BGPT: Paper Review: Expression landscape of the genetic hearing loss protein whirlin across human tissues and cell types

Fuel Your Discoveries

Quick Explanation Copied

Whirlin (WHRN/DFNB31): a body-wide expression atlas grounded in HPA RNA + IHC

This paper builds a “where is WHRN expressed?” map by integrating HPA bulk RNA (HPA + GTEx), HPA single-cell RNA (34 datasets; 154 cell types), and HPA validated antibody IHC on tissue microarrays, finding near-ubiquitous RNA presence (49/51 tissues) with strong cell-type enrichment (endocrine, ciliated, glandular epithelial) and protein staining patterns generally consistent with RNA.

Key skepticism: the work is an atlas (not a mechanistic study), so “expression here” does not imply “function here,” and IHC/FFPE TMA sampling + antibody specificity can shape which signals appear.

Long Explanation

Paper review (science-first): WHRN expression atlas

Target paper: “Expression landscape of the genetic hearing loss protein whirlin across human tissues and cell types”

Main claim

WHRN is broadly expressed across human tissues, with RNA-level enrichment driven by endocrine, ciliated, and specialized epithelial cell types, and IHC patterns generally consistent with RNA.

Evidence types

Bulk RNA consensus (HPA + GTEx), scRNA (34 datasets; 154 cell types), and antibody IHC on TMAs (normal + retina + cancers).

FIGURES (from the paper’s explicit reported counts; no invented numbers)

Figure A. WHRN RNA detection across tissues (49/51)

Data source (explicit counts in the manuscript text): 49/51 tissues detected.

Figure B. Highest IHC expression frequency reported for select cancer types

The manuscript explicitly states: lymphoma (10/11 high), colorectal (10/10 high), stomach ("high to medium"; high-only count not fully specified), thyroid (2/3 high).

Figure C. Experimental design “data fusion” overview (what is integrated)

The paper states it integrates bulk RNA (HPA consensus with GTEx), scRNA (HPA, multiple datasets), and IHC on TMAs using an antibody validated in the HPA workflow; data are available via HPA v25 portal. Antibody validation principles supporting skepticism:

What the paper actually did (and how strong that is)

1) Bulk RNA landscape: broad detection with organ-level peaks

The manuscript combines HPA internal bulk RNA with GTEx, building a normalized consensus across 51 tissues and reporting WHRN RNA present in 49 of 51 tissues.

It also highlights that inner ear tissues are not represented in these bulk datasets, so the “primary WHRN functional site” is missing from bulk-RNA inference.

2) Single-cell RNA: cell-type enrichment explains organ-level peaks

The study uses HPA scRNA data spanning 34 datasets and covering 154 cell types, and claims that WHRN expression is not uniformly distributed across cell types, with strongest scRNA levels in endocrine cells (e.g., adrenal cortex), second in ciliated cells, plus additional moderate enrichment in neuroendocrine and specialized epithelial subsets.

3) Protein validation via IHC: broad detection with cytoplasmic staining and cell-type patterns

The manuscript reports IHC on TMAs for 45 normal tissues plus retina samples (extended), stating WHRN is broadly detected with primarily cytoplasmic staining and intensity differences across tissues/cell types; it also emphasizes that retina staining supports antibody validation.

However, as an atlas paper, mechanistic inference is limited: expression alignment across modalities increases plausibility, but cannot establish causality for WHRN’s roles in non-sensory tissues. This limitation is structural to atlas-style studies.

4) Cancer: “mirrors normal tissue origin” with notable exceptions

For 20 common cancer types, the manuscript reports generally comparable WHRN staining to corresponding normal tissues, with variability between patients. It reports the strongest IHC in lymphoma (10/11 high) and colorectal cancer (10/10 high), and also notes strong cases in stomach (high-to-medium in 8/12) and thyroid (2/3 high).

5) Cell lines: WHRN RNA and some protein detectability in endogenous-expressing lines

The study analyzes a large HPA cell line panel and reports WHRN detection across a broad range of lines, with particularly high RNA in myeloma-derived lines (nTPM often above 65) translating into detectable WHRN protein based on proteomics evidence from a Pan-Cancer Atlas resource. It also notes endogenous WHRN in commonly used lines such as HEK293, HeLa, and MCF-7.

Skeptical critique: key blindspots & where the atlas can mislead

A) “Expression ≠ function.”

The paper’s strongest biological statements are framed as hypotheses (“may contribute to…”), which is appropriate. But users should resist the intuitive leap from “WHRN is present in endocrine/ciliated epithelial cells” to “WHRN is required for their specific biology,” because the study does not include perturbation experiments in those tissues.

B) Inner ear missing from bulk RNA → incomplete tissue triangulation

Because inner ear tissues are absent from the bulk RNA datasets, the atlas cannot quantify WHRN distribution in the primary sensory site using the same bulk-RNA framework.

C) Antibody/IHC specificity and scoring granularity

IHC relies on antibody specificity and on subjective/semiautomated manual annotation. The paper describes HPA workflows and antibody validation logic, but any residual risk remains: staining intensity scoring (0–3) can collapse heterogeneous signal distributions, and TMAs sample small representative cores. These risks are consistent with antibody validation methodology emphasizing need for rigorous evidence.

D) Cancer sampling: limited cases per tumor type and subtype heterogeneity

The manuscript itself acknowledges a constraint in cancer analyses: tumor sample numbers per type (up to ~12 patients) can be too small to draw robust biological conclusions across heterogeneous subtypes.

E) Cross-platform reconciliation still probabilistic

Even when RNA and IHC “agree,” differences in transcript/protein correlation, isoforms, and detection limits can still produce discordant outcomes. This is a general limitation of multi-omics atlases and antibody-based proteomics resources. The paper’s contribution remains valuable as an organized hypothesis generator.

Context: why these “non-sensory” expression patterns are plausible

WHRN/whirlin is known as a scaffolding protein with multiple interaction domains and established roles in sensory cell stereocilia organization and retina complexes; earlier work also indicates neural expression and pleiotropic roles beyond core sensory functions, making non-sensory expression observations biologically plausible.

The paper’s inferred mechanism themes—membrane-cytoskeletal organization, secretion/trafficking complexes, and PDZ-based signaling scaffolds—are consistent with whirlin’s domain architecture and known roles in specialized structures (stereocilia/photoreceptor complexes).

What would change my mind? (falsification targets)

If orthogonal protein assays (e.g., targeted proteomics with validated peptides) show that WHRN is not expressed in the cell types identified by HPA scRNA/IHC, then the atlas-driven “cell-type enrichment” story would weaken. This follows directly from the atlas nature and antibody limitations discussed via validation principles.
If future datasets that include inner ear tissues in comparable bulk/protein formats show WHRN absence or qualitatively different distribution, then the paper’s “broad expression beyond sensory systems” emphasis would need revision.
If cancer WHRN staining patterns in larger, subtype-stratified cohorts show strong cancer-specific upregulation independent of normal tissue origin, then the “mirrors normal tissue origin” interpretation would be falsified.

Buttons: continue exploring on BGPT

Author reviews

Feedback:

Updated: June 14, 2026