Review Reports - Sweetpotato-Derived Cosmetic Actives: Chemistry, Skin Mechanisms, and Formulation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Sweetpotato (Ipomoea batatas L.) Foliar Caffeoylquinic Acids and Root Pigments as Cosmetic Actives: Chemistry, Skin-Relevant Mechanisms, Formulation Constraints and a Validation Framework

Manuscript ID

cosmetics-4626352

This manuscript addresses a timely and relevant topic: the potential use of sweetpotato (Ipomoea batatas L.) leaves and pigmented roots as sources of cosmetic ingredients. Its main strength is that it goes beyond phytochemical description and discusses formulation stability, skin delivery, safety, regulatory aspects, and claim substantiation.

However, the manuscript requires substantial revision because some conclusions occasionally extend beyond the strength of the available evidence. The authors should more clearly distinguish between purified compounds and whole extracts, in vitro findings and skin/human evidence, mechanistic plausibility and demonstrated cosmetic efficacy, and UV absorption and validated sunscreen performance. The manuscript itself describes a six-tier framework progressing from extract characterization to claim substantiation.

- I recommend shortening the title to « Sweetpotato-Derived Cosmetic Actives : Chemistry, Skin Mechanisms, and Formulation »

- Identify the work clearly as a narrative review in the Abstract.

- The manuscript is described as a narrative review, but the literature-search approach is partly systematic. Please provide the search end date, period covered, main search terms, inclusion/exclusion criteria, and study-selection strategy. The chronological inconsistency between the 2025 manuscript format and the search period extending to August 2026 must also be corrected.

- Several skin-related effects are based on purified chlorogenic acid, neochlorogenic acid, or 3,4,5-tricaffeoylquinic acid. These findings should not be directly extrapolated to whole sweetpotato extracts. For major claims, specify the tested material, concentration, model, cytotoxicity, formulation relevance, and whether the active reaches the biological target. Separate purified-compound evidence from extract evidence.

- In vitro SPF and UV-absorbance measurements should be presented as preliminary evidence only. They are influenced by concentration, solvent, film thickness, method, and product color and are not equivalent to in vivo SPF or UVA protection. The conclusion that sweetpotato extracts are better positioned as antioxidant or SPF-supporting ingredients rather than stand-alone UV filters is appropriate and should be emphasized throughout.

-The discussion of CQA instability is strong, but the manuscript would benefit from a practical stability-testing scheme including accelerated and real-time testing, pH, heat, light, freeze–thaw cycles, oxygen exposure, packaging compatibility, HPLC monitoring of CQA isomers, and retention of biological activity.

-MIC values of 2.5–5 mg/mL are relatively high. Antibacterial activity, bactericidal activity, biofilm prevention, biofilm disruption, preservation efficacy, and microbiome modulation should be clearly distinguished. An in vitro biofilm effect should not automatically support preservative, antimicrobial, anti-acne, or microbiome claims.

-The manuscript correctly identifies limitations of mushroom tyrosinase. Please further explain that differences between fungal and human enzymes, assay interference, catechol oxidation, and substrate effects can produce misleading inhibition data. Human tyrosinase assays and melanocyte or reconstructed-skin confirmation are needed.

-Safety should be discussed separately for crude extracts, purified fractions, and isolated compounds. The authors should also consider pesticide residues, heavy metals, residual solvents, microbial contamination, chlorophyll degradation products, allergens, storage degradation products, and batch-to-batch variability.

-Regulatory status depends on market, finished-product composition, intended use, and wording of claims. UV absorption does not make a botanical extract an approved UV filter, and antimicrobial, acne-treatment, or SPF claims may move the product outside the cosmetic category. Regulatory citations should be updated and based primarily on official sources.

The manuscript has good potential and addresses an important gap between phytochemical research and realistic cosmetic development. However, the authors should ensure that the strength of each cosmetic claim matches the strength of the underlying evidence, particularly for whole extracts, SPF-related effects, antimicrobial activity, skin delivery, safety, and regulatory positioning.

I recommend : Major correction

Author Response

Response to Reviewer 1

I thank the reviewer for the precise review. All concerns have been addressed, and each claim is now matched to the strength of its evidence.

