Ultrasonics 54 (2014) 1020–1028

Contents lists available at ScienceDirect

Ultrasonics journal homepage: www.elsevier.com/locate/ultras

Structural and permeability sensitivity of cells to low intensity ultrasound: Infrared and fluorescence evidence in vitro Fabio Domenici a,⇑, Claudia Giliberti b,⇑, Angelico Bedini b, Raffaele Palomba b, Ion Udroiu c, Lucia Di Giambattista a, Deleana Pozzi d, Stefania Morrone e, Federico Bordi a, Agostina Congiu Castellano a,⇑ a

Sapienza University, Physics Department, Rome, Italy INAIL, Italian Workers’ Compensation Authority, Rome, Italy Roma Tre University, Department of Science, Rome, Italy d Sapienza University, Department of Molecular Medicine, Rome, Italy e Sapienza University, Department of Experimental Medicine, Rome, Italy b c

a r t i c l e

i n f o

Article history: Received 7 May 2013 Received in revised form 3 December 2013 Accepted 7 December 2013 Available online 14 December 2013 Keywords: Ultrasound Membrane sonoporation Cell uptake FTIR spectroscopy Fluorescence microscopy

a b s t r a c t This work is focused on the in vitro study of the effects induced by medical ultrasound (US) in murine fibroblast cells (NIH-3T3) at a low-intensity of exposure (spatial peak temporal average intensity Ita < 0.1 W cm2). Conventional 1 MHz and 3 MHz US devices of therapeutic relevance were employed with varying intensity and exposure time parameters. In this framework, upon cells exposure to US, structural changes at the molecular level were evaluated by infrared spectroscopy; alterations in plasma membrane permeability were monitored in terms of uptake efficiency of small cell-impermeable model drug molecules, as measured by fluorescence microscopy and flow cytometry. The results were related to the cell viability and combined with the statistical PCA analysis, confirming that NIH-3T3 cells are sensitive to therapeutic US, mainly at 1 MHz, with time-dependent increases in both efficiency of uptake, recovery of wild-type membrane permeability, and the size of molecules entering 3T3. On the contrary, the exposures from US equipment at 3 MHz show uptakes comparable with untreated samples. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction US is beginning to emerge as a powerful stimulus in numerous medical applications [1], including new methods for localised treatments as gene therapy and drug delivery [2–4]. The challenge is to drive the biological properties of cellular membranes as an efficient way to non-invasively transfer DNA, drugs or other molecules into target cells. These applications take advantage of the pivotal US bioeffect investigated by numerous in vitro experiments, which is commonly referred to as the sonoporation phenomenon [5,6]. In the absence of ultrasonic heating, US-induced bioeffects are commonly assumed to be caused by acoustic cavitation [7,8]. Cavitation is typically generated through the activation of small dissolved gas nuclei in the presence of an acoustic pressure field [9–11]. According to sonoporation studies [5,6], cavitation-promoted plasma membrane fractures can allow the uptake of poorly membrane-permeable exogenous vectors in viable cells, although ⇑ Corresponding authors. Tel.: +39 0649913503/0697893312; fax: +39 0697893304. E-mail addresses: [email protected] (F. Domenici), [email protected] (C. Giliberti), [email protected] (A. Congiu Castellano). 0041-624X/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ultras.2013.12.003

the lack of reversibility of the process results in cell death [12]. Some applications may benefit from killing cells [13], however drug delivery scenarios seek to maximise intracellular uptake while maintaining constant cell viability [14,15]. Although the ability of US to deliver fluorescent molecules into viable cells has been demonstrated in numerous in vitro studies, the mechanism and the quantitative dependence of bioeffects on acoustic parameters remains poorly understood [15–21]. In particular, low-intensities of US exposure (Ita < 0.1 W cm2) have been shown to induce bioeffects in cells without any evidence of inertial or stable cavitation being present [17]. Based on these findings, an intramembrane cavitation model, named bilayer sonophore (BLS), has recently been proposed [18] as a unique hypothesis to explain US-induced bioeffects at both non-cavitation (Ita < 0.1 W cm2) and cavitation (Ita > 0.1 W cm2) regimes. According to this model, the bilayer membrane is capable of absorbing mechanical energy from the US field and transforming it into expansions and contractions of the intramembrane space, overcoming the molecular attraction forces between the bilayer leaflets. This mechanism would induce different effects on the cell membrane, such as an increase of membrane permeability,

