1 Introduction

Chronic lymphocytic leukemia (CLL) is the most common leukemia in adults, accounting for approximately 30 % of all leukemia cases in European and North American countries with an incidence of 3–5 cases per 100,000 [1–3]. CLL is rare under 45 years of age, and its prevalence increases with age. The median age of patients at diagnosis is 70 years and only 10–15 % of the patients are diagnosed under 50 years of age [4, 5]. CLL is characterized by a clonal expansion of mature non-functional B-cells with a high expression of CD5, CD19, CD20 and CD23 and a low expression of surface immunoglobulins IgM, IgD, and CD79a compared to normal B-cells [6–8]. These malignant B-CLL cells represent over 99 % of peripheral blood mononuclear cells (PBMCs) in CLL patients. Approximately 2–5 % of CLL patients show a T-cell phenotype, and these patients have a less favorable prognosis than patients with a B-CLL phenotype [9]. CLL develops through increased proliferation of immature lymphocytes in lymphoid organs, which results from an increased expression of anti-apoptotic BCL-2 family proteins [10]. As a result, CLL cells can survive for months (unlike normal cells, which only survive for a few days), thereby decreasing the number of normal lymphocytes and inducing immunodeficiency [11].

CLL is a heterogeneous disease that, based on its clinical course and response to treatment, can be divided into indolent and aggressive forms [12, 13]. In case of an indolent form, the disease does usually not progress to a severe form, and the patient may survive for years without treatment [14]. In case of an aggressive form, the cell number may quickly double and, as a result, the disease may be fatal within a relatively short period of time [13, 15].

Currently several biomarkers are being used as CLL prognosticators, including elevated protein levels (e.g. TCL-1, ZAP-70, CD38), elevated RNA levels (e.g. CLLU1, LPL, miRNAs), gene mutations (e.g. TP53, SF3B1, BIRC3, NOTCH1) and epigenetic changes [16, 17]. Prognostic serum markers that can be used to predict the survival and response to treatment include increased lactate dehydrogenase (LD) levels, which are associated with a poor prognosis and a likelihood of progressing to Richter’s syndrome, increased thymidine kinase (TK) levels, which are associated with aggressive disease, and unmutated immunoglobulin heavy chain variable region (IGHV) genes, which are associated with a high risk for genomic aberrations [16, 18–20]. Evaluation of the IGHV mutation status and FISH are among the most reliable molecular tools used in routine diagnostics to date to detect clinically relevant genetic aberrations, including 11q-, 13q-, 17p- and +12 [21, 22]. The detection of these aberrations can be useful for the clinical management and proper treatment of CLL patients [23]. Only patients with progressive CLL require treatment, and this treatment results in an increased survival. A number of drugs that is currently used for CLL treatment is listed in Table 1, including fludarabine and cyclophosphamide. Fludarabine is a purine analog inducing the P53-dependent apoptotic pathway [11]. In addition to these conventional chemotherapeutic drugs, monoclonal antibodies, cell cycle inhibiting drugs and cell death inducing drugs, as well as immune modulating drugs (Table 1), are increasingly being used in the treatment of CLL [26, 27]. These drugs can be used in conjunction with fludarabine, cyclophosphamide and rituximab to enhance its efficacy and, thus, to improve patient survival [11].

Table 1 Drugs used in CLL treatment, their targets and mechanisms of action
Full size table

Activation of cellular signaling pathways, such as the NOTCH1, WNT, TLR/IL-1R, BCR and JAK/STAT pathways, has frequently been observed in CLL patients. Most of these pathways are involved in cellular proliferation, differentiation and survival, and in determining cell fate. Mutations in factors participating in these pathways may cause increased cellular proliferation and survival rates, as also resistance to apoptosis due to an altered expression of e.g. downstream cell cycle regulatory proteins [31, 32]. The mode of action of each of these signaling pathways is depicted in Fig. 1.

Fig. 1
figure 1

Signaling pathways in CLL cells. a. NOTCH signaling is initiated by a series of proteolytic cleavages that lead to the release of the intracellular domain (IC-NOTCH) from the membrane to the nucleus. In the nucleus, IC-NOTCH recruits MAML, RBPJ and EP300, forming a complex that drives the transcription of its target genes. b. Binding of Wnt to Frizzled and LPR5/6 results in phosphorylation of β-catenin, release from its multi-protein complex and translocation to the nucleus, where it forms a protein complex that drives the transcription of target genes such as c-MYC. c. Increased activity of the BCR signaling pathway leads to activation of MAPK and NF-κB pathways followed by phosphorylation of the PI3K and Src kinase and subsequent transcription of the anti-apoptotic BCL-2 and MCL-2 protein coding genes, which results in inhibition of apoptosis and increased proliferation and survival of CLL cells. d. After activation of TLR/IL-1R, the MYD88 adapter protein becomes phosphorylated, leading to activation of downstream kinases and proteins, including TRAF6 and NF-κB. Increased NF-κB activation leads to inhibition of apoptosis and increased CLL cell survival. e. The JAK/STAT pathway is associated with cytokines, and its persistent activity results in STAT3 phosphorylation and an increased transcription of genes encoding cell cycle proteins, including Cyclin D1 and Cyclin E, which leads to increased CLL cell proliferation. BCR B-cell receptor, TLR/IL-1R toll-like receptor/interleukin-1 receptor, JAK/STAT Janus kinases/signal transducers and activators of transcription, MAPK mitogen-activated protein kinase, NF-κB nuclear factor κB, MCL-2 myeloid cell leukemia-2

