In 2021 there was a press release from the National Institute of Health entitled "NIH supports 106 grants featuring high-risk, high-reward research." The research discussed were projects with a high risk of failure, but which might yield a high reward if they succeed. We read the following: "Supported research this year includes understanding how long-term memory might be encoded in the shape of folded DNA in our neurons, mining data from unconventional sources to reveal social determinants of suicide, establishing new paradigms to address the functional consequences of health disparities in drug development, and looking at the impact of high school and collegiate athlete injuries on long-term health."
The first project mentioned (apparently NIH Project # 1DP2MH129985-01 discussed on this page) is relevant to the question of whether scientists currently have any credible theory of the storage of long-term human memories. For many decades scientists have been telling us that long-term memories are stored in synapses. There has never been any robust evidence or any credible detailed theory backing up such an idea. Everything we know about synapses suggests that they are totally unsuitable for the task of storing memories that can last for 50 years. For example, the proteins in synapses have average lifetimes of only a few weeks or less. Synapses have a high degree of structural dependency on dendritic spines, which are short-lived things that do not last for years. No one has ever proven that a synapse lasts for years, and we have good reason for believing they do not last for years.
What was interesting about this NIH Project # 1DP2MH129985-01 is that it is a kind of "heresy" project that was totally contrary to the "orthodoxy" that our neuroscientists have been spouting for decades about memory. The project had the wildly speculative title "The epigenetic encoding of learning and memory," which is a research project title as speculative as "Extraterrestrial UFO mother-ships near Jupiter." The idea that human memories are encoded in the genome or the epigenome is an idea totally contrary to what neuroscientists have been telling us for decades, that memories are stored in synapses. The genome and the epigenome are found in the center of cells, in chromosomes of a cell nucleus. A synapse is a unit vastly tinier than a cell, outside of a cell or or on the outer edge of a cell. In the visual below depicting a neuron (one of the cells in the brain), the brown circle at the center is the location of the genome and the epigenome, and synapses (too small to show) would be located around the orange parts on the edges:
We see from the project page that the NIH has granted $1,434,188 of public funds for this recent project. The project page presents no detailed research project plan. We merely get a vague project description that left the researchers free to play around pretty much in any way they want. That description often resorts to speculation stated as if it were fact. Here is the description (I'll put in boldface the very speculative parts that are not at all statements of fact):
"The nervous system requires tight control of transcription for processes such as learning and memory formation. The field of epigenetics seeks to understand how changes to gene transcription occur in response to environmental cues and external signals such as those that our brains experience during learning. This proposal lies at the intersection of neuroscience and epigenetics, with a particular focus on chromatin biology. Chromatin is the complex of DNA and the histone proteins that wrap up DNA into complex structures, recruit key transcriptional regulators, and in doing so, control gene expression. In recent years, it has become clear that disruptions to chromatin regulation lead to a range of neurological and mental health disorders such as post- traumatic stress disorder (PTSD). However, we have a limited understanding of how chromatin functions in the brain or how its disruption can lead to disease. We will apply the tools and techniques of the epigenetics field to the study of neuronal function. In doing so, we hope to elucidate the molecular mechanisms that allow our brains to perform incredibly complex tasks and how disruption of these mechanisms can lead to neuronal dysfunction. We propose overcome long-standing hurdles in the field using a combination of novel techniques to reveal how the epigenetic landscape encodes the transcriptional changes that underlie memory formation. Specifically, we seek to uncover the transcriptional signature of memory formation and memory maintenance within single neurons in an in vivo context. We then will examine the epigenetic underpinnings of this transcriptional signature and manipulate specific components of the chromatin environment to define their contribution to learning and memory maintenance. First, in order to elucidate the gene program associated with learning, we will use single-nucleus RNA-sequencing in combination with mouse models that label the specific neurons activated during learning. This will allow us to examine the transcriptional programs activated in neurons that form a memory engram compared to their neighboring cells at various times after learning. Next, we will employ a quantitative biochemical approach uniquely available to our group as part of the Epigenetics Institute to characterize the chromatin landscape changes the occur during memory formation, memory maintenance, and reversal learning. Finally, we will modify the chromatin landscape by manipulating specific histone proteins in combination with numerous sequencing approaches to elucidate how chromatin controls learning and the transcriptional program. Employing this novel combination of techniques will allow us to uncover the mechanisms through which the epigenome encodes information within neurons to modify behavior both in the context of normal learning and in the context of maladaptive responses that lead to disorders such as PTSD. If successful, these methods will 1) identify the transcriptional signature that encodes a memory in neurons, 2) map how this signature is encoded by specific epigenetic regulatory mechanisms, and 3) define how the chromatin landscape affects memory formation and contributes to mental health disorders."
