Introduction: The brain is considered as an immune-privileged organ, yet innate immune reactions can occur in the central nervous system of vertebrates and invertebrates. Silkworm (Bombyx mori) is an economically important insect and a lepidopteran model species. The diversity of cell types in the silkworm brain, and how these cell subsets produce an immune response to virus infection, remains largely unknown. Methods: Single-nucleus RNA sequencing (snRNA-seq), bioinformatics analysis, RNAi, and other methods were mainly used to analyze the cell types and gene functions of the silkworm brain. Results: We used snRNA-seq to identify 19 distinct clusters representing Kenyon cell, glial cell, olfactory projection neuron, optic lobes neuron, hemocyte-like cell, and muscle cell types in the B. mori nucleopolyhedrovirus (BmNPV)-infected and BmNPV-uninfected silkworm larvae brain at the late stage of infection. Further, we found that the cell subset that exerts an antiviral function in the silkworm larvae brain corresponds to hemocytes. Specifically, antimicrobial peptides were significantly induced by BmNPV infection in the hemocytes, especially lysozyme, exerting antiviral effects. Conclusion: Our single-cell dataset reveals the diversity of silkworm larvae brain cells, and the transcriptome analysis provides insights into the immune response following virus infection at the single-cell level.

The insect central nervous system (CNS) comprises the cerebral ganglia, the subesophageal ganglion, and the segmental ganglia that form the ventral nerve cord [1, 2]. The cerebral ganglia and the subesophageal ganglion make up the insect brain, both of which are located in the head [1]. The insect brain, as a main neuroendocrine organ, plays crucial roles in the regulation of development and growth and the central control of all types of behavior [3, 4]. The insect brain is often used as an important model system for a comprehensive cellular, molecular, and genetic investigation of neuronal development and function due to the low number of nerve cells and other advantages. However, understanding the molecular, genetic, and cellular mechanisms that underlie brain organization and function still remains one of the most challenging problems of neurobiology. Additionally, the brain was long considered to be an immune-privileged organ [5]. Despite reports that innate immune pathways that involve Toll or Imd signaling, autophagy, phagocytosis, and RNAi could be activated in the Drosophila brain following injury, pathogen infection, and modeling of neurodegenerative disease, our understanding of the protective antimicrobial mechanisms in the Drosophila brain, let alone other insects, remains limited and requires more research effort [6].

As the model species of Lepidoptera and the only truly domesticated insect, the domesticated silkworm has been used for basic and applied research for a long time. However, the Bombyx mori nucleopolyhedrovirus (BmNPV), belonging to the Alphabaculovirus genus of the large DNA virus family Baculoviridae, is one of the main infectious agents to cause disease and lethality in the silkworm [7]. Many omics approaches (bulk-seq) have been applied to explore the interaction between BmNPV infection and the silkworm brain [4, 8, 9], but these mainly concerned the hyperactive behavior rather than the immune response induced by baculovirus infection. In fact, the cell types that constitute the brain of the silkworm and their immune response pathways are still unclear.

With the development of single-cell RNA sequencing (scRNA-seq) and single-nucleus RNA sequencing (snRNA-seq) technology, the cellular landscape of the brain has already been reported in different vertebrate and invertebrate species such as mouse [10, 11], zebrafish [12], grouper fish [13], Drosophila [14‒16], and the mosquito Aedes aegypti [17]. Further, the host immune response induced by virus infection has been investigated from the perspective of the single cell and cell subgroup in particular tissues [18, 19].

Here, we report the application and an initial analysis of snRNA-seq data to investigate the cellular diversity of the silkworm larval brain. We also analyzed the effect of BmNPV infection in the brain at the level of single-cell transcriptomics. Our single-cell dataset provides an indication of the extent of cellular diversity in the silkworm brain and presents a transcriptome map that illustrates the impact of BmNPV infection on the composition and function of brain cell clusters. Because BmNPV infection is known to alter brain function and induce hyperactive behavior [20, 21], many of the changes observed in the majority of the clusters may be a consequence of the reprogramming of the brain by the virus which will be discussed in more detail in a future submission (manuscript in preparation). Here, we focus on the immune response to BmNPV infection, which is mainly carried out by a minor cell type that corresponds to hemocyte-like cells. Our results reveal that hemocyte-like cells in the silkworm brain respond to BmNPV infection by the activation of genes that encode antimicrobial peptides (AMPs), indicating their crucial role in the defense against baculovirus infection.

Silkworm and Virus Infection

Larvae of silkworm (Dazao strain) were reared with fresh mulberry leaves and reared at a temperature of 28°C and relative humidity between 60 and 70%. Recombinant BmNPV-eGFP (Enhanced Green Fluorescent Protein), as a reporter virus, was constructed by the BmNPV-based Bac-to-Bac System (B. mori MultiBac) [22]. Newly molted fifth-instar silkworm larvae were injected with either 10 μL of BmNPV-eGFP (budded virus (BV), 1 × 106.2 TCID50/mL) or phosphate-buffered saline (PBS) (Negative Control, NC, USA).

Single Nucleus Preparation of Silkworm Brain in Suspension

At 96 h postinfection (hpi) of BmNPV-eGFP, brain tissue was collected from pools of 20 BmNPV-infected silkworms or 20 controls. Each of the two experimental groups was collected as a pooled sample that was respectively named B96h-BV and B96h-NC. The brain tissue was cleaned with DEPC water and stored immediately in liquid nitrogen until use. Silkworm brain tissue was transferred from the cryopreserved tube to a Dounce homogenizer, and 500 μL of precooled lysis buffer was added. The nuclei extraction was completed by Gene Denovo company (Guangzhou, China) and performed as previously described [23, 24]. Briefly, after homogenizing brain tissue, the homogenate was filtered through a 70-μm cell strainer and 50% iodixanol was added to gently mix the sample obtained by filtration. Then, a discontinuous iodixanol gradient solution (33–30%) was prepared prior to centrifugation and the filtered product was added. After centrifugation at 4°C, 10,000 g for 20 min, the white nuclear layer located at the 33% and 30% iodixanol interface was obtained and washed with nuclear cleaning buffer. After filtering through a 40-μm cell strainer, the nuclear suspension stained with trypan blue was assessed using a cell counting plate under a microscope, and the total amount, concentration, and ratio of nuclei with an intact nuclear membrane were calculated. The final target concentration of the nuclear suspension was 700–1,200 nuclei/μL.

Single-Nucleus Encapsulation and Sequencing

Single-nucleus encapsulation, complementary DNA library synthesis, RNA-sequencing, and raw data analysis were completed by Gene Denovo (Guangzhou, China). Nucleus suspensions were bar-coded and reverse-transcribed into snRNA-seq libraries using the Chromium Single Cell 3′ Gel Bead-in Emulsion (GEM), Library and Gel Bead Kit (10x Genomics, USA) according to the manufacturer’s protocol. Briefly, single-nucleus suspensions of silkworm brain were barcode-labeled and mixed with reverse transcriptase into GEMs. The indexed sequencing libraries were prepared using Chromium Single Cell 3′ Reagent Kits (v3) according to the manufacturer’s instructions. The final single cell 3′ library contained the standard Illumina paired-end constructs that begin and end with P5 and P7 primers used in the Illumina bridge amplification. The barcoded sequencing libraries were quantified using a standard curve-based qPCR assay with KAPA Library Quantification Kit (KAPA Biosystems, USA) and Agilent Bioanalyzer 2100 with a High Sensitivity DNA chip (Agilent, USA). Finally, library sequencing was performed by Illumina HiSeq 4000 with a custom paired-end sequencing mode of 26 base pairs for read 1 and 98 base pairs for read 2.

snRNA-Seq Data Processing and Analysis

The Cell Ranger Single Cell Software Suite (v6.1) was applied for quality control, sample demultiplexing, barcode processing, and single-cell 3′ gene counting. Raw data were first demultiplexed into FASTQ data by bcl2fastq software. FASTQ file quality control was performed using FastQC, and these data were aligned against the Nucleotide Sequence Database (https://www.ncbi.nlm.nih.gov/genbank/) using the NCBI Basic Local Alignment Search Tool (BLAST). After the initial quality control, low-quality sequences together with barcodes and unique molecular identifiers (UMIs) were removed.

