Lamarckian Evolution, part 2

Lamarckian EPIgenetic inheritance is well-accepted. Evidence for actual GENETIC inheritance is strong.

Last week, in Part 1, I explained the concepts behind these three models of evolution. Can habits and exposures during your lifetime affect your children? Darwin thought so, though this part of his legacy was written out of 20th century neo-Darwinian theory. I wrote about the plausibility of blind, random mutations for explaining the pace of evolution, and the diversity of adaptations, and I conceded that it may be impossible to make this argument quantitative.

In Part 2, below, we’ll see that epigenetic inheritance is fully Lamarckian, and this has been accepted, mainstream science for about 20 years. James Shapiro has done experiments demonstrating that in bacteria, Lamarckian inheritance is genetic, i.e., it affects the primary structure of DNA. This is easier to accept because bacteria have a single cell, so any modification of the genome is passed on to offspring; whereas in multi-celled “higher” animals and plants, an archival copy of the genome is preserved in the gonads exclusively for transmission to the next generation. How could feedback from life experience be applied to restructure the germline archival copy? In recent years, molecular mechanisms have come to light that make this plausible. At the end, I’ll propose experiments to look for Lamarckian inheritance in fish, plants, and worms.

---

Lamarckian epigenetic inheritance is now un-controversial, mainstream science

The term “epigenetic inheritance” refers to any mechanism that is not coded directly into DNA. The best-established kind of epigenetic inheritance occurs through molecular decorations and markers that surround the DNA and affect which genes are expressed and which are held in reserve for another time and place. Methylation of DNA and acetylation of histones are two of the best-studied markers that affect gene expression. (Histones are molecule-sized “spools” around which DNA is wound so that it doesn’t become tangled in storage. Methylation refers to a modification of one of the four letters in the DNA alphabet. The “C” in ACTG stands for cytosine, and when one extra carbon atom is added, it becomes 5-methylcytosine. Methylized cytosine is a universal language that says “keep the adjacent gene silent for now”. There are stretches of DNA that are made up just of C’s (paired to G’s on the other side of the double helix) and G’s (paired to C’s that can be methylated). They are called CpG islands and they carry no genetic information of their own, but exist to regulate expression of nearby genes.

Methylation and acetylation patterns can change in response to habitual activities and to the environment. These patterns are copied with the DNA – not quite so faithfully as the DNA itself – and can be passed from parent to offspring through multiple generations. This is a kind of temporary Lamarckian inheritance. It is indisputably Lamarckian, but seems to last for a handful of generations if it is not reinforced.

Examples

• The classic study followed children of Dutch mothers who were pregnant during the war-induced famine of 1944-45. Deprived of food, the mothers developed gene expression that was more frugal with the calories they ate, burning less and storing more (fat) for later. Children of the hungry Dutch mothers were at higher risk of obesity as a result of inheriting the mothers’ epigenetic adaptation. [ref]

• Children of obese mothers have greater risk of insulin resistance and diabetes [Ref]. (This inheritance is both genetic and epigenetic, and we have to trust that the authors of the studies cited here correctly separated the two with their statistical filter.)

• Traumatized mother mice are affected in their metabolic as well as their psychological responses, and these adaptations are detectable in the offspring of the traumatized animals out to at least the fourth generation [Ref]. Ten years ago, a Nature article was published about male mice that transmit the effect of trauma to their young, and two more generations beyond.

• The behavior passed to offspring can be quite specific. Mice that were shocked after exposure to a specific (pleasant) smell developed a panic response to that smell, and passed that specific response to their offspring for at least several generations. In this study, the authors traced the response to a particular methylation pattern.

• In this study, inbred plants were exposed to “drought conditions” (in a greenhouse) and they developed water-conserving adaptations, which were passed through methylation to the next generation of plants.

• “One of the most dramatic examples is with diethylstilbestrol, a synthetic nonsteroidal estrogen prescribed in the 1970s to prevent miscarriage in women with prior history. While the drug helped pregnancies to go to term, it induced severe developmental abnormalities and increased the risk for breast cancer and a rare form of adenocarcinoma in girls whose mothers were exposed to the drug during the first trimester of pregnancy. Furthermore, the risk of cancer appeared to be transmitted to the following generation. A clinical study reported that a 15-year-old girl whose maternal grandmother was exposed to diethylstilbestrol during pregnancy was diagnosed with a very rare case of small cell carcinoma in the ovary. Many more of maternal granddaughters than expected also developed ovarian cancer. Although these findings are among the first and need to be confirmed by further transgenerational studies, they suggest that the detrimental effect of a drug can be transmitted across generations. Such transgenerational effect of diethylstilbestrol was also observed in mice. Similar to humans, perinatal exposure to the drug induced abnormalities in uterine development and uterine cancer in first and second generations. These abnormalities were suggested to result from aberrant DNA methylation in a gene that controls uterine development and in uterine cancer genes.” [from Franklin and Mansuy, 2009]

