Original contribution
High Frequency Ultrasound Tissue Characterization and Acoustic Microscopy of Intracellular Changes

https://doi.org/10.1016/j.ultrasmedbio.2008.01.017Get rights and content

Abstract

The objective of this work is to investigate changes in the acoustic properties of cells when exposed to chemotherapy for monitoring treatment response. High frequency ultrasound spectroscopy (10-60 MHz) and scanning acoustic microscopy (0.9 GHz) were performed on HeLa cells (Ackermann et al 1954, Masters 2002) that were exposed to the chemotherapeutic agent cisplatin. Ultrasonic radio-frequency data were acquired from pellets containing HeLa cells after exposure to cisplatin to induce apoptosis. Scanning acoustic and laser fluorescence microscopy images were recorded from single HeLa cells exposed to the same drug. Data acquisition in both cases was performed at several time points throughout the chemotherapeutic treatment for up to 27 h. In the high frequency ultrasound investigation, normalized power spectra were calculated within a region-of-interest. A 20 MHz transducer (f-number 2.35) and a 40 MHz transducer (f-number 3) were used for the data collection in the high frequency ultrasound experiments. The backscatter coefficients, integrated backscatter coefficients, mid-band fit and spectral slope were computed as a function of treatment time to monitor acoustical property changes during apoptosis. Acoustic attenuation was measured using the spectral substitution technique at all time points. Spectral parameter changes were detected after 12 h of exposure and coincided with the initiation of cell damage as assessed by optical microscopy. Integrated backscatter coefficients increased by over 100% between 0 h and 24 h of treatment, with small changes in the associated attenuation (∼0.1 dB/[MHz cm]). Acoustic microscopy was performed at 0.9 GHz frequency. The cell structure was imaged using staining in laser fluorescence microscopy. All cells showed excellent correspondence between the locations of apoptotic nuclear condensation observed in optical imaging and changes in attenuation contrast in acoustic microscopy images. The time after drug exposure at which such changes occurred in the optical images were coincident with the time of changes detected in the acoustic microscopy images and the high frequency ultrasound experiments. (E-mail: [email protected])

Introduction

Within the last 50 years ultrasonic imaging has seen an increase in utilization for diagnostic and therapeutic purposes. Sound waves in the frequency range from 100 kHz to 15 MHz have been used in medical diagnostics as well as therapy. The most popular way of displaying backscatter information is B-mode imaging. This technique uses the envelope of detected ultrasound echoes from a region-of-interest (ROI) to create gray- scale images which display a cross-sectional mapping of the echo intensity. However, these images only use a fraction of the information contained in the signal, as the backscatter frequency and phase information is not used in conventional B-mode imaging.

Several investigators have suggested that the frequency dependent information in ultrasound echo signals can be related to acoustical and structural properties of tissue microstructure (Feleppa et al 1986, Feleppa et al 1986, Insana et al 1990, Lizzi et al 1988, Lizzi et al 1997a, Nicholas 1982). Spectral analysis of radio-frequency (RF) ultrasound signals can provide information about the tissue properties like the frequency dependent attenuation (Kuc 1985, Madsen et al 1999) the frequency dependence of the backscatter (Waters et al. 2003) and (using these two parameters) the effective scatterer size (Lizzi et al. 1986). In general, the frequency dependent attenuation is related to the bulk composition of the tissues, while the frequency dependence of the backscatter is related to the structure of the biological scatterers. Tissue characterization using ultrasound has been performed for in vivo prostate, testis (Feleppa et al 1996, Feleppa et al 1999, Feleppa et al 2000, Jenderka et al 1999, Lizzi et al. 1997b), liver (Lang et al 1994, Lu et al 1999, Lizzi et al 1997b, Oosterveld et al 1991), ophthalmologic tissue (Feleppa et al 1986, Feleppa et al 1986, Thijssen 1993, Verbeek et al 1994) and the myocardium (Dent et al 2000, Zuber et al 1999). Apart from characterizing tissue pathologies for diagnosis, ultrasonic tissue characterization has also been used for classifying the quality of meat, defined by the content of intramuscular fat (Brand et al 2002, Moerlein et al 2005). Most of these studies used ultrasonic backscatter signals to extract tissue acoustic properties and related them to specific pathologic alterations of the investigated specimen. Especially the group of Lizzi et al. (Lizzi et al 1983, Lizzi et al 1986, Lizzi et al 1997a, Lizzi et al 1997b) pioneered detailed theoretical formulations for extracting tissue acoustic information and the link to the underlying microstructure from backscattered ultrasonic signals.

