Novel Lipid Mediator 7S,14R-Docosahexaenoic Acid: Biogenesis and Harnessing Mesenchymal Stem Cells to Ameliorate Diabetic Mellitus and Retinal Pericyte Loss

1. Introduction

Type 2 diabetes, accounting for > 90% of diabetes worldwide, is a major public health problem and places a severe economic burden on health care systems. It is characterized by insulin resistance in peripheral tissues. As long as pancreatic β-cells are able to compensate for insulin resistance by enhancing insulin secretion and increasing β-cell mass, euglycemia can be maintained. However, when they are unable to meet the body’s demand for insulin because of genetic defects and/or exogenous insults, diabetes mellitus ensues [1,2,3,4]. Chronic hyperglycemia leads to glucose toxicity and worsening of impaired insulin secretion due to the overworking of pancreatic β-cells, which results in a decreased ability to secrete insulin. Hyperglycemia can also cause oxidative stress, which influences β-cell survival [5,6]. Reduced β-cell function has been observed in both obese and lean type 2 diabetic humans and in diabetic rodent models [7,8,9,10,11]. Therefore, prevention of the progression of pancreatic β-cell dysfunction in patients with diabetes mellitus should be a key in the long-term management of diabetes mellitus.

Hypoglycemic agents can significantly reduce the development of diabetic complications [12,13,14]; however, many of them cause adverse effects and sometimes result in progressive deterioration in β-cell function [15,16,17]. Recently, some strategies using antioxidative treatment have been shown to ameliorate diabetic mellitus. An angiotensin II type 1 receptor blocker was shown to decrease oxidative stress markers in β-cells, improve abnormalities of pancreatic islets in diabetes db/db mice [18,19,20]. Also, increasing levels of the antioxidant enzyme, glutathione peroxidase-1, specifically in β-cells, has been demonstrated to protect them against the adverse effects of chronic hyperglycemia in db/db mice [21,22].

Multiple differentiation capacity, paracrine functions, and immunomodulation action make the marrow mesenchymal stem cells (MSCs) a good candidate for prevention or cure of diabetic mellitus and complications [23,24,25,26]. Transplantation of bone marrow–derived cells increased levels of serum insulin and reduced blood glucose levels in hyperglycemic mice with streptozotocin-induced pancreatic tissue damage, which improved the metabolic state and survival of recipients [27,28]. Studies using rat MSCs [29,30] or human bone marrow–derived multipotent stromal cells [31,32,33] have shown similar results of enhanced insulin secretion and repair of pancreatic islets after streptozotocin treatment in rodents. Patients with type 2 diabetes received an intrapancreatic autologous bone marrow stem cell infusion; after 1 year, these patients showed improved metabolic control and reduced insulin requirements [34,35]. However, diabetic hyperglycemia and concomitant oxidative stress damage DNA, proteins, and lipids in various tissues, resulting in dysfunction of cells and enzyme systems. When MSCs are transplanted into diabetes, the inhospitable environment might damage MSC functions [36,37].

Diabetic retinopathy, a secondary microvascular complication of diabetes mellitus, is the leading cause of blindness in the United States among individuals aged 20 to 64 [38,39]. A persistent increase in blood glucose levels shunts excess glucose into the aldose reductase pathway in certain tissues, which converts sugars into alcohol (e.g., glucose into sorbitol, galactose to dulcitol). Pericytes of retinal capillaries seem to be affected by this increased level of sorbitol. Thus, the loss of function of pericytes results in weakness and eventual vascular outpouching of capillary walls, resulting in microaneurysms [38,40].

Docosahexaenoic acid (DHA) is an essential omega-3 fatty acid. Resolvin D series, neuroprotectins, and maresins are potent anti-inflammatory lipid mediators (LMs) naturally generated from DHA during inflammation and/or resolution of inflammation [41,42,43,44,45]. Neuroprotectin D1 and other DHA derived lipid mediators have been reported to protect retinal epithelial cells and neuronal-glial cells from apoptosis induced by oxidative stress [46,47,48,49], promote neuronal functions and prevent neural degeneration [50,51,52,53]. We identified a DHA-derived LM, 14S,21R-diHDHA, which promotes wound healing [54,55,56], inhibits MSC apoptosis, and enhances MSC functions to ameliorate acute kidney injury [57]. In 2009, Serhan et al. reported their discovery of maresin 1 (7R,14S-dihydroxy-4Z,8E,10E,12Z,16Z,19Z-DHA and its inflammation-resolving functions [58]. Inspired by this discovery, we identified and prepared a novel maresin 1 stereoisomer, 7S,14R-dihydroxy-4Z,8E,10Z,12E,16Z,19Z-DHA (7S,14R-diHDHA) from leukocytes and related enzyme incubations, as we predicted the therapeutic significance that the biofunction of maresin 1 is conserved in this stereoisomer and that it could promote MSC amelioration of multiple diseases and injuries. Maresins promote the beneficial bioactions of stem cells in the treatment of diseases [59]. Maresin 1 regulates insulin signaling in human adipocytes as well as in adipose tissue and muscle of lean and obese mice [59]. It inhibits hyperglycemia-induced ferroptosis [60]. DHA preserves visual function by maintaining correct disk morphology in retinal photoreceptor cells [61].

