3-Phosphoinositide-dependent kinase 1 drives acquired resistance to osimertinib

Abstract

Osimertinib sensitive and resistant NSCLC NCI-H1975 clones are used to model osimertinib acquired resistance in humanized and non-humanized mice and delineate potential resistance mechanisms. No new EGFR mutations or loss of the EGFR T790M mutation are found in resistant clones. Resistant tumors grown under continuous osimertinib pressure both in humanized and non-humanized mice show aggressive tumor regrowth which is significantly less sensitive to osimertinib as compared with parental tumors. 3-phosphoinositide-dependent kinase 1 (PDK1) is identified as a potential driver of osimertinib acquired resistance, and its selective inhibition by BX795 and CRISPR gene knock out, sensitizes resistant clones. In-vivo inhibition of PDK1 enhances the osimertinib sensitivity against osimertinib resistant xenograft and a patient derived xenograft (PDX) tumors. PDK1 knock-out dysregulates PI3K/Akt/mTOR signaling, promotes cell cycle arrest at the G1 phase. Yes-associated protein (YAP) and active-YAP are upregulated in resistant tumors, and PDK1 knock-out inhibits nuclear translocation of YAP. Higher expression of PDK1 and an association between PDK1 and YAP are found in patients with progressive disease following osimertinib treatment. PDK1 is a central upstream regulator of two critical drug resistance pathways: PI3K/AKT/mTOR and YAP.

Fig. 1. Effect of osimertinib on survival of NCI-H1975 and NCI-H1975-OsiR cells.

a SRB assay was performed using two sets of osimertinib concentrations for sensitive and resistant cells to determine the IC₂₀, IC₃₀ and IC₅₀. b Table shows the IC₂₀, IC₃₀, and IC₅₀ values of H1975 and H1975-OsiR cells and the cell doubling time. Data shown represent the mean ± SE of three independent experiments.

Fig. 2. Effect of osimertinib on humanized H1975 and H1975-OsiR xenografts.

Humanized mice were generated by human CD34 stem cells implantation. After mice become humanized with over 25% human CD45 cells, the NCI-H1975 and NCI-H1975-OsiR cells were injected subcutaneously. a Experimental strategy for mouse humanization, tumor cell inoculation, and osimertinib prolonged treatment, b Levels of human immune cell repopulation in humanized mice at different time points. c Tumor growth comparison between NCI-H1975-parental vs. NCI-H1975-OsiR and the effect of osimertinib on their growth. d Experimental strategy for mouse humanization, tumor cell inoculation, and osimertinib short term treatment. e Antitumor effect of osimertinib on tumor growth. In each experiment, N ≥ 5 humanized mice/group were used. The humanized mice experiments were repeated N = 3 times.

Fig. 3. RPPA Gene expression profile analysis in osimertinib treated residual xenograft tumors.

NCI-H1975 and NCI-H1975-OsiR xenograft tumors were developed in NSG mice. Osimertinib-resistant tumors were developed under continuous osimertinib pressure by treating the mice with osimertinib following tumor cell implantation. At the end of the treatment, the residual tumors were harvested and snap-frozen for RPPA analysis. a Treatment strategy shows the timeline for osiemretinib treatment and tumor harvest for RPPA, b Growth curve comparison between NCI-H1975 and NCI-H1975-OsiR tumors. c Effect of osimertinib on NCI- H1975 parental tumors (N = 6mice/group) at two different doses; 5 mg/kg and 10 mg/kg, d Effect of osimertinib on NCI-H1975-OsiR tumors (N = 6mice/group) at two different doses; 5 mg/kg and 10 mg/kg, e Osimertinib response at D39, f Pairwise comparison between NCI-H1975 and NCI-H1975-OsiR residual tumors after osimertinib treatment, left shows the heatmap, middle shows the volcano plot and right has the list of upregulated proteins in NCI-H1975-OsiR as compared with NCI-H1975 xenografts. g Heatmap (left), volcano plot (middle) show the differences in protein expression in NCI-H1975 residual tumors after prolonged osimertinib treatment vs. control tumors. The list of major upregulated proteins are shown on the right. The criteria of protein selection for significantly up- down-regulation were: 1. Significant in overall F-test (FDR-adjusted P-value < 0.05); 2. Significant in pairwise comparison. (FDR-adjusted P-value < 0.05); 3. The fold-change of >1.5 or < −1.5 indicates whether a gene is upregulated or downregulated.

Fig. 4. PDK1, 3-Phosphoinositide-dependent kinase 1, was upregulated in osimertinib-resistant xenograft tumors developed in humanized mice.

NSG mice were humanized by human CD34 stem cells implantation, after humanization, NCI-H1975 and NCI-H1975-OsiR cells were injected subcutaneously. The resistant xenograft tumors were developed under osimertinib pressure throughout the experiment. At the end, tumors were harvested and IHC was performed for PDK1 expression. a the level of PDK1 expression between Hu-H1975 and Hu-H1975-OsiR tumors; b the PDK1 signal intensity and PDK1-positive cell counts were quantitated; c the level of PDK1 expression was compared in NCI-H1975-OsiR tumors developed in humanized and non-humanized mice; d the signals were quantitated, and statistics were performed.

