If you took any chemistry classes ever, you’ve heard that benzene is particularly stable due to its aromaticity. Well, benzene is cute but it pales in comparison to the massive Kekulene. It’s so apolar and insoluble that you can only dissolve it in spicily hot solvents and need to measure your NMR spectra in custom-made solvents at 200 °C. But is Kekulene super-aromatic?
What is Super-Aromaticity?
Before we get ahead of ourselves on super-aromaticity, we need to understand aromaticity. In the 19th century, pioneering chemists started to explore the field by doing all kinds of random reactions. They were puzzled that benzene was so unreactive towards addition reactions even though it was assumed to have a high degree of unsaturation. August Kekulé first proposed the cyclohexatriene structure for benzene in 1865. You might wonder what the big deal is – but remember that at this point in history, structures for compounds were still lacking. In one of his reports, he said that a vision of ouroboros, the snake that eats it own tail, inspired him to think of the mesomeric structure.
The actual fundamentals and rationale behind aromaticity were discovered some 60 years later – with the famous Hückel rule on 4n+2 pi-electrons. His concepts were quite unrecognized for two decades – apparently also due to his lacking communication skills – but his contributions made him a cornerstone of organic and physical chemistry. Looking at molecular orbitals, it became that 4n+2 corresponded to a full set of binding molecular orbitals, resulting in higher stability of aromatic compounds.
Upping the ante, the concept of super-aromaticity envisions that macrocyclic conjugation in large, cyclic polycyclic aromatic hydrocarbons leads to an increased stabilization of the molecules. In 1951, the physical chemist McWeeny postulated the potential existence of Kekulene. In 1965, 100 years after Kekulé’s seminal work, first synthetic investigations were published and the molecule was named in Kekule’s recognition.
How Many Pi-electrons in Kekulene?
The key question for Kekulene was which electronic model best represented its structure: Is it a simple, localized structure consisting of 6 benzene rings? This is what McWeeny postulated already in 1951. Or was it rather a super-aromatic one that is based on two connected annulene rings, blue and purple, that both satisfy the 4n+2 rule?
The implications of this setup are quite significant, and one of them relates to diamagnetic anisotropy. When benzene is placed in an external magnetic field during NMR, its pi electrons circulate in the conjugated plane. This new ‘green’ magnetic field that opposes the external light blue magnetic field. If you are a benzene proton, you will feel the new magnetic field in the same direction. This leads to de-shielding and higher chemical shift in 1H-NMR. The same is true for an outer red proton in [18] annulene, where electrons are delocalized over the whole ring. What changes is however the experience for the inner, blue proton. Here, the induced magnetic field opposes the external one. This shields the proton and leads to a lower chemical shift, even negative in the case of annulene. The implications for Kekulene are clear. If global delocalization is a thing, we should see highly shielded inner protons like we see in annulenes.
Synthesis Of Kekulene (Staab, Diederich)
To answer what Kekulene looks like, we have to make it in the lab and characterize its properties. There’s only so much you can compute. We will first look at the landmark synthesis of Kekulene achieved by Staab and Diederich at the University of Heidelberg. Staab is most known for inventing the CDI reagent for hydroxyl and amine derivatization. Diederich, his PhD student at that time, majorly contributed to our current understanding of supramolecular chemistry and medicinal chemistry.
The synthesis by Staab and Diederich was based on almost 2 decades of work! It starts the nitration of meta-xylene and condensation with benzaldehyde. Next, a large-scale hydrogenation with 1.6kg of starting material in 50L of solvent, reduced the double bond. This set the stage for a nice cascade featuring a Pschorr reaction. It resembles a Sandmeyer reaction as it proceeds via oxidation of the aryl amine to the diazonium. This is reduced by copper, triggering an intramolecular cyclization reaction. The cascade happens on both sides in one pot, but yields the pentacycle in poor yield only.
To add more rings, they bromomethylated the pentacycle via an electrophilic aromatic substitution. Then, the benzylic bromide was converted into a thiol in two steps through nucleophilic substitution with thiourea and basic cleavage of the adduct.
To build the Kekulene scaffold, they coupled the two halves through double nucleophilic substitution. This reaction was performed under high-dilution conditions with only 1 mM concentration. This favored intramolecular closure to the ring instead of successive intermolecular reactions and led to a high yield of 60%.
The sulfur was a useful group to build the scaffold – but at some point, you have to remove it. To prepare this, the team methylated the dithiacyclophane, and then subjected it to base-mediated stevens rearrangement. This led to isomers of ring-contracted thio ethers.
Sulfur-ExtrusionReactions
To remove the sulfur completely, the team looked at different methods. The first approach was methylation to the sulfonium which is a leaving group, and can thus be eliminated. However, they had to explore another route as by-products were very difficult to separate from the desired product. This was accomplished by oxidation to the disulfoxide which was then pyrolysed at 450 °C.
As a side note, they also looked into the exotic Ramberg-Bäcklund contraction to give the olefin. It proceeds via alpha-halogenation of the sulfone and subsequent intramolecular substitution – leading to a three membered ring that can again eliminate SO2. It’s quite a miracle they even managed to isolate the 2 mgs or 1% yield of this reaction by preparative TLC. Obviously, they opted for the other approach instead of this one.
Finally, the product was photo-cyclized to yield octrahydro Kekulene. Here they found that using the saturated starting material was vital for success. They initially tried to install the double bond first and then perform the cyclization last, but this was unfruitful. Probably, creating the fully planar system makes it too rigid – instead, the aliphatic CH2 groups add some flexibility that is needed to enable photochemical reactivity.
Last, DDQ oxidized the octahydro derivative. Just to solubilize the reagents and achieve the reaction, they used trichlorobenzene as a solvent and let the reaction run for 3 days at 100 °C. They found Kekulene to be so insoluble that they had to recrystallize it from boiling, 400 °C hot triphenylene by “slowly” cooling to 300 °C – whatever slowly means here.
Is Kekulene super-aromatic?
Growing Kekulene crystals allowed them to investigate its molecular structure. They found very low variation in bond lengths, even for the inner protons. By looking at bond lengths, they derived that on the basis of X-ray analysis, Kekulene appeared not to have globally delocalized electrons and super-aromaticity.
What about NMR? Well, due to its low solubility, the team had to resort to creating deuterated trichlorobenzene and recording NMR spectra of saturated solutions at 200 °C to get anything usable. When they finally measured the compound, they showed the inner protons to be extremely de-shielded. Basically, the contrary of the delocalized annulene example we talked about at the start. The protons look like benzene protons, suggesting that the super-aromatic structure was incorrect.
Modern Synthesis Of Kekulene
So much for the oldschool chemistry. Remember the initial Pschorr reaction with low yield? The team of Perez envisioned a clever Diels-Alder short-cut by using this commercially available bistriflate. Addition of fluoride triggers elimination towards a triple bond, which can engage in a 4+2 cycloaddition with styrene to form a six-membered ring. After re-aromatization and release of the second triflate, another Diels-Alder reaction yields two isomers – one desired “cis”-like isomer, and a trans one. After some optimization, they managed to increase the yield to 28% with a 2:3 mixture of cis to trans. This means the net yield is also just around 11% – but you save yourself all other steps in the beginning of the synthesis.
Staab’s and Diederich’s synthesis truly stood the test of time, as the team also looked into synthetic alternatives after this step. It seems even modern methodologies could not improve or even re-create the original synthesis. The point of the Perez team by the way was to perform ultra high-resolution atomic force microscopy – basically making nice pictures of single molecules.
So, is Kekulene super-aromatic? Based on their findings and calculations, they also concluded that Kekulene does not have delocalized pi-electrons, and that the Clar model with 6×6 pi systems is the most appropriate one.
Chemists create even funkier Kekulene-like molecules like Septulene (google it’s structure!). But this is where we will stop for today. Catch you in the next one!
Watch the video on YouTube or read the written blog!
Molecules take very intriguing forms. You might know cubane, the literal molecular cube. Now, scientists recently reported the synthesis and characterization of 1-azahomocubane, one of the first of hopefully many more cubane analogues. Similar to weird species like Dewar benzene, these molecules can offer chemists new insights into effects and limits of ring strain.
Today we’ll start with a trulyelegantoldschool synthesis of simple cubane, and then dive into azahomocubane’s modern synthesis and properties. Doing so, we also learn: 1) how people make the absolutely mad octanitrocubane; 2) how chemists can use the unstable antiaromatic cyclobutadiene as a reaction partner; 3) why azide groups are very useful for rearrangement reactions and last, 4) why you should never throw away your old products.
Hard to believe, but cubane was first synthesized already in 1964. Philip Eaton, part of the first cubane synthesis, continued to be a driving force of cubane research. Almost 4 decades years later, he also led the first synthesis of octanitro-cubane. As you can imagine, this required some unpleasant reaction conditions – and of course, this thing goes boom. The introduction of nitro-groups worked via so-called interfacial nitration of a cubyl anion at the melting interface of frozen THF and N2O4. I’ll gladly pass on this. Also, the last step requires you to think: “mhm ah yes, let’s do an ozonolysis of a nitrosyl-heptanitro-cubane”.
Octanitrocubane was hypothesized to a potentially best-in-class non-nuclear explosive based on theory – but its experimental density was shown to lower. There are no public records of larger-scale synthesis and testing, so it’s just researched from a computational point and remains elusive.
Back to normale cubane – mindblowingly, just after 2 years, an incredibly efficient synthesis was published. As promised, it uses cyclobutadiene as a starting material, which seems crazy but also very logical given cubane is full of squares. You will know that this is an anti-aromatic compound and very unstable, so it is not possible to store it in a bottle. However, oxidizing this iron complex, which can be handled better, leads to release of cyclobutadiene within the reaction vessel – in this case, allowing for a [4+2] cycloaddition with this diketone which bears two bromo groups and another olefin – both essential to this efficient synthesis.
Note: The Favorskii reaction proceeds in a “quasi-Favorskii” mechanism. This is because the alpha position is not enolizable due to the high ring strain of the bridged system (Bredt’s rule).
After the endo cycloaddition, the two double bonds are well positioned to engage in a [2+2] photoaddition, obviously one of the key reactions available for construction of cyclobutane rings. Now we have this basket-like intermediate, which can undergo a base-mediated double quasi-Favorskii reaction, leading to ring contraction of the cyclopentane rings. This creates the cubane system, and now you can simply remove the acid groups in some step via decarboxylation. The yields are suspiciously consistent at 80% but even if they would be lower, it would not take away from the nice sequence and use of cyclobutadiene.
Very timely, there was a super recent publication on another, metal-free way to liberate cyclobutadiene – quite nice work and much better than creating some toxic iron compounds. This works via a retro [2+2] addition to release nitrogen and create reactive cyclobutadiene. [J. Am. Chem. Soc. 2023, 145, 10, 5631]
So the little cubane boy is done – but how do we create the much more complicated azahomocubane? Well, a logical intermediate would be a simple cubane-amine. Because it’s not 1964 anymore, we don’t have to create cubanes bottom-up, rather, we can simply buy them because there are crazy folks performing cubane synthesis on kilo scale.