F. Domenici et al. / Ultrasonics 54 (2014) 1020–1028

potentially facilitating the uptake of drugs and genes, and the enhancement of tissue permeability [18]. In this framework, microscopic and spectroscopic techniques offer the sensitivity and capability to detect morphological, structural and functional fine alterations of sonicated cells, as they are not yet significantly advanced [22–27]. In previous papers [28,29], we have shown how the effects of US on biological samples can be effectively revealed by means of Fourier Transform Infrared (FTIR) spectroscopy, a non-destructive technique which is able to monitor conformational and functional changes exhibited by specific subsets of macromolecules inside a cell population. Proceeding from this approach, here we combine FTIR spectroscopy, flow cytometry and fluorescent microscopy to investigate the sensitivity of biological cells undergoing therapeutic US in regimes of sub-cavitation. To this aim, we focus on the murine fibroblasts cell culture NIH-3T3 in representing a well characterised in vitro biological model to analyse alterations which may occur at both the structural and plasma membrane permeability level, in the presence of 1 and 3 MHz US at varying field intensities and times of exposure. Within these working conditions, we analyse the intracellular uptake of the fluoroprobe calcein, which is extensively considered a small model drug molecule (radius = 0.6 nm) also in relation to sonoporation, and use these results to better understand the FTIR analysis of lipid, protein and nucleic acids composing the cell. Thus, the question of whether the presence of disruptions in the cell membrane (pore formation) is consistent with the uptake of cellimpermeable molecules has been addressed. 2. Materials and methods 2.1. Cell and culture conditions The experiments were carried out using murine fibroblasts NIH3T3. The cells were cultured as a monolayer in a humidified atmosphere with 95% air and 5% CO2 at 37 °C in Dulbecco’s modified Eagle’s medium (DMEM, Sigma–Aldrich, St. Louis, MO) with 10% foetal bovine serum, 1% penicillin and 1% glutamine/streptomycin. Samples were prepared in well plates (FalconÒ Easy Grip™ tissue culture dish, 35  10 mm) at concentrations varying between 7  105 cells/ml and 9  105 cells/ml. Three millilitres of Phosphate Buffered Saline (PBS) Dulbecco’s Formula without calcium and magnesium were used in each experiment. Cell viability of both the untreated (control) and the US treated (sonicated) cell samples has been determined by the Trypan blue exclusion test, for all of the transducers-samples distances and at all times of exposures. Sample viability, before US exposure, was better than 95% for each trial.

monitored by thermocouple system (Lutron electronic enterprise co., LTD.) both inside and outside the Petri dish. A cell culture-treated Petri dish (9.6 cm2) containing three millilitres of PBS solution hermetically lidded was positioned at the water surface and inserted to half of its thickness in the waterbath, in line with the transducer, according to the scheme reported in Fig. 1. Taking into account therapeutic applications in which the position of the US source is fixed and the time of exposure changed at a selected nominal power and duty cycle, the biological samples were sonicated at three sample-transducer distances, called Source-dish Surface Distance (SSD) (see Fig. 1): 5, 10, 15 cm, and at each distance for 5, 15, 30, 45 and 60 min. The SSD distance was varied by decreasing/increasing the water level of the tank and in turn the position of the Petri dish. The characterisation of the acoustic field produced by the US sources at the two frequencies was performed using a needle hydrophone (Precision Acoustics) of 1 mm diameter (S.N. 1470) with a sensitivity of 1670.4 mV/MPa at 1 MHz and 958.2 mV/MPa at 3 MHz (±14%). In this work, the intensity of the acoustic field is provided in terms of Spatial Peak Temporal Peak (Itp), which represents the higher intensity value measured when the pulse is on and therefore it is particularly significant to correlate mechanical bioeffects induced by pulsed US; on the other hand, Ita represents the maximum spatial intensity measured when the pulse is on, mediated for the period of pulse repetition. Itp values transmitted through the cell monolayer as measured nearby the cell culture plate for 1 MHz US exposures are 0.11 W cm2 (SSD = 5 cm), 0.12 (SSD = 10 cm) and 0.09 W cm2 (SSD = 15 cm); for 3 MHz they are 0.06 W cm2 (SSD = 5 cm), 0.04 W cm2 (SSD = 10 cm) and 0.01 W cm2 (SSD = 15 cm). Because of the frequency dependence of US transmission coefficient, the measured values Itp at 3 MHz fall below those corresponding to 1 MHz, although the nominal intensity of the US generators at 3 MHz and 1 MHz was set at 100% and 75% of Imax, respectively. In terms of Ita, the measurements provide values Ita < 0.026 W cm2, which fall significantly below the recommended threshold of 0.1 W cm2 in the sub-cavitation regime. 2.3. FTIR spectroscopy The spectroscopic measurements were performed with a Jasco spectrophotometer FT-IR 410 in transmission mode; for each SSD and exposure time, cells were grown to confluence on a window

2.2. Ultrasound exposure system and experimental protocol For the US exposures, we used two conventional medical devices (Nuova Elettronica, Italy), consisting of two submersible piezoceramic circular transducers (6 cm diameter) tuned at 1 and 3 MHz, and the corresponding waveform generators. Such systems can work in continuous and pulsed modes; the duty cycle of both 1 MHz and 3 MHz US generators has been selected in High pulsed mode, in which the signal is delivered for 750 ms, followed by a pause of 250 ms. The power can vary in the range from 10% to 100% of the maximum power (maximum nominal intensity Imax = 2.5 W cm2), allowing for selection of the maximum value maintains the integrity of the monolayer after sonication: 100% of Imax for 3 MHz, 75% of Imax for 1 MHz. Ultrasonic transducers were placed at the bottom of a tank filled with degassed water; the temperature of the waterbath was kept constant at 25 °C as

1021

Fig. 1. Experimental set up for US exposures.

1022

F. Domenici et al. / Ultrasonics 54 (2014) 1020–1028

of CaF2 (diameter 2.5 cm) previously treated with a polylysine substrate in order to promote cell attachment. Cells on the CaF2 window were sonicated, washed in Dulbecco’s PBS and dried in a desiccator, then spectra were collected. For each trial, a control spectrum contemporary with sonicated samples spectra was acquired at 4 cm1 resolution and 64 interferograms. The IR spectra of the treated and untreated cells were acquired in the range of 4000–900 cm1, identifying the region of lipids (3000– 2800 cm1), Amide I and Amide II (1750–1480 cm1) and the region of nucleic acids (1300–950 cm1). All spectra were corrected for baseline, normalised intensity of Amide I and processed using the software OPUS 5.0 (Bruker Optik). For each sample, the average spectrum of three experiments was analysed.

