Differential gene expression during recall of behaviorally conditioned immune enhancement in rats: a pilot study

Animals

Male, Sprague Dawley (SD) rats (5-7 weeks, 160-250 g) were obtained from InVivos Pte Ltd, Singapore. Male animals were chosen to minimize the impact of hormone-related gene expression and behavioural readouts. Animals were housed at the Biological Resource Centre (BRC) in groups of two in individually ventilated cages and maintained on a 12h light/dark cycle with food and water ad libitum in accordance with the guidelines of the Agency for Science, Technology and Research (A*STAR) Animal Care and Use Committee. All efforts were made to ameliorate any suffering of animals; animals were provided with Nylabones (hard, non-toxic nylon), nestlets, and/or domes for environmental enrichment to ensure adequate welfare and psychological well-being. In accordance with Institutional Animal Care and Use Committee (IACUC) guidelines, animals were allowed to acclimatize for a minimum of three days and examined to confirm their health status prior to study commencement.

General conditioning procedure and tissue collection

Animals (N=30) were randomly allocated into two series (a and b), each containing 15 animals, and further divided into three experimental groups as shown in Table 1. A 3-day conditioning paradigm was used in the study as described previously¹¹ where camphor smell was used as the conditioned stimulus (CS) and poly I:C injection (i.p.) as the unconditioned stimulus (US). In the present study, camphor smell exposure was performed in a separate procedure room for 1h using 10 cm petri dishes filled with 5 g of camphor powder (Sigma #21310) placed above the cage (on top of the metal grid) and warmed up with a heat lamp. On day D0, animals from all three groups were exposed to camphor smell for 1h following exposure to saline (i.p.; 10 ml/kg; negative control group) or poly I:C (Sigma #P1530; i.p.; 1 mg/kg; test and positive control group).¹⁶ On day D2, animals from the test and negative control groups received saline injections (i.p.), following exposure to camphor smell for 1h, whereas animals from the positive control group were re-injected with poly I:C, only. After the last injection on day D2, animals from each group were sacrificed at 0h (n=2), +3h (n=2), +6h (n=2), +24h (n=2) and +48h (n=2) postinjection, and tissues were collected (hypothalamus, pituitary, adrenal glands, and spleen) and stored snap frozen in RNAlater (MilliporeSigma, Burlington, U.S.A.) for further analysis. Blood was collected for the isolation of peripheral blood mononuclear cells (PBMCs) and plasma was collected to perform cytokine and chemokine Luminex analysis. Animal experiments were conducted at the Biological Resource Centre (BRC), Biomedical Sciences Institutes, Agency for Science, Technology and Research (A*STAR), 20 Biopolis Way, #07-01 Centros Building, Singapore 138668.

Table 1. Animal groups and conditioning protocol including timeline.

Group number	Group name	1st treatment (Day D0)	2nd treatment (Day D2)
1	Test (n=5)	Camphor smell + Poly I:C	Camphor smell + Saline
2	Positive control (n=5)	Camphor smell + Poly I:C	Poly I:C
3	Negative control (n=5)	Camphor smell + Saline	Camphor smell + Saline

Sucrose preference test (SPT)

The SPT is a standard behavioural test to assess depression-like behaviour in rodents during the development of anti-depressive treatments. The SPT is based on anhedonia (lack of interest in rewarding stimuli), which is present in some forms of affective disorders including depression. In this test, the interest of animals in seeking out a sweet rewarding drink relative to plain drinking water is used and a bias toward the sweetened drink is typical, failure to do so is an indication of anhedonia/depression.¹⁷ The SPT was carried out on D-7 to D-1 (run-in phase for baseline data) as well as days D0 - D3. Rats were presented with two drinking bottles, one with plain drinking water and one with 2% (w/v) sucrose solution for 24 h. Except for days D2 and D3, all 30 animals were analyzed; due to sacrifize after treatment on D2, only 12 animals were analyzed on D2 and 6 animals on D3. The ratio of sweetened vs. plain water consumed during this period was calculated, and results were expressed as %sucrose preference.

