Link between glucagon-like peptide-1 receptor agonists and brain cell growth: a systematic review

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) was utilised to conduct this study (Page et al.,2021). Relevant articles were systematically searched using Web of Science, OVID (MedLINE, Embase, AMED, PsychInfo, JBI EBP), and PubMed from database inception to June 26, 2024. The search string used for the search included: (“GLP-1” OR “Glucagon-Like Peptide-1” OR “Glucagon-Like Peptide 1” OR “GLP-1 Agonist” OR “Glucagon-Like Peptide-1 Agonist” OR “Glucagon-Like Peptide 1 Agonist” OR “Semaglutide” OR “Ozempic” OR “Rybelsus” OR “Wegovy” OR “Dulaglutide” OR “Trulicity” OR “Exenatide” OR “Byetta” OR “Bydureon” OR “Liraglutide” OR “Lixisenatide” OR “Tirzepatide”) AND (“Neurogenesis” OR “Neuron* Proliferation). Separate searches were conducted on Google Scholar and from reference lists to ensure all articles relevant to the topic were captured.

Articles obtained from the systematic search were screened through the Covidence platform, wherein duplicate articles were removed (Covidence, 2024). Two reviewers (H.A. and Y.J.Z.) independently screened the titles and abstracts based on the inclusion and exclusion criteria (Table 1). Primary articles that reported on changes in neurogenesis as a result of GLP-1 prescription or administration were retrieved for full-text screening by two reviewers (H.A. and Y.J.Z.) (Table 1). All conflicts were resolved via discussion, and articles deemed eligible by both reviewers were selected for data extraction.

A piloted data extraction template was used to organise and obtain data from included studies. Information to be extracted was established a priori, including (1) author, (2) study type, (3) sample size, (4) stains, (5) outcomes of interest. Two independent reviewers (H.A. and Y.J.Z.) conducted data extraction, wherein all conflicts were resolved through discussion. Outcomes of interest pertained to changes in neurogenesis associated with GLP-1 prescription or administration.

Quality assessments were conducted using the SYRCLE’s risk of bias analysis tool for animal studies (Hooijmans et al., 2014). Relevant literature was assessed by two independent reviewers (H.A. and Y.J.Z.), wherein the risk of bias was evaluated, and all conflicts were resolved following discussions. Further information on inclusion and exclusion criteria, as well as a summary table, can be found as supplementary material (Tables 1 and 2).

A systematic search generated a total of 162 studies, wherein 7 duplicates were identified manually, and 60 duplicates were identified by Covidence. 95 studies underwent abstract and title screening, with 65 articles deemed irrelevant. In accordance with the inclusion and exclusion criteria, 30 full-text studies were assessed for eligibility, of which, 13 were excluded due to wrong outcomes (n = 8), wrong study designs (n = 3), wrong comparator (n = 1), and wrong intervention (n = 1), yielding a total of 17 studies for further analysis (Figure 1, Table S2↗). Although animal and human studies were eligible for inclusion, the search only yielded animal studies.

Quality assessment of the included studies was conducted using the SYRCLE’s risk of bias tool for animal studies (Hooijmans et al., 2014). Studies utilizing animal cell culture generated ‘not reported’ (NR) or ‘X’ notations, as some prompts were not applicable or not reported within the studies. However, as these studies derive cell culture from live animals, the SYRCLE’s risk of bias tool was used.

All of the included studies assessed preclinical literature, demonstrating low attrition and reporting bias. Common limitations of selected studies include insufficient detection bias and blinding procedures, including allocation concealment, random housing, and random outcome assessment domains.

The bias assessment in animal model studies revealed distinct patterns that correspond to their overall quality ratings. Studies assessed as “Good,” such as those conducted by Belsham et al., (2009); McGovern et al., (2012); Pathak et al., (2018); Sampedro et al., (2019); Solmaz et al., (2015); Weina et al., (2018); and Yang et al., (2019), generally exhibit a lower risk of bias with most items adequately reported, despite occasional limitations in areas like random housing or random outcome assessment. In contradistinction, studies rated as ‘Fair’, including those by Bertilsson et al.,(2007), Darsalia et al., (2012); Hamilton et al., (2011), Lennox et al., (2013), Parthsarathy & Hölscher (2013), Ren et al., (2021); and Zhao et al., (2022), often show a higher number of items marked as “Not Reported” or “No,” particularly in baseline characteristics, allocation concealment, and random housing. Additionally, the study by Hunter & Hölscher (2012) was rated ‘‘Poor’’ as it consistently demonstrate significant bias with multiple items inadequately reported, highlighting the necessity for rigorous methodology to ensure comparability and reliability in animal research (Table S1↗). This pattern emphasises the link between comprehensive reporting and a lower risk of bias, underscoring the crucial significance of thorough methodological transparency in producing high-quality, reliable scientific outcomes.

