The discovery of novel immunomodulatory medicinal plants by combination of historical text reviews and immunological screening assays

Ethnopharmacological relevance: With the advent of immunotherapies against cancers, autoimmune diseases and infections, there is a steady demand for novel medicines. New sources for discovery of potentially novel immunomodulatory compounds are therefore needed. Nature contains a large and diverse reservoir of novel compounds that can be exploited for their potential as new drugs, and exploring the pharmaceutical potential of medicinal plants used in traditional medicine is highly relevant. Aim of the study: We aimed with this study to explore usage of medicinal plants in Scandinavian folk medicine against diseases interpreted to involve the immune system, and to further screen water extracts from previously overlooked medicinal plants in order to discover potential new sources of immunomodulatory compounds. Materials and methods: We systematically investigated historical records dating back to the 1800s with an emphasis on plants used as treatment for wounds or diseases interpreted to be inflammatory. Of 74 candidate plants, 23 pharmacologically under-studied species were selected for further characterization. The plants were collected from their natural habitats in Southern Norway, air-dried, and subjected to boiling water and accelerated solvent extraction. The crude extracts were separated into polysaccharide-enriched fractions and C-18 solid phase extracted fractions. Immunological screenings were performed with all extracts and fractions. Monosaccharide composition and total phenolic content were determined and compared across all species. Results: We identified 10 species with clear immune activating effects and 8 species with immune inhibitory effects by comparing cytokine production by human peripheral blood mononuclear cells, primary human T-and NK-cell proliferation, and nitric oxide production from macrophages. Conclusions: With this study, we provide a comprehensive overview of Scandinavian medicinal plants and their usage, and our findings support an approach of combining historical sources with modern pharmacology in the discovery of plant sources containing potentially new pharmacological compounds.


Introduction
As far back as we can trace human activity and medical treatment, plants have been used in folk medicine for their medicinal benefits (Conrad et al., 2008;Newman et al., 2000;Pan et al., 2014;Petrovska, 2012;Porter, 1999), and self-medication with plants has also been recorded in other animals (de Roode et al., 2013).The close relationship between people and plants, which provide oxygen, food, shelter, clothing, ornamentals, and medicine is essential for human life and culture (Teixidor-Toneu et al., 2020).People tended to use plants found in near proximity of where they lived, and these plants inspired herbal medicines.Knowledge about medicinal plants and their uses has been preserved and passed on both orally and in written in different cultures and social groups all over the world (Chen et al., 2020;Colombo and Ammirati, 2011;Laroche, 2009;Rankin, 2013).Based on this long history of medicinal plant use, it is logical to assume that nature still retains a large potential for the discovery of new pharmacologically active compounds (Cravens et al., 2019;Li, Li and Smolke, 2018;Scannell and Bosley, 2016).While current pharmaceutical research is focused mainly on development of new synthetic compounds and exploration of pre-discovered molecules, there is a steady decline in the discovery of novel compounds.In light of ever increasing demands of novel, effective immunomodulatory drugs against cancer, infections and autoimmune diseases, this calls for new efforts into the discovery of novel compounds from natural resources (Atanasov et al., 2015;Grabley and Thiericke, 1999).
About a third of modern day essential medicine derives from natural products and 25% from medicinal plants and fungi (Calixto, 2019;Newman and Cragg, 2020).Examples are morphine from the opium poppy and paclitaxel from the bark of the Pacific yew tree.A Scandinavian example is the accidental discovery of cyclosporine from the fungus Tolypocladium inflatum W. Gams.found in the mountains of Southern Norway.Cyclosporine is now on the WHO list of essential medicines.Advances within the field of modern medicine and pharmacology have made it possible to investigate, in detail, the diversity of chemical structures available from natural sources and plant compounds.These new advances make it possible to discover bioactive compounds from medicinal plants and generate valuable pharmaceutical leads.
Traditionally, medicinal plants were prepared with liquor or as infusions or decoctions, and administered orally or applied externally.In this study we have chosen to focus on water extracts where phenolics and polysaccharides are abundant constituents, and can mediate potent immunomodulatory or anti-cancer effects (González-Gallego et al., 2014;Inngjerdingen et al., 2008;Ramberg et al., 2010).Polysaccharides are complex biological macromolecules, and are a structurally diverse class of molecules.They are composed of sugar monomers linked together with glycosidic bonds, with different degrees of branching, molecular weight and structural confirmation.In the last decades, a range of studies have shown that polysaccharides have promising immune modulating and anti-tumor effects (Meijerink et al., 2018).Polysaccharides are promising candidates for novel therapeutics as they are generally non-toxic, biocompatible, and biodegradable (Yin et al., 2019).Phenolic compounds are produced in plants as secondary metabolites with protective functions against e.g.ultraviolet radiation, pathogen aggression, and oxidative stress.They are the largest class of bioactive compounds found in plants (Annunziata et al., 2020), including flavonoids, hydrolysable tannins, lignans, stilbenes, coumarins, as well as smaller phenolic compounds.Several in vivo and in vitro studies have found polyphenols to have anti-inflammatory properties in animals and humans (Yahfoufi et al., 2018).
In this study, Norwegian and Swedish primary written sources from the 19th and 20th centuries were studied to comprehensively identify medicinal plants used in Scandinavian folk medicine against immunerelated diseases, infections, or wounds in order to identify new sources of compounds with immunomodulatory activity.From the identified plants, water extracts were prepared from a selected set of previously understudied medicinal plants and screened for immunomodulatory activities.

