Using controllable gene-editing systems in targeted cancer treatment: current uses and future possibilities

One of the most straightforward approaches to chemical induction involves placing the expression of Cas9 or sgRNA under the control of drug‐responsive promoters. The tetracycline (Tet)‐inducible system, comprising the reverse tetracycline‐controlled transactivator (rtTA) and the Tet‐responsive element (TRE), is perhaps the most widely used transcriptional platform for CRISPR regulation. In this system, the addition of doxycycline (Dox) induces conformational changes in rtTA that enable its binding to TRE, thereby activating transcription of the downstream Cas9 gene.22, 30 Dow and colleagues demonstrated that Dox‐inducible Cas9 expression in transgenic mice enabled efficient and reversible gene editing in vivo, providing a powerful platform for temporal genetic studies in cancer biology.22 Similarly, the cumate‐inducible system, based on the CymR repressor and the cumate operator sequence, has been adapted for conditional CRISPR/Cas9 expression in mammalian cells.31 These transcriptional induction systems offer the advantage of simplicity and compatibility with established genetic tools, but their temporal resolution is inherently limited by the kinetics of transcription, translation, and protein accumulation, typically requiring several hours to achieve maximal editing activity.

To achieve more rapid and precise control, several groups have developed post‐translational strategies in which Cas9 protein activity is directly modulated by small molecules.

Split‐Cas9 with chemically induced dimerization (CID): Zetsche et al.32 demonstrated that Cas9 can be split into two inactive fragments (N‐Cas9 and C‐Cas9) that are individually fused to rapamycin‐responsive dimerization domains (FKBP and FRB). In the absence of rapamycin, the two fragments remain separated and catalytically inactive; upon rapamycin administration, FKBP–FRB heterodimerization reconstitutes a functional Cas9 nuclease. This strategy has been further refined using alternative CID pairs such as the abscisic acid (ABA)‐responsive ABI/PYL1 system, providing orthogonal induction modalities.33

Intein‐mediated protein splicing: Intein‐based approaches involve the insertion of ligand‐dependent inteins into the Cas9 coding sequence, rendering the protein non‐functional until the appropriate small molecule triggers intein self‐splicing and reconstitution of full‐length Cas9. Davis et al.17 pioneered this concept by incorporating a 4‐hydroxytamoxifen (4‐HT)‐responsive intein into SpCas9, achieving dose‐dependent and temporally controlled editing with significantly reduced off‐target activity compared to constitutive Cas9. More recently, intein engineering and split‐intein strategies have been used to improve the inducibility, specificity, and portability of conditional Cas9 activity beyond the original 4‐HT design, providing alternative routes to tighten background activity and expand stimulus responsiveness.34

Destabilizing domain (DD)‐based regulation: The fusion of engineered destabilizing domains to Cas9 targets the protein for rapid proteasomal degradation in the absence of a stabilizing ligand. Upon addition of the cognate small molecule (e.g., Shield‐1 for the FKBP12‐derived DD, or trimethoprim (TMP) for the E. coli dihydrofolate reductase (ecDHFR)‐based DD), the DD undergoes conformational stabilization, preventing Cas9 degradation and enabling editing activity.19, 35 This strategy is particularly attractive for applications requiring reversible control, as removal of the stabilizing ligand leads to rapid Cas9 depletion.

Allosteric small‐molecule switches: Maji et al.36 developed an allosterically regulated Cas9 in which a small‐molecule‐dependent allosteric effector domain was engineered into the REC lobe of Cas9, such that ligand binding induces a conformational change that activates DNA cleavage. Similarly, anti‐CRISPR proteins (Acrs) have been adapted for chemically regulated inhibition: fusion of Acrs to degron domains enables ligand‐dependent degradation of the inhibitor, thereby activating CRISPR function.37, 38

Chemical control can also be exerted at the guide RNA level. Ligand‐responsive aptamers incorporated into the sgRNA scaffold can modulate its structure and function in response to small molecules such as theophylline or tetracycline.39, 40 In the “off” state, the aptamer sequesters critical sgRNA regions required for Cas9 binding or target recognition; ligand binding to the aptamer releases these structural constraints, restoring sgRNA functionality. Kundert et al.41 systematically engineered ligand‐responsive sgRNAs with tunable dynamic ranges, demonstrating their utility for dose‐dependent gene regulation.

