Perturbed Endocytosis as a Mechanism of CALM-Dependent Leukemogenesis: Identification of Specific CALM Domains Required for Internalization and Demonstration of Prolonged JAK2 Signaling in Response to GM-CSF.

Background: Clathrin Assembly Lymphoid Myeloid leukemia (CALM) gene rearrangements, in which CALM is fused to MLL or AF10 genes, are found in aggressive leukemias and lymphomas. Expression of MLL-CALM or CALM-AF10 fusion proteins immortalizes murine hematopoietic cells in vitro, correlating with leukemogenesis in vivo. While disrupted MLL or AF10 activity contributes to transformation, perturbation of normal CALM function may also play a role. The native CALM protein is primarily cytoplasmic and functions in Clathrin-Dependent Endocytosis (CDE). We have previously shown that overexpression of CALM-containing fusion proteins in COS7 cells impairs CDE of both Transferrin (TF) and Epidermal Growth Factor (EGF). These initial observations were made using a qualitative (but time-consuming) visual assay of Texas Red (TR)-labeled TF or EGF. We confirmed these results using a semi-quantitative radioactive assay (with ¹³¹I-TF or EGF) that, while sensitive, lacked specificity for transfected cells. Here we describe the development of a flow cytometry-based assay of CDE that permits specific quantitation in cells transfected with CALM-containing proteins. We also measure downstream effects of perturbed CDE on growth factor signaling.

Objectives:

1. To validate a novel flow cytometry-based CDE assay;

2. To identify CALM domains that play critical roles in CDE;

3. To analyze kinetics of phosphorylation of JAK2, a downstream target of growth factor signaling, during perturbed CDE.

Methods: COS7 cells transfected with GFP-tagged CALM-containing constructs were incubated at 4°C for 1h with AlexaFluor (AF) 633-TF or AF647-EGF to allow binding, followed by incubation at 25°C (5, 10, or 15 min) to permit internalization. Following acid buffer wash and fixing (1% formalin/PBS), internalized fluorescence was measured by flow cytometry. Geometric mean fluorescence intensities of GFP+/AF+ populations were normalized to respective 4°C controls. JAK2 phosphorylation was assessed by western blotting at various time intervals following GM-CSF stimulation of six different leukemia cell lines.

Results: Overexpression of CALM-containing constructs in COS7 cells resulted in a reduced rate of TF or EGF internalization. Native CALM (N-CALM^1–660) overexpression reduced the rate of TF and EGF internalization by 53%. The portion of CALM found in MLL-CALM fusions, CALM^256–660, reduced TF and EGF internalization by 66% compared with controls. CALM^436–583 overexpression reduced the rate of TF and EGF internalization by 43%. A series of CALM deletion constructs (CALM^256–492, CALM^256–502, CALM^337–531, CALM^593–660, and CALM^601–660) did not perturb internalization of TF or EGF. Thus, the presence of a 52 amino acid (aa) region from CALM aa 531–583 was required to perturb internalization. Phosphorylation of JAK2 was prolonged in the setting of CALM-AF10 expression (U937 and P31/Fuji cells) compared with cell lines that do not express CALM-AF10 (HL60, K562, MonoMac6, THP1), likely as a consequence of perturbed CDE.

Conclusions: We demonstrate that CALM-dependent perturbation of CDE in COS7 cells can be measured by a novel flow cytometry-based approach. Use of this rapid, quantitative assay narrows the specific region of CALM critical to CDE perturbation to 52 aa. We also show that signaling downstream of growth factors is enhanced during perturbed CDE in leukemia cell lines. These observations support an underappreciated role for CDE dysregulation in leukemogenic transformation.