The available hyperthermic techniques for cancer tumor therapy1, 2, 3, 4 possess low spatial selectivity in the heating of tumors and surrounding healthy tissues. One of the ways to improve the laser heating spatial selectivity is tumor tissue photothermal labeling by gold nanoparticles with different shapes and structures, such as nanoshells,5, 6, 7 nanorods,8, 9, 10 nanocages,11 and others.11, 12, 13, 14 By exposing nanoparticles to laser radiation near their plasmon-resonant absorption band, it is possible to produce local heating of nanoparticle labeled cells without harming surrounding healthy tissues. Such an approach has been developed over the last five years and it has been called plasmonic photothermal therapy (PPTT).15 The spectral tuning of nanoparticle resonance to the “therapeutic optical window” and achieving the desired ratio between the absorption and scattering efficiencies can be achieved by variation in the particle size, shape, and structure.16, 17, 18, 19, 20 In practice, almost all reports on PPTT used diode lasers with the wavelength around . Prior to laser treatment, the index matching agents (e.g., PEG diacrylate) may be applied over the tumor surface.21.
Gold nanoshells (GNSs) belong to a prospective class of optical adjustable nanoparticles with dielectric (silica) core included in a thin metallic (gold) shell.22 By an appropriate choice of the ratio of core diameter to shell thickness and surface functionalization, the so-called immunotargeted GNSs can be engineered that effectively absorb or scatter light at a desired wavelength from visible to infrared, and provide selective labeling of cancer cells.23 The surface PEGylation of GNSs ensures their stability against agglomeration at physiological conditions during systemic circulation.24, 25 The absorption cross section of a single GNS is high enough to provide a competition of nanoparticle technology with applications of indocyanine green dye—a typical photothermal sensitizer used in laser cancer therapy.26 Enhanced efficiency of absorption and scattering of GNSs have been used to develop a combined approach, allowing one to provide selective destruction of GNS-labeled carcinoma cells via PPTT and concurrent optical coherence tomography (OCT) imaging of tumor tissues.27
There are several papers on in vitro experiments concerning application of NSs to PPTT of cancer cells,5, 6, 28, 29, 30, 31 while the number of in vivo and ex vivo studies is quite limited. Moreover, although several preclinical reports on the utility of GNSs for PPTT and OCT applications are available,5, 21, 32, 33 there remain several areas still to be studied more thoroughly. First of all, the question is about controlled and localized hyperthermia without significant overheating of both tumor and surrounding normal tissues. The distribution of elevated temperature under PPTT treatment is determined by absorption of light by nanoparticles acting as point-wise local heat sources, and by thermal diffusion over surrounding tissues. Practically at PPTT one needs to provide the temperature increment in the range . To achieve these temperature increments, the GNS concentration of , the laser power density of , and the treatment exposures between 1 and were adjusted properly in the published papers. It should be noted that biological effects have nonlinear dependence on changes in the particle concentration and the delivered laser power density, which is defined by the type of tissue and thermoregulation ability of a living organism. A possible method for intelligent PPTT may be adaptive heating in real-time control mode, as described in Ref. 33. Such a technology implies a precise control of local temperature within tumor and surrounding tissues. The use of thermocouples21 provides a direct but invasive method, which can be used for laboratory experiments only. In addition, the thermocouples response may be affected by unwanted direct laser heating rather than by the thermal equilibrium with a local tissue volume. More sophisticated and nondirect techniques such as temperature-dependent phase transition of hydrogels34 or ultrasound imaging35 have been recently described. The temperature distribution was also measured by in vivo magnetic resonance temperature imaging (MRTI) with the proton-resonance frequency-shift method.21, 33 The temperature resolution of MRTI measurements is less than , which is close to the resolution of IR thermography. IR thermography is simple and less expensive but is restricted to retrieving surface-temperature data only. It needs the design of special algorithms for reconstruction of in-depth temperature profiles of tissue on the basis of temperature surface measurements.4
Precise control over the local temperature distribution is the key factor to be considered in the context of enhanced PPTT efficacy. Laser heating can result in both tumor necrosis (and possibly apoptosis) and in accelerated tumor growth, depending on the accuracy of heating and on the rise in tumor temperature on illumination with laser light. Specifically, heating up to may lead to the acceleration of biological reactions accompanied by the production of shock-heating proteins and by intense growth of the tumor.36 It is generally accepted that one needs to ensure a steep rise in tumor temperature followed by stability over a given time period. The use of plasmon-resonant particles in general meets the basic goals of PPTT related to controllable tumor heating. It is significant that gold nanoparticle light-mediated therapy is on the way to clinical trials in humans; however, extensive animal studies should be done prior.
