Establishment of a paternal preconception donepezil exposure model in rats
To investigate the potential effects of chronic donepezil exposure on offspring, we randomly assigned wild-type male SD rats to receive either donepezil (Done, 4 mg/kg) or saline (Veh) treatment for 21 days. One day after the last dose was delivered, the rats were housed with wild-type naïve female rats for mating (Fig. 1a). There was no significant difference in the body weight gain in the two groups, as indicated by the body weight on day 0 or day 21 (Fig. 1b, day 0, t (19) = 0.248, P = 0.807; day 21, t (19) = 0.028, P = 0.977). Donepezil has been reported to enhance short-term memory and cognition. Therefore, multiple behavioral tests were used to assess its effects after mating. In the open field test, both Done and Veh rats traveled comparable total distances in the arena (Fig. 1c, left, t (19) = 1.419, P = 0.172). There were no significant differences between the two groups in terms of the distance traveled in the center (Fig. 1c, middle, t (19) = 1.212, P = 0.241) or time spent exploring the central area (Fig. 1c, right, t (19) = 1.294, P = 0.211). Similarly, Done and Veh rats spent comparable amounts of time exploring the open arm of the elevated plus arm maze (Fig. 1d, t (19) = 1.204, P = 0.243). We then used operant-conditioned learning and novel object recognition tests to assess the learning capacity and short-term memory of both Done and Veh rats. The group of rats treated with donepezil achieved the goal of obtaining 100 food pellets in a Skinner box via lever pressing significantly faster than did the control group treated with Veh (Fig. 1e, c2 = 8.354, P = 0.004). In the novel object recognition tests, both groups of rats spent comparable time exploring both objects (Fig. 1f, middle, t (19) = 1.081, P = 0.293), and were able to distinguish the novel object from the familiar object (Fig. 1f, right, t (19) = 0.515, P = 0.613). These data indicate increased short-term learning induced by 21-day donepezil treatment in male adult rats.
Paternal donepezil exposure does not disrupt birth and weight gain in F1 or F2 progeny
F1 progeny were generated by mating with naïve female animals, as mentioned above, while the F2 generation was obtained by crossing male Done-F1 and Veh-F1 with naïve female rats (Fig. 2a). The numbers of F1 offspring from veh-treated fathers (Veh) and those from donepezil-treated fathers (Done) were comparable per litter (Fig. 2b, t (19) = 1.053, P = 0.305). The body weights of male and female F1 offspring were measured between postnatal week 2 and week 8 and were found to be comparable (Fig. 2c, Fweek × sex × group (6, 264) = 0.3048, P = 0.9341; group, F (1, 44) = 0.1268, P = 0.724). Similarly, there was no difference in litter size (Fig. 2d, t (14) = 0.571, P = 0.577) or body weight between these two F2 groups (Fig. 2e, Fweek × sex × group (6, 216) = 1.260, P = 0.2773; group, F (1, 36) = 0.06088, P = 0.807). These data indicate that paternal exposure to donepezil does not disrupt birth or weight gain in F1 or F2 progeny.
Paternal donepezil exposure enhanced learning and short-term memory in F1 progeny
Behavioral tests were performed on eight-week-old F1 rats. Similar to the results obtained for the F0 generation, the male and female F1 offspring sired by both Done and Veh rats traveled comparable total distances in the arena during the open field test (Fig. 3a, left, F group × sex (1, 50) = 3.769, P = 0.0579; group, F (1, 50) = 0.1540, P = 0.696; sex, F (1, 50) = 2.435, P = 0.125). Additionally, there was no significant difference in the distance traveled (Fig. 3a, middle, F group × sex (1, 50) = 0.8623, P = 0.358; group, F (1, 50) = 0.799, P = 0.376; sex, F (1, 50) = 2.387, P = 0.129) or time spent exploring the central area (Fig. 3a, right, F group × sex (1, 50) = 0.0402, P = 0.842; group, F (1, 50) = 0.383, P = 0.539; sex, F (1, 50) = 1.110, P = 0.297) between the two groups. Furthermore, both Done and Veh F1 rats spent a comparable amount of time exploring the open arm of the elevated plus arm maze (Fig. 3b, F group × sex (1, 50) = 0.257, P = 0.615; group, F (1, 50) = 1.505, P = 0.226; sex, F (1, 50) = 4.452, P = 0.040). In two-bottle choice tests, Done F1 consumed more 5% sucrose water, indicating a normal hedonic response and a greater reward value associated with sucrose (Fig. 3c, F group × sex (1, 50) = 0.0132, P = 0.909; group, F (1, 50) = 4.105, P = 0.048; sex, F (1, 50) = 0.000765, P = 0.978).
Additionally, while F1 progeny from the two groups were comparable in time spent exploring the objects (Fig. 3d, left, F group × sex (1, 50) = 0.2237, P = 0.638; group, F (1, 50) = 1.321, P = 0.256; sex, F (1, 50) = 2.731, P = 0.105), Done-F1 outperformed Veh-F1 in discriminating novel objects (Fig. 3d, right, F group × sex (1, 50) = 1.977, P = 0.166; group, F (1, 50) = 8.218, P = 0.0061; sex, F (1, 50) = 0.829, P = 0.367). Similarly in Y maze tests, they spent more time exploring the novel arm, and enters the novel arm more than Veh-F1 (Fig. 3e, middle, F group × sex (1, 50) = 0.0632, P = 0.803; group, F (1, 50) = 22.79, P < 0.0001; sex, F (1, 50) = 0.0933, P = 0.761; right, F group × sex (1, 50) = 0.167, P = 0.685; group, F (1, 50) = 11.09, P = 0.0016; sex, F (1, 50) = 1.152, P = 0.288). They were also significantly faster than Veh-F1 in achieving the goal of obtaining 100 food pellets in a Skinner box (Fig. 3f, c2 = 10.56, P = 0.0012). Importantly, no differences were observed between males and females in any of the behavioral assessments performed. These behavioral assessments suggest that paternal exposure to donepezil enhances learning and short-term memory in F1 progeny. Importantly, no significant sex bias was detected.
