BGPT: Paper Review: Efficient manipulation of gene dosage in human iPSCs using CRISPR/Cas9 nickases

Explore by Goal

Quick Explanation Copied

Core idea: paired CRISPR/Cas9 nickases are engineered (Cas9nVQR; NGA PAM) to efficiently generate scarless, footprint-free APP copy-number isogenic iPSCs, enabling dose-dependent readouts in neurons.
Most testable payoff: APP dosage restoration (trial/bi/monoallelic states) correlates with changes in Aβ secretion, Thr231 Tau phosphorylation, and a transcriptomic signature implicating an PEG3 → Bax → cytochrome-c → apoptosis axis.

Long Explanation

Paper Review (Visual-first): Efficient manipulation of gene dosage in human iPSCs using CRISPR/Cas9 nickases

Communications Biology • DOI: 10.1038/s42003-021-01722-0

Received 20 Mar 2020; accepted 15 Jan 2021 (per provided paper header).

1) Experimental pipeline (what they actually did)

Engineered nickase: Cas9 nickase variant “Cas9nVQR” is created to broaden PAM compatibility to NGA, increasing genome targetability relative to the standard nickase.
Scarless “double-nicking”: two sgRNAs (sgRNA1/sgRNA2) are paired; they require both functional nickases to create proximal single-strand events enabling desired genome editing.
Proof-of-concept genetic system: starting from an Alzheimer’s patient-derived iPSC line carrying APP duplication, they generate isogenic iPSC lines with APP monoallelic, biallelic, or triallelic knockout states.
Differentiation + readouts: iPSCs are differentiated into cortical neurons using an Ngn2-mediated protocol; then they quantify Aβ secretion, APP protein, Tau (Thr231), and apoptosis-related markers/pathways.

2) Visualizing the engineering payoff (targetability + cloning efficiency)

3) From cells to isogenic clones: efficiency checkpoints

After GFP-based sorting in iPSCs, they report editing efficiency 40 ± 3.5% in iPSCs.
From 12 screened iPSC clones, they identify 4 with potential editing at the target region (then validate by sequencing/CGH).

4) Do dosage changes translate into AD-relevant neuronal phenotypes?

Key reported dose-dependent rescues (directional)

Aβ secretion: compared with nondemented control (NDC), APP-duplication iPSCs secrete significantly more Aβ38/Aβ40/Aβ42; correction in isogenic Iso cells restores secretion toward NDC levels.
APP protein level: Iso neurons show APP protein decreases close to NDC, consistent with copy-number correction.
Tau pathology: APP-duplication neurons show higher Tau phosphorylation at Thr231 and altered Tau measures (pThr231/total Tau ratio, total Tau); Iso correction reduces these levels.

5) Mechanistic hypothesis: PEG3 → Bax-mediated apoptosis

This section is where the paper moves from “phenotypes” to “mechanisms,” so it benefits from the most skeptical appraisal.

Transcriptomics (isogenic DEGs): between APP and Iso neurons, the authors report 376 differentially expressed genes (208 up, 168 down at FDR < 0.05 and fold-change ≥ 1.5).
Shared candidates: they identify 207 “common” genes shared between isogenic (APP vs Iso) and non-isogenic (APP vs NDC) comparisons, suggesting APP-dose dependent effects rather than purely non-isogenic noise.
PEG3 selection: they focus on PEG3, reporting that PEG3 mRNA/protein is higher in APP neurons than NDC and returns toward baseline upon APP dosage correction in Iso neurons.
Functional apoptosis readout: they report reduced cleaved caspase-3-positive apoptotic cells in Iso compared with APP neurons.
Bax translocation / cytochrome-c release intervention (causality attempt): they report that inhibiting Bax translocation or cytochrome-c release reduces cleaved caspase-3 and propidium iodide-positive cells in APP neurons relative to vehicle.

6) Skeptical critique (what’s strong vs what’s missing)

Strong points (supported by the paper text you provided):

Isogenic design logic is explicit: they compare APP vs Iso (dose correction within the same editing system) and also compare APP vs NDC to estimate non-isogenic background influence, then focus on overlap genes.
Footprint-free framing and deep sequencing/Sanger/CGH validation are used to substantiate APP copy-number edits.
Mechanistic follow-up goes beyond marker association by perturbing Bax/cytochrome-c steps and quantifying apoptosis readouts.

Primary blind spots / uncertainties (from the provided text only):

Off-target assessment depth: the paper text you provided does not include a comprehensive genome-wide off-target risk report (e.g., unbiased DSB/indel profiling at scale). This is important because Cas9-based systems can produce off-target activity; nickases reduce the probability but do not eliminate it.
Single genetic background: the experiment is anchored on one APP duplication iPSC line and one nondemented control line (male donors), which limits generalization: dose-dependent phenotypes might depend on additional patient-specific variants.
Interpretation of PEG3→Bax mechanism: PEG3 is upregulated and Bax-pathway inhibition rescues apoptosis markers, but the paper text provided does not show direct “PEG3 causality” (e.g., PEG3 perturbation epistasis) as a necessary link. Thus, the model is plausible yet not uniquely determined from the presented information.
Cell-state specificity: neuronal outcomes are measured in a differentiated cortical neuron state (Ngn2-mediated; Cux1+ upper-layer enriched per their markers). Different differentiation maturity or cell-type composition could alter pathway emphasis.

7) What would most disprove the paper’s core claim?

Editing-specificity challenge: If APP dosage states were accompanied by substantial unaccounted genomic changes that drive Aβ/Tau/apoptosis readouts, the “dose causality” inference would weaken. The paper’s provided text doesn’t show exhaustive off-target/structural rearrangement profiling beyond CGH and target-region sequencing.
Phenotype replication across clones: If multiple independently corrected Iso clones did not show consistent dose-dependent rescue, clone-specific effects could dominate. The provided text mentions an “Iso” clone but does not quantify how many independently corrected clones reproduce each phenotype in detail.

8) Data availability signals (reproducibility anchors)

Deep-seq (editing): deposited in NCBI SRA as PRJNA681926.
RNA-seq (transcriptomics): GEO accession GSE160224.

9) Explore author-specific reviews

Feedback:

Updated: April 04, 2026