Ex vivo characterization of the circulating hemocytes of bed bugs and their responses to bacterial exposure

Abstract

Bed bugs (Cimex spp.) are urban pests of global importance. Knowledge of the immune system of bed bugs has implications for understanding their susceptibility to biological control agents, their potential to transmit human pathogens, and the basic comparative immunology of insects. Nonetheless, the immunological repertoire of the family Cimicidae remains poorly characterized. Here, we use microscopy, flow cytometry, and RNA sequencing to provide a basal characterization of the circulating hemocytes of the common bed bug, Cimex lectularius. We also examine the responses of these specialized cells to E. coli exposure using the same techniques. Our results show that circulating hemocytes are comprised of at least four morphologically distinct cell types that are capable of phagocytosis, undergo degranulation, and exhibit additional markers of activation following stimulation, including size shift and DNA replication. Furthermore, transcriptomic profiling reveals expression of predicted Toll/IMD signaling pathway components, antimicrobial effectors and other potentially immunoresponsive genes in these cells. Together, our data demonstrate the conservation of several canonical cellular immune responses in the common bed bug and provide a foundation for additional mechanistic immunological studies with specific pathogens of interest.

1. Introduction

The common bed bug, Cimex lectularius, is a hematophagous ectoparasite that feeds almost exclusively on humans and is ubiquitous across the globe (Davies et al., 2012). Although there are many toxicological, behavioral and genetic studies of bed bugs, less is known about other fundamental aspects of their biology. Bed bugs display physiological and ecological similarities to insects that are pathogen vectors (e.g. lice, kissing bugs), yet their ability to transmit disease agents in nature appears limited (Delaunay et al., 2011; Lai et al., 2016; Pietri, 2020). Some have speculated that this property could be linked to robust immune responses that are specific to bed bugs. Similarly, it is hypothesized that due to their unusual reproductive strategy of traumatic insemination (Reinhardt et al., 2015), bed bugs may have evolved a specialized immune system that is unique among hematophagous insect taxa. In fact, infections acquired through copulatory wounding can significantly reduce bed bug fitness and the insects possess a unique organ known as the spermalage that helps mitigate the costs of these infections (Reinhardt et al., 2005; Reinhardt et al., 2003). The spermalage is a membrane-bound sack located in the hemocoele at the site of copulatory wounding and it is replete with cells (hemocytes) that possess antimicrobial activity. In addition to the evolution of this organ, bed bugs appear able to preemptively upregulate some immune functions such as lysozyme activity based on predictable feeding cues that precede traumatic insemination (Siva-Jothy et al., 2019). This unusual ability further helps to reduce the costs associated with sexually acquired infections. Moreover, multiple hemipterans, including C. lectularius and the kissing bug Rhodnius prolixus, appear to encode a modified immune deficiency (IMD) signaling pathway that is sparser than that of holometabolous insects and lacks some components (Benoit et al., 2016; Salcedo-Porras et al., 2019; Zumaya-Estrada et al., 2018). Yet, despite the many interesting features of the bed bug immune system, its cellular and molecular functions remain poorly characterized.

The immune systems of mammals and insects share numerous conserved responses against foreign pathogens (Yuan et al., 2014; Sheehan et al., 2018; Kumar et al., 2018). Although insects do not exhibit true immune memory and innate immune priming following pathogen exposure remains only partially understood, innate responses to bacteria, viruses, parasites, and pathway-specific immune activators have been examined in numerous species and demonstrate substantial complexity. The innate immune system of insects includes a humoral component, comprised of antimicrobial peptides/enzymes, melanin, and reactive oxygen/nitrogen species, as well as a cellular component. The cellular response is carried out by specialized cells termed hemocytes. These are found primarily in the hemolymph (circulating hemocytes) and to a lesser extent in association with tissues (sessile hemocytes). Some molecular mechanisms of hemocyte function have been elucidated in well-studied taxa such as mosquitoes and fruit flies. These are reviewed elsewhere (Lavine and Strand, 2002; Hillyer, 2015; Hillyer and Strand, 2014) and will not be discussed here in detail. Generally, hemocytes perform three broad functions: phagocytosis, encapsulation and nodulation. They further contribute to humoral responses through the production and release of effector molecules via degranulation and secretion. In mosquitoes, hemocytes also demonstrate the ability to alter their tissue localization during infection (King and Hillyer, 2012; King and Hillyer, 2013) and hemocyte proliferation in response to a blood meal or infection stimulus has been documented (King and Hillyer, 2013; Bryant and Michel, 2014).

