Background
Infection by the Human Immunodeficiency Virus (HIV) has been considered a serious infectious disease particularly in Southern Africa, which harbors more than 2/3 of all worldwide cases of HIV [
1].
The emergence of several co-pathogens has aggravated this scenario in resource-limited settings [
2,
3]. Human T-lymphotropic virus type 1 (HTLV-1) has been implicated as a frequent co-pathogen in areas or groups where both viruses are prevalent [
4,
5]. In the southern Africa region where HIV is highly prevalent [
6], the prevalence of co-infection by HTLV-1 and HIV varies among countries and in several places have been reported be higher than 10% [
7‐
9]. A recent study conducted in Mozambique reported a co-infection prevalence rate of 4.5% among HAART naïve HIV positive individuals. Thus, it is conceivable that the impact of chronic infection by HTLV-1 on HIV disease progression is a relevant issue in AIDS research. Nevertheless, such an issue remains controversial, and published data are conflicting[
10‐
12]. If in one hand it was initially postulated that co-infected patients progress faster to AIDS[
13], further studies reported contradictory results[
10,
14].
The influence of HTLV-1 on HIV disease progression has been tightly linked, not only to several molecular events [
15,
16] but to its potential to induce high levels immune activation [
17]. Although the mechanisms by which chronic activation induced by HTLV-1 could potentially affect the progression to AIDS are not completely understood, we have learned from the natural history of HIV infection that chronic activation of the immune system takes part and triggers a number of cellular and molecular pathways related to CD4
+ T cell loss and immune deregulation [
18‐
22]. In this context, HTLV-1 per se induces a strong immune activation [
23,
24] that has been associated to immunosuppression, unresponsiveness and immune deregulation [
25,
26]. How the immune system behaves in the presence of both HIV and HTLV-1 remains to be clarified.
The concern regarding the clinical outcome as a result of co-infection by HTLV-1 and HIV has gained a special relevance in recent years, in face of the growing body of evidence showing that: a) co-infection is prevalent in several geographical regions in Southern Africa [
27,
28]; b) we and others showed that co-infected patients present stable CD4
+ T lymphocytosis irrespective of their progression to AIDS, which by certain extent could mask the immunosuppression with consequent inappropriate therapeutic decisions in terms of the initiation of Highly Active Anti Retroviral Therapy (HAART) and prophylaxis for opportunistic infections [
29,
30].
The situation has being aggravated by the fact that neither cure nor effective treatment is yet available for HTLV infection [
31] in such a way that clinicians are unable to control the effects exerted by the virus.
Previous studies have enrolled patients with different HAART experiences and some authors believe that this may partially explain the divergence of results obtained [
11]. Other studies were based on
in vitro manipulation [
15] or used simian models [
32]. In addition, no study has been so far conducted in Africa, the region carrying the greatest burden of HIV disease and where the epidemiology of HIV and other diseases is quite different [
1].
Altogether, these data raise the need to define to what extent HTLV-1 impacts the clinical progression to AIDS in an African setting. We evaluated herein HTLV-1 and HIV co-infected HAART naïve adult patients, in terms of T cell phenotype, further correlating the expression of activation markers and HIV-1 viral load. We showed that co-infected patients progress with higher levels of CD4+ T cells expressing activation markers and a massive loss of naïve cells, thus suggesting that co-infected patients progress faster to AIDS.
Methods
Study design and subjects
A case control study was conducted with participants consisting of three sets, namely, individuals co-infected by HIV and HTLV-1 (Co-infected), HIV mono-infected patients (HIV) and healthy controls (HC). Co-infected and HIV groups were recruited from an ongoing cohort of HIV infected patients followed in the Alto Maé Health Centre, in the city of Maputo, Mozambique. In the period between March and June 2006, 724 HIV 1/2 infected patients were invited to participate and 704 (97.4%) accepted to be part of this study. They were all screened for HTLV-1 infection and 32 patients (4.5%, 32/704) were founded to be co-infected by HTLV-1 and HIV. Three patients with a positive HTLV-1 antibody test did not return to collect their result and were excluded of the study. Co-infected were matched at a ratio 1:2 with HIV mono-infected by age, sex and HIV clinical stage system as defined by WHO [
33]. Co-infected and HIV mono-infected were matched without prior knowledge of CD4+T cell counts results.
Healthy controls were not matched by age and sex as we did with co-infected and HIV-mono-infected, since they were recruited on a consecutive basis from the routine blood donors at the blood bank of Maputo Central Hospital. In addition, most of blood donors are males and younger as demonstrated by two previous studies conducted at the same Blood Bank ([
34,
35]) Informed consent to participate in the study was requested to all participants, and the study was approved by the National Bioethics Committee in Mozambique and by the Sydney University Ethics Committee, Australia.
