LC-ESI-MS/MS-Based Comparative Metabolomic Study, Antioxidant and Antidiabetic Activities of Three Lobelia Species: Molecular Modeling and ADMET Study

Abstract

The hydroethanol (70%) extracts of three Lobelia species (L. nicotianifolia, L. sessilifolia, and L. chinensis) were analyzed using LC-ESI-MS/MS. Forty-five metabolites were identified, including different flavonoids, coumarin, polyacetylenes, and alkaloids, which were the most abundant class. By applying Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) based on LC-ESI-MS/MS analysis, the three species were completely segregated from each other. In addition, the three Lobelia extracts were tested for their antioxidant activities using a DPPH assay and as antidiabetic agents against α-glycosidase and α-amylase enzymes. L. chinensis extract demonstrated significant antioxidant activity with an IC50 value of 1.111 mg/mL, while L. nicotianifolia showed mild suppressing activity on the α-glycosidase activity with an IC50 value of 270.8 μg/mL. A molecular simulation study was performed on the main compounds to predict their potential antidiabetic activity and pharmacokinetic properties. The molecular docking results confirmed the α-glycosidase inhibitory activity of the tested compounds, as seen in their binding mode to the key amino acid residues at the binding site compared to that of the standard drug acarbose. Furthermore, the predictive ADMET results revealed good pharmacokinetic properties of almost all of the tested compounds. The biological evaluation results demonstrated the promising activity of the tested compounds, aligned with the in silico results.

Introduction

Lobelia is the most prevalent of the 29 genera in the Lobelioideae family, with a nearly worldwide distribution, primarily in tropical to warm climate zones. Asia contains approximately 12% of all Lobelia species, with 27 of them occurring only in East Asia.¹ Natural products have conventionally shown potential as drug leads for the medicinal, nutraceutical, and cosmetics industries to find health-promoting medicines.²⁻⁵ Traditional medicine has a long history of using numerous Lobelia species to treat various diseases, including asthma, pneumonia, bronchitis, and cough. Among them are L. inflata, L. cardinalis, L. nicotianifolia, L. chinensis, L. siphiliticaL., L. laxiflora, and L. pyramidalis Wall.⁶ The phytochemical studies of these species resulted in the discovery of numerous new bioactive compounds, such as alkaloids, polyacetylenes, flavonoids, phenolics, terpenoids, and coumarins.^7,8 Piperidine alkaloids, such as lobeline, lobelanidine, nor-lobeline, and nor-lobelanine, are their main constituents.⁹Lobelia is a complex genus in East Asia, making it difficult to distinguish species and infraspecific taxa due to their physical similarities. The distinguishing characteristics are limited to a few morphological qualities that are typically ineffective for species recognition, such as changes in plant color and size.¹⁰ Macroscopic resemblance resulted in a significant degree of misidentification among closely linked taxa.¹ However, there has been no research on metabolites’ identification or characterization.

Diabetes mellitus is a type of endocrine illness characterized by an increase in the blood glucose level. Most used medications are expensive or have unfavorable side effects. Therefore, alternative medicine has gained popularity in recent years.¹¹ Many traditional plants have been shown to be valuable in treating hyperglycemia, and some of these plants were studied, and their active components were identified.¹²

The present study employed the LC-ESI-MS/MS method to profile the chemical constituents of the alcoholic extracts of the aerial parts of three Lobelia species, namely, L. nicotianifolia, L. sessilifolia, and L. chinensis. A comparative chemical profiling was conducted by applying chemometric analysis (PCA and HCA) to distinguish between different Lobelia species and provide evidence for using the whole metabolic profile as the main key for their classification. In addition, the antioxidant and antidiabetic activities of the three species were explored. Furthermore, a molecular simulation study was performed on the promising compounds to predict their antidiabetic activity together with their pharmacokinetic properties.