In the raw data analysis, the fastq data were aligned to the latest published version of the silkworm genome (SilkDB3.0) using STAR with default parameters. For further counting of the UMI tags, the CellRanger count algorithm was used to generate single-cell gene counts for a single library to get the most stable and accurate clustering solutions for 10x Genomics scRNA-seq data [25]. Only confidently mapped, non-PCR duplicates with valid barcodes and UMIs were used to generate the gene-barcode matrix. For combining data from different libraries, a gene-cell-barcode matrix from each library was normalized by equalizing the read depth between libraries for merging using CellRanger to reduce the batch effect introduced by sequencing. For the higher-depth libraries, the samples were normalized to the sample sequencing depth. CellRanger version 2.0.0 and Seurat (v4.0.4) R package were used to filter out the low-quality cells. The following criteria were used to filter cells: (1) gene counts <400 or gene count>3,000 per cell and (2) UMI counts >20,000 per cell.

Canonical correlation analysis was used to normalize and filter the gene-barcode matrix to reduce feature dimensions. The normalized data were clustered using principal component analysis and visualized with t-distributed Stochastic Neighbor Embedding or Uniform Manifold Approximation and Projection (UMAP) in a two-dimensional space. The cell subset was grouped by graph-based clustering based on the gene expression profile of each cell in Seurat. According to the expression pattern of hallmark genes, cell clusters in silkworm brain were grouped into unsupervised categories.

To identify genes that were enriched in a specific cluster, the mean expression of each gene was calculated across all cells in the cluster. Each gene from the cluster then was compared to the median expression of the same gene from cells in all other clusters. Gene expression level in this study was calculated as log (1 + [UMI A ÷ UMI total] × 10,000). It should be noted that when performing gene quantification, “UMI Total” does not include “UMI BmNPV” (thus excluding viral genes).

Differentially Expressed Genes Analysis per Cluster

Differentially expressed genes (DEGs) analysis for each cluster was performed as in our previous study [18, 24]. Likelihood-ratio test was performed on single cluster cells against all other cells to identify DEGs in single clusters based on differential expression. Upregulated DEGs in each cluster were identified by the following criteria: (1) p value ≤0.01; (2) log2FC ≥0.36 (log2FC means log fold change of the average expression between the two groups); and (3) percentage of cells in a specific cluster where the gene is detected >25%. The top 5 genes in each unsupervised cluster were selected as the potential marker genes according to the result of DEGs analysis. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were further carried out based on gene expression levels to identify the main features of each cluster.

DEGs Analysis in Cell Clusters between BmNPV-Infected and BmNPV-Uninfected Brain

To explore the response of each cell cluster in brain to BmNPV infection, we further analyzed the DEGs between the BmNPV-infected group and the control group using Seurat software under the condition of a minimum of 1 cell per cluster. DEGs between the BmNPV-infected and control groups were identified by the following criteria: (1) |log2FC| ≥0.36; (2) p value_adj ≤0.05; and (3) percentage of cells where the gene is detected in a specific cluster >10%. Identified DEGs were subsequently subjected to GO and KEGG pathway enrichment analysis.

Knockdown of Lysozyme in Silkworm Larvae

Double-strand RNA (dsRNA) specifically targeting the lysozyme (BMSK0006899) mRNA was designed and used in knockdown experiments. The primers for dsRNA synthesis are listed in Table 1. T7 RiboMAXTM Express RNAi System (Promega, USA) was used to synthesize dsRNA, and the products were stored at −80°C until use. DsRNA-DsRed2 was used as control. Silkworm larvae (5th instar) were injected into the tail proleg with 5 μL of dsRNA (6 μg). The knockdown effect was detected by qPCR using brain samples from three animals selected randomly from all treated silkworms at 48 h postinjection of dsRNA. qPCR was performed on the Bio-Rad CFX96 Real-Time Detection System using iTaqTM Universal SYBR® Green Supermix Kit reagents (Bio-Rad, USA). The silkworm rp49 gene was used as an internal control. All primer sequences are listed in Table 1. Data analyses were performed using the 2−ΔΔCt method.

Table 1.

Primers for dsRNA synthesis and qPCR detection

Gene nameGene IDSequence of primers (5′-3′)
dsRNA-BmLysozyme NM_001043983.1/BMSK0006899 F: TAA​TAC​GAC​TCA​CTA​TAG​GGC​AAG​ACG​AAC​ACG​AAC​CGT​A 
R: TAA​TAC​GAC​TCA​CTA​TAG​GGC​TGG​CAG​TGG​TTC​TTC​CAA​C 
qPCR-BmLysozyme NM_001043983.1/BMSK0006899 F: TGT​CCT​CTG​CGT​TGG​TTC​TG 
R: TCG​AAG​CCA​TGC​TTC​CTC​AG 
qPCR-Bmrp49 BMSK0014093 F: CAG​GCG​GTT​CAA​GGG​TCA​ATA​C 
R: TGC​TGG​GCT​CTT​TCC​ACG​A 
qPCR-BmTIF4A BMSK0002090 F: GAA​TGG​ACC​CTG​GGA​CAC​TT 
R: CTG​ACT​GGG​CTT​GAG​CGA​TA 
qPCR-vp39 JQ991008.1 F: CTA​ATG​CCC​GTG​GGT​ATG​G 
R: TTG​ATG​AGG​TGG​CTG​TTG​C 
Gene nameGene IDSequence of primers (5′-3′)
dsRNA-BmLysozyme NM_001043983.1/BMSK0006899 F: TAA​TAC​GAC​TCA​CTA​TAG​GGC​AAG​ACG​AAC​ACG​AAC​CGT​A 
R: TAA​TAC​GAC​TCA​CTA​TAG​GGC​TGG​CAG​TGG​TTC​TTC​CAA​C 
qPCR-BmLysozyme NM_001043983.1/BMSK0006899 F: TGT​CCT​CTG​CGT​TGG​TTC​TG 
R: TCG​AAG​CCA​TGC​TTC​CTC​AG 
qPCR-Bmrp49 BMSK0014093 F: CAG​GCG​GTT​CAA​GGG​TCA​ATA​C 
R: TGC​TGG​GCT​CTT​TCC​ACG​A 
qPCR-BmTIF4A BMSK0002090 F: GAA​TGG​ACC​CTG​GGA​CAC​TT 
R: CTG​ACT​GGG​CTT​GAG​CGA​TA 
qPCR-vp39 JQ991008.1 F: CTA​ATG​CCC​GTG​GGT​ATG​G 
R: TTG​ATG​AGG​TGG​CTG​TTG​C 

The underlined sequence indicates the T7 promoter.

Subsequently, silkworm larvae (5th instar) were injected into the tail proleg with a mixture of 5 μL of dsRNA (6 μg) and 5 μL of BmNPV-eGFP (1.0 × 106.2 TCID50/mL). Brain samples were collected from silkworm larvae treated with either dsRNA-lysozyme or dsRNA-DsRed2 at 96 hpi for the analysis of the progression of BmNPV infection. The viral gene vp39 was used to detect viral mRNA abundance according to the method described in our previous study [26]. The silkworm TIF4A gene was used as an internal control for samples of tissues infected by BmNPV. Each group contained at least three silkworm larvae, and at least two independent animal experiments were performed. Statistical analyses were performed using Prism 8 software (GraphPad).

Single-Nucleus Transcriptomics Identifies 19 Distinct Cell Clusters in the Brain Collected from BmNPV-Infected and Control Silkworm Larvae

10x Genomics platforms were used to perform 3′ snRNA-seq on pooled cell nuclei collected from brain of BmNPV-infected silkworm larvae and PBS-treated controls at 96 hpi (Fig. 1a). After filtering out low-quality cells, a total 19,933 brain cells from BmNPV-infected and BmNPV-uninfected silkworms were profiled together and 19 distinct unsupervised clusters were obtained that were visualized using t-distributed Stochastic Neighbor Embedding (Fig. 1b, c). Among these silkworm brain cells, 10,706 and 9,227 cells were obtained from the control group (B96h-NC) and the BmNPV-infected group (B96h-BV), respectively (Fig. 1c). From cluster 0 (9,802 cells, 49.17%) to cluster 18 (20 cells, 0.10%), the number of cells gradually decreases in the unidentified clusters (Fig. 1c). However, the cells contained in cluster 0 (9,802, 49.17%) and cluster 1 (3,411, 17.11%) account for more than 60% of all brain cells (Fig. 1c). Furthermore, most of the brain cell clusters in silkworm were detected with a large number of upregulated DEGs, especially cluster 11 (1,087 DEGs) and 4 (819 DEGs) (Fig. 1d). The expression levels and the percentage of cells expressing the top five genes in each cluster of silkworm brain are shown in a dot plot (Fig. 1e). These top DEGs in each cluster need to be confirmed in future research and are proposed to be useful as new marker genes for the identification of specific cell types within the silkworm larvae brain.

