What Is Crispr-cas9? Its Mechanism, Technique and Full Name
In this article
CRISPR-Cas9 is a tool derived from a bacterial defense system against viruses that cuts DNA at a desired location. Contrary to a common claim, it was not discovered in 2013 —that was the year the system was first made to work in human cells. Today it is no longer merely a laboratory technique: an approved medicine has existed since 16 November 2023.
This article does three things. First, it corrects seven errors circulating in Turkish-language sources, with references—from the discovery date and which enzyme does what to the translation of the abbreviation. Second, it explains how the system works, including frequently omitted components: guide RNA, the PAM sequence and repair after cutting —where the actual “editing” occurs. Third, it asks the central oncology question: Where does CRISPR stand in cancer treatment?
Let us answer this last question first, because it is the most commonly misunderstood point: As of today, no CRISPR treatment is approved for cancer. The only approved CRISPR medicine, Casgevy, is for sickle cell anemia and beta thalassemia—neither is cancer. All oncology CRISPR programs remain at phase 1 or phase 1-2. This does not mean the field has failed; patient expectations must align with today’s reality .
First, seven corrections
| Circulating statement | The facts |
|---|---|
| "Discovered in 2013 CRISPR-Cas9" | Incorrect. The repeat sequences were observed by chance in 1987 by Yoshizumi Ishino and colleagues in E. coli. The abbreviation was coined in 2001–2002 . Its immune function was proposed in 2005 and experimentally demonstrated in March 2007 . 2013 was the year the system was first used for genome editing in human cells —not its discovery year. |
| "The Cas9 enzyme scans and cuts viral DNA fragments. The bacterium inserts these fragments into its own genome and creates memory." | The mechanism is confused. Memory formation is performed not by Cas9, but by Cas1 and Cas2 (with Csn2 also involved in type II systems). Cas9’s role is only in the interference stage. The sequence is also reversed: memory is written first (on the first encounter), cutting happens later (on reinfection). |
| "Using CRISPR, HIV DNA was removed from human cells successfully" | Too strong a claim. Integrated HIV DNA can be cut in the laboratory. However, in the EBT-101 phase 1/2 study presented in May 2024, among five participants, of those who stopped treatment, the virus returned in all three and antiretroviral treatment resumed. The treatment was found safe but cured no one. |
| "In mouse experiments on muscular dystrophy, the mutation was corrected and healthy offspring were born" | This statement could not be verified for this article and is misleading because it suggests inherited (germline) editing. It has been removed from the text. |
| "CRISPR = Regularly divided clusters of palindromic repeats" | The translation contains two errors. “Interspaced” means divided is not the meaning; rather, regularly separated by intervening sequences is intended; also, “Short” was omitted. A more faithful rendering is: "Clustered Regularly Interspaced Short Palindromic Repeats." |
| "Cas = CRISPR-associated system" | For the enzyme, the correct expansion is "CRISPR-associated protein"—Cas9 = CRISPR-associated protein 9. “CRISPR-associated system” describes the gene/locus family. |
| The article never mentions guide RNA, the PAM sequence or repair pathways at all | These constitute the system’s operating principle itself. Guide RNA determines the target, PAM permits cutting, and the actual “editing” occurs during repair after cutting . CRISPR cannot be explained without these three; they are covered separately below. |
The correct timeline
Mechanism: what actually happens in bacteria?
CRISPR is an immune system and, like every immune system, works in three stages : remembering, preparing and attacking. Turkish sources often compress these three stages into one sentence and assign the wrong task to the wrong enzyme.
| Stage | What happens? | Which proteins? |
|---|---|---|
| 1. Acquisition (adaptation) Writing memory |
During the first infection, a short fragment of the invader’s DNA is taken and inserted into the CRISPR array between two repeats as a new spacer . | Cas1 and Cas2 (with Csn2 also involved in type II). Cas1 is an integrase. Cas9’s only role here is to assist in selecting PAM-adjacent fragments. |
| 2. Expression (crRNA production) Preparing |
The CRISPR array is transcribed as a long pre-crRNA and processed into short molecules, each carrying one spacer, called crRNA molecules. | In type II, a separate small RNA, tracrRNA , pairs with the repeat region; RNase III cuts and releases mature crRNA. |
| 3. Interference Attacking |
On reinfection, the crRNA-loaded complex moves around the cell, searches for PAM, opens the DNA and pairs the spacer with the corresponding protospacer; if they match, it cuts both strands. | Cas9 —and only at this stage. |
Figure 1 · The three stages of CRISPR immunity
The bacterium first writes memory, then produces the guide and finally cuts. Cas9, which performs cutting, has no role in writing memory.
The difference between spacer and protospacer: Spacer is the foreign-derived sequence stored between repeats in the bacterium’s own CRISPR array—its memory record. Protospacer is the original matching sequence in the phage or plasmid genome—the target. Both describe the same sequence, but one is in memory and the other in the enemy.