Physical and neurological examination was performed by two medical doctors blinded for the HTLV-1 status. Socio-demographic data, sexual/reproductive history and clinical data were also recorded from each participant.
The study population consisted of 59 HIV, 29 co-infected and 16 healthy controls individuals. All HIV and co-infected individuals were naïve to HAART. The median age was 40 years (IQR, 34 - 48 years) for co-infected individuals, 41 years (IQR, 32 - 47) for HIV patients and 32 (IQR, 29 - 38) for HC, respectively. There was a predominance of the female gender among HIV and co-infected individuals (86,3% and 86.4% respectively) and predominance of the male gender among HC (80%; see table
1). Almost half of co-infected and HIV patients were classified as stage 2 for HIV clinical disease (58.4% and 56.0% for co-infected and HIV patients respectively) based on the WHO criteria [
33]. None of the co-infected or HIV patients was at stage 4 of HIV clinical disease (table
1). HAM/TSP or ATL cases were not detected in subjects of any groups.
Table 1
General characteristics of HIV/HTLV-1 co-infected, HIV mono infected and healthy Mozambican subjects
Median | 40.0 | 41.0 | 32.0 | |
IQR | 34.0 - 48.0 | 32.0 - 47.0 | 29.0 - 38.0 | 0.123* |
Gender
| | | | |
Male | 4 (13.8%) | 8 (13.6%) | 12 (80%) | |
Female | 25 (86.3%) | 51 (86.4%) | 3 (20%) | 0.000** |
HIV clinical stage***
| | | | |
I | 7 (24.1%) | 14 (23.7%) | | |
II | 17 (58.4%) | 33 (56.0%) | | |
III | 5 (17.2%) | 12 (20.3%) | | 0.587** |
Co-infected and mono-infected were comparable regarding clinical presentation of opportunistic diseases (table
2).
Table 2
Clinical presentation of study groups
Assymptomatic
| 7 (24.1) | 14 (23.7) | |
Papular pruritic eruptions
| 3 (10.3) | 14 (23.7) | |
Dermatitis
| 2 (6.9) | 2 (3.4) | |
Seborrhoeic dermatitis
| 2 (6.9) | 0 (0.0) | |
Folliculitis
| 0 (0.0) | 1 (1.7) | |
Herpes zoster < 5 years
| 3 (10.3) | 4 (6.8) | 0.570 |
TInea capitis
| 5 (17.2) | 8 (13.6) | |
Candidiasis
| 3 (10.3) | 4 (6.8) | |
Tuberculosis < 1 year
| 1 (3.5) | 4 (6.8) | |
Weight loss > 10%
| 3 (10.3) | 7 (11.9) | |
Chronic diarrhoea
| 0 (0.0) | 1 (1.7) | |
Blood samples
Ten milliliters of venous whole blood were requested from each volunteer. The blood was collected aseptically into a 5 ml vacuum tube with K3EDTA and a 5 ml vacuum tube for serum separation (Becton-Dickinson Vacutainer Systems, USA). Blood specimens were delivered at the laboratory within four hours of collection.
HIV Serology
All patients enrolled in this study were screened for anti-HIV 1/2 antibodies at the Voluntary Counseling and Testing (VCT) services of the Mozambican Health Center. Patients were tested for HIV according to the Mozambican National protocol consisting of a sequential algorithm of two immunochromatographic rapid tests. All individuals were first screened using the Determine HIV-1/2 test (Abbott Laboratories, Japan). All specimens reactive on the screening assay were further tested using the Uni-Gold HIV test (Trinity Biotech, Ireland). Individuals reactive on both assays were considered positive for HIV-1/2 infection.
HTLV Serology
All samples were screened for anti-HTLV-1+2 antibodies using the qualitative EIA Murex HTLV-1 + 2 (Murex Biotech Limited, UK). Specimens reactive on the EIA were confirmed by a Western blot assay (HTLV BLOT 2.4, Genelabs® Diagnostics, Switzerland). All patients with reactivity to antigens encoded by the GAG gene (p19 with or without p24) and to two antigens encoded by the ENV gene (GD21 and rgp46-I) were considered to be infected by HTLV-1 according to the instructions provided by the manufacturer. All HTLV positive samples in our study population were typed as HTLV-1 by Western blot.
T cell immunophenotyping
Phenotypic analysis of circulating T lymphocytes was performed through four colour flow cytometry on fresh EDTA-anticoagulated whole blood using a FACSCalibur™ flow cytometer (Becton-Dickinson Biosciences, USA). T cells counts were obtained based on a lyse-no wash protocol using CellQuest Software for acquisition (BD Biosciences). Cells were stained with the following fluorescent-labeled anti-human monoclonal antibodies CD4-APC, CD8-PerCP, CD25-PE, CD62L-FITC, CD45RA-FITC. CD45RO-PE, CD38-PE (all from BD Biosciences). Analyses were performed using the Summit software v 4.3.2, 2006 (Dako Cytomation, Inc., Fort Collins, USA).