The classification of hemocyte types by form or function is controversial and not well standardized across most insect taxa. Drosophila is a notable exception as hemocytes have been extensively characterized in this model and molecular markers have been established to distinguish cell types (Evans et al., 2014). Similarly, in mosquitoes, a multitude of studies have led to the consensus that three functional types exist: prohemocytes, granulocytes, and oenocytoids (Hillyer and Strand, 2014). Prohemocytes of mosquitoes are characterized by their small size and a large nucleus to cytoplasm ratio. They are largely thought to be progenitor cells that differentiate into other cell types, though they have also been shown to have phagocytic and lytic activity. Oenocytoids are large homogenous cells that function primarily through the phenoloxidase pathway, while granulocytes are polymorphic cells that exhibit phagocytic activity but also produce a number of lytic molecules and are involved in melanization. In contrast to mosquitoes, studies of six species of blood-feeding triatomine bugs (Order: Hemiptera) have revealed seven morphological hemocyte types, but not all are present across each of the species investigated (deAzambuja et al., 1991). In addition to the three hemocyte types found in mosquitoes, triatomines possess plasmatocytes and cystocytes (deAzambuja et al., 1991). Plasmatocytes are abundant, often spindleshaped cells that vary significantly in size. On the other hand, cystoytes are small, fragile cells that circulate in fewer numbers and form extracellular strands that participate in aggregation. Triatomines also possess adipohemocytes and giant cells (deAzambuja et al., 1991). However, giant cells are extremely rare across all triatomines, and adipohemocytes along with oenocytoids are lacking in both Panstrongylus megistus and Triatoma infestans (deAzambuja et al., 1991).

A short report on the circulating hemocytes of the tropical bed bug, Cimex hemipterus, described prohemocytes, granulocytes, oenocytoids, and plasmatocytes, consistent with triatomines (Sonawane and Sonawane, 2017). The same authors also noted the presence of two additional cell types, spherulocytes and thrombocytoids, but unfortunately did not provide high-resolution images to accompany any descriptions (Sonawane and Sonawane, 2017). Meanwhile, limited research using microscopy has shown that at least two unclassified but distinct types of hemocytes exist in Cimex lectularius, and hemocytes in the paragenital tract, hemolymph, and spermalage appear to phagocytose sperm (Reinhardt and Siva-Jothy, 2007). However, as previous studies in C. lectularius have focused on the roles of bed bug hemocytes during reproduction, very little is known about the molecular properties and functions of circulating hemocytes in the context of exogenous pathogenic challenges. Knowledge of the cellular immune system of bed bugs could provide insight into the factors that regulate their vector competence for human pathogens. In addition, understanding bed bug immune responses may inform the development of biological control strategies involving entomopathogens. For these reasons, and because few studies of the cellular immune system of insects in the family Cimicidae have been conducted, the aim of the present work was to characterize the varying subtypes of circulating bed bug hemocytes and their responses to bacterial challenge using brightfield microscopy, flow cytometry and RNA sequencing. Here, an ex vivo approach was employed to enable more controlled stimulation of hemocytes than in vivo while minimizing alternations in cell physiology that can accompany prolonged culture of hemocytes.

2. Material and methods

2.1. Bed bug rearing

The Cincinnati field strain of Cimex lectularius was used for all experiments. This strain is derived from insects collected in Cincinnati, OH in 2007 by technicians from Sierra Research Laboratories, Inc. (Modesto, CA). Bed bug colonies were maintained in plastic jars containing corrugated cardboard harborages at 28 +/− 1 °C and 60–70% relative humidity. The photoperiod was 12 h of light: 12 h of dark. Colonies were fed mechanically defibrinated rabbit blood (Hemostat Laboratories, Dixon, CA) once per week using a Hemotek membrane feeding system (Hemotek LTD, Blackburn, UK).

2.2. Hemocyte extraction

For all experiments, adult C. lectularius were starved for seven days and anesthetized in petri dishes by chilling on wet ice for approximately five minutes. The surface of the insects was then decontaminated by successive rinses in 10% bleach and deionized water and allowed to dry. Hemolymph containing circulating hemocytes was extracted by inserting a pulled glass capillary needle into the ventral thorax under a dissecting microscope. The collected hemolymph was quickly diluted in tubes of PBS on ice and immediately prepared for further analyses to prevent aggregation. Using this method, typically ∼3000–5000 cells could be obtained per individual insect as determined by counting on a hemocytometer, but significant qualitative variation in the volume of hemolymph present in individual insects was noted.

2.3. Microscopy of hemocytes

For initial characterization by microscopy and flow cytometry, hemocytes extracted from both male and female bed bugs were examined to ensure that all cell types were captured. Hemocytes were directly visualized by fixing and staining with Giemsa or by phase contrast while live. For Giemsa staining, a mix of freshly extracted hemocytes from both sexes was allowed to adhere onto a glass slide and air dry at room temperature for approximately 1 h (Kadota et al., 2003; Ivanina et al., 2017; Yu et al., 2016). Slides were then fixed in methanol, dried, and subsequently incubated with 5% Giemsa stain solution in methanol before rinsing with water and drying once again. Stained cells were photographed on an upright Zeiss M-1 Axio imager (Zeiss, Oberkochen, Germany) using a 63× oil immersion objective with an AxioCam HRc (Zeiss). Cell measurements were carried out with ImageJ v.1.52a. For live visualization, fresh hemocytes were extracted, diluted in insect cell medium (ESF 921, Expression Systems, Davis, CA), and placed into a hemocytometer for imaging on the same microscope using a 40× phase contrast objective.

