We divided 1,800 crickets into 8 groups and conducted the experiments with 75 crickets per group, 3 repetitions (8x75x3 = 1,800 crickets) ( Figure 1 ). In this study, immune priming was induced using two groups, the PBS-primed group and the heat-killed Bacillus thuringiensis (Bt)-primed group. Two days later, live Bt was injected again to create two groups ( Figure 1 ). Subsequently, the immune priming effect was assessed by examining survival rates, the immune activation of granulocytes, and conducting RNA-Seq analysis for both the groups. Next, three groups were primed with PBS, heat-killed Escherichia coli (E. coli), and heat-killed Bt. After 2 days, live Bt was injected into all groups, and survival rates were examined. Similarly, 9 days later, we injected live Bt into each primed group and assessed survival rates.

We collected 150 crickets from basement of a building or under rocks or log debris in mountain, mated them, and collected eggs. Female laid an average of 90 ± 10 eggs, and hatched young were reared in a sterilized incubator (MIR-553, Sanyo Electric Biomedical, Japan; 25 ± 1°C, 40 ± 10%, 16h light:8h dark) maintain a constant environment. We used second- and third-generation adults as experiment insects. We provided sterilized food (Chinese cabbage and wheat husk purchased from Milwormnara Ltd., Korea) and always maintained cleanliness. Healthy crickets consistently exhibit activity through antennae movement, while sick or weak crickets demonstrate diminished antennae mobility and a notable decline in overall activity levels. Therefore, when such crickets are found in a breeding cage, we immediately remove them.

The bacteria strains utilized in in this study mentioned were Bacillus thuringiensis (Bt) strain KACC 12061 and Escherichia coli (E. coli) strain KACC16628, obtained from the Korean Agricultural Culture Collection (KACC). Briefly, the B. thuringiensis and E. coli strains originated from a single frozen glycerol stock, which was streaked on Luria-Bertani Broth plates (25 g.L^-1 of LB base, Invitrogen 12795–027 supplemented with 15 g.L^-1 of agar-agar, Fisher BioReagents, BP1423–2) and NA Broth (Beef extract 3gL, Peptone 5gL, Difco/BD, 234000) plates at 37°C and 25°C overnight. Subsequently, 100 µL of the E. coli culture was inoculated into 5 mL of fresh LB broth and incubated under the same conditions until reaching an optical density (OD) of 1.3 at 600 nm, as determined by an Ultrospec 10 (Amersham Biosciences), corresponding to 1.04x10⁹ cfu/mL. The bacterial pellet obtained by centrifuging 5 mL of the culture (2 minutes, 4°C, 13,000 g) was resuspended in 1 mL of PBS buffer to achieve an OD of 1.3 for injection. Similarly, 100 µL of the B. thuringiensis culture was inoculated into 3 mL of fresh NA broth and incubated under the same conditions until reaching an OD of 1.3 at 600 nm, as determined by an Ultrospec 10 (Amersham Biosciences), corresponding to 1x10⁹ cfu/mL. After centrifugation of 5 mL of the culture (2 min, 4°C, 13,000 g), the bacterial pellet was resuspended in 1 mL of PBS buffer to achieve an OD of 1.3 for injection. The LD₅₀ (Lethal Dose, 50%) value was determined at an OD of 1.3 for G. bimaculatus and measured over a 96-hour period, with a dosage of 10 µL (1x10⁷ cfu/insect for Bt and 1.04x10⁷ cfu/insect for E. coli).

