Recovering Cucurbita pepo cv. ‘Lungo Fiorentino’ Wastes: UHPLC-HRMS/MS Metabolic Profile, the Basis for Establishing Their Nutra- and Cosmeceutical Valorisation

Food-waste is produced throughout all the food supply chain, with a large part already achieved at farm level. In fact, fruits and vegetables, which do not satisfy aesthetic demands, cannot be marketed, but their recovery could favour their valorisation for the obtainment of highly qualified goods. In this context, faulty zucchini fruits (cultivar ‘Lungo Fiorentino’), intended for disposal, were rescued as effective, inexpensive and bio-sustainable source for cosmeceutical purposes. Zucchini fruits underwent extraction and fractionation to obtain ZLF-O and ZLF-A extracts, which were chemically characterized by UHPLC-HRMS. ZLF-A extract, rich in flavonols and flavones, scavenged massively DPPH• and ABTS•+, and was not cytotoxic at doses up to 200 μg/mL. Thus, ZLF-A was incorporated into a base cream formula. Zucchini-based emulsion was deeply screened for its antiradical properties and cytotoxicity towards human keratinocytes and fibroblasts. ZLF-A-enriched cream, whose chemical stability was assessed over time and mimicking different storage conditions, was further tested on reconstructed epidermis disks (EpiskinTM). The recovery of valuable chemical substances from zucchini agro-food waste, complying with the principles of valorisation and sustainable development, can represent a new market force for local farmers. Data acquired were eager to convey a suitable reuse of nutraceuticals rich zucchini waste.


