NDE1 loss pushes progenitors caudally
The paper's central claim is that NDE1, a centrosome-associated scaffold protein already linked to microcephaly, helps keep early neural progenitors assigned to the forebrain. When NDE1 was knocked out in human pluripotent stem cells, organoids derived from them lost forebrain character and gained caudal character. This is a change in regional identity, not simply fewer cells: the mutants still retained progenitor populations, but those progenitors read out the wrong positional program.
The evidence comes from several angles that converge. Bulk RNA-seq deconvolution of unguided on-chip organoids showed decreased forebrain signatures and increased caudal tissue contributions. Rostral genes including EMX2, LHX2 and PAX6 were down-regulated, while caudal and ventral genes including EN2 and DLX1 were up-regulated, and real-time PCR corroborated the direction. In cortical organoids, single-cell RNA-seq mapped against a first-trimester human brain reference atlas showed reduced forebrain, telencephalic and diencephalic populations and increased midbrain, hindbrain, medulla, pons and cerebellar identities.
That combination matters because it separates two explanations readers often conflate. A smaller organoid could reflect cell death or stalled proliferation alone. Here Ki67-positive and pHH3-positive cells increased while cleaved Caspase-3-positive cells decreased, arguing against increased apoptosis, yet pVim-positive mitotic radial glia and TBR2-positive intermediate progenitors decreased. The authors interpret this as preserved but mis-specified progenitors, shifting toward mixed non-pallial and more caudal transcriptional states.
NDE1 is enriched in forebrain progenitors
Why expect a mitotic protein to influence regional identity at all? The study first located NDE1 in a published single-cell atlas of the first-trimester human developing brain. NDE1 was expressed in cycling progenitors and peaked in G2/M phase, similar to Aurora kinase AURKA, a canonical marker of that phase. Its expression was significantly higher in the forebrain than in the midbrain and hindbrain.
This expression pattern is the setup for the functional work, not proof of mechanism. It tells readers where to look: if NDE1 is enriched in forebrain progenitors and cell-cycle regulated, its loss could plausibly disturb both division timing and positional programs in that region. The paper then tests that by generating four NDE1 knockout hPSC lines, with no protein detected by Western blot in any line, and differentiating them into organoids.
The cell-cycle phenotype is real but should not be read as the whole story. Live imaging of H2B-mCherry and Lyn-GFP organoids showed prolonged mitosis, with significant effects on prophase, metaphase and telophase. NDE1 knockout on-chip organoids also grew more slowly and showed fewer folds than wild-type controls. Those findings place NDE1 among microcephaly genes that regulate the cell cycle, but they do not by themselves explain a change in regional identity.
A microfluidic chip exposes the patterning defect
Standard brain organoids do not reliably reproduce rostrocaudal patterning within a single structure, so the authors used the NeuroMorphoChip (NMC) system, which applies microfluidic morphogen gradients to generate organoids with forebrain-, midbrain- and hindbrain-like regions along a defined axis. This is the assay that makes the regional claim testable rather than inferred from gene lists.
Three markers with distinct domains anchored the readout: PAX6 marks neural progenitors in the forebrain and hindbrain but is largely absent from the midbrain; OTX2 defines anterior identity spanning forebrain and midbrain; HOXB1 marks caudal hindbrain. Wild-type NMC organoids showed the expected axial pattern, with OTX2 marking the rostral domain and HOXB1 the caudal domain. NDE1 knockout organoids disrupted that organization.
The spatial quantification is the most decision-relevant result. Wild-type organoids displayed a bimodal PAX6 pattern, with strong rostral and caudal domains separated by a relative trough in the mid-region. In NDE1 knockouts, rostral PAX6 was reduced and the caudal PAX6 domain became broader and more prominent, a significant shift across the full axis. OTX2 distribution shifted too, with a reduced rostral OTX2 domain and an expanded caudal HOXB1 region. In other words, the boundary between rostral and caudal identities loosened.
The rostral compartment itself is reprogrammed
A spatial shift could simply mean tissue moved. To test whether the rostral tissue changed its own transcriptional program, the authors isolated the rostral third of day 25 NMC organoids for bulk RNA-seq. The rostral compartment of NDE1 knockouts showed notable transcriptional reprogramming.
Genes enriched in wild-type samples included dorsal forebrain-associated factors NEUROG1 and BHLHE22. Knockout samples showed increased expression of genes associated with ventral or non-cortical identities, including LHX6, OTX1, BARHL2, GATA2, SIM1 and SOX8. This supports the authors' interpretation that loss of NDE1 changes what the rostral cells become, not merely how many cells survive there.
Read alongside the cortical organoid data, the picture is consistent: dorsal cortical identity is disrupted while cells move toward mixed non-pallial and more caudal transcriptional states. The paper's wording is careful here. It describes a redistribution of regional identity and a possible shift, and the mechanistic link from NDE1 to those transcriptional states is proposed rather than fully traced.
