Auswahl der wissenschaftlichen Literatur zum Thema „Reference-Free detected RNAs“

Abstract Chromosomal rearrangements resulting in expression of oncogenic chimeric proteins drive tumor progression. Gene fusions are promising targets for cancer therapy in various types of cancers. Therefore, faithful detection of gene fusions is essential in precision medicine. Previously, fusions have been detected using FISH, RT-PCR and RNA-seq. Targeted sequencing provides better sensitivity as only the regions associated with the driver genes are sequenced, increasing possible coverage depth, and reducing cost. We developed a targeted RNA-based lung cancer fusion panel utilizing 5’ RACE technology to construct a NGS library for detection of fusion transcripts present in total RNA from highly degraded samples, such as blood or FFPE samples. The panel is deliberately designed to cover 64 known actionable fusion genes as well as novel gene fusions. Using an in-house fusion RNA mix comprising of 52 fusion transcripts, the panel was optimized and balanced. A total of 40 lung cancer patient FFPE samples, one 6-fusion FFPE reference sample and three fusion positive cell line samples were used for validating the panel. Limit of detection (LoD) was determined using the in-house RNA mix and normal lung tissue RNA. The minimum RNA input to detect all 52 fusion transcripts was 10-4 ng (0.1 pg) of mixture of RNAs in background of 10 ng of normal lung tissue RNA. The panel successfully detected all six fusion transcripts with 10 ng of RNA from a 6-fusion reference FFPE sample. With this confirmation, we tested clinically relevant sample types which are damaged RNAs from FFPE material. This study demonstrates the performance of our newly developed NGS fusion gene panel with clinical samples. It accommodates highly degraded RNA samples with low amounts of input. The next step would be to expand the testing to include more samples and cell-free RNAs. Citation Format: Bongyong Lee, Alisha Babu, Aiguo Zhang, Michael Y. Sha. Detection of actionable lung cancer fusion genes with known and novel partners from highly degraded FFPE material [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2191.

Abstract Triple Negative Breast Cancers (TNBCs) are the most aggressive breast cancers and so far, benefited very little from personalized medicine. Unlike in hormone positive breast cancers, coding transcriptomic signatures like Lehmann’s failed to predict which TNBC patients will benefit from perioperative chemotherapy. Here, we analyzed the total transcriptome of a unique cohort of 44 TNBC tumors before neoadjuvant chemotherapy (NAC), including both patients that were chemosensitive (N=26) and chemoresistant (N=18) to following treatment. Differential gene expression analysis (DE-seq) was first performed on annotated genes and de novo Scallop RNA-profiled transcripts. Nearly half of detected differential transcripts were from non-coding genes that were mostly long non-coding (lnc)RNAs (90%). More than half differential lncRNAs were not annotated in gencode. Next, De-kupl, a new reference-free analysis of differential fragments of transcripts, was applied. This approach is based on 31-nucleotide long k-mers that are later merged into contigs without annotation bias. These differential contigs were able to separate correctly, in an unsupervised manner, all TNBC patients regarding their response to chemotherapy, outperforming our previous DE-seq analysis of gencode and scallop transcripts. Strikingly, 80% of the DE-kupl differential contigs were not derived from DE-seq differential genes. Furthermore, as BRCA 1/2 mutated TNBCs are now well known for their chemosensitivity, we excluded them from subsequent analysis and repeated the differential expression analysis. Our results showed that DE-kupl differential contigs once more perfectly classified all patients and outperformed the whole length transcripts from DE-seq. Finally, we applied the LASSO regression analysis machine learning method, on both all misregulated genes and all misregulated contigs in parallel, to generate minimal signatures characterizing pre-chemotherapy TNBC tumors that are going to be chemoresistant. We obtained an 8-contig-signature that overpassed the 11-gene-signature in the accuracy to discriminate chemosensitive and chemoresistant TNBC patients before chemotherapy. These results suggest that a contig based unreferenced differential analysis of TNBCs regarding their response status to NAC provides promising predictive biomarkers of early TNBC chemoresistance and thus need further confirmation in a larger and independent validation cohort. Citation Format: Nouritza Torossian, Dominika Foretek, Marc Gabriel, Linda Larbi Cherif, Charlotte Lecerf, Maud Kamal, Christophe Le Tourneau, Daniel Gautheret, Sergio Roman-Roman, Antonin Morillon. Reference free transcriptomic characterization of chemoresistant triple negative breast cancers provides a promising reservoir of predictive biomarkers of early chemoresistance [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2774.

