**3. Inflammatory biomarkers**

A biomarker has been defined as "a biological molecule found in blood, or other bodily fluids, or tissue which represents a sign of a normal or abnormal process of a condition or disease" [18]. Due to the rapid advancement in engineering, methodology and equipment, researchers are able to generate large amounts of data analyzing only small amounts of samples, using mass spectrometry techniques. In DR studies, several kinds of samples were used, such as tear, cornea, aqueous humor (AH), lens, vitreous humor (VH), retina and serum [19]. Although VH is in direct contact with the retina, it requires a highly invasive procedure, whereas the tears can be collected using non-invasive methods but interact indirectly with the retina [19].

### **3.1 Local laboratory biomarkers**

### *3.1.1 Tears*

The tears are composed of an aqueous solution of water, lipids, proteins and salt. The major proteins are lactoferrin, lysozyme, serum albumin, secretory immunoglobin A, lipocalin and lipophilin [20]. One study identified 1526 proteins using protemics approaches, most of them being shown to respond to insults such as cataract surgery or diseases like DM [21, 22]. Mainly because it contains proteins in high concentration (about 8 μg/μl) and it is easy to collect, tears have been a fluid of interest for many studies [21]. Herber et al. was one of the first to demonstrate that DM patients' tear composition is different from control subjects [23]. Azzaralo et al. found increased levels of lactotransferin, also known as lactoferrin, a glycoprotein with immunomodulatory, anti-microbial and anti-tumor effects [24]. Increased levels of monocyte chemoattractant protein-1 (MCP-1) were also identified in DM with or without DR. A previous study demonstrated elevated Interleukin-1 receptor anatagonist (IL-1ra) levels, which represent an anti-inflammatory factor inhibiting pro-inflammatory activity of IL-1beta [25]. A positive correlation between IL-6 levels and DR stage was found. IL-6 is involved in synthesis and the release of acute phase reactants and matrix metalloproteinase 9 (MMP-9), decreased tear production, having also a role in the release of IL-1β [26]. TNF-α is a cytokine able to induce the disruption of BRB by loosening the tight junctions between endothelial cells and between RPE cells, with a role in the leukocytes adhesion to the endothelium [13]. TNF-α level was increased in patients with PDR (13.5 pg./mL) compared to NPDR (2.8 pg./mL) [27]. The tear sample can be obtained using two techniques: the direct one using aspiration with the help of a micropipette or a microcapillary tube placed in the inferior fornix, or the indirect one, with the help of a Schirmer test strip or cellulose sponge.

#### *3.1.2 Aqueous humor*

Aqueous humor (AH) is produced by the ciliary body epithelium and represents the fluid in the anterior chamber of the eye. It is involved in the removal of metabolic wastes, in ocular immunity, it supplies oxygen and nutrients, and has a role in refraction. AH is composed of water, electrolytes and proteins, which although are present in relatively low concentrations compared to blood serum, are vital in

the maintenance of anterior segment homeostasis. AH is not in direct contact with the retina but proteins released from it can enter AH either through disrupted BRB, cilio-retinal circulation or through diffusion from the VH and the VH-AH barrier [28]. Proteomic analysis studying AH from patients with DR compared to subjects without DM, found several proteins associated with PDR like apolipoprotein A-I (APOA1), apolipoprotein A-II (APOA2), apolipoprotein A-IV (APOA4), and α-1-acid glycoprotein 1 (ORM1) [29] as well as selenoprotein P (SELENOP), and cystathionine β-synthase (CBS) [28].

IL-23 levels were increased in PDR patients, and IL-17 was also higher in NPDR and PDR compared to control subjects as shown by Wang et al. in 625 patients with T2DM [30]. IL-17 is a powerful pro-inflammatory cytokine capable of mobilizing neutrophils and inducing the secretion of IL-6 and IL-8 prostagladin E2 [30]. Regarding IL-23, it stimulates Th17 cells to secrete IL-17, which further increases the cycle of inflammation.

Concerning IL-10, the cytokine with immunosuppresive properties, capable of attenuating the antigen-presenting ability of APC, studies found contradictory results: decreased concentrations in DR compared to NPDR by Cheung et al. and Zhang et al., contrary to Hu et al. who reported increased levels [30]. It was also suggested that IL-10 levels could be correlated with central subfield thickness (CST), one study reporting higher IL-10 concentration in PDR (224,79 pg./mL) that in the control group (160/14 pg./mL) [31]. TGF-beta is involved in the differentiation of activated CD4<sup>+</sup> Th0 cells into Treg cells, but it is also involved in promoting differentiation of CD4<sup>+</sup> Th0 cells into Th17 cells in combination with IL-23 [32]. Zhang et al. found elevated and stable TGF-beta levels in DM compared to control [30].

