Human T lymphotropic virus types I and II proviral sequences in Argentinian blood donors with indeterminate Western blot patterns

Human T-cell lymphotropic virus (HTLV) seroindeterminate blood donors have been reported worldwide including Argentina. To investigate the significance of HTLV-I/II seroindeterminate Western blot (WB) patterns, we conducted an 8-year cross-sectional study. Of 86,238 Argentinian blood donors, 146 sera were reactive by screening tests. The WB results indicated that 20% were HTLV-I reactive, 8% HTLV-II reactive, 61% indeterminate, and 11% negative. The overall seroprevalence was 0.034% for HTLV-I, 0.014% for HTLV-II, and 0.103% for indeterminate. In 57 reactive specimens, HTLV-I/II provirus could be examined by type specific PCR for tax, pol, and env regions. When at least two gene fragments were amplified HTLV-I/II infection was considered confirmed. PCR results confirmed all WB seropositive samples for HTLV-I (n = 15), and HTLV-II (n = 7), and the only WB negative case was also PCR negative, showing a complete concordance between PCR and WB. However, of 34 WB seroindeterminate sera studied by PCR, in 5 was proviral DNA amplified. According to our criteria PCR confirmed one to be HTLV-I, and one HTLV-II, 3 remained indeterminate since only tax sequences were amplified. Among WB indeterminate samples tested by PCR, most of their serological profile showed reactivity to gag codified proteins but lacked env reactivities (70%). One sample with a WB gag pattern showed proviral tax sequences, but of the four samples with reactivity to env proteins GD21 (n = 3) or rgp46II (n = 1) PCR results indicated that one was HTLV-I, one was HTLV-II, and two were indeterminate (only tax sequences). In conclusion, the majority of HTLV-seroindeterminate WB donors exhibited a gag indeterminate profile lacking HTLV provirus, and were thus considered uninfected. However, seroreactivity to env proteins, in particular to GD21, may indicate infection and a follow-up study of each seroreactive blood donor should be considered.     

Primary isolation of human T-cell leukemia-lymphoma virus types I and II: use for confirming infection in seroindeterminate blood donors.

We describe the use of an immunofluorescence assay and coculture to confirm human T-cell leukemia-lymphoma virus (HTLV) infection. Peripheral blood mononuclear cells from 32 of 32 seropositive donors were positive in the immunofluorescence assay, and 63% of their cocultures produced p24 antigen. Specific antibodies distinguished HTLV type I (HTLV-I) from HTLV-II. HTLV-I or HTLV-II was isolated from donors with indeterminate serologic test results.     

Characterization of antibody reactivity to human T-cell lymphotropic virus types I and II using immunoblot and radioimmunoprecipitation assays.

We have characterized the immunoreactivity to human T-cell lymphotropic virus type I (HTLV-I) among 26,983 persons of various seroprevalence groups by using enzyme immunoassay, immunoblot (IB), and radioimmunoprecipitation assays (RIPA) in accordance with Public Health Service recommended guidelines for the interpretation of serologic test results for HTLV-I infection. IB-indeterminate serum specimens (n = 178) were reactive to HTLV-I gag proteins, and no serum contained only env reactivity. Overall, RIPA resolved 40% of IB-indeterminate serum samples; however, the probability that RIPA would confirm IB-indeterminate samples depended on the seroprevalence of the population tested. HTLV-I gag p19-only reactivity on IB was not a reliable marker of HTLV-I infection, while gag p24 reactivity on IB was clearly associated with positive seroreactive specimens. IB and RIPA tests did not clearly distinguish between HTLV-I and HTLV-II seroreactivities. These data emphasize that patterns of immunoreactivity to HTLV-I antigens are dependent upon the seroprevalence of the risk groups tested. In addition, RIPA detected antibodies to env proteins present in low titer in a substantial number of IB gag-only reactive sera and resolved the HTLV-I antibody status of these sera.     

