Chronic paternal morphine exposure increases sensitivity to morphine-derived pain relief in male progeny

Parental history of opioid exposure is seldom considered when prescribing opioids for pain relief. To explore whether parental opioid exposure may affect sensitivity to morphine in offspring, we developed a “rat pain scale” with high-speed imaging, machine learning, and mathematical modeling in a multigenerational model of paternal morphine self-administration. We find that the most commonly used tool to measure mechanical sensitivity in rodents, the von Frey hair, is not painful in rats during baseline conditions. We also find that male progeny of morphine-treated sires had no baseline changes in mechanical pain sensitivity but were more sensitive to the pain-relieving effects of morphine. Using RNA sequencing across pain-relevant brain regions, we identify gene expression changes within the regulator of G protein signaling family of proteins that may underlie this multigenerational phenotype. Together, this rat pain scale revealed that paternal opioid exposure increases sensitivity to morphine’s pain-relieving effects in male offspring.

directed each VFH towards the center of the plantar paw and pressed upward until the filament bent [50]. For the four natural stimuli and VFHs, a non-responsive animal did not respond within 2 s of stimulus delivery. For traditional measures using VFH (Figure 3), animals were placed in the same plexiglass enclosure and paw responses were scored as either withdrawn or not (binary outcome). Animals were stimulated 5 times with each of the following VFH: 1g, 2g, 4g, 6g, 8g, 10g, 15g, 26g, 60g, 100g, 180g, 300g. When animals withdrew their paw 2 or more times (40%) at a given force, the experiment stopped and the threshold was reported at that VFH force. If the threshold was below 10g, one more round of stimulation at 10g was performed to establish the percentage of responses at 10g. The purpose of that experiment was to collect data to assess another commonly reported measurement of pain in the scientific literature: proportion of responses at 10g of force. For experiments testing the antinociceptive properties of morphine using VFH stimulations and the novel pain scale, we used 10g, 100g and 300g of force. Each stimulus was measured on a different day and the order in which stimuli were presented was counterbalanced across animals. All experiments were carried out by an experimenter blind to experimental group.

Scoring hind paw withdrawal movement features
We extracted paw height and paw speed from the high-speed videos and processed with Photron FastCAM software. We scored paw height in centimeters as the distance from the mesh floor to the highest point following paw stimulation. We calculated paw speed as the distance, in centimeters, from initial paw lift to the highest point, divided by the time in seconds between the two points. The composite nocifensive score is a composite of four individual behavior features: orbital tightening, paw shake, paw guard, and jumping. For example, if a given animal displayed 1/4 of those features it would receive a composite nocifensive score of 1. We scored orbital tightening when the eyes went from fully open to partially or fully closed following stimulus application. We defined paw shaking as high frequency paw flinching. We defined jumping as three or more paws off the mesh floor at the same time following a stimulus application. Lastly, we defined paw guard as any abnormal orientation, or placement, of the paw back during the descent of the withdrawal following stimulus application. We were not blind to the strain when scoring behaviors of wild-type rats, as Long Evans have white coats with a black hood color, while Sprague Dawleys are white-coated. However, we are blind to the stimulus type and VFH forces.

Timeseries experimental paradigm
Following 2-minutes of habituation inside of the testing chamber, naïve adult male rats received mechanical stimulation to their hindpaw for approximately 1-2 seconds or until an apparent behavioral response was elicited with 1 of 3 stimuli: cotton swab, light pinprick or heavy pinprick. A different stimulus was used for different groups of rats during the 3-day experiment, with the initial stimulation serving as a baseline measure of pain sensitivity. Rats were then immediately injected with morphine (either 1mg/kg or 3mg/kg subcutaneously) and returned to their home cage. At 15-min and 60 min after the morphine injection, rats were returned to the testing chamber to receive hindpaw stimulation from the same stimulus. Behavioral responses were recorded using high-speed videography at baseline, 15min, and 60min post-injection.