Myonectin inhibits the inflammatory response triggered by lipopolysaccharide in macrophages and in myonectin-knockout mices, in which the proinflammatory gene expression was identified. Previous studies demonstrated decreased myonectin concentrations in AH belonging to T2DM especially with PDR but also in NPDR compared to control subjects [33, 34]. In refractory DME, IL-2, IL-8, PIGF and VEGF were found elevated but only IL-8 was correlated with responsiveness, exhibiting higher levels in the refractory group [31]. This result is supported by Jeon et al., who observed that after injecting triamcinolone intravitreally to patients unresponsive to IVB, the effect was positive while the efficacy was associated with IL-8 levels in the AH [35].

IL-6, another proinflammatory cytokine was found elevated in PDR and DME patients. Additionally, it induces VEGF expression, being involved in the breakdown of BRB and thus being involved in the pathogenesis of earlier stages of NPDR [36]. IL-6 is considered as a driving force because it changes the overall cytokine profile in the AH of DM patients. Moreover, the level of VEGF in AH was strongly correlated with VEGF level from the VH [37].

IFN-α is a known angiogenic inhibitor, with a protective role against retinal damage in diabetes. Cheung et al. found reduced IFN-α levels in DM patients compared to control subjects [38].

MCP-1 level in AH is believed to be a predictor for a worse visual acuity outcome of PDR patients after vitrectomy, while there was no significant correlation between VEGF levels and BCVA, as noted by Lei et al. [39]. MCP-1 stimulates the recruitment and activation of monocytes and macrophages which further leads to higher microvascular permeability or ischemia from vessel occlusions. A few studies have investigated the effect of intravitreal triamcinolone [40] or dexamethasone implant [41] on the intraocular inflammatory cytokines and concluded that both drugs could significantly decrease the level of MCP-1 in the AH.

#### *Local Inflammatory Biomarkers and Potential Inflammation-Targeting Therapies in Diabetic… DOI: http://dx.doi.org/10.5772/intechopen.99807*

Samples from AH are collected through a paracenthesis from patients undergoing cataract surgery, trabeculectomy or from post-mortem eyes. The volume collected ranges from 100 to 150 μL. The samples from living patients are more useful since the profile from post-mortem eye is different due to the accumulation of metabolic waste [26].

#### *3.1.3 Vitreous humor*

The vitreous humor is the gelatinous component of the posterior segment of the eye that gives its spherical shape. In addition to its role in structural support, the transparent nature of the vitreous body aids in light transmission to the retina. Because of its avascular nature, much of the proteins found in VH are derived from the retina itself [28]. Several proteins in the VH have been identified as biomarkers for different stages of DR, like components of the acute phase response, coagulation pathway, complement system and other inflammatory pathways (e.g., VEGF, amyloid-β A4 protein, metalloproteinase inhibitor 1, kininogen-1) [19].

Pigment epithelium-derived factor (PEDF) is a negative regulator of inflammatory processes, with lower levels in VH from DR patients compared to NDR. This finding suggest that beside an increase in proinflammatory cytokines, there is also a decrease in the anti-inflammatory ones [42]. Clusterin, a protein involved in the regulation of complement cascade, was found decreased in DR patients [43]. One of its function consists in an anti-inflammatory protection for the BRB [44]. Kallikrein, a known central component of pro-inflammatory kallikrein-kinin system, was found elevated in the VH from DR patients [45].

Anti-VEGF agents such as bevacizumab did not have a significant effect on VH inflammatory proteins as shown by Loukovaara et al. [46], as opposed to Zou et al. who studied VH proteome from patients treated with anti-VEGF agent ranibizumab and found decreased VEGF levels as well as a decrease in complement activation protein, acute inflammatory response and platelet degranulation [47]. Similarly, photocoagulation reduced the levels of pro-inflammatory protein osteopontin (SPP1), as shown by Wei et al. [48].

When exposed to intravitreal injection of IL-1β, retinal capillary endothelial cells degenerate [49], while co-administration of IL-1β, together with high concentrations of glucose, disrupt the RPE cells [50]. Increased levels of IL-1β were found in the VH collected from diabetic patients compared to control [51], underlining its role it the DR-related inflammatory processes.