• Here’s an example that’s not really Lamarckian, but that clearly demonstrates epigenetic inheritance. There’s a mutation in a gene called Kit that causes brown mice to have white spots. One copy of the gene is enough to cause the spots. So experimenters crossed a mother mouse with one copy of the gene with an un-mutated father mouse that had no spots. According to standard Mendelian genetics, we would expect that half the offspring of the cross would get the Kit mutant gene from their mother, and half would get the mother’s normal gene. So they expected half the offspring to have spots. The surprise was that all the offspring had spots. With DNA tests, they checked and, as predicted, only half the offspring had the mutated Kit gene. Still, they all had spots. Epigenetics! The experimenters figured out that the mutated gene signals the body to silence the other copy with methylation. So the offspring mice inherited a methylated version of the normal gene from their mothers. The methylation was copied along with the DNA. [Ref]

Here is a Stanford study that isolated an epigenetic component of longevity inheritance in worms.

Eva Jablonka is an Israeli geneticist who realized early the importance of epigenetic inheritance, and has been writing about the subject for 30 years. Here is a review article from 2009 in which she lists hundreds of examples of epigenetic inheritance.

We know that response to early trauma can have deep and lasting effects on people’s attitudes and relationships. The reality that trauma responses are passed on for untold generations has disturbing implications for sociology and politics. Will the wars and genocides of our time seed cycles of violence for many generations to come?

There are hopeful (and heroic) examples of people who have endured the most horrific torture in childhood, and find the inner strength to overcome their triggered responses to take on deeply empathetic projects as adults. Simone Weil, Elie Wiesel, Vera Sharav, and Viktor Frankl survived Auschwitz. Fifty voices tells the stories of women who endured Satanic abuse and ritual torture as infants. Epigenetics is a powerful determinant of behavior, but human resilience is yet more powerful.

But where did epigenetic inheritance come from?

A question which I have not seen to be addressed in the literature is this: The epigenetic inheritance mechanism is itself permanently installed, presumably with a basis in the genome. So how did the mechanism of epigenetic inheritance come to be? Here is a prime example of irreducible complexity! Copying the methylation state requires a whole different set of enzymes from copying the DNA bases. Epigenetic inheritance offers many potential advantages over the long term, but it is not an adaptation that offers fitness benefits in terms that neo-Darwinian theorists would recognize (survival or fertility). So the very existence of epigenetic inheritance is a challenge to the standard (dying) neo-Darwinian paradigm.

In addition, it is agreed in the mainstream that mutations increase in response to stress

Epigenetic inheritance is a well-accepted Lamarckian mechanism, but it is temporary, and doesn’t affect the DNA itself. Is there also Lamarckian influence on the DNA?

Normally, DNA is replicated accurately, with negligible errors, but in times of stress something different happens. It was once described as a breakdown of the cell’s proofreading facility under stress. But it has now become a mainstream idea that this is no accident, that the cell flails at random, trying wild cards when it is clear that the standard strategy is not working so well. Jim Shapiro goes further, and describes “conservation in times of successful growth as compared to active restructuring in times of stress.” [my emphasis] The stressed organism is not just trying more random things to see what works; it is actively and intelligently re-engineering its own genome. Shapiro’s position stands out from the crowd, and he has credentials that suggest we ought to listen. My belief is that he is pointing the way to the future.

For example, it was once thought that UV radiation damages chromosomes, a purely physical effect of high-energy photons. The truth that has emerged is that the cell detects the UV as a stressor, and mutates its own DNA, under metabolic control, as part of an adaptive response. This was discovered already in the 1950s by Swiss microbiologist Jean Weigle. Whether the mutations are random or whether they are part of a directed response to the radiative environment remains controversial.

True Lamarckism in bacteria

The classical model for investigating this question is the common E. coli bacteria. These bacteria can normally live on two kinds of sugar, glucose or lactose. For the purpose of experiment, a gene is disabled, preventing them from being able to digest lactose.

