Blood-Brain Barrier Neurosciences Sciences

[Neurosciences/BBB] Thiomersal and the blood-brain barrier: where does the science stand?

I decided to write down this article as one follower of this page asked me several questions about the blood-brain barrier (BBB), the scientific rationale of a “leaky” BBB and whether such claims were applicable for delayed vaccines schedule, with a special emphasis on ethylmercury and ethylmercury precursor “thiomersal” or “Thimerosal”.

Following a series of discussions through social media messaging, I thought it would be good to compile the science behind it and share the information.


1.What is the difference between inorganic mercury and organic mercury and why we used it for vaccines?

The first part of this post will focus on setting the science on mercury chemistry, understand the different forms and why we used one form of it inside vaccines.

Mercury (Hg) is considered as a heavy metal due to its high molecular weight (~200g/mol). It is also a particular type of metal called transition metal and it is known for its particular feature of being a liquid metal at room temperature.

Mercury exists as different ionized forms, such as Hg+ (as a bimercury form +Hg-Hg+) and Hg2+, the latter being the most commonly found, in particular as the inorganic mercury form HgCl2.

In addition to its inorganic form, mercury is also found as an organic form, either chemically bound to a methyl (CH3-) group (methylmercury or CH3-HgCl) or to a ethyl (CH3-CH2-) group (ethylmercury or CH3-CH2-HgCl).

Because of its bactericide activity, mercury has been used for a long time in medicine, with use of inorganic form of mercury (cinnabar) in Traditional Chinese Medicine. It was also used as a treatment of syphilis (from the 15th century until the discovery of penicillin) and as a dental amalgam for both its bactericidal properties and its compliance and longevity to the buccal environment compared to resin-based fillings.

Various organic formulation of mercury has disinfection of superficial wound (merbromin, also known as Mercurochrome) and as a preservative for vaccines (thiomersal or thimerosal, both names refers to the same compounds). The structure of thiomersal is the following:


Inside the body, thiomersal is metabolized into ethylmercury:


as well as elemental mercury (as HgCl2 form)

Elemental mercury (Hg), as any other heavy metals (lead, cadmium….) is highly toxic as it has a long biopersistence (it stays a long time inside the body) and by its chemical properties.

Inorganic mercury (as its HgCl2) is considered as the most toxic heavy metal, and classified as highly to extremely toxic (LD50 = 1mg/kg in rats via oral route, source: Methylmercury LD50=24-30mg/kg) is higher than inorganic mercury (LD50=23-30 mg/kg with some age variability in rats, with the highest tolerance noted in young rats  (Lin, Malaiyandi et al. 1975), whereas ethylmercury has an LD50 of 40mg/kg ( Thimerosal has the highest LD50 of all of them, with an LD50 of 70mg/kg (

Inside the body, thimerosal breaks down into various species but classically follows these metabolic steps as recently documented by Carneiro and colleagues (Carneiro, Oliveira Souza et al. 2014). In this study, the authors have injected a single intramuscular dose (20micrograms) of thimerosal and followed the accumulation inside different organs as well as the fate of it. They found that thimerosal rapidly accumulated into the kidney (30 minutes after injection) as ethylmercury and inorganic mercury forms. 70% of thimerosal metabolites were found in the kidney and reported the lowest concentrations in the brain. They also reported some pharmacokinetic parameters including half-lives (the time you need to remove 50% of a chemical out of your body) of 10.7 days and 45.8 days for the brain and kidney tissues respectively.

Thimerosal has been in vaccines since the 1930’s following the tragic death of 12 children during an diphtheria immunization campaign in 1928. Such deaths were due to tainted batch of vaccines that were preservative-free and following an expert panel, such lot was contaminated by a live strain of staphylococcus. ( Thiomersal concentrations in vaccines have been variable, with reported values oscillating between 0.001% to 0.01% (0.01g thimerosal/100mL solution). At such concentration, a vaccine shot (0.5mL) contains about 50 micrograms of thimerosal (~400g/mol) which represent about 25 micrograms of mercury element (~200g/mol). A common practice in modern medicine is to phase out and replace older generations with newer generations of medicines that have shown a better efficacy and/or lesser side effects.

