Advantages and disadvantages of mouse models of chronic lymphocytic leukemia in drug discovery

KEYWORDS:

• Chronic lymphocytic leukemia (CLL)
• patient-derived xenograft (PDX) models
• TCL1 mouse model
• transgenic mouse models

1. Introduction

Chronic lymphocytic leukemia (CLL) is a lymphoid malignancy characterized by the proliferation and accumulation of mature CD5⁺ B cells in the peripheral blood (PB), bone marrow (BM) and lymphoid tissues. When the number of CD5⁺ B cells is lower than 5 × 10⁸ cells/L, this entity is called monoclonal B lymphocytosis (MBL) and it is asymptomatic monoclonal or oligoclonal proliferation of B cells [1]. The disease may have a stable course but also become aggressive, with frequent relapses, or even transform into an aggressive lymphoma, typically diffuse large B-cell lymphoma (DLBCL) (Richter transformation). Two major molecular subgroups have been identified: those harboring unmutated immunoglobulin heavy-chain variable region (IGHV) genes (U-CLL, ≥98% identity with the germline) and those with mutated IGHV genes (M-CLL). Genomic and epigenomic studies have elucidated multiple aspects of the pathogenesis of the disease. Nowadays CLL should be considered as a complex disease in which genetic and epigenetic mechanisms cooperate with microenvironmental factors in the malignant transformation and in leukemia progression [2]. In parallel, new targeted therapies and management strategies have been developed. Due to this complexity and heterogenousity, the generation of adequate pre-clinical mouse model fully reflecting tumor biology has not yet been successfully achieved. Here we describe and discuss the mouse models developed for CLL, the most prevalent type of leukemia in adults in Western countries.

2. Mouse models in CLL

Several murine genetically engineered and human xenograft models for CLL have been reported. The New Zealand Black (NZB) mouse naturally develops a CLL-like disease without the induced oncogenes or deletion of regulatory regions. These mice have an MBL phenotype at an early age which always progresses to an aging-associated clonal CD5⁺ B-cell disease. Thus, this model could be useful to study the progression from MBL to CLL [3].

2.1. Genetically engineered

In 2002 was described the first transgenic model for CLL, the T-cell leukemia/lymphoma 1 transgenic (Eμ-TCL1) mouse [4] that mimic aggressive CLL. The Eμ-TCL1 is based on the exogenous expression of the human TCL1 gene under the control of the Igh variable region (V_H) promoter and the IgH intronic enhancer region (Eμ) to target TCL1 gene expression. Older mice (10–18 months) develop a clonal expansion of CD5⁺ B-cell that progress as a CLL-like clonal disease, resembling aggressive U-CLL with splenomegaly, hepatomegaly and expanded B cells in PB [5]. Disease progression in Eμ-TCL1 mice can be accelerated by adoptive transfers of splenocytes from Eμ-TCL1 murine CLL cells into wild-type recipients, being this model very useful for preclinical evaluation of novel therapies and to study the interplay of CLL with tumor microenvironment, being an important tool for studying CLL progression [6]. Recently, several groups have crossed transgenic Eμ-TCL1 mice with other genetically modified mice by modulating genes related to B-cell receptor (BCR) signaling, leukemia-environment interactions and proliferation and cell survival in order to achieve better knowledge of CLL biology [7–10]. Figure 1 describes all the Eμ-TCL1-related mouse models in CLL.

