Seven-week-old female C57BL/6J mice were purchased from Jackson Laboratory. CD11c-eYFP mice were donated by Dr. Nussenzweig at Rockefeller University. All measurements were performed in a blinded manner, (to control or experimental groups).

We used a neodymium-doped yttrium orthvanadate (Nd:YVO₄) laser (RMI Laser, Lafayette, CO). The 1064 nm laser can be set to emit either continuous wave (CW) output or nanosecond-duration pulses (PW) at a periodicity of 10 kHz, while at 532 nm the output is only PW. Average output powers were determined using a power meter for each illumination (Thorlabs). The beam profile for all exposures was flat, with a less than 50% variation in beam intensity from center to edge. The laser diameter on the skin was measured approximately 5 mm (0.2 cm²). Mice were depilated using a hair remover (Nair, Church & Dwight). The following day, the shaved skin of anesthetized mice was illuminated with the laser on 4 spots for ovalbumin (OVA) and 1 spot for influenza studies. The skin temperature was measured during the procedure using an infrared thermal imager (FLIR Systems).

For visual inspection, we observed for any signs of skin damage including blistering, bruising, crusting, edema, redness or swelling during and at 0, 1, 2, and 4 days after laser illumination. For skin histology, mice were heart-perfused with 4% paraformaldehyde before, or at 3, 6, 24, 48, and 96 hours after laser illumination. Five µm-thick paraffin sections were H & E stained and examined for microscopic tissue damage, and polymorphonuclear infiltration were quantitated on the slides in 5 randomized fields using Image J freeware (NIH).

Mice were injected intradermally (i.d.) using a 28 G insulin grade syringe (Kendal) with chromatographically purified OVA (10 µg in 10 µL saline per spot, 4 spots, Worthington), which was found to contain less than 1.75 EU/mg of endotoxin using the Limulus amebocyte lysate QCL-1000 (Cambrex). I.d. delivered alum- (Imject®, Thermo-Fisher) adjuvanted OVA, prepared per manufacturer's instructions, served as positive control. A dose of 5 µL of Imject® (200 µg aluminum hydroxide plus 200 µg magnesium hydroxide) was used per spot. Blood samples were drawn at 3, 6 and 12 weeks post-vaccination via retro-orbital bleeding.

Mice were injected i.d. with whole inactivated influenza virus A/PR/8/34 (H1N1) (1 µg in 10 µL saline, 1 spot, Charles River). Alum-adjuvanted vaccine served as a positive control. Blood samples were taken 28 days after immunization and 4 days post-challenge with an intranasal application of live influenza virus 4 weeks after vaccination as previously performed in the context of i.d. influenza vaccination [28]–[31].

Immulon™ 2 HB Flat Bottom Plates (Thermo-Fisher) were coated overnight with 1 µg of OVA at a concentration of 5 µg/mL or 0.2 µg of inactivated influenza virus at 1 µg/mL. Serially diluted mouse serum samples were added to the wells and incubated for 1 hour after the plates were blocked. Bound immunoglobulins were detected with the appropriate horseradish peroxidase-conjugated secondary antibody (goat antibody to mouse IgG [1:10,000, Sigma-Aldrich]; rat antibody to mouse IgG1 [1:4,000, SouthernBiotech]; rat antibody to mouse IgG2b [1:500, SouthernBiotech]; goat antibody to mouse IgG2c [1:4,000, SouthernBiotech]; goat antibody to mouse IgA [1:1,000, Sigma-Aldrich]; rat antibody to mouse IgE [1:1,000, SouthernBiotech]). In the case of IgE, the wells were further treated with ELAST ELISA Amplification System (Perkin Elmer) to improve sensitivity of the assay. At the end of the incubation, TMB substrate (1-Step Ultra TMB, Thermo-Fisher) was added and the reaction was stopped with 2 N sulfuric acid. The reproducibility of the assay was ascertained by applying mouse anti-ovalbumin IgG (Sigma-Aldrich) or a hyperimmune mouse serum to influenza to each plate. We measured the absorption at 450 nm using an ELISA reader (TECAN Sunrise™ plate reader, TECAN). For antibody titers to OVA and IgE antibody to influenza, a statistically defined endpoint antibody titer was determined with a confidence level of 99% as previously described [32]. For antibody titers to influenza except IgE isotype, a titer was designated as a serum dilution corresponding to an inflection point.

Mouse sera were analyzed for HAI titers by Charles River Avian Vaccine Services as described previously [33], [34].

