NEW YORK (GenomeWeb) — Through a combination of next-generation sequencing and CRISPR-Cas9-mediated genome editing, two sets of researchers homed in on the mechanisms of action for a handful of drugs and uncovered mutations that can lead to resistance to those drugs, as they reported in separate studies in Nature Chemical Biology yesterday.
The exact targets of a number of drugs aren't known, making it difficult for researchers to improve the efficacy of those drugs or to limit their toxicity. The 'gold standard' to prove the direct target of a drug typically involves showing that a mutation in a particular protein leads resistance to the drug.
But, as the two teams of researchers noted, being able to conduct such screens in mammalian or human cells is difficult because of a lack of efficient molecular biology tools.
Both teams turned to CRISPR-Cas9 genome editing to inactivate certain genes in mammalian cells and determine whether those mutations led to resistance to the drugs under study.
"[T]esting whether any single mutation can confer drug resistance in human cells typically involves transgene overexpression and may fail for several reasons, such as toxicity," Rockefeller University's Tarun Kapoor and his colleagues said in their paper. "We reasoned that direct genome editing would circumvent this major obstacle and developed an integrated approach for identification of drug targets."
Genome editing using CRISPR-Cas9 relies on a guide RNA that targets the nuclease to the corresponding DNA segment where it introduces double-stranded breaks into the DNA, disrupting that gene.
Kapoor and his colleagues developed a method they dubbed DrugTargetSeqR that combines sequencing, computational mutation discovery, and CRISPR-Cas9-based genome editing.
They developed this tool by analyzing ispinesib, an inhibitor of kinesin-5. They isolated a dozen drug-resistant clones — clones that were between 70- and 300-fold less sensitive to ispinesib than their parents. While four clones harbored known multi-drug resistance substrates, eight did not, and those eight seemed to resist the drug through a direct, rather than indirect, mechanism.
By sequencing the transcriptomes of the eight resistant clones and their parental clones, the researchers narrowed in on reads with mutations that were absent in the parental clones. The kinesin-5-encoding gene was mutated in each of the eight drug-resistant clones. The researchers uncovered three different mutations in that gene in those clones.
To determine whether any of those three kinesin-5 mutations is enough to confer ispinesib resistance, the researchers turned to the CRISPR-Cas9 system. They transfected HeLa cells with the Cas9 nuclease and synthetic guide RNAs and homologous template DNA with and without a substitution in the kinesin-5 gene.
Cells transfected with the mutant kinesin-5 — which was confirmed through the Surveyor mutation detection assays and Sanger sequencing of the locus — led to a more than 150-fold increase in resistance to ispinesib.
The substitutions mapped to the protein's drug-binding pocket, confirming a previous finding.
Kapoor and his colleagues also applied their approach to induce point substitutions that lead to resistance to the proteasome inhibitor bortezomib in two cancer cell lines.
"Together, these data indicate that this genome-editing protocol can overcome a major bottleneck in establishing the 'genetic' proof of a drug's physiological target," they said.
Further, by sequencing the ispinesib-resistant clones, the researchers found that the mutant kinesin-5 allele represents between 85 percent and all of the reads in each clone. However, the allele encoding kinesin-5 allele is heterozygous, leading the researchers to speculate that the wild-type kinesin-5 is silenced through an epigenetic mechanism.
Similarly, Yan Feng and a team at the Novartis Institutes for Biomedical Research found a CRISPR-Cas9 genome editing approach to be "highly efficient" for searching for drug-resistance alleles or rescuing drug sensitivity.
Through a forward genetic screen in the KBM7 cell line, Feng and colleagues searched for mutations linked to resistance to 6-thioguanine, an anti-cancer drug that is transformed into 6-thioguanine-monophosphate by hypoxanthine phosphoriosyltransferase (HPRT1). All the resistant clones, the researchers found, had nonsense mutations in one HPRT1 allele.
To validate this, they transfected HCT116 cells with CRISPR-Cas9 constructs targeting HPRT1, which, they reported, led to a marked increase in 6-TG resistance and confirmed the role of HPRT1 in 6-TG-induced cell death.
Feng and colleagues used the same approach to examine the antiprofilerative agent triotolide, which has been show to have an effect in mouse models of pancreatic cancer, though through an unknown mechanism.
Previous work, the researchers noted, implicated several possible targets for the drug, including ERCC3, dCTP pyrophosphatase, polycystin 2, and TAB1.
In the 26 triotolide-resistant clones the researchers studied, 15 had ERCC3 mutations and six had wild-type ERCC3 but had mutations in GTF2H4, which binds and actives the ATPase activity of ERCC3.
Using two Cas9 guide RNAs, the researchers targeted two different ERCC3 mutant alleles in a cell line. After transfection, the researchers noted, cells can attempt to repair the cuts through either nonhomolo¬gous end-joining or homologous recombination, leading to either knock-in or knock-out products. To enrich for knock-in products, the researchers exposed cells to a dose of triptolide to kill any sensitive cells.
Most of the resistant clones, they noted, had both a knock-in allele generated by homologous recombination and a knockout allele with a frameshift mutation generated by nonhomolo¬gous end-joining. They also found no wild-type alleles among the genotyped clones, leading them to hypothesize that the presence of even one allele is sufficient for triplolide sensitivity.
Introducing wild-type ERCC3 cDNA, the researchers reported, could reverse the trait.
"Although recessive drug-resistant alleles can be easily identi¬fied and verified in a lower-organism model system such as yeast, our approach, which combines haploid screening, next-generation exome sequencing, and CRISPR knockout and knock-in technologies, makes it practical in mammalian cells," Feng and colleagues said.
They said that such an approach might also be able to be applied to the study of nontoxic compounds by tweaking how phenotypic selection is conducted.
"To analyze mechanisms of action of non-cytotoxic agents, drug-resistant clones could be selected using reporter gene expression or a phenotype that can be readily measured for a few cells or for single cells, for example, by high-throughput microscopy," added the Rockefeller team in their paper.