Synthetic peptide-based immunoassays for distinguishing between human T-cell lymphotropic virus type I and type II infections in seropositive individuals

Until now, serologic tests that distinguish the closely related human T-cell lymphotropic virus types I (HTLV-I) and II (HTLV-II) infections have not been available. Synthetic peptide assays, employing peptides derived from the core and envelope proteins of HTLV-I and HTLV-II (SynthEIA and Select-HTLV tests), were evaluated for the ability to serologically discriminate HTLV-I and HTLV-II infections. Of 32 HTLV-I- and 57 HTLV-II-positive serum specimens from individuals whose infections were confirmed by polymerase chain reaction, the SynthEIA test categorized 29 (91%) as HTLV-I and 50 (88%) as HTLV-II, and 10 (11%) were nontypeable. In contrast, the Select-HTLV test categorized 32 (100%) as HTLV-I and 55 (96%) as HTLV-II, and 2 (2%) were nontypeable. The specificity of both the assays in seropositive serum specimens was 100% in that none of the specimens were incorrectly classified. Additional serum specimens obtained from clinically diseased patients from the United States (n = 8) and asymptomatic carriers and patients from Japan (an endemic population for HTLV-I; n = 40) were categorized as HTLV-I by at least one of the assays, while serum specimens from Guaymi Indians from Panama (an endemic population for HTLV-II; n = 13) were categorized as HTLV-II. Thus, peptide enzyme immunoassays appear to represent a simple technique employing chemically synthesized antigens for discrimination between antibodies of HTLV-I and HTLV-II.     

Prevalence of infection by human T-cell leukemia virus types I and II in southern Spain

To assess the spread of human T-cell leukemia virus (HTLV) type I and II in different population groups at potential risk of infection in Spain, a total of 756 subjects were studied: 453 belonging to groups at risk for retrovirus infection, 255 with diseases potentially linked to HTLV-I/II infection and 48 immigrants from endemic areas. An HTLV-I viral-lysate enzyme immunoassay (EIA) with a recombinant transmembrane envelope protein incorporated was used to screen serum samples. Reactive specimens were confirmed by Western blot strips spiked with recombinant proteins that differentiated HTLV-I from HTLV-II. Infection was then verified by the polymerase chain reaction (PCR). Serum samples from 19 of the 756 subjects analyzed (2.5 %) were reactive for HTLV by EIA. One of these was from an intravenous drug user (IVDU) in whom HTLV-II infection was confirmed by Western blot and PCR; a specimen from another IVDU showed Western blot reactivity for both retroviruses, but PCR results were negative. Lastly, Western blot confirmed the presence of HTLV in one of the immigrant subjects. Western blot did not verify HTLV infection in the remaining 16 cases, indicating a high rate of nonspecific anti-HTLV reactivity when a second-generation EIA screening test was applied. These results suggest that HTLV is present in Spain among populations at high risk for HTLV, although at a very low rate and restricted to intravenous drug users and individuals immigrating from endemic areas.     

Detection of antibodies to trans-activator protein (p40taxI) of human T-cell lymphotropic virus type I by a synthetic peptide-based assay.

Antibodies to human T-cell lymphotropic virus type I (HTLV-I) trans-activator protein (p40taxI) were determined in serum specimens from individuals infected with HTLV-I (n = 138) and HTLV-II (n = 19). Western blot (immunoblot) analysis using recombinant tax demonstrated the presence of anti-tax antibodies in 96% of patients (25 of 26) with HTLV-I-associated myelopathy, 43% of those (20 of 46) with adult T-cell leukemia, and 61% of asymptomatic HTLV-I blood donors (40 of 66); only one of the HTLV-II specimens reacted with the recombinant tax protein. Synthetic peptides (Tax8(106-125), Tax22(316-335), Tax-23(331-350), and Tax-24(336-353) representing the immunodominant epitopes of¿ p40taxI detected anti-tax antibodies in 66 (48%), 50 (36%), 66 (48%), and 64 (46%) of 138 HTLV-I-positive specimens, respectively. An enzyme immunoassay using an equimolar ratio of these four peptides allowed sensitive detection of anti-tax antibodies in 96% of patients (25 of 26) with HTLV-1-associated myelopathy, 52% of adult T-cell leukemia patients (24 of 46), and 62% of asymptomatic HTLV-1-infected donors (41 of 66). The synthetic peptide-based cocktail assay was HTLV-I specific, since none of the HTLV-II-infected specimens reacted with these peptides. Interestingly, the corresponding regions from the HTLV-II tax protein, Tax8II(106-125), and Tax-22II(312-331) did not react with either HTLV-II or HTLV-I specimens. Thus, a synthetic peptide-based assay composed of immunodominant epitopes located towards the amino terminus and the C terminus of p40taxI provides a reliable and sensitive assay for the detection of anti-tax antibodies in seroepidemiologic studies.     

Detection and differentiation of antibodies to human T-cell lymphotropic virus types I and II by the immunofluorescence method.