The tool’s three components: guide RNA, PAM and cutting
Guide RNA
| crRNA | A mature guide derived from a spacer, approximately 20 nucleotides long . It base-pairs with the target protospacer. |
|---|---|
| tracrRNA | A separate small RNA complementary to the repeat region. In type II systems it is required both for crRNA maturation and loading onto Cas9. Charpentier’s group, March 2011. |
| sgRNA (single guide RNA) |
An engineered version of crRNA and tracrRNA fused into a single molecule . This is the step that turns a bacterial immune system into a programmable tool —Jinek, Chylinski, Charpentier and Doudna, 2012. |
PAM: the system’s safety lock
PAM (protospacer adjacent motif) is a short DNA motif that must lie immediately beside the target sequence. Streptococcus pyogenes Cas9 has the standard PAM 5'-NGG-3' (N = any base). In the human genome, NGG occurs roughly every 8–12 bases, allowing most genes to be targeted.
Without PAM, Cas9 does not cut. PAM recognition is the first step permitting DNA opening and the start of base pairing.
PAM’s most elegant function is preventing the bacterium from cutting itself. Beside spacers in the bacterium’s CRISPR array, there is no PAM; beside protospacers in the phage genome, there is. Thus, the same crRNA that destroys a phage cannot target the bacterium’s own memory record. A two- or three-base motif solves the immune system’s problem of distinguishing “self” from “foreign.” Nature has prevented autoimmunity with a punctuation mark.
Where and how does cutting occur?
Cas9 creates the double-strand break 3 base pairs upstream of PAM . Two separate nuclease domains operate: HNH cuts the strand paired with crRNA, RuvC cuts the opposite strand. Approximately %85 of breaks are blunt-ended (blunt), while the rest are slightly staggered.
The actual “editing” occurs after cutting
A frequent conceptual error is this: CRISPR does not “cut and paste.” CRISPR only cuts. What determines the outcome—gene disruption or correction—is the cell’s own repair machinery. There are two pathways, and which one operates depends largely on the cell’s state, beyond your control.
| NHEJ (nonhomologous end joining) | HDR (homology-directed repair) | |
|---|---|---|
| Is a template required? | No | Yes —a sister chromatid or externally supplied donor DNA |
| Accuracy | Error-prone; often leaves small insertions/deletions (indels) behind | Because it uses a template, error-free |
| Outcome | An indel not divisible by three → frameshift → premature stop codon → gene disruption (knockout) | Correction of a point mutation or insertion of a new sequence |
| When does it work? | At every stage of the cell cycle, including in nondividing cells | Mainly in S and G2 phases —in other words, only in actively dividing cells |
| In nondividing cells (neurons, muscle) | Works | Not usable in practice |
Figure 2 · Two repair pathways after cutting
CRISPR only cuts. The repair pathway used by the cell determines whether the outcome is gene disruption or correction.
This table explains the field’s greatest technical limitation. A gene is easy to disrupt —NHEJ works everywhere. A gene is much harder to correct because HDR operates only in dividing cells and within a particular cell-cycle window, while competing with NHEJ for the same break. Therefore, most CRISPR treatments reaching the clinic to date are based on “switching something off,” rather than “correcting something” —including Casgevy. Why a mutation’s location within a gene matters is discussed in our article, but the logic works in reverse here: correcting a mutation is much harder than recognizing it.
Cas9 is not the only option
| Enzyme | Target | PAM | Cut | Feature |
|---|---|---|---|---|
| Cas9 (S. pyogenes) | DNA | 5'-NGG-3', on the 3' side | Blunt-ended double-strand break | Standard; requires crRNA + tracrRNA (or sgRNA) |
| Cas12a (Cpf1) | DNA | 5'-TTTV-3', on the 5' side | Staggered, sticky ends | Opens up AT-rich targets; does not require tracrRNA, processes its own pre-crRNA |
| Cas13 | RNA | No DNA PAM | RNA cleavage | RNA silencing, RNA editing and diagnostic tests |
| SaCas9 (S. aureus) | DNA | — | Double-strand break | Small enough to fit in a viral vector; used where SpCas9 does not fit |
How is it delivered into cells? Ex vivo and in vivo
The original article’s ex vivo / in vivo distinction is correct but incomplete: what really matters is which delivery vehicle is used.
| Vehicle | Destination | Use | Note |
|---|---|---|---|
| RNP electroporation (preassembled Cas9 protein + guide RNA) |
Into cell culture | Ex vivo gold standard —blood stem cells and T cells | Fastest action, shortest lifetime, lowest off-target risk |
| Lipid nanoparticle (LNP) | When given intravenously, mainly to the liver | The main in vivo nonviral route | Carries mRNA or RNP; transient expression |
| AAV | Tissue-directed (eye, nervous system, muscle) | In vivo | Packaging limit ~4,7 kb —requiring smaller Cas9 variants or dual-vector strategies |
| Lentivirus | Integrates into the host genome | Predominantly ex vivo —stem cell and CAR-T production | Provides permanent expression |
Casgevy: the world’s first approved CRISPR medicine
This section addresses the field’s most important development, absent entirely from the original article. Casgevy (active ingredient exagamglogene autotemcel, abbreviated exa-cel) was developed by Vertex Pharmaceuticals and CRISPR Therapeutics.