Introduction
A large amount of waste is produced along the entire agri-food supply chain, causing considerable concern for environmental sustainability, waste of resources and human health. Indeed, it is broadly demonstrated that agri-food by-products and wastes are rich in valuable bioactive compounds, whose valorisation could reach targeted applications in diverse biotechnological fields: from the pharmaceutical sector to the food industry, up to the rapidly evolving cosmetics sector [1]. This latter carefully requires innovative and functional natural compounds and extracts, as a clear response to consumers' concerns about synthetic substances [2]. In this context, polyphenols, thanks to their anti-oxidant and anti-inflammatory activities [3], as well as their anti-photocarcinogenic effects [4], are attractive ingredients for cosmetics [5], and plant oils are considered beneficial in wound healing In Table 1 UHPLC-HRMS and MS/MS data for all compounds were summarized, together with theoretical m/z values, calculated molecular formulas, errors (ppm) and the tentative assignment.
Compound 1 was tentatively identified as the O-hexoside of the flavonol myricetin. In fact, the deprotonated pseudomolecular ion at m/z 479.0826 was in accordance with the molecular formula C 21 H 20 O 13 (−1.0 ppm error). The neutral loss of 162/163 Da, typical of hexoses O-linked to an aglycone, led to the product ions at m/z 317.0313 and 316.0218, whose intensity ratio suggested 3-O-glycosylation [18].
The [M−H] − ion detected for compound 2 at m/z 755.2025 (−2.0 ppm error) suggested the molecular formula C 33 H 40 O 20 and a 'Rings and Double Bonds' (RDB) value equal to 14. Following collision induced dissociation, it gave rise to the main product ions at m/z 301.0355/300.0277, putatively ascribed to the flavonol quercetin. Thus, a neutral loss of 454/455 Da occurred, which was supposed to be the glyconic moiety of the molecule, likely a trisaccharide formed by one hexose and two deoxyhexoses. The presence in the TOF-MS 2 spectrum of very low fragment ions let us to hypothesize the connection among sugars (see supplementary materials, Figure S1). In fact, the first one at m/z 609.1484 ([M−H−146] − , loss of one deoxyhexose moiety), once formed, generated the ion at m/z 489.1059 through the loss of 120 Da, which resembles a cross-ring cleavage of hexose sugars. This evidence could indicate a connection between the remaining sugars which involves hexose C2 position and likely the anomeric deoxyhexose moiety [18].
In addition, compounds 3, 5 and 6 were also putatively characterized as quercetin glycosides. The deprotonated aglycone ions were formed in these cases following neutral losses likely attributable to a hexosylpentose (294 Da, e.g., sambubiose), a hexosyldeoxyhexose (308 Da) and a hexose (162 Da), respectively. The identity of metabolite 5, detected as the main constituent of ZLF-A fraction (Table S1), was further confirmed by comparison with data recorded for the pure commercial rutin.
Compounds 4, 10, 11, 13 and 14 shared the same deprotonated aglycone ions, likely attributable to the flavonol kaempferol, based on the occurrence of the peak at m/z 285.039 and other characteristic product ions at m/z 284.032 (deprotonated aglycone ions, deriving from the homolytic cleavage of the saccharidic unit), 255.03 and 227.03. Metabolite 4, with a molecular formula C 33 H 40 O 19 and RDB value equal to 14, could be identified as clitorin, a glycoside bearing a trisaccharide formed by one glucose and two rhamnose residues linked to C-2" and C-6" positions. This hypothesis found confirmation in the TOF-MS 2 spectrum, where the first product ion at m/z 575.1429 was generated by the neutral loss of 164 Da, likely a deoxyhexose residue, which in turn fragmented to give the deprotonated kaempferol ([M−H−164−290 Da] − ). To the best of our knowledge, this glycoside has never been described in C. pepo. Furthermore, the putative presence of the isomer robinin, already reported as constituent of zucchini by Iswaldi et al. [19], could be excluded. In fact, in this latter molecule one rhamnose unit is bound directly to kaempferol at C-7 position (aromatic ring A). Thus, as previously reported [20], TOF-MS 2 mass spectra of compounds 10 and 13 at m/z 593.1514 (or 593.1520) were in accordance with the presence of two O-hexosyldeoxyhexosyl derivatives of kaempferol, differing in the sugar position and/or in the interglycosidic linkage [21]. They have been already tentatively identified in zucchini [19]. Instead, [M−H] − ion for compound 11 at m/z 579.1358 was in accordance with a molecular formula C 26 H 28 O 15 (0.4 ppm error; RDB value of 13). Its fragmentation pattern highlighted the presence of a hexosylpentoside moiety bound to kaempferol at C3 position. Indeed, recently Abu-Reidah et al. [22] described kaempferol 3-sambubioside as a polyphenolic constituent of cucumber whole fruit extract. Finally, deprotonated compound 14 at m/z 447.0924 (−2.4 ppm error) with a molecular formula equal to C 21 H 20 O 11 and RDB value of 12 was tentatively identified as kaempferol 3-O-hexoside, based on its fragmentation pattern previously reported in literature [23].
TOF-MS 2 spectrum of deprotonated compound 7 at m/z 769.2183 showed the same neutral losses of 164 and 290 Da observed for metabolite 4, allowing us to hypothesize the occurrence of the same glycosidic moiety, linked in this case to a different aglycone (m/z 315.0514/314.0439), putatively identified as isorhamnetin (methylquercetin). The presence of the ion at m/z 299.0204, likely formed through the loss of a methyl radical from the aglycone ion, seemed to confirm our hypothesis. Metabolites 12 and 15 were also putatively identified as isorhamnetin glycosides. They were isomers, showing the same molecular formula equal to C 28 H 32 O 16 . In particular, from deprotonated pseudomolecular ions at m/z 623.1619 (or 623.1620) neutral losses of 308.1147 (or 308.1148) Da occurred to generate the aglycon ions, thus suggesting the presence of a hexosyldeoxyhexose (e.g., rutinose) moiety linked. A different glycosylation position was supposed, based on the different intensity ratio of Y 0 − and [24]. Compounds 8 and 9 could be O-glycosides of the flavone luteolin. In fact, in the collision induced dissociation process from deprotonated precursor ions at m/z 447.0925 and 593.1506 the same aglycone ion was generated at m/z 285.0395 (or 285.0404). The TOF-MS 2 spectra differed from those recorded for kaempferol derivatives for the absence of fragments typical of flavonols, such as m/z 271.0246, 255.0302, 227.0342. Based on neutral losses of 162 and 308 (162 + 146) Da, which led to the aglycone ions, these compounds were putatively identified as luteolin O-hexoside and luteolin O-hexosyl-deoxyhexoside, respectively. The occurrence of luteolin and its O-glycosides has been already reported in literature in zucchini, as well as in other plants belonging to Cucurbitaceae family [19,25].
The To the best of our knowledge glycosylated or acetylated anthraquinone derivatives have never been identified in zucchini, but they are quite common in other plants, such as Cassia species [26][27][28]. Metabolite 18 showed a fragmentation pattern, which resembles the one observed for ent-kaurene diterpene glycosides (e.g., steviol glycosides) previously detected in Stevia rebaudiana leaves [29]. In fact, the neutral loss of two hexose moieties from the deprotonated pseudomolecular ion at m/z 803.3713 (C 38 H 60 O 18 , 0.8 ppm error) gave rise to the ion at m/z 479.2673, which in turn underwent a further deglycosylation to produce the ion at m/z 317.2129. This latter was putatively identified as hydroxykaurenoic acid (e.g., steviol), in accordance with a molecular formula C 20 H 30 O 3 with RDB value equal to 6. To confirm this hypothesis, TOF-MS and MS 2 spectra of compound 18 and of rebaudioside A commercial standard were compared ( Figure S3). It is worth of note that kaurenoic acid is involved in the biosynthesis of several plant growth regulators, thus its derivatives can be mostly found in different plant species belonging to several families [30]. The occurrence in C. pepo of hydroxykaurenoic acid derivatives has been previously reported in literature [31].

Metabolic Profiling of ZLF-O Fraction
The UHPLC-HRMS analysis of ZLF-O extract (  Figure 1.    Table 2).  At higher retention times (>3.63 min) metabolites detected in the UHPLC-MS analyses could be C18 fatty acids differentiated by the unsaturation degree, except for compounds 23 and 26, which are constituted by 16 carbon atoms and could correspond to the hydroxypalmitic and palmitic acid, respectively [32]. The elution order reflects the polarity conferred by the presence/absence and by the number of OH groups on the carbon chain. Therefore, it is not surprising that the first eluted are trihydroxyoctadecadienoic and trihydroxyoctadecenoic acids (C 18 Figure 2).