Fig. 5. In vitro inhibition of PDK1 or PDK1 knockout increased osimertinib sensitivity and inhibited colony formation.

a NCI-H1975 and NCI-H1975-OsiR cells were treated with PDKi, BX 795, and dose-dependent inhibition of PDK1 and pPDK1 are shown; b osimertinib XTT was performed on NCI-H1975 and NCI-H1975-OsiR cells in the presence or absence of BX 795 and shows that BX795 renders NCI-H1975-OsiR sensitive to osimertinib; c osimertinib XTT with or without BX 795 on PDK1 knockout clone, NCI-H1975-OsiR-PDK1^-/-; d osimertinib XTT on re-expressing PDK1 into PDK1 KO clone, NCI-H1975-OsiR-PDK1^++/++ in presence or absence of BX 795. e Generation of CRISPR-cas9 mediated PDK1 knockout clone, NCI-H1975-OsiR-PDK1^-/- and rescue of PDK1 in the KO clone by stably expressing the PDK1 plasmid to make NCI-H1975-OsiR-PDK1^++/++ clones; f XTT assay shows the effect of PDK1 KO on osimertinib sensitivity; g XTT shows rescue of PDK1 expression in the PDK1 KO clone reverted the resistance to osimertinib. h Transient expression of PDK1 into NCI-H1975-parental cells and osimertinib XTT on transiently expressing PDK1 cells; i, j Colony formation assay shows osimertinib differential sensitivity among all clones. Data shown represent the mean ± SE of three independent experiments.

Fig. 6. In vivo inhibition of PDK1 by PDK1 inhibitor, BX795, enhanced osimertinib response in resistant PDXs.

Fresh osimertinib-resistant TC386-OsiR PDXs were implanted into NSG mice. When PDX sizes reached around 200mm³, PDXs bearing mice were randomized into different treatment arms for the treatment of osimertinib and BX 795. a End of the treatment harvested PDXs were checked for PDK1 and pPDK1 (S241) expression in TC386-OsiR PDXs treated with BX 795, osimertinib and osimertinib + BX 795. b Level of PDK1 and pPDK1 (S241) expression in osimertinib-sensitive TC386 PDX, and ositmertinib-resistant TC386-OsiR PDX tissues comparing with NCI-H1975-OsiR/PDK1^-/- and NCI-H1975-OsiR-PDK1^++/++ cells, c Osimertinib + BX795 treatment strategies in osimertinib-resistant PDXs, d Antitumor activity of Osimertinib + BX795 combination on TC386-OsiR PDXs, e Growth curves of TC386-OsiR PDXs bearing individual mice in different treatment groups. Each treatment group was N = 5 PDX bearing mice.

Fig. 7. PDK1 knock-out dysregulates AKT/mTOR signaling and promotes cell cycle arrest.

PDK1 KO cells, NCI-H1975-OsiR-PDK1^-/- and PDK1 re-expressing cells, NCI-H1975-OsiR-PDK1^++/++ were treated with osimertinib and upstream and downstream signaling molecules were investigated by the western blot. a AKT, b mTOR, and c PTEN expression and its quantitation in NCI-H1975-OsiR-PDK1^-/- and NCI-H1975-OsiR-PDK1^++/++ cells and alteration by osimertinib treatment. c PTEN expression in NCI-H1975-parental, NCI-H1975-OsiR, NCI-H1975-OsiR-PDK1^-/- and NCI-H1975-OsiR-PDK1^++/++ cells. d Cell cycle analysis of NCI-H1975-parental, NCI-H1975-OsiR, NCI-H1975-OsiR-PDK1^-/- and NCI-H1975-OsiR-PDK1^++/++ cells after osimertinib treatment. e Quantitation of cells in difference phases and its alteration by osimertinib treatment.

Fig. 8. PDK1 knock-out inhibits YAP expression and nuclear translocation.

a Western blot shows expression of total YAP, pYAP (S127), pYAP (S397) in osimertinib-sensitive NCI-H1975, resistant NCI-H1975-OsiR, NCI-H1975-OsiR-PDK1^-/- and NCI-H1975-OsiR-PDK1^++/++ cells, b Quantitation of pYAP (S127), pYAP (S397) and total YAP in NCI-H1975 isogenic cell lines, c Immunoflourescence images of nuclear translocation of YAP detected by immunostaining with pYAP (Tyr 357) antibody on osimertinib-sensitive and resistant cell lines as well as PDK1 knockout and re-expressing cells, d Quantitative analysis of nuclear YAP signals in four NCI-H1975 isogenic cell lines, e Active-YAP expression was evaluated by IHC using anti-active-YAP antibody on NCI-H1975 and NCI-H1975-OsiR xenograft tumors. f Active-YAP signals were quantitated from IHC sections by imageScopre using 15–20 10X images per group. g Level of YAP was determined from RPPA analysis of osimertinib-sensitive and resistant xenograft tumors (left panel) and osimertinib treated residual sensitive and resistant tumors (right panel).