This di-ester-cubane is not cheap and although it comes with the cubane, it also has two esters. Because we need the mono-substituted amino-cubane, we need to get rid of one of them. This is achieved via hydrolysis with just one equivalent of hydroxide and subsequent Barton-decarboxylation of the activated acid.
So, how do we get the amine from the ester? After hydrolysis, the acid was converted to the azyl azide. This prepares the Curtius-rearrangement which is facilitated by the strong energy gain of N2 release. The intermediate isocyanate can be trapped with various nucleophiles, in this case, with tert-butanol to give the N-Boc carbamate product.
Simple acidic exposure releases the cubyl amine. What now? It’s azide time again! Exposure to triflyl azide, a terrifyingly reactive azide transfer reaction, creates cubyl azide via substitution. Both of these azides come with a hefty explosion warning, so they are best handled only in solution and behind a blast shield. Adding some acid now, you guessed it, leads to another Curtius-like rearrangement with migration of the alkyl rest to displace N2. The resulting carbocation is then trapped by the acetic acid to give this product. After all of this, we now have a single nitrogen in the correct oxidation state.
Because we need the nitrogen in a corner position, there is more to be done. Once again, let’s do a sweet rearrangement. N-chlorination creates the opportunity for a new C-N bond formation, initiated by the fragmentation of the acetate. This is reminiscent of the Favorskii ring contraction we saw in normal cubane’s synthesis, just this time, the halogen leaving group is bound to nitrogen. The intermediary amide can be hydrolyzed and voila, we finally positioned nitrogen in a cyclobutane. Now, we just need to create the final linkage of the extra carbon and nitrogen. We have the ester carbon to work with, so the final steps used the methyl ester. Now it’s just a simple orgo freshman sequence of ester reduction, chlorination and intramolecular SN2 to get to aza-homocubane. As this final product is volatile, they prepared the salt form for easier handling.
So does azahomocubane go boom like octanitrocubane? No, clearly not – just because it looks strained, does not it can release nitrogen or CO2 as part of an explosion. It’s decomposition is much less exciting actually, exposure to acid or simple storage in the refrigerator led to ring opening of the cubane. This step was irreversible, a testament to the high instability of azahomocubane. It also makes you think that the last SN2 with the primary leaving group is probably one of the only reasonable ways to create the system in the first place.
In terms of geometry, the team obviously expected the product to be different than your generic tertiary amine. DFT calculations indicated that this was definitely less than ideal sp3 geometry, and the crystal structure – funnily found via serendipitous discovery – was consistent with this. You can see that the five-membered ring distorts the picture quite a bit, so homocubanes look more like baskets or houses, instead of cubes – if that makes sense.
What is the impact on basicity? The 1-azahomocubane nitrogen is not happy with strain energy being an order of magnitude higher – but this means that basicity is more than 10-fold lower! You might think that because there is more strain, the nitrogen might be happier to be present in some other configuration. However, looking nitrogen NMR chemical shift analysis showed that the nitrogen in azahomocubane is less electron rich compared to the other frameworks, which is aligned with the basicity trend. Notably, the hypothetical azacubane has 45 kcal/mol more strain energy than azahomocubane, which makes you wonder how long it will take to synthesize it in the lab.
Finally, they looked into hypothetical atom exchange and hypohomodesmotic reaction calculations. What is that? It’s basically a nerdy theory-crafting method of computing bond separation and formation energies, and the name is due to different sets of reaction conditions qualifying for different reaction types. For example, the hypohomodesmotic reactions shown here can be used to compute strain in cycloalkanes. Basically, you’re taking heats of formation of the individual molecules – which are known values – and figure out what is the enthalpy value Q needed to balance out the hypothetical reaction. You can see that while cyclohexane has a Q value of almost zero – because it is not strained at all – Q is much higher for cyclopentane, cyclobutane and cyclopropane as strain increases.
Applying this methodology to our question, we see that azahomocubane is significantly more stable than all-carbon homocubane. But why is that?
To shed some light on this, they modelled nitrogen lone-pair sigma star interactions with the carbon framework. This revealed that there is substantial hyperconjugation from the nitrogen with both adjacent cubic and basket handle C-C bonds, somewhat stabilizing the system compared to a normal homocubane. Both hyperconjugative effects and general orbital re-configuration are at play. For example, Bent’s rule describes orbital re-orientation once you add in electronegative substituents like fluorine or nitrogen. In such settings, orbitals with s character are pointed towards more electropositive substituents like hydrogen or carbon, so this leads to some changes in bond geometry and lengths.
This work doesn’t answer all our questions, but it shows that aza-variants of strained systems will be interesting playground for chemists. Catch you in the next one!
Did you know that the active chemical in magic mushrooms psilocybin might rewire and make the human brain more flexible, translating into major therapeutic benefit in depression and other conditions? This has led to Australia approving specific medicinal use of psychedelics as the first country.
You will be surprised to hear that psilocybin can be synthesized in just 4 steps leveraging quite simple chemical reactions. It was the legendary Albert Hofmann, who synthesized it first, and just like with his discovery of LSD reported a self-experiment. This didn’t just make him trip big time, but also unveiled some salient characteristics this molecule.
This story is not commonly talked about, so I hope you are in the mood for some 60-year-old German texts. Oh, and just like heroin, psilocybin was sold as a legit pharmaceutical (lol). If you enjoyed the discussion of history and syntheses of psilocybin, LSD, THCP or ibogaine, you will like this one too!
Today we will 1) review psilocybin’s richhistory, chemical structure, and mechanism of action, 2) look at its chemical synthesis, 3) dive into the scientific evidence behinds its therapeutic potential, and 4) last, explore its effect on brain networks.
History and Context
Psilocybin is a so-called tryptamine alkaloid found in a variety of mushrooms, but most potently in the genus Psilocybe. Due to its hallucinogenic and mystical effects, human use of psilocybin for medicinal and religious purposes dates to pre-historic times. Some notable artefacts include a Spanish cave painting – though maybe this person simply liked champignons – or more telling mushroom figurines and even statues from pre-historic Columbia. Supposedly, the mushrooms were called “God’s flesh” in an Aztec language, so you can be sure they were tripping big time.
Supposedly, the oldest indication of human use was found in an Algerian cave, estimated to date back to an insane 7000-9000 BC. It shows what we probably all drew at one point in our life: transcendental, bee-faced mushroom-shamans! Jokes aside, do you think that the interpretation on the far left is sensible compared to the original? Let me know in the comments! Some folks argue that it’s very easy to over-interpret these types of drawings and immediately think that magic mushrooms are encrypted everywhere. Also, don’t forget what exposure to the elements over thousands of years can do. What looks like a mushroom to us, might have represented something else. By the way, the esoteric drawings in the Algerian cave naturally make it also a prime site for ancient aliens enthusiasts.
Given Europeans prohibited mushrooms and other non-alcoholic intoxicants in the 16th century, their use was driven underground – so much so, that during the early 20th century western academics weren’t even sure if psychoactive mushrooms existed at all. They didn’t see the convincing Algerian shamanic cave yet, so they are excused. This changed with Gordon Wasson, a banker at JP Morgan and investigational mushroom enthusiast, who took some trips to remote Mexico with his wife in 1953. Albert Hofmann wrote that it took more than one trip for Wasson to participate in a mushroom ceremony instead of just watch, and that he might have been one of the first white folks to ever take them. A few years later in 1957, Wasson published the first ever broadly distributed article on magic mushrooms. I invite you to look through it yourself – the link is in the description. Not only is it nice to read through his personal expeditions and experiences; but the flipping through this 70-year-old magazine itself is quite hilarious, partly because it has nice oldschool ads for things like canned meatball pasta – or the inviting flavor feast of canned pork and beans. In 1957, the chemistry and biology behind mushrooms was still a mystery – but even though Wasson thought chemists have a long road to go for isolation and synthesis, this would change very soon.
Wasson was connected to the botanist Roger Heim, who as Wasson wrote, was a man with vast experience in the field of mushrooms. Heim managed to grow the mushrooms in his laboratory in Paris – so he’s like the founding father of shroom farming – and shared samples with the Swiss pharma company Sandoz where the legendary Albert Hofmann was still active, having already discovered LSD more than 10 years earlier in 1943. Hofmann soon isolated psilocybin and its active metabolite psilocin – but how?
He first tried something I mentioned in an old video talking about the mind-blowing isolation of 0.35mg of ciguatoxin, a marine natural product, from 125kg of moray eel guts. During that effort, the scientists injected their chromatographic fractions into mice and determined presence of ciguatoxin based on if the mouse would drop dead or not.
Hofmann’s Self Experiment with Psilocybin
After the unsatisfactory animal tests, the temptation got too big. He ate 32 medium-sized mushrooms – talk about not being cautious at all – and the ensuing effects left no doubts regarding the potency of these fungi.
After half an hour, he started to experience some serious sensory disruptions and hallucinations, apparently accepting the fact that his supervising doctor might sacrifice him to some ancients Aztec gods.
His trip peaked 1.5h after eating the mushrooms, and the entire psychedelic dream lasted about 6 hours. Clearly, he came back feeling quite euphoric.
This made him realize that potency of the mushrooms was not the problem, but that animals are less sensitive and clear in terms of their response to psychedelics. You can also see this in today’s research in animals, which uses doses that are usually more than 2-3 times higher than human doses. So, with this confidence and insight, Albert Hofmann consistently performed human testing of extracts to guide the rest of isolation, which worked quite well even though these were diluted samples and not dried mushrooms.
Which of Hofmann’s self-experiments do you like better: the LSD bicycle story or the mushroom dream? Let me know in the comments!
Wrapping up on history before we go to chemistry, Hofmann also rapidly developed the first synthesis of psilocybin in 1959. Thereafter, Sandoz distributed 2mg dosed tablets, nucleating several clinical studies in the 1960s and 1970s on mental disorders and other areas. However, as part of the “war on drugs” to control abuse of psychedelics, psilocybin was classified as a highly controlled Schedule 1 substance. Only after strict governmental controls were somewhat lifted by the 2000s, psilocybin became the subject of clinical investigations again; and recently picked up even more steam with the FDA granting psilocybin a breakthrough therapy designation in 2018.
structure and biochemical effects of psilocybin/ psilocin
Before we talk about these studies, we need to take a closer look at psilocybin. Comparing their chemical structures, you will realize that psilocybin and its active metabolite, psilocin, share the tryptamine core with various other compounds – for example with the endogenous neurotransmitter serotonin which modulates mood, learning and other things in humans; but also the hormone melatonin which modulates our sleep and circadian rhythm, and of course other psychedelics such as DMT. It’s always impressive to see that simple changes in chemical structure lead to massively different physiological mechanisms of action.
But wait, what’s the connection between psilocybin and psilocin? Well, just like aspiring for example, psilocybin is a pro-drug which is rapidly dephosphorylated under acidic conditions in the intestine and liver. After first metabolism, the resulting psilocin is much less hydrophilic and only now can cross the blood-brain barrier. Once in the brain, psilocin is believed to selectively bind to the serotonin 2A receptor.
