Background
The HTLV-1 virus is the etiological agent of two major diseases: adult T cell leukemia and the neurological disease HTLV-associated myelopathy / tropical spastic paraparesis (HAM/TSP), a progressive neurological and inflammatory disease of the central nervous system. Although many of the infected individuals are asymptomatic, approximately 2–5% of the infected individuals will develop HAM/TSP [
1]. This neurological disease could be the consequence of an inflammatory network that results in damage of the spinal cord [
2], but this association is still poorly understood.
Tax viral protein plays an important role in the regulation of the virus genome acting in proviral genome transcription by interacting with several cellular signaling pathways that modulate the expression of cytokine and chemokine genes [
3]. The IFN-γ secreted by HTLV-1 infected CD4
+T cells and the virus recognition by CD8
+T lymphocytes in the central nervous system induces production of other cytokines, such as the myelinotoxic TNF-α. This event, together with the hemodynamic changes and interactions mediated by adhesion molecules among circulating lymphocytes and endothelial cells, which contribute to the location of spinal cord injury, is known as a “bystander” damage hypothesis. The involvement of cytokines in the outcome of HAM/TSP is associated with the “bystander” damage hypothesis [
1]. Among the potential immunopathological findings, high levels of IFN-γ, TNF-α and IL-6 have been detected in patients with HAM/TSP, which seem to contribute for the breakthrough of the blood–brain barrier and resulting in immunopathology and neurological symptoms [
4]. In addition, HAM/TSP patients showed dysregulation in TGF-β signaling, affecting Treg function and contributing for disease pathology [
5]. Differences in IL-2, IL-4, IL-10, IL-12p70, TNF-α and IFN-γ levels were also found in the supernatants of cultured peripheral blood mononuclear cells (PBMC) from HAM/TSP and AC [
6]. Recent publication showed no difference in plasma cytokine levels among AC and HAM/TSP patients. However, cerebrospinal fluid levels of cytokines (ITAC, IFN-γ, IL-5, IL-8 and TNF-α) were higher in HAM/TSP compared with AC patients, indicating that those cytokines might be used as disease markers of neurologic manifestation in long-term HTLV-1 infected individuals [
7].
Single nucleotide polymorphisms (SNPs) might have influence in the cytokine production. It has been shown that
IFNG + 874 T/A [
8],
TNF -308G/A [
9,
10],
IL6 -174G/C [
11],
IL10–1082A/G,
-819C/T and
-592C/A [
12], and
TGFB at the codons
+ 10 T/C and
+ 25G/C [
13,
14] SNPs alter the expression of the related cytokines in pathological or physiological conditions. Many of these polymorphisms have been associated with inflammatory and infectious diseases [
15‐
19]. Indeed,
IFNG + 874A/T [
20],
IL6 -174G/C [
21], and
IL10 -592C/A and -819C/T [
22] have been associated with HAM/TSP development or HTLV-1 infection.
Because these specific polymorphisms are associated with changes in cytokines production and many inflammatory and infectious diseases, and those cytokines have important role in HTLV-1 infection, it would be important evaluate the association between these polymorphisms and HAM/TSP development. Studies addressing the role of SNPs that might alter cytokine production and HTLV-1 infection are scarce. Therefore, our aim was to determine in a case-control study if SNPs of pro- and anti-inflammatory cytokines in HAM/TSP and AC patients were related to disease outcome.
Methods
Study population
HAM/TSP patients and ACs were randomly recruited from a cohort of approximately 700 individuals attended at the Laboratory for Clinical Research in Neuroinfection, INI-FIOCRUZ, Rio de Janeiro, RJ, Brazil. The diagnosis of HAM/TSP patients was performed according to the World Health Organization diagnostic criteria [
23]. The patients were mainly from Rio de Janeiro, Brazil, and the follow-up mean was 128.85 ± 54.44 months. The Institutional Ethics Committee approved the study, and all subjects provided written informed consent. Demographic data included information regarding a self-identified skin color as described as white or non-white (black and mixed persons), sex and age.
Five milliliters of whole blood were collected in EDTA-containing tubes, and DNA extraction was carried out by a commercial kit from Puregene (Gentra Systems Inc., Minneapolis, MN, USA), following the manufacturer’s instructions.
