Sixteen previously untreated ATL patients (2 acute ATL, 2 ATL lymphoma, and 12 chronic ATL) received arsenic/IFN/zidovudine treatment. The patients’ characteristics are listed in Table 1. Clinical data, response to therapy and follow up data were previously reported for 10 of these patients 30 . Briefly, all patients initially presented with symptomatic disease requiring treatment. The most frequent symptoms were cutaneous manifestations with maculopapular rash, severe itching, and skin ulcerations. At day 30 and as previously reported 30 , treatment with arsenic, IFN, and AZT resulted in a good albeit partial response in all patients except one ATL lymphoma: 9 patients achieved partial response (PR) and 6 patients achieved very good partial response (VGPR), as described in Patients and Methods. One ATL lymphoma died from disease progression (Table 1). Interestingly, as previously reported in seven patients 30 , in the eleven patients for whom initial and Day 30 DNA was available, HTLV-I proviral load significantly decreased from an average of 1415 copy/μl of blood to 226 copy/μl (p<0.05). Most patients continued to improve their response and the best response was achieved within 2 to 4 months. Indeed, out of the 15 responding patients, 10 patients achieved CR, 2 patients achieved VGPR (solely because of the presence of 6% and 8% of atypical lymphocytes on peripheral blood smear, respectively), and three patients achieved PR (Table 1).

CR= complete remission; PR= partial response; VGPR= very good partial response; NR= No response; NA= not available.

We examined the effect of treatment on the relative distribution of the T cell subpopulations at day 30 (Table 2). Treatment with arsenic, AZT, and IFN decreased the average percentage of CD4⁺ cells (from 87 ± 14% to 77 ± 21%; p<0.01) and increased the average percentages of CD8⁺ cells (from 12 ± 12% to 16 ± 12%; p<0.01). In accordance with the observed response to therapy, a decrease in the average percentage of CD4⁺CD25⁺ cells was noted in all patients after treatment (from 47 ± 15% to 25 ± 14%; p<0.001), suggesting that this combination is mainly acting on the circulating malignant cells and HTLV-I infected cells (Table 2).

P values indicate the significance level for the pairwise comparisons between “Before” and “after”.

We investigated the early effect of the triple combination (arsenic/IFN/zidovudine) on the T_reg subpopulation by evaluating the transcript levels of Foxp3 and IL-10 in 10 healthy sero-negative blood donors, 10 asymptomatic HTLV-I healthy carriers, as well as in 16 ATL patients before and 30 days after treatment, before achievement of maximal response.

Untreated ATL patients displayed significantly higher transcript levels of Foxp3 (Figure 1A) and IL-10 (Figure 2A), relative to seronegative or healthy carrier individuals (p < 0.05). Notably, Foxp3 and IL-10 mRNA levels in untreated ATL patients were, respectively, at least 12 and 53 folds higher than seronegative and healthy carrier individuals (p < 0.05). High interindividual variation was observed among ATL patients for both Foxp3 and IL-10 transcript levels (Figures 1A and 2A, Table 3). Interestingly, treatment with arsenic/IFN/zidovudine significantly decreased transcript levels of both Foxp3 (Figure 1B) and IL-10 (Figure 2B). Because ATL cells express a T_reg phenotype, we normalized the expression levels of Foxp3 and IL-10 relative to the percentage of CD4⁺ CD25⁺ cells. ATL patients still showed significantly higher normalized transcript levels of Foxp3 and IL-10 relative to seronegative individuals (Figures 1C and 2C; p < 0.05). Normalized mRNA values of IL-10 (Figure 2D), but not Foxp3 (Figure 1D), significantly decreased in ATL patients after treatment (p < 0.05). Interestingly, standardized IL-10 transcript level negatively correlates with clinical response at Day 30 (r = −0.452, p < 0.05, spearman). This correlation remained valid even when adjusted for the confounder effects of gender and/or age. This suggests that the lower the IL-10 levels before therapy, the better the response after 30 days of therapy.

All values are expressed as percentage of human beta2-microglobulin used as internal control.

In order to confirm the decrease of IL-10 in treated patients, we assessed the secretion of this cytokine in the serum of patients by ELISA before and 30 days after treatment. Seven patients had significantly higher IL-10 serum levels relative to seronegative and healthy carrier individuals (Figure 2E; p < 0.001) and all displayed decreases in IL-10 serum levels after treatment (Figure 2F; p < 0.01). Interestingly, these seven patients are those who displayed the highest IL-10 mRNA expression before treatment (Figure 2A). IL-10 serum levels were low before and after treatment in the remaining patients (data not shown). These results suggest that treatment with arsenic/IFN/zidovudine decreases the proportion of ATL cells with T_reg phenotype as well as their IL-10 production.

