Increased Monocyte and T-cell activation in Treated HIV+ Ugandan Children: Associations with Gut Alteration and HIV factors

Abstract

Introduction:

The pathophysiology of immune activation and its mechanisms in children living with perinatally acquired HIV (PHIV) in sub-Saharan Africa has been understudied.

Methods:

We enrolled 101 children living with perinatally acquired HIV (PHIV) and 96 HIV negative controls (HIV-). All participants were between 10–18 years of age with no known active infections. PHIVs were on ART with HIV-1 RNA level ≤400 copies/mL. We measured plasma and cellular markers of monocyte activation, T-cell activation (expression of CD38 and HLA-DR on CD4+ and CD8+), oxidized lipids, markers of gut integrity and fungal translocation. Spearman correlations and linear regression models were used.

Results:

Overall median (Q1; Q3) age was 13 years (11; 15) and 52% were females. Groups were similar by age, sex and BMI. Median ART duration was 10 years (8; 11). PHIVs had higher monocyte and T-cell activation; higher sCD14 (p=0.01) and elevated frequencies of non-classical monocytes (p<0.001 for both). Markers of systemic inflammation (hsCRP), fungal translocation (BDG), intestinal permeability (zonulin) and oxidized lipids (ox LDL) correlated with monocyte and T cell activation in PHIV (≤0.05). After adjusting for age, sex, ART duration, protease inhibitor and non-nucleoside reverse transcriptase inhibitor use, a modest association between BDG and activated CD4+ T cells was observed (β=0.65, p<0.01). Oxidized LDL was inversely associated with activated T cells, inflammatory and non-classical monocytes (p<0.01).

Conclusion:

Ugandan children with perinatally acquired HIV with viral suppression have evidence of ongoing immune activation. Intestinal barrier dysfunction and fungal translocation may be involved in chronic immune dysfunction.

Introduction

Chronic immune activation is a hallmark of HIV and the causes of this residual immune dysregulation are likely multifactorial and may include ART drugs[1], micronutrient deficiencies[2], bacterial translocation from a damaged intestinal mucosal barrier[3], pro-inflammatory lipid profiles[4], including oxLDL [5], dysfunctional adipose tissue[6], and co-infections such as CMV[7]. Dysfunction in intestinal integrity and the resultant translocation of microbial products from the intestinal lumen to the systemic circulation appears to be a central factor in HIV-associated chronic immune activation in adults with HIV[8, 9].

Limited data exist on immune activation in children living with perinatally acquired HIV (PHIV) [10–13]. To our knowledge, no studies have assessed detailed cellular and soluble immune activation profiles in PHIV children and compared these profiles to those measured in HIV- children. Nor have studies explored the potential mechanisms behind the sustained immune activation in PHIV virally suppressed on ART from lower income settings, including the roles of oxidized lipoproteins and alterations in intestinal integrity on immune activation in this population. This is important because the effect of intestinal damage and bacterial translocation on immune activation may differ in children infected perinatally whose intestinal epithelium may be more permissive to antigens than that of adults [14]. A growing body of research also indicates that early life stress, economic insecurity, environmental enteropathy and air pollution also drive chronic inflammation ([7, 15–18]. These factors are all more prevalent in PHIV living in low-and-middle income countries (LMICs) specifically in eastern and southern Africa, where the majority of PHIV live [19] and where this study took place. Thus, evaluating the potential mechanisms of chronic immune activation and inflammation in LMICs is of particular importance in children as they age into adulthood. In this study, we assessed soluble and cellular markers of monocyte activation and T-cell activation, and focused on evaluating the associations with plasma markers of systemic inflammation, oxidized lipids, and gut integrity in PHIV and age and sex matched HIV- Ugandan children. The primary objective of this study was to determine whether monocyte and T cell activation markers were different by HIV status; second, to determine if ongoing immune activation was associated with alteration in gut integrity, translocation of fungal products, or oxidized lipoproteins.

