早期生活经历如何写入DNA

So it all came to life in a dark bar in Madrid. I encountered my colleague from McGill, Michael Meaney. And we were drinking a few beers, and like scientists do, he told me about his work. And he told me that he is interested in how mother rats lick their pups after they were born. And I was sitting there and saying, "This is where my tax dollars are wasted -- (Laughter) on this kind of soft science."

And he started telling me that the rats, like humans, lick their pups in very different ways. Some mothers do a lot of that, some mothers do very little, and most are in between. But what's interesting about it is when he follows these pups when they become adults -- like, years in human life, long after their mother died. They are completely different animals. The animals that were licked and groomed heavily, the high-licking and grooming, are not stressed. They have different sexual behavior. They have a different way of living than those that were not treated as intensively by their mothers.

So then I was thinking to myself: Is this magic? How does this work? As geneticists would like you to think, perhaps the mother had the "bad mother" gene that caused her pups to be stressful, and then it was passed from generation to generation; it's all determined by genetics. Or is it possible that something else is going on here?

In rats, we can ask this question and answer it. So what we did is a cross-fostering experiment. You essentially separate the litter, the babies of this rat, at birth, to two kinds of fostering mothers -- not the real mothers, but mothers that will take care of them: high-licking mothers and low-licking mothers. And you can do the opposite with the low-licking pups. And the remarkable answer was, it wasn't important what gene you got from your mother. It was not the biological mother that defined this property of these rats. It is the mother that took care of the pups. So how can this work?

I am an a epigeneticist. I am interested in how genes are marked by a chemical mark during embryogenesis, during the time we're in the womb of our mothers, and decide which gene will be expressed in what tissue. Different genes are expressed in the brain than in the liver and the eye. And we thought: Is it possible that the mother is somehow reprogramming the gene of her offspring through her behavior? And we spent 10 years, and we found that there is a cascade of biochemical events by which the licking and grooming of the mother, the care of the mother, is translated to biochemical signals that go into the nucleus and into the DNA and program it differently. So now the animal can prepare itself for life: Is life going to be harsh? Is there going to be a lot of food? Are there going to be a lot of cats and snakes around, or will I live in an upper-class neighborhood where all I have to do is behave well and proper, and that will gain me social acceptance? And now one can think about how important that process can be for our lives.

We inherit our DNA from our ancestors. The DNA is old. It evolved during evolution. But it doesn't tell us if you are going to be born in Stockholm, where the days are long in the summer and short in the winter, or in Ecuador, where there's an equal number of hours for day and night all year round. And that has such an enormous [effect] on our physiology. So what we suggest is, perhaps what happens early in life, those signals that come through the mother, tell the child what kind of social world you're going to be living in. It will be harsh, and you'd better be anxious and be stressful, or it's going to be an easy world, and you have to be different. Is it going to be a world with a lot of light or little light? Is it going to be a world with a lot of food or little food? If there's no food around, you'd better develop your brain to binge whenever you see a meal, or store every piece of food that you have as fat.

So this is good. Evolution has selected this to allow our fixed, old DNA to function in a dynamic way in new environments. But sometimes things can go wrong; for example, if you're born to a poor family and the signals are, "You better binge, you better eat every piece of food you're going to encounter." But now we humans and our brain have evolved, have changed evolution even faster. Now you can buy McDonald's for one dollar. And therefore, the preparation that we had by our mothers is turning out to be maladaptive. The same preparation that was supposed to protect us from hunger and famine is going to cause obesity, cardiovascular problems and metabolic disease. So this concept that genes could be marked by our experience, and especially the early life experience, can provide us a unifying explanation of both health and disease.

But is true only for rats? The problem is, we cannot test this in humans, because ethically, we cannot administer child adversity in a random way. So if a poor child develops a certain property, we don't know whether this is caused by poverty or whether poor people have bad genes. So geneticists will try to tell you that poor people are poor because their genes make them poor. Epigeneticists will tell you poor people are in a bad environment or an impoverished environment that creates that phenotype, that property.