Additionally, chemically modified sgRNAs with photolabile protecting groups (“caged” sgRNAs) or chemically cleavable modifications have been developed, although these approaches overlap with light‐inducible strategies and will be discussed in subsequent sections (Figure 2).42

Nihongaki et al.18 pioneered the photoactivatable CRISPR‐Cas9 (paCas9) system by fusing split Cas9 fragments to the blue light‐responsive positive Magnet (pMag) and negative Magnet (nMag) photodimers derived from Neurospora crassa Vivid protein. Blue light illumination (488 nm) induces pMag–nMag heterodimerization and Cas9 reconstitution, enabling light‐dependent genome editing in human cells. The same group subsequently developed an improved version, paCas9‐2, with enhanced efficiency and reduced background activity.43 An alternative approach utilized the CRY2–CIB1 blue light‐inducible dimerization pair from Arabidopsis thaliana for split‐Cas9 reconstitution.44

Beyond direct Cas9 activation, light‐inducible transcription systems have been used to control the expression of CRISPR components. The light‐activated (LA)‐CRISPR system employs a split Cre recombinase fused to blue light‐responsive dimerization domains; upon illumination, Cre‐mediated recombination removes a transcriptional stop cassette flanked by loxP sites, permitting Cas9 expression from an upstream constitutive promoter.44 Similarly, the EL222 blue light‐responsive transcription factor has been adapted to drive light‐dependent Cas9 expression.45

An elegant chemical biology approach involves the introduction of photolabile protecting groups (“cages”) on critical nucleotides within the sgRNA, thereby blocking Cas9 loading or target hybridization in the dark. Upon UV or near‐visible light irradiation, the photolabile groups are cleaved, restoring sgRNA activity. Jain et al.42 demonstrated that caging as few as three nucleotides in the seed region of the sgRNA was sufficient to abolish editing activity in the dark, with full restoration upon brief UV exposure. Similarly, Hemphill et al.46 developed photocaged anti‐CRISPR oligonucleotides that inhibit sgRNA function in the dark and release it upon light exposure.

Near‐infrared (NIR) light‐activated upconversion nanoparticle (UCNP) systems enable spatiotemporally precise control of CRISPR/Cas activity in deep tissues, overcoming limitations of conventional delivery. UCNPs convert tissue‐penetrating NIR light into localized UV/visible emissions to trigger photocleavable linkers or activate photosensitizers, facilitating endosomal escape and on‐demand release of CRISPR components. For instance, UCNP‐based nanomachines have been engineered for NIR‐controlled co‐delivery of CRISPR machinery and photosensitizers to enhance gene‐editing efficacy while enabling real‐time imaging guidance.47 Similarly, NIR‐upconversion theranostic platforms tailored for CRISPR‐Cas9/DNAzyme systems support precise tumour‐targeted gene regulation with minimized off‐target effects.48 Furthermore, UCNP‐integrated biosensors leverage CRISPR‐Cas12a for dual‐mode upconversion luminescence detection of nucleic acid biomarkers, enhancing diagnostic sensitivity in oncology applications.49

Despite these advantages, a representative problematic design is the use of UCNP‐assisted light‐inducible CRISPR activation for deep tumours. Although UCNPs can convert tissue‐penetrating NIR light into UV/visible emission and thereby activate photocaged or photosensitive CRISPR modules, as demonstrated in NIR‐responsive CRISPR/Cas9 nanocarrier systems, this design introduces several translational liabilities, including nanoparticle retention, uncertain long‐term clearance, local heating or phototoxicity, heterogeneous intratumoral distribution, and difficulty in standardizing the delivered optical dose.50 Thus, UCNP‐assisted systems are conceptually powerful for spatial control but remain less mature than chemically inducible or transient non‐viral delivery platforms for clinical oncology.

