The Nagasaki Prefecture, Japan (population, 1.5 million), is one of the hot endemic foci of human T-lymphotropic virus type I (HTLV-I) (1). The prevalence of HTLV-I carriers is ≈10% in the age group over 40 years old (40,000 individuals), ≈10 times the national average. The annual registry of adult T-cell leukemia (ATL) in the prefecture is ≈60 cases (estimated incidence, 100 cases), or 0.5% of the total deaths (1,2). An effective measure to control the endemic cycle of HTLV-I has been imperative, since practical ways to prevent or control ATL are not available.
MOTHER-TO-CHILD TRANSMISSION OF HTLV-I AS THE MAJOR PATHWAY
Our first report on mother-to-child transmission of HTLV-I in Nagasaki, Japan (1), revealed that HTLV-I seroprevalences of mothers and of children borne to carrier mothers were 3.7 and 22%, respectively. The product (0.8%) of the maternal prevalence and the transmission rate was close to the prevalence of children in the area (1%). The data strongly suggested that HTLV-I is transmitted mainly from mother to child. The prevalences of carriers' mothers were 12 of 13 (92%) in mothers of student carriers (age, 6-18 years) and 10 of 12 (83%) in mothers of pregnant carriers (1). Furthermore, the two of two pregnant carriers with seronegative mothers had been transfused for open heart surgery or for massive bleeding at the previous delivery. In Nagasaki, Japan, screening of blood donors commenced in 1987 essentially terminating transfusion-mediated transmission of HTLV-I. At this meeting, we learned that screening of donated blood has been commenced in several European countries even though the prevalence is very low. There remain many endemic countries where unscreened blood is still in use. For example, as many as 43% of hospitalized carrier children had seronegative mothers, but with histories of blood transfusions, in Gabon (3). Blood transfusions had probably been one of the major factors for mother-to-child transmission of HTLV-I even in Japan, since transfusions for anemic gravidas were common by 1980. Doctors should refrain from transfusion to young females, especially in endemic areas. HTLV-I may transmit from seronegative mothers: a breast-fed and seropositive child had been born to a mother (age, 23 years) who seroconverted between 26.5 and 29.5 years (4).
EVIDENCE FOR MILK-BORNE INFECTION
The first experimental evidence to support milk-borne infection of HTLV-I was the detection of infected cells in breast milk of carrier mothers (5). Intraoral inoculation of human ATL or breast-milk cells into common marmosets transmitted HTLV-I (6,7): the total inoculum size corresponded to only 200 ml of fresh human milk. Mother-to-child transmission of HTLV-I in rabbits further supported the conjecture (8,9). Cells rather than virions in the breast milk are responsible for the transmission of HTLV-I, since freeze-thawed milk did not transmit HTLV-I in culture (10) and 13 children borne to carrier mothers (11). Direct evidence for the port of entry has not been available, although lymphatic tissues in the oral cavity or Peyer's patches in the intestine are suspected. Milk-borne transmission may play a role also in the case of HTLV-II: 0 of 20 formula-fed children who were born to carrier mothers coinfected with human immunodeficiency virus (HIV) and HTLV-II in New York did not seroconvert (12).
THE ATL PREVENTION PROGRAM IN NAGASAKI, JAPAN
We classified elder siblings of the target children in the first report (1) by their feeding method. While 14 of 37 (38%) children breast-fed for >1 year were seropositive, 0 of 10 children never breast-fed were seropositive (p =.06 by Fisher's direct test, unidirectional) (13). Encouraged by the experimental transmission of HTLV-I in marmosets, we started preparations for an intervention study by distributing information and educating people in public health programs in the prefecture. In the summer of 1987, the ATL Prevention Program (APP) commenced in the Nagasaki Prefecture, Japan, after a 1-year pilot study (13). The purpose of the study was to obtain definite evidence that breast-feeding by carrier mothers is playing the major role in the endemic cycle of HTLV-I. In the APP, commercial laboratories are assigned the first screening. We retest positive and undetermined samples at the Department of Bacteriology, Nagasaki University, by both particle agglutination assay (PA; Fuji Rebio Co., Tokyo, Japan) and indirect immunofluorescent assay (IF) using an MT-2/CEM cell mixture as target (1). Still undetermined samples are confirmed by immunoblottings. We test all samples directly in the follow-up study. Our database system uses first and family name, birth date, husband's name, address, and phone number to identify each mother. The number of mothers screened each year decreased from 15,000 in 1989 to 10,000 in 1993 (Fig. 1). According to the prefectural census, ≈90% of gravidas enrolled for the screening. The gradual decline in screening size is consistent with the decreasing deliveries each year. The prevalence of carrier mothers was relatively constant, ≈4%.