Comment 1.0. Distinguish purified compounds from extracts, in vitro from skin or human evidence, mechanism from efficacy, and UV absorption from sunscreen performance.

Response. Done throughout. Each study is classified in the new Table 2 by material (extract; compound isolated from sweetpotato; same compound from another source), by model (cell-free or optical; skin cells; reconstructed or ex vivo skin; human), and by evidence class E1–E5. Mechanistic statements are labeled as plausibility, not efficacy (Sections 3.1 and 7.1; Conclusions).

Comment 1.1. Shorten the title.

Response. Accepted verbatim.

Comment 1.2. Identify the work as a narrative review in the Abstract.

Response. Accepted. The Abstract now opens with "This narrative review…" and separates extract evidence from purified-compound evidence.

Comment 1.3. Report the search method and correct the chronology.

Response. Section 1.1 now reports the databases (Web of Science, Scopus, PubMed, J-STAGE), the period (January 2010 to 31 August 2026), three Boolean concept blocks, inclusion and exclusion criteria, and two-stage screening. Pre-2010 sources are explained by backward citation tracking. The file uses the current template, and the header now reads 2026.

Comment 1.4. Separate purified compound from extract evidence; give material, concentration, model, cytotoxicity, formulation relevance and target exposure.

Response. Table 2 reports all of these for every study. Two statements were corrected: the 3,4,5-triCQA in [35] was prepared from Aconitum koreanum (class E3), and purified CQAs acted at 3–200 µM in monolayers, whereas no study shows that such concentrations reach viable epidermis (Section 4.3).

Comment 1.5. Present in vitro SPF and UV absorbance as preliminary.

Response. Section 3.2 now states that the leaf SPF of 24.65 is a Mansur absorbance index of ethanolic solutions (290–320 nm, 700 µg mL⁻¹) without UVA information, and that [17] measured UV absorbance of a cream, not SPF. A protection claim requires ISO 24444:2019 and ISO 24443:2021, or ISO 23675:2024, plus photostability data. The positioning "antioxidant and SPF-supporting, not a UV filter" appears in the Abstract, Sections 3.2 and 6, Table 6 and the Conclusions.

Comment 1.6. Add a practical stability-testing scheme.

Response. New Table 3, based on ISO/TR 18811 with ICH Q1A(R2) and Q1B conditions, covers real-time (25 °C/60% RH, to 24 months) and accelerated (40–45 °C) storage, pH bracketing (4.5–6.0), freeze–thaw cycling, ICH Q1B light exposure, oxygen and headspace control, packaging compatibility, ISO 11930 challenge testing, isomer-resolved HPLC-DAD/LC–MS, and retention of activity (≥80%). Section 4.1 explains why Arrhenius extrapolation fails: degradation follows Weibull kinetics, and acyl migration yields isomers that remain active.

Comment 1.7. MICs are high; distinguish antimicrobial, bactericidal, biofilm, preservative and microbiome endpoints.

Response. Section 3.3 was rewritten around five endpoints. Killing was reached only for P. aeruginosa and MSSA (MBC = MIC). Biofilm prevention was shown at ½ MIC, but disruption was not tested. Preservation requires ISO 11930 testing. No microbiome data exist. The MICs are 22–45-fold above the HaCaT non-cytotoxic limit (≈111 µg mL⁻¹), and the extract raised IL-6 about fourfold, so the term "immunomodulatory" was corrected. No preservative, antimicrobial, anti-acne or microbiome claim is made.

Comment 1.8. Explain the limits of mushroom tyrosinase assays.

Response. Section 5.1 now covers the structural differences between the fungal and human enzymes, the behaviour of catechols as alternative substrates, reduction of dopaquinone and o-quinone chemistry, and optical (475 nm) and copper-chelation artefacts. It gives a minimum control set and requires confirmation on human tyrosinase, followed by melanin content in normal human melanocytes and pigmented reconstructed epidermis.

Comment 1.9. Discuss safety by material type and contaminants.

Response. New Table 4 separates crude extracts, enriched fractions and isolated compounds. Section 4.4 sets cross-cutting limits for pesticide residues (LC–MS/MS, GC–MS/MS), heavy metals (ICP-MS; Article 17, Regulation 1223/2009), residual solvents (ICH Q3C, including deep eutectic solvents and ionic liquids), microbiological quality (ISO 17516), chlorophyll-derived photosensitizers (OECD TG 432/498), sensitization (OECD GL 497), storage degradants and batch-to-batch fingerprint similarity.