Tissue‐ and tumour‐specific promoters represent a straightforward transcriptional approach to restrict CRISPR activity to cancer cells. Commonly used cancer‐specific promoters include the human telomerase reverse transcriptase (hTERT) promoter, the survivin promoter, the alpha‐fetoprotein (AFP) promoter for hepatocellular carcinoma, and the prostate‐specific antigen (PSA) promoter for prostate cancer.27 Liang et al.51 demonstrated that driving Cas9 expression from the hTERT promoter restricted gene editing to telomerase‐positive cancer cells while sparing normal cells. However, the relatively weak transcriptional activity of many tumour‐specific promoters, compared with constitutive promoters such as CMV or EF1α, remains a significant challenge that may necessitate the use of transcriptional amplification strategies.27

Hypoxia, a hallmark of solid tumours, contributes to therapy resistance, metastasis, and immune evasion.52, 53 Hypoxia‐responsive CRISPR/Cas systems exploit hypoxia‐inducible factors (HIFs), particularly HIF‐1α—a master transcriptional regulator of cellular adaptation to low oxygen—to restrict genome‐editing activity or cargo release to hypoxic tumour regions. Current strategies include HIF‐responsive regulatory circuits, hypoxia‐activated nanoplatforms, tumour‐homing carriers delivering CRISPR/Cas9 against hypoxia‐associated targets, and oxygen‐generating or redox‐responsive systems combined with CRISPR‐mediated pathway modulation.51, 53, 54 These approaches may modulate hypoxia‐adaptive metabolic programs and immunosuppressive signalling, including the PD‐1/PD‐L1 axis, in specific experimental settings.51, 53 Overall, hypoxia‐responsive platforms convert a key tumour microenvironmental feature into a therapeutic trigger, potentially improving tumour selectivity while supporting combination with chemo‐, immuno‐, and phototherapeutic strategies.52, 54 Future work should focus on improving induction efficiency, delivery kinetics, and robustness across heterogeneous hypoxic niches.

Differential microRNA (miRNA) expression profiles between cancer cells and normal cells can be exploited to construct miRNA‐responsive CRISPR circuits. Hirosawa et al. engineered miRNA‐responsive Cas9 mRNAs containing miRNA target sites in the 5′‐ or 3′‐untranslated regions (UTRs)55; in cells expressing the cognate miRNA, Cas9 translation is repressed, while in cells lacking the miRNA (e.g., cancer cells with downregulated tumour‐suppressive miRNAs), Cas9 is expressed. Conversely, Acr can be placed under miRNA‐responsive control, where cancer‐specific miRNAs degrade the anti‐CRISPR mRNA and de‐repress CRISPR activity. Hoffmann et al.56 demonstrated a cell‐type‐specific CRISPR system controlled by endogenous miR‐21, which is overexpressed in many cancers, achieving selective gene editing in miR‐21‐high tumour cells.

The TME is characterized by elevated levels of reactive oxygen species (ROS), glutathione (GSH), and acidic extracellular pH (6.5–6.8) compared to normal tissues.57 These biochemical differences have been exploited for stimulus‐responsive delivery of CRISPR components. Liu et al. developed GSH‐responsive nanoparticles that release Cas9/sgRNA ribonucleoproteins (RNPs) selectively in the reducing intracellular environment of tumour cells.58 Similarly, pH‐sensitive lipid nanoparticles (LNPs) that undergo structural disassembly in the acidic TME or endosomal compartments have been engineered for tumour‐selective CRISPR delivery.59

Dynamic metabolic reprogramming represents another clinically relevant feature of the TME that may be exploited by inducible CRISPR/Cas systems. In bladder cancer, altered lipid metabolism, including fatty acid transport, fatty acid metabolism‐related molecular subtypes, and TERC‐driven fatty acid remodelling, has been associated with tumour progression, prognosis, immune microenvironment features, and cisplatin resistance.60, 61, 62 Because lipid metabolic pathways are also essential for normal tissue homeostasis, constitutive CRISPR‐mediated disruption of these axes may increase toxicity in non‐malignant cells. TME‐responsive or physically triggered inducible CRISPR platforms may therefore provide a safer strategy by restricting metabolic gene editing to specific tumour regions or therapeutic windows.