Full size image

Inhibitory agents (or activators) can be used to revert the apoptotic imbalance. For example, ABT 737 (a BCL-2 antagonist) can induce apoptosis in B-CLL cells, whereas DHMEQ (a NF-κB inhibitor) can induce expression changes in other genes involved in this pathway, including c-IAP, BFL-1, BCL-XL and c-FLIP, resulting in an increased therapeutic efficacy of fludarabine [33]. Nutlin mimics the molecular structure of P53 and inhibits the binding of MDM2 to it, thereby inducing apoptosis in CLL cells and eliciting synergistic effects with genotoxic drugs [34, 35]. Cell cycle inhibitors such as rapamycin and roscovitine can also induce apoptosis in CLL cells [12]. In spite of all this knowledge gathered, currently available treatment strategies have remained unsatisfactory, and further information on the pathogenesis of CLL is needed to increase the options.

By using advanced molecular (cyto)genetic techniques, our knowledge on the molecular events underlying CLL development has improved. While cytogenetic banding analysis allows the identification of less than 50 % of the (relevant) genetic abnormalities, fluorescence in situ hybridization (FISH) allows the detection of ~80 % of these abnormalities [36, 37]. In addition, comparative genomic hybridization (CGH) and single nucleotide polymorphism (SNP) microarray-based CGH (array CGH) analyses have revealed additional abnormalities. An advantage of the latter techniques is that they do not require proliferating cells [11, 38, 39]. More recently, also next-generation sequencing (NGS) techniques have contributed to a better knowledge of the genetic abnormalities present in CLL cells, including their heterogeneity [40, 41]. As yet, however, these latter abnormalities cannot be used to reliably predict disease progression [42].

In this review we focus on recent (epi)genetic findings in CLL, including microenvironmental factors, and on the evaluation of their role in disease etiology and progression, patient survival and response to treatment, including recently developed therapeutic options and their efficacies.

2 Microenvironment and drug resistance

The bone marrow (BM) entails niches that provide specific physiological environments for hematopoietic stem cells (HSCs) and other non-hematopoietic stem cells, such as mesenchymal stem cells (MSCs), that regulate their survival, maintenance and proliferation [43]. CLL cells in the BM interact with different types of cells, such as MSCs, monocyte-derived nurse-like cells (NLCs) and T-cells, collectively referred to as the “microenvironment” [44, 45]. Many factors produced by the CLL microenvironment are involved in its homing and trafficking, and it has been well-established that the microenvironment plays an important role in the pathogenesis of CLL [46–48].

B-cell surface receptor (BCR) stimulating cytokines, chemokines and adhesion molecules produced by the microenvironment in BM, lymph nodes and spleen play an important role in the accumulation, growth, survival and drug resistance of CLL cells [49]. CXCL12 and CXCL13 are chemokines that are constantly produced by MSCs and NLCs, and they can attract CLL cells that express high levels of its related receptors CXCR4 and CXCR5 and, by doing so, they can regulate the implantation and survival of leukemic cells in various tissues (Fig. 2) [50, 51].

Fig. 2
figure 2

Increased expression of the CXCR4 and CXCR5 receptors in CLL cells enhances the uptake of the chemokines CXCL12 and CXCL13 produced by MSCs and NLCs, thereby regulating the implantation and survival of leukemic cells in various tissues. Also, hyper-secretion of CCL3/CCL4 by CLL cells as a result of stimulation by CD40-L and BCR can induce the trafficking and implantation of accessory cells in the BM. The expression of BAFF, APRIL and CD31 proteins on NLC cells causes induction of anti-apoptotic pathways and survival of CLL cells. MSCs Mesenchymal stem cells, NLCs nurse-like cells, BAFF B-cell activating factor of the tumor necrosis factor (TNF) family, APRIL a proliferation-inducing ligand

