Antiflammatory activity and potential dermatological applications of characterized humic acids from a lignite and a green compost

Long-term exposure to air pollution has been associated with the development of some inflammatory processes related to skin. The goal of modern medicine is the development of new products with antiflammatory action deriving from natural sources to improve environmental and economic sustainability. In this study, two different humic acids (HA) were isolated from from lignite (HA-LIG) and composted artichoke wastes (HA-CYN) and characterized by infrared spectrometry, NMR spectroscopy, thermochemolysis-GC/MS, and high-performance size-exclusion chromatography (HPSEC), while their antiflammatory activity was evaluated on HaCaT cells. Spectroscopic results showing the predominance of apolar aliphatic and aromatic components in HA-LIG, whereas HA-CYN revealed a presence of polysaccharides and polyphenolic lignin residues. The HA application on human keratinocyte pre-treated with Urban Dust revealed a general increase of viability suggesting a protective effect of humic matter due to the content of aromatic, phenolic and lignin components. Conversely, the gene expression of IL-6 and IL-1β cytokines indicated a significant decrease after application of HA-LIG, thus exhibiting a greater antiflammatory power than HA-CYN. The specific combination of HA protective hydrophobic components, viable conformational arrangements, and content of bioactive molecules, suggests an innovative applicability of humic matter in dermatology as skin protectors from environmental irritants and as antiflammatory agents.

In the last decades, global air quality has shown a progressive decline, a direct effect of human activities such as biomass burning, industrial operations and vehicle emissions 1 . The World Health Organization (WHO) estimated that air pollution was implicated into 3.7 million deaths representing more than 10% of all-cause deaths in industrialized countries 2 , thereby rendering air pollution as the world's largest single environmental health risk 3 .
Recently, large epidemiological studies identified that exposure to air-borne pollution increases respiratory and cardiovascular morbidity 4,5 . Moreover, long-term exposure to air pollution has been associated with the development of certain cancers and may affect some processes related to skin, including aging, and heighten the clinical manifestations of psoriasis, acne, and atopic dermatitis 6 . In fact, skin is considered the first barrier of human body to the external factors and is constantly exposed to harmful compounds existing in the surrounding environment 7 .
The most common air pollutants comprise polycyclic aromatic hydrocarbons (PAHs), volatile organic compounds (VOCs, e.g., benzene), particulate matter (PM, most commonly PM2.5 and PM10), gaseous pollutants, (carbon-monoxide, nitric oxides, sulphur oxide, ozone, heavy metals) and indoor pollutants, as solid fuels consumption 8 . Although the molecular mechanisms relating air pollutions to skin diseases are not yet fully Infrared spectroscopy. The ATR spectra of HA-LIG and HA-CYN showed poorly diagnostic broad bands in the 3200-3400 cm −1 range, assigned to O-H and N-H stretching vibrations (Fig. 1). Both HA samples were characterized by shoulders at 2850 and 2920 cm −1 , associated to symmetric and asymmetric stretching vibration of CH 2 and CH 3 groups in long chain alkyl compounds 23 . The main infrared differences between the two HA were shown at smaller wavenumbers (Fig. 1). The prominent peak at 1565 cm −1 in the HA-LIG spectrum is related to ring vibrations in aromatic structures, while the less intense signal ay 1372 and 1217 cm-1 are currently assigned to C-O bending of ethers and phenolic compounds 24 . In the HA-CYN spectrum, the sharp wide band at around 1635 cm −1 may include both C=O stretching vibrations of carboxylate groups of alkanoic acids and the vibrations of conjugated and unconjugated double C=C bonds 23,25 . The small signals at 1450 and 1220 cm −1 derive from the bending of C-H and C-O groups in alkyl chains and phenolic components, respectively. Finally, a large incorporation of carbohydrates derivatives in HA-CYN from green compost was indicated by the intense band around 1013 cm −1 , due to the C-O stretchings in the alcohol functions in pyranoside rings 25 .

Off-line Pyr-TMAH-GC-MS.