Antioxidant Capacity
Extracts obtained from ZLF fractionation underwent Folin-Ciocalteu assay to estimate their relative total phenol content. This latter was expressed as gallic acid equivalents (GAE) per g of extract: ZLF-A was the fraction most enriched in (poly)phenols (118.6 ± 1.6 mg GAE g −1 ), followed by

Antioxidant Capacity
Extracts obtained from ZLF fractionation underwent Folin-Ciocalteu assay to estimate their relative total phenol content. This latter was expressed as gallic acid equivalents (GAE) per g of extract: ZLF-A was the fraction most enriched in (poly)phenols (118.6 ± 1.6 mg GAE g −1 ), followed by ZLF-O (109.2 ± 1.9 mg GAE g −1 ). ZLF-W1, deriving from the fractionation of the initial aqueous extract, contained the lowest amount of (poly)phenols (16.0 ± 2.0 mg GAE g −1 ). Furthermore, DPPH · and ABTS + methods were applied. ZLF parental extract, ZLF-W, and the derived ZLF-W1 proved to be ineffective as DPPH radical scavengers (data not shown). Similarly, ZLF-O did not show antiradical capacity ( Figure 3A), whereas ZLF-A exerted a dose-dependent radical scavenging capacity, with an ID 50 value equal to 92.8 µg/mL. The obtained results were also evaluated with reference to a positive standard (Trolox ® ), and the TEAC (Trolox ® Equivalents Antioxidant Capacity, µg Trolox ® per g of extract) value of 7.13 was calculated for ZLF-A.
be ineffective as DPPH radical scavengers (data not shown). Similarly, ZLF-O did not show antiradical capacity ( Figure 3A), whereas ZLF-A exerted a dose-dependent radical scavenging capacity, with an ID50 value equal to 92.8 μg/mL. The obtained results were also evaluated with reference to a positive standard (Trolox ® ), and the TEAC (Trolox ® Equivalents Antioxidant Capacity, μg Trolox ® per g of extract) value of 7.13 was calculated for ZLF-A. Analogously, ZLF-O was not able to efficaciously scavenge ABTS •+ , while ZLF-A revealed an extremely high activity, even at the lowest tested dose (12.5 μg/mL) ( Figure 3B). The observed RSC (%) values showed the need to further dilute the sample up to 0.78 μg/mL, thus highlighting that its massive antiradical potential was strongly dose-dependent in the concentration range 0.7-12.5 μg/mL. An ID50 value of 3.68 μg/mL was calculated, and the TEAC value was equal to 0.40 μg/mL. ZLF-O and ZLF-A chemical composition could explain their difference in antiradical capacity. In fact, ZLF-A, mainly enriched in flavonoids, was more likely to elicit radical trapping effects. Analogously, ZLF-O was not able to efficaciously scavenge ABTS •+ , while ZLF-A revealed an extremely high activity, even at the lowest tested dose (12.5 µg/mL) ( Figure 3B). The observed RSC (%) values showed the need to further dilute the sample up to 0.78 µg/mL, thus highlighting that its massive antiradical potential was strongly dose-dependent in the concentration range 0.7-12.5 µg/mL. An ID 50 value of 3.68 µg/mL was calculated, and the TEAC value was equal to 0.40 µg/mL. ZLF-O and ZLF-A chemical composition could explain their difference in antiradical capacity. In fact, ZLF-A, mainly enriched in flavonoids, was more likely to elicit radical trapping effects.
The relative addition of both ZLF-O and ZLF-A extracts in cosmetic emulsion ( Figure 4A) highlighted ZLF-A antiradical capability. In fact, ZLF-A enriched emulsion, whose optical microscopy images are in Figure 4 (panel D), appeared to be able to scavenge ABTS radical cation, reaching a RSC(%) value equal to 54%, when 0.10% w/w ZLF-A enriched the base emulsion ( Figure 4B). The dose-response effect was also highlighted in the DPPH radical scavenging capacity assay, whereas ZLF-O-based cream did not lead to effective response ( Figure 4C).
The relative addition of both ZLF-O and ZLF-A extracts in cosmetic emulsion ( Figure 4A) highlighted ZLF-A antiradical capability. In fact, ZLF-A enriched emulsion, whose optical microscopy images are in Figure 4 (panel D), appeared to be able to scavenge ABTS radical cation, reaching a RSC(%) value equal to 54%, when 0.10%w/w ZLF-A enriched the base emulsion ( Figure 4B). The dose-response effect was also highlighted in the DPPH radical scavenging capacity assay, whereas ZLF-O-based cream did not lead to effective response ( Figure 4C).

