Galectin-3 in pre-B acute lymphoblastic leukemia

Only 30–40% of adult patients with acute lymphoblastic leukemia (ALL) achieve long-term, disease-free survival and can be considered cured. The remainder will eventually relapse and die from disease progression or other ALL-related consequences.^1,2 Relapse most frequently originates in the bone marrow (BM),³ where stroma protects ALL cells against drug treatment.⁴

The development of drug resistance in ALL cells as occurs in vivo can be modeled by coculture of ALL cells in the presence of irradiated stromal cells during drug exposure. We have shown previously that this allows for the outgrowth of drug-insensitive leukemic cells (e.g. Zhang et al⁵), in a process that has been named environmental-mediated drug resistance. Microarray analysis on precursor B-lineage (pre-B) ALL cells from a BCR/ABL transgenic mouse model for Ph-positive ALL undergoing environmental-mediated drug resistance showed significantly increased expression of mRNAs related to stress and inflammation.⁶ Among others, we found that transcript levels of the Galectin-3 gene (Lgals3) were significantly increased upon development of resistance to the Bcr/Abl-targeted tyrosine kinase inhibitor nilotinib, and this was also seen in pre-B ALL cells isolated from the BM of BCR/ABL transgenic mice that had been treated with nilotinib for 8 days.⁶

Galectin-3, a 30-kDa protein without enzymatic activity, belongs to the family of lectins, which recognize and specifically bind (mostly extracellular) carbohydrate structures. Galectin-3 has a C-terminal carbohydrate-binding domain with preference for galactose and poly-N-acetyllactosamine-modified glycans. Its N-terminal domain promotes homotypic oligomerization. Because of these domains, extracellular Galectin-3 can promote the cross-linking and lattice formation between glycan-containing molecules on the cell surface.⁷

ALL cells were reported not to express Galectin-3.⁸ The induction of its transcripts in the pre-B ALL cells prompted us to further examine this. We first assessed expression of Galectin-3 protein in BM samples from ALL patients using fluorescence-activated cell sorting (FACS). Galectin-3 was clearly present in pre-B cells from BM of all four ALL patients (Figure 1a). Elevated serum levels of Galectin-3 were reported in some types of cancer such as metastatic colorectal cancer and melanoma.^9,10 To evaluate circulating levels of Galectin-3 in ALL patients, plasma samples from peripheral blood and BM were collected from ALL patients, as well as controls. Interestingly, we found that plasma levels of Galectin-3 in BM from ALL patients were significantly higher than healthy controls (P<0.05) (Figure 1b) as detected by enzyme-linked immunosorbent assay. To further investigate circumstances under which Galectin-3 levels are increased, we used human patient-derived pre-B ALL cells that were passaged in NSG mice and grown ex vivo in the presence of OP9 stromal feeder cells. In this system, the ALL cells adhere to the stromal cells, migrate underneath them and return into the medium in a dynamic fashion. Notably, cells that were harvested from underneath the OP9 layer, and had thus been in direct contact with OP9 cells, expressed a high amount of Galectin-3, while cells in suspension or cells cultured without OP9 for 24 h had low expression of cell surface or intracellular Galectin-3 as measured by FACS (Figure 1c). OP9 cells secrete Galectin-3 (Supplementary Figure 1) and part of the Galectin-3 detected on these ALL cells was of stromal origin (Fei F, Lim M, Groffen J, Heisterkamp N, manuscript in preparation).

Figure 1.

Expression of Galectin-3 in pre-B ALL samples. (a) Total (cell surface + intracellular) expression of Galectin-3 in CD19⁺ CD10⁺ double-positive cells from primary ALL patient BM samples as assessed by FACS in fixed permeabilized cells. Control, isotype-matched immunoglobulin. (b) Galectin-3 levels in plasma from peripheral blood or BM of normal individuals or ALL patients detected by enzyme-linked immunosorbent assay (n = 5/group) (*P<0.05). (c) FACS analysis of Galectin-3 expression on Ph-positive TXL2 and Ph-negative US7 human ALL cells without OP9 stroma or in suspension (supernatant), or underneath OP9 stromal cells. Surface (fixed or non-fixed cells) Galectin-3 or total Galectin-3 in fixed permeabilized cells, as indicated.

