Peripheral blood mononuclear cells (PBMCs) are often considered the industry standard for investigating many aspects of immunological response in a cell culture system. As a mixed population of T-lymphocytes, B-lymphocytes, natural killer cells, monocytes, macrophages and dendritic cells, PBMCs represent key cells involved in the genetic expression of cytokines leading to innate and adaptive immune and inflammatory responses. To investigate the cytokine response induced by immuno-stimulatory herbal extracts, PBMCs were treated with an herbal extract followed by microarray analysis on isolated cellular RNA. To date, the effect of immuno-stimulatory herbs has only been evaluated by observing the expression of a limited number of cytokines. Such studies can be misleading and do not provide a broad or complete understanding regarding the effect of the herbal extract on immune cells. In our studies, 1×107 cells were treated with 5 microL herbal extract for 18 hours. This concentration was based on an average human prescribed dose in relation to the approximate total blood volume (2.5 mL extract per 5 L blood). Following incubation with the extract, total RNA was purified from the cells and mRNA expression levels were evaluated by microarray analysis using an Affymetrix platform.
These studies were done using herbal extracts historically used to boost immune activity in patients, including Astragalus membranaceus
(Milk-Vetch Root, Huang qi), Sambucus cerulean
(Blue Elderberry), and Andrographis paniculata
(India Echinacea, King of Bitters). Since these extracts were prepared in an aqueous vehicle containing ethanol and glycerol, PBMCs were treated with identical concentrations of ethanol and glycerol in distilled water (ethanol) as a control. Our initial method to measure modulation in gene expression was done using scatter plot analysis. Treatment of PBMCs with the vehicle solution had almost no effect on gene expression when compared to untreated PBMCs (, Plot A). However, as shown in , treatment of PBMCs with Astragalus
extract led to the alteration of expression of many cellular genes (compare diffuse scattering of , Plot A to Plot B). In this figure, the diagonal lines off the center represent 2-, 3-, 10-, and 30-fold levels of induction or repression of gene expression. With Astragalus
treatment, the expression of several hundred cellular genes was altered, many to levels greater than 30-fold relative to ethanol treatment alone (, Plot B). The induction
of gene expression was greater than the level of gene repression
in regard to the relative fold-change and number of genes altered (146 genes induced with 58% induced greater than 10-fold vs 118 genes repressed with 21% repressed greater than 10-fold). A similar scatter plot appearance was observed after treatment with Sambucus
(, Plot C). After treatment of PBMCs with Andrographis
, similar fold changes of induction/repression were observed, but fewer genes appeared to be affected (, Plot D). Since all three of these immuno-stimulatory herbal extracts produced relatively similar scatter plot profiles, an extract prepared from the immuno-suppressive herb, Urtica dioica,
was studied. Treatment of cells with Urtica
led to very limited or low level changes in cellular gene expression (, Plot E). The lack of change in gene expression after treatment with Urtica
argues that the alteration in gene expression by the immuno-stimulating extracts (e.g., Astragalus)
was not due to a non-specific effect of any
botanical extract, but instead was a legitimate effect due to specific components present in the immuno-stimulating herbal extracts. It was not surprising that the Urtica
extract did not alter gene expression since the PBMCs were from a reportedly healthy individual and immune-suppressive effects would likely not be observed. This effect has been observed previously where oral administration of an Urtica
extract had no effect on cytokine expression in a healthy individual and only repressed cytokine expression following lipopolysaccharide induction 
Scatter plot representation of botanical extract regulation of gene expression.
Although PBMCs represent the primary immune cells involved in cytokine production, other cell types may also be involved in the immune and inflammatory responses associated with these botanical extracts. Therefore the intestinal cell line, Caco2 cells, were used and changes in gene expression measured following treatment with the botanical extracts. Results demonstrated no dramatic changes in gene expression following treatment with any of the botanical extracts (data not shown). Since PBMCs are key cells involved in cytokine production, these results suggest that the active constituents in these botanicals may interact with cells in the PBMC population leading to changes in cytokine expression.
