Chemical and nutritional characteristics of Cannabis sativa L. co‐products

Abstract Cannabis sativa L. is an annual herbaceous plant. It was used for centuries to obtain different products. In the last century, hemp cultivation was forbidden due to the psychoactive effects of tetrahydrocannabinol acid (THCA). In the last years, new strains, characterized by high cannabidiolic acid (CBDA) and low THCA level, were developed renewing the interest in hemp cultivation to obtain food or to extract essential oils from flowers. All these processes produce many residues with different chemical–physical characteristics. In order to evaluate their potential use also in animal nutrition, some hemp co‐products were evaluated. Two different co‐products of seed processes (flour and oil) and two co‐products obtained trimming the flowers, differing in granulometry were used. The samples were analysed for chemical composition and evaluated in vitro using the gas production technique with buffaloes' ruminal inoculum. All hemp co‐products showed interesting nutritional characteristics, such as crude protein content always higher than 20% on a dry matter basis, and high neutral detergent fibre concentration partially lignified. The in vitro gas production parameters at 120 h of incubation showed quite low fermentability testified by the low organic matter degradability and cumulative gas volume (OMD from 28.09 to 45.64% and OMCV from 110 to 164 ml/g, respectively). Also, the methane produced after 24 h of incubation was particularly low (from 1.78 to 11.73 ml/g dOM). These results could be due to the high lipid and ash amounts or to the CBDA content that probably affected the CH4 formation processes. According to preliminary results obtained by this study, it is possible to hypothesize that these co‐products could be useful to mitigate the methane production into the rumen. Further studies are necessary in order to evaluate the correct inclusion into the diet for ruminants.

to several reasons. First, the competition with other natural fibres such as cotton and jute for textile applications, and the intensive development of synthetic fibres. Moreover, Cannabis sativa L.
Moreover, cultivars of Cannabis sativa L. are rich in cannabidiolic acid (CBDA), the main phytocannabinoid of this plant particularly concentrated in its products (fibre and seeds) (Fiorini et al., 2019).
Nowadays, the renewal of hemp cultivation was increased throughout the world. In Europe, the cultivation areas for hemp increased over 50% (from 12,232 to 24,939 ha) from 2008 to 2018 (FAOSTAT, 2020). Indeed, hemp could be considered a valid alternative to the conventional crops produced in excess due to its limited environmental impact (less use of water and land exploitation).
Hemp provides many agricultural benefits, such as weed control, pest and disease resistance, fast growth, capacity to remove significant quantities of heavy metals from the soil (bioremediation), and to provide high biomass production with low inputs. These characteristics make C. sativa an adaptable, affordable, and ecological crop (Ranelli & Venturi, 2004).
The cultivation and processing of hemp have many purposes, especially in food production to obtain many products, such as decorticated seeds, flour and oil. These products are considered sources of essential fatty acids and dietary fibre and are used as ingredients for several food preparations such as biscuits, bread, and protein powder. The seeds of C. sativa present on average 25-35% of lipids and 20-25% of proteins. The fatty acid profile of hemp seeds includes ω3 and ω6 essential fatty acids; in particular, the oil shows an abundance of α-linolenic acid providing a ω6/ω3 ratio nearly to 2.5-3.0/1.0, within the range indicated as optimal for human dietary recommendations (Faugno et al., 2019). Additionally, hemp is also rich in natural antioxidants and other bioactive components (peptides, phenolic compounds, tocopherols, carotenoids and phytosterols) (Farinon et al., 2020).
During the flowers and seeds processing to obtain hemp prepared food various residues are produced. We have hypothesized that these co-products could be used also for ruminant nutrition considering their nutritional properties. The study aimed to characterize hemp co-products in terms of chemical composition and in vitro fermentation characteristics. The amounts of phytocannabinoids were also measured.

| Sampling
All samples were collected in a storage shed, sited in the province of Caserta (South Italy). Totally four different hemp co-products were evaluated. All evaluated substrates derive from the same cultivar (Futura 75). This cultivar, particularly widespread in South Italy, is usually sowed, for seed production, in March-April and mechanically harvested in September-October. While the flowers were obtained from previous harvesting (July-September), immediately after, the apical portion of the plant was dried 21-23°C for 10 days, subsequently, the flowers were mechanically separated (trimmed) and a lot of biomass were discarded. Two co-products were obtained after trimming the flowers and sieving at different granulometry (rough_ FR and thin_FT). This biomass was partially chopped, and, for the evaluation, we have separated by sieving the chopped biomass into two portions, higher or lower than 4 mm. Considering one kilo of biomass 565 and 435 g of FR and FT were respectively obtained.
Two co-products were obtained from the seed processing, which was made within 20 days from the harvesting. SF was the hemp oil meal obtained by cold mechanical extraction while, SO was the flour of the seeds discarded because they were too small or too light.

