The study adopted a laboratory-based experimental design.

The study was conducted in Zambia, a landlocked nation situated in sub-Saharan Africa. Covering a geographical area of 753,612 square kilometers, Zambia is located between latitudes 8° and 18° South and longitudes 22° and 34° East. Phytochemical analyses were conducted at the Department of Chemistry, University of Zambia, while the biochemical analyses were done at the Department of Physiological Sciences, School of Medicine, University of Zambia, Lusaka, Zambia.

The study population comprised Wistar rats obtained from the University of Zambia. This study adopted the guidelines for ethical conduct in the care and use of animals provided by the American Association of Psychologists. Ethical approval was sought from the University of Zambia Biomedical Research Ethics Committee (Ref. No. 2892-2022) and the National Health Research Authority (NHRA-1656/23/10/2024).

The sample size was calculated using R Studio. To determine the effect of the leaf extracts of L. edulis on glucose absorption in the everted rat jejunum, the sample size for Kruskal-Wallis for 7 groups was determined. The power level was set at 80%, the significance level at 5%, and a large effect size was assumed. The sample size for 7 groups was determined to be 42. To investigate the effects of the leaf extracts of L. edulis on glucose uptake in the isolated hemidiaphragm, the sample size for five groups was determined to be 30. Lastly, to analyze the effects of the leaf extracts of L. edulis on the histology of the pancreas, the sample size was determined to be 20 using the same method outlined above.

Lannea edulis whole plant was obtained from Chongwe, 47 km east of Lusaka, with the help of a field assistant. The plant was then taken to the Department of Biological Sciences at the University of Zambia for identification and authentication by a qualified botanist. A voucher specimen with accession number 22523 was deposited at the herbarium situated in the School of Natural Sciences, Department of Biological Sciences at The University of Zambia.

The leaves of L. edulis were washed with running water to eliminate surface contaminants and air-dried in the shade for 7 days. This was followed by blending the leaves to homogeneity. The blended leaves were then placed in hot water and allowed to sit for 30 min. The water was then filtered using MN615 150 mm filter paper. The filtrate was evaporated to dryness using a hot water bath at 40°C to obtain a sticky brownish residue. The residue was weighed, stored in air and water-proof containers, and then kept between 4°C and 8°C. A fresh preparation was made from this stock whenever required.

The percentage yield was determined as follows: P e r c e n t a g e   y i e l d = M a s s   o f   e x t r a c t   g M a s s   o f   g r o u n d   l e a v e s   g   X 100 %

The presence of various secondary metabolites in the leaf extracts of L. edulis was assessed using different standard procedures (Shaikh and Patil, 2020; Altemimi et al., 2017; Abuzaid et al., 2020; Nortjie et al., 2022; Pandey et al., 2014; Patil and Deshmukh, 2016). We assessed the presence of phenols, flavonoids, saponins, glycosides, steroids, terpenoids, alkaloids, carbohydrates, and amino acids.

Glucose uptake by the rat diaphragm was examined according to standard protocol (Mohamed et al., 2013). Rats weighing 150–200 g were sacrificed by cervical dislocation and dissected to remove the diaphragm. Each diaphragm was rinsed in ice-cold Tyrode’s solution to preserve the tissue, and then cut into two equal hemidiaphragms, which were gently blotted and weighed. The composition of Tyrode’s solution is as follows: NaCl (134 mM), KCl (2.68 mM), CaCl₂ (1.80 mM), MgCl₂ (1.05 mM), NaH₂PO₄ (417 µM), NaHCO₃ (11.9 mM), and glucose (5.56 mM). This solution is a balanced salt solution that is isotonic with interstitial fluid and is used in experiments to mimic normal physiological conditions (Barreto-Chang and Dolmetsch, 2009; Gilani et al., 2001). Each hemidiaphragm was placed in a test tube containing 2 mL of Tyrode’s solution and test sample with constant aeration by bubbling air throughout the experiment. Since Tyrodes’ reagent contains glucose, well-preserved tissues are expected to take up glucose from the solution. The tissues were constantly shaken at 37°C and incubated for 90 min.