The microarray results from treated PBMCs were confirmed by quantitative real-time PCR (qPCR). Ten representative genes were selected from the microarray expression profiles and changes in the expression level confirmed using qPCR. All ten genes tested using qPCR showed comparable gene expression levels as those obtained in the microarray analysis (). From our previous experience with microarray studies, significant variations between expression levels determined by microarray analysis or qPCR have not been observed 
Quantitative real-time PCR (qPCR) verification of microarray data.
In order to compare the gene expression profiles altered by the herbs, comparative scatter plots were done between the immunostimulatory herbs. As shown in (Plot A), the genes altered by Sambucus and Astragalus were very similar as indicated by a lack of diffuse scattering of the genes. Treatment with Astragalus did lead to additional changes in gene expression as compared to Sambucus, indicated by a few genes induced 3–10 fold specifically by Astragalus. These Astragalus specific alterations occurred toward unique genes as well as to genes regulated by both Astragalus and Sambucus but were induced to a higher expression level following Astragalus treatment (). More differences in gene expression were observed when comparing Astragalus to Andrographis or Sambucus to Andrographis (, Plot B and Plot C, respectively). Again, the genes altered by these comparative herbs were similar, but each herb also induced a unique subset of genes (data not shown and ). This data suggests that immuno-stimulating herbs may consistently affect a common set of genes while also having additional unique effects on gene expression.
Host gene expression regulated by Astragalus membranaceus treatment of PBMCs.
The cellular genes induced and repressed after treatment with Astragalus
were identified as (a) those genes which lie outside the scatter plot profile of untreated vs. ethanol treatment (, Plot A) and (b) those genes which were altered in gene expression at least 3-fold compared to ethanol treatment alone. From our experience and others on microarray studies, a fold-change of 2.5 or higher is typically considered significant; so limiting our fold-change to 3-fold or higher provides strong confidence in the validity that the expression of these genes was truly being altered 
. Following treatment of PBMCs with Astragalus
extract, approximately 150 genes were induced (data not shown). At least 65% of the genes that were induced by Astragalus
are known to have functions involved in the immune/inflammatory response. (Astra column) shows the immune/inflammatory genes induced upon treatment with Astragalus
Previous studies on Astragalus
have suggested that treatment leads to (a) increased phagocytic activity of macrophages, (b) an increase in proinflammatory cytokines IL1, IL6 and TNF, and (c) an increase in levels of lymphocyte stimulatory IL2 and IL2 receptor expression 
. Immune and inflammatory uses of Astragalus
include antimicrobial therapy 
, the treatment of leukemia and lung cancer 
, and wound healing 
. Based on literature and anecdotal evidence, Astragalus
treatment may be associated with increased risk of bleeding, and blood pressure lowering effects 
. The genes induced in our study by Astragalus
treatment agree well with these previous studies and the physiological effects of the herb on the body. In PBMCs, IL1alpha, IL1beta, IL6 and TNFalpha induced/interacting proteins (e.g., TNFAIP6) were highly induced, as well as genes involved in immune cell stimulation, chemotaxis, extravasation, and maturation/differentiation (). In addition, genes involved in increased risk of bleeding (e.g., thrombomodulin, monoamine oxidase), regulation of blood pressure (e.g., adrenomedullin, prostaglandin endoperoxide synthase 2 [Cyclooxygenase-2], monoamine oxidase), and wound healing (epiregulin, fibronectin type III, heparin-binding EGF-like growth factor, interleukin 24) were identified (). The effect of Astragalus
on the expression of most of these genes has not been previously shown, and may provide new insight into the physiological effects of this herb.
In support of the comparative scatter plots (), a very similar set of genes regulated by the Astragalus extract was also regulated by the Sambucus extract (, compare Astra column to Sambu column). This suggests that similar physiological activities are present in both botanical extracts, however, the active components may still be unique (i.e., conservation of function with diverse chemical constituents). Treatment of cells with the Andrographis extract also altered many genes in common with Astragalus, but overall, the level of induction of genes regulated by Andrographis was lower (, Andro column).