Accordingly, we hypothesized that MSCs can protect pancreas β-cells from the damage of diabetic oxidative stress and ameliorate retinal pericyte loss, and biogenic 7S,14R-diHDHA can induce MSC protective functions. This was further tested to be positive by our studies in vitro and by our experiments using db/db diabetic mouse model in vivo. 7S,14R-diHDHA was formed by the cytochrome P450 enzyme and 5-lipoxygenase (5-LOX) in tandem and by leukocytes [58,62,63,64]. 7S,14R-diHDHA induced MSC functions in amelioration of diabetes mellitus in db/db mice by lowering blood sugar and increasing glucose tolerance and blood insulin levels, and decreased the number of macrophages in islets, increasing β-cell viability, density and insulin secretion, and reducing pericyte loss in the retina. 7S,14R-diHDHA can enhance the MSC secretion of trophic growth factors. Overall, pretreatment of MSCs with 7S,14R-diHDHA, or related lipid mediators or structural mimics, may offer a new clinical strategy for improved treatment of diabetic mellitus and retinopathy as well as other diabetic complications.

2. Materials and Methods

Reagents

Racemic ±14-hydroxy-4Z,7Z,10Z,12E,16Z,19Z-DHAs (±14-HDHAs or 14S/R-HDHAs) and 5-lipoxygenase (5-LOX, human recombinant or potato) were supplied by Cayman (Ann Arbor, MI, USA). NaBH₄, interleukin (IL)-1β, TNF-α, and Escherichia coli lipopolysaccharide (LPS) were supplied by Sigma-Aldrich (St. Louis, MO, USA).

Cells

MSCs were isolated from the bone marrow of C57BL/6J mice (8 weeks old, female, Jackson Laboratory), as described in our publications [57]; more than 95% of isolated MSCs were positive for Sca-1 and CD29 based on flow cytometric analysis [54,57,65]. These MSCs possessed adipogenic and chondrogenic differentiation ability. They were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% mesenchymal supplement (Stem Cell Technologies, Vancouver, BC, Canada). Cells in passages 7 through 10 were used for the experiments. The Min6 mouse β-cell line (ATCC, Manassas, VA, USA) was cultured in RPMI1640 containing 10% fetal bovine serum (ATCC, Manassas, VA, USA).

Isolation of human neutrophils, monocytes, and lymphocytes

Human samples were managed following protocols approved by the Institutional Review Boards of our institutes. Established procedures were followed [48,66]. Human peripheral blood supplied by the Blood Center (New Orleans, LA, USA) was from a healthy donor who was unknown to us and who had not been on medication for > 2 weeks prior to donation. We centrifuged the blood in sodium citrate (180 g, 10 min, 23^oC); then centrifuged the supernatants (1100 g, 15 min, 23 °C). The plasma (the top layer) was removed. We washed the cell pellets with 7 mM EDTA. Neutrophils and mononuclear cells were separated from blood by a Ficoll-Hypaque gradient. We plated mononuclear cells onto plastic dishes to further separate the cells. The cells adhered to the dishes were monocytes, and nonadhered cells were lymphocytes. Neutrophils, monocytes, and lymphocytes prepared were of > 96% purity. Cell viability was > 95% according to trypan blue exclusion.