With psilocin looking quite like serotonin, it engages in similar interactions in the active site of the serotonin 2A receptor. This receptor is highly expressed in certain brain regions and part of a broader family of 14 sub-types of receptors which drive the incredibly broad biological functions of serotonin. Psilocin is believed to get its therapeutic effect from its selectivity for the 2A receptor over all other serotonin receptor types. However, there is also some research showing that affinities for 2A, 2C and 1A receptors are in the same order of magnitude – so clearly, the picture is much more complex.
These receptors are so called G-protein coupled receptors or GPCRs, which if activated lead to an array of complex downstream signaling cascades. These molecular mechanisms, most of them not well understood and involving other neurotransmitters like glutamate as well, result in structural and functional cellular changes which can translate into enhanced neuroplasticity.
Let’s remember, a typical psilocybin session is just 4-6 hours long with a peak of acute subjective effects after 60 to 90 minutes. Due to the high expression of serotonin 2A receptors in the visual cortex, individuals experience visual hallucinations – at high doses, even with eyes closed. This heterogeneous receptor expression in brain regions is something you should remember for later by the way. This is really interesting, because on top of these acute effects, there are these cellular changes we’ve mentioned earlier which happen on a much slower but also more durable timescale.
Neuroplasticity refers to the brain’s ability to change throughout life and can be driven by changes in cell structure but also changes in the efficacy of synaptic transmissions. From these different mechanisms, I just wanted to highlight one study on dendritogenesis.
This team of researchers looked at structural effects of psilocybin on mice. To validate effects and dosage, we once again have the good old head-twitch response. Within the frontal cortex, they found a significant density and size increase of dendritic spines, which are neuron protrusions which help signal transmission across the nervous system. Strikingly, a fraction of these new dendritic spines was still present after a month and seemed no different than normal spines. This is clear evidence for structural change.
But now comes another interesting thing. They pre-treated other mice with Ketanserin. This compound, you can see that its left half looks like a tryptamine analog, is also a strong 2A receptor antagonist which leaves many receptors inaccessible to psilocin activation. This blocking or knockdown is reflected in the lack of a head-twitch response, as you can see here. However, the structural remodeling took place nevertheless. It might be that this happens already at much lower concentrations of psilocybin, or potentially also proceeds via other mechanisms beyond serotonin 2A. Now that we understand psilocybin’s rich history and biochemical mechanisms, we can check out some chemistry.
The research we will talk about is a synthesis of psilocybin by scientists at the Usona Institute which runs early research and clinical trials for psilocybin and 5-methoxy DMT. This is a non-profit medical research organization which is pioneering the application of psychedelics to neurological disorders. They also have an investigational supply program of the psilocybin that they synthesize themselves. Let’s check out how this works.
Organic chemistry: Synthesis of Psilocybin and Psilocin
Psilocybin can be made from psilocin, a slightly simpler molecule, by essentially reversing what would happen in the body. Albert Hofmann’s first synthesis published in 1958 and most large-scale routes accomplish this in two steps. First, psilocin nucleophilically attacks an activated pyrophosphate. Interestingly, one of the benzyl groups hanging on the phosphate oxygens intramolecularly shifts to the nitrogen, creating a zwitter-ionic species. The nucleophilic substitution on phosphorous would have been ineffective without the benzyl or other ester groups, but now a hydrogenation is requiredto deprotect the intermediate to psilocybin. This approach has quite a few challenges, most notably that atom economy is quite bad. Starting from 700g of zwitterion, just 100g of psilocybin can be isolated. In addition, the benzyl migration step to nitrogen is quite poorly understood and behaves in funky ways which makes in-process monitoring tricky. Obviously, you don’t want to start randomly guessing reaction endpoints when you are doing a reaction on kilo-gram scale. This transformation is very dependent on reaction volume and temperature control; the authors highlight than in one reaction “gone bad”, they had to filter the product over 6 days.
Instead, the Usona team discovered a more efficient way. Let’s look at the full synthesis in just 4 steps from the commercially available starting material. 4-acetoxyindole undergoes a regioselective electrophilic aromatic substitution with oxalyl chloride to introduce the two-carbon chain. In a second step, the remaining acyl chloride facilitates an amide-formation with dimethylamine. Adding lithium aluminum hyride at forcing conditions, we exhaustively reduce both carbonyls on the nitrogen chain – liberating the tryptamine – as well as the acetyl group to give the free phenol in psilocin. Now, under optimized conditions, the authors found that using POCl3 followed by immediate hydrolysis introduce the phosphate. They were able to scale this very well, delivering a whopping 1.2kg of psilocybin in one go.
This phosphorylation proceeds by stepwise hydrolysis of the di-chlorointermediate. But as you can see, the reaction setup and work-up is quite complex at scale – and this step remains the most problematic one. For instance, they had to add 3 kilograms of celite powder to prevent the formation of sticky precipitate that messed up the stirring. The hydrolysis itself was achieved by quenching the di-chloro product into a cold THF/water mixture containing excess triethylamine. Keeping the slurry at sub-0 degrees for roughly 60min seemed to work well – in case of longer hold times, they observed decomposition to psilocin again. After some more work-up, they performed a recrystallization to isolate 1.2 kg of highly pure psilocybin.
Selected clinical data on psilocybin
So probably you’ve heard various people, including drug abusers themselves, throwing around the idea of psychedelics as the be-all end-all super medicines. Indeed, there is a growing body of evidence for psilocybin, but the focus is currently on depression – in particular, cancer-related and treatment-resistant depression, as well as anxiety. Most of these studies use one or two 25mg psilocybin doses separated by around 3 weeks; and the treatment is supported by targeted psychotherapy sessions as well – so these patients are not just taking psilocybin to their liking and magically getting better.
Also, all credible research is looking at pure psilocybin, instead of magic mushroom bites – obviously, this way the effective dose of the primary active ingredient and safety for patients can be controlled. You don’t want to give your patient some random mushroom and hope it was some good stuff. However, to give credit to nature where it’s due, there is the so called entourage effect which postulates that the sum of the parts in psychedelics might lead to a greater synergistic effect compared to the individual compounds. Some research has shown that magic mushrooms might be a more potent source of psilocybin than pure psilocybin itself – likely because there are other compounds and psilocybin-derivatives that have synergistic interactions and amplify efficacy.
There is really strong evidence from the largest psilocybin depression study to date published a few months back – and this research has triggered quite some broader public attention on psilocybin. Before we go there, I wanted to briefly showcase that some evidence in other disorders is quite promising as well:
One randomized clinical study compared two cohorts of around 50 heavy drinkers undergoing psychotherapy – compared to a tobacco use study which only had 15 patients, this sample is really not that bad. These black arrows indicate two sessions, where they either received psilocybin or a sedative which served as a quasi-placebo and blinding control. What they found was that the psilocybin group had much lower alcohol use even half a year later, also translating into fewer alcohol-related issues like physical problems or impulse control.
Let’s switch gears now, looking at this large, eye-catching study on treatment-resistant depression. Resistance in depression is a huge problem, as remission rates across courses of therapy drop by a large margin. The success rate of a first anti-depressant course is around 37% – which to be honest is shockingly low – but drops to the mid-teens after 2 or more courses. Once someone failed two different courses, their depression is classified as treatment-resistant.
This Phase 2 study included 233 patients and tested three doses: a high dose of 25mg – which isn’t crazy high really; a medium dose of 10mg and a 1mg dose which they call control – not really a true placebo but probably also too low to show any effect. Because the acute effects of psychedelics are very strong, the value of placebos and double blinding itself is questionable in these trials anyways, as each participant and staff member realize who’s actually tripping or not.
I said it a few times already – psychedelics can also be dangerous or counterproductive in some cases. If you are interested, you can read through the exclusion criteria for this study. Also, the study had participants discontinue anti-depressants and other prohibited medications during a wash-out period.
Looking across the ingoing randomized groups, 60-80% of patients were severely depressed as measured with MADRS depression scores, and around 80% had failed two treatments for their current depressive episode. Notably, less than 10% had already used psilocybin in their life – this might be important because some researchers have hypothesized that previous psychedelic exposure might impact trial enrollment and efficacy on individual patients.
The treatment consisted of a one-time psilocybin dose which was administered in a calm supervised environment, with patients chilling out to a nice playlist and laying around in introspective for a few hours.
Let’s look at the results: The medium dose of 10mg did not bring any changes compared to the low 1mg control dose – however, the 25mg group a mean drop of 12 points in the depression score, which was significantly lower than the other two.
Looking at the evolution over time, we can see steep acute drop followed by a slight rebound; however, the separation of the three groups seemed to persist over time. At 12 weeks, the changes in scores we saw translated into roughly 30% of patients dropping below a score of 10, which means they are depression-free. Considering we are talking about treatment-resistant depression here, this remission rate is very meaningful – also, we are just talking about a single psilocybin dose here.
Looking at safety, the 25mg group had few more adverse events – but this was primarily driven by headaches, nausea, and not serious AEs like suicidal ideation.
I hope this gave you a more detailed look at why some folks are getting excited about psilocybin in depression – but bigger and longer trials are needed to replicate these results at scale, and bring more insights into durability of response.
Neurostructural mechanism behind psilocybin
Finally, let’s play neuroscientists. One potential reason for psilocybin’s activity is connected to large-scale brain networks which govern all our thoughts and actions. There are many different networks depending on what model and definitions you consider, but three are important for us. The default mode network is a set of interconnected brain regions that are active when a person is not engaged in any specific task or actively thinking about something. This network is often referred to as the “resting state” network, as it is active when the brain is not focused on external stimuli – and thought to be involved in self-reflection, future planning as well as social cognition – so thinking about the mental states of other people. The executive network, on the other hand, is involved in goal-directed behavior, decision-making, and working memory. This network is active when a person is engaged in a task that requires concentration and effort, such as solving a complex problem or completing a difficult math equation. The salience network is another important network in the brain, which is involved in detecting and responding to important stimuli in the environment. This network is active when a person is presented with a novel or unexpected environmental or emotional stimulus, such as a loud noise or a sudden movement. These interconnected networks are associated with cognitive control and flexibility, ultimately governing processes like learning and task switching. In depression or disorders exhibiting cognitive inflexibility such as autism or OCD disorders, the interdependency and switching of these networks is impaired.
Research published in Nature Medicine in 2022 showed that two doses of psilocybin with psycho-therapy swiftly decreased symptoms in severely depressed patient, as shown by reductions in the BDI score. Nothing shocking for us – but what’s more interesting is that psilocybin showed strong significant decrease in brain network modularity as shown by fMRI brain imaging. In the bottom right chart, you can see that default mode network activation decreased while its integration with the executive and salient networks increased. In plain English, this kinda made their brain more flexible. They also had a control arm which administered a selective serotonin reuptake inhibitor antidepressant, which lacked this integration-inducing effect.
