Thursday, July 20, 2017

Future of Medicine - draft

( Forthcoming paper in Inference -International Review of Science )

The pursuit of an indefinite healthy lifespan, or healthspan, is the most noble challenge put before mankind. To get there, many bold technologies have been proposed and are now being  explored. Among them are two seemingly unique approaches that many believe have already achieved some preliminary measure of success: The first is to eliminate cellular senescence using ‘senolytic’ drugs. The second is to ‘fix’ mitochondria by ‘allotopically expressing’’ reserve copies of untarnished mitochondrial genes into the nuclear DNA.

Regrettably, neither of these methods stands a chance, for reasons I will detail in this essay. Fortunately, at the confluence of these seemingly isolate approaches to conquering aging there is a way forward. This interface, namely, the mitochondrial control of cellular senescence offers deep promise. The idea idea that mitochondria have a hand in controlling cell senescence is nothing new (1). For example, by toggling energy production in the cell between mitochondrial respiration and cytosolic glycolysis with various drugs or genetic manipulations we can tweak cellular metabolism and fate to our liking. At least for cells in a dish.

What I will offer here that is entirely new, is an incite into the surprising way in which mitochondrial control of cell senescence is exacted across the entire organism. Furthermore, I will describe how we can exploit these natural mechanisms to control the rate of aging at a systems level. Perhaps even more encouraging than mere talk, is that the bleeding edge in medicine has already pushed beyond drug and genetic therapies to devise optimal methods to therapeutically transform ailing tissue by introducing whole mitochondria.

The clear implication here is that the molecule or gene is not the fundamental medicinal unit of measure we will need in order to to out-engineer the ultimate malady; going forward, the ‘quantum of aging’ is the mitochondrion.

Many of the new techniques for artificial mitochondrial transfection, or mitoception, have only been applied in-vitro. One is a ‘photothermal nanoblade’ that uses laser pulses to induce bubble cavitation and mitochondrial uptake.  Another, termed magnetomitotransfer, outfits mitochondria with magnetic beads and directs them via magnetic fields (2). But getting mitochondria into cells does not seem to be the problem. Whether they are injected into culture dishes or into whole organs, blood, or cerebrospinal fluid, cells naturally scoop them up faster than fish flakes sprinkled into an aquarium. In other words, uptake of mitochondria seems to be universal (3).

Cells will use any and all manner of endocytotic mechanisms to get them. Cells will also build elaborate ‘membrane tunneling nanotubes’ reinforced with actin or tubulin to siphon them up when scarce, or unload them during excess. The interesting part for us is discovering the rules of the game -- who transfers to who, and when? Which cell types are the most needy and which the most generous? What really drives the epithelial-mesenchymal transition, or maturation in the bone marrow stem cell population? The answers have been coming in, and a lot of them have been coming from the biology of cancer (4).

Current clinical efforts to deploy mitochondria aim to treat everything from the whole body, by making three-parent embryos, to treating specific regions of damaged hearts by brute force arterial or muscular injection. (4a) What’s missing in all these efforts is an overarching theory to guide them. A theory encompassing how mitochondrial pools are endogenously created, mined, selected, and eliminated. Perhaps more importantly, how they naturally MOVE throughout the body. We hinted at such a theory last year in Part I of the story where we focused on excitability and the nervous system (5). In particular, how the directional transfer of mitochondria can establish polarized neurons and circuits.

Since that time, several advancements in our understanding of how the nervous system controls stem cells, tumors, and cell senescence have suggested we were on the right track. Before getting to some of these advances, let's dispense with those two outdated ideas mentioned above that now occupy and beleaguer much of aging research.

What’s wrong with senolytics?

Cellular senescence refers to cells that can no longer divide. The flip side of the senescence coin is that these cells don’t die either. Instead, they limp along in a crippled limbo lacking the most basic drive inherent in peaceful multicellularity. That drive is a willingness to self-destruct gracefully through apoptosis when you have outlived your useful lifespan. There are many ways cells become senescent. Neurons and heart cells are senescent by design -- they quiescently rest in the permanent G0  growth phase.

A more insidious cellular fate is something known as replicative senescence. This occurs when cells reach their so-called Hayflick limit. At this stage, the telomeres at the ends of the chromosomes have been eroded to the point where cell division is no longer possible. When this phenomenon was initially discovered, it appeared that artificial activation of telomerase genes might be used to rebuild telomeres. This would circumvent the built in limitations of the telomere counting mechanism and extend the division cycles of expired cells. Later on, it was found that some 85% of tumors evade senescence by doing exactly this. Researchers eventually admitted to themselves that an indiscriminate reactivation of telomerase would be folly.

Other common varieties of senescence include oncogene-induced, DNA damage, and stress induced senescence. There is also immunosenescence, where hematopoiesis declines with age resulting in a diminished adaptive immune response. This is a double whammy because a healthy immune system will eliminate not only cancer cells but also senescent cells. In other instances, senescence results when precancerous cells manage to halt potential runaway replication only to fall short of deploying the full apoptotic program that normally would remove them from the body.

The rapidly exploding new field of ‘senolytics’ aims to eliminate cells in these various debilitated states using senolytic drugs. Success so far hinges on several assumptions. The main one is that senescent cells possess something in common which can be used both to identify them, and also target some therapy towards them. Typically this would be either the expression of a very specific biomarker, or some defined pathway or sequence of protein operations that could be blocked or activated. The first major problem with senolytics is the underlying assumptions are simply invalid. There is no unique marker only expressed by senescent cells, and there is no pathway that can be uniquely targeted in just these cells. Fake news.

The second major problem is that the drugs that have been promoted as senolytics invariably turn out to have significant toxicity. There is a reason for this. The only way that any of them have managed to gain any traction in the medical or anti-aging communities is by virtue of the fact that they have been previously tested and approved by the FDA -- no small feat for any treatment in today’s regulatory pipeline. Generally speaking, we are dealing chemotherapies or anti-viral drugs.  The kicker here, is that these drugs end up being promoted as senolytics only because their side effects proved TOO toxic to continue to justify their use in killing tumors.