Biogenesis of novel 7S, 14R-DHA

14R-hydroxy-4Z,7Z,10Z,12E,16Z,19Z-DHA (14R-HDHA) was prepared from ±14-HDHAs as described in our previous publication [67] and the section below. We then incubated the 14R-HDHA (10 µg/incubation) with 5-LOX (3 h, 37°C) to identify the biogenic pathway of the novel maresin-1 stereoisomer, 7S,14R dihydroxyl DHA, which is expected to be produced by 5-LOX with 14R-hydroxyl preserved. A few grains of NaBH₄ powder were added at the end of incubation. For cellular generation of 7S,14R-diHDHA, human leukocytes (6 x 10⁶ neutrophils + 5 x 10⁵ monocytes + 3 x 10⁶ lymphocytes) were incubated (20 min, 37°C) in phosphate-buffered saline (PBS) containing 200 μg 14R-HDHA. This ratio of leukocytes resembles the typical leukocyte profile of healthy adults [68,69]. The cells were then stimulated (37°C, 30 min) with inflammatory factors (10 ng/mL TNFα, 10 ng/mL IL-1β, and 100 ng/mL LPS) as described in our previous publication, to promote the biosynthesis of lipid mediators [54,55,67].

Analysis and Isolation of 7S, 14R-diHDHA and of its precursor 14R-HDHA

14R-HDHA (> 98% pure) was isolated from purchased ±14-HDHAs using aqueous reversed-phase chiral liquid chromatography (aR chiral LC) coupled to an ultraviolet photo diode array detector and tandem mass spectrometer (aR chiral LC-UV-MS/MS) as we published previously [54,55,67]. The incubations were extracted and then analyzed or fractionated for 7S,14R-diHDHA. Briefly, A Chiralpak-IA chiral column (150 mm long x 2.1 mm ID x 5 μm, Chiral Tech, West Chester, PA, USA) was used for the aR chiral-LC separation. The mobile phase flowed at 0.2 mL/min, eluted as B (methanol:H₂O:acetic acid = 27:73:0.01) from 0 to 1 min, ramped from B:methanol 40:60 to B:methanol 20:80 by 50 min, ramped to methanol by 55 min, and then flowed as methanol. The instrument used included the system of Survey LC-PDA UV-LTQ linear ion trap MS/MS (Thermo Fisher, Waltham, MA, USA) [54,55] and the system of an Agilent 1100 LC system (HPLC-PDA, Agilent Technologies, Santa Clara, CA, USA) and a QTRAP 6500⁺ quadruple-linear trap MS/MS (Sciex.com, Framingham, MA, USA), with electrospray ionization for both systems [70]. The LC-UV effluents were split at a 1:20 ratio, where 1 portion went into a mass spectrometer that monitored the chromatography of 7S,14R-diHDHA and 20 portions were fraction-collected manually for the aR chiral LC peak of 7S,14R-diHDHA. The collected effluents containing 7S,14R-diHDHA were pooled together and evaporated for solvent removal under vacuum or nitrogen gas blow.

Protocols for the treatment of cells in vitro

MSCs (2 × 10⁵) were treated with 7S,14R-diHDHA (250 nM) for 12 h. The level of VEGF or HGF in supernatant was analyzed using ELISA kits (RayBiotech, Peachtree Corners, GA, USA, and R&D Systems, USA, respectively). Min6 β-cells were cultured to 60%–80% confluence (1.8 × 10⁵) in lower chamber of a 24-transwell plate (8 µm pore) pretreated with polylycine overnight. MSCs (3 × 10⁴) were cultured in the upper chamber of a 24-transwell pretreated with or without 250 nM 7S,14R-diHDHA for 4 h. Min6 β-cells and MSCs were cocultured in RPMI1640 for 8 h, and then treated with 100 or 200 µM H₂O₂ for 10 h. Cell viability was analyzed using the MTT method. Moreover, Min6 β-cells were treated for 4 h with conditioned medium (CM) of MSCs or 7S,14R-diHDHA-treated MSCs.

Mice

We used 16-week-old female C57BL/6J mice and diabetic db/db mice (BKS.Cg-m+/+leprdb/J, Strain #:000642) (Jackson Lab, Bar Harbor, ME). Animal protocols were approved by the Institutional Animal Care and Use Committee and Institutional Review Board of Louisiana State University Health Sciences Center, New Orleans, and followed the ARRIVE guidelines [71]. The studies were conducted in a blinded fashion.

Treatment procedures in vivo

MSCs were cultured with or without 250 nM 7S,14R-diHDHA for 4 h and then collected trypsinization (0.25% trypsin/0.53 mM ethylenediaminetetraacetic acid; Invitrogen, Carlsbad, CA), similar to the procedures that we used previously [57]. MSCs (1 × 10⁶) with or without treatment of 7S,14R-diHDHA were infused into 16-week-old db/db mice via tail vein every 5 days for 35 days. During the intervention period, body weight was measured every week. Nonfasting blood glucose was measured using a OneTouch Ultra glucose meter (LifeScan, Milpitas, CA) every week. The intraperitoneal glucose tolerance test (IPGTT) was conducted as follows. After 5 weeks of treatment, mice were injected intraperitoneally with glucose (2 g/kg body weight) after 8 h of fasting. Blood samples were taken at 0, 60, and 120 min from the tail vein, and blood glucose was measured. Sera were collected at 120 min, and insulin concentrations were measured using ELISA using mouse insulin as a standard.