The pyrogram of HA-LIG revealed a typical profile of a humic acid of geochemical origin characterized almost exclusively by alkyl and aromatic apolar components ( Figure S1) 26 . The main aromatic molecules identified in the first part of the spectrum, were related to benzene, toluene, m-and p-xylene, styrene, naphthalene and alkyl-naphthalene, phenolic structures, lignin compounds and benzoic acids. The peaks at longer retention times were assigned to long alkyl chains of alkanoic acids and alkane/alkene doublets (Table S1). The data of HA-LIG revealed its main hydrophobic features, with long alkyl chains and aromatic moieties being currently ascribed to thermal degradation products of selectively preserved apolar fractions. These components are due to the advanced structural modification of natural compounds during the accumulation processes of geo-organic sediments 27 . The pyrogram of HA-CYN was mainly composed by ester and ether derivatives of lignin fragments and linear and branched chain fatty acids followed by biopolyesters, carbohydrates derivatives and N-containing compounds (Table S2). The most abundant lignin monomers were the oxidized products of both di-and tri-methoxy www.nature.com/scientificreports/ phenylpropane molecules, with aldehydic-, ketonic-, benzoic-acid, and benzene-propenoic acids as main components ( Figure S1; Table S2). This finding combined with the lower yields of less altered lignin units indicated the prevalent incorporation in HA-CYN of bio-available lignin fragments from composted plant residues 18,19,25 . In the wide range of long chain alkanoic acids (Table S1), the largest abundance was that of the ubiquitous hexadecanoic and octadecanoic homologues, while a prevalence of even-carbon linear chains suggests a plant origin 25 . The presence of branched chain fatty acids found in the HA-CYN pyrogram was related to natural microbial constituents built up in the mature compost from which HA-CYN was extracted 25 . In respect to HA-LIG, the larger abundance of polar compounds in HA-CYN included methylated forms of mono-and oligosaccharides components and nitrogenated compounds, such as peptide, indole, and pyrrole derivatives (Table S2).
13 C-CPMAS NMR spectra. The 13 C-CPMAS NMR spectrum of HA-LIG ( Fig. 2) confirmed the predominance of apolar aliphatic and aromatic components which represented 28.7 and 42.2% of the total carbon content, respectively ( Table 2). The two main resonances at 33 and 128 ppm indicated CH 2 units in long alkyl chains and in C-substituted aromatic compounds, respectively, while the shoulder at 152 ppm is attributed to phenolic carbons (Fig. 2). A diversified C distribution among different functional groups was found in the solid-state NMR spectrum of HA-CYN ( Table 2). The distinct signals around the broad resonance at 30 ppm in the 0-45 ppm range, are due to of methyl and methylene groups in alkyl chains of various linear lipid compounds, while the less intense peaks at 40-45 ppm may be due to tertiary (CH) and quaternary (C-R) carbons in assembled rings of sterol, flavonoids and lignans 25 . The prominent resonance at 57 ppm is mainly assigned to methoxy substituents in aromatic lignin monomers, with an additional contribution of C-N bonds of peptidic moieties 18,19 . The signals in the 60-110 ppm chemical shift pertain to various C-O groups in carbohydrates and polysaccharides from plant tissues. The resonance from 110 to 145 ppm is related to aromatic carbons, followed by the one of aryl C-O groups (145-160 ppm) in phenolic and lignin derivatives, which are commonly corresponding to resonances in the 145-60 ppm range 24 . The signal at 174 ppm is due to carboxyl carbons in different compounds, such as fatty acids, esters and peptides. The molecular properties of HA samples may be inferred by the dimensionless structural parameters such as aromaticity (ARM) and hydrophobicity (HB/HI) indexes and alkyl (A/OA), and lignin (LigR) ratios (Table 2), that are based on integration of spectral intervals 18,19,25 .
