For now the power of genes to affect behavior is more evident in other species, whose nature we can actually manipulate using genetic tricks. One of the oldest, and most effective, of those tricks is artificial selection, which farmers have long used to increase milk yield in cows or wool quality in sheep. But its applications have not been limited to agriculturally valuable traits like these. Dogs are derived from wolves – possibly from wolf individuals that tended to hang around human settlements looking for a scrap and thereby conveniently assisting in garbage disposal. It is thought that they first laid claim to the title of "man's best friend" 10,000 years ago, roughly coinciding with the origin of agriculture. In the brief time since, the anatomical and behavioral diversity engendered by dog breeders has become literally a thing to behold. Dog shows are celebrations of the power of genes, with each breed effectively a genetic isolate – a freeze frame in the spectacular feature film of canine genetic diversity. Of course, the morphological differences are the most striking and fun to consider: the fluff ball called Pekingese; the enormous shaggy English mastiff, which sometimes weighs over 300 pounds; the stretched-out dachshund; the face-flattened bulldog. But it is the behavioral differences that I find most impressive.
Of course, not all dogs within a breed behave alike (or look alike), but typically individuals in the same breed have much more in common with one another than with specimens from other breeds. The Labrador retriever is affectionate and pliant; the greyhound is twitchy; the border collie will round up anything available if a flock of sheep is not to be had; the pit bull, as news reports occasionally remind us, is the canine embodiment of aggression. Some dog behaviors are so engrained as to have become stereotypes. Think of a pointer's elaborate "pointing" stance: that is no "stupid pet trick" taught each individual dog, but a hardwired part of the breed's genetic makeup. Despite their diversity, all modern dogs remain members of one species – meaning that in principle even the most apparently dissimilar pair can produce offspring, as one doughty male dachshund demonstrated in 1972 when he managed to inseminate a sleeping Great Dane bitch. Thirteen "Great Dachshunds" resulted.
While the basis of most behavior is surely polygenic – affected by many genes – a number of simple genetic manipulations in mice reveal that changing even a single gene can have a major behavioral effect. In 1999, Princeton neurologist Joe Tsien used sophisticated recombinant DNA techniques to create a "smart mouse" with extra copies of a gene producing a protein that acts as a receptor for chemical signals in the nervous system. Tsien's transgenic rodent performed better than normal mice in a battery of tests of learning and memory; for example, it was better at figuring out mazes and at retaining that knowledge. Tsien named the mouse strain "Doogie" after the precocious medic in the television show Doogie Howser, M.D. In 2002 Catherine Dulac at Harvard discovered that by deleting a gene from a mouse of her own she could affect the processing of chemical information contained in pheromones, the odors mice use to communicate. Whereas male mice will typically attack other males and attempt to mate with females, Dulac's doctored males failed to distinguish between males and females and attempted to mate with any mouse they encountered. Nurturing behavior in mice is also subject to a sex-specific manipulation of a single gene. Females instinctively look after their newborns, but Jennifer Brown and Mike Greenberg at Harvard Medical School found a way to short-circuit this native sense by knocking out the function of a gene called fos-B. Otherwise entirely normal, a mouse so altered simply ignores her offspring (see Plates 59 & 60).
Rodents have even given us a glimpse of the mechanistic basis of what in humans we call "love" (in rodents, it's less romantically termed "pair-bonding"). Mouselike voles are common throughout North America. Although they all look pretty much alike, different species have wildly different approaches to life. The prairie vole is monogamous, meaning that couples bond for life, but its close relative, the montane vole, is promiscuous: a male mates and moves on, and over the course of her life the female typically produces litters by several different males. What differences could underlie such widely divergent sexual strategies? Hormones provide the first half of the answer. In all mammals, oxytocin figures in many aspects of mothering: it stimulates both the contractions in labor and the production of breast milk for the newborn; thus it also plays a part in creating the nurturing bond between a mother and her young. Could the same hormone also engender another kind of bond, one between members of a prairie vole couple? In fact it does, together with another garden-variety mammalian hormone, vasopressin, which is primarily known for controlling urine production. But why is the montane vole, which also produces both hormones, such a randy little critter compared with its prairie cousin? The key, it turns out, is in the hormone receptors – the molecules that bind to circulating hormones, initiating a cell's response to the hormone signal.
Focusing on the vasopressin receptor, Tom Insel, a psychiatrist at Emory University, found a major difference between the vole species, not in the receptor gene itself but in an adjacent DNA region that determines when and where the gene is turned on. As a result the distribution of vasopressin receptors in the brain is very different in the prairie vole and the montane vole. But does this difference in gene regulation alone explain why one species is, in human terms, loving and the other cavalier? Apparently it does. Insel and his colleague Larry Young inserted the prairie vole vasopressin gene, complete with its next-door regulatory region, into a regular lab mouse (a species that is typically promiscuous, like the montane vole). Though the transgenic mice did not instantly become romantic pair-bonders, Insel and Young observed a marked change in their behavior. Rather than doing the typical mouse thing of mating with a female and then hightailing it unceremoniously, the transgenic male appeared in contrast tenderly solicitous of the female – in short, the addition of the gene, while not a guarantee for everlasting love, did seem to make the affected mouse less of a rat.
