Abstract
Introduction:
Our laboratory investigates changes in the respiratory pattern during systemic inflammation in various rodent models. The endogenous cannabinoid system (ECS) regulates cytokine production and mitigates inflammation. Inflammation not only affects cannabinoid (CB) 1 and CB2 receptor gene expression (Cnr1 and Cnr2), but also increases the predictability of the ventilatory pattern.
Objectives:
Our primary objective was to track ventilatory pattern variability and transcription of Cnr1 and Cnr2 mRNA, and of Il1b, Il6, and tumor necrosis factor-alpha (Tnfa) mRNAs at multiple time points in central and peripheral tissues during systemic inflammation induced by peritonitis.
Methods:
In male Sprague Dawley rats (n=24), we caused peritonitis by implanting a fibrin clot containing either 0 or 25×106 Escherichia coli intraperitoneally. We recorded breathing with whole-animal plethysmography at baseline and 1 h before euthanasia. We euthanized the rats at 3, 6, or 12 h after inoculation and harvested the pons, medulla, lung, and heart for gene expression analysis.
Results:
With peritonitis, Cnr1 mRNA more than Cnr2 mRNA was correlated to Il1b, Il6, and Tnfa mRNAs in medulla, pons, and lung and changed oppositely in the pons, medulla, and lung. These changes were associated with increased predictability of ventilatory pattern. Specifically, nonlinear complexity index correlated with increased Cnr1 mRNA in the pons and medulla, and coefficient of variation for cycle duration correlated with Cnr1 and Cnr2 mRNAs in the lung.
Conclusion:
The mRNAs for ECS receptors varied with time during the central and peripheral inflammatory response to peritonitis. These changes occurred in the brainstem, which contains the network that generates breathing pattern and thus, may participate in ventilatory pattern changes during systemic inflammation.
Keywords: gene expression, inflammatory response, control of breathing, ventilatory pattern variability
Introduction
Cannabinoid (CB) receptor 11–7 and 2 mRNA8–11 are expressed in the central and peripheral immune cells and tissues. In wild-type animals, the endogenous cannabinoid system (ECS) regulates cytokine production and mitigates pathophysiology in several inflammatory conditions.12–18 If the ECS is disrupted during the inflammatory response, the normal proinflammatory cytokine increase19–21 becomes destructive. For example, in the cecal ligation and puncture or the lipopolysaccharide (LPS)-elicited endotoxemia models, serum levels of tumor necrosis factor-alpha (TNF-α) and IL-6 increase more in CB2-knockout mice than wild-type littermates.
Furthermore, CB2-knockout mice exhibit more lung injury markers and have worse survival outcomes. Selectively activating the CB2 receptors can reverse these effects by reducing leukocyte, neutrophil, and macrophage recruitment and reducing proinflammatory cytokine production4,6,21–28 in the intestinal and lung tissue. Moreover, CB2 receptor activation results in improved lung function,29 whereas CB1 activation transiently depresses respiratory rate and oxygen saturation.19
We have found in acute lung injury, systemic injection of LPS, and peritonitis that predictability of ventilatory pattern variability (VPV) increases concomitantly with peripheral and central increases in proinflammatory cytokine production, including brainstem areas that encompass the respiratory circuitry.20,21 To determine if the ECS could be linked to the respiratory response to neuroinflammation, we examined the time course of CB1 or CB2 mRNA expression in association with the altered respiratory pattern observed in acute peritonitis.
In this study, we measured VPV and the gene transcription responses relevant to ECS functions after systemic infection. In this study, we tested the hypotheses that changes in mRNAs for cannabinoid receptor (Cnr) 1 and 2 in the brainstem depend on the duration of exposure to an Escherichia coli inoculant and that these changes are associated with increases in VPV predictability.
Methods
All experimental protocols followed NIH guidelines and were approved by the Institutional Animal Care and Use Committee at Case Western Reserve University.
Animal model
Methods were adapted from a previously described protocol.26 We implanted male Sprague Dawley rats (n=24) with fibrin clots containing 0 (n=12) or 25×106 (n=12) E. coli, strain ATCC 25922 (American Type Culture Collection, Manassas, VA) colony-forming units (CFU; Ec0 or Ec25, respectively) intraperitoneally. We recorded ventilatory pattern before clot implantation and for 1 h before euthanasia at 3, 6, or 12 h later. We euthanized rats with an overdose of anesthesia followed by a large pneumothorax. We harvested lung, heart, medulla, and pons, which were frozen immediately on dry ice before storage at −80°C.
Quantifying the ventilatory pattern
We used whole-body plethysmography to record ventilatory waveforms from awake, spontaneously breathing rats. As the rats inspire, the resulting pressure increase of heating and hydrating the inhaled air is reflected in the Plexiglas chamber and is measured as airflow across a pneumotach. This signal was amplified (Max II, Buxco Electronics, DSI, St. Paul, MN), acquired (Sampling rate=200 Hz, Power1401, CED, Cambridge, UK), and stored with acquisition software (Spike 2, CED) for analysis of VPV and dynamics.
We analyzed three 1-min epochs from each rat's baseline and pre-euthanasia recordings. Representative (10 s) ventilatory pattern sample traces are displayed in Figure 1. We calculated the mean and standard deviation of inspiratory duration (Ti), expiratory duration (Te), cycle duration (Ttot), and its coefficient of variation (CVttot). We also measured nonlinear characteristics of the VPV as described previously.23
FIG. 1.
Paired representative 10-s traces (top 2 rows Baseline [gray] and Pre-euthanasia [black] time point) and the measurements (bottom two rows of numbers). We implanted the rats with either a sterile agar pellet (Ec0, left column) or one that had 25×106 colony-forming units of Escherichia coli (Ec25, right column). These recordings are from a whole-body plethysmograph and are ventilatory waveforms of awake, spontaneously breathing rats during 60-min recording periods at baseline and pre-euthanasia. We analyzed three 1-min epochs from each recording for inspiratory duration (Ti), expiratory duration (Te), cycle duration (Ttot), and its coefficient of variation (CVttot). Also, we calculated nonlinear characteristics of VPV, specifically, MI and SampEn. Finally, we calculated a NLCI for each epoch by determining the difference between the surrogate datasets and the original data. For this, we generated 19 iAAFT surrogate datasets that were constrained to have a similar AC function (r) as the original data. The calculated data are from one of the three epochs. The group data (the means for Ec0 [n=4], Ec25 [n=4] at 3, 6, and 12 h) are in Table 1. AC, autocorrelation; iAAFT, iterative amplitude-adjusted Fourier Transform; MI, mutual information; NLCI, nonlinear complexity index; Tnfa, tumor necrosis factor α; SampEn, sample entropy; VPV, ventilatory pattern variability.
This analysis included autocorrelation (AC) function (r, at one cycle length), mutual information (MI), and sample entropy (SampEn) for both the original data and 19 iterative amplitude-adjusted Fourier Transform surrogate datasets. From these analyses, we calculated the nonlinear complexity index (NLCI) for each epoch as the difference in SampEn between original and the mean of the surrogate datasets. We averaged values derived from the three 1-min epochs and present the average value for each measurement.
Gene expression
Pons, medulla, lung, and heart tissue were thawed on wet ice and total RNA was extracted using the TRIzol reagent method.27 Isolated RNA concentration was assessed through the NanoDrop (ND-1000, Thermo Fisher Scientific, Waltham, MA) spectrophotometer and purity was assured by an A260/A280 ratio range between 1.9 and 2.0. Samples that did not obtain the necessary A260/A280 ratio were purified using the sodium acetate and ethanol precipitation protocol or the PureLink® RNA Mini Kit (Thermo Fisher Scientific, Waltham, MA). One microgram of total RNA was reversed transcribed into cDNA using the High-Capacity cDNA Kit (Thermo Fisher, Applied Biosystems, Waltham, MA).
Real-time quantitative polymerase chain reaction (qPCR) was performed in duplicate for each sample using a StepOne Plus (Thermo Fisher Scientific, Waltham, MA) thermocycler. Each sample run was performed in duplicate and also repeated a second time; the results of the two runs were averaged.