To characterize tissue heating effects, the morphological investigations in the framework of therapeutic pathomorphology should be provided. Although such an approach has been used for estimation of hyperthermic effects of gold nanorods in tumor cells,9 studies of morphological changes under in vivo conditions are quite rare5, 21 and they still need additional efforts to be completely understood.
The present work is focused on the investigation of heating kinetics, spatial temperature distribution, and morphological alterations in tissues dependent on laser irradiation parameters and nanoparticle concentration. We describe experimental results for phantoms and laboratory animals at several temperature regimes and related morphological tissue patterns.
Materials and Methods
The following reagents were used to provide protocol of gold nanoparticle synthesis: tetraethyl orthosilicate (TEOS, Aldrich, Saint Louis, Missouri) tetrakis (hydroxymethyl) phosphonium chloride (THPC, Fluka, Saint Louis, Missouri), 3-aminopropyltrimethoxysilane (APTMS, Aldrich), PEG-thiol (Nektar, San Carlos, California), absolute ethanol, tetrachloroauric acid (TCAA, Aldrich), potash (Reachim Company, Russia), and formaldehyde (Serva, Heidelberg, Germany). All chemicals were of research grade, and 25% aqua ammonia was of analytical grade. Ethanol was purified by additional distillation.
Gold Nanoparticle Characteristics and Quality Control
For experiments described in this work, the plasmonic nanoparticles (gold nanospheres and nanoshells) were fabricated in the Laboratory of Nanobiotechnology at the Institute of Biochemistry and Physiology of Plants and Microorganisms (IBPPM) RAS. The poly (ethylene glycol) molecules have been attached (using end thiol groups) to the surface of nano-particles to provide their defense from reticuloendothelial systems of the organisms under study.
Silica/gold nanoshells were fabricated as described in paper Ref. 37, with minor modifications related to reagent concentrations. First, silica nanoparticles were grown by reducing TEOS with in absolute ethanol. To do this, we mixed of absolute ethanol with 25% aqua ammonia and TEOS. For the preparation of gold seeds, of aqueous NaOH and of 80% THPC were added to of triply distilled water. The solution was vigorously agitated on a magnetic stirrer at , and after that of a 1% TCAA solution were added. Next, aminated silica particles were added to the gold-seed dispersion. Gold particles adsorb to the amine groups on the silica surface, resulting in silica nanoparticle covering by the gold colloid. Gold nanoshells were then grown by reacting with the silica/gold colloid particles in the presence of formaldehyde at room temperature. In this process, additional gold reduces on the adsorbed metal particles, which act as nucleation sites. The nanoshells were centrifuged and sonicated in HPLC-grade water before use.
The following PEGylation protocol was used for silica/gold nanoshells. The nanoshells’ solution was centrifuged at for to pellet the particles, decanted, and then nanoshells were resuspended in the same amount of water. One hundred microliters of potassium carbonate and of PEG-thiol solution were added per of the nanoparticles’ solution. The mixture was stored overnight at room temperature, then the centrifuged, decanted, and prepared PEGylated nanoshells were resuspended in water several times to remove the PEG-thiol excess.
Nanoshells have silica cores of in diameter with -thick gold shells. These parameters have been chosen to provide plasmon resonance in the IR region of the spectrum where diode lasers radiate.