2.4. Optical fluorescence microscopy For fluorescence measurements an inverted optical microscope Leica DM IL (Obj 10; 20; 40) equipped with Hg vapour lamp and filters DAPI/FITC/TRITC was used. The morphological changes in cellular samples exposed to US were captured by an Olympus digital camera for the acquisition of frames. The cells were incubated in Petri dishes until confluence and the sonications were performed in PBS: immediately prior to US application, a solution of a fluorescent dye, calcein (Molecular Probes, Eugene, OR, wavelengths of the maximum absorption and emission peak respective at 494 nm and 514 nm) was added to the wells to attain a final concentration of 10 lM [15]. Calcein is a green fluorescent molecule that cannot cross intact cell membranes. At the end of the sonication, in agreement with the literature [4], a recovery of 100 was scheduled, and then washes with PBS to eliminate calcein that had not been internalised were performed. Non-viable cells were identified by the uptake of propidium iodide (PI, 20 ll, MW = 668.4 Da and wavelength of maximum absorption peak and emission respective 535 nm and 617 nm), a probe that can be internalised through damaged plasma membranes and is intercalated into hydrophobic space of cellular DNA, emitting fluorescence. It was added immediately prior to analysis. Measurements using the fluoroprobes 40 kDa and 70 kDa FITCdextran were performed by a Zeiss camera and ZEN 2011 software (see for more detail in Supplementary Data).

2.6. Statistical analysis All experiments were performed in triplicate. Results are reported and displayed as mean ± standard deviation. Tests of significance were performed using one-way analysis of variance (ANOVA p < 0.05). The IR spectra were processed by taking the second derivative of the second Savitzky–Golay algorithm with 9 points smoothing and normalised intensity Amide I. The statistical analysis of spectroscopic data was carried out by the Principal components analysis (PCA) [26,30,31], a powerful data-mining technique that generates a new set of orthogonal variables: the principal components (PCs). Based on PCA, the higher the spatial separation between points in a score plot, the higher the level of dissimilarity in the spectroscopic parameter considered. 3. Results and discussion In this section, we report on potential cytotoxic and molecular structure effects of US exposure on 3T3 cells by using the viability assay and FTIR analysis, respectively. Their influence on cell permeability is then described in terms of efficiency of cellular uptake of the membrane-impermeable fluoroprobe calcein by Fluorescence and FACS techniques, and discussed by defining a US healthy therapeutic indicator. 3.1. Viability assay The Trypan blue viability test, applied to treated samples (1 MHz and 3 MHz US) at the three distances SSD = 5, 10, 15 cm and overall exposure times, was compared to the test applied to control sample; the results are reported in Fig. 2a and b for 1 MHz and 3 MHz, respectively. As shown in Fig. 2, all sonicated samples have a lower viability than the control, but none of the differences were statistically significant. Moreover, for exposure to

2.5. Flow cytometry Flow cytometry was performed using Fluorescence Activated Cell Sorting (FACS) cytometer FACSCalibur (BD Biosciences) equipped with two excitation lasers (argon k = 488 nm and visible red diode laser 635 nm); 10,000 events were recorded for each sample. Sample preparation and treatment steps were performed according to the fluorescent microscopy procedure and PI (10 ll) was used to identify non-viable cells. At the end of each sonication, the samples were detached from Petri dishes, centrifuged (1800 rpm) in PBS buffer, collected in tubes suitable for cytometry and labelled with PI, which was added just before analysis. Flow cytometry was used to detect the number of cells stained with calcein probe, indicating cell permeabilisation, and cells stained with PI, indicating necrotic cell. Cells emitting a fluorescent intensity higher than those of sham control were classified as permeabilised. Results were expressed in percentage of positive cells (% gated), using CellQuest software (Becton Dickinson). Of practical significance is the ratio of cells in which permeability increases when induced by US to those that are killed by the same ultrasound exposure; this is defined the ‘‘therapeutic ratio’’ (TR) for sonoporation [18].

Fig. 2. Bar charts of viability of NIH-3T3 cells treated with ultrasound at 1 MHz (a) and 3 MHz (b) on varying of distance and exposure time determined by Trypan blue exclusion test. The data represents the mean percentage ± standard deviation.

F. Domenici et al. / Ultrasonics 54 (2014) 1020–1028

1 MHz, a statistically significant negative correlation between the exposure time and the viability was shown at SSD = 5 cm (r = 0.9354, P = 0.0061) and for exposure to 3 MHz, at SSD = 10 cm (r = 0.8621, P = 0.0272). 3.2. FTIR analysis of cellular biochemical structure We studied the spectrum of the control sample in the wavenumber range from 4000 to 900 cm1 (Fig. 3); the assignment of the main absorption peaks found in the IR spectra is reported in Table 1 [28]. In order to detect structural changes in major biological molecules following cell exposure to 1 and 3 MHz US, the Ri spectral parameters were defined, according to the definitions reported in our previous paper [26]; thereby, the i index becomes L1 and L2 (RLi), P1, P2, P3, P4 (RPi) and finally A for Amide I (RA), with reference to the same bands in Table 1. By means of these indicators, the area of corresponding bands (Table 1) in sonicated and control samples can be compared, bearing in mind that the unit value of Ri means that the corresponding band area in a sonicated sample is comparatively unchanged. In particular, RL1 (defined as the ratio between the methyl antisymmetric stretching band areas of a sonicated sample and the corresponding control) may be a useful spectroscopic indicator in monitoring changes at the level of lipid packages, composition, and phase transitions of plasma membrane structure, as may actually be observed on cells in response to mechanical stressors like US [26,32–34]. The trends of the RL1 parameter vs. exposure times for different SSDs, obtained exposing samples to 1 MHz and 3 MHz US, are reported in Fig. 4. The PCA score plot, examined in parallel with the corresponding loading plots reported in Fig. 5, reveals that significant changes in the structures of membrane phospholipids may occur for 1 MHz US, at SSD = 5 cm for the sonication time of 600 and at SSD = 15 cm for t = 450 . In this line, for both 1 MHz and 3 MHz setups, the exposure configurations enlightened by PCA analysis reflecting significant changes of the Ri spectral parameters value are listed in Table 2. Notably, as shown in Table 2, the exposure to 1 MHz US equipment produces a comparatively greater number of configurations in terms of conformational changes of functional groups for each spectral region (compare 1st and 2nd row of Table 2, for 1 MHz and 3 MHz, respectively), and the corresponding Itp values

