Development and characterization of sotorasib-acquired resistant patient-derived xenografts (PDXs) and PDX-derived organoids (PDXOs).

A) A list of KRASG12C mutant-bearing NSCLC PDXs was developed in our PDX core. All the PDXs underwent whole-exome sequencing to identify their co-mutation status. B) Sotorasib treatment responses in four KRASG12C mutant NSCLC PDXs (TC247, TC303, TC314 and TC453). Tumors were treated with sotorasib (100 mg/kg) when tumor volume reached around 200 mm3. Control mice were treated with solvent. Sotorasib or solvent was administered orally QD for 21 days. Tumor volume changes in individual mice were calculated for treatment starting day 0. C) Third-generation sotorasib-acquired resistant TC303AR and TC314AR PDXs were shown and compared with their parental TC303 and TC314 PDXs. D) PDXOs were developed and characterized. Whole exome sequence (WES) in PDXOs showed the KRASG12C mutation status (right panel) and PDXOs purity was determined by flow cytometry for detecting any mouse cells contamination by staining with mouse MHC Class I antibody (H2kb) and human epithelial marker (EpCAM) (left panel). E) The list of PDXOs developed from their respective PDXs and their sensitivity to sotorasib and adagrasib were determined based on IC50 (µM). F) Cell viability assay was performed on PDXOs for sotorasib (Left panel) and adagrasib (Right panel) sensitivity. Bold black lines indicate the PDXOs that were sensitive and ash color lines indicate the PDXOs that were resistant to KRAS inhibitors. G) Sotorasib-acquired resistant PDXO303AR and PDXO314AR were developed and these acquired resistant PDXOs were treated with sotorasib (top panel) and adagrasib (bottom panel) and compared with their parental counterparts (PDXO303 & PDXO314). Data shown represent the mean ± SE of three independent experiments.

in-vitro & in-vivo characterization of sotorasib-acquired resistant isogeneic cell lines and acquired-resistant cell line-derived xenograft (CDX) development.

A) SRB (Sulforhodamine B) assay was performed to compare the sotorasib sensitivity between H23 (parental) vs H23AR (resistant) cells and H358 (parental) vs H358AR (resistant) cells. The fold of resistance is listed in the table (bottom panel). B) Cell viability assay showed the sensitivity of H23 & H23AR and H358 & H358AR isogeneic cells to adagrasib (top panel) and Opnurasib (bottom panel). C) The strategy for in-vivo resistance to sotorasib for H23AR xenograft tumors is shown. D) Evaluation of in-vivo resistance to sotorasib. H23 and H23AR cells were implanted into mice and treated with sotorasib for at least 3 weeks when tumor sizes reached approximately 200 mm3. H23AR cells were expanded in culture under sotorasib pressure before implantation. The antitumor effect of sotorasib was evaluated in both H23 xenograft (control tumor vs treated tumors) and H23AR xenograft tumor models (control vs sotorasib). Statistical significance was calculated. E) The antitumor effect of sotorasib on H358 parental tumors was evaluated and their statistical significance was calculated. F) The strategy shown here for the development of sotorasib-acquired resistant H358AR CDX. G) Third-generation sotorasib-resistant H358AR CDX showed no significant effect with sotorasib treatment whereas H358 parental xenograft tumors showed significant sensitivity to sotorasib. H) Sotorasib antitumor effect was calculated based on tumor volume shrinkage after 21 days of treatment where H358AR CDX showed no significant effect of sotorasib treatment. Each experiment was repeated three times. Statistics are shown at a significance level of p<0.05 unless otherwise noted. Data is shown as mean percentage ±SD, n>5 mice/group was used in each experiment. *, P < 0.05; **, P <0.005; *** P <0.0005.

Whole exome sequencing (WES) on acquired resistant cells.

Two sotorasib-resistant isogeneic cell lines were subjected to WES to identify additional mutations in KRAS. Three biological replicates for each sample were used. Top left; number of SNPs and Indels found in H358AR as compared with the parental H358 cells. Top right; SNPs and Indels found in H23AR cells as compared with H23 cells. Bottom; KRASG12C mutation is shown present in both H358 and H358AR cells.

Reactivation of KRAS and Upregulation of PI3K/AKT/mTOR Signaling in Acquired Sotorasib-Resistant PDX Models.