Parameters of fabricated gold nanoparticles were tested by spectral analysis (measuring of optical density) and by transmission electron microscopy (TEM). The average diameter of nanoshells of was evaluated using TEM micrograms of nanoshell suspensions (Fig. 1 ). Synthesized nanoshell suspensions had a sufficient narrow size distribution.
Spectral characteristics of gold nanoparticles calculated using the Mie theory are presented in Fig. 2a . Solid gold nanospheres of 15 and in diameter have extinction maximums at the wavelengths near . Optical properties of such particles are characterized by strong absorption and weak scattering. For providing tissue hyperthermia by laser diode, the best is using gold nanoshells with the ratio of core diameter to shell thickness as , because these particles have extinction maximums at this particular wavelength, which fits well with the optical transparency window of the tissue. The measured spectrum of optical density of fabricated silica/gold nanoshells is presented in Fig. 2b. There is a good correlation between calculated and measured spectra; however, the experimental curve is smoother due to particle size heterogeneity.
For in vivo experiments, the white laboratory rats, of age (weight ), were used. Animals were bred under SPF conditions and a barrier was maintained for animals involved in experiments. Drinking water and conventional food were provided ad libitum. Husbandry conditions were maintained according to all applicable provisions of the national laws. The experimental protocol was approved by an independent ethical committee prior to the study.
Investigation of thermal effects and alterations of tissue morphology induced by laser irradiation at subcutaneous and/or intramuscular injection of silica/gold nanoshell suspensions was carried out using white laboratory rats. In solution of designed silica/gold nanoshells , gold content was of . Similar concentration expressed in the number of particles, , was used in in vitro experiments with tissue phantoms. In in vivo studies, 48 animals were divided into two groups, 24 rats in each; rats from the first and the second groups were irradiated by continuous wave (cw) and pulse laser light, accordingly. Each of the 24 animals in each group was injected with a solution of gold nanoparticle suspension once subcutaneously and then in intramuscularly. The first injection was done subcutaneously in front of the abdominal wall, the second intramuscular one was done into the hip far enough from the first injection area. After two minutes elapsed after each injection, the skin area around the injection puncture was exposure to laser light to produce photothermal effects mediated by the inserted nanoparticles within the skin or within the muscle tissue. A delay and different locations of treated areas allowed us to consider tissue photothermal responses at subcutaneous and intramuscular gold nanoparticle delivery for the particular animal as independent ones. We used the diode laser from Opto Power Corporation (Tusca, Aricona), providing a mean wavelength of , optical-fiber output, and cw and pulse regimes. The cw laser output power was . In pulse mode, peak power was , with an on-off time ratio of 0.25, and pulse duration of . The preliminary estimation of heating kinetics and size of the heated area was carried out at different concentrations of nanoparticles in a test tube. For noninvasive measurement of spatial distribution of surface temperature of objects under investigation, the IR Imager IRISYS 4010, Infrared Integrated System Limited (Northampton, United Kingdom) was used. General anesthesia was used for providing experiments in vivo. Tissue samples for histological examination were collected after injection of nanoparticles before, after, and past laser heating. Heating duration was from . To prove the evidence of nanoparticle delivery into the tissue and to evaluate qualitatively their spatial distribution within the tissue, TEM and dark-field optical microscopy were used.
Imaging of Gold Nanoparticles
The dark-field optical microscopy was used for imaging of gold nanoparticles in tissue slices [Biolam-M microscope (LOMO, Saint Petersburg, Russia)].
For electron microscopic examinations, the cut-out strips of tissue of size were rapidly ground by blade to pieces not more than . Further fixation, dehydration, and subsequent impregnation with resin via standard procedures were carried out. Tissue thin slices were obtained by the ultratome Reichert (Austria), and thickness was from . Each slice was mounted to the grids without the base layers and contrasted by the saturated alcoholic solution of uranyl acetate at a temperature of during . Specimens were investigated using a Hitachi Hu-1a electron microscope (Japan) in transmission microscopy (TEM) mode.