All crickets were selected 24 hours after molting from young to adult. Additionally, we removed inactive or unhealthy individuals (less than ~2%) after the first injection and used only healthy individuals that were very active and feeding for the injection of live Bt. In all groups, the crickets were almost identical in age (24 hours after adult molting), health (active movement and no infection), average size (~25 ± 5 mm per cricket), and average weight (~1.1 ± 0.5g per cricket). Immune priming was performed by diluting the OD 1.3 obtained from the LD₅₀ value at a ratio of 1:999, and 2μl was inoculated (2x10³ cfu/insect for Bt and 2.08x10³ cfu/insect for E. coli). The reason for this is similar to preventive vaccination, where we aimed to ensure that there was no impact on survival due to the preventive inoculation. Therefore, we obtained values over the entire observation period of 96 hours where there was no difference between the control group and the survival group. Next, 48 hours after the immune cycle of crickets, all individuals were injected with 10μl of live pathogen at the concentration described in material and method section 2.3. Subsequently, survival analysis and microscopic examinations were conducted at different time intervals. For injection, the crickets were sterilized with 70% alcohol paper, and a finely pulled glass needle was shallowly inserted into the abdomen (Haematocrit capillaries, Sigma-Aldrich) (17, 18). The experiments were repeated three times independently.

Change in hemocyte count were conducted for group A-1 and B-1. Hemolymph was collected from each group by extracting 50 µL from each cricket. Subsequently, a total of 500 µL of hemolymph from 10 crickets was placed into a 1.5 mL microcentrifuge tube (Bioneer Ltd., Korea) containing 500 µL of coagulant solution (98 mM NaOH, 186 mM NaCl, 17 mM EDTA, and 41 mM citric acid, pH 4.5), followed by centrifugation (10 min at 4°C) to remove plasma (Beckman Inc., USA). Hemocytes (pellets) were washed three times with 1mL of sterile PBS buffer. To count total hemocyte (THC), we placed hemocytes on sterile disposal hemocyte meter slides (Ncubauer Improve, iNCYTO C-Chip DHC-N01; 10µL) and count them in four squares using a light microscope (Leica^® DMI 3000M; 40x objectives) and calculated as absolute numbers. The hemocyte count were calculated at 0, 12, 24, and 48 h post-infection.

For analysis using an optical microscope (Leica^® DM2500 and Leica^® DMI 3000B), 1,000 hemocytes per group were photographed and hemocytes image database for each group was constructed (LMD application software version 6.1, 2048x1536 pixels). We used this image data to classify hemocytes by size and shape (21–24). To analyze lysosome activation in granulocytes, we stained the hemocytes collected from each group using the acid affinity dye LysoTracker Red (7.5nM, Molecular Probes; Thermo Scientific, Waltham, USA) at room temperature for 30min, washed three times with PBS, and analyzed by a Leica DMI 3000B fluorescent microscope and flow cytometry (BD™ FACS Canto; BD Bioscience, San Jose, CA) (24). Red fluorescence was detected by channel FL3 (610/20 band-pass), 10,000 hemocytes were read per sample, and statistical processing was performed using BD Bioscience FASC Diva software (24). We provided the flow cytometry gating strategy as a Supplementary File ( Supplementary File S1 ). For scanning electron microscopy (SEM), hemocytes for each group were sedimented on Thermanox™ 125 coverslip coated with 0.1% poly-L-lysine (Thermo Fisher Scientific, Waltham, USA) for 3 days. Hemocytes attached to the cover slip were fixed using sodium cacodylate buffer (Na (CH3)₂AsO₂, 0.1M, pH 6.5; 50mM isopropanol 15%, Sigma-Aldrich) containing 2.5% glutaraldehyde. The coverslip was fixed again with 1% osmium tetroxide (OsO₄) and were dehydrated by gradual immersion in ethanol (70%~100%) and subjected to critical point drying (Leica^® EM CPD030). The coverslip was sputtered with gold and finally observed with an SEM Leo 420 scanning electron microscope (LEO Electron Microscopy Ltd., Cambridge, UK). For transmission electron microscope (TEM), hemocytes were fixed with 1% osmium tetroxide (OsO₄) and then the dehydrated using ethanol (70%~100%) and 100% propylene oxide. Hemocytes were embedded in araldite resin (Epon 812, Japan) and polymerization reaction was performed at 60°C for 48 hours. Hemocytes were ultrathin sectioned using an ultramicrotome (Pabisch Top Ultra 150) and collected on a 200mesh nickel grid. Ultrathin sections were stained with uranyl acetate, precipitated in citrate for 5min, and observed using a JEX-1230 transmission electron microscope (JEOL Ltd, Japan).