IL-8, responsible for activating neutrophils and T Lymphocytes, together with IL-6, a chemokine responsible for increased endothelial cell permeability by rearranging actin filaments and changing the shape of endothelial cells, were both found elevated in VH of PDR patients [52]. Their origin does not seem to be the serum, since higher levels were found in the vitreous, suggesting that macrophages, RPE cells, glial cells or monocytes from the vitreous of PDR patients, could produce citokines [53]. IL-10, another anti-inflammatory cytokine, with the ability of deactivating macrophages, was not found to be increased in PDR patients, the author of the study suggesting the lack of a counter balanced by the anti-inflammatory cytokines [54].

MCP-1, with its known role in promoting leukocyte recruitment and inducer of angiogenesis and fibrosis, was found elevated in DR VH, with levels being higher than in the serum [55]. Hyperglycemia seems to increase the formation of MCP-1 from RPE cells, retinal vascular endothelial cells and Muller's cells [56]. This finding further encouraged that MCP-1 gene polymorphism to be proposed as a potential risk factor for DR [57].

Stromal cell-derived factor-1 (SDF-1) is upregulated in damaged tissues and as a response, it mobilizes progenitor cells to promote repair [58]. Together with VEGF, it promotes endothelial progenitor cells from the bone marrow to the ischemic retina [59]. Butler et al. found increased SDF-1 concentrations in VH from DME or PDR patients, and moreover the increase was correlated with disease severity [60].

Extracellular High-Mobility Group Box-1 Protein (HMGB1) as a proinflammatory cytokine, leads to overexpression of TNF-α, MCP-1, ICAM-1, through activation of NF-β [61]. In addition, El-Asrar reported elevated levels of HMGB1 in VH belonging to PDR patients. They further reported that HMGB1 and its receptor for advanced glycation products (RAGE) are expressed by vascular endothelial cells and stromal cells in PDR fibrovascular epiretinal membranes [61].

Tumor Necrosis Factor-*α* (TNF-*α*) was found elevated in VH from DM patients, with a higher vitreous/serum ratio [51]. In addition, TNF-α was also found expressed in vascular endothelial and stromal cells in epiretinal membranes belonging to PDR eyes [62].

ICAM-1 levels in PDR were found increased, especially in active PDR compared to inactive PDR [61]. ICAM-1 was highly expressed in retinal and choroidal blood vessels, as well as in fibrovascular membranes, and its expression seems to correlate with the number of migrated neutrophils in the retina and choroid [6]. When ICAM-1 blocking was obtained, the leukostasis, endothelial cell death and also the vascular leakage in the retinal vessels was diminished.

When retinal capillary endothelial cells were exposed to high concentration of glucose, TNF-α or IL-1β, they produce the membrane-bound form of vascular adhesion protein-1(VAP1) and release the soluble vascular adhesion protein 1 (sVAP-1) [63]. Its role consists in leukostasis and leukocyte entrapment. High levels of sVAP-1 were found in the serum and VH of patients with PDR [64].

Samples from VH are collected through vitreous taps from patients that underwent PPV or after intravitreal injections. Simo et al. raised the problem of confounding factors that could lead to misinterpretation of result. Vitreous hemorrhage could induce a massive influx of serum proteins. Similarly, the disruption of BRB also leads to increased protein levels in VH. The first problem could be solved by rejecting samples with hemoglobin >5 mg/mL, while the second one, by correcting the intravitreal concentration of peptide under study for total vitreal proteins or calculating the vitreous/plasma ratio [56].

#### **3.2 Imaging retinal biomarkers**

#### *3.2.1 Hyperreflective retinal foci*

These lesions are found under the name of hyperreflective foci, dots or spots. Hyperreflective retinal foci (HRF) have been first described in wet AMD but they were also noted in retinal vein occlusion, MacTel type 2, central serous retinopathy, retinitis pigmentosa, Stargardt disease and Best disease. Different pathogenetic origins were hypothesized: initially they were considered precursor of hard exudates or lipid-laden macrophages [65], representing subclinical features of lipoprotein extravasation that were not observed on clinical examination, fundus photography, or fluorescein angiography, due to their small size [66]. In contrast to the previous studies, Lee et al. proposed that HRF represent activated microglial cells since they determined the presence of increased soluble CD14 in the AH. Soluble CD14 seems to be released by activated microglial cells [67, 68]. Other theories assumed that HRF represent debris of photoreceptors or RPE hyperplasia [65, 66].