Full size image

CXCR4 (CD184), which is highly expressed on the surface of CLL cells, is regulated by its ligand CXCL12 (SDF1), thereby inducing the chemotaxis and migration of these cells, as well as their resistance to apoptosis-inducing drugs [43]. Also, a pro-survival effect of CXCL12 on CLL cells has been noted [52, 53]. Increased levels of CXCL12 have significantly been associated with the mobilization of HSCs into the peripheral bloodstream [6], and proliferating Ki67-positive CLL cells have been found to express higher CXCR4 and CXCR5 levels than resting cells [54]. CXCR5 (CD185) acts as a receptor of the chemokine CXCL13, thereby regulating the homing of lymphocytes and their orientation in lymphatic follicles, and is continuously secreted by stromal cells in follicular lymphoid B-cell regions [55]. CLL cells express high CXCR5 levels, thereby causing further stimulation of these cells and a long term activation of the MAPK (ERK1/2) pathway [44]. It has been shown that CXCR5 plays an important role in the orientation and interaction between malignant B-cells and CXCL13-secreting stromal cells in lymphoid tissues [44]. Activation of the BCR signaling pathway causes clonal proliferation of normal and malignant B-cells. In response to BCR activation, CLL cells release the cytokines CCL3 and CCL4 (also called MIP-1α and β), which are likely to be involved in the recruitment of accessory cells like T-reg cells (Fig. 2) [50]. In addition to production of the chemokines CXCL12 and CXCL13, NLC cells express the B-cell activating factor of TNF family (BAFF), proliferation-inducing ligand (APRIL), CD31 and plexin-B1 proteins, which protect CLL cells against apoptosis and activate cell survival pathways (Fig. 2) [56]. CCL3 and CCL4 act as chemical attractants for monocytes and lymphocytes, and their expression is stimulated by BCR and CD40 ligands and suppressed by BCL6 [56, 57]. CLL patients exhibit higher levels of CCL3/CCL4 in their plasma compared to normal individuals, and the CCL3 plasma level has been associated with prognostic factors and treatment duration [24]. Increased secretion of CCL3/CCL4 by CLL cells can induce the trafficking and implantation of accessory cells, especially T-cells and monocytes, in the microenvironment [58]. CCL3 is secreted by CLL cells following activation of the BCR pathway. Inhibition by SYK inhibitors and PI3Kδ inhibits CCL3 secretion. It has been shown that the level of CCL3 in plasma of patients under treatment with these inhibitors reverts to normal after approximately 28 days [24].

ZAP-70 protein expression in CLL patients causes increased responses to the chemokines CCL19, CCL21 and CXCL12 which, in turn, results in an increased migration of CLL cells and an activation of pro-survival signals [24, 59]. The integrin VLA-4 (CD49d) is mainly expressed on the surface of hematopoietic cells, i.e., lymphocytes and monocytes, and plays a role in cell trafficking and implantation. VLA-4 also plays an important role in the adhesion of CLL and other leukemic cells to stromal cells and the extracellular matrix (ECM), and its increased expression in CLL cells has been shown to cause resistance to fludarabine [44, 45]. Thus, changes in expression of various receptors on CLL cells and differences in their affinities can stimulate the activity of relevant signaling pathways, resulting in increases in the survival, implantation and trafficking of these cells. In addition, the altered expression of cell adhesion molecules and the altered proliferation rate of CLL cells enables them to infiltrate into lymph nodes and endows them with the ability to develop lymphadenopathy. Awareness of these alterations may be of relevance for the treatment of patients [51].

3 Cytogenetic abnormalities as diagnostic and prognostic biomarkers

Several numerical and structural chromosomal aberrations in hematologic malignancies can be detected using karyotyping [60]. This cytogenetic technique requires dividing cells, has a relatively low resolution (20–30 Mb) and hampers the identification of small structural changes [60, 61]. A markedly increased sensitivity of this technique was reached by the application of high-resolution banding, through which the chromosomes are evaluated in the prometaphase state (3–5 Mb). This has enabled the identification and demarcation of various structural aberrations such as duplications, deletions and inversions [62]. Nevertheless, karyotyping of CLL cells only allows the detection of 40–69 % of all clinically relevant chromosomal abnormalities present, such as del(11q23), trisomy 12, del(13q14) and del(17p13), mainly due to the low in vitro proliferative capacity of CLL cells, even in the presence of B-cell mitogens [36, 37]. Therefore, high-resolution molecular (cyto)genetic techniques have been developed to more sensitively detect chromosomal aberrations, including multiplex ligation-dependent probe amplification (MLPA) [63] and SNP microarray-based comparative genomic hybridization (array-CGH) [60]. The latter technique allows the detection of small deletions or duplications as well as gross chromosomal imbalances, and has rapidly become a standard method for the evaluation of cytogenetic abnormalities in CLL [64, 65]. A major advantage of this technique over karyotyping is the rapid identification of any chromosomal loss or gain without requiring dividing cells [66, 67]. Since chromosomal abnormalities are encountered in ~80 % of CLL cases, the use of the above mentioned novel techniques can be applied to the identification of such abnormalities. Below, the most common chromosomal abnormalities and their characteristics are discussed. In Table 2 the most common chromosomal abnormalities are listed.