ERK signaling is aberrant but not uniformly lost
The molecular thread connecting NDE1 to patterning is ERK signaling. Gene set analysis of differentially expressed genes showed robust enrichment for ERK signaling in NDE1 knockout cells across developmental stages, peaking at the pluripotent stage and remaining elevated in brain organoids. Enrichment for nervous system development emerged at day 5 and became more notable at day 18.
The signaling defect is more specific than a simple on-off switch. NDE1-deficient cells showed a blunted ERK phosphorylation response after EGF stimulation, whereas FGF2 induced an abnormal sustained activation profile compared with the transient response in wild-type cells. The authors therefore describe altered activation dynamics in a ligand-dependent manner, and they explicitly note that the precise molecular intermediates linking NDE1-dependent ERK regulation to altered progenitor identity remain to be determined.
This is where readers should keep the paper's own boundary in view. The ERK enrichment and the ligand-specific phosphorylation profiles are measured findings. The claim that NDE1 sits at the interface of signaling pathways coordinating regional specification is the authors' model, supported by the rescue experiment that follows but not by a complete molecular chain.
ERK activation partially restores rostral PAX6
The functional test of the ERK link is a pharmacological rescue in human NMC organoids. Adding PMA, which activates the ERK pathway downstream, expanded the PAX6 rostral domain. The WNT inhibitor XAV939 did the same, and the two together produced a further increase. In the reported comparisons with the control condition, XAV alone gave p = 0.0324, PMA alone p = 0.0352, and XAV plus PMA p = 0.0141, from 3 independent experiments with group sizes of 7 to 9 organoids.
The rescue is partial, and the paper says so. Rostral PAX6 qPCR in organoids, normalized to RPS29 and pooled from two independent experiments with 8 to 10 samples per group, showed a Kruskal-Wallis P = 0.047 with pairwise Mann-Whitney tests marking P < 0.05 and P < 0.01 differences. Quantification of PAX6-positive cell distribution along the rostrocaudal axis compared wild-type, wild-type plus PMA, knockout and knockout plus PMA organoids, with permutation testing reporting P < 0.001 and P = 0.002 for the main contrasts.
The interpretation the authors draw is that the patterning defect is at least in part reversible and that ERK modulation could be a therapeutic entry point. That is a hypothesis-generating statement, not a demonstrated treatment. The readout here is rostral PAX6 expression in organoids, not brain size, cortical folding or behavior, and the WNT result is consistent with WNT's established role in promoting caudal fates rather than an independent NDE1 mechanism.
Mouse embryos confirm the shift, but weakly
Parallel analyses of Nde1 knockout mice confirmed disrupted regional patterning of the forebrain, which is important because it shows the phenotype is not an artifact of one organoid platform. The mouse lines were generated with CRISPR/Cas9 targeting exon 6; the line used in most studies carried an additional A producing a premature stop codon and a predicted 82-amino-acid protein. Embryos were analyzed at E10 to E12 without sex determination, because the authors state that sexual differentiation and hormone-dependent neural dimorphism are not present at those stages.
The species comparison is one of the paper's most useful qualifications. The authors report that patterning defects in human organoid systems were robust and consistently observed across NMC, on-chip and cortical organoids, whereas the corresponding mouse embryo phenotypes were comparatively subtle, with several regional identity markers unchanged. They interpret this as species-specific sensitivity, consistent with the greater complexity and prolonged neurogenic period of the human cortex.
That interpretation is plausible but remains an author explanation. The practical consequence for readers is that the strongest evidence for the caudal identity shift comes from human organoid systems, and the mouse data corroborate disruption without reproducing the same magnitude. Anyone extending these results to human disease timing or to mouse models should carry that asymmetry explicitly.
Where the NDE1 story is strongest
The paper's most defensible contribution is a specific, spatially measured claim: NDE1 knockout reduces the rostral PAX6 domain and expands caudal PAX6 and HOXB1 domains in patterned human organoids, while rostral tissue down-regulates dorsal forebrain genes and up-regulates ventral and non-cortical genes. That is a regional identity phenotype, distinct from the prolonged mitosis and reduced organoid growth that accompany it.
The ERK result then sets the priority for follow-up. Because PMA-driven ERK activation and WNT inhibition each expanded rostral PAX6, and combined treatment expanded it further, the patterning defect appears modifiable through signaling rather than fixed by the mutation. The paper's own framing places NDE1 among centrosomal and microtubule-associated proteins that integrate cytoskeletal organization with signaling, coordinating proliferation with spatial patterning cues.
What remains open is the step between ERK dynamics and progenitor identity. The authors state that the molecular intermediates are undetermined and require future mechanistic work, and the rescue is partial. For readers weighing the clinical relevance for NDE1-associated microcephaly, the key distinction is between a demonstrated organoid patterning phenotype with a partial pharmacological rescue and a proposed therapeutic entry point that has not been tested at the level of brain structure or organismal outcome.
Sources used in this answer
NDE1-Mediated Regulation of Neural Progenitor Regional Identity and Its Implication in Microcephaly
NDE1 knockout shifts human neural progenitors toward caudal identity in patterned organoids, coinciding with aberrant ERK signaling; ERK activation partially restored rostral PAX6 expression, and Nde1 knockout mice showed subtler forebrain patterning disruption.