Affymetrix microarray analysis

At the end of the study, mRNA from all tissues harvested at different time points was isolated by using an RNeasy Plus mini kit (QIAGEN, Hilden, Germany). RNA quality was assessed by using a Bioanalyzer 2100 (Agilent Technologies, Santa Clara, U.S.A.). Highly purified RNA was subjected to microarray analysis with a GeneChip ^TM Rat Gene 2.0 ST Array (Affymetrix, Inc, Santa Clara, U.S.A.) to identify differential gene expression during the recall of behaviourally conditioned immune responses. RNA labelling, hybridization, staining, and scanning were performed according to the manufacturer’s instructions. Briefly, 100 ng of total RNA from each sample was reverse transcribed to cDNA, followed by overnight in vitro transcription to generate cRNA which was further fragmented and labelled. The quality of cDNA and fragmented cRNA was assessed with an Agilent bioanalyzer. We used the Robust Multichip Average (RMA) model for array background correction and quantile normalization as described previously.^18–20 Probes were mapped to the Rattus norvegicus genome chip (RGSC 5.0/rn5) and results from all tissues were normalized using β-actin.

Analysis of differentially expressed genes

Gene expression results for all test and positive control samples were matched by tissue and time point with their respective counterparts in the negative control group and subtracted to derive single point gene expression differences (all data expressed as 2^x). Due to insufficient RNA quality after extended storage, we could not analyse all tissue samples by microarray and had to discard multiple samples (hypothalamus negative control +24h, pituitary test and positive control 0h and +3h, adrenal test +6h and +24h and positive control 0h and +3h, spleen +24h and +48h all samples); accordingly, we first limited the analysis to the average across all available time points and then examined single-point regulations with a focus on genes related to neurotransmitters in greater detail; in each type of tissue, a minimum of three individual time points were available. Following gene enrichment and comparison of test vs. negative and positive vs. negative control selection, all differentially expressed genes per tissue were then analysed by Reactome pathway analysis.²¹ As we hypothesized the origin of the message post recall was from the hypothalamus, we additionally leveraged QIAGEN Ingenuity Pathway Analysis (QIAGEN IPA) for a more detailed analysis of hypothalamic genes. All genes with an average regulation of >2fold in the hypothalamus of test animals were selected, and a subcellular network analysis was conducted using an IPA graphic interface.

Quantitative RT-PCR

The same RNA of all samples as used in the microarray analysis was also subjected to quantitative reverse transcription PCR (qRT-PCR) to confirm the observed differential gene expression. The qPCR laboratory work was conducted as a fee-for-service by the Qiagen service laboratory (Qiagen Life Science Service & Support, Hilden, Germany), using the RT² Profiler^TM PCR custom Array RAT (CLAR27943). Briefly, RNA was isolated, and cDNA was synthesized with Qiagen RT² First Strand Kit using 500 ng of RNA, 1 μl RT primer mix (oligo-dT and random hexamer primers), 4 μl Quantscript RT buffer (5X) and 1 μl Quantiscript Reverse Transcriptase (QuantiTect kit, Cat. No. /ID: 205311, Qiagen). qPCR reactions were then assembled using synthesized cDNA (1 μl), 5 μl RT² Profiler PCR Array SYBR Green master mix (Qiagen, Germany), 1 μl of each primer diluted to a 5 μM working solution and 1 μl sterile water and processed in a 96-well format including housekeeping genes, RNA and DNA controls, using Qiagen’s proprietary primer panel. The mRNA expression was determined using the 2−^ΔΔCT method.²² All gene expression values were normalized using Hypoxanthine phosphoribosyl transferase (HPRT), glyceraldehyde phosphate dehydrogenase (GAPDH), and β-actin as references for expression analysis of genes of interest.