We compared the sensitivities of the prototype human T-cell lymphotropic virus type I (HTLV-I)- and HTLV-II-transformed cell lines, MT2 and Mo-T, with that of an HTLV-II-infected cell line, clone 19, established in our laboratory, in the immunofluorescence (IF) test for detection of antibody to HTLV-I and HTLV-II. In addition, IF antibody titers with the three antigens were determined, and the results were compared with HTLV-I and HTLV-II typing by polymerase chain reaction (PCR). The MT2 cell line was more sensitive than the two HTLV-II cell lines for detecting HTLV-I antibody by IF, and clone 19 was more sensitive than Mo-T or MT2 for measuring HTLV-II antibody. In the titration study, the antigen that gave the highest titer correlated completely with the HTLV type determined by PCR, indicating that the relatively simple IF titration method can be used for differentiating HTLV-I and HTLV-II antibody in sera and plasmas.     

Evaluation of two commercial human T-cell lymphotropic virus western blot (immunoblot) kits with problem specimens.

We evaluated two commercial human T-cell lymphotropic virus (HTLV) Western blot (WB; immunoblot) kits, Cambridge Biotech Corp. (CBC) and Diagnostic Biotechnology Ltd. (DBL). Both methods employ HTLV type I (HTLV-I) viral lysate and rgp21. The DBL WB kit also distinguishes between HTLV-I and HTLV-II antibodies, using an HTLV-I-specific and an HTLV-II-specific recombinant. Fifty weakly reactive HTLV-II-positive plasma specimens which were falsely negative with the Abbott enzyme immunoassay (EIA) and 50 Ortho EIA false-positive samples were selected to determine sensitivity and specificity. The sensitivities of the CBC and the DBL WB kits were 90 and 68%, respectively. All positive samples reacted with rgp21 in both kits, but some did not display core bands. Five samples were typed as HTLV-I and four were typed as dual infection by the DBL WB kit. The specificities of the CBC and DBL kits were 48 and 70%, respectively. The most prevalent WB reaction with the negative samples was with the core protein, p19, followed by p24 and p28 for CBC and rgp21 and p28 for DBL. DBL had two false-positive interpretations, and CBC had none, rgp21 was the most sensitive antigen in both kits for the weakly reactive HTLV-II samples. If all samples not reacting with this protein were interpreted as WB negative, regardless of other bands, the specificity would improve to 90% for CBC and 86% for DBL.     

Comparison of Western immunoblot antigens and interpretive criteria for detection of antibody to human T-lymphotropic virus types I and II.

Twenty human T-lymphotropic virus type I (HTLV-I) antibody-positive sera from Japan, Hawaii, and the Marshall Islands and 15 HTLV type II (HTLV-II) antibody-positive sera from intravenous drug users in the United States were tested by immunoblotting with two recombinant HTLV-I proteins and three commercial kits to determine whether there were any differences in reactions between HTLV-I- and HTLV-II-positive sera by the Western immunoblot method and, also, to evaluate the ability of these reagents to detect HTLV-I- and HTLV-II-seropositive individuals by using the recommended Western blot interpretation. These sera were first extensively characterized by immunofluorescence, enzyme immunoassay, radioimmunoprecipitation assay, and Western blot using HTLV-I and HTLV-II viral lysates and an envelope (env) recombinant protein. Although both HTLV-I- and HTLV-II antibody-positive sera reacted with the env protein gp68, reactions with the gp46 env antigens appeared to be specific for HTLV-I. It was found that the use of either p19 or p24 core bands plus an env reaction instead of only the p24 plus env reaction (as presently recommended) increased the number of positive interpretations for HTLV-I but had no effect on the number of HTLV-II-positive interpretations.     

Seroprevalence of HTLV-1 and HTLV-2 among intravenous drug users and persons in clinics for sexually transmitted diseases.

