From time to time a report appears in the scientific literature that forces us to pay renewed attention to a trend in thinking that had been developing all along, but which, until the report was published, we could avoid thinking about. Such a paper appeared in the journal Science in July 1993. Summarizing several years of work by the laboratory of Dr. Dean Hamer, at the National Cancer Institute, the report was entitled “A Linkage between DNA Markers on the X Chromosome and Male Sexual Orientation.” In it, Hamer and his colleagues presented the first serious evidence at the molecular level that male homosexuality has heritable components, and that at least one of the genes underlying this trait is transmitted exclusively through the maternal parent. The response, both from the public and the scientific community, was almost instantaneous. The report was variously praised, rejected, feared, welcomed, or damned, depending on the audience. But it was not ignored.

The notion that sexual orientation might have a significant genetic (p. 240 ) component did not originate with the Hamer paper. The tendency for homosexuality to run in families had been noted for over sixty years. Many of the early studies were small and somewhat anecdotal, but more recent, larger studies have confirmed this general finding. A study published in 1986, for example, found that 22 percent of the male siblings of self-described gay men also described themselves as gay. That is well beyond what we would expect on a chance basis, since not more than 5 percent of the male population in the United States is estimated to be gay. Indeed, in the same study only 4 percent of the male siblings of an equivalent group of heterosexual males described themselves as gay. Overall, the frequency of gay brothers is three to six times higher for gay than for heterosexual males. It had even been noted in a few of these early studies that the trend toward male homosexuality appeared to run through maternal rather than paternal lineages. Homosexual behavior in females also tends to cluster in families, although there have been fewer studies and the data are less strong.

It had also been recognized that male homosexuals are more likely to have a gay brother rather than a lesbian sister, whereas lesbians are more likely to have a lesbian sister than a gay brother. In other words, male and female homosexuality tends to run in separate families. That was one piece of evidence suggesting that, to the extent homosexuality would be affected by genes, the genetic programs might not be entirely the same in men and women. Gay men and women also differ behaviorally. Males tend to be exclusively heterosexual or homosexual, and are only rarely bisexual. Women show a much broader spectrum of preferences, with a much higher percentage of nonheterosexual women showing varying degrees of bisexuality.

For those who did not want to believe in a genetic basis for homosexuality, the family studies could be interpreted on the basis of a common environment that might predispose to homosexual behavior, particularly in the sense that one homosexual son or daughter could tempt other siblings into the same sort of behavior. The twin studies were harder to dismiss. In seven studies of monozygotic and dizygotic twins, it was found that in 244 monozygotic twin pairs where one twin was homosexual, the other was homosexual as well 58 percent of the time, on average (Fig. 13.1). In 175 dizygotic twin pairs, both were gay 18 percent of the time, which is very close to the figure for nontwin siblings. Where it has been looked at in large enough numbers of subjects, (p. 241 ) these concordances were true of female twins as well as males, and of twins reared together or apart. One study included three sets of triplets in addition to monozygotic and dizygotic twin pairs. One triplet set contained a pair of monozygotic male twins, both of whom were gay, and a sister who was not. A second mixed set had two monozygotic female twins, both of whom were gay, and a heterosexual sister. The third set were all monozygotic males, and all three were gay. In a study of 57 male infants adopted into families in which at least one male child turned out to be gay, the adopted male child was also gay in 11 percent of cases.

The high incidence of both monozygotic twins being gay, particularly in those cases where they were reared apart, raised a serious challenge to either the family environment or the environment in general being the sole factor in determining a homosexual lifestyle. Analysis of the available twin data suggest the genetic component of homosexual behavior is about 50 percent. On the other hand, the finding of 11 percent homosexuality developing in adopted siblings, which is considerably higher than the 5 percent or so estimated for the population at large, raises some questions. This effect could be only apparent, due to the relatively small number of individuals involved, but it could also reflect an influence, direct or indirect, of siblings on one another in the determination of becoming gay. We could imagine, for example, that (p. 242 ) an individual with only a modest genetic predisposition for becoming gay could find that propensity strengthened in a home where a close sibling followed a clearly gay developmental pathway.

Figure 13.1 Sexual preference correlations in twin pairs (MZ, DZ), between adopted children and adoptive siblings (A), and in the general population (G).