Ultrasound‐responsive systems offer a non‐invasive, spatiotemporally precise platform for inducible CRISPR/Cas applications. These systems leverage ultrasound's deep tissue penetration and stimuli‐responsive nanocarriers to enable on‐demand release of genetic payloads or activation of gene circuits.63, 64 Recent advances in ultrasound‐triggered transgene circuits provide a foundation for spatiotemporally regulated CRISPR/Cas activation.65 Integration with TME cues or sonodynamic therapy may further enhance specificity and therapeutic synergy.66 However, clinical translation requires refinement in safety, scalability, stimulus calibration, and delivery efficiency.67

Thermal control systems enable spatiotemporally precise activation of CRISPR/Cas tools in oncology through non‐invasive external stimuli. Focused ultrasound (FUS) or NIR light induces localized mild hyperthermia (≈42°C), triggering Cas9‐mediated editing of apoptosis‐resistance genes (e.g., HSP70, BAG3) within tumours while sparing healthy tissue.68 This strategy disrupts tumour cell survival mechanisms and remodels the immunosuppressive microenvironment to enhance adoptive T‐cell infiltration and efficacy. Magnetothermal approaches similarly activate CRISPR‐Cas9 for targeted HSP70 knockout, achieving synergistic pro‐apoptotic effects.69 Engineered thermal gene switches leveraging heat‐shock promoters (e.g., HSP70) and optimized FUS platforms provide reversible, tunable control with real‐time temperature monitoring.70

Magnetic field‐responsive CRISPR/Cas systems may support spatiotemporally controlled genome editing in oncology by leveraging external magnetic stimuli to activate editing machinery preferentially within tumour regions. Magnetothermal approaches utilize alternating magnetic fields to induce mild hyperthermia (∼42°C), triggering dual CRISPR‐Cas9 editing to disrupt apoptosis‐resistance genes (e.g., HSP70) while synergistically sensitizing tumour cells to thermal apoptosis.69 Magnetomechanical strategies employ magnetic nanostructures to activate mechanosensitive channels (e.g., Piezo1), initiating Ca²⁺‐dependent expression of Cas9 for remote, non‐invasive gene regulation.71 Additionally, magnetic nanoparticle‐assisted platforms (e.g., MAGE, magnetoplexes) enhance delivery efficiency, biocompatibility, and on‐target accuracy of CRISPR components to tumour sites.72, 73 These systems may reduce off‐target exposure and collateral damage to healthy tissues, which are important concerns for translating CRISPR‐based therapies to clinical oncology (Figure 3).74

Traditional genetically engineered mouse models (GEMMs) based on Cre‐loxP technology require extensive breeding and are limited to pre‐engineered alleles. The advent of constitutive Cas9‐expressing mice (Rosa26‐Cas9) enabled somatic genome editing through viral delivery of sgRNAs, accelerating cancer model generation.23 However, inducible Cas9 mice offer further refinements by allowing precise control over the timing and duration of editing. Dow et al. generated Dox‐inducible Cas9 mice (Rosa26‐LSL‐rtTA; Col1a1‐TRE‐Cas9) that enable temporally controlled gene editing in any tissue accessible to Dox.22 Using this platform, the authors demonstrated sequential inactivation of tumour suppressors in the intestinal epithelium, faithfully modelling the adenoma‐carcinoma sequence of colorectal cancer.

Platt et al. and Sánchez‐Rivera et al. utilized Cre‐dependent Cas9 mice (Rosa26‐LSL‐Cas9) combined with tissue‐specific Cre drivers (e.g., Krt14‐Cre for skin, Pdx1‐Cre for pancreas) to model tumour initiation in defined compartments.23, 82 More recently, inducible Cas9 systems have been combined with multiplexed sgRNA libraries delivered via adeno‐associated virus (AAV) to generate autochthonous models harbouring combinatorial mutations in putative cancer genes, enabling in vivo genetic interaction mapping.83, 84

