Notebook

Search

SearchSearch

Recent writing

See 71 more →

Peptide-mediated delivery of CRISPR enzymes (RNPs) for the efficient editing of primary human lymphocytes - Dana Foss, Joseph Muldoon

Modified: Jan 01, 2025Created: Dec 22, 20248 min read

Background / Motivation: Justin Eyquem’s postdoctoral work at Sloan Kettering:

The old way of doing things: pseudo-random integration using Lentiviral & Gammaretroviral transduction

Targeting a CAR (chimeric antigen receptor) gene to the TRAC (T-cell Receptor Alpha Chain) locus

  1. Cas9 RNP makes cut at locus (arrow in picture)
  2. AAV6 delivers Homology Directed Repair (HDR) repair template to enable precise gene knock-in (instead of random integration), containing
    1. CAR transgene (1928z in diagram)
    2. Homology arms (LHA and RHA in diagram) - match sequences around the cut
    3. 2A and BGHpA are regulatory elements
      1. 2A: self-cleaving peptide sequence that, when translated, causes ribosome to skip making a peptide bond, resulting in separate proteins being produced instead of 1 long fusion protein
      2. BGHpA: bovine growth hormone polyadenylation signal; tells cell where to add poly-A tail to mRNA; poly-A tail for better mRNA stability, processing
  3. Result after HDR: repair of DSB and introduction of CAR into TRAC locus
    1. T-cell expresses your synthetic antigen receptor (CAR) instead of its endogenous antigen receptor (T-cell rec, abbrev. as TCR), so we can redirect the T cell’s specificity to a desired cancer antigen

TRAC targeting confers ideal CAR expression and improves anti-tumor potency

Mouse experiments showed big improvement in survival in NALM6 leukemia mice & much better homogeneity of expression (vs. standard approach: delivering CAR using retroviral vector. Read more here: CAR T)

TRAC-CAR T cell manufacturing for clinical trials

  • Clinical trials done at MSKCC and UCSF, but in both cases, T cells are edited by a CRISPR RNP after electroporation

PERC: Peptide-Enabled RNP Delivery for CRISPR Engineering

  • Problems
    • Electroporation is violent, cumbersome (requires hardware)
    • Primary T cells are extremely difficult to transfect/transduce
      • Lipofection (standard method) does not work
  • Goal: efficient, minimally-perturbative, non-toxic, safe, multiplex editing / T cell engineering

How it works

  1. Mix CRISPR RNPs w/ amphiphilic peptides (10-20 molar equiv.)
  2. Apply mixture from 1. to activated primary human T cells
    • T cells need to be activated for transfection to occur; activated T cells are more amenable to genetic manipulation compared to resting T cells. But why? toAdd
  3. Wait for 3-5 days
  4. Readout editing levels using flow cytometry and/or NGS for biallelic DNA editing

Peptide screening: identifying an effective peptide for RNP delivery to primary human T cells

Got candidate peptides from literature. Result of screen: A5K had highest KO efficiency among those screened and preserved cell viability

Editing data

  • x-axis: TCR KO, measured by flow cytometry These are all derivatives of Hemagglutinin 2 (HA2, found on surface of influenza virus) fused w/ TAT (from HIV-1)
  • E5-TAT
  • INF7-TAT
  • A5K

Cell viability data

  • Summary: PERC is a much gentler delivery method (results in much higher cell viability) compared to electroporated cells
  • Raw data:
    • Top bar = non-treated cells
    • PERC: No viability difference between non-treated and green-bar
    • Electroporated cells lose 1/3 of viable cells vs. non-treated

PERC+AAV to make TRAC-CAR-T cells: the results

  • Summary
    • PERC-treated cells result in improved cell viability & edited cell yield compared to electroporation
      • Cell yield is important for cell manufacturing?
    • 95% knockout and 76% knock-in efficiency with PERC
    • Even more pronounced effect with Cas12a RNP than Cas9. For Cas12a:
      • Editing efficiency (% of cells that successfully received the intended edit) on par w/ electroporation
      • Cell yield 2x higher w/ PERC vs. electroporation
  • Raw data (green = PERC, white = electroporation)
    • Cas9 RNP
    • Cas12a RNP

Flow cytometry data: shows PERC recapitulates the desired expression profile for TRAC-CAR

  • The graph
    • Top row = TRAC KO only
    • Bottom row = TRAC-CAR (CAR gene knocked into TRAC locus)
  • Shows PERC + AAV reproduces expected pattern of expression in TRAC-CAR cells (matches w/ expression profile of TRAC-CAR cells made from electroporation)
    • Pattern of expression/expression profile = how much certain proteins are present on the surface of T cells, specifically
      • TCR (T cell rec), shown on y-axis
      • CAR (chimeric antigen rec), shown on x-axis
  • We see
    • Main cell populations shift downward ⇒ both e-por and PERC reduce TCR expression (expected when targeting TRAC)
    • Note: interesting that there is a population of cells in the upper-left quadrant for PERC, TRAC-KO. this maybe suggests PERC is less efficient at TRAC KO but the reduction in efficiency has no functional impact on successful CAR integration. Maybe it’s not a reduction in efficiency but a result of less cell damage?