• If you put the mutated bacteria in a glucose medium, they do fine, and do not re-evolve the gene to digest lactose.

• If you starve them, with neither glucose nor lactose, they go into stress mode, increase their rate of experimental mutations, and re-evolve the gene to digest lactose.

• If you put them in a medium containing lactose but not glucose, it is claimed that they re-evolve the gene for lactose digestion more quickly. The perceived utility of digesting lactose stimulates them to acquire this ability efficiently.

But does this really happen? It has been a controversial claim for forty years. The experiments are not so easy to interpret, first because bacteria readily incorporate genes from their environment in the form of DNA loops called plasmids; and second because in the absence of lactose, it is hard to know whether just one bacterium out of many billions might have acquired the ability to digest lactose. [References: Original Nature article by John Cairns, 1988 proposing Lamarckian mutations in E coli; Davis 1989, a suggested mechanism; a quasi-Lamarckian view, 1990; a follow-on experiment 1996; another traditional explanation, 1997; Statemaster Encyclopedia article; a 2010 review of Lamarckism in bacteria; radical Lamarckian view based on quantum information]

From here, Shapiro takes a leap into full-blown Lamarckism

Shapiro has a thin, dense book called Evolution: A View from the 21st Century, in which he makes the case for a radical departure from the notion that evolution takes place via random mutations. He cites evidence that the “mutations” that appear under stress are far from random, that in fact the cell is re-arranging its own DNA, and doing so in a way that is much more likely than “random” to produce an adaptive response to the particular stress at hand. “Natural genetic engineering”, he calls it. He has spent much of his career documenting this effect in bacteria, but he claims that animals and plants have far more sophisticated abilities to rearrange their own DNA – it’s just that these are more difficult to see in the laboratory.

If Shapiro is right, then perhaps we can begin to understand the mystery of how evolution is so miraculously efficient as it seems to be. One way or another, we will have to leave traditional limits of neo-Darwinian evolutionary theory behind, and I believe that Shapiro’s work together with the literature of evolvability, provide the clearest roadmap we have for a new understanding of evolution.

Protists are single-celled eukaryotes, much more complex than bacteria, from which all multicelled life is descended. Protists also splice and dice their own DNA.

Do animals and plants edit their own genomes?

According to the textbooks, our bodies have developed from a single fertilized egg, and every cell in the body has the same DNA as that egg. The DNA in every cell in our bodies is identical (except for copying errors — rare somatic mutations).

This has been assumed without experimental support until 2012, when a team of researchers from Yale decided to test it out. They were surprised to find substantial variation from one tissue to another in the DNA of a single individual. They looked in particular for copy number variation, in which segments of the genome typically a few thousand BP long are duplicated. They found examples wherever they looked, and they uncovered evidence that this is not random but functional. For example, genes that are expressed in the pancreas have extra copies in pancreatic cells. Regulatory genes that operate at a high level were more likely to be duplicated than downstream genes or regions of non-coding DNA.

Most of the biological community still believes the textbooks, but this finding suggests that the body is capable of editing its own genome for functional purposes. The article says nothing about the mechanism by which it is accomplished, but whatever it is, it is not hard to imagine that that same mechanism is harnessed for a Lamarckian function.

Lamarckian restructuring of the genome

In this 2015 paper from Washington State biologists, no “new genes” are created as a result of life experience, but the genome is permanently changed with extra copies of existing genes. The authors exposed rats to a toxic fungicide, and confirm the previously-observed epigenetic changes in the rats, changes that are transmitted to their offspring. They then go on to breed the rats for three more generations, and note that there are extra copies of hundreds of genes, some of which are useful in the detox of the fungicide. These genetic changes appeared in the third generation after exposure, but they were absent in the first generation. They can’t be written off as mutagenic effects in the fungicide, because they were three generations removed from exposure.

This report does not claim creation of new genes or even new alleles, but it does include permanent changes to the germline DNA. The emerging view is that gene expression is more important in determining an organism’s structure and function (and fitness) than the precise form of the alleles (different versions of the same gene) themselves. 98% of our DNA is not genes but introns, the segments of DNA between genes that collectively determine the timing and circumstance of gene expression. A curious finding stressed by the authors of this study is that there is zero overlap between the areas of the genome that were epigenetically modified in Generation One after exposure and the areas of the genome that later produced extra copies in Generation Three. This suggests that the mechanism for this first example of Lamarckian genetic inheritance remains a complete mystery.