Most cross-sectional studies in endemic areas reveal a steep rise in prevalence by increasing age, especially in females. This profile often serves as evidence of male-to-female sexual transmission, however, several follow-up studies in Japan could not observe frequent enough seroconversions to explain the marked increase associated with aging (14,15). We consider the apparent steep rise in prevalence with aging to be somewhat dependent on a birth cohort effect, at least in Japan. In Japan, the real Westernization of lifestyles started in the 1950s.
In the APP, the average age of gravidas (including multiparas) was ≈30 years, and that of seropositive gravidas was consistently higher than that of seronegative gravidas in each test year (Table 1). The difference of the average ages increased from 1989 to 1993. This is consistent with the significant decrease in prevalence among gravidas born in 1940 to 1970 (Table 2). Over 90% of carrier mothers agreed to refrain from breast-feeding. We register ≈400 children delivered by carrier mothers each year and ask them to visit one of 17 designated pediatric clinics distributed in the prefecture at 6, 12, 18, 24, and 36 months after delivery. Half of these children enrolled in the follow-up study.
AGE OF CHILDREN FOR DIAGNOSIS OF HTLV-I TRANSMISSION
The age of children for serological diagnosis of HTLV-I transmission is important. The maternal antibody titer decreased at a rate of 1/10 every 2 months. At the age of 6 months, a detectable amount of maternal antibodies remained in ≈20% of the children. Rarely, maternal antibodies still remained even at the age of 12 months (16). Most breast-fed and seropositive children seroconvert by the age of 3 years, and no additional seroconversions were observed for another 15 years (17). Seroconversions after the age of 2 are infrequent, but have been reported (16,18). In our experience, most children who did not seroconvert by the age of 12 months remained seronegative by the age of 3 years. Therefore, we usually diagnose at the age of 12 months (19). Probably the age of 18 months is in the safe zone for diagnosis of HTLV-I transmission, at least for statistical purposes.
Table 3 shows the current results of the APP. Since most mothers chose formula-feeding rather than breast-feeding in the APP, we included children born before their mothers enrolled in the APP. While the incidence of seroconversion in children breast-fed over 12 months continued to be high, 37 in 235 (15.7%), that in children only formula-fed was as low as 41 in 1141 (3.6%) at the age of ≥18 months (p < 0.001). The rate of transmission in breast-fed children was not significantly different from previously published data from other laboratories, 15-25% (15,20). At this meeting, Tajima et al. reported essentially the same results obtained in Tsushima and Kami-Goto: <3% of formula-fed children were seropositive. The data indicated that refraining from breast-feeding by carrier mothers prevented ≈80% of mother-to-child transmission. However, this type of intervention cannot be a choice in developing countries, where one of the major factors in infantile death is diarrhea or malnutrition.
AN ALTERNATIVE PATHWAY OF MOTHER-TO-CHILD TRANSMISSION
Since 3-4% of children born to carrier mothers still remain seropositive even after complete formula-feeding, searches for an alternative pathway are in progress. Detection of antigen-positive cells in cord blood after culture (21) or by polymerase chain reaction (PCR) (22,23) raised the possibility of intrauterine infection. We also confirmed that 18 of 717 (2.5%) cord blood samples were positive by nested PCRs directed to the gag and tax/rex genes (24). However, none of seven children whose cord blood was PCR positive seroconverted by 24-48 months. Furthermore, zero of nine formula-fed children who were confirmed HTLV-I positive and whose cord blood samples were available in the stock had been PCR positive for HTLV-I in their cord blood. Thus, the presence of proviral DNA in the cord blood sample is not a hallmark of intrauterine infection (25). Since the amount of proviral DNA in the cord blood, if present, was ≈1/10 of that in the maternal blood, medical-office or laboratory contamination of maternal blood was unlikely. Mothers who transmitted HTLV-I to their children even though formula-feeding had significantly lower titers of anti-HTLV-I antibody. The data suggested that HTLV-I transmission is established after delivery, possibly by another perinatal pathway. Further study is necessary to elucidate the matter.
There have been reports claiming that short-term breast-feeding (<7 months) significantly decreased mother-to-child transmission of HTLV-I, from 20 in 139 (14.4%) to 4 in 90 (4.4%); p = 0.018) (20). They claimed that the incidence of infection in short-term breast-fed children was not significantly higher than that in formula-fed children, 9 in 158 (5.7%). Infected cells transferred via breast milk may be abortive for infection in the presence of enough maternal antibody. The addition of anti-HTLV-I-containing serum to cocultures of infected and uninfected cells inhibited infection with HTLV-I (20). In rabbit models, passively transferred IgG containing anti-HTLV-I protected animals from infection (26). This concept also explains why babies with HTLV-I proviruses in the cord blood are free from infection (25).