Cytotoxicity Screening of a Cosmeceutical Formulation Containing ZLF-A Extract
MTT test data on HaCaT cell line put the basis for the usage of ZLF-A extract in the development of a cosmeceutical preparation, which could guarantee a beneficial effect on skin cells, following a constant and prolonged application. To this purpose, the prepared cosmetic emulsions (0.10%w/w of added ZLF-A extract) underwent an in vitro screening, in order to highlight its cytotoxic potential towards HaCat human keratinocytes and MRC-5 human fibroblasts. Cells were exposed for 15′ or 2h 30′ to ZLF-A enriched emulsion. HaCat cell line proved to be moderately sensitive to the short-term exposure. In fact, it was observed that, after 15 min exposure time, ZLF-A enriched cream inhibited by 26.7 ± 0.8% the mitochondrial redox activity; its effect was slightly higher than that estimated for the base emulsion, used as control (24.5 ± 0.5%). When the exposure time increased, cell adaptation occurrence seemed to be defined as cell viability decrease was calculated equal to 16.7 ± 0.7%.
MRC-5 cell line showed an interesting responsiveness to the MTT assay, in that it revealed a more pronounced resistance to the treatment with the cosmeceutical preparation containing ZLF-A extract. In fact, the initial mitochondrial redox activity inhibition (19.6 ± 0.4%), evaluated at 15 min exposure time, decreased to 12.8 ± 0.7% after 2 h 30 min.
In order to further evaluate the safety of ZLF-A enriched cosmetic emulsion, an in vitro reconstructed human epidermis (RhE; EpiSkin ™ ), from normal human keratinocytes and therefore histologically similar to the in vivo tissue, was used. Due to the ban of animal experiments taking place in March 2009 (7th amendment of the cosmetic directive, EU 2003), for its ability to mimic the human skin this model represents an effective alternative [33]. After validation by the European Union (EU), it was included in the relevant EU Test Method Regulation (440/2008/EC) [34] and, also, certified by the Organization for Economic Cooperation and Development (OECD) [35]. Following manufacturer recommendations, MTT test was applied after exposure of the RhE to the emulsion containing 0.010 and 0.10%w/w of ZLF-A extract.
The mitochondrial redox activity (RA) of the cells constituting the reproduced epidermis was compared to those obtained with the crude extract and with Triton X-100, used as positive standard, able to induce skin irritation ( Figure 6).

Cytotoxicity Screening of a Cosmeceutical Formulation Containing ZLF-A Extract
MTT test data on HaCaT cell line put the basis for the usage of ZLF-A extract in the development of a cosmeceutical preparation, which could guarantee a beneficial effect on skin cells, following a constant and prolonged application. To this purpose, the prepared cosmetic emulsions (0.10% w/w of added ZLF-A extract) underwent an in vitro screening, in order to highlight its cytotoxic potential towards HaCat human keratinocytes and MRC-5 human fibroblasts. Cells were exposed for 15 or 2 h 30 to ZLF-A enriched emulsion. HaCat cell line proved to be moderately sensitive to the short-term exposure. In fact, it was observed that, after 15 min exposure time, ZLF-A enriched cream inhibited by 26.7 ± 0.8% the mitochondrial redox activity; its effect was slightly higher than that estimated for the base emulsion, used as control (24.5 ± 0.5%). When the exposure time increased, cell adaptation occurrence seemed to be defined as cell viability decrease was calculated equal to 16.7 ± 0.7%.
MRC-5 cell line showed an interesting responsiveness to the MTT assay, in that it revealed a more pronounced resistance to the treatment with the cosmeceutical preparation containing ZLF-A extract. In fact, the initial mitochondrial redox activity inhibition (19.6 ± 0.4%), evaluated at 15 min exposure time, decreased to 12.8 ± 0.7% after 2 h 30 min.
In order to further evaluate the safety of ZLF-A enriched cosmetic emulsion, an in vitro reconstructed human epidermis (RhE; EpiSkin ™ ), from normal human keratinocytes and therefore histologically similar to the in vivo tissue, was used. Due to the ban of animal experiments taking place in March 2009 (7th amendment of the cosmetic directive, EU 2003), for its ability to mimic the human skin this model represents an effective alternative [33]. After validation by the European Union (EU), it was included in the relevant EU Test Method Regulation (440/2008/EC) [34] and, also, certified by the Organization for Economic Cooperation and Development (OECD) [35]. Following manufacturer recommendations, MTT test was applied after exposure of the RhE to the emulsion containing 0.010 and 0.10% w/w of ZLF-A extract.
The mitochondrial redox activity (RA) of the cells constituting the reproduced epidermis was compared to those obtained with the crude extract and with Triton X-100, used as positive standard, able to induce skin irritation ( Figure 6). Molecules 2019, 24, x 14 of 22 Figure 6. Mitochondrial redox activity (RA (%)) estimated by MTT test in Episkin ™ reconstructed human epidermis exposed to ZLF-A crude extract and ZLF-A-enriched cosmeceutical formulation.
The results highlighted a very favourable outcome in terms of residual cellular redox activity following application. In fact, the RA% of the emulsion containing the lower extract amount (0.010%w/w) was equal to 86.5 ± 0.9% and increased to 93.6 ± 0.8% when the concentration was 10-fold higher. A similar trend was observed after the exposure to the crude plant complex, in that the increase of the extract tested dose resulted in an enhanced RA% (from 94.4 ± 0.6% up to 105.3 ± 0.6%), demonstrating that the treatment ensured cytoprotection, cell proliferation, preservation of mitochondrial dehydrogenase reduction activity, and even an increase in its intensity. Moreover, in both cases the residual mitochondrial redox activity values were about twice the value found for the reference standard (50.3 ± 0.7%).
Histological examination was carried out on reconstructed human epidermis (RhE). Tissues exhibit from four to seven viable layers comprising at least basal, suprabasal, spinous and granular cell layers, and stratum corneum. According data from antiradical tests and MTT assay, histological observations of EpiSkin reconstructed tissue treated with ZLF-A (0.10%w/w)-enriched cream revealed a good tissue integrity with compacted stratum corneum ( Figure 7A). Analogously, tissues' preservation was observed after their exposure with ZLF-A extract alone (0.10%w/w). In this context, the stratum corneum retained its open basket-weave appearance ( Figure 7B). Triton X-100 treatment defined tissue's integrity loss and death ( Figure 7C). The results highlighted a very favourable outcome in terms of residual cellular redox activity following application. In fact, the RA% of the emulsion containing the lower extract amount (0.010% w/w ) was equal to 86.5 ± 0.9% and increased to 93.6 ± 0.8% when the concentration was 10-fold higher. A similar trend was observed after the exposure to the crude plant complex, in that the increase of the extract tested dose resulted in an enhanced RA% (from 94.4 ± 0.6% up to 105.3 ± 0.6%), demonstrating that the treatment ensured cytoprotection, cell proliferation, preservation of mitochondrial dehydrogenase reduction activity, and even an increase in its intensity. Moreover, in both cases the residual mitochondrial redox activity values were about twice the value found for the reference standard (50.3 ± 0.7%).
Histological examination was carried out on reconstructed human epidermis (RhE). Tissues exhibit from four to seven viable layers comprising at least basal, suprabasal, spinous and granular cell layers, and stratum corneum. According data from antiradical tests and MTT assay, histological observations of EpiSkin reconstructed tissue treated with ZLF-A (0.10% w/w )-enriched cream revealed a good tissue integrity with compacted stratum corneum ( Figure 7A). Analogously, tissues' preservation was observed after their exposure with ZLF-A extract alone (0.10% w/w ). In this context, the stratum corneum retained its open basket-weave appearance ( Figure 7B). Triton X-100 treatment defined tissue's integrity loss and death ( Figure 7C). The results highlighted a very favourable outcome in terms of residual cellular redox activity following application. In fact, the RA% of the emulsion containing the lower extract amount (0.010%w/w) was equal to 86.5 ± 0.9% and increased to 93.6 ± 0.8% when the concentration was 10-fold higher. A similar trend was observed after the exposure to the crude plant complex, in that the increase of the extract tested dose resulted in an enhanced RA% (from 94.4 ± 0.6% up to 105.3 ± 0.6%), demonstrating that the treatment ensured cytoprotection, cell proliferation, preservation of mitochondrial dehydrogenase reduction activity, and even an increase in its intensity. Moreover, in both cases the residual mitochondrial redox activity values were about twice the value found for the reference standard (50.3 ± 0.7%).
Histological examination was carried out on reconstructed human epidermis (RhE). Tissues exhibit from four to seven viable layers comprising at least basal, suprabasal, spinous and granular cell layers, and stratum corneum. According data from antiradical tests and MTT assay, histological observations of EpiSkin reconstructed tissue treated with ZLF-A (0.10%w/w)-enriched cream revealed a good tissue integrity with compacted stratum corneum ( Figure 7A). Analogously, tissues' preservation was observed after their exposure with ZLF-A extract alone (0.10%w/w). In this context, the stratum corneum retained its open basket-weave appearance ( Figure 7B). Triton X-100 treatment defined tissue's integrity loss and death ( Figure 7C).

