Three-month-old female Hartley guinea pigs were obtained from and maintained in the Center for Laboratory Animals of Xiangya School of Medicine during all the experimental procedures before brain perfusion. Animals were housed in a vivarium with controlled temperature, humidity, and lighting (12:12-h light–dark cycle), with free access to standard laboratory chow and water. Fruit or fresh vegetables were also provided daily. The use of animals complied with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. All experimental procedures were approved by the Ethics Committee of Xiangya School of Medicine.

Animals were randomly assigned to a sham operation group (n = 8) and an experimental group receiving different dosages of ¹²⁵I seeds local radiation. In pilot studies, we implanted unilaterally three 0.8-mCi ¹²⁵I seeds between the scalp and skull above the temporal lobe for 60 days, which were found to have caused too severe neurotoxic injury to mature neurons. In the final experiments, 18 guinea pigs were randomly divided into six groups: A1, B1, C1, A2, B2, and C2 (n = 3). We administered 0.3 mCi of extracranial local radiation in group A1, 0.6 mCi in group B1, and 1.0 mCi in group C1 for 30 days. We also administered 0.3 mCi of extracranial local radiation in group A2, 0.6 mCi in group B2, and 1.0 mCi in group C2 for 60 days. The ¹²⁵I tubes were all prescription-ready products used for clinical brachytherapy.

We shaved the hair of the female guinea pigs in the operative region before surgery. Animals were then placed in a stereotaxic apparatus and were under anesthesia after injection with sodium pentobarbital [30 mg/kg, intraperitoneally (i.p.)]. After disinfection of the operation areas, we performed an incision with sterile instruments (Figure 1A) to find the bregma (Figure 1B). ¹²⁵I seeds were implanted and fixed with superglue on the skull surface 5 mm posterior relative to the bregma, 1 mm lateral to the sagittal suture, which corresponds to the middle part of the dorsal hippocampus. In the sham group, we implanted a titanium tube without the ¹²⁵I nuclide core, whereas the other operation procedures were identical to the experimental groups. After 30 or 60 days of extracranial radiation, animals were perfused transcardially, and the skulls were carefully inspected for any shedding titanium tubes (Figures 1C,D). All tubes in the operated animals were retrieved in proper collection apparatus (Figure 1).

For immunohistochemical experiments, animals were perfused via the ascending aorta with 4% paraformaldehyde in 0.01 M phosphate-buffered saline (pH 7.4, PBS) under overdose anesthesia (sodium pentobarbital, 100 mg/kg, i.p.). The whole brain was removed from the cranium and postfixed in the perfusion fixative for 2 days. The brains were transferred into solutions containing 15% and then 30% sucrose for cryoprotection. Before sectioning, a small portion of the surface cortex was removed from the contralateral cortex relative to the implanted side, for an orientation during microscopic examination (Figure 1I). The cerebrum was cut frontally in a cryostat at 30-μm thickness, with 24 sets of sections passing the rostrocaudal dimension of the hippocampus in all animals. The sections were collected serially in PBS in cell culture plates. Thus, each well contained equally distant (∼30 × 24 = 720 μm) sections from the septal to the temporo-occipital levels of the hippocampus in both hemispheres (Figure 1J).

From each brain, separate sets of hippocampal sections were immunohistochemically stained with the avidin–biotin complex (ABC) method using previously characterized commercial primary antibodies, including Ki67, doublecortin (DCX), and neuron-specific nuclear antigen (NeuN) (Xiong et al., 2008; Yang et al., 2015; Wan et al., 2019). The sections were first treated in PBS containing 5% H₂O₂ for 30 min and in PBS containing 5% normal horse serum with 0.3% Triton X-100 for 2 h, to minimize non-specific labeling. The sections were then incubated with rabbit anti-Ki67 (Vector Laboratories, Burlingame, CA, United States; #014-1107, diluted at 1:1,000), goat anti-DCX (Santa Cruz Biotech.; sc-8066, 1:1,000), or NeuN (Abcam; #ab134014,1:2,000), overnight at 4°C with gentle agitation on a rotator. After three washes with PBS, the sections were incubated with biotinylated pan-specific secondary antibodies (horse anti-mouse, rabbit, and goat IgGs) at 1:400 for 2 h and ABC reagents (1:400) (Vector Laboratories) for another 2 h. The immunoreactive product was visualized with 0.003% H₂O₂ and 0.05% 3,3’-diaminobenzidine (DAB). We washed the sections three times with PBS for 10 min between the incubations. The immunolabeled sections were placed on glass microslides, dehydrated with ascending concentrations of ethanol, cleared with xylene, and mounted with coverslips using a mounting medium. The Ki67⁺ and DCX⁺ immunolabeled sections were also counterstained with hematoxylin for the histological examination under a microscope.

