Antiosteoporosis Effect and Possible Mechanisms of the Ingredients of Fructus Psoraleae in Animal Models of Osteoporosis: A Preclinical Systematic Review and Meta-Analysis

Objective Fructus Psoraleae (FP) and its ingredients (IFP) have a variety of biological activities and are widely used to treat osteoporosis (OP). Herein, we conducted a systematic review to evaluate the efficacy of IFP for an animal model of OP from the current literatures. Potential mechanisms of IFP in the treatment of OP were also summarized. Materials and Methods We carried out a search for electronic literature in the PubMed, Chinese National Knowledge Infrastructure, EMBASE, Wanfang, Web of Science, Chinese Biomedical Literature Database, and Cochrane Library, as well as Chinese VIP databases targeting articles published from inception to June 2021. The inclusion criteria were animal studies that assessed the efficacy and safety of IFP for OP, regardless of publication status or language. The exclusion criteria included (1) other types of studies (in vitro studies, case reports, clinical trials, reviews, abstracts, comments, and editorials), (2) combination with other compounds, (3) compared with other traditional Chinese medicine, (4) not osteoporosis or bone loss model, (5) studies with insufficient data, (6) lack of a control group, and (7) duplicate publications. The modified Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Stroke (CAMARADES) 10-item quality checklist was used to evaluate the risk of bias of included studies. We computed the relative risk (RR) and the standard mean difference (SMD) for dichotomous outcomes and continuous outcomes, respectively. When heterogeneity was detected or there was significant statistical heterogeneity (P < 0.05 or I2 > 50%), a random-effects model was employed, followed by further subgroup analysis and metaregression estimations to ascertain the origins of heterogeneity. Otherwise, we used a fixed-effects model (P ≥ 0.05 or I2 ≤ 50%). The primary outcome measures were bone mineral density (BMD), serum osteocalcin(S-OCN), bone volume over total volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), bone maximum load, and elasticity modulus. The secondary outcome measure was the antiosteoporosis mechanisms of IFP. The STATA 12.0 software was used to analyze the data. Results Overall, 16 studies focusing on 379 animals were enrolled into the study. The risk of bias score of included studies ranged from 4 to 7 with an average score of 5.25. The present study provided the preliminary preclinical evidence that administration of IFP could significantly increase the S-OCN, BMD, BV/TV, and Tb.N while Tb.Th and Tb.Sp were remarkably decreased by IFP in OP model animals (P < 0.05). Moreover, IFP could significantly improve the bone biomechanical indicator bone maximum load and elasticity modulus (P < 0.05). In terms of the possible mechanisms of treatment of OP, IFP exerts anti-OP effects in animal models probably through osteoprotegerin/receptor activator of the nuclear factor-κB ligand/receptor activator of nuclear factor-κB (OPG/RANKL/RANK), peroxisome proliferator activated receptor γ (PPAR-γ)/Axin2/Wnt, antioxidative stress via forkhead box O3a (FoxO3a)/Axin2/Wnt, phosphatidylinositol 3-kinase/protein kinase B/mammalian target of rapamycin (PI3K/Akt/mTOR), estrogen-like effect, and gamma-aminobutyric acid/gamma-aminobutyric acid receptor (GABA/GABABRI) signaling pathway. Conclusion Taken together, the findings suggest the possibility of developing IFP as a drug or an ingredient in diet for the clinical treatment of OP. We recommend that rigorous, as well as high-quality, trials involving large sample sizes should be conducted to confirm our findings.


