2.1. Extraction Yields and Chemical Characterization of Labdanum and Fraction Extracts
In this study, labdanum resin, extracted according to the Andalusian process (Adl), yielded 5.79 ± 0.52% (dw/fw) (
Table 1). The Zamorean process (Zam) was performed as a control to evaluate the importance of alkaline extraction of labdanum resin. In fact, the Zam extraction yielded only 0.23 ± 0.07%, which is approximately 25 times lower than the yield of Adl resin. In this study, Zam resin was successfully isolated, however, using acidification to precipitate it, as was done for the Adl resin. Despite the difference of extraction yields, both resins rendered similar absolute yields of around 70% (dw/dw). Diterpenoid and flavonoid fractions were obtained by column chromatography from the Adl labdanum absolute, rendering 79.86 ± 0.76% and 10.64 ± 2.17% (dw/dw) yields, respectively. A remainder fraction, lost during the purification process, represented 9.48 ± 1.87% (dw/dw) of the absolute. As the yield of Zam labdanum resin was too low, it was not fractionated.
The chemical profile of the labdanum resins and purified fractions was assessed by GC-EI-MS with a prior TMS derivatization reaction of the mixtures. Fragment ion 73
m/
z was used to confirm TMS derivatives [
26]. Overall, chromatograms of extracts revealed 37 peaks (
Figures S1–S4) with a peak area above 2% in relation to the major peak, which depended on the extract (indicated in
Table 2). Using standards, eight peaks could be identified as typical labdane-type diterpenoids or flavonoid aglycones. Using the NIST library, only 3-phenylpropanoid acid compound could be identified as a TMS derivative, with an excellent fragmentation pattern match of 921 (peak 2, 9.64 min). Other compounds were identified using NIST library with a good match (800–900); however, most were methyl ester derivatives. Those compounds, identified as methyl ester derivatives, presented a significant amount of ion (73
m/
z), which means that they are most likely TMS derivatives. Interestingly, carboxylic acids (phenylpropanoids, fatty acids and diterpenoid acids) presented ion at 73
m/
z, but alcohols (rhododendrol and flavonoids) did not, which is inconsistent with what is described by Harvey and Vouros [
26].
As observed in
Table 2, the diterpenoid fraction (
Figure S2) was shown to be mostly composed by labdane-type diterpenoid acids, the flavonoid fractions (
Figure S3) were shown to be mostly composed by methylated flavonoid aglycones with apigenin and kaempferol skeleton, and the Adl labdanum absolute (
Figure S1) was shown to be mostly composed by labdane-type diterpenoid acids, fatty acids, and jaranol. Zam labdanum absolute (
Figure S4) was shown to be mostly composed of phenylpropanoids, fatty acids, and several non-identified compounds, and to have a different chemical profile to the Adl labdanum absolute.
2.2. Evaluation of Anti-Diabetic and Neuroprotective Potential of Labdanum and Fraction Extracts
Concerning the anti-diabetic activity, inhibition of the enzymes involved in the starch breakdown, which reduces glucose availability to intestinal absorption, was assessed. Inhibition of α-amylase and α-glucosidase activity by labdanum absolutes and purified fractions is presented in
Table 3 and
Table 4, respectively. Diterpenoid fraction was the extract with the highest α-amylase inhibitory activity (around 40 and 30% inhibition at 1 and 0.5 mg/mL, respectively) followed by Adl labdanum absolute. Flavonoid fraction presented low α-amylase inhibitory capacity (~4% at 1 mg/mL), and Zam labdanum absolute did not inhibited the enzyme (
Table 3). In contrast (
Table 4), Zam labdanum absolute produced the highest inhibitory effect on α-glucosidase activity (around 25% at 1 mg/mL), followed by Adl labdanum absolute (~14% inhibition at 1 mg/mL) and flavonoid fraction, which showed similar inhibitory effect. Diterpenoid fraction showed the lowest capacity to inhibit α-glucosidase activity (~6.5% inhibition, at 1 mg/mL).