Histological study

Mice were perfused with PBS at the end of the treatment. Pancreas and eyes were collected and fixed in 4% paraformaldehyde for 2 h, and cryosections (5 µm) were prepared. Sections were blocked with normal goat serum (10% in PBS) for 30 min at room temperature, and then incubated with the following primary antibodies: rabbit antimouse insulin, rabbit antimouse glucagon, rat antimouse F4/80 (Santa Cruz Biotechnology, Dallas, Texas, USA), and rabbit antimouse α-SMA (Millipore, Burlington, MA, USA). Primary antibodies were followed by incubation with the secondary antibodies: Cy3-labeled goat antirat immunoglobulin G (IgG), Cy5-, or fluorescein isothiocyanate (FITC)-labeled goat antirabbit IgG (Invitrogen, Carlsbad, California, USA). Normal rat IgG and rabbit IgG (Santa Cruz Biotechnology, Dallas, Texas, USA) were used as isotype controls. Nuclei were counterstained with Hochest 33342 (Invitrogen, Carlsbad, California, USA). To assess the α cell ratio, the image analysis software NIH ImageJ was used to calculate the pancreatic islet cell number, the insulin⁺ and glucagon⁺ cells in the pancreatic islets to determine the percent β-cell count and α-cell count (relative to the pancreatic islets). The macrophage infiltration in the pancreas was assessed using the number of F4/80⁺ cells per hpf. To assess retinal pericyte loss, retinal sections were stained with an antibody against α-SMA, which specifically marks pericytes in the capillaries. The number of α-SMA⁺ cells in the inner nuclear layer of the retinas was counted and normalized against the area of the inner nuclear layer.

Statistical analysis

All data were analyzed using one-way analysis of variance followed by Fisher’s least significant difference post hoc test and expressed as mean ± standard error of the mean. p < 0.05 was considered statistically significant.

3. Results

3.1. Biogenesis of novel 7S,14R-dihydroxy-4Z,8E,10Z,12E,16Z,19Z-DHA

14R-HDHA forms from DHA by P450 enzymatic conversion, as we observed previously [67]. We predicted that 5-LOX or 5-LOX rich leukocytes [58,62,63,64] convert 14R-HDHA to 7S,14R-dihydroxy-4Z,8E,10Z,12E,16Z,19Z-DHA where the chirality of C-14 is preserved as 14R, and the C-7 takes in the hydroxyl at S-configuration based on the well-established knowledge of 5-LOX [58,62,63,64]. This prediction was tested positive as we observed 7S,14R-diHDHA as a single aR-chiral LC peak at 50.6 min (Figure 1A), with its UV spectral triplet possessing a head at 270 nm and two shoulders at 261 and 281 nm representing its conjugated triene (Figure 1B) and its MS/MS spectrum at m/z 359 [molecular mass M-H]^- finger-printing its 7S,14R and double-bond locations (Figure 1C). The MS/MS ions m/z 359 [M-H]^-, 341 [M-H-H₂O]^-, 323 [M-H-2H₂O]^-, 315 [M-H-CO₂], 297 [M-H-H₂O-CO₂]^-, and 279 [M-H-2H₂O-CO₂]^- were consistent with one carboxyl and a molecular weight (M) of 360 Daltons. The fragment ions m/z 217, 141, 123 [141-H₂O]^-, and 113 showed a hydroxyl at the 7 position (C₇). The fragment ions m/z 221, 203 [221-H₂O]^-, and 177 [221-CO₂]^-, demonstrated another hydroxyl at the 14 position (C₁₄). These data and our previous observation on P450 formation of 14R-HDHA [67] support our prediction that P450 or 5-LOX act together, generating 7S,14R-diHDHA.

