The research team, led by Giana Izzo and Andrea Righini of Ospedale dei Bambini Vittore Buzzi in Milan together with colleagues at the University of Naples Federico II and the University of Milan, studied four fetuses in which fetal MRI had raised suspicion of a cortical malformation. After termination of pregnancy following multidisciplinary counseling, each fetus underwent post-mortem MRI within 24 hours of delivery, on fresh specimens stored at 4 to 5 degrees Celsius, followed by complete neuropathological workup including immunohistochemistry for markers such as NeuN, GFAP, CD34, CD3, CD8, and myelin basic protein. The four cases were deliberately chosen to represent the major etiological categories of polymicrogyria: an ischemic form, an infectious form caused by congenital cytomegalovirus, and two genetic forms driven by mutations in PIK3CA and ADGRG1/GPR56.

What makes the study technically remarkable is the imaging resolution achieved. The post-mortem examinations were performed on a 1.5-Tesla scanner using high-resolution T2-weighted fast spin-echo sequences with an in-plane resolution of roughly 0.26 square millimeters, approximately four times finer than what in vivo fetal MRI can achieve. At 21 weeks of gestation, the fetal cortex is organized into a transient laminar architecture, with distinct layers including the cortical plate, the subplate containing thalamocortical fibers, the intermediate zone, the periventricular zone, and the germinal zone. Post-mortem MRI could resolve these layers individually, allowing the researchers to detect precisely where and how the normal laminar pattern broke down in each case, something in vivo fetal MRI, limited by motion artifacts and field strength, could only hint at as vague cortical irregularity.

The first case illustrated the ischemic pathway. This fetus was the surviving co-twin of a monochorionic pregnancy complicated by twin-to-twin transfusion syndrome treated with laser therapy, after which the donor twin died in utero. Fetal MRI performed roughly two weeks after the procedure, following a prior scan that had shown no abnormalities, revealed volume loss in one hemisphere and a wart-like irregularity of the cortical plate in perisylvian frontoparietal regions. Post-mortem MRI sharpened the picture considerably, showing a wavy, irregularly thickened cortical plate that appeared to extend beyond its expected boundary, a markedly reduced subplate, and an irregular thickening of the leptomeningeal layer. Histology confirmed the classic festooned appearance of polymicrogyria and revealed foci of laminar necrosis, micro-regional disruption of the pial basement membrane, ectopic neurons and glial cells outside the cortex, macrophage infiltration, and proliferating meningeal vessels, all signatures of an acute, regionally selective vascular insult striking a developing brain.

The second case, driven by congenital cytomegalovirus infection confirmed by PCR of amniotic fluid, presented a fundamentally different pattern. CMV has a well-documented tropism for radial glial progenitor cells, and the imaging showed bilateral but strikingly asynchronous injury. One hemisphere already displayed an established polymicrogyria pattern, with a thickened irregular cortical plate, an indistinct subplate, a disrupted germinal zone, and a markedly hypointense rim along the cortical ribbon on T2-weighted images, corresponding to laminar necrosis and calcifications that matched areas of pronounced diffusion restriction seen on in vivo imaging. The other hemisphere showed an earlier stage of the same process, with edematous T2 hyperintensity, a thickened cortical plate poorly delineated from the subplate, and, notably, an undulating appearance of the thalamocortical fiber band, as if the cortical plate were crumpling onto the subplate beneath it. Immunohistochemistry confirmed diffuse laminar disorganization, intense reactive gliosis, lympho-macrophage infiltration on CD3 and CD8 staining, and, crucially, an intact pial basement membrane, distinguishing this inflammatory mechanism from the ischemic case.

The two genetic cases demonstrated that malformations arising from mutations can look entirely different from acquired forms. The third fetus carried a somatic mosaic PIK3CA mutation, identified through next-generation sequencing of a biopsy from an overgrown limb, and showed hemimegalencephaly with unilateral polymicrogyria. Post-mortem MRI revealed expansion and blurring of the intermediate zone and subplate, globally reduced T2 signal suggesting increased cellularity in every layer, and loss of the normal demarcation of thalamocortical fibers. Histology found an excess of mature neurons in the superficial cortical layers, an inversion of the normal inside-out neuronal layering, while the pial basement membrane and leptomeninges remained intact and no gliosis was present. The researchers attribute this profile to hyperactivation of the PI3K-AKT signaling pathway, driving progenitor hyperproliferation and impaired apoptosis rather than tissue destruction.

The fourth case involved a mutation in ADGRG1, also known as GPR56, a gene encoding a receptor expressed on the end feet of radial glial cells that anchors them to the pial basement membrane. When that anchorage fails, neurons overmigrate through gaps in the pial barrier, producing a mixed picture. Fetal MRI had shown frontal bumps along the cortical rim, subependymal heterotopic nodules, cerebellar vermian hypoplasia, and flattening of the pons. Post-mortem MRI delineated a thickened, irregular cortical plate with focal mushroom-like protrusions and a cobblestone-like surface at the frontal vertex, alongside polymicrogyria-like changes in parietal regions. Histology confirmed the coexistence of both patterns: festooning cortical bands with vessels entrapped in fused sulci characteristic of polymicrogyria, and complete laminar disorganization with meningeal vessels trapped in the subplate, pial basement membrane discontinuity, and neuronal overmigration characteristic of cobblestone cortex, placing this condition within the mechanistic spectrum of dystroglycanopathies.