Figure 1. Eμ-TCL-1 driven leukemia mouse models as a tool for studying CLL pathogenesis. Eμ-TCL-1 mouse model has been crossed with constructs to overexpress or to delete molecules related to leukemia-environment interactions, BCR (B-cell receptor) signaling, proliferation and survival [7–10]. cd44^−/-: homing cell adhesion molecule involved in cell–cell interactions; cxcr5^−/-: C-X-C chemokine receptor type 5 involved in B-cell migration to the lymph nodes; baff-Tg: B-cell activating factor belonging to the tumor necrosis factor (TNF) family. It is a potent B-cell activator of proliferation and differentiation; mif^−/-: macrophage migration inhibitory factor is related with innate immunity and survival of tumor-associated macrophages; fzd6^−/-: Frizzled 6 is associated with Wnt signaling pathway playing a key role in development, tissue-specific stem-cell maintenance and tumorigenesis; rhoh^−/-: Ras homolog gene family, member H is a small G protein that is involved in CLL BM homing and its engraftment; dnaRag1-Tg: recombination activating genes 1 is a protein involved in V(D)J rearrangement, xid: X-linked immunodeficiency mouse; pkcβ^−/- or pkcβ^+/-: protein kinase C beta type, protein involved in BCR signaling and overexpressed in CLL patients; xbp-1^−/-: X-box binding protein 1 is a transcription factor related with the ER-stress, reduced expression of XBP1 protein compromise the BCR signaling disadvantaging the leukemic cells survival; APRIL-Tg: a proliferation-inducing ligand is a protein from the TNF family that mediates for CLL cell survival and leukemogenesis; ROR-Tg: receptor tyrosine kinase-like orphan receptor 1 is an oncoembryonic antigen found on CLL cells; tir8^−/-: also known as single Ig IL-1 related receptor (SIGIRR) is involved in the inflammation pathway and the Toll-like receptor (TLR) signaling; id4^−/-: inhibitor of DNA binding protein 4 is a member of the dominant-negative basic helix-loop-helix transcription factor family that lacks DNA binding activity and has tumor suppressor function; p53^−/: tumor protein p53; c-Myc: c-myelocytomatosis oncogene product. (BM: bone marrow; BTK: Bruton’s tyrosine kinase; LN: lymph node; TAM: tumor-associated macrophages)

Other strategies have been the generation of transgenic mice carrying some of the most important genetic alterations reported in CLL such as 13q chromosome deletion, overexpression of several survival factors including B-cell lymphoma 2 (Bcl-2), tumor necrosis factor Receptor Associated Factor 2 (TRAF2), B-cell Activating Factor (BAFF) or A Proliferation Inducing Ligand (APRIL), deletions of Interferon Regulatory Factor 4 (IRF4) and ataxia telangiectasia mutated (ATM) or splicing factor 3B1 (SF3B1) mutations (Table 1).

Table 1. Transgenic mouse models of CLL non-TCL-1 related

Gene	Rationale	Mutation	Penetrance	Mouse strain	Reference
13q chromosome deletion	Most frequent deletion in CLL	mir-15a/16-1^−/- and mir-15a/16-1^floxed	25–30% development of MBL	NZB background	[11]
		Deletion of Dleu-2 and Dleu-5 (14qC3 minimal deleted region (MDR)^−/-	40–45% development of MBL, CLL and CD5^− NHLs	129/Sv-C57BL/6 mixed background	[12]
		Deletion of common deleted region (14qC3CDR^floxed)	40–45% development of CLL	129/Sv-C57BL/6 mixed background	[13]
BCL2	Juxtaposition of BCL2 next to IGH gene	bcl2 tg mice [t(14;18)]	No B-cell expansion Prolonged survival in vitro	SWR/J x SJL/J crossover	[14]
Traf2.DN	TRAF2 inhibition induce NF-κB activation	Traf2.DN^−/- a TNF receptor	No B-cell expansion Tg mice with greater resistance to apoptosis	C57BL/6 x CBA/J crossover	[15]
BCL2 x Traf.2DN	Similar to Traf2.DN with more splenomegaly	Overexpression of BCL2 and Traf2.DN	>6 months splenomegaly and 80% of mice die at 14 months	BALB/c x FVB/N crossover	[16]
APRIL-Tg	High levels in serum, associated with proliferation and survival	Overexpression of APRIL	Between 9 to 12 months mice developed mild lymphoproliferation of CD5^+ B1 cells in the peritoneal cavity	C57BL/6 background	[17]
Eµ-mir-29-Tg	miR-29a/b is overexpressed in CLL	Overexpression of APRIL and microRNA cluster miR-29a/b	At 24 to 26 months ~20% of mice developed leukemia and died. Aggressive CLL phenotype	C57BL/6 background	[18]
irf4^±	Regulator of B-cell development and function	Heterozygous mutation of IRF4	Accelerated disease	NZB background	[19]
Vh11 x irf4^−/-	IGHV11 knock-in mice lead to anti-PtC (autoantigen) development.	Crossed Irf4-deficient mouse (irf4^−/-) with IGHV11 knock-in mouse	Spontaneously developed CLL (5–10 months) at 100% penetrance. Aggressive phenotype	irf4^−/- mouse	[20]
ROR1-Tg	ROR1 is overexpressed in CLL cells	Overexpression of ROR1 with the mouse IgH promoter/ enhancer	After 15 months (3% to 65% mice) developed lymphocytosis and splenomegaly	C57BL/6 background	[21]
Sf3b1-K700E^fl/+ x Atm^fl/fl	SF3B1 and ATM are recurrently mutated in CLL	Heterozygous expression of Sf3b1(K700E) (Sf3b1-K700E^fl/+) and Atm deletion (Atm^fl/fl)	At 9 to 12 month mice developed lymphocytosis, splenomegaly and BM infiltration of B-cells carrying SF3B1 and Atm mutations.	C57BL/6 background	[22]

