A Role for Photobiomodulation in the Prevention of Myocardial Ischemic Reperfusion Injury: A Systematic Review and Potential Molecular Mechanisms

Myocardial ischemia reperfusion injury is a negative pathophysiological event that may result in cardiac cell apoptosis and is a result of coronary revascularization and cardiac intervention procedures. The resulting loss of cardiomyocyte cells and the formation of scar tissue, leads to impaired heart function, a major prognostic determinant of long-term cardiac outcomes. Photobiomodulation is a novel cardiac intervention that has displayed therapeutic effects in reducing myocardial ischemia reperfusion related myocardial injury in animal models. A growing body of evidence supporting the use of photobiomodulation in myocardial infarct models has implicated multiple molecular interactions. A systematic review was conducted to identify the strength of the evidence for the therapeutic effect of photobiomodulation and to summarise the current evidence as to its mechanisms. Photobiomodulation in animal models showed consistently positive effects over a range of wavelengths and application parameters, with reductions in total infarct size (up to 76%), decreases in inflammation and scarring, and increases in tissue repair. Multiple molecular pathways were identified, including modulation of inflammatory cytokines, signalling molecules, transcription factors, enzymes and antioxidants. Current evidence regarding the use of photobiomodulation in acute and planned cardiac intervention is at an early stage but is sufficient to inform on clinical trials.

When blood flow is restored, reperfusion causes additional injury, thought to be related to the surge of oxygen return. Yellon and Hausenloy 6 have suggested that reperfusion alone can contribute up to 30% to 40% of total infarct size following coronary artery occlusion. Microvascular obstruction has been suggested to cause damage during coronary events, however its contribution to infarct size, if any, remains unclear 4 . Under conditions of rapid re-oxygenation, the shift in ionic flux results in restoration of cellular pH and the rapid alteration of cellular pH, rather than the return of oxygen, may be the stimulus that activates processes leading to cell death 5 . A speed dependant relationship of pH restoration has been identified, where the intensity at which oxygen returns determines the amount of reactive oxygen species (ROS) released and the opening of the MPTP, allowing accumulated Ca 2+ into the cytoplasm. Disordered intracellular Ca 2+ /ROS balance ultimately leads to dysregulation of the MPTP and rupture of the sarcolemma 5,7,8 . A more severe form of apoptosis is oncosis, where cell death is characterised by cell swelling and karyolysis during MIR injury. Factors that mitigate against oncosis include the presence of melatonin 9 .
Secondary damage as a result of the innate immune response has been suggested as an immediate and delayed process that also contributes to infarct size. Persistent inflammation has been identified as harmful, preventing infarct repair 2 . However, the exact involvement of inflammation has yet to be fully elucidated 4 . Neutrophils, monocytes, and macrophages, which are responsible for remodelling and removal of dead or dying tissue, depend on specific spatiotemporal and quantitative signalling for their activation 10 . Modulation of these signalling processes, especially in the acute vulnerable period following MIR injury, presents another pathway for reducing infarct size.
MIR injury is also associated with potentially serious systemic effects, such as increased morbidity, increased mortality (9%) at one year 11 and neurological impairment, where it is estimated that only 10% of resuscitated patients are neurologically intact when discharged from hospital 7 . MIR injury as a sequela of cardiac surgery, may also lead to a greater rate of post-operative cognitive dysfunction (POCD) than other types of surgery 12 . Thus recovery from MI is the gold standard for effectiveness of treatments that precondition against MIR injury.
While there are many therapies to reduce the effect of ischemia, there has been less success in treating reperfusion injury, although a number of novel potential therapies have been proposed 1,11,13 . These include ischemic preconditioning, where brief episodes of ischemia followed by reperfusion, are introduced before a sustained ischemia. Ischemic preconditioning has been shown to reduce infarct size substantially (30-80%) and can last for 2 to 3 hours after the preconditioning event, as well as having a second window of protection that occurs 24 hours after preconditioning and can last for about 48 hours 14 . Ischemic post-conditioning, where the rate of reperfusion is slowed by short episodes of myocardial ischemia (using, for example, an angioplasty balloon) may also be effective in reducing MIR injury 14 . Remote ischemic conditioning, where brief non-lethal episodes of ischemia and reperfusion are applied to an area remote from the heart, such as an arm or leg, has also been shown to have some effect on MI size 6 . While ischemic conditioning has been demonstrated in animal models, there has been less success in the translation of this therapy to clinical trials 15,16 , although balloon angioplasty post-conditioning and hypothermia have shown some positive effects 17 . Ischemic preconditioning is also limited to conditions where the potential for MIR insult can be predicted, such as coronary artery bypass grafts (CABG) surgery, and ischemic post-conditioning is limited by the very short window (5 to 10 minutes) during which the treatment can be effective 14 .