| Chemical composition
The samples were ground with a 1-mm screen and analysed for chemical composition (dry matter, crude protein, ether extract and ash) as reported by AOAC (2015) procedures (ID number: 2001.12, 978.04, 920.39 and 930.05 for DM, CP, EE and ash, respectively).
Neutral detergent fibre, acid detergent fibre and free ash acid detergent lignin were also determined as indicated by Van Soest et al. (1991).

| Phytocannabinoid content assessment
The phytocannabinoid contents were detected at the University of Campania, Luigi Vanvitelli (Caserta, Italy), where the substrates were first ground using mortar and pestle in liquid nitrogen, and then extracted through an ultrasound-assisted maceration (UAM). For this purpose, an ultrasonic bath system was used (Branson UltrasonicsTM BransonicTM M3800-E, Danbury, CT, USA), with n-hexane as the extracting solvent. Three ultrasound cycles of 30 min each were carried out, with operating frequency set at 40 kHz, and solid to solvent ratio equal to 1:5. The obtained extracts were dried using a rotary evaporator (Heidolph Hei-Vap Advantage, Schwabach, Germany), and reconstituted in n-hexane for the UHPLC-ESI-QqTOF-MS and MS/MS analyses. Thus, the apolar hemp by-products extracts (10 mg/ml) were investigated using a Shimadzu NEXERA UHPLC system (Shimadzu, Tokyo, Japan) equipped with Luna Omega C18 column (50 × 2.1 mm i.d., 1.6 µm particle size). The mobile phase consisted of a binary solution composed of water (solvent A) and acetonitrile (solvent B), both acidified with formic acid (0.1% v/v).
A linear gradient was used as follows: 0-1 min, 38% B; 1-5 min, and Δ 9 -tetrahydrocannabinolic acid (THCA), were constructed. Both the compounds were previously isolated and fully characterized by means of spectroscopic and spectrometric techniques . Thus, working solutions of each standard, prepared by dilution from a stock solution, were injected into the UHPLC-ESI-QqTOF MS system under the same conditions as the samples .

| In vitro fermentation
All samples were incubated in a serum flask (seven replications per substrate ±1 g for each replication) with buffalo rumen fluid (10 ml) at 39°C under anaerobic conditions as indicated by Theodorou et al. (1994). The rumen liquor was collected, at slaughterhouse, from four buffaloes according to EU legislation (EU Council, 2004). The buffaloes fed a total mixed ration containing corn silage, oat hay and concentrate. All procedures involving animals were approved by the Ethical Animal Care and Use Committee of the University of Napoli Federico II (Prot. 2019/0013729 of 08/02/2019). The collected rumen fluids were placed inside to pre-heated thermos and transported within 2 h to the laboratory of Feed Evaluation of University of Napoli, Federico II. The rumen fluid was mixed and strained through four layers of cheese cloths and diluted in a buffered medium (75 ml), successively, the reducing agent (4 ml) was added into the flasks (Vastolo et al., 2020). On seven replication, three bottles for each substrate were utilized for cumulative gas production measurement (120 h of incubation), the remaining were used for methane production evaluation (24 h of incubation).
The gas produced during 120 h of incubation, into the fermenting cultures, was recorded 21 times (from 2 to 24 h of intervals) using a manual pressure transducer (Cole and Palmer Instrument Co, Vernon Hills, IL, USA). The cumulative volume of gas produced after 120 h of incubation was related to incubated OM (OMCV, ml/g). At the end of the incubation period, the fermentation liquor was analysed for pH using a pH meter (ThermoOrion 720 A+, Fort Collins, CO, USA). The organic matter degradability (OMD, %) was determined by weight difference of the incubated OM and the undegraded filtered (sintered glass crucibles; Schott Duran, Mainz, Germany, porosity # 2) residue burned at 550°C for 3 h.