2.4. Characterization of unstimulated hemocytes by flow cytometry

A BD Accuri™ C6 plus cell cytometer with a 488 nm excitation laser was used for all flow cytometry according to the manufacturer’s standard protocols (BD Biosciences, San Jose, CA). Unstained hemolymph samples from males or females were run separately for initial physical analyses and used for gating and removing autofluorescence during additional analyses, including those that employed fluorescent staining. Syto-9 (ThermoFisher) cellular nucleic acid stain was used to further discriminate hemocytes in the FITC channel. For Syto-9 staining, freshly extracted hemocytes were first fixed by adding chilled ethanol to hemolymph suspended in PBS to a final concentration of 70%. Cells were incubated for 10 min, the ethanol was removed by centrifugation at 1200 RPM for 5 min and the resulting pellet was resuspended in MACs buffer. Syto-9 was then added to the sample at a final concentration of 1:500, incubated for 10 min, and then removed by centrifugation. The pellet was resuspended again in MACs buffer for flow cytometry. At least 20,000 events were collected for every sample analyzed and three biological replicates were conducted for each sex.

2.5. Hemocyte morphology after exposure to heat-killed bacteria

Subsequent analyses of responses to bacteria were carried out using only male insects in order to preserve females and maintain long-term viability of the colony. Hemocytes were extracted from male bed bugs as previously described. Escherichia coli (H4H strain, BEI Resources, Manassas, VA) was grown to log phase at 25 °C in LB media with 50 µg/ml ampicillin. The bacteria were incubated at 60 °C for 10 min to heatkill them without extensive denaturation. Heat killed E. coli, a common immune elicitor, was then diluted in PBS and approximately 1.4 × 10⁵ bacteria were added to freshly extracted hemocyte pools (55,000–65,000 cells per replicate). The hemocytes were incubated with heat-killed bacteria for 30 min at room temperature. After treatment with heat-killed bacteria, hemocyte samples were processed into MACS buffer as previously described and immediately analyzed on the cytometer with at least 5000 events collected per sample. Untreated cells (∼20,000 cells per replicate) from the same pool and heat-killed E. coli alone were run as controls. Three biological replicates were conducted.

2.6. Cell cycle analysis

Male hemocytes were collected and heat killed E. coli (H4H) were added to treatment groups or PBS to control groups, according to the aforementioned protocols. After approximately 30 min of incubation with either heat killed E. coli or diluent, cells were fixed at room temperature for one hour by adding 100% ethanol dropwise until a final concentration of 70% ethanol was reached. Cells were then washed by centrifugation at 1200 rpm for 5 min. RNAse treatment was carried out to remove background by adding 10 µg/mL of RNAse A (ThermoFisher) and incubating for 10 min. Nucleic acid staining was conducted by adding propidium iodide (PI; Thermo Fisher), an indicator of DNA content, to a final concentration of 50 µg/mL in PBS. Cells were incubated in the dark at room temperature for 10 min before acquiring data on the BD Accuri C6 plus in the FL-3 channel. At least 10,000 events were collected per sample and three biological replicates were conducted. Cell cycle analysis was completed using FloJo v10.2.1.

2.7. Live bacteria phagocytosis assay

To examine phagocytosis of live bacteria by male hemocytes, a GFP-expressing strain of E. coli (Strain B21, Ward’s Science, Rochester, NY) was used. As strain H4H, E. coli B21 was grown to log phase at 25 °C in LB media with 50 µg/ml ampicillin. The log phase culture was then diluted in PBS and approximately 1 × 10⁵ live bacteria were added to freshly extracted hemocytes. After incubation for 30 min at room temperature, the cells were fixed with 4% paraformaldehyde (PFA) for 20 min on ice. The PFA fix was employed in the place of ethanol because it is a better fixative to prevent contamination of the cytometer for assays involving live bacteria. PFA and free bacteria were removed by centrifugation at 1200 RPM for 10 min and the resulting cell pellet was resuspended in MACs buffer for flow analysis. PFA fixed but untreated cells and E. coli were run for gating controls. The experiment was repeated three times.

2.8. Fluorescent bioparticle assay

To further verify the activities of male hemocytes following exposure to bacteria, E. coli-derived fluorescently labelled bioparticles (ThermoFisher) were added to freshly extracted hemolymph at a concentration of approximately 6 × 10³ particles/µL according to the manufacturer’s recommendations. The bioparticle treatment was conducted with or without trypan blue, which acts as a quencher of noninternalized bioparticles (Wan et al., 1993). After incubation for 1 h, free bioparticles were removed by centrifugation and the hemocyte pellet was resuspended in MACs buffer for flow cytometry as described above. Untreated hemolymph and E. coli bioparticles were run as gating controls. The experiment was repeated three times.