Leukocytes are major blood cells that are essential not only in immunity against infection but also in the protection against or repair of organ degeneration and dysfunction caused by injury, diabetic complication, aging, or other adverse conditions. This is demonstrated by leukocyte production of prohealing resolvins, maresins [58], and/or 14,21-diHDHAs [55,72,73]. To further test our first hypothesis, we incubated human leukocytes (monocytes + neutrophils + lymphocytes) in 14R-HDHA containing medium as we conducted previously [54], then activated the cells with factors involved in inflammation and organ degeneration. The incubations were then studied using aR chiral LC-UV-MS/MS. 7S,14R-diHDHA was found to be the peak at the chromatographic retention time (RT) of 50.7 min (Figure 1D) with UV and MS/MS spectra (Figure 1E and 1F, respectively) matching those of 7S,14R-diHDHA generated by 5-LOX from 14R-HDHA (Figure 1). Of note, a small peak at RT 49.7 min was found to be a 7,14-diHDHA because its UV and MS/MS spectra match those of 7S,14R-diHDHA in Figure 1E and F. This 7,14-diHDHA is generated from added 14R-HDHA as purified leukocytes alone without feeding exogenous substrate did not show detectable 7,14-diHDHA; its 14-hydroxyl should be 14R-hydroxyl because the chirality of its 14-hydroxyl is preserved when undergoing such enzymatic conversion. It had RT 49.7 min, shorter than 50.7 min of 7S,14R-diHDHA, thus, its 7-hydroxyl should be 7R because the aR chiral LC RT is shorter for a fatty acid derived R-hydroxyl compound than for its S-epiomer when other structural features are the same (Figure 1A) on the basis of our extensive studies and publications from others [74,75]. The 7R-hydroxylation is likely to be catalyzed by P450 enzyme(s) in the leukocytes because P450 can catalyze both R and S hydroxylation. The exact mechanism of this unknown aspect is beyond the scope of this study. In brief, 7S,14R-diHDHA was the major product, and 7R,14R-diHDHA was the minor one when feeding leukocytes with 14R-DHA. These results and our previous finding that P450 converts DHA to 14R-HDHA converged to the biogenic pathway depicted in Figure 1G that human leukocytes produced 7S,14R-diHDHA, while cell-possessed P450 and 5-LOX catalyzed the biosynthesis.

3.2. 7S,14R-diHDHA enhanced MSC function to regulate nonfasting blood glucose levels

The db/db mice began to exhibit obesity by 5 weeks of age, and body weight increased rapidly with age [9]. 16-week old db/db mice weighed twice as much as the control C57BL/6J mice (45.3 ± 1.7 vs. 22.3 ± 0.7 g, p < 0.05, Figure 2A). The weight of non-diabetic control mice increased a little from 16 to 21 weeks old (Figure 2A). In addition, treatment with MSCs or 7S,14R-diHDHA-treated MSCs had no influence on weight (Figure 2A). The nonfasting blood glucose levels in db/db mice were much higher than in normal control C57BL/6J mice (640.0 ± 21.4 vs. 120.7 ± 9.3 mg/dL, p < 0.05, Figure 2B). At the end of intervention, the blood glucose level did not change in MSC infusion groups, but infusion of 7S,14R-diHDHA-condictioned MSCs reduced the blood glucose level compared to the untreated MSC group (Figure 2B). Blood insulin concentration analysis revealed increased blood insulin levels in the MSC group. Moreover, 7S,14R-diHDHA augmented the MSC effect on insulin secretion (Figure 2C), thus demonstrating that 7S,14R-diHDHA improved MSC amelioration of type 2 diabetic mellitus.

3.3. 7S,14R-diHDHA treatment promoted the capacity of MSCs to improve glucose tolerance in db/db mice

IPGTT was performed to determine the effect of the treatments of MSCs and 7S,14R-diHDHA-treated MSCs on glucose tolerance. The blood glucose concentration at 120 min after glucose load was significantly reduced by MSC treatment (1078.0 ± 23.4 vs. 1290.0 ± 12.2 mg/dL, p < 0.05, Figure 3). In addition, 7S,14R-diHDHA-treated MSCs were more effective than untreated MSCs in reducing glucose concentration (667.0 ± 66.8 vs. 1078.0 ± 23.4 mg/dL, p < 0.05, Figure 3). These results suggest that the improvement of glucose tolerance by MSCs and 7S,14R-diHDHA-treated MSCs was at least in part due to improvement of β-cell function.