Monocytes represent a significant percentage of cells present in the PBMC population. They can mature into macrophages or differentiate into dendritic cells (DCs) upon migration into tissues. Monocytes that have matured into macrophages become more adherent to tissue culture surfaces and increase expression of CD11/CD18 integrins 
. To test for monocyte maturation, PBMCs were treated with Astragalus extract and assayed for the effect on cell attachment. As shown in (upper panels), treatment of PBMCs with Astragalus extract led to an increase in adherent cells. Quantitatively, the number of adherent cells was increased approximately 15-fold. These results suggest that treatment of PBMCs with an immuno-stimulating herb, such as Astragalus, not only increases the transcription of immune related genes, but that these gene products are inducing immune cell maturation and/or differentiation. It is known that IL4 and GM-CSF mature monocytes towards DCs and that IL6 directs monocyte maturation primarily towards macrophages 
. Upon treatment of PBMCs with Astragalus extract, IL6 expression was highly up-regulated (), whereas IL4 expression and GM-CSF was unchanged. These results suggest that the monocyte maturation observed upon treatment with Astragalus extract may be directed toward a macrophage lineage. This is in agreement with previous data in which the isolated APS polysaccharide from Astragalus was shown to induce macrophage activation 
. To investigate if this monocyte maturation was due to a direct effect of the extract on monocytes, monocyte to macrophage maturation following treatment with Astragalus extract was tested using the THP-1 monocyte cell line. Treatment of these cells with the Astragalus extract led to no significant increase in adherent cells as compared to untreated or ethanol treated cells suggesting the lack of a direct effect of the extract on monocyte to macrophage maturation (, lower panels). The maturation of monocytes observed in the PBMC population may be linked to cytokines released from other responder cells present in the PBMC population (e.g., T helper cells). To test for this, the cell-free culture media from untreated or Astragalus treated PBMCs (24 hours post treatment) was added to THP-1 cells. As shown in , an increase in adherent cells was observed in the THP-1 cells treated with the cell-free media from Astragalus treated PBMCs but not from untreated PBMCs. As a control, the cell-free media from PMA-treated PBMCs was added to THP-1 cells. Similar increases in adherent cells were observed using either cell-free media from PMA or Astragalus treated cells (). These results suggest that the Astragalus induced monocyte maturation in the PBMC population was due to a secondary effect following the release of cytokines from responder cells.
Astragalus membranaceus treatment of PBMCs led to monocyte maturation.
Maturation of THP-1 cells by cell-free media was further monitored by the expression of CD14 antigen and by expression of the β2 integrin component, CD11b, both of which are upregulated when this promonocytic cell line differentiates into mature monocyte/macrophage-like cells 
. Incubation of THP-1 cells with cell-free media from mock-treated PBMCs resulted in 15.4% of cells expressing CD11b while incubation with cell-free media from either Astragalus- or PMA-treated PBMCs resulted in 28.5% or 28.6%, respectively, of cells expressing CD11b. It has been previously shown that the presence of adhesion-associated molecules on monocytes does not necessarily correlate with adherence to a substrate, but that adherence requires intracellular signaling which conformationally activates the adhesion molecule 
. Mock-treated THP-1 cells may therefore express an inactive form of CD11b that does not mediate adherence (). Incubation of THP-1 cells with cell-free media from mock- or PMA-treated PBMCs resulted in 3.4% and <1% of cells expressing CD14, respectively. Incubation with cell-free media from Astragalus-treated PBMCs resulted in expression of CD14 in 50.3% of the cells. In addition, extracts from Astragalus
-treated PBMCs caused a 4.5-fold increase in the percentage of double-positive cells expressing both CD14 and CD11b as compared to mock-treated PBMCs. shows averaged levels of expression of total CD14 or CD11b positive cells and CD14+CD11b+ double positive cells, normalized to mock levels, from PBMCs isolated from two different patients. Although cell-free media from both Astragalus
- and PMA-treated PBMCs induced maturation as measured by cell adhesion and an increase in percentages of total CD11b+ cells, only media from Astragalus
-treated PBMCs induced expression of CD14.