In the APP, the seroconversion rate of children breast-fed for a short term (6.0%) was significantly lower than that for long-term breast-feeding (13.7% for >6 months and 15.7% for >12 months). The seroconversion rate of formula children was not significantly higher than that of short-term breast-fed children (p = 0.130). At the meeting Tajima reported data consistent with this conjecture. However, we are still reluctant to state that short-term breast-feeding is equally safe as formula-feeding, especially on an individual basis. A short-term breast-feeding strategy with careful control of the maternal antibody titer and duration of breast-feeding may minimize mother-to-child transmission even in developing countries.
PREDISPOSING FACTORS FOR TRANSMISSION IN CARRIER MOTHERS
If there are predisposing factors for HTLV-I transmission in mothers, this is another important area to explore. If the putative factor could discriminate nontransmitters from transmitters, mothers without the factor can rear children without much anxiety or distress. We reported previously that mothers with higher titers of PA antibody have a significantly higher tendency to transmit HTLV-I to children (27). Association of transmission with high anti-p40tax (28,29) and gp46 (30) titers has been reported. We have also reported that mothers with higher concentrations of infected cells in blood after culture have a higher tendency to infect children (13,19,31). This was consistent with our recent findings that the provirus load in the peripheral blood and the antibody titer by PA were correlated (p = 0.86) in Okinawan high-school students (32). Furthermore, using synthetic and cloned polypeptides of HTLV-I, the antigen correlated with transmission of HTLV-I has been explored (33). Transmitter mothers in the long-term breast-feeding group had significantly higher titers against env protein antibody than nontransmitters, while transmitter mothers in the formula-feeding group had significantly lower titers against env proteins. Antibody titers against gag and tax proteins did not show a significant difference in either group. The result for long-term feeders is consistent with previous reports, and that for formula-feeders is consistent with the putative protective effect of maternal antibody for HTLV-I transmission. Research on the possible role of human leukocyte antigen (HLA) haplotypes on HTLV-I transmission is under way. This type of genetic predisposing factor may explain the extraordinarily high proportion of HTLV-I carriers in families of ATL patients (34).
1. Hino S, Yamaguchi K, Katamine S, et al. Mother-to-child transmission
of human T-cell leukemia virus type-I. Jpn J Cancer Res
2. Kinoshita K, Ikeda S, Suzuyama J, et al. Annual incidence of adult T-cell leukemia
-lymphoma (ATL-L) from ATL virus-carriers in Nagasaki Prefecture. Nagasaki Med J
3. Delaporte E, Peeters M, Bardy JL, et al. Blood transfusion as a major risk factor for HTLV-I infection among hospitalized children in Gabon (Equatorial Africa). J Acquir Immune Defic Syndr
4. Umemoto M, Take H, Kuraya K. Vertical transmission of HTLV-I from a seronegative mother using the PA method. Acta Paediatr Jpn
5. Kinoshita K, Hino S, Amagasaki T, et al. Demonstration of adult T-cell leukemia
virus antigen in milk from three seropositive mothers. Gann
6. Yamanouchi K, Kinoshita K, Moriuchi R, et al. Oral transmission of human T-cell leukemia virus type-I into a common marmoset (Callithrix jacchus) as an experimental model for milk-borne transmission. Jpn J Cancer Res
7. Kinoshita K, Yamanouchi K, Ikeda S, et al. Oral infection of a common marmoset with human T-cell leukemia virus type-I (HTLV-I) by inoculating fresh human milk of HTLV-I carrier mothers. Jpn J Cancer Res
8. Uemura Y, Kotani S, Yoshimoto S, et al. Mother-to-offspring transmission of human T cell leukemia virus type I in rabbits. Blood
9. Hirose S, Kotani S, Uemura Y, et al. Milk-borne transmission of human T-cell leukemia virus type I in rabbits. Virology
10. Ando Y, Kakimoto K, Tanigawa T, et al. Effect of freeze-thawing breast milk on vertical HTLV-I transmission from seropositive mothers to children. Jpn J Cancer Res
11. Ando Y, Nakano S, Saito K, et al. Prevention
of HTLV-I transmission through the breast milk by a freeze-thawing process. Jpn J Cancer Res
12. Kaplan JE, Abrams E, Shaffer N, et al. Low risk of mother-to-child transmission
of human T lymphotropic virus type II in non-breast-fed infants. The NYC Perinatal HIV Transmission Collaborative Study. J Infect Dis
13. Hino S, Sugiyama H, Doi H, et al. Breaking the cycle of HTLV-I transmission via carrier mother's milk. Lancet
14. Hino S. Maternal-infant transmission of HTLV-I: implication of disease. In: Blattner WA, ed. Human retrovirology: HTLV
. New York: Raven Press, 1990:363-75.
