Safety assessment of herbal drugs as alternatives to treat non-psychotic mental diseases during pregnancy

Pregnancy is an extraordinary state for the expecting mother. Her body goes through many
changes to bring forth new life. Amidst the joy of bearing a child, up to 20% of pregnant women
will suffer from depression. Other ailments are anxiety, restlessness, stress, and sleeping issues,
which are categorized as non-psychotic mental diseases (NMDs). In order to prevent further
discomfort for the expecting mothers and negative birth outcomes like preterm birth or
preeclampsia, medications are prescribed to mitigate the symptoms. Typical treatments are
selective serotonin reuptake inhibitors (e.g., citalopram) and benzodiazepines (e.g., diazepam)
to treat mild depression and anxiety, respectively. Those synthetic drugs are for instance known
to cross the placental barrier, which might be one of many factors having physicians and
pregnant women discuss possible alternatives. Alternatives to treat NMDs are valerian,
lavender, hops, St. John’s Wort (SJW), and California poppy. Treatment with herbal medication
is as old as the history of medicine itself and remedies of natural origin have ever since been
considered trustworthy. This notion is also reflected in the different handling of synthetic drugs
versus herbal ones. The regulations worldwide differ. In some countries herbal medications are
categorized as dietary supplements. Although the Committee on Herbal Medicinal Products has
regulations in place for herbal medications; data on those products are often missing and most
herbal medication are simply labelled as not recommended during pregnancy and lactation.
The high trust in herbal medications and the lack of safety data at the same time, potentially
expose pregnant women and their unborn child to unknown risks.
To assess the safety of valerian, hops, SJW, and California poppy we first established
bioanalytical methods for the compounds attributed to the therapeutic effect, namely valerenic
acid, humulone, hyperforin, hypericin, and protopine. Lavender and its volatile constituent
linalool have been excluded from the following studies.
In a first step the bioanalytical methods allowed determination of foetal exposure to those single
compounds in an ex vivo placental perfusion model. We could show that protopine and valerenic
acid crossed the placental barrier, while hyperforin only transferred to a small extent into the
foetal circuit. Hypericin on the other hand, accumulated in the placenta.
To make an informed safety assessment, knowledge of potential metabolites is also needed. We
therefore investigated phase I and phase II metabolism for valerenic acid, protopine, and
humulone in human liver microsomes. For valerenic acid, we were able to identify five
dehydrovalerenic acid isomers, valerenic acid glucuronide, and oxoglucuronide. Protopine was
metabolised to dimethylcryptopine, and to a metabolite formed by catechol-O-methylation and
a third metabolite formed by ketone reduction. Finally, for humulone we could identifiy eleven
hydroxyhumulone isomers, three dihydroxyhumulone isomers, and a metabolite formed by
aliphatic hhydroxylation and ketone reduction. Those metabolites have been described in
human liver microsomes for the first time.
The third part of the project was dedicated to screening for potential mitochondria targeted liver
toxicity in HepG2 and HepaRG cells. Hypericin was, in term of toxic effects, the most promising
candidate. Hypericin turned out to be a potent mitochondria toxicant. It showed inhibition of
complexes I, III, and IV in the electron transport chain and a decrease in mitochondrial ATP
production, which leads to mitochondrial reactive oxygen species accumulation. Nrf2 was
activated, which eventually leads to necrosis.
In conclusion we could show that the foetus is directly exposed to valerenic acid and protopine,
while exposure to hyperforin is minimal. Further studies to determine the potential harm for the
unborn are still needed. An investigation for possible influences of the accumulation of
hypericin in the placenta (change in the viability of the placenta and therefore consequences for
the foetus) also needs to be carried out further since we now know the toxicant potential of
hypericin. Additionally, a study of the identified metabolites and their interactions with the
body are necessary. Finally, an assessment of the full extracts in vitro and in vivo is required
since those single compounds are rarely consumed alone, but as part of complex matrix in the
herbal medication.
With the obtained results we cannot conclude that the investigated compounds or the full
extracts are safe.