Alterations in visible light exposure modulate platelet function and regulate thrombus formation

Abstract

Background:

Variations in light exposure are associated with changes in inflammation and coagulation. The impact of light spectra on venous thrombosis (VT) and arterial thrombosis is largely unexplored.

Objectives:

To investigate the impact of altering light spectrum on platelet function in thrombosis.

Methods:

Wild-type C57BL/6J mice were exposed to ambient (mice^white, 400 lux), blue (mice^blue, 442 nm, 1400 lux), or red light (mice^red, 617 nm, 1400 lux) with 12:12 hour light:dark cycle for 72 hours. After 72 hours of light exposure, platelet aggregation, activation, transcriptomic, and metabolomic changes were measured. The ability of released products of platelet activation to induce thrombosis-generating neutrophil extracellular trap formation was quantified. Subsequent thrombosis was measured using murine models of VT and stroke. To translate our findings to human patients, light-filtering cataract patients were evaluated over an 8-year period for rate of venous thromboembolism with multivariable logistic regression clustered by hospital.

Results:

Exposure to long-wavelength red light resulted in reduced platelet aggregation and activation. RNA-seq analysis demonstrated no significant transcriptomic changes between mice^red and mice^white. However, there were global metabolomic changes in platelets from mice^red compared with mice^white. Releasate from activated platelets resulted in reduced neutrophil extracellular trap formation. Mice^red also had reduced VT weight and brain infarct size following stroke. On subgroup analysis of cataract patients, patients with a history of cancer had a lower lifetime risk of venous thromboembolism after implantation with lenses that filter low-wavelength light.

Conclusion:

Light therapy may be a promising approach to thrombus prophylaxis by specifically targeting the intersection between innate immune function and coagulation.

1 |. INTRODUCTION

Venous thromboembolism (VTE) is estimated to affect up to 10 million people worldwide each year [1]. Despite broad efforts to improve VTE prophylaxis protocols, VTE is the leading cause of preventable hospital death worldwide [2]. There is a need for novel approaches to prevention of VTE to improve hospital outcomes. Arterial thrombosis is similarly ubiquitous; in 2019 alone, there were 12.2 million incident cases of stroke worldwide [3]. Treatment and prophylaxis for each of these thrombotic diseases is suboptimal, as anticoagulants and thrombolytic therapies are often delayed or contraindicated due to the burden of bleeding [4–6].

The risk of arterial thrombotic events, such as stroke and myocardial infarction, has been found to oscillate with the day-light cycle [7–9]. There is also evidence that light modulates incidence of VTE in human populations [10,11]. Environmental light has been found to regulate many biological processes including cardiovascular homeostasis, hormone secretion, and metabolism [12–14]; all of these circadian physiologies exhibit oscillations that mirror the solar day and night. Previous studies have defined the physiology and biology through which light spectrum, illuminance, and duration modulate the inflammatory and immune response to sterile and septic insults. For example, in murine models of sepsis and ischemia-reperfusion, mice exposed to short-wavelength blue light exhibited improved bacterial clearance and survival and reduced tissue injury [15,16]. Additional evidence suggests that alterations in environmental light exposure may regulate circulating markers of thrombogenesis and fibrinolysis in both humans and rodents to optimize coagulation and reduce bleeding during periods of greatest activity and risk of injury [7,17–20].

Many of the mechanisms of arterial and venous thrombosis are shared. For example, the intersection of innate immune function and coagulation has been well described in both venous and arterial thrombosis [21,22]. Specifically, platelet interactions with neutrophils have been implicated as the key interface between inflammation and coagulation [23,24]. Further, neutrophils and megakaryocytes show inherent circadian oscillations [25,26]. It follows that altering circadian regulation of platelet–neutrophil interactions may augment coagulation. These interactions represent a novel area to target for alternative preventative strategies. Given previous data suggesting an effect of light exposure on both immunity and coagulation, we sought to investigate the impact of altering light spectrum in thrombosis. We hypothesized that alterations in wavelength of light exposure may alter arterial and venous thrombosis by influencing the interaction between innate immunity and coagulation.

2 |. RESULTS

2.1 |. Selective exposure to long-wavelength red light reduces venous thrombogenesis

To investigate the role of light on venous thrombosis, we used a validated murine model of VTE [27]. Mice were exposed to red light (peak 617 nm) (mice^red), blue light (peak 442 nm) (mice^blue), or white fluorescent light (100% transmittance; to represent the fluorescent light of hospital rooms [15]) (mice^white) for 72 hours (maintaining a standard 12-hour light/dark cycle) prior to inferior vena cava (IVC) ligation (Figure 1A, B). Twenty-four hours following IVC ligation, thrombus mass was significantly lower in mice^red (4.21 mg ± 5.11 mg) compared with mice^white (18.54 mg ± 5.97 mg; P < .0001) or mice^blue (21.46 mg ± 8.52 mg; P < .0001, Figure 1C).

FIGURE 1.