To perform RNA sequencing, we extracted fatbodies from group A, B, A-1, and B-1. We selected 10 crickets from each group and prepared them in one tube for each group. Since each group was repeated 3 times, fatbodies were extracted from a total of 30 crickets and prepared in 3 tubes per group (12 tubes in total). Total RNA was extracted using the RNeasy Plus Mini Kit (Qiagen, USA) (all operations were performed using sterile spaces and tolls to minimize contamination). To briefly summarize the procedures, 100mg of fat bodies were mixed with 1ml of TRIzol reagent (Thermo Fisher Scientific, Waltham, USA). Plastic beads of 0.6 mm, 2 mm, and 7 mm were added to the mixture, and it was homogenized for 10 minutes with a vortex (DAIHAN Ltd., Korea). The sample were centrifuged at 14,000 rpm for 3 min at 4°C and 1mL of supernatant was extracted. The extracted supernatant was mixed with 200 µL chloroform and reacted at 4°C for 20 min (Thermo Fisher Scientific, Waltham, USA). The sample were centrifuged at 14,000 rpm for 3 min at 4°C for 15min and then the supernatant (~600 µL) was loaded into a gDNA elimination column to remove gDNA (Thermo Fisher Scientific, Waltham, USA). The sample was centrifuged at 14,000 rpm for 1 min at 4°C and then total RNA was precipitated by adding 600 µL of cold isopropanol/70% ethanol (maintained at -20°C, Sigma-Aldrich, USA). The sample was centrifuged at 14,000 rpm for 1min at 4°C using RNeasy Plus Mini column, mixed with ~700 µL of RW1 buffer (10mM Tris-Hcl, 1mM EDTA, 50mM NaCl, 0.5% Tween-20, pH 8.5, Sigma-Aldrich, USA), and centrifuged at 14,000rpm for 1min. Finally, the sample was mixed with 500 RPE buffer (10mM Tris-Hcl, 80% ethanol, 0.5mM EDTA pH 7.5), centrifuged, and total RNA was extracted by placing ~50 µL RNeasy-free water in a new sterilized tube (Sigma-Aldrich, USA).

The cDNA library was generated from the total RNA, and paired-end sequencing was performed using the Illumina RNA-seq technology (Macrogen Ltd, Korea). Raw sequencing data was preprocessed using integrated software for adapter and quality trimming, Trim Galore v0.6.5. Reads with a Phred33 score of less than 20 and a length of less than 20 base pairs were filtered out, and the standard Illumina adapter sequence was trimmed if found. De novo full-length transcriptome reconstruction was performed by a software to reconstruct full-length transcriptome without a reference genome, Trinity v2.15.1, and ORF (open reading frame) prediction was performed by software to identified candidate coding region from de novo transcriptome assembly, TransDecoder v5.5.0 (25). Predicted ORFs were assessed using homology search (BLAST), protein domain identification (PFAM), gene ontology assignment, and functional classification (e.g., eggnog) of sequence data using software to annotate de novo transcriptome assembly, Trinotata v4.0.2 (26). Quantification of gene expression was performed using software to quantify expression of genes and transcripts from transcriptomic data, Salmon v1.10.2 (27).

The gene expression result was normalized and transformed using a function provided by DESeq2 v1.36.0, Variance Stabilizing Transformation (VST). Principal Component Analysis (PCA) was conducted on the transformed gene expression data using the prcomp function from the R core package, stats v3.6.2. Differentially expressed genes (DEGs) were defined using DESeq2 v1.36.0 with an adjusted P-value < 0.05 and |fold change| > 1.5 (28). Gene Ontology (GO) terms highlighted in significant DEGs were accessed using ClusterProfiler v4.4. PCA plots, Volcano plots, bar plots, box plots, dot plots, and heatmaps were generated using ggplot2 v3.3.6 and ComplexHeatmap v2.13.1. All statistical analyses and visualizations were performed in R v4.2.1 and the R Studio environment (29). Biological Process (BP) terms were visualized using the emapplot function from the ClusterProfiler package to show correlations between them. Each dot represents significantly representative BP terms, and the size of each dot reflects the number of genes associated with the corresponding BP term that significantly increased or decreased in expression. The color of each dot in the visualization indicates the corrected p-value of the corresponding BP terms. The distance between dots reflects the calculated similarity or dissimilarity between BP terms, considering the gene content associated with each term. Closer dots indicate higher similarity, while dots further apart suggest greater dissimilarity. BP terms are grouped based on the calculated distances. Terms with higher similarity are grouped together, forming clusters or groups.

