The hyperglycaemic state of diabetic patients leads to the excessive generation of free radicals (mainly reactive oxygen species: ROS) because of an increased cell mitochondrial activity [1, 2]. Consequently, oxidative stress plays an unquestionable role in the aetiology of diabetic complications, such as corneal neuropathies, cataract and retinopathies with different grades of severity [2, 3, 4, 5, 6]. In fact, thiobarbituric acid reactive substances (TBARS), F2 isoprostanes and 8-OH-guanosine, that are considered reliable biomarkers of oxidative stress, were present in serum from diabetic patients at significantly higher levels with respect to healthy subjects [4].

Very often, diabetic complications involving the ocular surface unit result in a corneal diabetic neuropathy due to sub-basal nerve plexus alterations [7]. Therefore, several clinical studies in diabetic patients have revealed a significant involvement of the ocular surface with alterations of the corneal and conjunctival epithelium and persistent epithelial defects [8, 9]. As a consequence, diabetic patients are generally considered at higher risk than non-diabetics for ocular surgery, with a prevalence of ocular surface complications [2, 10]. Diabetic patients usually present decreased corneal sensitivity associated to reduced tear secretion, tear film hyperosmolarity and alterations of corneal epithelial cell junctions, finally resulting in corneal epitheliopathy and neuropathy [11, 12, 13, 14]. Corneal involvement may manifest with punctate keratopathy, recurrent erosions, different grades of neurotrophic keratopathy and may be associated to wound healing delay and a higher risk of infection [14, 15, 16, 17]. Diabetic patients often complain of dry eye symptoms, such as burning, stinging and foreign body sensation. Additionally, increased conjunctival squamous metaplasia with low goblet cell density and corneal sub-basal nerve alterations may also occur [11, 12, 13, 17, 18, 19].

Alpha-Lipoic Acid (ALA) is a naturally occurring antioxidant also present in foods such as broccoli, spinach, yeast or meat, heart and liver [20]. Two sorts of antioxidants generally work in biological systems: those which mainly associate with a lipophilic phase (membranes or lipoproteins such as vitamin E and ubiquinols) and those operating in the aqueous phase such as ascorbate, glutathione and thioredoxin. ALA in its reduced form as dihydrolipoic acid (DHLA) can interact with both phases and is a bridge between them (Fig. 1A). Therefore, ALA and its reduced form DHLA, are considered powerful natural antioxidant agents with a scavenging capacity for many reactive oxygen species [21, 22]. However, naturally occurring ALA in tissues (normally found at concentrations ranging between 5 and 25 nmoles/gr) is mostly found covalently attached in an amide linkage with lysine residues on enzymes and thus becomes unavailable as antioxidant (Fig. 1B). Dietary supplementation increases unbound ALA, which can now act as a potent antioxidant and counteract oxidative stress [23]. Beside its antioxidant activity, ALA may increase tissue sensitivity to insulin and lowers blood glucose levels, probably by increasing glucose uptake in muscles and fat cells through the insulin signalling cascade [24]. Lipoic acid distributes to mitochondria where it serves as a critical co-factor for mitochondrial enzymes and is considered a powerful antioxidant, since it is a direct scavenger of ROS with the ability to regenerate endogenous antioxidants such as coenzyme Q10, glutathione and vitamins E and C (Fig. 1A). The additional chelating activity of metal ions gives to ALA high and proven properties to efficiently blunt ROS levels [25].

Fig. 1.

Lipoic acid and its metabolic effects in living bodies. (A) After its absorption, lipoic acid is reduced enzymatically by NADH or NADPH to dihydrolipoic acid. It can regenerate glutathione and vitamin C, which in turn help the regeneration of vitamin E, thus giving a strong contribution to the antioxidant defense of cells and tissues. (B) Lipoic acid is often found near the active sites of enzymes, usually bound to the peptide by a long, flexible amide linkage with a lysine residue, and is therefore unavailable for vitamin regeneration.

Reports describing the effects of ALA oral supplementation in diabetic patients confirm its capability to improve insulin sensitivity, to decrease glycemia and oxidative stress, and to induce clinical improvement of neuropathy [20, 25, 26, 27].