Genotyping
SNPs in five cytokine genes, TNFA-308G/A, TGFB at codons + 10 T/C and TGFB + 25G/C, IL10-1082A/G, -592C/A and -819C/T, IL6-174G/C, IFNG + 874 T/A were genotyped through sequence-specific primers-polymerase chain reaction technique (SSP-PCR) using a commercial kit by Cytokine Genotyping Tray (One Lambda, Inc., Canoga Park, CA, USA). This technique has allowed us genotyped eight SNPs of five cytokine genes, simultaneously, through PCR reaction. The SSP-PCR technique provides an accurate, simple, and economical means of detecting polymorphisms of these important pro- and anti-inflammatory cytokine genes.
Ninety-six well microtiter® trays with dried primers in each reaction well were filled with 10 μl of a mix containing dNTPs (8 μl, provided by the manufacturer), Taq polymerase (1 μl at 5 U/μl) and DNA sample (1 μl, 100 ng) per reaction well. The reaction was performed in a Veriti thermocycler (Applied Biosystems, Foster City, CA, USA) with the follow instructions: one cycle of 130 s at 96 °C, 60s at 63 °C; nine cycles of 10s at 96 °C, 60s at 63 °C; twenty cycles of 10s at 96 °C, 50s at 59 °C and 20s at 72 °C; and ending at 4 °C.
After completing the PCR reaction, the samples were transferred to 2.5% agarose gel stained with ethidium bromide (0.5 μg/ml), and were electrophoresed at 140–150 V. After electrophoretic run, the gel was transferred for an ultraviolet transilluminator and photographed for analysis. The presence or absence of gel band from each amplified reaction well determined the SNP identification.
Phenotype analysis for IL10-1082A/G, -819C/T and -592C/A SNPs were demonstrated as low (ACC/ACC, ACC/ATA, ATA/ATA), intermediate (GCC/ACC, GCC/ATA) and high (GCC/GCC) IL-10 producers, as described by the commercial kit.
Proviral load quantification
HTLV-1 PVL DNA from peripheral blood leucocytes was measured by real-time PCR assay (SmartCycle II; Cepheid) using the TaqMan system (Applied Biosystems, Foster City, California, USA), through the amplification of a 159-bp fragment of the
tax gene. As a reference, a standard curve was generated using the human
β-globin gene and DNA from the TARL-2 cell line, which contains a single copy of the provirus HTLV-1, to establish the
tax gene quantification. PVL was calculated in peripheral blood leukocytes (PBL) from each volunteer following the equation: [(copy number of
tax gene) / (copy number of β-globin gene/2)] × 100 [
24].
Statistical analysis
Proviral load statistical analysis between genotypes/phenotypes and disease groups was performed by an ordinary two-way ANOVA with Bonferroni post-test for multiple comparisons correction [
25], considering results with
p-values < 0.05 statistically significant. Nonparametric data were evaluated by Kruskal-Wallis and Mann-Whitney tests and a Spearman correlation, and association between qualitative variables (alleles, genotypes and phenotypes frequencies) was tested by chi-square, Cochran-Armitage or Fisher’s exact test.
P-values from tables with any significant result were corrected using Bonferroni correction for multiple comparisons. Results with a
p-value < 0.05 were considered statistically significant. All analyses were performed using the software GraphPad Prism 6.0 or R version 3.1.0 [
26].
Discussion
HAM/TSP could be the consequence of an inflammatory response of the host immune system; however, the majority of HTLV-1-infected individuals remain asymptomatic, indicating that HTLV-1 infection by itself is not sufficient to induce HAM/TSP. Neurons and glial cells are damaged by toxic or inflammatory products released from HTLV-1-infected T cells and the bystander damage by activated cytotoxic lymphocytes in the central nervous system [
27]. Cytokines production varies among individuals, due in part to genetic factors and, in particular, the presence of polymorphisms in important regulatory regions, such as promoter regions. In this study, we aimed to establish an association between polymorphisms in cytokine-related genes (
TNFA, IL6, IFNG, TGFB and
IL10) and disease outcome. Our results demonstrated that there was no association between
TNFA-308G/A,
IL6-174G/C,
IFNG + 874 T/A,
TGFB at codons
+ 10 T/C and
+ 25G/C,
and IL10-1082G/A, -819C/T and
-592C/A polymorphisms and development of HAM/TSP in our set of the Brazilian population. To the best of our knowledge, few studies addressed information on cytokine SNPs and HTLV-1 infection in the Brazilian population.