Results

LC-ESI-MS/MS Profiling of Lobelia Species

Identification of Lobelia species compounds via high-performance liquid chromatography–mass spectrometry (LC-MS/MS) has long been accepted as an authentication tool and standard analysis for the identification of plant metabolites and the profiling of various active plant compounds. Furthermore, LC-MS/MS is important in the quantification of various constituents; chemical confirmation and identification rely on both retention time and mass fragmentation.¹³

Forty-five compounds were quantified and tentatively identified in Table 1 and were arranged according to their chemical class. The identified metabolites were classified into 39 alkaloids, 4 flavonoids, 1 coumarin, and 1 polyacetylene. The majority of the detected compounds were found in L. nicotianifolia (32 compounds), followed by L. chinensis (27 compounds), and finally L. sessilifolia (21 compounds). The percentage identification ranged from 85.1 to 91.16% in the ESI-positive mode. L. chinensis was rich in lobechidine A 7.45%, radicamine A 10.96%, linarin 7.17%, and diosmin 18.36%. These results agreed with previously reported data.^14,15 While for L. nicotianifolia, the abundant compounds were lobeline 12.62%, 8-methyl-10-phenyl-dehydrolobelionol 11.99%, nor-lelobanidine 11.27%, and no flavonoids were detected. However, in L. sessilifolia, the major identified compounds were lobinaline 8.02%, 8,10-dietheyllobelionol 10.66%, radicamine A 10.24%, and diosmin 17.00%.

Table 1. LC-ESI-MS/MS Profiling of Three Lobelia Species (n = 3).

		L. nicotianifolia	L. chinensis	L. sessilifolia
class	compound	relative abundance (%)			[M + H]+	MS/MS	reference
alkaloids	lobinaline	2.80 ± 0.25		8.02 ± 0.62	387	186, 200, 201	(17)
	lobechidine A		7.45 ± 0.95		258	58, 94, 96, 150, 168, 184, 240	(17)
	nor-lobelanidine	1.51 ± 0.11			327	82, 143, 185, 186	(16)
	8,10-diethyl-hydroxylobelionol	0.79 ± 0.05		2.59 ± 0.01	258	96, 150, 168, 240	(16)
	lobechidine C		2.91 ± 0.01		228	96, 98, 138, 152, 156, 170, 210	(17)
	lobinine	3.64 ± 0.54			288	94, 96, 166, 198, 216	(16)
	8-methyl-10-ethyllobelidione	3.26 ± 0.06			226	58, 96, 154, 168	(16)
	8,10-dietheyllobelionol		6.65 ± 0.45	10.66 ± 1.22	242	96, 98, 152, 170, 224	(17)
	8-ethyl-10-propyl-hydroxylobelionol	2.62 ± 0.02			272	254, 182, 164, 96, 94	(16)
	8,10-diethyllobelidiol		1.02 ± 0.02	2.47 ± 0.32	244	81, 98, 152, 154, 170, 226	(16)
	radicamine A		10.96 ± 1.04	10.24 ± 1.07	256	82, 84, 138, 156	(17)
	lobeline	12.62 ± 1.10			338	320, 218, 216, 200, 105, 98, 96	(16)
	8,10-diethyl-norlobelionol	1.89 ± 0.10	0.65 ± 0.03		228	156, 138, 84	(16)
	8-methyl-10-propyllobelidione	4.08 ± 0.48			240	182, 168, 96, 58	(16)
	nor-allosedamine	1.43 ± 0.03	2.73 ± 0.32	1.96 ± 0.05	206	84	(16)
	allosedamine	4.83 ± 0.10			220	98	(16)
	8-ethyl-10-phenyl-hydroxylobelionol	1.14 ± 0.18	2.78 ± 0.11	1.95 ± 0.09	306	288, 234, 216, 198, 184, 166, 114, 112, 96, 94	(16)
	isolobinine	0.35 ± 0.05			288	94, 96, 166, 198, 216	(16)
	lelobanonoline	3.90 ± 0.10	3.09 ± 0.05	3.84 ± 0.12	368	98, 202
	8-phenyllobelone	1.42 ± 0.01	0.84 ± 0.04		218	112, 98, 84, 58	(16)
	8-methyl-10-phenyl-dehydrolobelionol	11.99 ± 0.87	0.68 ± 0.03	0.79 ± 0.04	274	256, 216, 198, 152, 146, 112, 96, 94, 58	(16)
	8-methyl-10-phenyllobelionol	2.74 ± 0.02			276	258, 218, 200, 154, 98, 96	(16)
	nor-lelobanidine	11.27 ± 0.95		0.82 ± 0.02	278	260, 156, 138, 84, 82	(16)
	lelobanidine	5.65 ± 0.45	1.26 ± 0.10		292	274, 202, 170, 98	(16)
	8-ethyl-10-phenyl-dehydrolobelionol	1.06 ± 0.06	1.52 ± 0.03	1.08 ± 0.02	288	270, 216, 198, 172, 166, 164, 146, 112, 96, 94, 58,	(16)
	8-ethyl-10-phenyllobelionol	0.61 ± 0.01			290	272, 218, 200, 168, 96	(16)
	8-ethyl-10-phenyl-norlobelidione	0.50 ± 0.01			274	202, 154, 82	(16)
	lobelanine	0.51 ± 0.01			336	216, 96	(16)
	radicamine B		0.91 ± 0.04		226	58, 96, 154, 168	(17)
	8-propyl-10-phenyl-nor-lobelionol	0.93 ± 0.03	1.57 ± 0.11		290	170, 152, 84, 82	(16)
	8-propyl-10-phenyllobelionol	1.84 ± 0.04	3.69 ± −0.21	3.07 ± 0.35	304	96, 182, 200, 218, 286	(16)
	nor-lobelanine	0.56 ± 0.01		1.04 ± 0.11	322	82, 202	(16)
	nor-lobeline (isolobelanine)	0.75 ± 0.05	1.07 ± 0.06	1.03 ± 0.07	324	204, 202, 186, 143, 82	(16)
	lobechinenoid		0.69 ± 0.04		508	137, 151, 180, 265, 285, 297, 311, 315, 323, 341, 478	(17)
	3-hydroxy-3-phenylpropionic acid (nor-allosedaminate)	0.46 ± 0.02	1.12 ± 0.03		354	188, 84	(16)
	lobelidine	2.31 ± 0.01		6.31 ± 0.65	338	218, 216, 98, 96	(16)
	8,10-diphenyl-hydroxylobelionol	1.85 ± 0.05	2.73 ± 0.13	3.09 ± 0.34	354	336, 232, 216, 214, 143, 110, 105	(16)
	3-hydroxy-3-phenylpropionic acid	0.56 ± 0.04	0.62 ±	2.23 ± 0.13	384	218, 112, 98	(16)
	lobelanidine	1.29 ± 0.05		3.16 ± 0.37	340	322, 218, 202, 98	(16)
flavonoid	Llinarin		7.17 ± 0.17		593	285, 447	(17)
	diosmin		18.36 ± 1.12	17.00 ± 0.98	609	301,463	(17)
	diosmetin		5.65 ± 0.22		301	153, 229, 258, 286	(17)
	3′-methoxyl-linarin		1.56 ± 0.06	2.56 ± 0.05	623	315, 477	(17)
polyacetylenes	lobetyolinin		1.17 ± 0.04	1.19 ± 0.17	559	155, 199, 217, 397	(17)
coumarin	tomentin		1.19 ± 0.09		443	96, 216, 336	(20)
total no. of identified compounds		32	27	21

