Rapid Proliferation of Dentate Gyrus Progenitors in the Ferret Hippocampus following Neonatal Lipopolysaccharide Exposure Open Access

Maternal immune activation, induced by bacterial and viral infections during pregnancy, has a pivotal influence on brain development. It increases the risk of neurodevelopmental disorders such as autism spectrum disorder (ASD), cognitive dysfunction, and attention-deficit/hyperactivity disorder [1‒4]. Neonatal infection in rodents is known to alter brain function, causing dysfunction of adult hippocampal neurogenesis in relevant to neuroinflammation-induced depression [5, 6]. The dentate gyrus (DG) of the hippocampus develops late ontogenetically and serves as a neurogenic niche for generating adult neural stem cells during the early postnatal period [7, 8]. Neural progenitors in the hippocampal DG transition from the embryonic to the fatty acid-binding protein (BLBP)-expressing adult type immediately after birth in mice [9]. Therefore, neonatal infections may alter the proliferation of DG progenitors and their subsequent transition to the adult type.

Lipopolysaccharide (LPS), an endotoxin of Gram-negative bacteria, plays a role in initiating innate immune responses [10, 11] and activates microglia releasing the proinflammatory and anti-inflammatory cytokines such as tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and interleukin-10 (IL-10) [10, 12]. LPS also regulates the proliferation and differentiation of neural stem cells/progenitors via the activation of toll-like receptor 4 (TLR4) [13, 14]. Notably, TLR4 knockout enhanced the proliferation and differentiation of adult hippocampal neural stem cells [15, 16]. This study aimed to elucidate the neurogenic effects of LPS on DG progenitors in the developing hippocampus. The hippocampal DG develops during mid- to late-gestational period in humans [17] and late-gestational to early postnatal periods in rodents [8] and ferrets [18]. Here, we used ferret neonates to investigate the acute effect of LPS on DG progenitors by administrating LPS directly to their pups. Using ferrets further allowed for easy evaluation of the density and immunohistochemical characteristics of the DG progenitors because they appeared abundantly in the hippocampal DG of ferrets at early postnatal ages [18].

Six male ferret infants, naturally delivered from 5 pregnant ferrets purchased from Japan SLC (Hamamatsu, Japan), were used in this study. The infants were reared with lactating dams (3–5 pups/dam) in stainless-steel cages (80 cm × 50 cm × 35 cm) maintained at 21.5 ± 2.5°C under 12-h artificial illumination in the Facility of Animal Breeding, Nakaizu Laboratory, Japan SLC. All lactating dams were fed a pellet diet (High-Density Ferret Diet 5L14; PMI Feeds, Inc., St. Louis, MO, USA) and provided with tap water ad libitum.

Three infants were given a subcutaneous injection of LPS at 500 µg/g body weight on postnatal days 6 and 7. The remaining three pups that were not exposed to LPS served as controls. To label the proliferating cells, 5-bromo-2-deoxyuridine (BrdU; Sigma-Aldrich, St. Louis, MO, USA) was administered intraperitoneally at 30 μg/g body weight to all ferret infants simultaneously with the second LPS injection. Two hours after BrdU injection, all animals were perfused with 4% paraformaldehyde in phosphate-buffered saline (PBS) under deep anesthesia with ∼2% isoflurane gas.

Cerebral hemispheres were immersed in a 30% sucrose-PBS solution overnight and then embedded in an optimal cutting temperature compound (Sakura Finetek Japan Co., Ltd., Japan) at −70°C. Coronal cryosections of the hemispheres at 100-µm thickness were cut using a Retratome (REM-700; Yamato Koki Industrial Co. Ltd., Asaka, Japan) equipped with a refrigeration unit (Electro Freeze MC-802A; Yamato Koki Industrial). All sections were collected in vials containing 4% paraformaldehyde in PBS.

Serial coronal sections of the dorsal hippocampus were subjected to immunofluorescence analysis. All immunohistochemical procedures were performed infloating mode. Sections were heated in Antigen Unmasking Solution H-3300 (pH 6.0; Vector Labs Inc., Burlingame, CA, USA) for 30 min in a 90°C water bath and then cooled at 4°C for 30 min. After preincubation with PBS containing 0.1% Triton-X 100 at 37°C for 1 h, immunofluorescence staining was performed using the primary and secondary antibodies listed in online supplementary Tables S1 and S2 (for all online suppl. material, see https://doi.org/10.1159/000546709). The primary antibodies used were highly specific for the ferret brain tissue [18‒20].