Literature review of primary written sources from rural Norway and Sweden
To identify Scandinavian medicinal plants, we reviewed primary written sources focusing on firsthand accounts from farmers, including lay practitioners and local healers, in rural Norway and Sweden during the 19th and 20th century.These accounts give recollections of what these societies and possibly their ancestors used as medicinal plants.Divided into sections for different illnesses and their traditional treatments, or by describing symptoms, these sources give accounts of what people did when sickness struck, what type of plants and other remedies they collected and how they prepared them for use.The sources have been systematized and made available in printed collections by scholars of various academic backgrounds and these collections form the basis of this study.The main collections documenting these firsthand accounts were published by Ingjald Reichborn-Kjennerud (1865-1949), medical doctor and medical historian, Ove Arbo Høeg (1898Høeg ( -1993)), professor in botany at the University of Oslo, and Erling Christophersen (1898Christophersen ( -1994)), Norwegian botanist, geographer and diplomat (Christophersen and Hjort, 1960;Høeg, 1975Høeg, , 1985;;Reichborn-Kjennerud, 1922, 1944, 1947).These collections were supplemented with works from the Swedish authors Lars Hammarin and Kristina Frølich (Frølich and Wille, 1921;Hammarin, 2013).Collectively, the sources made available from these authors were the focus of this study.During this study, it became obvious that plants for medicinal use were inconsistently documented during this period.People used what they had available and different common/vernacular names were often given depending on where in the country the plant was used.Most of the sources include scientific names, matched to the vernacular names used by people in the countryside.We used two books published by Høeg as the main references when linking vernacular and scientific names (Høeg, 1975(Høeg, , 1985)).Høeg has done extensive work in trying to track these different vernacular names and gathering them under their standard scientific name.Accepted botanical names were confirmed by consulting the Catalogue of Life Checklist (Bánki et al., 2021;Roskov et al., 2020).The following traditional therapeutic indications were chosen for the selection of plant species: Gastrointestinal related diseases interpreted to be inflammatory, other inflammatory related diseases, infections, and inflamed wounds.These conditions were identified from vernacular diseases like gastrointestinal-related problems (aches, cramps, constipation, and diarrhea), skin wounds and rashes.

Selection of study plants
From 74 plants identified from review of historical sources (Table 1), plants with few pharmacology or chemistry search results (<50) in PubMed or SciFinder were selected for further studies in order to test immunological activities in under-studied plant species.The search term used to find publications in PubMed was the scientific plant name + "Pharmacology".In SciFinder, only the scientific plant name was used as a search term, and the search hits were then refined by the category heading "Biology".For the search of chemical compounds, the search was further refined by "Substances in biology", while the search for immunological related publications were refined by the category "Immunology".SciFinder has a broader database compared to PubMed, and therefore resulted in a higher number of search hits.Thus, the publications encountered were assessed individually for each plant to determine the degree of relevance towards immunology and infections.Hypericum perforatum L. [Hypericaceae] was included amongst the short-listed plants as comparison against the pharmacologically less studied Hypericum maculatum Crantz [Hypericaceae].Juniperus communis L. [Cupressaceae], which is also well studied, was included due to its long-standing role as a Nordic medicinal plant, probably since the Viking Age (Teixidor-Toneu et al., 2021).Ulriksen et al.

