Toward personalized therapy for chronic lymphocytic leukemia

Abstract

The differences in clinical course of chronic lymphocytic leukemia could have an impact on variations in a patient’s response to therapy. Our published results revealed that thermal transition (95 ± 5°C) in differential scanning calorimetry profiles appear to be characteristic for the advanced stage of CLL. Moreover, a decrease/loss of this transition in nuclei from leukemic cells exposed to drugs ex vivo could indicate their diverse efficacy. It seems that the lack of changes in thermal profile could predict patient’s drug resistance. In this study, we demonstrate the results obtained after drug treatment of leukemic cells by calorimetry, apoptosis-related parameters involved in expression of genes using cDNA microarray and western blot. These data were compared with the patients’ clinical parameters before and after RCC therapy (rituximab + cladribine + cyclophosphamide). The complementary analysis of studied cases with opposite clinical response (CR or NR) revealed a strong relationship between clinical data, differences in thermal scans and apoptosis-related gene expression. We quantified expression of eight of apoptosis-related 89 genes, i.e., NOXA, PUMA, APAF1, ESRRBL1, CASP3, BCL2, BCL2A1 and MCL1. Particular differences in NOXA and BCL2 expression were revealed. NOXA expression in cells of patients who achieved a complete response to RCC therapy was 0.44 times higher in comparison to control ones. Interestingly, in the case of patients who did not respond to immunotherapy, NOXA expression was highly downregulated (RQ = 4.39) as compared with untreated cells. These results were confirmed by distinct cell viability, protein expression as well as by differences in calorimetry profiles.

Introduction

Despite the growing list of chronic lymphocytic leukemia (CLL) treatment options, this hematological neoplasm still remains incurable and highly heterogeneous.¹ The reason for this appears to be the variability in the clinical course of disease, which could originate from aberrations on a genetic level.^2,3 Moreover, the simultaneous coexistence of two populations of CLL cells—quiescent in peripheral blood and highly proliferative in patient’s bone marrow germinal centers, requires therapy effective for both resting and cycling CLL cell populations.^4-6 Over the past 20 y, an improvement in CLL treatment has occurred by application of antibiotics, corticosteroids, purine analogs and monoclonal antibodies.^7-11 The combined therapy of purine analogs with alkylator significantly increased patients’ response to administered therapy.^8,10,11 There are a large number of new anticancer compounds currently undergoing preclinical or clinical tests.¹² Some of them reflect their action toward uncycling and cycling cell populations.^6,7,12

Despite the increasing number of clinical diagnostic parameters and overall improvement of patients’ response to therapy, there are still patients who display a poor reaction to administered drugs.^13,14 The results revealed that in the group of CLL patients treated previously with chlorambucil or fludarabine, an increase in resistance to a second course of drug administration occurred.¹⁵ Therefore, there is a need to tailor therapy for patients as a strategy to avoid ineffective treatment, increase the complete remission rates and to improve patients’ survival rates.

Over the last decade, microarray technology has been an innovative tool to study gene expression. This technique provides detailed information about expression levels of simultaneously examined genes and could be helpful in treatment-monitoring of gene expression, particularly related to apoptosis. The limited data concerning apoptotic gene expression changes in CLL cells, mainly obtained from in vivo/ex vivo treatment.^16-18 In vitro-reported studies were conducted on CLL cells exposed to purine analogs and directed mainly to P53-related genes and drug resistance.^16,17

The aim of the present study was to evaluate the ex vivo potential of apoptosis induction in primary tumoral peripheral blood mononuclear cells (PBMCs) by RCM, i.e., rituximab + cladribine + mafosfamide (ex vivo active form of cyclophosphamide) using Vybrant Apoptosis Assay #4, differential scanning calorimetry and western blot (Mcl-1, Bcl-2, procaspase-3, Noxa). The obtained results compared the data from two cured patients receiving opposite clinical responses described as complete responder (CR) or non-responder (NR) after six cycles of immunochemotherapy. Moreover, the expression in vivo of eight of 89 apoptosis-related genes by qPCR microarray in CLL cells from patient’s blood before and 2 wk after first RCC (rituximab + cladribine + cyclophosphamide) cycle of administration was reported.

