Discovery of proangiogenic CD44+mesenchymal cancer stem cells in an acute myeloid leukemia patient’s bone marrow

To the Editor

Acute myeloid leukemia (AML) most commonly occurs in older adults [1] and has the lowest survival rate among all types of leukemia [2]. Mechanistically, AML is a hematopoietic cancer that involves a complex interplay among different types of bone marrow (BM) cells, cytokines, growth factors, and neural modulation—all of which contribute to the survival and growth of malignant Hematopoietic Stem Cells (HSCs) [3, 4]. Currently, most AML therapies have been developed to target malignant HSCs because of their significant role in initiating uncontrolled clonal proliferation and transforming the naïve BM niche to allow disease progression and treatment resistance [5]. However, the overall survival of relapsed or refractory AML patients has remained dismal for the past three decades [6]. Therefore, more research on the mechanisms of AML etiology and relapse are urgently needed to further the development of effective treatments.

A new concept of niche-induced oncogenesis has been proposed that is based on the observation that artificial genetic perturbation of mesenchymal stem cells (MSCs), and their progenies leads to leukemogenesis of myelodysplasia and AML in transgenic mice [7, 8]. In this study from an AML patient BM, for the first time, we discovered previously unrecognized CD90-CD13-CD44+ MSCs with tumorigenic proliferation and proangiogenic properties ex vivo.

The AML bone marrow donor is a 32-year-old female with a past medical history of diabetes and hypothyroidism. She was admitted to the emergency department with intermittent chest pain, shortness of breath, and gum bleeding, and diagnosed with AML with myelomonocytic differentiation. Peripheral blood smears showed leukocytosis with ~ 52% blasts/equivalents and marked thrombocytopenia. FISH showed normal cytogenetics. An AML panel was positive for mutations of DNMT3A, NPM1, and NRAS without FLT3-ITD.

During ex vivo isolation of MSCs from this patient BM, we observed two different cell populations. The early plastic-attached cells were CD90+CD13+CD44+ MSCs (Fig. 1(a1, a3, d1 upper panels)). The late attached cells in the non-adherent floating cells of MSC cultures became adherent after 7 or more days (Fig. 1(a2)). We named them mesenchymal cancer stem cells (MCSCs) because of their capabilities of multiple lineage differentiation (Supplementary Figure 1), rapid cell proliferation and self-renewal (Fig. 1b, c, Supplementary Figure 2), and release of high amounts of cytokines and growth factors reported to be essential for angiogenesis and blasts’ growth. At early passages (P), there were about 20% CD90-CD13-CD44+ cells within large-sized strong CD44+ MCSCs (Fig. 1d lower panels). MCSCs grew in plastic attachment patterns; however, the number of viable floating cells increased with passage (Fig. 1(A4)). After P7, most MCSCs were CD90-CD13-CD44+, expanded in floating cluster conditions (Fig. 1g, i), and generated tube-like vessel structures (Fig. 1j) by expressing endothelial cell biomarkers like VE-Cadherin/CD144 (Fig. 1k). To understand the interaction between MCSCs and AML blasts, we plated MCSCs or MSCs with GFP-MOLM-14 (AML blasts) together. After 24 h of 1:1 co-cultures, GFP-MOLM-14 were found in the middle of floating clusters and surrounded by MCSCs (Fig. 1m), in contrast to their position on top of MSCs when cultured with MSCs only (Fig. 1l). These data are consistent with a previous report that AML blasts integrate into vascular niches to progress and metastasize [9].

Finally, we performed a proteome assay of supernatants from MSC and MCSC cultures. The release of key angiogenic cytokines like VEGF and growth factors such as FGF basic and PDGF were significantly increased in the P7 MCSC culture (Fig. 2b). To evaluate therapeutic potential, we performed blocking experiments with anti-CD44 monoclonal antibodies, which stopped tumorigenic proliferation of P5 MCSCs, but not P7 MCSCs (Supplementary Figure 3). It is possible that the large amount of ICAM-1 (53-fold increase in P7 MCSCs versus P5 MCSCs, Supplementary Figure 3E) compensated the loss of CD44 as previously reported in CD44 null mice [10].

In summary, we provide novel evidence of MCSCs as the cellular origin of angiogenesis and cytokine release in an AML patient’s BM. More research with additional patients is needed to discover if this phenomenon is unique to this one AML patient or has broader ramifications for AML as a whole. Our data suggest that future therapeutic strategies should also be developed to target MCSCs and their cytokine release to achieve stringent complete remission and prevent AML progression and relapse. The existence of MCSCs might be applicable to other cancers as well.