ATM pathway. Clinically aggressive and advanced CLL is commonly associated with ATM and TP53 aberrations and typically manifests as bulky nodal disease and dense marrow infiltration, with associated marrow failure resulting from proliferative stimuli from the CLL microenvironment. The synergistic effect of the

SNX-7081 with fludarabine combination may function by modifying the CLL cells in their tumor microenvironment. In this context, Hsp90 inhibition has been shown to inhibit angiogenesis through abrogation of vascular endothelial growth factor (VEGF) production [37], which is critical for the survival of CLL cells in the marrow and lymph nodes. As these tissues are believed to include areas that promote CLL cell proliferation, our observations concerning the effects of SNX-7081 and 2-FaraA on cell cycle progression in the MEC1 and MEC2 lines may be particularly relevant in the treatment of CLL, with the higher proliferating cell fraction that occurs in progressive disease. The accumulation of cells in the G2/M phase suggests that SNX-7081 may promote 2-FaraA-induced accumulation of cells in G2/M, as described in a 2-FaraA sensitive myeloid leukemia line [38]. As well as the effects on cell cycle progression, we show that SNX-7081, in combination with 2-FaraA, has a synergistic and inhibitory effect on formation of the cell colonies associated with normal proliferation of the MEC cells. This is similar to findings with other drug combinations against other leukemic cell lines [19]. Taken together, the data from the cell cycle and colony formation assays demonstrate a marked synergistic effect of SNX-7081 in combination with 2-FaraA on the normal proliferation of MEC cells. In conclusion, our data demonstrate a strong in vitro synergy between the novel Hsp90 inhibitor SNX-7081 and fludarabine against hematological cell lines and CLL patient samples that are fludarabine-resistant as a consequence of lesions in the ATM/TP53 pathway. The combination of SNX-7081 and fludarabine may be effective in the treatment of aggressive and fludarabine-refractory CLL, particularly those with mutations and defects of the TP53 pathway, by re-establishing fludarabine-sensitivity.

Acknowledgements The study was funded by a grant from the Leukemia Foundation of Australia. The authors would like to thank Lyndsay Peters for his help with the flow cytometric experiments. Potential conflict of interest: Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.

References [1] Tam CS, O’Brien S, Wierda W, et al. Long-term results of the fludarabine, cyclophosphamide, and rituximab regimen as initial therapy of chronic lymphocytic leukemia. Blood 2008;112:975–980. [2] Tsimberidou AM, Tam C, Abruzzo LV, et al. Chemoimmunotherapy may overcome the adverse prognostic significance of 11q deletion in previously untreated patients with chronic lymphocytic leukemia. Cancer 2009;115:373–380. [3] Enoch T, Norbury C. Cellular responses to DNA damage: cell-cycle checkpoints, apoptosis and the roles of p53 and ATM. Trends Biochem Sci 1995;20:426–430. [4] Wattel E, Preudhomme C, Hecquet B, et al. p53 mutations are associated with resistance to chemotherapy and short survival in hematologic malignancies. Blood 1994;84:3148–3157. [5] Best OG, Singh N, Forsyth C, et al. The novel Hsp-90 inhibitor SNX7081 is significantly more potent than 17-AAG against primary CLL cells and a range of haematological cell lines, irrespective of lesions in the TP53 pathway. Br J Haematol 2010;151:185–188.


Hsp90 inhibition and fludarabine sensitivity 1375 [6] Jones DT, Addison E, North JM, et al. Geldanamycin and herbimycin A induce apoptotic killing of B chronic lymphocytic leukemia cells and augment the cells’ sensitivity to cytotoxic drugs. Blood 2004;103:1855– 1861. [7] Mahalingam D, Swords R, Carew JS, et al. Targeting HSP90 for cancer therapy. Br J Cancer 2009;100:1523–1529. [8] Cuni S, Perez-Aciego P, Perez-Chacon G, et al. A sustained activation of PI3K/NF-kappaB pathway is critical for the survival of chronic lymphocytic leukemia B cells. Leukemia 2004;18:1391–1400. [9] Kim YS, Alarcon SV, Lee S, et al. Update on Hsp90 inhibitors in clinical trial. Curr Top Med Chem 2009;9:1479–1492. [10] Solit DB, Ivy SP, Kopil C, et al. Phase I trial of 17-allylamino17-demethoxygeldanamycin in patients with advanced cancer. Clin Cancer Res 2007;13:1775–1782. [11] Solit DB, Osman I, Polsky D, et al. Phase II trial of 17-allylamino17-demethoxygeldanamycin in patients with metastatic melanoma. Clin Cancer Res 2008;14:8302–8307. [12] Chandarlapaty S, Sawai A, Ye Q, et al. SNX2112, a synthetic heat shock protein 90 inhibitor, has potent antitumor activity against HER kinase-dependent cancers. Clin Cancer Res 2008;14:240–248. [13] Jin L, Xiao CL, Lu CH, et al. Transcriptomic and proteomic approach to studying SNX-2112-induced K562 cells apoptosis and anti-leukemia activity in K562-NOD/SCID mice. FEBS Lett 2009;583:1859–1866. [14] Okawa Y, Hideshima T, Steed P, et al. SNX-2112, a selective Hsp90 inhibitor, potently inhibits tumor cell growth, angiogenesis, and osteoclastogenesis in multiple myeloma and other hematologic tumors by abrogating signaling via Akt and ERK. Blood 2009;113:846–855. [15] Bull EE, Dote H, Brady KJ, et al. Enhanced tumor cell radiosensitivity and abrogation of G2 and S phase arrest by the Hsp90 inhibitor 17(dimethylaminoethylamino)-17-demethoxygeldanamycin. Clin Cancer Res 2004;10:8077–8084. [16] Koll TT, Feis SS, Wright MH, et al. HSP90 inhibitor, DMAG, synergizes with radiation of lung cancer cells by interfering with base excision and ATM-mediated DNA repair. Mol Cancer Ther 2008;7:1985– 1992. [17] Georgakis GV, Li Y, Rassidakis GZ, et al. The HSP90 inhibitor 17-AAG synergizes with doxorubicin and U0126 in anaplastic large cell lymphoma irrespective of ALK expression. Exp Hematol 2006;34:1670–1679. [18] Munster PN, Basso A, Solit D, et al. Modulation of Hsp90 function by ansamycins sensitizes breast cancer cells to chemotherapyinduced apoptosis in an RB- and schedule-dependent manner. See: E. A. Sausville, Combining cytotoxics and 17-allylamino, 17demethoxygeldanamycin: sequence and tumor biology matters. Clin Cancer Res 7: 2155–2158, 2001. Clin Cancer Res 2001;7:2228–2236. [19] Pelicano H, Carew JS, McQueen TJ, et al. Targeting Hsp90 by 17AAG in leukemia cells: mechanisms for synergistic and antagonistic drug combinations with arsenic trioxide and Ara-C. Leukemia 2006;20:610–619. [20] Hallek M, Cheson BD, Catovsky D, et al. Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines. Blood 2008;111:5446–5456. [21] Stacchini A, Aragno M, Vallario A, et al. MEC1 and MEC2: two new cell lines derived from B-chronic lymphocytic leukaemia in prolymphocytoid transformation. Leuk Res 1999;23:127–136.