Table 2 Most frequent clinically relevant chromosomal aberrations in CLL
Full size table

13q-

The 13q14 deletion is the most common chromosomal abnormality and is found in 40–60 % of CLL cases [75, 79]. This deletion can be either heterozygous [mono-allelic (76 %)] or homozygous [bi-allelic (24 %)]. It has been shown that heterozygous deletions mostly occur in the early stages of the disease, and that homozygous deletions occur in the more advanced stages. Detailed analysis of the 13q14 deletion by SNP arrays has revealed the occurrence of two deletion types: type I, which encompasses the miR15a/16 locus but not the RB1 locus, and type II, which also encompasses the RB1 locus and is associated with a more aggressive disease type. Large deletions are often homozygous, whereas small deletions are usually heterozygous [12, 80]. CLL patients with del(13q) show a better prognosis and a longer overall survival than CLL patients with del(11q23) or del(17p) (see below;[69]). Also the percentage of CLL cells with del(13q) is associated with survival, i.e., a high percentage (> 80 %) of del(13q) cells results in a shorter survival compared to patients with a lower percentage (<80 %) of these cells [81–83]. In addition, it has been shown that patients with a higher percentage of del(13q) cells have a higher lymphocyte count and tend to exhibit a diffuse pattern of BM infiltration and splenomegaly. In addition, increased expression of genes involved in the MAPK signaling pathway and decreased expression of cell cycle arrest-associated genes (including CDKN2C, GAS2L1, GAS1, GAS7, ZAK,GAS2L3) has been observed in such patients, which may lead to increased cell proliferation. In contrast, decreased expression of apoptosis-associated genes (such as CASP6, CLU, DAPK1 and E2F1) has been found to result in a reduced apoptotic activity and an increased survival of CLL cells [81].

11q-

The 11q23 deletion is seen in 10–20 % of CLL cases, and most patients with this deletion are relatively young [73]. The deletion is associated with lymphadenopathy, a rapid disease progression, a poor response to treatment and a short overall survival time [84, 85]. The 11q23 deletion encompasses several genes, including the ATM tumor suppressor gene (position 11q23.3-23.1). The various functions of this gene include cell cycle checkpoint control, double-strand DNA break repair, resistance to oxidative stresses, regulation of the P53 protein, as well as protection of telomere regions. The role(s) of other genes located in the 11q23 region remains to be resolved [9, 86, 87].

17p-

This chromosomal deletion is usually associated with a high aggressiveness and a lack of response to therapy. Del(17p) is observed in 3–8 % of CLL cases, and is often associated with deletion of the TP53 tumor suppressor gene [25, 88]. TP53 plays an essential role in the induction of apoptosis and cell cycle arrest following DNA damage. Patients with 17p deletions or TP53 gene mutations are resistant to purine analogs. Standard treatment with fludarabine, cyclophosphamide and/or rituximab is, therefore, not very useful for this group of patients. In contrast, however, alemtuzumab is very effective in these patients, especially in combination with methylprednisolone [12, 89, 90]. Nevertheless, the overall survival of these patients is lower than those with less unfavorable genetic aberrations, and allogeneic stem cell transplantation is recommended for this group of patients in case of a good physical condition [12, 91].

Trisomy 12

This anomaly is encountered in 10–20 % of CLL cases, and is associated with a typical morphology and a typical immunophenotype [73, 92]. Small duplications have also been reported encompassing the 12q15 region, which harbors the MDM2 gene locus. Duplication of 12q22, encompassing the CLLU1 gene locus, serves as a prognostic risk factor in patients younger than 70 years of age [93, 94]. In addition, impaired Hedgehog (HH) signaling has often been found to be associated with trisomy 12. Increased activity of the HH signaling pathway in CLL patients, which is associated with factors such as SUFU, GLI1 and GLI3, has been found to be the leading cause of CLL disease progression [95, 96]. Treatment of these patients with fludarabine and cyclophosphamide, along with rituximab (anti CD-20), is thought to be effective due to a high level of CD20 expression in CLL cells [12, 26].

6q-

Deletion of the long arm of chromosome 6 is a relatively rare aberration occurring in approximately 6 % of CLL cases, and is associated with marked lymphocytosis with an abnormal morphology, splenomegaly, over-expression of CD38 and an unmutated state of the immunoglobulin heavy chain variable region (UM-IGHV) [68, 69].

Other rare chromosomal aberrations in CLL include 14q32 translocations (7 %), trisomy 3q27 (3 %), trisomy 8q24 (5 %), trisomy 18 and trisomy 19 [73].

4 Gene mutations as diagnostic and prognostic biomarkers

Currently, Sanger sequencing is considered as the first generation sequencing technology, whereas more recently developed advanced next generation sequencing (NGS) techniques are considered as second or third generation techniques [97] allowing, among others, comprehensive de novo sequencing, re-sequencing and SNP profiling. NGS has been used for the identification of structural changes, copy number variations, polymorphisms, point mutations and genomic anomalies previously undetectable by Sanger sequencing [40, 98]. Also, mRNA and micro-RNA expression profiling are applications of these latter NGS techniques [98–100]. In the recent past, NGS has also be used for a better delineation of complex DNA mutations in CLL [41, 42, 101].

