High‐intensity interval exercise is more efficient than medium intensity exercise at inducing neurogenesis Authors: Marvin Lambertus, Samuel Geiseler, Cecilie Morland Published: 2024 ABSTRACT AbstractAbstractThe neurogenic potential of the brain decreases during ageing, whereas the risk of neurodegenerative diseases and stroke rises. This creates a mismatch between the rate of neuron loss and the brain's capacity for replacement. Adult neurogenesis primarily occurs in the subgranular zone (SGZ) and the ventricular‐subventricular zone (V‐SVZ). Exercise enhances SGZ neurogenesis, and we previously showed that V‐SVZ neurogenesis is induced by exercise via activation of the lactate receptor HCA1. Here, we investigated how high‐intensity interval training (HIIT) and medium‐intensity interval training (MIIT) affect neurogenesis in these niches. Wild‐type (WT) and HCA1 knockout (KO) mice were randomized to sedentary, HIIT or MIIT (n = 5–8 per group) for 3 weeks. In the SGZ, HIIT increased the density of doublecortin (DCX)‐positive cells in WT mice by 85% (5.77±1.76 vs. 3.12±1.54 cells/100 µm, P = 0.013) and KO mice (67% increase; 7.91±2.92 vs. 4.73±1.63 cells/100 µm, P = 0.004). MIIT did not alter the density of DCX‐positive cells in either genotype. HIIT increased the density of Ki‐67‐positive cells only in KO mice (P = 0.038), whereas no differences in nestin‐positive cells were observed. In the V‐SVZ, HIIT increased the density of DCX‐positive cells in WT mice by 155% (117.79±39.72 vs. 46.25±19.96 cells/100 µm, P < 0.001) and MIIT increased the density of DCX‐positive cells by 80% (83.26±39.48 vs. 46.25±19.96 cells/100µm, P = 0.027). No exercise‐induced changes were observed in KO mice. Similar patterns were noted for Ki‐67 positive and DCX/Ki‐67 double‐positive cells in the V‐SVZ. These findings suggest that HIIT enhances neurogenesis more robustly than MIIT in both niches, with HCA1 playing a crucial role in V‐SVZ neurogenesis. Key points The neurogenic potential of the brain decreases with age, whereas the risk of neurodegenerative diseases and stroke increases, highlighting a mismatch between neuronal loss and replacement capacity.Exercise enhances neurogenesis in both the subgranular zone and the ventricular‐subventricular zone.High‐intensity interval exercise is more effective than medium‐intensity interval exercise at promoting neurogenesis in both the subgranular zone and the ventricular‐subventricular zone of wild‐type mice.The enhancement of neurogenesis in the ventricular‐subventricular zone is dependent on the activation of the HCA1 receptor, as evidenced by the ability of medium‐ and high‐intensity interval exercise to induce neurogenesis in wild‐type mice and the lack of this effect in HCA1 knockout mice.By contrast, neurogenesis in the subgranular zone is independent on the activation of the HCA1 receptor, highlighting that neurogenesis in the two major neurogenic niches are regulated differently. Keywords: adult neurogenesis, ageing, doublecortin, exercise, GPR81, HCA1 , Ki‐67, lactate, nestin, stroke Introduction The ability of the human brain to replace dead or damaged neurons is limited. Consequently, neurodegenerative disorders, stroke or trauma often result in uncompensated neuronal loss, which may lead to motoric and/or cognitive decline. The neurogenic potential of the brain decreases over the lifespan (Boldrini et al., 2018; Kempermann et al., 2018; Sorrells et al., 2018). Concurrently, the risk of neurodegenerative diseases and stroke increases with age (Hou et al., 2019). This creates an increasing mismatch between the rate at which neurons are lost and the brain's capacity to replace them. Interventions that enhance adult neurogenesis could effectively protect the brain from neuronal loss. Neurogenesis primarily occurs in two brain niches (Fig. 1A ): the ventricular‐subventricular zone (V‐SVZ) and the subgranular zone (SGZ). The V‐SVZ, which is located along the lateral ventricles (Fig. 1B ), represents the largest neural stem cell niche (Alvarez‐Buylla & Lim, 2004). New‐born neurons originating from the V‐SVZ can migrate towards the olfactory bulb, the striatum or the neocortex, depending on their anatomical origin within the V‐SVZ (Merkle et al., 2007). Neurogenesis in the SGZ (Fig. 1C ), on the other hand, gives rise to neurons that migrate internally in the granular cell layer of the hippocampal formation. When mature, these glutamatergic neurons are considered to be essential for learning and memory function (Lopez‐Rojas & Kreutz, 2016; Voss et al., 2019). Adult neurogenesis in the V‐SVZ and the SGZ A, localization of the V‐SVZ and SGZ in a mice sagittal section. B, neurogenesis in the V‐SVZ. Activated neural stem cells in the wall of the lateral ventricle express nestin and glial fibrillary acidic protein (GFAP). Upon activation, they readily begin to differentiate into transit amplifying cells (TACs) and proliferate, expressing the proliferation marker Ki‐67 and gradually losing their GFAP expression. The TACs will eventually differentiate to neuroblasts (type A cells) as they gradually lose their nestin expression and instead start to express doublecortin (DCX). Finally, these cells will migrate and differentiate further to mature neurons, expressing NeuN and other markers for adult neurons. C, neurogenesis on the SGZ. Activated neural stem cells (type 1 cells) in the granular cell layer of the dentate gyrus (DG) express nestin. The activated neural stem cells may proliferate to Ki‐67‐expressing type 2 intermediate progenitor cells (IPCs). These cells slowly begin to express DCX at the same time as losing