Abstract BACKGROUND Evaluation of the salivary transcriptome is an emerging diagnostic technology with discriminatory power for disease detection. This study explored massively parallel sequencing for providing nucleotide-level sequence information for each RNA in saliva. METHODS Transcriptome profiling with the SOLiD™ system was applied to RNA isolated from unstimulated cell-free saliva (CFS) and whole saliva (WS) from healthy human volunteers. Sequenced reads were aligned to human genome build 18 and the Human Oral Microbiome Database (HOMD). RESULTS Massively parallel sequencing enabled the acquisition of complete sequence information for each nucleotide position of the human salivary transcriptome through alignment to multiple sequence databases. Approximately 20%–25% of the sequenced reads from CFS aligned to the human genome, and approximately 30% of the sequenced reads aligned to the HOMD. We detected the expression of >4000 coding and noncoding genes in CFS and WS. Gene expression at different genome loci showed that the structural integrity of the transcripts for the annotated genes was preserved in saliva. CONCLUSIONS A single measurement provided RNA sequence information of gene transcript abundance for both coding and noncoding RNAs and identified sequences from >400 different microbial species within a single sample. Contrary to previous data suggesting that salivary RNA is highly fragmented, in our study the structural integrity of RNA was preserved. The high degree of sequence alignment to annotated exons and introns for each of the respective reference genomes, with sequence coverage spanning the full length of the genes, provides strong evidence that the salivary transcriptome is a complex RNA network.

Introduction Within the general population over 50 years, 3% is affected with a monoclonal gammopathy of undetermined significance (MGUS). Although MGUS generally demands no treatment, interest in treating premalignant stages to prevent progression to overt malignancies has grown. Better insight in processes underlying progression of MGUS is therefore warranted. Platelet RNA profiles have been shown to reflect information on the topical tumor environment in various cancer types, including multiple myeloma (MM). Our aim was to investigate whether IgM MGUS patients can be distinguished from healthy controls (HCs) using platelet RNA splicing profiles, and to relate IgM MGUS splicing profiles to genes associated with Waldenström's Macroglobulinemia (WM). Methods Twenty-four IgM MGUS patients with presence of a paraprotein confirmed through immunofixation consented to participation. The thromboSeq protocol for platelet RNA sequencing (RNA-Seq) based classification (Best et al. Nat Protoc 2019) was used to acquire platelet RNA profiles from all MGUS subjects. RNA was extracted from platelets isolated from whole blood. After reverse transcription and SMARTer amplification, 100 bp single-read sequencing was performed on the Illumina Hiseq 2500 platform. Raw sequencing data were processed with Trimmomatic, aligned to a reference genome with STAR and mapped with HTSeq. Mapping data of 29 age and blood storage time matched HCs were also included in the analysis. Differential expression of spliced RNAs expressed as log2 fold change (logFC) was assessed with a likelihood-ratio ANOVA and visualized in a heatmap constructed with differentially spliced RNAs that passed the particle swarm optimized (PSO) false discovery rate (FDR) threshold. RNA profiles were further explored and related to WM and MM by DAVID gene ontology enrichment analyses. Results Mean age in the MGUS and HC groups was 67.2 (SD 7.7) and 66.9 (SD 6.7) years. Thirteen of 24 MGUS patients suffered from IgM mediated anti-myelin associated glycoprotein (MAG) neuropathy; the other 11 were asymptomatic. Of 3,426 high-abundant spliced RNAs detected, 1,371 demonstrated significant differential expression (FDR <0.05) and allowed perfect distinction of MGUS from HCs after PSO (Fig. 1; P <0.0001). The two distinct MGUS phenotypes (neuropathy vs. asymptomatic) could not be distinguished. Top 50 overexpressed RNA isoforms were mainly involved in apoptotic processes, transcription regulation, cellular defense response and immune response regulation. WM associated genes (logFC in parentheses) BCL7C (1.35), CASP8 (1.75), CD79A (2.65), CD81 (1.58), DUSP1 (2.32), HLA-DRB1 (1.83), JAK3 (0.65), NFKBIA (3.59), NFKB2 (1.63), REL (1.35) and TNFRSF14 (0.75), as well as MM associated JUND (4.61) and CCNL1 (3.11), were overexpressed in MGUS. TNFAIP3, negatively regulating the NF-κB pathway, was highly overexpressed (logFC 4.07). JAK1 and JAK2 mRNAs were mildly though significantly depleted in the MGUS group. We also found overexpression of CCNL2 (1.81), MS4A1 (2.17) and WNK1 (0.79), which have been shown to be enriched in WM relative to IgM MGUS. Top 50 depleted RNAs were mainly involved in protein phosphorylation, intracellular signal transduction and response to tumor necrosis factor. CASP3 (-1.43), CASP4 (-0.62) and TNFSF4 (-0.99) mRNAs, of which overexpression is associated with WM, were depleted in MGUS. Conclusions For the first time, we have demonstrated widespread differential expression of spliced RNAs in platelets of MGUS patients as compared to healthy controls. We found evidence of overexpression of various genes that have been