Diffuse backscattering spectroscopy at the application of fiber optic spectrometer LESA-01 was used for noninvasive in vivo detecting of subcutaneously injected gold nanoparticles. The sample was illuminated with the source fiber (fiber core diameter of ), and diffuse reflected light was collected with six detector optical fibers , which were apart from the source fiber. Distance between the fiber probe tip and skin was . The detector fibers were coupled to a spectrometer, and a diffuse reflectance spectrum was collected over a wavelength range of . Prior to spectral analysis, recorded signals were corrected for system response.
An IR thermograph was used for noninvasive monitoring of surface temperatures. The thermal imaging system IRISYS 4010 is based on the uncooled bolometric matrix providing temperature recording in the range from , with sensitivity of . The system is sufficiently fast to provide animal studies in vivo. The memory flash of the thermal imager makes it possible to record up to 6000 thermograms with size. Range of focus is from to infinity. The minimal size of the object, whose temperature can be measured, depends on the distance between the IR imager and the object. The wavelength range of the thermal imager comprises , thus the scattered laser radiation does not affect thermal-vision measurements.
Model Experiments on Near Infrared Laser Heating Mediated by Gold Nanoparticles
A preliminary investigation using phantom studies and computer simulations are necessary to solve the inverse problem of reconstruction of in-depth temperature distribution on the basis of surface thermal imaging for the specified experimental conditions. For optimization of laser heating technology, in-depth temperature distributions were investigated using test tubes ( diam) containing solutions of silica/gold nanoshells with different concentrations (Fig. 3 ). Heating was provided at irradiating by a cw NIR diode laser for (total dose of , power density ). From the measurements, it follows that at high concentrations of particles the whole phantom volume is not heated directly by laser light, but only the part which is closer to the incident laser beam. The rest volume of nanoparticle solution is heated up indirectly due to a thermal diffusion process. Thus, inhomogeneity of heating could be expected. With lesser concentration of particles, the whole volume is heated directly by laser light, but this heating could be insufficient. Evident inhomogeneity of in-depth heating of nanoparticle solutions is followed from data presented in Fig. 3. The effective heating depth for highly concentrated solutions does not exceed . This puts restrictions on nanoparticle concentration for clinical applications for tumors with sizes bigger than . If the concentration decreases four times, the effective heating depth is increased two times. Uniform temperature distribution in the test tube is recorded only at the thinning of the nanoparticle solution 16 times. In this case, maximal temperature alterations do not exceed , while at a concentration of , temperature rise is . It is necessary to note that laser radiation is also absorbed by water, thus in control measurements with the test tube filled up by physiological solution, background heating was about .
The temperature measurements using a test tube allowed us to determine the upper limit of the temperature increase in tissues. The concentration of gold nanoparticles in tissues is conceivably lower than in the initial injected solution. To estimate temperature rise in tissues, experimental dependence of temperature rise on concentration of gold nanoparticles in a phantom can be used (see Fig. 4 ). Temperature kinetics for the solution of gold nanoparticles with a concentration of is shown in Fig. 5 . The observed temporal dependence of temperature rise is nonlinear. The most heating occurs over the first of laser irradiation, then temperature increase is saturated with time. After laser switch off, the temperature decreases more slowly than it increased at laser heating. Even after elapsed since the laser was switched off, the temperature of nanoparticle solution was about above the reference one. Evidently, the cooling rate depends on the thermal diffusivity of material and sample volume. The characteristic thermal time response of an object is defined by its dimension (the radius for a cylinder form) and the thermal diffusivity of its material :4
Experimental values for the thermal diffusivity of many soft tissues lie within the rather narrow range from (hydrated collagen containing 50% of water) to (pure water). Therefore, the characteristic thermal time response for our experiments with phantoms and tissues according to expression 1 is around . This value is in good agreement with our experimental results presented in Fig 5. Therefore, the real time of object overheating exceeds the laser exposure time. This finding is necessary to take into account to assign a duration of local laser hyperthermia mediated by gold nanoparticles.