Introduction
Osteoporosis (OP), is a systemic skeletal disease characterized by loss of bone mass and bone microarchitectural deterioration, resulting in increased bone fragility and a greater risk of fractures, especially in the spine, hip, and wrist [1,2]. Many risk factors are associated with OP, including age, race, smoking, alcohol, low physical activity, hormone-related factors, drugs such as glucocorticoids, low calcium, and vitamin D levels [3]. Reports have showed that about 9.9 million Americans suffer from OP, with an additional 43.1 million experiencing low bone mineral density [4]. In China, the prevalence of OP was 14.94% before 2008 and increased to 27.96% from 2012 to 2015, with the rate being higher in females relative to males [5]. Approximately, in the United States, 16% of men and 29.9% of women aged more than 50 years have OP on the basis of the diagnostic criteria of the National Bone Health Alliance [6]. With the aggravation of global population aging, OP has become a serious global health problem. OP has been attributed to decreased quality of life and increased risks of death along with an elevated burden on health systems economically [7]. Thus, the management of patients with OP is extremely urgent.
Up to now, calcium and vitamin D supplementation are the standard choices for OP treatment [8]. Pharmacological therapies, including bisphosphonates, denosumab, and teriparatide, are also recommended to reduce the risk of vertebral or hip fractures in patients with OP [9]. Estrogen therapy, menopausal estrogen plus progestogen therapy, or raloxifene is suitable for postmenopausal women [10]. However, despite the availability of numerous anti-OP medications with diverse pharmacological properties, as well as fixed-dose combination therapy, the targeted therapeutic effect is not attained in significant numbers of individuals with OP, and the mitigation of OP fracture has remained suboptimal [11,12]. Therefore, finding a drug that is effective and safe for osteoporosis is an important challenge for the industry.
Recently, the growing utilization of complementary and alternative medicine consists of herbal medicine in research and in clinical practice medicinal: they usually have few side effects and are easily accessible [13,14]. Herbal medicines are typically used as a complementary and adjunct therapy for a wide range of diseases such as OP [15]. Fructus Psoraleae (FP) is the dried matured fruits of Psoralea corylifolia Linn, which has a long history and a wide range of applications in Asia, particularly in China, Japan, and Korea for their effects on OP and bone fracture [16]. Studies have shown that FP has the effects of dilating coronary artery, increasing coronary blood flow, antitumor, protecting liver, estrogen-like, and anti-OP effects [17]. The ingredients of FP (IFP) are shown in Figure 1, which mainly contains furanocoumarins (mainly psoralen and isopsoralen), coumestrol (such as psoralidin), flavonoids (mainly corylifolin, corylifolinin, corylin, and bavachalcone), and phenolic terpenoids (such as bakuchiol). Many IFP have estrogen-like effects [18]. Recent reports have shown that phytoestrogens have anti-OP effects similar to natural estrogen with less estrogen-like side effects [19]. Several studies have demonstrated that multiple IFP may possess anti-OP effects both in vivo and vitro [20,21]. Corylifolin could prevent estrogen deficiency-induced bone loss in ovariectomized rats and induced primary human osteoblast differentiation [20]. Psoralen and bakuchiol improved osteoclast differentiation and bone resorption via inhibiting the protein kinase B (AKT) and activator protein-1 (AP-1) pathway [21]. However, the scattered evidence and uncertain mechanisms limited the application of FP and IFP in the clinic. Systematic review and meta-analysis of animal studies are considered to be a valuable tool to provide important insights into the validity of animal studies, improve the precision of estimated effects, and support further generalization to human clinical trials [22]. We speculated that IFP could exert anti-OP effects in animal models of OP. However, it is difficult to translate these beneficial effects of IFP from basic research to clinical application. Besides, the uncertain mechanisms and adverse drug reaction of IFP should also be taken into consideration, which were also the uncertainties and conflicts that underlie the hypotheticals. The evidence of studies was beneficial to find out a potential medication to prevent OP for healthy individuals and to treat OP for OP patients with less adverse drug reaction. Now clinically, FP has not been used to treat OP and there is a lack of clinical evidence, which is the focused clinical question we review and address in our study. Our study might provide a theoretical basis for the application of FP in OP. Therefore, we presented a systematic review and meta-analysis from the preclinical evidence of IFP in animal models of OP to summarize the significant outcomes on efficacy and mechanisms.

Methods
We used the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement (Table 1) to perform a systematic review and meta-analysis [23]. There are no protocols preregistered for this review.

Database and Search
Strategies. Electronic searches were performed in eight databases from their respective inception to June 2021: PubMed, EMBASE, Web of Science, Cochrane Library, Chinese National Knowledge Infrastructure, Chinese Biomedical Literature Database, Chinese VIP Database, and Wanfang Database. No language restrictions were applied. The following search terms were used in PubMed and were modified to suit other databases: "Fructus Psoraleae extract" AND "Osteoporosis." In addition, reference lists from the resulting publications and reviews were also searched carefully for the eligible studies. ies. Firstly, the titles and abstracts were screened to exclude the duplicated and apparently irrelevant ones or those that do not meet our inclusion criteria. After that, the remaining potential studies were full-text downloaded and reviewed. Any disagreement between the two above authors was sent and discussed with the third independent author.