We then assessed the effect of treatment on the Th2 subpopulation by evaluating the transcript levels of IL-4. ATL patients had significantly higher levels of IL-4 mRNA (Figure 3A; p < 0.05), but not of normalized IL-4 transcripts (Figure 3C), compared to healthy carriers and seronegative blood donors. Treatment with arsenic/IFN/zidovudine significantly decreased IL-4 and normalized IL-4 expression (Table 3 and Figure 3B and 3D; p < 0.05). Interestingly, the decrease in IL-4 expression was more pronounced in the patients who had the highest pretreatment transcript levels (Table 3 and Figure 3B and D).

We finally investigated the effect of arsenic/IFN/zidovudine treatment on the Th1 subpopulation by evaluating the transcript levels of IFN-γ and IL-2. ATL patients exhibited significantly lower mRNA levels of IFN-γ (Figure 4A) and normalized IFN-γ (Figure 4C), relative to seronegative or healthy carrier individuals (p < 0.05). Similarly, IL-2 (Figure 5A) and normalized IL-2 (Figure 5C) mRNA levels were lower in ATL patients, compared to seronegative individuals, with only normalized values being statistically significant (p < 0.001). Interestingly, treatment with arsenic/IFN/zidovudine significantly increased the expression and normalized transcript levels of IFN-γ (Table 3 and Figure 4B and D; p < 0.001) and IL-2 (Table 3 and Figure 5B and D; p < 0.01).

This study included 16 ATL patients. They were referred between 2006 and 2009 to the Hematology-Oncology Department of Ghaem and Imam Reza hospitals, Mashhad University of Medical Sciences, Iran. All ATL patients had serologic evidence of HTLV-I infection by Enzyme-Linked Immunosorbent Assay (ELISA). Confirmation of HTLV-I positivity was done by standard Polymerase Chain Reaction (PCR) (Data not shown). According to the Shimoyama classification criteria for ATL 18 , 12 patients had chronic ATL, 2 patients had acute ATL, while the 2 remaining patients had ATL lymphoma. The patient's characteristics are shown in Table 1. In addition to ATL patients, this study included 10 HTLV-I healthy carriers and 10 HTLV-I seronegative healthy volunteers prospectively selected from the Mashhad Blood Bank Center.

Treatment consisted of intravenous arsenic (10 mg/day, 5 days/week), subcutaneous IFN (Pooyesh Darou Pharmaceutical Co) (5 million units/day), and oral zidovudine (900 mg/day in 3 divided doses) and was administered to the patients for 30 days. Later on, patients received maintenance therapy with zidovudine and IFN. In case of toxicity, zidovudine and IFN were either transiently interrupted or their dose was reduced to 600 mg/day and 3 million units per day, respectively. Arsenic dose was not reduced in case of toxicity, but arsenic treatment was transiently interrupted.

This study was approved by the ethical committee of Mashhad University of Medical Sciences. Blood collection was performed on all patients and control subjects after signing informed consent forms in accordance with the Declaration of Helsinki. Peripheral blood mononuclear cells (PBMCs) and serum were obtained from all the patients before and after 30 days of treatment.

Complete remission (CR) was defined as a normalization of the CBC associated with a disappearance of all measurable tumors lasting at least one month. Patients with persistence of less than 5% atypical lymphocytes were also considered in CR as this situation may be seen in healthy carriers of HTLV-I. Very good partial response (VGPR) was defined as a normalization of the CBC associated with a disappearance of all measurable tumors lasting at least one month, but with persistence of more than 5% atypical lymphocytes on peripheral blood smear. Partial response (PR) was defined as a decrease of more than 50% in the number of leukemia cells and in the size of all measurable tumors. No response (NR) was defined as less than 50% decrease in the number of leukemia cells or in the size of any measurable tumor, or as disease progression.

Red blood cells were removed using a lysis solution (Becton Dickinson, San Diego). After fixation of white blood cells for 20 minutes in 0.5% paraformaldehyde solution, the cells were washed twice with PBS then immunostained with specific monoclonal antibodies and incubated in dark for 20 minutes. Monoclonal antibodies used for this study were: anti-CD4- peridinin chlorphyl protein (PerCP) , anti-CD3- phycoerythrin (R-PE), anti-CD25-fluorescein isothiocyanate (FITC) and anti-CD8-PE. All monoclonal antibodies (mAbs) were purchased from (IQ-product/Netherlands) and used at 1 in 10 dilution (10 μl of antibody in 100 μl of blood). Flow cytometry data was analysed using the CellQuest software (Becton Dickinson, San Diego). Results are presented as the relative fluorescence intensity and percentage of gated cell populations.

Total RNA was isolated from PBMCs using a TriPure Isolation Reagent (Roche Applied Science, Germany). cDNA was synthesized using a RevertAidTM H minus First Strand cDNA Synthesis Kit (Fermentas, Germany) and stored at −20°C until use. Primers were designed at exon-exon junctions (Beacon Designer http://WWW.premierbiosoft.com). All of the selected primers sequences were further analyzed with the Oligo software (http://WWW.cambio.co.uk/index.php). Beta 2 microglobulin (β2m) was used as internal housekeeping control gene to normalize the mRNA expression levels.