Mutations in specific genes have been reported to be associated with the etiology, progression and treatment of CLL. Some of these mutations are known to affect the activity of relevant cellular signaling pathways. The most frequent mutations and their roles are listed in Table 3. The B-cell surface receptor (BCR) encompasses the monomeric immunoglobulins IgM and IgD. The variable regions of these receptor molecules recognize foreign antigens, and Igα and Igβ are responsible for their intracellular signaling through phosphorylation of Src family kinases and subsequent activation of downstream effectors. Activation of the BCR signal transduction pathway is involved in the regulation of metabolism, survival, proliferation, differentiation and migration of various cell types, including CLL cells (Fig. 1c) [47, 108]. Therefore, the BCR signaling pathway may serve as a therapeutic target in CLL patients. Ibrutinib (a Bruton tyrosine kinase inhibitor), GS-1101, idelalisib (a PI3K-δ inhibitor) and fostamatinib (a spleen tyrosine kinase inhibitor) are among the inhibitors of this signaling pathway. They prevent its activation via inhibiting Src family kinases and, by doing so, result in apoptosis induction [16, 109]. Approximately 60–65 % of CLL cases show mutations in the immunoglobulin heavy chain variable region (IGHV), which can alter the affinity of BCR for antigens. The remaining 35–40 % of CLL cases lack IGHV mutations [25, 102]. IGHV gene mutations are markers for a favorable prognosis, i.e., unmutated CLL (U-CLL) is clinically more aggressive than mutated CLL (M-CLL) [110, 111]. The overall survival rate for U-CLL patients is 8–9 years, whereas that for M-CLL patients is more than 24 years. The unmutated cases are associated with the expression of CD38 and ZAP-70, which are markers for a poor prognosis [112, 113].

Table 3 Most frequent somatic mutations in CLL
Full size table

TP53 mutations

DNA damage induces the activation of a variety of protein kinases, such as ATM, ATR, Chk1, Chk2 and DNA-dependent protein kinases. These kinases can form complexes and, subsequently, phosphorylate P53, thereby releasing it from MDM2 (a negative regulator of P53) causing cell cycle arrest and activation of the apoptotic pathway [114]. TP53 gene mutations have been observed in 4–12 % of untreated CLL cases, and in most cases (~70 %) with 17p deletions [88, 102]. The presence of > 20 % cells with a TP53 deletion or mutation is associated with a poor prognosis, whereas the prognosis of patients with < 20 % of such cells is similar to the overall prognosis. It has amply been shown that the presence of TP53 mutations/deletions is associated with disease progression, a poor response to chemotherapy and a short survival of CLL patients [115, 116].

ATM mutations

The ATM gene is a member of phosphatidylinositol-3 kinase (PI3K) gene family and encodes a nuclear serine/threonine kinase. ATM expression increases following double strand DNA damage, causing arrest in the G1/S and G2/M phases of the cell cycle and activation of DNA repair pathways, as well as apoptosis induction in case of lack of suitable DNA repair [117]. ATM gene mutations have been reported in 12 % of CLL patients and in about 30 % of CLL patients with a 11q23 deletion [118–120]. Lack of ATM protein expression causes loss of P53 function, chronic severe oxidative stress, rapid telomere shortening, loss of cell cycle checkpoint control, defective DNA repair and, thereby, disease progression [121]. Individuals carrying mutations in this gene have a 70-fold increased risk to develop a lymphoid malignancy compared to individuals without such mutations. The two most common ATM mutations include p.R2691C and p.P2699S, occurring in the ATP-binding site of the ATM protein and causing defects in its kinase activity [28, 122]. Bi-allelic mutations or deletions of the ATM gene following primary treatment with alkylating agents or treatment with purine analogs have been associated with a poorer prognosis and response to treatment compared to mono-allelic mutations or deletions [123].

NOTCH1 mutations

NOTCH1 is a trans-membrane protein encoded by the NOTCH1 gene and regulates the expression of several downstream target genes, such as c-MYC and TP53, and genes in the NF-κB pathway such as SCL, GATA2 and RUNX1, which are required for normal hematopoiesis [124, 125]. NOTCH1 is involved in determining cell fate, growth, differentiation and self-renewal, as well as the maintenance of stem cells and apoptosis in many tissues during normal embryonic and postnatal development [126]. Activating NOTCH1 mutations in CLL cells lead to persistent expression of the NOTCH1 protein and its binding to the Jagged-1 (JAG1) ligand, resulting in an increased activity of this pathway. The NOTCH signaling pathway has been shown to play a role in the survival and resistance to apoptosis in CLL cells [127, 128]. NOTCH1 mutations in aggressive clinical phases of CLL have been associated with a poor survival, resistance to treatment, disease progression and an increased risk of progression to Richter’s syndrome [21, 41]. NOTCH1 activating mutations have been observed in 10 % of B-CLL cases and in 60 % of T-ALL cases [25, 129]. NOTCH1 mutations are associated with U-CLL and trisomy 12, and are associated with a poor prognosis [130, 131]. NOTCH1 mutations often represent frame shift mutations causing an increased stability and defective degradation of the NOTCH1 protein, as well as accumulation of an active isoform of NOTCH1 by creating an early stop codon in the C-terminal region of the second functional PEST domain [132]. C.7544-7545del CT is the most common NOTCH1 mutation occurring in ~ 80 % of the CLL cases [28, 102].