Background. The human T-cell lymphotropic virus Type I (HTLV-I) is associated with adult T-cell leukemia and myelopathy, whereas HTLV-II infection has uncertain clinical consequences. We assessed the seroprevalence of these retroviruses among intravenous drug users and among patients seen at clinics for sexually transmitted diseases (STD clinics). METHODS. We used serum samples that were collected in eight cities in 1988 and 1989 during surveys of human immunodeficiency virus infection among intravenous drug users entering treatment and persons seen in STD clinics. The serum samples were tested for antibodies to HTLV, and positive specimens were tested further by a synthetic peptide-based enzyme-linked immunosorbent assay to differentiate between HTLV-I and HTLV-II. RESULTS. Among 3217 intravenous drug users in 29-drug-treatment centers, the median seroprevalence rates of HTLV varied widely according to city (range, 0.4 percent in Atlanta to 17.6 percent in Los Angeles). Seroprevalence increased sharply with age, to 32 percent in persons over 44 years of age. HTLV infection was more common among blacks (15.5 percent) and Hispanics (10.7 percent) than among whites (4.1 percent), and it was strongly associated with a history of heroin injection (P less than or equal to 0.001). Among 5264 patients in 24 STD clinics, the median rates of HTLV infection were much lower (range, 0.1 percent in Atlanta and Newark to 2.0 percent in Los Angeles). Again, this infection was more common among intravenous drug users (7.6 percent) than among non-drug users (0.7 percent). Eighty-four percent of the seropositive samples from drug-treatment centers and 69 percent of those from STD clinics were due to HTLV-II infection (P = 0.03). CONCLUSIONS. HTLV infections are common among intravenous drug users and are primarily caused by HTLV-II. Among patients seen at STD clinics, HTLV is strongly associated with intravenous drug use, but the retrovirus is also prevalent among non-drug users.     

In vivo fluctuation of HTLV-I and HTLV-II proviral load in patients receiving antiretroviral drugs

HTLV-I and HTLV-II infect T lymphocytes. A high HTLV-I proviral load in peripheral blood mononuclear cells (PBMCs) has been associated with a higher risk of neurologic disease. For HTLV-II, large numbers of infected lymphocytes might contribute to accelerate the immunodeficiency and increase the risk of neuropathy in HTLV-II/HIV-1 coinfected people. We have examined the impact of antiretroviral drugs on HTLV proviral load, testing longitudinal samples collected from 1 HTLV-I infected patient suffering HTLV-I-associated myelopathy (HAM), and two HTLV-II/ HIV-1 coinfected subjects. The HAM patient showed a reduction greater than 2 log in the peripheral proviral load after being treated with zidovudine and lamivudine. In contrast, potent antiretroviral treatment in HIV-1/HTLV-II coinfected carriers produced an initial increase in the HTLV proviral load, which was followed by a reduction greater than 1 log thereafter. In conclusion, antiretroviral drugs seem to reduce HTLV proviral load, although in HIV-1 coinfected persons a transient increase in HTLV proviral load could reflect the rapid blocking of HIV-1 replication occurring in response to therapy, thus causing an increase in the number of circulating T lymphocytes carrying HTLV proviral DNA.     

Molecular characterization of human T-cell lymphotropic virus type 2 (HTLV-II) from people living in urban areas of Sao Paulo city: evidence of multiple subtypes circulation

BACKGROUND: In Brazil, human T-cell lymphotropic virus type I and type II (HTLV-I and HTLV-II) are co-circulating and possess approximately 65% homology, which results in high cross-reactivity in serological tests. Based on the detection of EIA and Western blot (WB) tests, HTLV serodiagnosis yields indeterminate results in high-risk population, with the true determination of HTLV-II prevalence requiring a combined serological and molecular analysis. Molecular analysis of HTLV-II isolates has shown the existence of four distinct subtypes: IIa, IIb, IIc, and IId. The aim of this study was to evaluate the routine EIA and WB used in Sao Paulo city, as well as molecular methods for confirmation of infection and HTLV-II subtype distribution. Results: Two hundred ninety-three individuals, who were enrolled in the HTLV out-clinic in Sao Paulo city, Brazil, between July 1997 and May 2003, were tested by EIAs, and positive sera 232 (79%) reactive by one of the tests. When these sera were tested by WB revealed 134 were HTLV-I, 28 HTLV-II, 4 HTLV-I/II, and 48 were indeterminate. Polymerase chain reaction (PCR) on the indeterminate group showed that 20 (42%) were HTLV-II and 28 were negative. From a total of 48 HTLV-II subjects with DNA available, restriction fragment length polymorphism (RFLP) of the env region revealed 47 HTLV-IIa and 1 HTLV-IIb. The phylogenetic analysis was performed on 23 samples, which identified 19 as subtype a, Brazilian subcluster, and 4 as subtype b. This is the first time HTLV-II subtype b has been described in Brazil. However, further studies, such as a complete nucleotide DNA sequencing, need to be done to confirm these findings.     

Serologic confirmation of simian T-lymphotropic virus type I infection by using immunoassays developed for human T-lymphotropic virus antibody detection.