Let that sink in for a moment. As it turns out, cutting edge senolytics are largely just unemployed layabouts in need of a job. This is madness.

To put some meat on this sweeping dismissal, we should probably consider some specifics in more detail. One recent paper in Nature recently hit the news cycle with considerable fanfare (6). The authors imply that we might clear senescent cells from auricular cartilage to cure osteoarthritis by injecting the mystery compound UBX0101 into joints. What exactly is UBX0101 I wonder? If you can somehow manage to get access to their locked-down paper, you can see that they are using the expression of the molecule p16INKa as a proxy for ‘senescent cells’. As mentioned, this is already suspect -- all ‘biomarkers’ are flawed.

The references in the paper for this UBX0101 indicate that the compound is actually ABT-263. This changing of getaway cars in the tunnel does not engender confidence, particular when we google ABT-263 and find out it is actually navitoclax. Although navitoclax does specifically induce apoptosis in senescent cells, it wreaks havoc on your platelets. Navitoclax was already on the radar of the extremely diligent anti-aging community as something many of them might want to experiment with. It had previously been billed as a safer successor to more indiscriminate cancer-killing drugs like obatoclax, or to natural flavanoids like quercetin and dasantinib which act similar molecular pathways that promote survival in senescent of cancer cells.

As it turns out, a contributor to one noteworthy and excellent blog, ‘Fight Aging’ (7), was at a conference where he spoke with someone from with the maker of the drug, Unity Biotechnology. This representative apparently indicated that UBX0101 wasn’t ABT-263 after all. Furthermore, they related that after the first review of the paper the actual name and structure was removed! Obviously people don’t expect as much these days from published ‘science’ as they used to. At least to me, this whole business sounds more like some kind of senolytic weed killer a company like Monsanto might market; Its actual chemical formulation changes from time to time when needed, but it is still offered to the public under a trade name like ‘RoundUp’.

What’s wrong with allotopic expression?  

Allotopic expression is a great idea on paper. It was proposed decades ago almost as soon as people realized our mitochondria are in fact endosymbiont proteobacteria that were engulfed by an archaebacterial host eons ago at the dawn of the eukaryotes. For all the recent talk of the importance of bacterial ‘microbiome’, meaning all the bacteria that crawl through our guts and on our skin, we should probably confess that our primary microbiome is our mitochondria. The mitochondriome, if you will. A most curious feature of this eukaryotic merger is that nearly all of the genes of the original symbiont were copied into nuclear DNA. As the story goes, they then somehow acquired laser-like mitochondrial localization sequences (MLS’s) to get the proteins back to the mitochondria.

Once mitochondria had a steady source of nuclear mitoproteins, their own genes were dropped from the mtDNA. In many mammals, for example, all that remains in the mtDNA is 13 protein subunits and a minimal set of tRNA and rRNA genes to translate them. These select proteins form the critical hydrophobic membrane spanning portions of various respiratory complexes. Proponents of allotopic expression fancy their mission as a man-made speedup that completes a natural evolutionary process. Namely, a world where copies of ALL the mitochondrial genes have been carted off and encoded in nuclear DNA. Rest assured, this is big government of the highest order. It is also a complete fiction.

Why might someone want this? The logic behind allotopic expression is that once safely ensconced away in the nucleus, mitogenes would be immune to corruption from the free radicals generated by respiration. The evolution of the nuclear versions of these genes would effectively be decoupled from the high substitution and deletion rates of their brethren still in the mtDNA.

The success of this strategy depends on the ability to re-target the proteins made by these genes back to the mitochondria. First they must be recoded to use the nuclear codon system. Then the hydrophobicity of the proteins themselves needs to be dialed back so that they don’t fold up prior to import. This sequence optimization step may also require adding code for translocator and protease recognition, and for proper insertion in the inner membrane, The mRNA code itself may also need to be optimized so that it is translated by cytosolic ribosomes localized at the mitochondrial surface. Getting this process to work for those few remaining proteins that nature herself had the most difficulty with would seem to be a tall order.

It is now known that cytosolic ribosomes conveniently bind to mitochondria at their TOM (Translocator on the Outer Membrane) complexes, typically right above crista junctions (7a). If some of the energy supplied to the nascent peptide elongated by the ribosome is also used to cotranslationally import the newly made proteins through TOMs, much of this whole idea of MLS 'sequence' starts to look like an artifact.

Remarkably, researchers at the SENS Institute reported that they had achieved some signs of successful allotopic expression for the ATP6  and ATP8 subunits of complex V. (8) Another group at Gensight has similarly reported success for the ND4 subunit of  complex I. If these studies pan out it is possible that some very sick people with major mitochondrial dysfunction, and their offspring, could see some benefit. Perhaps it should not be too surprising that some functional complexes can be built by saturating cells with nuclear copies of missing subunits. The problem, is in trying to extrapolate these findings into something that can improve normal people.

For one thing, if you alter the sequence away from naturally evolved ideal in order to pull off nuclear expression, your subunits just won’t be as good. For example, some bugs and many single celled protists manage to get by with far fewer genes in their mitochondria. Some have even offloaded their tRNAs or mRNAs to the nucleus and import them as well. For example, trypanosomes get all of their tRNA from the nucleus and must use the same TOM importers for both tRNAs and proteins. (8a). These creatures only manage to pull this off by making qualitative sacrifices in every product they jettison off the nucleus for cytosolic manufacture. Typically for proteins, this means making them less hydrophobic, as has been intentionally done by researchers to get allotopic expression.