Alkaloids and their derivatives were the major class of phytochemicals discovered using LC-ESI-MS/MS, which is consistent with the majority of the reported data on other Lobelia extracts.^16,17 Alkaloids were reported in most Lobelia species at 46.05% being of piperidine ring alkaloids with lobeline as the most reported in all Lobelia species.⁹

Wang et al. reported that the major bioactive compounds of L. chinensis were piperidine alkaloids along with flavonoids, lignans, coumarins, terpenes, and polyacetylenes. They also stated that lobeline was absent in L. chinensis and it should not be used as a marker in verification or qualitative analysis.¹⁸ Zimare et al. reported 18 phenolic compounds from L. nicotianifolia, which displayed a resemblance to various Lobelia species.¹⁹

LC-ESI-MS/MS-Based Chemometric Analysis

Both principal component analysis (PCA) and hierarchal cluster analysis (HCA) were used as chemometric tools to explore their abilities to discriminate between various Lobelia species, in addition to identifying any substantial correlation between them.²¹ A matrix of the total number of samples and their replicates (9 samples) multiplied by 45 variables (LC/MS peak area %) was constructed in MS Excel and then subjected to chemometric analysis (PCA and HCA). Owing to the diversity of variables qualitatively and quantitatively, PCA was applied as an initial phase to reduce the dimensionality of the multiple data sets in addition to removing the redundancy in the variables, utilizing raw data (peak area% for each compound as in Table 1. Figure 1a represents the PCA score based on the LC-MS metabolic profiles, revealing a significant statistical segregation among various species.

Figure 1.