We must not forget that human brain function remains a million miles removed from that of a mouse. No rodent, whether from the mountains or the prairies, has yet produced a major work of art. It nevertheless remains worthwhile to bear in mind that most sobering lesson of the Human Genome Project: our genome and the mouse's are strikingly similar. The basic genetic software governing both mice and men has not changed much over the 75 million years of evolution since our lineages separated.
Unable to target specific genes for inactivation or enhancement as mouse geneticists can, human geneticists must rely on what might be called "natural experiments" – spontaneous genetic changes that affect brain function. Many of the best characterized genetic disorders affect mental performance. Down syndrome, caused by having an extra copy of chromosome 21, results in lowered IQ and, in many cases, a disarmingly sunny disposition. People with Williams syndrome, caused by the loss of a small portion of chromosome 7, also have low IQs but are often preternaturally talented as musicians.
But these are instances in which the mental aspects of a given disorder are effectively by-products of systemwide dysfunction. Thus they tell us relatively little about the particular genetics of behavior. It's a bit like discovering that your computer doesn't work during a power outage. OK, so you know now that it requires electricity, but you've learned very little about the specifics of computer function. To understand the genetics of behavior we need to look at disorders that affect the mind directly.
Among the mental scourges that have attracted close attention from gene mappers, two of the most formidable are bipolar (or manic-depressive) disease (BPD) and schizophrenia. Both diseases have strong genetic components (identical-twin concordance for BPD is as high as 80 percent; for schizophrenia it is close to 50 percent), and both take a devastating toll on mental health worldwide. One in every hundred people is schizophrenic, and the figure is about the same for BPD.
As we have seen, mapping polygenic traits is difficult because each individual gene has only a small incremental effect, and the trait as a whole is often mediated strongl
y by the environment, as with both these afflictions. But the overall difficulty has also bred a bad habit among researchers: they tend to publish only positive results, failing to notify the field of eliminated possibilities. Compounding the problem is the converse: the understandable but ultimately counterproductive impulse to publish any correlation that appears, after countless other genetic markers have proved a dead end. Ideally, spotting a correlation should only be the beginning of more in-depth analysis to sort out meaningful results from statistical coincidence – after all, if we try enough markers, we should expect occasionally to see, even in the absence of genetic linkage, a correlation produced by chance alone. Too often the pressure to get results leads to premature pronouncements, which must later be withdrawn sheepishly after other groups fail to replicate the finding.
There are additional impediments to the gene hunt when mental illness is the target. Diagnosis, however psychiatric manuals may try to standardize it, is often more an art than a science. Cases may be identified on the basis of ambiguous symptoms, and so a portion of the individuals in a pedigree may be misdiagnosed; these false positives wreak havoc in the mapping analysis. Another complicating factor is that the disorders are defined and diagnosed according to their symptoms, and yet it is likely that a number of genetic causes result in a similar set of symptoms. Thus the genes underlying schizophrenia may differ from one case to the next. Even apparently clear differences between syndromes can prove messy when viewed through the microscope of genetics. Since 1957 we have known that BPD and unipolar disease (the condition characterized by depression alone) are genetically distinct syndromes, but, confusingly, there is some genetic overlap between the two: unipolar depression is much more common among relatives of BPD patients than in the overall population.
Partly for such reasons, the genetic culprits of mental illness have so far proved particularly elusive. A recent study reveals that as many as twelve chromosomes – half the total – have been shown through mapping analysis to contain genes contributing to schizophrenia. It's the same story for BPD, in which genes on ten chromosomes have been implicated. One interesting finding is that there does seem to be some overlap between the gene regions identified in the separate mapping studies of the two maladies. Perhaps, then, there are some genes responsible for the overall organization and structure of our brains. Malfunctions in these genes maybe the cause of the delusional or hallucinatory episodes common to both BPD and schizophrenia. The history of this research is full of high hopes brought low. A study will identify a strong correlation in one pedigree, and then subsequent research will fail to generalize the result in other populations. Such was the case in 1987, with a much-publicized study of BPD in the Amish: a promising connection with chromosome 11 did not fare well in follow-up studies. Stanford geneticists Neil Risch and David Botstein have aptly articulated these disappointments:
In no field has the difficulty [of mapping disease genes] been more frustrating than in the field of psychiatric genetics. Manic depression (bipolar illness) provides a typical case in point. Indeed, one might argue that the recent history of genetic linkage studies for this disease is rivaled only by the course of the illness itself. The euphoria of linkage findings being replaced by the dysphoria of non-replication [in other populations] has become a regular pattern, creating a roller coaster-type existence for many psychiatric genetics practitioners as well as their interested observers.