1023

(reported in Section 2.2) of 1 MHz US setup are comparatively lower accordingly. For each selected time of exposure, the 3 MHz US setup is thus predicted to transfer comparatively lower energy on 3T3 cells. As the US pulse moves through matter, it continuously loses energy [35,36]. This is generally referred to as attenuation. Among several factors which contribute to this reduction in energy (i.e. scattering and refraction interactions), absorption of the US energy by the material is undoubtedly the most significant. For a given biosystem through which it is passing, the attenuation (the rate of absorption) is directly proportional to the US frequency [36]. Hence, according to our measurements, at a given nominal intensity setting, by increasing US frequency from 1 MHz to 3 MHz, a corresponding reduction of pulse amplitude, and in turn, pressure affecting biochemical structures composing the cell can reasonably be expected. Nevertheless, other causes also directly related to the frequency employed cannot be ruled out here. However, concerning the severity of the significant effects which may occur for a selected US frequency, it may not simply be expected to increase with the ‘‘dose’’ of US exposure. In this respect, according to the analysis of the RL1 parameter shown in Figs. 4a, 5a and b, the samples treated at 15 cm, 450 and 1 MHz seem to exhibit a slightly more severe effect on lipid structure than that of 5 cm, 600 and 1 MHz, although the latter is considered a more feeble exposure condition in terms of both intensity and exposure time. A plausible, perhaps simplistic, hypothesis which could be helpful to resolve the apparent contradiction may concern the loss of isolated cells that are severely damaged by the US field due to their detachment from the IR substrate. Thereby, the RL1 values we reported may be underestimated, although the integrity of 3T3 cells onto the polylysine coated CaF2 substrate was optically assessed upon each US treatment. On the other hand, according to the BLS elastic model [18] mentioned in Section 1, it can also be determined that varying both intensity and time of US exposure may result in energy-dependent regime of oscillations. Thus, allowing for increasing membrane leaflet deformation and membrane detachment, the latter occurred when the tension capable of causing lipid bilayers to rupture was reached [18]. In this regard, it can reasonably be inferred that several cells may exhibit significant changes in RL1 parameter, reflecting some structural alteration (i.e. involving membrane lipid tail package, fluidity level, lipid composition), before/without

Fig. 3. Control spectrum of NIH-3T3 cells with the absorption bands.

1024

F. Domenici et al. / Ultrasonics 54 (2014) 1020–1028

Table 1 Peak labels, peak wavenumbers and relative assignment of the main IR absorption bands of cells. Label

Peak (cm1)

Assignment

L1 L2 L3 L4 A1 A2

2960 2924 2872 2852 1646 1541

P1

1454

P2 P3

1399 1244

P4

1085

P5

967

ACH3 asymmetric stretching, lipids, proteins ACH2 asymmetric stretching, lipids, proteins ACH3 symmetric stretching, lipids, proteins ACH2 symmetric stretching, lipids, proteins Amide I (AC@O stretching), proteins Amide II (ANAH bending, ACAN stretching), proteins ACH2 scissoring/ACH3 asymmetric bending, proteins, lipids ACOO symmetric stretching, proteins, lipids APO 2 asymmetric stretching, nucleic acids, phospholipids APO 2 symmetric stretching, nucleic acids, phospholipids APO 4 symmetric stretching, nucleic acids

sonopore effects take place. At the same time, in the fashion of Landau theory [37,38], where the pore is effectively formed, the membrane tension of the intact portions is expected to relax again. In other words, the ensemble of treated cells exhibiting comparatively lower levels of calcein uptake and in turn, lower sonopore formation, may provide higher structural membrane perturbation in terms of RL1, which is in agreement to that observed in Section 3.3 (Table 3, compare the uptake level of 15 cm, 450 and 5 cm 600 of 1 MHz US setup). By bearing these aspects in mind, it is not surprising that more severe exposure conditions in terms of US mechanical energy on lipid membranes may provide quite lower effects in terms of R1L parameter. On this line, even if the interpretation of FTIR results is quite complex and undoubtedly deserves further dedicated studies, we are confident that disclosure of changes at the lipid level by FTIR spectroscopy may be particularly useful when supported by the fluorescence combined approach. 3.3. Fluorescence and FACS evidences of cellular uptake The effects of alterations of membrane permeability induced by 1 and 3 MHz US, under the same exposure conditions specified above, have been investigated in greater detail with more selective microscopic techniques such as fluorescence microscopy and FACS using the fluorescent probe calcein. The flow cytometry allows cells labelled with the fluorophore calcein to be quantified. The results for 1 MHz US setup, expressed in percentage of positive cells (% gated) and reported in Table 3,