The researchers postulate that this decreased modularity or increased flexibility of brain networks are key to psilocybin’s mechanism of action. Depression is characterized by abnormally constricted network connectivity – psilocybin might ameliorate this by broadening mental states, in line with the liberating emotions some people feel. This effect differentiates psilocybin from other SSRIs, likely due to its more targeted effect on cortical 5-HT2A receptors which are highly expressed in regions linked to these networks. There are probably some additional mechanisms at play here like reduction of amygdala response, which is another pivotal area implicated in depression.
Another area of interest is the claustrum, a thin sheet of neurons located in the central cortex, which also shows a high level of 5-HT2A receptor expression. In other research, psilocybin was demonstrated to decrease activity or turndown this area which is highly interconnected and involved in setting attention and switching tasks. You can see that there are some different effects from psilocybin on the right claustrum, decreasing connectivity with some networks, but also increasing others. Put simply, shutting down the claustrum might explain why psychedelics help with re-setting rigid though patterns, as well as unveil new psychological insights through an effective re-wiring of brain networks.
A final interesting point is related to different people experiencing different types of trips, if you will. There is some degree of association between brain structure and personality dimension as well as risk factors or pharmacological treatment response. In one 2020 study, researchers found that the thickness of the anterior cingulate, a brain compartment connected to other areas known to be important for emotion or memory, can predict the emotional experience that people get from psilocybin, in terms of feeling strong emotions of bliss, unity or others. This might be one of dozens of other factors that dictate why people react quite differently to psychedelics or drugs.
Thank you for reading and hope you learned something new today!
Key references:
Single-Dose Psilocybin for a Treatment-Resistant Episode of Major Depression | NEJM 2022, 387, 1637
Therapeutic use of psilocybin: Practical considerations for dosing and administration | Front. Psychiatry, 2022, 13, 1040217
Psychedelics and Neuroplasticity: A Systematic Review Unraveling the Biological Underpinnings of Psychedelics | Front. Psychiatry, 2021, 12, 724606
Increased global integration in the brain after psilocybin therapy for depression | Nature Medicine 2022, 28, 844
Percentage of Heavy Drinking Days Following Psilocybin-Assisted Psychotherapy vs Placebo in the Treatment of Adult Patients With Alcohol Use Disorder | JAMA Psychiatry 2022, 79, 953
DARK Classics in Chemical Neuroscience: Psilocybin; ACS Chem. Neuroscience 2018, 9, 2438
Psilocybin acutely alters the functional connectivity of the claustrum with brain networks that support perception, memory, and attention | NeuroImage 2020, 218, 116980
Seeking the magic mushroom | Wasson’s 1957 publication in Life Magazine |
Die psychotropen Wirkstoffe der mexikanischen Zauberpilze (Albert Hofmann) | Verhandlungen der Naturforschenden Gesellschaft in Basel 1960, 71, 239
Research on Acute Toxicity and the Behavioral Effects of Methanolic Extract from Psilocybin Mushrooms and Psilocin in Mice; Toxins 2015, 7, 1018
Watch the video on YouTube or read the blog below!
You probably know that drugs can get very expensive. What do you think – what percentage of new drugs actually provide significant therapeutic benefit compared to existing treatments? And do you think that drugs with a low clinical benefit are indeed priced any cheaper than highly beneficial ones?
Beyond drug prices, we will also talk about cancer “approvals gone wrong”. These were shown only after their regulatory approval to lead to equal or even worse outcomes for patients. One of these debated drugs ultimately led to bankruptcy of the company who developed it. However, the molecule has some nice chemistry behind it. You know the channel – we will take any opportunity to study organic synthesis that we get.
Healthcare Spending and Drug Prices
Let’s face the facts – healthcare spending has grown much faster than gross domestic product – in the US and everywhere else. Drug costs are one driver behind this. The most extreme example is perhaps cancer, where we can observe almost exponential growth of drug costs over the last 50 years.
The US spent $4.2 trillion on healthcare in 2021. Who paid the bill, and on what? As you can see, government spending on Medicare and Medicaid is about 40%, whereas patients paid out-of-pocket for 10% of costs. On the other hand, hospital care received the bulk of spend with 31%, with prescription drugs coming in at 9%. Increasing drug costs are problematic, but as we see, they are not the only reason behind high healthcare spend.
Initially expensive branded drugs are becoming much cheaper after entry of generic drugs. Fortunately, this leads to more accessible, vital treatment options for patients. This is where we have the crux: While standards of care improved across all diseases, new medicines need to continuously drive improvement in therapeutic benefit.
Let’s imagine we have a new drug in a disease that affects the liver, demonstrating it slows loss of liver function by 25%. Sounds good in theory, but let’s assume this was only tested against placebo and already approved treatments have shown similar results in their trials. If we ignore any benefits on things like safety or durability of response, this means the relative therapeutic benefit of this therapy would be rather low. On the other hand, a new drug showing superior efficacy to other already approved treatments would be much better. Instead of “me-too” drugs, we would want any new drugs to significantly improve on existing drugs. The problem is that this has not been the case in past years.
The Therapeutic value of new drugs
The authors of the first piece of research looked at the last decade’s new drug approvals and their ratings of therapeutic value from health technology assessment bodies in France and Germany. These agencies assess the added benefits of a drug’s approved indication compared with existing therapies. This tells us whether there is major, considerable, minor, or no benefit of a new drug. The underlying criteria are slightly different across countries but capture the drug’s effect on reduction of disease duration or side effects, for example. If any of the two agencies rated a drug as providing considerable or major benefit, the authors consider it a high value drug in their analysis.
The sad result is that only 40 to 50% of drugs approved by the US FDA or European EMA regulator are high value drugs. But check out the second set of bars. These refer to approved uses in additional diseases or indications after a drug’s first approval. In these settings, are even less likely to be quote-unquote better drugs.
Drugs with multiple indications
But why can drugs have more than one indication? As most of you know, a single mechanism of action can apply to different diseases. For instance, immuno-oncology drugs like PD1 inhibitors can be applied across solid tumors. Keytruda, the soon-to-be world’s best-selling medicine, is approved in more than a dozen different tumor types. Beyond oncology, we have complement inhibitors like Soliris which you might remember if you watched the previous video on the world’s most expensive drugs. Because the complement immune pathway is involved in many diseases, it is possible that the same drug which saves kidneys might save eyes as well.
On one hand, patients benefit because they get one more option which might help them go into disease remission (not always). It also makes sense for pharmaceutical companies to get rewarded by bringing innovation to more patients. Because clinical development costs billions, they maximize product revenues during their limited window of exclusivity.
But the reality shatters this pleasant theory: Only 40% of drugs in supplementary indications bring significant new benefit. If we look at the relative rate, drugs are less and less likely to bring new therapeutic benefit in every follow-on indication.
Some caveats remain, like questions behind the precise logic of defining something as high therapeutic benefit. There are specific local factors, like France using their own therapeutic strategy as a criterion. But more important is that while a drug might not be high value on average, sub-segments of patients with specific mutations or other characteristics might respond very well. We know who might respond best for some drugs but for most, this is less clear.
Value of Accelerated approvals & Friends
This research confirms what others have found before. An interesting second analysis differentiated the therapeutic value of normal drugs and those with expedited approval. The exception is awarded for drugs which major promise in diseases with high unmet (e.g., certain cancers, rare diseases). The logic is obvious: imagine there is a drug which shows drastic tumor shrinkage in an early phase 2 clinical trial. Instead of waiting 3-4 years for larger and longer-term data, an accelerated approval might be lifesaving for many patients. Accelerated approvals require drug developers to run confirmatory trials in parallel. The hope is that the drug will show full efficacy and safety retrospectively. In a minute, we will talk about two cases where this unfortunately was not the case.
The publication at hand also included other expedited regulatory mechanisms in their analysis such as fast track and priority review, as well as breakthrough therapy designations.
So, what do you think is value-add of these expedited programs? As you might expect and can see in the much higher purple band, drugs under expedited mechanisms are more likely to have high therapeutic value.
Most program types are in a similar range, with breakthrough designations coming in highest – 60% are indeed high value. While normal approvals are at a crappy 13%, drugs with 3 or more designations are at 65%. These designations are awarded during development, so before the clinical product profile become clear. Thus, it’s obvious that we will never get to let’s say 90% for breakthroughs, simply because drugs which appear very promising in phases 1 or 2 might prove out to be only mediocre after a larger scale phase 3 trial. Still, regulatory agencies should give additional explanations or disclaimers to set more realistic expectations for patients and doctors.
Are Drugs getting More expensive?
Finally, let’s see what a 2020 study has to say, covering 65 cancer drugs across countries. The sad insight: low-benefit drugs are just as expensive as high-benefit medicines. You can see that the blue and red distributions are awfully close to each other, with no statistically significant differences between them. From this chart, you can also immediately see why the US is the major pharmaceutical market globally, with drug costs roughly 1.3-times higher than in the European countries.
In summary, new drugs do not always mean innovation, breakthroughs are not always proven to be true breakthroughs, and a high price does not mean high value. Did any of these statistics surprise you? On an optimistic side, we should remember that this is just theory and does not reflect the real-world use of these drugs. Patients will not switch to a new drug which works just as well as what they are already on, and payers would restrict coverage for high-price drugs with dubious efficacy.
Approvals gone wrong – Avastin in metastatic Breast Cancer
However, as we alluded to at the start, there are unfortunately cases where drugsare proven to be quote unquote useless. Let’s check out two examples, and then dive into some chemistry.
The first one is Avastin, which is a VEGF-targeting antibody which suppresses the growth of new blood vessels. This hinders supply nutrients to the cancer, slows its growth, and therefore slows cancer progression.
The FDA approved Avastin in a broad set of tumors. So, we can see the theme of supplemental indications mentioned in the first part. The problem really was around its accelerated approval in metastatic breast cancer (2008). At the time, the FDA’s Oncology Advisory committee, so an outside advisory panel, had actually recommended 5 to 4 against the approving Avastin in metastatic breast cancer. This mixed opinion was because Avastin at the time did not extend life.
We look at the details of the clinical study comparing Avastin as an add-on to chemotherapy. Overall survival – the time from trial entry to patient death – is the gold standard primary endpoint in cancer trials but takes more time and patients to collect. Avastin missed the mark on this important metric. Avastin’s OS is numerically higher, 26.7 months vs. 25.2 months, but this is not statistically significant as indicated by a p value of higher than 0.05.
However, progression free survival improved significantly. PFS is the time between treatment start and first evidence of disease progression, or also death. PFS data is available earlier than OS, so it serves as a surrogate endpoint. So essentially, Avastin reduced the speed of progression by more than 5 months, but patients were ultimately not living longer. Also, their subjective quality of life did not increase either. You might think: If the drug slows down progression, should we approve it nevertheless?
What was The Problem?
An additional problem was the overall 20% increase in adverse events with Avastin. Toxic effects like hypertension or infections increased significantly. This resulted in 6 deaths due to Avastin. So, considering all this evidence, the accelerated approval was quite optimistic but understandable given the lack of other treatments that showed similar PFS benefit.