2.9. Flow cytometry data analysis

Each flow cytometry experiment was repeated at least three times and all flow cytometry data analyses were carried out using FloJo v10.6.1 software (BD Biosciences). All samples were analyzed on FloJo by first gating cells of interest away from debris by size (FSC-A) vs granularity (SSC-A.) This population was then gated on FSC-A by FSC-H for doublet discrimination. Untreated hemocytes and control samples were used to set all gates (e.g. to remove autofluorescence) for any fluorescence or physical cell analyses. The gating strategy is further illustrated in (Fig S1).

2.10. RNA sequencing

Freshly extracted hemolymph from male bed bugs in PBS was pelleted by centrifugation at 2,000 for 5 min and re-suspended in PBS to wash the hemocytes. The hemocytes were divided into two groups which were treated with either heat-killed E. coli as described above, or with diluent (LB media), and incubated at room temperature for 1 h. After treatment, the cells were pelleted again by centrifugation at 1200 RPM for 10 min and then re-suspended in RLT lysis buffer (ThermoFisher) for RNA extraction. RNA was extracted using the RNeasy kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. The experiment was repeated with three independent cohorts of bed bugs. Total RNA extracted from the three replicates was pooled and used for RNA sequencing and analysis of gene expression as follows. RNA samples were first treated with DNAase (Ambion Life Technologies) according to the manufacturer’s standard procedure. Then, the Agencourt RNAClean XP purification protocol (Agencourt, Beverly, MA) was followed for RNA clean-up. The SMART-Seq Ultra Low Input RNA sequencing kit (Clontech Laboratories, Inc., Mountain View, CA) with 2 ng of input RNA from each sample was used to generate cDNA. 11 cycles of PCR were performed and 150 pg of cDNA was then used as input to prepare Nextera XT libraries according to standard protocol with 12 PCR cycles (Illumina, San Diego, CA). Sequencing of 75 bp paired-end reads was carried out on a mid-output flow cell with an Illumina NextSeq 550 (Illumina). All raw reads were deposited into the NCBI SRA and will be publicly available upon publication of this manuscript (PRJNA607805).

For analysis, the original FASTQ format reads were trimmed using the fqtirm tool (https://ccb.jhu.edu/software/fqtrim) to remove adapters, terminal unknown bases (Ns) and low quality 3′ regions (Phred score <30). The trimmed FASTQ files were then processed using FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc) for quality control. The trimmed FASTQ files were mapped and annotated to the bed bug reference genome (Benoit et al., 2016) (http://www.hgsc.bcm.edu/arthropods/bed-bug-genome-project) using CLC Genomics Workbench 12 for RNAseq analyses. Normalized expression values (TPM; Transcripts per Kilobase Million) were calculated for all transcripts by dividing read counts by the length of each gene in kilobases. The TPM values were used to measure transcript abundances and calculate fold changes in E. coli treated hemocytes relative to control (fold change < 0.5 or >2).

3. Results

3.1. Microscopy of hemocyte populations

Upon visualization of hemolymph stained with Giemsa, four distinct types of hemocytes were discerned based on size, morphology, and staining properties. We classified the hemocyte types based on previous classifications of stained hemocytes from various insects (Kadota et al., 2003; Yu et al., 2016; Hernandez et al., 1999; Kwong et al., 2014; Hong et al., 2018). Our observations included: granulocytes (Fig. 1A), plasmatocytes (Fig. 1B), prohemocytes (Fig. 1C), and oenocytoids (Fig. 1D). Granulocytes were generally round with a slightly irregular plasma membrane and were noted to have granules of varying number and size that stained dark purple or pink. The nuclei were peripheral and often multi-lobbed. These cells readily formed thin membrane projections (filopodia) when adhered onto glass, and typically ranged in diameter from 7 µm to 11 µm when fixed. On a qualitative basis, they were the most frequently observed type of hemocyte. Plasmatocytes were observed to have a smooth plasma membrane and contain central nuclei which were large in relation to the volume of the cytoplasm. They were similar in size to granulocytes and sometimes displayed thick membrane protrusions or a spindle shape. Prohemocytes contained large, densely stained central nuclei, but were smaller in size, typically ranging from 3.5 µm to 4 µm in diameter when fixed. They were similar in abundance to granulocytes and homogenous in appearance. Oenocytoids were seen as large oval-shaped cells with a dense nucleus and irregular plasma membrane. They sometimes contained translucent vesicles. These cells, which are known to be tissue associated in other insects (Hillyer and Strand, 2014), ranged in diameter from ∼10 µm to as large as 40 µm, had a variable nucleus: cytoplasm ratio, and were qualitatively less abundant in the circulation than other types. Phase contrast microscopy of live, unstained hemocytes revealed morphologies consistent with those identified by Giemsa staining, though some expected variation was apparent due to the lack of fixing and staining (Fig. 1E-H, Fig S2).

Fig. 1.

