Sickle cell disease is the most widespread monogenic genetic disease in the world [1]. It disproportionately affects populations in sub-Saharan Africa and their diaspora. According to the latest analysis of the Global Burden of Disease, more than 7.7 million people were living with sickle cell disease worldwide in 2021, compared to 5.46 million in 2000, and more than 515,000 children are born with the disease each year [2]. The African region bears the brunt of this burden, with an estimated average prevalence of 2% and nearly 400,000 affected children born annually, 50% to 80% of whom die before the age of five in the absence of widespread newborn screening and structured care [1]. Benin is among the most affected countries in the West African sub-region, with a birth incidence exceeding 2% [2]. Beninese health authorities estimate that more than 4% of the population carries the disease, and approximately 5.1% of children under five years of age admitted to the pediatric emergency department of the National University Hospital Center-Hubert Koutoukou Maga in Cotonou were affected in 2019 [3]. Beyond the numbers, the disease imposes a considerable burden on patients, their families and the health system through unpredictable vaso-occlusive crises, chronic hemolytic anemia, increased susceptibility to infections and the high cost of optimal care, largely out of reach for a large part of the population of sub-Saharan Africa. The inaccessibility of conventional treatments naturally leads populations to turn to traditional medicine.

A recent systematic review encompassing all the literature published between 2000 and 2022 documented 117 plant species with proven anti-sickling activity, with Africa alone accounting for nearly 87% of the publications on the subject [4]. Among these plants, Uvaria chamae P. Beauv. (Annonaceae) occupies a special place in traditional West African medicine, where its roots are used in the management of sickle cell disease. This anti-sickling activity has been experimentally confirmed. The methanolic extract of the roots inhibits and reverses erythrocyte sickling in vitro [5], and more recent work conducted in Benin specifically validated the anti-sickling activity of the aqueous extract of the plant’s roots [6], thus confirming the local traditional use. Different parts of the plant are traditionally used to treat various health problems such as digestive disorders, fever, dysmenorrhea, and cancer, and to promote wound healing, among many other applications [7] [8]. However, the question of its safety remains a concern. Several toxicological studies conducted on rats have indeed qualified the image of a harmless plant. A subchronic toxicity study of the ethanolic extract of the roots showed a significant increase in serum aspartate aminotransferase at the highest doses, associated with significant mortality in the treated animals, with only 43% surviving to the end of the protocol [9]. Furthermore, an evaluation of the cytotoxicity and genotoxicity of the methanolic extract of the leaves of this plant in mice revealed a dose-dependent reduction in the mitotic index of bone marrow cells as well as a significant increase in chromosomal aberrations compared to controls, concluding that the plant has both cytotoxic and genotoxic potential [10]. These results contrast with those of acute toxicity trials conducted orally, which generally place the median lethal dose above 5000 mg/kg, a threshold classically considered reassuring but which does not, on its own, allow us to rule out a cumulative or specific tissue risk. Similarly, phytochemical screenings carried out on different parts of the plant have highlighted, in addition to flavonoids sought for their therapeutic properties, the recurring presence of tannins, alkaloids, cardiotonic glycosides and cyanogenic glycosides [5] [11], all families of compounds whose potential toxicity, at high doses or in prolonged use, is well documented in the pharmacological literature.

Larval toxicity tests provide additional but equally contrasting insights. Lethality tests on Artemia salina larvae (Brine Shrimp Lethality Test), frequently used as a preliminary cytotoxicity screening, generally classify Uvaria chamae extracts as non-toxic, with high median lethal concentrations, on the order of 27.8 mg/mL [12]; a more recent evaluation confirms this conclusion of non-toxicity on the same model [13]. This result, obtained in a simple aquatic organism, cannot, however, be directly extrapolated to systemic toxicity in mammals. Therefore, there is currently an unresolved discrepancy between the different models for assessing the plant’s biological tolerance.