Preliminary in Vitro Stability Tests of ZLF-A Enriched Emulsion
During the formulation of a cosmetic product an investigation on its stability is necessary in order to guarantee the quality and safety for consumers in the foreseeable conditions of storage and use. Most cosmetic formulations are tested with emphasis on microbiological stability, whereas their physicochemical stability is only marginally investigated. Instead, a number of factors during the shelf-life of the product, such as temperature and pH changes, can affect the quality of the formulation. For a preliminary evaluation of the in vitro stability aliquots of 50 g of ZLF-A-enriched emulsion and of base cream were packaged in opaque white polyethylene flasks and stored at three different temperatures (4 ± 0.5 • C, 25 ± 2 • C and 37 ± 3 • C) and four storage periods (48 h, 7, 30 and 90 days). The room temperature (25 • C) was used to provide information about the behaviour of the product when stored under appropriate storage conditions, whereas the upper temperature mimics an accelerated aging. At each pre-determined time, samples underwent centrifugation, which produces stress in the sample, simulating an increase in gravity force and increasing the mobility of the particles, thus anticipating possible instabilities. These changes may appear in the form of precipitation, separation of phases, caking, or coalescence among others. Likely due to the proper homogenization during emulsion formulation [36], none of the cosmetic formulations exhibited phase separation, which evidenced their extremely high stability at mechanical stress, regardless of the storage temperature to which it has been exposed.
Monitoring the pH value is crucial for determining the emulsions' stability. In fact, pH changes indicate the occurrence of chemical reactions that can give an idea on the quality of the final product [37].
The pH values of all the samples ranged from 4.15 ± 0.09 to 6.03 ± 0.08 and remained quite constant during storage, although there was a slight pH decrease after 7 days. It is generally recommended that skin care products should exhibit a pH of 5-5.5 to not disturb the physiological pH of the skin. In contrast to the viable epidermal cell layers, the stratum corneum is known to have an acidic pH between 4.1 and 5.8 under physiological conditions-the acid mantle-critical for the maintenance of the barrier homeostasis, regulation of cohesion/desquamation and antimicrobial defence. However, in certain conditions, e.g., dry or aged skin, there is a significant increase of skin pH leading to impaired barrier function and reduced buffer capacity, hence fostering the development of pathological skin conditions. The increased pH may be reversed by exogenous acidification of the stratum corneum [38]. In the light of the above considerations, the prepared emulsion is not only suitable for topical application, but even may be the best option to improve the acid mantle and skin barrier function. Indeed, Cucurbita pepo fruits are a common, high-value fruit vegetable [39], whose anti-genotoxicity and chemopreventive potential were recently explored [40]. The usage of pumpkin in cosmetic sector is not uncommon as its seeds, extracted by cold pressure, produce an oil rich in antioxidant and antimicrobial components. Tocopherols, sterols and polyunsaturated fatty acids in pumpkin oil were found to make it a cosmetic formula able to provide protection against dermatological wound [41].