Representative images of hemocytes. (A-D) Giemsa stained hemocytes. 630× magnification. (E-H) Live, unstained hemocytes. 400× magnification. (A, E) Granulocyte with peripheral multi-lobbed nucleus (asterisk), cytoplasmic granules (arrows), and thin filopodia (arrowheads). (B, F) Plasmatocyte with large nucleus and membrane protrusion (arrow). (C, G) Densely-stained small prohemocyte. (D, H) Large oenocytoid with translucent vesicles (arrows).

3.2. Flow cytometry analysis of unstimulated bed bug hemocytes

Analysis of unstained hemolymph by flow cytometry discriminated multiple distinct populations of hemocytes by size (FSC-A) and granularity (SSC-A) in both males and females (Fig. 2A-D). The granularity of female derived hemocytes was observed to be significantly greater than that of male hemocytes based on median side scatter (Fig. 2A-B, paired t-test, p = 0.0076). Syto-9 cellular staining revealed similar populations of hemocytes by granularity (SSC-A) and size (FSC-A), supporting the results of the unstained analyses (Fig. 2E-H). In combination with morphological visualization, these data provide a baseline morphological profile of the circulating hemocytes of bed bugs.

Fig. 2.

Bed bug hemocyte analysis by flow cytometry. Hemolymph samples from bed bugs were collected and prepared for flow cytometry as described in the methods. (A-B) Granularity by SSC-A of unstained hemocytes from males and females. (C-D) Size by FSC-A of unstained hemocytes from males and females. (E-F) Granularity by SSC-A of Syto-9 stained hemocytes from males and females. (G-H) Size by FSC-A of Syto-9 stained hemocytes from males and females.

3.3. Cellular responses to bacterial exposure

Changes in immune cell size and/or granularity are hallmarks of activation that are conserved from mammals to insects (Mudoi et al., 2019; Yuan et al., 2014; Sheehan et al., 2018; Kumar et al., 2018). Accordingly, we next examined the characteristics of bed bug hemocytes following exposure to heat-killed E. coli, a common immune elicitor, to determine if these cells mount similar responses (Fig. 3). Flow cytometry analysis of hemocytes exposed to heat-killed E. coli for 30 min showed a significant decrease in granularity (SSC-A) compared to untreated control cells from the same pool (Fig. 3A-B, paired t-test, p = 0.055). This result indicated that degranulation events occurred in response to exposure to bacterial products. On the other hand, hemocyte size as measured by FSC-A was slightly but significantly increased after exposure to E. coli (Fig. 3C-D, paired t-test, p = 0.0066). When hemocytes were treated with E. coli-derived bioparticles tagged with AlexaFlour 488, the cells also showed a decrease in granularity (SSC-A) similar to that seen in response to heat-killed bacteria (Fig. 3E), providing further support that the cells degranulate in response to bacterial exposure.

Fig. 3.

Cellular responses to heat-killed bacteria. Hemocytes from male bed bugs were prepared as described in the methods, treated with heat-killed E. coli, and processed for flow cytometry. (A-B) Cell granularity (SSC-A) before and after exposure to bacteria. (C-D) Cell size (FSC-A) before and after exposure to bacteria. Median fluorescence intensities from three replicates were analyzed by paired t-test to identify significant differences between the two groups. (E) Granularity (SSC-A) was examined in male hemocytes exposed to E. coli-derived fluorescent bioparticles. Median fluorescence intensity shows a decrease in granularity following exposure, similar to that observed in hemocytes exposed to heat-killed E.coli. Normalized to mode to show percentages and account for varying cell counts between groups.

Previous preliminary work suggested that circulating and tissue associated hemocytes of bed bugs can undergo division in response to infection (Wilson and Siva-Jothy, 2014). Thus, we also investigated this phenomenon through propidium iodide (PI) staining of unstimulated and heat-killed E. coli exposed hemocytes ex vivo (Fig. 4). PI was used in these experiments because Syto-9 is a suitable cell-specific dye, but not a precise indicator of chromosomal content. Analysis of PI staining revealed a population distribution characteristic of eukaryotic cells undergoing the cell cycle. That is, two peaks indicative of the G0/G1 and G2 phases separated by a valley of varying DNA content indicative of S phase were observed (Fig. 4). Most cells in both stimulated and unstimulated groups were in the G0/G1 phase (91.3–99.6%), with smaller proportions undergoing DNA synthesis (S phase, 0.22–2.48%) or in G2 phase (0.29–5.7%). Moreover, in each of three biological replicates, stimulation with heat-killed bacteria resulted in a decrease in the proportion of cells in G0/G1 and increases in the proportion of cells in S and G2 phases. The ability of a subset of the circulating hemocyte population to undergo DNA replication was further supported by positive bromodeoxyuridine (BrdU) staining (Fig S3). Consistent with the results of PI staining experiments, in each of three biological replicates BrdU incorporation was greater in hemocytes treated with heat-killed bacteria than in unstimulated hemocytes.

Fig. 4.

