Comment on “Investigation of Possible Herb-Drug Interactions for the Treatment of
Cardiovascular Diseases”
This correspondence is intended as a brief report of an investigation that furthers
the data in previous reports published in this journal by McEwen[1 ] and Maione.[2 ]
Cardiovascular diseases (CVDs) are the leading cause of mortality and mainly account
for thrombotic events.[3 ] Herbal medicines are widely used for the treatment of CVDs due to traditional and
cultural beliefs and have so for many years. However, in several countries, herbal
medicines have relatively loose regulatory requirements and can be considered self-prescribed
medications.[1 ]
[4 ]
[5 ] In numerous experimental studies, herbal medicines have been found to have an effect
on biological mechanisms related to the cardiovascular system.[6 ]
[7 ] Hence, increasingly greater attention has been paid to potential risk of interactions
between conventional drugs and widespread active herbs.[4 ]
[8 ]
The article by McEwen[1 ] reviewed herbal efficacy and noted their potential to modify progression of CVD
via the modification of platelet function. Furthermore, Maione[2 ] investigated two single compounds of danshen (Salvia miltiorrhiza ) and reported their underlying mechanism on platelet function and hemostasis. McEwen[1 ]
[9 ] also suggested that the herb − drug interactions are not fully understood and outlined
underlying mechanisms of action and, in particular, potential risks that herbal medicine
create by affecting concomitant anticoagulant therapy. In terms of the complexity
and wide prevalence of herbal medicines, it is of great importance to explore the
biological mechanism of herb − drug interactions for reasons of safety. This correspondence
explains the possible biological mechanism of herb − drug interaction at the drug
targets of action.[10 ] To our knowledge, this issue has not been fully described in the existing literature.
In clinical practice, both Salviae miltiorrhizae radix et rhizoma (Danshen in Mandarin, DS) and Chuanxiong Rhizome (Chuanxiong in Mandarin, CX) are
herbs frequently used for invigorating blood circulation and eliminating stasis.[11 ]
[12 ] A pairing (a basic unit of complex herbal formulae) of the two herbs was one of
the most frequently compatible herbal pairs in best-selling herbal formulae released
by the China Association of Chinese Medicine in 2017.[13 ] Therefore, our investigation employs the herbal pair (DS − CX) to analyze the mechanism
of action and possible interaction with Western drugs. Moreover, to understand the
intrinsic herb − drug interaction on the molecular level, this investigation uses
the network of pharmacological approach[14 ] to integrate the abundant data which have accumulated from previous research on
herbal medicines and current approved anti-CVD drugs. For detailed methods, please
refer to the Appendix (also refer to [Fig. 1 ] and [Tables 1 ], [2 ]).
In this investigation, the “target − (pathway) − target” network clearly shows that
DS − CX interacts with the targets of anti-CVD Western drugs. DS − CX may interact
with 56 (33.9%) of targets of anti-CVD Western drugs. Totally, DS − CX's 384 compounds
may affect 567 biological molecules. Meanwhile, it is of significance to specify the
kind of interactions in the specific pathways, especially the main thrombotic pathways.
Our investigation found that two (prostaglandin-endoperoxide synthase 1 [PTGS1], integrin
subunit alpha 2b [ITGA2B]; 22.2%) out of nine Western medicine targets on platelet
activation are involved in target sets of DS − CX. Similarly, for the pathway of vascular
smooth muscle contraction, there are 6 of 17 drug targets (adenosine A2a receptor
[ADORA2A], ADORA2B, adrenoceptor alpha 1A [ADRA1A], ADRA1B, ADRA1D, potassium calcium-activated
channel subfamily M alpha 1 [KCNMA1]; 35.3%) are relevant to DS − CX. In addition,
for the pathway of complement and coagulation cascades, the corresponding number is
4 (F2, F10, Serpin [serine protease inhibitors] family E member 1 [SERPINE1], plasminogen
activator, urokinase [PLAU]; 36.4%) out of 11. Refer to the Appendix for more detailed
results of network analysis. Higher overlapping intensity between pharmaceutical and
herbal targets may mean higher risk in clinical practice of Western drugs.
For example, the verified corresponding targets of aspirin include prostaglandin G/H
synthase 1 (COX-1 [Cyclooxygenase-1]; gene name: PTGS1 ) and prostaglandin G/H synthase 2 (COX-2; gene name: PTGS2 ).[15 ]
[16 ] COX-1 mainly contributes to the arachidonic acid metabolism pathway and the platelet
activation pathway. COX-2 is mainly involved in the arachidonic acid metabolism pathway
directly relevant to thrombosis and the vascular endothelial growth factor (VEGF)
signaling pathway relevant to thrombosis indirectly. COX-1 and COX-2 might be targeted
by senkyunolide B, senkyunolide C, and senkyunolide E from CX, and phenanthraquinones
from DS, including tanshinone I, 1-dehydrotanshinone IIA, neocryptotanshinone, przewaquinone
A, isotanshinone I, nortanshinone, etc. Previous studies have partially supported
the potential effects of DS and CX on COX-1 and COX-2.[11 ]
[12 ]
[17 ]
With the wide prevalence of active herbal medicines, it is necessary to add an understanding
of the biological mechanism of herb − drug interaction. In general, this kind of herb − drug
interaction implies that patients simultaneously taking anti-CVD Western drugs and
DS − CX may be exposed to additional risks caused by accumulated effects on common
herb − drug targets. More importantly, this kind of risk is not usually visible and
difficult for clinical physicians to observe due to wide herbal utilization as undisclosed
self-medication. In this context, herb − drug relationships are worthy of further
exploration in the future, especially from experimental and clinical perspectives.