Sections were amounted onto gelatinized glass slides and air-dried at room temperature. Following several minutes in PBS, the sections were rinsed with distilled water and stained using a Nissl staining kit (1% cresyl violet staining, Solarbio, #G1430, China) for 60 min at 56°C in the dark. The samples were washed again with distilled water and differentiated in a Nissl differentiation solution until the background was light purple or colorless. The slides were dehydrated with 75, 80, and 90% ethanol sequentially for 2 min each, with absolute ethanol for 5 min two times, and with xylene for 10 min two times. They were then sealed with neutral gum and stored in the dark.

The immunolabeled sections were examined initially on an Olympus BX51 microscope (CellSens Standard; Olympus Corporation, Japan) to validate the immunolabeling quality. All sections were scanned using the 20 × objective of a Motic-Olympus microscope equipped with an automated stage and imaging system (Wuhan, Hubei, China), which could yield a final auto-focused, montaged, and magnification-adjustable image covering the entire area of a glass slide.

Low-magnification (2× and 2.5×) and high-magnification (10×, 40×) images over the area of interest (AOI) were extracted from the images for figure illustration. For the sections immunolabeled with DAB as the chromogen, the entire area of hippocampal formation was imaged using 2× objective for orientation and measurement of the length of the granule cell layer (GCL) using the NIH ImageJ software. We then counted Ki67⁺, DCX⁺ cells along the subgranular zone (SGZ) in the dorsal dentate gyrus (DG) (Figure 1K). For all the measurements, the AOI was defined as a band region equal to the thickness of the GCL but centered over the SGZ (Figures 1L,M). Thus, we opened the images with Photoshop software and drew two lines along the upper and low blades of the GCL (Figure 1K). This lined template was then moved toward the hilus by half of the distance of the GCL thickness. After enlarging the images on the computer screen, we counted the numbers of Ki67⁺ and DCX⁺ cells and calculated the sums for each hippocampal section. The length of the GCL was calculated by referring to the scale bar. Finally, the cell density, expressed as the number of cells per unit (mm) length of GCL, was calculated for each brain, using the data obtained from the second to the fifth equally spaced (rostral-to-caudal direction, Figure 1J) hippocampal sections from both hemispheres (the first section at the septal end showed no or only a small cross-sectional area of the DG and was therefore not included in the quantification). We used the dorsal hippocampus for quantification to keep the sampling region consistent.

Low-magnification (2.5×) and high-magnification (10×, 60×) images over the AOI were extracted from the images for figure illustration. These images were used to quantify the number of NeuN⁺ cells in the hippocampus. The GCL of the dorsal hippocampus and the stratum pyramidale (s.p.) in the CA1 region were randomly sampled using a method described recently (Ai et al., 2021). Thus, four or five 100 μm by 100 μm (10,000 μm²) squares in the GCL and the s.p were tagged. We then enlarged the images on the computer screen and counted the NeuN⁺ cells in the squares. The sums were calculated for each hippocampal section. The NeuN⁺ cell density, expressed as the number of cells per unit area (mm²) in the GCL and s.p., was calculated for each brain based on data obtained from the second to the fifth equally spaced (Figure 1J) sections from both hemispheres. As with the analysis of Ki67⁺ and DCX⁺ cells, we used the dorsal hippocampus for quantification to keep the sampling region consistent.