Thiomersal has been used for almost one hundred year in vaccines with no reported fatalities or morbidities with the exception of the retracted Lancet article from Andrew Wakefield that was claiming a causal link between autism and MMR vaccines (the original publication can be viewed here:

There are several flaws in this studies that are too lengthy to explain in this post and that have been already refuted many times by different studies. However, for a lay person one important criteria in a scientific study involving population study is the sample size (or n). It tells you how many individuals have been enrolled or followed in this study. In the Wakefield study, this number was 12. This is a very small number, especially if you are using it to make bold statements. There is a strong bias with small numbers because you may have a very homogenous group based on their age, gender, ethnicity and geographical localization. These different parameters matter when you deal with statistics and try to identify if a phenomenon or feature in a population is real or it if it just some background noise.

To give you an idea, the latest meta-analysis that investigated the relationship between autism and MMR vaccines by Taylor and colleagues (Taylor, Swerdfeger et al. 2014) in which by comparing studies done in over 1’000’000 children (in other means 100’000 times the sample size of the study by Wakefield) failed to find any association between thimerosal and autism, as well as any association between mercury compounds and increased incidence of autism, as well as no association between the administration route and increased risk of autism.

Why this matters? Because size matters. What can maybe true in a small population may not be true when you are expanding the population size and diversify the population studied. This is a common feature that happens in clinical trials. Clinical trials are the crucial steps for assessing the efficacy of a novel medicine. It is divided into three phases. In Phase I, we assess the toxicity and side effect of a drug candidate as well as its fate inside the human body, usually we are talking about a small number of healthy volunteers (n=12-20). This is usually following pre-clinical studies done in animals that already provided us with some information. If there is any severe side effects noted, the drug is immediately phased out.

After assessing the safety, clinicians proceed to Phase II trials in which a small number of patients are recruited and split into two groups: placebo (sugar pill) and treatment (drug candidate). Note that most of the case neither the patient nor the physician know which group they have been assigned. This is what we refer to the double-blind study. Only when the experimentation timeline is completed that the identity is revealed. Upon conclusive Phase II, a drug candidates move into Phase III in which we have now hundreds of patients distributed worldwide that receive the drug candidate and compare to placebo. This is where often a promising drug candidate that have shown some encouraging results in Phase II fails to show the same effects when the population size is increased.

This is why sample size matters, it allows to increase the statistical power to sort real and predictive effect from just pure luck or from a lucky strike.

2. The “leaky” blood-brain barrier in the newborn: fact or fiction?

A common thematic that comes from certain anti-vaxxers or those promoting a delayed vaccination schedule is that the newborn and infant blood-brain barrier is not fully mature and leaky, thus may present an increased risk to absorb and retain mercury or any other chemicals included inside vaccines formulation.

So, in this section we will dive deeper into that question and see what the current science says about it. For this section, I will rely on existing reviews on that topic that have been published in the last five years (Liebner, Czupalla et al. 2011, Ek, Dziegielewska et al. 2012, Lippmann, Al-Ahmad et al. 2013, Obermeier, Daneman et al. 2013).
Before we dig into the concept of developmental blood-brain barrier, it is important to define what is a blood-brain barrier (BBB).

The blood-brain barrier (BBB) is a component of the neurovascular unit, mostly found in brain microvessels.

As illustrated in the picture below, the BBB is formed by a multicellular unit formed by specialized brain microvascular endothelial cells (BMECs) lining the brain blood vessels. These BMECs makes tightly firmed cell-cell junctions, as noted by the presence of tight junctions (TJ) complexes that strictly restrict the diffusion of water, ions and small solutes between the blood tissue and the brain compartment.

For molecules to enter the brain, they have to be either very lipophilic (soluble in fat and oils) or have a set of carriers that can transport in and out the two compartments.

Even being a lipophilic compound is not sufficient to guarantee a transport as there is an array of drug efflux pumps that highly limit the diffusion across the BBB. We estimate about 95% of drugs cannot cross the BBB.