In addition to ischemic conditioning, there are a number of experimental pharmaceutical products with the potential to reduce MI size. Alpha-melanocyte-stimulating hormone (α -MSH) has shown some promise 18 , the use of volatile anaesthetics for cardioprotection during open heart surgery can produce modest effects 19 and the use of cyclosporin A 20 and exenatide 21 have been shown to produce reductions in MI size. The disadvantage of drug and anaesthetic intervention is the potential to introduce serious side effects, complicated by the proteostasis of the patient, as well as disease (e.g., diabetes) and interactions with other drugs 22 . These side effects can potentially be as severe as Parkinson's disease and Huntington's disease-like symptoms 7 .
In short, MIR injury involves a complex redox stress response, for which there is so far no effective treatment, although a number of novel treatments have been proposed. A goal for such treatment would be a mechanism to switch between the deleterious redox stress reactions, towards protective redox conditions. The complexity, however, of the switching mechanism has hindered an effective therapeutic regime 23 . A potential target for treatment intervention is the modification of the mitochondrial response to oxidative stress.
Photobiomodulation (PBM) is the low power (1-500 mW) non-thermal delivery of photons in the visible or near infrared spectrum (405-1000 nm) that elicits a beneficial biological response in cells and tissues 24 . PBM can include light emitting diodes (LED) and low-level laser therapy (LLLT). Phototherapy has a long history of application in medicine. The earliest scientific report was the use of red light in the treatment of smallpox scars by Neil Finsen in 1903 25 and his use of ultraviolet light for the treatment of lupus vulgaris, which resulted in a Nobel Prize in 1903. Laser light was used in radiation ulcer attenuation after the Chernobyl nuclear accident 26 and light has been used as a treatment for bilirubin dysfunction in neonates for many decades 27 . Interestingly, when light is used for premature neonate jaundice treatment, there is a concomitant effect in patent ductus arteriosus, where the systemic effect of the light application can cause cardiac vasodilation effects and prevents the closure of the patent ductus, which can be avoided by the use of light impermeable chest shielding 28 . Recently 670 nm light has been proposed as a treatment for oxygen-induced retinal disease in neonates 29 . PBM has been increasingly used for treatment of ulcers 30 , wounds 31 , neuro-inflammation 32,33 , pain 34 , lympoedema 35 , macular degeneration 36 and tendon healing 37 . The use of LLLT in various clinical applications has been reviewed by Chow et al. 34 in the area of chronic neck pain, by Khan and Arany 38 for wound healing, by Geneva et al. 39 for retinal disease, by Agrawal et al. 40 as a preconditioning treatment for various diseases and by Carlos et al. 41 for cardiac remodeling after myocardial infarction. Animal models have also suggested a role for the PBM treatment of neurodegenerative diseases 42 , such as Alzheimer's disease 43 , Parkinson's disease 44 , multiple sclerosis 45 and as a preconditioning treatment against post-operative cognitive dysfunction 46 . Natural light sources have also been shown to be important in recovery Scientific RepoRts | 7:42386 | DOI: 10.1038/srep42386 from spinal surgery 47 and management of cardiac risk patients 48 . An emerging body of experimental evidence as well as some clinical trials support the application of PBM in conjunction with routine cardiac interventions, which warrants a systematic review of PBM application and MIR injury.
The aim of this contribution is to report the results of a systematic review into the experimental evidence in tissue studies, animal studies and clinical trials for the use of PBM in the treatment, intervention and management of myocardial reperfusion injury, and to summarize the underlying mechanisms and metabolic signalling pathways found to underpin this effect. The results of the review can be used to inform scientists and clinicians on the potential scope of further experimental studies and clinical trials, the creation of best practice guidelines and the use of PBM in other areas of cardiovascular intervention.