| Date | Authority | Decision |
|---|---|---|
| 16 November 2023 | United Kingdom MHRA | The world’s first licensed CRISPR/Cas9 treatment. Sickle cell anemia and transfusion-dependent beta thalassemia, ages 12 and older |
| 8 December 2023 | United States FDA | With recurrent vaso-occlusive crises: sickle cell anemia, ages 12 and older |
| 16 January 2024 | United States FDA | Transfusion-dependent beta thalassemia, ages 12 and older |
| 13 February 2024 | European Commission | Conditional authorization, both indications, ages 12 and older |
| 1 July 2026 | United States FDA | Age threshold lowered to 2 years —for both indications |
MHRA and EU approvals are limited to patients for whom stem cell transplantation is appropriate and no HLA-matched related donor is available .
How does it work? The patient’s own CD34+ blood stem cells are collected and, in the laboratory, CRISPR/Cas9 edits the erythroid-specific enhancer region of the BCL11A gene . BCL11A normally suppresses fetal hemoglobin (HbF) ; removing this suppression allows cells to produce HbF again. Fetal hemoglobin does not sickle and replaces defective adult beta-globin, which is why the same edit works in both sickle cell anemia and beta thalassemia.
Note: the faulty gene is not corrected. Instead, a gene switched off after birth— another gene—is switched on again. This is a direct consequence of the NHEJ–HDR limitation discussed above: switching off is easier than correcting.
Figure 3 · Casgevy does not repair the faulty gene—it switches another gene back on
The beta-globin gene defective in sickle cell anemia is not touched at all. Instead, fetal hemoglobin, switched off after birth, is reactivated.
| Study | Patients | Result |
|---|---|---|
| CLIMB-121 severe sickle cell anemia |
35 patients treated, 17 evaluable | 16/17 (%94,1) free of vaso-occlusive crises for at least 12 consecutive months; 17/17 (%100) at least 12 months without crisis-related hospitalization. Average crisis-free duration 18,7 months (longest 36,5). At month 3, HbF exceeded %30 of the total, subsequently ~%40. |
| CLIMB-111 transfusion-dependent beta thalassemia |
48 patients treated, 27 evaluable | 24/27 (%88,9) transfusion-independent for at least 12 consecutive months, weighted average Hb ≥9 g/dL. Average duration 20,5 months (longest 40,7). Total Hb at month 3 was ≥11, and from month 6 was ≥12 g/dL. |
Follow-up in the two studies extended to 43,7 months . Cells engrafted in all patients, and no malignancies were reported. The safety profile was described as “consistent with busulfan myeloablative conditioning and autologous stem cell transplantation.”
But these figures are misleading without explaining what patients undergo. Casgevy is not a pill or an infusion; it is a stem cell transplantation process:
(1) Mobilization medication and stem cell collection by apheresis—up to one week, repeated if necessary. (2) Off-site cell manufacturing and release testing— up to six months. (3) Several days in hospital for busulfan conditioning chemotherapy. (4) A one-time intravenous infusion. (5) After infusion, 4–6 weeks in hospital. Early risks are bleeding and infection due to low platelets and white blood cells. And the manufacturer’s own patient information states: "You may not be able to become pregnant or father a child after receiving the conditioning medicine."
The United States list price is 2,2 million dollars. For the broader discussion of access and pricing, see our articles on oncology drug prices and the map of access to new medicines . Casgevy’s licensing and reimbursement status in Türkiye could not be verified for this article; no claim is made on this subject.
The next generation: base editing and prime editing
Cas9’s fundamental problem is that it creates a double-strand break—and, as the safety section below shows, most risk arises from that break. David Liu’s laboratory developed two methods to overcome this problem.
| CRISPR-Cas9 | Base editing | Prime editing | |
|---|---|---|---|
| Break type | Double-strand break | Single-strand nick | No double-strand break |
| What it can do | Cuts and leaves repair to the cell | C→T and A→G transitions (cannot perform transversions) | All 12 base substitutions, plus small insertions and deletions |
| Is template DNA needed | Yes for HDR | No | No |
| Publication | 2012–2013 | Liu laboratory (exact years could not be verified) | Anzalone et al., Nature, 2019 |
And the field’s most striking case: KJ Muldoon. An infant with severe carbamoyl phosphate synthetase 1 (CPS1) deficiency —a urea cycle disorder causing toxic ammonia accumulation in the blood. At Children’s Hospital of Philadelphia and Penn Medicine, a treatment designed specifically for his variant using base editing was delivered to the liver through lipid nanoparticles.
The first dose was given in February 2025, at approximately 6–7 months of age; two more doses followed in March and April 2025. The time from variant identification to manufacturing was less than six months. As of April 2025, no serious side effects had occurred, the baby tolerated more protein, required less medication and recovered from a viral infection without elevated ammonia. New England Journal of Medicine, 15 May 2025.
