Cytokine Profiles in Aqueous Humor and Plasma of HIV-infected Individuals with Ocular Syphilis or Cytomegalovirus Retinitis

ABSTRACT

Purpose: To characterize the immunologic profile in aqueous humor (AqH) of HIV-infected individuals with cytomegalovirus retinitis (CMVr) or ocular syphilis and to assess if AqH and plasma represent independent cytokine compartments.

Methods: Concentrations of 27 cytokines in AqH and plasma of HIV-infected individuals with CMVr (n = 23) or ocular syphilis (n = 16) were measured by multiplex assay. Cytokine profiles of both groups were compared.

Results: Individuals with CMVr had higher plasma concentrations of interleukin (IL)-7, IL-8, IL-10, interferon (IFN)-γ, IFN-α2, G-CSF, IP-10 and IL-1α; as well as higher AqH concentrations of IL-1α, IP-10 and GM-CSF than those with ocular syphilis. AqH and plasma levels correlated only for IP-10 in both ocular infections.

Conclusions: Individuals with CMVr had higher plasma cytokine levels than those with ocular syphilis. The immunologic profiles in AqH and plasma are independent. Therefore, AqH cytokine concentrations cannot be inferred from plasma cytokine concentrations in the population studied.

Keywords:

• AIDS
• cytokines
• cytomegalovirus retinitis
• HIV
• ocular syphilis

The eye is an immunologically privileged site characterized by modification of the innate and adaptive immune responses, immunologic ignorance, and peripheral tolerance derived in part from the ability of deviating humoral and cellular immunity to antigens (anterior chamber-associated immune deviation, ACAID). This immunosuppressive microenvironment allows the preservation of intraocular structures.¹ Inflammation caused by ocular infections (OIs), such as CMVr and syphilis, contributes to the destruction of ocular structures. These OIs are commonly associated with human immunodeficiency virus (HIV) infection,² and both are frequently observed in resource-limited settings. CMVr is an important cause of blindness in individuals with advanced HIV infection and is characterized by intraretinal hemorrhages, white zones of retinitis, retinal edema, and vasculitis.³ Ocular syphilis typically occurs at earlier stages of HIV infection and manifestations may include anterior uveitis, vitritis, retinitis, choroiditis, papillitis, or panuveitis.⁴

The levels of some cytokines involved in various OIs and idiopathic uveitis have been described in immunocompetent individuals,^5–7 and no correlation between aqueous humor (AqH) and plasma values has been found.^8,9 However, the description of cytokine profiles characterizing different OIs in HIV-infected population remains incomplete. High levels of IL-6 have been reported in HIV-infected individuals with active CMVr, while high levels of IL-12 were found in those with scarred CMVr and uveitis.¹⁰ Recent studies in individuals with HIV infection and CMVr reported cytokine profiles consistent with a Th1 response mediated by lymphocytes and monocytes/macrophages in AqH, with decreasing concentrations after intravitreal ganciclovir treatment.^11,12 In HIV-infected individuals with syphilis, elevated serum levels of IL-10 and TNF-α were described, with a decrease of both cytokines after treatment.¹³

The aim of this study was to investigate the cytokine and chemokine profile in AqH and plasma of HIV-infected individuals with CMVr or ocular syphilis. A panel of 27 cytokines was analyzed using a multiplex assay.