The large values of HB/HI and ARM indexes and of A/OA ratio shown by the HA-LIG sample further highlight the large hydrophobic character provided to this material by the combined contribution of alkyl and aromatic structures. Conversely, less aromatic and alkyl carbon content, concomitant to an enhanced amount of C-O and C-N containing compounds, determined smaller values of HB/HI and A/OA for HA-CYN (Table 2). A different molecular composition of aryl C between the two HA was also suggested by the LigR ratio (Table 2), that resulted greater for HA-CYN, thus indicating a significant presence of lignin residues, also confirmed by the correlation between signals of methoxyl groups (45-60 ppm) and those of aryl C-O molecules (140-160 ppm), that agrees with the selective incorporation of lignin derivatives in HA-CYN revealed by thermochemolysis analyses 18,19,25 .
Conversely, the decrease of LigR value in HA-LIG indicated a smaller content of the methoxyl groups of lignin  www.nature.com/scientificreports/ molecules and a prevalence of phenol and polyphenolic components derived by the geochemical processes that formed the lignite from which HA-LIG was extracted.

Carboxyl-C Phenol-C Aryl-C O-Alkyl-C Methoxyl-C Alkyl-C Structural indexes
1 H NMR spectra. 1 H NMR spectra of the HA-LIG and HA-CYN (Fig. 3) can be divided into three main regions: 0-3 ppm, alkylic region (mostly aliphatics components); 3-5.5 ppm, O-Alkyl region (prevalently oligoand polysaccharides); 6.5-8.5 ppm, aromatics region (mainly lignin components) 28,29 . The 1 H NMR spectrum of HA-LIG shows several sharp signals emerging from the large resonance in the 0.9-2.4 ppm range, a single predominant signal at 3.3 ppm in the O-Alkyl region, and a conspicuous broad resonance in the aromatic and heterocyclic region over the 7-8.5 ppm range. The appearance of this spectrum is in line with the indications of the solid-state spectrum and confirms the large hydrophobicity of this humic matter of geochemical origin that contains prevalently alkyl and aromatic compounds. Conversely, the HA-CYN spectrum reveals a less intense resonance in both alkyl and aromatic region, but a greater signal intensity in the O-alkyl regions. Again, the features of the HA-CYN spectrum confirms the significantly smaller hydrophobicity of this material, as already noted for the CPMAS spectrum, and its richness in polar aliphatic compounds, such as the saccharidic and peptidic moieties derived from artichoke residues transformed in the green compost from which HA-CYN was isolated.
High performance size exclusion chromatography (HPSEC) of HA. The HPSEC elution profiles before and after the treatment of humic solutions with acetic acid to lower the pH of the eluent from 7 to 3.5, has been used to evaluate the conformational structure of the loosely bound supramolecular associations of humic acids 30,31 . The chromatograms of both HA-LIG and HA-CYN before and after acetic acid additions are shown in Fig. 4. The two HA had a different HPSEC behaviour. In fact, while HA-LIG showed one single peak when www.nature.com/scientificreports/ injected at pH 7, the elution profile was disrupted in four different peaks after having adjusted the pH to 3.5 with AcOH before injection. Conversely, the elution profile of HA-CYN did not change between the two treatments revealing the same three peaks, although shifted to larger elution volumes due to the AcOH addition. The different HPSEC behaviour of the two HA was reflected by the parameters calculated from the chromatograms (Table 3). For example, the Weight Average molecular weight (Mw) of HA-LIG was about 11,960 Da for the diffuse peak of the chromatogram of the material injected at pH 7, while the four peaks after pH adjustment to 3.5 revealed Mw of about 21,800, 10,370, 4240 and 180, respectively, passing from the one at smaller elution volumes, corresponding to material of greater molecular size, to that of lesser molecular size at larger elution volumes. Conversely, in the case of HA-CYN, the MW of the three peaks maintained similar values before and after AcOH addition, although the values observed for the latter treatment were lower than for former one, thereby agreeing with the noted peaks shift to larger elution volumes (Fig. 4). These results may be interpreted with the supramolecular understanding of humus from soil that describes the conformational structure of the HA as being stabilized by dispersive weak interactions among relatively small heterogeneous molecules, which may be disrupted by AcOH addition, due to formation of intermolecular hydrogen bonds stronger than the  Table 3. Weight average (Mw) and number average (Mn) molecular weights, and polydispersity (P), as calculated from UV-detected HPSEC chromatograms for HA-LIG and HA-CYN, before and after addition of acetic acid (AcOH). Standard deviation was < 5%. www.nature.com/scientificreports/ previous dispersive forces 26,30,31 . While this is confirmed for HA-LIG, whose material distribution in four peaks suggests a supramolecular association of smaller molecules, the same cannot be said for HA-CYN. In fact, the latter conformational arrangement appears to be mostly stabilized by covalent-bonded biopolymers, that survived the microbial composting process and are resilient to the disruption capacity of the AcOH treatment 32 . These may well be the polysaccharidic, peptidic, and polyphenolic lignin fragments, whose presence is indicated by FTIR, NMR and thermochemolysis results. Altogether, these findings indicate that HA-LIG was composed of relatively smaller compounds than HA-CYN, thereby conferring to the former material a greater conformational flexibility than the latter.