Borrowing genes from others

Bacteria are promiscuous. Their genomes are in the form of thousands of circular snippets of DNA called plasmids, and they constantly exchange plasmids with each other. But eukaryotes (including all plants and animals) are more conservative. The great majority of the time, we make do with the genes passed to us by our two parents. Still, there are exceptions. In times of stress, especially, plants and animals can pick up genes from other organisms and assimilate them. This is called “horizontal gene transfer” or “lateral gene transfer”. This Nature paper claims that lateral gene transfer is a kind of Lamarckian adaptation. A stressed organism is likely to find itself in an environment shared by bacteria that are already pre-adapted to that stress, and they sometimes make use of this, picking up from local bacteria just the genes that they will need in order to adapt to the local stress.

Retrotransposons: A candidate mechanism for Lamarckian Inheritance

Transposons are stretches of DNA that have the ability to copy themselves and insert elsewhere. They were discovered by Barbara McClintock 80 years ago, and later interpreted as parasitic DNA. The theory was that they provide no benefit to the organism, but they have evolved the ability to be expert hitchhikers. How do they copy themselves? Some transposons do it directly, DNA copied to DNA. Others copy themselves by transcribing to RNA, and then the RNA transcribes itself back into DNA somewhere else. DNA is supposed to be copied to RNA but not the reverse, so RNA transcription to DNA is commonly called “retro”. Hence transposons that copy themselves through RNA are called retrotransposons. (The HIV virus and the COVID virus are examples of “retroviruses” that spread as RNA but turn to DNA inside the cell.)

How is the copying and reverse transcription accomplished? “Long” retrotransposons, or LINEs (Long Interspersed Nuclear Elements), actually contain a region that codes for the requisite enzyme; “short” retrotransposons, or SINEs, depend on the protein provided by LINEs.

LINEs and SINEs together constitute 30% of human DNA. The most common are a kind of SINE known as Alu elements. There are over 1 million Alu elements, together making up 11% of human DNA.

Most researchers writing about transposable elements (TEs), including retrotransposons, regard them as random or (worse) “parasitic DNA”. I suspect that evolution is more efficient than this, and that anything lasting tens of millions of years has a purpose, whether or not we are yet able to divine what that is. In the case of Alu elements, the purpose is to affect DNA transcription, not just epigenetically but by locating strategically, so as to promote or suppress particular genes. The placement of genes on a chromosome has a major effect on where and when that gene can be expressed. DNA spools and folds and unwinds in complex ways that are barely understood, but it is certainly true that the local neighborhood of the chromosome is a factor in gene expression. So copying and relocating retrotransposons certainly affects gene expression, and there is every reason to think that natural selection has harnessed this for the body’s many purposes.

If LINES and SINES are moved around in different organs of one body, changing gene expression from one tissue to another, this can be part of the effect announced in the 2012 Yale study cited above. But once we know that the body has this ability, we might suspect that the same mechanism is harnessed in the germ line as well, thereby supporting true Lamarckian genetic inheritance. This is the hypothesis of three molecular biologists at University of Rochester, in a paper titled Retrotransposons as regulators of gene expression.

Curiously, the article begins and ends with the assumption that TEs are parasites that have learned to copy themselves, and that organisms have learned to work around them. But in between, the article cites a great deal of evidence that TEs have acquired functions, and have evolved to be essential for life. Perhaps retrotransposons — like mitochondria — had a parasitic origin once upon a time, but long ago became part of the machinery of evolution. When I see this kind of disparity within a single article between the message of the actual data and the framing of that message in the title and conclusions section, I suspect that the article has been edited to get past the conservative prejudices of the reviewers — which, in this case, I know to be strong.

Alu elements tend to be rich in methylation sites (CpG islands) which are places where the most common, best-understood kind of epigenetic regulation takes place.

Retrotransposons actively copy themselves, thereby restructuring chromosomes, during development. This accounts for some variation in DNA in tissues (documented in the Yale article mentioned above). There is also active copying throughout life within the brain, which makes me wonder if learning might be accompanied by restructuring DNA in the brain.

In this short video, Carl Zimmer and Job Dekker explain the importance of the intricate way that DNA is folded over on itself, helping to determine which regions are transcribed and which remain locked up as heterochromatin. The stretches of TE DNA certainly affect transcription, and they are re-programmable during an organism’s lifetime. We might expect as a matter of course that the number and placement of TEs has been subjected to natural selection, and has become highly adaptive in a way that responds to experience during a lifetime.

Of course.