Serum specimens from diverse species of Old World monkeys, categorized as seropositive (n = 97) or seronegative (n = 23) for human T-lymphotropic virus (HTLV) infection, were tested by using recombinant env-spiked Western immunoblot assays and synthetic peptide assays for simultaneous detection and discrimination of simian T-lymphotropic virus (STLV) infection. Of the 97 seropositive specimens, 93 reacted with the recombinant transmembrane (r21env) protein and 90 reacted with a recombinant, MTA-1, derived from the central region of the external glycoprotein of HTLV-I (rgp46env), thus yielding test sensitivities of 96 and 93%, respectively. While 1 of the 23 negative monkey specimens reacted with r21env, none reacted with rgp46env, for overall specificities of 96 and 100%, respectively. Analysis of synthetic peptide-based immunoassays demonstrated that while 85 of 97 (88%) seropositive specimens reacted with HTLV-I-specific epitope (p19gag), none of the specimens reacted with HTLV-II-specific epitope (gp52env). These results show that recombinant envelope-spiked Western blots provide a simple means for serologic confirmation of STLV-I infection and that type-specific synthetic peptides can be used to confirm the virus type in seropositive monkey specimens.     

Nucleotide sequence of the protease-coding region in an infectious DNA of simian retrovirus (STLV) of the HTLV-I family

A provirus clone of simian T-cell leukemia virus isolated from a pigtailed monkey (PT-STLV), which is 90% homologous to HTLV-I, was shown to be biologically active in transfection assay. In transfected cells, gp61env, Pr55gag, and the mature gag proteins p24, p21, and p15 were detected, and type C particles were produced. The virus could be transmitted from the transfectants to recipient cells by cocultivation. In this biologically active provirus clone, a coding frame, possibly for protease, was identified between the gag and pol genes. The corrected sequence of the protease region of HTLV-I was also found to have a single open reading frame overlapping the gag and pol genes, although it has an amber codon in the middle of the frame. Thus, a single coding frame, which is different from those of gag and pol, is common to proteases of the HTLV family including HTLV-I.     

Development and evaluation of a human T-cell leukemia virus type I serologic confirmatory assay incorporating a recombinant envelope polypeptide.

A recombinant protein derived from the gp21 region of the human T-cell leukemia virus type I (HTLV-I) env gene was synthesized in Escherichia coli and purified by reversed-phase high-performance liquid chromatography. The purified protein was free of contaminating bacterial proteins and retained reactivity with human HTLV-I- and HTLV-II-positive sera and a gp21 monoclonal antibody. An immunoblot procedure using the recombinant polypeptide in conjunction with native viral proteins was more sensitive than the conventional immunoblot and radioimmunoprecipitation confirmatory assays for detection of antibodies to HTLV-I and HTLV-II env-encoded gene products. The recombinant protein was equally reactive with sera from polymerase chain reaction-confirmed HTLV-I or HTLV-II infections. Furthermore, on the basis of the differential reactivities of gp21-positive sera with the HTLV-I p19 and p24 gag-encoded proteins, an algorithm was proposed to distinguish exposure to HTLV-I from exposure to HTLV-II. These results establish the utility of a modified immunoblot assay incorporating a recombinant envelope polypeptide as an alternative to existing HTLV-I-confirmatory assays.     

Differential diagnosis of HTLV-I and HTLV-II infections by restriction enzyme analysis of 'nested' PCR products.

A 'nested' polymerase chain reaction (PCR) assay is described which is capable of detecting single copies of human T-cell lymphotropic virus (HTLV) in genomic DNA extracted from peripheral blood mononuclear cells (PBMCs). A single set of 'nested' oligonucleotide primers, based on the highly conserved tax/rex region of the viral genome, was able to detect both HTLV-I and HTLV-II proviral sequences in clinical samples of diverse geographical origins, from the United States, Great Britain, Japan, the Caribbean, Italy, Greece, Iraq and West Africa. Rapid discrimination between HTLV-I and HTLV-II infections was achieved by restriction enzyme analysis of unpurified second-round PCR products, even in those cases in which serological assays had failed to provide a definitive result. Over a 2-year period, a total of 53 HTLV infections (37 HTLV-I and 16 HTLV-II) were identified by this technique and complete concordance with serological typing, available in 41 cases, was observed.     