show a high calcein uptake in the sonicated samples (84–99%) compared to the control (2%), with a saturation effect for the internalisation of the fluorophore. As reported in Section 2.2 for 1 MHz device, although SSD distances varied from 5 cm to 15 cm, only slight changes in the corresponding measured Itp values were found, and the efficiency of uptake was substantially independent from the intensity accordingly. However, Table 3 also shows that at a fixed SSD the uptake efficiency is time-dependent. For instance, at SSD of 10 cm, it increases on varying exposition times from 300 to 600 . According to the sonoporation hypothesis [16,18], this could mean that the longer the time of exposure, the higher the probability of finding cells with membrane injury that are large enough to allow calcein to enter the cell. On the contrary, for US setup operating at 3 MHz, only low efficiency of calcein uptake (comparable with controls) has been measured. As discussed in Section 3.2, the higher coefficient of attenuation of US at 3 MHz (0.0195 dB/cm, water) than that at 1 MHz (0.0022 dB/cm, water) mainly accounts for the large gap between the major and minor Itp values measured when working with 3 MHz and 1 MHz equipment, respectively. This means that less energy is expected to be transmitted through the cell monolayer when working with 3 MHz, when all other conditions are equal [35,36]. The experimental configuration at 1 MHz frequency, SSD = 15 cm and t = 450 , being spectroscopically active and characterised by high level calcein uptake, has also been investigated by fluorescence microscopy. As expected, the representative images shown in Fig. 6b and c reveal that exposure to 1 MHz US promote the internalisation of the fluorescent probe calcein uniformly distributed in the cells in accordance with the literature [10]. Analogous to FACS results, when the 3 MHz setup is employed, only low calcein uptake levels can be observed (see Fig. 6d and e). As reported in Section 2.4 in order to allow the membrane repair response [4] once the US field is stopped, a recovery time of 100 was scheduled before rising to eliminate calcein that had not been internalised. As a control, we repeated the FACS experiment at SSD = 15 cm and 1 MHz set up, rinsing the Petri dishes of the calcein solution immediately after stopping the US field. Interestingly, in this case, we measured that the uptake efficiency of 3T3 was almost unchanged after 300 of exposure (a slight drop of less than 10 was detected), whereas after 600 of exposure, the rinse caused the loss of about half of the efficiency, together with a similar apparent increase in dead cells on the basis of PI probe. In this respect, it may be hypothesised that for low US exposure times, membrane embrasures can be formed, and, similar to that which is seen for transient pores [37,39], they reseal within a few seconds as the mechanical tension of the membrane relaxes; on the

Fig. 4. Changes in the RL1 parameter (relative to the lipid region) versus sonication times, pointed out according to the different SSDs employed at 1 MHz (a) and 3 MHz (b) US setup.

1025

F. Domenici et al. / Ultrasonics 54 (2014) 1020–1028

Fig. 5. PCA analysis (score plots and loading plots) applied to the second derivative of the IR spectra in the lipids region, at the sample-transducer distances of SSD = 5 cm (a), and SSD = 15 cm (b) for all the sonication times.

Table 2 Specific configurations of sample transducers distance (SSD) and exposure time (t), at which the two US frequencies of 1 MHz and 3 MHz have shown significant changes of the spectroscopic parameters as revealed by PCA analysis. Lipid region

Nucleic acid region

Amide I region

(3000–2800 cm1) RL1,2 L1,2sonicated/LCTRL

(1500–1300 cm1) RP1,2 P1,2sonicated/P1,2 CTRL

(1300–950 cm1) RP3,4 P3,4sonicated/P3,4CTRL

(1640 cm1) RA Rsonicated/RCTRL

1 MHz

SSD = 5 cm t = 600 SSD = 15 cm t = 450

SSD = 5 cm t = 600 SSD = 10 cm t = 300 SSD = 15 cm t > 300

SSD = 5 cm t = 150 SSD = 10 cm t = 300 , 60 SSD = 15 cm t = 450 , 600

SSD = 5 cm t = 150 SSD = 15 cm t = 150 , 300 , 600

3 MHz

SSD = 10 cm t = 150 , 600

SSD = 10 cm t = 300 , 600

SSD = 5 cm t = 600

–

Table 3 Results of flow cytometry measurements performed for 1 MHz US exposures. % Gated – 1 MHz 75% high SSD = 5 cm Itp = 0.11 W/cm2 CTRL 2.63 ± 1.1 t = 300 74.8 ± 2.9 t = 450 84.1 ± 4.1 t = 600 97.5 ± 0.9

SSD = 10 cm Itp = 0.12 W/cm2

SSD = 15 cm Itp = 0.09 W/cm2

49.4 ± 1.6 86.8 ± 1.5 99.3 ± 0.5

64.6 ± 1.3 87.0 ± 1.7 99.5 ± 0.5

contrary, by increasing the exposure time at a fixed intensity, and in turn, the amount of energy transferred in plasma membrane of 3T3 cells [16], the embrasure may progressively expand, allowing for calcein leakage as well as PI probe penetration into 3T3 cells, until the kinetics of membrane repair [4] takes place.

To analyse whether the integrity of 3T3 membranes was instead recovered effectively 100 after the end of the US treatment, and also upon the calcein fluoroprobe entering the 3T3 cell (see Fig. 6), an additional PI test was performed by fluorescence microscopy. As shown in Fig. 7 for SSD = 15 cm, t = 450 and 1 MHz (also representative of several different images obtained at the different 1 MHz US setup employed), there is sporadic evidence of both PI and calcein co-localised within a single 3T3 cell. Therefore, we can assert that the US-induced membrane permeability alteration allowing for the internalisation of calcein in 3T3 cells can be repaired and the cell can retain calcein and viability. However, we should stress that the recovery of wild-type membrane permeability discussed here does not mean that the wild-type structure was reached. As mentioned in Section 3.2, when US insult is considered on an equal footing of mechanical stress of plasma membrane, the bioeffect could involve other membrane features, such as physical state, biochemical

1026

F. Domenici et al. / Ultrasonics 54 (2014) 1020–1028

Fig. 6. FITC fluorescence together with low intensity transmitted phase contrast microscopy images of the monolayer of NIH-3T3 cells exposed to 1 MHz (b and c) and 3 MHz (d and e) US setup at SSD = 15 cm and t = 450 in presence of calcein 10 lM. For 1 MHz exposure: sham control (a), high fluorescence uptake of calcein (10 Obj) (b), and internalisation of the fluorophore in the cell (40 Obj) (c). For 3 MHz exposure: comparatively very low uptake of calcein (20 Obj) (d) and almost whole absence of internalisation of calcein (40 Obj) (e). Bright light and FITC fluorescence modality were used simultaneously in order to clearly allocate fluorescent and non-fluorescent structures.