The following TaqMan Gene® Expression Assays (Thermo Fisher, Applied Biosystems, Waltham, MA) were used: 18S ribosomal RNA (18S), glyceraldehyde-3-phosphate dehydrogenase (Gapdh), beta actin (Actb), Cnr1 and Cnr 2, interleukin (Il)-1beta (Il1b), Il6, and Tnfa. The most stable reference gene (RG) combination was calculated using NormFinder software (https://moma.dk/normfinder-software) and each gene of interest (GOI) was normalized to the geometric mean of the RG.
Gene expression comparisons were performed using the delta Ct (ΔCt) method, which is the difference in cycle threshold (Ct) for the GOI and the RG Ct value. The Ec0 group at each time point or the 3-h pre-euthanasia values for the Ec0 and Ec25 groups were used as the comparative controls for each relative analysis.
Statistical analyses
The mean respiratory measurement values obtained for each rat's baseline and pre-euthanasia recording were compared for each time point using a three-way repeated-measures analysis of variances (ANOVAs) (with Bonferroni corrections). The ΔCt values for each GOI were compared using a two-way ANOVA (with Bonferroni corrections). All analyses were conducted in the program SPSS Statistics version 25 (IBM, Armonk, NY). Values were considered significant if p≤0.05.
Results
Duration-dependent changes in the ventilatory patterns
We analyzed the mean values of the respiratory variables (Ti, Te, Ttot, CVttot, AC, MI, SampEn, and NLCI) from the baseline and pre-euthanasia at the 3, 6, and 12 h recordings (Table 1).
Table 1.
Time Dependence of Ventilatory Variables of Rats that Received Sterile (Ec0) or Infectious (Ec25) Clots
|
3 h
|
6 h
|
12 h
|
||||
|---|---|---|---|---|---|---|
| Baseline | Pre-euthanasia | Baseline | Pre-euthanasia | Baseline | Pre-euthanasia | |
| Ti (s) | ||||||
| Ec0 | 0.29±0.03 | 0.28±0.02 | 0.24±0.02 | 0.24±0.03 | 0.29±0.03 | 0.28±0.02 |
| Ec25 | 0.29±0.01 | 0.30±0.02 | 0.23±0.02 | 0.23±0.01 | 0.25±0.03 | 0.23±0.02 |
| Te (s) | ||||||
| Ec0 | 0.57±0.08 | 0.39±0.02 | 0.56±0.03 | 0.38±0.01 | 0.51±0.06 | 0.51±0.02 |
| Ec25 | 0.58±0.04 | 0.46±0.08 | 0.54±0.03 | 0.43±0.05 | 0.56±0.06 | 0.29±0.05 |
| Ttot (s) | ||||||
| Ec0 | 0.86±0.11 | 0.67±0.03 | 0.80±0.04 | 0.62±0.03 | 0.79±0.06 | 0.79±0.04 |
| Ec25 | 0.87±0.05 | 0.76±0.09 | 0.77±0.05 | 0.67±0.05 | 0.81±0.08 | 0.52±0.04 |
| CVttot | ||||||
| Ec0 | 0.14±0.02 | 0.25±0.08 | 0.18±0.02 | 0.27±0.04 | 0.12±0.03 | 0.15±0.02 |
| Ec25 | 0.22±0.04 | 0.16±0.03 | 0.18±0.03 | 0.19±0.04 | 0.15±0.03 | 0.09±0.02 |
| AC (r) | ||||||
| Ec0 | 0.54±0.04 | 0.52±0.13 | 0.52±0.06 | 0.49±0.08 | 0.64±0.09 | 0.50±0.02 |
| Ec25 | 0.56±0.05 | 0.62±0.08 | 0.43±0.07 | 0.48±0.05 | 0.49±0.08 | 0.79±0.09 |
| MI (bits) | ||||||
| Ec0 | 0.53±0.03 | 0.58±0.03 | 0.53±0.05 | 0.55±0.03 | 0.60±0.07 | 0.58±0.03 |
| Ec25 | 0.54±0.04 | 0.62±0.02 | 0.54±0.05 | 0.57±0.01 | 0.52±0.03 | 0.82±0.10 |
| SampEn (bits) | ||||||
| Ec0 | 1.31±0.02 | 1.31±0.05 | 1.34±0.01 | 1.34±0.03 | 1.22±0.05 | 1.26±0.02 |
| Ec25 | 1.31±0.02 | 1.27±0.02 | 1.33±0.03 | 1.25±0.05 | 1.27±0.02 | 1.00±0.07 |
| NLCI | ||||||
| Ec0 | 0.06±0.01 | 0.09±0.03 | 0.06±0.01 | 0.14±0.04 | 0.13±0.03 | 0.10±0.04 |
| Ec25 | 0.11±0.03 | 0.12±0.03 | 0.11±0.03 | 0.21±0.03 | 0.07±0.02 | 0.17±0.04 |
Ec0, no Escherichia coli; Ec25, 25×106 E. coli CFU Data: mean±SEM, n=4.
AC, autocorrelation; CFU, colony forming units; MI, mutual information; NLCI, non-linear complexity index; SampEn, sample entropy; SEM, standard error of the mean.
First, we compared the differences between baseline and its respective pre-euthanasia data at each time point for each group (Table 1 and 2 and; Figs. 1 and 2, in Figure 2 [significant differences denoted by color-coded horizontal lines with asterisks at the bottom of the graph]). In both groups, the rats that received sterile fibrin clots (Ec0) and experimental rats that received infectious clots (Ec25), TI did not change (Fig. 2A). In contrast, TE decreased in both Ec0 and Ec25 groups from respective baseline by 3 and 6 h but not at 12 h. For the Ec0 group, TE was no longer different from baseline at 12 h (Fig. 2B). However, for the Ec25 group, TE decreased from baseline at each time point and this difference appeared greatest at the 12-h time point (Fig. 2B).
Table 2.
Three-Way Analysis of Variance Comparing
| Ti | Te | Ttot | CVttot | AC | MI | SampEn | NLCI | ||
|---|---|---|---|---|---|---|---|---|---|
| Overall ANOVA | F=0.205; ns | F=3.98; p-=0.027 | F=3.14; p=0.055 | F=0.31; ns | F=1.88; ns | F=3.60; p-=0.038 | F=4.14; p-=0.024 | F=1.67NS | |
| Main effect | E. coli | F=0.77; ns | F=0.15; ns | F=0.41; ns | F=1.02; ns | F=0.39; ns | F=2.11; ns | F=7.30; p=0.01 | F=4.53; p=0.04 |
| Duration | F=5.83; p=0.006 | F=0.52; ns | F=1.91; ns | F=5.43; p=0.009 | F=2.79; p=0.075 | F=3.32; p=0.048 | F=14.0; p=0.0005 | F=1.26NS | |
| Inoculation | F=0.14; ns | F=25.65; p=0.0005 | F=18.35; p=0.0005 | F=1.02; ns | F=0.73; ns | F=7.65; p=0.009 | F=7.29; p=0.011 | F=7.37; p=0.010 | |
F, F-value; p, significance; ns, not significant p>0.1.
ANOVA, analysis of variance.
FIG. 2.
Time and bacterial burden-dependent changes in respiratory variables following inoculation. Values are paired (connected by color-coded lines): on the left, baseline values obtained the day before surgery, on the right, values by 3, 6, and 12 h after surgical implant of sterile pellets (Ec0, light-colored dashed lines) or pellets with 25 million E. coli cells (Ec25, dark-colored lines). We compare: (A) Inspiratory duration (TI), (B) Expiratory duration (TE), (C) Autocorrelation coefficient (AC, r) at one cycle length, (D) MI, (E) SampEn, and (F) NLCI. A three-way ANOVA revealed significant changes in TE [F(2,36)=3.98, p=0.027], MI [F(2,36)=3.60, p=0.038], and SampEn [F(2,36)=4.14, p=0.024], and a two-way ANOVA revealed significant changes in NLCI [F(1,36)=3.48, p=0.076]. See Table 1 and text for details. Mean±SEM, values for Ec0 (n=4), Ec25 (n=4). **p<0.05. ANOVA, analysis of variance.