Eligibility
2.5. Data Extraction. Two reviewers independently extracted data, and the third reviewer checked the consistency between them. A standard form was used; the extracted items included the general study information: the author's name(s), publishing date, animal species, age, gender, weight, sample size, OP modeling method, the use of anesthetics in the course of the experiment, the therapeutic regimen of the treatment and control groups, and primary and secondary outcomes and its intergroup differences. For continuous outcomes, we extracted the mean, standard deviation (SD), and participant number. The data in other forms was recalculated when possible to enable pooled analysis. If the study was involved in multiple intervention groups, we extracted data only for the group(s) involving IFP and the control group(s). Disagreements between two researchers were resolved by discussion. Whenever necessary, we contacted the authors of the studies for the missing data and additional information.
2.6. Quality Assessment of Included Studies. Two authors independently performed methodological quality and risk of bias assessment of the included studies using the CAMARADES 10-item quality checklist with minor modification [24]. The modification is listed as follows: D: blinded induction of model (group randomly after modeling), F: use of anesthetic without significant protective and toxic effects  Specify the methods used to collect data from reports, including how many reviewers collected data from each report, whether they worked independently, any processes for obtaining or confirming data from study investigators, and, if applicable, details of automation tools used in the process Page 9 Data items 10a List and define all outcomes for which data were sought. Specify whether all results that were compatible with each outcome domain in each study were sought (e.g., for all measures, time points, analyses), and if not, the methods used to decide which results to collect Page 9 10b List and define all other variables for which data were sought (e.g., participant and intervention characteristics, and funding sources). Describe any assumptions made about any missing or unclear information Page 9 Study risk of bias assessment 11 Specify the methods used to assess risk of bias in the included studies, including details of the tool(s) used, how many reviewers assessed each study and whether they worked independently, and, if applicable, details of automation tools used in the process Page 9 Effect measures 12 Specify for each outcome the effect measure(s) (e.g., risk ratio, mean difference) used in the synthesis or presentation of results Page 10 Synthesis methods

13a
Describe the processes used to decide which studies were eligible for each synthesis (e.g., tabulating the study intervention characteristics and comparing against the planned groups for each synthesis (item #5))  [25,26]. The evidence quality of included outcomes 5 Oxidative Medicine and Cellular Longevity was graded as high, moderate, low, or very low. RCTs were initially classified as having high-quality evidence. The quality of each outcome was downgraded for the following five factors: risk of bias, inconsistency, indirectness, imprecision, and publication bias. GRADE pro3.6.1 software was used for the data analysis and synthesis.

Statistical
Analysis. The data was collected and input into the STATA software (version 12.0; StataCorp, College Station, TX) for meta-analysis. A random-effects model was applied when heterogeneity was detected or the statistical heterogeneity was high (P < 0:05 or I 2 > 50%), and then, further subgroup study and metaregression analysis (the number of included studies was more than ten) were performed to detect the origin of heterogeneity. Otherwise, a fixed-effects model was used (P ≥ 0:05 or I 2 ≤ 50%). To test the strength and stability of the pooled results, we performed a sensitivity analysis by omitting the individual studies one by one. Moreover, the effect of publication bias was investigated by Egger's test. Standard mean difference (SMD) was calculated for continuous outcomes.

Study Selection.
We identified 130 potentially relevant hints from eight databases. Removing duplication of literature, there were 62 articles left. We excluded 10 studies that are not related to this study after reading the titles and abstracts in detail. Through reading the full text of 52 articles, 36 articles were excluded for at least one of the exclusion criteria. Finally, 16 articles [17,[27][28][29][30][31][32][33][34][35][36][37][38][39][40][41] were included for analysis. Figure 2 presents the detailed flow diagram of the search and selection process.