The experiments were repeated three times independently. We performed all statistical analyses with R, version 3.6.3 (30). The protection of immune priming, reflected by the frequency of living animals following a challenge at the LD₅₀, was tested with a global mixed-effects Cox proportional hazard regression model (coxme packages) (30). Comparisons on survival between two conditions is presented as a hazard ratio (HR) that scores survival rate of a test group against survival in a referent group. A HR is reported with its 95% confidence interval and a Wald test p-value with Benjamini-Hochberg correction for multiple comparisons reporting whether the HR significantly deviates from 1. For hemocyte count and analysis, Student’s t-test was used to compare differences between groups in flow cytometry results. The experiment was repeated three times. Error bars indicate Mean ± SEM. *P < 0.05 (t-test). NS indicates not significant.

The datasets generated and analyzed during this study have been uploaded to the GEO database (GSE266080).

For the purpose of conducting a comparative analysis of survival rates, two groups were initially administered PBS buffer and heat-killed Bacillus thuringiensis (Bt), respectively. Following a 2-day interval, all groups were subsequently exposed to live Bt ( Figure 2A ; Conducted in triplicate with 75 crickets per experiment). All crickets were of nearly identical age (24 hours post-adult molting), exhibited good health (active movement and absence of infection), possessed an average size of approximately 25 ± 5mm per cricket, and an average weight of approximately 1.1 ± 0.5g per cricket.

In the PBS-primed group, crickets showed increased mortality post-live Bt injection, with survival rates dropping to less than 10% beyond 60 hours. The data indicates a noteworthy decline in survival rates over time, with a substantial proportion of crickets failing to survive beyond 60 hours. In contrast, the heat-killed Bt-primed group exhibited a remarkable 58 crickets surviving after 60 hours, yielding an approximate 80% survival rate. The higher survival rate observed in the heat-killed Bt-primed group is attributed to the potential induction or maintenance of immune activation following the initial injection of heat-killed Bt. This priming effect is believed to contribute to an augmented and swift immune response upon subsequent exposure to live Bt.

After the injection of live Bt, hemocyte types and numbers were examined in each group ( Figures 2B ). In untreated crickets, the average hemocyte count is ~150/10µL hemolymph. The PBS-primed group displayed an average of ~250/10µL hemolymph after 12h, while the heat-killed Bt-primed group exhibited an average of ~650/10µL hemolymph ( Figure 2B ). Specifically, the PBS-primed group had an average of ~90 granulocytes, while the heat-killed Bt-primed group had ~330 granulocytes ( Figure 2B ). There were no significant differences in plasmatocytes. In the heat-killed Bt-primed group, elevated levels of prohemocytes, coagulocytes, and spherulocytes were observed compared to the PBS-primed group ( Figure 2B ). These findings suggest that the initial injection of heat-killed Bt may have induced or sustained cellular immune activation, leading to elevated hemocyte counts, particularly granulocytes.

