consist of Y Gilad et al. commune biol. https://doi.org/10.1038/s42003-021-01929-1 (2021).
Synthetic lethal screens support the discovery of novel anticancer drug targets1. In a recent issue of communication biologyGilad et al.2 performed a synthetic lethal CRISPR-Cas9 dropout screen on the human estrogen receptor-positive breast cancer (BC) cell line MCF-7. They aimed to identify targets that increase the sensitivity of MCF-7 cells to the small molecule inhibitor SI-12, which targets SRC-3, an essential transcription cofactor of the estrogen receptor. A key result of their screen showed that targeting certain olfactory receptors (ORs) in BC could confer anti-tumor effects. However, these ORs and a number of other hits are not expressed in MCF-7 cells, which challenges the establishment of the screen and justifies the inclusion of transcriptome data in the analysis pipeline of genetic screens.
The identification of novel drug targets is an important pillar to expand treatment options for cancer patients. Through their screen, Gilad et al. Identified olfactory receptors OR4D6 and OR5I1alongside a number of other genes, as potential targets in breast cancer patients2. Interestingly, other studies have linked olfactory receptors (ORs) to cancer biology in general3and also specifically in BC4.5. The studies by Weber et al.4 and Masjedi et al.5 performed transcriptome analyzes of both BC cell lines and BC tumor tissues to assess OR expression and many ORs were discovered. Surprisingly, however, this was not found in either study OR4D6 or OR5I1, used in the Gilad et al. identified ORs, in each of the BC-related biological samples, that contained MCF-7. We further investigated the expression of these two ORs in cancer cell lines using online expression databases (The Human Protein Atlas (https://www.proteinatlas.org/), the Broad Institute Cancer Dependency Map (https://depmap.org/ portal/) (version 21Q1)) which confirmed the lack of expression of these ORs in MCF-7 cells. Further investigation showed that a number of other targets included in Gilad et al.’s CRISPR-Cas9 screen. were also not expressed in this cell line (NDNF, S1PR1) according to the online databases.
To investigate whether the synthetic CRISPR-Cas9 lethal results extended beyond the MCF-7 cell line, Gilad et al. exposed additional BC cell lines (T-47D, BT-474, ZR-75-1 and MDA-MB-231) to SI-12 after temporarily silencing a number of the identified targets, including OR4D6, using siRNAs. Although less potent than in MCF-7 cells, it is silenced OR4D6 nevertheless sensitized three of these cell lines to treatment with SI-12, leading the authors to conclude that this gene could be an interesting therapeutic target in BC in general. The data from Weber et al. and Masjedi et al. showed no expression of this OR in these additional BC cell lines reported by Gilad et al. were used, which was confirmed by the data in the online expression repositories.
A possible explanation for the therapeutic effects of silencing the identified targets, which are apparently not expressed in the examined cancer cell lines, could be that their expression is induced by exposure to SI-12. Gilad et al. exposed cells to SI-12 after CRISPR Cas9 or siRNA-mediated gene interference. Therefore, the CRISPR-Cas9 screen appears to have selected a number of targets with no biological role in MCF7 cells. This indicates that Gilad et al. For the setup and execution of their screen, the internal experimental controls may not have been sufficient to accurately identify valid targets. Against this background, it is surprising that the siRNA validation experiments by Gilad et al. performed also showed that turning off the “unexpressed” targets conferred antitumor effects. Both CRISPR-Cas9 and RNAi technologies have been described as suffering from off-target effects, albeit to varying degrees6.7.
As high-throughput functional genetics screens found their way into biomedical research, much effort was put into optimizing the technical aspects and data analysis. However, a prerequisite for assessing the biological role of a gene in a given context is expression at physiologically relevant levels, an aspect that is equally important when performing functional screens to assign a value. Therefore, it is strongly recommended to determine the full-length transcriptome of the biological test sample under the experimental conditions of a genetic screen. This should be integrated into the data analysis pipeline to exclude transcripts with irrelevant expression levels. Also, excluding unexpressed targets avoids wasting time validating such “hits”. In this regard, publicly available transcriptome data from sources such as The Human Protein Atlas and the DepMap consortium provide a good starting point, although they relate to untreated samples.
The study by Gilad et al. illustrates, in our opinion, the need to integrate target gene expression data into the analysis pipeline of a functional genetic screen. Furthermore, it demonstrates the susceptibility of siRNA-based target validation to off-target effects, as silencing the unexpressed targets resulted in a phenotypic effect. An important recommendation for siRNA-based validation is to carry out rescue experiments6, which are not published by Gilad et al. In addition, methods were developed to distinguish between on- and off-target effects in siRNA- and CRISPR-Cas9-based screens8th. Rigorous follow-up of all recommendations made over the years to increase the validity of functional genetic screen results should lead to a more robust identification of promising drug targets. Although the results of other studies have confirmed an interesting role for ectopically expressed olfactory receptors in breast cancer, in our opinion those of Gilad et al. so it needs to be improved.
The data supporting the results of this study are available at https://www.proteinatlas.org/ and https://depmap.org/portal/.
Huang, A., Garraway, LA, Ashworth, A. & Weber, B. Synthetic lethality as a driver for cancer drug discovery. nat. Rev. Drug Discov. 1923-38 (2020).
Gilad, Y. et al. A genome-scale CRISPR-Cas9 dropout screen identifies synthetically lethal targets in SRC-3-inhibited cancer cells. commune biol. 4399 (2021).
Kalra, S. et al. Analysis of single-cell transcriptomes links odorant receptor enrichment to cancer cell differentiation status and prognosis. commune biol. 11506 (2020).
Weber, L. et al. Olfactory receptors as biomarkers in human breast cancer tissues. Front. unc. 8th33 (2018).
Masjedi, S., Onion, LJ & Giorgio, TD Olfactory receptor gene abundance in invasive breast carcinoma. Science. representative 913736 (2019).
Kaelin, WG Common pitfalls in preclinical validation of cancer targets. nat. Rev Cancer 17425-440 (2017).
Schuster, A. et al. RNAi/CRISPR screens: from a pool to a valid hit. Trends Biotechn. 3738-55 (2019).
Riba, A. et al. Explicit modeling of siRNA-dependent on- and off-target repression improves the interpretation of screening results. cell syst. 22182-193 (2017).
The authors declare no competing interests.
Peer Review Information
communication biology thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editors: Eve Rogers.
publisher’s Note Springer Nature remains neutral on published map jurisdiction claims and institutional affiliations.
About this article
Cite this article
Schultz, IJ, Coelingh Bennink, HJT Target expression is a relevant factor in synthetic lethal screens.
Communal Biol 5, 835 (2022). https://doi.org/10.1038/s42003-022-03746-6
By submitting a comment, you agree to our Terms of Service and Community Guidelines. If you find something abusive or not consistent with our terms or policies, please flag it as inappropriate.