In addition to this toxicological uncertainty, there is a major methodological gap, rarely taken into account in published studies: the intraspecific variability of Uvaria chamae. Studies conducted in Benin have indeed revealed substantial genetic and morphological diversity among the plant populations distributed throughout the national territory, with a structuring into distinct genetic groups according to geographical areas [14] [15]. However, almost all toxicity studies available to date have been conducted on single accessions, without taking into account the genetic and environmental variability that can modulate the phytochemical composition and therefore the biological tolerance profile of the extracts. A recent ethnobotanical study conducted in Benin highlights the importance of examining the biochemical and toxicological properties of the plant extracts to confirm or refute the virtues traditionally attributed to it [16]. This gap is all the more worrying as the roots of Uvaria chamae, the most toxic part according to several phytochemical screenings, are precisely those favored by traditional medicine in the treatment of sickle cell disease, potentially exposing patients already weakened by their chronic pathology to extracts whose biological tolerance has never been established on the scale of the different Beninese accessions.

It is in this context that the present study is justified. Faced with the widespread traditional use of Uvaria chamae in the management of sickle cell disease, but in the absence of solid toxicological data taking into account the intraspecific diversity of the plant in Benin, it appears essential to evaluate, in a systematic and comparative way, the biological safety of aqueous extracts of roots from several accessions representative of the national territory. This approach aims to bridge the gap between the pharmacological validation of the plant’s anti-sickle cell activity and its safety, an essential prerequisite for any prospect of therapeutic development.

This study aimed to evaluate the biological safety of aqueous extracts of five accessions of Uvaria chamae (Annonaceae), a medicinal plant whose roots are traditionally used in Benin for the management of sickle cell disease.

In the acute oral toxicity study, oral administration of the aqueous extract of each accession of Uvaria chamae at a single dose of 2000 mg/kg did not cause death in rats. During the experimental period, there were no signs of toxicity such as agitation, aggression, body twisting, convulsions, diarrhea, asthenia, drowsiness, anorexia or decreased mobility. We also noted no significant difference in urea levels, ALT and platelet count between D0 and D14 (P > 0.05). The absence of mortality at 2000 mg/kg is consistent with acute oral toxicity studies conducted in Nigeria on aqueous root [27] and leaf [28] extracts of Uvaria chamae. Serum creatinine increased in all five treated groups between D0 and D14, although not significantly. Because this rise was observed with all accessions, factors unrelated to treatment, such as growth or analytical variation between runs, are likely to contribute. The control group showed a rise of similar magnitude (from 6.2 ± 0.6 mg/L to 10.1 ± 1.6 mg/L), which suggests that this change is not attributable to the extracts. Renal histology showed no visible lesions and blood urea did not increase; together, these findings do not indicate overt acute renal injury, although a subtle functional effect cannot be ruled out. The AST level significantly decreased on D14 in the group of rats treated with the UV9 extract (P = 0.01). Compared with the control group, however, this decrease did not differ significantly (Dunnett’s test, P = 0.15), and the decrease in AST was significant only with UV13 (P = 0.03), a group with a high baseline value. A similar, though not statistically significant, decrease in AST was observed with UV7 and UV19, and ALT decreased in most groups, including the control group. The decrease in serum levels of these two markers, combined with normal liver parenchyma on histology, does not suggest hepatic cytolysis. However, it cannot be interpreted as a hepatoprotective effect, the demonstration of which would require an induced hepatotoxicity model such as carbon tetrachloride intoxication, used by Omoruyi et al. [29] to demonstrate the hepatoprotective effect of the aqueous extract of Uvaria afzelii root in the Wistar rat. In contrast, Legba et al. [30], in Benin, reported significant increases in AST and ALT after acute oral administration of Uvaria chamae root extracts, without histological abnormalities, and recommended using doses below 2000 mg/kg. Olumese et al. [27] reported no major atypia in the histological examinations of rat livers, apart from slight vascular congestion.

The mean hemoglobin level increased significantly in rats treated with extracts from accessions UV9 and UV13 (P = 0.02 and P = 0.04). Furthermore, a non-significant increase was observed in the remaining groups (UV2, UV7 and UV19). This observation could be compatible with stimulation of erythropoiesis by these accessions, but other explanations remain plausible, including transient hemoconcentration after administration of a single high dose or a physiological increase related to growth; differentiating between these would require reticulocyte counts and erythrocyte indices, which were not measured. The increase in hemoglobin also observed in the control group (P = 0.07) favors a physiological increase related to growth rather than an effect of the extracts. Increases in hemoglobin have also been reported in two studies conducted in Nigeria with root and leaf extracts of the plant [27] [28].

