Gene Therapy
CRISPR Therapy
Cas9 + guide RNA — programmable genome editing that reached clinical cure in 2023
CRISPR-Cas9 cuts DNA at a sequence specified by a 20-nucleotide guide RNA. The cell repairs the break — usually with errors (NHEJ) — to knock the gene out. Casgevy became the first approved CRISPR drug in December 2023.
- Guide length20 nt spacer + tracrRNA scaffold
- PAM (SpCas9)5'-NGG-3' on non-target strand
- On-target efficiency30-80% per locus
- First approvedCasgevy (exa-cel) · Dec 2023 · sickle cell
- Price~$2.2M per patient (one-time)
- NobelDoudna & Charpentier · Chemistry 2020
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Worked example — curing sickle cell with Casgevy
A 19-year-old with HbSS sickle cell disease has had three vaso-occlusive crises requiring hospitalization in the past year. She enrolls in the Casgevy commercial protocol. The therapy takes about six months from collection to infusion.
Day −60. She receives plerixafor and is mobilized; her CD34+ hematopoietic stem cells are collected by apheresis. The cells ship to a Vertex manufacturing site.
Manufacturing. Engineers electroporate the patient's CD34+ cells with Cas9 protein pre-loaded with a guide RNA targeting the erythroid-specific enhancer of BCL11A. Electroporation efficiency is ~80%; on-target indel rate at the GATA1 binding site is 70-90% across alleles. Edited cells are expanded and cryopreserved. Quality control: indel rate, off-target sequencing of >40 predicted sites, sterility, viability >70%, CD34 content.
Day −7. The patient is admitted for myeloablative conditioning with busulfan (3.2 mg/kg/day × 4 days). This destroys her existing marrow to make room for the edited cells. She becomes pancytopenic.
Day 0. Single IV infusion of her edited CD34+ cells (target dose ≥3 × 106/kg). The cells home to the marrow via CXCR4-CXCL12.
Day +30. Neutrophil engraftment. Edited erythroid progenitors begin producing red cells without BCL11A repression of HbF — fetal hemoglobin (HbF) climbs from <1% pre-treatment toward 30-50% of total hemoglobin.
Day +180. HbF stable at ~40%, total hemoglobin recovered to ~12 g/dL, no vaso-occlusive crises since infusion. In the CLIMB-121 trial, 28 of 29 evaluable patients (97%) were free of severe vaso-occlusive crises for ≥12 consecutive months.
A 35-year timeline
- 1987. Ishino describes repeat sequences (later named CRISPR) downstream of iap in E. coli — unknown function.
- 2002-2005. Mojica, Jansen, Bolotin link CRISPR to bacterial defense — spacer sequences match phage genomes.
- 2007. Barrangou and Horvath at Danisco show CRISPR adaptation experimentally in Streptococcus thermophilus.
- 2012. Jinek, Charpentier, Doudna publish Cas9 + tracrRNA + crRNA programmed with a single guide can cut any DNA in vitro.
- 2013. Zhang, Church show CRISPR works in mammalian cells.
- 2018. He Jiankui announces germline-edited CRISPR babies — global condemnation; he is jailed.
- 2020. Nobel in Chemistry to Doudna and Charpentier.
- 2021. Intellia NTLA-2001 (LNP-mRNA Cas9 vs TTR) — first systemic in vivo CRISPR in humans for transthyretin amyloidosis.
- December 2023. FDA approves Casgevy (exa-cel) for sickle cell — first CRISPR drug. UK MHRA approval one month earlier.
- 2024. Verve Therapeutics shows base-editing PCSK9 sustains LDL reduction in early human data.
Editing modalities at a glance
| Method | Targeting | Edit type | Strengths | Limitations |
|---|---|---|---|---|
| SpCas9 | 20 nt guide + NGG PAM | DSB; NHEJ knockout / HDR knock-in | Most active, well characterized | 4.2 kb, off-targets, DSBs activate p53 |
| SaCas9 | 21 nt guide + NNGRRT PAM | DSB; smaller — fits in single AAV | AAV-deliverable | PAM restrictive |
| Base editor (CBE/ABE) | Guide + PAM; deaminase window | C→T or A→G; no DSB | Single base, no DSB | Bystander edits, off-target RNA |
| Prime editor | pegRNA + nickase Cas9 + RT | Any single-base change; small indels | Most versatile, no DSB | Lower efficiency, large cargo |
| ZFN | Engineered zinc-finger protein | DSB | Compact, defined | Hard to design, slow |
| TALEN | Engineered TAL effector | DSB | Modular design | Larger than CRISPR |
Where CRISPR matters today
- Sickle cell and beta-thalassemia. Casgevy is now offered at qualified centers worldwide.
- Transthyretin amyloidosis. Intellia NTLA-2001 (LNP-mRNA) — single IV dose, >90% sustained TTR knockdown.
- Hereditary angioedema. NTLA-2002 in trials for KLKB1 knockout.
- Hypercholesterolemia. Verve VERVE-101 (PCSK9 base editing) — durable LDL reduction.
- Cancer. CRISPR-edited CAR-T cells; PD-1 knockout in tumor-infiltrating lymphocytes.
- HIV reservoirs. Excision BioTherapeutics in trials for CCR5/proviral targeting.
- Inherited blindness. Editas EDIT-101 (subretinal AAV-Cas9) for CEP290 LCA10 — first in vivo dosing 2020.
Common misconceptions
- CRISPR rewrites the DNA letter by letter. Standard Cas9 makes a break; the cell does the rewriting, and most repairs are imprecise indels.
- One injection of Casgevy and you're done. The patient is conditioned with chemotherapy and receives an autologous transplant — major procedure.