Even more problematic, is that there will be no regulation of genes that are haphazardly integrated into the nuclear DNA. Protists have been playing this game for millions of years and have achieved some level of nuclear regulation, but there is a cost to giving up local on site regulation in the mitochondria. Namely, these creatures are still just protists.

There has been a complete failure in the field to distinguish between what I would call ‘soft’ and ‘hard’ allotopic expression. Soft expression would be putting copies of just one or two genes into the nucleus, while hard expression would be all the genes. Similarly, there has been a failure to distinguish between ‘duplicating’ mtDNA in the nucleus and ‘replacing’ mtDNA. There is a reason for this: DUPLICATING mtDNA into the nucleus is REPLACING mtDNA, and when you replace mtDNA, you have no mitochondria left. Instead you have something else -- a mitosome or perhaps a hydrogenosome -- if that.

The nucleoid that contains the mtDNA is the heart of the mitochondria. Researchers have found that mitochochondrial ribosomes, or mitoribosomes, are assembled right on the nucleoid. Proteins made by these mitoribosomes are in turn co-translationally inserted right into the overlying inner membrane, leaving little wiggle room for wasted motions. By analogy to the nucleolus of the cell nucleus, this larger machine is now called the mitochondriolus (9).

While there are many good reasons why organelles retain this private genome, the mac daddy of all them is CoRR hypothesis. It stands for ‘colocation (of gene and gene product) for Redox Regulation of Gene Expression’ (10). In a nutshell, it says that organelles with elaborate electron transport chains need to have a direct vote in the expression of its major components in order to maintain redox balance in the bioenergetic membrane.

The initial evidence for the CoRR hypothesis came from observing how photosynthetic organelles tune the rates of electron transport through Photosystem I and II in response to changes in light quality. The redox state of electron carriers directly controls various protein modifications which ultimately determine how energy is distributed between the two photosystems. Local redox regulation of transcription adjusts the stoichiometry of photosystem components within the chloroplasts.

Similar redox regulation occurs in mitochondria, where the generation of radicals, direction of electron transport, and stoichiometry of individual respiratory complexes within supercomplexes is tightly controlled. When researchers first started to try allotopic expression little was understood about how these macro complexes were assembled. That situation has changed. For example, the precise location and sequence of assembly for individual submodules of complex I has recently been worked out for several species. Similar to the need for mito-encoded RNAs to seed assembly of mitoribosomes, the mito-encoded membrane protein subunits are critical control points for assembly of the complexes.

The academic community has been critical of private efforts to do allotopic expression to the point that they now mutually ignore each other. I asked Leo Nijtmans, the researcher who cracked the human complex I assembly process, to present the SENS paper to his peers at the recent EuroMit conference. Although still somewhat noncommittal, he did eventually get back to me with a list of questions and criticisms in the appendix section (11).

There seems to be little doubt that restriction or loss of organelle respiratory bioenergetics is accompanied by reduction or loss of organelle DNA. The prediction of the CoRR hypothesis is borne out in nature in the relic mitosomes and hydrogenosomes that have no genomes left. The converse, that loss of genome leads to loss of functional oxidative respiration would seem inevitable, the only question is how quickly things would unravel.

The apparent speed at which mutant mitochondria with naturally occurring major deletions to their genome can clonally expand to take over cells, organs, or bodies has been a main motivator of the allotopic expression initiative. Yet allotopic expression itself would in fact be a major mtDNA deletion by design. Not only would this be a public relations disaster for mitochondria, their reason to be would vanish.

The original deletion studies were done a fairly long time ago in a restricted class of cells. More recent studies have revealed more details regarding large deletions that deletions that result directly from transient ischemia. One common deletion takes out a 7.3 kb chunk that includes 11 proteins. The breakpoints occur in the ATP synthase eight and cytochrome b genes near direct repeat sequences and other hotspots of mutation. While allotopic expression, or even direct germline editing could theoretically be useful here, it seems that a simple mitochondrial transplantation is all that is necessary to fix things (11a).

The seemingly massive and disproportionate overhead of carrying nearly twice as many tRNAs as proteins is actually a primary feature of mitochondria rather than a bug. As genetic hybrids, these organelles exploit two completely separate and uniquely optimized transcription and translation systems. (In some species, there are even three or four separate genetic systems that have been retained within different organelles.) This meta-genetic cooperative works in tandem to simultaneously construct the respiratory complexes with the mitoribosomes cranking away in the matrix while the ribosomes work at the outer membrane facing the cytosol.

-Maybe add picture of dual ribosome system?,

Eukaryotes, which posses a bacterially-derived glycolytic pathway, use ribosomes that ultimately root from an archeal lineage. Mitoribosomes, on the other hand, stem from bacterial ribosomes. There is only one reported instance in the literature of mitoribosomes naturally translating nuclear RNAs, and that is for some sperm mitochondria (12) There is a marvellous new theory that explains how the composition of different kinds of ribosomes is precisely tailored to what they need to accomplish. The authors note that cytosolic ribosomes are autocatalytic in the sense that they beget other ribosomes. Except for a couple of rRNAs, mitoribosome parts are made by cytosolic ribosomes.

It was realized that the ribosome doubling time imposes significant constraints on the cell doubling time. The smallest and fastest translating ribosomes are those of bacteria which are under the most selective pressure for fast biogenesis. Their ribosomes, which contain the shortest ribosomal proteins and the highest mass percentage of rRNA (70%), require only six minutes to make a new set of ribosomal proteins. While their rRNAs vary greatly in size, the short proteins all turn out to be roughly the same length. The length of the rRNAs does not matter so much, only that they are made in the proper stoichiometry. Like the respiratory subunits encoded in mitochondria, this is ensured making single long transcripts which are later cleaved into final products.