So we moved to look into our cousins, the monkeys. My colleague, Stephen Suomi, has been rearing monkeys in two different ways: randomly separated the monkey from the mother and reared her with a nurse and surrogate motherhood conditions. So these monkeys didn't have a mother; they had a nurse. And other monkeys were reared with their normal, natural mothers. And when they were old, they were completely different animals. The monkeys that had a mother did not care about alcohol, they were not sexually aggressive. The monkeys that didn't have a mother were aggressive, were stressed and were alcoholics. So we looked at their DNA early after birth, to see: Is it possible that the mother is marking? Is there a signature of the mother in the DNA of the offspring?

These are Day-14 monkeys, and what you see here is the modern way by which we study epigenetics. We can now map those chemical marks, which we call methylation marks, on DNA at a single nucleotide resolution. We can map the entire genome. We can now compare the monkey that had a mother or not. And here's a visual presentation of this. What you see is the genes that got more methylated are red. The genes that got less methylated are green. You can see many genes are changing, because not having a mother is not just one thing -- it affects the whole way; it sends signals about the whole way your world is going to look when you become an adult. And you can see the two groups of monkeys extremely well-separated from each other. How early does this develop? These monkeys already didn't see their mothers, so they had a social experience. Do we sense our social status, even at the moment of birth?

So in this experiment, we took placentas of monkeys that had different social status. What's interesting about social rank is that across all living beings, they will structure themselves by hierarchy. Monkey number one is the boss; monkey number four is the peon. You put four monkeys in a cage, there will always be a boss and always be a peon. And what's interesting is that the monkey number one is much healthier than monkey number four. And if you put them in a cage, monkey number one will not eat as much. Monkey number four will eat [a lot]. And what you see here in this methylation mapping, a dramatic separation at birth of the animals that had a high social status versus the animals that did not have a high status.

So we are born already knowing the social information, and that social information is not bad or good, it just prepares us for life, because we have to program our biology differently if we are in the high or the low social status.

But how can you study this in humans? We can't do experiments, we can't administer adversity to humans. But God does experiments with humans, and it's called natural disasters.

One of the hardest natural disasters in Canadian history happened in my province of Quebec. It's the ice storm of 1998. We lost our entire electrical grid because of an ice storm when the temperatures were, in the dead of winter in Quebec, minus 20 to minus 30. And there were pregnant mothers during that time. And my colleague Suzanne King followed the children of these mothers for 15 years.

And what happened was, that as the stress increased -- and here we had objective measures of stress: How long were you without power? Where did you spend your time? Was it in your mother-in-law's apartment or in some posh country home? So all of these added up to a social stress scale, and you can ask the question: How did the children look? And it appears that as stress increases, the children develop more autism, they develop more metabolic diseases and they develop more autoimmune diseases. We would map the methylation state, and again, you see the green genes becoming red as stress increases, the red genes becoming green as stress increases, an entire rearrangement of the genome in response to stress.

So if we can program genes, if we are not just the slaves of the history of our genes, that they could be programmed, can we deprogram them? Because epigenetic causes can cause diseases like cancer, metabolic disease and mental health diseases.

Let's talk about cocaine addiction. Cocaine addiction is a terrible situation that can lead to death and to loss of human life. We asked the question: Can we reprogram the addicted brain to make that animal not addicted anymore? We used a cocaine addiction model that recapitulates what happens in humans. In humans, you're in high school, some friends suggest you use some cocaine, you take cocaine, nothing happens. Months pass by, something reminds you of what happened the first time, a pusher pushes cocaine, and you become addicted and your life has changed.

In rats, we do the same thing. My colleague, Gal Yadid, he trains the animals to get used to cocaine, then for one month, no cocaine. Then he reminds them of the party when they saw the cocaine the first time by cue, the colors of the cage when they saw cocaine. And they go crazy. They will press the lever to get cocaine until they die. We first determined that the difference between these animals is that during that time when nothing happens, there's no cocaine around, their epigenome is rearranged. Their genes are re-marked in a different way, and when the cue comes, their genome is ready to develop this addictive phenotype.

So we treated these animals with drugs that either increase DNA methylation, which was the epigenetic marker to look at, or decrease epigenetic markings. And we found that if we increased methylation, these animals go even crazier. They become more craving for cocaine. But if we reduce the DNA methylation, the animals are not addicted anymore. We have reprogrammed them. And a fundamental difference between an epigenetic drug and any other drug is that with epigenetic drugs, we essentially remove the signs of experience, and once they're gone, they will not come back unless you have the same experience. The animal now is reprogrammed. So when we visited the animals 30 days, 60 days later, which is in human terms many years of life, they were still not addicted -- by a single epigenetic treatment.