We investigated the correlation between the higher survival rate of the heat-killed Bt-primed group and the morphological changes and degree of lysosome activation in granulocytes. After encountering pathogens, the cricket granulocytes exhibit rapid morphological changes in their cell membrane (forming extra-cellular traps (ECTs); fan- or pod-like structures) and eliminate pathogens through phagocytosis, encapsulation, and nodulation (18). Therefore, we observed the formation for ECTs in granulocytes and quantified activation of lysosomes following phagocytosis using the fluorescence microscopy and fluorescence-activated cell sorting (FACS) ( Figure 3 ). First, we observed the degree of immune activation of granulocytes in group A (PBS buffer injection) and B (heat-killed Bt injection) over time (intervals of 4h, 12h, and 24h) ( Figures 3A-1~A-7, 3B-1~B-7 ). Figure 3A-1 illustrates the typical morphology of granulocytes 4 hours after the injection of PBS buffer. ECTs and pods (hair-like structures) related to immune activity were not observed on the cell membrane ( Figure 3A-1 ). Most granulocytes in group B were observed to be relatively larger in size, but similar to granulocytes in group A, no other morphological changes were observed ( Figure 3B-1 ). We collected hemocytes and examined lysosome activation at 4 hours, 12 hours, and 24 hours ( Figures 3A-2 ~ A-4, 3B-2 ~ B-4 ). In groups A and B, lysosomal activity began to be observed in granulocytes at 4 h post-injection, and we did not observe any differences in lysosomal activity at 12- and 24-h post-injection. To quantify the red-fluorescent signals of groups A and B, hemocytes were analyzed by flow cytometry at 4h, 12h, and 48 h post-injection ( Figures 3A-5~A-7, 3B-5~B-7 ). The average red fluorescent of group A was ~6.97% and that of group B was ~8.18%. The flow cytometry analysis was repeated three times ( Figures 3C-1~C-3 ). Based on the results, it appears that despite observing changes in the size of granulocytes, the statistical analysis or interpretation of the FACS results for granulocyte immunity did not reveal a significant difference between group A (injected with PBS) and group B (injected with heat-killed Bt).

Next, we performed a subsequent injection of live Bt into both groups A (previously injected with PBS) and B (previously injected with heat-killed Bt) two days after the initial injections (designated as group A-1 and B-1). To investigate morphological changes in granulocytes, we collected hemocytes at different time points (4 hours, 12 hours, and 24 hours) after the injection of live Bt ( Figures 3A-8~A-14, B-8~B-14 ). At 4 h post-injection, most granulocytes in group A-1 appeared to be relatively larger in size and some pods began to form ( Figure 3A-8 ; indicated by white arrows). On the other hand, granulocytes in group B-1 exhibited several very long pods (indicated by white arrow) and ECTs were also widely formed along the cell membrane, represented by yellow arrows ( Figure 3B-8 ). To investigate the activation of lysosomes, we also collected hemocytes at different time points (4 hours, 12 hours, and 24 hours) after the injection of live Bt. At 4 h post-injection, lysosomal activation was observed in both groups A-1 and B-1. Subsequently, at 12 h post-injection, a higher level of lysosomal activation was observed in group B-1 compared to group A-1. By the 24 h post-injection, group A-1 exhibited reduced lysosomal activation, while group B-1 maintained sustained lysosomal activation ( Figures 3A-9~A-11, B-9~B-11 ). To quantify the red-fluorescent signals of groups A-1 and B-1, hemocytes were analyzed by flow cytometry at 4h, 12h, and 48 h post-injection. In groups A-1 and B-1, lysosomal activity was first observed in granulocytes at 4 h post-injection and peaked at 12 hours ( Figures 3A-12, A-13, B-12, B-13 ). At 24 h post-injection, lysosomal staining in the A-1 and B-1 groups was measured to be 13.84% and 17.85%, respectively.