Cell cycle analysis after bacteria treatment. Hemocytes were treated with either heat killed E. coli strain H4H (orange) or PBS as a control (red) and the cell cycle was assessed by propidium iodide staining. All three replicates showed a decrease in cells in G0/G1, and increases in cells in S and G2 phases after 30-minute incubation with heat killed E. coli. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.4. Phagocytosis of live bacteria

To examine the phagocytic activity of bed bug hemocytes against bacteria, cells were incubated for 30 min with a strain of E. coli that constitutively expresses GFP (Fig. 5A). In these experiments, untreated control hemocytes showed no fluorescence (-FITC). Meanwhile, individual E. coli cells were weakly fluorescent. However, the hemocytes treated with E. coli had a median fluorescence intensity that was nearly double that of E. coli cells alone. The majority of hemocytes treated with bacteria were GFP positive (+FITC), but some showed no fluorescence, similar to untreated controls. These results demonstrated that some, but not all bed bug hemocytes are able to phagocytose bacteria. Further microscopy examination of hemocytes incubated with E. coli-derived fluorescent bioparticles enabled the direct visualization of phagocytosed material (Fig. 5B).

Fig. 5.

Phagocytic activity of hemocytes. Hemocytes from male bed bugs were prepared as described in the methods, treated with live GFP-expressing E. coli, and processed for flow cytometry as described in the methods. (A) Fluorescence of untreated hemocytes, GFP-expressing E. coli, and hemocytes incubated with E. coli measured in the FITC channel. Analysis shows increased fluorescence in treated hemocytes indicating uptake of bacteria. (B) Hemocytes from male bed bugs were exposed to E. coli-derived fluorescent bioparticles. Epifluorescence phase-contrast microscopy showed free fluorescent bioparticles as well as internalization of numerous particles by a hemocyte (arrow). 400× magnification.

3.5. Gene expression in hemocytes

RNA sequencing produced 149,639,158 raw reads for unstimulated hemocytes and 147,476, 464 reads for hemocytes stimulated with heat-killed E. coli. Initial analysis focused on previously identified immunity-related genes in the bed bug genome (Benoit et al., 2016). Among these, antimicrobial effectors, including defensin-like peptides, diptericin-like peptides, lysozymes and prophenoloxidases (PPOs) were expressed in circulating hemocytes (Table 1). The latter two were the most abundantly expressed immunity-related genes. Thioester-containing proteins (TEP), on the other hand, were minimally expressed. Components of the IMD and Toll signaling pathways (e.g. TAK1, Toll receptors, Cactus, Pellino) were also expressed to an appreciable degree. Notably, a predicted homolog of Caspar, an inhibitor of the IMD pathway (Kim et al., 2006), was not expressed in unstimulated or stimulated cells despite being present in the genome, which may have interesting implications for IMD pathway regulation in bed bug hemocytes.

Table 1.

Expression of immunity-related genes in hemocytes.

Gene ID	Gene Name	TPM Untreated	TPM Stimulated
LOC10661791	Defensin-like	98	124
LOC10661792	Defensin-like	53	81
LOC106664366	Diptericin-like	298	395
LOC112127760	Diptericin-like	40	43
LOC106663586	Lysozyme-like	16	10
LOC106667626	Lysozyme-like	7	9
LOC106666694	Lysozyme-like	64	42
LOC106669094	Lysozyme-like	19	20
LOC106663700	Lysozyme-like	470	804
LOC106673939	PPO-like	867	871
LOC106673564	PPO-like	4370	4190
LOC106661077	TEP	7	6
LOC106665617	TEP	20	16
LOC106665718	PGN-RP	170	154
LOC106673865	Caspar	0	0
LOC106666374	IKK1	3	1
LOC106672599	Posh	24	21
LOC106666550	TAK1	117	111
LOC106671087	Dome	47	43
LOC106664605	HOP	3	2
LOC106661660	STAT	65	49
LOC106667063	Rel	43	44
LOC106667016	Rel	45	27
LOC106670370	Toll	112	98
LOC106672311	Toll	42	33
LOC106672979	Toll	33	25
LOC106672317	Toll	1	1
LOC106662834	Toll	8	8
LOC106667081	Cactus	230	243
LOC106668016	Cactus	20	18
LOC106667707	Cactus	11	11
LOC106662592	MYD88	26	31
LOC106663343	Pelle	35	34
LOC106670316	Pellino	100	87
LOC106672583	Tube	24	21
LOC106671955	Spatzle	1	1
LOC106667522	Dicer	12	10
LOC106667539	Dicer	45	38
LOC106672312	Drosha	23	20
LOC106668600	Argonaute	68	71

^*TPM = transcripts per million rounded to the nearest integer.