Plant Extraction and Fractionation
Zucchini cv. 'Lungo Fiorentino' were harvested in July 2016 from a biological cultivation in Tuscany (latitude 43 • 42 32.0 N and longitude 10 • 57 45.8 E) and represent agro-wastes as they were out of gauge, and damaged by bad weather. The whole zucchini, once cleaned, underwent a freeze-drying procedure (FTS-System Flex-Dry, SP Scientific, Stone Ridge, NY, USA). 585 g of the lyophilized product were powdered by a rotary knife homogenizer and then wetted and subjected to maceration. To this end, three extraction cycles (30 min each) were performed in a 40 kHz ultrasonic bath (Branson M3800, Carouge, Switzerland) with 450 mL of EtOH for each sample. The parental extract obtained (ZLF, 10 g) was solubilised in water, and subsequently underwent discontinuous liquid-liquid extraction, using ethyl acetate as the extraction solvent. As a result, an organic fraction (ZLF-O) and an aqueous fraction (ZLF-W) were obtained. This latter was further chromatographed on Amberlite XAD-4 (h 70 cm, Ø 4.0 cm), eluting with water first (ZLF-W1) and then with MeOH (ZLF-A). The extraction/fractionation scheme is depicted in Figure 8.

Plant Extraction and Fractionation
Zucchini cv. 'Lungo Fiorentino' were harvested in July 2016 from a biological cultivation in Tuscany (latitude 43°42′32.0″ N and longitude 10°57′45.8″ E) and represent agro-wastes as they were out of gauge, and damaged by bad weather. The whole zucchini, once cleaned, underwent a freezedrying procedure (FTS-System Flex-Dry, SP Scientific, Stone Ridge, NY, USA). 585 g of the lyophilized product were powdered by a rotary knife homogenizer and then wetted and subjected to maceration. To this end, three extraction cycles (30 min each) were performed in a 40 kHz ultrasonic bath (Branson M3800, Carouge, Switzerland) with 450 mL of EtOH for each sample. The parental extract obtained (ZLF, 10 g) was solubilised in water, and subsequently underwent discontinuous liquid-liquid extraction, using ethyl acetate as the extraction solvent. As a result, an organic fraction (ZLF-O) and an aqueous fraction (ZLF-W) were obtained. This latter was further chromatographed on Amberlite XAD-4 (h 70 cm, Ø 4.0 cm), eluting with water first (ZLF-W1) and then with MeOH (ZLF-A). The extraction/fractionation scheme is depicted in Figure 8.

UHPLC-HRMS Analyses
ZLF-O and ZLF-A fractions were investigated for their metabolic content by UHPLC-ESI-QqTOF-MS/MS techniques. To this purpose, a Shimadzu NEXERA UHPLC system was used with a Luna ® Omega Polar C18 column (1.6 μm particle size, 150 × 2.1 mm i.d., Phenomenex, Torrance, CA, USA). ZLF-O elution was achieved with a linear gradient of water (A) and acetonitrile (B), both with 0.1% formic acid under a linear gradient 5 → 95% B in 10 min, followed by 2 min of re-equilibration time in the initial conditions. A different linear gradient was used for the analysis of ZLF-A fraction: 0-5 min, 5 → 10% B; 5-10 min, 10 → 15% B; 10-12 min, 15 → 20% B; 12-13 min, 20 → 40% B; 13-15 min, 40% B. Then, also in this case the starting conditions were restored and the column was allowed to re-equilibrate for 2 min. In both cases a flow rate of 0.5 mL min −1 and an injection volume of 2.0 μL were set.
MS analysis was performed using the AB SCIEX TripleTOF 4600 (AB Sciex, Concord, ON, Canada) system with a DuoSpray™ ion source operating in negative electrospray ionisation. The APCI probe of the source was used for fully automatic mass calibration using the Calibrant Delivery System (CDS). CDS injects a calibration solution matching polarity of ionization and calibrates the mass axis of the TripleTOF ® system in all scan functions used (MS or MS/MS). Data were collected by information dependent acquisition (IDA) using parameters as in Table 3. The following parameter settings were also used: declustering potential (DP), 70 V; ion spray voltage, −4500 V; ion source heater, 600 °C; curtain gas, 35 psi; ion source gas, 60 psi. Data processing was performed using the PeakView ® -Analyst ® TF 1.7 Software. Quantitative analysis of ZLF-A identified constituents was also