Quantification of Foxp3 and cytokine transcript expression was performed using real-time reverse transcriptase (RT) PCR in a rotor gene 6000, Corbett. Taqman method amplification was carried out for Foxp3 using the following primers and probes: Foxp3 sense: 5′ACTACTTCAAgTTCCACAACATgC-3′; Foxp3 anti sense: 5′gAgTgTCCgCTgCTTCTCTg-3′; Foxp3 probe: 5′TCACCTACgCCACgTTCATCCgCT3′; β2m sense: 5′CTTgTCTTTCAgCAAggACTgg-3′; β2m antisense: 5′CCACTTAACTATCTTgggCTgTg-3′; β2m probe: 5′TCACATggTTCACACggCAggCAT-3′. Thermal cycling conditions consisted of an initial step of 10 min at 95°C, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. Sybergreen amplification was carried out for interleukin-2 (IL-2), IL-4, IL-10, and IFN-γ using the following primers: IL-2 sense CTCACCAggATgCTCACATTTAAg; IL-2 antisense CTCCAgAggTTTgAgTTCTTCTTC; IL-4 sense: CACCgAgTTgACCgTAACAgAC; IL-4 antisense: CCCAggCAgCgAgTgTCC; IL-10 sense: TTgCTggAggACTTTAAgggTTAC; IL-10 antisense: CTTgATgTCTgggTCTTggTTCTC; IFN-γ sense: TgggTTCTCTTgg CTgTTACTg; IFN-γ antisense: gAgTTCCATTATCCgCTACATCTg; β2m sense: CTTgTCTTTCAgCAAggACTgg; β2m antisense: CCACTTAACTATCTTgggCTgTg. Sybergreen reactions were carried out in a final reaction volume of 20 μL using SYBR Premix EX Taq (RR041Q, TaKaRa). Thermal cycling conditions consisted of an initial step of 10 min at 95°C, followed by 40 cycles at 95°C for 10s, 60°C for 30s. For each run, a standard curve was generated using using a five-fold dilution series of a pooled cDNA for Foxp3, IFN-γ, IL-10, IL-2, IL-4 and β2m. The relative standard curves for both reference and target genes were generated by plotting the threshold cycle value versus the log of the dilution of the cDNA. PCR efficiencies (10^-1/slope^-1) was automatically calculated by the Rotor Gene Q system Software.

The HTLV-I viral copy number per μl of blood was calculated from the cell count and the average viral copy number per cell as assessed by quantitative PCR. Real-time quantitative PCR was performed on DNA extracted from peripheral blood mononuclear cells as previously described, using primers and Taqman probe positioned on tax gene and albumin gene for normalization 30 . TaqMan amplification was carried out in reaction volumes of 25 μL, with the use of the qPCR MasterMix (Eurogentec, Leuven, Belgium). Each sample was analyzed in triplicate with the use of 250 ng of DNA in each reaction. Thermal cycling was initiated with a 2-minutes incubation at 50°C, followed by a first denaturation step of 10 minutes at 95°C and then by 45 cycles at 95°C for 15 seconds and 58°C for 1 minute for tax (or 60°C for 1 minute for albumin).

Peripheral blood samples were transferred to serum-separating tubes and centrifuged at 1000g for 20 min after clot formation. The supernatants were carefully harvested, and immediately stored at −80°C until analysis. The serum IL-10 concentration was measured in duplicate by enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s instructions (R&D Systems, MN, USA). Absorbance of each well was measured spectrophotometrically at 450 nm. The amount of IL-10 protein in the samples was calculated using a reference plot established from serial dilutions of rh IL-10 protein as provided.

SPSS Version 16.0 and Microsoft Office Excel 2010 were used for statistical analyses. In scatter plots, means were compared using ANOVA with associated post-hoc tests: Dunnett t, Tukey, Student-Newman-Keuls (SNK), and Bonferroni tests. All analyzed samples fit normal distributions, except IFN-γ and IL-10 values in ATL patients. Therefore, both mean and median were plotted for ATL patient data, as shown in line diagrams, and were analyzed using parametric and nonparametric (Wilcoxon) paired sample t-tests, respectively, to compare values before and after treatment. In all cases, mean and median either both showed statistical significance or both not. For normally distributed data, pearson correlation was performed. Otherwise, spearman correlation was used and compared to pearson’s rho tested on log-transformed data, with similar results. Spearman rho was employed for correlations involving ordinal scale data. Statistical significance was reported when the P-value was ≤ 0.05, except in scatter plots comparing ATL patients before and after treatment, for which P-values were stated separately in line diagrams. *, P < 0.05; **, P < 0.01; ***, P < 0.001.