MATERIALS AND METHODS

Study Population

This study was conducted at the Center for Research in Infectious Diseases at the National Institute of Respiratory Diseases (INER), a referral center in Mexico City. Participants were enrolled in the study during a period of 2 years. Institutional Review Board (IRB)/Ethics Committee approval was obtained according to the Declaration of Helsinki. Written informed consent was obtained from all participants. A total of 39 HIV-infected individuals were included in this prospective, observational study. Of these, 23 had CMVr and 16 had ocular syphilis. Plasma and AqH samples were obtained from participants before specific treatment of ocular infection. The medical work-up at the INER for HIV-infected patients includes CD4 T-cell counts and HIV RNA load determinations. All individuals were examined by slit-lamp biomicroscope and indirect ophthalmoscope by a retina and uveitis specialist for detection of ocular manifestations of HIV infection. Diagnosis of OIs was based on clinical examination. CMVr was defined by white retinal lesions, intraretinal hemorrhages, vasculitis, and retinal edema in the absence of vitritis.³ Ocular syphilis was defined by one or more of the following signs: uveitis, vitritis, papillitis, retinitis, choroiditis, or panuveitis, with positive serum VDRL and FTA-ABS tests.⁴ After plasma and AqH samples were obtained, CMVr was treated with oral valganciclovir.^14,15 Patients with ocular syphilis received 2 weeks of treatment with intravenous penicillin.⁴

Plasma Samples

Peripheral EDTA anticoagulated blood (20 mL) was obtained from HIV-infected individuals at OI diagnosis. Plasma was separated by centrifugation and stored at –80ºC for further analyses.

Aqueous Humor Samples

AqH (100–150 μL) was obtained from HIV-infected individuals at OI diagnosis. AqH was collected by anterior chamber puncture with a 30-gauge needle, using topical anesthesia under standard aseptic and antiseptic techniques. Samples were immediately stored at –80ºC until use.

Cytokine and Chemokine Quantification

The concentrations of cytokines and chemokines in AqH and plasma were measured with a Luminex 200 instrument (Luminex Corporation, Austin, TX) using xMAP (Multi-Analyte Profile) technology.¹⁶

Data were analyzed by Milliplex Analyst software to determine cytokine concentrations (VigeneTech, Carlisle, MA, USA). The Milliplex human cytokine reagent kit (Millipore, Billerica, MA) was used to measure a cytokine panel consisting of: eotaxin, granulocyte colony-stimulating factor (G-CSF); granulocyte/macrophage colony-stimulating factor (GM-CSF); interferon alpha-2 (IFN-α2); interferon gamma (IFN-γ); IL-10, IL-12p40, IL-12p70, IL-13, IL-15, IL-17, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8; protein-10 induced by gamma interferon (IP-10); monocyte chemoattractant protein-1 (MCP-1); macrophage inflammatory protein-1 alpha (MIP-1α); macrophage inflammatory protein-1 beta (MIP-1β); RANTES (regulated on activation, normal T cell expressed and secreted); tumor necrosis factor alpha (TNF-α); and tumor necrosis factor beta (TNF-β).

Data Analysis

We first explored whether cytokine patterns of HIV-infected individuals with viral ocular co-infection (CMVr) differed from those with bacterial ocular co-infection (syphilis) using the Mann–Whitney test. Given the large number of assessments for each group, we used the false discovery-based q-value statistic to correct for multiple hypothesis testing.¹⁷ Statistical significance was set at p<0.05 and q<0.2. All analyses were performed with R statistical software version 3.0.2 and associated packages. We used the Kruskal–Wallis test for comparison of pre-treatment immunologic variables, such as CD4 T-cell counts and HIV RNA loads between groups. We determined correlations between AqH and plasma cytokine concentrations by using Spearman’s Rho. In order to describe the cytokine profiles in plasma and AqH of both groups, we carried out a principal component analysis (PCA) of all cytokine variables using FactoMineR.¹⁸

RESULTS

The group of 39 HIV-infected individuals with OI included three women and 36 men (median age was 33 years, interquartile range, IQR 27–39). Of those, 23 had CMVr (median CD4 count of 49 cells/μL, IQR 23–145, median HIV RNA load of 4.0 log₁₀ copies/mL, IQR 2.4–5.4); and 16 had ocular syphilis (median CD4 count of 317 cells/μL, IQR 218–480, median HIV RNA load of 3.2 log₁₀ copies/mL, IQR 1.6–5.0) (Table 1).

TABLE 1. Immunologic characteristics of study participants.