Plant material and study area
Plant material was collected in Southern Norway; primarily in the Oslo region (0-100 MASL) and mountain areas in Southern Norway (800-1000 MASL) (Table A1).Collection occurred during March to September 2020, depending on the historically described collection time points for each plant, or when the plants were in flower.The collection was focused on plant parts with reported usage in the historical sources (see Table 1 and Table A1).Plant materials were dried at room temperature and powdered with a blender machine (Vitamix Acenti 2300i).Plant identities were confirmed by botanists at the Natural History Museum, University of Oslo, and herbarium specimens were collected and deposited at the Natural History Museum herbarium (voucher numbers are given in Table A1).

Accelerated solvent extraction
Powdered plant materials (ca. 10 g) were extracted with an Accelerated solvent extraction ASE 350 instrument (Dionex, Sunnyvale, CA, USA) with 100 • C dH 2 O as solvent.The water was removed by freezedrying to give the crude extracts which were divided into three equal parts.

Polysaccharide precipitation
The crude extracts were precipitated in ethanol (75%, 48 h, 4 • C), followed by centrifugation (4400 rpm, 20 min), and filtration using a Büchner funnel.The precipitate was washed with ethanol, centrifuged (4400 rpm, 5 min) and filtrated again.The precipitate was washed with acetone, dissolved in water, and freeze-dried for 48 h to give the polysaccharide-enriched fraction (PS-fractions).

Solid phase extraction (SPE)
Solid phase extraction (SPE) to obtain fractions enriched in phenolics was performed as previously reported with small modifications (Ho et al., 2017).A Phenomenex strata C-18 2 g column (Phenomenex, Macclesfield, UK) was washed with methanol and conditioned with water.The third part of the crude extract was applied to the column and eluted with water, 50% methanol and 100% methanol.The methanol fractions were combined and evaporated under reduced pressure to give the SPE-fractions.

Monosaccharide composition analysis
PS-fractions (1 mg) were dried in a desiccator 24 h prior to methanolysis, which was performed as previously described (Chambers and Clamp, 1971;Wold et al., 2018).In short, the samples were subjected to methanolysis with 3M hydrochloric acid in anhydrous methanol for 24 h at 80 • C. Trimethylsilylated derivatives of the methyl glycosides obtained after methanolysis were analyzed using a Restek RTx-5 silica column (30 m, i.d.0.25, 0.25 μm film thickness) coupled to a Focus GC (ThermoFisher Scientific, Waltham, MA) with flame-ionization detection.The parameters used for the gas chromatography were identical to previous studies (Ellefsen et al., 2021;Wold et al., 2018).Helium was used as carrier gas and Chromelion software v.6.80 (Dionex Corporation, Sunnyvale, CA) was used to analyze the results.Mannitol was used as an internal standard.

Total phenolic content
Total phenolic content was determined using Folin Ciocalteu reagent (Singleton and Rossi, 1965).The SPE-fractions were dissolved in DMSO (0.1 and 0.5 mg/ml), and 20 μL added to a 96-well plate along with dH 2 O (50 μL).Folin Ciocalteu reagent (10%, 100 μL) was added to the samples, mixed and left for 5 min.Na 2 CO 3 -solution (20%, 30 μL) was added to the wells, followed by shaking and the plate was allowed to stand for 2 h.Finally, the absorbance was measured at 765 nm using a SpectraMax 190 Microplate Reader (Molecular Devices, San Jose, CA).The values were reported as milligrams of gallic acid equivalents per grams of extract (mg GAE/g extract).

Peripheral blood mononuclear cell (PBMC) isolation
Buffy coats were obtained from the blood bank at Oslo University Hospital upon informed consent according the declaration of Helsinki, and the study was approved by the South-Eastern Norway Regional Ethical Committee.PBMCs were separated by density gradient separation on Lymphoprep (Axis-Shield, Oslo, Norway) and spinning for 20 min at 650 g with no brakes.The mononuclear cells were collected and washed in PBS with 2% FBS, and viably frozen at − 80 • C.