Results and Discussion
Quality and Limitations of Studies. The review of articles was conducted according to PRISMA guidelines (http://www.prisma-statement.org) (Fig. 1). Since most of the articles reviewed pertain to animal or tissue experiments, with only one clinical trial, the extent to which results can be extended to human effects is necessarily limited. In the single clinical trial, assessment using the JADAD method 49 resulted in a score of 4, indicating high quality. The use of the SYRCLE risk of bias tool 50 to assess quality of animal studies indicated a high or unknown risk of bias for most studies in the majority of categories (Table 1 and Fig. 2). Only two categories were assessed as having a low risk of bias for the majority of studies. While reporting bias (item 9, Table 1) was assessed as low, the possibility of publication bias, in that results have been selectively reported, cannot be discounted, especially as no adverse effects or negative findings of the effect of PBM were described. It should also be kept in mind that it is difficult to translate results of animal studies to humans, due in part to the variations in physiology and anatomy between the species (for example, rats have fewer collateral coronary arteries than dogs and humans).

Characteristics of Studies.
Experimental designs in the studies reviewed ranged from in vitro cardiomyocyte cultures to a single human clinical trial. Of the 23 studies included, 17 (74%) were conducted using animal models (13 on rats, 2 on mice, 1 on rabbits and 1 on dogs and rats), 2 were conducted on isolated rat hearts, 3 on tissues or cell cultures and 1 was a clinical trial ( Table 2). The application site for PBM in animal studies varied with experimental design, with 6 studies using trans-thoracic application of PBM (with skin shaved or removed), 12 irradiating the myocardium directly and 3 studies irradiating sites distal to the heart (leg muscle or tibia). In the clinical trial there was a transthoracic administration of PBM. Studies on isolated tissue and cell samples involved irradiation without direct contact. PBM was administered as preconditioning treatment, as post-conditioning treatment, as treatment during surgery or reperfusion, or as a combination of treatments. Total intervention dose ranged from 0.6 J to 36 J, with a mean of 6.26 J. All trials used wavelengths between 630 nm and 830 nm, with the human trial using infrared light (810 nm) and the majority of animal trials used red light (at around 660 nm). All animal studies reported physical or histological outcomes after PBM intervention and 18 reported molecular changes.
Infarct Size Reduction, Histological Profile and Long-Term Effects. The most significant finding from this review was the positive effect of PBM on modulating infarct size and the improvement of cardiac remodelling [51][52][53][54][55][56][57][58][59][60][61][62][63] . These results were achieved using a range of experimental designs, wavelengths and dosages. Three studies 59-61 reported a reduction in total infarct size of greater than 60% with one 60 reporting MI size reduction of 76% when irradiation was applied to the tibia of the rat (remote preconditioning), compared to a 31% reduction when applied locally to myocardium. Yaakobi et al. 58 reported consistently reduced infarct size compared to controls throughout a 45-day follow-up period. The results agreed with other studies that have shown PBM can significantly decrease the narrowing of coronary arteries, decrease rates of restenosis after stenting

Attrition bias
Reporting bias
Yang et al. 51 ?    Hentschke et al. 71 Rats  procedures [64][65][66] and is a therapeutic option for severe medically refractory angina 67 and healing of sternotomy incisions 68 . Histological analysis in a number of studies revealed improvements in cardiomyocyte arrangement 51,53,58,69 , including reductions in dense collagen, reduced ventricular wall thinning and increased mesenchymal and cardiac stem cells. Gavish et al. 70 also reported that PBM modulated collagen synthesis in a porcine aorta model, along with changes to the levels of matrix metalloproteinase-2 (MMP-2), a regulator of collagen synthesis.

PBM Dose, Timing and Treatment Regime.
All studies showed the safety of PBM in animal and human subjects, including direct myocardial irradiation and the clinical trial did not report any side effects or adverse reactions in patient populations as a result of PBM treatment. Both LLLT and LED were shown to be effective as a source of PBM and all wavelengths tested showed positive results, including red and infrared wavelengths ( Table 2). As yet there have been no studies on super-pulsed laser infrared (904-980 nm) wavelengths. The reported effects of PBM on MIR injury were dose dependant and mirrored PBM effects that have been reported in the literature for treatment of conditions such as chronic pain, neurodegenerative disease, lymphoedema and macular degeneration [34][35][36]42 .