PCA score plot (a), biplot (b) based on LC-MS metabolic profiles of various Lobelia species based on the identified compounds displayed in Table 1.

The PCA score plot (Figure 1a) explained about 99% of the data set variation by the first two PCs, where PC1 accounted for 63% and PC2 for 36% of the variance. Various Lobelia species were segregated into three main groups in three single quadrants. Both L. sessilifolia and L. chinensis were located on the positive score values (right side of PC1) in the upper and lower quadrants, respectively. However, L. nicotianifolia was positioned on a negative score plot (the negative side of both PC1 and PC2). Figure 1b displays the biplot for both scores (t) and loading (p); the plot enabled the visualization of closeness and dissimilarities between various species in accordance with their metabolic profiles. The species positioned near different metabolites are patterned in the score plot on the basis of these metabolites. The biplot displayed that diosmetin, 8-ethyl-10-phenyl-dehydrolobelionol, and nor-allosedamine were the key metabolites responsible for the discrimination of L. chinensis; however, for L. sessilifolia species, 8,10-diethyllobelidiol and 3-hydroxy-3-phenylpropionic acid hydroxyallosedaminate were the main markers accountable for its segregation in a separate quadrant.

Regarding L. nicotianifolia, nor-lobelanidine, 8-methyl-10-phenyl-dehydrolobelionol, lelobanidine, and 8-ethyl-10-phenyl-nor-lobelidione were the leading compounds contributing to its discrimination. It was observed that the major compounds that exist in the highest concentrations in each species are not the chief markers responsible for the overall pattern in the PCA score plot. However, alkaloid profiles can provide diagnostic markers for the taxonomy of different Lobelia species. Therefore, the biplot confirmed the deep reputation of the whole metabolic profile in discrimination between various species and not merely the compounds existing in high concentration.

Furthermore, HCA was employed as an unsupervised pattern recognition method to endorse the results of PCA. Figure 2 displays the HCA dendrogram, confirming the segregation of various Lobelia species into three main clusters. Clusters I, II, and III presented L. nicotianifolia, L. chinensis, and L. sessilifolia, respectively. The HCA dendrogram revealed the closeness of L. chinensis and L. sessilifolia. The HCA results confirmed those of PCA.

Figure 2.

HCA dendrogram based on LC-MS metabolic profiles of various Lobelia species based on the identified compounds displayed in Table 1.

In this study, although Lobelia species are closely related regarding their taxonomical classification, PCA and HCA were applied successfully as powerful chemometrics tools to discriminate them based on their full chemical profiles. On conclusion, chemotaxonomic classification of different Lobelia species was achieved by clear metabolomic discrimination.

Antioxidant Activity

The antioxidant activities of the ethanol extracts of L. chinensis, L. sessilifolia, and L. nicotianifolia were evaluated in a dose-dependent manner, as shown in Figure 3. Among these extracts, the total ethanol extract of L. chinensis exhibited the highest antioxidant activity, as indicated by its lower IC₅₀ value of 1.111 mg/mL. In comparison, the IC₅₀ values for L. sessilifolia and L. nicotianifolia total ethanol extracts were found to be 1.242 and 1.615 mg/mL, respectively. The literature review for the antioxidant activity of Lobelia plants was due to their variety of flavonoid content.⁹ The methanol extract of L. chinensis reported higher antioxidant activity than the aqueous extract.²² However, polysaccharides isolated from L. chinensis reported antioxidant and anticancer activities.²³L. nicotianifolia methanol extract of the leaves reported higher antioxidant activity than roots methanol extract.²⁴

Figure 3.

DPPH inhibition percentage of different Lobelia species in a dose-dependent manner. Ls: L. sessilifolia, Ln: L. nicotianifolia, and Lc: L. chinensis.