Prophylactic long-wavelength light exposure reduces thrombogenesis. (A) Schematic demonstrating the visible light spectra. Transmittance measurements from red and blue filters using a Thermofisher evolution ultra-violet visible spectrophotometer, Evolution 600. (B) Schematic of the IVC ligation model. (C) Thrombus weight (milligrams, mg) following IVC ligation after exposure to white (gray circles), red (red squares), or blue (blue triangles) filtered light. Mean ± SEM, n = 13, ANOVA. Gross images of harvested thrombi from white light (upper panel), red light (center panel), or blue light (lower panel) exposed mice. (D) Thrombus weight following IVC ligation after exposure to preligation-light exposure or postligation-light exposure. Mean ± SEM, n = 5, ANOVA. (E) Gross images of control mice (129S1 WT) and blind mice (129S1/Vsx2^or-J). (F) Thrombus weight (milligrams, mg) following IVC ligation in 129S1/Vsx2^or-J mice after exposure to white, red, or blue filtered light. Mean ± SEM, n = 8, ANOVA. ANOVA, analysis of variance; IVC, inferior vena cava.

To further study the timing of altered lighting relative to thrombotic insult and thrombus size, mice were exposed to light either exclusively pre- or post-IVC ligation. Preligation exposure to red light resulted in significantly reduced clot weight compared with preligation white light exposure (mice^red 7.20 mg ± 6.83 mg vs mice^white 18.00 mg ± 3.39 mg; P < .01) and blue light exposure (23.60 mg ± 3.65 mg; P < .01). However, postligation light variation resulted in no significant differences between groups (Figure 1D). These findings suggest that selective exposure to long-wavelength red light, when applied to mice prior to a thrombotic insult, reduces thrombosis.

To determine whether these observations required an optic pathway as opposed to direct exposure to blood, we tested the murine model of venous thrombosis in a genetic model of blindness. Visual system homeobox 2 (Vsx2)^or-J/J mice exhibit completely degenerated retina by 12 weeks of age, as well as closed eyelids, resulting in blindness and inability to sense light (Figure 1E). Contrary to the wild-type mice, Vsx2^or-J/J mice displayed similar thrombus mass after exposure to red light (20.50 mg ± 4.87 mg) compared with white light (20.75 mg ± 3.37 mg, P = .99) and blue light (18.38 mg ± 3.38 mg, P = .54; Figure 1F). To test whether this was a platelet-autonomous effect, platelets from healthy human volunteers were exposed ex vivo to ambient or red light in a microfluidic thrombosis assay. No difference in thrombosis was observed based on the light wavelength used (Supplementary Figure S1). These data suggest that the reduction in thrombus burden seen with exposure to red light is optically mediated.

Next, we evaluated whether circadian activity, sleep, or metabolism changed with alterations in light characteristics and thus, potentially contributed to the observed differences in thrombus size. Activity levels, sleep-wake cycle, caloric intake, mass, and core body temperature were similar between groups (Supplementary Figure S2). Tail bleeding, a measure of hemostasis [28], as well as conventional laboratory measurements of coagulation (prothrombin time [PT]; partial thromboplastin time [PTT]; international normalized ratio [INR]) and fibrinolysis (D-dimer; plasminogen activator inhibitor-1 [PAI-1]) were not significantly different (Supplementary Figure S3). Steroid hormones (cortisol, estradiol, progesterone, T3, T4, and testosterone) and metabolic hormones (amylin, c-peptide, ghrelin, gastric inhibitory polypeptide, glucagon-like peptide 1, glucagon, insulin, leptin, pancreatic polypeptide, peptide YY, secretin) were also not significantly different between groups (Supplementary Figures S4 and S5).

2.2 |. Prophylactic, selective exposure to long-wavelength red light reduces ischemic stroke brain injury

To extend our findings beyond venous thrombosis, we used a well-characterized murine model of ischemic stroke [24] to characterize the effect long-wavelength light exposure has on arterial thrombosis. Twenty-four hours following middle cerebral artery (MCA) occlusion (Figure 2A), neurological and motor outcomes were measured, and brain infarct size was quantified. In line with the results from the VTE model, mice^red had significantly reduced brain infarct sizes (38.78 mm³ ± 17.55 mm³) compared with mice^white (73.31 mm³ ± 15.00 mm³; P < .01) or mice^blue (72.87 mm³ ± 26.89 mm³; P < .001; Figure 2B, C). This reduction in brain injury was accompanied by improved neurological function, as assessed by the Bederson test (Figure 2D) and the murine neurological severity score (Figure 2E). Similarly, motor function was improved in mice^red compared with mice^blue, as measured by rotarod testing (Figure 2F, G). Mice^red were compared with mice^white for each of the future experiments in the interest of reducing animal utilization and with the focus on the impact of high-wavelength light observed in these models.

FIGURE 2.

Prophylactic long-wavelength light exposure reduces thrombogenesis. (A) Schematic of the middle cerebral artery (MCA) occlusion model. (B) Stroke burden (infarct volume, mm²) following MCA occlusion after exposure to white (grey circles), red (red squares), or blue (blue triangles) filtered light. Mean ± SEM, n = 10–13, ANOVA. (C) Gross images of harvested brain specimens from white light (left panel), red light (center panel), or blue light (right panel) exposed mice. (D) Benderson score, (E) murine neurological severity score (F) Latency to turn and (G) latency to fall following MCA occlusion after light exposure. Mean ± SEM, n = 9–13, ANOVA. ANOVA, analysis of variance; MCA, middle cerebral artery.

