Transcriptomic-metabolomic reprogramming in EGFR-mutant NSCLC early adaptive drug escape linking TGFβ2-bioenergetics-mitochondrial priming

Abstract

The impact of EGFR-mutant NSCLC precision therapy is limited by acquired resistance despite initial excellent response. Classic studies of EGFR-mutant clinical resistance to precision therapy were based on tumor rebiopsies late during clinical tumor progression on therapy. Here, we characterized a novel non-mutational early adaptive drug-escape in EGFR-mutant lung tumor cells only days after therapy initiation, that is MET-independent. The drug-escape cell states were analyzed by integrated transcriptomic and metabolomics profiling uncovering a central role for autocrine TGFβ2 in mediating cellular plasticity through profound cellular adaptive Omics reprogramming, with common mechanistic link to prosurvival mitochondrial priming. Cells undergoing early adaptive drug escape are in proliferative-metabolic quiescent, with enhanced EMT-ness and stem cell signaling, exhibiting global bioenergetics suppression including reverse Warburg, and are susceptible to glutamine deprivation and TGFβ2 inhibition. Our study further supports a preemptive therapeutic targeting of bioenergetics and mitochondrial priming to impact early drug-escape emergence using EGFR precision inhibitor combined with broad BH3-mimetic to interrupt BCL-2/BCL-xL together, but not BCL-2 alone.

Keywords: EGFR; drug escape; inhibitor; lung cancer; resistance.

Figure 1. Principal component analysis (PCA) of precision therapy of EGFR-mutant drug-sensitive lung adenocarcinoma cells

1A. Segregation of HCC827 (deletion exon 19-EGFR) and H1975 (T790M/L858R-EGFR) cells in the two separate preclinical in vitro model systems. Transcriptome-wide profiling of early adaptive drug-escape in EGFR-mutant lung adenocarcinoma was analyzed using PCA. HCC827 cells were treated with erlotinib at 1 μM and H1975 cells were treated with CL-387,785 at 1 μM for 9 days followed by TKI washout for another 7 days in biologic triplicates. 1B. PCA of the gene expression data from the microarray gene expression profiling study as above in 1C. Transcriptomic profiling of early adaptive drug-escape against EGFR-TKI in (a) HCC827 and (b) H1975 lung adenocarcinoma cells - Clustering heat map analysis. 1D. Pathway analysis from the Affymetrix gene expression microarray studies in early adaptive drug-escape against EGFR-TKI in EGFR-mutant lung adenocarcinoma.

Figure 2. Autocrine TGFβ2 upregulation in lung adenocarcinoma early adaptive drug-escape correlated with EMT and stem cell signaling reprogramming

2A. HCC827 cells persisting under 9 days of erlotinib treatment displayed progressively downregulated expression of E-cadherin and remarkably increased expression of vimentin at day 9. Bright field microscopic images (right) showed HCC827 cells treated with exogenous TGFβ2 (5 ng/mL) for 9 days. 2B. Q-PCR results verified that the TGFβ2 gene expression was upregulated in HCC827 lung adenocarcinoma cells in adaptive escape against erlotinib. HCC827 cells were treated in culture without or with erlotinib for the indicated time durations, and TGFβ2 expression was elevated in both 8 hr (1.5 fold) and 9 days (~2.5 fold) cell groups when compared with control cells (DMSO treated) * p < 0.01. 2C. Immunofluorescence staining of TGFβ2 expression in HCC827 cells treated with erlotinib. Immunocytochemistry shows the TGFβ2 protein expression in cells treated with erlotinib at 3 different time points (8Hr, 9D and 9D+7D washout). DAPI nuclear staining (nuclear stain) is shown in the upper panel, TGFβ2 protein (green) staining alone is shown in the middle panel and both are merged and shown in the bottom panel. Scale bar - 10 μm. 2D. In vivo murine xenograft model of early adaptive drug-escape against erlotinib in HCC827 EGFR-mutant lung adenocarcinoma revealed upregulated TGFβ2 and depressed Ki-67 expression. IHC analysis of the post-treatment xenograft tumors using primary antibodies against human TGFβ2 and Ki-67 were performed and shown here (representative images). 2E. Heat map analysis of stem cell signaling genes expression in EGFR-mutant NSCLC.