In vitro studies have shown that HTLV-1 infected lymphocytes transpose the blood-brain barrier, changing its permeability by secretion of IL-1α and TNF-α, which increases the migration of lymphocytes through this site [
28]. This migration could enhance the inflammatory response in the nervous system, causing damage in neural cells in vivo. Polymorphism in the promoter region of the
TNFA-308G/A was described by Wilson et al. 1992 [
29], where the replacement of the allele guanine
TNFA (−308G or
TNF1) for adenine in the
TNFA allele
(−308A or
TNF2) has been associated with high production of this cytokine in European populations [
29]. The
TNFA-308G/A polymorphism was analyzed in adult T-cell leukemia/lymphoma versus AC in Japanese population, and no association between this SNP and disease outcome was found [
20]. In our set of individuals typed for
TNFA-308G/A polymorphism, we also did not find an association between this SNP and the HAM/TSP or AC groups. However, an association between the genotype
TNFA-308GG (low producer) and high PVL was observed. Starling et al. 2013 showed an inverse correlation between TNF-α production and PVL in HAM/TSP patients [
30], corroborating our results and indicating a possible protective role of TNF-α in the control of pro viral load. Our data suggest a possible effect of
TNFA-308GG genotype in PVL not related to disease outcome.
The
IL6 -174G/C is associated with changes in cytokine production, where G allele carriers show higher production of IL-6 compared to C allele carriers [
11]. Gadelha and colleagues [
21] described no association between HAM/TSP and AC in
IL6 -174G/C in a Northeast Brazilian population, although this SNP was a risk factor for HAM/TSP when compared with oligosymptomatic patients. Although none of the
IL-6 genetic background was associated with HAM/TSP outcome, we observed an increased level of PVL in HAM/TSP patients carrying the higher IL-6 producer (
IL6-174GG and CC) genotypes, indicating that a high level of this inflammatory cytokine could be related to a worse scenario in HAM/TSP patients. In addition, IL-6 and TGF-β may induce the differentiation of a Th17 cell profile and increase central nervous system inflammation [
31].
In HAM/TSP patients, there is a predominance of Th1 cytokines such as IFN-γ, reduction in Th2 cytokines such as IL-4 and IL-10, and increased production of neurotoxic cytokines affecting regions along the spinal cord [
32]. The IFN-γ serum concentration is higher in HAM/TSP individuals than in asymptomatic individuals [
6,
30]. Located in the first intron of the
IFNG gene, the SNP + 874
T/A contributes to IFN-γ production, with T allele carriers presenting higher production of IFN-γ compared to A allele carriers [
8]. Recently, Queiroz et al. 2018 showed higher prevalence of the
IFNG + 874 T/A T allele among AC compared to HTLV-1 symptomatic carriers, although no differences in genotypes distribution were detected [
20]. Queiroz et al. data agree with our results regard genotype distribution and PVL, differently of allelic distribution, where we did not find differences. This discordance might be due to different study approaches. Queiroz et al. enrolled in their study, HTLV-1 symptomatic carriers with diverse clinical manifestations, such as rheumatism, condition absent in our study. In addition, the patients of that study belonged to North region of Brazil, where differences in migration history and genetics background might occur compared to our study. Our results, in agreement with a previous study [
33], demonstrated no significant differences for allele and genotype frequencies of the
IFNG + 874
T/A SNP between HAM and AC. In our set of the Brazilian population, there was no statistical influence on any genotypes of
IFNG + 874 T/A and PVL in the HAM/TSP or AC groups. However, our results were in discordance with those published by Rocha-Junior et al. [
33], who reported a significant association between PVL and the
IFNG + 874AA genotype, where this SNP was associated with low PVL. In addition to the two populations belonging to Brazil, the study also had a group included by Rocha-Junior that was from São Paulo State, where high Asiatic frequency is present; our study groups did not include any Asiatic individuals.