Serial digital sectioning images at a 10-μm depth (section plane thickness: 1 μm; number of sections: 10) were acquired from the most superficial plane, where BrdU labeling and various immunostained markers were detectable. Images were captured using an Axio Imager M2 ApoTome.2 microscope with a ×20 objective, equipped with an AxioCam MRm camera (Zeiss, Gottingen, Germany) and Zen 2.3 blue edition software (Zeiss). A set of sectional images, 4-μm apart in the Z-direction (seventh and third from the superficial slices of the acquired images), were selected as the lookup and reference images, respectively. The dissector method using systematic random sampling was used to estimate the density of BrdU-labeled and immunostained cells, according to a previous report [18]. In the sections immunostained for each marker, frames with 12–24 square boxes (box size, 40 μm × 40 μm) were used from one section of both the left and right hemispheres. These frames systematically select the region of interest (ROI), which was randomly superimposed on the granular and subgranular layers (GL/SGL) and the hilus of the hippocampus in both the lookup and reference images. BrdU-labeled or immunostained cells were counted within the ROIs using the “forbidden line” rule. Their densities were calculated using the following formula: (cell density = Qn−/[a × b × t]) (Qn− = total number of BrdU-labeled and/or immunostained cells appearing within ROIs in the lookup images but not in the reference images, a = 12 to 24, which is the total number of ROIs in the lookup images from two sections [the left and right hemispheres] per animal, b = 40 μm × 40 μm areas of counting box, and t = 4 μm, which is the distance between the lookup and reference images). The percentage of BrdU-labeled and immunostained cells was estimated by adding the number of cells counted within all ROIs from all animals in the LPS-exposed and the control groups. The overlapping expression of each antigen was unclear because the percentage was estimated independently for each marker antigen.

Measurements from the left and right hemispheres were combined and the number of animals (n) was set to “3” in each group. Significant differences in BrdU-labeled and immunostained cell densities were statistically assessed using the repeated-measures two-way analysis of variance (ANOVA), with “region” (GL/SGL and hilus) and “group” (LPS-exposed and control groups) as factors. For post hoc testing, Scheffé’s test was performed when significant differences in group and/or region × group interactions were detected using a two-way repeated-measures ANOVA and simple main effects. The percentage of BrdU-labeled cells immunolabeled for markers was statistically assessed using the chi-square test. The total number of BrdU-labeled cells was defined as “n” for the chi-square test.

In this paper, we use the terms “immunopositive” or “positive labeling” (+) to refer to cells labeled with BrdU or marker antigens at high and medium levels. Figure 1a shows the immunofluorescence images of calbindin D28k with BrdU labeling in the dorsal hippocampus of LPS-exposed and control ferrets 2 h after BrdU injection. The histoarchitecture of the ferret hippocampus of neonates was depicted by calbindin D28k immunostaining that appeared in the DG granular and pyramidal neurons of the CA1 to CA4 fields [21]. There was no obvious difference in hippocampal histoarchitecture between LPS-exposed and control ferrets (Fig. 1a). BrdU+ cells were distributed throughout the molecular layer, GL/SGL, and hilus of the DG (Fig. 1b). Repeated-measures two-way ANOVA revealed a significant region-specific effect of LPS on BrdU+ cell density in the hippocampal DG of ferret neonates (F_{[1, 4]} = 14.668, p < 0.05). Post hoc testing indicated a significantly greater density of BrdU+ cells in the hilus (p < 0.001), but not in the GL/SGL, of LPS-exposed ferrets than in that of controls (Fig. 1c).

To identify the types of BrdU+ cells, we performed BrdU labeling with immunofluorescence staining using various marker antigens. The proportion of cells immunostained for the marker antigens was estimated in BrdU+ cells in the GL/SGL and hilus of the ferret neonates. No calbindin D28k+ granular neurons were labeled with BrdU in either the LPS-exposed or control ferrets (Fig. 1b; Table 1). In addition, BrdU+ cells were immunopositive for Sox2, a marker for DG progenitors [18], in a great majority (Fig. 2a; Table 1).

BrdU+ cells were immunostained for BLBP, a marker for adult-type DG progenitors [9], at 47.2% in the GL/SGL and 35.1% in the hilus of control ferrets (Fig. 2b; Table 1). The BLBP+ ratio of BrdU+ cells in the hippocampal DG of LPS-exposed ferrets was 18.5% in the GL/SGL and 14.9% in the hilus, which was significantly lower than that in controls (Fig. 2b; Table 1). Approximately half of the BrdU+ cells were immunopositive for S100, a marker for glial cells [22], in the GL/SGL and hilus of control ferrets (Fig. 2c; Table 1). Although the S100+ ratio of BrdU+ cells in LPS-exposed ferrets was 51.8% in the GL/SGL, which was not different from that in the controls, the ratio was 25.0% in the hilus of LPS-exposed ferrets, which was significantly lower than that in the controls (Fig. 2c; Table 1).