The dosage of radioactivity absorbed by the tissues from the initial source activity with different irradiation times was calculated with the following equation (Jones et al., 1995; Yu et al., 1999; Wang et al., 2017):

D_r = S_k∧[G(r,θ₀)/G(r₀,θ₀)]g(r)φ_an(r)

In this equation, S_k represents the air kerma strength, φ_an(r) is the ratio of the dose rate at distance r, G (r, θ₀) represents the geometry factor, and g(r) is the radial dose function. The absorbed radioactive dosages at the reference point varied among the experimental groups with the initial source activity, the distance from the ¹²⁵I seeds, and the follow-up times. In our research, we detected using a stereotactic instrument that the midpoint of the ¹²⁵I seed radiation source was about 5.0 mm from the middle part of the ipsilateral hippocampus relative to the implanted side, and 3.25 mm from the sagittal suture of the skull. The levels of absorbed dosages in the hippocampus in regions located 5.0 and 8.2 mm transverse from the midpoint of the ¹²⁵I seeds are summarized in Table 1.

Cell counting was performed by experimenters who were blinded to the animal grouping information, and the data were gathered by the primary experimenters for the quantitative and statistical analyses. The numeric density (number of cells/mm GCL length and number of cells/mm² GCL and CA1 areas) were calculated in Excel spreadsheets and expressed as mean ± standard derivation (SD). We entered the data from individual animals into Prism spreadsheets (GraphPad Prism 5, San Diego, CA, United States) under the corresponding groups and generated graphs accordingly. Statistical analyses were performed to compare the means between the groups using one-way analysis of variance (ANOVA) tests together with Bonferroni multiple-comparisons post hoc tests. The figures were assembled with Photoshop 7.1 and converted into TIFF files after adjusting the contrast and brightness in the entire panel as needed.

DCX⁺ labeling in neuronal somata and dendrites was microscopically prominent to observe (as compared to the Ki67 labeling in cell nuclei). Therefore, for an overall assessment of regional effect of radiation, we compared the amount of DCX⁺ labeling in the cerebrum across the rostrocaudal dimension and also between the two hemispheres in experimental animals relative to sham control. For reference, the distribution of DCX⁺ cells over the entire cerebrum is illustrated in Figure 1. Thus, as documented previously (Xiong et al., 2008, 2010), DCX⁺ cells occurred in the SVZ surrounding the lateral ventricle (LV) and the SGZ of the DG (Figures 2A–C,F–I), the two established principal sites of adult neurogenesis in mammalian brain. In addition, DCX⁺ cells were distributed along layer II across the piriform cortex, as well as neocortical regions (Figures 2A,D,E,H,I). It should be emphasized that the overall amount and distribution of DCX⁺ labeling around the LV, in the hippocampal formation, and across the cerebral cortex appeared in a symmetric manner between the two hemispheres, either viewed at low magnification (Figure 1A, arrows) or examined by closer observation (Figures 2B–I). Notably, different from primates and some non-primate species (Xiong et al., 2008; Cai et al., 2009; Zhang et al., 2009’ Piumatti et al., 2018; Ai et al., 2021), DCX⁺ immature neurons in the amygdala were not prominently present in guinea pigs (Figures 2A,D,E).

As illustrated, for example, in the cerebral sections from an animal with unilateral extracranial ¹²⁵I seed radiation at 0.6 mCi surviving 30 days, the amount of DCX⁺ immunolabeling along the SGZ of the DG appeared dramatically asymmetric between the two hemispheres. However, the labeling at the SVZ at the septal level appeared largely comparable between the two sides (Figure 3). Specifically, DCX⁺ labeling was greatly reduced in the ipsilateral relative to the contralateral DG, which were evident essentially across the entire rostral to caudal range of the hippocampal formation. Further, the reduced labeling in the ipsilateral side could be clearly observed over the dorsal and temporal subregions of the DG over the entire rostrocaudal range of the hippocampus (Figures 3A,D,I). It should be noted that the reduction of DCX⁺ profiles in the most ventral part of the ipsilateral DG appeared less prominent relative to the contralateral counterpart (Figures 3A,J,K). Furthermore, in comparison with the sections from age-matched shame control (Figures 2A,F–I), it appeared that the overall amount of DCX⁺ labeling in the SGZ in the ipsilateral hippocampus from the radiated animal was also reduced to a certain degree (Figures 3A,G,I,K).