Eukaryotic ribosomes are somewhere in the middle of the ribosome world -- they have more proteins, each of longer length, and have less of a need for rapidly synthesized rRNA content. At the far end of the spectrum are the mitoribosomes, which have the largest protein mass (80%) and the highest number of ribosomal protein subunits (~80), each with longer average length. The authors came up with a theory that neatly accounts these distributions. Specifically, a formula for the minimum fraction of time ribosomes spend on their own generation as a function of cell doubling time, number of proteins in a ribosome, and the time it takes a ribosome to make a new set of those proteins (13).

If theory can accurately predict how mitochondria should outfit and apportion their mitoribosomes, we need not consign these kinds of details to ’frozen accidents’. Perhaps we can generate theories for how mitochondria should conduct other business. In particular we should like similar theories of operation for each respiratory complex. Ribosomes may be more structurally complicated, but they are still basically enzymes -- enzymes which take the concept of cofactors to an extreme; The prosthetic groups of ribosomes are not iron-sulfur clusters, cytochromes, or flavins, but rather, a singular exchangeable system of virtualized cofactors (the tRNAs), that are in turn charged by a massive arsenal of tRNA synthetase enzymes.

Eventually we would desire that the respiratory and metabolic theory somehow adds up to a morphological theory of how mitochondria choose to fuse, fizz, die, and amble through different compartments in the cell. Beyond that, a budget to account for net creation and dissolution in the mitochondrial population as it migrates across cells and through the body at large. A couple things we know we should probably include at a fundamental level, are oxygen, the associated generation of radicals like superoxide, and perhaps a few nitrogen radicals.

To fully comprehend what mitochondria do it is instructive to examine how they get pressed into service in some of the unique anatomical specializations found across the animal kingdom. Whether it is the pit organs of vipers, shock box of the eel, or paracrystalline lens of the planarian eye, invariably these sensory exotica are packed to the hilt with unusual forms of mitochondria. In familiar organs, like hearts, cell-wide networks of mitchondria some 8000 strong fill 30 percent of the volume of the cell and tightly control contractility. They do this by rapidly uncoupling their intermitochondrial junctions when their own membranes are depolarized (13a). When forming a network, their individual cristae line up into  a continuous reticulum that spans the double membranes that separate them (13b). Incredibly, temperatures inside active mitochondria have apparently been apparently measured to hover around a scorching 50 degrees celsius (13c).

To look at one example in slightly more detail, consider constructing a theory of operation for the firefly light mantle (14). To turn on, the firefly needs to send boatloads of oxygen to peroxisomes where the luciferin based light reaction happens. This oxygen is freed up by shutting down mitochondria with nitric oxide (NO). Nerves release the transmitter octopamine onto nearby tracheal cells which form branching channels that penetrate the light mantle. The subsequent signal cascade leads to the generation of NO which effectively diffuses to mitochondria some 20um away in the cytoplasm of photocytes. Here NO shuts down oxygen use at cytochrome c oxidase (complex IV) by binding to its hemoprotein.

Removal of this inhibition requires electron transfer from haem. Curiously, the firefly flash itself may supply the off signal because inhibition of complex IV can be reversed by white light. The basic behavior of this relatively simple transduction pathway could be captured by static compartmental models involving O2, NO, and perhaps a few other molecules. If, however, we consider autonomous mitochondria capable of moving toward or away from peroxisomes, the situation gets complicated quickly.

There have been several intriguing attempts to capture global mitochondrial performance in terms of just a few parameters, in particular with regards to oxygen use and formation of radicals. One notable idea is that the crucial determinant of radical formation is something called the F/N ratio. This is ratio of electrons entering the respiratory chain via FADH2 vs. those from NADH. This ratio, along with radical formation, would be low during glucose breakdown and high during fatty acid breakdown. The longer the fatty acid, the higher the F/N. Ideally then, the longest fatty acids should be broken down in a separate place where the extra FADH2 won’t lead to excess mitochondrial radical formation.

If the power of a theory is what it can explain, we might say its usefulness would be what it can predict. In this case the theory offers an explanation both for the evolution of peroxisomes, and for the absence of fatty acid oxidation in long lived cells like neurons. It may also explain several eclectic mitochondrial refinements like carnitine shuttles, uncoupling proteins, and multiple antioxidant mechanisms linked to fatty acid oxidation. Depending on the species and tissue, peroxisomes perform many unique functions for the cell. One thing the theory predicts is that β-oxidation would be the most conserved and most ancient pathway across all peroxisomes. (15)

Another attempt to define mitochondrial performance has been to split operation into three primary operating modes for the purpose of quantifying how radical production, particularly superoxide anion (O2•−), occurs in each. In the first case mitochondria are in a normal mode of making ATP resulting in a more oxidized NADH pool, reduced membrane potential ( Δp ), and consequently a negligible O2•− production. In a second mode, mitochondria are not making ATP and consequently have a high Δp, a reduced CoQ (coenzyme Q) pool and CoQH2/CoQ ratio. This results in reverse electron transport through complex I and significant generation of  O2•−. In the third mode,  O2•− is also high, in this case resulting from a reduced NADH pool or high NADH/NAD+ ratio in the matrix(16). The ability to accurately measure O2•− in vivo, perhaps via EPR measurements would be vital to testing these kinds of theories (16).

Return to reality

A curious thing has unfolded in aging research that seems to have caught many off guard. That thing is the realization that free radicals (like dual genetic systems) are actually features rather than just bugs. What was the response of the allotopic community and SENS Foundation to this newsflash? I have not heard, but I can tell you what I think their response should have been. I would argue that rather than trying to allotopically diminish the role of our mitochondria, what we should be doing instead is trying to enhance them. In other words, give them more genes and more powers where it might make sense. This is something we can already do.

Genetic engineering of mitochondria is not as advanced as for the nucleus. There are not as many good restriction enzymes available, and advanced methods like CRISPR will not work for nucleoids because mitochondria lack the ability to do many kinds of DNA repair taken for granted in the nucleus. In particular, mitochondria do not do non-homologous end joining. But there are many new methods now in the pipeline to modify the mitochondrial genome that do not need the repair machinery (17). It is also possible to get new proteins to mitochondria without doing full blown allotopic expression.

