Targeting genome integrity dysfunctions impedes metastatic potency in non-small cell lung cancer circulating tumor cell-derived explants

Abstract

DNA damage and genomic instability contribute to non-small cell lung cancer (NSCLC) etiology and progression. However, their therapeutic exploitation is disappointing. CTC-derived explants (CDX) offer systems for mechanistic investigation of CTC metastatic potency and may provide rationale for biology-driven therapeutics. Four CDX models and 3 CDX-derived cell lines were established from NSCLC CTCs and recapitulated patient tumor histology and response to platinum-based chemotherapy. CDX (GR-CDXL1, GR-CDXL2, GR-CDXL3, GR-CDXL4) demonstrated considerable mutational landscape similarity with patient tumor biopsy and/or single CTCs. Truncal alterations in key DNA damage response (DDR) and genome integrity-related genes were prevalent across models and assessed as therapeutic targets in vitro, in ovo, and in vivo. GR-CDXL1 presented homologous recombination deficiency linked to biallelic BRCA2 mutation and FANCA deletion, unrepaired DNA lesions after mitosis, and olaparib sensitivity, despite resistance to chemotherapy. SLFN11 overexpression in GR-CDXL4 led to olaparib sensitivity and was in coherence with neuroendocrine marker expression in patient tumor biopsy, suggesting a predictive value of SLFN11 in NSCLC histological transformation into small cell lung cancer (SCLC). Centrosome clustering promoted targetable chromosomal instability in GR-CDXL3 cells. These CDX unravel DDR and genome integrity-related defects as a central mechanism underpinning metastatic potency of CTCs and provide rationale for their therapeutic targeting in metastatic NSCLC.

Keywords: Lung cancer; Mouse models; Oncology.

Figure 1. Establishment and characterization of CDX and CDX-derived cell lines.

(A) Schematic of available patient samples and established CDX and CDX-derived cell lines (red cross = not available). (B) CDX tumor growth curves. Indicated number of CTCs was injected in NSG mice. Palpable CDX tumors were obtained after 100, 200, 116, and 100 days in GR-CDXL1, GR-CDXL2, GR-CDXL3, and GR-CDXL4, respectively. (C) IHC characterization of patient L4 TB at baseline and disease progression, GR-CDXL4 CDX tumor and CDX-derived cell line. Representative images of HES, CK8/18, EpCAM, Ki67, Vimentin, TTF1, Chromogranin A, and Synaptophysin stainings are shown at a total magnification of ×200. Scale bar: 10 μm.

Figure 2. Evaluation of CDX-derived cell line metastatic potency in ovo and in vivo.

(A) Metastatic capacity of CDX-derived cell lines in the CAM. mCherry-expressing CDX-derived cells were implanted into the CAM, and metastatic fluorescent signal was analyzed at day 7. Representative fluorescence/CT images of GR-CDXL1–, GR-CDXL3–, and GR-CDXL4–generated tumors are shown (top). Quantitative analysis of average fluorescence intensity (bottom); each point represents a single embryo. (B) Metastatic capacity of CDX-derived cell lines in NSG mice. Luciferase-expressing CDX-derived cells were grafted into NSG mice by IC to generate metastases. Representative BLI images of GR-CDXL1–, GR-CDXL3–, and GR-CDXL4–generated tumors are shown (top). Quantitative analysis of average BLI intensity (bottom); each point represents a single mouse. Data are mean ± SEM; **P ˂ 0.01, ***P < 0.001 by Kruskall Wallis and post hoc Dunn’s test.

Figure 3. Comparative genomic analysis of biopsies and the CDX.

(A) Fraction of TB mutations detected and undetected in the CDX. (B) Fraction of TB driver mutations detected and undetected in the CDX. (C) Mutated driver genes and their amino acid sequence variation in the biopsy and the CDX. (D) Mutated driver genes and amino acid sequence variation in the TB only. (E) Fraction of CDX mutations issued or not from the TB. (F) Fraction of CDX driver mutations issued or not from the TB. (G) Mutated driver genes and amino acid sequence variation in the CDX only.

Figure 4. Heatmap of the CNA analysis of the TB, the CDX, and the CDX-derived cell lines.

Unsupervised hierarchical clustering of CNA profiles was performed. Copy gains are shown in red, and copy losses are shown in blue. CN, copy number.

Figure 5. Phylogeny of CDX and CDX-derived cell lines.

(A–D) Branches are unscaled, and their length is not proportional to the number of alterations occurring in the branch. The number of mutations (in dark) and the CNAs (gain in red and loss in blue) are mentioned on the branches of the tree. Only genes bearing driver truncal or acquired alterations (mutations or CNAs) are indicated.

Figure 6. DNA damage response activation in CDX-derived cell lines.

