1.

Introduction

Cells exposed to heat stress undergo multiple molecular alterations, including synthesis and augmented production of heat-shock proteins, protein misfolding or damaging, and energy depletion. These initial events possibly trigger complex interactions of metabolic mechanisms and signal transductions that determine cellular stress responses. Depending on the heat stress dose, comprising temperature and exposure time, the cell species, and the tissue architecture and environment, the stress responses lead to cell survival, apoptosis, or necrosis. In this context, heat stress represents a large array of various physical, biochemical, toxicological, and environmental stress factors.^{1, 2, 3} In the field of clinical cancer therapy, two major groups comprise thermotherapy: laser-induced thermotherapy inducing local temperatures between 60 and $100°\mathrm{C}$ ,^{4, 5} and hyperthermia working with temperatures below $45°\mathrm{C}$ .⁶ Combined modality treatments such as hyperthermia-assisted radiotherapy and/or chemotherapy are under intensive investigation.^{7, 8} A new thermotherapeutic approach has been published by Jordan that suggests injection of magnetic nanoparticles into the tumor, followed by application of alternating current magnetic fields, inducing a temperature elevation in the particle-loaded tissue.⁹

Dosimetry remains a critical factor in thermotherapy, since therapy responses might vary among different tissue species, tissue volumina, tissue entities, and tissue environments.¹⁰ With regard to dosimetry in medical thermotherapeutic applications, adequate knowledge about cellular and subcellular reactions to thermal stress conditions is essential. The molecular genetics of heat stress responses already have been studied extensively and reviewed by Takayama, Reed, and Homma,¹¹ and Moggs and Orphanides.¹² Representative examples of cellular heat-induced phenotypes using an Arrhenius-based approach have been studied in vitro by Dewhirst ¹⁰ and in vivo by Minet and Beuthan.¹³

In the present study, we investigated the stress responses of breast cancer cells MX1 under hyperthermal $(40°\mathrm{C})$ and heat stress (45, 50, or $56°\mathrm{C}$ ) conditions. Several different subcellular sites were targeted to characterize their impact on heat-induced phenomena.

The most versatile approach for probing stress-induced effects in cells is fluorescence microscopy, as it is one of the best established standard techniques in biomedical analytics and diagnostics. Fluorescence detection of specific structures and functions in cells is supported by an enormous diversity of commercial fluorescent probes and biosensors mainly tailored for specific targets as well as spectral properties.^{14, 15} Many organic fluorescent dyes, immunoconjugates, and GFP-based biosensors have been routinely used for imaging cells; recently, several bioactive nanoparticles have entered the field of experimental bioanalytics. Most interesting among those are magneto/optical nanoparticles,¹⁶ as well as luminescent quantum dots (Qdots).^{17, 18} Various surface modifications and conjugation strategies have been developed for improving stabilities and bioactivities of Qdots without altering their unique spectral properties resulting from semiconductor effects.^{19, 20, 21, 22} The bright and narrow-banded fluorescence emissions combined with broad excitation ranges and high absorptions of Qdots especially are relevant features concerning bioanalytical applications. Qdot fluorescence colors in the visible spectrum depend on the semiconductor material and mainly on particle diameters roughly ranging from $2\text{to}6\mathrm{nm}$ . The broad spectral variability of Qdots also makes them very interesting tools in many kinds of multiplexed labeling strategies.^{23, 24}

As shown previously, MX1 cell viability did not decrease after hyperthermal treatment at 40 or $42°\mathrm{C}$ , but after heat stress at temperatures of 45, 50, or $56°\mathrm{C}$ , cell viabilities had drastically diminished.²⁵ In this study the metabolic data were supplemented by molecular fluorescence imaging and high-resolution structural imaging. The following targets were compared using a defined heat stress protocol: plasma membrane, actin cytoskeleton, and attachment filaments, as shown in Fig. 1 . The microscopic techniques listed in Table 1 were used depending on the particular target.

Fig. 1

Schematic view of a MX1 breast cancer cell showing the investigation targets of this study.

Table 1

Microscopes and measurement configurations used for imaging various targets of heat-induced stress response in MX1 cancer cells. LP is a long-pass filter, BP is a bandpass filter, bw is black and white, and C is for color.