A small number of previous investigations have examined the frequency dependence of ultrasound backscatter combined with its relation to the underlying tissue structure in vivo in the frequency range above 15 MHz. However, recent developments in transducer technology and electronics enabled the availability and application of ultrasound in the frequency range above 20 MHz (Cloutier and Qin 1997, Foster et al 2000, Silverman et al 1995, Silverman and Lizzi 2001, Ursea et al 1998). As the ultrasound wavelength approaches the size of the cells, it is expected that ultrasound based techniques are more sensitive to structural changes at a cellular level. Indeed, experimental data and effective scatterer size estimates derived from the frequency dependence of ultrasonic backscatter signals, using models developed by Lizzi and coworkers, indicate that the size of the dominant scatter source is on the order of a cell or its' nucleus (Lizzi et al 1983, Kolios et al 2002, Kolios et al 2003).

The interaction of ultrasound with biological tissue is determined by the mechanical properties of the underlying structure. In soft tissues, this structure is composed primarily of a cellular matrix containing single cells in an aggregate. Ultrasonic scattering is caused by structures below and up to the order of the ultrasound wavelength. Scattering effects that can be observed in the frequency range from 10 to 100 MHz may have their origin in structures at a subcellular level (Briggs 1995). Thus, investigating the mechanical properties on a single cell base will potentially provide insight in the mechanism of scattering from cellular ensembles, such as tumors (Kolios et al. 2002, Czarnota et al. 1997). Scanning acoustic microscopy (SAM) enables the investigation of viable cells in vitro on a microscopic scale offering the potential to explore mechanical properties. Using frequencies up to 2 GHz resolutions of 1 μm can be achieved (Briggs 1992). A detailed description of these methods can be found elsewhere (Briggs 1992, Briggs 1995, Kundu et al 2000, Kundu 2004). Ultrasonic properties that can be measured in these experiments include sound velocity, sound attenuation, cell thickness but also the elasticity and stiffness of intracellular components (Bereiter-Hahn et al 1995, Bereiter-Hahn and Luers 1998, Zoller et al 1997).

Two different approaches are used in SAM: time resolved SAM (Briggs et al 1993, Lemor et al 2003, Weiss et al 2007) and V(z) (Bereiter-Hahn and Luers 1998, Kundu et al 1992) or V(f) (Kundu et al. 2000) techniques. Time resolved acoustic microscopy applies a short broad band pulse and seems the most promising approach for quantitative measurements. However, the analysis requires the separation of acoustic pulses which potentially overlap (Briggs 1995). Another method known as reflection confocal acoustic microscopy (applying a monochromatic tone burst) enables the acquisition of V(z) or V(f) signatures. A comprehensive theoretical background and contrast theory is readily available (Briggs 1992). However, these techniques require data collection at multiple defocus positions. This creates problems when the transducer approaches the cell as it causes shear stress due to the scanning motion. Moreover, the acquisition of data at multiple depths significantly increases the data acquisition time. As opposed to light microscopy, in SAM, the contrast arises from the mechanical properties of the cellular substructure and potentially other physical effects (surface wave modes on the substrate, interference fringes due to modulation of the substrate echo by the cell surface echo and so on). The rich contrast produced by acoustic property variations negates the need for the use of specialized and potentially toxic dyes that are used in light microscopy.

Major changes in cell morphology occur after chemotherapeutic treatment. Apoptosis is a type of cell death that is induced by the cell itself. The cell undergoes significant morphological changes, such as nuclear condensation and fragmentation as well as membrane blebbing and overall cell shrinkage. A range of techniques have been developed to identify cells undergoing apoptosis using biochemical assays. These techniques are widely used to assess and identify apoptosis in cultured cells and in fixed tissues. However, there are very few reliable techniques for the noninvasive assessment of apoptosis in live un-sectioned tissues. Yet, this is critical for assessing the response of tumors to anticancer therapies, like chemo- or radiation treatment. Since multiple ultrasonic imaging sessions do not carry the increased risks that other modalities used for apoptosis monitoring do, such as positron emission tomography (PET) (Lahorte et al. 2004), scanning procedures can be performed rapidly for monitoring the treatment response over an extended period of time.