In total, 4095 unique genes had a normalized expression value between 50 and 1400 in unstimulated hemocytes and 3939 unique genes were in the same range in stimulated hemocytes. Based on analysis of fold change we further identified 997 genes with a fold change >2 in treated hemocytes and 1106 genes with a fold change < 0.5 (Fig. 6). Among these, we selected those with >2.5 fold increase and a normalized expression value (TPM) of > 3 in treated hemocytes, yielding a list of 37 genes of additional interest that may potentially be upregulated in response to bacterial challenge (Supplementary Table 1). The pooling of RNA from biological replicates for technical reasons accounted for possible variation between experiments but did not allow for statistical comparison of differential expression between untreated and E. coli exposed hemocytes. As such, we were unable to obtain measures of certainty that fold changes in gene expression represent true biological responses. No less, these descriptive data provide additional transcriptional insight on the activities of circulating bed bug hemocytes and will be a useful resource to inform targeted studies of gene expression and function in the same cells.

Fig. 6.

Up and down regulation of gene expression in hemocytes exposed to heat-killed E. coli. Hemocytes from male bed bugs were prepared as described in the methods, treated with heat-killed E. coli, and harvested for RNA extraction and sequencing. Fold change was calculated based on the normalized expression values (TPM) of genes in control and treated samples (treated/control).

4. Discussion

Here, we describe the morphological features of the circulating hemocytes of C. lectularius and we establish the first baseline flow cytometric and transcriptomic profiles of these cells. In addition, we describe the conservation of several cellular immune responses to bacterial challenge in an important and taxonomically unique insect pest. This information expands current knowledge of bed bug hemocytes beyond the few previously identified functions in reproduction (Reinhardt and Siva-Jothy, 2007) and provides a methodological foundation for mechanistic studies.

Several distinct types of hemocytes were found in C. lectularius. Generally, the morphological features and abundances of these were consistent with those described in reports of some other insect species. For instance, as in C. lectularius, granulocytes are the most abundant type of circulating hemocyte in mosquitoes, several species of triatomines, and adult C. hemipterus (Hillyer and Strand, 2014; deAzambuja et al., 1991; Sonawane and Sonawane, 2017). They also readily form filopodia-like structures, as those of termites, crickets, and beetles (Hong et al., 2018; Kim et al., 2006; Hwang et al., 2015; Cho and Cho). Conversely, oenocytoids and giant cells are the rarest and most heterogenous in appearance (Hillyer and Strand, 2014; Cho and Cho; deAzambuja et al., 1991; Sonawane and Sonawane, 2017). Furthermore, the distinguishing small size and homogeneity of C. lectularius prohemocytes (3.5–4 µm) was similar to that of prohemocytes in mosquitoes and C. hemipterus (Hillyer and Strand, 2014; Sonawane and Sonawane, 2017). Notably, we did not observe thrombocytoids, which may be an artifact of the methods used in previous studies of C. hemipterus (Sonawane and Sonawane, 2017).

Unlike prior studies, our work quantitatively examined the physical features of the heterogenous circulating hemocyte populations of both male and female bed bugs. In an unstimulated state, the granularity of female hemocytes is significantly higher than that of male hemocytes. This sex-dependent variation could be due to a number of reasons. First, females may have a higher proportion of circulating granulocytes than males. Alternatively, the granular cells of females may be similar in abundance but contain larger numbers of granules than those of males. While our flow cytometry assays were not able to distinguish between these two possibilities, both explanations are consistent with the need of females to mount a more robust immune response to traumatic insemination than males (Reinhardt et al., 2015; Reinhardt et al., 2005; Reinhardt et al., 2003)

The responses of circulating bed bug hemocytes following bacterial exposure were also consistent with previous studies of other insects. For example, our flow cytometry analysis revealed a decrease in cell granularity following E. coli exposure and recent microscopy analyses of the hemocytes of Periplaneta americana showed that granulocytes undergo degranulation in response to fungal exposure (Mudoi et al., 2019). Similarly, in another member of the Hemiptera, R. prolixus, the vacuoles of circulating hemocytes undergo structural changes and the cells release effectors such as pro-phenoloxidase in response to challenge, which may manifest as shifts in granularity (Fruttero, 2016; Borges et al., 2008). The uptake of bacteria by a large proportion of circulating bed bug hemocytes in our studies is indicative of strong phagocytic capabilities in more than one cell type. Phagocytosis has been reported in diverse hemocyte types in many different insects (Hillyer and Strand, 2014; Oliver et al., 2011), including the tissueassociated hemocytes of C. lectularius, which phagocytose sperm (Reinhardt and Siva-Jothy, 2007). Phagocytic activity was also previously documented in R. prolixus hemocytes following bacterial challenge but was restricted to plasmatocytes (Borges et al., 2008). The hemocytes of bed bugs may differ from those of triatomines in this regard.