UHPLC-HRMS Analyses
ZLF-O and ZLF-A fractions were investigated for their metabolic content by UHPLC-ESI-QqTOF-MS/MS techniques. To this purpose, a Shimadzu NEXERA UHPLC system was used with a Luna ® Omega Polar C18 column (1.6 µm particle size, 150 × 2.1 mm i.d., Phenomenex, Torrance, CA, USA). ZLF-O elution was achieved with a linear gradient of water (A) and acetonitrile (B), both with 0.1% formic acid under a linear gradient 5 → 95% B in 10 min, followed by 2 min of re-equilibration time in the initial conditions. A different linear gradient was used for the analysis of ZLF-A fraction: 0-5 min, 5 → 10% B; 5-10 min, 10 → 15% B; 10-12 min, 15 → 20% B; 12-13 min, 20 → 40% B; 13-15 min, 40% B. Then, also in this case the starting conditions were restored and the column was allowed to re-equilibrate for 2 min. In both cases a flow rate of 0.5 mL min −1 and an injection volume of 2.0 µL were set.
MS analysis was performed using the AB SCIEX TripleTOF 4600 (AB Sciex, Concord, ON, Canada) system with a DuoSpray™ ion source operating in negative electrospray ionisation. The APCI probe of the source was used for fully automatic mass calibration using the Calibrant Delivery System (CDS). CDS injects a calibration solution matching polarity of ionization and calibrates the mass axis of the TripleTOF ® system in all scan functions used (MS or MS/MS). Data were collected by information dependent acquisition (IDA) using parameters as in Table 3. The following parameter settings were also used: declustering potential (DP), 70 V; ion spray voltage, −4500 V; ion source heater, 600 • C; curtain gas, 35 psi; ion source gas, 60 psi. Data processing was performed using the PeakView ® -Analyst ® TF 1.7 Software. Quantitative analysis of ZLF-A identified constituents was also carried out. Rutin (y = 125837x + 26331; R 2 = 0.9998) was used as reference standard for identified glycosylated flavonoids and rebaudioside A (y = 208034x − 1638.2; R 2 = 1) for the ent-kaurene diterpene glycoside. Calibration curves of both the standards were constructed, based on analyses performed in the same experimental conditions of ZFL-A fraction. Results were expressed as mean ± SD values of three independent measurements.

Determination of Total Phenols
Total phenol amount of ZLF parental extract and of the fractions derived was determined according to the Folin-Ciocalteau procedure [42] with slight modifications. Analyzed samples (0.25 and 0.50 mg/mL in DMSO) were mixed with 0.250 mL of the Folin-Ciocalteau reagent (FCR) and 2.25 mL of Na 2 CO 3 (7.5% w/v). After stirring the reaction mixture at room temperature for 2 h, 300 µL of each sample were transferred into a multiwell plate and the absorbance was read at 765 nm using a Wallac Victor 3 multilabel plate reader (PerkinElmer Inc., Waltham, MA, USA). The content of total phenols of the samples was expressed as milligram gallic acid equivalents (GAEs) per g of dried extract ± standard deviation (SD).

Determination of DPPH Radical Scavenging Capacity
DPPH radical scavenging capability was determined as previously reported [43] with slight modifications. Briefly, ZLF-O and ZLF-A, tested at different final concentration levels (12.5, 25.0, 50.0 and 100.0 µg/mL), were dissolved in a DPPH • methanol solution (9.4 × 10 −5 M) at room temperature. After stirring the reaction mixture at room temperature for 20 min, 300 µL of each sample were transferred into a multiwell plate and the absorbance was read at 520 nm using a Wallac Victor 3 multilabel plate reader (PerkinElmer Inc., Waltham, MA, USA). Results are the mean ± SD values. ID 50 , based on the percentage decrease of the initial DPPH • absorption by the different concentrations of the test samples, and TEAC (Trolox ® Equivalent Antioxidant Capacity) values were also calculated.

Determination of ABTS Radical Cation Scavenging Capacity
ABTS radical cation was generated by reacting ABTS (7.0 mM) and K 2 S 2 O 8 (2.45 mM). The mixture was allowed to stand in the dark at room temperature for 16 h, and then diluted with PBS (pH 7.4) in order to reach an absorbance of 0.70 at 734 nm. ZLF-O and ZLF-A extracts (12.5, 25.0, 50.0 and 100.0 µg/mL, final concentration levels) were dissolved in 1.0 mL of diluted ABTS •+ solution. After 6 min of incubation, 300 µL of each sample were transferred into a multiwell plate and the absorbance was read at 734 nm using a Wallac Victor 3 multilabel plate reader (PerkinElmer Inc., Waltham, MA, USA). Results, expressed in terms of the percentage decrease of the initial ABTS •+ absorption by the test samples, are the mean ± SD values [44]. ID 50 and TEAC values were also calculated.

Cell culture and Cytotoxicity Assessment
Human HaCat keratinocyte were grown in RPMI 1600 high glucose medium supplemented with 10% Newborn Calf Serum, 50.0 U/mL penicillin, and 100.0 µg/mL streptomycin, at 37 • C in a humidified atmosphere containing 5% CO 2 . Human MRC-5 lung fibroblasts cell lines were grown in the same conditions except for the medium, which was DMEM supplemented with 10% Fetal Bovine Serum. Cells were seeded in 96-multiwell plates at a density of 1.5 × 10 4 cells/well. After 24 h cells were treated with ZLF-O and ZLF-A extracts at 25.0, 50.0, 100.0, 200.0 and 500.0 µg/mL final dose levels. At 24, 48 and 72 h of incubation, inhibition of mitochondrial redox activity was determined by MTT cell viability test as previously described [45].

