Circulating LOXL2 Levels Reflect Severity of Intestinal Fibrosis and GALT CD4+ T Lymphocyte Depletion in Treated HIV Infection

Background: Incomplete immune reconstitution may occur despite successful antiretroviral therapy (ART). Gut-associated lymphoid tissue (GALT) fibrosis may contribute via local CD4+ T lymphocyte depletion, intestinal barrier disruption, microbial translocation, and immune activation. Methods: In a cross-sectional analysis, we measured circulating fibrosis biomarker levels on cryopreserved plasma from adult HIV-infected (HIV+) SCOPE study participants on suppressive ART who also had fibrosis quantification on recto-sigmoid biopsies. Relationships among biomarker levels, clinical and demographic variables, GALT lymphoid aggregate (LA) collagen deposition, and LA CD4+ T lymphocyte density were analyzed using simple regression. Biomarker levels were also compared to levels in HIV+ viremic SCOPE participants and a convenience sample of HIV-uninfected (HIV-) samples. Results: HIV+ aviremic participants (n = 39) were 92% male and 41% non-white, with median age 48 years, CD4+ T lymphocyte count 277 cells/mm3, and 17 years since HIV diagnosis. Most biomarkers were lower in HIV− (n = 36) vs HIV+ aviremic individuals, although CXCL4 levels were higher. HIV+ viremic individuals (N = 18) had higher median TGF-ß3, CIC-C1Q, and TIMP-1 (P < 0.05) and lower LOXL2 levels (P = 0.08) than HIV+ aviremic individuals. Only higher LOXL2 levels correlated with more GALT collagen deposition (R = 0.44, P = 0.008) and lower LA CD4+ T lymphocyte density (R = −0.32, P = 0.05) among aviremic individuals. Conclusions: Circulating LOXL2 levels may be a noninvasive measure of intestinal fibrosis and GALT CD4+ T lymphocyte depletion in treated HIV infection. LOXL2 crosslinks elastin and collagen, and elevated LOXL2 levels occur in pathologic states, making LOXL2 inhibition a potential interventional target for intestinal fibrosis and its sequelae.


INTRODUCTION
Despite control of HIV replication with antiretroviral therapy (ART), long-term failure to restore CD4 + T lymphocyte numbers has been reported in HIV-infected (HIV+) individuals [1,2] Additionally, poorer immune reconstitution has been associated with the development of non-AIDS comorbidities and decreased lifespan among HIV+ persons [3]. The pathophysiology of failure to achieve immunologic recovery is not fully understood; however, collagen deposition and fibrosis in lymphoid tissues, as observed in gut-associated lymphoid tissue (GALT) during HIV infection, may contribute to incomplete immune reconstitution by limiting GALT plasticity and immune cell homeostasis [4,5].
Fibrosis occurs when the normal healing response to local inflammation becomes perturbed, resulting in an imbalance between deposition and degradation of the extracellular matrix (ECM) [6]. Fibroblasts in tissues produce ECM components such as collagen and hyaluronic acid (HA). Anti-fibrotic and pro-fibrotic mediators such as chemokines, cytokines, and growth factors (especially transforming growth factor [TGF]-β 1 ) [7] direct tissue remodeling. ECM turnover also involves the innate and adaptive immune systems through production of inflammatory mediators [8]. Over time and in the face of a persistent inflammatory stimulus such as HIV infection, progressive fibrosis occurs that may ultimately compromise organ function.

Study Population
We utilized data and plasma samples from participants enrolled in the Observational Study of the Consequences of the Protease Inhibitor Era (SCOPE), a longitudinal cohort of HIV+ and HIV-uninfected adults followed in San Francisco, California. A subset of 73 SCOPE participants underwent recto-sigmoid biopsies and had stored plasma available for biomarker assessment. For this cross-sectional analysis, the primary population was further restricted to the subset of participants (n = 39) with HIV-1 RNA < 50 copies/mL. As secondary endpoints, a group of viremic (HIV-1 RNA > 50 copies/mL) HIV+ individuals (n = 18) from the SCOPE study and a convenience sample of unmatched HIV-uninfected (HIV−) individuals (n = 36) were used for comparison analysis of biomarker values.
The SCOPE cohort was approved by the University of California-San Francisco (UCSF) Committee on Human Research. All subjects provided informed consent prior to enrollment, and the human experimentation guidelines of UCSF were followed in the conduct of clinical research.