In general, the cytokine gene profile induced following treatment with Astragalus
was not strongly directed toward either a Th1 or Th2 response, but rather a more generalized or preparative immune/inflammatory response (see ). Most genes involved in defining a Th1 or Th2 response remain unchanged (eg. Th1: IFNgamma, IL18, SOCS1, STAT1, CSF2; Th2: IL13, IL4, IL5, ICOS). However, gene expression profiles have been defined regarding human monocyte to macrophage maturation and polarization toward M1 or M2 phenotypes 
. As shown in , the gene expression profile of Astragalus
treated PBMCs follows an M1 polarization (with both induced and repressed genes) with 83% of the genes defining the polarization of M1/M2 macrophage matching that of an M1 lineage. In agreement with this, COX-2 which was highly induced following treatment with Astragalus
(), has long been associated and induced in M1 macrophages 
. Likewise, COX-1, which is up-regulated in M2 macrophage 
, was slightly repressed following Astragalus
treatment (data not shown). These results suggest that Astragalus
treatment matured monocytes toward an M1 polarization.
Transcriptional profiling of M1/M2 macrophage polarization induced by Astragalus membranaceus.
Since the microarray data presented thus far represents treatment of a single population of PBMCs from a commercial source (Lonza, Inc. [identified as PBMC set I]), we wanted to confirm that similar gene modulation would occur in fresh human PBMCs. Fresh PBMCs were obtained and purified as described in the Materials and Methods
(identified as PBMC set II). Following treatment of these cells with Astragalus
extract, 123 genes were found to be induced (, PBMC set II), giving a similar response as that observed with the commercially obtained PBMCs, PBMC set I (146 genes induced). Like PBMC set I, the genes induced in PBMC set II primarily had an immune/inflammatory role (65% vs. 62%, respectively). Although many of the inflammatory genes induced were different between the two PBMC populations, over 30% of the induced genes were identical. This level of identity is high given that the PBMCs were from different individuals and obtained from fresh vs. frozen. The genes commonly induced are likely many of the key players involved in the known physiological responses to Astragalus
including those involved in immune cell activation, differentiation and chemotaxis (eg. CCL3, CCL4, CXCL1, CXCL2, CXCL3, IL1alpha, IL1beta, IL6 and IL8), as well as wound healing and blood coagulation (eg. coagulation factor III, prostaglandin endoperoxide synthase 2, thrombomodulin, and IL24) (see , panel A).
Different PBMC isolates led to similar changes in gene expression following treatment with Astragalus membranaceus.
Genes that were induced in only one of the PBMC sets are likely related to human to human genetic variations and/or physiological state of the individual at the time of the blood draw. Many of the uniquely induced genes represented functional families of genes. For example, in PBMC set I following treatment with Astragalus
, several metallothioneins, vanins, and integrins were induced (, panel B). In PBMC set II, these gene families remained primarily unchanged, however gene families including colony stimulating factors, dual specific phosphatases, and growth arrest and DNA-damage inducible were induced. Notably, PBMC set II also led to the unique induction of CD69, early growth response 1 and microRNA 155 which are all key proteins involved in the inflammatory response (, Panel B). Previous studies demonstrate that considerable natural variation in gene expression levels exists within and among human populations 
. When they compared individuals, gene expression variations occurred in an estimated 83% of genes. Genes which are differentially expressed among populations may be particularly relevant as candidates for disease, or in our studies, for differences in the immune response.
We also identified over 100 cellular genes which were repressed in expression levels after treatment with Astragalus (). In this data set, the majority of the genes (approximately 70%) were not directly involved in the inflammatory/immune response, but were genes involved in metabolic pathways such as lipid and carbohydrate metabolism (e.g. fructose-1,6-biphosphatase 1, fucosidase, lipase A, and alpha galactosyltransferase 1). This data supports the concept that these herbs are involved in the induction of a systemic inflammatory and immune response. Again, when PBMC set I and PBMC set II were compared, similar numbers and percentages of genes were repressed (), but for those genes involved in the inflammatory response, specific gene families to each PBMC set were repressed. For PBMC set I, several major histocompatability complex class II genes were repressed (, panel C). For PBMC set II, repressed gene families included leukocyte immunoglobulin receptors, C-lectin domain family members, neutrophil cytosolic factors and metallothioneins (, panel C). At this time, it is difficult to speculate on what role the suppression of these inflammatory gene families may play in the physiological effects of Astragalus.