Triplicate samples of each medium were taken for glucose estimation before the placement of the tissues and at the end of the experiment. The hemidiaphragms were randomly placed in one of the five groups (n = 5) as shown in Table 1. Equal numbers of left and right hemidiaphragms were used for controls and test samples, respectively and dosages of the leaf extract used in the study were established after preliminary tests were conducted using a similar study for reference (Mohamed et al., 2013). The results obtained were expressed as the amount of glucose utilized by the hemidiaphragms per gram of tissue during the 90-min incubation period. The glucose concentration was measured using an Accu-Chek glucose meter. The dosages used are commonly employed in studies involving the isolated hemidiaphragm (Mohamed et al., 2013).

Healthy experimental rats that weighed between 150 and 200 g were fasted for 16 h before the experiment. They were then sacrificed by a cervical dislocation, and dissected to remove the small intestine. The jejunum segments were excised and placed in ice-cold Tyrode’s solution, a physiological solution used to maintain tissue viability. Then, the segments were rinsed with Tyrode’s reagent, and everted onto a glass rod.

The segments were then cut into 5 cm pieces (n = 6), rinsed, and placed on filter paper to dry before being weighed. Each piece of the intestine was tied tightly around one end with a thread, while the other end only required a loose knot (Mohamed et al., 2013). Thereafter, 2 mL of Tyrode’s solution was injected into each sample using a slipping blunt needle, and the knot was tightened.

The jejunum samples were randomly placed in seven groups (n = 6), and then incubated in 30 mL Tyrode’s solutions as shown in Table 1. Metformin was used because different studies have shown that it inhibits intestinal glucose transport at varying concentrations (Horakova et al., 2019). The preparations were constantly aerated by bubbling air and kept at 37°C throughout the experiment. At the end of the 90-min incubation period, the glucose concentration of the inside sacs and the incubation media was determined using the Accu-chek glucose meter. Results were expressed as milligrams of glucose absorbed per gram of tissue (Mohamed et al., 2013).

Alloxan monohydrate (Sigma Aldrich) was used to cause pharmacological diabetes by intravenously injecting 60 mg/kg of Alloxan monohydrate dissolved in 0.9 percent w/v cold normal saline to overnight-fasted rats (12 h). The rats were then kept on a 10% glucose solution for 24 h to prevent hypoglycemia. After 72 h of injection, fasting blood glucose level was measured, and animals that developed more than 200 mg/dL glucose levels were considered diabetic. The rats were randomly allocated to five groups (n = 6) and subjected to different oral treatments for 14 days, as shown in Table 1. Toxicity studies have been conducted on the leaf extracts of the L. edulis, and the LD₅₀ was found to be greater than 6,000 mg/kg, and according to Hodge and Sterner, falls in the practically nontoxic range (Banda et al., 2018). This was referenced in this study to make sure the dosages of L. edulis used are safe and effective. At the end of the experimental period, rats from each group were surgically operated on under diethyl ether (1.9%) anesthesia (at 80 μL per litre of volume of a container). The pancreatic tissues were dissected out, and placed in 10% neutral buffered formalin and fixed in PBS containing 10% formalin. The tissues were then washed in running tap water, dehydrated in descending grades of isopropanol, and finally cleared in xylene. The tissues were thereafter embedded in molten paraffin wax, cut into transverse 5 μm sections of the mid-organ level, and stained with Hematoxylin-Eosin (Venkatesh et al., 2010; Mohamed et al., 2013; Al-Qudah et al., 2016; Atangwho et al., 2012). Histopathological changes in the pancreatic tissues were observed under a light microscope at a magnification of 400X. The number of islets and the number of β-cells in each islet were counted by three pathologists.