Acquired resistant TC303AR and TC314AR PDXs and their isogeneic TC303 and TC314 PDXs snap frozen samples underwent global and phospho-proteomics by mass spectrometry. The data were analyzed in the Biostatistics Department in the MD Anderson Cancer Center. A-B) Dendrogram plot and two-component analysis curve showed the clustering of the samples indicating the similarities and distances among samples. C) Heat-map was generated among the proteins that were statistically significantly upregulated or downregulated between parental TC314 and resistant TC314AR PDX samples (right panel) and parental TC303 vs resistant TC303AR PDX samples (left panel). D) Pathway Enrichment analysis was performed among significantly altered global proteins in TC314AR samples. A list of gene-sets enriched was shown in the top panel with their FDR values, and the enrichment of gene-sets associated with mTORC1 and KRAS signaling are shown in the bottom panel with their significant p-value and FDR. E) Enrichment analysis was performed for global proteomics for combined TC303AR and TC314AR data. A list of enriched gene-sets was listed in the right panel and the mTORC1 signaling enrichment graph was shown in the left panel. F) Pathway Enrichment analysis in phosphoproteomics in TC314AR PDX was performed. The enriched mTORC1 signaling is shown in the left panel, and the PI3K/AKT/mTOR pathway enrichment is shown in the right panel. FDR and p-values are displayed in the graphs. G) RPPA data derived from TC303AR and TC314AR isogenic PDX samples and the significantly altered proteins associated with MAPK and PI3K/AKT/mTOR pathways were analyzed and compared. The increased expression by fold-change of these proteins between TC303 parental vs TC303AR resistant samples (top) and TC314 parental vs TC314AR resistant samples (bottom) were shown. H) PDXs TC247, TC314, & TC453 were implanted into mice and when PDX tumor volume reached 200 mm3, tumors were treated with sotorasib (100 mg/kg) for 3 weeks. After 21 days of treatment, the treatment efficacy was determined based on percentage of tumor growth inhibition (details in method section) and categorized between responders and non-responders. PDX samples underwent RPPA analysis. Heat-maps (bottom panel) were generated based on the significantly altered proteins and compared between responder vs non-responder groups. Non-responder PDXs were grown again and treated with sotorasib for 21 days. After treatment, tumors were harvested and RPPA was performed. The RPPA data derived from sotorasib treated non-responders and control non-responders were compared and a heat-map was generated among significantly altered proteins between control vs treated among non-responder PDXs (top panel). N≥3 biological replicates were used for each PDX for Mass Spectrometry, and RPPA, N=5 PDX-tumors/treatment group and N ≥ 3 PDX samples from each treatment group used for RPPA analysis. 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).

Venn Diagram.

Mass spectrometry was performed for two sotorasib resistant isogenic PDXs (TC303AR and TC314AR) samples. Altered proteins from each pair of isogenic PDXs were used for the Venn Diagram. The percentage of commonly altered proteins is shown in the middle (white area).

Mass spectrometry analysis for phosphoproteomics in resistant PDXs.

Mass spectrometry was performed for two sotorasib resistant isogenic PDXs (TC303AR and TC314AR) samples. Phosphoproteomics data was analyzed for enrichment pathway analysis. Enrichment graph for PI3K/AKT/mTOR signaling in TC303AR PDX (Left) and Enrichment graphs for mTORC1 and PI3K/AKT/mTOR signaling in TC314AR+TC303AR PDXs combined data analysis (middle and right).

KRAS Reactivation in Sotorasib-Resistant Cells via ERK Phosphorylation.

H23, H23AR and H358 & H358AR cells were compared for their p-ERK1/2 expression. Both H23AR and H358AR resistant cells were cultured and maintained under sotorasib pressure. Western blots for total ERK and p-ERK in these cells are shown.

Pharmacological inhibition of PI3K signaling by copanlisib inhibits colony formation and downregulates PI3K/AKT/mTOR signaling molecules in acquired resistant cells and PDXOs.

