Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Lower CXCR3 expression in both patients with neovascular AMD and advanced stages of chronic myeloproliferative blood cancers

  • Charlotte Liisborg ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Validation, Writing – original draft, Writing – review & editing

    Affiliations Department of Ophthalmology, Zealand University Hospital, Roskilde, Denmark, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark

  • Vibe Skov,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Hematology, Zealand University Hospital, Roskilde, Denmark

  • Lasse Kjær,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Hematology, Zealand University Hospital, Roskilde, Denmark

  • Hans Carl Hasselbalch,

    Roles Conceptualization, Supervision, Writing – review & editing

    Affiliations Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark, Department of Hematology, Zealand University Hospital, Roskilde, Denmark

  • Torben Lykke Sørensen

    Roles Conceptualization, Supervision, Writing – review & editing

    Affiliations Department of Ophthalmology, Zealand University Hospital, Roskilde, Denmark, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark



Peripheral T cell CXCR3 expression has been found uniquely lower in patients having neovascular age-related macular degeneration (nAMD) than in healthy individuals. The CXCR3-axis has been shown to have angiostatic and antifibrotic properties. We have recently investigated systemic markers in patients with myeloproliferative neoplasms (MPNs) because of their higher prevalence of AMD, and we have observed higher systemic chronic low-grade inflammation and immunosenescence signs in MPNs with drusen (MPNd) compared to those with normal retinas (MPNn). The MPNs evolve in a biological continuum from early cancer-stages (essential thrombocytosis, polycythemia vera) to the advanced myelofibrosis stage. Especially myelofibrosis is characterized by bone marrow angiogenesis and fibrosis, similarly to retinal observations in nAMD. We speculate if we can find lower CXCR3 expression in MPNs, particularly myelofibrosis and if differences are seen between MPNd and MPNn. We also wanted to compare expression in nAMD and intermediate (i)AMD.


Patients in this cross-sectional study were 29 nAMD, 28 iAMD, 35 MPNd, and 27 MPNn. We performed flowcytometry on blood to measure CXCR3 expression.


CD8+CXCR3 expression in nAMD was 6,1%, significantly lower than in iAMD 16%, MPNd 11%, MPNn 12% (p-values<0.05). Similar results were seen for CD4+CXCR3 expression. We also found CXCR3 expression decreasing over the MPN-continuum. For instance, in myelofibrosis, intermediate monocytes expression was 6.2%, significantly lower than 18% in ET and 18% in PV (p-values<0.05).


We find CXCR3 downregulation on T-cells and some monocyte subset in nAMD compared to iAMD, MPNd, and MPNn, in line with previous nAMD studies. We also find CXCR3 downregulation in most monocyte subsets over the MPN continuum. Systemic leukocyte CXCR3 expression could both be involved in changes seen in the retina and the bone marrow. Further understanding the CXCR3-axis in AMD and MPNs may elucidate underlying pathogenic mechanisms and reveal new targets for treatment.


Age-related macular degeneration (AMD) is a disease that affects the photoreceptors (PR), the retinal pigment epithelium (RPE), Bruch’s membrane (BM), and the choroid [1]. The disease stages are divided into early-, intermediate (iAMD)- and late AMD [2]. All AMD stages are characterized by extracellular deposits (drusen) accumulation below or above the RPE. The vision-threatening late-AMD stages are divided into two clinical entities, the often more slowly evolving geographic atrophy and the faster developing neovascular AMD (nAMD). In nAMD, angiogenesis as choroidal neovascularization (CNV) is a central feature. The abnormal blood vessel growth that seems to be driven significantly by vascular endothelial growth factor (VEGF) results in fragile vessels that leak fluid and blood into the retina with devastating visual consequences [1, 3]. Eventually, fibrosis arises from the pre-existing CNV lesions [4].

Angiogenesis is a normal physiological process needed, for instance, in wound healing and tissue repair, but it can also be an outcome of pathological conditions such as in nAMD, chronic inflammation, cancer, and ischemia. Pro- and anti-angiogenic factors control vascular homeostasis. Inflammation and angiogenesis are interconnected and involved in many pathological conditions, with cytokines as regulators of these processes, exerting effects on receptors shared by many cells throughout the body, including leukocytes and endothelial cells [57]. Chemokines are chemoattractant cytokines, small signaling proteins with diverse functions secreted by cells. The most studied function is cell migration of leukocytes, but chemokines have numerous activities and work on several cell types. For instance, chemokines and their receptors play a role in the immune cell activation and differentiation and thus the inflammatory process [8]. In addition, many members of the CC- and CXC- families of chemokines and their corresponding receptors have angiogenic and angiostatic properties. The interferon (IFN)-γ inducible chemokines CXCL9, CXCL10, and CXCL11 are strongly associated with Th1 mediated immune responses [9] and are also potent angiogenesis inhibitors functioning through the chemokine receptor 3 (CXCR3). CXCR3 is therefore involved in differentiating naïve T cells into Th1 effector cells and plays an important role in T cell function and trafficking [10]. Many chemokines also “cross-talk” with the pro-angiogenic VEGF, for instance, CXCL10-mediated inhibition of VEGF [11]. In summary, chemokines are key players in innate and adaptive immunity both in health and disease, and given the association between inflammation and the CXCR3-axis, it may not be unexpected that this system plays a role in many autoimmune diseases [1214].