SF3B1mutations

The mRNA splicing process is catalyzed by spliceosomes, which encompass a series of small ribonucleoproteins (snRNPs), including U2. Splicing factor 3B1 (SF3B1) is an important component of U2-snRNP [102, 133]. Inactivating mutations in the SF3B1 gene, causing defective splicing of various mRNAs followed by defective protein synthesis, have been reported in 5–10 % of the CLL cases. These mutations often represent missense mutations occurring in hot spots in codons 662, 666, 700, 704 encoding the second HEAT domain [101, 133]. These mutations mostly occur in the final stages of CLL development, and are associated with ZAP-70 protein expression as well as 11q23 deletions and unmutated IGHV genes, along with adverse clinical symptoms and a poor prognosis [42, 101].

BIRC3 mutations

The BIRC3 protein participates in a complex together with the TRAF2 and TRAF3 proteins and acts as a negative regulator of MAP3K14 (an abnormal activator of the NF-κB signaling pathway). Inactivating BIRC3 mutations cause over-activation of the MAP3K14 pathway and increased NF-κB levels, resulting in enhanced proliferation and resistance to apoptosis [25, 105]. Most BIRC3 mutations in CLL occur in the C-terminal region of the second RING domain, which is essential for the proteosomal degradation of MAP3K14 by BIRC3 [102]. Overall, BIRC3 mutations are rare in CLL patients (~4 %), but BIRC3 inactivation is commonly seen in relapsed and fludarabine-refractory CLLs (~25 % of the cases) [28, 104].

MYD88 mutations

The interleukin-1 receptor/Toll like receptor (IL-1R/TLR) signaling pathway is involved in primary immune responses and homeostasis of B-cells, and its activation causes an increased production of pro-inflammatory cytokines [134, 135]. Recently, a correlation has been found between increased expression of TLR1, TLR6, TLR7, TLR9 and an increased expression of downstream effectors, including NF-κB, c-Jun and MAPK, in CLL patients [134, 136]. MYD88 serves as an adapter protein in the IL-1R/TLR signaling pathway [137]. It becomes phosphorylated upon activation of TLR and IL-1R and causes activation of NF-κB after activating kinases and downstream proteins such as TRAF6 (Fig. 1c) [103, 106]. MYD88 activating mutations have been shown to increase the activity of this pathway in CLL patients and to cause resistance to apoptosis [107]. MYD88 mutations are found in 3–5 % of CLL cases, as also in other B-cell malignancies such as diffuse large B-cell and marginal zone lymphomas [100, 138].

Other mutations reported in CLL include mutation in the KRAS, SMARCA2, NFKBIE and PRDK3 genes [12].

5 Epigenetic alterations as diagnostic and prognostic biomarkers

Epigenetic changes do not alter the DNA sequence itself but, instead alter DNA methylation and histone methylation or acetylation patterns. Several studies have highlighted the importance of these processes in CLL and other malignancies [139–141]. Activation of proto-oncogenes through hypo-methylation of DNA promoter sequences is a relatively common change observed in CLL. Conversely, promoter hyper-methylation has been found to play a role in the silencing of tumor suppressor genes [142], i.e., hyper-methylation of the P16 INK4A and P15INK4B gene promoters has been reported in a sub-group of CLL patients [143–145]. A high global level of DNA methylation in IGHV-unmutated CLL patients has been associated with a poor prognosis and patients with a high methylation index (MI) need to be treated, whereas those with a low MI require a ‘wait and see’ policy [28]. Hypo-methylation of the BCL-2 gene promoter in CLL patients results in its over-expression [146] and hypo-methylation of MDR1 and TCL1 gene promoters results in NF-κB activation, a phenomenon that is frequently seen in CLL patients [78]. Hyper-methylation of the TWIST-2 gene promoter is an important factor that is associated with TP53 gene silencing and has been reported to occur in CLL patients [145, 147]. It has also been shown that in patients with IGHV mutations, methylation of the CD38 and BTG4 gene promoters is associated with a favorable prognosis, whereas methylation of the HOXA4 gene promoter in these patients is associated with a poor prognosis. The expression level of ZAP-70 is related with methylation of specific CPG islands in the promoter of this latter gene [148–150].