We have identified three studies examining the effects of GLP-1 and neurogenesis (Table 2) (Lennox et al., 2013; McGovern et al., 2012; Sampedro et al., 2019). Lennox et al., (2013) identified an association between (Val⁸)GLP-1-Glu-RA administration and a 30% increase in BrdU-positive cells (p < 0.05) within the granular cell layer of the dentate gyrus of mice in comparison to saline controls (Lennox et al., 2013). The increase in BrdU-positive cells indicate a significant difference in mature neuronal populations between GLP-1 administration and saline controls (Lennox et al., 2013). This trend was in accordance with results from McGovern et al., (2012), wherein (Val⁸)GLP-1-Glu-PA administration was significantly associated with increases in BrdU-positive cells (p < 0.05) and DCX-positive cells (p < 0.05) in mice dentate gyrus (McGovern et al., 2012). Similarly, Sampedro et al., (2019) reported significant increases in neurogenesis ascertained by marker Ki-67 in the ganglion cell layer, inner nuclear layer, inner plexiform layer, outer nuclear layer, and the outer plexiform layer (p < 0.05) of obese mice treated with GLP-1 in comparison to controls (Sampedro et al., 2019). Notably, GLP-1 administration was also associated with upregulation of survival pathways glycogen synthase kinase-3 beta (GSK3β) (p < 0.05) and B-cell lymphoma-extra-large protein (Bcl-xL) (p < 0.05) (Sampedro et al., 2019). Taken together, these results suggest that GLP-1 increases neurogenesis in various regions of the brain and provides mechanistic insight on the pathways in which GLP-1 can modulate neurogenesis (Lennox et al., 2013; McGovern et al., 2012; Sampedro et al., 2019).

To understand the effects of GLP-1 RAs and neurogenesis in animal models, we have identified 13 studies (Table 2). Of the studies included herein, Hamilton et al., (2011) characterised the effect of both exenatide and liraglutide on neurogenesis.

Findings from Solmaz et al., (2015) identified significantly higher hippocampal neuronal count (p < 0.05) in mice treated with exenatide in comparison to saline (Solmaz et al., 2015). These results were replicated by Pathak et al., (2018), wherein exendin-4 (Ex-4) administration in mice was associated with significantly increased DCX-positive cell count in comparison to controls (p < 0.001). The aforementioned trends accord with the results by Darsalia et al., (2012) wherein Ex-4 treatment over two weeks was associated with a 50% in neuroblast production when compared to PBS-treated rats (p < 50%).

Furthermore, analysis using Ki67 staining identified that Ex-4 treatment over two weeks was associated with a two-fold increase in proliferating cells in the SVZ of the hippocampus and dentate gyrus (p < 0.05) (Darsalia et al., 2012). However, no effects of Ex-4 were reported when a NeuN/BrdU double marker was used to examine the effects of Ex-4 treatment (Darsalia et al., 2012). Notwithstanding, Hamilton et al., (2011) reported a 65 and 63% increase in BrdU-positive cells in Ex-4 treated obese (ob/ob) mice and diabetic (db/db) mice (mice with a point mutation causing leptin receptor deficiency) (p < 0.001) in comparison to saline controls, respectively (Hamilton et al., 2011; Suriano et al., 2021). High-fat diet mice treated with Ex-4 also exhibited a significant increase in DCX-positive neurons in comparison to saline-treated mice (p < 0.001) (Hamilton et al., 2011). These trends are further reinforced by findings from Bertilsson et al., (2007), wherein bis in die administration of Ex-4 for 7 days in rats was associated with a near two-fold and 70% increase in BrdU-positive cells in the SVZ and DCX-positive cells in the medial striatum, respectively (p < 0.01) (Bertilsson et al., 2007). Consistent with the aforementioned trends, results from Belsham et al., (2009) identified that Ex-4 treatment over a one-week period was associated with a two-fold increase in BrdU-positive cells (p < 0.05) in mice (Belsham et al., 2009). These findings suggest the presence of a positive association between exenatide and neurogenesis.

There is currently insufficient information on the effects of geniposide on neurogenesis. Notwithstanding, an experimental study by Sun et al., (2021) examined the effect of geniposide in mice. Findings suggest that geniposide mediated an increase in DCX-positive cells and dendrites in mice treated with geniposide in comparison to controls (p < 0.05). Due to the limited data available, a comprehensive evaluation of geniposide’s effect on neurogenesis cannot be made.