Mice were anesthetized and challenged intra-nasally with live influenza A/PR/8/34 at a dose of 1.5×10⁶ 50% egg infectious doses (EID₅₀), which is equivalent to 3×10³ 50% mouse lethal dose (MLD₅₀), in 30 µL saline 28 days after vaccination. Survival and body weight were monitored for 14 days post-challenge. Mice showing a hunched posture, ruffled fur, or greater than 20% body weight loss, or mice which were not eating or drinking, were considered to have reached the experimental end point and were euthanized [33], [34]. MLD₅₀ titers were determined by inoculating groups of 10 mice intranasally with serial 10-fold dilutions of virus using the Reed-Muench formula as previously described [33], [34].

Four days after the virus challenge, both sides of the lung were isolated and homogenized. The EID₅₀ m/L values were determined by serial titration of the lung homogenate in eggs by Charles River Avian Vaccine Services as described previously [33], [34].

Splenocytes were harvested 4 days after the virus challenge as previously performed in the context of i.d. influenza vaccination [28]–[31]. The 2×10⁶ splenocytes were re-suspended in 100 µl of media and incubated for 5 hours in the presence of an inhibitor of Golgi function (Golgi plug, BD Bioscience) and 1 µg/mL of influenza A MHC class I (NP_366–374, ASNENMETM, Anaspec), or II peptides (NP_311–325, QVYSLIRPNENPAHK). Multi-parameter surface staining for CD3ε, CD4, and CD8α (CD3ε: 145-2C11; CD4: RM4-5; CD8α: 53-6.7, BD) was performed, followed by fixation, permeabilization in Cytofix/Cytoperm™ (BD Bioscience) according to the manufacturer's instructions and intracellular staining for IFN-γ, IL-17A, IL-5 (IFN-γ: XMG1.2; IL-17A: TC11-18H10; IL-5: TRFK5, BD). Cell subpopulations, including influenza-specific IFN-γ, IL-17A, IL-5 producing CD3+CD4+T-helper cells and CD3+CD8+T-cytotoxic cells, were quantified as a percentage of total viable cells by flow cytometry using a BD 4 Laser LSR II (BD). Analysis was completed using FlowJo software (Tree Star).

Mice were injected with OVA labeled with Alexa Fluor 647 (OVA₆₄₇, 10 µg in 10 µl saline per spot, 4 spots in total, Invitrogen) with or without laser illumination. 24 hours after vaccination, skin-draining lymph nodes (dLNs) were harvested, minced, and stained for the cell surface and maturation markers CD11c, CD86, CD80, CD40, and I-A^b (CD86: GL1; CD80: 16-10A1; I-A^b: AF6-120.1; BD, CD40: 1C10; CD11c: N418, eBioscience). We then performed flow cytometry on CD11c-positive dendritic cells (DCs).

To examine temporal and spatial expression of CCL2 and CCL20, and to quantify DCs in laser-treated skin, we performed immunofluorescence and confocal microscopic analysis of laser-treated skin from CD11c-eYFP mice, in which DCs are intravitally labeled [35], [36]. Control animals received sham treatments in which they were anesthetized and shaved, but not treated with the laser. Mice were heart-perfused with 4% paraformaldehyde in PBS via the left ventricle before, or at 6 or 24 hours after 1 minute of CW 1064 nm laser treatment. The skin was then harvested and embedded in OCT compound (Tissue-Tek, Sakura). The primary antibodies used were rat anti-mouse CCL2 (1:50, R&D Systems) or rat anti-mouse CCL20 (1:10, R&D Systems). Goat anti-rat IgG (Dylight 549, 1:200, Jackson Immunoresearch) was the secondary antibody used. 10 µm sections were washed and incubated with the respective primary antibody overnight at 4°C followed by a wash and 1 hour incubation at room temperature with the secondary antibody. Appropriate negative controls were prepared by omission of the primary antibody. Following application of the secondary antibody, tissues were washed, counterstained with To-Pro-3 (1:5000, Invitrogen) and mounted. Digital images of immunofluorescence slides were obtained by means of confocal microscopy (Carl Zeiss LSM5 Pascal; Carl Zeiss, Inc.). YFP-positive DCs were manually counted in 5 randomized fields using Image J 1.43 freeware (NIH).