Cream Formulation
Base cream was prepared as an oil-in-water disperse system by heating separately the aqueous phase (45%) and the oily phase containing white wax (5%) and sweet almond oil (40%). ZLF-O and ZLF-A extracts, previously solubilized in pure ethanol, were incorporated into the base cream in two different amounts (0.010 and 0.10% w/w ). Then, the emulsion was further enriched with ascorbic acid (0.08%), quercetin (0.005%) and Lavandula officinalis L. essential oil (3.0%) for antioxidant, preservative and antimicrobic purposes.

Determination of Antiradical Activity
Base cream and the emulsions containing zucchini extracts were tested for their scavenging efficacy towards DPPH • and ABTS •+ . Samples were prepared with a protocol similar to that previously described by Pelle et al. [46]. Briefly, emulsions were dissolved in iPrOH (1% w/v) and 5-fold diluted with distilled water. Then, the radical probe (DPPH • or ABTS •+ ) was added and the spectrophotometric analysis were carried out as described for pure extracts (see Sections 2.5 and 2.6).

Cytotoxicity Assessment in an In Vitro Reconstructed Human Epidermis (EpiSkin™)
Reconstructed Human Epidermis (small, age day 17-0.50 cm 2 ) was removed from the agarose nutritive solution provided by the manufacturer, transferred in a 6-well plate containing fresh growth medium and incubated at 37 • C in a humidified atmosphere containing 5% CO 2 . After 24 h a small amount (60 µg/cm 2 ) of the base cream or of that enriched with ZLF-A extract was deposited onto the surface of the tissues and spread with a small paintbrush. The cultures were incubated at 37 • C for 24 h, after which mitochondrial redox activity inhibition was determined by MTT test.

Histological Analysis on Reconstructed Human Epidermis (EpiSkin™)
An important EpiSkin quality control criteria is based on histological examination [47]. After the treatment, tissues were rapidly immersed in Bouin's fluid. The samples were dehydrated through a graded ethanol series and finally embedded in paraffin. Then, paraffin sections (5 µm thickness) were stained with haematoxylin and eosin stain and examined under a light microscope (Leica DM LB) for histological evaluation. Photographs were taken using the Leica ICC50 HD digital camera [48][49][50].

Stability Tests
Stability tests were performed at different conditions, predictive of different climatic conditions, in order to explore their effects on the emulsions' storage. To this aim, samples were kept at 4 ± 0.5 • C, 25 ± 2 • C and 37 ± 3 • C. Phase separation was evaluated after 48 h, 7, 30 and 90 days of storage, centrifuging 1 g of each sample at 1500 rpm for 10 min, then at 300 rpm for 10 min and at 4000 rpm for 5 min using an Allegra™ 64R centrifuge (Beckman, Palo Alto, CA, USA), equipped with a F2402H rotor. The pH measurements were repeated at each storage time on all the emulsions kept at different conditions (1 g dissolved in 15 mL of pure water).

Conclusions
Fruits and vegetables, with their undeniable nutritional and nutraceutical value, are an essential part of daily diet, but agricultural processing, in spite of what comes on our tables, mainly based on consumer's demand, generates abundant raw materials, which can range from whole, and unexploited materials to fractions and/or mixtures produced by physical, thermal, chemical processing. Indeed, vegetable and fruit trimmings could be more valuable than the main products [51], thanks to their diversity in sugars, minerals, organic acids, dietary fibres and bioactive compounds, which could be similar to that in their edible and pleasantly palatable counterparts. Herein, the exploitation and re-evaluation of zucchini cv. 'Lungo Fiorentino' wastes, through ultrasound assisted maceration, an eco-friendly extraction processes, represent a valuable alternative to recycle disposable products in green cosmetics. The application of UHPLC-MRMS analytical method to identify polyphenols and other polar, as well as apolar compounds, for the first time, in Cucurbita pepo cv. 'Lungo Fiorentino' allowed the deep fruit investigation for its nutraceutical and cosmeceutical purposes. The recovery of fine and precious chemicals from zucchini wastes was in accordance with valorisation and sustainable development principles, and was prompt to convey a new skin care concept through products in which the nutraceutical zucchini value was further exploited for its cosmeceutical strength.
Supplementary Materials: The following are available online, Figure S1: TOF-MS 2 spectrum of compound 2 and the hypothesized structure of the product ion at m/z 609.1484, Figure S2: (a) TOF-MS 2 spectra of metabolites 16 and 17; (b) Proposed fragmentation pathway (measured exact mass of each fragment ion, as listed in Table 2, was at m/z value to within 5 ppm vs. its relative theoretical m/z value. This latter was reported below each structure), Figure S3: TOF-MS spectra of (a) the ent-kaurene diterpene glycoside (compound 18) tentatively identified in ZLF-A fraction; (c) rebaudioside A commercial standard. TOF-MS 2 spectra of (b) the ent-kaurene diterpene glycoside (compound 18) tentatively identified in ZLF-A fraction; (d) rebaudioside A commercial standard, Table S1: Quantitative data of ZLF-A identified constituents.