nestin expression as they differentiate to neuroblasts (type 3 cells). The neuroblasts may migrate throughout the granular cell layer and become mature granule neurons expressing NeuN and other markers for adult neurons. Created with BioRender.com. [Colour figure can be viewed at wileyonlinelibrary.com] Although hippocampal neurogenesis is crucial for spatio‐temporal memory, and its reduction may lead to memory loss, the role of V‐SVZ neurogenesis in memory is not well understood. Migration, maturation and integration of V‐SVZ neurons into the circuitry of the olfactory bulb is involved in odour‐recognition/olfactory memory (Enwere et al., 2004; Rochefort et al., 2002). Although distinct processes, olfactory memory may influence hippocampus‐dependent memory by giving contextual cues (Eichenbaum & Robitsek, 2009). Becuase the integration of odour information is regarded part of the higher order cognitive processing, it has been proposed that reduced olfactory memory correlates with cognitive decline in ageing (Eichenbaum & Robitsek, 2009). Physical exercise increases cerebral vascularization (Ding et al., 2006; Mattson et al., 2018; Morland et al., 2017; van Praag et al., 2005), induces adult neurogenesis (Brown et al., 2003; Itoh et al., 2011; Mastrorilli et al., 2017; van Praag, 2008) and ameliorates other forms of neuroplasticity associated with ageing and age‐related neurodegenerative diseases (Voss et al., 2019). Adult neurogenesis of the hippocampus is affected by exercise at several levels. It enhances neuronal proliferation, differentiation (Brown et al., 2003; Itoh et al., 2011; van Praag, 2008) and functional integration into the neuronal network (El‐Sayes et al., 2019; Farioli‐Vecchioli & Tirone, 2015; Liu & Nusslock, 2018; Voss et al., 2019). Increased neurogenesis coincides with other plasticity‐related changes such as spinogenesis, dendritic arborization and enhanced long‐term potentiation (Dietrich, 2013; Lin et al., 2012). Accordingly, exercise may decelerate or even prevent cognitive decline and dementia development in humans (Beydoun et al., 2014; Blondell et al., 2014; Hayes et al., 2015, 2016, 2017; Livingston et al., 2017; van Praag, 2009). Although exercise‐induced neurogenesis in the SGZ is a well‐established phenomenon, regulation of neurogenesis in the V‐SVZ by exercise is less obvious. In the V‐SVZ, exercise has primarily been demonstrated to rescue neurogenesis in mouse models with genetically or chemically reduced neurogenesis (Blackmore et al., 2012; J. C. Lee et al., 2016; Mastrorilli et al., 2017). However, we previously reported that exercise induces neurogenesis also in normal, healthy mice (Lambertus et al., 2020). Exercise after stroke also improves memory and motor function (Saunders et al., 2014; Zhang et al., 2013), as well as survival of new‐born neurons in the SGZ (Luo et al., 2007), but to what extent this contributes to the overall neuroprotective effects of exercise in stroke remains unknown. However, a recent study (Geiseler et al., 2024) found that lactate injections administered at 12 and 24 h after experimental stroke, causing lactate levels corresponding to that found after high‐intensity exercise (Morland et al., 2017), reduced lesion volumes in mice in a hydroxycarboxylic acid receptor 1 (HCA1)‐dependent manner. Most studies focusing on neurogenic effects of exercise have been performed with moderate or low intensity running/walking, but high‐intensity interval training (HIIT) has attracted more attention recently. Cardiovascular and metabolic effects of HIIT appear to be comparable (and in some respects superior) to what can be achieved by traditional low‐to‐moderate intensity exercise (Lucas et al., 2015). For cerebral effects of exercise, no consensus has been reached regarding their sensitivity to exercise intensity. This probably reflects the scarcity of studies directly comparing different running exercise regimes. One study demonstrated that HIIT was superior to moderate intensity continuous running (Constans et al., 2020) at inducing plasticity related genes in rats. Lactate increases in the blood during strenuous exercise such as HIIT and may gain access to the brain through monocarboxylate transporters (MCTs) at the blood–brain barrier. Lactate injections in mice induce some of the same effects on neurobioenergetics as those seen with exercise, and hence lactate has been suggested to act as an exercise mimetic (Cunnane et al., 2020; E et al., 2013; Morland et al., 2017). After uptake to the brain though MCT1 in the blood–brain barrier (Pierre & Pellerin, 2005), lactate may affect brain cells through two mechanisms: uptake into neurons to fuel metabolism (Magistretti & Allaman, 2018) or acting as an agonist at the HCA1 (also known as HCAR1 or GPR81), which is present in the brain (Bozzo et al., 2013; Hadzic et al., 2020; Lauritzen et al., 2014; Morland et al., 2017). HCA1 is a Gi‐protein coupled receptor and binds extracellular lactate. Receptor actions are therefore independent of lactate uptake into cells via MCTs (Morland et al., 2015). Intriguingly, lactate may affect memory through a combination of metabolic and signalling mechanisms: the learning process may be enhanced by lactate used as a metabolic substrate and memory consolidation may be induced by lactate through activation of HCA1 (Scavuzzo et al., 2020). Exercise increases the concentration of brain‐derived neurotrophic factor in brain and blood, and the concentration has been found in some studies to be proportional to the exercise intensity (Ferris et al., 2007; Saucedo Marquez et al., 2015). Intense exercise also