	CMVr (n = 23)		Ocular syphilis (n = 16)
	Median	IQR	Median	IQR
CD4+ (T cells/μL)	49	23–145*,^†	317	218–480*,^†
CD4+ (T cell %)	9.0	6.8–12.5	17.0	11.8–21.0
HIV RNA (copies/mL)	10,943	109–265 637*	21,037	39–148 469*
Log HIV RNA (copies/mL)	4.0	2.4–5.4	3.2	1.6–5.0

CMVr, cytomegalovirus retinitis; IQR, interquartile range.

*Kruskal–Wallis p<0.0001 for CD4 T cells/μL and p = 0.003 for HIV RNA copies/mL.

^†Bonferroni p = 0.00007.

Differential Cytokine Profiles between HIV-infected Individuals with CMVr or Ocular Syphilis

Concentrations of 27 cytokines in AqH and plasma are shown in Table 2. AqH and plasma samples of HIV-infected individuals with CMVr were compared with those affected by ocular syphilis (Table 3). Individuals with CMVr had significantly higher AqH concentrations of GM-CSF, IL-1α, and IP-10 (p<0.05, q<0.2) compared with those with ocular syphilis (Figure 1). Levels of IL-1α, IP-10, IFN-γ, IFN-α2, IL-10, IL-7, IL-8, and G-CSF were higher in plasma of individuals with CMVr (p<0.05, q<0.2) (Figures 1 and 2). We found undetectable levels of T_H2 cytokines, such as IL-4, IL-5, and IL-13; and of IL-12p40, IL-12p70, IL-15, IL-17, IL-1β, IL-2, IL-3, and TNF-β in AqH and plasma of individuals with CMVr or ocular syphilis.

TABLE 2. Cytokine concentrations in aqueous humor and plasma of study participants (median pg/mL).

	CMVr (n = 23)		Ocular syphilis (n = 16)
Cytokine	Plasma	AqH	Plasma	AqH
Eotaxin	66.9525	65.81	61.39	43.76
G-CSF	107.475	344.046667	45.1725	119.48
GM-CSF	4.085	59.98	2.44	15.47
IFN-2α	25.495	123	12.505	37.13
IFN-γ	6.22	24.51	3.84	4.47
IL-10	18.085	8.44	8.4	13.58
IL-12p40	59.1^a	5.05^a	59.1^a	5.05^a
IL-12p70	7.81^a	2.37^a	7.81^a	2.37^a
IL-13	7.81^a	15.36^a	7.81^a	15.36^a
IL-15	2.86^a	2.86^a	2.86^a	2.86^a
IL-17	3.11^a	3.14^a	3.11^a	3.14^a
IL-1α	20.655	991	12.475	88.92
IL-1β	1.61^a	5.03^a	1.61^a	5.03^a
IL-2	1.54^a	3.94^a	1.54^a	3.94^a
IL-3	1.99^a	190	1.99^a	1.66^a
IL-4	13.30^a	8.34^a	13.30^a	8.34^a
IL-5	1.65^a	3.99^a	1.65^a	3.99^a
IL-6	6.57	254	2.07	79.095
IL-7	6.3975	2.43^a	1.89	2.43^a
IL-8	14.2175	87.94	5.61	54.95
IP-10	2718	14772.44	972.75	4622.5
MCP-1	316.5	8759	253	3740.5
MIP-1α	26.39^a	30.33	26.39^a	41.1
MIP-1β	30.71	53.96	28.23	42.05
RANTES	4841	50.69^a	3952	50.69^a
TNF-α	26.6725	18.52	21.405	5.96
TNF-β	4.12^a	7.91^a	4.12^a	7.91^a

CMVr, cytomegalovirus retinitis; AqH, aqueous humor.

^aBelow detection limit.

TABLE 3. Comparison of cytokine concentrations in AqH and plasma of HIV-infected individuals with CMV retinitis or ocular syphilis (pg/mL).