These decreases in TE were reflected in TTOT (not shown, but these data are in Table 1). Also, in the Ec25 group, decreases in variability in TTOT and the ventilatory waveform were evident by 12 h as CVttot decreased (not shown) and r, the AC coefficient, increased (Fig. 2C). The decrease in variability could be associated with an increased predictability of the ventilatory pattern. We calculated MI, SampEn, and NLCI and in the Ec25 group, MI increased (Fig. 2D), SampEn decreased (Fig. 2E), and NLCI increased (Fig. 2F) by 12 h.
Next, we compared the respiratory measurements between the Ec0 and Ec25 groups at 3, 6, and 12 h postinoculation (Fig. 2, significant differences denoted by the black vertical bars on the right of the panels). Although this analysis was not as statistically sensitive as the paired comparisons, the results supported our findings of the paired comparisons. We found differences for TE, AC, MI, and SampEn between Ec0 and Ec25 groups only at 12 h.
Finally, we compared the Ec25 groups across the different pre-euthanasia time points (Fig. 2 significant differences denoted by the two-colored horizontal bars above the graph). MI was greater by 12 h compared with that at 3 and 6 h (Fig. 2D) and similarly, SampEn was less by12 h compared with 3 and 6 h (Fig. 2E).
RG combination expression
In tissue harvested from the rats (Ec0 and Ec25; n=24), the NormFinder software selected the most stable RG as 18S and Actb for the pons and Actb and Gapdh for the medulla, lung, and heart. No RG changes were observed across time points and bacterial burden in pons [F(2,18)=0.351, p=0.71], medulla [F(2,18)=0.097, p=0.91], lung [F(2,18)=0.177, p=0.84], and heart [F(2,18)=1.070, p=0.36] (Figs. 3, 4, Left Column).
FIG. 3.
Time-dependent qPCR analysis of RG combinations and cannabinoid receptor (Cnr) gene expression. The geometric mean expression of RG (left column) combinations, Cnr1 (middle column), and Cnr2 (right column) were compared in (A) pons, (B) medulla, (C) lung, and (D) heart tissue obtained from Ec0 (n=12) and Ec25 (n=12) rats. The RG did not have significant differences across time in pons (A), medulla (B), lung (C), and heart (D). In contrast, a two-way ANOVA revealed time interactions within each bacterial burden dose for Cnr1 and Cnr2 for: (A2 and 3) pons; (B2 and 3) medulla; (C2 and 3) lung; and (D2 and 3) for heart. See Table 3 and text for details. Mean±SEM, Significant comparisons by 3, 6, and 12 h harvest time points within Ec0 or Ec25 groups are indicated with **p<0.05, ***p<0.001. qPCR, quantitative polymerase chain reaction; RG, reference gene.
FIG. 4.
Bacterial burden-dependent qPCR analysis of RG combinations and cannabinoid receptor (Cnr) gene expression. The geometric mean expression of RG (left column) combinations, Cnr1 (middle column), and Cnr2 (right column) were compared for the (A) pons, (B) medulla, (C) lung, and (D) heart. Tissue obtained from Ec0 (n=12) and Ec25 (n=12) rats. No significant RG differences in pons, medulla, lung, or heart were observed. However, a two-way ANOVA revealed bacterial burden-dependent interactions by each harvest time point for Cnr1 and Cnr2: (A2 and 3) pons; (C2 and 3) lung; (D2 and 3) heart; and Cnr2 for medulla (B3). See Table 4 and text for details. Mean±SEM, Significant comparisons between Ec0 and Ec25 by 3, 6, and 12 h harvest time points are indicated with **p<0.05, ***p<0.001.
Duration dependence of cannabinoid receptor gene expression
In this analysis, we examined whether tissue-specific changes in mRNA for Cnr1 and Cnr2 correlate with changes in VPV. Figure 3 presents how duration after inoculation affected the expression of Cnr1 mRNA (Fig. 3, middle) and Cnr2 mRNA (Fig. 3, right) in the pons (Fig. 3A), medulla (Fig. 3B), lung (Fig. 3C), and heart (Fig. 3D). For data and statistics, see Table 3. Table 5 presents changes in Il1b, Il6, and Tnfa mRNA expression.
Table 3.
Two-Way Analysis of Variance Comparing Duration-Dependent Fold Change Differences in Cnr1 and Cnr2 Gene Expression in Pons, Medulla, Lung, and Heart
| |
Pons
|
Medulla
|
Lung
|
Heart
|
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 3 h | 6 h | 12 h | 3 h | 6 h | 12 h | 3 h | 6 h | 12 h | 3 h | 6 h | 12 h | ||||
| Cnr1 | |||||||||||||||
| Ec0 | 1.0±0.14 | 0.51±0.08** | 0.25±0.02*** | 1.0±0.12 | 0.54±0.09 | 1.51±0.03 | 1.0±0.13 | 0.15±0.03*** | 1.06±0.13 | 1.0±0.10 | 0.80±0.06 | 0.69±0.10** | |||
| Ec25 | 1.0±0.20 | 0.06±0.01*** | 0.02±0.003*** | 1.0±0.18 | 0.4±0.03*** | 0.29±0.04*** | 1.0±0.08 | 0.5±0.07*** | 0.50±0.06*** | 1.0±0.15 | 1.85±0.25*** | 3.36±0.57*** | |||
| Cnr2 | |||||||||||||||
| Ec0 | 1.0±0.15 | 0.91±0.09 | 0.89±0.12 | 1.0±0.14 | 1.04±0.17 | 0.68±0.08** | 1.0±0.14 | 0.08±0.01*** | 37.0±4.13*** | 1.0±0.12 | 4.04±0.37*** | 1.47±0.25** | |||
| Ec25 | 1.0±0.21 | 1.53±0.13** | 2.09±0.35** | 1.0±0.14 | 1.54±0.17** | 1.79±0.22** | 1.0±0.15 | 2.37±0.35*** | 20.4±2.57*** | 1.0±0.12 | 10.5±1.25** | 3.0±0.51** | |||
| Pons | Medulla | Lung | Heart | ||||||||||||
| Overall ANOVA | Cnr1 | F=63.62; p-=0.0005 | F=1.16; ns | F=70.32; p=0.0005 | F=35.69; p-=0.0005 | ||||||||||
| Main effect | Ec0 | F=33.93; p=0.0005 | F=68.32; p=0.0005 | F=178.05; p=0.0005 | F=3.94; p=0.038 | ||||||||||
| Time point | Ec25 | F=281.58; p=0.0005 | F=23.70; p=0.0005 | F=41.81; p=0.0005 | |||||||||||
| Overall ANOVA | Cnr2 | F=5.51; p=0.014 | F=9.86; p=0.001 | F=181.07; p=0.0005 | F=25.83; p=0.0005 | ||||||||||
| Main effect | Ec0 | F=0.24; ns | F=4.65; p=0.024 | F=739.39; p=0.0005 | F=47.97; p=0.0005 | ||||||||||
| Time point | Ec25 | F=8.07; p=0.003 | F=7.52; p=0.004 | F=185.20; p=0.0005 | F=27.94; p=0.0005 | ||||||||||
Data: mean±SEM, n=4. Significance code: ** p<0.05, *** p<0.001.
F, F-value; p, significance; ns, not significant.
Table 5.