There appears to be wide windows for the dose and timing of PBM treatment. In a number of studies, PBM was shown to be effective both as a preconditioning treatment and if applied during reperfusion or in the immediate post (4 hour) surgical window. There was also good evidence 60,71 that PBM administered to a distal area (tibia or leg muscle) produced a positive effect on infarct size. This is consistent with the abscopal effect seen in application of PBM 72 with, for example, neuroprotection where targeting remote tissues with PBM can decrease the symptoms of Parkinson's disease 73 . This abscopal or systemic effect has been attributed to the downregulation of pro-inflammatory cytokines and upregulation of anti-inflammatory cytokines 73,74 as well as to the proliferation of mesenchymal stem cells 59,60,75 , both of which PBM is known to influence 76 . It may also possibly be due to a systemic mitokine response (see below).
One study found that a higher PBM dose was more effective in the initial phase of MIR to reduce injury 58 , 804 nm LLLT at 12 mW/cm 2 having more effective cardioprotection than 6 mW/cm 2 . This is in contrast to the most effective dose given 1.5 hour after MIR being a combination of 12 mW/cm 2 and 5 mW/cm 2 , which was found to be superior to two doses of 12 mW/cm 2 , leading to the conclusion that LLLT may be less beneficial when given too frequently. The power density required to achieve a reduction in MI size was similar for rats, dogs and humans. For infrared 803-810 nm the dose window indicated is 5 to 15 mW/cm 2 (with 25 mW/cm 2 less effective) 53,77 , which equates to 1 to 6 J/cm 2 for rats 58 and 6 J/cm 2 for the human trial 78 . This is consistent with other studies of treatment with PBM that have demonstrated a biphasic dose response, where a therapeutic effect is achieved with an optimal dose within a wide dose window, outside of which there is no effect 79 . Oron et al. 53 stressed that the complex sequential physiological processes after MIR injury required different doses of PBM at different stages of reperfusion. PBM administered after the reperfusion event (post-conditioning) was found to be effective in a number of studies, even after a considerable delay. In fact, when a delayed post-conditioning treatment was omitted from the treatment regime, the effect on MI size was reduced 58 . Evidence for treatment protocols from a range of animal studies and the clinical trial would seem to suggest that a combination of preconditioning, immediate and post-application (with red or infrared wavelengths of LED or Laser) produce the most positive effects. All of these dose and timing factors would be important for administration of PBM in any planned clinical trial.
The achievement of a therapeutic dose in small experimental animals is not in question using near infrared wavelengths 80 , however there is the possibility that transthoracic PBM may not reach a therapeutic dose in cardiac tissue in humans. While the transmission of light to the heart does not appear in the literature per se, it can be inferred from other studies. For example, the use of PBM to treat traumatic brain injury has included studies that have demonstrated that transcranial LLLT can reach therapeutic doses 40 to 50 cm deep with infrared irradiation (800-830 nm), using head simulations 81 , formalin fixed 82 and unfixed human cadaver heads 83 . The average parasternal, skin-to-heart distance has been reported as 32.1 ± 7.9 mm 84 . In addition, in the clinical trial reviewed here, the transthoracic application of PBM resulted in changes to cardiac and inflammation markers 78 .
Mitochondrial Effects. The overall trend from the studies was a positive effect on mitochondrial respiration pathways, including an increase in the availability of ATP and nitric oxide (NO). Zhu et al. 85 noted 15% higher ATP levels in isolated hearts during induced cardioplegia when treated with PBM and Oron et al. 53 showed a 21% reduction in mitochondrial damage in an animal model, which, it was suggested, may account for sustained ATP production.