Three‐dimensional organoid cultures derived from patient tumours or normal tissues have emerged as powerful platforms for cancer modelling and drug testing.85 Inducible CRISPR systems have been applied to organoids to model stepwise tumorigenesis ex vivo. Drost et al.86 sequentially introduced oncogenic mutations (APC, TP53, KRAS, SMAD4) into normal human intestinal organoids using CRISPR/Cas9, recapitulating the multi‐hit model of colorectal cancer. The use of inducible Cas9 in organoid systems allows researchers to control the timing and order of mutation acquisition, facilitating studies of clonal evolution, epistatic interactions, and the minimal set of driver mutations required for malignant transformation.87

Inducible CRISPR‐based lineage tracing systems, such as the CARLIN (CRISPR array repair lineage tracing) mouse, employ Dox‐inducible Cas9 to generate cumulative, heritable barcode mutations in a synthetic target array, enabling retrospective reconstruction of clonal hierarchies and cell fate decisions during tumour evolution.88 These systems provide unprecedented insights into cancer stem cell dynamics, clonal competition, and the impact of therapeutic interventions on tumour architecture.

To improve translational relevance, inducible CRISPR/Cas systems should also be validated in patient‐derived xenografts (PDX), orthotopic tumour models, and humanized immune mouse models. PDX models better preserve patient‐specific tumour heterogeneity and can be used to evaluate tumour penetration, delivery efficiency, and editing responses in clinically relevant tumour architectures. Humanized immune mouse models are particularly important for testing inducible CRISPR strategies designed to enhance immunotherapy, such as checkpoint editing or CAR‐T sensitization. However, establishing stable inducible tumour models remains technically challenging because lentiviral packaging, clonal selection, basal leakiness, and inducer responsiveness may vary across cell lines and in vivo passages.

Inducible CRISPR/Cas systems may enable spatiotemporally controlled oncogene disruption to enhance precision in cancer therapy while potentially reducing off‐target exposure in healthy tissues. For instance, an enzyme‐inducible CRISPR (eiCRISPR) platform achieves cell‐selective genome editing via disease‐associated proteases, restricting Cas9 activity to malignant cells.89 Similarly, FUS‐inducible CRISPR systems, delivered via AAVs, disrupt telomeres to sensitize solid tumours to CAR‐T therapy.90 Targeting amplified oncogenes (e.g., MYCN in neuroblastoma), CRISPR‐Cas9 nickase exploits replication‐dependent DNA breaks for selective tumour cell killing.91

Inducible CRISPR/Cas systems offer precise spatiotemporal control for tumour suppressor gene (TSG) restoration in precision oncology. Catalytically dead Cas9 (dCas9)‐based activators, such as the dCas9‐NVPR complex, effectively restore expression of silenced TSGs (e.g., PER2 and ZNF382), significantly inhibiting malignant phenotypes with high genomic specificity.92 Dual CRISPR interference/activation (CRISPRi/a) strategies enable targeted reactivation of non‐mutated X‐linked TSG alleles silenced by X‐chromosome inactivation, presenting a tailored approach for breast cancer therapy.93 Furthermore, enhancer‐switching via CRISPR/Cas9‐mediated editing reactivates endogenous TSGs, providing an alternative epigenetic restoration strategy.94 Integration of inducible modules—such as enzyme‐responsive (eiCRISPR),89 far‐red light‐inducible Cas12a (FICA),95 or Tet‐regulated systems75—may improve the safety profile by restricting activation to tumour microenvironments(TMEs), potentially reducing off‐target exposure while enabling dynamic control over TSG re‐expression for personalized therapeutic intervention.

Acquired drug resistance is a major cause of treatment failure in oncology. Inducible CRISPR systems can be deployed to disrupt resistance‐conferring genes in a temporally coordinated manner with drug administration. For example, CRISPR‐dCas9‐based platforms enable targeted transcriptional regulation of resistance‐associated genes (e.g., enhancing platinum sensitivity via CT45 modulation) without genomic cleavage, improving drug susceptibility in lung cancer models.96 Genome‐wide CRISPR knockout and activation screens systematically identify resistance drivers (e.g., BCL2, BRIP1) and synthetic lethal targets across chemotherapeutics, guiding rational combination strategies.97, 98, 99 Innovations like SIBR‐Cas provide tight, inducible control over editing activity to reduce prolonged nuclease exposure and potentially limit off‐target effects during resistance gene targeting.100 Integrating these controllable systems with TME‐responsive delivery holds significant promise for overcoming therapy‐induced resistance while preserving normal tissue function in precision oncology.