Cui et al. examined the antioxidant activity of two L. chinensis extracts, stating that the antioxidant activity of both extracts was associated with the type and polarity of the extraction solvent. Demonstrating that the ethanol extract has higher activity than the water extract. It has been reported that antioxidant activity of natural products is attributed to synergism of their active constituents.²⁵ Hence, flavonoids, phenolic compounds, and tannins have antioxidant activity and can eliminate free radicals.²⁶⁻²⁸

The evident presence of flavonoids in both L. chinensis and L. sessilifolia has played an essential role in their antioxidant activity.²⁹

Zimare et al. reported that L. nicotianifolia leaf extracts possess good antibacterial, antioxidant, and anticancer properties.¹⁹ Diosmin, diosmetin, and linarin have been shown to have anti-inflammatory^30,31 and antioxidative^32,33 properties. Furthermore, the polyacetylene lobetyolinin has been reported to have antioxidant activity.³⁴

Both L. chinensis and L. sessilifolia alcoholic extracts were more potent as antioxidants than the L. nicotianifolia alcoholic extract. The superior antioxidant behavior observed in the extracts could be attributed to the presence of diosmin and diosmetin.²⁹ The ethanol extract of L. chinensis contained the highest percentage of diosmin (18.36 ± 1.12%), followed by L. sessilifolia (17.00 ± 0.98%), while it was absent in the L. nicotianifolia ethanol extract. On the other hand, diosmetin was exclusively detected in the L. chinensis extract with a content of 5.65 ± 0.22%. This finding suggests a correlation between the presence of flavonoids and the potent antioxidant activity of the L. chinensis ethanol extract compared to the other extracts. Despite the significant antioxidant activity observed in the ethanolic extracts of the three plants, it is worth noting that their activity was relatively weak when compared to that of a reference drug, Trolox, which exhibited a much lower IC₅₀ value of 0.009893 mg/mL, indicating its substantially stronger antioxidant potential in comparison to the plant extracts.

Antidiabetic Activity

The ethanol extracts from L. chinensis, L. sessilifolia, and L. nicotianifolia were tested for their possible ability to lower blood sugar. Specifically, the inhibitory effects of these extracts were studied on α-glucosidase and α-amylase enzymes. The results revealed notable differences among the extracts in terms of their inhibitory activities, in which α-glucosidase inhibitory activity may be explained by the presence of flavonoid glycosides or alkaloids as reported by Astiti et al.³⁵

The ethanol extract of L. nicotianifolia exhibited the highest α-glucosidase inhibitory activity, with an IC₅₀ value of 270.8 μg/mL. Following this, the L. sessilifolia extract demonstrated an IC₅₀ value of 364.5 μg/mL. However, the L. chinensis extract displayed no significant α-glucosidase inhibitory activity, as its IC₅₀ value exceeded 1000 μg/mL. In comparison, the reference drug acarbose exhibited an IC₅₀ value of 121.3 μg/mL, as shown in Figure 4.

Figure 4.

α-Glucosidase inhibitory activity on a dose-dependent manner. Ls: L. sessilifolia, Ln: L. nicotianifolia, and Acarbose: reference drug.

L. nicotianifolia ethanol extract showed mild activity against the α-glucosidase enzyme, followed by L. sessilifolia. This activity could be attributed to radicamine A, which suppresses the α-glucosidase enzyme because of its aromatic ring and demonstrates antidiabetic activity³⁶⁻³⁸ and lobeline.^39,15 Lobeline was found to be present in the ethanol extract of L. nicotianifolia at a concentration of 12.62 ± 1.10%, while it was absent in the other two extracts. This finding suggests a potential correlation between the presence of lobeline and the α-glucosidase inhibitory activity observed in the L. nicotianifolia extract, distinguishing it from the other extracts.

In contrast, all of the tested extracts failed to exhibit significant α-amylase inhibitory activity, as their IC₅₀ values exceeded 500 μg/mL. In comparison, the reference drug acarbose demonstrated an IC₅₀ value of 1.648 μg/mL, indicating its notable α-amylase inhibitory potential. These findings highlight the differential antidiabetic activities of the three ethanol extracts, with L. nicotianifolia showing the most promising α-glucosidase inhibitory activity, potentially attributed to the presence of lobeline. However, it is important to note that none of the extracts demonstrated significant α-amylase inhibitory activity compared to that of the reference drug, acarbose. Further investigations are warranted to elucidate the mechanisms underlying the observed activities and to explore the potential of these extracts for the development of antidiabetic agents.