One of the major photon receptors is generally believed to be the cytochrome C oxidase (COX) enzyme in the electron transport chain and stimulation of this enzyme sets in motion multiple cascading signal transduction pathways 86 . Heart tissue, however, responds differently from other tissues to mitochondrial COX signalling pathways, due to the tissue specific expression of signalling molecules 7 , which might make heart tissue more susceptible to ischemic and reperfusion injury than other tissues. The response of the mitochondria to reperfusion is a critical factor in the health or otherwise of myocardial cells and tissues. Protection of the mitochondria, particularly the MPTP, has been called "the Holy Grail" of cardioprotection 13 and phosphorylation events that regulate COX are ideal targets for therapeutic interventions 7 . The effect of PBM on COX is generally accepted 2 and is believed to be a major reason for the effect PBM on mitochondrial health (Fig. 3), due to the increased production of ATP and the activation of ATP dependent ion pumps 86,87 . Recently however, it has been demonstrated that near IR has no effect on isolated COX protein 88 , which may suggest a more complex interaction between light and the mitochondria, possibly an indirect effect. This was also implied in a study of retinopathy in diabetic mice 89 and in diabetic rats 90 , where there was no evidence found for a PBM effect on mitochondria. Never-the-less, an effect of PBM on the mitochondria was suggested by a number of the studies reviewed here (Fig. 3), with a positive effect of PBM on both mitochondrial respiration and mitochondrial retrograde signalling and major impacts on the regulation of Ca 2+ ion channel flux, NO production through COX and increased in ATP levels. These factors ensure the maintenance of mitochondrial retrograde signalling, which has been identified as a regulator of many cellular activities in both normal and pathological states 91 , including mitochondrial retrograde signalling in MIR 92 and the ischemic preconditioning effect of cardioprotection 40 . Mitochondrial homeostasis is important in the cell stress response 93 , which can be regulated by the level of cellular melatonin in the MIR injury response, where elevated melatonin is neuroprotective against myocardial cell oncosis 9 . Modulation of the mitochondrial retrograde signalling response would also have an effect on systemic metabolism through the mitochondrial unfolded protein response (UPRmt) 94,95 . Since proteomic stress can induce changes in redox stress across tissues, this may be important in the limitation of cardiomyocyte death and scar formation 96 as well as in organism wide neuroprotection. This may involve a mitokine response 95 , which could also have implications in preconditioning against ischemic reperfusion injury. The modulation of the UPRmt would restore homeostasis through mitochondrial hormesis (mitohormesis) (see Fig. 3).

Non-Mitochondrial Signalling by PBM.
Non-mitochondrial effects of PBM involve signalling pathways that originate from photon absorption and protein conformational modulation at the cell membrane (Fig. 3). Photons can be absorbed by ion channels (including TRPV1, K + and Ca 2+ channels), membrane receptors including tyrosine kinases and a variety of photoactive molecules described as opsins 38 . Downstream signal transduction is known to affect inflammatory cytokines 97 , heat shock proteins 98 , endothelial and axonal cytoskeleton morphology 99,100 , antioxidant IL10 101 , growth factors vascular epithelial growth factor (VEGF) and MMP-2 102 , inducible nitrite oxide synthase (iNOS) 103 , and superoxide dismutase (SOD) 104 as well as modifying ROS and NO. These have a variety of effects, such as improving tissue responses, including cardiac tissue, endothelial tissue and arterial lumen diameter.
An example of a signalling pathway modified by PBM in MIR is the suppression of the ASK1/p38/NF-κ B signalling pathway 71 , which is important in cardiomyocyte necrosis following myocardial infarction 2 and which has also been shown, in an in vitro model of wound healing, to enhance tissue regeneration and angiogenesis 105 . A pathway that was not monitored in any study reviewed here is the Akt/GSK-3β pathway 106 , involved in neuronal survival in traumatic brain injury. This pathway has been demonstrated to be modulated by PBM and to be neuroprotective against apoptosis induced by amyloid β peptide 107 . This pathway might be predicted to also be involved in protection against MIR injury by PBM.
PBM was shown in a number of the studies reviewed here to influence a variety of non-mitochondrial signal transduction pathways, which are linked to a decreased cell adverse stress response, improved tissue responses (decreased MI size, reduced restenosis, etc.) and a restoration of homeostasis (see Fig. 3). While these modulations of cell signalling can also be achieved using drug and anaesthetic strategies 5 , these treatments have the potential for serious side effects, as previously noted.