2.3 |. Selective exposure to long-wavelength red light reduces neutrophil extracellular trap formation

Previous reports have shown the importance of neutrophils in thrombosis and the formation of neutrophil extracellular traps (NETs) in thrombus formation (NETosis) [24,29,30]. Figure 3A, B demonstrates immunofluorescence images of harvested thrombus following light exposure and IVC ligation. Immunostaining of platelet and fibrinogen deposition showed no significant differences between groups Supplementary Figure S6. Despite similar levels of neutrophils in each harvested sample (Figure 3C), there are reduced NETs in the IVC-thrombus in mice^red compared with mice^white (Figure 3D) as measured by citrullinated histone 3 (citH3). The number of nucleated cells was not different between groups (P = .78). However, given this strategy of quantifying NETosis in nucleated cells, there is the potential for under-reporting any prior or ongoing NETosis at time of fixation. Thus, we also performed MPO-DNA ELISA to quantify the levels of NETs in the thrombus. Within the thrombus, red light-exposed mice had lower levels of MPO-DNA complex than ambient light-exposed mice (optical density values mice^red: 0.75 ± 0.45 vs mice^white 1.52 ± 0.35, P = .01; Figure 3H). Similarly, brain tissue from mice^red after MCA occlusion displayed lower levels of NETs than the brain tissue from mice^white (Figure 3F–I). Immunostaining of brain tissue revealed no change in the deposition of platelets (Supplementary Figure S7).

FIGURE 3.

Long-wavelength light exposure reduces NET accumulation in the thrombus and reduces platelet-induced NETosis. (A) Representative immunofluorescence image of a thrombus harvested from a white light-exposed mouse. Upper panel displays max projections of a confocal stack with all labeled channels. Lower panels display max projections of a confocal stack with individual channels labeled. Cellular nuclei in blue (Hoechst+), neutrophils in green (Ly6G), and citH3 in red. Scale = 100 μm. (B) Representative immunofluorescence image of a thrombus harvested from a red light exposed mouse. Scale = 100 μm. (C) Relative number of NETs per nucleated cell in white light (gray circles) compared with red light (red squares) exposed mice. NETs calculated in thrombus as colocalization of Hoescht, LY6G, and citH3 and standardized to Hoescht positive cells. All images had minimum 5 neutrophils per high powered field. Mean ± SEM, n = 9, Student’s t-test. (D) Relative number of neutrophils per nucleated cell. Neutrophils calculated in thrombus as colocalization of Hoescht and LY6G and standardized to Hoescht positive cells. All images had minimum 5 neutrophils per high powered field. Mean ± SEM, n = 9, Student’s t-test. (E) MPO-DNA complex in thrombus quantified with a capture ELISA. Data presented as optical density (OD) of light-exposed thrombus standardized to negative tissue control. (F) Representative immunofluorescence image of a brain harvested from a white light-exposed mouse. Scale = 20 μm. (G) Representative immunofluorescence image of a brain harvested from a red light exposed mouse. Scale = 20 μm. (H) Relative number of neutrophils per mm². Neutrophils calculated in thrombus as colocalization of DAPI and LY6G. Mean ± SEM, n = 4, Mann–Whitney U test. (I) Relative number of NETs per mm². NETs calculated in thrombus as colocalization of DAPI, LY6G, and citH3. Mean ± SEM, n = 4, Mann–Whitney U test. (J) NETosis was measured after coincubation of healthy control neutrophils with platelet releasate from ambient or red light exposed mice. Data are presented as MPO-DNA complexes after 4-hour coincubation with platelet releasate standardized to vehicle control. Mean ± SEM, n = 10, ANOVA. (K) NETosis was measured after coincubation of neutrophils from ambient or red light exposed animals with either PMA or healthy control platelet releasate. Data are presented as MPO-DNA complexes after 4-hour coincubation with platelet releasate standardized to vehicle control. Mean ± SEM, n = 6, 1-way ANOVA. ANOVA, analysis of variance; citH3, citrullinated histone 3.

The effect of red light on neutrophil function was further explored. Neutrophils from mice^red and mice^white were stimulated with phorbol myristate acetate (PMA) to induce NET formation. PMA-stimulated neutrophils from mice^red formed NETs at similar levels as PMA-stimulated neutrophils isolated from mice^white, suggesting that selective exposure to high-wavelength red light had no direct effect on neutrophil function (Figure 3J). As platelets are critical regulators of NET formation during venous thrombosis, we next asked if platelet releasate (released products of platelet activation) [31] from mice^red or mice^white altered NET formation in neutrophils isolated from wild-type (WT) animals. Platelet releasate isolated from mice^red and applied to neutrophils from mice^white [31] resulted in lower levels of NETosis compared with platelet releasate from mice^white (fold increase from WT: mice^red 1.08 ± 0.34 vs mice^white 2.14 ± 0.58, P < .01; Figure 3J). Importantly, platelet releasate from WT mice had a similar response on neutrophils isolated from either mice^red or mice^white, further indicating light exposure altered platelet-induced NET formation due to changes in platelet behavior (Figure 3K).