PERC minimally perturbs T cell gene expression (transcriptome)

  • Measured T cell mRNA levels using a NanoString CAR-T characterization panel (770 genes)whatIsThis
  • How could electroporation and PERC disturb the T cell transcriptome?
    • Elecktroporation: physical stress triggers stress response pathways, damages cellular structures, may trigger survival/death pathways, disrupts ion balance and membrane potential
    • PERC: chemical stress (exposure to DMSO), endosomal disruption for RNP delivery etc.
  • Data in volcano plots (red = significantly upregulated genes, blue = sig. downreg. genes, grey = non-significant changes. X-axis = magnitude of change, Y-axis = statistical significance)
    • 3 panels
      1. DMSO vs. n.t. (no treatment) - to isolate how much DMSO solvent alone affects T cells
        1. Treatment: T cells + DMSO (solvent used to dissolve peptides)
        2. Control: untreated T cells
      2. PERC vs. n.t. - to measure effect of PERC + DMSO on T cell mRNA levels
        1. Treatment: T cells + PERC (peptide + RNP + DMSO)
        2. Control: untreated T cells
      3. PERC vs. DMSO - to measure effect of PERC (no DMSO) on T cell mRNA levels
        1. Treatment: T cells + PERC (peptide + RNP + DMSO)A
        2. Control: T cells + DMSO
    • Interpretation
      • 1st panel: DMSO alone has some effect on T cell mRNA levels, but none are statistically significant
      • 2nd panel: PERC + DMSO shows stat sig changes in T cell transcriptome
      • 3rd panel: when controlling for DMSO (only examining effects of PERC on T cell transcriptome), almost all stat sig changes in T cell mRNA levels disappear
    • Conclusion: the PERC method itself (peptide + RNP, excluding DMSO) minimally disturbs T cell gene expression (transcription) levels, especially when compared to electroporation, see:

Sequential (multiple rounds of) editing is possible with PERC (but impossible w/ electroporation)

Sequential editing

  • An important benefit of sequential editing is that is reduces risk of chromosomal translocations (compared to simultaneous editing of multiple loci)
  • The state of sequential editing
    • :( Electroporation ⇒ cell death/viability issues: each round of electroporation causes a lot of cells to die, cells become more fragile after each round; after 3rd round of e por, most cells are dead
    • :) PERC ⇒ maintains high cell viability after multiple rounds of treatment

Demonstrated success of sequential PERC delivery (double-locus editing)

  • Successfully performed 2 rounds of sequential edits using RNP w/ better cell yield than electroporation and minimal translocations:
    1. TRAC-CAR knock-in, then
    2. B2M KO

Triple-locus editing: PERC shows improved edited yield over electroporation but lower editing efficiency

  • Performed in both CAR-T and T cells
    1. TRAC-CAR knock-in, then
    2. B2M KO, then
    3. CD5 KO
  • Observed
    • Decreasing editing efficiency with each round of editing (expected bc as you make more edits, the probability of one cell containing all edits goes down - multiplication rule of probability)
    • PERC has significantly higher edited cell yield, but lower intended edit efficiency % compared to electroporation
      • the higher edited cell yield compensates for the lower editing efficiency,
    • Phenotypic observations of cells: PERC maintains T cell differentiation state (resembles that of non-treated cells), whereas electroporation pushes T cells toward undesirable terminal state
      • Method: stained edited CD4+ and CD8+ T cells for several markers that indicate the differentiation state of those T cells after 3 rounds of editing

PERC produces similarly functional, potent (anti-tumor) CAR-T cells to its electroporation counterpart

  • Methodology: did 2 rounds of sequential edits (1. TRAC-CAR knock-in, then 2. B2M KO), then tested those CAR-T cells in NALM6 leukemia mouse line (used mice w/ similar tumor burden)
  • Result: similar mouse survival rate in those injected w/ e-por CAR-T cells vs. PERC CAR-T cells

Summary: PERC is an attractive alternative to electroporation for therapeutic T cell engineering

T-cell engineering methodComponentsPrecision genome engineeringEngineered cell yieldTranscriptomic impactSerial CRISPR multiplexingRequires dedicated hardware?
Lentiviral/gammaretroviral transduction (oldest method)

- viral vector		Just the virus (lentivirus or gamma retrovirus)NoHighMinimalN/ANo
Electroporated RNP with AAV HDRT

- RNP delivery: electroporation
- Gene template delivery: AAV		CRISPR RNP + AAVYesLimitedSubstantialImpracticalYes
PERC w/ AAV HDRT

- RNP delivery: PERC
- Gene template delivery: AAV		CRISPR RNP + peptide + AAVYesHighMinimalYes!No
  • all results described above apply to both CD4+ and CD8+ T cells (no significant difference between the 2 types of T cells)

Questions

  • How do peptides mediate cellular uptake and endosomal escape of RNP? No clear understanding of mechanism of endosomal escape at the moment
  • Peptides have cell-type specificity (e.g. some sequences work with liver cells but not T cells)
    • Need to screen first for different tissues to see if it works
  • Load size of PERC: PERC is able to deliver a base editor (beeg, size-wise)
  • Future work:
    • Wilson lab testing PERC on HSPCs
    • haven’t tested if PERC can deliver CRISPRi or CRISPRa yet

My Summary

Graph View

Backlinks