The Cancer/Testes (CT) Antigen HORMAD1 promotes Homologous Recombinational DNA Repair and Radioresistance in Lung adenocarcinoma cells

Abstract

The Cancer/Testes (CT) Antigen HORMAD1 is germ cell-restricted and plays developmental roles in generation and processing of meiotic DNA Double Strand Breaks (DSB). Many tumors aberrantly overexpress HORMAD1 yet the potential impact of this CT antigen on cancer biology is unclear. We tested a potential role of HORMAD1 in genome maintenance in lung adenocarcinoma cells. We show that HORMAD1 re-distributes to nuclear foci and co-localizes with the DSB marker γH2AX in response to ionizing radiation (IR) and chemotherapeutic agents. The HORMA domain and C-term disordered oligomerization motif are necessary for localization of HORMAD1 to IR-induced foci (IRIF). HORMAD1-depleted cells are sensitive to IR and camptothecin. In reporter assays, Homologous Recombination (HR)-mediated repair of targeted ISce1-induced DSBs is attenuated in HORMAD1-depleted cells. In Non-Homologous End Joining (NHEJ) reporter assays, HORMAD1-depletion does not affect repair of ISce1-induced DSB. Early DSB signaling events (including ATM phosphorylation and formation of γH2AX, 53BP1 and NBS1 foci) are intact in HORMAD1-depleted cells. However, generation of RPA-ssDNA foci and redistribution of RAD51 to DSB are compromised in HORMAD1-depleted cells, suggesting that HORMAD1 promotes DSB resection. HORMAD1-mediated HR is a neomorphic activity that is independent of its meiotic partners (including HORMAD2 and CCDC36. Bioinformatic analysis of TCGA data show that similar to known HR pathway genes HORMAD1 is overexpressed in lung adenocarcinomas. Overexpression of HR genes is associated with specific mutational profiles (including copy number variation). Taken together, we identify HORMAD1-dependent DSB repair as a new mechanism of radioresistance and a probable determinant of mutability in lung adenocarcinoma.

Figure 1

HORMAD1 redistributes to nuclear foci and co-localizes with the DNA DSB marker γH2AX in genotoxin-treated cancer cells. (a) Illustration depicting spatiotemporal organization of various CT antigens including HORMAD1, HORMAD2, SPO11, SYCE1 and SYCP1 during prophase I of meiosis (adapted from Bolcun-Filas and Schimenti, 2012). At the first meiotic division homologous chromosomes are joined via at least one crossover during the first prophase. Cross-overs are mediated via homologous recombination (HR) between the paired (homologous) chromosomes and the HR process is initiated via SPO11-induced DSB. A homology search juxtaposes the homologous chromosomes along their lengths and recombination is facilitated by the formation of the chromosome axis and the synaptonemal complex (SC). HORMADs associate with the unsynapsed chromosome axes and promote DNA DSB formation by the Spo11 endonuclease. (b) Immunoblot showing relative levels of HORMAD1 in various lung adenocarcinoma (H1299, A549, H2228, H358, H1359) and breast cancer (SUM159, MDA-MB436) cell lines. Please note that the protein sample in the A549 lane was from the same gel and immunoblot used to analyze all the other samples. An intervening ‘empty’ lane was excised from the digital image. (c) Plasmids encoding V5 epitope-tagged HORMAD1, HORMAD2, SPO11, SYCE1 and SYCP1 were transfected into H1299 lung carcinoma cells. 48 h later the transfected cells were irradiated (10 Gy) or sham-treated and 1 h later the subcellular distribution of each CT antigen in relation to γH2AX was analyzed by confocal microscopy. We used the microscopy image analysis software IMARIS (see Materials and Methods) to empirically measure and confirm that HA-HORMAD1 co-localized with γH2AX as shown in Supplementary Fig. S1. (d) HA-HORMAD1 was expressed in H1299 cells using a recombinant adenovirus. 24 h post-infection, some cultures were treated with 20 μM KU55933 for 1 h. Control and KU55933-treated cells were conditionally irradiated (10 Gy) and 1 h later the subcellular distribution of HA-HORMAD1 in relation to γH2AX was analyzed by confocal microscopy. The bar graph summarizes results of two independent experiments in which 100 cells were scored for HORMAD1 IRIF in the absence and presence of ATM inhibitor. Error bars indicate the standard deviation of results from two independent experiments. Quantification of Hormad1-53BP1 colocalization is presented in Supplementary Fig. 1B. (e) HA-HORMAD1 was expressed in H1299 cells using a recombinant adenovirus. 24 h post-infection, cultures were treated with various DNA-damaging agents. The subcellular distribution of HA-HORMAD1 in relation to γH2AX was analyzed by confocal microscopy at specific times (indicated in parentheses) after each genotoxin treatment: 5 μg/ml Cisplatin (6 h), 100 nM Camptothecin (6 h), 100 μM Etoposide (6 h), 10 Gy IR (1 h), 20 J/m² UVC (6 h). For each genotoxin treatment the treatment conditions selected (dose, time) are ones that we and others have shown are associated with DSB formation. The average Pearson’s correlation coefficients for co-localization of Hormad1 with γH2AX in response to different treatments are as follows: 0.7 for cisplatin, 0.58 for Camptothecin, 0.72 for Etoposide, 0.67 for IR, and 0.53 for UV. The images shown in panel E are representative of nuclei with focal patterns that were observed in 2 independent experiments.