Cytokine array analysis

Plasma protein levels for all animals and time points were evaluated as a fee-for-service by Singapore Immunology Network (SIgN), coordinated by Biomedical Research Council (#04-06 Immunos, Singapore), using the MILLIPLEX^® MAP rat pituitary endocrine multiplex assay (RPTMAG-86K) for ACTH and BDNF and MILLIPLEX^® MAP Cytokine/Chemokine Panel (RECYMAG65K27PMX, both MerckMillipore, Darmstadt, Germany) to simultaneously analyze the levels of 27 further cytokines/chemokines. Plasma samples were diluted to a protein concentration of 2 mg/ml, then analysed on a MAGPIX system (Luminex, Austin, U.S.A.) and quantified using MILLIPLEX® Analyst 5.1 software. Cytokines/chemokine measured included RANTES, GRO KC CINC-1, VEGF, Fraktalkine, LIX, MIP-2, G-CSF, Eotaxin/CCL11, GM-CSF, IL-1α, Leptin, MIP-1α, IL-4, IL-5, IL-1β, IL-2, IL-6, EGF, IL-13, IL-10, IL-12p70, IL-5, IL-17A, IL-18, MCP-1, IP-10, KC, VEGF, Fractalkine, LIX, MIP-2 and TNF-α.

Statistical analyses

All statistical analyses were conducted using RMA gene expression data for all available time points for test, positive, and negative control, respectively, applying a two-tailed, unpaired, homoscedastic Student t-test (Microsoft^® Excel^® for Microsoft 365 MSO, version 2022). A p-value of <0.05 was considered significant.

Ethical approval

This study was approved under protocol number BRC IACUC #151001 by the IACUC of the Biological Resource Centre (BRC, Biomedical Sciences Institutes, Agency for Science, Technology and Research (A*STAR), 20 Biopolis Way, #07-01 Centros Building, Singapore 138668) on 26 Feb 2015. A specific procedure amendment for the use of the sucrose preference test was approved on 03 Feb 2016.

Gene expression analysis

The present pilot study was designed to identify differentially expressed genes along the HPA-axis (hypothalamus, pituitary, adrenal, and spleen) during recall of the behaviourally conditioned immune response, using a conditioning paradigm to explore efferent pathways.¹¹ As the exact time point of gene alteration was unknown, we first looked at genes with maximum average regulation across all time points (0h, +3h, +6h, +24h, and +48h post recall) in all three studied groups and then looked at neurotransmitter-related regulations in greater detail, also considering single point regulations. Each analysis per group and tissue included a minimum of three independent time points (for exact samples analysed, see Methods).

Leveraging this approach to identify major upregulated genes in test vs. negative control across all tissues with a >2fold average regulation across all time points, we identified 11 upregulated transcripts in both the hypothalamus and the pituitary (Figure 1). In the adrenal and spleen, however, the number of genes upregulated using this approach was significantly higher. To identify meaningful candidates in these tissues, we excluded transcripts with unknown identity and focused only on genes with strong upregulation during the first six hours post recall and/or p<0.05, comparing expression across all time points in the test group with their respective counterparts in the negative control group, thereby reducing the number of transcripts from the adrenal glands and spleen to 18 and 20, respectively. We also excluded genes with a stronger upregulation in positive vs. negative control than the test except when also showing a strong regulation across multiple other tissues, e.g. Cxcl11 or Cxcl13. A similar approach was then applied to identify the most promising candidate genes for downregulation (Figure 2).

Figure 1. Genes upregulated across the hypothalamus, pituitary, adrenal, and spleen; differential gene expression shown as avg. across all time points for test vs. negative control and positive vs. negative control, respectively.

All data shown as 2^x.

Figure 2. Genes downregulated across the hypothalamus, pituitary, adrenal, and spleen; differential gene expression shown as avg. across all time points for test vs. negative control and positive vs. negative control, respectively.

All data shown as 2^x.

In addition to this, we focused on the up- and downregulation of neurotransmitter-associated genes such as receptors, transporters, and solute carriers, potentially involved in dopaminergic, serotonergic, opioid, cholinergic, glutamatergic, or GABAergic neurotransmission. We then selected for strong regulation across multiple tissues, even if not crossing an average >2fold regulation across all time points analysed, and limited the time frame for this analysis to the first six hours post recall, unless otherwise mentioned.