associated with WM. The differences were mainly in pathways related to apoptosis, signaling and immune response regulation. This is in line with findings that balanced elevation of free light chains, possibly due to a chronic inflammatory state, increases risk of developing a monoclonal gammopathy (Kumar et al. Blood Cancer J 2019). Overexpression of NFKBIA, NFKB2 andTNFAIP3 in MGUS possiblysuggests activation of the NF-κB pathway, a key step in MYD88 mediated progression to WM. JAK subtype expression levels may reflect that the JAK/STAT pathway, also important for progression to WM, has not (yet) been aberrantly activated in our MGUS subjects. Future research could be focused on studying the development of MGUS and MGUS progression to WM and MM using differentially expressed splice junctions in platelets. Disclosures Minnema: Amgen: Honoraria; Celgene Corporation: Honoraria, Research Funding; Gilead: Honoraria; Jansen Cilag: Honoraria; Servier: Honoraria.

Introduction More than half of patients with relapsed or refractory large B-cell lymphoma (R/R LBCL) will experience progressive disease after CD19 chimeric antigen receptor T-cell therapy (CART). Circulating tumor DNA profiling has been used to prognosticate patients before CART, but mainly has focused on identification of somatic genomic aberrations. Epigenetic changes, specifically aberrant DNA methylation, can alter transcriptional regulation of genes important for treatment resistance. Altered methylation features can be detected from cell-free DNA (cfDNA) in patients with LBCL and are associated with outcomes after frontline immunochemotherapy. In the current study, we used whole genome bisulfite sequencing (WGBS) to analyze the methylation landscape in cfDNA from patient plasma before CART and identified novel methylation signatures associated with poor treatment outcomes. Methods Patients with R/R LBCL who received standard-of-care CART with plasma samples drawn before lymphodepletion (LD) or at cell infusion were identified. The primary endpoint was 3-month overall response after CART. cfDNA was extracted from 3-5 ml of plasma and cfDNA libraries were prepared using a Zymo-Seq cell free WGBS kit. Each library was sequenced to 300 - 600 million read pairs at a read length of 150 bp, enabling an average CpG coverage of 7-13X per detected CpG. Unique alignment rate among all libraries were about or above 82%. Reads were adapter trimmed with Trim Galore and aligned to reference genome hg19 using Bismark. Methylation calling was performed by MethylDackel. Differentially methylated cytosines (DMC, FDR ≤0.05) and regions (DMR) from WGBS data with an absolute methylation difference of ≥ 0.1 between 3-month responders and non-responders were identified using DSS. Functional enrichment analysis of genes overlapped by DMRs was performed by gprofiler2. Results A total of 22 patients with plasma collected before LD (n=21) or at cell infusion (n=1) who received CART were identified. Median follow-up was 21.5 months (95% CI 12.5-NR). Three-month response was evaluable in 21 patients (1 pending evaluation). Overall response rate (ORR) was 77% (8 CR, 9 PR) and 55% (9 CR, 2 PR) at one- and 3-months post-CART, respectively. Median PFS was 6.5 months (95% CI 3.0-NR), and median OS was 57.7 months (95% CI 7.7-NR). Patient and disease characteristics are described in Table 1. Differential methylation analysis identified 22,815 DMCs, of which 1,043 were hypomethylated and 21,772 were hypermethylated in non-responders. The global average methylation values across 25 million cytosines were 0.807 and 0.786 for non-responders and responders, respectively (p=0.265). A total of 582 DMRs were identified, of which 34 were hypomethylated and 548 were hypermethylated in non-responders (Figure 1). Of these, 281 (48%) overlapped intergenic regions, 265 (46%) gene bodies, and 33 (6%) promoter regions. DMRs overlapping DNA repair gene PARP1 (areaStat 88.2), B-cell receptor signaling genes SYK (27.2), FYN (21.4), and PTPN6 (20.9), polycomb repressor complex genes JARID2 (25.2) and BCOR (21.4), tumor suppressor WT1 (47.1), immune mediators NRP1 (80.0), TRAF3 (56.7), PAK3 (28.4), IRAK2 (28.7), and TNFS14 (23.6), and non-coding RNAs MIR4520 (95.8) and VTRNA-2 (23.1) were hypermethylated in non-responders and IL17RA (-33.2) hypomethylated. Functional enrichment analysis on overlapped genes also identified common molecular functions involving guanyl nucleotide exchange factors (p=0.0007), GTPase regulation (p=0.001), and cGMP binding (p=0.007). Discussion This proof-of-concept study is the first to perform WGBS on pretreatment plasma from patients with R/R LBCL to associate methylation signatures with post-CART outcomes. We identified that non-responders were enriched in hypermethylated DMRs compared to responders, overlapping with genes important to lymphomagenesis, immune function, and GTP energy metabolism. Further work will expand this cohort, compare with matched-tissue to find tumor-specific markers, and identify candidate DMR targets for targeted bisulfite sequencing from pre- and post-treatment plasma samples to assess response. By identifying high-risk methylation signatures in the plasma of CART recipients, we may be able to identify patients to be prioritized for novel treatment strategies such as combination with hypomethylating agents to improve outcomes.