Both acquired resistant H23AR and H358AR were cultured and maintained under sotorasib pressure. A) SRB (Sulforhodamine B) assay was performed to compare the copanlisib sensitivity between H23 (parental) vs H23AR (resistant) cells and H358 (parental) vs H358AR (resistant) cells. IC50 values of copanlisib for each cell line were shown at the bottom of each respective graph. B) H23AR and H358AR cells were cultured in sotorasib containing media and plated for copanlisib treatment in different doses for 24h. For no treatment control samples, H23AR and H358AR were omitted from sotorasib for the last 24h, otherwise acquired cells were always maintained under sotorasib pressure. Western blot was performed for PI3K (p85), AKT and phospho-AKT (S473 & T308). C) Cell viability assay for copanlisib sensitivity on acquired resistant PDXO303AR and PDXO314AR isogeneic organoids and primary resistant organoids (PDXO780 & PDXO664). Cell viability of copanlisib was also compared with sotorasib on the sotorasib primary resistant organoids PDXO481, PDXO551, PDXO723. D) Inhibition of colony proliferation by copanlisib treatment at different doses on H23AR and H358AR isogeneic cells was shown in colony formation assay. E) Effect of copanlisib and MK2206, and F) copanlisib and everolimus combination on the inhibition of cell proliferation on both resistant H23AR and H358AR cells shown in colony formation assays. G-H) H23AR and H358AR cells cultured, maintained, and seeded in media containing sotorasib and cells were treated with copanlisib doses mentioned in the figure for 24h. For no treatment control samples, H23AR and H358AR were omitted from sotorasib for the last 24h, otherwise acquired cells were always maintained under sotorasib pressure. Western blot was performed for upstream (G) mTOR, p-mTOR, PDK, p-PDK and p-PTEN and downstream (H) p70S6, p-S6, p-GSK-3b, p-PRAS40, p-4E-BP1 molecules. I) Cell viability assay on PDXO780 & PDXO481 organoids to compare the sensitivity between copanlisib and everolimus. Each experiment was repeated three times. Data is shown as mean percentage ±SD, n=3.

PI3K knockout via CRISPR-Cas9 restores sotorasib sensitivity in acquired resistant cells by dephosphorylating 4E-BP1.

A) PI3K/p110α subunit was knocked out using CRISPR-Cas9 technology on H23AR. Five clones H23ARPI3K-KO-2-15, H23ARPI3K-KO-2-17, H23ARPI3K-KO-3-3, H23ARPI3K-KO-3-12 & H23ARPI3K-KO-2-1 from H23AR cells were selected and verified the PI3K expression on these clones by western blots. B) Cell viability assay was performed on these PI3K knockout clones generated from H23AR cells to compare the sotorasib sensitivity to determine the most sensitive and least sensitive clone. The sotorasib sensitivity on PI3K clones was also compared with the parental H23 and resistant H23AR cells. The statistical significance was compared among the clones and H23AR isogeneic cells shown in the table below the graph. C) PI3K/p110α subunit was knocked out by CRISPR-Cas9 technology on H358AR. Three clones H358ARPI3K-KO-1-4, H358ARPI3K-KO-2-4, H358ARPI3K-KO-2-7 from H358AR cells were selected and the knockout status of PI3K expression on these clones was verified by western blots. D) Cell viability assay was performed on these PI3K knockout clones generated from H358AR cells to compare the sotorasib sensitivity to H358AR resistant and H358 parental cells. E) H23, H23AR and the most sensitive PI3K knockout clone H23ARPI3K-KO-2-17 were seeded and treated with copanlisib for 24h and the downstream PI3K, AKT, p-AKT, p-S6 and p-4E-BP1 status was determined by the western blot. Effect of copanlisib treatment on downstream molecules and the level of expression was also compared among PI3K knockout and H23AR isogeneic cells. F) H23, H23AR and the least sensitive PI3K knockout clone H23ARPI3K-KO-3-3 were seeded and treated with copanlisib for 24h and the downstream PI3K, AKT, p-AKT, p-S6 and p-4E-BP1 status was determined by the western blot. The effect of copanlisib treatment on downstream molecules and the level of expression was also compared among PI3K knockout and H23AR isogeneic cells is shown. Each experiment was repeated three times. Data is shown as mean percentage ±SD, n ≥ 3.

mTORC1/2 mediated phosphorylation of 4E-BP1 is associated with acquired resistance and 4E-BP1 knockout via CRISPR-Cas9 restores sotorasib sensitivity.