Hypothalamus

In the hypothalamus, a total of 11 transcripts, detected by 12 probes, were upregulated >2fold when comparing the average regulation across all time points in the test group vs. negative control. Among these, Spp1, which codes for secreted immune modulator osteopontin (OPN), and Slc18a2, which codes for vesicular monoamine transporter 2 (VMAT2), demonstrated the strongest upregulation in the test group vs. negative control, without a comparable upregulation in the positive vs. negative control. Likewise, Otx2, which codes for orthodenticle homeobox 2 and was detected by two independent probes, also showed a strong upregulation in the test group vs. negative control without a strong upregulation in the positive vs. negative control as e.g. observed for Hapln4 or Cxcl11 (Figure 1).

In addition to Slc18a2, Otx2, and Spp1, 8 additional genes were found upregulated >2fold in the hypothalamus. Out of these genes, only Fzd6, Zic1, and Cbln1 exhibited a significantly higher level of upregulation in the test group vs. negative control. On the other hand, Fa2h, Tspan18, Sh2b2 and Hapln4 and Cxcl11 as already mentioned, either showed a weaker level of regulation or had a similar or even stronger regulation in the positive vs. negative control.

Both Otx2 and Spp1 have been shown to interact with Sox9 (SRY box transcription factor 9)²³ and like Fzd6 and Zic1, seem to play a role in Wnt/ß-catenin signaling. We, therefore, also analysed Sox-related gene expression and found that Sox7/8/9/10/11 and 13 were upregulated in the hypothalamus of the test group compared to the negative control. Notably, Sox11 showed the strongest immediate regulation post recall (Figure 4).

Looking at downregulation (Figure 2), among the genes with the strongest average regulation across all time points were dopamine receptors Drd1 and Drd2, inhibitory regulatory G-protein Rgs9, and Gpr88 (involved in negative regulation of opioid signaling), as well as thyroid hormone-binding protein crystallin-mu (Crym). Likewise, Gpr52, shown to have a modulatory role in dopamine receptor signaling, as well as Mlip, Tiam2, Kcnv1, Chrm1, Ephx4, Arhgap6, Tac3, Plk2, and Ano3 were strongly downregulated.

Looking in more detail at neurotransmitter-related regulation (Figure 3a), we detected an upregulation of inhibitory glycine- (Glra1, Slc6a9, Slc6a5) and GABA- (Slc6a11, Gabra6) and excitatory glutamate-related genes (Grik3, Grik4, Grm1, Grm4, Grin2d and 3b, Slc17a6 and Slc25a18). In addition, we saw upregulation of neuropeptide S (Nps), adenylate cyclase-activating polypeptide 1 (Adcyap1) plus upregulation of nicotinic cholinergic (Chrna4, Chrna2, and Chrna6) and opioid-transmission related genes (Pnoc, Pcsk1n; detected by two independent probes). Interestingly, while this upregulation was observed with multiple genes immediately post recall in the positive control, the same upregulation was only observed +3h post recall in the test group.

Figure 3a. Neurotransmitter-related genes upregulated in the hypothalamus; differential gene expression at 0h, +3h, +6h, +24h, and +48h for test vs. negative control and positive vs. negative control, respectively.

Grey lanes depict combinations not analyzed due to insufficient RNA quality. All data shown as 2^x.

This pattern of delayed regulation +3h in the test group vs. positive control was also seen with many downregulated, signaling-related transcripts (Figure 3b) such as enkephalin precursor proenkephalin Penk. We also observed a downregulation of Pcsk1 and prodynorphin Pdyn, opioid receptors kappa and mu (Oprk1, Oprm1), GABAergic- (Gabra4, Gabra5, Gabrad, Gabrarq), serotonergic- (Htr2a, Htr1d, and Htr6) and muscarinic cholinergic receptors (Chrm1), opioid regulators Rgs4 and 9, and neuropeptide Y (Npy; detected by two independent probes). Interestingly, the hypothalamus of test animals immediately post recall showed more than 30fold decrease in the expression of two of the most highly regulated genes, overall: pro-melanin-concentrating-hormone Pmch and vasoactive intestinal peptide Vip.

Figure 3b. Neurotransmitter-related genes downregulated in the hypothalamus; differential gene expression at 0h, +3h, +6h, +24h, and +48h for test vs. negative control and positive vs. negative control, respectively.

Grey lanes depict combinations not analyzed due to insufficient RNA quality. All data shown as 2^x.