Specimen collection and analysis
Rectal biopsy specimens were stained with Masson trichrome to identify collagen fibers and with CD4 antibody, as previously described [13]. Collagen abundance and the CD4 + T lymphocyte population were assessed using quantitative image analysis.

Clinical parameters
Clinical and demographic data and medical and medication history were extracted from the SCOPE database.

Circulating fibrosis biomarkers
Levels of fibrosis biomarkers were measured on EDTA plasma by ELISA or multiplex assay using commercially available kits. HA (Echelon Biosciences, Salt Lake City, UT), LOXL2 (R&D Systems, Minneapolis, Minnesota), and CICP and CIC C1Q (Quidel, San Diego, CA) were measured by ELISA. PAI-1, PCSK9, TGF-ß 1 -3 , MMP-2, MMP-9, TIMP-1, CXCL4, Chi3L1 (all R&D Systems, Minneapolis, MN) were measured using Luminex xMAP technology. All analytes except LOXL2 were measured according to the manufacturers' instructions. For LOXL2, the ELISA plates were coated overnight with capture antibody at room temperature. The plate was blocked with reagent diluent 1 (R&D Systems, Minneapolis, MN) for 1 hour and washed. Plasma was diluted 1:10 in PBS-Tween 10 and added to the plate. After incubating for 2 hours at 37° C and washing, detec-tion antibody was added for 2 hours at 37° C. The plate was washed, streptavidin was added and the plate was incubated for 20 minutes at room temperature. After washing, TMB was added (Sigma-Aldrich Corp., St. Louis, MO), the plate was incubated for 20 minutes at room temperature, and sulfuric acid was added to stop the reaction. The plate was read at 450 nm.

Statistical methods
Continuous variables are presented as medians and interquartile ranges (IQR), and nominal data are described as absolute values or percentages. Statistical analyses were performed using the Mann-Whitney U-test. Associations were assessed using simple regression and Spearman correlation coefficients. Significance was defined as two-sided α ≤ 0.05. All analyses were exploratory, without adjustment for multiple testing.

Biomarker level comparisons between HIV+ aviremic, HIV+ viremic and HIV-participants
Fibrosis biomarkers were also compared between viremic and aviremic participants (