In Vivo Near-Infrared Laser Hyperthermia Mediated by Gold Nanoshells
These experimental studies aim the in vivo determination of the spatial distributions of temperature corresponding to dif-ferent depths and concentrations of nanoparticles in tissues of white laboratory rats. The solution of silica/gold nanoshells ( , of gold) was injected subcutaneously or intramuscularly. The diffuse backscattering spectrum of the skin as a scattering medium is influenced by inserted gold nanoparticles due to the specific absorption spectrum of the particles (see Fig. 2). Evidently, the effect of nanoparticles on the backscattering spectrum should be lower as the depth of the layer of the inserted gold nanoparticles increases. Experimental diffuse reflectance spectra measured for rat skin with injected gold nanoparticles (Fig. 6 ) validate the possibility to detect gold nanoparticles within some depth in tissue. These spectra are different from the absorption spectrum of gold nanoparticle solution. The maximum of diffuse reflectance of the skin with gold nanoparticles is mismatched with the absorbance peak of gold colloid. The rise of backscattering intensity (curve 3) nearly for tissue with nanoshells is due to the backscattering of these particles. The absorption peaks [ for particles with size and for nanoshells with size (Fig. 2)] correspond to the minimal values (dips) of diffuse backscattering spectra. This dips are formed because of nanoparticle absorption of scattered radiation in inhomogeneous medium.
In vivo temperature spatial distributions were measured for cw and pulsed laser irradiation and different injection depths of nanoparticles into tissues (Fig. 7 ). It is seen that heating is happening more rapidly and temperature rise is substantially higher at laser irradiation of the skin with inserted nanoparticles than for skin free of particles. The rate of heating is very important for photothermal therapy of cancer. Temperature rise to is accepted as optimal for this purpose.38 Such temperatures admittedly stimulate apoptosis in tumor tissue, in contrast to heating in the range , which is very undesirable.39 In our experiments, by using nanoshells the time interval for tissue heating up to was , and in their absence it was more than tenfold prolonged, up to .
It is interesting to note that both temperature kinetic curves for cw and pulsed irradiation of tissue with nanoparticles have a local maximum at (Fig. 7). The origin of this maximum may be explained by the manifestation of compensatory reaction of organisms to temperature rise, which has a characteristic time of about . With the fast heating caused by nanoparticles, the inertial mechanism of heat regulation could not provide any immediate temperature compensation. In contrast, essentially good compensation is realized for laser heating without inserted particles, and some compensation is seen in a larger time scale for heating with inserted nanoparticles. Also important to note is that laser pulse heating is more controllable in a large time scale and has similar short time kinetics as cw. In that sense, pulsed irradiation could be preferable because it gives fast and self-limited temperature rise. The comparison of curves 2 and 4 (Fig. 7) shows that the increase of effectiveness of hyperthermia is possible by a combination of different heating modes. It is reasonable to use the prime cw mode with a higher power density of laser radiation, and switch over to pulse mode or cw mode with lower power.
At after laser action, biopsy samples of irradiated tissues were taken for histological investigation. Dark-field microscopy was used for imaging of nanoparticles in muscle tissue (Fig. 8 ).
In vivo thermal measurements and histological studies of the tissue before and after laser irradiation were compared. Figure 9 shows the thermorgams and structure of the skin and muscle tissue before and after laser action on tissue with the administrated nanoshells.
In the control samples, we did not observe any pathological changes [Figs. 9a and 9d]. For subcutaneously administrated nanoshells after laser irradiation, epidermis of the skin is partly absent with the formation of vesicles filled up by serous fluid. In the derma, we found severe edema with disorientation of collagen fibers [Fig. 9c]. In the muscle tissue we observed edema, hyperemia, and inflammatory infiltration by leukocytes. The control of the spatial distribution of the temperature of the skin surface using the IR imager showed that after of laser action, the maximal temperature in the center of the laser spot is about [Fig. 9b].