Two-Way Analysis of Variance Comparing Duration-Dependent Fold Change Differences in Il1b, Il6, and Tnfa Gene Expression in Pons, Medulla, Lung, and Heart
|
Pons
|
Medulla
|
Lung
|
Heart
|
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 3 h | 6 h | 12 h | 3 h | 6 h | 12 h | 3 h | 6 h | 12 h | 3 h | 6 h | 12 h | |
| Il1b | ||||||||||||
| Ec0 | 1.0±0.15 | 1.045±0.12 | 1.28±0.33 | 1.0±0.10 | 5.62±0.54 *** | 7.45±0.81*** | 1.0±0.14 | 0.20±0.04*** | 12.25±1.67*** | 1.0±0.14 | 4.93±0.38*** | 294.8±45.8*** |
| Ec25 | 1.0±0.12 | 0.45±0.03** | 2.39±0.40*** | 1.0±0.17 | 11.18±1.09*** | 9.56±1.24*** | 1.0±0.08 | 0.67±0.12** | 0.41±0.05*** | 1.0±0.13 | 10.51±1.25*** | 1.26±0.24 |
| Il6 | ||||||||||||
| Ec0 | 1.0±0.14 | 0.15±0.02 *** | 0.02±0.05 *** | 1.0±0.16 | 0.74±0.09** | 0.83±0.10 | 1.0±0.12 | 0.17±0.04 *** | 0.56±0.08** | 1.0±0.13 | 42.78±2.37 *** | 16.73±2.54 *** |
| Ec25 | 1.0±0.15 | 0.33±0.03*** | 11.75±2.51*** | 1.0±0.14 | 0.17±0.02 *** | 0.21±0.02*** | 1.0±0.15 | 0.70±0.10* | 0.94±0.11 | 1.0±0.15 | 0.28±0.04*** | 0.77±0.17 |
| Tnfa | ||||||||||||
| Ec0 | 1.0±0.19 | 2.32±0.28 *** | 0.74±0.07** | 1.0±0.11 | 1.72±0.23** | 1.36±0.13** | 1.0±0.16 | 5.05±1.09 *** | 25.09±2.90 *** | 1.0±0.10 | 36.93±2.27 *** | 108.3±16.8 *** |
| Ec25 | 1.0±0.17 | 1.93±0.19 *** | 2.68±0.38 *** | 1.0±0.15 | 4.48±0.37 *** | 6.81±0.73 *** | 1.0±0.16 | 56.31±7.29 *** | 0.99±0.13 | 1.0±0.12 | 0.53±0.07 *** | 4.94±1.04*** |
| Pons | Medulla | Lung | Heart | |||||||||
| Overall ANOVA | Il1b | F=13.64; p-=0.0005 | F=8.59; p-=0.002 | F=209.56; p-=0.0005 | F=744.01; p-=0.0005 | |||||||
| Main effect | Ec0 | F=0.88; ns | F=167.51; p=0.0005 | F=317.23; p=0.0005 | F=1114.94; p=0.0005 | |||||||
| Time point | Ec25 | F=35.17; p=0.0005 | F=7.52; p=0.004 | F=14.51; p=0.0005 | F=217.75; p=0.0005 | |||||||
| Overall ANOVA | Il6 | F=82.81; p=0.0005 | F=38.16; p=0.0005 | F=23.45; p=0.0005 | F=265.88; p=0.0005 | |||||||
| Main effect | Ec0 | F=42.33; p=0.0005 | F=2.53; ns | F=75.58; p=0.0005 | F=314.73; p=0.0005 | |||||||
| Time point | Ec25 | F=158.07; p=0.0005 | F=105.05; p=0.0005 | F=2.53; p=0.051 | F=37.90; p=0.0005 | |||||||
| Overall ANOVA | Tnfa | F=35.7; p=0.0005 | F=39.29; p=0.0005 | F=321.07; p=0.0005 | F=337.52; p=0.0005 | |||||||
| Main effect | Ec0 | F=39.01; p=0.0005 | F=8.80; p=0.002 | F=207.92; p=0.0005 | F=844.82; p=0.0005 | |||||||
| Time point | Ec25 | F=27.99; p=0.0005 | F=121.46; p=0.0005 | F=434.72; p=0.0005 | F=185.30; p=0.0005 | |||||||
Data: mean±SEM, n=4. p, significance. Significance code: *p<0.1, **p<0.05, ***p<0.001.
F, F-value; p, significance; ns, not significant.
In the Ec0 group, qPCR analysis revealed that Cnr1 mRNA expression (1) decreased in pons and medulla as postsurgical time increased (Fig. 3A, B, central column, left bar graphs); (2) decreased transiently by 6 h in lung (Fig. 3C), and (3) decreased in heart, but not until 12 h (Fig. 3D). In the Ec25 group, Cnr1 mRNA expression: (1) decreased in pons, medulla, and lung (Fig. 3A–C), and (2) increased progressively in heart (Fig. 3D).
In the Ec0 group, Cnr2 mRNA expression: (1) did not change across harvest times in pons (Fig. 3A, left column); (2) decreased by 12 h in medulla (Fig. 3B); (3) initially decreased by 6 h, but increased by 12 h in lung; and (4) increased transiently by 6 h, and then decreased by 12 h, but to a level still greater than 3 h in heart. In the Ec25 treatment group, Cnr2 mRNA expression increased progressively by 6 and by 12 h in each tissue (Fig. 3A–D).
Bacterial burden dependence of CB receptor gene expression
Because we observed tissue-specific decreases in Cnr1 and increases in Cnr2 expression, we tested the hypothesis that these changes are also modulated by bacterial dose. Figure 4 presents how the presence of inoculant at each harvest time point affected the expression of Cnr1 mRNA (Fig. 4, middle) and Cnr2 mRNA (Fig. 4, right) in the pons (Fig. 4A), medulla (Fig. 4B), lung (Fig. 4C), and heart (Fig. 4D). For data and statistics see Table 4. Table 6 presents changes in mRNA for Il1b, Il6, and Tnfa.
Table 4.
Two-Way Analysis of Variance Comparing Bacterial Burden-Dependent Fold Change Differences in Cnr1 and Cnr2 Expression in Pons, Medulla, Lung, and Heart
|
Pons
|
Medulla
|
Lung
|
Heart
|
|||||
|---|---|---|---|---|---|---|---|---|
| Cnr1 | Cnr2 | Cnr1 | Cnr2 | Cnr1 | Cnr2 | Cnr1 | Cnr2 | |
| 3 h | ||||||||
| Ec0 | 1.0±0.14 | 1.0±0.15 | 1.0±0.12 | 1.0±0.14 | 1.0±0.13 | 1.0±0.14 | 1.0±0.10 | 1.0±0.12 |
| Ec25 | 2.23±0.45*** | 0.32±0.07*** | 1.04±0.19 | 0.71±0.01** | 14.69±1.14*** | 0.84±0.13 | 0.69±0.11** | 0.85±0.11 |
| 6 h | ||||||||
| Ec0 | 1.0±0.16 | 1.0±0.10 | 1.0±0.08 | 1.0±0.16 | 1.0±0.20 | 1.0±0.21 | 1.0±0.09 | 1.0±0.08 |
| Ec25 | 0.26±0.03*** | 0.54±0.05** | 0.77±0.06 | 1.05±0.12 | 47.78±7.07*** | 26.02±3.81*** | 1.77±0.22*** | 0.39±0.05*** |
| 12 h | ||||||||
| Ec0 | 1.0±0.10 | 1.0±0.14 | 1.0±0.10 | 1.0±0.11 | 1.0±0.12 | 1.0±0.11 | 1.0±0.15 | 1.0±0.17 |
| Ec25 | 0.18±0.03*** | 0.75±0.13 | 0.86±0.11 | 1.87±0.23** | 1.55±0.78*** | 0.12±0.06*** | 3.37±0.57*** | 1.74±0.30** |
| Overall ANOVA | F=63.62; p-=0.0005 | F=5.51; p-=0.014 | F=1.16; p-=0.336 | F=9.86; p-=0.001 | F=70.32; p-=0.0005 | F=181.07; p-=0.0005 | F=35.69; p-=0.0005 | F=25.83; p-=0.0005 |
| Main effect E. coli | Cnr1 | Cnr2 | Cnr1 | Cnr2 | Cnr1 | Cnr2 | Cnr1 | Cnr2 |
| 3 h | F=63.62; p=0.0005 | F=63.62; p=0.0005 | F=0.08; ns | F=4.93; p=0.039 | F=537.81; p=0.0005 | F=1.09; p=0.310 | F=7.60; p=0.013 | F=1.20; ns |
| 6 h | F=63.62; p=0.0005 | F=63.62; p=0.0005 | F=3.43; p=0.080 | F=0.10; ns | F=1113.3; p=0.0005 | F=407.78; p=0.0005 | F=18.23; p=0.0005 | F=41.04; p=0.0005 |
| 12 h | F=63.62; p=0.0005 | F=63.62; p=0.0005 | F=1.11; ns | F=16.19; p=0.001 | F=280.00; p=0.0005 | F=22.47; p=0.0005 | F=83.29; p=0.0005 | F=14.10; p=0.011 |
Data: means±SEM, n=4. Significance code: ** p<0.05, *** p<0.001.