We have previously shown a significantly lower CD8+ T cell CXCR3 expression in patients with nAMD than in age-matched controls [15, 16]. Also, CXCL10 has been found elevated in the retinas of patients with AMD, and choroidal endothelial cells express its receptor CXCR3 [17]. Further, mice with genetic ablation of CXCR3 show exacerbated CNV formation [18].

We have recently studied systemic alterations in patients with myeloproliferative neoplasms (MPNs) because of their increased drusen and AMD prevalence [19]. The MPNs evolve in a biological continuum from essential thrombocythemia (ET) to polycythemia vera (PV) to the advanced myelofibrosis stage (MF). The earlier stages (ET & PV) are characterized by excess production of different myeloid terminally differentiated blood cell quantities, while the late MF stage show variable platelet and leukocyte counts and even anemia. Angiogenesis and bone marrow fibrosis are less pronounced in ET and PV but are essential in MF pathogenesis [20, 21]. We have reported, in MPNs, an association between drusen presence and chronic low-grade inflammation and drusen presence and signs of an aging immune system (immunosenescence). Patients with nAMD and MPNs share several characteristics, such as angiogenesis, fibrosis, and inflammation being involved in the pathogeneses.

Because CXCR3 is found to be uniquely lower in patients with nAMD, and this receptor has been shown to have anti-angiogenic and antifibrotic properties, the research questions/purposes for this study were: 1) Can we repeat our finding of low CXCR3 expression on leukocytes in patients with nAMD when comparing to iAMD? 2) since the MPNs are a biological continuum, can we find low CXCR3 expression in patients with dominant angiogenic and fibrotic stages of disease (which nAMD also could be considered as)? 3) We also wanted to compare the CXCR3 expression in nAMD to MPNs and measure the ligands of CXCR3: CXCL9, CXCL10, and CXCL11.

Materials and methods

Study design and participants

This cross-sectional study was approved by the Region Zealand Ethics Committee, Denmark, and adhered to the Helsinki declaration’s tenets. All participants provided written and oral informed consent. The study was performed at the ophthalmology- and hematology departments, Zealand University Hospital, Roskilde. Patients in the outpatient programs in the two departments were asked to participate. The participants consisted of 30 patients with nAMD, 30 iAMD (Beckmann Classification) [2], and 63 MPNs (WHO2016 criteria) [22]; of those 35 MPNs with early or intermediate AMD (MPNd) and 28 with normal retinas (MPNn). Inclusion was done between July 2018 and November 2020 [23, 24]. Exclusion criteria were patients having other concurrent cancer, inflammatory- or autoimmune diseases. Patients were also excluded if receiving immunomodulating therapy (Ruxolitinib, interferon-a), recent anti-VEGF therapy for nAMD (within the last eight weeks), or having a CRP greater than 15. Four patients were excluded post hoc, one nAMD (CRP>15), two iAMD (fulfilled GA criteria), and one MPNn (flow cytometric analyses failed).

Retinal imaging and clinical data

We obtained stereoscopic 45° color fundus photographs centered on the macula (model TRG-NW8, Topcon). We diagnosed nAMD according to standard procedures (OCT and angiography). Pupil dilation was done before photographing (tropicamide 1%). All participants were subjected to a questionnaire about their health status, medication, and lifestyle.

Blood sampling, flow cytometry, and immunoassays

The same investigator (C.L.) drew blood from antecubital veins in ethylene-diamine-tetraacetic-acid-coated (EDTA) tubes for flow cytometry and lithium-heparin-coated tubes for CRP analysis and immunoassays. Blood used to isolate plasma was centrifuged, and the plasma was immediately stored frozen (-80°C) for later immunoassay analyses.

We performed flow cytometric analyses within four hours. A Sysmex KX-21NTM (Sysmex Corporation) measured white blood count, and from this count, we calculated the blood needed to obtain 1.0 x 106 white blood cells in the test tube. We used a 1% lysis buffer (Nordic BioSite AB) to lyse erythrocytes. We washed the cells in BD FACS Flow isotonic buffer (BD Biosciences) and centrifuged the cells at 500xg for 5 min. We repeated the washing/centrifuging process additional two times. Cells were resuspended in isotonic buffer, and we added monoclonal antibodies, leaving them to incubate for 20min in the dark at room temperature. We used one flow cytometry panel for monocytes and one for T cells, and we used corresponding fluorochrome-matched isotype controls. In the T cell panel, we used Brilliant Violet V510 CD8 IgG1κ (BioLegend, 301048), Phycoerythrin (PE)/Cyanine7(Cy7) CXCR3 IgG1 (BD biosciences, 560831), Peridinin-chlorophyll-protein(PerCP) CD4 IgG2a (R&D Systems, FAB3791C). In the monocyte panel, we used Brilliant Violet V510 CD16 IgG1κ (BioLegend, 302048), PE/Cy7 CXCR3 IgG1 (BD biosciences, 560831), Pacific Blue CD14 IgG1κ (BioLegend, 325616). The following isotype controls from BioLegend were used: Brilliant Violet V510 IgG1 (400172), PE/Cy7 IgG1 (400126), Pacific Blue IgG1 (400151), PerCP IgG2a (400250). In the final preparation step, we washed and resuspended the stained cells in isotonic buffer and analyzed the cells on a BD FACSCantoII flow cytometer (BD Biosciences) with a gating size of 100,000 singlet leukocytes. We analyzed the flow data with Kaluza Analysis software (v. 2.1; Beckman Coulter). Fig 1 shows an example of the gating strategy.