2.4 |. Selective exposure to long-wavelength red light reduces platelet aggregation and activation

As platelet–neutrophil interactions are key regulators of NET formation in both venous and arterial thrombosis [24,27] and form in response to platelet activation, we asked if changes in light exposure altered platelet activation to mediate the observed reduction in NETs after venous thrombosis.

2.4.1 |. Aggregation

Despite similar platelet counts (Figure 4A), platelets from mice^red had reduced activation in response to collagen and adenosine diphosphate (ADP) compared with platelets from mice^white (collagen: mice^red 16.00 ohm ± 4.05 ohm vs mice^white 23.83 ohm ± 5.57 ohm, P = .02; ADP: mice^red 9.83 ohm ± 3.31 ohm vs mice^white 13.17 ohm ± 1.33 ohm, P = .04; Figure 4B). Cumulative aggregometry curves are displayed in Figure 4C.

FIGURE 4.

Long-wavelength light exposure alters platelet function. (A) Platelet count (×10³/μL) of white and red light exposed mice. (B) Aggregation (ohm) following white light (gray circles) or red light (red squares) exposure. Whole blood from each mouse was separately tested after activation with collagen, adenosine diphosphate (ADP), or thrombin. Mean ± SEM, n = 5–6, ANOVA. (C) Platelet aggregometry curves following activation with collagen. Mean ± SEM, n = 6. (D) Surface level activation of platelets following white or red light exposure. Isolated platelets from each mouse were treated with no agonist, convulxin, or thrombin. Data are presented as percent of CD41+ platelets positive for CD62 (P-selectin) or (E) Jon/A (GPIIbIIIa). Mean ± SEM, n = 4–5, ANOVA. (F) Representative cytograms for platelet activation with thrombin agonist. Y-axis represents CD41+ cells, X-axis represents CD62P+ and Jon/A+ cells. (G) Metabolomic comparisons for platelets isolated from mice^red or mice^white. Metabolite levels normalized and log transformed. N = 5. (H) Single-cell sequencing of megakaryocyte progenitor cells revealed differential expression in inflammation, oxidative stress, cytoskeleton, and coagulation related pathways. N = 3. padj, adjusted P value; NES, normalized enrichment score.

2.4.2 |. Activation

We also assessed platelet activation by measuring the expression of both CD62P+ (P-selectin) and JonA (GPIIbIIIa) on CD41+ platelets and found activation was reduced in platelets from mice^red compared with mice^white. Convulxin-activated platelets displayed lower levels of P-selectin (35.17% vs 53.17%; P = .02) and GPIIbIIIa (77.49% 88.63%; P = .02) in mice^red compared with mice^white (Figure 4D, E). Thrombin activation also resulted in lower levels of platelet activation in mice^red than in mice^white (P-selectin: 42.77% vs 68.51%, P = .02; GPIIbIIIa: 52.85% vs 85.12%, P < .01; Figure 4E). A representative gating strategy and cytograms are displayed in Figure 4F.

2.4.3 |. Transcriptomics

To determine if red light exposure altered platelet function via transcriptional changes, we performed RNA-seq analysis on CD41-bead isolated platelets from red and blue light-treated animals. RNA-seq analysis demonstrated no significant changes between red and blue light-treated animals, suggesting red light did not markedly alter the platelet transcriptome. The publicly available link for transcriptomics is available in Supplementary Material.

2.4.4 |. Metabolomics

Platelets were isolated and evaluated for metabolomic changes. There were global metabolome level changes in platelets exposed to red light compared with ambient light. After normalization, fatty acid metabolism, specifically, was significantly different in mice^red compared with mice^white. Oleic acid and linoleic acid, factors that are known to decrease platelet aggregation and activation [32,33], were significantly higher in mice^red compared with mice^white (oleic acid 1.58 fold increase, P = .03; linoleic acid 1.49 fold increase, P = .03; Figure 4G).

2.4.5 |. Single-cell sequencing

To determine where the observed changes in platelet activity occur, megakaryocyte progenitor cells were isolated from bone marrow and evaluated after differential light exposure. Supplementary Figure S8 displays the positive and negative selection methods to distinguish megakaryocyte progenitors. Mice^red showed significantly up-regulated pro-inflammation, cell death, and oxidative stress profiles in megakaryocyte progenitors compared with blue light (Figure 4H). This response was also significant in ambient light exposure but to a lesser extent with lower number of enriched pathways and normalized enrichment scores. Megakaryocyte progenitors from mice^red also showed significantly suppressed actin binding, extracellular binding and organization, platelet aggregation/activation, Rho GTPase activation (master regulators of the platelet cytoskeleton and platelet function) [34] and, MAPK (platelet activator via integrin) and adhesion molecule interactions [35] (Figure 4H). This response was also significant in ambient light exposure but to a lesser extent with lower number of enriched pathways and normalized enrichment scores. Specific surface markers for activated platelets such as von Willebrand factor and platelet factor 4 (PF4) were significantly downregulated in red light-exposed animals (Supplementary Figure S9) [36,37].