The BBB is a feature we refer as induced as pioneering studies using in vivo (animal studies) and in vitro  (Janzer and Raff 1987, Tao-Cheng, Nagy et al. 1987) have shown that astrocytes can induce such barrier properties in both endothelial cells isolated from brain microvessels and from non-brain microvessels. For a long time, we thought that astrocytes were the only key players in the induction of the barrier function. In addition, other studies that investigated anatomical and functional features noted changes between brain microvessels from embryonic stage compared to post-natal stage animals (Senjo, Ishibashi et al. 1986, Butt, Jones et al. 1990, Keep, Ennis et al. 1995, Kniesel, Risau et al. 1996) suggested that the fetal BBB was probably immature and that it gets the maturity only during late in gestation, when astrocytes differentiate from their precursor cells, undergo maturation and wrap brain microvessels with their astrocytic end-feet process.

However, more recent studies and techniques using genetically engineered animals (knock-out mice) found that indeed such BBB may be already present as pericytes, another cell types lining the vasculature has shown to play a role as important as astrocytes. Unlike astrocytes, these pericytes are already present during development much earlier than astrocytes (Braun, Xu et al. 2007, Virgintino, Girolamo et al. 2007, Daneman, Zhou et al. 2010, Al Ahmad, Taboada et al. 2011).

By the 3rd trimester of gestation, the developing fetus has already a functional BBB and capable to work as well as the adult BBB, in particular such concept has been extensively discussed by Pr. Norman Saunders (University of Melbourne, VIC, Australia)  that have been discussing that the misconception of a leaky BBB in developing brain was the consequence of inappropriate or flawed techniques that wrongly brought such conclusions (Saunders, Ek et al. 2011, Ek, Dziegielewska et al. 2012, Saunders, Liddelow et al. 2012, Saunders, Dreifuss et al. 2014, Saunders, Dziegielewska et al. 2016, Saunders, Habgood et al. 2016).

3. Mercury and the blood-brain barrier: what do we know?

As we have been discussed now, mercury exists into different forms and we have a functional BBB in both the newborn and the adult. So our next step is to understand and use the existing literature to assess the ability of mercury to cross the BBB.

In this post, we are discussing about thiomersal and its two major metabolites: ethylmercury and inorganic mercury. Another form of organic mercury that is problematic and see often and wrongly associated to ethylmercury is methylmercury.
Unlike ethylmercury (CH3-CH2-HgCl), methylmercury (CH3-HgCl) is a form of organic mercury that has serious toxicity issues.

Methylmercury is formed by bacterial metabolism from the biotransformation of inorganic mercury present in the environment into methylmercury. In particular such transformation is occurring preferentially in aquatic systems (Ullrich, Tanton et al. 2001). Unlike ethylmercury, methylmercury is a potent environmental pollutant from industrial processes that have shown ability to bioaccumulate in fish through the food chain and by the relative important half-life of methylmercury (~72 days in aquatic environment).

The uptake of methylmercury occurs through the L-system amino-acid transporter-1 (LAT-1, SLC7A5) both at the human intestinal wall and at the BBB (Wade and Brady 1981, Mori, Yamamoto et al. 2012).

An important caveat that needs to be noted is the relative paucity of the literature in regards of the BBB uptake of the different forms of mercury. Such caveat maybe explained by the relative poor in vitro models of the BBB lacking the tightness needed for studying the diffusion of metals across the BBB.

Two studies worth citing and further discussion: the study from Toimela and collagues (Toimela, Maenpaa et al. 2004) using a co-culture models based on immortalized cell lines and the study from Lohren and colleagues  co-authored by Pr. Hans-Joachim Galla (University of Muenster, Muenster, Germany) (Lohren, Bornhorst et al. 2015) that investigated the effect of mercury on the blood-CSF-barrier (CSF: cerebrospinal fluid). The blood-CSF-barrier (or BCSFB) is the second barrier of the central nervous system located in the choroid plexus.

In the study of Toimela and colleagues using RBE4 monolayers (an immortalized rat brain endothelial cell line), the authors found an impaired BBB function (as marked by increase in fluorescein permeability) for concentrations of HgCl2 and CH3HgCl of 100 micromole/L and 0.1micromole/L following 24 hours treatment. Cell toxicity were noted for HgCl2 and CH3HgCl on SH-5YSY (an immortalized human neuroblastoma cell line) concentrations of 10 micromole/L and 1 micromole/L respectively (24 hours incubation). Toxicity of HgCl2 and CH3HgCl were noted for values of 10 micromoles/L for both respectively.