Antiflammatory activity of HA: in vitro assays on HaCaT cells. Cells viability. Cell viability was assessed by the Trypan blue method with different increasing concentrations of HA-LIG and HA-CYN. The treatment of human keratinocyte cells with both humic extracts did not influence cells viability despite the maximum applied concentration ( Figure S2). Contrary to earlier works 33,34 , these results exclude a possible cytotoxic activity of these HA on human cells cultures. Additionally, HaCaT cells pretreated with both HA at the same concentration (50 µg mL −1 ) showed a statistically significant viability increase, as compared to the cells treated with only Urban Dust ( Figure S2). In fact, the cell viability after application of HA-CYN and HA-LIG was still 87% and 82% of that of the untreated control, respectively, whereas the inflammatory stimulation by Urban Dust decreased viability of HaCaT cells by 62%, as also shown by previous studies 35 . These results suggest a protective effect by humic matter towards human cells survival. A similar behaviour was reported by fulvic acids which apparently protected the activity of RAW 264.7 human cells cultures 36 .
The capacity of natural organic matter to protect the activity of human cells is an important prerequisite for the development of novel and sustainable antioxidant systems 37 . In fact, an oxidative cell stress typically indicates that the balance between ROS and the antioxidant system is compromised 38 , thus fostering an increase in cells free-radicals content that causes oxidative damages, such as DNA modification, and lipid oxidation or peroxidation 39 . It has been suggested that the antioxidant and protective role of humic materials may be related to their redox properties 40 , which may inactivate reactive oxygen species by their binding to the humic structure through a reversible and reproducible process 41 . This behaviour may well explain our results indicating that addition of humic matter induces a normal proliferation of HaCaT, cells despite the Urban Dust treatment 42 . However, no univocal structure-activity relationship was yet produced to explain the molecular bases of HA bioactivity against human diseases. In this work, the protection properties of HA-CYN are related to the large content of phenolic and lignin components and in accordance to previous studies on phenolic materials, in which monoand oligo-hydroxylated aromatic units were shown to exhibit protective activities 18,19,43 . Similarly, the protection activity of HA-LIG is be related to its hydrophobic composition as reported in both NMR and thermochemolysis results (Tables 2, S1, S2; Figs. 1, 3). While previous studies pointed out that the antimicrobial activities promoted by humic and Humic-like substances (HULIS) against plants pathogens depended on their specific molecular composition 18,19,44,45 , we show for the first time that HA isolated from composted agricultural biomasses exert protection properties on human cells and this depends on the molecular composition of the tested humic acid.
Effects of HA-LIG and HA-CYN on IL-1β and IL-6 gene expression after Urban Dust exposure. The keratinocyte model is commonly used to evaluate the effects of irritants in dermatology due the presence of these cells in the outer layers of the skin and their preliminary role in the encounter with foreign irritants 46 . Moreover, keratinocytes play a crucial role to maintain the barrier of the stratum corneum that protects the inner layers of skin cells 47 . Additionally, keratinocytes can produce a wide range of inflammatory mediators which further induce the upregulation and secretion of secondary mediators, including chemokines, resulting in the recruitment of leukocytes towards the damaged area of the skin 48 .