MTT cytotoxic assay
Toxicity of extracts was tested with the J774.A1 macrophage cell line via the MTT assay (Ferrari et al., 1990).Briefly, 100 μL (1 × 10 6 cells/ml) J774.A1 cells were plated in duplicates in 96-well flat bottom plates, and incubated overnight with samples at concentrations of 100 μg/mL, 10 μg/mL or 1 μg/mL.Medium alone served as negative control and reference value for live cells, and LPS (10 ng/mL) as positive control.Cells were next incubated with 10 μL MTT reagent (Roche Diagnostics, Germany) for 4 h at 37 • C, followed by addition of 100 μL MTT detergent solution (Roche Diagnostics).Cells were incubated in a 37 • C cell incubator overnight, then colorimetric detection was performed at OD570.Data are presented as percentage live cells relative to live cells in medium alone, calculated as the fraction of OD570 values in test samples versus medium alone (Table A.2).
Plant extracts were added to duplicate wells at 100 μg/mL, 10 μg/mL or 1 μg/mL.Medium alone served as negative control and LPS (10 ng/mL) as positive control and reference value.Cells were incubated overnight at 37 • C in a 5% CO 2 cell incubator, the supernatant was harvested, spun at 350 g for 2 min, and 50 μL supernatant was mixed with 50 μL Griess reagent A (1% sulfanilamide and 5% phosphoric acid) and incubated for 10 min at room temperature in the dark.Afterwards, 50 μL Griess reagent B (0.1% N-(1-napthyl)ethylenediamine in sterile water) was added to each sample and colorimetric detection was measured at OD 570 .The amount of NO secreted was calculated based on a NaNO standard curve, and NO-concentration of test samples were normalized against NO released from negative and positive control, and data presented as % NO release in test samples relative to LPS (Table A.2).

Enzyme-linked immunosorbent assay (ELISA)
Human TNF-α or human IFN-γ ELISA were run using a kit from Mabtech, Sweden.PBMCs were stimulated in duplicates for 16 h in 96well round-bottom plates at 1 × 10 6 cells/ml with 1, 10, or 100 μg/ml of test samples, medium alone as negative control, or 5 μg/ml ConA as positive control and reference value.Plates were coated overnight (4 • C) with capture antibodies in PBS using 96-well EIA/RIA Flat Bottom High Binding Plates (Corning).Plates were blocked for 1 h at room temperature with 200 μl/mL PBS with 0.05% Tween-20 and 0.1% BSA (Incubation buffer).Assay standards or test samples were added to appropriate wells and incubated for 2 h at room temperature.100 μL/ well of human TNF-α or human IFN-γ monoclonal detection antibody (Mabtech) diluted in incubation buffer (1 μg/mL) were added.After washing, streptavidin-HRP (Mabtech) was added, and the plates were developed with TMB substrate for 15 min followed by 0.2 M H 2 SO 4 .All washes were done using a BioTek ELx405 plate washer with 0.05% Tween-20 in PBS.Absorbance at 450 nm was measured using a Molecular Devices FlexStation 3 Reader.Data are presented as the concentration of cytokines in test samples relative to concentrations of cytokines measured in response to ConA (Table A.2).

Bioinformatics analysis
Hierarchical clustering was done using R studio version 4.1.0using Euclidean distance and complete linkage using the agglomeration method (Warnes et al., 2020).For each analysis, the average normalized value of each sample was used.

Comprehensive analysis of Scandinavian medicinal plants with usage against immune-related diseases and infections
The available historical records were reviewed with a focus on obtaining information on medicinal plants used in rural Norway and Sweden to treat ailments of skin and gut that were interpreted as wounds, infections or immune-related disorders.This led to the identification of 74 medicinal plants.A comprehensive overview of their traditional usage and the historical context of their usage is provided in Table A1, and is based primarily on the written sources made available by Høeg, Reichborn-Kjennerud and Hammarin (Hammarin, 2013;Høeg, 1975Høeg, , 1985;;Reichborn-Kjennerud, 1922, 1944, 1947).We further focused on plants with few past pharmacological studies, as assessed by the number of pharmacological studies identified by literature searches in PubMed (February-April 2020) and SciFinder (March-May 2020, December 2020) for each plant, Table 1 and Table A1.All 23 plants included in the study were subjected to water extraction to yield a crude extract, a polysaccharide enriched fraction, and a solid phase extracted fraction as described in the methods section.The yields of all extracts and fractions related to dry plant material are presented in Table A.3.Chemical and immunological characterizations were performed with all extracts and fractions.

Chemical characterization
Total phenolic content of the SPE-fractions was determined using the FC-reagent.The total phenolic content of the SPE-fractions is shown in Fig. 1, and presented as mg GAE/g.Among the analyzed plants, Alnus incana (L.) Moench [Betulaceae] had the highest phenolic content, with 150 mg GAE/g followed by H. maculatum and Vaccinium vitis-idaea L.
The monosaccharide compositions of the PS-fractions were analyzed by methanolysis and GC.The data are presented as percentage values of each monosaccharide related to the total amount of carbohydrate in the sample (Fig. 2).The 23 plants cluster into three main groups, driven primarily by differential percentages of glucose (Glc) versus galacturonic acid (GalA), arabinose (Ara), and galactose (Gal).Some of the PSfractions, such as those from V. vitis-idaea, Polypodium vulgare L.