We identified 6 studies examining the role of liraglutide in neurogenesis (Hamilton et al., 2011; Hunter & Hölscher, 2012; Parthsarathy & Hölscher, 2013; Weina et al., 2018; Salles et al., 2018; Zhao et al., 2022). Weina et al., (2018) reported a significant increase in the density of DCX-positive immature neurons in mice injected with corticosterone and liraglutide in comparison with mice injected with corticosterone and saline (p = 0.0456). Additionally, ob/ob mice treated with liraglutide exhibited a 65% increase in BrdU-positive cells, wherein db/db mice treated with liraglutide exhibited an 88% increase in BrdU-positive cells in comparison to saline controls (p < 0.001) (Hamilton et al., 2011). Similarly, a significant increase in neuronal count in the dentate gyrus of mice injected with liraglutide in comparison to controls (p < 0.05) was observed (Zhao et al., 2022). Simultaneously, liraglutide administration was also associated with increased activation of the Wnt/β-catenin pathway (p < 0.05) (Zhao et al., 2022). Additionally, liraglutide administration was associated with significant increases in cAMP levels in comparison to controls (p < 0.05) (Hunter & Hölscher, 2012). In a separate study by Salles et al., (2018), liraglutide administration was associated with increases in DCX-marked neuroblasts of wild-type mice (p < 0.05) and APP/PS1 mice (p < 0.01) in the SVZ when compared to saline treatment. In contradistinction, liraglutide did not significantly change levels of GFAP-marked astrocytes (p > 0.05) (Salles et al., 2018).-catenin pathway (p < 0.05) (Zhao et al., 2022). Additionally, liraglutide administration was associated with significant increases in cAMP levels in comparison to controls (p < 0.05) (Hunter & Hölscher, 2012). In a separate study by Salles et al., (2018), liraglutide administration was associated with increases in DCX-marked neuroblasts of wild-type mice (p < 0.05) and APP/PS1 mice (p < 0.01) in the SVZ when compared to saline treatment. In contradistinction, liraglutide did not significantly change levels of GFAP-marked astrocytes (p > 0.05) (Salles et al., 2018).

Parthsarathy & Hölscher (2013) identified that wild-type mice treated with liraglutide over a week exhibited increased cell proliferation at 3 months (90%), 6 months (63%), 12 months (114%), and 15 months (137%) of age (p < 0.05) in comparison to saline controls. Similarly, an increase in DCX-marked immature neurons was observed after liraglutide treatment in comparison to saline at 3 months (59%), 6 months (26%), 12 months (57%), and 15 months (61%) of age (p < 0.05) (Parthsarathy and Hölscher, 2013).

This trend was replicated in wild-type mice treated with liraglutide over 37 days, wherein significant increases in cell proliferation in the dentate gyrus were observed in comparison to saline controls at 3 months (20%; p = 0.0457), 6 months (22%; p = 0.0467), 12 months (36%; p = 0.0455), and 15 months (52%; p < 0.05) of age (Parthsarathy & Hölscher, 2013). Similarly, increased proliferation was observed using Ki67 immunostaining post-liraglutide treatment in comparison to saline treatment at 3 months (94%), 6 months (103%), 12 months (143%), and 15 months (122%) of age (Parthsarathy & Hölscher, 2013). Additionally, an increase in DCX-positive neurons post-liraglutide treatment in comparison to saline controls was also observed at 3 months (43%), 6 months (40%), 12 months (74%), and 15 months (68%) of age (p < 0.05) (Parthsarathy & Hölscher, 2013). Notwithstanding, no significant change in gliogenesis was observed in mice treated with liraglutide (p > 0.05) (Parthsarathy & Hölscher, 2013). Results suggest a positive association between liraglutide and neurogenesis, which may be mediated through changes in the Wnt/β-catenin pathway.-catenin pathway.

To examine the association between lixisenatide and neurogenesis, we identified two studies (Hunter & Hölscher, 2012; Ren et al., 2021). In an experimental study by Hunter & Hölscher (2012), mice treated with lixisenatide exhibited an 80% increase in neuronal proliferation marked by BrdU-positive cells in comparison to control (p < 0.01). Additionally, a 70% increase in proliferating neurons was observed in lixisenatide-treated mice in comparison to saline controls observed via DCX analysis (p < 0.05) (Hunter & Hölscher, 2012). Similarly, Ren et al., (2021) reported that intranasal lixisenatide administration was associated with increased numbers of BrdU and BrdU/DCX-marked cells in the olfactory bulb and hippocampus (p < 0.001). Taken together, these results suggest a positive association between lixisenatide and neurogenesis (Hunter & Hölscher, 2012; Ren et al., 2021).

There is insufficient evidence on the effect of semaglutide in neurogenesis. Notwithstanding, a study by Yang et al., (2019) reported that mice with middle cerebral artery occlusion administered with semaglutide had a significantly higher number of DCX-positive cells in comparison to saline administration (F = 25.277, p < 0.01) whilst also exhibiting increased expression of neurogenesis markers nestin, CXCR4, and SDF-1 (p < 0.01) (Yang et al., 2019). Contrastingly, semaglutide administration was also associated with a significant reduction of Iba1-positive cells in the hippocampus in comparison to saline treatment (p < 0.001) (Yang et al., 2019). Additionally, administration of semaglutide was associated with significant changes in apoptotic pathways (Yang et al., 2019). Notably, increases in proto-oncogene c-RAF (c-Raf) (p < 0.05), mitogen-activated protein kinase 1 (ERK2) (p < 0.01), B-cell lymphoma-2 (Bcl-2) (p < 0.05), and decreases in Caspase-3 (p < 0.05) were observed one week after semaglutide administration (Yang et al., 2019). Although these findings suggest that semaglutide is likely associated with increased levels of neurogenesis, further research is required to investigate changes in specific neuron types.