[22] Best OG, Gardiner AC, Majid A, et al. A novel functional assay using etoposide plus nutlin-3a detects and distinguishes between ATM and TP53 mutations in CLL. Leukemia 2008;22:1456–1459. [23] Orchard JA, Ibbotson RE, Davis Z, et al. ZAP-70 expression and prognosis in chronic lymphocytic leukaemia. Lancet 2004;363: 105–111. [24] Cossarizza A, Kalashnikova G, Grassilli E, et al. Mitochondrial modifications during rat thymocyte apoptosis: a study at the single cell level. Exp Cell Res 1994;214:323–330. [25] Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors. Adv Enzyme Regul 1984;22:27–55. [26] Gandhi V, Kemena A, Keating MJ, et al. Cellular pharmacology of fludarabine triphosphate in chronic lymphocytic leukemia cells during fludarabine therapy. Leuk Lymphoma 1993;10:49–56. [27] Strasser A, Harris AW, Jacks T, et al. DNA damage can induce apoptosis in proliferating lymphoid cells via p53-independent mechanisms inhibitable by Bcl-2. Cell 1994;79:329–339. [28] Shoemaker RH, Wolpert-DeFilippes MK, Kern DH, et al. Application of a human tumor colony-forming assay to new drug screening. Cancer Res 1985;45:2145–2153. [29] Goetz MP, Toft D, Reid J, et al. Phase I trial of 17-allylamino-17demethoxygeldanamycin in patients with advanced cancer. J Clin Oncol 2005;23:1078–1087. [30] Ramanathan RK, Trump DL, Eiseman JL, et al. Phase I pharmacokinetic-pharmacodynamic study of 17-(allylamino)-17demethoxygeldanamycin (17AAG, NSC 330507), a novel inhibitor of heat shock protein 90, in patients with refractory advanced cancers. Clin Cancer Res 2005;11:3385–3391. [31] Cysyk RL, Parker RJ, Barchi JJ Jr, et al. Reaction of geldanamycin and C17-substituted analogues with glutathione: product identifications and pharmacological implications. Chem Res Toxicol 2006;19: 376–381. [32] Morabito F, Stelitano C, Callea I, et al. In vitro drug-induced cytotoxicity predicts clinical response to fludarabine in B-cell chronic lymphocytic leukaemia. Br J Haematol 1998;102:528–531. [33] Zenz T, Krober A, Scherer K, et al. Monoallelic TP53 inactivation is associated with poor prognosis in chronic lymphocytic leukemia: results from a detailed genetic characterization with long-term followup. Blood 2008;112:3322–3329. [34] Zenz T, Mohr J, Eldering E, et al. miR-34a as part of the resistance network in chronic lymphocytic leukemia. Blood 2009;113:3801–3808. [35] Ha K, Fiskus W, Rao R, et al. Hsp90 inhibitor-mediated disruption of chaperone association of ATR with hsp90 sensitizes cancer cells to DNA damage. Mol Cancer Ther 2011;10:1194–1206. [36] Stingl L, Stuhmer T, Chatterjee M, et al. Novel HSP90 inhibitors, NVP-AUY922 and NVP-BEP800, radiosensitise tumour cells through cell-cycle impairment, increased DNA damage and repair protraction. Br J Cancer 2010;102:1578–1591. [37] Sanderson S, Valenti M, Gowan S, et al. Benzoquinone ansamycin heat shock protein 90 inhibitors modulate multiple functions required for tumor angiogenesis. Mol Cancer Ther 2006;5:522–532. [38] Sampath D, Shi Z, Plunkett W. Inhibition of cyclin-dependent kinase 2 by the Chk1-Cdc25A pathway during the S-phase checkpoint activated by fludarabine: dysregulation by 7-hydroxystaurosporine. Mol Pharmacol 2002;62:680–688.