Fig. 7. Transmitted light (a) and epiluminescence (b) microscopy details of a NIH-3T3 cell culture treated with PI upon exposure to US (1 MHz, SSD = 15 cm, 450 ) together with calcein. Pass band FITC emission filter was used to simultaneously capture calcein and PI emission band. Image was captured at 40 of magnification.

composition, fragility and mechanochemical signalling [32– 34,39–42] whose reversibility has not been clarified to date. Finally, the optimisation of the uptake process provides a balance between the desired effect of the US exposure (uptake) and the destructive effect (cells death); within this purpose, a ‘‘therapeutic ratio’’ has been defined (TR, Fig. 8) [19]. The results show that the TR value depends on the exposure conditions; for the experimental configurations investigated, the maximum of the TR was obtained for SSD = 10 cm t = 600 , in consideration of a high membrane permeability induced by exposure (99.27%) and reduced mortality of samples (25%), while the minimum is observed for SSD = 5 cm. The present results indicate that calcein is able to efficiently enter fibroblast cells demonstrating the actual possibility of operating US-mediated in vitro drug delivery at 1 MHz under a very low intensity of exposure. It is also worth noting that our findings, which were performed in sub-cavitating conditions (see Section 1), can be reasonably accounted for by the BLS model [18], in which the dynamic behaviour of the lipid membrane undergoing US is depicted. Simulations based on this model have in fact shown that

biological membranes exposed to 1 MHz can suffer a dynamic response capable of raising the elastic tension of the bilayer lipid membrane, close to the values required for the establishment of hydrophilic pores in the membrane [18]. According to the FTIR outcomes, the spectroscopic and microscopic results suggest that in sub-cavitation conditions at 1 MHz, pores are induced in the membrane, which constitute favourable conditions for the internalisation of the fluorophore. Moreover, according to our findings, once the sonopore is formed, the increased dimension of membrane pores should be assessed as the US-mediated energy transfer on plasma membrane proceeds [4,16,43]. In other words, the increase in exposure time of US may allow the delivery of molecules with higher molecular weights. In order to provide experimental evidences in favour of sizedependent uptake on 3T3 cells, we undertook a further fluorescence experiment in which 3T3 cells were exposed to 1 MHz US at fixed SSD of 10 cm; the time of exposure was selected as 300 and 600 in the presence of 10 lM of 40 kDa and 70 kDa FITC-labelled dextrans, respectively. Subsequently assays were left for

F. Domenici et al. / Ultrasonics 54 (2014) 1020–1028

1027

One of the main challenges in biomedicine is the improvement of clinical benefit, which can be obtained by US; the latter is currently being developed for cancer, cardiovascular and bloodbrain-barrier-limited treatments. In this framework, our results may provide useful for designing safe acoustic-targeted drug delivery methodologies. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ultras.2013. 12.003.

Fig. 8. Bar chart of Therapeutic Ratio (TR) (defined as the ratio of the number of fluorescent cells to the number of dead cells) as a function of the sonication time in the configurations investigated (1 MHz US). The data represents the mean percentage ± standard deviation.

100 after stopping the US treatment to allow the recovery of equilibrium. Once the Petri dish was carefully rinsed in PBS buffer we analysed several different areas using a total magnification of 100; the representative pictures are provided as Supplementary Data. The results clearly indicate that after 300 of exposition, a significant number of cells can internalise 40 kDa, but without evidence of the internalisation for 70 kDa dextrans (see Fig. S1a of Supplementary Data). After 600 of exposure, a significant internalisation of 70 kDa dextrans can occur (see Fig. S1b of Supplementary Data). In the literature, it has been reported that US and microbubble-targeted delivery of macromolecules such as dextrans in endothelial cells is regulated by the induction of endocytosis and pore formation [44,45]. After ATP depletion, reduced uptake of 4.4 kDa dextran and no uptake of 500 kDa dextran was reported [45]. Accordingly, our results show a clear size-dependent uptake thus suggesting that sonoporation represents the important transport mechanism even working below the threshold of US intensity necessary to initiate cavitation. However, beside sonoporation, we cannot exclude that further and concomitant endocytic pathways [41] can be altered when a specific physical stimulus occurs (i.e. mechanical forces induced by US) to promote the targeted delivery of peptides, DNA and other biotherapeutic compounds across cell membranes. In this respect, much work still has to be done towards a more comprehensive knowledge on mechanisms of membrane transport which are altered by very low ultrasound intensity.

4. Conclusions Changes in the molecular structure and permeability of murine fibroblasts in vitro induced by US using 1 MHz and 3 MHz equipment of therapeutic relevance have been described here under sub-cavitation regimes, using methods combining FTIR spectroscopy, flow cytometry and fluorescent microscopy. Our findings are consistent with previous reports on the occurrence of the intramembrane cavitation model, but go beyond these by disclosing that the calcein fluorophore uptake appears to be particularly efficient for sonication of the samples exposed to US setup at the frequency of 1 MHz, in which structural and functional changes may also be induced. Time-dependent increases in efficiency of uptake, recovery of membrane permeability, and size of molecules able to enter 3T3 has been enlightened here by working well below the threshold intensity required to initiate cavitation, and explained in terms of the elastic BLS model which was recently proposed.