Comparison of cytotoxic effects of plant extracts on cells
Prior to testing immunomodulatory effects of the samples, their potential cytotoxic effect was tested using the MTT assay that measures cellular viability.J447.1 mouse macrophages were incubated overnight with 1, 10, or 100 μg/ml of either crude extracts, PS-fractions, or SPEfractions.Cell viability was measured relative to J447.1 cultured in medium alone.The SPE-fractions from Sanicula europaea L. [Apiaceae] and P. saxifraga showed high cytotoxic effects at both 10 and 100 μg/ml (only 7.9% and 4.3% viability with 100 μg/ml) (Fig. 3).A relative high cytotoxic effect was also observed with the highest concentration of the PS-fraction from P. vulgare (25.3% viability).For the remaining plant extracts and fractions, the overall cell viability was within acceptable limits (well above 50%).

Immune-activating effects of plant extracts
To assess the immune activating effects, NO-secretion from macrophages was assessed.This test has been commonly used to assess proinflammatory activities of plant constituents (Wink et al., 2011).A    dose-dependent increase in NO in cell supernatants was generally observed in response to both PS-and SPE-fractions from many study-plants (Fig. 4A), but with only minor response from Potentilla erecta (L.) Raeusch [Rosaceae] and Rumex longifolius DC [Polygonaceae].The strongest NO release was observed with PS-fractions from Angelica archangelica L. subsp.archangelica [Apiaceae], D. mezereum, Ranunculus acris L. [Ranunculaceae], and Sedum acre L. [Crassulaceae] (83%, 90%, 170% and 168% relative to LPS stimulation with 100 μg/ml).For S. acre, we also observed high NO release induced by the crude extract at 10 or 100 μg/ml.For J. communis and V. vitis-idaea a high NO-release induced by the SPE-fractions was observed (80% and 67% relative to LPS stimulation with 100 μg/ml).
Next, proliferative effects on T cells and NK cells, both important mediators of immune responses, were tested.T cells classically respond to specific antigens through the T cell receptor, while NK cells express receptors that can recognize complex biomolecules.For T cells, a striking proliferative response was observed with the crude extract and the SPE-fraction from D. mezereum (79% and 84% relative to ConA stimulation at 100 μg/ml) (Fig. 4B).High proliferative activity was also observed with the PS-fraction from A. archangelica subsp.archangelica (71% relative to ConA stimulation at 100 μg/ml).As for T cells, NK-cell proliferation was observed for the PS-fraction from A. archangelica subsp.archangelica (249% relative to IL-2 stimulation at 100 μg/ml), but less strongly with the SPE-fraction from D. mezereum (Fig. 4C).Instead, proliferative effect was observed for the SPE-fractions of P. saxifraga (232% relative to IL-2 stimulation at 100 μg/ml) and Cochlearia officinalis L. [Brassicaceae] (227% relative to IL-2 stimulation at 10 μg/ml) and PS-fractions from P. saxifraga and H. maculatum (265% and 453% relative to IL-2 stimulation at 100 μg/ml) (Fig. 4C).
Further, secretions of TNF-α and IFN-γ from PBMC were assessed.We found generally a higher secretion of TNF-α compared to IFN-γ (Fig. 4D  and E).IFN-γ secretion was observed in response to crude extracts from Angelica archangelica subsp.litoralis (Fr.)Thell.[Apiaceae] and A. incana (257% and 57% relative to ConA, respectively at 10 μg/ml), the SPE-fraction from D. mezereum (both at 10 and 100 μg/ml), and very weakly in response to the PS-fraction from C. officinalis (95% and 37% relative to ConA, respectively at 100 or 10 μg/ml) (Fig. 4D).For TNF-α, there was generally secretion in response to the PS-fractions derived from all the plants, except H. maculatum and Gentiana purpurea L.