Methods

Study Design

This is a cross-sectional analysis of baseline data from an observational cohort study of PHIV and HIV- children prospectively enrolled at the Joint Clinical Research Center in Kampala, Uganda between September 2017 and May 2019. The study was approved by the Research Ethics Committee in Uganda, the Ugandan National Council of Science and Technology as well as the IRB of the University Hospitals Cleveland Medical Center, Cleveland, Ohio. Caregivers gave written informed consent. All participants were 10–18 years of age. PHIV participants were on ART for at least 2 years with a stable regimen for at least the last 6 months with HIV-1 RNA < 400 copies/mL. HIV- participants were either HIV- family members of the PHIV or recruited from the community using community liaison volunteers from the JCRC. All participants lived in Kampala or peri-urban surroundings. Evidence of self-reported or documented diarrhea or acute infection (malaria, tuberculosis, helminthiasis, pneumonia, meningitis) in the last 3 months, as well as moderate or severe malnutrition were excluded. Adolescents with pregnancy or intent to become pregnant were excluded.

Study Evaluations

Blood was drawn after an 8-hour fast. Blood was processed and plasma, serum, and PBMCs were cryopreserved for shipment to University Hospitals Cleveland Medical Center, Cleveland, Ohio.

Cellular markers of monocyte and T-cell activation

Monocytes and T-cells were phenotyped by flow cytometry as previously described by Dr. Funderburg[20]. CD4⁺ and CD8⁺ T-cell activation was measured by expression of CD38 and HLA-DR. Monocyte subsets were determined by the relative expression of CD14 and CD16.

Inflammation, soluble immune activation and gut markers

We selected intestinal biomarkers based on prior data in PHIV in Uganda as well as in adults living with HIV suggesting a potential role in cardiovascular disease (CVD) and inflammation. Zonulin (Promocell Germany), a marker of intestinal permeability and Beta D glucan (BDG, Mybiosource Inc. CA), a polysaccharide cell wall component of most fungal species, were found to be elevated in PHIV and correlated with systemic inflammation [21]. Intestinal fatty acid binding protein (I-FABP, R &D Systems, Minneapolis, Minnesota, USA) is considered a marker of enterocyte inflammation or damage and associated with metabolic complications of HIV[22, 23]. Soluble CD14 (sCD14, R &D Systems, Minneapolis, Minnesota) is a marker of monocyte activation, and is associated with mortality and progression of atherosclerosis[26]. We have shown that oxidized lipids (oxLDL) upregulates monocyte activation in HIV[24], making oxLDL a potentially important mediator on the causal pathway of monocyte activation.

Plasma markers of monocyte activation (sCD163), systemic inflammation (sTNFRI), high sensitivity C-reactive protein (hsCRP) and IL-6 were selected because they correlate with cardiometabolic complications, or drive other hallmarks of immune dysregulation in HIV. sCD163, sTNFRI, hsCRP, IL6 and ox LDL were measured by ELISA (R &D Systems, Minneapolis, Minnesota, USA, ALPCO, Salem, New Hampshire, USA and Mercodia, Uppsala, Sweden). The intra-assay variability ranged between 4–8% and inter-assay variability was less than 10% for all markers. All assays were performed at Dr. Funderburg’s laboratory at Ohio State University, Columbus, OH. Laboratory personnel were blinded to group assignments.

Statistical analysis

We studied quality of the data by running frequency analyses, and graphs. Median and interquartile range (IQR) are reported for continuous variables and numbers and percentages are reported for categorical variables. All variables were compared between groups using Wilcoxon ranks sum tests or Chi square, as appropriate. BMI-for-age z scores were determined using WHO 2007 reference values which is a reconstruction of the 1977 National Center for Health Statistics/WHO reference. T-cell and monocytes measures were presented as percent of activated cells, and presented using box plots by groups. The relationship between T-cells and monocytes, demographic variables, HIV-related variables and biomarkers were assessed using Spearman Correlation analyses.