So what did we learn about DNA? DNA is not just a sequence of letters; it's not just a script. DNA is a dynamic movie. Our experiences are being written into this movie, which is interactive. You're, like, watching a movie of your life, with the DNA, with your remote control. You can remove an actor and add an actor. And so you have, in spite of the deterministic nature of genetics, you have control of the way your genes look, and this has a tremendous optimistic message for the ability to now encounter some of the deadly diseases like cancer, mental health, with a new approach, looking at them as maladaptation. And if we can epigenetically intervene, [we can] reverse the movie by removing an actor and setting up a new narrative.

So what I told you today is, our DNA is really combined of two components, two layers of information. One layer of information is old, evolved from millions of years of evolution. It is fixed and very hard to change. The other layer of information is the epigenetic layer, which is open and dynamic and sets up a narrative that is interactive, that allows us to control, to a large extent, our destiny, to help the destiny of our children and to hopefully conquer disease and serious health challenges that have plagued humankind for a long time.

So even though we are determined by our genes, we have a degree of freedom that can set up our life to a life of responsibility.

Thank you.

(Applause)

这一切都是 在马德里的一个黑暗的酒吧里发生的。我遇到了麦吉尔大学的同事，迈克尔·梅尼。我们喝着几杯啤酒， 像所有科学家一样， 他跟我聊他的工作。他告诉我，他感兴趣的是 小鼠出生后，母鼠如何舔它们的婴儿。我呆住了，说：“我交的税就浪费在这儿了啊，—— （笑声） 就是这种软科学。”

他开始给我讲， 老鼠和人类一样， 用不同的方式舔抚她们的孩子。一些母亲做的很多， 一些母亲做的很少， 大多数介于两者之间。但是有趣的是， 当他跟踪这些小鼠直到成年—— 相当于人类的很多年后， 母亲去世后的很长时间。它们变成了完全不同的动物。那些受到大量舔舐和梳毛的动物， 即高舔舐和梳毛的动物， 不会紧张不安。它们的性行为不同。它们的生活方式 与没有被妈妈密切照顾的小鼠不同。

那么我问自己：这是魔法吗？怎么会这样的？遗传学家希望你这样想：也许那些妈妈拥有“坏妈妈”基因， 导致她的婴儿紧张不安， 然后这种基因代代相传；全部是遗传决定的。有没有别的可能？

在老鼠身上，我们可以 提出这个问题并解答它。所以我们做了交叉养育实验。基本上在一窝小鼠出生时 把它们分开， 分给两种不同类型的养母—— 不是生母，而是照顾小鼠的母鼠：高舔舐母鼠和低舔舐母鼠。另外一边是低舔舐小鼠。令人惊奇的答案是， 从母亲那里得到什么基因并不重要。定义这些老鼠的这种特征的不是生母， 而是照顾小鼠的母鼠。那么，这是怎么实现的？

我是一个表观遗传学家。我的兴趣是在胚胎期间，  我们在母体子宫里时， 如何用化学记号标记基因， 并决定 哪个基因在哪个组织中表达。大脑中表达的基因与肝脏和眼睛中的不同。我们怀疑：有没有可能 妈妈通过自己的行为 改写她后代的基因编码？我们花了10年， 发现有一系列的生物化学事件， 母鼠的舔舐和梳毛，妈妈的照顾 被转化成生物化学信号， 这信号进入细胞核并进入DNA， 改写了编码。所以现在这个动物能 为生活做更好的准备：生活会很艰难吗？会有很多食物吗？周围会有很多猫和蛇吗？还是会住在上层社区， 只要行为端正举止适当 就会赢得社会认可？现在可以想想这个过程 对我们的生命有多重要了。