General Characteristics of the Included Studies.
The characteristics of all the 16 included studies are summarized and shown in Table 2. All of them were published between 2008 and 2021. All studies involved a total of 379 subjects with 191 in the experimental group vs. 188 in the control group. As for animal species, Sprague-Dawley rats were used in ten studies [27-29, 31, 32, 34, 37-39, 41], Wistar rats in three [17,30,40], C57BL/6 mice in two [35,36], and ICR mice in one [33]. The weight of rats varied between 160 g and 290 g, while the weight of mice varied between 16.6 g  Oxidative Medicine and Cellular Longevity       [17,[37][38][39], and ketamine in one study [40]. Detailed information of IFP in each study is displayed in Table 3. With regard to the specific IFP, isopsoralen was reported in five studies [27,28,32,34,36], psoralen in seven studies [17, 29-31, 35, 37, 40], bakuchiol in two studies [33,41], psoralidin in one study [39], and bavachin in one study [38]. Fifteen studies [17,[27][28][29][30][31][32][33][34][35][36][37][38][39][40]  3.3. Study Quality. The CAMARADES 10-item quality checklist was adopted to judge the risk of bias of each study, and the number of criteria met varied from 4/10 to 7/10 with the average of 5.25. Detailed results of methodological quality are presented in Table 4. All the included studies were peer-reviewed publications; three studies [31,37,39] did not mention control of temperature. Two [32,34] of the included studies did not declare randomisation. The ways of blinding induction of model were reported in seven studies [29-31, 33, 35, 36, 41], and all of them reported that the animals were grouped randomly after modeling. No study mentioned the calculation of the sample size, and none used a blinding method during outcome assessment and an appropriate animal model. Ten studies [17, 27, 29, 32-34, 11 Oxidative Medicine and Cellular Longevity 37-40] used the anesthetic without protective and toxic effects on bones. Compliance with animal welfare regulations was not described in four studies [33,35,37,40], and the potential conflict of interests was not mentioned in four studies [30,32,35,40].

18
Oxidative Medicine and Cellular Longevity 3.6. Publication Bias and Sensitivity Analysis. Egger's test ( Figure 11) was used to assess the potential publication bias of the BMD-femur in this meta-analysis. The P values from Egger's tests indicated that there was no significant publication bias for BMD-femur (P = 0:416).
To determine the influence of each study on the pooled data for BMD-femur, BMD-lumbar spine, and BV/TV to verify the robustness of our results, sensitivity analysis was performed by omitting one study at a time and calculating the pooled data for the remaining studies. The results of the sensitivity analysis indicated that no significant effect was observed after excluding any single study, suggesting that the results of this meta-analysis were relatively robust ( Figure 12). 3.7. GRADE Assessment. The GRADE system was used to assess the level of evidence for the outcomes, which indicated moderate, low, or very low quality with methodological problems and heterogeneity problem. The GRADE evidence profiles are shown in Table 5.

Discussion
4.1. Summary of Evidence. This is the first preclinical systematic review and meta-analysis to estimate the efficacy and possible mechanism of IFP for the OP animal model. Sixteen high-quality studies involving 379 animals with the OP model were enrolled in the analysis. The primary findings of present systematic review illustrated that IFP could significantly increase the S-OCN, BMD, BV/TV, and Tb.N while Tb.Sp and Tb.Th were remarkably decreased by IFP in OP model animals. Moreover, IFP could significantly improve the bone biomechanical indicator bone maximum load and elasticity modulus. Therefore, the findings revealed that FP is a potential anti-OP drug through multiple mechanisms. However, the outcomes BMD-femur, BMD-lumbar spine, and BV/TV represented high heterogeneity in our metaanalysis. According to the results of subgroup analysis and metaregression in our study, the source of heterogeneity was from the sample size, OP modeling methods, kind of IFP, and dosages of IFP. Thus, more high-quality studies involving large sample sizes should be conducted to confirm our findings. Strengths. The strengths of this meta-analysis study consisted of a clearly defined research question, which reduced the bias in the choice of the included studies, fidelity, and consistency to a precise research approach that we designed before the meta-analysis, an in-depth search of the literature, the agreement between the two researchers regarding the entry data components, and the quality control appraisal of all the data. The quality of included original studies was relatively moderate, which suggests that the results were relatively reliable. Additionally, the number of trials and the overall sample size was comparatively large (16 trails with 379 animals). We performed subgroup assessments and metaregression evaluation to identify the origin of heterogeneity. Consequently, no publication bias was reported in this meta-analysis, and sensitivity estimation revealed that the findings of this meta-analysis are comparatively robust.