Redox State and Antioxidants. In addition to the respiratory chain, NO is also produced in a number of reactions outside of the mitochondria. A number of studies reviewed here demonstrated an increase in NO following PBM treatment, as well as increased levels of iNOS and endothelial nitric oxide synthase (eNOS). Tuby et al. 77 reported a significant elevation in iNOS, which increased with increasing PBM dose. The increase in iNOS levels was recorded after 2.5 hours and peaked at 2 days, which indicated an ongoing role for NO after the acute insult. Interestingly, enhanced levels of iNOS were also observed in non-infarcted but irradiated myocardium, which raises questions concerning the modulating effect of PBM on this enzyme in non-traumatic pathological states. Manchini et al. 63 however reported reductions in iNOS in the intervention group, 3 days post injury, although plasma nitrite and nitrate concentrations (NOx) were markedly higher, indicating a transient effect of PBM on NO levels.
In one study 62 , NO was shown to be released from nitrosyl heme proteins in the irradiation group and two other studies 54,55 reported an increase in NO from MbNO and HbNO, independent of nitric oxide synthase.
NO is important in the PBM myocardial protection during ischemic reperfusion injury, both in the cell stress response and also in regulating Ca 2+ homeostasis 108 and thus limiting the effect of mitochondrial Ca 2+ overload from MPTP. NO signalling may also lead to mitochondrial S-nitrosation, which slows reactivation of mitochondrial metabolism at the onset of re-oxygenation 7 . This slowing of reactivation in post-conditioning may be particularly important to mitigate against further MIR injury. Non-NOS sources of NO have been argued to be a major mechanism behind PBM protection against MIR injury 55 .
While two studies reported that levels of SOD increased within 1 hour post irradiation 110,111 , others 109,115 reported a reduction in SOD activity in PBM treated groups, when irradiation occurred 4 and 3 weeks after rats were exposed to ischemic insult. These results suggest that PBM induces an acute transient increase in SOD activity immediately after MIR injury. This was confirmed in an experiment where SOD activity was shown to double when isolated rat hearts were irradiated immediately after MIR injury 112 . Interestingly Malinovaskaya et al. found that while non-coherent light resulted in an increase in SOD, laser brought about a decrease in SOD activity 111 .
Yaakobi et al. 58 found a 2.2-fold increase of heat shock protein inducible factor (Hsp70i), 5 hours post irradiation using 810 nm. Zhang et al. 110 reported increased levels of gastrin-releasing peptide (GRP) 78 (a chaperone to Hsp70) at 1 hour, 1 day and 1 week, indicating a sustained up-regulation of this protein in the later stages of the tissue recovery cycle. Heat shock proteins have a role in mitochondrial membrane homeostasis and have an important role in protein folding, unfolding and the cell stress response 116 . Hsp70 is also an important factor in the amelioration of acute lung injury induced by gut ischemia in the rat model, where PBM up-regulates Scientific RepoRts | 7:42386 | DOI: 10.1038/srep42386 peroxisome proliferator-activated receptor-γ (PPARγ ), which in turn results in an increase in Hsp70 production and a decrease in inflammation and lung injury 85 .
Inflammatory Cytokines. Inflammatory cytokines regulate the cellular stress response and the inflammatory cascade associated with the ischemic event that causes injury 23 . The balance between pro and anti-inflammatory cytokines is important in determining prognosis after MIR injury, especially the balance in redox signalling molecules, which appears to act as a switch between protective oxidative signalling and damaging oxidative signalling 23 . While this switching mechanism is not yet fully understood, it appears that PBM is able to affect the switching in a positive way, resulting in cardioprotection. This was consistent across all studies and parallels the PBM effect on cytokines in non-cardiac studies, including macular degeneration 36 , wound healing 38 , muscle pre-conditioning 40 , gut ischemia and reperfusion 98 and neuroprotection against Alzheimer's and Parkinson's diseases 42 and, potentially, POCD 46 .
Six studies reported changes in levels of the inflammatory markers IL-1α , 1β , 2, 4, 6 and 8 after PBM 57,63,71,114 . In a human clinical trial with patients undergoing angioplasty and cardiac stenting (excluded from this review), Derkacz et al. 114 also found that IL-1β , and IL6 were reduced when irradiated with 808 nm with a total treatment dose of 9 J/cm 2 . In contrast to these findings, Manchini et al. 63 found strong increases in IL6 in both control and PBM groups (660 nm and 22.5 J/cm 2 ) three days post intervention. In a comparable study, Hentschke et al. 71 showed that a four-week delay in PBM application resulted in reduced IL6 levels in the group treated with 660 nm 21 J/cm 2 but significantly higher IL6 levels in the group treated with 3 J/cm 2 . Together these results suggest that there was a dose window and a dose dependant relationship between PBM and levels of IL6.