HIV-1 subtyping
Samples from co-infected and mono-infected individuals were sequenced for HIV-1 subtyping. A 297-bp fragment encompassing full sequence of the protease gene was amplified using nested PCR. The outer primers PRT15F 5' TGAAAGATTGTACTGAGAGACAG
G 3'/K2R 5' GTCAATGACATACAGAAGTTAGTGGGAAAA 3' were used in first-round PCR, and DP10F 5' CAACTCCCTATCAGAAGCAGGAGAAG 3'/RVP3R 5'-CCATACAATACTCCAGTATTTGCC-3' were used as inner primers during the second-round of nested PCR. The conditions for both rounds of PCR were described previously [
36]. Amplified DNA was quantified, purified and sequenced under PCR conditions described elsewhere [
37], using the primers from the second-round PCR. The genetic subtypes were determined using the amino acid sequences of the protease genes, deduced from the nucleic acid sequences, and were compared to a subtype B consensus sequence from the Stanford HIV Protease Sequence database
http://hivdb.stanford.edu/hiv/.
The analysis of the Mozambique sequence with HIV Drug Resistance database showed that all samples sequenced (20 mono-infected and 24 co-infected) were subtype C in protease gene with profile of similarity of 98% compared with sequence subtype C from the Stanford HIV Protease Sequence database.
Stools were collected into a sterile, wide mouth, leak-proof container with a tight fitting lid containing a preservative solution. Stools were kept at room temperature until delivery at the study setting within six hours of collection. Small amounts of stool specimens were placed on microscope slides and mixed with 0.9% NaCl solution to prepare wet mount smears. Slides were then examined under a light microscope to screen for cysts of Giardia lamblia, Entamoeba hystolitica, Entamoeba coli and Balantidium coli, ova of Ascaris lumbricoides, Trichuris trichiura and Ancylostoma duodenale and larvae of Strongyloides stercolaris.
Statistical analysis
Data was analyzed using the statistics package STATA 9.0 (College Station, Texas: StataCorp, USA, 2005).
Taking into consideration that the major goal of our study was to compare activation markers between co-infected versus HIV mono-infected patients, two sample comparison means with a ratio 1:2 (case:control) was used to determine the required sample size in these groups. Due to lack of information regarding comparison of activation markers between these groups, we calculated a sample size enough to detect a least a difference of 10% in the mean of these cells frequencies with a standard deviation of 14 at a significance level of 5%.
The Mann-Whitney test and the One Way Anova trend test were used to compare the differences among numerical variables in the three groups. Associations between categorical variables were determined using the Pearson Chi-square test, the Fisher exact test or the Chi-square trend test, as appropriated.
Discussion
HIV and HTLV-1 have emerged as common co-pathogens especially in areas or groups where both viruses are circulating [
5,
9,
38,
39]. Nevertheless, the impact of HTLV-1 on HIV disease progression is still a matter of debate with controversial results [
10‐
14,
30,
40]. Here, for the first time we conducted a case control study in an African setting, aiming to determine the impact of HTLV-1 infection on HIV disease. In fact, to our knowledge, the present study is pioneer in the region, since it was conducted in HAART naïve patients, on a well controlled cross-sectional basis.
Previous studies were conducted mainly in South and North Americas where the epidemiology of HIV infection and other diseases is quite different from that seen in sub-Saharan African countries [
29,
30].
We found no evidence of HTLV-2 in our study population. This is in keeping with two recent studies conducted among blood donors in Maputo city[
34,
35] and suggest that only HTLV-1 (but not HTLV-2) circulates in Mozambique.
As expected, co-infected individuals presented a stable CD4
+ T lymphocytosis irrespective of their progression to AIDS, contrasting with the depletion of CD4
+ T cells counts observed among HIV patients over time. To date, it is well established that cell immortalization and transformation induced by Tax and Rex proteins encoded by HTLV-1 genes constitute major events related to uncontrolled CD4
+ T cell growth and proliferation [
41,
42].
The intriguing progression to AIDS in the presence of normal or high levels of CD4+T cells counts suggest these to be functionally altered. Consensus exists that both HIV [
43,
44] and HTLV [
24,
25,
45] separately induce functional modifications on T cells populations, characterized among others by a decrease of naïve populations and higher levels of cell activation when compared with uninfected individuals.