increased the concentration of serum insulin‐like growth factor 1 (IGF‐1) (Kim et al., 2015) and we have previously shown that the exercise‐induced increase in vascular endothelial growth factor A depends on lactate signalling through HCA 1 (Morland et al., 2017). These findings suggest that intensities above the aerobic threshold may be required, and that lactate and growth factors released in response to HIIT may outcompete possible stress‐effects, and hence efficiently induce neurogenesis. We have previously reported that neurogenesis in V‐SVZ, but not in the SGZ, is dependent on activation of HCA1 by lactate (Lambertus et al., 2020). In the present study, we compare two commonly used exercise regimes; HIIT and medium‐intensity interval exercise (MIIT). We further compare wild‐type (WT) mice to HCA1 knockout (KO) mice to determine whether HCA1‐dependent neurogenesis varies with exercise intensity. Methods Animals The study was performed in strict accordance with the national and regional ethical guidelines. All animal handling and experiments were performed by FELASA (i.e. Federation of European Laboratory Animal Science Associations) certified personnel. The experiment was approved by the Animal use and care committee of the Institute of Basic Medical Sciences, Faculty of Medicine, University of Oslo, and by the Norwegian Animal Research Authority (FOTS 12521, FOTS 14204), and is reported in accordance with the Animal research: reporting in vivo experiments (ARRIVE) guidelines, version 2.0 (Percie du Sert et al., 2019). The HCA1 KO mice were generated as described (Ahmed et al., 2010); the strain was a gift from Professor Stefan Offermanns (Max Planck Institute for heart and lung research, Bad Neuheim, Germany) and was back‐crossed nine times with C57Bl/6N mice after arrival in our lab. The mice had access to food and water ad libitum and were housed in groups of one to nine animals per GreenLine cages (Sealsafe Plus GM500 or GR900), at the Department for comparative medicine at the Institute of Basic Medical Sciences, Faculty of Medicine, University of Oslo. The light photocycle comprised 12:12 h light/dark. The cages were enriched with nest‐building material and a toilet roll core or a plastic ‘house.’ Because we investigated the effects of treadmill exercise, running wheels were not installed as enrichments in the cages. HCA1 KO mice and WT mice were aged 8.5–18.5 weeks at the start of the experiment. All mice were weighed at the first day of the experiment and again at the beginning of the third week. The mice were semi‐randomly assigned to the three groups: HIIT, MIIT or a sedentary control group, by drawing lots. The number of animals in each group comprised: WT control: N = 8 (three males and five females), KO control: N = 7 (two males and five females), WT MIIT: N = 8 (two males and six females), KO MIIT N = 8 (two males and six females), WT HIIT: N = 9 (four males and five females), KO HIIT: N = 8 (three males and five females). The mice assigned to HIIT or MIIT received exercise intervention 5 days a week for 3 weeks. A maximal exercise capacity test (MECT) was performed on the second day of exercise, and again after 2 weeks of exercise. The HIIT and the MECT were performed as previously described (Morland et al., 2017) using a treadmill for rodents (Exer 3/6; Columbus Instruments, USA) at a 25° incline. For the MECT, the mice ran on the treadmill for 15 min at a speed 9.6 m min−1 (warm‐up), then the speed was increased by 1.8 m min−1 every 2 min. The mice were continuously monitored by the operator and the test was stopped when the mice refused to run further, despite being manually pushed gently on to the band or receiving electrical stimuli (<1.5 mA, maximally once per day). The HIIT and MIIT consisted of 10 × 4 min of running at a speed calculated based on the maximum capacity, with 2 min active rest (at a speed of 8.0 m min−1) between each interval. For the mice exposed to HIIT, the intervals were performed at 70% of the running speed obtained in the MECT. The MIIT was performed at 50% of the maximum running speed. During the intervals, the mice were stimulated to run for the designated time by gently tapping their tail or gently pushing them back on the band before they reached the grid at the bottom of the treadmill. The control mice were brought in the exercise room along with the others but received no exercise intervention. Exclusion criteria The animals were monitored daily according to a standardized scheme to ensure their well‐being throughout the experiment. The following exclusion criteria were set a priori and were described in the animal welfare protocol (FOTS 14204): at any sign of distress, such as weight loss of more than 10% of the bodyweight within a week, stereotypical behaviour, lack of grooming, etc., the animal was excluded from the experiment. For the animals exposed to exercise, additional exclusion criteria were set: if an animal performed worse than expected (based on the previous performance of the same animal) in one interval training session, additional breaks of a few seconds were given as needed. If the same animal continued to underperform for two consecutive exercise sessions, the animal could be injured or sick and was therefore withdrawn from the study. In total, three animals were withdrawn from the study because they met one of the exclusion criteria, all of them because of the last criteria (presumed injured or sick). These were: WT HIIT, two mice and WT MIIT, one mouse. Antibodies The primary antibodies used in the immunofluorescence labelling experiments were rabbit anti‐Ki‐67 (Abcam, Cambride, UK; AB15580; final dilution 1:500), guinea pig anti‐doublecortin (DCX) (Merck Rahway, NJ, USA; AB2253; final dilution 1:2000) and mouse anti‐nestin (Abcam; AB6142; final dilution 1:500). The secondary antibodies used were Alexa Fluor 647 anti‐rabbit (Invitrogen, Waltham, MA, USA; A21241; final dilution 1:500), Alexa Fluor 594 anti‐guinea pig (Invitrogen; A11076; final dilution 1:500) and Alexa Fluor 488 anti‐mouse (Invitrogen; A21202; final dilution 1:500). Tissue preparation At 6 h after the last exercise session, the mice were anaesthetized with zolazepam (3.3 mg mL−1), tiletamine (3.3 mg mL−1), xylazine (0.5 mg mL−1) and fentanyl (2.6 µg mL−1); 10 µL g−1 body, injected i.p.Under deep anaesthesia, the animals were transcardially perfused for 8 min with freshly made and filtered 4% formaldehyde in 0.1 m sodium phosphate buffer. After the fixation, the brain was gently removed from the scull and placed in the fixative solution overnight at 4°C. The following day, the brains were transferred to a 1:10 dilution of the fixative for storage at 4°C until further processing. Each brain was assigned an ID number (1–45) to ensure that the operator was blinded to the treatment groups and genotypes throughout experiments and analyses. Hence, to reduce potential confounding factors all procedures (the sectioning, immunolabeling, imaging and quantitative analysis) were performed in the order of the ID number, which was unrelated to the treatments or genotype. The brains were cut along the midline, and one hemisphere was allowed to sink in 30% sucrose (in 0.1 m phosphate buffer) overnight for cryoprotection. The brains were then cryosectioned at –20°C into 20 µm thick free‐floating sagittal sections with a sliding microtome (HM 450; Thermo Fisher Scientific, Waltham, MA, USA). The free‐floating sections were stored in phosphate buffer with 0.05 NaN3 at 4°C. Immunofluorescence For the immunostaining, the sections were rinsed in PBS for 2 × 10 min and then exposed to 0.1 m citrate buffer, pH 8.6, at 80°C for antigen retrieval. Next, the sections were rinsed again for 3 × 10 min before they were treated with blocking buffer (3% newborn calf serum, 1% bovine serum albumin, 0.05% Triton‐X in PBS) for 2 h to block unspecific antigen binding sites. The sections were then incubated with the primary antibodies diluted in blocking solution overnight with gentle shaking at room temperature. Then the sections were washed for 6 × 10 min with PBS and incubated with the secondary antibodies in blocking solution for 2 h with gentle shaking and covered from light. After that, the sections were washed in PBS for 3 × 10 min, followed by incubation with 4′,6‐diamidino‐2‐phenylindole (DAPI) for 15 min. Finally, the sections were washed with PBS for 3 × 10 min, mounted with Prolong Gold antifade reagent (Life Technologies, Carlsbad, CA, USA) and cover slipped. The mounted sections were stored at 4°C, protected from light, until confocal imaging. Of the sections intended for quantification of hippocampal neurogenesis, one KO HIIT was lost during the staining process and one WT HIIT was found to be from different stereological region. Of the sections intended for analysis of V‐SVZ neurogenesis, one WT control and one WT HIIT were found to be from outside of the narrow stereological range included in this study. From one of the WT HIIT section, the area containing the lateral ventricle was lost during the staining process. These sections were omitted from the analysis. Confocal microscopy Confocal Z‐stack images of the sections were taken with an Andor Dragonfly spinning disk confocal microscope (Oxford Instruments, Abingdon, UK) with a Zyla 4.2 camera at 40× magnification. Diode lasers with the wavelength 405, 488, 561 and 637 nm were used with a laser intensity of 24%, 15%, 7% and 6%, respectively, and 20 ms light exposure time for each channel. The images of SGZ and V‐SVZ from each section were taken in a Z‐stack with 0.5 µm thickness in each single optical image to obtain a series of 30–40 images per stack, which covered the whole 20 µm thickness of the section. Data analysis The images were examined with IMARIS, version 9.3.1 (Oxford Instruments) and Fiji, version 2.1.0/1.53c; Java 1.8.0_172 (Schindelin et al., 2012). The cells in the V‐SVZ of one sagittal section from each mouse were quantified along the full length of the lateral ventricle wall. The sections were carefully selected to be from the same area, 0.8 mm from the midline, with reference to the Allen Mouse Brain Atlas (https://mouse.brain‐map.org). The number of cells labelled with each neurogenesis marker in the SGZ was quantified along both blades of the granular layer. The densities of cells are expressed as the number of cells that were positive for each marker per 100 µm of ventricle wall length or 100 µm of granular layer length, respectively. DCX‐positive (in V‐SVZ and SGZ) and nestin‐positive (in SGZ) cells were identified as cells containing a DAPI‐positive nucleus, surrounded by DCX or nestin labelling (consistent with soma staining). The Ki‐67‐positive cells (in V‐SVZ and SGZ) were identified as DAPI‐positive nuclei that co‐expressed Ki‐67. Because of the high density of nestin‐positive cells in the V‐SVZ, individual nestin‐positive cells could not be manually counted. Consequently, in the V‐SVZ, nestin positive cells was quantified using a thresholding method in Fiji, and the area of the nestin positive profiles was measured as an estimate of the number of cells. The validity of this method was verified by performing the same type of analysis of DCX signal in the SVZ and comparing