Skin sections measuring 5×5 mm² and including both the epidermis and dermis were excised 6 hours after laser illumination. Total RNA was extracted using the RNeasy Fibrous Tissue Mini Kit (Qiagen) and reverse-transcribed using the RT² First Strand Kit (Qiagen). The samples were tested on an RT² Profiler™ PCR Array System (Qiagen) on an Mx3005™ Multiplex Quantitative PCR System (Stratagene). The fold change in mRNA expression over sham-treated controls was normalized against housekeeping genes and calculated following the 2^–ΔΔCT method.

We performed an open-label, single-center, single-arm study in healthy adults. We used a clinically-approved Q-YAG 5 laser emitting light at 1064 nm (Palomar). The laser was altered to operate at an average power level below 2 W with an exposure area of 0.5 cm², pulse duration of 3 nanoseconds, and pulse frequency of 10 Hz. We selected qualified subjects with either skin phototype V or VI [37]. We exposed each subject to a range of doses by increasing the average irradiance level of each exposure from 0.5 to 3.7 W/cm² in stepwise increments of about 0.2–0.4 W/cm² each for up to 120 seconds. While subjects were asked to immediately report any sensations during and after the exposure, the operator recorded any signs of visible skin damage. An individual test exposure was stopped if there were any indications of the subject's discomfort, pain, or distress, or if the investigator noted any signs of skin damage. At the end of the exposure, the operator acquired a digital photograph of the area and scored the skin sensations and damage. Subjects returned 2 days later for follow-up assessments of their skin conditions and an additional digital photograph of the area. Subjects came in after 2 weeks if the subject indicated any appearance of skin markings in the area of the laser exposure.

All animal procedures were performed following the Public Health Service Policy on Humane Care of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of Massachusetts General Hospital. The laser safety study in humans was performed in full conformity with the Declaration of Helsinki, the current International Conference for Harmonisation Good Clinical Practice (ICH-GCP) regulations and all applicable regulatory and ethical requirements. The study was approved by New England Institutional Review Board (NEIRB) and the registry number is #09-325. All subjects provided written informed consent before study-related procedures were performed.

We used the Mann Whitney U-test for the comparison of numerical values between 2 groups, and the Kruskal-Wallis followed by the Dunn's test for comparisons of more than three groups for all statistical analyses unless otherwise specified. Data were pooled from at least two independent experiments.

We first sought to establish the maximum non-tissue damaging dosages for the near-infrared (NIR) laser for both continuous wave (CW) and nanosecond-duration pulse wave (PW) mode as well as the previously described PW visible lasers [25], [26]. Mice received exposures at escalating irradiances (0.5 to 1.5 W/cm² for PW 532 nm laser; 0.5 to 6.0 W/cm² for CW or PW 1064 nm lasers) for durations up to 4 minutes. Skin damage was evaluated after illumination by visual inspection and histology. Maximum safe irradiances were considered to be those at which skin temperatures did not exceed 43°C and for which no visible or microscopic skin damage was apparent at any of the post-exposure evaluation times [27], [38]. We identified 1.0 W/cm² as the maximum safe irradiance for the PW 532 nm laser (Figure 1A) and 5.0 W/cm² for both the 1064 nm PW (Figure 1B) and CW lasers (Figure 1C). No visual damage such blistering, bruising, crusting, edema, redness or swelling damage was seen when the laser power was below the safe irradiance for each parameter for the PW 532 nm laser (Figure S1A) and 5.0 W/cm² for both the 1064 nm PW (Figure S1A) and CW lasers (Figure 1D). On a histological examination, no tissue damage or inflammatory response at any given time point was detected by H & E staining (Figure 1E) with minimal polymorphonuclear cell-infiltration in the skin after the administration of 5.0 W/cm² of CW 1064 nm laser (Figure 1F). Thus, we concluded that the dosages below 1.0 W/cm² for the PW 532 nm and 5.0 W/cm² for both the 1064 nm PW and CW lasers are non-tissue damaging and non-inflammatory.

In order to determine if the NIR laser dose used in mice is safe and tolerable in humans, we performed a clinical study (Figure 2A). Five subjects with skin phototypes V or VI were enrolled [37], and each subject received 16 exposures of a clinically-approved 1064 nm laser at escalating irradiances (0.5 to 3.7 W/cm²) for durations up to 2 minutes each (Table 1). All subjects tolerated the highest irradiance for 2 minutes (total dose 442 J/cm²) with no subject reporting severe skin sensations or distress (Table 2). Investigators noted no significant skin damage during any laser exposure (Figure 2B–D). Laser-induced skin damage is a function of skin heating, which depends upon the duration of exposure, wavelength and irradiance [38]. With the same wavelength and a similar order of irradiance and the duration of exposure, the heat generation in the animal and human studies is equivalent. Thus, we conclude that a NIR laser at equivalent irradiances and doses used in mice is safe and well tolerated in humans.