Comparing Dit and Flv fractions, while the Dit fraction presented higher potential in inhibiting α-amylase, the Flv fraction presented higher potential in inhibiting α-glucosidase. Of the compounds identified in
Table 2, apigenin, genkwanin and acacetin have been tested for their anti-α-amylase and anti-α-glucosidase activities [
27]. In fact, apigenin and genkwanin produced higher α-glucosidase inhibition when compared to α-amylase, supporting the results presented here (
Table 3 and
Table 4). Acacetin showed similar inhibitory activity for both enzymes [
27].
A variety of fungi- and plant-based metabolites belonging to several classes of compounds, including terpenoids and flavonoids, have shown α-amylase and α-glucosidase inhibitory activity [
20,
28,
29]. Acarbose is currently one of the drugs used to control hyperglycemia in individuals with type-2 diabetes by competitively and reversibly inhibiting intestinal α-glucosidases, a family of enzymes that hydrolyzes oligosaccharides into glucose and other monosaccharides, and non-reversibly inhibiting α-amylase, secreted by salivary glands and exocrine pancreas, that cleave complex polysaccharides (e.g., starch) into oligosaccharides [
20,
30]. We observed that the acarbose inhibitory effect was higher against α-amylase (~75% and ~87% inhibition at 0.5 mg/mL and 1 mg/mL, respectively) than against α-glucosidase (~70% and ~80% inhibition at 0.5 mg/mL and 1 mg/mL, respectively); this inhibitory trend is reported by other authors (e.g., [
31]).
Labdanum absolutes and fractions are not so effective in inhibiting α-amylase (
Table 3) or α-glucosidase (
Table 4) as acarbose; however, this extract and fractions have high potential as anti-diabetic agents, since the inhibitory effect against α-amylase and α-glucosidase is much higher than that of medicinal and aromatic plants commonly reported to have antidiabetic activities. These extracts (
Table 3 and
Table 4) show higher or similar antidiabetic activity than
Thymus pulegioides L. aqueous or hydroethanolic extracts (10% inhibition of α-glucosidase and none effect at α-amylase, at 0.5 mg/mL extracts) [
28], or orange thyme extracts that at 0.5 mg/mL inhibited 10% of α-amylase and 12% of α-glucosidase activity [
32].
Concerning
Cistus spp., hydromethanolic extracts obtained from
Cistus salviifolius L. and
Cistus monspeliensis L. aerial parts showed very high α-glucosidase (IC
50 ~0.01 mg/mL, for both plants) activity and high α-amylase (IC
50 ~0.6 and 0.7 mg/mL, for
C. salviifolius and
C. monspeliensis, respectively) activity [
22]. Aqueous infusion extracts of
C. ladanifer were reported to inhibit α-amylase (IC
50 ~1.3 mg/mL); however, these inhibitory assays [
33] were performed at different temperature conditions from the ones reported here. These results indicate that labdanum resin, in particular the diterpenoid fraction, has compounds with α-amylase activity (
Table 3), while the flavonoid fraction has compounds with higher activity against α-glucosidase (
Table 4).
AChE inhibition activity of labdanum absolutes and purified fractions is shown in
Table 5. Adl labdanum absolute showed the highest inhibition effect (~70% and 75% inhibitory activity, at 0.5 and 1 mg/mL, respectively), but Zam labdanum absolute only produced ~22% and 7% AChE inhibitory effect at 1 and 0.5 mg/mL, respectively. Adl labdanum fractions showed moderate AChE inhibitory activities, indicating that compounds present in both fractions have AChE target ability. Around 10% (dw/dw) of the Adl labdanum absolute was lost during the purification of diterpenoid and flavonoid fractions. GC-EI-MS analysis revealed that there are compounds present in significant amounts in Adl labdanum absolute, which are not present in both purified fractions (between peak 13, at 26.48 min, and peak 31, at 39.99 min,
Table 2); this may explain the higher activity of the absolute comparing to the sum of fractions activity. Several plant extracts and plant-derived compounds have been shown to possess AChE inhibition activity, mostly alkaloid compounds [
34], and phenolic rich extracts, such as the aqueous and hydroetanolic extracts of
T. pulegioides that, at 0.5 mg/mL, inhibited ~82% and 89% AChE, respectively [
28]. Concerning Cistus spp., essential oils of five species (
Cistus creticus,
Cistus salviifolius,
Cistus libanotis,
Cistus monspeliensis and
Cistus villosus) were reported to have AChE inhibitory activity with wide range of activity (at 0.5 mg/mL ~12% and ~90% inhibition for