References [1] P.A. Lewin, Quo vadis medical ultrasound?, Ultrasonics 42 (2004) 1–7 [2] C.M.H. Newman, T. Bettinger, Gene therapy progress and prospects: ultrasound for gene transfer, Gene Ther. 14 (2007) 465–475. [3] Y. Liu, J. Yan, M.R. Prausnitz, Can ultrasound enable efficient intracellular uptake of molecules? A retrospective literature review and analysis, Ultrasound Med. Biol. 38 (2012) 876–888. [4] R.K. Schlicher, H. Radhakrishna, T.P. Tolentino, R.P. Apkarian, V. Zarnitsyn, M.R. Prausnitz, Mechanism of intracellular delivery by acoustic cavitation, Ultrasound Med. Biol. 32 (2006) 915–924. [5] K. Hynynen, Ultrasound for drug and gene delivery to the brain, Adv. Drug. Deliv. Rev. 60 (2008) 1209–1217. [6] I.V. Larina, B.M. Evers, R.O. Esenaliev, Optimal drug and gene delivery in cancer cells by ultrasound-induced cavitation, Anticancer Res. 25 (2005) 149–156. [7] B.E. O’Neill, K.C.P. Li, Augmentation of targeted delivery with pulsed high intensity focused ultrasound, Int. J. Hyperther. 24 (2008) 506–520. [8] D.L. Miller, S.V. Pislaru, J.E. Greenleaf, Sonoporation: mechanical DNA delivery by ultrasonic cavitation, Somat. Cell. Mol. Genet. 27 (2002) 115–134. [9] T. Leighton, The Acoustic Bubble, Academic Press, New York, 1994. [10] K. Milowska, T. Gabryelak, Reactive oxygen species and DNA damage after ultrasound exposure, Biomol. Eng. 24 (2007) 263–267. [11] D. Pozzi, P. Fattibene, D. Viscomi, L. Di Giambattista, P. Grimaldi, I. Udroiu, A. Bedini, C. Giliberti, R. Palomba, A. Congiu Castellano, Use of EPR and FTIR to detect biological effects of ultrasound and microbubbles on a fibroblast cell line, Eur. Biophys. J. 40 (2011) 1115–1120. [12] D.M. Hallow, A.D. Mahajan, M.R. Prausnitz, Ultrasonically targeted delivery into endothelial and smooth muscle cells in ex vivo arteries, J. Control. Release 118 (2007) 285–293. [13] M. Ward, J. Wu, J.F. Chiu, Experimental study of the effects of optison concentration on sonoporation in vitro, Ultrasound Med. Biol. 26 (2000) 1169– 1175. [14] C.X. Deng, F. Sieling, H. Pan, J. Cui, Ultrasound-induced cell membrane porosity, Ultrasound Med. Biol. 30 (2004) 519–526. [15] H.R. Guzmán, D.X. Nguyen, A.J. McNamara, M.R. Prausnitz, Equilibrium loading cells with macromolecules by ultrasound: effect of molecular size and acoustic energy, J. Pharm. Sci. 91 (2002) 1693–2002. [16] H.R. Guzman, D.X. Nguyen, S. Khan, M.R. Prausnitza, Ultrasound-mediated disruption of cell membranes. II. Heterogeneous effects on cells, J. Acoust. Soc. Am. 110 (2001) 597–606. [17] W.D. O’Brien, Ultrasound-biophysics mechanisms, Prog. Biophys. Mol. Biol. 93 (2007) 212–255. [18] B. Krasovitski, V. Frenkel, S. Shoham, E. Kimmel, Intramembrane cavitation as a unifying mechanism for ultrasound-induced bioeffects, Proc. Natl. Acad. Sci. 108 (2011) 3258–3263. [19] R. Karshafian, P.D. Bevan, R. Williams, S. Samac, P.N. Burns, Sonoporation by ultrasound-activated microbubble contrast agents: effect of acoustic exposure parameters on cell membrane permeability and cell viability, Ultrasound Med. Biol. 35 (2009) 847–860. [20] R. Karshafian, S. Samac, P.D. Bevan, P.N. Burns, Microbubble mediated sonoporation of cells in suspension: clonogenic viability and influence of molecular size on uptake, Ultrasonics 50 (2010) 691–697. [21] M. Kinoshita, K. Hynynen, Key factors that affect sonoporation efficiency in vitro settings: the importance of standing wave in sonoporation, Biochem. Biophys. Res. Commun. 359 (2007) 860–865. [22] D.L. Miller, WFUMB safety symposium on echo-contrast agents: in vitro bioeffects, Ultrasound Med. Biol. 33 (2007) 197–204. [23] M.M. Forbes, R.L. Steinberg, W.D. O’Brien Jr., Examination of inertial cavitation of optison in producing sonoporation of chinese hamster ovary cells, Ultrasound Med. Biol. 34 (2008) 2009–2018. [24] K. Hensel, M.P. Mienkina, G. Schmitz, Analysis of ultrasound fields in cell culture wells for in vitro ultrasound therapy experiments, Ultrasound Med. Biol. 37 (2011) 2105–2115. [25] M. Diem, L. Chiriboga, H. Yee, Infrared spectroscopy of human cells and tissue. VIII. Strategies for analysis of infrared tissue mapping data and applications to liver tissue, Biopolymers 57 (2000) 282–290. [26] P. Grimaldi, L. Di Giambattista, S. Giordani, I. Udroiu, D. Pozzi, S. Gaudenzi, A. Bedini, C. Giliberti, R. Palomba, A. Congiu Castellano, Ultrasound-mediated