DISCUSSION
In this group of HIV+ aviremic individuals on suppressive ART, GALT fibrosis was universally present, with more fibrosis associated with lower CD4 + T lymphocyte percentage in LA. Additionally, we found significant correlations among GALT collagen deposition, LA CD4 + T lymphocyte frequency, and circulating LOXL2 levels.  HIV+ individuals may display an incomplete immune restoration despite control of HIV replication. Interestingly, all participants had a long history of HIV infection and our sample was enriched for those who persistently had CD4 + T lymphocyte counts < 350 cells/mm 3 (69%). This population can therefore be considered to consist primarily of immunologic non-responders (median CD4 + T lymphocyte count 277 cells/mm 3 ), and the lack of peripheral immune restoration may partially result from alterations of intestinal immunity. In the first weeks of HIV infection, gut mucosal CD4 + T lymphocyte depletion occurs and, with disruption of the mucosal barrier, permits increased microbial translocation. Over time, this leads to chronic stimulation of the adaptive and innate immune systems, perpetuating gut damage, local HIV replication, and CD4 + T lymphocyte depletion, culminating in fibrosis [14, [15]. Simultaneously, the lymphatic tissue is affected by structural remodeling, from hyperplasia to follicular involution [16,17] and/or fibrosis [13]. In this virologically suppressed population, we reported a high level of collagen deposition in all biopsy samples and a negative correlation between LA CD4 + T lymphocyte frequency and LA collagen deposition, consistent with previously described results from the complete SCOPE biopsy population [13]. These findings support GALT fibrosis as a contributor to incomplete CD4 + T lymphocyte cell restoration in chronic HIV infection despite suppressive ART [13,14].
This study is the first to explore the relationship between gut collagen deposition and circulating fibrosis biomarkers among HIV+ individuals on suppressive ART. Interestingly, we found that LOXL2 levels increase with LA collagen deposition and are inversely correlated with CD4 + T lymphocyte frequency in LA. LOXL2 is a metalloenzyme that plays an important role in the progression of fibrotic disease by promoting collagen and elastin cross-linkage, which is essential for the tensile strength of the ECM [18]. Interestingly, while LOXL2 plays an essential role in angiogenesis, wound healing, and scar formation, [19,20] it is not thought to be over-expressed in normal hosts [9,21] and its upregulation has been associated with the development of fibrotic, hematologic, and malignant diseases [9,10,19,[21][22][23][24][25].
Thus, our findings suggest a dynamic GALT remodeling process that may prevent complete immune restoration on ART. This hypothesis is supported by the fact that we also observed a strong positive relationship between duration of HIV infection and TIMP-1 levels, suggesting chronic and potentially escalating tissue remodeling [26]. We also observed significantly higher MMP-2 levels and MMP-2:TIMP-1 ratios in HIV+ individuals with CD4 + T lymphocyte counts < 350 cells/mm 3 , indicative of ongoing ECM remodeling [27]. While high MMP-2 and TIMP-1 levels have been associated with severity of liver fibrosis, [28,29] in this cohort levels did not differ by viral hepatitis co-infection status (data not shown).While we did not observe an association between circulating LOXL2 levels and circulating CD4 + T lymphocyte counts, the circulating CD4 + cells pool is derived from a number of tissues that may not be affected/equally affected by fibrosis, preventing a strong correlation.
In an exploratory subanalysis, comparison of fibrosis biomarker levels between aviremic and viremic HIV+ individuals demonstrated a trend toward lower LOXL2 levels in the viremic group. This result could be explained by the significantly higher nadir CD4 + T lymphocyte values among viremic HIV+ individuals, reflecting less immunologic damage and, potentially, lower inflammation and fibrosis in this group. Additionally, most fibrosis biomarkers were significantly lower in HIV− group compared to the HIV+ aviremic group, as expected. While CXCL4 levels were high-er among HIV− persons, this was also expected: we previously demonstrated among Multicenter AIDS Cohort participants that CXCL4 levels are lower among persons with failure to immune reconstitute on ART, [30] and in vitro data suggest that lower CXCL4 levels in HIV+ persons may be a marker of successful immune evasion by the virus [31]. Thus, persons with lower CXCL4 levels on suppressive ART may be reflective of greater immunologic devastation. Additionally, CXCL4 originates from activated platelets, and circulating levels are elevated in atherosclerosis and cardiovascular disease, [32,33] malignancy, [34][35][36] and systemic sclerosis [12,37]. Thus, many factors affect CXCL4 levels.
Finally, no significant difference in LOXL2 levels was observed between HIV− and HIV+ aviremic participants, which is not surprising as, given that LOXL2 upregulation is associated with pathologic states, levels are low in most persons. Notably, the HIV− group was a sample of stored plasma from unmatched HIV− individuals with unknown medical histories and no GALT biopsy data, so this finding must be interpreted with caution. Additionally, circulating LOXL2 levels have not been measured in many clinical cohorts, so the effect of factors such as age, sex, obesity, cardiovascular disease, and liver disease on LOXL2 levels are unknown. Thus, we believe that the lack of statistical difference in levels between HIV+ aviremic persons and unmatched HIV− persons does not take away from the strong relationship observe between circulating LOXL2 levels and the extent of GALT fibrosis and CD4 + T lymphocyte depletion observed in HIV+ aviremic adults.
Our study has several limitations: The cross-sectional design and small sample size may have prevented us from observing additional statistically significant relationships between other circulating fibrosis biomarkers and GALT collagen deposition, particularly for markers that may fluctuate more widely within individuals and/or have higher coefficients of variation. Additionally, we were not powered to look at gender differences in the observed relationships, as the vast majority of participants with available samples were males. The majority of our participants also identified as being of white race. Thus, our results cannot be generalized to women and/or the global HIV+ population. Finally, HIV− control samples came from an unmatched convenience sample of persons without clinical profiling.
In conclusion, LOXL2 levels increase with GALT lymphoid aggregate collagen deposition and are inversely correlated with lymphoid aggregate CD4 + T lymphocyte frequency in HIV+ adults on suppressive ART. Further studies are needed to assess the utility of circulating LOXL2 as a noninvasive marker of fibrotic tissue burden in treated HIV infection. Finally, specific inhibitors of LOXL2 should be explored for the prevention and treatment of fibrotic disease in chronic HIV infection.