Abstract Introduction: Colorectal cancer (CRC) ranks second in cancer morbidity, with approximately 60% of patients being diagnosed at an advanced stage. Here, we apply a multiomic strategy jointly leveraging plasma cell-free DNA (cfDNA) and multiple extracellular vesicle (EV)-derived analytes to enhance CRC diagnostic accuracy. Materials and Methods: Plasma was collected from 96 individuals, 48 CRC patients (stage II-IV) and 48 CRC-negative healthy controls. EVs and cfDNA were isolated from the same plasma sample to look at complementary information within the sample. We developed an EV RNA sequencing platform targeting mRNAs and long non-coding RNAs and sequenced to a depth of 50M reads per sample; cfDNA methylome profiling was sequenced to an equivalent depth. The Olink platform profiled proteins both directly from plasma and from isolated EVs. Expression of EV-derived RNA, splice variants, and proteins in addition to cfDNA methylation patterns were analyzed using Bio-Techne’s multiomic platform. Machine learning-based feature selection algorithms identified biomarker signatures from each analyte. Receiver-operator characteristic curves (ROC) were generated utilizing leave-one-out cross-validation of naïve Bayes classifier models to compute the area under the curve (AUC). Individual signatures were integrated to generate a multiomic classifier. Results: Differential gene expression (DEx) analysis identified ~1000 DEx genes with significant enrichment of CRC-relevant pathways. Splice variant analysis identified several genes previously implicated in CRC with differential isoform usage. Over 1.7 million differentially methylated bases were detected between CRC-positive and CRC-negative cohorts. Segmentation of the genome using out-of-bag reference data yielded ~1100 genomic segments, many of which overlap previously reported CRC biomarkers and identified unique information about the CRC status. Proteomic analysis identified several differentially abundant markers with high discriminatory power. Following feature selection analysis, AUCs obtained from top 10 features of RNA expression, splice variants, cfDNA methylation, and protein abundance with an integrative analysis identified a multi-analyte signature, with an overall AUC of 0.99. Conclusion: While biomarker signatures obtained from each analyte in this study resulted in effective separation of CRC status, the multiomic signature further improved the discriminatory power—underscoring the complementary nature of the signals. Thus, a multiomic biomarker discovery strategy leveraging cfDNA and EV cargo holds tremendous promise for minimally invasive screening and potential detection of early-stage cancers, which is currently under further investigation. Citation Format: Kyle Manning, Dulaney Miller, Yang Yang, Jeff Cole, Shuran Xing, Christopher Benway, Christian Ray, Sudipto Chakrabortty, Brian Haynes, Seth Yu, Johan Skog. Exosome based multiomics combined with cfDNA methylation reveals complementary signatures in blood based liquid biopsy that carry promise for minimally invasive colorectal cancer screening [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(7_Suppl):Abstract nr LB393.