For example, one group recently reported the ability to optically control mitochondrial membrane potential and ATP generation by transfecting cells with cDNAs for optically gated channels. (18) In this case it was just a matter of getting the right leader sequences so that the proteins would be put into the mitochondria instead of the plasma membrane. This is heady stuff. For one thing, it should also be possible to transfect mitochondria with optical indicators (18a). If we might briefly opine, the ‘Brain Machine Interfaces’ (BMIs) we have now are basically gimmicks (18b). On the other hand, an all-optical BMI using specially modified mitochondria would be the real deal.

Similarly enticing, would be the ability to swap out a few mitochondria to better match your metabolism to your environment. There are specific mitochondrial haplotypes that are ideally suited for different temperatures and altitudes. Need to burn brown fat in a warm sea-level location? There’s a hap for that. In particular, the mtDNA ‘uncoupling’ variants U5a and U4 that are common in Northern Europe reportedly give you greater on-demand fat burning ability. The downside is they may have a higher Parkinson’s risk associated with them (18c).

Today researchers and clinicians alike invariably still speak of ‘heteroplasmy’ (having more than one brand or haplotype of mitochondria in a cell or body) as a negative thing. This doesn’t have to be the case. There is no evidence that two heterogeneous populations of mitochondria cannot peacefully coexist and complement each other. The fate of mitochondria with a few artificial optoproteins inside the body would be unknown. Fortunately, we could watch them. We can even watch endogenous mitochondria that have not been modified in any way. For example, two-photon-excited fluorescence of NADH in mitochondria has been used to image mitochondrial reorganization in the skin, and distinguish basal cell carcinoma and melanoma (18d).

Charlie Gard

Charlie Gard was a child with severe mitochondrial depletion syndrome. He was made famous this past July after Donald Trump and Pope Francis made compassionate pleas on his behalf to an English hospital where he was being held. Eerily similar to the case of Justina Pelletier, a patient with mitochondrial disease at a Boston hospital, he was for all intents and purposes being kept as a medical prisoner. When Charlie’s parents wanted him to see an American doctor and try a seemingly rational (if a long shot) therapy to replace the nucleosides his mitochondria could not make, a court had to intervene to stop the English hospital from pulling the plug on his respirator.

Charlie had the misfortune of inheriting two mutated copies of a Ribonucleoside-diphosphate reductase (RRM2B) gene from his parents. RRM2B is needed in senescent cells for DNA repair and for mtDNA synthesis. Another form of ribonucleoside reductase is used in dividing cells. Unable to make enough mtDNA, what Charlie really needed was an infusion of good mitochondria.

The conventional wisdom has been that you can’t just inject extra mitochondria into your bloodstream like you would in a transfusion. For one thing, mitochondria are highly immunogenic by virtue of their bacteria-style formylated peptides and undermethylated CpG islands. CpG islands are regions of their genome that can be given extra methyl group tags by epigenetic regulatory enzymes. In fact, our immune system exploits this immunogenic property of mitochondria in order to self-activate (19).

Notwithstanding this settled science, a Chinese group recently found evidence that bloodstream injection of mitochondria in rodents can apparently fix certain kinds of brain damage (20). They reported zero immune activation in the acute phase for the first 2 hrs after injection. Intriguingly they discovered that cells from many different organs had taken up functional mitochondria from the initial injection stock. A likely explanation is that blood or endothelial cells absorb the mitochondria right after injection and pass them on to other distribution channels. I asked the researchers if they had found evidence of immune reaction in the longer chronic phase, and also suggested that they might try injecting mitochondria directly into the ventricular system of the brain. Both experiments are now planned.

If the mitochondria can pass the blood brain barrier, or alternatively can be supplied more directly to key access points in the nervous system, then many new possibilities open up. For one, like the testes, the nervous system is immunoprivileged. The peripheral immune system rarely gains access to the brain. Another advantage is that nerves reach out and touch many exclusive locations that capillary beds only reach by diffusion. For example, they directly contact sensory corpuscles, hair follicles, bone marrow sites of hematopoiesis, and niche stem cell populations in the epithelial sheets lining many organs in the body.

An important finding in these stem cell populations is that mitochondria directly control the fate and senescence of daughter cells. One way they do this is by sorting themselves out inside the cell according to age, membrane potential, and other indicators of status. The daughter cell that differentiates into a post-mitotic state asymmetrically receives more and younger mitochondria while the daughter that retains stem cell character retains fewer but older mitochondria (21). When cells or their progeny turn cancerous, it is frequently because their mitochondrial function has been compromised in some way. Furthermore, what often transforms a quiescence tumor into an invasive cancer after the initial metabolic pivot away from mitochondrial respiration is donation of fresh mitochondria from nearby cells (22).

The other major element that controls cell fate, and particularly cancer, is nerves. The list of nerve-controlled tumors has expanded to include multiple cancer types in multiple organs. Cholinergic innervation of stem cell crypts times the maturation, proliferation, and tumorigenicity of cells in the gut. Nerves specializing in other factors, like GDNF, control pancreatic cancers. Still other nerve channels feed basal cell carcinomas or melanomas in the skin. The control of cell senescence and cancer by nerves has several things in common with the control of tissue regeneration by nerves. Salamander digits and tadpole limbs readily regrow when when cut off, but all regenerating tissues seem to require something from nerves in order to do this. Take away the nerves, and regeneration stalls.

In order to tie these observations together we need to consider one further fact. That fact is that nerves take the donation and absorption of mitochondria to an extreme. As we will see shortly, we might say it is their bread and butter.

With that in mind, the full Monty is now laid bare;

The nervous system controls the rate of aging throughout the body, and maintains its populations of cells in various states of senescence and proliferation by apportioning mitochondria. In this light, cell senescence is not quite so irreversible as one might assume (23).