While the family and twin data suggest that homosexuality is a heritable trait, none of the data suggest the existence of anything we could call a definitive “ gay gene” which, if inherited, would cause an individual to be homosexual. Indeed, the family and twin studies make it clear that homosexuality, like any other behavioral trait, is not Mendelian (single-gene) but quantitative, almost certainly involving multiple genes. Like any other behavioral trait, there is also a substantial contribution of nongenetic factors, such as environment. On the other hand, the fact that environment rarely accounts for more than 50 percent of homosexuality makes it clear that environment is unlikely to be the sole determining factor. Nevertheless, even with the twin and adoption studies, those opposed on political, social, or moral grounds to the idea that homosexuality has a substantial genetic component could still find room to quibble.

Hamer’s study, if confirmed, makes such quibbling more difficult. Hamer and his research group focused on two questions. First, they wanted to determine whether the apparent tendency for male homosexuality to pass through maternal lineages was true. Males have only one X chromosome (they are XY with respect to the sex chromosomes), and their single X chromosome is always and only inherited from the mother. In males, altered allelic forms of genes on the X chromosome may cause the appearance of a trait that is not seen in females inheriting exactly the same chromosome. Hemophilia and certain immune deficiency diseases are examples of X-linked traits that show up almost exclusively in males. The reason is that females have two X chromosomes (e.g., they are XX), and unless both chromosomes have the same altered allele, the altered trait generally will not be expressed; it is overridden by the “normal” allele on the second X chromosome. So if a gene predisposing toward male homosexuality were indeed inherited exclusively through the mother it would have to be associated with the X chromosome, which would be a tremendous head start in ultimately isolating and identifying such a gene.

To address this question, Hamer carried out a detailed “pedigree” study of the families of two sets of homosexual male volunteers. Family lineage studies used in genetics normally focus on simple parent-to-offspring (p. 243 ) inheritance patterns; Hamer decided to explore some of the branches of the trees, as well as their main trunks. One set of seventy-six volunteers was randomly selected, in the sense that the investigators knew nothing of the incidence of homosexuality among family members. The second set was recruited by specifically advertising for brothers who were both homosexual; thirty-eight pairs were selected for the study. Through extensive questioning of the volunteers and of various family members, sixty-nine homosexual relatives were identified from the two sets, and interviews with the individuals identified by family members confirmed their sexual orientation.

The pedigree analysis of the seventy-six randomly selected volunteers showed that, as expected, the highest rate of homosexual orientation—13.5 percent—was in brothers (Table 13.1). Among other male relatives, only two groups had a higher level of homosexual orientation than would be expected by chance: maternal uncles (7.3% gay), and the sons of maternal aunts (7.7%). Among the paired homosexual brothers in the second set, again the only male relatives with a (p. 244 ) significant level of homosexuality were maternal uncles (10.3%) and the sons of maternal aunts (12.9%). As controls, Hamer looked at the frequency of gay men among the relatives of lesbian subjects used for a separate study, since all evidence suggests that homosexuality in gay men and lesbian women is genetically distinct. These results confirmed in a convincing way the transmission of a predisposition toward male homosexual behavior through maternally inherited genes.

The higher family incidence of homosexual orientation in the families containing two gay brothers suggested such families would be good candidates for tracing the genes involved. The clearly demonstrated maternal inheritance of these genes strongly suggested that they would be found on the X chromosome. Thus DNA samples from the homosexual brothers in forty different families, together with DNA from their mothers, where available, were subjected to a “chromosome scan”; they were typed for twenty-two DNA markers spanning the entire X chromosome. The mother in each case will have two X chromosomes, and the DNA markers (selected in the first place because they are highly allelic in the general population) will be in allelic form at most of the twenty-two marker loci. If there is a gene predisposing to homosexuality on the X chromosome, then gay brothers in the same family should have inherited the same X chromosome from their mother, and therefore the same allelic DNA markers located near that gene. The results showed that markers defining a region of the X chromosome called Xq28, located at one tip of the chromosome (Fig. 13.2), were inherited by 33 (82%) of the forty gay pairs tested. The likelihood that this could happen by chance is small.