Fig. 1.

Single-nucleus profiling of unsupervised cell subsets in the brain of silkworm larvae. a Schematic illustration of the experimental workflow. Twenty larvae for each group were injected with BmNPV-eGFP or PBS and sacrificed at 96 h posttreatment. Brains of the twenty larvae in each group were pooled as one sample for snRNA-seq analysis. In each single nucleus, mRNAs of host and virus were simultaneously measured, allowing comparison of the transcriptome profile in each cell subset under the infection or control condition. b t-distributed Stochastic Neighbor Embedding (t-SNE) projection representing the 19 unsupervised cell clusters identified in the brain of fifth-instar silkworm larvae (unified set of BmNPV infection and control samples). c Number of cells in each cluster and their proportional distribution in the total brain dataset. d Number of upregulated DEGs in each brain cluster. e Top 5 DEGs (x-axis) identified in each brain cluster (y-axis). Dot size represents the fraction of cells in the cluster that express the gene; intensity indicates the mean expression (Z-score) in expressing cells, relative to other clusters.

Fig. 1.

Single-nucleus profiling of unsupervised cell subsets in the brain of silkworm larvae. a Schematic illustration of the experimental workflow. Twenty larvae for each group were injected with BmNPV-eGFP or PBS and sacrificed at 96 h posttreatment. Brains of the twenty larvae in each group were pooled as one sample for snRNA-seq analysis. In each single nucleus, mRNAs of host and virus were simultaneously measured, allowing comparison of the transcriptome profile in each cell subset under the infection or control condition. b t-distributed Stochastic Neighbor Embedding (t-SNE) projection representing the 19 unsupervised cell clusters identified in the brain of fifth-instar silkworm larvae (unified set of BmNPV infection and control samples). c Number of cells in each cluster and their proportional distribution in the total brain dataset. d Number of upregulated DEGs in each brain cluster. e Top 5 DEGs (x-axis) identified in each brain cluster (y-axis). Dot size represents the fraction of cells in the cluster that express the gene; intensity indicates the mean expression (Z-score) in expressing cells, relative to other clusters.

Close modal

Identification of Specific Cell Types in the Brain of Silkworm

Based on the expression of the pan-neuronal marker synaptosomal-associated protein 25 (Snap25, BMSK0014784) and the glial cell pan-marker retinal homeobox protein Rx3 (ARX, BMSK0008644, the homolog of the Drosophila reverse polarity (repo) gene) (Fig. 2a), we preliminary divided the 19 unsupervised clusters into neuronal clusters (Snap25-high), glial clusters (ARX-high), and other type (Snap25/ARX low) (Fig. 2a, b). From the violin diagram of Snap25 and ARX, clusters 1, 4, 5, 6, 8, 9, 10, 11, 13, 14, 15, 16, and 18 were initially identified as neuronal clusters, and clusters 3 and 7 were initially identified as glial clusters (Fig. 2a). It should be noted that Snap25 was also expressed in some cells in clusters 3 and 7, while ARX was lowly expressed in some cells in clusters 0, 2, and 12 (Fig. 2a, b). While Snap25 is reported to be expressed in neurons to regulate exocytosis of synaptic vesicles [27], its other functions include the control of intracellular trafficking related to lysosome function and autophagy [28]. Expression of Snap25 at low levels in some cells may reflect the latter function. Bombyx BMSK0008644, on the other hand, is a gene that belongs of the Otx/Arx/Rax subgroup of paired-like homeodomain of transcription factors of which the Drosophila homolog Repo is generally used as a glial cell marker [29]. Repo is induced during glial cell differentiation by the nuclear protein Glial Cells Missing (GCM) that acts as a binary switch between neuronal and glial cell fates [30]. Cells with low expression of ARX may indicate a lower level of differentiation toward the glial cell fate.

Fig. 2.

Identification of specific cell types in the brain of silkworm larvae. a Violin plot showing the expression of two pan marker genes across the 19 cell clusters. Each cluster is color coded. Snap25 (BMSK0014784), pan-neuronal marker; ARX (BMSK0008644, the homologous gene of Drosophila repo gene), and glial cell pan-marker. b UMAP plots showing distribution of pan-neuron and pan-glial cell markers in distinct cell clusters. c Expression levels and the percentage of cells expressing markers in brain clusters 0–18 are shown as a dot plot. d UMAP plot displaying the KCs, glia cells, olfactory PNs, optic lobes neurons, hemocyte-like cells, muscle cell-like cells, and unannotated cells.

Fig. 2.

Identification of specific cell types in the brain of silkworm larvae. a Violin plot showing the expression of two pan marker genes across the 19 cell clusters. Each cluster is color coded. Snap25 (BMSK0014784), pan-neuronal marker; ARX (BMSK0008644, the homologous gene of Drosophila repo gene), and glial cell pan-marker. b UMAP plots showing distribution of pan-neuron and pan-glial cell markers in distinct cell clusters. c Expression levels and the percentage of cells expressing markers in brain clusters 0–18 are shown as a dot plot. d UMAP plot displaying the KCs, glia cells, olfactory PNs, optic lobes neurons, hemocyte-like cells, muscle cell-like cells, and unannotated cells.

Close modal

Since there are no recognized marker genes for various cell types in the silkworm brain, we then used the homologous genes for published cell-type-specific markers in the Drosophila brain to provisionally assign identities to these silkworm brain clusters [14, 15]. We identified clusters 4, 9, and 18 as Kenyon cells (KCs) based on the expression of Mef2 (BMSK0002529), Cubn (BMSK0003554), dlg1 (BMSK0007018), PCBP3 (BMSK0004996), rut (BMSK0005480), and PAX6 (BMSK0008051) (Fig. 2c, d). As already mentioned, clusters 3 and 7 were assigned as glia cells because of the significant upregulation of ARX (Fig. 2c, d). Based on the presence of DEGs LIM/Lhx1 (BMSK0009315), sox6 (BMSK0008039), sox1a (BMSK0001801), TTF-1 (BMSK0014528), and ct (BMSK0009668), we further assigned clusters 16, 14, 15, 11, and 6 as olfactory projection neurons (PNs) and cluster 5 as olfactory PN-like (Fig. 2c, d). We identified cluster 10 as optic lobe neurons based on DRGX (BMSK0006373) and LIM/LHX3 (BMSK0002038) expression (Fig. 2c, d). PGRP (BMSK0004739) and paralytic peptide (BMSK0007609) are typical markers of silkworm hemocytes [18] that were also here used to designate cluster 12 as hemocyte-like cells (Fig. 2c, d). Finally, cluster 17 was identified as muscle cell-like because of high expression of sls/kettin (BMSK0000201), Mlc1 (BMSK0003384), and Mhc (BMSK0004205) (Fig. 2c, d). Unfortunately, there is insufficient evidence to designate clusters 0, 1, 2, 8, and 13 as identifiable cell types (Fig. 2c, d).