John McCain

What about cancers of the brain itself? By one count there are now at least 80 treatments in various stages of testing and trial to defeat glioblastoma. Many of them use viruses like herpes, HIV, or polio, or even combinations of viruses, drugs, and immunotherapies against the tumor. A unique feature of glioblastoma is its resistance to all forms of treatment. When it is knocked down by drug, surgery, and radiation, it reorganizes and returns with a vengeance. Whether these gliomas originate from neural stem cells, glial progenitors, or astrocytes, the feature that defines them is their formation of one continuous syncytium (24).

By watching these networks evolve on timescales from minutes to years it was discovered that nuclei and mitochondria constantly scan the entire glioma through the membranous tubes that interconnect it. It has been shown that when cells are irradiated, mitochondria bear the brunt of the damage. (25) Incidentally, mitochondria may be the primary targets of damage by excessive ultrasound. It is believed that one source of the remarkable resistance of gliomas to radiation is the ability to rapidly repopulate and energize the network with fresh mitochondria. Something that advocates of allotopic expression might want to keep in mind, is that should they ever succeed in installing their nuclear backup copies of mtDNA, their tumors may not be as susceptible to radiation therapies.

As far as John McCain’s prognosis, one thing that could be in his favor would be for his brand of tumor to be an isocitrate dehydrogenase (IDH) mutant version. IDH mutants typically respond better to treatment and patients have longer mean survival, but they are a less common form of glioma. A related finding is that tumors with a certain deletion, the 1p/19q codeleted tumors, are invariably also IDH mutated. The interesting thing about these tumors is that they lack the elaborate intercellular membrane tubes, and do not form radiotherapy-resistant astrocytoma networks. Many of the behind the scenes links between nuclear alterations and mitochondrial dysfunction, and vice-versa, are now coming to light. One crafty mechanism the nucleus employs to sense mitochondrial status is to attach mitochondrial localization sequences to the front end of a dual purpose transcription factor, and nuclear localization sequences to the back end.

For example, if the nucleus reels off copies of a transcription factor known as ATFS1, but the mitochondrial can’t import it because their membrane potential is too weak for uptake, it will eventually find its way back to the nucleus where it can regulate genes to rev respiration back up (26). This level of feedback would be difficult to install in artificial allotopic expression.
While we lack a complete understanding of mito-nuclear regulatory crosstalk in cancers like glioblastoma, some inferences are now possible. The regular migration of nuclei and mitochondria within the gliomal syncytium might be considered to be a throwback to a much earlier time. One theory, due to Garg and Martin (27), places the symbiotic association with mitochondria at the founding of a multinucleated eukaryotic cell. They propose that eukaryotic chromosome division arose in a filamentous, syncytial ancestor. Individual nuclei inside the proto-syncytium with insufficient chromosome numbers could complement each other through mRNA in the cytosol, and generate new chromosome combinations through nuclear fusion (karyogamy). The theory explains why the mechanisms for eukaryotic chromosome separation are more conserved than those for cell division. It also explains the origin of sex, and accounts for sequence of the evolutionary appearance of meiosis and mitosis.

The authors show that the energy provided by mitochondria relieved a major constraint on the ability to produce tubulin for chromosome separation. The ability to make tubulin in sufficient quantities to nucleate microtubules also made things like neurons, axons, and organelle transport possible. A remarkable feature of the methods used to transmit mitochondria, namely cytoskeleton infused membrane nanotubes and endo/exocytosis, is that they look a lot like those features the brain normally employs in its daily operation. That is to say, axons, dendrites, and vesicle-mediated synaptic transmission.

It is not to difficult to imagine that in the early evolution of nervous system mitochondria began to push out these neuritic proto-appendages in order to escape predation and degradation by the lysosomes in the locale of the nucleus. Fully arborized neurons eventually rectified themselves into polarized circuits transmitting from dendrite to axon, with mitochondria along for the ride. One predictable side effect of this would be that neurons became selectors of mitochondria (28).

Much of the basic allometry of nervous systems (like relative lengths and branching patterns of dendrites and axons, or thickness and gyrification of cortex and cerebellum) can not be adequately explained by the electrotonic computational theory that traditional neuroscience continues to offer. Instead, these features may have been shaped more by what mitochondria within those neurons and circuits needed to accomplish. In the view, rapid signal transduction and ‘computation’ would have arose as a side effect of this primal function.

If neurons are devices to select mitochondria, we might ask when they require net creation and transmission of mitochondria, and when their sequestration and removal? Under what circumstances are whole mitochondria transferred, and when just their nucleoids? Similarly, neurons with distinct phenotypes may select the particular transmitter systems they use in successive legs of a neural circuit based on the metabolic needs of their mitochondria. This entails certain advantages and vulnerabilities for synapses competing for real estate within the brain. The immediate result of this transmitter specialization and competition for survival can be seen directly in the microstructure of grey matter; The thorny excrescences, synaptic ribbons, and spiney involutions of neuropil accrete into nested triads and glomeruli around their core  sources, directionally passing different metabolites between center and periphery. In the grey matter, this source is often a single apical dendrite or capillary, and in the white matter, a node or hillock.

Consider cells which use the transmitter dopamine. They are uniquely susceptible to loss of their mitochondria, particularly in Parkinson’s disease or when certain drugs are imported into their nerve terminals. While Parkinson’s is a blanket term, different forms of the disease can be distinguished by detection of mitochondria vented into the CSF. Incidentally, mitochondria are also found in the CSF during stroke (29). Stroke is interesting here because it is now known that astrocytes can rescue ailing neurons after their blood supply is interrupted by donating mitochondria (30). In idiopathic Parkinson’s disease, a high concentration of mtDNA in the CSF has been found in patients that carry the LRRK2G2019S mutation (31). To replenish mitochondria in a disease that uniquely targets certain transmitter systems like dopamine, one thing we would want is a circuit diagram for mitochondrial flow through the striatum.