Donepezil-sired F2 offspring show normalized learning compared to saline-sired controls
We conducted the same behavioral tests on the F2 offspring as on the F1 offspring. Notably, F2 offspring were sired by F1 offspring that were not subjected to any behavioral tests. The results showed no significant difference in the performance of open field tests(Fig. 4a, left, F group × sex (1, 37) = 0.844, P = 0.364; group, F (1, 37) = 0.243, P = 0.625; sex, F (1, 37) = 1.327, P = 0.257;middle, F group × sex (1, 37) = 0.332, P = 0.568; group, F (1, 37) = 0.159, P = 0.692; sex, F (1, 37) = 0.362, P = 0.551;right, F group × sex (1, 37) = 0.108, P = 0.745; group, F (1, 37) = 0.045, P = 0.833; sex, F (1, 37) = 0.375, P = 0.544), elevated plus maze (Fig. 4b, F group × sex (1, .37) = 0.357, P = 0.554; group, F (1, 37) = 2.354, P = 0.134; sex, F (1, 37) = 0.175, P = 0.678), two-bottle choice tests (Fig. 4c, F group × sex (1, 37) = 0.0349, P = 0.853; group, F (1, 37) = 1.035, P = 0.316; sex, F (1, 37) = 0.336, P = 0.567), novel object recognition tests (Fig. 4d, left, F group × sex (1, 37) = 0.0419, P = 0.839; group, F (1, 37) = 0.781, P = 0.383; sex, F (1, 37) = 3.656, P = 0.0636;right, F group × sex (1, 37) = 0.0429, P = 0.837; group, F (1, 37) = 0.746, P = 0.393; sex, F (1, 37) = 5.967, P = 0.0195), Y maze tests (Fig. 4e, left, F group × sex (1, 37) = 1.739, P = 0.195; group, F (1, 37) = 0.273, P = 0.605; sex, F (1, 37) = 2.514, P = 0.121;right, F group × sex (1, 37) = 0.342, P = 0.562; group, F (1, 37) = 0.191, P = 0.665; sex, F (1, 37) = 2.247, P = 0.142), or operant learning (Fig. 4f, c2 = 0.329, P = 0.566), between the F2 generation of veh-administered grandfathers (Veh) and those from donepezil-administered grandfathers (Done). These results suggest intergenerational but not transgenerational effects of learning and memory enhancement due to paternal preconception donepezil exposure on offspring.
Transcriptomic changes in the hippocampus of donepezil-treated offspring
We sampled hippocampal tissues from the Done-F1 and Veh-F1 groups for transcriptome sequencing and differential expression analysis. A total of 726 differentially expressed genes (DEGs; 418 upregulated and 308 downregulated) were identified under a p value ≤ 0.05 (Fig. 5a). First, functional overrepresentation was performed by comparing DEGs with genes expressed in the hippocampus as a background. Gene Ontology was used to analyze the function and distribution of these DEGs. The biological functions of DEGs were overrepresented in cell adhesion, morphogenesis, cell migration, and extracellular matrix composition, whereas the cellular component analysis revealed their distribution in the cell periphery, extracellular matrix, receptor complex, cell junction and vesicles(43). Because the brain is heterogeneous, we used a single-cell sequencing-based cell marker database, CellMarker 2.0(41), as a reference and predicted the cellular distribution of these DEGs. The results showed that the DEGs were primarily distributed in mesenchymal cells and microvascular pericytes (Fig. 5c). Interestingly, in contrast with the overrepresentation of pathways in nonneuronal cells of these DEGs, background genes were closely related to synapse function, with significant enrichment in synaptic long-term potentiation, the MAPK signaling pathway, and axon morphogenesis (Fig. 5b, left, yellow), and were primarily located in synapses, chromosomes, and the cytoplasm (Fig. 5b, right, yellow). These results suggest that paternal donepezil exposure might alter intercellular connections in the neurovascular units of offspring.
We also conducted a gene set enrichment analysis (GSEA) to identify potential disease associations of these genes. Comparison with the DisGeNET database revealed significant associations with autoimmune disease, cerebral ischemia, Alzheimer's disease, Parkinson's disease, and schizophrenia (Fig. 5d).
As donepezil-treated offspring show improvements in short-term memory and learning, it would be of potential interest to explore whether certain intelligence-related genes are altered. We retrieved a list of candidate genes related to intelligence and cognitive performance from two GWAS analyses of UK BioSample Bank samples(44) and (45)compared our DEGs with those candidates. Interestingly, out of the 726 DEGs, 130 collapsed (Fig. 5e-f). These genes were mainly involved in the Wnt, VEGF, and TNF pathways (Fig. 5g). Taken together, although no single gene or pathway was pinpointed, our data suggest epigenetic predisposition to learning and memory enhancement by genes involved in nonneural, extracellular regulation. Although our behavioral assessments of the F1 generation indicated improved learning and short-term memory in adulthood, there is concern about the potential for neurological disorders in middle and old age, which requires further behavioral and mechanistic investigations.