Optimizing deep learning-based segmentation of densely packed cells using cell surface markers.

BACKGROUND: Spatial molecular profiling depends on accurate cell segmentation. Identification and quantitation of individual cells in dense tissues, e.g. highly inflamed tissue caused by viral infection or immune reaction, remains a challenge. METHODS: We first assess the performance of 18 deep learning-based cell segmentation models, either pre-trained or trained by us using two public image sets, on a set of immunofluorescence images stained with immune cell surface markers in skin tissue obtained during human herpes simplex virus (HSV) infection. We then further train eight of these models using up to 10,000+ training instances from the current image set. Finally, we seek to improve performance by tuning parameters of the most successful method from the previous step. RESULTS: The best model before fine-tuning achieves a mean Average Precision (mAP) of 0.516. Prediction performance improves substantially after training. The best model is the cyto model from Cellpose. After training, it achieves an mAP of 0.694; with further parameter tuning, the mAP reaches 0.711. CONCLUSION: Selecting the best model among the existing approaches and further training the model with images of interest produce the most gain in prediction performance. The performance of the resulting model compares favorably to human performance. The imperfection of the final model performance can be attributed to the moderate signal-to-noise ratio in the imageset.

Optimizing Deep Learning-Based Segmentation of Densely Packed Cells using Cell Surface Markers.

Background: Spatial molecular profiling depends on accurate cell segmentation. Identification and quantitation of individual cells in dense tissues, e.g. highly inflamed tissue caused by viral infection or immune reaction, remains a challenge. Methods: We first assess the performance of 18 deep learning-based cell segmentation models, either pre-trained or trained by us using two public image sets, on a set of immunofluorescence images stained with immune cell surface markers in skin tissue obtained during human herpes simplex virus (HSV) infection. We then further train eight of these models using up to 10,000+ training instances from the current image set. Finally, we seek to improve performance by tuning parameters of the most successful method from the previous step. Results: The best model before fine-tuning achieves a mean Average Precision (mAP) of 0.516. Prediction performance improves substantially after training. The best model is the cyto model from Cellpose. After training, it achieves an mAP of 0.694; with further parameter tuning, the mAP reaches 0.711. Conclusion: Selecting the best model among the existing approaches and further training the model with images of interest produce the most gain in prediction performance. The performance of the resulting model compares favorably to human performance. The imperfection of the final model performance can be attributed to the moderate signal-to-noise ratio i the imageset.

Tissue-Resident-Memory CD8+ T Cells Bridge Innate Immune Responses in Neighboring Epithelial Cells to Control Human Genital Herpes.

Tissue-resident-memory T cells (TRM) populate the body's barrier surfaces, functioning as frontline responders against reencountered pathogens. Understanding of the mechanisms by which CD8TRM achieve effective immune protection remains incomplete in a naturally recurring human disease. Using laser capture microdissection and transcriptional profiling, we investigate the impact of CD8TRM on the tissue microenvironment in skin biopsies sequentially obtained from a clinical cohort of diverse disease expression during herpes simplex virus 2 (HSV-2) reactivation. Epithelial cells neighboring CD8TRM display elevated and widespread innate and cell-intrinsic antiviral signature expression, largely related to IFNG expression. Detailed evaluation via T-cell receptor reconstruction confirms that CD8TRM recognize viral-infected cells at the specific HSV-2 peptide/HLA level. The hierarchical pattern of core IFN-γ signature expression is well-conserved in normal human skin across various anatomic sites, while elevation of IFI16, TRIM 22, IFITM2, IFITM3, MX1, MX2, STAT1, IRF7, ISG15, IFI44, CXCL10 and CCL5 expression is associated with HSV-2-affected asymptomatic tissue. In primary human cells, IFN-γ pretreatment reduces gene transcription at the immediate-early stage of virus lifecycle, enhances IFI16 restriction of wild-type HSV-2 replication and renders favorable kinetics for host protection. Thus, the adaptive immune response through antigen-specific recognition instructs innate and cell-intrinsic antiviral machinery to control herpes reactivation, a reversal of the canonical thinking of innate activating adaptive immunity in primary infection. Communication from CD8TRM to surrounding epithelial cells to activate broad innate resistance might be critical in restraining various viral diseases.

Distinct populations of antigen-specific tissue-resident CD8+ T cells in human cervix mucosa.

The ectocervix is part of the lower female reproductive tract (FRT), which is susceptible to sexually transmitted infections (STIs). Comprehensive knowledge of the phenotypes and T cell receptor (TCR) repertoire of tissue-resident memory T cells (TRMs) in the human FRT is lacking. We took single-cell RNA-Seq approaches to simultaneously define gene expression and TCR clonotypes of the human ectocervix. There were significantly more CD8+ than CD4+ T cells. Unsupervised clustering and trajectory analysis identified distinct populations of CD8+ T cells with IFNGhiGZMBloCD69hiCD103lo or IFNGloGZMBhiCD69medCD103hi phenotypes. Little overlap was seen between their TCR repertoires. Immunofluorescence staining showed that CD103+CD8+ TRMs were preferentially localized in the epithelium, whereas CD69+CD8+ TRMs were distributed evenly in the epithelium and stroma. Ex vivo assays indicated that up to 14% of cervical CD8+ TRM clonotypes were HSV-2 reactive in HSV-2-seropositive persons, reflecting physiologically relevant localization. Our studies identified subgroups of CD8+ TRMs in the human ectocervix that exhibited distinct expression of antiviral defense and tissue residency markers, anatomic locations, and TCR repertoires that target anatomically relevant viral antigens. Optimization of the location, number, and function of FRT TRMs is an important approach for improving host defenses to STIs.

