One rare surface marker quietly decides what a salivary organoid is

What the work claims

Jeon and colleagues claim that NGFR identifies a conserved, surface-accessible marker of salivary gland epithelial stem/progenitor cells: a restricted basal-duct subpopulation that in humans is strongly enriched for organoid-forming activity, and that in mice behaves as an injury-responsive lineage contributing to both ductal and acinar compartments during post-injury regeneration. NGFR-enriched human organoids also engraft into radiation-injured glands of immunodeficient mice, organizing into duct-like structures that keep an outer NGFR-positive layer.¹

This is a primary cross-species study combining single-cell transcriptomics, prospective fluorescence-activated cell sorting, organoid functional assays, transplantation and genetic lineage tracing. The human regenerative claim rests on organoid and engraftment data; the in vivo regeneration claim is demonstrated only in mice.

How it works

Single-cell RNA-sequencing of two normal human submandibular glands (11,161 cells after quality control) resolved 17 populations, including five epithelial subtypes. Within the basal cluster, NGFR was the standout surface receptor: log2 fold-change of 3.95 with an adjusted p-value below 2.4 x 10^-182. It is a narrow gate. KRT5, the conventional basal marker, covers 77.9 percent of basal cells; NGFR covers 16.9 percent, and 93 percent of NGFR-positive cells sit inside the KRT5 compartment, marking basal cells associated with ducts rather than myoepithelium.¹

Developmental-potency analysis (CytoTRACE2) places NGFR-positive cells earlier: mean score 0.294 versus 0.243 for NGFR-negative basal cells (p < 0.0001), with an oligopotent fraction of 35.2 percent versus 19.1 percent, roughly two-fold. Slingshot trajectory analysis inferred basal-to-duct-to-acinar and basal-to-duct-to-mucous-acinar lineages with NGFR-positive cells concentrated at early pseudotime. Their transcriptional programme is adhesion- and growth-factor-centric: cell-substrate adhesion (adjusted p = 0.0017), wound healing, extracellular matrix organization, and PI3K-Akt, integrin, focal-adhesion and ECM-receptor pathways.

The functional gate is striking. NGFR-positive cells are about 5 percent of dissociated human gland cells, yet after seven days they form organoids at roughly 6.5 percent efficiency against about 0.2 percent for the NGFR-negative fraction - a roughly thirty-fold enrichment. The NGFR-positive fraction also outperformed two incumbent markers, CD24 and ITGA6, under identical conditions. In mouse glands the pattern recapitulates: Ngfr-positive cells are about 3 percent of dissociated cells, form organoids at roughly 5 percent efficiency, and rarely do so from the negative fraction.¹

Inside organoid culture, more than 95 percent of cells are NGFR-positive at the early stage, with expression retreating to an outer layer as acinar markers (AQP5, AMY1A) rise; re-sorting day-15 organoids shows NGFR-high cells have significantly higher secondary organoid-forming efficiency than NGFR-low cells. Transplanted into radiation-injured immunodeficient (NSGA) mice, human organoid cells engrafted for at least four weeks, confirmed by human E-cadherin, with an outer NGFR-positive layer resembling basal ductal organization and a Ki67-positive subset indicating retained proliferation.

In vivo, an Ngfr-CreERT2 knock-in crossed to a Rosa26-tdTomato reporter shows that under homeostasis the lineage stays within KRT7-positive ducts; after two weeks of duct ligation, release and a four-week recovery, tdTomato-positive cells appear in both KRT7-positive ducts and AQP5-positive acini, a subset expressing the mature acinar marker MIST1, with continuous epithelial structures running from ductal toward acinar regions.¹ Ngfr induction also follows a second, inflammatory injury (Concanavalin A). Notably, Ngfr-lineage contribution to epithelia was additionally observed in uterus, tongue, eye and trachea under homeostatic conditions.

Where a skeptic should push

The load-bearing inference for therapeutics is that organoid-forming enrichment equals regenerative capacity, and the paper does not fully carry it. The engraftment experiment is structural: four weeks, human E-cadherin positivity, layered organization. No saliva output, no acinar functional rescue, no comparison against NGFR-negative grafts. A marker that enriches colony formation may still be the wrong starting cell for restoring secretion, which is what a radiation-xerostomia patient actually needs.

Second, the human transcriptomic evidence is two donors, with NGFR-positive fractions of total cells differing more than five-fold between them (0.79 versus 4.15 percent). The lineage-tracing evidence for duct-to-acinar contribution is mouse-only and, critically, qualitative: the text reports increased acinar contribution after deligation but quantifies no fraction, and no clonal analysis establishes that single Ngfr-positive cells are multipotent rather than a mixed population of duct-committed and acinar-committed cells. Third, the comparison against CD24 and ITGA6 is reported as an efficiency ranking without numeric values in the text, so the size of the claimed advantage cannot be verified from the manuscript. Fourth, the cross-organ Ngfr lineage results are a double-edged finding: they support conservation and simultaneously warn that Ngfr marks distinct biological programmes in different epithelia, so nothing about salivary biology can be assumed from the marker alone.

Sorting progenitors changes what the model is

For organoid models of human organs, this paper is a case study in a quiet methodological revolution: marker-gated organoid production. Sorting on NGFR converts "salivary gland organoid" from a whole-tissue average into a defined progenitor product, which is exactly what reproducible disease modelling and regulated cell therapy require. The roughly thirty-fold enrichment and the marker's injury-responsiveness make it a plausible GMP-compatible starting fraction, and the outer-layer NGFR pattern after engraftment doubles as a positional quality-control readout for manufactured grafts.¹

But the same data contain a warning most readers will miss. In vivo, NGFR marks a rare, spatially restricted ductal population; in culture, more than 95 percent of cells are NGFR-positive within days, and the marker's selectivity has to be re-imposed by re-sorting mature organoids. Standard expansion therefore homogenizes the very heterogeneity the marker was defined by. For drug screening, the consequence is sharp: a salivary organoid model of Sjogren's pathology or radiation injury built from unsorted epithelium, and one built from NGFR-sorted progenitors, are different assays with different sensitivities, and the literature rarely says which it used.

The broader generalization failure is the cross-organ one. Because Ngfr-lineage cells also populate uterine, tracheal, ocular and lingual epithelia, the field's habit of treating a surface marker as a portable identity badge - "the NGFR progenitor", as if it were one cell type with one function - is not supported. Marker-to-function claims must be re-earned per organ. The opportunity is a potency-test template: prospective isolation, primary and secondary organoid-forming efficiency, engraftment with positional readouts - a ladder any organoid-derived cell product should climb before therapeutic claims. The threat is the shortcut: engraftment structure mistaken for functional repair, and a duct-biased progenitor mistaken for the acinar replacement the clinic needs.

The bottom line

Established: NGFR is a conserved, sortable marker of a rare basal-duct salivary population that is strongly enriched for organoid formation in humans and mice, and the mouse lineage contributes to ductal and acinar compartments after injury. Not established: that NGFR-sorted human cells restore gland function, that single cells are multipotent, or that the efficiency advantage over CD24 and ITGA6 is of the magnitude the narrative implies - those numbers are in a figure, not the text. What would settle it: competitive transplantation of NGFR-positive versus NGFR-negative fractions with a functional saliva readout, and clonal lineage tracing. As a marker-and-QC story it is strong; as a cell-therapy proof it is a beginning.