Comment 1.10. Regulatory status depends on market, composition, use and claims; update citations from official sources.

Response. Section 6 now opens with this principle and cites only official instruments: FD&C Act §201(g) and (i); OTC Monograph M020; MoCRA effective dates; Annex VI of Regulation 1223/2009; Recommendation 2006/647/EC; Directive 2001/83/EC; Regulation (EU) 528/2012; and Annex IV/CosIng and 21 CFR Parts 73/74 for colorants.

Comment 1.11. Match claim strength to evidence strength.

Response. Addressed through Responses 1.4–1.10. The Conclusions now state explicitly which claims the evidence supports and which it does not.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript offers a comprehensive and well‑structured synthesis of sweetpotato foliar caffeoylquinic acids (CQAs) and root pigments as cosmetic actives. The authors integrate phytochemical characterization, mechanistic pathways relevant to skin biology, formulation constraints, and regulatory considerations in a way that is both timely and highly relevant for cosmetic science. The review addresses a genuine gap in the field: botanical cosmetic ingredients are frequently marketed with limited compositional definition and insufficient mechanistic support, whereas sweetpotato provides well‑characterized phytochemistry and a credible upcycling rationale. This integration of chemistry, formulation science, and regulatory alignment is one of the manuscript’s strongest contributions. The work stands out through its detailed discussion of formulation challenges such as pH instability, acyl migration, catechol oxidation, and permeation limitations—topics rarely addressed in phytochemical literature. The use of human‑target enzyme and cell studies for individual CQAs adds mechanistic rigor beyond typical antioxidant‑focused botanical reviews. The tiered validation framework aligned with US and EU cosmetic regulations is particularly valuable, offering a structured approach to standardization, mechanistic validation, permeation testing, safety assessment, and claim substantiation. The sustainability argument, centered on upcycling leaf biomass discarded after root harvest, is compelling and well contextualized.

Despite these strengths, several limitations reduce the manuscript’s ability to function as a true “validation framework” for commercial development.

The narrative (non‑systematic) design introduces selection bias, limits reproducibility, and precludes transparent evidence grading. The review does not include a formal quality appraisal of the cited studies, and much of the mechanistic evidence is extrapolated from purified compounds or unrelated botanical matrices rather than sweetpotato‑specific data.

The economic feasibility, scalability, and regulatory burden associated with the proposed validation pathway are not discussed, although these aspects are essential for industry translation.

A central issue is the permeation data gap. Given the high polarity and molecular weight (>500 Da) of di‑ and triCQAs, it is unlikely that these compounds reach viable epidermal or dermal targets at bioactive concentrations without specialized delivery systems. In the absence of quantitative permeation data, the mechanistic pathways described in the review cannot be assumed to translate into topical efficacy. This gap should be explicitly acknowledged, and the authors may consider discussing potential delivery strategies or the need for future permeation studies.

The manuscript is rich in information, but the novel contribution relative to existing reviews on caffeoylquinic acids, anthocyanins, or botanical antioxidants is not fully explicit. It would strengthen the work to clarify what is genuinely new in this review, how sweetpotato‑derived actives differ mechanistically from other botanical sources, and which insights are specific to sweetpotato rather than generalizable to CQAs or anthocyanins. A clearer distinction between direct sweetpotato evidence, purified compound studies, extrapolated data, and surrogate enzyme assays (such as mushroom tyrosinase) would also improve transparency and support claim substantiation.

The proposed validation framework is promising, but it would benefit from a more clearly articulated workflow, with explicit identification of mandatory versus optional steps and a brief discussion of feasibility and regulatory implications. The sustainability section could be strengthened by including approximate quantities of leaf biomass discarded globally or preliminary life‑cycle assessment considerations, if available.

Author Response

Response to Reviewer 2

I thank the reviewer for recognizing the integrative value of the work. All seven concerns have been addressed.

Comment 2.1. Narrative design, selection bias and no quality appraisal.