It has been reported that fine dust (FD) elicited inflammatory responses in keratinocytes and transferred inflammatory responses to macrophages possibly mediated by the proinflammatory cytokine Interleukin-1β (IL-1β) 49 . IL-1β is a potent pro-inflammatory cytokine that is crucial for host-defence responses to infection and injury produced and secreted by the innate immune system cells, such as monocytes and macrophages 50 . It is produced without a signal sequence and does not follow the conventional route of protein secretion, but rather employs one or more non-conventional pathways of secretion 51 . Conversely, interleukin-6 (IL-6) is produced transiently in response to infection during its acute phase. The continuous dysregulated synthesis of IL-6 induced a pathological effect such as chronic inflammation and autoimmunity, although its normal expression is tightly controlled by transcriptional and post-transcriptional mechanisms 52 . Furthermore, IL-6 plays a crucial role in inducing an epithelial response to environmental inputs 35 .
In this work, the gene expression of IL-1β and IL-6 was assessed to verify the effective antiflammatory efficacy of HA-CYN and HA-LIG on HaCaT cells treated with Urban Dust, as suggested by preliminary viability experiments (Fig. 5). Our results revealed a significant increase of IL-1β and IL-6 gene expression after the stimulation by Urban Dust (Fig. 5a,b). Conversely, an application of Urban Dust on HaCaT cells after their treatment with either HA-LIG or HA-CYN induced a decrease of cytokines expression (Fig. 5a,b). HA-LIG was more effective than HA-CYN by inducing a less IL-1β expression, whereas the HA-CYN treatment exhibited a slightly larger anti-flammatory activity in the case of IL-6. Moreover, the analysis of markers levels of the inflammatory process in HaCaT cells suggested the beneficial effects of both humic extracts to reduce the inflammation induced by Urban Dust (Fig. 5a,b).
These results are in accordance with previous studies that reported an inhibitory effect of humic derivates on cytokines release associated with an in vitro 22,53 . The biochemical explanation of antiflammatory effects of HA was related to the inhibition of the 5-lipoxygenase pathway of the arachidonic acid (AA) cascade by naturally occurring HA as well as synthetic HA-like polymers 53  www.nature.com/scientificreports/ AA release by either natural sodium humates or humic-like products obtained by caffeic acid oxidation (KOP) supported the characterization of HA as antiflammatory materials 22 . On the other hand, it has been found that humic substances stimulated the release of pro-inflammatory cytokines, such as TNF-α in vitro, but only in the presence of exogenous lipopolysaccharides, thus suggesting that humic substances alone do not cause inflammation under normal conditions 54 . Additionally, potassium humates were reported to induce a positive effect on the proliferation of lymphocytes cells 20 . In this case, cells proliferation was increased as a response to autocrine growth pathway characterized by synthesis of cytokines and expression of T cell surface receptors for these mediators, as well as by the production of the primary autocrine growth factor, such as IL-2, and the enhanced expression of the IL-2 receptor (CD25) 55 . Our findings represent an innovative approach to relate the effect on HaCaT cells to the molecular composition of humic matter.
Several studies pointed out the occurrence of beneficial effects of humic matter, also from composted biomasses, on either plants or microorganisms at molecular, physiological and transcriptional levels 56 . A working hypothesis of the bioactive effects of humic matter relies on the concomitant combination of conformational structure and molecular composition on the interaction with a different type of cells 57 . In fact, the difference between HA-CYN and HA-LIG in antiflammatory activities on the treated HaCaT cells is explained by the different molecular and conformational properties of the employed HA, as indicated by NMR, thermochemolysis and HPSEC results. The inhibition of the release of IL-6 and IL-1β inflammatory cytokine by HA-LIG when added at 50 µg mL −1 before Urban Dust may be related to its dominant hydrophobic properties (large HB/HI index), whereas the slighter inhibition of the IL-1β cytokine by HA-CYN, at the same concentration of HA-LIG, may be accounted to its large content of polyphenolic lignin components.