In the clinical trial reviewed here 78 , repeated doses of 808 nm PBM resulted in reduced white blood cell, lymphocyte and neutrophil activity 5 days post cardiac artery bypass graft (CABG) surgery. In a rat reperfusion injury model 98 it was reported that an acute increase in myeloperoxidase (MPO) occurred after irradiation with 660 nm laser, indicating an increase in neutrophil granulocyte activity. Manchini et al. 63 found that kinin B2 receptor mRNA expression and Mas receptor protein expression was increased after MI and PBM, whereas PBM significantly decreased the kinin B1 and significantly reduced angiotensin-converting enzyme (ACE) mRNA expression, all of which affect vasodilation.

Growth Factor Modulation.
Three studies showed significant increases in expression of VEGF after PBM.
Tuby et al. 77 demonstrated significant increases in expression as early as 2.5 hours after treatment, which continued to 24 and 48 hours, reverting back to pre-treatment levels by 72 hours. There was also a dose dependent relationship, with power densities 5, 12, and 17 mW/cm 2 increasing expression rates by 2, 2.3, and 1.3 fold respectively. Zhang et al. 110 also found significant increases in VEGF 1 hour and 1 day post intervention, with no significant difference in levels after 1 week. This is consistent with other studies investigating the mechanisms of PBM action 102 . The increase in VEGF promotes the proliferation of endothelial cells and angiogenesis, important in the cell stress response under hypoxic conditions 117 or under high concentrations of ROS 94 . VEGF is also important in the homeostatic control of the unfolded protein response in the endoplasmic reticulum and in the mitochondria 94 . In addition, PBM has an indirect action on mesenchymal stem cell proliferation 60,69 and the interaction between mesenchymal stem cells and VEGF mRNA modulates cell adhesion and proliferation in a nutritionally deprived model 118 .

Cardiac Markers.
There was a reduction in cardiac enzyme marker troponin in the rat model 52 and creatine phosphokinase (CPK) in the rat 52 and the human study 78 . These are clinically important markers for cardiac damage, which may be important in the design of PBM trials for the evaluation of the success of dose and timing of PBM protocols.
Cytoskeleton Modulation by PBM. In a study on nutrient stressed rabbit aortic endothelial cells 99 , it was shown that irradiation with laser at 685 nm at 8 J/cm 2 for 7 days caused reorganisation of actin filament stress fibres and filament proliferation, such that the cell regained its structure to a pre-stressed state. This may be a similar mechanism to the reorganization of the cytoskeleton seen after PBM application to dorsal root ganglion nerves in pain blockade 46,100 ; a process strikingly similar to the cytoskeleton remodulation in the neuroprotective mechanism of rapid ischemic tolerance defence against NMDA excitotoxicity. These mechanisms have been reviewed by Liebert et al. 46 . This rapid ischemic tolerance mechanism may be relevant to the PBM effect in the animal model for infarct size reduction in MIR. It is often observed that rapid ischemic tolerance induce by PBM (such as in pain blockade) requires a higher dose in the initial stages, which is also the case in cardioprotection 77 , suggesting a similar mechanism.

Conclusion
Cardiac protection, both in clinical use and preconditioning applications, will become progressively more important with increasing numbers of cardiac procedures in the future and the further burden of cardiac disease, evident in the incidence of morbidity and potential mortality and the escalating cost of treatment. The evidence outlined in this review from a range of in vitro and in vivo animal studies, as well as one clinical trial, suggests that PBM may have a role as a cardioprotective agent against MIR injury and could protect against the initial cardiac ischemic event and the ongoing damage caused by reperfusion. PBM has been shown to affect a variety of signal transduction pathways that are critical to switching from the deleterious redox stress reactions that occur as a result of reperfusion, towards the more protective redox conditions that can limit injury and promote repair. This could ultimately lead to improved tissue responses, including reduced infarct size and lower rates of restenosis. PBM is non-invasive, simple to administer, inexpensive and has no known side effects, unlike other interventions that appear to have limited evidence of efficacy and potentially deleterious side effects. PBM could therefore be considered as a potential alternative to drug and anaesthetic pre-treatments for MIR injury. It also appears to be an effective post-conditioning therapeutic protocol, thus extending the available time for intervention to treat MIR injury.