2.2. Xenograft mouse models

In a xenograft model, human tumor cells or cell lines are transplanted subcutaneously (s.c.), intravenously (i.v.) or orthotopically into the organ type of origin in immunocompromised mice. Given the poor engraftment of primary CLL cells, several efforts have been made in the last decade to improve the engraftment into key organs of mice that mimic the typical spread of CLL in humans.

It has been reported that hematopoietic stem cells (HSCs) purified from patients with CLL displayed lymphoid-lineage gene priming and developed monoclonal or oligoclonal B-cell simulating MBL after xenogeneic transplantation, but other events are needed to develop a CLL [23].

The most commonly used strains are the immunodeficient NOD/SCID model, where the severe combined immunodeficiency (SCID) model has been transferred onto a diabetes-susceptible mouse, the non-obese diabetic (NOD) background. This model exhibits better engraftment of human cells due to dysfunctional NK cells, lowered cytokine production, and defective T or B cells. The other mouse strain used is the NSG (NOD scid gamma) mice. In this model, the NOD/SCID strain has been modified with a mutation in the gene encoding the interleukin-2 receptor common gamma chain (IL2Rγ^−/-). This model lacks mature B, T and NK cells, to avoid immune rejection of human cells [24].

The first successful studies have been achieved in 2007 [25] using the NOD/SCID mice, where after 4 weeks, an engraftment and growth of cells was achieved in spleen and peritoneal cavities (~90% to 95% of efficiency), BM (~54% of efficiency), and PB (~36% of efficiency) [25]. Later, in 2011, using NSG mice, they demonstrated that the successful xenograft of CLL cells was dependent on autologous T cells [26]. This model was refined by the addition of autologous T cells that have been activated previously in vitro [6,27]. Other modifications such as transfection of CLL cells with miR-15a/16-1 microRNAs (miRNAs) have also been tested [28]. Even though there are a few human CLL cell lines available, most of them have also been useful for xenografting into immunodeficient mice [24,29]. A summary of all these models is presented in Figure 2.

Figure 2. Xenograft models in CLL. The most relevant xenograft strategies in CLL are showed. The findings/advantages or disadvantages of each strategy are highlighted in green or red, respectively. Immunodeficiency of each mouse used is labeled in blue and italics. Models developed in a NSG mice using PBMCs or purified CD19⁺ cells, has allowed to discover that (i) splenic-hCD45⁺ cells engraftment correlated with CLL prognosis markers (i.e. U-IGHV genes) [24]; (ii) the murine splenic microenvironment sustained CLL cell proliferation like human lymph nodes, with induction of NF-κB and BCR signaling in the xenografted cells, which served as proof-of-concept to test the efficacy of ibrutinib on tumor–host interactions [30]; (iii) the restoration of miR-15a or miR-16-1 impairs the capacity of CD19⁺ CLL cells with biallelic 13q14 deletion for engraftment and growth in the spleen of NSG mice, although without synergy between both miRNAs and no effects in CLL clones harboring TP53 alterations (TP53 mut) [28], evidencing that the CLL cell proliferation capacity into murine splenic microenvironment also depends of non-genetic factors; (iv) co-xenotransplant of CLL PBMCs with allogenic APC (CD14⁺ or CD19⁺) promote the activation of autologous T-cells, which is crucial for CLL cells to engraft in secondary lymphoid tissues, survive, and proliferate mimicking key features of human CLL [26]. The engraftment of CLL cell lines have been achieved using JVM-3, MEC-1 [24] and MDA-BM5 [29]. To avoid the limited availability of CLL primary samples, PDX models in CLL and Richter syndrome have been described [31–33], allowing an exponentially expand few amount of primary CLL cells to study of tumor biology, discovery of novel therapeutic targets, and preclinical screening of drugs. (APC: antigen presenting cell; BCR: B-cell receptor; CLL: chronic lymphocytic leukemia; i.p.: intraperitoneal injection; i.v.: intravenous injection; NOD: non-obese diabetic mouse; NSG: NOD scid gamma mouse; PBMCs: peripheral blood mononuclear cells; PDX: patient-derived xenograft; s.c.: subcutaneous injection; SCID: severe combined immunodeficiency)