15. Tachibana N, Okayama A, Murai K, et al. Seroepidemiological analysis of the age specific prevalence of anti HTLV-I in Miyazaki Prefecture. Kansenshogaku Zasshi
1990;64:570-4 (in Japanese).
16. Hino S, Katamine S, Kawase K, et al. Intervention of maternal transmission of HTLV-I in Nagasaki, Japan
17. Kusuhara K, Sonoda S, Takahashi K, Tokugawa K, Fukushige J, Ueda K. Mother-to-child transmission
of human T-cell leukemia virus type I (HTLV-I): a fifteen-year follow-up in Okinawa, Japan
. Int J Cancer
18. Kajiyama W, Kashiwagi S, Hayashi J, et al. Study of seroconversion of antibody to human T-cell lymphotropic virus type-I in children of Okinawa, Japan
. Microbiol Immunol
19. Sugiyama H, Doi H, Yamaguchi K, Tsuji Y, Miyamoto T, Hino S. Significance of postnatal mother-to-child transmission
of human T-lymphotropic virus type-I on the development of adult T-cell leukemia
/lymphoma. J Med Virol
20. Takahashi K, Takezaki T, Oki T, et al. Inhibitory effect of maternal antibody on mother-to-child transmission
of human T-lymphotropic virus type I. The Mother-to-Child Transmission
Study Group. Int J Cancer
21. Satow Y, Hashido M, Ishikawa K, et al. Detection of HTLV-I antigen in peripheral and cord blood lymphocytes from carrier mothers. Lancet
22. Saito S, Furuki K, Ando Y, et al. Identification of HTLV-I sequence in cord blood mononuclear cells of neonates born to HTLV-I antigen/antibody-positive mothers by polymerase chain reaction. Jpn J Cancer Res
23. Fujino T, Fujiyoshi T, Yashiki S, Sonoda S, Otsuka H, Nagata Y. HTLV-I transmission from mother to fetus via placenta. Lancet
24. Kawase K, Katamine S, Moriuchi R, et al. Maternal transmission of HTLV-I other than through breast milk: discrepancy between the polymerase chain reaction positivity of cord blood samples for HTLV-I and the subsequent seropositivity of individuals. Jpn J Cancer Res
25. Katamine S, Moriuchi R, Yamamoto T, et al. HTLV-I proviral DNA in umbilical cord blood of babies born to carrier mothers. Lancet
26. Kataoka R, Takehara N, Iwahara Y, et al. Transmission of HTLV-I by blood transfusion and its prevention
by passive immunization in rabbits. Blood
27. Hino S, Doi H, Yoshikuni H, et al. HTLV-I carrier mothers with high-titer antibody are at high risk as a source of infection. Jpn J Cancer Res
28. Sawada T, Tohmatsu J, Obara T, et al. High risk of mother-to-child transmission
of HTLV-I in p40tax antibody-positive mothers. Jpn J Cancer Res
29. Kashiwagi S, Kajiyama W, Hayashi J, et al. Antibody to p40tax protein of human T cell leukemia virus 1 and infectivity. J Infect Dis
30. Wiktor SZ, Pate EJ, Murphy EL, et al. Mother-to-child transmission
of human T-cell lymphotropic virus type I (HTLV-I) in Jamaica: association with antibodies to envelope glycoprotein (gp46) epitopes. J Acquir Immune Defic Syndr
31. Kinoshita K, Amagasaki T, Hino S, et al. Milk-borne transmission of HTLV-I from carrier mothers to their children. Jpn J Cancer Res
32. Miyata H, Kamahora T, Iha S, Katamine S, Miyamoto T, Hino S. Dependency of antibody titer on provirus load in human T lymphotropic virus type I carriers: an interpretation for the minor population of seronegative carriers. J Infect Dis
33. Hino S, Katamine S, Miyamoto T, et al. Association between maternal antibodies to the external envelope glycoprotein of human T-lymphotropic virus type I. Maternal anti-env antibodies correlate with protection in non-breast fed children. J Clin Invest
34. Momita S, Ikeda S, Amagasaki T, et al. Survey of anti-human T-cell leukemia virus type I antibody in family members of patients with adult T-cell leukemia
. Jpn J Cancer Res
Proceedings of the VIIth International Conference on Human Retrovirology: HTLV