It has also been shown that epigenetic mechanisms are involved in changes in expression of miRNAs (see below) in CLL [151]. It has e.g. been shown that the miR-139 and miR-582 promoters are targets of significant methylation and, thus, silencing in CLL patients [78], whereas hypo-methylation of the miR-21and miR-29a promoters is associated with an increased expression of these miRNAs. Conversely, hyper-methylation of the miR-124-2 and miR-129-2 promoters has been associated with a decreased expression of these miRNAs [78, 152]. Hyper-methylation of the miR-9-3 promoter, a miRNA that has a tumor suppressor function, is frequently observed in CLL, causing silencing of its expression and sustained activation of the NF-κB signaling pathway, inducing resistance to apoptosis [153]. Hyper-methylation of the miR-34b/c promoter, leading to silencing, is also commonly seen in CLL [72].

Due to their reversible nature, epigenetic changes serve as prime treatment targets. The application of DNA methyl-transferase (DNMT) and histon deacetylase (HDAC) inhibitors, capable of reducing DNA methylation and histone deacetylation, may serve as a novel treatment strategy in CLL [16, 78]. The efficacy of several of these compounds has already been tested in vitro on CLL cells and are currently subject to clinical trials [154].

6 Micro-RNAs as prognostic biomarkers

Micro-RNAs are small non-coding RNA molecules causing mRNA translation inhibition and/or degradation through binding to its 3′untranslated regions (UTR) [155, 156]. Several studies have reported changes in expression of miRNAs in human malignancies [155, 157]. In CLL, specific miRNAs can be used as etiologic and prognostic (therapy-related) markers, as outlined below.

MiR-15a/16 acts as a tumor suppressor by inhibiting the expression of BCL-2 [76]. Since the miR-15a/16 locus is located in 13q14, deletion of this region in CLL (see above) is a main cause of its decreased expression. There may, however, also be other causes for this decreased expression, including regulatory TP53 gene mutations and deletions or mutations in pri-miR transcription factors [158, 159]. Decreased miR15a/16 expression occurs in a considerable proportion of CLL cases (68 %) [76, 160] and leads to increased BCL-2 expression and resistance to apoptosis, or increased expression of cell cycle-related genes, including CCND2, CCND3 and CDK4/6, which play an important role in the transition of cells from the G0/G1 to S phase [159, 161].

As a tumor suppressor, miR-29b inhibits the expression of several proto-oncogenes, including TCL1, MCL1, CDK2/6, SP-1, DNMT3a and DNMT3b [162, 163]. Reduced miR-29 expression and increased TCL1 expression in CLL patients has been associated with an aggressive phenotype, an unmutated IGHV status, over-expression of ZAP-70 and deletion of 11q23 [71, 146]. TCL1 proto-oncogene expression has been found to be increased in ~90 % of CLL patients, and is variably expressed during different stages of the disease. The TCL1 protein acts as a co-activator of the PI3K/Akt kinase, which causes apoptosis reduction by phosphorylation and inactivation of BAD (a pro-apoptotic protein). Additionally, activation of NF-κB (a pro-survival protein) by TCL1 has been found to increase the BCL-2 protein level and, as a consequence, to inhibit apoptosis in CLL cells [11, 164]. It has been found that TCL1 expression is regulated by miR-29 and miR-181, and down-regulation of these miRNAs has been shown to cause its increased expression, followed by a decrease in apoptosis [163]. Expression regulation of the MCL1 and BCL-2 family of anti-apoptotic proteins is mediated by miR-29, and reduced expression of miR-29 in CLL increases its resistance to apoptosis [165].

MiR-34a is a direct transcriptional target of the P53-pathway, and its increased expression leads to apoptosis and cell cycle arrest through the regulation of proteins such as BCL-2, cyclin D1 and E2, CDK4 and c-MYC, depending on the cell type [77, 155]. Several studies have shown that deletion or mutation of the TP53 gene can cause a decreased expression of miR-34a, defective DNA repair and apoptosis responses, and fludarabine resistance of CLL cells [159, 160]. The expression of miR-34a has been found to be decreased in patients with 17p deletions or TP53 mutations, causing a fludarabine refractory phenotype [18, 166]. Moreover, decreased expression of miR-34a, miR-29c and miR17-5p has been observed in aggressive subtypes of CLL with TP53 abnormalities, indicating a role of target genes of these miRNAs in CLL pathogenesis [165].

As an oncogene (oncomiR), miR-181 is involved in B-cell CLL development [31], and it has been found that decreased expression of miR-181 increases the levels of the MCL1 and BCL-2 anti-apoptotic proteins [167]. Specifically, miR-181b has been shown to regulate the expression of TCL1, and its reduced expression has been proposed as an unfavorable prognostic factor for CLL and to be related to refractoriness to treatment [19]. In CLL patients with 17p deletions, increased expression of miR-21 has been associated with a poor prognosis. Additionally, in this group of patients decreased expression of miR-223, miR-29b and miR-181 family members has been found to be more prevalent in aggressive than in indolent cases. So, it has been suggested that these miRNAs can be used as markers for the identification of del(17p)-positive CLL patients with an adverse prognosis [168, 169].