Response. The search is now reproducible (Section 1.1). Table 2 grades each study and appraises it on four criteria: material identity, concentration and units, cytotoxicity in the same system, and formulation relevance to target exposure. New Section 7.1 sets out the limitations. Meta-analysis is not feasible, because E1 studies are few and their endpoints (DPPH IC₅₀, Mansur SPF, MIC, MMP-1 expression, tyrosinase IC₅₀) are heterogeneous. GRADE and RoB 2 do not apply because no randomized trials exist.

Comment 2.2. Economic feasibility, scalability and regulatory burden.

Response. Addressed in new Section 5.3 and the relative-cost column of Table 5. Post-harvest drying is the main cost driver; we compare freeze-drying with low-temperature air-drying [74]. Scalable options include aqueous-ethanol extraction within ICH Q3C limits, resin enrichment of triCQA [18], and natural deep eutectic solvent microwave-assisted [75] or ionic-liquid [76] extraction, subject to proof of solvent-residue safety. The regulatory burden is set out by market: in the EU, the cosmetic product safety report with the SCCS margin of safety, which uses a 50% default absorption when OECD TG 428 data are missing; in the US, MoCRA registration, listing, and safety substantiation. SPF, anti-acne or antimicrobial claims trigger the drug or biocide route.

Comment 2.3. Permeation data gap for di- and triCQAs.

Response. The gap is now acknowledged explicitly (Section 4.3) and quantified (Table 1). Potts–Guy log kₚ values are −2.97 for caffeic acid, −5.34 for 5-CQA, −5.14 for 3,5-diCQA and −4.94 cm h⁻¹ for triCQA, with TPSA of 165–258 Å² and 6–8 hydrogen-bond donors. With pKa ≈ 3.4, only 2.5% (pH 5.0) to 0.8% (pH 5.5) is un-ionised. An upper-bound estimate for 0.5% 3,5-diCQA over 24 h is ≈1–4 µM in viable epidermis, against 3–200 µM active in vitro. The mechanisms are therefore not assumed to translate into topical efficacy. Surface-acting claims are the default; target-acting claims require delivery systems (phospholipid complexes, ethosomes, micro- and nanoemulsions, cyclodextrins, lipophilic esters) verified by OECD TG 428 under finite dose, with mass balance and isomer-resolved LC–MS/MS.

Comment 2.4. Clarify novelty and sweetpotato-specific insights.

Response. New Section 1.2 identifies four sweetpotato-specific features: a high share of di- and triCQAs, with preparative recovery of triCQA; diacylated YGM anthocyanins stabilized by intramolecular copigmentation; CQAs, anthocyanins, and carotenoids in one crop whose leaves are a by-product; and strong genotype × environment control of leaf CQAs. It also states the central insight: poly-acylation raises potency but lowers permeation and stability.

Comment 2.5. Distinguish direct, purified, extrapolated, and surrogate-assay evidence.

Response. The Table 2 classes map onto these categories: direct sweetpotato evidence (E1), purified compounds (E2 and E3), extrapolation (E5), and surrogate assays (E4). Mushroom tyrosinase data are classed as E4, screening only (Section 5.1).

Comment 2.6. Clarify the workflow, with mandatory versus optional steps, feasibility and regulatory implications.

Response. New Figure 5 and the Status column of Table 5 show that Tiers 1 and 5 are mandatory, Tier 4 is strongly recommended, and Tiers 2, 3, and 6 depend on the claim. Four go/no-go gates are defined: G1, specification and marker variation; G2, activity below the cytotoxicity threshold; G3, target exposure at or above the effective concentration, failing which claims are limited to surface action; and G4, margin of safety ≥100, no sensitization or phototoxicity, stability and preservation. Feasibility and regulatory implications are in Section 5.3.

Comment 2.7. Quantify discarded leaf biomass and consider LCA.

Response. The Introduction now states that vine biomass is on the same order of magnitude as root yield [5]. Global root production of ≈85–90 Mt per year [58] implies ≈9–13 Mt of dry foliage per year at 10–15% dry matter, and with CQAs at several per cent of dry weight, supply is not limiting. Section 7.2 outlines a cradle-to-gate LCA under ISO 14040/14044, with economic allocation, drying energy, avoided burdens, and soil-nutrient export.