Cell proliferation through forced engagement of c-Kit and Flt-3

A variety of receptors coexist on the surface membrane of hemopoietic cells. To test for potential functional interactions between different receptors, we have employed a system that allows the signaling domains of 2 different receptors to be forcibly juxtaposed. The system relies on coexpressing 2 fusion proteins in the interleukin-3 (IL-3)–dependent cell line Ba/F3.1 The first fusion protein contains the signaling domain of one receptor linked to the FK506 binding protein (FKBP12), whereas the second fusion protein contains the signaling domain of another receptor linked to the FKBP12–rapamycin-binding (FRB) domain of the FKBP12–rapamycin-associated protein (FRAP). Dimerization is achieved by the synthetic ligands FK10122,3 (or its synthetic analogue AP15104,5) and rapamycin.6FK1012 and AP1510 induce homodimerization of the FKBP12 fusions, whereas rapamycin directs heterodimerization between the FKBP12 and FRB fusions. Signaling is reflected in Ba/F3 cell growth in the absence of IL-3.

The c-Kit^3FKBP12 and c-Kit^1FKBP12constructs have been described previously.5 FRB domain–containing constructs were generously provided by Steffan Ho and were generated by inserting a XhoI/SalI linkered fragment of FRAP (residues 2026-2114) downstream of a myristylation-targeting sequence from c-Src, as described previously.2 Two FRB-containing constructs were made, one with 3 tandem copies of FRB and one with a single FRB domain, and were named pMFRB3E and pMFRB1E, respectively. A 1296-bp fragment encoding the cytoplasmic domain of murine c-Kit was freed from a pBS-based vector5 by XhoI digestion. The fragment was gel purified and subcloned into the SalI site of the pMFRB1E and pMFRB3E constructs to generate c-Kit^1FRB and c-Kit^3FRB. A 1313-bp fragment encoding the cytoplasmic portion of Flt-3 was amplified by polymerase chain reaction (PCR) using Pfu polymerase and a plasmid cDNA template (a gift from Ihor Lemishcka, Princeton University, NJ). The following oligonucleotide sequences were used for the amplification reaction: 5′-CCC CTC GAG TGC CAC AAA TAC AAA AAG C-3′ and 5′-CCC CTC GAG CTA ACT TCT TTC TCT GTG A-3′. Each primer contains added sequence to incorporate a XhoI restriction site at each end of the PCR product. The amplicon was digested with XhoI, gel purified, and subcloned into XhoI-digested pBluescript. After sequence confirmation using the PRISM system (Applied Biosystems, Foster City, CA), the fragment was released from pBluescript byXhoI digestion, gel purified, and subcloned into theSalI site of the pMFRB1E and pMFRB3E constructs. The plasmids generated were named Flt-3^1FRB and Flt-3^3FRB.

Electroporation, Western blot analysis, and cell proliferation assays were performed as described previously.3 

Studies were directed at determining whether forced apposition of c-Kit and Flt-3 could induce Ba/F3 cell growth. Constructs were generated in which sequences encoding the intracellular portions of c-Kit and Flt-3 were linked to one or more copies of either FKBP12 or FRB. We have previously reported that FK1012 and AP1510 induce proliferation in Ba/F3 cells that express a fusion protein containing 3 tandem copies of FKBP12 linked to the intracellular portion of c-Kit (c-Kit^3FKBP12) (Figure 1A, upper and middle panels). In contrast to the proliferative effects of FK1012/AP1510, these cells failed to proliferate in response to rapamycin (Figure 1A, lower panel), confirming that rapamycin is unable to activate the c-Kit^3FKBP12 fusion and lacks nonspecific proliferative effects in Ba/F3 cells. In fact, as expected,7 increasing concentrations of rapamycin were inhibitory to IL-3–dependent Ba/F3 cell growth (data not shown).