F, F-value; p, significance; ns, not significant.
Table 6.
Two-Way Analysis of Variance Comparing Bacterial Burden-Dependent Fold Change in Il1b, Il6, and Tnfa Gene Expression in Pons, Medulla, Lung, and Heart
|
Pons
|
Medulla
|
Lung
|
Heart
|
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Il1b | Il6 | Tnfa | Il1b | Il6 | Tnfa | Il1b | Il6 | Tnfa | Il1b | Il6 | Tnfa | |
| 3 h | ||||||||||||
| Ec0 | 1.0±0.15 | 1.0±0.19 | 1.0±0.30 | 1.0±0.10 | 1.0±0.16 | 1.0±0.11 | 1.0±0.14 | 1.0±0.12 | 1.0±0.16 | 1.0±0.14 | 1.0±0.13 | 1.0±0.10 |
| Ec25 | 2.74±0.33*** | 0.45±0.07** | 1.2±0.21 | 2.41±0.41*** | 2.78±0.40*** | 0.82±0.13 | 12.09±0.93*** | 12.94±1.90*** | 10.60±1.75*** | 1.13±0.56*** | 421.1±62.8*** | 7.20±0.88*** |
| 6 h | ||||||||||||
| Ec0 | 1.0±0.11 | 1.0±0.14 | 1.0±0.12 | 1.0±0.10 | 1.0±0.12 | 1.0±0.13 | 1.0±0.21 | 1.0±0.23 | 1.0±0.22 | 1.0±0.08 | 1.0±0.06 | 1.0±0.06 |
| Ec25 | 1.19±0.08 | 0.97±0.08 | 1.04±0.10 | 4.80±0.47*** | 0.63±0.08** | 2.15±0.18*** | 39.58±7.26*** | 52.02±7.97*** | 118.4±15.3*** | 8.92±1.06*** | 2.73±0.35*** | 0.10±0.01*** |
| 12 h | ||||||||||||
| Ec0 | 1.0±0.26 | 1.0±0.15 | 1.0±0.10 | 1.0±0.11 | 1.0±0.11 | 1.0±0.10 | 1.0±0.14 | 1.0±0.14 | 1.0±0.12 | 1.0±0.16 | 1.0±0.15 | 1.0±0.15 |
| Ec25 | 5.13±0.85*** | 15.92±3.4*** | 4.5±0.64*** | 3.10±0.40*** | 0.69±0.07** | 4.14±0.45*** | 0.41±0.05*** | 21.66±2.62*** | 0.42±0.05*** | 0.02±0.003*** | 19.34±4.16*** | 0.33±0.07*** |
| Overall ANOVA | F=13.64; p=0.0005 | F=82.81; p=0.0005 | F=35.7; p=0.0005 | F=8.59; p=0.002 | F=38.16; p=0.0005 | F=39.29; p=0.0005 | F=209.57; p=0.0005 | F=23.45; p=0.0005 | F=321.07; p=0.0005 | F=744.01; p=0.0005 | F=265.88; p=0.0005 | F=337.5; p=0.0005 |
| Main effect E. coli | Il1b | Il6 | Tnfα | Il1b | Il6 | Tnfα | Il1b | Il6 | Tnfα | Il1b | Il6 | Tnfα |
| 3 h | F=25.74; p=0.0005 | F=15.00; p=0.001 | F=2.63; ns | F=55.05; p=0.0005 | F=57.84; p=0.0005 | F=2.22; ns | F=231.08; p=0.0005 | F=310.49; p=0.0005 | F=223.16; p=0.0005 | F=132.40; p=0.0005 | F=1503.86; p=0.0005 | F=273.45; p=0.0005 |
| 6 h | F=0.75; ns | F=0.022; ns | F=0.071; ns | F=174.46; p=0.0005 | F=11.59; p=0.003 | F=35.05; p=0.0005 | F=503.19; p=0.0005 | F=739.7; p=0.0005 | F=912.2; p=0.0005 | F=310.13; p=0.0005 | F=41.39; p=0.0005 | F=360.90; p=0.0005 |
| 12 h | F=67.71; p=0.0005 | F=180.03; p=0.0005 | F=126.03; p=0.0005 | F=90.63; p=0.0005 | F=7.58; p=0.013 | F=120.40; p=0.0005 | F=29.84; p=0.0005 | F=448.03; p=0.0005 | F=30.34; p=0.0005 | F=1049.05; p=0.0005 | F=361.41; p=0.0005 | F=87.03; p=0.0005 |
Data: mean±SEM, n=4. p, significance. Significance code: *p<0.1, **p<0.05, ***p<0.001.
F, F-value; p, significance; ns, not significant.
Ec25 initially increased Cnr1 mRNA expression in the pons and lungs compared with Ec0 (Fig. 4A, C), but Cnr1 mRNA expression decreased by later time points in the pons. In the lung, Cnr1 mRNA expression remained increased by 6 and 12 h (Fig. 4A, C). In the medulla (Fig. 4B), Cnr1 mRNA expression did not change by 3, 6, and 12 h.
In the heart (Fig. 4D), Cnr1 mRNA expression initially decreased by 3 h, but then increased by 6 and 12 h. Cnr2 mRNA (Fig. 4, right) differed between Ec0 and Ec25 as follows: (1) by 3 h, it decreased in pons and medulla (Fig. 4A, B), but remained unchanged in the lung and heart (Fig. 4C, D); (2) by 6 h, Cnr2 mRNA expression decreased in the pons and heart (Fig. 4A, D), increased in the lung (Fig. 4C), and remained unchanged in the medulla (Fig. 4B); (3) and by 12 h, Cnr2 mRNA expression was unchanged in the pons (Fig. 4A), increased in the medulla and heart (Fig. 4B, D), and decreased in the lung (Fig. 4C).
VPV parameters compared with CB receptor gene expression
A regression analysis was performed comparing the coefficient of variation of respiratory cycle duration (CVttot) against the ΔCt Cnr1 mRNA (Table 7 and Fig. 5 top row, square symbols) and Cnr2 mRNA (Table 7, Fig. 5 bottom row, circular symbols) expression in the pons (Fig. 5, left column) and lung (Fig. 5, right column). Data from rats (n=4) within a group (colored-coded symbols) formed clusters. In the pons, weak and nonsignificant correlations, points tended to cluster. However, in the lung, the clusters were robust. We conclude the clusters support a common time dependence in the relationship between variability in respiratory cycle duration and Cnr1 and Cnr2 mRNA expression.
Table 7.
Correlation Coefficients (r) Between Ventilatory Pattern Variables and Cycle Thresholds for mRNA Expression of the RG and Cannabinoid Receptors
| Variable | mRNA | Pons | Medulla | Lung | Heart |
|---|---|---|---|---|---|
| NLCI | RG | 0.11 | −0.06 | −0.13 | −0.38* |
| Cnr1 | 0.44** | 0.47** | −0.09 | 0.28 | |
| Cnr2 | −0.04 | 0.28 | −0.33 | −0.11 | |
| CVttot | RG | 0.18 | −0.31 | 0.02 | −0.15 |
| Cnr1 | −0.38* | −0.35* | 0.37* | 0.33 | |
| Cnr2 | −0.25 | 0.05 | 0.49*** | −0.04 | |
| Te | RG | 0.14 | −0.32 | −0.11 | −0.13 |
| Cnr1 | −0.18 | −0.01 | 0.03 | 0.41*** | |
| Cnr2 | −0.06 | 0.51*** | −0.27 | 0.11 | |
| MI | RG | −0.09 | 0.62**** | −0.03 | 0.44** |
| Cnr1 | 0.03 | 0.06 | −0.32 | −0.20 | |
| Cnr2 | 0.09 | 0.06 | −0.24 | −0.03 | |
| SampEn | RG | 0.06 | −0.32 | −0.14 | −0.09 |
| Cnr1 | −0.42* | −0.39* | 0.41** | 0.41** | |
| Cnr2 | −0.27 | 0.06 | 0.60**** | 0.11 | |
| AC | RG | −0.08 | 0.37* | 0.01 | 0.34* |
| Cnr1 | 0.17 | 0.12 | −0.32 | −0.32 | |
| Cnr2 | 0.19 | −0.12 | −0.18 | −0.09 |
Cnr1, Cannabinoid Receptor 1; Cnr2, Cannabinoid Receptor 2; RG, Reference Gene. Symbols: p, significance. Significance code: *p<0.1; **p<0.05, ***p<0.001, ****p<0.0005.