3.4. 7S,14R-diHDHA induced MSC function to augment the ratio of β-cells and to reduce the ratio of α-cells in pancreatic islets

Figure 4 shows typical immunostaining patterns of pancreatic islet tissues in the nondiabetic normal C57BL/6J mice, diabetic db/db mice, and diabetic db/db mice infused with MSCs treated with or without 7S,14R-diHDHA. The islet of C57BL/6J mice, β-cells occupied most parts of the islets and existed at the center, whereas α-cells existed peripherally. However, db/db mice generally have hypertrophied pancreatic islets with a mixture of β-cells and α-cells in the islets. In addition, the ratio of β-cells to total islet cells in a pancreatic islet of db/db mice is less than that in C57BL/6J mice. Although MSC and 7S,14R-diHDHA-treated MSC infusions cannot change the distribution of β-and α-cells in the islets of db/db mice, MSC treatment can increase β-cell ratio compared with diabetic control mice (73.3 ± 1.4% vs. 67.8 ± 2.8%, p < 0.05, Figure 4), and 7S,14R-diHDHA can enhance MSC function to augment this ratio (78.2 ± 1.8% vs. 73.3 ± 1.4%, p < 0.05, Figure 4). In addition, 7S,14R-diHDHA can enhance MSC function to reduce α-cell ratio in islets of db/db mice (Figure 4).

3.5. 7S,14R-diHDHA enhanced MSC function to decrease the number of macrophages in islets.

Massive macrophage infiltration was observed in the islets of db/db mice compared to C57BL/6J mice, which suggests that inflammation may contribute to the dysfunctions of β-cells. In MSC-treated db/db mice, the number of macrophages in islets was reduced compared with nontreated db/db mice (Figure 5). Treatment with 7S,14R-diHDHA improved the anti-inflammatory action of MSCs.

3.6. 7S,14R-diHDHA augmented MSC function to increase min6 β-cell viability and insulin secretion

H₂O₂ dose-dependently reduced min6 β-cell viability was measured using the MTT method. When cocultured with MSCs, min6 β-cell viability was significantly increased under H₂O₂ condition (Figure 6A). Additionally, 7S,14R-diHDHA augmented the MSC effect on increasing min6 β-cell viability (Figure 6A). Furthermore, when min6 β-cells were treated with MSC conditioned medium, min6 β-cells secreted much more insulin (1.3 ± 0.1 vs. 0.9 ± 0.1 ng/mL, p < 0.05; Figure 6B), suggesting that MSCs can generate some trophic factors that trigger insulin release. 7S,14R-diHDHA-treated MSCs further enhanced insulin secretion from Min6 β-cell (1.7 ± 0.1 vs. 1.3 ± 0.1 ng/mL, p < 0.05; Figure 6B), which indicates that 7S,14R-diHDHA can promote MSCs to generate many more trophic factors acting on min6 β-cells to secrete more insulin.

3.7. 7S,14R-diHDHA treatment enhanced MSC secretion of trophic growth factors

MSCs generate an array of cytokines and growth factors [76,77]. Therefore, we postulated that 7S,14R-diHDHA treatment enhanced the beneficial effects of MSCs by augmenting their secretion of trophic factors. We observed that 7S,14R-diHDHA treatment increased the amount of the trophic growth factors VEGF and HGF secreted by MSCs into the cell culture medium (VEGF: 282.5 ± 9.6 vs. 233.0 ± 5.6 pg/mL, p < 0.05; 220.7 ± 13.2 vs. 152.3 ± 22.3 pg/mL, p < 0.05; Figure 7A, B).

3.8. 7S,14R-diHDHA treatment induced MSC function to decrease pericyte loss in the retina

Vasoregression is the primary pathogenic response of the retina to chronic hyperglycemia. Loss of capillary pericytes, followed by the formation of acellular, nonperfused capillaries, is an attempted compensation for retinal hypoxia [78]. Controlling mellitus can ameliorate diabetic complications; therefore, we want to know whether 7S,14R-diHDHA-treated MSCs can protect diabetic retinopathy by controlling hyperglycemia. The retinas of db/db mice showed a decrease in pericyte density compared with that of the C57BL/6J mice (Figure 8). However, MSCs can reduce pericyte loss in db/db mice. In addition, 7S,14R-diHDHA-treated MSCs can further increase the pericyte density compared to untreated MSCs (Figure 8), which suggests that 7S,14R-diHDHA enhance MSC paracrine function to ameliorate retinal pericyte loss in diabetes through adjusting hyperglycemia and acting pericyte directly.