As increased cellular levels of Galectin-3 are protective to several cancer cell types,^8,11,12 we elevated Galectin-3 levels in ALL cells by transduction of TXL2 cells with pMIG-Gal3 and control vector pMIG. As phosphorylation of serine residue 6 in Galectin-3 is needed for this protein to provide protection to BT-549 breast cancer cells against cisplatin-induced apoptosis,¹² we also compared the effect of a Gal3S6A mutant. Analysis of major signal transduction pathways in TXL2-pMIG, TXL2-pMIG-Gal3 and TXL2-pMIG-Gal3S6A cocultured with or without OP9 cells did not show major differences between these cells and their growth was also not statistically significantly different (Supplementary Figure 2). Interestingly, compared with TXL2-pMIG and TXL2-pMIG-Gal3 cells, TXL2-pMIG-Gal3-S6A cells were more vulnerable to spontaneous apoptosis when cultured without OP9 cells over a 3-day period (Figure 2a). Consistent with studies in other hematopoietic cell type-derived cancers including chronic myeloid leukemia, diffuse large B-cell lymphoma and multiple myeloma,^8,13,14 overexpression of Galectin-3 enhanced drug resistance of the TXL2 cells, to both nilotinib and vincristine, whereas TXL2-pMIG-Gal3S6A cells were more sensitive to drug treatment (Figure 2b and c). These results demonstrate that increased intracellular Galectin-3 expression promotes resistance to drug treatment, and that abrogation of its ability to be phosphorylated on S6 results in loss of such capability.

Figure 2.

Galectin-3 protects ALL cells against drug treatment. (a) TXL2-pMIG, TXL2-pMIG-Gal3 and TXL2-pMIG-Gal3S6A cells cultured without OP9 cells over a 3-day period. Apoptosis was measured using FACS (*P<0.05, **P<0.01). (b, c) TXL2 cells transduced with pMIG, pMIG-Gal3 or pMIG-Gal3S6A were treated with nilotinib (b) or vincristine (c). (d) Analysis of signal transduction in gal3 +/+ and gal3 −/− pre-B ALL cells as detected by western blot. Cells were grown without stroma. (e, f) Comparison of viable cell counts of gal3 +/+ and gal3 −/− pre-B ALL cells treated with nilotinib (e) or vincristine (f) for 72 h. Viable cells were counted by Trypan blue exclusion assay. Results shown are from one of three independent experiments with similar results.

To examine the effect of loss of endogenous Galectin-3 on ALL cells, we generated pre-B ALL cells from gal3 +/+ and gal3 −/− BM cells by transduction with Bcr/Abl using standard procedures.¹⁵ These cells do not require support from stroma, allowing an assessment of the function of endogenously produced Galectin-3. Transduction efficiency was confirmed by western blot (Supplementary Figure 3). Gal3 +/+ and gal3 −/− pre-B ALL cells had a similar proliferation rate as assessed by cell counting and cell cycle analysis (data not shown). We found that gal3 −/− cells had a specific attenuation of Erk pathway activation, as indicated by a significant decrease in pErk1/2 levels (Figure 2d). To compare drug sensitivity, gal3 +/+ and gal3 −/− pre-B ALL cells were treated with different concentrations of nilotinib or vincristine. As expected, nilotinib or vincristine treatment reduced viability of Bcr/Abl transformed pre-B ALL cells, but gal3 −/− pre-B ALL cells were significantly more sensitive to nilotinib or vincristine treatment than gal3 +/+ pre-B ALL cells (Figure 2e and f). Thus, our studies show that increased levels of Galectin-3, as found in BM plasma and pre-B ALL cells, provide protection to these leukemia cells against drug treatment.