it with the results obtained by manual counting (see Supporting information, Fig. S1). A test of homogeneity of variances (SPSS, version 26; IBM Corp., Armonk, NY, USA) gave P > 0.05, hence confirming that equal variance could be assumed. Consequently, the statistical analyses were performed using one‐way ANOVA with least significant difference (LSD) post hoc analysis using SPSS, version 26, with a 95% confidence level. A pairwise Student´s t test (SPSS) was used when comparing bodyweight before and after 3 weeks of exercise. All data are reported as the mean ± SD. The underlying data are available from the corresponding author upon reasonable request. Results Higher exercise intensity is required for SGZ neurogenesis In the SGZ of the hippocampus, we observed that HIIT increased the density of DCX‐positive cells in both WT and HCA1 KO mice (Fig. 2). In WT mice, HIIT induced an 85% increase in DCX‐positive cells, from 3.12 ± 1.54 cells/100 µm in the WT control group to 5.77 ± 1.76 cells/100 µm after HIIT (P = 0.013). In KO mice, HIIT resulted in a 67% increase in DCX‐positive cells, from 4.73 ± 1.63 cells/100µm in the KO control group to 7.91 ± 2.92 cells/100 µm after HIIT (P = 0.004). MIIT did not significantly increase the density of DCX‐positive cells in either genotype (WT MIIT: 4.21 ± 2.33 cells/100 µm vs. WT control. 3.12 ± 1.54 cells/100 µm; P = 0.29; KO MIIT: 5.22 ± 1.11 cells/100 µm vs. KO control: 4.73 ± 1.63 cells/100 µm; P = 0.622) (Fig. 2H ). The effects of exercise on markers of earlier neurogenic stages, Ki‐67 and nestin, were less clear. Although the densities of Ki‐67‐positive cells tended to increase after HIIT in both genotypes, the increase was statistically significant only in KO mice (WT HIIT: 1.57 ± 0.89 cells/100 µm vs. WT control: 1.15 ± 0.62 cells/100 µm, p 0.23, KO HIIT 2.00 ± 0.62 cells/100 µm vs. KO control: 1.24 ± 0.50 cells/100 µm in the, P = 0.038) (Fig. 2I ). MIIT did not significantly affect the density of Ki‐67‐positive cells in either genotype (WT MIIT: 0.84 ± 0.44 cells/100 µm vs. WT control: 1.15 ± 0.62 cells/100 µm P = 0.37, KO MIIT: 1.36 ± 0.77 cells/100 µm vs. KO control: 1.24 ± 0.50 cells/100 µm P = 0.74). There was no difference in the density of nestin‐positive cells between any of the groups in response to exercise (WT control: 2.47 ± 1.11 cells/100 µm; WT MIIT: 2.84 ± 1.74 cells/100 µm; WT HIIT: 3.21 ± 1.60 cells/100 µm; KO control: 2.84 ± 1.06 cells/100 µm; KO MIIT: 2.42 ± 0.52 cells/100 µm; KO HIIT: 3.10 ± 1.69 cells/100 µm) (Fig. 2J ). The effect of exercise at different intensities on neurogenesis in the SGZ of WT mice and HCA 1 KO mice A, sagittal section of a mouse brain showing the location of the SGZ in the hippocampal formation. B–G, confocal micrographs of the dentate hilus, including the granule cell layer and the SGZ of WT (B–D) and HCA1 KO (E–G) mice exposed to no exercise (control; B, E), or interval exercise at medium intensity (C, F) or high intensity (D, G). The sections were labelled for DCX (red), Ki‐67 (cyan), nestin (green) and DAPI (blue). H–J, bar charts showing the density of positively stained neurons (number of cells per 100 µm of granular cell layer length). Bars represent the mean density (WT; grey bars; HCA1 KO: white bars); error bars represent SD; the black dots represent data from one individual mouse. H, density of DCX‐positive cells. I, density of Ki‐67‐positive cells. J, density of nestin‐positive cells. *P < 0.05; (one‐way ANOVA; LSD post hoc test); WT, n = 8, 7, 6; KO, n = 7, 7, 7. Scale bar = 100 µm. [Colour figure can be viewed at wileyonlinelibrary.com] Neurogenesis of the V‐SVZ depends on exercise intensity and HCA1 activation In the V‐SVZ, DCX‐positive cells increased by 80% in WT mice exposed to MIIT (83.26 ± 39.48 cells/100 µm) compared to sedentary controls (46.25 ± 19.96 cells/100 µm; P = 0.027) (Fig. 3). HIIT induced an even more pronounced increase (155 %) in DCX‐positive cells compared to sedentary controls (WT HIIT: 117.79 ± 39.72 cells/100 µm vs. WT control: 46.25 ± 19.96 cells/100 µm; P = 0.00003). In HCA1 KO animals, the density of DCX‐positive cells was unaffected by exercises at any intensity (KO control: 55.62 ± 31.44 cells/100 µm; KO MIIT: 64.46 ± 28.44 cells/100 µm; KO HIIT: 63.87 ± 18.14 cells/100 µm) (Fig. 3H ). A similar pattern was observed for Ki‐67‐positive and double positive DCX/Ki‐67‐cells: In WT mice, HIIT increased Ki‐67‐positive cells by 83% (WT HIIT: 78.05 ± 15.06 vs. WT control: 42.70 ± 15.47 cells/100 µm; P = 0.002) (Fig 3I ) and DCX/Ki‐67‐positive cells by 118% (WT HIIT: 68.92 ± 13.78 vs. WT control: 31.65 ± 14.69 cells/100 µm; P = 0.0003) (Fig. 3J ). We found that MIIT was not sufficient to evoke an increase of Ki‐67‐positive cells (53.14 ± 22.41 cells/100 µm) (Fig. 3I ) or DCX/Ki‐67‐positive cells (41.28 ± 16.97 cells/100 µm) in the V‐SVZ (Fig. 3J ). Again, no neurogenic effect of exercise was observed in KO mice in the V‐SVZ (Ki‐67: KO control: 61.59 ± 18.50 cells/100 µm, KO MIIT: 59.20 ± 18.90 cells/100 µm, KO HIIT: 63.26 ± 15.28 cells/100 µm). It is noteworthy that the density of Ki‐67‐positive cells was higher, almost reaching statistical significance, in sedentary KO mice (61.59 ± 18.50 cells/100 µm) compared to sedentary WT mice (42.70 ± 15.48 cells/100 µm; P = 0.057). The effect of HIIT and MIIT on V‐SVZ neurogenesis in WT and HCA 1 KO mice A, sagittal section of a mouse brain showing the location of the V‐SVZ neurogenic niche. B–G, confocal micrographs of the ventricular‐subventricular zone of WT (B–D) and HCA1 KO (E–G) mice exposed to no exercise (control; B, E), or interval exercise at medium intensity (C, F) or