Here we found that co-infected individuals presented markedly lower expression of CD45RA
+ (a phenotypic marker of naïve T lymphocytes) on CD4
+ T cells. Naïve cells are considered the first cells to be depleted in the presence of immune activation [
46,
47] and represent one of the hallmarks of HIV infection [
48]. The magnitude and impact exerted by naïve T cells erosion on HIV disease progression remain to be defined. Although not fully understood, there is a consensus that for both HIV and HTLV-1, the loss of naïve cells has been linked, among others, to, (i) a homeostatic mechanism to replenish the cells being killed (ii) a massive recruitment of naïve cells, partially imposed by the mechanisms driving the activation and (iii) the impairment of T cell production [
19,
20,
46‐
49].
In our study, the erosion of the naïve compartment was further confirmed by evaluating the expression of CD62L, another marker for naïve T lymphocytes, usually lost upon activation. As expected, there was a dramatic loss of CD4+CD62L+ lymphocytes in the co-infected group, when compared to HIV mono-infected and HC groups. Importantly, these differences were further confirmed when we compared the groups in terms of naive cell absolute counts (data not shown) arguing against an indirect effect of higher percent counts of memory cells. Whether there is an impairment of T cell production, if they are dying faster or if more cells being recruited from the naïve T cell pool into activated/memory cell compartment remain to be determined.
Not surprisingly, this loss of naïve cells in co-infected individuals was accompanied by higher frequencies of memory and activated cells as measured by CD45RO
+ (memory), CD38
+ and CD25
+ (activated) cell markers. In fact, co-infected individuals presented with higher proportions of CD45RO
+ on CD4
+ T cells when compared to the HIV and HC groups. These findings are in agreement with previously data [
45,
46,
50,
51], Similarly, the relative number of CD4
+CD25
+ cells seen in co-infected patients was higher than what was found in HIV and HC individuals. It is conceivable that the increase of CD4
+CD25
+ cells is a consequence of the virus-driven induction of IL-2/IL-2 receptor expression by tax, as previously reported [
12,
23,
52]. Interestingly, the frequency of CD38
+cells within the CD8
+ T cell compartment but not in CD4
+ T cells was increased in co-infected and HIV when compared to HC. This is in keeping with the results showed in a case-control study conducted among HAM/TSP patients [
24]. Although we did not find differences in the frequency of CD38
+ cells, either in CD8
+ or CD4
+ T cells, we found that co-infected patients presented higher expression of CD38 in CD8
+ T cells (as ascertained by Median Fluorescence Intensity measurements) when compared to HIV patients. Nevertheless differences were not statically significant.
Noteworthy, increased expression of CD38 on the surface of CD8
+ T cells have long been considered an even better prognostic predictor of progression to AIDS and response to HAART than HIV viral load itself [
53,
54]. This is relevant due to the fact that such parameter is being proposed to be included in clinical settings to monitor HIV disease progression [
55].
It is now widely accepted that the presence of chronic activation is a major factor influencing the pathogenesis of HIV in Africa [
56]. HTLV-1 is a strong activator of immune system. Immune activation and exaggerated immune response has been demonstrated to be the main pathogenetic mechanism involved in the HTLV-1 associated inflammatory syndromes[
24,
57‐
59]. The immunodominant Tax protein encoded by HTLV transactivates and modulates a large number of genes playing a key role in triggering several pathways leading to cell activation[
60‐
62]. Available data demonstrate that a large proportion of asymptomatic carries progress with high levels of immune activation[
63].
On the basis of the patients' age and their HAART naïve status, we believe that HTLV-1 infection preceded HIV infection. Considering that individuals chronically infected by HTLV progress with immune activation, it is conceivable that these patients acquire HIV infection in a pre-activated immune milieu, and the presence of immune hyper activation not only turns them more susceptible to acquire HIV, but also to progress faster to a poor prognosis.
Cases and controls were matched by age and clinical stage (WHO) so that to be comparable in terms of clinical presentation (see table
2), Clinical staging system is performed on the basis of patient's clinical presentation. This information is important when interpreting the differences in the activation markers between these groups. Another aspect deserving discussion is the helminthic infection as a factor involved in immune activation, particularly in Southern Africa [
3,
64]. Accordingly, a differential presence of parasitic infection in our patients could bias our results. However, this does not seem to be the case since in all groups evaluated, the degree of protozoan and helminthic infections were similar. Of note, all samples sequenced in both groups were founded to be HIV subtype C, ruling out any linkage between HIV subtype in mono and co-infected groups, and immunological/clinical behavior.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ESG participated in the study design, data collection, data analysis, and writing the manuscript. NBB participated in the study design, data collection, data analysis and writing the manuscript. DRB participated in the study design, sample processing and data analysis. WiS participated in the study design, data analysis and writing the manuscript. SDSB participated in the study design, data analysis and writing the manuscript. CMA and AT participated in HIV subtyping and writing the manuscript. IVJ participated in the study design, data collection, data analysis and writing the manuscript. None of the authors have any financial or personal interest in the World Bank Quick Impact Fund. All authors have read and approved the final manuscript