More than 76% of the BrdU+ cells were immunopositive for proliferating cell nuclear antigen (PCNA), a proliferation marker [23], in both the GL/SGL and hilus, without a statistical difference between LPS-exposed and control ferrets (Fig. 3a; Table 1). Approximately 35% of the BrdU+ cells were immunostained for cleaved caspase 3 (cCasp3), a marker of apoptosis [24], in the hilus of control ferrets (Fig. 3b; Table 1). A significantly lower ratio of cCasp3+ to BrdU+ cells was observed in the hilus of LPS-exposed ferrets (17.3%) than in that of the controls (Fig. 3b; Table 1).

Next, the total density of cells immunopositive for each marker antigen was estimated and compared between LPS-exposed and control ferrets. The total densities of calbindin D28k+, Sox2+, BLBP+, S100+, PCNA+, and cCasp3+ cells did not differ between the two groups, except for the absence of calbindin D28k+ cells in the hilus (Fig. 4).

The present findings reveal that LPS facilitates the proliferation of DG progenitors, consistent with other progenitors, such as human neural stem cells derived from the diencephalic and telencephalic brain [13] and subventricular zone progenitors for the cerebral cortex [14, 25]. LPS attenuates the expressions of BLBP and S100 in post-proliferative DG progenitors with sustained Sox2 expression in the GL/SLG and/or hilus. The hippocampal DG serves as a neurogenic niche for the generation of adult neural stem cells during the early postnatal period in mice [7, 8]. BLBP is expressed in adult-type DG progenitors, including adult neural stem cells, which are transformed from BLBP-deficient embryonic DG progenitors [9]. Therefore, LPS may act to sustain embryonic-type characteristics of post-proliferative DG progenitors by preventing their transitions into adult-type and/or glial cell lineages. The effect of LPS on DG progenitors may vary between the embryonic and adult types. The proliferation and differentiation of adult hippocampal neural stem cells decreased immediately following LPS exposure [26] and were enhanced by TLR4 knockout [15, 16].

The current investigation further revealed that the apoptosis of post-proliferative DG progenitors was inhibited immediately after LPS exposure. This effect of LPS may be evoked via TLR4 but not via proinflammatory or anti-inflammatory cytokines released from microglia. LPS-mediated TLR4 activation prevents apoptosis of human neural stem cells in vitro [13]. Thus, LPS may act via TLR4 to prevent programmed cell death that occurs immediately following the proliferation of DG progenitors, as one of the LPS actions on sustaining embryonic-type characteristics of DG progenitors, such as reduced BLBP and S100 expressions in this study. On the contrary, neural apoptosis was induced by the proinflammatory cytokines released from activated microglia such as TNF-α and IL-6 [27, 28]. Notably, the survival of DG progenitors was shortened when exposed to LPS during proliferation [29]. Such long-term effect of LPS may be involved in prolonged alterations in the hippocampal inflammatory status by persistent dysregulation of genes associated with M1 and M2 microglial phenotypes [30]. Thus, the effect of LPS on the DG progenitor population may be biphasic: an early response to promote proliferation and prevent apoptosis via TLR4 activation and a delayed response to enhance apoptosis via proinflammatory cytokines released from persistently altered microglia.

Neonatal exposure to LPS causes depression-like behaviors in adult rodents [5, 6]. Altered adult neurogenesis in the hippocampus is a key factor causing major depressive disorders [5, 31]. Thus, the TLR4-mediated early response of DG progenitors to LPS, as observed in the current study, may not be essential for LPS-induced depression. Additionally, neonatal LPS exposure impairs social and cognitive behaviors in rats [5]. ASD-like social impairments are induced in ferrets by neonatal exposure to valproic acid [32], which enhances the proliferation of DG progenitors [18]. Therefore, the TLR4-mediated effects of LPS on DG progenitor proliferation may be associated with ASD pathogenesis. Further studies are required to elucidate the involvement of TLR-mediated LPS in ASD-like social impairments.

All experimental procedures were approved by the Institutional Animal Care and Use Committee of Tsukuba International University (approval code: 30-1). We also attempted to minimize the number and suffering of animals used in this study.

The authors declare no conflicts of interest.

This work was supported in part by the Japan Society for the Promotion of Science (JSPS) KAKENHI (grant No. 21K06741, to Kazuhiko Sawada).

All authors had full access to all data in the study and take full responsibility for the integrity of the data and accuracy of the data analysis: K.S.: conception and design of the study, drafting the article, and obtaining funds and K.S. and S.K.: acquisition of data, analysis and interpretation, and final approval of the version to be submitted.

The original contributions of this study are included in the article and online supplementary material. Further inquiries can be directed to the corresponding author.