Fig 1. Gating strategy for identifying CXCR3 expression on monocytes using Kaluza software.

a) First, we identified Singlets using forward height vs. area scatter. b) Next, we identified monocytes using a forward side scatter plot. c) Monocyte subsets were identified using the markers CD14 and CD16 (classical monocytes CD14+CD16-, intermediate monocytes CD14+CD16+, and non-classical monocytes CD14-CD16+). d) Finally, we used a negative isotype control (set to 1%) to distinguish the CXCR3 positive monocytes (shown in the overlay plot). The same procedure was done for lymphocytes, using the markers CD4+ and CD8+ for identifying CD4+ and CD8+ T cells.

Immunoassays were carried out at the Technical University of Denmark (DTU). According to the manufacturer’s instructions, chemokines were quantified with multiplex immunoassays from Meso Scale Discovery. We ran the tests in duplicates and determined the mean concentration and the coefficient of variation (CV). The mean CV values were between 1.9–3.6. The plates were read on a QuickPlex SQ120 (Meso Scale Discovery).


Data statistics were done with RStudio v.4.1.1. We based our power calculation on prior comparable immunologic studies of these patients, resulting in a sample size of a minimum of 26 in each group [23, 24].

We report normally distributed data as mean and 95% confidence interval (CI), and non-normal data as median and interquartile range (IQR). Normality was assessed with histograms and QQ-plots. Age dependency and correction were assessed with linear regression or robust linear regression. Multiple regressions were used to investigate effects on outcomes for variables in Table 1. For group comparisons, we used the independent samples t-test, Wilcoxon’s rank-sum test, One-way analysis of variance (ANOVA), or Kruskal Wallis test for continuous variables. We used the Chi-squared test or Fisher’s Exact test for categorical variables. P-values less than 5% were considered statistically significant. We did not use Bonferroni corrections since the comparisons were part of our preplanned hypotheses.


Study population

Table 1 shows patient characteristics. The median age in nAMD was 77 years (IQR:71–82), significantly older than the three other groups; iAMD 73 (IQR:68–76), p = 0.034; MPNd 72 (IQR:65–76), p = 0.0040; and MPNn 69 (IQR: 62–74) p<0.001. We found no difference in sex, body mass index, smoking habits, or comorbidities between the groups. The median alcohol consumption in MPNd was seven units per week, significantly higher than nAMD with 2 (IQR:0–7), p<0.001; iAMD 3 (IQR:0–7), p = 0.019, and MPNn 2 (IQR:0–8), p = 0.021. All outcomes described below were corrected for age where appropriate and alcohol consumption or other patient characteristics mentioned above did not influence outcomes reported in this work. The MPNs consisted of 39 PV-, 17 ET-, and six MF patients, and 53 (85%) had the JAK2V617F-mutation, four (6.2%) CALR, and one (1.5%) MPL. The median JAK2V617F allele burden was 33% (IQR:11–56) in MPNd and 17% (IQR:5.6–28) in MPNn (p = 0.088). Patients receiving hydroxyurea were similar in the MPN groups (p = 0.99), and all MPN received acetylsalicylic acid or other anticoagulant therapy. Patients receiving statins were similar across all groups (p-value = 0.58).

Leukocyte CXCR3

CXCR3 expression was not dependent on age on CD4+ T cells and monocytes and subsets but decreased with age on CD8+ T cells. For CD4+ T cells and monocytes, we used the Kruskal Wallis test and Wilcoxon rank-sum test for comparisons between groups, and we used robust linear regression to compare groups for CD8+ T Cells.

The CD8+ T cell CXCR3 expression in patients with nAMD was 6.1% (4.2–8.9) significantly lower than the other groups (iAMD 16% (12–22) p = 0.0025, MPNd 11% (8.9–15) p = 0.036, MPNn 12% (8.0–17) p = 0.012) (Fig 2A). A lower CD4+CXCR3 expression was also observed when comparing nAMD to the other groups, except for MPNn (p = 0.080) (Fig 2A). The intermediate monocyte CXCR3 expression in nAMD was 11% (4.9–17), significantly lower than 17% (9.1–25) in iAMD (p = 0.029), 19% (9.8–30) in MPNd (p = 0.013) and 17% (6.2–29) in MPNn (p = 0.0060). The same tendency was seen in monocytes, classical- and non-classical monocytes with significantly lower levels in nAMD than in some of the other groups (Fig 2B). No significant differences were found between MPNd and MPNn (Fig 2).

Fig 2. Lower CXCR3 expression on leukocytes in patients with nAMD.

Median frequency (%) of leukocytes and leukocytes expressing CXCR3 in patients with neovascular AMD (nAMD), intermediate AMD (iAMD), myeloproliferative neoplasms with drusen (MPNd), and myeloproliferative neoplasms with normal retinas (MPNn) a) Expression of CXCR3 on T cells. b) Expression of CXCR3 on monocytes. Bars represent the interquartile range. Statistical comparison between groups: Kruskal Wallis or if the outcome depended on age, robust linear regression. Wilcoxon rank-sum test for multiple comparisons. Significant P-values are shown in bold.