However, the systemic administration of ALA as a food supplement is expected to deliver only a small fraction to the eye, whereas if given as topical eye drops, a higher amount can be retained within the eye [28]. Therefore, given the ascertained role of oxidative stress on diabetic dry eye, the purpose of this study has been to investigate the effects of an original topical treatment with ALA-enriched artificial tears on the tear film and the ocular surface of diabetic patients with symptomatic dry eye disease.

Six dropouts were registered among the 70 patients enrolled: two patients from group 1 (treated), and four patients from group 2 (controls) did not show up at the two months’ control. Therefore, data are based on 33 patients from group 1 and 31 patients from group 2.

3.2 Confocal Microscopy

Fig. 2, panels A and C show confocal pictures of the tear film of two representative diabetic patients, one per each group, at enrolment. Before treatment an abnormal pattern with numerous irregular elements is clearly visible: numerous bright speckles and dark spots can be observed, likely due to the presence of dispersed lipid drops and cellular debris, sign of an abnormal tear film and the suffering of ocular surface epithelial cells. After two months of topical ALA/HPMC administration (Fig. 2B), a regular distribution of the tear film can be appreciated, with a smooth surface and no bright speckles or dark elements (p = 0.0078). Fig. 2D shows the tear film of the control patient after treatment for two months with HPMC alone, also showing a similar, although lesser, improvement (p = 0.031). The semiquantitative score assigned to the different samples are reported in Table 3. A significant improvement was observed in both groups, which were and remained statistically similar at the beginning (p = 0.22) and at the end (p = 0.14) of the treatment time.

Fig. 2.

Illustrative confocal images of the tear film in diabetic patients before and after treatment in both studied groups. (A) Group 1 (HPMC/ALA) before treatment: the tear film surface appears irregular with an altered lipid layer, numerous bright speckles, possibly due to exposed tips of microvilli and microplicae, and dark spots, produced by desquamating epithelial cells (arrows). (B) Group 1 after treatment: the tear film surface appears smooth with a regular distribution of the lipid layer and the absence of irregular elements. (C) Group 2 (HPMC alone) before treatment: the tear film surface shows an irregular pattern of the lipid layer with numerous presence of bright speckles (arrows). (D) Group 2 after treatment: evident improvement of the lipid layer with reduction of the irregularities can be observed. Scale bar 100 $\mu$m.

Table 3.Confocal analysis of tear film, the average score $\mathrm{\pm }$ SD is reported.

Treatment	Time	Score
GROUP 1 (ALA/HPMC)	Baseline	1.81 $\pm$ 0.75
	2 months	0.94 $\pm$ 0.77
	p value	0.0078${}^{\mathrm{\#}}$
GROUP 2 (HPMC)	Baseline	2.22 $\pm$ 0.67
	2 months	1.44 $\pm$ 0.53
	p value	0.031${}^{\mathrm{\#}}$
# Mann-Whitney test.

A healthy ocular surface unit is necessary for proper visual acuity and diabetes is known to be a strong risk factor for its integrity. In fact, the prevalence of dry eye disease in diabetic patients is about 50–53%, and corneal epithelial lesions are present in 47–64% of affected patients who frequently report ocular discomfort and visual impairment [1, 8]. Goebbels et al. [13] presented data showing that, in comparison to healthy subjects, insulin-dependent diabetic patients showed significantly decreased reflex tearing (Schirmer test I) and significantly increased evidence of conjunctival metaplasia. Basal tear flow (Schirmer test II) and TBUT were found to be unaffected. These data are consistent with the observation that diabetic neuropathy most often involves the ophthalmic nerve, thus triggering a corneal neuropathy with loss of sensitivity [39, 40], which in turn results in decreased tear secretion and altered composition of tear proteins [41, 42].

Given the wide occurrence throughout the world of dry eye disease, different treatment strategies have been developed, and proposed for the treatment of eye dryness, independent of its etiology [43]. They are based on the integration of one of the three components of the tear film: mucins, water and lipids; or contain molecules with direct or indirect antinflammatory and secretagogue effects. The main objective is the stabilization of the tear film (normalization of the BUT) and the recovery of the ocular surface. Hyaluronic Acid (HA) and Hydroxy-Propyl-Methylcellulose (HPMC) are among the most used hydrating and lubricant molecules to integrate the aqueous and mucinous layers of the tear film [44]. However, it is oxidative stress the main and an early event in the pathogenic mechanism of diabetes [45], which may influence the onset and progression of late complications, especially diabetic retinopathy [46], but also including dry eye [47]. Therefore, the use of antioxidants on the ocular surface of such patients appears to be an interesting therapeutic target.