To evaluate the feasibility of achieving rapamycin-inducible cell growth, we tested Ba/F3 cells expressing both c-Kit^3FKBP12and a second fusion containing the same c-Kit fragment linked to 3 copies of the FRB domain of FRAP (c-Kit^3FRB).8Fusions containing 3 copies of the drug-binding domain were expected to allow for the generation of both dimers as well as higher-order oligomers by chemical inducers of dimerization (CID). C-kit^3FKBP12 homo-oligomers, generated by the addition of FK1012, induced cell growth as expected (Figure 1B, upper panel). Rapamycin-mediated juxtaposition of the c-Kit^3FKBP12 and c-Kit^3FRB fusions also induced Ba/F3 cell growth (Figure1B, lower panel), although to a lesser extent than observed in response to FK1012. The relatively less intense cell growth observed in response to rapamycin may be due to the well-characterized antiproliferative effects of rapamycin.7 Alternatively, signaling by c-Kit^3FKBP12–c-Kit^3FRBhetero-oligomers may be less efficient, possibly as a consequence of suboptimal alignment between the 2 c-Kit signaling domains.

To test whether c-Kit can functionally interact with Flt-3, we generated Ba/F3 cell clones expressing both c-Kit^3FKBP12and Flt-3^3FRB fusions. Of note, fusions containing Flt-3 linked to 3 copies of FKBP12 produced factor-independent growth in Ba/F3 cells and were therefore not characterized further (data not shown). As illustrated in Figure 1C (lower panel), 2 clones coexpressing the c-Kit^3FKBP12 and Flt-3^3FRBfusions demonstrated proliferative responses to rapamycin.

As noted above, rapamycin-induced aggregation of the c-Kit^3FKBP12 and Flt-3^3FRB fusions would be expected to produce hetero-oligomeric complexes. To determine whether c-Kit–Flt-3 heterodimers were sufficient to induce Ba/F3 cell growth, we made 2 additional constructs in which c-Kit was linked to a single FKBP12 domain (c-Kit^1FKBP12) and Flt-3 was linked to a single FRB domain (Flt-3^1FRB). As shown in Figure 1D (lower panel), Ba/F3 cell clones coexpressing the c-Kit^1FKBP12 and Flt-3^1FRB fusions exhibited dose-dependent proliferative responses to rapamycin.

Studies directed at examining tyrosine phosphorylation following stimulation with either FK1012 or rapamycin were performed. Ba/F3 cells coexpressing the c-Kit^3FKBP12 fusion and the Flt-3^3FRB fusion exhibited constitutive tyrosine phosphorylation, including phosphorylation of the c-Kit^3FKBP12 fusion, as we have reported previously.5 FK1012 induced a modest increase in tyrosine phosphorylation, whereas rapamycin had no discernible effect (Figure2). We interpret the disparity between our cell-proliferation assays and signaling studies to indicate that although rapamycin-triggered signals are sufficient to induce cell proliferation, they fall below the threshold needed for inducing appreciable changes in tyrosine phosphorylation.

Receptor tyrosine kinases are segregated into approximately 20 structurally related subfamilies, and interactions between different members within a subfamily are well described, as exemplified by the 4 members of the epidermal growth factor receptor (EGFR) subfamily. The formation of heterodimers between different members of the EGFR family is a key mechanism for achieving signal diversity.9 Using the same system as described here, Muthuswamy and colleagues10 showed that c-Cbl phosphorylation is induced by ErbB1 homodimers, but not by ErbB1–ErbB2 heterodimers. In the hemopoietic system, functional interactions have been documented between c-Kit and the erythropoietin receptor.11Additionally, the erythropoietin receptor has been found to interact directly with the common β chain of the granulocyte-macrophage colony-stimulating factor and IL-3 receptors.12 

This method provides a general approach for identifying productive interactions between different growth factor receptor signaling domains. In the present study, we show that functional interactions between c-Kit and Flt-3 are possible. Whether similar interactions occur in nature is unknown. Once the potential for functional interactions between different receptors is revealed, these interactions can be further explored for their physiologic relevance.

We thank Ihor Lemischka for providing the Flt-3 cDNA, Steffan Ho for the FRB cDNA, Mike Gilman and Tim Clackson for providing FK1012 and AP1510, Joachim Deeg for providing rapamycin, and Barbara Richard for providing expert technical support.