It was shown previously that intracellular changes resulting from exposure to the chemotherapeutic agent cisplatin can be detected using high frequency ultrasound (Czarnota et al 1997, Czarnota et al 1999, Kolios et al 2002). These studies used acute myeloid leukemia (AML) cells. Whereas individual cells cannot be resolved in the frequency range between 10 MHz and 60 MHz, an increase in the backscatter intensity and changes in the spectral parameters of AML cell ensembles were observed. It is not clear however whether the morphological changes clearly observed in optical microscopy at the single cell level correspond to changes in the cell acoustical properties. It is tempting to relate the large changes in apoptotic nuclear structure to the ultrasound backscatter changes, as our experiments to date suggest the nucleus as the main scattering source of cell ensembles (Taggart et al 2007, Kolios et al 2002, Kolios et al 2003). However, we have also shown theoretically that changes in the spatial organization of the scatterers can also induce the observed changes in ultrasound backscatter (Hunt 2002) and potentially changes in spectral slopes (Hunt, personal communication, 2004). The AML cell line used in previous studies is not an epithelial cell line that forms solid tumors but a hematologic malignancy. However, the ultrasonic techniques our group is developing are intended to image solid tumors. Moreover, as the AML cell line is not adherent, it cannot be used in acoustic microscopy studies.

To investigate changes in ultrasound backscatter of an epithelial cell line and, also, further probe changes in acoustical parameters at a single cell level applying acoustic microscopy, a cervix carcinoma cell line (HeLa) was used (Ackermann et al 1954, Masters 2002). Cell death was induced by exposure to a chemotherapeutic drug and cell pellets were scanned at several time points after treatment using high frequency ultrasound (20 and 40 MHz transducers). Spectral parameter estimation methods were applied to the radio-frequency (RF) ultrasound signals, acquired from pellets of HeLa cells. For further investigation on a cellular level, a novel device that combined optical and acoustical microscopy was used. Laser fluorescence and scanning acoustic microscopy imaging was performed on treated HeLa cells. The acoustic microscopy applied ultrasound at 0.9 GHz center frequency with a pixel spacing of 0.16 μm, thereby allowing us not only to detect changes in intracellular acoustical parameters but also to compare these changes with the corresponding variations observed in the optical images of the same cell.

Section snippets

Cell culture preparation

In the experiments using both high frequency ultrasound and acoustic microscopy HeLa cells obtained from ATCC (American Type Culture Collection, Manassas, VA, USA) were used. This is a rapidly growing human cervix carcinoma cell line and has been previously explored to investigate changes during cell death (Maldonado et al. 1995). The chemotherapeutic agent cisplatin was used for inducing apoptosis in HeLa cells. HeLa cells exposed to cisplatin undergo cell death, presenting morphological and

High frequency ultrasound experiments

In Fig. 2, the normalized power spectra of backscattered ultrasound from HeLa cells undergoing cell death are shown for three experiment time points (pretreatment, 18 h, 24 h). Plotted are the results obtained from measurements with the 20 MHz and 40 MHz transducer. It can be seen from Fig. 2 that normalized power spectra overlap and show acceptable trend continuity for the two transducers at each time point, despite the different transducer geometries and attention path length through the

High frequency ultrasound experiments

In ultrasound spectroscopy, increases in the backscattered intensity from HeLa cells were observed after exposure to cisplatin. Similar results were obtained from ultrasonic imaging of acute myeloid leukemia (AML) (Czarnota et al 1997, Kolios et al 2002) and epithelial cells (HEp-2) (Brand et al. 2006) during treatment. These frequency dependent increases were 10 dB for AML (Kolios et al. 2002) and 7 dB for HEp-2 (Brand et al. 2006) cells. Similar increases in backscatter were measured for the

Acknowledgments

The authors acknowledge the financial support of the Whitaker Foundation (grants RG-01-0141 and transitional funds), the Natural Sciences and Engineering Research Council (NSERC, CHRP grant 237962–2000) and the Ontario Premier's Research Excellence Awards (PREA 00/5–0730). The VisualSonics ultrasound bio microscope was purchased with the financial support of the Canada Foundation for Innovation, the Ontario Innovation Trust and Ryerson University. The authors also thank Arthur Worthington,

References (59)

  • F.L. Lizzi et al.

    Comparison of theoretical scattering results and ultrasonic data from clinical liver examinations

    Ultrasound Med Biol

    (1988)
  • F.L. Lizzi et al.

