Developing a Rational, Optimized Product of Centella asiatica for Examination in Clinical Trials: Real World Challenges

Abstract

Botanical products are frequently sold as dietary supplements and their use by the public is increasing in popularity. However, scientific evaluation of their medicinal benefits presents unique challenges due to their chemical complexity, inherent variability, and the involvement of multiple active components and biological targets. Translation away from preclinical models, and developing an optimized, reproducible botanical product for use in clinical trials, presents particular challenges for phytotherapeutic agents compared to single chemical entities. Common deficiencies noted in clinical trials of botanical products include limited characterization of the product tested, inadequate placebo control, and lack of rationale for the type of product tested, dose used, outcome measures or even the study population. Our group has focused on the botanical Centella asiatica due to its reputation for enhancing cognition in Eastern traditional medicine systems. Our preclinical studies on a Centella asiatica water extract (CAW) and its bioactive components strongly support its potential as a phytotherapeutic agent for cognitive decline in aging and Alzheimer's disease through influences on antioxidant response, mitochondrial activity, and synaptic density. Here we describe our robust, scientific approach toward developing a rational phytotherapeutic product based on Centella asiatica for human investigation, addressing multiple factors to optimize its valid clinical evaluation. Specific aspects covered include approaches to identifying an optimal dose range for clinical assessment, design and composition of a dosage form and matching placebo, sourcing appropriate botanical raw material for product manufacture (including the evaluation of active compounds and contaminants), and up-scaling of laboratory extraction methods to available current Good Manufacturing Practice (cGMP) certified industrial facilities. We also address the process of obtaining regulatory approvals to proceed with clinical trials. Our study highlights the complexity of translational research on botanicals and the importance of identifying active compounds and developing sound analytical and bioanalytical methods for their determination in botanical materials and biological samples. Recent Phase I pharmacokinetic studies of our Centella asiatica product in humans (NCT03929250, NCT03937908) have highlighted additional challenges associated with designing botanical bioavailability studies, including specific dietary considerations that need to be considered.

Keywords: Centella asiatica; botanical; clinical trials; dietary supplement; placebo; reproducible; translation.

Conflict of interest statement

AS is an ad hoc consultant for Oregon's Wild Harvest. Services provided by Ashland Laboratories and Oregon's Wild Harvest were financed by the NIH grant R61AT009628 awarded to Oregon Health & Science University. Co-authors affiliated with Oregon's Wild Harvest (JM and JR) contributed to product design, manufacture and quality control of CAP. The CAP product and placebo described in this study were made for research purposes only and not for commercial use. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2022 Wright, McFerrin, Alcázar Magaña, Roberts, Caruso, Kretzschmar, Stevens, Maier, Quinn and Soumyanath.

Figures

Figure 1

Chemical structures of triterpenes and caffeoylquinic acids found in Centella asiatica. Structures were obtained from Chemspider.

Figure 2

Coloring excipient test and final CAP matching evaluation. (A) Three different coloring agents were dissolved in 8 oz. of water and compared against one another for color matching with CAW. (B) 0, 2, and 4 g CAP, containing the selected coloring agent and additional excipients for flavor matching, dissolved in 8 oz. of warm water.

Figure 3

Concentration of triterpenes and caffeoylquinic acids found in 0, 2, and 4 g Centella asiatica product as determined by LC-HRMS. Sachets of the Centella asiatica water extract products (CAP; n = 5 per dose) were extracted with methanol and analyzed for the content of active compounds (triterpenes and caffeoylquinic acids) using LC-HRMS in positive and negative ion mode against commercial reference standards. The content of triterpenes and caffeoylquinic acids per gram of Centella asiatica extract was identical for the 2 g and 4 g doses of CAP and showed low variability indicating successful and uniform manufacture of the two doses. None of the specific analytes were detected in the 0 g dose, confirming their absence at detectable levels in the placebo which was comprised solely of the excipients used.

Figure 4

A 32-day stability test of CAP 2 g and 4 g showing unchanged levels of triterpene and caffeoylquinic acid components at all temperatures as determined by LC-HRMS. Storage conditions: 25 ± 2°C/60% ± 5% relative humidity, accelerated 40 ± 2°C/75% ± 5% relative humidity, fridge (4°C), ambient temperature (RT), and freezer (−20°C) (n = 3–6 per condition).

Figure 5

Principal Component Analysis (PCA) of 32-day stability test of CAP 2 g and CAP 4 g. Samples were stored in special chambers held at 25 ± 2°C/60% ± 5% relative humidity or accelerated 40 ± 2°C/75% ± 5% relative humidity, at ambient temperature (RT), in a refrigerator (4°C), or in a freezer (−20°C) (n = 3–6 per condition). Chemical fingerprinting analysis of CAP storage stability samples by untargeted data dependent acquisition was performed using LC-HRMS as described earlier (50). The content of each sachet was suspended in 70% v/v methanol (100 mL) containing formic acid (0.1% v/v). Samples were sonicated for 2 hrs with strong shaking every 30 min at room temperature. The suspension (1 mL) was centrifuged (15,000 rpm, 10 min) and diluted 100 times before injection. QC samples were obtained by pooling equal aliquots of each sample. Principal component analysis was performed in Progenesis QI software (V 2.4). All m/z-signals that triggered MS2 experiments (5,849) were used after log-transformation and Pareto scaling. The first principal component represents the maximum variation through the data. Next, another axis representing the next highest variation within the data is added orthogonal to the first one, and is designed as the second principal component. Each marker represents a sample; identically colored markers are replicate samples (n = 6, −20C; n = 5, QC; and n = 3 for all other groups). The PCA plot shows that samples that have a high degree of similarity in their chemical fingerprints cluster closely together. Clustering of the QC samples in the center of the plot confirms LC-MS platform stability. The shift from the left bottom quadrant to the upper left quadrant (4 g sachets) and from right bottom quadrant to right upper quadrant (2 g sachets) indicates that the chemical fingerprints are sensitive to storage temperature. Targeted analyses of the CAP stability samples conducted in parallel (Figure 4) indicated CQA and TT levels were unaffected by storage temperature; thus the observed shifts hint that either excipients or other components of CAW were sensitive to storage temperature.

Figure 6

Percentage transitions toward light for sniffer flies treated with Centella asiatica water extract, Centella asiatica water extract product, excipients used in the manufacture of CAP or control food. Drosophila melanogaster fruit flies with a mutation in the sniffer gene sni1 allele were fed standard food (control) or standard food supplemented with either Centella asiatica water extract (CAW; 10 mg/mL), Centella asiatica water extract product (CAP; equivalent to CAW 10 mg/mL), or the matching placebo for CAP containing only excipients (Ex; equivalent to the amount in CAP) for 7 days. Fast phototaxis was performed with flies (data from both sexes combined) and compared to either control or Ex treatment. The number of tested flies is given above the bars and the SEM is indicated. *p < 0.05, ***p < 0.001.