As we learned, accelerated approvals require parallel studies which intend to confirm the drug’s benefit. However, the opposite happened.
Looking at the AVADO trial comparing two doses of Avastin, the extended PFS was still there. However, with just under 1 month difference between trial arms, this was significantly lower than expected. More importantly, patients on Avastin had numerically shorter OS. Because of the high p value, we can’t really say if this means Avastin is indeed worse or just equal to chemo. This led the FDA to withdraw the metastatic breast cancer indication for Avastin.
Interestingly, the European regulator EMA did not follow suit. In their view, benefits outweigh risks given positive albeit small PFS benefit and lack of statistically significant detrimental effect on OS.
Approvals gone wrong – rucaparib in 3rd line ovarian cancer
The second case study is Rucaparib, which was indicated, among other uses, for third-line treatment in ovarian cancer. It belongs to the class of PARP inhibitors which interfere with DNA repair mechanisms.
Like the results for avastin, rucaparib extended the median progression-free survival by 5 months. Patients with DNA repair deficient tumors responded even better as they are more prone to PARP inhibition. This seemed like a solid benefit at the time – but can you guess the problem?
Four years after its approval, now in 2022, more mature data on overall survival came in. The harsh reality – across different data cuts, patients on treatment again did not live longer. Again, we need to check the p value to see that even in the BRCA mutated sub-group, the potential strong responders, the seemingly beneficial effect is not statistically significant. This led to a voluntary withdrawal in the third-line indication by the company. Unfortunately, a somewhat desperate filing of rucaparib as a first-line therapy by the company did not work out either.
If you watched my previous video, you would remember that pharmaceutical companies burn a lot of money on development and operations. Clovis tried to cut costs through lay-offs and raise additional money, but ultimately had to throw in the towel and file for bankruptcy.
In summary, PFS benefit does not necessarily translate into true survival benefit – which is ultimately most important. Also, we saw that the FDA and EMA can reach different scientific conclusions looking at the same data.
organic synthesis of rucaparib
Finally, it’s chemistry time. Here we have two things to check out. First, we will look at just one part of Rucaparib’s original large scale process route, and second, go through a more efficient synthesis in full.
The process route, albeit long, has an interesting reaction early on. It starts with a nitration of this benzoic acid – which selectively nitrates meta to the acid. So far, nothing special – most of you should know the selectivity. Doing a nitration with 23L concentrated nitric acid is nothing so sneeze at, but we are more eager to see what happens after this step, and a simple esterification. It’s a Leimgruber-Batcho indole synthesis which leverages the moderate acidity at the benzylic position of these nitroarenes. In the first step, the benzyl anion is formed and attacks the reagent, dimethylformamide dimethylacetal. This adds one carbon to the system and after spontaneous elimination of methanol, a conjugated link is formed.
The second step of the indole synthesis is the reduction of the nitro-group. The liberated nucleophilic aniline can intramolecularly cyclize with the iminium, and again eliminate dimethylamine to create the aromatic indole. Although you might expect that immediate intramolecular attack in a 5-endo-trig fashion would give an anion that is very much stabilized through the ester group, I think that’s kinetically disfavoured work due to Baldwin’s rules.
Second synthesis of rucaparib
The remaining steps are not revolutionary so we will instead look at the very direct synthesis published in 2022. The starting material is already highly functionalized, but you can buy this commercially so it’s fair game. The first step is a Heck-reaction with the highly reactive aryl iodide. Next, the aryl amine was condensed with an aldehyde bearing the other half of rucaparib. At this point, we essentially already have all atoms that we need already, it’s just about linking them and getting to the right oxidation states.
To create the indole, the chemists used a cyanide-catalyzed imino-Strecker reaction – you might remember this one as a prime example of Umpolung chemistry. The mechanism starts with nucleophilic addition of cyanide to the aldimine. The negative charge first sits on nitrogen, but rapidly tautomerizes alpha to the nitrile. Now, we have the typical Stetter 1,4-conjugate addition onto the vinyl nitrile. Finally, just like we’ve seen in the previous Leimgruber-Batcho synthesis, elimination delivers the indole. This regenerates the catalytic cyanide which is why the reaction works with just 20 mol% of sodium cyanide.
Well, what do we do next with the nitrile? You’ve guessed it – we need to reduce it to the amine, and link it to the ester. Because typical hydrogenation did not result in any reaction, the authors looked for other reducing agents. Interestingly, the generation of nickel boride from nickel chloride and sodium borohydride worked smoothly and chemoselectively, leaving the ester group in peace. The lactamization occurred spontaneously in situ, so that was quite convenient as well. The ultimate step was a simple deprotection of the amine sitting at the other side of the molecule, completing the total 5-step synthesis.
That’s it for this time. As always, I will catch you in the next one.
Key References:
Therapeutic value of first versus supplemental indications of drugs in US and Europe (2011-20): retrospective cohort study: BMJ 2023, 382, e074166
Association between FDA and EMA expedited approval programs and therapeutic value of new medicines: retrospective cohort study: BMJ 2020, 371, m3434$
rices and clinical benefit of cancer drugs in the USA and Europe: a cost-benefit analysis: Lancet Oncology 2020, 21, 664
The US FDAs withdrawal of the breast cancer indication for Avastin (bevacizumab): Saudi Pharm J 2012, 20, 381
Efficacy and safety of bevacizumab in combination with docetaxel for the first-line treatment of elderly patients with locally recurrent or metastatic breast cancer: Results from AVADO: Eur J Cancer 2011, 47, 2387
Paclitaxel plus Bevacizumab versus Paclitaxel Alone for Metastatic Breast Cancer: N Engl J Med 2007, 357, 2666
Multikilogram Scale-Up of a Reductive Alkylation Route to a Novel PARP Inhibitor: OPRD 2012, 16, 1897
Watch the video on YouTube or read the blog below!
What is the most expensive drug in the world? Imagine a single dose of medicine priced at $3.5 MILLION. Since the approval of the gene therapy Hemgenix in late 2022, this is a reality. This eye-popping price is easy to scrutinize, especially if you know the truth behind most newly approved drugs. Drug manufacturers think these prices are still fair, with some even winding down operations due to pricing disagreements. Also, it couldn’t be more fitting that one of the companies we will talk about, who is selling a very simple drug for one million a year, very recently saw itself forced to lay off 25% of its staff, maintain loans and preserve cash burn to simply keep up its operations for another year.
After reading this post, you will know:
Why soaring prices do not always equal large profits. You will not understand why some therapies, as crazy as it sounds, cannot be sold for prices under hundreds of thousands of dollars
The answer to the question “what is the most expensive drug”?, as well as top contenders
The chemical synthesis of the most expensive small molecule drug, used for an ultra-rare disease occurring in 1 out of 40 million people
How to estimate drug sales for gene therapies
most expensive drug in the world – what is the bar?
Hemgenix approval dethroned the gene therapies Skysona and Zynteglo, both developed as rare disease therapies by the company BlueBird Bio. With costs per dose of 3 and 2.8 million respectively, they currently land on number 2 and 3. The Institute for Clinical and Economic Review, a non-profit organization publishing clinical and cost-effectiveness analyses of treatments, deemed the true value of Skysona to not be too far off. The company itself postulated normal standard of care consisting of regular transfusion treatments cumulate to total healthcare costs of $6million over a patients life time. Comparing this with Soliris, which had been the world’s most expensive drug for some years in the past decade, we can see that in certain diseases such as myasthenia gravis, a cost-effective price would require a more than 95% discount on the annual price.
So, million-dollar gene therapies are actually not bad in terms of value provided. We have also seen some companies offer outcome-based deals, where for example 80% of the price is refunded if the patient does not respond to therapy.
Gene therapies are also less-overpriced if we look at the cost that it takes to produce them. Producing the annual supply of the antibody Soliris, costs less than 1% of its price. In contrast, gene therapies require bespoke manufacturing and supply chains, with fully loaded manufacturing costs reaching over 200 thousand dollars. But maybe you’re wondering why companies wouldn’t sell them at a lower price let’s say, half a million, and still make a decent profit? First of all, these variable costs do not include the large capital investments required for manufacturing facilities.
Drug R&D Costs
But as anyone who had more than 30min of business school will tell you, there are many more direct and indirect costs behind drugs.
Currently, it costs roughly 2.3 billion dollars to research, develop and launch a new drug. As you can see, R&D costs have been trending up, roughly doubling during the last decade. Some drugs are even higher than 3-5 billion. These large upfront investments are why a company will typically command as high of a price as it can reasonably get, to try to maximize internal return on the dollars.
But if we look at the peak sales potential of drugs, we can see that this decreased instead of increased in recent years. Basically, development is getting more expensive while returns are getting smaller.
Don’t forget that there are other costs beyond the manufacturing cost we highlighted. Companies deploy various functional teams to drive uptake of drugs – the easiest to think about are sales and marketing. On a company level, selling, general and administrative cost are usually around 15-30% of net revenues. On a product level, this obviously depends on the lifecycle stage, with new drugs having much more investments behind them than established products which are more like cash cows.
Additional Factors: Discounts and risk
Another important thing to note are gross-to-net price discounts. You see, the list price you read in the news refers to a theoretical price which is actually never paid. In the extremely convoluted US healthcare system, manufacturers give discounts to various stakeholders involved in access, distribution, or other things. These typically untransparent gross-to-net discounts translate into actual net sales for companies which can be less than 50% of the original list price.
One of the underlying problems behind high development costs is the low probability of clinical success. Science and drug development is inherently high risk, high-reward. While big companies with more than a hundred projects will always get something out of their R&D funnel, smaller companies can get wiped out if several drugs turn out to simply not work as hoped. This is why no company will ever price drugs at minimal margins, because prices need to be high to compensate for large R&D costs AND offer protection against critical pipeline setbacks.
Lonafarnib: another expensive drug
This molecule was originally discovered by Schering-Plough more than 2 decades ago as an investigational cancer drug. After development for oncology was discontinued due to lack of efficacy, and Schering-Plough was acquired by Merck, a deal was struck with the small biotech Eiger Pharmaceuticals. Eiger originally wanted to develop the molecule for hepatitis D, but was also introduced to the Progeria Research Foundation by Merck. This non-profit research organization found that the mechanism underlying lonafarnib’s activity was also involved in progeria.
This heart-breaking disorder causes rapid-aging in children. With a prevalence rate of 1 case in 20 million individuals, this is an ultra-rare disease. Patients die at an average age of 13 years, perversely from heart attack or stroke – so the unmet need is massive.
This always fatal disease is caused by mutation in the gene coding for a protein called lamin A. A single cytosine-to-thymine mis-spelling is the most common mutation, and has critical effects.
Normal lamin A is modified with a farnesyl group which helps direct it to the nuclear lamina (a shipping tag). The tag is subsequently cut-off from normal lamin A. However, the progerin cannot be defarnesylated – meaning the un-natural shipping tag stays on the protein – and this interferes with a myriad of different cellular mechanisms. Because lonafarnib is a farnesyl transferase inhibitor, it prevents the addition of farnesyl groups to progerin to start with. Due to the missing shipping tag, the proteins do not reach their destination as easily but also do not accumulate and mess up the nuclear lamina either.