It was shown by wide-field fluorescence microscopy (WFFM) and laser scanning microscopy (LSM), respectively, that the actin cytoskeleton in MX1 cells already responded at comparatively mild stress conditions of $40°\mathrm{C}$ $\left(30\min \right)$ , while cell morphologies and plasma membrane integrities were not affected. Under increased stress temperatures, rearrangements of actin filaments were enhanced and accompanied by striking changes of cell shapes. Inline with cytoskeletal destabilization, the surface substructures increase with temperatures as visualized by atomic force microscopy (AFM). On severe heat stress at $56°\mathrm{C}$ $\left(30\min \right)$ , actin cytoskeletons as well as plasma membranes were completely destroyed, resulting in cell necrosis. Treatment at $50°\mathrm{C}$ for $30\min$ was not sufficient for induction of necrosis, but the cells exhibited altered morphological features as a result of active stress response, as shown by fluorescence microscopy and scanning electron microscopy (SEM).

2.

Material and Methods

All chemicals were purchased from Sigma-Aldrich Chemie GmbH, Germany, unless stated otherwise.

3.

Results

3.2.

Fluorescence Detection of F-Actin Cytoskeletons

Heat-induced reactions of F-actin cytoskeletons in MX1 cells were imaged by green fluorescence of Alexa Fluor® 488 conjugated to F-actin-bound phalloidin. Under increasing heat stress conditions, the F-actin components exhibited drastic structural alterations, as shown in Fig. 3 . In control cells [Fig. 3(a)], actin exhibited long filamentous structures forming intact networks within the cell bodies. Intact contacts with F-actin in plasma membranes were stabilizing the cell morphologies and intercellular connections. After treatment at $40°\mathrm{C}$ , F-actin networks already were destabilized, as evidenced by diminished filamentous structures; however, plasma membranes and intercellular contacts had not yet been affected. Corresponding with the SEM image [Fig. 2(c)], heat stress at $45°\mathrm{C}$ induced loosening of tissue as well as advanced disintegration of cytoskeletons, leading to shrinkage and rounding of cells. F-actin structures in cells stressed at $50°\mathrm{C}$ contained even more defects, where it was observed that the actin particles clustered in the cytoplasm and at membranes. A total breakdown of tissue cohesion and cytoskeletal structures resulted from heat stress at $56°\mathrm{C}$ .

Fig. 3

Fluorescence imaging of cytoskeletal F-actin in MX1 cells treated at various heat stress conditions as indicated. The fluorescence label is Alexa Fluor® 488 conjugated to actin-bound phalloidin (Molecular Probes, Netherlands): (a) control cells, (b) cells stressed at $40°\mathrm{C}$ , (c) cells stressed at $45°\mathrm{C}$ , (d) cells stressed at $50°\mathrm{C}$ , and (e) cells stressed at $56°\mathrm{C}$ ( $30\min$ each). Excitation range: BP $470\pm 20\mathrm{nm}$ , detection range: $\mathrm{LP}>520\mathrm{nm}$ . Details of the measurement setup are given in Table 1. The scale bar is $20\mu \mathrm{m}$ .

3.3.

Fluorescence Detection of Qdot-Labeled Plasma Membranes

In control cells and $40°\mathrm{C}$ -stressed cells, Qdot fluorescence was localized at the superficial areas of plasma membranes and intercellular connections [Figs. 4(a) and 4(b) ]. Fluorescence intensities were irregularly distributed along the plasma membranes, indicating variant compositions and frequencies of carbohydrate residues in the glycocalix. The intensities were higher in Fig. 4(b) than in Fig. 4(a) because of enhanced ConA binding. After stressing at 45 or $50°\mathrm{C}$ [Figs. 4(c) and 4(d)], cells again were smaller and rounder than in the control group [Fig. 4(a)]. The 3-D intensity profiles exhibited more regular fluorescence distributions than under milder stress conditions at $40°\mathrm{C}$ . Qdot fluorescence at the same time was also detected at the cytoplasmic leaflets of plasma membranes, supporting the presumption that plasma membranes were reorganized, permitting Qdot conjugates to accumulate inside plasma membranes. Nuclei in the cells of the $50°\mathrm{C}$ -stress group were intact to the extent that they did not exhibit fluorescence accumulation. Cells were severely damaged at $56°\mathrm{C}$ , presenting necrotic phenotypes because the structural integrities of plasma membranes and nuclear membranes were destroyed. Now Qdot fluorescence accumulated inside the cells [Fig. 4(e)].

