In this study, we examined the potential role of IGFBP3 as a mediator of the therapeutic effects of imatinib mesylate in GISTs. Our previous studies showed that IGFBP3 is up-regulated after imatinib treatment in a responsive GIST cell line (GIST882), and we provide evidence that IGFBP3 does indeed partially mediate GIST882 cell response to imatinib in vitro. In contrast, IGFBP3 has no effect on imatinib sensitivity in the responsive GIST-T1 cell line, which has no detectable endogenous IGFBP3 levels before or after imatinib exposure. Further, our studies, using both gain-of-function and loss-of-function approaches, reveal that IGFBP3 is an important modulator of cell viability in GISTs, but the effect is cell-dependent. Similar to what has been reported for epithelial cancers [23, 26–34], IGFBP3 also manifests dual functions on cell survival in GIST, a mesenchymal cancer.
Up-regulation of IGFBP3 has been observed in response to a variety of anti-cancer agents [18–23], including celecoxib . In addition, IGFBP3 potentiates the action of paclitaxel  and sensitizes cancer cells to the cytotoxic effects of gefitinib  and other chemotherapeutic agents . Because we observed IGFBP3 expression in GIST in response to imatinib , we hypothesized that IGFBP3 would mediate its anti-tumor effects. After manipulating IGFBP3 levels in two GIST cell lines, we observed a modulating effect on response in GIST882, suggesting that the induction of IGFBP3 is a significant, specific response to imatinib-induced stress. Failure to observe a similar response in GIST-T1 suggested that GIST-T1 cells are insensitive to IGFBP3. However, additional studies showed that IGFBP3 regulates GIST cell viability with opposing effects. Overexpression of IGFBP3 in GIST-T1 cells, which have no detectable endogenous IGFBP3 expression before or after imatinib, results in a loss of cell viability, demonstrating that IGFBP3 has growth inhibitory effects in this cell line. In contrast, we expected that the loss of IGFBP3 by neutralization or knockdown in GIST882 cells, which have increased IGFBP3 expression after imatinib, would have a protective effect on cell viability. However, our data shows that IGFBP3 down-modulation is cytotoxic, demonstrating that IGFBP3 is necessary for cell viability. Thus, in GIST882, IGFBP3 has two distinct roles, which may be attributed to a dose-dependent mechanism. Dual functions of IGFBP3 have been reported previously in cancers of the renal cells [26, 27], esophagus [28, 29], breast [30, 31], colon [32, 33], and prostate [23, 34], as well as in endothelial cells . The mechanism that determines the final outcome of IGFBP3 action is not well understood, though some studies suggest a role for post-translational modification , localization within specific cellular compartments [25, 40], extracellular matrix composition , or binding partner interaction [15, 42, 43]. Despite its dual effects on GIST cell viability, IGFBP3 appears to exert its effects through a KIT-independent mechanism, as imatinib-induced KIT inactivation has no effect on IGFBP3-mediated loss of cell viability in either GIST882 or GIST-T1 cells.
IGFBP3 expression is lost in many cancer cells [44–46], and reintroduction of the protein often results in cell death [21, 46, 47]. Similarly, our results show that IGFBP3 expression is not detectable in GIST-T1 cells but overexpression leads to loss of cell viability. Indeed, the growth inhibitory and pro-apoptotic effects of IGFBP3 are well established in a variety of in vitro and in vivo cancer models. On the other hand, IGFBP3 also has growth stimulatory effects [29, 33, 34, 38, 41, 48], depending on the cell type and context. Further, increased IGFBP3 expression has also been linked to renal cell carcinoma , breast cancer [31, 49], and metastatic melanoma , suggesting that IGFBP3 may contribute to tumorigenesis or disease progression. Here, we report that GIST882 cells, which have detectable IGFBP3 protein expression, require IGFBP3 for cell viability, confirming the notion that IGFBP3 may facilitate cancer cell proliferation and survival. Complete understanding of IGFBP3 requires investigations of its binding partners, post-translational modifications, and signal transduction pathways in vitro and in vivo.
One possible pathway through which IGFBP3 may exert its effects in GISTs is the IGF pathway. A number of recent studies have explored the IGF axis for prognostic and therapeutic value in GISTs. Braconi and colleagues reported that expression of IGF-1 and IGF-2 is correlated with poor prognosis and relapse, and that IGF-1R expression was strong in all cases . Furthermore, Tarn and colleagues reported that knockdown of IGF-1R was cytotoxic in GIST-T1 cells . IGFBP3 is the most abundant IGF binding protein in the circulation and is responsible for a majority of IGF transport . Because IGFBP3 has intrinsic IGF-binding activity that can act to sequester IGF from its cognate receptor , it is possible that using IGFBP3 as a therapeutic agent would be useful to GIST patients with abnormal IGF expression or IGF-dependent IGF-1R activation. Furthermore, if IGFBP3 is indeed acting through an IGF-dependent mechanism, a difference in the expression levels of IGF or IGF-1R or increased sensitivity to IGF might contribute to the differential IGFBP3-induced effects on cell viability and imatinib response in GIST882 or GIST-T1. Additional studies are needed to determine IGF and IGF-1R expression levels and IGF sensitivity in GIST cell lines and to further examine whether IGFBP3 functions through an IGF-dependent or IGF-independent mechanism in GIST.
In addition to its direct effects on cancer cells, IGFBP3, as a secreted protein, may also have paracrine effects on the tumor environment. Recent studies report that IGFBP3 regulates endothelial cell survival  and suppresses angiogenesis [55, 56]. Thus, it is possible that IGFBP3 further modulates the viability of GIST cells or alters their response to imatinib by targeting endothelial cells or other important cell types, such as macrophages, in the tumor microenvironment. However, the present study is limited to an in vitro cell culture system. Mouse model studies are needed to further investigate whether the effects of IGFBP3 extend to the GIST microenvironment.