Its significance goes far beyond one patient’s improvement: this case demonstrated that a treatment can be designed for a single person and manufactured within months. Today’s drug-development model is economically impossible for rare mutations; this model could change that. However, this is an experimental intervention, not an approved product.
CRISPR and cancer: where do we actually stand?
The original article mentions cancer as one of the fields where “CRISPR is considered a revolutionary approach.” That is true but incomplete. The concrete situation is:
The first United States CRISPR cancer study (Pennsylvania, 2020)
| Number of patients | 3 (advanced cancer) |
|---|---|
| What was done | In T cells, three genes were switched off —TRAC and TRBC (the cells’ own T-cell receptor chains) and PDCD1 (PD-1); then a lentiviral vector added a synthetic TCR specific to NY-ESO-1 . |
| Safety | No serious treatment-related adverse events |
| Persistence | Edited cells persisted for months —longer than previously documented; in one patient, at month 4, ~%20 of circulating T cells still carried multiple edits |
| Efficacy | No objective response. Two patients had stable disease, one a mixed response (some lesions shrank ~%50 for ~4 months, while others progressed). All had progressed by study end; one patient died. |
The value of this study lies in its safety message, not its treatment effect. It showed that T cells with three genes edited simultaneously could persist in humans for months without serious toxicity. But they did not shrink the tumor. It is important to maintain this distinction: “CRISPR was tested in cancer and found safe” is correct; “CRISPR treated cancer” is incorrect.
Allogeneic CAR-T: CRISPR’s most realistic oncology application
CAR-T therapy in its current form is autologous: the patient’s own T cells are collected, engineered and returned. The process takes weeks and is expensive. Allogeneic CAR-T can instead be produced from a healthy donor and used off the shelf —but donor T cells attack the patient’s tissues. CRISPR is being used to solve this problem: T-cell receptor and tissue-compatibility genes are switched off.
| Number of patients | 12, at four dose levels (30, 100, 300, 600 ×10⁶ CAR+ T cells) |
|---|---|
| Diseases | Relapsed/refractory B-cell malignancies: follicular lymphoma, marginal zone lymphoma, mantle cell lymphoma, large B-cell lymphoma |
| Overall response rate | 8/12 (%67) |
| Complete response | 6/12 (%50) |
| Response duration | One complete response exceeded 12 months; five responses lasted longer than 6 months |
| Toxicity | No dose-limiting toxicity; no grade 3 or higher infections. Cytokine release syndrome in 7/12, only grade 1–2; ICANS in 4/12, only grade 1–2; no GvHD observed |
| Regulatory | The FDA granted on 9 December 2024 for follicular lymphoma and marginal zone lymphoma RMAT status |
These data are promising, but involve 12 patients in phase 1. Responses across dose levels and absence of GvHD are meaningful, but there are no data on durability, long-term safety or comparison with autologous CAR-T. We discussed efforts to extend CAR-T to solid tumors in our articles on B7-H3 CAR-T in glioblastoma and B7-H3-targeted treatments .
In the laboratory: CRISPR screens and new targets
CRISPR’s greatest contribution to oncology so far may actually be finding targets , rather than treatment. Genome-wide CRISPR knockout screens switch off thousands of genes individually to determine which cancer cells cannot survive without which genes.
The canonical example is WRN helicase. Nature published a study in 2019 showing that cancers with microsatellite instability (MSI) depend on WRN helicase, whereas cells without MSI do not. Thus, WRN is a selectively lethal target in a molecularly defined patient group—MSI colorectal, endometrial and gastric cancers. This target emerged from a screening display, not from patients. We discussed AI-assisted drug discovery in our Virtual Biotech article; CRISPR screens are the experimental counterpart of that pipeline.
Safety: the honest section
Popular accounts usually compress CRISPR risks into “off-target cutting.” Yet some documented problems occur directly on target —cutting the correct site itself creates risk.
| Problem | Source | What was found? |
|---|---|---|
| The p53 problem | Haapaniemi et al., Nature Medicine, 11 June 2018 | Cas9 cutting triggers a p53-mediated DNA damage response and cell-cycle arrest. The implication is disturbing: the easiest cells to edit are those with a disrupted p53 pathway. Because p53 is a fundamental tumor suppressor, the authors explicitly recommended monitoring p53 function when developing CRISPR-based cell therapies. |
| Large deletions and complex rearrangements | Kosicki, Tomberg, Bradley, Nature Biotechnology 2018;36:765–771 | Cas9 double-strand breaks can create deletions spanning kilobases and complex rearrangements—which standard short-range PCR genotyping misses. |
| Chromothripsis | Nature Genetics 2021 | Chromosome fragmentation and incorrect reassembly, a phenomenon associated with cancer genomes, was reported as a consequence of Cas9 cutting an on-target, rather than off-target in the latter category. |
These findings do not mean CRISPR should not be used—but they mean we must understand three things.
First: Much of the risk arises from the double-strand break . Avoiding such breaks gives base editing and prime editing a theoretical safety advantage—and that is why the field is moving in that direction.