The Cell Heterogeneity between Infected and Uninfected Brain of Silkworm Larvae

We further compared the distribution of each silkworm larvae brain cluster in the BmNPV-infected and the control groups. The UMAP plots of Figure 3a, b display brain cells of PBS-treated and BmNPV-infected silkworm larvae at 96 hpi. All previously described cell types from cluster 0 to cluster 18 could be detected in the uninfected silkworm larvae brain (Fig. 3a) and BmNPV-infected brain (Fig. 3b). The cell subsets in the BmNPV-infected brain and the BmNPV-uninfected brain were consistent, and the difference was mainly due to the different number of cells in each cell subset (Fig. 3a, b). However, comparing the cellular landscape, we observe that cells of the BmNPV-infected and control groups are clustered into two distinct parts in the UMAP plot, indicating that the BmNPV infection caused a larger heterogeneity in the silkworm brain cell subsets (Fig. 3c). Quantitative analysis of the heterogeneity of individual cell subsets in the BmNPV-infected and BmNPV-uninfected groups is presented in Figure 3d. Only for cluster 0 and cluster 17 (muscle cell-like) is the proportion of cells in the normal brain higher than that in the BmNPV-infected brain. In all other cell subsets, the cellular proportion is increased after infection with the virus (Fig. 3d). Notably, cluster 0 is the largest cell cluster (83.63%) in the normal silkworm brain, but in the virus-infected brain, the cellular proportion of this cluster (9.2%) is decreased by ninefold at the late stage of infection (96 hpi) (Fig. 3d). From this analysis, it appears that the cellular distribution across subpopulations in the virus-infected brain is more uniform than in the normal brain. Compared with the normal brain, the cellular proportion was increased by a large margin in KC1 (cluster 4; 13.7 fold), KC2 (cluster 9; 14.7 fold), Glia 1 (cluster 3; 15.5 fold), Glia 2 (cluster 7; 87.8 fold), PN4 (cluster 11; 24.8 fold), and cluster 2 (60.9 fold) (Fig. 3d). More moderate increases were observed in PN5 (cluster 6; 3.7 fold), PN3 (cluster 15; 3.3 fold), PN2 (cluster 14, 3.4 fold), PN1 (cluster 16; 2.5 fold), PN-like (cluster 5; 6.1 fold), optic lobe neuron (cluster 10; 3.0 fold), hemocyte-like (cluster 12; 2.0 fold), cluster 1 (3.1 fold), cluster 8 (3.7 fold), and cluster 13 (3.0 fold) (Fig. 3d).

Fig. 3.

Comparison of brain cell heterogeneity between BmNPV-infected and BmNPV-uninfected control larvae. a, b UMAP displaying the brain cell clusters in the PBS-treated group (a) and the BmNPV-infected group (b). c UMAP displaying the difference in the distribution of cells in the brain clusters between the BmNPV-infected and control larvae at 96 hpi. d Table comparing the cellular distribution in each cluster between the BmNPV-infected and control groups at 96 hpi.

Fig. 3.

Comparison of brain cell heterogeneity between BmNPV-infected and BmNPV-uninfected control larvae. a, b UMAP displaying the brain cell clusters in the PBS-treated group (a) and the BmNPV-infected group (b). c UMAP displaying the difference in the distribution of cells in the brain clusters between the BmNPV-infected and control larvae at 96 hpi. d Table comparing the cellular distribution in each cluster between the BmNPV-infected and control groups at 96 hpi.

Close modal

Strikingly, cluster 0 is characterized by high differential expression of many ribosomal protein genes (data not shown), which is associated with cell growth [31, 32]. It can be assumed that such cells are an attractive host for baculovirus infection because of the requirements of translation for high virus protein production [33]. In our previous study of scRNA-seq of silkworm hemocytes, increased ribosomal protein was observed in heavily infected cells [18]. Analysis of baculovirus gene expression shows that the diminished cluster 0 has the second highest viral load among the clusters in infected animals (71% of the cells carried a medium viral load [5–20% of total UMIs] and 12% were highly infected [>20%]). The relative decrease in cluster 0 could be explained by cell death after virus infection or the induction of apoptosis pathways [33]. Alternatively, since cluster 0 cells are not very differentiated and may function as precursor cells, an accelerated differentiation program may have become induced by the virus. Brain infection by baculovirus has indeed been associated with hyperactive behavior although the molecular mechanisms for the behavioral change remain to be elucidated [20, 21]. In such case, the changes in the brain cell populations toward higher heterogeneity may reflect altered brain function that will benefit the propagation of the virus in the environment (manuscript in preparation).

Hemocyte-Like (Cluster 12) and Its Potential Antiviral Function in the Silkworm Larvae Brain

To characterize the silkworm larvae brain responses against to BmNPV infection, we calculated the differential expression of genes between cell subsets of the BmNPV-infected and control samples using Seurat (applied on clusters with at least 1 cell). Compared to the uninfected brain, it was found that the expression of hundreds of host genes was significantly altered after BmNPV infection in clusters 0, 1, and 11 (PN4), while fewer host genes were affected by viral infection in other subgroups (less than 100; Fig. 4a). In a few clusters (14, 15, 16), the number of DEGs is less than 15, while in cluster 7 only one DEG was identified. The high number of DEGs in (unannotated) clusters 0 and 1 coincides with the large changes in cellular abundance during virus infection (Fig. 3d). With respect to cluster 11, it can be speculated that this cluster represents an important target for the virus and it remains to be determined in future experiments whether this cell type plays a key role in the regulation of behavioral changes during the late phases of infection.

Fig. 4.

Overview of the DEGs analysis between BmNPV-infected and control groups within brain cell clusters. a Histogram showing all upregulated (red) and downregulated DEGs (green) in BmNPV-infected cells compared to control cells within selected clusters at 96 hpi. b Heatmaps of DEG-enriched pathways in each brain cell subset after BmNPV infection at 96 hpi. c Dot plot showing the expression of the AMP genes in each brain cluster at 96 hpi. The intensity gradient represents the expression level, while the size of the dots represents the percentage of cells expressing the indicated gene per cluster. d Violin plots showing the expression level of NF-κB p65 (dorsal¸ dl), NF-κB p110 (Relish), STAT5A, STING, Dicer-2, and AGO2 in each subgroup of the BmNPV-infected and the control groups at 96 hpi.

Fig. 4.

Overview of the DEGs analysis between BmNPV-infected and control groups within brain cell clusters. a Histogram showing all upregulated (red) and downregulated DEGs (green) in BmNPV-infected cells compared to control cells within selected clusters at 96 hpi. b Heatmaps of DEG-enriched pathways in each brain cell subset after BmNPV infection at 96 hpi. c Dot plot showing the expression of the AMP genes in each brain cluster at 96 hpi. The intensity gradient represents the expression level, while the size of the dots represents the percentage of cells expressing the indicated gene per cluster. d Violin plots showing the expression level of NF-κB p65 (dorsal¸ dl), NF-κB p110 (Relish), STAT5A, STING, Dicer-2, and AGO2 in each subgroup of the BmNPV-infected and the control groups at 96 hpi.

Close modal

In the heatmap of DEGs-enriched pathways, relatively few pathways (only one pathway related to the immune response) become apparent that correspond to the response to virus infection (Fig. 4b). One reason may be that most subgroups have few DEGs after BmNPV infection, resulting in insufficient power for enrichment to the corresponding pathways. It is worth noting that only DEGs in cluster 12 (hemocyte-like) were significantly enriched in the immune-related Toll and Imd signaling pathway (Fig. 4b). Furthermore, we found that AMPs, major effectors of insect humoral immunity, are almost exclusively and strongly induced in cluster 12 (hemocyte-like) (Fig. 4c). However, the violin plots of key genes from major antiviral pathways including dorsal (dl, BMSK0006818), NF-κB p110 (BMSK0005070, Relish), STAT5A (BMSK0005942), STING (BMSK0001114), Dicer-2 (BMSK0013303), and AGO2 (BMSK0006970) showed that these genes were not significantly upregulated in each brain cell cluster after BmNPV infection (Fig. 4d). Surprisingly, these genes were hardly differentially expressed in cells from cluster 12 (hemocyte-like) where AMPs were strongly induced (Fig. 4d). Because it is well known that many immune pathways are regulated at the posttranscriptional level [34‒36], we speculate that when cells in cluster 12 produce AMPs in response to BmNPV infection, key immune transcription factors such as Rel/Dorsal, NF-κB p110 (Relish), and STAT5A become activated by immune signaling cascades in the absence of a transcriptional response. On the other hand, PN4 (cluster 11) shows increases in transcripts of Relish, STAT5A, STING, and Dicer-2 and some of these genes also become induced in Glia1 (cluster 3) and Glia2 (cluster 7) (Fig. 4d). Relish is the transcription factor with the broadest expression among the different clusters, but its expression decreases upon infection in KC3, PN1, and PN2. Similarly, STING expression is significantly reduced in PN3, PN5, optic lobe neuron, cluster 1, and cluster 8. Relative high expression of Ago2 in Glia2 is abolished after BmNPV infection (Fig. 4d). Thus, dynamic changes are observed with respect to the expression of immune-related factors in the brain clusters during BmNPV infection.