Multivariable quantile (median) regression models were used to further explore the relationship between CD4+ T-cells, inflammatory and non-classical monocytes and biomarkers in PHIV. Demographic risk factors and HIV-related risk factors were considered for inclusion in the multivariable model. All statistical analyses were performed using software Stata 15.0 and R 3.5.2.

Results

Baseline

A total of 197 participants were enrolled, 101 PHIV and 96 HIV- children (Table 1). Overall, the median age was 13 years and 52% were females. There were no differences in age, gender and BMI between the groups.

Table 1:

Baseline characteristics

	PHIV N=101	HIV- N=96	p
Age (years)	12.93 [11.53, 14.71]	12.67 [11.08, 14.33]	0.24
Female sex (%)	54 (53)	50 (52)	0.96
Waist:hip ratio	0.87 [0.83, 0.90]	0.85 [0.82, 0.89]	0.02
Body mass index (kg/m2)	17.69 [15.82, 19.10]	18.02 [16.09, 19.46]	0.34
BMI-for-age Z score	−0.57 [−1.29, −0.01]	−0.33 [−1.04, 0.36]	0.066
HIV variables
Viral load< 20 copies/mL (%)	84 (86)
CD4 nadir (cells/µL)	619.50 [333, 1097]
CD4 cell count (cells/µL)	988 [638, 1307]
CD4 percent	34.50 [27, 41]
ART Duration (years)	9.88 [7.61, 11.08]
Nucleotide Reverse Transcriptase Inhibitor (%)
Abacavir	42 (47)
Lamivudine	1 (1)
Tenofovir	12 (14)
Zidovudine	33 (37)
Nevirapine (%)	18 (27)
Efavirenz (%)	46 (44)
Lopinavir/ritonavir (%)	27 (28)
Lipids
Total cholesterol (mg/dL)	152 [134, 172]	148 [131, 170]	0.49
LDL (mg/dL)	84 [68, 104]	83 [70, 105]	0.61
VLDL (mg/dL)	19 [13, 23]	16 [12, 20]	0.03
Triglycerides (mg/dL)	93 [66, 115]	82 [61, 102]	0.03
Cellular Markers of Monocyte Activation
sCD14 (pg/mL)	2110.70 [1759.38, 2568.92]	1675.39 [1445.09, 1956.71]	<0.001
sCD163 (pg/mL)	589.60 [404.24, 735.95]	684.97 [495.51, 854.47]	0.01

Median [Interquartile Range]

Bold values represent p<0.05

PHIV: children with perinatally acquired HIV, sCD14 and 63: soluble CD14 and 63

Among PHIV, 86% had viral load of <20 copies/mL, median CD4 cell count was 988 cells/µL, 72% were on a non-nucleotide reverse transcriptase inhibitor (NNRTI) regimen, 28% were on lopinavir/ritonavir (LPV/r) and 2 participants were on dolutegravir. All PHIV were on cotrimoxazole following local guidelines.

Immune activation and HIV disease factors

PHIV participants had higher sCD14 (Table 1), a higher proportion of non-classical CD14^dimCD16⁺ monocytes (Figure 1 a) and increased activation marker expression on CD4+ and CD8+ T-cells (Figure 1 b) compared to these indices in HIV- children.

Figure 1a: Monocyte activation between groups.

Dot plots showing the percentage of monocyte subsets in children with perinatally acquired HIV (PHIV) compared to HIV negative children (HIV-). Median % (Q1, Q3) CD14+CD16- monocytes was 68% (59, 78) vs 71% (64, 77) for PHIV and HIV- participants respectively. Median % CD14+CD16+ monocytes was 22% (15, 30) vs 23% (17, 27) for PHIV and HIV- participants respectively. Median % CD14dimCD16+ monocytes was 8% (6, 11) vs 7% (4, 9%) for PHIV and HIV- participants respectively.

Figure 1b: T cell activation between groups.