In contrast, the study by Lohren and colleagues investigated the effects of various concentrations of elemental mercury (HgCl2), methylmercury (CH3HgCl) and ethylmercury on primary choroid plexus epithelial cells. Following an incubation of 72 hours in presence of different concentrations, the authors found a decrease in cell viability for HgCl2, CH3HgCl and CH3CH2HgCl for concentrations of 40, 6 and 2 micromoles/L.
To put into context, we have to remember that the average vaccine content in thimerosal is 50 micrograms for an average volume of 0.5mL.

If we consider that the intramuscular injection yields a 100% bioavailability (all the thimerosal injected ends up in the systemic circulation) and consider such injection in a newborn (~200mL blood), the final circulating maximal circulation following a vaccine shot would be 50/200=0.25microgram/mL or 0.25 mg/L or (0.25/400) = 0.00625 mmol/L, that is equal to 0.625micromol/L. In the worst case scenario situation (one-shot injection in a newborn with a consideration that thiomersal remains non-metabolized), the peak plasma concentration is still three times less than those shown to have in vitro. Now you have to remember, thiomersal is rapidly degraded into ethylmercury and elemental mercury.

Now another interesting piece of data from the study is the relative absence of diffusion of elemental mercury across the BCSFB and the presence of an active efflux for both methylmercury and thiomersal. In other words, the BCSFB may act as a drainage system for both thiomersal and methylmercury, diminishing their probability to accumulate inside the brain tissue.

Finally the last study that is worth discussing in another study from Lohren and colleagues (Lohren, Blagojevic et al. 2015) using  immortalized human neuronal cell line LUHMES ( and astrocytes isolated from a glioblastoma multiforme (primary brain tumor patient CCF-STTG1 ( This is one major point of criticism is how reliable can be tumor derived cell lines to assess cell toxicity and viability (unless you are screening for novel anticancer drugs).

Again, this study supports that organic mercury have a higher cell toxicity than the elemental mercury, with about 10-fold increased toxicity in thiomersal compared to elemental mercury. Toxicity values were in the range of 1 micromole/L for neurons and 8 micromole/L for astrocytes. You have to keep in mind such toxicity was achieved for a prolonged (24 hours) and direct insult. As we have discussed, the amount of thiomersal or methylmercury reaching these neurons and astrocytes is surely much less as we know by now that both the BBB and BSCFB hinder the penetration of both molecules.

Thus the main source of concern comes from methylmercury rather than thiomersal or ethylmercury. Again, both studies failed to measure the toxicity of ethylmercury, that is the major thiomersal metabolite.

If expecting mothers and new parents would have to worry about, it is certainly methylmercury toxicity. Therefore, the primary source of methylmercury uptake that is of concern is uptake from fish and its effect on health. In particular, methylmercury is a legitimate cause of concern for gestating women and for infants.

The Food and Drug Administration (FDA) in collaboration with the Environmental Protection Agency (EPA) has written down an exhaustive report on the current situation of the impact of methylmercury on the brain development in the US and in other countries, such report can be downloaded here (
A take home message from the report are multiples. The average methylmercury concentration in US fish consumption is about 0.072 ppm (particle per millions), that’s the equivalent of saying 0.072mg/L or 72micrograms/L. The CDC upon EPA recommendation set methylmercury levels for drinking water to 2 parts per billions (ppb) or 2micrograms/L and from fish and seafood to 1ppm or 1mg/L (source: Unless someone eat astronomical amount of fish and seafoodon daily basis, the levels of methylmercury should not be a concern for expecting mothers and for infants.

4. Conclusion and summary

In this post, I have been covering the possible neurotoxicity of thiomersal on brain development. Several take home messages to take from this post:

1) Thiomersal introduction about 100 year ago was crucial to ensure vaccine safety and sterility, but as science progress thiomersal became an obsolete preservative and as early as 1997, the FDA was developing a framework to phase it out from the market. As today, only certain forms of dTaP vaccine has thiomersal. If thiomersal

2) The association between thiomersal and neurodevelopmental disorders as suggested by Andrew Wakefield has been since refuted by a number of meta-analysis. Of course, anti-vaxxers will not change their mind in front of evidence and either called a conspiracy or decided to move their goalpost to other vaccines components (such as the aluminum used in adjuvants, that will be discussed in another post, or formaldehyde).