Transformation of guinea pig leukocytes by coinfection with human T-cell lymphotropic virus types I and II

OBJECTIVE: The susceptibility of guinea pigs to human T-cell lymphotropic virus (HTLV) infection and of their cardiac blood mononuclear cells (CBMCs) to HTLV-induced transformation were investigated. STUDY DESIGN/METHODS: Guinea pig CBMCs were cocultured with HTLV-infected cell lines. Guinea pigs were then inoculated with transformed guinea pig CBMCs. RESULTS: The coculture experiment gave rise to a guinea pig cell line, GP-1, that was coinfected with both HTLV-I and HTLV-II as shown by immunofluorescence staining, electron microscopy, polymerase chain reaction (PCR) using primers specific for the pol region of each virus, and Southern blot hybridization. The GP-1 cell line expressed T-cell markers and monocyte/macrophage markers. Three guinea pigs given an intraperitoneal inoculation of GP-1 cells seroconverted for HTLV-I and became positive for HTLV-I, HTLV-II, or both, as confirmed by PCR. CONCLUSIONS: Guinea pigs and their CBMCs can be infected with HTLV-I and HTLV-II. This animal system may be useful as an experimental model of HTLV-I and HTLV-II infection.     

Immunologic cross-reactivity between structural proteins of human T-cell lymphotropic virus type I and the blood stage of Plasmodium falciparum.

To determine the serologic cross-reactivity between human T-cell lymphotropic virus type I (HTLV-I) and parasite antigens, we measured antibody responses against HTLV-I, Plasmodium falciparum, Plasmodium vivax, and Brugia malayi in serum specimens obtained from regions where malaria (n = 482) and filariasis (n = 101) are endemic. Analysis of immune reactivity to HTLV-I antigens showed that specimens from regions where malaria is endemic had significantly higher rates of enzyme immunoassay (EIA) reactivity (76 of 482 [15.8%] than those from regions where filariasis is endemic (0 of 101 [0%]). Western blot (immunoblot) analysis of the HTLV-I EIA-reactive specimens demonstrated predominant Gag reactivity (HTLV-Iind). Only two specimens each from Indonesia and Brazil and four specimens from Papua New Guinea had Env reactivity by radioimmunoprecipitation analysis. Furthermore, a positive correlation between HTLV-EIA and titers of antibody to the blood stage of P. falciparum (rs = 0.24, P < 0.005) was discerned; no correlation was observed between antibodies to the blood stage or the circumsporozoite protein of P. vivax and the circumsporozoite protein of P. falciparum. In addition, P. falciparum-infected erythrocyte lysate specifically abrogated binding of Gag-specific antibodies in HTLV-Iind specimens from regions where malaria is endemic without affecting binding in HTLV-I-seropositive specimens, suggesting that the immunologic cross-reactivity between HTLV Gag proteins and malaria parasites is restricted to the blood-stage antigens of plasmodia in specimens from regions where malaria is endemic. However, HTLV-seroindeterminate specimens from the United States did not demonstrate serologic cross-reactivity, suggesting that antigenic mimicry of HTLV proteins extends to other nonplasmodial antigens as well.     

Comparison of immunofluorescence, enzyme immunoassay, and Western blot (immunoblot) methods for detection of antibody to human T-cell leukemia virus type I.

A total of 3,349 serum samples were screened by the immunofluorescence (IF) method for antibody to human T-cell leukemia virus type I (HTLV-I). Only 9 of 2,409 specimens from selected individuals, blood bank donors, patients with encephalitis-meningitis, and human immunodeficiency virus antibody-positive homosexual or bisexual men were reactive by IF. In addition, 940 serum samples from intravenous drug abusers were tested by IF and also by an HTLV-I enzyme immunoassay (EIA) method. Of these, 222 (24%) were positive for both HTLV-I and HTLV-II antigens by IF, and 191 of these 222 were also reactive in the HTLV-I EIA. Of the 31 IF-positive, EIA-negative serum samples, 20 exhibited optical density readings greater than or equal to 70% of the positive cutoff in the EIA, and 29 samples reacted with 1 or more bands in the Western blot (immunoblot) test. An additional 10 specimens that were EIA negative reacted only with HTLV-I by IF. Differences in staining morphology and in reactions on HTLV-I and HTLV-II antigens before and after absorption of the serum specimens with HTLV-I and HTLV-II-infected cell pellets revealed six distinct serological patterns by IF. These results indicate that infections by HTLV-I or by another closely related retrovirus(es) occur in California. Further studies utilizing statistically valid sampling methods are needed to estimate true prevalence rates among various groups. IF and Western blot tests should supplement the EIA method to maximize sensitivity and specificity of test procedures.