When we investigated the MPN biological continuum (Fig 3), we found CXCR3 expression on CD8+ T cells significantly lower in PV with 11% (5.9–20) than in MF with 17% (15–24) (p = 0.011) but not than in ET with 15% (11–22) (p = 0.22). There was no difference between MPN subtypes on CD4+ T cells (p = 0.48) (Fig 3A). Fig 3B shows the CXCR3 expression on monocytes over the biological continuum. The expression on overall-monocytes and intermediate- and non-classical monocytes was in MF 4.0% (2.6–12), 6.2% (4.2–18) and 4.0 (3.4–5.2), significantly lower compared to ET with 14% (9.8–24), 18% (15–32) and 15% (9.0–29) (p = 0.047, p = 0.031, p = 0.0070, respectively). On overall monocytes and intermediate monocytes, the CXCR3 expression was also significantly lower in MF than PV with 9.5% (4.7–17) and 18% (8.8–32) (p = 0.023, p = 0.011). There was an overall tendency for decreasing levels across the biological continuum.

Fig 3. CXCR3 expression on leukocytes in MPN subtypes showing decreasing levels of CXCR3 on monocytes over the MPN biological continuum in monocytes.

Median frequency (%) of leukocytes and leukocytes expressing CXCR3 in patients with essential thrombocythemia (ET), polycythemia vera (PV), and myelofibrosis (MF) a) Expression of CXCR3 on T cells. b) Expression of CXCR3 on monocytes. Bars represent the interquartile range. Statistical comparison between groups: Kruskal Wallis or if the outcome depended on age, robust linear regression. Wilcoxon rank-sum test for multiple comparisons. Significant P-values are shown in bold.

We also investigated leukocyte percentages (Table 2). Patients with iAMD had a significantly higher lymphocyte percentage of 16% (CI:14–19) than the other groups (nAMD 11% (CI9.3–14) p = 0.040, MPNd 9.4% (CI:8.0–11) p<0.001, MPNn 10% (CI8.8–12) p<0.001). We also observed an intermediate monocyte percentage of 9.2% (CI:7.5–11) in MPNd, significantly higher than 6.3% (CI:4.7–8.5) in nAMD (p = 0.037) and 6.3% (CI: 5.2–7.6) iAMD (p = 0.0060). We found no differences in leukocyte percentages between MPN subtypes except for the lymphocyte percentage, which was higher in ET with 13% (CI:10–15) than PV with 9.3% (CI:8.0–11) (p = 0.050) and MF with 7.1% (CI:5.8–8.7) (p<0.001) (S1 File).

Table 2. Mean percentages of leukocytes in patients with nAMD, iAMD, MPNd, and MPNn.

Plasma CXCL9/-10/-11

The CXCL9- and CXCL10 levels increased with age (both p<0.001), and we, therefore, used robust linear regression to compare groups for these outcomes. We found no differences in the levels of CXCL9 and CXCL10 between any of the groups (Table 3). The levels of CXCL11 were not age-dependent, and we used the Kruskal Wallis test and Wilcoxon rank-sum test to compare groups. We found higher levels in both MPN groups than the AMD groups (nAMD-MPNd: p = 0.0050, nAMD-MPNn: p = 0.0028, iAMD-MPNd: p = 0.0026, iAMD-MPNn: p = 0.0035). There were no differences between subtypes of MPNs (ET, PV, and MF) in levels of CXCL9, CXCL10, and CXCL11 (Data not shown).

Table 3. Median plasma concentration of CXCL9, CXCL10, and CXCL11.


This study aimed to investigate leukocyte CXCR3 expression in patients with nAMD, iAMD, MPNd, and MPNn. The CXCR3 expression is interesting since it has been strongly related to angiogenesis- and fibrosis inhibition, and it is found to be lower in nAMD than in healthy controls. We wanted to investigate if this lower expression in nAMD was also seen when compared to iAMD. We also speculated if we could find a low CXCR3 expression over the MPN biological continuum, especially in the MPN subtype MF, which shows excessive bone marrow angiogenesis and fibrosis, as patients with nAMD show in the retina.


We found a lower CXCR3 expression in patients with nAMD than iAMD, MPNd, and MPNn, in CD4+ and CD8+ T cells and a tendency for a lower expression on all monocyte subsets, with a significantly lower expression on intermediate- and non-classical monocytes.

Our group previously found lower CXCR3 expression on peripheral CD8+ T cells in patients with nAMD than in a healthy control group, indicating a systemic CXCR3 dysregulation [15]. This study also finds a lower CXCR3 expression when comparing nAMD and iAMD. It would be interesting to investigate if we could find differences between iAMD and healthy controls.

The CXCR3-axis has various functions, such as regulating immune cell migration, differentiation, and activation [25, 26], and is expressed on numerous cells throughout the body, including leukocytes and endothelial cells [27].

The binding of CXCL9/10/11 to CXCR3 is known to have angiostatic and antifibrotic properties [28, 29]. For instance, CXCL10 binding to CXCR3 has been shown to reduce endothelial cell migration and the ability of these cells to form tubes [30]. CXCR3 is expressed in both human and mouse choroidal endothelial cells. A study found increased CXCL10 and CXCR3 expression in CNV tissue from mice with laser-induced CNVs. The same study found that CXCR3-deficient mice developed larger laser-induced CNV with more fluid leakage and macrophage infiltration than wild-type mice [18]. Other experimental studies on endothelial cells show that binding of CXCL10 to CXCR3 can have a counterproductive effect on VEGF stimulation of endothelial cells and decreases this activity, and further blocking of CXCR3 results in CXCL10 not being able to override the VEGF signal, which leads to angiogenesis [30, 31].