| End-products measurement
In order to determine the volatile fatty acids (VFA), the fermentation liquor was cooled at 4°C and, before analyses, centrifuged at 12,000 g for 10 min at 4°C (Universal 32R centrifuge, Hettich FurnTech Division DIY, Melle-Neuenkirchen, Germany); the supernatant (1 ml) was then mixed with 1 ml of 0.06 mol oxalic acid.

| Methane production evaluation
For each substrate, four flasks of seven were stopped at 24 h to measure the methane (CH 4 ) production as described by Guglielmelli et al. (2011); the relative end-products were also determined.
The gas-phase from each flask was sampled (3 ml) in duplicate with a gastight syringe and injected into a gas chromatograph (ThermoQuest 8000top Italia SpA, Rodano, Milan, Italy), equipped with a loop TC detector and a packed column (HaySepQ SUPELCO, 3/16-inch, 80/100 mesh).

| Data processing
To estimate the fermentation kinetics, for each bottle stopped at 120 h, the gas production profiles were fitted to the sigmoidal model (Groot et al., 1996): where G is the total gas produced (ml per g of incubated OM) at time t (h), A is the asymptotic gas production (ml/g), B is the time at which one-half of A is reached (h), and C is the curve switch. Maximum fermentation rate (R max , ml/h) and the time at which it occurs (T max , h) were calculated utilizing model parameters (Bauer et al., 2001): In particular, the in vitro parameters concerning 120 h of incubation (OMCV, OMD, T max , R max ) and the data related to the endproducts (pH, VFA, BCFA, N-NH 3 ) were statically analysed, as well as the results of the methane analysis (methane production as percentage of total gas, related to incubated organic matter, and related to degraded organic matter, respectively). The significance level was verified using HSD Tukey's test at p < 0.01 and p < 0.05.
The correlations between chemical composition values and fermentation parameters and between chemical compound and methane production were also evaluated (JMP®, Version 14 SW, SAS Institute Inc., Cary, NC, USA, 1989-2019). Table 1 shows the chemical composition of the tested samples.

| Chemical composition
The co-products of hemp seeds presented a significantly higher (p < 0.01) level of crude protein (>30% DM) than flower co-products.
The structural carbohydrates were quite high and variable for all the tested samples. Indeed, NDF values were higher (p < 0.01) in seeds' samples (>38% DM) than in flowers co-products (37 and 31% DM); lignin content was always higher than 6% DM. Moreover, both seeds co-products showed significantly higher (p < 0.01) values of all structural carbohydrates' fractions. While the ether extract content of flowers co-products was significantly (p < 0.05) higher than seed coproducts. Ash content of both seeds co-products was significantly (p < 0.01) lower compared to flower ones.

| Phytocannabinoid content
UHPLC-HRMS analysis on a polar extract from the hemp coproducts matrices provided qualitative information on their phytocannabinoids' content, whereas the main acidic phytocannabinoids, namely CBDA and Δ 9 -THCA, were quantized thanks to calibration curves available from the pure isolated reference compounds

| In vitro fermentation characteristics
The in vitro fermentation characteristics and kinetics are reported in

| End-products parameters after 120 h of incubation
In tVFA). The flower co-products showed significantly higher (p < 0.01) acetate and propionate production compared to seeds co-products.
The ratio acetate/propionate was significantly higher for SF compared to the other samples. Regarding butyrate production, the seeds co-products showed the highest production compared to flowers co-products.

| Methane production
The methane production registered after 24 h of incubation, the organic matter degradability and relative tVFA values are reported in