high intensity (D, G). The sections were labelled for DCX (red), Ki‐67 (cyan), nestin (green) and DAPI (blue). H–K, bar charts showing the density of positively stained neurons (number of cells per 100 µm of granular cell layer length). Bars represent the mean density (WT; grey bars; HCA1 KO: white bars); error bars represent SD; the black dots represent data from one individual mouse. H, density of DCX‐positive cells. I, density of Ki‐67‐positive cells. J, density of DCX/Ki‐67‐doble positive cells. K, density of nestin‐positive cells. *P < 0.05; **P < 0.01; ***P < 0.001 (one‐way ANOVA; LSD post hoc test);WT, n = 7, 7, 5; KO, n = 7, 8, 7. Scale bar = 50 µm. [Colour figure can be viewed at wileyonlinelibrary.com] Consistent with the results for the other neurogenesis markers, nestin expression increased in WT mice exposed to HIIT (WT HIIT: 20 720 ± 7414 µm2/100 µm vs. WT control: 9609 ± 3337 µm2/100 µm in WT controls; P = 0.00003) but not in WT mice exposed to MIIT (9542 ± 4065 µm2/100 µm; P = 0.975 for WT MIIT vs. WT controls) (Fig. 4K ). Nestin levels were unaffected by exercise in HCA1 KO mice (KO control: 11 434 ± 2389 µm2/100 µm; KO MIIT: 12.47 ± 3.99 µm2/100 µm; KO HIIT: 12 255 ± 2289 µm2/100 µm). Progression in bodyweight and exercise performance during the experiment A, bodyweight of the animals throughout the intervention period. B, maximum speed obtained in the MECT at week 1 and week 3. Filled lines represent WT mice; dotted lines represent KO mice, exposed to no exercise (blue), MIIT (red) or HIIT (green). All animals exhibited a significant increase in performance in the MECT after 2 weeks of exercise compared to the beginning of week 1 (P <0.05). [Colour figure can be viewed at wileyonlinelibrary.com] Progression in running speed and bodyweight was equal in HCA1 KO and WT The bodyweight of the animals was monitored throughout the exercise period (Fig. 4A ). The KO animals, on average had lower bodyweight at the start of the experiment (WT: 35.88 ± 4.96 g vs. KO 26.63 ± 4.12 g; mean ± SD; P = 0.0018; Student's t test), which potentially reflects the higher average age in this group (17 weeks vs. 10 weeks). Over the 3‐week exercise period, most animals maintained their weight, although a slight decrease in bodyweight was observed in WT animals exposed to exercise. The weight at the end of the experiment were (in percent of their starting weight): WT control: 101.15 ± 2.76%, WT HIIT 90.40 ± 8.48% (WT HIIT vs. WT control P = 0.002), WT MIIT: 87.71 ± 4.57% (WT MIIT vs. WT control; P = 0.0002). In KO animals, exercise affected bodyweight only in the MIIT group, albeit with lower significance. The weights (as a percentage of starting weight) were: KO control: 102.73 ± 8.57%; KO MIIT: 96.39 ± 2.77% and KO HIIT: 100.8 ± 5.83%. All animals significantly improved their exercise performance during the intervention period (Fig. 4C ). Notably, KO animals generally tended to perform better on the initial MECT. Discussion In the present study, we used well‐established markers for neuroblasts (DCX), neuroprogenitor cells (Ki‐67) and neural stem cells (nestin) to show that both HIIT and MIIT, 5 days a week for 3 weeks, increased neurogenesis in the SGZ (Fig. 2) and the V‐SVZ (Fig. 3) of mice. In both neurogenic niches, HIIT was more efficient than MIIT at inducing neurogenesis. Furthermore, we show that the effect of exercise on neurogenesis in the V‐SVZ was abolished in HCA1 KO mice, indicating that lactate signalling through HCA1 regulates neurogenesis specifically in this niche (Fig. 3). By contrast, exercise induced neurogenesis in the SGZ equally in both wild‐type and HCA1 KO mice, suggesting that SGZ neurogenesis is independent of HCA1 activation. In a model of neonatal hypoxia/ischaemia (HI), HCA1 was found to be important for the induction of the transcriptional response to HI (Kennedy et al., 2022). Interestingly, this HCA1‐dependent regulation of the cell cycle in response to HI was found in the SVZ but not in the SGZ. Although Kennedy et al. (2022) studied neonatal brains and did not investigate the effects of exercise, their findings are from a HI model where lactate is presumably elevated in the brain, allowing activation of HCA1. Hence, their findings related to cell cycle regulation may also be representative of the HCA1‐dependent effects of exercise reported in the present study, Further supporting the notion that HCA1 is a driver of HI‐induced neural proliferation, Kennedy et al. (2022) found that neurospheres from HCA1 KO mice developed fewer colonies than neurospheres from WT mice, indicating a lower ability for self‐renewal and/or proliferation in the HCA1 KO neural progenitors. The main findings of the present study are also consistent with a previous study (Lambertus et al., 2020) reporting that exercise‐induced V‐SVZ neurogenesis is HCA1‐dependent, whereas SGZ neurogenesis is not. In the same study, lactate injections alone (without exercise) could mimic the neurogenic effects in the V‐SVZ. In the present study, however, the observed effect of HIIT on neurogenesis in the V‐SVZ (1.5‐fold increase in DCX‐positive cells) was much larger than that observed in the previous study (0.5‐fold increase). The two studies used the same HIIT regime, but the mice were exercised for 3 weeks in the present study, as opposed to 7 weeks in the previous study. Two different (but not mutually exclusive) explanations for this difference may be that (1) exercise slightly loses efficacy over time and/or (2) prolonged HIIT results in increasing stress, which is known to negatively affect neurogenesis (Dioli et al., 2019; Fitzsimons et al., 