Figure 2

Defining functional domains of HORMAD1 in the DNA damage response. (a) Disorder probablility profile of HORMAD1 (predicted using the Protein Disorder Prediction System at http://prdos.hgc.jp/cgi-bin/top.cgi) and summary of deletion mutants used in this study. (b) H1299 cells were infected with adenovirus vectors encoding HA epitope-tagged wild-type (WT) and mutant forms of HORMAD1. 24 h later, cells were biochemically fractionated and the resulting soluble and chromatin extracts were analyzed by SDS-PAGE and immunoblotting with anti-HA antibody. The immunoblot shows relative expression of WT and mutant forms of HORMAD1 in soluble and chromatin compartments. (c) H1299 cells were infected with adenovirus vectors encoding HA epitope-tagged wild-type (WT) and mutant forms of HORMAD1. 24 h later, cells were analyzed for HORMAD1 distribution using confocal microscopy. The average Pearson’s correlation coefficients for co-localization of HORMAD1 with γH2AX were 0.50 for HORMAD1 WT, 0.57 for HORMAD1 SS 361,378 > AA, 0.53 for HORMAD1 Δ1-21, −0.23 for HORMAD1 Δ22-220, and 0.007 for Δ373-394. These results indicate that the HORMA domain and the extreme C-terminal amino acids of HORMAD1 are necessary for co-localization of HORMAD1 with γH2AX-containing IRIF. (d) H1299 cells were transfected with a plasmid encoding MYC-CCDC36. 24 h post-transfection, some cultures were irradiated (10 Gy) and 1 h later the subcellular distribution of CCDC36 in relation to 53BP1 was was analyzed by confocal microscopy. (e) H1299 cells were transfected with a plasmid encoding MYC-CCDC36 individually or in combination with an HA-HORMAD1 expression vector. 48 h later, cells were treated with IR (10 Gy) or were sham-irradiated. 1 h after irradiation cells were harvested and extracts were immunoprecipitated using MYC antibodies. MYC immune complexes or total ‘Input’ fractions were analyzed by SDS-PAGE and immunoblotting with anti-HA and anti-MYC antibodies.

Figure 3

HORMAD1 promotes Homologous Recombination and DNA damage tolerance in cancer cell lines. (a) Replicate cultures of H1299 cells were transfected with siRNAs targeting HORMAD1 or BRCH1 (or with non-targeting control oligonucleotides). Transfected cells were treated with the indicated dose range of IR and DNA damage sensitivities were measured by clonogenic survival assays. The lower panel is an immunoblot showing relative levels of HORMAD1 expression 48 h post transfection in the siCon- and siHORMAD1-treated cells. (b) Replicate plates of H1299 cells harboring the stably-integrated DR-GFP reporter construct were transfected with siRNAs (against HORMAD1, BACH1, or non-targeting siRNA). 24 h later the siRNA-treated cells were transfected with an ISceI expression plasmid (to induce DSB in the DR-GFP locus) or with an empty control vector. After 24 h cells were trypsinized and GFP-expressing populations (resulting from HR-mediated reconstitution of a silent GFP allele) were enumerated by flow cytometry. Supplementary Fig. S3A shows the original flow cytometry profiles corresponding to these HR assays. In QPCR analyses of mRNA from siBACH1-transfected cells, endogenous BACH1 transcript levels were reduced by 62% relative to control (siCon) cultures. (c) Replicate plates of H1299 cells harboring the stably-integrated EJ5-NHEJ reporter construct were transfected with siRNAs (against HORMAD1, LIG4, or non-targeting siRNA). Some cultures were also treated with the DNA-PK inhibitor NU7441 (20 μM). 24 h later the siRNA or DNA-PKi-treated cells were transfected with an ISceI expression plasmid (to induce DSB in the stably integrated reporter locus) or with an empty control vector. After 24 h cells were trypsinized and GFP-expressing cells (arising from NHEJ-mediated reconstitution of a silent GFP allele) were enumerated by flow cytometry. In QPCR analyses of mRNA from siLIG4-transfected cells, LIG4 transcript levels were reduced by 95% relative to control (siCon) cultures. (d) H1299 cells were transfected with siRNAs targeting HORMAD1 or with non-targeting control oligonucleotides. 48 h later cells were treated with 10 μM BrdU for 1 h to label actively-replicating DNA. Cells were then collected and stained sequentially using a FITC-labeled anti-BrdU antibody and PI. The labeled nuclei were analyzed by flow cytometry. The BrdU-positive (S-phase) populations are indicated by the dashed quadrant. (e) Replicate plates of H1299 cells harboring the stably-integrated DR-GFP reporter construct were transfected with siRNAs against CCDC36, BRCA1, or with non-targeting siRNA (siCon). 24 h later the siRNA-treated cells were transfected with an ISceI expression plasmid and HR activity was measured based on enumeration of GFP-positive cells as described for panel (b) above. In QPCR analyses of mRNA from siBRCA1-transfected cells, BRCA1 transcript levels were reduced by 96% relative to control (siCon) cultures. (f,g) Replicate cultures of H358 cells were transfected with siRNAs targeting HORMAD1 or BRCA2 (or with non-targeting control oligonucleotides). In QPCR analyses of mRNA from siBRCA2-transfected cells, levels of the endogenous BRCA2 transcript were reduced by 92% relative to control (siCon) cultures. Transfected cells were treated with the indicated dose ranges of camptothecin (f) or IR (g) and DNA damage sensitivities were measured by clonogenic survival assays. (h) Replicate cultures of H1299 cells were transfected with siRNAs targeting HORMAD1 or BRCA1 (or with non-targeting control oligonucleotides). Transfected cells were treated with the indicated dose ranges of olaparib and viability was determined by clonogenic survival assays. Error bars throughout this figure represent the standard error of the mean from three independent experiments that were performed in triplicate.