(A) Representative images of 53BP1 foci (red) in A549 and GR-CDXL3 cells (top). Proportion of S/G2 (cyclinA⁺) cells with more than 5 53BP1 foci (bottom). (B) Representative images of GH2AX⁺ (green) mitotic cells in GR-CDXL1 and A549 (top). Proportion of H2AX⁺ mitoses (bottom). (C) Representative images of 53BP1 NB (red) in A549 and GR-CDXL1 cells (top). Proportion of G1 (cyclinA^–) cells with more than 3 53BP1 NB (bottom). (D) Western blot analysis of the levels of p-CHK1, CHK1, FANCA, and PARP1 in CDX-derived and NSCLC cell lines. (E) Representative images of IR-induced RAD51 foci (red) in S phase (geminin⁺) GR-CDXL1 and GR-CDXL3 cells (top). Proportion of RAD51⁺/geminin⁺ NT and IR cells (bottom). (F) Representative images of APH-induced pRPA32 foci (green) in GR-CDXL1 and GR-CDXL3 cells (top). Level of pRPA32 foci per cell in NT and APH-treated (bottom). (G) Representative images of APH-induced pDNA-PK foci (red) in A549 and GR-CDXL1 cells (top). Level of pDNAPK foci per cell in NT and APH-treated (bottom). For E–G, we kept only A549 NSCLC cell line as a comparator, as the other control had equivalent levels of DNA damage. Data are shown as mean ± SD from at least 3 independent experiments; *P ˂ 0.05, **P ˂ 0.01, ***P ˂ 0.001, ****P < 0.0001 by Kruskall Wallis and post hoc Dunn’s test. NB, nuclear body; NT, nontreated, IR, irradiated; APH, aphidicolin. Scale bar: 10 μm.

Figure 7. Mitotic defects in CDX-derived cell lines.

(A) Metaphase spreads of GR-CDXL1, GR-CDXL3, and GR-CDXL4 chromosomes, shown at a total magnification of ×150. (B) Chromosome numbers. (C) Absolute copy number profiles showing whole-genome duplication of GR-CDXL3 (top) and GR-CDXL4 (bottom) cell lines. (D) IF analyses of mitotic defects in NSCLC and CDX-derived cells; yellow arrow indicates anaphasic bridge, and red arrow indicates lagging chromosome (left). Fraction of mitotic cells presenting defects (right). (E) Representative images of dual α-tubulin/centrin immunostaining revealing clustering of extra centrosomes in GR-CDXL3 cells (white arrows) (left). Proportion of cells presenting centrosome clusters (right). Statistical significance was assessed using Kruskall Wallis and post hoc Dunn’s test for B and E. Data are shown as mean ± SEM; *P ˂ 0.05, ***P < 0.0001 from n = 3 experiments.

Figure 8. In vitro, in ovo, and in vivo drug assays.

(A) Mean in vitro IC₅₀ values of cisplatin for control and CDX-derived cell lines. (B) Mean in vitro IC₅₀ values of olaparib. (C) Western blot showing SLFN11 expression levels in GR-CDXL1, GR-CDXL3, GR-CDXL4, and NSCLC cell lines. (D) qPCR for SLFN11 gene expression in A549 and CDX-derived cell lines normalized to GAPDH expression level. Data are fold change and are shown as mean ± SEM. n = 3 experiments; **P ˂ 0.01 by Kruskall-Wallis and Dunn’s test. (E) IHC staining of SLFN11 in patients L1 (CDX, cell line, and metastatic mouse tumor) and L4 (TB, CDX, cell line, metastatic mouse tumor) samples. Scale bar: 30 μm. (F) Mean in vitro IC₅₀ values of BYL719 for control and GR-CDXL3 cell line. For A, B, and F, data are shown as mean ± SD. n = 3 experiments; *P < 0.05, **P < 0.01, Kruskall-Wallis and Dunn’s test (A and B), 2-tailed unpaired t test with Welch’s correction (F). (G) Three-dimensional representative images at ID17 (left) and quantitative analysis of average fluorescent tumor foci (right) of GR-CDXL1 or GR-CDXL4 mCherry-expressing CAM tumors, treated or not with olaparib. (H) Luciferase-expressing GR-CDXL1 (left, upper panel) or GR-CDXL4 (left, lower panel) tumors treated with olaparib. Representative BLI images (left) and tumor volumes (right) obtained at indicated days of treatment are shown. (I) Tumors at day 32 (GR-CDXL1-Luc) and day 25 (GR-CDXL4-Luc). (J) Three-dimensional representative images obtained at ID17 (left) and quantitative analysis of average fluorescent tumor foci (right) of GR-CDXL3 mCherry-expressing CAM tumors treated or not with AZ82 and/or BYL719. For G and J, data are shown as mean ± SEM. n = 3 experiments; *P ˂ 0.05, **P ˂ 0.01, ***P ˂ 0.001, ****P < 0.0001 by Kruskall-Wallis and Dunn’s test. Each point represents a single embryo. (K) Representative BLI images of GR-CDXL3 luciferase–expressing mouse tumors treated or not with AZ82 and/or BYL719. Tumor volume is shown (lower panel). For H and K, data are shown as mean ± SEM. n = 5; *P ˂ 0.05, ***P ˂ 0.001, ****P ˂ 0.0001 by 2-way ANOVA. (L) GR-CDXL3 luciferase–expressing tumors obtained at day 28.