The body weight decreases observed with UV7 in the acute test (−14.1%) and with UV7 and UV19 in the subchronic test (−8.1% and −7.9%) deserve attention; the decrease with UV7 in the subchronic test was statistically significant (P = 0.03). The rats used were young and still growing, so that a stable or decreasing body weight over 14 to 28 days represents a deviation from the expected growth trajectory; by comparison, mean body weight of control rats increased by 3.4% over 14 days and by 7.1% over 28 days. A decrease of this magnitude may reflect reduced food intake, reduced palatability of the extract, gastrointestinal discomfort or a systemic effect; food consumption was not measured, which prevents these possibilities from being distinguished. With UV7, the weight decrease coincided with the largest transaminase increases in the subchronic test, including the only ALT increase that differed significantly from the control group, so a treatment-related effect of this accession cannot be excluded. These observations temper the conclusion of good tolerability for UV7 and, to a lesser extent, for UV19, and justify including food intake and body-weight gain relative to controls as endpoints in future studies.

For the study of subchronic oral toxicity, oral administration of the aqueous extract of each of the five Uvaria chamae accessions at a dose of 200 mg/kg per day for 28 consecutive days induced a significant decrease in body weight only with UV7, and a non-significant decrease with UV19, as discussed above. For the evaluation of renal toxicity, serum creatinine did not significantly increase between D0 and D28 (P > 0.05). Blood urea levels did not increase on D28 in the treated rat groups and, conversely, decreased significantly in the group receiving the UV9 extract (P = 0.03). These blood urea and creatinine results do not suggest renal toxicity, and the normal histology of the renal parenchyma observed in the subchronic oral toxicity test is consistent with this interpretation. Comparable observations have been reported with other aqueous extracts of medicinal plants administered orally for 28 days in rats. Tcheutchoua et al. [31] noted no significant effect on serum creatinine or uremia at doses of 200 to 800 mg/kg, with serum creatinine even decreasing significantly in males at doses of 200 and 400 mg/kg, without any link to hepatic, renal, pulmonary or splenic damage on histology. In contrast, Dossou-Agoin et al. [32], applying OECD Test Guideline 407 in male rats to aqueous extracts of Pedalium murex at the same dose levels in Benin, observed proteinuria and renal damage indicating potential nephrotoxicity, which shows that the absence of renal effects cannot be generalized from one plant extract to another. An isolated decrease in blood urea or creatinine levels, not accompanied by renal histological lesions, is generally not considered harmful.

Addressing the liver toxicity of Uvaria chamae extracts, the groups treated with extracts from accessions UV2 and UV7 showed a significant increase in AST and ALT (P < 0.05), and the UV19 group a significant increase in AST (P = 0.02); ALT also increased with UV13 and UV19, without reaching significance. In interpreting these changes, it should be noted that the absolute AST and ALT activities measured in this study (group means of about 1 to 18 U/L) were markedly lower than those generally reported for young Wistar rats with standard methods [25] [26], for example 61 to 153 U/L for AST and 23 to 67 U/L for ALT in female Wistar rats [26]. Such a systematic difference suggests an analytical offset related to the assay method, calibration or sample handling. Consequently, these values cannot be meaningfully compared with published physiological limits. The control rats also showed low activities (Table 2 and Table 4), which is consistent with an analytical offset affecting all samples rather than an effect of treatment. The changes must therefore be assessed against the concurrent control group. On this basis, the AST increases observed after 28 days largely reflect low baseline values in the treated groups: on D28, AST in the treated groups (9.80 to 18.24 U/L) was close to that of the control group (12.48 U/L), and none of the AST changes differed significantly from the control (Dunnett’s test, P ≥ 0.34). On D28, ALT was 0.66 to 1.37 times the control value in the UV2, UV9, UV13 and UV19 groups, but 2.05 times the control value in the UV7 group. Because an analytical offset may be additive, fold changes calculated from very low baseline values should not be over-interpreted, and the ALT increase with UV7, ALT being the more liver-specific enzyme in the rat [33], is the most robust signal of possible hepatocellular stress. The adversity of such changes should be judged by integrating clinical pathology with histopathology [34]-[36]. In the subchronic test, only the ALT increase with UV7 differed significantly from the control group (Dunnett’s test, P = 0.006), which identifies UV7 as the accession of greatest concern; given the low power of a design with three animals per group, the absence of significance for the other accessions does not exclude an effect. These results are consistent with those of Olumese et al. [9], who observed a decrease in urea levels and a significant increase in AST and ALT. The same team had previously reported, with the aqueous root extract, a significant increase in AST at 5000 mg/kg and in ALT at 2000 mg/kg [27]. In contrast, Tcheutchoua et al. [31] observed no significant variation in mean AST and ALT levels after 28 days of oral administration of an aqueous plant extract in rats, illustrating the variability of enzymatic responses depending on the accession and plant species considered.