B cells, antibody-secreting cells, and virus-specific antibodies respond to herpes simplex virus 2 reactivation in skin.

Tissue-based T cells are important effectors in the prevention and control of mucosal viral infections; less is known about tissue-based B cells. We demonstrate that B cells and antibody-secreting cells (ASCs) are present in inflammatory infiltrates in skin biopsy specimens from study participants during symptomatic herpes simplex virus 2 (HSV-2) reactivation and early healing. Both CD20+ B cells, most of which are antigen inexperienced based on their coexpression of IgD, and ASCs - characterized by dense IgG RNA expression in combination with CD138, IRF4, and Blimp-1 RNA - were found to colocalize with T cells. ASCs clustered with CD4+ T cells, suggesting the potential for crosstalk. HSV-2-specific antibodies to virus surface antigens were also present in tissue and increased in concentration during HSV-2 reactivation and healing, unlike in serum, where concentrations remained static over time. B cells, ASCs, and HSV-specific antibody were rarely detected in biopsies of unaffected skin. Evaluation of samples from serial biopsies demonstrated that B cells and ASCs followed a more migratory than resident pattern of infiltration in HSV-affected genital skin, in contrast to T cells. Together, these observations suggest the presence of distinct phenotypes of B cells in HSV-affected tissue; dissecting their role in reactivation may reveal new therapeutic avenues to control these infections.

High Community SARS-CoV-2 Antibody Seroprevalence in a Ski Resort Community, Blaine County, Idaho, US. Preliminary Results.

Community-level seroprevalence surveys are needed to determine the proportion of the population with previous SARS-CoV-2 infection, a necessary component of COVID-19 disease surveillance. In May, 2020, we conducted a cross-sectional seroprevalence study of IgG antibodies for nucleocapsid of SARS-CoV-2 among the residents of Blaine County, Idaho, a ski resort community with high COVID-19 attack rates in late March and Early April (2.9% for ages 18 and older). Participants were selected from volunteers who registered via a secure web link, using prestratification weighting to the population distribution by age and gender within each ZIP Code. Participants completed a survey reporting their demographics and symptoms; 88% of volunteers who were invited to participate completed data collection survey and had 10 ml of blood drawn. Serology was completed via the Abbott Architect SARS-CoV-2 IgG immunoassay. Primary analyses estimated seroprevalence and 95% credible intervals (CI) using a hierarchical Bayesian framework to account for diagnostic uncertainty. Stratified models were run by age, sex, ZIP Code, ethnicity, employment status, and a priori participant-reported COVID-19 status. Sensitivity analyses to estimate seroprevalence included base models with post-stratification for ethnicity, age, and sex, with or without adjustment for multi-participant households. IgG antibodies to the virus that causes COVID-19 were found among 22.7% (95% CI: 20.1%, 25.5%) of residents of Blaine County. Higher levels of antibodies were found among residents of the City of Ketchum 34.8% (95% CI 29.3%, 40.5%), compared to Hailey 16.8% (95%CI 13.7%, 20.3%) and Sun Valley 19.4% (95% 11.8%, 28.4%). People who self-identified as not believing they had COVID-19 had the lowest prevalence 4.8% (95% CI 2.3%, 8.2%). The range of seroprevalence after correction for potential selection bias was 21.9% to 24.2%. This study suggests more than 80% of SARS-CoV-2 infections were not reported. Although Blaine County had high levels of SARS-CoV-2 infection, the community is not yet near the herd immunity threshold.

Keratinocytes produce IL-17c to protect peripheral nervous systems during human HSV-2 reactivation.

Despite frequent herpes simplex virus (HSV) reactivation, peripheral nerve destruction and sensory anesthesia are rare. We discovered that skin biopsies obtained during asymptomatic human HSV-2 reactivation exhibit a higher density of nerve fibers relative to biopsies during virological and clinical quiescence. We evaluated the effects of HSV infection on keratinocytes, the initial target of HSV replication, to better understand this observation. Keratinocytes produced IL-17c during HSV-2 reactivation, and IL-17RE, an IL-17c-specific receptor, was expressed on nerve fibers in human skin and sensory neurons in dorsal root ganglia. In ex vivo experiments, exogenous human IL-17c provided directional guidance and promoted neurite growth and branching in microfluidic devices. Exogenous murine IL-17c pretreatment reduced apoptosis in HSV-2-infected primary neurons. These results suggest that IL-17c is a neurotrophic cytokine that protects peripheral nerve systems during HSV reactivation. This mechanism could explain the lack of nerve damage from recurrent HSV infection and may provide insight to understanding and treating sensory peripheral neuropathies.