Microscopically, the overall amount of Ki67⁺ cells in the SGZ was reduced in all the radiation animal groups relative to sham control, which appeared more evident in the ipsilateral side in general (Figures 2, 3). This change was visible in the sections from animals subjected to ¹²⁵I seed implantation surviving 30 days (Figure 4) as well as 60 days (Figure 5). The numeric densities, expressed as mean ± SD, of the subgranular Ki67⁺ cells in the dorsal hippocampus were measured for statistical analysis. The values in the animals surviving 30 days were 6.7 ± 0.7, 6.4 ± 0.6, 6.0 ± 0.6, 5.4 ± 0.5, 4.3 ± 0.5, 3.7 ± 0.4, and 1.9 ± 0.3 cells/mm of GCL length in the sham, 0.3 mCi-contral., 0.6 mCi-contral., 1.0 mCi-contral., 0.3 mCi-ipsil., 0.6 mCi-ipsil., and 1.0 mCi-ipsil. groups, respectively (Figure 6A). In the animals surviving 60 days, the numeric densities were 6.8 ± 0.6, 6.1 ± 0.7, 5.9 ± 0.5, 4.8 ± 0.7, 4.1 ± 0.2, 3.8 ± 0.7, and 1.2 ± 0.2 cells/mm in the sham, 0.3 mCi-contral., 0.6 mCi-contral., 1.0 mCi-contral., 0.3 mCi-ipsil., 0.6 mCi-ipsil., and 1.0 mCi-ipsil. groups, respectively (Figure 6A). Non-parametric tests (one-way ANOVA and Dunn multiple-comparisons tests) indicated that the mean densities were different among the different dosage groups surviving 30 and 60 days (30 days: p < 0.0001, F = 67.88; 60 days: p < 0.0001, F = 97.07). Post hoc tests reported a statistically significant difference between the sham group and the 1.0 mCi-contral., 0.3 mCi-ipsil., 0.6 mCi-ipsil., and 1.0 mCi-ipsil. groups at 30 and 60 days postsurgery, respectively (Figure 6A).

In microscopic examination, the overall amount of DCX-labeled profiles in the DG was overtly reduced in the sections from animals that received ¹²⁵I radiation relative to sham controls, with this change appearing more dramatic in the ipsilateral side in reference to the implants (Figures 7, 8). At high magnification, the dendritic processes of DCX⁺ immature neurons also appeared fewer and shorter in radiated groups relative to sham controls and in the ipsilateral relative to contralateral hippocampi within the radiated groups (Figures 5B,C,E,F,H,I,K,L,N,O,Q,R,T,U, 7B,C,E,F, H,I,K,L,N,O,Q,R,T,U, 8B,C,E,F,H,I,K,L,N,O,Q,R,T,U). A dose-related impact on the morphology of subgranular DCX⁺ cells was also noticeable, such that the remaining labeled immature neurons appeared smaller, had few or no dendrites entering the molecular layer, and were more sparsely aligned along the SGZ in the higher dosage relative to lower dosage groups (Figures 7, 8, right panel images).