In histology, the liver parenchyma of rats treated with extracts from the five accessions of Uvaria chamae did not show any visible atypia in the subchronic oral toxicity test. This biochemical-histological discrepancy is a well-documented phenomenon in experimental toxicology and underscores the need never to interpret an enzymatic parameter in isolation, but always in light of the overall functional and morphological profile [33] [35] [36]. The heterogeneity observed between accessions may reflect differences in phytochemical composition; however, because the extracts were not chemically characterized in this study, this explanation remains a hypothesis to be tested. Moreover, the toxicological profile of Uvaria chamae also depends on the organ sampled and the extraction solvent, as illustrated by the work carried out separately on the aqueous root extract [27], the ethanolic root extract [9] and the n-hexane fraction of leaves [37]. Comparisons between studies must therefore take these parameters into account before any generalizations are made. These results also call for caution in interpreting the data. Indeed, while the traditional use of Uvaria chamae to relieve sickle cell crises appears to present a low risk in view of the results of the acute toxicity test, repeated or prolonged administration, as might be envisaged in a long-term prevention strategy for sickle cell disease, would require monitoring of liver function. This precaution appears particularly relevant for UV7 and, to a lesser extent, for UV2 and UV19, which showed increases in liver enzymes relative to baseline.

Regarding hematological parameters, the number of leukocytes did not significantly vary on D28 in the different groups of rats compared to D0. However, total leukocyte counts alone do not assess immune function, and no conclusion on immunotoxicity can be drawn from this result. This result corroborates those obtained by Tcheutchoua et al. [31], who noted no significant variation in leukocyte count after 28 days of oral administration of an aqueous extract in rats.

Hemoglobin levels did not significantly change in the treated groups on D28, and no evidence of an effect on the erythrocyte lineage was detected. In the thrombocyte lineage, only the UV2 extract significantly increased the platelet count on D28 (P = 0.04). A significant increase in platelet count has also been reported in rats treated with Uvaria chamae root extract at a dose of 500 mg/kg [9]. This concordance could point to an effect of the species on platelet counts rather than a peculiarity of the UV2 accession; this remains speculative, since megakaryopoiesis and bone marrow were not examined.

In summary, no deaths, overt clinical signs or histological lesions of the liver and kidneys were observed. However, body weight decreased in some groups, notably with UV7, and, relative to baseline, repeated administration increased AST with UV2, UV7 and UV19 and ALT with UV2 and UV7, although only the ALT increase with UV7 differed significantly from the control group. Lokonon et al. [38], who studied aqueous root extracts from a single accession of Uvaria chamae at the same doses in Wistar rats, also observed no deaths or histological lesions, but found no change in AST or ALT; the transaminase increases observed here with some accessions illustrate the value of testing several accessions. After 28 days, analysis of hematological parameters showed no variation in leukocyte or hemoglobin levels and an isolated increase in platelets in the group treated with the UV2 accession, a finding also reported in Nigeria by Olumese et al. [9]. Overall, the data do not show overt subchronic toxicity at 200 mg/kg/day, but they are not sufficient to establish the safety of prolonged use.