Fig. 4

Fluorescence imaging of plasma membranes in MX1 cells labeled with Qdots (Quantum Dot Corporation, USA). The fluorescence distributions are imaged as 3-D signal intensity profiles: (a) control cells, (b) cells stressed at $40°\mathrm{C}$ , (c) cells stressed at $45°\mathrm{C}$ , (d) cells stressed at $50°\mathrm{C}$ , and (e) cells stressed at $56°\mathrm{C}$ ( $30\min$ each). Details of the measurement setup are given in Table 1.

3.4.

AFM Imaging of Plasma Membrane Topographies

AFM measurements focusing on the area roughness of cell surfaces were performed in dynamic force operation mode. The membrane surfaces of control cells were more or less regularly undulated without showing indentation or incisions [Fig. 5(a) ]. After treatment at $40°\mathrm{C}$ , the topography in Fig. 5(b) exhibited a more jagged and less regular structure than the control cell. The surface roughness was further increased after $45°\mathrm{C}$ heat stress, since globular protruding substructures led to spot-like cavities, as shown in Fig. 5(c). This effect was enhanced after treating cells at $50°\mathrm{C}$ [Fig. 5(d)], where the cavities were more stretched like clefts. Heat shock at $56°\mathrm{C}$ actually did not change the surface structures of the cells [Fig. 5(e)] compared to the $50°\mathrm{C}$ -treated cells, but the sizes of cavities were increased, appearing like holes. The area roughness of the cell shown in Fig. 5(e) was increased. The development of this temperature-dependent effect is represented by a quantitative evaluation of the mean surface area roughness $\left(S_a\right)$ illustrated in Fig. 6 . $S_a$ values were measured for the topographies shown in Fig. 5 using the following equation:^{28, 29}

$$S_a=\frac{1}{MN}\sum _{k=0}^{M-1}\sum _{l=0}^{N-1}\mid z(x_k,y_l)\mid .$$

Fig. 5

AFM measurements of heat stressed cell topographies. The xy range is $1.5\times 1.5\mu \mathrm{m}$ and the dynamic Z range is $150\text{to}153\mathrm{nm}$ in every image. (a) Control cell, (b) cell stressed at $40°\mathrm{C}$ , (c) cell stressed at $45°\mathrm{C}$ , (d) cell stressed at $50°\mathrm{C}$ , and (e) cell stressed at $56°\mathrm{C}$ ( $30\min$ each). All panels include the 3-D profile (left) of the measured surface area, a cross section (center), and the height image (right) with the line indication (arrow head) of the cross section. Details of the measurement setup are given in Table 1.

Fig. 6

Evaluations of the surface area roughness of MX1 cells treated at various heat stress conditions as indicated. The topographic measurements are shown in Fig. 5.

4.

Discussion

Heat as the classical form of stress factor induces manifold molecular changes in animal cells, including the synthesis of heat-shock proteins, leading to a wide variety of possible heat-shock responses. Although molecular genetics of stress responses are highly conserved among living organisms, the individual outcome (survival, apoptosis, or necrosis) always depends on many molecular, physiological, biochemical, and environmental situations and impacts. Different organisms, different cells species, even different subcellular components and molecules may exhibit varying heat sensitivities that might be correlated with diseased conditions.^{1, 2, 3} Therefore, the (patho)physiology of cellular heat stress responses is an extensive area of basic research.

We investigated the heat stress response of breast cancer cells MX1, analyzing various morphological and subcellular features requiring the application of different imaging techniques listed in Table 1. Our data supported a differentiation between continuous and discontinuous heat-induced effects on specific targets in MX1 cancer cells, which were evaluated in Table 2 . The destabilization of actin cytoskeletons and the increase of cellular surface roughness were continuously enhanced by increasing temperatures. On the other hand, the adhesion and contact behavior of cells, cell shapes, and plasma membrane integrities responded in discontinuous manners. While alterations of the cell morphologies obviously initiated at a threshold temperature of approximately $45°\mathrm{C}$ , the plasma membrane integrities were not demonstrably affected until approximately $56°\mathrm{C}$ as a consequence of the selective melting behavior of biological membranes.³⁰ Hence, active stress response and cell necrosis should be classified as discontinuous heat-induced processes.

Table 2

Evaluation of temperature-dependent increase of stress-induced alterations in cellular target structures referred to as enhancement of effect.