我们从祖先继承了DNA。DNA历史悠久。它在进化过程中演变。但它并不告诉我们， 你会出生在夏季白天长、 冬季白天短的斯德哥尔摩， 还是出生在 整年都白天黑夜一样长的厄瓜多尔。这对我们的生理有巨大的影响。所以我们认为， 也许是在生命早期发生的事， 通过母亲传递的那些信号， 告诉子女将要在怎样的社会环境中生存。如果环境艰难，你最好焦虑紧张， 或者如果环境轻松，你必定也不同。那世界会有很多光线还是很少光线？那世界会有很多食物还是很少食物？如果周围没有食物， 你最好让大脑发育成 一旦看到食物马上大吃一顿， 或者把拥有的每一块食物都 储存成脂肪。

所以，这很好。进化已经这样选择， 让我们固有的、陈旧的DNA  在新的环境中不断变化地发挥作用。但有些事可能会出错：例如，如果你出生在一个贫穷的家庭， 信号是，“你最好快吃， 你最好把遇到的每一块食物都吃掉。” 但是现在我们人类和人脑已经进化了， 已经进化得更快。现在你用一块钱就可以买麦当劳。因此，母亲为我们所做的准备 变得不适用了。本应保护我们免受饥苦的 同样的准备工作 却会导致肥胖症、 心血管问题和代谢疾病。所以，这个可以通过经历、 尤其是早期生活经历 来标记基因的概念 可以为我们提供 对健康和疾病的统一解释。

但这只在老鼠身上才正确吗？问题是，我们不能用人类做测试， 因为从道德上，我们不能 随机地为儿童设置逆境。所以，如果穷孩子养成某种特征， 我们不知道它是由贫穷引起的 还是因为穷人有不良的基因。所以，遗传学家试图告诉你 穷人之所以穷是因为基因让他们穷。表观遗传学家会告诉你 穷人所在的不良环境或贫穷环境 环境创造了那些表型，那种贫穷。

所以我们转而观察我们的表亲：猴子。我的同事斯蒂芬·苏米 以两种不同的方式饲养猴子：随机地将猴子与母亲分离， 通过护理员和代养条件 来培养她。所以这些猴子没有妈妈， 只有护理员。而其他猴子是正常的亲生妈妈养育的。当他们年老时，他们变成了完全不同的动物。有妈妈的猴子不嗜酒， 他们没有性暴力倾向。没有妈妈的猴子有攻击性，紧张， 还是酒鬼。我们在出生后的早期观察他们的DNA：看看有没有可能是妈妈在标记？后代的DNA里有没有妈妈的签名？

这些是14天的猴子， 这里看到的是我们研究 表观遗传学的现代方式。我们能把这些化学标记—— 称为甲基化标记， 以单核苷酸分辨率绘制到DNA上。我们能够绘制整个基因组。现在我们可以比较 有妈妈或没有妈妈的猴子了。这里是直观展示。你看到的是 较多甲基化的基因呈现红色。较少甲基化的基因呈现绿色。你能看到很多基因在改变， 因为没有妈妈不只是简单一件事—— 它影响所有的一切；它传递着关于你 成年后的世界 是什么样子的全部信号。你可以看到两组猴子 它们极其明显的大不相同。这种发育从多早开始？这些猴子已经没见过妈妈， 所以它们有社交经历。我们在出生的时刻也能 感受自己的社会地位吗？

所以在这个实验中， 我们采用具有不同 社会地位的猴子的胎盘。关于社会秩序，有趣的是 在所有的生物中， 他们都用等级制度构建社会。一号猴子是老板；四号猴子是苦力。在一个笼子里放四只猴子， 总是有一个老板，一个苦力。并且，有趣的是，一号猴子 比四号猴子健康得多。如果把它们放在一个笼子里， 一号猴子吃得不多。四号猴子能吃（很多）。在这个甲基绘制图中可以看到， 社会地位高的动物与社会地位低的动物 在刚出生时的明显不同 的对比图。

所以我们出生时已经知道社会信息， 该社会信息并无好坏之分， 它只是帮我们为生活做准备， 因为我们要根据所处的 高社会地位或低社会地位， 用不同方式给自己进行生物编码。

但如何在人类中研究这种现象？我们不能做实验， 不能给人类制造逆境。但上帝在用人类做实验， 它叫做自然灾害。

加拿大历史上最严重的自然灾害之一 发生在我来自的魁北克省。是1998年的冰暴。因为冰暴，我们整个电网坏掉， 当时是魁北克最冷的冬天， 温度在零下20至零下30。在那段时间有怀孕的妈妈们。我同事苏珊娜·金对这些妈妈们的孩子 跟踪调查了15年。