When we compared the CXCR3 expression in MPNs, we found no differences between MPNd and MPNn. However, when we investigated the biological continuum, a tendency of a lower expression across the continuum from ET/PV to MF was observed in all monocyte subsets, with significant differences between subgroups in intermediate- and non-classical monocytes. It seems that advanced stages of MPNs are also associated with a lower CXCR3 expression. Interestingly, angiogenesis and fibrosis are seen in later stages of MPN (especially MF) and in nAMD, and the two conditions simultaneously show a CXCR3 downregulation.

The CXCR3 receptor has three alternative splice variants which activate different intracellular signaling pathways. Which of the intracellular pathways are activated depends on the receptor, the binding ligand, and the type of cell or tissue studied, a phenomenon called “biased signaling” or “biased agonism”. Biased signaling adds complexity to the chemokine-receptor axis [32, 33]. Therefore, the three CXCR3 ligands are also thought to have different roles in T cell trafficking. CXCL9 and -10 seem to induce inflammation by Th1/Th17 cell polarization and recruitment, while CXCL11, which binds CXCR3 with the highest affinity, appears to induce T cell polarization into Th2 and regulatory T cells and can thereby play a role in restricting inflammation [34]. Thus, the ligand-CXCR3 axis may have opposing effects on regulating the biological T cell function. The CXCR3 expression on many different cells links cell-mediated immunity and angiogenesis. In addition to the direct effect of CXCR3 signaling on endothelial cells, the IFN-γ inducible chemokines CXCL9 and -10 induce a Th-1 type immune response by attracting CXCR3 expressing T cells (Th1 T cells), monocytes, and natural killer cells, which lead to the production of more IFN-γ, inducing increased expression of the IFN-γ inducible chemokines, which again can recruit and activate more CXCR3 expressing cells. This process of type 1 immune response occurring along with inhibition of angiogenesis (immunoangiostasis) in a positive feedback loop seems, from our results, dysregulated (lower) in nAMD and also in MPN with more advanced disease [6, 35, 36].

The systemic CXCR3 alterations could affect the retina. A study has shown that monocytes and endothelial cells can directly interact and synergistically induce the expression of CXCL10 [36]. Macrophages play an important role in immune regulation and pathogenic angiogenesis [37, 38], including the neovascularization seen in nAMD [37, 39, 40]. Tissue macrophages derive from blood monocytes, which can be classified as classical- (CD14+CD16-), intermediate (CD14+CD16+), and non-classical monocytes (CD14-CD16+). In particular, the intermediate monocytes, and especially those expressing the angiopoietin-2 receptor, Tie-2, have been characterized as highly pro-angiogenic cells, which secrete pro-angiogenic factors such as VEGF [41, 42]. Compelling evidence suggests that blood-derived monocytes can enter the subretinal space, and changes in monocytes’ expression of different receptors could influence the subretinal milieu [4345].

In addition to monocytes being able to cross the blood-retina-barrier, lymphocytes have been observed in the choroid of eyes from AMD patients [4648] and CD8 positive cells seem present to a higher degree in the macular choroid of patients with drusen [48].

We can only guess what the lower CXCR3 expression in nAMD and advanced MPN means. However, we may suggest that the low expression could result in altered T helper cell polarization and/or an imbalance between angiogenic and angiostatic factors and accompanying dysregulated angiogenesis (dysfunctional immunoangiostasis).


When we investigated the IFN-γ inducible chemokines, we found no differences in CXCL9 and -10 levels between groups, but both MPN groups had higher levels of CXCL11 than the two AMD groups. This could be interesting to investigate further in future studies. One hypothesis could be that the higher CXCL11 levels could be a compensatory mechanism for the high level of chronic inflammation seen with these diseases since, as mentioned, CXCL11 could play a part in restricting inflammation. There are limited in-vivo studies of CXCL11 functions, but studies have found elevated levels in patients with autoimmune disease [49, 50] and high levels of all three IFN-γ inducible chemokines have been associated with a poor prognosis in many types of cancers [51].

Many studies have investigated cytokine levels in the aqueous humor of patients with nAMD, and several cytokines have been found elevated, including CXCL9 and -10 [52, 53]. Further, levels of CXCL-10 and -11 in post-mortem eyes have been found to correlate with the presence of drusen [54]. Only a few studies have investigated serum/plasma levels of the IFN-γ inducible chemokines in patients with nAMD. One study found no differences in plasma levels of these chemokines between a healthy age-matched control group and patients with early AMD, nAMD, and geographic atrophy [15]. Another study found CXCL10 levels elevated in the serum of all AMD stages compared to healthy controls [17]. This study found the highest levels in intermediate AMD and a slight decrease in nAMD. The same study investigated donor’s eyes and found the CXCL10 expression increased in the RPE and is expressed by neovascular endothelial cells.


Several limitations must be kept in mind when interpreting this study’s results. The study is only observational, and further experimental studies are needed to infer causality. The groups compared are also small, especially the number of patients with MF. It is also worth remembering that the chemokines and cells discussed in this paper are part of a large, complex, and interacting system, and we are looking at a few isolated systemic differences.