2.5 |. Filtration of blue light reduces long-term VTE risk after cataract surgery

To identify whether these murine findings were translatable to human pathophysiology, we studied the incidence of thrombosis after cataract surgery in patients who received either blue light filtered intraocular lenses (BF-IOL), which results in approximately 50% less transmission of short-wavelength blue light, or conventional intraocular lenses (C-IOL), which transmit the entire visible spectrum (ie, restoration of blue light) [38]. Furthermore, 10 464 patients were included in the analysis (Figure 5A). Mean age was 75.1 (SD, 6.5) years and 6679 (63.8%) were women; 7035 (67.2%) underwent implantation with BF-IOL and 3429 (32.8%) underwent implantation with C-IOL (Figure 5B). The median duration of follow-up was 5.1 (IQR, 3.5–6.8) years. For the entire cohort, subjects implanted with BL-IOL by comparison with C-IOL had a lower adjusted risk of VTE (BL-IOL, 194 patients; C-IOL, 86 patients), though this did not attain statistical significance: adjusted hazard ratio (aHR), 0.91; 95% CI, 0.61 to 1.36; P = .65 (Figure 5C, D).

FIGURE 5.

Long-wavelength light exposure protects highest risk patients from VTE development. All patients who underwent cataract surgery at a multihospital health care center were evaluated for incidence of VTE (composite outcome of deep venous thrombosis and pulmonary embolism) development. (A) Flow diagram representing inclusion and exclusion criteria. (B) Patient characteristics for each lens group. (C) Cox regression curve demonstrating VTE free survival after cataract surgery (intraocular lenses) (D) Adjusted hazard ratios for risk of VTE after implantation with a blue light filtered lens. (E) In subgroup analyses, blue filtered lenses result in lower incidence of VTE development in patients with history of cancer. Data presented as a forest plot of relevant odds ratios related to venous thrombosis development following cataract surgery. n = 10 464. VTE, venous thromboembolism.

As a planned analysis, we a priori defined several subgroups for which the literature supported elevated risk of VTE [39–46]. In all, BL-IOL was associated with a reduced risk of VTE relative to C-IOL, which was significant for cancer (aHR, 0.53; 95% CI, 0.33–0.83; P = .01; P_interaction = .052). In patients with diabetes, BL-IOL was compatible with increased risk of VTE, however, this was not significant (aHR, 0.92; 95% CI, 0.57–1.13; P = .11) (Figure 5C, D).

3 |. DISCUSSION

Light is an integral regulator of inflammation and immune function, and ongoing investigations aim to elucidate the effects of light spectrum on mammalian biology and physiology. We and others have reported on the mechanisms by which interactions between innate immunity and platelet function contribute to thrombosis [23,24]. Thus, the study of light exposure on the intersection of immune activity and coagulation remains critical to understanding thrombogenesis. The present study reports that selective exposure to long-wavelength red light results in a significant reduction in thrombosis. This is consistent across differential exposure and animal breeds. We found reduced platelet aggregation and activation leading to a reduction in platelet-driven NET release by neutrophils, together resulting in a lower burden of thrombosis. This represents a unique way to modulate platelet–neutrophil interactions and subsequent thrombosis.

In both venous and arterial pathologies, we observed significant reductions in thrombus burden with prophylactic selective exposure to long-wavelength red light. An optic pathway is integral to this observation, as these effects were not observed in genetically blind mice or through direct exposure of light to blood flowing ex vivo. The mechanism of optic-mediated effects on platelets is not yet known and is topic of ongoing investigation. Multiple metabolomic, but not transcriptomic, changes were observed when studying platelets from red light-exposed mice. Specifically, fatty acid metabolism was altered, suggesting its role in light-modulated platelet function. Fatty acids have previously been implicated in platelet function by reducing PIP2/PIP to prevent aggregation induced by platelet-activating factor [32,47]. Mechanistically, red light has been found to increase fatty acid production, specifically oleic acid, which is known to reduce platelet activation [48]. With regards to the single-cell sequencing of megakaryocyte progenitor cells, light exposure altered the expression of pathways involving pro-inflammation, cell death, and oxidative stress profiles, all of which have been shown to be integral in platelet-mediated thrombosis [49,50]. These findings suggest possible mechanisms by which light affects platelet function in vivo. However, given multiple observed changes across many mechanistic pathways, the effect of light is most likely global and related to many cellular systems.

The importance of light as a regulator of biological processes has been well documented [15,16,51–53]. When exposed in vivo, the effects are often the result of proposed optic-neurophysical changes. Alternatively, direct in vitro red light exposure has been implicated as an anticoagulant to reduce device-related thrombosis [54]. However, no study has linked in vivo platelet function to the wavelength of light exposure, which has immediate translational potential as an alternative to thrombosis prophylaxis. Given the consequences of circadian gene activity are heavily reliant on optical exposure to light, it follows that platelet-mediated thrombosis may be modified by light wavelength transmission through an optical-circadian axis. Given circadian genes have been found to influence many different systems, the effect of visual light modification is likely multifactorial [12,55,56]. Further studies are necessary to elucidate the exact mechanism by which light modulates thrombosis. In our retrospective study of human cataract patients, while there were no differences in thrombosis risk for the entire cohort, cancer patients reduced rate of VTE. Cancer is associated with a 9-fold increase in VTE risk, due to a systemic hypercoagulable and proinflammatory state [57]. We propose that given this, the effect of light on risk of VTE is greater in magnitude. However, given the unique biology of cancer and its associated immunomodulation, there could be a separate mechanism placing them at higher risk.

