Chronic immune activation underlies multiple pathways along the course of HIV-1 infection, and a persistent inflammatory state is considered a driver of morbidity (cardiovascular, neurocognitive, and pulmonary diseases, neoplasia, genomic instability, immunosenescence, and premature aging) and mortality (1, 2). Interferon-inducible protein-10 (IP-10), also known as C-X-C motif chemokine ligand (CXCL)-10, and monokine induced by interferon-gamma (MIG), also known as CXCL-9, are pro-inflammatory chemokines originated from circulating myeloid cells (mainly monocytes and dendritic cells) and lymphoid organs (largely from the small intestine) in response to interferon. Both cytokines are ligands for the C-X-C chemokine receptor (CXCR)3, which is highly expressed on memory CD4+ T cells, a major cell target for HIV-1 (3). In addition, IP-10 can recruit susceptible T cells to lymph nodes, contributing to immune activation and dysfunction as well as helping to establish viral reservoirs (4–6). These chemokines have been previously explored in persons with HIV-1 (PWH) both in early and chronic infection (4–15). IP-10 and MIG are upregulated early after primary HIV-1 infection. IP-10 begins to increase on approximately day 6 after the first positive viral load assessment, which is a more rapid response than that observed for MIG (levels increase one month after infection) (7, 8, 11, 12). Indeed, IP-10 has been assessed as an accurate screening tool to detect acute HIV-1 infection in resource-limited settings (8). Elevated IP-10 levels are associated with rapid loss of CD4+ T cells upon HIV-1 infection and faster disease progression, even in individuals with CD4+ T cell count above 500 cells/mm³ (4, 11, 12, 16). Interestingly, we previously observed higher IP-10 and MIG levels in elite controllers who later lost viral control compared with persistent controllers, highlighting the interest of these cytokines as potential biomarkers to predict loss of viremic control (17).

After HIV-1 infection, a strong positive correlation of IP-10 with HIV-1 RNA and a negative correlation with CD4+ T cell count has been reported (4, 18). In turn, after antiretroviral treatment (ART) initiation, there is significant parallelism of plasma IP-10 and MIG levels decrease with HIV-1-RNA suppression and CD4 cell count recovery (15), in such a way that persistent levels of IP-10 can be associated with a CD4+ T cell count that fails to recover (19). However, no studies have assessed their dynamics after treatment initiation with integrase inhibitors.

We aimed to assess the dynamics in plasma IP-10 and MIG levels throughout the study period in a well-characterised cohort of PWH initiating current antiretroviral regimens including integrase inhibitors. Analyses were focused on particular scenarios of interest, stratifying according to age, AIDS-defining events, and CD4+ T cell count at baseline and HIV-1 RNA at month 12 after ART initiation.

In this prospective cohort analysis, IP-10 and MIG plasma levels showed an early, significant, and consistent decline just a few days after ART initiation and throughout 48 months.

There was significant global parallelism of plasma IP-10 and MIG decreases with HIV-1-RNA suppression and CD4+ T cell count recovery.

These findings are concordant with those previously reported in a French cohort of ART-naïve individuals, but where IP-10 and MIG plasma levels were measured only at baseline and 24 months after ART initiation (15). We were able, in addition, to show a significant reduction of these biomarkers at very early time points (20 days, specifically for IP-10), therefore recognizing both IP-10 and MIG as early ART response surrogate biomarkers. This strongly suggests that these biomarkers are impacted mainly by unsuppressed HIV-1 viremia and they decrease rapidly paralleling the significant HIV-1 RNA decay seen in this cohort just 1 month after ART initiation. The CD4 + T cell count recovery took a longer time, approximately 6 months.

At baseline, we found significantly higher IP-10 and MIG levels in PWH aged ≥ 50 years. A French cohort (n=139) reported similar baseline levels of IP-10 and MIG among age strata, but groups were defined as those aged ≤35 years, those with ages ranging from 35 to 44 years, and those aged ≥44 years (15). Of note, the number of individuals aged ≥50 years included in the upper stratum (>44 years) is unknown. Thus, the inclusion of individuals with ages ranging from 44 to 49 years in the same group could have biased the overall analysis. In our study, after ART initiation, both cytokines showed a sharp reduction in PWH aged ≥50 years, and at month 12, IP-10 levels matched those of PWH aged <50 years. However, the MIG level decline was slower and differences were still significant at month 12. Older age was associated with higher levels of IP-10 at month 24 after ART initiation (OR 1·6 per 10 years older, p=0·047, in multivariate analysis) in the previous French cohort (15). In addition, in an Australian cohort, plasma IP-10 levels were more likely to remain higher in older individuals with comorbidities (frailty, hypertension, and coronary heart disease), which correlates with other traditional cardiovascular disease surrogate plasma biomarkers, including C-reactive protein, IL-6, and fibrinogen (21). Importantly, older age ranges were not defined in either of these cohorts. Although these data could suggest that persistent increases in IP-10 would not be impacted by age by itself, but by existing comorbidities, in our multivariate analysis, older age (≥50 years) and comorbidities were not associated with persistent high levels of IP-10 at month 12 after ART initiation, and neither was it associated with high levels of MIG. The higher rates of HIV-1 RNA suppression achieved with current ART regimens could therefore overcome baseline differences seen in previous cohorts.