After injection of PBS buffer or heat-killed Bt, the morphology of granulocytes at 4 hours is nearly identical to wild-type granulocytes (data not shown). Next, we performed a subsequent injection of live Bt into both the PBS-primed group and the heat-killed Bt-primed group two days after the initial injections. To investigate morphological changes, we collected granulocytes at different time points (4 hours, 12 hours, and 24 hours) after the injection of live Bt ( Figures 4A-1~A-3, B-1~B-3 ). In the PBS-primed group, 4 hours after injection, pod structures (indicated by a white arrow) began to appear on the cell membrane, and ECTs started to form ( Figure 4A-1 ; ECTs surrounded by yellow lines). After 12 hours, ECTs were extensively formed around the cell membrane of most granulocytes, reaching the peak of cellular immune activation ( Figure 4A-2 ). After 24 hours, ECTs that had formed in granulocytes disappeared and were faintly observed ( Figure 4A-3 ; ECTs losing their shape indicated by white arrows). As seen in Figure 3A-11 , 24 hours after the live Bt injection, most granulocytes were observed to be unhealthy ( Figure 4A-3 ). In the heat-killed BT-primed group, pod structures and ECTs had already begun to actively form on the granulocyte cell membrane after 4 hours ( Figure 4B-1 ). By 12 hours after injection, the cell membranes of most granulocytes were completely transformed, and very large ECTs could be observed ( Figure 4B-2 ). Additionally, in contrast to the PBS-primed group, most granulocytes were observed to maintain ECTs and continue immune activity after 24 hours ( Figure 4B-3 ). The above results indicate that the heat-killed BT-primed group exhibited a faster cytological immune response to the re-injected activated pathogen (live Bt) and simultaneously, the duration of immunity was prolonged in the priming group.

Sequentially, we used TEM to observe the cytoplasmic features of granulocytes between groups ( Figures 4C-1 ~ 4C-3, 4D-1 ~ 4D-3 ). The nucleus of the cell was labeled with a green ‘N’ and the nuclear membrane was outlined with a gray line around the nucleus. The yellow arrows indicate black granules of various sizes present in granulocytes, while the blue arrows are believed to represent phagosomes involved in the phagocytosis of Bt. Red arrows are considered to indicate early endosomes that occur after phagosome formation or lysosomes that have fused with the phagosome. First of all, the main difference observed between granulocytes in groups is the notable presence of black granules of various sizes in heat-killed Bt-primed group ( Figures 4D-1 ~ D-3 ). These granules, recognized for containing various immune-related substances in mammalian white blood cells, were found in significant numbers within primed group (31). As indicated by the SEM results, morphological changes in the granulocytes of both groups were not observed solely after the initial injection (> 4 hours). However, a noteworthy finding was the significant production of granules within the cytoplasm of granulocytes in the heat-killed Bt-primed group (data not shown). In both groups, live Bt was reinjected and hemocytes were subsequently collected and photographed at intervals of 4 hours, 12 hours, and 24 hours, respectively. After 4 hours, phagosomes containing Bt fragments began to be observed in the granulocytes of both groups ( Figures 4C-1, D-1 ; indicated by blue arrows). Additionally, large phagosomes were also observed, suggesting their growth through fusion with each other and potentially combining with lysosomes (indicated by red arrows). After 12 hours, large lysosomes were observable in the cytoplasm of granulocytes in both groups. Furthermore, a substantial number of black granules were consistently observed in the granulocytes of the heat-killed Bt-primed group ( Figure 4D-2 ). After 24 hours, the cytoplasm of granulocytes in the PBS-primed group was observed to be filled with small-sized transparent granules, indicating cytoplasmic vacuolization (represented by purple arrows). Additionally, irregular disruption of the cell membrane was noted ( Figure 4C-3 ). In contrast, the granulocyte cytoplasm of the heat-killed Bt-primed group consistently showed considerable-sized phagosomes (represented by blue arrows), large lysosome granules (represented by red arrow), and black granules of various sizes (represented by yellow arrows). Furthermore, pod-like structures were consistently observed in the granulocyte cytoplasm of the heat-killed Bt-primed ( Figure 4D-3 ; represented by white arrow). In summary, the observed results indicate that granulocytes in the heat-killed Bt-primed group, when compared to PBS-primed group, maintained a healthy state and exhibited highly active immune activities even after 24 hours.

We conducted an analysis of the levels of gene activity to identify patterns and differences associated with the observed variations in survival rates and granulocyte activation among the experimental groups. To visualize the transcriptomic distances among the samples, dimension reduction using principal component analysis (PCA) was performed. In the 2-dimensional plot comprising the two most informative components, principal component 1 (PC1, explaining 47.9% of variance) and PC2 (explaining 12.9% of variance), the distances among the groups are clearly larger than the distances among samples within the same group ( Supplementary Figure S1A ). The samples in group A, injected only with PBS buffer, were positioned far from the samples in the other three groups. This indicates that the expression of genes in the samples injected with Bt differs significantly from those injected with PBS buffer only. To identify genes showing significant differences in their expressions between groups, we conducted differential gene expression (DEG) analysis and visualized the results using a volcano plot ( Supplementary Figures S1B-D ). Overall differences were observed in the gene expression patterns of each group, and these variances are believed to have influenced both the survival rate and the degree of cellular immune activation ( Supplementary Figure S1 ).