We know for a fact that methylation programming extends back to the germ line, and accounts for heritable epigenetics. Now that we have a glimpse of the retrotransposon mechanism, why wouldn’t we expect it also to feed back and restructure the germline DNA?

Call for new experiments

Scientific bias against Lamarckian inheritance is an anachronism. Some modes of Lamarckian evolution have been firmly established. The most general and most permanent form has never been tested competently. The last remaining argument against it was the difficulty of imagining a plausible mechanism. What we have learned about retrotransposons and genetic variation among different tissues of the same body removes that objection.

The time is ripe for a well-planned exploration of Lamarckian inheritance in various circumstances, with a variety of animal and plant species, coordinated over multiple laboratories worldwide. At this point a “surprising” result is to be expected.

Experiments can be done with large animals or plants; in this case the generation time is long and the lab population is small, so the challenge is to see an effect that may be small. Alternatively, experiments can be done with large populations of C elegans worms. In this case, the challenge is to clearly distinguish breeding from mutations.

1. There is a Mexican fish [Borowsky, 2023] that has a blind cave-dwelling variety and a sighted stream-dwelling variety. The experiment would be to cross a blind fish with a sighted fish, and hatch half the eggs in darkness and the other half in a fully-lighted tank. Breed the two populations separately. Collect offspring of the population raised in darkness and the population raised in light, and compare how many of them have eyes.
   
   It will then be necessary to analyze the two genomes to demonstrate that the difference is really genetic and not epigenetic, or perhaps some mixture of the two. We know that in wild populations, there is a genetic difference between the eyeless cave-dwelling fish and the sighted fish that are found in streams.
   

2. Choose a salt-tolerant annual plant. And grow 100 different plants in graded conditions of salinity, such that plant #1 is watered with no salt at all, and plant #100 gets a lethal salt concentration. Grow a second generation, 1 seed from each of the 100 plants, self-pollinated, and expose them all to a high but sub-lethal salt concentration. Measure the productivity of these 100 plants, and see if their productivity correlates with the salt to which their parent was exposed.
   
   Use sequencing to check that the difference is genetic and not only epigenetic.
   

3. There are known single-gene mutations in C elegans worms that confer tolerance to, respectively, Zn and Cu, both of which are normally toxic to the worms. A million worms can easily be grown in a test tube.
   
   Grow generations of C elegans worms in high, sub-lethal concentrations of Zn. In parallel, grow generations of worms in high, sublethal concentrations of Cu.
   
   After each generation, siphon off half the worms from each population and expose half of each half to what would be lethal concentrations of Zn and to lethal concentrations of Cu. This is a way to assay for the mutations that confer resistance to the respective toxicities.
   
   The Lamarckian prediction is that in the populations grown in Zn have a higher probability of mutating to Zn immunity and the worms grown in Cu will have a higher probability of mutating to Cu immunity. You do these two experiments in parallel to allow for the fact that both the Cu and Zn are mutagens that increase the probability of new adaptations.

I am personally interested in seeing any of these three experiments performed and written up, and I know of a funding source, if you or a colleague are interested in doing the work. Please reach out to me in the comment section below.

Summary

In the 20th century, evolutionary theory was taken over by a mathematical theory known as the New Synthesis or Population Genetics or neo-Darwinism. It is much too narrow a framework for understanding the diversity of phenomena in the biosphere, yet it remains the default model of evolution today. Conservatism aside, the strongest argument against Lamarckian inheritance has been that there was no plausible mechanism by which experience could feed back to change DNA structure. This has changed in recent years.

• Lamarckian epigenetic inheritance has been accepted into the textbook canon.

• James Shapiro has demonstrated that bacteria can “engineer” their genomes.

• Stressed animals are more likely to give rise to offspring that are genetically different from themselves. Edge science tells us that the differences are likely to be adaptive.

• Modern understanding is that gene expression is more important than particular alleles (gene versions) for most structure and functional issues affecting fitness.

• A big portion of the genome in animals and plants is transposable elements (TEs) — stretches of DNA that copy themselves into different places on different chromosomes.

• A series of recent articles has documented that TEs have major effects on gene expression, and there is now every reason to think that the way they distribute themselves is functional, and therefore evolvable.

• All this together provides a plausible mechanism for Lamarckian inheritance that is genetic as well as epigenetic.

Time is ripe for real experimental tests of Lamarckian inheritance in plants and animals. With modern DNA analysis and other techniques, it shouldn’t be hard. The main barrier is that the subject is taboo.