Te, expiratory duration.
FIG. 5.
Regression of the coefficient of variation for respiratory cycle duration (Ttot; CVttot) versus the change in cycle threshold (ΔCt) for mRNA for endocannabinoid receptor 1 (Cnr1; top panels) and 2 (Cnr2; bottom panels) expressed in the pons (left panels) and lung (right panels). The different symbols represent the different groups, with each rat represented individually. Symbols represent the means for the three analyzed epochs at each pre-euthanasia time point (n=3) of a rat (n=4) for Ec0 and for Ec25. For details, see Table 7 and text.
Due to indications of common dynamics in respiratory frequency and mRNA for CB receptors, we evaluated the relationship between SampEn and Cnr1 mRNA in the pons (Fig. 6A), medulla (Fig. 6B), lung (Fig. 6C) and heart (Fig. 6D). For statistics, see Table 7. Again, the time points formed clusters and the distribution of clusters varied; for example, in the lung, the cycle threshold (Ct) for Cnr1 mRNA decreased, almost uniformly, in inoculated rats. The surprise was in the pons, the clusters were tighter, and the correlation coefficient greater, and Ct for Cnr1 mRNA depended on the survival time. Compared with 3 h, the ΔCt increased possibly by 6 h, but definitely by 12 h, as SampEn decreased.
FIG. 6.
Regression of SampEn versus the change in cycle threshold (DCt) for endocannabinoid receptor 1 (Cnr1) mRNA expressed in the Central Nervous System (A: pons, B: medulla) and Peripheral Tissues (C: lung, D: heart). Symbols represent the mean for the three analyzed epochs at each pre-euthanasia time point (n=3) of a rat (n=4) for Ec0 and for Ec25. The negative correlation of SampEn to the Ct for Cnr1 complements Figure 5 with CVttot, even though these respiratory variables can change independently. For details, see Table 7 and text.
Changes in mRNAs for proinflammatory cytokines compared with those for CB receptors
As stated in the Introduction, VPV becomes more predictable in conditions that elicit systemic and brainstem inflammation. In Table 8, we provide the correlation coefficients between the mRNA for proinflammatory cytokines and variables of the respiratory pattern to provide a context for the Cnr1 and Cnr2 correlations. We expected that increase in the mRNA for proinflammatory cytokines would be correlated to NLCI, but this occurred only the mRNA for Il6 in the medulla (Table 8). In contrast, Il1b and Il6 were correlated positively to CVttot and SampEn and negatively to AC coefficient.
Table 8.
Correlation Coefficients Between Ventilatory Pattern Variables and Cycle Thresholds for Il1b, Il6, and Tnfa mRNA
| Variable | mRNA | Pons | Medulla | Lung | Heart |
|---|---|---|---|---|---|
| NLCI | Il1b | −0.26 | −0.59**** | −0.15 | −0.40** |
| Il6 | 0.26 | 0.40** | −0.20 | −0.38* | |
| Tnfa | −0.43** | −0.41** | −0.41** | −0.01 | |
| CVttot | Il1b | 0.57**** | 0.32 | 0.25 | 0.07 |
| Il6 | 0.45** | −0.09 | 0.35* | 0.40** | |
| Tnfa | 0.19 | 0.27 | 0.24 | 0.21 | |
| Te | Il1b | 0.24 | 0.09 | −0.33 | −0.39* |
| Il6 | 0.30 | −0.08 | 0.30 | 0.47** | |
| Tnfa | 0.38* | 0.30 | −0.32 | −0.22 | |
| MI | Il1b | −0.34* | −0.08 | −0.23 | 0.05 |
| Il6 | −0.37* | −0.12 | −0.39* | −0.22 | |
| Tnfa | −0.04 | −0.08 | −0.09 | −0.22 | |
| SampEn | Il1b | 0.60**** | 0.36* | 0.28 | 0.09 |
| Il6 | 0.54*** | −0.09 | 0.44** | 0.36* | |
| Tnfa | 0.23 | 0.30 | 0.29 | 0.20 | |
| AC | Il1b | −0.53*** | −0.12 | −0.05 | 0.16 |
| Il6 | −0.36* | 0.00 | −0.37* | −0.25 | |
| Tnfa | −0.23 | −0.17 | −0.03 | −0.15 |
Il1b, Interleukin lβ; Il6, Interleukin 6; Tnfa, tumor necrosis factor α. Symbols: p, significance. Significance code: *p<0.1; **p<0.05, ***p<0.001, ****p<0.0005.
Finally, we correlated mRNA for proinflammatory cytokines to Ct Cnr1 (Fig. 7) and Ct Cnr2 (Fig. 8). We expected that microbial product, such as LPS or Pathogen-Associated Molecular Patterns (PAMPs) would be a common driver to mRNA for proinflammatory cytokines and endocannabinoid receptors, both of which are present in the central nervous system. Although Il1b, Il6, and Tnfa were correlated to Cnr1 in the pons and medulla, the strongest positive correlations between these variables were in the lung (Fig. 7). In contrast, Cnr2 is associated with peripheral tissue. Il1b, Il6, and Tnfa were not correlated to Cnr2 in the pons but were correlated to Cnr2 in the medulla. However, the strongest correlation for Cnr2 was with Il1b in the lung.
FIG. 7.
Linear regression of mRNA for proinflammatory cytokines (Il1b (top row), Il6 (middle row), and Tnfa (bottom row)) against mRNA for Cnr1 for the pons (left column), medulla (center column), and lung (right column). Cnr1 is present in the CNS. Even though the range of ΔCt for Cnr1 is small in the medulla, ΔCt Cnr1 is highly correlated to ΔCt for the proinflammatory cytokines. In the periphery, the ΔCt Cnr1 in the lung is highly correlated to ΔCt Il1b and ΔCt Il6. Also, for the lung, the rats that received the infectious pellets had decreased ΔCt Cnr1 and decreased ΔCt Il6. For details, see text. CNS, central nervous system.
FIG. 8.
Linear regression of mRNA for proinflammatory cytokines (Il1b [top row], Il6 [middle row], and Tnfa [bottom row]) against mRNA for Cnr2 for the pons [left column], medulla [center column], and lung [right column]). Cnr2 is not prevalent in the CNS. Furthermore, the ranges of changes in Ct are minimal for proinflammatory cytokines and Cnr2. Consequently, most of the correlation are not or weakly significant. In contrast, in the lung, the Ct ranges for proinflammatory cytokines and for Cnr2 are broader and generally, the infected rats have decreased Ct for both. However, only the correlation between Ct for Il1b and Cnr2 is significant. For details, see text.
Discussion
We tested a hypothesis that ventilatory variability of the CB system responds to microbial products that alter variability of the respiratory pattern. We confirmed that Ec25 increases pattern predictability as indicated by decreases in SampEn and increases MI in postinoculation recordings. In general, changes in the respiratory variables were associated with increases (Cnr1 mRNA) and decreases (Cnr2 mRNA) in the expression of mRNA for the receptors of the ECS in pons, medulla, and lung. We also demonstrated a potentially complex interdependence in the influences between the expression of mRNA for proinflammatory cytokines and for the endocannabinoid receptors.
Standard measures of the respiratory pattern focus on the timing of the cycle. For instance, cycle duration decreases and respiratory frequency increases during systemic infection. Nonlinear measures of VPV focus on the predictability of waveform pattern and on time-dependent characteristics of the waveform, which requires the use of surrogate datasets to distinguish stochastic from time-dependent variability.
