Smoking While Pregnant: Title and subTitle BreakTransplacental Mutagenesis of the Fetus by Tobacco Smoke

The International Agency for Research on Cancer (IARC) has identified tobacco smoking as the cause of cancer at more organ sites (11) than any other human carcinogen.¹ Eight of these organ sites have been examined for smoking-associated genotoxic effects, and such effects have been found in all 8 sites.² The sites are the oral/nasal cavity, esophagus, pharynx/larynx, lung, pancreas, myeloid organs, bladder/ureter, and uterine cervix. To this long list of organs in adults in which tobacco smoking causes genotoxic effects might now be added the somatic epithelial cells of fetuses carried by mothers who smoke. As described by de la Chica and colleagues³ in this issue of JAMA, smoking during pregnancy was reported to be associated with increased chromosomal instability in amniocytes collected by amniocentesis. Such results, if substantiated, would provide direct evidence of tobacco-associated intrauterine mutagenesis and could have important implications for the immediate and long-term health effects of children born to mothers who smoke.

After interviewing 800 women, the authors identified 25 mothers for whom smoking was considered to be their only significant exposure to a mutagenic agent; the authors also identified 25 nonsmoking mothers who had little if any reported passive exposure to tobacco smoke. The smokers had a more than 10-year history of smoking and smoked more than 10 cigarettes per day while pregnant. A standard cytogenetic approach was used to identify and classify chromosomal alterations. A G-banding method was used to identify the locations of the alterations. The key finding reported by the authors was an approximately 3.5-fold increase in structural chromosomal aberrations in the amniocytes from smokers compared with those from nonsmokers (P = .002), with deletions and translocations accounting for a majority of these changes. In addition, the authors reported that chromosomal band 11q23 was involved in aberration formation more frequently in amniocytes from smokers than in those from nonsmokers (P<.05). The authors proposed that there could be a link between the observation of an increase in breakpoints at this site and a potential increase in childhood leukemia. However, the experimental design and results of the study do not permit a firm conclusion regarding this point.

The findings of de la Chica et al represent the first report of tobacco smoke–associated genotoxic damage in fetal epithelial cells from mothers who smoke. As such, these data add to the set of related and indirect data that have implicated tobacco smoke as a transplacental mutagen to the developing fetus. Indirect evidence for the presence of tobacco smoke–related mutagens in the fetal environment was obtained nearly 20 years ago by the demonstration that the cervical mucus of smokers was more mutagenic in the Salmonella (Ames) assay than that of nonsmokers.⁴ Likewise, organic extracts of amniotic fluid from smokers induced higher frequencies of sister chromatid exchanges (SCEs) in cultured mammalian cells than did those from nonsmokers.⁵ Analytical studies have found elevated levels of tobacco-specific mutagens or cotinine in amniotic fluid, fetal urine, or fetal blood from smoking mothers compared with nonsmoking mothers,^{5 - 9} providing direct evidence of the presence of such compounds in utero. Protein adducts of tobacco-specific compounds have been identified in cord blood,^{10 - 12} indicating that such compounds interact covalently with fetal macromolecules, with potentially mutagenic consequences.

Evidence of transplacental mutagenicity was obtained more than 10 years ago with the finding of elevated HPRT mutant frequencies in cord blood from smokers relative to that from nonsmokers using the autoradiographic method¹³ ; similar findings were obtained later for marijuana smokers.¹⁴ Although maternal smoking did not increase the HPRT mutant frequency in T lymphocytes of newborns based on the more definitive T-cell cloning assay,¹⁵ it did cause a change in mutation spectra (increased deletion frequency) in T cells from newborns of active¹⁶ as well as passive smokers.^{17 - 18} Molecular cytogenetic methods demonstrated that in utero exposure to tobacco smoke also increased frequencies of chromosomal translocations in cord blood of mothers who smoke.¹⁹ The results of de la Chica et al in amniocytes are consistent with the molecular cytogenetic and HPRT studies in cord blood in that a majority of the chromosomal mutations in amniocytes from mothers who smoked were observable deletions and translocations. Beyond these somatic-cell effects, evidence has been reviewed that provides some indirect support for the ability of tobacco smoke to be a germ-cell mutagen in the oocytes and sperm of parents who smoke, with the possibility that such mutations could be transmitted to their children.²

Although consistent with these previous reports, the study by de la Chica et al contains several limitations that urge caution and suggest that their findings should be viewed as highly preliminary. One limitation is lack of a direct or indirect exposure assessment, such as urinary cotinine measurements of the mother and DNA or protein adduct measurements of both maternal and fetal tissues.

A more serious limitation is illustrated by the figures in the article by de la Chica et al, which show that the observed chromosomal aberrations were of the chromatid type. These aberrations are produced by errors of DNA replication during S phase and must have been formed during the last DNA replication prior to the observed metaphase. Because the amniocytes were cultured prior to fixation, the observed chromatid aberrations were formed while the cells were in culture and not in the fetus. Although carryover into culture of some small portion of amniocytes with unrepaired DNA damage might have accounted for some of the observed aberrations, assigning all or most of these aberrations to smoking is not correct. The aberration frequencies of just a few samples of amniocytes from smokers were outside the frequency range for nonsmokers, suggesting that cell cycle and culture conditions may have influenced the results and accounted for the few outliers observed. It seems unlikely that smoking-induced DNA damage was retained in the cells for several days to several weeks and was converted into chromatid aberrations at the last S phase in culture.

The authors demonstrated that the 3-year higher mean maternal age of the smokers vs the nonsmokers was not a statistically significant factor in the higher observed frequency of aberrations in amniocytes from smokers vs nonsmokers. However, the frequency of reciprocal translocations in an individual increases with age,²⁰ as do the frequencies of other aberration classes²¹ ; this might influence the outcome if the analyses of de la Chica et al were based on the range of values in smokers vs nonsmokers rather than on the mean values. The background frequency of chromosomal aberrations in peripheral lymphocytes²¹ is generally reported to be lower than that observed for amniocytes in the present study, but the basis for this difference remains unexplained.

A final caution relates to the reported link between a specific breakpoint (11q23) among smokers and a possible risk for childhood leukemia. The numbers of observations of an 11q23 event are small (5 among all smokers), and many such observations are needed to be convincing. Most important, a single chromatid aberration at a band in a G-banding study of amniocytes cannot be extrapolated to an adverse health effect in the resulting child.

That so few studies have examined the genotoxic effects of maternal smoking on the fetus reflects the difficulty of such studies in terms of screening and selecting study participants, obtaining fetal cells for analysis, and the limited availability of assays and end points that have been available for use. Although obtaining suitable patients and cells may continue to be problematic, the development of new methods and end points, especially genomic methods, offers hope for progress in this important area of public health. A definitive demonstration of transplacental mutagenesis of fetal epithelial cells by maternal smoking requires further studies that take into account the concerns outlined here. In the meantime, the message to women based on the published literature remains clear: smoking during pregnancy can be hazardous for both the fetus and the mother.