RNA sequencing:
To have an overview of samples and RNA-seq reads, the quality control (QC) step was done using Fastqc software (version 0.11.8) [51]. Based on the Fastqc software report, the Trimmomatic (version 0.39) software [52] was applied to remove non-paired reads and identified adaptors. For RNA-seq analysis, the suggested pipeline by Sahraeian et al. was used [ ] as previously reported [48]. Briefly speaking, the reads were mapped on rn6 genome assembly using the Hisat2 [53] as the alignment step, the assembled by StringTie [54]. The differentially expressed genes (DEGs) were selected by applying Deseq2 library [55] in R (version 4.0.1) [56].
The conditions to define DEGs were a 50% change in the expression (|log 2 Fold change| > 0. 58) of genes and an adjusted p-value < 0.1. R statistical software (version 4.0.3) was also used for downstream analysis and visualization of the RNA-seq analysis output, including, but not limited to drawing heatmaps and Venn diagrams.
To find the DEGs' transcription factors, HOMER [57] was used to identify potential transcription factors responsible for DEGs. The applied criteria were − 2000 to + 1000 bp of the transcriptional start site and a length of 8-12 bp for TFs.
Pathways (from KEGG) and gene ontology terms (biological process, molecular function and cell compartment) by adjusted p-value < 0.1 were selected for further analysis. Also, in the case of REACTOME as another source of enrichment analysis, the ReactomePA package in R was used [58]. To compare samples' gene expression, agnostically, the Rank-Rank Hypergeometric Overlap (RRHO) analysis was used. This algorithm steps within two gene lists ranked by the p-value of differential expression observed in two experiments and estimating the number of overlapping genes. Subsequently, a heatmap is created that shows the strength and pattern of correlation between two expression profiles [59]. This analysis was done using the RRHO2 library in R [60] to compare gene expression between Nac, VTA and PAG for overlap in up-and downregulated DEGs, respectively. This research includes calculations carried out on Temple University's HPC resources and thus was supported in part by the National Science

Statistical differences in composite nocifensive score, paw speed and paw height of rats in responses to innocuous (CS and DB) and noxious stimuli (LP and HP)
Hind paw stimulation with CS or DB produced significantly lower composite nocifensive scores than LP or HP across both sex and strain (Supplemental Figure 1A; Long Evans Female:

Linear transformation and machine learning estimates the pain-probability of a given response on a trial-by-trial basis
Machine learning approaches allowed for consistent prediction of pain probability on a trial-bytrial basis (Supplemental Figure 2). Principal Component 1 scores of CS and HP trials were used to train a support vector machine. We chose CS and HP trials because they triggered "nonpainful" or "painful" behaviors with high confidence and the corresponding PCA scores showed the most consistent patterns across strain and sex. The trained SVM then predicted the probability of being "pain-like" for other trials. Quite analogously to the PCA based approach, we consistently separated innocuous versus noxious behavioral responses.

Von Frey hair filaments do not elicit pain-like responses
Applying VFHs stimulation of varying forces resulted in responses that registered in the nonpain domain for nocifensive scores, paw height and velocity.
There were no significant difference in composite nocifensive score across any of the VFHs for male or female SD or LE rats (Supplemental Figure  Step (1): calculate PC1 score of each trial by performing PCA on the z-scores from Table 1. Step (2): Split PC1 scores into training and testing sets.
Step (3): train SVM with PC1 scores of training data (CS and HP).
Step (4) Supplemental Figure 3 . Sub-second temporal mapping of rat composite nocifensive score and paw kinematic behavioral profile in response to Von Frey hair filaments. (A) Composite nocifensive score of Long Evans and Sprague Dawley female and male rats (40 rats in total; 10 per group) following stimulation with Von Frey hair filaments of varying forces: 0.008, 10, 100, and 300 g. The score is a composite measurement of eye grimace/orbital tightening, paw shake, jumping, and paw guard. For instance, animals featuring 3 of the 4 behaviors are assigned a score of 3 for that particular trial. (B) Hind-paw kinematic movements evoked by the same VFH filaments. Paw speed of the first paw raise of the stimulated paw is the distance from the initial paw lift to the highest point divided by the time in seconds between the two points.   Table 5. List of differentially expressed genes in the PAG, comparing male saline-sired to male morphine-sired offspring. Supplemental Table 6. Enrichment analyses of DEGs in PAG comparing male saline-sired to male morphine-sired progeny.
Supplemental Figure 5. Baseline composite nocifensive score and paw kinematic behavioral profile of first-generation offspring derived from saline-exposed and morphine-exposed sires in response to application of mechanical stimuli to the hind paw (A) Baseline composite nocifensive score drug-naïve firstgeneration male derived from sires exposed to wither saline or morphine. The score is a composite measurement of eye grimace/orbital tightening, paw shake, jumping, and paw guard. For instance, animals featuring 3 of the 4 behaviors are assigned a score of 3 for that particular trial. (B) Hind paw kinematic movements evoked by the same VFH filaments. Paw speed of the first paw raise of the stimulated paw is the distance from the initial paw lift to the highest point divided by the time in seconds between the two points. (C) Paw height is the distance from the mesh floor to the highest point following paw stimulation.