We find a CXCR3 downregulation on T-cells and some monocyte subset in patients with nAMD compared to iAMD, MPNd, and MPNn, in line with previous nAMD studies [15, 16]. We also find a CXCR3 downregulation in most monocyte subsets over the MPN continuum, a disease group having increased drusen/AMD prevalence. The advanced MF stage has characteristics like nAMD with angiogenesis and fibrosis and, at the same time, low CXCR3 expression. Therefore, systemically leukocyte CXCR3 expression could both be involved in changes seen in the retina and the bone marrow.

The principal nAMD feature is CNV formation. Angiogenesis regulating mechanisms are complex, but VEGF is a dominant stimulator, and AMD is treated with anti-VEGF eye injections, which have proven effective. However, many patients still lose vision despite treatment, and the need for new targets and treatment is evident, especially therapy that can be initiated earlier in the disease progresses. Our findings could support a role of CXCR3 in AMD pathogenesis, and further understanding the CXCR3-axis and immune cell trafficking in AMD may elucidate underlying pathogenic mechanisms and reveal new treatment targets. Targeting chemokines or their receptors may be an effective therapeutic approach, as suggested for other inflammatory and neoplastic diseases [5558]. Since chemokines have cross-talk with other angiogenic factors such as VEGF, combined targeting of VEGF and chemokines may have synergistic effects.


For help with immunoassays, the authors would like to thank Sanne Schou Berger from the Technical University of Denmark.