HRF are initially present in the inner retinal layers, where resting retinal microglia are normally located and after they are activated, they subsequently migrate to

#### *Local Inflammatory Biomarkers and Potential Inflammation-Targeting Therapies in Diabetic… DOI: http://dx.doi.org/10.5772/intechopen.99807*

the outer retinal layers [66, 69]. Another argument that HRF are in fact microglia is that both are primarily found in the inner retina in DR without edema and show an extended distribution into the outer retina in DME [68]. By contrast, Bolz et al. described that HRF are distributed throughout all retinal layers in eyes with different types of diabetic macular edema (DME), diffuse or cystoid [70]. Uji et al. reported the presence of HRF in the outer retina in 53.7% cases while in the inner retina in 99.1% cases in eyes with DME [71]. Uji et al. further suggested an origin from degenerated photoreceptors or macrophages engulfing the former, since HRF were found in the outer retina closely associated with disrupted external limiting membrane or IS/OS line [71].

HRF appears on OCT as small discrete, well-circumscribed, dot-shaped hyperreflective intraretinal lesions (**Figure 1**) [69]. Their size is less than 30 μm, and as opposed to hard exudates they do not induce back-shadowing, while their reflectivity is similar to retinal nerve fiber layer and equal or greater than the retinal pigment epithelium (RPE) band [67, 72].

As opposed to HRF, hard exudates are thought to form from microaneurysms, which is the major source of leakage in focal edema [68]. Other theories even suggested that HRF can occur even before DR development, representing a neurodegenerative process [69]. Both the presence and number of HRF was increased in diabetics versus control subjects, especially in those with signs of DR, but interestingly, none of the patients had signs of DME, hard exudates nor subclinical signs of DME on OCT examinations [68].

The size and also the number of HRF seem to decrease after treatment with anti-VEGF or corticosteroids [71]. The decreased microglial activity following anti-VEGF therapy could contribute to the reduced loads of inflammatory cytokines, which further translates into decreased number of HRS and reduced CMT [68]. Moreover, HRS was the first feature to disappear or reduce after anti-VEGF therapy [66]. One study revealed that an initial presence of larger number of HRF responded better to dexamethasone implants than anti-VEGF therapy but a higher number of HRF is associated with early recurrence of DME after steroid implant [67].

One study found that eyes with HRF in the outer retinal layers presented a greater disruption in EZ (IS/OS) lines at baseline but they improved more than the eyes without the foci at 12 months. Taking into account this result, the authors further suggested that greater VA improvement could be partially explained by foveal photoreceptor restoration in eyes with HRF in the outer retinal layers [65].

#### **Figure 1.**

*Spectral domain OCT linear scan in the macula of a diabetic patient with PDR. The yellow arrows indicate HRF while red arrows indicate hard exudates.*

#### *3.2.2 Sub-foveal neurosensory detachment*

Hypoalbuminemia lowers the intravascular osmotic pressure and together with increased hydrostatic pressure can cause retention of fluid in the subretinal space. The role of subretinal fluid (SRF) in final visual and anatomical outcomes is still unclear. There is a number of studies which have shown that the presence of SRF is associated with good anatomical and functional gains. Contrary to this, there are other studies which have reported the association between the presence of SRF with poor visual gains [73]. Results from RESTORE study [74] and post hoc analysis from RISE/RIDE study [75] proved the protective role of SRF. These studies revealed the correlation between the presence of baseline SRF with better visual gains at the end of 1 year. These studies also showed a positive impact from SRF in response to ranibizumab therapy. Eyes with SRF seem to have increased IL-6 levels, which signifies active inflammation in these eyes, suggesting that inflammation has an important role in the development of this phenotype of DME [76]. Sonoda et al. found that eyes with subretinal detachment (SRD) presented increased intravitreal levels of IL-6, IL-8 and VEGF [77]. Discontinuation of the external limiting membrane in the SRD type of DME induces cellular damage, which further attracts scavenger cells to the retina, producing IL-6 (**Figure 2**) [78]. The migration of activated microglia in the outer segments of the retinal layer could support the production of IL-6 and subsequently the accumulation of SRF. The SRD type of DME is associated with an increased number of HRF and significant functional impairment of the macula [78]. The presence of numerous HRF is also associated with poorer glycemic control in patients without significant DME [79]. The levels of ICAM 1 are associated with the height of SRF, indicating that increased vascular permeability by inflammatory mediators results in SRF [80]. An important finding was that DME eyes with SRF respond significantly better to dexamethasone implants and the authors consequently supported the use of dexamethasone implants in patients with DME with SRF [81]. Interestingly, a recent study assessing the change of parameters in DME with SRD after treatment, reported that intravitreal injection of steroid like dexamethasone implant showed greater decrease in number of RHF compared to anti-VEGF agent such as ranibizumab [82].