    Statistical framework for ultrasonic spectral parameter imaging

    Ultrasound Med Biol

    (1997)
  • Z.F. Lu et al.

    Ultrasound backscatter and attenuation in human liver with diffuse disease

    Ultrasound Med Biol

    (1999)
  • R.H. Silverman et al.

    Three-dimensional high-frequency ultrasonic parameter imaging of anterior segment pathology

    Ophthalmology

    (1995)
  • L. Taggart et al.

    Ultrasonic characterization of whole cells and isolated nuclei

    Ultrasound Med Biol

    (2007)
  • J.M. Thijssen

    The history of ultrasound techniques in ophthalmology

    Ultrasound Med Biol

    (1993)
  • D.H. Turnbull et al.

    Ultrasonic characterization of selected renal tissues

    Ultrasound Med Biol

    (1989)
  • R. Ursea et al.

    Correlation of high-frequency ultrasound backscatter with tumor microstructure in iris melanoma

    Ophthalmology

    (1998)
  • K.R. Waters et al.

    Parametric analysis of carotid plaque using a clinical ultrasound imaging system

    Ultrasound Med Biol

    (2003)
  • E.C. Weiss et al.

    Imaging of focal contacts of chicken heart muscle cells by high-frequency acoustic microscopy

    Ultrasound Med Biol

    (2007)
  • A.E. Worthington et al.

    Changes in ultrasound properties of porcine kidney tissue during heating

    Ultrasound Med Biol

    (2001)
  • M. Zuber et al.

    Myocardial tissue characterization in heart failure by real-time integrated backscatter

    Eur J Ultrasound

    (1999)
  • W.W. Ackermann et al.

    Growth characteristics of poliomyelitis virus in HeLa cell cultures – Lack of parallelism in cellular injury and virus increase

    J Exp Med

    (1954)
  • J. Bereiter-Hahn et al.

    Mechanical basis of cell shape: Investigations with the scanning acoustic microscope

    Biochem Cell Biol

    (1995)
  • J. Bereiter-Hahn et al.

    Subcellular tension fields and mechanical resistance of the lamella front related to the direction of locomotion

    Cell Biochem Biophys

    (1998)
  • S. Brand et al.

    Estimation of intramuscular fat content (IMF) in the musculus longissimus dorsi (LD) of pigs by analysis of RF echo signals obtained from a clinical B-mode device

  • S. Brand et al.

    Ultrasonic assessment of death in HEp2 cells using spectral and wavelet based analysis of backscattered RF-signals

  • A. Briggs

    Acoustic microscopy, in monographs on the physics and chemistry of materials

    (1992)
  • A. Briggs

    Advances in acoustic microscopy

    (1995)
  • Cited by (48)

    • Acoustic Histology with Specific Dyes and Antibodies

      2019, Ultrasound in Medicine and Biology
    • Automatic Cataract Hardness Classification Ex Vivo by Ultrasound Techniques

      2016, Ultrasound in Medicine and Biology
      Citation Excerpt :

      The most common representation for the backscattering signals is the B-Scan image (Szabo 2014). This technique uses the envelope of the ultrasound echoes from a region of interest (ROI) to create gray-scale images as an intensity mapping (Brand et al. 2008; Huang et al. 2007a; 2007b; Paunksnis et al. 2007; Raitelaitienė and Paunksnis 2005; Tabandeh et al. 2000; Tsui and Chang 2007; Tsui et al. 2007; 2008; 2011). However, due to the processing complexity, B-Scan images commonly used in medical diagnosis contain less information than the original signals (Brand et al. 2008; Despotovic et al. 2008).

    • Quantification of ultrasonic scattering properties of in vivo tumor cell death in mouse models of breast cancer

      2015, Translational Oncology
      Citation Excerpt :

      The first high-frequency QUS studies of cell death in cancer cells induced by a chemotherapeutic agent demonstrated that apoptosis causes an increase in ultrasound echogenicity [18,19]. These ultrasound changes were later quantified by system-independent parameters including midband-fit (MBF) and spectral slope (SS), both of which demonstrated an increase as a result of cell death [20]. This observation was experimentally determined to be related to the nuclear condensation and fragmentation that occur during apoptosis.

    View all citing articles on Scopus

    Video Clips cited in this article can be found online at: http://www.umbjournal.org.

    View full text