BACKGROUND MicroRNAs (miRs) are small, noncoding RNAs that target genes involved in tumor development and progression. In the current study, we investigated the use of circulating miR concentrations as biomarkers in the serum of breast cancer patients. METHODS We analyzed serum samples from 120 patients with primary breast cancer after surgery and before chemotherapy (M0, classified into 3 subgroups of 40 patients with progesterone/estrogen-positive, HER2-positive, and triple-negative cancer), 32 patients with overt metastasis (M1), and 40 healthy women. Using quantitative TaqMan MicroRNA PCR, we measured the relative concentrations of 6 circulating microRNAs (miR-10b, -17, -34a, -93, -155, and -373) known to be relevant for tumor development and progression. The data were correlated with clinicopathologic risk factors, with particular reference to HER2 and hormone receptor status of the primary tumor and the presence of metastases. RESULTS The relative serum concentrations of circulating miR-34a [P = 0.013, area under the curve (AUC) 0.636], miR-93 (P = 0.001, AUC 0.699), and miR-373 (P = 0.0001, AUC 0.879) were significantly different between M0 breast cancer patients and healthy women, whereas miR-17 (P = 0.002, AUC 0.679) and miR-155 (P = 0.0001, AUC 0.781) were differently expressed between M0 and M1 patients. Increased concentrations of miR-373 were associated with negative HER2 status of the primary tumor (P = 0.0001). Deregulated concentrations of miR-17 (P = 0.019) and miR-34a (P = 0.029) were detected in patients with progesterone/estrogen receptor–positive and –negative status, respectively. CONCLUSIONS Our findings indicate that serum concentrations of deregulated microRNAs may be linked to a particular biology of breast carcinomas favoring progression and metastatic spread.

Les cancers du sein triple négatifs (CSTN) sont un sous-type hétérogène représentant 12% à 24% de tous les cancers du sein. Ils ont les plus mauvais pronostics et touchent souvent les femmes jeunes. Le traitement au stade localisé repose essentiellement sur la chimiothérapie, sans thérapie ciblée (hors cas de mutations germinales de BRCA). Quasiment toutes les patientes reçoivent la même chimiothérapie néoadjuvante (CNA) avec anthracyclines et taxanes, ce qui grève leur survie en l’absence de réponse complète pathologique (RCp). L’intensification thérapeutique est la tendance actuelle, notamment avec l’addition d’immunothérapie, afin d’améliorer taux de RCp et survie. Les signatures d’expression génique standard ne pas utilisables en pratique clinique pour prédire la chimiorésistance des CSTNs à la CNA, alors qu’elles permettraient de sélectionner les patientes pour une intensification thérapeutique. D’où l’intérêt d’explorer les transcrits issus des 90% du génome restant, constitué de régions non codantes et non référencées. Parmi eux, les ARNs longs non-codant (lnc) qui se caractérisent par une longueur d’au moins 200 nucléotides présentent l’intérêt pour certains d’entre eux d’avoir une expression tissu- voire cancer- spécifique. De plus, certains ARNs lnc ont un rôle démontré dans différents mécanismes de chimiorésistance. Les ARNs lnc ne sont cependant pas bien annotés dans le génome humain et de nouveaux transcrits non référencés, codant ou non, et de nouvelles isoformes de gènes connus, sont découverts quotidiennement.Mon 1er objectif de thèse était d’évaluer le transcriptome non référencé en tant que réservoir potentiel de biomarqueurs prédictifs de chimiorésistance des CSTN à la CNA. Nous avons analysé une cohorte de 78 CSTNs avant CNA, et comparé les cas chimiosensibles (chS) et chimiorésistants (chR) selon le score RCB (Residual Cancer Burden). Nous avons comparé l’analyse d’expression génique différentielle standard (DE-seq) sur les gènes annotés, et sur de nouveaux ARNs lnc détectés grâce à un profiler ARN, et une analyse non biaisée par les annotations génomiques, de fragments de transcrits différentiels. L’analyse non biaisée a permis d’obtenir la meilleure séparation des cas chS et chR dans notre cohorte exploratoire. Nous avons ensuite évalué cette approche sur une cohorte indépendante, et nous l’avons optimisée, en considérant les régions génomiques d’expression différentielle. Ceci nous a permis d’obtenir une signature reproductible entre les deux cohortes. Au total, ces résultats montrent le potentiel d’une approche non référencée pour générer une signature de chimiorésistance précoce des CSTN. Des validations supplémentaires sont nécessaires sur des cohortes plus larges.Mon 2nd objectif de thèse était d’évaluer les ARNs lnc en tant que potentiels acteurs/cibles thérapeutiques dans les CSTNs chR. Nous avons sélectionné les ARNs lnc surexprimés dans les CSTNs chR pre-CNA (versus les CSTNs chS pre-CNA) et dans les CSTNs chR post-CNA (versus les CSTNs chR pre-CNA). En intégrant leurs niveaux et