The most thermosensitive structure analyzed here was the actin cytoskeleton that responded to a slightly elevated temperature of $40°\mathrm{C}$ . While cell morphology, adhesion behavior, and plasma membrane integrity were not detectably affected at this stress temperature, the surface roughness increased as a consequence of cytoskeletal reorganization. In erythrocytes, actin monomers have been shown to be thermally more labile than filamentous actin.³¹ Similar observations were made in microglial cells when heat-induced appearance of membrane ruffles were suggested to result from cytoskeletal reorganization involving actin as the most heat-responsive component.³²

In our previous study, we demonstrated that mild heat stress at $40°\mathrm{C}$ leads to a transient metabolic activation of MX1 cells, resulting in enhanced cell viability.²⁵ At higher temperatures (45 and $50°\mathrm{C}$ ), destabilization of actin filaments increased, accompanied by drastic shrinkage and rounding of cell shapes. This implies that dissolution of solid tumor tissue might be induced by heat at $45°\mathrm{C}$ or higher, but complete tumor destruction might not occur, since effective devitalization of MX1 cells was only achieved after treatment at $56°\mathrm{C}$ . In medical thermotherapy, the escape of vital cells from the tumor mass should be avoided to minimize the risk of metastases. Optimal thermotherapy protocols would therefore have to be developed to realize therapy that would define effective temperatures and temperature distributions in the target tissues without damaging adjacent healthy structures. Also, the $S_a$ data derived from the AFM measurements in Fig. 5 denote a temperature-dependent growth of surface roughness (Fig. 6), although this phenomenon was not seen in SEM images in Fig. 2 when a pan-cellular magnification was selected. The morphological changes of cells under thermal stress generally corresponded well when the differently prepared samples imaged by different microscopic techniques were compared and did not result from preparative artefacts.

Fig. 2

Scanning electron microscopy of MX1 cells treated at various heat stress conditions as indicated: (a) control cells, (b) cells stressed at $40°\mathrm{C}$ , (c) cells stressed at $45°\mathrm{C}$ , (d) cells stressed at $50°\mathrm{C}$ , and (e) cells stressed at $56°\mathrm{C}$ ( $30\min$ each). Details of the measurement setup are given in Table 1.

Destructive perforation of plasma membranes occurred after treatment at $56°\mathrm{C}$ ,²⁵ as demonstrated by SEM imaging [Fig. 2(e)] and Qdot fluorescence [Fig. 4(e)] for whole cells. A detail study measured by AFM in Fig. 5(e) matched the results previously stated.

While SEM in our study was the most suitable approach for morphological and structural characterization of heat-stressed cells, the AFM provided detailed imaging of subcellular topographies in the dynamic force mode. Although SEM and AFM offer high-resolution imaging, both techniques are restricted to cell surfaces or surfaces of isolated structures, and both lack target specificity. Hence, correct interpretations of SEM or AFM imaged structures always require profound knowledge and experience. These shortcomings in principle are by-passed using fluorescence microscopy applicable to a wide variety of specific labels sensing superficial or intracellular targets.^{14, 15, 17} The limited lateral resolution of conventional WFFM is reduced by confocal LSM, offering a greater resolution by the factor $\sqrt{2}$ if the confocal detector pinhole is smaller than the Airy disk produced by a fluorescent point.³³ The best lateral resolutions so far can be achieved by near-field microscopy, which includes AFM technologies. A combination of near-field optics and fluorescence detection has been realized with scanning near-field optical microscopy (SNOM). The SNOM technique offers high-resolution imaging of topographies and site-specific detection of fluorescence signals simultaneously, making localization of single sensor molecules possible.^{34, 35, 36} Several studies have demonstrated that supplementing AFM measurements with fluorescence images by any means is a common experimental approach in cytology when a SNOM is not available.^{37, 38, 39, 40} In SNOM investigations, Qdots should be especially useful as fluorescent labels, since they emit comparatively high signal intensities, exhibit photostability for several minutes under continuous illumination, and their broad absorption spectrum is not restricted to specific excitation sources.⁴¹

The long-term photostability and high emission intensity of luminescent Qdots especially support the tracking of dynamic processes in living cells, like receptor-mediated transmembrane signal transduction⁴² and the lateral diffusion dynamics of glycine receptors in neuronal cells.⁴³

5.

Conclusion

In MX1 breast cancer cells, differential thermosensitivities of subcellular structures are detected in the heat-stress range $40\text{to}56°\mathrm{C}$ . Heat-induced reactions are shown to increase continuously on the fine-structural level concerning the actin cytoskeleton and the cell surface roughness resulting from molecular and supra-molecular alterations. On the other hand, the cell morphologies and the tissue architecture are observed to change discontinuously under the same heat-stress regimen. We therefore conclude that further investigations on different cell types under various heat-stress conditions will support the development of optimized thermotherapy protocols.