It has been shown that TGF- β signaling is critical for Foxp3 expression and T regulatory cells functions in HAM/TSP patients [
34]. HAM/TSP patients showed low levels of TGF-β receptor II (TGF-βRII) and Smad7 (a TGF-β–inducible gene) in CD4+ T cells, when compared to healthy donors. In addition, TGF-βRII expression was inversely correlated with the HTLV-1 PVL. This evidence suggests that HTLV-1 can modulate the immune tolerance affecting both regulatory and effector T cells and contributing to the pathogenesis of HAM/TSP [
5]. HTLV-1 infection in patients with Sjogren’s Syndrome led to an enhanced serum level of TGF-β and may be important for increased HTLV PVL [
35]. The haplotype TG from the SNPs of
TGFB at the codons
+ 10 T/C and
+ 25G/C has been associated with high production of this cytokine [
13]. We did not observe association with these SNPs and HAM/TSP development. However, in our study, we observed higher PVL in HAM/TSP patients TGF-β high producers (+10TT or TC, and + 25GG) compared to AC, suggesting a role of this cytokine in control of the viral replication and disease prognosis. Additional factors might affect the expression of TGF-β during the HTLV-1 infection, changing the gene expression levels regardless the polymorphisms influence [
5,
36,
37]. To the best of our knowledge, there is no previous work describing the
TGFB SNP in HAM/TSP patients.
HTLV-1 infection in HAM/TSP patients does not alter the expression of IL-10 when compared with AC patients [
38]. SNPs in the
IL10 gene are controversially associated with the HAM/TSP outcome. A Japanese population showed an association between the
IL10 -592A allele and a protective effect, reducing risk of HAM/TSP [
39]; however, this SNP was considered to be a risk factor for developing HTLV-1 infection and disease in an Iranian population [
22]. No association between
IL10 SNPs and HAM/TSP outcome was observed in a Brazilian population [
21] in a high-risk HTLV-1 prevalence area with 9.4/1000 habitants [
40]. In our set of Brazilian population enrolled from Rio de Janeiro City, an intermediate HTLV-1 prevalence risk area, with 4.7/1000 habitants [
40] there was no association with the disease outcome. At position
IL10–1082A/G, an Iranian population showed no association with HAM/TSP in a co-dominant genetic model [
22]. There were no previous reports describing the SNP in this position in Brazilian HTLV-1-infected individuals. Our results showed a predominance of intermediate producers of
IL10-1082AG and
GG genotypes among the HAM/TSP individuals (OR 2.42 [1.2–4.97]
p = 0.01), predisposing to disease outcome, although the data lost significance after Bonferroni correction for multiple comparisons. Low-producer haplotypes (
IL10-1082A, − 819 T and
-592A) have been previously associated with AC or HAM/TSP compared with healthy controls [
40]; however, no association was described between AC and HAM/TSP groups, as described in this study. Our results demonstrated that only
IL10–819CC and
-592AA carriers, presenting high levels of IL10 production, have the highest PVL when the HAM/TSP group was compared with the AC group.
The infection caused by HTLV-1 elicits a robust immune response with many factors affecting the cytokines gene expression, such as genetic population’s background, virus subtypes, immunomodulation and individual health status. Indeed, it has been shown that different subtypes of HTLV-1 are region restricted, besides the genetic background of ethnicities across the world [
41]. This fact could explain association between polymorphisms of cytokine genes, such as
IL10 -592C/A and
IL10–1082A/G, and HTLV-1 infection in some specific populations. On the other hand, some polymorphisms, such as
TNF -308G/A, IFNG + 874 T/A, IL6 -174G/C, are repeatedly not associated with the disease, regardless genetic background population, suggesting irrelevant role in the HTLV-1 infection. However, more studies in different populations are needed to confirm this hypothesis. Despite no association with disease development, some of these polymorphisms in cytokine genes, such as
IFNG + 874 T/A, TNF -308G/A, IL6 -174G/C, TGFB + 10 T/C and + 25G/C, IL10–819C/T and
-592C/A might change PVL, indicating a role of this polymorphisms in the controlling of the viral replication and disease prognosis.
Our study might be analyzed considering its limitations such as the number of patients enrolled and the unavailability to measure cytokines concentrations in plasma or cultured cells in the patients. As we used convenience samples, we had limitations regard sample size and biological material to perform the analysis. Despite these polymorphisms are well established to modulate the cytokines production, is not clear its role on the HTLV-1 infection. Future studies addressing the role of these polymorphisms in different populations, associated with the cytokines production, might clarify the role of these functional SNPs in the HTLV-1 infection.