Conditional CRISPR/Cas systems may enable more spatiotemporally controlled immune checkpoint disruption to enhance tumour‐specific immunotherapy while potentially reducing systemic toxicity. For example, an HSP70 promoter‐driven CRISPR/Cas9 nanosystem (F‐PC/pHCP) achieves tumour‐selective PD‐L1 genomic knockout, triggering potent T‐cell activation, inhibiting primary and distant tumours, and establishing long‐term immune memory.101 Similarly, FUS‐inducible CRISPR platforms allow non‐invasive, localized editing to disrupt checkpoints and prime solid tumours for CAR‐T cell therapy.90 Enzyme‐inducible CRISPR (eiCRISPR) further refines specificity by activating editing only in disease‐associated microenvironments via self‐blocked sgRNA unmasking.89 These inducible strategies show how controlled immune checkpoint modulation can overcome limitations of conventional blockade therapies.

Inducible CRISPR has been applied to engineer “armored” CAR‐T cells with enhanced functionality. Examples include: (i) inducible knockout of endogenous TCR and HLA genes to generate universal, allogeneic CAR‐T cells with reduced graft‐versus‐host disease (GvHD) risk102; (ii) conditional disruption of T‐cell exhaustion‐associated genes (e.g., TOX, NR4A family) to sustain effector function in the TME10; and (iii) inducible activation of synthetic cytokine circuits that augment CAR‐T cell persistence and tumour infiltration.103 The temporal control afforded by inducible systems is particularly important for CAR‐T cell engineering, where the timing of gene edits relative to T cell activation and expansion critically influences manufacturing outcomes.

In situ tumour vaccination (ISV) leverages endogenous tumour‐associated antigens (TAAs) at the primary tumour site to stimulate systemic, tumour‐specific adaptive immunity while minimizing systemic toxicity.104, 105 However, its efficacy is frequently constrained by the immunosuppressive TME, insufficient immunogenic cell death (ICD), and inadequate antigen cross‐presentation.106, 107 Integrating inducible CRISPR/Cas systems offers a precision strategy to overcome these barriers: TME‐responsive nanoplatforms enable programmable release of CRISPR/Cas9 to disrupt immune checkpoints (e.g., PD‐L1, PTPN2) or silence immunosuppressive genes, thereby reprogramming the TME and amplifying adaptive immune responses.108 CRISPR activation (CRISPRa) screens further identify targets to enhance antigen presentation and T‐cell infiltration,106, 109 while manganese‐based or hydrogel‐based delivery systems facilitate localized, stimulus‐triggered genome editing for robust in situ vaccination. These advances position inducible CRISPR/Cas tools may become useful components of next‐generation personalized cancer immunotherapy.

Light‐activated CRISPR systems can be co‐delivered with photosensitizers for photodynamic therapy (PDT) or photothermal agents for photothermal therapy (PTT), enabling simultaneous gene editing and physical tumour destruction. Li et al.110 developed a multifunctional nanoplatform combining gold nanorods (for PTT), UCNPs (for NIR‐to‐UV light conversion), and photocleavable CRISPR/Cas9 components, achieving synergistic anti‐tumour effects through concurrent gene disruption and hyperthermia. The shared light trigger ensures spatial and temporal coordination of the two therapeutic modalities.

CRISPR/Cas systems synergize with chemotherapy to overcome resistance and enhance tumour‐specific cytotoxicity. Nanoparticle‐mediated co‐delivery of CRISPR components and chemotherapeutics (e.g., cisplatin or doxorubicin) reconfigures cellular drug responses, sensitizes resistant cells via targeted gene knockout (e.g., PRKDC, FBXO44), and targets cancer stem cell‐associated programs while potentially reducing off‐target exposure.111, 112, 113, 114 Engineered inducible platforms, such as i‐CRISPR incorporating DNA repair inhibitors, enable spatiotemporally controlled DNA damage to amplify chemotherapy efficacy in a tumour‐selective manner.115 These combinatorial strategies reflect a broader move in precision oncology toward individualized and mechanism‐guided treatment combinations.