In a follow-up study over the next two years, Hamer’s group checked their initial results with the Xq28 DNA markers on a second group of thirty-three families having two nontwin gay brothers, and extended their research to include thirty-six families that had two nontwin lesbian daughters. In these studies, they used additional markers more closely spaced around the Xq28 region, in order to define more precisely the chromosomal region influencing male homosexuality. They found that a preponderance of gay brothers (albeit a lower 67%) shared the same Xq28-associated markers; these markers defined a narrower QTL in Xq28 called DXS52 (for the nearest marker). This time they also checked the DNA of heterosexual brothers (where they existed) of the gay subject brothers; only 22 (p. 245 ) percent of the nongay brothers had the same DXS52 markers as their gay brothers. When the same analytical methods were applied to lesbian sisters and their heterosexual sisters, no preferential distribution of X chromosome markers of any kind was found. A lesbian was just as likely to share the same DXS52 markers with her heterosexual sister (56% of the time) as with her lesbian sister (58% of the time.) Even females with the appropriate DXS52 markers on both X chromosomes were not necessarily lesbian. Thus it would appear that the X-chromosome gene predisposing to homosexuality in males, while certainly present in females, has no influence on the development of female homosexuality.

Figure 13.2 The human X chromosome, showing region identified by Hamer’s putative DXS52-linked gene. Note also the position of the MAOA gene referred to in Chapter 9.

Given the polygenic nature of homosexuality, made clear by numerous inheritance studies, we assume the DXS52 marker defines only one of a number of QTL scattered throughout the genome that influence homosexuality. While it is clear that the DXS52 QTL is involved in predisposing some males toward a male sexual preference, it is equally clear that the gene or genes at this QTL are neither necessary nor sufficient for the development of male homosexuality. And it apparently has no effect at all on the development of female homosexuality, (p. 246 ) which is consistent with previous thinking that male and female homosexuality may be genetically different in at least some respects. In males, the fact that a number of gay brother pairs did not both show preferential inheritance of the DXS52 markers is a clear indication that other genes must be involved. And the 22 percent of the heterosexual brothers in the follow-up study who had the predisposing DXS52 marker but were not gay also indicates that the DXS52-associated gene must require other genetically or environmentally controlled factors to trigger homosexual behavior.

The existence of a gene in the XQ28 region affecting homosexuality has recently been challenged by a Canadian study that failed to find preferential inheritance of markers in this region in male homosexuals. This study did not challenge the notion that homosexuality has a heritable component or that it is linked to the X chromosome, but the researchers disagree about the previous locus. Studies are now looking at how subjects were selected for these and other studies. Hamer purposely selected families where maternal inheritance seemed to be involved. He willingly admits that such individuals may not be representative of all male homosexuals, and that there will likely be many other genes involved in the heritability of this trait. Hamer is now searching the DNA in the region of DXS52 for candidate genes. It is certainly possible that at least one gene, with a significant influence on at least some forms of male homosexuality, may eventually be found at this locus.

Although it will likely be several years before the presence or absence of a DXS52-associated gene will be confirmed or disproved, there is already a good deal of informal speculation about what such a gene—or any other gene associated with homosexuality—might govern. As we saw earlier, the default gender in mammals, including humans, is female. The presence of a gene called sry, located on the Y chromosome, is absolutely required for the development of maleness. This gene shuts down female development, and induces the development of male sexual characteristics; in the absence of sry, an individual will always be female. Clearly neither primary nor secondary sexual characteristics are affected by Hamer’s gene; gay males are indistinguishable in these respects from heterosexual males. However, most of the Y chromosome genes evolved from genes on the X chromosome; it could be that an X chromosome gene evolutionarily related to sry (p. 247 ) affects certain behavioral aspects of female reproductive function, and under certain circumstances can influence male mate selection as well.

It is unlikely that the DXS52-linked gene affects male reproductive hormones such as testosterone; there is no discernible difference in either the timing of appearance or in the levels of these hormones between gay and heterosexual men.^* The possibility of different allelic variants in the receptor for testosterone predisposing to homosexual behavior has been examined, but no correlation of a particular testosterone receptor with heterosexuality or homosexuality could be found. The lack of involvement of sex hormones generally in predisposing to homosexual behavior is apparent in the fact that solid indicators of this behavior can be discerned in both males and females well before the increases in these hormones during adolescence. “Gender-atypical” behavior as a forerunner of homosexuality has been noted for over a hundred years. Individuals often display preferences for opposite-gender playmates and play activities, and identify with opposite-gender role models, well before such activities could have a reproductive association. Gender-atypical behavior is apparent in children even before they enter school, and is an accurate predictor of future sexual preference across many different world cultures. We do not know of course that this is the aspect of homosexuality governed by Hamer’s X-linked gene, but if it is, we would expect it to act at a very early stage in development, possibly even from birth.

