A Human Stem Cell-Derived Neurosensory-Epithelial Circuitry on a Chip to Model Herpes Simplex Virus Reactivation

Abstract

Both emerging viruses and well-known viral pathogens endowed with neurotropism can either directly impair neuronal functions or induce physio-pathological changes by diffusing from the periphery through neurosensory-epithelial connections. However, developing a reliable and reproducible in vitro system modeling the connectivity between the different human sensory neurons and peripheral tissues is still a challenge and precludes the deepest comprehension of viral latency and reactivation at the cellular and molecular levels. This study shows a stable topographic neurosensory-epithelial connection on a chip using human stem cell-derived dorsal root ganglia (DRG) organoids. Bulk and single-cell transcriptomics showed that different combinations of key receptors for herpes simplex virus 1 (HSV-1) are expressed by each sensory neuronal cell type. This neuronal-epithelial circuitry enabled a detailed analysis of HSV infectivity, faithfully modeling its dynamics and cell type specificity. The reconstitution of an organized connectivity between human sensory neurons and keratinocytes into microfluidic chips provides a powerful in vitro platform for modeling viral latency and reactivation of human viral pathogens.

Keywords: herpes simplex virus; keratinocytes; latency; microfluidics; organoids; reactivation.

Figure 1

Coculture of DRGOs and keratinocytes on a chip. (A) Graphical representation of microfluidics chip and bright-field image (20× magnification) of axonal extensions from DRGO chamber to NHEK chamber through microgrooves. (B) Bright-field microscopy images of cells cocultured in microfluidic chips, with neuron-to-cell connection (white triangles). Scale bars: DRGO, 200 μm; HaCaT/NHEK/axons, 100 μm. (C) Immunofluorescence analysis of free nerve ending neuron-to-cell connection (arrowheads) using K14 (red), NF200 (white), synapsin (green, up), and vGlut1 (green, down). Total nuclear DNA is counterstained with Hoechst (blue). Scale bar, 50 μm.

Figure 2

Global gene expression profile of HSV receptors in DRGO. (A) Supervised gene expression heatmaps showing the HSV receptor genes differentially expressed either in DRG vs. iPSC or in DIV 40 DRGO vs. iPSC; the correlation between samples is also shown as an unsupervised hierarchical clustered dendrogram on the side. (B,C) Gene expression heatmaps showing the differentially expressed genes belonging to aggregated GO categories associated with heparan sulfate biosynthetic pathways (B) and integrin-mediated signaling pathway (C); the correlation between samples is also shown as an unsupervised hierarchical clustered dendrogram on the side.

Figure 3

Single-cell transcriptional profile of HSV receptors in DRGO. (A) Uniform Manifold Approximation and Projection (UMAP) plot displaying multidimensional reduction and clustering of single-cell RNA-seq data from DIV 80 DRGOs showing clusters of mature sensory neurons (proprioceptors C3, nociceptors C4, mechanoreceptors C8), satellite cells C9–C11, and Schwann cells C12 and C13 (modified from [20]). (B–I) UMAP plots highlighting normalized expression values of NECTIN1 (B), NRP1 (C), SDC1 (D), HS3ST3B1 (E), ITGB8 (F), ITGA6 (G), ITGA5 (H), and ITGA8 (I). (J) Heatmap showing normalized expression values of cell lineage-specific genes within the different clusters.

Figure 4

HSV-1 latency and reactivation protocol setup. (A) ACV phenotypic assay performed on HSV-1 laboratory strain (HF) and recombinant virus (KOS). (B) Schematic representation of different protocol timing tested on SH-SY5Y and DRGOs, using both virus strains. (C) Immunofluorescence analysis of SH-SY5Y and DRGOs during virus latency and reactivation. Virus protein gC (red, first column) is transcribed only during productive infection, while ICP0 promoter (green, second column) is active during both stages of infection. Total nuclear DNA is counterstained with Hoechst (blue); the last column shows the merge of green and red signals. Scale bars: SH-SY5Y, 30 μm; DRGO, 200 μm. (D) Levels of ICP0 (green fluorescent signal) and gC (red fluorescent signal) expression measured as integrated density values for SH-SY5Y and DRGOs during latency and reactivation. Mean ± SD is reported, * p < 0.05, *** p < 0.001, **** p < 0.0001. (E) Heatmap showing virus gene expression analysis of SH-SY5Y and DRGOs during latency and reactivation using both HSV-1 strains (dark red = high expression; white = no expression). Each condition was tested in quadruplicate.

Figure 5

Testing HSV-1 latency in the microfluidic culture system. Graphical representation for HSV-1 latency establishment directly in DRGOs following anterograde virus spread (EXP #1, (A,B)) or retrograde–anterograde virus spread (EXP #2, (C,D)). In EXP #1, organoids were infected (A) and latency was obtained through ACV addition to the culture medium. Thermal stress (B) led to controlled reactivation, and anterograde transport of virus particles resulted in NHEK lytic infection. EXP #2 instead was carried out by HSV-1 latency establishment indirectly in DRGOs (C). NHEK cells were infected, and retrograde transport of virus particles resulted in HSV-1 latency establishment in organoids. Then, thermal stress (D) led to controlled virus reactivation, and anterograde transport of virus particles resulted in NHEK lytic infection.

Figure 6

Anterograde virus spread experiment (EXP #1). Virus gene expression profiling for HSV-1 latency establishment (A) directly in DRGOs. Organoids were infected and latency was obtained through ACV addition to the culture medium. Thermal stress led to controlled reactivation (B), and anterograde transport of virus particles resulted in NHEK lytic infection. (C) Live imaging of DRGOs and NHEKs showing recombinant HSV-1 during latency (pICP0 active, green) and reactivation (pICP0 and pgC active, yellow). Scale bar, 100 μm. (D) Levels of ICP0 (green fluorescent signal) and gC (red fluorescent signal) expression measured as integrated density values for DRGOs and NHEKs, during latency and reactivation. Mean ± SD, * p < 0.05, **** p < 0.0001, n = 6 independent experiments (12 DRGOs).

Figure 7

Retrograde–anterograde virus spread experiment (EXP #2). Virus gene expression profiling for HSV-1 latency establishment (A) indirectly in DRGOs. NHEK cells were infected, and retrograde transport of virus particles resulted in HSV-1 latency establishment in organoids. Thermal stress led to controlled reactivation (B), and anterograde transport of virus particles resulted in NHEK lytic infection. (C) Live imaging of DRGOs and NHEKs showing recombinant HSV-1 during latency (pICP0 active, green) and reactivation (pICP0 and pgC active, yellow). Scale bar, 100 μm. (D) Levels of ICP0 (green fluorescent signal) and gC (red fluorescent signal) expression measured as integrated density values for DRGOs and NHEKs, during latency and reactivation. Mean ± SD, *** p < 0.001, **** p < 0.0001, n = 6 independent experiments (12 DRGOs).