Figure 4. Otx and Sox genes up- and downregulated in the hypothalamus; differential gene expression at 0h, +3h, +6h, +24h, and +48h for test vs. negative control and positive vs. negative control, respectively.

Grey lanes depict combinations not analyzed due to insufficient RNA quality. All data shown as 2^x.

Pituitary

In the pituitary of the test group and positive controls, we found a heterogenous pattern of expression in steroidogenic, nuclear, and metabolism-related genes. Except for Fam111a and Nt5c3b, most of the genes like Cyp11b1/2, Plk5, unknown transcript RGD1564409, and Myo15 showed an upregulation preferentially in the test group but not in positive controls; however, as we had to exclude the first two time points from the pituitary, most of the upregulation of these genes was observed only +24h post recall. Contrary to this pattern in the test group, we detected a very strong upregulation of interferon-stimulated genes (ISGs) in the positive vs. negative control such as Cxcl9/10/11, Gbp5, Rsad2, Ccl2 and 12, Isg15, Oasb1a/b, Ifit2/3, in some cases exceeding regulations by factor 100.

Looking at downregulation, we only observed four genes with an average gene expression downregulated >2fold: ubiquitin-conjugating enzyme Ube2U, Efcab1, Hist2h2ab, and Fndc9.

With respect to neurotransmission, we detected an upregulation of ionotropic glutamate receptors like kainate receptor (Grik2) at +6h and monoamine transporter Slc18a1 (VMAT1, endocrine variant of CNS-specific and hypothalamus-upregulated VMAT2), GABA receptors (Gabbr1, Gabre) and opioid binding protein-like (Opcml) upregulation in the pituitary +24h post recall of the test group. At the same time, tryptophan hydroxylase Tph1 and GABA-related transporter Slc6a1 were downregulated at +6h post recall, while cholinergic nicotinic neurotransmission-related genes as Chrna9 and Slc44a4, neuron-derived neurotrophic factor Ndnf and proline transporter Slc6a20 were downregulated +24h post recall (Figure 9).

Adrenal

In the adrenal glands, 19 genes were found to be upregulated >2fold on average in the test group compared to their negative control counterparts. Most of the upregulated genes observed were involved in thermogenesis and metabolic pathways like carbohydrate metabolism (Pck1), lipid metabolism (Cidec, Thrsp), triglyceride catabolism (Fabp3), and heat-generation related functions (Ucp1; Thermogenin), which showed a strong and immediate upregulation in the test group.

In addition to the above-mentioned genes, we observed an upregulation of tumor necrosis factor signaling genes (Tnfaip6, Nr4a2), stress-responsive gene (Ddit4), gene encoding serine/threonine protein kinase (Sik1), thyroid hormone responsive protein (Thrsp), nuclear receptor (Nr4a2), proto-oncogene (Junb), and cell death-inducing DFFA-like effector c (Cidec).

Looking at downregulation, we observed 9 transcripts with an average gene expression downregulated >2fold: Phosphodiesterase (Pde01a), Gdpgp1, mgc11619, LOC31693, Rho family GTPase (Rnd3), Guanosine monophosphate receptor (Gmpr), Myosin protein (Myo16), and Golgi vesicular membrane trafficking protein (Bet1l).

Related to neurotransmission in the adrenal glands of test animals, we found an immediate upregulation of adrenergic receptors (Adra1a, Adrb3), neuropeptide Y receptor (Npy4r), purinergic signaling receptor (P2rx5) as well as cholinergic nicotinic receptor (Chrna7). We also observed an upregulation of endorphin precursor proenkephalin (Penk), opioid receptor mu (Oprm1), neuron-derived neurotrophic factor (Ndnf), and dopamine receptor (Drd2) (Figure 10). We further saw an upregulation in the inhibitory neurotransmission-associated genes like GABA receptors (Gabrg1, Gabrg2), excitatory neurotransmitter glutamate/D-serine amino acid transporter (Slc1a4), and muscarinic cholinergic receptor 1 (Chrm1). Moreover, we observed single-point downregulation for opioid-related convertase (Pcsk1/2), vesicular glutamate transporter (Slc17a6), GABA transporters (Slc6a1, Slc6a11), and serotonin receptor (Htr3a).

