Prenatal PM2.5 exposure impairs spatial learning and memory in male mice offspring: from transcriptional regulation to neuronal morphogenesis

Abstract

Background: As one of the environmental risk factors for human health, atmospheric fine particulate matter (PM_2.5) contributes to cognitive deterioration in addition to respiratory and cardiovascular injuries. Recently, increasing evidence implicates that PM_2.5 inhalation can affect neurological functions in offspring, but the sex-specific outcomes and the underlying biological processes are largely unknown.

Objectives: To observe the influence of prenatal PM_2.5 exposure on cognitive performance in offspring, to elucidate the neuronal morphological alterations and possible transcriptional regulation based on mRNA-sequencing (mRNA-Seq) data after birth, and to determine the key components of PM_2.5 contributing to the adverse effects.

Methods: Pregnant C57BL/6J mice were exposed to sterile saline or PM_2.5 suspension. Morris water maze test was used to assess the cognitive function in weanling offspring. Microscopic observation was applied to detect neuronal morphogenesis in vivo and in vitro. The cortex tissues from male offspring were collected on postnatal days (PNDs) 1, 7, and 21 for mRNA-Seq analysis. The organic and inorganic components of PM_2.5 were separated to assess their contributions using primary cultured neurons.

Results: Prenatal PM_2.5 exposure impaired spatial learning and memory in weanling male mice, but not female mice. The sex-specific outcomes were associated with mRNA expression profiles of the cortex during postnatal critical windows, and the annotations in Gene Ontology (GO) of differentially expressed genes (DEGs) revealed that the exposure persistently disrupted the expression of genes involved in neuronal features in male offspring. Consistently, axonal growth impairment and dendritic complexity reduction were observed. Importantly, Homeobox A5 (Hoxa5), a critical transcription factor regulating all of the neuronal morphogenesis-associated hub genes on PNDs 1, 7, and 21, significantly decreased in the cortex of male offspring following PM_2.5 exposure. In addition, both inorganic and organic components were harmful to axonal and dendritic growth, with organic components exhibiting stronger inhibition than inorganic ones.

Conclusion: Prenatal PM_2.5 exposure affected spatial learning and memory in male mice by disrupting Hoxa5-mediated neuronal morphogenesis, and the organic components, including polycyclic aromatic hydrocarbons (PAHs), posed more adverse effects than the inorganic components.

Keywords: Fine particulate matter (PM2.5) prenatal exposure; Homeobox A5 (Hoxa5); Neuronal morphogenesis; Spatial learning and memory; mRNA expression profiles.

Fig. 1

Effects of prenatal PM_2.5 exposure on spatial learning and memory in weanling offspring. (a) Morris water maze diagram. Learning performance was assessed by (b) escape latency, (c) distance travelled, (d) thigmotaxis, (e) percentage decrease of latency, and (f) swimming speed during the first 5 days. Memory function was assessed by (g) the total frequency across platform zone in the target quadrant, (h) the time in quadrant (%), (i) the number of times crossing the quadrant, and (j) distance travelled in four quadrants during the memory probe test. n = 13–27 from 10 litters

Fig. 2

Analysis of mRNA expression profiles in the cortex of male offspring following prenatal PM_2.5 exposure. (a) Bar graphs showing the number of cortical DEGs at different windows. Chord plot depicting the relationship between DEGs and neurodevelopment-related biological process GO terms on PNDs 1 (b), 7 (c) and 21 (d). Neurodevelopment-related GO terms: biological process (e) and cellular component (f) in male offspring. n = 3 from 3 litters

Fig. 3

Effects of prenatal PM_2.5 exposure on morphological development of axons in male offspring. (a) Representative images of primary cortical neurons on DIV2. Bar = 50 μm; (b) Average neurite length; (c) Axon length; and (d) Average neurite number. n = 96 (Veh) and 95 (PM_2.5) neurons from 3 to 5 litters. (e) Representative images of primary cortical neurons on DIV5. Bar = 50 μm; (f) Average number of axon branches; (g) Average axon branch length at different branch orders. n = 50 (Veh) and 52 (PM_2.5) neurons from 3 to 5 litters

Fig. 4

Effects of prenatal PM_2.5 exposure on morphological development of dendrites and mRNA expression of synaptic markers in male offspring. (a) Representative images of Golgi-Cox-stained dendrites on PNDs 7 and 21. Bar = 200 μm; (b) Sholl analyses of dendritic complexity on PND 7 (n = 21 neurons from 3 mice of different litters); (c) Sholl analyses of dendritic complexity on PND 21 (n = 21 neurons from 3 mice of different litters). (d) Representative images of Golgi-Cox stained dendritic spines on PNDs 7 and 21. Bar = 10 μm. (e) The percentage of each spine morphological category (filopodium, thin, stubby, and mushroom) on PNDs 7 and 21 (n = 15 neurons from 3 mice of different litters). (f) Quantification of dendritic spine density was calculated as the number of spines per 10-µm dendrite length (n = 17–20 neurons from 3 mice of different litters). (g) mRNA expression of Psd95 on PNDs 1, 7, and 21. (h) mRNA expression of Syp on PNDs 1, 7, and 21. (n = 4–6 from 4 to 6 litters)

Fig. 5

Transcriptional regulation of neuronal morphological development-associated DEGs in male offspring following prenatal PM_2.5 exposure. (a) The numbers of DEGs associated with the morphological development of axons, dendrites, and synapses at different windows, with genes selected based on FDR < 0.05. (b-d) PPI of DEGs associated with axons, dendrites, and synapses on PNDs 1, 7 and 21, with inner circles representing the hub genes. (e) The upstream transcriptional factors of the hub genes on PNDs 1, 7 and 21. (f) Relative mRNA levels of C/EBPbeta, cFos and Hoxa5 on PNDs 1, 7 and 21 (n = 4–6 from 4–6 litters). (g) Immunofluorescence staining of HOXA5 (red fluorescence) on PNDs 1, 7 and 21, with neurons stained with NeuN (green fluorescence) and nuclei fluorescence) and nuclei stained with DAPI (blue fluorescence) (n = 3 from 3 litters), Bar = 50 μm. (h) Quantification of the immunofluorescence intensity of Hoxa5 on PNDs 1, 7 and 21. (i) A diagram showing the binding sites of Hoxa5 with the hub genes on PNDs 1, 7 and 21 using the JASPAR database

Fig. 6

Effects of inorganic components (IC) and organic components (OC) of PM_2.5 on neuronal morphology and gene expression. (a and b) The cytotoxicity of IC and OC to primary cultured cortical neurons. (c) Representative images of primary cortical neurons on DIV2. Bar = 50 μm; (d-f) The average neurite length, axon length, and average neurite number. n = 97 (Veh), 97 (IC) and 97 (OC) neurons from 3 to 5 litters. (g) Representative images of primary cortical neurons on DIV5. Bar = 50 μm; (h and i) The average number of axon branches and the average length of axon branches at different branches. n = 52 (Veh), 52 (IC) and 52 (OC) neurons from 3 to 5 litters. (j) mRNA expression of genes related to neuronal morphological development (n = 3 from 3 litters). Data are presented as mean ± SEM from three independent experiments. *p < 0.05, **p < 0.01 for comparisons between the PM_2.5 exposure group and the vehicle group; ^#p < 0.05, ^##p < 0.01 for comparisons between the IC group and the OC group