Association between major depressive disorder and pro-inflammatory cytokines and acute phase proteins among HIV-1 positive patients in Uganda

This was a cross-sectional study nested within a bigger study that was investigating the association between MDD and psychosocial risk factors in HIV [8, 17]. Study participants were ART naïve patients attending two specialised HIV clinics run by the AIDS Support Organisation (TASO) at Entebbe (semi-urban) and Masaka (predominantly rural) [18]. The MRC/UVRI Uganda Research Unit (the host research institution) has established research collaboration with these two study clinics. Out of 1100 participants enrolled for the main study, a random sample of 201 participants was selected using tables of random numbers to provide plasma samples for the immunological sub-study.

A structured standardised questionnaire administered by trained psychiatric nurses was used to collect the socio-demographic and clinical data for the mother study. The following socio-demographic factors that were relevant to the immunological sub-study were collected age, gender, highest educational attainment, religion, employment status, marital status and socio-economic index (SEI) which was constructed from commonly available household items in a typical Ugandan households, with higher scores reflecting higher socio-economic status [19]. Clinical data collected included the CD4 count and an assessment for major depressive disorder (MDD) using the Mini International Neuropsychiatric interview (M.I.N.I. Plus), a structured interview based on the DSM-IV and ICD-10 classifications [20].

Plasma IL-6, TNF-α and CRP concentrations were measured by ELISA. Anti human TNF-α (MAb TNF3/4) and IL-6 antibodies (MAb 13A5) were purchased from MABTECH AB, Sweden whereas the anti human CRP antibody (Part 842676) was purchased from Research & Diagnostics (R &D) systems, USA.

Briefly, 96-well immunolon 2HB micro plates (Thermo Labsystems, USA) were coated with anti human monoclonal antibodies (IL-6 at 0.25 μg/ml, TNF-α at 1.5 μg/ml and CRP at 2 μg/ml) at 100 μl per well and incubated over night at 4 °C. The following day, the coated plates were washed three times using 200 μl of PBS containing 0.05% Tween, pH 7.4 (Sigma-Aldrich, USA) and blocked for 1 h (at room temperature) with 200 μl of PBS and 0.05% Tween 20 (Sigma-Aldrich, USA) containing 0.1% MACS BSA buffer (Miltenyi Biotec, Germany). The stored plasma samples were thawed on the bench, diluted using the provided ELISA Diluent (1,2 for IL-6 and TNF-α)/Reagent Diluent (at 1:6400 for CRP) and 100 μl added in duplicates and then incubated at room temperature for 2 h. The plates were washed four times with 200 μl of PBS containing 0.05% Tween 20 and 100 μl of anti human IL-6 (MAb 39C3), TNF-α (MAb TNF5) or CRP (Part 842677) biotinylated detection antibodies (0.5 μg/ml for both IL-6 and TNF-α then 125 ng/ml for CRP) were added and incubated at room temperature (1 h for both TNF-α and IL-6 and 2 h for CRP). After the plates were washed as stated above, 100 μl of Streptavidin-HRP was added (diluted in incubation buffer at 1:2000 for both IL-6 and TNF-α, and 1:200 for CRP) and incubated (1 h for both IL-6 and TNF-α, and 20 min for CRP) at room temperature. After further washes as stated above, 100 μl of TMB-substrate (KPL, USA for the IL-6 and TNF-α and R & D systems, USA for CRP) was added for colour development. The assay reaction was stopped by adding 1 M sulphuric acid (100 μl to the IL-6 and TNF-α and 50 μl to the CRP) and the optical densities were measured immediately using a microplate reader (Biotek uQuant, USA) set to 450 nm. Any samples whose C.V were greater than 15% were repeated.

The study obtained ethical approval from the Uganda Virus Research Institute’s Science and Ethics Committee and the Uganda National Council of Science and Technology. Only study clinic attendees who were not impaired in anyway and had full cognitive insight were asked to participate in this study. Study participants who agreed to part in this study provided written consent after being provided with adequate information about the study. Respondents found to have significant psychiatric problems were referred to the psychiatric departments at Entebbe district hospital (at the semi-urban site) and Masaka regional referral hospital (rural site) for further assessment and management.

Table 1 shows the socio-demographic and clinical factors of the 201 participants in this study broken down by whether or not they had MDD. Overall MDD prevalence was 62 (30.8%). More than half of the participants came from TASO Masaka and participants from Masaka were more likely to have MDD than participants from TASO Entebbe (41.1% vs. 18.1%). Almost 80% of participants were female and female participants had a slightly higher prevalence of MDD than male participants (32.5% vs. 24.4%). About one third of participants were aged 18–29 years, one third were aged 30–39 and one third were aged 40 or over. There was no clear trend of MDD with age. Almost two thirds of participants had primary level as the highest level of education, while almost one quarter had secondary or higher education; those with secondary or higher education had a lower prevalence of MDD than those with primary or no education. Half of the participants were currently married; the prevalence of MDD was highest among the widowed and lowest among the currently married. Over half of the participants were Catholic; there was no clear pattern of MDD varying between different religions. Almost 37% of participants worked as traders, artisans or in the transport sector, with a further 30% involved in farming or fishing. MDD was highest among participants who were unemployed or retired. The prevalence of MDD decreased with increasing socio-economic status. About one quarter of participants had a baseline CD4 count of below 350 cells/μl (and thus qualified for ART under the then prevailing ART guidelines) with a further quarter having CD4 counts between 350 and 499 cells/μl. There was no clear pattern of MDD increasing or decreasing with CD4 count.

Figure 1 shows the bivariate associations between the inflammatory proteins and MDD. The IL-6 concentration was significantly higher among MDD participants compared to those without MDD (p < 0.0001; Median 22.4 pg/ml for MDD participants n = 62 and 5.762 pg/ml for participants with no MDD n = 139; Fig. 1B). There was a trend suggesting higher TNF-α concentrations among MDD participants compared to those without MDD (p = 0.0765; median 22.98 pg/ml for MDD participants n = 62 and 12.41 pg/ml for participants with no MDD n = 139; Fig. 1C). There was no trend between CRP and MDD (p = 0.7499; Fig. IA).

According to Table 2, participants in Masaka were more likely to have MDD than those in Entebbe. There was no evidence that gender, age or education level influenced MDD. Adjusting for the confounders there was no evidence that MDD was associated with the level of CRP. IL-6 and TNF-α were associated with MDD (Table 2) in that the odds of MDD increased with increasing IL-6, but in a non-linear fashion. Having any TNF-α decreases the odds of MDD, but as the level of TNF-α increases, the odds of MDD increase. Participants with low levels of TNF-α showed reduced risk of MDD compared to participants with no TNF-α, but as the level of TNF-α increase, the risk of MDD increases, and in particular participants with high levels of TNF-α (of 500 or above) were at a significantly increased risk of MDD. In summary, the odds of MDD increased with increasing levels of IL-6; participants with low levels of TNF-α were at reduced risk of MDD compared to participants with no TNF-α, but as the level of TNF-α increased, the risk of MDD increased, and in particular participants with high levels of TNF-α (of 500 or above) were at a significantly increased risk of MDD.