Appendix—Methods and Results
Network pharmacology has been widely used to explain complex biological system and
herbal medications.[18 ]
[19 ] This investigation employed the approach of network pharmacology to explain the
possible biological mechanism of herb − drug interaction at the drug targets of action.
First, the data of compounds of danshen–chuanxiong (DS − CX) were preliminarily collected
from chemical databases, including TCM database@Taiwan,[20 ] traditional Chinese medicine-information database (TCM-ID),[21 ] Herbal Ingredients’ Targets Database (HIT),[22 ] and traditional Chinese medicine systems pharmacology database (TCMSP),[23 ] together with the supplements from PubMed literature. The data of targets or putative
targets of DS − CX compounds were obtained from PubChem,[24 ] TCMSP,[23 ] and similarity ensemble approach (SEA).[25 ] Moreover, we collected information around drugs used for the treatment of cardiovascular
diseases (CVDs), mainly including anti-CVD small molecules and biological products
approved by the U.S. Food and Drug Administration (FDA) by September 2017.[26 ] The data of targets of these conventional drugs were extracted from the databases
of Drugbank[15 ] and Drugcentral.[16 ] The enrichment analysis of targets of DS − CX and conventional drugs was carried
out by STRING[27 ] to acquire the target − pathway associations.
Based on the target − pathway associations, a “target − (pathway) − target” (TPT)
network was constructed with the reference to our previous work.[14 ] In this TPT network, a node represents a DS − CX's or pharmaceutical target distinguished
by different colors, and an edge indicates that both of connected nodes are involved
in at least one of the same pathways. By highlighting targets relevant to specific
pathways and neglecting irrelevant ones, the TPT network can be shown as various subnetworks
specific to different pathways.
[Fig. 1A ] shows the whole TPT network and various TPT subnetworks specific to different anti-thrombotic
pathways, that is, the platelet activation pathway ([Fig. 1B ]), the vascular smooth muscle contraction pathway ([Fig. 1C ]), and the complement and coagulation cascades pathway ([Fig. 1D ]).
Fig. 1 The TPT (target − [pathway] − target) networks; (A ) whole network, (B ) subnetwork in the platelet activation pathway, (C ) subnetwork in the vascular smooth muscle contraction pathway, (D ) subnetwork in the complement and coagulation cascades pathway. CVD, cardiovascular
disease; DS–CX, danshen–chuanxiong.
As identified in [Fig. 1A ], there exists a large number of biological molecules relevant to DS − CX; more importantly,
many of them are also targets of conventional anti-CVD drugs. This type of overlap
seems to be more distinct in the vascular smooth muscle contraction pathway and the
complement and coagulation cascades pathway in comparison with the platelet activation
pathway, as identified in [Fig. 1B–D ]. In order to measure quantitatively, we established an adjusted Jaccard's index
(JI) which denotes how many percentages of compounds or targets of anti-CVD drugs
are covered by DS − CX. In [Table 1 ], values in columns “W,” “DS,” “CX,” and “DS–CX” represent the number of corresponding
compounds or targets. As [Table 1 ] shows, there is only one common compound of anti-CVD drugs and DS − CX, but targets
between them have a high overlap.
Moreover, [Table 2 ] shows target-based interaction between anti-CVD drugs and DS − CX in different thrombotic
pathways. Values in the middle of the two columns indicate the corresponding number
of anti-CVD drug targets or common targets, which are used to calculate the JI. The
relatively higher JI values in [Fig. 1C, D ] further demonstrate the observation in [Fig. 1 ] that herb − drug interaction between anti-CVD drugs and DS − CX may be more serious
in the vascular smooth muscle contraction pathway and the complement and coagulation
cascades pathway.
Table 1
A general comparison of conventional anti-CVD drugs and DS − CX
Western drugs (W)
Danshen (DS)
Chuanxiong (CX)
DS − CX
JI
Compounds
217
188
220
384
0.005[a ]
Targets
165
439
339
567
0.339
Abbreviations: CVD, cardiovascular disease; DS–CX, danshen–chuanxiong; FDA, Food and
Drug Administration; JI, Jaccard's index.
a Adenosine is the only common compound of anti-CVD drugs and DS − CX. It was approved
in 1989 by FDA for cardiac therapy.
Table 2
Target-based interaction between anti-CVD drugs and DS − CX
Anti-CVD drug targets
Common targets
JI
Whole network [Fig. 1A ]
165
56
0.339
Subnetwork [Fig. 1B ]
9
2
0.222
Subnetwork [Fig. 1C ]
17
6
0.353
Subnetwork [Fig. 1D ]
11
4
0.364
Abbreviations: CVD, cardiovascular disease; JI, Jaccard's index.