Spleen

While analysing the gene expression in the spleen, the expression of 15 transcripts was found to be upregulated on average >2fold across all time points measured in the test group as compared to negative controls (Figure 1). These genes included Ig-like domain-containing protein MOR5S4 (“Hiramoto factor”), lymphocyte antigen 6 complex (Ly6al), ubiquitin-like modifier (Isg15), similar to paired-Ig-like receptor A1 LOC690948, signal regulatory protein delta (Sirpd), chemokines Cxcl10 and Ccl12, oligoadenylate synthetase (Oasl), interferon signaling molecules as Ifitm3 and Ifit3, oxidized low-density lipoprotein (Olr1), B-cell receptor Cd72, folate receptor 2, RT-1BP, and erythrocyte protein band Ebp4.1. Most of the upregulated genes identified in the spleen were either involved in immune activation or regulation.

Likewise, many genes downregulated in the spleen across all time points measured were related to immune regulatory functions such as G protein signaling regulator (Rgs13) and Killer cell lectin receptor (Klrd1). In addition, we observed a strong downregulation of olfactory receptor Olr1516, detected by two probes, Smarce1l, UDP-Gal:betaGlcNAc beta 1,4-galactosyltransferase polypeptide 4 (B4galt4) and two further uncharacterized transcripts (LOC10254, LOC10091).

Looking at gene expression regarding neuronal and endocrine signaling (Figure 11), we detected an upregulation of opioid signaling (Pcsk2 and Pcsk1; detected by two independent probes and opioid receptor Sigmar1), nicotinic acetylcholine receptor alpha 3 (Chrna3), ionotropic glutamate receptor delta 2 (Grid2), DNA directed RNA polymerase II (Polr2m), nmDA glutamate receptor ionotropic 1A (Grin1a), D-serine transporter (Slc1a4) and glutamate decarboxylase 2 (Gad2). We also observed downregulation of opioid receptor delta 1 (Oprd1), adrenomedullin 2 (Adm2), cholinergic receptor muscarinic 2, and both glutamatergic (Grik1, Gad2, Grm4), GABAergic (Gabbr1, Gabrp) and serotonergic receptors (Htr1d). Interestingly, we observed an upregulation of neuropeptide Y (Npy) and a downregulation of neuropeptide S (Nps) in the spleen, which is exactly the opposite of what was seen in the hypothalamus.

Taken together, gene expression data obtained from the spleen were summarized by a strong upregulation of immunomodulatory genes, chemo- and cytokines, and glutamatergic as cholinergic neurotransmission related genes. Similar to hypothalamus, we saw a pattern of immediate regulation in the spleen for positive control vs. negative control of opioid-related transcripts, such as Pcsk1 or Gpr88, while the same upregulation in the test group vs. negative control was only observed 3h later.

Validation of microarray gene expression data by using qRT-PCR

To validate the gene expression results obtained by rat full genome microarray analysis, we picked a representative subset of 32 genes including housekeeping genes, interleukins, ISGs, and chemokines, preferably regulated across multiple tissues (Table 2).²⁴ The mRNA quantity of those genes was analysed by qRT-PCR and correlation with the corresponding microarray results was assessed. While not all genes correlated perfectly, a strong overall correlation across multiple tissues and expression levels (r ² = 0.8377) was detected (Figure 5).

Figure 5. Scatter plot correlation analysis of 32 differentially expressed genes in all four tissues analyzed using Affymetrix chip analysis and qRT-PCR.

Plasma cytokine/chemokine levels post recall of the behaviorally conditioned immune response

Plasma levels of ACTH were found to be increased during the first six hours post recall in both test group and positive control animals, with a peak at +6h in the latter. Similarly, we observed an increased plasma level of IFN-γ immediately post recall in the test group and 0h to +3h in the positive control. IL-1α showed a similar yet less pronounced response in the plasma of both test and positive control animals, peaking at 0h and +6h post recall while IL-1β showed a similar profile to IFN-γ, peaking at 0h to +3h in the test and three hours later in the positive control. The most pronounced changes were observed for the plasma levels of chemokines GRO/KC/Cinc-1 (Cxcl1), IP-10 (Cxcl10), MCP-1 (Ccl2), MIP-1α (Ccl3), Rantes (Ccl5), and TNF-α, with a strong peak at +6h post recall, both in test group and positive control animals (Figure 6).

