Modelling acute myeloid leukemia (AML): What's new? A transition from the classical to the modern

Abstract

Acute myeloid leukemia (AML) is a heterogeneous malignancy affecting myeloid cells in the bone marrow (BM) but can spread giving rise to impaired hematopoiesis. AML incidence increases with age and is associated with poor prognostic outcomes. There has been a disconnect between the success of novel drug compounds observed in preclinical studies of hematological malignancy and less than exceptional therapeutic responses in clinical trials. This review aims to provide a state-of-the-art overview on the different preclinical models of AML available to expand insights into disease pathology and as preclinical screening tools. Deciphering the complex physiological and pathological processes and developing predictive preclinical models are key to understanding disease progression and fundamental in the development and testing of new effective drug treatments. Standard scaffold-free suspension models fail to recapitulate the complex environment where AML occurs. To this end, we review advances in scaffold/matrix-based 3D models and outline the most recent advances in on-chip technology. We also provide an overview of clinically relevant animal models and review the expanding use of patient-derived samples, which offer the prospect to create more "patient specific" screening tools either in the guise of 3D matrix models, microphysiological "organ-on-chip" tools or xenograft models and discuss representative examples.

Keywords: 3D model; AML, cancer; Biomaterials; Bone marrow; Matrix; Organ-on-chip; Preclinical models; Scaffold; Xenograft.

Fig. 1

Schematic representation of the femur bone and relative inner hierarchical structure. (Left) Long bones such as the femur consist of two main parts: the parts at the top and bottom, are called the proximal and distal epiphysis, respectively, and a central tubular one called diaphysis. (Center) The cross section of the femur at the diaphysis shows the different layers the bone is composed of. The outer layer is the periosteum which covers much of the bone structure and is anchored to it through perforating fibers called Sharpey’s fibers. The inner layer is the endosteum at the boundary with the marrow cavity. (Top Right) Small round tubular lamellar units termed Haversian systems or osteons run longitudinally along the bone length in long bones. The typical rigidity and mechanical strength they confer on the bone depends on the different orientation of the concentric lamellae. The cavity in the osteons known otherwise as the Haversian canal, hosts nerve and blood vessels for nutrient supply. (Bottom Right) Bones hosts two types of bone marrow: the ‘red’ and the ‘yellow’ marrow which are differently specialized. The ‘red’ or ‘hematopoietic’ marrow is where hematopoiesis occurs and surrounds externally the ‘yellow’ marrow which fills the hollow part of the marrow cavity. The ‘yellow’ fatty marrow is rich in adipocytes but also contains MSCs. Created with BioRender.com

Fig. 2

Revised model for human HSC hierarchy. In the classic model for the human HSC hierarchy long-term hematopoietic stem cell (LT-HSCs) sit at the top of the hierarchy and differentiate into multipotent progenitors (MPP). Downstream of MPP, separation into common myeloid progenitors (CMP) and common lymphoid progenitors (CLP) occurs. CMP can generate granulocyte-monocyte progenitors (GMP) and megakaryocyte-erythrocyte progenitors (MEP), while lymphoid progenitors form T, B, NK and dendritic cells. Further GMP differentiate into granulocytes and monocytes and MEPs generate megakaryocytes and erythrocytes. In a revised model, HSCs can differentiate directly into MEP by bypassing CMP (here represented as MEP bypass route). Redrawn and modified from Tajer et al. [135] under the terms of the CC BY 4.0 creative commons licence http://creativecommons.org/licenses/by/4.0/

Fig. 3

Schematic illustration of the cells involved in the maintenance of the red BM niche with simplified representation of the interplay occurring in healthy and leukemic states. HSCs reside mainly within the BM and frequently localize adjacent to blood vessels. In healthy marrow (left), some of the HSCs are in a quiescent state, some others differentiate or operate self-renewal for repopulation. Many different cells either hemopoietic and non-hemopoietic in origin coexist with the HSCs in the BM niche and actively take part in their maintenance, differentiation, and self-renewal processes. Osteoblasts initially were thought to regulate HSC maintenance via CXCL-12 and SCF but recently MSCs were found to be implicated. When leukemogenesis occurs (right), the resulting effects cause the disruption of cell molecular signaling pathways in the niche. The altered BM microenvironment offers the leukemic cells protection and contributes to the development of the disease. Leukemic cells adhere to the endothelium through soluble adhesion factor E-selectin and VCAM-1. CAR cells offer protection to leukemic cells via CXCL-12/CXCR4 interaction. An inflammatory state initiates in the BM and soluble cytokines (e.g., IL-6, IL-8, IL-1β) are released. Loss of the HSC pool happens through reduction of CXCL-12 and SCF. Created with BioRender.com

Fig. 4

Schematic representation of the different approaches to study samples from patients. General examples of typical workflows depending on starting material including analysis of patient blood or marrow samples (Top). Patients’ cancer cells can be isolated from the samples and used as the starting material to undertake personalized cancer studies by establishing in vivo, ex vivo and in vitro models which include common and novel approaches like the use of organ-on-chip devices (Bottom). Created with BioRender.com

Fig. 5

Microfluidic device used for the detection of MRD in blood samples from patients. A) The blood collected from patients is analyzed in three microfluidic devices each of those differently coated with monoclonal antibodies (mAbs) to allow the detection of leukemic cells through recognition of the surface antigens CD33, CD34 and CD117 respectively. B) Images of the 50 coated sinusoidal channels in the device, in particular the inlet of a channel. The images highlight the coating with anti-CD33 mAbs, which appear false colored in red. C) Schematic representation of the mechanisms of recognition and isolation of antigen presenting cells. CD33 + cells are selectively retained in the channels while other blood cells flow freely through the device without any retention. Selected cells are then immuno-stained followed by fixation and staining of the nuclei with DAPI. D) Details of the mechanisms of release of the retained cells from the surface of the channels. The cells are collected into flat-bottomed wells and imaged afterwards using semi-automated fluorescence microscopy. Reproduced from ref [309] with permission from the Royal Society of Chemistry

Fig. 6

Engineered BM-on-a-chip used to grow HSPC from host. The BM on-chip consists firstly of a PDMS device (1 mm high × 8 mm in diameter), either with one or two openings, and contains a cylindrical cavity filled with bone inducing materials DBP, BMP-2 and collagen type I gel. The eBM was implanted subcutaneously in the back of a mouse for 8 to 12 weeks. After removal, the eBM was cultured in a microfluidic chip device. A) Schematic representation of the different steps involved in the development of the BMa-on-chip from the development and manufacturing of the eBM to the cultivation of the eBM in a separate microfluidic device. B) Images showing the eBM and microfluidic device. Top (prior to implantation), PDMS device with bone-inducing materials contained in its central cavity. Center (8 weeks after implantation), newly formed white bone surrounds pink marrow. Bottom, BM-on-a-chip microfluidic device used to culture the eBM in vitro. C) Low-left, histological images, and relative high magnification views of sections of the eBM stained with H&E in the PDMS device with two openings (top) or lower opening (center). The images are taken 8 weeks after implantation in the host. Control: cross-section of BM in a normal adult mouse femur (bottom). Scale bars, 500 and 50 μm for low and high magnification views, respectively. D) 3D reconstruction of micro-CT data from eBM 8 weeks after in vivo implantation (average bone volume was 2.95 ± 0.25 mm.³; n = 3). Scale bar, 1 mm.). Reproduced with permission [312] Copyright 2014, Springer Nature