Biphasic Dose Response in Low Level Light Therapy

1.2. Evidence for effectiveness of LLLT

Since 1967 over 100 phase III, randomized, double-blind, placebo-controlled, clinical trials (RCTs) have been published and supported by over 1,000 laboratory studies investigating the primary mechanisms and the cascade of secondary effects that contribute to a range of local tissue and systemic effects.

RCTs with positive outcomes have been published on pathologies as diverse as osteoarthritis (Bertolucci and Grey 1995; Ozdemir et al. 2001; Stelian et al. 1992), tendonopathies (Bjordal et al. 2006b; Stergioulas et al. 2008; Vasseljen et al. 1992), wounds (Caetano et al. 2009; Gupta et al. 1998; Ozcelik et al. 2008; Schubert et al. 2007), back pain (Basford et al. 1999), neck pain (Chow et al. 2006; Gur et al. 2004), muscle fatigue (Leal Junior et al. 2008a; Leal Junior et al. 2008b), peripheral nerve injuries (Rochkind et al. 2007) and strokes (Lampl et al. 2007; Zivin et al. 2009); nevertheless results have not always been positive. This failure in certain circumstances can be attributed to several factors including dosimetry (inadequate or too much energy delivered, inadequate or too much irradiance, inappropriate pulse structure, irradiation of insufficient area of the pathology), inappropriate anatomical treatment location and concurrent patient medication (such as steroidal and non-steroidal anti-inflammatories which can inhibit healing) (Aimbire et al. 2006; Goncalves et al. 2007).

2.2. Action Spectrum and Tissue Optics

One important consideration should involve the optical properties of tissue. There is a so-called “optical window” in tissue, where the effective tissue penetration of light is maximized. This optical window runs approximately from 650 nm to 1200 nm. (Figure 2). The absorption and scattering of light in tissue are both much higher in the blue region of the spectrum than the red, because the principle tissue chromophores (hemoglobin and melanin) have high absorption bands at shorter wavelengths, tissue scattering of light is higher at shorter wavelengths, and furthermore water strongly absorbs infrared light at wavelengths greater than 1100-nm. Therefore the use of LLLT in animals and patients almost exclusively involves red and near-infrared light (600–1100-nm) (Karu and Afanas’eva 1995).

Open in a separate window

FIGURE 2

Absorption spectra of the main chromophores in living tissue on a log scale showing the optical window where visible and NIR light can penetrate deepest into tissue.

Phototherapy is characterized by its ability to induce photobiological processes in cells. Exact action spectra are needed for determination of photoacceptors as well as for further investigations into cellular mechanisms of phototherapy. The action spectrum shows which specific wavelength of light is most effectively used in a specific chemical reaction (Karu and Kolyakov 2005). The fact that defined action spectra can be constructed for various cellular responses confirms the first law of photobiology described above (light absorption by specific molecular chromophores).

2.3. Mitochondrial Respiration and ATP

Current research about the mechanism of LLLT effects inevitably involves mitochondria. Mitochondria play an important role in energy generation and metabolism. Mitochondria are sometimes described as “cellular power plants”, because they convert food molecules into energy in the form of ATP via the process of oxidative phosphorylation (see Figure 3 for an illustartion of the mitochondrial respiratory chain).

Open in a separate window

FIGURE 3

Mitochondrial respiratory chain consisting of contains five complexes of integral membrane proteins: NADH dehydrogenase (Complex I), succinate dehydrogenase (Complex II), cytochrome c reductase (Complex III), cytochrome c oxidase (Complex IV), and ATP synthase (ComplexV).

The mechanism of LLLT at the cellular level has been attributed to the absorption of monochromatic visible and NIR radiation by components of the cellular respiratory chain (Karu 1989). Several pieces of evidence suggest that mitochondria are responsible for the cellular response to red visible and NIR light. The effects of HeNe laser and other illumination on mitochondria isolated from rat liver, have included increased proton electrochemical potential, more ATP synthesis (Passarella et al. 1984), increased RNA and protein synthesis (Greco et al. 1989) and increases in oxygen consumption, membrane potential, and enhanced synthesis of NADH and ATP.

2.4. Cytochrome c oxidase and nitric oxide release

Absorption spectra obtained for cytochrome c oxidase (Cox) in different oxidation states were recorded and found to be very similar to the action spectra for biological responses to light (Karu and Kolyakov 2005). Therefore it was proposed that Cox is the primary photoacceptor for the red-NIR range in mammalian cells (Karu and Kolyakov 2005).