X‐ray‐guided CRISPR/Cas9 (X‐CC9) enables tumour‐localized activation under irradiation, selectively reprogramming tumour‐associated macrophages (M2‐to‐M1) while sparing healthy tissues.116 CRISPR‐based functional screens identify synthetic lethal targets linked to radiation resistance in glioblastoma and head and neck cancers, revealing novel radiosensitization pathways.117, 118 Knockout of atypical PARP genes (e.g., PARP9/12/13/14) via CRISPR/Cas9 significantly enhances radiosensitivity in colorectal cancer models.119 Integrating inducible editing with radiotherapy supports spatiotemporally controlled interventions, advancing personalized combinatorial regimens to overcome therapeutic resistance.120 Among current applications, ex vivo immune‐cell engineering and locoregional treatment of accessible solid tumours appear to be more realistic near‐term translational scenarios, whereas systemic in vivo genome editing remains constrained by delivery heterogeneity, tumour penetration, immune responses, and difficulties in monitoring editing outcomes across lesions (Table 2).

The field currently lacks standardized metrics to compare inducible systems across platforms. Consensus is needed on: ‐ Minimum acceptable dynamic range (fold‐change between ON/OFF states) in solid tumours, not just cell culture ‐ Maximum tolerable leakiness (uninduced activity as % of maximal activity) for different therapeutic contexts ‐ Editing efficiency thresholds required for phenotypic impact in heterogeneous tumour populations ‐ Standardized tumour models for head‐to‐head comparisons (e.g., orthotopic vs. subcutaneous, immunocompetent vs. immunodeficient)

Without these benchmarks, we cannot objectively assess whether a “tighter” system with 10% editing efficiency is superior to a “leakier” system with 60% efficiency. The community should establish a consortium to define these parameters through systematic comparative studies, similar to what the ENCODE project achieved for functional genomics.

Current inducible systems face a paradox: the more components required for tight control (split proteins, regulatory RNAs, stabilizing ligands), the harder they become to deliver efficiently. We must prioritize: ‐ All‐in‐one compact architectures: Engineering miniature Cas variants (CasΦ, Cas12f) with built‐in inducibility through rational protein design or directed evolution, eliminating the need for separate regulatory modules. ‐ Stimulus‐responsive delivery vehicles that ARE the inducible switch: Rather than delivering pre‐assembled inducible systems, develop nanocarriers where the act of cargo release itself constitutes the induction event (e.g., TME‐triggered disassembly releases active Cas9 RNPs). ‐ Tissue‐specific AAV capsids with endogenous promoters: Combine viral tropism engineering with tumour‐specific transcriptional control to achieve dual‐layer selectivity without payload expansion.

The goal is not merely incremental improvement but a fundamental redesign: inducible systems must become simpler to deliver, not more complex.

Before any inducible CRISPR cancer therapy enters human trials, we need long‐term (≥6 month) safety data in immunocompetent large animals (pigs, non‐human primates) addressing: ‐ Chronic immunogenicity: Do repeated inductions trigger adaptive immune responses that eliminate edited cells or cause systemic inflammation? ‐ Off‐target accumulation: Does transient Cas9 activity truly reduce off‐target editing, or do repeated induction cycles lead to cumulative mutagenic burden? ‐ Resistance evolution: Can tumours evolve to evade inducible systems (e.g., by silencing tumour‐specific promoters or upregulating anti‐CRISPR‐like inhibitors)? ‐ Biodistribution of inducers: For systemically administered chemical or physical triggers, what are the long‐term consequences of repeated exposure?

These studies are expensive and time‐consuming, but they are non‐negotiable. The field cannot afford a clinical failure due to unforeseen toxicity that could have been detected in appropriate preclinical models.