We construct surrogate datasets by shuffling the original data but only use the surrogate datasets that replicate (within less than a 2.5% tolerance) the auto-correlation function of the original dataset. Thus, the surrogate datasets account for the random variability of the original data but destroy its time-dependent properties. While MI, and SampEn are nonlinear measures of the predictability of the waveform, NLCI measures the predictability of waveform that is due to its time-dependent characteristics. In this study, we report the AC, MI, and SampEn of the original dataset.
As breathing becomes more regular, the AC coefficient at 1 cycle length will increase, the MI (a measure of the statistical dependence in the dataset) will increase, and SampEn (a measure of self-similarity in the dataset) will decrease.
NLCI evaluates the difference in SampEn between the original and surrogate datasets and defines a measure of the time-dependent portion of total “self-similarity” in the waveform. This component of total variability increases during systemic inflammation and may be due to a decrease in synaptic efficacy at the first-order synapse between vagal sensory input and neurons in the nucleus tractus solitarii. Thus, an increase in NLCI along with decreases in SampEn and increases in AC and MI, all support decreases in efficacy of sensory input and other sources of inputs, and the remaining variability of the rhythm depends on intrinsic sources of the rhythm generating network, such as an internal feedback loop between the dorsolateral pons and ventrolateral medulla.
Time-dependent changes in gene expression and VPV
We assessed the transcriptional mechanisms of the ECS during systemic inflammation by correlating increased VPV predictability (Fig. 2) with coincident gene expression changes (Fig. 3) in pons, medulla, lung, and heart. Previous reports indicate that systemic inflammation reduces dynamic and static lung compliance and a respiratory index measure. Moreover, Cnr2 mRNA expression and activation of its cognate receptor in lung tissue are important in mitigating proinflammatory cytokine production following an inflammatory insult.25 These data support our VPV observation in Ec25, but not Ec0, rats that SampEn decreased and MI increased compared with the earliest harvest time point.
Interestingly, our qPCR analysis (Fig. 3) demonstrates that a simultaneous time-dependent decrease in Cnr1 mRNA expression in pons and medulla may be required for the observed VPV changes. Also, sustained decreases in Cnr1 mRNA and increases in Cnr2 mRNA expression correlated with a more predictable ventilatory waveform. Our findings support and extend previous results demonstrating the relationship between altered ventilatory factors resulting from acute lung injury and extend the observations on proinflammatory cytokine production to include ECS gene expression changes.
Bacterial burden-dependent changes in gene expression and VPV
We further refined the association between increased VPV predictability to include decreased respiratory cycle times (Fig. 2) and tissue-specific gene expression changes (Fig. 4) for Ec0 and Ec25 groups by 12 h postinoculation. Although we found significant gene expression changes by 3 and 6 h postinoculation between these groups, we did not observe concomitant respiratory changes. The unique time-dependent mRNA expression patterns observed in Figure 3 were repeated in this dose-dependent analysis by 12 h postinoculation.
The substantial fold change increase in Cnr1 and Cnr2 mRNA expression by 6 h (Fig. 4) in the Ec25 lung samples is indicative of their low expression in Ec0 lung samples. Collectively, our data suggest that opposite changes in Cnr1 and Cnr2 mRNA expression in pons and lung, as well as the increase in Cnr2 mRNA in the medulla, are important for the production of the observed VPV predictability effects.
Moreover, these observations point to the intimate interaction between the lungs and the brainstem areas responsible for breathing pattern generation. Of note, the central activation of CB1 receptors has been shown to worsen respiratory measures and increase markers of organ failure.22,24,28 By contrast, systemic activation of CB2 receptors can improve neurogenic pulmonary edema and the respiratory index.30 These CB receptor-mediated effects may also reduce the link between inflammatory mediators and the direct stimulation of vagal C-fibers to alter the breathing pattern.
Finally, our model of peritonitis using agar/fibrin clots with bacteria is meant to release microbial PAMPs gradually as opposed to an injection of LPS, which presents an acute load. We did determine that E. coli are viable after implanting the pellet. Our protocol focused on the mRNA expression of Cnr1 and Cnr2, but this is only the one factor in series of steps in protein synthesis, post-transcriptional modification, cell membrane localization, and degradation. The correlations present among the mRNAs for endocannabinoid receptors, proinflammatory cytokines, and respiratory pattern variables indicate that these factors are associated. The rationale for the correlations was the concept that a common driver, such as a microbial product, will affect these variables simultaneously.
In summary, our data support the assertion that the severity of the inflammatory insult alters the effectiveness of the ECS to enhance survival outcomes. Compared with the Ec0 group, the magnitude of the inflammatory insult in the Ec25 group may have been sufficient to activate an innate immune response through a preponderance of alveolar macrophages31 or eosinophils.32 The overall dynamic Cnr1 mRNA and Cnr2 mRNA expression levels may also reflect possible dendritic cell33,34 and T cell35–40 activation. A future study in this model is needed to confirm protein expression levels and receptor activity of the immune cells described above.
Conclusion
In this model of systemic infection, dynamics in Cnr1 and Cnr2 mRNA expression were reciprocal in pons, medulla, and lung, and parallel in heart tissue. Moreover, these expression patterns were associated with a decrease in VPV. The Cnr mRNA expression changes in pulmonary tissue and central tissues responsible for breathing pattern generation, in conjunction with worsening ventilatory measures, may hold potential as a biomarker informative of an innate, adaptive immune response transition in the setting of this model.
Abbreviations Used
- AC
autocorrelation
- ANOVA
analysis of variance
- CFU
colony-forming units
- CNS
central nervous system
- CVttot
coefficient of variation for TTOT
- ECS
endogenous cannabinoid system
- iAAFT
iterative amplitude-adjusted Fourier Transform
- MI
mutual information
- NLCI
nonlinear complexity index
- qPCR
quantitative polymerase chain reaction
- RG
reference gene
- SampEn
sample entropy
- SEM
standard error of the mean
- Te
expiratory duration
- Tnfa
tumor necrosis factor α
- VPV
ventilatory pattern variability
Author Disclosure Statement
No competing financial interests exist.
Funding Information
The study was funded by the National Institutes of Health 5T32HL007913-18, National Institutes of Health U01 EB021960, and Veteran Affairs Research Service I01BX004197.
Cite this article as: Horton K-KA, Campanaro CK, Clifford C, Nethery DE, Strohl KP, Jacono FJ, Dick TE (2023) Cannabinoid receptor mRNA expression in central and peripheral tissues in a rodent model of peritonitis, Cannabis and Cannabinoid Research 8:3, 510–526, DOI: 10.1089/can.2021.0085.
