The glucose-deprivation network counteracts lapatinib-induced toxicity in resistant ErbB2-positive breast cancer cells

Abstract

Dynamic interactions between intracellular networks regulate cellular homeostasis and responses to perturbations. Targeted therapy is aimed at perturbing oncogene addiction pathways in cancer, however, development of acquired resistance to these drugs is a significant clinical problem. A network-based computational analysis of global gene expression data from matched sensitive and acquired drug-resistant cells to lapatinib, an EGFR/ErbB2 inhibitor, revealed an increased expression of the glucose deprivation response network, including glucagon signaling, glucose uptake, gluconeogenesis and unfolded protein response in the resistant cells. Importantly, the glucose deprivation response markers correlated significantly with high clinical relapse rates in ErbB2-positive breast cancer patients. Further, forcing drug-sensitive cells into glucose deprivation rendered them more resistant to lapatinib. Using a chemical genomics bioinformatics mining of the CMAP database, we identified drugs that specifically target the glucose deprivation response networks to overcome the resistant phenotype and reduced survival of resistant cells. This study implicates the chronic activation of cellular compensatory networks in response to targeted therapy and suggests novel combinations targeting signaling and metabolic networks in tumors with acquired resistance.

Figure 1

Initial characterization of lapatinib-resistant cell line. (A) Percent change in cell numbers in response to increasing doses of lapatinib in parental and resistant SKBR3 cells. (B) Cell-cycle and apoptosis analysis of the parental and resistant cells in response to 1 μM lapatinib after 4 days. (C) Immunoblotting of ErbB signaling pathway before and after 1 μM lapatinib treatment. Immunoblotting of phosphorylation sites on ErbB1, 2 and 3 family members as well as the downstream signaling proteins AKT, MAPK1, 2, mTOR, S6Kinase and S6 before and after lapatinib treatment.

Figure 2

Network analyses of gene expression data. (A) A heatmap of Edge Flux (EF) values with highest variance across the six conditions (see text and Materials and methods). (B) A network plot of the interactions in K1. Nodes are colored by the relative gene expression values of respective genes in resistant cells (see the red-green color key), and edges are colored according to the type of interaction. (C) Functional enrichment scores of highest scoring GO functional classes in the network of 500 highest EF values in resistant cells relative to parental cells. Enrichment score was calculated as the −log of hypergeometric distribution function. (D) Heatmap of gene expressions of some genes in our data set previously implicated in glucose deprivation response.

Figure 3

Glucose deprivation response phenotype in lapatinib resistance. (A) Immunoblotting of key members of glucose deprivation response in parental and resistant cells HSPA5, HK2, IRE1, pJNK, pAMPK, PERK and p38. (B) Immunoblotting of the AKT and AMPK phosphorylation sites on TSC2 in parental and resistant cells. (C) Glucose uptake flux analysis (*P<0.05), (D) lactate production flux analysis (*P<0.05), (E) ratio of lactate to glucose (*P<0.05) and (F) total ATP content of parental and resistant cells after 24 h of lapatinib treatment (*P<0.01). (G) Change in cell numbers of parental SKBR3 cells treated with increasing doses of lapatinib in a media with normal and no glucose (*P<0.01). (H) Change in cell numbers of BT474 cells treated with increasing doses of lapatinib after prolonged incubation in a media with normal (2 g/l) and low (0.25 g/l) glucose (*P<0.01).

Figure 4

Correlation of glucose deprivation response with clinical relapse rates in ErbB2-positive breast cancers. (A) Network of genes with highest correlation with relapse in ErbB2-positive tumor patients (see Materials and methods). Node coloring reflects the strength of correlation (COX regression z-score, see color key). Edge coloring is same as in Figure 2A. (B) Kaplan–Meier plots of relapse-free survival of ErbB2-positive patients segregated based on their expression of some key members of glucose deprivation response in lapatinib-resistant cells that are also present in the network in Figure 4A. (C) Functional enrichment scores of highest scoring GO functional categories in the network of highest EF values calculated by NetWalk analysis of COX regression values. Functional enrichment scores are calculated as −log of the hypergeometric distribution function.

Figure 5

Identification and treatment of resistant cells with drugs reversing the hypoglycemic response phenotype. (A) Gene expression profiles of 12 genes in the glucose deprivation response gene set that are specifically upregulated in the resistant cells. Tested genes are HSPA5, NDRG1, HK2, SLC2A10, HYOU1, HMGCS1, SRPR, ALDH3 A2, UGCGL1, GYS1, TXNIP and GLRX. The plot shows the position of each of these genes (indicated by a black vertical line) in the whole distribution of gene expression values in response to indicated drugs. Note specific positioning of most of these genes in the lower part of the distribution with pyrvinium. (B) Distribution of average ranks of 10 000 random draws of 12 genes in the pyrvinium data. The arrow indicates the average rank of the GD (glucose deprivation response) set. (C) Change in cell numbers of the parental and resistant cells in response to increasing doses of pyrvinium after 3 days of treatment (*P<0.01). (D) Western blots of parental and resistant cells before and after treatment with lapatinib probed for LC3-II, a lipidated LC3 molecule, a marker of active autophagy. (E) Change in cell numbers of the parental and resistant cells in response to increasing doses of bafilomycin A (*P<0.01).