2016); for example, by increasing the circulating levels of glucocorticoids (Numakawa et al., 2017; Planchez et al., 2020). It should be noted that the duration of the exercise regimen (3 weeks) does not allow us to investigate whether the higher numbers of Ki‐67‐positive and DCX‐positive neuroprogenitor cells produced in response to exercise result in an increased number of neurons that reach maturation and are integrated into the existing neural network because this process takes 4–8 weeks (Aimone et al., 2014; Kempermann et al., 2015). For the same reason, we cannot conclude whether the maturation of DCX‐positive neurons differs between WT mice and HCA1 KO mice. Nevertheless, we show that the pool of DCX‐positive cells expanded more than the Ki‐67‐positive pool, and the effects on the nestin‐positive early progenitor cells were less pronounced. This suggest that exercise promotes the proliferation and survival of neuroblasts rather than expanding the entire stem/progenitor cell pool. Although exercise did not increase the density of Ki‐67‐positive cells in KO mice, we observed that sedentary KO mice had a tendency towards more Ki‐67 positive cells than sedentary WT mice. Although not reaching statistical significance, this observation allows the possibility that compensatory effects may have occurred as a result of the lack of HCA1, and that these effects may mask any exercise‐induced increase of neurogenesis in KO mice. This difference between sedentary WT and sedentary HCA1 KO mice was much less pronounced with the other neurogenesis markers in the present study, but a tendency towards increased DCX‐positive cells in the KO was also reported in a previous study (Lambertus et al., 2020). Taken together, our data demonstrate that HIIT induced neurogenesis more efficiently than MIIT in both major neurogenic niches of the brain. However, we cannot determine whether there is a peak threshold beyond which higher intensity exercise no longer yields additional neurogenic benefits, considering that excessive intensity may elevate stress levels detrimental to new‐born neurons. Nevertheless, our results suggest that exercise above the lactate threshold is required for optimal neurogenic effects. The HCA1 dependency of V‐SVZ neurogenesis may be a key mechanism in stroke‐induced neurogenesis (Zhang et al., 2013) because lactate levels in the brain and the cerebrospinal fluid (CSF) increase in response to cerebral stroke. Whether HCA1‐dependent neurogenesis is important in stroke remains to be investigated experimentally. The mice used in the present study have the gene encoding HCA1 knocked out in all cells of the body. With a known location of HCA1 in several peripheral cells (Ahmed et al., 2010; Hoque et al., 2014; Wise et al., 2003) in addition to the brain, we cannot exclude that the effects seen in our study represent (in part) HCA1 activation in peripheral organs, resulting in secondary effects on the brain. As demonstrated by in previous studies, rejuvenation of the brain, including enhanced neurogenesis and cognition, may be regulated by cues present in the circulation (Middeldorp et al., 2016; Villeda & Wyss‐Coray, 2013; Villeda et al., 2011, 2014). Growth factors and hormones such as VEGF, IGF‐1, insulin and irisin are able to cross the blood–brain barrier and affect neurogenesis (Bouchard & Villeda, 2015; Cotman et al., 2007; Palmer et al., 2000; Pereira et al., 2007; Tari et al., 2019), although whether they are increased in the blood in response to HCA1 activation is not known. Increased cerebral blood flow per se, as a result of increased angiogenesis after long‐term high‐intensity exercise (Morland et al., 2017), might facilitate the delivery of growth factors from the systemic environment to neurogenic niches We have previously shown that HCA1 activation increases cerebral angiogenesis (Morland et al., 2017). Nevertheless, in the present study, neurogenesis did not follow angiogenesis in the hippocampus: angiogenesis was HCA1‐dependent (Morland et al., 2017), whereas neurogenesis occurred with exercise irrespective of HCA1 (present study; Lambertus et al., 2020). Which mechanisms underly the differential regulation of neurogenesis in the V‐SVZ vs. the SGZ in response to HCA1 activation? This question remains to be investigated experimentally. We have, however, demonstrated that HCA1 is present in the pial fibroblast (Morland et al., 2017), as well as in the choroid plexus, the tela choroidea and the neuroepithelial lining of the dorsal third ventricle, where fibroblast and ependymal cells were found to express the receptor (Hadzic et al., 2020). These HCA1‐expressing cells along the ventricle can sense fluctuations in lactate and may release growth factors into the CSF in response. In isolated pial fibroblasts, activation of HCA1 leads to a rapid activation of the Akt/phosphoinositide 3‐kinase pathway (Lambertus et al., 2020), potentially inducing growth factor release (Ma et al., 2020; Y. J. Lee et al., 2016). Hypothetically, growth factors in the CSF would have an easier access to the cells of the V‐SVZ than to the SGZ, and hence stimulate neurogenesis in this niche selectively. To our knowledge, HCA1 has not been demonstrated on neural stem cells, making direct effects of lactate on these cells less probable but not impossible. The physiological implications of neurogenesis in the V‐SVZ remain largely unresolved. Progenitor cells formed in the V‐SVZ migrate to the olfactory bulb and contribute to odour‐recognition/olfactory memory (Enwere et al., 2004; Rochefort et al., 2002). The extent to which odour recognition provides