What we have here (in the boldface parts) are statements of an unfounded and wildly speculative theory: the contrarian idea that memories are stored in chromatin (consisting of DNA and proteins surrounding it) and an associated epigenome (consisting of kind of chemical marks next to parts of DNA) . Such statements are made in a matter-of-fact manner, as if such a "yet-to-reach-first-base" theory was fact. The not-yet-popular theory being suggested is one very different from what neuroscientists have been claiming for decades. For decades, neuroscientists have been telling us that memory formation occurs through "synapse strengthening," not through "transcriptional changes." We see no mention of the word "synapses" or "synaptic" in the quotation above.
The boldface above states an idea that makes no sense. "Transcriptional signatures" are transitory fleeting fluctuating biomarkers of the rates at which particular genes are being expressed. Conversely, for a long-term memory to be encoded in a brain there would need to be some all-but-miraculous effect that caused learned information or sensory experience to be permanently stored as brain states or synapse states, rather like letters being written into clay. Referring to "the transcriptional signature that encodes a memory in neurons" is rather like saying the words from your lips are a tape recorder that permanently store what you are saying. But since "transcriptional signatures" bear no resemblance to sensory experience, it's far worse, and would be more like making the double-goofy claim that your heart rate fluctuations are a tape recorder that record all the words you speak.
We should be extremely suspicious and skeptical whenever scientists suddenly start giving some new answer to a fundamental question, an answer completely different from the answer they have been dogmatically declaring for years. For example, if scientists were to suddenly start telling us that galaxies are not held together by gravity (as they've been telling us for decades), but by, say, “dark energy pulsations,” we should be extremely skeptical that the new explanation is correct. In this case, there are very good reasons why the speculations in boldface above cannot be right.
- "Autism spectrum disorder risk genes have convergent effects on transcription and neuronal firing patterns in primary neurons" (link). This is some laboratory study failing to mention memory or learning.
- "Histone variants: expanding the epigenetic potential of neurons one amino acid at a time." The paper is behind a paywall. The abstract says nothing relevant, other than the bare claim that histone variants have a role in memory.
- "Control of striatal circuit development by the chromatin regulator Zswim6" (link). This paper makes no mention of memory or learning.
- "Histone variant H2BE enhances chromatin accessibility in neurons to promote synaptic gene expression and long-term memory" (link). The study provided no good evidence to back up its claim that H2BE has any relation to memory. To test the claim, the study produced "H2BE knockout" mice, whose performance was compared to normal test mice. The first test used was a "novel object recognition" test that is not an effective way of judging animal memory when a blinding protocol is not followed, and the result reported was an unimpressive "p < .05" (widely regarded as a marginal result easy to obtain by chance). In the test scientists attempted to judge how much time a mouse spends exploring a type of object it has already been exposed to, using manual scoring -- we are told, "Time spent interacting with each object was manually analyzed." Such a test is not a reliable way of judging memory in rodents whenever there is a failure to follow a blinding protocol, and no mention is made that a blinding protocol was followed. The unreliability of "novel object recognition" types of tests is discussed in my post here. The second and third tests used (Figures 6I and 6J) were tests requiring a judgment of "freezing behavior," and such tests are utterly unreliable in judging whether an animal recalled, for reasons discussed at length here. These second and third tests as presented in this paper look fishy, as there is no real difference between the "H2BE knockout" mice and the control mice in a "freezing behavior" claimed test of memory taken after an interval of 24 hours or 48 hours (Figures 6I and 6J), with the authors having to retry the test at a 14-day or 15-day interval before they could get something they claimed as a positive result. That has a "keep testing until you can claim the desired result" smell to it. The paper has no robust evidence that the "H2BE knockout" produced a harm in the memory of the tested mice, and its reported results at the 24 hour interval and the 48 hour interval suggest that the knockout did not harm the memory of the mice.
- "A nociceptive amygdala-striatal pathway for chronic pain aversion" (link). No mention is made of learning or memory in this paper.
- "Loss of DOT1L function disrupts neuronal transcription, animal behavior, and leads to a novel neurodevelopmental disorder" (link). The authors first discuss 11 humans who had variations in something called DOT1L The paper claims that 2 of these 11 had "intellectual disability" without giving us any specifics. This does not constitute any evidence of a relation between this DOT1L and memory. The paper also claims that DOT1L modification in zebrafish had some effect on their cognitive performance. But the claim has no clear reference to memory, and the claim is not demonstrated, because the number of zebrafish tested is way too small, being only 3 zebrafish. There is then a claim that DOT1L modifications have some effect on mice. But no clear claim is made of an effect on memory performance.
- "Histone variant H2BE controls activity-dependent gene expression and homeostatic scaling" (link). The research discussed does not involve learning or memory.
- "SARS-CoV-2 disrupts host epigenetic regulation via histone mimicry" (link). The research discussed does not involve learning or memory.
- "Identification of a transcriptional signature found in multiple models of ASD and related disorders" (link). The research discussed does not involve learning or memory. It merely makes a passing reference to memory, claiming (without providing any specifics) that a "histone code" plays a role in memory.
No comments:
Post a Comment