Nitric oxide produced in the mitochondria can inhibit respiration by binding to Cox and competitively displacing oxygen, especially in stressed or hypoxic cells (Brown 2001). Increased nitric oxide (NO) concentrations can sometimes be measured in cell culture or in animals after LLLT due to its photo release from the mitochondria and Cox. It has been proposed that LLLT might work by photodissociating NO from Cox, thereby reversing the mitochondrial inhibition of respiration due to excessive NO binding (Lane 2006). Figure 4 illustrates the photodissociation of NO from its binding sites on the heme iron and copper centers where it cometively inhibits oxygen binding and reduces necessary enzymic activity, thus allowing an immediate influx of oxygen and resumption of respiration and generation of reactive oxygen species.

Open in a separate window

FIGURE 4

When NO is released from its binding to heme iron and copper centers in cytochrome c oxidase by the action of light, oxygen is allowed to rebind to these sites and respiration is restored to its former level leading to increased ATP synthesis.

2.5. NO signaling

In addition to NO being photodissociated from Cox as described, it may also be photo-released from other intracellular stores such as nitrosylated hemoglobin and nitrosylated myoglobin (Shiva and Gladwin 2009). Light mediated vasodilation was first described in 1968 by R F Furchgott, in his nitric oxide research that lead to his receipt of a Nobel Prize thirty years later in 1998 (Mitka 1998). Later studies conducted by other researchers confirmed and extended Furchgott’s early work and demonstrated the ability of light to influence the localized production or release of NO and stimulate vasodilation through the effect NO on cyclic guanine monophosphate (cGMP). This finding suggested that properly designed illumination devices may be effective, noninvasive therapeutic agents for patients who would benefit from increased localized NO availability

2.6. Reactive oxygen species and gene transcription

Reactive oxygen species (ROS) and reactive nitrogen species (RNS) are involved in the signaling pathways from mitochondria to nuclei. Reactive oxygen species (ROS) are very small molecules that include oxygen ions such as superoxide, free radicals such as hydroxyl radical, and hydrogen peroxide, and organic peroxides. They are highly with biological molecules such as proteins, nucleic acids and unsaturated lipids. ROS form as a natural by-product of the normal metabolism of oxygen and have important roles in cell signaling (Storz 2007), regulating nucleic acid synthesis, protein synthesis, enzyme activation and cell cycle progression (Brondon et al. 2005). LLLT was reported to produce a shift in overall cell redox potential in the direction of greater oxidation (Karu 1999) and increased ROS generation and cell redox activity have been demonstrated (Alexandratou et al. 2002; Chen et al. 2009b; Grossman et al. 1998; Lavi et al. 2003; Lubart et al. 2005; Pal et al. 2007; Zhang et al. 2008). These cytosolic responses may in turn induce transcriptional changes. Several transcription factors are regulated by changes in cellular redox state. But the most important one is nuclear factor B (NF-B). Figure 5 illustrates the effect of redox-sensitive transcription factor NF-κB activated after LLLT and is instrumental in causing transcription of protective and stimulatory gene products.

Open in a separate window

FIGURE 5

Reactive oxygen species (ROS) formed as a result of LLLT effects in mitochondria may activate the redox-sensitive transcription factor NF-κB (relA-p50) via protein kinase D (PKD).

2.7. Downstream cellular response

Although the underlying mechanism of LLLT are still not completely understood, in vitro studies, animal experiments and clinical studies have all tended to indicate that LLLT delivered at low doses may produce a better result when compared to the same light delivered at high doses. LLLT can prevent cell apoptosis and improve cell proliferation, migration and adhesion at low levels of red/NIR light illumination (see Figure 6).

Open in a separate window

FIGURE 6

The downstream cellular effects of LLLT signaling include increases in cell proliferation, migration and adhesion molecules. Cell survival is increased and cell death reduced by expression of proteins that inhibit apoptosis.

LLLT at low doses has been shown to enhance cell proliferation in vitro in several types of cells: fibroblasts (Lubart et al. 1992; Yu et al. 1994), keratinocytes (Grossman et al. 1998), endothelial cells (Moore et al. 2005), and lymphocytes (Agaiby et al. 2000; Stadler et al. 2000). The mechanism of proliferation was proposed to involve photostimulatory effects in mitochondria processes, which enhanced growth factor release, and ultimately led to cell proliferation (Bjordal et al. 2007). Kreisler et al showed (Kreisler et al. 2003) that the attachment and proliferation of human gingival fibroblasts were enhanced by LLLT in a dose-dependent manner. LLLT modulated matrix metalloproteinase activity and gene expression in porcine aortic smooth muscle cells (Gavish et al. 2006). Shefer at el. showed (Shefer et al. 2002) that LLLT could activate skeletal muscle satellite cells, enhancing their proliferation, inhibiting differentiation and regulating protein synthesis.