Pancreatic islet area was quantified using ImageJ (NIH). The pixel-to-micron scale was calibrated based on the microscope camera sensor and optical system, yielding a scale of 0.0875 µm/pixel. Using the freehand selection tool, individual islets of Langerhans were outlined manually, and their cross-sectional areas were measured in µm². A minimum of four islets per treatment were analyzed, and mean islet areas were compared across experimental groups using One-Way ANOVA.

The data was initially collected in MS Excel 2019 and exported to Stata version 17 for analysis. Data normality was determined using box plots, and it was found to be not normally distributed. This distribution was due to the small sample size that the study used. Given that data were not normally distributed, the median and interquartile range were used to summarize it. The Kruskal-Wallis non-parametric test and One-Way ANOVA were used to determine whether there were statistically significant differences between two or more intervention groups for glucose uptake and glucose absorption. Tukey’s HSD test and Dunn’s test with Bonferroni correction were used for post hoc testing. Data are presented as mean ± SD unless otherwise specified; graphical representations show mean ± SEM. The results were considered statistically significant if p value was less than 0.05.

Lannea edulis whole plant was collected from Chongwe, Zambia, with the help of a field assistant. The plant was then taken to the University of Zambia, Department of Biological Sciences, for professional identification and authentication after which it was assigned the unique accession number 22523. Following leaf extract preparation, the percentage yield of 13.1% was obtained; this is less than the amount reported previously (Banda et al., 2018). The L. edulis aqueous leaf extract was then used for phytochemical and biochemical analyses, as illustrated in Figure 1.

Several studies have reported the presence of phytochemicals in the leaf extracts of L. edulis (Banda et al., 2018; Kumar et al., 2023; Queiroz et al., 2003). To confirm this, the leaf extracts of L. edulis were screened for the presence of alkaloids, saponins, steroids, terpenoids, glycosides, flavonoids, phenols, tannins, carbohydrates, amino acids, and proteins. Our data shows that phytochemicals were present in varying amounts (Table 2).

The hemidiaphragm model is an ex-vivo system commonly used to study muscle glucose metabolism and evaluate anti-diabetic agents (Buse et al., 1976; Sabu and Subburaju, 2002; Koski and Max, 1974; Bati et al., 2017; Fernandes et al., 2007). To study the effects of L. edulis on glucose uptake in the isolated rat hemidiaphragm, five groups were established as shown in Materials and Methods. Our data, as illustrated in Table 3, indicates that the median glucose uptake in milligrams in the isolated rat hemidiaphragm model was highest in the insulin 1 IU/mL group [9.51 (9.33–10.75)] and lowest in the control group [4.98 (4.61–5.83)]. This data was further visualized using box plots (Figure 2). Our results show that the data were not normally distributed, indicating that the uptake of glucose in the isolated rat hemidiaphragm model varied across treatment groups. Given that our data was not normally distributed, we performed a Kruskal-Walli’s test to check if glucose uptake across groups was significantly different (Table 4). Our results show that there was a significant difference between the intervention groups and the negative control in the uptake of glucose in the hemidiaphragm model (p = 0.001). We then carried out a Dunn’s test to determine which of the interventions had a statistically significant effect. A post hoc test using Dunn’s test with Bonferroni correction, as shown in Table 5, showed that glucose uptake in the hemidiaphragm model was significantly higher in the L. edulis 1 mg/mL (p = 0.0029), L. edulis 2 mg/mL (p = 0.0479), and insulin 1 IU/mL (p = 0.0007) groups compared to the control. Additionally, the L. edulis 1 mg/mL group (p = 0.0266) and the insulin 1 IU/mL group (p = 0.0082) had significantly higher glucose uptake than the insulin 1 IU/mL + L. edulis 2 mg/mL group. This data suggests chemical antagonism between the leaf extracts of L. edulis and insulin.