Evidence for Unique Characteristics of Hemocyte-Like Cluster 12 Cells

In our previous analysis of hemocytes of the hemolymph in silkworm larvae, 13 distinct clusters corresponding to hemocyte subtypes were identified and marker genes were proposed to characterize individual clusters [37]. Similarly, potential marker genes were predicted for the hemocyte-like cells in the brain (Fig. 1), which include three genes that encode factors that are commonly found in the hemocytes of the hemolymph (small heat-shock peptide, apolipoprotein-like and 30-kDa lipoprotein). However, one brain-specific marker gene (BMSK0009578) encodes a protein of the phosphatidylethanolamine-binding protein (PEBP) family that was not encountered in the hemolymph hemocytes. PEBPs are highly conserved proteins with diverse functions that include the regulation of neuronal development and the immune response [38, 39]. Furthermore, cluster 12 hemocyte-like cells in uninfected brain are enriched for the expression of ribosomal proteins (data not shown), which suggests a function in cell growth, as is observed for the cells of cluster 0. Finally, hemocyte-like cells in the brain show a robust increase in the expression of a large variety of AMP genes following BmNPV infection (attacins, gloverins [one gloverin gene is considered as a specific marker for cluster 12], lysozymes, moricins, cecropins, and others; Figure 5d), while only cecropins are enriched in hemolymph hemocytes. These observations argue for unique properties for the hemocyte-like cells in the brain, which may indicate that cluster 12 represents macrophage-like cells that are permanently resident in the brain and developed specific functions for brain maintenance (see discussion). This population increases twofold following baculovirus infection and it remains to be experimentally established whether the increase is caused by the proliferation of resident cells or recruitment from the hemolymph. The latter scenario is not very likely, however, because of the high viral loads observed in the hemocytes of the hemolymph [18]. BmNPV infection in the brain clearly is much more limited, which likely results from the presence of the blood-brain barrier (BBB) (discussed below) in combination with the low neurotropic character of baculovirus infection.

Fig. 5.

Lysozyme plays an antiviral role as an effector in the immune response of silkworm brain hemocytes. a UMAP displaying the distribution of cluster 12 (hemocyte-like cells) in the total silkworm brain population. b, c GO (b) and KEGG (c) analysis was performed on DEGs from cluster 12 (hemocyte like cells) after BmNPV infection. d Volcano plot showing that most of the DEGs induced by BmNPV infection in brain hemocytes are immune-related genes, especially AMP genes. e The violin plot further showed that lysozyme (BMSK0006899) expression is strongly induced after BmNPV infection only in brain hemocytes. f Silkworm larvae were injected with lysozyme-specific dsRNA and expression of lysozyme was quantified by qPCR after 48 h injection. g Six μg of dsRNA-lysozyme and 5 μL of BmNPV-eGFP (1 × 106.2 TCID50/mL) were co-injected into fifth-instar larvae. The expression of the viral gene vp39 was significantly downregulated at 96 hpi when endogenous lysozyme was knocked down. Each bar represents the mean ± standard deviation. ****p < 0.0001.

Fig. 5.

Lysozyme plays an antiviral role as an effector in the immune response of silkworm brain hemocytes. a UMAP displaying the distribution of cluster 12 (hemocyte-like cells) in the total silkworm brain population. b, c GO (b) and KEGG (c) analysis was performed on DEGs from cluster 12 (hemocyte like cells) after BmNPV infection. d Volcano plot showing that most of the DEGs induced by BmNPV infection in brain hemocytes are immune-related genes, especially AMP genes. e The violin plot further showed that lysozyme (BMSK0006899) expression is strongly induced after BmNPV infection only in brain hemocytes. f Silkworm larvae were injected with lysozyme-specific dsRNA and expression of lysozyme was quantified by qPCR after 48 h injection. g Six μg of dsRNA-lysozyme and 5 μL of BmNPV-eGFP (1 × 106.2 TCID50/mL) were co-injected into fifth-instar larvae. The expression of the viral gene vp39 was significantly downregulated at 96 hpi when endogenous lysozyme was knocked down. Each bar represents the mean ± standard deviation. ****p < 0.0001.

Close modal

Lysozyme Plays an Antiviral Role as an Effector in the Immune Response of Silkworm Brain Hemocyte

Although cluster 12 (hemocyte-like) constitutes a small proportion of all cells in the silkworm brain (Fig. 3, 5a), the induction of expression of AMP genes indicates an important role in the antiviral response of the brain during BmNPV infection (Fig. 4c). The GO biological process analysis also showed that the DEGs in cluster 12 were mainly enriched for immune-related terms such as defense response to bacterium, antibacterial humoral response, innate immune response, and response to biotic stimulus (Fig. 5b). KEGG analysis illustrated that the pathway with the most enriched DEGs in cluster 12 is Toll and Imd signaling pathway (Fig. 5c). A Volcano plot shows that most of the DEGs induced by BmNPV infection in cluster 12 are immune-related genes, especially AMPs (Fig. 5d). Among these AMP genes, lysozyme (BMSK0006899) has been shown to inhibit BmNPV replication during late-stage infection [40]. The violin plot further showed that lysozyme (BMSK0006899) expression is strongly induced after BmNPV infection only in brain cluster 12 (hemocyte-like) (Fig. 5e). When dsRNA-lysozyme was injected into the silkworm tail foot, the expression of lysozyme was significantly decreased in the brain at 48 h postinjection (Fig. 5f), implying that RNAi can be applied to knock down genes in the brain of silkworm. To further confirm the antiviral role of lysozyme during BmNPV infection of the brain, 6 μg of dsRNA-lysozyme and 5 μL of BmNPV-eGFP (1 × 106.2 TCID50/mL) were co-injected into fifth-instar B. mori larvae. The expression of the viral gene vp39 was significantly upregulated at 96 hpi when endogenous lysozyme was knocked down (Fig. 5g). These results suggest that the expression of lysozyme in hemocyte-like cells plays an important role in the antiviral immune response of the silkworm brain.

BmNPV is a common silkworm pathogen that also infects the CNS [41]. However, our understanding of the cellular composition and function of the silkworm brain is still very limited. Identifying the specific cell types in the silkworm brain is crucial for understanding the pathogenesis of BmNPV infection and the host immune response induced by the virus.

In this work, we report a comprehensive single-cell atlas of the silkworm larvae brain during BmNPV infection. We performed snRNA-seq and identified 19 transcriptionally distinct cell subtypes in the silkworm brain, including 3 KC, 2 glia cell, 6 olfactory projection neuron (PN), 1 optic lobe neuron, 1 hemocyte-like, 1 muscle cell-like, and 5 unannotated clusters (Fig. 2c, d). Obviously, at this stage of the analysis, many cell clusters have not been identified due to the lack of sufficient information of marker gene for silkworm brain cells. In fact, in the initial identification of Drosophila midbrain cells, a large number of cells were also unrecognized [14].

In contrast to the highly vascularized mammalian CNS, the insect brainfloats in the hemolymph [42]. Nevertheless, the BBB of insects is considered homologous to that of the vertebrates [43] and is achieved by tight junctions among glial cells instead of endothelial cells as in vertebrates [44]. The BBB ensures canonical brain function in insects and vertebrates by maintaining ionic integrity of the neuronal bathing fluid [45]. Viruses can cross the BBB by several mechanisms, such as direct infection of the endothelial/glial barrier cells, trans- or paracellular trafficking, or entry via infected macrophages [46, 47]. The only literature so far suggests that the tracheal system is responsible for disseminating BmNPV infection in the B. mori CNS and that the tracheal branches allow virions to pass through the BBB [48]. Recently, a study demonstrated that invasion of BmNPV into silkworm brain tissue could cause severe brain damage [41]. In the infected brain, the number of hemocyte-like cells was increased by twofold (Fig. 3). Although it is formally possible that the hemocyte-like cells (cluster 12) are infiltrated from the hemolymph after brain damage from viral disease, it is considered less likely according to clues obtained in this study (but see also below). On the other hand, it is clear that there are resident hemocytes in the silkworm brain under normal physiological conditions, which has also been confirmed in the brain of other insects [15, 16, 49, 50]. In Drosophila and other insects, the BBB is exclusively formed by glial cells that regulate the access of hemocytes to the brain [43]. Although we also identified glial cells (clusters 3 and 7) in the silkworm brain, we were not able to link them to a role in the BBB function or the regulation of hemocyte entry.