3.2. Biphasic Response—irradiance

As early as 1978 Endre Mester observed a “threshold phenomenon” after laser irradiation of lymphocytes in vitro (Mester et al. 1978). Peter Bolton in 1991 irradiating macrophages with two different irradiances (W/cm²) but the same energy density (J/cm²) recorded different results (Bolton et al. 1991). Karu (Karu and Kolyakov 2005) found a dependence of stimulation of DNA synthesis rate on light intensity at a constant energy density 0.1 J/cm² with a clear maximum at 0.8 mW/cm². In another study (Karu et al. 1997) the same group found no less than seven maxima in the dose vs. biological effect curves using a pulsed 810-nm diode laser. Four different biological models were used: luminol-amplified chemiluminescence measured in nucleated cells of murine spleen (splenocytes), bone marrow (karyocytes), and murine blood and adhesion of HeLa cells cultivated in vitro. The peaks coincided for all four models. Anders conducted the widest ranging in-vitro study (on normal human neural progenitor cells) with four different energy density groups, each group tested across a range of six different irradiance parameters (Anders et al. 2007) Table 3.

In 1979 Ginsbach found that laser stimulation of wound closure had “no reciprocity relation”. His controlled experiments on rats with He-Ne laser at an energy density of 4 J/cm² found stimulation at an irradiance of 45 mW/cm² but not at 12.4 mW/cm² (Ginsbach 1979). Uri Oron (Oron et al. 2001a) showed different reductions of infarct size after induced heart attacks in rats. Keeping energy density constant and varying the irradiance he found that the beneficial effects were maximumal at 5 mW/cm² and significantly less effect both at lower irradiances (2.5 mW/cm²) and also at higher irradiances (25 mW/cm²). Ray Lanzafame (Lanzafame et al. 2007) conducted a study varying irradiance and interval on laser-induced healing of pressure ulcers in mice. Energy density (5 J/cm²) was fixed but four different irradiance (0.7 – 40 mW/cm²) parameters were tested with a significant improvement only occurring for 8 mW/cm²

We know of only one human clinical trial which varied irradiance but this trial kept treatment time the same so energy density (J/cm²) did not remain the same. This RCT by Hashimoto on the treatment of the stellate ganglion to reduce pain in patients with post herpetic neuralgia of the facial type. This study compared the effects of 830-nm lasers delivering 60 mW, 150 mW and placebo, each applied for 3 minutes to the anterior aspect of the lateral process of the 7th cervical vertebrae. Each patient had three treatments (one treatment, three consecutive days), each treatment was with a different laser or placebo. The study was properly blinded and randomized. There was a significant difference in skin temperature of the forehead and in recorded pain scores. The greatest improvements were for the 150mW laser (Hashimoto et al. 1997).

There have been several systematic reviews and meta analyses of RCTs and these have revealed some irradiance dependant effects: Bjordal published a review of LLLT for chronic joint disorders and identified 14 RCTs of suitable methodological quality, 4 of which failed to report a significant effect because the irradiance was either too high or too low, and/or delivered insufficient energy, the remaining eight studies all produced positive effects (Bjordal et al. 2003). Tumilty reviewed 25 LLLT RCTs of tendinopathies,13 of which (55%) failed to produce a positive outcome, all of these negative/inconclusive studies that recorded irradiance (or could subsequently be established) had delivered an irradiance in excess of the guidelines set by the World Association for Laser Therapy (www.walt.nu) (Tumilty et al. 2009).

3.3. Biphasic Response—time or energy density

Again, Peter Bolton’s study mentioned in 3.2 above had an energy density aspect showing a different response for each of the irradiances used. For the 400mW/cm² study he found increasing energy density from 2.4 J/cm² to 7.2 J/cm² increased fibroblast proliferation, in the 800 mW/cm² group increasing energy density from 2.4 J/cm² to 7.2 J/cm² decreased fibroblast proliferation (Bolton et al. 1991). Anders’ study also mentioned in 3.2 above looked at four energy density groups, and for the irradiance parameters that produced significant results increasing energy density increased neurite length (Anders et al. 2007) Table 3. Yamaura and colleagues found a biphasic dose response in MTT activity in rheumatoid arthritis synoviocytes after 810-nm laser with a peak at 8 J/cm² and less effect at lower and higher fluences (Yamaura et al. 2009). Loevschall measured human oral mucosal fibroblast cell proliferation by incorporation of tritiated thymidine after varying fluences of 812-nm laser delivered at 4.5 mW/cm² and found a biphasic dose response with a distinct peak at 0.45 J/cm² (Loevschall and Arenholt-Bindslev 1994). Another study (al-Watban and Andres 2001) looked at chinese hamster ovary and human fibroblast proliferation after various fluences of He-Ne laser delivered at a constant irradiance of 1.25 mW/cm². Again they found a clear biphasic dose response with a peak at 0.18 J/cm². Zhang et al (Zhang et al. 2003) found a biphasic dose response in human fibroblast cell numbers after treatment with varying fluences of 628-nm light, with a maximum increase of 30% after 0.88 J/cm² and an actual reduction appearing at 9 J/cm². Brondon and colleagues (Brondon et al. 2005) found that two treatments per day caused a bigger increase than 1 or 4 treatments per day measuring proliferation index in human HEP-2 and murine L-929 cell lines. They used a 670 nm light emitting diode device with an irradiance of 10 mW/cm² and each single treatment was 5 J/cm² and the course was stopped after 50 J/cm² had been given (at 10, 5 or 2.5 days).