One primary goal in CLL therapy is to reach CR with intensive treatment options. However, there is no clinical evidence that therapy intervention in CLL can improve outcome, even if prognostic parameters are considered. Herewith, we describe the result sets of two patients with newly diagnosed CLL who underwent RCC therapy and reached the distinct clinical response, i.e., complete remission or lack of response.

Results

Complete responder (CR)

CLL cells from blood of patient PZ who achieved CR displayed mutated status of IgHV (97.57%). After immunochemotherapy, the complete regression of lymph nodes and lack in spleen and liver enlargement was observed (Table 1). The strong decrease of patient PZ’s cell viability within leukemic cells exposed to RCM (32.7%) in comparison to control ones (76.6%) was seen (Fig. 1A). Importantly, in thermal scans of nuclei from PBMCs incubated for 48 h with RCM, a strong decline of thermal transition at about 95° ± 5°C, in contrast to control ones, occurred (Fig. 1B). It should also be underlined that in the immunoblots of drug-treated cell lysates, the dropping of antiapoptotic Mcl-1 and Bcl-2 protein expression and slight proteolysis of procaspase-3 detected with appropriate antibodies were noticed. Simultaneously, an elevated Noxa expression was also seen in the control, untreated as well as exposed to RCM leukemic cells (Fig. 1C). The patient’s overall reactivity manifested by the changes of parameters related to apoptosis as well as the observation of strong decreases in thermal transition at 95 ± 3°C are consistent with our previously reported data,¹⁸ and reveal their predictive value for clinical response for CLL patients much earlier, after ex vivo testing of their PBMC sensitivity to planned treatment options.

Table 1. Clinical characteristics of patients.

Initials	PZ	RJ
Age/sex	65/M	54/M
Rai	I B	II A
Blood count (before treatment) WBC Hb PLT	174.63 x 103/μl 13.7 g/dl 193.0 x 103/μl	126.27 x 103/μl 11.1 g/dl 136.0 x 103/μl
Blood count (assessment of response) WBC Hb PLT	3.00 x 103/μl 11.6 g/dl 108 x 103/μl	1.87 x 103/μl 13.5 g/dl 83 x 103/μl
IgHV mutational status	Mutated	Unmutated
CD 38	20%	21%
β2-microglobulin (mg/L)	4.20	8.78
LDH (U/L) (before treatment)	225	260
Cytogenetics	del 11q, del 13q	del 11q
Lymph nodes diameter (before treatment)	Cervical region 2 × 1 cm Axillary region 3 × 4 cm Inquinal region 2 × 2 cm	Cervical region 3 cm Axillary region 3 cm Inquinal region 3 cm
Lymph nodes diameter (during assessment of response)	Complete regression of peripheral and central lymph nodes	Complete regression of peripheral lymph nodes and < 50% regression of central lymph nodes
Spleen and liver (before treatment)	Not enlarged	Spleen enlarged to 8 cm below the left costal margin; Liver enlarged to 4 cm below the left costal margin;
Spleen and liver (during assessment of response)	Not enlarged	Spleen decreased to 2 cm below the left costal margin (CT = 15 cm); Liver enlarged to 3 cm below the left costal margin;
Treatment response*	CR MRD+	NR

The cut-off values for CD38 was 30%.

No p53 mutation.

^*RCC treatment response after 6 cycles.

Figure 1. Viability of CLL cells (A), differential scanning calorimetry profiles of nuclear fraction (B) and changes in expression of apoptosis-related proteins (Mcl-1, Bcl-2, procaspase-3, Noxa) in PBMC lysates (C), isolated from blood of patients PZ and RJ, incubated for 48 h with and without RCM combination. Immunoblot analysis of PBMCs after their exposure to drug combination were analyzed by alkaline phosphatase method in the presence of appropriate antisera. Protein lysates (50 μg) were loaded into 12.5% polyacrylamide gels; β-actin was used as loading control; Cp^ex - excess heat capacity.

The differences in studied gene expression, before and after treatment, were considered significant only if RQ values was > 2 or < -2 (indicative of upregulation or downregulation, respectively). The heatmap shows the hierarchical clustering of 89 apoptosis-related genes (Fig. 2).