		CMVr (n = 23)		Syphilis (n = 16)
Cytokine	Compartment	Median	IQR	Median	IQR	p value (<0.05)^a	q value (<0.2)^b
GM-CSF	AqH	59.98	12.46–95.64	15.47	5.92, 38.44	0.0262	0.1571
IL-1α	AqH	991	48.87–1389	88.92	29.12–221.4	0.033	0.15
IP-10	AqH	14772.44	12430–14772.44	4622.5	2692–10480	0.002	0.046
G-CSF	Plasma	107.475	31.06–213.8	45.1725	15.51–58.179	0.032	0.157
IFN-α2	Plasma	25.495	14.66–72.29	12.505	7.358–20.21	0.022	0.151
IFN-γ	Plasma	6.22	3.84–10.09	3.84	3.84–3.87	0.003	0.047
IL-10	Plasma	18.085	12.59–36.33	8.4	4.58–19.179	0.034	0.157
IL-1α	Plasma	20.655	9.902–92.73	12.475	2.87, 19.04	0.043	0.157
IL-7	Plasma	6.3975	2.428–7.858	1.89	1.89–4.605	0.007	0.067
IL-8	Plasma	14.2175	8.968–23.58	5.61	4.458–10.54	0.003	0.047
IP-10	Plasma	2718	1903–4729	972.75	726.4–2839	0.001	0.046

CMVr, cytomegalovirus retinitis; AqH, aqueous humor; IQR, interquartile range.

^aWilcoxon test.

^bStorey’s q for multiple comparisons.

Only significantly different values are shown.

FIGURE 1. Cytokine concentrations in plasma and aqueous humor. Cytokine levels were plotted for HIV-infected individuals with cytomegalovirus retinitis (CMVr) or with ocular syphilis. Only cytokines showing significant differences are presented in the figure (Wilcoxon sign-rank test p<0.05; Storey q<0.2).

FIGURE 2. Cytokine concentrations in plasma. Cytokine levels were plotted for HIV-infected individuals with cytomegalovirus retinitis (CMVr) or with ocular syphilis. Only cytokines showing significant differences are presented in the figure (Wilcoxon sign-rank test p<0.05; Storey q<0.2).

Comparison of AqH and Plasma Cytokine Profiles

Concentrations of RANTES and IL-7 were significantly higher in plasma than in AqH in both OIs. In contrast, levels of MIP-1α, IFN-γ, MIP-1β, GM-CSF, IFN-α2, IL-8, IL-3, IL-1α, G-CSF, IL-6, MCP-1, and IP-10 were significantly higher in AqH than in plasma. There was a significant positive Spearman correlation (p<0.01, q<0.19) in plasma and AqH concentrations only for IP-10 (Rho = 0.48; Table 4).

TABLE 4. Comparison of aqueous humor and plasma cytokine profiles (median pg/mL).

Cytokine	AqH-plasma	p value (<0.05)^a	q value (<0.2)^b	Rho	p value (Rho)	q value (Rho)
RANTES	–4872.31	0.0000000009	0.00000006	–0.1224	0.511	0.73
IL-7	–0.41	0.01	0.03	–0.0449	0.810	0.74
MIP-1α	9.03	0.0003	0.002	–0.4068	0.023	0.31
IFN-γ	16.18	0.0005	0.003	0.0214	0.908	0.75
MIP-1β	18.96	0.008	0.02	0.0986	0.597	0.74
GM-CSF	43.01	0.0000001	0.000005	0.1932	0.297	0.72
IFN-α2	54.52	0.002	0.01	–0.0235	0.900	0.75
IL-8	68.42	0.000001	0.00003	0.0161	0.931	0.75
IL-3	110.055	0.00009	0.0007	NA	NA	NA
IL-1α	143.18	0.00001	0.0001	0.2420	0.189	0.67
G-CSF	173	0.00009	0.0007	0.1218	0.513	0.73
IL-6	206.95	0.000003	0.00004	–0.0359	0.847	0.74
MCP-1	8273.5	0.0000000009	0.00000006	0.2121	0.251	0.70
IP-10	8509	0.000000004	0.0000002	0.4795	0.006	0.18

AqH, aqueous humor; Rho, Spearman’s rank correlation coefficient (n = 31).