Taken together, the four cases allow the authors to propose a set of differentiating imaging features that map onto underlying mechanisms. Symmetry of involvement separates the bilateral, asynchronous infectious form from the vascular-territory-limited ischemic form and the hemispheric overgrowth of PIK3CA disease. The integrity of the pial basement membrane divides the cases sharply: it is focally breached in ischemic injury and diffusely disrupted in ADGRG1-related disease, but preserved in both CMV infection and PIK3CA-related overgrowth. The presence of an extracortical layer, the pattern of laminar alteration, the state of the subplate, and associated anomalies such as cerebellar hypoplasia or hemispheric enlargement each carry diagnostic weight. In the CMV case, the undulating thalamocortical fiber band stands out as a particularly evocative early marker, suggesting that disturbance of the cortical plate-subplate interface may be a harbinger of abnormal folding before any overt gyral anomaly appears.

The authors are careful about the limits of their work. The series comprises only four deliberately selected fetuses, examined retrospectively and without blinded image interpretation, and post-mortem correlation was available only in cases of pregnancy termination, which may not represent the full clinical spectrum of polymicrogyria. No quantitative measurements of cortical thickness were performed, the fetal imaging was acquired at 1.5 Tesla rather than the higher-resolution 3-Tesla systems now optimized for fetal work, and the interval between prenatal and post-mortem scanning may have introduced tissue changes affecting comparison. Direct translation of post-mortem findings to routine in vivo fetal MRI remains constrained by differences in acquisition conditions and tissue characteristics. The patterns described are therefore framed as hypothesis-generating observations that require validation in larger, independent cohorts with standardized quantitative approaches.

Even with those caveats, the conceptual payoff is substantial. The study reinforces the view, increasingly dominant in developmental neuropathology, that polymicrogyria is a final common pathway rather than a single entity, shaped by genetic mutations, vascular insults, viral infection, and metabolic disturbance acting at different moments of corticogenesis. It also elevates the pial surface and its basement membrane to a central orchestrating role in normal cortical development, since the integrity or breakdown of that interface emerges as the key anatomical hinge separating the major pathogenic routes. For families facing a prenatal diagnosis of a cortical malformation, the practical hope is that recognizing these early signatures, particularly the state of the laminar pattern and the pial boundary, will sharpen differential diagnosis, guide targeted genetic testing, and ultimately improve the counseling offered for future pregnancies, turning a descriptive label into a mechanistic explanation.

Subject of Research: Early laminar disruption in fetal incipient polymicrogyria characterized by post-mortem MRI and neuropathological correlation

Article Title: Laminar pattern disruption in fetal incipient polymicrogyria: insights from post-mortem magnetic resonance imaging and neuropathological correlation

Article References: Izzo, G., Tortora, M., Toto, V., Lanna, M., Doneda, C., Arrigoni, F., Tortora, F., Parazzini, C., & Righini, A. (2026). Laminar pattern disruption in fetal incipient polymicrogyria: insights from post-mortem magnetic resonance imaging and neuropathological correlation. Pediatric Radiology. https://doi.org/10.1007/s00247-026-06770-4

Image Credits: AI Generated

DOI: 10.1007/s00247-026-06770-4

Keywords: polymicrogyria, fetal MRI, post-mortem MRI, cortical development, neuropathology, cytomegalovirus, PIK3CA, ADGRG1, pial basement membrane, cortical lamination, twin-to-twin transfusion syndrome, Pediatric Radiology

UBAP2L first attracted attention from geneticists through the patients themselves. Rare variants in the UBAP2L gene have been repeatedly identified in individuals diagnosed with neurodevelopmental disorders, including autism spectrum disorder, a condition that affects communication, social interaction, and learning in millions of families worldwide. What those human genetic findings could not establish, however, was mechanism. Correlation between a mutated gene and a disorder does not reveal whether the protein acts during embryonic brain construction, in mature neurons, or in both. The new study was designed to answer precisely that question by removing the gene selectively from the developing nervous system of laboratory mice and watching what breaks.

The team used a well-established genetic engineering strategy known as the Nestin-Cre conditional knockout system. Nestin is a marker gene active in neural stem and progenitor cells, so pairing it with the Cre recombinase enzyme allows researchers to delete Ubap2l specifically and efficiently in the embryonic nervous system while leaving the gene intact everywhere else in the body. The researchers generated two groups of animals: heterozygous mutants, called cHET, carrying one functional copy of the gene, and homozygous conditional knockouts, called cKO, in which both copies were deleted from neural tissue. This design also mirrors the human genetic situation, where many patients carry a single damaged copy.