Insect hemocytes have been proposed to be the equivalent of the myeloid-like blood cells in vertebrates [51]. A large proportion of insect hemocytes have “macrophage-like” properties [37], and it is increasingly recognized that the macrophage lineage is conserved in metazoans from insects to vertebrates [52, 53]. Interestingly, macrophages can be divided into two functionally, developmentally and genetically distinct lineages [52, 53]. “Resident” macrophages originate during embryonic development and remain permanently associated with specific tissues in which they are capable to perform specialized tasks, e.g., synaptic pruning for brain-resident macrophages [54]. In insects, a second macrophage lineage is established during the larval stage in the hematopoietic organs [55, 56]. Similar to hematopoietic stem cells in vertebrates, hematopoietic organs can release large numbers of hemocytes into the hemolymph in response to infection [57]. The hemocyte-like cells that are identified in the brain tissue are distinct from the different hemocyte subsets isolated from the larval hemolymph (13 different clusters in uninfected larvae broadly corresponding to granulocytes, oenocytoids, plasmatocytes, and spherulocytes [18]). Brain hemocyte-like cells in cluster 12 are characterized by high differential expression of a variety of AMP genes (cecropin, attacin, moricin, lebocin, gloverin, enbocin, lysozyme; Fig. 5d), even in the noninfected condition, in contrast to hemocytes from the hemolymph where only the expression of cecropins is observed [18, 37]. This observation is consistent with additional functional roles for AMPs in neuronal function and development that were revealed more recently [58, 59]. AMPs have structural similarities with neuropeptides and some of the latter exhibit antimicrobial activity [60]. In Drosophila, long-term memory is modulated by neuronally expressed AMPs [61]. AMPs are also speculated to have important roles in neurodegeneration and Alzheimer’s disease in humans [58, 59]. From this point of view, we are more inclined to believe that the hemocyte-like cells of cluster 12 represent brain-specific resident macrophages that regulate neuron function and homeostasis through the expression of AMPs. Such a predominant nonimmunological role may also explain the lack of increased expression of regulators of the immune response, at least in the noninfected condition. After BmNPV infection of brain cells, however, AMP gene expression is significantly increased (log2FC values from 3 to 5), which strongly indicates an antiviral function that was corroborated by the effects of silencing of lysozyme on BmNPV replication in the brain (Fig. 5).

Increased expression of AMPs was also observed in our analysis of infection of fat body by BmNPV by snRNA-seq although in this tissue the expression of key immune regulators (e.g., Relish, STAT) was also increased [24]. Some of these AMPs have been predicted to have antiviral potential in our previous study [62]. An antiviral effect against AcMNPV infection was reported for gloverin [63]. Similarly, B. mori C-lysozyme (silkDB 3.0 ID: BMSK0006899) has been shown to have anti-BmNPV effects both in vivo and in vitro [40]. We further demonstrated that inhibition of lysozyme (BMSK0006899) results in increased BmNPV replication in the silkworm brain (Fig. 5). It has been reported that innate immune reactions also take place in the CNS to restrict pathogen replication and to induce antimicrobial protection [64, 65]. Thus, the induction of the expression of AMPs is strongly suggested to be a strategy for cluster 12 (hemocyte-like cells) to exert its antiviral function in the brain.

The data therefore indicate that hemocyte-like cells in the silkworm brain may have both neuronal and immunological functions. A recruitment of hemocytes from the hemolymph is considered less likely not only because of their distinct gene expression profiles [37] but also because of the much higher levels of virus replication in the hemolymph. At 3 dpi, a large majority of hemocytes in the hemolymph have high viral loads [18], while at 4 dpi viral loads in brain cells are mostly medium or low (data not shown; this study). High levels of viral replication have an obvious impact on the physiology of hemocytes (reprogramming to virion production), which is predicted to affect considerably their capacity to invade tissues and carry out repair of tissue damage. Most hemocyte-like cells in the brain (>90%) have only moderate viral loads (5–20% of total UMIs) at 4 dpi (data not shown), while almost all hemocyte clusters in the hemolymph predominantly have high viral loads (>20% of total UMIs) at 3 dpi [18] (in both studies infection is initiated by hemolymph injection with identical MOI). Only cluster 14 in hemolymph is moderately infected, but this subtype represents highly differentiated cells (plasmatocytes) at low abundance (<5%). Thus, infection in hemolymph occurs both more rapidly and more intensely than in brain which likely will inactivate the repair function of hemolymph hemocytes.

However, we do not have conclusive experimental evidence to prove whether the hemocyte-like cells in the silkworm brain that respond to viral infection are resident hemocytes or hemocytes infiltrated from the hemolymph. Since the marker genes we use to identify hemocyte-like cells in the brain are consistent with some marker genes of hemocytes in hemolymph, we cannot distinguish these cells through antibodies or probes. Actually, our focus in this study is not on the source of hemocyte-like cells in the brain, but rather on the presence of hemocyte-like cells in the silkworm brain, which produce antiviral responses. As for the source of hemocytes in insect brains, this is another scientific issue and a very interesting one. To my knowledge, it has not yet been discussed in fruit flies. The characterization and the source of this cell type will require more in-depth and systematic research in the future.

In conclusion, for the first time (to our knowledge), our snRNA-seq data provide a resource for a comprehensive systems-level understanding of the silkworm larvae brain and its response to virus infection. Six cell types including KCs, glial cells, olfactory PNs, optic lobes neurons, hemocyte-like cells, and muscle-like cells were identified in the silkworm brain. We found that extensive AMP gene expression is a defining characteristic of brain hemocytes. After BmNPV infection, AMP production became strongly upregulated and silencing of lysozyme increased BmNPV replication, indicating that AMP genes have a defensive function against baculovirus infection.

Thanks are due to Gene Denovo Corp. for its help in bioinformatics analysis. We also thank Jingjing Ning for her help in project coordination.

An ethics statement was not required for this study type, no human, or animal subjects or materials were used.

The authors declare that they have no conflicts of financial interest.

This work was supported by the Natural Science Foundation of Guangdong Basic and Applied Basic Research Fund (2022A1515012657); National Natural Science Foundation of China (31872426); Guangzhou Science and Technology Plan (202201010369); Guangdong Provincial Promotion Project on Preservation and Utilization of Local Breed of Livestock and Poultry (No. 2018-143); and South China Agricultural University high-level talent launch project.

M.F. designed the study scheme, collected and analyzed data, and drafted the manuscript. S.F., J.Z., J.X., W.L., and Y.H. helped with sample preparation and experiment validation. L.S. revised the manuscript. J.S. participated in coordination of the study and revised the manuscript. All authors read and approved the final manuscript.

Sequencing data have been deposited in China National Center for Bioinformation (CNCB) under the accession number CRA006827 (https://ngdc.cncb.ac.cn/search/?dbId=gsa&q=%20CRA006827&page=1).