In our study, PWH with CDC stage C had higher baseline IP-10 and MIG levels than those with CDC stage A. According to our data, IP-10 increases have been described in PWH with opportunistic infections (tuberculosis and cryptosporidiosis) (7). Baseline levels of IP-10 (but not MIG) were significantly higher in individuals with AIDS-defining events in the aforementioned French cohort, and this condition was not associated with subsequent persistently elevated levels of both cytokines (non-adjusted analysis) at month 24 after ART initiation (15). In our cohort, we were able to find a sharp decrease in both cytokines upon ART initiation in PWH with baseline CDC stage C, and no significant differences were subsequently observed between CDC stages C vs. A, at month 6 after ART initiation. However, CDC stage C was identified as an independent variable associated with a risk of having high levels of IP-10 at month 12 after ART initiation, in the adjusted multivariate analysis.

Interestingly, we observed that PWH with CD4+ count <350 cells/mm³ at ART initiation had higher baseline IP-10 and MIG levels than those with CD4+ count ≥350 cells/mm³. The difference was significant for IP-10 and borderline significant for MIG. Upon ART initiation, both cytokine levels decreased and matched those of PWH with CD4+ count ≥350 cells/mm³ at month 6 and thereafter. A Brazilian cohort found that individuals with CD4+ count <350 cells/mm³ at ART initiation showed higher IP-10 levels compared to those with CD4+ count >350 cells/mm³. In that cohort, individuals who initiated ART with CD4+ count >350 cells/mm³ achieved normalization of plasma IP-10 to levels similar to those of healthy individuals after undergoing ART for a median of 12 months. However, the evolution of IP-10 at 12 months was not compared between PWH with CD4+ count < or > 350 cells/mm³ at ART initiation (22).

In our study, there were no differences in IP-10 and MIG plasma levels among individuals with TND or TD HIV-1 RNA below 20 copies/mL or those with low-level viremia (20-199 copies/mL) at month 12 after ART initiation. This is in agreement with data retrieved from a Swedish case-control study, where IP-10 did not show significant differences between individuals with detected viremia (50-999 copies/mL) and those with persistent virological suppression (HIV-1 RNA <50 copies/mL) after 12 months following ART initiation (23). However, we had limited power in the number of individuals with these characteristics, which precludes a stronger conclusion regarding this association.

Interestingly, on day 10 after ART initiation, MIG (but not IP-10) levels showed a significant decrease in individuals who initiated an INSTI-based regimen (vs. a b-PI-based one). This could be influenced by the well-documented significantly faster initial HIV-1 RNA decay kinetics of INSTIs (24, 25). Importantly, this suggests that MIG is more impacted than IP-10 by this initial plasma HIV-1 RNA decay under INSTIs, and more influenced by that than by the CD4+ T cell count recovery or homeostasis. Even though the differences were only significant at day 10, we saw slower decreases in MIG values with b-PI, which remained numerically higher than those in INSTI-treated individuals up to 12 months. In a French cohort, the decline in IP-10 and MIG levels was significantly smaller with ritonavir-boosted atazanavir than with efavirenz, while no significant difference was found between ritonavir-boosted lopinavir and efavirenz at 24 months after initiating ART (26). However, data were lacking regarding currently recommended INSTI as a third drug. In this regard, in an Italian cohort that divided the individuals into groups according to the third drug at the time of enrollment, the decline in IP-10 plasma levels was significantly lower among individuals receiving non-nucleoside reverse transcriptase inhibitors compared to the b-PI and INSTI group after at least 12 months following ART initiation (27).

Despite both IP-10 and MIG being pro-inflammatory biomarkers, the subtle differences found in their plasma dynamics in the present study could suggest potentially diverse pathophysiological underlying mechanisms. In contrast to MIG, IP-10 (known as being impacted as prematurely as 6 days after primary HIV-1 infection) showed significant declines as early as 20 days after ART initiation. Intriguingly, IP-10 levels remained high at month 12 in PWH with CDC-stage C at HIV diagnosis. A higher reservoir size (partly mediated by IP-10) (6, 28) could contribute to these high IP-10 levels in this subset of individuals. In this regard, it has been postulated that IP-10 could be produced as part of a second wave of cytokine responses, following toll-like receptor (TLR)7/8 stimulation and activation of antigen-presenting cells, though not to the same extent as following primary virus stimulation (12).