The everted rat jejunum is an ex-vivo experimental model that mimics the physiological environment of the small intestine and is often used to study early-stage absorption (Kellett and Helliwell, 2000; Mace et al., 2007; Morgan et al., 2007; Alam et al., 2012; Emoto et al., 2000; Wilcock and Bailey, 1991). To evaluate the effects of L. edulis on the absorption of glucose in the everted rat jejunum, seven groups were formed, as shown in Materials and Methods. Our data, as shown in Tables 6, 8 indicates that the median absorption of glucose in the everted rat jejunum model was highest in the negative control group [5.26 (4.95–5.37)] and lowest in the L. edulis group (0.5 mg/mL) [3.38 (3.27–3.54)]. This data was further visualized using a box plot (Figure 3). Our data shows that there were variations in absorption of glucose by different treatment groups, suggesting that some interventions might have inhibited glucose absorption in the everted rat jejunum. The data was skewed, an indication the data was not normally distributed; this suggests that the extent of inhibition was not dependent on the concentration of the interventions. Since our data was not normally distributed, we performed a Kruskal-Walli’s test to confirm the differences in glucose absorption observed in Figure 3. As shown in Table 7, our results indicate that there was a significant difference in glucose absorption in the inverted rat jejunum model between the interventions and the negative control (p = 0.001). This data suggests that some interventions lowered the amount of glucose absorbed. Additionally, we carried out a Dunn’s test to determine which of the interventions had a statistically significant effect. A post hoc test using Dunn’s test with Bonferroni correction as shown in Table 8 indicated that glucose absorption in the everted rat jejunum was significantly inhibited by L. edulis 0.5 mg/mL (p < 0.001), metformin (1.5 mg/mL) (p < 0.001) and metformin (3.0 mg/mL) (p = 0.0043). Further, L. edulis (0.5 mg/mL) significantly inhibited glucose absorption than L. edulis (0.25 mg/mL) and (1 mg/mL) groups (p = 0.0273 and p = 0.0136). Lannea edulis (0.5 mg/mL) significantly lowered glucose absorption than the metformin group (0.75 mg/mL) (p = 0.007). The absorption of glucose was significantly higher in the metformin (0.75 mg/mL) group than in the metformin (1.5 mg/mL) group (p = 0.0098). Collectively, this data shows that the leaf extracts of L. edulis significantly decreased glucose absorption in the everted rat jejunum at a concentration of 0.5 mg/mL.