1.
Reichert
H
,
Boyan
G
.
Building a brain: developmental insights in insects
.
Trends Neurosci
.
1997
;
20
(
6
):
258
64
. .
2.
Oland
LA
,
Tolbert
LP
.
Roles of glial cells in neural circuit formation: insights from research in insects
.
Glia
.
2011
;
59
(
9
):
1273
95
. .
3.
Gan
L
,
Liu
X
,
Xiang
Z
,
He
N
.
Microarray-based gene expression profiles of silkworm brains
.
BMC Neurosci
.
2011
;
12
:
8
. .
4.
Wang
GB
,
Zheng
Q
,
Shen
YW
,
Wu
XF
.
Shotgun proteomic analysis of Bombyx mori brain: emphasis on regulation of behavior and development of the nervous system
.
Insect Sci
.
2016
;
23
(
1
):
15
27
. .
5.
Louveau
A
,
Harris
TH
,
Kipnis
J
.
Revisiting the mechanisms of CNS immune privilege
.
Trends Immunol
.
2015
;
36
(
10
):
569
77
. .
6.
Lye
SH
,
Chtarbanova
S
.
Drosophila as a model to study brain innate immunity in health and disease
.
Int J Mol Sci
.
2018
;
19
(
12
):
3922
. .
7.
Rohrmann
GF
.
Baculovirus molecular biology
.
Bethesda (MD)
;
2019
.
8.
Wang
G
,
Zhang
J
,
Shen
Y
,
Zheng
Q
,
Feng
M
,
Xiang
X
, et al
.
Transcriptome analysis of the brain of the silkworm Bombyx mori infected with Bombyx mori nucleopolyhedrovirus: a new insight into the molecular mechanism of enhanced locomotor activity induced by viral infection
.
J Invertebr Pathol
.
2015
;
128
:
37
43
. .
9.
Wang
G
,
Na
S
,
Qin
L
.
Screening of Bombyx mori brain proteins interacting with protein tyrosine phosphatase of BmNPV
.
Arch Insect Biochem Physiol
.
2020
;
105
(
2
):
e21732
. .
10.
Han
X
,
Wang
R
,
Zhou
Y
,
Fei
L
,
Sun
H
,
Lai
S
, et al
.
Mapping the mouse cell atlas by microwell-seq
.
Cell
.
2018
;
173
(
5
):
1307
. .
11.
Han
X
,
Wang
R
,
Zhou
Y
,
Fei
L
,
Sun
H
,
Lai
S
, et al
.
Mapping the mouse cell atlas by microwell-seq
.
Cell
.
2018
;
172
(
5
):
1091
107.e17
. .
12.
Raj
B
,
Wagner
DE
,
McKenna
A
,
Pandey
S
,
Klein
AM
,
Shendure
J
, et al
.
Simultaneous single-cell profiling of lineages and cell types in the vertebrate brain
.
Nat Biotechnol
.
2018
;
36
(
5
):
442
50
. .
13.
Wang
Q
,
Peng
C
,
Yang
M
,
Huang
F
,
Duan
X
,
Wang
S
, et al
.
Single-cell RNA-seq landscape midbrain cell responses to red spotted grouper nervous necrosis virus infection
.
PLoS Pathog
.
2021
;
17
(
6
):
e1009665
. .
14.
Croset
V
,
Treiber
CD
,
Waddell
S
.
Cellular diversity in the Drosophila midbrain revealed by single-cell transcriptomics
.
Elife
.
2018
;
7
:
7
. .
15.
Davie
K
,
Janssens
J
,
Koldere
D
,
De Waegeneer
M
,
Pech
U
,
Kreft
L
, et al
.
A single-cell transcriptome atlas of the aging Drosophila brain
.
Cell
.
2018
;
174
(
4
):
982
98.e20
. .
16.
Li
H
,
Janssens
J
,
De Waegeneer
M
,
Kolluru
SS
,
Davie
K
,
Gardeux
V
, et al
.
Fly Cell Atlas: a single-nucleus transcriptomic atlas of the adult fruit fly
.
Science
.
2022
;
375
(
6584
):
eabk2432
. .
17.
Cui
Y
,
Behura
SK
,
Franz
AWE
.
Cellular diversity and gene expression profiles in the male and female brain of Aedes aegypti
.
BMC Genomics
.
2022
;
23
(
1
):
119
. .
18.
Feng
M
,
Xia
J
,
Fei
S
,
Peng
R
,
Wang
X
,
Zhou
Y
, et al
.
Identification of silkworm hemocyte subsets and analysis of their response to baculovirus infection based on single-cell RNA sequencing
.
Front Immunol
.
2021
;
12
:
645359
. .
19.
Dai
M
,
Sun
H
,
Zhao
L
,
Wu
Q
,
You
B
,
Xu
F
, et al
.
Duck CD8(+) T cell response to H5N1 highly pathogenic avian influenza virus infection in vivo and in vitro
.
J Immunol
.
2022
;
209
(
5
):
979
90
. .
20.
Hoover
K
,
Grove
M
,
Gardner
M
,
Hughes
DP
,
McNeil
J
,
Slavicek
J
.
A gene for an extended phenotype
.
Science
.
2011
;
333
(
6048
):
1401
. .
21.
van Houte
S
,
Ros
VI
,
Mastenbroek
TG
,
Vendrig
NJ
,
Hoover
K
,
Spitzen
J
, et al
.
Protein tyrosine phosphatase-induced hyperactivity is a conserved strategy of a subset of baculoviruses to manipulate lepidopteran host behavior
.
PLoS One
.
2012
;
7
(
10
):
e46933
. .
22.
Yao
L
,
Wang
S
,
Su
S
,
Yao
N
,
He
J
,
Peng
L
, et al
.
Construction of a baculovirus-silkworm multigene expression system and its application on producing virus-like particles
.
PLoS One
.
2012
;
7
(
3
):
e32510
. .
23.
Del-Aguila
JL
,
Li
Z
,
Dube
U
,
Mihindukulasuriya
KA
,
Budde
JP
,
Fernandez
MV
, et al
.
A single-nuclei RNA sequencing study of Mendelian and sporadic AD in the human brain
.
Alzheimers Res Ther
.
2019
;
11
(
1
):
71
. .
24.
Feng
M
,
Xia
J
,
Fei
S
,
Huang
Y
,
Lai
W
,
Swevers
L
, et al
.
Single-nucleus sequencing of fat body reveals distinct metabolic and immune response landscapes in silkworm larvae after Bombyx mori nucleopolyhedrovirus infection
.
J Immunol
.
2023
;
211
(
1
):
140
53
. .
25.
Freytag
S
,
Tian
L
,
Lonnstedt
I
,
Ng
M
,
Bahlo
M
.
Comparison of clustering tools in R for medium-sized 10x Genomics single-cell RNA-sequencing data
.
F1000Res
.
2018
;
7
:
1297
. .
26.
Wu
H
,
Xia
J
,
Fei
S
,
Wang
Y
,
Zhang
M
,
Guo
Y
, et al
.
BmCH25H, a vertebrate interferon-stimulated gene(ISG) homolog, inhibits BmNPV infection dependent on its hydroxylase activity in Bombyx mori
.
Insect Sci
.
2023
;
30
(
2
):
321
37
. .
27.
Wu
Y
,
Gu
Y
,
Morphew
MK
,
Yao
J
,
Yeh
FL
,
Dong
M
, et al
.
All three components of the neuronal SNARE complex contribute to secretory vesicle docking
.
J Cell Biol
.
2012
;
198
(
3
):
323
30
. .
28.
Mu
Y
,
Yan
X
,
Li
D
,
Zhao
D
,
Wang
L
,
Wang
X
, et al
.
NUPR1 maintains autolysosomal efflux by activating SNAP25 transcription in cancer cells
.
Autophagy
.
2018
;
14
(
4
):
654
70
. .
29.
Ahn
HJ
,
Jeon
SH
,
Kim
SH
.
Expression of a set of glial cell-specific markers in the Drosophila embryonic central nervous system
.
BMB Rep
.
2014
;
47
(
6
):
354
9
. .
30.
Yuasa
Y
,
Okabe
M
,
Yoshikawa
S
,
Tabuchi
K
,
Xiong
WC
,
Hiromi
Y
, et al
.
Drosophila homeodomain protein REPO controls glial differentiation by cooperating with ETS and BTB transcription factors
.
Development
.
2003
;
130
(
11
):
2419
28
. .
31.
Naora
H
.
Involvement of ribosomal proteins in regulating cell growth and apoptosis: translational modulation or recruitment for extraribosomal activity
.
Immunol Cell Biol
.
1999
;
77
(
3
):
197
205
. .
32.
Polymenis
M
.
Ribosomal proteins: mutant phenotypes by the numbers and associated gene expression changes
.