How do we create mitochondrial maps for the body?

A significant problem with many common chemotherapies is they take a huge toll on the hematopoietic system that generates new blood cells. It turns out that the drugs destroy the adrenergic nerve endings that contact stem cells niches in the bone marrow to promote their survival. Without these sympathetic nerves, proliferation and differentiation of hemotpoietic cells grinds to a  halt (32). In acute myeloid leukemia (AML) the bone marrow is  infiltrated by a clonally expanding population poorly differentiated blast cells. Compared to normal CD34+ hematopoetic stem cell progenitors, these cells are overloaded with mitochondria and rely on their oxidative phosphorylation to generate ATP. This contrasts with most common tumors which generate their ATP by ’aerobic glycolysis’ in accordance with the Warburg hypothesis.

It was recently discovered that the mitochondria responsible for transforming malignant AML blasts are supplied by bone marrow stromal cells (BMSCs). This heterogenous population of local stromal cells includes precursors of endothelial cells, osteoclasts, osteoblasts, adipocytes, and fibroblasts. The way the AML blasts get the mitochondria is by extending tunneling nanotubes to the BMSCs. The way that blasts make the nanotubes is by generating superoxide radical with NADPH oxidase (33).

We just described how sympathetic nerves trophically support the marrow stem cell pool, and how the marrow cells support AML blasts. What do we get when when put these two observations together?  One exemplary possibility we can infer is a multi-hop mitochondrial transfer circuit from nerve to marrow cell to AML blast.
Better living by defeating male pattern baldness

Under the Warburg effect, cancer cells oxidize a decreased fraction of the pyruvate (generated by glycolysis) in their mitochondria. Instead they ferment it to lactate to generate NAD to rerun glycolysis. The lactate then fuels other pathways or is excreted. One strategy against colon and prostate cancers has been to re-express or upregulate various carriers that ferry pyruvate into mitochondria (34).
Hair follicle stem cells (HFSCs) utilize glycolytic metabolism to produce significantly more lactate than other cells in the epidermis. This lactate appears critical for HFSC activation because deleting lactate dehydrogenase blocks their activation while deleting mitochondrial pyruvate carriers accelerates it. Mitochondria would therefore directly influence the unique pattern of oscillation between proliferation and quiescence that is seen in follicular cells. Balding men everywhere recently got some very good news; Blocking mitochondrial pyruvate carriers with a molecule called UK-5099 regrows hair (35).

Nerves like to feed on lactate.  During aerobic exercise the brain acts largely as a processor for muscle released lactate. Axons use a monocarboxylate transporter called MCT2, and astrocytes a different one called MCT1. Both work similarly to the mitochondrial pyruvate carriers, but the direction of transport depends on local conditions. Evidence from the optic nerve suggests that astrocytes release lactate and axons gobble it up (36). This trophic pathway parallels a reciprocal transexudation of mitochondria from axon to glial cell in the optic nerve head (37).

As with the above example for bone marrow stem cells, we might pool these observations about hair and nerves to infer another potential mitochondrial circuit from nerve to follicle. In this circuit nerves would be attracted to follicles by the abundant lactate they produce, and control their cycle of activation via mitochondria.

Where does all this leave aging?

For people with serious mitochondrial dysfunction or depletion, knowing the status of their mitochondriome is of vital importance. It is difficult to tell what is going on in the entire body just by sampling the blood, but it is the easiest place to begin. It has not escaped the notice of many people in the anti-aging community that the status of the mitochondriome might be of considerable interest to anyone. This information would not simply be the amount of mitochondria in your blood cells and a static sequence of your mtDNA. Rather, it would be a living sequence database that is continually updated from many sources to document how heteroplasmy evolves in the body.

Heteroplasmy is a woefully neglected issue in medicine. Consider current efforts to make transplantable organs from pigs. To make them compatible with humans, researchers have gone to great lengths to delete sequences known as PERVs (porcine porcine endogenous retroviruses) from the pigs (38). Little mention is made that putting a pig liver into a human would introduce a bucketload of pig mitochondria. Experiments done a couple of decades ago have given some indication that human cells may do okay with heteroplasmic mitochondria from other species. One researcher from Advanced Cell Technology created hybrid cow-human embryos by fusing his own cells with cow eggs that had their nuclei removed. The cells, which retained cow mitochondria along with the human nucleus were in fact quite viable (39).

To get an idea how the mitochondriome is evaluated with our current technology, consider some recent advances in understanding and treating Friedrich’s ataxia. This neurodegenerative disease caused by inherited deficiency of the mitochondrial protein Frataxin. Frataxin makes the iron-sulfur clusters that are needed by respiratory complexes I,II, and III. These clusters are also needed for DNA replication and repair. By measuring nucleoid copy number as ratio of mtDNA over nuclear DNA by quantitative PCR it was found that frataxin mutations cause deficiencies in mitochondrial biogenesis (40).

A second study showed that a drug now used to treat multiple sclerosis called dimethyl fumarate (DMF), could boost mitochondrial numbers. The authors suggest that DMF might be used to treat a broad class of mitochondrial deficiencies. If so, this would probably the first actual mitochondria-boosting drug, despite the claims of many supplements on the market. To see if it is actually working, one might simply count mitochondria in bone marrow derived platelets in the blood (41).

Ultimately any drug given to the body at large is going to affect metabolisms all throughout it. As far as dealing with cell senescence and aging there simply has been no drug that has been unambiguously proven to be worth the trouble -- there are always significant side effects. Controlling the life cycle of cells, tissues, and the body itself will require controlling the nervous system and mitochondria distributed by it.