As with the aforementioned quantification of Ki67⁺ cells, we measured densities of the subgranular DCX⁺ cells per unit length of the GCL, given that their somata resided along the SGZ in a linear fashion. The densities of DCX⁺ cells were estimated to be 44.2 ± 4.9, 29.5 ± 3.3, 22.1 ± 2.5, 6.3 ± 0.7, 7.4 ± 0.8, 3.7 ± 0.4, and 1.7 ± 0.6 cells/mm in the sham, 0.3 mCi-contral., 0.6 mCi-contral., 1.0 mCi-contral., 0.3 mCi-ipsil., 0.6 mCi-ipsil., and 1.0 mCi-ipsil. groups surviving 30 days postsurgery, respectively. The values were 42.6 ± 4.5, 14.2 ± 1.5, 10.6 ± 1.1, 2.0 ± 0.2, 5.3 ± 0.6, 2.1 ± 0.2, and 0.4 ± 0.7 cells/mm in the sham, 0.3 mCi-contral., 0.6 mCi-contral., 1.0 mCi-contral., 0.3 mCi-ipsil., 0.6 mCi-ipsil., and 1.0 mCi-ipsil. groups at 60 days postsurgery, respectively. Statistically (one-way ANOVA), there existed significant differences of the means between the 30- and 60-day surviving groups (30 days: p < 0.0001, F = 157.7; 60 days: p < 0.0001, F = 252.3). Post hoc tests showed a significant difference of the means in the sham animal groups relative to individual radiated animal groups at both the 30- and 60-day surviving time points (Figure 6B).

We assessed potential radiation effect on NeuN⁺ cells and Nissl stain cells in the GCL of DG and the s.p. of CA1. Again, we chose the CA1 region of dorsal hippocampus for a consistent quantitative analysis, considering it is closer to the radiation site than CA3. We also analyzed in the GCL of dorsal DG for observational and quantitative analysis. As mentioned earlier, in our preliminary experiments, we noticed microscopically evident neurotoxic damage to mature hippocampal neurons based on NeuN and Nissl stain (Supplementary Figures 2, 3). We expected that lower radiation dosages used in the formal experiments would cause less or no apparent injurious effect on NeuN⁺ neurons and Nissl-stained cells.

Indeed, there were no overt microscopic changes regarding the laminar organization and thickness of the CA1 region and DG while examining the NeuN and Nissl-stained sections from animals with 0. 3-, 0. 6-, and 1.0-mCi radiation surviving 30 and 60 days, relative to sham controls (Figure 9 and Supplementary Figures 4–7). Specifically, no disruption of cellular arrangement, irregularity of staining intensity, and nuclear breakage or condensation were visible in the NeuN⁺ neuron (Figure 9 and Supplementary Figures 4, 5) and Nissl-stained profiles (Supplementary Figures 6, 7) in the s.p. and GCL in both the ipsilateral and the contralateral sides in the radiated animals in comparison with controls. However, by closer examination, the NeuN⁺ nuclei in the s.p. and GCL appeared to be less densely packed in the radiated animals with 1.0-mCi treatment surviving 30 and 60 days, relative to sham control (Figures 9D,H,L,P,T,X).

Quantitatively, the numeric densities (mean ± SD, number of cells per square mm) of NeuN⁺ cells were 7,675.0 ± 153.5, 7,583.3 ± 145.7, 7,523.3 ± 206.5, 7,410.0 ± 200.8, 7,496.7 ± 231.6, 7,336.7 ± 220.1, and 6,876.7 ± 280.4 in the GCL in the sham, 0.3 mCi-contral., 0.6 mCi-contral., 1.0 mCi-contral., 0.3 mCi-ipsil., 0.6 mCi-ipsil., and 1.0 mCi-ipsil. groups surviving 30 days, respectively. The densities of the NeuN⁺ cells in the s.p. of CA1 were 5,123.3 ± 220.1, 4,986.7 ± 179.3, 4,870 ± 285.1, 4,690 ± 175.8, 4,780 ± 183.6, 4,653.3 ± 225.0, and 4,063.3 ± 190.4 cells/mm² of s.p. area in the sham, 0.3 mCi-contral., 0.6 mCi-contral., 1.0 mCi-contral., 0.3 mCi-ipsil., 0.6 mCi-ipsil., and 1.0 mCi-ipsil. groups surviving 30 days, respectively. The density values for the 60-day surviving animals were 7,622.5 ± 204.2, 7,543.3 ± 214.6, 7,436.7 ± 309.9, 7,276.7 ± 165.0, 7,363.3 ± 180.1, 7,153.3 ± 195.0, and 6,767.6 ± 133.2 cells/mm² in the GCL, and were 5,202.5 ± 212.2, 4,993.3 ± 211.3, 4,820 ± 262.1, 4,743.3 ± 215.0, 4,733.3 ± 181.5, 4,540 ± 251.2, and 3,903.3 ± 270.3 cells/mm² in s.p. of CA1, in the sham, 0.3 mCi-contral., 0.6 mCi-contral., 1.0 mCi-contral., 0.3 mCi-ipsil., 0.6 mCi-ipsil., and 1.0 mCi-ipsil. groups, respectively. Statistically, there was a significant difference in the means between the groups at 30- and 60-day time points (CA1-30d: p = 0.0003, F = 8.94; DG-30d: p = 0.0003, F = 5.45; CA1-60d: p < 0.0001, F = 16.20; DG-60d: p = 0.0009, F = 7.29). Post hoc tests indicated a significant difference for the sham groups in comparison with the 0.6 mCi-ipsil. group of CA1-60d, the 1.0 mCi-ipsil. group of CA1-30d, DG-30d, CA1-60d, and DG-60d, whereas no difference existed for the remaining paired groups (Figures 6C,D).