To date, few eye drops formulations have been proposed to specifically treat the ocular surface disease (OSD) of diabetic patients. N-acetyl-carnosine has been shown to be an antioxidant prodrug, with the potential to treat several ocular pathologies in which the oxidative stress is a main agent [48]. Naltrexone (an opioid antagonist and a mild antioxidant) or Rebamipide (a scavenger of free radicals and a COX2 inhibitor) have been formulated as eye drops to treat diabetic OSD [49, 50]. Most recently, Glucagon-Like Peptide-1 (GLP-1) formulated in eye drops have been used in a rat model system of diabetes-induced retinal abnormalities, showing that the enhancement of the antioxidant defense system in the diabetic retina has a neuroprotective action favouring DNA repair and neuron cells proliferation [51]. However, the great advantage of using ALA as an antioxidant comes from its very same chemical nature, in that it recycles between a reduced and an oxidized form, and both the oxidized and reduced forms of ALA exhibit antioxidant properties [52] so that this antioxidant system gets exhausted in a much longer time than other antioxidant systems. This property can be very critical in the management of diabetic dry eye, which recognizes oxidative stress due to a long-lasting hyperglycaemic state as a main etiological cause.

Systemic supplementation with ALA has already been suggested to have positive effects both in the prevention and treatment of diabetes and its systemic complications [53]. Most recently, ALA treatment has even been proposed to attenuate the symptoms of COVID-19 in diabetic patients [54]. Along this line, several other experimental studies have shown the positive effects of ALA in patients affected by dry eye and diabetic retinopathy [55]. In fact, from a mechanistic point of view, ALA can down regulate the expression of MMP-9 in the corneal epithelium and induce a better antioxidant defence on the ocular surface. Moreover, being endowed with a strong direct antioxidant effect, it can efficiently counteract oxidative stress-induced corneal surface erosion and damage to the lachrymal glands. Through the inhibition of the enzyme N-acetylglucosamine transferase, of the pro-inflammatory transcription factor NFk-B and the direct alleviation of oxidative stress, ALA may contribute to the prevention of diabetic retinopathy. Similarly, it can activate the transcription factor Nrf-2, which regulates the expression of antioxidant defence and PKA in retinal ganglion cells [55]. Indeed, Nrf-2 activation has been proposed as a strategy to prevent diabetic complications [56]. Clinical trials of ALA given as food supplement on diabetic complications such as neuropathy, nephropathy and cardiomyopathy returned favourable results, especially for pain attenuation in neuropathy, and contrast sensitivity in retinopathy [57]. More specifically, clinical trials conducted in pre-retinopathic diabetic patients showed ALA with genistein and vitamins could protect the retinal cells and decrease the inflammatory effect in diabetic patients [58]. In order to enhance the specific activity of ALA on the eye, a nanomicellar formulation of ALA as eye drops has been described, with the aim of treating diabetes-associated ocular surface disease. Such formulation showed an improved pharmacodynamics, with enhanced stability and permeability [59].