Immune-inhibitory effects by plant extracts
We next evaluated the immune inhibitory effects of the extracts and fractions, by testing their ability to inhibit LPS-mediated NO-release from macrophages, or ConA-mediated cytokine release from PBMCs.An inhibitory effect on NO-release from macrophages were observed in response to the SPE-fractions of A. incana, A. archangelica subsp.litoralis, A. dioica, S. acre, Phyllodoce caerulea (L.) Bab [Ericaceae] and S. europaea (91%, 127%, 98%, 105%, 55% and 182% reduction compared to LPS alone at 100 μg/ml) (Fig. 5A).Low or negligible inhibitory activity was observed with the PS-fractions.In contrast, rather synergistic effects were observed for PS-fractions from P. multiflorum, S. acre, and R. acris (55%, 50% and 86% at 100 μg/ml).Measuring inhibition of ConAmediated cytokine release, a higher inhibitory effect was again observed for SPE-fractions compared to PS-fractions (Fig. 5B and C).Of note, samples from D. mezereum showed synergy with ConA for both TNF-α and IFN-γ secretion (855% and 1162% relative to LPS alone at 100 μg/ml), but this synergistic effect was not observed for LPSmediated NO-release from macrophages.
Finally, we performed an unsupervised hierarchical clustering of all samples with the immunological assay datasets (Fig. 6).This analysis demonstrated that D. mezereum exhibited the most potent immune stimulatory effect, while samples from A. incana, H. perforatum and S. europaea displayed the most potent immune inhibitory effects.

Discussion
We describe here the traditional use and pharmacological properties of pharmacologically under-studied medicinal plants used in Scandinavian folk medicine against inflammatory related diseases, infections, wounds, or diverse gastrointestinal ailments.Chemical characteristics with respect to polysaccharide and phenolic contents were matched against assays measuring immune modulatory activities in order to assess whether our approach successfully could select plants for further in-depth pharmacological studies.
Water extraction of medicinal plants yields crude extracts with a great diversity of natural products, including among others carbohydrates and phenolics.Flavonoids and other polyphenols are known to contribute to anti-inflammatory properties (Yahfoufi et al., 2018) while polysaccharides may contribute to immune activation (Meijerink et al.,