As part of human sexual development, there clearly must be a mechanism in both sexes that favors, in sexually mature adults, selection of a mate of the opposite gender. In the development of male homosexuality, there is not only a strong sexual attraction to males, but a fairly categorical rejection of females as mates. This rejection of the opposite sex is much less evident in nonheterosexual females, as reflected in the higher proportion who are bisexual. There is strong evidence in other mammals that mate selection is not learned behavior, and there is no reason to believe that it is in humans. Whether this mechanism is the same in males and females, we do not know. Evidence from both (p. 248 ) human and animal studies suggests that this mechanism is in fact genetically acquired. Ultimately, a genetic understanding of both male and female homosexuality will have to include an explanation of this change in the mate selection component of sexuality.

Is there evidence for homosexual behavior in animals other than humans, or is such behavior a unique attribute of the human brain? Many investigators have tried to answer this question over the years. Most scientists agree that homosexual behavior, beyond random same-sex playing and exploration in young animals, does not occur in animal species below the primates (humans, apes, monkeys, lemurs, tarsiers, and marmosets). In some species of primates there is overt behavior that is difficult in many ways to distinguish from human homosexual behavior, in both males and females. However, because we have no way of plumbing the psychological elements that are such an important part of human homosexuality, these interactions in nonhuman primates are usually referred to as same-sex sexual behavior.

In New World primates, which are evolutionarily older, same-sex sexual behavior does not get beyond the juvenile play stage observed in other animal orders. It is in some of the more recently evolved Old World species, such as apes and chimpanzees, that we begin to encounter startlingly human-like, same-sex sexual behaviors. Couples of the same sex often establish and maintain a long-term, stable, exclusive relationship. There may even be competition for same-sex mates, with displays of aggression and apparent jealousy. These couples engage in complex sexual behaviors, including genital-to-genital contact (particularly among females), and mutual genital manipulation.

Nevertheless, there is still disagreement about the extent to which primate behaviors of this sort are equivalent to human homosexuality. A key element, often overlooked, is that there are no well-characterized instances of preferential same-sex activity; in every case that has been examined, if given adequate access to a partner of the opposite sex, primates will abandon same-sex for opposite-sex interactions. Often same-sex interactions reflect inadequate access to mates of the opposite sex stemming from dominance structures within a social unit. Interestingly, even within the constrained same-sex sexual behavior paradigm, there is a difference in male versus female activities. Male same-sex behaviors almost always occur in the juvenile and adolescent years, whereas females rarely engage in same-sex behavior during these (p. 249 ) years; they almost always engage in such activities as adults, even after they have begun breeding with males and bearing young.

These behaviors are clearly distinct from anything found in lower animal orders, and may well represent at least the beginnings of the corresponding neurological and behavioral patterns in humans. Whether or not these behaviors cluster in familial lines in nonhuman primates—whether they are heritable—has never been determined.

The existence of genes predisposing to homosexuality poses an interesting challenge to current evolutionary theory. The conventional view is that when new allelic variants of genes arise through various mutational processes, whether or not these alleles spread from the individuals in which they arise into the general population depends upon how natural selection acts on them. The only way a new allele can be selected for further propagation within a species is if it somehow increases the reproductive success of the individuals that inherit it. And of course, therein lies the rub: How could an allele that diverts individuals from reproductive activities be established and maintained within a species over evolutionary time? Homosexual behavior in humans is at least as old as recorded history, and the existence of something very close to homosexual behavior in many of the higher primates suggests it has probably been present and maintained in humans from close to the time of our origin as a distinct species. How might this have happened? Homosexual men and women produce only about one-fifth as many children as their heterosexual counterparts. Ordinarily this would lead to rapid elimination of the underlying alleles in short order. But that doesn’t seem to be happening. Could it be that some of the genetic alleles predisposing to homosexual behavior confer (or conferred in the past) an advantage on human beings that we have yet to discern?