Altogether, these results suggest PWH who start ART on time (CD4+ T cell count ≥350 cells/mm³) and achieve levels of HIV-1 viremia below 200 copies/mL after undergoing ART for 12 months, have limited levels of chronic inflammation to impact MIG and IP-10 12 months after ART initiation. Nearly half of PWH in this cohort had CD4+ counts <350 cells/mm3 at baseline. Initiating ART during the first year of HIV infection reduces numerous biomarkers of inflammation and immune activation that are impacted by different inflammatory pathways. This response is more remarkable during acute HIV infection, potentially because of an early effect on attaining a smaller HIV-1 latent reservoir (partly mediated through IP-10) (6) and viral diversity (28). A baseline CDC-stage C is associated with a greater HIV reservoir and higher levels of persistent inflammation, with high IP-10 levels at month 12 after ART initiation in our study. These data suggest that IP-10 would be more impacted than MIG by a greater HIV reservoir. In turn, MIG seems to be more impacted than IP-10 by initial plasma HIV-1 RNA decay, as evidenced by a greater decline with INSTI-based regimens (vs. a b-PI-based one).

Overall, these findings reinforce the recommendation of early ART, prioritizing INSTI-based regimens, resulting in a decrease in chronic immune activation and inflammation.

At baseline, PWH aged >50 years, a CDC-stage C, or CD4+ T count <350 cells/mm³ had higher levels of both IP-10 and MIG levels. All these variables are associated with late HIV diagnosis. This finding reinforces the need to achieve timely diagnosis of individuals with occult or unknown HIV infection. This study adds information regarding early and late time points in the dynamics of IP-10 and MIG plasma levels, as well as in current scenarios in clinical practice: PWH aged ≥50 years, INSTI-based ART, and low-level and TND viremia.

Further analyses are required on the potential impact on plasma IP-10 and MIG and subsequent chronic immune dysfunction of new ART regimens, including dolutegravir-based two-drug regimens and long-acting formulations, as well as in people with persistent immune discordance.

Our study has limitations. The small sample size could underpower some of the analyses, especially those evaluating particular scenarios. Despite internal quality control and the prospective design, in a real-life cohort study, errors and heterogeneity in data gathering and recording cannot be ruled out. Finally, IP-10 and MIG could be impacted by other simultaneous uncontrolled conditions, such as viral or bacterial coinfections, which could act as potential confounding factors.

In conclusion, in this prospective cohort, IP-10 and MIG arise as soluble plasma biomarkers that showed an early, significant, and consistent decline in their plasma levels just a few days after ART initiation, paralleling the HIV-1 RNA decay and CD4+ T cell recovery. PWH with ages ≥50 years, CDC stage C, or PWH with CD4+ count <350 cells/mm³ had higher baseline IP-10 and MIG levels, with sharp subsequent declines. All of them matched their counterparts several months following ART initiation. PWH treated with INSTI-based regimens had faster declines of MIG as early as day 10. Overall, these findings suggest that both IP-10 and MIG are sensitive biomarkers of early antiretroviral treatment response, and are mainly impacted by plasma HIV-1 viremia, low CD4+ counts, and late ART initiation. Starting ART with CD4+ T cell count >350 cells/mm³ had limited impact on MIG or IP-10 levels. Once ART has been initiated, having TD HIV-1 RNA below 20 copies/mL or even low-level viremia (<200 copies/mL) did not impact MIG or IP-10 levels. These issues merit further investigation in larger cohorts and evaluation of new antiretroviral treatment strategies.

The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

The study was approved by the regional Ethic Committee (CAEIG), an external scientific and ethics committee of the HIV Biobank (Spanish HIV/AIDS Research Network).The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

HÁ: Conceptualization, Writing – original draft, Writing – review & editing. AG-V: Writing – review & editing. AM: Writing – review & editing. AS-A: Writing – review & editing. BC-C: Methodology, Writing – review & editing. AP-G: Writing – review & editing. JA-D: Methodology, Writing – review & editing. IM-B: Methodology, Writing – review & editing. MG-R: Writing – review & editing. SM: Writing – review & editing. TA: Writing – review & editing. MM-A: Writing – review & editing. EB: Writing – review & editing. CG: Writing – review & editing. JML: Conceptualization, Writing – original draft, Writing – review & editing. EP: Conceptualization, Writing – original draft, Writing – review & editing.

Please see the Supplementary file Data Sheet 2 for the full list of members.