Although far from a cure, the simple farnesylation-block has significant clinical effects. Treated patients have significantly lower mortality risk and over the course of 11 years, live roughly 2.5 years longer. It also had a good safety profile with small number of discontinuations.
How Expensive is it?
As this is a significant improvement in ridiculously small patient population, we can start to understand why this medicine costs around one million a year, despite being a simple small molecule. This makes it the most expensive small molecule drug. There are not that many sales to be made – as many of the 400 children living with progeria worldwide were already dosed in your clinical trials. Of course, Eiger could theoretically sell it for cheaper, but you also must consider cost of the several clinical trials which lonafarnib was tested in, including its oncology history. So even though the price looks Machiavellian, the company actually lost 200 million dollars in the last three years.
The current R&D spend is focused on the development of lonafarnib in hepatitis D, which is a much bigger commercial opportunity but will also come with lower pricing. They also have additional pipeline projects, so all this cash drain resulted in them recently communicating a 25% workforce reduction. In summary, we see that having the most expensive small molecule drug doesn’t necessarily make life easy.
Chemical Synthesis
Lonafarnib was synthesized on a up to mind-blowing 100kg scale when it was planned as an oncology drug. If you compare the key intermediate to the target, you will notice that we need to introduce an aryl bromide group. How is this done selectively? If you ever had some electrophilic aromatic substitution theory, you would know that other positions are more electronically activated due to the present chloride substituent.
The team approached this starting with a single nitration, giving rise to two nitro regioisomers. Regardless of their position, the effect is the same. A novel reduction system with catalytic iodide proved to reduce the nitro group as well as ketone, saving a step. A mix of hypo-phosphorous and phosphorous acid avoided halogen side reactions while reducing both groups. However, they also observed severe foaming at the beginning of the reaction. For more control, they first added H3PO3 to reduce the nitro group, and then H3PO2 to reduce the ketone. This system proved better than hydrogenation or metal catalysis which would have decomposed the aryl halide groups and introduced potential contaminants.
With the amine in place, the desired bromination product is favored through either ortho-direction in one of the isomers, or para-direction in the other.After removal of the amino groups, the most acidic site is between the aryl rings. The addition of LDA, quinine and this electrophile led to a chiral alkylation, proceeding with impressive 95% ee. Gratifyingly, they were able to recycle the quinine – quite important if you run things on a double-digit kilo scale. Next, a diastereomer salt formation enriched ee to 99%. Last, a final amide formation introduced the missing part – and a one-pot Boc deprotection and urea formation gave the product.
Hemgenix in hemophilia B: The world’s most expensive drug
Now that the chemists are happy, we can pivot to the most expensive drug. For this, we need to understand a disease called hemophilia B, an inherited bleeding disorder caused by a deficiency of the coagulation factor IX. This disorder is characterized by frequent and recurrent bleeding into joints or soft tissue, leading to chronic pain, disability, and impaired quality of life.
The current standard-of-care for haemophilia B includes life-long on-demand and prophylactic replacement therapy with FIX concentrates. This replacement therapy is effective at reducing bleeding episodes and is well tolerated – although some patients develop antibodies, making them resistant to replacement therapy. These repeated injections become extremely expensive over a patients lifetime, with most estimates well beyond $10 million. ICER says 3 million dollars would be a very reasonable price – so Hemgenix is not too far off. So clearly this one-time treatment has a lot of impact, but how does this work biochemically?
How do Gene therapies work?
Gene therapies use engineered viruses as carriers or so-called vectors of genetic information to correct a patient’s genetic code, ultimately restoring the proper functions of vital proteins. For Hemgenix, this vector is adeno-associated virus or AAV-based. It’s like a train or truck delivering cargo. As viruses are doing virus things, they can enter cells and integrate or transduct the missing gene into the cell.
Compared to normal viruses, these AAV are modified to lose their replicative abilities, rendering them safe as therapies for humans. In addition, a DNA regulatory element called promoter limits the transgene expression in only the desired tissue. In this case, it’s the liver, as factor 9 is produced in hepatocytes and released into circulation from there. The actual factor 9 gene being incorporated is basically a factor 9 protein on steroids, which is 7-times more activity than wildtype due to a single mutation. This allows for a lower AAV vector dose which in turn decreased averse immune reactions by patients.
As you might expect, Hemgenix was superior to replacement therapy in a mid-sized trial with 54 patients. This led to a significantly lower rate of bleeding events as the primary endpoint. 96% of patients were able to stop replacement therapy. This transformational benefit shows in the increased factor IX activity of patients. Severe hemophilia is characterized with untreated factor IX activity below 5%. This also approved, with immediate and sustained increase post-treatment. Given the adult population, the approval is only in patients over 18 years. In summary, instead of continuously replacing factor 9 exogenously multiple times a week, incorporation of the missing gene solves the problem for good – or at least for more than 20 years.
How can we estimate drug sales?
Hemgenix is the world’s most expensive drug – so will it lead to the highest profits?
You might know that total yearly sales equal sales per year multiplied with the cost per patient. We already know the latter – but what is the true number of patients that will get treatment?
Well, there’s only approximately 6000 people with hemophilia B in the US. Let’s assume this population grows by 1% each year. Do you remember if we can treat all of them? Nope! The approval is only for adult patients, which might be around 80% of patients, and only in cases with severe hemophilia B, maybe around 60% of the patients. So 6K times 80% times 60% – some of you might do it in their head –is 2’880. But why is the number I have here different? Well, I mentioned some patients develop antibodies or inhibitors to factor IX replacement therapy. Because Hemgenix does not have data in these patients, we will exclude them too. This reduces the eligible pool by 2%.
Gene THerapies Are Unique
Now just multiply this number with price? Think again! Will all patients use Hemgenix, immediately? Clearly not – most of them will never receive it due to limited coverage due to the high cost, and the uptake will only gradually increase with time.
It’s probably fair to say that not much more than 10% of patients will ever get Hemgenix, and we also need to assume a build over time – let’s say a one or two percentage points every year. Hemgenix is a one-time treatment. It practically cures patients or it “does not work”. Therefore, more patients being treated reduces the addressable pool.
But hey, this therapy will also not be the only one. There is another hemophilia B gene therapy (Pfizer/Spark). Even if we assume that the first-to-market therapy will dominate, the second product will also have some penetration. Obviously, this all depends on its relative efficacy, safety and importantly price. Here, I simply assumed that the new product would have 50% of Hemgenix’ uptake, delayed by one year.
This means the blue line gives us the actual pool we can treat with each product in every year, factoring in the total number of patients previously treated with Hemgenix and the Pfizer/Spark product. I hope you could follow.
Now we can take the number of new patients treated with Hemgenix, calculated by multiplying the phased penetration with the untreated addressable pool, and multiply it with the price per patient. There are also some less important pricing assumptions.
How Profitable is The World’s most expensive drug?
Do you notice anything about the result? The number of Hemgenix-treated patients actually peaks after just a few years, maybe even staying below 200 patients per year. This translates into peak sales of $630M.
Definitely nothing to sneeze at, but we also have to be real here – many mega-blockbusters with lower pricing are much more lucrative for companies due to steady revenues and higher number of patients. Throw the quarter million-per unit production cost on top, program development cost, and all your other functional and SG&A cost … Just because something is the world’s most expensive drug, doesn’t mean it’s highly profitable or overly “capitalistic”.
And let’s once again re-iterate the topic of low probability of success. With bad luck, developers can easily risk all profits of one drug while financing others. Even the simplest medicines can simply fail to demonstrate efficacy – but the development of innovative therapies like gene or cell therapies is another beast. Very recently, the FDA put a clinical hold on Arcellx’s Phase 2 investigating a CAR-T cell therapy following the death of a patient. Just half a year prior, Gilead invested over $300M to secure development and commercialization rights. This shows how fast the profits gained on some products can disappear.
Thank you for sticking through to the end! As always, I hope you’ve learned something!
Key references
Lonafarnib: First Approval | Drugs. 2021; 81(2): 283
A Novel Iodide-Catalyzed Reduction of Nitroarenes and Aryl Ketones with H3PO2 or H3PO3: Its Application to the Synthesis of a Potential Anticancer Agent | Org. Lett. 2011; 13 (19): 5220
Impact of farnesylation inhibitors on survival in Hutchinson-Gilford progeria syndrome | Circulation 2014;130(1): 27
Gene Therapy with Etranacogene Dezaparvovec for Hemophilia B | N Engl J Med 2023; 388(8): 706
Etranacogene dezaparvovec for hemophilia B gene therapy | Therapeutic Advances in Rare Disease. 2021;2. doi: 10.1177/26330040211058896
Structure-activity relationship study to improve cytotoxicity and selectivity of lonafarnib against breast cancer cells | Arch Pharm (Weinheim) 2023; 356(4): e2200263
Deloitte | Measuring return from pharmaceutical innovation 2022 – IQVIA | Global Trends in R&D 2021
Ibogaine has built a reputation as an anti-addiction magic bullet. Even the Wolf of Wallstreet is vouching for it, lol. While drug manufacturers are settling lawsuits, the state of Kentucky recently announced they might use double-digit millions for ibogaine research.
Maybe you’ve heard of syntheses and promising effects of psychedelics psilocybin, LSD, THCP or MDMA. This will be just as interesting!
However, the clinical development of psychedelics is not as rosy as some of you might expect. There is an increasing number of case reports with severe and even deadly adverse events at high doses. Thus, scientists pursue next-generation molecules that unify life-changing efficacy with superior safety.
Join me on a journey to learn the biochemistry, therapeutic promise, and chemical synthesis of ibogaine and psychedelics-inspired medicines. How can we even know if these drugs might help, let’s say, heroin addiction?
Let’s start with the basics. What is ibogaine? Iboga comes from the bitter root bark of the Tabernanthe Iboga rainforest shrub native to West-Central Africa. Beyond traditional medicine, iboga also has a long-rooted – pun intended – importance to spiritual practices. From a Western perspective, its ritual use was first documented by French and Belgian explorers in the 19th century. Early on, high iboga doses were shown to induce powerful states of mind but also have toxic side effects. On the other hand, tribal hunters used much smaller quantities as mild stimulants. These guys were already microdosing before it was cool.
History of ibogaine and context:
Its recent history is reminiscent of other substances as it also meets what you could call the “20th century psychedelics starter pack”. Was iboga once sold commercially as a dubious extract, just like psilocybin or heroin? Check. Did the CIA run unsettling experiments as we’ve seen with LSD, in search of agents for warfare or mind control? You bet. And did the FDA classify ibogaine as a devilish Schedule 1 drug – to the dismay of people like Howard Lotsof who started to report anecdotal evidence of potent anti-addictive effects? Check. Although it was indeed abused by athletes as a doping agent, this classification dealt a blow to ibogaine investigations. While some early clinical studies were funded in the 1990s, many were terminated, and progress was quite sluggish.