Second: Clinical data are reassuring for now. In Casgevy studies, no malignancy was reported during follow-up of up to 43,7 months; no serious treatment-related adverse events occurred in the Penn cancer study. But this follow-up is short for answering the cancer-risk question —and busulfan in the conditioning regimen itself carries a known secondary malignancy risk.
Third: Documented clinical consequences of off-target cutting and regulators’ official positions on long-term cancer risk from edited cell therapies could not be verified for this article. This does not mean “there is no risk,” but “this question has not yet been answered” .
He Jiankui and the prohibition on inherited editing
| 26 November 2018 | He Jiankui announces the birth of twin girls whose genomes were edited with CRISPR. The target gene is CCR5; the aim is HIV resistance. |
|---|---|
| Reaction | 122 Chinese scientists issue a statement calling the work “unethical” and “crazy.” Nobel laureate David Baltimore , chairing the Hong Kong summit, expresses “horror and sadness”; on 28 November the organizing committee calls the procedure “irresponsible” and a violation of international norms. Criticism focuses on inadequate informed consent, undisclosed risks and lack of peer oversight. |
| 30 December 2019 | Shenzhen Nanshan People’s Court convicts him of illegal medical practice: three years in prison and a fine of 3 million yuan. Two colleagues receive two years and 18 months. |
| April 2022 | He is released after completing his sentence. |
| Today | The World Health Organization, through an expert committee established in 2018, has concluded that inherited human genome editing remains too risky and should not be carried out . China amended its Civil Code and Criminal Law in 2020 to prohibit human gene editing for reproductive purposes, raising penalties to up to seven years. |
This episode also holds a lesson for oncology. What He Jiankui did was not technically difficult—CRISPR’s low cost and simplicity were its most celebrated features. The barrier lay in institutions, not technology; and one person circumvented those institutions. All edited cell therapies offered to patients with cancer today are somatic—they affect only that person, not future generations. This distinction must be clear when speaking to patients: Receiving CAR-T or a similar treatment does not change your children’s genomes.
Nobel and patents: whose invention?
The 2020 Nobel Prize in Chemistry was awarded equally to Emmanuelle Charpentier and Jennifer A. Doudna “for the development of a method for genome editing” . Because the prize is limited to three people, many contributors were excluded, generating debate.
| Who | Contribution |
|---|---|
| Feng Zhang | Made CRISPR-Cas9 work in human cells (3 January 2013). Against the United States–EU patent dispute, the Academy’s preference for basic over applied science was interpreted as a “major message.” |
| Virginijus Šikšnys | Reached the same biochemical conclusion; his paper was submitted before Jinek and colleagues’ but published later because it was rejected without peer review. He shared the 2018 Kavli Prize in Nanoscience with Doudna and Charpentier. |
| Francisco Mojica | Named CRISPR and proposed its immune function (2005) |
| Rodolphe Barrangou, Philippe Horvath | Experimentally demonstrated the function (2007) |
Sourced critical accounts of Mojica, Barrangou and Horvath’s exclusion could not be verified for this article; the dates of their contributions were verified. For more about the Nobel process, see our article on the 2020 Nobel Prize in Chemistry .
The patent battle is still not over. The Berkeley side (CVC) filed on 25 May 2012, and the Broad Institute on 12 December 2012 ; Broad obtained the first patent through accelerated examination in April 2014 . The United States Patent Trial and Appeal Board ruled on 15 February 2017 and on 28 February 2022 in Broad’s favor. However, on 12 May 2025, the Federal Court of Appeals reviewed the Board’s assessment of “conception” and vacated it and remanded the case: the wrong legal standard had been applied. The correct standard is whether the invention was “sufficiently complete for a person skilled in the art to put it into practice without further invention.” The priority question remains unresolved today.
DROZDOGAN Academy Commentary
The most striking aspect of updating this article was that the information needing correction was outdated and overly optimistic, rather than wrong . A CRISPR text written in 2018–2020 cannot include the field’s most important development—an approved medicine. But it also fails to show its most important limitation: there is still no approved CRISPR treatment for cancer.
⚡ Why it matters
Patients with cancer and their relatives hear about CRISPR and understandably ask, “Is this an option for me?” Today’s honest answer is no—but not yet is the answer. The most realistic oncology application is CRISPR editing not the tumor but the immune cell : allogeneic CAR-T. In CTX112’s 12-patient phase 1 study, a %67 response rate and absence of GvHD are meaningful signals—but 12 patients are 12 patients.
✅ The strength
Casgevy demonstrated that gene editing is no longer a promise but a licensed product —and the figures are good: in sickle cell anemia, 16/17 patients were crisis-free for more than a year, and in thalassemia, 24/27 patients were transfusion-independent. No malignancies were reported during follow-up of up to 43,7 months. More striking still is the KJ Muldoon case: a treatment written for one baby in less than six months. This is evidence that could reshape the economics of drug development for rare diseases.