We next tested whether NIR laser treatment at safe irradiances and doses could enhance immune responses to vaccination in mice as compared to the licensed adjuvant, alum. First, mice received 1- or 4-minute exposures to CW or PW 1064 nm or PW 532 nm lasers on the back skin. Immediately thereafter, mice received an i.d. injection of OVA. Alum-adjuvanted OVA i.d. served as positive controls as i.d. injection of alum has been used to increase the efficacy of vaccines both in mice [39]–[42] and humans [43], [44]. Notably, the 1-minute CW 1064 nm laser treatment induced the highest antibody titer among all the tested parameters, which were significantly higher at all time points than both the non-adjuvanted controls (Figure 3A, 3 weeks: P<0.01; 6 weeks: P<0.05, 12 weeks: P<0.05). The anti-OVA specific IgG antibody titer in the previously explored 4-minute PW 532 nm laser-treated group was not significantly higher than those in non-adjuvanted controls at any time point (Figure 3A). There is no significant difference between CW 1064 nm laser-treated and alum-adjuvanted groups at any time point (Figure 3A). We did not find a relationship between maximal skin surface temperature during laser treatment and antibody titer (Figure 3B), suggesting that heat generation as a result of laser exposure does not play a significant role in enhancing immune responses. We selected the 1-minute CW 1064 nm laser exposure as the most effective and efficient immune adjuvant for subsequent experiments.

Previous studies indicate that licensed and experimental immunologic adjuvants activate DC-mediated innate immune responses resulting in robust adaptive immune responses [3], [10], [45]–[47]. To investigate if the NIR impacts DC trafficking and function, we injected fluorescently labeled OVA i.d. into mice and assessed both OVA-positive DCs in skin-dLNs and their activation status. The NIR laser treatment induced an up-regulation of maturation markers including MHC class-II, CD40, and CD86 compared to controls injected i.d. with OVA only (Figure 4A–C, MHC II: P = 0.013; CD40: P = 0.010: CD86: P<0.037), but did not increase CD80 expression or OVA-positive DC populations above those of the controls (Figure 4D and 4E). To further determine the impact of NIR laser exposure on DC trafficking in the skin, we treated the skin of CD11c-YFP transgenic mice, in which DCs are intravitally labeled [35], [36], with a 1-minute exposure of the CW 1064 nm laser. We observed an over 2-fold increase in the concentration of DCs in both the epidermal and dermal areas of exposure, reaching a maximum 6 hours after treatment (Figure 5A and 5B, no laser control vs. laser treated in epidermis: 589±125 vs. 1,045±113/mm², P = 0.038; dermis: 195±36 vs. 415±62/mm², P = 0.040), and returning to the baseline level of non-treated skin by 24 hours. These data suggest that key mechanistic elements for the NIR laser include migrational and functional changes of DCs in skin and draining lymph node.

After establishing the beneficial effect of NIR laser upon DCs, we sought to identify the mechanisms contributing to the migration and activation of DCs by the NIR laser. Skin cells function as sentinels for damage or pathogen invasion by releasing pro-inflammatory cytokines to recruit and condition APCs including DCs [45]. We therefore tested the effect of NIR laser upon chemokine production and signaling. We measured the expression of 160 genes related to inflammatory cytokines, their receptors, and inflammasomes using qPCR, 6 hours after the 1-minute CW 1064 nm laser treatment, Gene expression for a selective set of cytokines, including Ccl2 and Ccl20, increased significantly (Figure 5C and Table S1). CCL2 and CCL20 protein expression has been shown to be involved in DC migration and recruitment [48]. 6 hours after the laser treatment, CCL2 was expressed in the epidermal and dermal regions, possibly in keratinocytes, fibroblasts and mast cells (Figure 5D, top row). CCL20 was expressed sporadically in the dermis, possibly in mast cells (Figure 5D, middle row). The expression of both CCL2 and CCL20 declined at 24 hours, matching the timing of DC localization to the laser-treated skin. Taken together, these data indicate that the NIR laser adjuvant stimulates the expression of a defined set of cytokines and chemokines which collectively could induce functional and migrational changes in DCs in the skin.