3) Contrary to some logical fallacy, the concept of “leaky” blood-brain barrier in infants is not supported anymore. This concept was considered in the light of science about 30 years ago but since modern techniques refuted this concept. Therefore, using such outdated concept to refuse or delay a vaccination schedule is not only fallacious but clearly underlie a serious issue of continuous medical education (CME) from healthcare professionals proponent of refusing vaccinations or promoting a delayed schedule (CME as part of an important continuous education to ensure healthcare professionals are up-to-date with scientific knowledge, such CME are earned as credits by attending national and international conferences and symposia, with a minium number of credits to be fulfilled within a certain period).

4) The nature of the mercury discussed is very important: elemental mercury is not methylmercury. Methylmercury is not ethylmercury. Thiomersal has shown some toxicity in vitro but it remains important to put into context. However, pharmacokinetic data suggest that thiomersal is rapidly degraded into ethylmercury and inorganic mercury, the latter being the least neurotoxic), as well as the presence of defense mechanism at the BBB and BCSFB.

5) If there is one form of mercury that is non-occupational that expecting moms should worry about is the methylmercury found in fish and seafood. However, the beneficial gains of such food category rich in poly-unsaturated fatty acids (PUFAs, such as omega-3 and omega-6) is essential for ensuring the proper fetal brain development. A equilibrated diet with the advice of the reputable dieticians and physician will ensure the optimal use of fish and seafood without having to worry about detrimental effects due to methylmercury.

 5. References

Al Ahmad, A., C. B. Taboada, M. Gassmann and O. O. Ogunshola (2011). “Astrocytes and pericytes differentially modulate blood-brain barrier characteristics during development and hypoxic insult.” J Cereb Blood Flow Metab 31(2): 693-705.

Braun, A., H. Xu, F. Hu, P. Kocherlakota, D. Siegel, P. Chander, Z. Ungvari, A. Csiszar, M. Nedergaard and P. Ballabh (2007). “Paucity of pericytes in germinal matrix vasculature of premature infants.” The Journal of neuroscience : the official journal of the Society for Neuroscience 27: 12012-12024.

Butt, A. M., H. C. Jones and N. J. Abbott (1990). “Electrical resistance across the blood-brain barrier in anaesthetized rats: a developmental study.” J Physiol 429: 47-62.

Carneiro, M. F., J. M. Oliveira Souza, D. Grotto, B. L. Batista, V. C. de Oliveira Souza and F. Barbosa, Jr. (2014). “A systematic study of the disposition and metabolism of mercury species in mice after exposure to low levels of thimerosal (ethylmercury).” Environ Res 134: 218-227.

Daneman, R., L. Zhou, A. A. Kebede and B. A. Barres (2010). “Pericytes are required for blood-brain barrier integrity during embryogenesis.” Nature 468(7323): 562-566.

Ek, C. J., K. M. Dziegielewska, M. D. Habgood and N. R. Saunders (2012). “Barriers in the developing brain and Neurotoxicology.” Neurotoxicology 33(3): 586-604.

Janzer, R. C. and M. C. Raff (1987). “Astrocytes induce blood-brain barrier properties in endothelial cells.” Nature 325: 253-257.

Keep, R. F., S. R. Ennis, M. E. Beer and A. L. Betz (1995). “Developmental changes in blood-brain barrier potassium permeability in the rat: relation to brain growth.” J Physiol 488 ( Pt 2: 439-448.

Kniesel, U., W. Risau and H. Wolburg (1996). “Development of blood-brain barrier tight junctions in the rat cortex.” Brain Res Dev Brain Res 96: 229-240.

Liebner, S., C. J. Czupalla and H. Wolburg (2011). “Current concepts of blood-brain barrier development.” Int J Dev Biol 55(4-5): 467-476.

Lin, F. M., M. Malaiyandi and C. R. Sierra (1975). “Toxicity of methylmercury: effects on different ages of rats.” Bull Environ Contam Toxicol 14(2): 140-148.