事实是，随着压力增大—— 我们这里有对压力的客观度量：你停电多长时间？在哪里度日？在婆婆的小公寓还是 奢侈的乡间别墅？所有这些加起来得出社会压力评级， 你可以问的是：孩子们看上去如何？显然，随着压力增大， 孩子们的自闭症更多， 他们的代谢疾病更多， 自身免疫性疾病也更多。我们绘制甲基化状态， 再一次看到，随着压力增大 绿色基因变成红色， 随着压力增大，红色基因变成绿色， 基因组响应于压力进行完全重排。

所以，如果我们能编写基因， 如果我们不仅仅做我们基因历史的奴隶， 如果基因可以编写， 那我们能抹掉对基因的编写吗？因为表观遗传因素会导致癌症、 代谢疾病 和精神健康疾病等。

我们来谈谈可卡因成瘾。可卡因成瘾是一种可怕的状态 会导致死亡和丧命。我们问了一个问题：我们可以重新编写上瘾的大脑， 让上瘾的动物不再有瘾吗？我们使用可卡因成瘾模型， 该模型概括了人类成瘾的情况。对于人类，你上高中时， 某些朋友建议你用点可卡因， 你用了可卡因，没发生什么事。几个月后，突然想起第一次的事， 毒贩子在卖可卡因， 然后你上瘾了，然后生活彻底改变。

我们用老鼠做同样的事。我同事加尔·雅迪 他训练这些动物习惯可卡因， 然后一个月没有可卡因。然后他提醒动物们 第一次看到可卡因的场景， 提示方法是 看到可卡因时笼子的颜色。动物们疯了。它们一直按动会提供可卡因的杠杆， 直到死亡。我们首先确定的是， 这些动物之间的区别 是在什么都没发生的那段时间里 周围没有可卡因时， 它们的表观基因组重新排列。它们的基因以不同的方式重新标记， 当提示出现时， 它们的基因组已经准备好 发展出成瘾表型。

然后我们用药治疗这些动物， 药物要么增加DNA甲基化—— 对象是表观基因标记物， 要么减少表观基因标记。我们发现，如果增加甲基化， 这些动物会更疯狂。它们对可卡因更渴望。但如果减少DNA甲基化， 动物不再有瘾。我们把这些动物重新编码了。表观基因药物与其他药物 的基本区别是， 通过表观基因药物， 我们在根本上去除了经历的印记， 一旦去除印记， 它不会回来，除非再经历一遍。现在动物被重新编码。所以在30天、60天后—— 等于人类的很多年， 再造访这些动物时， 它们仍然没有瘾—— 只通过单一的表观基因治疗。

那么关于DNA，我们学到了什么？DNA不仅仅是一系列的字母; 它不只是一个脚本。DNA是一部动态电影。我们的经历正在写入这部电影中， 它是互动的。就像用遥控器看电影一样， 你在用DNA来观看你的人生。你可以去掉一个演员、增加一个演员。因此，尽管遗传学具有确定性， 你仍可以控制基因的表达方式， 这为我们面对某些致命疾病的能力 例如癌症、心理疾病等 提供了非常乐观的信息， 提供了新的方法， 可以把这些疾病看作适应不良。如果我们能进行表观基因干预， 我们就可以去掉一个演员，让电影倒退 并设置新的故事线。

所以我今天告诉你的是， 我们的DNA实际由两个部分组成， 两层信息。一层信息是古老的， 从数百万年的进化演变而来。它是固定的，很难改变。另一层信息是表观遗传层， 它是开放和动态的， 并设置了一个可以互动的故事线， 让我们能够在很大程度上 控制自己的命运， 帮助改变我们下一代的命运， 并且有希望征服那些 长期困扰人类的 疾病和严峻的健康挑战。

所以即使我们 已经被我们的基因决定， 我们还是有一定程度的自由， 能够把自己的生命设置成 有责任担当的生命。

谢谢。（掌声）

来源：TED演讲

长按识别二维码可关注该微信公众平台

往期回顾