2018
).When such crude extracts are screened in in vitro systems, these classes of compounds may mask or counterbalance each other's activities.Therefore, the water extract was fractionated into a polysaccharide enriched fraction and a fraction enriched in phenolics.The historical sources gave generally no detailed information about how the decoctions or infusions were prepared.The plants were therefore extracted with boiling water using the same conditions for all plants, and a modern pressurized liquid extraction instrument was chosen as it ensured the same conditions for all samples.
It was interesting to note that plants with high immune inhibitory activity (A. incana, H. perforatum, S. europaea, R. longifolius, and A. dioica) were all reported in the sources as wound-healing remedies.In contrast, plants with primarily immune activating activities (D. mezereum, S. acre, A. archangelica species, and P. saxifraga) shared usage against the cold or upper airway infections that could support a function in promoting viral immune defense.Some plants, such as A. eupatoria, Euphrasia.officinalis L. [Orobanchaceae] and G. purpurea, showed low activities with our screens, and these plants shared a usage against diarrhea and other gastrointestinal-related ailments that could indicate that effects are not mediated through the immune system.Surprisingly, Menyanthes trifoliata L. [Menyanthaceae] and J. communis showed little activity in the screens despite multiple reported indications and broad use in folk medicine.
Plant-derived polysaccharides have been shown to induce macrophage activation measured as NO-production, release of TNF-α, and/or increase in phagocytic activity (Yin et al., 2019).Especially plants rich in pectic type polysaccharides have shown potent activity (Beukema et al., 2020;Inngjerdingen et al., 2008;Vogt et al., 2016).The observed general activation of macrophages by PS-fractions in our study was thus expected.Interestingly, the observed macrophage stimulatory effects of PS-fractions from different plants was quite diverse, and the observed effects are likely related to differences in structures and monosaccharide composition.Pectic polysaccharides consist of a backbone of GalA (homogalacturonan), but also of branched rhamnogalacturonan regions (RG-I and RG-II).RG-I has a backbone of rhamnose and GalA units, and is decorated by side chains consisting of primarily Ara and Gal (Wusigale et al., 2020).Among strong inducers of NO-release, the polysaccharide fractions of D. mezereum, S. acre, and R. acris are enriched in GalA, Ara and Gal which indicates presence of pectic structures.However, A. archangelica subsp.archangelica shows low enrichment in these monosaccharides, yet the PS-fraction of this plant has strong macrophage stimulatory activity.S. acre and R. acris induced TNF-α secretion from PBMCs, with the PS-fraction being the most potent TNF-α inducer.To date, little research has been performed on polysaccharides of these two plants.From the presented screening, S. acre appears to contain a considerable amount of mannose (Man) (13%), which could indicate the  presence of hemicelluloses like glucomannan or galactomannan in the plant.Polysaccharides from S. acre have not been studied and further detailed analysis is necessary.R. acris contained considerable amounts of GalA, Ara and Gal which could indicate the presence of pectic polysaccharides.Likely, the fine structure, in terms of backbone and branching, of the polysaccharides present in plant extracts is determining the functional effects.
While the PS-fractions generally were more potent in terms of NOproduction than the SPE-fractions isolated from the same plant material, the SPE-fractions from H. perforatum and V. vitis-ideaea showed higher NO-release capacity compared to the PS-fractions.In fact, the SPE-fractions from these two plants were the most potent SPE-fractions across all tested plants.The activity is likely reflected in specific compounds, as there were no general association between high phenolic content and macrophage activation by the SPE-fractions.Interestingly, we noted that H. perforatum showed more potent immunomodulating activity than the closely related, and much less studied, H. maculatum.This could be related to slight differences in chemical structures between the two plants, although the overall phenolic content and monosaccharide composition of the polysaccharides were very similar.
A. archangelica subsp.litoralis and A. archangelica subsp.archangelica are very similar with regards to their high content of Glc.However, minor differences in amounts of GalA, Man and 4-O-Me GlcA could potentially contribute to the observed differences in biological effects of the PS-fractions obtained from these two plants.A more potent effect on TNF-α secretion was observed with the PS-fraction of A. archangelica subsp.litoralis.In contrast, the PS-fraction from A. archangelica subsp.archangelica induced higher proliferation of T cells and NK cells, as well as NO-production from macrophages.Although no studies have been conducted on the polysaccharides of A. archangelica, heteropolysaccharides isolated from the Chinese angelica Angelica sinesis, an important Chinese medicinal plant, have previously been reported to induce secretion of NO and TNF-α in macrophages (Sun et al., 2005;Yang et al., 2006).While the PS-fractions of both angelica subspecies induced varying degrees of immune activation, we observed that the SPE-fraction of A. archangelica subsp.litoralis strongly suppressed macrophage NO-release and TNF-α secretion from PBMCs.This was not evident with the SPE-fraction from A. archangelica subsp.archangelica.To date, no pharmacological studies have been conducted on A. archangelica subsp.litoralis, but A. archangelica subsp.archangelica has a longstanding tradition in Norwegian folk medicine as an "angel plant" with miraculous curative power (Teixidor-Toneu et al., 2020).Previous studies have identified active compounds in extracts from the plant, in particular furanocoumarins, coumarins and terpenes (Kaur and Bhatti, 2021).A wide range of effects of these compounds are reported, including neurological related diseases (Dahija et al., 2014;Gorick and Melzig, 2013;Kupchan and Baxter, 1975;Prakash et al., 2015;Ren et al., 2017;Vidal et al., 2012;Zhang et al., 2007), anti-cancer activity, antiviral activity, and gastro-protective activity (Altınyay et al., 2015;Fraternale et al., 2014;Fraternale et al., 2018;Joshi, 2016;Kaur and Bhatti, 2021;Krasilnikova et al., 2018;Li, Webster, Johnson and Gray, 2015;Prakash et al., 2015).
T cells are highly antigen specific through recognition of peptides in complex with MHC-molecules, and not believed to be directly stimulated through plant polysaccharides.However, antigen-independent activation occurs through lectins such as ConA via cross-linkage of the T cell receptor complex.T-cell proliferative activity was only observed in presence of PS-fractions from A. archangelica subsp.archangelica and A. dioica.The monosaccharide composition of the PS-fractions from these two plants are apparently quite different, thus the structural component mediating the observed T-cell proliferative effects is as yet uncertain.
Interestingly, the SPE-fraction, but not the PS-fraction, from D. mezereum uniquely induced high T-cell proliferation, as well as high secretion of IFN-γ and TNF-α by PBMCs.IFN-γ is secreted selectively by T and NK cells, while TNF-α is broadly secreted by macrophages, NK cells, and T cells.In light of this dichotomy, it is interesting to note that IFN-γ was induced almost exclusively by the D. mezereum SPE-fraction.This indicates a unique mode of action by as yet uncharacterized compounds in the D. mezereum SPE-fraction.Several bioactive compounds from D. mezereum have been characterized.Aqueous alcohol extracts from D. mezereum and the isolated compound mezerein possess anti-leukemic activity, and the diterpenoids daphnetoxin and gniditrin are shown to have cholesterol-lowering activity (Kupchan and Baxter, 1975;Vidal et al., 2012;Zhang et al., 2007).However, no immune modulating activity has been reported.Moreover, D. mezereum has been reported to have strong toxic effects due to the presence of diterpen esters, such as gniditrin (Gorick and Melzig, 2013).It should be noted that many of the assumed toxic compounds are highly lipophilic, and are thus not expected to be found in the water extract studied here.
Immune inhibitory activities were also observed in the SPE-fractions from A. incana, S. europaea, and P. saxifraga.Phenolic compounds isolated from A. incana have previously been reported to have antiinflammatory properties (Dahija et al., 2014;Joshi, 2016;Prakash et al., 2015;Ren et al., 2017).Previous studies have attributed the anti-inflammatory effects of A. incana to the diarylheptanoid glycoside oregonin, which has been found to reduce inflammation in macrophages and structurally resembles the well-known antioxidant curcumin (Krasilnikova et al., 2018).