Limitations of the Study

This work has several limitations that should be kept in mind when interpreting the results.

First, the small number of animals per group (n = 3) results in low statistical power: non-significant differences cannot be interpreted as evidence of the absence of an effect, and isolated significant results should be interpreted with caution. In addition, the paired comparisons with baseline were not adjusted for multiple comparisons, and Dunnett’s test adjusts for the comparisons with the control within each parameter but not across the many parameters examined.

Second, the subchronic toxicity trial was conducted at only one dose (200 mg/kg/day), which does not allow for establishing a dose-response relationship or determining a no-observed-adverse-effect level (NOAEL). Both studies were adaptations of OECD Test Guidelines 425 and 407 rather than fully compliant studies, and only female rats were used.

Third, food and water consumption, organ weights, reticulocyte counts and erythrocyte indices were not measured, and the extracts were not chemically characterized, so that differences between accessions cannot be linked to their composition. In addition, the low absolute transaminase activities limit comparison with published reference values, and the calibration records of the transaminase assays were not available for this revision.

Finally, the 28-day exposure period does not guarantee the safety of prolonged chronic use, and the results obtained in healthy Wistar rats are not directly applicable to sickle cell patients, whose liver and kidneys are already under stress from the disease. Chronic toxicity studies, at multiple doses and combined with phytochemical analysis of the extracts, are needed to complete this work.

The main objective of this study was to systematically and comparatively evaluate the safety of aqueous extracts of Uvaria chamae roots from several representative accessions in Benin. This approach aims to bridge the gap between the pharmacological validation of the plant’s anti-sickle cell activity and its safety, an essential prerequisite for any potential therapeutic use. Acute and subchronic oral toxicity tests revealed no deaths or overt clinical signs of toxicity in Wistar rats, and histological examinations showed no visible atypia in the hepatic and renal parenchyma of rats treated with root extracts from the five accessions. However, several biochemical changes were observed, in particular increases in AST (UV2, UV7 and UV19) and ALT (UV2 and UV7) relative to baseline after 28 days, of which only the ALT increase with UV7 differed significantly from the control group, together with body weight decreases with UV7 and UV19. UV7 was therefore the accession of greatest concern, as it also caused a significant decrease in body weight after 28 days.

Overall, these results provide preliminary evidence that aqueous extracts of Uvaria chamae roots are tolerated at the doses tested, namely 2000 mg/kg as a single dose and 200 mg/kg/day for 28 days. Given the small sample size of three animals per group, the single subchronic dose assessed, and the absence of an established dose-response relationship, these findings remain exploratory. Liver function monitoring is recommended in cases of repeated or prolonged administration, particularly for UV7.

Chronic toxicity studies, at multiple doses and coupled with phytochemical characterization of accessions, are necessary before any prospect of therapeutic development.

Not applicable.

HWH, FA and JEL participated in the experimental work and in drafting and revising the manuscript. TDMA, RCH, GA, CN, EM, GD, PAT, AM, CA and AAEA participated in reviewing and validating the manuscript. MS supervised the entire study, controlled and revised the statistical analysis, and reviewed, corrected, validated, and translated the manuscript. All authors read and approved the final version of the manuscript.

This study received no specific funding from public, commercial, or non-profit organizations.

The data generated and analyzed during this study are available from the corresponding author upon reasonable request.

During the preparation of this manuscript, the authors used the Claude assistant (Anthropic) for editorial and technical assistance, including: language editing and proofreading of the English text; consolidation and formatting of the reference list; verification and correction of citations and bibliographic references; verification of internal numerical and terminological consistency throughout the manuscript; reformatting of the manuscript to comply with the submission requirements of the journal; and, during the revision, execution of the statistical calculations (descriptive statistics, paired Student’s t-tests, one-way ANOVA and Dunnett’s tests) on the individual raw data provided by the authors, using the SciPy library in Python, according to the analysis plan defined in response to the reviewer. The authors reviewed, verified, and approved all AI-assisted edits and calculations and take full responsibility for the accuracy and content of the publication. No artificial intelligence tool was used to generate experimental data, and no such tool is listed among the authors.