⚠️ The weakness
Three areas. Biology: CRISPR made cutting easier, not correction—HDR works only in dividing cells, so most treatments reaching the clinic “switch off” rather than “correct”; even Casgevy does not repair the faulty gene, but switches off another gene. Safety: p53 selection, kilobase deletions and chromothripsis are on-target problems, and the long-term cancer-risk question remains unanswered. Access: A list price of 2,2 million dollars, manufacturing lasting up to six months, busulfan conditioning, 4–6 weeks in hospital and risk of permanent infertility—this is not a pill but a transplantation process.
🩺 What to tell the patient
(1) There is currently no approved CRISPR treatment for cancer; all oncology programs remain investigational. (2) Participating in a clinical trial is not the same as receiving an approved treatment; the primary goal of phase 1 studies is safety, not tumor shrinkage. (3) Somatic editing is not passed to future generations—CAR-T-like treatment does not alter your children’s genomes. (4) Be extremely cautious about unlicensed, fee-charging centers promising “cancer treatment with CRISPR”; no such approved service exists today. (5) The field’s real progress is in hematology and rare diseases; this may signal future progress in cancer, but does not guarantee it.
❓ Open questions
Do edited cell therapies carry a secondary malignancy risk ten years later ? Our longest follow-up for this question is 43,7 months, and the conditioning regimen’s own risk confounds the issue. Second: do base and prime editing truly reduce this risk by eliminating double-strand breaks, or merely create a different risk profile? Third, and most important for oncology: can allogeneic CAR-T be equally effective as autologous CAR-T? The cost and access advantages of an off-the-shelf product might be worthwhile even with a small loss of efficacy—but that trade-off has not yet been measured.
Conclusion: CRISPR’s story began in 1987 with a few repeat sequences nobody understood and reached a licensed medicine in 2023—thirty-six years. That interval shows how slowly “revolution” works. Today we have a genuinely effective treatment for two inherited blood diseases, an experimental intervention written for a baby within months, and encouraging but very early phase 1 oncology data. The correct statement is not “CRISPR will defeat cancer,” but “CRISPR may work by editing the cell that attacks the cancer cell rather than the cancer cell itself,” and the evidence so far involves only 12 patients. The memory bacteria wrote against viruses produced a medicine; it may also produce one for cancer, but has not yet done so.
Frequently asked questions
- When was CRISPR discovered?
- Repeat sequences were observed in 1987 by chance; the abbreviation was coined in 2001–2002 , immune function proposed in 2005 and demonstrated in 2007 , programmable scissors shown in 2012 , and the system made to work in human cells on 3 January 2013 . The common claim “discovered in 2013” is incorrect.
- What does CRISPR stand for?
- Clustered Regularly Interspaced Short Palindromic Repeats —“Clustered Regularly Interspaced Short Palindromic Repeats.” The frequent Turkish translation “divided” is incorrect: “interspaced” means the repeats are separated by intervening sequences . Cas , when referring to the enzyme, means “CRISPR-associated protein.”
- What is PAM, and why does it matter?
- A short DNA motif that must be immediately beside the target sequence; for S. pyogenes Cas9, it is 5'-NGG-3'. Without PAM, Cas9 does not cut. Its most elegant function is that absence of PAM beside records in the bacterium’s own CRISPR array prevents it from cutting itself.
- Is CRISPR really “cut and paste”?
- No— CRISPR only cuts. The cell’s own repair machinery determines the outcome. NHEJ is template-free and error-prone, usually silencing the gene; HDR uses a template and can make corrections, but works only in dividing cells within a particular cell-cycle window. This makes “disrupting a gene” easy and “correcting a gene” difficult.
- Is there an approved CRISPR medicine?
- Yes, one: Casgevy (exagamglogene autotemcel). United Kingdom 16 November 2023, United States 8 December 2023 (sickle cell anemia) and 16 January 2024 (beta thalassemia), Europe 13 February 2024. In the United States, on 1 July 2026 the age threshold was lowered to 2.
- Is there an approved CRISPR treatment for cancer?
- No. All oncology CRISPR programs are at phase 1 or phase 1-2. The most promising area is CRISPR-produced allogeneic CAR-T cells; CTX112’s 12-patient phase 1 study reported %67 overall response and %50 complete response. The FDA on 9 December 2024 granted RMAT status. This is a development-acceleration designation, not approval.
- Has CRISPR treated HIV?
- No. Integrated HIV DNA can be cut in the laboratory. However, in EBT-101 phase 1/2 research, among three participants who stopped antiretroviral therapy, the virus returned in all three and treatment resumed—one maintained suppression for 16 weeks. The treatment was found safe but cured no one, because not all cells harboring latent HIV could be reached.
- Will CRISPR treatment be passed to my children?
- No. All gene-editing treatments currently administered to patients are somatic —they affect only the treated person’s body cells and are not passed to future generations. Editing reproductive cells (inherited editing) is considered too risky by the World Health Organization and is prohibited in many countries, including China.
- Does CRISPR cause cancer?
- This remains an open question, and the honest answer is “not yet known.” There are three documented concerns: Cas9 cutting selects cells with a disrupted p53 pathway (2018), can create kilobase deletions (2018), and chromothripsis was reported as an on-target consequence (2021). Conversely, clinical data are reassuring for now: no malignancy was reported in Casgevy studies with follow-up up to 43,7 months. However, this is short for answering the cancer-risk question.