| DISCUSS ION
Regarding chemical composition, the data of hemp co-products of this study agreed to that reported in literature (Alaru et al., 2011;Gibb et al., 2005;Vonapartis et al., 2015) for flower and seeds coproducts. In this regard, hemp co-products chemical composition, particularly of seeds, could be comparable to soybean meal ones (Semwogerere et al., 2020). Indeed, taking into count the high protein content and considering the high amount of essential amino acids reported by House et al. (2010) for similar substrates, all tested co-products could be evaluated as a source of high-quality protein.
Moreover, the crude protein amount of hemp co-products is higher than the endorsed ruminant dietary requirement for maintenance and growth (CSIRO, 2007). In particular, these co-products seem to be able to satisfy the nutritional requirements for buffaloes in early lactation as indicating by Bartocci et al. (2002). As reported by Mustafa et al. (1999)   Abbreviations: SF, refusal of flour production; SO, refusal of oil extraction; FR, refusal of trimming of flower (rough); FT, refusal of trimming of flower (thin); OMD, organic matter degradability; OMVC, cumulative volume of gas related to incubate organic matter; R max , maximum fermentation rate; T max , time at which R max occurs.
high content of structural carbohydrates could make available a high energy level in these substrates. Nevertheless, the particularly high amount of lignin could limit the availability of these resources (House et al., 2010). However, the active molecules present in the hemp The substrates SO and SF showed the highest OMD and the lowest OMCV values. The most intense fermentation could be due to the high crude protein concentration probably, considering that protein fermentation did not produce gas (Abreu and Bruno Soares, 1998). When the microorganisms are not able to ferment the carbohydrates, they have to use other nutrients such as protein, in order to survive. Also the highest BCFA proportion an N-NH 3 production registered by these substrates after 120 h of incubation could indicate a protein fermentation. Moreover, the higher values of BCFA and ammonia were significantly related to the protein content (BCFA vs. CP: r = 0.813, p < 0.05 and N-NH 3 vs. CP: r = 0.727, p < 0.05); indeed, no correlation was found between BCFA and N-NH 3 production, as described by Davila et al. (2013) this could be due to the specific amino acids profile. In this regard, Isobutyrate and Isovalerate production are the result of branched-chains amino acids metabolism, such as valine, leucine and isoleucine; they can be hydrolyzed and fermented to phenols, and biogenic amines (i.e., indole, skatole, 4-ethylphenol, p-cresol) (Musco et al., 2018). Semwogerere et al. (2020) reported the amino acid profile of seeds and different co-products indicating that leucine and valine were among the main represented amino acids in hemp.
The shape representing the fermentation rate of the hemp coproducts indicates that FR is a quite slow substrate, probably due to some component that does not facilitate the enzymatic attack by microorganisms. On the contrary, FT and SF show a rapid fermentation process, due to an easily fermentable component. About the SO, the curve is in an intermediate position.
Regarding the methane production, the data of this study seem particularly low compared to that one obtained in previous studies (Calabrò et al., 2012;Guglielmelli et al., 2011) and reported by other authors (Tuyen et al., 2013;Elghandour et al., 2016), probably due to the different experimental condition in terms of hours of incubation. The limited methane production could be ascribed to specific nutritional characteristics of hemp co-products or to the TA B L E 4 In vitro fermentation end products evaluated after 120 h of incubation (n = 6)  Shibata and Terada (2010) indicated that the supplementation of unsaturated fatty acids (linoleic, α-linolenic and oleic acids) in the ruminant's diet inhibits CH 4 production into the rumen.
No correlations were observed between OMD and cannabinoids or methane production and cannabinoids. All these results suggest that acidic phytocannabinoids, mainly CBDA and TCHA could affect the fermentation pathways without limiting the organic matter degradability.

| CON CLUS ION
These preliminary results underline how the use of residues of hemp processing could be a valid nutrient resource in the diet of ruminants. Indeed, hemp co-products showed interesting nutritive values as demonstrated by the results of the chemical composition.
However, in vitro data (i.e. OMD and OMCV) indicate as these substrates are quite low utilized by the rumen microbes. However, the low methane values suggest as these samples could be used in the ruminant diets to contain gas emission, and therefore, the environmental impact. Further studies are needed to evaluate if it is possible to improve the hemp structural carbohydrates utilization fermentability and to investigate the mechanisms that determine the trend registered in vitro. Moreover, could be interesting to conduce in vivo studies to understand in what doses these co-products could be included in ruminants' diet and if they could be useful to mitigate the environmental impact of intensive farm.

ACK N OWLED G EM ENTS
The authors want to thank Dr Maria Ferrara for her technical support in chromatographic analysis at the Department of Veterinary Medicine and Animal Production, University of Napoli Federico II, Italy.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R S ' CO NTR I B UTI O N S
SC and MIC designed the research. AV collected the data. AV, BIK and SP made the laboratory analysis. BIK and AV analysed the data.
BIK and AV wrote the manuscript. SC, SP and MIC reviewed and edited the manuscript. All the authors read and approved the final manuscript.

A N I M A L WE LFA R E S TATE M E NT
All procedures involved animals were approved by the Ethical Animal Care and Use Committee of the University of Napoli Federico II (Prot. 2019/0013729 of 08/02/2019).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets used and/or analysed during the current study are available from the corresponding author on reasonable request.