Without denying these difficulties I am extremely hopeful that we are right now entering an era of genetic analysis that will soon take us beyond this irritating game of "now we have it, now we don't." Two innovations hold the key. First, the "candidate gene" approach to finding the genes. With both the complete human genome sequence and a rudimentary functional understanding of many genes finally in hand, we can narrow our search as never before, homing in on genes with functions related to a given disorder. In the case of BPD, for example, a condition apparently connected with a fault in the mechanism by which the brain regulates its concentration of certain chemical neurotransmitters like serotonin and dopamine, we might choose to concentrate on genes that produce neurotransmitters or their receptors. Having chosen our candidate gene, we simply compare its sequence in affected and unaffected individuals to determine whether or not a particular variant might correlate with the disorder. In 2002, Eric Lander's team at MIT's Whitehead Institute surveyed seventy-six BPD candidate genes. Only one – a gene encoding the brain-specific nerve growth factor, the neurochemical tested as a possible treatment for Lou Gehrig's disease (see chapter 5) – proved to correlate with the disorder. But one truly relevant gene can be extremely valuable. The one found resides on chromosome 11, apparently vindicating the original Amish study, which long ago implicated the same region of the chromosome in BPD.
Technological improvements underlie the other reason for my optimism about the hunt for these elusive genes. To detect the subtle effect of a particular gene, we need extra-sensitive statistical analyses, which themselves require very large data-sets. Only with the advent of high-throughput sequencing and genetic-typing technologies have we had the capacity to collect appropriate data for huge numbers of markers from huge numbers of people. Not surprisingly, such industrial-scale genetic analysis is beyond the reach of most academic labs, so we will see biotech companies bankrolled by the pharmaceutical industry come to play an increasingly prominent role in this area. In 2002, two such companies, Genset in France and deCODE in Iceland, identified separate genes implicated in schizophrenia. These discoveries are a major step forward: because we have now fingered actual genes – as opposed to merely mapping an effect to a region of a chromosome – we can study gene function to learn about the biochemical basis of the disorder. Strikingly, both genes are involved in regulating the function of a particular neurotransmitter, glutamate.
With these new approaches – candidate genes and super-powerful genetic mapping – I am confident that we will soon uncover the major genes contributing to BPD and schizophrenia. Hopefully that will lead to improved treatments, as well as to a better understanding of how genes govern the workings of our brain.
For traits about whose neurochemical basis we have no clue, however, the roller-coaster ride of euphoric expectations dysphorically dashed is likely to continue. This has often been the case in studies of nonpathological behavior. Dean Hamer's 1993 analysis of the genetics of male homosexuality provides a case in point. It caused quite a stir when he found a particular region on the X chromosome that seemed to correlate with being gay. If being gay were proven to be as much a function of genes as is, say, skin color, then perhaps antidiscrimination legislation applicable to skin color was equally applicable to gays. Hamer's finding, however, has not withstood the test of time. Nevertheless, I suspect that as we develop more statistically powerful means of analysis (and learn to recognize and discount weaker correlations), we will indeed eventually identify some genetic factors that predispose us to our respective sexual orientations. But this should not be taken as purely determinist conjecture; environment is never to be discounted and a predisposition does not a predetermination make.
My pasty complexion may predispose me to skin cancer but, absent the effects of ultraviolet input from the environment, my genes are merely a matter of potential.
Hamer's other high-profile discovery looks more robust. He looked into the genetics underlying the urge for novelty, one of five key "personality dimensions" identified by psychologists. Do you cower in a corner when your routine gets disrupted? Or do you go out of your way to avoid a rut, subjecting yourself to an ever-changing kaleidoscope of new adventures? These, of course, are the extremes. Hamer's evidence pointed to a slight but significant effect of variation in a gene underlying a receptor for the brain signal molecule dopamine. Some attempts to replicate this result have failed, but others have extended it, finding the same gene implicated in particular types of novelty seeking, including drug abuse.
Violence, too, can be viewed through the lens of g
enetics. Some people are more violent than others. That's a fact. And violent behavior may be governed by a single gene interacting with environmental factors. This does not, of course, mean that we all carry a "violence gene" (though it's likely that most violent individuals do possess a Y chromosome), but we have identified at least one simple genetic change that can lead to violent outbursts. In 1978, Dr. Hans Brunner, a clinical geneticist at University Hospital in Nijmegen, Holland, learned of a family whose men tended to border on mental retardation and were prone to aggressive episodes. Thirty years earlier, in an attempt to document this "curse," a relative had compiled an extensive dossier on the family's woes. Brunner brought the survey up-to-date. He found eight men in the clan who, despite coming from different nuclear families, evinced similar patterns of violence. One had raped his sister and subsequently stabbed a prison guard; another used his car to hit his boss after being mildly reprimanded for laziness; two others were arsonists.
Dna: The Secret of Life Page 44