Figure 4

HORMAD1 promotes DSB resection to promote HR. (a) H1299 cells were transfected with siCon or siHORMAD1. 48 h post-transfection cells were irradiated (10 Gy) and 1 h later nuclei were analyzed for RAD51 and γH2AX distribution using confocal microscopy. The number of cells containing RAD51 IRIF was reduced to 33% in HORMAD1-depleted cells when compared with control cultures (p < 0.001). Error bars represent the standard deviation and were derived from three independent experiments. (b) Replicate cultures of control, HORMAD1-depleted or BRCA1-depleted cells were irradiated (10 Gy). At different times post-IR, chromatin fractions were analyzed by SDS-PAGE and immunoblotting with the indicated antibodies. (c) H1299 cells were reverse transfected with siCon or siHORMAD1-5′UTR (which targets the untranslated region of the endogenous HORMAD1 mRNA but not the ectopically expressed HORMAD1 transcript). 16 h post-transfection cells were infected with advenovirus vectors encoding WT and mutant forms of HORMAD1. 48 hours post infection, cells were irradiated (10 Gy) and harvested 1 h later for SDS-PAGE and immunoblotting analysis with the indicated antibodies. (d) H1299 cells were transfected with siCon, siHORMAD1, or siCtIP. 48 h post-transfection cells were irradiated (10 Gy) and 1 h later nuclei were analyzed for RPA34 and 53BP foci using immunofluorescence confocal microscopy. 53BP1 focus formation was not significantly affected by HORMAD1 depletion. The experiment shown here was performed in parallel with the experiment presented in Supplementary Fig. 4. Immunoblots shown in Supplementary Fig. S4 validates efficient ablation of HORMAD1 and CtIP expression by these siRNAs. The histogram shows quantification of RPA IRIF data from multiple experiments. The percentage of cells containing RPA foci is counted. Depletion of CtIP or HORMAD1 significantly reduced the RPA- positive population of IR-treated cells by approximately 50% (relative to control siRNA transfected cultures). Error bars represent the standard deviation and were derived from three independent experiments.

Figure 5

Genomic Correlates of HORMAD1 Expression in the TCGA Lung Adenocarcinoma (LUAD) dataset. (a) Boxplot of normalized HORMAD1 RNA expression in tumor and tumor-adjacent normal tissue samples in the LUAD samples. (b,c) Correlation between RNA expression and genomic copy number variation across all expressed. genes in LUAD and breast cancer samples respectively. HORMAD1 is indicated in green, other known HR genes are indicated in red, and NHEJ genes are indicated in blue. (d) Boxplot of genomic copy number variations by high vs low HORMAD1 expression. (e) Barplot of deletion frequency and length by high vs low HORMAD1 expression. (f) Heatmap of correlations between HORMAD1 and gene signature scores across DNA repair pathway gene set. (g) Plot showing correlation of HORMAD1 RNA expression and other CTAs in the LUAD dataset, showing no significant correlation between their expressions. (h) Kaplan-Meier plot of LUAD patients stratified by HORMAD1 expression.

Figure 6

Scheme depicting role of HORMAD1 in DSB signaling and repair. Our immunofluorescence microscopy analyses (Fig. 4) show that formation of CtIP, RAD51 and RPA p34-containing IRIF is reduced in HORMAD1-depleted cancer cells (whereas 53BP1 and p95/NBS1 foci remain intact in the absence of HORMAD1). We infer that HORMAD1 contributes to HR pathway activation at (or prior to) the CtIP-mediated resection step, but does not affect the proximal stages of DSB signaling. HORMAD1-depletion leads to attenuation of DR-GFP/HR reporter activity (Fig. 3), prolonged IR-induced ATM signaling (Fig. 4), and radiosensitivity (Fig. 3), further demonstrating that HORMAD1 promotes the HR pathway.