4.3.
Limitations. Some limitations that may affect the accuracy of the study should be considered. Firstly, the included primary studies had some intrinsic and methodological shortcomings: (1) Only 14 trials had sufficient information on the generation of random allocation. (2) The blinding procedure and sample size calculation were not reported or remained unclear in some studies, making it a challenge to bias findings unintentionally or intentionally and to help allow the credibility of study conclusions. Secondly, selection bias was unavoidable because only eight frequently used databases were searched for English and Chinese language   Heterogeneity (I 2 > 50%, P < 0:05) was found. 3 No details of random protocol were reported. 4 Some of the included studies lack allocation concealment.

22
Oxidative Medicine and Cellular Longevity studies. Therefore, the potentially relevant studies published in other languages could have been left out. Thirdly, the absence of negative studies might have led to the true effect of IFP being overestimated. Fourthly, though the metaregression and subgroup analysis were done, the high heterogeneity of BMD-femur, BMD-lumbar spine, and BV/TV could not be neglected. The OP modeling methods, the specific kind of IFP, dosage of IFP, administration approaches, and period of IFP treatments differed remarkably in the included studies. This heterogeneousness could compromise the viability of our findings. Fifthly, most of the included studies in the meta-analysis were conducted in China, a potential limitation to the generalizability of our findings. Sixthly, the overall quality of evidence of this study was low (Table 5). Finally, many of the included studies suffer from significant sources of bias; this also will jeopardize the validity of results.

Implications.
High-quality methodologies of studies are the cornerstones of translating animal research into clinical drug treatments for human disease [42]. The score (mean 5.25) by prudent assessment of included studies was better than that of most studies of traditional Chinese medicine [43]. There were limitations in terms of blinding and sample size calculation. The blinding methods in the animal model establishment and outcome assessment were usually seen as technical difficulties for most studies. A sample size calculation could avoid the waste of resources caused by oversize and the imprecision of study result by undersize; the specific steps could be referred from the literature [44]. Besides, the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines are aimed at improving the quality of research reports by guiding complete and transparent reporting of in vivo animal research. Employing of experimental animals with comorbidities such as advanced age, obesity, hyperglycemia, or other risk factors may be more in line with the physiology of OP patients and may be helpful for the clinical translation of experimental results. These should be adopted in the future study management of IFP for OP. Using different animal models at different research stages of disease is crucial to study the pathophysiology and treatments [45]. Factors that need to be considered include pathogenesis of model, availability of the animals, technical requirements, and cost and ethical considerations [46]. According to the pathogenesis, animal models of OP can be divided into two types: models with increased bone resorption as the dominant mechanism (such as ovariectomized OP model, disused OP model, tretinoin induction OP model, nutritional OP model, and glucocorticoid OP model) and models with reduced bone formation as the dominant mechanism (such as senile OP model) [47]. This study comprehensively includes the ovariectomized OP model and nonovariectomized OP model (tretinoin induction OP model and glucocorticoid OP model) to evaluate the efficacy and mechanisms of IFP for OP. The results of subgroup analysis suggested that the ovariectomized OP model group showed better effect size than the nonovariectomized OP model group in regard to BMD-femur (SMD = 5:409 ± 3:193 versus SMD = 1:000 ± 1:408, respectively, P = 0:027, Figure 9(b)), which suggests that the different OP model methods may be the source of high heterogeneity. In clinical practice, postmenopausal OP has a similar pathogenesis with ovariectomized OP model [48]. The ovariectomized OP model has become the most widely used animal model to study OP [49], and it also was the most adopted model in eligible studies. Thus, we suggest an ovariectomized OP model be adopted to assess OP in future studies.
The results of subgroup analysis regarding the specific kind of IFP in BMD-femur demonstrated that the prosalen group gave a higher effect size than the isoprosalen group and neither the isoprosalen nor the prosalen group (SMD = 2:063 ± 3:533 versus SMD = 6:839 ± 2:090 versus SMD = 1:276 ± 0:963, respectively, P = 0:021, Figure 9(c)), indicating that different kinds of IFP might be the origin of high heterogeneity. Psoralen is one of the most major ingredients of FP and was also the most used IFP in the included studies (reported in 7 studies) because it has the highest content in FP and is easy to extract from FP. Besides, psoralen has been reported to possess potential anti-OP effects in animal OP models through multiple signal pathways [17,29,35]. For example, psoralen could significantly increase BMD of femur and vertebra, serum levels of calcium, osteocalcin, N-terminal propeptide of type 1 procollagen (P1NP), bone morphogenetic protein 2 (BMP2), and vascular endothelial growth factor (VEGF) via phosphatidylinositol 3kinase/protein kinase B/mammalian target of rapamycin (PI3K/Akt/mTOR) in postmenopausal rats [29]. Isoprosalen is another furanocoumarin compound of FP and is derived from psoralen. Analogously, several studies have shown that isoprosalen exerted mighty protective effects of OP in animal OP models [27,28,36]. Isopsoralen could increase BMD and improve serological indicators inhibiting Axin2/peroxisome proliferator activated receptor (PPAR-γ) signaling pathway and activating Wnt signaling pathway in rats with glucocorticoid-induced OP [27]. Other IFP including bakuchiol, bavachin, and psoralidin were also reported in our included studies; however, the number of studies was relatively less and their effect size was lower than prosalen and isoprosalen. Therefore, prosalen and isoprosalen may be recommended as potential candidates of anti-OP drugs in the future studies. However, their safety and toxicity should be taken into consideration, which will be the direction of future research.