Also miR-155 acts as an oncogene and its over-expression has been reported in several solid tumors, including lung, colon and breast carcinomas, as well as in hematological malignancies, including AML and CLL [70, 170]. It has been reported that miR-155, and other miRNAs such as miR-182 and miR-96, regulate several components of the PI3K signaling pathway [171]. Additional studies have shown that miR-155, and also miR-181a and miR-223, are involved in the normal development and differentiation of lymphoid and myeloid cells, and that their altered expression is associated with the development of its respective corresponding leukemias, i.e., increased expression of miR-155 and miR-181a causes enhanced survival and proliferation in both hematopoietic lineages [172].

MiR-221/222 expression has been found to be increased in many solid tumors, including lung, thyroid and hepatocellular carcinomas. The primary target of miR-221/222 is P27, which functions as a tumor suppressor by inhibiting cell cycle progression [173]. In CLL, increased miR-221/222 expression has been found to cause resistance to fludarabine [77]. Additional studies have indicated that miR-223 may serve as a prognostic biomarker that may be used to predict the overall survival (OS) and time to first treatment (TFT) [174, 175].

The miR-146a locus is located on chromosome 5, and its expression is increased in some inflammatory diseases and malignancies, including CLL. Several studies have indicated a relationship between miR-146a expression and tumorigenesis through inappropriate regulation of inflammatory responses [176]. Increased miR-146a expression has been associated with cell migration, motility and adhesion [168, 177], and increased transcription of the integrin encoding genes ITGA4 and ITGB2. These latter genes are involved in CLL via the inhibition of α4β1 and αLβ2 heterodimers in trans-endothelial migration (TEM) during tissue invasion. Therefore, it has been suggested that inhibitors of this miRNA may be useful as therapeutic agents in CLL [176, 177].

Although miRNAs may, as yet, not be very useful as diagnostic indicators, they can be used as biomarkers for predicting response to therapy and survival, and also to better understand the pathogenesis of CLL cells [18, 178]. Since miRNAs can function both as oncogenes and tumor suppressor genes during leukemogenesis, miRNA-based therapies have recently been explored as therapeutic options, and some progress has indeed been made in this field (Table 1). It has e.g. been shown that chemically engineered anti-miRNA oligonucleotides (AMOs or antagomirs) can specifically bind to certain miRNAs and thereby inhibit their function [178]. The use of these oligonucleotides in conjunction with traditional treatment protocols may provide novel treatment strategies for (sub-groups of) CLL patients [29, 30].

7 Conclusions and future directions

CLL is a clinically heterogeneous disease with survival rates ranging from months to decades. Also genetically CLL is heterogeneous, as exemplified by the occurrence of various gross chromosomal alterations (including 11q-, 13q-, 17p- and +12), immunoglobulin heavy chain gene mutations, mutations in several signal transduction pathway genes including NOTCH1, BIRC3, SF3B1, TP53, and alterations in miRNA expression patterns. Also serum levels of several proteins, including B2M, TK and LD, and the expression levels of several immunophenotypic markers, such as CD38 and ZAP-70, may be altered. It has been found that there is a strong relationship between specific genetic aberrations and the clinical course of CLL. Specifically, the expression of unmutated immunoglobulin heavy chain variable region genes, chromosomal abnormalities such as del(17p) and del(11q), and the expression of the ZAP-70 and CD38 proteins have been associated with a poor prognosis. Mutations in tumor suppressor genes, such as TP53 and ATM, have been associated with a more aggressive course of the disease and, in addition, to resistance to standard treatment protocols. Since CLL patients usually bear several genetic abnormalities, a combination of various treatments is often used, and its choice should be tailored to the abnormalities present. For example, patients with del(17p) and/or a TP53 mutation are resistant to fludarabine, cyclophosphamide and rituximab, whereas alemtzumab combined with methylprednisolone has been shown to be highly effective. The identification of new biomarkers will be of help for a proper patient stratification and treatment choice and, in parallel, the efficacy of new drugs should be tested in corollary with the presence of these new biomarkers (i.e., molecular lesions) in CLL patients.

In addition to genetic biomarkers, also epigenetic changes and miRNA expression alterations have been addressed as prognosticators. Although not yet routinely used as diagnostic markers, miRNAs can be applied as biomarkers to predict response to treatment and survival in patients and to better understand disease etiology. Because of the reversible nature of epigenetic changes and the dual role of miRNAs as both oncogenes and tumor suppressor genes in leukemogenesis, they are considered as proper therapeutic targets. Already, a number of them has been tested in vitro with satisfactory results and, based on this, included in clinical trials. Therefore, a further understanding of the (epi)genetic basis in CLL is expected to lead to an improved clinical management through specifically designed targeted therapies.