  1. 1. Fleckenstein M, Keenan TDL, Guymer RH, Chakravarthy U, Schmitz-Valckenberg S, Klaver CC, et al. Age-related macular degeneration. Nat Rev Dis Prim 2021;7:31. pmid:33958600
  2. 2. Ferris FL, Wilkinson CP, Bird A, Chakravarthy U, Chew E, Csaky K, et al. Clinical classification of age-related macular degeneration. Ophthalmology 2013;120:844–51. pmid:23332590
  3. 3. Ambati J, Ambati BK, Yoo SH, Ianchulev S, Adamis AP. Age-related macular degeneration: Etiology, pathogenesis, and therapeutic strategies. Surv Ophthalmol 2003;48:257–93. pmid:12745003
  4. 4. Little K, Ma JH, Yang N, Chen M, Xu H. Myofibroblasts in macular fibrosis secondary to neovascular age-related macular degeneration—the potential sources and molecular cues for their recruitment and activation. EBioMedicine 2018;38:283–91. pmid:30473378
  5. 5. Romagnani P, Lasagni L, Annunziato F, Serio M, Romagnani S. CXC chemokines: the regulatory link between inflammation and angiogenesis n.d.
  6. 6. Groom JR, Luster AD. CXCR3 in T cell function. Exp Cell Res 2011;317:620–31. pmid:21376175
  7. 7. Groom JR, Luster AD. CXCR3 ligands: redundant, collaborative and antagonistic functions. Immunol Cell Biol 2011;89:207–15. pmid:21221121
  8. 8. Luster AD. Chemokines—Chemotactic Cytokines That Mediate Inflammation. N Engl J Med 1998;338:436–45. pmid:9459648
  9. 9. Rotondi M, Lazzeri E, Romagnani P, Serio M. Role for interferon-gamma inducible chemokines in endocrine autoimmunity: an expanding field. J Endocrinol Invest 2003;26:177–80. pmid:12739748
  10. 10. Kuo PT, Zeng Z, Salim N, Mattarollo S, Wells JW, Leggatt GR. The Role of CXCR3 and Its Chemokine Ligands in Skin Disease and Cancer. Front Med 2018;5:271. pmid:30320116
  11. 11. Keeley EC, Mehrad B, Strieter RM. Chemokines as mediators of tumor angiogenesis and neovascularization. Exp Cell Res 2011;317:685–90. pmid:21040721
  12. 12. Karin N. CXCR3 Ligands in Cancer and Autoimmunity, Chemoattraction of Effector T Cells, and Beyond. Front Immunol 2020;11:976. pmid:32547545
  13. 13. Lacotte S, Brun S, Muller S, Dumortier H. CXCR3, inflammation, and autoimmune diseases. Ann N Y Acad Sci 2009;1173:310–7. pmid:19758167
  14. 14. Antonelli A, Ferrari SM, Giuggioli D, Ferrannini E, Ferri C, Fallahi P. Chemokine (C-X-C motif) ligand (CXCL)10 in autoimmune diseases. Autoimmun Rev 2014;13:272–80. pmid:24189283
  15. 15. Falk MK, Singh A, Faber C, Nissen MH, Hviid T, Sørensen TL. Dysregulation of CXCR3 Expression on Peripheral Blood Leukocytes in Patients With Neovascular Age-Related Macular Degeneration. Investig Opthalmology Vis Sci 2014;55:4050. pmid:24812555
  16. 16. Singh A, Subhi Y, Nielsen MK, Falk MK, Matzen SMH, Sellebjerg F, et al. Systemic frequencies of T helper 1 and T helper 17 cells in patients with age-related macular degeneration: A case-control study. Sci Rep 2017;7:1–9.
  17. 17. Mo FM, Proia AD, Johnson WH, Cyr D, Lashkari K. Interferon γ–Inducible Protein-10 (IP-10) and Eotaxin as Biomarkers in Age-Related Macular Degeneration. Investig Opthalmology Vis Sci 2010;51:4226.
  18. 18. Fujimura S, Takahashi H, Yuda K, Ueta T, Iriyama A, Inoue T, et al. Angiostatic Effect of CXCR3 Expressed on Choroidal Neovascularization. Investig Opthalmology Vis Sci 2012;53:1999. pmid:22408007
  19. 19. Liisborg C, Nielsen MK, Hasselbalch HC, Sørensen TL. Patients with myeloproliferative neoplasms and high levels of systemic inflammation develop age-related macular degeneration. EClinicalMedicine 2020;26:100526. pmid:33089124
  20. 20. Ponce CC, de Lourdes Lopes Ferrari Chauffaille M, Ihara SSM, Silva MRR. Increased angiogenesis in primary myelofibrosis: latent transforming growth factor-β as a possible angiogenic factor. Rev Bras Hematol Hemoter 2014;36:322–8. pmid:25305163
  21. 21. Ponce CC, de Lourdes F. Chauffaille M, Ihara SSM, Silva MRR. The relationship of the active and latent forms of TGF-β1 with marrow fibrosis in essential thrombocythemia and primary myelofibrosis. Med Oncol 2012;29:2337–44. pmid:22200991
  22. 22. Barbui T, Thiele J, Gisslinger H, Kvasnicka HM, Vannucchi AM, Guglielmelli P, et al. The 2016 WHO classification and diagnostic criteria for myeloproliferative neoplasms: document summary and in-depth discussion. Blood Cancer J 2018;8:15. pmid:29426921
  23. 23. Liisborg C, Skov V, Kjær L, Hasselbalch HC, Sørensen TL. Patients with MPNs and retinal drusen show signs of complement system dysregulation and a high degree of chronic low-grade inflammation. EClinicalMedicine 2022;43:101248. pmid:35128362
  24. 24. Liisborg C, Skov V, Kjær L, Hasselbalch HC, Sørensen TL. Retinal drusen in patients with chronic myeloproliferative blood cancers are associated with an increased proportion of senescent T cells and signs of an aging immune system. Aging (Albany NY) 2021;13:25763–77. pmid:34954692
  25. 25. Strieter RM, Polverini PJ, Kunkel SL, Arenberg DA, Burdick MD, Kasper J, et al. The Functional Role of the ELR Motif in CXC Chemokine-mediated Angiogenesis. J Biol Chem 1995;270:27348–57. pmid:7592998
  26. 26. Lasagni L, Francalanci M, Annunziato F, Lazzeri E, Giannini S, Cosmi L, et al. An alternatively spliced variant of CXCR3 mediates the inhibition of endothelial cell growth induced by IP-10, Mig, and I-TAC, and acts as functional receptor for platelet factor 4. J Exp Med 2003;197:1537–49. pmid:12782716
  27. 27. Groover MK, Richmond JM. Potential therapeutic manipulations of the CXCR3 chemokine axis for the treatment of inflammatory fibrosing diseases. F1000Research 2020;9:1197. pmid:33145014
  28. 28. Tager AM, Kradin RL, LaCamera P, Bercury SD, Campanella GS V, Leary CP, et al. Inhibition of pulmonary fibrosis by the chemokine IP-10/CXCL10. Am J Respir Cell Mol Biol 2004;31:395–404. pmid:15205180
  29. 29. Angiolillo AL, Sgadari C, Taub DD, Liao F, Farber JM, Maheshwari S, et al. Human interferon-inducible protein 10 is a potent inhibitor of angiogenesis in vivo. J Exp Med 1995;182:155–62. pmid:7540647