The mouse retina does differ from the human retina, in being deficient in red light–sensitive cones; thus, it has long been assumed that mice perceive less brightness at longer wavelengths exceeding 620 nm [58]. However, several independent studies support that rodents still see the external environment in red light and are not red light blind [59,60]. We observed the preservation of circadian oscillation in thermoregulation, sleep-wake cycles, and eating in mice exposed to red light. Thus, red light cannot be equated with no light (ie, complete darkness throughout the light:dark cycle); rather, it differs from blue or white light in how it regulates the optic-neurophysiological pathways that govern platelet biology. Nonetheless, it remains unknown whether the observed impact of light wavelength on thrombosis is derived from the selective exposure to high-wavelength light or, rather, the absence of lower wavelengths. Whether red light reduces thrombosis or whether lower wavelengths promote thrombosis is not answered by our data presented here.

Several limitations exist for this study. While we demonstrate changes to platelet function and subsequent thrombosis are attenuated by light exposure through an optical pathway, future studies are needed to determine the specific mechanistic changes within a platelet that occur in response to prophylactic red light. Additionally, the murine models of thrombosis that are used here are primarily inductive, and we are therefore unable to determine the true natural incidence of venous or arterial thrombus after red light exposure. To elucidate this uncertainty, future studies will need to pilot selective exposure to high-wavelength light in high-risk human patients to determine its prophylactic antithrombotic effects. Regarding the human cataract cohort analysis, there are additional limitations. Given the retrospective nature of this analysis, there is an inherent confounding bias that is unavoidable. The authors have controlled for this by adjusting for patient-level factors to avoid unanticipated confounding effects. Additionally, the human venous data capture the effects of chronic changes in light exposure to thrombotic risk, whereas the animal models tested represent acute changes in light exposure. Future studies are required to test the effects regarding the timing of exposure.

In summary, we report a thrombotic prophylactic strategy through exposure to long-wavelength red light. We demonstrate that this observation is associated with reduced platelet activation and subsequently reduced NET formation, leading to reduced thrombus burden in both venous and arterial models. Furthermore, we find that red light–mediated protection against thrombosis is driven optically by altering optical exposure to light in mice. We speculate based on the lower thrombotic risk we observed in cancer patients with blue light-blocking lenses that humans can similarly manifest an antithrombotic phenotype based on ocular input. Thus, altering and optimizing light exposure is a promising target for thrombus prophylaxis and a potentially impactful future direction for clinical use.

4 |. METHODS

4.1 |. Animals

C57BL/6J mice and 1129S1/Sv-Vsx2^or-J/J mice were purchased from Jackson Laboratories (The Jackson Laboratory). 1129S1/Sv-Vsx2^or-J/J mice contain a mutation in Vsx2 (visual system homeobox 2) resulting in phenotypic blindness due to microphthalmia, early inhibition of retinal development, and absence of the optic nerve. Mice were housed in accordance with the University of Pittsburgh (Pittsburgh, Pennsylvania, United States) and National Institutes of Health (Bethesda, Maryland, United States) animal care guidelines.

4.2 |. Light exposure

All experiments were performed in a room maintained on the same day-night cycle the animals were housed in (12 hours light, 12 hours dark, lights on from 07:00 to 19:00) at ambient temperature of 25 °C ± 2 °C. Animals were given unlimited access to water and pellets. We tested red light, blue light, and white light (to represent the fluorescent light of hospital rooms [15]). Two Day-Light Classic DL930 light (3 × 36–W compact fluorescent: 4000 Kelvin, 10 000 lumens at 30 cm) were used (Uplift Technologies) as previously described [31]. Lamps were fitted with blue (peak emission 442 nm) and red (peak emission 617 nm) filters (Lee Filters) and positioned over each cage to ensure identical illuminances of approximately 1400 lux when measured at the middle of the cage; we have previously reported that the intensities are also similar [12]. Light transmission was confirmed with spectrophotometry (ThermoFisher, Figure 1A). We included a third lighting condition: white fluorescent lighting, similar to that found in hospital settings (400 lux, no predominant spectrum). We measured room and cage temperatures and determined only a modest heat effect from the lights with no differences noted between the different wavelength exposures.

Mice were exposed to the light of interest for 72 hours prior to IVC ligation, carotid injury, or euthanization. This time point was chosen to maximize number of circulating platelets exposed to exclusively augmented light, while avoiding a circadian shift in the animals’ cycle [61,62]. For experiments that did not involve IVC ligation, mice were euthanized after 72 hours of light exposure and whole blood was collected as described. For experiments involving IVC ligation or arterial occlusion, mice recovered with continued light exposure for 24 hours until euthanization. Throughout the experiments, mice were maintained on 12 hours light, 12 hours dark cycle. Illuminance for each experiment was measured and confirmed with a hand-held digital lux meter (Digital Light Level Meter LX1330B, Mastech).