The different conformational stability of HA revealed by the changes in HPSEC elution profiles after AcOH addition may also represent an additional factor that contributes to the antiflammatory behavior of HA (Table 3, Fig. 4). However, in the case of cells viability (Figure S2), both HA-LIG and HA-CYN showed the same positive protective effect, despite the former was found to be composed of smaller molecular associations than the latter. This indicates that the HA molecular composition, rather than the conformational stability, is responsible for the effectiveness of the protection of cells viability. In particular, it is the content of hydrophobic components in HA-LIG, as shown by the large values of HB/HI, ARM and A/OA indexes, and the amount of lignin-derived polyphenols in HA-CYN, as revealed by the great LigR ratio (Table 2), to make both materials effective in cell protection. This finding is in line with previous studies in which phenolic components were shown to suppress human chronic diseases, and with other ones that reported that the hydrophobicity and phenolic content, rather than molecular size, stimulated the activity of plant root cells. Conversely, the greater effect of HA-LIG in the suppression of the expression of both IL-1β and IL-6 cytokines (Fig. 5a,b) may be explained by the lesser conformational stability of this material than HA-CYN that exhibits a more rigid conformational structure. In fact, the HA-LIG conformation could be disrupted into smaller humic superstructures by interacting with HaCaT cells and concomitantly release small bioactive phenolic molecules that significantly reduces the inflammation. Conversely, in the case of HA-CYN, its observed conformational stability prevents a similar break down in Limitations of study. The aim of this study was to investigate the anti-inflammatory properties of two characterized humic acids (HAs) extracted from two natural sources. Despite the molecular differences between HA were extensively explained, we hypothesize that a possible limitation of this work consists in having used only the humic acid fraction of compost humus, while fulvic acids alone or a mixture of the two (HA and FA) may reveal different anti-inflammatory processes. However, for this very reason, the observed relationship between molecular structure and biological activity of HA assumes an even greater role in our work since it describes a specific anti-inflammatory activity of the HA fraction. An additional shortcoming of our work may be due to have evaluated the anti-inflammatory activity on a single cell line and quantified a limited number of key inflammatory cytokines. Our preferential selection of IL-6 and IL-1β as two inflammatory markers was based on previous studies that have shown their role as highly sensitive molecules of the cytokine pathway to the inflammatory stimulation 58,59 , thereby justifying their choice as preliminary biomarkers of humic matter bioactivity.
These considerations imply that further studies are required to explore in greater details the beneficial properties of humic materials in medical applications 60 and identify the molecular and cellular mechanisms underpinning their anti-inflammatory proprieties. Nevertheless, we believe in the pioneering validity of our study since our unprecedented results can stimulate further investigations of the anti-inflammatory activity of well characterized humic matter on human keratinocytes.

Conclusions
Medical applications of humic substances are related to the anti-viral, antiflammatory, antioxidant, anti-tumoral and anti-toxin proprieties of these materials. Our findings showed antiflammatory activities of two different HA isolated from either a non-renewable geochemical deposit, such as lignite, or a renewable and sustainable source like green compost. Their difference in molecular compositions was evaluated by FTIR spectrometry, NMR spectroscopy, off-line thermochemolysis associated to GC-MS, while their conformational structure was assessed by HPSEC. Our results from specific assays suggest that HA antiflammatory activities on HaCaT cells were partially related to their hydrophobic properties conferred by aromatic compounds and polyphenolic lignin fragments and partially on the conformational stability of their supramolecular associations.
This work innovatively evaluated HA bioactivity on a keratinocyte model after stimulation by Urban Dust and on the consequent HaCaT cellular production of IL-1β and IL-6 cytokines, by attempting a relation between the HA molecular properties and their observed bioactivities. It is suggested that HA bioactive properties are due not only to their degree of hydrophobicity that promotes their adhesion on the surface of the target cells and increase general cells viability, but also to the conformational arrangement of their molecular complexity. We found that HA associated in less stable conformation may facilitate the release of bioactive components when in interactions with cells and decrease cellular expression of inflammatory cytokines. We believe that this preliminary identification of the necessary element to draw a structure-activity relationship between well characterized HA and their bioactivity represents an important step forward in supporting the potential exploitation of natural sustainable materials in the dermatological field.