Figure 2. Microarray analysis indicating changes in expression of 89 apoptosis-related genes in two CLL samples following immunotherapy (top); boxplots of the DDCT values of samples from patients PZ and RJ (bottom). The rows and the columns represent individual genes and mRNA samples, respectively. The relative level of gene expression is depicted according to the color scale shown below the matrix. The accuracy of reading in the presented experiments is illustrated by boxplots of the signal (signal DDCT ~-logRQ).

The spectacular differences in NOXA and BCL2 expression levels were detected (Table 2). In the PBMCs of patient PZ, NOXA expression was 1.28 times higher in comparison to control ones (before treatment). However, the expression of MCL1 about four times lower was detected in the case of patient PZ, in comparison to the related its expression for RJ (RQ = 8.60).

Table 2. Alterations in selected gene expression in two CLL cell samples (patients: PZ and RJ) after 2 wk of RCC therapy illustrated as fold change and ΔΔCt values in reference to control samples (before therapy).

Gene description	PZ		RJ
		ΔΔCt	RQ	ΔΔCt	RQ
NOXA (Phorbol-12-myristate-13-acetate-induced protein 1; proapoptotic)	0.44	1.28	-4.39	161.35
CASP3 (Caspase 3; proapoptotic)	-0.32	2.75	-0.43	3.08
BCL2 (B-cell CLL/lymphoma 2; antiapoptotic)	-0.35	2.84	-3.54	69.60
MCL1 (Myeloid cell leukemia sequence 1; antiapoptotic)	-0.78	4.35	-1.45	8.60
PUMA (P53 upregulated modulator of apoptosis; proapoptotic)	-0.19	2.42	-0.44	3.13
BCL2A1 (BCL2-related protein A1; antiapoptotic)	-0.72	4.11	-1.30	7.36
APAF1 (apoptotic protease activating factor 1; proapoptotic)	-0.76	4.29	-0.41	3.01
ESRRBL1 (IFT57-intraflagellar transport 57 homolog; proapoptotic)	0.47	1.24	-3.01	40.58

ΔΔCt, a value of the difference in expression of a gene between two examined groups (before and after RCC treatment); RQ, fold change (fold difference); RCC, rituximab + cladribine + cyclophosphamide

Non-responder (NR)

Patient (RJ), who did not respond to therapy (NR), reflected unfavorable unmutated status of IgHV (Table 1). After PBMCs exposure to RCM, the moderate decrease of PBMC viability, i.e., 69.5%, in comparison to control ones, 93.1% (Fig. 1A), occurred. Concomitantly, a slight decline of thermal transition at about 95 ± 5°C in nuclei from PBMCs exposed to RCM was observed (Fig. 1B). Western blot analysis of lysate proteins revealed the dropping of Mcl-1 and Bcl-2 expression (Fig. 1C). In the same blot, a high expression of uncleaved procaspase-3 was noticed. Moreover, a very low basal expression of Noxa in the lysates of untreated CLL cells that diminished in leukemic cells exposed to drugs was found. NOXA expression level was highly downregulated (RQ = 4.39) as compared with untreated cells (Table 2). The similar trend was seen for antiapoptotic genes BCL2, BCL2A1 and proapoptotic ESRRBL1. The complementary studies conducted ex vivo for patient RJ indicated its lower reactivity following the treatment in vivo in comparison with patient PZ. Interestingly, we have noticed only slight changes in gene expression of CASP3, PUMA and APAF1 in the studied PBMC samples of both patients.

Discussion

The integration of pro-survival microenvironmental signals and variations in the intrinsic apoptotic machinery may contribute to CLL cell resistance to apoptosis induction as well as reaching insufficient efficacy of anticancer therapy.¹⁹ Moreover, the remarkable heterogeneity in clinical course of this disease, as well as distinct patients’ response to therapy, still revealed a serious problem in CLL treatment.²⁰ To date, knowledge on various pro- and antiapoptotic gene expression in response to clinically used apoptosis-induced drug(s) in this type of leukemia is scarce. It is plausible that the delicate balance between anti- and proapoptotic protein level could determine the fate of leukemic cells.^19,20 The scientists underlined the special importance of Mcl-1 as well as Bcl-2 expression levels for inhibition of apoptosis in CLL cells related to their binding potential to proapoptotic proteins, thereby preventing mitochondrial outer membrane potential obstruction due to apoptosis inhibition.²¹