Before we can understand medicinal effects, we need to take step back again from history. Iboga bark is not a pill, so it contains numerous natural products. This table from a mass spec study just shows ones over 1% – so the full list is long. Like we’ve mentioned for psilocybin, it could be that some of these phytochemicals support some sort of entourage effect of iboga. As the major alkaloid present with 2% of total bark weight, ibogaine is our primary molecule of interest. Here’s a fun fact some of you might find interesting: iboga even contains yohimbine, an alkaloid used as a dietary fat burning supplement.
Biochemical Effects of Ibogaine
In the body, ibogaine has a half life of roughly 7 hours. After ingestion, metabolization through a demethylation kicks in, catalysed by several cytochrome P450 enzymes. The resulting noribogaine with the free phenol group is more persistent. With an even longer half-life, it’s quite evident why ibogaine usually results in psychoactive effects over 24 hours, longer than most other psychedelics. Despite intensive research, we still do not understand these molecule’s mechanisms properly.
I mean, just look at this table – I’m you will agree that it seems complex! Unlike psilocybin or LSD, ibogaine does not get its hallucinogenic properties due to serotonin 2A receptor activation. This sets ibogaine apart from classical psychedelics.
Noribogaine displays sub-micromolar agonistic affinity to the kappa opiod receptor. This profile is reminiscent of the hallucinogenic natural product Salvinorin A, present in the leaves of the Mexican Salvia plant.
Noribogaine is also a strong partial agonist of the related mu opiod receptor – this is the target of classic opiod analgesics such as morphine and fentanyl, commonly used as sedatives or to treat severe pain. These agents are usually very dangerous, highly addictive substances – they are behind the extensive opioid overuse in the US. But as we will see, due to the breadth of molecular mechanisms implicated, ibogaine-derived substances could be helpful in overcoming opioid dependency.
Another key mechanism is the inhibition of NMDA-receptors, similar to drugs like ketamine and even alcohol. This might explain the dissociative effects of ibogaine shared with these other agents. NMDA receptors are glutamate-gated ion channels which drive neural processes like learning, memory, and neuroplasticity. I’m not saying that randomly taking drugs can help neurodegenerative diseases like Alzheimer’s. It’s probably counterproductive, but there is a molecular link here.
I wanted to highlight two other mechanisms – firstly, inhibition of serotonin and dopamine transporters. The 2 micro-molar Ki value for ibogaine and Noribogaine essentially match the affinity of amphetamines. Ibogaine differs from these notorious drugs of abuse as the serotonin uptake inhibition is non-competitive. This and other reasons are why ibogaine has a lower abuse potential than cocaine, another inhibitor of this class. This mechanism might drive ibogaine’s effect on mood and psychological performance.
Finally, the nicotinic acetylcholine receptor activity is perhaps most likely accounting for the anti-addictive property of ibogaine. Ibogaine is a non-competitive antagonist at several receptor subtypes, most notably the alpha 3 beta 4. This receptor is an important part of reward pathways. Blocking it can dampen dopaminergic activity and reduce self-administration of various drugs.
Beyond this, ibogaine also induces upregulation of GDNF. This is a crucial neurotrophic factor that promotes survival and plasticity of neurons, amongst others. This effect likely drives the attenuation of drug craving and use by ibogaine.
What is the Evidence for Ibogaine?
Now we’ve seen that ibogaine bridges several different classes of psychoactive substances. This translates into promising clinical efficacy, particularly in substance-use disorder. Most ibogaine studies lack rigorous clinical study design – however, there are good data in opioid and cocaine craving.
Let’s briefly check out the largest study, comparing self-reported mood and drug craving measures of opiod or cocaine dependent patients. Strikingly, after an oral dose of ibogaine, patients reported significantly lower levels of drug craving. This is measured through a questionnaire which tests patients’ confidence in ability to quit, emotionality and other factors. In addition, depressive symptoms got better as well. These improvements continued to grow after one month follow-up, indicating potentially quite durable benefits. Many other conditions have preliminary data but we will not talk about them here.
In any case, the upsides look quite promising. What about the downsides?
Ibogaine’s complex pharmacology leads to considerable potential to generate adverse effects. In rats, high doses led to degeneration of neurons. They did not replicate this in primates, so it might be species dependent and less worrisome. High doses have also led to tremors and convulsions in rats.
Much more importantly, ibogaine can also negatively affect the cardiovascular system by prolonging the QT interval of the heart. This comes from strong inhibition of hERG potassium ion channels. These channels coordinate the heart’s beating through repolarization of cardiac neuromuscular junctions. Abnormally QT intervals increase risk of developing heart rhythms problems and even sudden cardiac death.
That’s why alarming reports of life-threatening complications associated with ibogaine have been accumulating. As you can see here, even young people with no other substance use are at risk. Due to the longevity of the metabolite Noribogaine we mentioned, cardiac adverse events may also occur several days. In some cases it can even be weeks after intake of a single dose of ibogaine.
The goal is not to test ibogaine mindlessly in dozens of conditions, potentially giving patients sudden cardiac arrests. Instead, we should explore safer, ibogaine-related molecules to unlock its therapeutic potential. This research needs to elucidate the underlying mechanisms of actions. If promising, the drugs should be translated into robust, objective clinical trials in humans.
So how can we shift the balance towards better safety at similar or even better efficacy?
Ibogaine Variants: 18-Methoxy-Coronaridine
The first attempt at this is an investigational molecule is 18-MC. 18-Methoxy-coronaridine is a modified ibogaine with an additional methoxy and methyl ester group. It is synthesized differently than ibogaine, so stay tuned for the last chemistry section.
These new functionalities impact the pharmacological profile a lot. For instance, low activity at sigma sites reduces risks of neurotoxicity, while lack of activity on serotonin transporters means that 18-MC is not hallucinogenic. Interestingly, the activity at the alpha 3 beta 4 nicotinic receptor is much lower, but 18-MC is much more selective for this sub receptor than ibogaine. So, we can see that in some cases, a lower affinity is not bad if it is more targeted.
A more complicated point is also that this table only shows binding affinity – but sometimes, an equally strong affinity expressed as Ki can have a much higher IC50 value, which reflects true inhibition. Unlike ibogaine however, 18-MC does not increase GDNF expression, the additional factor believed to be critical for neuroplasticity, so their mechanisms of action are potentially distinct. Overall, 18-MC seems to have a much narrower spectrum of actions. In theory, this drives a greater therapeutic index – meaning the effective dose is much lower than a potentially harmful dose.
Regarding cardiotoxicity, 18-MC inhibits hERG channels roughly 3- to 4-times weaker than ibogaine. It’s not fully clear whether this is enough to abolish the arrythmia and cardiac adverse events – just shortly, we will check out another analog which is even better.
The clinical fate of 18-MC is not clear either. The biotech MindMed – don’t confuse it with Mind Cure – completed a Phase 1 trial last year with a solid number of patients dosed. Initial data was positive with good tolerability and no serious adverse events. They also planned a larger proof of concept trial. However, they paused it due to financial reasons with new financing and partnering required to advance the program.
Instead, the company is focusing their efforts on the development of LSD in phase 2 for anxiety and ADHD, and MDMA pre-clinically for autism spectrum disorder. As we have seen in previous videos on this channel, these drugs might be very promising in these conditions. So, who knows – strong data could resurrect MC-18. Drop me a comment if you want an update on these programs in future!
Ibogaine Variants: Tabernanthalog
In any case, fortunately there has been a promising addition to the analog roster. A 2021 paper in Nature reported the results of another quest into ibogaine analogs. Instead of throwing more groups on ibogaine like MC-18, the logic here was to simplify ibogaine’s structure, thereby improving accessibility and elucidating which features are most important for activity. In case of the ibogaine skeleton, you can envision two different simplified ring systems – one in light green and one in blue.
Out of many compounds, the most promising is “tabernanthalog“, featuring a shifted methoxy group compared to ibogaine. Before we check out why this molecule seemed to hit the sweet spot of safety and therapeutic effect – do you have an idea how to synthesize TBG?
Even though it’s quite sizeable, it requires only one step, a Fischer indole synthesis. This reaction links this substituted phenyl hydrazine with the seven-membered ketone, creating the tricyclic TBG. The mechanism is part of many undergrad courses. The initial condensation reaction forms a phenyl hydrazone which isomerizes to the enamine form, drawn here. Upon protonation, we have a sigmatropic rearrangement which creates the C-C bond. After re-aromatization, the nucleophilic amine drives C-N bond formation via the aminal – which eliminates ammonia under acidic catalysis. We will review full syntheses of ibogaine in the final section of this video – but you can already guess that making TBG in a single step with 55% yield is infinitely easier than synthesizing ibogaine from scratch.
What are Effects Of Tabernanthalog?
First up is hallucinogenicity. While appreciated by some folks, pharmaceuticals should not elicit hallucinations. Seasoned channel viewers will recognize the classic head-twitch response assay to test for hallucinogenic potential of molecules. As a positive control, we have 5-methoxy DMT which is strongly hallucinogenic, reflected in the frequent head-banging of mice. In red is IBG – this is not ibogaine but rather the simplified version with the methoxy at a constant position. Even lower than IBG was TBG in blue with essentially no hallucinogenic potential. So, these were quite some sleepy mice instead of the energetic headbangers for 5-MeO-DMT.
Remember ibogaine’s adverse cardiac effects, mediated by the hERG channel? Both simplified analogs have much weaker inhibitors than ibogaine. The simple shift of the methoxy position between IBG to TBG comes with an additional 7-fold reduction in IC50 value. The overall 150-fold weaker binding gives TBG its promise as a quote unquote “safer ibogaine”. Obviously, this is much better than the 3-4-fold difference between ibogaine and 18-MC, the first analog we talked about.
So, safety is just one part of the equation – but does TBG also bring similar positive effects? Here is where we want to review a few interesting experiments, starting with neural plasticity.
This is the ability of neural networks to change through growth or reorganization. One way to look at it is the growth of dendrites – these are a nerve cell’s extensions which propagate electrical stimuli. Exposure of rat neurons to ibogaine, IBG or TBG all lead to more dense dendritic spines.
We can distinguish if dendritic growth is due to slower break-down of spines, or instead by a higher rate of formation. Both DOI and TBG drive growth in the same manner – they accelerate the formation of new dendritic spines.
Do these psychoplastogenic effects translate into behavioral or anti-addictive effects for TBG as well? We pointed out, there are anecdotal and initial clinical reports that ibogaine can reduce alcohol or opioid use. For this analysis, you unfortunately must make mice alcoholic by giving them the option of binge drinking. After a standard 7-weeks protocol, they compared alcohol consumption between two groups. Mice who proceeded as usual (blank) and mice who received TBG prior to the drinking session. The latter group had much lower alcohol intake both during the initial part of the consumption test, as well as acutely over the following days.
The team observed similar effects when looking at heroin as another substance with high abuse potential. Here, TBG administration also led to a much lower heroin intake – seen on the left graph – and also seeking behavior – as seen on the right graph in terms of number of lever presses during their experiment.