Lippmann, E. S., A. Al-Ahmad, S. P. Palecek and E. V. Shusta (2013). “Modeling the blood-brain barrier using stem cell sources.” Fluids Barriers CNS 10(1): 2.

Lohren, H., L. Blagojevic, R. Fitkau, F. Ebert, S. Schildknecht, M. Leist and T. Schwerdtle (2015). “Toxicity of organic and inorganic mercury species in differentiated human neurons and human astrocytes.” J Trace Elem Med Biol 32: 200-208.

Lohren, H., J. Bornhorst, H. J. Galla and T. Schwerdtle (2015). “The blood-cerebrospinal fluid barrier–first evidence for an active transport of organic mercury compounds out of the brain.” Metallomics 7(10): 1420-1430.

Mori, N., M. Yamamoto, E. Tsukada, T. Yokooji, N. Matsumura, M. Sasaki and T. Murakami (2012). “Comparison of in vivo with in vitro pharmacokinetics of mercury between methylmercury chloride and methylmercury cysteine using rats and Caco2 cells.” Arch Environ Contam Toxicol 63(4): 628-636.

Obermeier, B., R. Daneman and R. M. Ransohoff (2013). “Development, maintenance and disruption of the blood-brain barrier.” Nat Med 19(12): 1584-1596.

Saunders, N. R., J. J. Dreifuss, K. M. Dziegielewska, P. A. Johansson, M. D. Habgood, K. Mollgard and H. C. Bauer (2014). “The rights and wrongs of blood-brain barrier permeability studies: a walk through 100 years of history.” Front Neurosci 8: 404.

Saunders, N. R., K. M. Dziegielewska, K. Unsicker and C. Joakim Ek (2016). “Delayed astrocytic contact with cerebral blood vessels in FGF-2 deficient mice does not compromise permeability properties at the developing blood-brain barrier.” Dev Neurobiol.

Saunders, N. R., C. J. Ek, M. D. Habgood, P. Johansson, S. Liddelow and K. M. Dziegielewska (2011). “Assessing blood-cerebrospinal fluid barrier permeability in the rat embryo.” Methods Mol Biol 686: 247-265.

Saunders, N. R., M. D. Habgood, K. Mollgard and K. M. Dziegielewska (2016). “The biological significance of brain barrier mechanisms: help or hindrance in drug delivery to the central nervous system?” F1000Res 5.

Saunders, N. R., S. A. Liddelow and K. M. Dziegielewska (2012). “Barrier mechanisms in the developing brain.” Front Pharmacol 3: 46.

Senjo, M., T. Ishibashi, T. Terashima and Y. Inoue (1986). “Correlation between astrogliogenesis and blood-brain barrier formation: immunocytochemical demonstration by using astroglia-specific enzyme glutathione S-transferase.” Neurosci Lett 66: 39-42.

Tao-Cheng, J. H., Z. Nagy and M. W. Brightman (1987). “Tight junctions of brain endothelium in vitro are enhanced by astroglia.” J Neurosci 7: 3293-3299.

Taylor, L. E., A. L. Swerdfeger and G. D. Eslick (2014). “Vaccines are not associated with autism: an evidence-based meta-analysis of case-control and cohort studies.” Vaccine 32(29): 3623-3629.

Toimela, T., H. Maenpaa, M. Mannerstrom and H. Tahti (2004). “Development of an in vitro blood-brain barrier model-cytotoxicity of mercury and aluminum.” Toxicol Appl Pharmacol 195(1): 73-82.

Ullrich, S. M., T. W. Tanton and S. A. Abdrashitova (2001). “Mercury in the Aquatic Environment: A Review of Factors Affecting Methylation.” Critical Reviews in Environmental Science and Technology 31(3): 241-293.

Virgintino, D., F. Girolamo, M. Errede, C. Capobianco, D. Robertson, W. B. Stallcup, R. Perris and L. Roncali (2007). “An intimate interplay between precocious, migrating pericytes and endothelial cells governs human fetal brain angiogenesis.” Angiogenesis 10: 35-45.

Wade, L. A. and H. M. Brady (1981). “Cysteine and cystine transport at the blood-brain barrier.” J Neurochem 37(3): 730-734.


4 replies on “[Neurosciences/BBB] Thiomersal and the blood-brain barrier: where does the science stand?”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s