4. Discussion

Leukocytes produced 7S,14R-diHDHA while P450 and 5-LOX catalyzed the biosynthesis

Taking our current results and previous findings [67] together revealed that leukocytes and their 5-LOX and P450 can produce a novel 7S,14R-dihydroxy-4Z,8E,10Z,12E,16Z,19Z-DHA or 7S,14R-diHDHA, a maresin-1 stereoisomer from DHA, as depicted in Figure 1G. We previously found that human P450, including 2D6 (CYP2D6), can catalyze the 14R-hydroxylation of DHA, where the contribution of autooxidation is negligible [67]. We observed that human leukocytes can convert 14R-HDHA to 7S,14R-diHDHA. Moreover, we observed that human 5-LOX catalyzed the biosynthesis of 7S,14R-diHDHA from 14R-HDHA by generating 7S-hydroxy in 14R-HDHA. Therefore, 5-LOX and P450 of leukocytes are accountable for the biogenesis of 7S,14R-diHDHA since 5-LOX is a major LOX and P450 is readily available in leukocytes [62,63,64]. The geometries of double-bonds outside of the region from C7 to C14 and the C14 chirality of 14R-HDHA are conserved after the transformation to 7,14-diHDHA, based on our results and reported analogous data for eicosanoids [48,79]. Thus, 7S,14R-diHDHA produced by leukocytes or by sequential catalysis of P450 and 5-LOX is 7S,14R-dihydroxy-4Z,8E,10Z,12E,16Z,19Z-DHA.

Harnessing mesenchymal stem cells by 7S,14R-diHDHA to ameliorate diabetic mellitus and retinal pericyte loss

The etiologies for type 1 and type 2 diabetes differ. Type 1 diabetes is a chronic autoimmune disease hallmarked by immune-mediated destruction of the pancreatic β-cells. T lymphocytes are aberrantly activated by the antigen-presenting cells in the pancreatic draining lymph nodes, and the activated T lymphocytes then circulate, target, invade the islets and destroyed the β-cells [80,81]. However, type 2 diabetes, hallmarked by underlying insulin resistance, is also characterized by defects in glucose-responsive insulin secretion in addition to an eventual decline in β-cell function [4,82], and oxidative stress caused by hyperglycemia is considered to be the main reason leading β-cell apoptosis [3,4,6,10,11,83].

MSCs have multiple differentiation and paracrine functions [25]. It has been confirmed that MSC transplantation into STZ-induced type 1 diabetic mouse and NOD models can enhance islet regeneration, lower blood sugar, and increase blood insulin levels [30,32,33,84]. Although MSCs can be recruited to the pancreas, less than 3% of cells migrate to the pancreas [31,85] confirmed the direct differentiation capacity of MSCs is not the main mechanism, which suggests that the paracrine function may confer tissue-repairing function to MSCs. MSCs can generate many bioactive factors, such as TGF-α, HGF, IDO, PGE₂, which play important roles in modulating immuno-cell functions, such as inhibiting T cell proliferation, dendritic cell activation, and inflammatory cytokine generation [86,87]. On the basis of generation immunomodulating factors, MSCs inhibit the onset of type 1 diabetes and reverse β-cell functions [23,88]. However, type 2 diabetes is different from type 1 diabetes, and immunomodulating functions may not be the reason for MSCs repairing damaged β-cells in type 2 diabetes. Increased leukocyte infiltration is associated with the onset of type 2 diabetes [89]. Depletion of macrophages by clodronate in liposomes reduced pancreatic invasion of macrophages and destruction of islet β-cells [90,91]. Our results revealed that MSCs can inhibit macrophage infiltration into islets. The increased oxidative stress observed in islets of diabetic states seems to be one of the main factors of deteriorated β-cell function. We have found that MSCs can inhibit macrophage generation of the inflammatory cytokines TNF-α and ROS [57], which indicates the anti-inflammatory functions of MSC protecting β-cell functions in type 2 diabetes.

The islet capillary network is about five times denser than the capillary network of exocrine tissue. The enriched vascularity meets the requirement for the quick secretion of insulin from β-cell in response to blood glucose levels [92]. In diabetic db/db mice, the number of endothelial cells was reduced compared with normal nondiabetic mice. VEGF is one of the most important angiogenic cytokines[93]. We have confirmed that MSCs can generate VEGF and promote angiogenesis in wounds, which promotes wound healing in db/db mice [54]. MSCs promote vascularity in islets of db/db mice may also through generating VEGF.