Figure 6. Heat map showing levels of IFN-γ, ACTH, GRO/KC/CINC-1, MIP-1α, RANTES, TNFα, MCP-1, and IP-10 in the plasma of test, positive control, and negative control animals.

Sucrose preference test (SPT) and anhedonia

Regarding future studies, we also assessed behavioural data during recall of the conditioned immune response employing the sucrose preference test to evaluate potentially conditioned depression-like anhedonia. During the run-in phase from day -7 to -1, a constant and strong preference (>90%) for the sucrose solution was observed in all study groups. As shown in Figure 12, we observed a drop in sucrose preference in both the test group and the positive control compared to D-1 after poly I:C injection; the reduction was much less pronounced in the negative control, receiving only saline injections. Following a return to normal on D1, we observed another drop in all three groups on D2 with further reduction in sucrose preference in all animal groups, potentially due to the repeated injection and handling stress (Figure 12).

Figure 7. Network chart, depicting potential interaction between upregulated genes connected to Otx-2, Spp1, and Slc18a2 in the hypothalamus and potential intermediaries (Qiagen 2020-2022 Ingenuity Pathway Analysis).

Figure 8. Illustration showing the possible efferent pathway involved in the immune regulation during recall of behaviorally conditioned immune response.

Figure 9. Neurotransmitter-related genes upregulated in the pituitary; differential gene expression at 0h, +3h, +6h, +24h, and +48h for test vs. negative control and positive vs. negative control, respectively.

Grey lanes depict combinations not analyzed due to insufficient RNA quality. All data shown as 2^x.

Figure 10. Neurotransmitter-related genes upregulated in the adrenal; differential gene expression at 0h, +3h, +6h, +24h, and +48h for test vs. negative control and positive vs. negative control, respectively.

Grey lanes depict combinations not analyzed due to insufficient RNA quality. All data shown as 2^x.

Figure 11. Neurotransmitter-related genes upregulated in the spleen; differential gene expression at 0h, +3h, +6h, +24h, and +48h for test vs. negative control and positive vs. negative control, respectively.

Grey lanes depict combinations not analyzed due to insufficient RNA quality. All data shown as 2^x.

Figure 12. Sucrose preference test (SPT).

Water and 2% sucrose solution were presented to all animals, and the relative consumption of sucrose solution is shown as % (100% = exclusively sucrose solution). The consumption over 24h was measured and analysed for each day.

Limitations of the study

The present study has several limitations: first, not all tissue samples could be analysed due to insufficient RNA quality. Furthermore, due to the character of a pilot study, we consciously used a limited amount of animals (2) per group and time point, respectively, and only a limited number of cytokines and signaling peptides were analysed on protein level.

Another limitation considers the design of the trial groups: our negative control was built in a way that animals were exposed to camphor smell and then injected with saline on both day D0 and D2 as opposed to an alternative design with poly I:C injection on D0 and saline injection on D2. While this latter design would have had the advantage of observing residual poly I:C injection effects on D2 in the negative control, we decided against this approach as the injection procedure itself, due to its pain, might have been perceived by the animals as a conditioning stimulus and thus resulted in a partial recall. Future studies might explore a different design and/or explore injection under anesthesia to limit the injection effect.

Independently of this, we believe that the gene expression data generated in this study, in combination with plasma protein data, provide strong evidence for a role of Otx2, Vmat2, and osteopontin – in concert with opioid and monoaminergic signaling – as key efferent messengers during the recall of the conditioned immune enhancement, confirming earlier results by Hiramoto, Ghanta, and Hsueh.^{9–12,37,38,40}

Reporting guidelines

Zenodo: ARRIVE checklist for “Differential gene expression during recall of behaviorally conditioned immune enhancement in rats: a pilot study”, https://doi.org/10.5281/zenodo.7086375.⁵⁷

Data are available under the terms of the Creative Commons Attribution 4.0 International “No rights reserved” data waiver (https://creativecommons.org/share-your-work/public-domain/cc0/)