Dot plots showing the median percentage of activated CD4+ and CD8+ T cells expressing CD38 and or HLA-DR in children with perinatally acquired HIV (PHIV) compared to HIV negative children (HIV-). Horizontal lines denote interquartile range. A) Median (Q1,Q3) % CD4+CD38+ T cells was 11.6% (10.2, 13.2) vs 11.6% (9.9, 13.2) for PHIV vs HIV- participants respectively. B) Median CD8+CD38+ T cells was 8.1% (6.7, 10.2) vs 8.3 (5.8, 10.1) for PHIV vs HIV- participants respectively. C) Median CD4+HLA-DR+ T cells was 11.2 (9.6, 12.7) vs 9.2 (7.6, 10.1) for PHIV vs HIV- participants respectively. D) Median CD8+HLA-DR + T cells was 8.1% (7.2, 9.4) vs 7.2% (6.2, 7.8) for PHIV vs HIV- participants. E) Median CD4+CD8+HLA-DR T+ T cells was 4.1 (2.9, 5.3) vs 3.6 (2.6, 4.6) for PHIV vs HIV- participants. F) Median CD8+CD38+HLA-DR+ T cells was 8.9% (6.7, 14.7) vs 8.3 (5.8, 12.7) for PHIV vs HIV- participants.

There was no correlation between ART duration, PI or NNRTI use and plasma, cellular markers of monocyte activation or T-cell activation (p≥0.32). There was a correlation with nadir CD4+ cell count and traditional (CD14+CD16-) and non-classical monocytes, as well as CD4 and CD8+ T-cells expressing HLA-DR and CD38 (p≤0.034).

Association between markers of gut integrity/microbial translocation and immune activation

As shown in Figure 2a, markers of systemic inflammation (hsCRP), fungal translocation (BDG), intestinal permeability (zonulin) and oxidized lipids (ox LDL) correlated with monocyte and T- cell activation in PHIV. In HIV-, however, only markers of systemic inflammation correlated with monocyte or T-cell activation (Figure 2b).

Figure 2: Correlation between immune activation and markers of inflammation, gut integrity and oxidized lipoprotein.

Heat map showing the correlation between T cells and monocyte activation and measures of inflammation, gut integrity and oxidized lipids in A) HIV negative children and B) children living with perinatally acquired HIV (PHIV).

*for p <0.05; ** for p<0.01; *** for P<0.001

T cells and monocyte represent % lymphocytes and % monocytes

BDG: Beta-d-glucan (pg/mL), hsCRP: high sensitivity C reactive protein (ng/mL), IFAB-P: intestinal fatty acid binding protein (pg/mL), IL-6: interleukin 6 (pg/mL), oxLDL: oxidized LDL, sCD14 and 163: soluble CD14 and soluble CD163 (pg/mL), sTNFR-I: soluble tumor nectrosis factor α I (pg/mL)

Separate quantile regression models examined the relationship between gut biomarkers and oxidized LDL, and selected T-cell and monocyte activation in PHIV after adjusting for potential demographic and HIV related confounders (Table 2). No associations were found with age, sex or PI use in any of the models (p≥0.06). However, NNRTI use was consistently associated with higher % of non-classical monocytes (β=0.03, p≤0.03). There was a significant association between BDG and activated CD4+ T cells. We found an inverse relationship between oxidized LDL and activated T cells and monocytes.

Table 2:

Association with CD4 activated T cells and non-classical monocytes using quantile regression analysis within PHIV