In Silico Study

2.4.1. Molecular Docking Study

A molecular docking study was applied on the nine promising antidiabetic compounds named: radicamine A, 8-methyl-10-ethyllobelionol, 8,10-diethyl-lobelionol, nor-lelobanidine, linarin, diosmin, lobechidine A, lobinaline, and lobeline. Docking results using the C-DOCKER algorithm were used to investigate the binding mode of the tested compounds at the binding site of the human lysosomal acid α-glucosidase (PDB ID: 5NN8).⁴⁰ The docking interaction results of most of the newly tested ligands and the downloaded ligand acarbose revealed the main reported binding interactions, from the previous literature with the essential amino acid residues at the active site: Trp376, Tyr378, Leu405, Trp481, Asp518, Met519, Phe525, Asp616, Trp618, Phe649, Leu650, His674, and Leu678.^41,42

The (−) C-DOCKER interaction energy (E) results of the docked compounds showed a range of 36.33–54.82 kcal/mol and were compared to those of the acarbose ligand (E = 78.78 kcal/mol). The 2D diagram of the binding interaction of the compounds with the essential amino acids in the active site was studied (Figure 5).

Figure 5.

I: 2D docking diagram showing the binding interactions with α glucosidase (PDB ID: 5NN8): (A) acarbose (E = −78.78 kcal/mol), (B) diosmin (E = −54.82 kcal/mol), (C) lobeline (E = −51.30 kcal/mol), (D) linarin (E = −48.51 kcal/mol), (E) lobechidine A (E = −45.31 kcal/mol), and (F) radicamine A (E = −42.63 kcal/mol). (Color code—green dotted lines: hydrogen bond, orange dotted lines: attractive charge, pink dotted lines: Pi-alkyl bonds, and purple dotted lines: Pi alkyl). II: 2D diagram showing the binding interactions on α glucosidase (PDB ID: 5NN8): (A) 8,10-diethyl-lobelionol (E = −40.89 kcal/mol), (B) nor-lelobanidine (E = −37.42 kcal/mol), (C) 8-methyl-10-ethyllobelionol (E = −37.34 kcal/mol), and (D) lobinaline (E = −36.33 kcal/mol). (Color code—green dotted lines: hydrogen bond, orange dotted lines: attractive charge, pink dotted lines: Pi-alkyl bonds, and purple dotted lines: Pi alkyl).

The 3D overlay of the binding interactions of ligand acarbose with the key amino acid residues at the active site with each of lobeline and radicamine A is demonstrated in Figure 6, showing comparable interactions and hence suggesting a promising antidiabetic effect via α-glucosidase inhibitory activity.

Figure 6.

Acarbose ligand (purple) overlay with (A) lobeline (cyan), (B) radicamine A (blue) with amino acids: Asp282 green, Leu283 blue, Asp518 faint pink, Met519 grayish blue, Arg600 blue, Asp616 red, leu650 and His674 hot pink, visualized by Paymol.

Predictive ADMET Study

An ADMET study was conducted via Discovery Studio 4.0 Software on the nine tested compounds: radicamine A, 8-methyl-10-ethyllobelionol, 8,10-diethyl-lobelionol, nor-lelobanidine, linarin, diosmin, lobechidine A, lobinaline, and lobeline.

This study was used to evaluate the pharmacokinetic profile and investigate the druglikeness of the tested compounds. Investigations were based mainly on the given chemical structure of the molecule and were correlated to several calculated parameters, including aqueous solubility level, atom-based Log P98 (A LogP 98), blood brain barrier level, cytochrome P450 2D6 (CYP2D6), hepatotoxicity probability, absorption level, plasma protein binding level (PPB Level), and 2D polar surface area (ADMET 2D PSA) (Figure 7).⁴³

Figure 7.

ADMET plot of the tested compounds: calculated PSA_2D versus A log P98 properties.

The results demonstrated in Table 2 showed that all the compounds exert BBB penetration levels between 3 and 4, which indicates that they could not pass BBB, except lobinaline, which showed high BBB penetration. Most of the compounds had an absorption level range of (0–1), indicating a good-to-moderate level of human intestinal absorption. Also, most of the compounds showed an ADME aqueous solubility level values range in the range (3–5), which confirms their good aqueous solubility. While the Log P parameter is used to determine the lipophilicity of the compound, it is evaluated together with the polar surface area (PSA) parameter, as compounds not exceeding the value of 5 usually attain a probability of well absorption. Also, the (PSA) property is an important key for describing the drug’s bioavailability. Passive absorption is detected for molecules with PSA <140 and is known to be of lower bioavailability. Hereby, most of the compounds were predicted to show good oral absorption, including linarin and diosmin, with good bioavailability results in the PSA range of 216.589 and 237.405. On the other side, compounds 8,10-diethyl-lobelionol, nor-lelobanidine, lobechidine A, lobinaline, and lobeline showed no hepatotoxicity. Also, most of the compounds are determined to be cytochrome P450 2D6 (CYP2D6) noninhibitors, except nor-lelobanidine, lobinaline, and lobeline. Only lobeline showed a probability of plasma protein binding (Table 2).