A) The level of expression of p-GSK-3B, p-S6, p-4E-BP1 and 4E-BP1 in H23 parental, H23AR resistant, and a series of PI3K knockout clones H23ARPI3K-KO-2-15, H23ARPI3K-KO-2-17, H23ARPI3K-KO-3-3, H23ARPI3K-KO-3-12 was compared (left panel). The sotorasib sensitivity was compared among the most sensitive H23ARPI3K-KO-2-17, the least sensitive H23ARPI3K-KO-3-3 clones and H23AR isogeneic cells (right panel). The association between 4E-BP1 expression and sotorasib sensitivity among sensitive and resistant clones was determined. B) 4E-BP1 knockout clones were generated by the CRISPR-Cas9 technology on H23AR cells. The clones were confirmed based on the complete absence of 4E-BP1 expression in western blots. H23AR4E-BP1-KO-1-1 clone was selected after screening several clones. C) Similarly, 4E-BP1 knockout clones were generated by CRISPR-Cas9 technology on sotorasib least sensitive H23ARPI3K-KO-3-3 cells to make the double PI3K and 4E-BP1 knockout clones. Several clones were screened for their 4E-BP1 expression and H23ARPI3K-KO-3-3/4E-BP1-KO-1-2, H23ARPI3K-KO-3-3/4E-BP1-KO-1-4, H23ARPI3K-KO-3-3/4E-BP1-KO-3-4 clones were selected based on no expression of 4E-BP1. D) Cell viability assay was performed for sotorasib sensitivity on 4E-BP1 knockout clone H23AR4E-BP1-KO-1-1 that was compared with parental H23 and H23AR cells. Statistical significance was determined based on p-value calculation, which is shown in the table below the graph. E) Cell viability assays were performed for sotorasib sensitivity on 4E-BP1 knockout clones H23ARPI3K-KO-3-3/4E-BP1-KO-1-2, H23ARPI3K-KO-3-3/4E-BP1-KO-1-4, H23ARPI3K-KO-3-3/4E-BP1-KO-3-4 and their sensitivity was compared with that of resistant H23AR cells. Statistical significance was determined based on p-value calculation, which is shown in the table below the graph. F) The effect of everolimus, an mTORC1 inhibitor on expression of p-S6 and p-4E-BP1 (S65) in resistant H23AR and sensitive PI3K knockout clone H23ARPI3K-KO-2-17 treated with everolimus for 24h. G) The effect of AZD8055, an mTORC1/2 dual inhibitor, on the expression of p-S6, p-4E-BP1 (S65) and p-4E-BP1 (T37/46) on parental H23, resistant H23AR, sotorasib most sensitive PI3K knockout clone H23ARPI3K-KO-2-17 and the least sensitive PI3K knockout clone H23ARPI3K-KO-3-3 after AZD8055 treatment for 24h. H) A cell viability assay was performed for AZD8055, mTORC1/2 dual inhibitor, and its combination with sotorasib on H23, H23AR, H358AR and H23ARPI3K-KO-3-3 clone. Drug synergy was calculated based on their combination index (CI) on both H23AR and H358AR cells. Combination Index (CI) is synergistic if the CI value is <0.7 (synergism if CI = 0.3-0.7, strong synergism if CI = 0.1-0.3, very strong synergism if CI <0.1). I) Cell viability assay on H23AR and H358AR cells to compare the cytotoxicity effect among Everolimus, AZD8055, and sotorasib. Each experiment was repeated three times. Data is shown as mean percentage ±SD, n ≥ 3.

Overcoming resistance by targeting PI3K-AKT-mTOR signaling with copanlisib or sapanisertib in acquired resistant NSCLC PDX, CDX, and xenograft Tumors.