As a last notable effect, we look at TBG’s impact on depression. We can investigate this through a “forced swim test”. “Less depressed” mice will spend more time in motion, somewhat reflecting their drive and will to live. Even though also quite controversial, all marketed antidepressants increase swimming time in the FST – so the test is legit. The researchers performed two tests – one 24 hours after administration of TBG, and a second after one week of rest. This time, the blank positive control bar is ketamine, an effective anti-depressant. During the first test, both ketamine and TBG reduced immobility. Adding ketanserin once again abolished the effect as you can see in red. Interestingly, ketamine’s effects seemed more durable, as it still led to significant lower immobility one week after administration. TBG on the other hand looked more like the vehicle control.
We discussed previously that ibogaine and its metabolite noribogaine interact with numerous biological targets. Unlike Noribogaine, TBG or IBG showed no activity at opioid receptors. Perhaps, the higher selectivity could lead to a better drug profile down the line. On the other hand, the control experiments which ketanserin already showed us that serotonin receptors are vital for TBG’s activity. A more detailed screening revealed that TBG is both an agonist of the serotonin 2A receptor – but also an antagonist to the serotonin 2B receptor. Drop a comment if you need some more explanations on how to read these charts. The interesting thing here is that many 2A agonists are also 2B agonists, which can lead to side effects like heart valve disease. 5-methoxy DMT is a key example – as you can see in the orange plot, it inhibits both receptors in a similar manner.
Outlook on Tabernanthalog
This case study was rather simple on the synthetic design part of things. Still, I think it’s really fascinating that TBG looks like ibogaine but seems to behave differently mechanistically. Although much work on translational science into humans and dosing optimization is required, TBG might be able to overcome ibogaine’s safety limitations and unlock the potential of this class of drugs.
And last, a brief note on Mindcure. This biotech company was pursuing the development of ibogaine, garnering some attention from professional and private investors such as the chap we saw during the intro. They supposedly were on track to have fully synthetic GLP supply of ibogaine ready by end of last year – but ironically, just two weeks after, reported the result of a strategic review – the discontinuation of all activities. The psychedelic pharmaceutical market can be quite volatile, and funding challenges in recent years have definitely not helped these companies either.
As a random side note, their website was dubious from the start as they didn’t get the molecular structure of ibogaine right – unless they were showing some other analog which I missed.
So – all in all, there are some promising evolutions, but progress is sluggish. I expect that we are still far away from regulatory approvals. Instead, emerging clinics in countries where ibogaine is legal will continue to draw visits from abroad. They might be helpful for some individuals as a last resort but come at a risk of sketchy medical practices and questionable patient safety.
On the positive side, we do see increased state and federal interest in ibogaine due to the opioid problem, and psychedelics more broadly. For instance, the state of Kentucky is currently considering the allocation of 42 million dollars for ibogaine research. Out of a much bigger pocket of almost a billion in settlement funds, this looks like money well spent on larger and broader clinical trials.
Organic Chemistry: Retrosynthesis of Ibogaine
Now we will discuss not one or two, but three different approaches towards the ibogaine framework – as well as the synthesis of 18-MC.
From a retrosynthetic perspective, given the high complexity of the ibogaine scaffold, there are various disconnections that lead to sensible synthetic approaches. A quite straight-forward option uses a Fischer indole synthesis with a simpler ketone. However, most approaches include the indole from the start to guide the synthesis. One method we will review uses transition metal catalysis, while others harness the electrophilic reactivity of the indole. The gram-scale synthesis we will look at uses yet another approach based on nucleophilic substitution at the aliphatic nitrogen. Note that these syntheses focus on ibogamine – which is ibogaine lacking the methoxy group – because it is not a controlled substance.
First total synthesis of ibogamine (Büchi)
Let’s start with the pioneering first total synthesis of ibogamine, published by Büchi in 1965. It started from this pyridinium salt, which was reduced to the diene. This prepared the Diels-Alder reaction with methyl vinyl ketone which nicely builds the iso-quinuclidine core of ibogamine. Next, some redox and functional group interconversions produce the following intermediate. There are quite a few things going on, so we won’t go into it in detail – but this is a nice exercise for motivated viewers. Now, hydrogenation of the benzyl protecting group released the nucleophilic amine, which was coupled to this indole, bearing an acyl chloride.
The next task was to create the central C-C bond to connect the rings. This was achieved in two steps – under acidic conditions, the indole electrophilically attacks the adjacent ketone, and the resulting adduct was reduced with Zinc and acid. A few more steps were needed to get all ducks in order. First, a reduction removed the acetate protecting group and partially reduced the amide. To get to the fully aliphatic amine, they had to take a detour due to the reactivity of the system. Elimination of the hydroxy group with base temporarily cleaved the isoquinuclidine ring.
The link was regenerated by reduction with Zinc, which is mediated through the unsaturated ketone. Finally, a Wolff-Kishner reduction with hydrazine removed the ketone and gave ibogamine. All in all, not bad for 1965, but can we make this more efficient?
Modern Synthesis of Ibogaine
That’s exactly what the second synthesis is about. It starts off with a Palladium-catalysed heteroannulation to forge a highly functionalized indole. You will note that the ring contains a methoxy group, so this is indeed a synthesis of proper ibogaine. Next, two iodide groups were introduced – first at the indole by treatment with electrophilic NIS, and second at the aliphatic position by deprotection and SN2.
This reactive iodide should remind you of the acyl chloride we saw in the 1965 synthesis – again, it allows the introduction of the isoquinuclidine through another substitution. This can be made in a similar fashion as we saw in the 1965 synthesis as well – so these syntheses have some parallels. Interestingly, the authors noted that when using potassium carbonate as a base, there was a significant degree of intramolecular cyclization to the cyclopropane. This could be suppressed by using caesium carbonate instead. Finally, the indole and isoquinuclidine were bridged through a reductive Heck coupling, which after elimination already gave our product ibogaine. This synthesis is definitely more efficient and direct – but is there anything cooler?
Large-Scale Synthesis of Ibogaine
Three time’s a charm today. Quite recently, a paper described the gram-scale synthesis of ibogamine in just nine steps and an impressive 24% overall yield. Most notably, this approach would provide ample material to pursue even more synthetic analogs, particularly ones than are more complex than TBG.
The synthesis started from this vinylogous ester. First, a simple silylation protected the primary alcohol. Then, a Stork-Danheiser transposition with the Grignard reagent formed an enone, now bearing the ethyl group present in ibogamine. Through a Mitsunobu coupling, this fragment was linked to an indole bearing an amine. So, this contrasts with the previous syntheses where we had an electrophilic indole partner, this one is nucleophilic.
The ketone was then selectively reduced under Luche conditions and acetylated to create an activated allylic system. This set the stage for pivotal Friedel-Crafts reaction – which as we’ve seen can be mediated by Bronsted or Lewis acids. After some screening, the chemists found decent conditions with catalytic camphorsulfuric acid and lithium perchlorate at a 5M concentration. This meant that the scale-up would require massive amounts of perchlorate. Initial optimization attempts were not fruitful, as they either had to keep the quantity of perchlorate or dilute the mixture to unpractical 0.001M. Ultimately, after trying enough conditions, they got good conditions employing only 2 equivalents of magnesium perchlorate.
Finally, the only thing left was the formation of the C-N bond on the isoquinuclidine – you might remember that we highlighted this as a key retrosynthetic disconnection at the start. First, the double bond which remained from the enone was hydroborated and activated as a mesylate. And last, the nitrogen was deprotected which triggered the intramolecular SN2 reaction to give ibogamine. The whole exercise delivered 1.1g of pure ibogamine in one go.
Synthesis of 18-Methoxycoronaridine
To conclude our journey, let’s check out the initial synthesis of 18-MC starting from tryptamine – as you can imagine, it will be more complex than the ibogaine syntheses given the two additional functional groups.
The route starts off with a condensation of tryptamine to the ketone of this fragment. If you are still awake, you will notice that this ester group ultimately ends up in 18-MC. Due to the alpha-chloro group, the product can undergo an intramolecular substitution, creating a transient aziridine, and rearrange to the expanded 7-membered ring. Then, the double bond can be reduced, and the nitrogen protected.
The unique thing about this system is that upon heating, a retro 1-4 addition can fragment the ring, liberating the free amine and the alpha, beta unsaturated ester. Why is this helpful? Well, by condensing the amine with this aldehyde, a dearomative Diels-Alder reaction can be triggered. As a side note – that it really matters down the line – note that because the intermediary (e)-enamine was preferred, the product has the substituents in trans positions. Also, the newly introduced piece features the methoxy group we want to have in 18-MC.
So now, it’s all about linking up the rings properly. First, a conjugate reduction regenerates the aromatic indole and releases the quaternary carbon. Next, a hydrogenation unveils the amine, which upon deprotection of the aldehyde forms yet another cyclic enamine.
Redrawing this structure, we realize that just a final ring closure is needed to create 18-MC. This was achieved by simply heating in toluene because the additional ester group proved quite handy. It’s likely that an intramolecular proton shift facilitates formation of an anion and iminium, which can react to create the quaternary centre and deliver 18-MC. This first synthesis from 2001 did seem a bit random – and there are more efficient routes that are more analogous to ibogamine – but I thought it was nice that they used the additional ester to guide the approach.
This concludes our ibogaine journey. I hope you learned several new interdisciplinary science facts today!
References on Ibogaine, Ibogamine, Tabernanthalog & 18-MC
DARK Classics in Chemical Neuroscience: Ibogaine | ACS Chem. Neurosci. 2018, 9, 2475
Phytochemical characterization of Tabernanthe iboga root bark and its effects on dysfunctional metabolism and cognitive performance in high-fat-fed C57BL/6J mice
A systematic literature review of clinical trials and therapeutic applications of ibogaine | Journal of Substance Abuse Treatment 2022, 138, 108717
Ibogaine Detoxification Transitions Opioid and Cocaine Abusers Between Dependence and Abstinence: Clinical Observations and Treatment Outcomes| Front. Pharmacol., Sec. Neuropharmacology 2018, 9: 00529
The Anti-Addiction Drug Ibogaine and the Heart: A Delicate Relation | Molecules 2015, 20, 2208
18-Methoxycoronaridine (18-MC) and Ibogaine: Comparison of Antiaddictive Efficacy, Toxicity, and Mechanisms of Action Annals of New York Academy of Sciences 2000, 914, 369
A non-hallucinogenic psychedelic analogue with therapeutic potential | Nature 2021, 589, 474
The Total Synthesis of (±)-Ibogamine and of (±)-Epiibogamine | JACS 1965, 87, 2073
Total synthesis of ibogaine, epiibogaine and their analogues | Tetrahedron 2012, 68, 7155
Gram-Scale Total Synthesis of (±)-Ibogamine | Org Lett 2023, 25, 4567
Chemical Synthesis and Biological Evaluation of 18-Methoxycoronaridine (18-MC) as a Potential Anti-addictive Agent | Current Med Chem CNS Agents 2001, 1, 113