spécificités d’expression, leurs localisations génomiques et les données préexistantes suggérant une fonction potentielle, nous avons retenu trois ARNs lnc (AL450326.1, LINC02609 et MIR503HG). Nous avons évalué l’impact de l’inactivation de leur expression sur la cytotoxicité du Docetaxel dans un modèle de lignée cellulaire de CSTN. L’inactivation de l’expression de chacun des trois ARNs lnc a induit une cytotoxicité accrue du Docetaxel. Une clonogénicité spontanée diminuée et une mort cellulaire accrue sous Docetaxel ont de plus été observées après la déplétion d’AL450326.1 et de LINC02609. Au total, nous avons identifié trois ARNs lnc jouant un rôle dans la chimiorésistance des CSTN. Des tests fonctionnels supplémentaires sont nécessaires pour en décrypter les mécanismes, avec pour objectif à long terme d’identifier de nouvelles approches thérapeutiques contrant la chimiorésistance des CSTNs à la CNA
Triple-negative breast cancers (TNBC) represent a heterogeneous subtype of breast cancers including 12% to 24% of all cases, having the poorest prognoses and often affecting young women. Treatment at localized stage is mainly based on chemotherapy, with no targeted therapy (except germline BRCA mutated patients). Nearly all patients receive the same Neo-Adjuvant Chemotherapy (NAC) with anthracyclines and taxanes, that badly impacts survival in the absence of pathological complete response (pCR). Therapeutic intensification, notably with addition of immunotherapy, is the current trend to increase pCR rate and improve survival. Standard gene expression signatures have failed to provide effective tools to predict TNBC chemoresistance, probably due to their incomplete nature, as they are mostly based on expression of protein coding genes and/or referenced transcripts and up to date there is no clinically useful transcriptomic signature predicting TNBC chemoresistance to NAC. Such a predictive signature would allow patient selection for therapeutic intensification. Therefore, it is important to explore the remaining 90% of the genome consisting of non-coding and non-referenced regions. One class of non-coding RNAs that is of great interest are long non-coding (lnc) RNAs, that are at least 200 nucleotides long, some of them being specifically expressed in cancer. Moreover, some lncRNAs have been shown to be implicated in different mechanisms of chemoresistance. LncRNAs are not fully well annotated in the human genome and new unreferenced transcripts, coding or not, and new isoforms of known genes are discovered daily.Therefore, the first goal of my PhD was to assess reference-free transcriptome as a potential reservoir of predictive biomarkers of TNBC chemoresistance. A cohort of 78 TNBCs before NAC was analyzed, comparing chemosensitive (chS) and chemoresistant (chR) cases based on international Residual Cancer Burden (RCB) score. A standard differential gene expression analysis (DE-seq) on annotated genes, and on new lncRNAs detected with a de novo RNA-profiler, and a reference-free analysis of differential fragments of transcripts without annotation bias were compared. Reference-free approach showed best separation of chS and chR patients in the training cohort. Further, based on comparison with an independent validation cohort, an optimized approach was proposed, where specific genomic regions with differential expression were selected. This technique gave a reproducible signature of chemoresistance between the two cohorts. In all, these results show the potential of a reference-free approach to generate a transcriptomic signature as predictive biomarker of early TNBC chemoresistance. Further investigation is needed to validate the signature using larger validation cohorts.The second objective of my PhD was to assess lncRNAs as potential actors/therapeutic targets in chR TNBCs. For that we selected lncRNAs upregulated in chR pre-NAC TNBCs (compared with chS pre-NAC TNBCs) and in chR post-NAC TNBCs (compared with chR pre-NAC TNBCs). Considering lncRNAs level and specificity of expression, genomic position, and pre-existing data of their potential function, three lncRNAs (AL450326.1, LINC02609 and MIR503HG) were retained for functional analysis. By knocking down levels of these lncRNAs in TNBC cell line model, an impact on Docetaxel cytotoxicity was assessed. All three lncRNAs knock downs showed an improved Docetaxel induced cytotoxicity. Knock down of AL450326.1 and LINC02609 resulted in a decreased spontaneous clonogenicity and increased Docetaxel induced cell death, giving a first indication of their mode of action. In all, we identified three lncRNAs playing a role in NAC chemoresistance. Further functional studies will allow to decipher the mechanisms by which the identified lncRNAs affect chemoresistance with the ultimate goal to identify new therapeutic approaches to circumvent NAC chemoresistance of TNBCs