1) Mitochondria in Cell Senescence: Is Mitophagy the Weakest Link?
   Mitochondrial Dysfunction Meets Senescence
   An attempt to prevent senescence: A mitochondrial approach

2) Artificial Mitochondria Transfer: Current Challenges, Advances, and Future Applications

3) Mitochondria Know No Boundaries: Mechanisms and Functions of Intercellular Mitochondrial       Transfer

4) Nerve Dependence: From Regeneration to Cancer

4a) Myocardial Rescue with Autologous Mitochondrial Transplantation in a Porcine Model of Ischemia/Reperfusion

   Critically reviewed on hacker News

6) Local clearance of senescent cells attenuates the development of post-traumatic osteoarthritis and creates a pro-regenerative environment

7a)  Visualization of cytosolic ribosomes on the surface of mitochondria by electron cryo‐tomography

8) Stable nuclear expression of ATP8 and ATP6 genes rescues a mtDNA Complex V null mutant

8a) tRNAs and proteins use the same import channel for translocation across the mitochondrial outer membrane of trypanosomes

10) Why chloroplasts and mitochondria retain their own genomes and genetic systems:    Colocation for redox regulation of gene expression

11) The Assembly Pathway of Mitochondrial Respiratory Chain Complex I

Regulation of Mitochondrial Complex I Biogenesis in Drosophila Flight Muscles

Leo Nijtman’s criticisms to the SENS allotopic paper  
1. What are the specific tricks they did to improve the allotopic expression (compared to earlier work).
2. Why did they only use the ATP8 mutant for the allotropic (stable?). Why not also wt cells, rho zero, or NARP (8993) cybrids?
3. I think that it would be elegant to use classical import assays to show the import of the allotopic gene. Perhaps also Prot K treatment to show that the proteins are really imported.
4. Blue native gels show co-migration with HMW complexes (not necessarily CV). It would have been good if they had also used other approaches to show the incorporation of the allotopic protein. e.g co-IP or complexome profiling could be informative.
5. I was also surprised that there was no low molecular weight species of A8 (unincorporated in the complex) in the BN gels.
6. Also why does the stable expressed allotopic gene does not show a unprocessed species (uncleaved presequence).

11a) Mitochondrial transplantation for therapeutic use

12) Mammalian sperm translate nuclear-encoded proteins by mitochondrial-type ribosomes

13) Ribosomes are optimized for autocatalytic production

13a) Power Grid Protection of the Muscle Mitochondrial Reticulum

13b) Trans-mitochondrial coordination of cristae at regulated membrane junctions

13c) Mitochondria Are Physiologically Maintained At Close To 50 C

14) Nitric oxide and control of firefly flashing

15) On being the right Q: Shaping Eukaryotic evolution'

18) Optogenetic control of mitochondrial metabolism and Ca2+signaling by        mitochondria-targeted opsins

18b) Pure optical detection of spikes for the ultimate brain machine interface

18a) Multibeam femtosecond optical transfection for the ultimate brain interface

18b) Pure optical detection of spikes for the ultimate brain machine interface

18c) The Impact of Mitochondrial DNA and Nuclear Genes Related to Mitochondrial Functioning on the Risk of Parkinson’s Disease

18d) Imaging mitochondrial dynamics in human skin reveals depth-dependent hypoxia and malignant potential for diagnosis

Platelets release mitochondria serving as substrate for bactericidal group IIA-secreted phospholipase A2 to promote inflammation

Mesenchymal stem cells use extracellular vesicles to outsource mitophagy and shuttle microRNAs

20) Intravenous administration of mitochondria for treating experimental Parkinson's disease

Isolated Mitochondria Transfer Improves Neuronal Differentiation of Schizophrenia-Derived Induced Pluripotent Stem Cells and Rescues Deficits in a Rat Model of the Disorder

21) Asymmetric apportioning of aged mitochondria between daughter cells is required for           stemness

22) Horizontal transfer of mitochondria in sickness and in health

23) How the nervous system controls tumor growth   

24) Brain tumour cells interconnect to a functional and resistant network.

     A malignant cellular network in gliomas: potential clinical implications    

Mitoception: Transferring Isolated Human MSC Mitochondria to Glioblastoma Stem Cells.

25) Effects of ionizing radiation on mitochondria

26)Mitochondrial and Nuclear Accumulation of the Transcription Factor ATFS-1 Promotes OXPHOS Recovery during the UPRmt

27) Mitochondria, the Cell Cycle, and the Origin of Sex via a Syncytial Eukaryote Common Ancestor

28) The Origins of Polarized Nervous Systems

29) Extracellular Mitochondria in Cerebrospinal Fluid and Neurological Recovery After Subarachnoid Hemorrhage

30) Transfer from astrocytes to axon in stroke

31) Mitochondrial DNA in CSF distinguishes LRRK2 from idiopathic Parkinson's disease
     Leucine-rich repeat kinase 2 (LRRK2) also called dardarin, from the Basque word                    ‘dardara’, which means trembling

32) Chemotherapy-induced bone marrow nerve injury impairs hematopoietic regeneration.

33) NADPH oxidase-2 derived superoxide drives mitochondrial transfer from bone marrow stromal cells to leukemic blasts

34) Inhibition of pyruvate carrier widely explored in cancer, prostate, colon cancer, and warburg effect repessor in general

35) Lactate dehydrogenase activity drives hair follicle stem cell activation

36) Transfer of glycogen-derived lactate from astrocytes to axons via specific monocarboxylate transporters supports mouse optic nerve activity.

37) Transcellular degradation of axonal mitochondria

38) Inactivation of porcine endogenous retrovirus in pigs using CRISPR-Cas9

39) Researchers claim embryonic mix of human and cow

40) Frataxin deficiency impairs mitochondrial biogenesis in cells, mice and humans

41) Dimethyl Fumarate Mediates Nrf2-dependent Mitochondrial Biogenesis in Mice and Humans.