Studies regarding the clinical efficacy of different therapeutic options used for CLL treatment have already been performed; however, there is no data specifically focused on the possible modification of the gene expression profile (GEP) after the administration of drugs.¹⁰ Very recently we have documented the differences in apoptotic GEP among CLL cells exposed to fludarabine or cladribine.²²

Hallaert et al.²³ reported that Mcl-1 differently binds to proapoptotic proteins, Noxa and Bim, in untreated and seliciclib-treated CLL cells. The cited authors showed that Noxa is required to drive the mitochondrial apoptosis upon this novel CDK inhibitor treatment. Noxa protein remained bound to Mcl-1 in seliciclib exposed CLL cells, while Bim is dissociated. They suggested the sequence of events in seliciclib-treated leukemic cells after Mcl-1 degradation, following by Noxa-dependent Bim displacement, its liberation and finally Bax activation and forming large multimeric pores in mitochondria.²³ The described findings are consistent with our results reflecting a high-basal Noxa expression in PBMC cells of patient PZ, who reached CR after RCC administration. However, we have observed a low expression of Noxa before drug exposure of patient RJ’s cells and its lack after cell exposure to RCM.

As previously identified,²⁴ the appearance of additional thermal transition at about 95 ± 5°C zone in cell nuclei obtained from PBMC samples from patients’ blood could be proposed as a new parameter for the advanced stage of CLL. Interestingly, the decrease of this transition appears to have a prognostic significance and could reflect patients’ potential sensitivity to the treatment in vivo. On the other hand, the lack of differences in thermal scans of nuclei isolated from CLL cells ex vivo exposed to drugs for 48 h may suggest patients’ resistance to therapy.

It must be underlined that results obtained ex vivo matched with apoptosis-related gene profiling from the same patients administered with RCC in vivo. We have found that ex vivo and in vivo expression of proapoptotic NOXA could be a significant factor related to different patient’s response to the applied therapy. A high expression of Noxa in patient PZ’s cells seems to correlate with its activity in inhibition of antiapoptotic proteins in mitochondrial outer membrane promoting Noxa proapoptotic activity. Accordingly, in PBMCs from blood of patients who achieved CR to RCC therapy, a strong decline in thermal transition at 95 ± 5°C, accompanied by the increase of main transition at about 83 ± 3°C, was observed.²⁴ The differences in thermal scans of nuclear fraction from drug-treated CLL cell samples seems to be a consequence of chromatin remodeling during the apoptosis process.

Based on the cell viability results, we perceived the strong correlation between DSC and western blot results (expression/proteolysis level of apoptosis-related proteins) accompanied by high relevance to gene expression profiling data and patients’ clinical monitoring in vivo. The highest differences in NOXA and BCL2 expression on gene and protein expression levels between both examined in vivo and ex vivo patients were revealed.

We also matched the results of ex vivo or in vivo experiments with clinical parameters and prognostic markers for these patients. In accordance with accepted hematological clinical parameters for patient PZ’s, who achieved CR to the applied therapy, IGHV-mutated status was confirmed, while patient RJ, who did not reached remission, indicated unfavorable unmutated status of this gene variable region. It is plausible that unmutated IgHV status, and the presence of genomic aberrations (del 11q), reflect link with progressive type of CLL and patients’ shorter survival.²⁵

Our findings suggest the value of ex vivo treatment of CLL cell samples before patient’s drug administration to choose the most effective manner of therapy.

In conclusion, special attention should be paid to ex vivo data on CLL cell viability, and some apoptotic events in drug-treated cells together with molecular status of apoptosis-related genes. Additionally, the promising results demonstrate the advantage of differential scanning calorimetry in the evaluation of CLL therapy and reveal that its application could facilitate the choice of effective treatment for individual patients. The validation of presented data by examination of larger groups of CLL patients is needed.