The anatomical consequences were unambiguous. Adult cKO mice weighed significantly less than their normal littermates, and histological examination of their brains revealed a marked reduction in the number of mature cortical neurons. The cortex, the folded outer layer of the brain responsible for perception, cognition, and voluntary movement, is built in an inside-out sequence: the earliest-born neurons settle in the deepest layers while successive waves of newborn neurons migrate past them to form the outer layers. At embryonic day 18.5, near the end of gestation, the knockout mice showed abnormal cortical lamination, meaning this precisely choreographed layering had gone wrong. The blueprint of the cortex had been corrupted at its source.

Tracing the defect further back in time, the researchers examined embryos at embryonic day 14.5, the height of the cortical neurogenic period when neural progenitor cells are dividing furiously to produce the neurons that will populate the cortex. At this stage, the abundance of cortical neural progenitor cells was already reduced, and the proliferation of the remaining progenitors was impaired. In other words, the root cause of the later cortical disorganization was not a failure of neurons to migrate or survive, but a failure of the stem cell pool to expand and deliver its full quota of neurons on schedule. Fewer progenitors dividing less efficiently means fewer neurons, and fewer neurons means a thinner, mislayered cortex.

Behavioral testing translated the cellular damage into functional terms. The cKO mice exhibited impaired social interaction and deficits in learning and memory, phenotypes that resonate strongly with the diagnostic features of autism spectrum disorder and related neurodevelopmental conditions. Importantly, the cHET mice, with only one copy of the gene deleted, showed milder behavioral and cognitive impairments. This dose-sensitive pattern, where gene copy number tracks with symptom severity, is a hallmark of many neurodevelopmental disorders and strengthens the argument that the mouse model faithfully recapitulates the human condition associated with UBAP2L variants.

To understand what was happening inside the cells, the researchers performed transcriptomic profiling, sequencing the RNA of the embryonic day 14.5 cortex to measure the activity of every gene. The differentially expressed genes clustered around three biological themes: transcriptional regulation, neuronal differentiation, and cell proliferation, with proliferation-related genes predominantly downregulated. Pathway analysis sharpened the picture further, revealing dysregulation of three major signaling cascades known to steer brain development: the PI3K-AKT pathway, which controls cell growth and survival; the MAPK pathway, which transmits proliferative signals; and the Wnt pathway, a master regulator of embryonic patterning and progenitor behavior. The convergence of these three pathways on a single deleted gene suggests that UBAP2L sits at or near a hub coordinating the proliferative program of neural progenitors.

The most mechanistically revealing experiments came from direct cell-cycle analysis. By examining cell-cycle phases in the embryonic cortex and running flow cytometric analysis on UBAP2L-deficient cells in culture, the team demonstrated that loss of Ubap2l produces two distinct cell-cycle defects: the S phase, during which DNA is replicated, became abnormally shortened, and cells accumulated in the G2/M checkpoint, indicating arrest before or during mitosis. A shortened S phase raises the specter of incomplete or error-prone DNA replication, while G2/M arrest halts division altogether. Together, these defects would throttle the output of the progenitor pool exactly when the cortex can least afford it. The cell cycle, it turns out, is where UBAP2L exerts its essential function.

The rescue experiments delivered the study’s most compelling verdict. When the researchers re-introduced wild-type UBAP2L into the deficient cells, the cell-cycle defects were reversed, confirming that the phenotype was specifically caused by the absence of this protein rather than by collateral damage from the engineering process. Critically, however, UBAP2L variants associated with neurodevelopmental disorders in patients failed to rescue cell-cycle progression. This is the functional smoking gun: the very mutations found in affected individuals abolish the protein’s ability to support normal cell division, directly linking patient genetics to a concrete cellular failure mode.

The broader implications reach well beyond a single gene. UBAP2L is known to participate in RNA processing and the assembly of membraneless cellular compartments called stress granules, and this study now positions it as a bridge between the cell’s gene-expression machinery and the proliferative demands of embryonic neurogenesis. For the growing number of families carrying UBAP2L variants, the work offers a biological explanation for their children’s conditions and a defined cellular process, progenitor cell-cycle control, as a potential point of therapeutic intervention. More generally, the study adds to a compelling emerging theme in neuroscience: many autism-related genes converge not on synapses or neurotransmitters, as once assumed, but on the fundamental mechanics of how the embryonic brain builds itself, one dividing progenitor at a time.

Subject of Research: The role of the UBAP2L protein in neural progenitor proliferation, cortical development, and autism-related neurodevelopmental disorders

Article Title: Neural-specific UBAP2L deficiency causes developmental delay and cortical abnormalities in mice

Article References: Neural-specific UBAP2L deficiency causes developmental delay and cortical abnormalities in mice. (n.d.). https://doi.org/10.1007/s00018-026-06416-6

Image Credits: AI Generated

DOI: 10.1007/s00018-026-06416-6

Keywords: UBAP2L, autism spectrum disorder, neurodevelopmental disorder, cerebral cortex, neural progenitor cells, cell cycle, conditional knockout mice, PI3K-AKT pathway, MAPK pathway, Wnt signaling, cortical lamination, developmental delay