The current study used ¹²⁵I isotope-filled tubes of the identical physical property but eliciting varying doses of radioactivity. The radioactivity was considered as the ultimate factor that led to the aforementioned changes in Ki67⁺ and DCX⁺ cells evidently in the hippocampal DG and, to a lesser extent, the changes noticeable in NeuN⁺ mature neurons and Nissl-stained profiles. By examining the extent of local tissue trauma (Lieberwirth et al., 2016), one could further cross-validate the existence of tissue injury that should be radioactivity-dependent. Therefore, we documented the cranial surface images over the implantation sites after animal perfusion and before brain removal. Overall, ¹²⁵I seeds with lower radioactivity (e.g., 30 mCi), or shorter animal surviving time (e.g., the 30 days groups), were found to have milder local skull trauma (data not shown). Shown, for example (Figures 10A–D), in animals with 0.6-mCi radiation surviving 60 days, a fairly small and shallow depression (∼0.2 mm in diameter) of the cranial surface was found at the site of ¹²⁵I tube implantation. In an animal with 1.0-mCi radiation surviving for 60 days, a larger and noticeably deeper skull depression (∼0.4 mm in diameter) was observed at the implantation sites (Figures 10E–H). It should be noted that in all animals included in the formal experiments in the current study, the skull damage did not penetrate the cranial bone (the parietal bone). However, in the animal with three 0.8-mCi seed implants that survived for 60 days, the skull damage was significant, with the cranial depression approximately 1.5 mm in size and the underneath brain tissue visible (Figure 1G).

¹²⁵I is a radioisotope that emits γ rays with a low average photon energy of 28.5 keV, with a half-life of approximately 60 days (Krishnaswamy,1978). As elaborated in the Introduction, ¹²⁵I seeds are commonly used in brachytherapy of many solid tumors (Schwarz et al., 2012; Wang et al., 2016; Zhang et al., 2016; Hirose et al., 2017; Brun et al., 2018; Vigneault et al., 2018; Zhu et al., 2019; Jia et al., 2020; Li et al., 2020; Watson et al., 2020). Given that radiation in general can result in inhibition or elimination of neurogenesis by blocking cell proliferation (Mizumatsu et al., 2003; Tan et al., 2011), we sort to establish a method to block AHN via localized extracranial application of ¹²⁵I seeds using guinea pigs as experimental model. Guinea pigs are widely available and economically affordable laboratory animals. They have a larger brain than small laboratory rodents (e.g., mice and rats) and therefore could be better to allow the analysis of a space-related effect of radiation in the brain. As with mice and rats, DCX⁺ neuroblasts and immature neurons are present at the SVZ and SGZ relating to adult neurogenesis. DCX⁺ immature neurons are also distributed around layer II of the piriform cortex and the entire neocortex (Xiong et al., 2008; Bonfanti and Nacher, 2012; La Rosa et al., 2020). Therefore, this animal could be used to explore the response of different types of cerebral DCX⁺ cells to radiation (Xiong et al., 2010). The present study aimed to establish a “proof of principle” of a new and relatively simple approach for radioablation of adult neurogenesis in mammalian brain. Therefore, we focused on the hippocampal region and evaluated the effect by microscopic analysis of cellular profiles representing proliferative cells (Ki67) and developing/differentiating neurons (DCX) (Del Bigio, 1999; Kee et al., 2002). To identify an effective radiation dosage that can inhibit AHN yet yield limited damage to mature neuronal and non-neuronal tissue/cells, we designed multidosage experiments and also analyzed the hippocampal sections with NeuN immunolabeling and Nissl stain, as well as the skull regions around the ¹²⁵I implants.