Next, we visualized the relationship between the top four Gene Ontology (GO) terms based on significance. Specifically, in the comparison between group A-1 and group B-1, among the significant GO terms identified in group B-1, GO terms 1, 2, and 3 were grouped together based on similar functions. The identified genes are complex, encompassing diverse metabolic and signaling pathways, including those related to immunity. This underscores the importance of maintaining a balanced and regulated cellular environment to ensure homeostasis in energy metabolism and lipid localization ( Figures 5A-1 ). Additionally, GO term 4 was associated with genes directly involved in insect innate immunity, such as Toll pathways, pattern recognition receptors (PRRs), and cytokines. Toll pathways, PRRs, and cytokines act as integral components of the insect immune system, working together to detect, signal, and coordinate immune responses against pathogens. Their intricate relationships form a robust defense mechanism critical for maintaining immune homeostasis. Down-regulated GO terms in group B-1 were primarily associated with muscle-related genes ( Figures 5B, B-1 ). Comparing group A to group B, overexpressed GO terms in group B were associated with genes related to immune-related cell proliferation, differentiation, and developmental processes ( Supplementary Figures 2SA, 2SA-1 ). Down-regulated GO terms were primarily linked to pigment and metabolism ( Supplementary Figures 2SB, 2SB-1 ). When comparing group B to group B-1, the GO terms mainly identified were related to melanin synthesis. Down-regulated GO terms in this comparison were associated with muscle-related genes ( Supplementary Figures 2SC, 2SC-1, 2SD, 2SD-1 ).

Utilizing ‘bacteria’ as a keyword in the Dotplot analysis, overexpressed GO terms in group B-1 compared to group A-1 were confirmed to be related to antibacterial humoral response, detection of bacterial lipopeptides, and cellular response to bacteria lipopeptides ( Figure 6A ). Notably, antibacterial humoral response (GO:0019731)-related genes were significantly upregulated in group A-1, B, and B-1 compared with group A ( Figure 6B ). Normalization count analysis further confirmed high expression levels of attacin, defensin, lysozyme, and GNBP-like genes exclusively in group B-1 ( Figure 6C ). Comparing the results, specifically between groups A and B, it was determined that the expression of genes inducing cell morphological changes commenced in group B. Additionally, when comparing A-1 and B-1, it was observed that group B-1 exhibited overexpression of various genes related to both humoral and cellular immunity.

We analyzed the priming effect and survival when the pathogen used for immunity induction differed from the reinvading pathogen. Group A, B, and C were injected with heat-killed Bt, heat-killed E. coli, and PBS buffer, respectively. Subsequently, all groups were re-injected with live Bt to assess survival rates. Survival rates were evaluated for two- and nine-day intervals between injections. Two days later, in the control group C injected with live Bt, the survival rate was approximately 10% ( Figure 7A ). After 60 hours, the survival rate for groups A and B was approximately 60%, and the surviving crickets remained highly active and healthy ( Figure 7A ). Next, live Bt was injected 9 days after the first injection to examine survival rates among groups ( Figure 7B ). Surprisingly, while group C and B exhibited low survival rates of ~10% until 60 hours after injection, group A demonstrated a high survival rate of ~60% after 60 hours ( Figure 7B ). To confirm the results in Figure 7 , the survival survey was replicated three times ( Supplementary Figure S3 ). During short-term immune priming, there is an observed elevated immune response to both homologous and heterogeneous pathogens (indicating a cross-priming effect). However, it is presumed that crickets exhibit an enhanced immune response exclusively to the homologous pathogen used during long-term immune priming.