  30. 30. Yates-Binder CC, Rodgers M, Jaynes J, Wells A, Bodnar RJ, Turner T. An IP-10 (CXCL10)-Derived Peptide Inhibits Angiogenesis. PLoS One 2012;7:e40812. pmid:22815829
  31. 31. Bodnar RJ, Yates CC, Wells A. IP-10 blocks vascular endothelial growth factor-induced endothelial cell motility and tube formation via inhibition of calpain. Circ Res 2006;98:617–25. pmid:16484616
  32. 32. Berchiche YA, Sakmar TP. CXC Chemokine Receptor 3 Alternative Splice Variants Selectively Activate Different Signaling Pathways. Mol Pharmacol 2016;90:483–95. pmid:27512119
  33. 33. Steen A, Larsen O, Thiele S, Rosenkilde MM. Biased and G Protein-Independent Signaling of Chemokine Receptors. Front Immunol 2014;5:277. pmid:25002861
  34. 34. Karin N, Wildbaum G, Thelen M. Biased signaling pathways via CXCR3 control the development and function of CD4 + T cell subsets. J Leukoc Biol 2016;99:857–62. pmid:26657511
  35. 35. Kasama T, Muramatsu M, Kobayashi K, Yajima N, Shiozawa F, Hanaoka R, et al. Interaction of monocytes with vascular endothelial cells synergistically induces interferon gamma-inducible protein 10 expression through activation of specific cell surface molecules and cytokines. Cell Immunol 2002;219:131–9. pmid:12576031
  36. 36. Balestrieri ML, Balestrieri A, Mancini FP, Napoli C. Understanding the immunoangiostatic CXC chemokine network. Cardiovasc Res 2008;78:250–6. pmid:18252760
  37. 37. Sunderkötter C, Steinbrink K, Goebeler M, Bhardwaj R, Sorg C. Macrophages and angiogenesis. J Leukoc Biol 1994;55:410–22. pmid:7509844
  38. 38. Owen JL, Mohamadzadeh M. Macrophages and chemokines as mediators of angiogenesis. Front Physiol 2013;4:159. pmid:23847541
  39. 39. Chen M, Lechner J, Zhao J, Toth L, Hogg R, Silvestri G, et al. STAT3 Activation in Circulating Monocytes Contributes to Neovascular Age-Related Macular Degeneration. Curr Mol Med 2016;16:412–23. pmid:27009107
  40. 40. Lim LS, Mitchell P, Seddon JM, Holz FG, Wong TY. Age-related macular degeneration. Lancet (London, England) 2012;379:1728–38.
  41. 41. Zawada AM, Rogacev KS, Rotter B, Winter P, Marell R-R, Fliser D, et al. SuperSAGE evidence for CD14++CD16+ monocytes as a third monocyte subset. Blood 2011;118:e50–61. pmid:21803849
  42. 42. Dalton HJ, Armaiz-Pena GN, Gonzalez-Villasana V, Lopez-Berestein G, Bar-Eli M, Sood AK. Monocyte subpopulations in angiogenesis. Cancer Res 2014;74:1287–93. pmid:24556724
  43. 43. Gupta N, Brown KE, Milam AH. Activated microglia in human retinitis pigmentosa, late-onset retinal degeneration, and age-related macular degeneration. Exp Eye Res 2003;76:463–71. pmid:12634111
  44. 44. Guillonneau X, Eandi CM, Paques M, Sahel J-A, Sapieha P, Sennlaub F. On phagocytes and macular degeneration. Prog Retin Eye Res 2017;61:98–128. pmid:28602950
  45. 45. Lad EM, Cousins SW, Van Arnam JS, Proia AD. Abundance of infiltrating CD163+ cells in the retina of postmortem eyes with dry and neovascular age-related macular degeneration. Graefes Arch Clin Exp Ophthalmol 2015;253:1941–5. pmid:26148801
  46. 46. Penfold PL, Killingsworth MC, Sarks SH. Senile macular degeneration. The involvement of giant cells in atrophy of the retinal pigment epithelium. Invest Ophthalmol Vis Sci 1986;27:364–71. pmid:3949464
  47. 47. Penfold P, Killingsworth M, Fraco SS. An ultrastructural study of the role of leucocytes and fibroblasts in the breakdown of Bruch’s membrane. Aust J Opthalmology 1984;12:23–31. pmid:6732655
  48. 48. Ezzat M-K, Hann CR, Vuk-Pavlovic S, Pulido JS. Immune cells in the human choroid. Br J Ophthalmol 2008;92:976–80. pmid:18577650
  49. 49. Han JH, Suh C-H, Jung J-Y, Ahn M-H, Han MH, Kwon JE, et al. Elevated circulating levels of the interferon-γ–induced chemokines are associated with disease activity and cutaneous manifestations in adult-onset Still’s disease. Sci Rep 2017;7:46652. pmid:28436448
  50. 50. Antonelli A, Ferrari SM, Frascerra S, Di Domenicantonio A, Nicolini A, Ferrari P, et al. Increase of Circulating CXCL9 and CXCL11 Associated with Euthyroid or Subclinically Hypothyroid Autoimmune Thyroiditis. J Clin Endocrinol Metab 2011;96:1859–63. pmid:21470996
  51. 51. Lee K-S, Chung W-Y, Park J-E, Jung Y-J, Park J-H, Sheen S-S, et al. Interferon-γ-Inducible Chemokines as Prognostic Markers for Lung Cancer. Int J Environ Res Public Health 2021;18. pmid:34501934
  52. 52. Minaker SA, Mason RH, Lahaie Luna G, Bapat P, Muni RH. Changes in aqueous and vitreous inflammatory cytokine levels in neovascular age‐related macular degeneration: a systematic review and meta‐analysis. Acta Ophthalmol 2021;99:134–55. pmid:32602185
  53. 53. Liu F, Ding X, Yang Y, Li J, Tang M, Yuan M, et al. Aqueous humor cytokine profiling in patients with wet AMD. Mol Vis 2016;22:352–61. pmid:27122966
  54. 54. Jiang K, To E, Cui JZ, Cao S, Gao J, Matsubara JA. Drusen and Pro-inflammatory Mediators in the Post-Mortem Human Eye. J Clin Exp Ophthalmol 2012;3:208. pmid:24977103
  55. 55. Jacquelot N, Duong CPM, Belz GT, Zitvogel L. Targeting Chemokines and Chemokine Receptors in Melanoma and Other Cancers. Front Immunol 2018;9:2480. pmid:30420855
  56. 56. Turner JE, Steinmetz OM, Stahl RA, Panzer U. Targeting of Th1-associated chemokine receptors CXCR3 and CCR5 as therapeutic strategy for inflammatory diseases. Mini Rev Med Chem 2007;7:1089–96. pmid:18045212
  57. 57. Hensbergen PJ, Wijnands PGJTB, Schreurs MWJ, Scheper RJ, Willemze R, Tensen CP. The CXCR3 targeting chemokine CXCL11 has potent antitumor activity in vivo involving attraction of CD8+ T lymphocytes but not inhibition of angiogenesis. J Immunother 2005;28:343–51. pmid:16000952
  58. 58. Sorensen TL. Targeting the chemokine receptor CXCR3 and its ligand CXCL10 in the central nervous system: potential therapy for inflammatory demyelinating disease? Curr Neurovasc Res 2004;1:183–90. pmid:16185193