4.3 |. Venous thrombosis model

C57BL/6J, Vsx2/J, or Vsx2^or-J/J male mice, littermates, age 10 to 14 weeks, were utilized in all experiments. VTE was induced using a validated model of venous stasis, IVC ligation [34]. Briefly, mice were anesthetized, underwent midline laparotomy, and the IVC was ligated with 7–0 monofilament suture immediately below the renal veins, allowing for renal vein outflow. All visible side branches were ligated with 7–0 monofilament suture as well. For visible lumbar branches that were not able to be safely ligated, cautery was used to disrupt venous flow. Mice were then allowed to recover and later sacrificed 24 hours after ligation. Thrombi, which included the vessel wall and thrombus, were excised, and weighed immediately. Thrombi were processed for immunofluorescent (IF) staining and ELISA analysis. Mice were exposed to ambient light only during the procedure (<10 minutes), and subsequently exposed to either ambient, red, or blue light for the entirety of the recovery period.

4.4 |. Transient MCA occlusion stroke model

Transient MCA occlusion (tMCAO) and functional testing were performed as described previously [23]. Briefly, occlusion of the right MCA was achieved by inserting a standardized monofilament (Doccol Corp) via the right internal carotid artery to occlude the origin of the right MCA. The occluding monofilament was left in situ for 60 minutes. Induction of stroke was confirmed by neurological assessment while the MCA was occluded. Anesthesia was induced by inhalation of 5% isoflurane and maintained by inhalation of 2% isoflurane. Buprenorphine was administered one hour before surgery and every 12 hours as needed. Brains of dead mice were visually checked for surgical complications and stained with 2,3,5-triphenyl-tetrazolium chloride when possible, as described below, to confirm ischemic stroke-related mortality.

Twenty-four hours poststroke onset, mice were subjected to the modified Bederson test and the modified neurological severity score to assess neurological and motor function. Analysis of motor outcomes was done on an accelerated rotarod. The speed of the rod increases with time (start: 2.5 rpm, final: 30 rpm), and the amount of time, until the animal stops walking (latency to turn) or falls from the device (latency to fall), is recorded (in seconds).

After functional testing, ischemic stroke brain damage was quantified by staining 2-mm-thick coronal brain sections with 2% 2,3,5-triphenyl-tetrazolium chloride (T8877; Sigma-Aldrich) to distinguish unaffected brain tissue from infarcted tissue. Stained slices were photographed, and infarct areas (white) were measured using Image J software (National Institutes of Health) by an operator blinded for treatment.

4.5 |. Measurement of platelet aggregation

Whole blood impedance aggregometry was performed using a ChronoLog aggregometer (Model 700) according to the manufacturer’s instructions [35]. In addition, 10 μM ADP and 5 μg/mL collagen were used as aggregation agonists. Amplitude (ohm), area under the curve (ohm × min), slope (ohm/min), and lag time (min) were calculated. Area under the platelet aggregation curve was used as a summary measure of the platelet aggregometry curve consistent with prior publications [63,64].

4.6 |. Measurement of platelet activation

Circulating activated platelets were measured from washed platelets by staining with the following antibodies: CD41-APC (MWREG30, ThermoFisher), CD62P (P-selectin)-FITC (Wug.E9, Emfret), JonA GPIIbIIIa-PE (M023–2, Emfret). Stained platelets were acquired via flow cytometry (LSR-Fortessa; Becton Dickinson). In all experiments, 100 μL of washed platelets (10 × 10⁷ per mL) were incubated with either convulxin (50 ng, Santa Cruz; sc-20255) or thrombin (0.05 units, ChronoLog).

4.7 |. Generation of platelet releasate

Platelet releasate was harvested as previously described [30]. Briefly, a washed platelet pellet was isolated from citrate anticoagulated whole blood with serial centrifugation. Platelet counts were adjusted to 500 000/μL and activated with thrombin (0.1 units/mL). After a 30-minute incubation period in Tyrode’s buffer, platelet-secreted factors were collected and stored at −80°C before use.

4.8 |. Neutrophil isolation

Mouse neutrophils were isolated from the bone marrow of femurs and tibias using Ficoll Hypaque density centrifugation as previously described [65,66]. Only isolates with >97% purity and viability were used, as determined by Turk’s staining and Trypan Blue exclusion. Neutrophils were then plated with 1 × 10⁴ cells per well in a 6-plate well culture plate in RPMI medium.

4.9 |. Ex vivo neutrophil activation

Isolated neutrophils were treated with platelet releasate generated from mice^red or mice^white, PMA (250 nM), or with Tyrode’s buffer vehicle controls; each in duplicate. NETosis was quantified after a 4-hour treatment [31]. Treatment groups were standardized to vehicle control in both red light and white light groups. Fold increase from control was compared between groups.

In a separate experiment, neutrophils were isolated from mice^red and mice^white. Cells were plated and stimulated with either PMA or platelet releasate from healthy controls, each in duplicate. NETosis was quantified and reported as above.