Materials and methods
Extraction of humic acids (HA). Humic acids were extracted from both a North Dakota lignite (HA-LIG) and a green compost made by artichoke residues (HA-CYN) 18,19 . The green compost was produced in the composting facility of the Experimental Farm of the University of Napoli Federico II at Castel-Volturno (CE) and was obtained by mixing residues of artichoke production with woodchips from poplar trimming at 70/30 w/w ratio. The on-farm composting processes were performed on static piles spread on an air insufflation system formed by perforated rubber tubes and a rotative pump. The composting processes lasted 100 days with an initial 45 days of active phase where the pile temperature reached about 70 °C. At the end of composting process, the piles were randomly sampled to obtain a final amount of 1 kg for each compost. The organic materials were air dried, sieved at 2 mm and stored a 4 °C until further analyses.
HA were obtained by suspending 100 g of humic matter (artichoke compost and North Dakota lignite, respectively) in 1000 mL of 1 M KOH solution and mechanically shaking the mixture for 24 h. The suspension was then centrifuged at 7000 rpm for 20 min and the supernatant filtered through glass-wool. The extraction was repeated twice. The combined extracts were acidified to pH 1 with 6 M HCl and allowed to settle for 24 h. Then, the samples were centrifuged at 4000 rpm for 20 min and the HA residues dialyzed (1kD cutoff Spectrapore membranes) against deionized water until Cl-free. The HA were then freeze-dried for further analytical characterization. www.nature.com/scientificreports/ a furnace (Barnstead Thermolyne 21,100). The products released by thermochemolysis were transferred online by a helium flow (20 mL min −1 ) into a series of two chloroform (50 mL) traps kept in ice/salt baths. The extracts were concentrated by rota-evaporation. The residue was resuspended in 1 mL of chloroform and transferred in a glass vial for GC-MS analysis. The GC-MS analyses were performed with a Perkin-Elmer Autosystem XL by using a RTX-5MS WCOT capillary column, (Restek, 30 m × 0.25 mm; film thickness, 0.25 μm) that was coupled, through a heated transfer line (250 °C), to a PE Turbomass-Gold quadrupole mass spectrometer. The chromatographic separation was carried out with the following program: 60 °C (1 min isothermal), rate 7 °C min −1 to 320 °C (10 min isothermal). Helium was applied as carrier gas at 1.60 mL min −1 , the injector temperature was at 250 °C, the split-injection mode had a 30 mL min −1 of split flow. Mass spectra were obtained in EI mode (70 eV), scanning in the range 45-650 m/z, with a cycle time of 1 s. Compound identification was based on comparison of mass spectra with the NIST library database, previous published spectra and standard.

Elemental analyses and FTIR
NMR spectroscopy of HA. A 300 MHz Bruker Avance spectrometer, equipped with a 4 mm wide-bore MAS probe, was used to obtain solid-state HA spectra. Each fine-powdered sample (5 mg) was packed into a 4 mm zirconium rotor, provided with a Kel-F cap, and spun at a rate of 13,000 ± 1 Hz.  www.nature.com/scientificreports/ USA) was performed to assess gene expression of IL-6 and IL-1β. PCR primers were designed based on published sequences, and their specificity was verified with BLAST alignment search. The amount of mRNA for a given gene in each sample was normalized to the amount of mRNA of 18S reference gene in the same sample.
Statistics. All statistical analyses were performed using GraphPad Prism 4.0 (GraphPad Software Inc, La Jolla, CA, USA). Data that passed the normality test were analysed with two-tailed, t test. Values of P < 0.05 were considered significant.

Data availability
All data generated or analysed during this study are included in this published article or in the accompanying Supplementary Information files.