Table 2. In Silico ADMET Prediction Results.

name	ADMET Sol level	ADMET A LogP 98	ADMET BBB level	CYP2D6 prediction (non-inhibitor)	hepatotoxic probability	absorption level	PPB binding prediction	A LogP 98	PSA 2D
1-radicamine A	5 (high)	0	4	FALSE	TRUE	1 (moderate)	FALSE	–1.807	92.192
2-8-methyl-10-ethyllobelionol	5 (high)	0	3	FALSE	TRUE	0 (good)	FALSE	–0.465	44.834
3-8,10-diethyl-lobelionol	5 (high)	0	3	FALSE	FALSE (non toxic)	1 (moderate)	FALSE	–0.536	41.32
4-nor-lelobanidine	4 (optimal)	0	3	TRUE	FALSE (non toxic)	0 (good)	FALSE	0.769	44.834
5-linarin	3 (good)	0	4	FALSE	TRUE	3	FALSE	–0.153	216.589
6-diosmin	2 (low)	0	4	FALSE	TRUE	3	FALSE	–0.395	237.405
7-lobechidine A	5 (high)	0	3	FALSE	FALSE (non toxic)	0 (good)	FALSE	–0.83	62.135
8-lobinaline	2(low)	0	0 (high penetration)	TRUE	FALSE (non toxic)	0 (good)	FALSE	3.806	14.527
9-lobeline	3 (good)	0	2	TRUE	FALSE (non toxic)	0 (good)	TRUE	2.373	41.32

Conclusions

The metabolomic profiling of the tested species belonging to the genus Lobelia has shown their richness in alkaloidal secondary metabolites, which play an important role in their activities. Although lobeline was found only in L. nicotianifolia extract, PCA proved the significant importance of the whole metabolic profile in the segregation between different Lobelia species. The total ethanol extract of L. chinensis showed the highest antioxidant activity and no activity against the α-glucosidase enzyme, in contracts with L. nicotianifolia. Thus, their use should be based on their activity or active constituents. In addition, molecular docking study results suggested the antidiabetic α-glucosidase inhibitory activities of most of the compounds, including mainly lobeline and radicamine A. In silico ADMET results predicted their good pharmacokinetic profile too. To further support the profiling method of the selected species, we may carry out the isolation of the major components responsible for these biological activities in the future.

Methods

Plant Material

The seeds of the three Lobelia species, namely, L. nicotianifolia, L. sessilifolia, and L. chinensis, were collected from Mercara hill in area Western Ghat regions of Karnataka, India. The seeds were planted in March in the medicinal farm of Mepaco-Medifood and were collected during Fall. Plant material was authenticated by Dr. Labib, T. consultant of plant taxonomy, exdirector of Al-Orman botanical Garden, Giza, Egypt.

Preparation of Plant Extracts

100 g of L. nicotianifolia, L. sessilifolia, and L. chinensis aerial parts were powdered, macerated 3 times each with 200 mL of 70% hydro alcohol for 1 h in a shaker. The total hydroalcoholic extracts were collected and evaporated via a rotatory evaporator at 40 °C, yielding green-black extract of 0.183, 0.533, and 0.628 g of L. nicotianifolia, L. sessilifolia, and L. chinensis, respectively,

Identification of Metabolites in Hydroalcoholic Extracts Using LC-ESI-MS/MS

The hydroalcoholic extracts of three plant extracts were dissolved in a methanol HPLC grade using a membrane disc filter 0.2 μm for filtration. Ten μL of the sample was injected into an RP C-18 column (5 μm, 125 mm × 4 mm). Gradient mobile phase methanol and water acidified with 0.1% formic acid were employed at a 0.2 mL/min flow rate. The total program time is 35 min. Mass spectra were detected in the ESI negative and positive ion modes. Using the molecular weight, the fragmentation pattern of the mass spectrum, and comparison with previously published data, compounds were tentatively identified.⁴⁴

Chemometric Analysis

The obtained information from LC-MS was subjected to chemometric analysis. Principal Component Analysis (PCA) was first utilized as a primary step for interpretation of the data to portray an outline for all species discrepancies and to identify markers accountable for this difference.^45,46 Hierarchal cluster analysis (HCA) was then employed to allow clustering of different species. The clustering pattern was constructed by a single linkage method. PCA and HCA were achieved utilizing the SIMCA-P version 13.0 software package (Umetrics, Umeå, Sweden).