1028

[27]

[28]

[29]

[30]

[31]

[32] [33]

[34]

[35]

F. Domenici et al. / Ultrasonics 54 (2014) 1020–1028 structural changes in cells revealed by FTIR spectroscopy: a contribution to the optimization of gene and drug delivery, Spectrochim. Acta Part A 84 (2011) 74–85. J. Kong, S. Yu, Fourier transform infrared spectroscopic analysis of protein secondary Structures, Acta Biochim. Biophys. Sin. (Shanghai) 39 (2007) 549– 559. L. Conti, P. Grimaldi, I. Udroiu, A. Bedini, C. Giliberti, L. Giuliani, R. Palomba, A. Congiu Castellano, Effects induced in cells by ultrasound revealed by ATR–FTIR spectroscopy, Vib. Spectrosc. 52 (2010) 79–84. L. Di Giambattista, P. Grimaldi, I. Udroiu, D. Pozzi, G. Cinque, M.D. Frogley, A.M. Cassarà, A. Bedini, C. Giliberti, R. Palomba, L. Giuliani, S. Buogo, A. Giansanti, A. Congiu Castellano, The ultrasound effects on non tumoral cell line at 1 MHz therapeutic frequency, J. Phys.: Conf. Ser. (JPCS) 279 (2011) 1–6. A. Bombalska, M. Mularczyk-Oliwa, M. Kwasny, M. Wlodarski, M. Kaliszewski, K. Kopczynskia, M. Szpakowska, E.A. Trafny, Classification of the biological material with use of FTIR spectroscopy and statistical analysis, Spectrochim. Acta Part A 78 (2011) 1221–1226. E. Robotti, M. Bobba, A. Panepinto, E. Marengo, Monitoring of the surface of paper samples exposed to UV light by ATR-FT-IR spectroscopy and use of multivariate control charts, Anal. Bioanal. Chem. 388 (2007) 1249–1263. D.C. Lee, D. Chapman, Infrared spectroscopic studies of biomembranes and model membranes, Biosci. Rep. 6 (1986) 235–256. M.A. Haidekker, N. L’Heureux, J.A. Frangos, Fluid shear stress increases membrane fluidity in endothelial cells: a study with DCVJ fluorescence, Am. J. Physiol. Heart Circ. Physiol. 278 (2000) 1401–1406. D. Berchtold, M. Piccolis, N. Chiaruttini, I. Riezman, H. Riezman, A. Roux, T.C. Walther, R. Loewith, Plasma membrane stress induces relocalization of Slm proteins and activation of TORC2 to promote sphingolipid synthesis, Nat. Cell. Biol. 14 (2012) 542–547. S. Umchid, Frequency dependent ultrasonic attenuation coefficient measurement, in: The 3rd International Symposium on Biomedical Engineering (ISBME), 2008, pp. 234–238.

[36] P. Sprawls, Physical Principles of Medical Imaging, second ed., Medical Physics Publishing Corporation, 1995. [37] O. Sandre, L. Moreaux, F. Brochard-Wyart, Dynamics of transient pores in stretched vesicles, Proc. Natl. Acad. Sci. USA 96 (1999) 10591–10596. [38] M.-T. Lee, W.-C. Hung, F.-Y. Chen, H.W. Huang, Mechanism and kinetics of pore formation in membranes by water-soluble amphipathic peptides, Proc. Natl. Acad. Sci. USA 105 (2008) 5089–5092. [39] E. Karatekin, O. Sandre, H. Guitouni, N. Borghi, P.-H. Puech, F. Brochard-Wyart, Cascades of transient pores in giant vesicles: line tension and transport, Biophys. J. 84 (2003) 1734–1749. [40] B. Sinha, D. Köster, R. Ruez, P. Gonnord, M. Bastiani, D. Abankwa, R.V. Stan, G. Butler-Browne, B. Vedie, L. Johannes, N. Morone, R.G. Parton, G. Raposo, P. Sens, C. Lamaze, P. Nassoy, Cells respond to mechanical stress by rapid disassembly of caveolae, Cell 144 (2011) 402–413. [41] V. Lionetti, A. Fittipaldi, S. Agostini, M. Giacca, F.A. Recchia, E. Picano, Enhanced caveolae-mediated endocytosis by diagnostic ultrasound in vitro, Ultrasound Med. Biol. 35 (2009) 136–143. [42] M. Chachisvilis, Y.-L. Zhang, J.A. Frangos, G protein-coupled receptors sense fluid shear stress in endothelial cells, Proc. Natl. Acad. Sci. USA 103 (2006) 15463–15468. [43] E. Gelderblom, F. Wolbers, N. De Jong, A. Van den Berg, M. Versluis, Timeresolved high-speed fluorescence imaging of bubble-induced sonoporation, POMA 19 (2013) 1–6. [44] T. Kodama, M.R. Hamblin, A.G. Doukas, Cytoplasmic molecular delivery with shock waves: importance of impulse, Biophys. J. 79 (2000) 1821–1832. [45] B.D.M. Meijering, L.J.M. Juffermans, A. van Wamel, R.H. Henning, I.S. Zuhorn, M. Emmer, A.M.G. Versteilen, W.J. Paulus, W.H. van Gilst, K. Kooiman, N. de Jong, R.J.P. Musters, L.E. Deelman, K. Otto, Ultrasound and microbubbletargeted delivery of macromolecules is regulated by induction of endocytosis and pore formation, Circ. Res. 104 (2009) 679–687.