The number of well-designed clinical trials is limited (one in this review), but the available evidence from animal and tissue studies suggest that further clinical trials are warranted. The timing and dosage of PBM appears to be complex, with different doses required at different stages of reperfusion, in order to achieve optimal treatment outcomes; something that must be considered when planning clinical trials. The cardiac markers of troponin, CPK and other novel biochemical indices such as osteopontin 119 could be used as markers of MIR injury in these future trials. Evidence would suggest that a combination of preconditioning, immediate and post application treatment would be appropriate, with a potential role for remote application of PBM to produce a cardioprotective preconditioning abscopal effect.

Methods
Search Strategy. This review was conducted according to PRISMA guidelines and the overall search strategy is shown in Fig. 1. The search of published articles was conducted on the 28 th of August 2015. PubMed, Ovid (OvidMedline), Scopus, and Web of Science, journal databases were searched with restrictions set to 1995-current, and English publications only (see Supplementary Table 1). Keywords used for the search included "laser, low level" and was combined with "myocardial". Initial searches were expanded by adding derivatives: "ischemia"; "infarct"; "rupture"; "arrest"; and "failure". Preliminary search results indicated that certain keywords such as "cardiovascular" and "disease" returned an excessive number of results when searching databases. These keywords were not used in order to limit the number of irrelevant papers captured during the initial search, which may have meant that studies that included these keywords were overlooked. Duplicate articles from the database search results were removed. Titles and abstracts of each paper were screened, and irrelevant articles removed. The reference lists of relevant papers were searched for additional articles, which were then added to the list of articles.
Eligibility Criteria. Once all full-text copies were obtained, each article was submitted to the eligibility criteria set out by authors (Supplementary Table 2). Articles investigating therapeutic, procedural or methodological applications of PBM were accepted, including application of PBM as an adjunct to surgery, or with standard medical management in a clinical population. The application of the irradiation could be to any part of the body. Due to the small number of clinical trials available, research conducted using animal models, isolated tissues and cell cultures were also accepted. Experimental design had to be representative of either ischemic and/or reperfusion injury. The primary outcome of the review was changes to mortality, cardiac tissues or cells or a change to molecular markers of cardiac function. Secondary outcomes were changes to other molecular markers, such as signalling molecules, redox markers or cytokines. All studies had to report a minimum of wavelength, power output and dose intervention parameters, or the missing parameter had to be calculable using alternate parameters, such as fluency and power density. Studies using combination therapies were omitted to ensure clarity when collating treatment effects. Populations with diseases or co-morbidities, with the exception of those directly involving MIR pathophysiological processes, were excluded to ensure conformity of examined population. Likewise, investigations involving organisms or cell structures with atypical genetics, or those containing disease-causing microorganisms were excluded to avoid false negative results.
Language and Time Restrictions. Language was restricted to English. Preliminary searches returned 21 articles not available in English. Titles of these papers were screened and a further 12 papers identified as potentially relevant. A sample of these were selected and translated into English. Upon review by AL, it was deemed that these articles were not suitable for inclusion. Examination of the publication trend indicated an increase in article number after 2004, with 50% of included articles published since 2010. Articles were therefore limited to those published after 1995. While the decision to only review articles after 1995 limited the number of articles captured, it is highly likely that key insights and findings from before 1995 would have been re-examined in later articles.
Qualitative assessment. Twelve articles did not meet the inclusion criteria. These articles are shown, along with reasons for exclusion, in Supplementary Table 3. The remaining studies were then divided between AK, AL and NG for data extraction. The experimental design, including population studied, number, outcome measures, treatment protocol, and laser dose parameters were recorded. All recorded data was collated into results tables and synthesized (BB and AK) to identify analogous results. Laser dose parameters were standardized by conversion into total dose administered, using the formula from Bjordal et al. 120 . Qualitative analysis of the clinical trial was conducted using the method of Jadad & McQuay 49 . Risk of bias for animal studies was assessed using the SYRCLE tool of Hooijmans et al. 50 . Interpretation of the articles in this review is necessarily descriptive, due to the limited number but wide range of studies that have investigated the effect of PBM in MIR injury.