C. creticus and
C. libanotis, respectively) [
35]. In this work, we report, for the first time, the inhibition of AChE by a
C. ladanifer extract, the labdanum resin (absolute and fractions) and, thus, its neuroprotective potential.
Furthermore, since alkaloids are not reported for this species, this activity may be associated with compounds belonging to the flavonoid or terpenoid classes. In fact, compounds belonging to these classes of phytochemicals have shown promising potential as AChE inhibitors [
36,
37]. Research on the inhibitors of AChE is important for the development of drugs with neuroprotective activity, since decreased (or altered) cholinergic function is observed in many neurodegenerative disorders, such as Alzheimer’s disease, senile dementia, Parkinson’s disease, and others [
25]. Rivastigmine (based on physostigmine chemical structure, from
Physostigma venenosum) and galanthamine (an alkaloid from
Galanthus woronowii) are among the most prescribed cholinergic enhancers, acting as AChE inhibitors, and are derived from natural sources [
38]. Finding new AChE inhibitors is, thus, essential for the development of new pharmaceutical drugs. It is also relevant, but to a lesser extent, for the development of useful toxins for agriculture, principally those selectively toxic for arthropods and not to vertebrates [
25,
39].
The anti-AChE activity has also been previously described for some of the compounds identified in
Table 2, or for similar compounds. Octadecanoic acid, whose methyl ester and hydropropyl ester derivatives were identified in Adl and/or Zam extracts, was shown to inhibit AChE activity [
40]. Regarding flavonoid derivatives, Jurčević Šangut et al. [
27] reported the anti-AChE activity of apigenin, acacetin and genkwanin, all identified in Flv fraction (
Table 2). In addition, hydroxycinnamic acids were also shown to be potential neuroprotective agents through the inhibition of AChE activity.
2.3. Anti-Proliferative/Cytotoxic Activity of Labdanum Absolutes and Fractions
Anti-proliferative or cytotoxic activity of labdanum absolutes and fractions was performed against two relevant cell lines, HepG2 (human hepatocellular carcinoma cell line) and Caco-2 (human colorectal adenocarcinoma cell line), namely if medicinal preparations are aimed for ingestion purposes. Caco-2 cells represent the absorptive enterocytes, and HepG2 cells the first pass hepatocytes. Caco-2 and HepG2 cells were exposed to different concentrations of both labdanum absolutes (Adl and Zam absolutes) and of Adl fractions (diterpenoid and flavonoid fractions), for 24 h and 48 h; the results are shown in
Figure 1, and the calculated IC
50 values (i.e., the concentration that inhibits 50% of cell viability/proliferation) are shown in
Table 6. Sigmoid curves of HepG2 and Caco-2 cells viability data versus extracts concentration are presented in
Figures S5 and S6, respectively. The Hill slope parameter and the R
2 of the sigmoid curves are presented in
Table S1. As shown in
Figure 1, Zam absolute induced no cytotoxicity at the concentrations tested (25–200 µg/mL), so the sigmoid regressions could not be projected. Regarding the other extracts, sigmoid regressions presented a R
2 higher than 0.9 at all conditions, considering them a good fit to extract the IC
50 parameter.
As observed in
Figure 1, Zam labdanum absolute did not show cytotoxic activity against both cell lines at the maximum tested concentration (200 µg/mL). On the other hand, Adl labdanum resin and its purified diterpenoid and flavonoid fractions presented a dose-dependent cytotoxicity against both cell lines. At 100 µg/mL, Adl labdanum resin and fractions reduced cell viability to values below 30% of control, independently of the cell line, incubation time, and extract.