- Is base editing different from CRISPR?
- It uses the same targeting system but does not create a double-strand break: it nicks one strand and converts C→T or A→G. Prime editing can make all 12 base substitutions and small insertions/deletions without creating any break. These methods are theoretically safer because most risk arises from double-strand breaks.
Glossary
- CRISPR
- Short palindromic repeat sequences in bacterial and archaeal genomes, separated by foreign DNA fragments. The memory record of an adaptive immune system.
- Cas9
- CRISPR-associated protein 9. An RNA-guided DNA endonuclease acting only at the interference stage.
- Cas1 / Cas2
- Proteins responsible for adding new spacers to the CRISPR array—writing memory. Cas1 is an integrase.
- Spacer
- A foreign-derived sequence stored between repeats in the bacterium’s CRISPR array. A memory record.
- Protospacer
- The original sequence in the phage or plasmid genome matching the spacer. The target.
- PAM
- Protospacer adjacent motif. A short motif required immediately beside the target; for S. pyogenes Cas9, 5'-NGG-3'. A prerequisite for cutting.
- crRNA
- A guide RNA approximately 20 nucleotides long, derived from one spacer.
- tracrRNA
- A separate small RNA required for crRNA maturation and loading onto Cas9 in type II systems.
- sgRNA
- crRNA and tracrRNA fused into a single molecule. The engineering step that makes CRISPR a programmable tool.
- NHEJ
- Nonhomologous end joining. Template-free, error-prone repair leaving small indels and usually silencing the gene.
- HDR
- Homology-directed repair. Uses a template, is error-free, but works mainly in dividing cells.
- Indel
- A small insertion or deletion. Causes a frameshift if not divisible by three.
- Ex vivo
- Removing cells from the body, editing them and returning them.
- In vivo
- Delivering editing components directly into the patient’s body.
- Base editing
- A method converting a single base without creating a double-strand break (C→T, A→G).
- Prime editing
- A “search and replace” system that writes new genetic information directly without breaks or donor DNA.
- Allogeneic CAR-T
- CAR-T produced from a healthy donor rather than the patient’s own cells. CRISPR is used to switch off genes that prevent donor cells from attacking the patient.
- RMAT
- Regenerative Medicine Advanced Therapy. An FDA designation to accelerate development of regenerative medicine products; not approval.
- Chromothripsis
- A chromosome shattering and reassembling in the wrong order. A phenomenon seen in cancer genomes.
Transparency note.
Verified: Ishino’s 1987 observation and organism; Mojica’s 1993–2000 studies; the abbreviation being coined in 2001–2002 and identification of cas1–cas4; three groups demonstrating spacer–phage matching in 2005 and Mojica’s immune hypothesis; Barrangou–Horvath’s March 2007 experimental proof; Marraffini–Sontheimer 2008; Moineau’s December 2010 finding on cutting position; the Charpentier group’s March 2011 tracrRNA discovery; Jinek et al. Science 337:816–821 citation and sgRNA fusion; Gasiunas–Šikšnys PNAS 4 September 2012 and the 18 April 2012 rejection story; Zhang and Church’s simultaneous publications on 3 January 2013 and the %2–38 rate in Church’s paper; Science’s selection on 18 December 2015 and its twice being runner-up; the three-stage mechanism, division of roles between Cas1/Cas2/Csn2 and Cas9, spacer–protospacer definitions; crRNA/tracrRNA/sgRNA definitions; the PAM sequence, genomic frequency and self-cleavage prevention; cleavage 3 bases upstream of PAM, HNH/RuvC division of labor and %85 blunt-ended breaks; NHEJ and HDR characteristics and HDR’s S/G2 restriction; Cas12a, Cas13 and SaCas9 features; delivery routes and AAV’s ~4,7 kb limit; Casgevy’s five regulatory decisions and dates, the BCL11A erythroid enhancer mechanism, CLIMB-121 and CLIMB-111 results, 43,7-month follow-up without reported malignancy, patient-process steps, infertility warning and 2,2 million dollar United States list price; the Anzalone 2019 citation for prime editing and base editing’s origin in Liu’s laboratory; the KJ Muldoon case’s diagnosis, institutions, dose dates, manufacturing in less than six months and the 15 May 2025 NEJM publication; the Penn phase 1 study design, three genes, three patients and lack of objective responses; CTX112 phase 1 data and the 9 December 2024 RMAT date; Haapaniemi 2018 p53, Kosicki 2018 large-deletion and 2021 chromothripsis findings; EBT-101 phase 1/2 results; the WRN/MSI Nature 2019 finding; dates of the He Jiankui events, sentence and release information, the statement by 122 scientists and Baltimore’s response; the Nobel 2020 citation and official rationale; Šikšnys’s 2018 Kavli Prize; patent case dates and the 12 May 2025 Federal Appeals decision.