As a powerful antioxidant, ALA chelates transition metal ions, increases the levels of other antioxidants such as glutathione and vitamin E and C, therefore resulting in the scavenging of ROS and other free radicals. Based on these properties, the systemic use of ALA has been proposed for the treatment of diabetic peripheral neuropathy [20, 25, 26]. Since oxidative stress is clearly involved also in the pathogenesis of ocular surface alterations such as in the case of diabetic dry eye, the topical use of strong antioxidant drops could be expected to improve tear film stability and surface integrity. In fact, in the present study, we have shown a better improvement of the ocular surface in the group of patients treated with the combination of HPMC and ALA, with respect to the group of patients taking only HPMC eye drops. It is interesting to note that although the subjective perception of improvement (as measured by the OSDI score) was similar between the two groups of patients, the average improvement in group 1 appeared better than the one measured in group 2 (Fig. 3). Moreover, the objective signs improved more in the group of patients treated with the association of HPMC and ALA (Fig. 3). We observed a 30% average improvement in BUT with HPMC/ALA (p = 0.004) vs. a 13% average improvement with HPMC alone (p = 0.021). Even more interesting was the finding of a significant 40% average reduction (p = 0.02) of the corneal epithelial staining in patients who received HPMC/ALA with respect to the control group, in which no significant variation was observed (p = 0.65). Therefore, the strong antioxidant effects of ALA added to the hygroscopic and lubricant activities of HPMC improved the efficacy of the eye drop combination. In fact, HPMC alone was indeed effective, but not on all the parameters under evaluation, and not with the same intensity as the combination with ALA. Only tear production and corneal sensitivity were not affected by either treatment, likely due to the high variability among the enrolled subjects, and to the fact that their average starting values were not in the pathologic range. On the other hand, even though the average quantity of tears was not increasing, tear quality appeared to be improving, since its stability as measured by the BUT was increasing, in parallel with corneal epithelial cells integrity as evidenced by the confocal microscopy analysis of the tear film. This technique was applied to tear film analysis by Mather in 1997 [60], however with a poor resolution of the images, likely due to the fact that they used a contact confocal procedure, which is not optimal for the observation of the tear film [61]. The technique was in fact perfected by the pioneering work of Smolek et al. [37] using non-contact confocal microscopy, who described the regular lipid layer distribution over a smooth surface with no bright speckles of the tear film in normal eyes. On the contrary, in dry eye the bright speckles - possibly due to exposed tips of microvilli and microplicae - could be observed with numerous dark erosions produced by desquamating epithelial cells. In a following paper, Stonecipher and Green confirmed the previous observations by evaluating the tear film in normal eyes and in moderate to severe dry eye [38]. In normal eyes the lipid layer was found to be evenly distributed over the eye surface, whereas in moderate dry eye it appeared spotted by the presence of many tear film debris; in severe dry eye, many desquamated epithelial cells with a poor tear film lipid layer with a rapid BUT could be seen. In our study, the confocal analysis of the tear film before any treatment showed severe alterations, in agreement with its instability as evidenced by the low BUT. These observed anomalies could be attributed to a thinner lipid layer, and the dark elements may likely derive from cell debris or mucous layer alterations. In our study - the first to apply non-contact confocal microscopy of the tear film in a comparative treatment analysis - we could also appreciate the same irregular pattern described by the previous authors, with numerous bright speckles corresponding to the exposed tips of microvilli as it happens in moderate and severe dry eye, and in good agreement with the more intense corneal fluorescein staining of the enrolled subjects, due to desquamating cells and deep erosions. In images taken after the two months of treatment, we could see a clear improvement of the ocular surface and the tear film, with a smooth lipid layer, reduced bright speckles, dark erosions and cellular elements in both treatment groups. However, patients in group 1, treated with the combination of ALA/HPMC displayed a 48% average improvement (p = 0.0078), better than the 35% average improvement seen in control group patients (p = 0.031) (Fig. 3), in agreement with the significant improvement of the corneal staining for patients of group 1, but not of control group 2. This confocal investigation of the tear film is a relatively new technique, with scant literature publications, so that no standard reference values (in terms of a semiquantitative analysis) are available.

Fig. 3.

The average variation of different parameters before and after two months of treatment with either ALA/HPMC or HPMC alone eye drops has been calculated as a percentage, dividing the average value obtained at the end of treatment by the initial value, as reported in Table 2.

To our knowledge, this is the first report on the topical use of a strong antioxidant such as ALA in diabetic patients with dry eye symptoms. More studies will be necessary to confirm and extend these preliminary observations, which have been done in a short time frame. Nonetheless, we believe that the additional benefits of such therapy could already be considered for diabetic patients with ocular surface disorders. Furthermore, the topical use of ALA could be extended to all dry eye patients in view of the fact that oxidative stress is more than often a key element in the etiology of dry eye.

Conception and design—AMR, PA. Administrative support—AMR, PA. Provision of study materials or patients—AMR, PA. Collection and assembly of data—RS, GWO, EIP, GAS. Data analysis and interpretation—AMR, RS, GWO, EIP, GAS, DR. Manuscript writing—AMR, DR, PA. Final approval of manuscript—All authors.

A single blind, randomized, prospective study was conducted in accordance with the tenets of the Declaration of Helsinki and was approved by the Ethical Committee of Messina with Prot.N°106/18.

Not applicable.

This research received no external funding.

DR is a full time employee of Fidia Pharmaceuticals, the company selling the product Tioretin under analysis in this study. The other authors declare no conflict of interest.