At intramuscular administration of the nanoshells, we revealed moderate edema with defibrillation of collagen fibers in muscle tissue [Fig. 9f]. In the connective tissue, we observed inflammatory leukocyte infiltration and edema. In the case of intramuscular particle introduction, we did not observe significant changes in color or structure of tissue surface during the irradiation time. However, the surface temperature was approximately higher than in the control measurements without nanoparticles. Thus, the laser thermolysis can be achieved in the region of nanoparticle localization without damage of the surface tissues. This conclusion is confirmed by histological examinations.
Figure 10 shows aggregates (clusters) of nanoshells (marked by white arrows) injected immediately in tissue before laser action. There are evident signs of cell damage: the nucleus contains vacuoles (V), the cytoplasmic membrane (CM) is ruinous, and cytoplasm is pouring out into extracellular space. Near the cell there are unstructured necrotic masses (NM), maybe remains of destroyed cells.
Gold nanoshells with maximum extinction in the NIR spectral range are synthesized. Spectral characteristics of design nanoparticles are in good agreement with theoretical ones calculated on the basis of the Mie theory. Parameter verification (size and shape) is executed by TEM. By experimental investigations using a test tube, the dependences of temperature rise of solutions on concentration of nanoparticles are measured, as well as the depth and time interval of laser heating. It is found that at cw laser heating, the time needed to reach the steady-state temperature conditions is . The results of studing laser heating kinetics allow us to optimize clinical protocol for cancer treatments via inducing pathologi-cal cell apoptosis or their photothermolysis by NIR laser irradiation. For instance, both cw and pulse laser heating have similar short time kinetics (temperature rise up to over time no more than ); however, due to compensatory response of the organism, pulsed laser heating is more controllable than cw on a large time scale. In that sense, pulsed irradiation (on-off time ratio of 0.25, and pulse duration of ) is preferable because it gives fast and prolonged self-limited temperature rise. Temperature distributions and corresponding morphological alterations in tissues caused by laser heating at different depths of inserted nanoparticles are analyzed. Gold nanoshells mediated laser destruction of muscle tissue is achieved in vivo in the region of nanoparticle localization (at a depth of ) without any damage to surface tissues. In spite of similar power densities used in pulse and cw mode, the nanoparticle mediated laser heating of tissue is different in these two modes. The combined mode of laser operation for optimization of laser heating could be useful. It is found that the elevated temperature of the object under study is kept for some time after laser action. This finding is necessary to account for assigning duration of local laser/gold nanoparticle hyperthermia in clinical studies. The presented results allow us to estimate the concentration of nanoparticles needed for providing desirable temperature elevation and the potentiality of controllable laser hyperthermia of deep tissue layers without damage to overlying tissue layers. In particular, optimal nanoparticle concentration that provides smooth heating of tissue volume allows for the treatment of tumors with sizes no bigger than . Thus, to provide smooth heating of big tumors at high concentrations of nanoparticles, multibeam irradiation could be effective.
The back reflectance spectral measurements demonstrate the possibility of in vivo visualization of nanoparticle-filled tissue regions within the depth of at least a few millimeters from the skin surface. Therefore, fiber optic diffuse spectroscopy is a promising tool for noninvasive monitoring of selective accumulation of nanoparticles in tumors in clinical studies.
This research was supported by grants of RFBR (numbers 07-02-01434a and 08-02-00399a), RF Program on the Development of High School Potential (number 2.1.1/4989 and 188.8.131.52/2950), the President of the RF for the support of Leading Scientific Schools (number 208.2008.2), FP7-ICT-2007-2 PHOTONICS4LIFE (number 224014). B. Khlebtsov was supported by grants from the President of the Russian Federation (MK) and RFBR (numbers 07-04-00301a, 07-04-00302a, and 09-02-00496-a). N. Khlebtsov was supported by a grant from the Presidium of RAS Program “The Basic Sciences—to Medicine.”