Conclusion
In summary, a wealth of information on Scandinavian folk medicine was collected from 19th-and 20th-century sources, forming the foundation of this study.These historical descriptions of ailments treated by each plant allowed for the recognition of plants with possible immunomodulatory activity.However, details on the preparations of plants for medicinal usage were sometimes lacking or vague.These missing details made it difficult to predict the nature of the compounds and bioactivity that may have been responsible for a plant's place in a remedy.While the focus on polysaccharides and polyphenols in this investigation may not have captured the most relevant compounds for every plant's recorded immunomodulatory function, overall we were able to detect the predominant immune activating or inhibitory activities with most of the studied plant extracts.This could underscore the feasibility of our selection approach.Our findings thus support an approach of combining historical sources with modern pharmacology in the discovery of plant sources containing potentially new pharmacological compounds.
[Polypodiaceae], and Pimpinella saxifraga L. [Apiaceae], contained mainly Glc, which is most likely due to the high abundance of starch.PSfractions that consisted of considerable amounts of GalA were Daphne mezereum L. [Thymelaeaceae], P. multiflorum, and Agrimonia eupatoria L. [Rosaceae], which could indicate the presence of pectins.The more unusual monosaccharide 4-O-methyl glucuronic acid (4-O-Me GlcA) was found in relatively high amounts in Antennaria dioica (L.) Gaertn [Asteraceae].

Fig. 1 .
Fig. 1.Total phenolic content in SPE-fractions expressed as gallic acid equivalents (mg of GAE/g of extract).

Fig. 2 .
Fig. 2. Illustrative representation of the monosaccharide composition of PSfractions obtained from each plant.Data are presented as the percentage of each monosaccharide in relation to total amount of carbohydrates in the sample, and hierarchical clustering of monosaccharides versus plant species.

Fig. 3 .
Fig. 3. Assessment of cellular toxicity of extracts.Cell viability of the J774.1 macrophage cell line, measured by the MTT assay for each plant extract and fraction.Data are presented as percentage live cells of treated cells relative to untreated cells.

Fig. 4 .
Fig. 4. Immune activating effects of plant extracts and fractions.A) NOsecretion from J477.1 macrophages for each plant extract or fraction, presented as percentage values relative to LPS stimulation.B) T-cell and C) NK cell proliferation after 6-day cultures of CFSE-labelled PBMCs with indicated plant extracts and fractions.Data presented as percentage values relative to ConA stimulation for T cells and IL-2 stimulation for NK cells.D) IFN-γ and E) TNF-α secretion measured via ELISA in supernatants from overnight cultures of PBMCs with indicated plant extracts and fractions.Data are presented as percentage relative to PMA/ionomycin stimulation.

Fig. 5 .
Fig. 5. Immune inhibitory effect of plant extracts and fractions.A) Percentage reduced NO-secretion after culturing J477.1 macrophages over night with plant extracts and fractions, relative to LPSstimulation or medium alone.Reduction in B) TNFα and C) and IFN-γ secretion measured by ELISA in supernatants from PBMCs cultured overnight in plant extracts and fractions, relative to PMA/Ionomycin and medium alone.Percentage values are measured relative to LPS stimulation alone and PBMC without stimulation.

Fig. 6 .
Fig. 6.Hierarchical clustering of all plant extracts and fractions with immunological assays.Red indicates reduced immune activity, while blue color indicate activating effect.Plant extracts and fractions, and immunological assays is colored coded.The immunological assays are given along the x-axis, and plants along the y-axis.(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)