Allelic variants of genes associated with homosexuality are not the only ones whose survival is something of a mystery. It is becoming increasingly clear that many of the mechanisms underlying aging and death are under genetic control. It is difficult to perceive the reproductive advantage of those genes. There are also gene variants that cause hereditary diseases such as cystic fibrosis; as discussed in Appendix II, it has been estimated that 20 percent of the Caucasian population carries a defective allele of this gene. That is an extraordinary frequency for an allele that causes sterility and premature death in (p. 250 ) those unfortunate enough to inherit two copies. The best guess at present is that variants of genes that interfere with reproduction in the homozygous form may actually confer some unperceived benefit to individuals who carry only a single copy, the so-called “heterozygote advantage.” In some cases, such as sickle-cell anemia, we in fact know the advantage tendered by single copies of the errant alleles. But whether, or to what extent, heterozygote advantage could affect selection of genes involved in a quantitative trait like homosexuality is not clear.

Like research into the genetic basis of mental function, research into the genetic basis of human sexual preference—and specifically, a possible genetic explanation of homosexuality—has been criticized on two quite separate grounds. There have been, as expected, valid scientific criticisms of Hamer’s study, mostly focusing on the assumptions made by the researchers about the relationship between paired sibling studies and inheritance of complex genetic traits. Hamer’s group has responded to these criticisms in exchanges of opinion in scientific journals and at scientific meetings, as is perfectly normal in such cases. Everyone agrees that more data must be collected to ascertain exactly how genes may be involved, and doubtless future research designs may be modified somewhat in response to criticisms that have been made.

But research of this type inevitably evokes strong reactions of a political and social, as well as a scientific, nature. Some elements of society would very much like to believe that homosexuality is a lifestyle choice and thus a moral, rather than a biological, issue. Some would like to prove beyond a reasonable doubt that homosexuality is as genetically driven as any other human behavior, and that the element of choice plays little if any role in becoming homosexual. The lack of significant choice in becoming homosexual could provide a basis for legal protections extended to homosexuals as a class, and many people oppose this.

Yet defining homosexuality as a normal human behavioral variant, with a strong genetic component, will be a double-edged sword even for those who stand to gain the most from such a recognition. Some would like to see the new view of homosexuality used to define it as a disease, in the sense that we now think of alcoholism as a disease. And as discussed in the previous chapter and in Appendix II, one of the greatest fears about human genetics in general is that once we have isolated (p. 251 ) and deciphered every gene in the human genome, and know exactly what it does, there will be enormous temptation to combine that knowledge with our increasing ability to control reproduction, to favor transmission of those very same genes. It is not at all unimaginable that heterosexual couples with a family history of homosexuality may want to screen the embryos they produce for the underlying genes, and discard embryos they consider at risk for homosexuality. This is an entirely valid fear; the technology is already there, and it is only a matter of time until we know the identity of the genes involved.

Another legitimate worry on the part of some is that the same knowledge that would allow selective elimination of potentially homosexual embryos could be used to determine the probable sexual orientation of adults on the basis of a simple genetic test. This takes us into the much larger arena of concerns about genetic privacy. Health and life insurers might want to use genetic tests to estimate risks of HIV infection, just as they now routinely screen, using standard medical tests, for a range of other health risks prior to issuing insurance policies or determining premiums. Employers hiring into a particular workplace context may decide they want to use genetic information to determine someone’s sexual orientation, in the same way they routinely use psychological and personality tests prior to hiring individuals. These are questions we as a society must address, and soon. Hamer and his colleagues were clearly thinking of these questions when they wrote their first paper on the DXS52 locus; here is their closing paragraph:

Our work represents an early application of molecular linkage methods to a normal variation in human behavior. As the Human Genome Project proceeds, it is likely that many such correlations will be discovered. We believe that it would be fundamentally unethical to use such information to try to assess or alter a person’s current or future sexual orientation, either heterosexual or homosexual, or other normal attributes of human behavior. Rather, scientists, educators, policy makers and the public should work together to ensure that such research is used to benefit all members of society.

The overwhelming majority of the scientific community would agree. We must never allow individuals defined by molecular genetics as somehow different from the “norm,” on whatever basis, to be put at risk for selective and potentially prejudicial treatment. To the extent (p. 252 ) that information about genetic constitution is ever used to identify, isolate, or diminish any one segment of society to the advantage of another, all of the potential benefits of our knowledge of human genetics could be lost in a political backlash the likes of which we have not seen since the closing days of eugenics. These benefits are too important to all of us to let this happen. Every segment of society—physicians, scientists, legal experts, ethicists, and most important, common citizens—must make itself aware of the underlying problems and possible solutions, and join in the debate that will decide these issues as we enter into the twenty-first century.