Our results show that by placing ¹²⁵I seeds on the cranial area above the temporal lobe of the cerebrum, a region-selective blocking effect on HAN is achievable. Thus, DCX⁺ cells in the DG are apparently reduced microscopically across the entire rostrocaudal extent of the hippocampus. The loss of subgranular DCX⁺ neurons and processes are evident in the dorsal as well as temporal portions of the DG, whereas those in the ventral DG partially remained. The latter implicates a lesser impact of the extracranial radiation to adult neurogenesis in the deeper part of hippocampal formation. A distance-dependent effect is further conceivable by the fact that the loss of DCX⁺ cells is greater in the ipsilateral than the contralateral hippocampus in reference to the radiation source. It should be noted that the amount of DCX⁺ cells at the SVZ at the levels of the striatum appears to be largely symmetric between the two hemispheres, indicating a lesser impact of the temporal cranial radiation on subventricular adult neurogenesis at a cerebral level rostral to the hippocampal formation. It should be noted that unilateral cerebral X-radiation can simultaneously block neurogenesis in both the SVZ and SGZ in the ipsilateral hemisphere (Xiong et al., 2010). Taken together, our results indicate that localized extracranial ¹²⁵I seed radiation can produce site-specific or space-limited inhibition of adult neurogenesis in guinea pig brain.

Our quantitative analysis of Ki67⁺ and DCX⁺ cells in the dorsal DG indicates that the effect of extracranial ¹²⁵I radiation on HAN is dose-dependent. Thus, the extent of decrease in the densities of Ki67⁺ and DCX⁺ cells in the DG tends to increase as the radiation dosage rises from 0.3 mCi to 0.6 mCi, and to 1.0 mCi. This trend is seen in both the ipsilateral and contralateral hippocampi as compared to sham operation controls at the 30- and 60-day surviving time points. Hypothetically, one may expect that radiation effect on HAN should be also time-dependent. However, in the present study, the reduction in the densities of Ki67⁺ and DCX⁺ cells in the dorsal DG are already dramatic at the 30-day surviving time point, although a noticeable further decline is seen at the 60-day time point. This finding appears to suggest that a 30-day ¹²⁵I radiation could yield a largely “saturated” or celling effect on the blockage of HAN. This result could be informative for the design of future studies using this method.

As mentioned in the Introduction, one concern for the use of radiation to block adult neurogenesis in the brain is the neurotoxic damage to mature neural components such as neurons, glia, and perhaps other cell types (e.g., vascular cells). This could be an issue involving confounding effect in functional studies on the role of adult neurogenesis. In our pilot experiment of this study, we did observe significant alterations in the morphology of NeuN⁺ neurons and Nissl-stained cells in the hippocampal formation in animals surviving 60 days with ∼2.4 mCi (3 × 0.8 mCi) radiation. However, our microscopic observation and quantitative analysis suggest that 30- and 60-day ¹²⁵I radiations at 0.3–1.0 mCi appear to cause insignificant, if any, histological or neuronal damage even in the ipsilateral hippocampus. Specifically, neither quantitative nor quantitative changes are found in the NeuN⁺ neurons in the CA1 and DG in the 0.3- and 0.6-mCi groups surviving 30 and 60 days relative to sham controls. Accordingly, 0.3- and 0.6-mCi extracranial ¹²⁵I radiations can apparently ablate HAN, but do not result in overt injurious impact on mature neuronal population.