#### *3.2.3 Hyperreflective choroidal foci*

The analysis of the choroid could be an important tool in the assessment of progressing DR. Choroidal parameters such as luminal area (LA), stromal area (SA) and total choroidal area (TCA) were found decreased in diabetic patients compared to control by a recent study [83]. Hyperreflective choroidal foci (HCF) is a recently described entity in DR. It is postulated that HCF represent HRF migrated from the retina, into the choroidal layers, which was permitted by ELM and EZ disruption (**Figure 3**). The hypothesis that HCF are actually migrated HRF, is supported by the study of Roy and al. where all 59 eyes with HCF had HRF as well and there was no eye which had HCF in the absence of HRF on SD-OCT [84]. Presence of HCF could be used as a biomarker of disease severity in eyes with DR [84]. HCF were found to be present significantly more in PDR eyes than NPDR eyes [67].

#### *3.2.4 Intraretinal cystoid spaces*

Inner BRB is affected by elevated levels of VEGF leading to increased vascular permeability, which will result in a decreased osmotic gradient, extracellular fluid accumulation, and cyst formation [67]. VEGF is also responsible for the depletion of the occludin in RPE which will further disrupt the tight junction in the outer

*Local Inflammatory Biomarkers and Potential Inflammation-Targeting Therapies in Diabetic… DOI: http://dx.doi.org/10.5772/intechopen.99807*

#### **Figure 2.**

*Spectral domain OCT linear scan in the macula of a diabetic patient with PDR. The small arrow indicates sub-foveal detachment. In the magnified segment of the image, the tall arrows indicate intact external limiting membrane (ELM) while the curly brackets indicate a discontinuous ELM.*

#### **Figure 3.**

*Spectral domain OCT linear scan in the macula of a diabetic patient with PDR. Right upper image: White-onblack OCT scan. Right lower image: Black-on-white OCT scan. The white arrows head indicate hyperreflective choroidal foci. The white and black arrow highlights a group of hyperreflective retinal foci.*

BRB [85]. Damian and al. found decreased RPE thickness in almost all quadrants in the PDR-DMO group as opposed to NPDR-DMO group, where the quadrants number was lower [86]. This finding was explained by a disruption of the RPEphotoreceptors complex due to ischemia and level of inflammation that varies according to the stage of DR. Cysts larger than >200 μm in the outer nuclear layer (ONL) are seen in advanced stages of DME and may cause irreversible damage to visual functions. Due to their location, they damage photoreceptor cells and disrupt inner segment–outer segment (IS/OS) junction (**Figure 4**). Regarding their content, there is paucity of information in the literature. Liang and al. hypothesize that fibrin and other inflammatory cells may fill these spaces [87]. Treatment with anti-VEGF therapy leads to a decrease in the number and size of ONL cysts by decreasing the permeability of the inner BRB. This was associated with improvement in BCVA and microperimetric retinal sensitivity. Hyperreflective signals within the cyst are associated with severe disruption of the BRB [67].

#### *3.2.5 Foveal hyperautofluorescence*

Fundus autofluorescence (FAF) is a rapid, noninvasive imaging technique that may give new insights into the evaluation of DME. Short wavelength FAF derives its signal mainly from lipofuscin in the RPE. Long wavelength autofluorescence

#### **Figure 4.**

*Spectral domain OCT linear scan in the macula of a diabetic patient with PDR. The white arrows indicate intraretinal cystoid spaces.*

or near-infrared (NIR) FAF derives its signal from melanin which is present in RPE and choroid. Melanin accumulates in the apical parts of the RPE cells and is thought to protect the RPE. In DR, local ocular inflammation and oxidative stress lead to increased amount of lipofuscin and decreased amount of lutein and zeaxanthin in the macula. This is responsible for increased FAF signal in subjects with DME. In addition, activation of microglia in DM could cause oxidation of proteins and lipids. Histologic studies have found lipofuscin to accumulate in microglia [67].