DHA is a typical ω-3 long-chain poly-unsaturated fatty acid, it can be transformed enzymatically into potent inflammation-resolving lipid mediators, such as Resolvin D series [41], neuroprotectin/protectins [46,47,48,49], maresins [58,59], maresin-likes [67], and 7S,14R-diHDHA. We also found a DHA-derived lipid mediators 14S,21R-diHDHA, which can enhance MSC functions to promote wound healing and repair acute kidney injury [54,57]. 7S,14R-diHDHA is a new lipid mediator that has a potential anti-inflammatory function. In this study, we observed that 7S,14R-diHDHA can promote MSC paracrine function to generate VEGF and HGF. HGF has anti-inflammatory properties; it can increase IL-10 levels and reduce IL-1β levels in pancreatitis or diabetes [94,95,96]. These data suggest that 7S,14R-diHDHA promotes MSC functions to ameliorate diabetic mellitus mainly by enhancing MSC paracrine functions. These results are consistent with reports that maresins promote the beneficial bioactions of stem cells in the treatment of diseases [59]. Maresin 1 regulates insulin signaling in human adipocytes as well as in adipose tissue and muscle of lean and obese mice [59]. Maresin 1 inhibits hyperglycemia-induced ferroptosis [60].

Diabetic retinopathy is a clinically well-defined, sight-threatening, chronic microvascular complication that eventually affects virtually all patients with diabetes. Diabetic retinopathy is characterized by gradually progressive alterations in the retinal microvasculature, which start with the loss of the two cellular components of retinal capillaries: the pericyte, a vessel support cell, and the endothelial cell, and pericytes disappear before endothelial cells start to vanish [39,40,97]. db/db mice develop hyperglycemia starting at 8-week age as a result of excessive food consumption [98,99]. They show early signs of diabetic retinopathy, such as thickening of capillary basement membrane at 22 weeks [100] and loss of pericytes in retinas at 26 weeks, followed by endothelial cell loss at 34 weeks [101]. In our study, pericyte loss occurred as early as 22 weeks of age in db/db mice. MSCs can inhibit pericyte loss. Moreover, 7S,14R-diHDHA-treated MSCs further markedly increased pericyte density compared with untreated MSCs. This suggests that 7S,14R-diHDHA enhances MSC paracrine function to ameliorate retinal pericyte loss in diabetes by adjusting for hyperglycemia and/or acting pericyte directly. This enhancement is likely to occur through MSC secretion of VEGF, as 7S,14R-diHDHA can significantly promote MSC production of VEGF (Figure 7), and VEGF is thought to be an important cytokine inducing vasoregeneration in diabetic retinopathy [78]. When MSCs were infused into diabetic mice, MSCs have an anti-inflammatory function, reducing oxidative stress. Under this condition, VEGF or HGF generated from MSCs may trigger the PI3K/AKT signaling pathway and maintain retinal vascular integrity [102,103], while 7S,14R-diHDHA treatment can enhance MSC function via this pathway. The mechanism for the amelioration of diabetic damage of retinal pericytes by 7S,14R-diHDHA-treated MSCs is also likely attributable to their amelioration of diabetes, especially diabetic hyperglycemia.

Future directions

The scope of this pilot study provides multiple opportunities for future research. 5-LOX could convert DHA to 7S-hydroxyl DHA first, followed by P450 14R-hydroxylation in sequential enzymatic catalysis to generate 7S,14R-diHDHA from DHA. However, the exact P450 enzymes involved in 14R-hydroxylation are unknown. The appearance of 7S,14R-diHDHA in the blood, pancreas, and other organs in vivo in animals and humans under healthy and diabetic conditions is likely to be transient, which needs to be studied systematically at different time points. We will study the actions of 7S,14R-diHDHA alone. Future studies should also evaluate the effect of our treatments on the other aspects of diabetes and on diabetic retinopathy.

5. Conclusions

We identified a novel DHA-derived 7S,14R-dihydroxy-4Z,8E,10Z,12E,16Z,19Z-DHA (7S,14R-diHDHA), a maresin-1 stereoisomer, biosynthesized by leukocytes and related P450 and 5-LOX enzymes [62,63,64]. 7S,14R-diHDHA promoted MSC paracrine functions in db/db mice in vivo to ameliorate diabetic mellitus through improving β-cell function, lowering blood sugar, improving glucose tolerance, and increasing blood insulin levels. It augmented the proportion of β-cells and reduced that of α-cells to total islet pancreatic cells, and decreased the number of macrophages in islets in db/db mice in vivo. Also, the lipid mediator enhanced MSC function to increase min6 β-cell viability and insulin and MSC secretion of trophic growth factors. Furthermore, 7S,14R-diHDHA treatment induced MSC function to decrease pericyte loss in the retinas of diabetic db/db mice in vivo. Taken together, pretreatment of MSCs with 14S,21R-diHDHA or related LMs or structural mimics may offer a new clinical strategy for improved treatment of diabetic mellitus and diabetic complications.