Possible
Mechanisms. Systemic review of preclinical studies is conducive to understand comprehensively pathological mechanisms of disease and pharmacological effects of drugs [50]. The possible mechanisms of IFP that mediated anti-OP effects in the included studies are summed up as follows: (1) Osteoprotegerin/receptor activator of nuclear factor-κB ligand/receptor activator of nuclear factor-κB (OPG/ RANKL/RANK) signal pathway: IFP could highly increase OPG secretion and reduce RANKL expression, resulting in an enhancement in OPG/RANKL ratio, which leads to a decrease in the number and activity of osteoclasts [30,31,34]. (2) PPAR-γ/Axin2/Wnt signal pathway: IFP suppressed the PPAR-γ/Axin2 signaling pathway involved in lipid metabolism, blocked the inhibition of Axin2 on the Wnt signaling pathway, upregulated the expression of β-catenin protein, and made β-catenin form a complex with nuclear transcription factors after entering the nucleus. Eventually, β-catenin activated the Wnt signaling pathway to regulate the lipid metabolism of rat bone marrow, thereby affecting bone metabolism and promoting bone formation [27]. Besides, the inhibition of PPAR-γ resulted in the increase of Runx2 expression, which facilitates osteogenic differentiation of bone mesenchymal stem cells in rats [36]. (3) Antioxidative stress through the forkhead box O3a (FoxO3a)/Axin2/ Wnt signal pathway: IFP inhibited the activation of the FoxO3a signaling pathway through its antioxidant effect; meanwhile, it upregulated the expression of β-catenin, bound it to FoxO3a, initiated the transcriptional program that regulates bone tissue cell apoptosis and eliminates reactive oxygen species (ROS), which promotes the generation and formation of osteoblasts. Then, the expression of Axin2 was downregulated by removing excessive ROS, the PPARγ/Axin2 signaling pathway was inhibited to participate in lipid metabolism and the inhibition of Axin2 on the Wnt signaling pathway was blocked, which ultimately exerts its anti-OP effect [32]. (4) PI3K/Akt/mTOR signaling pathway: IFP restrained the expression of PI3K/Akt/mTOR in the rat's femur and increased the levels of BMP2 and VEGF, which improves bone formation and angiogenesis [29]. (5) Estrogen-like effect: IFP had the similar effect of phytoestrogen on inhibiting bone resorption by participating in the binding of estrogen receptor. On the one hand, IFP may promote the synthesis and secretion of estrogen outside the ovary [40,41]. On the other hand, it may enhance the secretion of thyroid calcitonin to exert it anti-OP effect [39]. (6) Gamma-aminobutyric acid/gamma-aminobutyric acid receptor (GABA/GABA B RI) signaling pathway: IFP was observed to reduce the production of GABA and GABA B RI to inhibit the activity of osteoblast [38]. The mechanism diagram is summarized in Figure 13.

Conclusion
This preclinical systematic review provided preliminary evidence that IFP was capable of partially exerting anti-OP effects in animal models probably through the OPG/ RANKL/RANK, PPAR-γ/Axin2/Wnt, antioxidative stress via FoxO3a/Axin2/Wnt, PI3K/Akt/mTOR, estrogen-like effect, and GABA/GABA B RI signaling pathway. Taken together, the findings suggest the possibility of developing IFP as a drug for the clinical treatment of OP.