^aWilcoxon

^bStorey’s q for multiple comparisons.

Cytokine Profile Analysis by Principal Components Analysis (PCA)

We selected the dimensions that explained >5% of the variance for the PCA of baseline data for plasma and AqH. We found that the dimensions were formed by multiple variables, including chemokines (eotaxin, IL-8, IP-10, MCP-1, MIP-1α, MIP-1β, and RANTES); T_H1 cytokines (IL-12p40, IL-12p70, IL-1α, IL-1β, IL-2, IL-6, TNF-α, and TNF-β); T_H2 cytokines (IL-4, IL-5, IL-10, and IL-13); stimulatory factors (G-CSF, GM-CSF, IL-3, and IL-7) and antivirals (IFN-α2 and IFN-γ). These cytokines were not grouped into a single function.

There was no correlation between the dimensions in AqH and plasma, indicating that the patterns of cytokines in the two compartments are independent. By comparing the main dimensions of cytokine concentrations in AqH and plasma of individuals with CMVr versus those with ocular syphilis, we determined that baseline cytokine profiles were specific for each OI (Figure 3).

FIGURE 3. Principal component analysis (PCA). Graphics of each measurement (subject) in the two main dimensions of the PCA. (A) shows AqH dimensions in ocular syphilis and cytomegalovirus retinitis (CMVr). (B) shows plasma dimensions in ocular syphilis and CMVr. (C) represents the main type of cytokine response in AqH (SF, stimulatory factors). (D) represents the main type of cytokine response in plasma. For each group, an ellipse of 95% reliability, calculated by 100 repetitions with replacement (bootstrapping), was drawn.

DISCUSSION

To our knowledge, this is the first study comparing the immunologic profile in AqH and plasma of HIV-infected individuals with CMVr versus those with ocular syphilis. We used additional multivariate analysis to more comprehensively explore the dynamics of AqH in CMVr and ocular syphilis. There is limited information about the differences and similarities in cytokine patterns between viral and bacterial ocular diseases in the context of HIV infection, as most studies have been performed in autoimmune diseases or in immunocompetent individuals.^5–8 As CMV infection and syphilis are both able to modify the immune response,^19,20 and these are the most common ocular infections diagnosed at our center, we compared their cytokine profiles in AqH and plasma of HIV-infected individuals with these OIs. Our study showed that the AqH cytokine profile is independent in this compartment, and is different from the profile observed in plasma. This observation is common to AqH studies, and advocates the use of AqH for studying these diseases.^8,9,21,22 Individuals with CMVr had higher concentrations of GM-CSF, IL-1α, and IP-10 in AqH. A trend towards a T_H1 response cytokine was observed in both OIs, and not only in those with CMVr, as previously reported.^11,12 Plasma IL-1α, IP-10, IFN-γ, IFN-α2, IL-10, IL-7, IL-8, and G-CSF were present in both OIs, with higher concentrations in CMVr than in ocular syphilis.

We found higher concentrations of IP-10 in AqH and plasma of individuals with CMVr. IP-10 is a chemokine largely secreted by monocytes²³ that promotes chemotaxis of activated T-, B-, and antigen presenting cells to the inflamed site.²⁴ In fact, a pathogenic role for IP-10 in the recruitment and activity of T cells into the eye has been described in active uveitis.²⁵ Recent studies also suggest that higher plasma IP-10 levels correlate with lower CD4+ T cell counts and higher proportion of activated Treg cells in immune non-responders with late antiretroviral therapy initiation,²⁴ suggesting deregulated inflammation. A role for the Treg-derived IL-10 cytokine should also be considered, as we and others¹² found higher levels of this immunoregulatory cytokine in the plasma of individuals with CMVr, which has been shown to interfere with effective anti-CMV immune responses.²⁶