Lopes-Martins showed a biphasic response to LLLT on the number of mononuclear cells that accumulate in pleural cavity after carrageenan injection. The results showed neutrophil influx mice treated with three different laser fluencies at 1, 2.5 & 5 J/cm²) with 2.5 having the greatest effect (Lopes-Martins et al. 2005).

As stated in 3.2 above, Hashimoto reported on the laser treatment of the stellate ganglion to reduce pain in patients with post herpetic neuralgia of the facial type. The study compared the effects of 830-nm lasers delivering 60mW, 150mW and placebo, The greatest improvements were for the 150mW laser (Hashimoto et al. 1997). Again as stated in 3.2 above, there have been several systematic reviews and meta analyses of RCTs and these revealed some energy density dependant effects (Bjordal et al. 2003; Tumilty et al. 2009).

4.2. Mouse wound healing

In an in vivo study (Demidova-Rice et al. 2007) we used a set of fluences of 635-nm (+/−15-nm) light delivered from a filtered lamp. The model was a full thickness dorsal excisional wound in BALB/c mice treated with a single exposure to light 30 minutes after wounding. These fluences were 1, 2, 10 and 50 J/cm² delivered at constant fluence rate of 100 mW/cm² and taking 10, 20, 100 and 500 seconds respectively. In this model the untreated wound tends to expand for 2–3 days after it was made, but even a brief exposure to light soon after wounding, reduces or stops the expansion of the wound and the integrated time course of the wound size can therefore be significantly reduced. Our hypothesis is that fibroblasts in the edge of the wounded dermis can be transformed into myofibroblasts, and the contractile nature of these cells with their smooth muscle actin fibers prevents the wound expanding. It should be noted that the fibroblast-myofibroblast transition can be mediated by NF-κB activation (Watson et al. 2008). As shown in Figure 11 there was a biphasic dose response with positive effects (difference in integrated area under the curve of time course of wound size compared to no treatment control) seen in low doses with a clear maximum seen at 2 J/cm², and the high dose of 50 J/cm² actually gave a worsening of the wound healing time curve i.e. there was a greater expansion of the wound compared with non-treated controls.

Open in a separate window

FIGURE 11

Biphasic dose response in measured difference in integrated area under the curve of time course of wound size compared to no treatment control, with a clear maximum seen at 2 J/cm², and the high dose of 50 J/cm² gave a worsening of the wound healing time curve.

5.1. Excessive ROS

As discussed in 2.5 the light mediated generation of reactive oxygen species has been observed in many in vitro studies and has been proposed to account for the cellular changes observed after LLLT via activation of redox sensitive transcription factors (Chen et al. 2009a). The evidence of ROS mediated activation of NF-κB in MEF cells presented in 4.1 provides additional support for this hypothesis (Chen et al. 2009a). It is well-accepted that ROS can have both beneficial and harmful effects (Huang and Zheng 2006). Hydrogen peroxide is often used to kill cells in vitro (Imlay 2008). Other ROS such as singlet oxygen (Klotz et al. 2003) and hydroxyl radicals (Pryor et al. 2006) are thought to be harmful even at low concentrations. The concept of biphasic dose response in fact is well established in the field of oxidative stress (Day and Suzuki 2005). If the generation of ROS can be shown to be dose dependent on the delivered energy fluence this may provide an explanation for the stimulation and inhibition observed with low and high light fluences.

5.2. Excessive NO

The other mechanistic hypothesis that is put forward to explain the cellular effects of LLLT relates to the photolysis of nitrosylated proteins that releases free NO (see section 2.6). Again the literature has many papers that discuss the so-called two-faced or “Janus” molecule NO (Anggard 1994; Lane and Gross 1999). NO can be either protective or harmful depending on the dose and particularly on the cell or tissue type where it is generated (Calabrese 2001a).

Research in the Hamblin laboratory is supported by the US NIH (grants R01CA/AI838801 and R01AI050875 to MRH).