Immune surveillance by CD8αα+ skin-resident T cells in human herpes virus infection.

Most herpes simplex virus 2 (HSV-2) reactivations in humans are subclinical and associated with rapid expansion and containment of virus. Previous studies have shown that CD8(+) T cells persist in genital skin and mucosa at the dermal-epidermal junction (DEJ)--the portal of neuronal release of reactivating virus--for prolonged time periods after herpes lesions are cleared. The phenotype and function of this persistent CD8(+) T-cell population remain unknown. Here, using cell-type-specific laser capture microdissection, transcriptional profiling and T-cell antigen receptor ß-chain (TCRß) genotyping on sequential genital skin biopsies, we show that CD8αα(+) T cells are the dominant resident population of DEJ CD8(+) T cells that persist at the site of previous HSV-2 reactivation. CD8αα(+) T cells located at the DEJ lack chemokine-receptor expression required for lymphocyte egress and recirculation, express gene signatures of T-cell activation and antiviral activity, and produce cytolytic granules during clinical and virological quiescent time periods. Sequencing of the TCR ß-chain repertoire reveals that the DEJ CD8αα(+) T cells are oligoclonal with diverse usage of TCR variable-ß genes, which differ from those commonly described for mucosa-associated invariant T cells and natural killer T cells. Dominant clonotypes are shown to overlap among multiple recurrences over a period of two-and-a-half years. Episodes of rapid asymptomatic HSV-2 containment were also associated with a high CD8 effector-to-target ratio and focal enrichment of CD8αα(+) T cells. These studies indicate that DEJ CD8αα(+) T cells are tissue-resident cells that seem to have a fundamental role in immune surveillance and in initial containment of HSV-2 reactivation in human peripheral tissue. Elicitation of CD8αα(+) T cells may be a critical component for developing effective vaccines against skin and mucosal infections.

An effector phenotype of CD8+ T cells at the junction epithelium during clinical quiescence of herpes simplex virus 2 infection.

Herpes simplex virus 2 infection is characterized by cycles of viral quiescence and reactivation. CD8(+) T cells persist at the site of viral reactivation, at the genital dermal-epidermal junction contiguous to neuronal endings of sensory neurons, for several months after herpes lesion resolution. To evaluate whether these resident CD8(+) T cells frequently encounter HSV antigen even during times of asymptomatic viral infection, we analyzed the transcriptional output of CD8(+) T cells captured by laser microdissection from human genital skin biopsy specimens during the clinically quiescent period of 8 weeks after lesion resolution. These CD8(+) T cells expressed a characteristic set of genes distinct from those of three separate control cell populations, and network and pathway analyses revealed that these T cells significantly upregulated antiviral genes such as GZMB, PRF1, INFG, IL-32, and LTA, carbohydrate and lipid metabolism-related genes such as GLUT-1, and chemotaxis and recruitment genes such as CCL5 and CCR1, suggesting a possible feedback mechanism for the recruitment of CD8(+) T cells to the site of infection. Many of these transcripts are known to have half-lives of <48 h, suggesting that cognate antigen is released frequently into the mucosa and that resident CD8(+) T cells act as functional effectors in controlling viral spread.

Evasion of the mucosal innate immune system by herpes simplex virus type 2.

Understanding the mechanisms by which herpes simplex virus (HSV) evades host immune defenses is critical to defining new approaches for therapy and prevention. We performed transcriptional analyses and immunocytochemistry on sequential biopsy specimens of lesional tissue from the acute through the posthealing phases of recurrent mucocutaneous HSV-2 infection. Histological analysis of these biopsy specimens during the acute stage revealed a massive infiltration of T cells, as well as monocytes/macrophages, a large amount of myeloid, and a small number of plasmacytoid dendritic cells, in the dermis of these lesional biopsy specimens. Type I interferon (IFN-beta and IFN-alpha) was poorly expressed and gamma IFN (IFN-gamma) potently induced during time periods in which we detected abundant amounts of HSV-2 antigens and HSV-2 RNA. IFN-stimulated genes were also markedly upregulated, with expression patterns that more closely matched those in primary human fibroblasts treated by IFN-gamma than those in fibroblasts treated by IFN-beta. Transcriptional arrays of the same lesional biopsy sites during healing and at 2 and 4 weeks posthealing revealed no HSV nucleic acids or antigen; however, there was persistent expression of IFN-gamma, with very low levels of IFN-beta and IFN-alpha. The findings of extremely low levels of IFN-alpha and IFN-beta, despite the presence of a large number of cells capable of synthesizing these substances, suggest a potent alteration in host defense during HSV-2 infection in vivo. HSV-2's blockade of the innate immune system's production of type I IFN may be a major factor in allowing the virus to break through host mucosal defenses.