The baseline fasting blood glucose levels were within the normal range, measuring between 3.3 mmol/L and 4.2 mmol/L before diabetes induction. Alloxan, used to induce diabetes in the current experiment, selectively destroys pancreatic β-cells (Lenzen, 2008; Malaisse et al., 1982). Diabetes was confirmed by post induction blood glucose levels of more than 200 mg/dL. Further, the experimental diabetic rats exhibited classical signs of diabetes mellitus, including polydipsia, polyuria, observable lethargy, and progressive weight loss. To analyze the effect of L. edulis leaf extracts on the histological morphology of β-cells in the pancreas of rats, we prepared pancreatic sections from the five groups as described in Materials and Methods, following induction of diabetes and administration with the extracts. Our data shows that the sections obtained from the normal control group treated with normal saline exhibited normal pancreatic histology and normal islet structure (Figure 4A). Furthermore, our data showed that the pancreatic sections from the alloxan-induced diabetic group treated with normal saline showed very few islets, with a notable reduction in size (Figure 4B). The high-power view photomicrographs of the pancreatic sections from the alloxan-induced diabetic group treated with vitamin C (150 mg/kg) showed islets close to normal size with signs of β-cell regeneration (Figure 4C). The high-power view photomicrographs of the pancreatic sections from the alloxan-induced diabetic group treated with L. edulis (500 mg/kg) showed normal-sized islets with regeneration of β-cells (Figure 4D). The photomicrographs of sections obtained from the alloxan-induced diabetic group treated with a lower dose of L. edulis (100 mg/kg) showed reduced islet size and reduced number of β-cells (Figure 4E). To confirm if there were significant differences in the size of islets among the five treatment groups, we quantified the islet areas as earlier described and then performed a One-Way ANOVA and Tukey HSD test (Figure 4E). Our data revealed significant differences in islet area across the experimental groups (One-Way ANOVA, p < 0.0001). As expected, diabetic control rats exhibited a marked reduction in islet area (mean: 3,968.83 ± 61.18 µm²) compared to normal controls (mean: 12,559.92 ± 36.48 µm²), confirming alloxan-induced β-cell destruction (p < 0.001, Tukey HSD test). Treatment with vitamin C (150 mg/kg) significantly increased islet area to 9,358.81 ± 281.25 µm² (p < 0.001 vs. diabetic control), indicating partial regenerative potential. Notably, administration of L. edulis extract at 100 mg/kg modestly improved islet area (6,794.73 ± 26.41 µm²), while the 500 mg/kg dose almost completely restored islet size (12,202.05 ± 87.13 µm²), with no statistically significant difference from the normal control group (p > 0.05). These findings demonstrate a dose-dependent β-cell protective and regenerative effect of L. edulis extract in diabetic rats. To confirm if there were significant differences in the number of islets and the number of β-cell among the five treatment groups, we performed the Kruskal-Walli’s test. The results comparing the number of β-cells indicated a statistically significant difference between the groups (χ² (6) = 27.204, p = 0.0001) (Table 9). This suggests that at least one of the treatment groups had a significantly different number of β-cells from the others. For the number of islets, the analysis also revealed a statistically significant difference among these groups (χ² (6) = 25.227; p = 0.0001) (Table 9). These results indicated that at least one of the treatment groups had a significantly different distribution (or median, assuming similar distribution shapes) of the number of islets compared to the others. We performed Dunn’s test to identify specific differences in β-cell count among the various treatment groups following a significant Kruskal-Walli’s test (Table 10). The group treated with Vitamin C (150 mg/kg) showed significantly higher β-cell count compared to L. edulis (100 mg/kg) (p = 0.0058). The diabetic control group treated with normal saline (1 mL) had significantly lower β-cell ranks than the groups treated with Vitamin C (150 mg/kg) (p = 0.0001) and L. edulis (500 mg/kg) (p = 0.0091). This is consistent with other studies that have reported the regenerative effects of Vitamin C on the pancreas (Gui et al., 2023). The normal control group treated with normal saline (1 mL) showed significantly higher β-cell ranks compared to L. edulis (500 mg/kg) (p = 0.0140), L. edulis (100 mg/kg) (p = 0.0004), and especially diabetic control group treated with normal saline (1 mL) (p = 0.00001). No significant differences were observed between the groups treated with Vitamin C (150 mg/kg) and L. edulis (500 mg/kg) (p = 0.0893). These findings suggest that L. edulis at the dosage of 500 mg/kg could have regenerative effects on the β-cells of the pancreas. Additionally, Dunn’s test was performed to identify specific differences in islets among the various treatment groups following a significant Kruskal-Walli’s test (Table 11). The results show that there was no significant difference in islet ranks between the groups treated with Vitamin C (150 mg/kg) and L. edulis (500 mg/kg) (p = 0.0812). The diabetic control group treated with normal saline had significantly lower islet ranks compared to the groups treated with L. edulis (500 mg/kg) (p = 0.0253) and L. edulis (100 mg/kg) (p = 0.0366). These unadjusted results show that the groups treated with L. edulis generally lead to higher islet ranks compared to the diabetic control group, suggesting possible regenerative and augmentation effects of the leaf extracts of L. edulis on the islets. Together, this data suggests that at higher doses, the leaf extracts of L. edulis have an antioxidant and regenerative effect on the pancreas.