Figure 3. Downregulation of glycolytic regulatory enzymes expression through TGFβ2 signaling to promote pro-survival mitochondrial-priming

3A. Induction of autocrine TGFβ2 cytokine expression in HCC827 cells persisted after 9 days of erlotinib treatment undergoing adaptive drug-escape, correlated with downregulated expression of the key glucose metabolism regulatory enzymes GPI, PGK1 and ENO2, as well as the cell cycle progression regulator TIMELESS. (a): HCC827 cells were treated without (0 hr) or with erlotinib for 8 hr, 9 days, and 9 days of TKI followed by 7 days of drug washout. Actin (bottom panel) was included as loading control. (b): Gene expression heat map signature showing adaptive suppression in gene expression of glucose metabolism regulatory Warburg genes GPI, PGK1 and ENO2 in HCC827 cells (upper panel) treated with erlotinib and H1975 cells (lower panel) treated with CL-387,785. TKI treatment conditions: a, 0 hr (untreated control); b, 8 hr; c, 9 days and d, 9 days of TKI followed by 7 days of drug washout. 3B. Immunoblot analysis of the effect of TGFβ2 on HCC827 and PC-9 cells in the expression levels of the glucose metabolism regulatory enzymes: GPI, PGK1, ENO2, and the mitochondrial prosurvival marker BCL-2/BCL-xL. (a) HCC827 cells were treated with exogenous TGFβ2 (5 ng/ml) for 9 days, at increasing concentration of the cytokine as indicated for immunoblotting. (b & c) HCC827 and PC-9 cells under similar treatment conditions as above were subjected to immunoblotting. 3C. Gene expression heat map signature of suppressed Warburg glycolytic enzyme genes in early adaptive EGFR-mutant NSCLC cells under targeted EGFR-TKI treatment. HCC827 cells treated with erlotinib (upper panel) and H1975 cells treated with CL-387,785 (lower panel). PKM2, pyruvate kinase M2; LDHA, lactate dehydrogenase A; ENO1, enolase 1; TPI1, triosephosphate isomerase 1; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. TKI treatment conditions: a, 0 hr (untreated control); b, 8 hr; c, 9 days and d, 9 days of TKI followed by 7 days of drug washout. 3D. (a) SiRNA-specific silencing of TGFβ2 (3 days) in erlotinib-treated HCC827 cells for 9 days compared with the untreated, 9-day treated erlotinib and the mock control-treated HCC827 cells. (b & c) siRNA knockdown of TGFβ2 shows a significant reduction in mRNA expression of TGFb2, BCL-2, and BCL-xL. BIM showed no significant difference in expression upon siRNA silencing of TGFβ2.

Figure 4A

Glucose metabolism, TCA cycle and glutaminolysis reprogramming. 4B. Erlotinib-treated HCC827 cells were treated with and without glutamine for 9 days followed by three days of no treatment. 4C. Relative number of viable early adaptive resistant cells without glutamine was significantly reduced compared with the cells treated with glutamine.

Figure 5. Branched-chain amino acid catabolism, and Lipid metabolism reprogramming in early adaptive drug-resistant HCC827 EGFR-mutant cells under erlotinib inhibition

Figure 6. Early adaptive transcriptomic-metabolomic cellular reprogramming under EGFR-TKI precision therapy to promote resistant drug-escape

6A, The schematic diagram highlights the central role of autocrine TGFβ2 in EGFR oncogene-addicted NSCLC. 6B, Cells were then treated with erlotinib (1 μM) for up to 9 days, with intermittent inhibitor replenishment every 3 days. After that, cells were further treated for 3 additional days using (a) diluent alone, (b) erlotinib alone, ABT-199 without (c) or with (f) continuing erlotinib, ABT-263 without (d) or with (g) continuing erlotinib, obatoclax without (e) or with (h) continuing erlotinib. HCC827 cells were also treated concurrent at the beginning of the experiment using erlotinib (1 μM) plus the respective BH3 mimetics (2 μM) as above (i: ABT-199, j: ABT-263, and k: obatoclax). 6C, Spectrometric quantification of the different groups including the untreated (A), treated with erlotinib (B: 12 days 1 μM Erlotinib), treated with erlotinib and ABT-263 (C: 12 days 1 μM Erlotinib + 3 days 2 μM ABT-263; D: 9 days 1 μM Erlotinib + 3 days 2 μM ABT-263; E: 12 days 1 μM Erlotinib + 12 days 2 μM ABT263), treated with erlotinib and ABT-199 (F: 12 days 1 μM Erlotinib + 3 days 2 μM ABT-199; G: 9 days 1 μM Erlotinib + 3 days 2 μM ABT-199; H: 12 days 1 μM Erlotinib + 12 days 2 μM ABT-199) and treated with Obatoclax (I: 12 days 1 μM Erlotinib + 3 days 2 μM Obatoclax; J: 9 days 1 μM Erlotinib + 3 days 2 μM Obatoclax; K: 12 days 1 μM Erlotinib + 12 days 2 μM Obatoclax). (ns - not significant; ** p < 0.002; *** p < 0.001).