4.10 |. NETosis quantification

NETosis was quantified with a sandwich ELISA detecting myeloperoxidase (MPO)-DNA complexes [67] (Hycult Biotech). After treatment with the specified activators, cell supernatant was aspirated and incubated on a 96-well plate from a commercial MPO ELISA kit (Hycult biotech) and then NETs were measured using MPO (Roche) associated with DNA ELISA (Roche). Samples were analyzed in a plate reader at 405 nm. Each sample was plated in triplicate, and values were reported as average of the triplicates.

NETosis was also quantified within the thrombus. First, thrombi were suspended in lysis buffer, dissociated by sonication, and centrifuged at 7000 g for 10 minutes. The supernatant was then concentrated 10-fold with Amicon Ultra Centrifugal Filter Units (10 kDa; Merck Millipore), lysed with lysis buffer, and standardized to a protein level of 100 μg protein per well (ThermoFisher). MPO-DNA capture ELISA was then performed to quantify NETosis.

4.11 |. Thrombus immunofluorescence

Thrombi were fixed in 2% paraformaldehyde for 2 hours and then switched to 30% sucrose 24 hours. The thrombi were then frozen and 6-μm sections were blocked in 20% normal goat serum. The samples were then incubated overnight with a combination of the following primary antibodies: citH3 (5 μg/mL, rabbit IgG; Abcam ab5103); Ly6G-APC (1:1000, BD Pharmogen 560599), CD41 (5 μg/mL, rat IgG; Abcam ab 33661); fibrinogen (2 μg/mL, sheep IgG; Abcam 61352). Sections were then incubated with Alexa 488-conjugated F-actin phalloidin (1:500, Invitrogen) in the presence of the following secondary antibodies depending on the primary antibody pairing: Cy5–conjugated goat anti-rat IgG (1:1000, Jackson Immunoresearch 112–165-167); Cy3-conjugated goat anti-rabbit IgG (1:1000, Jackson Immunoresearch 111–165-003); Cy3-conjugated goat anti-sheep IgG (1:1000, Jackson Immunoresearch 111–165-003) for 1 hour. A Hoechst nuclear stain was applied for 30 seconds and slides were prepared for imaging. Imaging conditions were maintained at identical settings within each antibody-labeling experiment with original gating performed using the negative control. Large area images in X and Y were obtained through the Z plane and are presented as a mean intensity projection of Z using a Nikon A1 confocal microscope. Quantification was performed using NIS Elements Software. NETs were quantified as Hoechst+/Ly6G+/citH3+ and the treatment groups were compared by dividing the NETs by the number of nucleated cells (Hoechst+) on images acquired with a 20× magnification with a 2-fold digital zoom per field. Similarly, neutrophils were quantified as Hoechst+/Ly6G+ cells.

4.12 |. Brain immunofluorescence

Twenty-four hours after stroke, mice were euthanized, and the brains were dissected, snap-frozen in optimal cutting temperature compound and cryosectioned into 10-μm slices. Before immunofluorescence staining, slides were fixed in 4% paraformaldehyde and blocked in 3% donkey serum with 0.5% Tween20. As primary antibodies, we used goat anti-MPO (2 μg/mL, AF3667; R&D Systems), CD41 (2 μg/mL, MWReg30; ThermoFisher) and rabbit anti-human citrullinated Histone H3 (2 μg/mL, ab5103, Abcam). As secondary antibodies, we used donkey anti-rabbit AF488 (2 μg/mL, R37118; ThermoFisher) and donkey anti-goat AF546 (2 μg/mL, A-11056; ThermoFisher), when appropriate. DAPI was used as a nuclear counterstain (300 nM, D1306; Life Technologies). Images were acquired using a high-resolution, confocal reflection microscope (Olympus IX81, FV300; Olympus).

4.13 |. Cellular changes to visible light alterations

Detailed methodology regarding transcriptomic, metabolomic, and single-cell sequencing experiments is included in the Supplementary Methods.

4.14 |. Thrombosis risk in human cataract patients

Detailed methodology regarding the retrospective evaluation of thrombosis risk in human cataract patients is included in the Supplementary Methods. A STROBE checklist is provided in Supplementary Table S1.

4.15 |. Statistical analysis

For comparisons between groups, Student’s t-test or the Mann–Whitney U test as well as analysis of variance or the Kruskal-Wallis test for the analyses of continuous variables, when appropriate. Differences between categorical variables were calculated with the use of chi-squared or Fisher’s exact test, when appropriate. All data are reported and presented as mean ± SD or SE of the mean. Two-sided P values of <.05 were considered statistically significant. For a power of 80% and an alpha level of .05, a minimum of 7 animals in each group were needed to achieve adequate power for the primary outcome.

4.16 |. Study approval

All animal experiments were approved and conducted in accordance with the guidelines set forth by the Animal Research and Care Committee at the University of Pittsburgh (Institutional Animal Care and Use Committees [IACUC] protocol #21110037) and the University of Utah (IACUC protocol #21–09012) and followed the Animal Research: Reporting of In Vivo Experiments guidelines for reporting animal research [33]. Human studies were approved by The University of Pittsburgh Institutional Review Board (STUDY21060010). It was determined that a waiver to obtain informed consent was appropriate for this retrospective study.

DATA AVAILABILITY

All data are available in the main text or the Supplementary Materials.