A common handicap of primary CLL cells xenografts models is often the incompatibility between the large number of cells required and the limited availability of CLL samples. This is avoided by the establishment of patient-derived xenograft (PDX) models, able to exponentially expand a relative few starting primary CLL cells. Most CLL-PDX models have been established by the initial xenograft of primary CLL cells with activated autologous T cells into NSG mice, which has been clinically useful to distinguish between spontaneous and treatment-induced clonal selection in CLL [31–33].

3. Expert opinion

CLL is a well-defined lymphoid neoplasm with very heterogeneous biological and clinical behavior. New models that recapitulate the key features of CLL are needed to study CLL biology and to test novel drugs in CLL. Due to the facility to obtain primary CLL cells, a high number of studies to test the efficacy of new drugs have been done, and we have learnt that CLL cells die spontaneously in vitro and that they need external factors to survive. Thus, although new in vitro systems to check the efficacy of drugs have been developed (co-culture with stromal cells or cytokines in a long-term culture to induce the proliferation of CLL cells), these systems only partially recapitulate the response of CLL patients and probably new 3D models are needed.

The transgenic Eμ-TCL1 mouse model has been widely used to study pathophysiology, clonal evolution, the interplay of CLL with tumor microenvironment and drug efficacy in aggressive CLL. Unfortunately, exome sequencing of CLL samples from the Eμ-TCL1 mice showed that although mutations in mice are frequently subclonal and heterogeneous and upon transplantation a clonal evolution was observed, they lack any of the most recurrent mutations detected in human CLL cells [34]. Then, the development of new genetically engineered mouse models reflecting the different recurrent genetic lesions with clinical impact described in CLL is warranted as they will help to improve the knowledge of CLL complexity and response to new therapies. One of the disadvantages of these transgenic mice is that they cannot be used to test some new immunotherapy drugs that are not expressed in mouse.

On the other hand, the development of more immunodeficient mice and the knowledge that CLL cells are dependent on autologous T cells for engraftment in mice have improved the methodology to generate xenograft CLL models, without the bias of humanization or irradiation. These models could be considered as an autologous T humanized mice. Currently, the main problem of these models is that tumor development is highly variable and is unable to completely recapitulate the complex host stroma–tumor interactions, which depend on autocrine and paracrine interaction among cells with the adaptive and innate immune system; although it seems that murine cells are able to provide a supportive microenvironment for xenografted CLL cells. Theorically, these models cannot be used for pre-clinical testing of immunomodulatory compounds, but taking into account the kinetics of B- and T-cell engraftment in these mice, recently the efficacy of a CD19xTCR antibody has been demonstrated [35]. In addition, SCID mice exhibit defects in DNA repair that prevent the testing of some cytotoxic compounds and nude mice demonstrate complete fragility which could hamper their ability to tolerate new therapies. Thus, the establishment of PDXs able to exponentially expand a relative few starting primary cells presenting the most recurrent mutations in CLL could be of great interest. Due to the low frequency of these mutations among CLL patients, little data exist about the impact of these mutations on the pathobiology of CLL. These studies might help to study clonal evolution in CLL cells and use these expanded cells to check the efficacy of new drugs targeting important altered biological pathways (e.g. RNA splicing, Notch1, BCR and Toll-like receptors [TLRs]) in CLL alone or in combination with approved drugs, both in vivo and in vitro. The generation of new 3D models recapitulating the stroma–tumor interactions that exist in CLL may help to achieve a more effective and personalized treatment for CLL patients.

Declaration of interest

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers in this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

The authors have been supported by the Spanish Ministry of Science and Innovation [RTI2018-094584-B-I00] and was cofounded by the European Regional Development Fund (ERDF) and the CERCA program from Generalitat de Catalunya, Centro de Investigación Biomédica en Red de Cáncer (CIBERONC) [CB16/12/00334] and Generalitat de Catalunya [2017 SGR 1009]. H Playa-Albinyana is a recipient of a predoctoral fellowship from the Spanish Ministry of Science, Innovation and Universities [FPU19/03110].