Antioxidant Activity

The antioxidant activity of various Lobelia was assessed using the (DPPH) 2,2-diphenyl-1-picrylhydrazyl assay. Initially, stock solutions of the samples were 50 mg/mL in DMSO. Then, the samples were further diluted in methanol to obtain a final concentration of 5000 μg/mL. Subsequently, additional dilutions were prepared at concentrations of 5000, 2500, 1250, 625, and 312.5 μg/mL in methanol. A Trolox stock solution with a concentration of 20 μg/mL was prepared in methanol for comparison. From this stock solution, five concentrations were prepared: 12.5, 7.5, 6.25, 2.5, and 1.25 μg/mL. To initiate the DPPH assay, 100 μL of freshly prepared DPPH reagent (0.1% in methanol) was added to 100 μL of each sample in a 96-well plate. The plate was then incubated at room temperature for 30 min in the dark. Following incubation, the reduction in DPPH, indicative of antioxidant activity, was measured by assessing the color intensity at 540 nm using a microplate reader (FluoStar Omega). The recorded data were analyzed using Microsoft Excel, and the IC₅₀ value, representing the concentration required to inhibit 50% of the DPPH activity, was calculated using GraphPad Prism 9 software. This was achieved by converting the concentrations to their logarithmic values and fitting them to a nonlinear regression equation known as the “log (inhibitor) vs normalized response–variable slope equation”^47,48

Data are represented as means ± SD according to the following equation:

Antidiabetic Activity

The antidiabetic activity of different Lobelia species using α-glucosidase and α-amylase inhibitory reported assays was explored.⁴⁹ To conduct the experiments, a stock solution of acarbose was prepared at a concentration of 2 mM in phosphate buffer (100 mM, pH = 7). Final concentrations of 1000, 500, 250, 125, and 62.5 μM (equivalent to 645, 322.5, 161.25, 80.625, and 40.31 μg/mL) were prepared in water. The samples were dissolved in DMSO at a final concentration of 50 mg/mL. Dilutions were then prepared in methanol to obtain the following final concentrations for the α-glucosidase assay: 400, 200, 100, 50, and 25 μg/mL; and for the α-amylase assay: 500, 250, 125, 62.5, and 31.25 μg/mL. Enzymatic activity was assessed by measuring the release of substrates at 405 nm using a microplate reader from Onega (USA). Overall, the study employed α-glucosidase and α-amylase inhibitory assays to investigate the antidiabetic activity of the samples. The enzymatic activity was determined by measuring the release of substrates at 405 nm using a microplate reader. The percentage of inhibition of α-glucosidase and α-amaylase was calculated according to the equation:

In Silico Study

Molecular Docking Study

A molecular docking study was applied on the promising antidiabetic compounds named: radicamine A, 8-methyl-10-ethyllobelionol, 8,10-diethyl-lobelionol, nor-lelobanidine, linarin, diosmin, lobechidine A, lobinaline, and lobeline. The crystal structure of human lysosomal acid-alpha glucosidase in complex with acarbose (PDB ID: 5NN8) was downloaded from Protein Data Bank.⁴⁰ Clean protein was successfully done, and hydrogen atoms were added to complete any missing amino acid residues. The unneeded water molecules were also removed. Force field was applied with CHARMm and partial charge MMFF94. Protein was prepared and then minimized. Also, the active site was defined. Acarbose ligand was removed before the docking step. The new ligands were first prepared, and then docking was performed using the C-DOCKER docking algorithm.

Predictive ADMET Study

In silico ADMET study using Discovery Studio 4.0 Software were conducted to evaluate the pharmacokinetic properties of all tested compounds.⁴³

This work was funded by King Saud University Researchers Supporting Project number (RSP-2024R294), King Saud University, Riyadh, Saudi Arabia. Therefore, all the authors acknowledge, with thanks, RSP for technical and financial support.

The authors declare no competing financial interest.