Predictorsb	Coef.	SE	p-value	95%CI
				L.L.	U.L.
%CD4+CD38+HLADR+
Zonulin (ng/mL)	0.275	0.206	0.188	−0.139	0.690
BDG (pg/mL)	0.655	0.146	<0.001	0.362	0.949
IFAB-p (pg/mL)	0.171	0.221	0.443	−0.274	0.615
Oxidized LDL (mU/L)	−0.409	0.206	0.053	−0.824	0.006
%CD14dimCD16+
Zonulin (ng/mL)	0.822	0.704	0.251	−0.606	2.250
BDG (pg/mL)	−0.603	0.399	0.139	−1.412	0.205
IFAB-p (pg/mL)	0.825	0.675	0.230	−0.544	2.193
Oxidized LDL (mU/L)	−1.617	0.490	0.002	−2.610	−0.624
%CD14+CD16+
Zonulin (ng/mL)	0.018	1.293	0.989	−2.604	2.640
BDG (pg/mL)	0.932	0.870	0.291	−0.833	2.696
IFAB-p (pg/mL)	−0.808	1.413	0.571	−3.674	2.057
Oxidized LDL (mU/L)	−5.011	1.370	0.003	−7.791	−2.232

^aQuantile (median) regression equation for biomarkers: Q_τ(O) = β₀(τ) + β₁(τ) Age + β₂(τ) Male + β₃(τ) ART Duration in years + β₄(τ) Protease use + β₅(τ) NNRTI use + β6(τ) p6, where O outcome variable is either %CD4⁺CD38⁺HLADR⁺or %CD14^dimCD16⁺ or %CD14⁺CD16⁺, and

^bPredictors p6 is one of these markers: Zonulin, BDG, IFABD-p and Oxidized LDL, τ = 0.5 for median regression. β is the associated regression coefficient. All continuous variables or biomarkers were rescaled for a tangible regression coefficient estimate by dividing its standard deviation, except for age and ART duration. Coef., SE, p, L.L., and U.L. represent regression coefficients, standard error of the estimates, p-value, lower limit, and upper limit of 95% confidence interval (CI) of the regression coefficients, respectively. Bold values represent p≤0.05.

DISCUSSION

To the best of our knowledge, this is the first comprehensive study investigating T-cells and monocyte activation in PHIV on ART in sub-Saharan Africa and in age and sex matched uninfected children. Despite benefits of ART and viral suppression, Ugandan PHIV have evidence of T cell (both CD4+ and CD8+) and monocyte activation. Modest associations were observed between zonulin, BDG, and T cell activation. Interestingly, contrary to our hypothesis and the findings in adults, oxidized lipoproteins were inversely associated with chronic immune activation in ART treated PHIV.

Despite the high prevalence of HIV in children living in sub-Saharan Africa, there are limited data on immune activation in PHIV in the setting of ART. With universal ART recommendation, understanding immunologic perturbations that are associated with disease progression that persist despite viral suppression in PHIV is paramount to reducing long term complications as children age into early adulthood. PHIV in this study have both increased CD4+ and CD8+ T cell activation compared to levels in HIV negative children, supporting prior findings from the US, Spain and Uganda[9–11, 13]. In addition, we found that Ugandan PHIV have higher levels of the non-classical monocytes, unlike what has been found in US children[10, 25]. PHIV in our study did not demonstrate a normalization in immune activation on ART compared to activation levels in negative participants, unlike what has been shown in Kenyan PHIV 6 months post ART[26]. Although our findings could be secondary to different methodology, small sample size in these previous reports, or differing mode of HIV acquisition (horizontal vs vertical), they could also be due to differences in ethnicity, genetics, nutritional or infectious factors in these populations. Thus, there is a need for further studies in African PHIV and for exercising caution when extrapolating data from PHIV in higher income settings to PHIV in lower income settings.

Despite a mean ART duration of 10 years and successful viral suppression, Ugandan PHIVs in this study have ongoing immune dysfunction. Similarly to adults living with HIV, there is a decline in mortality rate in PHIV[27] and an increase in non-AIDS co-morbidities[28] due to successful ART. Immune activation can promote viral replication, but most importantly, it is a main driver of non-AIDS complications in adults with HIV[29]. The limited data on the role of heightened immune activation in PHIV suggest that T cell activation may contribute to neurocognitive impairment[25] and monocyte activation to insulin resistance[30] and warrants further investigation.