In order to further understand the temporal effect of Astragalus treatment on immune cells, PBMCs were treated with the botanical extract and microarray analysis performed on RNA samples prepared at 3, 8 and 18 hours post treatment. shows scatter plot representation of gene expression levels of ethanol treated PBMCs in comparison to cells treated with the Astragalus extract for the indicated times. As shown, the cellular transcriptional response to the herbal extract was rapid with a dramatic induction/repression of several hundred genes by 3 hours post treatment (, panel A). By 8 and 18 hours post treatment, fewer genes were induced/repressed, but still represented 100–200 genes (, panels B and C). The induction of genes by Astragalus was biphasic with different families of genes being induced at different times post treatment. As shown in , at 3 hours post treatment, there was a predominance of genes (approximately 40%) involved in transcriptional, translational and signal transduction processes with only 26.6% of genes having an inflammatory role. However, by 8 and 18 hours post treatment, approximately 70% of the genes induced were involved in the immune/inflammatory response (). The immune genes which were induced at 3 hours post treatment typically continued to be induced through 18 hours, including various chemokines (CCL3, CCL4, CCL20) and interleukins (IL1beta and IL6) (). Interestingly, there are differences in the immune/inflammatory genes induced between 8 and 18 hours post treatment. Genes involved in the type I interferon response, such as interferon-induced 44 and interferon stimulated exonuclease 20 kDa, were induced only at the 8 hour time point (). Likewise, several interleukins were only induced at the 18 hour time point, including IL12B and IL24 (). Since PBMCs contain a variety of immune cells, these temporal effects may be due to constituents present in the herbal extract interacting with (a) responder cells which subsequently induce/release cytokines and lead to transcriptional responses in other effector cells, or (b) cells which induce cellular receptors/signal transduction processes which, after expression, can respond to additional herbal constituents leading to inflammatory gene induction.
Temporal regulation of gene expression in PBMCs following Astragalus membranaceus treatment.
Lipopolysaccharides (LPS) are major components of the outer membrane of Gram-negative bacteria and highly potent activators of the innate immune response. The interaction of LPS via CD14 with cells, such as macrophages, granulocytes, and dendritic cells, leads to the synthesis of a multitude of inflammatory mediators 
. However, strong inflammatory responses to LPS can be hazardous and may lead to endotoxin shock and death 
. Therefore the presence of LPS or endotoxin contamination in pharmaceutical drugs and medicines is a common concern. The herbal extracts used in these studies were assayed for the presence of endotoxin using a modified Limulus Amebocyte Lysate kit. As shown in , herbal extracts historically reported and used to stimulate the immune response (Astragalus membranaceus, Sambucus cerulea, Andrographis paniculata, Echinacea angustifolia
and Glycyrrhiza glabra
) had highly elevated levels of endotoxin (>100 EU/ml). Those herbal extracts historically used to repress immune activity (Urtica dioica
and Tylophora asmatica
) had much lower endotoxin levels (<100 EU/ml). This correlation between activity and endotoxin concentration may suggest that the presence of LPS was not due to contamination of the extract, but instead was a common constituent of the immune-stimulatory extracts and as such, may play a role in immune regulatory activity. In order to confirm that the LPS present in the extracts was not due to contamination during harvesting or extract preparation, multiple extracts of Astragalus
were prepared using different lots of plant material from the suppliers. As shown in , all the extracts of Astragalus
had high endotoxin levels (>100 EU/ml) and all the extracts of Urtica
had low endotoxin levels (<100 EU/ml). This further supports that the presence of LPS in the botanical extracts was not due to exogenous bacterial contamination during harvesting, handling, or extract preparation, and instead was likely due to the lysis of endophytic bacteria which colonize the internal tissue of the plants in a mutualistic and asymptomatic fashion. Of the nearly 300,000 plant species that exist on earth, it is believed that each individual plant is host to one or more bacterial endophytes 
. Endophytes are defined as endosymbionts that live within a plant without causing any apparent disease. Endophytes are not only found in the roots and in the rhizosphere, but are often found to colonize the intercellular spaces and have been isolated from all plant compartments, including seeds 
. A vast number of these endophytes, including Rhizobium sp.
which has been isolated from Astragalus
, are Gram-negative bacteria 
. Based on this, it is expected that many botanical extracts would contain LPS due to lysis of these bacterial endophytes.
Endotoxin concentration present in botanical extracts.