References
- 1. Herkenham M, Lynn AB, Johnson MR, et al. Characterization and localization of cannabinoid receptors in rat-brain —a Quantitative Invitro Autoradiographic Study. J Neurosci. 1991;11:563–583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Herkenham M, Lynn AB, Little MD, et al. Cannabinoid receptor localization in brain. Proc Natl Acad Sci U S A. 1990;87:1932–1936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Matsuda LA, Bonner TI, Lolait SJ. Localization of cannabinoid receptor mRNA in rat brain. J Comp Neurol. 1993;327:535–550. [DOI] [PubMed] [Google Scholar]
- 4. Galiegue S, Mary S, Marchand J, et al. Expression of central and peripheral cannabinoid receptors in human immune tissues and leukocyte subpopulations. Eur J Biochem. 1995;232:54–61. [DOI] [PubMed] [Google Scholar]
- 5. Bouaboula M, Rinaldi M, Carayon P, et al. Cannabinoid-receptor expression in human leukocytes. Eur J Biochem. 1993;214:173–180. [DOI] [PubMed] [Google Scholar]
- 6. Laprairie RB, Kelly ME, Denovan-Wright EM. The dynamic nature of type 1 cannabinoid receptor (CB(1)) gene transcription. Br J Pharmacol. 2012;167:1583–1595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Matias I, Bisogno T, Di Marzo V. Endogenous cannabinoids in the brain and peripheral tissues: regulation of their levels and control of food intake. Int J Obes (Lond). 2006;30 Suppl 1:S7–S12. [DOI] [PubMed] [Google Scholar]
- 8. Cabral GA, Marciano-Cabral F. Cannabinoid receptors in microglia of the central nervous system: immune functional relevance. J Leukoc Biol. 2005;78:1192–1197. [DOI] [PubMed] [Google Scholar]
- 9. Gong JP, Onaivi ES, Ishiguro H, et al. Cannabinoid CB2 receptors: immunohistochemical localization in rat brain. Brain Res. 2006;1071:10–23. [DOI] [PubMed] [Google Scholar]
- 10. Onaivi ES, Ishiguro H, Gong JP, et al. Discovery of the presence and functional expression of cannabinoid CB2 receptors in brain. Ann N Y Acad Sci. 2006;1074:514–536. [DOI] [PubMed] [Google Scholar]
- 11. Van Sickle MD, Duncan M, Kingsley PJ, et al. Identification and functional characterization of brainstem cannabinoid CB2 receptors. Science. 2005;310:329–332. [DOI] [PubMed] [Google Scholar]
- 12. Alhouayek M, Lambert DM, Delzenne NM, et al. Increasing endogenous 2-arachidonoylglycerol levels counteracts colitis and related systemic inflammation. FASEB J. 2011;25:2711–2721. [DOI] [PubMed] [Google Scholar]
- 13. Di Sabatino A, Battista N, Biancheri P, et al. The endogenous cannabinoid system in the gut of patients with inflammatory bowel disease. Mucosal Immunol. 2011;4:574–583. [DOI] [PubMed] [Google Scholar]
- 14. Rajesh M, Mukhopadhyay P, Batkai S, et al. CB2-receptor stimulation attenuates TNF-alpha-induced human endothelial cell activation, transendothelial migration of monocytes, and monocyte-endothelial adhesion. Am J Physiol Heart Circ Physiol. 2007;293:H2210–H2218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Zhao Y, Liu Y, Zhang W, et al. WIN55212-2 ameliorates atherosclerosis associated with suppression of pro-inflammatory responses in ApoE-knockout mice. Eur J Pharmacol. 2010;649:285–292. [DOI] [PubMed] [Google Scholar]
- 16. Costola-de-Souza C, Ribeiro A, Ferraz-de-Paula V, et al. Monoacylglycerol lipase (MAGL) inhibition attenuates acute lung injury in mice. PLoS One. 2013;8:e77706. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Liu AP, Yuan QH, Zhang B, et al. Cannabinoid receptor 2 activation alleviates septic lung injury by promoting autophagy via inhibition of inflammatory mediator release. Cell Signal. 2020;69:109556. [DOI] [PubMed] [Google Scholar]
- 18. Liu Z, Wang Y, Zhao H, et al. CB2 receptor activation ameliorates the proinflammatory activity in acute lung injury induced by paraquat. Biomed Res Int. 2014;2014:971750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Iring A, Hricisak L, Benyo Z. CB1 receptor-mediated respiratory depression by endocannabinoids. Respir Physiol Neurobiol. 2017;240:48–52. [DOI] [PubMed] [Google Scholar]
- 20. Jacono FJ, Mayer CA, Hsieh YH, et al. Lung and brainstem cytokine levels are associated with breathing pattern changes in a rodent model of acute lung injury. Respir Physiol Neurobiol. 2011;178:429–438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Young BP, Loparo KA, Dick TE, et al. Ventilatory pattern variability as a biometric for severity of acute lung injury in rats. Respir Physiol Neurobiol. 2019;265:161–171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Byrne K, Cooper KR, Carey PD, et al. Pulmonary compliance: early assessment of evolving lung injury after onset of sepsis. J Appl Physiol. 1990;69:2290–2295. [DOI] [PubMed] [Google Scholar]
- 23. Dhingra RR, Jacono FJ, Fishman M, et al. Vagal-dependent nonlinear variability in the respiratory pattern of anesthetized, spontaneously breathing rats. J Appl Physiol. 2011;111:272–284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Gui H, Sun Y, Luo ZM, et al. Cannabinoid receptor 2 protects against acute experimental sepsis in mice. Mediators Inflamm. 2013;2013:741303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Litvin DG, Dick TE, Smith CB, et al. Lung-injury depresses glutamatergic synaptic transmission in the nucleus tractus solitarii via discrete age-dependent mechanisms in neonatal rats. Brain Behav Immun. 2018;70:398–422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Namas RA, Namas R, Lagoa C, et al. Hemoadsorption reprograms inflammation in experimental gram-negative septic peritonitis: insights from in vivo and in silico studies. Mol Med. 2012;18:1366–1374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Nolan RL, Teller JK. Diethylamine extraction of proteins and peptides isolated with a mono-phasic solution of phenol and guanidine isothiocyanate. J Biochem Biophys Methods. 2006;68:127–131. [DOI] [PubMed] [Google Scholar]
- 28. Tschop J, Kasten KR, Nogueiras R, et al. The cannabinoid receptor 2 is critical for the host response to sepsis. J Immunol. 2009;183:499–505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Fujii M, Sherchan P, Soejima Y, et al. Cannabinoid receptor type 2 agonist attenuates acute neurogenic pulmonary edema by preventing neutrophil migration after subarachnoid hemorrhage in rats. Acta Neurochir Suppl. 2016;121:135–139. [DOI] [PubMed] [Google Scholar]
- 30. Leite-Avalca MCG, Staats FT, Verona D, et al. Cannabinoid CB1 receptor antagonist rimonabant decreases levels of markers of organ dysfunction and alters vascular reactivity in aortic vessels in late sepsis in rats. Inflammation. 2019;42:618–627. [DOI] [PubMed] [Google Scholar]
- 31. Lai CJ, Ruan T, Kou YR. The involvement of hydroxyl radical and cyclooxygenase metabolites in the activation of lung vagal sensory receptors by circulatory endotoxin in rats. J Appl Physiol. 2005;98:620–628. [DOI] [PubMed] [Google Scholar]
- 32. Csoka B, Nemeth ZH, Mukhopadhyay P, et al. CB2 cannabinoid receptors contribute to bacterial invasion and mortality in polymicrobial sepsis. PLoS One. 2009;4:e6409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Hunninghake GW, Gadek JE, Kawanami O, et al. Inflammatory and immune processes in the human lung in health and disease: evaluation by bronchoalveolar lavage. Am J Pathol. 1979;97:149–206. [PMC free article] [PubMed] [Google Scholar]
- 34. Staiano RI, Loffredo S, Borriello F, et al. Human lung-resident macrophages express CB1 and CB2 receptors whose activation inhibits the release of angiogenic and lymphangiogenic factors. J Leukoc Biol. 2016;99:531–540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Chouinard F, Lefebvre JS, Navarro P, et al. The endocannabinoid 2-arachidonoyl-glycerol activates human neutrophils: critical role of its hydrolysis and de novo leukotriene B4 biosynthesis. J Immunol. 2011;186:3188–3196. [DOI] [PubMed] [Google Scholar]
- 36. Chouinard F, Turcotte C, Guan X, et al. 2-Arachidonoyl-glycerol- and arachidonic acid-stimulated neutrophils release antimicrobial effectors against E. coli, S. aureus, HSV-1, and RSV. J Leukoc Biol. 2013;93:267–276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Oka S, Ikeda S, Kishimoto S, et al. 2-arachidonoylglycerol, an endogenous cannabinoid receptor ligand, induces the migration of EoL-1 human eosinophilic leukemia cells and human peripheral blood eosinophils. J Leukoc Biol. 2004;76:1002–1009. [DOI] [PubMed] [Google Scholar]
- 38. Kaplan BL. The role of CB1 in immune modulation by cannabinoids. Pharmacol Ther. 2013;137:365–374. [DOI] [PubMed] [Google Scholar]
- 39. Matias I, Pochard P, Orlando P, et al. Presence and regulation of the endocannabinoid system in human dendritic cells. Eur J Biochem. 2002;269:3771–3778. [DOI] [PubMed] [Google Scholar]
- 40. Borner C, Hollt V, Sebald W, et al. Transcriptional regulation of the cannabinoid receptor type 1 gene in T cells by cannabinoids. J Leukoc Biol. 2007;81:336–343. [DOI] [PubMed] [Google Scholar]