Regarding the IC
50 parameter, it is evident that Caco-2 cells are more susceptible than HepG2 cells to all extracts. In general, at 24 h exposure, the diterpenoid purified fraction presented on average higher cytotoxicity against Caco-2 and HepG2 cells, as indicated by the lower IC
50 values (
Table 6). Although at 48 h incubation, the cytotoxic activity of Adl labdanum resin was on average higher (lower IC
50), against both cell lines, the cytotoxic activity was not statistically different (α > 0.05) from the flavonoid and diterpenoid for Caco-2 cells, or from the diterpenoid for HepG2 cells and fractions. The Caco-2 cell line used is commonly used as a model of the intestinal epithelial barrier, because it retains the typical properties of absorptive enterocytes [
41], and the HepG2 cell line is used as a model to predict hepatotoxic potential of substances because hepatocyte features and properties are maintained [
42]. Given the results observed in this study, labdanum resin is potentially toxic to the gastrointestinal tract if ingested and to the liver if absorbed. However, having in account that these are cancer cell lines, this cytotoxic/anti-proliferative activity could be regarded as a cytostatic action and, thus, these extracts and their individual compounds (e.g., methylated flavonoid aglycones and labdane-type diterpenoids) are worth being studied in the future as anti-cancer agents.
To the best of our knowledge, this is the first work reporting the anti-proliferative activity of
Cistus ladanifer labdanum resin absolute and of its purified fractions. However, Barrajón-Catalán et al. [
43] reported the cytotoxicity effect of a hot-water extract from the milled shoots of
C. ladanifer against several human cancer cell lines, showing IC
50 values between 0.49 and 16.10 mg/mL, much higher than those obtained in this study (which were lower than 0.1 mg/mL). These differences might be related to the extract composition, as hot-water extracts are rich in polyphenols, flavonoids and tannins [
43], and labdanum (
Table 2) is also rich in diterpenoids, which enhance cytotoxicity [
44]. Gaweł-Bęben et al. [
45] reported that
C. ladanifer aerial parts extracted with methanol (at different proportions with water) rendered extracts with cytotoxic activity against A375 (human malignant melanoma), with IC
50 values between 100 and 200 µg/mL, depending on the extract, but were ineffective against human squamous cell carcinoma (ACC-15) cells; IC
50 > 500 µg/mL, higher values than the results shown in this study.
Cistus ladanifer extracts (acetone/water and ethanol) only slightly affected skin fibroblast viability [
46]. Extracts from other
Cistus spp. were reported to have anti-proliferative activity, as is the case of the ethanol extract of
Cistus creticus subsp.
creticus L. shoots that was cytotoxic on cervix carcinoma (HeLa), breast apocrine carcinoma (MDA-MB-453) and melanoma (FemX) cancer cells, with IC
50 reaching 80.83 μg/mL, 76.18 μg/mL, and 87.52 μg/mL, respectively [
47], values that are identical to those obtained in HepG2 cells at 24 h exposure (
Table 6).
Cistus incanus L. extracts, obtained with different organic solvents, rich in polyphenols were tested against human colon adenocarcinoma (HCT116) cells, showing IC
50 values between ~140 μg/mL and 400 μg/mL, depending on the extract [
48] that, although not so effective as the extracts obtained here (
Table 6), corroborates the anti-proliferative activity against colon cancer. This comparison with the literature suggests that Adl labdanum absolute, but not Zam labdanum absolute, has higher cytotoxic activity than
C. ladanifer whole plant extracts (water or organic solvent extracts). It is worth emphasizing that, considering the IC
50 values obtained here, and according to NCI guidelines, Adl extracts and fractions are considered moderately active (IC
50 0.02–0.20 mg/mL) as anti-cancer agents.
Therefore, in this study, we provide a first assessment of labdanum resin bioactivities using in vitro enzymatic and cell-based assays. While the resin and its fraction present anti-diabetic, neuroprotective and anti-proliferative potential, these results should be confirmed in a broader set of cell models, as well as in more complex experimental models, such as 3D cultures and in vivo models. These would allow us to confirm both their pharmacological potential and their safety profile.