Not verified: Online publication and submission dates of Jinek et al.’s paper; Science’s two years naming CRISPR runner-up; the full citation for the Barrangou–Horvath 2007 paper; whether the abbreviation was agreed in 2001 or 2002 (both are reported); sourced criticism of Mojica, Barrangou and Horvath’s exclusion from the Nobel; NHEJ–HDR efficiency ratios; the first publication years for base editing; the citation for Ihry et al.’s 2018 p53 paper; documented clinical consequences of off-target cleavage; regulators’ official views on long-term cancer risk in edited cell therapies; CTX110 and CTX130 data; Caribou, Intellia and Beam oncology programs (including BEAM-201); details of DepMap; Casgevy’s price outside the United States and the number treated to date; NTLA-2001’s target gene; exact citations for WHO’s 2021 governance documents; Casgevy’s licensing and reimbursement status in Türkiye and whether CRISPR-based clinical trials exist in Türkiye —no claims have been made on these subjects. The original statement “muscular dystrophy was corrected in mouse experiments and healthy offspring were born” was not included because it could not be verified and suggests inherited editing.
Source of the criticisms: The assessments that “2013 was the application year, not the discovery year,” “CRISPR cuts rather than cutting and pasting,” “most treatments reaching the clinic switch things off because that is easier than correction,” “Casgevy is a transplantation process, not a pill,” and “the correct cancer statement is editing the cell attacking the tumor rather than the tumor” belong to this article. In contrast, the p53 selection warning, EMA’s and FDA’s own wording, the Penn study’s lack of response and the chromothripsis finding are statements from the sources themselves.
Conflict of interest declaration: We have no connection with any medicine, product or company named in this article.
References
- Ishino Y, Shinagawa H, Makino K, Amemura M, Nakata A. Journal of Bacteriology 1987;169(12):5429–5433.
- Broad Institute. "CRISPR Timeline." broadinstitute.org
- Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. "A Programmable Dual-RNA-Guided DNA Endonuclease in Adaptive Bacterial Immunity." Science 2012;337(6096):816–821.
- Gasiunas G, Barrangou R, Horvath P, Šikšnys V. PNAS 2012;109:E2579–E2586.
- Cong L, et al. Science 2013;339(6121):819–823; Mali P, et al. Science 2013;339(6121):823–826. Both on 3 January 2013.
- AAAS. "Science selects CRISPR genome-editing tool as 2015 Breakthrough of the Year." 18 December 2015. aaas.org
- The Nobel Prize in Chemistry 2020. nobelprize.org
- Vertex Pharmaceuticals / CRISPR Therapeutics. Announcements from MHRA (16 November 2023), FDA (8 December 2023 and 16 January 2024), European Commission (13 February 2024) and FDA age expansion (1 July 2026).
- Vertex Pharmaceuticals. CLIMB-121 and CLIMB-111 pivotal study results, EHA 2023 data cutoff.
- European Medicines Agency. Casgevy EPAR. ema.europa.eu
- Stadtmauer EA, June CH, et al. "CRISPR-engineered T cells in patients with refractory cancer." Science 2020. doi:10.1126/science.aba7365
- CRISPR Therapeutics. CTX112 phase 1 data, ASH 2024; RMAT announcement 9 December 2024.
- Anzalone AV, et al. Prime editing. Nature 2019.
- Penn Medicine and Children’s Hospital of Philadelphia. KJ Muldoon case announcements; New England Journal of Medicine, 15 May 2025.
- Haapaniemi E, Botla S, Persson J, Schmierer B, Taipale J. "CRISPR–Cas9 genome editing induces a p53-mediated DNA damage response." Nature Medicine 2018;24(7).
- Kosicki M, Tomberg K, Bradley A. "Repair of double-strand breaks induced by CRISPR–Cas9 leads to large deletions and complex rearrangements." Nature Biotechnology 2018;36:765–771.
- "Chromothripsis as an on-target consequence of CRISPR–Cas9 genome editing." Nature Genetics 2021.
- Excision BioTherapeutics. EBT-101 phase 1/2 interim results, ASGCT, May 2024.
- "WRN helicase is a synthetic lethal target in microsatellite unstable cancers." Nature 2019.
- United States Court of Appeals for the Federal Circuit, decision 22-1653, 12 May 2025.
Sources
CRISPR-Cas systems for editing, regulating and targeting genomes.
Nature Biotechnology 32, 347–355 (2014)
2. Chen Liang, Mark A. Wainberg, Atze T. Das and Ben Berkhout.
CRISPR/Cas9: a double-edged sword when used to combat HIV infection.
Retrovirology 2016.
3. Rafal Kaminski, Yilan Chen, Tracy Fischer, Ellen Tedaldi, Alessandro Napoli, Yonggang Zhang, Jonathan Karn, Wenhui Hu & Kamel Khalili.
Elimination of HIV-1 Genomes from Human T-lymphoid Cells by CRISPR/Cas9 Gene Editing.
Scientific Reports 6, Article number: 22555 (2016)
Was this article helpful?
Your feedback helps us improve our content.
Reader comments
Share your comment
Your name and comment are published only after administrator approval. Your email address is never shown to visitors.
Be the first to comment.