Immune activation in this population is likely the consequence of several mechanisms, including: HIV and ART related factors, antigenic stimulation by other viruses (CMV, EBV), a compromised intestinal integrity, gut dysbiosis, and abnormal production of reactive oxygen species. In our study, ART duration and PI use were not associated with either T cell or monocyte activation, however proportional representation of non-classical monocytes (CD14^dimCD16+) was associated with NNRTI use. Although the etiology for this association is unclear, it is consistent with findings from trials showing that soluble markers of monocyte activation sCD14 and sCD163 do not consistently decrease with initiation of ART containing either NNRTIs or PIs[31]. Integrase inhibitors may have a larger effect on reducing inflammation[32], and with the recent universal recommendation to use integrase inhibitors as first line therapy[33], additional studies will be needed to determine whether there is an inflammatory benefit to using this class of drugs.

We specifically explored the associations among T cell and monocyte activation and intestinal integrity, fungal translocation, and oxidized LDL because our group has shown previously that 1) Ugandan PHIV have altered intestinal permeability and fungal translocation[21], 2) fungal translocation as measured by BDG is associated with inflammation and immune activation in adults living with HIV[34] and 3) that oxidized lipoproteins may drive immune activation in adults living with HIV [5, 35]. We found that zonulin, a marker of intestinal permeability and BDG, a marker of fungal translocation correlated with activated CD8 T cells and monocyte subset proportions. These findings confirm our hypothesis and further support the ongoing relationship between immune activation, fungal translocation and the ongoing disruption in the intestinal barrier despite viral suppression. Since our group found that oxLDL is a main driver of systemic inflammation and monocyte activation in adults[24, 35], we assessed the correlation between oxLDL and measured biomarkers. Surprisingly, contrary to what is known in adults, we found that increased CD4+ T cell activation and monocyte activation was associated with lower oxidized LDL. Oxidized LDL can modulate innate and adaptive immunity[36]. HIV pathogenesis differs among adults, children, and Ugandan PHIV with normal BMI and low LDL cholesterol, as oxidized LDL does not appear to be in the pathway of immune activation in younger populations. There may be different reactive oxygen species and oxidative pathways in this population that could explain this relationship. Further studies are needed to elucidate whether oxidized lipoproteins or other oxidative components contribute to immune activation in PHIV as they age.

Surprisingly, we found that sCD163 was higher in uninfected participants. Although the selected participants were healthy, this may represent heightened inflammation due to unmeasured confounders including early life stress, economic insecurity, environmental enteropathy [7, 15–18]. In addition, we hypothesize that this finding could also be the result that PHIV participants have access to routine health care and are on co-trimoxazole. The role of co-trimoxazole in mitigating mortality and improving health outcomes may be attributed to its anti-inflammatory properties: in the ARROW trial, PHIV children who were randomly assigned to discontinue co-trimoxazole had persistent increase in several inflammatory biomarkers[37].

Our study has several limitations. As in all cross-sectional studies, we cannot prove causal relationships or exclude the possibility of residual confounding. Although we excluded participants with malaria, tuberculosis, and diarrhea, we did not evaluate other viral etiologies such as CMV or EBV which could contribute to immune activation in this population. We did not measure zonulin using a monoclonal antibody assay and caution should be used when interpreting zonulin ELISA assays[38]. Additionally we did not collect stool samples to be able to assess the gastrointestinal microbiome composition.

Conclusions

Despite these limitations, this is the most comprehensive study describing monocyte and lymphocyte activation in PHIV on ART with viral suppression compared to age and sex matched uninfected participants in Uganda. Our findings highlight that immune activation (lymphocyte and monocyte) persist in Ugandan PHIV despite ART. Our data support the hypothesis that gut integrity and microbial translocation may play a role in ongoing immune activation in this population. Longitudinal follow up of this cohort is ongoing and is needed to confirm these findings and may help determine the consequences of ongoing immune activation in African PHIV and whether it is associated with non-AIDS complication as children age. A deeper understanding of the mechanisms causing immune activation in ART treated African PHIV is crucial.