The increase in hemocyte size upon stimulation with E. coli is a classical marker of immune cell activation in vertebrates and has been transiently observed by flow cytometry in Anopheles mosquitoes post blood-feeding (Bryant and Michel, 2014; Bryant and Michel, 2016). It may also be an indicator of hemocyte differentiation (Bryant and Michel, 2016; Smith et al., 2015; Rodrigues et al., 2010) or cell division. Although we did not directly visualize dividing hemocytes, staining of hemocytes with the nucleic acid indicator propidium iodide revealed a population structure indicative of an active cell cycle, as did the incorporation of BrdU. We interpret these results to mean that a small proportion of bed bug hemocytes proliferate under basal conditions and in response to infection, confirming previous reports (Wilson and Siva-Jothy, 2014), but it remains unknown if hemocytes are undergoing symmetric division, asymmetric division, or change in ploidy without mitosis (Borges et al., 2008; King and Hillyer, 2013; Bryant and Michel, 2014). While the ability of hemocytes to proliferate in adult insects is debated, it has been documented in numerous insect species across different orders, including R. prolixus (King and Hillyer, 2013; Ghosh et al., 2018; Anderl et al., 2016; Ryan and Karp, 1993; Jones, 1967; Kwon et al., 2014). Microscopy observations in C. hemipterus were also suggestive of the existence of transitional forms between prohemocytes and plasmatocytes (Sonawane and Sonawane, 2017). Therefore, the details and biological importance of hemocyte differentiation and proliferation in bed bugs should be investigated further through the use of techniques such as anti-tubulin staining (King and Hillyer, 2013).

Lastly, RNAseq documented the expression of canonical immune signaling pathway components and antimicrobial effectors in circulating hemocytes. Further analyses also identified numerous transcripts that may potentially be upregulated in response to bacterial challenge. While most of these transcripts mapped to annotated, protein coding genes in bed bug genome (Benoit et al., 2016), many mapped to uncharacterized protein coding genes. Surprisingly, genes encoding immunity-related proteins such as antimicrobial peptides that are commonly expressed by hemocytes of other invertebrate species (Bartholomay et al., 2004; Xie, 2014; Smith et al., 2019) did not make our list of potentially upregulated genes based on normalized expression values and fold changes. This omission could be due to several factors, one being that these genes are only slightly upregulated in response to bacterial challenge or are regulated post-transcriptionally. For example, two defensin-like genes (Benoit et al., 2016) (LOC10661791 and LOC106661792) were expressed in both control and treated hemocytes. However, with normalized expression values of 98 to 124 and 53 to 81 respectively, these did not meet the 2.5-fold-change cutoff. Similar situations were noted for a prophenoloxidaselike gene (LOC106673939), diptericin-like gene (LOC106664366), and lysozyme (LOC106663700), which is a critical component of the humoral immune response in C. lectularius (Siva-Jothy et al., 2019; Bellinvia et al., 2020). The fold changes we observed in immune effectors are generally consistent with those observed upon injection of bacteria or feeding of NF-kB inhibitors in R. prolixus (Salcedo-Porras et al., 2019; Vieira et al., 2018). Nonetheless, it is also possible that these genes require longer periods of bacterial challenge to upregulate to higher levels in bed bugs.

Several genes involved in metabolism appeared on the list of potentially upregulated genes, including a trehalose transporter, glycogen debranching enzyme, and lipase. The potential upregulation of such genes in stimulated hemocytes is consistent with metabolic shifts that are known to occur concomitant with immune responses across a range of animal taxa, including insects (DiAngelo et al., 2009; Chambers et al., 2012; Dionne et al., 2006; Pietri et al., 2016). In fact, Drosophila hemocytes upregulate aerobic glyoclysis to meet the energetic demands of responding to pathogenic challenges (Krejcova et al., 2019). Although largely descriptive, the transcriptomic data reported here extend the limited body of information that is available regarding global gene expression responses to external stimuli in bed bugs (Mamidala et al., 2012; Narain et al., 2015; Koganemaru et al., 2013) and provide a useful resource for the community. In the future, it will be critical to verify our results and determine if the genes we identified contribute to bed bug immunity through targeted analyses of gene expression and function (e.g. qRT-PCR, RNAi) in the context of infection.

Now that baseline data on bed bug hemocytes and their responses to a common bacterium are available, additional studies using our methodology may shed light on important facets of bed bug immunology and control. For example, examining hemocyte responses to human pathogens that bed bugs may encounter in nature (e.g. Rickettsia, Borrelia, Bartonella) could provide insight on the factors that influence vector competence, or lack thereof (Delaunay et al., 2011; Mediannikov, 2013; El-Hamzaoui et al., 2019; Colaud et al., 2014). Our finding that bed bug hemocytes do not express Caspar is particularly intriguing, as loss of function Drosophila mutants exhibit increased resistance to bacterial infection, suggesting that perhaps a lack of Caspar expression contributes to pathogen resistance in bed bugs (Kim et al., 2006). Further, in the context of infection with bacterial entomopathogens of bed bugs previously identified by our group (Pietri and Liang, 2018), inhibiting specific hemocyte responses such as phagocytosis and DNA replication, or the expression of immunoresponsive genes, could be investigated as a means to improve the efficacy of microbial control agents. As a limitation of the analyses presented here is that they were all conducted on heterogenous cell populations, efforts should be directed at optimizing sorting and separation techniques to individually analyze the functions of the distinct hemocyte types. Investigating additional responses that were beyond the scope of the present study, such as aggregation, encapsulation, and nodulation using specific assays is also a priority (Fruttero, 2016).