Although individuals with CMVr have a more advanced HIV infection than individuals with ocular syphilis, and plasma cytokine concentrations seem to be determined mostly by the stage of HIV disease and in a lesser extent by the type of ocular infection, differences found in the AqH compartment reflect a local response with unique signatures for each OI. Previous studies^11,12 and ours, have found a strong ocular inflammatory response associated with CMVr, suggesting an important role of monocytes producing innate mediators for Th1-driven anti-viral responses in the context of a severely depleted CD4+ T cell pool. Indeed, direct participation of macrophages in other CMV-related ocular diseases has been recently confirmed.²⁷ The inflammatory mechanism in the context of ocular syphilis remains to be described. Our study contributes to the elucidation of this mechanism by suggesting a milder Th1 pro-inflammatory environment with several assayed cytokines observed in significantly lower levels, both in plasma and AqH, in comparison with CMVr.

Even though a limited area of the retina is usually infected with CMV, elevated concentrations of cytokines may contribute to the severe damage and destruction of retinal layers observed in individuals with CMVr. The consequent retinal atrophy or the presence of retinal detachment, with the release or migration of autoantigens, may amplify the proinflammatory response perpetuating the intraocular inflammatory damage.²⁸ Previous studies have reported elevated levels of AqH TNF-α and IFN-γ in CMV and herpes zoster virus retinitis, and it has been considered that both cytokines may contribute to persistence of viral retinitis and retinal destruction trough FasL-mediated apoptosis.^29,30

In sum, our results support previous findings indicating elevated levels of AqH IP-10 and MCP-1 in individuals with CMVr,¹¹ and we corroborated the notion of elevated concentrations of MCP-1 and IL-8 previously reported in AqH of individuals with uveitis,³¹ as well as the intraocular production of IL-8 by the ciliary processes.³²

In contrast with a study reporting increased concentrations of IL-17 in peripheral blood and cerebrospinal fluid of individuals with neurosyphhilis,³³ we found undetectable levels of this cytokine in plasma.

An important limitation of our study was the ethical impediment for including AqH samples of a control group constituted by HIV-infected individuals without OIs. By consequence, we were unable to distinguish the cytokines associated to OIs from those associated to HIV-infection. Another limitation of this study is that CMV load levels were not assessed, considering that viral load correlates with the concentration of some cytokines in the AqH.¹²

We confirmed that the AqH cytokine profile was independent from the plasma profile in HIV-infected individuals with CMVr or ocular syphilis. There was no correlation between plasma and AqH cytokines/chemokines, except for IP-10. By consequence, we consider that AqH cytokine concentrations cannot be inferred from plasma cytokine concentrations in HIV-infected individuals with CMVr or ocular syphilis. The different profiles in AqH and plasma possibly reflect the local pathological mechanisms underlying the complex pathogenesis of uveitis, but may also be related to compartment diffusion.

Intravitreal drugs targeting different cytokines for the treatment of local inflammation in chronic non-infectious uveitis have given promising results.^34–38 Identification of the cytokines involved in the pathogenesis of different OIs in HIV-infected individuals has important implications for the development of drugs targeting these cytokines, particularly in cases of persistent inflammation after successful treatment of active ocular infections.

DECLARATION OF INTEREST

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

ACKNOWLEDGMENTS

We thank Dr. Barbara Gastel (M.D., M.P.H., Texas A&M University and Health Science Center) for valuable editing suggestions.

FUNDING

This work was supported by grants from the Mexican Government [Comisión de Equidad y Género de la Honorable Cámara de Diputados de la LXI Legislatura de México].

Additional information

Funding

This work was supported by grants from the Mexican Government [Comisión de Equidad y Género de la Honorable Cámara de Diputados de la LXI Legislatura de México].