Adult neurogenesis is a large topic in neuroscience, relating to basic mechanistic research, to pathogenesis of certain brain disorders, and to the development of therapeutic strategies for some neurological and psychological diseases (Shors et al., 2001; Cho et al., 2015; Radad et al., 2017; Fares et al., 2019; Abrous et al., 2021). Despite significant progress, much work is needed in each of the above frontiers. In general, ablation (knockout, knockdown or silencing in other terms, as used in genetic and molecular studies) is an important experimental approach to define the functional output of a particular biological system, from molecular to behavioral levels. The findings discussed above represent initial experimental data in support of a potential utility of localized extracranial ¹²⁵I radiation in the investigation into HAN in mammalian brain. Besides the importance of extracranial ¹²⁵I radiation leading to stereotactic and dose-dependent effect on adult neurogenesis elaborated previously, this method has other advantages on its application in animal experiments. For instance, the surgical procedure is rather simple and, with practice, can be completed shortly even under a brief inhalation anesthesia. This would cause fewer anesthesia-/operation-related animal mortalities and also minimize animal stress, which could be of importance in some experimental studies, such as that aimed to explore the functional implications of HAN during pregnancy, including the development of motherhood behavior in some mammals (Wan et al., 2019, 2021).

Different regions of the hippocampus appear to be engaged in different cognitive activities (Shors et al., 2002; Bannerman et al., 2004; Fanselow and Dong, 2010; Strange et al., 2014; Huckleberry et al., 2018; Trompoukis and Papatheodoropoulos, 2020). It is suggested that the dorsal hippocampus may perform primarily cognitive functions, whereas the ventral part is more related to the processing of stress, emotion and affect (Fanselow and Dong, 2010; Kheirbek and Hen, 2011). HAN in different regions of the longitudinal axis of the hippocampus may differentially participate in learning and memory and mode responses (Levone et al., 2021). At present, the ¹²⁵I tubes were placed around the middle area of the temporal bone at the posterior part of the cranium. We see the radiation effect predominantly on the Ki67⁺ and DCX⁺ at the SGZ of the dorsal and temporal subregions of the DG in the ipsilateral side. These labeled cells were also reduced in the dorsal DG of the contralateral hippocampus relative to sham control according to cell count. The Ki67⁺ and DCX⁺ at the SGZ in the most ventral part of the DG appeared to be maintained largely. By adjusting the implant location and ¹²⁵I dosage, it appears possible to allow the radiation effect projecting specifically or predominantly to the bilateral dorsal (e.g., a midline placement) hippocampus and temporal/ventral hippocampus (e.g., bilateral and lateral placement on the cranium, respectively). This type of surgical design could allow investigations into the role of HAN in different parts of the hippocampus in processing various cognitive and behavioral functions.

Another application of the extracranial ¹²⁵I radiation method is to explore the extent, time course, and cellular feature of the recovery of HAN in mammalian brain, which may be also clinically relevant to radiation encephalopathy (Kickingereder et al., 2014; Weusthof et al., 2021; Zhang et al., 2021). The SGZ contains a pool of neural stem cells that are quiescent or undergoing proliferation, with the latter event leading to both neuronal and glial genesis (Del Bigio, 1999; Fares et al., 2019; Gage, 2019). However, less clear is whether, and if so, the extent of which, radiation would impact the populations of neural stems, the neuronal and glial fates of the precursor cells, and the time course and degree of neurogenesis recovery (Jinno, 2011; Kohler et al., 2011; Amrein, 2015). The relatively long lifespan of guinea pigs might allow a follow-up of morphological and behavioral recovery over an extended time, if needed.

In summary, we explored the feasibility of using extracranial ¹²⁵I seed implantation to block hippocampal adult neurogenesis using guinea pigs as animal model, and cell proliferation (Ki67) and immature neuronal (DCX) markers as cause–effect indexers. Observational and quantitative analyses suggest that this approach can lead to stereotactic radioablation of hippocampal adult neurogenesis without dramatically damaging the mature neuronal populations. This method may be optimized to explore the functions of adult neurogenesis in mammalian brain.