Open Biol
.
2020
;
10
(
8
):
200114
. .
33.
Nguyen
Q
,
Nielsen
LK
,
Reid
S
.
Genome scale transcriptomics of baculovirus-insect interactions
.
Viruses
.
2013
;
5
(
11
):
2721
47
. .
34.
Hua
X
,
Li
B
,
Song
L
,
Hu
C
,
Li
X
,
Wang
D
, et al
.
Stimulator of interferon genes (STING) provides insect antiviral immunity by promoting Dredd caspase-mediated NF-κB activation
.
J Biol Chem
.
2018
;
293
(
30
):
11878
90
. .
35.
Jiang
L
.
Insights into the antiviral pathways of the silkworm Bombyx mori
.
Front Immunol
.
2021
;
12
:
639092
. .
36.
Hu
Z
,
Zhu
F
,
Chen
K
.
The mechanisms of silkworm resistance to the baculovirus and antiviral breeding
.
Annu Rev Entomol
.
2023
;
68
:
381
99
. .
37.
Feng
M
,
Swevers
L
,
Sun
J
.
Hemocyte clusters defined by scRNA-seq in Bombyx mori: in silico analysis of predicted marker genes and implications for potential functional roles
.
Front Immunol
.
2022
;
13
:
852702
. .
38.
Wang
Z
,
Bu
J
,
Yao
X
,
Liu
C
,
Shen
H
,
Li
X
, et al
.
Phosphorylation at S153 as a functional switch of phosphatidylethanolamine binding protein 1 in cerebral ischemia-reperfusion injury in rats
.
Front Mol Neurosci
.
2017
;
10
:
358
. .
39.
Wang
S
,
Guo
H
,
Zhu-Salzman
K
,
Ge
F
,
Sun
Y
.
PEBP balances apoptosis and autophagy in whitefly upon arbovirus infection
.
Nat Commun
.
2022
;
13
(
1
):
846
. .
40.
Chen
TT
,
Tan
LR
,
Hu
N
,
Dong
ZQ
,
Hu
ZG
,
Jiang
YM
, et al
.
C-lysozyme contributes to antiviral immunity in Bombyx mori against nucleopolyhedrovirus infection
.
J Insect Physiol
.
2018
;
108
:
54
60
. .
41.
Li
Y
,
Zhang
J
,
Zhao
S
,
Wu
X
.
BmNPV-induced hormone metabolic disorder in silkworm leads to enhanced locomotory behavior
.
Dev Comp Immunol
.
2021
;
121
:
104036
. .
42.
Hindle
SJ
,
Bainton
RJ
.
Barrier mechanisms in the Drosophila blood-brain barrier
.
Front Neurosci
.
2014
;
8
:
414
. .
43.
Kubick
N
,
Klimovich
P
,
Bienkowska
I
,
Poznanski
P
,
Lazarczyk
M
,
Sacharczuk
M
, et al
.
Investigation of evolutionary history and origin of the Tre1 family suggests a role in regulating hemocytes cells infiltration of the blood-brain barrier
.
Insects
.
2021
;
12
(
10
):
882
. .
44.
Limmer
S
,
Weiler
A
,
Volkenhoff
A
,
Babatz
F
,
Klambt
C
.
The Drosophila blood-brain barrier: development and function of a glial endothelium
.
Front Neurosci
.
2014
;
8
:
365
. .
45.
Carlson
SD
,
Juang
JL
,
Hilgers
SL
,
Garment
MB
.
Blood barriers of the insect
.
Annu Rev Entomol
.
2000
;
45
:
151
74
. .
46.
Daniels
BP
,
Holman
DW
,
Cruz-Orengo
L
,
Jujjavarapu
H
,
Durrant
DM
,
Klein
RS
.
Viral pathogen-associated molecular patterns regulate blood-brain barrier integrity via competing innate cytokine signals
.
mBio
.
2014
;
5
(
5
):
e01476
14
. .
47.
Suen
WW
,
Prow
NA
,
Hall
RA
,
Bielefeldt-Ohmann
H
.
Mechanism of West Nile virus neuroinvasion: a critical appraisal
.
Viruses
.
2014
;
6
(
7
):
2796
825
. .
48.
Torquato
EF
,
Neto
MH
,
Brancalhao
RM
.
Nucleopolyhedrovirus infected central nervous system cells of Bombyx mori (L.) (Lepidoptera: bombycidae)
.
Neotrop Entomol
.
2006
;
35
(
1
):
70
4
. .
49.
Sheng
L
,
Shields
EJ
,
Gospocic
J
,
Glastad
KM
,
Ratchasanmuang
P
,
Berger
SL
, et al
.
Social reprogramming in ants induces longevity-associated glia remodeling
.
Sci Adv
.
2020
;
6
(
34
):
eaba9869
. .
50.
Zhang
W
,
Wang
L
,
Zhao
Y
,
Wang
Y
,
Chen
C
,
Hu
Y
, et al
.
Single-cell transcriptomic analysis of honeybee brains identifies vitellogenin as caste differentiation-related factor
.
iScience
.
2022
;
25
(
7
):
104643
. .
51.
Hartenstein
V
.
Blood cells and blood cell development in the animal kingdom
.
Annu Rev Cell Dev Biol
.
2006
;
22
:
677
712
. .
52.
Lazarov
T
,
Juarez-Carreno
S
,
Cox
N
,
Geissmann
F
.
Physiology and diseases of tissue-resident macrophages
.
Nature
.
2023
;
618
(
7966
):
698
707
. .
53.
Lazarov
T
,
Juarez-Carreno
S
,
Cox
N
,
Geissmann
F
.
Publisher Correction: physiology and diseases of tissue-resident macrophages
.
Nature
.
2023
;
619
(
7970
):
E51
. .
54.
Paolicelli
RC
,
Bolasco
G
,
Pagani
F
,
Maggi
L
,
Scianni
M
,
Panzanelli
P
, et al
.
Synaptic pruning by microglia is necessary for normal brain development
.
Science
.
2011
;
333
(
6048
):
1456
8
. .
55.
Lanot
R
,
Zachary
D
,
Holder
F
,
Meister
M
.
Postembryonic hematopoiesis in Drosophila
.
Dev Biol
.
2001
;
230
(
2
):
243
57
. .
56.
Nardi
JB
,
Pilas
B
,
Ujhelyi
E
,
Garsha
K
,
Kanost
MR
.
Hematopoietic organs of Manduca sexta and hemocyte lineages
.
Dev Genes Evol
.
2003
;
213
(
10
):
477
91
. .
57.
Anderl
I
,
Vesala
L
,
Ihalainen
TO
,
Vanha-Aho
LM
,
Ando
I
,
Ramet
M
, et al
.
Transdifferentiation and proliferation in two distinct hemocyte lineages in Drosophila melanogaster larvae after wasp infection
.
PLoS Pathog
.
2016
;
12
(
7
):
e1005746
. .
58.
Hanson
MA
,
Lemaitre
B
.
New insights on Drosophila antimicrobial peptide function in host defense and beyond
.
Curr Opin Immunol
.
2020
;
62
:
22
30
. .
59.
Stuart
BAR
,
Franitza
AL
,
Lezi
E
.
Regulatory roles of antimicrobial peptides in the nervous system: implications for neuronal aging
.
Front Cel Neurosci
.
2022
;
16
:
843790
. .
60.
Brogden
KA
,
Guthmiller
JM
,
Salzet
M
,
Zasloff
M
.
The nervous system and innate immunity: the neuropeptide connection
.
Nat Immunol
.
2005
;
6
(
6
):
558
64
. .
61.
Barajas-Azpeleta
R
,
Wu
J
,
Gill
J
,
Welte
R
,
Seidel
C
,
McKinney
S
, et al
.
Antimicrobial peptides modulate long-term memory
.
PLoS Genet
.
2018
;
14
(
10
):
e1007440
. .
62.
Feng
M
,
Fei
S
,
Xia
J
,
Labropoulou
V
,
Swevers
L
,
Sun
J
.
Antimicrobial peptides as potential antiviral factors in insect antiviral immune response
.
Front Immunol
.
2020
;
11
:
2030
. .
63.
Moreno-Habel
DA
,
Biglang-awa
IM
,
Dulce
A
,
Luu
DD
,
Garcia
P
,
Weers
PM
, et al
.
Inactivation of the budded virus of Autographa californica M nucleopolyhedrovirus by gloverin
.
J Invertebr Pathol
.
2012
;
110
(
1
):
92
101
. .
64.
Klein
RS
,
Garber
C
,
Howard
N
.
Infectious immunity in the central nervous system and brain function
.
Nat Immunol
.
2017
;
18
(
2
):
132
41
. .
65.
Klein
RS
,
Hunter
CA
.
Protective and pathological immunity during central nervous system infections
.
Immunity
.
2017
;
46
(
6
):
891
909
. .