- Off-target editing is solved. Better guides and high-fidelity Cas9 reduce it; unbiased detection is still active research.
- CRISPR edits germline. All approved therapies are somatic (ex vivo or in vivo non-germline). Germline editing remains banned in most jurisdictions.
- CRISPR cures everything. Most diseases are polygenic, require precise edits HDR can't reliably deliver, or have no good delivery to the target tissue.
- Cas9 is the only option. Cas12, Cas13, base editors, prime editors, and CRISPRoff (epigenetic silencing) extend the toolkit.
Frequently asked questions
How does Cas9 find its target?
Cas9 carries a single guide RNA (sgRNA) — a chimeric ~100 nt RNA with a 20-nucleotide spacer at the 5' end (the targeting sequence) fused to a tracrRNA scaffold that holds Cas9. The complex slides and hops along DNA, transiently melting it at every encountered PAM (protospacer adjacent motif). For S. pyogenes Cas9 (SpCas9) the PAM is 5'-NGG-3' on the non-target strand. If the PAM is found, Cas9 strand-invades the adjacent 20 bp; if the spacer matches, Cas9 undergoes a conformational change and its two nuclease domains (RuvC, HNH) each cut one strand, leaving a blunt double-strand break 3 bp upstream of the PAM. Mismatch tolerance increases toward the 5' end of the spacer.
What is NHEJ vs HDR?
Non-homologous end joining (NHEJ) is the default repair pathway — active throughout the cell cycle, fast, but error-prone. Ku70/Ku80 bind the break, DNA-PKcs is recruited, end-processing nucleases and polymerases fill or chew back ragged ends, and ligase IV/XRCC4 ligate. Insertions or deletions of 1-50 bp commonly result, producing frameshifts that knock genes out. Homology-directed repair (HDR) is precise but only active in S/G2 phase when a sister chromatid is available. By providing a synthetic template flanked by homology arms, HDR can install single nucleotide changes, insert tags, or knock in entire transgenes — but efficiency is low (often <10%) and dependent on cell type.
How is Casgevy used to treat sickle cell?
Sickle cell disease is caused by a single A→T mutation in HBB that swaps glutamate for valine at position 6 of beta-globin. Casgevy (exa-cel, made by Vertex/CRISPR Therapeutics) does not fix the HBB mutation directly. Instead, it disrupts the erythroid enhancer of BCL11A — a transcription factor that normally silences fetal hemoglobin (HbF) after birth. With BCL11A knocked out in erythroid cells, HbF is reactivated, and HbF doesn't sickle. Process: harvest patient's hematopoietic stem cells, electroporate with Cas9 RNP + sgRNA ex vivo, expand, condition the patient with busulfan, infuse edited cells. FDA approved December 2023; price ~$2.2 million per patient. >90% of patients are free of severe pain crises in trials.
How efficient is CRISPR editing?
On-target editing efficiency typically ranges 30-80%, depending on locus accessibility, chromatin state, sgRNA quality, delivery method, and cell type. RNP (ribonucleoprotein) delivery of pre-assembled Cas9/sgRNA tends to outperform plasmid or mRNA delivery and reduces persistence/off-targets. Off-target editing is the main safety concern — Cas9 can tolerate 1-4 mismatches at distal positions. Mitigations: bioinformatic sgRNA selection (e.g. CRISPRoff, GuideScan), high-fidelity Cas9 variants (eSpCas9, SpCas9-HF1, HiFi Cas9), shorter sgRNAs, ribonucleoprotein dosing, and unbiased off-target detection (GUIDE-seq, CIRCLE-seq, Discover-seq).
What about base and prime editing?
Base editors fuse a catalytically dead or nickase Cas9 to a deaminase to install single base changes without a double-strand break. Cytosine base editors (CBE) convert C→T (and G→A on the opposite strand); adenine base editors (ABE) convert A→G (and T→C). Verve Therapeutics has used ABE delivered by LNP-mRNA to knock out PCSK9 in liver hepatocytes for permanent LDL lowering — a single-shot cholesterol treatment in trials. Prime editing uses a Cas9 nickase fused to a reverse transcriptase, with a longer guide (pegRNA) carrying both the target sequence and the desired edit. Prime editing can install any single nucleotide change or short insertion/deletion without a double-strand break or repair template.
How is Cas9 delivered to cells?
Ex vivo: harvest cells, deliver Cas9 by electroporation (RNP, mRNA, or plasmid) or lentiviral transduction, return edited cells to patient. This is how Casgevy and most CAR-T programs work. In vivo: more challenging. AAV vectors can deliver Cas9 + sgRNA but SpCas9 is ~4.2 kb — close to AAV's 4.7 kb cargo limit. Smaller orthologs (SaCas9, ~3.2 kb) leave room for promoter and sgRNA. Lipid nanoparticles carrying Cas9 mRNA + sgRNA reached patients first in Intellia's NTLA-2001 for transthyretin amyloidosis — single IV infusion knocked out hepatic TTR with >90% reduction in serum TTR sustained for years.
Where did CRISPR come from?
CRISPR (clustered regularly interspaced short palindromic repeats) was first noticed in 1987 in E. coli (Ishino), then catalogued in many bacteria. Mojica, Jansen, and others showed in 2002-2005 that the repeat-interspersed spacers match phage and plasmid DNA — CRISPR is a bacterial adaptive immune system. Barrangou and Horvath (2007) experimentally proved adaptation in Streptococcus thermophilus. Doudna and Charpentier (2012) showed Cas9 + tracrRNA + crRNA could be programmed with a single guide RNA to cut any sequence in vitro. Zhang, Church, and others adapted Cas9 to mammalian cells in 2013. The 2020 Chemistry Nobel went to Doudna and Charpentier.