cues to enhance hippocampus‐dependent memory is unclear, but some researchers suggest that olfactory memory deficits in rodents may parallel human age‐related cognitive decline (Eichenbaum & Robitsek, 2009; Enwere et al., 2004). In primates, including humans, progenitor cells from the V‐SVZ may migrate to other brain regions, such as the striatum (Bédard et al., 2002; Ernst & Christie, 2005; Ernst et al., 2014), potentially replacing neurons lost during neurodegenerative disorders (Androutsellis‐Theotokis et al., 2008; Zachrisson et al., 2011). In the present study, we did not detect DCX‐positive cells in the stratum (data not shown), although this might reflect that our experiments were performed in healthy young animals, and we cannot exclude that low‐level migration or HCA1‐induced neurogenesis might still contribute to striatal function in ageing or neurodegenerative conditions. Because we demonstrate that HCA1 KO mice and WT mice performed equally well in MECT during the first and the last week of the intervention (Fig. 4), differences in running speed or exercise intensity did not account for the HCA1 dependency of neurogenesis in the V‐SVZ. As expected, the mice rapidly increased their maximal exercise capacity during 3 weeks of exercise (Lambertus et al., 2020; Morland et al., 2017) and the percentage of increase was virtually equal in both genotypes. Furthermore, on average, the WT animals were some weeks older than the KO animals, although all animals were younger adults. The older average age of the WT mice compared to KO mice should not account for the observed increase in exercise‐induced neurogenesis because neurogenesis typically decreases with age. Our data suggest that high intensity is required for the most effective induction of neurogenesis. HIIT may not be feasible in elderly people at risk for neurodegenerative disorders, nor in stroke patients, where enhanced neurogenesis may be particularly important to replace dead or dying neurons. Based on the HCA1‐dependent neurogenesis detected in this study, one may speculate whether neurogenesis may be augmented by combining moderate‐intensity exercise with treatment with lactate or other HCA1 agonists. Additional information Competing interests The authors declare that they have no competing interests. Author contributions C.M. designed the study. M.L. performed the in vivo studies, with some help from S.G. and the people listed in the Acknowledgements section. M.L. planned these studies and C.M. and M.L. harvested organs. M.L. performed immunohistochemistry as well the analysis of neurogenesis in the S.G.Z. and V.‐S.V.Z. (supervised by C.M. and S.G.). M.L. and C.M. wrote the draft manuscript and prepared the figures. All authors discussed the analysis and results, critically revised the manuscript and approved the final version of the manuscript submitted for publication. Funding Norges Forskningsråd (Forskningsrådet): Marvin Lambertus, Samuel Geiseler, Cecilie Morland, 262647. Supporting information Peer Review History Supplementary Figure 1 Analysis of the area covered by DCX‐expressing cells in V‐SVZ Bars represent the mean density (WT; grey bars; HCA1 KO: white bars); error bars represent SD; the black dots represent data from one individual mouse. The area covered by DCX‐expressing cells in V‐SVZ gave similar results to those obtained when the cells expressing DCX were counted manually (Fig. 3H ). This demonstrates the suitability of measuring the area of positive profiles for nestin (as presented in Fig. 3J ), when individual nestin‐positive cells were too densely packed to be manually counted. Acknowledgements This work was supported by grants from the Research Council of Norway: 262647. We thank Stefan Offermanns for providing the HCA1 KO mice. The images were obtained at the NorMIC IBV imaging facility. We are grateful to Linda Thøring Øverberg, Gezime Seferi, Karl Martin Forbord and Ghazal Sajedi for valuable help with the exercise regime. Biographies Marvin Lambertus obtained his PhD in neuroscience in 2021. He is currently a postdoctoral researcher at Linköping University, Sweden, specializing further in neuroscience. Samuel Geiseler is a physiologist who obtained his PhD in 2016. His research interest is focused on intrinsic adaptations involved in hypoxia tolerance in the brain. Cecilie Morland is Professor of Pharmacy and group leader at the University of Oslo, Norway. Her research group focusses on understanding the influence of exercise on brain plasticity and metabolic health. Handling Editors: Paul Greenhaff & Madeleine Lowery The peer review history is available in the Supporting Information section of this article (https://doi.org/10.1113/JP287328#support‐information‐section). Data availability statement The original data are available from the corresponding author upon reasonable request. References Associated Data Supplementary Materials Peer Review History Supplementary Figure 1 Analysis of the area covered by DCX‐expressing cells in V‐SVZ Bars represent the mean density (WT; grey bars; HCA1 KO: white bars); error bars represent SD; the black dots represent data from one individual mouse. The area covered by DCX‐expressing cells in V‐SVZ gave similar results to those obtained when the cells expressing DCX were counted manually (Fig. 3H ). This demonstrates the suitability of measuring the area of positive profiles for nestin (as presented in Fig. 3J ), when individual nestin‐positive cells were too densely packed to be manually counted. Data Availability Statement The original data are available from the corresponding author upon reasonable request.