We next examined the adjuvant effect of visible and NIR lasers in a murine influenza vaccination and lethal challenge model and compared them with alum. Mice received a single laser dose and were injected i.d. with whole inactivated influenza virus A/PR/8/34. We also used alum-adjuvanted vaccine to examine whether the laser adjuvant can induce responses comparable to a licensed chemical adjuvant, as i.d. injection of alum has been used to increase the efficacy of vaccines both in mice [39]–[42] and humans [43], [44]. The CW 1064 nm laser significantly augmented pre-challenge IgG and IgG1 titers compared to the non-adjuvanted group (Figure 6A–C, IgG: 1064 nm vs. controls: P = 0.012, IgG1: 1064 nm vs. controls: P = 0.040). Alum also produced elevations in IgG1 titers that were greater than non-adjuvanted controls (P<0.0001). The IgG2c responses were similar among all test groups. Post-challenge, the CW 1064 nm laser significantly augmented anti-influenza IgG, IgG1 and IgG2c titers compared to the non-adjuvanted group (Figure 6D–F, IgG: P = 0.001; IgG1: P = 0.026; IgG2c: P = 0.023). In comparison, the PW 532 nm laser augmented only IgG and IgG2c titers (IgG: P = 0.028; IgG2c: P = 0.003), and alum increased only IgG1 titers (P<0.001). The finding that alum elicited an IgG1-biased response is consistent with published literature showing that alum induces a profoundly polarized T helper type 2 (T_H2) immune response with consequently elevated IgE production and hypersensitivity in mice [3], [10], [15], a finding replicated in this study (Figure 6G, P<0.0001 compared to the non-adjuvanted group). In contrast, the laser adjuvants did not increase IgE responses to the vaccine (Figure 6G). Influenza-specific IgA in the lung homogenate was not detected in any experimental group (data not shown), suggesting mucosal IgA does not contribute to protection in this model. These data indicate that NIR laser treatment produces a mixed T_H1-T_H2 immune response to influenza vaccination without enhancing an IgE response.

Previous studies suggest that effector CD4+ and CD8+ T cell-mediated immune responses contribute to protection from influenza [49], [50]. To determine whether the NIR laser elicits these cell-mediated immune responses, we re-stimulated splenocytes from influenza-challenged mice ex vivo with influenza peptides and then assessed the expression of cytokines in T-cell subpopulations. Influenza-specific CD4+IFN-γ+ T-cell subpopulations induced by a major histocompatibility complex (MHC) class-II peptide were significantly increased in the 1064 and 532 nm laser-treated groups compared to non-adjuvanted controls (Figure 7A, CW 1064 nm laser: P = 0.044; PW 532 nm laser: P = 0.003). Only the CW 1064 nm laser also significantly increased influenza-specific CD4+IL-5+ T-cell subpopulations (Figure 7B, P<0.003). No test groups showed significantly increased responses of influenza-specific CD4+IL-17+ T-cell subpopulations to MHC class-II peptide, or CD8+IFN-γ+ T-cells to MHC class-I peptide, compared to the non-adjuvanted controls (Figure 7C and 7D). These data show the unique ability of the non-tissue damaging NIR laser to induce a systemic T_H1-T_H2 immune response to an inactivated influenza vaccine.

After determining the quality of the laser-induced immune response, we sought to test the ability of this strategy to generate a protective response against lethal influenza virus challenge. Mice were vacciniated and allowed to rest for four weeks. Subsequently, naïve and vaccinated mice were challenged intra-nasally with homologous live influenza virus and monitored for survival time. In the lethal challenge model, the CW 1064 nm laser-treated group showed a marked decrease in lung viral titers by a factor of 10^1.9 compared to the non-adjuvanted group at 4 days after challenge (Figure 8A, P = 0.025), while the PW 532 nm laser-treated group failed to show any significant impact. In addition, the single one minute duration CW 1064 nm laser treatment consistently conferred better protective immunity while the PW 532 nm laser treatment did not, as determined by survival time after viral challenge (Figure 8B, P = 0.036). Body weight loss upon viral challenge in CW 1064 nm laser-treated group was smaller than in PW 532 nm laser-treated group (Figure S2). There was no significant difference in protection and body weight loss between CW 1064 nm laser-treated and alum-adjuvanted groups. Consistent with protection and antibody levels, clinically relevant HAI titer levels in the CW 1064 nm laser group were higher than in the non-adjuvanted, 532 nm laser-treated and alum-adjuvanted groups in pre-challenge serum (Figure 8C), and higher than the non-adjuvanted group and comparable to those in the alum-adjuvanted group in post-challenge serum (Figure 8D). These data support the view that NIR laser induces protective immune responses to an inactivated influenza vaccine.