A) In vitro cell viability assay was performed to evaluate the synergistic effect of sotorasib + copanlisib on H23AR, H358AR and H23ARPI3K-KO-3-3 cells. Cells were treated with sotorasib, copanlisib, and their combination. In combination, very low fixed copanlisib dose (either 0.05µM or 0.03µM) was combined with variable doses of sotorasib and cell viability assays were performed after 72h of treatment. The combination index (CI) was calculated and presented in the table listed under the graph. Combination Index (CI) is synergistic if CI value is <0.7 (synergism if CI = 0.3-0.7, strong synergism if CI = 0.1-0.3, very strong synergism if CI <0.1). B-D) In vivo antitumor efficacy of copanlisib and sotorasib combination on acquired resistant H358AR CDX tumors. Mice were engrafted with freshly harvested 3rd-generation acquired resistant H358AR CDX tumor tissues into experimental mice. When the tumor volumes reached approximately 200 mm3, mice were randomized into four treatment groups (control, sotorasib (100 mg/kg), copanlisib (6 mg/kg) and sotorasib + copanlisib combination) and treated for at least 3 weeks. B) Treatment strategy is shown, C) Individual mouse response for each group. D) Antitumor effect of copanlisib and sotorasib combination on H358AR CDX tumors. E) The residual tumor tissues from H358AR CDX tumors were harvested and subjected to RPPA analysis. The level of p-Src, p-mTOR, p70.S6, GSK-3b, p-S6, 4E-BP1, p-4E-BP1 (T37/46) and p-4E-BP1 (S65) expression in residual tissues was determined and compared among the treatment groups. F-H) Antitumor effect of sotorasib + copanlisib combination on acquired resistant H23AR xenograft tumors. F) The treatment strategy is shown where 5-8 million cells/mouse were subcutaneously injected. When tumor sizes reached 200 mm3, tumors were treated for 3 weeks with sotorasib (50 mg/kg), copanlisib (6 mg/kg), and the combination. G) Antitumor effect is shown. H) Individual mouse responses were compared with mice from every treatment group. I) Antitumor effect of sotorasib + copanlisib combination on acquired resistant TC314AR PDX tumors. Mice were engrafted with freshly harvested 3rd-generation acquired resistant TC314AR PDX tumor tissues into experimental mice. When the tumor volumes reached approximately 200 mm3, mice were randomized into four treatment groups (control, sotorasib (50 mg/kg), copanlisib (6 mg/kg) and sotorasib+copanlisib combination) and treated for 3 weeks. Tumor growth curves were generated for each treatment group from the beginning to the end of the treatment. J) Individual mouse responses were compared with mice from every treatment group. K) Antitumor effect of sotorasib + sapanisertib combination on acquired resistant TC314AR PDX tumors. Mice were engrafted with freshly harvested 3rd-generation acquired resistant TC314AR PDX tumor tissues into experimental mice. When the tumor volumes reached approximately 200 mm3, mice were randomized into four treatment groups (control, sotorasib (50 mg/kg), sapanisertib (1 mg/kg) and sotorasib + sapanisertib combination) and treated for 3 weeks. Tumor growth curves were generated for each treatment group from the beginning to the end of the treatment. L) The antitumor data were analyzed by a statistical program, CombPDX, to evaluate the synergistic antitumor effect between sotorasib and sapanisertib on this TC314AR PDX model. Top left) shows the individual mouse response. Top right) Treatment effect for combination, sotorasib alone, and sapanisertib alone treatment. The treatment effect was calculated based on percentage of tumor growth inhibition with 95% confidence interval. Bottom panels) The combination index (CI) was calculated by using two different methods, including Highest Single Agent analysis (HAS) and Bliss Independent (BI) analysis with 95% confidence interval. In-vivo experiments were repeated three times with 5-7 mice/group used in each experiment. Statistics are shown at a significance level of p<0.05 unless otherwise noted. Data is shown as mean percentage ±SD, n=5. *, P < 0.05; **, P <0.005; *** P <0.0005.

Treatment effect on TC314AR PDX by sotorasib + Copanlisib combination treatment.

The resistant TC314AR PDX tumors were treated with the combination and single agents. The antitumor effect data were analyzed by CombPDX analysis software to determine treatment effect. The treatment effect is measured as the percentage of tumor growth inhibition with 95% confidence interval.

The analysis of the synergistic effect of sotorasib + copanlisib combination treatment on TC314AR PDX tumors.

The resistant TC314AR PDX tumors were treated with the combination and single agents. The synergistic antitumor effect was calculated by analyzing the combination index (CI) analyzed by CombPDX analysis software to determine the synergy between these two drugs. Highest single agent method with 95% confidence interval was used to determine the combination index (CI). CI values <1 are synergistic.

Cell viability assay for sotorasib + sapanisertib combination on H23AR and H358AR cells.

The experiment was done three times. Data is shown as mean percentage ±SD, n=3

Cell viability assay for sotorasib + sapanisertib combination on PDXO303AR and PDXO314AR organoids.

The organoids were treated with sotorasib in combination with sapanisertib. Combination index (CI) for cell viability was calculated using the CalcuSyn software using Chou-Talalay method. If the CI value is <0.7, the combination is synergistic, if the CI value is between 0.3-0.1, the combination is strongly synergistic, and if CI value is <0.1, the combination is very strongly synergistic. The experiment was done three times. Data is shown as mean percentage ±SD, n=3