Materials and Methods

The peripheral blood samples from two previously untreated progressive CLL patients with leukocytosis 174 and 120 × 10⁹/L, respectively, were tested. The diagnosis of the disease and clinical staging determination were established according to standard clinical, immunological and cytological IWCLL criteria.²⁶ The study was approved by the Local Ethics Committee of the Medical University of Lodz (No. RNN/143/10/KE) and the patients signed a declaration of consent.

The patients were of 65 and 54 y old with diagnosed IB and IIB CLL stage, respectively. Table 1 summarizes clinical data for both patients, who indicate no p53 mutation and opposite IgHV mutational status.

Blood samples of both patients were collected prior to therapy (experiments ex vivo). The next day, both patients underwent therapy with RCC.²⁷ Clinical response to therapy after six cycles of RCC administration was estimated by NCI Sponsored Working Group criteria.²⁶

The cytogenetic analysis of four of the most frequently detected in CLL aberrations, i.e., del(13)(q14), trisomy 12, del(11)(q22) and del(17)(p13) by fluorescent in situ hybridization (FISH) was evaluated.^28,29 PCR amplification of IGHV-IGHD-IGHJ rearrangements and immunoglobulin sequence analysis were performed as previously described.³⁰ Sequence data were made using the international ImMunoGeneTics (IMGT) information system database and tools: the sequence with a nucleotide identity of 100% as compared with the germline was considered unmutated, while the sequence with an identity of 97,57% as mutated.³¹

Mononuclear cells from peripheral blood of patients, before RCC administration, were cultured at final density of 3 × 10⁶ cells/ml in RPMI-1640 medium and incubated without (control) or with combination of cladribine (C), mafosfamide (M) and rituximab (R) for 48 h, at concentrations described previously.^18,32 Determination of living cell numbers was performed by flow cytometry using Vybrant Apoptosis Assay #4 (Molecular Probes, Inc.). Preparation of whole-cell lysates and nuclear fraction was conducted as previously reported.¹⁸ The control and drug-treated CLL cells were lysed. DSC analysis of the nuclear fraction isolated from leukemic PBMCs after incubation for 48 h with or without anticancer agents was made as described previously.¹⁸ Nuclear fraction samples were transferred into steel pans. Calorimetric studies were performed on a Seatram TG-DSC 111 calorimeter from 20– 120°C at a scanning rate - R (K/s) of 5 K/min. The analyzed parameter heat flow - HF (J/s) was used to excess heat capacity - Cp^ex estimation, according to formula: Cp^ex = HF/R × m; where m is the DNA content (g) in the probe.

Protein expression (Mcl-1, Bcl-2, procaspase-3 and Noxa) were analyzed using western blot technique as reported previously.^18,33

Total RNA was extracted from cells using RiboPure Blood Kit (Ambion) according to the manufacturer’s instructions. cDNA was synthesized from total RNA using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) by priming with random hexamers (0.5 μg of RNA was transcribed to cDNA). Gene expression profiling was carried on with the 384 TaqMan^® Low Density Array (Human Apoptosis Panel) (Applied Biosystems) for PCR analysis of 92 gene transcripts (89 examined, three housekeeping genes). cDNA samples were subjected to Real-Time Quantitative PCR (RQ-PCR) in duplicate in the TaqMan^® 7900HT Sequence Detection System (Applied Biosystems). The relative expression of each gene was quantified by the comparative cycle threshold (Ct) method (ΔΔCt, DDCT), using 18S as an endogenous control.³⁴ The difference in expression of genes was considered as the fold change values, defined as the differential gene expression before and after treatment. The complete microarray data are available at Gene Expression Omnibus (accession number GSE33925, samples 17–20). The differences in expression of examined genes between both patients were observed for eight of 89 genes, i.e., NOXA, PUMA, APAF1, ESRRBL1, CASP3, BCL2, BCL2A1 and MCL1.

The hierarchical clustering algorithm used was obtained in accordance with the average-linkage clustering method.³⁵

Glossary

Abbreviations:

CLL

chronic lymphocytic leukemia

DSC

differential scanning calorimetry

RCC

rituximab + cladribine + cyclophosphamide

RCM

rituximab + cladribine + mafosfamide

PBMCs

peripheral blood mononuclear cells

CR

complete response/responder

NR

non-response/non-responder

GEP

gene expression profile

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.